Point-Scanning Confocal and Multiphoton Approaches

Confocal microscopy has become a centrally important tool to the research microscopist. The technology allows optical sectioning of cells and tissues and recovery of focused, low noise images. The basic principal behind confocal microscopy is to deliver illumination and collect the returned signal through a pinhole of defined size that rejects out-of-focus light. Clearly however, this simple system will only collect light from a single point in the specimen. To collect a full image of the specimen, it is necessary to move the point source relative to the specimen. In fact, in the first usable design, developed in the late 1950s (1), the pinhole and illumination stayed stationary, and the specimen was scanned beneath the pinhole using a scanning stage. However, this approach is extremely slow (due to the mass of the stage) and not a currently useful solution. The first practical application (2) of the technology we use today was introduced 20 years ago. This instrument, which as a commercial device became the Bio-Rad MRC 500, moved the illumination relative to the specimen using two galvanometer-driven mirrors (one scanning rapidly, at approximately 1000 times/s in one axis, and the other scanning slowly, at 1/s in the orthogonal axis). This raster-based approach built a two-dimensional image in a similar fashion to a simple television image. Since this seminal development, there have been innumerable reinventions of the same conceptual design, some using a single mirror scanning in both axes, others a combination of stepping motors and galvonometers. All devices were relatively slow and, apart from the devices made by the “big four” microscope companies (Leica, Olympus, Nikon, and Zeiss), none have survived to the current day.

Concurrently, there has been a constant need to develop machines that could work faster with high signal-to-noise ratios and with multiple fluorescent markers. This need was principally driven by physiologists, who required the ability to examine ionic reporters in real-time in living systems. Scanheads that were developed in these early days were extraordinary in complexity, but were fundamentally limited by the sensitivity of the detectors. A greater problem, however, was the performance limitations in the computers of the time, which were challenged both from the point of view of data throughput and storage.

Today, the majority of the problems that limited the use of conventional bench-mounted confocal microscopy have been resolved, and while point-scanning laser-based confocal microscopes are only sold by the four major microscope companies, all the available systems are extremely sensitive, flexible, and have highly sophisticated and useful analysis software. Furthermore, high-speed confocal imaging is becoming much more practical, with the ability to collect images at frame rates of over 100 frames/s, at high resolution, and with multiple colors. These systems either work in line scan mode rather than point scan or use multiple focal volumes that are rapidly scanned across the sample. More recently, these latter technologies have been facilitated by major developments in electron multiplier array technologies that have much higher sensitivity than cameras available even 5 years ago.

The investigator, however, continues to demand more flexibility and utility from the confocal microscope. In particular, there is a constant need to image deeper into tissues or to image specimens that do not conform to the constraints of a conventional microscope platform. Deep tissue imaging in living animals has, to some degree, been solved using multiphoton approaches. This technology, developed at Cornell in 1990 (3), uses ultrafast pulsed long wavelength lasers as the illumination source. If two long wavelength photons arrive at a fluorophore essentially simultaneously, they act synergistically, donating the photonic energy of a shorter wavelength photon, and hence exciting the fluorophore. One reason this approach allows deeper penetration is that long wavelength light is absorbed and scattered much less than short wavelength light. Furthermore, as fluorophore excitation only occurs at the focal plane of the lens, no pinhole is needed on the emission side of the device. As such, even if the emission light is scattered dramatically by the tissue, it can still reach the detector and provide useful signal since precise focusing of the light through a pinhole is not required. However, penetration is still limited to a few hundred microns and generally to samples that can be placed conveniently on a conventional microscope stage.