Novel Combinatorial MicroRNA-Binding Sites in AAV Vectors Synergistically Diminish Antigen Presentation and Transgene Immunity for Efficient and Stable Transduction

Abstract

Recombinant adeno-associated virus (rAAV) platforms hold promise for in vivo gene therapy but are undermined by the undesirable transduction of antigen presenting cells (APCs), which in turn can trigger host immunity towards rAAV-expressed transgene products. In light of recent adverse events in patients receiving high systemic AAV vector doses that were speculated to be related to host immune responses, development of strategies to mute innate and adaptive immunity is imperative. The use of miRNA binding sites (miR-BSs) to confer endogenous miRNA-mediated regulation to detarget transgene expression from APCs has shown promise for reducing transgene immunity. Studies have shown that designing miR-142BSs into rAAV1 vectors were able to repress costimulatory signals in dendritic cells (DCs), blunt the cytotoxic T cell response, and attenuate clearance of transduced muscle cells in mice to allow sustained transgene expression in myofibers with negligible anti-transgene IgG production. In this study, we screened individual and combinatorial miR-BS designs against 26 miRNAs that are abundantly expressed in APCs, but not in skeletal muscle. The highly immunogenic ovalbumin (OVA) transgene was used as a proxy for foreign antigens. In vitro screening in myoblasts, mouse DCs, and macrophages revealed that the combination of miR-142BS and miR-652-5pBS strongly mutes transgene expression in APCs but maintains high myoblast and myocyte expression. Importantly, rAAV1 vectors carrying this novel miR-142/652-5pBS cassette achieve higher transgene levels following intramuscular injections in mice than previous detargeting designs. The cassette strongly inhibits cytotoxic CTL activation and suppresses the Th17 response in vivo. Our approach, thus, advances the efficiency of miRNA-mediated detargeting to achieve synergistic reduction of transgene-specific immune responses and the development of safe and efficient delivery vehicles for gene therapy.

Keywords: adeno-associated virus vectors; antigen presenting cells; gene therapy; miR-142; miR-223-3p; miR-652-5p; miR-BS; microRNA.

Figure 1

Selection and in vitro screening of candidate miRNA binding sites (miR-BS). (A) Schematic illustration of rAAV.OVA expression vectors. The expression of the OVA transgene is driven by the CB6 promoter. Two copies of miR-BSs are cloned between the transgene and the rabbit β-globin poly A (RBG pA). The OVA expression cassette is flanked by inverted terminal repeats (ITRs) at both ends. (B) An ideal miR-BS candidate, upon delivery to myocytes and myoblasts, which do not express the corresponding miRNA, is expected to undergo transcription and translation to produce high levels of transgene. In contrast, when this vector is delivered to miRNA-expressing immune cells, like macrophages and DCs, miRNA binding to the transgene mRNA leads to translational inhibition and transcript degradation, resulting in minimal transgene protein production. (C) Summary of the in vitro screening of individual miR-BS candidates represented as a heat map with OVA expression denoted as fold change (log₂) with respect to expression vectors lacking any miR-BS. (D) The in vitro screening of miR-BS combinations summarized as a heat map with relative OVA levels represented as fold change (log₂) (n = 5). C2C12, mouse myoblasts; dC2C12, mouse myocytes; RAW264.7, mouse macrophages; JAWSII, mouse DCs.

Figure 2

Incorporation of miR-223BSs and miR-652BSs boosts in vivo OVA production and suppresses antibody development. (A) Endogenous miRNA expression levels in cultured mouse myoblasts (C2C12), myocytes (dC2C12), macrophages (RAW264.7), DCs (JAWSII), bone marrow derived macrophages (BMDM), primary mouse hepatocytes, and Kupffer cells as quantified by reverse transcription quantitative PCR (RT-qPCR) (n = 3). rAAV1 expression vectors were injected by i.m. on day 0 followed by serum collection every week for an eight-week period. (B, C) ELISA quantification of circulating OVA expression (B) and anti-OVA IgG1 (C) (1 × 10¹¹ GCs/mouse, n = 10). Single gradient heat map representing respective analyte levels (n = 5). (D–F) ddPCR detection of rAAV vector genome copies in injected skeletal muscle (D), spleen (E), and liver (F) at eight weeks post-injection (n = 5). Values represent mean ±SD. *p < 0.05, ***p < 0.001, one-way ANOVA with Tukey’s post hoc test.

Figure 3

Incorporation of miR-142BS, miR-223BS, and miR-652BS into rAAV1.OVA transgene vectors reduces macrophage and DC activation. rAAV1.OVA expressing vectors with or without miR-BSs (1 × 10¹¹ GCs/mouse) were delivered by i.m. injections into C57BL/6 mice. Four weeks after injection, cells were isolated from TAs, lymph nodes, spleens, and livers and stained for macrophage markers (CD11b+, F4/80+) and activated DCs (CD11c+, CD80+, CD86+) followed by flow cytometry analysis. Relative frequencies of macrophage populations (A–D) and activated DCs (E–H) are represented as box plots with means, first and third quartile boundaries, and whiskers indicating max and min values. (n = 5). Mock = AAV1.empty capsid. p values were determined by one-way ANOVA with Tukey’s post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 4

miR-142BS-, miR-223BS-, and miR-652-mediated detargeting suppresses OVA specific CD8 T cell response. C57BL/6 male mice, six weeks old, were i.m. injected with mock, empty capsid, or rAAV1.OVA vectors with or without miR-BSs (1 × 10¹¹ GCs/mouse). Mice were sacrificed four weeks after treatment and cells were isolated from TAs, lymph nodes, spleens and livers. Cells were then stained for CD8 T cell markers or with anti-CD8a/H-2Kb SIINFEKL tetramer and quantified by flow cytometry (n = 5). Relative frequencies of CD8+ T cells (A–D) and OVA-specific CD8+ T cells (E–H) are depicted as box plots with means, first and third quartile boundaries, and whiskers indicating max and min values (n = 5). p values were estimated by one-way ANOVA with Tukey’s post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 5

miR-BS incorporation diminishes transgene-specific Th1 and Th17 inflammatory responses. (A, B) Estimation of IFN-γ and TNF-α response to OVA stimulation (5 μg/mL) by splenocytes isolated from mice four weeks after vector injection. Three days after treatment, supernatants were collected and quantitated by ELISA (mean ± SD, n = 5). p values were determined by ANOVA with Tukey’s post hoc test. Splenocytes isolated two and four weeks post-AAV1 injection were stimulated for 24 hours with OVA (5 μg/mL) and stained for detecting either Th1 population (CD4+) (C, D) or Th17 cell population (CD4+IL-17A+) (E, F) and analyzed by flow cytometry (Box plots with means, first and third quartile boundaries, and whiskers indicating max and min values; n = 5). (E, F) Levels of secreted Th17-specific cytokines, IL-17 and IL-21- were assessed after stimulation of splenocytes (harvested four weeks post-injection) for 72 hours with OVA (5 μg/mL) by ELISA (mean ± SD, n = 5). *p < 0.05, **p < 0.01, ***p < 0.001. p values were estimated by one-way ANOVA with Tukey’s post hoc test.

Figure 6

miRNA-mediated detargeting acts independent of Treg immunosuppression and reduces tissue clearance by downregulation of OVA-specific CTL response. (A–C) Cells from TA muscle, lymph nodes and spleen were harvested from rAAV1 injected C57BL/6 mice four weeks after treatment and stained for Treg markers (CD4+, CD25+, FOXP3+). The frequencies of Tregs were quantified by flow cytometry and displayed as box and whisker plots (n = 5). (D) C57BL/6 male mice, six weeks old, were i.m. injected with rAAV1.OVA vectors with or without miRBS (1 × 10¹¹ GCs/mouse, n = 5) and sacrificed two weeks after injection to harvest muscles. Tissue sections were stained for H&E (upper panels, original magnification: 20x), CD8 and granzyme B (lower panels: DAPI, blue; red, granzyme B, red; and CD8, green; original magnification: 40x). Scale bars = 200 μm (H&E images), 50 μm (fluorescence images). (E) Quantification of H&E images for nuclear infiltrates in whole tissue section at original magnification of ×20. (F, G) Quantification of CD8 (F) and granzyme B (G) images in four fields at original magnification of 40x. p values were estimated by one-way ANOVA with Tukey’s post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001.