Increased Wnt5a in squamous cell lung carcinoma inhibits endothelial cell motility

Abstract

Background: Angiogenesis is important both in normal tissue function and disease and represents a key target in lung cancer (LC) therapy. Unfortunately, the two main subtypes of non-small-cell lung cancers (NSCLC) namely, adenocarcinoma (AC) and squamous cell carcinoma (SCC) respond differently to anti-angiogenic e.g. anti-vascular endothelial growth factor (VEGF)-A treatment with life-threatening side effects, often pulmonary hemorrhage in SCC. The mechanisms behind such adverse reactions are still largely unknown, although peroxisome proliferator activator receptor (PPAR) gamma as well as Wnt-s have been named as molecular regulators of the process. As the Wnt microenvironments in NSCLC subtypes are drastically different, we hypothesized that the particularly high levels of non-canonical Wnt5a in SCC might be responsible for alterations in blood vessel growth and result in serious adverse reactions.

Methods: PPARgamma, VEGF-A, Wnt5a, miR-27b and miR-200b levels were determined in resected adenocarcinoma and squamous cell carcinoma samples by qRT-PCR and TaqMan microRNA assay. The role of PPARgamma in VEGF-A expression, and the role of Wnts in overall regulation was investigated using PPARgamma knock-out mice, cancer cell lines and fully human, in vitro 3 dimensional (3D), distal lung tissue aggregates. PPARgamma mRNA and protein levels were tested by qRT-PCR and immunohistochemistry, respectively. PPARgamma activity was measured by a PPRE reporter system. The tissue engineered lung tissues expressing basal level and lentivirally delivered VEGF-A were treated with recombinant Wnts, chemical Wnt pathway modifiers, and were subjected to PPARgamma agonist and antagonist treatment.

Results: PPARgamma down-regulation and VEGF-A up-regulation are characteristic to both AC and SCC. Increased VEGF-A levels are under direct control of PPARgamma. PPARgamma levels and activity, however, are under Wnt control. Imbalance of both canonical (in AC) and non-canonical (in SCC) Wnts leads to PPARgamma down-regulation. While canonical Wnts down-regulate PPARgamma directly, non-canonical Wnt5a increases miR27b that is known regulator of PPARgamma.

Conclusion: During carcinogenesis the Wnt microenvironment alters, which can downregulate PPARgamma leading to increased VEGF-A expression. Differences in the Wnt microenvironment in AC and SCC of NSCLC lead to PPARgamma decrease via mechanisms that differentially alter endothelial cell motility and branching which in turn can influence therapeutic response.

Keywords: Angiogenesis; Lung cancer; NSCLC; PPARgamma; Wnt.

Fig. 1

VEGF-A expression in PPARgamma KO mice and human lung tumors. a, Hematoxylin eosin staining and immunofluorescent staining of the lung of wild type C57BL/6 and PPARgamma knock-out (KO) mice show a significantly increased level of mVEGF-A in KO lung tissues. Scale bars, 200 μm and 50 μm. Intensity data are representation of three independent experiments as mean ± SEM. b, Resected human lung AC and SCC revealed significant PPARgamma decrease and VEGF-A expression increase. Data are presented as mean ± SEM. One-way ANOVA, post hoc Bonferroni; n = 11 and n = 12 per groups. P < 0.05 was considered as significant, * p < 0.05, ** p < 0.01, *** p < 0.001

Fig. 2

VEGF-A expression following modification of PPARgamma activity in A549 lung adenocarcinoma cell line tranfected with PPRE control or reporter plasmid. a, Mimicking beta-catenin dependent canonical Wnt pathway activation using 10 mM LiCl led to significant decrease in PPRE reporter activity. b, 10 mM LiCl treatment induced VEGF-A mRNA expression compared to PPRE control cells and also c, 10 mM LiCl increased VEGF-A protein levels. Error bars, SEM. One-way ANOVA, post hoc Bonferroni; n = 4. Scale bars, 20 μm. d, VEGF-A mRNA expression decreased after 10 μM PPARgamma agonist treatment (RSG), while 10 μm PPARgamma specific antagonist (GW9662) increased VEGF-A transcript levels. Independent samples t-test, n = 3. e, VEGF-A protein level shows similar pattern after 10 μm RSG and 10 μm GW9662 treatment. Fluorescence intensity are representations of three different experiments as mean ± SEM. One-way ANOVA, post hoc Bonferroni; n = 3. Scale bars, 20 μm. P < 0.05 was considered as significant, * p < 0.05, ** p < 0.01, *** p < 0.001

Fig. 3

Transcript analysis of Wnt5a, miRNA and VEGF-A in primary human lung cancer samples of AC and SCC and 3D in vitro lung aggregate cultures. a, Wnt5a mRNA is significantly upregulated in SCC compared to both normal lung and AC specimens. Error bars, SEM. One-way ANOVA, post hoc Bonferroni; n = 11 and n = 12 per groups. b, Immunohistochemical staining of Wnt5a in primary resected AC and SCC samples, n = 3 per groups. c, miR-27b and miR-200b expression levels are significantly lower in AC compared to SCC. Error bars; SEM. Independent samples t-test, n = 5 per groups. d, miR-27b is up-regulated by rhWnt5a in 3D lung aggregate cultures, while neither miR-27b nor miR-200b was affected by rhWnt11. Error bars; SEM. One-way ANOVA, post hoc Bonferroni, n = 6. e, VEGF-A and miR-27b expression levels after 10 μM RSG (PPARgamma agonist) and 10 μM GW9662 (PPARgamma antagonist) and combination treatment with rhWnt5a. Error bars; SEM. Independent samples t-test n = 3. P < 0.05 was considered as significant, * p < 0.05, ** p < 0.01, *** p < 0.001

Fig. 4

The effect of Wnt5a on VEGF-A induced endothelial cell activation and motility. a, CD105 mRNA expression is significantly higher in primary AC compared to SCC samples. Error bars, SEM. One-way ANOVA, post hoc Bonferroni; n = 11 and n = 12 per groups. b, Flow cytometric analysis of CD105 protein expression in CD31 positive endothelial cells in primary AC and SCC samples. n = 6 per groups. c, Flow cytometric analysis of CD105 levels in normal and high VEGF-A microenvironment in 3D lung aggregate tissues has also shown an increase of activation marker CD105 in VEGF-A^high tissues. The double positive (CD105/CD31) cell population was considered as activated endothelial cells. Independent samples t-test, n = 6. 1 μg/ml rhWnt5a treatment had no effect on the VEGF-A induced endothelial cell activation measured by the double positive (CD105/CD31) cell population identified by flow cytometric analysis. Independent samples t-test, n = 6. d, Localization of endothelial cells was identified by immunoflurescent staining of CD31 and analyzed by confocal microscopy in 3D lung tissue aggregates. In VEGF-A^normal microenvironment endothelial cells remained diffuse in the tissue. Under VEGF-A excess endothelial cell migrated towards the source (VEGF-A^high fibroblasts) of the signal in the center of the aggregate tissue. 1 μg/ml rhWnt5a treatment of VEGF-A^high tissue aggregates inhibited endothelial cell accumulation in the center of the aggregate. Bar chart represents the quantification of endothelial cell distribution. Relative area of CD31+ endothelial cells are compared to total field. Percentages were calculated as relative area of endothelial cells/area of total field *100. Error bars; SEM. Independent samples t-test n = 3. Representative images of three independent experiments are shown. Scale bars, 50 μm. e, HMVEC-L transwell migration assay. Endothelial cells migrate significantly faster towards VEGF-A^high fibroblast, while 1 μg/ml rhWnt5a can reverse the effect of elevated VEGF-A level. Scale bar 100 μm. One-way ANOVA, n = 3. P < 0.05 was considered as significant, * p < 0.05, ** p < 0.01, *** p < 0.001

Fig. 5

Summary of the role of Wnt dependent regulation of PPARgamma in tumor angiogenesis