Search

Background: In vitro cell observation has been widely used by biologists and pharmacologists for screening molecule-induced effects on cancer cells. Computer-assisted time-lapse microscopy enables automated live cell imaging in vitro, enabling cell behavior characterization through image analysis, in particular regarding cell migration. In this context, 3D cell assays in transparent matrix gels have been developed to provide more realistic in vitro 3D environments for monitoring cell migration (fundamentally different from cell motility behavior observed in 2D), which is related to the spread of cancer and metastases. Methodology/Principal Findings: In this paper we propose an improved automated tracking method that is designed to robustly and individually follow a large number of unlabeled cells observed under phase-contrast microscopy in 3D gels. The method automatically detects and tracks individual cells across a sequence of acquired volumes, using a template matching filtering method that in turn allows for robust detection and mean-shift tracking. The robustness of the method results from detecting and managing the cases where two cell (mean-shift) trackers converge to the same point. The resulting trajectories quantify cell migration through statistical analysis of 3D trajectory descriptors. We manually validated the method and observed efficient cell detection and a low tracking error rate (6%). We also applied the method in a real biological experiment where the pro-migratory effects of hyaluronic acid (HA) were analyzed on brain cancer cells. Using collagen gels with increased HA proportions, we were able to evidence a dose-response effect on cell migration abilities. Conclusions/Significance: The developed method enables biomedical researchers to automatically and robustly quantify the pro- or anti-migratory effects of different experimental conditions on unlabeled cell cultures in a 3D environment. © 2011 Adanja et al., SCOPUS: ar.j, info:eu-repo/semantics/published

The aim of this study was to compare the efficacy of local drug delivery by inhalation to intravenous delivery in a B16F10 melanoma metastatic lung model. Temozolomide was formulated as a suspension, which was elaborated and evaluated in terms of particle size, shape and agglomeration. An endotracheal administration device was used to aerosolise the suspension. This mode of delivery was evaluated at different temozolomide concentrations and was optimized for the uniformity of delivered dose, the droplet size distribution and the distribution of droplets in vivo. Of the particles in the stabilised suspension, 79% were compatible with the human respirable size range, and this formulation retained 100% in vitro anticancer activity as compared to temozolomide alone in three distinct cancer cell lines. The pulmonary delivery device provided good reproducibility in terms of both the dose delivered and the droplet size distribution. Most of the lung tissues that were exposed to aerosol droplets contained the particles, as revealed by fluorescent microscopy techniques. The global in vivo antitumour activity of the inhaled temozolomide provided a median survival period similar to that for intravenous temozolomide delivery, and three out of 27 mice (11%) survived with almost complete eradication of the lung tumours. The present study thus shows that inhalation of a simple liquid formulation is well tolerated and active against a very biologically aggressive mouse melanoma pulmonary pseudo-metastatic model. This inhalation delivery could be used to deliver other types of anticancer drugs., JOURNAL ARTICLE, info:eu-repo/semantics/published

We present here a method giving a robust segmentation for in vitro cells observed under standard phase-contrast microscopy. We tackle the problem using the watershed transform. Watershed transform is known for its ability to generate closed contours and its extreme sensitivity to image borders. One main drawback of this method is over- segmentation. In order to circumvent this, marked watershed based on the "modified gradient" method has been developed. However, the choice of the watershed mark locations is critical and their inadequacy may cause wrong results. Similarly to randomization and combination procedures used in the machine learning field, the present paper promotes the use of an assembly of marked watershed transforms, in order to increase the segmentation robustness. This results in the definition of candidate segmentations margins (expressed in terms of object border confidence) from which final segmentation can be chosen by means of thresholding.