In vivo gene delivery: Why lentiviral vectors are a compelling alternative to AAV

AUTHORs: Glenda Joanne Dickson, Ph.D., FARZIN FARZANEH, ph.D., nicholas ostrout, ph.d., ViroCell Biologics

Abstract

As the use of gene therapy further expands into systemic and chronic diseases, the tools used to deliver genetic material are under growing scrutiny. While adeno-associated virus (AAV) vectors established early clinical success, their limitations in payload capacity, delivery efficiency, redosing, and manufacturing are becoming harder to ignore. Lentiviral vectors (LVVs), historically most frequently used in ex vivo applications for gene modification, are reemerging as a powerful gene delivery technology for in vivo gene therapeutics, with technical capabilities that are differentiated from those of AAV. The increased number of in vivo applications of LVV is being enabled by innovation and development, through more efficient vector system designs, manufacturing processes, and accompanying analytics. With a high cargo capacity, programmable expression systems, cell targeting, and advancing safety features, LVVs are well-suited for complex therapeutic strategies. Recent developments in targeting, manufacturing, and regulatory control have positioned them to meet the evolving demands of next-generation gene therapies.

The in vivo imperative

As gene therapy moves into its next chapter, the focus is shifting from ex vivo gene modification applications like chimeric antigen receptor T (CAR-T) cell therapies for cancer to in vivo treatments aimed at a wider spectrum of cancer, multifactorial, or chronic conditions, like autoimmune disorders. This progression brings with it new demands, not just in terms of clinical indication but in the complexity and precision required of the gene delivery vehicle itself. In vivo applications require, and must contend with, large quantities of vector due to systemic delivery, with increased focus on biodistribution, tissue selectivity, immune evasion, and long-term safety in ways that provide additional challenges to those experienced with ex vivo approaches.

These challenges have sharpened scrutiny of the viral vectors used to deliver therapeutic payloads. Adeno-associated virus (AAV) vectors initially became the preferred platform for in vivo use owing to their relatively low immunogenicity, non-integrating behavior, propensity of academic researchers with access to AAV vectors, and successful early clinical programs. However, as gene therapy moves toward treating broader patient populations and more complex diseases, the limitations of AAV, particularly its small payload capacity, pre-existing immunity, limitations in effectiveness in dividing cells or tissues and challenges with redosing, and massive dose requirements due to manufacturing challenges are becoming harder to overlook.

In this changing environment, lentiviral vectors (LVVs) are gaining renewed attention. While originally primarily favored for ex vivo applications due to their integrating nature and stable expression in both dividing and non-dividing cells, LVVs are increasingly being engineered for safe and efficient use for in vivo gene delivery. With innovations in targeting, regulation, and manufacturing, lentiviral vectors are emerging as a strong alternative to AAV that offers greater payload flexibility – hence greater potential for substantially greater complexity of regulation, including multiple payloads in combinations of constitutive expression of some genes, and expression of other genes in response to physiological cues. Other advantages include modular design, lower dose requirements, and improved performance in applications where AAV may fall short.

The rise (and constraints) of AAV

AAV earned its early reputation as the leading vector for in vivo gene therapy by addressing many of the safety and regulatory concerns associated with earlier delivery platforms. Its non-pathogenic nature, episomal persistence (i.e., lack of integration into the host genome), and initially low immunogenicity made it an attractive choice for in vivo administration. AAV also benefits from a diverse library of naturally occurring and engineered serotypes, each with varying tissue tropisms, offering some degree of targeting flexibility. These features contributed to its adoption in high-profile gene therapy programs, including approved therapies, such as Luxturna for retinal disease and Zolgensma for spinal muscular atrophy.

However, AAV’s early dominance for in vivo applications is now being reevaluated in light of both clinical experience and the rising demands of next-generation gene therapies. One of the most significant limitations of AAV is its payload capacity, which is capped at approximately 4.7 kilobases, constraining the use of AAV in therapies that require larger or multiple genes, regulatory elements, or safety switches. This size constraint has proven especially problematic for the delivery of genome editing systems like CRISPR/Cas9, as well as for combination therapies that aim to deliver multiple functional payloads in a single vector. While some researchers have moved to using multiple AAVs per therapy, or a combination of AAV and Virus Like Particles (VLPs), this multiple vector approach increases costs and complexity of the final drug product. As such, LVV’s provide a compelling solution due to their larger payload capacity.

Equally concerning is the growing evidence of immune responses to AAV capsid proteins. Pre-existing immunity in the patient population, combined with immune activation following administration, can limit vector efficacy, preclude redosing, and in some cases contribute to adverse events, such as hepatotoxicity. These immune barriers are particularly challenging in systemic indications that require systemic delivery, high vector doses or repeated administration.

From a manufacturing perspective, AAV poses several technical and economic challenges. The production process often yields a high proportion of empty capsids lacking the therapeutic genome, which must be carefully separated from active vector to reduce side effects, ensure potency and consistency. Moreover, scale-up is resource-intensive, with low per-particle potency requiring large-scale bioreactor volumes to meet clinical dose requirements.

Taken together, these issues are prompting many in the field to reconsider AAV’s role as the default choice for in vivo gene therapy. While still valuable in certain contexts, AAV’s structural limitations, immunological profile, and production constraints have opened the door to alternatives better suited for the complexity and scale of the field’s future ambitions.

LVV: Built for complexity

LVVs offer a compelling alternative to AAV, particularly in applications that demand flexibility, large payloads, and long-term performance. One of the most immediate advantages of LVV is a higher payload capacity. With the ability to accommodate 8-10 kilobases of genetic material (and some pushing payloads even higher) – double that of AAV – LVVs are uniquely well suited for therapies that involve large genes, multiplexed constructs, or the inclusion of additional regulatory and safety elements. This capacity enables the use of more sophisticated gene circuits, including multigene cassettes, conditional expression systems, and self-regulating payloads.

LVVs also exhibit a key functional advantage in their ability to achieve persistent expression in dividing cells, or in tissues where cell division is at a low level. Because they integrate into the host genome, they are not subject to dilution in proliferating tissues, making them ideal for hematopoietic, epithelial, or regenerative medicine applications where target cells undergo rapid turnover, or in tissues with low but measurable turnover rates. In contrast, AAV’s episomal expression tends to wane over time in such settings, necessitating higher doses or repeated administration, both of which carry added risks.

While concerns around integration have historically limited the in vivo use of LVVs, modern vector design has significantly reduced these risks. Self-inactivating long terminal repeats (SIN-LTRs and heterologous promoters for expression of the vector genome), combined with the use of constitutive or tissue specific promoters for the expression of some or all of the therapeutic genes, combined with targeted integration strategies, and the option of non-integrating LVVs for certain applications, all contribute to a safe profile. Notably, LVVs have not been associated with hepatotoxicity, a growing concern in high-dose AAV programs, particularly for liver-targeted therapies.

Crucially, LVVs are particularly adaptable to the demands of complex therapeutic designs. They support the use of tissue-specific promoters and enhancers, allowing expression to be confined to desired cell types or physiological contexts. They also accommodate autoregulatory systems that can help to fine-tune gene expression in response to environmental signals or internal feedback loops. Safety switches, such as inducible, and / or context dependent expression of suicide genes, can be incorporated into the same vector backbone, offering an added layer of control over cell fate as a potential intervention in the event that such a potential safety signal arises. 

This modularity makes LVVs especially attractive for the next generation of gene therapies, which increasingly require combinatorial, programmable payloads tailored to dynamic disease states. Rather than designing around vector limitations, developers using LVVs can focus on optimizing therapeutic logic, with the vector acting as a flexible, tunable and robust delivery platform.

Integration, or not: Addressing the safety narrative

One of the most persistent concerns surrounding LVVs in in vivo therapeutic applications has been the risk of insertional mutagenesis. Early-generation integrating vectors, particularly in the context of gamma retroviral-based systems, raised valid safety alarms due to their tendency to insert into the genome at sites close to the transcriptional start site and with vector designs that frequently maintained the use of the viral promoter with powerful enhancer effects, potentially activating surrounding oncogenes. Since the emergence of a number of such safety signals in the early 2000’s a vast body of evidence has been generated showing that the risks of insertional mutagenesis by a retro- or lentiviral vector is multi-factorial – including the influence of the background disease, and the precise vector used e.g. use of SIN-LTRs substantially reduces the risk, and LV have an integration profile that subtly differs from RV, in terms of exhibiting no preference for integration close to transcriptional start sites. There was also the theoretical risk of the vector insertion leading to disruption of essential regulatory sequences, although there is little evidence for this mechanism resulting in a safety signal, presumably because the human genome is diploid. This risk, in the context of the ‘at the time’ limitations in understanding of the vector systems, represented a major reason for why AAV, an episomal vector, became the preferred choice for in vivo applications despite its inherent limitations.

However, the landscape has changed significantly. Advances in vector design, and a greater mechanistic understanding of the causes of insertional mutagenesis, have addressed many of the safety concerns associated with integration. Modern LVVs incorporate self-inactivating (SIN) long terminal repeats (LTRs), which remove the viral enhancer and promoter sequences from the LTRs, reducing the potential for insertional activation of nearby genes. This SIN configuration has become the industry standard and is now routinely used in both ex vivo and in vivo settings to mitigate genotoxic risk. In addition, much care is now taken in selection of the promoter(s) used to express the gene of interest (GOI), where non-viral promoters are typically associated with a high degree of safety versus their viral counterparts (which often possess very high enhancer activity). There are also technologies emerging that can be used to prevent unwanted effects within the vector payload, such as splicing, or inhibitory, unwanted or toxic effects of gene-of-interest expression during vector production.

Beyond SIN design, developers now have the option of using non-integrating LVVs (NILVs). These vectors carry mutations in the integrase enzyme that drastically reduce genomic integration, resulting in the viral payload remaining primarily episomal. While episomal maintenance is less durable in dividing cells, NILVs can be highly effective in post-mitotic tissues, such as muscle, brain, or retina, making them an attractive option for many in vivo therapeutic contexts that do not require long-term proliferation of transduced cells.

Integration is particularly desirable for persistent expression in dividing cell populations and retention of integration capability may also be important in general for long-term / permanent gene expression, and new strategies are making it possible to favour vector insertion towards defined genomic “safe harbors.” These include well-characterized loci like AAVS1 or ROSA26, which are known to tolerate integration without disrupting host cell function or triggering adverse events. Site-specific integration can be achieved through technologies such as engineered integrases, recombinases, or CRISPR-based systems, offering developers a path towards durable expression with minimized risk. Incorporation of such additional technologies will necessitate additional study to gain a full understanding of the strengths and weaknesses of such approaches, to ensure that there are not additional / different risks associated with their adoption.

ViroCell’s approach further enhances the safety and controllability of LVV systems through the inclusion of built-in regulatory safeguards. These include autoregulatory circuits that modulate gene expression in response to intracellular feedback loops or exogenous inducers and kill switch mechanisms designed to eliminate transduced cells if abnormal behavior is detected. Such tools provide an added layer of confidence, especially for in vivo applications where long-term persistence and biological unpredictability present greater challenges.

Taken together, these innovations redefine the risk-benefit equation for LVVs. Integration is no longer an all-or-nothing proposition; it is a controllable feature that can be turned off, redirected, or tightly regulated depending on the needs of the therapy. As all these capabilities continue to mature, LVVs are poised to take on a broader role in in vivo gene therapy, no longer limited by a safety narrative rooted in outdated vector technologies or an incomplete understanding of the associated risk factors. 

Smarter targeting: Ligands and pseudotypes

Precise targeting is advantageous for the safety and efficacy of in vivo gene therapies, where systemic (e.g. IV) delivery must result in selective transduction of disease-relevant tissues or cell types. Historically, AAV has been favored for this purpose due to its large serotype library, with specific variants offering different tropisms that enable increased targeting of muscle, liver, CNS, and other tissues, coupled with less efficient targeting of unintended cells or tissues. However, serotype-based targeting is not ‘all or nothing’ but only a bias towards certain target cells and away from others. In addition, many AAV serotypes display significant species specificity, meaning tropism observed in preclinical models often fails to translate effectively in humans, and such targeting typically results in an increased propensity to target certain tissues and less for other tissue (i.e. not an ‘all or nothing’ effect). Additionally, pre-existing immune responses to AAV capsids can constrain both efficacy and redosing potential.

LVVs, in contrast, offer a more customizable and programmable framework for achieving tissue specificity. One approach is pseudotyping, where the native HIV-1 envelope protein is replaced with glycoproteins from other viruses to change the cellular tropism of the vector. Pseudotypes, such as RD114-TR, GaLV-TR, or measles virus H/F glycoproteins, have been shown to preferentially target hematopoietic, epithelial, or neuronal cells, respectively. These alternatives not only alter the entry pathway of the vector but also help reduce immunogenicity and improve compatibility with stable cell line development.

Beyond pseudotyping, ViroCell is pioneering the use of ligand display technology to bring an entirely new level of targeting precision to LVVs. In this system, packaging cells are engineered to express specific protein ligands, such as stem cell factor (SCF), B7.1(CD80), or truncated low-affinity nerve growth factor receptor (ΔLNGFR), on their surface. During vector assembly, these ligands are incorporated into the viral envelope during budding, allowing the resulting vector to engage with specific receptors on the target cell surface. Importantly, this approach does not require any modification to the vector genome itself, preserving regulatory simplicity and transgene integrity.

In addition to enhancing tissue selectivity, ligand display also facilitates downstream purification. Functional groups, such as His-tags or biotin acceptor peptides, can be fused to display ligands, enabling affinity-based capture and purification of lentiviral vectors through immobilized metal affinity chromatography (IMAC) or streptavidin-based methods. This improves vector purity, reduces contaminating proteins and particles, and enables more scalable manufacturing processes.

Together, these capabilities transform LVVs from passive gene carriers into programmable delivery systems capable of navigating biological environments, selectively engaging target cells, and delivering payloads with molecular precision. This level of control is increasingly necessary as gene therapies move into complex disease contexts, where specificity is not just a performance enhancement but a prerequisite for clinical success.

Manufacturing: From bottleneck to advantage

Manufacturing remains one of the most critical — and often overlooked — factors influencing the viability of in vivo gene therapies at scale. Despite its clinical track record, AAV has proven difficult to manufacture efficiently. The production process often results in a high proportion of empty capsids, which dilute potency and increase the total particle burden delivered to patients. This not only drives up manufacturing costs but also raises the risk of dose-related toxicities, particularly in the liver. Moreover, the high vector doses required to achieve therapeutic effect exacerbate these issues, leading to tighter regulatory scrutiny and greater clinical uncertainty.

LVVs offer several inherent advantages in this domain. First, they tend to exhibit higher potency per particle, due to more efficient gene transfer and expression, especially in dividing cells. This enables lower effective doses in many applications, reducing the overall manufacturing burden and the patient’s exposure to viral components. LVVs also demonstrate superior genome packaging fidelity. LVV particles produced are functional, significantly improving yield and consistency compared with AAV.

Further gains come from innovations in process development for LVV. Unlike traditional adherent systems used for early LVV manufacturing, modern production now relies on suspension-adapted cell lines grown in serum-free, chemically defined media. As outlined in ViroCell’s suspension cell line development program, this transition enables industrial-scale manufacturing in stirred-tank bioreactors, facilitating GMP compliance, reproducibility, and cost-efficiency. The move away from serum and antibiotics also anticipates future regulatory requirements for defined, xeno-free systems, supporting broader adoption of in vivo LVV platforms.

At the heart of ViroCell’s strategy is a dual-phase production platform that aligns vector development with the realities of translational timelines. In early stages, clients can rapidly generate vector using a license-free packaging line combined with a single plasmid transfer vector, supporting research or preclinical batches in a matter of weeks. Later, this setup can be converted seamlessly into a fully integrated stable producer line for GMP manufacturing, avoiding the need to re-optimize or requalify core elements of the system.

A critical enabler of this transition is ViroCell’s use of regulated VSVg expression, a tightly controlled system that mitigates the cytotoxic effects of this otherwise potent envelope protein. By regulating VSVg at the transcriptional and translational level, the company achieves safe, stable production of high-titer vectors even in fully integrated cell lines. This capability overcomes a key historical barrier to LVV scalability and opens the door to broader adoption in in vivo applications.

Together, these innovations reposition LVV not as a manufacturing challenge, but as a strategic asset. With a more efficient process, reduced dose requirements, and scalable platforms built from the ground up for GMP readiness, LVV production, particularly via approaches like ViroCell’s modular framework, offers a viable, future-proof alternative to the bottlenecks plaguing AAV-based therapies.

Strategic use cases: Where LVVs win

As in vivo gene therapy expands into more complex and heterogeneous disease areas, LVVs are emerging as a better fit for many applications where AAV’s limitations are increasingly prohibitive. The value of LVVs becomes particularly evident in therapeutic contexts that require larger payloads, persistent expression, precise regulation, or adaptable delivery strategies.

Autoimmune diseases are a prime example. These conditions often require fine-tuned modulation of immune pathways over extended periods, making persistent and regulatable gene expression essential. LVVs can accommodate multigene constructs that include both effector and regulatory genes, along with tissue-specific promoters and safety switches that adapt expression in real time. Their ability to carry and independently regulate multiple transgenes offers the complexity and control needed to rebalance immune function without tipping into overstimulation or suppression.

Cardiovascular and central nervous system (CNS) diseases present another area where LVVs have distinct advantages. These tissues are largely composed of non-dividing, post-mitotic cells, where genome integration is not always necessary or desirable. NILVs offer durable episomal expression in such cells, avoiding risks of insertional mutagenesis while still delivering long-term benefit. Combined with ligand-based targeting and pseudotyping, LVVs can home in on disease-relevant tissues and achieve therapeutic transduction with fewer off-target effects.

In vivo CAR-T and genome-editing therapies further highlight the need for programmable payload delivery. LVVs excel here by supporting conditional expression systems, such as inducible promoters, autoregulation circuits, and kill switches. These tools allow therapies to be modulated based on disease state, treatment phase, or emergent safety concerns, capabilities that are simply not feasible within AAV’s limited genome space. The same features also enable tunable gene-editing approaches where activity must be tightly controlled to avoid unintended edits or toxicities.

More broadly, emerging therapeutic strategies are increasingly outgrowing what AAV can deliver. As the field moves toward combinatorial payloads, engineered cell interactions, and context-responsive gene circuits, LVVs stand out for their modular design and capacity to support advanced programming. Their role is no longer confined to ex vivo therapies; they are becoming indispensable tools in the push toward systemic, precise, and patient-adapted in vivo interventions.

Conclusion: The next standard?

The debate in gene therapy is no longer about displacing AAV entirely; it’s about whether AAV alone can support the growing complexity of in vivo treatment strategies. As therapies expand beyond rare monogenic diseases into chronic, multifactorial conditions, the demands on delivery vectors increase in kind: larger payloads, smarter regulation, safer control, and scalable production are no longer optional but foundational.

LVVs, particularly in non-integrating formats and with targeted ligand modifications, are increasingly equipped to meet these challenges. Their greater payload capacity, programmability, and adaptability make them well-suited for a new generation of therapeutics that require precision without compromise. With continued innovation in vector design, control systems, and manufacturing, including modular, dual-phase platforms like those developed by ViroCell, LVVs are shedding their historical limitations and emerging as highly versatile tools for in vivo gene delivery.

Regulators, developers, and clinicians alike are beginning to recognize that the future of gene therapy will not be built on a one-vector-fits-all approach. Instead, it will require platforms that can evolve alongside therapeutic ambitions. Lentiviral vectors are no longer limited to niche applications but fully poised to become a new standard in systemic, scalable, and sophisticated gene delivery.

April 2026