Centrosome reduction in newly-generated tetraploid cancer cells obtained by separase depletion

Abstract

Tetraploidy, a common feature in cancer, results in the presence of extra centrosomes, which has been associated with chromosome instability (CIN) and aneuploidy. Deregulation in the number of centrosomes triggers tumorigenesis. However, how supernumerary centrosomes evolve during the emergence of tetraploid cells remains yet to be elucidated. Here, generating tetraploid isogenic clones in colorectal cancer and in non-transformed cells, we show that near-tetraploid clones exhibit a significant increase in the number of centrosomes. Moreover, we find that centrosome area in near-tetraploids is twice as large as in near-diploids. To evaluate whether centrosome clustering was occurring, we next analysed the number of centrioles revealing centriole amplification. Notwithstanding, more than half of the near-tetraploids maintained in culture do not present centrosome aberrations. To test whether cells progressively lost centrioles after becoming near-tetraploid, we transiently transfected diploid cells with siRNA against ESPL1/Separase, a protease responsible for triggering anaphase, to generate newly near-tetraploid cells. Finally, using this model, we assessed the number of centrioles at different time-points after tetraploidization finding that near-tetraploids rapidly lose centrosomes over time. Taken together, these data demonstrate that although most cells reduce supernumerary centrosomes after tetraploidization, a small fraction retains extra centrioles, potentially resulting in CIN.

Figure 1

Assessment of CIN levels by FISH in 2N and 4N isogenic models. (a) Representative images of 2N (top) and 4N (bottom) DLD-1 isogenic clones after FISH using centromeric probes specific for chromosomes 4 (green), 6 (red) and 10 (yellow). DAPI was used for nuclear counterstaining. (b–e) Graphs illustrate percentage of cells with corresponding number of FISH signals for chromosomes 4, 6 and 10 for one 2N and two 4N clones of DLD-1 (b), RKO (c) and SW837 (d), and one 2N and one 4N RPE1 clones (e). A total of ~200 nuclei were analysed for each clone.

Figure 2

Enlarged centrosomes, indicative of centriole amplification events, are observed in 4N cells. (a) Representative images of 2N (top) and 4N (bottom) cells after pericentrin (green) immunostaining. DAPI was used for nuclear counterstaining. Inserts represent image amplifications in black and white of the green channel (anti-pericentrin). (b–e) Dot plots depicting centrosome area (µm²) in individual cells from one 2N and one 4N clone for DLD-1 (b), RKO (c), SW837 (d) and RPE1 (e) cell lines. Black lines denote median with interquartile range centrosome area for each clone (n = 100 centrosomes/clone). (f) Representative images of 4N cells immunostained for centrin3 (green) and cyclin D1 (red) showing increasing numbers of centrin3 dots from left to right: 2 centrioles (left side), 3 centrioles (middle-left), 4 centrioles (middle-right) and >5 centrioles (right side). DAPI was used for nuclear counterstaining. Inserts represent image amplifications in black and white of the green channel (anti-centrin3). (g–j) Bar plots showing percentage of cells in G1 phase with corresponding number of centrioles per cell for one 2N and two 4N clones of DLD-1 (g), RKO (h) and SW837 (i) cell lines, and for one 2N and one 4N RPE1 (j) clones. A total of ~750 cells were analysed for each clone. Data are reported as means ± SD. * represents P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 (n.s., not significant).

Figure 3

Deregulation of genes associated with CIN in 4N cells. (a) Venn diagram of differentially expressed genes contributing to genome doubling. Blue represents the 159 most differentially expressed genes between 4N and 2N clones obtained from genome-wide transcriptome profiling (fold-change >3 and Q < 0.05), and red represents the 70 top-ranking genes with the highest CIN score. (b) Graph showing the fold-change assessed by real time RT-qPCR of the gene ESPL1 in DLD-1, RKO, SW837 and RPE1 4N cells compared to their 2N counterparts. GAPDH was used as a housekeeping gene. Dashed red line represents the cut-off for overexpression.

Figure 4

Silencing of ESPL1 induces tetraploidization. (a) Relative expression (%) of ESPL1 after transient transfection with negative control and ESPL1 siRNAs in 2N and 4N DLD-1 (left) and RKO (right) cells. GAPDH was used as a housekeeping gene for normalization. Data are reported as means ± SD (n = 4 independent experiments/cell line). (b) Immunoblot showing decreased expression of separase after inducing gene silencing by siRNA against ESPL1 for 96 h. GAP120 was used as protein loading control. Blotting for separase and the loading control GAP120 was performed from the same gel after stripping the membrane. (c) Representative images of immunofluorescence against separase (red) comparing negative control (left) and cells treated with siRNA against ESPL1 (right). DAPI was used for nuclear counterstaining. (d) Bar plot showing the quantification of immunofluorescence staining in interphase nuclei. A minimum of 40 fields of view from two different slides of each condition (corresponding to a minimum of 150 nuclei) were analysed using ImageJ. Data are reported as mean ± SD. (e) Graph depicting significantly greater cell viability reduction in 4N compared to 2N DLD-1 cells after transient transfection with ESPL1 siRNA. Non-specific siRNA-treated cells were used as a negative control. Data are expressed as means ± SD (n = 4 independent experiments) (f) Representative FISH images of 2N DLD-1 cells before (top) and after (bottom) transient transfection with siRNA against ESPL1. FISH assays include centromeric probes for chromosome 4 (green), 6 (red) and 10 (yellow). DAPI was used for nuclear counterstaining. (g,h) Stacked bar graphs illustrate the percentage of cells with the corresponding number of signals for chromosomes 4, 6 and 10 for 2N DLD-1 (g) and RKO (h) cells transfected either with negative control or ESPL1 siRNA. More than 500 nuclei were analysed for each condition and cell line. i,j) Dot plot depicting number of chromosomes in individual cells from DLD-1 (i) and RKO (j) cells transfected either with negative control or ESPL1 siRNA. Black lines denote median chromosome number with interquartile range for each condition (n > 100 metaphases/condition). * represents P < 0.05 and ****P < 0.0001.

Figure 5

Centrosome amplification in cells after transiently silencing ESPL1. (a) Representative images of 2N DLD-1 cells immunostained with pericentrin (green) and cyclin D1 (red) antibodies. Cells transfected with negative control siRNA (top) and with siRNA against ESPL1 (bottom). DAPI was used for nuclear counterstaining. Inserts represent image amplifications in black and white of the green channel (anti-pericentrin). (b,c) Graphs depicting centrosome number in individual DLD-1 (b) and RKO (c) cells transfected either with negative control or ESPL1 siRNA. A total of ~250 cells were analysed for each condition for both cell lines. (d) Centrin3 (green) and cyclin D1 (red) co-immunostaining representative images of DLD-1 2N cells transfected with negative control or ESPL1 siRNA. While negative control transfected cells display two centrioles (top), a cell transfected with ESPL1 siRNA shows four centrioles (bottom). DAPI was used for nuclear counterstaining. Inserts represent image amplifications in black and white of the green channel (anti-centrin3). (e,f) Graphs showing the number of centrioles in 2N DLD-1 (e) and RKO (f) cells transfected with negative control or ESPL1 siRNA. A total of ~750 cells were analysed for each condition of DLD-1, and ~500 cells for each condition of RKO. Data are reported as means ± SD. * represents P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 (n.s., not significant).

Figure 6

Centriole reduction in newly-generated 4N cells. (a) Box and whisker plot depicting the number of centrioles per cell at 0, 24, 48, 96 and 144 h after tetraploidization. More than 500 nuclei were analysed for each time-point. Data are reported in boxes which extend from the 25th to 75th percentiles and whiskers from the smallest value up to the largest. Black lines denote median (4 at 0 h, 4 at 24 h, 3 at 48 h, 2 at 96 h, and 2 at 144 h) and “+” indicates the mean. Differences between time-points are significant (P < 0.0001). (b) Plots showing time-course experiments to characterise the number of centrioles. A reduction of centriole number throughout the different time-points after tetraploidization was observed. A total of ~500 nuclei were analysed for each time-point. Data are reported as means ± SD. (c) Dot plot depicting the nuclear area (µm²) at different time-points after tetraploidization. No significant differences were found between different time-points. Black lines denote mean ± SD (n = 100 nuclei/time-point). (d) Stacked bar graph showing the frequency of bipolar and multipolar anaphases. Note that significant differences were only found at 0 h after siRNA washout. Data are reported as mean (n = 200–400 anaphases/time-point). **** represent P < 0.0001.