Institute of Molecular Genetics, Academy of Sciences of the Czech Republic, Prague 6-Dejvice, Czech Republic; *Department of Anatomy and Developmental Biology, University College London, London, UK
Experience with confocal reflection imaging indicated that the original Noran Odyssey with composite video out-put (no longer in production) had apparently been the best system for backscattered light imaging of living cell inter-nal dynamics. Rather than choose other systems, or the then-existent Noran upgrades (Odyssey XL, OZ), the bet-terment of the existing Odyssey appeared to be the right ap-proach. The improvement included introduction of a 633 nm red HeNe laser and direct digital recording of the com-posite video signal in full PAL. Due to the lower dispersion of red light, we can obtain deeper penetration into the liv-ing cells, or small spheroids composed of up to 100 cells, or even into small fragments of normal or tumour tissues.
The images have sharper contrast. Direct digital recording is achieved with a Sony DCR-TRV320E Camcorder with unblocked composite video input. The necessary movie editing and separation of images for further processing and animation is achieved in combination with a Pinnacle System Studio DV card (fire wire connection) and pro-gram. Such a system minimises information losses that previously occurred when the signal from VHS tapes was digitised. Direct computer grabbing is generally out of consideration because we do not know in advance of any
ing for signal attenuation in confocal microscope images. Proc Microsc Soc Amer 58thAnnual meeting, Philadelphia, (2000) 4. Al-Kofahi K, Lasek S, Turner JN, Roysam B: Rapid automated
3-D tracing of neurons from confocal image stacks. Proc Mi-crosc Soc Amer 58thannual meeting, Philadelphia, (2000) 5. Becker DE, Ancin H, Szarowski DH, Turner JN, Roysam B:
Au-tomated 3-D montage synthesis from laser-scanning confocal images: Application to quantitative tissue-level cytological analysis. Cytometry, 25, 3, 235–245 (1996)
6. Holmes TJ, Bhattacharyya S, Cooper JA, Hanzel D, Krishna-murthi V, Lin W, Roysam B, Szarowski DH, Turner JN: Light Microscopic Images Reconstructed by Maximum Likelihood Deconvolution. Handbook of Confocal Microscopy, (Ed. Paw-ley J). Plenum Press, New York, (1995)
Scanning oblique illumination for three-dimen-sional image representation in light microscopy A. BOYDE, C.W. HEWITT,*G.L. GREENBERG†
Department of Anatomy and Developmental Biology, University College London, London, UK; *Department of Surgery, Robert Wood Johnson Medical School, Camden, NJ; †Edge 3D Imaging LLC, 441 N. 5th Street, Philadelphia, PA, USA
Real time 3-D images can be obtained from conven-tional (non-confocal, single objective) high-resolution transmission light microscopic systems by controlling the aperture of illumination, for example, by half-aperturing, such that each eye sees information deriving from opposite oblique ray bundles (1,2). Here we address the question of how much 3-D information can be obtained from tempo-ral display sequences of images, that is, using motion rather than stereo-parallax.
Switching back and forth between the views obtained from the left, both, the right, both, and again the left off-axis illumination channels gives the impression of a tilting or rocking object, but backwards and forwards rocking dis-plays of 3-D data sets are annoying. Motion parallax depth cues are best appreciated when the movement is continu-ous in one sense. Obvicontinu-ously we cannot keep looking at the same scene if we “fly past” it, so that continuous rotation of the real or apparent object is required. Since we cannot rotate other than very specially contrived samples on their own axes under a high numerical aperture light micro-scopic objective lens, we can exclude a real rotatory motion of a cylindrical sample as a realistic possibility.
We surmised that changing the direction of incidence of an oblique illuminating cone such that it spins whilst tilted with respect to the mean optic axis of the microscope sys-tem would create the illusion of the sample tilting contin-uously, or of the successive layers within the imaged vol-ume moving past each other. We have made and evaluated several practical microscopic systems using this simple principle, and all worked. The simplest form of obscuring aperture is a pie-sector opening in an opaque disc placed at any of the relevant conjugate planes, the most accessible
experiment of image series what information is worth-while retaining.
The provision of the best physiologic conditions is an es-sential requirement for the investigation of living cells.
This required the construction of a suitable housing for the microscope body that would ensure maintenance of the important survival parameters, including temperature and nutrition, as well as allowing for easy experimental pro-cedures. After fulfilling these requirements, a plethora of motions inside living cells could be observed, and these motions could be tentatively linked to cellular activity sub-sequent to various experimental manipulations.
The relationships between fast intracellular motion (FIM) and Brownian motion were clarified by experimen-tal treatments known to change biological processes re-versibly or irrere-versibly. Rapid exposure to ice cold medium temporarily but reversibly slows the speed of FIM. Hypo-tonic medium raises the speed of FIM, whilst hyperHypo-tonic media slow it. The FIM rate is largely unchanged by rapid cooling and warming. A degree of heating which will even-tually cause cell shrinkage and detachment temporarily increases the speed of FIM. Fixation with glutaraldehyde initially reduces and then arrests FIM within about 3 min without changing the appearance of the structure under observation.
PowerPoint has revolutionised the chances for public demonstration of the range of phenomena of FIM in vari-ous primary normal and neoplastic cells and cell lines, and their changes under conditions threatening survival and their reactions during rescue.
Acknowledgments: Conversion of the Noran Odyssey VRCSLM to operation with a red laser was funded by a Royal Society Grant to AB. This work was also supported by grant No. 304/99/0368 from the Grant Agency of the Czech Republic and a Senior Visiting Fellowship from The Anatomical Society of Great Britain and Ireland for PV.
Red laser video-rate scanning confocal microscopy in vivo: experience in studies of renal tubular function
M. SIMEONI,*†‡ R.J. UNWIN,* D.G. SHIRLEY,*
G. CAPASSO,† A. BOYDE‡
*Centre for Nephrology, University College London, Middlesex Hospital, London, UK; †Second University of Naples, Naples, Italy; ‡Department of Anatomy and Developmental Biology, University College London, London, UK
Confocal microscopy is now in widespread use in biol-ogy to improve fluorescence imaging, mostly using slow scanning systems to document preserved cells and tissues in which motion in the subject and therefore the speed of image acquisition are no problems, but also at greatly
re-duced morphologic resolution to study fast ion fluxes in live cells in vitro. We attempted to create a system to improve structural and functional imaging at high temporal resolu-tion in vivo, which is an absolute requirement where parts of the whole subject are changing very rapidly. We recon-figured a video rate laser scanning confocal microscope with the aim of making improvements in its performance in studying live tissues in vivo at a moderate scale of reso-lution (e.g., a field of view of 200 µm) and in examining the behaviour of intracellular organelles in living cells at ex-tremely high resolution (e.g., a field of view of 5 µm ob-tained using a 100/1.4 objective and 10× zoom magnifica-tion). Our Noran Odyssey system was originally fitted with an argon ion laser with a dominant emission at 488 nm (blue-green), particularly suited for exciting green-yellow fluorescence, with additional lines at 458 and 514 nm. By fitting a 633 nm HeNe red laser, we reduce diffuse scatter and increase the depth at which we can image into tissues, at the same time reducing the tissue damage due to the laser radiation. Reflection mode confocal imaging at video rate is already remarkably impressive: by being able to use more power with less cellular damage, we achieve effec-tively noise-free imaging.
A particular aim of this study was to examine the appli-cation of video-rate scanning confocal microscopy to the nephron in vivo. This was done by direct visualisation of proximal and distal tubular segments following treatment with three classes of diuretics whose principal sites of ac-tion are established. Rats were anaesthetised and surgi-cally prepared as for in vivo renal micropuncture, with the left kidney exposed, freed of fat, and placed on a support-ing acrylic platform for imagsupport-ing ussupport-ing a 40/1.0 oil immer-sion objective lens with glass coverslips sandwiched to 400 µm. We used optical sectioning at depths of 10 to 50 µm below the intact kidney capsule to observe superficial proximal and distal renal tubular segments and blood flow.
Blood cells can easily be identified passing rapidly through peritubular capillaries.
Early and late proximal and segments can be distin-guished on the basis of the reflectivity of the brush border, and this is absent in distal segments. Baseline measure-ments of tubule section diameter were made and analysed on image frames grabbed on-line, or from video recordings.
After acute intravenous injection of mannitol (500 mg/kg), increases in both proximal and distal tubular diameters were observed. However, after frusemide (2 mg/kg) or hy-drochlorothiazide (25 mg/kg), the increase in diameter was confined to the distal tubules.
We have also demonstrated ample fluorescence of Cy5 at 650–670 nm, both using the free dye and coupled pep-tides to study their uptake, but we require a greater range of affordable far-red fluorescing materials. However, in the interim, we are able to switch to the argon ion laser to study a larger series of shorter-wavelength labelled sub-stances.
Studying the living tissues with video-rate scanning con-focal microscopy abolishes the fixation, embedding,
sec-tioning, and staining artefacts of conventional histology.
Making comparison with the use of the 488 nm laser light, the epithelial proximal cells are more sharply defined in video-rate confocal reflection scanning using the red laser, and this, therefore, represents a significant advance in phys-iologic imaging. The approach holds promise for future in-vestigations of dynamic tubular morphology and function in vivo.
Acknowledgments: The authors thank The Royal Society, The National Kidney Research Fund and St Peter’s Trust for Kidney, Bladder and Prostate Research for financial support.
Limits to the precision of optical sectioning in live-cell confocal microscopy
J. PAWLEY
Zoology Department, University of Wisconsin-Madison, Madison, WI, USA
Confocal microscopy is inextricably linked to the con-cept of the optical section. Optical sectioning occurs be-cause light originating from out-of-focus planes is ex-cluded by the pinhole diaphragm in front of the photodetector. It is so firmly embedded in the psyche of modern biological microscopy that few stop to ask whether the “plane-of-focus” is an actual geometrical plane like a mechanical section, or merely “the surface described by the array of points at which the laser beam reaches best focus.”
In fact, this best focus plane will be something close to a geometrical plane only if the specimen is optically homo-geneous (i.e., has the same refractive index: R.I.). For spec-imens such as embedded cells, this is a fair approximation.
However, as more and more studies are performed on liv-ing cells, it is becomliv-ing clear that few livliv-ing specimens are optically homogeneous. This should not come as a surprise:
we have been viewing living cells using phase contrast and DIC for years. The contrast in these images reflects RI variations within cells.
The present work gives some indication of how serious this problem can be. Cheek cells were prepared fresh, stained with Acridine Orange and viewed in a 70 µm high chamber made of a coverslip separated from the slide by 4 dots of dried nail polish. This specimen was then viewed in the XZ plane using both fluorescent and reflected or backscattered light (BSL) (Fig 1). There is some loss of sig-nal with depth, caused primarily by the increasing effects of spherical aberration, but general features of the cells can be seen in both fluorescent and reflected light: partic-ularly nuclei and features in the cortex that outline the cell margin. However, the most noticeable feature of this image pair is the tremendous difference in appearance between the refection image of the near and far glass-water interfaces.
The image of the near side at the left approximates a
straight line. (The image is saturated because the surface signal is high compared to that from cellular components).
The image of the far side is not a straight line. In some cases, deep “holes” in it can be correlated with the presence of overlying structures such as a nucleus (Fig. 2). Because we know that the surface of the side is essentially flat, the image of this surface should be uniformly bright. The fact that it isn’t is an indication that the optical section is not flat, particularly under refractile features such as nuclei. Mea-surements from Fig. 1, suggest that the presence of the nu-cleus has displaced the “optical section” downward by about ~6 µm for all planes below it. Perhaps this explains the common observation that one seldom records fluores-cent features on the far side of a nucleus.
FIG. 1 Two XZ images of a fresh cheek cell made simultaneously using reflected light (left) and fluorescent light from Acridine Orange (right). In the left image, notice the difference in appearance be-tween the refection image of the near and far glass/water interfaces.
The image of the near side (left) approximates a straight line, that from the far side (right) shows distortion caused by the optical properties of the overlying nucleus, (the large blob in the fluorescence image).
FIG. 2 A three-dimensional projection of a reflected light image of a sample similar to that shown in Fig. 1, but oriented to show the distor-tion of what should be a uniform representadistor-tion of the lower glass/water interface in relation to the overlying cell nucleus causing it.
As living cells necessarily have nuclei and other refrac-tive organelles, it is not clear how we can avoid the se-quellae of their optical properties. In some cells these fea-tures are naturally less optically disruptive: cornea and lens come to mind. For other tissues, it would be well to keep in mind that the visibility of peri-nuclear structures may depend strongly on whether or not they are located on the side nearest to the objective. When looking far below the surface of a tissue, it maybe worth the effort to try to arrange that the foreground contains as few nuclei as pos-sible.