Introduction: High-grade serous carcinoma (HGSC) is not only the most common form of ovarian cancer, but also the deadliest. HGSC accounts for over 68% of all gynecologic cancer and is estimated to claim over 13,000 lives in 2024 [1][2][3]. Initial survival rates are high with approximately 85% of patients surviving one year post-diagnosis; these rates plummet over time, declining to a 5-year survival rate of 32%, and a 10-year survival rate of 15% [4].
Patients undergoing treatment for HGSC achieve remission in most cases [2], which seems contradictory to the long-term survival rates. These low survival rates can be attributed to the high relapse rates of HGSC. Patients have a 70%-95% chance of recurrence [5]. Recurrent tumors are more aggressive, less responsive, and lead to poorer outcomes than the original tumors [6].
Initial treatments for HGSC can be successful in minimizing the original tumors during treatment, but they do not eradicate all tumor cells. These residual cell populations, though dormant during remission, may become proliferative once more and form recurrent tumors. Developing a deeper understanding of how modern cancer treatments impart arrest onto HGSC cells will provide insights on modulating these treatments to fully eradicate each sub-tumor cell population, thereby preventing further post-remission tumor development.
The two cell lines chosen for this experiment differ in their production of Cyclin E1, a key regulator of the G1 phase. Understanding differences– and similarities– in the arrest states of these models can also inform targeted therapies based upon individual patients' unique tumor signatures.
Materials and
Methods: The OVCAR-3 and OVCAR-8 cell lines were used as HGSC model cell lines. Both cell lines are homologous recombination-proficient. OVCAR-3 has a Cyclin E1 (cycE1) gene amplification.
Cells were grown in RPMI Medium 1640 with varying compositions of fetal bovine serum (15% for OVCAR-3 and 10% for OVCAR-8) and 1% Penicillin/Streptomycin (Gibco, USA). Cells were incubated at 37°C and 5% CO₂.
96-well plates (CellVis, USA) were prepared with 8 drugs with varying concentrations ranging from 0.5 nM to 10 uM for dose response curves. Drugs analyzed were those common in front-line cancer treatment. The drugs examined are detailed with their mechanism of action in Table 1.
Cells were plated 24 hours prior to treatment; samples incubated 116 hours prior to incorporation of EdU for 4 to gauge DNA synthesis. Chemical fixation with 4% paraformaldehyde and permeabilization with 0.3% Triton X-100 was completed prior to Hoechst staining. Finally, the samples were processed in accordance to instructions from the EdU Cell Proliferation Kit (Thermo Fisher C10337).
High-resolution fluorescence images were obtained using the Leica Biosystems THUNDER Imager. Resulting images were quantified and analyzed using a custom Python-based pipeline [7]. DAPI images were analyzed to produce Integrated DNA, a metric of total genetic content within each cell. EdU images were analyzed to find the percentage of cells which were positive for EdU, indicating synthesis of new DNA.
Results, Conclusions, and Discussions: EdU staining for the samples occurred across berzosertib, CVT-313, palbociclib, and PF-06873600 for both OVCAR-3 and OVCAR-8 cells. Dose response curves generated can be found in Table 2.
The effects between cell lines are mostly consistent; responses to berzosertib and PF-068700 were both met with complete halting of active proliferation, as indicated by zero positive EdU cells at the maximum concentration.
The first difference is seen in CVT-313, in which the final efficacy of OVCAR-3 cells is lower than their OVCAR-8 counterparts. This may be due to a need for a wider range of concentrations, as the full sigmoid curve associated with dose response curves is not seen here.
The major difference in response is seen in palbociclib, a CDK4/6 inhibitor. While OVCAR-8 experiences a slight drop in proliferation, OVCAR-3 increases in overall EdU positivity. This may be due to its cycE1 amplification; with a higher, more constant level of cycE1, the cell does not need the effects of CDK4/6 to activate its transcription through the RB pathway. This leads to continued proliferation as cycE1 and CDK2 come together to progress the cell through proliferation.
The DNA content of OVCAR-3 and OVCAR-8 cells vary. One variation occurs in PF-068700, where OVCAR-3 cells accumulate gradually to a bulk G2-state, while the OVCAR-8 cells acclimate mostly in a 2n state indicating pre-S phase.
Ongoing dose response curves are in place to optimize the data shown here; the integrated DNA data for the OVCAR-8 samples were clearly unimodal, which is not representative of cell populations including a 2n (pre-S Phase) group, and a 4n (post-S-Phase) group. This may be due to extreme cell death during treatment; further optimization of density and doses are being tested. Ongoing multiplexed immunofluorescence experiments are also being conducted to uncover specific molecular signatures of each of these response states. Using 40+ biomarkers to show various identifiers of cell state and phase, specific key molecules that qualify as candidates for targeted therapies may be uncovered. After identification of these features, modulation experiments will be performed to test potential combinations’ efficacy.
