Missense Mutant p53 Transactivates Wnt/β-Catenin Signaling in Neighboring p53-Destabilized Cells through the COX-2/PGE2 Pathway

TP53 is one of the most frequently mutated genes in a wide range of cancers, including colorectal cancer (1, 2), and the TP53 mutations were found in about 80% of colorectal cancer cases that are associated with metastasis (3). These results suggest that TP53 mutation plays an important role in the late stage of cancer progression. The target molecules of p53 regulate a variety of biological processes, including cell-cycle arrest, DNA repair, and senescence, which may prevent tumor progression as a tumor suppressor (4). On the other hand, approximately 75% of TP53 mutations in cancer are missense mutations in hot spots, resulting in the expression of mutant p53 with single amino acid substitutions, and such mutant p53 is thought to acquire a novel oncogenic function (5, 6).

Such gain of function of mutant p53 has been demonstrated in mouse genetic studies (7, 8). Trp53-null mutant mice spontaneously develop soft-tissue sarcomas and lymphomas, whereas mice expressing missense-type mutant p53 develop adenocarcinomas in the intestine and lung. It has also been reported that mutant p53 induces stem cell properties in breast and colorectal cancer cells (9, 10). In addition, we showed that the Trp53^R270H missense mutation (corresponding to human TP53^R273H) causes submucosal invasion of intestinal tumors in Apc^Δ716-mutant mice through activation of the inflammatory and Wnt signaling pathways (11). We also constructed a malignant intestinal tumor model by introducing Apc^Δ716, Kras^G12D, Tgfbr2^−/−, and Trp53^R270H mutations to generate quadruple-mutant mouse model (12) and showed that missense-type mutant p53 promotes the survival and clonal expansion ability of single cell–dissociated cancer cells, which may contribute to colonization of disseminated cells in distant organs (13). It has also been reported that in vivo growth of Trp53^R248Q mouse tumors depends on sustained expression of mutant p53 (14). Similar results also showed that ablation of mutant p53^R248Q suppressed growth and invasion of mutant p53–driven tumors (15). These results suggest a possibility of gain-of-function mutant p53 as a therapeutic target molecule.

Although the genetic evidence for gain of function of mutant p53 has been accumulated, the underlying mechanisms of mutant p53–induced tumor promotion have not been fully understood yet. It has been shown that missense mutant p53 dramatically alters the transcriptome in tumor cells through modification of chromatin regulation by upregulation of chromatin regulatory genes or interaction with the SWI/SNF complex (16, 17). However, the precise molecular mechanisms of how mutant p53 promotes malignant phenotypes are still unclear. Moreover, there is intratumor heterogeneity in stabilization of mutant p53 (18), and it has not yet been understood whether there is interaction in oncogenic function between p53-stabilized and -destabilized cells.

In this study, we used mouse intestinal tumor–derived organoids that carried Apc, Kras, and Tgfbr2 mutations together with Trp53^R270H or Trp53^−/− mutation to identify a tumor-promoting mechanism of mutant p53 and interaction between mutant p53-stabilized and -destabilized cells. Through the organoid experiments, we found that mutant p53^R270H–expressing cells transactivate Wnt/β-catenin signaling of co-cultured Trp53^−/− cancer cells through induction of the COX-2/prostaglandin E2 (PGE2) pathway. Accordingly, these results suggest that the COX-2/PGE2 pathway induced by stabilized mutant p53 activates Wnt/β-catenin signaling in neighboring p53-destabilized cells, which may contribute to malignant progression of tumor tissues.

An AKTP^R270H organoid was established from intestinal tumors of quadruple mutant Apc^Δ716 (A), Kras^G12D (K) Tgfbr2^−/− (T), and Trp53^R270H (P) mice (12). AKTP^Null organoids were generated by the disruption of wild-type (WT) Trp53 in AKT triple-mutant cells using the p53 CRISPR/Cas9 plasmid (13). Trp53-disrupted cells were selected with 10 µmol/L nutlin-3 (Cayman Chemical). These organoid cells were cultured on two-dimensional dishes with Advanced DMEM/F-12 medium (Gibco) supplemented with 10% FBS, 5 µmol/L anaplastic lymphoma kinase inhibitor (A83-01; Tocris Bioscience), 5 µmol/L GSK inhibitor (CHIR99021; Tocris Bioscience), and 10 µmol/L ROCK inhibitor (Y27632; Wako).

For suppression experiment of Trp53^R270H in AKTP^R270H cells, siRNA for Trp53 or control-siRNA (Silencer select predesigned siRNA, Ambion, Thermo Fisher Scientific) and siRNA for Trp53 (ON-TARGETplus Mouse Trp53 siRNA-SMART pool, Dharmacon, Horizon) were transfected by Lipofectamine RNAiMAX (Thermo Fisher Scientific). Knockdown efficiency was examined by immunoblotting. Wnt signaling activation was examined after 9 hours of siRNA transfection.

For co-culture experiments, 2 × 10² AKTP^Null and AKTP^R270H cells were mixed and co-cultured, or 4 × 10² AKTP^Null or AKTP^R270H cells were cultured (monoculture) on Ultra-Low Attachment culture plates (Corning). After 72 hours, the size of organoids was measured using ImageJ (NIH) on the photographs. All cell culture experiments were repeated three times unless otherwise noted.

HCT116 (#CCL-247, RRID: CVCL_0291), SW480 (#CCL-228, RRID: CVCL_0546), SW620 (#CCL-227, RRID: CVCL_0547), and Caco-2 (#HTB-37, RRID: CVCL_0025) were obtained from ATCC. LOVO (RCB1639, RRID: CVCL_0399) and COLO320 (RCB1193, RRID: CVCL_1989) were obtained from RIKEN BRC. Cells were maintained in high-glucose DMEM (Gibco) supplemented with 100 U/mL penicillin, 100 μg/mL streptomycin (Lonza), and 10% FBS (HyClone). Cells were maintained in the culture for maximum of 15 passages. Cells were monitored monthly for Mycoplasma contamination, and cell lines were authenticated using short tandem repeat analysis (Labcorp).

For measurement of Wnt signaling activity, the cells were transfected with Super 8× TOPFlash (RRID: Addgene_12456) or Super 8× FOPFlash (RRID: Addgene_12457; Addgene) using Lipofectamine LTX (Thermo Fisher Scientific). After 24 hours of transfection, the luciferase activity was measured using a Dual-Luciferase Reporter Assay system (Promega). For measurement of NF-κB activation, the cells were transfected with a pNL3.2.NF-κB-RE vector (Promega) using Lipofectamine LTX (Thermo Fisher Scientific). The luciferase activity was measured using a Nano-Glo Dual-Luciferase Reporter Assay System (Promega).

For measurement of Wnt signaling activity in co-culture experiment, AKTP^Null cells were transfected with Super 8× TOPFlash (RRID: Addgene_12456) or Super 8× FOPFlash (RRID: Addgene_12457). At 24 hours after transfection, the culture medium was changed, and then AKTP^Null cells were co-cultured with AKTP^R270H cells. After 24 hours of co-culture, the luciferase activity was measured.

For COX-2 inhibition, cells were cultured with 5 µmol/L celecoxib (Selleck Chemicals). For EP2 and EP4 inhibition, cells were cultured with 100 nmol/L PF-04418948 (Selleck Chemicals) or 10 µmol/L RQ-15986/CJ-042794 (AskAt Inc.), respectively (19). For PGE2 stimulation, 1 µmol/L PGE2 (Cayman Chemical) was added at 9 hours after transfection of the reporter vector.

Venus cDNA was replaced with luciferase cDNA of the TOPFlash vector (Upstate) to generate TOP-Venus and transfected to AKTP^Null cells using Lipofectamine LTX (Thermo Fisher Scientific). Transfectant TOP-Venus AKTP^Null cells were selected with 200 to 500 ng/mL puromycin. A total of 2 × 10² TOP-Venus AKTP^Null and AKTP^R270H cells were mixed and co-cultured, or 4 × 10² TOP-Venus AKTP^Null cells were monocultured on Ultra-Low Attachment culture plates (Corning). After 72 hours of culture, spheroids were picked up and embedded in Matrigel (Corning), and after another 72 hours of culture, spheroids were immunostained.

For 5-ethynyl-2'-deoxyuridine (EdU) labeling, we used 2 × 10² AKTP^Null cells and Venus-labeled AKTP^R270H cells that were previously described (20). These cells were co-cultured on Ultra-Low Attachment culture plates (Corning). After 48 hours of culture, cells were treated with 10 mmol/L EdU for 90 minutes. After fixation with 4% paraformaldehyde, organoids were embedded in Matrigel, and EdU-labeled AKTP^Null cells were detected using Click-iT EdU cell proliferation assays (Thermo Fisher Scientific) and a confocal microscope (TCS SP8, Leica Microsystems). The numbers of EdU-labeled cells were counted at three different horizontal levels of each organoid, and the mean EdU labeling index was calculated.

Short hairpin RNA (shRNA) lentiviral vectors for Ctnnb1 in pLKO.1-puro (MISSION shRNA, Sigma-Aldrich) were co-transfected to HEK293T cells with pCMV-VSV-G (Addgene, #8485, RRID: Addgene_8454) and pCMV-dR8.2 dvpr (Addgene, #8455, RRID: Addgene_8455) using PEI MAX (Polysciences) for packaging. Lentiviruses were infected to AKTP cells using polybrene (Sigma-Aldrich), and shRNA-expressing cells were selected with puromycin. Cell growth of shRNA lentivirus–infected cells was examined using CellTiter-Glo 2.0 Cell Viability Assay (Promega).

C57BL/6 mice (male; 6 weeks of age) were purchased (The Jackson Laboratory). AKTP^R270H or AKTP^Null organoids were treated with trypsin, and 3 to 5 × 10⁵ cells were injected into the mouse spleen. Liver metastatic tumors were examined at 4 weeks after the injection (n = 3) as follows. All animal experiments were carried out according to the protocol approved by the Committee on Animal Experimentation of Kanazawa University, Japan.

Whole-liver tissues were collected from C57BL/6 mice. Human colorectal cancer samples were obtained from patients who underwent surgical resection at Ishikawa Prefectural Central Hospital, Japan. The experiments using human tissues were approved by the Human Genome/Gene Analysis Research Ethics Committee of Kanazawa University at #2021-002-603, and written informed consent was obtained from the patients. The tissues were fixated in 4% paraformaldehyde, paraffin-embedded, and sectioned at 4 μm thickness. For IHC, antibodies against E-cadherin (R&D Systems, Cat. #AF748, RRID: AB_355568), mouse p53 (Leica Biosystems, CM5, Cat. #NCL-p53-CM5p, RRID: AB_563933; Cell Signaling Technology, 1C12, Cat. #2524, RRID: AB_331743), human p53 (Santa Cruz Biotechnology, DO7, Cat. #47698, RRID: AB_628083), β-catenin (Cell Signaling Technology, D10A8, #8480, RRID: AB_11127855), and COX-2 (Cayman Chemicals, polyclonal antibody, aa584-598, RRID: AB_10079419) were used as primary antibodies. Staining signals were visualized using the VECTASTAIN Elite Kit (Vector Laboratories). For fluorescence IHC, Alexa Fluor 594– or Alexa Fluor 488–conjugated antibodies (Molecular Probes, Fluor 594, #A21207, RRID: AB_141637; Fluor 488, #A21206, RRID: AB_2535792) were used as the second antibody. The immunostained specimens were examined using a confocal microscope (Leica TCS SP8, Leica Microsystems). The number of p53-stabilized and β-catenin nuclear-accumulated cells were counted in three to four independent microscopic fields, and the mean number per tumor was calculated.

The organoids were immunostained using antibodies against mouse p53 (Leica Biosystems, CM5, Cat. #NCL-p53-CM5p, RRID: AB_563933; Cell Signaling Technology, 1C12, Cat. #2524, RRID: AB_331743) and COX-2 (Cayman Chemicals, polyclonal antibody, aa584-598, RRID: AB_10079419) and Alexa Fluor 594– or Alexa Fluor 488–conjugated antibodies (Molecular Probes, Fluor 594, RRID: AB_141637; Fluor 488, RRID: AB_2535792) for the secondary antibody. The immunostained organoids were examined using a confocal microscope (Leica TCS SP8, Leica Microsystems).

Organoid cells were lysed in lysis buffer, and protein samples were separated in 10% SDS-PAGE. Antibodies against mouse p53 (Cell Signaling Technology, 1C12, Cat. #2524, RRID: AB_331743) and COX-2 (Cayman Chemicals, polyclonal antibody, aa584-598, RRID: AB_10079419) were used. Antibodies for GAPDH (FUJIFILM Wako, Cat. #016-25523, RRID: AB_2814991) and β-actin (Wako, 6D1, #010-27841, RRID: AB_2858279) were used as the internal control. The enhanced chemiluminescence detection system (GE Healthcare) was used to detect the signals.

Trp53 genotyping in AKTP^R270H cell lines was performed using a TaqMan SNP Genotyping assay (Thermo Fisher Scientific). The custom TaqMan probes for WT Trp53 (VIC dye) and mutant Trp53^R270H (FAM dye) were designed in exon 8 including the R270 codon of mouse Trp53. Sequences of TaqMan probes are as follows:

Forward primer: CCG​GAT​AGT​GGG​AAC​CTT​CTG

Reverse primer: TCT​TCT​GTA​CGG​CGG​TCT​CT

The TaqMan probe to detect WT Trp53 (VIC dye) is CTT​TGA​GGT​TCGTGT​TTG​T, and the TaqMan probe to detect mutant Trp53^R270H (FAM dye) is TTT​GAG​GTT​CATGT​TTG​T. The PCR results were analyzed using Allele discrimination in Agilent AriaMx (Agilent Technologies).

For RNA sequencing (RNA-seq) data of AKTP^R270H and AKTP^Null organoids and AKTFP^R270H/LOH and AKTFP^R270H/+ organoids, we used public datasets of the DNA Data Bank of Japan (accessions #PRJDB5631 and #PRJDB15570).

The data were analyzed using two-sided unpaired t tests unless otherwise mentioned and presented as the means ± SD. Statistical analyses for Fig. 1A, B, and G were performed using a one-way ANOVA test. The Fisher exact test was used for Fig. 1G. A value of P < 0.05 was considered statistically significant. GraphPad Prism 9 (GraphPad) was used for statistical analyses. All data were reproduced with at least two independent experiments and at least three biological replicates.

The data generated in this study are available upon request from the corresponding authors.

We previously established an intestinal tumor–derived organoid, AKTP, in which Apc^Δ716, Kras^G12D, Tgfbr2^−/−, and Trp53^R270H mutations were introduced (12). To investigate the role of oncogenic function of mutant p53^R270H, we further constructed Trp53^−/− AKTP organoid by disrupting Trp53 using the CRISPR/Cas9 (13). Here, we named AKTP organoid lines with Trp53^R270H or Trp53^−/− mutations as AKTP^R270H or AKTP^Null, respectively. We used three independently established AKTP^R270H lines in this study, and genomic PCR analysis revealed that all lines had two copies of the Trp53^R270H mutant alleles, indicating the loss of WT Trp53 by copy-number neutral loss of heterozygosity (LOH; Supplementary Fig. S1).

We have previously shown that the inflammatory pathway and Wnt/β-catenin signaling are activated in the Apc^Δ716Trp53^R270H double-mutant intestinal tumor–derived organoids compared with that in the simple Apc^Δ716-mutant tumor organoids (11). Therefore, we examined the activation levels of NF-κB in the inflammatory pathway and Wnt/β-catenin signaling in the AKTP cells using reporter assays. We found that NF-κB activation was at the similar levels between AKTP^R270H and AKTP^Null organoid cells (Fig. 1A). In contrast, the levels of Wnt/β-catenin signaling were significantly higher in all AKTP^R270H compared with AKTP^Null organoid lines (Fig. 1B). Moreover, transfection of siRNA against Trp53 mRNA into AKTP^R270H cells significantly suppressed Wnt signaling activation with a decrease in p53 expression levels (Fig. 1C). These results indicate that mutant p53^R270H promotes Wnt/β-catenin signaling activation.

We confirmed p53^R270H-induced Wnt signaling activation by other methods. First, we performed an Ingenuity Pathway Analysis using RNA-seq data of AKTP^R270H and AKTP^Null organoids (13) and found that the Wnt signaling pathways (CTNNB1 and TCF7L2 pathways) were significantly activated in AKTP^R270H (Supplementary Fig. S2). In addition, immunostaining of cultured AKTP^R270H organoids and histologic sections of liver metastasis of AKTP^R270H organoid cells showed β-catenin stabilization and nuclear accumulation, which were rarely found in AKTP^Null organoid cultures and metastatic tissues (Fig. 1D and E).

We further examined the relationships between mutant p53 and Wnt/β-catenin signaling using TP53 WT and TP53-mutant human colorectal cancer cell lines (21). As expected, stabilized p53 was detected in the missense-type TP53-mutant SW480, SW620, and COLO320 cells but not in TP53 WT HCT116 and LOVO cells (Fig. 1F). Notably, Wnt/β-catenin signaling is significantly higher in TP53-mutant cells than in TP53 WT cells (Fig. 1G). However, the Wnt/β-catenin signaling is also high in the TP53-null mutant Caco-2 cells (22). We further examined the primary colorectal cancer tissues by IHC using tissue microarray. In the p53 staining–positive colorectal cancer group, β-catenin nuclear accumulation was found in 39% of tissues, whereas the ratio was significantly decreased to 17% in p53-negative colorectal cancer tissues (Fig. 1H and I). These results collectively suggest that the Wnt signaling activation in human colorectal cancer may involve, though not exclusively, the presence of missense-type mutant p53, similar to that found in mouse intestinal tumors.

It has been reported that there is heterogeneity in p53 stabilization among TP53-mutant cells (18). We also found the heterogeneity in p53 stabilization in both human colorectal cancer and mouse AKTP^R270H liver metastasis by IHC (Fig. 2A). In p53-positive human colorectal cancer, p53 stabilization was found in approximately 80% and 72% of tumor cells in the primary and liver metastases, respectively (Fig. 2A and B). Similarly, in AKTP^R270H metastatic tumors, p53 stabilization was found in approximately 50% to 66% of two different cell lines in average. Specifically, not all cancer cells with p53 mutations have stabilized p53 protein, and there is intratumor heterogeneity in p53 stabilization. As stabilization of mutant p53 is thought to be a prerequisite for the oncogenic function, these results suggest that there is intratumor heterogeneity in the oncogenic function of mutant p53 in the p53-mutant tumors.

We therefore examined whether the stabilization of p53 and β-catenin occurs simultaneously in the same cells or not by double-fluorescence immunostaining. In the AKTP^R270H metastatic tumors, we found double-positive cells for β-catenin and p53 as expected; however, we also found certain populations of single-positive cells for β-catenin or p53 [Fig. 2C (top) and D (left)]. Similar immunostaining patterns, namely, double-positive and single-positive cells for p53 and β-catenin were also found in human colorectal cancer tissues [Fig. 2C (bottom) and D (center)]. In contrast, nuclear accumulation of β-catenin was rarely detected in AKTP^Null metastatic tumors [Fig. 2D (right)]. Taken together, these results suggest that mutant p53 activates Wnt/β-catenin signaling not only in the p53-stabilized cells but also in the p53-destabilized cells possibly through a non–cell autonomous mechanism.

To investigate a possible mechanism for Wnt signaling activation in cancer cells in which the mutant p53 protein is destabilized and undetectable, we used AKTP^Null as a model for these p53-destabilized cells in the p53-mutant tumors. After co-culture with AKTP^R270H cells, we examined the Wnt signaling activity in AKTP^Null cells using TOPFlash reporter assay (Fig. 3A). AKTP^Null cells (AKTP^Null-bottom cells in Fig. 3B) carry an Apc mutation; thus, reporter assay showed basal activation of Wnt signaling (Fig. 3B). Importantly, the TOPFlash level in AKTP^Null-bottom cells was significantly increased when co-cultured with AKTP^R270H-top cells compared with that in the co-culture with AKTP^Null-top cells or the monoculture control, suggesting a paracrine mechanism for Wnt activation in AKTP^Null cells by AKTP^R270H cells (Fig. 3B).

We previously performed RNA-seq of AKTFP^R270H/LOH (Trp53^R270H with loss of WT Trp53 by LOH) and AKTFP^R270H/+ (Trp53^R270H/+ heterozygous) organoids and found that the expression level of Ptgs2 encoding COX-2 was significantly upregulated in AKTFP^R270H/LOH cells (Fig. 3C; ref. 23). As the mutant p53 stabilization is significantly increased by loss of WT p53 by LOH (24, 25), the expression data suggest that COX-2 upregulation is caused by stabilized mutant p53 function. As expected, we confirmed increased levels of COX-2 expression in all AKTP^R270H organoid lines by immunoblotting compared with those in AKTP^Null cells (Fig. 3D). Furthermore, we confirmed COX-2 expression in AKTP^R270H cells by immunostaining, although the levels of COX-2 expression seemed to vary (Fig. 3E). On the other hand, COX-2 was not detected in AKTP^Null cells. Liver metastatic tumors of AKTP^R270H cells also showed COX-2 expression in p53-stabilized tumor cells (Fig. 3F). We also found COX-2–expressing macrophages but not α-smooth muscle actin (α-SMA)-expressing myofibroblasts in the tumor stroma of the liver metastatic tumors of AKTP^R270H cells (Fig. 3G). COX-2 is a rate-limiting enzyme for prostaglandin biosynthesis, and several studies have shown an interaction between the COX-2/PGE2 pathway and Wnt/β-catenin signaling activation (26–28). Importantly, treatment with celecoxib, a selective COX-2 inhibitor, dramatically suppressed Wnt/β-catenin signaling activation in AKTP^Null-bottom cells co-cultured with AKTP^R270H-top cells (Fig. 3H). In contrast, direct stimulation of AKTP^Null cells with PGE2 significantly increased Wnt signaling activity (Fig. 3I). Furthermore, suppression of PGE2 signaling by treating cells with the EP2/EP4 receptor inhibitor suppressed Wnt signaling activation in AKTP^Null-bottom cells that were co-cultured with AKTP^R270H-top cells (Fig. 3J). The similar results were obtained in the two independently established AKTP^Null organoid lines (Fig. 3I and J). Accordingly, these results indicate that AKTP^R270H cells transactivate Wnt/β-catenin signaling in AKTP^Null cells through induction of the COX-2/PGE2 pathway, and PGE2 signaling through the EP2 and/or EP4 receptor is vital for Wnt activation in AKTP^Null cells. These results suggest that in p53-mutant tumors, cancer cells with stabilized mutant p53 activate Wnt/β-catenin signaling in neighboring p53-destabilized cells through the COX-2/PGE2 pathway. It is also possible that macrophage-secreted PGE2 contributes to Wnt/β-catenin signaling activation in p53-destabilized cells in the liver metastatic tumors, although it remains to be further investigated.

Interestingly, Wnt signaling activity in AKTP^R270H cells was not affected by treatment with the COX-2 inhibitor or EP2/EP4 inhibitor (Supplementary Fig. S3). We previously showed that missense p53 mutations resulted in Wnt/β-catenin pathway activation with increased promoter accessibility (11). It is possible that Wnt/β-catenin signaling is activated in the p53-stabilized AKTP^R270H cells by the COX-2/PGE2 pathway–independent mechanism.

We further examined the Wnt signaling activity by imaging analysis using TOP-Venus reporter–transfected AKTP^Null cells that express Venus in response to Wnt/β-catenin signaling. Notably, the number of Venus-positive AKTP^Null cells increased significantly when co-cultured with AKTP^R270H cells compared with that of control AKTP^Null cell monocultures (Fig. 4A and B). Similar results were obtained in the two independent AKTP^Null organoid lines. These results confirmed that AKTP^R270H cells transactivate the Wnt signaling activity of AKTP^Null cells.

We next analyzed the cell proliferation rate of monocultured AKTP^Null cells and co-cultured AKTP^Null cells with AKTP^R270H-Venus cells by EdU labeling assay. Fluorescent immunostaining of the organoids revealed that the number of EdU-labeled AKTP^Null cells increased significantly when co-cultured with AKTP^R270H cells compared with that of monocultured cells (Fig. 4C and D). The similar results were obtained in the two independent AKTP^Null organoid lines.

The size of organoids developed in the Matrigel was significantly larger in AKTP^R270H compared with AKTP^Null cells, indicating a higher proliferation rate of AKTP^R270H cells (Fig. 4E and F). Interestingly, the size of mixed organoids of AKTP^Null and AKTP^R270H cells was intermediate; however, it tended to be of similar size to AKTP^R270H cells at least in one AKTP^Null organoid line [Fig. 4F (left)]. These results indicate that AKTP^R270H cells promote the proliferation of AKTP^Null cells.

Finally, we examined a role of Wnt/β-catenin signaling in AKTP^R270H cell proliferation. Notably, partial suppression of β-catenin expression by shRNA transfection significantly suppressed proliferation in the independent AKTP^R270H organoid lines (Fig. 4G and H). Accordingly, it is possible that transactivation of Wnt/β-catenin signaling in AKTP^Null cells may contribute to increased proliferation.

Recently, it has been reported that removal of mutant p53 genes in human cancer cells did not affect the proliferation and survival of cancer cells including colorectal cancer (29). On the other hand, we previously showed that mutant p53 enhances the clonal expansion and colony formation ability of the single cell–dissociated tumor cells, suggesting increased stemness of the p53-mutant cancer cells (13). As is generally thought, Wnt signaling is essential for the maintenance of tissue and tumor stem cells (30). Therefore, it is possible that mutant p53 plays an important role in colonization of the disseminated tumor cells in the liver through Wnt/β-catenin activation rather than growth of established tumors. Consistently, it has been reported that p53 mutations induce stem cell properties in cancer cells (9, 10).

Although the precise molecular mechanisms underlying the gain-of-function activity by mutant p53 remain unclear, previous studies have indicated the potential involvement of epigenetic modification by mutant p53, which resulted in significant transcriptome changes, including VEGFR2 upregulation (16, 17). Additionally, our recent findings suggest that the upregulation of the transcription factor Hmga2 by mutant p53 contributes to partial epithelial-to-mesenchymal transition and cluster migration of cancer cells (23). However, these mutant p53–induced oncogenic functions are considered cell intrinsic mechanisms. Namely, it was thought that only mutant p53–stabilized cells carry malignant phenotypes. However, in the present study, we demonstrate for the first time that mutant p53–stabilized cells may activate Wnt/β-catenin signaling of p53-destabilized cells by a paracrine mechanism. Therefore, it seems that mutant p53 enhances Wnt/β-catenin signaling of whole-tumor tissues through cell-autonomous and non–cell autonomous mechanisms.

In this study, we elucidated that mutant p53 induces COX-2 expression, and the downstream product PGE2 activates EP2/EP4 receptor signaling in surrounding p53-destabilized cells, leading to the activation of Wnt/β-catenin signaling (Fig. 5). Previously, we showed that the missense-type p53 mutation in the intestinal tumor cells leads to a dramatic transcriptomic shift through an increase in promoter accessibility (11), and inflammatory pathways, including TNFα/NF-κB, are significantly activated in p53-mutant AKTP^R270H cells (13). TNFα has been shown to induce COX-2 expression via NF-κB activation (31). Accordingly, it is possible that the TNFα/NF-κB pathway may be a potential driver of COX-2 induction in AKTP^R270H cells.

The COX-2/PGE2 pathway has been reported to activate Wnt signaling in colorectal cancer cells (26, 27), and the interaction between PGE2 and Wnt signaling has been shown to be important in the regulation of hematopoietic stem cells and organ regeneration processes (28). These results suggest that COX-2/PGE2-induced Wnt activation is thought to play an important role in the proliferation of colorectal cancers harboring p53 mutations. However, we could not elucidate the mechanism of PGE2-induced Wnt/β-catenin pathway activation in the present study. Both AKTP^R270H and AKTP^Null cells lack the WT Apc gene, resulting in constitutive stabilization of β-catenin. Accordingly, it is possible that PGE2 signaling may be involved in the regulation of nuclear localization or the degradation process of β-catenin, although it remains to be examined. In this study, we found that infiltrating macrophages in the metastatic tumor stroma also express COX-2. It is thus possible that macrophage-derived PGE2 may contribute to Wnt signaling activation in the p53-destabilized cancer cells, although this requires further investigation.

Accumulating evidence from genetic studies has revealed that the COX-2/PGE2 pathway plays a critical role in colorectal cancer development by generating an inflammatory microenvironment (32, 33). In addition, activation of the COX-2/PGE2 pathway in fibroblasts surrounding intestinal crypts has recently been reported to activate YAP via EP4 receptors on intestinal epithelial cells, thereby promoting tumorigenesis (34). These results suggest that activation of the COX-2/PGE2 pathway in mutant p53–stabilized cells may have significant effects on tumorigenesis possibly beyond activation of the Wnt/β-catenin pathway in the neighboring p53-destabilized cells. Accordingly, it is possible that the COX-2/PGE2 pathway is an effective therapeutic target against both p53-stabilized and -destabilized cells in TP53-mutant cancer.

M. Nakayama reports grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan during the conduct of the study. M. Oshima reports grants from the Japan Agency for Medical Research and Development and Ministry of Education, Culture, Sports, Science and Technology of Japan during the conduct of the study. No disclosures were reported by the other authors.

M. Nakayama: Funding acquisition, investigation, writing–review and editing. H. Saito: Investigation. K. Murakami: Methodology. H. Oshima: Investigation. M. Oshima: Conceptualization, supervision, writing–original draft, writing–review and editing.

We thank Manami Watanabe and Ayako Tsuda for their technical assistance. This work was supported by Japan Agency for Medical Research and Development (22ck0106541h0003; 22gm4010012h0002; M. Oshima), and Grants-in-Aid for Scientific Research (A; 22H00454; M. Oshima and C; 20K07585 and 23K06631; M. Nakayama) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.