by Jeffrey M. Perkel
There's no shortage of insights that researchers can draw from looking at cells or their components in isolation. Sometimes, though, the biology can only truly be appreciated in its entirety—seeing the forest for the trees, so to speak.
Enter in vivo imaging, an approach in which biology is studied at an organismal, as opposed to cellular, level. With in vivo imaging, researchers can quite literally shine a light on their animal subjects— generally mice or rats—and derive an answer in the form of a map of emitted light intensity as a function of anatomic position.
"The reason for shifting to whole-animal imaging from any other kind of biological approach is that it allows you to do the analysis in the most complex, most physiologically relevant way possible," says David Daniels, global product leader for pre-clinical imaging at PerkinElmer. "You cannot get all that information in cell-based assays, because they are individual cells separated from the biology one is interested in."
There are other, more practical benefits, too. Because in vivo imaging is non-invasive, animals may be tracked over time, for instance to watch as a tumor develops. That not only simplifies data analysis (as you can collect longitudinal information for each animal), it also reduces the number of animals needed, as they aren't being sacrificed during the experiment.
For instance, explains Anna Christensen, imaging product manager at Caliper Life Sciences, suppose you are interested in a cell that expresses a particular receptor protein. In cell culture, or using biochemical assays, you can study that protein or even its functionality at a particular moment in time—the point at which the cells or proteins were harvested, say. But with in vivo imaging, you can observe it in the functional context of the animal itself, over the long term.
"You basically get the umbrella view of how one specific functional element affects the complete organism through a long period of time, rather than a snapshot, and that is what microplate or slide-based technologies give you," Christensen says.
Researchers interested in in vivo imaging of small animals have a wide range of modalities available, from downsized versions of the medical options of PET, SPECT, CT and MRI to the lab-only tools of fluorescence and bioluminescence—the subject of this guide. Applications run the gamut from stem cell biology and development to cancer, cardiovascular disease and infection. All that's required is an appropriate animal model, a molecular tracer—and of course, an imager.
The basics of in vivo imaging
Although there's a great deal of design, research and development involved in their development and manufacturing, in vivo imagers fundamentally comprise a dark box, a tray to hold the animal, a light source and a camera.
The imagers come essentially in two flavors: Those that can image one modality, and those that can collect multiple types of data in a single unit (multimodal imagers).
LI-COR Biosciences' Pearl® Impulse and UVP's iBox imagers are examples of the former; they focus exclusively on fluorescence. But Carestream Molecular Imaging's MS FX PRO offers four modalities in a single unit: fluorescence, luminescence, radioisotopic (using radio-tracers) and X-ray imaging. Caliper Life Sciences’ IVIS imagers handle fluorescence, luminescence and Cerenkov imaging. (Cerenkov, says Christensen, is an imaging modality in which radioactive PET tracers are detected optically—think of the bluish glow observed in photos of nuclear reactions. Cerenkov imaging allows researchers to test drive in animals the PET tracers they ultimately intend to use in humans. "We recognize this application as one of the fastest growing areas of novel optical imaging technologies," she says.)
Single-mode imagers can be a smart choice for those who are certain that the single modality is all they'll ever need. But multimodality offers an additional advantage: The ability to overlay distinct datasets to put the optical signal into anatomic context. After all, a fluorescent or luminescent image is nothing more than a false color heat map on a black background (or overlaid on a white-light image). But even that only illustrates the animal itself. Picture a multicolor bull’s-eye drawn atop a 2D photo of a mouse; to know precisely where in the body the signal emanates from, you need, at a minimum, X-ray or some other anatomic dataset.
Often this kind of data integration is performed at the software level. Some systems, though, such as PerkinElmer's FMT systems, provide a hardware solution, as well. "We can take the mouse cassette used in the FMT and then move it into a PET, CT, SPECT or MRI," Daniels explains. "DICOM data output allows the researcher to co-register [the datasets] based on the fiduciary marks on top of the mouse cassette."
Optical configurations
Whole-animal imagers can use either of two basic configurations. Some illuminate the specimen from above (reflected, or epi-illumination); others image from below the animal (transillumination). Some, like Caliper's IVIS Spectrum, can image from either direction.
Carestream Molecular Imaging's MS FX PRO is an epi-illumination system that illuminates and images from below the animal. According to Seth Gammon, worldwide in vivo product manager at Carestream, this strategy offers several advantages. "One reason we image from the bottom is you get a nice flat focal plane, because the mouse settles due to gravity," Gammon says. "And you can also image multiple mice at once with uniform illumination." When imaging multiple mice simultaneously from above, the animals can cast shadows on one another, complicating imaging.
Another variable is the excitation light itself. Although many systems use broad-spectrum excitation light, others use lasers. The PerkinElmer FMT 2500 LX quantitative tomography system and LI-COR Pearl® Impulse imager both use laser excitation, for instance. In contrast, UVP's iBox fluorescence imager line uses a Xenon arc lamp to excite over a range of wavelengths from the long-wave ultraviolet into the near infrared. Carestream Molecular Imaging's MS FX PRO, Biospace Lab's Photon Imager and Caliper Life Sciences' IVIS imagers also use Xenon arc lamps.
Yet according to Gammon, anyone serious about in vivo imaging should avoid the relatively shortwave end of the spectrum, as biological tissues tend to absorb and diffract that light.
"There's a really simple mental demonstration," Gammon says. "If you ever put a flashlight up to your hand, what light comes out? Red. Blue and green light is absorbed. So we support [fluors] from visible to near infrared, but we recommend people migrate as quickly as they can from blue/green to near-IR, simply because of the photophysics of getting the near-infrared (NIR) light out."
Many commercial fluorophores absorb and emit in or near the NIR (LI-COR's IRDye700 and IRDye800, for instance). But few fluorescent proteins do; they tend to be excited by blue or green light and emit in the green or red spectral region. As a result, they are a poor choice for in vivo imaging, says Jeff Harford, senior product marketing manager at LI-COR Biosciences.
"A lot of people, because they are using fluorescent proteins in vitro, think they can use them in vivo," Harford says. "But there's a need for education there, helping people understand that when using fluorescence [in vivo], you need to be in the near-infrared."
Cameras
When it comes to whole-animal imaging, sensitivity is key. All imagers include some form of CCD, but there are differences.
For instance, UVP's iBox Spectra, a relatively low-frills manual imager intended for screening studies, uses an uncooled 2-megapixel color CCD for rapid imaging. But the more automated iBox Scientia and iBox Explorer Imaging Microscope (intended for imaging through surgical window or flaps) include more sensitive, cooled black-and-white CCD. Options are available from about 4 to 8 megapixels, says Sean Gallagher, UVP’s chief technology officer.
Biospace Lab uses an intensified CCD (iCCD) in its optical platform, Photon Imager. iCCD is a combination of a fast CCD camera and a light intensifier. According to chief executive officer Olivier Merle, the iCCD enables the instrument to count photons and thus provide highly quantitative results. "The advantage [of that approach] is sensitivity," he says. "Every single photon that reaches the intensifier is actually amplified by one million, and then it becomes visible by the CCD video camera. So this technology gives us very high sensitivity."
2D vs. 3D
Most imagers, whether they image from above or below, generate a planar image, flattening the three-dimensional animal into a two-dimensional projection. Some, however, can generate or at least approximate a 3D dataset.
PerkinElmer's FMT, for instance, uses an imaging mode called "quantitative tomography." Three-dimensional images in this case are calculated by estimating the depth of fluorescent signal based on the properties of the emission light. "Each excitation [event] gives off a number of different emissions," Daniels explains. "We capture each one, and based on the angle and the amount of diffraction that occurs, [we] know whether the fluorescent signal is at the surface, in the middle or at the bottom of the animal."
Caliper's IVIS instruments can generate 2D and 3D images in both fluorescence and luminescence modes, says Christensen, again by quantitatively reconstructing signal location and amount and overlaying that information on a topographic rendering of the animal. The IVIS software can either co-register microCT or MRI scans for precise anatomical reference or use a built-in mouse "atlas," which it morphs to the shape of the subject animal. The user can choose up to 24 different organs to overlay with the optical reconstruction to give anatomical reference to the signal. "It's not exact," she says, "but if you are looking for a sense of where your signal is without co-registration to a second modality, you can get a good approximation of where your signal is coming from."
Carestream's MS FX PRO doesn't enable 3D imaging, but it does offer the ability to image an animal from any direction using a module called MARS (Multimodal Animal Rotation System). That means researchers can locate the animal's fluorescent signal wherever it is—even if it is, say, strongest on the right side of the animal. "With MARS you can sweep through a series of angles to find the sweet spot for peak binding intensity in each animal," says Gammon. "It provides a 360-degree view of the animal."
Buying considerations
Before making a purchasing decision, be sure to consider the kind of imaging you intend to do (fluorescence or luminescence, or both), the dyes you'll be using—you'll want to make sure your system has filters to support that choice—and of course, price. Other variables include:
Flexibility. Make sure the system is upgradable to meet your future needs. Whether that means new fluorophores or new imaging modalities, it makes sense to consider the changing imaging landscape. Biospace Lab's Photon Imager is a modular system in which each imaging modality is available as a separate module. The advantage of that approach, says Merle, is that is reduces up-front cost and increases flexibility. The company already offers modules for fluorescence and bioluminescence on non-anesthetized animals, for instance, and new modules are on the way. Next up, says Merle, are modules to support X-ray imaging (September 2011) and opto-acoustics (2013).
Speed/throughput. Although in vivo imaging is generally fast, if you plan to screen hundreds of animals you'll want to make sure the system can handle it. How quickly can the system take an image? And how many animals can it image at once? The UVP iBox® Scientia can handle up to six mice, says Gallagher. Carestream's MS FX PRO can handle five, says Gammon.
Ease of use. As with all imaging systems, there's a learning curve. Make sure the curve isn't too steep. LI-COR's Pearl imager, says Harford, is as easy to use as a single push-button operation. "I've seen people get trained for three days on an instrument on how to do simple software analysis," he says. "Those same people can be trained in three minutes on the Pearl."
Sensitivity. One popular application for in vivo applications is oncology—specifically, monitoring tumor progression. Though it's easy enough to see a large, fluorescent or bioluminescent tumor, can your system detect it when it's very small? Can it detect micrometastases? Bioluminescence is far more sensitive than fluorescence, says Christensen, in part because there is minimal auto-luminescent background to subtract (as opposed to autofluorescence). "There are [publications] out there where they are able to see a handful of cells, down to five cells in an animal," she says. A related issue is dynamic range: Will the system be able to image the animal over a wide range of signal intensity without adjusting imaging settings? LI-COR's Pearl has a 22-bit dynamic range, says Harford. "That means you can see very dilute signals, but if you also have a lot of signal in other parts of the body, such as the liver, it won't be blown out [that is, saturated] in the same image."
Multimodality. Finally, consider your multimodality requirements. Will you need both fluorescence and luminescence? What about other modalities, such as X-ray, microCT and MRI? Carestream's Albira system offers PET, SPECT and CT imaging to go along with the four modalities in its MS FX PRO. "By partnering with us, you can get any of seven imaging modalities in two instruments," says Gammon. Consider also whether the software that comes with your instrument can perform the required data overlays. If not, you may need to invest in some third-party software.
The images at the top of the page are from PerkinElmer's in vivo Technology page.