Annexin A2 facilitates endocytic trafficking of antisense oligonucleotides

Abstract

Chemically modified antisense oligonucleotides (ASOs) designed to mediate site-specific cleavage of RNA by RNase H1 are used as research tools and as therapeutics. ASOs modified with phosphorothioate (PS) linkages enter cells via endocytotic pathways. The mechanisms by which PS-ASOs are released from membrane-enclosed endocytotic organelles to reach target RNAs remain largely unknown. We recently found that annexin A2 (ANXA2) co-localizes with PS-ASOs in late endosomes (LEs) and enhances ASO activity. Here, we show that co-localization of ANXA2 with PS-ASO is not dependent on their direct interactions or mediated by ANXA2 partner protein S100A10. Instead, ANXA2 accompanies the transport of PS-ASOs to LEs, as ANXA2/PS-ASO co-localization was observed inside LEs. Although ANXA2 appears not to affect levels of PS-ASO internalization, ANXA2 reduction caused significant accumulation of ASOs in early endosomes (EEs) and reduced localization in LEs and decreased PS-ASO activity. Importantly, the kinetics of PS-ASO activity upon free uptake show that target mRNA reduction occurs at least 4 hrs after PS-ASOs exit from EEs and is coincident with release from LEs. Taken together, our results indicate that ANXA2 facilitates PS-ASO trafficking from early to late endosomes where it may also contribute to PS-ASO release.

Figure 1.

ANXA2 localizes to LEs in different cell types treated with PS-ASOs. (A) Representative immunofluorescence images of indicated cells stained for ANXA2 (green) and PS-ASO (red). Cells were incubated with 2 μM PS-ASO (IONIS ID 446654) for 8 hrs prior to staining. The lower panels are images of cells not treated with PS-ASO stained for ANXA2. Nuclei were stained with DAPI (blue). Scale bars, 10 μm. Arrows indicate the co-localization of ANXA2 and PS-ASO. (B) Western analyses for ANXA2 protein in A431 cells treated with control luciferase siRNA (Luci-si) or ANXA2-specific siRNA (ANXA2-si). Grp78 served as a control for loading. (C and D) Cells treated with control luciferase siRNA (blue) or ANXA2-specific siRNA (red) were incubated with PS-ASOs targeting either Drosha or Malat1 RNAs for 16 hrs, and levels of (C) Drosha and (D) Malat1 RNA were analyzed by qRT-PCR. Results were presented relative to no PS-ASO control. (E) ANXA2 protein levels in MHT cells treated with control siRNA (Luci-si) or with siRNAs specific for mouse ANXA2 mRNA (ANXA2-si1 and ANXA2-si2). Grp78 served as a loading control. (F and G) MHT cells treated with control siRNA (Luci-si, blue) or ANXA2 siRNAs (ANXA2-si, red; or ANXA2-si2, green) were incubated with PS-ASOs targeting SRB1 mRNA (F) or Malat1 RNA (G), and the RNA levels were analyzed by qRT-PCR. Results were presented relative to no PS-ASO control. The error bars for qRT-PCR assay represent standard deviations from three independent experiments. P-values were calculated based on two-tailed, paired t-test.

Figure 2.

Reduction of ANXA2 does not affect PS-ASO uptake. (A) Western analysis for ANXA2 protein levels in control HeLa cells or cells treated with an ANXA2-specific siRNA. GAPDH served as a loading control. (B and C) Control (Luci-si) or ANXA2 siRNA (ANXA2-si)-treated HeLa cells were incubated with PS-ASOs targeting either Drosha or Malat1 RNA for 16 hrs, and RNA levels were determined using qRT-PCR. Results were presented relative to no PS-ASO control. (D) HeLa cells treated with control (Luci-si, white bars) or ANXA2 siRNA (ANXA2-si, gray bars) were incubated with Cy3-labeled PS-ASO (IONIS ID 446654) for 3 hrs. PS-ASO uptake was analyzed by flow cytometry; results were presented as relative fluorescence units (RFU). The error bars represent standard deviations from three independent experiments. (E) Western analysis for ANXA2 protein co-isolated with PS-ASO (IONIS ID 116847) and PO-ASO (IONIS ID 364617). (F) Representative images of immunofluorescence staining of ANXA2 and LAMP1 in HepG2 cells incubated with 2 μM GalNAc-PS-ASO (upper panel) or GalNAc-PO-ASO (lower panel) for 8 hrs. Arrows indicate the co-localization of ANXA2, PS-ASO, and LAMP1. Scale bars, 10 μm.

Figure 3.

ANXA2 localization to LEs is PS-ASO dose-dependent and occurs shortly after PS-ASO treatment begins. (A) Representative images of immunofluorescent staining for ANXA2 and Rab7 in HeLa cells incubated with different concentrations of Cy3-labeled PS-ASO (IONIS ID 446654) for 8 hrs. Images for PS-ASO (red) and ANXA2 (green) were merged to show the co-localization. Merged images for Rab7 (red) and ANXA2 (green) were shown in right panels. The arrow indicates co-localization of PS-ASO, ANXA2, and Rab7 in LEs. The nuclei were stained with DAPI (blue). Scale bars, 10 μm. (B) Quantification of ANXA2-enriched loci per cell. The average number of loci in 20 cells is plotted as a function of PS-ASO concentration. The error bars represent standard deviations. (C) ANXA2 co-localization with PS-ASOs in HeLa cells incubated with 2 μM Cy3-labeled PS-ASO for different times. The co-localization of ANXA2 with PS-ASOs is exemplified with arrows. Scale bars, 20 μm. In magnified images, scale bars are 2 μm.

Figure 3.

ANXA2 localization to LEs is PS-ASO dose-dependent and occurs shortly after PS-ASO treatment begins. (A) Representative images of immunofluorescent staining for ANXA2 and Rab7 in HeLa cells incubated with different concentrations of Cy3-labeled PS-ASO (IONIS ID 446654) for 8 hrs. Images for PS-ASO (red) and ANXA2 (green) were merged to show the co-localization. Merged images for Rab7 (red) and ANXA2 (green) were shown in right panels. The arrow indicates co-localization of PS-ASO, ANXA2, and Rab7 in LEs. The nuclei were stained with DAPI (blue). Scale bars, 10 μm. (B) Quantification of ANXA2-enriched loci per cell. The average number of loci in 20 cells is plotted as a function of PS-ASO concentration. The error bars represent standard deviations. (C) ANXA2 co-localization with PS-ASOs in HeLa cells incubated with 2 μM Cy3-labeled PS-ASO for different times. The co-localization of ANXA2 with PS-ASOs is exemplified with arrows. Scale bars, 20 μm. In magnified images, scale bars are 2 μm.

Figure 4.

ANXA2 and PS-ASO co-localization is independent of S100A10. (A) S100A10 was stained in control HeLa cells (upper panel) or cells incubated with 2 μM Cy3-labeled PS-ASO for 8 hrs (lower panels: scale bar, 10 μm). S100A10 co-localization with PS-ASO and ANXA2 is exemplified using an arrow. (B) Western blot for ANXA2 and S100A10 recovered from lysates of either control or PS-ASO-treated cells by immunoprecipitation with an ANXA2 antibody (a). Silver stained SDS-PAGE gel for ANXA2-associated proteins immunoprecipitated with an ANXA2 antibody (b). Cells were treated with 2 μM PS-ASO (IONIS ID 116847) for 16 hrs; co-immunoprecipitated (IP) proteins and 10% of input lysate (input) were analyzed. (C) Western analysis for S100A10 protein levels in HeLa cells treated with either a luciferase siRNA (Luci-si) or two different siRNAs specific for S100A10 mRNA (S100A10-si). Grp78 served as a loading control. (D and E) Cells treated with control (blue) or siRNAs for S100A10 (S100A10-si1, red; S100A10-si2, green) were incubated with PS-ASOs targeting either Drosha (D) or Malat1 (E) for 16 hrs, and the RNA levels were quantified using qRT-PCR. The error bars represent standard deviations from three independent experiments. (F) Western analysis for ANXA2 and S100A10 proteins in cells treated with indicated siRNAs. Grp78 served as a loading control. G) Immunofluorescence staining of ANXA2 protein in cells treated with control siRNA (left panel) or S100A10 siRNA (right panel) then with Cy3-PS-ASO for 8 hrs. The arrows indicate co-localization of ANXA2 (green) and PS-ASO (red). Nuclei were stained with DAPI (blue). Scale bars, 10 μm.

Figure 5.

ANXA2 associated with LEs is resistant to high salt treatment. (A) Western analyses for proteins present in endosomal fractions isolated from A431 cells treated or not treated with 2 μM PS-ASO (IONIS ID 116847) for 16 hrs. Markers of EEs (Rab5), LEs (Alix), endoplasmic reticulum (Grp78), and mitochondria (P32) were evaluated. ANXA2, ANXA3 and ANXA4 were probed sequentially. The signal intensity of ANXA2, quantified using ImageJ, was reported relative to levels of Rab5 and Alix. The ANXA2 amounts were presented as a ratio of amounts in EE versus LE and as percentages in PS-ASO-treated cells compared to that in control cells. (B) LE fractions were treated with 2 M NaCl buffer for 30 min, and total, membrane pellet, and wash were analyzed by Western blotting. ANXA2 protein levels were quantified using ImageJ, normalized to Alix, and the relative levels of ANXA2 in the pellet and released fractions were given.

Figure 6.

ANXA2 is present with PS-ASOs inside LEs. (A) HeLa cells incubated with 2 μM Cy3-PS-ASO (IONIS ID 446654) for 8 hrs were co-stained for ANXA2 and LAMP1. ANXA2 and PS-ASO in a LAMP1 enclosed structure are indicated by arrows. Scale bars, 2 μm. (B) HeLa cells were incubated with 2 μM Cy3-PS-ASO (IONIS ID 446654) for 8 hrs, and cells were treated with U18666A for an additional 16 hrs. Cells were co-stained with ANXA2 and LAMP1. Scale bars, 5 μm. For magnified areas, scale bars, 1 μm.

Figure 7.

PS-ASOs gain activity after they escape from LEs. (A) Representative images of immunofluorescent staining for EEA1 (green in panel a), Rab7 (green in panel b and c), and LAMP1 (green in panel d, e and f) in HeLa cells incubated with Cy3-labeled PS-ASOs (IONIS ID 446654) for indicated times. Images for PS-ASO (red), EEA1, Rab7, and LAMP1 (green) were merged to show the co-localization, which were circled. The nuclei were stained with DAPI (blue). Scale bars are indicated. (B) Kinetic studies on PS-ASO activity for Drosha and Malat1 RNA reduction. A431 cells were incubated with PS-ASOs targeting either Drosha or Malat1 for 2 hrs after which PS-ASOs were removed. Cells were then collected at 2, 6, 10 and 14 hrs after PS-ASO treatment for activity assay and the RNA levels were quantified using qRT-PCR. The error bars represent standard deviations from three independent experiments.

Figure 8.

PS-ASO transport from early to late endosomes is impaired in ANXA2-deficient cells. (A and B) HeLa cells treated with a Luc-si or ANXA2-si for 48 hrs were incubated with 2 μM Cy3-PS-ASO (IONIS ID 446654) for 2 hrs, and co-stained with (A) EEA1 or (B) Rab7. The PS-ASO-positive EEA1- and Rab7 -stained structures are circled; EEA1- and Rab7 -stained loci that lack significant PS-ASO signals are indicated by arrows. Scale bars, 10 μm. In magnified regions shown in right panels, scale bars are indicated in the figures. (C–E) The PS-ASO-positive (C) EEs, (D) LEs, and (E) LAMP1-stained endosomes/lysosomes were counted in 15 cells and the percentage of the PS-ASO-positive organelles were calculated relative to the total numbers of the corresponding organelles. (F) The total numbers of the indicated organelles per cell. The error bars represent standard deviations. P-values were calculated based on unpaired t-test. P < 0.0001,****; N.S., not significant.

Figure 9.

Proposed model of ANXA2-mediated PS-ASO trafficking and release. PS-ASOs (black lines) enter EEs where ANXA2 (yellow open circle) is present. ANXA2 mediates biogenesis of PS-ASO-containing MVBs. ANXA2 may bind PS-ASO-containing cargo to travel into MVBs or may be recruited to the late endosomal membrane as ANXA3 (blue open circle) and ANXA4 (green open circle) through interaction with lipids. PS-ASO release mainly occurs in MVBs/LEs. Early endosomal ANXA2 facilitates PS-ASO trafficking from early to late endosomes for PS-ASO release and late endosomal ANXA2 could mediate PS-ASO release at LEs.