Gene Therapy and Gene Editing — Achievements, Clinical Application, and Challenges

1 Introduction

Gene therapy and gene editing are therapeutic approaches whose primary aim is to alter the genetic material of the cells in which the cause of a particular disease resides. They involve introducing correct genetic material that is missing or altered, activating or inactivating genes involved in the development of a disorder, or directly modifying genes in the patient's cells.

First and foremost, the development of these approaches offers the possibility of treating and even curing various genetic diseases that were beyond the reach of conventional medicine. Beyond this, almost all other diseases carry a certain genetic "trace" that can be acted upon in this way. Unlike conventional therapies, which most often act at the level of symptoms, gene therapy makes it possible to act on the molecular basis and the very cause of a disease, which expands the boundaries of therapeutic possibility.

This paper covers significant advances in the field of gene therapy and gene editing, analysing their clinical application as well as the regulatory, economic, and ethical challenges that determine how far these achievements will actually reach patients. Particular attention is devoted to the safety aspects of these therapies, which remain the subject of ongoing research alongside the expansion of clinical use.

The paper begins with a brief overview of the fundamentals of genetics relevant to gene therapy and editing, so that the therapeutic principles described in the chapters that follow are clear even to readers for whom genetics is not a primary field of work.

1.1 Fundamentals of genetics

The genetic code forms the basis of life. Deoxyribonucleic acid (DNA) contains the information necessary for the synthesis of proteins — the main executors of biological functions — as well as of various types of ribonucleic acids (RNA), which serve in the regulation and carrying out of cellular activities. Genetic information is recorded in the form of the sequence of nucleotides in the DNA molecule and is identical in all cells of a single organism. Individual segments of the DNA molecule that contain instructions for the synthesis of specific proteins or RNA molecules are called genes, and the complex processes of transcription and translation enable these instructions to be converted into functional molecules.

The type and quantity of proteins synthesised in a cell rest on a complex system of gene expression control. This process depends on numerous regulatory elements within the DNA itself — such as promoters, enhancers, and silencers — as well as on proteins called transcription factors, which, by responding to external and internal signals, determine which genes will be activated or inhibited. In this way, each individual cell has a specific set of active genes that allows it to acquire the structure and function characteristic of the organ system in which it is located.

From a molecular standpoint, diseases can arise in two basic ways: as a consequence of changes in the DNA sequence, or of disturbances in the regulation of gene expression. Both mechanisms lead to the same outcome — a change in the quantity, structure, or function of proteins and RNA molecules, which disrupts the normal functioning of the cell.

Changes in the DNA sequence are called mutations. Depending on the type of cells in which they arise, mutations may be germline or somatic. Germline mutations arise in reproductive cells and are passed on to offspring, whereby all cells of the new organism carry the same genetic change. This is how hereditary diseases develop, such as cystic fibrosis or haemophilia A. Somatic mutations, on the other hand, arise during the organism's lifetime in somatic cells and are not passed on to offspring. It is important to emphasise that the significance of somatic mutations stems from the fact that even in cells that do not divide, genetic information is continuously used for the synthesis of molecules necessary for their functioning. Somatic mutations are particularly significant in the development of malignant diseases, where the gradual accumulation of mutations in genes that regulate the cell cycle leads to uncontrolled cell proliferation.

Besides changes in the DNA sequence, diseases can also arise from disturbances in the regulation of gene expression, in which case the nucleotide sequence itself remains unchanged. What then changes is the level, timing, or location of expression of particular genes. An important group of these disorders comprises complex diseases, such as ischaemic heart disease, in which disturbances in the regulation of gene expression play a significant role in pathogenesis.

Contemporary research in molecular biology has made possible the development of therapies that can correct these errors at the level of the gene — gene therapy and gene editing.

2 Aim of the paper

This diploma thesis is conceived as a review paper whose fundamental aim is a systematic presentation of contemporary achievements, possibilities, and limitations of gene therapy and gene editing. Given that over the last decade this field has travelled the road from an experimental concept to clinically approved drugs, there is a need for a review of the available literature covering both the technological and the safety, regulatory, economic, and ethical aspects of these therapeutic approaches. The paper is based on an analysis of relevant professional and scientific literature, including primary research articles, reviews published in scientific journals, and official documents of regulatory bodies such as the European Medicines Agency (EMA) and the U.S. Food and Drug Administration (FDA). The ultimate aim of the paper is to provide a comprehensive picture of the current state of the field, with indications of possible directions for further development. In doing so, it underscores the importance of gene therapy and gene editing in contemporary medical practice. At the same time, it points out that their further success does not depend solely on scientific and technological achievements, but also on the broader social context in which these therapies are applied.

3 Gene therapy and gene editing

Gene therapy and gene editing are therapeutic approaches whose ultimate aim is the functional modification of genes involved in the development of a particular disease. One of the conceptually clear definitions of gene therapy states that it "involves the treatment of disease by introducing specific genetic material into the patient's body, which directly affects a change in the function of the target cells" [1].

It should be noted that there are therapeutic approaches based on RNA molecules that act at the level of gene expression regulation, but without permanently altering the genetic material. Approaches based solely on such regulation, such as small interfering RNAs (siRNA) and antisense oligonucleotides, will not be the primary focus of this paper [2].

Gene therapy in the narrower sense is traditionally defined as a therapeutic approach in which the effect is achieved by introducing recombinant nucleic acids that bring about a change in gene expression, whereby the therapeutic action is directly tied to the presence or expression of the introduced nucleotide sequence [3]. This understanding, which long dominated the European regulatory framework, primarily encompasses strategies based on the addition or replacement of genes, most often by means of viral vectors.

By contrast, gene editing involves the precise alteration of the host's existing nucleotide sequences, without the necessary introduction of new recombinant DNA that would be permanently expressed in the cell [4]. These approaches, including the best-known CRISPR–Cas9 system, base editing, and prime editing, achieve their therapeutic effect through direct modification of the genome, with the transient presence of proteins and RNA. The most widely applied is the CRISPR–Cas9 system, which in natural conditions functions as part of the immune response of bacteria and archaea. In gene editing, this system is used as a molecular tool that makes targeted breaks in the host's DNA, thereby enabling the precise alteration or inactivation of a desired gene.

Although gene editing, from a biological and functional standpoint, represents a continuation and extension of the concept of gene therapy, regulatory frameworks — particularly in the European Union — have historically not clearly defined and delineated these approaches. In newer guidelines, a tendency can be observed to include gene editing within the broader concept of gene therapy, which confirms the need for further alignment of regulatory terminology with contemporary scientific and technological development [5]. This will be considered in more detail in the chapter on the challenges of applying gene therapy.

Generally speaking, methods of treatment that act directly at the level of the gene can be classified into several basic approaches:

1. Introducing correct genes into the patient's cells that do not contain a particular gene in functional form

In this way, adequate synthesis is established of a protein that had until then been absent or present in insufficient quantity, and that is needed for the physiological course of a particular cellular process. This approach has been present in medical practice the longest and is used in the majority of drugs approved to date. The therapeutic goal is achieved by introducing the necessary DNA sequence into the target cell by means of an appropriate carrier (vector), or by directly modifying the target gene using gene editing techniques [6,7].

2. Regulation of genes within the patient's cells

Regulation refers to managing gene expression without altering the DNA sequence itself. This involves increasing or decreasing the amount of protein a particular gene encodes, by acting either on the RNA molecules produced by its transcription or on regulatory elements. The most common approaches include therapies based on RNA molecules, such as RNA interference and antisense oligonucleotides, by which a reduction in gene expression is achieved [8–10]. Regulation can also be achieved by CRISPR-mediated approaches that do not alter the DNA sequence, such as CRISPRi and CRISPRa, which reversibly change the transcription of the targeted gene [11,12].

3. Elimination of defective genes

The permanent disabling of a gene is achieved by gene editing techniques, by which a target gene carrying a harmful mutation or a pathologically increased expression is functionally inactivated [13]. Unlike regulatory approaches, which reduce gene expression temporarily, elimination represents a permanent solution. This approach makes it possible to functionally "switch off" the target gene, thereby permanently preventing the synthesis of a harmful or overly abundant protein.

4 Gene therapy

Traditional gene therapy is a therapeutic approach in which the treatment of a disease is based on introducing a functional copy of the gene of interest into the patient's cells, with the aim of compensating for the lost or damaged gene function. This approach is referred to as gene addition, that is, functional gene replacement. The therapeutic effect is achieved not by altering the existing genome, but through the expression of the introduced gene, independently of the state of the host's gene locus and its regulatory elements. As the most frequently applied strategy among approved drugs, gene addition will serve in this chapter as the basis for presenting the general principles and stages of gene therapy, while approaches based on genome editing are considered separately.

For gene therapy to be successful, it is necessary that the nucleic acid that forms the core of the therapy contains the appropriate genetic sequence, that the therapeutic product reaches the target tissue, that controlled expression of the introduced gene is enabled, as well as its long-lasting action in the cells. These requirements determine the basic components of gene therapy, which will be considered in the chapters that follow.

4.1 Recombinant nucleic acids as therapeutic agents

In gene therapy, the therapeutic effect is achieved by introducing into the target cell a nucleic acid that contains the gene of interest or a regulatory element. The therapeutic gene is most often designed using software and synthesised in vitro, together with the necessary regulatory elements that make up a functional genetic construct [14–16].

After synthesis, the gene is incorporated into a plasmid — a small, circular, double-stranded DNA molecule of bacterial origin that can be replicated independently. This produces recombinant DNA, a molecule that contains genetic material from different biological sources: the synthetic gene and the bacterial plasmid. The plasmid serves as a molecular carrier that enables the synthesised gene to be multiplied into a large number of copies using the bacterial replication apparatus [14].

It is important to distinguish between a plasmid and a vector. A plasmid is a DNA molecule used as a starting structure from which the therapeutic gene is transferred into more complex delivery systems, whereas a vector is a broader functional concept denoting a system capable of transferring genetic material into the patient's target cell and enabling its expression.

The reason for using a vector is the inability of therapeutic recombinant DNA to spontaneously cross the cell membrane, owing to its large molecular mass and negative charge. Even if it does reach the cell, it is quickly degraded by cellular endonucleases [7].

4.2 Vectors for gene transfer

Since a plasmid on its own is not a means of delivering a gene into the patient's cells, the next step in designing a therapeutic product involves the selection and use of a vector. Vectors enable the efficient transfer of the therapeutic gene from the laboratory environment into the host's cells and its appropriate expression. Gene therapy uses viral and non-viral vectors, which differ from one another in the efficiency of gene transfer, safety profile, and ability to ensure long-lasting expression of the therapeutic gene.

An ideal vector for gene therapy should possess a range of characteristics that would ensure the safety and efficacy of the therapy. From the standpoint of efficacy, what is needed is a sufficient capacity to accommodate therapeutic genes, high efficiency of transduction of target cells (including post-mitotic ones), as well as long-lasting and stable expression of the introduced gene specifically in the target cell type. From the safety standpoint, it is desirable that the vector not integrate randomly into the host genome, that it not provoke an immune response or inflammation, and that it not be pathogenic. In addition to biological characteristics, of practical importance are the possibility of production in large quantities and an acceptable cost, which are key preconditions for wider clinical application [7,17,18].

4.2.1 Viral vectors in gene therapy

The natural ability of viruses to introduce their genetic material into the cells of other species makes them particularly suitable for use in gene therapy, where they represent the most frequently used systems for delivering therapeutic genetic material. For therapeutic purposes, viruses are modified so that the genes responsible for pathogenicity and replication are removed or inactivated, after which an appropriate therapeutic genetic sequence is inserted into the viral genome. In this way, biological carriers are obtained that retain the natural ability to enter cells but lack the capacity to cause disease or to multiply autonomously in the organism.

One of the key characteristics of viral vectors is their tropism — that is, the natural tendency of a virus to recognise and efficiently infect certain cell types or tissues. Tropism is determined above all by the structure of the proteins on the surface of the viral capsid or envelope, which bind to specific receptors on the host cell membrane. Different viruses, and even different variants of the same virus (serotypes), may show a different affinity for particular cells, such as hepatocytes, neurons, or myocytes. This natural selectivity is used in gene therapy to deliver the therapeutic gene precisely to target tissues, with minimal expression in other tissues and organs, which increases the efficacy of the therapy and reduces the risk of adverse effects.

Depending on the type of virus, viral vectors can provide long-lasting or temporary expression of the therapeutic gene, in that their genetic material either integrates into the host genome or is maintained in episomal form (DNA that is located in the cell nucleus but is not integrated into the host genome). The most frequently used viral vectors in gene therapy include adenoviral (AV), adeno-associated viral (AAV), and lentiviral (LV) vectors. Each of these systems possesses specific characteristics with respect to gene transfer capacity, tropism for particular cell types, duration of expression, and safety profile.

Although viral vectors enable high efficiency of gene delivery, their use is also associated with certain limitations, such as a potential host immune response, limited capacity for the uptake of large genetic sequences, and the risk of insertional mutagenesis with lentiviral vectors. These limitations represent some of the key challenges of contemporary gene therapy and will be considered in more detail in the chapter on the safety of gene therapy.

4.2.1.1 Adenoviral vectors

Adenoviral vectors are among the earliest developed viral vectors in gene therapy. They are DNA viruses whose genome consists of a double-stranded DNA molecule. The advantages of adenoviral vectors are as follows: they are characterised by high transduction efficiency of both dividing and post-mitotic cells, tropism for a large number of different tissues, and a relatively large capacity to accommodate therapeutic genes. In addition, reliable systems are available for their production in large quantities.

The genetic material introduced by adenoviral vectors remains in episomal form — it does not integrate into the host genome. Expression of the therapeutic gene is strong but most often temporary, which makes adenoviral vectors suitable for applications requiring an intense but time-limited therapeutic effect, such as oncological indications and immunomodulation [19].

The main limitation of the use of adenoviral vectors is their pronounced immunogenicity, which manifests itself in two ways. First, since most of the general population has previously been exposed to adenoviruses, existing neutralising immunity can lead to rapid clearance of the vector and a reduction in therapeutic efficacy. Second, the administration of high doses can provoke an acute systemic inflammatory response with serious safety consequences. The best-known such case is the death of Jesse Gelsinger, a patient who in 1999 took part in a clinical trial of gene therapy for ornithine transcarbamylase deficiency, and in whom the administration of a high dose of an adenoviral vector provoked a systemic immune response that led to multi-organ failure and a fatal outcome [20,21]. This event significantly slowed the development of the field and is one of the reasons why adenoviral vectors are today applied in a limited number of clinical contexts, primarily in those in which a temporary but strong therapeutic effect is sufficient.

4.2.1.2 Adeno-associated viral vectors

Adeno-associated viruses are today the most frequently used viral vectors. The widespread use of AAV vectors rests on several key advantages: the genetic material introduced by AAV vectors largely remains in episomal form, AAV are not associated with any known human disease, and AAV serotypes are numerous and of differing tropism [1,22]. Thus, for example, AAV9 shows a pronounced ability to cross the blood–brain barrier and transduce neurons [23], which makes it suitable for the therapy of diseases of the central nervous system, while AAV8 has a pronounced affinity for hepatocytes and is most often used in the therapy of liver diseases [24]. This capacity for targeted delivery of the therapeutic gene to specific tissues is one of the most important properties of AAV vectors.

By contrast, the fundamental limitation of the AAV system stems from the very nature of the virus — small DNA viruses with a single-stranded genome can accommodate only about 4.7 kilobases of therapeutic genetic material, which is significantly less than the capacity of adenoviral or lentiviral vectors. This limitation poses a challenge in treating diseases whose causative genes are large, as is the case with Duchenne muscular dystrophy, where the gene for dystrophin exceeds the capacity of AAV vectors. As a solution, shortened variants of the therapeutic gene have been developed, such as the gene for microdystrophin [25], which retain the basic functional domains and can be accommodated within a single AAV vector.

A second significant limitation is the immunogenicity of the AAV capsid. Although considerably less pronounced than with adenoviruses, a significant percentage of the general population carries neutralising antibodies against particular AAV serotypes, with a prevalence that depends on the serotype and the geographic region. The presence of these antibodies can significantly reduce the efficacy of the therapy or make repeat administration of the same vector impossible, which is a particular problem in therapies whose effect wanes over time owing to the division of target cells. In addition to the humoral response, the systemic administration of high doses of AAV vectors is also associated with rarer but serious adverse effects such as hepatotoxicity, thrombotic microangiopathy, and damage to the dorsal root ganglia of the spinal cord [24].

Despite these limitations, AAV vectors are a pillar of contemporary in vivo gene therapy. Several AAV-based drugs have been approved for clinical use, and the specific indications and mechanisms of their action will be considered in the chapter on the clinical application of gene therapy.

4.2.1.3 Lentiviral vectors

Lentiviral vectors belong to the retrovirus family and were developed on the basis of HIV-1. They are RNA viruses which, after entering the cell, transcribe their genome into DNA by means of reverse transcriptase, and then permanently integrate that DNA into the genome of the target cell [26]. To ensure the safety of their use, lentiviral vectors have been significantly modified relative to the wild type of the virus — the genes necessary for replication and pathogenicity have been removed, and the components required for packaging the vector have been separated onto several plasmids, which prevents the emergence of replication-competent viruses during production [27,28]. Their capacity for stable integration makes them particularly suitable for therapies where long-lasting expression of the therapeutic gene is required, especially in dividing cells such as haematopoietic stem cells and T lymphocytes [1]. Another advantage is their relatively large capacity for the therapeutic gene (about 8–10 kilobases) [17].

The fundamental limitation of lentiviral vectors is the risk of insertional mutagenesis. Given that the viral DNA integrates randomly into the host genome, there exists a theoretical possibility of activating proto-oncogenes or disrupting tumour suppressor genes. This risk was pronounced with the earlier-studied gamma-retroviral vectors and led to several cases of leukaemia in clinical trials in children with X-linked severe combined immunodeficiency (X-SCID) [29]. With contemporary lentiviral vectors, the risk has been significantly reduced thanks to the use of a self-inactivating (SIN) design, in which the viral regulatory elements that could trigger the expression of neighbouring host genes after integration have been removed [30]. In clinical studies with modern lentiviral vectors, no clinically significant clonal growth of transduced cells has been observed [17].

In contemporary gene therapy, lentiviral vectors are the vector of choice for ex vivo approaches (when cells are modified outside the patient's body), and particularly for therapies that require the modification of haematopoietic stem cells.

4.2.2 Non-viral vectors in gene therapy

Non-viral vectors represent a significant alternative to viral systems. Among non-viral vectors, the most prevalent are lipid nanoparticles (LNP) and N-acetylgalactosamine (GalNAc) conjugates. LNP are spherical structures used for the delivery of siRNA, mRNA, and CRISPR-Cas9 components. Unlike viral vectors, in which the therapeutic sequence is placed into the viral capsid through the natural mechanisms of viral replication and assembly, these vectors bind the nucleic acid through electrostatic interactions between positively charged lipids or polymers and the negatively charged phosphates of the nucleic acid. This interaction leads to the encapsulation of the therapeutic construct within the nanoparticle, which simultaneously protects the genetic material from degradation by endonucleases in physiological fluids and enables its passage across the cell membrane [31].

GalNAc conjugates use a different principle compared with LNP — instead of encapsulation within a nanoparticle, the therapeutic nucleic acid is covalently bound directly to a sugar molecule that is naturally taken up from the circulation into the liver. The asialoglycoprotein receptor on hepatocytes physiologically serves to remove "worn-out" glycoproteins, recognising their exposed GalNAc residues. By binding GalNAc directly to an siRNA or antisense oligonucleotide, the therapeutic molecule is effectively "presented" to the liver as a natural substrate and selectively accumulates in hepatocytes.

The route of administration and distribution in the organism depend on the formulation and the indication. LNP are most often administered intravenously, intramuscularly (as with the mRNA COVID-19 vaccines), or locally; after systemic administration, they accumulate predominantly in the liver owing to the absorption of apolipoprotein E from the circulation, which directs them towards hepatocytes via LDL receptors. GalNAc-conjugated drugs are administered subcutaneously and, thanks to their high specificity for hepatocytes, also end up in the liver [19].

The main limitations of non-viral vectors are their lower efficiency compared with viral vectors, their limited tissue distribution (their predominant accumulation in the liver narrows the possibilities for their clinical application), and the challenges in the efficient delivery of larger complexes such as the Cas9 ribonucleoprotein. Despite this, LNP technology has already made possible the personalised CRISPR therapy of KJ Muldoon described in the chapter on clinical applications, illustrating the growing clinical significance of this class of vectors [32].

4.3 Targeted delivery of the therapeutic gene

A key condition for successful gene therapy is that the therapeutic gene reach the appropriate cell type. Inadequate delivery can significantly limit the therapeutic effect and increase the risk of adverse effects. Targeting can be achieved by selecting a vector with appropriate tropism, as well as by the mode of administering the therapy.

4.3.1 Ex vivo and in vivo approaches

Gene therapy can be carried out in two different ways — in vivo and ex vivo. The ex vivo approach involves isolating target cells from the body of the patient or a donor, their genetic modification under laboratory conditions, the multiplication of the modified cells in cell culture, and the return of those cells to the patient's body [1,33]. This approach is most often used in diseases affecting the cells of the blood and the immune system, where haematopoietic stem cells or T lymphocytes are modified. The main advantage of the ex vivo approach is a high degree of control over the process of genetic modification — it is possible to select only successfully transduced cells, verify their quality, and ensure an appropriate dose before returning them to the patient. In addition, the patient's exposure to the vector itself is limited, which reduces the risk of a systemic immune response. On the other hand, ex vivo therapy requires a complex laboratory infrastructure, is more expensive, and is technically more demanding to produce [33,34]. The most frequently used vectors for the ex vivo approach are lentiviral vectors, owing to their ability to integrate stably into the genome of dividing cells [17,18].

In the in vivo approach, the therapeutic vector is administered directly into the patient's body, most often by intravenous infusion or local injection into the target tissue [1,35]. This approach is particularly useful in diseases in which the target cells are difficult to extract and manipulate outside the organism, such as neurons, hepatocytes, or retinal cells [36,37]. Targeting the appropriate tissue in these cases relies on the tropism of the vector used and the mode of administration, whereby successful transduction is achieved without the need to isolate and process cells in the laboratory. For these reasons, in vivo administration is most often carried out using AAV vectors, whose favourable safety profile and ability to transduce certain tissues in a targeted manner enable efficient delivery of the therapeutic gene with a reduced risk of adverse effects [35,37].

4.3.1.1 Local and systemic administration

The in vivo approach involves two modes of administering the therapy — local and systemic — depending on the location of the target tissue and the clinical characteristics of the disease.

In local administration, the vector is introduced directly into the target tissue — for example, by subretinal injection into the eye [38]. Local administration makes it possible to achieve high concentrations of the vector in the target tissue with minimal systemic exposure, thereby reducing the risk of an immune response and of adverse effects in other organs. Such an approach is optimal in diseases affecting anatomically confined and accessible tissues [39].

Systemic administration, most often by intravenous infusion, is used when the therapeutic gene needs to reach several tissues or tissues that are less accessible to local injection. Thanks to the tropism of particular AAV vector serotypes, it is possible to achieve preferential transduction of certain tissues. Nevertheless, systemic administration involves the use of significantly higher doses of the vector, which increases the risk of a systemic immune response and of toxicity [40].

4.4 Expression of the therapeutic gene and duration of the therapeutic effect

Once the therapeutic nucleic acid reaches the target cell, its expression must be adequately regulated in order to achieve the desired therapeutic effect. The level of expression is determined by a combination of factors such as the regulatory elements built into the vector, the mode of maintaining the genetic material in the cell, and the characteristics of the target tissue.

4.4.1 Regulation of the expression of the introduced gene

The expression of the introduced gene is controlled by regulatory elements, above all promoters and enhancers, which are built into the vector together with the therapeutic gene. A promoter is a DNA sequence that determines the site at which transcription begins and the strength of expression, while enhancers additionally amplify transcription. The choice of regulatory elements is one of the key decisions in the design of the therapeutic product, since it directly affects both the efficacy and the safety of the therapy.

According to their mode of regulation, promoters can be divided into two basic groups [41]. Constitutive promoters enable constant expression of the therapeutic gene in all cell types the vector reaches and are most often used when the therapeutic gene needs to be active in a large number of different tissues. An example is the cytomegalovirus promoter, which provides strong and predictable expression [42]. Tissue-specific promoters, on the other hand, enable expression of the therapeutic gene only in certain cell types. Such is the promoter of the RPE65 gene, which is used in the therapy of retinal diseases and enables expression predominantly in the cells of the retinal pigment epithelium [38]. The use of tissue-specific promoters restricts the expression of the therapeutic gene to the desired tissue, reduces the risk of an immune response to the therapeutic protein, and reduces potential adverse effects in other tissues.

It is important to emphasise that excessive expression of the therapeutic gene can in itself pose a risk. In certain indications, levels of the therapeutic protein higher than physiological ones can act toxically. Thus, in preclinical studies on primates and pigs, it was shown that high expression of the SMN protein after the administration of an AAV vector can cause toxicity of the dorsal root ganglia of the spinal cord and proprioceptive deficits [43]. For this reason, in the design of the therapeutic product, the combination of promoter, vector dose, and mode of administration is carefully chosen so as to achieve a therapeutically effective but also safe level of expression.

4.4.2 Duration of the therapeutic effect

The duration of the therapeutic effect is largely determined by the way in which the therapeutic genetic material is maintained in the cell. Integrative vectors incorporate the therapeutic gene into the host genome, which enables its stable inheritance during cell divisions and often ensures a lasting therapeutic effect. By contrast, AAV vectors remain in episomal form and, in tissues with active proliferation, are gradually lost during cell divisions, which can lead to a decline in the therapeutic effect over time.

5 Gene editing

Gene editing is one of the tools by which the aims of gene therapy in the broader sense are achieved — a method by which the structure of the genetic sequence is directly altered. Using various types of specific nucleases, a break is induced in one or both strands of the DNA molecule at the desired sites. This damage is registered by proteins called DNA damage sensors, which triggers the physiological mechanism of DNA repair [44]. There are two main types of repair mechanism — non-homologous end joining and homology-directed repair [45].

Non-homologous end joining (NHEJ) is a less precise mechanism that directly connects the two ends of a double-strand DNA break [46]. Under physiological conditions this joining often takes place without changes in the sequence, but with repeated breaks at the same site there arise small errors — the addition or loss of a few nucleotides (so-called indel mutations), as well as larger deletions and chromosomal rearrangements [47–49]. Such changes most often disrupt the reading frame of the gene, which leads to the synthesis of a messenger RNA with a premature stop codon (which is removed by the cellular mechanism of nonsense-mediated decay) [50] or of a non-functional protein. In this way the target gene is inactivated, which forms the basis of one line of therapy aimed at "switching off" genes [13].

Homology-directed repair (HDR) is a considerably more precise mechanism that uses the sister chromatids as a template for repair, and is therefore, under physiological conditions, active only in dividing cells [51,52]. For therapeutic purposes, this pathway is used by introducing into the cell, simultaneously with the induction of the double-strand break, a synthetic DNA template that contains the correct sequence of the desired gene. Repair of the break with the help of that template enables the precise replacement of the mutated segment or the introduction of an entire gene, in both cases restoring the function of the target gene [13]. The restriction of this mechanism to actively dividing cells poses a significant obstacle to its application in post-mitotic cells such as neurons, cardiomyocytes, and hepatocytes — tissues that are frequent targets of gene therapy. This limitation is one of the key factors that has spurred the development of alternative approaches to gene editing that do not require cell division [53], which will be discussed below.

5.1 Systems for gene editing

The described mechanisms of inducing double-strand DNA breaks and of their repair form the common basis of contemporary technologies for gene editing. Before the development of the CRISPR system, gene editing relied on zinc-finger nucleases (ZFNs) and TALEN systems, which likewise used specific nucleases to create DNA breaks at desired sites. However, both systems required the design of entirely new proteins for each new gene locus — a process that is extremely technically demanding, expensive, and time-consuming, so their wider application in research and clinical practice was impractical [54].

It is precisely for this reason that the appearance of the CRISPR–Cas9 system marked a turning point in the field of gene editing. CRISPR–Cas9 set itself apart as the dominant technology in research and clinical practice. The reason for this is the conceptual simplicity of this system — targeting the desired gene locus is achieved by synthesising a short RNA complementary to the target DNA, instead of designing new proteins. In addition, the system is exceptionally adaptable, allows the simultaneous targeting of several loci, and is considerably more accessible in terms of technical and financial requirements compared with the older methods [54,55]. For this reason, the mechanism of the CRISPR–Cas9 system and its application in gene editing are considered separately below.

5.2 The CRISPR–Cas9 system

The CRISPR–Cas9 system is part of the adaptive immune response of bacteria and archaea, which enables the recognition and neutralisation of pathogens. Namely, upon a first infection, short segments of the pathogen's DNA (most often of a bacteriophage) are incorporated into the host genome in the form of so-called spacer sequences, which are permanently incorporated into the bacterial CRISPR gene locus and are specific to the pathogens the microorganism has previously encountered. The CRISPR locus consists of spacer sequences located between repeating DNA motifs. Spacers enable the specific recognition of genetically similar pathogens during subsequent infections, while the DNA motifs (repeats) that separate individual spacer sequences have a structural and regulatory role. Cas proteins, including the Cas9 endonuclease, are encoded by separate cas genes located in the vicinity of the CRISPR locus.

During repeated exposure to a pathogen, the CRISPR locus is transcribed into a long precursor RNA, which is further processed into short CRISPR RNAs (crRNA), each of which contains a sequence complementary to a particular foreign DNA. The crRNA molecules, in a complex with an auxiliary trans-activating crRNA (tracrRNA) and the Cas9 protein, connect with the DNA sequence of the pathogen present through complementary base pairing. In order for binding to the target sequence and the endonuclease action of Cas9 that will inactivate the pathogen to occur, the presence of a short PAM motif (protospacer adjacent motif) in the foreign DNA is also necessary, which ensures the distinction between one's own and foreign genetic information. For the most frequently used Streptococcus pyogenes Cas9, the PAM has the form NGG, where N can be any base (A, T, C, or G) and G are two guanine bases. After correct recognition, the Cas9 protein induces a double-strand break in the DNA molecule of the pathogen, which leads to its functional inactivation and the protection of the cell [56].

This natural mechanism has been adapted for the purposes of gene editing. The adaptation involves designing an artificial synthetic RNA that will take on the role that crRNA and tracrRNA have in the bacterial system. The synthesised RNA is called the guide RNA and is used to "programme" the CRISPR system. The guide RNA is designed to be complementary to the desired sequence upon which one wishes to act, to which it binds, after which the Cas9 protein induces a double-strand break in the DNA molecule. Given that an unmodified bacterial Cas9 protein is generally used, it is necessary that the above-mentioned PAM motif also be present on the human cell's gene. Although its presence in this case is not significant for distinguishing one's own DNA from foreign DNA, it is necessary for the activation and function of the Cas9 protein. With this in mind, the guide RNA is designed to contain about twenty nucleotides located immediately upstream of a PAM (NGG or another) that occur naturally within the gene sequence to be manipulated. In this way, breaks are formed in human DNA and it becomes possible to activate, in a targeted manner, the cellular mechanisms of DNA repair — non-homologous end joining or homology-directed repair — which enables the inactivation of a gene, the correction of mutations, or the precise introduction of new genetic sequences [54].

5.3 Base editing and prime editing

Although CRISPR–Cas9 has contributed considerably to gene editing, its application is accompanied by certain limitations. Above all, there is the risk of unwanted indel mutations arising from non-homologous joining, while the applicability of homology-directed repair is greatly limited in post-mitotic cells (those that no longer divide). Advanced variants of gene editing that do not require the induction of double-strand DNA breaks or cell division are under development.

5.3.1 Base editing

Base editing enables the direct conversion of one nitrogenous base into another within the target sequence, without introducing double-strand DNA breaks and without the need for a DNA template. This technology uses an inactivated Cas9 protein conjugated with a deaminase enzyme. The Cas9 component may be fully inactivated — so-called "dead" Cas9 (dCas9) — or only partially inactivated in the form of Cas9 nickase (nCas9), which makes a single-strand break instead of a standard double-strand DNA break. The deaminase is an enzyme that chemically alters a nitrogenous base — depending on the type of base editor, a cytidine deaminase is used for the conversion of cytosine to thymine (C→T editors) or an adenine deaminase for the conversion of adenine to guanine (A→G editors). These two components are joined together into a single functional protein, which is directed to the target site in the genome with the help of the guide RNA.

When this protein reaches the target sequence, the Cas9 component separates the double-stranded DNA and exposes the single-stranded portion on which the deaminase performs the chemical modification of the appropriate base. After deamination, the Cas9 nickase makes a single-strand break on the opposite DNA strand — the one that does not contain the modified base. This break prompts the cellular mismatch repair mechanism to use the modified strand as a template, whereby the change is permanently incorporated into both strands of the DNA molecule. In this way, a precise, permanent replacement of a single base is achieved without inducing a double-strand break and without the need for an external DNA template [57,58].

5.3.2 Prime editing

Prime editing further expands the possibilities of precise gene editing. This system combines the Cas9 nickase with a reverse transcriptase and a specially designed guide RNA (prime editing guide RNA — pegRNA) that contains within itself the sequence serving as the template for the change. The reverse transcriptase uses the pegRNA as a matrix for the synthesis of a new DNA sequence directly at the site of the single-strand break, by which all types of substitutions, as well as small insertions and deletions, can be carried out with high precision, and without double-strand DNA breaks [59].

After the synthesis of the new sequence, a break arises at the site of editing in the form of two competing DNA ends — the newly synthesised end that contains the desired change and the original end with the unchanged sequence. Cellular enzymes of the flap endonuclease family remove the original, unchanged end, while DNA ligase joins the newly synthesised end with the rest of the DNA strand. In this way the desired change is incorporated into one DNA strand, after which the cellular mismatch repair mechanism uses the changed strand as a template to repair the opposite one, whereby the new sequence is permanently incorporated into both strands of the DNA molecule [60].

The limitations of this technology include a somewhat lower efficiency compared with the classical CRISPR–Cas9 system, the need for a more complex design of the pegRNA, as well as potential off-target effects that are still the subject of intensive research [61].

6 Clinical application of gene therapy and gene editing

The development of gene therapy from an idea to approved drugs is one of the significant advances of contemporary medicine. The lasting treatment of a disease through a single intervention at the level of the genome is becoming a reality for an ever-greater number of patients. Up to the time of writing this paper, several gene therapy products have received regulatory approval from the FDA, EMA, or equivalent agencies, covering a broad spectrum of diseases — from rare monogenic disorders to malignant diseases [62,63].

6.1 The first approved therapies

The first commercially approved gene therapy drug in the world was Gendicine — an adenoviral vector carrying the gene for the tumour suppressor protein p53, approved in 2003 in China for the treatment of head and neck cancer [64]. The approval of Gendicine is a historically significant moment, but later experience showed that adenoviral vectors have significant limitations, above all owing to the strong immune response triggered in reaction to their presence, which reduces the efficacy and duration of the therapeutic effect.

In Europe, the first gene therapy to receive approval was Glybera, a drug based on the AAV1 vector that delivers the gene for lipoprotein lipase to patients with a hereditary deficiency of this enzyme, approved by the EMA in 2012 [65]. Glybera entered the history of medicine in another way as well, as it became known as the "first million-euro drug" [66], which sparked debate about the availability and cost of gene therapy. This drug was withdrawn from the market in 2017 owing to the small number of patients treated, difficulties in its administration, and its high price [67]. Its short lifespan on the market did not diminish its scientific significance; it opened the door to the development of the next generation of AAV-based drugs that followed.

6.2 Haematological disorders

Haematological diseases are among the first in which gene therapy achieved clinically significant results. The reason for this lies in the characteristics of the diseases themselves and the accessibility of the target cells. Haematopoietic stem cells can be relatively easily isolated from the patient's blood or bone marrow, modified under laboratory conditions, and then returned to the patient's body.

6.2.1 Sickle cell anaemia and beta-thalassaemia

Casgevy (exagamglogene autotemcel) occupies a special place in the history of gene therapy as the first approved drug to use CRISPR/Cas9 gene editing technology, approved at the end of 2023 for the treatment of sickle cell anaemia and transfusion-dependent beta-thalassaemia [68,69]. To understand the mechanism of action of this drug, it is first necessary to understand the nature of the diseases it treats. Both sickle cell anaemia and beta-thalassaemia arise owing to a disturbance in the structure or synthesis of beta-globin, a protein that is a constituent part of haemoglobin in the erythrocytes of adults. The therapeutic goal is the activation of the synthesis of foetal haemoglobin, which does not contain beta-globin. Namely, during foetal development, erythrocytes produce foetal haemoglobin, which contains gamma-globin instead of beta-globin. Foetal haemoglobin is functional and does not undergo the change in shape characteristic of sickle cell anaemia. After birth, the production of foetal haemoglobin is gradually shut down under the influence of the BCL11A protein, which acts as a transcriptional repressor of the gene for gamma-globin.

Casgevy acts directly on this biological mechanism. Haematopoietic stem and progenitor cells are isolated from the patient's blood and then edited ex vivo using the CRISPR/Cas9 system. The target site of editing is the enhancer of the BCL11A gene that is active exclusively in erythroid cells, which means that its inactivation does not disrupt the function of BCL11A in other tissues, such as immune cells. With the loss of this enhancer, BCL11A can no longer sufficiently suppress the genes for gamma-globin, so erythroid cells once again begin to produce foetal haemoglobin. The modified stem cells are returned to the patient's body after myeloablative conditioning, a procedure by which the existing bone marrow cells are removed in order to ensure space for the engraftment of the edited cells. Clinical trials showed that 91% of patients became independent of transfusions, and in patients with sickle cell anaemia an almost complete absence of the vaso-occlusive crises characteristic of this disease was recorded over a period of 12 months or more [68,69].

In addition to Casgevy, drugs based on gene therapy without genome editing are also available for the treatment of these diseases. Zynteglo (betibeglogene autotemcel), a lentiviral ex vivo therapy, introduces a functional copy of the gene for beta-globin into the autologous haematopoietic stem cells of a patient with beta-thalassaemia [70]. Lyfgenia (lovotibeglogene autotemcel) uses a similar approach for the treatment of sickle cell anaemia, but introduces the gene for a variant of beta-globin that prevents the polymerisation of haemoglobin S [71].

6.2.2 Haemophilia A and B

Haemophilia A and B are hereditary coagulation disorders caused by a deficiency of factor VIII and factor IX, respectively. Patients with these disorders are dependent on lifelong infusions of coagulation factors — a therapy that is expensive, demanding, and does not provide a lasting solution. Gene therapy offers the possibility of a single treatment that would establish lasting production of the missing factor.

Hemgenix (etranacogene dezaparvovec), an AAV5-based therapy, delivers the highly active Padua variant of factor IX to hepatocytes by means of a single intravenous infusion [72]. The Padua variant of factor IX possesses eightfold greater activity than normal factor IX, which is particularly important bearing in mind that AAV vectors cannot infect all liver cells, so the higher activity of the protein compensates for the uneven distribution of the vector. BBM-H901 is another drug that acts on a similar principle, approved in China in 2025 [73]. Roctavian (valoctocogene roxaparvovec) is an AAV5-based therapy for haemophilia A, in which a shortened but still functional copy of the gene for factor VIII is delivered to hepatocytes [74].

6.2.3 Primary immunodeficiencies and leukodystrophies

Strimvelis (for combined immunodeficiency caused by adenosine deaminase deficiency (ADA-SCID) [75]), Skysona (for cerebral adrenoleukodystrophy [76]), and Libmeldy (for metachromatic leukodystrophy [77]) are therapies based on a similar principle. The patient's autologous CD34+ haematopoietic stem cells are transduced ex vivo with an integrative viral vector carrying a functional copy of the missing gene, after which the modified cells are returned to the patient's body and become a permanent source of the missing protein.

ADA-SCID is a genetic disease in which the absence of the enzyme ADA leads to the accumulation of toxic metabolites that destroy lymphocytes, leaving the patient almost without a functional immune system. Cerebral adrenoleukodystrophy arises from a mutation in the ABCD1 gene, whose product enables the breakdown of long-chain fatty acids in peroxisomes. Without it, these fatty acids accumulate in the nervous system and lead to the progressive destruction of myelin. Metachromatic leukodystrophy arises owing to a deficiency of the enzyme arylsulfatase A (ARSA), which leads to the accumulation of sulfatides in the nervous system and the gradual loss of the myelin sheath of neurons. In all three drugs, the modified CD34+ cells migrate to the tissues affected by the disease (including the central nervous system) and ensure the continuous production of the missing protein.

6.3 Neurological and muscular disorders

6.3.1 Spinal muscular atrophy

Spinal muscular atrophy (SMA) is a genetic disease caused by a mutation in the SMN1 gene, which encodes a protein necessary for the survival of motor neurons. Without this protein, the motor neurons in the spinal cord gradually degenerate, which leads to progressive muscle weakness. The most severe form of this disease, SMA type I, until recently led to a fatal outcome in the first years of life.

Zolgensma (onasemnogene abeparvovec), an AAV9-based gene therapy, delivers a functional copy of the SMN1 gene to motor neurons by means of a single intravenous infusion [78]. The key characteristic that makes this therapy possible is the natural tropism of the AAV9 capsid for neurons and its ability to cross the blood–brain barrier after intravenous administration. Once it reaches the motor neurons, the vector remains in episomal form and ensures long-lasting expression of the SMN1 gene. Clinical trials showed that children treated with this drug before the onset of symptoms in a large number of cases achieve motor milestones (sitting, standing, and walking independently) that were unimaginable before the introduction of gene therapy.

In addition to Zolgensma, Spinraza (nusinersen) is also available for the treatment of SMA — an ASO therapy that does not alter the genome but modifies the splicing of the SMN2 gene, increasing the production of functional SMN protein [79]. While Zolgensma offers a single treatment, Spinraza requires lifelong intrathecal dosing but is approved for all ages and types of SMA, unlike Zolgensma, which is approved for the treatment of SMA type I in children up to two years of age [80]. These two therapies are not mutually exclusive; in certain cases they are considered even as a complementary approach.

6.4 Ophthalmological disorders

The eye is in many respects an ideal site for the application of gene therapy. It is relatively isolated from the rest of the organism thanks to the blood–retinal barrier, has limited immune surveillance, is easily accessible for local administration, and can be monitored non-invasively by various imaging methods. In addition, many retinal diseases are monogenic, which means that the cause of the disorder is known and the target of the therapy clearly defined.

6.4.1 Inherited retinal dystrophies

Luxturna (voretigene neparvovec) is an AAV2-based gene therapy for the treatment of inherited retinal dystrophy caused by biallelic mutations of the RPE65 gene, including Leber congenital amaurosis [38]. RPE65 is an enzyme in the cells of the retinal pigment epithelium that regenerates 11-cis-retinal, a molecule necessary for photoreceptors to detect light. Without functional RPE65, the photoreceptors gradually degenerate, which leads to progressive loss of vision and blindness.

Luxturna is administered by subretinal injection following vitrectomy and delivers a functional copy of the RPE65 gene directly into the cells of the retinal pigment epithelium. Phase III clinical trials showed that patients who received Luxturna achieve a statistically and clinically significant improvement in the ability to orient themselves in space at various levels of illumination, in light sensitivity, and in the visual field compared with the control group, whereby these effects were lasting over a four-year follow-up period [38,81]. Luxturna marked a turning point in the field of gene therapy as the first in vivo AAV-based therapy approved in the USA [82].

6.4.2 Idiopathic macular telangiectasia type 2

Encelto (revakinagene taroretcel), approved by the FDA in March 2025, represents a particularly interesting, innovative approach to gene therapy that differs in several important respects from all the previously described drugs [83]. Idiopathic macular telangiectasia type 2 (MacTel type 2) is a progressive degenerative disease of the macula that arises above all owing to the loss of Müller glial cells in the central zone of the retina, which secondarily leads to the degeneration of photoreceptors and loss of central vision. The disease affects adults, and until the appearance of Encelto there was no approved treatment.

The way in which this drug acts is unique; instead of delivering a gene directly into the patient's cells, Encelto uses an encapsulated cell therapy approach. An implant of only about 6 millimetres in size is surgically placed into the vitreous body of the eye and contains allogeneic retinal pigment epithelium cells that have been genetically modified to continuously produce recombinant human ciliary neurotrophic factor (rhCNTF). These cells are housed within a semi-permeable polymer capsule that has two functions: on the one hand, it allows CNTF to pass into the vitreous body of the eye, and on the other, it protects the cells from the host's immune system [84].

CNTF is a protein that supports the survival, differentiation, and function of neurons, including photoreceptors. It is thought that the rhCNTF released from the implant initially acts on the Müller glial cells, inducing a cascade of events that may help the survival of photoreceptors, although the exact mechanism of action has not yet been fully clarified [85].

The encapsulated cell therapy approach that Encelto uses opens up interesting possibilities for the future of gene therapy. By using allogeneic rather than autologous cells, and by protecting those cells from the immune system with a capsule, the need for individualised preparation of the drug for each patient is eliminated — which is one of the greatest practical challenges of ex vivo therapies based on the patient's own cells. In addition, the implant can be surgically removed should adverse effects arise.

6.5 Oncological applications

Tumours arise as a consequence of accumulated genetic changes that disrupt the control of cell growth. By acting on these changes, gene therapy can be directed straight at the molecular cause of the disease. Moreover, it is not necessary for oncological therapies based on gene therapy to permanently alter the genome; it is enough for them to achieve a therapeutic effect that lasts long enough for the immune system or other mechanisms to eliminate the tumour. The greatest contribution of gene therapy to oncology, however, does not come from direct intervention on tumour cells, but from the genetic modification of the patient's immune system so that it itself recognises and destroys the tumour.

The first oncological therapy based on a genetically modified virus was Imlygic (talimogene laherparepvec), approved in 2015 for the treatment of unresectable melanoma [86]. It is a genetically altered herpes simplex virus that selectively replicates in tumour cells and leads to their lysis, with the additional expression of GM-CSF, which enhances the local immune response. Imlygic is formally classified as an oncolytic viral therapy, not classical gene therapy, but it is historically significant as the first approved drug of this kind.

6.5.1 CAR-T cell therapy

Chimeric antigen receptor T-cell (CAR-T) therapy represents the most significant contribution of gene therapy to oncology. The basic idea behind this approach is the modification of the patient's T lymphocytes so that they recognise tumour cells and become activated precisely upon contact with them. T lymphocytes are isolated from the patient's blood and then transduced under laboratory conditions with a viral vector (most often lentiviral) carrying the gene for the chimeric antigen receptor (CAR). The CAR is a synthetic receptor that joins an antigen-binding domain — usually a derivative of an antibody specific for a tumour-associated antigen — with intracellular signalling domains that activate the T cell as soon as contact is made with the target antigen. It is important to emphasise that the CAR does not exist in nature; rather, it is an entirely synthetic molecule constructed by joining elements that are not physiologically functionally connected [87]. The modified cells are multiplied ex vivo to a therapeutic number and injected into the patient, where they selectively destroy the tumour cells that express the target antigen [88,89].

Kymriah (tisagenlecleucel) was the first approved CAR-T drug, intended for the treatment of paediatric B-cell acute lymphoblastic leukaemia resistant to standard therapy [90]. Kymriah uses a lentiviral vector to introduce a CAR construct directed against the CD19 antigen, which is located on the surface of B lymphocytes and B-cell malignancies. The approval of Kymriah opened the way for several similar anti-CD19 CAR-T therapies for various types of B-cell malignancies, which differ from one another in the choice of vector, the design of the signalling domains, and the approved indications, but share the same basic therapeutic principle [91,92].

A significant step forward is represented by CAR-T therapies directed at the BCMA antigen (B-cell maturation antigen), which is expressed on the surface of plasma cells and the malignant cells of multiple myeloma. Carvykti (ciltacabtagene autoleucel) and Abecma (idecabtagene vicleucel) are drugs that have shown exceptional results in patients with refractory multiple myeloma [93,94]. In the pivotal CARTITUDE-1 study, Carvykti achieved an objective therapeutic response in 97% of patients, with complete remission in almost 80% [93]. These results are particularly significant, since until a few years ago multiple myeloma was considered an incurable disease.

Tecelra (afamitresgene autoleucel) is a drug whose approach differs from the classical CAR-T approach — it involves genetically modified T lymphocytes that express a T-cell receptor (TCR) specific for the MAGE-A4 antigen [95]. Unlike the CAR construct, which recognises antigens on the cell surface, the TCR recognises peptide fragments presented in the context of MHC molecules, which makes it possible to target intracellular antigens. MAGE-A4 belongs to the group of cancer-testis antigens — it is expressed in synovial sarcoma and a number of other tumours, but in healthy tissues of the adult organism only in the testes, which makes it a selective target. Tecelra thereby became the first approved TCR-T drug for the treatment of a solid tumour and opened up the possibility of applying this approach beyond haematological malignancies.

6.5.2 Other oncological applications

In addition to cellular therapies such as CAR-T, in which the patient's immune system is modified ex vivo, in vivo gene therapies have also been developed that deliver the therapeutic gene directly into the tumour tissue. Adstiladrin (nadofaragene firadenovec) uses a non-replicating adenoviral vector for the intravesical delivery of the gene for interferon-alfa 2b into the urothelium of the bladder in patients with BCG-unresponsive non-muscle-invasive bladder cancer. The local expression of interferon-alfa enhances the antitumour effect directly in the bladder tissue, with minimal systemic exposure and without the significant adverse effects characteristic of systemic immunotherapy [96].

6.6 Dermatological applications

6.6.1 Dystrophic epidermolysis bullosa

Dystrophic epidermolysis bullosa (DEB) is a hereditary skin disease caused by mutations in the COL7A1 gene, which encodes the structural protein responsible for connecting the epidermal and dermal layers of the skin — collagen type VII. In the absence of functional collagen VII, the layers of the skin separate easily, which leads to a pronounced fragility of the skin that manifests itself in the appearance of extremely painful blisters and wounds that heal slowly. Conventional treatment is exclusively supportive and comprises wound care and pain control.

Vyjuvek (beremagene geperpavec) is a topical gene drug approved for the treatment of wounds arising as a consequence of this disease [97]. It is a genetically modified herpes simplex virus type 1 carrying the COL7A1 gene. The drug is applied directly to the wounds in the form of a gel, whereby the therapeutic gene is introduced into the skin cells, which then temporarily express collagen VII and enable the healing of wounds. Since the HSV-1 genetic material remains in episomal form and does not integrate into the genome, the therapeutic effect is temporary and requires continuous application for each new wound, which is at once both a limitation and a safety advantage of this approach.

Zevaskyn (prademagene zamikeracel) uses a different approach for the same genetic defect [98]. It is an ex vivo cell therapy in which keratinocytes are isolated from the patient's skin and then transduced with a retroviral vector carrying a functional copy of the COL7A1 gene. The modified cells are multiplied and formed into cell sheets that are surgically applied to the patient's chronic wounds [99]. Unlike Vyjuvek, Zevaskyn is administered once, and the duration of the therapeutic effect is measured in months to years, thanks to the fact that the retroviral vector integrates stably into the genome of the keratinocytes, whereby the corrected gene is passed on to the cells arising from their division as well.

6.7 Personalised gene therapy

In February 2025, a newborn with a rare deficiency of carbamoyl phosphate synthetase 1 (CPS1) became the first patient to receive a personalised CRISPR therapy developed specifically for its mutation [32]. CPS1 deficiency is a hereditary metabolic disorder that impairs the liver's ability to metabolise proteins, leading to an accumulation of ammonia in the blood that can cause serious brain damage. About half of newborns with this diagnosis die in the first months of life, and until then the only lasting therapeutic option was liver transplantation. The patient, KJ Muldoon, received a therapy based on base editing delivered by lipid nanoparticles to the hepatocytes, with the aim of correcting the specific point mutation in the CPS1 gene. The therapy was developed in just six months from diagnosis, which is an unprecedented speed in the history of gene therapies; the complete cycle from identification of the mutation, design of the guide RNA, preclinical testing, and regulatory authorisation usually takes about four years [100]. After three consecutive infusions the patient began to tolerate larger quantities of protein in the diet and continued to develop normally. The case of KJ Muldoon is the first example of an in vivo CRISPR therapy developed for a single individual, whereby it also opened questions that go beyond medicine itself — such a model of treatment is difficult to replicate at a systemic level under the existing regulatory framework, which relies on clinical trials with a large number of participants and high approval costs. These limitations motivated proposals for new regulatory pathways, which are described in the chapter on regulation.

7 Safety and adverse effects

Although gene therapy has over the last decade achieved significant clinical progress and made possible even the cure of diseases that until recently were considered lethal, each of the therapeutic approaches carries with it certain risks. An understanding of these risks is necessary for the safe application of existing therapies and for the further development of the field.

7.1 Immune response to viral vectors

Although the viruses used for therapeutic purposes are to a significant extent genetically modified, they remain foreign particles for the host's immune system. The immune response can considerably reduce the efficacy of the therapy and, in certain situations, endanger the patient's safety. The character and intensity of this response differ significantly between individual types of vector.

The immunogenicity of adenoviral vectors and the consequences of their use have already been discussed in the chapter on viral vectors. Here the specificities of the immune response to AAV will be considered more broadly, since these form the basis of the majority of the therapies approved today.

With AAV vectors, immunogenicity manifests itself in two forms. A large percentage of the general population carries neutralising antibodies against particular AAV serotypes, arising as a consequence of contact with the wild-type virus in childhood [40,101]. In addition, after the administration of the vector, a secondary immune response develops, which involves both a humoral and a cellular component, and which makes repeat administration of the same therapy impossible. This poses a problem in diseases where the therapeutic effect wanes over time, and cross-reactivity between serotypes additionally limits alternative strategies [101].

A challenge is posed by the systemic administration of high doses of AAV vectors. It can lead to thrombotic microangiopathy — which presents with acute kidney injury, haemolytic anaemia, and thrombocytopenia [102]; hepatotoxicity, which manifests itself as an elevation of liver enzymes [103,104]; and damage to the dorsal root ganglia of the spinal cord [105].

A specific form of toxicity associated with AAV vectors that is not primarily mediated by an immune response is precisely damage to the dorsal root ganglia of the spinal cord, which contain the bodies of sensory neurons. AAV vectors show a pronounced tendency to transduce precisely these neurons, which leads to excessive expression of the therapeutic gene, cellular stress, and gradual degeneration. In most patients these effects are subclinical and are detected by electrophysiological tests or magnetic resonance imaging. Strategies for reducing the risk include lowering the vector dose and incorporating binding sites for microRNAs specific to sensory neurons into the vector genome, whereby the expression of the therapeutic gene is selectively silenced precisely in those cells, with expression preserved in the target tissues [105].

Contemporary protocols include various strategies to reduce the risk of a harmful immune response. The approaches involve careful selection of patients without pre-existing antibodies and the prophylactic administration of corticosteroids or other immunosuppressive regimens [101].

7.2 Immune response to the therapeutic protein

In addition to the immune reaction to the vector itself, a significant safety challenge can also be posed by the immune response to the therapeutic protein itself. This problem is particularly pronounced in patients with so-called null mutations, in whom the target gene is completely non-functional, so the immune system has never been exposed to the protein that gene encodes. In such cases the immune system recognises the therapeutic protein as a foreign antigen and develops a specific response against it, which can neutralise the therapeutic effect. The phenomenon was first described in detail in clinical trials of gene therapy for haemophilia, where the administration of an AAV vector expressing factor IX led to a cell-mediated immune response against the transduced hepatocytes [104]. Similar observations were also recorded in clinical studies for Duchenne muscular dystrophy, where the immune response against microdystrophin led to the appearance of myositis in several patients [106]. As a result of these events, the criteria for including patients in clinical trials for this drug were changed. Patients whose mutations affect parts of dystrophin present in the therapeutic gene are now excluded from studies, owing to the elevated risk that their immune system will recognise microdystrophin as a foreign antigen [25,106].

7.3 Safety profile of CAR-T therapies

The most common serious adverse effect of this therapy is cytokine release syndrome (CRS). CRS is an acute systemic inflammatory reaction characterised by an elevated body temperature, a drop in blood pressure, a decrease in oxygen saturation, and potential organ dysfunction. It arises when the infused CAR-T cells, activated by contact with the tumour antigen, release cytokines that then activate macrophages and other cells of the immune system. Consequently, large quantities of pro-inflammatory cytokines are released, above all IL-6 (which is primarily of macrophage origin), IFN-γ, and TNF-α [107]. CRS has been recorded in the majority of patients treated with CAR-T therapy, but in most cases it is of a mild or moderate degree and is successfully treated with tocilizumab, a monoclonal antibody directed against the IL-6 receptor, and with corticosteroids [107,108].

In addition to CRS, immune effector cell-associated neurotoxicity syndrome (ICANS) is also clinically significant; it manifests itself with confusion, aphasia, encephalopathy, and, in more severe cases, convulsions. The mechanism by which ICANS arises has not yet been fully clarified, but it is thought to involve the activation of the endothelium of the central nervous system and a transient disruption of the blood–brain barrier [108].

A specific safety challenge associated with CAR-T therapies is the immunogenicity of the CAR construct itself. In the majority of CAR-T therapies approved to date, the antigen-binding part is derived from murine antibodies, which in a portion of patients leads to the development of an anti-CAR humoral (antibody) and cellular (cytotoxic T lymphocyte) immune response. Although pre-existing antibodies do not significantly affect the initial response to the therapy, the cellular response that develops after infusion can directly eliminate the CAR-T cells and is associated with a weaker response to repeated doses [109]. For this reason, CAR-T therapies with humanised or fully human antigen-binding parts are increasingly being developed.

7.4 Insertional mutagenesis

Vectors that integrate into the host genome carry a risk of insertional mutagenesis, whose historical significance and mechanism have already been mentioned in the chapter on lentiviral vectors. The most dramatic case was recorded in clinical trials of gene therapy for X-SCID, where a gamma-retroviral vector integrated near the LMO2 oncogene and caused the development of leukaemia in 6 of 20 treated children [29,110,111].

With contemporary lentiviral vectors, thanks to the SIN design and the use of weaker internal promoters, the risk of clinically significant genotoxicity has been considerably reduced, and analyses of integration sites do not show preferential integration near oncogenes [17]. Nevertheless, the field is not entirely safe — several cases of myelodysplastic syndrome have been recorded in patients treated with Skysona for cerebral adrenoleukodystrophy [111,112]. These findings led to an additional tightening of clinical practice and post-marketing monitoring.

AAV vectors, although they predominantly remain in episomal form, can in some cases integrate into the genome. Studies on animal models have shown that AAV integration can be associated with the development of hepatocellular carcinoma, particularly in the presence of predisposing factors such as young age or pre-existing liver disease [113,114]. To date no clear causal link has been established between AAV therapy and malignancy in humans, but this risk is carefully monitored through long-term post-marketing studies.

7.5 Off-target effects in gene editing

Gene editing technologies carry risks that differ from those of classical gene therapy with viral vectors. The Cas9 enzyme, although directed at a specific sequence by means of the guide RNA, can induce double-strand breaks at other sites in the genome as well that contain similar sequences — so-called off-target effects [115]. These unwanted breaks can be repaired by the non-homologous end joining mechanism, which often leads to indel mutations that can disrupt the function of other genes. In the worst cases, off-target effects can also lead to chromosomal rearrangements and the loss of larger segments of a chromosome [47].

In addition to off-target effects, unwanted changes can also arise at the target site itself. Base editing and prime editing were developed precisely as an attempt to overcome these limitations, but these systems too have specific shortcomings — base editing can cause so-called bystander edits, that is, unnecessary changes to neighbouring bases in the target region, while the efficiency of prime editing is lower relative to the classical CRISPR–Cas9 system [53].

Additionally, Cas9-induced double-strand breaks activate a p53-mediated DNA damage response that in healthy cells leads to apoptosis or arrest of division. The consequence of this is that within a population of successfully edited cells, those with a weakened p53 pathway survive disproportionately more often — and since p53 mutations are among the most common oncogenic events, this poses a potential risk [116,117].

In combination with viral vectors, CRISPR carries an additional risk — when an AAV vector carrying the CRISPR components integrates at the site of a double-strand break caused by Cas9, there is a significantly elevated rate of integration of viral DNA into the genome, which can have long-term genotoxic consequences [118].

7.6 Duration of the therapeutic effect and the need for repeat administration

Although gene therapies are often presented as therapies administered once, the available clinical data show that this approach may not be sustainable for all indications. With AAV-based therapies, which retain the genetic material in episomal form, the therapeutic effect can wane over time owing to the division of the target cells. An illustrative example is Roctavian for haemophilia A: in the pivotal clinical trial, factor VIII activity was significantly elevated at the end of the first year, but in the following years it gradually decreased [74,119]. This decline in expression is specific to Roctavian and differs from the more stable expression recorded with other AAV therapies, which indicates that the duration of the effect is not a universal property of all AAV approaches but depends on the specific vector and target tissue [120]. The repeat administration of AAV vectors is additionally prevented by the development of neutralising antibodies after the first dose, even when a different serotype is used, owing to frequent cross-reactions between serotypes. The development of strategies that would enable safe repeat administration is one of the most important directions of further research.

8 Challenges in the application of gene therapy

Although gene therapy is increasingly applied in clinical practice, its application opens up a series of challenges that do not concern the therapeutic principle itself, but the context in which these drugs are developed, approved, financed, and made available to patients. These challenges are of a regulatory, economic, and ethical nature and to a considerable extent determine how far contemporary achievements in this field will actually reach the patients for whom they are intended.

8.1 Regulatory framework and definitions

The precise technical definition of the European Medicines Agency (EMA) states that a gene therapy is a biological medicinal product that meets two criteria. The first is that it contains an active substance that contains or consists entirely of a recombinant nucleic acid of biological origin that is administered to humans with the aim of regulating, repairing, replacing, adding, or deleting a genetic sequence. The second is that its therapeutic effect is directly caused by the recombinant sequence it contains or is the product of the gene expression of that sequence [3]. In the European regulatory framework, gene therapy belongs to the broader category of advanced therapy medicinal products (ATMP), together with cell therapies and tissue engineering.

This definition has several shortcomings. Namely, it does not encompass forms of gene editing in which the administered medicinal product does not contain recombinant nucleic acids, but rather only molecules that bring about a change within the genetic material of cells or laboratory-synthesised short nucleic acids [2]. It also excludes antisense oligonucleotides and siRNA, despite the fact that they directly affect gene expression. Although gene editing conceptually belongs to gene therapy, it is not explicitly encompassed by this official definition. There is a gap between scientific achievements in this field and the regulatory categorisation, which continues to widen with the approval of newer technologies such as CRISPR–Cas9, base editing, and prime editing [121].

This need has been partly met in the most recent guideline for the conduct of clinical trials of advanced therapy medicinal products of the European Medicines Agency [122], where a place is explicitly given to gene editing as well. It is indicated that in in vivo editing, the molecules that carry out only the editing are considered the active substance — which encompasses not only RNA and DNA molecules, but also proteins, ribonucleoproteins, and other molecules necessary for this process. In ex vivo editing, the active substance is the prepared genetically modified cells, while the molecules used for their modification are treated as starting material.

The U.S. Food and Drug Administration (FDA) takes a broader and conceptually more flexible stance. According to its definition, gene therapy encompasses all therapeutic products that achieve their biological effect through the transcription or translation of introduced genetic material, or through the specific alteration of the host's genetic sequences. Examples of products classified as gene therapy include nucleic acids, genetically modified microorganisms, and artificially constructed nucleases with a specific site of action, which are used for gene editing [123]. In this way, gene editing is explicitly considered a type of gene therapy in the broader sense, which represents a notable difference relative to the European approach, which was historically more narrowly defined around recombinant nucleic acids.

In addition to the different scope of the definition, the European and American regulatory systems also differ in the subcategorisation of gene therapies. In the European Union, gene therapies make up one of four groups of ATMPs (alongside somatic cell therapies, tissue engineering, and combined advanced therapies), whereas the FDA recognises only two broader categories — gene therapy and cell therapy [2]. Various countries outside Europe and the USA also apply different approaches, whereby certain regulatory frameworks have permitted the approval of particular gene therapies significantly earlier than in Western countries — the best-known example being Gendicine, which was approved in China as far back as 2003.

Differences in regulatory philosophy directly affect the practical development of drugs. Manufacturers often have to adapt clinical programmes to different regulatory agencies in parallel, which increases costs and prolongs the time it takes for a therapy to reach patients. There are active efforts towards international harmonisation through initiatives such as the International Council for Harmonisation [124], but the specificity of gene therapies, particularly those based on gene editing, still leaves room for differences that are bridged only slowly.

In addition to international harmonisation, regulatory frameworks are also gradually adapting to the specificities of personalised gene therapy, where each patient is the carrier of a unique mutation. In February 2026, the FDA proposed the introduction of a new regulatory pathway called the plausible mechanism pathway, which makes it possible for several patients with different mutations but a common clinical manifestation (such as disorders of the same metabolic pathway or severe combined immunodeficiencies) to be included in a single clinical trial [100,125]. Under conventional rules, every guide RNA unique to a patient would be treated as a new drug requiring a separate set of preclinical and clinical trials. That process takes an average of four years and costs over 25 million US dollars, which for a newborn with a serious genetic disease is both unacceptably long and economically unsustainable. The proposed approach relies on the fact that only the first CRISPR therapy in a trial requires the complete series of tests, while subsequent ones, with minimal changes in the sequence of the guide RNA, require only supplementary testing. Estimates by the authors of this approach indicate that the time to clinical application could be shortened to just three months, with a cost per patient lower than 250,000 dollars. Similar changes are also announced by the British MHRA, which suggests that regulatory flexibility in the field of personalised gene therapy is becoming a global trend [100].

8.2 The cost of and access to gene therapy

Gene therapies are among the most expensive drugs in the history of medicine. The cost of most approved single-administration therapies exceeds two million dollars per patient, while certain drugs exceed even three million [126]. Hemgenix for haemophilia B is currently the most expensive approved drug in the world, with a price of approximately 3.5 million dollars per dose [127]. Casgevy for sickle cell anaemia and beta-thalassaemia costs about 2.2 million dollars, Lyfgenia about 3.1 million, Zolgensma for spinal muscular atrophy 2.1 million, while the prices of Roctavian for haemophilia A and Skysona for cerebral adrenoleukodystrophy also fall within the same range [126].

At first glance, such prices seem disproportionate, but there are several factors that explain them. First, the development of gene therapies is a lengthy and extremely expensive process that requires a decades-long research-and-development phase, complex preclinical and clinical studies, as well as a specialised production infrastructure, given that viral vectors are difficult to produce in large quantities while maintaining consistent quality and purity [126]. Second, the target patient populations are often small, since these are rare genetic diseases, which means that the development costs must be covered by a relatively small number of users. Third, the therapies are administered once. The manufacturers' argument meant to justify the high price is that the price should reflect the total value achieved by avoiding lifelong conventional treatment, which can also amount to millions of dollars per patient [128].

From the patient's perspective, this argument does not solve the practical problem of availability. Even in countries with developed health insurance systems, access to these therapies is often slowed by complex approval procedures and by the sum that has to be co-paid. In low- and middle-income countries, where the great majority of patients with some of the diseases for which these drugs are approved are located, such as sickle cell anaemia, contemporary gene therapies are practically unavailable [126,127]. This leads to a paradoxical situation in which drugs that represent the pinnacle of contemporary medicine are available above all in those systems in which conventional therapies are already relatively well developed. A serious question arises as to whether revolutionary scientific achievements truly fulfil their purpose if they are available only to a minority of the world's population.

Various strategies are being developed to improve the availability of these therapies. Outcome-based pricing models entail that the full price of the drug is paid only if the therapy achieves a predefined clinical effect, while in the event of failure a partial or full refund is made. Payment in instalments makes it possible for the costs to be distributed over several years, thereby reducing the financial impact on health systems and insurance companies. Special funds for rare diseases, as well as international solidarity mechanisms, represent additional approaches by which an attempt is made to ensure access to therapies in countries that cannot finance them themselves. In addition, there are ever-louder calls for transparency in the formation of prices and for a public debate on the extent to which high prices reflect the actual costs of development, and to what extent the profit margin of the industry — particularly bearing in mind that a significant part of the basic research that made these therapies possible was financed by public funds [128].

8.3 Ethical questions

The clinical application of gene therapy and gene editing opens up a series of ethical questions that go beyond the classical dilemmas of medical ethics. The most significant among them concerns the editing of germline cells, that is, interventions on embryos or germ cells. Such interventions represent a fundamentally different approach from somatic editing, since they can permanently alter the genetic make-up of the human species.

The scientific community, through international forums, has articulated the position that the clinical application of germline editing is not acceptable until the questions of safety, efficacy, and social consensus are resolved [129]. Although these documents did not have legal force, they represented a clear position of the scientific community — a position that in 2018 the Chinese scientist He Jiankui directly violated with his experiment [130]. Using CRISPR–Cas9 technology, he carried out the editing of the CCR5 gene on embryos with the intention of providing them with resistance to HIV, and this led to the birth of two girls with a modified genome. The case provoked almost universal condemnation from the scientific community owing to the insufficiently justified medical reasons for the intervention, inadequate informed consent, ethical failings in the design of the study, and the unknown long-term consequences of the changes carried out [131]. He Jiankui was dismissed from his university and excluded from the international scientific community, and the case accelerated international efforts towards clearer regulation of editing [132]. In most countries of the world, the clinical application of germline editing is today explicitly prohibited, while research on embryos that will not be implanted is permitted under strictly defined conditions [130].

This case illustrates the broader question of the limits of the application of powerful technologies that are technically available before society has reached a consensus on their acceptability. In addition to germline editing, ethical questions also arise in the context of somatic therapies — informed consent in the case of single, irreversible interventions, and decision-making in the paediatric population, where parents decide on behalf of children who are not themselves able to give consent, and who will live their whole lives with the outcome of the therapy [133].

There are also broader questions about the boundary between "treatment" and "enhancement" of human traits. Although the currently approved therapies are directed exclusively at the treatment of serious genetic diseases, the boundary towards potential application for non-therapeutic purposes is not always clear and remains the subject of ongoing public and scientific debate.

9 Concluding remarks

Over the past decades, gene therapy has travelled the road from an experimental concept to a clinically validated therapeutic approach. What in the 1990s were ambitious trials that more often yielded lessons than results has led to more than several dozen gene therapies being approved in the USA and the European Union for diseases that until recently were considered incurable — from inherited retinal dystrophies and spinal muscular atrophy, through haemoglobinopathies and haemophilia, to various haematological malignancies. The development of the CRISPR–Cas9 system and its derivatives, base and prime editing, has extended the therapeutic scope to precise interventions at the level of individual nucleotides, and the case of KJ Muldoon showed that a personalised therapy developed for a single individual is possible and feasible.

Alongside this progress, the field is actively working to overcome the limitations observed during clinical application and trials. The development of humanised CAR constructs is attempting to solve the problem of the immunogenicity of CAR-T therapies; the SIN design of lentiviral vectors has drastically reduced the risk of insertional mutagenesis; new AAV serotypes and strategies for suppressing anti-AAV immunity open up the possibility of repeat administration in patients in whom the therapeutic effect has waned; the in vivo delivery of gene editors by lipid nanoparticles enables corrections without the need for ex vivo manipulation of cells. Regulatory frameworks are gradually adapting to the specificities of this field, with initiatives such as the FDA's plausible mechanism pathway opening the way for the faster application of personalised therapies.

In the years to come, an even greater diversity of therapeutic approaches can be expected, along with the expansion of clinical application beyond rare monogenic diseases towards more complex polygenic disorders, autoimmune diseases, and potentially even some aspects of ageing. The greatest remaining question will not be of a technical nature, but rather a question of access — how to make a technology that is becoming ever more technically sophisticated available to patients outside the wealthier health systems. Gene therapy today represents a clinical reality, but its transformation into a systemically accessible clinical practice represents the next phase in the development of the field.