Picture This

Leeuwenhoek's microscope launched the optics race, and the drive to image the ever smaller, in greater detail, and with greater meaning. Today, one of the newest moves in drug discovery is to actually watch a cell go about its daily routine.

"The explosion of molecular imaging as a collection of biological tools, strategies, and hardware/software and the need for such an undertaking in biomedicine coalesced at this same moment in time," says David R. Piwnica-Worms, MD, PhD, director of Molecular Imaging Center, Washington University, St. Louis, Mo. CCD cameras became sensitive enough to detect extremely low photon flux; fluorophores, reporter constructs, and transgenic organisms became widely available.

In the pre-molecular imaging era, the instrument often dictated the agenda. "People tended to align themselves with a technology and then go around and try to answer every possible question with that technology," Piwnica-Worms explains. "If you have a hammer, everything starts to look like a nail. Occasionally it's a good fit, but often it's not." Now the imaging field is undergoing a restructuring of thought; the biological question comes first and its attributes dictate the approach.

Consider apoptosis. "We were interested in a non-invasive, dynamic, real-time read-out of apoptosis," says Piwnica-Worms. "So we chose the well-characterized protease, caspase-3." The strategy was to develop a molecule that could cross the cell membrane; avail itself of caspase-3 cleavage; and report back that cleavage had occurred. First they tried a radiolabel—but without success. The nonspecific binding was higher than anticipated. And since radiolabels are always 'on', background noise was unacceptably high. That's when they turned to light. "One of the advantages of optical imaging is you can design quenching strategies," he says. The final design used a self-quenching reporter peptide that when intact remained unseen, but when cleaved by caspase-3, sent up its signature flare. (For details see: Biochemistry. 2007 Apr 3; 46(13):4055-65)

Taken by STORM
As informative as optical imaging can be, light has its limitations. "It has only touched a relatively small subset of biology problems so far," says Xiaowei Zhuang, PhD, professor of chemistry and physics, Harvard University, Cambridge, Mass. And the disadvantage is resolution. "These days, people are more interested in mechanisms—at the gene level, the molecular level—and these things are typically much smaller than the resolution of light microscopy."

And so comes the STORM (sub-diffraction-limit imaging by STochastic Optical Reconstruction Microscopy). "We're excited about STORM because we think it can reach molecular resolution using optical microscopy," says Zhuang. STORM builds an optical image through the orchestration of photon emissions of individual, switchable fluorescent molecules such that each point of light, due to its proximity to the next in time and space, can be resolved to a lateral resolution of 20nm.

For instance, consider a tangle of neurons. "Often a neuron interacts with many axons and forms multiple synapses, and these synaptic spaces are often smaller than the diffraction limit, especially in baby animals," says Zhuang. "Without the proper resolution, neuronal relationships are hard to discern. "Electron microscopy (EM) used to be the solution. Now, when it comes to intracellular details, optical imaging may even have an advantage over EM due to the molecular specificity afforded by fluorescent probes."

Zhuang also notes that there are several different means besides STORM to obtain super-resolution optical images, including STimulated Emission Depletion microscopy (STED), Structure Illumination Microscopy (SIM) and Photo-Activated Light Microscopy (PALM).

MIMS the Word
For some, the information is the image, for others the importance lies in how the image was formed. "I don't like the word 'image'," says Claude Lechene, MD, principal investigator, National Resource for Imaging Mass Spectrometry, National Institutes of Health/National Institutes of Biomedical Imaging and Bioengineering (NIH/NIBIB), Cambridge, Mass. "In fact, when I take an image, each pixel has its address. But the content of counts at that pixel contains exquisitely quantitative information." The information refers to the presence of stable isotope labels, and the technique is Multi-isotope Imaging Mass Spectrometry, or MIMS. This technique, evolved from a similar concept used for material sciences, takes advantage of the existence of stable, non-toxic isotopes, such as 15N.

What MIMS is actually seeing are secondary ions that have been sputtered from a sample by a primary ion beam: the beam hits and some fraction of newly ionized atoms are reflected back.

But, what does all this mean for resolution? For lateral XY resolution, about 35nm; meaning you can discriminate very easily in a sub-cellular compartment. As for the Z axis, "When the sample is being sputtered, you actually peel off a layer of its composition, and this is enormously small (≅0.5nm)," he explains. Applications of the technique thus far range from pulse-chase, small molecule drug-target interaction to tracking the lineage of transplanted stem cells over time—theoretically for years.

Going with the flow
The essence of pharmacotherapy is to affect change, and the need to observe that change in real-time has never been greater. "That’s something I believe is under-used, especially in industry . . . How living cells respond to stimuli on a high-throughput level—such a capability is extremely powerful." says Philip Lee, PhD, director of research at CellASIC, Richmond, Calif."

Lee is enabling this process through development of a microfluidic platform for live-cell screening: a chip with the ability to keep cells alive, well-fed, contamination-free, and amenable to experiments while under microscopic scrutiny. "By using microfluidic technology (defined as the manipulation of volumes ranging down to the nanoliter scale), we can do high-throughput, fluidic control techniques for kinetic studies, changes of media, change of drugs, flow mixing—basically anything to do with a flow basis that is not possible right now—with micro-titer plates," he explains.

The ability to rapidly switch reagents is of particular interest to Lee: "If you have a 96-well plate . . . in order to exchange a liquid you would have to remove one liquid and flow in the new one, which I think would be really challenging with most automated platforms. With microfluidics, all the units are essentially attached to their own micropipettes that we can run through highly-defined sequences."

Previous designs have also altered the dynamics of the cells themselves, as efforts to scale down chamber size has lead to aberrant changes in morphology or other behavioral characteristics. But Lee’s cells are happy, and you can see it. "The ideal application for our system is microscopy because essentially the surface you’re imaging through is just a cover glass—a standard 1.5mm thickness. This enables 100X oil immersion, fluorescence with confocal . . . it will work with basically any device that takes a slide." The length of time to acquire the image is up to the investigator. For example, Lee has looked at cell death response to various agents for days at a time without removing the chip from microscopic view. "Right now, the only limitation is how good your temperature control is on your stage—if you have good control you should be able to image for a week with no problem," Lee says.