The AHR target gene scinderin activates the WNT pathway by facilitating the nuclear translocation of β-catenin

ABSTRACT

The ligand-activated transcription factor aryl hydrocarbon receptor (AHR) regulates cellular detoxification, proliferation and immune evasion in a range of cell types and tissues, including cancer cells. In this study, we used RNA-sequencing to identify the signature of the AHR target genes regulated by the pollutant 2,3,7,8-tetrachlorodibenzodioxin (TCDD) and the endogenous ligand kynurenine (Kyn), a tryptophan-derived metabolite. This approach identified a signature of six genes (CYP1A1, ALDH1A3, ABCG2, ADGRF1 and SCIN) as commonly activated by endogenous or exogenous ligands of AHR in multiple colon cancer cell lines. Among these, the actin-severing protein scinderin (SCIN) was necessary for cell proliferation; SCIN downregulation limited cell proliferation and its expression increased it. SCIN expression was elevated in a subset of colon cancer patient samples, which also contained elevated β-catenin levels. Remarkably, SCIN expression promoted nuclear translocation of β-catenin and activates the WNT pathway. Our study identifies a new mechanism for adhesion-mediated signaling in which SCIN, likely via its ability to alter the actin cytoskeleton, facilitates the nuclear translocation of β-catenin.

This article has an associated First Person interview with the first authors of the paper.

Summary: The transcription factor AHR drives the expression of the actin-severing protein scinderin, which, in turn, promotes nuclear translocation of β-catenin, suggesting a novel mechanism by which AHR promotes growth of colon cancer cells.

INTRODUCTION

Aryl hydrocarbon receptor (AHR) is one of the most well-studied sensors for environmental toxins, including polyaromatic hydrocarbons and other xenobiotics such as 2,3,7,8-tetrachlorodibenzodioxin (TCDD) (Denison and Nagy, 2003). These small molecules function as ligands for AHR, driving its nuclear translocation and the transcriptional activation of AHR target genes. In addition, several endogenous small molecules, such as the tryptophan-derived metabolites kynurenine (Kyn) and 6-formylindolo(3,2-b)carbazole (FICZ), and the plant-derived compound quercetin also function as ligands for AHR (Denison and Nagy, 2003; Murray et al., 2014; Van der Heiden et al., 2009).

In the absence of ligands, AHR is maintained in the cytoplasm by a complex containing heat shock protein 90 (HSP90) and X-associated protein 2 (XAP2) (Feng et al., 2013; Fukunaga et al., 1995). Binding of AHR to its ligands causes a conformational change that leads to the exposure of the nuclear localization signal of AHR (Ikuta et al., 2000). This results in the translocation of AHR to the nucleus where it binds to aryl hydrocarbon receptor nuclear translocator (ARNT) to form a transcriptionally active complex. This complex then binds to xenobiotic response elements (XREs) present in the promoters of its target genes, including those involved in detoxification pathways, to activate their transcription (Matthews et al., 2005; Yao and Denison, 1992).

In addition to regulating the xenobiotic response, AHR is also implicated in coordinating several critical pathways, including adaptive immunity, gut homeostasis and the secretion of cytokines (Rothhammer and Quintana, 2019). AHR expression is elevated in tumor tissues of multiple origins including lung, breast and colon (Koliopanos et al., 2002; Kubli et al., 2019; Lafita-Navarro et al., 2018; Opitz et al., 2011; Tsay et al., 2013; Venkateswaran et al., 2019; Xie et al., 2012). Nevertheless, the role of AHR in colon cancer is complex. Studies have shown that the deletion of AHR from colonic cells leads to increased proliferation and exacerbates the growth of colon cancer tumor models (Han et al., 2021; Murray et al., 2014; Narasimhan et al., 2018). In addition, ablation of AHR was shown to accelerate intestinal tumor formation by stabilizing β-catenin through mechanisms unrelated to the transcriptional activity of AHR (Ikuta et al., 2013; Kawajiri et al., 2009).

The role of AHR in cancer is likely influenced by its binding to specific ligands and the landscape of oncogenes or tumor suppressors of each cell type. Our laboratory previously found that AHR expression is transcriptionally induced by the universal oncogene MYC in colon cancer cells and that AHR supports MYC-induced proliferation (Lafita-Navarro et al., 2018) by activating the expression of genes necessary for biomass production (Lafita-Navarro et al., 2018, 2020a,b). MYC also promotes the uptake of tryptophan and its conversion into Kyn in colon cancer cells (Venkateswaran and Conacci-Sorrell, 2020; Venkateswaran et al., 2019), leading to nuclear translocation of AHR in these cells (Venkateswaran et al., 2019). AHR activation by TCDD was found to drive colon cancer cell survival and migration (Xie et al., 2012), and expression of a constitutively active AHR caused stomach cancer (Andersson et al., 2002). Additionally, the TDO2-AHR pathway is necessary for colon cancers to metastasize to the liver (Miyazaki et al., 2022), and AHR promotes chemoresistance in cancers (Li et al., 2021; Stanford et al., 2016; Wu et al., 2018). Moreover, inhibition of the Kyn–AHR axis can improve anti-cancer immunity, preventing colitis-associated cancer (Zhang et al., 2021).

The diversity of the cellular functions exhibited by AHR highlights the need to identify tumor-specific genes directly regulated by AHR and its ligands. To investigate the effects of exogenous and endogenous ligands of AHR on the transcriptional profile of colon cancer cells, we performed RNA-sequencing (RNA-seq) comparing colon cancer cells exposed to TCDD, Kyn and the AHR inhibitor CH223191. This approach identified a six-gene signature containing cytochrome P450 family 1 subfamily A member 1 (CYP1A1), aldehyde dehydrogenase 1 family member A3 (ALDH1A3), ATP-binding cassette subfamily G member 2 (ABCG2), adhesion G-protein-coupled receptor F1 (ADGRF1) and the gene encoding the actin-severing protein scinderin (SCIN) as being activated by both Kyn and TCDD in colon cancer cells. TCDD activated additional genes, including genes previously shown to function as oncogenes. Among the genes identified in our screen, SCIN was found to be necessary for cell growth due to its newly discovered ability to activate the WNT pathway.

RESULTS

Transcriptional signatures regulated by TCDD and Kyn in colon cancer cells

Elevated levels of AHR were detected in many tumors deposited in the Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/), including colon cancer (Fig. S1A), and AHR expression in the colon coincided with increased expression of the oncogene MYC (Fig. S1B). These data support findings from the previously published studies that documented elevated levels of AHR in colon cancer cell lines and patient samples (Lafita-Navarro et al., 2018; Venkateswaran and Conacci-Sorrell, 2020; Venkateswaran et al., 2019). Therefore, we selected colon cancer cells as a model system to study AHR transcriptional activity upon ligand stimulation. To select a colon cancer cell line for transcriptional profiling, we performed real-time quantitative PCR (RT-qPCR) for AHR in multiple colon cancer cell lines including DLD1, RKO, HCT115, HCT116 and HT29. DLD1 cells expressed the highest levels of AHR among the analyzed cell lines (Fig. 1A). Knocking down AHR through siRNA transfection in colon cancer cells led to a reduction in the viability of DLD1 and HT29 cells (Fig. 1B,C; Fig. S1C) but not HCT116 cells (Fig. 1D). DLD1 cells stably expressing an AHR shRNA displayed a similar decrease in proliferation as with transient knockdown (Fig. S1D). Although there was nuclear AHR, the majority of AHR was localized in the cytoplasm of DLD1 cells as demonstrated by western blotting of the nuclear and cytoplasmic fractions (Fig. S1E). Because AHR mRNA and cytoplasmic protein levels were elevated in DLD1 cells, we chose DLD1 to identify transcriptional targets of AHR that are regulated by ligand stimulation.

Fig. 1.

RNA-seq identifies transcriptional signatures regulated by TCDD, Kyn and the AHR inhibitor CH223191. (A) RT-qPCR analyses for AHR normalized to the housekeeping gene RPS18 in a panel of colon cancer cell lines. (B–D) Relative proliferation determined by absorbance at 460 nm of DLD1 (B), HT29 (C) and HCT116 (D) cells after transfection with siRNA to knockdown AHR. These experiments were repeated twice. Data show the mean±s.e.m. *P<0.0001 (two-tailed unpaired t-test). (E–G) Volcano plots showing the differentially expressed genes upon incubation with TCDD (1 nM) (E), Kyn (10 µM) (F) and CH223191 (10 µM) (G). (H–J) Gene Ontology analyses showing the major pathways that are regulated upon incubation with TCDD (H), Kyn (I) and CH223191 (J). (K) Venn diagram showing the overlap of genes regulated by TCDD, Kyn and CH223191. (L) Heatmap of genes regulated by Kyn, TCDD and CH223191 after 8 h of treatment. For all analyses, log₂FC≤0.585 and adjusted P-value≤0.05 were applied. (M) Venn diagram showing the overlap of genes regulated by TCDD and Kyn. (N) Heatmap of genes regulated by TCDD, Kyn and CH223191 after 8 h of incubation. (O) Molecular functions of the genes that are differentially expressed.

To compare the signature of genes induced by endogenous and exogenous AHR ligands, we performed RNA-seq of DLD1 cells incubated with vehicle (DMSO), TCDD or Kyn for 8 h (Fig. 1E,F,H,I; Tables S1 and S2). The 8-h incubation time was chosen to increase the probabilities of identifying early transcriptional responses to TCDD and Kyn. Using a cutoff of log₂[fold change (FC)]≤0.585 and adjusted P-value≤0.05, we identified 40 genes regulated by TCDD (39 upregulated and one downregulated) (Fig. 1E) and six genes upregulated by Kyn (Fig. 1F). The higher number of genes that were activated by TCDD could be explained by a previous finding that demonstrated TCDD binds to AHR with higher affinity than Kyn (Denison et al., 2002). It is also possible that TCDD might have the ability to maintain a sustained activation of AHR, leading to the activation of a larger number of genes. Despite the different number of target genes activated by TCDD and Kyn, Gene Ontology analyses found that the most significant pathway regulated by both ligands was xenobiotic response (Fig. 1H–J). Thus, these data suggest that both endogenous and exogenous ligands support the ability of AHR to activate genes necessary for xenobiotic response.

In addition to activating Kyn-regulated genes (Fig. 1F,I), TCDD also activated genes involved in multiple cellular processes (Fig. 1L) including cell growth, such as fibroblast growth factor-binding protein 1 (FGFBP1), prospero homeobox 1 (PROX1), proto-oncogene serine/threonine protein kinase pim-1 (PIM1) and alkaline phosphatase placental type (ALPP) (Fig. S2A). FGFBP1 is a fibroblast-secreted growth factor previously shown to play key roles in pancreatic and colorectal carcinogenesis (Tassi et al., 2006). PROX1 is a transcription factor essential during development, and increased levels of PROX1 have been observed in a variety of cancers including colon, breast and liver cancer (Elsir et al., 2012; Petrova et al., 2008). PIM1 is a serine/threonine kinase and an oncogene identified in T-cell lymphomas and pancreatic cancer shown to increase cell cycle regulation and cellular proliferation (Tursynbay et al., 2016). ALPP is an alkaline phosphatase with ectopic expression observed in pancreatic cancer (Dua et al., 2013). Our data provide molecular evidence that TCDD activates several potential oncogenes and signaling pathways likely in an AHR-dependent manner. Because TCDD is a common and persistent environmental pollutant, it is possible that exposure to TCDD might contribute to several types of cancers, as previously proposed by others (Poland et al., 1982). Indeed, pro-proliferative effects of TCDD have been shown in cell lines and rodent models of liver cancer (Poland et al., 1982).

To study the transcriptional outcomes of AHR inhibition in colon cancer cells, we performed RNA-seq in DLD1 cells treated with CH223191 (Table S3), a competitive inhibitor that prevents the binding of AHR to its ligands (Zhao et al., 2010). We have previously shown that colon cancer cells, which display elevated levels of AHR and Kyn, but not normal human colonic epithelial cells, are sensitive to CH223191 (Venkateswaran et al., 2019).

Using the same approach as described for TCDD and Kyn, we compared the transcriptional signature regulated by CH223191 using DMSO as a control. This approach identified 302 genes regulated by CH223191 (240 upregulated and 62 downregulated) (Fig. 1G,J). Unexpectedly, treatment with CH223191 caused broader effects on the transcriptional signature of DLD1 cells, leading to both activated and repressed genes. CH223191 only inhibited the expression of eight genes induced by TCDD, including three genes that were also induced by Kyn (Fig. 1K). Gene Ontology analyses revealed that CH223191 modulated the expression of genes that belong to multiple categories, including tRNA synthesis and mTOR signaling (Fig. 1J). Although CH223191 has been proposed to be a potent inhibitor of TCDD-induced nuclear translocation of AHR (Zhao et al., 2010), it is possible that CH223191 has additional effects that could account for its ability to inhibit cell growth. Further studies on the off-target effects of CH223191 are necessary to define its specificity and clinical value. Furthermore, the development of new and improved inhibitors of AHR ligand binding could lead to novel and more specific strategies to treat cancer.

Activation of AHR by TCDD and Kyn promotes the transcription of a common signature of genes

To characterize the canonical signature of ligand-activated AHR genes in colon cancer cells, we focused on the six genes co-regulated by Kyn and TCDD (Fig. 1M,N). This signature comprised an interesting and diverse group of genes: CYP1A1, ALDH1A3, ABCG2, ADGRF1, SCIN and ITPR1 (Fig. 1O). CYP1A1, ALDH1A3 and ABCG2 are involved in different aspects of drug metabolism or xenobiotic response. CYP1A1 metabolizes small molecules, leading to their clearance (e.g. etoposide and irinotecan) or to their conversion into carcinogens (e.g. benzopyrenes) (Androutsopoulos et al., 2009; Badawi et al., 2001; McFadyen et al., 2004). ALDH1A3 is critical for the maintenance of cancer stem cells and is involved in the detoxification of aldehydes generated from alcohols and lipid peroxidation (Duan et al., 2016). ABCG2 is a well-characterized drug efflux pump that plays a vital role in resistance to chemotherapy (e.g. irinotecan and methotrexate) (Robey et al., 2007). ABCG2 was previously shown to be regulated by AHR in esophageal squamous cell carcinoma (To et al., 2012) and upon its activation with pesticides in livestock (Kuhnert et al., 2020).

ADGRF1 is a G protein-coupled receptor involved in transmembrane signaling (Lee et al., 2016). ITPR1 is an ion channel involved in the release of Ca²⁺ from the endoplasmic reticulum. Finally, the actin-severing protein SCIN, also known as adseverin, plays a vital role in cytoskeletal organization and is important for cancer cell proliferation, migration, epithelial-to-mesenchymal transition and CDC42-induced filopodia formation (Chen et al., 2014; Messai et al., 2014; Zunino et al., 2001).

To determine whether this gene signature was directly regulated by AHR, we searched for XREs (AHR-binding sites) in their regulatory regions. Bioinformatics analyses predicted that all six genes contained one or more XREs (Fig. S3A–F), further indicating that these could be direct transcriptional targets of AHR. The localization of these XREs in relation to the transcription start site is listed in Fig. S3G. By examining a chromatin immunoprecipitation-sequencing dataset deposited in the ENCODE database (https://www.encodeproject.org/experiments/ENCSR412ZDC/), we found that among the six genes, two (CYP1A1 and ITPR) were strongly bound by AHR in HepG2 cells (Fig. S3A,F), whereas the other four displayed a weaker interaction (Fig. S3B–E).

AHR ligands promote the expression of the six-gene signature identified by RNA-seq

To validate the induction of these genes by AHR ligands, we performed RT-qPCR for CYP1A1, ALDH1A3, ABCG2, ADGRF1, SCIN and ITPR1 in DLD1 cells incubated with Kyn, FICZ, TCDD and quercetin (Fig. 2A) for 4 h and 8 h. All ligands were confirmed to promote the translocation of AHR into the nucleus of DLD1 cells (Fig. 2B). Kyn activated CYP1A1, ALDH1A3, ABCG2 and ADGRF1 more efficiently at 4 h than at 8 h. TCDD, however, had more dramatic effects at 8 h, potentially explaining the elevated number of target genes regulated by TCDD identified by RNA-seq at the 8-h time point. Incubation with FICZ resulted in wider induction of all the target genes at both 4 h and 8 h (Fig. 2C–H). Quercetin activated CYP1A1 and ADGRF1 at both 4 h and 8 h, but none of the other target genes. Indeed, quercetin was the least potent ligand in driving nuclear translocation of AHR in DLD1 cells (Fig. 2B). CYP1A1 was the gene that was the most readily activated by all four ligands. FICZ was the strongest inducer of CYP1A1, followed by Kyn, TCDD and quercetin. Furthermore, to evaluate the expression of these six genes in response to AHR ligands in other cell lines, we performed RT-qPCR for the same genes in HCT116, HCT15 and HT29 cells incubated with DMSO or Kyn (20 µM) (Fig. 2I–N). These results demonstrated that these six genes are likely regulated in a cell type-specific manner, with ALDH1A3 induced by Kyn in all tested lines and ABCG2 induced in none. Thus, our results identified a signature of genes regulated by ligands used to stimulate AHR translocation in colon cancer cells. These results suggest that different ligands of AHR have unique kinetics of AHR activation and that this should be considered when interpreting the potency of AHR ligands.

Fig. 2.

Activation of AHR by TCDD and Kyn promotes the transcription of a common signature of genes. (A) Molecular structures of the different ligands of AHR (Kyn, TCDD, FICZ and quercetin). (B) Western blot analysis showing the translocation of AHR upon treatment with DMSO, 1 nM TCDD, 20 µM Kyn, 1 µM FICZ and 10 µM quercetin for 30 min in DLD1 cells. Images are representative of three experiments. (C–H) RT-qPCR validation of the expression of the genes regulated by Kyn. DLD1 cells were treated with DMSO, 1 nM TCDD, 20 µM Kyn, 1 µM FICZ and 10 µM quercetin for 4 h and 8 h. Expression of the specified genes was normalized to RPS18. *P≤0.05 for comparison between DMSO and each ligand at 4 h or 8 h (two-tailed unpaired Student's t-test). (I–N) RT-qPCR analyses for the expression of the indicated genes in colon cancer cell lines. HCT116, HCT15 and HT29 cells were incubated with DMSO or 20 mM Kyn for 8 h. Expression of the specified genes was normalized to RPS18. *P≤0.05 comparison between DMSO and Kyn for each cell line (Student's t-test). Data show the mean±s.e.m.

Transcriptional activation of CYP1A1, ALDH1A3, ABCG2, ADGRF1 and SCIN by TCDD depends on AHR

To ascertain whether the induction of the genes identified to be regulated by TCDD and Kyn by RNA-seq was dependent on AHR, we generated stable DLD1 cells expressing an shRNA targeting AHR or an empty vector. We then confirmed that AHR mRNA (Fig. 3A) and protein (Fig. 3B) were downregulated in cells expressing the shRNA targeting AHR. Importantly, incubation with ligands had no dramatic effects on the expression levels of AHR (Fig. 3A,B). These cell lines were then incubated with DMSO or TCDD, and the expression of the six-gene signature was measured by RT-qPCR 8 h after ligand exposure (Fig. 3C–H). This experiment revealed that CYP1A1, as expected, was induced by TCDD, and that this induction was dramatically reduced upon AHR knockdown, confirming that the induction of CYP1A1 by TCDD requires AHR activity in colon cancer cells (Fig. 3C). Similarly, ALDH1A3, ABCG2 and SCIN exhibited increased expression upon incubation of DLD1 cells with TCDD, which was dependent on AHR expression (Fig. 3D,E,H). ITPR1, however, was the only gene of the colon cancer signature that did not display significant regulation by TCDD in the validation experiments, although it was regulated by FICZ (Fig. 2H, Fig. 3G). Owing to their more robust regulation, we decided to focus on CYP1A1, ALDH1A3, ABCG2 and SCIN for further investigation.

Fig. 3.

Modulation of AHR activity by shAHR, TCDD and Kyn regulates the transcription of a common signature of genes in colon cancer cells. (A) RT-qPCR for AHR in DLD1 cells stably expressing empty vector or shRNA for AHR in the presence or absence of TCDD. (B) Western blot for AHR in DLD1 cells stably expressing empty vector or shRNA for AHR in the presence or absence of TCDD. (C–H) RT-qPCR of genes regulated by AHR. DLD1 cells transduced with empty vector or a shRNA targeting AHR were incubated with DMSO or 1 nM TCDD for 24 h. Expression of the specified genes was normalized to RPS18. Fold change values were obtained by normalizing to pLKO-expressing control cells treated with DMSO. *P≤0.05 for comparison between DMSO and TCDD; *P≤0.05 for comparison between pLKO and shAHR (two-tailed unpaired Student's t-test). (I) Western blot for AHR, CYP1A1, SCIN, ABCG2, ALDH1A3, β-catenin and tubulin in AHR CRISPR KO DLD1 cells. (J) Relative proliferation of DLD1 control and AHR CRISPR KO cells with Crystal Violet staining 4 days after seeding. **P<0.0065; ****P<0.0001 (ordinary one-way ANOVA, Tukey's multiple comparisons test, with a single pooled variance, and two-tailed unpaired Student's t-test). Data show the mean±s.e.m. Images are representative of three experiments.

To further validate the importance of AHR on the regulation of these genes, we used CRISPR to generate AHR knockout (KO) lines using two independent AHR short guide RNAs (sgRNAs), which were compared to cells expressing sgControl. Expression of CYP1A1, SCIN and ABCG2 was attenuated in AHR KO cells, although ALDH1A3 levels were only slightly affected (Fig. 3I). These data confirm that AHR regulates CYP1A1, SCIN and ABCG2 even in the absence of ligand stimulation. Additionally, we found that AHR KO cells displayed significantly lower proliferation than control cells, thereby further confirming the role of AHR in promoting cell proliferation in DLD1 cells (Fig. 3J).

Examining the TCGA database, we found no correlation between the expression of CYP1A1, ALDH1A3, ABCG2, ADGRF1, SCIN and ITPR1 and patient survival for colon cancer (Fig. S4A). Moreover, the expression of these genes was not altered with the tumor stage (Fig. S4B). We propose that CYP1A1, ALDH1A3, ABCG2, ADGRF1 and SCIN form a signature of genes expressed transiently alone or in combination in colon cancer cells upon exposure to AHR ligands.

SCIN is required for the growth of colon cancer cells

To determine the biological importance of CYP1A1, ALDH1A3, ABCG2, ADGRF1 and SCIN for colon cancer cell growth, we used two independent siRNAs to knock down the expression of each of these genes in DLD1 cells and measured cell viability. ITPR1 was not evaluated for this experiment due to its minor regulation by TCDD and Kyn (Fig. 3G, Fig. 2H). The ability of each siRNA to reduce gene expression was validated using RT-qPCR 3 days after siRNA transfection in DLD1 cells (Fig. 4A–E). Although siRNAs for every target gene were effective at reducing gene expression (Fig. 4A–E), only SCIN knockdown caused a significant reduction in the growth of DLD1 cells with both siRNAs (Fig. 4F–J). Knockdown of some genes such as CYP1A1, ABCG2 and ADGRF1 showed a marginal decrease in cell proliferation for only one of the siRNAs (Fig. 4F,H,I).

Fig. 4.

SCIN is necessary for colon cancer cell viability. (A–E) RT-qPCR analyses of DLD1 cells transfected with siControl or siRNAs targeting each of the indicated genes. Cells were harvested 3 days after siRNA transfection. Experiments were repeated three times. (F–J) Relative proliferation of DLD1 cells quantified by Crystal Violet staining 6 days after transfection with control siRNA or siRNAs targeting the indicated genes. Experiments were repeated three times. (K) RT-qPCR for SCIN in colon cancer cell lines. (L) Relative proliferation of HCT116 and HT29 cells 5 days after transfection with siControl or siRNA targeting SCIN. Experiments were repeated twice. (M) Western blots of DLD1 cells stably expressing SCIN compared with control cells (Vector). Images are representative of three experiments. (N) Relative proliferation of DLD1 cells stably expressing SCIN compared with control cells. *P≤0.05 for comparison between siControl and siRNAs targeting the indicated genes (ordinary one-way ANOVA, and Tukey's multiple comparisons test, with a single pooled variance). Data show the mean±s.e.m.

To determine whether the sensitivity to CYP1A1, ALDH1A3, ABCG2, ADGRF1 and SCIN knockdown was altered in the presence of AHR ligands, we silenced each of the five genes with siRNAs (Fig. 4A–E) and incubated cells in the presence of DMSO, TCDD and Kyn (Fig. S5A–E). These experiments demonstrated that SCIN knockdown also reduced the viability of DLD1 cells grown in the presence of TCDD and Kyn (Fig. S5E), whereas ADGRF1 only reduced cell viability in the presence of Kyn. This suggests that SCIN and, to a lower degree, ADGRF1 are necessary for the growth of DLD1 cells.

Comparing the expression of SCIN by RT-qPCR in a panel of unstimulated colon cancer cell lines including DLD1, RKO, HCT15, HCT116 and HT29, we found that DLD1 expresses the highest levels of SCIN (Fig. 4K), in agreement with elevated levels of AHR in this cell line (Fig. 1A), followed by HT29 and HCT116. Therefore, to obtain a more diverse sample, we examined the importance of SCIN for the growth of HCT116 and HT29 cells. Given that HCT15 cells were isolated from the same patient as DLD1 (Fig. S5H; information obtained from ATCC), we did not include HCT15 in further experiments. Interestingly, HT29 cells exhibited the most dramatic induction of SCIN upon ligand stimulation (Fig. 2M). Knocking down SCIN in HT29 and in HCT116 cells reduced cell viability (Fig. 4L), demonstrating that SCIN expression is broadly required for colon cancer growth. Nevertheless, stable shRNA-mediated knockdown of SCIN using two independent shRNAs did not significantly affect cell proliferation (Fig. S5F,G). It is likely that stable knockdowns allowed cellular adaptations to compensate for the reduction in SCIN levels. Our data are in agreement with the data from other studies showing that SCIN is necessary for the growth and aggressive behavior of colon cancer cells (Lin et al., 2019).

Expression of SCIN had modest but significant effects in the proliferation of colon cancer cells (Fig. 4M,N). Although SCIN was first found to be highly expressed in normal cells of the kidneys and intestines (Lueck et al., 1998), its function is now recognized to be necessary for the proliferation of multiple cancer cell types (Chen et al., 2014; Lin et al., 2019; Zunino et al., 2001).

SCIN promotes nuclear translocation of β-catenin

To examine the cellular localization of V5-tagged SCIN that was ectopically expressed in DLD1 cells, we performed immunofluorescence using a anti-V5 antibody and DAPI. We found that SCIN was localized in the cytoplasm, cellular periphery and cell–cell junction domains, where actin also localizes (Fig. 5A). Hence, ectopically expressed SCIN displayed localization that is expected for the endogenous protein. Given the importance of SCIN in regulating the actin cytoskeleton, we asked whether SCIN expression affected actin dynamics. We first generated empty vector- and SCIN-expressing lines to also express GFP-tagged F-tractin, an actin sensor. We then used time-lapse microscopy to determine whether the F-actin cytoskeleton was affected in SCIN-expressing cells. Examining filopodia, cell shape and stress fibers, we found no measurable changes upon comparing populations of cells expressing SCIN or an empty vector (Fig. 5B; Fig. S6A). To quantify the ability of cells expressing SCIN or an empty vector to attach to substrates, we measured the attachment of these cells 30 min after seeding (Fig. S6B). Despite the addition of extracellular matrix to increase adhesion of these cells, there was no significant change in attachment in SCIN-expressing cells (Fig. S6B). There was a very modest trend towards decreased adhesion upon SCIN expression.

Fig. 5.

SCIN expression promotes nuclear translocation of β-catenin. (A) Immunofluorescence analyses showing SCIN (V5 tag) and the nucleus (DAPI staining). (B) Immunofluorescence analyses for actin and β-catenin in DLD1 cells expressing empty vector or SCIN. Scale bars: 10 μm. (C) Western blots for β-catenin and tubulin in lysates of DLD1 cells stably expressing vector (empty), SCIN, pLKO (vector for shRNA), shSCIN#22 and shSCIN#23; showing that total β-catenin levels are not affected by SCIN expression. (D) Western blots of nuclear (‘N’) and cytoplasmic (‘C’) fractions of DLD1 cells stably expressing vector or SCIN, immunoblotted for β-catenin, tubulin and histone H3. (E) Western blot of nuclear (‘N’) and cytoplasmic (‘C’) fractions of DLD1 cells stably expressing pLKO, shSCIN#22 or shSCIN#23 immunoblotted for β-catenin, tubulin and histone H3. (F) Western blot of nuclear (N) and cytoplasmic (C) fractions of DLD1 cells control and CRISPR AHR KO cells immunoblotted for β-catenin, tubulin and histone H3. (G) Schematic representation of the β-catenin responsive reporter TOPFlash and its negative control FOPFlash. (H) DLD1 cells stably expressing vector or SCIN were transfected with TOPFlash or FOPFlash and the luciferase reporter activity was measured 48 h later. Experiments were repeated three times. (I) DLD1 cells stably expressing empty vector (pLKO) and shRNA for SCIN were transfected with a TOPFlash or FOPFlash and processed as in H. Experiments were repeated three times. (J) DLD1 cells stably expressing vector or SCIN were transfected with TOPFlash or FOPFlash and treated with 10 µM ICG-001, and luciferase reporter activity was measured 48 h later. Results were normalized by FOPflash levels. (K) Western blots of control (vector) and SCIN-expressing DLD1 cells analyzed for SCIN, MYC, cyclin D1, axin 2 and tubulin. (L) Western blots of control (pLKO) and shSCIN #22 DLD1 cells analyzed for SCIN, MYC, cyclin D1, axin 2 and tubulin. (M) Growth curve of control (vector) DLD1 cells after treatment with DMSO, or 5 or 10 µM ICG-001. (N) Growth curve of SCIN DLD1 cells after treatment with DMSO, or 5 or 10 µM ICG-001. (O) Quantification of SCIN expression in colon cancer patients comparing normal and tumor samples. (P) Example of tumor sample displaying low levels of SCIN expression measured by immunohistochemistry for SCIN. (Q) Representative immunohistochemistry for SCIN and β-catenin in biopsies from a patient with colon cancer containing normal and tumor tissues of the same patient. Scale bars: 10 μm. Data show the mean±s.e.m. Images are representative of n=3. *P<0.05 (one-way ANOVA multiple comparison test).

SCIN has been shown to be involved in collagen degradation in MCF7 cells, which was proposed to affect matrix invasion and metastasis (Tanic et al., 2019). Previous studies found that elevated SCIN expression is associated with the metastatic potential of colon cancer (Lin et al., 2019). To determine the importance of SCIN in the migration of DLD1 cells, we performed scratch assays comparing cells with SCIN overexpression or knockdown. In both cases, SCIN upregulation or downregulation did not affect cell migration (Fig. S6C,D).

Actin interacts intracellularly with cell–cell adhesion complexes containing E-cadherin, β-catenin and α-catenin, and is part of a complex for cell adhesion that anchors the transmembrane protein E-cadherin. Upon activation of the WNT pathway, β-catenin translocates into the nucleus, where it acts as a transcription factor in conjunction with members from the T-cell factor/lymphoid enhancer factor (TCF/LEF) family to drive the expression of genes involved in cell proliferation and migration (MacDonald et al., 2009). In colon cancer cells (Fig. S5H), mutations in the degradation machinery of β-catenin, including AXIN and APC, together with glycogen synthase kinase 3 (GSK3) and casein kinase 1 (CK1) can increase β-catenin levels and its nuclear activity. Other factors can also affect the signaling pool of β-catenin. For example, increased cell–cell adhesion, which occurs in cells grown at high density, recruits β-catenin to the membrane, whereas single-cell culture facilitates nuclear translocation of β-catenin (Conacci-Sorrell et al., 2003). This led us to hypothesize that SCIN, by altering the actin cytoskeleton, might affect the soluble pool of β-catenin.

By performing immunofluorescence using an antibody for β-catenin, we found that expression of SCIN led to an increase in the nuclear pool of β-catenin (Fig. 5B). Total levels of β-catenin were unaffected by SCIN upregulation and downregulation (Fig. 5C). Nuclear and cytoplasmic fractionations of control or SCIN-expressing DLD1 cells demonstrated that SCIN expression promoted an increase in the nuclear pool of β-catenin (Fig. 5D). Additionally, nuclear and cytoplasmic fractions of control versus SCIN knockdown cells demonstrated that the reverse was also true, with SCIN knockdown cells displaying slightly lower nuclear β-catenin levels than control cells.

We probed β-catenin levels in nuclear and cytoplasmic fractions from AHR knockout cells compared to control cells to determine whether a loss of AHR decreased nuclear β-catenin levels. Nuclear fractions of AHR knockout cells displayed a visible decrease in β-catenin (Fig. 5F), which supports the model that AHR can affect the nuclear β-catenin pool through the regulation of SCIN.

To measure the transcriptional activity of β-catenin, we used a reporter gene that contains TCF/LEF-binding sites fused with luciferase (TOPFlash). To account for specificity of the WNT pathway, the activity values obtained for TOPFlash activity were normalized by the values obtained by a mutant form of this motif that cannot bind TCF/LEF factors (FOPFlash). Experiments were performed as previously described (Hao et al., 2019) (Fig. 5G; Fig. S6). Similar to the increase found in nuclear β-catenin in SCIN-expressing cells, we found that WNT signaling activity indicated by TOPFlash measurements was also elevated (Fig. 5H; Fig. S6F). Conversely, knocking down SCIN led to a modest reduction in TOPFlash activity (Fig. 5I; Fig. S6G), which is in line with the modest effects on the nuclear pool of β-catenin upon SCIN knockdown (Fig. 5E).

To explore whether the effects of SCIN upregulation could be attenuated by decreasing β-catenin or TCF transcription, we treated SCIN-expressing and control cells with ICG-001, a Wnt/β-catenin signaling inhibitor. ICG-001 decreased TOPFlash/FOPFlash values of both control and SCIN-expressing cells (Fig. 5J). Despite β-catenin inhibition, SCIN-expressing cells still showed increased activity of β-catenin. These cells also showed reduced proliferation when treated with ICG-001 (Fig. 5M,N). Further analysis of downstream β-catenin target genes was addressed by western blotting (Fig. 5K,L). These showed a a very slight increase of AXIN2 in SCIN-expressing cells compared to control cells, and a decrease in MYC and cyclin D1 (CCND1) upon SCIN knockdown.

Examining normal and tumor samples from a patient with colon cancer, we found that normal tissues did not express detectable levels of SCIN in the colon. However, matching samples of a subset of patients (38%, Fig. 5O) displayed elevated levels of SCIN in the tumors (example of samples from a patient in Fig. 5P,Q). Other patients had low or undetectable levels (Fig. S6G). SCIN co-localized with nuclear β-catenin signal (Fig. 5Q). Using a collection of 37 human biopsy samples, we found that 38% of them had elevated levels of SCIN (Fig. 5O).

Our working model is that ligand-dependent activation of AHR leads to the expression of a signature of target genes that promote drug clearance (such as CYP1A1, ABCG2 and ALDH1A3) and cell growth (such as SCIN). SCIN, by altering the actin cytoskeleton, allows the nuclear translocation of β-catenin either by releasing it from junctions or by additional unknown mechanisms. Nuclear β-catenin activates the WNT pathway to drive cell growth (Fig. 6).

Fig. 6.

Model for the activation of AHR target genes in the presence of Kyn and TCDD. After ligand activation, AHR translocates into the nucleus where it dimerizes with ARNT. Activation of AHR regulates ALDH1A3, ADGRF1, CYP1A1, ABCG2 and SCIN. SCIN interacts with the actin cytoskeleton, possibly releasing β-catenin from junctions. Free β-catenin translocates into the nucleus and regulates genes in the WNT pathway.

DISCUSSION

Our work demonstrates that AHR ligands can induce overlapping signatures of genes in the same cell line, although the group of AHR targets that are regulated is likely cell-type specific. Both Kyn and TCDD promote the expression of genes involved in xenobiotic response and growth. It is possible that endogenously produced ligands such as Kyn are in part responsible for chemoresistance through the activation of AHR. This is supported by several studies which demonstrate a correlation between IDO1 expression and chemoresistance (Campia et al., 2015). As inducing the expression of genes such as CYP1A1, ALDH1A3 and ABCG2 has been shown to play an active role in eliminating the byproducts of the chemotherapeutic agents such as irinotecan, etoposide and methotrexate (Zheng, 2017), this connection is further bolstered. Therefore, preventing the activation of AHR using pharmacological inhibitors in conjunction with chemotherapy might offer a means to help improve patient outcomes.

Of the six genes identified to be transcriptionally activated by AHR ligands in colon cancer cells, we have identified SCIN as an important contributor to cell growth and as a novel WNT pathway activator. Beyond its known actin-severing properties, SCIN possesses the ability to selectively enrich the nuclear β-catenin pool and promote TCF-mediated transcription. We demonstrated that expression of SCIN leads to an increase in cancer cell growth and a correlative increase in nuclear β-catenin. These findings are in agreement with recent studies that demonstrated that AHR activation resulted in increased nuclear β-catenin in breast cancer stem cells, in which AHR expression was elevated (Al-Dhfyan et al., 2017; Mohamed et al., 2019). It is possible that SCIN activity also contributes to WNT pathway activation in breast cancer. Interestingly, a recent study identified how an APC mutation activating β-catenin could generate an oncogenic axis that, in turn, activates AHR in in colon cancer (Park et al., 2020). Therefore, although multiple mechanisms inducing β-catenin accumulation might exist, AHR and β-catenin do appear to have a prevalent crosstalk that is still being investigated.

High SCIN expression could represent a strategy for tumors to stimulate sufficient WNT pathway activity, which might explain why high SCIN expression correlates to metastasis in colon cancer (Lin et al., 2019). Although there is much information on how transcriptional activity regulates the cytoskeleton, there are limited examples of how morphological changes might translate into signaling and transcriptional events (Park et al., 2020). Our work proposes that alterations in SCIN levels and activity are a new mechanism by which cytoskeletal changes regulate transcriptional activity of the WNT pathway.

MATERIALS AND METHODS

Cell culture

The colon cancer cell lines DLD1, HCT116, HCT15, RKO and HT29 were grown in Dulbecco's modified Eagle medium (DMEM, Sigma, D5796) containing 4.5 g/l glucose, L-glutamine and sodium pyruvate and supplemented with 10% fetal bovine serum (Sigma, F2442). All cells were purchased from American Type Culture Collection and tested for contamination before experiments. For attachment assays, DLD1 cells expressing an empty vector or SCIN were seeded in a 96-well plate (one million cells/well), with or without coating of 1% collagen (Thermo Fisher Scientific, CB-40236) or 1% Matrigel (Thermo Fisher Scientific, CB-40230). After 30 min of incubating at 37°C, cells were washed with PBS and attached cells were fixed with methanol and stained with 0.1% Crystal Violet. The dye was stripped using 10% acetic acid and absorbance at 595 nm was measured. Cell motility was assessed using the scratch-wound-healing method [https://www.abcam.com/ps/products/242/ab242285/documents/Wound-Healing-Assay-protocol-book-v1a-ab242285%20(website).pdf]. Cells were seeded at 90% confluency in 24-well plates (Abcam, ab242285). Using the area of the scratch as the cell motility parameter, we quantified the results by using the online software ImageJ.

For AHR CRISPR knockout cell lines, AHR sgRNA CRISPR/CAS9 All-in-One Lentivector or Scrambled sgRNA CRISPR/Cas9 All-in-One Lentivector was transfected together with lentiviral packaging plasmids pSPAX2 (Addgene #12260) and pMD2g (Addgene #12259) to HEK293FT cells. The conditional media of the transfected cells (containing viruses) were filtered with a 0.45 µM filter (vwr, 28145-505) and used to infect DLD1 cells. Knockout cells were selected with 10 µg/ml puromycin, and single clones were picked to confirm the knockout. All sgRNAs were obtained from Applied Biological Materials.

RNA-seq analysis

DLD1 colon cancer cells incubated with DMSO, TCDD (Cambridge Isotope Laboratories) or Kyn (Sigma-Aldrich) for 1 h were harvested and subjected to RNA-seq as previously described (Lafita-Navarro et al., 2018). The quality of sequencing reads were evaluated using NGS QC Toolkit (v2.3.3) (Patel and Jain, 2012) and high-quality reads were extracted. The human reference genome sequence and gene annotation data hg38 were downloaded from Illumina iGenomes (https://support.illumina.com/sequencing/sequencing_software/igenome.html). The qualities of RNA-sequencing libraries were estimated by mapping the reads onto human transcript and ribosomal RNA sequences (Ensembl release 89) using Bowtie (v2.3.2) (Langmead and Salzberg, 2012). STAR (v2.5.2b) (Dobin et al., 2013) was used to align the reads onto the human and viral genomes. SAMtools (v1.9) (Li et al., 2009) was used to sort the alignments, and HTSeq Python package was used to count reads per gene (Anders et al., 2015). DESeq2 R Bioconductor package was used to normalize read counts and identify differentially expressed genes (Gentleman et al., 2004; Anders and Huber, 2010). KEGG (Kanehisa et al., 2017) pathway data were downloaded using the KEGG API (https://www.kegg.jp/kegg/rest/keggapi.html) and Gene Ontology (GO) data were downloaded from NCBI FTP (ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/gene2go.gz). The enrichment of differentially expressed genes to pathways and GO terms were calculated by Fisher's exact test using the statistical package R.

Cell line production

Recombinant lentiviruses were produced by transfecting HEK293T Phoenix-amphotropic packaging cells with pMD2G (VSV-G protein, Addgene, #12259), pPAX2 (lentivirus packaging vector, Addgene, #12260) and lentiviral constructs using Lipofectamine 3000 (Thermo Fisher Scientific, L3000015). Recombinant lentiviruses were harvested 48 h and 72 h after transfection, filtered, and used to infect the colon cancer cells, which were selected with puromycin. shRNA SCIN lentiviral bacterial stocks were purchased from Sigma-Aldrich (TRCN0000116622–TRCN0000116626). For infections, ∼50,000 DLD1, HCT116 and HT29 cells were seeded into six-well plates with 1 ml medium and 2 µl polybrene (Santa Cruz Biotechnology, sc-134220). Each well received either 1 ml of 293T-produced lentivirus [pLKO (empty vector, Sigma Millipore, SHC216) or shSCIN]. Two days later, the medium was exchanged for fresh medium containing antibiotics needed for selection (10 or 5 mg/ml puromycin or 5 mg/ml blasticidin). For ectopic expression of SCIN, the vector was purchased from GeneCopoeia (EX-Z6199-LX304). GFP-F-tractin (9–52 amino acids of the enzyme IPTKA30 fused to eGFP, obtained from the Danuser laboratory, UT Southwestern Medical Center, USA) was introduced in the cells using the pLVX lentiviral system (Clontech) and selected using antibiotic resistance to puromycin.

Western blotting

Cells were lysed in RIPA buffer (25 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, protease and phosphatase inhibitors, MG132) for total protein extracts. Cytoplasmic extracts were isolated using buffer A [10 mM HEPES, 60 mM KCl, 1 mM EDTA, 0.075% (v/v) NP-40, protease and phosphatase inhibitors, MG132]. The remaining pellets are then washed twice with buffer A and resuspended in RIPA, sonicated and centrifuged at 21,130 g for 15 min. Nuclear fractions were collected from the supernatants. All protein lysates were processed for western blotting, and the samples were run on a gradient acrylamide gel (4–12%) and transferred to a nitrocellulose membrane. Membranes were blocked with 10% milk or 5% bovine serum albumin (BSA) in Tris-buffered saline and 0.05% Tween 20 (TBS-T) for 1 h at room temperature or overnight, and incubated overnight with primary antibodies dissolved in TBS-T containing 1% BSA. Membranes were washed the next day thrice with TBS-T for 10 min each before adding the corresponding secondary antibodies dissolved in 5% milk or 1% BSA in TBS-T for 1 h at room temperature. Membranes were once again washed three times with TBS-T for 10 min and then visualized using the Bio-Rad Chemidoc system. The following primary antibodies were used: anti-AHR polyclonal antibody (1:1000, Enzo Life Sciences, 50-201-1848), anti-histone H3 antibody (1:2000, Cell Signaling Technology, 4499S), anti-GAPDH antibody (1:1000, Abcam, ab9484), anti-β-catenin antibody (1:2000, Cell Signaling Technology, 2698S), anti-cyclin D1 antibody (1:1000, Cell Signaling Technology, 2978S), anti-Axin2 antibody (1:1000, Abcam, ab109307), anti-SCIN antibody (1:500, Santa Cruz Biotechnology, sc-376136), anti-CYP1A1 antibody (1:1000, Sigma-Aldrich, SAB2108545), anti-ABCG2 antibody (1:1000, Cell Signaling Technology, 42078S), anti-ALDH1A3 antibody (1:1000, Novus Biologicals, NBP2-15339), anti-tubulin antibody (1:5000, Sigma-Aldrich, T6199-200 μl) and anti-MYC antibody (1:1000, Abcam, ab32072). The following secondary antibodies were used: for IF, Alexa Fluor 594 (Fisher, A11058), Alexa Fluor 488 (Fisher, A11008), and for WB mouse secondary m-IgGk BP-HRP (Santa Cruz Biotechnology, sc-516202), sheep anti-mouse IgG1-HRP (Abcam, ab6808), donkey anti-rabbit IgG HRP (Abcam, ab16284).

RT-qPCR

For RT-qPCR, RNA was extracted using the RNA extraction kit (QIAGEN). Complementary DNA was generated by Superscript III (Invitrogen) and used with SYBR Green PCR master mix (Applied Biosystems) for RT-qPCR analysis. The mRNA content was normalized to that of RPS18. All assays were performed using an Applied Biosystems Prism 7900HT sequence detection system and Bio-Rad CFX. For each mRNA assessment, RT-qPCR analyses were repeated at least twice with three technical replicates each. The following primers were used for RT-qPCR: GAPDH F Human, 5′-TCGTGGAAGGACTCATGACCA-3′, and GAPDH R Human, 5′-AGGCAGGGATGATGTTCTGGA-3′; ITPR1 F1 Human, 5′-GCGGAGGGATCGACAAATGG-3′, and ITPR1 R1 Human, 5′-TGGGACATAGCTTAAAGAGGCA-3′; ITPR1 F2 Human, 5′-CCACAGACGCAGTGCTACTC-3′, and ITPR1 R2 Human, 5′-GTCCCCAGCAATTTCCTGTTT-3′; SCIN F1 Human, 5′-ATGGCTTCGGGAAAGTTTATGT-3′, and SCIN R1 Human, 5′-CATCCACCATATTGTGCTGGG-3′; ADGRF1 F1 Human, 5′-ACAGGGGAAACATCACAGCCA-3′, and ADGRF1 R1 Human, 5′-AAGGATGACACAGCTGCCTCA-3′; CYP1A1 F1 Human, 5′-TCGGCCACGGAGTTTCTTC-3′, and CYP1A1 R1 Human, 5′-GGTCAGCATGTGCCCAATCA-3′; ALDH1A3 F1 Human, 5′-TGAATGGCACGAATCCAAGAG-3′, and ALDH1A3 R1 Human, 5′-CACGTCGGGCTTATCTCCT-3′; and ABCG2 F1 Human, 5′-CAGGTGGAGGCAAATCTTCGT-3′, and ABCG2 R1 Human, 5′-ACCCTGTTAATCCGTTCGTTTT-3′.

Cell growth analyses

Cells were seeded at ∼10,000 cells/ml in 24-well plates in triplicates and at ∼25,000 cells/ml in six-well plates in triplicates. After 24 h, cells were transfected with Lipofectamine RNAiMAX (Life Technologies, 13778100) and 40 nM siRNA. After 96 h, the cells were fixed with methanol and stained with 0.1% Crystal Violet dissolved in 20% methanol. Cells were washed three times and de-stained with 10% acetic acid, and the absorbance of this solution was measured at 595 nm to obtain relative proliferation. Alternatively, cells were seeded at ∼2000 cells/well in 96-well plates in octuplicates with 5 pmol siRNA. After 96 h, 10% WST-8 medium (Thermo Fisher Scientific, 50-190-5565) was added to the cells, which were then incubated for 1.5 h, and analyzed by measuring absorbance at 460 nm (Fig. 1B–D, Fig. 4N, Fig. 5M,N). For the ligand experiments, cells were seeded and treated in the same manner: 24 h after transfection with siRNA, the cells were exposed to Kyn (20 µM), TCDD (1 nM) and DMSO (control). The following siRNAs were used (all from Sigma-Aldrich): SASI_Hs02_00325415 (siSCIN A), SASI_Hs02_00325416 (siSCIN B), SASI_Hs02_00321614 (siITPR1 A), SASI_Hs02_00321615 (siITPR1 B), SASI_Hs01_00243010 (siADGRF1 A), SASI_Hs01_00243011 (siADGRF1 B), SASI_Hs01_00129096 (siALDH1A3 A), SASI_Hs01_00129097 (siALDH1A3 B), SASI_Hs01_00067924 (siCYP1A1 A), SASI_Hs01_00067925 (siCYP1A1 B), SASI_Hs01_00136087 (siABCG2 A), SASI_Hs01_00136088 (siABCG2 B), SASI_Hs02 00332182 (siAHR) and SIC001-10NMOL Mission siRNA Universal Negative Control #1.

TOPFlash luciferase assays

For the LEF/TCF promoter activity assays, ∼1.5×10⁵ empty vector (and pLKO) DLD1 cells and DLD1 cells with SCIN expression or knockdown seeded in triplicate in 12-well plates were transfected with the M50 Super 8× TOPFlash (Addgene #12456) with seven TCF/LEF binding sites (Korinek et al., 1997) or control M51 Super 8× FOPFlash plasmid (Addgene #12457) with seven copies of mutated TCF/LEF consensus sequence (Korinek et al., 1997; Veeman et al., 2003). Cells were lysed with Glo Lysis Buffer (Promega) 48 h after transfection. Luciferase activity was measured with ONE-Glo Luciferase Assay System (Promega) and normalized with FOPFlash samples. All experiments were performed at least two times and each experiment contained three technical replicates. Statistical analyses were performed by ordinary one-way ANOVA, Tukey's multiple comparisons test, with a single pooled variance, and two-tailed unpaired Student's t-test and statistical significance was established by a P≤0.05 for every experiment.

Human tissue samples

De-identified normal and tumor colonic samples were obtained in full compliance of ethical regulations and with consent from all subjects by D. Thomas Brabletz at Nikolaus-Fiebiger-Center for Molecular Medicine University Erlangen-Nuernberg. Tissue samples of colorectal cancer cases (all cases were G2 or G3 adenocarcinomas) were retrieved from the archives of the Dept. of Pathology Erlangen. All samples were from patients operated on in 2003 or earlier, additionally the study was performed according to guidelines mentioned in the Declaration of Helsinki. Patient identity was anonymized, and informed consent was not required at that time. The usage for immunohistochemical analyses of the samples was approved by the local ethics committee (approval 374-14).

Immunohistochemistry, immunofluorescence and light-sheet microscopy

Tissue sections (4 mm) were deparaffinized, dehydrated and pretreated for 22 min in 10 mM citrate buffer pH 6 (citric acid monohydrate; cat. no. #C-7129 Sigma-Aldrich + 0.05% Tween-20) at 121°C in a pressure cooker. Primary antibodies diluted in Dako diluent with background reducing components (from Dako cat.no.S3022) (anti-scinderin antibody; 1:150, Sigma-Merck, #HPA024264) were added to the section and incubated overnight at 4°C. TBS with 0.05% Tween-20 was used to wash the slides twice. Slides were developed with the EnVision System (Dako) and DAB (Cell Signaling DAB substrate kit #8059S) for visualization, according to the manufacturer’s instructions. Immunofluorescence stainings were performed on glass coverslips (0.16 to 0.19 mm). Cells were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, and blocked with 5% BSA for 1 h at room temperature. Samples were incubated with primary antibody for 1 h at room temperature or overnight at 4°C. Alexa Fluor secondary antibodies (1:500) were used. DAPI (1 µg/ml), phalloidin (1:100, Thermo Fisher Scientific, PI12833) and V5 (1:250, Cell Signaling, 13202) were used to visualize the nucleus, actin filaments and SCIN, respectively. For light-sheet microscopy, imaging was performed with cells plated on 35-mm glass-bottomed dishes (P35G-1.5-14 C, Mattek). Imaging was done using an oblique plane microscope (OPM, laser scanning confocal Zeiss LSM880 inverted +2-photon and tissue sections were visualized using a Leica Microscope DM5500B) that uses a bespoke glass-tipped tertiary objective with optical shearing as described before (Chang et al., 2021; Sapoznik et al., 2020). Imaging was done in a chamber at 37°C providing 5% CO₂ with an exposure time of 20 min. The deconvolution was performed in 3D by using the blind deconvolution routine in MATLAB R2020a with a synthetic point-spread function and ten iterations.

Chemicals and inhibitors

TCDD, Kyn, FICZ (Sigma-Aldrich), CH-223191 (S7711, Selleck Chemicals), IGC-001 (Selleckchem, S2662) and quercetin (Cayman Chemical) were dissolved in DMSO. Equivalent volumes of DMSO were used in control conditions.