MOM#3: LIGHT SHEET MICROSCOPY (LSM) Stanley Cohen, MD

Transcription

1 MOM#3: LIGHT SHEET MICROSCOPY (LSM) Stanley Cohen, MD Introduction. Although the technical details of light sheet imaging and its various permutations appear at first glance to be complex and require some mathematical formalism for true understanding, the basic principles are very simple. The basic idea of LSM is to illuminate the sample with a thin sheet of light and to detect the emitted signal with an orthogonally arranged wide-field detection device. This allows the rapid accumulation of a large array of data points while minimizing effects of photo-bleaching, which can occur when a focal beam of intense laser light is used, as well as its phototoxic effects when living tissue is being examined. Perhaps the greatest impediment to an overview of the field is not its inherent complexity, but rather the alphabet soup of four letter acronyms that are used to describe its various permutations. As confocal technologies mature LSM will become more accessible and increasingly useful for biomedical studies in general and pathology in particular because it facilitates 3D spatial visualization of complex biological phenomena. General Principles of Confocal Laser Scanning (CLSM) Microscopy. In order to appreciate the power of light sheet microscopy it is helpful to begin with a brief overview of traditional confocal microscopy, which was developed to overcome the limitations of conventional fluorescence microscopy. Major limitations of fluorescence microscopy include the loss of contrast and resolution due to fluorescence from out-of-focus structures, and light scatter as the illuminating beam travels through an inhomogeneous medium (the specimen sample). Diffraction effects also create a significant constraint for conventional microscopy. Diffraction results when a beam of light is spread out as a result of passing through a narrow aperture or across an edge, resulting in interference between the wave forms produced. In a conventional microscope this diffraction occurs as the wave-fronts pass through the circular aperture at the rear focal plane of the objective. This leads to a diffraction barrier (or diffraction limit) that restricts the ability of traditional optical instruments to distinguish between two objects separated by a lateral distance less than approximately half the wavelength of light used to image the specimen. This diffraction limit was initially described by Ernst Abbe in 1873, and is known as the Abbe limit. In a confocal laser scanning microscope, the specimen is illuminated by a beam of laser light focused to a diffraction-limited spot within the specimen. This reduces the fluorescent light emitted in regions outside the focal plane of the objective. The emitted fluorescent light is separated from the incident exciting laser light and then passed through a confocal pinhole, which rejects the fluorescent light generated outside the focal plane as well as most of the light scattered by the specimen. The specimen image is generated by scanning the spot across the specimen and processing the resulting signals by computer to create the actual digital image, and thus, in a manner of speaking, confocal microscopy represents one of the earliest examples of

2 computational imaging in pathology. Principles of confocal microscopy and its extensions are discussed in detail in (1). Evolution of Confocal Microscopy. There have been many improvements in CLSM since its inception, including but not limited to spinning disk confocal microscopy and multiphoton confocal microscopy. In spinning disk microscopy a series of points in parallel across the specimen are scanned by a spinning disk with an array of pinholes arranged in spiral patterns. In multiphoton microscopy, a high excitation laser is rapidly pulsed so as to induce two-photon excitation within a small volume, obviating the need for a pinhole on the collection side. Even more exciting (pun intended) is the use of various quenching techniques such as FRET (fluorescence resonance energy transfer) to visualize molecular interactions with high spatial resolution. Super-resolution by confocal microscopy. The typical lateral resolution of a confocal microscope is about nm and about nm in the axial plane. 4pi microscopy utilizes two opposed objective lenses focused within the same imaging volume leading to interference of two spherical wavelengths. Computational processing can then improve axial resolution and bring it to the same magnitude as lateral resolution. However, technology has progressed much further than this. Confocal microscopy gives us the ability to bypass, if not overcome, the diffraction barrier by exploiting loopholes in the laws of optics. This ability to resolve details beyond conventional diffraction limits is known as super-resolution. As can be expected from such a hot area of research, there have emerged a wide variety of methods to achieve super-resolution. It is at first difficult to approach the literature in this area because it is characterized by an alphabet soup of techniques including STED (stimulated emission depletion), PALM (photo-activated localization microscopy), STORM (stochastic optical reconstruction microscopy), SOFI (super-resolution optical fluctuation imaging), SIM (superresolution structured illumination microscopy), etc. We will only consider STED as an example of the ingenious approaches that have been developed. STED utilizes two lasers, one to excite the fluorophores as usual, and another shaped to yield a concentric ring around the excited beam. The wavelength of this second laser is selected to inactivate the fluorophores at the periphery of the excited region before they fluoresce. This, in effect, reduces the diameter of the exciting beam below the diffraction limit. Light Sheet Microscopy (LSM).The basic idea of LSM Is to illuminate the sample with a thin sheet of light and to detect the emitted signal with an orthogonally arranged wide-field detection arm. There are several ways of accomplishing this. One approach is to illuminate the sample sheet of light parallel to the focal plane of an objective. This results in optical sectioning, with low photobleaching and low phototoxicity for living specimens. The volumetric data can be acquired either by scanning the sample through a stationary light sheet/detection objective or by moving the light sheet/objective synchronously to scan a stationary object. In both cases, the volumetric scanning rate is limited by physical motion requirements. Data from multiple planes can be used for 3D reconstruction. Light sheet microscopy has been exploited in examining whole living specimens such as zebrafish and other animals. Techniques for clearing tissues have allowed imaging of brain neurons in tissue specimens, and with suitable probes, functional data have been 2

3 obtained, but with only approximately one second temporal resolution. More recently, improvements in light sheet microscopy have allowed for good subcellular resolution in tissue specimens. A very good collection of articles on light sheet microscopy can be found in (2). One technical problem with LSM is the vast amount of data that are generated to create the three dimensional reconstructions, especially since existing and potential modifications to LSM utilizing multiphoton and super resolution techniques compound this difficulty. This problem is now tractable thanks to advances in computation and data storage. However, these advances make it possible to collect so much data that it slows down the process of image acquisition, thus imposing real-time limitations on experimental design. For this reason, a great deal of attention has been paid to developing techniques for fast volumetric imaging, of which only three will be discussed here (SPED, SCAPE, and mesbim). SPED refers to spherical aberration assisted extended depth-of-field light sheet microscopy (3). In general, in LSM lateral resolution is determined by numerical aperture and the axial resolution by light sheet thickness. The latter is dependent on the point-spread function (PSF). The PSF is the three dimensional image of a point-like object under the microscope, created by diffraction limitations that don t allow visualization of the exact point. Extending the PSF thus extends depthof-field. SPED combines a large volumetric field of view via an extended depth of field with the optical sectioning of LSM, which enables scanning of thousands of volumes/second, limited solely by camera acquisition rate. This involves extending the depth of field by building upon optical methods that induce spherical aberrations, thus turning a limitation (aberration) into an advantage, since it has been observed that spherical aberration results in PSF elongation. This is effected by inserting a thick block of altered refractive index material between the objective and specimen, thus introducing a uniform spherical aberration. The process by which optical degradation is minimized under these conditions is complex. Another fast capture technique is known as SCAPE swept confocally aligned planar excitation microscopy (4). SCAPE is a hybrid between LSM and confocal scanning microscopy that uses a single objective lens for both illumination and detection, making sample positioning and alignment much simpler than with conventional LSM. Basically, it utilizes an oblique light sheet swept through the sample with a polygonal scanning mirror. The illuminated plane always stays aligned with a stationary camera. It requires no sample mounting or translation, and thus makes it especially useful for the study of freely moving living samples, for example migrating cells or intact (small) organisms. Imaging speeds are an order of magnitude better than traditional LSM. Many of the visualization strategies currently available to us involve tradeoffs, such as between high-resolution, isotropy (resolution parity in all directions), wide field of view, difficulties in real time capture, etc. An additional difficulty is the problem of examining cellular and tissue patterns in the context of the microenvironment in which they exist in vivo. Most of the initial applications of LSM have involved Gaussian beams to illuminate the target, but newer techniques, such as mespim ( micro-environmentally selective plane illumination microscopy ) utilize Bessel beams that are propagation invariant in that they maintain their cross-sectional profile over long distances, and provide extended depth of focus. Just as there is a difference between lateral and 3

4 axial resolution in conventional confocal microscopy, there is an inverse relationship between field of view and axial (Z) resolution. The use of Bessel beams is one strategy to minimize this latter effect. A discussion of the difference between Gaussian and Bessel beams in illumination is beyond the scope of this discussion, but detailed comparisons may be found in (5) and (6); the former reference is provided to illustrate that not all scientific advances are confined to the past 20 years. The use of Bessel beams is one of the features enabling mespim. These and other principles of mespim are described in detail by Welf et al (7), a paper by the originators of this technique. mespim allows the imaging of cells in the context of the collagen matrix of their microenvironment over large fields of view. It also allows studies of cell clusters. In (7), examples of imaging of the morphological diversity of melanoma cells in their unperturbed environment are presented. Here, the combination of mespim with computer analysis was shown to enable automatic detection and tracking of dynamic 3D morphological structures, further improving the processes of data collection and analysis. Summary and Overview. Light sheet microscopy has resulted in the ability to generate highly detailed three-dimensional reconstruction of both structural and functional images from living and preserved tissues. These approaches have been used with great success in studies of the normal and abnormal nervous system and in embryogenesis as well as many other areas of research. In pathology, techniques for using LSM in formalin-fixed paraffin-embedded sections allow for a variety of studies, of which (8) provides an example involving comparisons of invasive and noninvasive breast cancer. By capturing large fields of view in real time, LSM can help sort out complexities of cell and tissue interactions. Additionally, when coupled with 3D visualization techniques based on biochemical and/or biophysical resolution, such as mass spectroscopic imaging (MALDI), atomic force (ATF) microscopy imaging and Raman imaging, it will be possible to obtain important insights as to how chemical and physical processes are distributed and how these patterns determine ultimate normal and abnormal function. Further Reading: 1. Fritzkey, L. and Lagunoff, D. Advanced Methods in Fluorescence Microscopy, in Biophotonics in Pathology, Cohen. S., editor, 2013, IOS Press, pp Tomer, R. et al. SPED LSM: Fast mapping of biological system structure and function. Cell, 163: , Bouchard, M.B. et al. Swept confocally-aligned planar excitation (SCAPE) microscopy for highspeed volumetric imaging of behaving organisms. Nature Photonics 9: , See also for a non-technical discussion. 5. Durnin, J., Micelli, J.J. and Eberly, J.H. Comparison of Bessel and Gaussian beams. Optical Letters 13: 79-80,

5 6. Nowack, R.L. A tale of two beams. Stud. Geophys. Geod. 56: , Welf, E.F. et al. Quantitative Multiscale Cell Imaging in Controlled 3D Microenvironments. Developmental Cell 36: , Bucur, O. et al. Abstract 3477: 3D morphological hallmarks of breast carcinogenesis: Diagnosis of non-invasive and invasive breast cancer with LSM. Cancer Res. 75:3477,