Search

Live imaging of explants containing MCFDCISO cells (H2B:mCherry, red, nuclei) with or without 578T cells (H2B:GFP, green, nuclei) showing cell motility.

Defining how interactions between tumor subpopulations contribute to invasion is essential for understanding how tumors metastasize. Here, we find that the heterogeneous expression of the transcription factor ΔNp63 confers distinct proliferative and invasive epithelial-to-mesenchymal transition (EMT) states in subpopulations that establish a leader–follower relationship to collectively invade. A ΔNp63-high EMT program coupled the ability to proliferate with an IL1α- and miR-205–dependent suppression of cellular protrusions that are required to initiate collective invasion. An alternative ΔNp63-low EMT program conferred cells with the ability to initiate and lead collective invasion. However, this ΔNp63-low EMT state triggered a collateral loss of fitness. Importantly, rare growth-suppressed ΔNp63-low EMT cells influenced tumor progression by leading the invasion of proliferative ΔNp63-high EMT cells in heterogeneous primary tumors. Thus, heterogeneous activation of distinct EMT programs promotes a mode of collective invasion that overcomes cell intrinsic phenotypic deficiencies to induce the dissemination of proliferative tumor cells.Significance:These findings reveal how an interaction between cells in different EMT states confers properties that are not induced by either EMT program alone.

Supplemental Figures S1-S5 and Supplemental Legends

Contains Supplementary Figures S1-S7, Supplementary Figure Legends and Supplementary Methods.

Supplementary tables related to miRNA and mRNA expression analysis.

Supplementary tables related to reagents used.

Defining how interactions between tumor subpopulations contribute to invasion is essential for understanding how tumors metastasize. Here, we find that the heterogeneous expression of the transcription factor ΔNp63 confers distinct proliferative and invasive epithelial-to-mesenchymal transition (EMT) states in subpopulations that establish a leader–follower relationship to collectively invade. A ΔNp63-high EMT program coupled the ability to proliferate with an IL1α- and miR-205–dependent suppression of cellular protrusions that are required to initiate collective invasion. An alternative ΔNp63-low EMT program conferred cells with the ability to initiate and lead collective invasion. However, this ΔNp63-low EMT state triggered a collateral loss of fitness. Importantly, rare growth-suppressed ΔNp63-low EMT cells influenced tumor progression by leading the invasion of proliferative ΔNp63-high EMT cells in heterogeneous primary tumors. Thus, heterogeneous activation of distinct EMT programs promotes a mode of collective invasion that overcomes cell intrinsic phenotypic deficiencies to induce the dissemination of proliferative tumor cells. Significance: These findings reveal how an interaction between cells in different EMT states confers properties that are not induced by either EMT program alone.

Epithelial-to-mesenchymal transition (EMT) contributes to tumor cell survival, immune evasion, migration, invasion, and therapy resistance. Across human cancer, tumors that are high grade, poorly differentiated, and have undergone EMT carry a worse prognosis with a higher likelihood of metastasis. AXL, a receptor tyrosine kinase, drives EMT and is implicated in tumor progression, metastasis, and therapy resistance in multiple cancer types including pancreatic cancer and breast cancer. TANK-binding kinase 1 (TBK1) is central to AXL-driven EMT yet, the mechanism of how TBK1 induces EMT remains unclear. Here, we report that AXL activation stimulates TBK1 binding and phosphorylation of AKT3. TBK1 activation of AKT3 drives binding and phosphorylation of slug/snail resulting in protection from proteasomal degradation and translocation of the complex into the nucleus. We show that nuclear translocation of AKT3 is required for AXL-driven EMT and metastasis. Congruently, nuclear AKT3 expression correlates with worse outcome in aggressive breast cancer. To advance AKT3 as a therapeutic target, an AKT3-isoform selective allosteric small molecule inhibitor, BGB214, was developed. BGB214 inhibits AKT3 nuclear translocation, EMT-TF stability, AKT3-mediated invasion of breast cancer cells and reduces tumor initiation in vivo. Our results suggest that AKT3 nuclear activity is an important feature of AXL-driven epithelial plasticity and that selective AKT3 inhibition represents a novel therapeutic avenue for treating aggressive cancer.SignificanceNuclear AKT3 activity is an important feature of AXL-TBK1 driven EMT and metastasis, thus selective AKT3 targeting represents a novel approach to treat aggressive cancer.

BACKGROUND Postpartum hemorrhage remains one of the largest causes of maternal morbidity and mortality in the United States. OBJECTIVE The aim of this paper is to use machine learning techniques to identify patients at risk for postpartum hemorrhage at obstetric delivery. METHODS Women aged 18 to 55 years delivering at a major academic center from July 2013 to October 2018 were included for analysis (N=30,867). A total of 497 variables were collected from the electronic medical record including the following: demographic information; obstetric, medical, surgical, and family history; vital signs; laboratory results; labor medication exposures; and delivery outcomes. Postpartum hemorrhage was defined as a blood loss of ≥1000 mL at the time of delivery, regardless of delivery method, with 2179 (7.1%) positive cases observed. Supervised learning with regression-, tree-, and kernel-based machine learning methods was used to create classification models based upon training (21,606/30,867, 70%) and validation (4630/30,867, 15%) cohorts. Models were tuned using feature selection algorithms and domain knowledge. An independent test cohort (4631/30,867, 15%) determined final performance by assessing for accuracy, area under the receiver operating curve (AUROC), and sensitivity for proper classification of postpartum hemorrhage. Separate models were created using all collected data versus models limited to data available prior to the second stage of labor or at the time of decision to proceed with cesarean delivery. Additional models examined patients by mode of delivery. RESULTS Gradient boosted decision trees achieved the best discrimination in the overall model. The model including all data mildly outperformed the second stage model (AUROC 0.979, 95% CI 0.971-0.986 vs AUROC 0.955, 95% CI 0.939-0.970). Optimal model accuracy was 98.1% with a sensitivity of 0.763 for positive prediction of postpartum hemorrhage. The second stage model achieved an accuracy of 98.0% with a sensitivity of 0.737. Other selected algorithms returned models that performed with decreased discrimination. Models stratified by mode of delivery achieved good to excellent discrimination but lacked the sensitivity necessary for clinical applicability. CONCLUSIONS Machine learning methods can be used to identify women at risk for postpartum hemorrhage who may benefit from individualized preventative measures. Models limited to data available prior to delivery perform nearly as well as those with more complete data sets, supporting their potential utility in the clinical setting. Further work is necessary to create successful models based upon mode of delivery and to validate the findings of this study. An unbiased approach to hemorrhage risk prediction may be superior to human risk assessment and represents an area for future research.

Pancreatic ductal adenocarcinoma (PDA), a leading cause of cancer-related death in the United States, has a high metastatic rate, and is associated with persistent immune suppression. AXL, a member of the TAM (TYRO3, AXL, MERTK) receptor tyrosine kinase family, is a driver of metastasis and immune suppression in multiple cancer types. Here we use single-cell RNA-sequencing to reveal that AXL is expressed highly in tumor cells that have a mesenchymal-like phenotype and that AXL expression correlates with classic markers of epithelial-to-mesenchymal transition. We demonstrate that AXL deficiency extends survival, reduces primary and metastatic burden, and enhances sensitivity to gemcitabine in an autochthonous model of PDA. PDA in AXL-deficient mice displayed a more differentiated histology, higher nucleoside transporter expression, and a more active immune microenvironment compared with PDA in wild-type mice. Finally, we demonstrate that AXL-positive poorly differentiated tumor cells are critical for PDA progression and metastasis, emphasizing the potential of AXL as a therapeutic target in PDA. Implications: These studies implicate AXL as a marker of undifferentiated PDA cells and a target for therapy.

BackgroundPostpartum hemorrhage remains one of the largest causes of maternal morbidity and mortality in the United States.ObjectiveTo utilize machine learning techniques to identify patients at risk for postpartum hemorrhage at obstetric delivery.Study DesignWomen aged 18 to 55 delivering at a major academic center from July 2013 to October 2018 were included for analysis (n = 30,867). A total of 497 variables were collected from the electronic medical record including demographic information, obstetric, medical, surgical, and family history, vital signs, laboratory results, labor medication exposures, and delivery outcomes. Postpartum hemorrhage was defined as a blood loss of 1000 mL at the time of delivery, regardless of delivery method, with 2179 positive cases observed (7.06%).Supervised learning with regression-, tree-, and kernel-based machine learning methods was used to create classification models based upon training (n = 21,606) and validation (n = 4,630) cohorts. Models were tuned using feature selection algorithms and domain knowledge. An independent test cohort (n = 4,631) determined final performance by assessing for accuracy, area under the receiver operating curve (AUC), and sensitivity for proper classification of postpartum hemorrhage. Separate models were created using all collected data versus limited to data available prior to the second stage of labor/at the time of decision to proceed with cesarean delivery. Additional models examined patients by mode of delivery.ResultsGradient boosted decision trees achieved the best discrimination in the overall model. The model including all data mildly outperformed the second stage model (AUC 0.979, 95% CI 0.971–0.986 vs. AUC 0.955, 95% CI 0.939–0.970). Optimal model accuracy was 98.1% with a sensitivity of 0.763 for positive prediction of postpartum hemorrhage. The second stage model achieved an accuracy of 98.0% with a sensitivity of 0.737. Other selected algorithms returned ≥ models that performed with decreased discrimination. Models stratified by mode of delivery achieved good to excellent discrimination, but lacked sensitivity necessary for clinical applicability.ConclusionsMachine learning methods can be used to identify women at risk for postpartum hemorrhage who may benefit from individualized preventative measures. Models limited to data available prior to delivery perform nearly as well as those with more complete datasets, supporting their potential utility in the clinical setting. Further work is necessary to create successful models based upon mode of delivery. An unbiased approach to hemorrhage risk prediction may be superior to human risk assessment and represents an area for future research.CondensationMachine learning methods can be successfully utilized to predict nearly three-quarters of women at risk of postpartum hemorrhage when undergoing obstetric delivery.AJOG at a GlanceWhy was the study conducted? To determine patients at risk for postpartum hemorrhage using modern machine learning techniques on a robust data set directly derived from the electronic medical recordWhat are the key findings? Using 28 predictor features, the model successfully classified 73.7% of patients who ultimately had a postpartum hemorrhage using information available prior to deliveryMany previously identified risk factors for postpartum hemorrhage were not included in the final model, potentially discounting their contribution to hemorrhage riskModels stratified by delivery method achieved good to excellent discrimination but noted lower sensitivity and need further investigationWhat does this study add to what is already known? This study represents the largest cohort directly-derived from the electronic medical record to use machine learning techniques to identify patients at risk for postpartum hemorrhage

BACKGROUND Vasa previa represents a rare prenatal finding with potentially life-threatening risk to the fetus. OBJECTIVE This study aimed to describe the natural history of prenatally diagnosed vasa previa and evaluate the association between antenatally diagnosed vasa previa and adverse obstetrical and neonatal outcomes. STUDY DESIGN This was a multicenter descriptive and retrospective study of patients diagnosed prenatally with vasa previa on transvaginal ultrasound in the New York City Maternal-Fetal Medicine Research Consortium centers between 2012 and 2018. Outcomes evaluated included persistence of vasa previa at the time of delivery, gestational age at delivery, indications for unplanned unscheduled delivery, and neonatal course. RESULTS A total of 165 pregnancies with vasa previa were included, of which 16 were twin gestations. Forty-three cases (26.1%) were noted to resolve on subsequent ultrasound. Of the remaining 122 cases with persistent vasa previa, 46 (37.7%) required unscheduled delivery. Twin gestations were nearly 3 times as likely to require unscheduled delivery as singleton gestations (73.3% vs 25.2%; P

Cancer/testis (CT) antigens are proteins whose expression is normally restricted to germ cells yet aberrantly activated in tumors, where their functions remain relatively cryptic. Here we report that ZNF165, a CT antigen frequently expressed in triple-negative breast cancer (TNBC), associates with SMAD3 to modulate transcription of transforming growth factor β (TGFβ)-dependent genes and thereby promote growth and survival of human TNBC cells. In addition, we identify the KRAB zinc finger protein, ZNF446, and its associated tripartite motif protein, TRIM27, as obligate components of the ZNF165-SMAD3 complex that also support tumor cell viability. Importantly, we find that TRIM27 alone is necessary for ZNF165 transcriptional activity and is required for TNBC tumor growth in vivo using an orthotopic xenograft model in immunocompromised mice. Our findings indicate that aberrant expression of a testis-specific transcription factor is sufficient to co-opt somatic transcriptional machinery to drive a pro-tumorigenic gene expression program in TNBC.

Significance Uterine carcinosarcoma (UCS) is an aggressive endometrial cancer variant distinguished from endometrial adenocarcinoma (EC) by admixed malignant epithelial and mesenchymal components (carcinoma and sarcoma). The molecular events underlying UCS are enigmatic, as cancer gene mutations are generally shared among UCS/EC. We take advantage of genetic approaches in mice to show that inactivation of Fbxw7 and Pten results in UCS through spontaneous acquisition of mutations in a third gene (Tp53), arguing for strong biological selection and synergism in UCS. We used this UCS model including tumor-derived cell lines to show that Fbxw7 loss drives epithelial–mesenchymal transition, explaining Fbxw7’s role in UCS. This model system argues that simultaneous genetic defects in 3 distinct pathways (Fbxw7, Pten/PI3K, Tp53) converge in UCS genesis., Uterine carcinosarcoma is an aggressive variant of endometrial carcinoma characterized by unusual histologic features including discrete malignant epithelial and mesenchymal components (carcinoma and sarcoma). Recent studies have confirmed a monoclonal origin, and comprehensive genomic characterizations have identified mutations such as Tp53 and Pten. However, the biological origins and specific combination of driver events underpinning uterine carcinosarcoma have remained mysterious. Here, we explored the role of the tumor suppressor Fbxw7 in endometrial cancer through defined genetic model systems. Inactivation of Fbxw7 and Pten resulted in the formation of precancerous lesions (endometrioid intraepithelial neoplasia) and well-differentiated endometrioid adenocarcinomas. Surprisingly, all adenocarcinomas eventually developed into definitive uterine carcinosarcomas with carcinomatous and sarcomatous elements including heterologous differentiation, yielding a faithful genetically engineered model of this cancer type. Genomic analysis showed that most tumors spontaneously acquired Trp53 mutations, pointing to a triad of pathways (p53, PI3K, and Fbxw7) as the critical combination underpinning uterine carcinosarcoma, and to Fbxw7 as a key driver of this enigmatic endometrial cancer type. Lineage tracing provided formal genetic proof that the uterine carcinosarcoma cell of origin is an endometrial epithelial cell that subsequently undergoes a prominent epithelial–mesenchymal transition underlying the attainment of a highly invasive phenotype specifically driven by Fbxw7.

Simple Summary Pancreatic cancer is a challenging disease to treat effectively. Fibroblasts associated with pancreatic cancer contribute to disease progression by secreting factors that enhance tumor cell survival and help tumor cells avoid detection by the immune system. This overview focuses on a chemokine, CXCL12, produced by cancer-associated fibroblasts and how CXCL12 signaling enhances pancreatic cancer progression by contributing to various hallmarks of cancer including, but not limited to, tumor growth and evasion of immune response. These pro-oncogenic functions of CXCL12 make it an attractive target in pancreatic cancer. We discuss the different approaches in development to therapeutically target CXCL12 and finally propose a novel approach, the use of the farnesyl transferase inhibitor tipifarnib to inhibit CXCL12 expression in pancreatic fibroblasts. Abstract Pancreatic ductal adenocarcinoma (PDAC) is a disease with limited therapeutic options and dismal long-term survival. The unique tumor environment of PDAC, consisting of desmoplastic stroma, immune suppressive cells, and activated fibroblasts, contributes to its resistance to therapy. Activated fibroblasts (cancer-associated fibroblasts and pancreatic stellate cells) secrete chemokines and growth factors that support PDAC growth, spread, chemoresistance, and immune evasion. In this review, we focus on one such chemokine, CXCL12, secreted by the cancer-associated fibroblasts and discuss its contribution to several of the classical hallmarks of PDAC and other tumors. We review the various therapeutic approaches in development to target CXCL12 signaling in PDAC. Finally, we propose an unconventional use of tipifarnib, a farnesyl transferase inhibitor, to inhibit CXCL12 production in PDAC.

Genes that are normally biased towards expression in the testis are often induced in tumor cells. These gametogenic genes, known as cancer-testis antigens (CTAs), have been extenstively investigated as targets for immunotherapy. However, despite their frequent detection, the degree to which CTAs support neoplastic invasion is poorly understood. Here, we find that the CTA genes SPANX-A/C/D and CTAG2 are coordinately induced in breast cancer cells and regulate distinct features of invasive behavior. Our functional analysis revealed that CTAG2 interacts with Pericentrin at the centrosome and is necessary for directional migration. Conversely, SPANX-A/C/D interacts with Lamin A/C at the inner nuclear membrane and is required for the formation of actin-rich cellular protrusions that reorganize the extracellular matrix. Importantly, SPANX-A/C/D was required for breast cancer cells to spontaneously metastasize to the lung, demonstrating that CTA reactivation can be critical for invasion dependent phenotypes in vivo. Moreover, elevated SPANX-A/C/D expression in breast cancer patient tumors correlated with poor outcome. Together, our results suggest that distinct CTAs promote tumor progression by regulating complementary cellular functions that are integrated together to induce invasive behavior.

Tumor cells can engage in a process called collective invasion, in which cohesive groups of cells invade through interstitial tissue. Here, we identified an epigenetically distinct subpopulation of breast tumor cells that have an enhanced capacity to collectively invade. Analysis of spheroid invasion in an organotypic culture system revealed that these “trailblazer” cells are capable of initiating collective invasion and promote non-trailblazer cell invasion, indicating a commensal relationship among subpopulations within heterogenous tumors. Canonical mesenchymal markers were not sufficient to distinguish trailblazer cells from non-trailblazer cells, suggesting that defining the molecular underpinnings of the trailblazer phenotype could reveal collective invasion-specific mechanisms. Functional analysis determined that DOCK10, ITGA11, DAB2, PDFGRA, VASN, PPAP2B, and LPAR1 are highly expressed in trailblazer cells and required to initiate collective invasion, with DOCK10 essential for metastasis. In patients with triple-negative breast cancer, expression of these 7 genes correlated with poor outcome. Together, our results indicate that spontaneous conversion of the epigenetic state in a subpopulation of cells can promote a transition from in situ to invasive growth through induction of a cooperative form of collective invasion and suggest that therapeutic inhibition of trailblazer cell invasion may help prevent metastasis.

The objective of this study was to determine whether the use of cyanoacrylate skin glue following subcuticular skin closure was associated with a decrease in wound outcomes in comparison with subcuticular closure plus Steri-strips at cesarean delivery.This was a retrospective cohort study of patients undergoing cesarean delivery at a single center over a two-year period. The primary outcome of wound infection and secondary outcomes of wound separation and composite wound complication rate were assessed throughout the six-week postpartum period.Of 660 women who met inclusion criteria, 35 (5.3%) experienced a wound infection and 90 (13.6%) experienced a wound separation. The composite wound complication rate was 16.4% (n = 108). Of the 515 cases with a skin coverage method noted, use of skin glue was associated with a marginal decrease in wound infections (p = 0.057), as well as a significantly reduced incidence of wound separation (p = 0.03) and composite wound complications (p = 0.006).Cyanoacrylate skin glue may be superior to Steri-strips for wound separation and composite wound complication rates when utilized with subcuticular suture at the time of cesarean delivery and may yield some benefit for prevention of wound infection.

EWS/FLI1-activated transcript 2 (EAT-2)A and EAT-2B are single SH2-domain proteins, which bind to phosphorylated tyrosines of signaling lymphocyte activation molecule family receptors in murine NK cells. While EAT-2 is a positive regulator in human cells, a negative regulatory role was attributed to the adapter in NK cells derived from EAT-2A–deficient 129Sv mice. To evaluate whether the genetic background or the presence of a selection marker in the mutant mice could influence the regulatory mode of these adapters, we generated EAT-2A–, EAT-2B–, and EAT-2A/B–deficient mice using C57BL/6 embryonic stem cells. We found that NK cells from EAT-2A– and EAT-2A/B–deficient mice were unable to kill tumor cells in a CD244- or CD84-dependent manner. Furthermore, EAT-2A/B positively regulate phosphorylation of Vav-1, which is known to be implicated in NK cell killing. Thus, as in humans, the EAT-2 adapters act as positive regulators of signaling lymphocyte activation molecule family receptor-specific NK cell functions in C57BL/6 mice.

Cell identity signals influence the invasive capability of tumor cells, as demonstrated by the selection for programs of epithelial-to-mesenchymal transition (EMT) during malignant progression. Breast cancer cells retain canonical epithelial traits and invade collectively as cohesive groups of cells, but the signaling pathways critical to their invasive capabilities are still incompletely understood. Here we report that the transcription factor ΔNp63α drives the migration of basal-like breast cancer (BLBC) cells by inducing a hybrid mesenchymal/epithelial state. Through a combination of expression analysis and functional testing across multiple BLBC cell populations, we determined that ΔNp63α induces migration by elevating the expression of the EMT program components Slug and Axl. Interestingly, ΔNp63α also increased the expression of miR-205, which can silence ZEB1/2 to prevent the loss of epithelial character caused by EMT induction. In clinical specimens, co-expression of various elements of the ΔNp63α pathway confirmed its implication in motility signaling in BLBC. We observed that activation of the ΔNp63α pathway occurred during the transition from noninvasive ductal carcinoma in situ to invasive breast cancer. Notably, in an orthotopic tumor model, Slug expression was sufficient to induce collective invasion of E-cadherin–expressing BLBC cells. Together, our results illustrate how ΔNp63α can drive breast cancer cell invasion by selectively engaging promigratory components of the EMT program while, in parallel, still promoting the retention of epithelial character. Cancer Res; 75(18); 3925–35. ©2015 AACR.

The signaling lymphocyte activation molecule (SLAM) family of receptors and their associated signaling adaptors play a pivotal role in the regulation of various stages of cellular immunity. They regulate lymphocyte-lymphocyte interactions involved in both cell-mediated and humoral immune responses. Recent evidence indicates that members of this family of receptors and signaling intermediates are also involved in autoimmunity. These include strictly correlative studies showing increased expression of various family members in immune effectors involved in rheumatoid arthritis and in inflammatory bowel disease, as well as more direct evidence (from various knockout strains of mice) for their role in autoimmune processes such as experimental allergic encephalomyelitis and lupus. Additional studies defining naturally occurring polymorphic variations in the SLAM family show a direct role in initiating the break in tolerance that is an essential step in the progression towards autoimmunity.

Diverse genetic and epigenetic traits can promote variant behaviors in subpopulations of cells within primary tumors. In some cases, these heterogeneous subpopulations may compete against each other for clonal dominance to improve tumor fitness. By comparison, the extent to which tumor cell subpopulations cooperate during malignant progression is not well understood. Here, we find that a commensal relationship can exist between epigentically distinct subpopulations of tumor cells during collective invasion. Using quantitive phenotypic analysis and live-cell imaging, we discovered that a subpopulation of cells, termed trailblazer cells, has an enhanced ability to collectively invade. Trailblazer phenotype was heritable and trailblazer subpopulations were derived from 5 triple-negative breast cancer cell lines. The percentage of trailblazers cells within a population ranged from between 0.5% and 90% of the parental population. Gene expression profiling identified a 29-gene expression signature that distinguished trailblazer cells from non-trailblazer cells and identified patients with poor outcome, A cohort of genes trailblazer signature genes were necessary for the formation of actin-rich projections that facilitate path generation during collective invasion. In addition, the trailblazer gene DOCK10 was necessary for spontaneous metastasis to the lung. The path generating function of a small subpopulation of trailblazer cells was sufficient to induce the collective invasion of large population of naturally occurring non-trailblazer cells in organotypic culture. Trailblazer cells were also sufficient to induce the invasion of a mouse model of ductal carcinoma in situ. Taken together, our results suggest that a rare subpopulation of tumor cells with an enhanced ability to collectively invade can promote metastatic progression through cell autonomous and non-cell autonomous mechanisms. Thus, our results support a paradigm shift towards identifying both the cell autonomous and the non cell-autonomous regulatory mechanisms of cell biological processes within heterogeneous tumors to understand neoplastic progression and develop new intervention modalities. Citation Format: Amanda Prechtl, Jill Westcott, Tuyen Dang, Erin Maine, Gray Pearson. Non-cell autonomous regulation of collective invasion be an epigenetically distinct subpopulation of tumor cells. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3161. doi:10.1158/1538-7445.AM2014-3161

The invasion of tumor cells out of ductal epithelium is a poorly understood process that dramatically increases the odds of patient death. We have defined a new non cell-autonomous control mechanism for neoplastic invasion that trigger tumors to switch from a benign to a malignant stage of growth. Using high resolution live-imaging and functional studies we show that an epigenetically distinct and autonomously invasive subpopulation of tumor cells induces the invasion of a large bulk population of autonomously noninvasive tumor cells by adopting migratory principles observed during sprouting angiogenesis. The molecular dissection of this cooperative interaction revealed a new “invasion signature” of 29 genes that is distinct from traditional epithelial-to-mesenchymal transition markers. Patients expressing this invasion signature had a worse clinical outcome, which suggests that this commensal relationship between tumor cell subpopulations contributes to tumor progression. Functional analysis showed that multiple components of the invasion signature were specifically required for the cooperative induction of invasion, but not cell motility, thus revealing new roles for these genes as non-cell autonomous regulators of invasion. These results are of broad interest because they are the first demonstration of a commensal relationship between tumor cell subpopulations during the induction of invasion. Further, these findings establish a mechanistic underpinning that obviates the notion that the exclusive role of heterogeneity in tumors is to act as a source for unique traits that can be selected for and expanded to improve the cell autonomous fitness of neoplastic cell populations. Thus, our results support a paradigm shift towards identifying both the cell autonomous and the non cell-autonomous regulatory mechanisms of cell biological processes within heterogeneous tumors to understand neoplastic progression and develop new intervention modalities. Citation Format: Gray Pearson, Amanda Prechtl, Jill Westcott, Tuyen Dang, Erin Maine. A commensal relationship between tumor cell subpopulations promotes collective invasion. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 4930. doi:10.1158/1538-7445.AM2013-4930 Note: This abstract was not presented at the AACR Annual Meeting 2013 because the presenter was unable to attend.