Article Figures and data Abstract eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Metastasis is a major cause of cancer mortality. We generated an autochthonous transgenic mouse model whereby conditional expression of MYC and Twist1 enables hepatocellular carcinoma (HCC) to metastasize in >90% of mice. MYC and Twist1 cooperate and their sustained expression is required to elicit a transcriptional program associated with the activation of innate immunity, through secretion of a cytokinome that elicits recruitment and polarization of tumor associated macrophages (TAMs). Systemic treatment with Ccl2 and Il13 induced MYC-HCCs to metastasize; whereas, blockade of Ccl2 and Il13 abrogated MYC/Twist1-HCC metastasis. Further, in 33 human cancers (n = 9502) MYC and TWIST1 predict poor survival (p=4.3×10−10), CCL2/IL13 expression (p<10−109) and TAM infiltration (p<10−96). Finally, in the plasma of patients with HCC (n = 25) but not cirrhosis (n = 10), CCL2 and IL13 were increased and IL13 predicted invasive tumors. Therefore, MYC and TWIST1 generally appear to cooperate in human cancer to elicit a cytokinome that enables metastasis through crosstalk between cancer and immune microenvironment. eLife digest Cancer develops when cells in the body gain mutations that allow them to grow and divide rapidly and uncontrollably. As the disease progresses these cancer cells develop the ability to spread around the body. This process of spreading, called metastasis, is responsible for most cancer-related deaths in humans, but no current treatments target it. Mutations that increase the levels of two proteins known as MYC and TWIST1 in cells cause many human cancers. In healthy adult cells, normal levels of MYC and TWIST1 act as key regulators that switch thousands of genes on or off. TWIST1 is known to control the movement and spread of cells in the embryo. However, it is not known how MYC and TWIST1 work together to promote the metastasis of cancer cells. To address this question, Dhanasekaran, Baylot et al. used mice to investigate the roles of MYC and TWIST1 in the metastasis of cancer cells. The experiments showed that these two proteins work together to reprogram mouse cancer cells to release signal molecules known as cytokines. These molecules convert immune cells known as macrophages to a tumor-friendly state that allows cancers cells to spread around the body. Inhibiting two cytokines known as CCL2 and IL13 prevented the cancer cells from moving. Further experiments analyzed tumor samples from around 10,000 human patients with 33 different cancers. This revealed that patients that had higher levels of MYC and TWIST1 proteins in their tumors also had increased levels of CCL2 and IL13, more activated macrophages and were less likely to recover from their cancer. The findings of Dhanasekaran, Baylot et al. suggest that MYC and TWIST1 may instigate metastasis in many human cancers, and therapies targeting specific cytokines may prevent these cancers from spreading around the body. Furthermore, screening blood for the levels of cytokines may help to identify the cancer patients who would benefit from such therapies. Introduction Tumorigenesis is caused by specific oncogenes but tumor progression often involves the acquisition of metastasis (Chaffer and Weinberg, 2011). Metastasis occur when tumor cells gain the ability to invade, migrate and colonize distant sites, and this accounts for most of the morbidity and mortality associated with cancer (Mehlen and Puisieux, 2006). Many studies have examined human clinical specimens and/or tumor-derived cell lines to discern mechanisms of metastasis, and identified the role of specific genes (Ji et al., 2007; Sun et al., 2018; Zhu et al., 2017) and also the role of the tumor microenvironment (Fidler, 2003; Kalluri, 2016; Kim et al., 2017; Whitfield and Soucek, 2012). However, to date, there are very few mouse models that exhibit spontaneous metastasis, and even fewer in vivo models where the stepwise progression from a non-metastatic to metastatic cancer can be studied. Such a model would provide a tractable approach for studying specific mechanisms of metastasis, particularly the role of the immune microenvironment. Innate immune cells, especially tumor associated macrophages (TAMs), are known to contribute to metastasis through multiple mechanisms including effects on angiogenesis, production of specific cytokines, suppression of the immune system, and induction of epithelial-mesenchymal transition (EMT) (Gonzalez et al., 2018; Lu et al., 2011; Qian and Pollard, 2010; Wan et al., 2014). The specific discrete events in the cancer cell that modulate the tumor immune microenvironment and enable metastasis are not clear. The MYC oncogene is a transcription factor that is one of the most commonly activated oncogenes in the pathogenesis of many types of human cancer including HCC (Schaub et al., 2018; Dang, 2012; Gabay et al., 2014). Previously, we used the Tet System to generate a conditional transgenic mouse model for MYC-induced HCC that we and others have used to study mechanisms of oncogene addiction (Settleman, 2012) and identify potential therapies (Dhanasekaran et al., 2018; Kapanadze et al., 2013; Ma et al., 2016; Shachaf et al., 2004). But murine MYC-driven HCC do not metastasize. Twist1 is a transcription factor that is important during embryogenesis for normal cellular migration (Lee et al., 1999; Thisse et al., 1987). Twist1 has been shown to be an important gene product that can enable mouse and human tumor cell lines to acquire the ability to metastasize associated with EMT (Thiery et al., 2009; Xu et al., 2017). Here we used the Tet System to conditionally express Twist1 in combination with MYC to show that their co-expression leads to widely metastatic and invasive HCC. We use this powerful in vivo model to uncover a surprising mechanism by which MYC and Twist1 drive metastasis. Cancer cell-intrinsic properties like proliferation, apoptosis or invasiveness were not different between the non-metastatic MYC-HCC and the metastatic MYC/Twist1-HCC. Instead, metastatic progression was dependent on the ability of MYC and Twist1 to dramatically reprogram the tumor innate immune microenvironment. Together, MYC and Twist1 induce the cancer cell to secrete cytokines like Ccl2 and Il13 that lead to recruitment and polarization of macrophages respectively, thus causing metastasis. Systemically, administering Ccl2 and Il13 is sufficient to cause metastasis of MYC-HCC and, conversely blocking these specific cytokines profoundly inhibits metastasis in MYC/Twist1 HCC. Our results are broadly generalizable to 33 different human cancers and predict invasive cancer in a pilot clinical study. Results Twist1 induces spontaneous metastatic progression of MYC-driven HCC in vivo We first generated a transgenic mouse using the Tet system that conditionally expresses Twist1 in a liver specific manner (LAP-tTA/TRE-Twist1/Luc). We crossed TRE-Twist1/Luc mice which harbored the Twist1 and firefly luciferase (luc) genes under the control of a bidirectional tetracycline responsive element (TRE), with the LAP-tTA mice which contain the tetracycline-controlled transactivator protein (tTA) driven by the liver-enriched activator protein (LAP) promoter (Tran et al., 2012). Twist1 transgenic mice (LAP-tTA/TRE-Twist1/Luc) exhibited no disease nor gross or microscopic pathology for as long as 18 months of observation thus demonstrating that Twist1 did not play a role in autochthonous tumorigenesis when overexpressed in the liver (Figure 1—figure supplement 1a). To examine the influence of Twist1 on tumor progression, LAP-tTA or LAP-tTA/TRE-Twist1/Luc mice were crossed with TRE-MYC (Shachaf et al., 2004) (Figure 1a) to generate transgenic mice that inducibly expressed MYC alone (MYC mice) or co-expressed MYC, Twist1 and luciferase (Luc) in a liver-specific manner (MYC/Twist1 mice) (Figure 1b). We induced transgene expression in adult mice at 6 weeks of age (Figure 1b). In vivo, Twist1 transgene expression was confirmed to be confined to the liver by measuring the luciferase reporter by bioluminescence imaging (BLI) (Figure 1c). We followed in vivo tumor progression with serial cross-sectional imaging. Both MYC and MYC/Twist1 mice were observed to develop multifocal liver cancer, while only MYC/Twist1 mice developed lung metastases (Figure 1d). MYC/Twist1 mice were moribund with HCC sooner and had a median survival of 25 months compared to 32 months in MYC mice (p<0.001, Figure 1e). MYC mice rarely exhibited metastasis even after extended observation (2%, n = 50, Figure 1f); whereas, MYC/Twist1 mice regularly exhibited rapid onset of metastasis with high penetrance (90%) -metastases to the lungs (70%), peritoneum (60%) and lymph nodes (20%) (n = 50, Figure 1f). Thus, Twist1 combined with MYC expression in liver cells elicits HCC metastasis. Figure 1 with 1 supplement see all Download asset Open asset Twist1 induces spontaneous metastatic progression of MYC driven HCC in vivo. (a) Mouse model of MYC induced HCC where MYC is under the control of a tetracycline responsive element (TRE) which contain the tetracycline-controlled transactivator protein (tTA) driven by the liver-enriched activator protein (LAP). Doxycycline (Dox) can be used to inactivate oncogene expression in adult mice. (b) Mouse model of MYC/Twist1-induced HCC which inducibly co-expressed MYC, Twist1 and firefly luciferase in a hepatocyte specific manner. (c) Bioluminescent imaging (BLI) confirms in vivo rapid induction of oncogenes by demonstrating liver specific luciferase expression upon withdrawal of Dox. (d) Serial cross sectional imaging of MYC- (n = 10) and MYC/Twist1-HCC (n = 10) using MRI scan for the abdomen and CT scan for the lungs demonstrate step-wise tumor progression. Both MYC and MYC/Twist1 mice develop multifocal liver tumors but only the latter develops lung metastases. (e) Kaplan Meier survival curves show that MYC/Twist1 mice (n = 16) had significantly shorter survival than MYC mice (n = 12) (**p<0.01) while Twist1 transgenic mice (n = 10) remained healthy. (f) Pie charts show incidence of metastasis in MYC- (n = 50) and MYC/Twist1 transgenic mice (n = 50). (g) Comparison of liver weights between MYC (n = 3) and MYC/Twist1 tumor bearing mice (n = 4) and control mice (n = 3) which were kept on Dox throughout (**p<0.01). (h) Gross and histopathologic appearance of tumors in MYC transgenic model confirming HCC. Lungs do not show any metastases. (i) Representative images showing MYC/Twist1-HCC have histologic appearance of HCC and lung histology shows metastatic disease. (j) Representative images from Immunohistochemistry to show phospho histone three (PH3) expression in MYC- and MYC/Twist1-HCC with quantification of IHC staining. (k) Representative images from Immunohistochemistry to show Cleaved Caspase three (CC3) expression in MYC- and MYC/Twist1-HCC with quantification of IHC staining. A simple explanation for our results is that Twist1 was inducing more rapid onset and thereby progression of tumorigenesis. Against this possibility, the tumor burden in the liver was not statistically different between MYC and MYC/Twist1 mice (Figure 1g). Also, there was no difference in the gross or microscopic appearance of MYC- and MYC/Twist1-HCC (Figure 1h–i). The MYC- and MYC/Twist1-HCC tumors were confirmed to be HCC by a pathologist and by expression of hepatocyte marker glutamine synthetase (Figure 1—figure supplement 1b). We considered that Twist1 could be influencing MYC expression levels, but MYC levels were similar between the two tumor models, while Twist1 was only overexpressed in the MYC/Twist1-HCC (Figure 1—figure supplement 1c-d). Tumor cell proliferative index (phospho histone three expression) and apoptosis (cleaved caspase three) between MYC- and MYC/Twist1-HCC were not different (Figure 1j–1k). Primary tumor-derived cell lines from MYC- and MYC/Twist1-HCC did not show any difference in migratory capacity (Figure 1—figure supplement 1e). Lastly, Twist1 is a regulator of epithelial-mesenchymal transition (EMT) 26,37, but we did not observe significant differences in the expression of multiple epithelial and mesenchymal markers between MYC and MYC/Twist1 tumors (Figure 1—figure supplement 1f). Therefore, Twist1 drives metastasis of MYC-induced HCC without affecting primary tumor burden, MYC expression, tumor cell proliferation, apoptosis, invasiveness or EMT markers. MYC and Twist1 cooperate to remodel the tumor immune microenvironment The influence of Twist1 on global gene expression was measured in MYC- and MYC/Twist1-HCC (n = 5) using next generation sequencing (NGS) based RNA sequencing. Through unsupervised hierarchical clustering using principal component analysis (PCA), MYC- and MYC/Twist1-HCC were found to have overall distinct, non-overlapping expression profiles that clustered separately (Figure 2a). A comparative analysis identified 514 genes (220 up and 294 down) that were differentially expressed between MYC-HCC and MYC/Twist1-HCC (p<0.001, q < 0.05, fold change ≥2) (Figure 2b, Supplementary file 1). Functional pathway analysis revealed the top biological processes upregulated in MYC/Twist1-HCC involved inflammatory responses including leukocyte infiltration, myeloid cell and granulocyte recruitment (Figure 2b, Figure 2—figure supplement 1a, Supplementary file 2). CIBERSORT (Newman et al., 2015) identified M2 macrophages to be significantly enriched in MYC/Twist1 tumors, of the 22 immune subsets analyzed (Figure 2c). MYC/Twist1-HCC exhibited a 15-fold shift in the ratio of M2 to M1 macrophages when compared to MYC tumors (Figure 2c). No significant differences in other major immune compartments were seen including- B cells, T cells, NK cells, dendritic cells, neutrophils, or mast cells (Figure 2d). Increased macrophage infiltration in MYC/Twist1 primary and metastatic tumors (Figure 2e) with no change in neutrophils or CD4 T cells infiltration was confirmed by IHC (Figure 2—figure supplement 1b). TAMs isolated from primary MYC/Twist1-HCC had increased macrophages of the M2 phenotype (Cd206High/Arg1High), more specifically a M2a phenotype (Figure 2f). Induction of MYC and Twist1 expression is associated with tumor initiation and rapid onset of macrophage infiltration in early tumors which increases during tumor progression. Conversely, the inactivation of MYC and Twist1 in tumors shows rapid and complete tumor regression in all observed mice (n = 20) within 2 weeks (Figure 2—figure supplement 1c) and also prompt exodus of macrophages (Figure 2g–2h). Figure 2 with 1 supplement see all Download asset Open asset MYC and Twist1 cooperate to remodel the tumor immune microenvironment. (a) Principal component analysis (PCA) showed that MYC- (n = 5) and MYC/Twist1-HCC (n = 5) overall had distinct, non-overlapping expression profiles. Volcano plot shows comparative analysis of differentially expressed genes between the MYC-HCC and MYC/Twist1-HCC. (b) Ingenuity pathway analysis of differentially expressed genes between MYC/Twist1- (n = 5) and MYC-HCC (n = 5) used to identify top biological processes upregulated in MYC/Twist1-HCC. (c) Comparison of relative percentage of monocyte and macrophage subpopulations, derived using CIBERSORT analysis, between MYC- (n = 5) and MYC/Twist1-HCC (n = 5) (*p<0.05). (d) Comparison of relative abundance of major immune subsets between MYC- (n = 5) and MYC/Twist1-HCC (n = 5) (p=ns). (e) Representative images from Immunohistochemistry staining for F4/80 in MYC and MYC/Twist1 normal liver (n = 4), primary tumor (n = 4) and lung (n = 4) with quantification in bar graph (*p<0.05, **p<0.01). (f) Macrophages were isolated from primary tumors and expression level of M2 markers (Cd206, Arg1) and M1 markers (iNos, Ccr2) was compared between MYC- and MYC/Twist1-HCC (*p<0.05). (g) Representative images from IHC staining for MYC and F4/80 in MYC/Twist1 HCC in temporal sequence from early tumor (n = 5) to later tumor (n = 5) and tumor regression (n = 5) upon oncogene inactivation. (h) Quantification of F4/80 staining in early tumor vs late tumor vs regressed MYC/Twist1-HCC (**p<0.01). TAMs have been shown to increase the migratory capacity of tumor cells (Lin et al., 2001). Conditioned media from TAMs isolated from MYC/Twist1-HCC but not MYC-HCC increased the invasiveness of both MYC- and MYC/Twist1-HCC tumor cells in vitro (Figure 3a–3c). Figure 3 with 1 supplement see all Download asset Open asset Coordinate expression of both MYC and Twist1 are necessary for inducing metastasis. (a) Schematic of the experiment to extract tumor associated macrophages from the primary tumors of MYC- (n = 5) and MYC/Twist1-HCC (n = 5). Conditioned media from macrophages extracted from MYC-HCC (n = 3 biological replicates with three technical replicates each) or MYC/Twist1-HCC tumors (n = 3 biological replicates with three technical replicates each) was used to treat either MYC- or MYC/Twist1-HCC cells. (b) Wound healing assay was performed in MYC- or MYC/Twist1-HCC cells which were treated with conditioned media derived from TAMs isolated from primary MYC- or MYC/Twist1-HCC. (c) Bar graphs show quantification of percentage wound closure at 24 hr (**p<0.01). (d) Schematic showing generation of cell lines which constitutively express MYC and/or Twist1 (blue font) upon transgene inactivation with Dox treatment. (e) H and E (1X and 20X) of lungs of mice injected with the four different cell lines as shown in their individual titles. (f) Quantification of lung metastatic burden upon intravenous (IV) injection of cell lines expressing either MYC or Twist1 alone or both (***p<0.001). To discriminate the independent roles of MYC and Twist1 in metastasis we developed cell lines where we could modulate the expression of MYC and Twist1 separately. Primary cell lines from MYC/Twist1 HCC were retrovirally transduced with MYC and/or Twist1, such that upon inactivation of transgene expression with Doxycycline, they now constitutively expressed MYC and/or Twist1 (Figure 3d). We confirmed that treatment of these cell lines with doxycycline resulted in the continued expression of constitutive MYC and/or Twist1 by qPCR (Figure 3—figure supplement 1a-b). We observed that the inactivation of either MYC or Twist1 abrogated the ability of the cells to develop lung metastasis when injected intravenously in NOD scid gamma (NSG) mice, while cells expressing both MYC and Twist1 led to development of extensive lung metastasis with prominent macrophage infiltration (Figure 3e–3f, Figure 3—figure supplement 1c). Thus, MYC and Twist1 cooperate, and are both required to induce metastasis of HCC by a macrophage dependent mechanism. Tumor associated macrophages are required for Twist1 to induce metastasis We determined if TAMs are required for Twist1 to drive metastasis on MYC-HCC in vivo. Primary tumor-derived cell lines which conditionally express MYC or MYC/Twist1 (Figure 4—figure supplement 1a) were re-introduced in vivo either by orthotopic transplantation into the liver or intravenous injection. Orthotopic implantation (Figure 4a) of MYC/Twist1- but not MYC-HCC tumor cells in NSG mice led to pulmonary and intrahepatic metastases with extensive macrophage infiltration (Figure 4b–c, Figure 4—figure supplement 1b). Macrophage depletion with clodronate liposomes (Moreno, 2018) but not control liposomes, in mice orthotopically transplanted with MYC/Twist1-HCC had reduced intrahepatic (p=0.0006, FC 4.4) and lung metastases (p<0.0001, FC 8.8) (Figure 4e–f). Quantification of BLI signal at the end of treatment did not show statistical difference in densitometry between control treated and clodronate treated mice (Figure 4e). Note, clodronate was confirmed to remove macrophages but not affect tumor cells (Figure 4—figure supplement 1c-d). A reduction in the number of macrophages was confirmed by IHC for F4/80 in normal liver, tumor and lungs (p<0.001) (Figure 4—figure supplement 1e-f). Figure 4 with 1 supplement see all Download asset Open asset Tumor associated macrophages are required for Twist1 to induce metastasis of MYC-HCC. (a) Experimental scheme- MYC and MYC/Twist1 cells were implanted orthotopically and metastatic burden in liver and lung assessed after 4 weeks. (b) Representative BLI imaging, gross organ appearance, histopathology of liver (10X) and lungs (1X) from mice orthotopically implanted with MYC (n = 5) and MYC/Twist1 cells (n = 5). (c) Comparative quantification of liver and lung metastatic burden between MYC (n = 5) and MYC/Twist1 orthotopic HCC (n = 5). (**p<0.01). (d) Experimental model of orthotopic MYC/Twist1-HCC treatment either with control liposomes or clodronate liposomes for 4 weeks for macrophage depletion. (e) Representative BLI imaging, BLI quantification, gross organ appearance, histopathology of liver and lungs from MYC/Twist1 orthotopic HCC bearing mice treated with either control liposomes (n = 5) or clodronate liposomes (n = 4). (f) Comparative quantification of liver and lung metastatic burden between MYC/Twist1 orthotopic HCC bearing mice treated with either control liposomes (n = 5) or clodronate liposomes (n = 4) (**p<0.01). (g) Experimental scheme- MYC and MYC/Twist1 cells were injected intravenously and metastatic burden in lung assessed after 4 weeks. (h) Representative BLI imaging, gross organ appearance, histopathology of lungs from mice intravenously injected with MYC (n = 5) and MYC/Twist1 cells (n = 4). (i) Comparative quantification of lung metastatic burden between MYC (n = 5) and MYC/Twist1 intravenously injected HCC (n = 4). (**p<0.01). (j) Experimental model of intravenous MYC/Twist1-HCC treatment either with control or clodronate liposomes for 3 weeks for macrophage depletion. (k) Representative BLI imaging, and lung histopathology from MYC/Twist1 intravenous HCC injected mice treated with either control liposomes (n = 4) or clodronate liposomes (n = 5). (l) Comparative quantification of liver and lung metastatic burden between MYC/Twist1 intravenously injected HCC bearing mice treated with either control liposomes (n = 4) or clodronate liposomes (n = 5) (**p<0.01). To evaluate if macrophages are required for the colonization step of metastasis, we used the lung trap assay. Intravenous injection of MYC/Twist1-HCC but not MYC-HCC cells resulted in pulmonary metastases associated with macrophage infiltration (Figure 4g–i, Figure 4—figure supplement 1g). Clodronate depletion of macrophages, almost completely abrogated pulmonary metastasis (p=0.0003, FC 4.3) (Figure 4j–4l). Therefore, Twist1 elicits metastasis of MYC-HCC and promotes the invasiveness and colonization of metastases by a macrophage-dependent mechanism. MYC and Twist1 reprogram the crosstalk between cancer cells and macrophages To evaluate if cytokines secreted by cancer cells mediate MYC and Twist1 driven macrophage recruitment and polarization, we evaluated the impact of tumor cell derived conditioned media on non-polarized macrophage cell lines. Conditioned media derived from MYC/Twist1- but not MYC-HCC cells (Figure 5a) was sufficient to promote the migration of macrophages towards cancer cells (Figure 5b). Conditioned media from MYC/Twist1- but not MYC-HCC cells was able to elicit changes in the morphology of macrophages to resemble M2 phenotype (McWhorter et al., 2013) (Figure 5c) and increased expression of M2 markers: Cd206, Arg1 and Cc3cr1 (Figure 5c), but not M1 markers: iNos, Ccr2, Ifnar2 (Figure 5—figure supplement 1a). A multiplex ELISA for 38 cytokines was performed (Figure 5d) identifying that MYC/Twist1-HCC cells had increased secretion of cytokines: Il13, Ccl2, Ccl5, Ccl7 and Cxcl1 (p<0.05; Fold change ≥2, mean ≥20 ng/ml) (Figure 5e, Figure 5—figure supplement 1b). These five cytokines were confirmed to be transcriptionally upregulated in MYC/Twist1- vs. to MYC cells by qPCR (Figure 5e). Figure 5 with 1 supplement see all Download asset Open asset MYC and Twist1 reprogram the cytokinome to induce macrophage recruitment and polarization. (a) Experimental scheme- conditioned media (CM) from MYC or MYC/Twist1 cells was used to treat non polarized macrophages for 48 hr. Following that, macrophage migration or polarization was assessed. (b) Transwell macrophage migration across a membrane insert when treated with CM from MYC-cells (n = 3 biological replicates, with three technical replicates each) or MYC/Twist1-cells (n = 3 biological replicates, with three technical replicates each). Bar graph shows quantification of migrated cells. (c) Morphologic appearance of macrophages treated with CM from MYC or MYC/Twist1 cells. Expression of M2 markers (CDd06, Arg1, Cx3cr1) in macrophages treated with CM from MYC (n = 3 biological replicates, with three technical replicates each) or MYC/Twist1 cells (n = 3 biological replicates, with three technical replicates each). (*p=0.05). (d) Experimental scheme- the CM from MYC or MYC/Twist1cells were analyzed using Luminex-plate based multiplex ELISA assay. (e) Heatmap showing expression levels of top five differentially secreted cytokines in CM of MYC (n = 3 biological replicates, with three technical replicates each) or MYC/Twist1 cells(n = 3 biological replicates, with three technical replicates each) by ELISA. Second heatmap showing mRNA expression levels of top five cytokines between MYC or MYC/Twist1 cells by qPCR. (f) Experimental scheme- Co-culture of MYC/Twist1 cells and macrophages separated by a chamber to evaluate chemotaxis of macrophages towards the cancer cells was performed. Neutralizing antibodies to individual cytokines or control antibody were added to the CM of MYC/Twist1 cells. (g) Transwell chamber migration assay of macrophages in the upper chamber toward the MYC/Twist1 cells in the lower chamber. MYC/Twist cells CM was treated with control antibody or neutralizing antibody to Il13, Ccl2, Ccl5, Ccl7 or Cxcl1 respectively. (n = 3 biological replicates, with three technical replicates each)(**p<0.01). (h) Experimental scheme- CM of MYC/Twist1 cells treated with control antibody or neutralizing antibody to Il13, Ccl2, Ccl5, Ccl7 or Cxcl1 respectively was added to non-polarized macrophages for 48 hr. (i) Macrophage polarization was assessed by qPCR for M2 markers (ARG1, CD2016) and M1 markers (Ifnar2, Ccr2)(n = 3 biological replicates, with three technical replicates each) (***p<0.001). The role of individual cytokines in macrophage recruitment was assessed. Antibodies that neutralize Ccl2, Ccl5, Ccl7 or Cxcl1 inhibited the ability of conditioned media from MYC/Twist1-HCC to promote migration of macrophages (Figure 5f–h). Neutralization of Ccl2 decreased migration 7-fold (p<0.0001), Ccl5 1.2-fold (p=0.001), Ccl7 1.8-fold (p=0.001) and Cxcl1 2.1-fold (p=0.0002). Neutralizing Il13 (p=ns, FC 1.2) did not affect macrophage migration. Also, we found that the neutralization of Ccl2 inhibited the recruitment of macrophages into MYC/Twist1-spheroids by 3D culture (Figure 5—figure supplement 1c). Next, the role of cytokines on macrophage polarization was determined (Figure 5h). We found that neutralization of Il13 blocked M2 polarization by 50-fold reduction in Cd206 expression (p<0.0001) and 7-fold decrease in Arg1 (p<0.0001) without any change in M1 markers. Neutralization of Ccl5 led to a 4-fold reduction in Cd206 (p<0.001) and 1.1-fold decrease in Arg1 (p<0.05) and Cxcl1 led to 2.4-fold decrease in Cd206 (p<0.001) without significant change in Arg1 (p=ns) (Figure 5i). Conversely, adding the cytokines Il13, but not Ccl2, to co-cultured MYC-HCC cells increased M2 markers expression (Figure 5—figure supplement 1d). MYC and TWIST1 are both transcription factors, so we evaluated if they epigenetically regulated CCL2 and IL13 expression. We identified MYC and TWIST1 promoter binding upstream of human CCL2 and IL13 protein-coding genes in Gene Transcription Regulation Database (GTRD), a meta-analysis of Chip-seq experiments (Dreos et al. 2017; Yevshin et al., 2019) (Figure 5—figure supplement 1e). Both MYC and Twist1 demonstrated binding at multiple sites in the promoter regions of CCL2 and IL13 in the ChIP-seq data from several different cancer cell lines (Figure 5—figure supplement 1e). We also looked for MYC and Twist1 promoter binding sites in mouse Ccl2 and Il13 promoters using motif finding analysis of the public data from JASPAR (Bryne et al., 2008) and Eukaryotic promoter database (EPD) (Dreos et al. 2017). Again, we found multiple potential MYC and Twist1 transcription factor binding sites for both Ccl2 and Il13 (Figure 5—figure supplement 1f). These data suggest that MYC and Twist1 cooperate to transcriptionally regulate expression of Ccl2 and Il13 in the cancer cells. Ccl2 and Il13 are sufficient and required for metastasis of HCC We examined if Ccl2 and/or Il13 are sufficient to elicit metastasis in vivo. Orthotopic transplants of non-metastatic MYC-HCC in NSG mice were treated with either PBS (control), or with recombinant Ccl2 alone or Il13 alone or their combination for 4 weeks (Figure 6a). Control mice did not develop metastatic nodules even though scattered, single cells were found in the lungs (Figure 6b). No mice treated with Ccl2 alone (p=0.390, FC = 3) or Il13 alone (p=0.99, FC = 1) exhibited intrahepatic or pulmonary metastases (Figure 6c, e and g). All orthotopic MYC-HCC mice treated with the combination of Ccl2 and Il13 developed intrahepatic metastases (p=0.008, FC 2.5) and multifocal pulmonary metastases (p=0.02, FC = 104.4) (Figure 6c, e and g). Hence, both Ccl2 and Il13 are sufficient to elicit MYC-HCC to metastasize, even if Twist1 is not expressed in the tumor cells.
<p>Evaluation of ZL-2201 antiproliferative activity <i>in vitro</i>. <b>A,</b> Chemical structure of ZL-2201, a potent and selective DNA-PK inhibitor. <b>B,</b> Concentration-dependent response to ZL-2201 in M059J and M059K glioblastoma cancer cells measured by CTG after 6-day of treatment. The graph represents the average inhibitory (IC<sub>50</sub>) values (<i>n</i> = 3). <b>C,</b> Concentration-dependent response to ZL-2201 in CRISPR KO of ATM in A549 and FaDu cancer cells measured by CTG after 6 days of treatment. The graph represents the average inhibitory (IC<sub>50</sub>) values (<i>n</i> = 2–5). <b>D,</b> IC<sub>50</sub> responses to ZL-2201 in ATM mutant cell lines (black bars) versus ATM WT A549 cell line (gray bar). Cell growth was measured by CTG after 6 days of treatment. The graph represents the average inhibitory (IC<sub>50</sub>) values (<i>n</i> = 2–6). NCI-H1703 (<b>E</b>) and A549 (<b>F</b>) cancer cells were treated with bleomycin (10 μmol/L) for 2 hours followed by the addition of increasing concentration of ZL-2201 for 2 hours. Whole-cell lysates were harvested, and concentration-dependent inhibition of DNA-PK protein were analyzed by Simple Western.</p>
'%3e%3ccircle r='20' fill='%23008'/%3e%3ccircle r='17.5' fill='%23fff'/%3e%3ccircle r='3.5' fill='%23008'/%3e%3cg id='d'%3e%3cg id='c'%3e%3cg id='b'%3e%3cg id='a' fill='%23008'%3e%3ccircle r='.875' transform='rotate(7.5 -8.75 133.5)'/%3e%3cpath d='M0 17.5.6 7 0 2l-.6 5L0 17.5z'/%3e%3c/g%3e%3cuse xlink:href='%23a' transform='rotate(15)'/%3e%3c/g%3e%3cuse xlink:href='%23b' transform='rotate(30)'/%3e%3c/g%3e%3cuse xlink:href='%23c' transform='rotate(60)'/%3e%3c/g%3e%3cuse xlink:href='%23d' transform='rotate(120)'/%3e%3cuse xlink:href='%23d' transform='rotate(-120)'/%3e%3c/g%3e%3c/svg%3e)