DDW Spring 2021

Innovation & Technology | imaging

machine learning has been applied to image processing for deep learning, also called neural networks, of patterns within images to speed up and extract complicated information.” How imaging is furthering drug discovery and development Imaging has advanced to create efficiencies and benefits to workflow in drug discovery and development. “A raft of technologies introduced in recent years has reduced drug candidate attrition, accelerated development workflow, and helped identify more effective therapies,” says Amirmansour. “A particular area of focus has been the adoption of more physiologically relevant cell models used during early selection, based on the assertion that those more closely resembling human physiology generate a truer indication of efficacy ahead of clinical trials. This has led to the development of 3D cell culture methods, often using human primary or differentiated stem cells, aggregated into tissue or organ-like structures associated with the disease indication.” According to Amirmansour, assays associated with these 3D models tend to quantify specific phenotypes of the cell aggregate,

which are often best analysed by quantitative automated microscopy. In fact, he adds, phenotypic screening associated with automated imaging has supplanted industrialised, target- based, small molecule screening to a large degree. “As automated imaging/microscopy and image analytics have evolved, cellular imaging has become the preferred method for other areas in the drug discovery process (both upstream and downstream of screening), including new target identification, compound profiling, toxicology, and the study of the mechanism of action of test molecules.” Hawryluk adds that the use of imaging throughout the drug discovery workflow allows for efficient and quantitative measurements that complement various genetic and biochemical data and enhance mechanistic studies by providing location information that cannot be detected using non-imaging methods. ѦSpecifically, when using imaging within the cell culture process, this helps ensure the quality control of your cultures while the 3D information enables a more detailed analysis. Doing the detailed work earlier in the drug discovery process can help reduce the risk of failure and, therefore, costs and time,” she explains.

period, requiring technology to ensure that specimens are non- perturbed during imaging - even if the experiments continue for days.” He notes that image analysis algorithms are continually being refined to ensure that microscopy is not just a useful tool for structural analysis, but for quantifying cellular function too. “High content screening, developed over twenty years ago, enabled the use of quantitative microscopy in drug discovery ў a field where throughput demands brought the microplate format to the microscope. Nevertheless, these systems remain cost prohibitive to the average laboratory. More recently, the emergence of benchtop automated imagers with similar capability, yet costing a fraction of the price of a conventional HCS platform, have democratised quantitative microscopy,” he says. Dr. Joanna Hawryluk is Associate Product Manager for Research Imaging and Microscopy at Olympus. She explains that the imaging market – including Olympus – is constantly looking at ways to develop new technologies that will contribute to improving the entire imaging workflow from data acquisition to analysis. “On the data acquisition side, the current focus is on improving image quality and acquisition speed to increase throughput while maintaining the sample’s integrity. In other words, improving the signal-to-noise ratio, the measure of image quality based on signal intensity and background noise,” she says. “For example, in traditional confocal technology, pinholes are used to obtain sharp, high-resolution images, but the inherent loss of light from the pinhole(s) makes it challenging to capture images with a high signal-to-noise ratio. It also exposes your sample to large amounts of out-of-focus light, increasing the likelihood of photobleaching. This makes

it difficult to image thick samples such as spheroids (3D cultures), as they require a higher excitation light to produce a fluorescent signal strong enough to be detected.” To improve light detection, says Hawryluk, several spectral- based detection methods have been developed in which a grating or prism is used to create a spectrum of emitted light. “While spectral-based detection methods enable fine tuning of your emission spectrum, it comes at the cost of sensitivity. Spectral-based detection provides greater flexibility, but more traditional filter-based detection is typically more efficient. The current focus is to improve the sensitivity of spectral detection while maintaining imaging versatility.” Olympus’ FLUOVIEW FV3000 confocal microscope with TruSpectral detection technology, uses a volume phase hologram (VPH). “VPH gratings allow light to pass through a grating to diffract it into its spectral components to allow for spectral unmixing, without compromising on sensitivity,” adds Hawryluk. “Improved image quality leads to more robust and reliable data. But these data need to be processed quickly and intuitively,” she says. “Recently,

Figure 1: Evolution of drug discovery process from target-based to rich phenotypic assays. Image courtesy of Biotek Instruments

37

Drug Discovery World | Spring 2021

Made with FlippingBook HTML5