Oxaliplatin-Induced Senescence in Colorectal Cancer Cells Depends on p14ARF-Mediated Sustained p53 Activation

Simple Summary Chemotherapy can lead to cellular senescence in tumor cells. Here we demonstrate that oxaliplatin induces senescence in p53-proficient colorectal cancer (CRC) cells and leads to the G2-phase arrest in all lines studied (HCT116p53+/+, HCT116p53−/−, LoVo, SW48, and SW480). At early times the p53-competent lines activate p53 and p21CIP1, however, at later times, only LoVo cells showed sustained p53/p21CIP1 activation, accompanied by a strong induction of senescence and senescence-associated secretory phenotype (SASP) factors. Opposite to LoVo, the p53/p21CIP1 response and senescence induction is much weaker in the other p53-proficient cells, due to p14ARF deficiency. LoVo cells express p14ARF protein and siRNA-mediated knockdown of p14ARF significantly reduces sustained p53/p21CIP1 activation and senescence. Vice versa, ectopic expression of p14ARF enhances oxaliplatin-induced senescence in SW48 and SW480 cells. Our data show that oxaliplatin-induced senescence in CRC cells depends on p53 proficiency; however, a significant induction can only be observed upon p14ARF-mediated p53 stabilization. Abstract Senescence is an important consequence of cytostatic drug-based tumor therapy. Here we analyzed to which degree the anticancer drug oxaliplatin induces cell death, cell cycle arrest, and senescence in colorectal cancer (CRC) cells and elucidated the role of p53. Oxaliplatin treatment resulted in the G2-phase arrest in all CRC lines tested (HCT116p53+/+, HCT116p53−/−, LoVo, SW48 and SW480). Immunoblot analysis showed that within the p53-competent lines p53 and p21CIP1 are activated at early times upon oxaliplatin treatment. However, at later times, only LoVo cells showed sustained activation of the p53/p21CIP1 pathway, accompanied by a strong induction of senescence as measured by senescence-associated β-Gal staining and induction of senescence-associated secretory phenotype (SASP) factors. Opposite to LoVo, the p53/p21CIP1 response and senescence induction was much weaker in the p53-proficient SW48 and SW480 cells, which was due to deficiency for p14ARF. Thus, among lines studied only LoVo express p14ARF protein and siRNA-mediated knockdown of p14ARF significantly reduced sustained p53/p21CIP1 activation and senescence. Vice versa, ectopic p14ARF expression enhanced oxaliplatin-induced senescence in SW48 and SW480 cells. Our data show that oxaliplatin-induced senescence in CRC cells is dependent on p53 proficiency; however, a significant induction can only be observed upon p14ARF-mediated p53 stabilization.


Introduction
The tumor suppressor p53 is, to date, the most important transcription factor, regulating cell cycle arrest, cell survival, and cell death. A specific form of cell cycle arrest is cellular senescence, which was initially described as a permanent cell cycle arrest that limits the life span of cultured human fibroblasts [1]. Contrary to quiescence, which is defined as a temporary cell cycle arrest, senescence is not reversible in response to proliferative conditions (for review see [2][3][4][5]). Cellular senescence can be induced by telomere shorten-Center Mainz. Pifithrin α (PTHα, P4359, Sigma-Aldrich, St. Louis, MO, USA) was used at 30 µM. For silencing CDKN2a, predesigned siRNA (s217, Thermo Fisher Scientific) and control human non-silencing siRNA (Silencer Select Predesigned siRNA Negative Control #1 siRNA; Ambion, Austin, TX, USA) were used. For re-expressing p14 ARF , SW48 and SW480 cells were transfected with the pcDNA3-myc-ARF plasmid (19930, Addgene, [14]) or, as control, pcDNA3 using Effectene Reagent (Qiagen, Hilden, Germany).

Determination of Cell Death, Cell Cycle Progression, and Colony Formation Capacity
To determine cell death and cell cycle distribution, cells were incubated for different times after drug exposure, stained with propidium iodide (PI) and subjected to flow cytometry, as described [24]. The data were analyzed using the CellQuestPro/FACS DIVA Software (BD Canto II). For colony formation assays, 1 × 10 3 cells were seeded per 6-cm dish and 24 h later exposed to anticancer drugs, as described [25]. Colonies were fixed in methanol and stained with 1.25% Giemsa/0.125% crystal violet three weeks after exposure.

Determination of Senescence
For detection of senescence via β-Gal staining, cells were washed twice with PBS, afterwards fixed with 2% formaldehyde, 0.2% glutaraldehyde in PBS. After washing with PBS, cells were stained (40 mM citric acid/phosphate buffer pH 6.0, 150 mM NaCl, 2 mM MgCl 2 , 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 0.1% x-Gal) overnight at 37 • C. Cells were then washed with PBS and overlaid with 70% glycerine. Micrographs were acquired and analyzed using the ECHO Rebel microscope (Discover ECHO INC., San Diego, CA, USA). For detection of senescence via C 12 FDG staining, cells were exposed to oxaliplatin for indicated times. Thereafter, medium was removed and 4 mL fresh medium containing 10% FBS and 300 µM chloroquine was added per 10 cm dish. Cells were incubated for 30 min at 37 • C. Finally, 33 µM C 12 FDG (ImaGene Green™ C 12 FDG lacZ Gene Expression Kit) was added and incubated for 90 min at 37 • C. The medium was removed; cells were washed with PBS, trypsinized, and resuspended in PBS. Senescence was measured at the BD FACSCanto™ II flow cytometer.

Preparation of RNA and RT-qPCR
Total RNA was isolated using the Nucleo Spin RNA Kit (Machery and Nagel, Düren, Germany). One µg total RNA was transcribed into cDNA (Verso cDNA Kit, Thermo Scientific) and qPCR was performed using GoTaq ® qPCR Master Mix (Promega, Madison, WI, USA) and the CFX96 Real-Time PCR Detection System (Bio-Rad, Munich, Germany). The analysis was performed using CFX Manager TM software. Non-transcribed controls were included in each run, expression was normalized to GAPDH and ACTB; the untreated control was set to one. The specific primers are listed in Table A1.

Preparation of Cell Extracts and Western-Blot Analysis
Whole-cell extracts were prepared by direct lysis with 1x loading buffer (Roti-Load ® , Roth) heated to 95 • C. Proteins were separated by SDS-PAGE and transferred onto a nitrocellulose membrane (Amersham) by western blotting, blocked in 5% BSA dissolved in TBS-Tween and incubated with specific primary antibodies and secondary peroxidase conjugated antibodies (Table A2). Detection was performed using the iBright CL1000 (Invitrogen) system.

Statistical Analysis
The data were evaluated using Student's t-test and were expressed as a mean ± SD * p ≤ 0.1 significant, ** p ≤ 0.01 very significant, *** p ≤ 0.001 highly significant. Statistical analyses were performed using GraphPad Prism version 6.01 for Windows, GraphPad Software, La Jolla, CA, USA.

Oxaliplatin Induces Cell Death and Cell Cycle Arrest
To analyze whether (and to which degree) oxaliplatin induces senescence in CRC cells, we utilized five different cell lines differing in their p53 status. Thus, HCT116 p53+/+ , LoVo and SW48 are p53-proficient (wild type); SW480 harbors a p53-stabilizing mutation, but retains p53-dependent transcriptional activity. Opposite, HCT116 p53−/− cells were deficient for p53. First, we analyzed sensitivity of these cell lines to oxaliplatin to identify moderate toxic concentrations. The results show that among the cell lines, LoVo were most resistant, showing no toxicity up to 10 µM oxaliplatin, 96 h after exposure. After treatment with 25 µM oxaliplatin, LoVo cells showed~40% dead cells ( Figure 1A). The same level of toxicity was achieved in HCT116 p53+/+ and SW480 cells upon exposure to 5 µM, as well as in SW48 cells after treatment with 10 µM oxaliplatin. In SW48 and SW480 cells, the first signs of toxicity were observed after exposure to 2.5 µM oxaliplatin. Opposite, HCT116 p53+/+ already showed toxicity after 0.5 µM oxaliplatin. Direct comparison between the isogenic p53-proficient and p53-deficient HCT116 cell lines showed that particularly HCT116 p53+/+ cells were highly sensitive to oxaliplatin. Similar to HCT116 p53+/+ cells, HCT116 p53−/− cells showed toxicity after 0.5 µM oxaliplatin, however, the toxicity did not significantly rise upon treatment with higher oxaliplatin concentrations. Independent of the cell line, an accumulation of the cells in the G2-phase was observed ( Figures 1B and A1 for histograms). The weakest accumulation in G2 was observed in LoVo cells.

Oxaliplatin Induces Proliferation Arrest and Senescence
To analyze at which concentrations oxaliplatin abrogates cellular proliferation, we performed colony formation assays (Figure 2A). A significant loss of the proliferation ability (≤10%) was observed in HCT116 p53+/+ , HCT116 p53−/− and SW48 cells after treatment with 2.5 µM oxaliplatin. SW480 and LoVo cells already showed a dramatic loss in colony formation after exposure to 0.5 µM oxaliplatin. Next, senescence induction was analyzed using microscopy-based detection of β-Gal positive cells ( Figure 2B) and FACS-based detection of C 12 FDG positive cells ( Figure 2C). Among all cell lines, a significant induction of senescence was only observed in LoVo cells. However, the fraction of senescent cells was rather low, reaching only 10-15%. Experiments were performed in triplicates and differences between treatment and control were statistically analyzed using Student's t test (not labeled = not significant, * p < 0.1 ** p < 0.01, *** p < 0.001). Concerning cell cycle distribution, at concentrations > 2.5 µM, a significant (* or **) decrease of cells in G1 and increase in the G2-phase was observed. Experiments were performed in triplicates and differences between treatment and control were statistically analyzed using Student's t test (not labeled = not significant, * p < 0.1 ** p < 0.01, *** p < 0.001). Concerning cell cycle distribution, at concentrations > 2.5 µM, a significant (* or **) decrease of cells in G1 and increase in the G2-phase was observed. ity (≤10%) was observed in HCT116 , HCT116 and SW48 cells after treatment with 2.5 µM oxaliplatin. SW480 and LoVo cells already showed a dramatic loss in colony formation after exposure to 0.5 µM oxaliplatin. Next, senescence induction was analyzed using microscopy-based detection of β-Gal positive cells ( Figure 2B) and FACS-based detection of C12FDG positive cells ( Figure 2C). Among all cell lines, a significant induction of senescence was only observed in LoVo cells. However, the fraction of senescent cells was rather low, reaching only 10-15%.  (C) Senescence was measured by flow cytometry using C 12 FDG staining 120 h after oxaliplatin exposure. (B,C) Experiments were performed in triplicates and differences between treatment and control were statistically analyzed using Student's t test (not labeled = not significant, * p < 0.1, ** p < 0.01, *** p < 0.001).

Oxaliplatin Induces Cell Death, Cell Cycle Arrest, and Senescence at Late Time Points
Since the low frequency of senescent cells could be explained by oxaliplatin-induced senescence being a very late response, we analyzed cell death, cell cycle arrest, and senescence one week (168 h) after treatment. As expected, the observed toxicity was higher than after 96 h. Thus,~20-30% of LoVo, SW48, SW480 and HCT116 p53−/− cells died after treatment with 2.5 µM oxaliplatin ( Figure 3A). Again, HCT116 p53+/+ were most sensitive, showing~75% dead cells after 2.5 µM oxaliplatin. In analogy to shorter exposure to oxaliplatin, an accumulation of the cells in the G2-phase was also observed after longer exposure ( Figure 3B). Similar to the early time points, LoVo cells showed a significant induction of senescence upon treatment with 2.5 and 5 µM oxaliplatin, reaching~40% ( Figure 3C). Those oxaliplatin concentrations also induced senescence in SW48 and SW480 cells; however, the frequency was much lower, with 10% in SW48 and 15% in SW480 cells.
senescence being a very late response, we analyzed cell death, cell cycle arrest, and senescence one week (168 h) after treatment. As expected, the observed toxicity was higher than after 96 h. Thus, ~ 20-30% of LoVo, SW48, SW480 and HCT116 p53-/-cells died after treatment with 2.5 µM oxaliplatin ( Figure 3A). Again, HCT116 p53+/+ were most sensitive, showing ~75% dead cells after 2.5 µM oxaliplatin. In analogy to shorter exposure to oxaliplatin, an accumulation of the cells in the G2-phase was also observed after longer exposure (Figure 3B). Similar to the early time points, LoVo cells showed a significant induction of senescence upon treatment with 2.5 and 5 µM oxaliplatin, reaching ~40% ( Figure 3C). Those oxaliplatin concentrations also induced senescence in SW48 and SW480 cells; however, the frequency was much lower, with 10% in SW48 and 15% in SW480 cells. Experiments were performed in triplicates and differences between treatment and control were statistically analyzed using Student's t test (not labeled = not significant, * p < 0.1, ** p< 0.01, *** p < 0.001). Concerning cell cycle distribution, at concentrations > 2.5 µM, a significant (* or **) decrease of cells in G1 and increase in the G2-phase was observed. (A-C) Experiments were performed in triplicates and differences between treatment and control were statistically analyzed using Student's t test (not labeled = not significant, * p < 0.1, ** p< 0.01, *** p < 0.001). Concerning cell cycle distribution, at concentrations > 2.5 µM, a significant (* or **) decrease of cells in G1 and increase in the G2-phase was observed.

Oxaliplatin Induces Upregulation of SASP and SCAP Factors
To verify the results obtained by β-Gal and C 12 FDG staining, we next measured the activation of the senescence-associated secretory phenotype (SASP), which is characterized by the induction and secretion of different cytokines [28] and represents a clinically relevant mechanism affecting the outcome of cancer therapy. In order to analyze whether oxaliplatin can induce the SASP, the time-dependent expression of different SASP components was analyzed ( Figure 4).
To verify the results obtained by β-Gal and C12FDG staining, we next measured the activation of the senescence-associated secretory phenotype (SASP), which is characterized by the induction and secretion of different cytokines [28] and represents a clinically relevant mechanism affecting the outcome of cancer therapy. In order to analyze whether oxaliplatin can induce the SASP, the time-dependent expression of different SASP components was analyzed ( Figure 4).  In HCT116 p53+/+ , SW48 and SW48 cells a weak induction of IL8 and CCL2 (2-4-fold) was observed after exposure to 2.5 µM oxaliplatin. A weak induction of CCL2 in LoVo cells was observed at 120 h (3-fold), whereas a strong induction of IL8 (8-fold) was observed both at 96 and 120 h. Furthermore, LoVo cells also exhibited an upregulation of CCL8 and IL6 (4-8 fold); induction of IL1α was observed in HCT116 p53+/+ , SW480 and LoVo cells.
Overall, similar to the results on senescence induction obtained by β-Gal and C 12 FDG staining, the strongest activation of the SASP factors was observed in LoVo cells.
An important mechanism leading to senescence induction and its maintenance is the upregulation of anti-apoptotic factors, rendering senescent cells resistant to apoptosis. This mechanism is referred to as senescent-cell anti-apoptotic pathway (SCAP) which might be caused by senescence-associated mitochondrial dysfunction (SAMD) (for review see [29,30]). In order to analyze whether oxaliplatin induces the SCAP factors, the timedependent expression of different anti-apoptotic factors was analyzed by real-time qPCR ( Figure A2). In all cell lines, a weak (2-fold) transcriptional activation of BcL-2 was observed at 120 h. SW480 cells exhibited an 8-fold Bcl-2 induction. Of note, a strong difference in the expression of BIRC2 (encodes c-IAP1), BIRC3 (encodes c-IAP2) and BIRC5 (encodes Survivin) was observed in LoVo cells, with a 16-fold upregulation of BIRC2 and BIRC3 and an 8-fold downregulation of BIRC5. This pronounced transcriptional repression of the Survivin gene also shows that the cells must be significantly compromised in their proliferation potential, presumably undergoing senescence [31,32]. Beside anti-apoptotic factors, we also compared the transcriptional activation of pro-apoptotic factors among the CRC cell lines upon oxaliplatin ( Figure A3). Also, in this case, clear differences were observed. Corresponding to the high toxicity, an induction of the genes coding for the p53-dependent pro-apoptotic factors FASR, PUMA, and NOXA, as well as of the p53 target MDM2 and the AP1-dependent pro-apoptotic factor FASL was observed in HCT116 p53+/+ cells, whereas HCT116 p53−/− cells only showed induction of the FASL mRNA. Interestingly the highest induction of all these factors, and of the p53-dependent pro-apoptotic factor BAX, was observed in LoVo cells, indicating a more efficient activation of the p53 pathway than in SW48 and SW480 cells.

Activation of p53 and p21 CIP1 upon Oxaliplatin Treatment
One possibility to explain the different efficiency of oxaliplatin to induce senescence in CRC cells would be a dissimilar activation of the DDR. Thus, we next analyzed phosphorylation/activation of p53, as well as induction of p21 CIP1 , 24 and 48 h after exposure to 2.5 µM oxaliplatin. As shown in Figure 5A, the initial induction of p53 and its phosphorylation is comparable between SW48, SW480, and LoVo cells. Moreover, in HCT116 p53+/+ cells, induction and phosphorylation was observed, which was however rather weak, as compared to the other cells. As expected, HCT116 p53−/− cells showed no expression of p53. Concerning expression of p21 CIP1 , SW48, SW480, LoVo and HCT116 p53+/+ , but not HCT116 p53−/− cells showed a pronounced induction. The data indicate that the initial induction of p53 and p21 CIP1 is comparable between all p53-proficient cell lines and, therefore, cannot explain the differences in senescence induction between these cells. Next, we analyzed the persistence of p53 activation and p21 CIP1 induction 120 h after oxaliplatin exposure, showing only in LoVo cells a sustained activation of p53 and upregulation of p21 CIP1 ( Figure 5B). Similar results were also observed at mRNA level. Whereas in LoVo cells a 20-fold induction of p21 CIP1 was observed, SW48, SW480, and HCT116 p53+/+ cells showed only a 2-3-fold induction, and in HCT116 p53−/− cells the amount of p21 CIP1 mRNA was reduced ( Figure 5C). The data suggest that LoVo cells utilize a specific mechanism reinforcing the induction of p53 response.

Expression of p14 ARF and p16 INK4A upon Oxaliplatin Treatment
Two crucial factors involved in senescence induction and its maintenance are p14 ARF and p16 INK4a , which are encoded by different reading frames of the CDKN2A gene. Thus, we next analyzed the expression and induction of p16 INK4a and p14 ARF 120 h after exposure to oxaliplatin ( Figure 6A). qPCR experiments showed a 3-fold upregulation of p16 INK4a in HCT116 p53+/+ cells but no induction in the other cell lines. In addition, striking differences were obtained at the basal level expression. Thus, comparing the Cq values revealed only in HCT116 cells a reliable expression of the p16 INK4a mRNA (Cq: 28), whereas the Cq values for SW48, SW480, and LoVo (Cq: 36,36,38, respectively) were much higher, indicating a low p16 INK4a expression. Similar results were obtained concerning p14 ARF expression. In this case a strong p14 ARF expression was observed in LoVo (Cq: 26) and HCT116 (Cq: 24, 25) cells. Apart from mRNA expression, we also analyzed the epigenetic silencing of p14 ARF and p16 INK4a ( Figure 6B). The results obtained by methylation-specific PCR (MSP) revealed a complete methylation and, thereby, silencing of p16 INK4a in SW48, SW480, and LoVo cells, whereas HCT116 cells were hemi-methylated. Since only SW48, SW480, and LoVo cells underwent senescence, an impact of p16 INK4a on senescence induction can be excluded. Of note, also HCT116 cells are p16 INK4a deficient, independent of mRNA expression; in HCT116 cells one allele is silenced via methylation, whereas the other allele is mutated, showing a base deletion in the codon 33 within the promoter/exon 1β region, generating a stop codon at codon 47 [33]. The same is also true for p14 ARF in HCT116 cells. Interestingly, opposite to SW48 and SW480 cells, the p14 ARF promoter of LoVo cells is hemi-methylated ( Figure 6B). We should note that for SW48 and LoVo cells the p14 ARF methylation and mRNA expression status is in line with the literature, whereas SW480 have been reported to be unmethylated [34], although in our hands, they are methylated for p14 ARF . Since no mutations are known for p14 ARF in LoVo cells, we verified the p14 ARF and p16 INK4a status by immunodetection, showing that among all CRC cell lines tested only LoVo express p14 ARF protein, indicating LoVo cells to be p14 ARF proficient ( Figure 6C). Expression of p16 INK4a was observed in none of the cell lines (data not shown).

Impact of p53 and p14 ARF on Oxaliplatin-Induced Cellular Senescence
To analyze whether the expression of p14 ARF is associated with sustained activation of the p53/p21 CIP1 response in LoVo cells, p14 ARF was downregulated via siRNA. Furthermore, to verify that oxaliplatin-induced senescence requires p53, p53 was inhibited using 30 µM pifithrin α (PTHα). Under these conditions, inhibition of p53 completely blocked the oxaliplatin-induced increase in p21 expression, and knockdown of p14 ARF was accompanied by an abrogated oxaliplatin-induced stabilization and phosphorylation of p53 as well as by abrogated p21 CIP1 induction ( Figure 7A), indicating that p14 ARF is involved in stabilization and phosphorylation of p53 and, thereby, in the sustained upregulation of p21 CIP1 . Next, we measured the oxaliplatin-triggered induction of senescence upon inhibition of p53 and knockdown of p14 ARF ( Figure 7B). As expected, inhibition of p53 nearly completely abolished senescence induction. Most importantly, also knockdown of p14 ARF significantly reduced onset of senescence to a level observed in SW48 and SW480 cells.
ported to be unmethylated [34], although in our hands, they are methylated for p14 ARF . Since no mutations are known for p14 ARF in LoVo cells, we verified the p14 ARF and p16 INK4a status by immunodetection, showing that among all CRC cell lines tested only LoVo express p14 ARF protein, indicating LoVo cells to be p14 ARF proficient ( Figure 6C). Expression of p16 INK4a was observed in none of the cell lines (data not shown).

Impact of p53 and p14 ARF on Oxaliplatin-Induced Cellular Senescence
To analyze whether the expression of p14 ARF is associated with sustained activation of the p53/p21 CIP1 response in LoVo cells, p14 ARF was downregulated via siRNA. Furthermore, to verify that oxaliplatin-induced senescence requires p53, p53 was inhibited using 30 µM pifithrin α (PTHα). Under these conditions, inhibition of p53 completely blocked the oxaliplatin-induced increase in p21 expression, and knockdown of p14 ARF was accom- ylation of p53 as well as enhanced p21 induction was only observed in p14 ARF expressing SW48 and SW480 cells after oxaliplatin treatment. Similar to the phosphorylation of p53 and induction of p21 CIP1 , a significantly enhanced frequency of senescence was only observed in p14 ARF expressing SW48 and SW480 cells after oxaliplatin treatment ( Figure 7D). Obviously, re-expression (rescue) of p14 ARF enhances stability of p53, however it seems not to be sufficient to induce senescence in absence of DNA damage.

Discussion
Colorectal cancer (CRC) is caused by old age [13] and lifestyle factors, such as diet (red meat, processed meat, and alcohol) [19], smoking [23], obesity, and lack of physical activity, and is also associated with diseases, e.g., inflammatory bowel disease [14,15]. Both sexes combined, colon cancer comprises for 6.1% of all cancer cases and for 9.2% of To analyze whether expression of p14 ARF could enhance senescence frequency in SW48 and SW480 cells, the cells were transiently transfected with a p14 ARF expression plasmid [14] ( Figure 7C). Under these conditions, p14 ARF was detectable in SW48 and SW480 cells 120 h after transfection. The expression of p14 ARF was accompanied by an increased stabilization of p53 independent of oxaliplatin treatment. Opposite, enhanced phosphorylation of p53 as well as enhanced p21 induction was only observed in p14 ARF expressing SW48 and SW480 cells after oxaliplatin treatment. Similar to the phosphorylation of p53 and induction of p21 CIP1 , a significantly enhanced frequency of senescence was only observed in p14 ARF expressing SW48 and SW480 cells after oxaliplatin treatment ( Figure 7D). Obviously, re-expression (rescue) of p14 ARF enhances stability of p53, however it seems not to be sufficient to induce senescence in absence of DNA damage.

Discussion
Colorectal cancer (CRC) is caused by old age [13] and lifestyle factors, such as diet (red meat, processed meat, and alcohol) [19], smoking [23], obesity, and lack of physical activity, and is also associated with diseases, e.g., inflammatory bowel disease [14,15]. Both sexes combined, colon cancer comprises for 6.1% of all cancer cases and for 9.2% of cancer related deaths. In addition, rectal cancer comprises for 3.9% of all cancer cases and 3.2% of cancer related deaths [35,36]. Thus, CRC represents the third most frequent cancer and the second frequent cause of cancer-related death worldwide; the five-year survival rate amounts to 65% [35,36]. First-line therapy consists of fluoropyrimidines, such as 5-fluorouracil (5-FU) or capecitabine alone or in combination with leucovorin. Furthermore, either oxaliplatin (FOLFOX) or irinotecan (FOLFIRI) are co-administered [37,38]. The second-line therapy is performed by switching FOLFOX into FOLFIRI or the other way around [39]. FOLFOXbased therapy [40] has improved the response rate from 20 to 50% compared with 5-FU alone, and also the survival rate of metastatic patients [41,42]. During FOLFOX-based therapy, 85 mg oxaliplatin pro m 2 /body surface area is applied intravenous every two weeks for a total of 12 cycles.
Opposite to cisplatin, which is known to induce senescence in different tumor entities [43][44][45][46], the data concerning oxaliplatin-induced senescence are very limited. Thus, an induction of senescence was observed in the mouse cell line CT26 and in the rat cell line PROb [47]. Moreover, oxaliplatin treatment of tumor-bearing rats induced the expression of the SASP factors IL6, IL8, CXCL1, CXCL2, and MMP [47]. In human cancer cell lines, mostly negative results were obtained. Thus, HCT116 and SW480 cells induced senescence upon doxorubicin, 5-FU and irinotecan but not upon oxaliplatin exposure [48] and senescence could be strongly induced in various cancer cell lines (A549, SH-SY-5Y, HCT116, MDA-MB-231, and MCF-7) by doxorubicin, irinotecan, and methotrexate, whereas oxaliplatin and 5-FU failed to do so [49]. Therefore, it is not yet clear whether oxaliplatin can induce senescence in CRC cells and whether p53 is implicated in this pathway. Here, we show that oxaliplatin can induce senescence in CRC cells; however, this is highly cell line-specific.
To shed light on the question concerning oxaliplatin-induced senescence, we analyzed cell death, senescence induction as well as transcriptional activation of SASP factors and factors associated with SCAP and apoptosis in p53-proficient (p53 wild-type) SW48, HCT116 p53+/+ , and LoVo cells. Moreover, we included in the study SW480 cells, which harbor a p53 stabilizing mutation but retain p53-dependent transcriptional activity, and p53-deficient HCT116 p53−/− cells. All cell lines reacted to oxaliplatin with the induction of cell death, LoVo and HCT116 p53−/− cells exhibited the weakest response, whereas toxicity was most profound in HCT116 p53+/+ cells. The difference between the isogenic HCT116 cells is in line with data showing that p53 sensitizes cells against oxaliplatin-induced cell death [50]. Irrespective of the cell line, the cells accumulated in the G2-phase of the cell cycle. Again, the weakest response was observed in LoVo cells. Correspondingly, the proliferation capacity determined via colony formation was most strongly abrogated in LoVo cells. In line with this, only LoVo cells showed induction of senescence, determined via β-Gal and C 12 FDG staining 120 h after oxaliplatin exposure. Similar results were obtained at later time points (168 h), at which the fraction of senescent cells was further enhanced in LoVo cells. At those later times, SW48 and SW480 cells also displayed a weak, but significant, induction of senescence. The induction of senescence observed in LoVo cells was accompanied by transcriptional activation of SASP components CCL8, IL6, and IL8, and of the anti-apoptotic SCAP factors BIRC2 and BIRC3. Interestingly, in LoVo cells, the highest increase of different p53-dependent pro-apoptotic factors (BAX, FASR, PUMA, NOXA) and of the p53 target MDM2 was observed, pointing to a more pronounced activation of the p53 pathway, in comparison with the other cell lines.
To elucidate the molecular mechanism underlying the differential activation of senescence between the p53-proficient CRC cells, we compared activation of the DDR 24 and 48 h upon oxaliplatin. In all proficient cell lines, but not in HCT116 p53−/− cells, activation/phosphorylation of p53 and induction of p21 CIP1 was observed at early time points after oxaliplatin treatment. Importantly, at later times (120 h), the activation/phosphorylation of p53 and induction of p21 CIP1 was only observed in LoVo cells, indicating differential regulation of the p53/p21 CIP1 pathway among the p53-proficient lines.
Beside p21 CIP1 , p16 INK4a and p14 ARF also play important roles in cellular senescence. However, in oxaliplatin-induced senescence in CRC cells, an involvement of p16 INK4a could be excluded since all cell lines are p16 INK4a deficient. In contrast, we observed that LoVo cells are proficient for p14 ARF , showing a hemimethylated promoter, as well as a strong expression of the mRNA and the protein. Opposite to p16 INK4a , the role of p14 ARF in senescence is less clear. Thus, ectopic expression of p14 ARF induced a senescent phenotype in normal human fibroblasts [51] and p19 ARF (homolog of human p14 ARF ) deficient mouse fibroblasts do not show oncogenic H-Ras induced senescence [52]. However, the impact of p14 ARF may differ between human and rodent fibroblasts, at least during oncogenic senescence. Hence, p14 ARF was upregulated by c-Myc in senescent MEFs [53], but not upon H-Ras in human fibroblasts [54]. Moreover, ectopic expression of p14 ARF was shown to arrest p21 CIP1 deficient MEFs [55], but not p21 CIP1 deficient human fibroblasts [54]; Ras can upregulate p14 ARF in rodent cells [56,57], but not in human fibroblasts [54]. Since it was previously shown that c-Myc-induced senescence in human primary cells requires stabilization of p14 ARF [58], the observed differences can also be caused by differences in the oncogenes used for senescence induction. Concerning the impact of p14 ARF on DNA damage-induced senescence it has been suggested that p14 ARF plays a role in the long-term maintenance of p53 activity following ionizing radiation [59]; however, its expression seems not to be dependent on DNA damage [15]. Our data suggest that p14 ARF may be important for oxaliplatin-induced senescence in CRC cells. Therefore, the impact of p14 ARF on p53 stability and senescence maintenance was analyzed by knockdown approach. As positive control, p53 was inhibited. The data clearly show that p53 inhibition nearly completely blocks oxaliplatin-dependent p21 CIP1 expression and senescence. Importantly, knockdown of p14 ARF abrogated oxaliplatin dependent expression/activation of p53 and of p21 CIP1 at later times and significantly reduced senescence to a level similar to SW48 and SW480. Vice versa, ectopic expression of p14 ARF enhanced oxaliplatin-induced senescence in SW48 and SW480 cells. Of note, the importance of p14 ARF for anticancer drug-induced senescence seems not to be restricted to oxaliplatin since, in the case of the topoisomerase I (TOP1) inhibitor irinotecan LoVo cells showed the strongest induction of senescence ( Figure A4).
Overall, our data clearly show that the assumption that oxaliplatin cannot induce senescence in CRC cells is not true. We observed induction of senescence in three out of four p53-proficient CRC lines. However, the efficiency in inducing senescence strongly differed between these cell lines. Thus, we could show that proficiency in p14 ARF mediates sustained p53 activity/expression and p21 CIP1 expression and, thereby, reinforces senescence induction. Of note, in parallel also an increased activation of p53-dependent pro-apoptotic pathways was observed. However, the p53-dependent activation of apoptosis seems to be efficiently blocked in LoVo cells, which could be explained by the strong induction of BIRC2/c-IAP1 and BIRC3/c-IAP2, probably representing the central component of the SCAP during oxaliplatin-induced senescence. This, however, has to be addressed in future work. Our data also suggest that p14 ARF might influence the decision between cell death and senescence in primary CRC. Moreover, differential epigenetic silencing of p14 ARF and/or p16 INK4A among cell lines from different labs could explain the heterogeneity in the data published concerning anticancer-drug-induced senescence. Opposite to cell lines, most primary CRCs are p14 ARF proficient. Thus, meta-analysis showed p14 ARF methylation in~25% of all cases [60]. The impact of p14 ARF on chemotherapy-induced senescence in CRC will be an important topic for future studies.

Acknowledgments:
The authors would like to thank Bert Vogelstein (Johns Hopkins University, Baltimore, ML, USA) for providing HCT116 p53+/+ and HCT116 p53−/− cell lines, and Birgit Rasenberger for excellent technical assistance.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.  Figure A1. Cell death and cell cycle distribution were measured by flow cytometry 96 h upon exposure to oxaliplatin in PI-stained CRC cells. Representative histograms were shown for untreated cells (con) and cells exposed to 2.5 µM oxaliplatin. Figure A1. Cell death and cell cycle distribution were measured by flow cytometry 96 h upon exposure to oxaliplatin in PI-stained CRC cells. Representative histograms were shown for untreated cells (con) and cells exposed to 2.5 µM oxaliplatin.