In this section we discuss general aspects of obtaining, slicing, and maintaining the human tissue for both neocor-tical and hippocampal specimens. In addition, we will consider some of the ethical and health-hazard issues involved in handling human tissue.
Obtaining Human Brain Tissue from Epilepsy Surgery Resections
Human tissue samples in most cases result from surgical procedures aimed at treating focal, pharmacoresistant epileptic disorders. These samples derive from partial temporal or extratemporal lobectomy, excision of lesions, or selective hippocampectomy. Rarely, complete lobectomies or functional or anatomic hemispherectomies are undertaken, during which tissue for experimental purposes can be obtained. In some studies, specimens from tumor resections have been investigated (Lücke et al., 1995; Patt et al., 1996; Straub et al., 1992; Williamson et al., 2003).
Location and Source of the Tissue and Its Relationship to the Epileptogenic Zone
It is critical to obtain a detailed record of precisely where the resected tissue comes from. In the case of neocortical resections, this is most easily accomplished with a diagram or photograph showing the resection site. When tissue is removed from deeper structures (e.g., hippocampus), it is useful to get a verbal description from the surgeon regarding the location and extent of the resection. If the tissue is to be divided (e.g., with some going to the clinical pathologist), it is important to maintain a record of which part of the tissue is retained for the slice experiment. In all cases a key piece of information is how the resected tissue is related to the epileptogenic zone. When electrocorticography has been carried out as part of the surgery, electrode positions should be documented relative to the cortical surface so that spiking and nonspiking areas can be identified (Avoli and Olivier, 1989; Colder et al., 1996; Williamson et al., 1995; Schwartzkroin et al., 1983). If electroencephalographic (EEG) monitoring has been carried out before the surgery (e.g., with implanted strips or grids), an attempt should be made to relate the resected tissue to the site(s) of seizure initiation.
Clinical, Imaging, and Pathology Background Data
A record of presurgical clinical workup is also helpful. The investigator must know patient age and sex, the onset of seizure activity (at what age and for how many years), the clinical seizure diagnosis (seizure type, frequency), and current and past medications. Also useful is information from magnetic resonance imaging (MRI) or other imaging procedures (especially if that helps to localize the resection with respect to abnormal structure and/or function). The pathology report based on the resected tissue is also important and may complement the investigator's own histology efforts.
Although surgical outcome is not necessarily a determinant of whether the resected tissue is actually "epilepto-genic," it is helpful to learn whether the surgical procedure led to cessation of seizures. Presumably, if the answer to this question is yes, the investigator can conclude that the resected tissue contributed to the patient's epileptic state.
The general approach to obtain the tissue is similar across laboratories. As the patient is undergoing surgery, the investigator sets up a tissue-processing station in or adjacent to the operating room. It is useful for the experimenter and the surgeon to have conferred about the nature of the resection beforehand so that the surgeon can deliver the requested samples without the need for discussion at the time of surgery. Tissue samples (often more than one) are excised and immediately placed in a beaker filled with ice-cold
(4° C) artificial cerebrospinal fluid (ACSF) saturated with a 95% O2/5% CO2 gas mixture (pH 7.4). The composition of the ACSF, as with animal slice experiments (Dingledine, 1984), varies slightly from group to group. A basic ACSF solution consists of (in mM): NaCl, 124-129; KCl, 2-4; CaCl2, 1.6-2.4; MgSO41.3-2, Na^PO4 1.24-1.25, NaHCOs 21-26 and glucose, 10 (e.g., Schwartzkroin et al., 1983). Instead of NaH2PO4, KH2PO4 can be used; in this instance, the KCl concentration should be 2mM. Some groups (e.g., Kivi et al., 2000; Kohling et al., 1998b) have used one or more of the following variations as protective solution for transport and slicing procedures of the tissue: (1) 0.1 mM a-tocopherol as radical scavenger; (2) high-Mg2+ concentration (addition of 2mM MgCl2) to reduce excitability; (3) low Na+ concentrations (omitting NaCl completely and replacing it with 200 mM sucrose) to reduce neuronal toxicity; and (4) low (1 mM) CaCl2 levels to protect against hypoxia-induced depolarization and Ca2+-mediated cytotoxicity.
Tissue handling, from its removal from the patient through all experimental manipulations, must be carried out in accordance with the institutional review board policies for experiments with humans (or human materials). In addition, appropriate precautions should be taken to protect the experimenters from human pathogen exposure. Although rare, the danger of exposure to tissue containing pathogens (e.g., prions associated with Jacob-Creutzfeldt disease) warrants special precautions. Thus it is important not only to adhere to common-sense laboratory practices (e.g., use of gloves, eye shields, etc) but also to clean and disinfect all surfaces exposed to the tissue. It is also highly recommended to have a separate set of instruments (and a separate slice chamber) devoted to human studies.
Tissue samples may come from the neurosurgeon in small blocks (of approximately 1 cm3) or as larger resections (e.g., en bloc hippocampus). If the tissue is to be used for functional (e.g., electrophysiologic or imaging) studies, it should be rapidly sliced into thinner sections (maximally 600 mm in thickness) to ensure proper oxygen supply by diffusion from the surrounding ACSF. In all cases, special care should be taken to slice the block of resected tissue as quickly as possible, maximally within 5 minutes after excision. Preferably this tissue preparation is carried out at a site adjacent to the operating room.
Coronal slices are cut with a conventional chopper (e.g., Mclllwain type) or with a vibratome (e.g., Campden, Leica, TSE, WPI, etc.), using the same procedures as used for making slices from animal tissue. If network activity is to be investigated, thicker (450-600 mm) slices should be obtained to maintain a maximum degree of connectivity within the slices; for isolated neurones (see later), slices can be made thinner. In some cases, intense gliotic reaction will render the tissue hard and rubbery and make it difficult to cut through the tissue to make thin, even slices. In such cases, the experimenter is left to his or her own devices. It is usually the case, however, that such tissue will yield little useful electrophysiologic data because it will be difficult to penetrate with microelectrodes.
After slicing, the tissue can be transferred directly to the recording chamber, or it can be maintained viable at room temperature in carbogenated ACSF (placing slices on a nylon mesh immersed in a fluid-filled beaker). Good viability, particularly for longer transport of the tissue slices from the operating room to the laboratory, can be achieved by using a portable chamber with ACSF at 28° C (Kohling et al., 1996a). Such a chamber, illustrated in Figure 1, consists of multiple wells (with nylon-mesh bottoms) and a central funnel through which the carbogen-bubbled ACSF can rise. The fluid circulation stabilizes the slices on the nylon mesh. Recording chambers are identical to those used for animal tissue slices and may be of either a submerged type, with slices resting on the bottom of an ACSF; a perfused flow chamber; or a standard interface-type chamber with the tissue resting on a nylon mesh at the interface of humidified carbogen and ACSF.
Recognition of slice orientation within the tissue chamber is essential for carrying out interpretable experiments. If hippocampal slices are used, identification of the different hippocampal subregions is often visually possible, although the convoluted nature of the very anterior tip of hippocampus may make such identification difficult. In many cases, the hippocampus may show severe atrophy of areas CA3 and CA1 (Ammon's horn sclerosis, or AHS), it may be damaged from surgical procedures, or it may be intensely gliotic. These factors can determine which areas are accessible and appropriate for further investigation. In studying neocortical samples, it is useful to not only identify pial and white matter edges but also to estimate cortical layers.
Electrophysiologic Monitoring Field Potential Recordings
Extracellular recording of network activity can be performed in human tissue very much the same way as it is done in animal preparations. To obtain field potential recordings, low impedance (1-2MW) glass microelectrodes filled with ACSF usually are used. For special purposes (e.g., current-source-density analyses, which necessitate multiple electrode recordings at short equidistant positions), insulated, etched tungsten or platinum-nickel wire (diameter, 30-40 mm) or carbon fibers can be used (Cohen et al., 2002; Kohling et al., 1999; Louvel et al., 2001). The disadvantage of such electrodes is that that high-frequency signal components (>50Hz) are not detected because of the impedance
properties, and direct coupled (DC) recordings are not possible because of the polarizing properties of metal or carbon. Field potential recordings have been used to address several experimental issues in human tissue slices (see section on Characteristics of the Activity Generated by Human Epileptic Neurons).
The first intracellular studies in human tissue were performed with sharp-electrode recordings. For impalements of both neocortical and hippocampal cells, sharp microelec-trodes with a resistance of 30 to 150MW and filled with 2 to 4M of K-acetate or K-methylsulphate were used. Several experimental issues can be addressed with this technique:
1. Resting and passive properties of neurons. Most investigators have concluded that resting membrane potential, input resistance, and time constants of human neurons are not different from corresponding cells in animal preparations. Most laboratories failed to observe any conspicuous increase in intrinsic bursting; firing properties on current injection generally did not differ from region—specific
"controls" from rodents (or monkeys) (Avoli and Olivier, 1989; Foehring et al., 1991; McCormick and Williamson, 1989; Schwartzkroin, 1987; Schwartzkroin and Prince, 1976; Tasker et al., 1996). One exception to this generality is that Dietrich et al. (1999a) found that some dentate granule cells in AHS tissue appeared to display properties of hilar interneurones.
2. Spontaneous synaptic activity. Spontaneous activity was demonstrated to occur regularly in human slices, apparently without sufficient network synchronization to generate field potential discharges. Such activity, particularly in mesial structures, was found to be dependent on gluta-matergic and g-aminobutyric acid (GABA)ergic transmission (Knowles et al., 1992; Schwartzkroin and Haglund, 1986; Schwartzkroin and Knowles, 1984). An interesting finding was that in mesial temporal tissue where such network synchronization was seen, this spontaneous activity reflected bursting-discharging neurones in which there was a positively-shifted GABA reversal potential (and hence a depolarizing GABA response) (Cohen et al., 2002).
3. Synaptic activation by focal stimulation. For such studies, a stimulating electrode (monopolar or bipolar) is placed into pathways afferent to the recorded cell. In the hippocampus, these investigations have focused mainly on perforant path activation of granule cells (Dietrich et al., 1999a; Isokawa et al., 1991). In neocortical tissue, stimulation of underlying white matter or focal activation of tangential association fibers have been used (Avoli and Olivier, 1989; Avoli et al., 1994a; Schwartzkroin et al., 1983; Strow-bridge et al., 1992). These studies revealed graded evoked bursts and, in rare cases, paroxysmal depolarization shifts or very prolonged excitatory postsynaptic potentials with relatively weak inhibitory activity.
For further analysis of both synaptic responses and modulation of voltage-gated channels, voltage-clamp experiments have been performed in human tissue slices, either using switch-clamp amplifiers and sharp electrodes (McCormick and Williamson, 1989, Wuarin et al., 1992), or patch-clamp electrodes (2-4MW, with Lucifer yellow filling for cell identification (Isokawa et al., 1997, Isokawa, 1998). These studies revealed the existence of pharmacologically isolated NMDA-receptor mediated currents and the existence of Mg2+-block, as known from animal experiments (Wuarin et al., 1992) and modulation of different K+ currents via acetylcholine, adenosine, and other modulators (McCormick and Williamson, 1989). Perhaps more importantly, NMDA-receptor mediated responses were particularly variable in amplitude, whereas GABA-mediated inhibitory potentials were reduced after high-frequency activation (Isokawa, 1998; Isokawa et al., 1997). These results underscore an increased excitability of the human epileptic hippocampus. Also, human astrocytes have altered properties (i.e., a larger proportion of AMPA-receptor flip variant, enhanced inward-rectifying K+ currents similar to juvenile rodent astrocytes, and even generation of slow action potentials) in sclerotic hippocampus (Bordey and Sontheimer, 2004; Hinterkeuser et al., 2000; Schröder et al., 2000).
As an extension of field potential recordings, the concentrations of ions in the extracellular space can be monitored using modified, double-barrelled micropipettes. The methods used to manufacture these electrodes have been described in detail in studies on animal tissue, where this technique has been in use for many years (Heinemann et al., 1977). Briefly, double-barrelled theta-glass capillaries are pulled (tip diameter, 2-6 mm), and one channel is backfilled with of NaCl solution equimolar to the ACSF to be used as field potential and reference channel. The tip of the other barrel channel's tip is silanized, filled with ion exchanger (typically Fluka 60398, 60031, or Corning 477317 for K+ and Fluka 21196, 21048, or 21191 for Ca2+) and backfilled with corresponding KCl or CaCl2 solutions (100mM). Recordings with these electrodes have shown that although basal levels of K+ are relatively normal in human neocorti-
cal or hippocampal tissue (Kivi et al., 2000; Kohling et al., 1998b), K+ spatial buffering can be disturbed, particularly in sclerotic hippocampus (Kivi et al., 2000a). High levels of K+ are also reached in human neocortical tissue when the tissue exhibits pathological function, such as spreading depression (Avoli et al., 1991; Gorji et al., 2001); however, these levels do not substantially differ from those seen in animal preparations undergoing similar treatments.
Another technique, again adapted from animal experiments, is the use of acutely isolated neurons from neocortex or hippocampal tissue. This approach has been used to investigate voltage-gated currents, which are difficult to be analyze in situ because of space-clamp problems. The methods of cell isolation are essentially the same as in animal tissue (see Chapter 2). First, slices are cut from the resected block and then microdissected to yield the areas of interest. These slabs are then incubated with proteases, for example, trypsin (type XI, 0.5mg/ml for 1-2 hours in ACSF at 29°C) or pronase (2-3mg/ml, 25 minutes, 28°C, in piperazine diethanesulfonic acid (PIPES)-buffered ACSF) and then washed with PIPES- or 4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid (HEPES)-buffered solution. To isolate neurons, the predigested tissue slabs are triturated through fire-polished pipettes and then incubated in appropriate solutions (Beck et al., 1996, 1997a, b, 1998; Cummins et al., 1994; Remy et al., 2003; Ruschenschmidt et al., 2004; Vreugdenhil et al., 1998, 2004).
These studies have characterized current properties in human cells and studied the mechanism of action of antiepileptic drugs. Although K+, Na+ and Ca2+ currents are essentially similar to those described in animal tissue, some peculiarities have been identified. For instance, human subicular cells possess a large, persistent Na+ current that may predispose them to bursting (Vreugdenhil et al., 2004); moreover, effects of carbamazepine on Na+ currents are reduced in cells from sclerotic, but not from nonsclerotic tissue (Remy et al., 2003, Vreugdenhil et al., 1998).
Optical imaging techniques can be employed in human tissue in vitro very much the same way as in animal preparations. Both voltage-sensitive dyes and intrinsic optical imaging can be used to monitor spatiotemporal patterns of network activity. For voltage-sensitive imaging, the styryl dye RH795 has already been employed (1 hour incubation in 12.5 mg/ml, 450-|mm- thick slices; (Kohling et al., 2000, 2002)); other fast dyes of the ANEP-type should also be informative. This method has the advantage of yielding high temporal resolution, but it allows for limited recording times because of phototoxicity effects. Studies using this technique have shown that spontaneous network activity goes along with heterogeneously distributed excitation within the neuronal network which it can be initiated in minimal foci up to 50 ms before activation of the rest of the network (Kohling et al., 2000, 2002). Furthermore, about 30% of the slices tested displayed epileptic responses after focal stimulation (Straub et al., 2003).
Intrinsic optical imaging makes use of alterations in reflectance and absorbance properties of brain tissue as a consequence of neuronal activity. These changes, as recorded in in vitro slices, presumably reflect modifications in the extracellular space, for example, as a function of K+ accumulation and cell swelling (Andrew et al., 1996, Holthoff and Witte, 1996; MacVicar and Hochman, 1991). To date the temporal resolution of this method appears to be somewhat limited; however, it does not necessitate any tissue manipulation other than illumination. Hence, it can be used to monitor spread of activity not only in vitro but also in the patient's brain in the operating room (as a diagnostic tool for guidance of tissue resection) (Haglund et al., 1992, Haglund and Hochman, 2004).
A large number of histochemical and immunochemical techniques have been used to characterize resected human tissue. It is beyond the scope of this chapter to describe such analyses, but they are a critical part of the study of human tissue and provide information that can help to interpret the electrophysiologic data (see chapter 50). The importance of identifying individual neurons from which intracellular data have been obtained is noteworthy. A number of dye injection techniques for cellular labeling have been used, including fluorescent dyes (e.g., Lucifer yellow) and tracers to be detected histologically (e.g., biocytin, neurobiotin). These dyes are added to the intracellular recording solution both in sharp and patch pipettes.
A number of nonelectrophysiologic techniques can be used to address changes of intrinsic excitability or synaptic physiology in human tissue. Although these methods do not directly show functional effects, they are nevertheless valuable for identifying the density of receptors or the abundance and localization of channel proteins. In many cases, and in contrast to functional studies, control tissue is available (e.g., from autopsy), so that direct comparisons between epileptic and nonepileptic tissue can be drawn. Receptor autoradiography using [3H]-tagged ligands have revealed up-regulation of AMPA-receptors in human epileptogenic neocortex (Zilles et al., 1999). Immunohistochemical exper iments with antibodies directed against specific GABA and metabotropic glutamate receptor (mGluR) subtypes have shown up-regulation of specific GABA and mGluR subunits (Lie et al., 2000; Loup et al., 2000). Finally, by employing a similar approach, Lie et al. (1999) discovered that Ca2+ channel b1 and b2 subunits are increased in sclerotic hippocampus.
Molecular methods can also provide information on the mRNA expression of receptors and channels in the human tissue. These studies involve the use of reverse transcription reaction followed by polymerase chain reaction (PCR) and by restriction enzyme assays. Such experiments have shown that the relative amount of edited RNA of the AMPA receptor GluR2 was significantly increased in the hippocampal tissue, whereas no changes were found in neocortical tissues from epilepsy patients (Musshoff et al., 2000) and that NMDA-receptor mRNA splicing was unchanged compared with autopsy control material (Vollmar et al., 2004). Similar changes have been found in neurones and in hippocampal astocytes, where real-time PCR has revealed an increase in flip-to-flop ratio of the GluR1 AMPA-receptor mRNA matching functional results (Seifert et al., 2004). An alternative approach involves expresson monitoring by micro-arrays (Becker et al., 2002, 2003). Using this method, investigators have found some genes are up-regulated in human epileptic hippocampus (ataxin-3 and glial acidic fib-rillary protein), whereas some are down-regulated (e.g., calmodulin) (Becker et al., 2002). In these studies, the findings were pinpointed to cell populations by using single-cell real-time PCR. Another application of molecular techniques to the analysis of ligand-gated currents in human tissue consists of injecting mRNA or cell membranes extracted from epileptic patients into frog oocytes (Palma et al., 2002, 2004). This procedure led to the expression of ionotropic receptors for GABAa, kainate, and AMPA. These investigators have reported that GABAA receptor-mediated currents in oocytes injected with "epileptic" mRNA or cell membranes are characterized by a strong run-down after repetitive ligand applications; this phenomenon can be abolished by phosphatase inhibitors (Palma et al., 2004).
Was this article helpful?