Genome-wide CRISPR/Cas9 screening identifies a targetable MEST-PURA interaction in cancer metastasis

Abstract

Background: Metastasis is one of the most lethal hallmarks of esophageal squamous cell carcinoma (ESCC), yet the mechanisms remain unclear due to a lack of reliable experimental models and systematic identification of key drivers. There is urgent need to develop useful therapies for this lethal disease.

Methods: A genome-wide CRISPR/Cas9 screening, in combination with gene profiling of highly invasive and metastatic ESCC sublines, as well as PDX models, was performed to identify key regulators of cancer metastasis. The Gain- and loss-of-function experiments were taken to examine gene function. Protein interactome, RNA-seq, and whole genome methylation sequencing were used to investigate gene regulation and molecular mechanisms. Clinical significance was analyzed in tumor tissue microarray and TCGA databases. Homology modeling, modified ELISA, surface plasmon resonance and functional assays were performed to identify lead compound which targets MEST to suppress cancer metastasis.

Findings: High MEST expression was associated with poor patient survival and promoted cancer invasion and metastasis in ESCC. Mechanistically, MEST activates SRCIN1/RASAL1-ERK-snail signaling by interacting with PURA. miR-449a was identified as a direct regulator of MEST, and hypermethylation of its promoter led to MEST upregulation, whereas systemically delivered miR-449a mimic could suppress tumor metastasis without overt toxicity. Furthermore, molecular docking and computational screening in a small-molecule library of 1,500,000 compounds and functional assays showed that G699-0288 targets the MEST-PURA interaction and significantly inhibits cancer metastasis.

Interpretation: We identified the MEST-PURA-SRCIN1/RASAL1-ERK-snail signaling cascade as an important mechanism underlying cancer metastasis. Blockade of MEST-PURA interaction has therapeutic potential in management of cancer metastasis.

Funding: This work was supported by National Key Research and Development Program of China (2021YFC2501000, 2021YFC2501900, 2017YFA0505100); National Natural Science Foundation of China (31961160727, 82073196, 81973339, 81803551); NSFC/RGC Joint Research Scheme (N_HKU727/19); Natural Science Foundation of Guangdong Province (2021A1515011158, 2021A0505030035); Key Laboratory of Guangdong Higher Education Institutes of China (2021KSYS009).

Keywords: CRISPR/Cas9 screening; Cancer metastasis; Esophageal cancer; Noncoding RNA; Protein interactome.

Fig. 1

Integrative analysis of genome-wide CRISPR/Cas9 screening and gene profiling of highly metastatic cell model identifies MEST as a driver of cancer metastasis. (a) Schematic diagram illustrating the CRISPR/Cas9 screening of key regulators of cancer metastasis. (b) Comparison of invasion ability of GeCKOv2 cells and input cells after three rounds of invasion selection. (c) Cumulative frequency of sgRNAs in GeCKOv2 cells and input cells. (d) A heatmap displaying the 8927 significantly depleted sgRNAs targeting 1489 genes. (e) A total of 17 genes were overlapped between the data from genome-wide CRISPR/Cas9 screening and RNA-seq of highly invasive ESCC subline. (f) Scatter plot showing the sgRNA abundance from CRISPR/Cas9 screening and the expression fold change from RNA-seq of the 17 overlapped genes. (g) Comparison of read counts of individual sgRNAs targeting MEST in GECKOv2 and input cells. (h, i) Diagram illustrating the generation of highly metastatic cancer cells (KYSE150-Luc-LM5) by using a serial intravenous injection mouse model. (j) Western blot showing the expression of MEST in KYSE150-Luc-LM5 and KYSE150 cells, as well as KYSE410-I6 and KYSE410 cells. Bars, SD; ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001, the student's t test.

Fig. 2

Clinical and biological significance of MEST in ESCC metastasis. (a) qRT-PCR analysis and expression pattern of MEST in 40 pairs of ESCC tumor and normal tissues. (b) Expression pattern of MEST in 242 ESCC tissues and 212 paired normal tissues. (c) Kaplan–Meier analysis curves of MEST expression in 242 ESCC patients. (d) Violin plot showing the expression of MEST in primary ESCC without (n = 133) or with metastasis (n = 28) in The Cancer Genome Atlas (TCGA). (e, f) Diagram showing the establishment of a patient-derived xenograft (PDX) metastasis model and comparison of MEST expression in the nine PDXs. (g) Expression pattern of MEST in 40 ESCC tumors and corresponding metastatic tissues. (h) Boyden chamber assay showing the effect of MEST on the invasive abilities of KYSE150 and KYSE410 cells. (i) Bioluminescence imaging and quantification showing that MEST-overexpressing ESCC cells had a higher potential to metastasize to the lungs in mice than control cells (n = 6/group). (j) Boyden chamber assay showing the effect of MEST knockdown on the invasion of ESCC cells. (k) Effect of MEST silencing on tumor metastasis (n = 6/group). Bars, SD; ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001, the student's t test.

Fig. 3

MEST activates ERK signaling through the regulation of SRCIN1 and RASAL1 expression. (a) IPA analysis showing that MEST may activate the ERK signaling pathway. (b) Western blot showing the expression of p-ERK and ERK when the expression of MEST was manipulated. (c) Boyden chamber showing the effect of MEST on ESCC invasion in the presence or absence of the MEK inhibitor U0126 (10 μM) for 24 h. (d, e) The mRNA and protein expression of SRCIN1 and RASAL1 upon ectopic expression of MEST in KYSE150 and KYSE410 cells. (f) Boyden chamber assay showing the invasion of MEST-overexpressing cells when the expression levels of SRCIN1 and RASAL1 were manipulated. (g) Bioluminescence imaging and quantification showing the metastatic potential of KYSE150-Luc-MEST-CON, KYSE150-Luc-MEST-SRCIN1, KYSE150-Luc-MEST-RASAL1, and vector control cells (n = 6/group). (h) Western blot showing the expression of p-ERK, p-MEK, p-RAF and snail in MEST-overexpressing cells when the expression levels of SRCIN1 and RASAL1 were manipulated. (i) Association between MEST and SRCIN1 or RASAL1 in ESCC. Bars, SD; ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001, the student's t test.

Fig. 4

MEST interacts with PURA to activate the ERK-snail pathway. (a) Diagram showing the approach to identify the overlapping proteins that interact with MEST and simultaneously regulate the transcription of SRCIN1 and RASAL1. (b) Coimmunoprecipitation showing the binding of PURA with MEST. (c) 3D modeling showing the interaction between MEST and PURA. (d) Diagram showing the key amino acids selected to investigate the interaction between MEST and PURA, as well as the mutation design for expressing mutant proteins. (e) The interaction between PURA and wild-type or mutant MEST, as well as MEST and wild type or mutant PURA in KYSE150-Luc cells. (f) qRT-PCR assay showing the effect of PURA overexpression on the mRNA expression levels of SRCIN1 and RASAL1 in KYSE150 and KYSE410 cells. (g) Putative binding sites of PURA in the promoter regions of SRCIN1 and RASAL1 were identified by in silico prediction, and the enrichment of PURA in the promoter regions was determined by ChIP. (h, i) Western blot showing the effect of MEST overexpression on the expression levels of SRCIN1, RASAL1, p-ERK, p-MEK, p-RAF and snail in PURA-deficient ESCC cells with or without reintroduction of PURA. (j, k) Boyden chamber and experimental metastasis assays were performed to examine the invasive and metastatic potential of PURA-deficient ESCC cells when both MEST and PURA or MEST alone were overexpressed. Bars, SD; ∗∗, P < 0.01; ∗∗∗, P < 0.001, the student's t test.

Fig. 5

Promoter hypermethylation results in deregulation of miR-449a in ESCC. (a) Diagram showing the strategy to screen potential miRNAs that target MEST and inhibit cancer metastasis. (b) TargetScan bioinformatics algorithm showing the base pairing between miR-449a and the 3′UTR of MEST. (c) Comparison of miR-449a expression in KYSE150-Luc-LM5 or KYSE410-I6 cells with that in their respective parental cells by TaqMan miRNA PCR assay. (d) Representative image showing the expression of miR-449a in ESCC tumor and matched normal tissue (left panel). Expression pattern of miR-449a in primary tumor (n = 242) and matched normal tissue (n = 212) (right panel). (e) The correlation between MEST and miR-449a expression in 242 ESCC tumors examined is shown. (f) Kaplan–Meier analysis of overall survival of 242 ESCC patients stratified according to tumor miR-449a expression. (g) Representative image and expression pattern of miR-449a in primary ESCC tumor tissues (n = 40) and matched metastatic tissues (n = 40). (h) The methylation status of the miR-449a promoter in immortalized normal esophageal epithelial cell lines and ESCC cell lines was examined by whole genomic methylation sequencing. Each plot represents the methylation percentage of each site. (i) qRT-PCR analysis was used to detect the expression level of miR-449a in KYSE150 and KYSE410 cells treated with 5′-aza-2'deoxycytidine (5-Aza). (j) Detection of promoter hypermethylation in immortalized normal esophageal epithelial cell lines, ESCC cell lines and matched tumor tissues and nontumor tissues by methylation-specific PCR. M: methylated allele; U: unmethylated alleles. Bars, SD; ∗∗, P < 0.01; ∗∗∗, P < 0.001, the student's t test.

Fig. 6

miR-449a targets MEST to inhibit cancer invasion and metastasis. (a) MEST expression was determined by Western blotting in ESCC cells with stable overexpression or knockdown of miR-449a. (b, c) Boyden chamber and bioluminescence imaging showing the effect of miR-449a overexpression on tumor invasion and metastasis. (d) Ectopic expression of MEST abolished the suppressive effect of miR-449a on ESCC cell invasion and metastasis. (e) Diagram illustrating the site mutations introduced in the reporter plasmid containing the MEST 3′UTR. The lower panel shows the luciferase activity of ESCC cells cotransfected with miR-449a and wild type or mutant MEST 3′UTR. (f) Experimental scheme and bioluminescence imaging showing the effect of systemically delivered 20 μg miR-449a oligonucleotide on tumor metastasis (n = 6/group). (g) H&E staining showing the metastatic niches in the lung tissues (n = 3). (h) The effect of miR-449a delivery on tumor metastasis was monitored in the multiorgan metastasis model (n = 6/group). (i) H&E staining showing the metastasis niches in multiple organs, including the lung, liver, kidney and spleen (n = 3). Bars, SD; ∗∗, P < 0.01; ∗∗∗, P < 0.001, the student's t test.

Fig. 7

Identification of G699-0288 as a lead compound to disrupt the MEST-PURA interaction and cancer metastasis. (a) Diagram showing the approach to screen candidate compounds targeting MEST to suppress cancer metastasis. (b) The top 20 compounds were used to treat KYSE410-Luc-I6 cells, and the inhibitory effect on cell invasion was examined by the Boyden chamber assay. (c) SPR analysis revealing the binding of G699-0288 to the MEST protein. (d) The workflow of the modified ELISA screening system. In brief, the interaction between the recombinant proteins PURA-GST and MEST-His in the presence or absence of inhibitors was analyzed. (e) The G699-0288 compound, but not the inhibitor#3, significantly blocked the binding of MEST with PURA protein. (f) The structure of the G699-0288 compound. (g) A co-IP assay was performed to determine the interaction of MEST and PURA in the absence or presence of inhibitors. (h) Boyden chamber assay showing the effect of G699-0288 on ESCC cell invasion. (i) Bioluminescence imaging and quantification of lung metastasis in mice intravenously injected with KYSE150-Luc-LM5 cells and treated with G699-0288 (5 mg/kg). (j) Experimental scheme of the PDX metastasis model and microscopic images of the dissected lungs from the mice treated with G699-0288 and vehicle. (k) Experimental scheme illustrating the animal experimental design. Bioluminescence imaging and quantification of lung metastasis when wild type or mutant MEST protein was re-overexpressed in MEST-deficient cells in the presence or absence of G699-0288 (5 mg/kg) treatment (n = 6/group) (l). Schematic diagram summarizing the MEST-PURA as a key mechanism of cancer metastasis, which can be blocked by G699-0288. Bars, SD; ∗∗, P < 0.01; ∗∗∗, P < 0.001, the student's t test.