br contact mediated stromal e
contact-mediated stromal eﬀects on tumor progression and drug ef-fectiveness. In addition to contact-related mediation, secreted factors from stromal KX2-391 can also contribute to cancer evasion of drug-induced apoptosis. One well-studied factor is hepatocyte growth factor (HGF) secreted by TAFs. HGF not only increases drug resistance by activating both mitogen-activated protein kinase (MAPK) and phosphatidylino-sitol-3-OH kinase (PI(3)K)-AKT signaling pathways, but also stimulates expression of self-renewal genes to expand the cancer stem cell popu-lation that naturally possesses this drug resistant feature [4,76]. The 3D coculture tumor model can easily capture the secreted factors-regulated drug resistance by tracking the factor expression and exchange during drug screening.
In vitro assessment of drug eﬃcacy in cell-based assays highly relies on an accurate analysis of cellular responses such as growth, mor-phology, migration and gene expression pattern. However, cell detec-tion remains a challenge for 3D coculture assays . A variety of LIVE/DEAD reagents (e.g. WST-1 and calcein-AM) have been utilized for viability assessment in coculture tumor models [77,78]. Although comparable results versus 2D systems were obtained in some studies [30,79], conventional viability assays are unable to distinguish two or more cell types that are mixed within the same 3D matrix or multi-cellular aggregates. The reagent permeability into the interior of tumor spheroids is also limited. Moreover, phase contrast microcopy, which was widely used to monitor the morphological change of tumor spheroids in 3D assays, often resulted in false positive hits and mis-judgments . To address these problems, immunocytochemistry (ICC) or immunohistochemistry (IHC) staining integrated with confocal microscopy have also been used . However, these methods involve complicated operational procedures and high costs, which greatly limit their application in automation and high throughput platforms. On the other hand, most of the traditional detection methods require cell lysis for data capture and thus consume intensive labor for screening and yet can only provide endpoint data. In this work, we developed a simple and non-invasive dual fluorescence system that can simultaneously monitor the growth of MCF-7 and NIH-3T3 cells in a coculture. The EGFP and Ds-red fluorescence signals can be detected with a commer-cial plate reader for high throughput drug screening without significant cross interference in the PET-based 3D system (Fig. 1). Also, the cyto-toxic eﬀects of diﬀerent antitumor agents on both MCF-7 and NIH-3T3 cells were successfully captured by changes in the culture fluorescence (Figs. 4–6). These results demonstrated that the dual fluorescence system can provide high-sensitivity and robust cell quantification in the tumor coculture model.
Lastly, the potential side eﬀects of tamoxifen, oxaliplatin and cis-platin were evaluated using the cytotoxicity assay of NIH-3T3 cells in 3D coculture (Fig. 6). Their stronger cytotoxic eﬀects against MCF-7 cells than NIH-3T3 cells suggested that these drugs were relatively cancer-specific agents. In the past decade, mouse fibroblasts have been increasingly used as normal cell line to evaluate the cancer drug toxi-city. For example, Flis and Splawinski utilized Balb/c 3T3 cells as re-ference to evaluate the cancer specificity of COX inhibitor sulindac
IC50 (μM) of tamoxifen, oxaliplatin and cisplatin to MCF-7 and NIH-3T3 cells cultured in diﬀerent conditions.
Cell type MCF-7
Drug Tamoxifen Oxaliplatin Cisplatin
Tamoxifen Oxaliplatin Cisplatin Viabilitya
a IC50 estimated from viable cell count at 4 days.
b IC50 estimated from the specific growth rate during the 4-day culture period.