Jane,S.D.; Phenna,S. & Chad,J.E.*
*Author for correspondence and enquiries:-
Biosciences Imaging Group
Department of Physiology & Pharmacology
School of Biological Sciences
Biomedical Sciences Building
University of Southampton
Bassett Crescent East
Southampton SO16 7PX
UK
Tel: 01703 594292
Fax: 01703 594319
Email: jchad@uk.ac.soton.mail
WorldWideWeb: http:\\www.neuro.soton.ac.uk
Keywords:- Ratiometric UV Confocal Laser Scanning Microscopy, Calcium, Hippocampus, Pyramidal neuron preparation.
Summary:-
A method is described for increasing cell yields of dissociated hippocampal neurones from rat brain. Dye loading using AM ester techniques is assessed in relation to hippocampal neurones. UV confocal laser scanning microscopy of hippocampal neurones is examined in conjunction with the ratiometric calcium sensitive fluorescent probe Indo-1. The effects on the intracellular pH of neurones caused by adding drugs to the external bathing solution is tested and its subsequent effect on the accuracy of calcium ion concentration estimations using Indo-1 are examined using the ratiometric pH sensitive dye Snafl-calcein.
1.0 INTRODUCTION
Intracellular calcium ion activity (Ca++i) plays a major role in controlling cellular processes and the predicted (Chad et al. 1984; Simon & Llinas 1985) and inferred (Smith & Augustine 1988) spatial localisation of Ca++i transients responsible for this signalling, lead to the necessity to be able to measure the subcellular pattern of Ca++i changes. The confocal laser scanning microscope (CLSM; White et al. 1987) has permitted the imaging of thin optical sections of cells with greatly reduced out-of-focus contamination. However, until recently, the usual laser wavelengths used were in the visible region of the spectrum, requiring the use of dyes such as Fluo-3 (Minta et al. 1987;1989) for the detection of changes in Ca++i (Segal & Manor 1992). This method suffered from the inability to use ratiometric techniques (Tsien & Poenie 1986) in order to control for unequal dye distribution or bleaching. We have been using a modified Bio-Rad MRC-600 confocal imaging system equipped with an ultra-violet (UV) argon-ion laser in order to permit the ratiometric imaging of Ca++i using the fluorescent dye Indo-1 (Grynkiewicz et al. 1985). The Ca++i dependence of the emission spectrum for Indo-1 means that, with the use of a dichroic filter and two matched photomultiplier tubes, it is possible to record both high and low wavelength values for the ratio for any region simultaneously rather than having to swap excitation filters as for the ratiometric calcium sensitive probe Fura-2 (Grynkiewicz et al. 1985). Therefore, the system has the potential for submicron spatial resolution, and millisecond time resolution. Unfortunately, both can not be realised together, but trade-offs can be made to suit the experimental design.
Pyramidal neurones dissociated from rat hippocampus (Kay & Wong 1986; Chad et al. 1991) can generate voltage and drug gated responses under patch clamp conditions (Kay 1991; Stanford et al. 1992), and can be shown to be viable by their exclusion of propidium iodide and uptake of acetoxymethyl-esters of fluorescent dyes. Glutamate acting through the metabotropic receptor causes activation of G-proteins and the release of Ca++ from internal stores in these cells (Miller 1991), but the kinetics and spatial distribution of these changes have not been determined, or compared to the Ca++i transients produced by voltage gated calcium channels. Intracellular calcium stores can also be released by a number of pharmacological agents such as caffeine (eg. Jane et al. 1995) and thapsigargin (Thastrup et al. 1990). We have therefore attempted to develop a method for the determination of the three-dimensional (3D) distribution of changes in Ca++i during neuronal responses.
2.0 METHODS
2.1 Cell preparation:-
The methodology selected for the preparation of dissociated hippocampal neurones was dependent upon the subsequent experimental procedures they were to undergo. First our basic procedure for preparation of hippocampal neurones will be described then possible improvements identified which, although increasing cell numbers, could limit their eventual use.
The adopted protocol is basically that described by Chad et al.(1991). Briefly, 8-12 day old Wistar rats were decapitated and the brain rapidly removed to iced, oxygenated, artificial cerebrospinal fluid (aCSF: NaCl 118mM; NaHCO3 26mM; KCl 3mM; KH2PO4 1.25mM; MgSO4 1mM; CaCl2 2.5mM and glucose 10mM). When the tissue had cooled the hippocampi were removed using blunt dissection techniques. Transverse sections of approximately 450痠 were cut using a McIlwain tissue chopper. The hippocampal slices were maintained for a minimum of 1 hour at room temperature (18-20oC) in an interface chamber containing aCSF. Following this equilibration period hippocampal slices were incubated with 0.25mg/ml pronase E (Sigma type XIV) and 0.25mg/ml thermolysin (Sigma type X) for approximately 30mins in continuously oxygenated PIPES buffer (NaCl 120mM; KCl 5mM; MgCl2 1mM; CaCl2 5mM; PIPES 20mM; glucose 25mM and pH 7.0 with NaOH). Following incubation the slices were repeatedly washed in oxygenated PIPES buffer to remove the proteolytic enzymes. The final step of the cell preparation process was the mechanical disruption or trituration of the slices. This involved passing the slices through a series of pasteur pipette tips of decreasing diameter. The solution is passed through each tip approximately 20-30 times taking care to avoid producing bubbles. At the end of this procedure a milky solution of dissociated cells is produced. Centrifugation can be used at this stage to eliminate debris though we have not included this step in our protocol.
Figure 1
Assessing the viability of acutely dissociated hippocampal neurones with the
membrane impermeant fluorescent dye propidium iodide. Transmission (left) and
fluorescent image (right) pairs for two neurones. Upper pair: The neurone has
been intensely labelled by propidium iodide (10-15然: Molecular Probes)
particularly in the nucleus. This indicates that the plasma membrane has been
compromised and that propidium iodide has been able to bind and fluorescently
label (right hand image) the neurone's nucleic acids.
Lower Pair: 
A pyramidal
neurone clearly visible in the transmission (left hand) image shows no
fluorescent labelling (right hand image) despite exposure to propidium iodide,
indicating that the plasma membrane is intact and the neurone viable. Neuronal
cell bodies are approximately 10痠 in diameter.
2.2 Indo-1 loading:-
Dissociated neurones were loaded with dye by incubation with the acetoxymethyl-ester (AM) of Indo-1 (3然; Molecular Probes) in aCSF at pH 7.4. Initially 1mg/ml of detergent (Pluronic F-127) was added to the dye solution to assist solubility. Incubation times from 10 minutes to 60 minutes wereinvestigated to determine the optimal time for dye loading.
Prior to imaging, an aliquot of the dye-loaded cell suspension was transferred to an imaging chamber on the CLSM stage and allowed to "plate out". The bottom of the perspex imaging chamber consisted of a poly-L-lysine [0.5mg/ml] coated glass coverslip to which cells adhered. Subsequent perfusion with aCSF removed any cells that had not adhered to the coverslip and which might otherwise hinder imaging.
2.3 Ratiometric UV-CLSM:-
The CLSM used in this study was a BIO-RAD MRC-600 series instrument equipped with a 2W argon-ion laser (Coherent Enterprise, Innova Technology) modified to emit two main laser lines in the UV part of the spectrum at 351nm and 364nm using 50mW power. Increasing the UV output caused the emission of extra laser lines at 488nm and 514nm in addition to those produced in the UV. Appropriate filters were used to isolate the excitation wavelengths, the main excitation line, 351nm was used to activate Indo-1 near to its excitation peak (355nm) and make simultaneous recordings of the dual wavelength fluorescent emissions centred at 405nm and 485nm (using Bio-Rad Filters In-1 and In-2).
Figure 2 Determination of dissociated neurone viability by patch clamp recordings of barium currents through calcium channels: Whole cell patch clamp was used to measure voltage-gated membrane currents. The control extracellular recording solution was Gey's balanced salts solution (GBSS), and the electrode solution contained (in mM) Gluconate, 100; HEPES, 40; KCl, 9.6; TEA-Cl , 3; D-AP, 5; EGTA, 10; ATP-Mg, 2; GTP, 0.3; leupeptin, 0.1; pH adjusted to 7.3 with CsOH (approx. 80mM). Following the creation of a giga-ohm seal and membrane breakthrough, the currents recorded at successive times after time zero (T=0) in response to a step to 0mV from the holding potential of -80mV were recorded. Examples (leak corrected) are shown labelled with the changes to bathing solution. T=0: Fast inward current is followed by a slow outward current. T=30s: The extracellular solution was replaced with an ionic solution containing (in mM)- CaCl2, 5; NaCl, 106; KCl, 2.5, MgCl2, 2, TEA-Cl, 25; D-AP, 6; glucose, 20; PIPES, 20; pH adjusted to 7.2mM with CsOH. Following addition of TEA, D-AP and Cs to block K+ currents, the inward current is enhanced and the outward current reduced. T=60s: addition of TTX 0.05然 to block the fast inward Na+-current, and equimolar replacement of Ca++ with Ba++ (5mM) lead to a prolongation of the inward current. T=2min: Continued exposure produced a slow onset inward current with little inactivation. T=5min: The isolated Ba++current through voltage-gated Ca++-channels has a slow onset (>5ms to peak) little or no inactivation and a large slow tail current, typical of an L-type response. These responses could be maintained for up to an hour. (Key:- TEA - Tetra Ethyl Ammonium-Cl; D-AP - D-Amino-pyridine, ; TTX - Tetrodotoxin ) (data courtesy of Stanford,I.M., 1992).
Black levels:-
Prior to recording images the background "black" levels were manually zeroed for each photomultiplier tube to ensure that images and therefore subsequent ratiometric calculations were directly comparable. This was achieved by monitoring the output voltage of the photomultiplier tubes on an oscilloscope and adjusting the black level gain dial of each tube until a zero level was attained.
Objective lenses:-
We have routinely used one of three objective lenses to record our fluorescence data, a Nikon x60(oil) Plan-Apo NA(numerical aperture)=1.4 for high magnification investigations of single cells or parts of cells; a Nikon x40 Fluor NA=0.85, used for high magnification scans of individual cells and a Nikon x10 Fluor NA=0.5, for low power scans of fields of cells. The amount of light transmitted and hence the brightness of the fluorescence image obtained by lenses of the same specification can vary - it is advised that a series of lenses are tested prior to selection of a lens for purchase. Lenses were often used in conjunction with an electronic zoom facility provided by the Bio-Rad COMOS software package which enabled a further magnification factor of up to x8 where required.
Temporal and Spatial Resolutions:-
Theoretically, the x-y resolution at 351nm should be approximately 1.4x better than that at 488nm (resolution % (0.61.Wavelength)%NA). Practically, chromatic aberrations incurred meant that any improvement gained by using shorter wavelength UV excitation was not always obvious to the user.
Resolution in the z plane of a CLSM is determined by a number of factors including wavelength of laser light being used and the numerical aperture of the objective lens. These factors are approximately related by the equation:-
Z½ = 0.45% / n{1-cos[sin-1(NA/n)]}
where Z½ is the distance in the z direction at which the image intensity drops to half its original maximal level; % is the wavelength of light; n is the refractive index of the immersion medium and NA is the numerical aperture of the objective lens. Confocal aperture diameters may also be adjusted to alter the volume of space from which a fluorescent signal can be collected, and hence alter the z-resolution. The confocal aperture settings must be the same for each photomultiplier tube. Signal strength for both channels must be sufficient to allow high and low wavelength fluorescence data to be ratioed ie. intensities must be greater than zero. Confocal aperture settings (cfa) of between 5 and 10 were routinely used to obtain suitable fluorescent signals, this has the effect of limiting the optical sectioning ability of the objective lens from theoretical best of %0.35痠 to a working value of approximately 3痠 for our most frequently used lens the x40 Nikon Fluor. This means that the narrowest optical section possible for the x40 lens under these conditions is about 3痠 (Figure 3), it is therefore of no advantage to optically section cells at smaller z intervals. An advantage of imaging thicker optical sections is to allow the fluorescent signal to be gathered from a greater volume maximising the signal strength and henceallowing a lower, potentially less damaging, laser intensity to be used.
Figure 3
Assessment of confocality and indo-emission spectra:
A: Histogram of apparent mean image z-depth (% s.e.mean) for a 0.2痠 diameter fluorescent bead (Molecular Probes) versus confocal aperture (cfa) setting. Image depth was measured as the distance between the first and last sequential optical sections recorded at 0.1痠 intervals along the z-axis of the spherical bead in which an image appeared (n=5 per cfa setting). At the smallest apertures the z-resolution was 500nm.
B: In order to assess the Ca++-sensitivity and emission fluorescence dependence, we undertook tests of indo-1 in a solution of similar osmolarity and ionic composition to intracellular solution. The figure shows the emission spectra for excitation at 355nm (matched to the Argon ion laser UV line), when the dye was mixed with different Ca++-buffer solutions (from Molecular Probes). The dye concentration (5然) was set to be much smaller than that of the Ca++-buffers (1-10mM EGTA- Molecular Probes Calibration kit C-3009) in order to avoid altering the free Ca++ concentration. Progressive changes in spectra were observed as the Ca++ concentration was increased from zero to 1.35然. This reveals the sensitivity of the system in the 100nM range and the saturation at >1然 as would be expected for a dye with a KD of 200nM.
Image collection protocol:-
Pairs of images were simultaneously collected at a size to gain maximal information (384x512 pixels each) at a rate of %1 frame per 3 seconds. For a z-stack imaged at 2.5痠 intervals this would routinely require 30-90 seconds per complete 3D data set. Faster acquisition of image pairs was possible if the image size was reduced increasing image capture rates to sub-second time frames. Images were recorded and stored on computer hard drive for off- line analysis at a later date and from there transferred to optical disc for permanent storage.
Calibration of Indo-1:-
Calibrations of calcium ion concentrations were carried out following experiments: first, neurones were exposed to a cocktail of ionophores and uncouplers including ionomycin [10然] to provide a maximal fluorescence intensity (Fmax) value and then by exposure to the cocktail which in this case contained EGTA [10mM] in addition to the previous cocktails ingredients to chelate free calcium and provide a value for the minimum fluorescence intensity (Fmin). Fmax and Fmin values could then be substituted into the relevant equation(s) to provide a means of determining the Ca++i for a particular image ratio intensity value.
Image processing:-
Stored image pairs were ratioed using computer software developed within the group. Essentially, images were ratioed (405nm/485nm) and the resultant image masked using binarised images constructed from sequentialimage files in the original data set. Images were usually 3x3 median filtered to improve background noise prior to ratioing and rendered into false colour using a look-up table designed to make changes in intensity both continuous and easier to distinguish by eye (Clarke & Leonard 1989). Image sets were then compared to ascertain if, and where, changes in the distribution of Ca++i had taken place.
In order to allow interactive analysis and display of the the sub-cellular localisation of Ca++ changes, the large 3D image data sets (%0.35MB per optical slice) were transferred to a Silicon Graphics Crimson workstation and examined using custom written modules in the "Explorer" (Silicon Graphics) data visualisation environment. These allow the simultaneous display of all z-sections in registration and with perspective. The slice separation and orientations could then be altered to produce the most informative view of the dataset. The image was then transferred to a postscript file for printing.
2.4 pH and its effect on Ca++i estimations using Indo-1:-
Indo-1 is known to be pH sensitive (Morris et al., 1994) therefore it is advisable to check whether the intracellular pH environment remains constant during the course of an experiment or whether it could be changing and thus possibly distorting the accuracy of Ca++i reporting by the dye. To address this problem the ratiometric pH sensitive fluorescent probe Snafl-calcein AM (Molecular Probes) was employed in place of Indo-1 to determine if any changes were occurring in the internal pH of neurones under identical experimental constraints and procedures. Neurones were loaded with Snafl-calcein AM [3然] for 30 minutes at room temperature (ca.18oC) and were then treated as cells loaded with Indo-1. Calibrations were performed by exposing neurones to solutions of known pH contained within a cocktail of ionophores.
3.0 RESULTS
3.1 Cell yields:-
The protocol described in 2.1 provided approximately 1900 healthy cells (assessed by their ability to exclude propidium iodide) per hippocampal slice though this yield can be greatly improved with small modifications to the protocol especially by modification of the aCSF composition. Removal of NaCl and its replacement with equiosmolar sucrose in both bicarbonate and PIPES aCSF is probably the single most effective modification to increase cell yields. Decreasing cell excitability by this method, a three-fold increase in cell numbers (6200 cells/slice) can be realized. The major drawback of this modification is that there is a rapid loss of previously healthy neurones on their return to the normal NaCl containing aCSF. The usefulness of this procedure therefore does not extend to experiments in which a physiological aCSF is required. An alternative way of damping down cell excitability is to include tetrodotoxin (TTX;[0.5然]) in solutions used throughout cell preparation, yields of approximately 3400 cells per slice can be achieved using this method. Although this overcomes the problem of returning cells to a normal medium it leaves Na+ channels blocked, an undesirable problem for many experimental procedures.
Replacing extracellular CaCl2 with equimolar sucrose throughout cell preparation produced yields of approximately 600 cells per slice. Addition of cadmium ions [50然] to the normal aCSF (blocking voltage-dependent Ca++ channels) dramatically reduced the amount of Ca++ able to enter the neurone whilst leaving the ion available for its structural role(s). This protocol produced yields of up to 6000 cells per slice.
Figure 4-
Effect of ionic composition on viable cell yield.
Histograms to show the yield of viable neurones per hippocampal slice when various preparation protocols are employed (modifications to the standard saline solutions are indicated for each column). Asterisks indicate statistical significance (* P<0.01; ** P<0.001). Minimising Ca++ entry during the dissociation, by either Cd++ block or reduction of voltage-gated Ca++ influx by prevention of depolarising Na+ currents (TTX block or Na + removal) produces an increased neurone yield. However, reintroduction of normal saline solutions can cause subsequent degeneration to some of these neurones.
Cell yields may be at least doubled by utilizing the NMDA receptor blocker D-APV or the AMPA receptor blocker CNQX producing around 2500 cells per hippocampal slice. Blockade of the metabotropic receptor by MCPG was found to increase cell yields to approximately 1400 cells per slice.
The age of the animals used for the preparation of dissociated cells was found to be critical. Cell yields are much greater from neonates than from adult animals due to their greater resistance to anoxia. Throughout these experiments animals of between 8 and 12 days of age were used providing constant and reliable cell numbers.
3.2 Propidium iodide:-
Propidium iodide is a fluorescent cell-impermeant DNA-binding dye widely used to assess cell viability. Healthy cells displayed no traces of fluorescent emissions (maximum @ 617nm) when excited by either 488nm or 514nm laser light whereas cells whose membranes had been compromised showed a strong fluorescence primarily in the nuclear region. Studies to determine the viability of dissociated cells using propidium iodide have provided valuable information about the detection of healthy cells using light microscopy. CLSM is able to simultaneously record and display transmission and fluorescence images of a cell (Figure 1). This enables direct comparison of the morphology of a neuron and whether or not it has taken up propidium iodide (ie. is dead/dying) and shows fluorescence. Healthy cells usually appear to be enshrouded in light when viewed by normal light microscopy, unhealthy neurones appeared swollen or shrunken and usually lacked any sign of the aforementioned birefringent "halo". Generally, and with practice, visual inspection using light microscopy at the initial stages of an experiment was sufficient to enable dead/unhealthy cells to be avoided.
3.3 Indo-1 loading:-
Preliminary experiments showed that maximal loading occurred after approximately 60 minutes, but that 30 minutes incubation at room temperature was the most successful protocol, producing healthy, dye-loaded cells. Longer loading times resulted in uptake of Indo-1 into the nucleus and cytoplasmic organelles, a result mimicked by shorter loading times at higher temperatures. Increasing the loading concentration of Indo-1 AM in order to give greater signals should be treated with caution due to the actions of Indo-1 as a buffer. This could result in Ca++ activity changes not being observed due to the spare buffering capacity of the dye.
It was found that the addition of Pluronic F-127 as a detergent to aid solubility of dyes did not always appear to enhance the loading characteristics of Indo-1 and therefore it was usually omitted.
Figure 5
Measurement of intracellular Ca++ in dissociated pyramidal neurones:
Confocal laser scanning microscopy using a modified Biorad 600 fitted with a UV-Argon ion laser allows the ratiometric determination of Ca++i throughout a single plane through a neurone.
Right hand panel: The upper monochrome images show the indo-1 emission at two different wavelengths, left 405nm and right 485 nm, recorded simultaneously from the two channels of the CLSM. Both channels show theconcentration of fluorescent dye in the neurone. The lower wavelength, left-hand image, indicates high Ca++ by higher fluorescence and the higher wavelength, right-hand image fluorescence is reduced by Ca++i. Ratioing the two images produces the lower, main colour image of the pyramidal neurone. The ratio values have been scaled in to the range of 0-255 and are represented as colours in accordance with the colour bar shown. Lowest values correspond to 10nM whilst the highest values on the colour bar represent 1然, the scale is effectively logarithmic. The blue horizontal line across the images denotes the localisation of the data depicted graphically in figure 6.
Left-hand panel: The CLSM can be used to measure a time-sequence of changes in several cells. In the field of cells shown there are several pyramidal neurones, most of which begin with a low resting level of Ca++i . The level of Ca++ is colour coded as for the main image (blue, approx. 100nM). Ratiometric images were taken every 120s, and between the top, first, frame and the next in sequence ryanodine (3然) was added. The Ca++ levels gradually increased in all the cells, saturating the dye (white) in a number. Additional images can be seen on http://www.neuro.soton.ac.uk.
Acknowledgments: Thanks to Biorad Ltd. and %ios Scientific Publishers Ltd.
who co-sponsored colour plate reproduction costs
3.4 Confocal recording and the effects of calcium mobilizers:-
Once a healthy dye-loaded neuron had been selected images were recorded. Image pair recording was only problematical when, very infrequently, black levels drifted and required re-zeroing during the course of an experiment.
Calcium mobilizing agents caused changes in the intensity of image pairs collected and hence in Ca++i following drug exposure. Different drugs may cause different patterns and intensities of change. The Ca++i under control conditions was routinely estimated to be 50-100nM whilst the caffeine induced elevation reached levels of approximately 1000nM in the cell body. Cells that looked healthy but that had an abnormally high Ca++ ratio at the start of an experiment were rejected prior to further experimentation because this was a sign of impending cell death.
3.5 Effect of pH:-
The intracellular pH of neurones may be dramatically affected by protocols adopted during the course of an experiment. pH measurements using the ratiometric pH sensitive fluorescent probe Snafl-calcein was used to measure resting (control) and experimental intracellular pH values following application of drugs to dissociated hippocampal neurones. Control values were approximately pH 7.2 % 0.8 (mean % S.D.). The addition of 1S,3R-ACPD ([100然]; 5minutes) produced a shift towards acid pH with average measurements indicating that pH increased to 6.6 % 1.0 (mean % S.D.).
Figure 6
Graphical representation of a the Ca++ data from a single line passing through the neurone. The digital data obtained from the CLSM can be shown graphically as well as with images. The upper plot shows the fluorescence intensity values measured at different points along a line passing through the middle of the neurone shown in the main panel of figure 5 (indicated by the blue line in the image). This is a sequence of voxel values ranging from 0-255, ie. the range of the AD converter which accepts data from the photo-multiplier tubes. NB. The black values have been accurately set and the full AD voltage range is being utilised. Both the 485nm and 405nm emissions are greater from the neurone than from the background, showing dye distribution, but they vary in separate ways. Division of the 405nm/485nm emissions from each voxel (3-D pixel) gives the lower plot which is effectively log Ca++i plotted against distance, showing the greatest Ca++ level in the cell soma.
4.0 DISCUSSION
4.1 Cell preparation:-
Modifications to the composition of the buffers used in the preparation of dissociated hippocampal neurones appeared to be the single most important factor involved in producing viable cells in large numbers that are suitable for calcium imaging. For most applications the use of aCSF buffer was adequate providing approx 2000 viable neurones per hippocampal slice.
Ca++ plays a major role in the process of cell death and one might expect its removal from the external medium to produce a dramatic increase in cell yields; this however is not the case highlighting the apparent requirement of divalent cations for membrane integrity. A compromise can be achieved by adding cadmium ions [50然] to the normal aCSF to block voltage-dependent Ca++ channels dramatically reducing the entry of Ca++ into neurones whilst leaving the ion available for its structural role(s). This method provided about 6000 cells per slice.
Production of large numbers (thousands) of neurones by the blockade of various channel species can be useful if, for example, the cells are required for antibody mapping of a particular receptor species not interfered with by the preparation protocol.
Generally, if dissociated neurones are required they should be produced in such a way that their preparation will not interfere with subsequent experiments. For morphological studies or receptor/ion channel distribution studies those protocols which give the greatest cell yields could be used provided that the focus of the study has not been blocked during cell preparation. For example it would not be prudent to study Na+ channel distribution using a fluorescently labelled ligand if TTX was a component of the aCSF used during their preparation! For electrophysiological studies the best possible advice is to use normal aCSF; a slightly smaller cell yield is a fair compromise to make for suitable cells.
One final point to mention concerns the age of animals used for the preparation of dissociated cells. Cell yields are much greater from neonates than from adult animals due to their greater resistance to anoxia. It is therefore advantageous to prepare cells from young animals where prudent to do so. For example, comparing data obtained from adult rat brain slices with neonatal dissociated cells could be invalid due to changes in the numbers and types of receptors/channels that are expressed throughout development.
4.2 Confocal Ratioing:-
Initial ratioing and filtering was done using either pre-written CoMOS (BioRad) software or custom-written programmes constructed within the SOM (BioRad) software package.
Ratioing of data pairs was found to be reliable and consistent. Standardization of experimental approach and of the settings used for laser intensity and aperture sizes etc. ensured that the multiplication factor used to make full use of the 256 intensity levels available to us for ratioing were also constant.
Of those problems encountered "edge effects" (the production after ratioing of a very high ratio intensity along an edge of a cell) were perhaps the most common. Edge effects are caused when one of the pair of images to be ratioed has a very low Ca++ concentration and hence one image will register this by a high intensity image and the other image by a very low intensity - if the low intensity pixels reach zero the computer will need to divide by zero at some stage and will create a value of infinity for the subsequent ratio by default - giving a false high intensity (=255) reading instead of the expected low reading.
Movements of the cell being imaged was a rare problem that could have serious consequences if a particular data set was to be viewed for example as a projection or as sequential optical slices with perspective. Movements were particularly detrimental to viewing neurones in 3D. Usually perfusion at a rate of 1ml/min was sufficient to weed out any cells which had not adhered to the poly-L-lysine coated coverslip but occasionally one slipped through the net. Movement, even if slight, was enough to cause us to abandon that particular experiment as some of the data manipulations we chose to perform on the data set required cells to remain stationary.
4.3 Calcium Activity:-
A change in the Ca++i was judged to have occurred when levels altered by an amount distinguishable by eye with respect to control levels (usually at least doubling the baseline Ca++i levels to ca.>200nM). Small changes in Ca++i were indistinguishable from baseline fluctuations in control readings. In general, detecting a change by eye was sufficient to convince us that a change had occurred the implementation of a pseudo- colour look-up-table made detection far easier.
Different drugs acting at different receptor types may be able to access different pools of intracellular Ca++ and therefore it may be possible to image different 3D patterns of Ca++ activity. In addition, the activation of different classes of receptor will almost certainly have a profound effect on the amounts of Ca++ that are released and the time course of its mobilization.
4.4 pH and its effect on Indo-1:-
The addition of drugs to neurones can dramatically alter the measured values of their intracellular pH and hence affect the accuracy of calcium ion concentration estimations by the calcium sensitive ratiometric dye Indo-1 (Morris et al. 1994). Simple pH measurements, for example using the ratiometric pH sensitive dye Snafl-calcein, to determine if any such changes are caused by the protocols adopted during the course of an experiment would seem an important consideration to make in the subsequent discussion ofalterations in Ca++i. It is useful to know intracellular pH when, for example, 1S,3R-ACPD (a known calcium mobilizing agent) is added to dissociated neurones as the pH can clearly be seen to alter and become more acidic. The marked increase in acidity would cause a decrease in the affinity of Indo-1 for calcium and therefore lead to an under-estimation in the concentration of Ca++i. The effects of pH can however be compensated for and a more accurate estimation of Ca++i arrived at (Phenna et al. 1995).
4.5 Data visualization:-
Data visualization is a very important aspect of imaging - however, the presentation of data in such a way as to maximise its potential can potentially be both time consuming and expensive.
The most obvious requirement is the use of colour. Pseudo-colour spectra are essential when trying to discriminate subtle changes in intensity by eye. The human visual system is better equipped to discern changes in colour than it is to register subtle changes in the degree of contrast that an image may possess when it is only registered in black and white. As mentioned earlier ( 2.3 Ratiometric UV-CLSM: Image Processing) we have used a standardised continuous pseudo-colour spectrum that has been optimised for visual contrast and resolution (Clarke & Leonard 1989). In comparison to other pseudo-colour look-up tables used to show changes in fluorescence intensity the spectrum of Clarke and Leonard (1989) makes changes noticeably easier to detect.
The use of a Silicon Graphics Crimson© workstation and custom written modules in the "AVS" (© Advanced Visual Systems) data visualisation environment allowed the simultaneous display of all z-sections in registration with perspective. This software is available to Universities at reduced costs via the Chest agreement. Examples of these and other images can be seen on our WorldWideWeb site (http://www.neuro.soton.ac.uk).
The slice separation and orientation could then be altered to produce the most informative view of the dataset. The image was then stored on disc and/or transferred to a postscript file for printing using either a thermal imaging printer or photographed onto colour transparencies. Using this method images can be produced which give information on the distribution of Ca++ in three dimensions whilst retaining the 3D perspective of the neuron.
Various mathematical filters were tested to judge their effectiveness at improving the overall quality of image presentation without obviously distorting the data. It was found that for our particular applications the use of a 3x3 median filter prior to ratioing greatly improved background noise/scatter levels surrounding isolated neurones without seriously affecting Ca++ intensity data.
5.0 CONCLUSIONS
1) UV confocal laser scanning microscopy can theoretically produce submicron spatial resolution in all three dimensions.
2) The emission spectra shift of Indo-1 allows ratiometric estimations of the Ca++i activity.
3) Simultaneous recordings of two emission wavelength bands, permits millisecond temporal resolution.
4) In practice, spatial and temporal resolutions have to be traded off against each other.
5) Subcellular standing gradients of Ca++i can be observed.
6) Caffeine produces an increase in Ca++i presumably by the release of calcium from intracellular stores.
7) Intracellular pH values may change during the course of an experiment following application of drugs - this can alter the accuracy of Ca++i concentrations reported by Indo-1.
8) UV-ratiometric confocal microscopy offers a powerful means of analysing the subcellular compartmentalisation of second messenger Ca++i signalling which has great significance in enhancing our understanding of cellular function.
Acknowledgements:-
This work has been supported by the Science and Engineering Research Council, Medical Research Council, Wellcome Trust, Action Research, Wessex Medical Trust and the University of Southampton.
6.0 REFERENCES:-
Chad,J.E.; Eckert,R. & Ewald,D.(1984)
Kinetics of calcium dependent inactivation of calcium current in voltage clamped neurones of Aplysia californica. J.Physiol. 347: 279-300
Chad,J.E.; Stanford,I.; Wheal,H.V.; Williamson,R. & Woodhall,G. (1991) In "Cellular Neurobiology; A Practical Approach" , Eds. Chad,J.E. & Wheal,H.V. pp.19-36 Academic Press
Clarke,F.J.J. & Leonard,J.K. (1989)
Proposal for a standardised continuous pseudo-colour spectrum with optimal visual contrast and resolution. in "An International Conference On Image Processing and its Applications" IEE Conference Publications 703, pp.687-691.
Grynkiewicz,G.; Poenie,M. & Tsien,R.Y.(1985)
A new generation of Ca2+ indicators with greatly improved fluorescence properties. J.Biol.Chem. 260: 3440-3450
Jane,S.D.; Chad,J.E. & Wheal,H.V. (1992)
Determination of the resolving ability of confocal microscopy in measuring the three dimensional morphology of neurones within rat brain slices.
Neuroscience Letters 42:S40
Jane,S.D.; Phenna,S.; Tunwell,R.E.A.; Lai,F.A. & Chad,J.E. (1995)
Caffeine-induced elevations of intracellular Ca2+ in rat hippocampal neurones, and immuno-histological evidence for ryanodine receptor isoforms.
J.Physiol. 487P: 21-22
Kay,A.R.(1991)
Inactivation kinetics of calcium currents of acutely dissociated CA1 pyramidal cells of the mature guinea pig hippocampus. J.Physiol. 437:27-48
Kay,A.R. & Wong,R.K.S.(1986)
Isolation of neurons suitable for patch clamping from adult mammalian central nervous systems. J.Neuroscience Methods 16: 227-238
Miller,R.J.(1991)
Metabotropic excitatory amino acid receptors reveal their true colors.
TIPS. 12: 365-367
Minta,A.; Harootunian,A.T.; Koa,J.P.Y. & Tsien,R.Y.(1987)
New fluorescent indicators for intracellular sodium and calcium.
J.Cell Biol. 105: 89A
Minta,A.; Koa,J.P.Y. & Tsien,R.Y.(1989)
Fluorescent indicators for cytosolic calcium based on rhodamine and fluorescein chromophores. J.Biol.Chem. 264: 8171-8178
Morris,S.J.; Wiegmann,T.B.; Welling,L.W. & Chronwall,B.M. (1994)
Rapid simultaneous estimation of intracellular calcium and pH.
Methods in Cell Biology 40: 183-220
Phenna,S.; Jane,S.D. & Chad,J.E.(1995)
Increased perinuclear Ca2+ activity evoked by metabotropic glutamate receptor activation in rat hippocampal neurones. J.Physiol. 486.1: 149-161
Segal,M. & Manor,D.(1992)
Confocal microscopic imaging of [Ca2+]i in cultured rat hippocampal neurons following exposure to N-methyl-D-aspartate. J.Physiol. 448: 655-676
Simon,S.M. & Llinas,R.R.(1985)
Compartmentalization of submembrane calcium acting during calcium influx and its significance in transmitter release. Biophys.J. 48: 485-498
Smith,S.J. & Augustine,G.J.(1988)
Calcium ions, active zones and synaptic transmitter release.
TINS. 11: 458-464
Stanford,I.M. (1992)
'Studies on synaptic and ionic mechanisms involved in the regulation of excitability in central neurones.'- Ph.D. Thesis; Southampton University, UK.
Stanford,I.M.; Caddick,S.J.; Wheal,H.V. & Chad,J.E.(1992)
GABA B receptor modulation of high voltage activated barium currents in acutely dissociated rat hippocampal neurones.
Soc. Neurosci. Abstr. 18: p432; 188.3
Thastrup,O.; Cullen,P.J.; Drobak,B.K.; Hanley,M.R. & Dawson,A.P.(1990)
Thapsigargin, a tumour promoter, discharges intracellular Ca2+ stores by specific inhibition of the endoplasmic reticulum Ca2+- ATPase.
PNAS. 87,2466-2470
Tsien,R.Y. & Poenie,M.(1986)
Fluorescence ratio imaging: a new window into intracellular ionic signalling.
TIBS. 11,450-455
White,J.G.; Amos,W.B. & Fordham,M.(1987)
An evaluation of confocal versus conventional imaging of biological structures by fluorescence light microscopy. J.Cell Biol. 105: 41-48
Converted by Andrew Scriven