Gene therapy for erythroid congenital disorders  

Major issues that we are trying to address

Luisa Jung GT 2021

In the past 20 years, gene therapy for hematopoietic stem cells (HSCs) has made significant progress, leading to the approval of treatments for hemoglobinopathies and a large number of ongoing clinical trials worldwide. The β-hemoglobinopathies were prioritized because of their high frequency and severity, but other genetic disorders such as α-hemoglobinopathies, Shwachman-Diamond and Diamond-Blackfan syndromes, Congenital Dieserythropoietic Anemia, Pyruvate Kinase and Phosphoglycerate Kinase deficiencies, Congenital Erythroblastic Porphyria and Hereditary Sideroblastic Anemia are curable through CD34+ cell transplants or gene therapy in HSCs.

The rarity of these diseases poses a significant technical and economic challenge for the field, as developing mutation-specific therapies for them using current technologies would be prohibitively expensive due to the large number of rare nucleotide substitutions, indels, and deletions that underlie these disorders.

One possible solution to this conundrum is to use CRISPR/Cas9 to knock in therapeutic transgenes to cure these disorders. Compared to lentiviral therapy, knockins carry no risk of insertional mutagenesis and result in predictable levels of expression. Compared to CRISPR/Cas9-based gene repair, gene editing, and enhancer editing, knockins have the advantage of being mutation-agnostic. Therefore, while each method has its own advantages and disadvantages, and may be most appropriate for specific diseases, knockins could become an important tool in the increasingly large gene therapy arsenal.

Issues that need to be solved to implement the knockin approach include finding the best integration sites for the therapeutic transgenes and developing more efficient methods to perform knockins.

One of the focus of the lab is to develop knockin approaches in HSCs, usign the α- and β-hemoglobinopathies as models.

Gene therapy for the hemoglobinopathies

Globin cluster Farashi

Regulation of the globin genes: Hemoglobin (Hb) is a tetramer, composed of two α-like and two β-like globin chains. Globin chains are encod-ed by three α-like genes (HBZ, HBA1 and HBA2) on chr 16 and five beta-like genes (HBE1, HBG1, HBG2, HBD and HBB) on chr. 11 that are expressed differentially from embryogenesis to adulthood, resulting in the production of multiple hemoglobin tetramers at different stages of development.

Expression of the β-like globin genes is controlled by the Locus Control Region (LCR), a super-enhancer composed of 5 DNAse I hypersensitive sites located 6 to 25 kb upstream of the HBE1 gene. Globin switching results from the formation of DNA loops that bring the LCR in close proximity to the first accessible β-like globin gene promoter.

Early in development, the LCR interacts with the HBE1 gene, which is the closest gene to the LCR in the locus. As development proceeds, the HBE1 gene becomes inaccessible to the LCR, leading to activation of the HBG1 and HBG2 genes and a switch from embryonic hemoglobin (HbE) to fetal hemoglobin (HbF) expression. Later in development BCL11a contributes to silencing of the HBG1 and HBG2 genes which frees the LCR which can then activate the HBD and HBB genes, leading to the replacement of HbF by the adult hemoglobins (HbA and HbA2).

The α-like globin genes are also under the control of upstream enhancers located between 10 and 48 kb upstream of the genes, but there is only one switch from HBZ to HBA1 and HBA2. The HBA1 and HBA2 genes coding for the α1 and α2 globin chains start being expressed early in development and remain ex-pressed throughout life.

Hemoglobinopathies Mutations in any of the globin genes result in hemoglobinopathies which are classically divided into the thalassemias, which are quantitative disorders of globin gene synthesis and qualitative mutations leading to the production of structural hemoglobin variants with a propensity to form Hb polymers (such as in SCD), to precipitate, or to have altered oxygen transportation properties. About 6% of the human population carries a pathogenic globin variant with considerable quality of life and economic costs. This high gene frequency is explained by the fact that carriers of hemoglobinopathies are partially protected from falciparum malaria.

β-hemoglobinopathies: More than 350 mutations causing a β-hemoglobinopathy have been reported and the global incidence of carriers is about 1.5%. SCD affects more than 100,000 patients in the U.S and between 300,000 and 400,000 neonates globally each year. Other severe hemoglobinopathies are individually less frequent but together represent a considerable health burden. Treatment options depend on severity and include transfusions, iron chelation and drugs that increase HbF production or alter oxygen affinity. In the most severe forms, such BT major (caused by complete lack or very severe decrease in the production of β-globin chains), or SCD, hematopoietic stem cell transplantation is a curative option, but the lack of matched donors and the risks associated with graft failure limit its applicability.

Current method for gene therapy for the β-hemoglobinopathies: In the last decade, gene therapy has become an alternative to transplantation. The first approach to reach the clinic was lentiviral therapy, a gene addition approach, in which HIV-derived vectors carrying segments of the LCR and a therapeutic globin transgene are transduced ex vivo into mobilized patients' CD34+ cells, resulting in random integration of the vector. The transduced cells are then transplanted back into the patient after myeloablation. Results have been good for β-thalassemia (BT) but more limited for sickle cell disease (SCD). While there have been few reports of insertional mutagenesis, the levels of expression achieved are often too low to completely compensate for the globin deficiency, leading to improvement in the clinical situation but not a cure.

More recently, CRISPR-based trials aimed at repairing the mutated globin gene by homologous recombination or reactivating the γ-globin genes by silencing the BCL11a gene through editing of its erythroid enhancer have shown impressive results. The Bcl11a approach is currently the most advanced, and according to a press release in June 2022, HbF responses in 44 patients with BT and 31 with SCD were extremely encouraging. Nevertheless, the HbF levels reached after treatment were not sufficient for a complete cure in all patients, suggesting that some patients might benefit from alternate strategies.

α-thalassemia (AT): AT is caused by insufficient number of functional α-globin alleles compared to four functional copies in healthy individuals and most commonly occurs due to deletional alleles or mutations. The severity of the disease is proportional to the number of affected α-globin alleles. With three mutated alleles, patients have hemoglobin H (HbH) disease and may require blood transfusions for survival. Patients with four mutated alleles have Bart's hydrops fetalis (BartHF), which is generally lethal in utero, although the number of surviving BartHF patients is increasing due to advanced perinatal and neonatal care. Severe forms of HbH disease can lead to cirrhosis, hypertension, hypothyroidism, diabetes, and infertility. Although the prognosis for AT is similar to that for BT, the effects of α-globin gene mutations manifest throughout fetal and adult life, while mutations in the β-globin gene manifest only postnatally after γ-globin gene silencing.

Over 400 forms of deletional or non-deletional mutations of AT have been identified and an estimated 5% of the world's population has at least one deletion or mutation associated with AT. The highest prevalence is observed in Vietnam, followed by Cambodia, Laos, Thailand, Malaysia, Greece, and China. In the US, the incidence is rapidly increasing due to immigration from Southeast Asia. In some geographical areas, the frequency of patients with AT is similar to or higher than that of patients with BT, making AT an increasing public health burden on the existing medical system.

Safe harbors

Site-specific genomic integration has long been used in bacteria and in yeast to study transgenes due to its technical simplicity. In mammalian cells, transgenesis was initially only possible throuh DNA transfection or retroviral transduction, which led to random integration of the transgenes in the genome, carrying the risk of insertional mutagenesis. In addition to this risk, random integration is associated with variable transgene expression and silencing because the transgenes often interact with endogenous regulatory elements located in close proximity to the site of integration. These position effects have been an impediment to genetic engineering in general, and for gene therapy in particular because variations in expression levels decrease therapeutic efficacy.

Recombinase-Mediated Cassette Exchange (RMCE) More than 20 years ago, the Bouhassira lab developed RMCE, a CRE recombinase-based method to efficiently integrate transgenes at pre-determined genomic sites and used the method to study the β-globin LCR. These and other studies led to the development of lentiviral vectors that were used to provide the first proof-of-principle for gene therapy for SCD in mouse models, and shed light on how the interaction of transgenes and endogenous sequences resulted in variation of expression for both the exogenous and endogenous genes.

safe harbors: Building on these and other studies, Papapetrou et al. coined the term safe harbors, i.e. ideal genomic sites for integration of therapeutic transgenes. Safe harbors should provide high-level expression and be far from any genes whose disruption would be deleterious (such as oncogenes or other important genes). Five criteria aimed at keeping the transgene 50kb away from any transcription unit, and 300kb away from those involving cancer, were defined, but few such harbors have been found in the human genome.. More recently, the same author reported that these 5 criteria might be inadequate because advances in our understanding of the genome have revealed that 300kb might be too small and because of the increasing understanding of the 3D organization of the genome and of the importance of non-coding RNA and regulatory regions.

Although imperfect, three safe harbors, the AAVS1, the CCR5, and the ortholog of the Rosa26 genomic loci, have been used extensively for basic and translational studies but not yet in the clinic. A recent study has suggested that the human ribosomal DNA locus might also be a good safe harbor.

Gene knockins

With the successive development of zinc-finger nucleases, TALEN, and CRISPR/Cas9 technology, gene knockouts in hematopoietic stem cells have been achieved at increasingly higher efficiency. Gene editing approaches are also becoming more efficient, but gene knockins have remained particularly difficult without selection.

Since most genetic disorders are caused by a large number of rare SNPs, indels, or large deletions for which most patients are compound heterozygous, the lack of an efficient knockin strategy is a major impediment for gene therapy because designing repair or editing strategies for hundreds of mutations is not practical, can be difficult for compound heterozygous, and is not possible for deletions of entire genes.

Site-specific knockin of therapeutic transgenes, which is a mutation agnostic approach, is an attractive option because the same construct and gRNAs can be applied to most of the patients affected by mutations in a specific gene.

One of the project in the lab is to develop the β-globin cluster region as a safe harbor because of its genomic location in a gene-poor region and because the extensive knowledge that has been accumulated on the transcriptional regulation of this locus makes it possible to predict the effect of transgene insertion. Several methods are being tested to increase knockin efficiency in HSCs.