Aspirin exerts high anti-cancer activity in PIK3CA-mutant colon cancer cells

Evidence suggests that nonsteroidal anti-inflammatory drug aspirin (acetylsalicylic acid) may improve patient survival in PIK3CA-mutant colorectal carcinoma, but not in PIK3CA-wild-type carcinoma. However, whether aspirin directly influences the viability of PIK3CA-mutant colon cancer cells is poorly understood. We conducted in vitro experiments to test our hypothesis that the anti-proliferative activity of aspirin might be stronger for PIK3CA-mutant colon cancer cells than for PIK3CA-wild-type colon cancer cells. We measured the anti-proliferative effect of aspirin at physiologic concentrations in seven PIK3CA-mutant and six PIK3CA-wild-type human colon cancer cell lines. After exposure to aspirin, the apoptotic index and cell cycle phase of colon cancer cells were assessed. In addition, the effect of aspirin was examined in parental SW48 cells and SW48 cell clones with individual knock-in PIK3CA mutations of either c.3140A>G (p.H1047R) or c.1633G>A (p.E545K). Aspirin induced greater dose-dependent loss of cell viability in PIK3CA-mutant cells than in PIK3CA-wild-type cells after treatment for 48 and 72 hours. Aspirin treatment also led to higher proportions of apoptotic cells and G0/G1 phase arrest in PIK3CA-mutant cells than in PIK3CA-wild-type cells. Aspirin treatment of isogenic SW48 cells carrying a PIK3CA mutation, either c.3140A>G (p.H1047R) or c.1633G>A (p. E545K), resulted in a more significant loss of cell viability compared to wild-type controls. Our findings indicate that aspirin causes cell cycle arrest, induces apoptosis, and leads to loss of cell viability more profoundly in PIK3CA-mutated colon cancer cells than in PIK3CA-wild-type colon cancer cells. These findings support the use of aspirin to treat patients with PIK3CA-mutant colon cancer.


INTRODUCTION
Colorectal cancer is the second leading cause of cancer-related deaths in the U.S. [1], and the prognosis for individuals with advanced colorectal cancer remains poor [2,3].Accumulating evidence indicates that colorectal cancer is a heterogeneous group of diseases, and that responsiveness to treatment varies from patient to patient [4][5][6][7].This heterogeneity necessitates precision medicine approaches that take tumor molecular characteristics into account in order to predict an individual's response to a specific agent.
Aspirin (acetylsalicylic acid) is one of the most commonly used medications worldwide as an effective analgesic, antipyretic and cardiovascular prophylactic agent.Several observational studies and randomized controlled trials have shown that regular use of aspirin after a diagnosis of colorectal cancer is associated with a superior clinical outcome [8][9][10][11][12][13][14].Randomized trials of aspirin for prevention of cardiovascular and cerebrovascular events also suggest that regular aspirin use has the potential to reduce colorectal cancer incidence and mortality [15,16].Recently, the U.S. Preventive Services Task Force (USPSTF) recommended the use of aspirin for prevention of colorectal cancer in individuals at a 10% increased risk for cardiovascular disease compared with the general population [17], while cautioning against the potential harms associated with regular aspirin use including gastrointestinal pain and bleeding [18].Therefore, it is of particular interest to identify subgroups of individuals who are most likely to benefit from aspirinbased therapeutic strategies through the development of informative tumor biomarkers [14].
To test this hypothesis, we assessed the effect of aspirin treatment on cell cycle arrest and apoptosis using 13 colon cancer cell lines with known PIK3CA mutation status and isogenic colon cancer cells with knock-in PIK3CA-activating mutations.Additionally, our study identified cell models that may serve as ideal in vitro models to further clarify the molecular mechanisms of aspirin's anti-tumor effects in PIK3CA-mutant colorectal cancer.

PIK3CA-mutant colon cancer cells
To test the hypothesis that aspirin's anti-cancer effects might differ by PIK3CA mutation status, we investigated the anti-proliferative activity of aspirin at physiologically attainable concentrations in seven PIK3CA-mutant and six PIK3CA-wild-type colon cancer cell lines (Table 1) using the 3-(4,5-dimethylthiazol-2yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2Htetrazolium inner salt (MTS) assay.Initially, we examined the effect of a variety of aspirin doses (0, 0.5, 1, 2, 4, 6, 8, 10, and 12 mM) on cell proliferation after exposure to aspirin for 12 to 96 hours in two PIK3CA-mutant (HCT15 and HCT116) and PIK3CA-wild-type (SW480 and SW620) colon cancer cell lines.We found that the most prominent effects of aspirin on proliferation occurred between 48 and 72 hours of exposure.Therefore, all subsequent experiments were performed after treatment with aspirin for 48 and 72 hours.The dose-response curves of PIK3CA-mutant and PIK3CA-wild-type colon cancer cell lines were generated (Figure 1) and the half maximal inhibitory concentration (IC 50 ) values of each PIK3CA subgroup were compared (Supplementary Figure 1A).
PIK3CA-wild-type cells after treatment for 48 and 72 hours.Aspirin treatment also led to higher proportions of apoptotic cells and G0/G1 phase arrest in PIK3CA-mutant cells than in PIK3CA-wild-type cells.Aspirin treatment of isogenic SW48 cells carrying a PIK3CA mutation, either c.3140A>G (p.H1047R) or c.1633G>A (p.E545K), resulted in a more significant loss of cell viability compared to wild-type controls.Our findings indicate that aspirin causes cell cycle arrest, induces apoptosis, and leads to loss of cell viability more profoundly in PIK3CA-mutated colon cancer cells than in PIK3CAwild-type colon cancer cells.These findings support the use of aspirin to treat patients with PIK3CA-mutant colon cancer.
PIK3CA-mutant colon cancer cell lines were considerably more sensitive to the anti-proliferative effects of aspirin than PIK3CA-wild-type cells (Figure 1A).Aspirin did not show differential effects according to BRAF or KRAS mutation status (Figure 1B).Next, in order to validate that PIK3CA-mutant colon cancer cell lines were specifically sensitized to aspirin but not to common DNA damaging chemotherapeutic agents, we investigated the IC 50 values of 5-fluorouracil and cisplatin in our thirteen colon cancer lines from a compound sensitivity database of the Cancer Cell Line Encyclopedia (CCLE) [35].However, no significant differences were found between PIK3CAmutant and PIK3CA-wild-type colon cancer cell lines (Supplementary Figure 1B).

Aspirin induces apoptosis and G 0 /G 1 cell cycle arrest in PIK3CA-mutant cells
We evaluated the effects of aspirin on induction of apoptosis and cell cycle distribution in three PIK3CAmutant cell lines (HCT15, HCT116, and SW948) that were particularly sensitive to aspirin exposure in the MTS assay and three PIK3CA-wild-type colon cancer cell lines (COLO205, SW480, and SW620) that were resistant to aspirin exposure.Cells were treated with 0, 1, 2.5, and 5 mM aspirin for 48 and 72 hours after plating.Aspirin treatment led to a more profound induction of apoptosis in colon cancer cells harboring a PIK3CA mutation than in those with wild-type PIK3CA (Figure 2).We next assessed the percentage of cells in each phase of the cell cycle within these PIK3CA mutant and wild-type colon cancer cells (Figure 3).All PIK3CA-mutant cell lines underwent a dose-dependent increase in G 0 /G 1 cell cycle arrest after aspirin exposure for 48 and 72 hours.This effect was not observed in PIK3CA-wild-type colon cancer cell lines.

Mutations in PIK3CA oncogene sensitize colon cancer cells to aspirin
The PIK3CA c.3140A>G (p.H1047R) and c.1633G>A (p.E545K) somatic mutations are commonly found in colorectal carcinoma, and were present in the colon cancer cells (HCT15 and HCT116, respectively) included in the published studies [36,37].In order to further confirm our findings and ensure that the observed PIK3CA-mutant-specific effect was not confounded by the genomic backgrounds of the selected cell lines, we investigated whether activation of PI3K via knock-in of the c.3140A>G (p.H1047R) or c.1633G>A (p.E545K) mutation could sensitize a PIK3CA wild-type colon cancer cell line to aspirin treatment.Two isogenic cell lines were derived from parental SW48 cells, each of which carried constitutively active mutant alleles with mutations at either c.3140A>G [SW48 (PIK3CA c.3140A>G/+)] or c.1633G>A [SW48 (PIK3CA c.1633G>A/+)].The IC 50 values of each group showed that aspirin treatment of isogenic SW48 cells carrying either PIK3CA mutation resulted in a statistically significant loss of cell viability of up to 47% relative to parental SW48 cells (P = 0.031) (Figure 4).

DISCUSSION
We conducted this study to provide in vitro experimental evidence supporting the recent molecular pathological epidemiology studies that suggest tumor PIK3CA mutation status as a biomarker to predict benefits from aspirin therapy for colorectal cancer [9,12,13,21,38].In this study, we have shown that physiologically attainable concentrations of aspirin can exert stronger anti-cancer effects on PIK3CA-mutant colon cancer cells relative to PIK3CA-wild-type colon cancer cells.
Our results are consistent with a recent report by Zumwalt et al [39] that clarified the relationship of PIK3CA mutations in colorectal cancer cells and aspirin-induced chemoprevention in vitro and in vivo.In that study, eight colon cancer cell lines and engineered HCT116 cells with PIK3CA kinase domain mutant allele knockout were used to assess the effects of aspirin on cell proliferation and cell-cycle distribution.Our current study further investigated 13 commonly used colon cancer cell lines and two isogenic cell lines with heterozygous knock-in of either of PIK3CA mutations c.3140A>G (p.H1047R) and c.1633G>A (p.E545K), both of which are commonly observed in colorectal carcinoma.In addition, aspirin treatment of breast cancer cells carrying mutations in PIK3CA at either exon 9 (c.1633G>A) or exon 20 (c.3140A>G) also resulted in a significant decrease of cell viability [40].
Although multiple lines of evidence indicate that aspirin use after colorectal cancer diagnosis is beneficial, the toxicities of therapy, particularly in older patients, have limited its use in routine clinical practice [14,19,41].In 2016, the USPSTF gave a "B" recommendation (high certainty that the net benefit is moderate or moderate certainty that the net benefit is moderate or substantial) for routine aspirin use for colorectal cancer prophylaxis in U.S. adults between the  ages of 50 and 59 with a greater than 10% 10-year risk of cardiovascular events [17].The USPSTF recommendations also highlight the need to clarify the mechanisms by which aspirin prevents development of colorectal cancer [18].There are currently several clinical trials underway (including ADD-ASPIRIN [42], ASCOLT [43], and ASPIRED [44]) that are attempting to clarify the association between aspirin use and survival after diagnosis of colorectal cancer, as well as distinguishing the subgroups of individuals for whom the benefits outweigh the harm.Thus, it is essential to confirm the efficacy of biomarkers to predict benefits from aspirin therapy through multifaceted approaches, including clinical trials and in vitro or in vivo experimental models [14].The present study serves this purpose by improving our understanding of the use of tumor PIK3CA mutation status for prediction of patient response to aspirin therapy.
PIK3CA mutations (or PIK3CA amplifications [45]) may result in constitutive activation of PI3K and the downstream AKT pathway, enhancing PTGS2 activity and prostaglandin E 2 synthesis and leading to inhibition of apoptosis in colorectal cancer cells [11].Recent in vitro and in vivo studies have suggested that aspirin may suppress cancer cell growth and induce apoptosis through activation of protein kinase A (PRKA, also referred to as AMPK), inhibition of MTOR downstream signaling, and inhibition of PI3K-induced prostaglandin E 2 synthesis [46][47][48].The present study was designed to detect the potency of aspirin therapy in colon cancer cell lines according to major somatic driver mutations such as PIK3CA, BRAF, and KRAS mutations.The results demonstrate that aspirin may selectively cause G 0 /G 1 cell cycle arrest, induce apoptosis, and inhibit cell growth in human colon cancer cells with PIK3CA mutations in vitro.Our findings in the apoptotic and cell cycle assays are consistent with the data on the previous in vitro studies [39,40,47,48].The reduced viability of PIK3CA-mutant colon cancer cells was also confirmed in colon cancer cells containing the heterozygous PIK3CA knock-in mutations c.3140A>G (p.H1047R) and c.1633G>A (p.E545K).Previous retrospective analyses investigating aspirin use in colon cancers according to PIK3CA mutation status did not specify which PIK3CA mutations were included [10,12].The present study is the first to show that the isogenic conversion of a PIK3CA-wild-type colon cancer cell into a PIK3CA-mutant cell line is sufficient to promote sensitization to aspirin.PIK3CA c.3140A>G (p.H1047R) and c.1633G>A (p.E545K) mutations are common in colorectal carcinomas.Heterozygous knock-in c.3140A>G and c.1633G>A mutations in the PIK3CA gene were found to activate multiple oncogenic pathways and promote cell growth and invasion in vitro for human breast cancer and colon cancer cell lines [37,49].Notably, breast cancer cells carrying a c.3140A>G (p.H1047R) mutation at exon 20 were less sensitive to aspirin than those carrying a c.1633G>A (p.E545K) mutation at exon 9 [40].This phenomenon may be attributed to differences in the aberrant activation of signaling pathways.However, few, if any, differences in phenotype were observed according to PIK3CA mutation variants.
The relationship between somatic or germline genetic variants and the chemopreventive use of aspirin is not limited to patients with PIK3CA-mutant colorectal cancer.Lizaka and colleagues utilized a genome-wide complementary DNA microarray containing 23,040 genes to analyze the time-dependent alteration of gene expression in response to two nonsteroidal anti- inflammatory drugs (NSAIDs), sulindac and aspirin, in SNU-C4, SW480, and SW948 colon cancer cell lines [50].Interestingly, they found that PIK3C2A, which belongs to the PIK3 family, was down-regulated only in aspirinsensitive cells.We previously performed a genome-wide analysis of the interactions between single nucleotide polymorphisms (SNPs) and aspirin use in relation to colon cancer risk, which identified two SNPs on chromosomes 12 and 15 that differentially correlate with aspirin and NSAID effects on colon cancer prevention [51].One, on chromosome 12, is located upstream from PIK3C2G, which belongs the PIK3 family.Combined, these findings further support the use of PIK3CA mutation status as a biomarker for precision aspirin chemoprevention and adjuvant therapy strategies for colon cancer.
Our current study has limitations.As our colon cancer cell panel did not include all possible activating alterations in PIK3CA, we were unable to elucidate the potency of aspirin treatment on colon cancer cells with PIK3CA alterations other than the ones examined.In addition, our current cancer cell panel was limited to immortalized colon cancer cell lines due to the lack of reasonable PIK3CA-mutant and PIK3CA-wild-type rectal cancer cell line pairs in the American Type Culture Collection (ATCC).Taking into account the potential heterogeneity of tumor behavior between colon and rectal carcinomas, anti-cancer effects of aspirin according to PIK3CA mutation status in rectal cancer cells and human primary cells should be examined in future studies.In the present study, the aspirin concentrations utilized in apoptosis and cell cycle analyses are relatively high (as high as 5 mM) compared to typical plasma concentrations of salicylate (0.19 to 0.63 mM), a major metabolite derived from aspirin.Thus, the doses may not be attainable through administration of standard doses (81 or 325 mg) of aspirin.However, the dose-dependent nature of the response to aspirin treatment suggests that some effects may be experienced in vivo even at low doses.Furthermore, Turturro et al. reported that the serum levels of aspirin could safely reach concentrations of up to 10 mM [40], indicating that the 5 mM aspirin dose is physiologically achievable in humans.Our in vitro models did not take into account the critical role of the tumor stromal microenvironment.Accumulating evidence indicates that regular aspirin use may synergize with other immunomodulatory pathways including the immune checkpoint blockade in colorectal cancer [52][53][54][55][56].Further investigations are warranted to analyze interactions of multiple cell types that exist in the tumor microenvironment, and to examine the dual blockade of the PTGS2 and immunosuppressive pathways.Another drawback is that our current study did not utilize functional assays to investigate potential molecular mechanisms through which aspirin may exert stronger anti-cancer effects on PIK3CA-mutant colon cancer cells than PIK3CA-wild-type cells.Nonetheless, our study hypothesis was based on human clinical studies [8-13, 57, 58].It is important to perform functional studies to explore potential mechanisms underlying our findings in the future.We point out an important synergism of experimental research and molecular pathological epidemiology (MPE) research [59][60][61] as replicating biological complexities of human tumors in experimental models is a substantial challenge.In addition, it is of particular interest to investigate modifying effects of other endogenous and exogenous factors (such as diet, lifestyle, microbiota, germline genetics, and immunity) on medications (such as aspirin) in relation to tumor molecular characteristics [4,6,7] in integrative pharmaco-MPE research [62,63] in the future.
In summary, our findings provide valuable in vitro data to further support the use of aspirin in patients with PIK3CA-mutant colorectal cancer.Further investigations into the precise molecular mechanisms associated with aspirin treatment of cancers with mutant PIK3CA are warranted.

Cell proliferation MTS assay
Cells were seeded in 96-well microplates at a density of 4x10 4 cells/ml and cultured overnight; they were treated with a range of aspirin (Sigma-Aldrich, St. Louis, MO, USA) doses (0, 0.5, 1, 2, 4, 6, 8, 10 and 12 mM) in dimethyl sulfoxide (DMSO).These doses are representative of physiologically attainable plasma concentrations as previously reported [40].Control cells received DMSO only.Cell proliferation was assessed after 48 hours and 72 hours from initial drug exposure using the MTS assay (Promega, Madison, WI, USA) according to the manufacturer's instructions.Cell proliferation is directly proportional to the absorbance at 490 nm by a formazan product that is bio-reduced from MTS in living cells.All experiments were carried out in hexaplicate and were repeated at least three times independently.

Figure 2 :
Figure 2: Aspirin treatment results in more apoptosis in PIK3CA-mutant human colon cancer cell lines.PIK3CA-mutant human colon cancer cell lines (HCT15, HCT116, and SW948) and PIK3CA-wild-type colorectal cancer cell lines (COLO205, SW480, and SW620) were incubated with aspirin (1, 2.5, and 5 mM) or DMSO for 48 and 72 hours.Apoptosis analysis was performed by flow cytometry with Annexin V-FITC/PI staining.The percentage of Annexin V-FITC/PI positive cells after aspirin treatment at each dose and time point for all cell lines were normalized by DMSO group and compared according PIK3CA status.Student's t-test was performed to determine significance.The data shown represent mean ± standard deviation of three replicates.* P value < 0.05.** P value < 0.01.DMSO, dimethyl sulfoxide; MUT, mutation; WT, wild-type.

Figure 3 :
Figure 3: Aspirin leads to a higher proportion of cells in G 0 /G 1 phase arrest in PIK3CA-mutant human colon cancer cell lines.PIK3CA-mutant human colon cancer cell lines (HCT116, HCT15, and SW948) and PIK3CA-wild-type colorectal cancer cell lines (COLO205, SW620, and SW480) were incubated with aspirin (0, 1, 2.5, and 5 mM) for 48 and 72 hours.Cell cycle analysis was performed by flow cytometry with PI staining.The percentage of G 0 /G 1 phase cells of each aspirin treatment point was normalized by DMSO group and was compared according PIK3CA status.Student's t-test was performed to determine significance.The data shown represent mean ± standard deviation of three replicates.* P value < 0.05.** P value < 0.01.DMSO, dimethyl sulfoxide; MUT, mutation; WT, wild-type.