Hippo cooperates with p53 to maintain foregut homeostasis and suppress the malignant transformation of foregut basal progenitor cells

Significance

Basal progenitor cells play an essential role in maintaining the homeostasis of the foregut, including both the esophagus and forestomach, by safeguarding epithelial integrity. Our research demonstrates that Mst1/2, the core components of Hippo signaling, are crucial for preserving foregut epithelial homeostasis through the Yes-associated protein (Yap)-dependent regulation of basal cell expansion. Loss of Mst1/2 accelerates the onset of foregut squamous cell carcinoma (SCC) in a Yap-dependent manner. Additionally, the combined loss of Mst1/2 and p53 is sufficient to drive the initiation of foregut SCC. These findings uncover important regulatory mechanisms controlling the function of foregut basal cells, squamous epithelial homeostasis, and the onset of foregut SCC.

Abstract

Basal progenitor cells serve as a stem cell pool to maintain the homeostasis of the epithelium of the foregut, including the esophagus and the forestomach. Aberrant genetic regulation in these cells can lead to carcinogenesis, such as squamous cell carcinoma (SCC). However, the underlying molecular mechanisms regulating the function of basal progenitor cells remain largely unknown. Here, we use mouse models to reveal that Hippo signaling is required for maintaining the homeostasis of the foregut epithelium and cooperates with p53 to repress the initiation of foregut SCC. Deletion of Mst1/2 in mice leads to epithelial overgrowth in both the esophagus and forestomach. Further molecular studies find that Mst1/2-deficiency promotes epithelial growth by enhancing basal cell proliferation in a Yes-associated protein (Yap)-dependent manner. Moreover, Mst1/2 deficiency accelerates the onset of foregut SCC in a carcinogen-induced foregut SCC mouse model, depending on Yap. Significantly, a combined deletion of Mst1/2 and p53 in basal progenitor cells sufficiently drives the initiation of foregut SCC. Therefore, our studies shed light on the collaborative role of Hippo signaling and p53 in maintaining squamous epithelial homeostasis while suppressing malignant transformation of basal stem cells within the foregut.

Basal progenitor cells serve as crucial stem cell reservoirs in multiple organs, including the skin, esophagus, trachea, and bladder (1–6). In the esophagus, basal cells exhibit the ability to proliferate to replenish the stem cell pool and can also differentiate into suprabasal cells as they ascend to form stratified epithelium. At the molecular level, basal cells express multiple markers such as p63 and Krt5, whereas suprabasal cells express Krt4 and Krt13 (1, 7). The terminally differentiated epithelial cells in the apical layers show high levels of Involucrin and Loricrin, eventually undergoing enucleation and shedding into the lumen. Basal cells play a crucial role in both the development and homeostatic maintenance of the esophageal epithelium. Excessive proliferation can lead to basal cell hyperplasia, disrupting squamous epithelial homeostasis (8). Despite its significance, the precise mechanisms regulating esophageal basal cell proliferation and its role in homeostasis have largely remained unknown. Given the structural and functional similarities between mouse and human esophageal epithelium, genetic mouse models have been extensively utilized to study esophageal morphogenesis and homeostasis (9, 10). Notably, the forestomach mucosa of mice is also composed of stratified squamous epithelium and is considered the dilated distal part of the esophagus (9, 11).

Esophageal squamous cell carcinoma (ESCC) is a deadly disease that caused approximately 500,000 deaths worldwide in 2020 (12). Unraveling the factors that trigger its onset holds the potential to develop effective preventive and curative strategies. Previous studies have identified basal cells as the cell of origin for ESCC (11, 13), and the malignant transformation of basal progenitor cells into tumor cells can be driven by a combination of overexpressing oncogenes and ablating tumor suppressors. For example, overexpression of Sox2, a putative SCC marker, collaborates with Stat3 to induce the malignant transformation of foregut basal cells (13). Additionally, p53, an important tumor suppressor, has been reported to be mutated in 56 to 96% of ESCC cases in different studies (14–17). p53 mutant basal cells in the mouse esophageal epithelium display advantages in expansion over the wild-type (WT) cells when exposed to oxidative stress from low-dose ionizing radiation (18). A combination of p53 loss and oncogene Kras overexpression in basal cells leads to the development of aggressive forestomach SCC (11). Despite the significance of oncogenes in ESCC, their genetic alternations are only partially observed in human ESCC samples (17). For example, the chromosome segment 3q26.33 containing SOX2 is amplified in only 15% of human ESCC samples, and KRAS has been reported in less than 12% of ESCC samples (17, 19, 20), implying the involvement of other genes in ESCC onset. The role of these additional genes and their interaction with p53 in foregut SCC initiation warrants further investigation.

The Hippo signaling pathway is an important regulatory mechanism that controls a diverse array of biological processes, including cell proliferation, organogenesis, tissue homeostasis, and carcinogenesis (21, 22). In mammals, the serine/threonine kinases Mst1 and Mst2 (Mst1/2) represent core components of the Hippo signaling pathway (21). Inactivation of Mst1/2 results in the dephosphorylation and stabilization of Yap, leading to its translocation into the nucleus, where it cooperates with TEADs to activate the expression of downstream targets (21). In our previous study, we found that Yap promotes esophageal growth by upregulating basal cell proliferation, whereas its ablation leads to esophageal atresia during mouse embryonic development (23). However, the role of Hippo signaling in regulating basal cell function and esophageal homeostasis at postnatal stages is unknown. Additionally, prior in vitro studies employing human cancer cell lines have indicated that Yap promotes the progression of ESCC (24), suggesting the involvement of the signaling in the ESCC progression. Nonetheless, the role of Hippo signaling in the initiation of ESCC remains unexplored. Furthermore, these in vitro studies lack the critical components of the tumor microenvironment, such as immune cells and stroma, that are critical for tumorigenesis (25). Hence, leveraging genetic mouse models to dissect the role of Hippo signaling components in vivo will significantly enhance our comprehension of ESCC initiation.

In this study, we demonstrate that Hippo signaling maintains the homeostasis of both the esophageal and forestomach epithelium while collaborating with p53 to repress the onset of foregut SCC. Using genetic mouse models, we show that Mst1/2 deficiency leads to basal cell hyperplasia and overgrowth of the foregut squamous epithelium. Through molecular and histological analyses, we unravel that Mst1/2 ablation promotes basal cell proliferation via Yap. Significantly, Mst1/2 deficiency accelerates the onset of the carcinogen-induced SCC in the forestomach in a Yap-dependent manner. Notably, a combined ablation of Mst1/2 and p53 is sufficient to initiate foregut SCC. Lastly, MST1/2 expression levels are lower, while YAP expression levels are higher in the human ESCC samples than in the normal esophageal epithelium, indicating the clinical relevance of the mouse genetic studies.

Results

Deletion of Mst1/2 Dysregulates the Homeostasis of the Esophageal and Forestomach Epithelium.

While the Hippo signaling pathway has been shown to regulate tissue homeostasis in multiple organs (26–28), its role in esophageal homoeostasis remains undetermined. We first used immunostaining to reveal that Mst1 and Mst2 were both expressed in the mouse esophageal epithelium (Fig. 1 A and B). To assess their function in the homeostatic maintenance of the esophageal epithelium, we generated Mst1/2 double knockout (DKO) mice, p63CreERT2;Mst1^loxp/loxp;Mst2^−/−, in which Mst1/2 were deleted in the basal cells and their derivatives in the foregut squamous epithelium upon tamoxifen administration (29, 30) (Fig. 1 C–E). Remarkably, the Mst1/2 DKO mice displayed an increase in esophageal epithelial thickness and basal cell expansion compared to WT mice (Fig. 1 F and G). It is worth noting that the esophageal epithelium of p63CreERT2 or Mst2-deficient mice remained normal (SI Appendix, Fig. S1). Similarly, epithelial overgrowth and basal cell hyperplasia were also observed in the forestomach squamous epithelium of Mst1/2 DKO mice (Fig. 1 H and I). Together, these results support that Mst1/2 are essential for maintaining the homeostasis of the stratified squamous epithelium in the esophagus and forestomach.

Fig. 1.

Deletion of Mst1/2 dysregulates the homeostasis of the esophageal and forestomach epithelium. (A and B) The expression of Mst1 (A) and Mst2 (B) in the esophageal epithelium. (Scale bar: 10 µm.) (C) A schematic diagram of the procedure for generating Mst1/2 conditional knockout in the esophageal epithelium. Mst1/2 DKO, p63CreERT2;Mst1^loxp/loxp;Mst2^−/−. (D and E) Western blot analysis shows efficient Mst1 and Mst2 knockout in the esophageal epithelium of Mst1/2 DKO mutants. (F and G) Representative H&E-stained sections of the WT and Mst1/2 DKO esophageal epithelium (F) and the quantification of the thickness of the esophageal epithelium (G). Dotted lines indicate the basal cell layer. Data represent mean ± SD (WT, n = 6; Mst1/2 DKO, n = 9). ***P < 0.001, by unpaired, two-tailed Student’s t test. Note the increase in basal cell number and the thickness of the epithelium in the Mst1/2 DKO mice. (Scale bar: 20 µm.) (H) Representative gross morphology of the stomach. Black dotted lines denote the squamous-columnar junction. Note the hyperplasia in the forestomach of Mst1/2 DKO mice (black arrow). (Scale bar: 2 mm.) (I) H&E-stained sections of the WT and Mst1/2 DKO forestomach. Note the basal cell hyperplasia in the forestomach epithelium. (Scale bar: 200 μm.)

Mst1/2 Regulate Multiple Biological Processes to Modulate the Homeostasis of the Squamous Epithelium.

To elucidate the mechanisms underlying Mst1/2 modulation of epithelial homeostasis, esophageal epithelia were isolated from Mst1/2 DKO and WT mice and subjected to bulk RNA sequencing (Fig. 2A). A Pearson correlation coefficient analysis of transcription profiles revealed consistent expression patterns across different samples in both Mst1/2 DKO and WT mice (Fig. 2B), indicating high biological reproducibility. Notably, principal component analysis and gene expression hierarchical clustering delineated a distinct expression profile in Mst1/2 DKO mutants compared to WT mice (Fig. 2 C and D). Differentially expressed gene (DEG) analysis identified 569 up-regulated and 842 down-regulated genes, which were at least twofold altered in the mutants (Fig. 2E). Gene Ontology enrichment analysis revealed that the DEGs were involved in multiple biological processes such as DNA replication, mitotic cell cycle phase transition, regulation of cell growth, epithelial cell migration, and keratinocyte differentiation, among others (Fig. 2F). KEGG pathway analysis showed a downregulation of the Hippo signaling pathway but an upregulation of signaling involved in cell cycle, DNA replication, and chemical carcinogenesis (Fig. 2G). Moreover, gene set enrichment analysis (GSEA) revealed that Mst1/2 DKO led to an upregulation of the Yap-activated gene signature (Fig. 2H). Noteworthy up-regulated cell cycle positive regulators include Ccng1, Ccne1, Ccne2, Ccnd3, and Ccnc (31, 32) (Fig. 2I). Collectively, these results show that Mst1/2 regulate multiple biological processes to modulate squamous epithelial homeostasis.

Fig. 2.

Mst1/2 regulate multiple biological processes to modulate the homeostasis of the squamous epithelium. (A) Schematic diagram showing esophageal epithelium harvested from WT or p63CreERT2;Mst1^loxp/loxp;Mst2^−/−(KO) mice administered with tamoxifen. RNA was isolated from the esophageal epithelium and sequenced. (B and C) Pearson correlation coefficient analysis (B) and principal component analysis (C) were performed based on the transcriptional expression profiles of the KO versus WT samples. (D and E) DEGs between KO and WT mice were presented by hierarchical clustering (D) and a volcano plot (E). (F and G) Gene Ontology enrichment (F) and KEGG pathway (G) analysis of the DEGs between KO and WT mice. (H) GSEA indicates upregulation of the Yap-activated gene signature. (I) Heatmap of the transcript levels of cell cycle positive regulators Ccng1, Ccne1, Ccne2, Ccnd3, and Ccnc in the mouse esophageal epithelium. Note that these genes are upregulated in the Mst1/2 DKO mutants. Transcript levels were determined by RNA sequencing and presented as FPKM values. WT, wild type; KO, knockout; DKO, double knockout. n = 5 per genotype in (A–I).

Mst1/2 Deficiency Promotes Basal Progenitor Cell Proliferation through Yap in the Esophageal and Forestomach.

Given the epithelia hyperplasia in the foregut epithelium and elevated expression of cell cycle positive regulators, we conducted p63 and Ki67 immunostaining on the tissues. We found that the proliferating basal cells (p63⁺Ki67⁺) were significantly increased in the esophageal epithelium of Mst1/2 DKO mutants (Fig. 3A), indicating the significant role of Mst1/2 in restraining basal cell expansion. Additionally, we noted an upregulation of Yap expression levels, a key downstream effector of Mst1/2, in the Mst1/2 DKO esophageal epithelium (Fig. 3B). To probe further, we generated Mst1/2 Yap triple knockout (TKO) mutants, which exhibited a significant decrease in proliferating basal cells (Fig. 3 A and B), suggesting that Yap acts downstream of Mst1/2 to promote basal cell proliferation. Moreover, the increase in basal cell proliferation led to enhanced stratification in Mst1/2 DKO mutants, whereas the stratification was significantly reduced in Mst1/2 Yap TKO mice (Fig. 3 C and D). These findings demonstrate that Mst1/2 deficiency promotes basal progenitor cell proliferation in the esophagus in a Yap-dependent manner. Notably, the esophageal epithelium of Yap KO mutants appeared normal, indicating a dispensable role of Yap in maintaining the esophageal epithelium (SI Appendix, Fig. S1). Similarly, enhanced basal cell proliferation and epithelial stratification were observed in the forestomach epithelium of Mst1/2 DKO mutants, whereas significantly mitigated upon Yap deletion (Fig. 3 E–G). Together, these results show that Mst1/2 deficiency deregulates the homeostasis of the foregut squamous epithelium by augmenting basal cell proliferation mediated through Yap.

Fig. 3.

Mst1/2 deficiency promotes basal progenitor cell proliferation through Yap in the esophageal and forestomach. (A) Immunofluorescence staining of basal cell marker p63 and cell proliferation marker Ki67 in the esophageal epithelium. Note the increased proliferating basal cells in the Mst1/2 DKO mutants while reduced proliferating basal cells with Yap deletion. Data represent mean ± SD (n = 9 per genotype). ***P < 0.001, by unpaired, two-tailed Student’s t test. (Scale bar: 20 µm.) (B) IHC of Yap in the mouse esophageal epithelium. Note the increased protein levels of Yap in the p63CreERT2;Mst1^loxp/loxp;Mst2^−/−(Mst1/2 DKO) mutants and the loss of Yap expression in the p63CreERT2;Mst1^loxp/loxp;Mst2^−/−;Yap^loxp/loxp(Mst1/2 Yap TKO) mutants. WT, wild type. (Scale bar: 20 μm.) (C and D) Immunofluorescence staining of p63 and Krt4 in the esophageal epithelium (C) and quantification of the Krt4⁺ esophageal epithelium thickness (D). Note the increased Krt4⁺ epithelial thickness in the Mst1/2 DKO mutants but reduced Krt4⁺ epithelial thickness with Yap deletion. Data represent mean ± SD (n = 3 per genotype). **P < 0.01, by unpaired, two-tailed Student’s t test. (Scale bar: 20 µm.) (E and F) Immunofluorescence staining of basal cell marker p63 and cell proliferation marker Ki67 in the forestomach epithelium (E) and quantification of p63⁺Ki67⁺ proliferation basal cells (F). Note the increased proliferating basal cells in the Mst1/2 DKO mutants while reduced proliferating basal cells with Yap deletion. Data represent mean ± SD (n = 9 per genotype). ***P < 0.001, by unpaired, two-tailed Student’s t test. (Scale bar: 200 µm.) (G) Immunofluorescence staining of p63 and Krt4 in the forestomach epithelium. WT, wild type; Mst1/2 DKO, p63CreERT2;Mst1^loxp/loxp;Mst2^−/−; Mst1/2 Yap TKO, p63CreERT2;Mst1^loxp/loxp;Mst2^−/−;Yap^loxp/loxp. (Scale bar: 20µm.)

Mst1/2 Deficiency Accelerates the Initiation of Foregut SCC in a Yap-Dependent Manner.

Since basal cell hyperplasia that developed in the Mst1/2 DKO mutants is an important precancerous condition of SCC (33), our further investigation aimed to determine whether Mst1/2 deficiency would accelerate the initiation of foregut SCC. To evaluate this, mice were fed water containing 4-Nitroquinoline 1-Oxide (4-NQO), a carcinogen used to induce foregut SCC (34) (Fig. 4A). Significantly, 42% of Mst1/2 DKO mice developed SCC in the forestomach within 8 wk, whereas WT and Mst1/2 Yap TKO mice only displayed normal to dysplasia symptoms (Fig. 4 B–D). After 16 wk, all Mst1/2 DKO mutants developed SCC, with 62.5% displaying invasive phenotypes (Fig. 4 B–D). These tumors were characterized by high expression of SCC diagnostic markers, including Sox2, Krt5, Krt14, p63, and Itga6 (35–37) (SI Appendix, Fig. S2 A–D). In contrast, only 25% of WT mice developed SCC (Fig. 4 B–D). Immunostaining analysis showed a substantial increase in proliferating tumor cells in Mst1/2 DKO mice, whereas Yap deletion mitigated the proliferation (Fig. 4 E–H). Furthermore, the squamous differentiation in the SCC of Mst1/2 DKO mutants was diminished compared to the WT mice, a reversal observed in Mst1/2 Yap TKO mice (Fig. 4 I and J). These findings together indicate that Mst1/2 deficiency accelerates the initiation of the foregut SCC through Yap.

Fig. 4.

Mst1/2 deficiency accelerates the initiation of foregut SCC in a Yap-dependent manner. (A) Schematic illustrating 4-NQO-induced squamous cell carcinoma in the forestomach. Tmx, tamoxifen. (B) Representative gross morphology of the forestomach of mice fed 4-NQO water. Note that the tumors started to form at 8 wk in the forestomach of Mst1/2 DKO mice, while Mst1/2 Yap TKO mutants showed a delayed tumor initiation. (Scale bar: 2 mm.) (C) Representative H&E-stained forestomach sections of mice fed 4-NQO water. Note the carcinoma in situ that formed in the Mst1/2 DKO mice. (Scale bar: 400 µm.) (D) Percentage of the indicated types of lesions in the forestomach of mice fed 4-NQO water (WT, n = 9 for 8 wk, n = 8 for 16 wk; Mst1/2 DKO, n = 12 for 8 wk, n = 8 for 16 wk; Mst1/2 Yap TKO, n = 8 for 8 wk). CIS, carcinoma in situ; SCC, squamous cell carcinoma. (E–H) Immunofluorescence staining of p63 and Ki67 and quantification of p63⁺Ki67⁺ tumor cells in the forestomach of mice fed 4-NQO water for 8 wk (E and F) and 16 wk (G and H). Note that p63⁺Ki67⁺ tumor cells were increased in the Mst1/2 DKO mice, while the cells were reduced in the Mst1/2 Yap TKO mutants. Data represent mean ± SD (n = 9 per genotype). **P < 0.01, ***P < 0.001, by unpaired, two-tailed Student’s t test. (Scale bar: 100 µm.) (I and J) Immunofluorescence staining of p63 and Krt4 in the forestomach epithelium of mice fed 4-NQO water for 8 wk (I) and 16 wk (J). Note the reduced Krt4⁺ differentiated cells in the Mst1/2 DKO mutants. (Scale bar: 50 µm.) WT, wild type; Mst1/2 DKO, p63CreERT2;Mst1^loxp/loxp;Mst2^−/−; Mst1/2 Yap TKO, p63CreERT2;Mst1^loxp/loxp;Mst2^−/−;Yap^loxp/loxp.

Mst1/2 Cooperate with p53 to Repress the Initiation of the Foregut SCC.

Frequent mutations of the tumor suppressor p53 are a hallmark in ESCC cases (14, 17). Here, our investigation delved into whether loss of both Mst1/2 and p53 is sufficient to initiate foregut SCC. To address this, we generated p63CreERT2;Mst1^loxp/loxp;Mst2^−/−;p53^loxp/loxp(Mst1/2 p53 TKO) mice to induce Mst1/2 and p53 deletion upon tamoxifen administration (Fig. 5A). Strikingly, 80% of Mst1/2 p53 TKO mice developed SCC after 16 wk, and all mice exhibited SCC in the forestomach by 24 wk, with 60% displaying invasive phenotypes (SI Appendix, Fig. S3 A–D and Fig. 5 A–D). Conversely, Mst1/2 DKO and p53 KO mice only displayed hyperplasia or dysplasia by 16 or 24 wk, while all WT mice remained normal (SI Appendix, Fig. S3 A–D and Fig. 5 A–D). Immunostaining showed that Mst1/2 p53 TKO mice exhibited the highest number of proliferating tumor cells (Fig. 5 E and F) but the lowest levels of differentiation (Fig. 5G). Therefore, these results demonstrate that Mst1/2 collaborate with p53 to repress the initiation of foregut SCC.

Fig. 5.

Mst1/2 cooperate with p53 to repress the initiation of the foregut squamous cell carcinoma. (A) Schematic illustration of experiments. Mice were administered three doses of tamoxifen every 4 wk. (B and C) Representative gross morphology of the forestomach of mice (B) and percentage of the indicated types of lesions (C). Note that tumors exclusively formed in the forestomach of Mst1/2 p53 TKO mice, while only dysplasia or hyperplasia developed in the Mst1/2 DKO or p53 KO mutants (WT, n = 5; Mst1/2 DKO, n = 4; p53 KO, n = 4; Mst1/2 p53 TKO, n = 5). CIS, carcinoma in situ; SCC, squamous cell carcinoma. (Scale bar: 2 mm.) (D) Representative H&E-stained forestomach sections of mice. (Scale bar: 200 µm.) (E and F) Immunofluorescence staining of p63 and Ki67 and quantification of p63⁺Ki67⁺ tumor cells in the forestomach of mice. Note the highly increased proliferating tumor cells in the Mst1/2 p53 TKO mutants. Data represent mean ± SD (n = 9 per genotype). **P < 0.01, ***P < 0.001, by unpaired, two-tailed Student’s t test. (Scale bar: 50 µm.) (G) Immunofluorescence staining of p63 and Krt4 in the forestomach epithelium of mice. Note the reduced Krt4⁺ differentiated cells in the Mst1/2 p53 TKO mutants. (Scale bar: 50 µm.) WT, wild type; Mst1/2 DKO, p63CreERT2;Mst1^loxp/loxp;Mst2^−/−; Mst1/2 p53 TKO, p63CreERT2;Mst1^loxp/loxp;Mst2^−/−;p53^loxp/loxp.

The Expression of MST1/2 Is Down-Regulated while YAP Is Up-Regulated in Human ESCC.

Finally, we conducted immunohistochemistry (IHC) staining on MST1/2 and YAP to determine their expression levels in ESCC samples and normal esophageal epithelium from patients. Our analyses revealed significantly lower expression levels of MST1 and MST2 in ESCC samples compared to normal samples (Fig. 6 A–D and SI Appendix, Fig. S4 A and B). This indicates a negative association of MST1/2 expression levels with human ESCC development, aligning with their suppressive role in the onset of foregut SCC in murine models. Conversely, the expression of YAP was significantly increased in the ESCC samples compared to normal samples (Fig. 6 E and F and SI Appendix, Fig. S4C), indicating a positive correlation between YAP expression level and human ESCC development, consistent with Yap promoting ESCC initiation following Mst1/2 deletion in mice. Furthermore, we analyzed the genetic alterations in the major components of the Hippo signaling pathway in human ESCC patients using The Cancer Genome Atlas database (38). Among the 347 ESCC patient samples assessed, 146 (42%) displayed genetic alterations involving 29 genes (SI Appendix, Fig. S5). Specifically, MST1, MST2, and YAP showed genetic alteration frequencies of 2.5%, 4%, and 5%, respectively. Notably, genetic changes observed in FAT1-4 and LATS1/2, acting as upstream and downstream positive regulators of MST1/2, respectively, mainly comprised deep deletion, missense, and truncating mutations. In contrast, the downstream transcriptional coactivator TAZ showed consistent amplification across cases. These findings emphasize the clinical relevance and implications of Hippo signaling in human ESCC pathogenesis.

Fig. 6.

The expression of MST1/2 is down-regulated while YAP is upregulated in human ESCC. (A and B) Representative MST1 IHC images (A) and quantification of MST1 IHC scores (B). Note that the expression of MST1 is down-regulated in the ESCC samples. (C and D) Representative MST2 IHC images (C) and quantification of MST2 IHC scores (D). Note that the expression of MST2 is down-regulated in the ESCC samples. (E and F) Representative YAP IHC images (E) and quantification of YAP IHC scores (F). Note that the expression of YAP is upregulated in the ESCC samples. Data represent mean ± SD (Normal, n = 42; ESCC, n = 44). **P < 0.01, ***P < 0.001, by unpaired, two-tailed Student’s t test. IHC, immunohistochemistry; ESCC, esophageal squamous cell carcinoma. (Scale bar: 50 µm.)

Discussion

In this study, we have revealed that Mst1/2 maintain esophageal and forestomach epithelial homeostasis and cooperate with p53 to repress foregut SCC initiation. Ablation of Mst1/2 in mice dysregulates the homeostasis of the esophageal and forestomach squamous epithelium. Gene expression profiling shows that Mst1/2 regulate foregut homeostasis by controlling various biological processes and cell cycle positive regulators. Our histological analysis further demonstrates that Mst1/2 deficiency disrupts foregut epithelial homeostasis by upregulating basal cell proliferation in a Yap-dependent manner. Significantly, the absence of Mst1/2 accelerates the initiation of foregut SCC in a carcinogen-induced mouse model. More importantly, we establish that the concurrent deficiency of Mst1/2 and p53 is sufficient to drive foregut SCC initiation. Lastly, the expression of MST1/2 is down-regulated, whereas the expression of YAP is up-regulated in the human ESCC samples compared to normal tissues, suggesting their significant involvement in ESCC pathogenesis.

Previous studies have shown that Hippo signaling is involved in maintaining the homeostasis of multiple tissues, such as the intestinal epithelium and liver (26–28). Yet, the role of Hippo in esophageal homeostasis remains undetermined. Here, we first used immunostaining to show that Mst1/2, the core components of Hippo signaling, were expressed in the esophageal epithelium. Subsequent genetic ablation of Mst1/2 within the epithelium of the esophagus and forestomach, the counterpart of the human distal esophagus, resulted in increased epithelial thickness and basal cell hyperplasia. These findings support the essential role of Hippo signaling in maintaining the homeostasis of the foregut epithelium. Further RNA sequencing analysis revealed over 1,400 genes with at least a twofold change in expression levels in the esophageal epithelium of Mst1/2 DKO mutants. Notably, the DEGs in Mst1/2 DKO mutants were involved in multiple biological processes, including DNA replication and mitotic cell cycle phase transition. Specifically, cyclins, such as Ccng1, Ccne1, and Ccnd3, known as cell cycle positive regulators (32), were significantly upregulated with the loss of Mst1/2. Histological analysis further highlighted a significant increase in basal progenitor cell proliferation within the esophageal and forestomach epithelium of Mst1/2 DKO mutants, contrasting with reduced proliferation upon Yap deletion. Consequently, stratification was enhanced in Mst1/2 DKO mutants, a trend counteracted by Yap ablation. These observations align with prior studies illustrating that Mst1/2 deficiency can boost cell proliferation, leading to the overgrowth of various tissues such as the liver and the intestine, by upregulating the expression of cyclins (39–42). Collectively, these results indicate that Mst1/2 maintain foregut epithelial homeostasis by curbing basal cell expansion in a Yap-dependent manner.

As basal cell hyperplasia is an important precancerous characteristic of ESCC (33), we extended our study by utilizing a 4-NQO-induced foregut SCC mouse model to investigate the potential involvement of Hippo signaling in the onset of foregut SCC. Our findings showed that by 8 wk, 42% of Mst1/2 DKO mice had developed foregut SCC, whereas no WT mice exhibited SCC at this stage. By 16 wk, all Mst1/2 DKO mice had developed SCC, with 62.5% of the tumors displaying invasive phenotypes, in contrast to only 25% of SCC development in WT mice. However, the removal of Yap significantly reduced the rates of SCC occurrence. These findings demonstrate that Mst1/2 repress the initiation of foregut SCC, depending on Yap. Immunostaining indicated a substantial upregulation of tumor cell proliferation in Mst1/2 DKO mutants, which was significantly downregulated upon Yap ablation. Furthermore, epithelial differentiation was lowered by the deletion of Mst1/2, which was significantly reversed by the deletion of Yap. These phenotypes developing in the Mst1/2 DKO mutants are consistent with the hallmark traits of human ESCC, specifically characterized by high basal cell proliferation but low squamous differentiation within tumors (43). Therefore, these findings support that the inactivation of Hippo signaling in basal progenitor cells accelerates the initiation of foregut SCC.

Our further studies found that deleting Mst1/2 in the basal cells only led to hyperplasia or dysplasia in the foregut epithelium by 24 wk but failed to progress to SCC. Despite the frequent mutation rates of tumor suppressor p53 and its potential implications in human ESCC development (11, 14, 17, 18), ablating p53 in the foregut basal cells also only resulted in hyperplasia or dysplasia after 24 wk. These results suggest that Mst1/2 DKO or p53 deficiency alone is insufficient to drive foregut SCC, at least at this stage. Therefore, we generated Mst1/2 p53 TKO in the basal cells to determine whether the inactivation of both Hippo and p53 can initiate the foregut SCC. Strikingly, 80% of Mst1/2 p53 TKO mice developed SCC in the forestomach by 16 wk, and all the mice exhibited SCC by 24 wk. The forestomach of these mice appears more prone to developing SCC compared to the esophagus, potentially due to microenvironmental factors, including enhanced inflammation induced by exposure to various insulants such as gastric acids (9, 13, 44). These results support that Hippo and p53 cooperate to repress the malignant transformation of basal cells, and the deficiency of both signals is sufficient to drive foregut SCC initiation. These findings are significant as they demonstrate that the simultaneous loss of a tumor-suppressing signal, coupled with the absence of p53, can autonomously initiate foregut SCC. This complements previous studies that only indicated foregut SCC initiation through overexpressing oncogenes such as Kras or Sox2 (11, 13).

Lastly, we performed immunostaining on human ESCC tissues and found that MST1 and MST2 were both down-regulated in tumors compared to normal tissues, suggesting a negative correlation between MST1/2 expression level and human ESCC occurrence. On the contrary, an upregulation of YAP expression in human ESCC tissues indicated a positive association of YAP levels with human ESCC development. This aligns with the important role of Yap in mediating foregut SCC initiation in the carcinogen-treated Mst1/2-deficient mice. It also further corroborates previous findings that Yap promotes the growth of human cancer cells using in vitro assays (24). Furthermore, our analysis revealed prevalent genetic changes in key Hippo signaling components among human ESCC samples. Notably, MST1/2 and YAP were genetically altered. Moreover, the upstream and downstream regulators of MST1/2, including FAT1-4 and LATS1/2, predominantly exhibited deletions and mutations. In contrast, the downstream transcriptional coactivator TAZ largely displayed amplifications. Investigating these additional components will be critical for fully understanding the role of Hippo signaling in initiating foregut SCC. In addition to genetic changes, epigenetic regulation likely contributes to modulating MST1/2 expression in human ESCC. A prior study identified hypermethylation of the genomic region encompassing MST2 in human ESCC (45). In prostate cancer cells, DNA methylation within the promoter region and EZH2-mediated H3K27me3 repress MST1 expression (46). Investigating the direct impact of these various epigenetic regulations on MST1/2 expression and their role in ESCC will be crucial for future studies. Hence, our findings hold important clinical significance.

Overall, our study has identified an essential role of Hippo signaling in maintaining the homeostasis of the foregut epithelium by restricting basal cell proliferation. Moreover, we have demonstrated that Hippo signaling cooperates with p53 to suppress the onset of foregut SCC. These findings provide significant insights into the mechanisms that underlie basal cell function, foregut epithelial homeostasis, and tumorigenesis.

Materials and Methods

p63CreERT2, Mst1^loxp/loxp, Mst2^−/−, Yap^loxp/loxp, and p53 ^loxp/loxp mouse lines used in this study have been previously described (29, 47–50). All animal experiments were conducted in compliance with protocols approved by the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University. The study on human ESCC tissue arrays was approved by the ethics committee of Fujian Medical University Union Hospital. Written consent was obtained from all the enrolled patients.

Detailed protocols for human ESCC tissue array and gene mutation analysis, mouse model of foregut SCC, hematoxylin and eosin staining, immunofluorescence and IHC staining, microscopy imaging, isolation of the esophageal epithelium, RNA and protein isolation, western blot analysis, RNA sequencing, differential gene expression, Gene Ontology enrichment, KEGG, and GSEA analysis, and quantification and statistical analysis can be found in SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

The RNA sequencing data have been deposited in the Gene Expression Omnibus (GEO) database (accession number GSE248501) (51). All other data are included in the manuscript and/or SI Appendix.