TP53 mutation-mediated genomic instability induces the evolution of chemoresistance and recurrence in epithelial ovarian cancer

Background Genomic instability caused by mutation of the checkpoint molecule TP53 may endow cancer cells with the ability to undergo genomic evolution to survive stress and treatment. We attempted to gain insight into the potential contribution of ovarian cancer genomic instability resulted from TP53 mutation to the aberrant expression of multidrug resistance gene MDR1. Methods TP53 mutation status was assessed by performing nucleotide sequencing and immunohistochemistry. Ovarian cancer cell DNA ploidy was determined using Feulgen-stained smears or flow cytometry. DNA copy number was analyzed by performing fluorescence in situ hybridization (FISH). Results In addition to performing nucleotide sequencing for 5 cases of ovarian cancer, TP53 mutations were analyzed via immunohistochemical staining for P53. Both intensive P53 immunohistochemical staining and complete absence of signal were associated with the occurrence of TP53 mutations. HE staining and the quantification of DNA content indicated a significantly higher proportion of polyploidy and aneuploidy cells in the TP53 mutant group than in the wild-type group (p < 0.05). Moreover, in 161 epithelial ovarian cancer patients, multivariate logistic analysis identified late FIGO (International Federation of Gynecology and Obstetrics) stage, serous histotype, G3 grade and TP53 mutation as independent risk factors for ovarian cancer recurrence. In relapse patients, the proportion of chemoresistant cases in the TP53 wild-type group was significantly lower than in the mutant group (63.6% vs. 91.8%, p < 0.05). FISH results revealed a higher percentage of cells with >6 MDR1 copies and chromosome 7 amplication in the TP53 mutant group than in the wild-type group [11.7 ± 2.3% vs. 3.0 ± 0.7% and 2.1 ± 0.7% vs. 0.3 ± 0.05%, (p < 0.05), respectively]. And we observed a specific increase of MDR1 and chromosome 7 copy numbers in the TP53 mutant group upon disease regression (p < 0.01). Conclusions TP53 mutation-associated genomic instability may promote chromosome 7 accumulation and MDR1 amplification during ovarian cancer chemoresistance and recurrence. Our findings lay the foundation for the development of promising chemotherapeutic approaches to treat aggressive and recurrent ovarian cancer.


Background
Genomic evolution is one of the basic features of human cancers and endows malignant cells with the ability to survive and tolerate rigorous microenvironments, such as anoxia, ischemia and starvation, as well as resist attacks from immune cells [1][2][3][4]. Therefore, during chemotherapy, cancer cells undergo genomic changes to adapt to chemotherapeutics, a phenomenon referred to as single-or multi-drug chemoresistance [5,6]. According to Darwinian evolutionary theory, evolutionary process can be attributed to the presence of genetic mutations between parental and offspring generations [7]. Thus, somatic mutations, and the consequent genomic instability may be an important driving force for the development of chemoresistance in malignant tumors. Genomic instability is defined as an increased rate of DNA alterations [8]. TP53 is a checkpoint molecule that maintains genomic stability, prevents cell mitosis and induces apoptosis following abnormal chromosome segregation or chemical damage to DNA sequences [9,10]. The absence or mutation of TP53 promotes two types of genomic instability, chromosomal and amplification instability, resulting in daughter cells that exhibit large amount of aneuploidy or abnormal chromosomes [11,12]. Moreover, other related genes, such as pyruvate kinase isoform M2 (PKM2), breast cancer 1 (BRCA1) and homolog of the Schizosaccharomyces pombe cell cycle checkpoint gene (RAD17), are affected by genomic instability caused by TP53 gene mutation, resulting in an improved microenvironment that facilitates the ability of a cell to endure [13,14]. Thus far, TP53 gene mutations have been verified in more than 50% ovarian cancer patients, primarily in patients with high-grade serous carcinoma whose prognosis is worse [15,16]. In a recent rigorous reassessement, 100% of high-grade serous ovarian cancers exhibited TP53 mutations [17].
The evolutionary behavior of ovarian cancer includes short-term recurrence after anti-cancer treatments, which is the primary reason for poor prognosis. This recurrence is not only closely linked to a late FIGO stage and a high pathological grade, but is also fueled by chemoresistance [18,19]. TP53 mutations impart oncogenic proprieties, as exemplified by genomic instability, the deregulation of cell cycle progression, and multiple chemotherapeutic resistance [13,20]. In addition, overexpression of the multi-drug resistance gene MDR1 confers multi-drug resistance in almost all the solid tumors via the PI3K/Akt/Nrf2, Wnt/β-catenin, and HIF-1α/MDR1 pathways [21][22][23][24]. Along this line, it has been revealed that MDR1 and epithelialmesenchymal transition (EMT) were present in the P53-null breast cancer cell line MCF-7, which was resistant to doxorubicin [25]. However, whether aberrant expression of the MDR1 gene is associated with cancer genomic instability caused by TP53 gene mutation remains unknown and requires in-depth exploration.
In the current study, we analyzed the relationship between TP53 gene mutation and the evolution of chemoresistance in detail in ovarian cancer patients. We found that TP53 mutation promoted chromosome aneuploidy and induced MDR1 amplification instability. What's more, these patients were prone to relapse after chemotherapy. Thus, our data advocated to formulate clinically individualized treatments for ovarian cancer patients.

Study population
A total of 161 epithelial ovarian cancer patients were recruited from the Department of Obstetrics and Gynecology, Ren Ji Hospital, Shanghai, China, between June 2003 and December 2009. Patients met the following criteria: no neoadjuvant chemotherapy, treatment with cytoreductive surgery, a diagnosis of epithelial ovarian cancer confirmed by pathology and a standardized platinum-based chemotherapy after surgery. This research was approved by the Ethics Committee (No. RJ2015-087 k) of Ren Ji Hospital, Shanghai Jiao Tong University, School of Medicine, and informed consents were obtained from all patients or their direct relatives. The following information was recorded: clinicopathological characteristics, chemoresistance to platinum and patient prognosis. Patients' survival time was defined by the time from diagnosis until the date of death, or the last day of follow-up for surviving patients.

TP53 mutation analysis
Paraffin-embedded cancer specimens were dewaxed, rehydrated and stained with hematoxylin and eosin (HE) to identify the cancerous regions. An LMD6500 Laser Microdissection System (Leica, Wetzlar, Germany) was used to capture cancer cells, and DNA was extracted using a QIAamp DNA FFPE Tissue Kit (Qiagen, Valencia, CA, USA). DNA samples were then analyzed using a Qubit dsDNA HS Assay Kits (Life Technologies, Waltham, MA, USA). Targeted exon sequencing was performed with a MiSeq Reagent Kit v2 (Illumina, San Diego, CA, USA), which enables the detection of low levels of mutation from smaller amounts of DNA. Subsequently, we prepared amplified DNA libraries with a Gene Read DNA Libraries I Core Kit (Qiagen, Valencia, CA, USA), and all libraries were diluted to the designed range for cluster generation via the Illumina platform. Deep-sequencing was then performed using MiSeq (Illumina, San Diego CA, USA).

Immunohistochemistry
Paraffin-embedded tumor tissue sections (4 μm) were collected on slides, and antigen retrieval was conducted by microwaving the samples at >90°C after they were dewaxed and rehydrated. Slides were then blocked with 5% bovine serum albumin (BSA) for 1 h to reduce nonspecific binding. The specimens were incubated with the following primary antibodies for 2 h at 1:100 dilution: monoclonal mouse anti-human P53 (Clone DO-7, SANTA CRUZ, Dallas, Texas, USA) or monoclonal mouse anti-human MDR1 (SANTA CRUZ, Dallas, Texas, USA). Specimens were then incubated with a secondary antibody horseradish peroxidase (HRP)-conjugated goat anti-mouse lgG polyclonal antibody (Zhongshan, Beijing, China, dilution of 1:100) for 1 h. 3,3'-diaminobenzidine tetrahydrochloride (DAB; Zhongshan, Beijing, China) was used as a chromogen, and hematoxylin was used to counterstain the slides. Semiquantitative staining evaluation which included the fraction and intensity of stained cells, was performed by two pathologists in a blinded fashion.
Quantitative analysis of cellular DNA DNA smear image cytometry was performed on 20 patients with epithelial ovarian cancer. Deep-frozen cancer tissues were first cut into 8 μm-thick frozen sections. Laser capture microdissection was performed to dissect the cancer islets, which were digested with 0.25% trypsin (Biyuntian, Beijing, China) and 0.2% collagenase IV (Sigma-Aldrich, St Louis, MO, USA). Samples were filtered through a 70 μm screen mesh, fixed for 1 h in Bohm-Sprenger liquid (Mike Audi Corporation, Xiamen, China) and smeared to prepare single-cell-layer slides. The slides were then stainedwith Feulgen DNA dye (Mike Audi Corporation, Xiamen, China) for 75 min. DNA images were captured with a microscope, and data were analyzed withMotiCytometer and MotiClassify software (Mike Audi Corporation, Xiamen, China).

Cell cycle analysis
We collected additional fresh cancer tissue samples from 50 patients, microdissected the cancerous regions and digested the samples as described aboveto prepare ovarian cancer cell suspensions. Cells were then fixed with 70% ethanol overnight, and stained with propidium iodide (PI, 10 μg/ml) containing RNase (1 mg/ml) at 4°C for 30 min. Flow cytometry analysis of cellular DNA content was performed using a FC500 MPL flow cytometer (Beckman Coulter, Brea, CA, USA) with a total of 10,000 events per sample.

Fluorescence in situ hybridization
After paraffin-embedded tumor slides were dewaxed, fixed and rehydrated, samples were denatured via microwave exposure, and digested with pepsin (Sigma-Aldrich, St Louis, MO, USA). Then, slides were labeled with MDR1/CEN7 (centromere of chromosome 7) dualfluorescent probes (Zytolight, Bremerhaven, Germany), mounted with rubber cement (Best-Test, Galesburg, IL, USA), and allowed to hybridize at 37°C overnight. The next day, the slides were stained with DAPI and analyzed with a Leica DM2500 fluorescence microscope.

Statistic analysis
Differences in survival based on clinicopathological characteristics were assessed using a log-rank test. Correlations between the presence of TP53 mutation and clinicopathological factors were analyzed using Fisher's exact test, while nucleus diameter, DNA index and copy number were analyzed using Student's t-test. The relationships among ovarian cancer clinicopathological characteristics, TP53 mutation and recurrence were analyzed by performing multivariate logistic regression analysis. All analyses were performed with SPSS 19.0 software, and p < 0.05 was considered statistically significant.

Results
The relationship between the presence of TP53 mutation and patient clinicopathological characteristics This study included 161 epithelial ovarian cancer patients recruited from the Department of Obstetrics and Gynecology at Shanghai Ren Ji Hospital. To confirm P53 immunostaining as a surrogate marker for the presence of TP53 mutation [26,27], TP53 nucleotide sequencing and P53 immunostaining results were compared on samples obtained from 5 serous ovarian cancer patients. Next-generation sequencing analysis identified nonsynonymous TP53 mutations in 4 high-grade patients, while 1 low-grade patient exhibited wild-type TP53 (Table 1). Of the 4 patients with mutant TP53, 1 patient contained a nonsense mutation leading to the complete disruption of P53 expression, evidenced by negative staining for P53, which was referred to as a class I mutation. 2 patients exhibited a missense mutation or a frameshift mutation, resulting in partially damaged P53 function and a strong positive signal of P53 staining (moderate-to-strong staining cells in greater than 60% fraction), referred to as a class II mutation. Additionally, 1 patient exhibited a class III mutation comprising a frameshift mutation with <10% TP53 mutation abundance. Immunohistochemistry showed positive staining for P53 (10%-50% moderate staining cells) (Fig. 1). The patient with wild-type TP53 displayed weak positive staining for P53 (<10% positive cells). As the patient with a class III mutation contained only a small amount of TP53 mutant tissue within the entire tumor, and nearly unimpaired P53 protein function, this sample was approximately classified into the TP53 wild-type group, whereas class I and class II mutations were considered as the TP53 mutant group. Based on the results of the above analysis, P53 IHC may be used as a surrogate marker of TP53 mutation, and immunohistochemistry was performed for all 161 patients to identify the presence of TP53 mutation in our subsequent studies. A total of 123 patients harbored mutant TP53, while 38 patients exhibited wild-type TP53. The clinical characteristics of these patients are listed in Table 2. Parity, peritoneal metastasis, FIGO stage, tumor grade, CA125 level, the presence of residual tumor, and chemoresistance to platinum were related to the prognosis of ovarian cancer patients. Moreover, TP53 mutant patients tended to have low parity, increased peritoneal metastasis, late FIGO stage and high tumor grade.
The relationship between TP53 gene mutation and DNA abnormalities in cancer cell nuclei HE staining revealed the relative uniformity of cancer cell nucleus size in the TP53 wild-type group, and the ratio of the largest and smallest nuleus diameters was 3.4 ± 0.3, which was remarkably smaller than the diameter ratio of 5.4 ± 0.8 for the mutant group (p < 0.05) (Fig. 3a). Quantitative detection of the DNA index with DNA-specific dye was performed in ovarian cancer cell smears from 20 epithelial ovarian cancer patients. The incidence of polyploidy (>2-fold) and aneuploidy (1.1-1.9-fold) in the mutant group was 78.2 ± 7.0%, which was significantly higher than that in the wild-type group (56.0 ± 3.8%, p < 0.05) (Fig. 3b). In addition, PI was used  to stain ovarian cancer cell nuclei in samples from 50 patients, and the incidence of polyploidy and aneuploidy in the TP53 mutant group was much higher than that in the wild-type group (p < 0.01) (Fig. 3c).
Analysis of MDR1 copy number and expression in recurrent ovarian cancer tissues Immunohistochemistry was performed on samples obtained from patients whose cancers had relapsed after complete remission following chemotherapy, and the ratio of patients exhibiting elevated expression of MDR1 in relapsing tissues compared with in the treatmentnaïve lesions in the TP53 mutant group was 88.5% (54/ 61 cases), which was significantly higher than the ratio for the TP53 wild-type group (36.4%, 4/11 cases) (p < 0.01) (Fig. 5). In addition, the FISH results were consistent with the IHC results. The proportions of cancer cells containing 5-6 copies and >6 copies of MDR1 in the TP53 wild-type recurrent group were 36.0 ± 8.0% and 3.6 ± 0.7%, respectively, with no differences detected in the pre-treatment samples. However, in the mutant group, the proportions of cells containing 5-6 copies and >6 copies of MDR1 were 46.0 ± 10.3% and 14.2 ± 2.7%, which were significantly higher than the proportions of cells in the corresponding primary tissues. Furthermore, in the TP53 mutant relapse group, the proportion of cancer cells with >4 copies of chromosome 7 was significantly higher (5.4 ± 0.8)% than the proportion before recurrence (p < 0.01). In comparison, there was no obvious change for the copy of chromosome 7 in the wild-type relapse group (0.5 ± 0.07)%. Centromere FISH analysis revealed the accumulation of chromosome 7 in TP53 mutant patients during cancer evolution.

Discussion
Multi-drug resistance contributes to ovarian cancer relapse following the initial chemotherapy response [29,30]. Despite of increased doses of first-line chemotherapy drugs (such as platinum and paclitaxel) or the application of second-line chemotherapy drugs (such as topotecan, gemcitabine, liposomal doxorubicin and docetaxel), patient prognosis is poor and the average duration of survival is less than 1 year [31,32]. Aberrant MDR1 expression has been observed in most chemoresistant tumors. A better understanding of the molecular mechanism linking MDR1 and TP53, the most common mutated gene in ovarian cancer, may facilitate accurate prediction of the efficacy of standard chemotherapy and In the current study, we validated the use of immunohistochemical analysis to identify the presence of TP53 mutation in ovarian cancer, in accordance with previously published data from Yemelyanova [26]. TP53 mutation closely correlated with immunohistochemical staining patterns for P53 (strong and diffuse expression and complete lack of expression). P53 overexpression, as revealed by immunostaining, was associated with missense or frameshift mutations in TP53, whereas nonsense mutations resulted in protein truncation and a complete lack of immunostaining. Importantly, in addition to non-mutant TP53 tissue, some ovarian cancer samples contained a low-abundance frameshift TP53 mutation that resulted in weak (<10%) positive or positive (10%-50%) signals for P53 staining. Such samples were generally considered as wild-type TP53. There are two possible explanations for this phenomenon. First, direct sequencing may not be the most accurate gene testing method (massive parallel sequencing increases the accuracy of somatic mutation detection), and thus our detection of the low-abundance mutant tissue was a coincidence. Second, the presence of a low-abundance mutation in a tiny region of tumor tissue still results in adequate P53 protein function without impairment. In previous studies, the overall sensitivity for TP53 mutation detection based on P53 immunostaining in highgrade serous ovarian cancer was 99%, and for epithelial ovarian cancer, the accuracy of TP53 mutation diagnosis based on immunohistochemistry was as high as 95% [26,27].
Based on our findings, regardless of the presence of TP53 mutations, most relapse cases were first diagnosed as stage IIb-IIIc. The reason may be that tumor reductive surgeries can not completely remove cancer cells, and the remaining tumor cells seed recurrence. Additionally, according to our analysis, the primary reason for chemotherapeutic drug resistance and the relapse of ovarian cancer was P53 dysfunction caused by a gene mutation (class I/II mutation). The pathological grades among these TP53 mutant relapse cases were generally higher than the grades for the wild-type relapse patients, and in terms of the histotype, mutant relapse were generally associated with non-serous cancer. Regarding chemoresistant behavior, drug resistance in the wild-type group primarily occurred during the first chemotherapy treatment after surgery (based on the proportions of partial and no remission patients), while in the mutant group, chemoresistance also emerged during the second chemotherapy treatment after relapse, with the exception of the high relapse ratio during the first chemotherapy period. The subsequent chemoresistance in the TP53 mutant group may represent adaptive evolution as a result of the selective environmental pressure on ovarian cancer cells after the first round of chemotherapy. In addition, this result is the direct evidence of accelerating tumor chemoresistance evolution once P53 function is impaired.
A relationship between the presence of TP53 mutation and the 5-year survival of ovarian cancer patients has been reported in many studies. In a recent analysis, the hazard ratio (HR) of P53 status on survival was only 1.47 (95% CI: 1.33-1.61), and HR differed with various conditions (such as late stage and serous cancer cases). Even in late stage cases (stages III, IV), the HR for TP53 mutation decreased to 0.91 (95% CI: 0.59-1.39), indicating a possible protective effect, and serum P53 autoantibodies [HR: 1.09 (95% CI: 0.55-2.16)] were not associated with OS [33,34]. This result is likely attributable to the constrained P53 function in high-grade tumors and in late-stage serous ovarian cancer, a hypothesis that need to be tested in future studies. In the current study, TP53 mutation was not an independent factor for the 5-year survival compared with other factors (stage, grade and histotype) according to our multivariate analysis. In contrast, wild-type TP53 was a crucial factor that correlated with disease progression by preventing chemoresistance in ovarian cancer cells. Despite the proposal that TP53 mutations impact ovarian cancer recurrence in certain studies, in current study, the in-depth delineation of TP53 mutation and MDR1 copy number analysis permitted us to identify a new mechanism by which recurrence of ovarian cancer is accelerated through the evolution of chemoresistance due to TP53 mutation. Once TP53 is mutated, cancer cells with DNA damage are not able to activate the apoptosis program. Instead, cells remain in the G1 or G2 phase in anticipation of DNA repair [35,36], enabling cancer cells to augment their drug resistance capacity. Notably, cancer cell DNA content and nucleus size were markedly different in the TP53 mutant and wild-type groups, reflecting remarkable increases in the amounts of cells exhibiting polyploid and aneuploid DNA in the TP53 mutant group, which resulted in exacerbated genomic instability and a high degree of malignancy. Furthermore, we detected larger numbers of cancer cells with >6 MDR1 copy numbers and >4 copies of chromosome 7 in the TP53 mutant group. Regardless of the chemotherapeutic choice, MDR1 copy number in the mutant group was higher than the wild-type group. Based on these results, we predicted that cancer cells with high MDR1 copy numbers that were confronted with a selective environmental pressure (for example, chemotherapy after surgery) were more likely to survive and seed recurrence. To investigate this hypothesis, we compared MDR1 copy numbers and MDR1 expressions in original and relapse cancer tissues. The results were as expected: MDR1 protein levels were noticeably increased in relapse tissues, particularly in the TP53 mutant patients. Upon selective pressure exerted by chemotherapy, the number of MDR1 copies was remarkably upregulated in the TP53 mutant group, conceivably caused by the accumulation of chromosome 7. Conversely, MDR1 copy numbers in relapse patients in the TP53 wild-type group were not altered, and less chemoresistance was associated with second-round chemotherapy. We attributed recurrence in the wild-type group to other factors, such as late stage or large residual lesions. Our findings confirmed that TP53 mutation displayed an independent effect on ovarian cancer recurrence after complete remission due to effective chemotherapy. Thus, the detection of TP53 mutation is equally as important as clinical prognostic indicators (FIGO stage, histotype, grade) for predicting ovarian cancer recurrence.

Conclusion
The presence of TP53 mutation in ovarian cancer exacerbates genomic instability and promotes the expression of MDR1, which subsequently activates chemoresistance. For these patients, particularly those with an earlier stage (stages IIb and IIc) of cancer, non-serous cancer or a lower grade (G1、G2) of cancer, the administration of high-dose chemotherapeutics during the first round of chemotherapy should be considered to prevent cancer cells from rapidly generation of chemoresistance leading to short-term recurrence.