Rationale, Experimental Work, and Barriers to Success

Conceptualization of hematopoietic stem cell transplantation dates back to 1945 when Owen found that dizygotic cattle twins shared placental circulation. Later, Billingham et al. demonstrated that early fetal exposure to antigens allowed for tolerance to the exposed antigen, meaning that postnatal treatment was better tolerated. These investigators isolated cells from various organs in one strain of mice, and these were then injected into fetuses of a different strain. After birth, these mice were able to tolerate skin grafts from mice of the transplanted strain. This early study set the groundwork for future research exploring the fetal environment as an avenue for tolerance to foreign antigens.

IUHCT was later performed in anemic mice with a mutation in the c-kit gene, which is responsible for the differentiation, proliferation, and survival of hematopoietic stem cells. Using a transplacental injection of hematopoietic stem cells, it was shown that stem cells from either fetal livers or adult bone marrow were able to correct the anemia. Mice that were more anemic supported the engraftment of stem cells, whereas mice that were not anemic did not engraft. Further studies demonstrated that engraftment of stem cells after IUHCT could be achieved in mice that had no stem cells or mice with severe combined immunodeficiency (SCID), in which a genetic defect prevents the proper maturation of T and B cells. Mice that were deficient in stem cells demonstrated donor-derived multilineage engraftment, whereas SCID-transplanted mice were only able to produce donor-derived lymphocytes. These were the first studies to show that IUHCT could lead to successful engraftment if the host was deficient in that particular lineage of cell, whereas host cell competition limited the effectiveness of IUHCT and explained why IUHCT in immunocompetent hosts had not led to successful engraftment.

Studies of stem cell engraftment after IUHCT in animals without inherent defects in their hematopoiesis showed engraftment below clinically relevant levels including mice, dogs, goats, and primates. Further research has improved techniques for delivery of transplanted cells. Intrahepatic and intravascular cell delivery allows greater numbers of hematopoietic stem cells to be transplanted and results in increases in overall chimerism. Reaching levels of 1%–2% engraftment result in donor-specific tolerance of donor cells. Mice that reach these levels of chimerism have been found to be able to accept donor skin grafts and do not show immunoreactivity to donor antigens.

Even in the fetal environment, an immune response can still limit donor cell engraftment. In fetal mice, although the fetal immune system matures later than in humans, the main culprit for rejection of transplanted cells has been shown to be the maternal immune system. A barrier to engraftment are maternal T cells that cross the placenta. In human fetuses, in which there is earlier maturation of the immune system, it has been demonstrated that the trafficking of maternal antigens to the fetus makes fetal T cells less reactive to the mother. Thus, transplantation of maternal stem cells into fetuses should overcome the maternal immune response to the transplantation, particularly because the human fetus should have T cells specific for these maternal antigens.

Studies of In Utero Stem Cell Transplantation in Large Animal Models

One of the major early successes with IUHCT was in the sheep model. Using an intraperitoneal injection of fetal stem cells, allogeneic engraftment could be shown in 75% of recipients with engraftment as high as 30%. The sheep continued to have chimerism 9 months after transplantation and never developed evidence of graft-versus-host disease. This early success leads to a great deal of enthusiasm; however, initial clinical application of IUHCT failed for numerous diseases.

Studies in the pig model were more encouraging, and it was demonstrated that IUHCT recipients were able to develop donor-specific tolerance and to tolerate donor-matched kidney transplants without immunosuppression. In studies of a canine model of leukocyte adhesion deficiency (CLAD), dogs underwent IUHCT via an intraperitoneal injection, which resulted in low levels of engraftment in two dogs and improvement in the disease phenotype. The dogs were then able to undergo a booster transplant to improve their chimerism to clinically relevant levels. Around this time, discoveries made in the murine model of IUHCT were used to further improve donor cell engraftment, specifically the use of maternal cells. Dogs were then treated with maternally derived donor stem cells via intracardiac injection. Chimerism was significantly improved with the new approach, and 21 of 24 dogs had engraftment greater than 1% with an average of 11%. This verified that the barriers found in the mouse model were limiting in the large animal models as well.

Early Clinical Experience in Humans

There has been little clinical success in IUHCT in humans due in large part to the unforeseen barriers to engraftment that were only identified after the early trials. At this time, 26 cases of IUHCT have been attempted in human fetuses for a variety of diseases with only bare lymphocyte syndrome and X-linked SCID, demonstrating clinically relevant engraftment and disease amelioration. It appears that in all successful cases, engraftment occurred due to the lack of host immune cell competition. Despite limited success, numerous lessons were learned from these early experiences and multiple improvements have been made based in part on the identified barriers found in animal studies. Fetal access has improved significantly. Ultrasound guidance allows for safe access to the umbilical vein and allows for a more efficient intravascular delivery of hematopoietic stem cells. In addition, it is now recognized that a protocol based on transplantation of a high dose of maternal-derived stem cells, injected intravascularly, has the best chance of allowing engraftment. Even if clinically significant levels are not achieved, a low level of chimerism could allow a postnatal “booster” transplantation with minimal conditioning. This approach has been used in mouse models of thalassemia and sickle cell disease.

Hemoglobinopathies

Hemoglobinopathies are a group of clinical disorders caused by genetic defects that cause either an abnormal structure of hemoglobin or insufficient production. IUHCT can potentially correct or lessen the disease burden of any disease that results from defective hematopoiesis, including hemoglobinopathies. Hemoglobinopathies are ideal targets for IUHCT by virtue of their overall prevalence in the general population and their deficiency arising from within the hematopoietic stem cell population. Postnatal allogeneic stem cell transplantation for sickle cell disease has been demonstrated as curative for patients with symptomatic sickle cell disease and thalassemia. The most prevalent clinically severe hemoglobinopathy is sickle cell anemia. Its clinical manifestations and disease burden vary greatly from individual to individual. The disease is characterized by vasoocclusive crises that require prolonged hospitalization. The overall cost of sickle cell disease in the United States was estimated at $460,151 per person and this disorder afflicts approximately 100,000 Americans. There have been several advances that have improved and prolonged the lives of patients with sickle cell disease ; however, hematopoietic stem cell transplantation remains the only curative treatment. Unfortunately, postnatal stem cell transplantation carries a lifetime risk of graft-versus-host disease and requires myeloablative preconditioning. Although clinical experience with high chimerism levels is sparse, in the preclinical and few clinical cases, graft-versus-host disease is rarely observed when the amount of mature T cells is controlled for.

Thalassemias are also common; gene frequencies are estimated to range from 2.5% to 15% in the tropics and subtropics. α-Thalassemia major manifests with severe anemia in utero, including hydrops fetalis and fetal demise. If left untreated, α-thalassemia major (ATM) is lethal, but fetal therapy with in utero transfusions can be lifesaving for these fetuses, and fetuses who are treated with in utero transfusions can survive with reasonable neurologic outcomes. This disease represents an ideal target for in utero therapy because it is fatal in utero without treatment. Because the fetus already needs to receive an invasive procedure in utero to survive, the HSC transplantation can be performed at the same time as the transfusions.

Lysosomal Storage Disorders

Lysosomal storage disorders are a group of diseases that result from a genetic deficiency in an enzyme required for the normal metabolic functions of the lysosome. Lysosomes are therefore unable to break down their complex substrates, which then accumulate within the cells leading to cellular dysfunction. Consequences of the disease include intellectual disability, skeletal dysplasias, pulmonary insufficiency, and, in severe cases, hydrops fetalis and in utero fetal demise. Patients with certain lysosomal storage disorders can be treated after birth with enzyme replacement therapy (ERT). The deficient enzyme is transfused and taken up by various cells, decreasing the extent of cellular damage. This approach is limited, however, and ERT does not appear to cross the blood-brain barrier and does not significantly improve the central nervous system deterioration. Additionally, the deficient enzyme is seen as foreign by the patient’s immune system. Postnatal therapy is limited by the development of antibodies against the exogenous enzyme, which leads to decreased effectiveness of the enzyme and allergic responses. As a result, immunosuppression is required for continued enzyme therapy.

To improve outcomes in patients with lysosomal storage disorders, stem cell transplantation has been performed in multiple lysosomal storage disorders including Hurler disease, Batten disease, metachromatic leukodystrophy, Krabbe disease, and I-cell disease. Clinical results have varied based on the disease, which stem cells were transplanted, age at transplantation, and chimerism levels, but overall clinical results have been promising. In Hurler disease specifically, improvements were noted in survival, cognitive development, preservation of hearing, corneal clouding, and respiratory support requirements. Earlier age at transplantation and posttransplant enzyme levels were predictive of outcome improvement.

In utero hematopoietic stem cell transplantation coupled with in utero enzyme replacement therapy could greatly improve outcomes for this group of patients for multiple reasons. First, many affected fetuses die before birth or before enzyme replacement therapy can be initiated. Second, fetal accumulation of toxic metabolic by-products leads to numerous developmental insults before the possible initiation of enzyme replacement. Neurologic outcomes may be improved with earlier therapy, particularly during fetal development, either by providing the enzyme before formation of the blood brain barrier or by early receptor-mediated transport. Third, fetal exposure to the exogenous enzyme may induce tolerance, which would make postnatal therapy more efficacious.

Osteogenesis Imperfecta

Osteogenesis imperfecta (OI) is a genetic connective tissue disorder that is caused by defects in the proper synthesis of type I collagen. The disease is characterized by fragile bones among multiple other ailments, and there are multiple forms with significant variability in clinical presentation, from perinatal death to a normal life span. Current medical therapies seek to improve bone strength, although efficacy is limited and, in the most severe form, is available too late to improve survival. In utero therapy has been focused on the use of mesenchymal stem cells because of their nonimmunogenic nature and ability to engraft and form bone. In utero therapy with mesenchymal stem cells for OI is particularly appealing because it takes advantage of the chemotactic properties of MSCs that guide them to sites of injury (intrauterine fractures are common) and the small fetal size (30–100 g at 14–16 weeks’ gestation) allows for lower stem cell dosing and greater potential proliferation throughout gestation.

In utero stem cell transplantation for OI has been performed previously. In 2003, Westgren et al. performed in utero stem cell transplantation at a gestational age of 29 weeks. They found engraftment of 5% and no intrauterine fractures. In 2005, another transplantation was performed at 32 weeks’ gestation with 7% chimerism. At 2 years of age, the patient had suffered three fractures.

In Utero Gene Therapy

In utero gene therapy (IUGT) is another potential fetal molecular therapy for single-gene disorders with encouraging preclinical data. The general strategy for gene therapy involves packaging genetic sequences within inactivated viral vectors, which retain the ability to transduce genetic material to host cells. This takes advantage of existing viral infectious machinery but with the theoretical removal of treatment toxicity. As described in more detail below, some vectors have the ability to integrate into the host genome, allowing for transmission to daughter cell populations and long-term transgene expression. This section details the therapeutic advantages of using this treatment during the fetal period, outlines existing research with preclinical and postnatal clinical models, and describes some of the risks and potential adverse effects of this treatment modality.

Avoidance of Disease Onset

Prototypical diseases for in utero gene therapy are those that manifest early in life, including the thalassemias, hemophilia, and severe combined immunodeficiency (SCID). Many of these diseases result in severe and irreversible damage before birth. These diseases typically require lifelong treatment that is noncurative. Correction in utero has the potential to improve the quality of life for the patient and family. The impact of genetic disease was assessed in a study out of the Case hospital system in Ohio; review of the hospital’s records showed that 96% of children with chronic underlying disorders had a clear genetic etiology or susceptibility, with one-third of pediatric ICU admissions related to diseases with a genetic basis, such as cystic fibrosis and sickle cell disease. Liver diseases such as urea cycle disorders typically present with severe neonatal manifestations in the first week of life; they can result in irreversible multiorgan damage and death. In a study on patients with severe hemophilia, prophylactic treatment with clotting factors earlier in life resulted in lower incidence of arthropathies secondary to bleeding episodes later in life.

Unique Immunologic Advantage

The fetal immune system provides several distinct advantages for gene therapy with respect to immune responses to the viral vector as well as the proteins encoded by the transgene. Fetuses have theoretically never encountered the common viral vectors for gene therapy; thus it is unlikely that preformed antibodies will exist against viral vectors. In adulthood, both innate and adaptive immune responses have been demonstrated in response to the administration of these vectors. Administering viral vectors in a fetal host before the formation of these antibodies could potentially avoid inflammatory reaction and associated vector neutralization.

Immune responses to the proteins encoded by the transgene can also limit the efficacy of postnatal therapy. Preclinical models have shown that adults are capable of developing an immune response that limits the effectiveness of viral transgene. In a study on rhesus macaques, different doses of human factor IX were given via first-generation adenoviral vector to determine whether sustained expression of the human coagulation factor could be achieved. Human factor IX was detectable for several weeks but disappeared along with development of high titers of antifactor IX antibody. The relative immaturity of the fetal immune system could allow for immunologic tolerance and the avoidance of the above complications.

Many preclinical experiments have supported this idea of a tolerant fetal environment. Direct intraperitoneal injection of retroviral vector encoding β-galactosidase in fetal sheep led to development of postnatal tolerance to the protein product. This was demonstrated by administering a postnatal booster of the protein and showing that the animals have a blunted ability to form an antibody response to the protein. A similar study in mice involved in utero yolk sac injection of adenoviral vector carrying human factor IX transgene. Injection of the human protein in adult mice resulted in persistent expression of this protein as well as undetectable antifactor IX antibody levels, whereas control mice developed high antibody titers and associated loss of factor IX expression. This induction of tolerance facilitates postnatal repeat therapy as well, because both vector and vector transgene are introduced with in utero treatment.

These preclinical experiments have been expanded to nonhuman primates as well, where in utero treatment led to long-term expression of human factor IX, with evidence of tolerance induction. AAV5 and 8 vectors were used to deliver human factor IX in late gestation with macaques, which led to liver tropism and stable expression for more than 6 years. Challenge with postnatal AAV led to subclinical levels of anti-AAV antibody expression, indicating that tolerance had been induced with prenatal treatment. There was no evidence of toxicity from observed viral integration, nor was there evidence of germline transmission.

Crossing Blood-Brain Barrier

The blood-brain barrier provides a unique challenge in postnatal therapy—peripheral administration of drugs and therapies such as stem cell transplant or gene therapy generally does not cross this barrier. In utero gene therapy does show promise in the central nervous system—direct intraventricular injection of adenoviral vector carrying the missing enzyme in mucopolysaccharidosis VII (MPS7) resulted in widespread gene expression and amelioration of disease in early life. Other studies have used postnatal full-body irradiation in mice to circumvent the blood-brain barrier, a toxic adjunctive procedure that would be avoided if in utero treatment can provide access to the brain before formation of that barrier. Several AAV vectors have been demonstrated to successfully cross the blood-brain barrier in the perinatal period in murine and nonhuman models. Neonatal treatment in the mouse model with the AAV9 vector carrying the SMN gene has led to phenotypic rescue of spinal muscular atrophy. In the nonhuman primate model, both fetal and early neonatal treatments with AAV9 vector successfully transduced a wide range of systemic cells including neurons as well. Gaining access to the blood-brain barrier with in utero treatment, especially without toxic systemic treatments such as whole-body radiation, is another potential benefit to this approach.

Access to Stem Cell Populations

Similar to the rationales cited in fetal stem cell therapy, IUGT has the potential to access stem cell populations during a time when they exist in higher relative frequency to other cells. Usage of viral vectors that integrate into the host genome would allow for significant expansion of the viral transgene into daughter cells. An experiment using VSV-G pseudotype equine infectious anemia with a transgene for β-galactosidase in mice resulted in multiorgan gene expression that was uniquely found to be in clusters of cells. This suggested clonal expansions of originally transduced cells. IUGT with lentiviral delivery of lacZ has resulted in transduction of muscle stem cells (satellite cells) in a murine muscular injury model.

Hemophilia

There have been recent promising results in safety and efficacy of IUGT with preclinical models of hemophilia, suggesting that this may be a good candidate for initial IUGT clinical trials. Although the disease is no longer debilitating, it does relegate patients to a lifetime of frequent infusions, with alloantibody development reducing treatment efficacy. With hemophilia A, the relatively large size of the factor VIII coding region has led to limited gene therapy applications, as vectors such as AAV have limited packaging capacity. A recent study overcame this obstacle, and six of seven patients had normal long-term factor VIII expression levels. No liver toxicity of neutralizing antibodies was developed, and although anti-AAV5 capsid antibodies were detected, no cellular immune responses were found. These exciting phenotypic improvements after single doses, as well as the lack of significant adverse effects, indicate that hemophilia may be an ideal first target for IUGT. There were cases in these hemophilia studies where existing arthropathies secondary to recurrent bleeds could not be reversed despite successful long-term transgene expression—IUGT would address the disease before the development of such chronic sequelae.

In cases of severe human hemophilia B, another recent landmark study was published where a single injection of AAV-mediated hyperfunctional factor IX variant gene led to high levels of functional factor IX, along with decreased bleeding rates and factor infusion requirements. Of the 10 patients treated, 8 patients did not require any further factor infusions over the follow-up period of over a year, and 9 patients had no further bleeding episodes after vector treatment. Although low titers of neutralizing antibody titers were detectable, the patients universally had lasting factor IX activity (35.5 ± 18.7% of the normal value) throughout the study period. The improvement in bleeding episodes and factor dependence in all patients provides further evidence that gene therapy had significant potential with this disease.

Lysosomal Storage Disorders

As described above, lysosomal storage disorders are good candidates for in utero treatment because phenotypic manifestations often occur before birth. Although early clinical studies with retrovirally transduced stem cells were unsuccessful, more recent efforts with AAV and retroviral-mediated vector in conjunction with liver-specific promoters have shown improved efficacy. Retroviral vectors have been used in MPS7 to successfully transduce up to 20% of hepatocytes and have resulted in normal long-term expression of the missing B-glucuronidase. These high expression levels were associated with phenotypic improvements in affected dogs. As the collective prevalence of lysosomal storage disease is common (1 in 9000), early diagnosis and treatment with gene therapy vectors has the potential to ameliorate storage accumulation before development of systemic symptoms.

Severe Combined Immunodeficiencies

Severe combined immunodeficiencies are attractive targets for IUGT because they are typically associated with a defective hematopoietic niche that confers a survival advantage to treated cells. In mice with absence of both B and T cells, bone marrow–derived stem cells successfully reconstituted functioning B and T cells. This approach has been successful for multiple variants of SCID and has been extended into the clinical setting as well. X-linked SCID patients transplanted with autologous CD34+ stem cells that had been transduced with retroviral-mediated gamma chain cDNA resulted in immunologic reconstitution in 9 of 10 patients. These resultant reconstituted T cells were demonstrated to function normally in response to immunologic challenges.

Hemoglobinopathies

Using a similar technique to that described above, gene-modified HSCs have shown efficacy in both preclinical and human β-thalassemia models. Initial mouse models using lentiviral vectors resulted in correction of anemia and red cell morphology. The first human trial involved a β-thalassemia case where CD45+ cells were transfected with lentiviral-delivered β-globin—this led to long-term transfusion independence attributed to the expansion of multiple HSC clones. The study was extended to other patients with varying degrees of β-thalassemia severity, with correction of all but the most severe forms—this variant had partial correction with the treatment. As a result of these initial successes, multiple trials are now underway using postnatal gene therapy for both thalassemias and sickle cell anemias.