Acknowledgments: to Professor Evelyne Kohli and all the reviewers for their careful proofreading and suggestions for amendments.

The discoveries of the Austrian monk Gregor Mendel in 1865, and especially their significant rediscovery at the beginning of the 20th century by Hugo de Vries, Carl Correns, and Erich Tschermak von Seysenegg, marked the beginnings of what would later be called genetics.¹ Initially focused on understanding heredity, this emerging science quickly incorporated the concept of biological mutation. In the 1920s and 1930s, Hermann Joseph Muller advanced this field by discovering the mutagenic effect of ionizing radiation, further developing his theory of induced mutagenesis.² At that time, Muller was a defender of eugenics,³ a widely accepted idea in the Anglo-Saxon world.⁴ In this context, as genetics was emerging and eugenics was progressing, science fiction authors began to perceive the potential and especially the danger of these discoveries. Thus, the idea of genetic manipulation of human beings took root in novels rather than in laboratories. The most famous novel—and undoubtedly the first work to evoke the genetic manipulation of humans—was Brave New World by Aldous Huxley,⁵ published in 1932.⁶ During this period, Huxley, the author, became aware of the ethical risks and social consequences of the genetic manipulation of humans by humans. In his dystopian work, he depicted a totalitarian eugenic society, where genetically modified and conditioned individuals lived to occupy specific roles, without any freedom. It is likely that Aldous Huxley was influenced by his brother, Julian Huxley,⁷ an evolutionary biologist, who was one of the pioneers of transhumanism.⁸

Meanwhile, the “one gene/one enzyme” principle,⁹ established by George Beadle and Edward Tatum in 1941,¹⁰ and especially the identification of DNA as the carrier of genetic information in 1953, by James Watson, Francis Crick, Rosalind Franklin and Maurice Wilkins¹¹ paved the way for a deeper understanding of genetics and molecular biology. The theoretical prediction of mRNA in 1961 by the French scientists and Nobel laureates François Jacob and Jacques Monod, along with their pioneering work on genetic regulatory mechanisms, further advanced this knowledge.¹² This was later confirmed experimentally in the same year by François Jacob, Sydney Brenner, and Matthew Meselson.¹³

However, while science fiction authors had been discussing the genetic manipulation of human beings since the 1930s, it was not until 1963 that a scientist, Joshua Lederberg, raised this possibility from a scientific perspective.¹⁴ These ideas were further developed in 1970 by Stanfield Rogers,¹⁵ and later in 1972 by Theodore Friedmann and Richard Roblin,¹⁶ who proposed replacing defective DNA with exogenous¹⁷ DNA.

From Concept to Reality

In the 1970s, Stanfield Rogers attempted genetic manipulation by administering the Shope papilloma virus to patients suffering from an arginase enzyme deficiency,¹⁸ with the aim of restoring arginase production, as the virus coded for the enzyme. However, this attempt proved to be a failure.¹⁹ In the 1980s, Martin Cline conducted an experiment aimed at treating thalassemia, a genetic blood disorder that affects hemoglobin production, leading to anemia and other serious health issues. He attempted to insert a healthy β-globin gene (a key component of hemoglobin) into bone marrow cells. The modified cells were subsequently transplanted with the objective of producing normal hemoglobin in patients. Not only was this a failure, but the clinical trial was also subjected to severe ethical and scientific criticism.²⁰

In 1989, S.A. Rosenberg, K. W. Culver, W. F. Anderson, and R. M. Blaese conducted an experiment utilizing a modified retrovirus, as a tool, to introduce a specific gene into immune cells known as tumor-infiltrating lymphocytes (TILs). These immune cells, which naturally target cancer cells, were genetically modified to carry a gene that made them resistant to the antibiotic neomycin. This process was used to track the modified cells in the body and ensure that the introduced gene was successfully incorporated. Although the neomycin resistance gene itself didn’t treat cancer, the technique marked a significant advancement toward the use of gene therapy to modify immune cells for cancer treatment.²¹

The first true human gene therapy trials for therapeutic purposes began in the 1990s, and notably we can cite the trial by Culver, Anderson and Blaese.²² They sought to treat severe combined immunodeficiency (SCID) linked to mutations in the gene coding for adenosine deaminase (ADA), an enzyme essential for lymphocyte (white blood cells) function. This condition compromises the immune system from birth, rendering it non-functional and necessitating that affected individuals live in a sterile environment. As a result, this condition is colloquially referred to as “bubble baby disease.” Two children underwent treatment, with gene therapy proving partially effective in one of the cases.²³ Subsequently, the number of clinical trials increased significantly, rising from 1 in 1989 to 67 in 1995, and reaching 96 by 2000.²⁴ These various trials ultimately led to Strimvelis®, a gene therapy medicinal product authorized by the European Union in 2016.

In 2000, one of the most significant clinical trials took place, conducted by Prof. Alain Fischer, which enabled the treatment of another “bubble baby disease” (X-linked SCID / SCID-X1).²⁵ This trial treated twenty children, five of whom unfortunately later developed leukemia, due to insertional mutagenesis mediated by the vector used.²⁶ The trial was halted to investigate the underlying cause of the conditions and to administer cancer treatment to the patients, while both the French government and the U.S. FDA suspended similar trials.²⁷ However, the vector was subsequently changed and new clinical trials were conducted, many years later, yielding favorable results.²⁸

From a Need for Regulation to the Emergence of a Definition

In response to growing activity in the field of gene therapy, driven by increasingly promising trial results and rapid advancements, the European legislator introduced an initial definition of gene therapy medicinal products through Directive 2003/63/EC²⁹ amending Directive 2001/83/EC,³⁰ also referred to as the “Community code relating to medicinal products for human use.”
The 2003 definition was written as follows:

gene therapy medicinal product shall mean a product obtained through a set of manufacturing processes aimed at the transfer, to be performed either in vivo or ex vivo, of a prophylactic, diagnostic or therapeutic gene (i.e., a piece of nucleic acid), to human/animal cells and its subsequent expression in vivo. The gene transfer involves an expression system contained in a delivery system known as a vector, which can be of viral, as well as non-viral origin. The vector can also be included in a human or animal cell.

At this time, the first gene therapy medicinal product (GTMP), Gendicine®³¹ was authorized, not in Europe, but in China for the treatment of head and neck cancers.³²

In 2007, European Regulation 1394/2007³³ established the category of advanced therapy medicinal product (ATMP), which encompasses gene therapy medicinal products, somatic cell therapy medicinal products, and tissue engineered products. Additionally, combined ATMPs are also included. However, this regulation did not change the definition of gene therapy medicinal products as written in 2003. The definition was subsequently amended by Commission Directive 2009/120/EC, issued on 14 September 2009.³⁴ The third recital of the aforementioned directive explicitly states: “The definitions and detailed scientific and technical requirements for gene therapy medicinal products and somatic cell therapy medicinal products should be updated.”³⁵

Accordingly, the definition was revised as follows:

Gene therapy medicinal product means a biological medicinal product which has the following characteristics:
a) it contains an active substance which contains or consists of a recombinant nucleic acid used in or administered to human beings with a view to regulating, repairing, replacing, adding or deleting a genetic sequence;
b) its therapeutic, prophylactic or diagnostic effect relates directly to the recombinant nucleic acid sequence it contains, or to the product of genetic expression of this sequence.
Gene therapy medicinal products shall not include vaccines against infectious diseases.

The analysis of this definition allows us to identify several criteria that must all be satisfied:³⁶

• A content or nature criterion: it must be a biological medicinal product consisting of recombinant nucleic acid.
• A function or objective criterion: it must be intended to regulate, repair, replace, add, or delete a genetic sequence.
• A means or mode of action criterion: its therapeutic, prophylactic, or diagnostic effect must directly depend on the sequence of recombinant nucleic acid used.
• A negative destination criterion: the product cannot be intended to prevent infectious diseases (i.e., it cannot be a vaccine against infectious diseases).

The first ATMP meeting these criteria and authorized within the European Union in 2012 was Glybera® Alipogene tiparvovec.³⁷ Today, many other products have reached the market, notably CAR-T cells (6 on the market as of September 2024, see Table 1). It is important to highlight that the limited number of products withdrawn were not due to safety concerns, but rather for commercial reasons such as challenges with the reimbursement negotiations.³⁸

Trade name	INN	Indication	MA date	Marketed status
Casgevy®	Exagamglogene autotemcel	beta-Thalassemia and Sickle Cell disease	2024	yes
Hemgenix®	etranacogene dezaparvovec	Hemophilia B	2023	yes
Roctavian®	valoctocogene roxaparvovec	Hemophilia A	2022	yes
Upstaza®	eladocagene exuparvovec	Hemophilia A	2022	yes
Carvykti®	ciltacabtagene autoleucel (CAR-T cell)	Multiple myeloma	2022	yes
Breyanzi®	lisocabtagene maraleucel (CAR-T cell)	lymphomas – diffuse large B-cell – high-grade B-cell – primary mediastinal large B-cell – follicular grade 3B	2022	yes
Abecma®	idecabtagene vicleucel (CAR-T cell)	Multiple myeloma	2021	yes
Skysona®	elivaldogene autotemcel	treatment of early cerebral adrenoleukodystrophy	2021	Withdrawn in November 2021 at the request of the MA holder
Libmeldy®	atidarsagene autotemcel	metachromatic leukodystrophy	2020	yes
Tecartus®	Brexucabtagene autoleucel (CAR-T cell)	mantle cell lymphoma; acute lymphoblastic leukemia	2020	yes
Zolgensma®	onasemnogene abeparvovec	Spinal muscular atrophy 5q	2020	yes
Zynteglo®	betibeglogene autotemcel	Transfusion-dependent β-thalassemia transfusions	2019	Withdrawn in May 2022 at the request of the MA holder
Luxturna®	voretigene neparvovec	Leber’s amaurosis	2018	yes
Yescarta®	Axicabtagene ciloleucel (CAR-T cell)	lymphomas: – diffuse large B-cell; – high-grade B-cell; – primary mediastinal large B-cell; – follicular	2018	yes
Kymriah®	Tisagenlecleucel (CAR-T cell)	acute lymphoblastic leukemia diffuse: – diffuse large B-cell – follicular	2018	yes
Strimvelis®	Non applicable	Severe combined immunodeficiency due to adenosine deaminase deficiency (ADA-SCID)	2016	yes
Imlygic®	talimogene laherparepvec	melanoma	2015	yes
Glybera®	alipogene tiparvovec	hyperlipoproteinemia	2012	Withdrawn in October 2017 at the request of the MA holder.

An Ossified Definition

However, while the initial legal definition of ATMP evolved between 2003 and 2009, since 2009, it has remained unchanged. Thus, 15 years have passed, and despite significant advancements in scientific knowledge and techniques, this definition has remained stagnant, raising several questions. However, the wording of the definition of a gene therapy medicinal product (GTMP) has a direct impact on whether or not products can be legally categorized as gene therapy medicinal products. Some technologies that are emerging, or will soon emerge, such as CRISPR/Cas9 administered in vivo or small RNAs, do not fully meet at least one criterion of the GTMP definition. As a result, if they do not fall under this definition, they are considered as “biological medicinal products” or even “chemical medicinal products,” and the regulatory requirements for their market authorization differ accordingly. Indeed, categorization as GTMP entails a number of regulatory constraints to be applied in terms of design, toxicology, pharmacology, manufacturing, risk management, etc.

The wording of the definition raises a number of questions relating to the intrinsic nature of gene therapy medicinal products (I), particularly the fact that, at present, all the criteria of the definition (namely the biological nature and the recombinant nucleic acid criteria) must be strictly met for a product to be classified as a GTMP.

The wording of the definition also raises questions regarding the mode of action and the function of gene therapy medicinal products (II). Some products that would seem to fall within the scope of gene therapy are in fact excluded, such as medicinal products incorporating “suicide genes” or medicinal product candidates that use direct genome editing tools. Additionally, by law, vaccines against infectious diseases based on nucleic acids are also excluded.

It therefore appears that the definition of GTMP needs to be revised (III), which is important in light of the numerous innovations coming to market that would not be classified as GTMPs. If these innovations reach the market, they could fall under a legal categorization that may not protect patients as effectively as the GTMP classification.

I. The Nature of Gene Therapy Medicinal Products, As Imposed by the Definition, Could Compromise Patient Safety and, to a Lesser Extent, Obstruct Innovation

The definition of Gene Therapy Medicinal Products (GTMPs) is very strict and imposes specific characteristics on a drug that is a candidate for such categorization. Specifically, the active substance must be of biological origin (A), and this substance must consist of recombinant nucleic acid (B). Furthermore, the inclusion of all nucleic acids within the scope of the definition, particularly RNA, is called into question.

A) The Imposed Biological Nature

The definition of a GTMP states that it is a biological medicinal product. This requires GTMPs to comply with the criteria inherent to this status.

The biological medicinal product was first defined by the aforementioned directive 2003/63/EC, and is now defined as:

a product, the active substance of which is a biological substance. A biological substance is a substance that is produced by or extracted from a biological source and that needs for its characterization and the determination of its quality, a combination of physicochemical-biological testing, together with the production process and its control.

This is a definition with cumulative criteria. In fact, in 2012, the French Conseil d’État³⁹ ruled that it is not enough for the active substance of the medicinal product to be derived from a biological source for it to be qualified as a biological medicinal product; it must also undergo physicochemical-biological tests.⁴⁰

Thus, a GTMP must meet these criteria: its active substance must be biological, meaning it must:

• Be produced from or extracted from a biological source;
• Its characterization and determination of quality require a combination of physicochemical-biological tests, as well as knowledge of its manufacturing process and control.

1) The Obstacle of the Biological Nature

Even before discussing the second sub-criterion of content, i.e., that a GTMP (in addition to being biological) consists of or contains recombinant nucleic acid, it is worth noting that the status of a biological medicinal product excludes all molecules derived from chemical synthesis. In other words, nowadays, no medicinal product derived from chemical synthesis can currently claim the status of a GTMP.

However, some of these chemical molecules meet the function criteria of the definition, such as regulation, repair, or replacement of a gene sequence. Indeed, some chemical molecules are capable of modifying the expression of a gene sequence, such as Evrysdi® (risdiplam), which can modify the splicing of mRNA and thus increase the expression of the SMN (Survival of Motor Neuron) protein, which is deficient in spinal muscular atrophy.⁴¹

In this sense, by acting through the regulation of a gene sequence, they would indeed be considered gene therapy medicinal products (GTMPs) according to the functional criteria. However, as they do not adhere to the imposed nature of being a biological molecule (and a nucleic acid), they cannot claim this categorization. Some may argue that older anti-cancer molecules, which are not biological, are also capable of interacting with gene sequences, particularly DNA. For example, intercalating agents such as doxorubicin or alkylating agents such as cyclophosphamide interact directly with DNA, but their action is destructive to DNA, preventing replication. Newer chemical molecules, such as risdiplam, have a much more precise and targeted action, which is in the spirit of gene therapy. Their exclusion from the GTMP category raises questions about their evaluation. Indeed, the regulatory framework for GTMPs, beyond that of ATMPs, is undoubtedly the most stringent, particularly in terms of the evaluation requirements outlined in the guidelines. Therefore, it seems imperative that these treatments follow all or part of the GTMP guidelines to ensure a high level of safety for patients.

2) The Biological or Chemical Nature of Nucleic Acids

According to the GTMP definition, the active substance is a recombinant nucleic acid. In other words, it can be DNA or RNA⁴² or, hypothetically, a chimeric⁴³ combination of the two (chimeric polynucleotides).⁴⁴

This means that recombinant nucleic acids can only be obtained from a biological source. To date, most nucleic acids are indeed produced from or extracted from a biological source. In the case of RNA, in vitro transcription is considered involving a biological source, as the Committee for Advanced Therapies (CAT) has classified several candidate drugs using in vitro transcribed RNA as GTMPs.⁴⁵

The problem arises with chemically synthesized nucleic acids. It is possible to synthesize DNA or RNA without using a biological system, but by using a chemical system. However, although it is technologically possible to produce DNA and RNA by chemical synthesis, the cost of doing so remains significant.⁴⁶ These costs are expected to fall in the coming years, making the industrial use of these chemically synthesized nucleic acids possible. As a result, they could potentially be incorporated into medicines alongside biological nucleic acids. However, there is a major problem: since these nucleic acids are not biological, they do not meet the criteria of the legal definition of a gene therapy medicinal product. In other words, a medicinal product containing a synthetic nucleic acid as an active substance could meet all the GTMP criteria except for being biological. Consequently, the GTMP guidelines would not apply, as would certain safety requirements.

While the production of long nucleic acids by chemical synthesis is still in its early stages, the production of short polynucleotides is already a reality. In particular, small interfering RNAs⁴⁷ (siRNAs) and small activating RNAs⁴⁸ (saRNAs) are such molecules produced by chemical synthesis.

Although these are RNA molecules, they are small and chemically synthesized⁴⁹ and therefore cannot claim GTMP status; they would be classified as “simple” medicinal products. However, new siRNAs or saRNAs can be produced biologically, known as BioRNAs (bioengineered RNA) (previously known as BERAs for biological bioengineered RNA agents).⁵⁰ As these are biological nucleic acids, they could at least be classified as a biological medicinal product, and maybe as GTMP. This could lead to identical siRNAs or saRNAs on the market in terms of sequence, with the former being simple medicinal products and the latter being biological medicinal products and possibly GTMPs. This lack of consistency raises patient safety concerns as these two types of medicinal products would not have followed the same guidelines.⁵¹

In conclusion, two medicinal products with similar functions could appear on the market, one being a GTMP because the nucleic acid is of biological origin, and the other being a “simple” medicinal product. In addition to the inconsistency that this case presents, the trials outlined in the Gene Therapy Medicinal Products Guideline,⁵² in particular for insertional mutagenesis, would not necessarily be carried out, raising concerns about patient safety. One might imagine that the EMA would at least recommend these trials, provided that the manufacturer uses the centralized procedure (CP), which is rarely the case unless the procedure is mandatory.⁵³

The inherently biological nature of GTMPs is not the only problem with the definition; it also states, without further explanation, that the nucleic acid must be recombinant, without detailing this specific point.

B) A Nucleic Acid Required to Be Recombinant

The second criterion that must be met to be categorized as a GTMP is the recombination of the nucleic acid (it should be noted that this criterion is cumulative the biological nature criterion). So, the—biological—substance is necessarily a recombinant nucleic acid (DNA, RNA or chimeric construct, as mentioned above).

1) The Absence of a Legal Definition of a Nucleic Acid

The first difficulty arises from the very concept of nucleic acid. Nucleic acids are macromolecules composed of linear chains of nucleotides linked together by phosphodiester bonds; depending on the nature of the sugar (ose) constituting the nucleotides, a distinction is made between DNA and RNA.⁵⁴

However, beyond these points of consensus, some definitions add further characteristics. Some authors consider nucleic acids to be of high molecular weight,⁵⁵ and some specify a minimum molecular weight of 20 kDa⁵⁶ or 25 kDa.⁵⁷ It is also sometimes stated that nucleic acids necessarily consist of “long” chains of nucleotides⁵⁸ or many nucleotides,⁵⁹ without specifying the number.

In the absence of a legal definition of a nucleic acid, the classification of certain molecules is complicated. For example, are certain molecules of interest, such as dinucleotides⁶⁰ (two nucleotides), nucleic acids? Their function could be that of gene therapy, particularly in terms of repairing or regulating a gene sequence.

The same question could be asked about small non-coding RNAs, such as the biological siRNAs and saRNAs (BioRNAs) mentioned above, which consist of about twenty nucleotides. Is that enough to be considered a nucleic acid under the GTMP definition? Their molecular mass is generally less than 20 kDa. However, it can be argued that the fact that siRNAs and saRNAs are RNAs, even though they are small (less than 20 kDa), makes them de facto nucleic acids. Furthermore, most researchers classify siRNAs and saRNAs as nucleic acids.⁶¹ It would therefore seem logical that, for these small molecules, classification as GTMPs should be required to ensure the highest level of safety for patients.

It should also be noted that long non-coding RNAs with a size of more than 200 nucleotides are appearing in the literature.⁶² Among their various functions is the biological regulation of a genetic sequence,⁶³ which could bring them within the definition of GTMP. The situation is therefore complex, and regulators need to clarify the concept of nucleic acid.

2) The Need to Specify the Concept of Recombination

The use of the term “recombinant” is not obvious, given the lack of scientific consensus on the concept of recombination. In fact, it would have been useful to have a legal definition or, alternatively, a definition in guidelines.

Genetic recombination is a natural process involving the exchange of a sequence between two fragments of DNA or RNA nucleic acid (for example, in RNA viruses). The nucleic acid produced is therefore recombinant. However, this definition does not seem to apply to gene therapy medicinal products. Indeed, the CAT has ruled that an oncolytic chimeric adenovirus “simply” obtained by a bioselection process is not an ATMP, as it has no added genetic material and therefore its therapeutic effect is not mediated by any recombinant nucleic acid.⁶⁴ We can therefore deduce from this opinion that the term “recombinant” means that a nucleic acid is not recombinant if it has undergone natural recombination, and that the nucleic acid must be artificially manipulated to become recombinant.

This seems to be confirmed by Annex IA of Directive 2001/18/EC on the deliberate release into the environment of genetically modified organisms, which provides a definition of recombinant DNA techniques:

the formation of new combinations of genetic material by the insertion of nucleic acid molecules produced by whatever means outside an organism, into any virus, bacterial plasmid or other vector system and their incorporation into a host organism in which they do not naturally occur but in which they are capable of continued propagation.

It therefore seems important to distinguish natural recombination from artificial recombination, which is achieved by human intervention.⁶⁵ In other words, if the biological active substance is a naturally recombinant nucleic acid, the product will be classified as a biological medicinal product. However, if the recombination is artificial, it will be considered a GTMP. Once again, take note that GTMP status carries regulatory requirements for evaluation that are far more stringent than those for a simple biological medicinal product.

3) The Exclusion of Genome-Editing Tools

The legally imposed nature of nucleic acid for GTMP leads to another very important issue: it effectively excludes direct genome editing tools such as CRISPR/Cas9, transcription activator-like effector nucleases (Talens), zinc finger nucleases (ZFNs), EMRHE meganucleases, and so on.⁶⁶ With regard to CRISPR/Cas9, the CAT decided in 2021 that nanoparticles containing gene editing tools (CRISPR/Cas9 and simple guide RNAs) is not ATMPs and in this case not GTMPs. It is true that the tool integrates a strand of RNA, a nucleic acid, but this strand is generally produced by chemical synthesis… and above all, the action is carried out by the enzyme. The same applies to Talens and ZFNs, which are enzymes, biological enzymes, but not recombinant nucleic acids. Nevertheless, these tools can serve as simple reagents in the manufacturing of a GTMP, but they can also be considered medicinal products in their own right when directly injected via a lipid nanoparticle. In this case, due to their intrinsic functions and the associated risks, they should be classified as GTMPs. Unfortunately, this is not the case due to their nature (they are not recombinant nucleic acids). In this particular case, the opposition between science and law is an aberration. On the other hand, and this is a subtle point, if the patient is injected with mRNA encoding CRISPR/Cas9 or other tools, then yes, this is a GTMP, as the CAT ruled in 2021 and 2023.⁶⁷

Thus, the biological nature of GTMP, as enshrined in the legal definition, raises significant issues regarding the potential production of nucleic acids by chemical synthesis. Similarly, the “nucleic acid” nature excludes genome-editing tools from GTMP status, and the term “nucleic acid” itself should be defined. Furthermore, the lack of a precise definition of “recombinant” could also pose challenges.

The nature of the product is not the only difficulty in defining GTMP. In fact, the parameters of function and mode of action described in the legal definition raise questions about current and future technological developments. So, the mode of action of this type of medicinal product, which is also described in the definition, is a matter of some concern.

II. Questionable Mode of Action and Function for Gene Therapy Medicinal Products

Some products that would seem to fall within the scope of gene therapy are, in fact, excluded. This is due to the fact that the legal definition of gene therapy medicinal products requires that the therapeutic, prophylactic, or diagnostic effect is directly dependent on the nucleic acid sequence. This raises the question of the exclusion from the scope of gene therapy medicinal products of certain genetically modified cells whose modification does not have a direct therapeutic, prophylactic, or diagnostic effect (A). Additionally, vaccines against infectious diseases are excluded from the scope of gene therapy medicinal products, while cancer treatments are not (B).

A) The Uncertainties of the Function and Mode of Action Criteria

The GTMP includes in its definition the function of “regulating, repairing, replacing, adding or deleting a genetic sequence.”

As we have seen, certain molecules derived from chemical synthesis are capable of regulation or repair. Similarly, genome-editing tools, which are also not classified as GTMPs, can replace, add, or delete a sequence. This demonstrates that function alone does not define the concept of GTMPs, and that the nature of the active substance is indeed predominant.

But even if the criterion of the nature of the medicinal product is met, it is sometimes difficult to find one’s way. There are two possible interpretations of the addition of a genetic sequence:

• One interpretation is that the addition of a gene sequence is a permanent addition, which is the historical vision of gene therapy: the addition of DNA.
• A more modern interpretation, corresponding to that of the EMA is that the simple fact of injecting a recombinant nucleic acid sequence is sufficient. In other words, the administration (injection) is synonymous with addition.⁶⁸ Here, there is no notion of permanence.

The CAT’s vision, in which “addition” and “administration” are synonymous, anchors mRNA in the field of GTMPs.⁶⁹ Indeed, the lifespan of the mRNA, once administered, is short: mRNA has no long-lasting action.

Furthermore, the European definition of GTMP adds a very specific mode of action criterion: “Its therapeutic, prophylactic or diagnostic effect relates directly to the recombinant nucleic acid sequence it contains, or to the product of genetic expression of this sequence.”

This means that the recombinant nucleic acid sequence must have a therapeutic (curative), prophylactic (preventive) or diagnostic effect. In other words, a genetic manipulation that does not have a therapeutic, prophylactic or diagnostic effect will not be classified as a GTMP. This is not a hypothetical situation, but a very real one. Indeed, some genetic manipulations are aimed, for example, at adding a suicide gene (thymidine kinase or inducible caspase-9 techniques, etc.).⁷⁰ One example is Zalmoxis®, which has now been withdrawn from the market.⁷¹ In this case, allogeneic T-lymphocytes capable of restoring the immune system of a bone marrow transplant recipient were genetically modified (by adding a kind of self-destruct system to avoid the graft-versus-host reaction they can sometimes induce). The therapeutic effect is provided by the T-lymphocyte; the added genetic sequence does not carry the activity and so Zalmoxis was not a GTMP. Another example is the genetic modification of an oncolytic virus or protozoan so that it does not replicate in healthy cells but only in cancer cells. As in the previous example, this manipulation also has no therapeutic activity and therefore excludes the product from the scope of GTMPs.⁷²

In the case of the genetically modified T lymphocyte and protozoan, the medicinal product will not be a GTMP but rather another ATMP: a somatic cell therapy medicinal product (sCTMP), which ensures a high level of safety for patients but does not have the same development requirements as a GTMP. However, the oncolytic virus, which has been genetically modified to prevent it from replicating in healthy cells, does not qualify for sCTMP status because it does not consist of cells. It is instead classified as a biological medicinal product with fewer requirements than an ATMP.

This non-inclusion of genetically modified cells and viruses, because the action is not carried by the modified sequence, poses another logical problem. The International Non-proprietary Names (INN) for gene therapy medicinal products, defined by the WHO since 2005, are in principle made up of two words, sometimes combined. There are two types of gene therapy: gene therapy based on genetically modified cells, and gene therapy based on a viral vector, plasmid, or bacterium. In the INN, the first word refers to the gene therapy, and the second to the vector or cell, depending on the case. The first word ends with the suffix -gene or -gen, making it immediately clear that it is a gene therapy, as in talimogene laherparepvec (an in vivo gene therapy for unresectable melanoma) or tisagenlecleucel (CAR-T cells for certain hematological malignancies).

The problem is that the WHO does not take into account the notion of direct action of the gene sequence when defining INNs for gene therapy medicinal products. For example, Rivogenlecleucel, is an experimental medicine (currently in clinical trials) based on allogeneic cells modified to carry a suicide gene (in this case, caspase 9). Under European regulations, this product would be classified as a sCTMP and not as a GTMP, even though its INN clearly states that it is a gene therapy. This creates an inconsistency and an additional difficulty for healthcare professionals and for manufacturers to qualify the product they are developing and to know the applicable rules.

This really raises the question of whether or not to maintain this subtle European distinction between a genetically modified cell, which is a GTMP if the modification has a therapeutic, prophylactic or diagnostic action, and a sCTMP if the modification does not (safety requirements for sCTMPs being slightly less stringent).

In addition to the function or mode of action, the definition of GTMP includes a negative destination criterion.

B) A Negative Destination Criterion: The Exclusion of Vaccines Against Infectious Diseases

The GTMP definition includes a negative destination criterion: “Gene therapy medicinal products shall not include vaccines against infectious diseases.”

In other words, a product that meets both the definition of a GTMP and that of a vaccine will be classified as a vaccine. Therefore, it is important to understand the legal definition of a vaccine, what might not be considered a vaccine, and why the legislator has chosen to classify certain products that could potentially be GTMPs as vaccines. As we will see, this categorization between vaccine and GTMP has significant impacts, particularly in terms of safety.

1) The Definition of Vaccine

According to directive 2001/83/EC,⁷³ a vaccine is an immunological medicinal product intended to provide active immunity (in contrast to a serum, which provides passive immunity). In other words, a vaccine is able to “train” the patient’s immune system to develop its own defense against an infectious agent.

A number of vaccines are mentioned in this definition: cholera vaccine, BCG,⁷⁴ polio vaccine and smallpox vaccine. It is clear from these examples that vaccines in the regulatory sense are only medicinal products that confer active immunity against the pathogen of an infectious disease. Thus, from a legal point of view, a vaccine is limited to the prevention of a disease caused by an infectious agent. This view is undoubtedly explained by the age of the European legal definition of a vaccine, which dates back to 1975 and was introduced by Directive 75/319/EEC.⁷⁵ This view is also confirmed by the 11th edition of the European Pharmacopeia: “Vaccines for human use are preparations that contain antigenic substances capable of inducing a specific and active immunity against the infecting agent or the toxin or the antigen produced by it.”⁷⁶ In March 2023 an agreement was made for the writing of new monographs on mRNA vaccines, including a specific monograph: “mRNA Vaccines for human use (5.36).”⁷⁷ Indeed, the monograph for “traditional” vaccines did not meet the control requirements necessary for mRNA vaccines. This new monograph should facilitate the understanding of the requirements for manufacturers wishing to develop vaccines based on this technology.

2) A Restrictive Definition of Vaccines

This European legal vision of vaccines, strictly limited to the fight against infectious diseases, clashes with a semantic usage that seeks to include in the field of vaccines medicinal products that prevent and/or cure a non-infectious disease, such as cancer. While certain vaccines, such as those against human papillomavirus or hepatitis B, can effectively prevent indirectly cancer, the term “cancer vaccine” is often used to designate quite different medicinal products: it is possible to induce active immunity against cancer cells, for example, by using dendritic cells or by using RNA or DNA. Some scientists use the term “cancer vaccine” without hesitation,⁷⁸ although under European law this is generally an ATMP. Moreover, categorization as a GTMP, in this specific case, seems to offer greater protection than categorization as a vaccine. For example, biodistribution studies are mandatory for GTMPs, and oncogenesis testing is often more thorough than for vaccines.

Dendritic cells are used to fight cancer, and one medicinal product has even been granted MA (withdrawn in the EU for commercial reasons, but still active in the USA). The drug in question is sipuleucel (Provenge®), often presented as “a vaccine against prostate cancer.”⁷⁹ In the EU, this medicinal product was a sCTMP (and not a GTMP) and could not claim vaccine status because prostate cancer is not an infectious disease.

The use of mRNA to create active immunity against cancer is also a promising technology and a number of clinical trials are underway.⁸⁰ Two pharmaceutical companies, Moderna and Merck have developed an anti-melanoma product in phase 3 trials at the Centre Georges-François Leclerc in Dijon.⁸¹ This experimental treatment is subject to a PRIME (PRIority Medicines) procedure within the EU, with a view to accelerating evaluation (150 instead of 210 days). Two other pharmaceutical companies, BioNTech and Roche began a clinical trial in 2023 to treat pancreatic cancer.⁸² In both cases of mRNA therapy, the aim is to produce individualized neoantigens to activate antigen-presenting cells. In the EU, these treatments would be classified as GTMPs and not as vaccines.

It would be preferable for scientists to avoid referring to these RNA medicinal products as “vaccines,” as this creates confusion within the scientific community, the legal sphere, and among the general public. Some have suggested calling them “immunotherapies” (since they induce immune responses), but this term is very broad and is currently used in medical parlance to refer more specifically to checkpoint inhibitors.⁸³ The term individualized neoantigenic (mRNA) therapy therefore seems more appropriate to avoid any confusion with vaccines regarding the legal qualification and applicable regime.

It is therefore understandable that certain products, such as cancer “vaccines,” are legally outside the scope of vaccines, at least in Europe, but are generally ATMPs. However, some products may meet both the definition of a GTMP and that of a vaccine (against an infectious disease). Typically, these are mRNA anti-infective vaccines. In this specific case, the negative destination criterion applies. So why are they classified as vaccines and not as GTMPs? The answer is not simple.

3) The Distinction Between GTMPs and Vaccines: an Issue of Coherence and Safety, Justified by the Specific Characteristics of Vaccines

The decision to classify a medicinal product in one category rather than another, apart from the logic inherent in the nature of the product itself, generally stems from a concern for safety: the guidelines applicable to development and manufacture will not be the same.

It was likely the intended use of the product that mattered: The vaccine is usually administered to a healthy person for prophylactic purposes, to prevent infection. Additionally, it is sometimes given on a large scale as part of a vaccination campaign. In this context, vaccines are distinct medicinal products, and it was undoubtedly for the sake of consistency that the legislator in 2007 chose to classify prophylactic medicinal products against infectious diseases as vaccines. There is a certain logic in prioritizing the destination of the product over its mode of action (mRNA or DNA). However, this does not preclude the application of some or all GTMP guidelines during product development. For example, and in response to certain criticisms, the biodistribution studies required for GTMPs should also be applied to mRNA vaccines. This would not only enhance knowledge and improve the safety profile of the product.

Some vaccines also have another feature of particular interest in terms of safety: an EU Member State may ask an Official Medicines Control Laboratory (OMCL), usually under the supervision of the competent national authority, to carry out an official batch release (OCABR). Article 114 of the consolidated Directive 2001/83/EC provides that a Member State may, for public health reasons, require the marketing authorization holder of a new vaccine, in particular if it is to be used as part of a public health immunization program, to submit samples of each batch to the OMCL before it is actually placed on the market, ensuring a very high level of safety for patients.⁸⁴ Thus, the regulations concerning vaccines and those pertaining to GTMPs are both highly protective of patients, albeit for different reasons.

Moreover, while the European decision to include both DNA and RNA in the field of gene therapy strengthens patient safety. The complex consideration of GTMPs’ mode of action raises questions. It appears misaligned with the INN, posing a challenge for international harmonization. Conversely, the decision to exclude anti-infective vaccines from the scope of GTMPs may be justified, but it is the use of the term “cancer vaccine” by scientists and healthcare professionals that tends to cause confusion among scientists and legal experts alike.

A comprehensive reflection and the establishment of an international consensus are essential to address this specific issue. The current definition of Gene Therapy Medicinal Products (GTMPs) presents significant challenges for both safety and innovation. On one hand, it no longer adequately corresponds to the rapid pace of technological advancements in the field. This misalignment creates regulatory gaps, hindering the development of new therapeutic solutions. On the other hand, some technologies are exempt from GTMP regulation, potentially compromising patient safety.

In this regard, it seems important to explore the possibilities for evolving the definition of GTMPs to encompass technological advancements and ensure patient safety.

III. What Solutions Can Be Found to Adapt the Definition of GTMP to Technological Evolutions?

A partial solution to the challenges of defining GTMPs could come from the International Council for Harmonisation (ICH)⁸⁵ (A). However, this remains only a partial solution, and we propose several recommendations to modify the GTMP definition (B).

A) The ICH Solution

In the S12 guideline,⁸⁶ the ICH has outlined a form of definition for gene therapy medicinal products, or at least a definition of what is and what is not within the scope of gene therapy. This text solves certain consistency and scientific problems and would have the advantage of harmonizing different definitions and visions, but it does not solve all the problems.

ICH guideline S12 on nonclinical biodistribution considerations for gene therapy products, yet adopted by the EU, defines the scope of “gene therapy products.” It states that gene therapy products are:

products that mediate their effect by the expression (transcription or translation) of transferred genetic materials. Some examples of GT products can include:

• purified nucleic acid (e.g., plasmids and RNA), microorganisms (e.g., viruses, bacteria, fungi) genetically modified to express transgenes (including products that edit the host genome),

• and ex vivo genetically modified human cells.

Products that are intended to alter the host cell genome in vivo without specific transcription or translation (i.e., delivery of a nuclease and guide RNA by non-viral methods) are also covered in this guidance. Although not currently considered GT in certain regions, the principles outlined in this guideline are also applicable to oncolytic viruses that are not genetically modified to express a transgene.

This guideline does not apply to prophylactic vaccines.

Chemically synthesized oligonucleotides or their analogues, which are not produced using a biotechnology-based manufacturing process, are also outside the scope of this guideline.

This definition therefore states that the main mechanism of a “gene therapy product” is linked to the expression of the transferred genetic material. It is interesting to note that this wording bypasses the nature of the product: it does not specify that it is a biological product or a nucleic acid. However, it should be noted that these products must act by means of transcription or translation, which makes it possible to understand that this refers to DNA or RNA (mentioned in the examples). (See Table 2).

The precision of “transferred genetic material” also corresponds to the idea that human intervention is required, but does not mention the principle of recombination. In addition, the text includes in the field of gene therapy direct genome editing tools per se (as medicinal products), which, although they do not act by means of transcription or translation, are undeniably products that deserve to be included in the field of gene therapy medicinal products in order to ensure better development and monitoring because of the risks that this technology may potentially induce.

Conversely, the inclusion of non-genetically modified oncolytic viruses in the definition of GTMPs does not seem appropriate, but it should be remembered that this is a scope guideline rather than a true legal definition. Additionally, the text excludes prophylactic vaccines, i.e., anti-infective vaccines.

Finally, it excludes synthetic oligonucleotides, but not those produced by biotechnology.

What is meant by “oligonucleotide”? Oligonucleotides are not defined in the guideline glossary. Oligonucleotides are short nucleic acids.⁸⁷ They seem to cover both antisense oligonucleotides (ASOs) and small RNAs such as SiRNAs and SaRNAs,⁸⁸ which are sometimes even grouped together under the term of “oligonucleotide-based therapeutics”⁸⁹ or more simply “oligonucleotide drugs.”⁹⁰

Once again, the question of biological nature arises, which was conveniently avoided at the beginning of the ICH “definition,” but which comes back like a boomerang for these products. By excluding from the scope of the guideline oligonucleotides that “are not produced using a biotechnology-based manufacturing process,” the wording implies that oligonucleotides produced by biotechnology would be included, provided that they act by “expressing (transcribing or translating) genetic material.” The situation seems complex, and these points need to be clarified.

It should be noted that the ICH outlined this kind of definition with the objective of delineating the scope of the S12 guideline, not with the intent of altering any legal definition within the member states of the ICH. Consequently, this may explain the presence of some minor inconsistencies in the text. Since this kind of definition remains, from our point of view, perfectible, we have proposed several recommendations to modify the GTMP definition.

	EU	ICH
Source	Biological	Not specified
Nature	Recombinant nucleic acid	Transferred genetic material
Function/purpose	regulate, repair, replace, add or delete a genetic sequence	Not specified
Means/mode of action	its effect is therapeutic, prophylactic or diagnostic; this effect depends directly on the recombinant nucleic acid sequence used.	The effect is obtained by the expression (transcription or translation) of transferred genetic materials. By exception, direct genome editing tools are included.

B) Proposals for Modifying the Definition of GTMPs

The initial measure that we proposed is to reconsider the biological nature of GTMP in order to anticipate the imminent arrival of chemically synthesized DNA or RNA. Without a revision of the definition on this point, chemically synthesized nucleic acids could in principle de facto escape the controls inherent in gene therapy medicinal products. The EMA is fully aware of this, and the CAT, in a meeting minutes dating from 2020, explained that “such long chain mRNAs cannot yet be produced via chemical synthesis. However, when this becomes possible, the regulatory status of such synthetic RNAs needs to be considered, as it should be avoided having similar products being covered by different legal frameworks.”⁹¹ In order to address this issue, both the EMA and the European Commission have proposed that “The Commission supported EMA’s interpretation that RNA derived products [it should read mRNA here] should be considered as biologicals, even if not derived from a biological source.”⁹² While this position is to be welcomed, it raises the question of the legal force of this interpretation. If the EMA has the possibility of producing some kind of soft law through guidelines, can it issue a doctrinal position, a sort of praetorian construction, which runs counter to a European directive or regulation?⁹³

This first measure could have far-reaching consequences, particularly the inclusion of other products currently marketed as “simple” medicinal products within the scope of GTMPs.

The removal of the biological nature from the European definition of a GTMP could also have other impacts, particularly in terms of intellectual property. A biological GTMP could be copied by a chemically synthesized GTMP. When the intellectual property and data protection of the MA expire, it is possible to copy a medicinal product. However, depending on the nature of the product, the copy will be different. In the case of a chemical or chemically synthesized medicinal products, it is possible to create a generic medicinal product. Conversely, if the medicinal product to be copied is biological in nature, a similar biological medicinal product, more commonly known as a biosimilar, is created.⁹⁴ This is because it is not possible to perfectly copy a biological product perfectly due to differences linked in particular to the variability of the raw material or to the medicinal product’s manufacturing processes, which necessitates the production of additional clinical data. Therefore, in principle, a biological medicinal product cannot have a generic (as understood for non-biological medicinal products). However, this was without taking into account the scientific and regulatory ingenuity of certain generic companies. Indeed, when the biological molecule is relatively simple, it is now possible to produce it by chemical synthesis. This was demonstrated with teriparatide, a protein fragment of 34 amino acids corresponding to part of parathormone.⁹⁵ The reference medicinal product is a biological drug (the amino acid sequence is produced on E. coli using the recombinant DNA technique). Some laboratories have copied the medicinal product by means of biosimilars, sometimes by changing the production technique (i.e., on Pseudomonas fluorescens). However, other laboratories have chemically synthesized the amino acid sequence. Thus, the drug produced is not biological but rather chemically synthesized, making it a generic drug… whose substitution is automatic by the pharmacist, unlike biosimilars, much to the dismay of the laboratories producing biosimilar teriparatide.⁹⁶ If we transpose this example to the GTMP, as it is possible to produce nucleic acids by chemical synthesis, it would be possible for a reference GTMP to be biological and to be copied by means of a generic GTMP.

The second measure could be to replace the expression “recombinant nucleic acid” with “polynucleotide substance” in the legal definition of GTMP. The term “nucleic acid” would be deliberately replaced by the expression “polynucleotide substance,” which would include all nucleotide groupings, regardless of their mass or length, and thus avoid ill-defined terms such as “nucleic acids” or “oligonucleotides.” This proposal does not deliberately exclude non-coding oligonucleotides from the scope of the definition. Indeed, we believe that these products, which regulate gene expression, deserve to be better regulated through their inclusion as GTMP. However, if we wish to retain the notion of recombination in the definition, we would need to provide a little clarity. This could be achieved by clearly defining, for example, in a guideline or reflection paper, what is meant by “recombinant.”

The third essential measure would be to amend the definition to include genome-editing tools per se which, scientifically, are gene therapies and require strict supervision.

The fourth measure could be to simplify the mode of action of the GTMP. This would entail that any genetic modification of a cell would result in the medicinal product being categorized as a GTMP. This would avoid having genetically modified cells categorized as sCTMPs. It would also provide greater consistency with regard to INNs. In practical terms, this would mean deleting the passage that states that the effect depends directly on the recombinant nucleic acid sequence used. It seems unnecessary to retain the phrase “therapeutic, prophylactic or diagnostic” in the definition of a medicinal product, as this is inherent in the definition of a medicinal product.⁹⁷

Finally, it is understandable that certain products which may also fall within the scope of GTMPs should be classified as vaccines. However, it would be appropriate to consider the definition of a vaccine, which dates back to 1975. Should the definition be confined to the prophylaxis of infectious diseases, or should it be expanded to include other non-infectious diseases such as cancer? This question cannot be resolved in this article, which focuses on the GTMP. For the time being, we propose to let vaccines for infectious diseases outside the scope of GTMPs.

We suggest the following de lege feranda formulation:

A gene therapy medicinal product is any medicine consisting of a [recombinant] polynucleotide substance, whose action is to regulate, repair, replace, add, or delete a genetic sequence. Furthermore, medicinal products intended to modify the genome of the host cell in vivo without specific transcription or translation should also be considered gene therapy medicinal products.

Vaccines against infectious diseases are not included in gene therapy medicinal products.

In the meantime, until the European legislator changes the definition of GTMP, it should be necessary to apply the GTMP guidelines to the other products that we believe should fall within the scope of GTMP. This would avoid the temptation for some laboratories to use certain technologies (especially chemical synthesis) rather than others in order to avoid having to comply with the very strict GTMP requirements. Patient safety must therefore be our only guide.

Two additional solutions should be considered, either in conjunction with or as alternatives to modifying the definition. First, it may be worthwhile to grant the European Medicines Agency (EMA) a more substantial role in soft law, allowing it to clarify certain aspects of the definition through guidelines. Second, the reform of European pharmaceutical law⁹⁸ includes the creation of regulatory sandboxes,⁹⁹ which could serve as practical solutions for addressing isolated technological advancements that fall outside the current scope of the GTMP definition, thereby avoiding legal coercion.¹⁰⁰ These sandboxes would prevent the need for frequent revisions of the definition, avoiding a “patchwork” approach, especially if the technology in question is either isolated or likely to multiply and become more permanent. Of course, both of these solutions will require further study.

---

1 P. Lachaume, et al., “Rapide histoire de la génétique : de Mendel à Jacob et Monod,” in Génétique formelle : méthodes et exercices corrigés. 2021, Ellipses, Paris, p. 9-15. M.-D. Grmek, “Le centenaire des lois de Mendel,” Annales, 1967. p. 845-848.

2 Ionizing radiation refers to energy forms capable of damaging DNA. This process can cause mutations in the DNA of living organisms. This effect is central to understanding how radiation exposure may lead to genetic disorders (but also cancers), a concept further developed in the theory of induced mutagenesis.

3 Eugenics refers to the practice or advocacy of improving the genetic quality of a human population, historically through selective breeding or sterilization. Over times, it has been associated with controversial and unethical policies aimed at controlling reproduction to eliminate perceived genetic “defects.”

4 A. Tétry, “Uller Hermann Joseph” (1890-1967), in Encyclopædia Universalis [en ligne], consulté le 29 mars 2024.

5 Aldous Huxley (1894–1963) was a British writer and philosopher best known for his dystopian novel Brave New World (1932), which explores the implications of technological and social engineering. Huxley’s works often address themes of social control, individuality, and the potential consequences of scientific advancement.

6 A. Huxley, Brave New World, 1932, Chatto and Windus.

7 Julian Huxley (1887–1975) was a British biologist, philosopher, and the first Director-General of UNESCO, and a founding member of the WWF. He was a leading figure in the modern synthesis of evolutionary biology. Huxley’s work significantly influenced the fields of biology and international science policy.

8 J. S. Huxley, Essays of a humanist, London, 1964.

9 The “one gene/one enzyme” principle, proposed by George Beadle and Edward Tatum in the 1940s, suggests that each gene in an organism controls the production of a single enzyme, which is essential for a specific biochemical process. Their experiments showed that mutations in individual genes could disrupt enzyme function, linking genes directly to biological activity. While later research expanded this idea to include all types of proteins, the principle was key in connecting genetics to the functioning of living organisms.

10 G.W. Beadle, and E. L. Tatum, “Genetic Control of Biochemical Reactions in Neurospora,” Proc Natl Acad Sci USA, 1941, 27(11), pp. 499–506.

11 J. D. Watson and F. H. Crick, “Molecular structure of nucleic acids; a structure for deoxyribose nucleic acid” Nature, 1953, 171 (4356), pp. 737-8.

12 F. Jacob and J. Monod, “Genetic regulatory mechanisms in the synthesis of proteins,” J Mol Biol, 1961, 3, pp. 318-56.

13 Brenner, S., F. Jacob, and M. Meselson, “An unstable intermediate carrying information from genes to ribosomes for protein synthesis,” Nature, 1961, 190, pp. 576-581.

14 Indeed, his thought was that the goal of molecular biology will be, one day, to manipulate human DNA by directly controlling nucleotide sequences (the building blocks of DNA) to insert specific, desired genes. It predicts that eventually, scientists will be able to chemically attach custom DNA sequences to a virus, which could then be used to deliver these genes into cells. A. P. Cotrim and B. J. Baum, “Gene therapy: some history, applications, problems, and prospects,” Toxicol Pathol, 2008, 36(1), pp. 97–103.

15 S. Rogers, “Skills for genetic engineers,” New Scientist, 1970, 45 (686): p. 29.

16 T. Friedmann and R. Roblin, “Gene therapy for human genetic disease?,” Science, 1972, 175 (4025), pp. 949-55.

17 A DNA sequence that does not belong to the patient.

18 Arginase deficiency is a rare genetic condition that affects children and occurs when a mutation in the ARG1 gene leads to insufficient production of the enzyme arginase. As a result, the body is unable to properly process and eliminate ammonia, a toxic substance, leading to symptoms such as leg stiffness, slowed growth, cognitive delays, and delayed developmental milestones. Without treatment, it can lead to serious brain damage.

19 H. G. Terheggen, et al., “Unsuccessful trial of gene replacement in arginase deficiency,” Z. Kinderheilkd, 1975, 119 (1), pp. 1–3.

20 Martin Cline’s experiment was widely criticized for its serious ethical breaches. After his initial attempt to use the gene therapy technique in mice failed, Cline made a significant and risky leap by proceeding to test the method in humans, without adequate evidence of its safety or efficacy. Furthermore, he did so without proper regulatory approval, raising concerns about patient consent and safety. E. Beutler, “The Cline Affair,” Molecular Therapy, 2001, 4(5), pp. 396-397. J. A. Wolff and J. Lederberg, “An early history of gene transfer and therapy,” Hum Gene Ther, 1994, 5(4), pp. 469-80.

21 S.A. Rosenberg, et al., “Gene Transfer into Humans — Immunotherapy of Patients with Advanced Melanoma, Using Tumor-Infiltrating Lymphocytes Modified by Retroviral Gene Transduction,” New England Journal of Medicine, 1990, 323 (9), pp. 570–578.

22 K. W. Culver, W. F. Anderson and R. M. Blaese, “Lymphocyte gene therapy,” Hum Gene Ther, 1991, 2(2), pp. 107-9.

23 L. M. Muul, et al., “Persistence and expression of the adenosine deaminase gene for 12 years and immune reaction to gene transfer components: long-term results of the first clinical gene therapy trial,” Blood, 2003, 101 (7), pp. 2563-9.

24 S. L. Ginn, et al., “Gene therapy clinical trials worldwide to 2017: An update,” The Journal of Gene Medicine, 2018, 20(5), p. e3015.

25 The X-SCID (X-linked Severe Combined Immunodeficiency) is a genetic disorder caused by a defect in the gamma chain of the IL-2 receptor, which plays a critical role in the immune system. This defect prevents the body from producing functional immune cells, leaving individuals highly vulnerable to infections. Without treatment, even common infections can become life-threatening.

26 A. Fischer, S. Hacein-Bey-Abina, and M. Cavazzana-Calvo, “20 years of gene therapy for SCID,” Nat Immunol, 2010, 11(6), pp. 457-60. M. Cavazzana-Calvo, et al., “Gene therapy of human severe combined immunodeficiency (SCID)-X1 disease,” Science, 2000, 288 (5466), pp. 669-72. S. Hacein-Bey-Abina, et al., “LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-X1,” Science, 2003, 302 (5644), pp. 415-9.

27 M. L. Edelstein, M. R. Abedi, and J. Wixon, “Gene therapy clinical trials worldwide to 2007 – an update,” J Gene Med, 2007, 9(10), pp. 833-42.

28 E. Blanco, et al., “Immune Reconstitution After Gene Therapy Approaches in Patients With X-Linked Severe Combined Immunodeficiency Disease,” Front Immunol, 2020, 11.

29 Commission Directive 2003/63/EC of 25 June 2003 amending Directive 2001/83/EC of the European Parliament and of the Council on the Community code relating to medicinal products for human use, OJ L 159, 27.6.2003, pp. 46–94.

30 Directive 2001/83/EC of the European Parliament and of the Council of 6 November 2001 on the Community code relating to medicinal products for human use, OJ L 311, 28.11.2001, pp. 67–128.

31 Gendicine® was designed to treat head and neck squamous cell carcinoma. It works by delivering the p53 tumor suppressor gene directly into tumor cells using an adenovirus vector. This gene produces a protein that triggers the self-destruction of cancer cells or enhances the immune system’s ability to fight them. The goal is to target cancer cells specifically while minimizing damage to healthy tissue.

32 F. Arabi, V. Mansouri and N. Ahmadbeigi, “Gene therapy clinical trials, where do we go? An overview,” Biomedicine & Pharmacotherapy, 2022, 153 ; Qi L., et al., “Twenty years of Gendicine® rAd-p53 cancer gene therapy: The first-in-class human cancer gene therapy in the era of personalized oncology,” Genes & Diseases, 2024, 11(4), pp. 101155.

33 Regulation (EC) No 1394/2007 of the European Parliament and of the Council of 13 November 2007 on advanced therapy medicinal products and amending directive 2001/83/EC and Regulation (EC) No 726/2004, OJ L 324 10.12.2007, p. 121.

34 Commission Directive 2009/120/EC of 14 September 2009 amending Directive 2001/83/EC of the European Parliament and of the Council on the Community code relating to medicinal products for human use as regards advanced therapy medicinal products, OJ L 242, 15.9.2009, pp. 3–12.

35 Commission Directive 2009/120/EC of 14 September 2009 amending Directive 2001/83/EC of the European Parliament and of the Council on the Community code relating to medicinal products for human use as regards advanced therapy medicinal products, OJ L 242, 15.9.2009, pp. 3–12.

36 M. Guerriaud, “Les Médicaments de Thérapie Innovante – Statut juridique (Fascicule 61-70),” JurisClasseur Droit pharmaceutique, LexisNexis, 2022.

37 The drug consisted of a viral vector containing the human lipoprotein lipase gene variant, aimed at treating familial lipoprotein lipase deficiency. This is a rare genetic disorder that impairs the body’s ability to break down fat particles in the blood, leading to symptoms such as severe abdominal pain, recurring inflammation of the pancreas (pancreatitis), and the risk of life-threatening complications.

38 T. Bolden, “Why is the number of advanced therapy medicinal product (ATMP) withdrawals from the European market rising?,” Remap Consulting, 2022 [cited September 2024].

39 The Conseil d’État is one of the two highest courts in France, serving as the supreme court for administrative justice and advising the government on legal matters.

40 Conseil d’État, 1^re/6^e SSR, 17/02/2012, n^o 332509, 2012.

41 European Medicines Agency, “Assessment Report for Evrysdi® risdiplam (EMA/216061/2021),” Committee for Medicinal Products for Human Use (CHMP), 2021.

42 DNA (deoxyribonucleic acid) is the molecule that carries the genetic blueprint of living organisms, encoding the instructions for building and maintaining cells. RNA (ribonucleic acid), on the other hand, is primarily responsible for translating these instructions into proteins. The key difference between them lies in their structure and function: DNA is double-stranded and stores genetic information long-term, while RNA is single-stranded and acts as a messenger, transferring genetic data from DNA to the protein-making machinery of the cell.

43 B.T. Kren, et al., “Correction of the UDP-glucuronosyltransferase gene defect in the gunn rat model of crigler-najjar syndrome type I with a chimeric oligonucleotide,” Proceedings of the National Academy of Sciences of the United States of America, 1999, 96(18), pp. 10349–10354.

44 A polynucleotide is a long chain of nucleotides, which are the basic building blocks of DNA and RNA. Each nucleotide consists of a sugar, a phosphate group, and a nitrogenous base. When many nucleotides are linked together, they form a polynucleotide. R.A. Morgan and R.M. Blaese, “Gene therapy: lessons learnt from the past decade. Interview by Clare Thompson,” BMJ (Clinical research ed.), 1999, 319 (7220), pp. 1310-1310.

45 For example, in vitro transcribed RNA comprising mRNA encoding human interleukin 2 (IL‑2) and an antagonist RNA (aptamer) targeting VEGF-A. European Medicines Agency, “Scientific recommendation on classification of advanced therapy medicinal products.” EMA/140033/2021, 2023.

46 In 2024, 100 kilobases cost approximately €6.5 to €8.5 / High Quality Gene Synthesis – Twist Bioscience [cited September 2024]. R. A. Hughes and A. D. Ellington, “Synthetic DNA Synthesis and Assembly: Putting the Synthetic in Synthetic Biology,” Cold Spring Harb Perspect Biol, 2017, 9(1).

47 siRNAs bind to specific complementary mRNAs, which are then degraded and can no longer be translated.

48 Unlike siRNAs, which repress gene expression, saRNAs activate gene expression by binding to specific sequences in the promoter or other biological regulatory regions of the DNA, leading to increased transcription of the targeted genes.

49 M. Guerriaud and E. Kohli, “RNA-based drugs and regulation: Toward a necessary evolution of the definitions issued from the European union legislation,” Frontiers in Medicine, 2022, 9.

50 G.M. Traber and A.M. Yu, “RNAi-Based Therapeutics and Novel RNA Bioengineering Technologies,” J Pharmacol Exp Ther, 2023, 384 (1), pp. 133–154.

51 M. Guerriaud and E. Kohli, “RNA-based drugs and regulation,” op. cit.

52 European Medicines Agency, Guideline on the quality, non-clinical and clinical aspects of gene therapy medicinal products (EMA/CAT/80183/2014), CAT, 2018.

53 The centralized marketing authorization procedure, managed by the EMA, is mandatory for certain categories of medicinal products, including orphan drugs, advanced therapy medicinal products (ATMPs), and products derived from biotechnology. It is also required for innovative treatments and for products intended for the treatment of specific diseases, such as cancer, HIV/AIDS, diabetes, neurodegenerative conditions (e.g., Alzheimer’s and Parkinson’s diseases), and autoimmune disorders.

54 Académie nationale de Pharmacie, “Nucléique (acide)” — acadpharm.

55 Académie nationale de Médecine, “Acide Nucléique,” 2020.

56 S. Müller, Nucleic Acids from A to Z: A Concise Encyclopedia, 2008, Wiley.

57 J. Kruh, et al., “Nucléiques (acides),” in Encyclopædia Universalis [online], 2024.

58 Encyclopædia Britannica, Nucleic acid, 2024.

59 Académie Nationale de Médecine, “Acide Nucléique,” op. cit.

60 S. C. Jang, et al., “ExoSTING, an extracellular vesicle loaded with STING agonists, promotes tumor immune surveillance,” Communications Biology, 2021, 4(1), p. 497.

61 M. Egli, and M. Manoharan, “Chemistry, structure and function of approved oligonucleotide therapeutics,” Nucleic Acids Research, 2023, 51(6), pp. 2529–2573. T. C. Roberts, R. Langer, and M. J. A. Wood, “Advances in oligonucleotide drug delivery,” Nature Reviews Drug Discovery, 2020, 19(10), p. 673-694.

62 T. Pedrazzini, “Le cœur des ARN non codants – Un long chemin à découvrir,” Médecine/sciences, 2015, 31(3), pp. 261-267.

63 J.S. Mattick, et al., “Long non-coding RNAs: definitions, functions, challenges and recommendations,” Nature Reviews Molecular Cell Biology, 2023, 24(6), pp. 430–447.

64 European Medicines Agency, “Scientific recommendation on classification of advanced therapy medicinal products,” EMA/348841/2012, 2012.

65 For more information on GMOs, see Estelle Brosset’s article in the present dossier.

66 These genome editing tools can be used to producing GTMP, but are not currently GTMP per se.

67 Bacteriophage particles carrying recombinant DNA, which encodes CRISPR/Cas circuits that enable specific killing of E.coli bacteria by introducing programmed double-strand breaks in the bacterial DNA. Intended to treat bloodstream E. coli infection in neutropenic patients with hematological malignancy. 19/02/2021 / E. coli containing a plasmid encoding a CRISPR-Cas system that targets three polyketide synthase (PKS) island genes. Intended to prevent the progression of the familial Adenomatous Polyposis disease / European Medicines Agency, “Scientific recommendation on classification of advanced therapy medicinal products,” EMA/140033/2021, 2024.

68 For example, on 24/05/2024, the CAT considered that mRNA encoding ARCUS nuclease is a product administered to human beings with the purpose of adding a genetic sequence. See European Medicines Agency, “Scientific recommendation on classification of advanced therapy medicinal products,” EMA/140033/2021, 2024.

69 ibidem.

70 P. Tiberghien, et al., “Administration of herpes simplex–thymidine kinase–expressing donor T cells with a T-cell–depleted allogeneic marrow graft,” Blood, 2001, 97(1), pp. 63–72. C. Bonini, et al., “HSV‑TK Gene Transfer into Donor Lymphocytes for Control of Allogeneic Graft-Versus-Leukemia,” Science, 1997, 276 (5319), pp. 1719–1724.

71 As with almost all ATMP withdrawals, Zalmoxis was withdrawn from the market due to a lack of commercial profitability and challenges related to its adoption in European healthcare systems, rather than for safety or clinical efficacy reasons.

72 M. Guerriaud, et al., “Are genetically modified protozoa eligible for ATMP status? Concerning the legal categorization of an oncolytic protozoan drug candidate,” Nature Gene Ther, 2024.

73 Directive 2001/83/EC of the European Parliament and of the Council of 6 November 2001 on the Community code relating to medicinal products for human use, OJ L 311, 28.11.2001, pp. 67–128.

74 Bacillus Calmette-Guérin, i.e., the vaccine against tuberculosis.

75 Second Council Directive 75/319/EEC of 20 May 1975 on the approximation of provisions laid down by Law, Regulation or Administrative Action relating to proprietary medicinal products.

76 European Directorate for the Quality of Medicines and HealthCare, “Vaccine for human use 04/2022:0153,” in European Pharmacopeia, 11th Edition Council of Europe, Editor, 2022.

77 EDQM, “Ph. Eur. Commission kicks off elaboration of three general texts on mRNA vaccines and components,” European Directorate for the Quality of Medicines & HealthCare, 2023, Strasbourg.

78 M. Saxena et al., “Therapeutic cancer vaccines,” Nature Reviews Cancer, 2021, 21(6), pp. 360–378. M. J. Lin, et al., “Cancer vaccines: the next immunotherapy frontier,” Nature Cancer, 2022, 3(8), pp. 911–926. M. A. Morse, W. R. Gwin, and D. A. Mitchell, “Vaccine Therapies for Cancer: Then and Now,” Targeted Oncology, 2021, 16(2), pp. 121–152.

79 The treatment consists of collecting leukocytes (white blood cells) from the patient by leukapheresis (the cells in question are peripheral blood mononuclear cells, including dendritic cells, which are antigen-presenting cells). These cells are then brought into contact, in the laboratory, with a recombinant fusion protein: prostatic acid phosphatase (PAP, representing the tumor antigen) fused to granulocyte-monocyte colony-stimulating factor (GM-CSF). The dendritic cells are then activated and able to present the tumor antigens to T lymphocytes, once reinjected into the patient; EMA, “Assessment Report on PROVENGE. Common name: Autologous peripheral blood mononuclear cells activated with PAP-GM-CSF (sipuleucel-T),” EMA/440011/2013, CHMP, 2013. I. Ouzaid, and V. Ravery, “[Sipuleucel-T: a prostate cancer vaccine: ‘instructions for use’ for urologists],” Prog Urol, 2011, 21(9), pp. 595-8.

80 T. Fan et al., Therapeutic cancer vaccines: advancements, challenges, and prospects. Signal Transduction and Targeted Therapy, 2023, 8(1), pp. 1–23. E. Dolgin, “Personalized cancer vaccines pass first major clinical test,” Nature Reviews Drug Discovery, 2023, 22(8), pp. 607–609.

81 Combining V940 mRNA with pembrolizumab (immunotherapy marketed on its own under the name Keytruda®) / H.E.C.T.O.R. H.E.C.T.O.R. — V940-001. 2024 [cited 15/03/2024]. J.-B. Gervais, “Un premier vaccin à ARNm contre le mélanome disponible dès 2025 ?,” Univadis, 2023 [cited 15/03/2024].

82 Here, the treatment of the pancreatic ductal adenocarcinoma involves a therapy combining an mRNA called autogene cevumeran, which can encode up to 20 neoantigens per patient, with anti-PD-L1 immunotherapy (atezolizumab) and a more conventional chemotherapy, a modified version of the FOLFIRINOX® protocol. L. A. Rojas, et al., “Personalized RNA neoantigen vaccines stimulate T cells in pancreatic cancer,” Nature, 2023, 618 (7963), pp. 144–150.

83 Although it can also include certain interleukins, antibodies such as anti-EGFR, anti-HER, etc.

84 Directive 2001/83/EC, op. cit. M. Guerriaud and E. Kohli, “RNA-based drugs and regulation”, op. cit.

85 The ICH (International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use) is an international organization that gathers European, American and Japanese regulatory authorities and members of the pharmaceutical industry to promote the harmonization of pharmaceutical regulations worldwide. Its main objective is to develop internationally accepted guidelines and standards for the quality, safety, efficacy and rational use of medicinal products. M. Guerriaud and E. Kohli, “RNA-based drugs and regulation”, op. cit.

86 ICH and EMA, ICH S12 Guideline on nonclinical biodistribution. Considerations for gene therapy products – Step 5 (EMA/CHMP/ICH/318372/2021), Committee for Medicinal Products for Human Use, Editor, 2023.

87 T. C. Roberts, R. Langer, and M. J. A. Wood, “Advances in oligonucleotide drug delivery,” op. cit. N. Toub, et al., “Innovative nanotechnologies for the delivery of oligonucleotides and siRNA,” Biomedicine & Pharmacotherapy, 2006, 60(9), pp. 607–620.

88 N. Toub, et al., ibid.

89 M. Egli, and M. Manoharan, “Chemistry, structure and function of approved oligonucleotide therapeutics,” op. cit.

90 T. C. Roberts, R. Langer, and M. J. A. Wood, “Advances in oligonucleotide drug delivery,” op. cit.

91 Committee for Advanced Therapies, “Minutes of the meeting on 02–04 December 2020,” EMA/CAT/162004/2021, 2021.

92 Committee for Advanced Therapies, “Minutes of the meeting on 15–17 July 2020. EMA/CAT/510852/2020 2020.”

93 M. Guerriaud, “Regulatory Sandboxes Could Solve the Regulatory Problems Encountered in Europe and Arising from Innovation in Biological Medicinal Products,” Pharmaceutical Medicine, 2023.

94 The advantage of both generic and biosimilar products lies in the reduction in healthcare costs thanks to a lower price, itself the result of limited development costs.

95 Parathormone (PTH) is a hormone made by the parathyroid glands that helps control the amount of calcium in the blood. It raises calcium levels by pulling calcium from bones, helping the kidneys hold on to calcium, and boosting vitamin D, which allows the body to absorb more calcium from food.

96 M. Guerriaud, “Regulatory Sandboxes,” ibid. Bundesinstitut für Arzneimittel und Medizinprodukte, “Public assessment report – decentralized procedure Teriparatid-ratiopharm 20 μg/80 ml, solution for injection [DE/H/4291/01/DC] & [DE/H/4292/01/DC],” 2017. p. 16.

97 Directive 2001/83/EC of the European Parliament and of the Council of 6 November 2001 on the Community code relating to medicinal products for human use, OJ L 311, 28.11.2001, pp. 67–128.

98 Proposal for a Regulation of The European Parliament and of the Council laying down Union procedures for the authorisation and supervision of medicinal products for human use and establishing rules governing the European Medicines Agency, amending Regulation (EC) No 1394/2007 and Regulation (EU) No 536/2014 and repealing Regulation (EC) No 726/2004, Regulation (EC) No 141/2000 and Regulation (EC) No 1901/2006. COM/2023/193 final.

99 In the aforementioned proposal, regulatory sandboxes are defined as “a regulatory framework during which it is possible to develop, validate and test in a controlled environment innovative or adapted regulatory solutions that facilitate the development and authorisation of innovative products which are likely to fall in the scope of this Regulation, pursuant to a specific plan and for a limited time under regulatory supervision.”

100 M. Guerriaud, “‘Regulatory Sandboxes’ Could Solve the Regulatory Problems Encountered in Europe and Arising from Innovation in Biological Medicinal Products,” Pharmaceut Med, 2024, 38(1), pp. 19-23. M.-D. et G. Campion, “Bacs à sable réglementaires en santé : des initiatives diversifiées à la recherche d’un cadre fédérateur au service du bien public,” RGDM – Panorama de Droit pharmaceutique 2022, 2023, 10 , pp. 233-248.

---

Mathieu Guerriaud, « The Legal Definition of Gene Therapy Medicinal Products: Addressing Challenges for Innovation and Safety », Définitions et concepts du biodroit [Dossier], Confluence des droits_La revue [En ligne], 07 | 2025, mis en ligne le 7 juillet 2025. URL : https://confluencedesdroits-larevue.com/?p=4180.