Mechanisms of AAV neutralization by human alpha-defensins
Data files
Jul 14, 2025 version files 2.44 MB
-
2025_AAVDefensinMechanism_Fig01A_analysis.prism
67.87 KB
-
2025_AAVDefensinMechanism_Fig01B_analysis.prism
60.87 KB
-
2025_AAVDefensinMechanism_Fig01C_analysis.prism
71.19 KB
-
2025_AAVDefensinMechanism_Fig01C_stoichiometry.prism
83.37 KB
-
2025_AAVDefensinMechanism_Fig01C_stoichiometry.xlsx
18.94 KB
-
2025_AAVDefensinMechanism_Fig01D_analysis.prism
68.59 KB
-
2025_AAVDefensinMechanism_Fig01D_stoichiometry.prism
82.24 KB
-
2025_AAVDefensinMechanism_Fig01D_stoichiometry.xlsx
18.96 KB
-
2025_AAVDefensinMechanism_Fig02A_analysis.prism
460.46 KB
-
2025_AAVDefensinMechanism_Fig02B_analysis.prism
360.76 KB
-
2025_AAVDefensinMechanism_Fig02C_analysis.prism
94.58 KB
-
2025_AAVDefensinMechanism_Fig02D_analysis.prism
95.25 KB
-
2025_AAVDefensinMechanism_Fig03B_analysis.prism
57.20 KB
-
2025_AAVDefensinMechanism_Fig03D_analysis.prism
50.15 KB
-
2025_AAVDefensinMechanism_Fig03F_analysis.prism
26.77 KB
-
2025_AAVDefensinMechanism_Fig03H_analysis.prism
23.17 KB
-
2025_AAVDefensinMechanism_Fig03K_analysis.prism
69.30 KB
-
2025_AAVDefensinMechanism_Fig04A_analysis.prism
128.06 KB
-
2025_AAVDefensinMechanism_Fig04B_analysis.prism
110.68 KB
-
2025_AAVDefensinMechanism_Fig04C_analysis.prism
32.85 KB
-
2025_AAVDefensinMechanism_Fig04F_analysis.prism
56.24 KB
-
2025_AAVDefensinMechanism_Fig04G_analysis.prism
56.12 KB
-
2025_AAVDefensinMechanism_Fig05C_analysis.prism
44.32 KB
-
2025_AAVDefensinMechanism_Fig06A_and_Fig06B_analysis.prism
161.28 KB
-
2025_AAVDefensinMechanism_Fig06E_analysis.prism
52.95 KB
-
2025_AAVDefensinMechanism_Fig06F_analysis.prism
69.11 KB
-
README.md
14.56 KB
Abstract
Antiviral immunity compromises the efficacy of adeno-associated virus (AAV) vectors used for gene therapy. This is well understood for the adaptive immune response. However, innate immune effectors like alpha-defensin antimicrobial peptides also block AAV infection, although their mechanisms of action are unknown. To address this gap in knowledge, we investigated AAV2 and AAV6 neutralization by human neutrophil peptide 1 (HNP1), a myeloid alpha-defensin, and human defensin 5 (HD5), an enteric alpha-defensin. We found that both defensins bind to AAV2 and inhibit infection at low micromolar concentrations, similar to our prior studies of AAV6. While HD5 prevents AAV2 and AAV6 from binding to cells, HNP1 does not. However, AAV2 and AAV6 exposed to HD5 after binding to cells are still neutralized, indicating an additional block to infection. Accordingly, both HD5 and HNP1 inhibit externalization of the VP1 unique domain of both AAV2 and AAV6, which contains a phospholipase A2enzyme required for endosome escape and nuclear localization signals required for nuclear entry. Consequently, both defensins prevent AAV from reaching the nucleus. Disruption of intracellular trafficking of the viral genome to the nucleus is reminiscent of how alpha-defensins neutralize other non-enveloped viruses, suggesting a common mechanism of inhibition. These results will inform the development of vectors capable of overcoming these hurdles to improve the efficiency of gene therapy. AAVs are commonly used as gene therapy vectors due to their broad tropism and lack of disease association; however, host innate immune factors, such as human alpha-defensin antimicrobial peptides, can hinder gene delivery. Although it is becoming increasingly evident that human alpha-defensins can block infection by a wide range of nonenveloped viruses, including AAVs, their mechanism of action remains poorly understood. In this study, we describe for the first time how two types of abundant human alpha-defensins neutralize two AAV serotypes, AAV2 and AAV6. We found that one defensin prevents AAV binding to cells, the first step in infection, while both defensins block a critical later step in AAV entry. Our findings support the emerging idea that defensins use a common strategy to block infection by DNA viruses that replicate in the nucleus. Through understanding how innate immune effectors interact with and impede AAV infection, vectors can be developed to bypass these interventions and allow more efficient gene delivery.
Description of the data and file structure
Files are provided for each figure as follows:
Figure 1A: 2025_AAVDefensinMechanism_Fig01A_analysis.prism
Scans of 96-well plates of HeLa cells infected with green fluorescent protein-expressing AAV2 were obtained using a multi-mode plate scanner from 4 independent replicates. The experimental variable was human defensin 5 (HD5). From these images, total well fluorescence was enumerated using Fiji by applying a rolling ball background subtraction and measuring integrated density limited to threshold. Sample wells were normalized to control wells (no HD5), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 1B: 2025_AAVDefensinMechanism_Fig01B_analysis.prism
Scans of 96-well plates of HeLa cells infected with green fluorescent protein-expressing AAV2 were obtained using a multi-mode plate scanner from 3 independent replicates. The experimental variable was human neutrophil peptide 1 (HNP1). From these images, total well fluorescence was enumerated using Fiji by applying a rolling ball background subtraction and measuring integrated density limited to threshold. Sample wells were normalized to control wells (no HNP1), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 1C: 2025_AAVDefensinMechanism_Fig01C_analysis.prism, 2025_AAVDefensinMechanism_Fig01C_stoichiometry.prism, and 2025_AAVDefensinMechanism_Fig01C_stoichiometry.xlsx
Sensorgrams were obtained from a surface plasmon resonance instrument from 2 independent replicates and were visualized in GraphPad Prism 10.4.1 (analysis.prism). AAV2 was coupled to the CM5 sensor chip using amine coupling, and 9451 response units (RUs) were immobilized. The experimental variable was human defensin 5 (HD5). RU from these sensorgrams were obtained for each HD5 concentration at 180 sec and analyzed to determine Bmax, the total number of binding sites (stoichiometry.prism). Bmax values were used to calculate the stoichiometry of HD5 bound to AAV at equilibrium (.xlxs).
Figure 1D: 2025_AAVDefensinMechanism_Fig01D_analysis.prism, 2025_AAVDefensinMechanism_Fig01D_stoichiometry.prism, and 2025_AAVDefensinMechanism_Fig01D_stoichiometry.xlsx
Sensorgrams were obtained from a surface plasmon resonance instrument from 2 independent replicates and were visualized in GraphPad Prism 10.4.1 (analysis.prism). AAV2 was coupled to the CM5 sensor chip using amine coupling, and 9451 response units (RUs) were immobilized. The experimental variable was human neutrophil peptide 1 (HNP1). RU from these sensorgrams were obtained for each HNP1 concentration at 180 sec and analyzed to determine Bmax, the total number of binding sites (stoichiometry.prism). Bmax values were used to calculate the stoichiometry of HNP1 bound to AAV at equilibrium (.xlxs).
Figure 2A: 2025_AAVDefensinMechanism_Fig02A_analysis.prism
The experimental variable was human defensin 5 (HD5). For each of 4 independent replicates, AAV2 bound to HeLa cells at each concentration of HD5 was quantified by qPCR against a standard curve. Samples were normalized to controls (no HD5), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism). Curves from Fig01A and Fig02A were compared using the extra-sum-of-squares F test (.prism).
Figure 2B: 2025_AAVDefensinMechanism_Fig02B_analysis.prism
The experimental variable was human neutrophil peptide 1 (HNP1). For each of 3 independent replicates, AAV2 bound to HeLa cells at each concentration of HNP1 was quantified by qPCR against a standard curve. Samples were normalized to controls (no HNP1), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 2C: 2025_AAVDefensinMechanism_Fig02C_analysis.prism
Scans of 96-well plates of HeLa cells infected with green fluorescent protein-expressing AAV2 were obtained using a multi-mode plate scanner from 4 independent replicates. The experimental variable was human defensin 5 (HD5). From these images, total well fluorescence was enumerated using Fiji by applying a rolling ball background subtraction and measuring integrated density limited to threshold. Sample wells were normalized to control wells (no HD5), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism). Curves from Fig01A and Fig02C were compared using the extra-sum-of-squares F test (.prism).
Figure 2D: 2025_AAVDefensinMechanism_Fig02D_analysis.prism
Scans of 96-well plates of HeLa cells infected with green fluorescent protein-expressing AAV2 were obtained using a multi-mode plate scanner from 4 independent replicates. The experimental variable was human neutrophil peptide 1 (HNP1). From these images, total well fluorescence was enumerated using Fiji by applying a rolling ball background subtraction and measuring integrated density limited to threshold. Sample wells were normalized to control wells (no HNP1), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism). Curves from Fig01B and Fig02D were compared using the extra-sum-of-squares F test (.prism).
Figure 3B: 2025_AAVDefensinMechanism_Fig03B_analysis.prism
Scans of immunoblots of AAV2 immunoprecipitated with the A1 and A20 antibodies and probed sequentially with the B1 antibody and with goat anti-mouse Alexa Fluor 488-conjugated secondary antibody were obtained using a multi-mode plate scanner from 3 independent replicates. Experimental variables were no defensin (ND), human defensin 5 (HD5), and a form of HD5 with aminobutyric acid substituted for the six cysteine residues in the linear sequence of HD5 (ABU). The background subtracted volume of the VP3 band for each sample from each blot was quantified using AzureSpot Pro (v2.1.097) software. The ratios of these volumes for the A1 immunoprecipitated samples divided by the A20 immunoprecipitated samples from all replicates were analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 3D: 2025_AAVDefensinMechanism_Fig03D_analysis.prism
Scans of immunoblots of AAV2 immunoprecipitated with the A1 and A20 antibodies and probed sequentially with the B1 antibody and with goat anti-mouse Alexa Fluor 488-conjugated secondary antibody were obtained using a multi-mode plate scanner from 3 independent replicates. Experimental variables were no defensin (ND) and human neutrophil peptide 1 (HNP1). The background subtracted volume of the VP3 band for each sample from each blot was quantified using AzureSpot Pro (v2.1.097) software. The ratios of these volumes for the A1 immunoprecipitated samples divided by the A20 immunoprecipitated samples from all replicates were analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 3F: 2025_AAVDefensinMechanism_Fig03F_analysis.prism
Scans of immunoblots of AAV2 immunoprecipitated with the A1 and A20 antibodies and probed sequentially with the B1 antibody and with goat anti-mouse Alexa Fluor 488-conjugated secondary antibody were obtained using a multi-mode plate scanner from 3 independent replicates. Experimental variables were no defensin (ND) and human defensin 5 (HD5). The background subtracted volume of the VP3 band for each sample from each blot was quantified using AzureSpot Pro (v2.1.097) software. The ratios of these volumes for the A1 immunoprecipitated samples divided by the A20 immunoprecipitated samples from all replicates were analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 3H: 2025_AAVDefensinMechanism_Fig03H_analysis.prism
Scans of immunoblots of AAV2 immunoprecipitated with the A1 and A20 antibodies and probed sequentially with the B1 antibody and with goat anti-mouse Alexa Fluor 488-conjugated secondary antibody were obtained using a multi-mode plate scanner from 3 independent replicates. Experimental variables were no defensin (ND) and human neutrophil peptide 1 (HNP1). The background subtracted volume of the VP3 band for each sample from each blot was quantified using AzureSpot Pro (v2.1.097) software. The ratios of these volumes for the A1 immunoprecipitated samples divided by the A20 immunoprecipitated samples from all replicates were analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 3K: 2025_AAVDefensinMechanism_Fig03K_analysis.prism
Scans of 96-well plates of HeLa cells infected with green fluorescent protein-expressing AAV2 were obtained using a multi-mode plate scanner from 3 independent replicates. The experimental variables were no defensin (ND), human defensin 5 (HD5), a form of HD5 with aminobutyric acid substituted for the six cysteine residues in the linear sequence of HD5 (ABU), and human neutrophil peptide 1 (HNP1). From these images, total well fluorescence was enumerated using Fiji by applying a rolling ball background subtraction and measuring integrated density limited to threshold. Sample wells were normalized to control wells (ND), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 4A: 2025_AAVDefensinMechanism_Fig04A_analysis.prism
Infection data: Scans of 96-well plates of HeLa cells infected with green fluorescent protein-expressing AAV6 were obtained using a multi-mode plate scanner from 4 independent replicates. The experimental variable was human defensin 5 (HD5). From these images, total well fluorescence was enumerated using Fiji by applying a rolling ball background subtraction and measuring integrated density limited to threshold. Sample wells were normalized to control wells (no HD5), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Binding data: The experimental variable was human defensin 5 (HD5). For each of 4 independent replicates, AAV6 bound to HeLa cells at each concentration of HD5 was quantified by qPCR against a standard curve. Samples were normalized to controls (no HD5), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 4B: 2025_AAVDefensinMechanism_Fig04B_analysis.prism
Infection data: Scans of 96-well plates of HeLa cells infected with green fluorescent protein-expressing AAV6 were obtained using a multi-mode plate scanner from 3 independent replicates. The experimental variable was human neutrophil peptide 1 (HNP1). From these images, total well fluorescence was enumerated using Fiji by applying a rolling ball background subtraction and measuring integrated density limited to threshold. Sample wells were normalized to control wells (no HNP1), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Binding data: The experimental variable was human neutrophil peptide 1 (HNP1). For each of 3 independent replicates, AAV6 bound to HeLa cells at each concentration of HNP1 was quantified by qPCR against a standard curve. Samples were normalized to controls (no HNP1), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 4C: 2025_AAVDefensinMechanism_Fig04C_analysis.prism
The experimental variables were the ADK1a antibody and wheat germ agglutinin (WGA). For each of 6 independent replicates, AAV6 bound to HeLa cells in each condition was quantified by qPCR against a standard curve. Samples were normalized to controls without inhibitor (None), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 4F: 2025_AAVDefensinMechanism_Fig04F_analysis.prism
Scans of immunoblots of AAV6 immunoprecipitated with the A1 and ADK1a antibodies and probed sequentially with the B1 antibody and with goat anti-mouse Alexa Fluor 488-conjugated secondary antibody were obtained using a multi-mode plate scanner from 4 independent replicates. Experimental variables were no defensin (ND), human defensin 5 (HD5), and human neutrophil peptide 1 (HNP1). The background subtracted volume of the VP3 band for each sample from each blot was quantified using AzureSpot Pro (v2.1.097) software. The ratios of these volumes for the A1 immunoprecipitated samples divided by the ADK1a immunoprecipitated samples were calculated. The A1/ADK1a ratios for HD5 and HNP1 samples were normalized to the ratio for the ND control. Data from all replicates were analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 4G: 2025_AAVDefensinMechanism_Fig04G_analysis.prism
Scans of 96-well plates of HeLa cells infected with green fluorescent protein-expressing AAV6 were obtained using a multi-mode plate scanner from 4 independent replicates. The experimental variables were no defensin (ND), human defensin 5 (HD5), and human neutrophil peptide 1 (HNP1). From these images, total well fluorescence was enumerated using Fiji by applying a rolling ball background subtraction and measuring integrated density limited to threshold. Sample wells were normalized to control wells (ND), and the normalized data from all replicates was analyzed using GraphPad Prism 10.4.1 (.prism).
Figure 5C: 2025_AAVDefensinMechanism_Fig05C_analysis.prism
Peak values corresponding to the melting temperature (Tm) of AAV2 were derived from thermograms for 3 independent replicates. The experimental variables were pH, no defensin (ND), human defensin 5 (HD5), and human neutrophil peptide 1 (HNP1). Data from all replicates were analyzed using GraphPad Prism 10.4.1 (.prism).
Figures 6A, 6B, 6E, and 6F: 2025_AAVDefensinMechanism_Fig06A_and_Fig06B_analysis.prism, 2025_AAVDefensinMechanism_Fig06E_analysis.prism, and 2025_AAVDefensinMechanism_Fig06F_analysis.prism
DyLight 488 intensity for HeLa cells infected with DyLight 488-labeled AAV2 was measured from micrographs using CellProfiler. The experimental variables were time post-infection (p.i.), human defensin 5 (HD5), and human neutrophil peptide 1 (HNP1). Control cells were infected in the absence of defensin (No defensin, ND). A minimum of 35 cells were analyzed for each condition for each of 4 independent replicates. Data from all replicates were analyzed using GraphPad Prism 10.4.1 (.prism). For Figures 6A and 6B, the fraction of total DyLight 488 signal that was located in the nucleus was calculated for each cell. For Figure 6E, the ratio of DyLight 488 signal in the nucleus compared to the perinuclear space was calculated for each cell, and the log of this ratio was analyzed. For Figure 6F, the total DyLight signal is graphed for each cell.