Harnessing BET-Bromodomain Assisted Nuclear Import for Targeted Subcellular Localization and Enhanced Efficacy of Antisense Oligonucleotides

Abstract

Antisense oligonucleotides (ASOs) are a promising class of therapeutics designed to modulate gene expression. Both key mechanisms of action for ASOs operate in the nucleus: splice-switching ASOs modify pre-mRNA, processed in the nucleus, and mRNA-degrading ASOs require RNase H, an enzyme predominantly active in the nucleus. Therefore, to achieve maximal therapeutic efficacy, ASOs require efficient nuclear delivery. In this work, we have synthesized ASO conjugates for active nuclear import, by covalent conjugation with a potent and proven small-molecule nuclear importer, (+)-JQ1. (+)-JQ1 is a well-characterized high-affinity binder for members of the BET bromodomain family of proteins and was recently shown to transport cytoplasmic proteins into the nucleus. Our (+)-JQ1-ASO conjugates outperformed their unmodified counterparts for both splice-switching and mRNA knockdown in the nucleus, across different molecular targets, backbone chemistries, and cell lines. In addition, we show that the improvement in on-target efficacy correlates with increased nuclear localization of the (+)-JQ1-modified ASOs by subcellular fractionation and immunocytochemistry. Notably, we improved the performance of Oblimersen, a BCL-2 ASO drug that failed in phase-III clinical trials. (+)-JQ1-Oblimersen showed increased effectiveness in an acute myeloid leukemia cell model, showing that this therapeutic may merit re-evaluation. This work demonstrates that the covalent modification of ASOs with a small-molecule nuclear importer can significantly improve target engagement and pave the way for more effective therapeutics.

1

Schematic demonstrating the improved activity of (+)-JQ1-ASO conjugates over unconjugated ASOs. (+)-JQ1-ASO conjugates improve splice-modulation and RNase H-mediated knockdown through increased nuclear accumulation.

2

Covalent (+)-JQ1-SSO modification enhances slice-switching activity. (a) Sequence and chemical modifications for SSO used in the HeLa pLuc 705 cell line. (b) Splice-switching mechanism in HeLa pLuc 705 cells. (c) Synthesis of (+)-JQ1-SSO conjugate using copper-catalyzed click chemistry. (d) Luminescence values for SSO and (+)-JQ1-SSO activity, transfected with Lipofectamine 2000, at 24 h at concentrations indicated. In all cases, luciferase activity was measured and normalized to untreated cells. (e) Luminescence values for SSO and (+)-JQ1-SSO activity, using gymnotic delivery, at 96 h at concentrations indicated. In all cases, luciferase activity was measured and normalized to untreated cells. Competition assay between (+)-JQ1-ASO conjugate and excess small molecule, (+)-JQ1 at (f) 5 nM and (g) 200 nM. Three biological replicates are shown as diamonds for each condition (each from three technical replicates). The vertical bars represent the mean and the error bars the standard deviation. ** represents p < 0.05, *** represents p < 0.01, n.s. represents p values that are not significant.

3

Covalent-(+)-JQ1 modification of an ASO enhances RNase H-mediated knockdown. (a) Sequence and chemical modifications for the MALAT1 gapmers used. (b) Mechanism of RNase H-mediated degradation of the lncRNA MALAT1 by an ASO, localized in the nucleus. (c) RT-qPCR data for MALAT1 knockdown upon Lipofectamine transfection of (+)-JQ1- and unconjugated-gapmer in HEK293T cells for 24 h at concentrations indicated. (d) RT-qPCR data for MALAT1 knockdown upon Lipofectamine transfection of (+)-JQ1- and unconjugated-gapmer in HEK293T cells at 200 nM for 6, 12, and 24 h. (e) RT-qPCR data for MALAT1 knockdown upon gymnosis of (+)-JQ1- and unconjugated-gapmer in HEK293T cells for 96 h at concentrations indicated. (f) MALAT1 knockdown observed in competition assay in the presence of 200 nM (+)-JQ1. For all RT-qPCR, three biological replicates are shown as diamonds for each condition (each from three technical replicates). The vertical bars represent the mean and the error bars the standard deviation. ** represents p < 0.05, *** represents p < 0.01, n.s. represents p values that are not significant. (g) Representative immunocyochemistry of HEK293 cells transfected with 200 nM of unconjugated and (+)-JQ1-modified MALAT1 gapmers indicated for 24 h, using antibodies against the PS modifications (green) and α/β-tubulin (red). Arrows indicate ASO-containing puncta. Images are maximum intensity projections generated from Z-stacks; magnification 63×, scale bars as indicated. (h) Quantification of the fluorescent signal ratio between the PS immunopositive signal (AF488) within Hoechst-stained nuclei from HEK293 cells treated with the ASO concentrations as shown. Biological replicates represent random fields of view per condition. The vertical bars represent the mean and the error bars the standard deviation. *** represents p < 0.01.

4

(+)-JQ1-G3139 outperformed the unconjugated G3139, a late-stage clinical ASO. (a) Sequence and chemical modifications for Oblimersen (G3139) used. (b) Mechanism of RNase H-mediated degradation of BCL-2 mRNA by a gapmer ASO. (c) RT-qPCR data of BCL-2 knockdown upon (+)-JQ1-G3139 and unconjugated-G3139 Lipofectamine transfection in HEK293Ts for 24 h at concentrations indicated. (d) RT-qPCR data of BCL-2 knockdown upon (+)-JQ1-G3139 and unconjugated-G3139 Lipofectamine transfection in HEK293Ts at 500 nM at 6, 12, and 24 h. (e) RT-qPCR data of BCL-2 knockdown upon (+)-JQ1-G3139 and unconjugated-G3139 Lipofectamine upon gymnotic delivery in HEK293Ts at 96 h at the concentrations indicated. (f) Western blot of BCL-2 levels upon treatment with G3139 and (+)-JQ1-G3139 upon transfection with Lipofectamine at 24 h at concentrations indicated. Normalized to GAPDH expression levels. (g) Reduction of enhanced BCL-2 knockdown observed in competition assay in the presence of 5 nM and 200 nM (+)-JQ1. (h) Percent Cy3 fluorescence normalized to total fluorescence in nuclear or cytoplasmic fraction upon (+)-JQ1-G3139 and unconjugated-G3139 Lipofectamine transfection in HEK293Ts for 6, 12, and 24 h at concentrations indicated. For RT-qPCR, three biological replicates are shown as diamonds for each condition (each from three technical replicates). The vertical bars represent the mean and the error bars the standard deviation. ** represents p < 0.05, *** represents p < 0.01, n.s. represents p values that are not significant. (i) Representative immunocyochemistry of HEK293 cells transfected with 500 nM of unconjugated and (+)-JQ1-modified G3139 for 24 h using antibodies against the PS modifications (green) and a/β-tubulin (red). Arrows indicate ASO-containing puncta. Images are maximum intensity projections generated from Z-stacks; magnification 63×, scale bars as indicated. (j) Quantification of the fluorescent signal ratio between the PS immunopositive signal (AF488) within Hoechst-stained nuclei from HEK293 cells treated with the ASO concentrations as shown. Biological replicates represent random fields of view per condition. The vertical bars represent the mean and the error bars the standard deviation. ** represents p < 0.05, *** represents p < 0.01.

5

(+)-JQ1-G3139 shows enhanced knockdown and chemosensitization in THP-1 cells, a clinically relevant model for AML. (a) Schematic for G3139 mechanism of action in the THP-1 AML cell line. (b) RT-qPCR data of BCL-2 knockdown upon (+)-JQ1-G3139 and unconjugated-G3139 upon electroporation for 48 h at concentrations indicated. (c) Reduction of enhanced BCL-2 knockdown observed in competition assay in the presence of 5 nM and 200 nM (+)-JQ1. (d) Western blot of BCL-2 levels following treatment with G3139 and (+)-JQ1-G3139 upon two rounds of electroporation at 96 h at concentrations indicated. Normalized to GAPDH expression levels. (e) Cell viability upon treatment with G3139 and (+)-JQ1-G3139 via electroporation for 48 h as single agents or in combination with cytarabine, at concentrations indicated. For RT-qPCR, three biological replicates are shown as diamonds for each condition (each from three technical replicates). The vertical bars represent the mean and the error bars the standard deviation. ** represents p < 0.05, *** represents p < 0.01, n.s. represents p values that are not significant.