Improvement of Canine XSCID by In Vivo Gene Therapy Using RD114/TR-Pseudotyped SIV Lentiviral Vectors.

Abstract

X-linked severe combined immunodeficiency (XSCID) is characterized by profound immunodeficiency (dysfunctional B cells; absence of T and NK cells) and early mortality, caused by mutations in the IL2RG gene encoding the common gamma chain (γc) of receptors for interleukins (IL)-2,-4,-7,-9,-15 and -21. The standard therapy for XSCID in infants is a T cell depleted bone marrow transplant with either none or very modest conditioning. While 5-yr survival is >95% with HLA-matched sibling donors, most patients receive a haploidentical graft from a parent, resulting in lower survival and less robust immune reconstitution that may be limited to only the T cell lineage. Ex vivo autologous stem cell gene therapy has emerged as an alternate treatment capable of achieving substantial to complete immune reconstitution in infants without a sibling donor. However, some children have developed lymphocytic leukemia, which appears in part to be related to vector insertional mutagenesis. The gammaretroviruses currently in use have potent enhancers in the LTR and have a predilection to insert at the 5′ end of genes. Self-inactivating (SIN) lentivirus vectors may be advantageous because they do not show this property and can be constructed with internal promotors that have less enhancer activity together with insertion of insulators. We have constructed SIN simian immunodeficiency viral vectors (SIVmac) encoding human γc (hγc), with or without a double copy chicken insulator core element in the 3′ LTR. In addition we have pseudotyped the vector with a chimeric RD114 envelope to enhance targeting of hematopoietic stem cells (HSC) and avoid the cytotoxicity of the VSV-G envelope traditionally used with lentivectors. We have previously shown that it is possible to correct a dog model of XSCID using a RD114-pseudotyped gammaretroviral vector encoding the dog γc for in vivo gene therapy. Unlike the mouse model of XSCID which lacks B cells, the dog model more closely resembles the phenotype in humans. We decided to use our in vivo dog model for preclinical testing of safety and efficacy of our SIV hγc vectors. Viral particles were produced by transient transfection of 293T cells with a 4-plasmid system and concentrated by high-speed centrifugation. 30 mls (average 2.4–3×107 viral particles in total) was injected IV into 2–5 day-old pups. Transgene marking in blood lymphocytes was detected as early as 2 wks after treatment, increased within the first 6–8 wks and became relatively stable thereafter. The absolute lymphocyte count was normalized in one dog (injected on the second day after birth) by wk 6 (4000, 76% hγc+), and improved in two other dogs (422, 29% hγc+ at 18 wks; 498, 27% hγc+, at 16 wks) which had received the same amount of virus over 2 days. Furthermore, up to 5% of the myeloid lineage showed gene marking at 10–18 wks after viral delivery, indicating that early committed progenitors or HSCs had been transduced. It is noteworthy that our finding that the hγc can improve the disease phenotype in XSCID dogs makes it an excellent large animal model for preclinical evaluation of vectors. In summary, we have demonstrated that in vivo delivery of SIV lentiviral vectors expressing hγc efficiently reconstitute the T-lymphoid compartment in the XSCID canine model. We plan to monitor these dogs closely for potential adverse events.