September 15, 2014 (Vol. 34, No. 16)

MaryAnn Labant

Complex biology cannot be modeled with simple, standalone in vitro systems. More sophisticated, physiologically relevant systems are needed.

For example, systems that maintain cellular environments that closely mimic disease may improve translation of results, increase clinical development efficiency, and reduce downstream drug attrition.

Tissue models are preferred but in some circumstance are not available. Optionally, induced pluripotent stem (iPS) cells and other kinds of stem cells can be differentiated into various cell populations. After a cell population is derived and functionality is shown to mimic in vivo conditions, the cells can be used to screen a compound library to identify promising compounds. Although differentiation protocols can take up to two months, a large number of cells can be produced.

“A lot of the work in iPS cells goes into deriving them and putting together protocols to make sure the resultant cells are as close as possible to the adult ones. We are adapting a frozen-cell approach to remove batch-to-batch variation in stem cells. Then we know each aliquot is exactly the same,” said Mark Slack, Ph.D., vice president of in vitro pharmacology at Evotec.

“High-throughput screening is still the source of hits for our clinical programs,” he added. “Our desire is to develop more complex human systems to demonstrate that the compound does what we want it to do. If possible, we use primary human tissue.”

Although it is too early to say if cellular models have an effect on drug attrition, more assays have been added to the testing cascade to help mitigate risk. Emphasis is placed on demonstrating target engagement and modulation. Program dependent, new assays are developed and integrated into the cascade.

Improved bioinformatics tools incorporating biological knowledge are needed. The ability to efficiently mine and annotate mass datasets of gene-expression profiling and proteomics data remains a bottleneck.

In Vitro Phenotypes

Human-stem-cell-derived cells provide a way to access patient-specific backgrounds, such as phenotypes of non- or super-responders in clinical trials. Cell samples taken from the patients are reprogrammed into iPS cells, and then used to study differences relevant to the pathology.

“This is an incredible opportunity to molecularly understand why a patient responds or not, and it may allow us to derive a biomarker,” explained Roberto Iacone, M.D., Ph.D., senior principal scientist, Roche Innovation Center Basel.

“Ultimately, the goal is to correlate the in vitro phenotype to the clinical phenotype. Then you can use the biomarker to preselect a population. If we know patients will be responsive, the chance of seeing a difference in the clinical trial will be much higher.”

This goal may be achieved with the assistance of CRISPR technology for genome editing. Were a battery of CRISPR guide RNAs to be created, a library of mutations could be developed to investigate and correlate the loss of a gene with an in vitro phenotype, and allow target identification early in preclinical work.

Cell models derived from iPS cells are not as mature as the cells in the body, and the pathways in these models may lack the nuances seen in vivo. Although it is generally agreed that they are better than rodent cells, iPS cells must be pushed toward a specific maturation endpoint optimized for specific individual bioassays.

“Tissue contains many cell types that interact. Three-dimensional organoids, which are not organs but have some structure, may be feasible models in five years. If the different cells are orchestrating together, they will produce the right amount of proteins that they need to mature, and we can ask more complex questions,” added Dr. Iacone. “But it takes time. We are still learning how to organize these organoids.”

Controlling Cell Variability

Cells respond to their environment. Unless they are properly controlled during continuous culture, cells can be a major contributor to bioassay variability.

Parameters that may affect cells in continuous culture include seeding and harvest densities, doubling times, and passage frequencies, as well as the freezing process. Upfront characterization of such parameters helps determine the conditions that control the cells and their impact on bioassay performance.

The best approaches are quantitative, not qualitative. “Upfront work [points to] the easiest route. You do not want to waste resources on troubleshooting assays down the road,” commented Teresa Surowy, Ph.D., research manager, bioassays, Promega.

“You need to understand which steps in the cell-preparation process are most important and to optimize conditions. At Promega, we sometimes refer to our ‘thaw-and-use’ cells as reagents. We want them to be as good as a biochemical reagent in terms of consistency of performance.”

Lots, or banks, of single-use vials of thaw-and-use cells provide convenience and uniformity run to run; the same cell lot can be used for bioassays over an extended time period.

For example, classic antibody-dependent cell-mediated cytotoxicity (ADCC) bioassays use natural killer (NK) cells collected from human donors as the effector cells, which mediate ADCC. These can vary in ability to perform ADCC, increasing assay variability and making it more difficult to reliably quantitate activity.

“ADCC bioassays are also quite tedious,” says Ulrike Herbrand, Ph.D., scientific officer in the bioanalytics department at Charles River Biopharmaceutical Services. “Fortunately, the acceptance of alternatives such as reporter-based assays is rising within the scientific and regulatory authority communities, because the suitability of such approaches could be proven successfully.

The Promega ADCC Reporter Bioassay replaces NK cells with a dual-engineered thaw and use cell line that reports on pathway activation of ADCC mechanism of action.

The dual-engineered effector cell line has one of either variant of FcγRIIIa receptor, the receptor in humans that mediates the ADCC response, as well as an engineered NFAT (nuclear factor of activated T cells) response element linked upstream of firefly luciferase gene. Target cell-bound antibody binds to FcγRIIIa receptor, the V158 or F158 variant, and pathway activation occurs within the effector cells. The NFAT response is activated during the ADCC pathway activation, just as in classic ADCC; the amount of luciferase produced reflects the amount of activation.

The use of genetically engineered FcγRIIIa V158 and F158 effector cell lines in ADCC reporter bioassays offers the unique opportunity to evaluate the impact of the FcγRIIIa variants on the in vitro bioactivity of therapeutic antibodies, according to Dr. Herbrand. “This information is highly beneficial for biocomparability studies as well as for the evaluation of the impact of process changes.” she explains.


Artist’s rendering of Promega’s ADCC Reporter Bioassay depicts activation and readout from NFAT (nuclear factor of activated T cells) pathway in the dual-engineered effector cell line. ADCC: antibody-dependent cell-mediated cytotoxicity.

Improving Gene Delivery Systems

Enhancements to the BacMam technology, a gene delivery system for cultured mammalian cells using baculovirus, boost viral tropism. Modifications to the gp64 coat glycoprotein in the vector backbone permit more efficient delivery to primary, stem, and many immortalized cancer cells.

In addition, the woodchuck hepatitis virus post-transcriptional regulatory (WPRE)element was added to increase expression duration. BacMam now provides transient expression for a couple of weeks. Often with other transfection reagents, transduction only occurs for a couple of days unless a stable cell line is developed.

“This is unique. Most viral transfection methods lead to stable cell lines and integrate into the genome of the cell, which then leaves a footprint of the virus on the cells,” stated George Hanson, Ph.D., senior staff scientist, Life Technologies.

Essentially nontoxic, the cells are not adversely affected even when they have large amounts of BacMam particles. The system, which is capable of carrying a very large insert size, is amenable to human genes with introns 8–10 kb or larger. Published literature cites examples of BacMam delivering up to 38 kb to a cell.

Although some particles are commercially available, the technology does require upfront time investment in particle creation. However, once created, the particles can be added to a large amount of cells, enabling next-day compound library screening and negating the need to maintain multiple stable cell lines.

BacMam facilitates the study of calcium channels. These channels are composed of multiple subunits required for the channel to assemble and function correctly; some subunits are very large, up to 6 kb. Creating a stable cell line is difficult because multiple subunits need to be stably integrated and expressed at the right concentrations. The number of copies of subunits varies.

A more straightforward process, BacMam solutions can be developed for each subunit and added to the cells at the appropriate dilution.

Protein Interactions

Based on a proximity ligation assay that uses antibodies conjugated with oligos, the Duolink technology permits visualization and localization of a single protein-protein interaction per cell. A pair of primary antibodies is used, and secondary antibodies are added. If the antibodies are in close proximity (40 nm), the oligos attach. Rolling circle amplification amplifies the DNA at the site of interaction, which allows 500–1,000× detection probes to then recognize the single-stranded DNA, generating an intense fluorescent spot.

“Signal is only generated when two antibodies interact with two closely proximal epitopes. Consequentially, the noise in the assay is reduced enormously,” explained Caleb Hopkins, product manager, protein biology, Sigma Life Science. “Duolink can localize a small number of events in a cell, or in a subset of cells in a tissue sample. Time duration studies can also be undertaken to study transient interactions. And other interactions such as phosphorylation events can be examined as well.”

RNA interference (RNAi) set the stage for high-throughput cell-based assays; high-throughput screens are routinely done using lentiviral-delivered small hairpin RNA (shRNA). Genome-editing technologies and proteomic tools, such as Duolink, add to the spectrum of tests. Still, no single screening technology exists that will provide a final answer. The careful triaging of technology is important. Although they are becoming more elegant, tools are still evolving.

Sigma also offers cell engineering and assay services. Projects fall into two basic categories, knock-outs and knock-ins, and they can range in complexity from knock-out of a single gene to changing of a single base to introducing multiple different reporters in different loci in the same cell.

“We have developed a suite of techniques for screening cells for different genetic changes, allowing us to choose the assay appropriate for the job. We are able to identify cells with a single nucleotide change, even if this occurs at a frequency of less than 1 in 100,000,” asserted Ian Lyons, Ph.D., associate research fellow,  Sigma Life Science. “In this way, we can routinely introduce, and find, just about any change we want, and in almost any cell type.”

Biologics Drug Discovery

MedImmune’s biologics discovery process relies on phage and ribosome display or the more classic in vivo immunization route to generate antibodies. During the phage-display process, antibodies, in a single-chain fragment variable (scFv) format are expressed in Escherichia coli and extracted via an osmotic shock process using high-strength sugar/salt buffers and a series of centrifugation steps to yield crude scFv samples for testing in high-throughput assays.

According to Lesley Jenkinson, scientist 1, antibody discovery and protein engineering, MedImmune, the crude samples contain bacterial debris, endotoxin, as well as high concentrations of sugars and salts. All of these present tolerance issues in cell-based assays in terms of signal window and assay variability.

Traditionally, antibody high-throughput screening has been performed using biochemical assays that have a higher tolerance to this sample type. A newly developed process using a mammalian cell-expression system enables the expression of a large number of clones as immunoglobulin G (IgG) rather than scFv.

Advances in this system have included the use of new vectors with improved mammalian promoters that have high and consistent IgG expression levels, allowing assays to be run with a lower sample volume. In addition, IgG samples are expressed in an “assay friendly” tissue culture medium, which allows an overall reduction of the sample tolerance issues in cell-based assays.

An additional benefit of using the mammalian expression system is that the avidity of IgG binding compared to scFv can be more beneficial for binding to certain targets such as complex membrane proteins, allowing identification of hits that would have been missed in the monovalent-binding scFv format.

Tolerance issues have not been entirely solved; different cell lines respond to sample buffers in different ways. Sample tolerance must still be assessed during the assay-development phase to ensure assays are robust in the presence of the crude samples.


MedImmune’s biologics discovery process relies on phage and ribosome display or the more classic in vivo immunization route to generate antibodies.

Previous articleHuman Stem Cell Clock Reset to Zero
Next articleSix Websites You Need to Bookmark: September Picks