p53 missense mutation enhances OVE growth in vitro
To generate a model of precursor HGSC cells, we used the previously described OVE4 and OVE16 cell populations that were independently isolated from the oviducts of female FVB/n mice19. To confirm the origin of these lines, expression of the secretory oviductal epithelial cell marker Pax8 was assessed by immunoblot (Fig. 1A). In both OVE cell lines, Trp53 was stably knocked out by CRISPR-Cas9 genome editing, and the human p53 missense mutant p53R175H was stably expressed by lentiviral transduction as verified by immunoblotting (Fig. 1B,C).
Aberrant proliferation is a prominent feature of HGSC precursor lesions. p53 signatures form from secretory cell outgrowth, and epithelial stratification arises in STICs4,5. To assess if Trp53 mutations increase proliferative potential in OVE cells, doubling time was calculated. When cultured in complete media, OVE4 and OVE16 monolayer cells expressing p53R175H exhibit modest decreases in doubling time compared to OVE4-Trp53ko and OVE16 parental cells, respectively (Fig. 1D). When cultured in nutrient-depleted conditions, both deletion of Trp53 and expression of p53R175H provide a growth advantage in OVE cells, with OVE16-p53R175H demonstrating higher proliferation than both OVE16 parental and OVE16-Trp53ko cells (Fig. 1E).
To further assess the transformed state of OVE cells, we measured colony formation in soft agar. The total number of colonies > 1000 µm2 was unchanged due to Trp53 deletion or expression of p53R175H (Fig. 1F-H). However, p53R175H enabled a drastic increase in colony growth in a small subset of colonies, demonstrated by in increase in colony size in OVE4-p53R175H compared to OVE4 and OVE4-Trp53ko cells (Fig. 1F,I). OVE16 and OVE16-Trp53ko cells possessed minimal colony formation with only 2 and 3 colonies meeting the size threshold, respectively, while OVE16-p53R175H cells demonstrated a trend towards increased colony size (Fig. 1F,J). Overall, p53R175H expression is promoting a more transformed in vitro phenotype in OVE cells.
p53R175H enhances OVE spheroid growth and viability
In order to colonize the ovary, STIC lesion cells must survive following detachment from the ovary. To recapitulate this event in our system, we seed single cells in flat-bottom or round bottom Ultra-Low Attachment (ULA) plates to generate bulk spheroids or single uniform spheroids, respectively (Fig. 2A). When cultured in flat-bottom ULA plates, OVE cells spontaneously form spheroids with varying size and morphology based on p53 status, with expression of p53R175H producing the largest spheroids for OVE4 and OVE16 cells (Fig. 2B). In agreement with this, OVE4-p53R175H and OVE16-p53R175H bulk spheroids possess increased viable cells at day 3 compared to parental and Trp53ko counterparts (Fig. 2C,D).
When cultured in round-bottom ULA plates, OVE4 cells with either Trp53 mutation formed smoother, more uniform spheroids compared to the parental line, which exhibits cells on the periphery that do not incorporate into the spheroids (Fig. 2E). In OVE16 cells, p53R175H expression produces much larger spheroids compared to parental and Trp53ko cells. Using the luminescence-based CellTiter-Glo viability assay, cell survival of single spheroids was measured over time. OVE4 cells had similar viability regardless of p53 status at day 5, but by day 10 spheroids with either Trp53 mutation had increased viability compared to parental cells (Fig. 2F). In OVE16 cells, p53R175H expression increased viability compared to parental cells across all time points. OVE16-Trp53ko spheroids had comparable viability to OVE16-p53R175H at day 3, but failed to maintain this viability at day 7 (Fig. 2G). Further interpretation of this data implies that cells within OVE16-p53R175H spheroids may be proliferating, as the relative number of viable cells increased over time.
Based on the morphology of spheroids formed in round-bottom ULA plates, it appears as though parental OVE4 spheroids are less compact compared with spheroids possessing p53 mutations. To address this phenotype directly, spheroid compaction was calculated by measuring viability of day-3 single spheroids and dividing by spheroid size. Surprisingly, OVE4-p53R175H spheroids had a decreased compaction score compared to OVE4 and OVE4-Trp53ko spheroids (Fig. 2H). With respect to spheroid cell viability, OVE4 and OVE16 spheroids with p53R175H expression produced similar phenotypes (Fig. 2C,D,F,G). However, OVE16-p53R175H spheroids had increased compaction compared to parental and Trp53ko spheroids (Fig. 2I), representing an opposite phenotype to that of OVE4 spheroids. These results suggest the ability to form compact spheroids is not required for maintaining viability.
Upon surviving detachment from the fallopian tube, spheroids must adhere to the ovary and invade the underlying stroma to form a primary tumour. To assess the ability of OVE spheroids to invade and grow in an extracellular matrix, pre-formed single spheroids were embedded in Matrigel and spheroid size was measured over two weeks. Spheroids formed by OVE4-p53R175 had increased size compared to parental and Trp53ko spheroids already at the time of adding Matrigel (Fig. 2J, K). While OVE4-p53R175H spheroids displayed modest growth, OVE4-Trp53ko spheroids increased in size nearly 4-fold by day 14. Parental OVE4 spheroids had limited growth. In OVE16 spheroids, both Trp53 deletion and expression of p53R175H increased spheroid size at day 14 compared to parental spheroids, with Trp53 deletion displaying rapid growth as early as day 4 (Fig. 2J, L). Interestingly, in addition to increased spheroid growth in Matrigel, OVE spheroids with Trp53 deletion or expression of p53R175H demonstrated a qualitative increase in clonogenicity. Upon addition of Matrigel to the spheroids, some of the cells at the periphery of the spheroid detached as single cells embedded in Matrigel. Parental cells that exfoliated from the spheroid arrested or died off, while either Trp53 mutation enabled some of these cells to form new colonies. This phenomenon occurred in only a subset of spheroids, but was exclusive to spheroids with Trp53 mutations (Supplemental Figure S1). Overall, Trp53 deletion and p53 missense mutation provide unique pro-tumorigenic spheroid phenotypes, with both types of mutations enhancing properties relevant to multiple steps in early HGSC development.
p53 mutations alter OVE sensitivity to carboplatin
To determine if p53 mutations desensitize OVE cells to a commonly used HGSC therapeutic, carboplatin dose-response curves were generated for adherent OVE4 and OVE16 cell lines (Fig. 3A, B). OVE16 lines had a similar IC50 regardless of p53 status, while OVE4 cells expressing p53R175H had decreased sensitivity compared to parental OVE4 cells (Fig. 3C, D). Next, OVE4 and OVE16 bulk spheroids were treated with their respective parental adherent carboplatin IC50, and spheroid cell viability was measured. Interestingly, spheroids formed by all 3 OVE4 lines showed no difference in viability between treated and untreated spheroids, despite altered sensitivity due to p53R175H in adherent culture (Fig. 3E). In contrast, OVE16-p53R175H spheroids treated with carboplatin had decreased viability compared to untreated spheroids (Fig. 3F). This is consistent with the proliferative phenotype observed in Fig. 2G, as carboplatin targets proliferating cells. Parental and Trp53ko OVE16 cells showed no difference in viability between treated and untreated spheroids.
Parental OVE spheroids have enrichment of apoptosis and immune-related pathways
To assess transcriptional dysregulation due to p53 mutation, we performed RNA sequencing on total RNA isolated from OVE spheroids. Principal component analysis (PCA) demonstrated tight clustering of technical replicates, validating the quality of the RNA sequencing data (Fig. 4A). Principal component 1, which accounts for 63.3% of variance in gene expression, clusters the samples based on their parental origin. Principal component 2 clusters samples based on the status of p53, but only accounts for 12.5% of variance. These data indicate that OVE4 and OVE16 cell lines with either p53 mutation are more similar to their parental cell line than they are to their respective OVE counterpart with the same p53 mutation.
To identify pathways altered by p53 mutation that may be driving observed transformation and spheroid phenotypes, we performed gene set enrichment analysis (GSEA). More specifically, we used the Hallmarks gene sets from the Molecular Signatures Database to identify p53-dependent alterations in well-defined gene signatures that are common to both OVE4 and OVE16 spheroids. Parental OVE spheroids displayed 12 gene sets enriched compared to Trp53ko spheroids (Fig. 4B) and 12 gene sets enriched compared to p53R175H spheroids (Fig. 4C). Interestingly, several immune-related gene sets were enriched in parental spheroids compared to spheroids with Trp53 deletion or expression of p53R175H (Supplemental Figure S2).
We decided to pursue the apoptosis hallmark further, as it was down in spheroids expressing p53R175H but not down due to Trp53 deletion, indicative of a potential GOF of p53R175H. Additionally, a reduced capacity to induce apoptosis may explain the observed increase in spheroid viability due to p53R175H, as a form of apoptosis called anoikis is activated by loss of cellular attachment20. OVE4 and OVE16 parental spheroids had normalized enrichment scores of 1.30 and 1.45, respectively, compared to p53R175H counterparts for the apoptosis hallmark (Fig. 4D). To validate the GSEA analysis, transcript expression was measured for genes that were driving enrichment of the apoptosis hallmark in parental spheroids compared to spheroids expressing p53R175H (Fig. 4E, F). Spheroids expressing p53R175H consistently had reduced expression of apoptosis-related genes compared to parental spheroids, while deletion of Trp53 resulted in variable expression including some genes with increased expression.
p53 is required for the intrinsic and extrinsic apoptosis pathway in OVE spheroids
Apoptosis can be activated intrinsically through mitochondrial outer membrane permeabilization (MOMP) or extrinsically through the binding of extracellular or cell membrane-bound ligands to death receptors21–23. To determine whether the intrinsic or extrinsic apoptosis pathway is contributing to the observed apoptosis phenotype, we assessed expression of proteins involved in each pathway. In the intrinsic pathway, MOMP is dependent on the relative abundance of two families of outer mitochondrial membrane proteins. Pro-survival Bcl-2 family proteins antagonize MOMP, while pro-apoptosis Bax family proteins promote MOMP. A third family of pro-apoptosis BH3-only proteins can bind to and sequester pro-survival Bcl-2 family proteins to increase the relative abundance of Bax family proteins. Based on transcript expression from the RNA-seq analysis, we looked at one example of each of the families of proteins involved in the intrinsic pathway (Fig. 5A, B). The BH3-only member Puma was increased in parental OVE4 and OVE16 spheroids compared to spheroids with Trp53 deletion or expression of p53R175H. With respect to Bax family member Bax and Bcl-2 family member Bcl-xl, there was no difference among OVE16 spheroids, while OVE4-p53R175H had increased expression of both compared to parental OVE4 spheroids. In the extrinsic pathway, expression of the Trail ligand was decreased due to Trp53 deletion and expression of p53R175H in OVE4 spheroids, with OVE4-Trp53ko spheroids displaying the lowest expression (Fig. 5C, D). In OVE16 spheroids, deletion of Trp53 resulted in decreased Trail expression compared to OVE16 and OVE16-p53R175H spheroids. Importantly, all OVE spheroids expressed the Trail receptor Trailr2.
The intrinsic and extrinsic pathway converge on cleavage and activation of Caspase-3 and Caspase-7, and subsequent cleavage of Parp. To assess if altered upstream intrinsic and extrinsic apoptosis signaling is influencing cleavage of downstream substrates, we measured their cleaved products (Fig. 6A, B). In OVE4 spheroids, expression p53R175H produced the lowest cleavage of Caspase-3, while Trp53 deletion resulted in increased cleavage compared to parental spheroids. OVE16 and OVE16-p53R175H spheroids had lower Caspase-3 cleavage compared to OVE16-Trp53ko. Caspase-7 cleavage demonstrated similar trends to Caspase-3, with Trp53 deletion showing increased cleavage. Parp cleavage was similar among OVE4 spheroids, while Trp53 deletion produced the highest cleavage among OVE16 spheroids. To validate that increased caspase cleavage is facilitating increased caspase activity, we measured cleavage of the caspase consensus sequence DEVD in OVE spheroids. OVE4-p53R175H spheroids had reduced caspase activity compared to OVE4 parental and OVE4-Trp53ko spheroids (Fig. 6C). In OVE16 spheroids, expression of p53R175H decreased caspase activity compared to deletion of Trp53ko (Fig. 6D). Overall, expression of p53R175H is blocking apoptosis to a higher degree than Trp53 deletion.