The Coming of Age of Gene Therapy for the Treatment of Human Diseases: A Regulatory Perspective

Abstract

After many years of promising clinical results splashed with some serious adverse events, gene therapy has finally reached maturity, as demonstrated by the increasing number of medicinal products approved for commercialization by regulatory authorities. The approved products tackle monogenetic inherited diseases as well as cancer, include both in vivo and ex vivo approaches, and comprise mostly gene additions but also a genome-edited product, demonstrating proof of concept for most gene therapy modalities. Uncertainties still remain, especially on their long-term safety and efficacy, which can only be solved with time. These successes should not lead to self-complacency but rather stimulate the development of necessary improvements concerning manufacturing or the safety and efficacy profile. Here, we review the different categories of gene therapy medicinal products and highlight potential areas for improvement. Products approved for commercialization are taken as the basis for the discussion, since information on their assessment is publicly available. New products and manufacturing approaches under development are also reviewed, with an emphasis on the regulatory challenges expected for some of them.

INTRODUCTION

Gene therapy (GT) applications to treat human diseases are booming.¹ This is the result of several decades of development with impressive clinical successes accompanied by the appearance of very severe side effects.^1,2 Setbacks were followed by intense basic research, which allowed reaching the current state, where many products are showing clinical efficacy associated with acceptable safety profiles. This is reflected by the increasing number of GT medicinal products approved for commercialization by regulatory agencies.^3,4 But approved products only represent the tip of the iceberg of the many clinical trials that are ongoing around the world.⁵

There is no consensus definition for gene therapy medicinal products (GTMPs), so here we will use the one described in the European legislation,⁶ which excludes vaccines for the prevention of infectious diseases as well as small synthetic nucleic acids. Currently approved GTMPs include both products for in vivo application and ex vivo genetically modified autologous cells (namely, chimeric antigen receptor [CAR]-T cells and genetically modified hematopoietic stem and progenitor cells [HSPC]). Until recently, all approved products used the traditional gene addition approach with different viral vectors: gamma retrovirus or lentivirus for ex vivo application and recombinant adeno-associated virus (rAAV) for most of the in vivo products. Some approved in vivo GTMPs contain other vectors, such as Imlygic (talimogene laherparepvec), which has an oncolytic herpes virus as the active substance. More recently, the European Medicines Agency (EMA), the U.S. Food and Drug Administration (FDA), and other agencies approved Casgevy, a product in which clustered regularly interspaced short palindromic repeats (CRISPR)/Cas 9 technology is used to allow the expression of fetal globin in patients with beta-thalassemia and sickle cell disease.^4,5

These successes validate GT (both gene addition and genome editing) as an excellent tool to treat diseases, including many with an unmet medical need, and justify the high number of clinical trials currently ongoing.⁵ However, the run to tackle new indications should not shadow the need to improve existing technologies, especially with regard to manufacturing and characterization. For instance, the increased number of patients potentially treatable with CAR-T cell products (due to new indications and/or their move to earlier lines of treatment) makes the current manufacturing approaches (especially for autologous products) difficult to sustain.⁷ Also, the many applications of rAAV vectors warrant improvements in high-scale manufacturing to increase yield, reduce impurities, and, hopefully, lower the cost of goods.⁸

Here, we review the different categories of GTMP and highlight some of the main regulatory issues as well as improvements needed or under development for each category of products.

CAR-T CELLS

T-cell transfer has emerged as a promising strategy primarily to treat cancer that persists after the initial lines of treatment. T-cell transfer therapy involves the isolation and modification of patients’ T cells ex vivo to improve their ability to fight tumor cells upon readministration.⁹ There are several strategies to develop T-cell transfer therapies, many of which require a genetic modification of the cells with, for example, a recombinant CAR.^10,11

CAR-T cells have improved cytotoxic activity by directly recognizing antigens expressed on-target cells without involvement of major histocompatibility complex (MHC) presentation.⁹ Typically, to generate CAR-T cells, T cells derived from the patient (autologous) or a healthy donor (allogeneic) are isolated from a leukapheresis material, then stimulated in vitro with a cytokine cocktail and finally transduced with replication-incompetent viruses (derived from lentiviruses or gamma retroviruses) carrying the CAR mRNA sequence. Once T cells are transduced, the retrotranscriptase activity of the viral vector generates a complementary DNA (cDNA) sequence of the CAR, which is integrated into the transduced T-cell genome by means of a viral integrase. Finally, the CAR is expressed and displayed on the T-cell membrane as a functional receptor.¹²

The canonical structure of a CAR includes an extracellular domain for antigen recognition, a hinge and a transmembrane region, and an intracellular signaling domain, which initiates the activation cascade.¹³ This is the structure of the first-generation CAR-T cells, in which the extracellular region is derived from the antigen-binding domain of an immunoglobulin (single-chain variable fragment, scFv) and the intracellular region is the signaling domain of CD3ζ. This initial structure has been further optimized, including the addition of costimulatory domains to improve CAR activity. Several elements are combined to create receptors that, after ligand binding in target cells, trigger a signal to activate T cells. Depending on the particular combination of these elements (recognition, hinge, and costimulatory/signaling domains), the transmitted signal can be modulated. The promoter that drives CAR expression is also important. Therefore, the CAR should be carefully designed by choosing the optimal combination of elements to optimize CAR-T cell function. Thus, some developers are designing new types of CAR by tailoring the antigen recognition domains, choosing the costimulatory domains based on the desired properties of the CAR, or modulating its function by adding other domains such as cytokines, cytokine domains, other costimulatory domains, and checkpoint inhibitory domains.⁹

The success of CAR-T cells in clinical trials to treat hematological malignancies led to the authorization of several products by EMA,⁴ all directed against CD19 or B cell maturation antigen (BCMA). CAR-T cells directed against B cell antigens are also currently being tested, with promising results, against autoimmune diseases mediated by autoantibodies.¹⁴ Some research is also targeting the use of CAR-T cells for the treatment of cardiac fibrosis/cardiac injury.^15,16 Other strategies include the development of CAR-T cell therapies against solid tumors and the combination of CARs with genome editing (GE) approaches to optimize CAR-T cell activation by silencing the expression of interfering molecules on the T-cell membrane.¹⁷

Several issues may compromise the success of autologous CAR-T cells. First, oncologic patients usually receive different lines of treatment before the infusion of CAR-T cells, and this may affect T-cell proliferation and/or functional properties. Allogeneic cells obtained from healthy donors can overcome these problems, but this approach requires eliminating the expression of molecules, such as the endogenous TCR, MHC, and others (e.g., CD7, CD52) to avoid undesired immune reactions that could lead to the elimination of the CAR-T cells or the induction of graft-versus-host disease. An additional problem associated with autologous CAR-T cells is the logistics of transporting biological material from the patient to the manufacturing site and back to the patient. These transport steps need to be carefully controlled to keep the cells in an optimal state and maintain traceability, which adds costs, unforeseeable risks, and vein-to-vein time. Manufacturing close to the patient by, for example, a decentralized manufacturing setting¹⁸ could help solve some of these problems, although these approaches pose many regulatory challenges. Thus, the current Good Manufacturing Practice (GMP) guidelines for Advanced Therapy Medicinal Products (ATMP) applicable in the EU¹⁹ state that such an approach should include an approved central site for the oversight of all the decentralized sites in order to guarantee that both the manufacturing process and product quality are consistent in all sites. Other issues to consider include the demonstration of product and process comparability between all sites or process validation at each site. Manufacturers should discuss their decentralized manufacturing approach with regulators before its implementation.

Although the manufacturing of CAR-T cells appears straightforward, there are many factors that need to be carefully controlled to obtain a product of the appropriate quality and optimal functional properties. Thus, raw materials should be carefully selected and controlled, in-process controls should be implemented during the manufacturing process, an extensive product characterization should be performed, specifications should be established (including a well-designed potency assay) to consistently guarantee a product of the desired quality, and product stability should be assured from manufacturing to administration. Manufacturing controls are always necessary, even if part or all of the process is performed in an automated device. Changes in starting materials, raw materials, reagents, or the manufacturing process can have significant effects on the functionality of CAR-T cell products, and hence, the impact of the changes should be tested on appropriate comparability exercises.²⁰ The successful manufacturing of any CAR-T cell is dependent on the quality of the starting materials, especially those used in the genetic modification, such as the viral vector carrying the transgene. A commonly found deficiency in clinical trial applications is the lack of information on the manufacturing and control of this critical starting material, frequently obtained from a third-party manufacturer that sometimes refuses to share this information with the CAR-T cell manufacturer. For clinical trials, it is still possible to submit a separate dossier directly from the manufacturer of the starting material, but for marketing authorization, the applicant should include all information in the dossier.

After administration, CAR-T cells are expected to proliferate upon interaction with cells expressing the target antigen. Their proliferation capacity and persistence are influenced by factors such as the T-cell subtypes present in the cell source and in the final product and by the degree of differentiation of these cells. In addition, the duration of the ex vivo expansion step, the cytokines and other reagents used, and the type of signal transmitted by the (co)stimulatory domains play an important role in the functionality of CAR-T cell products, such that stronger signals or prolonged expansion times could lead to T-cell exhaustion, compromising the efficacy of the product.^21

–24 Some developments are in progress aiming at shortening the manufacturing time of CAR-T cells, thereby preserving the ability of cells to expand further.²⁵

In addition to exhaustion, other factors, such as the loss of the target molecule on tumor cells or the presence of tumor cells that do not express the target molecule, can also compromise efficacy.²⁶ To overcome these problems, CAR-T cells specific for more than one target are in development.²⁷

Other approaches are aimed at introducing the CAR construct directly into the patient by infusing specific viral vectors or encapsulated mRNA carrying the machinery necessary for T-cell transduction in vivo. ²⁶

In addition to CD19 and BCMA ligands, other target molecules are being explored for the treatment of different blood malignancies. Targets expressed in solid tumors are also being investigated, although this is a more difficult task, as there are few tumor-specific molecules, and the use of targets that are also expressed in healthy tissues can lead to off-tumor toxicity. Furthermore, it is difficult for CAR-T cells to penetrate the tumor, and the expansion and survival of CAR-T cells in the tumor is limited and even abolished due to the immunosuppressive microenvironment. Several approaches are being investigated to overcome these difficulties, such as intratumor administration, modification of CAR-T cells to make them express cytokines, abolishing/regulating the expression of PD1, or increasing the expression of enzymes able to degrade extracellular matrix.²²

Apart from their proven efficacy, CAR-T cell therapies are not exempt from toxicities, such as cytokine release syndrome or immune effector cell-associated neurotoxicity syndrome. These toxicities oscillate from mild to very severe and, if not properly managed, can be fatal. Other risks related to CAR-T cells are those associated with lymphodepletion, B cell aplasia, tumor lysis syndrome, and anaphylaxis.²⁸ Furthermore, a theoretical risk of insertional oncogenesis exists, as for all genetically modified cells. In fact, recent cases of malignant T-cell neoplasia in patients who had received CAR-T cells have been described.²⁹ However, the correlation between the development of these secondary T-cell malignancies and vector insertion is still unclear. Further investigation of these events and a thorough follow-up of patients receiving CAR-T cell therapies are required before reaching a conclusion.

Despite these undesirable reactions, there is currently a consensus in regulatory agencies that the benefits of CAR-T cell therapies outweigh their risks in the oncology indications approved so far. However, this conclusion can change for other indications. Strategies are being investigated to ameliorate the toxic effects associated with CAR-T cell therapies. Thus, systems are being studied to regulate CAR expression in cells, such as an on/off switch system, combinatorial CARs, inhibitory CARs, and expression of suicide genes in the CAR-T cells.²⁷ Nonviral delivery systems are also under development to diminish the risk of insertional mutagenesis.²⁶

OTHER CELL-BASED GTMPs

Beyond CAR-T cells

As an alternative to CAR-T cells, other cell-based gene therapies are being developed, including CAR-natural killer (NK) or T-cell receptor (TCR)-T cells.³⁰ NK cells can recognize and kill tumor cells independently of MHC antigen presentation, but they can also be modified to express a CAR. However, CAR-NK cells present some disadvantages when compared to their CAR-T cell counterparts that should be addressed before they can proceed beyond initial clinical development. These differences include a reduced transduction efficiency, short in vivo persistence, and lower resistance against freeze/thaw stress, as well as an extensive donor dependency.³¹ Trying to overcome some of these limitations, new developments are focusing on the use of NK cells generated from induced pluripotent stem cells (iPSCs). iPSCs offer a more reproducible cell source for clinical application as they enable the generation of homogenous cell banks and exhibit an increased proliferation ability as well as an enhanced transduction efficiency.³² iPSC-derived CAR-NK cells have shown promising cytotoxicity against refractory tumors in some preclinical studies³³ but still face important technical limitations like the absence of protocol standardization and nonoptimal differentiation procedures that may hinder the implementation of GMP manufacturing.³⁴

TCR-T-cell therapy focuses on the expression of tumor-specific, naturally occurring T-cell receptors in the patient’s T cells.³⁵ Through this approach, TCR-T cells can recognize intracellular and extracellular antigens presented by the MHC, broadening the applicability of these products, especially against solid tumors. However, because of this same property, the development of TCR-T-cell therapies is currently restricted by MHC compatibility.

Expression of CARs in other cell populations, such as γδ T cells or even macrophages and neutrophils, is also being explored.³⁶ In addition, there are some clinical developments with genetically modified dendritic cells for the treatment of several oncologic malignancies. Some of these developments pursue genetic modification of the dendritic cells to alter their signaling pattern or target them to specific tumor antigens to generate stronger immunological responses.³⁷

Genetically modified HSPC

HSPCs are multipotent, undifferentiated cells that can develop into all the different types of blood cells, including both the myeloid and lymphoid lineages. HSPCs mainly inhabit the bone marrow but can also be found in peripheral blood and the umbilical cord. Because of their wide availability, easy access, and developing potential, these cells have traditionally been employed as sources for different clinical applications.³⁸

Five different GTMPs consisting of genetically modified HSPCs have already received centralized marketing authorization in the EU (Skysona, Libmeldy, Zynteglo, Strimvelis, and Casgevy), albeit two of them (Skysona and Zynteglo) were later withdrawn from the market due to commercial issues.

Despite the demonstrated clinical efficacy of these treatments, administration of genetically modified HSPC has been associated with the development of severe hematological malignancies directly caused by insertional mutagenesis of the integrating vector.³⁹ Although this adverse genotoxicity was mainly observed when employing early-generation gamma retrovirus vectors, this can also occur with improved lentivirus vectors that have self-inactivating transcriptionally silent LTRs (SIN),⁴⁰ as shown with the recent cases reported on clinical trials with Skysona. ⁴¹ To diminish the potential risk of insertional mutagenesis that is inherent to the integrative vectors, vector design should be carefully considered. In addition to the use of SIN vectors, the selection of constitutively weak or even lineage-specific cellular promoters should be considered.⁴²

In addition, new emerging methods like GE are also being explored for the introduction of specific, targeted genetic modifications in the therapeutic cells. One example that has recently received marketing authorization from EMA, FDA, and other international regulatory agencies is Casgevy. A theoretical risk of oncogenesis caused by the gene editing process in this type of product (e.g., off-target edits) still exists and should be considered and addressed during product development (see below).

Another characteristic of the use of genetically modified HSPC is that, due to their self-renewal properties, a lengthy long-term safety follow-up is required. This is supported by the occurrence of some adverse genotoxic events caused by insertional mutagenesis up to 15 years postinfusion.⁴²

Recombinant bacteria

Some ongoing clinical developments with GTMPs consist of recombinant living nonpathogenic prokaryotic cells.⁴³ When designing these organisms, careful attention should be paid to key parameters such as bacterial genetic stability, residence time, replication capacity, and clearance to increase the reproducibility of their therapeutic properties and minimize their potential risks. In addition, the risk of horizontal transfer of antibiotic resistance genes to other residing microorganisms should be addressed and prevented if possible. Currently, some of these developments are being explored clinically for the treatment of a broad number of conditions, but none has yet received commercial approval in either the EU or the United States. However, the clinical development of recombinant living bacteria still faces some limitations, like the complex determination of dose–response mechanisms and the frequent dosing required by noncolonizing microbes with short-residence times, so further investigation in this field is warranted.

rAAV-BASED GTMPs

rAAVs are currently the main vectors of choice for in vivo GT, mainly due to their limited ability to integrate into the patient cells’ genome and their lower innate immunogenicity.

Innovations in rAAV vector design have enabled improvements in transduction efficiency, tissue-targeting specificity, and vector safety profile (e.g., by reducing immunogenicity or hepatotoxicity). However, other characteristics of rAAV manufacture still need to be optimized, particularly aiming at yielding higher infectious viral titers and improving the impurity profile. Long-term expression is another important feature that has led to a relevant/wider use of rAAVs in the clinic. The downside has been the putative risk of oncogenicity associated with long-term expression (see below). An additional safety risk to avoid when preparing rAAVs is the presence of replication-competent AAVs (RCAs).

rAAV vector design

Several improvements in vector design have been developed over the years and recently implemented in clinical products. On the one hand, optimization of the vector includes strategies to overcome the limited cargo size, increasing the packing capacity of the vectors, capsid optimization to evade the host immune response and/or achieve the desired tropism, and the use of tissue-specific promoters to direct gene expression to the target cells. On the other hand, optimization of the transgene sequence is normally required to fit into the limited cargo size of the vectors and to achieve sequence improvement using specific variants of therapeutic interest or codon optimization to increase translation efficiency and avoid host immune response. Some of these strategies are summarized below.

The limited cargo size restricts the use of rAAV in many indications, as the size available for transporting exogenous sequences is limited to around 4.4 kb. Several strategies have been put forward to ameliorate this limitation. Investigations focusing on exploring the upper limit of AAV packing concluded that an rAAV could package up to 5.2 kb, which corresponds to 500 bp more than the usual wild-type AAV size of 4.7 kb.⁴⁴ The size limitation is still an important problem that has led to the use of reduced transgene sequences and the removal of native regulatory regions (promoters, enhancers). -

The combination of capsids from different virus serotypes has enabled the increase in cargo size. Working with rAAV2/1-5 showed that it is possible to go even up to 6 kb, although above 5.3 kb, the transduction efficacy was dramatically reduced. The reason is a postentry step that blocks the infection by larger genome-containing virions, which can be bypassed with proteasome inhibitors. In a study with up to 8.9 kb in different vectors (rAAV2/1–5,7–9), the best results were achieved with rAAV2/5 vectors, suggesting that the best suited for higher capacity are the AAV5 capsids. Indeed, this is a serotype with differences from the others, although recent studies of these differences do not support a volume high enough to package a genome 50% larger than the typical AAV capacity. Additional studies with AAV2 and AAV5 have failed to package more than 5.2 kb.⁴⁵ That is not that far from previous studies with AAV8, which only managed to package up to 5.6 kb at the cost of losing infectivity.

The use of two AAV vectors allows packaging the full sequence of large transgenes. A limitation of this approach is the need to achieve coinfection of the target cells with both rAAV vectors. The reconstitution of the native sequence is typically mediated by homologous recombination or trans-splicing. A successful example of this dual AAV delivery strategy was recently reported for the human OTOF gene to restore the hearing loss in patients carrying mutations in the Otoferlin gene.⁴⁶ The two rAAV vectors strategy has also been used for other objectives, such as the vectoring of the two elements of the CRISPR/Cas9 GE technology.⁴⁷ Recently, three rAAVs have been used to express large dystrophins using protein trans-splicing mediated by split inteins.⁴⁸

Vectors using the self-complementary AAV (scAAV) strategy allow the acceleration and enhancement of gene expression (potentially allowing for lowering the vector doses), at the price of reducing the cargo size by half (to 2.2 kb), rendering this strategy impractical for most applications. In addition, this approach seems to increase innate immune responses⁴⁹ and transgene product-specific CD8+ T-cell and antibody responses.⁵⁰ Despite these limitations, the scAAV strategy has been used in Zolgensma, a commercially approved self-complementary AAV9 vector.⁵¹

Capsid optimization is pursued with different objectives: -

To evade the host immune response. More than 50% of the global population has neutralizing antibodies to AAV serotypes 1, 5, 6, 8, 9, and Rh74var. As some of these have been used in EU-authorized rAAV-based GTMPs (see Table 1), there is room for improvement in the choice of capsids.

To direct the entry of rAAVs to the desired target cells when using capsids with a restricted tissue specificity, which also avoids their biodistribution to unnecessary/undesired tissues. The existence of at least a dozen human AAV serotypes (and more from compatible non-human primates) with different tissue tropism⁵² has led to different strategies to retarget rAAVs, such as pseudotyping by combining the serotype of the viral genome with capsids from “exotic” serotypes. Other strategies are transcapsidation, which produces mosaic or chimeric capsids, or introducing more complex manipulations such as new capsids generated by directed evolution or rational design, and the use of bispecific molecules. Although cell tropism is affected by other interactions, such as intracellular trafficking and viral uncoating, so far EU-authorized rAAV-based GTMPs have used simple pseudotyping to modify the vectors, leading to a variety of capsids (see Table 1). In the recently EU-authorized Beqvez, which uses the Rh74var capsid, directed mutagenesis was applied to evolve the parental capsid (Rh74) toward specific amino acid residues more conserved among AAV serotypes traditionally considered hepatotropic.

Table 1.

AAV-Based Gene Therapy Medicinal Products Authorized in the EU

Date authorized	NamG	Serotype AAV	Vector design	Production platform	Promoter	Transgene/indication	Dose (vg/kg)	vg/mL
2012 (up to 2017)	Glybera (alipogene tiparvovec)	AAV1	Standard AAV	Insect cells/baculoviruses	CMV IE promoter	LPL^S447X /familial lipoprotein lipase deficiency (LPLD)	1.0 × 10¹²	3 × 10¹²
2018	Luxturna (voretigene neparvovec)	AAV2	Standard AAV	Mammalian cells/3 plasmids	CMV enhancer + CB promoter	hRPE65 /vision loss due to inherited retinal dystrophy caused by confirmed biallelic RPE65 mutations (Leber congenital amaurosis and retinitis pigmentosa)	1.5 × 10¹¹	5 × 10¹²
2020	Zolgensma (onasemnogene abeparvovec)	AAV9	scAAV	HEK293/3 plasmids	CMV enhancer + CB promoter	SMN1 /spinal muscular atrophy (SMN1 deficiency)	1.1 × 10¹⁴	2 × 10¹³
2022	Upstaza (eladocagene exuparvovec)	AAV2	Standard AAV	HEK293/2 plasmids	CMV IE promoter	hAADC /severe aromatic L-amino acid decarboxylase (AADC) deficiency	1.8 × 10¹¹ (total)	5.6 × 10¹¹
2022	Roctavian (valoctocogene roxaparvovec)	AAV5	Standard AAV	Sf9 insect cells/baculoviruses	Hybrid human liver-specific promoter HLP	hFVIII-SQ /severe hemophilia A (congenital factor VIII deficiency)	6.0 × 10¹³	2 × 10¹³
2023	Hemgenix (etranacogene dezaparvovec)	AAV5	Standard AAV	Sf9-derived insect cells/baculoviruses	Liver-specific promoter LP1	hFIX-Padua /severe and moderately severe hemophilia B (congenital Factor IX deficiency)	2.0 × 10¹³	1 × 10¹³
2024	Beqvez (fidanacogene elaparvovec)	AAV Rh74var	Standard AAV-directed mutagenesis capsid	HEK293/3 plasmids	Liver-specific promoter and hepatic control center	hFIX-Padua /severe and moderately severe hemophilia B (congenital factor IX deficiency)	5.0 × 10¹¹	0.79–1.21 × 10¹³

Data collected from the public EPARs and SmPCs (available at the EMA website).

AAV, adeno-associated virus; CB, chicken-β-actin; CMV, cytomegalovirus; EPAR, European public assessment report; SmPC, summary of product characteristics.

An important area of improvement for the in vivo administration is the use of tissue-specific promoters to drive the expression of transgene cassettes in particular tissue types, in contrast with commonly used promoters that drive strong and ubiquitous expression, such as the cytomegalovirus (CMV) or the chicken beta-actin promoters. This is an important safety aspect, as it avoids overexpression and off-target expression while sustaining an appropriate expression level in the right tissue. The three rAAV-based EU-authorized GTMPs encoding clotting factors (Roctavian, Hemgenix, and Beqvez) use liver-specific promoters, highlighting the importance of this issue for this particular class of products. Interestingly, an equivalent vector to Zolgensma, but with the native human promoter of its transgene, SMN1, seems to offer better safety and improved efficacy in a mouse model (the new construct also includes a codon-optimized transgene).⁵³

Genetic engineering has been used to reduce transgene size to fit the small cargo space in rAAV vectors. In Roctavian, the Factor VIII sequence is not the native form but a functional sequence approximately 40% smaller than the wild-type (FVIII-SQ), in which the B domain (907 aa out of a total protein of 2332 aa) was replaced by a 14 amino acid “SQ” linker sequence. Although this reduced FVIII is functional in hemostasis, the B domain is by no mean useless.⁵⁴

Improvements of the transgene sequence have included the use of specific sequence variants of enhanced therapeutic value. For example, Glybera used a natural gain-of-function gene variant, LPL-S447X. Other improvement is the use of codon optimization that allows the increased translation of the original native sequence. Hemgenix and Beqvez are codon-optimized and use the high-activity Padua variant of the factor IX gene (FIX-R338L). Roctavian also uses a codon-optimized transgene.

rAAV manufacturing strategies

Main rAAVs manufacture strategies involve either transient transfection of HEK293 cells with plasmids or baculovirus infection of insect cells. These manufacturing strategies show low efficiency of viral packing and assembly, which leads to a high proportion of empty or incomplete particles in the final product.⁸ This is more significant in the insect cells/baculovirus system compared to the transient transfection of mammalian cells with plasmids, as in each system, there are different post-translational modifications of the capsid and viral genome modifications that impact the viral particle assembly. Each system also generates different host cell proteins impurities with differential immunogenicity potential. All that leads to a higher potency of the human cell-produced rAAVs than the baculovirus/Sf9 vectors.⁵⁵

The potential connection between high viral titer (dose) and adverse events is a subject of concern in the clinical use of rAAV-based products. Adverse events related to the dose could include hepatotoxicity, thrombotic microangiopathy (TMA), and neurotoxicity. These risks may be attributed to the excessive activation of the complement pathway, which is a component of the innate immune system, further highlighting the potential for an immunological response in carriers of AAV.⁵⁶ TMA after administration of high doses of rAAVs has been reported in Zolgensma-treated patients (0.6%).⁵¹

As a consequence of the poor yield obtained with the current rAAV manufacturing systems, high amounts of vectors are required to ensure a sufficient number of fully active viral particles. Thus, the higher the dose used, the higher the number of empty particles that are coadministered, which may also contribute to those adverse effects. Duan (2023) considers the use of intravenous doses of more than 5 × 10¹³ vg/kg risky,⁵⁷ comparing cases of death in Zolgensma and other rAAV-based GTMPs under clinical trials. From a regulatory perspective, manufacturing improvements would be desirable to achieve a higher proportion of fully active/infectious particles, including optimization of downstream purification steps, strategies to enhance vector genome integrity, implementing mechanisms to improve capsid assembly, etc.

Whereas AAVs have been extensively studied and have been well-known viruses for decades, some aspects of their biology related to the encoded ORFs and virus encapsidation continue to be areas of active research. As examples of recent developments, the viral assembly-activating protein was described in 2010, and the membrane-associated accessory protein in 2019. These proteins are involved in capsid assembly and viral release. Therefore, further investigations on capsid assembly may be relevant to develop better manufacturing strategies to achieve higher functional (infectious) titers.

RCAs presence and detection

rAAV vectors used as GTMPs are engineered to be unable to replicate. Several strategies are considered during vector design to reduce the risk of RCAs formation, such as delivering the viral genes in different plasmids. Nevertheless, formation of RCAs can occur during vector manufacturing through homologous and nonhomologous recombination, and, therefore, regulators request that the presence of RCAs be examined in rAAV-based GTMPs.

The current standard method that manufacturers use to test the presence of RCAs in an rAAV preparation is a cell-based assay followed by quantitative polymerase chain reaction. The method is a limit test, as it uses a defined number of infectious particles as a positive control, representing the limit of detection (LoD) of the assay. Validation requirements are very limited for this type of assay, and the LoD is not standardized. Consequently, testing results among different manufacturers are heterogeneous as the LoD varies. Optimization and standardization of the assay might need to be implemented to obtain homogeneous values among different rAAV preparations.

Ideally, the absence of RCAs (<1 RCA/dose) should be demonstrated. Low levels of RCAs can be accepted provided that a proper risk assessment is presented showing negligible risk for the patients.

Long-term expression and the risks of oncogenicity and immunogenicity

Protein expression in rAAV-based GTMPs used to be considered “transient” in contrast to the “stable” transduction achieved by retro- and lentiviruses. Nevertheless, extended gene expression is achieved in nondividing cells from rAAV vector genomes present in the nucleus as episomes.

Data from recent investigations in dogs, humans, and macaques lead to the speculation that sustained gene expression is mediated by integrated rAAV vector genomes. This hypothesis/finding needs to be further investigated to determine the extent of this integration and its potential contribution to the long-term therapeutic effect.⁵⁸ However, this may result in a potential risk of oncogenicity if integration occurs within or nearby certain genes. The extent of genome integration and the oncogenic potential of these vectors were already mentioned and discussed in an EMA-relevant document²⁰ after the findings described in 2007 of an increased rate of hepatocellular carcinoma in neonatal mice treated with an rAAV. The ensuing debate concluded that integration of rAAV does occur in humans, but at a low frequency, and that the oncogenic risk is largely specific to mice.⁵⁹

Correlated to the potential risk of oncogenicity associated with vector integration in the cellular genome, the lasting expression of the rAAV genome in the target cells might represent a potential risk of immunogenicity.

mRNA-BASED GTMP

The rapid development of two effective mRNA vaccines against SARS-CoV-2 gave a great impulse to the development of mRNA products to protect against other infectious diseases (vaccines) and to treat different medical conditions such as cancer, autoimmune and metabolic disorders, and cardiovascular and genetic diseases (GTMPs).^60

–63 In particular, the relatively short production time makes mRNA-based products the ideal candidates for personalized cancer immunotherapies.^64
–66

Concerning mRNA-based GTMPs, most of the products included in preclinical or clinical trials are in vivo therapies, mainly in oncology indications using mRNAs expressing neoantigens or tumor-associated antigens.^60

–66 However, there are also ex vivo therapies where cells are treated with mRNA before being infused into the patient, for example, CAR-T cells.^67,68

Here we briefly review the mRNA manufacturing process, the quality concerns, future perspectives, and regulatory aspects of this type of GTMP.

Manufacture of mRNA products

Compared to other GTMPs, the mRNA manufacturing process is straightforward and highly similar among different manufacturers. Briefly, the mRNA drug substance (DS) manufacturing process starts with the in vitro transcription (using a cell-free enzymatic system) of a linear DNA template (linearized plasmid or PCR product) encoding the desired therapeutic polypeptide, followed by DNA and protein digestions, different chromatography steps (e.g., oligo dT affinity chromatography), various stages of ultrafiltration and diafiltration, and final filtration and dispense.

Although there are some approaches in which the naked mRNA is used directly (e.g., ex vivo use),⁶⁰ in most of the products under investigation, the mRNA is transported by a vehicle system.^60,61 Currently, the most used vehicle system is the lipid nanoparticles (LNPs), which are composed of a lipid mixture (cationic ionizable lipid, polyethylene glycol [PEG], phospholipids, and cholesterol) that surrounds the mRNA.^60,61,69 For mRNA drug product (DP) manufacturing, typically the lipids and mRNA DS are mixed, obtaining the LNPs that are subjected to a buffer exchange and concentration steps, several filtrations, a final sterile filtration, and an aseptic filling in the container closure system. For mRNA DP with other vehicle systems, the manufacturing process is similar.

Independent of the manufacturing processes, the mRNA DS and DP should be deeply characterized, and release specifications established for an appropriate quality assurance.⁷⁰

Problems affecting the quality of the mRNA

Although the mRNA manufacturing process is not very complex, there are aspects in different steps that could affect the quality of these products and, consequently, their clinical performance. Here we list the most common and studied problems and propose some solutions: •

The mRNA design is directly related to the translation efficiency, which is linked to the product efficacy.^60,61 The mRNA is composed of five functional regions, including the 5′ cap (or alternative), the 3′ poly(A) tail, the open reading frame (ORF) of the polypeptide of interest, and flanking untranslated regions (UTRs). Different approaches are carried out to increase the mRNA translation efficiency as well as the mRNA stability involving all the structural parts of the mRNA,^60,61,71,72 namely ORF codon optimization, use of modified ribonucleotides (the most used is N¹-methylpseudouridine instead of uridine that also decreases innate immunogenicity), sequence optimization of the 5′- and 3′-UTRs as well as of the poly(A) tail, use of modified 5′ cap, and 3′ poly(A) tail with different lengths. The clinical success of the product is highly dependent on the design of the mRNA sequence.

•

The duration of protein expression/production from the mRNA is relatively short due to the degradation of the mRNA in the cells.^60,61,72 Different approaches are under investigation to reduce degradation, for example, the use of circular mRNA to protect it from exonuclease degradation, increasing the stability and half-life.^73

–77 Investigation on this aspect is especially relevant for the development of mRNA-based GTMPs for the treatment of chronic diseases to reduce the number of doses.

•

A recent study has shown that incorporation of N¹-methylpseudouridine instead of uridine into the mRNA in Comirnaty sometimes results in a + 1 ribosomal frameshift that elicits cellular responses in vaccinated subjects to +1 frameshifted products.⁷⁸ To avoid this effect, manufacturers should use specific software to predict possible +1 frameshift sites in the mRNA sequences from their products and remove them by replacing nucleotides (codon optimization). As control of the effectiveness of the codon optimization, translated products obtained in vitro should be evaluated, as well as immune responses produced in model animals after treatment with this mRNA GTMPs. Alternatively, the use of different modified ribonucleotides could be explored.

•

Impurities produced during manufacturing could affect the safety and/or efficacy of the mRNA GTMP. There are currently several methods for mRNA purification, and manufacturers must choose their best options for obtaining mRNA DPs with a high level of purity. For instance, mRNA molecules with short-length poly (A) tails have reduced stability, and formation of mRNA–lipid adducts disrupts mRNA translation, negatively impacting the activity of LNP-formulated mRNA products.^60,72 Regardless of the purification methods, the impurities should be characterized by orthogonal analytical methods due to their different sensitivity. Among the impurities, residual DNA template, residual dsRNA, mRNA with short poly(A) tails and/or incorrect 5′ cap, degraded mRNA, and RNA–lipid adducts should be measured and controlled during DS and/or DP release. Depending on the manufacturing process, additional impurities may be included in the DS and DP release specifications.

•

The design and development of an assay that demonstrates the potency of these mRNA GTMPs can be challenging. The test should demonstrate that the mRNA is delivered to the target cells (with/without delivery systems) and translated into the therapeutic polypeptide. Then, because mRNA GTMPs are used to treat diverse conditions, different functionality tests should be developed that reflect the mechanism of action. Thus, for protein replacement, function recovery should be demonstrated. For cancer treatments where the mRNA expresses personalized neoantigens, once the translated polypeptide is processed by the proteasome into peptides, they are expected to be loaded onto specific Class I and Class II HLA molecules for presentation on the surface of antigen-presenting cells to induce specific immune responses. In this case, the development of a common potency test (measure of the specific immune response) for batch release is even more challenging (e.g., specific reagents needed for the neoantigens of each patient). In these cases, regulators will request at least suitable surrogate tests.

•

Cancer cells have mutated proteins that contain tumor-specific antigens and neoantigens. For personalized products, manufacturers use specific software for neoantigen’s selection. This software should be well described in the application dossier and properly validated for its intended use, as part of the manufacturing process.

•

Hypersensitivity to vehicle components is a major concern in repeated use. Strategies to decrease the required number of doses and/or reduce the mRNA quantity per dose are under study.⁷³ For example, the use of self-amplifying mRNA (sa-mRNA) in experimental models has shown that lower mRNA doses can yield an equivalent protein expression and long-lasting efficacy.⁷⁹ A vaccine against SARS-CoV-2 (Kostaive) using this technology has recently been approved in the EU.

•

Manufacturers should investigate different excipients to increase the stability of these products and to allow their storage at higher temperatures (currently, the approved mRNA products are stored at temperatures around −80°C), which would facilitate distribution and lower costs.

Future perspectives and regulatory status

Although the development of mRNA-based therapies started in the nineties,^60,62,63 the marketing authorization of two mRNA vaccines against SARS-CoV-2 in December 2020 and January 2021 gave a great impulse to the development of these products, not only in the vaccine field but also as GTMPs (see above).

Regarding mRNA-based GTMPs, a high number of clinical developments (including preclinical phase) are ongoing for the treatment of different medical conditions. Although some of the preclinical and clinical results described are encouraging, developers are advised to continue analyzing all the information of each product, starting with their quality aspects.

From a regulatory point of view, EMA as well as the European Directorate for the Quality of Medicines and Healthcare (EDQM) are working on the drafting of a guideline (guideline on the quality aspects of mRNA vaccines [Draft] [EMA/CHMP/BWP/82416/2025]) and general chapters (DNA templates for the preparation of mRNA substances [Ph. Eur. 5.40], mRNA substances for the production of mRNA vaccines for human use [Ph. Eur. 5.39], and mRNA vaccines for human use [Ph. Eur. 5.36]), covering quality aspects of mRNA vaccines. Although mRNA GTMPs are out of the scope of these official documents, some of their guidance will be applicable until specific guidelines for mRNA GTMPs are prepared.

OTHER IN VIVO GTMPs

Apart from rAAV vectors and encapsulated mRNA, other strategies are used for in vivo gene delivery. These include recombinant adenovirus, lentivirus, herpesvirus, or poxvirus. In vivo gene delivery using plasmid vectors is mostly residual.⁸⁰

In the context of cancer treatment, oncolytic viruses (OVs) and viral vectors encoding for tumor antigens or immune modulators are being explored. One example is nadofaragene firadenovec, a recombinant adenovirus carrying the IFNα gene, for which a marketing authorization application is currently being evaluated for the treatment of bladder cancer by EMA.⁸¹

Regarding OVs, their conditioned replication is expected to present with low systemic toxicities as compared to, for example, the immune deregulation induced with the use of immune checkpoint inhibitors (ICI).⁸² Regarding their oncolytic activity, OV ability to induce immunogenic cell death with release of danger-associated molecular pattern signals from tumor cells and their capacity to remodel an immunologically “cold” tumor microenvironment into a “hot” one, with infiltrating immune cells, has received increasing attention.⁸³

OVs have been intensively researched in many clinical trials performed over two decades⁸⁴ demonstrating safety but only moderate efficacy. So far, there is only one authorized OV in the EU for late-stage melanoma therapy (Imlygic).⁸⁵ In most cases, OV carries transgenes intended to stimulate the immune system: GM-CSF, ScFv-PD1, αCD47, IL12, IFNα, or bispecific T-cell engagers (BiTE) exhibiting a synergistic action with the immune-stimulatory effects triggered by the virus,⁸⁶ enhancing the antitumor response. Several studies have shown that combinations of OVs and other existing cancer therapies, such as T cells and ICIs, result in improved therapeutic efficacy and overcome the potential resistance against either individual treatment.⁸⁶

Recombinant viral vectors for in vivo gene delivery are also being investigated for further pathologies. For instance, Vyjuvek (Beremagene geperpavec) is an HSV-1 delivering two copies of the COL7A1 gene for the treatment of Dystrophic Epidermolysis Bullosa, for which a marketing authorization has been recently granted in the EU.⁸⁷

Regarding the quality regulatory requirements for this kind of GTMPs, some aspects must be demonstrated at the virus seed lot level, such as the tumor-specific cytolytic activity, when pertinent, and the absence of wild-type viral particles. For most products, vectors should be developed to be replication incompetent, and the absence of replication-competent viruses should be confirmed for each lot. For OV, modifications introduced for conditioning or avoiding replication in nontumor cells should be confirmed, if possible, in primary cells, also taking into consideration the permissiveness of the animal species to the parental virus. For integrative vectors, integration capacity should be determined, and additional safety considerations regarding the risk of integration-driven mutagenesis should be taken into account. Demonstration of transgene functionality is also considered crucial for these products. Impurities such as empty particles, nongenomic DNA, and impurities related to the transgene should be analyzed and kept to a minimum.

GENOME EDITING

GE allows precise alterations in the genome, which can be a powerful tool for in vivo and ex vivo therapies. Different GE technologies have been developed since the end of the 20th century, including meganucleases, megaTALs, zinc finger nucleases, transcription activator-like effector nucleases, and the CRISPR/Cas RNA-guided nucleases and derivatives. The development of the CRISPR/Cas tools has particularly revolutionized the field, since they can recognize their targets by simple interactions between the target DNA and a single-guide RNA molecule that localizes the Cas protein, providing a very effective, versatile, and cost-effective manner to engineer the cells’ genome. The CRISPR/Cas tools have been quickly translated into the clinic, with more than 100 clinical trials currently ongoing. Fewer clinical trials are also exploring the other nucleases mentioned above, but since they require customized proteins or DNA-binding domains, their use implies much longer development times and higher costs than the CRISPR/Cas products.⁸⁸

Manufacture and testing of the GE tools for clinical use

GE normally requires a modifying enzyme (delivered as a DNA expression cassette, mRNA, or protein) and, when using the CRISPR/Cas systems, a guide RNA (gRNA) that forms ribonucleoproteins (RNPs) with the modifying enzyme. If required, a donor sequence can be added for insertion at the cutting site. The regulatory classification (and hence the regulatory requirements) of the different GE components will depend on the type of therapy (e.g., ex vivo vs. in vivo).⁸⁹

For ex vivo therapies, where genetically modified cells contain the genome edition, the GE components are considered starting materials, which are delivered into cells by physicochemical methods (such as transfection or electroporation), recombinant viral vectors, or nonviral vectors. The genetically modified cells should be manufactured under full GMP conditions, but the GE components, classified as starting materials, can be manufactured following the GMP principles, according to the EU legislation.⁹⁰

For in vivo therapies, the GE components are normally considered the DS and can be administered by nonviral vectors (commonly LNPs) or by recombinant viral vectors (commonly AAVs). In this case, full GMP is required for their manufacturing.

The manufacture and testing of the GE components and delivery systems should follow the requirements for each specific product type, some of them discussed in the different sections of this review. In addition, there are specific quality regulatory requirements that the GE products should comply with: •

The intended modifying mechanism should be thoroughly described in the application dossier for regulators to review. Assays should be in place to measure the on-target editing efficiency of the desired GE modification. The manufacturers should also implement biological assays representative of the proposed mechanism of action (the result of the intended edition).

•

Assays for on-target/off-target edits should be conducted and considered as part of process development and characterization. A range of state-of-the-art in silico and in vitro methods⁹¹ should be validated and used to map on/off edits. The extent of this mapping should scale with the risk of the medicinal product. In addition, the target cells should be analyzed for large DNA-fragment inversions, deletions, duplications, or chromosomal rearrangements.

•

The impurities should be thoroughly characterized and controlled, as they may affect safety. For example, residual GE tools in ex vivo therapies could edit undesired cells when administered to patients or induce an immune response. The purity of the gRNAs is also very important to avoid undesired off-target edits.

•

For ex vivo therapies, genetic stability studies of the modified cells should be performed. For in vivo applications, the possibility of editing nontarget cells, particularly germline cells, should also be evaluated.

•

For in vivo approaches, if there is more than one GE editing tool (e.g., a nuclease and a donor sequence), the delivery approach for each component should be described. This is an important factor to consider for potency evaluation of the combination from a regulatory perspective.

Apart from the considerations listed above, there are challenges associated with GE that need to be considered for its clinical applications. One of the challenges comes from genome variability, which limits on- and off-target predictions and may jeopardize therapies’ universality. The use of a variety of human genome sequences in the in silico predictions should help in reducing the uncertainties. Another challenge comes from the unpredictability of the DNA repair after the double-strand breaks (DSBs), which may lead to a lower than expected rate of on-target mutations or to the generation of unwanted on-target mutations.^92,93 This requires extensive characterization of each GE product to achieve a good understanding of the biological processes associated with the edition and may require the implementation of additional ad hoc testing for clinical use.

Present and future of GE therapies: A regulatory view

The pace from discovery to clinical development of precise, sequence-driven GE tools has accelerated over the past decades. This progress culminated in the development of Casgevy, the first approved medicinal product based on CRISPR/Cas.^94,95 New CRISPR/Cas-based tools that avoid the generation of DSBs and reduce off-target mutagenesis risks are currently entering clinical trials, including base⁹⁶ and prime editors.⁹⁷ Development of bridge RNAs shows promise, offering a new versatile tool for programmable GE.⁹⁸ Mitochondrial genome editors^99,100 and epigenetic editors (which do not change the genome but only the expression of certain genes)¹⁰¹ are also in development. Although some of these newer tools could, theoretically, have a reduced risk of genotoxicity, more experience is needed before regulators can allow for a reduced quality and non-clinical package for entering clinical trials.

Innovative targeted delivery systems, such as functionalized nanoparticles (e.g., with linked antibodies), are also being developed¹⁰² to reduce systemic exposure for in vivo treatments. Characterization of new GE tools and their biological activities (both on-target and off-target) will be essential to ensure their safety and efficacy. Another foreseen field for innovation is extracorporeal administration (ex vivo perfusion of autologous or allogenic organs) to avoid the systemic exposure of the patient to the GE tools.¹⁰³

GE has opened a new and exciting clinical avenue, but some uncertainties need to be resolved to unlock its full clinical potential. These advancements bring along regulatory challenges. Until the field matures, it is difficult (and probably counterproductive) to prepare specific guidelines. For now, the best regulatory agencies can do is to evaluate GE-based medicinal products using a science-based approach, assessing the benefits and risks on a case-by-case basis. Developers are encouraged to approach regulatory agencies as early as possible to discuss their clinical plans with GE tools to receive adapted advice.

CONCLUSION

Gene therapy has consolidated in the clinic, as demonstrated by the increasing number of approved GTMP and ongoing clinical trials. The wide availability of viral and nonviral vectors with an acceptable safety profile for the traditional gene addition approaches, together with the advent of GE, allows tackling diseases with an unmet medical need but also developing innovative approaches to treat diseases with alternative treatments available. However, as reviewed in this paper, manufacturing and regulatory challenges still exist, even for the most consolidated GT approaches. In the case of autologous CAR-T cell therapies, the expected use in new indications and earlier lines of treatment exposes the challenges of scaling out (as opposed to scaling up) manufacturing as well as distribution and supply chain with the traditional centralized manufacturing settings. Decentralized manufacturing could take manufacturing close to the patients, but consistency between all sites should be maintained, which poses many regulatory challenges.

rAAV vectors have limitations with respect to, for example, cargo size. In addition, manufacturing needs optimization to reduce impurities and improve yield. Other viral and nonviral vectors will probably increase their clinical use as in vivo delivery systems, but the little GMP manufacturing experience with most of these vectors (especially at high scale) will also be a challenge as demand increases.

GE is quickly entering the clinic, especially CRISPR/Cas-based technologies. Concerns still remain with regard to unintended (off-target) modifications and potential genotoxicity, which require extensive quality and preclinical characterization. As more clinical experience is gained, regulatory requirements will likely be lowered if safety is demonstrated.

In summary, clinical success with GTMP should not undermine the need for improved manufacturing and control strategies. This will help solve some of the current regulatory uncertainties, facilitating a faster development of new products and indications.

Footnotes

AUTHORS’ CONTRIBUTIONS

All authors participated in writing different sections of the article. All authors read and approved the final article.

DISCLAIMER

The opinions presented here belong exclusively to the authors and should not be considered as the official opinions of Agencia Española de Medicamentos y Productos Sanitarios or the European Medicines Agency.

AUTHOR DISCLOSURE

No competing financial interests exist for any of the authors.

FUNDING INFORMATION

No funding was received for this article.

References

Wang

, Stevens

, Flotte

. Gene therapy then and now: A look back at changes in the field over the past 25 years. Mol Ther J Am Soc Gene Ther, 2025.

Anguela

, High

. Entering the modern era of gene therapy. Annu Rev Med, 2019; 70:273–288.

CAT quarterly highlights and approved ATMPs. 2025. Available from: https://www.ema.europa.eu/en/documents/committee-report/cat-quarterly-highlights-approved-atmps-february-2025_en.pdf [Last accessed: April 29, 2025 ].

Approved cellular and gene therapy products. FDA; 2025. Available from: https://www.ema.europa.eu/en/documents/committee-report/cat-quarterly-highlights-approved-atmps-february-2025_en.pdf [Last accessed: April 29, 2025 ].

Ginn

, Mandwie

, Alexander

, et al. Gene therapy clinical trials worldwide to 2023-an update. J Gene Med, 2024; 26(8):e3721.

Regulation-1394/2007-EN-EUR-Lex. 2025. Available from: https://eur-lex.europa.eu/eli/reg/2007/1394/oj/eng [Last accessed: April 29, 2025 ].

Capolla

, Rasool

, Toffoli

, et al. CAR-T Cell Manufacturing for Hematological and Solid Tumors: From the Preclinical to Clinical Point of View. Cancer Med, 2025; 14(5):e70726.

Tran

, Tai

. Profiling AAV vector heterogeneity & contaminants using next-generation sequencing methods. 2023. Available from: https://www.insights.bio/cell-and-gene-therapy-insights/journal/article/3088/Profiling-AAV-vector-heterogeneity-contaminants-using-next-generation-sequencing-methods [Last accessed: April 29, 2025 ].

Waldman

, Fritz

, Lenardo

. A guide to cancer immunotherapy: From T cell basic science to clinical practice. Nat Rev Immunol, 2020; 20(11):651–668.

10.

Kuwana

, Asakura

, Utsunomiya

, et al. Expression of chimeric receptor composed of immunoglobulin-derived V regions and T-cell receptor-derived C regions. Biochem Biophys Res Commun, 1987; 149(3):960–968.

11.

Maher

, Brentjens

, Gunset

, et al. Human T-lymphocyte cytotoxicity and proliferation directed by a single chimeric TCRzeta/CD28 receptor. Nat Biotechnol, 2002; 20(1):70–75.

12.

Zhao

, Cao

. Engineered T cell therapy for cancer in the clinic. Front Immunol, 2019; 10:2250; doi: 10.3389/fimmu.2019.02250

13.

Eshhar

, Waks

, Gross

, et al. Specific activation and targeting of cytotoxic lymphocytes through chimeric single chains consisting of antibody-binding domains and the gamma or zeta subunits of the immunoglobulin and T-cell receptors. Proc Natl Acad Sci USA, 1993; 90(2):720–724.

14.

Scherlinger

, Nocturne

, Radic

, et al.; Club for Innovative Immunotherapies in Immune-mediated Inflammatory diseases (C3I). CAR T-cell therapy in autoimmune diseases: Where are we and where are we going? Lancet Rheumatol, 2025; 7(6):e434–e447.

15.

Aghajanian

, Kimura

, Rurik

, et al. Targeting cardiac fibrosis with engineered T cells. Nature, 2019; 573(7774):430–433.

16.

Rurik

, Tombácz

, Yadegari

, et al. CAR T cells produced in vivo to treat cardiac injury. Science, 2022; 375(6576):91–96.

17.

Eyquem

, Mansilla-Soto

, Giavridis

, et al. Targeting a CAR to the TRAC locus with CRISPR/Cas9 enhances tumour rejection. Nature, 2017; 543(7643):113–117.

18.

Shah

, Krull

, Odonnell

, et al. Promises and challenges of a decentralized CAR T-cell manufacturing model. Front Transplant, 2023; 2:1238535; doi: 10.3389/frtra.2023.1238535

19.

Guidelines on Good Manufacturing Practice specific to Advanced Therapy Medicinal Product. 2025. Available from: https://health.ec.europa.eu/document/download/e1786e5d-1587-42af-b8ff-6a810ccb6836_en?filename=pharm731_2ib_atmps_guidelines.pdf [Last accessed: April 29, 2025 ].

20.

Advanced therapy medicinal products: Overview. Available from: https://www.ema.europa.eu/en/human-regulatory-overview/advanced-therapy-medicinal-products-overview [Last accessed: April 29, 2025 ].

21.

Song

, Milone

. Pharmacology of chimeric antigen receptor-modified T cells. Annu Rev Pharmacol Toxicol, 2021; 61:805–829.

22.

Salas-Benito

, Berger

, Maus

. Stalled CARs: Mechanisms of resistance to CAR T cell therapies. Annu Rev Cancer Biol, 2023; 7(1):23–42.

23.

Meyran

, Zhu

, Butler

, et al. TSTEM-like CAR-T cells exhibit improved persistence and tumor control compared with conventional CAR-T cells in preclinical models. Sci Transl Med, 2023; 15(690):eabk1900.

24.

Rafiq

, Hackett

, Brentjens

. Engineering strategies to overcome the current roadblocks in CAR T cell therapy. Nat Rev Clin Oncol, 2020; 17(3):147–167.

25.

Meermeier

, Bergsagel

, Chesi

. Next-generation therapies for multiple myeloma. Annu Rev Cancer Biol, 2024; 8:351–371.

26.

Furqan

, Shah

. Multispecific CAR T cells deprive lymphomas of escape via antigen loss. Annu Rev Med, 2023; 74:279–291.

27.

Guedan

, Ruella

, June

. Emerging cellular therapies for cancer. Annu Rev Immunol, 2019; 37:145–171.

28.

Schmidts

, Wehrli

, Maus

. Toward better understanding and management of CAR-T cell-associated toxicity. Annu Rev Med, 2021; 72:365–382.

29.

Levine

, Pasquini

, Connolly

, et al. Unanswered questions following reports of secondary malignancies after CAR-T cell therapy. Nat Med, 2024; 30(2):338–341.

30.

Pan

, Farrukh

, Chittepu

VCSR

, et al. CAR race to cancer immunotherapy: From CAR T, CAR NK to CAR macrophage therapy. J Exp Clin Cancer Res, 2022; 41(1):119.

31.

, Zhao

, Li

, et al. From CAR-T cells to CAR-NK cells: A developing immunotherapy method for hematological malignancies. Front Oncol, 2021; 11:720501.

32.

Goldenson

, Hor

, Kaufman

. iPSC-derived natural killer cell therapies—expansion and targeting. Front Immunol, 2022; 13:841107.

33.

Hermanson

, Bendzick

, Pribyl

, et al. Induced pluripotent stem cell-derived natural killer cells for treatment of ovarian cancer. Stem Cells, 2016; 34(1):93–101.

34.

Karagiannis

, Kim

S-I

. iPSC-derived natural killer cells for cancer immunotherapy. Mol Cells, 2021; 44(8):541–548.

35.

National Heart, Lung, and Blood Institute (NHLBI). A phase I study of HERV-E TCR transduced autologous T cells in patients with metastatic clear cell renal cell carcinoma. clinicaltrials.gov; 2024. Available from: https://clinicaltrials.gov/study/NCT03354390 [Last accessed: March 27, 2025 ].

36.

Labanieh

, Mackall

. CAR immune cells: Design principles, resistance and the next generation. Nature, 2023; 614(7949):635–648.

37.

JCI Insight—Efficacy of intracellular immune checkpoint-silenced DC vaccine. 2025. Available from: https://insight.jci.org/articles/view/98368 [Last accessed: March 27, 2025 ].

38.

Lee

, Hong

S-H

. Hematopoietic stem cells and their roles in tissue regeneration. Int J Stem Cells, 2020; 13(1):1–12.

39.

Stein

, Ott

, Schultze-Strasser

, et al. Genomic instability and myelodysplasia with monosomy 7 consequent to EVI1 activation after gene therapy for chronic granulomatous disease. Nat Med, 2010; 16(2):198–204.

40.

Hacein-Bey-Abina

, Von Kalle

, Schmidt

, et al. LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-X1. Science, 2003; 302(5644):415–419.

41.

Cancer cases in bluebird’s gene therapy trials. Nat Biotechnol, 2024; 42:1626.

42.

A systematic review and meta-analysis of gene therapy with hematopoietic stem and progenitor cells for monogenic disorders. Nature Communications. 2025. Available from: https://www-nature-com-s.web.bisu.edu.cn/articles/s41467-022-28762-2 [Last accessed: March 27, 2025 ].

43.

Developing a new class of engineered live bacterial therapeutics to treat human diseases. Nature Communications. 2025. Available from: https://www-nature-com-s.web.bisu.edu.cn/articles/s41467-020-15508-1 [Last accessed: March 27, 2025 ].

44.

Bennett

, Mietzsch

, Agbandje-McKenna

. Understanding capsid assembly and genome packaging for adeno-associated viruses. Future Virol, 2017; 12(6):283–297.

45.

, Yang

, Colosi

. Effect of genome size on AAV vector packaging. Mol Ther, 2010; 18(1):80–86.

46.

, Tan

, Zhang

, et al. AAV-mediated gene therapy restores hearing in patients with DFNB9 deafness. Adv Sci Weinh Baden-Wurtt Ger, 2024; 11:e2306788.

47.

Liang

S-Q

, Navia

, Ramseier

, et al. AAV5 delivery of CRISPR/Cas9 mediates genome editing in the lungs of young Rhesus monkeys. Hum Gene Ther, 2024; 35(19–20):814–824.

48.

Tasfaout

, Halbert

, McMillen

, et al. Split intein-mediated protein trans-splicing to express large dystrophins. Nature, 2024; 632(8023):192–200.

49.

Martino

, Herzog

, Anegon

, et al. Measuring immune responses to recombinant AAV gene transfer. Methods Mol Biol Clifton NJ, 2011; 807:259–272.

50.

, Töpfer

, Lin

S-W

, et al. Self-complementary AAVs induce more potent transgene product-specific immune responses compared to a single-stranded genome. Mol Ther, 2012; 20(3):572–579.

51.

Zolgensma. European Medicines Agency (EMA). 2020. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/zolgensma [Last accessed: April 29, 2025 ].

52.

Gao

, Vandenberghe

, Alvira

, et al. Clades of Adeno-associated viruses are widely disseminated in human tissues. J Virol, 2004; 78(12):6381–6388.

53.

Xie

, Chen

, Ma

, et al. Improved gene therapy for spinal muscular atrophy in mice using codon-optimized hSMN1 transgene and hSMN1 gene-derived promotor. EMBO Mol Med, 2024; 16(4):945–965.

54.

Pipe

. Functional roles of the factor VIII B domain. Haemoph Off J World Fed Hemoph, 2009; 15:1187–1196.

55.

Rumachik

, Malaker

, Poweleit

, et al. Methods matter: Standard production platforms for recombinant AAV produce chemically and functionally distinct vectors. Mol Ther Methods Clin Dev, 2020; 18:98–118.

56.

Laforet

. Thrombotic microangiopathy associated with systemic adeno-associated virus gene transfer: Review of reported cases. Hum Gene Ther, 2025; 36(3–4):64–76.

57.

Duan

. Lethal immunotoxicity in high-dose systemic AAV therapy. Mol Ther, 2023; 31(11):3123–3126.

58.

Tai

PWL

. Integration of gene therapy vectors: A risk factor for tumorigenesis or another commensal property of adeno-associated viruses that benefits long-term transgene expression? Hum Gene Ther, 2023; 34(21–22):1074–1076.

59.

Sabatino

, Bushman

, Chandler

, et al.; American Society of Gene and Cell Therapy (ASGCT) Working Group on AAV Integration. Evaluating the state of the science for adeno-associated virus integration: An integrated perspective. Mol Ther, 2022; 30(8):2646–2663.

60.

Qin

, Tang

, Chen

, et al. mRNA-based therapeutics: Powerful and versatile tools to combat diseases. Signal Transduct Target Ther, 2022; 7(1):166–135.

61.

Zhang

, Tang

, Chen

, et al. mRNA vaccines in disease prevention and treatment. Signal Transduct Target Ther, 2023; 8(1):365.

62.

Conde

, Langer

, Rueff

. mRNA therapy at the convergence of genetics and nanomedicine. Nat Nanotechnol, 2023; 18(6):537–540.

63.

Sahin

, Karikó

, Türeci

. mRNA-based therapeutics—developing a new class of drugs. Nat Rev Drug Discov, 2014; 13(10):759–780.

64.

, Wang

, Peng

, et al. mRNA vaccine in cancer therapy: Current advance and future outlook. Clin Transl Med, 2023; 13(8):e1384.

65.

Lorentzen

, Haanen

, Met

, et al. Clinical advances and ongoing trials on mRNA vaccines for cancer treatment. Lancet Oncol, 2022; 23(10):e450–e458.

66.

Xie

, Shen

, Gao

, et al. Neoantigens: Promising targets for cancer therapy. Signal Transduct Target Ther, 2023; 8(1):9.

67.

Meister

, Look

, Roth

, et al. Multifunctional mRNA-based CAR T cells display promising antitumor activity against glioblastoma. Clin Cancer Res, 2022; 28(21):4747–4756.

68.

Xiao

, Lai

, Yuan

, et al. mRNA-based chimeric antigen receptor T cell therapy: Basic principles, recent advances and future directions. Interdiscip Med, 2024; 2:e20230036.

69.

Yang

, Mixich

, Boonstra

, et al. Polymer-based mRNA delivery strategies for advanced therapies. Adv Healthc Mater, 2023; 12(15):e2202688.

70.

ICH Q6 5B. Specifications: Test procedures and acceptance criteria for biotechnological/biological products. 2025. Available from: https://www.ema.europa.eu/en/documents/scientific-guideline/ich-q-6-b-test-procedures-and-acceptance-criteria-biotechnologicalbiological-products-step-5_en.pdf [Last accessed: April 29, 2025 ].

71.

Chen

, Liu

, Guo

, et al. Branched chemically modified poly(A) tails enhance the translation capacity of mRNA. Nat Biotechnol, 2025; 43(2):194–203.

72.

Packer

, Gyawali

, Yerabolu

, et al. A novel mechanism for the loss of mRNA activity in lipid nanoparticle delivery systems. Nat Commun, 2021; 12(1):6777.

73.

Rohner

, Yang

, Foo

, et al. Unlocking the promise of mRNA therapeutics. Nat Biotechnol, 2022; 40(11):1586–1600.

74.

Santer

, Bär

, Thum

. Circular RNAs: A novel class of functional RNA molecules with a therapeutic perspective. Mol Ther, 2019; 27(8):1350–1363.

75.

Wesselhoeft

, Kowalski

, Parker-Hale

, et al. RNA circularization diminishes immunogenicity and can extend translation duration in vivo. Mol Cell, 2019; 74(3):508–520.e4.

76.

Unti

, Jaffrey

. Highly efficient cellular expression of circular mRNA enables prolonged protein expression. Cell Chem Biol, 2024; 31(1):163–176.e5.

77.

Chen

, Wang

, Belk

, et al. Engineering circular RNA for enhanced protein production. Nat Biotechnol, 2023; 41(2):262–272.

78.

Mulroney

, Pöyry

, Yam-Puc

, et al. N1-methylpseudouridylation of mRNA causes +1 ribosomal frameshifting. Nature, 2024; 625(7993):189–194.

79.

Papukashvili

, Rcheulishvili

, Liu

, et al. Self-amplifying RNA approach for protein replacement therapy. Int J Mol Sci, 2022; 23(21):12884.

80.

Lundstrom

. Viral vectors in gene therapy: Where do we stand in 2023? Viruses, 2023; 15(3):698.

81.

Applications for new human medicines under evaluation. Ref. num. EMA/51244/2025. February 2025. European Medicines Agency. 2025. Available from: https://eur-lex.europa.eu/eli/reg/2007/1394/oj/eng [Last accessed: April 29, 2025 ].

82.

Yin

, Wu

, Han

, et al. Immune-related adverse events of immune checkpoint inhibitors: A review. Front Immunol, 2023; 14:1167975.

83.

Palanivelu

, Liu

C-H

, Lin

L-T

. Immunogenic cell death: The cornerstone of oncolytic viro-immunotherapy. Front Immunol, 2022; 13:1038226.

84.

Lauer

, Beil

. Oncolytic viruses: Challenges and considerations in an evolving clinical landscape. Future Oncol, 2022; doi: 10.2217/fon-2022-0440

85.

Rehman

, Silk

, Kane

, et al. Into the clinic: Talimogene laherparepvec (T-VEC), a first-in-class intratumoral oncolytic viral therapy. J Immunother Cancer, 2016; 4:53.

86.

Porter

, Rosewell Shaw

, Jung

, et al. Oncolytic adenovirus armed with BiTE, cytokine, and checkpoint inhibitor enables CAR T cells to control the growth of heterogeneous tumors. Mol Ther, 2020; 28(5):1251–1262.

87.

Vyjuvek. European Medicines Agency (EMA). 2025. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/vyjuvek [Last accessed: March 27, 2025 ].

88.

D-Y

, Li

L-Q

, Liu

J-JG

. Nucleases in gene-editing technologies: Past and prologue. Natl Sci Open, 2023:2; doi: 10.1360/nso/20220067

89.

van Haasteren

, Li

, Scheideler

, et al. The delivery challenge: Fulfilling the promise of therapeutic genome editing. Nat Biotechnol, 2020; 38(7):845–855.

90.

EMA/246400/2021 Questions and answers on the principles of GMP for the manufacturing of starting materials of biological origin used to transfer genetic material for the manufacturing of ATMPs. 2025. Available from: https://www.ema.europa.eu/en/documents/other/questions-and-answers-principles-gmp-manufacturing-starting-materials-biological-origin-used-transfer-genetic-material-manufacturing-atmps_en.pdf [Last accessed: April 29, 2025 ].

91.

Zhang

, Li

, Sun

, et al. Perfecting targeting in CRISPR. Annu Rev Genet, 2021; 55:453–477.

92.

Rasul

, Hussen

, Salihi

, et al. Strategies to overcome the main challenges of the use of CRISPR/Cas9 as a replacement for cancer therapy. Mol Cancer, 2022; 21(1):64.

93.

Teboul

, Herault

, Wells

, et al. Variability in genome editing outcomes: Challenges for research reproducibility and clinical safety. Mol Ther, 2020; 28(6):1422–1431.

94.

Frangoul

, Locatelli

, Sharma

, et al.; CLIMB SCD-121 Study Group. Exagamglogene autotemcel for severe sickle cell disease. N Engl J Med, 2024; 390(18):1649–1662; doi: 10.1056/NEJMoa2309676

95.

Locatelli

, Lang

, Wall

, et al.; CLIMB THAL-111 Study Group. Exagamglogene autotemcel for transfusion-dependent β-thalassemia. N Engl J Med, 2024; 390(18):1663–1676.

96.

Gaudelli

, Komor

, Rees

, et al. Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage. Nature, 2017; 551(7681):464–471.

97.

Anzalone

, Randolph

, Davis

, et al. Search-and-replace genome editing without double-strand breaks or donor DNA. Nature, 2019; 576(7785):149–157.

98.

Durrant

, Perry

, Pai

, et al. Bridge RNAs direct programmable recombination of target and donor DNA. Nature, 2024; 630(8018):984–993.

99.

Mok

, de Moraes

, Zeng

, et al. A bacterial cytidine deaminase toxin enables CRISPR-free mitochondrial base editing. Nature, 2020; 583(7817):631–637.

100.

Cho

S-I

, Lee

, Mok

, et al. Targeted A-to-G base editing in human mitochondrial DNA with programmable deaminases. Cell, 2022; 185(10):1764–1776.e12.

101.

, Larson

, Gilbert

, et al. Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell, 2013; 152(5):1173–1183.

102.

Yang

, Im

, Chung

, et al. An Antibody-CRISPR/Cas conjugate platform for target-specific delivery and gene editing in cancer. Adv Sci Weinh Baden-Wurtt Ger, 2024; 11:e2308763.

103.

Mesaki

, Yamamoto

, Juvet

, et al. CRISPR-Cas genome editing in ex vivo human lungs to rewire the translational path of genome-targeting therapeutics. Hum Gene Ther, 2024; 35(11–12):374–387.