by Angelo DePalma

Life scientists seeking the "big picture" are recognizing the value of cell imaging techniques. Automated cell imaging arose from the need for more sophisticated drug discovery tools, principally at pharmaceutical and biotech companies. According to John Anson, Ph.D., who heads Lead Discovery Development at GE Healthcare (Cardiff, Wales), pharmaceutical and biotechnology companies are still imaging's biggest consumers, but academic labs are getting interested as well.

Imaging provides multiplicity, multiplexing, and a higher-order view than individual assays. Automation adds the component of speed and high throughput. Together, these benefits provide the basis for quantitation and statistical validation. "Greater numbers equals more replicates equals better statistics," says Dr. Anson.

click the image to enlarge GE Healthcare's GFP (Green Fluorescent Protein) can be used in drug discovery-related applications to track the movement of proteins within living cells.

Quantifying image data is not as much a function of the imaging instrument as of data management, particularly the algorithm that compares one cellular state to another and deconstructs the CCD image into numerical data. "Before-after analysis is critical," adds Dr. Anson. "Image analysis software was the 'forgotten' piece of the puzzle linking instrument and biology."

Where biochemical assays require a priori knowledge of both target and the most likely answer, cell imaging allows scientists to investigate and pose questions whose answers are completely unknown.

Cell imaging would not have progressed beyond simple photomicroscopy without automated microscopy, robotic workstations for sample prep, enabling biology such as fluorescent reporters, and software to make sense of it all. "The combination of green fluorescent protein and automation has been the game-changer for cell imaging," says Dr. Anson. GE Healthcare's cell imaging systems reflect their integrated approach to instrumentation. The IN Cell 3000 high-end, high-throughput, rapid imaging device is a mainstay of high-throughput pharmaceutical work. IN Cell 1000 is aimed at assay development and academic laboratories.

Top-level view
Cell imaging provides top-level views of complex cellular processes which were previously studied piecemeal through biochemical assays. For example, transcription is traditionally approached one pathway, transformation, or molecule at a time. "It's a reductionist approach that works quite well," says Michael A. Mancini, PhD, Associate Professor at the Baylor College of Medicine (Houston, TX). But now, armed with rapid, digitized imaging methods, transcription may be observed as a unified "event" rather than a chain of related processes.

Dr. Mancini has developed a quantitative imaging approach that answers multiple questions about transcription through a single experiment, from one set of images. The one-step approach avoids the effort involved in obtaining information on transcription, DNA binding, hormone switching, and other events separately. The technique was developed jointly by Dr. Mancini and Q3DM, an instrument company purchased by Beckman Coulter in 2003. Q3DM's contribution was a high-throughput, automated microscope imaging system which Beckman now manufactures and markets. The image files from microtiter plates are huge -- tens of gigabytes. Multiple investigations are possible with each image or group of images by addressing them through different analytic criteria. "You don't need to do the experiment over," Dr. Mancini notes.

Dr. Mancini believes imaging methods like his could lead to unifying, aggregate perspectives on important biological phenomena. "We're trying to get a much bigger view of what's happening by looking at multiple events at the same time, in one cell," he told Bioscience Technology. "A couple of years ago the idea of setting up 384 conditions for an experiment was unthinkable. Now it's possible."

Devil in the details
Cell imaging's major hurdle thus far has been not hardware but software for analyzing huge image files, says Jeffrey H. Price, MD, PhD CEO, a founder of Q3DM and now CEO of Vala Sciences (La Jolla, CA). Vala develops platform-independent software that analyzes and processes images from any microscope. The company's first product is software that performs membrane measurement. The company also develops reagent kits, the first of which helps screen compounds for inhibition or activation of protein kinase C-a.

click the image to enlarge Figure 1. Examples of the patterns displayed by different proteins within cells. These images were collected using a confocal fluorescence microscope, which can collect three-dimensional images of cells. The colors in the images represent staining for DNA (red), total protein (blue), and a specific protein (green). The name of the protein or organelle that contains this specific protein is shown above each panel. A fully automated computer system developed by Dr. Murphy's group (see text) can recognize the differences between all of these patterns with greater than 98% accuracy. Even more importantly, it can discriminate the patterns of similar proteins, such as those for the proteins giantin and gpp130, that human observers cannot distinguish.

Casey Laris, Product Manager for Cell Imaging and Analysis at Beckman Coulter (Fullerton, CA), worked for Dr. Price at Q3DM and now manages products Beckman acquired from his former company. Among them are Vi-Cell, an affordable image-based cytometer for viability measurements, the higher-end FC-500, and the IC 100 (late of Q3DM), an imaging cytometer. In contrast to flow cytometers, which count cells mechanically, imaging cytometers pick cells out within an image field electronically, virtually.

Mr. Laris sees imaging's automation and multiplexing capabilities transforming cell biology into a quantitative science. "We can see the same subcellular detail as before but now we can attach numbers to it," he says. "In the past biologists could sit there and watch a protein migrate from the cytoplasm to the nucleus, but they couldn't walk away and have an instrument measure this phenomenon 200,000 times."

Targeted therapies
Pathology, one of the earliest and most recognizable cell imaging applications, made great strides with the advent of inexpensive, powerful personal computers. As computing technology advanced, pathology laboratories found they could digitize images and analyze them electronically as well, which opened the door to mining images for details inaccessible to the human eye through unassisted light microscopy.

Clarient (formerly ChromaVision, San Juan Capistrano, CA) claims its ACIS (automated cellular imaging system) was the first instrument to provide non-academic pathologists with immunohistochemistry-based cell imaging. ACIS, which consists of a digital microscope and proprietary software, is used for tissue scoring, rare event detection, object detection and counting, integrated optical density analysis, and tissue array analysis. Most important, the instrument standardizes interpretation of common image-based cell assays. Clarient is pushing ACIS for drug discovery applications, particularly to quantify cell proliferation.

Today's push towards targeted therapies has opened a new front for cell imaging in pathology. Targeted or personalized therapies, which treat a particular genotype within a broad disease category (e.g. "prostate cancer"), are typically prescribed with a companion confirmatory assay. "If you have a targeted therapy you need an assay," says Ken Bloom, MD, medical director at Clarient. Nearly 100 targeted therapies are in Phase III testing and close to 350 are in Phase II. Genetic analysis is one way to identify patients who might benefit from targeted treatments, but cellular imaging could be faster, cheaper, and more accurate.

According to Dr. Bloom, targeted therapies will benefit from image-based techniques that measure protein location, quantity, and aggregation. Imaging has the potential to score multiple targets in multiple cellular compartments, as opposed to single-target solution-based tests that support today's approved targeted therapies.

Last year Clarient entered a clinical research collaboration with UCLA to investigate ACIS for counting circulating endothelial cells in breast cancer patients receiving combined treatment with Herceptin (trastuzumab) and Avastin (bevacizumab), both targeted therapies.

Come together
Researchers at Carnegie Mellon University (Pittsburgh, PA) have developed an automated tool that locates and groups fluorescently-labeled proteins within cells and organelles. According to lead researcher Robert F. Murphy, PhD, Professor of Biological Sciences, this new capability is one tool that will enable a new field, "location proteomics," which describes and relates protein locations to various cellular events, e.g. apoptosis, tumorigenesis, or response to a drug or other stimulus.

Location proteomics first establishes protein locations under normal conditions, then compares these protein maps to diseased or perturbed states. Today's imaging techniques allow tracking many proteins simultaneously, in one experiment.

"It's critical, when trying to understand how cells work, to know where the players are," commented Dr. Murphy.

Fanqing Chen (pictured) and Daniele Gerion have harnessed the powers of nanotechnology to image the interior of cell nuclei.

Many important cellular functions are accompanied by protein translocation. For example, transcription factors migrate from the cytoplasm to the nucleus during transcription. Another group of migratory proteins, glucose transporters, exist inside vesicles under low-glucose conditions but move to the vesicle surface in response to cellular glucose uptake. The extent and speed of this migration could signal how a diseased cell responds to glucose in diabetes, or the effectiveness of a novel diabetes drug in development.

Prof. Murphy's approach employs sets of what he terms subcellular location features (SLFs), which describe protein locations within a cell image. SLFs serve as a type of filter, which may be applied by computer multiple times or in combinations after the complex images are acquired. Dr. Murphy likens his method to applying criteria of color and surface smoothness to separate different types of fruit on a conveyor. SLFs use high-power confocal microscopes to detect and measure location-related properties such as shape, texture, edge qualities and contrast against background.

Although reporter molecules are the approach of choice for cell imaging, new ways to light up cells are on the way. Scientists at Lawrence Berkeley National Laboratory (LBNL; Berkeley, CA) are using nanoprobes to track and image long-term cellular phenomena such as DNA repair. Fanqing Chen and Daniele Gerion use quantum dots (qdots) nano-scale crystalline semiconductors, consisting of no more than a few thousand atoms, which fluoresce at different wavelengths when illuminated by laser light. The stable, nontoxic particles persist within the nucleus much longer than conventional fluorescent labels, without harming cells or fading.

Nanometer-wide qdots are surface-silanized for water-solubility, and possess functional chemistries (thiols, amines, carboxylate, and aldehyde) for conjugation with almost any biological molecule or entity. Qdots easily light up macro events, such as pathogens, within cells, as well as molecular events like single-nucleotide polymorphisms. Two years ago French researchers used qdots to track migration of glycine, a neurotransmitter, from one neuron to another by checking qdot fluorescence every 100 milliseconds over 20 minutes. LBL scientists have already used qdots to track nuclear events for up to a week and hope to get inside other, smaller organelles.