ABSTRACT
Cells that lead collective invasion can have distinct traits and regulatory programs compared to the cells that follow them. Notably, a specific type of epithelial-to-mesenchymal transition (EMT) program we term a “trailblazer EMT” endows cells with the ability to lead collective invasion and promote the opportunistic invasion of intrinsically less invasive siblings. Here, we sought to define the regulatory programs that are responsible for inducing a trailblazer EMT in a genetically engineered mouse (GEM) model of breast cancer. Analysis of fresh tumor explants, cultured organoids and cell lines revealed that the trailblazer EMT was controlled by TGFβ pathway activity that induced a hybrid EMT state characterized by cells expressing E-cadherin and Vimentin. Notably, the trailblazer EMT was active in cells lacking keratin 14 expression and evidence of trailblazer EMT activation was detected in collectively invading cells in primary tumors. The trailblazer EMT program required expression of the transcription factor Fra1, which was regulated by the parallel autocrine activation of the epidermal growth factor receptor (EGFR) and extracellular signal regulated kinases (ERK) 1 and 2. Together, these results reveal that the activity of parallel TGFβ and EGFR pathways confers cells with the ability to lead collective invasion through the induction of a trailblazer EMT.
INTRODUCTION
Epithelial-to-mesenchymal transition (EMT) program activation is an established feature of primary tumors that promotes tumor progression and metastasis (1). Rather than being a single program, there are a wide range of EMTs that are regulated by diverse activation mechanisms and have unique molecular properties (2,3). This variability allows specific EMT programs to confer disOnct functional traits, including invasive behavior (4), tumor initiating activity (5), suppression of the immune system (6) and resistance to therapeutics (7). Different EMT programs can be active in tumor subpopulations (4,8), raising the potenOal for complementary functions of cells in distinct EMT states. Indeed, specific types of EMT programs that influence tumor population properties have begun to be discovered (9). Notably, there are EMTs in which cells are endowed with the ability to influence the invasive behavior of non-EMT induced siblings (10,11). A “trailblazer EMT” program confers cells with an enhanced ability to iniOate the collective invasion of adherent multicellular units through the extracellular matrix (ECM) (12). Collecttive invasion is the most frequent type of invasion detected in primary tumors (13–15). collective invasion may also contribute to the dissemination of tumor clusters that seed metastases (16,17). Thus, understanding trailblazer EMTs is necessary to understand a common mode of local dissemination and a potential mode of metastasis.
During collective invasion, leading tumor cells extend actin-rich protrusions that deliver proteas-es and exert physical forces to reorganize the surrounding ECM into parallel fibrils (18,19). The lead cell then migrates along the newly created microtracks into the ECM. Additional cells subsequently invade by following the microtracks created by the lead cell (20). A trailblazer EMT confers cells with an enhanced ability to promote the formation of microtracks (12,21). Notably, cells that promote microtack generation induce the collective invasion of “opportunist” siblings that have not undergone a trailblazer EMT and have a relatively diminished intrinsic invasive ability (12,22,23). The interaction between trailblazer and opportunist cells can promote the progression from ductal carcinoma in situ (DCIS) to invasive breast cancer (IBC) in an orthotopic tumor model (21). Consistent with results obtained with these experimental models, the interaction of distinct subpopulations is observed during the initial transition from DCIS to IBC in breast cancer patients (24,25). Trailblazer cells are also capable of metastasizing to the lung (12). Signaling programs that are specifically required for trailblazer cells to initiate collective invasion are required for metastasis and their expression correlates with poor breast cancer patient outcome (12,26), underscoring the critical function of trailblazer subpopulations during tumor progression.
The trailblazer EMT program includes genes that are specifically required for cells to iniOate collective invasion, but are dispensable for other types of cell movement, including migration along two-dimensional surfaces (12,26). One of these trailblazer EMT genes is the guanine nucleotide exchange factor (GEF) DOCK10, which activates Cdc42 and N-Wasp to promote actin polymerization and protrusion formation. Others include PDGFRA, VASN and ITGA11 (12). Thus, the functional programs that directly contribute to the unique properties of a trailblazer EMT have begun to be defined. By comparison, the signaling programs that iniOate a trailblazer EMT are poorly understood. To address this question we defined factors that promote a trailblazer EMT in the C3TAg genetically engineered mouse model of breast cancer (27). The tumors that develop in this model are molecularly similar to basal type breast cancer, which is frequently classified as triple-negative (no estrogen receptor, progesterone receptor or HER2 expression) (28,29). Through analysis of protein expression in primary tumors and functional testing in freshly derived tumor organoids, cultured organoids and primary cell lines we discovered that TGFβ induces a trailblazer EMT in hybrid cells with canonical mesenchymal and epithelial traits. Importantly, expression of the transcription factor Fra1 is necessary for this trailblazer EMT. Fra1 is not regulated by TGFβ pathway activity. Instead, the autocrine activation of EGFR and ERK1/2 is sufficient to promote Fra1 expression and permit a trailblazer EMT. Together, these results reveal that the integration of parallel signaling pathways coordinates the induction of a trailblazer EMT.
METHODS
Cell Culture
Cells were generated from C3TAg tumors as indicated, cultured as described (12,30) and verified by Powerplex genotyping. Cells were tested for mycoplasma (Lonza, LT07-703) prior the creation of freezedown stocks. Cells from these frozen stocks were routinely used within 25 passages. Tumor organoids were propagated in Organoid Media (DMEM/F12 (Corning, 10-092-CV), B27 (Fisher, 17504001), ITS (Lonza, 17-838Z), non-essenOal amino acids (Sigma, M7145), 10 ng/ml FGF2 (Peprotech, 100-18C) 50 ng/ml EGF (Peprotech, AF-100-15), Y27632 (Peprotech, 1293823) N-acetylcysteine (MP Biomedicals, 194603) and A83-01 (Peprotech, 9094360) (Table S1). Organoids were passaged at least once per week by dissociating tumor organoids with Dispase (Sigma, SCM133) and TryPLE (Gibco, 12605-010) into single cell suspensions. The cell suspensions in plated at a density of 200,000-500,000 cells per well in a 24-well ultra low adhesion plate in Organoid Media. Virus was produced and cells were infected to generate stable cell lines as described (31).
3D culture experiments
Growth factor reduced Matrigel (Corning, 10-12 mg/ml stock concentration, #354230) and rat tail (Corning, #354236) collagen I were used for invasion assay experiments. Vertical invasion assays and experiments in 3D culture were performed and quanOfied as described previously using a Matrigel/Collagen I matrix (3-5 mg/ml Matrigel and 1.8-2.1 mg/ml Collagen I) (12,26). A 120 μm span on the z-axis is shown for the vertical invasion assays. For spheroid cluster experiments employing cell lines, 96-well Nunclon Sphera low adhesion plates (Thermo Scientific, 174925) were used to form clusters. 1000 cells per well were plated and incubated at 37°C for 24-72 h. Clusters were then resuspended in 30 μl of Matrigel/Collagen I mix and plated on 20 μl of a basement layer of Matrigel/Collagen I and allowed to invade for 24 h. For 1339 organoid invasion assays, 200,000-500,000 cells were plated per well in a 24-well ultra low adhesion plate in Organoid Media. The cells were allowed to clump overnight. Clumps were then resuspended in 30 μl of Matrigel/Collagen I mix and plated on 20 μl of basement layer of Matrigel/Collagen I and allowed to invade for 24 h-48 h.
Immunoblot analysis, IF and IHC
Experiments and analysis were performed as described (12) using antibodies detailed in Table S2.
siRNA experiments
Cells were transfected with 50 nM of siRNA using RNAiMax transfection reagent (Invitrogen) for 48-72 h. The siRNAs were from Dharmacon and Sigma. Cells in all conditions designated as “Control” were transfected with a pool of siRNAs that dties not target human genes. The details of the sequences and catalog numbers for each siRNA are located in Table S3.
Quantitative real-time PCR
Experiments were performed as described (12) using primer sequences listed in Table S4.
Mice
The C3-Tag [FVB-Tg(C3-1-TAg)cJeg/JegJ] mice (27) were a gir from Anna Riegel. Mice were housed and bred in accordance with a protocol approved by the Institutional Animal Use and Care Commisee at Georgetown University in compliance with the NIH Guide for the Care and Use of Laboratory animals. Female mice were used for all analyses.
Tumor explants
The largest tumors from female C3TAg mice were minced and tumors were digested for up to 120 min at 37°C in a mixture of 1mg/ml Collagenase, 2U/μl DNase, 5% FBS in DMEM/F12. Digested tumors were pelleted at 80 × g for 1 min and the supernatant was discarded. Tumor organoids were then rinsed up to 5 Omes in 10 ml 5% FBS in DMEM/F12. Organoids were plated in a mixture of 30 ul of 2.4 mg/ml Rat tail collagen (354236, Corning) and 3 mg/ml growth factor reduced Matrigel onto a base layer of 20 μl of Collagen I/Matrigel. The organoids in ECM were overlaid with DMEM/F12 supplemented with 1% FBS, 1X ITS (Sigma, I3146), non-essential amino acids (Sigma, M7145) and 10 ng/ml FGF2 (Peprotech, 100-18C). Organoids were allowed to invade for 48 h, fixed and imaged unless otherwise indicated (12).
Statistical Methods
Data with a normal distribution determined by Shapiro-Wilk test were analyzed by two tailed Student’s t-test (Graphpad Prism). Data that did not pass a normality test were analyzed by Mann-Whitney U test.
RESULTS
K14-high and K14-low C3Tag tumor cells lead collective invasion
To understand how a trailblazer state is induced we investigated the invasive properties of C3TAg tumor organoids. Consistent with previous observations (21,32), there was heterogeneity in the invasive behavior of organoids derived from C3TAg tumors (Fig. 1A). Intrinsic variability in invasive properties was also observed arer tumors were dissociated into smaller cellular clusters (<5 cells) and allowed to grow in 3D culture for 4 days, with many actively growing organoids having less invasive character, indicating that diminished fitness was not the cause of reduced invasion (Fig. 1A and S1A). Clonal populations of B6 and D6 cells derived from a C3TAg tumor also had distinct invasive properties (Fig. S1B). Given this variability we next determined how cells with different intrinsic invasive properties interacted with each other. InteresOngly, the noninvasive B6 cells were induced to opportunistically invade behind invasive cells from fresh C3TAg tumors, or the sibling invasive D6 clone (Fig. 1B and S1B). This indicated that the invasive cells in C3TAg tumors and invasive clonal populations can function as trail-blazer cells that lead the invasion of intrinsically less invasive opportunist siblings. Indeed, the noninvasive B6 clone expressed the transcription factor DNp63 (Fig. S1C), which is detected in opportunist cells derived from human cell lines. DNp63 expressing cells also opportunistically invade in C3TAg tumors and C3TAg tumor explants (21).
A) Model depicting tumor processing and analysis of organoids. Representative images show C3TAg organoids 24 h arer embedding in ECM. Dissociated cells were analyzed arer 4 days of growth. Violin plot shows quantification of circularity (Mann Whitney test, n=2 tumors). B) Representative images of multicellular clusters of the C3TAg-derived noninvasive B6 clone (H2B:mCherry) and C3TAg cells from fresh tumors. Cells were clustered in low adhesion wells alone or at a 1:1 ratio prior to embedding in ECM (n=3 tumors). C) K14 (magenta) and K8 (cyan) expression in C3TAg organoids. Graph shows quantification of keratin expression in the leading invasive cells (mean±SEM, n=3 tumors, unpaired Student’s t test). D) K14 (magenta) and K8 (cyan) expression in C3TAg primary tumors. E) Invasion of 1863 spheroids arer K14 and K8 depletion. Graph shows area of invading nuclei indicated by the image mask (mean±SEM, n=2, unpaired Student’s t test). Scale bars, 50 μm.
The activation of a basal gene expression program exemplified by the expression of K14 correlates with invasion in the PyMT GEM model of luminal B type breast cancer (32). Invading cells in the C3TAg model express K14 (Fig. 1C). However, we found that K14 expression alone did not correlate with trailblazer behavior in C3T organoids (Fig. 1C). This was indicated by invasion being led by K8-high/K14-low cells in explants and primary tumors (Fig. 1D). Moreover, depletion of K14 from 1863 cells derived from a C3TAg tumor only modestly reduced the extent of invasion, indicating that high K14 expression is not essenOal for C3TAg trailblazer cells to lead collective invasion (Fig. 1E). Thus, our data indicated that additional signaling programs beyond the control of K14 expression conferred a trailblazer EMT in C3TAg tumor cells.
A TGFβ1 induced hybrid EMT confers C3Tag cells with the ability to lead invasion
To test how EMT state influenced the invasive properties of C3Tag tumor cells, we determined how TGFβ1, a canonical inducer of EMT (9), influenced the invasive behavior the 1339 organoid line derived from a C3Tag tumor (Fig. 2A). Soluble TGFβ1 enhanced the extent of collective invasion of 1339 organoids within 48 h of treatment (Fig. 2A). In addition, tumors that develop from 1339 organoids arer orthotopic injection were more invasive than the 1339 organoid line itself, consistent with TGFβ1 in the microenvironment inducing a trailblazer EMT in vivo (Fig. S2A). Inhibitors of a receptor for TGFβ1, TGF-BR1, blocked the TGFβ1 induced trailblazer EMT in 1339 organoids (Fig. 2A). A cell line (1863) also retained epithelial properties when established with a TGFBR1 inhibitor in the culture media, whereas cells derived from the same tumor (1863T) lost E-cadherin expression and displayed increased Vimentin expression, further indicated the ability of TGFβ1 to induce a trailblazer EMT (Fig. S2B). A similar adoption of canonical mesenchymal features was observed in the trailblazer D6 clone that was also derived in the absence of a TGFBR1 inhibitor (Fig. S2C). Brief exposure of 1 h to TGFβ1 is sufficient to confer a stable EMT state through a switch to a Ecadherin-low/Vimentin high state and makes cells independent of TG-BFR activity (33). However, TGBFR1 inhibition suppressed the basal invasion of 1339 organoids and the invasion of C3TAg tumor explants, showing that TGBFR1 activity is required to sustain a trailblazer state (Fig. 2A and B). These results indicated that trailblazer state in C3TAg tumor populations did not require a stable switch to fully mesenchymal state or independence from TGFBR1 activity. TGFβ1 induces gene expression changes through promtiong the TGBFR1 dependent phosphorylation of the transcription factors SMAD2 and SMAD3 (34). The TGFβ1 induced invasion of 1863 cells required SMAD3 expression, indicating that TGFβ1 regulated gene expression program contributed to the observed trailblazer EMT (Fig. 2C and S2D). SMAD3 depletion also reduced the invasion of mesenchymal 1863T cells, with similar results observed using 2 disOnct siRNAs, showing that these cells also retained a requirement for TGFβ1 pathway activity (Fig. S2E and F). VimenOn mRNA expression in 1339 organoids was increased within 48 h of TGFβ1 treatment, indicating that there was mesenchymal expression program activation (Fig. 2D). However, cells retained E-cadherin expression and canonical EMT-TF expression was unchanged (Fig. 2D). These results indicated that invading cells were in a hybrid EMT state. Analysis of C3T primary tumors showed that Vimentin expressing cells leading invasion at the tumor stromal boundary, indicaOng that EMT activation can occurs in collectively invading cells in the C3T primary tumors (Fig. 2E).
A) Work-flow of the development of organoid lines from C3TAg tumors. The TGFBR1 inhibitors A83-01 (500 nM) and SB43152 (1 uM) suppress both intrinsic and TGFβ1 induced invasion of organoid lines. Violin plot shows quantification of circularity (Mann Whitney test, n=2 for SB43152, n>4 for A83-01). B) Model showing C3TAg tumor processing and treatment. Representative images show that A83-01 reduces the invasion of C3TAg organoids. Violin plot shows area of invading nuclei (Mann Whitney test, n=3 tumors). C) SMAD3 depletion suppresses the invasion of 1863 spheroids (mean±SEM, n=2, unpaired Student’s t test). D) qPCR results showing that TGFβ1 (2ng/mL) treatment for 48 hrs induces VimenOn while expression of additional canonical EMT markers was unaltered (mean±SD, n=2). E) Vimetin expression (magenta) in C3TAg primary tumor cells indicated by SV40 (cyan) and K8 (white) expression. Scale bars, 50 μm.
Fra1 and ZEB1/2 expression promote a trailblazer EMT state
To understand factors that control trailblazer EMT programs we next determined how canonical EMT regulatory factors influenced invasion in C3TAg tumors. TGFβ1 induced EMTs can involve suppression of miR200 family miRNAs (33). Indeed, transfection of 1863 and 1863T cells dramaOcally suppressed the invasion of spheroids and vertical invasion into the ECM from a cell monolayer (Fig. 3A and S3A and B). miR200 suppresses ZEB1/2 expression in 1863T cells (Fig. 3B), consistent with previous results in other models (35). Thus we further tested the potential function for ZEB1 and ZEB2 transcription factors in regulating the trailblazer state in cells derived from C3Tag tumors. Indeed, the combined depletion of both ZEB1/2 reduced 1863 spheroid invasion, consistent with the diminished invasion detected in cells in which ZEB1/2 are suppressed by exogenous miR200 (Fig. 3C and 3D). By comparison, siR-NAs targeting other canonical EMT TFs did not alter invasive behavior (FIg. S3C). TGFβ1 signaling is integrated with other signaling networks, including programs that are regulated by AP-1 family transcription factors, which collaborate to regulate ZEB1/2 expression. Depletion of Fra1 reduced the TGFβ1 induced the invasion of 1863 spheroids (Fig. 3E and F). Four disOnct individual siRNAs and a second siRNA pool targeOng Fra1 and a also suppressed the vertical invasion of 1863T cells (Fig. 3F, S3D and E). Fra1 depletion reduced the invasion of D6 cells as well (Fig. S3F). Fra1 interacts with c-Jun to regulate gene expression as an AP-1 complex (36). Indeed, c-Jun was required for the invasion of 1863T cells (Fig. S3G), suggesting Fra1 containing AP-1 complexes regulate the expression of genes necessary for trailblazer invasion. However, depletion of Fra1 did not alter ZEB1/2 expression, indicating that Fra1 regulated the trail-blazer EMT state in independent of controlling ZEB1/2 expression in these cells (Fig. 3G). Interestingly, Fra1 expression in 1339 organoids was not regulated by acute TGFβ1 stimulation and and was not dependent on TGFBR1 activity (Fig. 3H and I). This indicated that the induction of Fra1 by a parallel signaling pathway is necessary for TGFβ1 to induce a trailblazer EMT state.
A) Transfection of a miR200c-3p mimic decreases invasion of 1863 spheroids (mean±SD, n=2, unpaired Student’s t test). B) qPCR showing that miR200c-3p suppresses Zeb1 and Zeb2 mRNA expression (mean±SEM, n=2). C) Simultaneous depletion of Zeb1 and Zeb2 reduces the invasion of 1863 spheroids (mean±SEM, n=3). D) qPCR showing the depletion of Zeb1 and Zeb2 arer siRNA transfection (mean±SEM, n=2). E) Fra1 depletion decreases TGFβ1 induced invasion of 1863 spheroids. Graphs shows the area of invading nuclei (mean±SEM, n=5, unpaired Student’s t test). F) Fra1 mRNA is depleted by two separate Fosl1 siRNA pools (mean±SEM, n=3 for pool 1 and n=2 for pool 2). G) qPCR showing that Fra1 depletion dties not reduce Zeb1 and Zeb2 expression (mean±SD, n=2). H) qPCR showing that TGFβ1 treatment (2ng/mL for 48 hrs) dties not increase Fra1 expression in 1339 organoids (mean±SD, n=2). I) qPCR showing that the TGFBR1 inhibitor A83-01 does not reduce Fra1 expression (mean±SD, n=2). Scale bars, 50 μm.
EGFR and ERK1/2 activity are necessary for Fra1 expression and the induction of trailblazer EMT
We next determined how Fra1 expression was regulated in C3Tag tumor cells. ERK1/2 can promote Fra1 transcription and phosphorylation dependent stabilization (37). Treatment with trame-Onib, an inhibitor of MEK1/2, the upstream activators of ERK1/2, reduced Fra1 expression in 1863 cells (Fig. 4A). Inhibition of MEK1/2 did not alter the phosphorylation of SMAD2 or ZEB1 expression, showing that ERK1/2 specifically regulates Fra1 expression. Consistent with these results, MEK1/2 inhibition with trameOnib reduced TGFβ1 induced invasion of 1339 organoids (Fig. 4B). Notably, the clinically approved MEK1/2 inhibitor Trametinib inhibited the invasion of explants derived from 4 different tumors (Fig. 4C). Consistent with a requirement of ERK1/2 for invasion, active phosphorylated ERK1/2 was detected in collectively invading cells in C3TAg primary tumors (Fig. 4D). ERK1/2 activity is cyclical in response to the activation of negative feedback loops (38). Thus, persistent ERK1/2 phosphorylation is not necessarily expected in all collectively invading cells. HER2 amplification can activate ERK1/2 and synergize with TGFβ1 signaling to promote invasion (39). However, C3TAg tumors do not have amplified HER2 expres-sion (28). Autocrine activation of growth factor receptors can promote ERK1/2 phosphorylation (40). Indeed, the EGFR inhibitor Erltionib suppressed Fra1 expression (Fig. 4E). Erltionib also suppressed the TGFβ1 induced invasion of 1339 organoids, indicaOng that autocrine EGFR activity and promoted the Fra1 expression that was necessary for invasion (Fig. 4F).
A) Upper immunoblot shows Fra1 and phosphorylated Fra1 expression in 1863 cells treated with 50 nM of the MEK1/2 inhibitor Trametinib. ERK1/2 phosphorylation is reduced upon treatment. Lower immunoblot shows that Trametinib treatment does not reduce Zeb1 expression or SMAD2 phosphorylation (n=2). B) Representative images showing that Trametinib (50 nM) treatment suppresses the TGFβ1 induced invasion of 1339 organoids. Violin plot shows circularity quantification (Mann Whitney test, n=4). C) Representative images showing that Trametinib (50 nM) treatment suppresses the invasion of fresh C3Tag tumor organoids. Violin plot shows quantification of area of invading nuclei (Mann Whitney test, n=4 mice). D) Phosphorylated ERK1/2 (magenta) in C3TAg primary tumor cells indicated by SV40 (cyan) and K8 (white) expression. Scale bars, 50 μm. E) qPCR showing that inhibition of EGFR (Erltionib) and ERK1/2 activity (TrameOnib) suppresses Fra1 expression (mean±SD, n=2). F) Treatment with the EGFR inhibitor Erltionib (1 uM) reduces TGFβ1 induced invasion of 1339 organoids. Violin plot shows circularity quantification (Mann Whitney test, n=4). Scale bars, 50 μm.
DISCUSSION
Through investigating the requirements for invasion in primary tumor cells, we have revealed the molecular underpinnings of a hybrid EMT program that confers tumor cells with the ability to iniOate collective invasion. It is well established that EMT programs confer tumor cells with invasive properties (41). One of the defining asributes of EMTs is the loss of epithelial character (42). This shir from an epithelial phenotype includes the reduced expression of certain cell-cell adhesion proteins (43). The loss of cell-cell cohesion facilitates the detachment of single cells, which can migrate through the ECM utilizing proteolytic and force dependent mechanisms (44). While conceptually straigh{orward, a model centered on the highly mesenchymal tumor populations invading as single cells dties not account for the invasive properties of the majority of invasive populations in primary tumors (31). Tumor cells predominantly engage in multicellular collective invasion and retain epithelial traits, including E-cadherin expression. Indeed, E-cadherin expressing cells are fully capable of collective invasion (30,45). These observations have contributed to a revised understanding of EMTs as being variable in their molecular properties as opposed to a single program that ultimately leads to a fully mesenchymal state (46). This recognition of the existence of hybrid EMTs has created a new challenge of defining their function in tumors. Our results show that there is a hybrid trailblazer EMT state. Analysis of human cancer cell lines had shown a correlation with a full mesenchymal state (12,18). However, there was no indication that the loss of epithelial character was necessary for a trailblazer EMT. Thus, the specific asributes of trailblazer EMTs has remained unclear. Our results suggest that the unique properties of a trailblazer EMT that promote ECM reorganization can be acquired while cells retain epithelial character. Thus, our results support future investigations focused on regulatory processes that support features of collective invasion rather than only relying on canonical EMT markers to define the contribution of EMTs to invasive cell phenotypes.
The induction of a trailblazer EMT in the C3TAg model promoted the invasion of cells with low or undetectable levels of K14 expression. K14 expression is a hallmark of collectively invading cells in the PyMT GEM model of breast cancer and K14 positive cells collectively invade in breast cancer patient tumors (32). Consistent with our observations here, we have found that K14 expression is not required for human trailblazer breast cancer cells to collectively invade (12). This suggests that the specific requirements for a trailblazer EMT may depend on cell intrinsic properties that vary between tumor type. We found that expression of the transcription factor Fra1 was necessary for a trailblazer EMT in cells derived from C3TAg tumors. In addition, inhibition of the autocrine signaling pathway that promotes Fra1 expression blunts the TGFβ1 induced trailblazer EMT in cultured C3TAg organoids and the invasion of fresh organoids from primary tumors. TGFβ1 signaling did not modulate Fra1 expression in C3TAg tumor cells. Thus, the control of Fra1 was a cell intrinsic property in the context in which we evaluated collective invasion. Whether Fra1 controls a regulatory program distinct from the program active in K14 expressing cells, and thus providing an alternative type of trailblazer EMT is an interesting line of future investigation.
In summary, we have found that the integration of TGFβ1 and EGFR-ERK1/2-Fra1 pathways is necessary for a trailblazer EMT program. Our results suggest that disruption of either pathway is sufficient to blunt the collective invasion of tumor organoids. Thus, our results suggest the potential use of existing therapeutic agents to limit breast cancer cell collective invasion.
ACKNOWLEDGEMENTS
Work was supported by NIH R01CA218670, Georgetown Women and Wine (G. Pearson), NIH T32CA009686 (S. Camacho) and NIH P30CA051008 for shared resources (G. Pearson).
Footnotes
The authors declare no potential conflicts of interest.
