Alexplorer, B.S., M.Ed.
Proposal for the Degree of
MASTER OF SCIENCE
Dept. of Biological Sciences
UNIVERSITY OF NORTH TEXAS
Networks of cultured neurons exhibit spontaneous activity consisting of both single action potentials and rapid series of spikes termed bursts. When grown on micro-electrode arrays (MEAs), such activity can be monitored at multiple sites for extended periods. Recently, is has been demonstrated that such networks are pharmacologically histiotypic; i.e., they mimic the general responses of the parent tissue (Xia and Gross, 2003; Xia et al., 2003; Keefer et al. 2001).
As a consequence, platforms incorporating neuronal networks have been proposed as experimental platforms for pharmacological and toxicological studies as well as for use as tissue-based biosensors (Gross et al., 1995). As spontaneously active systems that display a rich repertoire of spatio-temporal patterns, networks on MEAs also represent test beds for systematic studies of internal network dynamics, pattern generation, and pattern processing.
One important target for study of these mechanisms are voltage-dependent calcium channels (VDCCs), a diverse family of transmembrane proteins involved in the regulation of cellular excitability and calcium homeostasis (Jimenez et al., 2000). Because the calcium concentration outside the cell is four orders of magnitude higher than inside, the control of this current via calcium channels in neurons continues to be an important area of research.
One class of VDCCs, the L-type, are especially prominent in cardiac and nervous tissue, although they are also found in skeletal and smooth muscle. The L-type calcium channel blockers (CCBs) verapamil and diltiazem, are widely prescribed for the treatment of hypertension, stroke, acute ischemia, migraine, and epilepsy (Jimenez et al., 1999). However, they also have effects on learning and memory in the central nervous system (Borroni et al., 2000)
In this study, MEAs will be employed to examine the influence of the L-type VDCC antagonists verapamil and diltiazem on the spontaneous activity of cultured neuronal networks. L-type Ca2+ currents have a high threshold (-10 mV) and are slow in their activation and inactivation (Squire et al., 2003). They give rise to persistent ionic currents that contribute to the generation of action potentials and may be involved in the establishment of plateau potentials, which have a great influence on bursting.
Significance: The significance of the expected results lies in a better understanding of network activity and its dependence on Ca2+ channels. Such data may allow pharmacological manipulations that provide more stable and quantifiable network patterns. Stable patterns greatly facilitate pharmacological and even toxicological investigations and appear to be essential for neuronal network applications as biosensors. In all application areas, network changes are compared to a native (reference) activity states where a high pattern variability contributes to a reduced system sensitivity, especially at low concentrations of neuroactive agents. Finally, the acceptance of neuronal networks on MEAs as viable platforms for pharmacology, toxicology, biosensors, and even for basic neurobiological investigations requires validation experiments that allow quantitative comparisons of networks responses in culture with those reported from animal experiments. This study will also provide such validation data.
This purpose of this research is to study the effects of the dihydropyridines verapamil and diltiazem on network activity for the purpose of (a) system validation via comparisons to responses in vivo, and (b) manipulation of network bursting to provide greater temporal stability to network burst patterns.Specific Aims:
SA 1. To determine the main features of network activity (such as spike and burst production, burst duration, burst periods, and spikes in bursts) affected by diltiazem and verapamil.
Rationale: A qualitative dependence of bursting patterns (i.e., burst duration and burst period) in cell culture on calcium currents has been established (Rhoades and Gross, 1994). Such a dependence also is generally accepted for in vivo systems. This dependence needs to be quantified as a function of dihydropyridine concentration.
SA 2. Establish whether tissue-specific effects are generated by these two dihydropyridines by examining and comparing responses from frontal cortex and spinal cord cultures (i.e., whether a particular tissues is more significantly influenced).
Rationale: Tissue specificity in culture is emerging as an important network response feature. This validity of this system can be further assessed via comparison to in vivo data on tissue specificity.
SA 3. Investigate culture sensitization or desensitization by a quantitative comparison of responses to single applications and sequential, small step applications.
Rationale: The method of application can influence network responses and corrupt EC50 values. Although observed for many compounds, it has never been studied quantitatively. This specific aim will help develop a methodology for compounds that induce rapid changes in receptor sensitivity.
Neuronal Cell Cultures
Frontal cortex and spinal cord tissue are cultured by the CNNS staff onto MEAs from embryonic 14-day Hsd:ICR mice. Frontal cortex cultures are grown in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 5% horse serum approximately 4 weeks after seeding; spinal cord in Minimum Essential Medium (MEM) supplemented with 10% horse serum, then transferred to 5% serum. Medium changes are performed twice a week in which 50% of the medium is removed and replaced with fresh stock. Incubators contain an atmosphere of 90% air and 10% CO2 maintained at 37°C. Cultures under these protocols are routinely regarded as mature 3-4 weeks following seeding on MEAs (Fig. 1). MEAs are fabricated in-house by CNNS technicians. Techniques for fabrication have been described previously (Gross, 1979; Gross et al., 1985; Gross & Kowalski, 1991).
|Figure 1: Cell culture procedures for spinal cord tissue (similar procedures are followed for frontal cortex tissue; see Materials and Methods). (Figure created, 1999)|
MEAs with cultures are placed in stainless steel chambers (Gross and Schwalm, 1994; Fig. 2) and maintained at 37°C on an inverted microscope stage. pH of the medium is maintained at 7.4 by applying a constant stream of humidified air also containing 10% CO2 through the cover of the chamber. The top of the chamber cover is composed of glass coated with indium tin oxide heated by passing current through it, thus preventing the formation of condensation and allowing continual observation. The chamber cover may be temporarily removed to apply drugs or to obtain samples of the medium. Additionally, medium changes may be accomplished via plastic syringes through ports in the chamber walls.
||Figure 2: Assembly of the open recording chamber. The assembled chamber is then placed on the stage of the inverted microscope and 32 pre-amplifiers are attached to either side (CNNS archives).|
Recording Set-up and Data Acquisition
MEAs are connected to two 32 channel preamplifiers which feed action potentials (spikes) to a Multichannel Acquisition Processor (MAP) system (Plexon, Inc., Dallas, TX). Each spike produced by these units is recorded to a .plx file as a time-stamp (indicating when the spike occurred). This system allows spikes to be visualized by RASPUTIN software (Plexon, Inc.; Fig. 3) which allows the selection of individual electrodes, the setting of total system gain (typically between 10k and 14k in this laboratory), and the discrimination of individual neuronal units via template-matching. Units are selected only if their signal to noise ratio (SNR) was greater than 2:1.
|Figure 3: Multi-channel activity display. In the left panel, the last 40 seconds of simultaneous network activity is shown. Coordination across channels is evident. On the right, unit discrimination is monitored across 32 channels.|
For the purpose of fulfilling the specific aims of this study two general experimental protocols will be employed: one for testing the effects of single applications of a drug, the other for exploring dose-response relationships across a series of applications for the purpose of developing a dose-response curve. In the latter case, cultures will first be washed with fresh medium, after which their activity allowed to stabilize. The activity recorded after the culture stabilizes will be used as the reference (i.e., control) activity. Generally this will run for 45 minutes to two hours. Following stabilization, a “low” concentration of drug will be applied (“low” defined as producing <20% change in network activity), and the network activity will again be allowed sufficient time to stabilize (i.e., activity remains level for a period of 30 minutes or more). Stable activity will then be recorded for approximately 45 minutes before the next application. Applications will be continued in this manner until activity is completely inhibited. The culture will then be washed with fresh medium from the same stock as that used during the “reference” wash followed by a second wash 10 to 15 minutes later in order to ensure that any residual, freshly dissociated molecules of the drug are removed.
For single drug applications, the same protocol will be followed except that the only application of the drug was at a concentration sufficiently high enough to produce a noticeable effect (e.g., inhibition greater than 50%). Because few individual applications will be required (compared with the protocol above), the duration of each episode will by extended to a length of several hours in order to better assess stability of the drugs effects across time.Data Analyses
Histograms of spiking and bursting activity as well as a time-series events across channels are viewed in NeuroExplorer (NEX; Plexon, Inc.; Fig. 4) to identify any units that may have “dropped out” (e.g., died or lost contact with the electrode). These histograms will also reveal the presence (or absence) of subpopulations with a different responses to the agents tested.
|Figure 4: Burst parameters identified on a simulated integrated recording of two bursts of action potentials. BA= Burst Amplitude, BP = Burst Period, BD = Burst Duration, and IBI = Interburst Interval.|
Data files will then processed by I-Burst (Jones et al., 2001), which derives a number of measures of network activity such as spike and burst rate, burst duration, interburst interval, etc. (Fig. 5). I-Burst stores these values in a series of files (one for each unit discriminated) which can then be exported to Excel via an in-house macro (Jimenez, 2002). This macro gives a population mean for each variable for each 60 second bin of time. The averaged measures of network activity (spikes/min, bursts/min, burst duration, burst period, etc.) are then plotted across time.
|Figure 5: Global response of a spinal cord culture to the application of an additional 3 uM of verapamil for a total of 9 uM (the 3 uM drug addition was made at the beginning of the plot; culture had been exposed to a total of 6 uM verapamil in preceding applications). For all units verapamil typically produces a similar effect on spike rate in terms of magnitude and latency until onset (Drug addition was made at the beginning of the plot). Graphs represent histograms of spikes per minute for approximately 67 minutes during the development of a dose-response curve. Maximum number of spikes varies by graph from 100 to 4000 spikes per minute. Those profiles that do not show response to the drug may be dominated by noise (note that these have a very low spike count. (from JB53h)|
Dose-response relationships will then be examined by generating dose-response curves in Microcal Origin (v.6.0) using the average of the period of stable activity following each application of the drug. IC50 values are determined automatically from curve fitting methods. Tests for statistical significance will be performed in SPSS (v.10.1).
Effects on main measures of activity
Both verapamil and diltiazem inhibit spiking and bursting activity in both FC and SC cultures (Fig. 5). However, inhibition is too broad a term for the specificity of these agents’ effects. Specifically, the reduction in the overall activity of the network is the result of changes to the burst period and burst duration. That is, bursts are reduced in their duration and in the frequency of their occurrence in a dose-dependent manner (Fig. Figs. 7 and 8). Interestingly, networks remain coordinated across units even when the "bursts" on the active channels were reduced to as little as one or two spikes each.
|Figure 6: Representative dose-response relationship between verapamil and spiking and bursting activity in a spinal cord culture. Duration of experiment was 370 minutes; n = 55 units. Digits at top indicate cumulative concentration of verapamil applied (uM). Fresh medium was applied before the beginning of the experiment and at the end of at the points indicated by "W" ("wash"). s|
|Figure 7: A dose-response relationship between diltiazem and burst duration and burst period demonstrated in a spinal cord culture. Bursts diminish in duration while the period from one burst to the next steadily expands with increasing concentrations of the agent until no bursts can be identified in the remaining activity.|
|Figure 8: Activity of a spinal cord culture under the influence of serial additions of verapamil. Several features of the activity change in a dose-dependent manner. Overall, spikes and bursts are less frequent with the bursts growing shorter in duration with greater intervals of time between each successive burst. (data from JB38)|
Both verapamil and diltiazem appear to inhibit network activity in both FC and SC tissue across all units equally in terms of delay before onset of effect, magnitude of depression and partially recovery, and time to achieve a stable, inhibited plateau of activity (Fig. 4). No subpopulations with alternate response patterns have been found (e.g., excitatory, disinhibitory, etc.).Stability of the effect of L-type VDCC antagonists
After application of the drug, activity levels remained nearly constant for spike and burst rate (Fig. 9) as well as burst period, duration (Fig. 10), and integrated amplitude for considerable lengths of time (300 mins = max. duration tested) indicating that that these agents were not being appreciably internalized nor broken down.
|Figure 9: The effect of a single application of 8 uM verapamil on the spike and burst rates in a spinal cord culture.|
|Figure 10: The effect of a single application of 8 uM verapamil on the burst duration and period in a spinal cord culture. Units for BD and BP are in seconds.|
Cultures followed the same response trend when a second dose-response series was attempted (Fig. 11), indicating that the effects of these drugs are fully reversible in the short term and have no lasting effects on later responses.
|Figure 11: Intra-culture comparison of network response to diltiazem and verapamil performed on a spinal cord network (46 div). Key: N = native activity; W = wash with fresh medium; BCC = 40 uM bicuculine; # = did not record for 121 minutes. All other values are uM diltiazem or verapamil as indicated. Note that verapamil is more potent at shutting off the network.|
Efficacy and Potency
Both verapamil and diltiazem have been able to completely inhibit network activity in cultures of both FC and SC tissue (fewer than 10 spikes/min across all channels). However, verapamil was several times more potent than diltiazem at achieving complete inhibition.Single vs. Multiple Drug Applications
A routine observation in our laboratory is that cultures typically respond more strongly to a drug applied all at once than to the same concentration achieved cummulatively through a series of applications. Early data indicate this is also the case with verapamil.
Borroni AM, Fichtenholtz H, Woodside BL, Teyler TJ. (2000) The role of voltage-dependent calcium channel long-term potentiation (LTP) and NMDA LTP in spatial memory. J Neuroscience 20:9272-6.
Gross, GW. (l979) Simultaneous single unit recording in vitro with a photoetched laser deinsulated gold multi-microelectrode surface. IEEEE Trans. Biomed. Eng. BME-26: 273-279.
Gross, GW, Wen W, Lin J. (l985) Transparent indium-tin oxide patterns for extracellular, multisite recording in neuronal cultures. J. Neurosci. Meth. 15: 243-252.
Gross GW, Kowalski JM. (1991) Experimental and theoretical analyses of random network dynamics. In: Neural Networks, Concepts, Application and Implementation, Vol. 4 (Antognetti and Milutinovic, eds.), Prentice Hall, N.J., pp 47-110.
Gross GW, Harsch A, Rhoades BK, Gopel W. (1997) Odor, drug and toxin analysis with neuronal networks in vitro: extracellular array recording of network Responses. Biosens Bioelectron 12:373-93.
Gross, GW & Schwalm, FU. (1994) A closed chamber for long-term electrophysiological and microscopical monitoring of monolayer neuronal networks. J. Neuroscience Methods 52: 73-85.
Jimenez C, Bourinet E, Leuranguer V, Richard S, Snutch TP, Nargeot J. (2000). Determinants of voltage-dependent inactivation affect Mibefradil block of calcium channels. Neuropharmacology 39:1-10.
Jimenez, M. (2001) IBurst Macros. (In-house software.)
Jones, CV, Raymond D, Weil J. (2001) IBurst. (In-house software.)
Keefer EW, Gramowski A, Gross GW. (2001) NMDA receptor-dependent periodic oscillations in cultured spinal cord networks. J Neurophysiol. 86:3030-42
Rhoades BK, Gross GW. (1994) Potassium and calcium channel dependence of bursting in cultured neuronal networks. Brain Research 643: 310-318.
Squire, Bloom, McConnell, Roberts, Spitzer and Zigmond eds. (2003) Fundamental Neuroscience; 2nd ed;
Xia Y, Gopal KV, Gross GW. (2003) Differential acute effects of fluoxetine on frontal and auditory cortex networks in vitro. Brain Research 973:151-60.
Xia Y, Gross GW. (2003) Histiotypic electrophysiological responses of cultured neuronal networks to ethanol. Alcohol 30:167-74.
|Back to the index|