The AAV9 receptor and its modification to improve in vivo lung gene transfer in mice

Abstract

Vectors based on adeno-associated virus (AAV) serotype 9 are candidates for in vivo gene delivery to many organs, but the receptor(s) mediating these tropisms have yet to be defined. We evaluated AAV9 uptake by glycans with terminal sialic acids (SAs), a common mode of cellular entry for viruses. We found, however, that AAV9 binding increased when terminal SA was enzymatically removed, suggesting that galactose, which is the most commonly observed penultimate monosaccharide to SA, may mediate AAV9 transduction. This was confirmed in mutant CHO Pro-5 cells deficient in the enzymes involved in glycoprotein biogenesis, as well as lectin interference studies. Binding of AAV9 to glycans with terminal galactose was demonstrated via glycan binding assays. Co-instillation of AAV9 vector with neuraminidase into mouse lung resulted in exposure of terminal galactose on the apical surface of conducting airway epithelial cells, as shown by lectin binding and increased transduction of these cells, demonstrating the possible utility of this vector in lung-directed gene transfer. Increasing the abundance of the receptor on target cells and improving vector efficacy may improve delivery of AAV vectors to their therapeutic targets.

Figure 1. Effect of SA on AAV vector binding.

(A) Schematic of N- and O-linked glycans for each CHO cell line (Pro-5, Lec-2, and Lec-8; ref. 17). Pretreating Pro-5 cells with NA alone or NA and β-gal should produce glycan structures similar to those seen in Lec-2 and Lec-8 cells, respectively. (B–D) SA was removed from the surface of (B) Pro-5, (C) HEK293, and (D) Huh-7 cells by treatment with NA from Vibrio cholerae. AAV vectors (5 × 10⁹ genome copies; MOI, 10⁴) were applied to NA-treated and untreated cells and incubated at 4°C for 1 hour. After extensive washing, total DNA was isolated and bound vector was quantified by qPCR. *P < 0.001.

Figure 2. AAV9 dependence on galactose for binding and transduction of CHO cells.

(A and B) AAV2, AAV6, or AAV9 expressing ffLuc was added to Pro-5, Lec-2, or Lec-8 cells and incubated at 4°C for 1 hour. (A) Total DNA was isolated to determine bound vector genome copies by qPCR or (B) cells were incubated at 37°C for 48 hours and analyzed for ffLuc expression. (C and D) AAV2 and AAV9 were applied to NA-treated Pro-5 cells in the presence of various lectins to compete for AAV binding (C) or transduction (D). RCA was not used in transduction studies because of its toxicity to the cells. (E and F) AAV2 and AAV9 were added to Pro-5 cells that were treated with either NA or both NA and β-gal to assess AAV binding (E) and transduction (F). *P < 0.001. For C and D, statistical significance was determined compared with the no lectin control.

Figure 3. GMA analysis of AAV9 binding.

AAV9 capsids were screened for binding to 465 different glycans based on the average relative fluorescence, with the top 5 glycans that bound AAV9 indicated (number indicates glycan identifier). Error bars represent the SD of glycan binding. The structures of the top 3 glycans with high specificity of binding and their representative illustrations are shown in the bottom panel. The average RFU and %CV for each glycan were as follows (glycan identifier: average RFU, %CV): 415: 633, 12; 297: 590, 19; and 399: 482, 17.

Figure 4. AAV9 transduction of murine conducting airway following pretreatment with NA.

C57BL/6 mice were given an intranasal instillation of 10¹¹ genome copies of AAV9 (A and B) or AAV6 expressing nLacZ (C and D) either 1 hour after intranasal instillation of 100 mU NA or simultaneously with the NA treatment. AAV9 vector administered without NA was used as a control. At day 21 after administration, lungs were harvested and sections stained for β-gal expression. Lung sections were examined at both ×100 magnification (A and C) and ×200 magnification (B and D) for each condition. (E) nLacZ-positive cells in the conducting airway were quantified for each group as the average number of nLacZ-positive cells ± SD per ×200 field of view. *P < 0.001 compared with the no NA control.

Figure 5. Expression of galactose in the cells of the murine conducting airways.

C57BL/6 mice were treated with PBS (A–D) or 100 mU NA (E–H) in a total of 30 μl delivered intranasally. Lungs were inflated and removed 1 hour later, and thin sections (8 μm) were stained with (A, C, E, and G) rhodamine-RCA, (B, D, F, and H) fluorescein-SNA, and (C, D, G, and H) DAPI. Sections were examined by wide-field ×200 magnification (A, B, E, and F) and confocal microscopy (C, D, G, and H). Scale bar: 20 μm. (I–K) Lung sections of mice treated intranasally with NA were stained for (I) galactose expression using rhodamine–RCA lectin and (J) α-tubulin expression to stain cilia. (K) Overlay of galactose and α-tubulin staining. Sections were examined at ×400 magnification.

Figure 6. Correlation of galactose expression and AAV vector transduction in mouse organs.

(A–C) Comparison of RCA staining of (A) muscle, (B) heart, and (C) liver with AAV vector transduction efficiency after IV injection of AAV9.CB.nLacZ at 2 different doses (10¹¹ and 10¹² GC) at day 21. (D) RCA staining of capillaries in brain and costaining with an antibody against CD31 as an endothelial marker. Scale bars: 100 μm.