Microscopy, one of biology's most mature technologies, is enjoying a resurgence thanks to fluorescence tagging, live cell experimentation, image acquisition and processing, and software that ties everything together. Although high-end imaging systems can be quite expensive, the cost of high-throughput microscopy has fallen, making the technique accessible for most biologists. At the same time microscopes and image analysis platforms have followed the track of other instruments in improved ease of use. Or as Stan Schwartz of Nikon Instruments, Inc. (Melville, NY), puts it, "Our customers want to be experts in science, not microscopy. That's our job."
Examining total internal reflection of fox lung fibroblast cells labeled with vinculin fused to mKusabira Orange fluorescent protein using Nikon Inc.’s Perfect Focus
system at the National High Magnetic Field Laboratory at Florida State University.

Automation, new light sources, and software for acquiring and analyzing images are at the top of the list of trends in microscopy. Automation covers a lot of ground, and is valued differently depending on the type of laboratory (academic, pathology, clinical lab, industrial service group). Automation improvements range from sample preparation to image acquisition, auto-focusing, and such seemingly straightforward tasks as moving slides across the stage to capture images more rapidly. Novel light source technology, including bright field (transmitted white light), fluorescence excitation sources, and LED (light emitting diode) lighting improve image resolution through intense, controlled illumination.

Still bigger than a shoebox
Unlike mass spectrometers, which at one time occupied large rooms but now fit inside a shoe box, high-end microscopy and imaging are still expensive propositions that demand a high level of technical expertise. Universities and medical centers therefore establish microscopy service centers or shared resources. The H. Lee Moffit Cancer Center (Tampa, FL) is home to one such facility, serving physicians and basic researchers. The Moffit shared resource holds the range of optical systems, from simple light microscopes with no CCD (charge-coupled device) detector, to advanced slide-scanning systems that handle both wide field and fluorescence microscopes.

Atlas slide:
Murine tissues
presented in a
tissue microarray, c/o Sanger Institute.
According to Ed Seijo, Manager of Shared Resources at Moffit, fluorescence microscopy on wide-field scopes is the center's most popular service, but confocal imaging is rapidly becoming the technique of choice for high-end applications. "Journals are increasingly asking for confocal images," he says. Confocal microscopy provides high-resolution, blur-free images and 3-D reconstructions. The technique is useful for co-location experiments, for example, the imaging of two proteins that are individually labeled and located in different regions of a cell. "Wide-field images, where the object under investigation is illuminated from above or below, tend to be fuzzier," Seijo says. A confocal microscopes' 3D capabilities can recreate a cell's outer features in full dimensional splendor, a useful capability for work on live, whole cells.

More than 95% of Seijo's "customers" are biologists, with an occasional materials scientist or physicist visiting his facility. "The big crossover is with nanoparticles and quantum dots," he told Bioscience Technology.

As expected, the cost differential for confocal microscopy can be significant. A wide-field fluorescence microscope may be had for about $50,000. Expect to pay about double that figure for fully automated systems. Confocal instruments fetch $250,000 for basic models, but bells and whistles can increase the price substantially. The Moffit center recently received a grant to purchase a new Leica confocal model for $515,000 (Leica Microsystems, Inc., Bannockburn, IL). That price tag easily doubles for the next level of sophistication, the two-photon confocal microscope, which penetrates more deeply into tissue.

Atlas cores: Transmission
microscope antibody profiles from animal tissue, c/o Sanger Institute. Click here to enlarge.
Live cell imaging has come on strong at imaging facilities like Moffit's. Both basic and applied researchers now prefer cells to biochemical assays because the former possess greater biological context. Thanks to improvements in instrumentation, hardware, and software, these experiments are now within reach of most biologists. Since cells dislike visible light, fluorescence tagging has been an enabling technology for live cell work. Coupled with high-resolution confocal methods, researchers can now visualize receptors on cell surfaces around the entire circumference of a cell.

Live cell assays present researchers with the task of keeping cells alive during successive cell cycles - imagine cell culture and all its refinements, at micro scale. "Researchers want cells to 'think' they're in their normal environment," says Stan Schwartz of Nikon. At the same time, researchers need cells to remain in focus as they expand, shrink, and multiply. Nikon's PE2000 Perfect Focus Inverted Microscope automatically adjusts the focus to correct even minute changes in object distance.

The Microscopy and Image Analysis Facility in the Department of Neurobiology and Anatomy at West Virginia University (Morgantown, WV) grew from a shared instrumentation grant 10 years ago to the multi-system resource it is today. The Facility's systems include five imaging systems serving biologists, agricultural scientists, and researchers from outside universities. Systems include a top-line Zeiss LSM 510 confocal instrument (Carl Zeiss Microimaging, Inc., Thornwood, NY) and an MBF Bioscience (Williston, VT) bright field system. Close to 80% of projects handled by the Facility, says director Jeffrey Altemus, involve fluorescence microscopy.

According to Altemus, there is considerable lag time between acquisition and full utilization for sophisticated microscopes. The Facility must conduct outreach of a sort to inform researchers of new capabilities, but even then investigators are not keen on adapting their research to new equipment. He cites a microdissector that prepares cell-sized samples for microscopic analysis, received three months ago, that is only now drawing attention from UVW biologists.

Climbing Mount Everest
Most of the developments in microscopy and imaging are driven by the need for higher throughput. The Atlas of Protein Expression Project, underway at the Welcome Trust Sanger Institute (Cambridge, UK), is a high-throughput protein expression project that relies on automated image capture of tissue sections. The goal is to map and catalog protein expression and distribution at the cellular level.

Researcher's capture and analyze image data automatically using Applied Imaging Corp.'s (San Jose, CA) Ariol automated imaging and image analysis system. The system captures images and analyzes data from chromogenic and fluorescence-tagged tissue microarrays, at a throughput of up to half a million core images per month. The slides are then annotated within an Oracle database (a process that urgently needs to be automated, and a task that will probably require international cooperation).

The goal of Atlas is to produce a comprehensive catalog of protein expression in mouse and human tissue. "It's a bit like climbing Everest," jokes senior project leader Tony Warford, Ph.D. "If you're going to call something an 'Atlas', it had better include everything." Automation and tissue microarrays make this work possible, as analogous high-throughput techniques enabled the Human Genome Project. Arrays allow sample miniaturization, which in turn permits running ambitious experiments on a reasonable number of slides; automation is therefore essential for capturing and analyzing data.

Ten years ago, this type of work would have taken much more time and many more slides. A project involving 20 tissues, for example, might involve 80 slides, each of which would need to be immunostained, washed, and analyzed individually under a microscope. After capturing images, researchers would need to analyze each micrograph individually. "Today we can do the same panel on one slide," Warford says, "and feed multiple slides to the instrument for overnight analysis."

The onset of data overload, common with all early-stage automation technologies, has been the impetus for automating the annotation of slides. Warford believes the technology for automated annotation exists, but will need to be developed through an international consortium of academia, government, and corporations.

Industrial collaboration
In 2003, Sanger entered a collaboration with Applied Imaging to develop the Ariol platform for tissue microarrays. The relationship has since expanded to include other aspects of imaging automation.

According to Iqbal Habib, Technical Product Manager, Pathology, at Applied Imaging, image system automation and software integration have improved workflow by allowing data to move between instruments, systems, and databases. Part of this equation is simplifying image acquisition and analysis through application-specific algorithms tailored to find nuclei, cytoplasm, or membranes on microscope slides. End users, Habib says, should not have to learn how to construct these scripts, or know the intimate technical details of image analysis, to run complex experiments.

Applied Imaging's Ariol platform, mentioned in the work at Sanger, automatically scans and analyzes tissue sections on slides in both bright field and fluorescence modes, simultaneously or independently. Ariol intimately integrates scanning with analysis, analyzing samples as it feeds slides into the field. "By the time the user looks at the screen the analysis is complete," says Habib. Applied imaging refers to itself a "medical device" company since some methods used on the Ariol are FDA approved, through the 510k route, as diagnostics. Users may also create and store proprietary applications for later execution.

On the dot
The overlap between materials scientists and engineers of various stripes with biology represent a growing user base for microscopy and image analysis. This trend can only grow as nanomaterials increasingly cross into biology and medicine through dendrimers, liposomes, fullerenes, carbon nanotubes, and quantum dots. Prof. Tania Vu at Oregon Health Sciences University (Portland) demonstrates how imaging facilitates the marriage of nanophase materials and biology. Quantum dots (QDs) are nanoscale (smaller than 1 μm) semiconductor particles, on approximately the same size domain as macromolecules, which fluoresce when illuminated with an interrogating light source. The excitation and emission colors and intensities depend on the material and the QDs dimensions, which gives rise to nearly limitless possibilities for imaging proteins and subcellular structures.

Vu attached two biomolecules to the surface of cadmium selenide QDs: an antibody to a surface receptor to get the quantum dot into nerve cells, and nerve growth factor (NGF) for shuttling the particle through cells. In earlier studies, cells failed to internalize quantum dots larger than 1 μm. Here, surface receptors recognized the QD-antibody complex and readily took it into the cell as part of normal receptor recycling. Once inside, the cell recognized the NGF on the QD surface and began shuttling the particle as it would unbound NGF.

Using a conventional fluorescence microscope, it is possible to image several different types of QDs, emitting at different colors, simultaneously and in the same cells. Thus, one could label QDs with various growth factors or other chemoattractant molecules and watch their migration into various cellular components under normal and diseased or treated conditions.