CD117 Antibody-Drug Conjugate Conditioning Allows Efficient Engraftment of Gene-Modified CD34+ Cells with Fertility Preservation in a Rhesus Gene Therapy Model

Hematopoietic stem cell (HSC) gene therapy is curative for multiple genetic diseases; however, it is limited due to toxicities from chemotherapy-based conditioning. To target HSCs directly, we developed an anti-human CD117 (c-KIT)-targeted antibody-drug conjugate (CD117-ADC) conjugated to a DNA-binding payload that cross-reacted with non-human primates. To investigate the activity of CD117-ADC in non-human primates, we administered a single injection of ADC without HSC infusion, resulting in a >99% depletion of CD34+CD90+ bone marrow cells in cynomolgus macaques (n=2 of 3 in 0.1 mg/kg and n=3 in 0.3 mg/kg); similar to the ablation observed with 4 doses of myeloablative busulfan (Bu) (n=3). With escalating ADC doses (0.2-0.6 mg/kg), the optimal timing of HSC infusion (6 days in 0.2 or 0.3 mg/kg and 10 days in 0.4mg/kg) was determined in rhesus macaques (n=1 each). The animal dosed 0.6 mg/kg required euthanasia on day 14 due to irreversible hematopoietic suppression and severe liver toxicity. In our preliminary transplantation study, dosing 0.2 mg/kg ADC to rhesus macaques (n=2) resulted in bone marrow depletion and modest-level engraftment of gene-modified cells (vector copy number per cell (VCN) 0.05±0.01) (ASH 2019).

Therefore, we evaluated higher single doses of CD117-ADC in rhesus macaques (0.3 mg/kg for ZL13 and ZJ62, and 0.4 mg/kg for H635 and H96G, n=4) and directly compared to Bu conditioning (5.5 mg/kg x 4 days, n=2). Mobilized rhesus CD34+ cells (ADC 6.4±2.9e6/kg vs. Bu 4.1±0.1e6/kg, n.s.) were transduced with a lentiviral vector for stable fetal hemoglobin (HbF) induction (Sci Transl Med. 2021). These cells (in vitro VCN 10.1±3.8 vs. 10.2±7.3, n.s.) were transplanted into autologous animals after ADC or Bu conditioning. The time course of blood cell count suppression and recovery was similar between ADC and Bu animals, except for a strong reduction of platelet counts observed only with ADC (ASH 2021). Two months post-transplant, substantial levels of vector-containing granulocytes were observed long-term in 1 of 2 animals receiving 0.3 mg/kg ADC (ZJ62, VCN 0.16) and in both animals receiving 0.4 mg/kg (H635, VCN 0.46 and H96G, VCN 0.21) at a similar range to the two Bu animals (0.44±0.17, n.s.). Robust HbF induction was detected in three ADC animals with substantial VCN as measured by both HbF-positive cells (F-cells 8.5±1.8% vs. 13.7±5.8%, n.s.) and HPLC measurement of total HbF (8.0±2.9% vs. 11.1±5.2%, n.s.). In ZL13 (0.3 mg/kg ADC), lower gene marking (VCN 0.02) was observed, along with lower HbF induction (F-cells 1.0% and HbF amounts 0.9%). Importantly, ADC resulted in minimal off-target toxicity, unlike those associated with Bu (such as infertility, veno-occlusive disease, pulmonary dysfunction, mucositis, or seizures). High diversity of gene-modified cells (Simpson's diversity index >0.99) without any large clone (maximal integration site 0.3-3.7%) was observed by integration site analysis 1-year post-transplant with both ADC (0.3-0.4 mg/kg) and Bu.

Infertility is a major complication of conventional myeloablative conditioning that limits the uptake of curative therapies; therefore, we investigated fertility in CD117-ADC animals. When monitored daily for menstrual bleeding, the menstrual cycles of all 4 female macaques (ZL13, ZJ62, H635, and H96G) were maintained post-ADC conditioning (0.92±0.11 per month) in contrast to our previously transplanted animals following myeloablative Bu (0.23±0.11, n=4, p<0.01) and total body irradiation (0.33±0.3, n=6, p<0.05). To confirm fertility, all transplanted animals post-ADC including females (n=4) and males (n=2) were housed for mating at a rhesus breeding colony. Two ADC females (ZL13 and H635) gave birth, and an ADC male (12U032) was paired with a pregnant female. Importantly, an ADC female (H635) was paired with an ADC male (13U047).

In summary, a single-injection CD117-ADC allows for efficient engraftment of gene-modified CD34+ HSCs in a rhesus gene therapy model, achieving similar levels as the current clinical standard of care, myeloablative Bu conditioning. Robust HbF induction was also confirmed at the protein level in our rhesus gene therapy model with ADC conditioning. The transplanted animals with ADC maintained their fertility. Overall, these data indicate that an ADC-based targeted approach offers safer conditioning and could improve the risk-benefit profile in HSC gene therapy.