A deep learning model of tumor cell architecture elucidates response and resistance to CDK4/6 inhibitors.

Cyclin-dependent kinase 4 and 6 inhibitors (CDK4/6is) have revolutionized breast cancer therapy. However, <50% of patients have an objective response, and nearly all patients develop resistance during therapy. To elucidate the underlying mechanisms, we constructed an interpretable deep learning model of the response to palbociclib, a CDK4/6i, based on a reference map of multiprotein assemblies in cancer. The model identifies eight core assemblies that integrate rare and common alterations across 90 genes to stratify palbociclib-sensitive versus palbociclib-resistant cell lines. Predictions translate to patients and patient-derived xenografts, whereas single-gene biomarkers do not. Most predictive assemblies can be shown by CRISPR-Cas9 genetic disruption to regulate the CDK4/6i response. Validated assemblies relate to cell-cycle control, growth factor signaling and a histone regulatory complex that we show promotes S-phase entry through the activation of the histone modifiers KAT6A and TBL1XR1 and the transcription factor RUNX1. This study enables an integrated assessment of how a tumor's genetic profile modulates CDK4/6i resistance.

Cancer Mutations Converge on a Collection of Protein Assemblies to Predict Resistance to Replication Stress.

Rapid proliferation is a hallmark of cancer associated with sensitivity to therapeutics that cause DNA replication stress (RS). Many tumors exhibit drug resistance, however, via molecular pathways that are incompletely understood. Here, we develop an ensemble of predictive models that elucidate how cancer mutations impact the response to common RS-inducing (RSi) agents. The models implement recent advances in deep learning to facilitate multidrug prediction and mechanistic interpretation. Initial studies in tumor cells identify 41 molecular assemblies that integrate alterations in hundreds of genes for accurate drug response prediction. These cover roles in transcription, repair, cell-cycle checkpoints, and growth signaling, of which 30 are shown by loss-of-function genetic screens to regulate drug sensitivity or replication restart. The model translates to cisplatin-treated cervical cancer patients, highlighting an RTK-JAK-STAT assembly governing resistance. This study defines a compendium of mechanisms by which mutations affect therapeutic responses, with implications for precision medicine. SIGNIFICANCE: Zhao and colleagues use recent advances in machine learning to study the effects of tumor mutations on the response to common therapeutics that cause RS. The resulting predictive models integrate numerous genetic alterations distributed across a constellation of molecular assemblies, facilitating a quantitative and interpretable assessment of drug response. This article is featured in Selected Articles from This Issue, p. 384.

Evaluation of large language models for discovery of gene set function.

Gene set analysis is a mainstay of functional genomics, but it relies on curated databases of gene functions that are incomplete. Here we evaluate five Large Language Models (LLMs) for their ability to discover the common biological functions represented by a gene set, substantiated by supporting rationale, citations and a confidence assessment. Benchmarking against canonical gene sets from the Gene Ontology, GPT-4 confidently recovered the curated name or a more general concept (73% of cases), while benchmarking against random gene sets correctly yielded zero confidence. Gemini-Pro and Mixtral-Instruct showed ability in naming but were falsely confident for random sets, whereas Llama2-70b had poor performance overall. In gene sets derived from 'omics data, GPT-4 identified novel functions not reported by classical functional enrichment (32% of cases), which independent review indicated were largely verifiable and not hallucinations. The ability to rapidly synthesize common gene functions positions LLMs as valuable 'omics assistants.

Evaluation of large language models for discovery of gene set function.

Gene set analysis is a mainstay of functional genomics, but it relies on manually curated databases of gene functions that are incomplete and unaware of biological context. Here we evaluate the ability of OpenAI's GPT-4, a Large Language Model (LLM), to develop hypotheses about common gene functions from its embedded biomedical knowledge. We created a GPT-4 pipeline to label gene sets with names that summarize their consensus functions, substantiated by analysis text and citations. Benchmarking against named gene sets in the Gene Ontology, GPT-4 generated very similar names in 50% of cases, while in most remaining cases it recovered the name of a more general concept. In gene sets discovered in 'omics data, GPT-4 names were more informative than gene set enrichment, with supporting statements and citations that largely verified in human review. The ability to rapidly synthesize common gene functions positions LLMs as valuable functional genomics assistants.

The Association Between Metabolic Derangement and Wound Complications in Elective Plastic Surgery.

INTRODUCTION: The incidence of metabolically unhealthy obesity is rising nationally. In this study, we compare wound and overall complications between metabolically unhealthy obese and healthy patients undergoing elective plastic surgery and model how operative time influences a complication risk. METHODS: Patients undergoing elective breast and body plastic surgery procedures in the 2009-2019 National Surgical Quality Improvement Program (NSQIP) dataset were identified. Complications were compared between metabolically unhealthy obese (body mass index [BMI] > 30 with diabetes and/or hypertension) versus metabolically healthy obese patients (BMI > 30 without diabetes or hypertension). Logistic regression was used to model the probability of wound complications across operative times stratified by metabolic status. RESULTS: Of 139,352 patients, 13.4% (n = 18,663) had metabolically unhealthy obesity and 23.8% (n = 33,135) had metabolically healthy obesity. Compared to metabolically healthy patients, metabolically unhealthy patients had higher incidence of wound complications (6.9% versus 5.6%; P < 0.001) and adverse events (12.4% versus 9.6%; P < 0.001), in addition to higher 30-d readmission, returns to the operating room, and length of stay (all P < 0.001). After adjustment, BMI (Odds ratio [OR] 7.86), hypertension (OR 1.15), and diabetes (OR 1.25) were independent risk factors for wound complications (all P < 0.001). Among metabolically unhealthy patients, the operative time was log-linear with a wound complication risk (OR 1.21; P < 0.001). CONCLUSIONS: Diabetes and hypertension are additive risk factors with obesity for wound complications in elective plastic surgery. Among patients with metabolically unhealthy obesity, a risk of wound complications increases logarithmically with operative time. This distinction with regard to metabolic state might explain the unclear impact of obesity on surgical outcomes within existing surgical literature.

Suppression of Fibrinolysis and Hypercoagulability, Severity of Hypoxemia, and Mortality in COVID-19 Patients: A Retrospective Cohort Study.

BACKGROUND: COVID-19 causes hypercoagulability, but the association between coagulopathy and hypoxemia in critically ill patients has not been thoroughly explored. This study hypothesized that severity of coagulopathy would be associated with acute respiratory distress syndrome severity, major thrombotic events, and mortality in patients requiring intensive care unit-level care. METHODS: Viscoelastic testing by rotational thromboelastometry and coagulation factor biomarker analyses were performed in this prospective observational cohort study of critically ill COVID-19 patients from April 2020 to October 2020. Statistical analyses were performed to identify significant coagulopathic biomarkers such as fibrinolysis-inhibiting plasminogen activator inhibitor 1 and their associations with clinical outcomes such as mortality, extracorporeal membrane oxygenation requirement, occurrence of major thrombotic events, and severity of hypoxemia (arterial partial pressure of oxygen/fraction of inspired oxygen categorized into mild, moderate, and severe per the Berlin criteria). RESULTS: In total, 53 of 55 (96%) of the cohort required mechanical ventilation and 9 of 55 (16%) required extracorporeal membrane oxygenation. Extracorporeal membrane oxygenation-naïve patients demonstrated lysis indices at 30 min indicative of fibrinolytic suppression on rotational thromboelastometry. Survivors demonstrated fewer procoagulate acute phase reactants, such as microparticle-bound tissue factor levels (odds ratio, 0.14 [0.02, 0.99]; P = 0.049). Those who did not experience significant bleeding events had smaller changes in ADAMTS13 levels compared to those who did (odds ratio, 0.05 [0, 0.7]; P = 0.026). Elevations in plasminogen activator inhibitor 1 (odds ratio, 1.95 [1.21, 3.14]; P = 0.006), d-dimer (odds ratio, 3.52 [0.99, 12.48]; P = 0.05), and factor VIII (no clot, 1.15 ± 0.28 vs. clot, 1.42 ± 0.31; P = 0.003) were also demonstrated in extracorporeal membrane oxygenation-naïve patients who experienced major thrombotic events. Plasminogen activator inhibitor 1 levels were significantly elevated during periods of severe compared to mild and moderate acute respiratory distress syndrome (severe, 44.2 ± 14.9 ng/ml vs. mild, 31.8 ± 14.7 ng/ml and moderate, 33.1 ± 15.9 ng/ml; P = 0.029 and 0.039, respectively). CONCLUSIONS: Increased inflammatory and procoagulant markers such as plasminogen activator inhibitor 1, microparticle-bound tissue factor, and von Willebrand factor levels are associated with severe hypoxemia and major thrombotic events, implicating fibrinolytic suppression in the microcirculatory system and subsequent micro- and macrovascular thrombosis in severe COVID-19.

CRISPR/Cas9-Mediated BRCA1 Knockdown Adipose Stem Cells Promote Breast Cancer Progression.

BACKGROUND: The tumor microenvironment within the breast is rich in adipose elements. The interaction between adipose cells and breast cancer is poorly understood, particularly as it pertains to patients with genetic susceptibility to breast cancer. This study focuses on the phenotype of human adipose-derived stem cells with the BRCA1 mutation and the effect they may have on breast cancer cell behavior. METHODS: CRISPR/Cas9 was used to generate de novo BRCA1-knockdown human adipose-derived stem cells. The effect of the BRCA1 knockdown on the adipose-derived stem cell phenotype was compared to wild-type adipose-derived stem cells and patient-derived breast adipose-derived stem cells with known BRCA1 mutations. Interactions between adipose-derived stem cells and the MDA-MB-231 breast cancer cell line were evaluated. RESULTS: BRCA1-knockdown adipose-derived stem cells stimulated MDA-MB-231 proliferation (1.4-fold increase on day 4; p = 0.0074) and invasion (2.3-fold increase on day 2; p = 0.0171) compared to wild-type cells. Immunofluorescence staining revealed higher levels of phosphorylated ataxia telangiectasia-mutated activation in BRCA1-knockdown cells (72.9 ± 5.32 percent versus 42.9 ± 4.97 percent; p = 0.0147), indicating higher levels of DNA damage. Beta-galactosidase staining demonstrated a significantly higher level of senescence in BRCA1-knockdown cells compared with wild-type cells (7.9 ± 0.25 percent versus 0.17 ± 0.17 percent; p < 0.0001). Using quantitative enzyme-linked immunosorbent assay to evaluate conditioned media, the authors found significantly higher levels of interleukin-8 in BRCA1-knockdown cells (2.57 ± 0.32-fold; p = 0.0049). CONCLUSIONS: The authors show for the first time that the BRCA1 mutation affects the adipose-derived stem cell phenotype. Moreover, CRISPR/Cas9-generated BRCA1-knockdown adipose-derived stem cells stimulate a more aggressive behavior in breast cancer cells than wild-type adipose-derived stem cells. This appears to be related to increased inflammatory cytokine production by means of a DNA damage-mediated cell senescence pathway.

Adipose stem cell crosstalk with chemo-residual breast cancer cells: implications for tumor recurrence.

PURPOSE: Most triple-negative breast cancer (TNBC) patients exhibit an incomplete response to neoadjuvant chemotherapy, resulting in chemo-residual tumor cells that drive tumor recurrence and patient mortality. Accordingly, strategies for eliminating chemo-residual tumor cells are urgently needed. Although stromal cells contribute to tumor cell invasion, to date, their ability to influence chemo-residual tumor cell behavior has not been examined. Our study is the first to investigate cross-talk between adipose-derived stem cells (ASCs) and chemo-residual TNBC cells. We examine if ASCs promote chemo-residual tumor cell proliferation, having implications for tumor recurrence. METHODS: ASC migration toward chemo-residual TNBC cells was tested in a transwell migration assay. Importance of the SDF-1α/CXCR4 axis was determined using neutralizing antibodies and a small molecule inhibitor. The ability of ASCs to drive tumor cell proliferation was analyzed by culturing tumor cells ± ASC conditioned media (CM) and determining cell counts. Downstream signaling pathways activated in chemo-residual tumor cells following their exposure to ASC CM were studied by immunoblotting. Importance of FGF2 in promoting proliferation was assessed using an FGF2-neutralizing antibody. RESULTS: ASCs migrated toward chemo-residual TNBC cells in a CXCR4/SDF-1α-dependent manner. Moreover, ASC CM increased chemo-residual tumor cell proliferation and activity of extracellular signal-regulated kinase (ERK). An FGF2-neutralizing antibody inhibited ASC-induced chemo-residual tumor cell proliferation. CONCLUSIONS: ASCs migrate toward chemo-residual TNBC cells via SDF-1α/CXCR4 signaling, and drive chemo-residual tumor cell proliferation in a paracrine manner by secreting FGF2 and activating ERK. This paracrine signaling can potentially be targeted to prevent tumor recurrence.

Chemotherapy enriches for an invasive triple-negative breast tumor cell subpopulation expressing a precursor form of N-cadherin on the cell surface.

BACKGROUND: Although most triple-negative breast cancer (TNBC) patients initially respond to chemotherapy, residual tumor cells frequently persist and drive recurrent tumor growth. Previous studies from our laboratory and others' indicate that TNBC is heterogeneous, being composed of chemo-sensitive and chemo-resistant tumor cell subpopulations. In the current work, we studied the invasive behaviors of chemo-resistant TNBC, and sought to identify markers of invasion in chemo-residual TNBC. METHODS: The invasive behavior of TNBC tumor cells surviving short-term chemotherapy treatment in vitro was studied using transwell invasion assays and an experimental metastasis model. mRNA expression levels of neural cadherin (N-cadherin), an adhesion molecule that promotes invasion, was assessed by PCR. Expression of N-cadherin and its precursor form (pro-N-cadherin) was assessed by immunoblotting and flow cytometry. Pro-N-cadherin immunohistochemistry was performed on tumors obtained from patients pre- and post- neoadjuvant chemotherapy treatment. RESULTS: TNBC cells surviving short-term chemotherapy treatment exhibited increased invasive behavior and capacity to colonize metastatic sites compared to untreated tumor cells. The invasive behavior of chemo-resistant cells was associated with their increased cell surface expression of precursor N-cadherin (pro-N-cadherin). An antibody specific for the precursor domain of N-cadherin inhibited invasion of chemo-resistant TNBC cells. To begin to validate our findings in humans, we showed that the percent cell surface pro-N-cadherin (+) tumor cells increased in patients post- chemotherapy treatment. CONCLUSIONS: TNBC cells surviving short-term chemotherapy treatment are more invasive than bulk tumor cells. Cell surface pro-N-cadherin expression is associated with the invasive and chemo-resistant behaviors of this tumor cell subset. Our findings indicate the importance of future studies determining the value of cell surface pro-N-cadherin as: 1) a biomarker for TNBC recurrence and 2) a therapeutic target for eliminating chemo-residual disease.

Nuclear basic fibroblast growth factor regulates triple-negative breast cancer chemo-resistance.

INTRODUCTION: Chemotherapy remains the only available treatment for triple-negative (TN) breast cancer, and most patients exhibit an incomplete pathologic response. Half of patients exhibiting an incomplete pathologic response die within five years of treatment due to chemo-resistant, recurrent tumor growth. Defining molecules responsible for TN breast cancer chemo-resistance is crucial for developing effective combination therapies blocking tumor recurrence. Historically, chemo-resistance studies have relied on long-term chemotherapy selection models that drive genetic mutations conferring cell survival. Other models suggest that tumors are heterogeneous, being composed of both chemo-sensitive and chemo-resistant tumor cell populations. We previously described a short-term chemotherapy treatment model that enriches for chemo-residual TN tumor cells. In the current work, we use this enrichment strategy to identify a novel determinant of TN breast cancer chemotherapy resistance [a nuclear isoform of basic fibroblast growth factor (bFGF)]. METHODS: Studies are conducted using our in vitro model of chemotherapy resistance. Short-term chemotherapy treatment enriches for a chemo-residual TN subpopulation that over time resumes proliferation. By western blotting and real-time polymerase chain reaction, we show that this chemotherapy-enriched tumor cell subpopulation expresses nuclear bFGF. The importance of bFGF for survival of these chemo-residual cells is interrogated using short hairpin knockdown strategies. DNA repair capability is assessed by comet assay. Immunohistochemistry (IHC) is used to determine nuclear bFGF expression in TN breast cancer cases pre- and post- neoadjuvant chemotherapy. RESULTS: TN tumor cells surviving short-term chemotherapy treatment express increased nuclear bFGF. bFGF knockdown reduces the number of chemo-residual TN tumor cells. Adding back a nuclear bFGF construct to bFGF knockdown cells restores their chemo-resistance. Nuclear bFGF-mediated chemo-resistance is associated with increased DNA-dependent protein kinase (DNA-PK) expression and accelerated DNA repair. In fifty-six percent of matched TN breast cancer cases, percent nuclear bFGF-positive tumor cells either increases or remains the same post- neoadjuvant chemotherapy treatment (compared to pre-treatment). These data indicate that in a subset of TN breast cancers, chemotherapy enriches for nuclear bFGF-expressing tumor cells. CONCLUSION: These studies identify nuclear bFGF as a protein in a subset of TN breast cancers that likely contributes to drug resistance following standard chemotherapy treatment.

Ascites Increases Expression/Function of Multidrug Resistance Proteins in Ovarian Cancer Cells.

Chemotherapy resistance is the major reason for the failure of ovarian cancer treatment. One mechanism behind chemo-resistance involves the upregulation of multidrug resistance (MDR) genes (ABC transporters) that effectively transport (efflux) drugs out of the tumor cells. As a common symptom in stage III/IV ovarian cancer patients, ascites is associated with cancer progression. However, whether ascites drives multidrug resistance in ovarian cancer cells awaits elucidation. Here, we demonstrate that when cultured with ascites derived from ovarian cancer-bearing mice, a murine ovarian cancer cell line became less sensitive to paclitaxel, a first line chemotherapeutic agent for ovarian cancer patients. Moreover, incubation of murine ovarian cancer cells in vitro with ascites drives efflux function in these cells. Functional studies show ascites-driven efflux is suppressible by specific inhibitors of either of two ABC transporters [Multidrug Related Protein (MRP1); Breast Cancer Related Protein (BCRP)]. To demonstrate relevance of our findings to ovarian cancer patients, we studied relative efflux in human ovarian cancer cells obtained from either patient ascites or from primary tumor. Immortalized cell lines developed from human ascites show increased susceptibility to efflux inhibitors (MRP1, BCRP) compared to a cell line derived from a primary ovarian cancer, suggesting an association between ascites and efflux function in human ovarian cancer. Efflux in ascites-derived human ovarian cancer cells is associated with increased expression of ABC transporters compared to that in primary tumor-derived human ovarian cancer cells. Collectively, our findings identify a novel activity for ascites in promoting ovarian cancer multidrug resistance.

Syngeneic Murine Ovarian Cancer Model Reveals That Ascites Enriches for Ovarian Cancer Stem-Like Cells Expressing Membrane GRP78.

Patients with ovarian cancer are generally diagnosed at FIGO (International Federation of Gynecology and Obstetrics) stage III/IV, when ascites is common. The volume of ascites correlates positively with the extent of metastasis and negatively with prognosis. Membrane GRP78, a stress-inducible endoplasmic reticulum chaperone that is also expressed on the plasma membrane ((mem)GRP78) of aggressive cancer cells, plays a crucial role in the embryonic stem cell maintenance. We studied the effects of ascites on ovarian cancer stem-like cells using a syngeneic mouse model. Our study demonstrates that ascites-derived tumor cells from mice injected intraperitoneally with murine ovarian cancer cells (ID8) express increased (mem)GRP78 levels compared with ID8 cells from normal culture. We hypothesized that these ascites-associated (mem)GRP78(+) cells are cancer stem-like cells (CSC). Supporting this hypothesis, we show that (mem)GRP78(+) cells isolated from murine ascites exhibit increased sphere forming and tumor initiating abilities compared with (mem)GRP78(-) cells. When the tumor microenvironment is recapitulated by adding ascites fluid to cell culture, ID8 cells express more (mem)GRP78 and increased self-renewing ability compared with those cultured in medium alone. Moreover, compared with their counterparts cultured in normal medium, ID8 cells cultured in ascites, or isolated from ascites, show increased stem cell marker expression. Antibodies directed against the carboxy-terminal domain of GRP78: (i) reduce self-renewing ability of murine and human ovarian cancer cells preincubated with ascites and (ii) suppress a GSK3α-AKT/SNAI1 signaling axis in these cells. Based on these data, we suggest that (mem)GRP78 is a logical therapeutic target for late-stage ovarian cancer.

Model of tumor dormancy/recurrence after short-term chemotherapy.

Although many tumors regress in response to neoadjuvant chemotherapy, residual tumor cells are detected in most cancer patients post-treatment. These residual tumor cells are thought to remain dormant for years before resuming growth, resulting in tumor recurrence. Considering that recurrent tumors are most often responsible for patient mortality, there exists an urgent need to study signaling pathways that drive tumor dormancy/recurrence. We have developed an in vitro model of tumor dormancy/recurrence. Short-term exposure of tumor cells (breast or prostate) to chemotherapy at clinically relevant doses enriches for a dormant tumor cell population. Several days after removing chemotherapy, dormant tumor cells regain proliferative ability and establish colonies, resembling tumor recurrence. Tumor cells from "recurrent" colonies exhibit increased chemotherapy resistance, similar to the therapy resistance of recurrent tumors in cancer patients. Previous studies using long-term chemotherapy selection models identified acquired mutations that drive tumor resistance. In contrast, our short term chemotherapy exposure model enriches for a slow-cycling, dormant, chemo-resistant tumor cell sub-population that can resume growth after drug removal. Studying unique signaling pathways in dormant tumor cells enriched by short-term chemotherapy treatment is expected to identify novel therapeutic targets for preventing tumor recurrence.

Evidence for phenotypic plasticity in aggressive triple-negative breast cancer: human biology is recapitulated by a novel model system.

Breast cancers with a basal-like gene signature are primarily triple-negative, frequently metastatic, and carry a poor prognosis. Basal-like breast cancers are enriched for markers of breast cancer stem cells as well as markers of epithelial-mesenchymal transition (EMT). While EMT is generally thought to be important in the process of metastasis, in vivo evidence of EMT in human disease remains rare. Here we report a novel model of human triple-negative breast cancer, the DKAT cell line, which was isolated from an aggressive, treatment-resistant triple-negative breast cancer that demonstrated morphological and biochemical evidence suggestive of phenotypic plasticity in the patient. The DKAT cell line displays a basal-like phenotype in vitro when cultured in serum-free media, and undergoes phenotypic changes consistent with EMT/MET in response to serum-containing media, a unique property among the breast cancer cell lines we tested. This EMT is marked by increased expression of the transcription factor Zeb1, and Zeb1 is required for the enhanced migratory ability of DKAT cells in the mesenchymal state. DKAT cells also express progenitor-cell markers, and single DKAT cells are able to generate tumorspheres containing both epithelial and mesenchymal cell types. In vivo, as few as ten DKAT cells are capable of forming xenograft tumors which display a range of epithelial and mesenchymal phenotypes. The DKAT model provides a novel model to study the molecular mechanisms regulating phenotypic plasticity and the aggressive biology of triple-negative breast cancers.

Nuclear Snail1 and nuclear ZEB1 protein expression in invasive and intraductal human breast carcinomas.

Snail1 and ZEB1 are transcriptional repressors that drive tumor initiation and metastasis in animal models. Snail1 and ZEB1 are frequently coexpressed in tumor cell lines, suggesting that these factors may cooperate to promote tumor progression. However, coexpression of these transcriptional repressors in primary human cancer specimens has not been investigated. Previous studies assessed expression in primary breast cancers of Snail1 messenger RNA, which does not reflect Snail1 activity because Snail1 is subject to posttranslational modifications that inhibit its nuclear localization/activity. In the current study, using breast tumor cell lines of known Snail1 and ZEB1 expression status, we developed immunohistochemistry protocols for detecting nuclear Snail1 and nuclear ZEB1 proteins. Using these protocols, we assessed nuclear Snail1 and nuclear ZEB1 expressions in primary human breast cancers of varying subtypes (n = 78). Nuclear Snail1 and estrogen receptor α expressions were inversely associated in primary breast cancers, and nuclear Snail1 was expressed in approximately 80% of triple-negative breast cancers (lacking estrogen receptor α, progesterone receptor, and human epidermal growth factor receptor 2 overexpression). In contrast, nuclear ZEB1 was expressed at a significantly lower frequency in these breast cancers. Notably, nuclear Snail1 protein was detected in 45% of ductal carcinoma in situ specimens (n = 29), raising the important possibility that nuclear Snail1 expression in early stage breast lesions may predict future development of invasive breast cancer. Collectively, our studies demonstrate frequent expression of nuclear Snail1, but not nuclear ZEB1, in invasive, triple-negative breast cancers as well as in intraductal carcinomas.

Autocrine Semaphorin3A stimulates eukaryotic initiation factor 4E-dependent RhoA translation in breast tumor cells.

Translation of the small G protein RhoA in neurons is regulated by the eukaryotic translation initiation factor eIF4E. Here we show that this translation factor also regulates RhoA expression and activity in breast cancer cells. The introduction of eIF4E into breast tumor cells increased RhoA protein levels, while expression of an eIF4E siRNA reduced RhoA expression. Previous studies indicate that the axon repulsion factor Semaphorin3A (Sema3A) stimulates the eIF4E-dependent translation of RhoA in neurons, and breast tumor cells support autocrine Sema3A signaling. Accordingly, we next examined if autocrine Sema3A signaling drives eIF4E-dependent RhoA translation in breast cancer cells. The incubation of breast tumor cells with recombinant Sema3A rapidly increased eIF4E activity, RhoA protein levels, and RhoA activity. This Sema3A activity was blocked in tumor cells expressing an shRNA-specific for the Sema3A receptor, Neuropilin-1 (NP-1), as well as in cells incubated with an eIF4E inhibitor. Importantly, RhoA protein levels were reduced in Sema3A shRNA-expressing compared to control shRNA-expressing breast tumor cells, demonstrating that autocrine Sema3A increases RhoA expression in breast cancer. Considering that Sema3A suppresses axon extension by stimulating RhoA translation, we next examined if the Sema3A/RhoA axis impacts breast tumor cell migration. The incubation of control breast tumor cells, but not RhoA shRNA-expressing cells, with rSema3A significantly reduced their migration. Collectively, these studies indicate that Sema3A impedes breast tumor cell migration in part by stimulating RhoA. These findings identify common signaling pathways that regulate the navigation of neurons and breast cancer cells, thus suggesting novel targets for suppressing breast tumor cell migration.

Autocrine semaphorin3A stimulates alpha2 beta1 integrin expression/function in breast tumor cells.

The axon repulsion factor semaphorin3A (SEMA3A) and its receptor neuropilin-1 (NP-1) are expressed in breast tumor cells, and function as suppressors of tumor cell migration. Based on the knowledge that both SEMA3A and the alpha2beta1 integrin suppress breast tumor cell migration, we studied the impact of SEMA3A signaling on alpha2beta1 integrin expression/function. The incubation of breast tumor cells with SEMA3A increased alpha2 and beta1 integrin levels, and stimulated tumor cell adhesion to the alpha2beta1-binding matrix protein collagen I. Conversely, reducing SEMA3A expression in breast tumor cells decreased alpha2beta1 levels and collagen adhesion. The ability of SEMA3A to increase tumor cell adhesion to collagen was dependent on both the SEMA3A receptor NP-1 and the glycogen synthase kinase-3. The incubation of breast tumor cells with SEMA3A disrupted the actin cytoskeleton, and reduced both tumor cell migratory and invasive behavior. Importantly, using an alpha2beta1-neutralizing antibody, we demonstrated that SEMA3A suppression of tumor cell migration is dependent on alpha2beta1. Our studies indicate that expression of the alpha2beta1 integrin, a suppressor of metastatic breast tumor growth, is stimulated in breast tumor cells by an autocrine SEMA3A pathway.

Vascular endothelial growth factor-A stimulates Snail expression in breast tumor cells: implications for tumor progression.

The E-cadherin transcriptional repressor Snail is a prognostic marker for metastatic breast carcinoma, as well as a critical determinant of tumor growth and recurrence. We define a non-angiogenic, autocrine function for the vascular endothelial growth factor-A (VEGF-A) in regulating Snail expression in breast tumor cells. The transfection of well-differentiated breast tumor cells with VEGF-A increases Snail mRNA and protein levels, resulting in reduced E-cadherin expression. Conversely, reducing endogenous VEGF-A expression in poorly differentiated breast tumor cells by siRNA transfection decreases Snail levels. Our studies demonstrate that VEGF and the VEGF receptor Neuropilin-1 increase Snail expression by suppressing the Glycogen Synthase Kinase-3 (GSK-3), an established inhibitor of Snail transcription and protein stability. The VEGF-A neutralizing antibody Avastin was recently approved by the FDA for the treatment of metastatic breast cancer. We present the provocative finding that beyond its anti-angiogenic activity, Avastin can reduce Snail expression in breast tumor cells. Collectively, this work describes a novel autocrine function for VEGF in breast tumor cells in driving the expression of Snail, a breast tumor progression factor. Based on our demonstration that Avastin reduces Snail expression in breast tumor cells, we propose that the treatment of early stage breast cancer patients with Avastin may impede tumor progression.

The alpha6beta4 integrin can regulate ErbB-3 expression: implications for alpha6beta4 signaling and function.

The integrin alpha(6)beta(4) has been shown to facilitate key functions of carcinoma cells, including their ability to migrate, invade, and evade apoptosis. The mechanism involved seems to be a profound effect of alpha(6)beta(4) on specific signaling pathways, especially the phosphatidylinositol 3-kinase (PI3K)/Akt pathway. An intimate relationship between alpha(6)beta(4) and growth factor receptors may explain this effect of alpha(6)beta(4) on signaling. Previously, we showed that alpha(6)beta(4) and ErbB-2 can function synergistically to activate the PI3K/Akt pathway. Given that ErbB-2 can activate PI3K only when it heterodimerizes with other members of the epidermal growth factor receptor family, these data imply that other receptors cooperate in this process. Here, we report that alpha(6)beta(4) can regulate the expression of ErbB-3 using several different models and that the consequent formation of an ErbB-2/ErbB-3 heterodimer promotes the alpha(6)beta(4)-dependent activation of PI3K/Akt and the ability of this integrin to impede apoptosis of carcinoma cells. Our data also support the hypothesis that alpha(6)beta(4) can regulate ErbB-3 expression at the translational level as evidenced by the findings that alpha(6)beta(4) does not increase ErbB-3 mRNA significantly, and that this regulation is both rapamycin sensitive and dependent on eukaryotic translation initiation factor 4E. These findings provide one mechanism to account for the activation of PI3K by alpha(6)beta(4) and they also provide insight into the regulation of ErbB-3 in carcinoma cells.