The patch-clamp technique is an extremely powerful method for studying electro- physiological properties of biological membranes.68 First described by Neher and Sakmann in 1976,69 it is now routinely used in laboratories around the world. The power of the patch-clamp technique comes from the fidelity with which electrophys- iological signals may be recorded. The technique was not widely used by pain researchers until the beginning of this last decade, at which point researchers began to use it to characterize electrophysiological properties and chemosensitivity of neurons involved in the transduction and transmission of nociceptive stimuli.
Patch-clamp techniques have been applied most widely in the pain field to the study of primary afferent neurons. Consequently, application of this technique to the study of primary afferent neurons will be the focus of this section. However, the
technique has been successfully employed for the study of neurons of the CNS in dissociated neurons,70,71 slice preparations,72–78 and more recently, in vivo.79 Two approaches utilizing the patch-clamp technique to study CNS neurons will be dis- cussed at the end of this chapter.
8.6.1 CELL-ATTACHED PATCH-CLAMP RECORDINGFROM
AFFERENT TERMINALS
The closest any investigator has come to cell-attached patch recording of an afferent terminal was recently reported by Brock and colleagues, who recorded extracellular activity from the peripheral terminals of corneal afferents.80 However, this approach has been used to record ion channel activity from C-fiber axons as well as the afferent cell body.81,82 If single-channel recordings are obtained, this approach may enable generation of the most detailed picture of the biophysical properties of the ion channel under study. Furthermore, because it is possible to record from specific sites on a neuron, it is possible to obtain information concerning the relative distribution of ion channels. However, this approach is probably the most technically difficult and labor intensive of all the approaches discussed here.
Unmyelinated axons are small and delicate, so access to isolated axons is not achieved easily. Ion channels in myelinated axons have to be studied at nodes of Ranvier unless axons are demyelinated to allow access to the membrane-bound ion channels normally covered by myelin; a process that may by itself modify the behavior of these ion channels. Furthermore, given the propensity of neurons to selectively distribute ion channels throughout the plasma membrane, the probability of obtaining single-channel recordings is low. Finally, a lot of data are required to generate a representative picture of the behavior of any given channel, and therefore, using this approach to identify injury-induced changes in channel activity is even more difficult.
8.6.2 PATCH-CLAMP RECORDING FROM SOMATA OF
PRIMARY AFFERENT NEURONS
Since many investigators will choose to study the cell body of sensory neurons in vitro, it is worth considering the extent to which it is reasonable to extrapolate findings obtained in the cell body to the afferent terminal. While there will always be legitimate concerns about such extrapolation, several lines of evidence suggest that the cell body of acutely isolated sensory neurons in vitro is a valid model for the afferent terminal in vivo. First, with few exceptions, receptors and ion channels that are transported to the peripheral or central terminals of sensory neurons are present and functional in the plasma membrane of the cell body in vitro. Moreover, many of the receptors for agents present on the cell body of sensory neurons in vitro83–86 appear pharmacologically similar to those near the peripheral and central arbors.87–92 Second, it is possible to induce changes in the excitability of the cell body in vitro with the same manipulations that induce changes in the peripheral terminal excitability in vivo. For example, the cell body in vitro is sensitized by inflammatory mediators such as PGE2.93–98 This observation indicates that the cellular
components (i.e., receptors, second messengers, ionic currents, etc.) necessary to achieve inflammatory mediator-induced increases in nociceptor excitability are present in the cell body in vitro. Third, the sensory neuron cell body in vitro can be induced to release neurotransmitters.99 While it has yet to be determined whether the mechanisms underlying transmitter release from the cell body are analogous to those mediating transmitter release from central and peripheral terminals, it is note- worthy that neurotransmitter release from primary afferent terminals and the cell body are Ca2+-dependent.99
Since a detailed discussion of the theory and practice of patch-clamp electro- physiological recording may be found elsewhere,68,100 this section will address some of the practical issues that must be considered before patch-clamp recording may be performed.
8.6.3 DISSOCIATION PROTOCOLS FOR STUDYING ISOLATED
SENSORY NEURONS
8.6.3.1 Choice of Enzymes
While there are a few investigators who employ a non-enzymatic approach to obtaining isolated neurons, the vast majority of protocols involve at least one enzyme (most commonly collagenase) and often a combination of enzymes.101 The trade-off in the choice of enzymatic treatments is between speed (minimizing the time between tissue harvest and recording) and preservation of plasma membrane proteins. More aggressive enzymatic treatments that produce more rapid dissociation require shorter exposure time, but they increase the likelihood that plasma membrane proteins (i.e., the ion channels to be studied) may be altered or destroyed. Investigators who are new to the purchase of enzymes should not be surprised by the number of different enzyme preparations available. Companies that specialize in enzymes will prepare them within a range of specific activity (specific activity is defined in several different ways depending on the substrate and reaction endpoint). While these ranges in activity are useful in narrowing the choice of enzyme, there may be a lot of variability between preparations of the same enzyme, so it is necessary to test the enzyme under the specific conditions in which it will be used. Thus, it may be helpful to order several small aliquots of an enzyme with different lot numbers, test these, and then order a large quantity of the lot that gave the best results.
Whatever enzyme combination is employed, some mechanical perturbation of the ganglia is usually necessary, particularly when adult animals are used, to facilitate enzyme access to neurons. Thus, manual removal of the connective tissue surround- ing the ganglia with sharpened jeweler’s forceps under a dissecting microscope is recommended, prior to exposing ganglia to enzymes. Since sensory neurons within the trigeminal ganglia are somatotopically organized101 and some of the neurons of interest may be located very superficially, it may be necessary to avoid removal of the connective tissue from these ganglia. Alternatively, ganglia may be cut into small pieces with scissors prior to enzyme exposure.
Several different approaches have been employed to further decrease the time that the ganglia must be exposed to digestive enzymes. Shaking water baths
maintained at 37°C are typically used; however, it is possible to further decrease incubation times by employing a more aggressive shaking system such as that provided by the Nutator™ or the Belly-Dancer™ shaker. We have also had success bubbling the ganglia/enzyme mix with carbogen (5% CO2, 95% O2), which serves the dual purpose of agitating the ganglia and maintaining pH (if a bicarbonate- buffered medium is used with the enzymes). Serum is omitted from the enzyme solution if bubbling is employed to avoid excessive frothing.
8.6.3.2 Choice of Medium
The choice of medium depends on the nature of the specific questions to be addressed and the nature of the experimental protocol. Many factors can influence the properties of the neurons to be studied. Which factors are necessary depends on the age of the animal from which the cells were obtained and the time period over which the neurons will be studied after they are harvested. For example, nerve growth factor is necessary for the survival of sensory neurons obtained from embryos,102 but is also involved in the regulation of phenotypic properties of sensory neurons from the adult.103 Similarly, Delree and colleagues104 have demonstrated that even brief expo- sure of sensory neurons to serum profoundly influences the properties of sensory neurons maintained in culture. The simplest solution to this concern is to study the neurons for as short a time as possible after removal from the animal. For short- term culture, investigators will often choose to provide for only the most minimal of survival needs. Thus, a minimal essential medium (MEM) is often sufficient to maintain cells for several hours. Investigators have also employed Ham’s F-12, F-14, or a combination of MEM and F-12 with success in maintaining sensory neurons in culture for short periods of time. The addition of serum or serum substitutes (see below) to these media enables the maintenance of sensory neuronal culture for weeks. Neurons will also survive in an even more restrictive medium, as several groups routinely maintain neurons in a physiological saline solution with glucose prior to recording.105,106 To further limit the possibility of phenotypic changes associated with dissociation, sensory neurons may be stored at room temperature or lower.101,107
8.6.3.3 The Use of Serum
Many cell types are maintained in culture in a medium containing 10% serum.
Sensory neurons will also thrive in medium containing 10% serum (generally fetal bovine serum, but other sera have also been shown to work well). Our own experience is that if it is necessary to maintain sensory neurons in culture for longer than 24 h, the use of serum or a serum substitute becomes increasingly important to keep the cells healthy. The problem with serum is that its contents are partially unknown, the contents of different lots may vary, and as noted above, even short exposure to serum may result in phenotypic changes.104 We routinely use heat-inactivated serum in an attempt to mitigate some of these concerns, but this is clearly not a perfect solution.
Consequently, investigators have turned to serum substitutes. Several of these are commercially available and appear to enable the maintenance of healthy sensory
neurons.108 The other two solutions that have been employed to address the serum problem are the use of a completely defined medium,104 or omitting serum and maintaining neurons for a short period of time in serum-free media.109
8.6.3.4 Plating Substrates
To study several neurons in a single field of view under conditions where the extracellular solution is continually exchanged, it is often useful to have neurons adhere to the bottom of a recording chamber. To facilitate neuronal adhesion, the plating surface is often coated with a charged molecule. Because of its ease of use, poly-d-lysine is probably used most commonly. Typically, the plating surface is soaked in a 0.1-mg/ml solution for 10s of minutes to hours, washed with water, and then air dried. Because poly-lysine may be toxic, care must be taken to thoroughly wash excess poly-lysine from the surface prior to drying. Due to the potential toxicity of poly-lysine, we have routinely employed poly-L-ornithine (0.1% solution) for 30 min.95 Surfaces coated with poly-L-ornithine need only be air dried. We add laminin (5 àg/ml), an extracellular matrix protein to the poly-L-ornithine prior to coating, to further promote adhesion to poly-L-ornithine-coated surfaces (the draw- back of the laminin is that it promotes neurite extension, thereby limiting the time during which neurons may be adequately voltage clamped after plating). Other charged molecules such as conconavalin-A are also employed with success.109 8.6.3.5 Obtaining a “Pure” Neuronal Culture
Sensory ganglia contain several different cell types in addition to neurons, and thus, obtaining a “pure” neuronal culture may be difficult. We have routinely avoided attempting to obtain a pure neuronal culture because non-neuronal cells appear to provide substances (such as nerve growth factor, NGF) that help maintain healthy neuronal cultures. However, if it is necessary to obtain a pure culture, there are several techniques that may be useful to clean up the neurons. Probably the most effective technique we have found is pre-plating the neurons. That is, following dissociation, cells are plated onto poly-L-ornithine-coated tissue culture plates and left to adhere for 2 to 4 h. The plating medium is then collected along with medium used to wash plates several times. This medium is centrifuged and the pellet is re- suspended and plated again onto fresh poly-L-ornithine/laminin-coated tissue culture plates. This second plating will contain primarily neurons, as the majority of non- neuronal cells will be left adhering to the first plates. Neuronal cultures may be further purified over time if mitotic inhibitors are included in the solution, as this will preferentially lead to the loss of non-neuronal cells. Of note for sensory neurons, it is preferable to use compounds other than cytosine arabinoside as a mitotic inhibitor, because this compound has been shown to inhibit the actions of NGF.110 8.6.3.6 Use of Trophic Factors
Whether or not to add trophic factors to the plating medium is becoming a more complicated question as the number of trophic factors increases along with our understanding of the various roles these compounds play in the physiology and
pathophysiology of the nervous system. Historically, investigators have routinely added NGF to culture medium in concentrations ranging between 5 and 5000 ng/ml.
Given our current knowledge of the role NGF plays in mediating phenotypic changes in sensory neurons,39,111–117 the amount of NGF added to the culture medium may clearly influence the electrophysiological properties of the neurons studied. More recently, glial-cell-derived neurotrophic factor (GDNF) has been shown to control expression of several ion channels in a subpopulation of sensory neurons.118,119 Thus, at this time, the questions of which trophic factors to add and at what concentrations have yet to be clearly answered. If neurons are studied soon after dissociation it is probably fair to argue that the impact of trophic factors (whether added or omitted) is minimal. However, if longer culture times are to be used, the impact of specific culture conditions, including the presence or absence of trophic factors, will have to be determined with respect to the specific properties of interest.
Many of the issues discussed above with respect to the dissociation and plating of sensory neurons also apply to the dissociation and plating of CNS neurons.
However, there are several additional issues that must be considered.
8.6.3.7 Time in Culture
Many CNS neurons do not survive when plated on typical substrates such as lysine or ornithine. Consequently, investigators who wish to study isolated dorsal horn or brain- stem neurons must study neurons within hours after dissociation and plating,70,71 or plate such neurons on support cells.101,120–122 Astrocytes grown to confluency appear to provide a suitable substrate to maintain CNS neurons in culture.101 Eckert and colleagues provide a detailed approach to the isolation and maintenance of astrocytes.101 8.6.4 UNIQUE APPLICATIONS
MacDermott and colleagues120–123 have studied dissociated sensory and dorsal horn neurons in combination through an approach that has enabled them to characterize, at an electrophysiological level, events controlling the release of transmitter from the central terminals of primary afferent neurons. To create a situation in vitro where one or two sensory neurons form synapses on one or two dorsal horn neurons, these investigators establish micro-islands of dorsal horn neurons. This technique was originally developed by Segal and Furshpan124 for the study of hippocampal neurons and was subsequently adapted by MacDermott and colleagues.121 To obtain micro- islands, cover-slips are first coated with poly-D-lysine and then dipped in 0.5%
agarose and air dried. Dry cover-slips are then sprayed with rat tail collagen (2 mg/ml in acetic acid) with an atomizer. Cover-slips are sterilized and then plated with astrocytes. The astrocytes will adhere to the collagen, thus forming the basis of the micro-island. From 3 to 7 days after plating the astrocytes, dorsal horn and DRG neurons are plated on top of the astrocytes at a density of 10 to 30 K dorsal horn neurons and 30 to 50 K DRG neurons per dish. Once established, it is possible to record from pairs of dorsal horn and sensory neurons using the electrophysiolog- ical responses of the dorsal horn neuron as an assay with which to assess events controlling release of transmitter from the sensory neuron.121–123