Automated interpretation of time-lapse quantitative phase image by machine learning to study cellular dynamics during epithelial-mesenchymal transition.

SIGNIFICANCE: Machine learning is increasingly being applied to the classification of microscopic data. In order to detect some complex and dynamic cellular processes, time-resolved live-cell imaging might be necessary. Incorporating the temporal information into the classification process may allow for a better and more specific classification. AIM: We propose a methodology for cell classification based on the time-lapse quantitative phase images (QPIs) gained by digital holographic microscopy (DHM) with the goal of increasing performance of classification of dynamic cellular processes. APPROACH: The methodology was demonstrated by studying epithelial-mesenchymal transition (EMT) which entails major and distinct time-dependent morphological changes. The time-lapse QPIs of EMT were obtained over a 48-h period and specific novel features representing the dynamic cell behavior were extracted. The two distinct end-state phenotypes were classified by several supervised machine learning algorithms and the results were compared with the classification performed on single-time-point images. RESULTS: In comparison to the single-time-point approach, our data suggest the incorporation of temporal information into the classification of cell phenotypes during EMT improves performance by nearly 9% in terms of accuracy, and further indicate the potential of DHM to monitor cellular morphological changes. CONCLUSIONS: Proposed approach based on the time-lapse images gained by DHM could improve the monitoring of live cell behavior in an automated fashion and could be further developed into a tool for high-throughput automated analysis of unique cell behavior.

Transcriptional crosstalk between TGF-ß and stem cell pathways in tumor cell invasion: role of EMT promoting Smad complexes.

Tumor cells undergoing the epithelial-mesenchymal transition (EMT) acquire the capacity to migrate, invade the stroma and metastasize. EMT cells also acquire stem cell characteristics suggesting crosstalk between EMT and stem cell pathways and contribution of the EMT process to the generation of cancer stem cells. Indeed, transforming growth factor-beta (TGF-ß), a major inducer of EMT, cooperates with stem cell pathways like Wnt, Ras, Hedgehog and Notch to induce EMT. A molecular basis for this cooperative signaling is indicated by recent data showing that many EMT associated transcription factors like Snail1, Zeb1/2, Twist, ß-catenin, Lef/TCF, Foxc2 and AP-1 interact with Smads and form EMT promoting Smad complexes (EPSC) engaged in both repressing epithelial genes and activating mesenchymal genes. Thus, formation and activation of EPSC seems to represent a point of convergence between EMT and stem cell pathways. Here, we review our current understanding of the mechanisms involved in the transcriptional crosstalk between TGF-ß and stem cell pathways and discuss how a fundament for the activation of these mechanisms may lead to the induction of EMT in tumors.

A SNAIL1-SMAD3/4 transcriptional repressor complex promotes TGF-beta mediated epithelial-mesenchymal transition.

Epithelial-mesenchymal transition (EMT) is essential for organogenesis and is triggered during carcinoma progression to an invasive state. Transforming growth factor-beta (TGF-beta) cooperates with signalling pathways, such as Ras and Wnt, to induce EMT, but the molecular mechanisms are not clear. Here, we report that SMAD3 and SMAD4 interact and form a complex with SNAIL1, a transcriptional repressor and promoter of EMT. The SNAIL1-SMAD3/4 complex was targeted to the gene promoters of CAR, a tight-junction protein, and E-cadherin during TGF-beta-driven EMT in breast epithelial cells. SNAIL1 and SMAD3/4 acted as co-repressors of CAR, occludin, claudin-3 and E-cadherin promoters in transfected cells. Conversely, co-silencing of SNAIL1 and SMAD4 by siRNA inhibited repression of CAR and occludin during EMT. Moreover, loss of CAR and E-cadherin correlated with nuclear co-expression of SNAIL1 and SMAD3/4 in a mouse model of breast carcinoma and at the invasive fronts of human breast cancer. We propose that activation of a SNAIL1-SMAD3/4 transcriptional complex represents a mechanism of gene repression during EMT.

Neutralized adenovirus-immune complexes can mediate effective gene transfer via an Fc receptor-dependent infection pathway.

Neutralization of adenovirus (Ad) by anti-Ad neutralizing antibodies in serum involves formation of Ad-immune complexes that prevent the virus from interacting with target cells. We hypothesized that Ad-immune complexes likely contain viable Ad vectors which, although no longer capable of gaining access to receptors on target cells, may be able to express transgenes in cells bearing Fc receptors for immunoglobulins, i.e., that antibody-based "neutralization" of Ad vectors may be circumvented by the Fc receptor pathway. To test this hypothesis, we expressed the Fcgamma receptor IIA (FcgammaR) in A549 lung epithelial cells or human dermal fibroblasts and evaluated gene transfer in the presence of human neutralizing anti-Ad serum. FcgammaR-expressing cells bound and internalized copious amounts of Ad, with a distinct population of internalized Ad trafficking to the nucleus. The dose-response curves for inhibition of gene transfer revealed that FcgammaR-expressing cells required a more-than-10-fold higher concentration of anti-Ad serum to achieve 50% inhibition of Ad-encoded beta-galactosidase expression compared with non-FcgammaR-expressing cells. The discrepancy between neutralization of Ad during infection of FcgammaR-expressing cells and neutralization of Ad during infection of non-FcgammaR-expressing cells occurred with either heat-inactivated or non-heat-inactivated sera, was blocked by addition of purified Fc domain protein, and did not require the cytoplasmic domain of FcgammaR, suggesting that immune complex internalization proceeded via endocytosis rather than phagocytosis. FcgammaR-mediated infection by Ad-immune complexes did not require expression of the coxsackie virus-Ad receptor (CAR) since similar data were obtained when CAR-deficient human dermal fibroblasts were engineered to express FcgammaR. However, interaction of the Ad penton base with cell surface integrins contributed to the difference in neutralization between FcgammaR-expressing and non-FcgammaR-expressing cells. The data indicate that complexes formed from Ad and anti-Ad neutralizing antibodies, while compromised with respect to infection of non-FcgammaR-expressing target cells, maintain the potential to transfer genes to FcgammaR-expressing cells, with consequent expression of the transgene. The formation of Ad-immune complexes that can target viable virus to antigen-presenting cells may account for the success of Ad-based vaccines administered in the presence of low levels of neutralizing anti-Ad antibody.

Cytokine-mediated downregulation of coxsackievirus-adenovirus receptor in endothelial cells.

Endothelial cells have the ability to change their complement of cell surface proteins in response to inflammatory cytokines. We hypothesized that the expression of the coxsackievirus-adenovirus receptor (CAR), a viral receptor and putative cell-cell adhesion molecule, may be altered during the response of endothelial cells to inflammation. To test this hypothesis, we evaluated CAR protein and mRNA levels in human umbilical vein endothelial cells after they were exposed to tumor necrosis factor alpha, gamma interferon, or a combination of the two cytokines. Flow cytometric and Western blot analyses indicated that cytokine treatment led to a synergistic decrease in CAR protein expression. A Western blot analysis showed that CAR levels decreased to 16% +/- 4% or 1% +/- 4% of the CAR protein levels in untreated cells with either 24 or 48 h of cytokine treatment, respectively. Quantitative reverse transcription-PCR demonstrated that the combination treatment caused CAR mRNA levels to decrease to 21% +/- 12% or 5% +/- 3% of the levels in untreated cells after a 24- or 48-h cytokine treatment, respectively. Reduced CAR expression led to a decrease in adenovirus (Ad) binding of 80% +/- 3% (compared with untreated endothelial cells), with a subsequent decrease in Ad-mediated gene transfer that was dependent on the dose and duration of cytokine treatment but not on the dose of Ad. A similar decrease in CAR protein level and susceptibility to Ad infection was observed in human microvascular endothelial cells, while CAR expression on normal human bronchial epithelial cells or A549 lung epithelial cells was less affected by cytokine treatments. Taken together, the data demonstrate that inflammatory cytokines decrease CAR mRNA and protein expression with a concomitant decrease in Ad binding, reflecting the impact of cell physiology on the function of CAR and the potential effect of inflammation on the ability of Ad to transfer genes to endothelial cells.