Mending RAG2: gene editing for treatment of RAG2 deficiency

In this issue of Blood Advances, Pavel-Dinu et al¹ demonstrated successful immune reconstitution in a mouse model of severe combined immunodeficiency (SCID) using RAG2-deficient human hematopoietic stem cells that were genetically corrected with a CRISPR/Cas9 system. By insertion of a codon-optimized transgene within the native RAG2 locus, their approach provides a potential universal gene therapy that could be used with most RAG mutations.

Autosomal recessive mutations in RAG1 and RAG2 genes are the second most common cause of SCID in the United States, resulting in low to absent T and B cells and normal numbers of natural killer (NK) cells.² Hypomorphic RAG mutations lead to a broad range of immune deficiency disorders, which include Omenn syndrome, combined immunodeficiency, and autoimmunity with granulomatous disease.³ Although newborn screening has enabled detection of SCID in many countries, unscreened patients may not present until onset of infections or autoimmune disease, which may be delayed in onset.⁴ Hematopoietic stem cell transplantation is potentially curative, and although the use of a matched related donor can achieve acceptable immune reconstitution without conditioning in some patients, this is not an option for most patients. Reduced intensity or myeloablative conditioning is typically required for good immune reconstitution when using haploidentical or unrelated marrow donors for patients with RAG SCID but involves risk of chemotherapy-related toxicity.⁵ 

Ex vivo gene therapy allows for the use of a patient’s autologous hematopoietic stem cells, minimizing exposure to chemotherapy and negating risk of acute and chronic graft-versus-host disease. Gene addition have been used in many previous clinical trials via integrating viral vectors, with evidence of efficacy and improved safety with modern lentiviral vectors.⁶ However, lentiviral gene addition is not a viable option for many highly regulated genes. Previous investigators have demonstrated immune reconstitution of RAG1 deficiency via lentiviral gene addition in a murine model, with only subtle defects in the T-cell repertoire after gene therapy.⁷ Although a study of lentiviral-based gene therapy for RAG1 deficiency is ongoing, gene insertion using CRISPR/Cas9 represents an attractive option to provide the corrected RAG gene under the endogenous promoter.

In this article, Pavel-Dinu et al used CRISPR/Cas9 editing for targeted insertion of a codon-optimized RAG2 transgene (coRAG2) within the RAG2 coding region. They used an adeno-associated virus 6 (AAV6) vector for coRAG2 delivery to this region via homologous recombination in RAG2-null cells. They showed high specificity for the RAG2 gene with no increase in off-targeting vs mock-edited cells. Gene-modified hematopoietic stem cell progenitors (HSPCs) showed normal development of multilineage hematopoietic progenitor cells.

Next, frozen peripheral blood HSPCs from a patient with RAG2 deficiency with compound heterozygous mutations (RAG2^null HSPCs) were treated with CRISPR/Cas9-AAV6. A higher multiplicity of infection of AAV6 (5000) resulted in viral toxicity and reduced viability of gene-corrected HSPCs. This was ameliorated by use of a p53 inhibitor (i53) to inhibit nonhomologous end joining, which allowed a lower AAV6 multiplicity of infection of 2500 to be used. This CRISPR/Cas9-AAV6 + i53 system resulted in gene correction by homologous recombination (HR-GT) of 19.2% of RAG2^null HSPCs. HR-GT HSPCs were transplanted into irradiated NSG-SGM3 mice. Twenty-two weeks after transplantation, there was 4% human chimerism in bone marrow cells. Mice that received transplantation with HR-GT cells showed increased production of immature and mature B cells compared with mice that received transplantation with untreated RAG2^null cells and were comparable with mice that received transplantation with healthy HSPCs. The HR-GT B cells were polyclonal, producing a full range of immunoglobulin heavy chains.

The majority of NSG-SGM3 mice that received transplantation with HR-GT HSPCs had detectable T-cell development in bone marrow, with 3 developing splenic T-cell numbers comparable with those that received transplantation with healthy human HSPCs. The T cells in these 3 mice were single-positive CD4⁺ and CD8⁺ cells with a normal polyclonal distribution of T-cell receptors, including normalized use of 5’ TRAV and 3’ TRAJ genes that are absent in patients with hypomorphic RAG2. Mice that received transplantation with HR-GT cells also showed normalization of NK cells with reduction in CD56^bright NK cell skewing compared with mice that received transplantation with RAG2^null cells. Overall, Pavel-Dinu et al showed feasibility of using CRISPR/Cas9-AAV6 with a p53 inhibitor to universally correct HSPCs with RAG2 mutations, resulting in restoration of lymphoid development when transplanted into SCID mice.

The efficiency of homology-directed repair (HDR) for correction of mutations in HSPCs using CRISPR/Cas9 has remained an ongoing challenge. Nondividing cells, such as HSPCs, do not express high levels of the homologous recombination enzymes needed to insert the corrected sequence after CRISPR/Cas9-mediated double-strand breakage, resulting in decreased frequency of HDR gene correction and increased frequency of nonhomologous end joining. Pavel-Dinu used i53 to optimize the rate of HDR, achieving nearly 20% correction. This yield is similar to other preclinical studies using CRISPR/Cas9 and other gene editors for correction of inborn errors of immunity, including X-linked SCID, CD40L deficiency, and others (see table).^8-15 Toxicity associated with recombinant AAV6 may be partly resolved via use of a good manufacturing practices-grade vector. It is not clear what correction rate is necessary in engrafted HSPCs to provide robust, long-lasting immune reconstitution, although ex vivo lentiviral gene therapy trials for X-linked SCID suggest that even a small fraction of gene-corrected HSPCs may be curative, because low vector copy numbers in CD34⁺ marrow stem cells were adequate to facilitate T-cell reconstitution.¹⁶ 

This article has shown that genome editing with CRISPR/Cas9 is a promising modality for potential correction of the full spectrum of RAG2 deficiency, with the ability to maintain physiologic regulation of the corrected gene. Future studies may help to address remaining questions, including the minimal threshold of corrected HSPCs to achieve a lasting cure and the degree of conditioning required to achieve it.

Conflict-of-interest disclosure: M.D.K. is an author for Elsevier (UpToDate) and has received research funding from Chiesi Pharmaceuticals. E.D.H. declares no competing financial interests.