Dr J. Mark Treherne, Dr Christian Pernstich and Dr Michael Jones examine how the controlled release of proteins will improve the reproducibility of cell-based drug discovery screens?
Cell-based screening systems are now routinely incorporated as pivotal decision-making assays much earlier into the drug discovery process across most therapeutic areas and increasingly so for more complex cellular assays. Although such assays systems can be more predictive of the desired therapeutic effects than simpler cell-based or cell-free screens, many apparently attractive new therapeutics continue to fail later in clinical trials. Complex assays now need to become sufficiently robust and reproducible to improve upon the current unacceptably high rates of compound attrition in downstream development. Production methods for expansion with bioprocessing technologies can result in significant batch-to-batch variations but precision control of protein concentrations, especially for growth factors, are emerging to overcome the constraints of more conventional bioprocessing methodologies. Here, we review the challenges and solutions required to solve the reproducibility concerns with the production of human cellular assay systems and their practical use in drug discovery. We propose that plug-and-play sustained release platforms for growth factors are reconfiguring the way that cell-based assay systems are produced to be deployed more effectively in drug discovery screens.
The biopharmaceutical sector is critically dependent on the productivity of its research expenditure to underpin future growth and has a mounting need to source innovative solutions for improving human cell-based assays. A report from Visiongain in 2019 estimated that the global drug discovery outsourcing market reached $23 billion in 2018 and forecasted a compound annual growth rate of 12% through to 2028, when the market could reach an estimated value of $67 billion. Despite these levels of investment, current therapeutic pipelines are still not meeting the expected clinical outcomes required for registering new drugs. Pressure to improve industrial productivity largely results from high levels of pipeline attrition with more than 90% of drugs tested in clinical trials not demonstrating statistically significant health benefits. The number of new drugs approved per billion US dollars spent on research has halved roughly every nine years from 1950 to 2010, falling by around 80-fold in inflation-adjusted terms1. The referenced 2012 review analysed the problem and introduced the concept of “Eroom’s Law”, which is Moore’s Law spelt backwards! The new ‘law’ referred to processes that are getting steadily slower and more difficult to execute over time (the opposite of Moore’s Law, which was originally the observation that the number of transistors in a dense integrated circuit doubled about every two years).
Although this may be an overly pessimistic analysis, the relatively few successful products that make it on to the market are having to compensate for too many failures, which are often weeded out too late in the lengthy development process.Late stage failures are the most expensive. A more recent analysis has illustrated how both assay validity and reproducibility correlate across a population of simulated screening and disease models2. The review’s authors hypothesised that screening and disease models with higher “predictive validity” are a credible and practical solution to the problem. They conclude that “perhaps there has also been too much enthusiasm for reductionist molecular models, which have insufficient predictive validity”, suggesting an increased role for more complex cell-based screening systems being incorporated as pivotal decision-making assays much earlier into the drug discovery process.
A further review3produced some disturbing evidence that “most published research findings are false”, which may be an excessively ominous conclusion. Nevertheless, the review exemplifies the extent to which the “reproducibility crisis” is prevalent across research activities beyond drug discovery and is not just limited assay precision. The overall conclusion of these previous reviews is that there is a scientific and commercial need to increase the predictive power of in vitroassays. However, this objective will only become effective when the statistical validity of cell-based assays, especially more complex ones, are sufficiently robust to mitigate the current reproducibility concerns.Only then can more robust data from such assays be used to drive improved pivotal decision-making in early drug discovery.
A brief history of cell-based assays in drug discovery
Drugs were originally tested in explanted human tissues in vitroon a routine basis from the 1950s onwards4. However, such cultures were often variable and not consistently available in the quantities and consistency required for drug discovery. Immortal cell lines derived from tumours were also developed for scientific research in the 1950s to circumvent the variability concerns found with fresh tissues and primary cultures5. Although these early cell lines continued to transform over time outside the human body, they did represent a significant advance over primary cultures of human cells, which were in limited supply, as it took significant effort and time to culture them reproducibly at scale. Since then, a plethora of other cell lines have been developed but they are typically grown on plastic (or coated plastic) as flattish monolayer cultures. Spheroids are typically derived from induced pluripotent or clonal cells that have been previously grown in monolayers or in suspension and then assembled into 3-dimensional (3D) compatible culture systems. Spheroids do enable cells to communicate with each other as well as their surroundings, providing some similarities to an in vivo3D environment6. So-called organoids, on the other hand, are typically derived from a one or a few stem cells, which can then self-organise into organ-like structures in culture, owing to their self-renewal and differentiation capacities7.Typically, organoids are grown from either pluripotent or adult stem cells, which are seeded and maintained in 3D for the entirety of their culture, often in hydrogels that mimic the growth conditions of the extracellular matrix in vivo8. Such studies with colorectal organoids, for example, were significantly enabled by the discovery that the human colonic crypt could renew itself ex vivo9,10(Figure 1). 3D cultures of human colonic crypts, the functional unit of the colonic epithelium, can be used to study stem cell biology, as well as tissue physiology and pharmacology. Crypts can be transformed into organoids during long-term culture in vitro10. Importantly, it was found that single adult stem cells could be stimulated by certain G-protein-coupled receptor ligands to build crypt-villus structures in vitro, which opened up the field to produce immortal colorectal organoid lines11. However, the effective screening of compound or antibody libraries requires large and consistent batches of monolayered cells, spheroids and organoids, which often require precise localised control of growth factors enabling them to grow consistently in tissue culture. Maintaining the correct spatial cellular resolution within complex heterogeneous organoids by the precise localised control of growth factor release is potentially an important solution for maintaining heterogeneity in culture.
Bioprocessing requirements for complex assays at scale
If grown on an industrial scale, complex cellular cultures can now be used more routinely for compound screening in drug discovery12but production quality control is critical when screening compound or antibody libraries, especially in comparison with more conventional cell cultures. Current production methods for the expansion of cells with conventional bioprocessing technologies can result in significant batch-to-batch variation and the precision control of proteins concentrations, especially for growth factors, is required to overcome the constraints of more conventional methods. The concept of continuous bioprocessing has been applied mainly to the manufacture of recombinant proteins. However, there are also applications for continuous flow methods to grow large numbers of human adherent cells for screening in drug discovery. For example, there are smart, multifunctional surface coatings capable of controlling the attachment, proliferation and subsequent self-detachment of human cells. Such systems allow the maintenance of cell cultures under improved steady-state growth conditions13. However, these biomanufacturing strategies still rely on continuously perfusing growth factors, other proteins or animal sera as part of the culture media, which is both wasteful and sub-optimal for some growth conditions.Maintaining relevant cellular spatial resolution is particularly important, especially for complex cultures such as organoids.
Sources of growth factor variability contributing to experimental noise
Recombinant growth factors are typically manufactured in E. colior mammalian cells. Following manufacture, they are purified from other proteins by preparative High-Performance Liquid Chromatography (HPLC). After elution from the column, the purified protein sometimes requires refolding. Typical levels of purity in commercial preparations are 98-99%. This process is complex and results in variable levels of growth factor bioactivity driving batch-to-batch inconsistency. The effects of this variability are amplified by the rapid degradation of many growth factors in cell culture. In sera free culture, these effects are pronounced. For example, FGF2 (Fibroblast Growth Factor) can be completely depleted in serum-free culture after 24 hours. Cell cultures require regular media changes, mainly because growth factors have become inactive. However, the effect of small changes in timing of these changes is difficult to control. More importantly, cells become stressed by an oscillating environment in culture, whereby they are successively satiated with and then starved of growth factors.
Use of novel platform technologies in bioprocessing
Novel plug-and-play sustained release platforms for growth factors are now beginning to reshape the way complex cellular assay systems, such as organoids, are grown in bioreactors.Many proteins, especially growth factors and cytokines, when used in culture media, degrade quickly, rapidly losing their bioactivity. Protein instability in aqueous solution hampers consistent bioprocessing outputs and significantly limits the physiological effects of proteins, such as growth factors, to maintain optimal growth conditions. Development of a technology that can continuously replenish active proteins from a localised microscopic store has been a significant challenge, as there is a need for greater precision and localised control over the growth of cells. Furthermore, approaches such as the encapsulation of proteins into hydrogels and other slow-dissolving matrices are emerging but their utility is still hampered by issues such as unpredictable compatibilities with the protein in question, loss of protein function due to harsh manufacturing processes and the burst release of the embedded protein. Such technical solutions are particularly important for maintaining cellular spatial resolution for complex cultures, such as organoids. The Polyhedrin Delivery System (or PODS) technology14, for example, has now made the objective of a micro-depot for the sustained release of proteins a practical reality. A single POD is shown diagrammatically in Figure 2. This technology is allowing cell culture systems to be reconfigured to provide greater certainty and consistency. PODS benefit from a streamlined production system which ensures high levels of batch-to-batch consistency. PODSgenerate protein co-crystals which provide sustained release, continuously replenishing proteins from millions of local microscopic stores, which can be placed next to (or at a distance from) cells, either randomly or in precise locations. In a similar manner to natural protein release from cells, these micro-depots release a steady stream of bioactive protein. This protein can either be limited to its local surroundings, dispersed more widely or made to form a concentration gradient to guide polarised cellular growth. These crystals do not suffer from burst release of their protein contents. At the core of this novel technology is a polyhedrin protein. The specific protein used encases cargo proteins within transparent, cubic, micro-sized crystals, which are much smaller than cells. Such crystals are shown in Figure 3. These protein crystals form admixtures of the polyhedrin and cargo proteins, which slowly degrade releasing the biologically active cargo protein over time. PODS are also robust and will withstand physical and chemical stress, so they can be handled easily and made to release their intact cargo protein over days, weeks or even months. Using such systems can readily create a steady-state growth factor environment in microscopic detail to enable the correct 3D self-assembly of organoids and other complex cultures.PODS can be attached and immobilised to, or within, 2-dimensional surfaces as well as 3D scaffolds, such as microcarriers or electrospun material, and positioned by hand or by using an instrument, such as a 3D cell printer.For example, PODS can also be embedded inside a microcarrier, such as aliginate/gelatin microspheres, with cell growing on the external surface of the microsphere, which is illustrated in Figure 4.
Use of growth factor micro-depots in cell-based assays
Complex cellular systems for use as research tools can now be reliably supplied as frozen organoid lines, for example in cryovials, ready for seeding into various screening assay formats for drug discovery6,7,8. However, for some other assays also based on complex cultures, the sustained release of growth factorsis important for organoid formation grown in situ. For example, mouse embryonic stem cells can be cultured with the addition of both PODS containing BDNF (Brain-Derived Neurotrophic Factor) and GDNF (Glial-Derived Neurotrophic Factor) crystals. During the 10-day period of growth factor treatment, only a single half-media change was needed and no new PODS were added.The protocol used is illustrated graphically in Figure 5. This new method was found to enable enhanced retinal ganglion cell yields compared with standard recombinant growth factors with only a single addition the growth-factor-containing PODS throughout a 10- to 14-day culture period, as is demonstrated in Figure 6. It was concluded by the investigators that the healthier phenotype of organoids observed was most likely due to the reduced handling disturbance and the consistent growth factor levels achieved by culturing them with the slow-release crystals.
Conclusion
By exploiting new emerging technologies, complex assays can now be grown with greater consistency to become sufficiently robust and reproducible to potentially improve the unacceptably high rate of compound attrition in downstream development. New production methods for expansion with bioprocessing can reduce batch-to-batch variation by the precision control of growth factor concentrations, which are overcoming the some of the constraints of more conventional bioprocessing methodologies. These new solutions are required to solve the reproducibility concerns with the production of human cellular systems and their practical use in drug discovery. Novel plug-and-play sustained release platforms for growth factors are now reshaping the way cell-based assay systems are produced at scale and then deployed more effectively in drug discovery.
Figure 1Images illustrating the development of patient-matched colonic organoids and tumouroids generated from the non-involved mucosa and tumour of a colorectal cancer patient, respectively. These images are provided by the courtesy of Dr Mark Williams from the School of Biological Sciences, University of East Anglia, Norwich Research Park, UK, which are grown from the human colonic organoid biobank for the discovery of novel drugs.
Figure 2Legend:A PODS nanocrystal shown as a schematic, to illustrate the maintenance of steady-state bioavailability of cargo proteins over extended periods for applications in 2D and 3D cell culture.
Figure 3 Legend: PODSproteins as viewed by a scanning electron microscope and demonstrating their regular cuboidal structure. The proteins self-assemble around a specified cargo protein, such as a growth factor, within the expression production system. The scale bar is 10µm.
Figure 4Legend: The image shows PODS containing fluorescent green protein glowing inside an aliginate/gelatin microsphere with C2C12 cells, which are an immortalised mouse myoblast line, growing on the surface of the microsphere.
Figure 5Legend:Diagramatic illustration of the assay protocol described in the text.
Figure 6Legend: Quantification of total differentiated retinal ganglion cells with the neurochemical marker RBPMS (RNA-Binding Protein with Multiple Splicing). The cell population data in the y-axis are plotted as percentage of total cell number (n=4).These data are provided by the courtesy of Dr Julia Oswald and Dr Petr Baranov from the Department of Ophthalmology, Harvard Medical School, Schepens Eye Research Institute, Massachusetts Eye and Ear, Boston, USA.
Main image credit: ThisIsEngineering RAEng
Volume 22, Issue 1 – Winter 2020/21
About the authors
Dr J. Mark Treherne has been involved in the biopharmaceutical industry for over 25 years and previously led a drug discovery research group at Pfizer’s research facility in the UK, including using stem-cell derived lines for screening compounds. In 1997, he co-founded Cambridge Drug Discovery, who automated cell-based assays for high-throughput screens. He has recently focused on pluripotent stem cells and organoids grown from adult stem cells, as well as research tools enabling cell culture.
Dr Christian Pernstich is a protein biochemist, who has worked on protein design, purification and characterisation to developing biochemical and biophysical assays. Dr Pernstich joined Cell Guidance Systems in 2016 to co-ordinate the development of the PODS technology platform, leading a team of scientists, which has launched more than 100 PODS products onto market. He has a PhD in protein biochemistry from the University of Bristol, which was awarded in 2010.
Dr Michael Jones founded Cell Guidance Systems in 2010. His research interests have focused on lung cancer, genomics and regenerative medicine, publishing over 30 papers. In his earlier career, Dr Jones cloned and characterised genes that have provided therapeutic targets and for lung cancer and refined its classification. He worked in Japan between 1990-1992 and 1995-2010, returning to the UK in 2010 to establish Cell Guidance Systems.
References:
1Scannell et al. (2012) Nature Reviews Drug Discovery, 11: 191–200.
2Scannell & Bosley (2016) PLoS ONE 11(2): 1-21.
3Ioannidis (2005) PLoS ONE 2(8): 696-701.
4Wright et al. (1957) New England Journal of Medicine, 257: 1207-11.
5Scherer et al. (1953) Journal of Experimental Medicine, 97: 695–710.
6Treherne et al. (2018) Drug Discovery World, Winter 2018/19 Issue: 8-13.
7Treherne (2020) Cancer Science & Research, 3(1): 1-4.
8Badder et al. (2020) PLoS ONE, 15(8):1-20.
9Reynolds et al. (2007) Journal of Physiology 582: 507-524.
10Reynolds et al. (2014) Gut 63(4): 610-21.
11 Sato et al.(2009) Nature, 459(7244): 262–5
12Stewart-Miller & George (2020) Drug Discovery World, 21(3): 23-29.
13 Miotto et al.(2017) Applied Materials & Interfaces, 9: 41131−41142.
14 Matsuzaki et al. (2019) Biomolecules 9(10): 510
