October 1, 2008 (Vol. 28, No. 17)
The Field Is Growing Fast and the Impact on Development and in the Clinic Is Significant
The demand for understanding cells on a microscopic level continues to fuel development for new imaging platforms and analysis software. This is being further buoyed by high-content and high-throughput protocols. Optical imaging is an emerging field that holds promise to help make drug discovery and development faster and more efficient.
Speakers at the Informa Global Imaging Summit, which will be held in December in Cologne, Germany, will illustrate how technologies in this field are rapidly evolving and making their way into the clinic to assess disease progression and monitor drug efficacy.
For example, scientists from PerkinElmer plan to address some of the major challenges of high-content screening (HCS) for cellular imaging with new image analysis software and platforms.
“Two of the biggest bottlenecks include setting up the high-content screening assays and then dealing with vast amounts of data,” explains Gabriele Gradl, Ph.D., global product leader, high content screening at PE Cellular Technology Germany. In addition, as HCS becomes more accepted, researchers are pushing the boundaries of its possibilities.
“People are trying to use their instruments to their fullest potential. So, the ability to do data analysis in two and three dimensions along with high-content screening, is something that has driven us to set up a cellular imaging business,” states Paul Orange, Ph.D., strategic development leader, cellular imaging and analysis at Improvision.
Opera™ is a confocal microplate imaging reader that provides automated simultaneous high-speed and high-resolution screening. Assay applications include: whole cell fluorescence, cell signaling, gene expression, membrane receptor, translocation, and morphology. The accompanying image analysis software, Acapella™, provides high-speed, two-dimensional, and high-content image analysis.
“This software enables you to look at different aspects of data in different ways,” says Dr. Orange. It comes with a set of ready-made application solutions or “scripts,” and is also flexible for new algorithm development via an open architecture.
Columbus™ Gallery is a database system that allows HCS multichannel images to be stored and accessed by multiple users. It can import, export, and manage images from a wide variety of sources. A system to enable reanalysis of data images, called Columbus Conductor, will be available sometime this fall.
Another platform, Volocity Visualisation software, helps during assay development and provides 3-D imaging solutions. Data is acquired from high-speed image capture and archived on a hard drive. It provides parallel processing and video streaming, and visualizes and analyzes structure and function of biological samples. Algorithms eliminate noise and blur.
Model Tumor Cell Invasion
Researchers at AstraZeneca have developed more predictive in vitro, in vivo, and in silico models to improve efficacy of potential drugs at an early stage.
“We do advanced image analysis on assays to obtain information on how drugs are affecting the cells’ phenotype. Some of the challenges with 2-D assays are throughput and large amounts of data generated from image analysis. When we try to translate that into a 3-D cell system, we have a whole host of new challenges,” explains Neil Carragher, Ph.D., associate principal scientist at AstraZeneca.
“This includes developing 3-D gels of the cells,” Dr. Carragher says, “which has to be done manually since most of the automated equipment can’t handle the 3-D gels; making it especially challenging for large target validation studies or large compound screens.”
Dr. Carragher’s group has also developed a number of image analysis algorithms to capture a drug’s physiological effect on the cell, rather than looking at a single target enzyme, called phenotype profiling. In order to look at different depths, his group has revised its image analysis approaches to account for an extra dimension—the Z focal plane.
The Z-series is a collection of images taken on a sample at different focal planes to construct a Z-stack, which provides the necessary spatial resolution to fully examine 3-D samples. To determine whether their 3-D assays are more predictive than 2-D samples, Dr. Carragher says, they are conducting in vivo imaging studies with real-time image analysis and comparing those results to 3-D data obtained from in vitro assays.
“We’ve seen several cases where our assay is not replicating what we see in vivo, so we’ve begun to re-engineer those assays,” Dr. Carragher notes. “The final endpoint is a multiparametric signature; we’re almost getting a fingerprint for every compound.” This information is used to identify how the compound is affecting the phenotype of the cells.
Dr. Carragher says that one of the active areas in cellular imaging is the development of new optical imaging instruments for in vivo assays. His group uses multiphoton confocal microscopy for deep tissue imaging in animals to monitor fluorescently labeled proteins or cells.
Predicting Preclinical Toxicity
Toxicology assays are becoming key to the drug discovery process. Researchers at University College Dublin are assessing compound cytotoxicity to predict their potential for human toxicity. This involves analyzing the structure of the compound and its target, what implications the target has for safety, and potential off target pharmacology.
There are some limitations to this process, says Peter O’Brien, Ph.D., DVM, lecturer in clinical pathology at the college. “It’s fairly new for toxicologists to be working in drug discovery. Historically, cytotoxicity testing has not been seen as being very effective in predicting human toxicity potential. There wasn’t a lot of technology available and the cell-based models were not effective. In addition, current technology for high content analysis for cytotoxicity is expensive and sophisticated. We’ve been defining the methodology for several years.” However, he adds, it’s slowly catching on.
“High-content analysis has reopened the subject and has potential for assisting in safety assessment,” Dr. O’Brien says. Pharmaceutical companies are conducting toxicology assessments in the discovery phase—Pfizer recently announced it is using HC analysis in safety screening in discovery. Dr. O’Brien says this may have the potential to lower the current drug attrition rate of about 33% failures due to safety, down to about 10%. This will require toxicological expertise early in the discovery phase.
“Safety pharmacology studies are now typically done in discovery phase as well as off target pharmacology studies,” says Dr. O’Brien. He adds that safety biomarkers can be used to monitor toxicity before it causes clinical symptoms, enabling dose changes or a change in therapeutic strategy.
The image intensity in most fluorescent microscopes is proportional to the concentration of the fluorophore being used. This provides localization information, but the resolution is limited by diffraction of about a micron, depending on the wavelength.
“What we’re trying to do is get more information from that sample,” explains Paul French, Ph.D., head of photonics group, Imperial College, London. “We are applying multidimensional fluorescence imaging and analysis to tissue autofluorescence.” His group is accomplishing this by developing and enhancing instruments, such as combining multibeam, multiphoton microscopes with fluorescent lifetime imaging.
A portable wide-field fluorescent lifetime imaging (FLIM) scanner based on time-gated imaging is being developed to provide faster in vivo diagnostic screening. This has potential to monitor disease and treatment response by using the autofluorescence of tissues. By evaluating the fluorescent lifetime of the autofluorescence, it has potential to show changes in cancerous vs. normal tissues.
FLIM can be quite slow because it must collect a lot of photons to make an accurate measurement and the widely used technique of time-correlated single-photon counting imposes a maximum photon detection rate. Wide field imaging can be faster because it collects photons for all the pixels in parallel.
“We are combining wide field time-gated imaging with Nipkow spinning disc microscopy to do high-speed FLIM to accelerate research based on FLIM and FLIM-FRET,” states Dr. French.
His group is working toward goals to develop clinical diagnostic instruments exploiting autofluorescence, to apply techniques like FLIM and FRET to automated imaging in a 96-well format, and to move to super resolution using a technique called stimulated emission depletion and combined with FLIM for protein-protein interaction analysis on a small scale to look at microclusters of signaling.
Optical imaging is being applied in the cancer drug development process to provide important information. Various in vivo imaging techniques are used in pre-clinical models to monitor disease progress and therapeutic efficacy, and to demonstrate target-oriented drug effects.
Werner Scheuer, Ph.D., research leader, preclinical imaging, Roche Diagnostics, is utilizing near infrared optical imaging because it has several advantages over that of other methods such as: no radioactive isotopes required, acquisition time between one and five seconds, and good penetration of tissue—roughly one centimeter. In addition, it is much less expensive, and training technical personnel can be accomplished in a reasonable time period.
His group is using optical imaging for pharmacodynamic (Pd) studies and pharmacokinetic (Pk) studies. Quenched probes, which are activated by tumor-associated proteases and allow monitoring of cell viability, are used for Pd studies. Therapeutic antibodies labeled with organic fluorophores like Cy5 are used for Pk studies. Monitoring signal intensity of the tumor area provides an indication of how fast the labeled antibody enters the tumor tissue and biodistribution of labeled antibodies to organs can be followed up. Verification of tumor-associated surface molecule expressed in different human xenografts in vivo is possible by using fluorescence labeled monoclonal antibodies. This approach allows the selection of an appropriate xenograft for further evaluation of an antitumor efficacy of target-orientated drugs.
One of the main limitations of optical imaging is exact quantification. “If you want to know exact quantification of tumor cell destruction, that’s hard to perform. But, since we’ve been using it for the past three years, it can accelerate the drug development process.” Dr. Scheuer says he is applying optical imaging in attempts to monitor apoptosis in vivo. “We are looking for other techniques or other proteins to identify apoptotic cells over time in tumor carrying mice.”
Studies are under way to evaluate labeled Annexin V for monitoring induction of apoptosis in vivo, allowing earlier estimates of drug efficacy. Also, quantum dots are suitable to address specific questions due to their photophysical properties like stronger signals and no bleaching.
“This field is growing fast, and the impact of optical imaging is becoming greater in the drug development process and in the clinic,” Dr. Scheuer says.