Oxidative stress contributes to hepatocyte growth factor-dependent pro-senescence activity of ovarian cancer cells

Author’s Accepted Manuscript Oxidative stress contributes to hepatocyte growth factor-dependent pro-senescence activity of ovarian cancer cells Justyna Mikuła-Pietrasik, Paweł Uruski, Martyna Pakuła, Konstantin Maksin, Sebastian Szubert, Aldona Woźniak, Eryk Naumowicz, Dariusz Szpurek, Andrzej Tykarski, Krzysztof Książek

PII: DOI: Reference:

www.elsevier.com

S0891-5849(17)30664-0 http://dx.doi.org/10.1016/j.freeradbiomed.2017.06.015 FRB13369

To appear in: Free Radical Biology and Medicine Received date: 5 February 2017 Revised date: 14 June 2017 Accepted date: 23 June 2017 Cite this article as: Justyna Mikuła-Pietrasik, Paweł Uruski, Martyna Pakuła, Konstantin Maksin, Sebastian Szubert, Aldona Woźniak, Eryk Naumowicz, Dariusz Szpurek, Andrzej Tykarski and Krzysztof Książek, Oxidative stress contributes to hepatocyte growth factor-dependent pro-senescence activity of ovarian cancer cells, Free Radical Biology and Medicine, http://dx.doi.org/10.1016/j.freeradbiomed.2017.06.015 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Oxidative stress contributes to hepatocyte growth factor-dependent pro-senescence activity of ovarian cancer cells

Justyna Mikuła-Pietrasik1, Paweł Uruski1, Martyna Pakuła1, Konstantin Maksin2, Sebastian Szubert3, Aldona Woźniak2, Eryk Naumowicz4, Dariusz Szpurek3, Andrzej Tykarski1, Krzysztof Książek1*

1

Department of Hypertensiology, Angiology and Internal Medicine, Poznań University of

Medical Sciences, Długa 1/2 Str., 61-848 Poznań, Poland. E-mails: [email protected] (J.M-P); [email protected] (PU); [email protected] (MP; [email protected] (AT); [email protected] (KK) 2

Department of Clinical Pathology, Poznań University of Medical Sciences,

Przybyszewskiego 49 Str., 60-355 Poznań, Poland. E-mails: [email protected] (KM), [email protected] (AW) 3

Division of Gynecological Surgery, Poznań University of Medical Sciences, Polna 33 Str,

60-535 Poznań, Poland. E-mails: [email protected] (SS); [email protected] (DS) 4

General Surgery Ward, Medical Centre HCP, 28 czerwca 1956 r. 223/229 Str., 61-485

Poznań, Poland. E-mail: [email protected] (EN)

Running title: Ovarian cancer cells induce senescence of HPMCs

*Correspondence: Prof. Krzysztof Książek, Department of Hypertensiology, Angiology and Internal Medicine, Poznań University of Medical Sciences, Długa 1/2 Str., 61-848 Poznań, Poland, Phone +48 61 854-92-99, Fax: +48 61 854-90-86, E-mail: [email protected] 1

Abstract The cancer-promoting activity of senescent peritoneal mesothelial cells (HPMCs) has already been well evidenced both in vitro and in vivo. Here we sought to determine if ovarian cancer cells may activate senescence in HPMCs. The study showed that conditioned medium (CM) from ovarian cancer cells (OVCAR-3, SKOV-3, A2780) inhibited growth and promoted the development of senescence phenotype (increased SA-β-Gal, γ-H2A.X, 53BP1, and decreased Cx43) in HPMCs. An analysis of tumors isolated from the peritoneum of patients with ovarian cancer revealed an abundance of senescent HPMCs in proximity to cancerous tissue. The presence of senescent HPMCs was incidental when fragments of peritoneum free from cancer were evaluated. An analysis of the cells’ secretome followed by intervention studies with exogenous proteins and neutralizing antibodies revealed hepatocyte growth factor (HGF) as the mediator of the pro-senescence impact of the cancer cells. The activity of cancerous CM and HGF was associated with an induction of mitochondrial oxidative stress. Signaling pathways involved in the senescence of HPMCs elicited by the cancer-derived CM and HGF included p38 MAPK, AKT and NF-B. HPMCs that senesced prematurely in response to the cancer-derived CM promoted adhesion of ovarian cancer cells, however this effect was effectively prevented by the cell protection against oxidative stress. Collectively, our findings indicate that ovarian cancer cells can elicit HGF-dependent senescence in HPMCs, which may contribute to the formation of a metastatic niche for these cells within the peritoneal cavity.

Key words: cellular senescence; hepatocyte growth factor; mesothelial cells; ovarian cancer; peritoneum

2

Introduction The role of human peritoneal mesothelial cells (HPMCs) in the intraperitoneal development of ovarian cancer metastases is indisputable [1,2]. The significance of these cells increases even more when they become senescent. It has been shown that senescent HPMCs which accumulate in the peritoneal cavity [3] stimulate various aspects of ovarian cancer cell progression in vitro and support the formation of solid ovarian tumors in mouse peritoneum in vivo. Mechanistically, the cancer-promoting activity of senescent HPMCs has been found to be associated with p38 MAPK-dependent intensification of the production of certain surfaceassociated and soluble proteins which support cancer cell adhesion [4], proliferation, migration, and invasion [5], as well as tumor neoangiogenesis [6]. Taking into account the explicit tumor-promoting activity of senescent HPMCs, we designed a project based on the hypothesis that ovarian cancer cells colonizing the peritoneal cavity may stimulate senescence in nearby normal cells (HPMCs) and thus create a permissive microenvironment for their further growth. This idea is based on the results of Yang et al., who showed that ovarian cancer cell-derived growth-related oncogene 1 (GRO-1) may induce p53-dependent senescence in fibroblasts adjacent to the cancer cells. Significantly, fibroblasts senesced in response to GRO-1 developed the capacity to facilitate tumor expansion – a which feature that disappeared when they became immortalized [7]. Here we show that ovarian cancer cells are capable of inducing senescence in HPMCs in a mechanism associated with the activity of hepatocyte growth factor (HGF). Mechanistically, we show the role of HGF-dependent oxidative stress and identify signaling pathways contributing to this phenomenon. Last but not least, we support our in vitro observations with an ex vivo analysis showing the incidence of senescent HPMCs in fragments of the peritoneum encompassed by ovarian cancer metastasis and in pathology-free areas.

3

Methods Chemicals Unless otherwise stated, all chemicals and plastics were from Sigma (St. Louis, MO). Recombinant HGF and HGF neutralizing antibody (cat #MAB294) were from R&D Systems (Abingdon, UK). Signaling pathway inhibitors: MG132, API-1 and SP600125 were purchased from Tocris Bioscience (Ellisville, MO, USA), whereas SB202190 was from Cell Signaling Technology (Danvers, MA, USA).

Patients The study included an analysis of tumors excised from the peritoneal cavity from 8 patients with serous ovarian cancer (stage III) as well as fragments of tumor-free peritoneum from the same patients. The tissues were fixed in 4% formalin, embedded in paraffin and cut into 3 μm sections. Deparaffinization, rehydration and epitope retrieval were conducted with Envision Flex Target Retrieval Solution (Dako, Glostrup, Denmark). The cancerous tissue was identified using standard H+E staining. Mesothelial cells lying in proximity of the cancer cells were identified according to positive results of immunohistochemistry against the D2-40 antigen, with the specific anti-D2-40 antibody (cat #M361901-2, Dako) [8]. Antigen visualization was performed using Envision Flex (Dako). The procedures involving human subjects were in accordance with the Helsinki Declaration of 1975. The study was approved by an institutional bioethics committee (consent number 543/14) and all patients gave their informed consent.

Cell cultures HPMCs were isolated by enzymatic digestion of omentum, obtained from 12 patients undergoing abdominal surgery. The age of the donors ranged from 26 to 32 years old. The 4

cells were propagated as described in [9] and identified as pure mesothelial by their typical cobblestone appearance at confluence and uniform positive staining for cytokeratins and HBME-1 antigen. The study was approved by an institutional bioethics committee (consent number 754/13) and all patients gave their informed consent. The ovarian cancer cell lines SKOV-3 and A2780 were purchased from the ECCC (Porton Down, UK) and the OVCAR-3 line was obtained from the ATCC (Rockville, MD). The cancer cells were cultured as described in [4]. Their identity was confirmed by analysis of NCBI database.

Conditioned media and exogenous, recombinant HGF Two types of conditioned media (CM) were collected and applicated to cultured HPMCs: an autologous CM generated by young (first passage) HPMCs and cancer-derived CM produced by OVCAR-3, SKOV-3 and A2780 cells. Irrespective on cell type, 3x105 of cells was seeded into 25 cm2 flasks and maintained until reaching 80% confluency. Afterwards, the cells were washed with phosphate-buffered saline (PBS) and a fresh growth medium was added for 72 h to generate the CM. In order to collect serum-free CM for the immunoenzymatic assays, the cells that reached 80% confluency were washed with PBS and serum starved for 72 h. Samples of CM were centrifuged, filtered through a 0.2 µm pore size filter, and stored at –80 °C until required. During subsequent experiments, HPMCs were subjected to 25% CM diluted in the standard growth medium for 72 h and then certain biological properties of the cells were evaluated. In some experiments, biological properties of HPMCs were also tested upon their exposure to exogenous, recombinant HGF for 72 h. The cytokine was used at doses corresponding to its level in the CM generated by ovarian cancer cells, that is: 2.5 ng/ml (OVCAR-3), 6.5 ng/ml (SKOV-3), and 9.5 ng/ml (A2780).

5

HPMC proliferation and senescence Proliferation of first passage HPMCs exposed to autologous and cancer-derived CM was estimated using Cell Proliferation Kit I (PromoKine; Heidelberg, Germany). In order to examine the effect of the CM on the cumulative number of population doublings (CPD), the HPMCs were serially passaged until their ability to divide was exhausted, as described in [10]. The activity of senescence-associated -galactosidase (SA-β-Gal) in cell cultures was assessed according to Gary and Kindell [11], while in the paraffin sections it was assessed according to Dimri et al. [12]. Immunofluorescence of -H2A.X and 53BP1 in cell cultures was evaluated using methodologies described in [13] with anti-γ-H2A.X (cat #ab2893, Abcam, Cambridge, UK) and anti-53BP1 antibody (cat #NB100-304, Novus Biologicals, Abingdon, UK), respectively. The presence of -H2A.X in the paraffin sections was detected as described in [14] with the use of anti--H2A.X antibody (cat #10856-1-AP, Proteintech, Manchester, UK). Expression of connexin 43 (Cx43) was performed in HPMCs fixed in paraformaldehyde and then incubated with anti-Cx43 antibody (cat #ab11370, Abcam, 1:100, overnight) and Alexa Fluor DyLight 488 (Abcam; 1:500, 1h). In the paraffin sections, Cx43 was stained using Connexin-43 Polyclonal Antibody (cat #15386-1-AP, Proteintech, 1:200) and the reaction was visualized using the Novolink Polymer Detection System (Novocastra Reagents, Wetzlar, Germany). Pictures of immunoreactions were taken using the Axio Vert.A1 microscope (Carl-Zeiss, Jena, Germany) and fluorescence was quantified with a SynergyTM 2 spectrofluorometer (BioTek Instruments, Winooski, VT, USA).

Oxidative stress The generation of mitochondrial superoxides in HPMCs treated with autologous and cancer-derived CM or with exogenous HGF was examined using a fluorescent dye, MitoSOX Red (Thermo Fisher Scientific, Waltham, MA USA). In brief, the cells were stained with 0.1 6

M MitoSOX Red for 15 min and then the fluorescence was recorded at excitation wavelength of 390 nm and emission wavelength of 580 nm using a SpectraMax M4 plate reader (Molecular Devices, Sunnyvale, CA). The parallel measurement at excitation wavelength of 510 nm was performed to monitor nonsuperoxide-dependent processes that can oxidize HE probes to ethidium [15]. The production of hydrogen peroxide was examined in cells transfected with either a cytosolic or a mitochondrially targeted Hydrogen Peroxide detection vector (HyPer; Evrogen, Moscow, Russia) which is a genetically encoded fluorescent sensor allowing highly specific detection of hydrogen peroxide. The transfection was conducted using 1.5 μg/ml of HyPer expressing vectors in 0.25% Lipofectamine 2000 (Thermo Fisher Scientific). Green fluorescence released by the cells was quantified using the SpectraMax M4 plate reader with excitation wavelength of 500 nm and emission wavelength of 516 nm [16]. The magnitude of oxidative modifications of cellular lipids was evaluated according to a quantification of 8-isoprostane (8-iso Prostaglandin F2α), using 8-Isoprostane ELISA kit from Cayman Chemical (Ann Arbor, MI, USA), as per manufacturer’s instructions.

Mitochondrial metabolism Mitochondrial membrane potential (ΔΨm) was measured in cells probed with 1µM 5,5',6,6'-tetrachloro-1,1',3,3'-tetraethylbenzimidazolylcarbocyanine iodide (JC-1), which is a lipophilic cationic fluorescent dye exhibiting potential-dependent accumulation in mitochondria [17]. JC-1 accumulates in the functional mitochondria forming red fluorescent J-aggregates (at 590 nm), while mitochondria de-energization results in the accumulation of green fluorescent monomers (at 535 nm). The rise in the green/red fluorescence intensity ratio indicates a drop of the mitochondrial membrane potential values. Mitochondrial mass was measured after cell staining with 10 µM of 10-n-nonyl-acridine orange (NAO) for 10 min at 7

37° C in the dark. The fluorescence was recorded at excitation wavelength of 435 nm and emission wavelength of 535 nm. Both mitochondrial membrane potential and mitochondrial mass were analysed as described in [18]. Enzymes involved in the production of mitochondrial ROS (cytochrome c oxidase and NADH dehydrogenase) and biogenesis of mitochondria (peroxisome proliferator-activated receptor gamma coactivator-1 alpha; PGC-1α) were analysed using commercial kits purchased from Wuhan EIAab Science Co., Ltd (Wuhan, China), as per manufacturer’s instructions.

Cell secretome Concentrations of GRO-1, IL-1β, IL-6, HGF, PAI-1, TGF-β1, and TNFα in CM from cancer cells and HPMCs were determined with appropriate DuoSet® Immunoassay Development kits (R&D Systems) according to the manufacturer’s instructions.

Signaling pathways The effect of autologous and cancer-derived CM or exogenous HGF on the activation of signaling pathways associated with cellular senescence (AKT, JAK1, JNK, PI3K, NF-B, and p38 MAPK) was analyzed using specific ELISA-based kits purchased from Active Motif (Carlsbad, CA, USA) as per manufacturer’s instructions.

Intervention studies In some experiments, SA-β-Gal was quantified in response to exogenous HGF or upon a CM pre-incubation (for 4 h) with HGF neutralizing antibody (1 µg/ml). The enzyme’s activity was also analyzed in HPMCs pre-incubated (for 4 h) with ROS spin-trap scavenger N-tertbutyl-alpha-phenylnitrone (PBN, 800 µM) and with HPMCs pre-incubated (for 4 h) with 8

chemical inhibitors of senescence-associated signaling pathways, i.e. with SB202190 (the inhibitor of p38 MAPK, 10 µM), MG132 (the inhibitor of NF-B, 10 µM), API-1 (the inhibitor of AKT, 50 µM), and SP600125 (inhibitor of JNK, 10 µM).

Adhesion assay Adhesion of calcein AM-probed ovarian cancer cells to HPMCs subjected to autologous and cancer-derived conditioned medium was quantified according to a fluorescence-based method described in [19]. During the experiment, the exposure of some HPMCs to the conditioned media was preceded with a pre-incubation with the PBN (800 µM) for 4 h.

Statistics Statistical analysis was performed using GraphPad Prism 5.00 (GraphPad Software, San Diego, USA). The means were compared with t-test. The results were expressed as means  SD. Differences with a P-value <0.05 were considered to be statistically significant.

9

Results Ovarian cancer cells induce the senescence of HPMCs in vitro The effects of conditioned medium (CM) generated by ovarian cancer cells were compared with those elicited by autologous HPMC-derived CM in terms of HPMC growth and senescence. The experiments showed that CM collected from OVCAR-3, SKOV-3, and A2780 cells inhibited proliferation of HPMCs in short-term conditions (Fig. 1A) and reduced the number of divisions completed by those cells (CPD) before reaching senescence upon their prolonged exposure (Fig. 1B). These anti-proliferative effects of cancer cell-derived CM coincided with the induction of SA-β-Gal (Fig. 1C) and the activation of DNA damage response, as evidenced according to increased incidence of DNA damage foci, i.e. γ-H2A.X and 53BP1. Finally, HPMCs exposed to the cancerous CM were characterized by decreased expression of junctional protein Cx43 (Fig. 2).

Senescent HPMCs accompany cancer cells within peritoneal ovarian tumors Comparative analysis of two areas of the peritoneal cavity from several patients with ovarian cancer metastases, i.e. regions encompassed by the pathology and those free from cancer cells, was performed. The study showed that HPMCs lying near the cancerous tissue displayed such features of senescence as increased expression of SA-β-Gal and γ-H2A.X and decreased expression of Cx43. When regions distant from the metastasis were analyzed, the incidence of HPMCs showing both positive and negative markers of senescence was only sporadic (Fig. 3).

HGF as a candidate mediator of the pro-senescence activity of ovarian cancer cells The biochemical composition of cancer-derived CM and HPMC-derived autologous CM was examined in order to identify plausible mediators of cancer cell-dependent senescence of 10

HPMCs. The analysis was focused on seven proteins that, according to the available literature, may have some potential to induce senescence, i.e. GRO-1[7], HGF [20], IL-1β [21], IL-6 [22], PAI-1 [23], TGF-β1 [10], and TNFα [24]. Importantly, the candidate mediator/s should fulfill one eligibility criterion, i.e. its/their level in the CM from three lines of ovarian cancer cells must be significantly greater than in the CM from HPMCs. Among cytokines whose concentration in various CMs was presented in Table 1, only HGF fulfilled the above-mentioned requirement, and was thus treated in subsequent research as a potential mediator of cancer cell pro-senescence activity.

HGF induces senescence in HPMCs Exogenous, recombinant form of HGF was applied to HPMCs at doses corresponding to its level in cancer cell-derived CM in order to verify the capacity of this cytokine to induce cellular senescence. The study showed that HGF increased the expression of SA-β-Gal in a dose-dependent manner (Fig. 4A) and that the pre-incubation of CM with a specific anti-HGF antibody prevented its capability to up-regulate the enzyme. Under these conditions, the activity of SA-β-Gal declined to the level characterizing cells subjected to the autologous, HPMC-derived CM (Fig. 4B).

Cancer-derived CM and HGF trigger senescence of HPMCs via the induction of oxidative stress The HPMCs were exposed to cancer-derived and autologous CM to quantify the production of ROS. The study showed that cells subjected to cancer-derived CM generated considerably more mitochondrial superoxides and hydrogen peroxide than their counterparts treated with CM from HPMCs. The level of cytosolic hydrogen peroxide was, at the same

11

time, unchanged. Similarly, increased generation of both types of ROS was also observed when the cells were exposed to an exogenous HGF (Tab. 2 and Tab. 3). These effects were accompanied by up-regulated level of cytochrome c oxidase. The activity of NADH dehydrogenase was, in turn, unaltered. Further analysis of mitochondria revealed that the mass of the organelles was markedly increased, which coincided with increased activity of an enzyme involved in mitochondria biogenesis, PGC-1α. At the same time, the level of mitochondria energization measured according to the values of the inner membrane potential (ΔΨm) in the cells exposed to the CM or the HGF were comparable with those characterizing the control groups (Tab. 2 and Tab. 3). Last but not least, the cells subjected to either cancerous CM or the exogenous HGF were characterized by increased concentration of the product of lipid peroxidation, 8-isoprostane (Tab. 2 and Tab. 3). Intervention studies with the ROS spin-trap scavenger, PBN, showed that when HPMCs subjected to samples of CM or to recombinant HGF were protected against oxidative stress by their pre-incubation with this compound, the initial induction of senescence (SA-β-Gal) in these cells was markedly reduced (Fig. 5A-B).

CM/HGF-induced senescence involves activation of p38 MAPK, AKT, and NF-B Six signaling pathways associated with cellular senescence, i.e. AKT, PI3K, p38 MAPK, NF-B [25], JAK1 [26], and JNK [27], were evaluated in HPMCs subjected to cancer cellderived CM. It has been found that CM from either OVCAR-3 or SKOV-3 cells activated AKT, JNK, NF-B, and p38 MAPK, whereas CM from A2780 cells activated NF-B and p38 MAPK (Fig. 6A). The same pathways were re-analyzed in HPMCs treated with exogenous HGF in order to identify pathways whose activation overlapped under both regimens (HGF vs. CM). The analysis showed that HGF used at 2.5 ng/ml (corresponding to OVCAR-3 cells) activated NF12

B, PI3K, and p38 MAPK; when used at 6.5 ng/ml (corresponding to SKOV-3 cells) it activated AKT, JNK, NF-B, and p38 MAPK; and when used at 9.5 ng/ml (corresponding to A2780 cells) it activated JNK, NF-B, PI3K, and p38 MAPK (Fig. 6B). Intervention studies with specific inhibitors of the signaling pathways revealed that the activation of senescence (SA-β-Gal) in HPMCs treated with cancer cell-derived CM can be abolished upon prior an inactivation of p38 MAPK and NF-B (in the case of CM from OVCAR-3 and A2780 cells), and the inactivation of p38 MAPK, NF-B, and AKT (in the case of CM from SKOV-3 cells) (Fig. 6C).

HPMCs senesced prematurely in response to cancer-derived CM promote ovarian cancer cell adhesion in oxidative stress-dependent mechanism In order to verify if premature senescence of HPMCs elicited by the CM generated by ovarian cancer cells may exert a cancer-promoting activity, the efficiency of cancer cell adhesion to these cells was evaluated. The experiment showed that, indeed, all three ovarian cancer cell lines tested adhered to HPMCs pre-exposed to cancer-derived CM more effectively than to their counterparts incubated with the autologous CM. However, when the HPMCs were protected against oxidative stress with the PBN shortly before the application of the cancerous CM, the adhesion-promoting effect of this medium disappeared (Fig. 7).

13

Discussion According to the “seed and soil” theory of carcinogenesis, cancer cells that leave their primary spot and reach a given distant organ need a hospitable tissue environment to successfully form a metastasis [28]. In the peritoneal cavity, which is the primary place for the dissemination of ovarian tumors [29], the metastatic niche is generated by senescent HPMCs [5]. In this paper, we show that ovarian cancer cells may force the HPMCs to adopt the senescence phenotype, which may help them to metastasize more efficiently. This finding stems from the observation that conditioned medium generated by the cancer cells elicits a spectrum of phenotypic features of cellular senescence in HPMCs, including reduced proliferative potential [30], increased expression of SA-β-Gal [12], activation of senescencerelated DNA damage response (the presence of γ-H2A.X and 53BP1 foci [31]), and reduced expression of the junctional protein, Cx43 [32]. More significantly, an analysis of the paraffin sections of the peritoneum from patients with ovarian cancer revealed that HPMCs adjacent to the cancerous tissue displayed several signs of senescence, whereas the presence of senescent HPMCs was incidental in areas free from the pathology. This observation implies that cancer cells shed from the surface of the ovary form peritoneal tumors preferentially in areas that are rich in senescent HPMCs, or that cancer cells floating in the peritoneal fluid actively strengthen the feasibility of their implantation by sending soluble, pro-senescence signals towards the HPMCs. As per current knowledge, both scenarios seem to be equally plausible and it cannot be excluded that they occur simultaneously. Such a paracrine mechanism of senescence spreading has already been described between different kinds of somatic cells [33], and a similar phenomenon proceeding in the cancer-cell-to-normal cell direction has only been reported for ovarian cancer cells and fibroblasts [7].

14

Mechanistic examinations focused on the ovarian cancer cell secretome followed by intervention studies with exogenous proteins and neutralizing antibodies identified HGF as the mediator by which all ovarian cancer cell lines that were tested induced senescence of HPMCs. The activity of HGF has been recognized to be associated with increased oxidative stress, as evidenced according to measurements of ROS and products of lipid peroxidation. As per the role of HGF in cellular senescence, the causative involvement of this cytokine was so far unknown, although its elevated production was reported in senescent fibroblasts [34]. At the same time, there is evidence that HGF may mediate some interactions between HPMCs and cancer cells. For example, HGF has been found to stimulate HPMC-dependent ovarian cancer cell adhesion, proliferation, and migration, which in the face of our current findings may indicate that these effects proceeded in the senescence-dependent mechanism [20]. Our findings are also in line with results of Matte and colleagues, who found that HGF present in ovarian cancer-derived ascites may modulate behaviour of HPMCs, in particular their migratory potential [35]. The link between HGF and ROS is, in turn, more recognized, as a stimulation of oxidative stress by this cytokine was described for both normal [36] and transformed cells [37]. Detailed studies aimed at identifying an exact source and mechanism of increased ROS productions showed that the cancerous CM and HGF consistently enhance the generation of superoxides and hydrogen peroxide in mitochondria. Increased formation of the ROS was probably a result of increased activity of oxidase cytochrome c, an enzyme involved in mitochondrial respiratory chain reactions, whose up-regulation was also linked with increased efflux of oxidants from mitochondria during senescence of fibroblasts [38]. Interestingly, increased release of mitochondrial ROS was also associated with increased biogenesis of mitochondria, likely due to increased activity of PGC-1α [39]. At the same time, mitochondria in the HPMCs subjected to the cancerous CM and HGF did not express signs of 15

de-energization, namely decreased ΔΨm, which may in turn indicate that the induction of oxidative stress-dependent senescence in these cells was not related to the so-called retrograde signaling response in which the intensified mitochondria biogenesis is a compensatory cell reaction to their decreased ability to generate ATP [17]. An analysis of signaling pathways underlying the activity of cancer cell-derived conditioned medium/HGF allowed to show that all cancer cells initiated in HPMCs p38 MAPK and NF-B signaling, and the SKOV-3 cells activated additionally the AKT pathway. The induction of HPMC senescence in a mechanism involving p38 MAPK is in line with our previous study, which showed that blocking this kinase rejuvenated late-passage presenescent HPMCs and inhibited their ability to support the development of ovarian tumors in the mouse peritoneal cavity [5]. The role of NF-B is in agreement, in turn, with the contribution of this transcription factor (next to p38 MAPK) in the development of the senescence-associated secretory phenotype in senescent HPMCs [9]. The role of AKT signaling in the course of HPMC senescence has never been reported before; however, activation of this kinase has been described during senescence of fibroblasts [40]. It can be speculated that HGF activated the above-mentioned pathways in an oxidative stressdependent manner, as activation of p38 MAPK in HPMCs was prevented by the ROS scavenger [4], and NF-B and AKT appeared to be activated by oxidants [41,42]. To finally verify whether HPMCs that prematurely senesced in response to the activity of the conditioned medium generated by ovarian cancer cells exert some cancer-promoting features, the efficiency of cancer cell adhesion to these cells was evaluated. This phenomenon was chosen because it plays a critical role in very initial steps of intraperitoneal cancer progression [1] and because the pro-adhesive capabilities of senescent HPMCs have already been well recognized [4]. Current experiments showed that, indeed, the efficiency at which ovarian cancer cells adhered to HPMCs treated with the cancerous medium was remarkably 16

elevated. The causative role of oxidative stress in this process confirmed, in turn, the reduction of the cancer cell adhesion upon the HPMCs pre-incubation with ROS scavenger. Collectively, our results shed new light on the pathophysiology of peritoneal cavity colonization by ovarian cancer cells. In particular, they indicate that cancer cells may actively contribute to the formation of their metastatic bed by inducing the tumor-promoting senescence phenotype in HPMCs.

Conflict of interest The authors declare no conflict of interest.

Acknowledgments This work was supported by a grant from the National Science Centre, Poland (2014/15/B/NZ3/00421).

Abbreviations CM, conditioned medium; OCCs, ovarian cancer cells; CPD, cumulative number of population doublings; Cx43, connexin 43; H2O2, hydrogen peroxide; HGF, hepatocyte growth factor; HPMCs, human peritoneal mesothelial cells; JC-1, 5,5',6,6'-tetrachloro-1,1',3,3'tetraethylbenzimidazolylcarbocyanine iodide; NAO, N-nonyl acridine orange; NF-B, nuclear factor kappa B; PBS, phosphate buffered saline; PGC-1α, peroxisome proliferator-activated receptor gamma coactivator-1 alpha; RFU, relative fluorescence units; ROS, reactive oxygen species; SA-β-Gal, senescence-associated β-galactosidase

17

References

[1] H.A. Kenny, C.Y. Chiang, E.A. White, E.M. Schryver, M. Habis, I.L. Romero, A. Ladanyi, C.V. Penicka, J. George, K. Matlin, A. Montag, K. Wroblewski, S.D. Yamada, A.P. Mazar, D. Bowtell, E. Lengyel, Mesothelial cells promote early ovarian cancer metastasis through fibronectin secretion, J. Clin Invest 124 (2014) 4614-4628. [2] J. Mikula-Pietrasik, P. Sosinska, M. Kucinska, M. Murias, K. Maksin, A. Malinska, A. Ziolkowska, H. Piotrowska, A. Wozniak, K. Ksiazek, Peritoneal mesothelium promotes the progression of ovarian cancer cells in vitro and in a mice xenograft model in vivo, Cancer Lett. 355 (2014) 310-315. [3] P. Sosinska, J. Mikula-Pietrasik, M. Ryzek, E. Naumowicz, K. Ksiazek, Specificity of cytochemical and fluorescence methods of senescence-associated beta-galactosidase detection for ageing driven by replication and time, Biogerontology. 15 (2014) 407413. [4] K. Ksiazek, J. Mikula-Pietrasik, K. Korybalska, G. Dworacki, A. Jorres, J. Witowski, Senescent peritoneal mesothelial cells promote ovarian cancer cell adhesion: the role of oxidative stress-induced fibronectin, Am. J. Pathol. 174 (2009) 1230-1240. [5] J. Mikula-Pietrasik, P. Uruski, P. Sosinska, K. Maksin, H. Piotrowska-Kempisty, M. Kucinska, M. Murias, S. Szubert, A. Wozniak, D. Szpurek, S. Sajdak, K. Piwocka, A. Tykarski, K. Ksiazek, Senescent peritoneal mesothelium creates a niche for ovarian cancer metastases, Cell Death. Dis. 7 (2016) e2565. [6] J. Mikula-Pietrasik, P. Sosinska, E. Naumowicz, K. Maksin, H. Piotrowska, A. Wozniak, D. Szpurek, K. Ksiazek, Senescent peritoneal mesothelium induces a proangiogenic phenotype in ovarian cancer cells in vitro and in a mouse xenograft model in vivo, Clin Exp Metastasis 33 (2016) 15-27. 18

[7] G. Yang, D.G. Rosen, Z. Zhang, R.C. Bast, Jr., G.B. Mills, J.A. Colacino, I. MercadoUribe, J. Liu, The chemokine growth-regulated oncogene 1 (Gro-1) links RAS signaling to the senescence of stromal fibroblasts and ovarian tumorigenesis, Proc. Natl. Acad. Sci. U. S. A 103 (2006) 16472-16477. [8] T. Terada, Immunohistochemical profile of normal mesothelium and histiocytic/methothelial hyperplasia: a case report, Int. J Clin Exp Pathol. 4 (2011) 631-636. [9] J. Mikula-Pietrasik, P. Sosinska, K. Maksin, M.G. Kucinska, H. Piotrowska, M. Murias, A. Wozniak, D. Szpurek, K. Ksiazek, Colorectal cancer-promoting activity of the senescent peritoneal mesothelium, Oncotarget. 6 (2015) 29178-29195. [10] J. Mikula-Pietrasik, P. Uruski, K. Aniukiewicz, P. Sosinska, Z. Krasinski, A. Tykarski, K. Ksiazek, Serum from Varicose Patients Induces Senescence-Related Dysfunction of Vascular Endothelium Generating Local and Systemic Proinflammatory Conditions, Oxid. Med Cell Longev. 2016 (2016) 2069290. [11] R.K. Gary, S.M. Kindell, Quantitative assay of senescence-associated betagalactosidase activity in mammalian cell extracts, Anal. Biochem. 343 (2005) 329334. [12] G.P. Dimri, X. Lee, G. Basile, M. Acosta, G. Scott, C. Roskelley, E.E. Medrano, M. Linskens, I. Rubelj, O. Pereira-Smith, J. Campisi, A biomarker that identifies senescent human cells in culture and in aging skin in vivo, Proc. Natl. Acad. Sci. U. S. A 92 (1995) 9363-9367. [13] J. Mikula-Pietrasik, P. Sosinska, M. Murias, M. Michalak, M. Wierzchowski, M. Piechota, E. Sikora, K. Ksiazek, Resveratrol Derivative, 3,3',4,4'-Tetrahydroxy-transStilbene, Retards Senescence of Mesothelial Cells via Hormetic-Like Prooxidative Mechanism, J. Gerontol. A Biol. Sci. Med Sci. 2014) 19

[14] C. Xie, L.Y. Xu, Z. Yang, X.M. Cao, W. Li, N.H. Lu, Expression of gammaH2AX in various gastric pathologies and its association with infection, Oncol Lett. 7 (2014) 159-163. [15] K.M. Robinson, M.S. Janes, M. Pehar, J.S. Monette, M.F. Ross, T.M. Hagen, M.P. Murphy, J.S. Beckman, Selective fluorescent imaging of superoxide in vivo using ethidium-based probes, Proc. Natl. Acad. Sci. U. S. A 103 (2006) 15038-15043. [16] Z. Ungvari, N. Labinskyy, P. Mukhopadhyay, J.T. Pinto, Z. Bagi, P. Ballabh, C. Zhang, P. Pacher, A. Csiszar, Resveratrol attenuates mitochondrial oxidative stress in coronary arterial endothelial cells, Am. J Physiol Heart Circ. Physiol 297 (2009) H1876-H1881. [17] J.F. Passos, G. Saretzki, S. Ahmed, G. Nelson, T. Richter, H. Peters, I. Wappler, M.J. Birket, G. Harold, K. Schaeuble, M.A. Birch-Machin, T.B. Kirkwood, T. von Zglinicki, Mitochondrial dysfunction accounts for the stochastic heterogeneity in telomere-dependent senescence, PLoS. Biol. 5 (2007) e110. [18] J. Mikula-Pietrasik, A. Kuczmarska, B. Rubis, V. Filas, M. Murias, P. Zielinski, K. Piwocka, K. Ksiazek, Resveratrol delays replicative senescence of human mesothelial cells via mobilization of antioxidative and DNA repair mechanisms, Free Radic. Biol. Med. 52 (2012) 2234-2245. [19] J. Mikula-Pietrasik, P. Sosinska, K. Ksiazek, Resveratrol inhibits ovarian cancer cell adhesion to peritoneal mesothelium in vitro by modulating the production of alpha5beta1 integrins and hyaluronic acid, Gynecol. Oncol 134 (2014) 624-630. [20] J. Mikula-Pietrasik, P. Uruski, K. Matuszkiewicz, S. Szubert, R. Moszynski, D. Szpurek, S. Sajdak, A. Tykarski, K. Ksiazek, Ovarian cancer-derived ascitic fluids induce a senescence-dependent pro-cancerogenic phenotype in normal peritoneal mesothelial cells, Cell Oncol (Dordr. ) 39 (2016) 473-481. 20

[21] N. Uekawa, A. Nishikimi, K. Isobe, Y. Iwakura, M. Maruyama, Involvement of IL-1 family proteins in p38 linked cellular senescence of mouse embryonic fibroblasts, FEBS Lett. 575 (2004) 30-34. [22] M. Sapochnik, M.R. Haedo, M. Fuertes, P. Ajler, G. Carrizo, A. Cervio, G. Sevlever, G.K. Stalla, E. Arzt, Autocrine IL-6 mediates pituitary tumor senescence, Oncotarget. 2016) [23] D.J. Elzi, Y. Lai, M. Song, K. Hakala, S.T. Weintraub, Y. Shiio, Plasminogen activator inhibitor 1--insulin-like growth factor binding protein 3 cascade regulates stress-induced senescence, Proc. Natl. Acad. Sci. U. S. A 109 (2012) 12052-12057. [24] O. Beyne-Rauzy, C. Recher, N. Dastugue, C. Demur, G. Pottier, G. Laurent, L. Sabatier, M. Mansat-De, V, Tumor necrosis factor alpha induces senescence and chromosomal instability in human leukemic cells, Oncogene 23 (2004) 7507-7516. [25] Y. Xu, N. Li, R. Xiang, P. Sun, Emerging roles of the p38 MAPK and PI3K/AKT/mTOR pathways in oncogene-induced senescence, Trends Biochem. Sci. 39 (2014) 268-276. [26] M. Xu, T. Tchkonia, H. Ding, M. Ogrodnik, E.R. Lubbers, T. Pirtskhalava, T.A. White, K.O. Johnson, M.B. Stout, V. Mezera, N. Giorgadze, M.D. Jensen, N.K. LeBrasseur, J.L. Kirkland, JAK inhibition alleviates the cellular senescence-associated secretory phenotype and frailty in old age, Proc. Natl. Acad. Sci. U. S. A 112 (2015) E6301-E6310. [27] P. Spallarossa, P. Altieri, C. Barisione, M. Passalacqua, C. Aloi, G. Fugazza, F. Frassoni, M. Podesta, M. Canepa, G. Ghigliotti, C. Brunelli, p38 MAPK and JNK antagonistically control senescence and cytoplasmic p16INK4A expression in doxorubicin-treated endothelial progenitor cells, PLoS. One. 5 (2010) e15583.

21

[28] E. Fokas, R. Engenhart-Cabillic, K. Daniilidis, F. Rose, H.X. An, Metastasis: the seed and soil theory gains identity, Cancer Metastasis Rev. 26 (2007) 705-715. [29] E. Lengyel, Ovarian cancer development and metastasis, Am. J. Pathol. 177 (2010) 1053-1064. [30] K. Ksiazek, J. Mikula-Pietrasik, A. Jorres, J. Witowski, Oxidative stress-mediated early senescence contributes to the short replicative life span of human peritoneal mesothelial cells, Free Radic. Biol. Med. 45 (2008) 460-467. [31] V. Turinetto, C. Giachino, Multiple facets of histone variant H2AX: a DNA doublestrand-break marker with several biological functions, Nucleic Acids Res 43 (2015) 2489-2498. [32] M. Statuto, C. Bianchi, R. Perego, M.U. Del, Drop of connexin 43 in replicative senescence of human fibroblasts HEL-299 as a possible biomarker of senescence, Exp Gerontol. 37 (2002) 1113-1120. [33] J. Mikula-Pietrasik, P. Sosinska, J. Janus, B. Rubis, M. Brewinska-Olchowik, K. Piwocka, K. Ksiazek, Bystander senescence in human peritoneal mesothelium and fibroblasts is related to thrombospondin-1-dependent activation of transforming growth factor-beta1, Int. J. Biochem. Cell Biol. 45 (2013) 2087-2096. [34] M. Miyazaki, E. Gohda, K. Kaji, M. Namba, Increased hepatocyte growth factor production by aging human fibroblasts mainly due to autocrine stimulation by interleukin-1, Biochem. Biophys. Res Commun. 246 (1998) 255-260. [35] I. Matte, D. Lane, C. Laplante, P. Garde-Granger, C. Rancourt, A. Piche, Ovarian cancer ascites enhance the migration of patient-derived peritoneal mesothelial cells via cMet pathway through HGF-dependent and -independent mechanisms, Int. J Cancer 137 (2015) 289-298.

22

[36] P.V. Usatyuk, P. Fu, V. Mohan, Y. Epshtein, J.R. Jacobson, J. Gomez-Cambronero, K.K. Wary, V. Bindokas, S.M. Dudek, R. Salgia, J.G. Garcia, V. Natarajan, Role of cMet/phosphatidylinositol 3-kinase (PI3k)/Akt signaling in hepatocyte growth factor (HGF)-mediated lamellipodia formation, reactive oxygen species (ROS) generation, and motility of lung endothelial cells, J Biol. Chem. 289 (2014) 13476-13491. [37] R. Jagadeeswaran, S. Jagadeeswaran, V.P. Bindokas, R. Salgia, Activation of HGF/cMet pathway contributes to the reactive oxygen species generation and motility of small cell lung cancer cells, Am. J Physiol Lung Cell Mol. Physiol 292 (2007) L1488L1494. [38] R.G. Allen, M. Tresini, B.P. Keogh, D.L. Doggett, V.J. Cristofalo, Differences in electron transport potential, antioxidant defenses, and oxidant generation in young and senescent fetal lung fibroblasts (WI-38), J. Cell Physiol 180 (1999) 114-122. [39] V.S. LeBleu, J.T. O'Connell, K.N. Gonzalez Herrera, H. Wikman, K. Pantel, M.C. Haigis, F.M. de Carvalho, A. Damascena, L.T. Domingos Chinen, R.M. Rocha, J.M. Asara, R. Kalluri, PGC-1alpha mediates mitochondrial biogenesis and oxidative phosphorylation in cancer cells to promote metastasis, Nat. Cell Biol. 16 (2014) 99215. [40] M.V. Astle, K.M. Hannan, P.Y. Ng, R.S. Lee, A.J. George, A.K. Hsu, Y. Haupt, R.D. Hannan, R.B. Pearson, AKT induces senescence in human cells via mTORC1 and p53 in the absence of DNA damage: implications for targeting mTOR during malignancy, Oncogene 31 (2012) 1949-1962. [41] Y. Wang, X. Huang, H. Cang, F. Gao, T. Yamamoto, T. Osaki, J. Yi, The endogenous reactive oxygen species promote NF-kappaB activation by targeting on activation of NF-kappaB-inducing kinase in oral squamous carcinoma cells, Free Radic. Res 41 (2007) 963-971. 23

[42] J. Ahn, M. Won, J.H. Choi, Y.S. Kim, C.R. Jung, D.S. Im, M.L. Kyun, K. Lee, K.B. Song, K.S. Chung, Reactive oxygen species-mediated activation of the Akt/ASK1/p38 signaling cascade and p21(Cip1) downregulation are required for shikonin-induced apoptosis, Apoptosis. 18 (2013) 870-881.

24

Tables

Table 1 Comparative analysis of the HPMC and cancer cell secretome autologous vs. OCCs-derived CM

soluble agent (pg/105 cells)

HPMCs

OVCAR-3

SKOV-3

A2780

GRO-1

5481±449

3904±338

4807±272

51±10

HGF

1357±256

8509±983*

26854±2776*

37848±4812*

IL-1β

3.5±0.3

1.9±0.4

3.8±0.3

2.6±0.1

IL-6

2494±1055

28±3

36±5

65±37

PAI-1

24023±1316

9911±2780

2949±356

176±39

TGF-β1

57±4

27±6

41±6

60±5

TNFα

20±6

17±4

18±3

18±2

The analysis was focused on identifying agents whose secretion by the cancer cells was greater than by the HPMCs. If this condition was fulfilled, statistical analysis for the respective groups was performed. The asterisks indicate a significant difference (P>0.05) as compared with autologous CM generated by the HPMCs. The experiments were performed with HPMCs established from 8 different donors. Cancer cells were used in octuplicates. CM – conditioned medium; OCCs – ovarian cancer cells

25

Table 2 Changes in oxidative stress-related parameters in HPMCs exposed to autologous and cancer-derived conditioned medium Parameter (units) mitochondrial superoxides (RFU/105 cells) mitochondrial H2O2 (RFU/105 cells) cytosolic H2O2 (RFU/105 cells) 8-isoprostane (ng/ml) inner membrane potential ΔΨm mitochondrial mass (RFU/105 cells) cytochrome c oxidase (ng/ml) NADH dehydrogenase (ng/ml) PGC-1α (ng/ml)

autologous vs. OCCs-derived CM HPMCs

OVCAR-3

SKOV-3

A2780

5850±428

7747±144*

8129±99**

8617±212**

12738±265

16273±432*

18273±291**

21823±827**

8732±253

9182±625

8172±647

9271±1023

37±3

76±9***

63±7**

54±2**

1.15±0.06

1.15±0.02

1.12±0.05

1.08±0.08

277±18

381±42**

539±48***

472±36***

2.1±0.1

2.5±0.1*

2.5±0.1*

2.8±0.1*

22.4±2.1

23.4±0.9

22.5±1.0

22.9±1.7

15.9±0.7

23.1±0.4**

25.2±1.3**

26.4±0.2**

The asterisks indicate significant differences as compared with HPMCs exposed to autologous conditioned medium (* - P<0.05, ** - P<0.01, *** - P<0.001). The experiments were performed using HPMCs obtained from 6 different donors. Samples of CM were collected from 6 separate HPMC cultures and 6 separate cultures of every cancer cell line. The results are expressed as mean ± SD. CM – conditioned medium; H2O2 – hydrogen peroxide; OCCs – ovarian cancer cells; PGC-1α - peroxisome proliferator-activated receptor gamma coactivator-1 alpha; RFU – relative fluorescence units

26

Table 3. Changes in oxidative stress-related parameters in HPMCs exposed to exogenous, recombinant HGF Parameter (units) mitochondrial superoxides (RFU/105 cells) mitochondrial H2O2 (RFU/105 cells) cytosolic H2O2 (RFU/105 cells) 8-isoprostane (ng/ml) inner membrane potential ΔΨm mitochondrial mass (RFU/105 cells) cytochrome c oxidase (ng/ml) NADH dehydrogenase (ng/ml) PGC-1α (ng/ml)

HGF 0 ng/ml

2.5 ng/ml

6.5 ng/ml

9.5ng/ml

4738±89

6847±254*

8263±311**

8726±119**

10283±99

14382±182*

16992±162**

19230±253**

7283±827

8172±771

7717±625

8001±552

31±4

72±1***

69±3**

49±6*

1.22±0.02

1.17±0.03

1.19±0.03

1.18±0.03

248±30

344±32**

383±23**

395±21***

2.1±0.1

2.5±0.2*

2.5±0.1*

2.7±0.2*

21.8±0.9

14.4±1.6*

19.5 ±1.9

21.8±1.8

17.4±0.8

22.6±0.9*

20.9±0.7*

24.7±0.4**

HGF was used in concentrations corresponding to its level in conditioned medium generated by ovarian cancer cells (2.5 ng/ml - OVCAR-3; 6.5 ng/ml - SKOV-3; 9.5 ng/ml - A2780). The asterisks indicate significant differences as compared with HPMCs not exposed to HGF (* - P<0.05, ** - P<0.01, *** - P<0.001). The experiments were performed using HPMCs obtained from 6 different donors. The results are expressed as mean ± SD. H2O2 – hydrogen peroxide; PGC-1α - peroxisome proliferator-activated receptor gamma coactivator-1 alpha; RFU – relative fluorescence units

27

Figure legends

Fig. 1. Effect of autologous and cancer-derived conditioned medium on proliferation (A), replicative lifespan (B), and activity of SA-β-Gal (C) in HPMCs. The asterisks indicate significant differences as compared with HPMCs exposed to autologous conditioned medium (* - P<0.05, ** - P<0.01, *** - P<0.001). The experiments were performed using HPMCs obtained from 8 different donors. Samples of CM were collected from 8 separate HPMC cultures and 6 separate cultures of every cancer cell line. The results are expressed as mean ± SD. CM – conditioned medium; CPD – cumulative population doublings; OCCs – ovarian cancer cells; RFU – relative fluorescence units.

Fig. 2. Effect of autologous and cancer-derived conditioned medium on the presence of γ-H2A.X and 53BP1 foci, and the expression of connexin 43 (Cx43) in HPMCs. Representative images (A) and results of quantitative analysis of γ-H2A.X (B), 53BP1 (C), and Cx43 (D). The asterisks indicate significant differences as compared with HPMCs exposed to autologous conditioned medium (* - P<0.05, ** - P<0.01, *** - P<0.001). The experiments were performed using HPMCs obtained from 8 different donors. Samples of CM were collected from 8 separate HPMC cultures and 6 separate cultures of every cancer cell line. The results are expressed as mean ± SD. CM – conditioned medium; OCCs – ovarian cancer cells; RFU – relative fluorescence units. Magnification x 200, objective aperture 0.35.

Fig. 3. Analysis of the presence of senescent HPMCs in the peritoneum of patients with peritoneal metastases of ovarian tumors. A layer of HPMCs is marked with a white arrow. Senescent cells were recognized according to positive reactions for SA-β-Gal (green cytoplasm – exemplary cells are marked with black arrows) and γ-H2A.X (brown nuclei), and 28

negative reactions against Cx43 (brown nuclei). HPMCs were identified according to a positive reaction against the D2-40 antigen (brown nuclei). Cancer (-) images refer to areas of the peritoneum that were free from cancer, whereas cancer (+) images represent regions in which the HPMCs were lying near cancerous tissue. The study includes a comparison of specimens obtained from the same patient. The images shown in this figure are representative examples obtained from 8 patients. The arrows indicate exemplary positive reactions. Ad – adipose tissue; Ca – cancerous tissue. Magnification x 50, objective aperture 0.15.

Fig. 4. Analysis of the role of HGF in the induction of SA-β-Gal in HPMCs. Panel A shows the effect of exogenous, recombinant human (rh) HGF on the expression of SA-β-Gal in HPMCs. The cytokine concentrations used in the study refer to the HGF level in conditioned medium generated by ovarian cancer cells (2.5 ng/ml - OVCAR-3; 6.5 ng/ml SKOV-3; 9.5 ng/ml - A2780). The asterisks indicate significant differences as compared with HPMCs not exposed to HGF (* - P<0.05, ** - P<0.01). Panel B shows the effect of the HGF neutralization in autologous and cancer cell-derived conditioned medium on the expression of SA-β-Gal in HPMCs. The asterisks indicate significant differences as compared with HPMCs exposed to conditioned medium not subjected to the HGF neutralizing antibody (* - P<0.05, ** - P<0.01, *** - P<0.001). The experiments were performed using HPMCs obtained from 8 different donors. Samples of CM were collected from 8 separate HPMC cultures and 6 separate cultures of every cancer cell line. The results are expressed as mean ± SD. CM – conditioned medium; OCCs – ovarian cancer cells; RFU – relative fluorescence units

Fig. 5. Effect of autologous and cancer-derived conditioned media (A), and recombinant HGF (B) on the activity of SA-β-Gal in HPMCs in the presence or absence of the ROS scavenger, PBN. Asterisks indicate significant differences as compared with HPMCs 29

exposed to the autologous conditioned medium or not treated with HGF (* - P<0.05, ** P<0.01, *** - P<0.001). Hashtags indicate significant differences as compared with HPMCs not protected by the PBN (# - P<0.05, ## - P<0.01, ### - P<0.001). The experiments were performed using HPMCs obtained from 8 different donors. Samples of CM were collected from 8 separate HPMC cultures and 6 separate cultures of every cancer cell line. The results are expressed as mean ± SD. CM – conditioned medium; OCCs - ovarian cancer cells; RFU – relative fluorescence units

Fig. 6. Analysis of signaling pathways underlying ovarian cancer cell-related induction of senescence in HPMCs. Panel A shows the effect of autologous and cancer-derived conditioned medium on the activation of six signaling pathways in HPMCs. The asterisks indicate significant differences as compared with HPMCs exposed to autologous conditioned medium (* - P<0.05, ** - P<0.01). Panel B shows the effect of exogenous, recombinant HGF on the activation of the same six pathways in HPMCs. The cytokine is used at the concentrations corresponding to its level in the conditioned medium generated by ovarian cancer cells (2.5 ng/ml - OVCAR-3; 6.5 ng/ml - SKOV-3; 9.5 ng/ml - A2780). The asterisks indicate significant differences as compared with HPMCs not exposed to the HGF (* P<0.05, ** - P<0.01, *** - P<0.001). Panel C shows the induction of senescence (SA-β-Gal) in HPMCs pre-incubated with specific inhibitors of certain signaling pathways upon their exposure to cancer cell-derived conditioned medium. The asterisks indicate significant differences as compared with HPMCs exposed to intact (not treated with inhibitors) CM from cancer cells (* - P<0.05, ** - P<0.01, *** - P<0.001). The experiments were performed using HPMCs obtained from 8 different donors. Samples of CM were collected from 8 separate HPMC cultures and 6 separate cultures of every cancer cell line. The results are expressed as

30

mean ± SD. CM – conditioned medium; OCCs – ovarian cancer cells; RFU – relative fluorescence units

Fig. 7. Adhesion of ovarian cancer cells to HPMCs exposed to autologous and cancerderived conditioned medium. OVCAR-3 (A), SKOV-3 (B), and A2780 (C) cancer cells were seeded on top of HPMCs pre-incubated with the conditioned media. The exposure of some HPMCs to the media was preceded by their treatment with PBN. Asterisks indicate significant differences as compared with HPMCs exposed to the autologous conditioned medium (* - P<0.05). Hashtags indicate significant differences as compared with HPMCs not protected by the PBN (# - P<0.05, ## - P<0.01). The experiments were performed using HPMCs obtained from 8 different donors. Samples of CM were collected from 8 separate HPMC cultures and 6 separate cultures of every cancer cell line. The results are expressed as mean ± SD. CM – conditioned medium; OCCs – ovarian cancer cells; RFU – relative fluorescence units

31

32

33

34

35

36

ovarian cancer cells

conditioned medium (HGF)

 cancer cell adhesion

young HPMCs senescent HPMCs

 ROS   

p38 MAPK NF-B AKT

37

Highlights     

Conditioned medium (CM) from ovarian cancer cells induces senescence in HPMCs. Hepatocyte growth factor (HGF) mediates the pro-senescence activity of cancerous CM. Activity of cancerous CM/HGF is associated with the induction of oxidative stress. Signaling pathways underlying activity of CM/HGF include AKT, p38 MAPK and NF-B. HPMCs senesced in response to cancerous CM/HGF promote ovarian cancer cell adhesion.

38