KCNQ potassium channels modulate Wnt activity in gastro-oesophageal adenocarcinomas

Abstract

Voltage-sensitive potassium channels play an important role in controlling membrane potential and ionic homeostasis in the gut and have been implicated in gastrointestinal (GI) cancers. Through large-scale analysis of 897 patients with gastro-oesophageal adenocarcinomas (GOAs) coupled with in vitro models, we find KCNQ family genes are mutated in ∼30% of patients, and play therapeutically targetable roles in GOA cancer growth. KCNQ1 and KCNQ3 mediate the WNT pathway and MYC to increase proliferation through resultant effects on cadherin junctions. This also highlights novel roles of KCNQ3 in non-excitable tissues. We also discover that activity of KCNQ3 sensitises cancer cells to existing potassium channel inhibitors and that inhibition of KCNQ activity reduces proliferation of GOA cancer cells. These findings reveal a novel and exploitable role of potassium channels in the advancement of human cancer, and highlight that supplemental treatments for GOAs may exist through KCNQ inhibitors.

Figure 1.. KCNQ genes are highly altered in gastro-oesophageal adenocarcinomas.

(A) Oncoprint of genetic alterations in the KCNQ/E gene family, and a set of known gastro-oesophageal adenocarcinoma driver genes. * represents FDR Q-value < 0.05 co-occurrence of alterations. (B) Chromosome 8q24.12-23 showing gene density, and identified genes that are recurrently amplified. * represents genes that are known drivers in human cancer.

Figure S1.. KCNQ genes are highly genetically altered in gastro-oesophageal adenocarcinomas (GOAs), related to Fig 1.

(A) Missense mutations in KCNQ genes in GOA cancers by tumour stage. (B) Percentage of types of genetic alterations in KCNQ/KCNE genes per patient (left) and per tissue (right). (C) dN/dS values for KCNQ genes in GOAs.

Figure S2.. Mutations in KCNQ genes are under selection, related to Fig 2.

(A) Schematic of the KCNQ protein topology. (B, C, D) Mutational lollipop plots for (B) KCNQ2, (C) KCNQ4, and (D) KCNQ5; purple highlights represent significant clusters (Q < 0.05) as defined by the non-random mutational clustering algorithm. (E) Root mean square deviation plots for simulations of KCNQ1-5 homology models in a POPC bilayer using 100 ns of atomistic molecular dynamics. (F) Structure of single subunit of KCNQ1-5 colours by mutational frequency in gastro-oesophageal adenocarcinomas. (G) Structure and 3D clusters in KCNQ1 (left) and KCNQ3 (right); P-value was determined through permutation-based statistical test (see the Materials and Methods section). (H) Results of sidekick simulations for (left) KCNQ1 and (right) KCNQ3.

Figure 2.. Mutations in KCNQ genes in gastro-oesophageal adenocarcinomas alter channel function.

(A) Mutational clustering for KCNQ1 (top) and KCNQ3 (bottom), coloured lines represent observed versus expected dN/dS ratio, and purple highlights represent statistically significant (non-random mutational clustering Q-value < 0.05) clusters of mutations. (B) Rendering of the pore region of KCNQ1. (Left) Mutations modelled are highlighted. (Right) HOLE analysis of the pore region of KCNQ1 WT (black) and mutations in cluster 1.2 inset is the smallest distance in the channel gate for each mutation.

Figure 3.. KCNQ expression alters the gastro-oesophageal adenocarcinoma cell phenotype.

(A) RNA-seq expression for KCNQ1 and KCNQ3 in our patient cohorts. (B) Multivariate Cox regression analysis of KCNQ1 in gastro-oesophageal adenocarcinomas. (C) Kaplan–Meier analysis of upper and lower 50% of patients with gastric adenocarcinoma subset by KCNQ1 gene expression. (D) Relative confluence of cell growth in WT- versus KCNQ3-overexpressing (OE) OE33 and FLO1 cell lines. (E) Relative confluence of cell growth in WT versus KCNQ1 knockout (KO) OE33 and FLO1 cell lines. (F, G, H, I) Images from mouse stomach tissue. Blue represents CellTiter-Blue, red represents KCNQ1, and green represents KCNQ3. (F, G, H, I) Images shown are (F, G), normal Stomach; (H), benign adenoma; (I), metastatic adenocarcinoma. Scale bar represents 25 µm.

Figure S3.. KCNQ1 and KCNQ3 expression is linked to clinical outcome in gastro-oesophageal adenocarcinomas (GOAs) (related to Fig 3).

(A) Multivariable Cox regression analysis of KCNQ1, KCNQ3, and a series of GOA driver genes in OAC. (B) Multivariable Cox regression analysis of KCNQ1, KCNQ3, and a series of GOA driver genes in STAD. (C, D, E, F, G) Kaplan–Meier analysis for patients split by top and bottom 50% of expressers for (C) KCNQ3 in GOA, (D) KCNQ3 in OAC, (E) KCNQ3 in STAD, (F) KCNQ1 in OAC, and (G) KCNQ1 in STAD.

Figure S4.. Manipulation of KCNQ expression in vitro and assessment of KCNQ expression in mouse models, related to Fig 3.

(A) Plasmid used for KCNQ1 KO (left) Sanger sequencing confirming KCNQ1 KO in OE33 cell lines (right). (B) Plasmid used for KCNQ3 overexpression (left) and Western blot validation of KCNQ3 expression in OE33 (right). (C) Confocal microscopy of KCNQ3 OE OE33 showing the presence of a GFP tag. (D) IncuCyte assay confluence time course for KCNQ1 KO or KCNQ3 OE FLO-1 and OE33 cell lines. (E) Cell area (in pixels) for OE33 WT cells and KCNQ3 OE OE33. (F) Microarray expression levels of KCNQ and KCNE genes in stomach cancer mouse model. (G) Expression of KCNQ1 (left) and KCNQ3 (right) in patients by tumour stage (American Joint Committee on Cancer tumour staging). *p represents a t test < 0.05.

Figure 4.. KCNQ activity mediates β-catenin signalling.

(A) PROGENY pathway correlation significance with KCNQ3 RNA expression. (B) Imaging of β-catenin localisation (top—silver) and nuclear staining (bottom—blue) for WT OE33 (left) and KCNQ3 overexpression (OE) OE33 cell lines (right). (C) Enrichment of hallmark gene sets by gene set enrichment analysis (GSEA) for WT versus KCNQ3 OE OE33 cells. (D) String analysis of top five transcription factors identified by GSEA TFT gene sets. Genes enriched for GO biological processes identify β-catenin signalling (Q < 0.001). (E) GSEA enrichment plot for GO biological processes: Regulation of Establishment of Planar Cell Polarity applied to WT OE33 cell lines versus KCNQ3 OE OE33. (F) GO biological process enrichment significance for significantly (Q < 0.05) differentially expressed genes in KCNQ3 WT versus KCNQ3 OE OE33. (G) Heatmap of genes involved in the most enriched GO molecular function (cadherin binding) for KCNQ3 WT versus OE OE33. (H) Imaging of E-cadherin (red) and N-cadherin (orange) in WT OE33 (left) versus KCNQ3 OE OE33 (right), and form factor calculation for microscopy images, N = 1,371.

Figure S5.. KCNQ gene expression impacts WNT signalling, related to Fig 4.

(A) −log₁₀ (P-value) for PROGENY pathways correlated against KCNQ1 expression in human gastro-oesophageal adenocarcinomas. (B) Clustering of top and bottom KCNQ3 (left) and KCNQ1 (right) 25 expressing patients with gastro-oesophageal adenocarcinomas by WNT pathway genes. *p represents the permutation clustering test (see the Materials and Methods section). (C) Transcription factor enrichment scores for KCNQ3 OE versus WT OE33. Enrichment was determined by gene set enrichment analysis against the TFT gene set. (D) Top GO biological processes enriched for differentially expressed genes in KCNQ3 OE versus WT OE33. (E) Top GO molecular function enriched pathways for differentially expressed genes in KCNQ3 OE versus WT OE33.

Figure 5.. Patients high and low for KCNQ3 alter similar signalling pathways to OE of KCNQ3 in OE33.

(A) Venn diagram of overlap between enriched pathways in cell lines (KCNQ3 WT versus KCNQ3 OE OE33) and patients (highest 25 versus lowest 25 patients by KCNQ3 expression in OAC). Overlap p represents using cell line pathways as custom set in g-profiler. (B) −log₁₀ Q-values for the top 10 overlapping pathways between cell lines and patients.

Figure S6.. OE33 cell lines modulated for KCNQ1 accurately reflect patients, related to Fig 5.

(A) Venn diagram of overlap between enriched pathways in cell lines (KCNQ1 WT versus KCNQ1 KO OE33) and patients (highest 25 versus lowest 25 patients by KCNQ1 expression in OAC). Overlap p represents using cell line pathways as custom set in g-profiler. (B) −log₁₀ Q-values for the top 10 overlapping pathways between cell lines and patients.

Figure 6.. KCNQ channels are potential therapeutic targets in gastro-oesophageal adenocarcinoma.

(A) Relative confluence plots for OE33 (left) and FLO1 (right) cell lines exposed to linopirdine (purple) or amitriptyline (orange). Cell lines are either WT for KCNQ3 (light) or KCNQ3 overexpression (OE) (dark). (B) Venn diagram of overlapping GO biological processes enriched in ctrl versus linopirdine-exposed KCNQ3 OE OE33 cells (purple) and ctrl versus amitriptyline-exposed KCNQ3 OE OE33 cells. (C) REVIGO clustered GO biological process terms associated with overlap between ctrl versus linopirdine- and amitriptyline-exposed KCNQ3 OE OE33 cells. (D) Fold change of MYC, CCND1, CDH1, and E2F1 in ctrl versus linopirdine-exposed (purple) and ctrl versus amitriptyline-exposed (orange) KCNQ3 OE OE33 cells. * represents Q-value < 0.05. (E) Speculative mechanism of KCNQ3 activity on gastro-oesophageal adenocarcinoma cells, and their inhibition by linopirdine.

Figure S7.. Timecourse confluence plots and pathway enrichment for linopirdine- and amitriptyline-treated KCNQ3 OE OE33, related to Fig 6.

(A) Confluence plots for OE33 and FLO-1 cell lines WT and KCNQ3 OE exposed to linopirdine and amitriptyline. (B) GO molecular function −log₁₀ (P-values) for genes differentially expressed in ctrl versus linopirdine-treated KCNQ3 OE OE33. (C) GO molecular function −log₁₀ (P-values) for genes differentially expressed in ctrl versus amitriptyline-treated KCNQ3 OE OE33.