Loss of pRB causes centromere dysfunction and chromosomal instability

Abstract

Chromosome instability (CIN) is a common feature of tumor cells. By monitoring chromosome segregation, we show that depletion of the retinoblastoma protein (pRB) causes rates of missegregation comparable with those seen in CIN tumor cells. The retinoblastoma tumor suppressor is frequently inactivated in human cancers and is best known for its regulation of the G1/S-phase transition. Recent studies have shown that pRB inactivation also slows mitotic progression and promotes aneuploidy, but reasons for these phenotypes are not well understood. Here we describe the underlying mitotic defects of pRB-deficient cells that cause chromosome missegregation. Analysis of mitotic cells reveals that pRB depletion compromises centromeric localization of CAP-D3/condensin II and chromosome cohesion, leading to an increase in intercentromeric distance and deformation of centromeric structure. These defects promote merotelic attachment, resulting in failure of chromosome congression and an increased propensity for lagging chromosomes following mitotic delay. While complete loss of centromere function or chromosome cohesion would have catastrophic consequences, these more moderate defects allow pRB-deficient cells to proliferate but undermine the fidelity of mitosis, leading to whole-chromosome gains and losses. These observations explain an important consequence of RB1 inactivation, and suggest that subtle defects in centromere function are a frequent source of merotely and CIN in cancer.

Figure 1.

Loss of pRB induces CIN. (A,B) FISH analysis with probes for chromosomes 6 and 8 revealed a high degree of aneuploidy in populations of pRB-depleted RPE-1 (sh-Rb) cells that persisted following several weeks of chronic depletion. Both control and sh-Rb populations exhibited a modal copy number of two for both chromosomes 6 and 8. (C) Analysis of individual segregation events revealed high rates of chromosome segregation errors in pRB-depleted cells (0.87% and 0.64% segregation error/division for chromosomes 6 and 8, respectively). (D) Following chronic depletion of pRB, aneuploid cells continued to proliferate (prometaphase cell) (top panel) and to missegregate chromosomes (late anaphase/telophase cell) (bottom panel). Chromosome 6 is shown in green, and chromosome 8 is shown in red throughout. Bar, 5 μm.

Figure 2.

pRB loss causes centromeric dysfunction. (A) RPE-1 cells lacking pRB exhibited >20% increase in intercentromeric distance. (B) This increase was not restored by inhibition of PLK1, but was partially corrected by microtubule depolymerization. (C) Chromosome alignment was also impaired, and the width of the metaphase plate was increased ∼40% in cells lacking pRB. Error bars represent standard error (SEM) of three independent experiments. (D) Centromeres were distorted when pRB was depleted, such that the barbell configuration (staining the kinetochores and associated centromeric region) seen with ACA staining in control cells (insets in top panels) was often bent (insets in bottom panels) in cells lacking pRB. Kinetochores are shown in red, microtubules are shown in green, and centrioles are shown in yellow (A) or blue (D). Similar results were obtained with siRNAs or shRNAs; the data shown were obtained from shRNA-treated cells. Bars, 10 μm. Insets are 4× enlargements.

Figure 3.

pRB loss disrupts sister chromatid cohesion. (A) Cells depleted of pRB (shRb) have an increased incidence of premature loss of sister chromatid cohesion (examples of separated sister are indicated by asterisks [*]). (B) Line scans through arms of sisters that maintain cohesion revealed an increase in interchromosomal distance in response to pRB loss. Each colored line represents a line scan through the arms of paired sister chromatids, with each chart depicting five line scans within a single representative cell. Prolonged mitotic arrest increased the interchromosomal distance (C) (error bars represent standard error [SEM]; P < 0.05 between control and sh-Rb cells at both 3 and 20 h) and promoted the complete loss of cohesion (D), as measured by the number of cells with one to three or four or more unpaired chromosomes.

Figure 4.

pRB loss compromises the fidelity of mitotic progression. Formation of a bipolar spindle (A,D), and progression into anaphase (B) following release from nocodazole-induced mitotic arrest is delayed when pRB is depleted (shRb). (A) The state of spindle organization is characterized as disorganized microtubule arrays (dark blue), multipolar spindles (light blue), and bipolar spindles (yellow). Numbers represent percentage of mitotic cells exhibiting specified spindle structure. P < 0.01 for differences in spindle structure and anaphase progression at time points 40 and 60 min between control and shRb samples. (C) pRB-depleted cells that enter anaphase exhibit an increase in the incidence lagging chromatids over control cells. (D) Examples of representative spindle structure at specified time points following nocodazole washout. Microtubules are shown in green, and DNA is shown in blue. Bar, 10 μm.

Figure 5.

pRB depletion impairs cohesin localization. (A) Western blot analysis shows that the chromatin-bound fraction of cohesin components Rad21 and SMC3 and the condensin II component CAP-D3, but not the condensin I component CAP-D2, were reduced in pRB-depleted (shRb) cells. (B) Punctate Rad21 staining was apparent in control cells after nuclear envelope breakdown (NEB) and during mitosis, but was dramatically reduced in cells depleted of pRB. Rad21 is shown in orange, chromosomes are shown in blue, and centrosomes are shown in red. Bar, 10 μm. (C) The levels of cleaved Rad21 (indicated by an asterisk [*]) did not increase in shRb-depleted cells.

Figure 6.

Loss of RBF1 in Drosophila causes chromatid cohesion defects and aneuploidy, similar to that seen in pRB-depleted human cells. (A,B) Mitotic chromosome spreads from dissected larval neuroblasts show frequent loss of sister chromatid cohesion (A) and increased incidence of aneuploidy (B). (W) Wild type; (M) rbf1 mutant. Error bars represent standard error (SEM) of three individual animals per condition. (C) FACs analysis of whole larval neuroblasts from wild-type and rbf1 mutant animals similarly reveals that the rbf1 mutants exhibit an increase in aneuploid cells. (D) dRad21 banding on salivary gland polytene chromosomes is decreased in rbf1 mutant animals.

Figure 7.

Centromeric localization of condensin II is compromised in pRB-depleted cells. (A) Chromatin-associated CAP-D3 is reduced in cells depleted of pRB by siRNA, but not in cells depleted of Rad21 (control vs. pRB depletion: P < 0.0001) (quantification of immunofluorescence from Supplemental Fig. 6). (B,C) Centromeric CAP-D3 localization is reduced in pRB-depleted cells to <30% of that seen in control cells (control vs. pRB depletion: P < 0.0001). Microtubules are shown in green, kinetochores are shown in red, and CAP-D3 is shown in blue. (D) The depletion of CAP-D3 causes an increase in intercentromeric distance similar to that observed in pRB-depleted cells. All error bars represent standard error (SEM) of three independent experiments.