No abstract

BackgroundReliable human in vitro blood–brain barrier (BBB) models suitable for high-throughput screening are urgently needed in early drug discovery and development for assessing the ability of promising bioactive compounds to overcome the BBB. To establish an improved human in vitro BBB model, we compared four currently available and well characterized immortalized human brain capillary endothelial cell lines, hCMEC/D3, hBMEC, TY10, and BB19, with respect to barrier tightness and paracellular permeability. Co-culture systems using immortalized human astrocytes (SVG-A cell line) and immortalized human pericytes (HBPCT cell line) were designed with the aim of positively influencing barrier tightness.MethodsTight junction (TJ) formation was assessed by transendothelial electrical resistance (TEER) measurements using a conventional epithelial voltohmmeter (EVOM) and an automated CellZscope system which records TEER and cell layer capacitance (CCL) in real-time.Paracellular permeability was assessed using two fluorescent marker compounds with low BBB penetration (sodium fluorescein (Na-F) and lucifer yellow (LY)). Conditions were optimized for each endothelial cell line by screening a series of 24-well tissue culture inserts from different providers. For hBMEC cells, further optimization was carried out by varying coating material, coating procedure, cell seeding density, and growth media composition. Biochemical characterization of cell type-specific transmembrane adherens junction protein VE-cadherin and of TJ proteins ZO-1 and claudin-5 were carried out for each endothelial cell line. In addition, immunostaining for ZO-1 in hBMEC cell line was performed.ResultsThe four cell lines all expressed the endothelial cell type-specific adherens junction protein VE-cadherin. The TJ protein ZO-1 was expressed in hCMEC/D3 and in hBMEC cells. ZO-1 expression could be confirmed in hBMEC cells by immunocytochemical staining. Claudin-5 expression was detected in hCMEC/D3, TY10, and at a very low level in hBMEC cells. Highest TEER values and lowest paracellular permeability for Na-F and LY were obtained with mono-cultures of hBMEC cell line when cultivated on 24-well tissue culture inserts from Greiner Bio-one® (transparent PET membrane, 3.0 μm pore size). In co-culture models with SVG-A and HBPCT cells, no increase of TEER could be observed, suggesting that none of the investigated endothelial cell lines responded positively to stimuli from immortalized astrocytic or pericytic cells.ConclusionsUnder the conditions examined in our experiments, hBMEC proved to be the most suitable human cell line for an in vitro BBB model concerning barrier tightness in a 24-well mono-culture system intended for higher throughput. This BBB model is being validated with several compounds (known to cross or not to cross the BBB), and will potentially be selected for the assessment of BBB permeation of bioactive natural products.

No abstract

No abstract

In many real world applications, labeled data are in short supply. It often happens that obtaining labeled data in a new domain is expensive and time consuming, while there may be plenty of labeled data from a related but different domain. Traditional machine learning is not able to cope well with learning across different domains. In this paper, we address this problem for a text-mining task, where the labeled data are under one distribution in one domain known as in-domain data, while the unlabeled data are under a related but different domain known as out-of-domain data. Our general goal is to learn from the in-domain and apply the learned knowledge to out-of-domain. We propose a coclustering based classification (CoCC) algorithm to tackle this problem. Co-clustering is used as a bridge to propagate the class structure and knowledge from the in-domain to the out-of-domain. We present theoretical and empirical analysis to show that our algorithm is able to produce high quality classification results, even when the distributions between the two data are different. The experimental results show that our algorithm greatly improves the classification performance over the traditional learning algorithms.