Ectopic expression of Snail confers epithelial cells nitidine chloride resistance via induction of ABCA1 expression

(A) Enhancement of growth inhibitory effect of nitidine chloride (NC) on human renal carcinoma cells by co-treatment with an ABCA1 inhibitor CsA. Cells were grown with or without 20 μM NC and 10 μM CsA for 24 h, and cell number was determined with CCK-8. Fold change in cell number by NC was shown as relative to controls without NC. Results are shown as mean of at least four independent experiments ± SD (Tukey-Kramer’s multiple comparison test). (B-E) Immunoblot analysis of whole cell lysate of human renal carcinoma cell lines. A representative result from three independent experiments is shown (B). Quantitative analyses of Snail (C) and ABCA1 (D) expression are presented. A scatter plot illustrating the correlation between ABCA1 and Snail expression levels is also shown (E). Quantitative data are shown as mean of at least three independent experiments ± SD (Tukey-Kramer’s multiple comparison test). (F) Comparison of ABCA1 expression between normal tissue and primary tumor of indicated subtypes of renal cancers analyzed using UCSC Xena. Bars indicate means (Welch’s t-test). (G) Immunohistochemistry of a surgically extracted renal tissue from a patient with Fuhrman grade 3 primary ccRCC, indicating upregulation of ABCA1 in the lesion site. The ABCA1 staining was verified by the accumulation of known signals in renal tubules and glomeruli in normal tissue (41). Scale bars, 250 mm (left), 50 mm (right). (H) Quantification of ABCA1 signal from surgically extracted renal tissues from three patients for each Furhman grade (Tukey-Kramer’s test; see also Figure1—figure supplement 2). (I) Phase contrast images of EpH4 wild-type cells (top) and Snail-overexpression cells (bottom, EpH4-Snail). Scale bars, 50 μm. (J) Immunofluorescence images of EpH4 and EpH4-Snail cells. The cells were fixed Scale bars, 20 μm. (K) Acquisition of NC resistance by exogenous expression of Snail in EpH4 cells and enhancement of 20 mM NC effect on EpH4-Snail cells by co-treatment with 10 mM CsA. Data are presented as in Figure 1A (Tukey-Kramer’s multiple comparison test). (L) Immunoblot analyses of whole cell lysates of EpH4, EpH4-Snail, E-cadherin KO and α-catenin KO EpH4 cells. (M) Immunofluorescence images of EpH4 and EpH4-Snail cells. Cells were fixed with MeOH. Scale bars, 20 μm. (N,O) Immunoblot analyses of whole cell lysates of MDCK II cells (N) expressing KRAS G12V treated with 5 μg/ml TGFβ for 0, 3 and 8 days and human esophageal carcinoma cell lines TE-15 and TE-8 (O). Graphs on bottom shows analyses from three biological replicates (Tukey’s multiple comparison test (N) and Student’s t-test (O)). Phase contrast images are also shown on top (O). Scale bars, 50 mm.

Cancer types that exhibit tumor specific upregulation of ABCA1 transcription

(A-C) Comparison of mRNA transcription of ABCA1 (A,B) and Snail (C) in different cancer types. Expression data from GDC TCGA database were analyzed using USCS Xena platform (Welch’s t-test).

Comparison of ABCA1 expression in different progression grades of kidney cancer cases

Immunohistochemistry of surgically extracted renal tissues from patients with Fuhrman grade 13 primary ccRCC, indicating elevation of ABCA1 signal at the lesion site of higher-grade ccRCC. All samples were collected from independent patients. The ABCA1 staining was verified by the accumulation of known signals in renal tubules and glomeruli in normal tissues (41). Scale bars, 100 mm.

Alteration of cellular cholesterol distribution induces ABCA1 expression in hybrid E/M cells

(A) Immunofluorescence images of cells stained with the LXRa+b antibody. Scale bars, 10 μm. Cells were fixed with 4% PFA. Comparison of the abundance of LXR is also shown (bottom). LXR signal intensity overlapped with DAPI signals was determined for three independent fields of view (Student’s t-test). (B) Immunoblot analyses of whole cell lysates of EpH4-Snail cells treated with 1 μM GSK-2033 for 24 h (n = 4, Student’s t-test). (C) Enhancement of growth inhibitory effect of nitidine chloride (NC) by co-treatment with an LXR inhibitor GSK2033. Cells were treated with or without 1 μM GSK2033 for 24 h, then cultured with or without 20 μM NC and 1 μM GSK2033 for 24 h, and cell (D) number was determined with CCK-8. Fold change in cell number by NC was shown as relative to controls without NC (Student’s t-test). (E) Immunoblot analyses of whole cell lysates of EpH4-Snail cells treated with indicated concentration of simvastatin or 0.1% DMSO (‘0’ mM) in DMEM supplemented with 10% normal or lipid-free fetal bovine serum (FBS) for 48 h (n = 3, Tukey-Kramer’s test). (F) Subcellular distribution of cholesterol stained with filipin. Scale bar, 10 mm. Intracellular signal to surface signal ratio was determined for three independent fields of view (Student’s t-test). (G) Live cell images of lipid droplets (LDs) visualized by Lipi-Green dye. (H) Quantification of LD signal per cell. The Lipi-Green signals of all optical sections were summed and divided by the number of cells counted using DAPI images. Shown are means of three different fields of view ±SD. (I) Subcellular distribution of Topfluor-(TF-)cholesterol. Living cells were incubated with 5 mg/ml TF-cholesterol for 1 h, then stained with Lipi-Deepred and Nuc Blue. Scale bar, 20 mm. (J) Observation of filipin in EpH4-Snail cells expressing either EGFP-ADRP (lipid droplets) or mRFP-LAMP1 (lysosome). Scale bar, 20 mm. (K) Comparison of cholesterol content relative to phospholipid (Student’s t-test). (K,L) Comparison of cell surface cholesterol (D4) and sphingomyelin (lysenin). Wild-type EpH4 cells expressing H2B-iRFP and EpH4-Snail cells were co-cultured and stained with recombinant lipid-binding proteins and CellMask Green (K). Quantification of fluorescent signals of lipid binding probes (L). Signal intensities per cell, relative to those of wild-type cells, were quantified and analyzed (Student’s t-test). (M–O) Observation of lipid order of plasma membrane using FpCM-SO3. See also Figure 2—figure supplement 2. Wild-type EpH4 cells expressing H2B-iRFP and EpH4-Snail cells were co-cultured and stained with 1 mM FpCM-SO3 for 10 min. The hue (Red/Blue ratio) – brightness (intensity) images of FpCM-SO3 are shown with the confocal image of H2B-iRFP (magenta) (M). Normalized histogram of Red/Blue ratio values within FpCM-SO3 positive pixels (N). Comparison of mean Red/Blue ratio values (P) (Student’s t-test).

Contribution of transcription factors involved in ABCA1 transcription in epithelial cells

(A) Immunoblot analysis of whole cell lysates of EpH4-Snail (without treatment) and EpH4 cells treated with inhibitors of PI3K-Akt pathway (5 mM GSK2334470: PDK1, 50 mM LY294002: PI3K, or 1 mM Wortmannin: PI3K for 24 h). Comparison of protein levels of ABCA1 and phospho-FoxO3a is also shown (Tukey-Kramer’s test). (B) Immunoblot analysis of whole cell lysates of EpH4 and EpH4-Snail cells (Student’s t-test). Quantitative analysis of c-myc expression is also shown.

Effect of ABCA1 knockout on nitidine chloride resistance of EpH4-Snail cells

(A) Immunoblot analysis of whole cell lysates of wild-type and two lines of ABCA1 KO EpH4-Snail cells (#1 and #2). (B) Effect of ABCA1 knockout on growth inhibitory effect of nitidine chloride. Cells were grown with or without 20 μM NC for 24 h, and cell number was determined with CCK-8. Fold change in cell number by NC was shown as relative to controls without NC (Tukey-Kramer’s test).

Visualization and analyses of lipid order using FπCM-SO3

(A) Schematic illustrating the alteration of fluorescence spectra of solvatochromic dyes in lipid bilayers with different lipid compositions. A high content of sphingolipids and cholesterol induces an ordered acyl chain orientation. The dyes exhibit red-shifted fluorescence in disordered membranes and blue-shifted fluorescence in ordered membranes. (B) Schematic showing lipid-order visualization and analysis using FpCM-SO3. Living cells cultured on glass-bottom dishes were stained with 1 mM FpCM-SO3 for 10 min. The dye was excited with a 405 nm laser, and emission signals of 400–500 nm (Blue) and 500–700 nm (Red) were separated using a variable dichroic mirror and simultaneously detected by two photomultipliers. The raw images were Gaussian-blurred, Red/Blue ratio and mean intensity of both channels were calculated. For visualization of lipid order, a ratio-intensity image was generated in which the hue represents the Red/Blue ratio and the brightness corresponds to the signal intensity (bottom left). For quantitative comparison, a threshold was set based on the fluorescence intensity of FpCM-SO3, and histograms of the Red/Blue ratio were generated from positive pixels (bottom center). For statistical assessment, mean Red/Blue ratio values of positive pixels were obtained and compared (bottom right).

Effects of Snail-induced EMT on sphingomyelin profile

(A) Comparison of sphingomyelin (SM) content relative to phospholipid content. (B) Comparison of cholesterol (Chol)/SM ratio between EpH4 and EpH4-Snail cells. (C) Fatty acid composition of SM determined using LC-MS. All peaks corresponding to d18:1-even fatty acid-SMs were included in the analysis and each peak value is expressed as a percentage. n.d., peak not detected. (D) Pie charts of SM chain length profile from c. SMs were classified into long-chain fatty acid (LCFA, orange) (≤20) and very long-chain fatty acid (VLCFA, purple) (>20) SMs. (E) Chain length specificity of Elovls and CerSs responsible for fatty acid profile of sphingolipids reported (25, 26). Elovls catalyze a rate-limiting step of fatty acid elongation cycle, and CerSs catalyze N-acylation of sphingoid bases to produce ceramides. (F) Comparison of transcription levels of ELOVLs and CERSs. Notable decreases in expressions of ELOVL7 and CERS3 in EpH4-Snail cells are indicated by asterisks. (G) Quantification of transcription levels of ELOVL7 and CERS3 relative to that of ZO-1, whose expression does not change between EpH4 and EpH4-Snail (14). (Student’s t-test). (H) Effect of supplementation of C22:0 ceramide (Cer22:0) on ABCA1 expression. EpH4-Snail cells were treated with 10 mM ceramide or 0.4% vehicle (dodecane/ethanol = 2/98) for 24 h, then whole cell lysates were analyzed by immunoblot. (I) Immunoblot analyses of whole cell lysates of mouse epithelial (MTD-1A, CSG) and fibroblast (L-929, NIH3T3) cells. (J) Comparison of Chol/SM ratio of EpH4, EpH4-Snail and normal fibroblasts (Dunnett’s test). (K) Comparison of Chol/SM ratio between TE-15 and TE-8 (Student’s t-test). (L) Scatter plot of Chol content against SM content used in B, H and I. Dashed lines indicate Chol/SM = 1.0 (blue) and 1.5 (red). The region of Chol/SM > 1.5 is shown with red background. (M) Possible mechanism of induction of ABCA1 expression and lipid droplet enlargement through decrease in VLCFA-SM. In EpH4 (epithelial) cells, a certain amount of choleseterol is sequestered through interaction with SM, organizing ordered plasma membrane (PM). In EpH4-Snail (hybrid E/M) cells, the lower level of VLCFA-SM biosynthesis leads to a decrease in plasma membrane SM content, resulting in membrane disorder. The increase in SM-unbound form of cholesterol in PM induces its internalization and activation of cellular cholesterol sensors, resulting in to lipid droplet enlargement and enhanced ABCA1 expression.

Comparison of SM and cholesterol content.

(A,B) Comparison among EpH4, EpH4-Snail, and fibroblasts. (C,D) Comparison between esophageal cancer cell lines TE-15 (Snail-negative) and TE-8 (Snail-positive). n = 3, ** p < 0.01, * p < 0.05, n.s., p ≥ 0.05 by Tukey’s multiple comparison test (A,B) and Student’s t-test (C,D).

Selective growth inhibition of Snail-positive cells by an ACAT inhibitor TMP-153

(A) Schematic illustrating the molecular mechanisms for handling excess cholesterol to avoid cell death. Cells eliminate excess cholesterol through efflux mediated by transporters including ABCA1 or isolation into lipid droplets (LD) via esterification of cholesterol by acyl-CoA: cholesterol acyltransferases (ACATs) to avoid cell death. CsA and TMP-153 inhibit ABCA1 and ACATs, respectively. (B) Immunoblot analysis of whole cell lysates of human kidney cell lines. Quantitative analysis of SOAT1 expression is also shown. (C-F), Effects of treatments with CsA (C,E) or TMP-153 (D,F) to cellular growth of EpH4 and EpH4-Snail (C,D) or human kidney cancer cell lines (E,F). Cells were treated with drugs for 24 h and relative cell number to 0.1% DMSO control was determined by CCK-8. Results are shown as mean of at least three independent experiments ± SD. (G-I) TMP-153 inhibits the growth of tumor xenografts composed of Snail-positive renal cancer 786-O cells. Experiments were performed as described in Materials and Methods. In brief, 50 mg/kg TMP-153 was administered intraperitoneally to nude mice bearing 786-O tumor xenografts. After 21 days of the first administration, the overall appearance (G) and the excised tumors (H) of control and treated mice were compared. (I) Tumor growth curve in the xenograft model. Results are shown as mean of four biological replicates ± SD (Student’s t-test).

Selective growth inhibition of a Snail positive esophageal cancer cell line by combined treatment of CsA and TMP-153

(A-C) Effects of treatments with CsA (a), TMP-153 (b) or combination of 50 mM CsA and indicated concentration of TMP-153 (c) to cellular growth of human esophageal cancer cell lines TE-15 and TE-8. Cells were treated with drugs for 24 h and relative cell number to 0.1% DMSO control was determined by CCK-8. n ≥ 3, error bars, SDs. (one-sided Student’s t-test). (D-F), Combined treatment of CsA and TMP-153 inhibits the growth of TE-8 tumor xenografts. Experiments were performed as described in Materials and Methods. In brief, 10 mg/kg CsA and 10 mg/kg TMP-153 were simultaneously administered intraperitoneally to nude mice bearing TE-8 tumor xenografts. After 36 days of the first administration, the overall appearance (D), the excised tumors (E), and the tumor size (F) of control and treated mice were compared. n = 5, error bars, SDs (Student’s t-test).