Link between DNA damage and centriole disengagement/reduplication in untransformed human cells

Abstract

The radiation and radiomimetic drugs used to treat human tumors damage DNA in both cancer cells and normal proliferating cells. Centrosome amplification after DNA damage is well established for transformed cell types but is sparsely reported and not fully understood in untransformed cells. We characterize centriole behavior after DNA damage in synchronized untransformed human cells. One hour treatment of S phase cells with the radiomimetic drug, Doxorubicin, prolongs G2 by at least 72 h, though 14% of the cells eventually go through mitosis in that time. By 72 h after DNA damage we observe a 52% incidence of centriole disengagement plus a 10% incidence of extra centrioles. We find that either APC/C or Plk activities can disengage centrioles after DNA damage, though they normally work in concert. All disengaged centrioles are associated with γ-tubulin and maturation markers and thus, should in principle be capable of reduplicating and organizing spindle poles. The low incidence of reduplication of disengaged centrioles during G2 is due to the p53-dependent expression of p21 and the consequent loss of Cdk2 activity. We find that 26% of the cells going through mitosis after DNA damage contain disengaged or extra centrioles. This could produce genomic instability through transient or persistent spindle multipolarity. Thus, for cancer patients the use of DNA damaging therapies raises the chances of genomic instability and evolution of transformed characteristics in proliferating normal cell populations.

Figure 1

Doxorubicin induced DNA damage leads to centriole disengagement during prolonged G2 phase. (A) Diagram of experimental protocol. Left image shows a 10X field of cells stained for γH2AX (red) and DNA by Hoechst (blue) 4 hours after the 1 hour Doxorubicin pulse. Right image is an enlargement of a portion of left image showing range of γH2AX labeling. Scale bars= 40µm. (B) Incidence of G2 cells with disengaged centrioles at 24, 48, and 72hrs after DNA damage. Disengagement determined by a 1:1 ratio of Centrin:C-Nap1 spots. Histogram bars indicate the average from at least 3 experiments with 200 cells counted for each condition. Error bars are one standard deviation. (C) Representative images of centrioles in control G2 cells and cells with disengaged centrioles 48hrs after Doxorubicin treatment. GFP-centrin (green), C-Nap1 (red). Scale bar=1µm. Images are maximum intensity point projections from Z series images. (D) Graph representing the distances between mother and daughter centrioles in control G2 cells and cells 24hrs after Doxorubicin treatment. Each dot represents the distance between one pair of mother-daughter centrioles. Microns between centrioles are shown along the X axis. Forty centriole pairs were measured for each condition.

Figure 2

Disengaged centrioles display markers of maturation after DNA damage. Numbers shown represent percentage of cells that contain 0, 2, or 4 foci of SAS-6 (top), CEP170 (middle), or γ-tubulin (bottom) associated with centrin foci for control G2 cells and cells 48 hours after Doxorubicin treatment. Percentages are based on 150 cells observed per condition. Images of centrioles in control cells and disengaged centrioles in Doxorubicin treated cells show GFP-centrin in green and SAS-6, CEP170, or γ-tubulin in red. Chosen images represent most prevalent phenotype for each condition. Scale bars=1µm. Images are maximum intensity point projections from Z series images.

Figure 3

Centriole disengagement after DNA damage is due to the independent activities of Plk and APC/C. (A) Diagram of experimental protocol. (B) Incidence of G2 cells with disengaged centrioles at 24 and 48 hours after addition of Doxorubicin with indicated treatments. Disengagement determined by a 1:1 ratio of Centrin:C-Nap1 spots. Histogram bars indicate the average from at least 3 experiments with 200 cells counted for each condition. Error bars are one standard deviation. *p≤ 0.05, **p≤ 0.01, ***p≤ 0.001, determined by a two-tailed unpaired Student’s t test. (C) Representative images of centrioles in a control G2 cell and in cells 48 hours after Doxorubicin treatment or 48hours after treatment with Doxorubicin plus either 200nM BI 2536, 12µM proTAME, or both. GFP-Centrin (green), C-Nap1 (red). Scale bar=1µm. Images are maximum intensity point projections from Z series images.

Figure 4

DNA damage leads to low levels of centriole reduplication during prolonged G2. (A) Representative images of centrioles in a control G2 cell and extra centrioles in a cell 48hrs after Doxorubicin. GFP-centrin (green), γ-tubulin (red). Scale bar=1µm. Images are maximum intensity point projections from Z series images. (B) Incidence of cells with more than four centrin foci colocalizing with γ-tubulin at 24, 48, and 72hrs after addition of Doxorubicin. Histogram bars indicate the average from at least 3 experiments with 200 cells counted for each condition. Error bars are one standard deviation. (C) Correlative phase contrast/immunofluorescence images of G2 arrested cells exhibiting more than four centrin foci colocalizing with γ-tubulin (upper panels) or C-Nap1 (lower panels). Phase contrast images were taken at 10X magnification and corresponding immunofluorescence images were acquired at 100X magnification. Arrows depict cell that was followed. GFP-centrin (green), γ-tubulin (red), C-Nap1 (red). Scale bars= 20µm and 1µm respectively. hr:min shown in upper left corner of each frame represents time after Doxorubicin pulse. Fluorescence images are maximum intensity point projections from Z series images.

Figure 5

Low incidence of centriole reduplication after DNA damage is due to p53-mediated inhibition of Cdk2 activity. (A) Correlative phase contrast/immunofluorescence images of a p53 knock-down cell 34hrs after Doxorubicin treatment showing extra centrioles. Phase contrast image was taken at 10X magnification and corresponding fluorescence images were taken at 100X magnification. GFP-centrin (green), γ-tubulin (red). Scale bars= 20µm and 1µm respectively. Fluorescence images are maximum intensity point projections from Z series images. (B) Incidence of cells with >4 centrin foci with indicated treatments 34hrs after Doxorubicin treatment. Histogram bars indicate the average from at least 3 experiments with 50 cells counted for each condition. Error bars are one standard deviation. *p≤ 0.05, determined by a two-tailed unpaired Student’s t test. (C) Representative images of a control cell 14hrs after mitotic shake-off stained for p21 (upper row) and cells 72hrs after Doxorubicin treatment stained for p21 (middle row) and p27 (bottom row). Inserts are magnifications of all centrioles in each cell shown. GFP-centrin (green), p21 (red), p27 (red). Merge panels include DNA stained by Hoechst (blue). Scale bar=20µm. Images are maximum intensity point projections from Z series images.