Pro-cancerogenic effects of spontaneous and drug-induced senescence of ovarian cancer cells in vitro and in vivo: a comparative analysis

Senescent cancer cells in ovarian tumors in vivo

Expression of SA-β-Gal, the commonly accepted biochemical marker of cellular senescence, was quantified in tumors from 30 chemotherapy-naïve patients with EOC and from 30 patients who received CPT and PCT prior cytoreduction. Each tumor was cut into 10 specimens which were subjected to cytochemical detection of the enzyme. We found that SA-β-Gal staining was present in 53% of patients who were not subjected to chemotherapy. For as many as 90% of patients treated with CPT and PCT we found SA-β-Gal-positively stained areas in at least 2 out of 10 specimens analyzed (Fig. 1A). Upon further analysis of every SA-β-Gal-positive tumor area we discovered that the magnitude of cellular senescence within cancerous tissue (Fig. 1B) and the intensity of the enzyme staining (Fig. 1C) were significantly higher in tumors from patients experienced chemotherapy.

Fig. 1figure 1

Presence of senescent, SA-β-Gal-marked ovarian cancer cells in tumors in vivo. Quantification of senescent cancer cell frequency in tumors from chemotherapy-naïve patients and from patients treated with CPT + PCT before cytoreduction (A). Determination of SA-β-Gal staining area in cancerous tissue within tumors from patients displaying signs of cellular senescence (B). Representative pictures demonstrating green SA-β-Gal staining in both groups of patients (C). The measurements were performed using tumors obtained from 30 different patients per group. The results depicted on panel B are expressed as the means ± SEM. ** p < 0.01 vs. No drugs. Magnification × 400; bar = 50 μm

Molecular characteristics of drug-induced senescence in vitro and in vivo

In order to establish an in vitro model of drug-induced senescence of pEOCs which will resemble in vivo conditions in patients subjected to chemotherapy, we have recently created and optimized a regimen of cell exposure to CPT and PCT [11]. In this study, we compared a wide array of senescence-associated parameters between pEOCs whose senescence was induced by CPT + PCT in vitro and those derived from patients treated with CPT + PCT in vivo and for which senescence was triggered by serial passages in the absence of drugs in vitro. The results of this comparative analysis are shown in Table 1. It reveals that generally speaking all the tested parameters characterizing the course of senescence are equal between these two groups. This includes the percentages of cells bearing SA-β-Gal( +)/γ-H2A.X( +) phenotype, changes in the expression of cell cycle inhibitors (p16, p21, p53), lack of either telomere erosion or telomerase (hTERT) activity decline, non-telomeric localization of DNA damage foci, growth arrest in G2/M phase of cell cycle, and only sporadic apoptosis. Interestingly, however, young, early-passage pEOCs from patients who underwent chemotherapy in vivo were characterized by a significantly higher fraction of senescent [SA-β-Gal( +)/γ-H2A.X( +)] cells and cells expressing p16 cell cycle inhibitor compared with young cells from chemotherapy-naïve individuals.

Table 1 A comparison of senescence-associated parameters in pEOCs senesced upon their exposure to carboplatin (CPT) and paclitaxel (PCT) in vitro and in vivoSenescence-associated secretory phenotype in cancer cells

Another functional feature of pEOCs which was analyzed and compared between spontaneous senescence of cancer cells, drug-induced senescence in vitro, and replicative senescence of cells from patients subjected to CPT + PCT in vivo was the development of senescence-associated secretory phenotype (SASP). To this end, 24 proteins engaged in a variety of aspects of cancer cell progression, such as: angiogenesis, ECM remodeling and invasion, inflammation, proliferation, and migration was quantified in conditioned media (CM) from young and senescent cells. The results we obtained showed that the SASP profile was the most intense with respect to spontaneously senescent pEOCs. For these cells, the secretion of ANG1, CXCL8/IL-8, FGF5, VEGF, ADAM12, PDGF-D, tPA, TGF-β1, TIMP-1, TSP-1, CCL2/MCP-1, ICAM-1, IL-6, VCAM-1, CCL11, CXCL12/SDF-1, EGF, HGF, IGF-1, and NRP-1 was significantly higher with respect to in vitro CPT + PCT-treated cells, senescent cells from patients subjected to CPT + PCT in vivo, or both (Table 2).

Table 2 Senescence-associated secretory phenotype in pEOCs undergoing spontaneous and drug-induced senescence in vitro and in vivo

Interesting effect was found regarding ANG1, FGF5, ADAM12, tPA, TIMP-1, TSP-1, CCL2/MCP-1, IL-6, VCAM-1, CCL11, CXCL1/GRO-1, CXCL5, EGF, and NRP-1 whose baseline production by young cells from patients undergoing chemotherapy in vivo was significantly higher than for young cells from chemotherapy-naïve donors.

Senescent cancer cell-driven progression of non-senescent cancer cells in vitro

Three models of pEOCs senescence were compared regarding the ability their autologous CM to promote vital elements of cancer cell expansion, that is adhesion, proliferation migration, invasion, and epithelial-mesenchymal transition (EMT). Our comparative analysis of proliferating (non-senescent) pEOCs adhesion to normal peritoneal mesothelial cells (PMCs) and peritoneal fibroblasts (PFBs) revealed that CM derived from all types of senescent pEOCs stimulates their attachment to normal cells more effectively than CM from young cells, albeit the strongest effect was recorded in both cases for CM generated by spontaneously senescent pEOCs (Fig. 2A, B). As per cell proliferation, only CM from spontaneously senescent pEOCs was able to promote this process more than the secretome of young cells (Fig. 2C). The migration (towards the chemotactic activity of CM) and invasion (towards CM through the Basement Membrane Extract) of cancer cells were also fueled by all kinds of senescent pEOCs, however, similarly to adhesion, the strongest stimulation was displayed by spontaneously senescent cells (Fig. 2D, E). The efficacy of migration and invasion of cancer cells upon exposure to CM from young cells obtained from chemotherapy-subjected patients was significantly stronger than CM from chemotherapy-naïve individuals. Last but not least, CM from spontaneously senescent cells was the sole capable of decreasing the expression of occluding (a negative marker of EMT). At the same time, CMs from all types of senescent cells increased the expression of vimentin (a positive marker of EMT), however, the activity of CM from spontaneously senescent cells was the most pronounced (Fig. 2F, G).

Fig. 2figure 2

Effect of senescent pEOCs secretome on ovarian cancer cell progression in vitro. Analysis of pEOCs-derived conditioned medium effect on adhesion (to PMCs – A and to PFBs—B) proliferation (C), migration (D), and invasion (D) of non-senescent ovarian cancer cells. Quantification of occludin (F) and vimentin (G) expression as markers of EMT. Results derive from 6–8 independent experiments with pEOCs obtained from different donors. The results are expressed as the means ± SEM. * p < 0.05; ** p < 0.01 vs. “Young in vitro” cells; #p < 0.05 vs. “Sen in vitro” cells; Ψp < 0.05 vs. “Young/CPT + PCT in vivo” cells. RFU – Relative Fluorescence Units

Pro-angiogenic activity of senescent cancer cells

Three angiogenic reactions of vascular endothelial cells (HUVECs), that is proliferation, migration, and invasion were tested in response to CM generated by young and senescent pEOCs. Proliferation of HUVECs was stimulated by all three types of senescent pEOCs but the effect exerted by spontaneously senescent cells was the greatest (Fig. 3A). Migration of endothelial cells was supported exclusively by spontaneously senescent cells (Fig. 3B), whereas invasion was promoted by CM from spontaneously senescent cells and those for which senescence was elicited by CPT + PCT in vitro, with the strongest effect on the side of the spontaneously senescent cells (Fig. 3C).

Fig. 3figure 3

Effect of senescent pEOCs secretome on angiogenic behavior of vascular endothelial cells in vitro. Analysis of pEOCs-derived conditioned medium effect on proliferation (A), migration (B), and invasion (C) of vascular endothelial cells (HUVECs). Results derive from 6–8 independent experiments with pEOCs obtained from different donors. The results are expressed as the means ± SEM. * p < 0.05; ** p < 0.01 vs. “Young in vitro” cells; #p < 0.05 vs. “Sen in vitro” cells. RFU – Relative Fluorescence Units

Senescence-related functional characteristics of normal peritoneal cells subjected to senescent cancer cells

CMs from spontaneous and drug-inducible (in vitro) pEOCs were applied to young PMCs and PFBs to determine the induction of their senescence and the capacity of CM generated by normal cells under such conditions to support cancer cell proliferation, migration, and invasion. The quantification of SA-β-Gal-dependent fluorescence revealed that both types of senescent pEOCs induce the enzyme in PMCs and PFBs, albeit the most robust induction of senescence was exerted in both cases by CM from spontaneously senescent cells (Fig. 4A, E). Subsequently, CM generated by PMCs and PFBs exposed to CM from senescent pEOCs stimulated proliferation (Fig. 4B, F), migration (Fig. 4C, G), and invasion (Fig. 4D, H) of non-senescent ovarian cancer cells. For all the three tested phenomena, the effects related to spontaneously senescent pEOCs were considerably stronger.

Fig. 4figure 4

Paracrine effects of senescent pEOCs on senescence and pro-cancerogenic activity of normal PMCs (A-D) and PFBs (EH). Quantification of SA-β-Gal activity in PMCs and PFBs exposed to pEOCs-derived conditioned medium (A, E). Proliferation (B, F), migration (C, G), and invasion (D, H) of non-senescent ovarian cancer cells in response to autologous PMCs- and PFBs-derived conditioned medium upon their preexposure to conditioned medium generated by young and senescent pEOCs. The experiments were performed using pooled PMCs and PFBs from 6 different donors and ovarian cancer cells from 6 different patients. The results are expressed as mean ± SEM. The results are expressed as the means ± SEM. * p < 0.05; ** p < 0.01; *** p < 0.001 vs. “Young” cells; #p < 0.05 vs. “Sen” cells. RFU – Relative Fluorescence Units

Secretory properties of normal peritoneal cells subjected to senescent cancer cells

Taking into account the fact that senescent pEOCs induce SA-β-Gal in normal PMCs and PFBs, and that senescent cells typically display SASP, 19 proteins relevant for cancer cells progression was quantified in PMCs- and PFBs-derived CM upon their pre-incubation with CM generated by young and senescent pEOCs. For PMCs, all the tested proteins were hypersecreted by senescent cells, irrespective of the type of senescence. In addition, the release of 8 out of 19 proteins (ANG1, CXCL8, VEGF, PDGF-D, tPA, TGF-β1, IL-6, VCAM-1) was significantly higher in response to CM from spontaneously senescent pEOCs, the release of 9 molecules was comparable in both groups, and the release of 2 proteins (CXCL12/SDF-1 and NRP-1) was higher in response to drug-inducible senescent pEOCs (Table 3). As per PFBs’ secretome, all but two proteins (FGF5, CCL11) were hypersecreted upon treatment with CM from senescent pEOCs. The release of 12 out of 17 remaining up-regulated molecules was higher in response to CM from spontaneously senescent pEOCs, and the release of CXCL-8/IL-8, VEGF, PDGF-D, tPA, and TGF-β1 by spontaneously senescent and drug-inducible senescent cells was comparable (Table 4).

Table 3 Secretory phenotype of peritoneal mesothelial cells subjected to a conditioned medium generated by young and senescent ovarian cancer cellsTable 4 Secretory phenotype of peritoneal fibroblasts subjected to conditioned medium generated by young and senescent ovarian cancer cellsSenescence of ovarian cancer cells and tumor growth in vivo

The immunocompromised Scid mice were used to generate xenografts at the i.p. injection of young pEOCs, spontaneously senescent, and drug-inducible senescent cells. Senescent cells were mixed 1:1 with young cells. At the end of the 21-day experiment, tumors developed in mice injected with young and spontaneously senescent cells, albeit in the latter group the incidence (5/5 animals vs. 2/5 animals) and the total weight of tumors were significantly higher (Fig. 5). The experiment also included three other groups of animals in which CPT + PCT were administered in vivo (every 3 days) with an injection of young and senescent pEOCs. In those animals, chemotherapy inhibited the tumor growth either in young pEOCs or spontaneously senescent pEOCs xenografts (Fig. 5).

Fig. 5figure 5

The influence of senescent pEOCs on ovarian tumor development in mouse peritoneal cavity in vivo. Comparison of the total weight of tumors that developed intraperitoneally 21 days after the implantation of young or senescent + young pEOCs (1:1 ratio) (A). Representative pictures of animals and excised tumors from each group (B)

Senescence of ovarian cancer cells and the expression of genes responsible for drug resistance

Expression of 10 genes associated according to literature with cancer cell resistance to platins and taxanes was investigated at mRNA level using qPCR. These included ABCB1/MDR1, ABCC4, AKT1, PLK2, BIRC5/Survivin, CHEK1, CHEK2, PHB1/Prohibitin-1, PHB2/Prohibitin-2, and CCND1/cyclin D1. An analysis of these transcripts showed that senescence of pEOCs, irrespective of the kind, is not associated with any changes in the expression of the tested genes (Table 5).

Table 5 Expression of mRNA for genes associated with drug-resistance in young and senescent pEOCs

留言 (0)

沒有登入
gif