The major application of the confocal microscope is in the improved imaging of thicker sections of a wide variety of specimen types. The advantage of the confocal approach results from the capability to image individual optical sections at high resolution in sequence through the specimen. A number of different imaging modes are used; all rely on the optical section as their basic image unit.
Single Optical Sections
The optical section is the basic image unit in confocal microscopy methods. Data can be collected from fixed and stained specimens in single, double, triple-, or multiple-wavelength illumination modes, and the images collected from multiple-labeled specimens will be in register with each other (if an objective lens with adequate correction for chromatic aberration is used). Minor registration errors can usually be corrected using digital image processing methods. Most laser scanning confocal microscopes (LSCMs) take approximately 1 second to acquire a single optical section, although several acquisitions are usually averaged by the software to improve signal-to-noise ratio. The time of image collection will of course vary with the size of the image in pixels and the speed of the system computer. When saved, a typical 8-bit image of 768 x 512 pixels in size will require about 0.3 Mb of storage space.
Presented in Figure 1 are optical sections collected simultaneously at three different excitation wavelengths (488, 568, and 647 nanometers) using a single krypton/argon laser. The specimen is a fruit fly third instar wing imaginal disk labeled for three genes involved with patterning the wing. The three genes imaged and their respective fluorochrome labels are (a) vestigial (fluorescein - 496 nanometers); (b) apterous (lissamine rhodamine - 572 nanometers); and (c) CiD (cyanine 5 - 649 nanometers). The merged composite of the three spatial expression domains of the wing patterning genes is shown in the lower right (image (d)).
Time-Lapse and Live Cell Imaging
Time-lapse studies of living cells are enhanced by the improved resolution of imaging with the LSCM. Early studies of cell locomotion were carried out using 16 mm movie film with a clockwork intervalometer coupled to the camera, and more recently using a time-lapse video cassette recorder, optical memory disk recorder, or video capture card. Now the LSCM can be used to collect single optical sections at preset time intervals.
Imaging living tissues with the LSCM is substantially more difficult than imaging fixed specimens, and is not always a practical option because the specimen may not tolerate the conditions involved. Table 1 lists some of the factors to be considered in imaging live and fixed cells with the LSCM. Some specimens simply will not physically fit on the stage of the microscope, or they cannot be kept alive on the stage during observation. The phenomenon or structures of interest may not be accessible to the objective lens field of view. For example, the wing imaginal disks of the fruit fly develop too deeply in the larva to be imaged, and when dissected out, they cannot be grown in culture. Therefore, the only method currently available to image gene expression in this type of tissue is to dissect, fix, and stain imaginal disks taken from different specimens at various stages of development.
Imaging Fixed and Living Cells with the LSCM
Successful imaging of live cells requires extreme care to be taken throughout the imaging process to maintain tolerable conditions on the microscope stage. Photo damage from the illuminating laser beam can be cumulative over multiple scans so the exposure to the beam should be kept to the minimum necessary to acquire the image. Antioxidants such as ascorbic acid are commonly added to the culture medium to reduce oxygen, which can be released in the excitation of fluorescent molecules, causing free radicals to form and kill the cells. It is usually necessary to carry out extensive preliminary control experiments to assess the effects of light exposure on the fluorescently labeled cells, keeping detailed notes on all the imaging parameters, whether they are thought to be relevant or not. Following the imaging tests, the continued viability of the living specimens should be evaluated. Embryos, for example, should continue their normal development following the imaging process, and any abnormalities caused by the imaging or the fluorochromes used should be determined. Time-lapse imaging of a living fruit fly embryo injected with calcium green is presented in Figure 2. The series of images shows the change in distribution of the fluorescent probe over time.
Specific requirements for life have to be met for each cell type that is to be imaged. Some cells such as insect cells can usually be maintained at room temperature in a large enough volume of the appropriate medium. Most cell types, however, require a stage heating device and possibly a perfusion chamber in which the proper carbon dioxide balance can be maintained during the time they are on the microscope stage. Choosing a cell type that is more amenable to the conditions of imaging in the LSCM can avoid many experimental problems. Potential problems have been reduced significantly by improvements in modern confocal instruments. Their increased photon efficiency, higher numerical aperture (brighter) objective lenses, and less phototoxic dyes for labeling has made live cell confocal analysis a practical option. The best approach is to use the least amount of laser power that allows imaging and to collect the images as quickly as possible. If the pinhole aperture is opened wider than for non-living fixed specimens to speed up the image acquisition, post-imaging deconvolution can sometimes be relied upon to restore lost image quality.
Many physiological processes and events take place faster than they can be captured by most LSCMs, which have image acquisition rates typically on the order of one frame per second. LSCMs using acousto-optical devices and a slit for scanning are faster than the galvanometer-driven point scanning systems, and are more practical for physiological studies. These faster designs combine good spatial resolution with good temporal resolution, which may be 30 frames per second at full screen resolution, or near video rate. The slower point scanning microscope systems can achieve the best temporal resolution only by scanning a much reduced area on the specimen. If full spatial resolution is required, the frames must be collected less frequently, losing some temporal resolution. The confocal systems using disk scanning or oscillating mirror scanning methods are also capable of imaging fast physiological or other transient events.
Z-Series and Three-Dimensional Imaging
A z-series is a sequence of optical sections collected at different levels perpendicular to the optical axis (the z-axis) within a specimen. Z-series are collected by coordinating step-by-step changes in the fine focus of the microscope with sequential image acquisition at each step. The steps in focus are usually accomplished by a computer-controlled stepping motor that changes focus by predetermined increments. A macro program in the computer can be used to acquire and save an image, change the focus by the programmed distance in the specimen, acquire and save a second image, change the focus again, and so on until the programmed number of images have been collected.
Several images may be extracted from a z-series taken through a region of interest, and merged in an image processing program to highlight the cells of interest. The z-series can also be displayed as a montage of images, such as those shown in Figure 3. This type of image combination and display, and many other image operations, are standard features of most current commercial image acquisition and processing software packages. The images chosen for the figure (Figure 3) are representatives of a larger series showing even smaller increments along the z-axis. The green-emitting stain localizes the peripheral nervous system of a fruit fly embryo that was labeled with the antibody designated 22C10.
It can be conceptually difficult to visualize complex interconnected structures from a series of several hundred optical sections taken through the volume of a specimen with a LSCM. Once collected, however, a z-series is ideal for further processing into a three-dimensional representation of the specimen using volume visualization techniques. This approach is now commonly used to elucidate the relationships between the structure and function of tissues in biological and medical studies. It is important that the images are collected at an appropriate z-step size of the motor that changes focus, so that the actual depth of the specimen is reflected in the image. As long as the specimen itself does not move during the acquisition of the images, the z-series produced in the LSCM will be in perfect register, and saved in digital format, they are relatively easily processed into a three-dimensional representation of the specimen. Figure 4 presents a comparison of a single optical section (a) with a z-series projection (b), and illustrates the value of this technique in visualizing the fruit fly peripheral nervous system stained with the antibody 22C10.
The step size taken by the stepper motor, and set up by the microscope operator, is related to the optical section thickness, but they may not have the same value. The optical section thickness refers to the thickness of the section of sample imaged by the microscope, and depends upon the objective lens and the diameter of the pinhole used. In some cases the focus step size and the optical section thickness do have the same value, however, and this may be a source of confusion.
Following acquisition of a z-series file it is usually exported into a computer three-dimensional reconstruction program designed specifically for processing confocal images. Such software programs run at extremely high speeds on graphics workstations, or, with current faster processors and large amounts of RAM, can be run quite effectively on personal computers or the workstation of the confocal microscope. The three-dimensional software packages can be used to produce either a single three-dimensional representation of the specimen or a movie sequence compiled from different views of the specimen that can produce the effect of rotating or other spatial transformation that enhances the appreciation of the specimen's three-dimensional character. The software allows various length, depth, and volume measurements to be made, and specific parameters of the image such as opacity can be interactively changed to reveal structures of interest at different levels in the specimen.
Another way in which a series of optical sections from a time-lapse sequence might be utilized is to process the data into a three-dimensional representation so that time is the z-axis. This is a useful approach for visualizing physiological changes during organism development. An example in which this method has been used is in the elucidation of calcium dynamics in developing sea urchin embryos. Color-coding optical sections taken at different depths is a simple method of displaying three-dimensional information. In practice a color (usually red, green, or blue) is assigned to each optical section obtained at a different depth in the specimen, then the colored images are merged and the colors manipulated to achieve the desired effect using an image processing program.
Living tissue preparations or other specimens exhibiting dynamic phenomena present the possibility of using the LSCM to collect time-lapse sequences of three-dimensional data to be presented with time as the fourth dimension. Z-series data collected at time intervals will produce 4-dimensional data sets, three spatial dimensions (x, y, and z) with time as the fourth dimension, which can be viewed using a 4D viewer program. Such programs allow stereo pairs taken at each time point to be constructed and played back as a movie, or, alternatively three-dimensional reconstructions for each time point can be processed and displayed as a movie or a montage.
If a profile view is needed of a specimen, such as a vertical slice of an epithelial layer, an x-z section can be produced in one of two ways. The profile can be constructed by scanning a single line across the specimen (the x-axis) at different z-axis depths by stepper motor control of focus changes, then displaying the series as a merged image. Another method is to use a cut plane option in a three-dimensional reconstruction program to extract the profile from an existing z-series of optical sections. In construction of the images of butterfly wing epithelium in Figure 5, the laser was scanned across a single line (horizontal black line in the left-hand image) at different z-axis positions, or depths, progressing into the specimen. The x-z image presented in Figure 5 was built up and displayed by the confocal imaging system. The wing epithelium is made up of two epithelial layers, but as fluorescence intensity drops off at greater depths in the specimen, only the upper layer is clearly visualized.
Reflected Light Imaging
Reflected, or backscattered, light imaging was the imaging mode used in all of the early confocal microscopes. Many specimens can be viewed in the LSCM in an unstained state using reflected light, or the specimen can be labeled with a probe that is highly reflective, such as immunogold or silver grains. An advantage of the reflected light method, especially for living tissue samples, is that photobleaching is not a problem. Some types of probes may attenuate the laser beam, and another potential problem is that in some microscopes, internal reflections can occur from optical elements in the light path. The reflection problem is not present in the slit or multi-beam versions of LSCM, and in instruments where it is troublesome, use of polarizers or imaging away from the artifact and off the optical axis can alleviate it.
Transmitted Light Imaging
Any of the transmitted light imaging modes commonly employed in microscopy can be used in the LSCM, including phase contrast, differential interference contrast (DIC), dark field, or polarized light. A transmitted light detector is used to collect light passing through the specimen, and a fiber optic light guide transmits the signal to one of the photomultiplier tubes in the microscope system's scan head. The transmitted light images and confocal epifluorescence images can be acquired simultaneously using the same illumination beam, ensuring that all of the images are in registration. When the images are combined or merged using image processing software, the precise location of labeled cells within the tissues can be mapped. An informative approach in some studies is to combine a transmitted light, nonconfocal image of a specimen with one or more confocal fluorescence images of labeled cells in the same specimen. Use of this approach would allow, for example, determining the spatial and temporal aspects of the migration of a subset of labeled cells within a population of unlabeled cells for a period of hours or even years.
A color transmitted light detector has now been introduced that collects the signal transmitted in the red, green, and blue (RGB) color channels to create a real color image in a way that is similar to some digital color cameras. Such a detector is especially useful to pathologists, who are accustomed to viewing true colors in tissues in transmitted light and overlaying these images with fluorescence data.
Stephen W. Paddock - Laboratory of Molecular Biology, Howard Hughes Medical Institute, University of Wisconsin, Madison, Wisconsin 53706.
Thomas J. Fellers and Michael W. Davidson - National High Magnetic Field Laboratory, 1800 East Paul Dirac Dr., The Florida State University, Tallahassee, Florida, 32310.