ADAR1-mediated regulation of melanoma invasion

Abstract

Melanoma cells use different migratory strategies to exit the primary tumor mass and invade surrounding and subsequently distant tissues. We reported previously that ADAR1 expression is downregulated in metastatic melanoma, thereby facilitating proliferation. Here we show that ADAR1 silencing enhances melanoma cell invasiveness and ITGB3 expression. The enhanced invasion is reversed when ITGB3 is blocked with antibodies. Re-expression of wild-type or catalytically inactive ADAR1 establishes this mechanism as independent of RNA editing. We demonstrate that ADAR1 controls ITGB3 expression both at the post-transcriptional and transcriptional levels, via miR-22 and PAX6 transcription factor, respectively. These are proven here as direct regulators of ITGB3 expression. miR-22 expression is controlled by ADAR1 via FOXD1 transcription factor. Clinical relevance is demonstrated in patient-paired progression tissue microarray using immunohistochemistry. The novel ADAR1-dependent and RNA-editing-independent regulation of invasion, mediated by ITGB3, strongly points to a central involvement of ADAR1 in cancer progression and metastasis.

Fig. 1

Reduced ADAR1 expression leads to increased melanoma invasion. ADAR1 reduced expression in four melanoma cell lines following ADAR1 silencing (ADAR1-KD, KD) or negative control (control, CNT) as determined by a qRT-PCR selective for p110 and p150 ADAR1 variants and b left: western blot using antibodies against ADAR1 and β-actin as loading control (grouping of images from different gels). Results are representative experiment out of three biologically independent performed; right: western blot for ADAR1 of 624mel cells treated with IFNα. The exponent numbers indicate the amount of international units (IU) used; c invasive behavior of ADAR-KD vs. control cell systems (as detailed above). Invasion was performed for 24 h using Boyden chamber assay and monitored by standardized XTT assay or membrane fixation and staining. Error bars indicate ± SEM. Numbers in the bars represent the absolute cell count of invading cells. Representative microphotographs of thincerts are shown; d correlation between % invasion and ADAR1 expression quantified by qPCR and normalized to reference melanocytes, in 10 cell lines. Correlation coefficient was determined with Spearman’s test. Asterisks represent P values: *P < 0.05; **P < 0.01 (two-tailed t-test)

Fig. 2

ADAR1-regulated invasion is mediated by ITGB3. a Normalized ADAR1 and ITGB3 expression level in 38 low-passage primary cultures of metastatic melanoma presented as 1/ΔCt. Correlation was calculated using Pearson test. b ITGB3 mRNA levels (qRT-PCR) after silencing of ADAR1 with shRNA (ADAR1-KD) or negative control (control) in four melanoma lines. Results are expressed as fold above negative control. The mean ± SE of three experiments on independent RNA purifications, each performed in triplicates, is shown. c Surface levels of ITGB3 expression, tested by flow cytometry in ADAR1-KD or control of each of the indicated cell lines. Background is isotype control; ADAR1-KD increased invasion is inhibited by blocking of ITGB3 function. d Stable ADAR-KD (KD) and control (CNT) cell lines or e control cell lines only were pre-incubated for 1 h with 10 or 30 µg/ml, respectively, anti-ITGB3 blocking antibody (blocking AB) or control IgG (control AB) and plated on the upper chamber of Boyden chamber assay. The number of invaded cells was evaluated with XTT standardized assay 24 h post seeding. Numbers in the bars represent the absolute cell count of invading cells. Data are presented as the means ± SD from three independent experiments. Statistical significance was determined by Student’s t-tests. *P < 0.05, **P < 0.01 (two-tailed t-test)

Fig. 3

ITGB3 expression is directly controlled by miR-22. a Venn diagram showing the number of differentially regulated miRNA, which are both known as tumor suppressor, demonstrate reduced expression due to silencing of ADAR1 and predicted to target ITGB3. The 15 identified miRNAs are indicated below the diagram; b dual luciferase assays. UTRs (UTR-I or UTR-II) and MUT-UTRs (UTR-MUT-A, UTR-MUT-B, and UTR-MUT-AB) denote ITGB3 3′-UTR segments containing the reference sequence or mutated sequence in the binding site of the respective miR; c effect of miR-22 overexpression (miR-22 OX+) or negative control (miR-22 OX−) on the invasion rate and ITGB3 expression in four melanoma lines with (+) or without (−) ADAR1 silencing (AD1-KD). The color code for each treatment scenario is identical for invasion and ITGB3 FACS data; d effect of anti-miR-22 (αmiR-22+) or negative control (αmiR-22−) on the invasion rate and ITGB3 expression in four melanoma lines with (+) or without (−) ITGB3 silencing (siITGβ3). The color code for each treatment scenario is identical for invasion and ITGB3 FACS data. Numbers above the bars represent the absolute cell count of invading cells; e effect of overexpression of miR-22 or control on mRNA of ITGB3 in four indicated melanoma lines. f Expression of both pri- and mature miR-22 form in four ADAR1-KD cell lines, as indicated in the figure, was examined by qRT-PCR. Results represent the mean ± SE of three biologically independent experiments, each performed in triplicates. Asterisks represent P values: *P < 0.05; **P < 0.01; ***P < 0.001 (two-tailed t-test)

Fig. 4

FoxD1 overexpression leads to decreased melanoma invasion. a The impact of reduced ADAR1 expression (ADAR1-KD, KD) on miR-22 promoter was evaluated by normalized luciferase activity of ADAR1-KD cells as compared to control cells (control, CNT) expressing naive promoter (pNaive) or mutated promoter (pMut) with point mutations at the FoxD1 designate binding site. b Expression of FoxD1 in ADAR-KD melanoma cell lines, as indicated in the figure, as compared to control cells was determined by b qRT-PCR and c western blot. d The expression of FoxD1 in four melanoma cell lines, as indicated, that were stably transduced with FoxD1 overexpression plasmid (FoxD1-OX, FxD1) or Mock plasmid (MOCK, MK) was examined by western blot. e The impact of increased FoxD1 expression on miR-22 promoter was evaluated by normalized luciferase activity of FoxD1-OX cells as compared to Mock cells expressing naive promoter (pNaive) or mutated promoter (pMut) with point mutations at FoxD1 designate binding site. f miR-22 expression was evaluated by qRT-PCR; expression of ITGB3 was determined by qRT-PCR and by extracellular staining with an anti-ITGB3-FITC-conjugated antibody and the corresponding isotype control. g Invasive behavior of cells using standardized Boyden chamber assay. The number of cells was evaluated for 24 h post seeding. Numbers in the bars represent the absolute cell count of invading cells. Results for a, b and e–g represent the mean ± SE of three biologically independent experiments, each performed in triplicates. Results for c and d and FACS stains in f are representative experiment out of three performed. Asterisks represent P values: *P < 0.05; **P < 0.01; ***P < 0.001 (two-tailed t-test)

Fig. 5

Silencing of PAX6 leads to decreased melanoma invasion. The impact of reduced ADAR1 expression (ADAR1-KD) on a ITGB3 promoter was evaluated by normalized luciferase activity of ADAR1-KD (KD) cells as compared to control cells (control, CNT) expressing naive promoter (pNaive) or mutated promoter (pMut) with point mutations at the FoxD1-binding site. Expression of PAX6 in ADAR-KD melanoma cell lines, as indicated in the figure, as compared to control cells was determined by b qRT-PCR and c western blot; d the expression of PAX6 in four melanoma cell lines, transfected with PAX6 siRNA (siPAX6 and siPX6) or control siRNA (siCNT and siCT), was examined by western blot. In c and d we used antibodies against vinculin as loading control. The impact of decreased PAX6 expression on e ITGB3 promoter was evaluated according to normalized luciferase expression in siPAX6 as compared to siCNT-treated cells expressing naive promoter (pNaive) or mutated promoter (pMut) when point mutations were made at the PAX6-binding site; f expression of ITGB3 was determined by qRT-PCR and by extracellular staining of ITGB3, followed by flow cytometry analysis. g Invasive behavior of the indicated cells using standardized Boyden chamber assay. The number of cells was evaluated for 24 h post seeding. Numbers in the bars represent the absolute cell count of invading cells; h chromatin immunoprecipitation for PAX6 followed by qPCR measurements of ITGB3 promoter region that contains the PAX6-binding site and of a control downstream sequence from the ITGB3 coding region. The Y-axis denotes the ratio between the qPCR measurements; i progression tissue microarray (TMA) was stained the indicated proteins. The staining for nuclear proteins is denoted as percent (bars represent the mean for each progression stage) while surface ITGB3 is denoted as staining intensity (individual cases for each progression stage are shown); j staining results of a representative patient. Results for a, b and e–h represent the mean ± SE of three biologically independent experiments, each performed in triplicates. Results for c and d are of a representative experiment out of three performed. TMA included triplicate cores for each case and analyzed with Wilcoxon signed rank test. Asterisks represent P values: *P < 0.05; **P < 0.01; ***P < 0.001 (two-tailed t-test)

Fig. 6

ADAR1-dependent regulation of invasion is RNA-editing independent. ADAR1 constructs used for functional assays: OX-P110 (overexpression); ΔCAT-P110; CAT-MUT-P110; and ADAR1 with mutated RNA-binding domains (RBD-mut) and Staufen1 fragments amplified and cloned into pCDNA3 and transfected into 624mel cells. a Their expression, relative to MOCK (Control), was detected by western blot (ADAR1 full length is 110 kDa and lacking the catalytic domain the size is reduced to 64 kDa). β-Actin or vinculin served as controls; b ratio of normalized ADAR1 in each indicated cell line relative to Mock (n-rADAR1 stands for normalized-ratio). The various transfectants were tested for the following: c invasion rate using standardized Boyden chamber assay monitored for 24 h after seeding. Numbers above the bars represent the absolute cell count of invading cells; d ITGB3 expression was examined by qRT-PCR and e by extracellular staining with an anti-ITGB3-FITC-conjugated antibody over isotype control; f pri- and mature miR-22 expression were monitored by qRT-PCR and g, h FoxD1 and Pax6 expression were evaluated using western blot. Results of c, d and f, g represent the mean ± SE of three biologically independent experiments, each performed in triplicates. The numbers in or next to blots indicate molecular marker. Results for a, e, and h are representative of three biologically independent experiments. Asterisks represent P values: *P < 0.05; **P < 0.01 (two-tailed t-test)