In vivo HSC transduction in rhesus macaques with an HDAd5/3+ vector targeting desmoglein 2 and transiently overexpressing cxcr4

Abstract

We developed a new in vivo hematopoietic stem cell (HSC) gene therapy approach that involves only IV injections and does not require myeloablation/conditioning and HSC transplantation. In this approach, HSCs are mobilized from the bone marrow into the peripheral bloodstream and transduced with IV injected helper-dependent adenovirus (HDAd) vectors. A fraction of transduced HSCs returns to the bone marrow and persists there long term. Here, we report desmoglein 2 (DSG2) as a new receptor that can be used for in vivo HSC transduction. HDAd5/3+ vectors were developed that use DSG2 as a high-affinity attachment receptor, and in vivo HSC transduction and safety after IV injection of an HDAd5/3+ vector expressing green fluorescent protein (GFP) in granulocyte colony-stimulating factor/AMD3100 (plerixafor)-mobilized rhesus macaques were studied. Unlike previously used CD46-targeting HDAd5/35++ vectors, HDAd5/3+ virions were not sequestered by rhesus erythrocytes and therefore mediated ∼10-fold higher GFP marking rates in primitive HSCs (CD34+/CD45RA-/CD90+ cells) in the bone marrow at day 7 after vector injection. To further increase the return of in vivo transduced, mobilized HSCs to the bone marrow, we transiently expressed cxcr4 in mobilized HSCs from the HDAd5/3+ vector. In vivo transduction with an HDAd5/3+GFP/cxcr4 vector at a low dose of 0.4 × 1012 viral particles/kg resulted in up to 7% of GFP-positive CD34+/CD45RA-/CD90+ cells in the bone marrow. This transduction rate is a solid basis for in vivo base or prime editing in combination with natural or drug-induced expansion of edited HSCs. Furthermore, our study provides new insights into HSC biology and trafficking after mobilization in nonhuman primates.

Graphical abstract

Figure 1

Analysis of DSG2 expression on HSCs. (A) Flow cytometry for CD46 and DSG2 on human CD34⁺ cells from 2 different healthy G-CSF–mobilized donors. (B) Schematic of HDAd vector capsids and genomes. The chimeric capsids contain either affinity-enhanced Ad35 or Ad3 fiber knob domains. Both vector genomes contain an EF1α-mgmt^p140k/GFP expression cassette used before. Both proteins are linked via a self-cleaving picornavirus 2A peptide. One set of vectors has an additional CMV promoter–human cxcr4 expression cassette. The 4 vectors used in this study are HDAd5/35++GFP, HDAd5/3+GFP, HDAd5/35++GFP/cxcr4, and HDAd5/3+GFP/cxcr4. (C) In vitro CD34⁺ cell transduction studies. Cells were infected with HDAd5/35++-GFP and HDAd5/3+-GFP virus at a multiplicity of infection of 2000 vp/cell with and without pre-incubation with recombinant soluble CD46 or DSG2 (10 μg/mL) for 1 hour. GFP expression was analyzed 24 hours' postinfection. (D) CD46 and DSG2 flow cytometry on human and rhesus peripheral RBCs stained with antibodies that recognize the receptors in both species (anti-human CD46 mAb clone M177 from Santa Cruz Biotechnology [Dallas, TX] and anti-human DSG2 polyclonal antibody AF947 from R&D Systems [Minneapolis, MN]). (E) Transduction of 293 cells with HDAd5/35++GFP and HDAd5/3+GFP in the absence and presence of human (huRBCs) and rhesus (rhRBCs) RBCs (details are provided in the Materials and methods). Shown is the percentage of GFP-positive cells measured 2 days after transduction.

Figure 2

In vivo HSC transduction studies in rhesus macaques with HDAd5/35++GFP, HDAd5/3+GFP, HDAd5/35++GFP/cxcr4, and HDAd5/3+GFP/cxcr4. (A) Schematic of the experiment showing the injection time of the mobilization drugs (G-CSF and AMD3100) and cytokine prophylaxis drugs (dexamethasone, anakinra, and tocilizumab), as well as the injection time of the HDAd vector at a dose of 0.4 × 10¹² vp/kg. The time point of HDAd injection is taken as “day 0, 0 hours.” HDAd5/35++GFP– and HDAd5/3+GFP–injected animals were monitored for 7 days. HDAd5/35++GFP/cxcr4– and HDAd5/3+GFP/cxcr4–injected animals were observed for 8 weeks to capture late effects of cxcr4 expression. Bone marrow aspirates (BMAs) were performed at days 3, 7, and (for HDAd5/35++GFP/cxcr4 and HDAd5/3+GFP/cxcr4) at weeks 4 and 8. Blood samples were taken at 0 hours (before HDAd injection) and 2, 6, 9, and 12 hours on day 0; on days 1, 2, 3, and 7; and then weekly. (B) HSC mobilization. Shown are numbers of primitive (CD34⁺/CD45RA^–/CD90⁺) HSCs in peripheral blood. HDAd vectors were IV injected at day 0, 0 hours. Notably, there was no correlation between the weight of the animals and the efficacy of mobilization. The weights were 9.70 kg (HDAd5/35++GFP), 6.64 kg (HDAd5/3+GFP), 7.13 kg (HDAd5/35++GFP/cxcr4), and 7.19 kg (HDAd5/3+GFP/cxcr4). (C) Vector clearance from blood. HDAd vector genome copies in serum samples were measured by using quantitative polymerase chain reaction.

Figure 3

Analysis of transduction in peripheral blood cells. (A) VCN per cell in PBMCs (black lines) and total blood cells (RBCs+PBMCs, red lines, measured only for HDAd5/35++GFP/cxcr4– and HDAd5/3+GFP/cxcr4–injected animals). (B) Analysis of transgene mRNA levels relative to mRNA levels of the housekeeping gene GAPDH. Human mgmt^P140K mRNA data are shown by black lines; human cxcr4 mRNA levels are shown by red lines. Note that the scale of the y-axis for the HDAd5/3+GFP/cxcr4–injected animal is different. (C) Percentage of GFP⁺ HSCs in peripheral blood white cells. The scale of the y-axis for the HDAd5/3+GFP/cxcr4–injected animal is different.

Figure 4

GFP expression in bone marrow (BM) HSCs and MNCs. (A) Return of mobilized HSCs to the BM. Shown is the percentage of primitive (CD34⁺/CD45RA^–/CD90⁺) HSCs in total BM MNCs measured in BM samples at days 3 and 7. (“Day 0” is the day of HDAd injection.) (B) In vivo transduced primitive HSCs that returned to the BM. Shown is the percentage of GFP-positive CD34⁺/CD45RA^–/CD90⁺ HSCs in BM aspirates collected at days 3 and 7. (C) Shown are the data from panel B as fold increase, taking transduction with the HDAd5/35++GFP vector as “1.” (D) Percentage of GFP-positive CD34⁺/CD45RA^–/CD90⁺ HSCs relative to total BM MNCs. (E) Data extracted from panel D showing the effect of cxcr4 expression on the GFP⁺ HSCs in the BM (comparison of HDAd5/35++GFP vs HDAd5/35++GFP/cxcr4 and HDAd5/3+GFP vs HDAd5/3+GFP/cxcr4). (F) Representative GFP immunofluorescence on a BM CD34⁺ cytospin of the HDAd5/3+GFP/cxcr4–injected animal. CD34⁺ cells were isolated from a BM aspirate (after RBC lysis) by human CD34 magnetic-activated cell sorting, spun on a glass slide, and then stained with a GFP–fluorescein isothiocyanate antibody. GFP signals are green. Nuclei were stained with 4′,6-diamidino-2-phenylindole and appear in blue. Scale bar = 50 μm.

Figure 5

Safety of IV HDAd injection into mobilized rhesus macaques. (A) Serum IL-6 levels measured by cytometric bead array. HDAd injections (day 0, 0 hours) are indicated by red arrows. (B) Serum TNFα levels. (C) Selected hematologic parameters. The normal range is indicated by the gray shading. THOU, thousands.

Figure 6

Percentage of GFP-positive cells in lineage–positive cells. Percentage of GFP⁺ CD3⁺, CD14⁺, and CD20⁺ lineage cells in PBMCs (A) and bone marrow MNCs (B).