A myeloid cell-binding adenovirus efficiently targets gene transfer to the lung and escapes liver tropism

Abstract

Specific and efficient gene delivery to the lung has been hampered by liver sequestration of adenovirus serotype 5 (Ad5) vectors. The complexity of Ad5 liver tropism has largely been unraveled, permitting improved efficacy of Ad5 gene delivery. However, Kupffer cell (KC) scavenging and elimination of Ad5 still represent major obstacles to lung gene delivery strategies. KC uptake substantially reduces bioavailability of Ad5 for target tissues and compensatory dose escalation leads to acute hepatotoxicity and a potent innate immune response. Here, we report a novel lung-targeting strategy through redirection of Ad5 binding to the concentrated leukocyte pool within the pulmonary microvasculature. We demonstrate that this leukocyte-binding approach retargets Ad5 specifically to lung endothelial cells and prevents KC uptake and hepatocyte transduction, resulting in 165,000-fold enhanced lung targeting, compared with Ad5. In addition, myeloid cell-specific binding is preserved in single-cell lung suspensions and only Ad.MBP-coated myeloid cells achieved efficient endothelial cell transduction ex vivo. These findings demonstrate that KC sequestration of Ad5 can be prevented through more efficient uptake of virions in target tissues and suggest that endothelial transduction is achieved by leukocyte-mediated 'hand-off' of Ad.

Figure 1. Systemically administered knob-deleted Ad.MBP enhances lung gene transfer and diminishes liver transduction independent of knob-deletion

(a, b) Efficacy of lung targeting was evaluated for Ad.MBP at low viral doses. Mice were injected (i.v.) with 10⁷–10¹⁰ VP of Ad.MBP.Luc and biodistribution of Luc gene transfer (a), and lung:liver and lung:spleen Luc gene expression ratios (b) were determined 24 h post-injection. Data are means ± s.d. for n=3 mice per group. ***P<0.001 for lung compared to all other tissues. (c) Biodistribution of Luc gene transfer in mice injected (i.v.) with 1.5 × 10¹⁰ VP of Ad.MBP.Luc or Ad.iMBP.Luc control 24 h prior to assessment. Data are means ± s.d. for n=3 mice per group for a representative experiment. **P<0.01 compared to Ad.iMBP, ***P<0.001 compared to Ad.iMBP, ^###P<0.001 compared to Ad.MBP Lung.

Figure 2. Systemically administered Ad.MBP enhances lung gene transfer and diminishes liver transduction compared to Ad5

Biodistribution of Luc gene transfer in mice injected (i.v.) with 2.5 × 10¹⁰ VP of Ad.MBP.Luc or Ad5.Luc control 24 h prior to assessment. Data are means ± s.d. for n=3 mice per group for a representative experiment. ***P<0.001 compared to Ad5, ^###P<0.001 compared to Ad.MBP Lung.

Figure 3. Liver sequestration of Ad.MBP virus particles is diminished due to first-pass effect through the lung after systemic administration

(a–c) Biodistribution of VP was assessed to evaluate sequestration and first-pass effect of Ad.MBP. (a) Mice were injected (i.v.) with a 1:1 ratio of Ad5:Ad.MBP (4 × 10¹⁰ total VP) and sacrificed at either 5 or 45 min post-injection. Lungs and livers were removed and genomic DNA was extracted for subsequent PCR-based measure of Ad genome content. (b) Some animals were perfused with PBS via the right ventricle (Perf) immediately before tissue isolation. Non-perfused mice (No Perf). Spiked samples contain either Ad5, or Ad.MBP, or the 1:1 viral cocktail added directly to naïve lung tissue prior to genomic DNA isolation. (c) Mice were also injected directly into the left ventricle (LV) and sacrificed at 5 min post-injection to compare to i.v. injected (TV) mice. Spiked samples in (a) and (c) contain the 1:1 viral cocktail added directly to naïve tissues prior to genomic DNA isolation. Gray and black arrowheads indicate positions for Ad5 and Ad.MBP PCR products, respectively. Data are representative of n=3–5 mice per group.

Figure 4. Ad.MBP targets gene transfer to the pulmonary endothelium

(a–f) Cell populations transduced by Ad.MBP were evaluated by flow cytometry of collagenase digested lungs from mice injected (i.v.) with 2.5 × 10¹⁰ VP of Ad.MBP.GFP 24 h prior. (a) Transduced (GFP⁺) cells from Ad.MBP injected mouse lungs were gated (R1) based on the staining profile of lung cells from uninjected (naive) controls (see Supplementary Figure S2 for Ad5 staining profile). Representative scatter plots are shown. (b) Percent of total lung cells transduced (GFP⁺) by Ad5.GFP or Ad.MBP.GFP, gated as in (a). Data are plotted with means for n=3–7 mice. **P<0.01. (c, d) Analysis of GFP⁺ lung cells gated in (a) for expression of the markers CD11b, CD11c, and Ly6G. Populations were gated as follows: non-myeloid (Non-My; CD11b⁻Ly6G⁻CD11c⁻), non-alveolar macrophages or monocytes (Mon/Mac; CD11b⁺Ly6G⁻CD11c⁻), neutrophils (PMN; CD11b⁺Ly6G⁺), and alveolar macrophages (AM; CD11c⁺CD11b⁻Ly6G⁻). (e, f) Analysis of GFP⁺ lung cells gated in (a) for expression of the markers CD45 and CD31. Populations were gated as follows: total leukocuytes (Leuk; CD45⁺), endothelial cells (EC; CD31⁺), and other stromal cells (Stromal; CD45⁻CD31⁻). Data in (c) and (e) are means ± s.d. for n=5 and n=7 mice, respectively. ***P<0.001. Representative scatter plots are shown in panels (d) and (f). Gating schema were based on lung staining patterns from uninjected controls (see Supplementary Figure S2).

Figure 5. Ad.MBP lung gene transfer is localized to alveolar capillaries

(a–d) Mice were injected (i.v.) with 2.5 × 10¹⁰ VP of Ad.MBP.GFP 24 h prior to sacrifice. Panels show staining for the GFP protein (green fluorescence), von Willebrand factor (vWF; red), and nuclei (DAPI; blue). (a) Uninjected control mice. (b) Mice injected with Ad5.GFP control virus. (c, d) Mice injected with Ad.MBP.GFP virus. Although the majority of GFP expression is limited to the vWF⁻ microvasculature (c, d), white arrowheads demonstrate infrequently transduced ECs from vWF⁺ large vessels (d). Scale bar, 40 μm. Images are representative of n=2–3 mice per group.

Figure 6. ECs are refractive to direct binding by Ad.MBP and myeloid cell-binding is a prerequisite to EC transduction

(a–c) Assessment of Ad5 or Ad.MBP binding at 4 °C to collagenase digested lung cells (500 VP per cell). (a) Percentage of Ad-bound cells within various lung cell populations by flow cytometry. A representative experiment is shown. (b) Evaluation of CD45 (a panleukocyte marker) expression on Ad.MBP-bound cells. (c) Cytospins of Ad.MBP-bound cells sorted by FACS [R1 gate in (b)]. Unbound cells [R2 gate in (b)] were also isolated for comparison. (d, e) Transduction of EC lines with Ad5.GFP or Ad.MBP.GFP at increasing multiplicity of infection (MOI). After 24 h, the percent of MCEC (d), or MiPMVEC (e) transduced were analyzed by flow cytometry. (f) Transduction of MCEC with Ad-bound BMCs. Increasing numbers of BMCs from hCAR⁺ transgenic mice were pre-incubated with 5 × 10⁸ VP of Ad5.GFP or Ad.MBP.GFP and then added to MCEC for ‘spinfection’ for 10 min at 4 °C. Plates were then moved to 37 °C and analyzed by flow cytometry after 24 h. Data in (d–f) are means ± s.d. of representative experiments (n=3) with triplicate samples. *P<0.05, **P<0.01, ***P<0.001.