The rational modulation of autophagy sensitizes colorectal cancer cells to 5‐fluouracil and oxaliplatin

Colorectal cancer (CRC) is the third most common and deadliest cancer globally. Regimens using 5‐fluorouracil (5FU) and Oxaliplatin (OXA) are the first‐line treatment for CRC, but tumor recurrence is frequent. It is plausible to hypothesize that differential cellular responses are triggered after treatments depending on the genetic background of CRC cells and that the rational modulation of cell tolerance mechanisms like autophagy may reduce the regrowth of CRC cells. This study proposes investigating the cellular mechanisms triggered by CRC cells exposed to 5FU and OXA using a preclinical experimental design mimicking one cycle of the clinical regimen (i.e., 48 h of treatment repeated every 2 weeks). To test this, we treated CRC human cell lines HCT116 and HT29 with the 5FU and OXA, combined or not, for 48 h, followed by analysis for two additional weeks. Compared to single‐drug treatments, the co‐treatment reduced tumor cell regrowth, clonogenicity and stemness, phenotypes associated with tumor aggressiveness and poor prognosis in clinics. This effect was exerted by the induction of apoptosis and senescence only in the co‐treatment. However, a week after treatment, cells that tolerated the treatment had high levels of autophagy features and restored the proliferative phenotype, resembling tumor recurrence. The pharmacologic suppression of early autophagy during its peak of occurrence, but not concomitant with chemotherapeutics, strongly reduced cell regrowth. Overall, our experimental model provides new insights into the cellular mechanisms that underlie the response and tolerance of CRC cells to 5FU and OXA, suggesting optimized, time‐specific autophagy inhibition as a new avenue for improving the efficacy of current treatments.

Colorectal cancer (CRC) is among the three most prevalent types of cancer and one of the most frequent cause of cancer-related deaths globally, reaching more than 1 million people per year. 1 Usually, the CRC originates from an initial benign adenomatous polyp, which turns into an advanced adenoma with high-grade dysplasia (stage I and II tumors).At these stages, the disease is considered curable, but if untreated, it can accumulate new molecular alterations and spread to the regional lymph nodes (stage III) or establish metastasis in to distant sites (stage IV). 2 The first-line therapy for colorectal neoplasms depends on the stage of the disease.Stage I or II are treated by surgical excision, while the treatment of stage III and IV tumors consists of surgery followed by adjuvant chemotherapy.Recent advances have improved the survival of colorectal neoplasms patients in the initial stages of the disease, but stage IV remains incurable. 3linically, one of the most effective pharmacological treatments for stage IV CRC tumors is the protocol called FOLFOX, combining 5-fluorouracil (5FU), Oxaliplatin (OXA) and leucovorin. 4,5As a single treatment, OXA has demonstrated minimal to negligible efficacy on CRC patients. 6Additionally, compared to 5FU alone, the cotreatment improved the median overall survival for metastatic CRC from 11 months to up to 3 years. 7he 5FU is the most important chemotherapeutic used in CRC.It is an antimetabolite drug that inhibits the thymidylate synthetase, thus blocking the conversion of deoxyuridylic acid to thymidylic acid and, finally, the DNA biosynthesis.Moreover, 5FU metabolites can incorporate into the DNA and RNA, leading to mispairing and failures in the transcription, respectively. 8OXA, in turn, induces DNA interstrand cross-linking, leading to DNA brakes. 9The cellular responses acutely triggered by 5FU or OXA in CRC cells mainly include cell cycle arrest and apoptosis. 10,11However, the response of cancer cells to 5FU or OXA varies based on their genetic status. 11,12evertheless, in vitro investigations into the cellular mechanisms underlying the response and resistance of CRC cells to FOLFOX have predominantly focused on the period immediately following treatment, despite this co-treatment being administered over 2 days with a 2-week patient recovery period. 2,13As a result, the kinetics and interplay between cellular mechanisms during the recovery period have been overlooked.In this study, we sought to evaluate both the acute and chronic response of CRC cells to 5FU + OXA using a schedule that mimics the clinical treatment.Our main objective was to understand the cellular mechanisms triggered by these drugs alone or combined over time, including not only the period of exposure to them, but also the subsequent days without their presence.

| Cell culture
HCT116 and HT29 cell lines were kindly provided by Dr. Annette Larsen from INSERM (Institut National de la Santé et de la Recherche Médicale, Paris, France).The genetic status of key tumor suppressors genes and oncogenes in these cell lines is shown in Table S1 (data from Ilyas et al. 14 and Liu & Bodmer 15 ).Cells were maintained in Dulbecco's modified Eagle medium (DMEM) low glucose supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin and 0.1% gentamicin, at 37°C and 5% CO 2 in a humidified incubator.

| Cell treatments
The CRC cell lines were subjected to the following doses: HCT116-10 µM of 5FU or 5 µM of OXA or both together; HT29-20 µM of 5FU or 10 µM of OXA or both together.The selected doses were based on previous studies of 5FU [16][17][18] and OXA, [19][20][21] and remain below the maximum plasma concentration (C max ) limit for both drugs (5FU, C max : 420 µM 22,23 ; OXA, C max : 90 µM). 24,25ontrol cells were exposed to the vehicle (dimethyl sulfoxide) at a maximum concentration of 0.5% (vol/vol).Hereafter, all analyses performed at the end of the drugexposure period (48 h) were referred to as "acute" analyses or responses (Figure S1-black dot).For longterm experiments, after the drug-exposure period (48 h), the remaining cells were washed twice with phosphatebuffered saline (PBS) 1x, harvested, counted, reseeded in drug-free medium (DFM) for more 14 days, representing the recovery period of patients in the FOLFOX clinical schedule (Figure S1, colored dots).
Autophagy was inhibited using 3-methyladenine (3MA) 2 mM (M9281, Sigma-Aldrich) in two ways: as a pretreatment (1 h before chemotherapeutics-Figure S1, blue arrowheads) or as three sequential expositions of 1 h on each the 3rd, 4th, and 5th day after the end of drugs exposition (Figure S1, green arrowheads), which corresponds to the peak of autophagy features.On both conditions, after each incubation of 1 h, the 3MAcontaining medium was withdrawn, and a complete DMEM was added to the cells.

| Cumulative population doubling (CPD)
After the treatment (48 h), cell counting was performed in a Neubauer chamber to assess the acute cytotoxicity.Subsequently, the remaining cells were seeded in DFM.After 3, 5, 7, 10, and 15 days, the number of cells and the CPD were determined, as previously described 26 according to the formula PD = [ln N(t) − lnN(to)]/log 2, where N(t) is the number of cells per well at the time of the cell counting (passage) and N(to) is the initial number of seeded cells.The sum of PDs was then plotted in a graph against the time of culture to assess the long-term proliferation profile through CPD.

| Clonogenic assay
For the clonogenic assay, cells were treated with 5FU, OXA, or co-treatment for 48 h, maintaining the remaining cells in DFM.After 7 days in this condition, these cells were harvested, counted, and reseeded at 10 2 cells/ well density in a 6-wells plate.After growing for more 12 days, colonies were fixed with cold methanol and stained with 1% crystal violet (see Figure S1).The clonogenic 6-wells plates were photographed to count the number of colonies with at least 50 cells and to measure their area using the Image-Pro Plus 6.0 (IPP6) software.

| Cell cycle analysis
For cell cycle analysis, 48 h after treatment, the remaining cells were harvested and centrifuged at 1400 rpm for 5 min.The supernatant was discarded, and the pellet was resuspended with cold PBS 1x and centrifuged at 1400 rpm for 5 min.After discarding the supernatant again, the cells were fixed with cold ethanol 70% (vol/vol in PBS) under soft vortexing and stored at −20°C for 2 h for appropriate fixation.Fixed cells were centrifuged at 1400 rpm for 5 min, washed once with PBS, and the pellet was stained with a solution containing 50 μg/mL propidium iodide (PI), 0.1% Triton X-100 and 50 μg/Ml RNAse for 30 min, in the dark, at room temperature.Marked cells were analyzed using the Attune flow cytometer (Thermo Fisher Scientific).Doublets were excluded in a "PI area versus PI width" plot.

| Annexin V and PI staining
Apoptosis was assessed using the Annexin V-FITC and PI (kit da Santa Cruz Biotechnology), according to the manufacturer's protocol.Briefly, cells were treated with 5FU, OXA or co-treatment for 48 h, and after maintained in DFM.Three days later (see Figure S1), the supernatant and harvested cells were transferred to a tube and centrifuged at 1400 rpm for 5 min.The pellet was washed once with 1x PBS and centrifuged at 1400 rpm for 5 min.The supernatant was discarded, and the annexin binding buffer containing annexin (2.5 μL/sample) and PI (3 μM/sample) was added to the pellet.Cells were incubated at room temperature, in the dark, for 15 min.Marked cells were analyzed using the Attune flow cytometer.

| SA-β-gal activity measurement by C12-FDG flow cytometry
For senescence measurement, cells were stained with 5-dodecanoylaminofluorescein di-beta D-galactopiranoside (C12-FDG, Life Technologies), a fluorescent substrate of the Senescence-Associated Acid β-Galactosidase (SA-β-gal) that emits green fluorescence when cleaved by the enzyme. 27Briefly, cells were treated with 5FU, OXA, or co-treatment for 48 h, followed by the maintenance of remaining cells in DFM.After 7 days in this condition, cells were incubated with C12-FDG for 2 h at 33 μM at 37°C.Marked cells were then trypsinized and analyzed using the flow cytometer.Results are presented as the percentage of SA-beta-gal-positive cells.

| SA-β-gal activity measurement by X-gal chromogenic staining
For X-gal staining, after treatments (48 h) cells were washed twice with PBS 1x and replated at a density of 2 × 10 4 cells/well, in a 12-wells plate.After 7 days, cells were tested for SA-β-gal activity as described. 28Briefly, cells were washed with PBS 1x, fixed with 2% paraformaldehyde for 30 min at room temperature, and incubated with fresh SA-β-gal staining solution (1 mg/mL X-gal substrate (Sigma-Aldrich), 40 mM citric acid/sodium phosphate (pH 6.0), 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM NaCl, and 2 mM MgCl) for 6 h at 37°C.Then, cells were marked with a solution containing 300 nM DAPI and 0.1% triton X-100 (vol/vol in PBS) for 30 min at room temperature.Results are presented as ratio of SA-β-gal-positive cells to total cells.

| Detection of acidic vesicular organelles
Acridine orange (AO) is an acidotropic probe that has been used as a marker of late autophagy. 29,30o evaluate the kinetics of autophagy in response to treatment, cells were treated with 5FU, OXA, or co-treatment for 48 h, followed by the maintenance of remaining cells in DFM.At indicated days (see Figure S1), cells were harvested and incubated with 2.7 μM of AO for 15 min, in DMEM, at room temperature and after analyzed by flow cytometry.Data are presented as the percentage of AO-positive cells.Red fluorescence intensity was also measured to assess the intensity of late autophagy in single cells.The area under the curve was calculated using the GraphPad Prism software.

| Flow cytometry for SQSTM1 levels
Cells were treated with 5FU, OXA, or co-treatment for 48 h, followed by the maintenance of remaining cells in DFM.After 5 or 10 days, cells were harvested and fixed with 4% formaldehyde (vol/vol in PBS) for 10 min at room temperature, followed by 1 min on ice.After this, cells were permeabilized with cold methanol 90% for at least 30 min on ice.Then, cells were centrifuged and resuspended in incubation buffer (0.5% bovine serum albumin [BSA] in PBS) for 10 min, followed by addition of primary antibody anti-SQSTM1 (Abcam), in a final dilution of 1:200, for 3 h at room temperature.Thereafter, cells were washed with incubation buffer once, followed by incubation with secondary antibody conjugated to phycoerythrin (PE), diluted 1:1000, for 30 min at room temperature in the dark.Finally, cells were washed once with incubation buffer, resuspended, and analyzed by flow cytometry.

| Stmeness immunophenotyping CD133 and CD44 flow cytometry
To assess the prevalence of CD133 and CD44-positive cells, cells were treated with 5FU, OXA or co-treatment for 48 h, followed by medium changing and culturing in DFM.The percentage of cancer stem cells (CSCs) was analyzed after 7 days at this condition.Briefly, cells were dissociated and washed once with PBS 1x containing ethylenediaminetetraacetic acid 2 mM and 0.5% of BSA for 1 h.After this, cells were incubated with PE-conjugated anti-CD133 and FITCconjugated CD44 (Miltenyi Biotec) 1:100 (vol/vol) and 1:200 (vol/vol) respectively, for 10 min.Finally, cells were washed once with incubation buffer, resuspended, and analyzed by flow cytometry.

| Statistical analysis
All experiments were performed at least three times independently.For comparison of the averages, we conducted t test or analysis of variance followed by Tukey test.Analyses were performed using the SPSS 18.0 software program.All p values under 0.05 were considered significant.

| 5-FU and OXA co-treatment exert an additive cytotoxic effect only in the long term in CRC cells
We initially assessed the acute effect of the co-treatment with 5FU + OXA comparing with both drugs alone.We defined the doses of 20 µM for 5FU and 10 µM for OXA for both cell lines based on previous studies and the serum dose reached in the blood, as mentioned in the materials and methods section.For HT29 cells, these doses allowed the development of our experimental design (Figure S1); however, they were very cytotoxic for the HCT116 cell line, so the evaluations of cellular responses during the 2 weeks after treatment were impossible due to the low number of tolerant cells.Due to this, for HCT116 cells, we used the doses of 10 µM for 5FU and 10 µM for OXA.We found a significant decrease of cells with viable phenotype after treatment for all analyzed conditions (Figure 1A,B), an increase in the number of round-shape and fragmented cells (Figure S2), and an increase in intracellular granularity (Figure 1A).For both cell lines, after 48 h the co-treatment did not reduce cell number in higher extent than 5FU alone (Figure 1B).On the contrary, the co-treatment was more potent in reducing cell growth of both cell lines during the following 2 weeks after the end of treatment compared to treatments with drugs alone (Figure 1C,D).Indeed, in comparison to cells treated with single drugs, the main control population practically disappeared in the co-treatment (Figure S3A and S3C, dashed circle; Figure S3B and S3D).However, corroborating the CPD, this population reappeared at day 10 in both cell lines.For all these phenotypic changes, the HCT116 strain was more sensitive even to half the concentrations used for the HT29 strain.CPD and flow cytometry data suggest around day 7 after treatment as the return point of population growth.We then performed a clonogenic assay to verify the proliferative capacity of single cells at that time (Figure 1E).In both cell lines, the co-treatment was significantly more effective than the isolated drugs in reducing the clonogenic capacity of the cells (Figure 1F) and the size of resulting colonies (Figure 1G).These results suggests that co-treatment reduced the clonogenicity of CRC cells in comparison to treatments alone and that the remaining cells in the co-treatment are less proliferative than surviving cells in the single treatments.
CSCs appear to be related to the resistance of cancer to therapy. 31CD133 and CD44 are the most used markers for CSCs analysis in CRC, [32][33][34][35] been associated with cancer progression, 36 poor prognosis 37 and resistance to therapy. 38Then, we investigated a possible enrichment of CSCs after single treatments or co-treatment (Figure 1H and Figure S3E,F).We observed a large enrichment of CD133 and CD44-positive cells after 7 days of OXA treatment for both cell lines (Figure 1H).On the other hand, there was no enrichment of CSCs after cotreatment in both cell lines, suggesting that the cotreatment avoid the enrichment of CSC observed after treatments with single drugs.

| OXA and 5FU affects cell cycle distribution in CRC cells
Aiming to investigate the mechanisms underlying the combined effect of 5FU and OXA, we assessed cell cycle distribution after 48 h of treatment.OXA induced the accumulation of cells in G2/M and S-G2/M in HCT116 (Figure 2A) and HT29 cells (Figure 2B), respectively.5FU and the co-treatment induced the accumulation of cells in the G1 phase in the HT29 cell line, which was not observed in HCT116 cells.In conclusion, the effect of 5FU on the cell cycle prevails in the co-treatment; the acute modulation of the cell cycle does not correlate with the chronic response of CRC cells to the treatments.

| 5FU and OXA do not exert an additive proapoptotic effect in CRC cells
It is well known that 5FU and OXA trigger apoptosis in cancer cells.Indeed, cell phenotype suggested apoptosis induction by 5FU and co-treatment, in higher extent in HCT116 cells than in HT29 (Figure S2).Thus, we objectively assessed apoptosis triggering through annexin V and PI labeling 3 days after de end of treatment.5FU triggered high levels of apoptosis in both cell lines, while significant levels of apoptosis in response to OXA were observed only in HCT116 cells.However, the percentage of apoptotic cells did not differ between 5FU and the cotreatment revealing the absence of combined effect in apoptosis induction (Figure 2C,D).

| Co-treatment induces senescence in CRC cell lines
CPD (Figure 1C,D) and flow cytometry data (Figure S3A-D) suggested a population growth arrest after treatments, mainly up to day 7 after treatment.Thus, other antiproliferative mechanism than apoptosis could be influencing the reduction in cell number, such as senescence.To evaluate this process, at day 7 after the end of treatment, the remaining cells were incubated with the β-galactosidase enzyme substrate, C12-FDG, which emits green fluorescence when cleaved, followed by flow cytometry.Complementarily, we also evaluated the activity of the enzyme through incubation with the chromogenic substrate x-gal followed by image acquisition and analysis.We observed an increase in the percentage of SA-β-gal-positive cells in response to OXA or 5FU, to a greater extent in HCT116 cells than in HT29 cells but with no difference between cotreatment and single treatments (Figure S4A,B, respectively,-blue staining; Figure S4C,D-green area).
We then looked at cell size (forward-scatter [FSC]) combined with C12-FDG staining.Through this, we found that only in HCT116 cells exposed to the cotreatment, but not in HT29, the SA-β-gal-positive cells acquired an enlarged phenotype (Figure 2E,F).Altogether, these data suggest that the co-treatment, but not single treatments, was able to induce an increase in typical senescent markers in CRC cells, in higher extent in HCT116 than in HT29 cells.

| 5FU and OXA co-treatment induce a transient increase of autophagy markers at the long-term
As shown in FSC × side-scatter (SSC) data from flow cytometry, we observed a transitory increase in intracellular granularity (i.e., SSC), reaching a peak from days 3 to 5 after the end of treatments (Figure S3A and S3C).It is plausible to infer that cells that survived the acute treatment triggered a dynamic stress response associated with the increase of intracellular vesicles, like autophagosomes.We then screened autophagy through AO staining. 30Corroborating data from SSC, in the long term we observed a transient increase in the percentage and the intensity of AO-positive cells in a greater extent to co-treatment compared to single treatments (Figure 3A,B).
To complement the AO data, we evaluated SQSTM1 (p62) protein levels using flow cytometry (Figure S5A-D).SQSTM1 is the main adapter that marks cellular components to be degraded.It binds to LC3 protein and incorporated into the autophagosome. 39After its fusion with lysosomes, p62 is degraded, so that a reduction in its levels demonstrates that the autophagic flux is functioning properly in the cell. 29,40We found a significant reduction in SQSTM1 levels 5 days after treatment, with greater intensity for co-treatment.Already 10 days after treatment, SQSTM levels were similar to the levels of control cells, with the exception of co-treatment, corroborating the AO and SSC data.

| The rational inhibition of autophagy sensitizes CRC cells to 5FU and OXA
Autophagy may play a dual role in the carcinogenic process by suppressing tumor initiation but favoring tumor progression. 41In response to therapy, autophagy acts, in most cases, as a prosurvival mechanism, thus favoring tolerance and tumor recurrence. 42With the results from AO in mind, we next pharmacologically inhibited autophagy with 3MA.The treatment with 3MA before the chemotherapeutics (3MA pre) did not alter the long-term growth of CRC cells (Figure 3C and 3E, dashed lines in the left graphs), Notably, autophagy inhibition during its peak of activation on days 3-5 (3MA peak), to AO staining (Figure 3A,B) and SSC (Figure S3A,B), strongly reduced the CPD of both cell lines in all treatments but to a greater extent in the cotreatment (Figure 3C and 3E, dashed lines in the right graphs).The comparison between the final CPD at day 15 without autophagy inhibition, 3MA-pre, and 3MA peak is shown in Figure 3D and 3F.Autophagy inhibition during its peak of activation also reduced the clonogenic capacity of CRC cells exposed to all treatments (Figure S5E,F).Altogether, these results suggest that the remaining cells after treatment seem to have triggered cytoprotective autophagy, and the rational inhibition of this mechanism strongly sensitized them to die.

| DISCUSSION
Although regimens using 5FU and OXA remain the firstline treatment for CRC, tumor growth recurrence remains the major challenge in clinics.Here, we evaluated acute and long-term responses of CRC cells to 5FU and OXA alone or combined.The co-treatment triggered many cellular responses, including cell cycle changes followed by apoptosis and senescence.Despite the expressive acute response, all treatments maintained a small subpopulation of cells presenting a viable phenotype, which resulted in a population regrowth resembling a clinical recurrence.These cells that survived acute treatments triggered a transitory increase in autophagy features, in higher levels to the cotreatment than single treatments.Notably, autophagy seemed to be the phenotype most likely linked to the survival of a subset of CRC cells since its inhibition sensitized tolerant cells and reduced cell regrowth.
After genotoxic stress, cells can undergo distinct cell fates due to differences in their genetic background or epigenetic changes, which characterizes the tumoral heterogeneity. 43In this study, we used two cell lines carrying different genetic status in driver genes associated with CRC prognosis and response to therapy.Some of these dissimilarities may explain the differential sensitivity and cell line-specific outcomes we observed.TP53 may be the most critical gene to define the differential sensitivity to apoptosis induced by 5FU and OXA. 44][49][50] Considering effector molecules, OXA triggers apoptosis by increasing the Bax/Bcl-2 ratio in a TP53-dependent manner. 51Complementarily, the resistance to 5FUinduced apoptosis is mediated by the maintenance of high rates of antiapoptotic versus proapoptotic proteins. 52ther drivers than TP53 may also contribute to the differential sensitivity and outcomes in response to 5FU and OXA.CRC cells mutated for BRAF, which is the case The rational modulation of autophagy sensitizes colorectal cancer cells to 5FU and OXA.HCT116 and HT29 cells were treated with 5FU and OXA for 48 h, followed by reseeding and growth in drug-free medium (DFM).(A, B) Kinetics of acridine orange (AO) staining in HCT116 and HT29 cells, respectively.Cells were stained at days 3, 5, 7, 10, and 15 after the end of treatments.Top graph: percentage of AO-positive cells.
Bottom graphs-intensity of AO red fluorescence.Boxes-relative area under the curve (AUC) to each condition.(C, E) Cumulative population doubling (CPD) of 48 h of treatment with 5FU and OXA, followed by growing in DFM, combined with two different autophagy inhibition regimens: left graphs-a pretreatment with 2 mM of 3MA for 1 h (3MA pre), 24 h before the treatment with 5FU and OXA; right graphs-or the rational inhibition with 2 mM of 3MA in the peak of autophagy induction, at days 3, 4, and 5 (3MA peak) after the end of treatment with 5FU and OXA.The CPDs were analyzed along 15 days after the end of treatment with 5FU and OXA.(D, F) Comparative analysis of the end CPDs at day 15 after the end of treatments for HCT116 and HT29, respectively.5FU, 5-fluorouracil; OXA, Oxaliplatin.
for HT29 cells, tend to be more resistant to apoptosis 53 due to the overexpression of the antiapoptotic protein MCL-1 54 and the oncogene YAP. 55In contrast to BRAF status, CRC cells mutated for KRAS, like HCT116, may be more sensitive to apoptosis to 5FU and OXA. 56,57inally, mutations in the APC tumor suppressor, the most frequent genetic alteration in sporadic CRC, can also lead to apoptosis resistance in CRC cells. 58,59In agreement with this, HT29 cells (APCmut) were less sensitive to apoptosis induced by 5FU and OXA than HCT116 cells (APCwt), while restoring wild-type APC can sensitize HT29 cells to apoptosis. 581][62][63] In HCT116 but not in HT29 cells, OXA increases TP53-p21 and p53-p14ARF after 48 h of treatment, which are classical pathways involved in senescence entry. 51,64Corroborating this, we found that only HCT116 cells (TP53wt) acquired a full senescent phenotype (i.e., increased cell size concomitant to increased SA-beta-gal activity and an arrest in cell proliferation) in response to the co-treatment, which could explain the most delayed regrowth for co-treated HCT116 cells among all conditions tested.The differential status of RAS between cell lines (HCT116 KRAS mut ; HT29 KRAS wt ) may also contribute to increased sensitivity to senescence in HCT116 cells. 65Interestingly, although the HT29 lineage has mutations in BRAF and APC, which could facilitate the entry into senescence after treatments, loss of TP53 may be dominant in the absence of the senescent phenotype in these cells.
Thus, the genetic profile of the HT29 lineage, which presents mutations for three genes associated with sensitivity to apoptosis in CRC (TP53, BRAF, and APC), may result in lower sensitivity to the treatments.Notwithstanding, the differential sensitivity and cellular outcomes in response to therapy must depend on integrating multiple genes.Regarding this, CRCs can be classified based on broader genetic characteristics as CIN (Chromosomal INstability), characterized by loss of the APC gene, aneuploidy, and chromosomal alterations, or MSI (Microsatellite Instability), which is a hypermutable phenotype resulting from loss of several DNA mismatch repair proteins. 66,67This classification may also help to explain the different outcomes observed in our cells since HT29 cells are classified as CIN.In contrast, HCT116 cells are classified as MSI, thus leading to defective DNA repair, the accumulation of mutations and increased sensitivity to DNA damage agents. 68,69ere, we observed that cells that survived the treatment triggered high levels of autophagy features throughout the week following treatment.1][72] However, these studies did not investigate (1) the dynamics of autophagy in tumor cells after drug exposition, mimicking the recovery period of patients, and (2) the most effective strategy of autophagy modulation to reduce cell resistance and regrowth.We demonstrated that after treatment withdrawal, there is an even more significant increase in autophagy in cells that survived the acute treatment.This dynamic corroborates previous results from our group, demonstrating a long-term increase in autophagy mainly after temozolomide removal in glioblastoma cells 73 and vincristine, 28 and doxorubicin in breast cancer. 74Considering that autophagy can play a cytoprotective role in cancer cells, we inhibited autophagy using two strategies to reduce the cells that survived treatments due to high autophagy levels: before treatments or during the peak of autophagy features.While the pretreatment with 3MA did not alter the long-term cell growth, the rational suppression of autophagy during its peak strongly reduced the CRC cells' growth and clonogenic capacity.
Therefore, our experimental design resembling the clinical regimen may be fundamental to bringing preclinical protocols closer to clinical regimens for decision-making in clinical trials.Indeed, Augustine and colleagues showed that in vivo response correlates well with the long-term (12 days or more) but not with acute in vitro data. 75Likewise, here we found that OXA had little to no chronic effect in CRC cells, which agrees with the observation that OXA had no significant activity as a single therapy in patients with advanced CRC. 6 However, unlike suggested by our results, most clinical trials use autophagy inhibitors concomitantly with chemotherapeutics, without considering the kinetics of autophagy induction after treatment. 42,76ome results obtained here have potential for the design of preclinical and clinical studies in CRC management.Most of the driver genes discussed here can modulate autophagy in CRC cells.Oncogenic mutations or the overexpression of wild-type BRAF or KRAS induce autophagy in CRC cells, [77][78][79][80] while TP53 deficiency is associated with autophagy suppression. 81,82owever, unlike the differential sensitivity to apoptosis and senescence, tolerant cells from both lineages showed similar levels and kinetics for autophagy after treatments, regardless of genetic differences.Even more importantly, both cell lines showed high and similar sensitivity to autophagy inhibition.Although this cannot be said for all cases of CRC, it suggests that rational modulation of autophagy after treatment with 5FU and OXA is a promising alternative to reduce tolerance and recurrence of CRC cells harboring different molecular backgrounds.
Another clinically relevant observation relies on the enrichment of senescent cells in the tumor population.
Initially, senescence induction may lead to an early stagnation of tumor growth due to the cell growth arrest.In the long term, however, senescent cells may contribute to increased aggressiveness by non-senescent tumor cells 83 and the suppression of the antitumor immune response through the secretion of molecules into the tumor microenvironment. 84,852][93] Thus, CRC patients with a molecular background susceptible to senescence, such as the HCT116 lineage, could benefit from the use of senolytic therapies after chemotherapy, a strategy that has already been tested in other tumor types. 94inally, the use of different drug concentrations in the cell lines due to the differential acute sensitivity is also translationally relevant since patients may have varied pharmacokinetics, culminating in varying levels of serum drug concentration even in response to similar doses of chemotherapy administered.Furthermore, different patients might be treated with different doses of chemotherapeutics, considering health status, age, comorbidities, or other interpersonal variables.Nevertheless, as raised, despite having different acute sensitivity, HCT116 and HT29 cells demonstrated similar levels of tolerance-associated phenotypes over the 2 weeks after treatment, including the minimum point of the CPD curve, clonogenicity, CSC enrichment, and levels or kinetics of autophagy features.This similarity was fundamental since we primarily aimed to evaluate mechanisms of tolerance triggered after 5FU and OXA.
In conclusion, our data suggest that CRC cells that survived the acute treatment with 5FU and OXA triggered high levels of autophagy, which can be rationally inhibited to increase the toxicity of the treatment.The clinicaltrials.gov database has more than 90 registered clinical trials using autophagy inhibitors in cancer, of which nine are in CRC.Therefore, these studies should consider the most appropriate time to inhibit autophagy to potentiate the treatment's cytotoxic effect in clinics.Based on our results, it is plausible to infer that the rational combination of 5FU + OXA with autophagy inhibitors considering the most promising order or time of administration must be the focus for future preclinical and clinical studies of CRC therapy.These strategies could lead to a better therapeutic response with a potential reduction in the risk of tolerance and recurrence of CRC.

F I G U R E 1
The co-treatment of 5FU and OXA transiently reduces CRC cells proliferation.CRC cells were treated with 5FU (HCT116: 10 µM; HT29: 20 µM), OXA (HCT116: 5 µM; HT29: 10 µM) or co-treatment with mentioned doses for 48 h and the remaining cells were kept in drug-free medium (DFM).(A) Forward scatter (FSC) versus side scatter (SSC) flow cytometry graphs after 48 h of treatment; numbers below the graphs: average ± standard deviation values of FSC and SSC in relation to control.(B) Percentage of cell number in relation to control after 48 h of treatment.Data are shown as percentage.(C, D) Cumulative population doubling (CPD) of HCT116 and HT29 cells, after 48 h of treatments, respectively.(E) Representative images of colonies from clonogenic assay.(F) Number of colonies at the end of 12 days of growth in relation to the control.Data are shown as percentage.(G) Colony area plot assessed using the Image Pro Plus software.Each marker represents a colony of at least 50 cells.(H) Percentage of CD133-and CD44-positive cells evaluated 7 days after the end of treatments.*p < 0.05, **p < 0.01, ***p < 0.001 in relation to control.# p < 0.05, ## p < 0.01, ### p < 0.001 in relation to 5FU.$$$ p < 0.001 in relation to OXA.CRC, colorectal cancer; 5FU, 5-fluorouracil; OXA, Oxaliplatin.

F
I G U R E 2 5FU and OXA affects cell cycle dynamics and trigger cell death and cellular senescence in CRC cells.HCT116 and HT29 cells were treated with 5FU and OXA for 48 h, followed by reseeding and growth in drug-free medium (DFM).(A, B) Cell cycle histograms and distribution after 48 h of treatment.PI, propidium iodide.(C, D) Annexin V-FITC/PI stain assay performed 3 days after the end of treatments.Left dot plots: representative plots of one experiment.Blue area-annexin-positive/PI-negative cells; red area-annexin-positive/ PI-positive cells.Right bar plots: percentage of cells in each quadrant (average ± standard deviation).(E, F) Dot plot graphs from flow cytometry of SA-β-gal activity versus forward scatter (FSC), 7 days after the end of treatments.Left dot plots: representative plots of one experiment.Blue area-C12-FDG-positive cells with low FSC; purple area-C12-FDG-positive cells with high FSC.Right bar plots: Values represent the percentage of cells in each quadrant (average ± standard deviation).*p < 0.05, **p < 0.01, ***p < 0.001 in relation to control.5FU, 5-fluorouracil; OXA, Oxaliplatin.