Loss of STIM2 in colorectal cancer drives growth and metastasis through metabolic reprogramming and PERK-ATF4 endoplasmic reticulum stress pathway

Abstract

The endoplasmic reticulum (ER) stores large amounts of calcium (Ca²⁺), and the controlled release of ER Ca²⁺ regulates a myriad of cellular functions. Although altered ER Ca²⁺ homeostasis is known to induce ER stress, the mechanisms by which ER Ca²⁺ imbalance activate ER stress pathways are poorly understood. Stromal-interacting molecules STIM1 and STIM2 are two structurally homologous ER-resident Ca²⁺ sensors that synergistically regulate Ca²⁺ influx into the cytosol through Orai Ca²⁺ channels for subsequent signaling to transcription and ER Ca²⁺ refilling. Here, we demonstrate that reduced STIM2, but not STIM1, in colorectal cancer (CRC) is associated with poor patient prognosis. Loss of STIM2 causes SERCA2-dependent increase in ER Ca²⁺, increased protein translation and transcriptional and metabolic rewiring supporting increased tumor size, invasion, and metastasis. Mechanistically, STIM2 loss activates cMyc and the PERK/ATF4 branch of ER stress in an Orai-independent manner. Therefore, STIM2 and PERK/ATF4 could be exploited for prognosis or in targeted therapies to inhibit CRC tumor growth and metastasis.

Keywords: ATF4; ER Ca2+; ER stress; PERK; SERCA2; STIM2; cMyc; colorectal cancer; metabolism; mitochondria.

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

(A, B) TCGA data showing (A) STIM1 and (B) STIM2 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, E) TCGA data analysis of (D) STIM1 and (E) STIM2 mRNA levels in early (stage I/II) and late (III/IV) stages of CRC progression. (F, G) Kaplan plot of STIM2 showing survival of (F) colon adenocarcinoma (COAD) and (G) rectal adenocarcinoma (READ) patients. (H, I) Kaplan plot for STIM1 showing survival of (H) COAD and (I) READ patient. Kruskal-Wallis ANOVA was performed to test single variables between the two groups. *p<0.05, **p<0.01, and ***p<0.001.

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

(A) Schematic representation of intrasplenic injection of luciferase tagged HCT116, STIM1 KO #01, STIM2 KO #15, and DKO #22 clones of HCT116 cells. The cells were injected at 5 ×10⁵ cells/mice into the spleen of male NOD-SCID mice. (B, C) Male NOD-SCID mice injected with luciferase tagged HCT116, STIM1 KO #01, STIM2 KO #15, and DKO #22 clones of HCT116 cells (B) representative image with three mice per group, and (C) quantification of whole-body luminance (Scale bar 2.0 in). (D, E) Six weeks after intrasplenic injection of HCT116, STIM1 KO #01, STIM2 KO #15, and DKO #22 clones of HCT116 cells, the mice were sacrificed, and the primary tumor at the injection site was imaged, (D) representative image of primary tumors, and (E) quantification of primary tumor weight. (Scale bar 0.5 cm). (F-I) Post intrasplenic injection of HCT116, STIM1 KO #01, STIM2 KO #15, and DKO #22 clones of HCT116 cells, organs of intrasplenic injected mice were harvested, and luminance was measured (F) representative image of liver, lung, and colon showing metastasis, and quantification of total luminance in (G) liver (Scale bar 1 cm), (H) lung (Scale bar 0.5 cm), and (I) colon (Scale bar 0.5 cm). (J) Survival of NOD-SCID mice injected with HCT116, STIM1 KO #01, STIM2 KO #15, and STIM DKO #22 clones of HCT116 cells (p< 0.0003). ANOVA followed with a post hoc Tukey test to compare between groups. *p<0.05, **p<0.01 and ***p<0.001

Figure 3.. Loss of STIM2 enhances CRC proliferation and metastasis

(A, B) Spheroid formation assay showing (A) representative image of spheroid formed by HCT116, and clones of STIM1 KO, STIM2 KO, and DKO in HCT116 cells, (B) quantification of the spheroid area at 0 and 72 hr. (Scale bar 300 μm) (C) Quantification of the spheroid area of DLD1 and clones of STIM1 KO, STIM2 KO, and DKO in DLD1 cells. (D, E) Quantification of (D) normalized migration and (E) percent invasion of HCT116 and clones of STIM1 KO, STIM2 KO, and DKO in HCT116 cells using Boyden chamber assay. (F, G) Quantification of (F) normalized migration and (G) percent invasion of DLD1 and clones of STIM1 KO, STIM2 KO, and DKO in DLD1cells using Boyden chamber assay. (H, I) RT-qPCR analysis of mRNA levels of (H) MMP16 and (I) MMP14 in HCT116, STIM1 KO, STIM2 KO, and DKO clones of HCT116 cells. (J, K) RT-qPCR analysis of mRNA levels of (J) MMP16 and (K) MMP14 in DLD1 and clones of STIM1 KO, STIM2 KO, and DKO clones of DLD1 cells. (L, M) Quantification of (L) normalized migration and (M) percent invasion of HCT116 and clones of Orai1 KO, and Orai TKO in HCT116 cells using Boyden chamber assay. ANOVA followed with a post hoc Tukey test except for F-M where the paired t-test was used to compare between groups. *p<0.05, **p<0.01 and ***p<0.001

Figure 4.. STIM2 regulates mitochondrial biogenesis and metabolism in CRC cells

(A) Heatmap showing significantly increased levels of glycolysis pathway metabolites in HCT116 STIM2 KO #15 and HCT116 DKO #22 compared to HCT116, and HCT116 STIM1 KO #01STIM1 KO #01 cells. (B) Extracellular acidification rate (ECAR) of HCT116 and clones of STIM1 KO, STIM2 KO, and DKO in HCT116 cells. (C, D) Measurement of (C) glucose consumption and (D) lactate generation in HCT116 and clones of STIM1 KO, STIM2 KO, and DKO in HCT116 cells. (E) Oxygen consumption rate (OCR) in HCT116 and STIM1 KO #01, STIM2 KO #15, and DKO #22 clones of HCT116 cells. (F) Heatmap showing significantly increased levels of citric acid cycle metabolites in HCT116 STIM2 KO #15 and HCT116 DKO #22 compared to HCT116, and HCT116 STIM1 KO #01 cells. (G, H) Measurement of GLUT1 protein level (G) representative western blot probed with anti-GLUT1 and GAPDH antibody and (H) densitometric analysis of GLUT1 normalized to GAPDH in HCT116 and clones of STIM1 KO, STIM2 KO, and DKO in HCT116 cells. (I-M) Ultrastructure of mitochondria analyzed by electron microscopy, (I) representative electron micrograph showing mitochondrial structure, measurement of (J) area, (K) circularity, (L) perimeter, and (M) mitochondrial number in HCT116 and STIM1 KO #01, STIM2 KO #15, and DKO #22 colons of HCT116 cells. (N, O) Flow cytometric analysis of MitoTracker intensity in (N) HCT116 and clones of HCT116 STIM1 KO, HCT116 STIM2 KO, and HCT116 DKO, and (O) DLD1, and clones of DLD1 STIM1 KO, DLD1 STIM2 KO, and DLD1 DKO cells. (P, Q) RT-qPCR analysis of mitochondrial DNA (mtDNA) in (P) HCT116 and clones of HCT116 STIM1 KO, HCT116 STIM2 KO, and HCT116 DKO, and (Q) DLD1, and clones of DLD1 STIM1 KO, DLD1 STIM2 KO, and DLD1 DKO cells. Statistical significance was calculated using ANOVA followed by a post hoc Tukey test to compare multiple groups except for C, D, H, P and Q, where the paired t-test was used to compare between groups. *p<0.05, **p<0.01 and ***p<0.001

Figure 5.. Loss of STIM2 leads to transcriptional reprogramming of CRC cells

(A-D) Volcano plot showing a comparison of differentially expressed genes between (A) STIM2 KO #15 and HCT116, (B) STIM2 KO #31 and HCT116, (C) DKO #22 and HCT116, and (D) DKO #34 and HCT116. The X-axis represents log₂ fold change and Y axis −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 represented in black and upregulated in blue or green. (E, F) The pathway analysis showing plot between normalized enrichment score and nominal p-value. The positive enrichment score represents upregulated pathway, and negative enrichment shows downregulated pathways in (E) STIM2KO #15 vs. HCT116 and (F) DKO #22 vs. HCT116. (G, H) GSEA analysis between (G) HCT116 and STIM2 KO #15 and (H) HCT116 and DKO #22 show a positive correlation in the enrichment of hallmark of Myc target version 1 genes. (I, J) GSEA analysis between (I) HCT116 and STIM2 KO #15 and (J) HCT116 and DKO #22 show a positive correlation in the enrichment of hallmark of Myc target version 2 genes. (K, L) GSEA analysis between (K) HCT116 and STIM2 KO #15 and (L) HCT116 and DKO #22 show a positive correlation in the enrichment of hallmark of unfolded protein response (UPR) genes. (M, N) GSEA analysis between (M) HCT116 and STIM2 KO #15 and (N) HCT116 and DKO #22 shows a positive correlation in the enrichment of hallmark of epithelial to mesenchymal transition (EMT) genes.

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

(A, B) ER Ca²⁺ measurement using genetically encoded R-CEPIA1_ER. The ER Ca²⁺ depletion was stimulated with 300 μM ATP in 0 mM Ca²⁺ for 10 minutes. The graph represents the mean ± S.E.M. of (A) HCT116, STIM1 KO #01, STIM2 KO #15, and DKO #22 clone of HCT116, and (B) HCT116, STIM1 KO #02, STIM2 KO #31, DKO #34 clones of HCT116 cells. (C, D) Quantification of (C) Basal ER Ca²⁺, (D) ER Ca²⁺ depletion rate 3 min post stimulation in 0 Ca²⁺, HCT116 (n=350), STIM1 KO #01 (n=342), STIM1 KO#02 (n=210), STIM2 KO #15 (n=150), STIM2 KO #31 (n=170). DKO #22 (n=250), and DKO #34 (n=150) clones of HCT116. (E, F) ER Ca²⁺ measurement using genetically encoded R-CEPIA_ER. The ER Ca²⁺ depletion was stimulated cells were stimulated with 300 μM ATP in 0 mM Ca²⁺ for 10 minutes. The graph represents the mean ± S.E.M. of (E) DLD1, STIM1 KO #03, STIM2 KO #04, and DKO #01 clone of DLD1, and (F) DLD1, STIM1 KO #05, STIM2 KO #09, DKO #07 clones of DLD1 cells. (G, H) Quantification of (G) Basal ER Ca²⁺, (H) ER Ca²⁺ depletion rate 3 min post stimulation in 0 Ca²⁺, in DLD1 (n=140), STIM1 KO #03 (n=54), STIM1 KO#05 (n=85), STIM2 KO #04 (n=50), STIM2 KO #09 (n=40). DKO #01 (n=35), and DKO #07 (n=40) clones of DLD1. (I-K) RT-qPCR analysis of mRNA levels of (I) SERCA1, (J) SERCA2, and (K) SERCA3 in HCT116, STIM1 KO, STIM2 KO, and DKO in HCT116 cells. (L-O) Measurement of SERCA2 protein level by probing the western blots with anti-SERCA2 and GAPDH antibody (L and N) representative blot probed and (M, and O) densitometric analysis of SERCA2 normalized to GAPDH in HCT116 and clones of STIM1 KO, STIM2 KO, and DKO in HCT116 cells. (P, Q) Measurement of SERCA2 protein level by probing the western blots with anti-SERCA2 and GAPDH antibody (P) representative blot, and (Q) densitometric analysis of SERCA2 normalized to GAPDH in DLD1 and clones of STIM1 KO, STIM2 KO, and DKO in DLD1 cells. All experiments were performed ≥three times with similar results. Statistical significance was calculated using one-way ANOVA followed by a post hoc Tukey test except for I-Q, where the paired t-test was used to compare between groups. *p<0.05, **p<0.01, ***p<0.001

Figure 7.. STIM2 deletion promotes cMyc stabilization and increases PPARɣ levels

(A) Representative western blot probed with anti- phospho-cMyc, cMyc, phospho-ERK, ERK, and GAPDH. (B, C) Densitometric analysis of (B) phospho-cMyc and (C) phospho-ERK in HCT116, STIM1 KO #01, STIM2 KO #15, DKO #22 clone in HCT116 and DLD1, STIM1 KO #03, STIM2 KO #04, DKO #01 clones in DLD1 cells. (D-G) PGC1α levels determined by western blot (D, F) blot probed with anti- PGC1α and GAPDH antibody and (E, G) densitometric analysis of PGC1α levels in HCT116, DLD1, STIM1 KO, STIM2 KO, and DKO clones of HCT116 and DLD1 cells. (H, I) PPARɣ levels quantified using (H) western blot probed with anti-PPARɣ and GAPDH antibody, and (I) densitometric analysis of PPARɣ levels in HCT116, DLD1, STIM1 KO, STIM2 KO, and DKO clones of HCT116 and DLD1 cells. All experiments were performed ≥three times with similar results. Statistical significance was calculated using paired t-test. *p<0.05, **p<0.01, ***p<0.001

Figure 8.. Loss of STIM2 activates the ATF4-dependent ER stress pathway

(A, B) Quantification of BiP protein levels using (A) western blot probed with anti-Bip and GAPDH antibody, and (B) densitometric analysis in HCT116, DLD1, STIM1 KO, STIM2 KO, and DKO clones of HCT116 and DLD1cells. (C, D) Quantification of ATF4 protein levels using (C) western blot probed with anti-ATF4 and GAPDH antibody, and (D) densitometric analysis in HCT116, STIM1 KO, STIM2 KO, and DKO clones of HCT116. (E, F) Quantification of IRE1α protein levels using (E) western blot probed with anti- IRE1α and GAPDH antibody, and (F) densitometric analysis in HCT116, STIM1 KO, STIM2 KO, and DKO clones of HCT116. (G, H) Quantification of ATF6 protein levels using (G) western blot probed with anti- ATF6 and GAPDH antibody, and (H) densitometric analysis in HCT116, STIM1 KO, STIM2 KO, and DKO clones of HCT116. (I, J) Quantification of phospho-eIF2α protein levels using (I) western blot probed with anti-phospho-eIF2α, eIF2α, and GAPDH antibody, and (J) densitometric analysis in HCT116, STIM1 KO, STIM2 KO, and DKO clones of HCT116. (K, L) Quantification of phospho-PERK protein levels using (K) western blot probed with anti-phospho-PERK, PERK, and GAPDH antibody, and (L) densitometric analysis in HCT116, STIM1 KO, STIM2 KO, and DKO clones of HCT116. All experiments were performed ≥three times with similar results. Statistical significance was calculated using paired t-test . *p<0.05, **p<0.01, ***p<0.001