Review Article

Advances in Microscopy Techniques

Daniel B. Schmolze, BS; Clive Standley, PhD; Kevin E. Fogarty, MS; Andrew H. Fischer, MD

microscopy N Context.Advances in morphologic enable visualization of a broad range of new features. Objective.To review and illustrate advances in microscopy with relevance to pathologists. Data Sources.Literature review and new observations. Results.Fluorescence microscopy enables multiantigen detection; allows novel optical-sectioning techniques, with some advantages compared to paraffin sectioning; and permits live-cell imaging. Live-cell imaging allows pathologists to move from a period when all diagnostic expertise was reliant on interpreting static images to a period when cellular dynamics can play a role in diagnosis. New techniques have bypassed by about 100-fold what had FLUORESCENCE MICROSCOPY In traditional transmitted light microscopy, it is the removal of light from the image that provides information. There is only so much light that can be removed to provide more information. In fluorescence microscopy, a virtually limitless addition of light to the image provides information. The spectra of the light from fluorescent microscopy can be separated (reviewed in Levenson et al1), giving fluorescence microscopy many dimensions for conveying information. It is a straightforward process to digitally invert a fluorescent image (to have a white background) and to pseudocolor fluorescent nucleic acid and protein stains (Figure 1, A and B) to closely resemble a standard hematoxylin-eosin (H&E) section (Figure 1, C). A possible application for this approach would be to perform multiple simultaneous immunofluorescent labelings on 1 section, while displaying a digital H&E image. With conventional light microscopy, a practical limit of about 3 to 4 colors can be usefully resolved, including any needed counterstain (eg, hematoxylin). Thus, with conventional transmitted microscopy, 2 or possibly 3 color immunohistochemical labels could be used, with only hematoxylin as a counterstain. However, colocalized antigens are difficult to separate in transmitted light microscopy. On the other hand, if immunofluorescence is used, at least 3 colocalized antigens can be separately
Accepted for publication April 13, 2010. From the University of Massachusetts Medical School, Worcester (Mr Schmolze); and the Departments of Physiology (Dr Standley and Mr Fogarty) and Pathology (Dr Fischer), University of Massachusetts Medical School, Worcester. The authors have no relevant financial interest in the products or companies described in this article. Reprints: Andrew H. Fischer, MD, Department of Pathology, University of Massachusetts Medical School, Biotech Three, 1 Innovation Dr, Worcester, MA 01605 (e-mail: andrew.fischer@umassmemorial.org). Arch Pathol Lab MedVol 135, February 2011

long been believed to be a limit to the resolution of light microscopy. Fluorescence resonance energy transfer (FRET) appears capable of visualizing diagnostically relevant molecular events in living or fixed cells that are immeasurable by other molecular techniques. We describe applications of 2-photon microscopy, FRET, structured illumination, and the subdiffraction techniques of nearfield microscopy, photoactivated localization microscopy, stochastic optical reconstruction microscopy, and stimulated emission depletion microscopy. Conclusion.New microscopy techniques present opportunities for pathologists to develop improved diagnostic tests. (Arch Pathol Lab Med. 2011;135:255263) detected; at the same time, a pseudo-H&E image of the same sample can be displayed by using 2 additional colors in the spectrum. In a possible application, fine-needle aspirations of lymph nodes could be immunofluorescently stained with multiple antibodies, to provide the information afforded by flow cytometry, but could still be viewed digitally with standard, light microscopelike morphology. Several commercial platforms exist for these manipulations, as well as for extracting spectral information and classifying objects (reviewed in Levenson et al1). LIVE-CELL IMAGING Transmitted light microscopy requires use of dyes to visualize nuclear detail; however, no dyes are known to be suitable for live-cell imaging with transmitted light microscopy. Phase contrast and differential interference contrast microscopy techniques cannot display subcellular morphology, particularly within thick tissue samples (Figure 2, B). Fluorescent dyes capable of penetrating viable cells are available for staining nuclei or particular organelles, as are a wide variety of fluorescently labeled proteins that can be expressed in living cells through use of viral vectors (Figure 2, A and C). The fluorescently stained tissue can subsequently be fixed and conventionally stained (Figure 2, D and E). All of a pathologists criteria for cancer diagnosis are still based on static images. If experience at other levels of biology is a guide, it will be much easier to learn about the abnormal physiology of pathologic tissue if dynamic images are studied. The dynamics of cancer cells will most likely provide a fundamentally new set of diagnostic features. The nuclear lamina irregularity, diagnostic of papillary thyroid carcinoma, appears to be due to a dynamic deformation of the nuclear lamina during interphase,2 and this dynamic property appears fundamentally different from the irregularity that develops
Advances in Microscopy TechniquesSchmolze et al 255

Figure 1. Fluorescence microscopy can provide the same information as conventional hematoxylin-eosin (H&E) microscopy. A conventional paraffin section of an ampullary villous adenoma was deparaffinized and rehydrated to 50% alcohol, then simultaneously stained with 49,6diamidino-2-phenylindole (DAPI) (A) and eosin (B) without washing out any excess dye. The image was inverted and pseudocolored to match the colors of an H&E section by using Photoshop (C). We have found that [Ru(bpy)3]Cl2 can be used like eosin and has a useful narrower emission spectrum (original magnification 3100).

during a period of days in neutrophils, or possibly other benign cells that show nuclear irregularity in static images. Live-cell imaging in animal models has demonstrated a direct correlation between tumor cell migration and seeding of the bloodstream.3 Further, dynamic studies have shown an interaction between tumor cells and macrophages in the process of intravasation of tumor cells into vessels.3 These interactions could not be directly

predicted based on static images, and they point toward limitless future diagnostic applications of the dynamic properties of tumor cells. THE DIFFRACTION LIMIT TO RESOLUTION IN CONVENTIONAL MICROSCOPY It has been known since the 19th century that the wave nature of light imposes limits on optical resolution. The

Figure 2. Transmitted light microscopy of thick tissues does not permit visualization of cell structure, whereas fluorescent proteins expressed in living cells are easy to resolve. Living normal human thyroid tissue was visualized after being in organ culture for 18 hours (AC) or after being fixed at 36 hours (D and E). Green fluorescent protein (GFP) bearing replication-incompetent adenovirus was added at the initiation of the organ culture. The GFP signal is shown in (A). The GFP in this example is unconjugated and therefore diffuses throughout infected cells. Green fluorescent protein fused to particular proteins, for example, lamins, permits visualization of subcellular detail (not shown). Barely discernable thyroid follicles are evident by differential interference contrast microscopy in B. C shows the overlay of A and B. Histologically and cytologically, the tissue appears normal after 36 hours (D) and GFP can be detected by immunohistochemistry in the paraffin-embedded tissue section (E) (original magnifications 3200). 256 Arch Pathol Lab MedVol 135, February 2011 Advances in Microscopy TechniquesSchmolze et al

Figure 3. Airy disk and point spread function. Because of diffraction, a microscope produces a blurry Airy disk (A) as the image of a point source of light. The surface plot of this disk defines the point spread function (B), whose shape in turn determines resolution.

effect occurs as light waves cause interference on their way through the optical system and is known as the diffraction limit. The practical consequence of this effect is that a point source of light will produce a fuzzy circle when viewed through a microscope. This characteristic pattern is known as an Airy disk (Figure 3, A) and consists of a central bright dot surrounded by increasingly dimmer circles of light. The corresponding intensity plot is called the point spread function, or PSF (Figure 3, B). The shape of the PSF determines resolution. If the PSF is very narrow or the separation between the points is wide, then 2 PSFs will add up in such a way that the peaks can still have a trough and the 2 images can be resolved. Ernst Abbe addressed the physics and mathematics of the PSF in the 19th century4 and defined an equation that yields the smallest distance x over which 2 objects can be resolved: x~l=2n sin a, 1

defined by a different PSF, x 5 l/(n sin2a), and the practical Abbe limit is about 500 nm. Until the 1990s, the Abbe limit was widely thought of as an absolute barrier to resolution by light microscopy. DIFFRACTION-LIMITED MICROSCOPY TECHNIQUES Before discussing newer optical methods that achieve molecular-level resolution, it is worth looking at some useful techniques that remain bound by the Abbe limit. Confocal Microscopy In standard epifluorescence microscopy, used in most laboratories for immunofluorescence of renal and skin biopsies, or fluorescence in situ hybridization of individual chromosome probes, no mechanism exists to restrict out-of-focus light from being imaged. At high magnifications or with thick sections, blurry light from outside the focal plane contributes to the image, significantly reducing contrast and perceived resolution. This effect occurs in a different, less obtrusive manner in transmitted light microscopy. A laser scanning confocal microscope focuses exciting light onto a small area of the specimen, and it images the resulting emitted light through a pinhole that only allows light from the focal plane to pass through (Figure 5). As a result, an optical section of the specimen at the focal plane is obtained by scanning the focused spot across the specimen. By changing the focal plane, such sections can be imaged at different depths in the specimen, and a 3-D representation can be reconstructed. Confocal microscopy was first described in the 1950s by Minsky5 and is now a relatively mature technology with many commercial implementations. The price and complexity have steadily decreased, and recently a benchtop model has been released.6 Most designs use a set of mirrors to scan a laser across the specimen. Although suitable for some live-cell imaging applications, an alternative spinning disk design collects the image faster, allowing for multiple images per second. The spinning disk design uses a set of concentrically arranged pinholes in a spinning disk, effectively allowing multiple confocal scans to take place simultaneously. Unfortunately, a relatively powerful illumination source is usually needed in all confocal designs, since the pinhole passes only a small amount of light, and this can be damaging to live cells.
Advances in Microscopy TechniquesSchmolze et al 257

where l is the wavelength of light, n is the refractive index of the medium and a is half the angular aperture of the lens. As shown in Figure 4, the angular aperture determines how much of the wave front can be captured by the lens. A material of high refractive index (for example immersion oil, see Figure 4) will bend more light into the field of view of the objective lens, acting like an increase in a to improve resolution. The Abbe equation shows that resolution increases as the wavelength decreases. The higher resolution of an electron microscope (less than 1 nm) relates to the very short wavelength of electrons. Compared to new microscopy techniques (described below), whose resolution is starting to approach that of electron microscopy, the technique of electron microscopy does not easily yield 3-dimensional information; moreover, immunolocalization by electron microscopy is difficult and live-cell imaging for structural dynamics is impossible. The denominator of equation (1) is known as the numerical aperture (NA) of the lens, so the equation can be rewritten as: x~l=2NA 2

With high refractive index media and high NA lenses, the Abbe limit is roughly 200 nm (0.2 mm) in the X-Y plane. For objects above and below the plane, the resolution is
Arch Pathol Lab MedVol 135, February 2011

Figure 4. Concept of numerical aperture. Both a large value of a (half the angular aperture of the lens) and a high refractive index (n) will serve to increase the amount of light from the specimen that is captured by the lens.

Figure 5. How a confocal microscope captures emitted light from only 1 plane. The dichroic mirror serves to reflect the exciting light while allowing the emitted fluorescence to pass through. The pinhole blocks light from outside the focal plane. 258 Arch Pathol Lab MedVol 135, February 2011

Two-Photon Microscopy In the usual case of fluorescence emission, a fluorophore that absorbs a single photon of light at a certain wavelength moves to a higher energy state. When the fluorophore returns to its ground state, it emits a photon with a longer wavelength than that of the photon absorbed. Instead of exciting the fluorophore with 1 photon, 2 photons with half the energy (twice the wavelength) can achieve the same excitation of the fluorophore. A 2-photon microscope operates by using this principle.7 It uses a pulsed infrared laser that is focused to a spot in the specimen. This laser requires a high instantaneous intensity at each pulse, since the absorption of 2 photons simultaneously is a rare event, but the average intensity is low. The illumination intensity falls off rapidly above and below the focal point, and since the chance for absorption of 2 photons increases with the square of the intensity, the probability of absorption falls steeply (roughly to the sixth power) above and below the focal plane. Thus, only fluorophores in a thin volume are excited. When this spot is scanned across the specimen, only light from this thin excitation plane is collected and an optical section is produced (Figure 6). Whereas, in a confocal microscope, fluorophores outside the focal plane are excited but prevented from being imaged by a pinhole, in a 2-photon microscope, absorption is confined to a small volume in the focal plane. The absorption plane can be as thin as about half a micron. Outside this plane, the sample is exposed only to a relatively mild infrared beam with a low average intensity. Thus, photobleaching and phototoxicity is restricted to the plane of focus, and the sample as a whole is exposed to much less potentially damaging radiation, making the technique well suited to live-cell imaging. A second advantage of 2-photon imaging is that the infrared excitation source penetrates more deeply into
Advances in Microscopy TechniquesSchmolze et al

Figure 6. Two-photon microscopy. A pinhole is not needed, since the pulsed laser achieves the necessary high intensity to simultaneously excite fluorophores with 2 photons at 1 plane.

tissue than the visible or ultraviolet light needed for confocal or conventional epifluorescence microscopy. Although the emitted light is still scattered by tissue, it appears feasible to image to a depth of about 200 mm into the tissue and still display diagnostic cellular-level morphology with the 2-photon technique. Confocal optical sectioning is reasonably crisp only to a depth of about 100 mm in whole tissue. The concept of 2-photon microscopy was first described in 1931.8 Costs and complexity have significantly declined in recent years, but microscopes remain expensive (largely owing to the cost of the pulsed laser). Nevertheless, 2photon imaging is being explored for various clinical applications, for example in phototherapy for cancer and macular degeneration.9,10 Two-photon imaging can allow optical sections to be cut, analogously to a paraffin section, on unfixed (living) or fixed unembedded tissues (Figure 7). Fluorescent dyes are available that can be taken up by living cells to display nuclear morphology (eg, 49,6-diamidino-2-phenylindole [DAPI]). The thickness of the optical section can be as thin as about 1 mm, comparing favorably to a paraffin section. Compared to confocal microscopy, it should be possible to cut such optical sections more deeply into the tissue because of the higher penetration of the exciting infrared light. The advantage of optical sectioning compared to paraffin sectioning is the decreased technical steps required for imaging (there is no paraffinization or physical sectioning, and the staining itself can take place in a liquid fixative or tissue culture media without the need for washing out the fluorescent dyes). Another major
Arch Pathol Lab MedVol 135, February 2011

advantage is that the plane of imaging can be precisely chosen without wasting tissue. Applications would appear to be particularly useful for small biopsies in which a tangential optical section on the various sides of the biopsy specimen could allow diagnosis. Diagnostic features for many common problems (eg, determining the presence of invasion in breast biopsy specimens) typically span a diameter of only 200 mm.11 4Pi Microscopy The Abbe limit refers to resolution in the X-Y plane. Resolution in the Z plane is much lower for all microscope designs previously described. Confocal and 2-photon microscopy are limited to about 500 nm of resolution in the Z axis. The Abbe equation shows that an objective lens with a large angular aperture will produce a higherresolution image than one with a smaller angular aperture. 4Pi microscopy places 2 identical objective lenses on either side of the specimen, thereby effectively doubling the angular aperture and thus the resolution in the Z plane of the 2 lenses, to about 200 nm (Figure 8).12 4Pi microscopy remains diffraction limited but can be paired with subdiffraction techniques such as stimulated emission depletion, discussed below. Spatially Modulated Illumination Microscopy When 2 densely patterned images are combined, an interference effect called a Moire pattern can result. This effect is often undesirable (for example, showing up in scanned magazine images). The Moire effect is exploited in spatially modulated illumination (SMI) by using a
Advances in Microscopy TechniquesSchmolze et al 259

Figure 7. Demonstration of 2-photon microscopy optical sectioning of ovarian cancer in a liquid fixative. Two-photon image, about 5 mm in thickness, of a cluster of ovarian cancer cells from ascites, stained with 49,6-diamidino-2-phenylindole (DAPI) and eosin in CytoRich Red fixative (Thermo Scientific, Hudson, New Hampshire), without washing out any excess dye. Pseudo hematoxylin-eosin staining was accomplished as in Figure 1 (original magnification, approximately 31000).

Figure 8. 4Pi microscopy principle. Two lenses are placed on opposing sides of the specimen, thereby effectively doubling the angular aperture and the resolution.

pattern of fine lines as the illumination source.13,14 When this illumination pattern (Figure 9, B) interferes with the detail emitted from or transmitted by the fine structure of the sample (Figure 9, A), a new pattern results (Figure 9, C). The Moire pattern is always more coarsely patterned than the sample, and so it may be resolvable even if the fine structure of the sample is not. The Moire pattern does not necessarily actually resemble the sample, but by collecting a series of images in which the illumination pattern has been slightly altered (eg, by simply shifting its position), the underlying structure can be computed. The interference effect only occurs at the focal plane, and as the illumination pattern is changed, the out-of-focus light remains invariant and can thus be subtracted from the final image. In this way, efficient optical sectioning is possible. Although the technique requires the collection of multiple images, the computation can be performed quickly, enabling live-cell imaging in conjunction with fast cameras. Spatially modulated illumination is commercially available as an add-on to most epifluorescence microscopes at a cost that is considerably cheaper than that of confocal microscopy.15 Spatially modulated illumination is limited by diffraction in that there is a limit to the detail of the illumination pattern. Like 4Pi microscopy, SMI can be combined with other methods to yield further resolution enhancement. One technique merges SMI and stimulated emission depletion to yield subdiffraction imaging with a method termed nonlinear structured illumination.16 SUBDIFFRACTION TECHNIQUES With lenses of the largest possible angular aperture, in media with the highest possible refractive index, the Abbe
260 Arch Pathol Lab MedVol 135, February 2011

equation limits optical resolution to about 200 nm. For a long time, this limit was thought to be unsurpassable, but recently several microscopy methods have been developed that achieve far higher resolution. In fact, the diffraction limit was never a hard limit, and several caveats exist that point the way toward subdiffraction imaging. First, equation (2) only applies in the so-called far field, where the light rays are allowed to travel a certain distance before being imaged (roughly, longer than the wavelength of light). This is because very close to a source (in the near field) the action of so-called evanescent nonpropagating waves predominates, and these waves are not accounted for by the usual rules of interference.17 Secondly, molecular interactions between different fluorophores can be visualized. Finally, the Abbe equation only applies if 2 objects with an identical interaction with light are simultaneously visualized. If there can be a distinction between the 2 objects in the spectra of their emitting photons, or their temporal release of photons, the Abbe limit does not apply, and in fact there is no theoretic limit to light microscopy resolution, even in living cells! Near-Field Scanning Optical Microscope Microscopy techniques have been developed that exploit all of these caveats to break the diffraction limit. The near-field scanning optical microscope scans a tiny probe across the surface of a specimen at distances over which near-field effects come into play. The resolution is limited only by the size of the probe, which can be as small as 10 nm.18 A similar diffraction-limited technique called total internal reflection fluorescence microscopy generates a near-field wave at the surface of a specimen.19 This wave then excites fluorophores over a very narrow distance (Figure 10). Both techniques have been successfully used to visualize dynamic events, in the range of 10 nm, occurring at the plasma membrane of living cells.
Advances in Microscopy TechniquesSchmolze et al

Figure 9. The Moire effect in structured illumination microscopy. A sample containing fine detail (A) when illuminated with a regularly patterned image (B) gives rise to a coarsely patterned Moire image (C). The original fine detail can be computed through collection of several Moire images. An important effect is that the out-of-plane component of the image is able to be computationally subtracted. Reprinted with permission from Gustafsson.15 Copyright 2005, National Academy of Sciences, United States.

Fluorescence/Forster Resonance Energy Transfer Fluorescence/Forster resonance energy transfer (FRET) exploits a different physical principle to obtain high spatial precision.20 Two fluorophores interact in proportion to the sixth power of their intermolecular separation. Within a few diameters of fluorescent molecules, one excited molecule (the donor) can transfer its excited state to a different nonexcited fluorescent molecule (the acceptor). No emission of light is involved in this transfer, and the acceptor will proceed to fluoresce at its characteristic emission wavelength. The degree of separation of the 2 fluorochromes (in the range of 1 nm) can be estimated from the decrease in donor fluorescence and a

Figure 10. Near-field microscopy with the total internal reflection fluorescence microscope. The exciting light is angled such that it is completely reflected when it reaches the sample. A near-field wave is generated, which excites fluorophores in a very narrow range. The resulting fluorescence is imaged as usual. Arch Pathol Lab MedVol 135, February 2011

corresponding increase in acceptor fluorescence. Donor and acceptor probes can be attached to 2 different molecules of interest, or to 2 different domains of a single protein, making FRET a useful tool for measuring proteinprotein interactions or conformational changes at the molecular level (Figure 11). Fluorescence/Forster resonance energy transfer is used in a nonmorphologic context as a read-out for human papilloma virus (HPV) detection in the Third Wave Technologies of Hologic Inc (Bedford, Massachusetts). In this case, 2 adjacent FRET fluorophores on a nucleic acid probe become separated through the action of a specific nuclease that cleaves a site between them. Fluorescence/Forster resonance energy transfer also has morphologic applications in anatomic pathology, with the potential to characterize diagnostic molecular events that are invisible to existing light microscopy or existing nonmorphologic molecular assays. An important example is provided by the signaling pathway from receptor tyrosine kinases to RAS, RAF, and, ultimately, MAPK. There is strong evidence for activation of this pathway (particularly at points upstream and downstream of RAS) in papillary thyroid carcinomas.2123 Paradoxically, activation of RAS is common in follicular-type thyroid tumors, a tumor with distinctly different morphologic and clinical features.23 Immunohistochemistry and Western blotting have not been able to distinguish papillary thyroid carcinomas with tyrosine kinase or BRAF activations from follicular-type tumors or normal thyroid.23,24 The solution to the paradox of how the same signaling pathway to MAPK can cause 2 different responses (apparently in the same cell of origin) could be the existence of differences in the precise temporal and/or spatial aspects of signaling (reviewed in Murphy and Blenis26). The power of FRET is its ability to combine cellular-level localization with dynamic information about molecular interactions, and FRET applications are beginning to dissect out the precise spatial and temporal involvement of events in the MAPK pathway that accounts for paradoxes such as this.27,28 By analogy to the findings in PC12 neuroblastoma cells, one may expect papillary thyroid carcinomas to have a sustained low-level activation of MAPK, whereas follicuAdvances in Microscopy TechniquesSchmolze et al 261

Figure 11. Fluorescence resonance energy transfer. When 2 fluorochromes remain separated by more than a few molecular distances, the exciting light causes only the first fluorochrome to fluoresce. As the distance is decreased, the excited state is transferred to the acceptor molecule, which proceeds to fluoresce at a different wavelength.

Figure 12. Photoactivated localization microscopy and stochastic optical reconstruction microscopy. Within a large population of fluorophores (A), an initial pulse of light (B) causes a sparse subset of the fluorophores to become activated (C). A second pulse of light then causes the activated fluorophores to actually fluoresce (D). The process is repeated until the location of every fluorophore in the sample is individually recorded.

lar-type tumors with RAS mutations would be expected to have a transient high level activation of MAPK.26 It is conceivable that FRET-based assays could find use some day in diagnosis of specific oncogenic signals or molecular interactions that are otherwise invisible by mutational analysis or immunohistochemistry. Fluores cence/Forster resonance energy transfer applications on viable cells provide dynamic, temporal information and would typically require some in vitro manipulations (eg, adenoviral-induced expression of green fluorescent protein fusions, see Figure 2), but static FRET can also be performed on fixed material after labeling with informative FRET pairs.29 Photoactivated Localization Microscopy and Stochastic Optical Reconstruction Microscopy Even if a point source of light is converted by an optical system into a fuzzy Airy disk, repeated sampling of an individual source of light can statistically show the location of the center of its Airy disk. The statistical sampling defines localization precision, a concept distinct from optical resolution. Localization precision is not limited by the Abbe limit, and statistical sampling p improves resolution by a factor of 1 N, where N is the number of sampled photons. Photoactivated localization microscopy30 and stochastic optical reconstruction microscopy31 are related applications of this concept. Within a large population of fluorophores (Figure 12, A), a lowintensity laser excitation first activates rare fluorophores to become susceptible to a second excitation laser pulse (Figure 12, B and C). The first pulse does not actually induce fluorescence; it merely brings the fluorophore to a state where it can become repeatedly excited
262 Arch Pathol Lab MedVol 135, February 2011

to fluoresce. The first pulse is weak enough to activate only about 1 fluorophore molecule in the area of a laser excitation (an area about 300 nm in diameter). By following the first low-intensity activation pulse with a second excitation pulse (Figure 12, D), individual molecules within the 300-nm area fluoresce a population of photons. The population of photons scatters according to a point spread function, and the center of the function can be estimated. The process is repeated through successive 300-nm areas until the location of every fluorescent molecule in the sample is multiply sampled. A high-resolution image is then computed. The process can take many hours to complete, and is thus not yet suitable for studying cellular dynamics, but resolution of about 15 nm has been achieved.31 Stimulated Emission Depletion Stimulated emission depletion32 is notable in that a commercial product has recently been announced.33 Stimulated emission depletion makes use of 2 separate geometries and wavelengths of pulsed light. The first pulse of light is focused to the smallest possible spot (about 300 nm in diameter) and excites all of the fluorochromes within this spot, just like a conventional confocal microscope. A second pulse of light is focused into a donut-shaped image, with the hole of the donut centered on the first pulse (Figure 13). The second pulse has a wavelength that quenches fluorescence induced by the first pulse. The subdiffraction resolution of stimulated emission depletion is due to nonlinear optical effects34 that come into play as the intensity of the quenching beam increases. The end result is a point-by-point scan through the sample at a much higher resolution than conventional
Advances in Microscopy TechniquesSchmolze et al

Figure 13. Stimulated emission depletion microscopy. An initial excitation beam is followed closely by a second beam shaped like a donut, which serves to quench the fluorescence of the first beam. The result is an excitation spot smaller than diameter that can be achieved by diffraction of a single source of light.

techniquesabout 20 nm. Stimulated emission depletion can be applied to green fluorescent proteins and their derivatives, enabling live-cell imaging, but it is currently too slow for visualizing live-cell dynamics. IMPLICATIONS FOR PATHOLOGY At the finest levels of biology it is easy to appreciate how a change in protein structure affects the function of a protein, and it is easy at the tissue level or organ level to appreciate how changes in structure lead to changes in physiology. At the cellular level, the relation between structure and function is less obvious. For example, it is not obvious why nuclear structure should differ between papillary thyroid carcinoma and follicular-type thyroid tumors, or why nucleolar enlargement would be characteristic of prostatic intraepithelial neoplasia. The cellular level is the bottleneck for understanding mechanisms of disease. We can know the gene that is responsible for certain diseases (eg, Huntington disease) and we can know the tissue-level structural changes that result from the gene, but precise understanding of disease mechanisms is lost at the cellular level. Advances in imaging at the cellular to molecular level, especially techniques that allow dynamic imaging of living cells, should enable pathologists to play a central role in uncovering the elusive structure-function relationships that underlie and define diseases.
References
1. Levenson RM, Fornari A, Loda M. Multispectral imaging and pathology: seeing and doing more. Expert Opin Med Diagn. 2008;2:10671081. 2. Fischer AH, Taysavang P, Jhiang SM. Nuclear envelope irregularity is induced by RET/PTC during interphase. Am J Pathol. 2003;163(3):10911100. 3. Wyckoff JB, Wang Y, Lin EY, et al. Direct visualization of macrophageassisted tumor cell intravasation in mammary tumors. Cancer Res. 2007;67(6): 26492656. 4. Abbe EC. Neue Apparate zur Bestimmung des Brechungs - und Zerstreuungsvermogens fester und fluussiger Korper. Maukes Verlag. 1874. 5. Minsky M. Memoir on inventing the confocal scanning microscope. Scanning. 1988;10:128138. 6. Olympus. New FluoView FV10i Confocal Laser Scanning Microscope from Olympus. http://www.nanowerk.com/news/newsid513565.php. Accessed February 4, 2010. 7. Diaspro A, Chirico G, Collini M. Two-photon fluorescence excitation and related techniques in biological microscopy. Q Rev Biophys. 2005;38(2):97166. 8. Goppert-Mayer M. Uber Elementarakte mit zwei Quantensprungen. Ann Phys. 1931;401(3):273294. 9. Bhawalkar JD, Kumar ND, Zhao CF, Prasad PN. Two-photon photodynamic therapy. J Clin Laser Med Surg. 1997;15(5):201204. 10. Ogawa K, Kobuke Y. Recent advances in two-photon photodynamic therapy. Anticancer Agents Med Chem. 2008;8(3):269279.

11. Istvanic S, Fischer AH, Banner BF, et al. Cell blocks of breast FNAs frequently allow diagnosis of invasion or histological classification of proliferative changes. Diagn Cytopathol. 2007;35(5):263269. 12. Hell S, Stelzer E. Properties of a 4Pi confocal fluorescence microscope. J Opt Soc Am. 1992;A9(12):21592166. 13. Neil MAA, Juskaitis R, Wilson T. Method of obtaining optical sectioning by using structured light in a conventional microscope. Opt Lett. 1997;22(24):1905 1907. 14. Gustafsson MG. Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy. J Microsc. 2000;198(pt 2):8287. 15. Qioptiq LINOS, Inc. Qioptiq Web site. http://www.qioptiqlinos.com/. Accessed February 4, 2010 16. Gustafsson MG. Nonlinear structured-illumination microscopy: wide-field fluorescence imaging with theoretically unlimited resolution. Proc Natl Acad Sci U S A. 2005;102(37):1308113086. 17. Girard C, Dereux A. Near-field optics theories. Rep Prog Phys. 1996;59(5): 657699. 18. Durig U, Pohl DW, Rohner F. Near-field optical-scanning microscopy. J Appl Phys. 1986;59(10):33183327. 19. Fish KN. Total internal reflection fluorescence (TIRF) microscopy. Curr Protoc Cytom. 2009;chapter 12:unit 12.18. 20. Clegg RM. Fluorescence resonance energy transfer. Curr Opin Biotechnol. 1995;6(1):103110. 21. Fagin JA. How thyroid tumors start and why it matters: kinase mutants as targets for solid cancer pharmacotherapy. J Endocrinol. 2004;183(2):249256. 22. Michaloglou C, Vredeveld LC, Mooi WJ, Peeper DS. BRAF(E600) in benign and malignant human tumours. Oncogene. 2008;27(7):877895. 23. Nikiforova MN, Lynch RA, Biddinger PW, et al. RAS point mutations and PAX8-PPAR gamma rearrangement in thyroid tumors: evidence for distinct molecular pathways in thyroid follicular carcinoma. J Clin Endocrinol Metab. 2003;88(5):23182326. 24. Pomerance M, Quillard J, Chantoux F, Young J, Blondeau JP. High-level expression, activation, and subcellular localization of p38-MAP kinase in thyroid neoplasms [comment in J Pathol. 2006;210(1):133134]. J Pathol. 2006;209(3): 298306. 25. Zuo H, Nakamura Y, Yasuoka H, et al. Lack of association between BRAF V600E mutation and mitogen-activated protein kinase activation in papillary thyroid carcinoma. Pathol Int. 2007;57(1):1220. 26. Murphy LO, Blenis J. MAPK signal specificity: the right place at the right time. Trends Biochem Sci. 2006;31(5):268275. 27. Abankwa D, Hanzal-Bayer M, Ariotti N, et al. A novel switch region regulates H-ras membrane orientation and signal output. EMBO J. 2008;27(5): 727735. 28. Moreto J, Llado A, Vidal-Quadras M, et al. Calmodulin modulates H-Ras mediated Raf-1 activation. Cell Signal. 2008;20(6):10921103. 29. Periasamy A, Wallrabe H, Chen Y, Barroso M. Chapter 22quantitation of protein-protein interactions: confocal FRET microscopy. Methods Cell Biol. 2008;8956989598. 30. Betzig E, Patterson G, Sougrat R, et al. Imaging intracellular fluorescent proteins at nanometer resolution. Science. 2006;313(5793):16421645. 31. Rust M, Bates M, Zhuang X. Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM). Nat Methods. 2006;3(10):793795. 32. Hell S, Wichmann J. Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy. Opt Lett. 1994;19(11):780782. 33. Leica Microsystems. Leica TCS STEDconfocal microscopes: Leica Microsystems. http://www.leica-microsystems.com/products/confocal-microscopes/ details/product/leica-tcs-sted. Accessed February 4, 2010. 34. Boyd R. Nonlinear Optics. 3rd ed. New York, NY: Academic Press; 2008.

Arch Pathol Lab MedVol 135, February 2011

Advances in Microscopy TechniquesSchmolze et al

263

Copyright of Archives of Pathology & Laboratory Medicine is the property of College of American Pathologists and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.