Key Points
- •
Gene therapy for fetuses may provide a therapeutic advantage over postnatal treatment for certain congenital diseases.
- •
Three types of fetal gene therapy are available: direct ‘somatic’ fetal gene therapy, in utero transplantation of gene corrected fetal stem cells and maternal gene therapy
- •
In preclinical animal models of congenital disease, direct fetal gene therapy can bring a phenotypic cure.
- •
Minimally invasive ultrasound-guided injection techniques can be used to target gene therapy to fetal organs in animal models.
- •
Observed risks of fetal gene therapy include insertional mutagenesis, germline gene transfer, vector toxicity, the fetal and maternal immune response to the vector, and maternal and fetal morbidity and mortality.
- •
Currently, direct fetal gene therapy and in utero transplantation of gene-corrected fetal stem cells remains at the experimental stage.
- •
Maternal gene therapy is close to being translated into the clinic.
Introduction
Gene therapy has now arrived as a major step forward in the treatment of select severe genetic diseases. Already lentiviral vectors have been used to cure severe combined immune deficiency, and trials of adeno-associated viral vectors show efficacy to treat haemophilia B (factor IX deficiency). Many of the original barriers to effective postnatal gene therapy have been overcome. These include difficulty targeting the appropriate organ, a robust immune response to the therapy in adults and low-level expression of the therapeutic protein. In some diseases, gene therapy may be most effective when it is given early in neonatal life. When congenital disease pathology occurs during fetal development, then treating the fetus may be the best solution.
Preclinical studies in animal models in the last 18 years have shown that fetal gene therapy can cure severe genetic disease. More recently, structural anomalies in fetuses have been prevented using a gene therapy approach. For some nongenetic conditions, timely expression of a particular protein, for example, during the last third of pregnancy, may alleviate pathology. The arrival of noninvasive prenatal diagnosis using circulating fetal DNA in the maternal blood allows clinicians to detect a congenital disease in fetuses as early as 10 weeks of gestation. This gives couples time to consider their options, including a possible in utero treatment to correct the genetic disorder.
What Is Gene Therapy?
Gene therapy uses the delivery of genetic material, generally DNA, into cells to generate a therapeutic effect by correcting an existing abnormality or providing cells with a new function. Vectors are the vehicles that are used to carry the genetic material into the cell. The two major classes of gene therapy methods use either recombinant ‘genetically altered’ viruses (viral vectors) or naked DNA or DNA complexes (nonviral methods). The genetic material that is transferred by this type of genetic engineering is called a transgene, and the introduction of the transgene has the potential to change the phenotype of a patient. The transgene contains (i) a promoter, which is a regulatory sequence that will determine where and when the transgene is active (e.g., a liver promoter to switch on the transgene only when it is in a liver cell); (ii) an exon, a protein coding sequence (usually derived from the cDNA for the protein of interest, e.g., human factor IX cDNA for treatment of haemophilia B); and (iii) a stop sequence to turn off the transgene. In the cell, the transgene is expressed producing a transgenic protein that has the therapeutic effect.
Somatic gene therapy treats an individual patient by insertion of genes into non germline cells that are either outside the body ( ex vivo ) (e.g., haematopoietic stem cells grown in culture) or in vivo (e.g., intravascular injection); pluripotential stem cells or differentiated cells may be targeted. Fetal application of gene therapy can be directly to the fetus or to fetal stem cells for autologous transplantation ( Fig. 46.1 ). Gene therapy can even be given to the mother when maternal pathology affects the fetus in utero (e.g., in the case of uteroplacental insufficiency). Germline gene therapy would target oocytes or spermatocytes and potentially might eradicate inherited diseases in future generations. From the earliest days of gene therapy, however, correcting the germline was considered to be neither scientifically nor medically justifiable, technically unsafe and unpredictable and therefore ethically unacceptable.
The Potential Advantages and Disadvantages of Fetal Gene Therapy
The advantages of fetal gene therapy over postnatal treatment are summarised in Table 46.1 . Production of clinical grade vector is time consuming and expensive, and the small size of the fetus could lead to increased vector biodistribution at the same vector dose as an adult. Organs that are difficult to target after birth may be more easily accessible during fetal life because of their developmental stages or relative immaturity.
| Advantage of Fetal Gene Therapy | Disease Example or Effect |
|---|---|
| Correct disease in the fetus when pathology begins before birth | Congenital MPS Fetal application may prevent the brain damage that occurs before birth |
| Targets a rapidly dividing population of stem cells that exist in the fetus | Provides a large population of transduced cells to produce a better therapeutic effect |
| Higher vector-to-target cell ratio because the fetus is small | Less vector needed to achieve the same therapeutic effect so more cost effective |
| Target organs that are difficult to access after birth | Epidermolysis bullosa The fetal epidermis may be targeted in utero using gene therapy. It undergoes remodelling by programmed cell death to be replaced postnatally by mature keratinocytes that prevent gene transfer. |
| Induce tolerance to the transgene protein product | Better therapeutic effect by avoiding the development of an immune response that occurs in postnatal gene therapy |
| Disadvantage of Fetal Gene Therapy | Disease Example or Effect |
| Possible increased risk for genetic modification of the germline | Gene transfer to male sheep sperm seen after first trimester injection but not second or third trimester exposure |
| Possible increased risk for insertional mutagenesis related to particular vector types (EIAV lentivirus) | Hepatocellular carcinoma in adult mice after fetal gene transfer |
| Vector toxicity | Fetal ascites and death seen after injection of VSVG-pseudotyped lentivirus into fetal sheep |
| Maternal immune response to vector may prevent gene transfer to fetus | Preexisting maternal antibodies to AAV5 serotype prevented gene transfer to fetal macaques |
| Maternal and/or fetal morbidity and mortality | Intrauterine procedures could cause fetal loss and maternal and fetal morbidity |
A major obstacle to postnatal gene therapy has been the development of an immune response against the transgenic (therapeutic) protein or the vector itself, particularly when gene therapy is aiming to correct a genetic disease in which complete absence of a gene product is observed. Some individuals have preexisting antibodies to the viral vector that will prevent long-term expression of the transgenic protein, limiting therapeutic efficacy and preventing repeated vector administration. For example, preexisting neutralising antibodies against adeno-associated virus (AAV) serotype 2 have been shown to interfere with AAV2 vector-mediated factor IX (FIX) gene transfer to the liver. Delivering foreign protein to the fetus can take advantage of immune tolerance which is induced during fetal life, a concept that was first proposed nearly 60 years ago. Induction of tolerance depends on the foreign protein being expressed sufficiently early in gestation, probably by 12 to 16 weeks, before the immune system is fully developed and expression maintained at a detectable level within the fetus for presentation to the thymus at the correct time. For human gestation, transgenic protein expression will need to last at least 6 months if the vector is given early in pregnancy, which limits the types of viral vectors that can be applied. Proof-of-principle studies have shown long-term expression of proteins at therapeutic levels and induction of immune tolerance in both small and large animals and cured congenital disease in some animal models.
Vectors for Fetal Gene Transfer
An ideal vector for prenatal gene therapy is one that can produce long-term regulated and therapeutic expression of the transferred gene through the use of a single and efficient gene delivery method and is safe to the mother and fetus, thus allowing incorporation into clinical practice. For example, a vector carrying the β-globin gene should deliver and express the gene only to erythroid specific cells and lineages. These and other essential characteristics are described in Table 46.2 .
| Characteristic | Reason |
|---|---|
| Highly efficient, regulated transgenic protein expression | Provide therapeutic levels of protein expression |
| Time and duration of transgenic protein expression to suit disease | Example (1): long-term transgenic protein expression for a monogenic disorder requires protein expression to last the lifetime of the individual (e.g., haemophilias) Example (2): transient transgenic protein expression for a developmental or obstetric disorder requires protein expression at a critical window of fetal growth (e.g., fetal growth restriction) Example (3): transient expression of a transgenic protein at a specific development time point to treat a structural defect (e.g., facial cleft) |
| Specific vector tropism to target organ | Avoid systemic gene transfer |
| Large carrying capacity of transgene insert | Accommodate therapeutic gene and any required regulatory elements |
| No toxicity | Safe for mother, fetus and future progeny |
| No immunogenicity | Avoid generating a fetal immune response |
| No mutagenic properties | Safe for fetus and future progeny |
| Replication incompetent | Inability to replicate itself after injection in vivo |
| Manufacturing process appropriate for GMP | Sufficient quantities of clinical grade vector available at a reasonable cost |
The most commonly tested vectors in fetal gene therapy preclinical studies have been adenovirus and AAV, lentivirus and retrovirus vectors. These and other less commonly used vector systems are described in Table 46.3 .
| Vector | DNA | Efficiency | Tropism | Advantages | Disadvantages | Prenatal Considerations |
|---|---|---|---|---|---|---|
| Nonviral DNA | No limit | + | Limited | Low toxicity Low immunogenicity | Low transduction efficiency | Expression may not last through gestation |
| Adenovirus | 7.5 kB | +++ | Depends on serotype | Can grow to high titre Highly efficient gene transfer Clinical safety and efficacy data long term in adults | Short-term expression & immunogenic | Case reports of an association with some fetal abnormalities |
| Helper-dependent adenovirus | 35 kB | +++ | Broad | Low immunogenicity, high capacity, long-term expression in quiescent cells | Inefficient production | |
| Adeno-associated virus | 4.7 kB generally | ++ | Depends on sub-type | Long-term expression Low immunogenicity Very high titre Can target central nervous system via systemic injection Clinical safety and efficacy data long term in adults | Liver toxicity in adult trials due to anti-capsid T cells. Risk for hepatocellular cancer. Androgens increase gene transfer (transduction level higher in males>females) | Some subtypes associated with miscarriage Low level integration into active genes so theoretical mutagenesis risk. |
| Retrovirus | 10 kB | + | Depends on pseudotyping | Long-term gene transfer Clinical safety and efficacy data long term in neonates | Potential for insertional mutagenesis. Infect dividing cells only. | Risk for germ-line transmission and insertional mutagenesis Virus inactivated by amniotic fluid |
| Lentivirus | 10 kB | ++ | Depends on pseudotyping | Long-term gene transfer Infects dividing and non-dividing cells | Potential for insertional mutagenesis | Risk for germ-line transmission and insertional mutagenesis |
| Nonintegrating lentivirus | 10 kB | ++ | Depends on pseudotyping | Insertional mutagenesis unlikely | Short term expression | Rapidly dividing fetal cells may result in long-term low transgenic protein expression |
Manipulating the vector structure and the transgene can alter vector properties. Pseudotyping, for example, involves changing the virus capsid (outer covering) for one of a different serotype or of a completely different virus, thus altering its ability to infect particular cell types or organs. Using alternative enhancer-promoters can improve gene transfer to specific organs or tissues. A promoter is a site on DNA to which RNA polymerase can bind and initiate transcription, and an enhancer is a regulatory sequence that can elevate levels of transcription from an adjacent promoter. These can be derived from the genomes of mammals, viruses or other organisms and can even be manipulated to allow regulatable gene expression if required.
Some vectors, such as lentivirus and retrovirus vectors, contain integrases, an enzyme that enables the virus genetic material to be integrated into the DNA of the infected cell. This ensures that daughter cells contain the viral genetic material after cell division. Other vectors such as adenovirus are nonintegrating. Many replication-deficient lentiviruses are based on the immunodeficiency virus, and there is the theoretical possibility of reversion to the wild type. In third- and fourth-generation lentivirus vectors, however, the risk for in vivo generation of replication competent viruses is reduced by removal of the tat gene. Modification of virus elements, such as mutating the integrase in lentiviral vectors, renders it incapable of integrating and greatly reduces the risk for insertional mutagenesis. Clinical grade production of vectors are tested rigorously for replication competent viruses. For more detailed information on vectors relevant to fetal gene therapy, the authors refer readers to other references.
The effect of fetal exposure on vectors is important to consider because many routes of fetal application require delivery into fluid compartments such as the serum, airways or amniotic fluid (AF). Human serum can inactivate retroviruses and AF inhibits retrovirus infection. Altering vector production can make them more robust. Lentivirus, AAV and adenovirus vectors are relatively immune to damage.
Gene Editing
Gene editing is a process of insertion, deletion or replacement of DNA at a specific site in the genome of a cell which is achieved in the laboratory using engineered nucleases also known as molecular scissors. It is an attractive alternative approach to correct gene defects because it avoids the introduction into the cell or genome of the extra DNA or RNA components that viral or nonviral vector approaches require. There are a number of different possible strategies. Zinc finger nucleases (ZFNs) are artificial restriction enzymes engineered to target desired DNA sequences within complex genomes. Transcription activator–like effector nucleases (TALENs) use DNA-recognition modules that recognise single base pairs. Reports of successful applications to genomic targets are appearing at an accelerating rate. RNA-guided engineered nucleases (RGENs) derived from the bacterial clustered regularly interspaced short palindromic repeat (CRISPR)-Cas (CRISPR-associated) system are now available. CRISPR/Cas-mediated genome editing has been successfully demonstrated in zebrafish and bacterial cells, giving high-precision genome editing. Gene editing could be particularly attractive to target specific stem cell groups in an in utero stem cell gene therapy (IUSCGT) approach.
Selecting the Right Disease for Fetal Gene Therapy
As with any new therapeutic modality, the risks of fetal gene therapy are not well characterised. Careful thought must be given to decide on the right disease to select for a first-in-woman trial. When vectors are given directly to the fetus for correction of genetic disease, there has been guidance given by the National Institutes for Health Recombinant DNA Advisory Committee ( Box 46.1 ).
The disease should be associated with serious morbidity and mortality risks for the fetus either in utero or postnatally.
There should be no effective postnatal therapy, or there is a poor outcome using available postnatal therapies.
The treatment should correct all serious abnormalities that are associated with the disease.
The disease must be definitively diagnosed in utero and have a well-defined genotype–phenotype relationship.
There should be an animal model that recapitulates the human disease or disorder for testing of in utero gene transfer.
Some of the diseases that may be suitable for fetal treatment are listed in Table 46.4 . Preclinical studies of direct fetal gene transfer are encouraging. Fetal application of gene therapy in mouse models of congenital disease such as haemophilia A and B, congenital blindness, Crigler-Najjar type 1 syndrome and Pompe disease (glycogen storage disease type II) have shown phenotypic correction of the condition. For structural anomalies, transient transduction of the periderm via intra-amniotic delivery of adenoviral vector encoding transforming growth factor (TGF) β3 prevents cleft palate in a mouse model of disease. For obstetric conditions that affect the fetus, maternal uterine artery injection of adenovirus containing the vascular endothelial growth factor (VEGF) gene improves fetal and lamb growth in growth-restricted sheep pregnancies.
| Disease | Therapeutic Gene Product | Target Cells/Organ | Age at Onset | Incidence | Life Expectancy |
|---|---|---|---|---|---|
| Cystic fibrosis | Cystic fibrosis transmembrane conductance regulator | Airway and intestinal epithelial cells | Third trimester of pregnancy | 1:4000 | Mid-30s |
| Duchenne muscular dystrophy | Dystrophin | Myocytes | 2 yr | 1:4500 | 25 yr |
| Spinal muscular atrophy | Survival motor neuron protein | Motor neurons | 6 mo (type 1) | 1:10,000 | 2 yr |
| Haemophilia | Human factor VIII or IX clotting factors | Hepatocytes | 1 yr | 1:6000 | Adulthood with treatment |
| β-Thalassaemia | Globin | Erythrocyte precursors | <1 yr | 1:2700 | <20 yr in developing countries |
| Lysosomal storage disease (e.g., neuronopathic Gaucher disease) | Glucocerebrosidase | Hepatocytes | 9.5 yr | 1:9000 overall 1:59,000 | <2 yr |
| Urea cycle defects (e.g., ornithine transcarbamylase deficiency) | Ornithine transcarbamylase | Hepatocytes | 2 days | 1:30,000 overall 1:105,000 | 2 days (severe neonatal onset) |
| Severe combined immunodeficiency | γ c Cytokine receptor (X-linked SCID) | Haematopoietic precursor cells | Birth | 1:1,000,000 | <6 mo if no bone marrow transplant |
| Epidermolysis bullosa (e.g., dystrophica) | Type VII collagen | Keratinocytes | Birth | 1:40,000 | Adulthood |
| Severe fetal growth restriction | Vascular endothelial growth factor | Trophoblast | Fetus | 1:500 | Days |
The ideal fetal gene therapy would be able to effectively treat a serious congenital disease with a single direct fetal vector injection, providing sufficient transgenic protein expression from one injection and maintenance of this therapeutic expression for the rest of the individual’s life. Progress in the treatment of a group of inherited conditions, the lysosomal storage disorders, is discussed in detail here to illustrate recent progress.
The Lysosomal Storage Diseases as Candidate Diseases
The lysosomal storage diseases are inherited deficiencies of lysosomal enzymes that lead to intracellular substrate accumulation. In mucopolysaccharidosis (MPS) type VII, for example, a deficiency of β-glucuronidase activity leads to accumulation of glycosaminoglycans in lysosomes, resulting in an enlarged liver and spleen, restricted growth, developmental delay and death from cardiac failure. The disease process starts in utero. Although rare, MPS type VII has been a disease of choice to investigate gene therapy because of the availability of a mouse and dog model.
Correction of the MPS phenotype theoretically requires only low levels of the therapeutic gene product. Neonatal injection of a retrovirus vector in MPS VII dogs and mice transduces hepatocytes, with uptake of the enzyme from the circulation by other organs. The treated animals did not develop cardiac disease or corneal clouding, and skeletal, cartilage and synovial disease was ameliorated. Nonviral mediated gene transfer to the liver of MPS I and VII mice also improved the phenotype. Still, the major challenge remains to target the brain, which currently requires multiple brain injections with accompanying risks, and immunosuppression to prevent pan-encephalitis that develops secondary to an immune response to the transgenic protein. Widespread correction of the pathological lesions in an MPS VII mouse has been observed with AAV gene transfer, a vector that elicits less of an immune response.
Fetal gene delivery is an alternative strategy. Injection of adenovirus into the cerebral ventricles of fetal mice led to widespread and long-term gene expression throughout the brain and the spinal cord. In the same study, delivery of a therapeutic gene to the cerebral ventricles of fetal MPS type VII mice prevented damage in most of the brain cells before and until 4 months after birth. A similar study using an AAV vector had comparable results but with longer expression.
From a translational perspective, direct vector administration into the fetal brain or ventricles through the fetal skull for prenatal gene transfer is technically difficult using minimally invasive injection techniques, although this has been achieved in nonhuman primates and sheep (AL David, unpublished work) under ultrasound guidance. In contrast, ultrasound-guided access to the human fetal circulation is commonly used for fetal blood sampling and transfusions in clinical practice, with minimal fetal loss rate or complication. Newer AAV vectors of serotypes 2/9 have an astonishing ability to transduce cells of the nervous system after systemic injection in neonatal mice, cats and nonhuman primates. Furthermore, intrahepatic umbilical vein injection in fetal macaques of AAV 2/9 gives comprehensive transduction of the central nervous system (CNS), including all areas of the brain and retina, and the peripheral nervous system, including the myenteric plexus. Systemic transduction was also achieved using these AAV serotypes in fetal mice and macaques, particularly to epithelial and muscle cells. A fetal gene therapy approach using AAV therefore has the potential to treat congenital disorders affecting a wide range of body systems, including the CNS.
Fetal Growth Restriction as a Candidate Disease
Severe fetal growth restriction (FGR) affects 1 in 500 pregnancies and is a major cause of neonatal morbidity and mortality. The underlying abnormality in many cases is placental insufficiency, whereby the normal physiology process of trophoblast invasion that converts the uterine spiral arteries into a high-flow large conduit for blood fails to occur. There is no therapy to rescue poor uteroplacental circulation or improve fetal growth. FGR is commonly diagnosed when fetal measurements fall below the expected gestational age charts. Abnormally increased uterine artery vascular resistance is classically seen in midgestation.
A targeted approach to the uteroplacental circulation is needed because intravascular infusion of sildenafil citrate, a nitric oxide donor drops systemic blood pressure and had detrimental effects on growth-restricted sheep fetuses. In the pregnant sheep, transient local overexpression of VEGF mediated via adenovirus vector injection into the uterine arteries increased uterine artery blood flow and significantly reduced vascular contractility. VEGF expression was confined to the perivascular adventitia of the uterine arteries, together with new vessel formation, supporting the local effect of gene transfer. These effects are long term, lasting from midgestation (80 days) through to term (145 days). In an FGR sheep model in which uterine blood flow is reduced by 35% in midgestation, uterine artery injection of the same dose of Ad.VEGF significantly improved fetal growth in late gestation, ( Fig. 46.2 ) and lambs continued to thrive during the neonatal and early postnatal period. The effect has been replicated in FGR guinea pigs.
