The temporal dynamics of chromosome instability in ovarian cancer cell lines and primary patient samples

Abstract

Epithelial ovarian cancer (EOC) is the most prevalent form of ovarian cancer and has the highest mortality rate. Novel insight into EOC is required to minimize the morbidity and mortality rates caused by recurrent, drug resistant disease. Although numerous studies have evaluated genome instability in EOC, none have addressed the putative role chromosome instability (CIN) has in disease progression and drug resistance. CIN is defined as an increase in the rate at which whole chromosomes or large parts thereof are gained or lost, and can only be evaluated using approaches capable of characterizing genetic or chromosomal heterogeneity within populations of cells. Although CIN is associated with numerous cancer types, its prevalence and dynamics in EOC is unknown. In this study, we assessed CIN within serial samples collected from the ascites of five EOC patients, and in two well-established ovarian cancer cell models of drug resistance (PEO1/4 and A2780s/cp). We quantified and compared CIN (as measured by nuclear areas and CIN Score (CS) values) within and between serial samples to glean insight into the association and dynamics of CIN within EOC, with a particular focus on resistant and recurrent disease. Using quantitative, single cell analyses we determined that CIN is associated with every sample evaluated and further show that many EOC samples exhibit a large degree of nuclear size and CS value heterogeneity. We also show that CIN is dynamic and generally increases within resistant disease. Finally, we show that both drug resistance models (PEO1/4 and A2780s/cp) exhibit heterogeneity, albeit to a much lesser extent. Surprisingly, the two cell line models exhibit remarkably similar levels of CIN, as the nuclear areas and CS values are largely overlapping between the corresponding paired lines. Accordingly, these data suggest CIN may represent a novel biomarker capable of monitoring changes in EOC progression associated with drug resistance.

Fig 1. Images Depicting Diploid and Aneuploid States within Primary EOC Cells.

(A) Representative high resolution images presenting two diploid EOC nuclei counterstained with DAPI (blue) and labeled with CEPs 8 (red), 11 (green) and 17 (yellow). Presented are the widefield fluorescence images (top) and the corresponding rendered images (bottom) generated and employed to assess both nuclear areas and CS values. Note that within each nucleus, two foci per chromosome evaluated are present indicating the cells are diploid for chromosomes 8, 11 and 17. For illustrative purposes, the individual (CS₈, CS₁₁ and CS₁₇) and combined CS (CS_C) values have been included within each rendered image. (B) Representative high resolution, widefield fluorescence images (top) and the corresponding rendered images (bottom) of an aneuploid nucleus. Note that four foci are present for each CEP evaluated indicating the cell is tetraploid for chromosomes 8, 11 and 17. For illustrative purposes, the CS values have been included within each rendered image.

Fig 2. CIN is Associated with Primary EOC.

(A) Timeline (months) presenting the six collection times (samples B, C, D, E, H and I) from patient EOC18 (who declined treatment options) prior to death. (B) Scatter plot (left) presenting the nuclear area distributions (black circles) and the associated interquartile ranges (25^th, 50^th and 75^th percentiles; red horizontal lines). Cumulative distribution frequency graph (right) depicting nuclear areas arranged from smallest to largest. (C) A scatter plot (left) presenting the CS_C distributions of samples B to I, and the corresponding cumulative CS_C distribution frequencies (right). Note that the absolute values of gains and losses are presented for the combined CS values (CS_C; CEPs 8, 11 and 17), and that a CS_C value of 0 identifies a diploid nucleus for each chromosome evaluated (i.e. 2 copies of CEPs 8, 11 and 17). (D) Scatter plots presenting the gains and losses in CS₈ (left), CS₁₁ (middle), and CS₁₇ (right) for each nucleus analyzed within each sample. Note that both gains (positive values) and losses (negative values) are shown for each individual CEP evaluated. (E) Cumulative distribution frequency graphs for CS₈ (left), CS₁₁ (middle) and CS₁₇ (right).

Fig 3. CIN is Dynamic and Changes in Response to Frontline EOC Chemotherapy.

(A) A timeline (months) presenting the collection times for samples B, C, G and H from EOC73 relative to Carboplatin/Paclitaxel treatment. (B) Scatter plot depicting the nuclear area distribution for each sample with the interquartile ranges indicated in grey. (C) Scatter plot depicting the CS_C values for each nucleus evaluated within the indicated samples. (D) Scatter plots presenting the gains and losses in CS₈ (left), CS₁₁ (middle) and CS₁₇ (right) for each nucleus analyzed within sample.

Fig 4. CIN Increases in Recurrent Disease.

(A) Timeline (months) indicating the time of collection for samples A, C, D and F from patient EOC13 relative to treatments, surgery and recurrence. (B) Scatter plot depicting the nuclear area distributions and interquartile ranges (grey). (C) Scatter plot presenting the overall distribution of CS_C values within each sample. (D) Scatter plots for each of the individual CS values; CS₈ (left), CS₁₁ (middle) and CS₁₇ (right).

Fig 5. Extensive Levels of CIN are Associated with Recurrent Disease.

(A) Timeline (months) presenting the collection times of the six samples collected from patient EOC140 relative to surgery, disease recurrence and death. (B) Scatter plot presenting the overall distribution of nuclear areas from each sample with the interquartile range indicated in grey. (C) Scatter plot of CS_C values. Note the expanded range for the overall distributions. (D) Scatter plots presenting CS₈ (left), CS₁₁ (middle) and CS₁₇ (right) for each nucleus evaluated.

Fig 6. The Levels of CIN Increase in Aggressive and Platinum Resistant Disease.

(A) Timeline (months) indicating the collection times for samples B, F, G and H from patient EOC16 relative to treatments, platinum resistance, metastasis and death. (B) Scatter plot for nuclear areas from the indicated samples (interquartile ranges indicated in grey). (C) Scatter plot for the CS_C values. (D) Scatter plots presenting the overall distribution of CS₈ (left), CS₁₁ (middle) and CS₁₇ (right) from each nucleus evaluated within the indicated samples.

Fig 7. PEO1 and PEO4 Cells Harbor Similar Levels of CIN.

(A) Scatter plot depicting the nuclear area distribution for PEO1 (sensitive) and PEO4 (resistant) cells with the interquartile ranges (25^th, 50^th and 75^th percentiles) identified in grey. (B) Scatter plot depicting the CS_C distribution for nuclei in PEO1 and PEO4 cells. (C) Scatter plots presenting the gains and losses of CEP 8 (CS₈; left), 11 (CS₁₁; middle) and 17 (CS₁₇; right) for each nucleus analyzed in PEO1 and PEO4.

Fig 8. Mean CS Values in Primary EOC Patient Samples and Cell Line Models.

(A) Graph presenting the mCS values calculated for each patient and sample isolated from each patient, and those of the paired EOC cell line models (PEO1/PEO4 and A2780s/A2780cp). Note that the mCS value in each of the initial patient samples is <2 and remains low in all subsequent samples with the exception of EOC16 and EOC140, which exhibit striking increases in mCS values prior to succumbing to the disease. Also, note that the carboplatin resistant cell lines (PEO4 and A2780cp) both exhibit small decreases in mCS values relative to their sensitive counterparts. (B) Timelines presenting the sample collection times for each EOC patient presented in (A) relative to disease progression and treatment(s).