Optical recording experiments of population neurons in brain slices
Optical recording experiments of population neurons in brain slices
Optical recording involves the use of molecular indicators whose optical properties (absorption or fluorescence) vary according to the parameters of cellular activity. Available indicators include reagents that are sensitive to membrane potential (Vm), to calcium concentration, or to the concentration of other ions (H+, Na+, K+, Mg2+, Zn2+, and Cl-). Optical recording has several advantages over conventional methods.
Operation method
Optical recording experiments of population neurons in brain slices
Principle
Figure 16-5 shows a flowchart of the decisions to be made before starting a new experimental application. The first step is to decide which type of indicator to use based on the need for recording: VSD for recording membrane potential, CaSD for recording calcium ion concentration, and the corresponding ion-sensitive indicator for recording the concentration of other ions, and then to select the specific indicator. For VSD, the following points should be considered: whether to choose a light-absorbing or fluorescent indicator; the desired spectral properties of the indicator (determined by the availability of filters, light sources, and tissue properties such as autofluorescence); and the number of experiments and errors associated with the loading of dyes for different samples. For ion-sensitive dyes, it is important to consider the affinity of the dye required for the experiment, and the kinetic properties of binding to ions. Also to be considered is whether to use selective injection of the dye into specific tissue structures or to use the immersion bath method; although other methods are mentioned in this section on voltage-sensitive dyes, VSD is usually performed using only the immersion bath method, and the next step is to select the type of microscope (upright or inverted) and objective lens. The advantages and disadvantages of differently constructed microscopes have already been discussed. For the objective lens, the main considerations should be its type (air or oil mirror for inverted microscopes, water mirror for upright microscopes), the desired magnification and numerical aperture, and the working distance. The choice of filter should be based on the characteristics of the indicator. Finally the type of detector has to be selected. All of the issues discussed above should be carefully considered before making a final decision. Also, it is highly recommended that a single photodiode be used first to test the stability of the planned experiment with respect to the amount of indicator to be loaded, the size of the signal, and the intensity of the light reaching the detector (see the Troubleshooting section for details). Below, we give a few specific examples of optical recordings after the final decision has been made after the above consideration process. Move makings The following is a brief description of the procedure used in our laboratory to prepare transverse slices of the hippocampus; however, any other method is feasible as long as the brain slices are viable. The appropriate staining solution should be prepared before the brain slices are cut. 1. The animal is anesthetized with methoxyflurane and is rapidly guillotined with a guillotine. 2. Immediately place the removed brain in an ice bath solution and cool for 3-5 min. 3. The hippocampal region was isolated from the brain. 4. Put the hippocampal region into a vibrating slicer (Vibratome1000, TPI) and cut out the brain slices with a thickness of 4OOum from the middle 1/3 of the hippocampus. 5. Before staining, the slices were immersed in a room temperature bath solution, but if CaSD bath staining was used, the slices were immediately placed in the staining solution. Below we will describe the operational procedures for recording evoked or spontaneous signals from dye-loaded brain slices using the various experimental designs mentioned above. Specific experimental details aside, there are several things that must be done for any optical recording experiment. First, the autofluorescence of the tissue is to be measured; as mentioned earlier, this is especially important when the indicator used needs to be excited with short wavelength light. When selectively injecting CaSD into specific tissues of a brain slice into the axon bundles, the autofluorescence can be recorded in regions of the same brain slice that have not been injected with the dye and that are similar to the recording area, or the autofluorescence of the recording area itself can be measured prior to the injection of the dye in order to correct for the signal. In the case of loading dye into a brain slice by the immersion bath method, the autofluorescence of the corresponding area on another of the same animal can be recorded. It is essential that the intensity of the illumination light remains constant between the measurement of autofluorescence and the actual recording. Another thing to do in every experiment is to record photobleaching. Optical indicator molecules are bleached during illumination, meaning that the fluorescence they emit is diminished. It is impossible to know whether these molecules are actually bleached or trapped and temporarily not fluorescing. Without considering its precise mechanism, photobleaching is a monoexponential function process, so it is easy to compensate for it by correcting for it. The simplest way to correct for photobleaching is to record a period of time in the absence of any activity (neither induced nor spontaneous). During this time the parameters (A/D conversion speed, recording time, shutter open time) should be the same as when the data was actually collected. Another approach is that, since photobleaching is a monoexponential function process, if the shutter is left open long enough before the actual recording, the photobleaching will be close to steady state at the time of recording. This latter approach is useful when recording spontaneous activity. The steps involved in recording evoked activity with a single photodiode are relatively straightforward—we usually use an oil immersion objective (Achroplan 50x, numerical aperture of 0.9, Zeiss). We use a multi-purpose card (12-bit, 50 kHz, DAS-50, Keithley) to perform analog-to-digital conversion of the data and to provide digital signal lines to control the shutter and electrical stimulator. The signals coming out of the microelectrodes traditionally used to record extracellular field potentials, as well as the signals coming out of the single photodiode, are converted to digital signals by this card. Figure 16-7 shows the steps involved in a typical experiment. The first step is to determine the appropriate autofluorescence. To record evoked responses with a photodiode array, we typically use a 10x and numerical aperture of 〇.5 objective (Zeiss). We used a multi-purpose I/O card (Flash12, StrawberryTree) to A/D convert the data and provide digital signal lines to control the PDM amplifier, shutter, and electrostimulator. The A/D converter on this card has an 8-channel, 12-bit, 400kHz, 256k sample-in-bit memory; it also has eight TTL input/output (I/O) lines and a two-channel A/D converter. The signals from the microelectrodes conventionally used to record extracellular field potentials are fed into the PDM amplifier and connected to all the optical channels shown in Figure 16-4. Figure 16-8 shows the steps involved in a typical experiment. Separate AC and DC measurements are necessary for the reasons already discussed above. Because the indicator we usually use with the PDM is excited with long wavelength light, in which case autofluorescence has a much smaller effect on the signal than the bias caused by the amplifier, it is not necessary to correct for autofluorescence. The hardware and basic operating procedures required to record spontaneous events with PDM are similar to those for recording evoked events with PDM. The main difficulty is the lack of an event to trigger data acquisition. Methods that are typically used for continuous recording of spontaneous activity by microelectrodes are usually not used because of the excessive amount of data (>10000 channels). We used two methods to record spontaneous activity with PDM. The first method uses software to obtain the probability of spontaneous activity (Coiom and saggau 1994; Sinha et aL1995): a segment of data of a pre-determined length is acquired and displayed; when the researcher sees an event of interest, the acquisition of data is stopped and the last segment is saved. This method is suitable for recording relatively frequent events, e.g., frequencies greater than 0.2 Hz; however, it is difficult to reliably record events that occur much less frequently. Some minor improvements to this method can make it possible to reliably record events that occur at frequencies lower than 0.05 Hz. Figure 16-9 shows a flowchart of the improved method. A segment of data is acquired (usually 0.5 to 1 s), with only the electrical signal channel;^ the output is separated and displayed. This process is repeated until the researcher observes spontaneous activity and triggers the computer; the hit segment of data (usually 2~3 segments) that was captured before the trigger is separated for output, display, and can be further processed and stored. The first step of optical recording data processing is to correct the bias caused by autofluorescence and instrumentation. For a single photodiode, it is sufficient to simply subtract these biases from all optically recorded data points; for PDM, these biases are subtracted from the DC fluorescence representing static fluorescence. All optical signals are to be expressed as the fluorescence change intensity divided by the fluorescence intensity at rest (AF/F). This representation helps to eliminate biases due to differences in dye concentration, illumination intensity, and light sensitivity of the various elements of the photodetector. For a single photodiode, fluorescence at rest (F) is a simple average of the portion of the data before any activity is observed in the data recording, especially before the onset of stimulation; subtracting F for each data point during the recording process yields the fluorescence change F. For a PDM, DC fluorescence denotes the resting fluorescence, while AC fluorescence denotes the fluorescence change (AF). For CaSD, an increase in AFAF represents an increase in [ Ca2+ ]i . For VSD, a decrease in AF/F means depolarization; therefore, all VSD data are flipped so that depolarization corresponds to an upward deflection. For both detectors, correction for photobleaching is done by first calculating the AFAP of the record data and the photobleaching data, and then subtracting the photobleaching data AF/F point-for-point from the record data AiVF for each data point of a single photodiode; for PDM data, this phase-subtracting correction is done for each element. We quantify the optical signal using two different quantities. For both VSD and CaSD, the Jardine magnitude can be used. For VSD alone, the mean window amplitude (MWA) (Albowitz, Kuhnt 1995) is also a useful quantity. Simply put, MWA is the average value of the VSD signal over a given time window. Unlike signal amplitude, MWA reflects the variability of a segment of activity as well as its maximum amplitude. In addition, for small signals, it has the added benefit of reducing the effect of noise: when calculating this time average, the noise time horizon is shorter than the time horizon over which the MWA is calculated, thus reducing the effect of the noise component. It is important to note that calculating the MWA of the CaSD signal is meaningless because the CaSD signal is affected by the kinetics of the indicator and also by transient changes in calcium ion concentration. The quantity that is useful for the CaSD signal is the first-order derivative of the signal: if the CaSD signal is proportional to [Ca2+]i and the release of calcium ions from the intracellular calcium reservoirs is significantly small, then the derivative should be proportional to the flow of calcium ions. Alternatively, if dyes with fast kinetics are used, especially those with low affinity for calcium ions, then the duration of this derivative is roughly comparable to the time of inward flow of transfer ions (Sinhaetal. 1997). The techniques described above have allowed us to study various aspects of hippocampal brain slices, including synaptic transmission, modulation of transmitter release, plasticity, and epileptic-like activity (WuandSaggau1994a,b,1995,1997;QianandSaggau1997a,b;ColomandSaggau1994; Sinhaetal.1995,1997)〇 Below we present some examples from these studies as an illustration of the use of various techniques. The reader is referred to the specific literature for more examples and details. Figure 16-10 shows experiments in which CaSDFura-2, which can be selectively transported to presynaptic terminals of pyramidal cells in the CA3 region (Fig. 16-6), was injected by pressurization within the lateral branches of Schaffer. Signals were recorded in the radial layer (stratum radiatum) of the CAl. Selective loading is confirmed when only axons of CA3 area pyramidal cells are seen to extend from the injection site to the recording site; some interneurons and glial cells may also have protrusions that extend into this region, but they are much less numerous than axons. Selective loading is also confirmed by blocking postsynaptic activity with the glutamate receptor antagonists CNQX and D-APV, which block postsynaptic responses but do not affect CaSD signaling, confirming that the signal originates from the presynapse. FIGS. 16-10B and 16-10C show the effect of CaSD signaling on Kd (and ;^. ^bff) dependence of the same calcium ion concentration transient measured with high affinity CaSD (Fura-2) and low affinity CaSD (Furaptra) with different results. Both indicators were used to measure calcium ion transient changes in CA3-CA1 presynaptic terminals evoked by a single action potential. The low-affinity indicator Furaptra apparently had faster kinetics. A combined model and experimental study (Sinhaetal. 1997) suggests that after a single action potential, the high-affinity indicator Fura-2 may be locally saturated by locally high concentrations of [ Ca2+ ]; near the cell membrane, while its overall amplitude remains proportional to the local concentration. In this way, Fura-2 can be used to study presynaptic calcium ion transients evoked by single action potentials. However, to study calcium ion transient changes evoked by multiple action potentials, it is more astute to choose a low-affinity indicator such as Furaptra. Our laboratory has used this technique extensively to study which depletion channels are involved in transmitter release at CA3-CA1 synapses, how presynaptic calcium ion flow changes during synaptic plasticity, and the role that presynaptic calcium ions play in the modulation of transmitter release (WuandSaggau1997). We also used this technique to selectively load dye to other tissues in hippocampal brain slices; others have used a similar technique (applying local diffusion of CaSD rather than injection) to selectively load dye to hippocampal brain slices and cerebellar brain slices. Injections of CaSD into the alveus can selectively load dye on presynaptic vesicular cells in the CAl region (RegehrandTank1991;WuandSaggau1994a). Injection of CaSD into the hilus can load dye on the presynaptic terminals of mossy fibers in the A3 area (RegehrandTank1994;QianandSaggauunpublished). Also this method allows selective loading of dye to presynaptic endings of cerebellar parallel fibers (Mintzetd. 1995). Figure 16-11 shows experiments in which hippocampal brain slices were bathed with CaSD calcium orange and a single electrical stimulus was applied to Schefferce to evoke and record calcium ion transients in the CAl region. Because the indicator is used by immersion, it is not selectively loaded into pre- and postsynaptic neuronal tissue. This is confirmed by applying the ionotropic glutamate receptor antagonists CNQX and D-APV to block the postsynaptic response without altering the presynaptic activity. The contribution of individual cellular structures to the overall CaSD signaling is dependent on two major factors: i) the magnitude of the [ Ca2+L change in that structure; ii) the volume of that structure Vi The second factor arises because the number of indicator molecules in a structure is roughly proportional to the volume of that structure-for a detailed discussion of the contribution of different cellular structures to the CaSD signal, see Sinha et al. (1995). Figure 16-11C shows how this technique can be used to obtain information on the time-space properties of neuronal activity. The CaSD signal first appears in the dendritic region in close proximity to the stimulus site, where the presynaptic terminal is located (shown to the left in the figure). From the stimulus site, the signal traveled across the feldsheet to the stractumpyramidale, the layer where the cytosol is located, which is the layer where the postsynaptic cytosol is located. When there is more CaSD loading into the glial cells (Albowitzetal.1997), there is little effect on this transient change because the [Ca2+];change in the glial cells is much slower. Optical recording techniques are well suited for characterizing the temporal-spatial response to complex activity in neuronal populations, such as epileptiform spontaneous activity. We have used this technique extensively to study various aspects of epileptiform spontaneous activity in hippocampal slices of interictal epilepsy (ColomandSaggau1994;SinhaetaL1995,1996). Figure 16-12 shows examples of experiments in which synchronized activity of hippocampal interictal neuronal networks was recorded with VSDRH-414 and PDM. This synchronization activity was generated spontaneously after the addition of the K+ channel antagonist 4-aminopyridine (4-AP, lOO/imol/L), and the addition of the ionotropic glutamate receptor antagonists CNQX and D-APV enabled the isolation of this synchronization activity from the interictal epilepsy-like spontaneous activity (Michelson and Wong1994; Sinhaetal.1996). This response is mediated by a depolarizing response in interneurons associated with GABAa receptors. Optical recording techniques allowed us to study the temporal and spatial characteristics of this response and to compare it with the spontaneous activity of interictal healing-like pain (Sinhaetal.1996). We are now further investigating the characteristics of this response, the types of interneurons that produce it, and the corresponding mechanisms. As mentioned above in the discussion of [Ca2+]; and CaSD, the activity recorded by a single element of the PDM when using VSD is actually a weighted average of the instantaneous membrane potentials of all structures within the area covered by that recording element. For VSDs present only at the cell surface, the weighting is limited to the tissue surface area Si For a detailed discussion of the contribution of individual cellular structures to the VSD signal, see Sinha et al. (1995). Because the VSD signal gives an average of the membrane potentials of multiple cellular structures, it is possible that some groups of structures are hyperpolarized while others are depolarized, and their activities cancel each other out to create the illusion of no activity. This is not a major problem for the CaSD signal, as there is very little physiological activity that would cause [Ca2Ii to fall rapidly from resting levels. It is also important to note that the AC coupling constant (1OOms) of the PDM may have an effect on the signal due to the relatively slow nature of the recorded activity - a way to avoid this problem is to use an analog/digital converter with a sufficiently high intensity resolution, which can be recorded as if it were recorded with a single photodiode in DC-coupled mode only. Note: This technique is not sensitive enough to record the spontaneous activity of a single neuron; only the activity of a group of neurons can be recorded. For more product details, please visit Aladdin Scientific website.
