In this issue, Persons and colleagues (page 506) report another advance in the now rapidly moving field of gene therapy for the β-chain hemoglobinopathies, β-thalassemia (β-thal), and sickle cell disease (SCD). The hemoglobinopathies are the most common human inherited diseases, and while the milder forms are increasingly amenable to drug therapies, the only real cure for the severe forms of these diseases has been bone marrow transplantation (BMT). But the scarcity of human leukocyte antigen (HLA)-matched donors and the high morbidity associated with complete myeloablation have limited the use of BMT for the treatment of SCD and β-thal. Recent breakthroughs in vector design allowing stable transfer of globin genes that can be expressed at therapeutic levels in red cells (May et al, Nature. 2000;406:82-86; Pawlick et al, Science. 2001;294:2368-2371) have made the hemoglobinopathies candidates for gene replacement therapy. In mouse models, the introduction of a globin gene (β-globin for β-thal and γ-globin to inhibit sickling in SCD) into a portion of autologous hematopoietic stem cells (HSCs) has led to permanent cures of β-thal or SCD mice receiving transplants (May et al, Blood. 2002;99:1902-1908; Rivella et al, Blood. 2003;101:2932-2939; Persons et al, Blood. 2003;101:2175-2183). Combined with the observations that stable mixed chimerism was associated with the successful cure of severe β-thal or SCD in a subset of human recipients of transplants, it appears that 25% to 50% of corrected cells are sufficient for a full cure. However, the low frequency of gene transfer into human hematopoietic stem cells (about 1%) and the morbidity of full myeloablation in patients with hemoglobinopathies have prevented the application of the recent advances in globin gene therapy to humans.

Persons et al have addressed these problems using 2 different mouse models of stem cell gene transfer. In the first model, they gave β-thal intermedia mice a nonmyeloablative conditioning regimen and transplanted into them a small number of normal bone marrow cells that were transduced with a retrovirus vector containing the MGMT gene, which confers resistance to O6-benzylguanine (BG). The resulting bone marrow chimeras resemble low-level engraftment of HLA-matched normal cells after partial myeloablation and a low frequency of gene transfer. Prior to treatment the animals receiving transplants were indistinguishable from β-thal mice. Following treatment with BG, the level of transduced normal cells rose from less than 10% to 56% in 6 of 10 animals with a concurrent normalization of all red cell indices.

In the second model, Persons et al introduced a lentivirus vector containing both a human γ-globin gene and the MGMT gene into mouse bone marrow cells and transplanted them into recipient mice. After BG selection the number of γ-globin producing red blood cells increased from a pretreatment level of less than 1% to more than 60% in 5 of 7 animals.

The 2 studies demonstrate a conservative approach that dramatically lowers the risk of transplantation-related complications to the patient and does not require high rates of HSC transduction. The most pressing issue facing Persons et al and other investigators in this field is to develop vectors that express higher levels of globin per vector copy, so that the maximum amount of globin protein can be produced from the minimum number of insertion events. In addition, while the MGMT selection is quite powerful, the frequency of gene transfer to human cells must still be improved to allow the most efficient treatment. I predict that these problems will be solved in the near future and that the first clinical trails for β-thal and SCD will resemble those described in this issue by Persons et al.