A guide to cell-based assays in drug discovery

©2016 Macmillan Publishers Limited, part of Springer Nature.

In a recent perspective article published on Nature Reviews Drug Discovery (link), various members of the European Cell-Based Assays Interest Group summarized limitations of traditional cell-based disease models and discuss how patient-derived cultures, induced pluripotent stem cell (iPSC) technology and 3D co-cultures, complemented by single-cell imaging, microfluidics and gene editing technologies, could improve clinical significance of  preclinical drug testing. In this view, limitations of standard cell-line screens and in vivo xenografts contributed to the failure of many drug candidates that did not show clinical efficacy:
​

---

• long-term culture of cell lines is likely to result in a drift in their genetic and epigenetic characteristics;
• common media containing fetal serum and high concentrations of nutrients are not designed to promote cellular identity;
• oxygen levels in standard incubators are higher than in organs and tissues and affect cell metabolism and ROS production;
• plastic substrates on which cells are grown as monolayers are resistant to mechanical stresses, unlike most substrates in the human body;
• absence of extracellular matrix prevents activation of signaling pathways that would have profound impact on cell function;
• cultures of a single cell type prevent to take into consideration biological process which involve the interaction of multiple cell types.

The design of new cell-based models should promote a more physiologically relevant environment and recapitulate key features of the disease of interest, as intratumour heterogeneity, poor drug penetrance, host-stroma-tumour interactions, and the cancer stem cell niche:

1. Patient-derived primary cancer cell cultures should overcome various of the disadvantages of cell lines, but it would be needed a robust and flexible culture method. Indeed, potential drugs have been already tested against well-characterized patient-derived primary cultures, but without a direct impact on treatment. Furthermore, application of high-throughput screenings using primary cultures is still challenging.
2. In diseases where tissue availability is a limiting factor for drug screening, iPSC technology would bring several advantages. iPSC derived cultures of cardiomyocites, neurons, as well as lung and intestinal tissues, have been developed as several disease models and have been already used in drug screening. However, the usage of iPSC technology focused on monogenic, hereditary diseases rather than spontaneous forms of diseases which are characterized by multiple genetic and epigenetic alterations.
3. Three-dimensional (3D) cell culture models would overcome various of the limitations of classic cultures, favouring cell-cell and cell-ECM interactions. However, animal-derived basement membrane extract (BME) hydrogels used to form 3D organoids have batch-to-batch variations which prevent to obtain reproducible results. Synthetic biomaterials have been developed, but the lack of ECM proteins still limits their physiological relevance.
4. Organ-on-a-chip and microfluidic technologies can offer higher levels of biological complexity and clinical relevance with the potential for high-throughput screenings, although these models are still early in their development.

These new cell-based assays can be complemented by recent technological advances that would exploit their potential:

• CRISPR-Cas9 genome editing allows to deepen in the mechanism of action of drugs, as well as to develop new cell-based assay models when combined with iPSC technology. It can also be used to directly mutate oncogenes and tumor suppressor genes to develop custom-designed ex vivo models to focus on signaling cascades.
• Automated microscopic imaging provides functional data points and allows cell-based screening assays with 3D models and heterogeneous co-cultures. 
• Integrative bioinformatic tools support a move to  a system pharmacology approach, wherein phenotypic response is linked to genome, epigenome, transcriptome and proteome, which allows to view drug targets as a part of an integrated biological network.

​
The authors conclude that these advances are well placed to support a new era of phenotypic drug discovery (generation of hit or lead molecules without any prior knowledge of the target) and to enhance also conventional target-directed drug discovery, if more predictive in vitro models of greater clinical relevance are developed.

We fully agree with the conclusions stated by the authors. Indeed, we developed a cytotoxicity and proliferation assay that take advantage of live-cell imaging to quantify at the same time number of live and dead cells before and after treatment and provide insight on the mechanism of action of drugs. Also, we offer a flexible and reproducible cytotoxicity assay on 3D cultures using a synthetic biomaterial to let reproducible formation of a single spheroid in each well. Despite cell:ECM interaction is missing, it provides useful informations on the role of cell:cell interactions on drug potency and on drug penetrance. 

Finally, we would like to emphasize that, despite it is evident that classic cell-based assays can not fully recapitulate clinical efficacy of tested drugs, a big issue on the reproducibility of cancer papers has been raised in the last years. In order to assess the clinical relevance of in vitro studies, it would be first necessary to cross out all the unreproducible results, which is not possible at the moment. If the majority of cancer papers can not be reproduced, as Amgen researchers in 2012 and Baker and Dolgin recently suggested, failure of in vitro studies to translate into the clinical practice is likely to be consequence also of the poor quality behind many of them.

References
​1. Gunjan Sinha; Downfall of Iniparib: A PARP Inhibitor That Doesn’t Inhibit PARP After All. J Natl Cancer Inst 2014; 106 (1): djt447. doi: 10.1093/jnci/djt447
2. Begley CG & Ellis LM. Drug development: Raise standards for preclinical cancer research. Nature 483, 531–533 (29 March 2012) doi:10.1038/483531a
3. Baker M & Dolgin E. Cancer reproducibility project releases first results. Nature 541, 269–270 (19 January 2017) doi:10.1038/541269a