Control of cell proliferation by memories of mitosis

Abstract

Mitotic duration is tightly constrained, and extended mitosis is characteristic of problematic cells prone to chromosome missegregation and genomic instability. We show here that mitotic extension leads to the formation of p53-binding protein 1 (53BP1)-ubiquitin-specific protease 28 (USP28)-p53 protein complexes that are transmitted to, and stably retained by, daughter cells. Complexes assembled through a Polo-like kinase 1-dependent mechanism during extended mitosis and elicited a p53 response in G₁ that prevented the proliferation of the progeny of cells that experienced an approximately threefold extended mitosis or successive less extended mitoses. The ability to monitor mitotic extension was lost in p53-mutant cancers and some p53-wild-type (p53-WT) cancers, consistent with classification of TP53BP1 and USP28 as tumor suppressors. Cancers retaining the ability to monitor mitotic extension exhibited sensitivity to antimitotic agents.

Fig. 1.. Control of G1 progression and cell proliferation by memories of mitosis.

(A) Schematic highlighting that prolonging mitosis beyond a threshold (~3 times longer than normal mitosis) leads to G1 arrest of daughter cells that requires p53, p21, 53BP1 and USP28. (B) Plots of the fate of daughter cells as a function of mother cell mitotic duration for WT and CDKN1AΔ (p21 knockout) RPE1 cell lines. Mother cells with mitotic durations between 30 and 400 minutes were generated by treating an asynchronous cell population with a reversible inhibitor of spindle assembly for 6 hours. Mother cells were imaged in the presence of the inhibitor to measure mitotic duration. After inhibitor washout, mother cells completed mitosis and the resulting daughter cells were imaged for 48h to determine whether they divided again (grey), arrested (red) or underwent apoptosis (black, not observed for RPE1 cells). Each bar represents a single daughter cell, with bar height representing the mitotic duration of its mother, and color representing its fate. Percentage of arrested daughter cells produced by mothers that spent ≥100 minutes in mitosis is noted above the black lines. See also Fig. S1A,B. (C) Stills from timelapse movies monitoring in situ-tagged p21-mNeonGreen and H2B-RFP. Monastrol treatment and washout was used to prolong mitosis to different extents and p21-mNG expression was monitored in daughter cells. Panels show mother cells of different mitotic durations (left) and one of their daughters (right). (D) (left) Box-and-whiskers plot of peak daughter cell p21-mNG expression in G1, as a function of mother cell mitotic duration. (right) G1 duration of daughter cells produced by mother cells of the indicated mitotic durations. Mean and 95% CI are indicated. p-values in (D) are from t-tests (**: p<0.01; ***: p<0.001; ****: p<0.0001). See also Fig. S1C–F. (E) Panels from time-lapse movies of representative control and iCometΔ cells expressing H2B-RFP. NEBD: nuclear envelope breakdown. (F) (left) Mitotic duration at different days after induction of Comet knockout, which extends mitosis by slowing the disassembly of mitotic checkpoint complexes; mean and SD are indicated. (right) Frequency of daughter cells that arrest produced by mothers with 60-90 min mitotic duration on different days following induction of Comet knockout. The Comet knockout was also induced and analyzed in USP28Δ (shown here) and TP53sh (Fig. S2D) backgrounds. See also Fig. S2 and Fig. S3A–G. Scale bars in (C) & (E), 5 μm.

Fig. 2.. Mitosis-specific formation of stopwatch complexes transmits memory of extended mitotic duration to daughter cells.

(A) Schematic highlighting the requirement for 53BP1 and USP28 for mitotic stopwatch function. (B) Immunoblots monitoring solubility of 53BP1 and USP28 in asynchronous and mitotic cell extracts. WCE: Whole Cell Extract. GAPDH and histone H3 are soluble and chromatin-bound insoluble controls, respectively. See also Fig. S4A. (C) – (E) Analysis of 53BP1 immunoprecipitates from: (C) Asynchronous cells or cells treated to arrest at indicated cell cycle stages; (D) Asynchronous cells and cells treated to induce DNA damage or prolong mitosis. (E) Cells of the indicated genotypes treated to prolong mitosis. Treatment details are indicated in Fig. S4B. Inputs are soluble supernatants; IP: immunoprecipitate; “No Ab.” is a beads-only control. α-tubulin and GAPDH serve as loading controls. See also Fig. S4C,D. (F) (top) Schematic showing the location of the G1560K point mutation in 53BP1’s Tudor domain that disrupts its interaction with USP28 (24) and was introduced by base editing of the endogenous TP53BP1 locus; see Fig. S4E). (bottom) Analysis of 53BP1 immunoprecipitates from cells treated to prolong mitosis as described in Fig. S4B. (G) Functional analysis of the mitotic stopwatch for the indicated engineered mutant lines. The control graph is the same as in Fig. 1B. (H) Analysis of 53BP1 immunoprecipitates from cells treated to prolong mitosis or following release from synchronization at the G2-M boundary using a CDK1 inhibitor into an unperturbed mitosis. Treatment details outlined in Fig. S5A. Anti-pS10 H3, which monitors mitosis-specific phosphorylation of histone H3 on Ser10, was used as a marker for mitosis and GAPDH served as a loading control. (I) Analysis of 53BP1 immunoprecipitates from synchronized cells held in mitosis for ~2 or ~8h using the protocol schematized on the top. GAPDH serves as a loading control. Lanes shown are from a single exposure of the same immunoblot. See also Fig. S5B. (J) Analysis of the stability of stopwatch complexes following their formation in prolonged mitosis. CDKN1AΔ cells were employed to avoid the G1 arrest observed following release from extended mitosis (Fig. 1B) and were treated as indicated in the schematic above the blot. See also Fig. S5D. 12h and 24h after release from mitosis represents late G1/S and G2 phases of the daughter cells’ cell cycle. GAPDH serves as a loading control. Progression of daughter cells into the next mitosis was confirmed by live imaging (see Fig. S5E).

Fig. 3.. PLK1 kinase activity is central to the formation of stopwatch complexes that are transmitted to daughter cells.

(A) (top left) Experimental approach and list of inhibited mitotic and DNA damage kinases. Plots show results of functional analysis of the mitotic stopwatch for the indicated conditions. See also Fig. S6 and Fig. S7A–D. (B) Analysis of 53BP1 immunoprecipitates from cells treated to prolong mitosis with and without PLK1 inhibition. For treatment details, see Fig. S7E. (C) Analysis of PLK1 activity-dependent mitotic phosphorylation sites in the 53BP1 BRCT region. Schematic above indicates the 3 sites mutated to non-phosphorylatable alanine. Immunoblot below shows analysis of 53BP1 immunoprecipitates from cells treated to prolong mitosis, comparing WT RPE1 and TP53BP1Δ RPE1 cells after introduction of transgenic WT and phospho-site mutant forms of 53BP1. See also Fig. S8A–H. Lanes shown are from a single exposure of the same immunoblot; the full blot is shown in Fig. S8D. (D) Analysis of p53 levels by immunostaining following 4-day treatment with PLK4i, which delays spindle assembly (Fig. S3A). Transgene expression is heterogenous; thus, cells with comparable transgene expression are shown. (E) Quantification of p53 fluorescence signal in nuclei, comparing DMSO and PLK4i treatment, for the indicated conditions. Due to heterogeneity of transgene expression in TP53BP1Δ cells, 53BP1 signal intensity was used to first select cells with comparable expression (Fig. S8H) and p53 signal was then quantified. 10^th-90^th percentile of measured values, normalized to the average value in DMSO-treated TP53BP1Δ +WT condition, are plotted. p-value is from a t-test (****: p<0.0001). (F) Schematic summary of mitotic extension being encoded by PLK1 activity-dependent formation of stopwatch complexes that are stably inherited by daughter cells where, depending on their abundance, they either trigger immediate proliferation arrest or impart a memory of prolonged mitosis into the subsequent cell cycle.

Fig. 4.. The mitotic stopwatch is compromised in cancers and influences efficacy of anti-mitotic agents.

(A) (left) Schematic of diverse tissue-of-origin cancer-derived cell lines annotated as expressing wildtype (blue) or mutant (orange) p53; for experimental confirmation of p53 status see Fig. S10A. (right) Representative plots analyzing mitotic stopwatch function in p53-wildtype cancer-derived cell lines. See also Fig. S10B–E. (B) Plot of daughter cell arrest/death for mother cell mitotic durations above the indicated thresholds in the p53-wildtype and p53-mutant cancer cell lines. (C) (top) USP28 schematic showing mutations identified in p53-wildtype cancer lines that lack stopwatch function. (bottom) USP28 immunoblot for the indicated cell lines; α-tubulin is a loading control. See also Fig. S10F,G and Fig. S11. (D) (left) Mean relative proliferation in PLK4i versus DMSO plotted for successive 4-day intervals in a passaging assay for 15 p53-wildtype and 3 p53-mutant cancer lines. See also Fig. S12. p53-wildtype lines are grouped based on the functionality of their mitotic stopwatch. (right) Mean relative proliferation in PLK4i of CHP134 neuroblastoma cells and derived isogenic clonal lines with mutations/knockdown of stopwatch complex components. Error bars are the SD (n=3). See also Fig. S13A–C. (E) Mean relative proliferation of parental CHP134 cells and derived lines knocked down or mutated for stopwatch components in 2 nM Taxol (left) or 40 nM CENPEi (right); n=9 for controls and 3 for other conditions. Error bars are the SD. p-values are from pair-wise t-tests comparing derived cell lines to the parental line (****: p<0.0001). See also Fig. S13D.