- Email: [email protected]
In vivo FM: using conventional fluorescence microscopy to monitor retinal neuronal death in vivo
Research Update
spines depend on their dendritic location. Neuron 33, 425–437 21 Petrozzino, J.J. et al.(1995) Micromolar Ca2+ transients in dendritic spines of hippocampal pyramidal neurons in brain slices. Neuron 14, 1223–1231 22 Markram, H. et al. (1998) Competitive calcium
TRENDS in Neurosciences Vol.25 No.9 September 2002
binding: implications for dendritic signaling. J. Comput. Neurosci. 5, 331–348 23 Matsuzaki, M. et al. (2001) Dendritic spine geometry is critical for AMPA receptor expression in hippocampal CA1 pyramidal neurons. Nat. Neurosci. 4, 1086–1092
441
Fritjof Helmchen Abteilung Zellphysiologie, Max-PlanckInstitute für medizinische Forschung, Jahnstr. 29, 69120 Heidelberg, Germany. *e-mail: [email protected]
Response: Raising the speed limit Knut Holthoff, David Tsay, Ania Majewska and Rafael Yuste We agree with the criticisms raised by Dr Helmchen and with his balanced and critical review of the topic. The problem he raises about our assumed slow diffusion constant of Calcium Green-1, a number that came out of our fluorescence-recovery after photobleaching (FRAP) measurements [1,2], is important and has also worried us. We would also like to praise the detailed measurements of Sabatini et al. and their ingenious use of the variance method to estimate diffusional equilibration [3]. At the same time, we would like to remark that the overall difference between our interpretation and that of Sabatini et al. is not as large as it seems – both papers essentially agree on the importance of Ca2+ pumps in controlling spine Ca2+ dynamics. Indeed, our kinetic predictions for the zero-exogenous-buffer condition match well with those of Sabatini et al., and we even explicitly propose that the pumps will be stronger than diffusional pathways in this case. In fact, the agreement between these two sets of results, using very different methods and analyses, is remarkable. Ca2+ extrusion by spine pumps, first described by Majewska et al. [1], appears in our data to vary among different spines. Based on the large heterogeneity in spine
structure and function, at this point we are cautious not to extrapolate to a simple scenario for all spines. Indeed, one of the major lessons learned from our past work is that of spine heterogeneity. As Dr Helmchen remarks, the importance of diffusional pathways is likely to be major in stubby spines (those with no neck), and we provide direct evidence for this in our Fig. 5e. Also, long duration Ca2+ accumulations produced by trains of action potentials, excitatory postsynaptic potentials (EPSPs), or the long clearing times present in thick dendrites, could allow diffusional pathways to shape considerably Ca2+ kinetics in spines. More measurements under physiological conditions are clearly necessary to understand the roles of Ca2+ extrusion and diffusion in vivo. Finally, we would also like to point out a few typographic errors present in our manuscript (although not in the code for the model and, therefore, not affecting our conclusions). Specifically, the extrusion rate coefficients used in the model were 30 µl/(cm2*s) and 7.5 µl/(cm2*s) for dendrite and spine, respectively. Also, the factor of (1 + Km + Kf ) should be substituted by 1 in Equation 12, and D should be marked Deff in Equation 21.
References 1 Majewska, A. et al. (2000) Mechanisms of calcium decay kinetics in hippocampal spines: role of calcium pumps and calcium diffusion through the spine neck in biochemical compartmentalization. J. Neurosci. 20, 1722–1734 2 Holthoff, K. et al. (2002) Calcium dynamics of spines depend on their dendritic location. Neuron 33, 425–437 3 Sabatini, B.L. et al. (2002) The life-cycle of Ca2+ ions in dendritic spines. Neuron 33, 439–452
Knut Holthoff* Physiologisches Institut, LudwigMaximilians-Universität München, Pettenkoferstr. 12, 80336 München, Germany. *e-mail: [email protected] David Tsay Dept of Biological Sciences, Columbia University, 1212 Amsterdam Avenue, New York, NY 10027, USA. Ania Majewska Massachusetts Institute of Technology, 45 Carleton Street, E25–235, Cambridge, MA 02139, USA. Rafael Yuste Dept of Biological Sciences, Columbia University, 1212 Amsterdam Avenue, Box 2435, New York, NY 10027, USA.
Techniques & Applications
In vivo FM: using conventional fluorescence microscopy to monitor retinal neuronal death in vivo Solon Thanos, Lars Indorf and Rita Naskar Post-traumatic death of mature retinal neurons occurs in glaucoma and after optic nerve injury. The death is a dynamic process that can be fully analyzed with methods that monitor changes over time. We have coupled the development of retrogradely transportable fluorescent dyes with modification of conventional http://tins.trends.com
epifluorescence microscopy to manipulate and visualize rat retinal neurons in vivo. The method is a relatively new concept and has potential for the monitoring of retinal conditions, such as glaucoma or optic nerve transection, and for evaluation of neuroprotective strategies in the near future.
The challenge of studying neuronal development, re-modeling, cell migration or process outgrowth within the natural habitat of neurons has led to the introduction of in vivo imaging techniques. Until the 1980s, vital dyes for in vivo studies were limited to conventional dyes, such as methylene
0166-2236/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved.
442
Research Update
Box 1. Modifying a fluorescence microscope (FM) for in vivo imaging in the retina • Use any conventional FM with epifluorescence (a 50–100 HBO mercury bulb is sufficient) and remove the microscopic stage. • Fix a non-flexible contact lens to the microscope lens of choice (2.5× to 10×) (Fig. 1b). • Create a head-holder for small animals (rats, mice), so that the head can be fixed by the front teeth (Fig. 1a). • Anesthetize an animal in which neurons have been prelabeled (any anesthetic is suitable). • Place the animal on a platform that can be raised, which replaces the microscope stage. • Keep the cornea moist and aplanate with a glass coverslip. • Perform in vivo FM imaging, and record results by video or photography. • Either keep the animal under anesthesia for long-term imaging, or reanesthetize to repeat imaging.
TRENDS in Neurosciences Vol.25 No.9 September 2002
(a)
(c)
Photographic camera
Eye of observer
Fluorescence source
(b)
Microscope lens
(d) I
blue, that were injected into the circulation; only sporadic studies were performed to visualize neurons [1]. The advent of new staining techniques, in particular with vital fluorescent dyes emitting at different spectral wavelengths, provided novel opportunities for application [2,3]. Fluorescent dyes can either be transported within axons and dendrites [4–6] or injected into the cell bodies of neurons or their precursors. In addition, they can be selectively taken up from the surroundings [2], or phagocytosed by macrophages [3] to label these cells in vivo. These features offer the opportunity of adopting dynamic strategies, in addition to the classical static studies with sections and whole-mounts. Consequently, since the 1980s numerous dynamic techniques have emerged in the field of in vivo imaging. These approaches were developed in the more-accessible peripheral nerves of anesthetized rodents, allowing neurons to be monitored with an image-intensifying camera over time [4,5]. They provided crucial information on various aspects of the neuromuscular terminals, sensory endings, ganglionic synapses and glial cells [2,7–9]. In parallel studies, growth cones were labeled in lower vertebrates, for following axonal guidance [7–11], observing single growth cone movements http://tins.trends.com
Contact lens (–3 to –8 diopters) Applanation of cornea
II
Rat eye Retina TRENDS in Neurosciences
Fig. 1. Experimental set up for in vivo fluorescence microscopy. (a) Anesthetized adult rats received local injections of the fluorescent dye 4Di-10ASP into the left superior colliculus (SC) to retrogradely label those retinal ganglion cells (RGCs) projecting into a particular area according to their retinotopic position. The SC was exposed by established microsurgical methods. Re-anesthetized rats were treated with topically applied mydriatic drops (Neosynephrin-Pos 5%, Ursapharm) and placed on a three-dimensionally adjustable microscope holder (the original microscope had been modified, with its own object-table and the corresponding condensors having been removed). (b) A non-flexible contact lens of –3.0 to –8.0 diopters was directly attached and fixed below the microscope lens (arrow) to prolong its focal length, enabling a fundoscopic view of the labeled retina. (c) Optics of the modified microscope: In addition to the contact lens, a planar glass slide was gently placed on the cornea enlarging the retinal field. Various microscope lenses can be used to detect RGCs in vivo. However, the most suitable lens was the 5× lens, which results in a final magnification of 50× to 100× and allows identification of individual cells. (d) Ganglion cells can be directly observed and photographed. The flexible apposition of the glass slide on the cornea, in conjunction with the three-dimensionally moveable head-holder, permits selection of the area to be observed by using blood vessels (arrows) and the papilla nervi optici (p) as landmarks. The device also allows repeated observations of exactly the same areas and cells over several days and weeks. Scale bar, 50 µm.
at a high-resolution and determining patterns of terminal arborization during development. The variety of such dynamic approaches has been substantially enriched by time-lapse imaging using image-intensifying cameras, or simply by repeatedly observing the same fiber terminals after anterograde staining using laser scanning confocal microscopes [2]. Analysis has now moved from peripheral nerves to the developing and regenerating visual system in different species [10,12], migrating neural crest
cells [13–15], transgenic and GFP-transfected cells [16] and functional re-modeling of dendritic terminals in various cortical areas [17]. Other uses include Ca2+-sensitive imaging [18], with applications to diseases such as epilepsy, to re-innervations and to pharmacological treatments [2]. Non-toxic carbocyanine dyes with different spectral features [2,3,14,19], Ca2+-sensitive probes [7,8] and derivatives of rhodamine are also used, according to the scope of the particular experiment [2].
Research Update
TRENDS in Neurosciences Vol.25 No.9 September 2002
The retinal ganglion cell: a representative of the CNS
Besides enabling the visualization of naturally occurring developmental or plastic events, in vivo imaging is suitable for studying experimentally induced perturbations within subsets of CNS neurons. For example, neuronal cell death is not considered to be a physiological event in the mature brain, unless external trauma or pathological circumstances such as stroke, ischemia, infections, aging or toxic agents force the neurons to die. Isolated, mechanically induced axotomy close to or distant from the cell body results in immediate interruption of axonal flow, disconnects the perikaryon from postsynaptic neurons, and probably deprives the neuron of target-derived neurotrophic factors. Many of the observations on the fate of cells after such traumatic events have been obtained using the paradigm of mature retinal ganglion cells (RGCs) [20–22] because the retina is ontogenetically part of the CNS, the anatomy and physiology of which have been well characterized. RGCs can be labeled retrogradely (i.e. from their central target areas) with a variety of fluorescent dyes and the time course of cell degeneration can be monitored in whole-mount preparations, in tissue sections or, as has been recently demonstrated in vivo, by using confocal neuroimaging laser microscopy [23–25]. This technique provided the possibility of following specific changes in axotomized cells (e.g. pre-apoptotic swelling [25]) but its availability remains limited to only a few researchers. What are the limits of conventional in vivo FM?
Our aim was to establish a convenient method of examining pre-labeled retinal neurons in the living animal and we modified a conventional Zeiss fluorescence microscope (FM) for this purpose (Box 1). The modified microscope, termed ‘in vivo FM’ (Fig. 1) consists of a non-flexible contact lens of −3.0 to −8.0 diopters (available from commercial suppliers) attached to the lens of a conventional FM (Fig. 1b,c). This attachment alters the focal length of the microscope lens, allowing large fields of the retina to be visualized at final magnifications of 50× to 100×, which are sufficient for individual RGCs of various http://tins.trends.com
Injection of fluorescent dye into the superior colliculus
5–8 days
443
Retrograde staining of ganglion cells
SC
Manipulation of cells (e.g. axon cut, glaucoma)
(a)
In vivo imaging over hours
Days to months
(b)
(c)
Return animal to cage for repeating imaging
(d)
Long-term imaging and histology
(e)
(f) RGC
MG TRENDS in Neurosciences
Fig. 2. Experimental steps towards visualization of retinal ganglion cells (RGCs) in vivo. The text boxes explain the intervention, and the photographs depict the images obtained with the in vivo FM. The black arrows in the photomicrographs (a–c) show an axotomized RGC disappearing within 30 min of observation. The black arrows in (d,e) show a cell that disappeared a few days after optic nerve injury. (f) shows the resolution that can be achieved; a ‘dying’ RGC and microglial cell (MG) are depicted.
sizes to be distinguished from one another clearly (Fig. 1d). The device also includes a head-holder, which can be rotated along three axes, making it possible to examine the entire retinal surface. To obtain planar images, it is crucial to place a glass slide on the cornea (applanation; Fig. 1a,c). The original camera and video attachments remained unaltered. Rats or other animals of similar size (e.g. mice, hamsters or chickens) are equally suitable for the imaging. The fluorescent dye 4Di-10Asp (4-[4-didecylaminostyryl]N-methyl-pyridinium iodide) (Molecular Probes, Eugene, OR, USA), or any other carbocyanine dye or rhodamine derivative, can be used to label
retrogradely small groups of RGCs within the retina (Fig. 1d) [22]. The number of cells labeled depends on the scope of the subsequent experiment and can range from few cells to the total RGC population. Repeated relocation of the cells is easy, by using blood vessels as landmarks (Fig. 1d), and/or by adjusting the head-holder using the same coordinates. This approach has allowed the disappearance of individual neurons, for example after either cut of the optic nerve or induction of glaucoma (which also results in RGC loss [22]), to be monitored (Fig. 2). Several aspects of degeneration (e.g. speed of death and
444
Research Update
location of the dying cells with respect to other cells, such as microglia) can be addressed [22]. In addition, the same tissue can be processed further for histo-morphological examination following conclusion of the in vivo imaging sessions. Comparison with similar techniques
This convenient and relatively simple in vivo FM provides a non-invasive procedure for assessing the fate of individual retinal neurons (which lie at a considerable depth in mammals and are not accessible using conventional microscopy techniques, especially in the living animal). The method is inexpensive compared to optical imaging using Ca2+-sensitive dyes [7,18] and two-photon-laser microscopy [26], is applicable to all fluorescence microscopes irrespective of make, and only involves minor modifications of an existing set up. Fluorescent dyes with different excitation and emission properties can be used in similar ways, providing they are conducive to retrograde axonal transport. The resolution obtained is similar to that obtained using the in vivo confocal (ICON) set up [23–25], but lower than that achieved with histological sections or retinal whole mounts. However, the relative advantage over ICON is that the observer can directly visualize and evaluate the cells of interest, instead of having to use a photo-detector, a confocal pinhole and scanning reflectors [23–25]. This practicability makes possible the fast and reproducible analysis of identical cells, and also allows examination over several days to months after transfection (using genes encoding fluorescent proteins [11,16] or trangenic animals [16]). The spherical shape of the mammalian eye and its size of several millimeters can be overcome by fixing contact lenses to the microscope objective. The problem of the spherical eye can also be circumvented by atraumatically applanating the cornea with a glass cover slip and by using the flexible head-holder, so that the head of the animal can be rotated as desired along different axes. Future perspectives
The technique described in this paper has implications for several studies associated with neuronal fate within the http://tins.trends.com
TRENDS in Neurosciences Vol.25 No.9 September 2002
retina, especially when attempting to unravel the mechanisms involved in neuronal degeneration or prevent cell death by pharmacological interventions and neuroprotective strategies. Animal models of ophthalmic diseases, such as glaucoma, optic neuritis, ischemia or photoreceptor dystrophy, can also be examined. Finally, because the eye (and especially the retina) has emerged as a classical and widely used neuronal tissue in which to study modern methods of creating fluorescent transgenic animals, this simple method should allow the large-scale screening of transgenic offspring within a reasonable time. Acknowledgements
We thank Ilka Romann for skilful technical assistance, Magdalena Pinheiro for typing the manuscript and Marliese Wagner and Susanne von der Heide for preparing the photographs. The advice of P. Heiduschka and M. Pavlidis with the processing of photographs is acknowledged. The work was supported by ‘The Centre for Interdisciplinary Clinical Research’ (IZKF, project E3) and by the BMBF (Neurotraumatology, Project II). References 1 Holländer, H. (1966) Method for microscopic observation of a single motor nerve fiber in the living frog. Z. Wiss. Mikrosk. 67, 156–170 2 Lichtman, J.W. and Fraser, S.C. (2001) The neuronal naturalist: watching neurons in their native habitat. Nat. Neurosci. 4, 1215–1220 3 Thanos, S. et al. (1994) Old dyes for new scopes. The phagocytosis-dependent long-term fluorescence labelling of microglial cells in vivo. Trends Neurosci. 17, 177–182 4 Purves, D. and Hadley, R.D. (1985) Changes in the dendritic branching of adult mammalian neurones revealed by repeated imaging in situ. Nature 315, 404–406 5 Purves, D. and Voyvodic, J.T. (1987) Imaging mammalian nerve cells and their connections over time in living animals. Trends Neurosci. 10, 398–404 6 Purves, D. and Lichtman, J.W. (1987) Synaptic sites on reinnervated nerve cells visualized at two different times in living mice. J. Neurosci. 7, 1492–1497 7 Meister, M. and Bonhoeffer, T. (2000) Tuning and topography in an odor map on the rat olfactory bulb. J. Neurosci. 21, 1351–1360 8 Lichtman, J.W. et al. (1985) Multiple innervation of tonic endplates revealed by activity-dependent uptake of fluorescent probes. Nature 314, 357–359 9 Cohen, C.S. and Fraser, S.E. (1995) Effects of brain-derived neurotrophic factor on optic axon branching and remodelling in vivo. Nature 378, 192–196
10 Jontes, J.D. et al. (2000) Growth cone and dendritic dynamics in zebrafish embryos: early events in synaptogenesis imaged in vivo. Nat. Neurosci. 3, 231–237 11 Moritz, O.L. et al. (1999) Fluorescent photoreceptors of transgenic Xenopus laevis imaged in vivo by two fluorescent techniques. Invest. Ophthalmol. Vis. Sci. 40, 3276–3280 12 Dawson, A.J. and Meyer, R.L. (2001) Regenerating optic fibers correct large-scale errors by random growth: evidence from in vivo imaging. J. Comp. Neurol. 434, 40–55 13 Bronner-Fraser, M. and Fraser, S.E. (1988) Cell lineage analysis reveals multipotency of some avian neural crest cells. Nature 335, 161–164 14 Fishell, G. et al. (1995) Optical microscopy. 3. Tracking fluorescently labelled neurones in developing brain. FASEB J. 9, 324–334 15 Frank, E. and Sanes, J.R. (1991) Lineage of neurons and glia in chick dorsal root ganglia: analysis in vivo with a recombinant retrovirus. Development 111, 895–908 16 Feng, G. et al. (2000) Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP. Neuron 28, 41–45 17 Lendai, B. et al. (2000) Experiencedependent plasticity of dendritic spines in the developing rat barrel cortex in vivo. Nature 404, 876–881 18 Grinvald, A. et al. (1986) Functional architecture of cortex revealed by optical imaging of intrinsic signals. Nature 324, 361–364 19 Cullander, C. (1999) Fluorescent probes for confocal microscopy. Methods Mol. Biol. 122, 59–73 20 Berkelaar, M. et al. (1994) Axotomy results in delayed death and apoptosis of retinal ganglion cells in adult rats. J. Neurosci. 14, 4368–4374 21 Garcia-Valenzuela, E. et al. (1993) Apoptosis in adult retinal ganglion cells after axotomy. J. Neurobiol. 24, 431–438 22 Naskar, R. et al. Detection of early neuronal degeneration and accompanying microglial responses in the retina of a rat model of glaucoma Invest. Ophthalmol. Vis. Sci. (in press) 23 Sabel, B.A. et al. (1997) In vivo confocal neuroimaging (ICON) of CNS neurones. Nat. Med. 3, 244–247 24 Engelmann, R. and Sabel, B.A. (1999) In vivo imaging of mammalian central nervous system neurones with the in vivo confocal neuroimaging (ICON) method. Methods Enzymol. 307, 563–570 25 Rousseau, V. et al. (1999) Restoration of vision III: soma swelling dynamics predicts neuronal death or survival after optic nerve crush in vivo. Neuroreport 10, 3387–3391 26 Denk, W. et al. (1990) Two-photon laser scanning fluorescence microscopy. Science 248, 73–76
Solon Thanos* Lars Indorf Rita Naskar Dept of Experimental Ophthalmology, University Eye Hospital Münster, Domagkstraße 15, D-48149 Münster, Germany. *e-mail: [email protected]
spines depend on their dendritic location. Neuron 33, 425–437 21 Petrozzino, J.J. et al.(1995) Micromolar Ca2+ transients in dendritic spines of hippocampal pyramidal neurons in brain slices. Neuron 14, 1223–1231 22 Markram, H. et al. (1998) Competitive calcium
TRENDS in Neurosciences Vol.25 No.9 September 2002
binding: implications for dendritic signaling. J. Comput. Neurosci. 5, 331–348 23 Matsuzaki, M. et al. (2001) Dendritic spine geometry is critical for AMPA receptor expression in hippocampal CA1 pyramidal neurons. Nat. Neurosci. 4, 1086–1092
441
Fritjof Helmchen Abteilung Zellphysiologie, Max-PlanckInstitute für medizinische Forschung, Jahnstr. 29, 69120 Heidelberg, Germany. *e-mail: [email protected]
Response: Raising the speed limit Knut Holthoff, David Tsay, Ania Majewska and Rafael Yuste We agree with the criticisms raised by Dr Helmchen and with his balanced and critical review of the topic. The problem he raises about our assumed slow diffusion constant of Calcium Green-1, a number that came out of our fluorescence-recovery after photobleaching (FRAP) measurements [1,2], is important and has also worried us. We would also like to praise the detailed measurements of Sabatini et al. and their ingenious use of the variance method to estimate diffusional equilibration [3]. At the same time, we would like to remark that the overall difference between our interpretation and that of Sabatini et al. is not as large as it seems – both papers essentially agree on the importance of Ca2+ pumps in controlling spine Ca2+ dynamics. Indeed, our kinetic predictions for the zero-exogenous-buffer condition match well with those of Sabatini et al., and we even explicitly propose that the pumps will be stronger than diffusional pathways in this case. In fact, the agreement between these two sets of results, using very different methods and analyses, is remarkable. Ca2+ extrusion by spine pumps, first described by Majewska et al. [1], appears in our data to vary among different spines. Based on the large heterogeneity in spine
structure and function, at this point we are cautious not to extrapolate to a simple scenario for all spines. Indeed, one of the major lessons learned from our past work is that of spine heterogeneity. As Dr Helmchen remarks, the importance of diffusional pathways is likely to be major in stubby spines (those with no neck), and we provide direct evidence for this in our Fig. 5e. Also, long duration Ca2+ accumulations produced by trains of action potentials, excitatory postsynaptic potentials (EPSPs), or the long clearing times present in thick dendrites, could allow diffusional pathways to shape considerably Ca2+ kinetics in spines. More measurements under physiological conditions are clearly necessary to understand the roles of Ca2+ extrusion and diffusion in vivo. Finally, we would also like to point out a few typographic errors present in our manuscript (although not in the code for the model and, therefore, not affecting our conclusions). Specifically, the extrusion rate coefficients used in the model were 30 µl/(cm2*s) and 7.5 µl/(cm2*s) for dendrite and spine, respectively. Also, the factor of (1 + Km + Kf ) should be substituted by 1 in Equation 12, and D should be marked Deff in Equation 21.
References 1 Majewska, A. et al. (2000) Mechanisms of calcium decay kinetics in hippocampal spines: role of calcium pumps and calcium diffusion through the spine neck in biochemical compartmentalization. J. Neurosci. 20, 1722–1734 2 Holthoff, K. et al. (2002) Calcium dynamics of spines depend on their dendritic location. Neuron 33, 425–437 3 Sabatini, B.L. et al. (2002) The life-cycle of Ca2+ ions in dendritic spines. Neuron 33, 439–452
Knut Holthoff* Physiologisches Institut, LudwigMaximilians-Universität München, Pettenkoferstr. 12, 80336 München, Germany. *e-mail: [email protected] David Tsay Dept of Biological Sciences, Columbia University, 1212 Amsterdam Avenue, New York, NY 10027, USA. Ania Majewska Massachusetts Institute of Technology, 45 Carleton Street, E25–235, Cambridge, MA 02139, USA. Rafael Yuste Dept of Biological Sciences, Columbia University, 1212 Amsterdam Avenue, Box 2435, New York, NY 10027, USA.
Techniques & Applications
In vivo FM: using conventional fluorescence microscopy to monitor retinal neuronal death in vivo Solon Thanos, Lars Indorf and Rita Naskar Post-traumatic death of mature retinal neurons occurs in glaucoma and after optic nerve injury. The death is a dynamic process that can be fully analyzed with methods that monitor changes over time. We have coupled the development of retrogradely transportable fluorescent dyes with modification of conventional http://tins.trends.com
epifluorescence microscopy to manipulate and visualize rat retinal neurons in vivo. The method is a relatively new concept and has potential for the monitoring of retinal conditions, such as glaucoma or optic nerve transection, and for evaluation of neuroprotective strategies in the near future.
The challenge of studying neuronal development, re-modeling, cell migration or process outgrowth within the natural habitat of neurons has led to the introduction of in vivo imaging techniques. Until the 1980s, vital dyes for in vivo studies were limited to conventional dyes, such as methylene
0166-2236/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved.
442
Research Update
Box 1. Modifying a fluorescence microscope (FM) for in vivo imaging in the retina • Use any conventional FM with epifluorescence (a 50–100 HBO mercury bulb is sufficient) and remove the microscopic stage. • Fix a non-flexible contact lens to the microscope lens of choice (2.5× to 10×) (Fig. 1b). • Create a head-holder for small animals (rats, mice), so that the head can be fixed by the front teeth (Fig. 1a). • Anesthetize an animal in which neurons have been prelabeled (any anesthetic is suitable). • Place the animal on a platform that can be raised, which replaces the microscope stage. • Keep the cornea moist and aplanate with a glass coverslip. • Perform in vivo FM imaging, and record results by video or photography. • Either keep the animal under anesthesia for long-term imaging, or reanesthetize to repeat imaging.
TRENDS in Neurosciences Vol.25 No.9 September 2002
(a)
(c)
Photographic camera
Eye of observer
Fluorescence source
(b)
Microscope lens
(d) I
blue, that were injected into the circulation; only sporadic studies were performed to visualize neurons [1]. The advent of new staining techniques, in particular with vital fluorescent dyes emitting at different spectral wavelengths, provided novel opportunities for application [2,3]. Fluorescent dyes can either be transported within axons and dendrites [4–6] or injected into the cell bodies of neurons or their precursors. In addition, they can be selectively taken up from the surroundings [2], or phagocytosed by macrophages [3] to label these cells in vivo. These features offer the opportunity of adopting dynamic strategies, in addition to the classical static studies with sections and whole-mounts. Consequently, since the 1980s numerous dynamic techniques have emerged in the field of in vivo imaging. These approaches were developed in the more-accessible peripheral nerves of anesthetized rodents, allowing neurons to be monitored with an image-intensifying camera over time [4,5]. They provided crucial information on various aspects of the neuromuscular terminals, sensory endings, ganglionic synapses and glial cells [2,7–9]. In parallel studies, growth cones were labeled in lower vertebrates, for following axonal guidance [7–11], observing single growth cone movements http://tins.trends.com
Contact lens (–3 to –8 diopters) Applanation of cornea
II
Rat eye Retina TRENDS in Neurosciences
Fig. 1. Experimental set up for in vivo fluorescence microscopy. (a) Anesthetized adult rats received local injections of the fluorescent dye 4Di-10ASP into the left superior colliculus (SC) to retrogradely label those retinal ganglion cells (RGCs) projecting into a particular area according to their retinotopic position. The SC was exposed by established microsurgical methods. Re-anesthetized rats were treated with topically applied mydriatic drops (Neosynephrin-Pos 5%, Ursapharm) and placed on a three-dimensionally adjustable microscope holder (the original microscope had been modified, with its own object-table and the corresponding condensors having been removed). (b) A non-flexible contact lens of –3.0 to –8.0 diopters was directly attached and fixed below the microscope lens (arrow) to prolong its focal length, enabling a fundoscopic view of the labeled retina. (c) Optics of the modified microscope: In addition to the contact lens, a planar glass slide was gently placed on the cornea enlarging the retinal field. Various microscope lenses can be used to detect RGCs in vivo. However, the most suitable lens was the 5× lens, which results in a final magnification of 50× to 100× and allows identification of individual cells. (d) Ganglion cells can be directly observed and photographed. The flexible apposition of the glass slide on the cornea, in conjunction with the three-dimensionally moveable head-holder, permits selection of the area to be observed by using blood vessels (arrows) and the papilla nervi optici (p) as landmarks. The device also allows repeated observations of exactly the same areas and cells over several days and weeks. Scale bar, 50 µm.
at a high-resolution and determining patterns of terminal arborization during development. The variety of such dynamic approaches has been substantially enriched by time-lapse imaging using image-intensifying cameras, or simply by repeatedly observing the same fiber terminals after anterograde staining using laser scanning confocal microscopes [2]. Analysis has now moved from peripheral nerves to the developing and regenerating visual system in different species [10,12], migrating neural crest
cells [13–15], transgenic and GFP-transfected cells [16] and functional re-modeling of dendritic terminals in various cortical areas [17]. Other uses include Ca2+-sensitive imaging [18], with applications to diseases such as epilepsy, to re-innervations and to pharmacological treatments [2]. Non-toxic carbocyanine dyes with different spectral features [2,3,14,19], Ca2+-sensitive probes [7,8] and derivatives of rhodamine are also used, according to the scope of the particular experiment [2].
Research Update
TRENDS in Neurosciences Vol.25 No.9 September 2002
The retinal ganglion cell: a representative of the CNS
Besides enabling the visualization of naturally occurring developmental or plastic events, in vivo imaging is suitable for studying experimentally induced perturbations within subsets of CNS neurons. For example, neuronal cell death is not considered to be a physiological event in the mature brain, unless external trauma or pathological circumstances such as stroke, ischemia, infections, aging or toxic agents force the neurons to die. Isolated, mechanically induced axotomy close to or distant from the cell body results in immediate interruption of axonal flow, disconnects the perikaryon from postsynaptic neurons, and probably deprives the neuron of target-derived neurotrophic factors. Many of the observations on the fate of cells after such traumatic events have been obtained using the paradigm of mature retinal ganglion cells (RGCs) [20–22] because the retina is ontogenetically part of the CNS, the anatomy and physiology of which have been well characterized. RGCs can be labeled retrogradely (i.e. from their central target areas) with a variety of fluorescent dyes and the time course of cell degeneration can be monitored in whole-mount preparations, in tissue sections or, as has been recently demonstrated in vivo, by using confocal neuroimaging laser microscopy [23–25]. This technique provided the possibility of following specific changes in axotomized cells (e.g. pre-apoptotic swelling [25]) but its availability remains limited to only a few researchers. What are the limits of conventional in vivo FM?
Our aim was to establish a convenient method of examining pre-labeled retinal neurons in the living animal and we modified a conventional Zeiss fluorescence microscope (FM) for this purpose (Box 1). The modified microscope, termed ‘in vivo FM’ (Fig. 1) consists of a non-flexible contact lens of −3.0 to −8.0 diopters (available from commercial suppliers) attached to the lens of a conventional FM (Fig. 1b,c). This attachment alters the focal length of the microscope lens, allowing large fields of the retina to be visualized at final magnifications of 50× to 100×, which are sufficient for individual RGCs of various http://tins.trends.com
Injection of fluorescent dye into the superior colliculus
5–8 days
443
Retrograde staining of ganglion cells
SC
Manipulation of cells (e.g. axon cut, glaucoma)
(a)
In vivo imaging over hours
Days to months
(b)
(c)
Return animal to cage for repeating imaging
(d)
Long-term imaging and histology
(e)
(f) RGC
MG TRENDS in Neurosciences
Fig. 2. Experimental steps towards visualization of retinal ganglion cells (RGCs) in vivo. The text boxes explain the intervention, and the photographs depict the images obtained with the in vivo FM. The black arrows in the photomicrographs (a–c) show an axotomized RGC disappearing within 30 min of observation. The black arrows in (d,e) show a cell that disappeared a few days after optic nerve injury. (f) shows the resolution that can be achieved; a ‘dying’ RGC and microglial cell (MG) are depicted.
sizes to be distinguished from one another clearly (Fig. 1d). The device also includes a head-holder, which can be rotated along three axes, making it possible to examine the entire retinal surface. To obtain planar images, it is crucial to place a glass slide on the cornea (applanation; Fig. 1a,c). The original camera and video attachments remained unaltered. Rats or other animals of similar size (e.g. mice, hamsters or chickens) are equally suitable for the imaging. The fluorescent dye 4Di-10Asp (4-[4-didecylaminostyryl]N-methyl-pyridinium iodide) (Molecular Probes, Eugene, OR, USA), or any other carbocyanine dye or rhodamine derivative, can be used to label
retrogradely small groups of RGCs within the retina (Fig. 1d) [22]. The number of cells labeled depends on the scope of the subsequent experiment and can range from few cells to the total RGC population. Repeated relocation of the cells is easy, by using blood vessels as landmarks (Fig. 1d), and/or by adjusting the head-holder using the same coordinates. This approach has allowed the disappearance of individual neurons, for example after either cut of the optic nerve or induction of glaucoma (which also results in RGC loss [22]), to be monitored (Fig. 2). Several aspects of degeneration (e.g. speed of death and
444
Research Update
location of the dying cells with respect to other cells, such as microglia) can be addressed [22]. In addition, the same tissue can be processed further for histo-morphological examination following conclusion of the in vivo imaging sessions. Comparison with similar techniques
This convenient and relatively simple in vivo FM provides a non-invasive procedure for assessing the fate of individual retinal neurons (which lie at a considerable depth in mammals and are not accessible using conventional microscopy techniques, especially in the living animal). The method is inexpensive compared to optical imaging using Ca2+-sensitive dyes [7,18] and two-photon-laser microscopy [26], is applicable to all fluorescence microscopes irrespective of make, and only involves minor modifications of an existing set up. Fluorescent dyes with different excitation and emission properties can be used in similar ways, providing they are conducive to retrograde axonal transport. The resolution obtained is similar to that obtained using the in vivo confocal (ICON) set up [23–25], but lower than that achieved with histological sections or retinal whole mounts. However, the relative advantage over ICON is that the observer can directly visualize and evaluate the cells of interest, instead of having to use a photo-detector, a confocal pinhole and scanning reflectors [23–25]. This practicability makes possible the fast and reproducible analysis of identical cells, and also allows examination over several days to months after transfection (using genes encoding fluorescent proteins [11,16] or trangenic animals [16]). The spherical shape of the mammalian eye and its size of several millimeters can be overcome by fixing contact lenses to the microscope objective. The problem of the spherical eye can also be circumvented by atraumatically applanating the cornea with a glass cover slip and by using the flexible head-holder, so that the head of the animal can be rotated as desired along different axes. Future perspectives
The technique described in this paper has implications for several studies associated with neuronal fate within the http://tins.trends.com
TRENDS in Neurosciences Vol.25 No.9 September 2002
retina, especially when attempting to unravel the mechanisms involved in neuronal degeneration or prevent cell death by pharmacological interventions and neuroprotective strategies. Animal models of ophthalmic diseases, such as glaucoma, optic neuritis, ischemia or photoreceptor dystrophy, can also be examined. Finally, because the eye (and especially the retina) has emerged as a classical and widely used neuronal tissue in which to study modern methods of creating fluorescent transgenic animals, this simple method should allow the large-scale screening of transgenic offspring within a reasonable time. Acknowledgements
We thank Ilka Romann for skilful technical assistance, Magdalena Pinheiro for typing the manuscript and Marliese Wagner and Susanne von der Heide for preparing the photographs. The advice of P. Heiduschka and M. Pavlidis with the processing of photographs is acknowledged. The work was supported by ‘The Centre for Interdisciplinary Clinical Research’ (IZKF, project E3) and by the BMBF (Neurotraumatology, Project II). References 1 Holländer, H. (1966) Method for microscopic observation of a single motor nerve fiber in the living frog. Z. Wiss. Mikrosk. 67, 156–170 2 Lichtman, J.W. and Fraser, S.C. (2001) The neuronal naturalist: watching neurons in their native habitat. Nat. Neurosci. 4, 1215–1220 3 Thanos, S. et al. (1994) Old dyes for new scopes. The phagocytosis-dependent long-term fluorescence labelling of microglial cells in vivo. Trends Neurosci. 17, 177–182 4 Purves, D. and Hadley, R.D. (1985) Changes in the dendritic branching of adult mammalian neurones revealed by repeated imaging in situ. Nature 315, 404–406 5 Purves, D. and Voyvodic, J.T. (1987) Imaging mammalian nerve cells and their connections over time in living animals. Trends Neurosci. 10, 398–404 6 Purves, D. and Lichtman, J.W. (1987) Synaptic sites on reinnervated nerve cells visualized at two different times in living mice. J. Neurosci. 7, 1492–1497 7 Meister, M. and Bonhoeffer, T. (2000) Tuning and topography in an odor map on the rat olfactory bulb. J. Neurosci. 21, 1351–1360 8 Lichtman, J.W. et al. (1985) Multiple innervation of tonic endplates revealed by activity-dependent uptake of fluorescent probes. Nature 314, 357–359 9 Cohen, C.S. and Fraser, S.E. (1995) Effects of brain-derived neurotrophic factor on optic axon branching and remodelling in vivo. Nature 378, 192–196
10 Jontes, J.D. et al. (2000) Growth cone and dendritic dynamics in zebrafish embryos: early events in synaptogenesis imaged in vivo. Nat. Neurosci. 3, 231–237 11 Moritz, O.L. et al. (1999) Fluorescent photoreceptors of transgenic Xenopus laevis imaged in vivo by two fluorescent techniques. Invest. Ophthalmol. Vis. Sci. 40, 3276–3280 12 Dawson, A.J. and Meyer, R.L. (2001) Regenerating optic fibers correct large-scale errors by random growth: evidence from in vivo imaging. J. Comp. Neurol. 434, 40–55 13 Bronner-Fraser, M. and Fraser, S.E. (1988) Cell lineage analysis reveals multipotency of some avian neural crest cells. Nature 335, 161–164 14 Fishell, G. et al. (1995) Optical microscopy. 3. Tracking fluorescently labelled neurones in developing brain. FASEB J. 9, 324–334 15 Frank, E. and Sanes, J.R. (1991) Lineage of neurons and glia in chick dorsal root ganglia: analysis in vivo with a recombinant retrovirus. Development 111, 895–908 16 Feng, G. et al. (2000) Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP. Neuron 28, 41–45 17 Lendai, B. et al. (2000) Experiencedependent plasticity of dendritic spines in the developing rat barrel cortex in vivo. Nature 404, 876–881 18 Grinvald, A. et al. (1986) Functional architecture of cortex revealed by optical imaging of intrinsic signals. Nature 324, 361–364 19 Cullander, C. (1999) Fluorescent probes for confocal microscopy. Methods Mol. Biol. 122, 59–73 20 Berkelaar, M. et al. (1994) Axotomy results in delayed death and apoptosis of retinal ganglion cells in adult rats. J. Neurosci. 14, 4368–4374 21 Garcia-Valenzuela, E. et al. (1993) Apoptosis in adult retinal ganglion cells after axotomy. J. Neurobiol. 24, 431–438 22 Naskar, R. et al. Detection of early neuronal degeneration and accompanying microglial responses in the retina of a rat model of glaucoma Invest. Ophthalmol. Vis. Sci. (in press) 23 Sabel, B.A. et al. (1997) In vivo confocal neuroimaging (ICON) of CNS neurones. Nat. Med. 3, 244–247 24 Engelmann, R. and Sabel, B.A. (1999) In vivo imaging of mammalian central nervous system neurones with the in vivo confocal neuroimaging (ICON) method. Methods Enzymol. 307, 563–570 25 Rousseau, V. et al. (1999) Restoration of vision III: soma swelling dynamics predicts neuronal death or survival after optic nerve crush in vivo. Neuroreport 10, 3387–3391 26 Denk, W. et al. (1990) Two-photon laser scanning fluorescence microscopy. Science 248, 73–76
Solon Thanos* Lars Indorf Rita Naskar Dept of Experimental Ophthalmology, University Eye Hospital Münster, Domagkstraße 15, D-48149 Münster, Germany. *e-mail: [email protected]