Fig 1.
Influence of the topology and composition of the matrix on BIK expression.
Control (CTRL) and MT1-MMP (MT1) expressing MCF-7 cells were cultured for 48h in different experimental settings. (A) For 2D cultures, the cells were plated directly on plastic (Plastic), on a thin coat of dried COL1 (Dried COL1), on top of a 3D gel of polymerized COL1 (COL1 2D), overlaid with a 3D COL1 gel (COL1 Overlay). For 3D cultures, the cells were embedded into a COL1 gel (COL1 3D). (B) CTRL and MT1 MCF-7 cells were cultured on plastic or embedded within 3D COL1 (COL 3D) or 3D Matrigel (Mat 3D). BIK expression was quantified by semi-quantitative RT-PCR. Relative expression levels were obtained after normalization for the 28S rRNA levels. Data are means ± SEM (n = 3). * p<0.05, ** p<0.01, *** p<0.001 MT1 versus CTRL; # p<0.05, ### p<0.001 matrix versus Plastic (two-way ANOVA with Bonferroni post tests; *, genotype effect; #, matrix effect).
Fig 2.
Hierarchical clustering of probes and samples.
Control (CTRL) and MT1-MMP (MT1) expressing MCF-7 cells were cultured for 24, 48 and 72h on 2D plastic (Plastic) or within 3D COL1 (Col3D). RNA was extracted from each sample and gene expression values measured using the Illumina Human HT-12 BeadChip array. (A) Unsupervised hierarchical clustering of the 12 samples. The length and the subdivision of the branches display the relatedness of the samples. Hierarchical clustering was performed using Euclidian as distance measure and average linkage. A complete version of this hierarchical clustering is available as S2 Fig. (B) Heat map representation of normalized signal intensity values (log2) for probes altered by ≥ 1.8-fold (for CTRL versus MT1 or 3D COL1 versus 2D plastic, in at least one time point). Red represents relative expression greater than the median expression level across all samples, and blue represents an expression level lower than the median expression level. The colour intensity represents the magnitude of the deviation from the median. The dendrogram at the left provides a measure of the relatedness of the probe expression profile in each sample.
Fig 3.
Transcriptomic alterations induced in CTRL and MT1 MCF-7 cells when shifted from 2D plastic to 3D COL1.
Control (CTRL) and MT1-MMP (MT1) expressing MCF-7 cells were cultured for 24, 48 and 72h on 2D plastic (Plastic) or within 3D COL1 (Col3D). RNA was extracted from each sample and gene expression values measured using the Illumina Human HT-12 BeadChip array. (A) Probes with a fold change ≥ 1.8 in at least one cell type were selected and displayed as a heat map based on unsupervised hierarchical clustering. Each column represents a cell line at a specific time point and each row represents a probe. Red colour indicates genes that were up-regulated and green colour indicates genes that were down-regulated. Black indicates genes whose expression is unchanged in 3D COL1 as compared to 2D Plastic. (B) Number of probes modulated in response to the shift from 2D plastic to 3D COL1. (C) Venn analysis showing the overlap of genes at the three time points in CTRL and MT1 cells. Numbers in italics, red, and underlined represent up-, contra-, and down-regulated genes, respectively. Numbers in brackets refer to the numbers of genes modulated at each time point. (D) Venn diagram analysis of genes similarly modulated at the three time points in CTRL and MT1 MCF-7 cells in response to 3D COL1. Numbers in brackets refer to the numbers of genes modulated at each time point. Percentages represent the proportion of genes present in each area of the diagrams. Venn analyses were performed using gene symbols.
Table 1.
Top 5 significant GO terms enriched in individual lists of 3D COL1-regulated genes.
Table 2.
Bio-functions associated with the "3D COL1 core signature".
Table 3.
Gene ontology terms associated with the 3D COL1 signature.
Fig 4.
Transcriptomic alterations induced by MT1-MMP in MCF-7 cells maintained on 2D plastic.
Control (CTRL) and MT1-MMP (MT1) expressing MCF-7 cells were cultured for 24, 48 and 72h on 2D plastic. RNA was extracted from each sample and gene expression values measured using the Illumina Human HT-12 BeadChip array. (A) Probes with a fold change ≥ 1.8 between CTRL and MT1 cells in at least 1 point in the time course were selected and displayed as a heat map based on unsupervised hierarchical clustering. Red colour indicates genes that were up-regulated and green colour indicates genes that were down-regulated. Black indicates genes whose expression is unchanged in MT1 cells as compared to CTRL cells. (B) Venn diagram showing the overlap of modulated genes at the three time points. Numbers in italics, red, and underlined represent up-, contra-, and down-regulated genes, respectively. Numbers in brackets refer to the numbers of genes modulated at each time point. Venn analyses were performed using gene symbols. (C) Bio-function analysis of the dataset of genes differentially regulated upon MT1-MMP expression was performed in IPA. The significance is expressed as a negative log of p-value, which was calculated using the right-tailed Fisher's Exact Test. The orange line represents the threshold p value of 0.05.
Fig 5.
Transcriptomic alterations induced by MT1-MMP in MCF-7 cells embedded in 3D COL1.
Control (CTRL) and MT1-MMP (MT1) expressing MCF-7 cells were cultured for 24, 48 and 72h in 3D COL1. RNA was extracted from each sample and gene expression values measured using the Illumina Human HT-12 BeadChip array. (A) Probes with a fold change ≥ 1.8 between CTRL and MT1 cells in at least 1 point in the time course were selected and displayed as a heat map based on unsupervised hierarchical clustering. Red colour indicates genes that were up-regulated and green colour indicates genes that were down-regulated. Black indicates genes whose expression is unchanged in MT1 cells as compared to CTRL cells. (B) Venn diagram showing the overlap of modulated genes at the three time points. Numbers in italics, red, and underlined represent up-, contra-, and down-regulated genes, respectively. Numbers in brackets refer to the numbers of genes modulated at each time point. Venn analyses were performed using gene symbols. (C) Bio-function analysis of the dataset of genes differentially regulated upon MT1-MMP expression was performed in IPA. The significance is expressed as a negative log of p-value, which was calculated using the right-tailed Fisher's Exact Test. The orange line represents the threshold p value of 0.05.
Fig 6.
Transcriptional regulators of the MT1-MMP modulated genes in 3D COL1.
The IPA Upstream regulator analysis identified transcription factors with direct actions on differentially expressed target genes. The different molecules are presented by cellular localization. HIF1A and EPAS1 were predicted to be activated (or to have increased activity) in MT1 cells relative to CTRL cells. Genes in red and green are up- and down-regulated in response to MT1-MMP expression, respectively. E: expression; PD: protein-DNA binding; RB: regulation of binding; T: transcription.
Fig 7.
3D COL1 modulates the expression of small nuclear RNAs (snRNAs)-coding genes.
Control (CTRL) and MT1-MMP (MT1) expressing MCF-7 cells were cultured for 24, 48 and 72h on 2D plastic (Plastic) or within 3D COL1 (Col3D). RNA was extracted from each sample and gene expression values measured using the Illumina Human HT-12 BeadChip array. Microarray data were expressed as fluorescence intensities. Dashed line represents the background fluorescence.
Fig 8.
Growth conditions determine cell morphology.
Representative maximum intensity projections of confocal z-stacks immunofluorescence images of control (CTRL) and MT1-MMP expressing (MT1) MCF-7 cells cultivated in 2D culture plates (A, B) or embedded in 3D COL1 (C, D). F-actin (red) and DRAQ5 (blue). Insets: higher magnification of 3D COL1 embedded cells. Arrows, actin-rich vesicles surrounding CTRL cells; arrowheads, cytoplasmic protrusions extending from MT1 cells into the COL1 gel.
Fig 9.
Influence of Src Family Kinase inhibitors on 3D COL1-embedded MCF-7 and ZR-75–1 cells.
MCF-7 and ZR-75–1 cells were cultured for 48h on 2D plastic (Plastic) or within 3D COL1 (COL3D) in the presence of PP2 (5 μM), saracatinib (1 μM) or vehicle (DMSO 0.1%). (A) RNA was extracted from each sample and BIK mRNA levels were quantified by semi-quantitative RT-PCR. Relative expression levels were obtained after normalization for the 28S rRNA levels. (B) Apoptosis was quantified by Cell Death Detection ELISAPLUS. Data are means ± SEM (n = 3). # p<0.01 Col3D versus Plastic; * p<0.05, *** p<0.001 treatment versus vehicle (one-way ANOVA analysis with Boneferroni post test). (C) Representative maximum intensity projection of serial confocal optical sections through MCF-7 cells embedded in 3D COL1 gels and treated with saracatinib (1 μM) or vehicle for 48h. F-actin (red) and DRAQ5 (blue). Arrows: actin-rich vesicles; Arrowheads: un-retracted cytoplasmic protrusions.
Table 4.
Pharmacological inhibition of 3D COL1-induced BIK expression in MCF-7 cells.
Fig 10.
3D COL1 and MT1-MMP regulate DDR1 activation and cleavage.
Control (CTRL) and MT1-MMP (MT1) expressing MCF-7 cells were cultured for 24h on 2D plastic (Plastic) or within 3D COL1 (Col3D) in the presence (+) of BB-94 (1 μM), DDR1-IN-1 (1 μM) or vehicle (DMSO 0.1%). 3D COL1 gels were mechanically disrupted and lysed in RIPA buffer. Lysates were resolved by reducing 8% SDS-PAGE followed by immunoblot analysis. Blots were probed with phospho-DDR1 (Tyr792) antibody (A) and then reprobed with antibodies directed against the cytosolic juxtamembrane domain of DDR1 (B), BIK (C) or β-actin (D), as a loading control. Relative abundances of phosphorylated DDR1, 120-kDa full-length, 62-kDa C-terminal forms of DDR1 and BIK were quantified and normalized with respect to β-actin expression. Data are means ± SEM (n = 3). # p<0.05, ## p<0.01, ### p<0.001 treatment versus vehicle-treated cells plated on plastic; * p<0.05, *** p<0.001 MT1 versus CTRL cells (one-way ANOVA analysis with Boneferroni post test) (E). Lysates of 3T3 cells transfected with human DDR1b cDNA [101], which contains both 120-kDa full-length and 62-kDa C-terminal DDR1 species were included as a positive control.
Fig 11.
Inhibition of DDR1 tyrosine kinase activity prevents the 3D COL1-mediated induction of BIK.
MCF-7 (A) and ZR-75–1 (B) cells were cultured for 48h on 2D plastic (Plastic) or within 3D COL1 (COL3D) in the presence of DDR1-IN-1 (1 μM) or vehicle (DMSO 0.1%). RNA was extracted from each sample and BIK mRNA levels were quantified by semi-quantitative RT-PCR. Relative expression levels were obtained after normalization for the 28S rRNA levels. (C) MCF-7 cells were pre-treated with EGFP or DDR1 esiRNAs for 48h and cultured within 3D COL1 during 24h. esiRNA efficacy was analysed by western blotting. Blots were probed with an antibodies directed against the cytosolic juxtamembrane domain of DDR1 or β-actin (D), as a loading control. Relative abundances of DDR1 were quantified and normalized with respect to β-actin expression. Lysates of 3T3 cells transfected with human DDR1b cDNA [101], which contains both 120-kDa full-length and 62-kDa C-terminal DDR1 species were included as a positive control. Apoptosis was quantified by Cell Death Detection ELISAPLUS. Data are means ± SEM (n = 3). ## p<0.01, ### p<0.001 Col3D versus Plastic; ** p<0.01, *** p<0.001 treatment versus vehicle (one-way ANOVA analysis with Boneferroni post test).
Fig 12.
Schematic drawing of 3D COL1-initiated apoptotic process.
In the absence of MT1-MMP, triple helical type I collagen activates DDR1 (yellow star). Activated DDR1-triggers a signalling pathway culminating in the transcriptional up-regulation of BIK and apoptosis. The exact implication of Src family kinase (SFK) members remains unclear (question marks). Catalytically active MT1-MMP can cleave type I collagen, preventing its recognition by DDR1. MT1-MMP is also able to shed the extracellular domain of DDR1. Both cleavages impair DDR1 activation, thereby preventing BIK up-regulation and apoptosis.