Mutant p53 disrupts antioxidant defense in fallopian tube epithelium via GSTAs suppression: A pathway to serous tubal carcinogenesis.

We assessed three different TMAs of HGSOC cases using image analysis of digital images (Aperio Spectrum software, Fig 1A ). Tissue cores were annotated to exclude tumor stroma from the analysis. Based on output data, the majority (150/162, or 92.6%) of HGSOC cases analyzed had lower expression of GSTA2 than the normal FTE expression (1.87%–34% of GSTA2+ cells) (p<0.0001) ( Fig 1B – D ). In the TCGA HGSOC data set, only 2.5–7% (n = 316) of the cases had a decrease or increase in mRNA levels of the GSTA1/2/3/4/5 genes (TCGA 2011) ( Supplementary Figure 1 ). In 26% of HGSOC cases, a shallow deletion, possibly a heterozygous deletion, at this genomic locus indicated that the overall decrease in RNA expression did not account for genomic loss at this locus. As TP53 is mutated in 95–99% of HGSOC and is found early in pre-neoplastic, histologically normal cells ( Kuhn, Kurman et al., 2012 , Salani, Kurman et al., 2008 ), GSTA2 protein expression was also analyzed in STIC, to determine how early expression of the protein is impacted. All 15 STIC cases studied showed loss of GSTA2 expression ( Fig 1E ). In the STIC cases shown, GSTA2 expression is lower in cells expressing p53 and MIB1 (Ki67). In FTE, STIC and HGSOC, overexpression of p53 reflects dominant negative p53 mutations. As shown in an example ( Fig 2A ), GSTA2 expression occurs in an intermittent pattern along the layer of fimbria epithelia, corresponding to a typical distribution of ciliated cells among the secretory FTE cells. To determine if GSTA2 expression is low because of inherently low expression in secretory epithelial cells or from aberrant p53 expression, FTE cells (4364-FTES and 4384-FTES) were isolated and examined for GSTA2 and p53 expression ( Fig 2B ). Primary FTES cells express GSTA2 but not p53. Transfection of cells with a gain of function p53 mutant expression construct (p53-R249S) resulted in decreased GSTA2 levels. In comparison, GSTA2 expression is greater in a cilia-enriched FTE cell line (5002-FTE) compared to a secretory FTE cell line (4364-FTES) ( Supplementary Figure 2 ). However, neither normal FTE ciliated nor secretory cells exhibit positive p53 expression that indicates a stabilized mutant p53 protein, as shown for 5002-FTE, 4364-FTES, and 4387-FTES ( Supplementary Figure 2 ) In the 4364 FTE-p53R249S cells and established cancer cell line (OCI-P5x), GSTA2 levels were significantly decreased compared to the whole fallopian tube lysate (5002) and normal FTES cells ( Fig 2B and Supplementary Figure 2 ). Moreover, γH2Ax (Ser139) protein levels were elevated in the FTE-p53R249S lines ( Fig 2B ), indicating increased DNA damage To assess whether the γH2Ax foci formation was a result of increased ROS in the FTE, we measured both basal and ROS induced by H 2 O 2 (50 μM) in the FTE cell lines ( Fig 2C ). The basal level of ROS was significantly higher in the FTES-p53R249S cells compared to the FTES. Upon H 2 O 2 treatment, ROS levels rose significantly in the FTES-p53R249S cells compared to the FTES ( Fig 2C ). We next assessed the DNA damage repair response in two independent FTE cell lines (n=3 replicates each) by treating them with H 2 O 2 (50 μM) for 10 mins. Cells were then washed and assessed for γH2Ax foci formation. Shortly after DSB induction, both the normal FTES and FTES-R249S cell lines exhibited a significant increase in γH2Ax foci formation (p=0.001) and demonstrated an increase in γH2Ax foci in two different FTE cases analyzed ( Fig 2D ). In the untreated FTES-p53R249S lines, cells had more γH2Ax foci than both FTES cells alone (4364-FTES, p=0.0072; 4392-FTES, p=0.0370). The differences between FTES and FTES-shP53 (p53 downregulated) lines did not reach statistical significance (4364-FTES, p=0.8782; 4392-FTES, p=0.7376) ( Fig 2D , 2E ). Statistical differences were also not found between FTES-shP53 and FTES-p53R249S cell lines, indicating that DSB repair response is activated independently of the type of p53 mutation (gain of function (GOF) or loss of function). Changes in damage response proteins in FTES cells with and without p53 GOF were examined by immunoblotting, which clearly showed a basal increase of γH2Ax (Ser139) in FTES-p53R249S cells. ATM kinase (a known redox sensor) mediates γH2Ax phosphorylation. Phosphorylated (phospho)-ATM (Ser1981) in FTE cells harboring p53R249S increased in response to DSB induction by H 2 O 2 ( Fig 2F ). In addition, checkpoint kinase 2 (Chk2), a serine/threonine kinase and tumor suppressor activated by double-strand breaks, is phosphorylated (Thr68) in FTE cells with p53R249S, signifying that these cells have chronic and possibly unrepairable DSB. Both basal and H 2 O 2 -induced ROS were significantly higher in FTES-p53R249S compared to FTES cells for FTE cases 4432, 4387, and 4392 ( # 4387-FTES vs 4387-FTES (+50 μm H 2 O 2 , p=0.0355, & 4392-FTES vs 4392-FTES (+50 μM H 2 O 2 , p=0.0293)), *p<0.05, **p<0.005, ****p<0.0001) ( Fig 2C ). The relative fluorescence units (RFU) were also measured for 4392-FTES, 4387-FTES and 4387-p53R249S untreated and treated with 50 μM of H 2 O 2 . A significant difference in RFU was measured between H 2 O 2 -untreated cell lines, with higher levels of RFU being observed in cell lines in the treatment condition and with the p53R249S mutation. These findings suggest that the presence of the p53 mutation, along with induced DNA damage via ROS, increases the susceptibility of cells to oxidative stress and dysregulation of cellular mechanisms that protect against disease pathogenesis. Cell cycle analysis performed using BrdU and PI demonstrated that control cells (FTES-33991) were primarily found in the G1-phase (55.7% of the cell population). A p53DN R175H mutation and knockdown of p53 decreased the proportion of cells in the G1 phase to 43.9% and 47.2%, respectively. These results demonstrate that GSTA2 protein decreased independently of p53 mutation type (R175H, R249S, or shP53). Furthermore, knockdown of GSTA2 decreased the proportion of cells in the G1 phase to 44.6% ( Supplementary Fig 3 ). The decrease in GSTA2 protein expression in primary FTE cells from normal FTE to cancer correlates with results from immunohistochemistry studies. Cell cycle analysis performed using BrdU and PI demonstrated that control cells (FTES-33991) were primarily found in the S-phase (61.4% of cell population). A p53DN R175H mutation or knockdown of p53 decreased the proportion of cells in the S phase to 10.8% and 10.4% respectively. Furthermore, knockdown of GSTA2 decreased the proportion of cells in the S phase to 14.1%. All cells are thought to have inherent and robust safeguard mechanisms which negate the deleterious effects of ROS. Phase I genes produce proteins that oxidize xenobiotic species, whereas phase II genes produce proteins that conjugate the species created from phase I reactions ( Justenhoven 2012 ). Oxidative stress is a known contributor to cancer development due to ROS impacting normal cellular processes. Further, using the FTE and HGSOC dataset ( GSE28044 ) with luteal and follicular phases included in the metadata, we observe variable expression of GSTA 1–4 in histologically normal FTE and HGSOC ( Supplementary Figure 4 ). We therefore analyzed mRNA expression of several antioxidant enzymes involved in phase I and phase II defenses in the FTE cells pre- and post- H 2 O 2 treatment ( Fig 3A , Supplementary Figure 5A ). The qPCR analyses revealed that FTES-p53R249S cells have markedly decreased antioxidant response to H 2 O 2 compared to FTES with wild type p53 proteins ( Fig 3B , Supplementary Figure 5B ). Next, we studied gene homologs of GSTA2 to understand the full extent of antioxidant gene response in the presence of ROS-induced DNA damage. GSTA2 is a member of a 22-gene superfamily consisting of 7 homologs that are all clustered in tandem on chromosome 6p12.2 and play an essential role in functional antioxidant response. ROS damage was induced with 50 μm H 2 O 2 and antioxidant homolog expression was compared between the treatment and non-treatment conditions. The expression levels of antioxidant homolog genes SOD1, SOD2, GSTA1, GSTA2, GPX2, and GPX3 were compared in FTES and FTES-p53R249S in the treated and untreated H 2 O 2 conditions. The 4364-FTES cell line demonstrated a significant increase in mRNA levels of the phase I and II antioxidant genes post H 2 O 2 treatment. Only GSTA1 showed a significant increase in FTES-p53-R249S. The qPCR analyses revealed that FTES-p53R249S cells have markedly decreased antioxidant response to H 2 O 2 compared to FTES with wild type p53 proteins ( Fig 3A – B ). Additionally, GPX2 and GSTA1 showed a 6000- and 20,000-fold change response, respectively, in wild type FTES compared to other glutathione peroxidases and glutathione-S-transferases. SOD1 and SOD2 also showed significant up-regulation in both FTES (p<0.0001) and FTES-p53R249S cell lines (p=0.001). FTES-R249S cells in general showed a muted antioxidant response of both phase I (SODs) and phase II enzymes (GPXs and GSTAs) once challenged by H 2 O 2 ( Fig 3B ). IF depicted the expression of γH2Ax and acetylated tubulin of normal case 5217 FTE both untreated and treated with 50 μM of H 2 O 2 on transwell filters. The IF confirms a marked increase of gH2Ax foci after H 2 O 2 treatment. Non-ciliated cells had a greater number of gH2Ax foci compared to ciliated cells. Our previous study determined that acetylated-tubulin and GSTA2 are predominantly expressed in ciliated epithelial cells ( Sowamber, Nelson et al. 2020 ) ( Fig 3C ). These results suggest that phase I and phase II antioxidant mechanisms are present in the FTES with either mutated or non-mutated p53. The results further suggest an essential role of these genes in the pathogenesis of HGSOC and the effect of mutated p53 in dysregulating the antioxidant response in the FTE.

The study protocol for collection of tissue and clinical information for all patients was approved by the University Health Network (UHN) Research Ethics Board and the IRB of the Miller School of Medicine at the University of Miami. All tissue donors provided written informed consent authorizing collection and use of tissue for research purposes. For primary cultures, donors were consented through the UHN Biobank and the Biospecimen Shared Resources (BSSR) at the Sylvester Comprehensive Cancer Center (SCCC). Three types of tissues were used: 1) histologically normal FTE, 2) FTE and HGSOC tissue microarrays (TMAs), and 3) a curated cohort of FTs with STIC lesions as described previously ( George, Greenaway et al. 2011 , Milea, George et al. 2013 , George, Milea et al. 2015 ). In brief, fimbriae and ampullae were collected after salpingectomy and incubated for 24–48 hours at 4°C in pronase and DNase as previously described ( Karst and Drapkin 2012 , George, Milea et al. 2015 , Sowamber, Nelson et al. 2020 ). Information about the cases is provided in Supplementary Table 1 . FTE cells were propagated on collagen IV (60 μg/ml) coated plates in DMEM-F12 (1:1) with 2% Ultroser G (Crescent Chemical). To generate FTE cells with mutations in or downregulated for p53 and downregulated for GSTA2, cells were infected with one of the following vectors according to published protocols ( Karst and Drapkin 2012 ): lentiviral dominant negative p53-R249S, p53-R175H, p53-R273H, or shP53 vector (Addgene #19119, 22936, 22934, 22935); ; or shGSTA2 (GE Dharmacon; RHS4533-EG2939). Briefly, a confluent well of FTE cells in a 24-well dish was trypsinized and cells were resuspended in 2 ml of FTE media, then transferred to a single well in a 12-well collagen-coated plate. Polybrene was added to 8 μg/mL and sufficient vector was added to achieve 2.5×10 5 infectious units/well. The plate was centrifuged at 1100×g for 30 min at 37 °C, then incubated overnight, with medium replaced with fresh FTE medium after 72 hours. Positively transfected cells were selected with the appropriate selection agent according to the vector supplier’s instructions. Cells were cultured on filters as described previously ( Karst and Drapkin 2012 , George, Milea et al. 2015 , Sowamber, Nelson et al. 2020 ) . The OCI-P5x cells (ATCC #CRL-3552) are a human ovarian cancer cell line originally derived from a papillary serous adenocarcinoma and were cultured in complete OCMI-L medium (US Biological Life Sciences #506390) containing 5% FBS and 25 ng/ml cholera toxin. Cells were grown on 8-well Lab Tek II chamber slides coated with collagen IV and fixed in 4% paraformaldehyde (PFA) for 5 minutes, permeabilized with 0.3% Triton-X/PBS, then blocked with 5% goat serum (Gibco). The fixed cells were incubated overnight at 4°C with the following primary antibodies: Bcl2 (Abcam ab692), CK18 (Dako M7010), Pax8 (ProteinTech 10336-1-AP), acetylated-tubulin (Sigma T6793), UBF (Santa Cruz sc-13125), GSTA2 (Abcam ab199115), γH2Ax (R&D Systems MA1-2022), vimentin (Abcam ab24525), or DNA Damage antibody kit (Cell Signaling #9947). Samples were visualized with appropriate fluorophore-labeled secondary antibodies (Jackson ImmunoResearch, R&D Systems). Sections (5 μm) of the profiled formalin-fixed, paraffin-embedded (FFPE) specimens were stained for Ki67, p53, CK7, and GSTA2 using standard procedures ( Sowamber, Nelson et al. 2020 ). The following antibodies were used: Ki67 (Novocastra MIB1, 1:100), p53 (Abcam pAb-1801, 1:250), and GSTA2 (Abcam ab199115, 1:100). Appropriate negative and positive controls were performed to determine the specificity of antibodies. Once stained, all slides were scanned at 40× magnification using the ScanScope XT slide scanner to create digital images. Images were annotated to include only the epithelial components of all cores and an automated image analysis program employing a nuclear algorithm (Spectrum Plus, Aperio, Inc.) was used to quantify the intensity of staining and percentage of cells positive ( George, Greenaway et al. 2011 , Milea, George et al. 2013 , Sowamber, Nelson et al. 2020 ). For each case, there were two independent regions analyzed. GSTA2 analysis results that included intermediate and high intensity staining (+2, +3) were categorized as positive, and no and low staining (0 to +1) were classified as negative. Individual patient samples (uninfected FTE secretory cells (FTES), FTE cells infected with shP53, TP53 R249S, and/or R273H) were plated at 4×10 3 cells per well in an 8-well Lab Tek II chamber slide previously coated overnight with collagen IV. After incubating cells overnight, FTES were incubated with 50 μM H 2 O 2 for 10 mins at room temp, 100 μl of pre-warmed media plus 100 μl of H 2 O 2 (50 μM), or media only for control wells. After 10 mins of H 2 O 2 treatment, the media was removed and cells were washed once with fresh media, then 200 μl of media was added per well and the cells were allowed to recover for 45 mins in the incubator at 37°C. FTE cells were washed once with 1× PBS and fixed for γH2Ax immunofluorescence staining as described (above). A Zeiss Axio Observer Z1 microscope was used to obtain images of FTE at 63×. ImageJ was used to count γH2Ax positive and negative cells. Total γH2Ax positive and negative FTE cells were expressed as a percentage of the total cell count for treated and untreated cells. Total RNA was extracted using the TRIzol ® Reagent (ThermoFisher 15596-026) according to manufacturer’s instructions. RNA (1 μg) was reverse transcribed to cDNA using the QuantiTect Reverse Transcription Kit (Qiagen 205311). RT-qPCR was performed using Qiagen QuantiNova SYBR Green PCR Kit (Qiagen 208054) and CFX96 Touch Real Time System (BioRad 1855195). Each experiment was performed in triplicate with two independent patient derived cells. Results were analyzed using the comparative Ct method normalizing to a control sample and housekeeping primers GAPDH and β-actin ( Supplementary Table 2 ). Two-way ANOVA was performed using GraphPad Prism version 10 for Mac, GraphPad Software, (La Jolla California USA, www.graphpad.com ). FTE cells were cultured in triplicate per cell type in a BD Falcon 96 well flat bottom black polystyrene plate (Corning, cat. 353219) overnight at 5 × 10 4 /100 μl per well in DMEM-F12 1:1 with 2% Ultroser G (Crescent Chemical). FTE cells were washed once with PBS and incubated with 100 μl of ROS Red working solution (Abcam ab186027) for 1 hour at 37 °C/5% CO 2 . FTE cells were incubated with 50 μM H 2 O 2 for 10 mins at room temperature to induce ROS and then the fluorescence intensity change was measured at Ex/Em = 520/605 nm on an Infinite 200Pro multimode plate reader (Tecan). FTE and ovarian cancer cell lines were lysed in RIPA buffer (Pierce #89900) supplemented with complete protease inhibitor (Roche #04693116001) and phosphatase inhibitor (Roche #04906837001) and allowed to rotate at 4 °C for 30 mins to allow for complete lysis. Cell lysates were then centrifuged at 14K rpm for 10 mins at 4 °C. Total protein concentration was determined by the Bradford protein assay; absorbance was measured using a Bio-Tek μQuant spectrophotometer at 750 nm. For standard immunoblots, 10–20 μg of protein was loaded. Primary antibodies were diluted at 1:500 or 1:1000. The following antibodies were used: Pax8 (ProteinTech 10336-1-AP), GSTA2 (Abcam ab199115), γ H2Ax (R&D Systems MA1-2022, 967116), p53 (Abcam ab1101), pATM (Cell Signaling #5883), pCHK2 (Cell Signaling Technology (CST) #2187), and BRCA1 (Santa Cruz sc642). Secondary antibodies were diluted at 1:5000 (CST #7076 and #7074). Actin-conjugated HRP (1:2000; Santa Cruz sc-1615) or monoclonal beta-actin (Sigma #A228) was used as a loading control. For assessment of cell cycle regulation, cells were washed in 1× PBS and 10 mM BrdU-APC was added to cells in a dark environment. Plates with BrdU were then incubated at 37 °C for 4 hours. Cells were then washed twice with ice cold PBS, trypsinized with 0.25% trypsin dissociation reagent, and neutralized for counting. Nuclei preparation and staining were performed by adding 0.08% Pepsin and 2N HCl and IFA/0.5%Tween20. Samples were incubated in the dark and 100 ml anti-BrdU-APC was added while samples were incubated on ice. Propidium iodide (5 mg/ml) was added to cells and incubated on ice for 15 min. Flow cytometry was carried out on BD FACSCalibur (BD Biosciences). Data was analyzed using FlowJo v10.

The fallopian tube is well accepted as the predominant site of origin of high-grade serous carcinoma. Precursors of HGSOC, which include the p53 signature and STIC, a direct precursor of HGSOC, occur most frequently at the distal fimbriated end of the fallopian tube (FTE), the region of the tube most directly exposed to the events of ovulation. Therefore, the fimbria is considered the high-risk epithelial zone prone to transformation and the ampulla is the low-risk epithelial zone. Significant epidemiological data links a higher number of ovulatory cycles with risks to developing ovarian cancer ( Gaona-Luviano, Medina-Gaona et al. 2020 ). The ovaries produce significant reactive oxygen species (ROS) after the luteinizing surge to induce ovulation ( Shkolnik, Tadmor et al. 2011 ). The incessant ovulation hypothesis described by Fathalla suggests that continuous cycles of chemokines, cytokines, and sex hormone production, along with repeated rupturing of the ovarian surface epithelia, are responsible for ovarian cancer formation ( Fathalla 2013 ). As a result, fallopian tube cells must have robust mechanisms to control cellular redox reactions to protect their macromolecules from damage. GSTA2 is a phase II antioxidant enzyme that detoxifies electrophilic compounds such as carcinogens, therapeutic drugs, environmental toxins, and reactive oxygen species, through conjugation with glutathione (GSH) ( Singh and Reindl 2021 ). As a member of the glutathione-S transferase phase II enzyme family, GSTA2 showed distinct and unique expression in ciliated cells and acetylated tubulin-positive FTE cells compared to secretory cells. In vitro and ex vivo fallopian tube cells showed a relationship between GSTA2 protein expression and γ H2Ax activity when exposed to DNA damaging agents. qRT-PCR on RNA from independent normal FTE cells confirmed enhanced expression of the GST-genes (GSTA1–5) when exposed to H 2 O 2 , but expression was suppressed in the presence of mutant p53. The relationship between GST-gene expression and p53 has not yet been resolved. Interestingly, based on our results by immunoblot, Pax8 expression levels after treatment with H 2 O 2 are decreased ( Fig 2F ). Kambe et al. documented that Pax8 protein is oxidized at cysteine 31 in the DNA binding domain in the presence of diamide Pax8 was also shown to be regulated by redox system thioredoxin ( Kambe, Nomura et al. 1996 , Marshall, Merchant et al. 2000 ), implicating Pax8 activity in oxidative/redox stresses in the FTE. SEPW1, a gene involved in oxidation-reduction reactions, was identified as a Pax8 target ( Adler, Corona et al. 2017 ). Our analysis revealed differential transcriptome profiles of GSTA2 in normal FTE and cancer FTE and within the low-risk and high-risk zones of the fallopian tube. GSTA2 was significantly upregulated in the high-risk fimbria when compared to the ampulla. Further supplementary data of RNA expression for other GSTs in FTE and HGSOC as well as data from TCGA show several of the family members show lower overall expression like GSTA4 as well as shallow deletions. Understanding its essential role in the quenching of xenobiotic factors and electrophilic species, along with its deregulation in cancer cells, allows for a hypothesis to be made about its function against ROS in cancer FTE. The loss of GSTA2 and its homologs transcription within the microenvironment of the high-risk distal end of the fallopian tube provides an opportunity for precancerous STIC lesions to undergo malignant transformation into HGSOC. Additionally, decreased GSTA2 expression yields sustained exposure to and loss of protection against ROS, such as chemokines, resulting in further damage to DNA, proteins, and lipids. Conditions of increased exposure to ROS are especially detrimental in the context of BRCA1/2 mutation carriers with loss-of-function mutations in genes responsible for genomic stability (Zámborszky et al. 2016). The presence of the BRCA1/2 mutation along with an inhibitory p53 and GSTA2 mutation leaves the opportunity for STIC lesions to undergo metastatic transformation, leading to HGSOC.

The persistent oxidative stress experienced by the fimbria of FTE cells due to ovulation requires a robust antioxidant response in this region to prevent subsequent DNA damage and eventual transformation. This response is compromised in STIC and HGSOC, as demonstrated with a decrease in GSTA2 expression. Our findings show that mutant p53 is a contributory mechanism by downregulating GSTA2 and the family of genes, expression and impairing the DNA damage response. These changes increase the susceptibility of FTE cells to oxidative stress and therefore represent some of the earliest events before malignant transformation. Further understanding of this mechanism is a potential target for interventions aimed at prevention in patients with high-risk for ovarian cancer disease.

High grade serous ovarian carcinoma (HGSOC) is the most common subtype of epithelial ovarian cancer (EOC). HGSOC presents at advanced stages in women and is characteristically aggressive, leading to lower rates of survival in those diagnosed with the disease. The early precursor lesions are still poorly described, yet the site of origin of HGSOC is likely the fimbria of the fallopian tube epithelium (FTE) ( Perets, Wyant et al. 2013 , George and Shaw 2014 , Labidi-Galy, Papp et al. 2017 , Shih, Wang et al. 2021 ). The pre-cancerous p53 signature, which is defined as strongly positive p53 nuclear immunostaining in a consecutive section of otherwise benign appearing epithelia ( Folkins, Jarboe et al. 2008 ), is predominantly found in the fimbria and thought to be a predecessor to the neoplastic lesions known as serous tubal intraepithelial carcinoma (STIC) lesions. The FTE are exposed to sustained cellular stress from the presence of reactive oxygen species (ROS) and ovulation-induced hormonal cycling. Stressors have the potential to cause an unregulated cell cycle, irregular metabolism, and increased exposure to inflammatory signatures ( George, Milea et al. 2012 , Fathalla 2013 , George and Shaw 2014 ). Conditions of increased ROS further affect the normal function, survival, and growth of epithelial cells and provide the opportunity for cancerous cells to proliferate via the evasion of cell death programs. Further, deficiencies in the double strand break (DSB) repair pathway are implicated early in the development of HGSOC ( Levanon, Ng et al. 2009 ). TP53 is mutated in 95–99% of HGSOC and is found early in pre-neoplastic, histologically normal cells ( Salani, Kurman et al., 2008 , Cancer Genome Atlas Network 2011 , Kuhn, Kurman et al., 2012 ). TP53 is considered the surveyor of the genome, while BRCA1 and BRCA2, tumor suppressors, are essential in the DSB repair pathway and act to maintain genome stability by resisting changes caused by DNA damage. BRCA1/2 pathogenic mutations are implicated in HGSOC, and the BRCA1/2 genes are deficient in 20–22% of HGSOC tumors ( Norquist, Harrell et al. 2015 , Norquist, Brady et al. 2018 ). This causes a loss of their highly conservative method of HR and increased errors in DNA repair. The presence of these mutations and aberrant p53 function mediate transformation of FTE cells. Glutathione S-transferase A1 and Glutathione S-transferase A2 (GSTA1 and GSTA2) are two enzymes that quench ROS and are differentially expressed between normal fimbria and ampulla of the fallopian tube (FT) ( Sowamber, Nelson et al. 2020 ). Given that ovulation includes a burst of follicular fluid, filled with chemokines, cytokines, and hormones such as estrogen, progestin, and androgens, and GSTAs are highly expressed in the high-risk zone of FT, we posit that these enzymes play an important role in ROS homeostasis, and that this process is implicated in serous carcinogenesis. The combination of mutant p53-driven dysregulation of GSTA and the increase in ROS primes the FTE cells for transformation. Here, we studied the expression of GSTA2 in the earliest identifiable precursor lesions (the p53 signature and STIC) and established HGSOC and modeled the expression in primary FTE cell lines to identify molecular effects that drive early changes in the FTE towards HGSOC pathogenesis.

Supplementary Table 1: Characteristics of donors of fallopian tubes used for in vitro experiments. Supplementary Table 2 : Primer sequences obtained from ( Dannenmann, Lehle et al. 2015 ) Supplementary Figure 1. Data retrieved from cBioPortal ( Cerami, Gao et al. 2012 ) for the Ovarian Serous Cystadenocarcinoma data set (N = 316) demonstrated that mRNA levels of GSTA 1–5 are significantly reduced in HGSOC ( http://bit.ly/2l6pL0s ). The TCGA HGSOC data represents an increase and decrease of mRNA levels of GSTA 1/ 2 / 3 / 4 / 5 genes. In the TCGA HGSOC data set, 2.5–7% of the cases had a decrease or increase in mRNA levels of the GSTA1/2/3/4/5 genes. 26% of HGSOC cases showed a shallow deletion at this genomic locus indicating that the overall decrease in RNA expression did not account for genomic loss at this locus TCGA HGSOC data represents an increase and decrease of mRNA levels of GSTA 1–5 genes. In the TCGA HGSOC data set, 2.5–7% of the cases had a decrease or increase in mRNA levels of the GSTA1/2/3/4/5 genes. 26% of HGSOC cases showed a shallow deletion at this genomic locus indicating that the overall decrease in RNA expression did not account for genomic loss at this locus ( http://bit.ly/2l6pL0s ). TCGA HGSOC data represents an increase and decrease of mRNA levels of GSTA 1–5 genes. In the TCGA HGSOC data set, 2.5–7% of the cases had a decrease or increase in mRNA levels of the GSTA1/2/3/4/5 genes. 26% of HGSOC cases showed a shallow deletion at this genomic locus indicating that the overall decrease in RNA expression did not account for genomic loss at this locus ( http://bit.ly/2l6pL0s ). Supplementary Figure 2. Mutant p53 downregulates GSTA2 and impairs the DNA damage responses in different primary FTES lines with A. modification of p53 using different mutations and B . GSTA2 expression in primary (5002) and p53 mutant (4364) and ovarian cancer cell lines. Supplementary Figure 3. GSTA2 modulates cell cycle position independently of p53 expression. A . Immunoblot analysis of primary fallopian tube cells with either a p53 dominant negative mutation at R175H (p53DN R175H) or a p53 gene knockdown (shp53). 33991-FTES, 33991-p53R175H, 33991-shP53, 34778-FTES, 34778-p53R175H, and 34778-shP53 cells demonstrating changes in GSTA2 expression. B . Cell cycle analysis by FACS in 33991-FTES, 33991-p53DN R175H, 33991-shP53, and 33991 shGSTA2_1 cells. C. Cell cycle analysis by FACS in 34778-FTES, 34778-p53DN R175H, and 34778-shP53 cells. Supplementary Figure 4. GSTA 1–4 expression varies among normal FTE cells in the luteal and follicular phases and HGSOC cells. The relative expression of Phase I and II antioxidant genes was examined in normal FTE cells in the luteal or follicular phases and in HGSOC. A. GSTA gene family, GSTA 1–4; B. GSTP and GSTM 1, 3, 5 genes; C. GPX 2, 3, 7 genes; and D. SOD1 and SOD2. Supplementary Figure 5 : A-B . qPCR data showing fold changes in the mRNA levels of SOD1, SOD2, GSTA1, GSTA2, GPX2 and GPX3 in 4387-FTES and 4387-FTES-p53R249S with and without H 2 O 2 treatment. There were significant (p<0.05) mRNA increase of phase I and II antioxidant genes in FTES following H 2 O 2 treatment, only GSTA1 showed a significant increase in FTES-p53R249S. C. IF of ƴ-H2Ax in normal case 4392-FTES, untreated and treated with 50μM of H 2 O 2 . There were more ƴ-H2Ax foci after H 2 O 2 treatment and in proximity to GSTA2 expression within the nucleus.