Why Light-Sheet Microscopy?
The Importance of Volumetric Imaging.
Historically, 3D imaging has been performed with spinning disk and laser scanning confocal microscopes. However, these microscopes indiscriminately illuminate regions above and below the focal plane and thus incur high rates of photobleaching and phototoxicity. This problem is further exacerbated by their reduced duty cycle of illumination (e.g., the percentage of the image acquisition duration that a region of the specimen productively contributes to image formation), which necessarily requires a compensatory increase in laser power. In contrast, depending on their design, LSFMs largely restrict illumination to a 2D focal plane of interest within a specimen, and owing to high quantum efficiency and massively parallel scientific cameras (e.g., 4x106), the laser power can be reduced without compromising the imaging speed or signal to noise ratio. To illustrate, a common voxel dwell time for a laser scanning confocal microscope is 1µs, and thus it takes ~4.16s to capture a 2048x2048 voxel image. For a LSFM, an identical region can be imaged ~5ms, 832-fold faster, yet with a per-voxel dwell time that is 5,000-fold longer. Consequently, LSFM allows for very delicate imaging with high spatiotemporal resolution.
Challenges with 3D Imaging.
To gain insight into cell biological events at subcellular scales, volumetric imaging technologies must be combined with molecularly specific labels, biosensors and opto- and chemogenetic tools, as well as computer vision analyses that translate non-human interpretable 5D (x, y, z, \(\lambda\), t) datasets into biological insight. This requires that several key criteria be met:
To gain quantitative insight, the event of interest must be Nyquist sampled in both space and time. For example, the GTPase cycle times of Rho, Rac, and Cdc42 are as short as 5s, and occur on the micron scale. This requires that a complete volume be acquired every 2.5s, with a spatial resolution <500nm. And ideally, to avoid cell morphology-dependent intensity artifacts, the resolution should ideally be isotropic or nearly isotropic in X, Y, and Z. Indeed, this is particularly important for signaling events taking place at or on the plasma membrane, as is the case for GTPases.
To multiplex cellular readouts, the microscope must be compatible with simultaneous multicolor excitation and detection. For microscopes designs that leverage spatial light modulators within their excitation train, this is not possible.
The microscope must have an uncompromised detection sensitivity to reduce the impact of ectopic expression of signaling active proteins, and to reduce photobleaching and toxicity. This problem is particularly exacerbated for the latter, as 10-100 image slices must be acquired per time point to construct a complete imaging volume.
When imaging cells in extracellular matrix environments (which can migrate in any spatial dimension) or a developing embryo, one must also maximize the field of view of the microscope (>100 x 100 x 100 microns). Indeed, many leading LSFMs are designed to image very small fields of view (e.g., 25µm).
And lastly, to be compatible with downstream analytical routines, one cannot use iterative deconvolution or structured illumination routines to improve microscope performance, as these methods alter the numerical statistics of the data.
Why build a microscope?
Importantly, the technology necessary to achieve multiplexed volumetric imaging with advanced probes and computer vision analyses already exists. However, owing to the need to produce aesthetically attractive, highly engineered, and serviceable optical systems that deliver a large return on investment, microscope manufacturers are incredibly conservative when it comes to adopting emerging technologies. Consequently, most microscopes take >7 years to commercialize, and are immediately considered obsolete by the microscopy community owing to scientific advances that take place during the interim. One exception is the LLSM, which was sub-licensed by Zeiss to 3i months after its seminal publication. Nonetheless, even here, it took an additional 6 years for Zeiss to release their consumer- friendly model, which prohibitively costs ~$1M USD. And lastly, a highly convoluted and entangled patent landscape adds further delays and limits start-up companies from accessing the consumer market. For example, despite their own limited development of the technology, Leica has exclusively licensed a patent for off-axis tertiary imaging systems, thereby effectively blocking commercialization of oblique plane microscopy. Consequently, commercialization not only delays technology adoption, but can serve as an active impediment to the dissemination of potentially transformative and cutting-edge technologies.