Adeno-associated virus (AAV) and lentivirus are the two most widely used viral vectors in gene therapy, and they differ in almost every dimension that matters: how they deliver genetic material, whether they permanently alter a cell’s DNA, what size of gene they can carry, which tissues they reach, and what safety risks they pose. AAV vectors dominate in vivo gene therapy, where the treatment is injected directly into a patient’s body, while lentiviral vectors are the workhorses of ex vivo therapy, where a patient’s cells are removed, modified in the lab, and returned. Choosing between them is less about which is “better” and more about which fits a particular disease, target tissue, and treatment strategy.
What Each Vector Actually Is
AAV is derived from a tiny, non-pathogenic virus that naturally infects humans but has never been linked to any disease. Its genome is small, consisting of single-stranded DNA, and the engineered vectors used in therapy have had all viral genes stripped out, leaving only the DNA packaging signals and whatever therapeutic gene researchers want to deliver. Lentiviral vectors, by contrast, are built from the backbone of HIV-1, which understandably raises eyebrows. But decades of engineering have removed the genes responsible for viral replication and disease, producing vectors that retain HIV’s remarkable ability to enter cells and integrate DNA into chromosomes while being incapable of causing infection.1PubMed Central. Lentiviral Vectors: From Wild-Type Viruses to Efficient Multi-Functional Delivery Vectors Both platforms allow stable, long-term expression of a therapeutic gene in cells that do not divide, such as neurons.2PubMed Central. Lentiviral Vectors and Adeno-Associated Virus Vectors: Useful Tools for Gene Transfer in Pain Research
Integration Versus Episomal Persistence
This is arguably the single biggest functional difference. Lentiviral vectors integrate their genetic cargo directly into the host cell’s chromosomes. That means the therapeutic gene becomes a permanent part of the cell’s DNA and is faithfully copied every time the cell divides. For therapies targeting blood stem cells, which divide frequently and replenish the entire blood system, integration is essential: without it, the corrected gene would be diluted away within weeks.3PubMed Central. Ex Vivo Gene and Cell Therapy in Hematopoietic Stem Cells
AAV vectors work differently. The therapeutic DNA they deliver mostly persists as circular structures called episomes, which float inside the cell nucleus without stitching themselves into the chromosomes. In cells that rarely or never divide, like those in the brain, eye, or liver, these episomes can stick around for years, providing durable gene expression. But in dividing cells, episomes get lost over time as cells split and the extra DNA is not replicated along with the chromosomes.4ScienceDirect. Adeno-associated viral vector integration: implications for long-term efficacy and safety This is why AAV is a natural fit for tissues with low cell turnover but a poor choice when you need the gene to survive in rapidly dividing cell populations.
AAV can occasionally integrate into the host genome at low rates. Whether that rare integration poses a meaningful cancer risk in humans has been debated for years, and the evidence so far is mixed. The theoretical concern is that viral DNA landing near an oncogene could switch it on, but this remains far less common with AAV than with vectors that integrate by design.5ScienceDirect. Adeno-associated viral vector integration: implications for long-term efficacy and safety
Cargo Capacity
AAV has a strict size limit. Its capsid (the protein shell that packages the DNA) can hold roughly 4.7 kilobases of genetic material, including the regulatory sequences needed to drive expression. Many human genes fit comfortably within that space, but some important ones do not. The dystrophin gene mutated in Duchenne muscular dystrophy, for instance, is far too large, forcing researchers to use truncated “mini” or “micro” versions that deliver partial function. Lentiviral vectors are more generous, accommodating inserts of about 8 to 10 kilobases. That extra room means larger genes, more complex regulatory elements, or even multiple genes can be packed into a single vector. For diseases caused by large gene defects, lentiviral vectors sidestep a packaging problem that AAV simply cannot solve without workarounds.
How They Reach Different Tissues
AAV comes in a wide family of natural serotypes, each with a different protein coat that determines which cell types it preferentially enters. AAV9, for example, crosses the blood-brain barrier and reaches motor neurons, which is why it is used in the spinal muscular atrophy drug onasemnogene abeparvovec. AAV8 homes to the liver. A comprehensive survey of ten naturally occurring serotypes found distinct tropism profiles across organs in mice, with some serotypes efficiently transducing muscle, others the central nervous system, and others glandular tissue.6Molecular Therapy. Comprehensive survey of natural AAV serotype tropism in mice This built-in diversity gives researchers a menu of options when picking a vector for a specific organ.
Lentiviral vectors lack that natural tissue-targeting variety, but they compensate through a trick called pseudotyping. By swapping the envelope protein on the vector’s surface with one borrowed from a different virus, researchers can redirect where the lentiviral particle goes.7PubMed Central. Pseudotyping Lentiviral Vectors: When the Clothes Make the Virus The most common envelope used is VSV-G, from vesicular stomatitis virus, which gives the vector an extremely broad host range and the ability to enter almost any mammalian cell type. For more targeted delivery, other envelopes can narrow the range. Vectors pseudotyped with rabies-derived glycoproteins exploit that virus’s natural tropism for the central nervous system.8PubMed Central. Altering the tropism of lentiviral vectors through pseudotyping Experiments comparing pseudotypes in the mouse brain found that Mokola virus envelope was actually the most efficient at transducing brain structures like the striatum and thalamus, while Ebola virus envelope produced no central nervous system transduction at all.9Molecular Therapy. Targeted Transduction Patterns in the Mouse Brain by Lentivirus Vectors Pseudotyped with VSV, Ebola, Mokola, LCMV, or MuLV Envelope Proteins The flexibility is impressive, though it adds complexity to manufacturing.
Immune Responses and Re-Dosing
Both vectors trigger immune responses, but the nature and consequences differ. AAV vectors are generally considered to have low immunogenicity, yet a substantial fraction of the human population already carries antibodies against common AAV serotypes from natural childhood exposure. Even very low levels of these pre-existing antibodies can neutralize the vector before it reaches target cells, blocking the therapy entirely.10PubMed Central. Pre-existing anti-adeno-associated virus antibodies as a challenge in AAV gene therapy This means patients must be screened for antibodies before treatment, and some are excluded from AAV-based trials altogether. After a first dose, the immune system mounts a robust response that makes re-dosing with the same serotype essentially impossible with current technology. That is a serious limitation for therapies that might need a booster years later as episomal expression wanes.
Lentiviral vectors also provoke immune activation. Their entry into cells and the process of reverse transcription trigger innate immune pathways through Toll-like receptors, which are part of the body’s early alarm system.11Journal of Virology. HIV-1 lentiviral vector immunogenicity is mediated by Toll-like receptor 3 (TLR3) and TLR7 However, because lentiviral vectors are most often used ex vivo, the immune challenge plays out differently. Cells are modified in the lab and washed before being returned to the patient, so the patient’s immune system never encounters the vector particles directly. In the rare cases where lentiviral vectors have been delivered systemically (directly into the bloodstream), studies in mice and monkeys found that a second injection was completely inactivated by the immune response to the first, even when months had passed between doses.12Molecular Therapy. Systemic Delivery of Lentiviral Vectors in Neonatal and Adult Mice and Rhesus Monkeys: Dosing Schedule, Biodistribution, and Vector Inactivation Re-dosing, then, is a challenge for both platforms when vectors are delivered in vivo.
Safety Profiles and Toxicity Concerns
Each vector carries its own set of risks, and neither is risk-free. For AAV, the biggest safety concern in recent years has been severe toxicity at high systemic doses. Despite remarkable success in treating spinal muscular atrophy, fatalities have occurred from liver, kidney, heart, or lung failure after high-dose intravenous AAV therapy.13PubMed Central. Lethal immunotoxicity in high-dose systemic AAV therapy The liver appears to be especially vulnerable because it acts as a sink for systemically delivered AAV, absorbing enormous quantities of vector. Innate and adaptive immune responses to the vector capsid play a critical role in this toxicity, and complications like thrombotic microangiopathy, a dangerous blood-clotting disorder, have been observed in some patients. Anti-complement therapy and hemodialysis have been used to manage these events.14Molecular Therapy. Lethal immunotoxicity in high-dose systemic AAV gene therapy These risks are dose-dependent and have pushed the field toward engineering capsids that transduce target tissues more efficiently, so lower doses can be used.
For lentiviral vectors, the primary safety worry is insertional mutagenesis, the possibility that the vector integrates into or near a cancer-related gene and switches it on. Early gene therapy trials using older-generation retroviral vectors (closely related to lentiviral vectors) caused leukemia in several patients, which cast a long shadow over the field. Modern self-inactivating (SIN) lentiviral vectors were designed to address this. They lack the strong enhancer sequences that drove oncogene activation in earlier designs. In vivo genotoxicity assays have shown that SIN vectors are substantially less genotoxic than their predecessors, though residual risk remains: the internal promoter driving the therapeutic gene can still act as a weak enhancer, and the strength of that promoter correlates with the likelihood of activating nearby oncogenes.15PubMed Central. Uncovering and dissecting the genotoxicity of self-inactivating lentiviral vectors in vivo Using moderate-strength promoters and adding insulator elements reduces this risk further, though it does not eliminate it entirely.
Where Each Vector Dominates in the Clinic
AAV vectors are the platform of choice for in vivo gene therapy targeting organs that are difficult or impossible to access ex vivo. Approved products use AAV to treat spinal muscular atrophy (delivered intravenously to reach motor neurons), inherited retinal dystrophy (injected under the retina), and hemophilia (targeting the liver). The platform’s strengths in this space include high transduction efficiency in non-dividing cells, tissue-specific targeting through serotype selection, and an established clinical safety profile at appropriate doses.16PubMed Central. In vivo applications and toxicities of AAV-based gene therapies in rare diseases Most AAV gene therapies in development target monogenic rare diseases, where replacing a single defective gene is the therapeutic goal.
Lentiviral vectors have carved out a complementary niche. They power the ex vivo modification of blood stem cells for conditions like sickle cell disease, beta-thalassemia, and severe combined immunodeficiency. In these treatments, a patient’s stem cells are harvested, transduced with a lentiviral vector carrying the corrective gene, and then re-infused after the patient receives conditioning chemotherapy to make room for the modified cells. Because the vector integrates, every daughter cell that arises from a successfully modified stem cell inherits the therapeutic gene. Lentiviral vectors are also the backbone of most CAR-T cell therapies, where a patient’s T cells are engineered to recognize and kill cancer cells. The durability of lentiviral integration is what makes these living drugs possible.17PubMed Central. Ex Vivo Gene and Cell Therapy in Hematopoietic Stem Cells
Manufacturing and Cost
Both vectors are notoriously expensive to produce, but the manufacturing challenges differ. AAV production generates a mixture of “full” capsids (containing the therapeutic DNA) and “empty” capsids (protein shells with no DNA inside). Empty capsids are not therapeutic and may worsen immune responses, so separating them from full capsids is a critical and difficult step. The two types of particle are nearly identical in size and surface charge, making chromatographic separation a painstaking optimization problem.18PubMed. Rational downstream development for adeno-associated virus full/empty capsid separation – A streamlined methodology based on high-throughput screening and mechanistic modeling Novel approaches like ultrafiltration with track-etched membranes have shown promise, achieving high recovery of full capsids by exploiting the subtle size difference between the two particle types.19PubMed Central. A Novel Method for Separating Full and Empty Adeno-Associated Viral Capsids Using Ultrafiltration Multi-column chromatography systems have also dramatically improved yield and throughput compared to single-column methods.20PubMed. Separation of empty and full adeno-associated viral capsids by MCSGP: From model-based design to implementation
Lentiviral vector manufacturing faces a different set of headaches. The particles are physically fragile compared to AAV, losing infectivity quickly during processing. Rapid, high-resolution purification is needed to preserve functional titers, and the process also generates non-functional byproducts like empty particles and extracellular vesicles that must be removed.21PubMed Central. Challenges and Opportunities in Lentivirus Viral Vector Manufacturing for In Vivo Applications The instability of lentiviral particles makes scaling up production harder than for AAV, which is one reason lentiviral therapies have largely stayed in the ex vivo space, where smaller vector quantities are needed compared to systemic in vivo dosing. Novel bioreactor platforms are being explored to reduce manufacturing costs for both vector types, with some showing cost-of-goods reductions of roughly 20 to 50 percent compared to conventional stirred-tank systems.
Next-Generation Engineering
Neither platform is standing still. The AAV field is investing heavily in capsid engineering to overcome its two most pressing limitations: pre-existing immunity and the need for high doses. Rational design, directed evolution, and machine learning are being combined to create synthetic capsids that evade neutralizing antibodies, transduce target tissues more efficiently, and avoid off-target organs like the liver.22PubMed Central. Advances in AAV capsid engineering: Integrating rational design, directed evolution and machine learning One recent effort used directed evolution to develop AAV capsids with enhanced delivery to Schwann cells in peripheral nerves, achieving not only higher transduction but also greater resistance to antibody neutralization and reduced liver accumulation.23Molecular Therapy Nucleic Acids. Directed evolution of novel AAV capsids for enhanced delivery to mouse and human Schwann cells
On the lentiviral side, a major innovation is the development of non-integrating lentiviral vectors (NILVs). These retain the cell-entry machinery and cargo capacity of standard lentiviral vectors but carry a mutated integrase enzyme that prevents chromosomal insertion. The therapeutic DNA instead persists as episomes, similar to AAV. This reduces the risk of insertional mutagenesis while preserving the ability to deliver large genes.24PubMed Central. Non-Integrating Lentiviral Vectors in Clinical Applications: A Glance Through NILVs enable sustained expression in non-dividing cells and open the door to using lentiviral vectors for in vivo applications where integration would be an unacceptable risk.25PubMed Central. Integrase deficient lentiviral vector: prospects for safe clinical applications It is an interesting convergence: AAV engineers are trying to make their episomal vector more durable and tissue-specific, while lentiviral engineers are building a version that behaves more like an episomal vector. The two platforms are, in a sense, evolving toward each other.
When the Choice Is Not Obvious
For many therapeutic goals, the choice between AAV and lentivirus is clear-cut. If you need to deliver a gene directly to the retina, liver, or central nervous system in a living patient, AAV is the established option with regulatory precedent and a deep clinical track record. If you need to permanently modify blood stem cells or engineer T cells outside the body, lentiviral vectors are the standard. But some scenarios sit in the gray zone. Gene therapy for muscle diseases, for example, requires reaching enormous amounts of tissue and demands high systemic doses of AAV, which is where toxicity becomes most concerning. Lentiviral vectors could theoretically be delivered systemically for such indications, and manufacturing advances for in vivo lentiviral applications are being actively pursued, but the platform lacks the clinical track record for that route.
Gene editing has further complicated the landscape. Both AAV and lentiviral vectors can deliver CRISPR components to cells, but they do so differently. AAV’s small cargo capacity makes it tricky to fit the entire CRISPR machinery into a single vector, sometimes requiring two vectors to be co-delivered. Lentiviral vectors have room but bring the integration risk. Non-integrating lentiviral vectors offer an appealing middle ground for transient delivery of editing tools, since the editing components only need to be present briefly to make their permanent cut in the genome. This is one area where NILVs could expand the lentiviral platform’s reach beyond its traditional ex vivo territory.
The cost picture also matters. Approved AAV gene therapies have carried some of the highest price tags in medicine, driven in part by the manufacturing challenges of producing clinical-grade vector at the enormous quantities needed for systemic dosing. Lentiviral-based therapies are expensive too, but for different reasons: the ex vivo workflow is labor-intensive, requiring cell collection, transduction, quality testing, and re-infusion for each individual patient. Neither platform has solved the cost problem, and both are betting on next-generation manufacturing to bring prices down enough for broader access.