High-Content Analysis of Breast Cancer Using Single-Cell Deep Transfer Learning

High-content analysis has revolutionized cancer drug discovery by identifying substances that alter the phenotype of a cell, which prevents tumor growth and metastasis. The high-resolution biofluorescence images from assays allow precise quantitative measures enabling the distinction of small molecules of a host cell from a tumor. In this work, we are particularly interested in the application of deep neural networks (DNNs), a cutting-edge machine learning method, to the classification of compounds in chemical mechanisms of action (MOAs). Compound classification has been performed using image-based profiling methods sometimes combined with feature reduction methods such as principal component analysis or factor analysis. In this article, we map the input features of each cell to a particular MOA class without using any treatment-level profiles or feature reduction methods. To the best of our knowledge, this is the first application of DNN in this domain, leveraging single-cell information. Furthermore, we use deep transfer learning (DTL) to alleviate the intensive and computational demanding effort of searching the huge parameter's space of a DNN. Results show that using this approach, we obtain a 30% speedup and a 2% accuracy improvement.

Identification of estrogen-responsive DNA sequences by transient expression experiments in a human breast cancer cell line.

The expression of a hybrid gene formed by the promoter region of the Xenopus laevis vitellogenin gene B1 and the CAT coding region is regulated by estrogen when the gene is transfected into hormone-responsive MCF-7 cells. Furthermore, the 5' flanking region of the gene B1 alone can confer inducibility to heterologous promoters, although to a varying extent depending on the promoter used. Deletion mapping of he vitellogenin hormone-responsive sequences revealed that a 13 bp element 5'-AGTCACTGTGACC-3' at position -334 is essential for estrogen inducibility. We have shown previously that this 13 bp element is present upstream of several liver-specific estrogen-inducible genes.

Prognostic effect of epithelial cell adhesion molecule overexpression in untreated node-negative breast cancer.

PURPOSE: Epithelial cell adhesion molecule (Ep-CAM) recently received increased attention not only as a prognostic factor in breast cancer but also as a potential target for immunotherapy. We examined Ep-CAM expression in 402 consecutive node-negative breast cancer patients with long-term follow-up not treated in the adjuvant setting. EXPERIMENTAL DESIGN: Ep-CAM expression was evaluated by immunostaining. Its prognostic effect was estimated relative to overexpression/amplification of HER-2, histologic grade, tumor size, age, and hormone receptor expression. RESULTS: Ep-CAM status was positive in 106 (26.4%) patients. In multivariate analysis, Ep-CAM status was associated with disease-free survival independent of age, pT stage, histologic grade, estrogen receptor (ER), progesterone receptor (PR), as well as HER2 status (P = 0.028; hazard ratio, 1.60; 95% confidence interval, 1.05-2.44). Recently, so-called triple-negative (HER-2, ER, and PR) breast cancer has received increased attention. We noticed a similar association of Ep-CAM with disease-free survival in the triple-negative group as for the entire cohort. CONCLUSION: In this study of untreated breast cancer patients, Ep-CAM overexpression was associated with poor survival in the entire cohort and in the subgroup of triple-negative breast cancer. This suggests that Ep-CAM may be a well-suited target for specific therapies particularly in HER-2-, ER-, and PR-negative tumors.

Breast cancer and melanoma cell line identification by FTIR imaging after formalin-fixation and paraffin-embedding.

Over the past few decades, Fourier transform infrared (FTIR) spectroscopy coupled to microscopy has been recognized as an emerging and potentially powerful tool in cancer research and diagnosis. For this purpose, histological analyses performed by pathologists are mostly carried out on biopsied tissue that undergoes the formalin-fixation and paraffin-embedding (FFPE) procedure. This processing method ensures an optimal and permanent preservation of the samples, making FFPE-archived tissue an extremely valuable source for retrospective studies. Nevertheless, as highlighted by previous studies, this fixation procedure significantly changes the principal constituents of cells, resulting in important effects on their infrared (IR) spectrum. Despite the chemical and spectral influence of FFPE processing, some studies demonstrate that FTIR imaging allows precise identification of the different cell types present in biopsied tissue, indicating that the FFPE process preserves spectral differences between distinct cell types. In this study, we investigated whether this is also the case for closely related cell lines. We analyzed spectra from 8 cancerous epithelial cell lines: 4 breast cancer cell lines and 4 melanoma cell lines. For each cell line, we harvested cells at subconfluence and divided them into two sets. We first tested the "original" capability of FTIR imaging to identify these closely related cell lines on cells just dried on BaF2 slides. We then repeated the test after submitting the cells to the FFPE procedure. Our results show that the IR spectra of FFPE processed cancerous cell lines undergo small but significant changes due to the treatment. The spectral modifications were interpreted as a potential decrease in the phospholipid content and protein denaturation, in line with the scientific literature on the topic. Nevertheless, unsupervised analyses showed that spectral proximities and distances between closely related cell lines were mostly, but not entirely, conserved after FFPE processing. Finally, PLS-DA statistical analyses highlighted that closely related cell lines are still successfully identified and efficiently distinguished by FTIR spectroscopy after FFPE treatment. This last result paves the way towards identification and characterization of cellular subtypes on FFPE tissue sections by FTIR imaging, indicating that this analysis technique could become a potential useful tool in cancer research.

High dose chemotherapy with autologous hematopoietic stem cell support for solid tumors other than breast cancer in adults.

Since the early 1980s high dose chemotherapy with autologous hematopoietic stem cell support was adopted by many oncologists as a potentially curative option for solid tumors, supported by a strong rationale from laboratory studies and apparently convincing results of early phase II studies. As a result, the number and size of randomized trials comparing this approach with conventional chemotherapy initiated (and often abandoned before completion) to prove or disprove its value was largely insufficient. In fact, with the possible exception of breast carcinoma, the benefit of a greater escalation of dose of chemotherapy with stem cell support in solid tumors is still unsettled and many oncologists believe that this approach should cease. In this article, we critically review and comment on the data from studies of high dose chemotherapy so far reported in adult patients with small cell lung cancer, ovarian cancer, germ cell tumors and sarcomas.

Breast implant-associated anaplastic large cell lymphoma: two distinct clinicopathological variants with different outcomes.

BACKGROUND: ALK-negative anaplastic large cell lymphoma associated with breast implant (i-ALCL) has been recently recognized as a distinct entity. Among 43 830 lymphomas registered in the French Lymphopath network since 2010, 300 breast lymphomas comprising 25 peripheral T-cell lymphomas (PTCL) were reviewed. Among PTCL, ALK-negative ALCL was the most frequent and all of them were associated with breast implants. PATIENTS AND METHODS: Since 2010, all i-ALCL cases were collected from different institutions through Lymphopath. Immuno-morphologic features, molecular data and clinical outcome of 19 i-ALCLs have been retrospectively analyzed. RESULTS: The median age of the patients was 61 years and the median length between breast implant and i-ALCL was 9 years. Most implants were silicone-filled and textured. Implant removal was performed in 17 out of 19 patients with additional treatment based on mostly CHOP or CHOP-like chemotherapy regimens (n = 10/19) or irradiation (n = 1/19). CHOP alone or ABVD following radiation without implant removal have been given in two patients. The two clinical presentations, i.e. effusion and less frequently tumor mass correlated with distinct histopathologic features: in situ i-ALCL (anaplastic cell proliferation confined to the fibrous capsule) and infiltrative i-ALCL (pleomorphic cells massively infiltrating adjacent tissue with eosinophils and sometimes Reed-Sternberg-like cells mimicking Hodgkin lymphoma). Malignant cells were CD30-positive, showed a variable staining for EMA and were ALK negative. Most cases had a cytotoxic T-cell immunophenotype with variable T-cell antigen loss and pSTAT3 nuclear expression. T-cell receptor genes were clonally rearranged in 13 out of 13 tested cases. After 18 months of median follow-up, the 2-year overall survival for in situ and infiltrative i-ALCL was 100% and 52.5%, respectively. CONCLUSIONS: In situ i-ALCLs have an indolent clinical course and generally remain free of disease after implant removal. However, infiltrative i-ALCLs could have a more aggressive clinical course that might require additional therapy to implant removal.

Synergistic antitumor effect of TRAIL and adriamycin on the human breast cancer cell line MCF-7

The aim of the present study was to determine the effect of the combination of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) and adriamycin (ADM) on the human breast cancer cell line MCF-7 and to identify potential mechanisms of apoptosis. Cell viability was analyzed by the MTT assay and the synergistic effect was assessed by the Webb coefficient. Apoptosis was quantified using the annexin V-FITC and propidium iodide staining flow cytometry. The mRNA expression of TRAIL receptors was measured by RT-PCR. Changes in the quantities of Bax and caspase-9 proteins were determined by Western blot. MCF-7 cells were relatively resistant to TRAIL (IC50 >10 µg/mL), while MCF-7 cells were sensitive to ADM (IC50 <10 µg/mL). A subtoxic concentration of ADM (0.5 µg/mL) combined with 0.1, 1, or 10 µg/mL TRAIL had a synergistic cytotoxic effect on MCF-7 cells, which was more marked with the combination of TRAIL (0.1 µg/mL) and ADM (0.5 µg/mL). In addition, the combined treatment with TRAIL and ADM significantly increased cell apoptosis from 9.8% (TRAIL) or 17% (ADM) to 38.7%, resulting in a synergistic apoptotic effect, which is proposed to be mediated by up-regulation of DR4 and DR5 mRNA expression and increased expression of Bax and caspase-9 proteins. These results suggest that the combination of TRAIL and ADM might be a promising therapy for breast cancer.

RNAi-mediated knockdown of FANCF suppresses cell proliferation, migration, invasion, and drug resistance potential of breast cancer cells

Fanconi anemia complementation group F protein (FANCF) is a key factor, which maintains the function of FA/BRCA, a DNA damage response pathway. However, the functional role of FANCF in breast cancer has not been elucidated. We performed a specific FANCF-shRNA knockdown of endogenous FANCF in vitro. Cell viability was measured with a CCK-8 assay. DNA damage was assessed with an alkaline comet assay. Apoptosis, cell cycle, and drug accumulation were measured by flow cytometry. The expression levels of protein were determined by Western blot using specific antibodies. Based on these results, we used cell migration and invasion assays to demonstrate a crucial role for FANCF in those processes. FANCF shRNA effectively inhibited expression of FANCF. We found that proliferation of FANCF knockdown breast cancer cells (MCF-7 and MDA-MB-435S) was significantly inhibited, with cell cycle arrest in the S phase, induction of apoptosis, and DNA fragmentation. Inhibition of FANCF also resulted in decreased cell migration and invasion. In addition, FANCF knockdown enhanced sensitivity to doxorubicin in breast cancer cells. These results suggest that FANCF may be a potential target for molecular, therapeutic intervention in breast cancer.

Role of the protein tyrosine phosphatase DEP-1 in Src activation and the mediation of biological cell functions of endothelial and breast cancer cells

L’implication des protéines tyrosines phosphatases (PTPs) dans la régulation de la signalisation et la médiation des fonctions cellulaires a été bien établie dans les dernières années. Cependant, les mécanismes moléculaires par lesquels les PTPs régulent les processus fondamentaux tels que l’angiogenèse demeurent méconnus. Il a été rapporté que l’expression de la PTP DEP-1 (Density-enhanced phosphatase 1) augmente avec la densité cellulaire et corrèle avec la déphosphorylation du récepteur VEGFR2. Cette déphosphorylation contribue à l’inhibition de contact dans les cellules endothéliales à confluence et diminue l’activité du VEGFR2 en déphosphorylant spécifiquement ses résidus catalytiques Y1054/1059. De plus, la plupart des voies de signalisation en aval du VEGFR2 sont diminuées sauf la voie Src-Gab1-AKT. DEP-1 déphosphoryle la Y529 de Src et contribue à la promotion de la survie dans les cellules endothéliales. L’objectif de cette thèse est de mieux définir le rôle de DEP-1 dans la régulation de l’activité de Src et les réponses biologiques dans les cellules endothéliales. Nous avons identifié les résidus Y1311 et Y1320 dans la queue C-terminale de DEP-1 comme sites majeurs de phosphorylation en réponse au VEGF. La phosphorylation de ces résidus est requise pour l’activation de Src et médie le remodelage des jonctions cellules-cellules dépendantes de Src. Ce remodelage induit la perméabilité, l’invasion et la formation de capillaires en réponse au VEGF. Nos résultats démontrent que la phosphorylation de DEP-1 sur résidu tyrosine est requise pour diriger la spécificité de DEP-1 vers son substrat Src. Les travaux révèlent pour la première fois un rôle positif de DEP-1 sur l’induction du programme angiogénique des cellules endothéliales. En plus de la phosphorylation sur tyrosine, DEP-1 est constitutivement phosphorylé sur la thréonine 1318 situé à proximité de la Y1320 en C-terminal. Cette localisation de la T1318 suggère que ce résidu pourrait être impliqué dans la régulation de la Y1320. En effet, nous avons observé que la T1318 de DEP-1 est phosphorylée potentiellement par CK2, et que cette phosphorylation régule la phosphorylation de DEP-1 sur tyrosine et sa capacité de lier et d’activer Src. En accord avec ces résultats, nos travaux révèlent que la surexpression du mutant DEP-1 T1318A diminue le remodelage des jonctions cellules-cellules et par conséquent la perméabilité. Nos résultats suggèrent donc que la T1318 de DEP-1 constitue un nouveau mécanisme de contrôle de la phosphorylation sur tyrosine et que ceci résulte en l’activation de Src et l’induction des fonctions biologiques des cellules endothéliales en réponse au VEGF. Suite à ces travaux dans les cellules endothéliales qui démontrent un rôle positif de DEP-1 dans la médiation des réponses angiogéniques, nous avons voulu approfondir nos connaissances sur l’implication potentielle de DEP-1 dans les cellules cancéreuses où l’activité de Src est requise pour la progression tumorale. Malgré le rôle connu de DEP-1 comme suppresseur tumoral dans différents types de cancer, nous avons émis l’hypothèse que DEP-1 pourrait promouvoir les fonctions biologiques dépendantes de Src telles que la migration et l’invasion dans les cellules cancéreuses. Ainsi, nous avons observé que l’expression de DEP-1 est plus élevée dans les lignées basales de cancer du sein qui sont plus invasives comparativement aux lignées luminales peu invasives. Dans les lignées basales, DEP-1 active Src, médie la motilité cellulaire dépendante de Src et régule la localisation des protéines impliquées dans l’organisation du cytosquelette. L’analyse d’un micro-étalage de tissu a révélé que l’expression de DEP-1 est associée avec une réduction tendencielle de survie des patients. Nos résultats proposent donc, un rôle de promoteur tumoral pour DEP-1 dans la progression du cancer du sein. Les travaux présentés dans cette thèse démontrent pour la première fois que DEP-1 peut agir comme promoteur des réponses angiogéniques et du phénotype pro-invasif des lignées basales du cancer du sein probablement du à sa capacité d’activer Src. Nos résultats suggèrent ainsi que l’expression de DEP-1 pourrait contribuer à la progression tumorale et la formation de métastases. Ces découvertes laissent donc entrevoir que DEP-1 représente une nouvelle cible thérapeutique potentielle pour contrer l’angiogenèse et le développement du cancer.

Effects of 17β-estradiol and tamoxifen on the induction of chromosomal abnormalities and on the expression of her2 gene in breast cancer cell lines

This study covers an area of great importance in the research of breast cancer, related to the study of the effects of both estrogens (E2) and anti-estrogens (Tamoxifen) on chromosomes and of modulation of gene expression. Considering that breast cancer is a very heterogeneous disease and that patients respond differently to treatment, the identification of chromosomal abnormalities as well as genes responsive to 17β-estradiol (E2) and Tamoxifen (TAM) could provide the necessary framework to understand the complex effects of this hormone in target cells and could explain, at least in part, the development of cellular resistance to TAM treatment and the subsequent best therapeutic option. In this order of ideas, we determined the effects of E2 and TAM on the chromosomes and on the modulation of gene expression in four breast cancer cell lines, which represent three of the five subtypes of breast cancer known at present. The results are presented in six chapters - each one has a group of the results achieved around the cytogenetic characteristics and gene expression profiles of four cell lines and the effects of E2 and TAM incubation on those. The first chapter describes the main features of breast cancer, furthering the use and effects of E2 and TAM treatment.

Oestrogenic activity of p-hydroxybenzoic acid (common metabolite of paraben esters) and methylparaben in human breast cancer cell lines

This paper addresses the question of whether p-hydroxybenzoic acid, the common metabolite of parabens, possesses oestrogenic activity in human breast cancer cell lines. The alkyl esters of p-hydroxybenzoic acid (parabens) are used widely as preservatives in consumer products to which the human population is exposed and have been shown previously to possess oestrogenic activity and to be present in human breast tumour tissue, which is an oestrogen-responsive tissue. Recent work has shown p-hydroxybenzoic acid to give an oestrogenic response in the rodent uterotrophic assay. We report here that p-hydroxybenzoic acid possesses oestrogenic activity in a panel of assays in human breast cancer cell lines. p-Hydroxybenzoic acid was able to displace [H-3]oestradiol from cytosolic oestrogen receptor of MCF7 human breast cancer cells by 54% at 5 x 10(6)-fold molar excess and by 99% at 10(7)-fold molar excess. It was able to increase the expression of a stably integrated oestrogen responsive reporter gene (ERE-CAT) at a concentration of 5 x 10(-4) M in MCF7 cells after 24 h and 7 days, which could be inhibited by the anti-oestrogen ICI 182 780 (Faslodex, fulvestrant). Proliferation of two human breast cancer cell lines (MCF7, ZR-75-1) could be increased by 10(-5) M p-hydroxybenzoic acid. Following on from previous studies showing a decrease in oestrogenic activity of parabens with shortening of the linear alkyl chain length, this study has compared the oestrogenic activity of p-hydroxybenzoic acid where the alkyl grouping is no longer present with methylparaben, which has the shortest alkyl group. Intrinsic oestrogenic activity of p-hydroxybenzoic acid was similar to that of methylparaben in terms of relative binding to the oestrogen receptor but its oestrogenic activity on gene expression and cell proliferation was lower than that of methylparaben. It can be concluded that removal of the ester group from parabens does not abrogate its oestrogenic activity and that p-hydroxybenzoic acid can give oestrogenic responses in human breast cancer cells. Copyright (C) 2005 John Wiley & Sons, Ltd.

Crosstalk with insulin and dependence on PI3K/Akt/mTOR rather than MAPK pathways in upregulation of basal growth following long-term oestrogen deprivation in three human breast cancer cell lines

Background: MCF-7, T-47-D, ZR-75-1 human breast cancer cell lines are dependent on oestrogen for growth but can adapt to grow during long-term oestrogen deprivation. This serves as a model for identification of therapeutic targets in endocrine-resistant breast cancer. Methods: An overlooked complication of this model is that it involves more than non-addition of oestrogen, and inadequate attention has been given to separating molecular events associated with each of the culture manipulations. Results: Insulin and oestradiol were shown to protect MCF-7 cells against upregulation of basal growth, demonstrating a crosstalk in the growth adaptation process. Increased phosphorylation of p44/42MAPK and c-Raf reflected removal of insulin from the medium and proliferation of all three cell lines was inhibited to a lesser extent by PD98059 and U0126 following long-term oestrogen/insulin withdrawal, demonstrating a reduced dependence on the MAPK pathway. By contrast, long-term oestrogen/insulin deprivation did not alter levels of phosphorylated Akt and did not alter the dose-response of growth inhibition with LY294002 in any of the three cell lines. The IGF1R inhibitor picropodophyllin inhibited growth of all MCF-7 cells but only in the long-term oestrogen/insulin-deprived cells was this paralleled by reduction in phosphorylated p70S6K, a downstream target of mTOR. Long-term oestrogen/insulin-deprived MCF-7 cells had higher levels of phosphorylated p70S6K and developed increased sensitivity to growth inhibition by rapamycin. Conclusions: The greater sensitivity to growth inhibition by rapamycin in all three cell lines following long-term oestrogen/insulin deprivation suggests rapamycin-based therapies might be more effective in breast cancers with acquired oestrogen resistance. Keywords Akt, breast cancer cells, endocrine resistance, insulin, MAPK, MCF-7 cells, mTOR, oestrogen, oestrogen-deprived, PI3K, picropodophyllin, rapamycin, T-47-D cells, ZR-75-1 cells

Chickpea (Cicer arietinum) and other plant-derived protease inhibitor concentrates inhibit breast and prostate cancer cell proliferation in vitro

The soybean-derived protease inhibitor, Bowman-Birk inhibitor (BBI), is currently showing great promise as a novel cancer chemopreventive agent. In contrast to the wealth of research conducted on this compound, the anticancer effects of protease inhibitors isolated from other leguminous sources have received limited attention. In the current study, 7 protease inhibitor concentrates (PICs) were isolated from various leguminous sources (including soybean) and characterized. The effects of PICs on the proliferation of breast and prostate cancer cells were investigated in vitro. Chickpea PIC significantly inhibited the viability of MDA-MB-231 breast cancer and PC-3 and LNCaP prostate cancer cells at all concentrations tested (25-400 μg/ml). In addition, kidney bean (200, 400 μg/ml), soybean (50, 100 μg/ml), and mungbean (100, 200 μg/ml) PICs inhibited LNCaP cell viability. These findings suggest that leguminous PICs may possess similar anticancer properties to that of soybean BBI and deserve further study as possible chemopreventive agents.

Diallyl disulfide-induced apoptosis in a breast-cancer cell line (MCF-7) may be caused by inhibition of histone de-acetylation

The health benefits of garlic have been proven by epidemiological and experimental studies. Diallyl disulphide (DADS), the major organosulfur compound found in garlic oil, is known to lower the incidence of breast cancer both in vitro and in vivo. The studies reported here demonstrate that DADS induces apoptosis in the MCF-7 breast-cancer cell line through interfering with cell-cycle growth phases in a way that increases the sub-G0 population and substantially halts DNA synthesis. DADS also induces phosphatidylserine (PS) translocation from the inner to the outer leaflet of the plasma membrane and activates caspase-3. Further studies revealed that DADS modulates the cellular levels of Bax, Bcl-2, Bcl-xL and Bcl-w in a dose-dependent manner, suggesting the involvement of Bcl-2 family proteins in DADS induced apoptosis. Histone deacetylation inhibitors (HDACi) are known to suppress cancer growth and induce apoptosis in cancer cells. Here it is shown that DADS has HDACi properties in MCF-7 cells as it lowers the removal of an acetyl group from an acetylated substrate and induces histone-4 (H4) hyper-acetylation. The data thus indicate that the HDACi properties of DADS may be responsible for the induction of apoptosis in breast cancer cells.