Loss of STIM2, but not of STIM1, drives colorectal cancer metastasis through metabolic reprogramming and the ATF4 ER stress pathway

Abstract

The large amounts of calcium (Ca²⁺) stored in the endoplasmic reticulum (ER) and the controlled release of this Ca²⁺ store into the cytosol regulate many cellular functions, and altered ER Ca²⁺ homeostasis induces ER stress. Stromal-interacting molecules 1 and 2 (STIM1/2) are homologous ER-resident Ca²⁺ sensors that synergistically activate cytosolic Ca²⁺ influx through Orai channels to promote Ca²⁺-dependent changes in gene expression and ER Ca²⁺ refilling. Here, we demonstrated that reduced abundance of STIM2, but not that of STIM1, was associated with poor prognosis in colorectal cancer (CRC). STIM2-deficient CRC cells showed enhanced ER Ca²⁺ content in a manner dependent on the ER Ca²⁺ pump SERCA2, increased expression of genes associated with protein translation, and transcriptional and metabolic rewiring. STIM2 deficiency in CRC xenografts led to increased tumor size, invasion, and metastasis. STIM2 loss activated the expression of genes involved in ER stress responses in a manner dependent on the chaperone BiP and the transcription factor ATF4 and independent of Orai channels. These results suggest that loss of STIM2 may inform CRC prognosis.

Figure 1.. Low STIM2 expression in colorectal cancer is associated with poor patient prognosis

(A and B) TCGA data showing STIM1 (A) and STIM2 (B) mRNA levels in the unmatched tumor and adjacent normal tissue. (C) Analysis of STIM1 and STIM2 mRNA levels in matched tumor (n=20) and adjacent normal tissue (n=20) of CRC patients by RT-qPCR. The patient samples were obtained from Pennsylvania State University Hershey Medical Center. (D and E) Kaplan plot correlating STIM1 mRNA to survival of colon adenocarcinoma (COAD) (D) and rectal adenocarcinoma (READ) (E) patients. (F and G) Kaplan plot correlating STIM2 mRNA expression to survival of COAD (F) and READ (G) patients. *p<0.05, **p<0.01, and ***p<0.001 by Kruskal-Wallis ANOVA.

Figure 2.. STIM2 loss enhances tumor growth and metastasis in mice

(A) Schematic representation of intrasplenic injection of luciferase tagged parental HCT116, STIM1 KO #01, STIM2 KO #15, or STIM DKO #22 HCT116 clones. The spleens of male NOD-SCID mice were injected with 5 × 10⁵ cells/mice. (B and C) Male NOD-SCID mice were injected with luciferase tagged parental HCT116 (n=28) or STIM1 KO #01 (n=27), STIM2 KO #15 (n=12), or STIM DKO #22 (n=10) HCT116 clones. Representative luciferase luminescence images of three mice per group (B) and quantification of whole-body luminance (C) are shown. Scale bar, 2.0 inches (D and E) Six weeks after intrasplenic injection of parental HCT116 cells (n=28) or STIM1 KO #01 (n=27), STIM2 KO #15 (n=12) or STIM DKO #22 (n=10) HCT116 clones, mice were sacrificed and the primary tumor at the injection site was imaged. Representative image of primary tumors (D) and quantification of primary tumor weight (E) are shown. Scale bar, 0.5 cm. (F to I) After intrasplenic injection of parental HCT116 (n=28) cells, or STIM1 KO #01 (n=27), STIM2 KO #15 (n=12), or STIM DKO #22 (n=10) HCT116 clones, organs were harvested, and luminance was measured. Representative images of liver, lung, and colon showing metastasis (F) and quantification of total luminance in liver (G), lung (H), and colon (I) are shown. Scale bars, 1 cm for liver and colon and 0.5 cm for lung. (J) Survival of NOD-SCID mice injected with parental HCT116 cells (n=30) or STIM1 KO #01 (30), STIM2 KO #15 (30), or STIM DKO #22 (15) HCT116 clones (p<0.0003). *p<0.05, **p<0.01 and ***p<0.001 by paired t-test.

Figure 3.. Loss of STIM2 enhances CRC migration, invasion, and metastasis

(A and B) Representative images of spheroids formed by parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones (A), and quantification of the spheroid area at 0 and 72 hr (B). Scale bar 300 μm, n=3 biological replicates per group. (C and D) Quantification of normalized migration (C) and percent invasion (D) of parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones using Boyden chamber assays. n=3 biological replicates per group. (E and F) GSEA analysis reveals enrichment of the hallmark epithelial–mesenchymal transition (EMT) gene set in STIM2 KO #15 HCT116 clone (E) and STIM DKO #22 HCT116 clone (F) compared to parental HCT116 cells. (G) RT-qPCR analysis of MMP16 mRNA levels in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n=3 biological replicates per group. (H) RT-qPCR analysis of MMP16 mRNA levels upon MMP16 silencing in parental HCT116 or STIM2 KO HCT116 clones. n=3 biological replicates per group. (I and J) Quantification of normalized migration (I) and percent invasion (J) of parental HCT116 or STIM2 KO HCT116 clones upon silencing of MMP16. n=4 biological replicates per group. (K and L) Quantification of normalized migration (K) and percent invasion (L) of parental, Orai1 KO, or Orai TKO HCT116 clones using Boyden chamber assay. n=4 biological replicates per group. *p<0.05, **p<0.01, and ***p<0.001 by ANOVA followed with a post hoc Tukey test except for paired t-test for (G to L).

Figure 4.. Loss of STIM2 leads to transcriptional reprogramming of CRC cells

(A to D) Volcano plot showing a comparison of differentially expressed genes between STIM2 KO #15 and parental HCT116 cells (A), STIM2 KO #31 and parental HCT116 cells (B), STIM DKO #22 and parental HCT116 cells (C), and STIM DKO #34 and parental HCT116 cells (D). The X-axis shows log₂ fold changes, and the Y-axis shows -log₁₀ (P-value). The threshold in the plot corresponds to P- value <0.05 and log 2-fold change <−1.0 or > 1.0. The significantly downregulated genes are black and upregulated genes are blue or green. n= 3 biological replicates per group. (E and F) Pathway analysis showing plot between normalized enrichment score and nominal p-value. The positive enrichment score represents upregulated pathways, and negative enrichment shows downregulated pathways in STIM2 KO #15 clone compared to parental HCT116 cells (E) and in STIM DKO #22 clone compared to parental HCT116 cells (F). (G and H) GSEA analysis reveals enrichment of the hallmark of Myc target version 1 gene set in STIM2 KO #15 HCT116 clone (G) and STIM DKO #22 HCT116 clone (H) compared to parental HCT116 cells. (I and J) GSEA analysis reveals enrichment of the hallmark of Myc target version 2 gene set in STIM2 KO #15 HCT116 clone (I) and STIM DKO #22 HCT116 clone (J) compared to parental HCT116 cells. (K and L) GSEA analysis reveals enrichment of the hallmark of unfolded protein response (UPR) gene set in STIM2 KO #15 HCT116 clone (K) and STIM DKO #22 HCT116 clone (L) compared to parental HCT116 cells.

Figure 5.. STIM2 regulates glycolysis in CRC cells

(A) Schematics showing metabolites produced during glycolysis and the TCA cycle. (B) Heatmap for glycolysis pathway metabolites in parental HCT116 cells or STIM1 KO #01, STIM2 KO #15, or STIM DKO #22 HCT116 clones. n=3 biological replicates per group. (C) Extracellular acidification rate (ECAR) of parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n= 3 biological replicates per group. (D and E) Measurement of glucose consumption (D) and lactate generation (E) in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n=4 biological replicates per group. (F and G) Representative Western blot probed for GLUT1 and GAPDH (F) and densitometric analysis of GLUT1 normalized to GAPDH (G) in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n=3 biological replicates per group. (H) TCGA data obtained from UALCAN analysis portal showing GLUT1 protein levels in the tumor and unmatched adjacent normal tissue of CRC patients. (p<0.0271 by t-test considering unequal variance). (I and J) Representative Western blot for GLUT1, tubulin, and GAPDH (I) and densitometric analysis of GLUT1 normalized to GAPDH (J) in parental HCT116 cells and STIM2 KO HCT116 clones with GLUT1 silencing. n=3 biological replicates per group. (K) Quantification of the spheroid area in parental HCT116 cells and STIM2 KO HCT116 clones with GLUT1 silencing compared to control. n=3 biological replicates per group. (L) Heatmap for citric acid cycle metabolites in parental HCT116 cells or STIM1 KO #01, STIM2 KO #15, or STIM DKO #22 HCT116 clones. n=3 biological replicates per group. (M) Oxygen consumption rate (OCR) in parental HCT116 cells or STIM1 KO #01, STIM2 KO #15, or STIM DKO #22 HCT116 clones. n=3 biological replicates per group. *p<0.05, **p<0.01, and ***p<0.001 by ANOVA followed by a post hoc Tukey test except for paired t-test in (G), (J), and (K).

Figure 6.. Loss of STIM2 leads to increased basal ER Ca²⁺

(A) ER Ca²⁺ measurement using genetically encoded R-CEPIA1_ER. ER Ca²⁺ depletion was induced with 300 μM ATP in 0 mM Ca²⁺ for 10 minutes. Means ± S.E.M. are shown for parental HCT116 cells or STIM1 KO #01, STIM2 KO #15, or STIM DKO #22 HCT116 clones. (B) Quantification of basal ER Ca²⁺ in parental HCT116 cells (n=350), or STIM1 KO #01 (n=342), STIM1 KO#02 (n=210), STIM2 KO #15 (n=150), STIM2 KO #31 (n=170), STIM DKO #22 (n=250), or STIM DKO #34 (n=150) HCT116 clones. (C to E) RT-qPCR analysis of mRNA levels of SERCA1 (C), SERCA2 (D), and SERCA3 (E) in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n=3 biological replicates per group. (F to I) Representative Western blot for SERCA2 and GAPDH (F and H) and densitometric analysis of SERCA2 normalized to GAPDH (G and I) in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n=5 biological replicates per group. (J) TCGA data obtained from UALCAN analysis portal, showing SERCA2 protein levels in the tumor and unmatched adjacent normal tissue of CRC patients. (p<0.0000106 by t-test considering unequal variance) *p<0.05, **p<0.01, ***p<0.001 by paired t-test except for one-way ANOVA followed by a post hoc Tukey test for (B).

Figure 7.. Loss of STIM2 activates the ATF4-dependent ER stress pathway

(A and B) Representative Western blot for BiP and GAPDH (A) and densitometric analysis of BiP normalized to GAPDH (B) in parental HCT116 or DLD1 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 and DLD1 clones. n=4 biological replicates per group. (C and D) Representative Western blot for ATF4 and GAPDH (C) and densitometric analysis of ATF4 normalized to GAPDH (D) in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n=4 biological replicates per group. (E and F) Representative Western blot for IRE1α and GAPDH (E) and densitometric analysis of IRE1α normalized to GAPDH (F) in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n=3 biological replicates per group. (G and H) Representative Western blot for ATF6 and GAPDH (G) and densitometric analysis of ATF6 normalized to GAPDH (H) in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n=3 biological replicates per group. (I and J) Representative Western blot for phospho-eIF2α, eIF2α, and GAPDH (I) and densitometric analysis of phospho-eIF2α normalized to eIF2α (J) in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. n=3 biological replicates per group. (K and L) Representative Western blot for phospho-PERK, PERK, and GAPDH (K) and densitometric analysis of phospho-PERK normalized to PERK (L) in parental HCT116 cells or STIM1 KO, STIM2 KO, or STIM DKO HCT116 clones. (M and N) Representative Western blot for ATF4, tubulin, and GAPDH (M) and densitometric analysis of ATF4 normalized to GAPDH (N) in parental HCT116 cells and STIM2 KO HCT116 clones with ATF4 silencing. n=3 biological replicates per group. (O to Q) Measurement of spheroid size (O), migration (P), and invasion (Q) of parental HCT116 cells and STIM2 KO HCT116 clones with ATF4 silencing. n=4 biological replicates per group. (R and S) TCGA data obtained from UALCAN analysis portal showing ATF4 mRNA levels (p< 0.00000000000375 by t-test considering unequal variance) (R) and BiP protein levels (p<0.0000000000216 by t-test considering unequal variance) (S) in the tumor and unmatched adjacent normal tissue of COAD patients. (T) Kaplan plot correlating ATF4 mRNA expression to survival of COAD patients. (U) Loss of STIM2 leads to increased ER Ca²⁺ though increased SERCA2 levels, causing ATF4-dependent ER stress, which leads to transcriptional reprogramming and metabolic transformation to increase tumorigenesis and metastasis.