Stemness marker ALDH1A1 promotes tumor angiogenesis via retinoic acid/HIF-1α/VEGF signalling in MCF-7 breast cancer cells

Abstract

Background: Aldehyde dehydrogenase 1A1 (ALDH1A1), a member of aldehyde dehydrogenase family, is a marker of stemness in breast cancer. During tumor progression cancer stem cells (CSCs) have been reported to secrete angiogenic factors to orchestrate the formation of pathological angiogenesis. This vasculature can represent the source of self-renewal of CSCs and the route for further tumor spreading. The aim of the present study has been to assess whether ALDH1A1 controls the output of angiogenic factors in breast cancer cells and regulates tumor angiogenesis in a panel of in vitro and in vivo models.

Methods: Stemness status of breast cancer cells was evaluated by the ability to form turmorspheres in vitro. A transwell system was used to assess the angiogenic features of human umbilical vein endothelial cells (HUVEC) when co-cultured with breast cancer cells MCF-7 harboring different levels of ALDH1A1. Under these conditions, we survey endothelial proliferation, migration, tube formation and permeability. Moreover, in vivo, MCF-7 xenografts in immunodeficient mice allow to evaluate blood flow, expression of angiogenic factors and microvascular density (MVD).

Results: In MCF-7 we observed that ALDH1A1 activity conferred stemness property and its expression correlated with an activation of angiogenic factors. In particular we observed a significant upregulation of hypoxia inducible factor-1α (HIF-1α) and proangiogenic factors, such as vascular endothelial growth factor (VEGF). High levels of ALDH1A1, through the retinoic acid pathway, were significantly associated with VEGF-mediated angiogenesis in vitro. Co-culture of HUVEC with ALDH1A1 expressing tumor cells promoted endothelial proliferation, migration, tube formation and permeability. Conversely, downregulation of ALDH1A1 in MCF-7 resulted in reduction of proangiogenic factor release/expression and impaired HUVEC angiogenic functions. In vivo, when subcutaneously implanted in immunodeficient mice, ALDH1A1 overexpressing breast tumor cells displayed a higher expression of VEGF and MVD.

Conclusion: In breast tumors, ALDH1A1 expression primes a permissive microenvironment by promoting tumor angiogenesis via retinoic acid dependent mechanism. In conclusion, ALDH1A1 might be associated to progression and diffusion of breast cancer.

Keywords: Aldehyde dehydrogenase 1A1; Angiogenesis; Breast cancer cells; Stemness; Vascular endothelial growth factor.

Fig. 1

Expression and activity of ALDH1A1 in breast cancer cells. a RT-PCR analysis of ALDH1A1 in breast cancer cells grown with 10% FBS. b Western blot analysis for ALDH1A1. β-Actin was used as loading control. Gel shown is representative of three experiments with similar results. c Variation of ALDH1A1 activity, measured by the formation of NADH in tumor cells. Breast tumor cell lysates were pretreated with CM037 (50 μM, 10 min) and absorbance at 340 nm (corresponding to NADH production) was measured. **p < 0.01 vs untreated cells

Fig. 2

MCF-7 ALDH1A1 affects in vitro stemness. a Representative images of tumorspheres (4x magnification) showing morphology of spheroids grown on ultra-low attachment plate. Scale bar, 100 μm. b Representative images of tumorspheres (4x magnification) of MCF-7 Scr, MCF-7 ALDH1A1KD and MCF-7 ALDH1A1⁺, showing morphology of spheroids grown on ultra-low attachment plate. Scale bar, 100 μm. b1, b2, b3. Representative images of tumorspheres (10x magnification) of MCF-1 Scr, MCF-7 ALDH1A1KD and MCF-7 ALDH1A1⁺, showing morphology of spheroids grown on ultra-low attachment plate. Scale bar, 100 μm. c Quantification of MCF-7 tumorspheres. Tumorspheres area were calculated using ImageJ Software. Ten pictures for each well were quantified. Tumorspheres> 10.000 pixel square were considered. **p < 0.01 vs MCF-7 Scr. ###p < 0.001 vs MCF-7 ALDH1A1KD. (n = 3). d Western blot analysis of stemness markers CD133 and KLF4 in MCF-7 Scr, MCF-7 ALDH1A1KD, and MCF-7 ALDH1A1⁺ tumorspheres. e Western blot analysis of ALDH1A1, HIF-1α and VEGF in MCF-7 Scr, MCF-7 ALDH1A1KD and MCF-7 ALDH1A1⁺ tumorspheres

Fig. 3

MCF-7 ALDH1A1 regulates angiogenic factor output via retinoic acid signalling. a Angiogenic factor release evaluated by ELISA plate array in supernatants of MCF-7 treated with CM037 (1 μM) for 48 h. The experiment was performed 2 times in duplicate. b MCF-7 cells were exposed to CM037 at different concentrations (1 and 10 μM) for 18 h and western blot was carried out. β-Actin was used to normalize loading. c Cells were treated with CM037 (1 μM, 18 h) and VEGF levels were measured by ELISA assay in MCF-7 conditioned media. After 18 h supernatants were harvested and cells fixed, stained and counted. The number of counted cells was not significantly different. Data are reported as pg/ml. **p < 0.01 vs untreated cells. d RT-PCR analysis of VEGF in MCF-7 Scr, MCF-7 ALDH1A1KD and MCF-7 ALDH1A1⁺ cultured in medium with 1% FBS for 48 h. Data are reported as ΔCt (Ct gene of interest-Ct Housekeeping gene). ***p < 0.001 vs MCF-7 Scr. ###p < 0.001 vs MCF-7 ALDH1A1KD. e Western blot analysis of VEGF and HIF-1α in MCF-7 exposed or not to CoCl₂ (100 μM, 72 h, 1% FBS). β-Actin was used as loading control. Gel shown is representative of three experiments with similar results. f Quantification of blots reported in e. *p < 0.05 vs MCF-7 Scr. **p < 0.01 vs MCF-7 Scr. ###p < 0.001 vs MCF-7 ALDH1A1KD. g Soluble VEGF was detected by ELISA in media conditioned by MCF-7 cells. Cells were seeded in 24-well plates at density 3 × 10⁴ cells/well. After 48 h the supernatants were harvested and cells fixed, stained and counted. The number of counted cells was not significantly different. Data are reported as pg/ml. **p < 0.01 vs MCF-7 Scr. ##p < 0.01 vs MCF-7 ALDH1A1KD. h HIF-1α and VEGF expression evaluated by western blot in MCF-7 ALDH1A1KD cells exposed for 48 h (1 μM) to exogenous retinoic acid. i HIF-1α and VEGF expression in MCF-7 ALDH1A1⁺ treated with RAR antagonist (AGN193109) and RXR antagonist (UVI 3003) for 48 h (each at 1 μM). β-Actin was used as loading control. Gel shown is representative of three experiments with similar results. j VEGF and CD133 expression in MCF-7 transiently silenced for HIF-1α. β-Actin was used as loading control. Gel shown is representative of three experiments with similar results

Fig. 4

MCF-7 ALDH1A1 regulates endothelial angiogenic features in VEGF dependent manner. a Viability of MCF-7 (Scr, ALDH1A1KD, ALDH1A1⁺) exposed to exogenous serum (10% FBS) or VEGF (2 and 20 ng/ml) at 72 h and evaluated by MTT assay. Data are reported as absorbance at 540 nm. ***p < 0.001 vs 0.1% FBS group. b MCF-7 were co-cultured with HUVEC for 48 h (1% FBS) in presence of Bevacizumab (100 ng/ml); HUVEC were fixed, stained and counted (5 fields random for well). Data are reported as number of HUVEC counted/well. (n = 3). **p < 0.01 vs HUVEC co-cultured with MCF-7 Scr without Bevacizumab. ###p < 0.001 vs HUVEC co-cultured with MCF-7 ALDH1A1⁺ without Bevacizumab. §§p < 0.01 vs HUVEC co-cultured with MCF-7 Scr without Bevacizumab. ^^^p < 0.001 vs HUVEC co-cultured with MCF-7 ALDH1A1KD. c Tumor cells were co-cultured with MCF-7 for 18 h (1% FBS) in presence of Bevacizumab (100 ng/ml). Data are reported as % area of migration ratio (% of area at 18 h/area at 0 h). **p < 0.01 vs HUVEC co-cultured with MCF-7 Scr without Bevacizumab. #p < 0.05 vs MCF-7 ALDH1A1⁺ without Bevacizumab. §p < 0.05 vs HUVEC co-cultured with MCF-7 Scr without Bevacizumab. ^^p < 0.01 vs HUVEC co-cultured with MCF-7 ALDH1A1KD. d Quantification of branching points of HUVEC seeded in Matrigel layer and co-cultured MCF-7 for 18 h (1% FBS). The results represent the media of 5 pictures. **p < 0.01 vs HUVEC co-cultured with MCF-7 Scr without Bevacizumab. ##p < 0.01 vs MCF-7 ALDH1A1⁺ without Bevacizumab. §§p < 0.01 vs HUVEC co-cultured with MCF-7 Scr without Bevacizumab. ^^^p < 0.001 vs HUVEC co-cultured with MCF-7 ALDH1A1KD. e Representative pictures of HUVEC network (4x magnification). f Tumor cells were seeded at the bottom of 12-well plates with HUVEC in transwells. The cells have been maintained in co-culture until HUVEC monolayer formation in presence or not of Bevacizumab (100 ng/ml). (n = 3). *p < 0.05 vs HUVEC co-cultured with MCF-7 Scr without Bevacizumab. ##p < 0.01 vs MCF-7 ALDH1A1⁺ without Bevacizumab. §p < 0.05 vs HUVEC co-cultured with MCF-7 Scr without Bevacizumab. ^^p < 0.01 vs HUVEC co-cultured with MCF-7 ALDH1A1KD. g HUVEC were co-cultured with MCF-7 until confluent in presence, or not of Bevacizumab (100 ng/ml). Immunofluorescent images for VE-Cadherin were obtained by confocal microscope (TCS SP5 Leica). Scale bars, 50 μm

Fig. 5

ALDH1A1 affects tumor growth and vascular flow in MCF-7 tumor xenograft in athymic nude mice. MCF-7 cells (1 × 10⁷ with 50% v/v of Matrigel) were injected s.c. in flank of athymic female nude mice. β-estradiol were injected (3 mg/kg), every 7 days i.m.. All mice were sacrificed at day 23. Tumor volumes were detected twice a week using a caliper and calculated by the formula: shortest diameter × longest diameter × thickness of the tumor in mm (n = 6 animals per group). a Tumor volume at day 23. **p < 0.01 vs Scr group. ##p < 0.01 vs ALDH1A1⁺ group. b Tumor mass at day 23. The tumors were weighted immediately after isolation from mice. **p < 0.01 vs Scr group. ##p < 0.01 vs ALDH1A1⁺ group. c Power Doppler imaging in tumors using 3D Power Doppler imaging VisualsonicVevo 2100 at the day of sacrifice. The tumor volume and percent vascularity are calculated. Red areas indicate blood flow. Images are representative of six mice per group. d Quantification of tumor vascularity (as %) by VisualsonicVevo 2100 before the sacrifice. ***p < 0.001 vs Scr group. ###p < 0.001 vs ALDH1A1⁺ group. e Quantification of tumor volumes by VisualsonicVevo 2100 at the day of sacrifice. ***p < 0.001 vs Scr group. ###p < 0.001 vs ALDH1A1⁺ group

Fig. 6

ALDH1A1 influences tumor angiogenesis and VEGF production in vivo. a Evaluation of VEGF, HIF-1α and ALDH1A1 RNA in tumor samples. Frozen tumors were homogenized and RNA was extracted to perform RT-PCR analysis of VEGF, HIF-1α and ALDH1A1 mRNA. Data are reported as ΔCt (Ct gene of interest-Ct Housekeeping gene). Each bar is the mean of 6 different tumors. The experiment was repeated two times. *p < 0.05 vs Scr group. **p < 0.01 vs Scr group. #p < 0.05 vs ALDH1A1KD group. ###p < 0.001 vs ALDH1A1KD group. b Evaluation of VEGF and ALDH1A1 proteins in tumor samples. Tissues were harvested, homogenized and sonicated. Subsequently, proteins were extracted and western blot was performed. β-Actin was used as loading control. The experiment was repeated two times. c Evaluation of mRNA for CAIX (HIF-1α target gene) and stemness markers (SOX2, NANOG, OCT-4 and TWIST) in tumor samples. Each bar is the mean of 6 different tumors. The experiment was repeated two times. #p < 0.05 vs ALDH1A1KD group. ##p < 0.01 vs ALDH1A1KD group. ###p < 0.001 vs ALDH1A1KD group. d Evaluation of HIF-1α and stemness markers (CD133, KLF4 and SOX2) proteins in tumor samples. The experiment was repeated two times. e Quantification of blots reported in d. *p < 0.05 vs Scr group. #p < 0.05 vs ALDH1A1KD group. ##p < 0.01 vs ALDH1A1KD group. f Quantification of microvessel density by human CD31 staining (magnification 20x) was done counting 5 random fields for section, each slide having five sections. **p < 0.01 vs Scr group. ##p < 0.01 vs ALDH1A1⁺ group. g Representative images of immunostaining for CD31 (red) and DAPI (blue) in tumor sections from Scr (left), ALDH1A1KD (center) or ALDH1A1⁺ (right) mice. Pictures report different vessel densities in tumors. Magnification 20x. Scale bar, 50 μm. h Representative images of double-immunostaining for CD31 (red) and NG2 (green) in tumor sections from Scr (left), ALDH1A1KD (center) or ALDH1A1⁺ (right) mice. DAPI staining is blue. Magnification 40x. Scale bars, 50 μm