p21Waf1 expression is regulated by nuclear intermediate filament vimentin in neuroblastoma

Abstract

Background: Human neuroblastoma (NB) cell lines may present with either one of the so-called S-and N-subtypes. We have previously reported a strong correlation between protein expression levels of vimentin, an S-subtype marker, and the p21Waf1 cyclin-dependent kinase inhibitor. We here investigated whether this correlation extend to the mRNA level in NB cell lines as well as in patients' tumors. We also further explored the relationship between expression of vimentin and p21, by asking whether vimentin could regulate p21 expression.

Methods: Vimentin and p21 mRNA levels in NB cell lines as well as in patients' tumors (n = 77) were quantified using Q-PCR. Q-PCR data obtained from tumors of high risk NB patients (n = 40) were analyzed in relation with the overall survival using the Log-rank Kaplan-Meier estimation. siRNA-mediated depletion or overexpression of vimentin in highly or low expressing vimentin cell lines, respectively, followed by protein expression and promoter activation assays were used to assess the role of vimentin in modulating p21 expression.

Results: We extend the significant correlation between vimentin and p21 expression to the mRNA level in NB cell lines as well as in patients' tumors. Overall survival analysis from Q-PCR data obtained from tumors of high risk patients suggests that lower levels of p21 expression could be associated with a poorer outcome. Our data additionally indicate that the correlation observed between p21 and vimentin expression levels results from p21 transcriptional activity being regulated by vimentin. Indeed, downregulating vimentin resulted in a significant decrease in p21 mRNA and protein expression as well as in p21 promoter activity. Conversely, overexpressing vimentin triggered an increase in p21 promoter activity in cells with a nuclear expression of vimentin.

Conclusion: Our results suggest that p21 mRNA tumor expression level could represent a refined prognostic factor for high risk NB patients. Our data also show that vimentin regulates p21 transcription; this is the first demonstration of a gene regulating function for this type III-intermediate filament.

Figure 1

Vimentin-p21 mRNA correlation in NB cell lines and tumors - Association between p21 expression and clinical evolution. A. mRNA quantification of vimentin (left panel) and p21 (right panel) relative expression level after GAPDH normalization, in 10 high risk NB cell lines, using Sybr Green Q-PCR. The S- or N-subtype of the NB cell lines is indicated. This result was confirmed after EF1α normalization and by TaqMan Q-PCR quantification, using two different genes for normalization (18 S and TFRC) (data not shown). Bars and error bars represent the means and standard deviations, respectively. B. Representation of the p21 mRNA relative expression level according to that of vimentin, quantified by Sybr Green Q-PCR after 18 S normalization, in 77 NB tumor samples collected at the time of diagnosis (primary tumor biopsies or metastatic bone marrows, all containing more than 70% of tumor cells) (Clinical data are presented in Additional file 1, Table S1). A cDNA sample from the SK-N-SH cell line was used as a reference to integrate all the data together in a unique analysis. The Spearman statistical test was performed with the Graph Pad Prism Software version 4.0. CI: Confidence Interval. C. Left panel: Representation of the p21 mRNA relative expression level according to that of vimentin, quantified by TaqMan Q-PCR after 18 S normalization, in the 40 high risk NB tumor samples of the cohort. The Spearman statistical test was performed as in 1B. This result was obtained using either Sybr-Green or TaqMan technology with two normalization genes, 18 S and GAPDH, or 18 S and TFRC, respectively (data not shown). Right panel: The survival analysis of patients older than one year of age with a stage 4 NB (N = 40) was performed using the Log-rank Kaplan-Meier estimation with the Graph Pad Prism Software version 4.0. The overall survival was defined as the time in months between diagnosis and death or last follow-up (excluding patients who died from toxicity). Patients were separated into two distinct groups according to their p21 mRNA tumor relative expression level quantified by Q-PCR as in the left panel. The median of p21 tumor expression was chosen as a cut-off between high (> median, black line) and low (< median, grey line) expression. HR: Hazard Ratio, CI: Confidence Interval.

Figure 2

Vimentin participates in p21 transcriptional regulation. A. Vimentin or p21 gene expression was monitored by Q-PCR analysis (upper panel), as described in Figure 1A. Vimentin or p21 protein levels in total cellular extracts were assessed by immunoblotting, α-Tubulin being used as loading control (lower panel). Cells were left untransfected (Unt.) or were transfected with the siRNAs targeting vimentin (siVIM) or control siRNA (siCTRL). Cells were either untreated (-) or NCS-treated (+). This figure is representative of 3 independent experiments. NCS: neocarzinostatin. B. Subcellular localization of soluble vimentin assessed by differential fractionation (C, Cytoplasmic; N, nuclear) followed by Western blot analysis in SK-N-SH, CHP-212, GI-MEN, SJN-B1 and Kelly cells lines. Topoisomerase IIα and α-Tubulin are used as controls for the nuclear and cytoplasmic soluble fractions, respectively. C. Analysis of p21 promoter activity using luciferase assay in indicated S- or N-type-like NB cell lines, either downregulated (siVIM) or not (siCTRL) for vimentin expression (upper panel), or overexpressing (pcDNA3.1-VIM) or not (CTRL: pcDNA3.1) vimentin (lower panel). Comparisons were assessed with the Mann-Whitney statistical test (* and *** stand for p < 0.05 and p < 0.005 respectively). The graph is representative of 3 independent experiments. D. Upper panel: Western Blot analysis of vimentin expression in indicated cells transiently transfected with pcDNA3.1-VIM or the empty vector (pcDNA3.1). β-actin is used as loading control. Lower panel: Vimentin cellular distribution (C, Cytoplasmic soluble; N, nuclear soluble; Ins., Insoluble cytoskeletal/matrix-associated fractions) in vimentin overexpressing cells, assessed by differential cellular fractionation followed by Western Blot analysis. Topoisomerase IIα and α-Tubulin are used as controls for the nuclear and cytoplasmic soluble fractions, respectively.