Novel Immune Response Evasion Strategy to Redose Adeno-associated Viral Vectors and Prolong Survival in Surfactant Protein-B-Deficient Mice

Abstract

Surfactant protein-B (SP-B) deficiency is a lethal neonatal respiratory disease with few therapeutic options. Gene therapy using adeno-associated viruses (AAVs) to deliver human SFTPB cDNA (AAV-hSPB) can improve survival in a mouse model of SP-B deficiency. However, the effect of this gene therapy wanes. Gene therapy efficacy could be prolonged if AAV vectors could be redosed, but readministering vectors is hindered by an immune response that includes TLR9 (Toll-like receptor 9) recognition of unmethylated CpG DNA motifs in the AAV genome. One strategy to mitigate TLR9 recognition of AAV is to incorporate decoy nucleotide sequences within the AAV genome. In this work we examined if an AAV containing these TLR9 inhibitory oligonucleotide sequences (AAV-hSPB_TLR9i) could mitigate the immune response sufficiently to redose AAV in the lungs and prolong the survival of SP-B-deficient mice. Indeed, AAV-hSPB_TLR9i could be redosed multiple times, which significantly improved survival in our mouse model. This was partially a result of long-term increased SFTPB RNA and SP-B protein expression. Conversely, redosing AAV-hSPB resulted in the rapid death of SP-B-deficient mice after the second AAV dose. TLR9 inhibition enabled readministration by avoiding the broad stimulation of genes belonging to multiple pathways in the host immune and inflammatory responses, including components of the IFN pathways. Thus, redosing of AAV vectors in the lungs using TLR9 inhibitory sequences is a promising strategy for prolonging gene therapy efficacy, with a proof of concept for AAV readministration in a clinically relevant mouse model of SP-B deficiency.

Keywords: adeno-associated virus; immune evasion; pulmonary atelectasis; pulmonary surfactant; somatic gene therapy.

Figure 1.

Incorporation of TLR9 inhibitory sequences permits adeno-associated virus (AAV) readministration. (A) Schematic of the recombinant AAV2 vector genome containing human (cDNA) SFTPB pseudotyped with the AAV6.2FF capsid coat (AAV-hSPB). (B) Schematic of AAV-hSPB incorporating TLR9 io sequences 5′ and 3′ to the transgene (AAV-hSPB_TLR9i). (C) Study design to assess median survival and changes to body weight after delivery of 10¹¹ vector genomes (vg) per mouse of AAV-hSPB or AAV-hSPB_TLR9i to inducible Sftpb^−/− mice. Control mice were untreated and maintained on doxycycline (Dox) or removed from a Dox diet. Mice treated with AAV (single and repeat doses) were maintained on Dox (0.625 g/kg Dox hyclate; Teklad) for seven days after AAV administration. On D7, the Dox diet was replaced with a regular chow diet. In the readministration groups (repeat), 10¹¹ vg per mouse of AAV-hSPB or AAV-hSPB_TLR9i was administered every 90 days. (D) Kaplan-Meier survival curves (all survival curve P values are calculated from pairwise comparisons using the two-tailed log-rank, Mantel-Cox test with Bonferroni correction; significance is set at P = 0.0033 for six groups and 15 comparisons). (E) Weekly percentage change in body weight over the course of the survival study. D = Day; io = inhibitory oligonucleotide; TLR9 = Toll-like receptor 9.

Figure 2.

Detection of SFTPB expression three hours after the delivery of AAV-hSPB or AAV-hSPB_TLR9i. (A) Schematic of the study design. Lung tissue from Sftpb^−/− mice was snap frozen three hours after the administration of 1 × PBS (n = 3), 10¹¹ vg per mouse of AAV-hSPB (n = 4), or 10¹¹ vg per mouse of AAV-hSPB_TLR9i (n = 4). All mice remained on a doxycycline diet throughout the duration of the experiment (three hours). RNA was isolated from lung tissue, and relative gene expression analyses was performed using qRT-PCR and nCounter using the Mouse Host Response Panel CodeSet. (B) TaqMan qRT-PCR of murine Sftpb expression relative to 1 × PBS–treated mice (P values are calculated using an ordinary one-way ANOVA with Dunnett’s multiple-comparisons post hoc test; *P < 0.05). (C) Human SFTPB expression relative to 1 × PBS–treated mice. ns = not significant.

Figure 3.

Long-term transgene expression and lung structure and function after AAV-hSPB or AAV-hSPB_TLR9i treatment (Tx). (A) Study design to assess lung structure and function after single and repeat administrations of AAV-hSPB or AAV-hSPB_TLR9i in Sftpb^−/− mice. Control mice were untreated and maintained on Dox or removed from Dox 3 days before harvest. In the readministration groups, AAV-hSPB or AAV-hSPB_TLR9i was delivered 90 days after the initial dose. All mice treated with AAV were maintained on Dox for 7 days after vector administration and then placed on regular chow diet for 90 days before lung harvest. (B) Kaplan-Meier survival curves. Dox and Dox* refer to the time points when Dox was removed from all AAV-treated and untreated mice, respectively. (C) Percentage body weight changes for the duration of this experiment. (D) qRT-PCR (TaqMan) of murine Sftpb expression relative to 1 × PBS–treated mice (P values are calculated using an ordinary one-way ANOVA with Dunnett’s multiple-comparisons post hoc test; *P < 0.05, ***P = 0.0003, and ****P < 0.0001). (E) Human SFTPB expression relative to 1 × PBS–treated mice. (F) Representative IB of proSP-B protein expression in total lung homogenate from mice in each of the six Tx groups. Vinculin was used as the loading control. (G) Quantification of SP-B protein expression from five mice in each of the six Tx groups normalized to vinculin (P values are calculated using an ordinary one-way ANOVA with Dunnett’s multiple-comparisons post hoc test; *P < 0.05). H&E = hematoxylin and eosin; WGJ = Wright-Giemsa Jenner.

Figure 3.

Long-term transgene expression and lung structure and function after AAV-hSPB or AAV-hSPB_TLR9i treatment (Tx). (A) Study design to assess lung structure and function after single and repeat administrations of AAV-hSPB or AAV-hSPB_TLR9i in Sftpb^−/− mice. Control mice were untreated and maintained on Dox or removed from Dox 3 days before harvest. In the readministration groups, AAV-hSPB or AAV-hSPB_TLR9i was delivered 90 days after the initial dose. All mice treated with AAV were maintained on Dox for 7 days after vector administration and then placed on regular chow diet for 90 days before lung harvest. (B) Kaplan-Meier survival curves. Dox and Dox* refer to the time points when Dox was removed from all AAV-treated and untreated mice, respectively. (C) Percentage body weight changes for the duration of this experiment. (D) qRT-PCR (TaqMan) of murine Sftpb expression relative to 1 × PBS–treated mice (P values are calculated using an ordinary one-way ANOVA with Dunnett’s multiple-comparisons post hoc test; *P < 0.05, ***P = 0.0003, and ****P < 0.0001). (E) Human SFTPB expression relative to 1 × PBS–treated mice. (F) Representative IB of proSP-B protein expression in total lung homogenate from mice in each of the six Tx groups. Vinculin was used as the loading control. (G) Quantification of SP-B protein expression from five mice in each of the six Tx groups normalized to vinculin (P values are calculated using an ordinary one-way ANOVA with Dunnett’s multiple-comparisons post hoc test; *P < 0.05). H&E = hematoxylin and eosin; WGJ = Wright-Giemsa Jenner.

Figure 4.

TLR9 inhibitory sequences broadly mitigate host immune and inflammatory response genes. (A) Study design to assess the host murine immune and inflammatory response by analyzing gene expression changes after single and repeat administrations of 10¹¹ vg per mouse of AAV-hSPB or AAV-hSPB_TLR9i in Sftpb^−/− mice. (B) Volcano plot and bar graph of gene expression changes (from ROSALIND) in no-treatment off-Dox (n = 4) mice relative to no-treatment on-Dox (n = 4) mice. The off-Dox group had its Dox diet removed for three days before lung harvest. All volcano plots display P_adj using the Benjamini-Hochberg method for controlling false discovery rate. (C) Gene expression changes in mice treated with a single dose of 10¹¹ vg per mouse of AAV-hSPB (n = 4). (D) Gene expression changes in mice treated with a single dose of 10¹¹ vg per mouse of AAV-hSPB_TLR9i (n = 4). (E) Gene expression changes in mice treated with a repeat dose of 10¹¹ vg per mouse of AAV-hSPB (n = 4). (F) Gene expression changes in mice treated with a repeat dose of 10¹¹ vg per mouse of AAV-hSPB_TLR9i (n = 4). (G) Heatmap of undirected global significance scores (from nSolver) in each treatment group relative to mice on Dox. As indicated by the color key, dark purple indicates less and light yellow indicates more gene expression changes in pathways relative to untreated mice on doxycycline. (H) Heatmap of directed global significance scores. As indicated by the color key, dark purple indicates decreased and tan indicates increased gene expression in pathways relative to untreated mice. NK = natural killer; NO = nitric oxide; P_adj = false discovery rate–adjusted P value.

Figure 4.

TLR9 inhibitory sequences broadly mitigate host immune and inflammatory response genes. (A) Study design to assess the host murine immune and inflammatory response by analyzing gene expression changes after single and repeat administrations of 10¹¹ vg per mouse of AAV-hSPB or AAV-hSPB_TLR9i in Sftpb^−/− mice. (B) Volcano plot and bar graph of gene expression changes (from ROSALIND) in no-treatment off-Dox (n = 4) mice relative to no-treatment on-Dox (n = 4) mice. The off-Dox group had its Dox diet removed for three days before lung harvest. All volcano plots display P_adj using the Benjamini-Hochberg method for controlling false discovery rate. (C) Gene expression changes in mice treated with a single dose of 10¹¹ vg per mouse of AAV-hSPB (n = 4). (D) Gene expression changes in mice treated with a single dose of 10¹¹ vg per mouse of AAV-hSPB_TLR9i (n = 4). (E) Gene expression changes in mice treated with a repeat dose of 10¹¹ vg per mouse of AAV-hSPB (n = 4). (F) Gene expression changes in mice treated with a repeat dose of 10¹¹ vg per mouse of AAV-hSPB_TLR9i (n = 4). (G) Heatmap of undirected global significance scores (from nSolver) in each treatment group relative to mice on Dox. As indicated by the color key, dark purple indicates less and light yellow indicates more gene expression changes in pathways relative to untreated mice on doxycycline. (H) Heatmap of directed global significance scores. As indicated by the color key, dark purple indicates decreased and tan indicates increased gene expression in pathways relative to untreated mice. NK = natural killer; NO = nitric oxide; P_adj = false discovery rate–adjusted P value.

Figure 5.

TLR9 inhibition has no effects on humoral immunity and does not affect the presence of immune cells in the lungs. (A) B-cell abundance scores (from ROSALIND) from each treatment group on the basis of nCounter cell-type profiling. (B) T-cell abundance scores. (C) Grouped summary graph of the mean serum anti-AAV6.2FF antibody titers with SD. Antibody concentrations were determined using the reciprocal serum dilution method. (D) Representative multiplexed immunohistochemical images of lymphoid-lineage cells including CD4 (T-helper cells) and CD8 (cytotoxic T cells) cells from whole left lung lobes from mice treated with repeat doses of AAV-hSPB or AAV-hSPB_TLR9i. Yellow boxes specify the region of magnification from the original lung image. DAPI counterstaining is in blue. Scale bars, 100 μm. Quantification of each cell type as total cell counts in whole left lung slices with n = 4 per treatment group (P values are calculated using a two-tailed unpaired t test). (E) Representative multiplexed immunohistochemical images of myeloid-lineage cells, including CD11b (neutrophils and natural killer cells), CD86 (M1 macrophages), CD163 (M2 macrophages), and F4/80 (murine macrophages) cells from mice treated with repeat doses of AAV-hSPB or AAV-hSPB_TLR9i. Yellow boxes specify the region of magnification from the original lung image. DAPI counterstaining is in blue. Scale bars, 100 μm. Each cell type was quantified as total cell count in whole left lung slices with n = 4 per treatment group.

Figure 5.

TLR9 inhibition has no effects on humoral immunity and does not affect the presence of immune cells in the lungs. (A) B-cell abundance scores (from ROSALIND) from each treatment group on the basis of nCounter cell-type profiling. (B) T-cell abundance scores. (C) Grouped summary graph of the mean serum anti-AAV6.2FF antibody titers with SD. Antibody concentrations were determined using the reciprocal serum dilution method. (D) Representative multiplexed immunohistochemical images of lymphoid-lineage cells including CD4 (T-helper cells) and CD8 (cytotoxic T cells) cells from whole left lung lobes from mice treated with repeat doses of AAV-hSPB or AAV-hSPB_TLR9i. Yellow boxes specify the region of magnification from the original lung image. DAPI counterstaining is in blue. Scale bars, 100 μm. Quantification of each cell type as total cell counts in whole left lung slices with n = 4 per treatment group (P values are calculated using a two-tailed unpaired t test). (E) Representative multiplexed immunohistochemical images of myeloid-lineage cells, including CD11b (neutrophils and natural killer cells), CD86 (M1 macrophages), CD163 (M2 macrophages), and F4/80 (murine macrophages) cells from mice treated with repeat doses of AAV-hSPB or AAV-hSPB_TLR9i. Yellow boxes specify the region of magnification from the original lung image. DAPI counterstaining is in blue. Scale bars, 100 μm. Each cell type was quantified as total cell count in whole left lung slices with n = 4 per treatment group.

Figure 5.

TLR9 inhibition has no effects on humoral immunity and does not affect the presence of immune cells in the lungs. (A) B-cell abundance scores (from ROSALIND) from each treatment group on the basis of nCounter cell-type profiling. (B) T-cell abundance scores. (C) Grouped summary graph of the mean serum anti-AAV6.2FF antibody titers with SD. Antibody concentrations were determined using the reciprocal serum dilution method. (D) Representative multiplexed immunohistochemical images of lymphoid-lineage cells including CD4 (T-helper cells) and CD8 (cytotoxic T cells) cells from whole left lung lobes from mice treated with repeat doses of AAV-hSPB or AAV-hSPB_TLR9i. Yellow boxes specify the region of magnification from the original lung image. DAPI counterstaining is in blue. Scale bars, 100 μm. Quantification of each cell type as total cell counts in whole left lung slices with n = 4 per treatment group (P values are calculated using a two-tailed unpaired t test). (E) Representative multiplexed immunohistochemical images of myeloid-lineage cells, including CD11b (neutrophils and natural killer cells), CD86 (M1 macrophages), CD163 (M2 macrophages), and F4/80 (murine macrophages) cells from mice treated with repeat doses of AAV-hSPB or AAV-hSPB_TLR9i. Yellow boxes specify the region of magnification from the original lung image. DAPI counterstaining is in blue. Scale bars, 100 μm. Each cell type was quantified as total cell count in whole left lung slices with n = 4 per treatment group.

Figure 6.

Readministration of AAV-fLuc_TLR9i enhances luciferase expression. (A) Schematic of AAV–firefly luciferase (fLuc) containing TLR9 io 5′ and 3′ to the transgene (AAV-fLuc_TLR9i). (B) Study design to assess transgene (firefly luciferase) expression. In the readministration group, AAV-fLuc_TLR9i was delivered every 90 days for a total of three doses. In Vivo Imaging System (IVIS) was performed 7 days before the first AAV dose (−7), and at Days 7, 21, 88, 97, 111, 139, 173, 187, 208, and 229 after AAV administration. Serum was collected at Day 217, and the lungs were harvested at Day 230. (C) Mean percentage change in body weight over the duration of the experiment. (D) Bioluminescence in either untreated mice or mice administered single or repeat doses of 10¹¹ vg per mouse of AAV-fLuc_TLR9i. (E) Quantification of the total flux as photons per second from the IVIS images in the thorax (lung) region at Days 7, 97, and 187. Data are presented as mean with SD (all P values are calculated using an ordinary one-way ANOVA with a post hoc Tukey multiple-comparisons test; *P = 0.0332 and **P = 0.0021).