Monitoring neurotransmitter amino acids in vivo by microdialysis with on-line flow gated capillary electrophoresis

MISSING IMAGE

Material Information

Title:
Monitoring neurotransmitter amino acids in vivo by microdialysis with on-line flow gated capillary electrophoresis
Physical Description:
x, 182 leaves : ill. ; 29 cm.
Language:
English
Creator:
Witowski, Steven Richard, 1971-
Publication Date:

Subjects

Subjects / Keywords:
Brain microdialysis   ( lcsh )
Capillary electrophoresis   ( lcsh )
Amino acids -- Analysis   ( lcsh )
Chemistry thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Chemistry -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 2000.
Bibliography:
Includes bibliographical references (leaves 167-181).
Statement of Responsibility:
by Steven Richard Witowski.
General Note:
Printout.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 025876844
oclc - 47102984
System ID:
AA00020441:00001

Full Text











MONITORING NEUROTRANSMITTER AMINO ACIDS IN VIVO BY
MICRODIALYSIS WITH ON-LINE FLOW GATED CAPILLARY
ELECTROPHORESIS













By

STEVEN RICHARD WITOWSKI


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2000














ACKNOWLEDGMENTS


I would like to thank my research advisor, Dr. Robert T. Kennedy, for his

scientific support. I also acknowledge Dr. Thomas Vickroy for his help with the in vivo

aspects of this work, both procedure and design. In addition I would like to thank the

Kennedy research group, both past and present, for informal discussions that solved more

problems than could be counted.

















TABLE OF CONTENTS

page

ACKN OW LED GM EN TS.......................................................................... ii

LIST OF ABBREVIATION S........................................................ ...........vi

A B STR A C T .......................... ................... .. .......... ............. ...... ... ........ix

CHAPTERS

1 IN TRO DU CTIO N ................. ............................ ................... ........ .... I1

M icrodialysis .................................................................. .............. 2
Analysis of Dialysate ................................................... .............. ........... ...5.
O n-Line Analysis........................................ .... ............................... ... .6
Glutatmate and Aspartate in the Brain...................................... ....... .... 8

2 TEMPORAL RESOLUTION IN MICRODIALYSIS.................... ..... ..... 14

Intro ductio n .................. ............................... .................... ............... 14
E xperim mental .................................. ............ ................. .... .... ........ ... 15
R results ................................................................................ .... ..... 19
D discussion ............................................................. ............... ...22

3 IN VIVO MONITORING OF GLUTAMATE: THE ROLE OF
ASCORBIC ACID IN THE BRAIN....................................... ..... ..... 34

Introduction.................................................................. ...............34
Experim mental .......................................... ........ ......... ... ........... ...35
Results .................. ....................................... ............ ........... 39
D iscussion..................................................................................... 4 1

4 CHARACTERIZATION OF METABOTROPIC RECEPTOR
CONTROL AND UPTAKE ON GLUTAMATE AND
ASPARTATE IN THE HIPPOCAMPUS.................................. ..... ..... 46

Introduction ...................................... .. ....... .... ........... .. ... ........... 46
Experim mental .............................................. ....................... ...........48












R eDiscu ssion ...... ............................................................... ........ ...... 50
Discussion ................................................................... 54


5 REGIONAL DIFFERENCES IN THE STRIATUM AND
HIPPOCAMPUS: EVIDENCE FOR VOLUME
TRANSMISSION............................................................

Introduction .............. ....................................................
E xperim ental.......................... .......................................
R esu lts .........................................................................
D discussion ....................................................................

6 THE ROLE OF GLUTAMATE IN THE PILOCARPINE MODEL
OF SEIZURE....................... ..........................................

Introduction ..................................................................
Experimental......................... .........................................
R results .........................................................................
D iscussion....................................................................

7 DEVELOPMENT STRATEGIES FOR THE HIGH SPEED
CHROMATOGRAPFHIC SEPARATIONS...............................


............... 69

............... 69
............... 70
............... 7 3
............... 76


............... 87

............... 87
............... 9 1
............... 9 5
.... ......... 10 1


............. 118


T heory ........................................................................... . .... 119
Non-Porous Particles............................................... .......... ....... 120
Capillary Columns ................................................... ..................... 126
M onoliths.................................................................................... 128
Temperature .................................................................................... .. 130
Electrochrom atography....................................................... ........... 133

8 COMPARISON OF PRESSURE AND
ELECTROOSMOTICALLY DRIVEN FLOW IN CAPILLARIES
PACKED WITH NON-POROUS PARTICLES FOR HIGH-
SPEED SEPARATIONS.................................................... ......... 137

Introduction .................................................................. ............. 137
Experim ental.............................................. ....... ....... ...... ........... 138
R esu lts ......................... ............................................. ... ............ 142
D discussion ..................................................................... .... .......... 146

9 SUMMARY AND FUTURE DIRECTIONS.......................................... 152

Sum m ary .................................................................................. 152
Future D directions ............................................. .. .......................... 156


iV











APPENDICES

A TROUBLESHOOTING THE INSTRUMENT............................ 160

B MICRODIALYSIS PROBE CONSTRUCTION ............... .............162

REFEREN CES................................................................................. 167

BIOGRAPHICAL SKETCH ......................................................... ........182


.1
















LIST OF ABBREVIATIONS


AA ascorbic acid

AChE acetylcholinesterase

ACPD (IS, 3R)- l-aminocyclopentane-trans- 1,3-dicarboxylic acid

aCSF artificial cerebral spinal fluid

ALS amyotrophic lateral sclerosis

AMPD 2-amnino-2-methyl- 1,3-propanediol

AP anterior-posterior

Asp aspartate

BME 0-mercaptoethanol

CE capillary electrophoresis

CEC capillary electrochromatography

D diffusion coefficient

Deff effective diffusion coefficient

dp particle diameter

DV dorsal-ventral

EAA excitatory amino acid

EAAC excitatory amino acid carrier

EAAT excitatory amino acid transporter

EEG electroencephalography









EGTA bis-(aminoethyl)glycoether-N,N,N',N'-tetraacetic acid

EOF electroosmotic flow

GABA y-aminobutyric acid

GLAST glutamtae -aspartate transporter

GLT glutamate transporter

Glu glutamate

h reduced plate height

HV high voltage

IAA iso-ascorbic acid

i.d. inner-diameter

iGluR ionotropic glutamate receptor

i.v. intravenous

k' capacity factor

% tortuosity factor

LC liquid chromatography

LIF laser induced fluorescence

LSOP 1-serine-o-phosphate

MCPG (RS)-a-methyl-4-carboxyphenylglycine

mGluR metabotropic glutamate receptor

ML medial-lateral

v reduced velocity

NDA naphthalene dicarboxyaldehyde

NT neurotransmitter










o.d. outer diameter

ODS octadecylsilane

OPA ortho-phthaldialdehyde

PDC L-trans-pyrrolidine-2,4-dicarboxylic acid

PI polyphosphorinositide

PILO pilocarpine

PMT photomultiplier tube

p column to particle diameter ratio

TAPS N-tris[hydroxymethyl]methyl-3-amino propane sulfonic acid

TTX tetrodotoxin

u linear velocity

UV ultraviolet















Abstract of Dissertation Presented to the Graduate School
Of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of the Doctor of Philosophy

MONITORING NEUROTRANSMITTER AMINO ACIDS IN VIVO BY
MICRODIALYSIS WITH ON LINE FLOW GATED CAPILLARY
ELECTROPHORESIS

By

Steven Richard Witowski

December 2000

Chairman: Robert T. Kennedy
Major Department: Chemistry

Microdialysis is a sampling technique based on the diffusion of analyte molecules

across a semipermeable membrane and has been widely used to probe brain chemistry.

This method has previously been coupled to on-line flow gated capillary electrophoresis

for the analysis of the excitatory amino acid neurotransmitters glutamate and aspartate in

the rat striatum with temporal resolution of 7 to 14 s. By eliminating system dead

volumes, resolution has improved to 3.5 s and is in fact limited by the microdialysis

probe itself These improvements have allowed the study of smaller, more complex brain

regions while maintaining similar or improved performance characteristics.

Electrical stimulation of neuronal paths was used to elicit extracellular changes in

glutamate and aspartate to study their regulation in both the striatum and hippocampus.

A possible role for ascorbic acid in the brain is to heteroexchange with glutamate during

reuptake by glutamate transporters. In heteroexchange, as glutamate is uptaken into a










cell, ascorbic acid is released into the extracellular space. Electrical stimulation of the

prefrontal cortex with monitoring in the striatum shows that when ascorbic acid is added

to the extracellular space via reversed dialysis both basal and stimulated release of

glutamate increase. This effect is not seen with the addition of iso-ascorbic acid, which

has the same anti-oxidant activity but is not transportable.

A characterization of metabotropic glutamate receptors and uptake in the

hippocampus by electrical stimulation of the perforant path shows that there is weak tonic

autoreceptor control and strong uptake. These results are in contrast to previous results in

the striatum where there was strong autoreceptor control that is normally inactive with

less uptake. To determine the relative overall clearance rates in striatum and

hippocampus, exogenous D-aspartate was introduced into each region. It was found that

diffusion played a greater role in clearance in the striatum as compared to hippocampus.

Taken together with the data on receptor activity, it was concluded that volume

transmission was more likely in the striatum than hippocampus.

The role of glutamate in seizure generation and maintenance is unclear. The

effects of the seizure agent pilocarpine on both glutamate and aspartate in the

hippocampus in chloral hydrate anesthetized rats were studied. Different treatments with

pilocarpine (intravenous injection or direct infusion) produced different effects on

glutamate and aspartate. Direct infusion with a lithium pretreatment consistently resulted

in a decrease in both transmitters. The infusion of pilocarpine during electrical

stimulation inhibited the immediate change in glutamate seen in control stimulations but

produced a delayed and prolonged change after 20 min of infusion. A change in

extracellular Glu is not seizurogenic, but pilocarpine does affect Glu neurons.













CHAPTER 1
INTRODUCTION


Early analyses in the brain involved dissection of the region of interest followed

by time consuming extraction and purification steps. These measurements offer

snapshots of the overall analyte concentration in the tissue sampled. The source of the

chemical change can be unclear and may be from any number of processes including

intracellular stores, metabolic processes, trauma from the dissection, or a direct result of

the process of interest. Additionally, the analyte may be diluted or lost during the

extraction and purification steps.

With the advent of small carbon paste electrodes, it became possible to monitor a

chemical in the brain of a living animal (KIS73). The first of such experiments

monitored ascorbic acid which is easily oxidized and at a high concentration in the brain

(and therefore easily measured). As the analytical methods improved, it became possible

to measure different and more varied compounds via electrochemistry. Addition of a

negatively charged polymer (nation) coating allowed for the analysis of catacholamines

(GER84). Fast scan voltammetry with carbon fiber microelectrodes can measure changes

in dopamine on the ms time scale (EWI82, WIG76). Additionally, the use of selective

enzyme coatings allows the measurement of non-electroactive compounds such as

glutamate (Glu), glucose, lactic acid, ethanol, proteins, and drugs and their metabolites

(see RUZ96, TR095, VAD92 for reviews).








2

The main advantages of electrodes are that they are small, and can therefore

monitor specific brain locations, and that they offer very high temporal resolution, how

often a measurement can be made. Glu sensitive electrodes have anywhere from I min

(KUL99) to 1 s (HU94) temporal resolution while carbon fiber microelectrodes have

millisecond resolution.

Unfortunately these electrodes have several drawbacks. The first is that there is a

background signal associated with the measurements. Because of this background it is

difficult to measure a basal concentration of the analyte and low levels or small changes

may be obscured. Another disadvantage of electrodes is that one is never sure if only the

analyte of interest is being measured. A change in signal may be the result of a change in

background current or a change in the concentration of a similar chemical. For example,

with Glu selective electrodes, there are small peptides such as Glu-Glu or glutathione

(Glu-Cys-Gly) that may interact with the electrode; in addition, the total amino acid

concentration (mM range) in the extracellular space of the brain may affect the response

of the electrode. Also, an electrode can monitor only one chemical even though several

may be changing and interacting with each other.



Microdialysis

An alternative to electrodes is microdialysis (ROB91). Microdialysis is a

sampling technique in which small molecules diffuse across a semipermeable membrane

(typically regenerated cellulose or polycarbonate) based on concentration gradients into a

continuously refreshed medium called the perfusion fluid; see Figure 1-1. Probes can be

implanted into virtually any part of the body including skin, internal organs, veins, and of










interest here, the brain. Because microdialysis is only a sampling technique, any

detection scheme can be employed to monitor a single compound or several compounds.

The type of molecule that can be sampled depends on the membrane, usually

regenerated cellulose or polycarbonate. Typical probes have molecular cutoff ranges of

anywhere from a few thousand to one hundred thousand Daltons, meaning that molecules

with a molecular weight below the cutoff will be able to diffuse across the membrane.

One distinct advantage of the technique is that dialysate is considered to be a clean

sample. The technique provides samples free of tissue or blood cells that often require

significant steps to remove. As the molecular weight cutoff decreases, more molecules

will be excluded. A low molecular weight cutoff membrane will provide a sample free of

enzymes and proteins that may interfere with or degrade the analyte of interest. A rule of

thumb is the molecular weight cutoff should be at least 10 times greater than the

molecular weight of the analyte.

The amount of analyte that will diffuse across the dialysis membrane is dependent

on tip length of the probe, flow rate of the perfusion fluid, diffusion coefficient of analyte

in tissue, and type of membrane (BUN90). The most common measurement of this

process is relative recovery, which is defined as the concentration of analyte in the

dialysate divided by the true concentration of analyte in the sample and is usually

expressed as a percentage. For example, with a 30% recovery a 10 PM analyte in the

brain would be 3 pM in dialysate. Just as a molecule can diffuse into the probe, it can

also diffuse out of the probe. This technique, called reversed dialysis, is used commonly

to deliver a drug or other chemical directly to the brain. In this case instead of recovery,

the term extraction is used. Under steady state conditions recovery and extraction are








equal. The absolute, or mass, recovery is defined as the total mass of analyte collected

over a set period of time (for example: 20 pmol/UgL/20 min). As flow rate increases,

absolute recovery increases while relative recovery decreases. A shorter tip length will

decrease both absolute and relative recovery.

Microdialysis probes come in three general shapes. The first, termed in-line or a

dialysis tube (Figure 1-2A), is in the shape of a tube with a section replaced by the hollow

fiber. This type of probe is used to sample large areas typically under the skin or in an

internal organ. The probe is implanted in the region with a needle and passed completely

through the skin or organ. The second style of probe, a loop probe (Figure 1-2B), is

similar to the in-line probe described above, but in this case instead of passing through a

region, the probe is folded so that the inlet and outlet are side by side. The sampling

region is smaller than the first type of probe described; these probes are generally used in

the brain. The last type of probe is called a side-by-side probe (Figure 1-2C). Here, two

smaller capillaries are sleeved inside the dialysis fiber. One end of the fiber is sealed.

Perfusion fluid enters the membrane at the bottom of the probe and leaves through the

other capillary. The side-by-side probe is a simple version of the concentric probe

(Figure 1-2D) in which the inlet line is sleeved inside the outlet. The ends are

constructed so that the inlet and outlet are separate. The concentric probe is difficult to

construct and is therefore often replaced by the side-by-side probe.

Each of the three types has its advantages. The dialysis tube, as mentioned above,

is useful for sampling of large areas such as skin and offers recoveries close to 100%

because of its size. The loop probe is smaller than the in-line probe and therefore is used

to sample smaller regions. Its drawback is that while its length is smaller, its diameter is








5

double that of the others and may cause more damage to the surrounding tissue. The side

by side probe is the smallest of the three and should cause the least amount of damage.

The drawback to this probe is that because of its size, it has poorer recoveries than the

other two. The selection of a probe should be based on the specific experimental

parameters accounting for all the factors discussed above.

The advantage of microdialysis is that because of the associated separation,

multiple analytes can be measured, possible interferent problems are removed, and basal

levels can be determined. Additionally, extremely sensitive detectors can be utilized to

monitor very low levels of analyte. The major disadvantage of microdialysis is that

because of requirements of the separation or detection method, temporal resolution is

poor. It will be the focus of this dissertation to achieve electrode-like response with

microdialysis and to use the technique to monitor brain activity.



Analysis of Dialysate

The most common method to analyze dialysate is by liquid chromatography with

either electrochemical or UV detection (see DAV99, TON99, DAV97, VAN95 for

reviews). Liquid chromatography is used most often because of the variety of developed

separation methods and its ruggedness. Unfortunately it is not selected because it is

highly amenable to microdialysis. Conventional LC columns of 4.6 mm inner diameter

(i.d.) typically require 20 to 30 gL of sample for injection which corresponds to a

sampling time of 20 to 30 min with a dialysis flow rate of I UL/min. While this sampling

rate is adequate for monitoring slow changes such as drug effects or diurnal fluctuations,

it is unacceptable if processes that change in a few minutes or seconds, such as










neurological transmission, are to be measured.

An obvious way to improve the sampling time, or temporal resolution, in this case

is to increase the perfusion flow rate. If the flow rate is increased to 10 UL/min from the

original 1 tL/min 30 uL will be collected in 3 minutes. The drawback to this approach is

that the relative recovery decreases with increasing flow rate. The 10 fold improvement

will result in a 10 fold decrease in the concentration of the analyte. In some analyses, this

loss may be acceptable; however, when monitoring analytes at low concentrations, the

decrease in recovery may interfere with the measurement. In these cases the sensitivity

of the detector will limit how often a measurement can be made. For this reason high

sensitivity detection methods such as electrochemistry and laser-induced fluorescence

have become popular. By coupling micro and capillary LC columns that require less

volume to these ultra-sensitive detection techniques low level analytes and short

sampling times can be obtained. For example, with a capillary column, electrochemical

detection, and special handling methods it was possible to monitor the 16 amino acids

present in the brain with 10 s sampling (BOYOO). The drawback to high sampling rates

is that a large number of samples are generated. In the example given above, 100 s of

monitoring generated 10 samples. With an analysis time of 30 min per sample, 100 s of

monitoring required 5 h to analyze.



On-Line Analysis

Manually collecting samples is labor intensive, can be difficult depending on the

sampling rate, and is less precise than automated collection. Additionally, because

analysis can be time consuming, samples are usually frozen and stored for a considerable










length of time. The freeze-thaw process can cause loss of analyte due to degradation and

adsorbtion. A way to avoid these problems is on-line analysis where the dialysis stream

is loaded directly into a sample loop or is discretely sampled and injected onto a

separation column.

One potential drawback of on-line microdialysis is that the separation time of the

analysis is the limit to temporal resolution as opposed to sample volume or mass

sensitivity of the detection scheme. A separation time of 30 minutes would only allow 30

minute temporal resolution and would offset most of the advantages of the on-line

analysis. As the speed of the separation is increased, the benefits of an on-line instrument

become greater. The use ofmicrobore columns and small injection volumes have

allowed for sampling times as short as 30 s with microdialysis with on-line LC (CHU87,

CHE95, NEW94).

On-line capillary electrophoresis (DAV97 for review) is not as straightforward as

on-line liquid chromatography but offers unique advantages. Sample loops and injection

valves do not lend themselves to CE because of the very small volumes injected onto the

capillary and the high voltage required for the separation. One method developed for on-

line CE is the gap-junction method developed in the Lunte group (HOG94). Here,

dialysate flows into a 60 nL internal loop injection valve which can be switched to

introduce as little as 10 nL into the system to be derivatized and injected onto the CE

column. The actual injection is performed by aligning the transfer line with the

separation capillary with a small gap in between. The gap junction is grounded to

complete the electrical circuit. With this method Glu and aspartate (Asp) have been

monitored in 2 min intervals (ZHO95B).









An on-line flow gated capillary electrophoresis instrument with laser-induced

fluorescence instrument has been previously developed in this lab for the analysis of the

neurotransmitter amino acids Glu and Asp capable of 5 s sampling intervals (LAD98).

The flow-gated interface is based on one introduced by Lemmo and Jorgenson (LEM91).

Briefly, the dialysis outlet and inlet of the separation capillary are placed a small distance

apart (from 30 to 300 rin, depending on flow rates). A perpendicular cross flow of

electrophoresis buffer sweeps the dialysis away from the separation capillary so that only

buffer is introduced onto the column. For injection, the cross flow is stopped for a short

period of time (from a few milliseconds to a few seconds) which allows the dialysis to

reach the separation capillary. The applied potential injects a plug of dialysate. The

injection is terminated when the cross flow is resumed.



Glutamate and Aspartate in the Brain

We have chosen to use the microdialysis flow-gated capillary electrophoresis

instrument described above to study glutamate and aspartate in the rat brain. These two

amino acids are the two major excitatory neurotransmitter amino acids in the mammalian

central nervous system (ORR93). It is generally accepted that Glu functions in this

manner and there is a wealth of data to support this idea. Asp is also believed to be a

neurotransmitter but its evidence is less convincing (PAL89). Previous work in this lab in

the striatum provides evidence for Asp's neurotransmitter role (LAD98).

A glutamatergic synapse is diagrammed in Figure 1-3. The signal propagates

through a cell by a change in action potential and is dependent on the opening and closing

of sodium/potassium ion channels. This action potential then causes the opening of gated










calcium channels. The influx of calcium into the cell allows the vesicular release of

neurotransmitters into the synaptic cleft to continue the signal at the next neuron (JES93).

Once in the cleft the neurotransmitter can interact with 1) autoreceptors that will either

promote or inhibit further Glu signaling or 2) Glu transporters on either glia or neuron to

clear Glu from the cleft. The neurotransmitter may also diffuse out of the cleft and

interact with receptors and transporters on other cells. It is what happens in and around

the synaptic cleft that is the focus of this work.

The major signal mediators of the glutamnatergic system are ionotropic and

metabotropic glutamate autoreceptors (iGluR and mGluR). This work will focus on the

mGluR only. There are currently 8 known mGluRs divided into three sub groups; all of

which are differentially located throughout the brain and have been mapped by

immunostaining (SHI97). Group I mGluRs (subtypes 1 and 5) are coupled to

polyphosphoinositide (PI) hydrolysis and serve to amplify post synaptic responses

(CON94). They are located both pre- and post-synaptically and are typically at the

periphery of the post synaptic densities (ABE92). Both subtypes are also differentially

located throughout the brain with group 5 primarily located in the hippocampus,

neostriatum, and cerebral cortex. Additionally group 5 has also been found on cultured

astrocytes. Group II (subtypes 2 and 3) and group III (subtypes 4, 6, 7, and 8) are

negatively coupled to cAMP and are considered to be inhibitory in nature (CON94). All

subtypes are predominately located in the nerve terminals but mGluR3 is also found in

astrocytes (BRU94).

A set of five glutamate transporters (GLAST, GLT, EAAC, EAAT4, and EAAT5)

are responsible for removing Glu and Asp from the synaptic cleft and extracellular space









(DEE97). GLT is the most prevalent of the transporters, located mostly on astrocytes

(CHA95), and is in fact one of the most prevalent of all brain membrane proteins at

approximately I % of all proteins (DAN90). The hippocampus and cerebral cortex have

the highest levels of GLT (PIN92). GLAST is found exclusively on glial cells and has its

highest concentration in the cerebellum. EAAC is located primarily on neurons, but

away from the nerve terminals, and is found throughout the brain (COC97). EAAT4 is

located in the hippocampus and neocortex at very low concentrations (FUR97) while

EAAT5 is primarily in the retina (ARR97).

For normal brain function, it is required that both receptors and transporters work

properly. Because of their excitatory nature, increased levels of group I mGluR have

been associated with several neurological diseases from ALS to Parkinson's to epilepsy

(see BOR99 for review); conversely, group II and III mGluR activity has been shown to

be neuroprotective because of its inhibitory effects (BRU98). High extracellular Glu has

been shown to be excitotoxic and may be due to inactive or reversed transporters,

especially in an energy deprived system (SAN99). In addition to elevated group I

mGluR, decreased transport has also been found in patients with ALS (ROT92). Because

of their neurodegenerative or neuroprotective characteristics, mGluR and Glu transporters

are drug candidates for a variety of neurological diseases.

By using the analytical methods for the high-speed in vivo analysis of Glu and

Asp developed previously and in the work presented here, the interactions between Glu,

mGluR, and Glu transporters are studied. Once the normal and stressed activity of the

glutamatergic system is known it will be possible to guide the development of treatments

for a variety of neurological diseases.















Inlet--^







T
I
1- 4mm


x


x
4

x


x


x

x


00 Outlet






x



x


200 grm
Figure 1-1. Microdialysis probe. Perfusion fluid enters the inlet, passes
through a semipermeable membrane and to the outlet. A concentration gradient
across the membrane allows analytes to pass into the probe.







/,


U


Figure 1-2. Types ofmicrodialysis probes. A) In-line. B). Loop. C) Concentric.
D) Side-by-side.


I-=Ir-


=m


(;P








Axon


Na', K'

4Glu


Gin


Ca2 -+


Figure 1-3. Action of glutamate at a neuron.


(11, III)


Space
















CHAPTER 2
TEMPORAL RESOLUTION IN MICRODIALYSIS


Introduction

Temporal resolution in microdialysis with conventional analysis methods, such as

LC, is limited by volume or mass requirements for detection in an off-line analysis or

separation time in an on-line analysis. The best temporal resolution reported in

microdialysis coupled to a LC separation is 10 s (BOYOO). This measurement utilized

capillary columns with electrochemical detection and preconcentration to load the highest

possible mass onto the column. Coupled to a flow-gated interface, on-line capillary

electrophoresis with laser-induced fluorescence detection is capable of monitoring Glu

and Asp with 7 to 14 s temporal resolution (LAD96) and has been used to monitor the

two neurotransmitters in vivo with 5 s sampling (LAD98). In that study temporal

resolution was assumed to be limited by the separation time but no studies were

performed to confirm it.

With such high sampling rates, it now becomes important to know the limits to

temporal resolution of the microdialysis probe itself The time the perfusion fluid

remains in contact with the dialysis membrane, diffusion across the membrane, and

diffusion in transfer lines may all play a role in the lower limit to temporal resolution. As

discussed in Chapter 1, decreasing probe length and increasing flow rates will improve

temporal resolution at the expense of relative recovery.


I A










In order to determine the limit to temporal resolution associated with

microdialysis, the temporal characteristics of probes constructed by the method of Church

and Justice (CHU87) are rigorously studied with respect to tip length and flow rate. The

effects of these parameters on relative recovery are also studied. Additionally, the

sources of temporal broadening of the on-line flow-gated capillary electrophoresis

instrument discussed above will be determined and eliminated to obtain the best possible

temporal resolution with microdialysis.

Finally, an advantage of microdialysis probes most often stated by the

manufacturer is that probes can be reused for chronic measurements or to reduce inter-

animal variability and cost. Here, the effects of probe reuse in different animals on both

relative recovery and basal levels of Glu are examined.



Experimental

Probe Construction and Equilibration

Side by side microdialysis probes 1, 2, 3, and 4 mm long were constructed

following the methods developed by Church and Justice (CHU87) as detailed in

Appendix 2. Probes consisted of 9 cm of 20 jtm i.d. and 100 gm o.d. fused silica capillary

as the inlet and 8 cm of 20 pm i.d. and 100 um o.d. as the outlet. The dialysis membrane

was regenerated cellulose (Spectrum, Houston, TX, USA) with 18 kD cutoff and the tip

sealed with polyimide sealing resin (Alltech Associates, Deerfield, IL, USA). Active

areas were 200 pm o.d. Before use, all probes were perfused with 70% ethanol in water

for 30 min at 1.2 guL/min followed by perfusion with artificial cerebrospinal fluid (aCSF)

that consisted of 145 mM NaCI, 2.68 mM KC1, 1.01 mM MgSO4, and 1.22 mM CaCI2,









for a minimum of 30 min.

Measurement of Temporal Resolution and Relative Recovery of the Probe

For experiments determining the temporal resolution and relative recovery of side

by side microdialysis probes, the setup depicted in Figure 2-1A was used. aCSF was

perfused through the probe via a mechanical syringe pump (Harvard, Holliston, MA,

USA) at flow rates of 0.22, 0.60, 1.17, or 2.29 pl/min. The active area of the probe was

immersed in 10 ml of stirred aCSF at 37 C and the outlet connected to a 50 mrn i.d. 360

pim o.d. fused silica capillary with a variable wavelength UV detector (Spectraphysics,

Anaheim, CA, USA) set at 261 nm 16 cm down column.

One hundred pL of 10 mM ascorbic acid in aCSF was injected with a 500 kL

Hamilton gas tight syringe into the stirred CSF. The temporal resolution was calculated

by the time for the absorbance to rise from 10% to 90% of the maximum value. Five

repetitions were performed for each probe at each flow rate and 3 probes of each length

were used. Relative recovery for a probe was determined using the average absorbance

at each flow rate divided by the absorbance of an aliquot of solution taken from the 10

mL stirred CSF solution pumped directly into the 50 p.m i.d. capillary.

Measurement of Temporal Resolution of the On-Line System

A schematic of the on-line system is shown in Figure 2-1B. The outlet of the

probe was 8 cm, the mixing tee was either from Valco (Figure 2-2A) (Houston, TX,

USA) or a zero dead volume tee constructed in house (for dimensions, see Figure 2-2B)

as follows. Approximately 5 cm lengths of 50 p4m i.d. by 360 prm o.d. fused silica

capillaries were cut so that the end was straight and flat. Under a stereomicroscope

(Nikon, Japan) the ends were ground to 45. The tee itself consisted of a standard Valco










housing with the center made of Lucite and constructed by the University of Florida

Chemistry Department Machine Shop. The ground capillaries were sleeved into 1/16"

o.d. Teflon tubing and inserted into the tee so that the ends fit together flush. If one

capillary did not fit flush, it was removed and remade. Once in place the capillary ends

were cut to be flush with the holding nut. The dialysis and derivatization lines were

connected by Teflon sleeves. The reaction capillary was 16 cm by 75 Pm i.d., 360 gm

o.d. The "derivatization fluid" was aCSF at the same flow rate as the dialysis fluid.

Experiments were performed in the same manner as above.

Capillary Electrophoresis with Laser-Induced Fluorescence

The CE-LIF system for monitoring dialysate was similar to that described

previously (LAD96, LAD97) and is diagrammed in Figure 2-3. Dialysate was

derivatized on-line by pumping 10 mM o-phthaldialdehyde/40 mM P-mercaptoethanol

(OPA/BME) in 35 mM borate at pH 10.5 at 0.60 uL/min via a syringe pump (Harvard

apparatus, Holliston, MA, USA) into a tee attached to the outlet of the dialysis probe.

The dialysate and OPA/BME were allowed to mix and react for 50 s as they flowed out

of the tee and through a 16 cm long by 75 pim i.d. fused silica reactor capillary.

Derivatized dialysate was pumped from the reactor capillary to a flow-gated interface

which allowed automatic injection onto the electrophoresis capillary. Dialysate was

electrokinetically injected for 40 ms at the separation voltage every 3.5 s.

Electrophoresis was carried out inside a 5 cm long (3.5 cm from injection point to

detector) by 10 gm i.d. capillary. Voltage was applied by grounding the flow-gated

interface and applying -14.5 kV to a buffer reservoir that was connected to the outlet of

the separation capillary by 20 cm long by 150 pm i.d. linker capillary. The








electrophoresis buffer was 40 mM carbonate adjusted to pH 9.5 with NaOH.

Fluorescence detection was performed on the electrophoresis column with an

epillumination fluorescence microscope (Carl Zeiss, Hanover, MD, USA) as described in

detail elsewhere (HER90, SHU93). Fluorescence was excited with 1.3 mW of the 354

nm line of a He-Cd laser (Liconix, Carlsbad CA, USA). Emission over 400 n was

collected with a 40x oil immersion objective and detected by a PMT (CRG Instruments,

Austin, TX, USA). Data was collected via a personal computer (Gateway, Sioux City,

SD, USA) and data acquisition board (AT-MIO-16F-5, National Instruments, Austin, TX,

USA) at 1 kHz and low-pass filtered at 200 Hz. All collection and analysis software was

written in house in Lab Windows (National Instruments).

Surgical Procedures

Male Sprague-Dawley rats (250-350 g) were anesthetized with chloral hydrate by

an initial subcutaneous injection of 100 mg/ml H20 at a dose of 400 mg/kg followed by

booster injections of 100 mg/kg at 30 min intervals until the rat no longer exhibited a

withdrawal reflex to limb pinch. Additional injections were given as needed to maintain

no withdrawal reflex. Microdialysis probes were stereotaxically implanted in the dorsal

striatum at AP +0.2 mm, ML -3.0 mm, DV -6.5 mm relative to bregma (SWA98). Probes

were implanted over a period of 15 to 20 min in 200 gtm steps in order to minimize tissue

damage. Recordings were made at least 1 h post implantation and after basal levels were

found to stabilize.

For repetitive implantations of the same probe, a single 2 mm tip probe was

conditioned as above and used in a single rat for approximately 4 to 6 h on four

consecutive days. During experiments the flow rate was 0.60 pl/min. After use, the










outside of the probe was rinsed thoroughly with aCSF and then perfused with aCSF at a

low flow rate overnight. The probe was perfused with EtOH/H20 only prior to initial use.

All experimental uses of laboratory rats in the present study were reviewed and

approved by the University of Florida Institutional Animal Care and Use Committee and

conform with policies and procedures set forth by U.S. Public Health Service Policy on

Humane Care and Use of Laboratory Animals.



Results
Temporal Resolution of Microdialysis Probes

The response time of the probe is defined as the time for the signal to rise from 10

to 90% of the height of the step (see inset in Figure 2-1). The expected trend in response

time is that as flow rate increases or tip length decreases, response time will decrease but

is not seen over the entire range as expected. A plot of response time vs. tip length

(Figure 2-4) shows that for the two slowest flow rates (0.22 and 0.60 pl/min) or longest

tip lengths tip lengths (3 and 4 mm) the effect is linear. Response time reaches a

minimum between 1-1.5 s with 1 mm tip lengths at flow rates of 1.17 and 2.29 uL/min

and 2 mm tip lengths at 2.29 uL/min. The measurement is likely limited by the mixing

time after the injection of ascorbic acid and not a real limit due to the probe. In fact, in

initial experiments ascorbic acid was introduced via a Brinkmann pipette and the lower

limit of temporal resolution was more than 2 s. By using the Hamilton syringe, the

sample could be introduced faster and hence decreased the lower limit.

Relative Recovery of Microdialysis Probes

Overall flow rates and tip lengths the relative recovery changed as expected,

doubling the flow rate or halving the tip length resulted in halving the recovery (see










Figure 2-5). The recovery ranged from nearly 100% at 0.22 gl/min with a 4 mm tip

probe down to approximately 6% at 2.29 .dl/min with a 1 mm tip probe.

Temporal Response of the On-Line Instrument

The temporal resolution measured in the configuration used for on-line CE was

much worse than the microdialysis probe by itself There are two possible sources for

broadening in the system, diffusion in the capillaries and dead volume. The calculated

broadening due to diffusion in the capillaries was insignificant indicating that dead

volume was the problem. Dead volume can be considered a region where fluid becomes

stagnant due to eddies in flow or a region where fluid can pool and mix. The system has

one such region, the Valco reaction tee (see Figure 2-2A). Fluid enters via a 40 Pm i.d.

capillary and exits via a 75 gm i.d. capillary, both sleeved in a 1/16th in. o.d. Teflon

holder. Additionally, the internal volume of the 150 um i.d. tee is 42 nL, seemingly

small but virtually identical to the volume inside the active region of a 2mm long probe

(45 nL). The capillary to tee interface may provide an unswept region because of the

sudden change in diameter and the additional volume will cause extra mixing. These two

factors make the Valco reaction tee a likely source of temporal broadening.

In order to confirm the tee as a broadening source, a zero dead volume tee (see

Figure 2-2B) was constructed using fused silica capillaries ground at 45 angles and held

flush inside a Lucite housing. With this configuration there is virtually no dead volume

throughout the system. Figure 2-6 compares the temporal resolution of the system with

the Valco tee to the zero dead volume tee. As can be seen, the Valco tee was responsible

for a significant amount of temporal broadening. With a 3 mm tip at 1.2 ml/min, the

response time with the Valco tee was nearly 250% (9.7 vs. 4.0 s) greater than that with a









zero dead volume tee. Response time of the system with the zero dead volume tee was

equal to that of a microdialysis probe only.

Possible Broadening Inside the Probe

Even though the system after the microdialysis probe does not contribute any

significant broadening, it is possible that the microdialysis probe has an unswept volume

or that there is broadening due to diffusion across the membrane itself. If rapid diffusion

across the dialysis membrane can be assumed, the theoretical temporal resolution should

be the time it takes to replace all of the perfusion fluid in the active area of the probe.

Figure 2-7 compares the temporal resolution at a single flow rate to the calculated time to

clear the active area, represented by the solid line. As can be seen, the experimentally

determined resolution agrees extremely well with the calculated value.

Finally, to determine any temporal broadening associated with the on-line

derivatization reaction or glow-gated interface, the in vitro resolution with a Glu and Asp

standard was determined. A typical electropherogram of a standard solution of 2.5 PM

Glu and 1.4 WM Asp is shown in Figure 2-8. The probe size and flow rate used in this

experiment were 2 mm and 0.60 ul/min, which should correspond to a temporal

resolution of approximately 5 s. As seen in Figure 2-9, both Glu and Asp reach

maximum height in 2 injections at 3.5 s per injection. This time frame corresponds to a

temporal resolution between 3.5 and 7.0 s, exactly what was determined for the

microdialysis probe only.

Reuse of the Microdialysis Probe

The effect of multiple uses of a single probe is shown in Figure 2-10. As

expected the relative recovery of the probe decreases with each consecutive day. The










initial relative recovery of this particular probe was 35.6 % and decreased to 21.7 % by

the fourth day. Brain concentration of Glu on the first day was 670 nM, by the fourth day,

the measured basal Glu increased over ten fold to 7.6 pM. In brain concentrations are

used here to account for the change in relative recovery.



Discussion

The results presented here show that microdialysis is a sampling method that is

highly amenable to and capable of high-speed analyses. Temporal resolution of the

microdialysis probe itself was shown to be as low as 1 s, depending on the flow rate and

probe size. It is likely that a better testing method will result in a more accurate and

lower limit to temporal resolution. With a tip length that was not apparently limited by

mixing time, temporal resolution was equal to the time for the active area of the probe to

be completely swept by perfusion media. Based on this result, a 1 mm tip probe and a

flow rate of 2.4 pl/min should provide a temporal resolution of 500 ms. Such resolution

does however have its drawbacks. With improvement in temporal resolution there is a

corresponding loss in relative recovery. The same 1 mm tip probe at 2.4 4l/min and 500

ms temporal resolution has a relative recovery of just 6%. This means that a 1.0 PM

analyte in the brain is only 60 nM in dialysate.

The CE system used here has a limit of detection for dialysate of approximately

30 nM and can reproducibly inject every 3.5 s, so therefore would barely be able to detect

the analyte and would not be capable of taking advantage of this temporal resolution. It

again becomes necessary to improve the analysis system. With proper instrument design

it should be possible to obtain sub-second temporal resolution with microdialysis and on-










line flow gated capillary electrophoresis.

Even though the lower limits of temporal resolution can not be taken advantage of

in the work carried out here; the knowledge of the limitations of both the system and

microdialysis will make better designed experiments possible. For example, probe

lengths and flow rates can be chosen to maximize both temporal and spatial resolution as

well as relative recovery. The same as one minute sampling does not take advantage of 5

s temporal resolution due to high flow rates and small tip lengths, a 1 s separation is

unnecessary with 5 s temporal resolution.

With these particular probes, it appears that reuse will result in decreased

performance of the probe as well as an increase in measured basal levels. It may be

possible to explain both phenomena by a layer of gliosis that develops on the probe after

implantation in the brain. This layer of cells in fact forms on any object implanted in the

brain for a length of time, including metal electrodes and fused silica capillary. The layer

of cells could serve as an added diffusion barrier that retards Glu from crossing the

membrane. The increase in basal concentrations may be the result of additional damage

caused by a thicker probe or some sort of interaction between the gliosis and the brain.













Dialysis Pump, CSF



37 C
Probe


Stir


Plate


Stir Bar
7'\


L
Syringe Pump, CSF
Mixing


Dialysis Pump, CSF Ll UV



37 OC -- I |
Probe Stir Bar


Stir Plate


Response Time


Time (s)

Tee


Figure 2-1. Schematics for temporal resolution studies. A) Set up for probe only
B) Set up for on-line analysis.




























B













Figure 2-2. Diagram of mixing tees. A) Valco tee. The volume inside is 42 nl.
B) Zero dead volume tee. 50 pm i.d. x 360 gm o.d. capillaries are ground to
45 degree angles and butted against each other inside the tee. Once secured in place
capillaries are cut to be flush with the nut.









Derivatization Pump

| 10.6 U/min


Dialysis Pump


340


Reaction Tee


Separation
Capillary


Cross-Flow Pump Switching Valve


loam x 5 cm
2900 V/cm





Buffer


Microdialysis Probe


To Waste


Figure 2-3. Microdialysis with on-line flow gated capillary electrophoresis


Flow-Gated Interface


















20-


16-


12-


8-


-- 0.22 pl/min
-- 0.60 di/min
-U-1.17 pi/min
-4-2.29 ti/min


Tip Length (mm)


Figure 2-4. Effect of tip length and flow rate on temporal
resolution. At high flow rates and small tip lengths resolution levels
off around 1 s. This limit is likely due to the mixing time of the
solution. Error bars represent one S. E. M. (n = 3)



















100-


1 mm


80-

60-

40-

20-


| .=


1 2
Flow Rate (pUL/min)


100


3 mm


Flow Rate (pL/min)


2 mm


0 1 2 3
Flow Rate (gL/min)


0 1 2
Flow Rate (pUL/min)


Figure 2-5. Effect of flow rate and probe tip length on relative recovery. Recovery decreases
as flow rate increases and tip length decreases over the range tested. Error bars represent one
S. E. M. (n = 3).


100.

680-

60'*

.. 40-

20-

0


100.

80'

60

40-

20-


- I

























0 D. V. Tee
Res = 4.0 s


Valco Tee
Res = 9.7 s


I . I I I. I .I I . I I I I. I. I
0 10 20 30 40 50
Time (s)
Figure 2-6. Temporal resolution of the on-line system, Comparison
of Tees. As seen here the dead volume due to the Valco tee more
than doubles temporal resolution. Conditions: 3 mm tip probe, flow
rate 1.17 gd/min, 100 gl injection of 10 mM ascorbic acid into 10 ml
of CSF at 37 0C. The calculated time in the active area of the probe
is 3.7 s.
















12

10 Experimental




6 -
0
V 4-

Calculated Time in Probe
0 -------------
2-

0
0 1 2 3 4
Tip Length (mm)

Figure 2-7. System temporal resolution compared to calculated resolution. At
0.60 pUL/min the temporal resolution of the system with a zero dead volume tee is
equivalent to the calculated dwell time in the active area of the microdialysis probe.
Error bars represent one S. E. M. (n = 5)





































0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

time (s)

Figure 2-8. In vitro electropherogram of 2.5 pM Glu and 1.4 PM
Asp. Labelled peaks are as follows 1) reagent peaks, 2) 200 JIM
glycine, 3) glutamate, 4) aspartate. Figures of merit: Glu: N = 34,600,
S/N = 160, Asp: N = 44,000, S/N = 100. Microdialysis probe: 2 mm
tip, Flow Rate: 0.6 pl/min. See text for separation conditions.




















1200-

- ;1000

.800

S600-

S400

-- 200
;290


-- Glu
- Asp


120


150


Time (s)

Figure 2-9. Response of CE-LIF to step change in Glu and Asp.
Points represent the peak height of analyte and are from
electropherograms collected every 3.5 s. With 2 mm tip probes at 0.60
pl/min, temporal resolution is between 3.5 and 7.0 s. Arrow indicates
change, the delay in response is due to the dwell time of the system.
See text for separation conditions.







































-*- Basal


45

40

35

30

25

20


Day
-- Recovery


Figure 2-10. Reuse of a single microdialysis probe in 4 rats on
consecutive days. Tip length was 2 mm with a flow rate of 0.6 p.L/min.
Note concentration is given as in brain to correct for recovery changes.
















CHAPTER 3
IN VIVO MONITORING OF GLUTAMATE: THE ROLE OF ASCORBIC ACID IN
THE BRAIN


Introduction

Ascorbic acid (AA), present as the anion ascorbate in vivo, is a ubiquitous

substance throughout the body with a wide range of functions. Its most widely recognized

role is as an antioxidant, reducing free radicals formed during oxidative metabolism

(BEY94, ROS93). This role seems to be especially important in the brain where a high

level of oxidative activity is required for the metabolism and release of neurotransmitters

(GOT94). Concentrations of AA in the extracellular fluid are between 200 and 500 JIM

(GON81, ONE82, SCH82) and are unevenly distributed throughout the brain (MIL82,

STA84). Regions of highest concentration are the hippocampus, hypothalamus and

cerebral cortex. Both the striatum and hippocampus have concentration gradients in the

individual region (BAS91, MIL82). Intracellular concentrations can be up to 10 mM and

also vary by region and cell type (neuron vs. glia) (RIC97).

An additional role for AA has recently been proposed in which it heteroexchanges

with Glu during reuptake at Glu transporters (GRU93, REB97). In this scenario, AA

would act as a charge carrier to preserve the cell potential (neuron or glia) as Glu is

removed from the extracellular space. The majority of the evidence is indirect and

involves the effect of Glu on AA. This first line of evidence for heteroexchange is that

direct infusion of Glu will cause an increase in extracellular AA measured by










voltammetry (ONE84). This effect is transient, with AA returning to basal levels within

3 to 5 min (PIE93) and can be blocked by Glu uptake inhibitors (CAM91). Additionally,

electrical stimulation of the glutamatergic pathway in the hippocampus causes an

increase in AA (ONE84, CAM91). Glu uptake inhibitors have also been shown to

attenuate the release ofAA (GRU93, CAM91).

In this chapter we study the direct effect of adding 500 pM AA to the brain via

reverse dialysis on basal Glu and how it affects the electrically stimulated release of Glu

by stimulation of the prefrontal cortex. Electrical stimulation will be used because it

provides a neuronal source of Glu (LAD98) which will then be uptaken into the

surrounding cells. If heteroexchange occurs, the excess AA in the extracellular space

should retard uptake due to an altered equilibrium. This effect may result in an increase

in basal Glu and/or stimulated overflow because of reduced uptake. To differentiate

between heteroexchange and oxidative metabolism, the effects of 500 pM iso-ascorbic

acid (IAA) on basal and stimulated levels are also studied. IAA is the D-isomer of AA

and has the same antioxidant properties as AA but is not transportable.



Experimental

Chemicals and Reagents

All chemicals were purchased from Sigma Chemical (St. Louis, MO USA). All

fused silica capillary tubing was from Polymicro Technologies (Phoenix, AZ, USA) and

had a 360 pm outer diameter (o.d.) unless noted otherwise.

Capillary Electrophoresis and Laser-Induced Fluorescence Detection

The CE-LIF system for monitoring dialysate was the same as that described in










Chapter 2. Dialysate was derivatized on-line by pumping 10 mM o-phthaldialdehyde/40

mM P3-mercaptoethanol (OPA/BME) in 35 mM borate at pH 10.5 at 1.2 iUL/min via a

syringe pump (Harvard Apparatus, Holliston, MA, USA) into a tee (Valco, Houston, TX,

USA) attached to the outlet of the dialysis probe. The dialysate and OPA/BME were

allowed to mix and react for 18 s as they flowed out of the tee and through a 16 cm long

by 75 gm i.d. fused silica reactor capillary. Derivatized dialysate was pumped from the

reactor capillary into a flow-gated interface which allowed automatic injection onto the

electrophoresis capillary. The distance between the reaction and separation capillaries

was approximately 60 gm. Dialysate was electokinetically injected for 40 ms at the

separation voltage every 5 s. Electrophoresis was carried out inside a 5 cm long (3.5 cm

from injection point to detector) by 10 gm i.d. capillary. Voltage was applied by

grounding the flow-gated interface and applying -13 kV to a buffer reservoir that was

connected to the outlet of the separation capillary by 20 cm long by 150 ptm i.d. linker

capillary. The electrophoresis buffer was 40 mM carbonate adjusted to pH 9.5 with

NaOH.

Fluorescence detection was performed on the electrophoresis column with an

epillumination fluorescence microscope (Carl Zeiss, Hanover, MD, USA). Fluorescence

was excited with 1.3 mW of the 354 nm line of a He-Cd laser (Liconix, Carlsbad CA,

USA). Emission over 400 nm was collected with a 40x objective and detected by a PMT

(CRG Instruments, Austin, TX, USA). Data was collected via a personal computer

(Gateway, Sioux City, SD, USA) and data acquisition board (AT-MIO-16F-5, National

Instruments, Austin, TX, USA) at 1 kHz and low-pass filtered at 200 Hz. All collection

and analysis software was written in house in Lab Windows (National Instruments).








Microdialysis

Side by side microdialysis probes were constructed as described in Appendix 2.

Probes consisted of 9 cm of 20 pm i.d. and 100 pgm o.d. fused silica capillary as the inlet

and 8 cm of 20 gpm i.d. and 100 gm o.d. as the outlet. The dialysis membrane was

regenerated cellulose (Spectrum, Houston, TX, USA) with 18 kD cutoff and the tip

sealed with polyimide sealing resin (Altech Associates, Deerfield, IL, USA). Active

areas were 200 pgm o.d. and 4 mm long. Before use, all probes were perfused with 70%

ethanol in water for 30 min at 1.2 gL/min followed by perfusion with artificial

cerebrospinal fluid (aCSF) that consisted of 145 mM NaCI, 2.68 mM KC1, 1.01 mM

MgSO4, and 1.22 mM CaCI2, for a minimum of 30 min. Perfusion flow rate was 1.2

pL/min and was driven by a syringe pump (Harvard apparatus, Holliston, MA, USA) for

all experiments. Calibration and in vitro recovery was performed before implantation by

placing the probe in Glu and Asp standards dissolved in aCSF held at 37 C. For

calibration and in vivo experiments, the outlet capillary of the probe was connected to the

reactor tee described above for on-line CE-LIF analysis.

Surgical Procedures

Male Sprague-Dawley rats (250-350 g) were anesthetized with chloral hydrate by

an initial subcutaneous injection of 100 mg/ml H20 at a dose of 400 mg/kg followed by

booster injections of 100 mg/kg at 30 min intervals until the rat no longer exhibited a

withdrawal reflex to limb pinch. Additional injections were given as needed to maintain

no withdrawal reflex. Microdialysis probes were stereotaxically implanted in the dorsal

striatum at AP +0.2 mm, ML -3.0 mm, DV -6.5 mm relative to bregma (SWA98).

Electrical stimulations were performed by placing a single platinum electrode (1 mm








diameter) in direct contact with the surface of the prefrontal cortex (AP +3.5 mm, ML

+1.5 mm). Probes were implanted over a period of 15 to 20 min in 200 pmn steps in order

to minimize tissue damage. Recordings were made at least 1 h post implantation and

after basal levels were found to stabilize.

All experimental uses of laboratory rats in the present study were reviewed and

approved by the University of Florida Institutional Animal Care and Use Committee and

conform with policies and procedures set forth by U.S. Public Health Service Policy on

Humane Care and Use of Laboratory Animals.

Electrical Stimulation and Pharmacological Treatments

Electrical stimulation of the perforant path consisted of application of a 10 s train

of 0.5 ms, 20 Hz, 80 V square wave pulses (Grass Medical Instruments, Quincy, MA). To

evaluate the effects of pharmacological treatment on the signal from electrical

stimulation, 500 gM AA or IAA in aCSF was administered by reverse dialysis. Prior to

treatment, at least 2 control stimulations were recorded in each rat, with 30 min between

each stimulation. After control stimulations, the dialysis media was switched to the

additive media. 60 min after initiation ofperfusion with either AA or IAA, a set of 2-3

identical electrical stimulations was again performed at 30 min intervals.

Data Analysis

Electropherograms were analyzed using a statistical moments program written in

house (Lab Windows). Glu and Asp peaks were marked manually and dialysate

concentration determined by in vitro calibration. The 100 % basal level was measured as

the average concentration of analyte in 10 electropherograms immediately prior to

stimulation. Elevated concentrations due to AA were determined by comparison to the








analyte concentration before application of the additive. Stimulated overflow was

calculated as the total mass of analyte collected by the microdialysis probe above the

average pre-stimulus level. A two-tailed student's t-test was used to determine all

statistical differences. All values are reported as mean one standard error, n is given as

the number of rats for basal concentrations or number of stimulations for overflow.



Results

Basal Conditions

Basal levels in dialysate determined by the 4 mm side-by-side probes were 250

85 nM for Glu and 95 10 nM for Asp (n = 7) respectively. Relative recoveries were 40

2% and 43 2% for Glu and Asp respectively. In the basal electropherogram shown in

Figure 3-1, it can be seen that several peaks besides Glu and Asp are present. Several of

them have been identified by comparison to the migration time of amino acid standards.

From left to right, they are o-phosphoserine a mGluR agonist (from the previous

injection), arginine, glycine an inhibitory amino acid (partially resolved from the large

neutral peak) glutathione an antioxidant, and o-phosphoethanolamine an osmo-regulator,

glutamate, and aspartate. The other peaks could not be conclusively identified. These

peaks do not occur in a 40 amino acid standard (Sigma Chemical) and may be small

peptides (2-3 amino acids) or degradation products from larger peptides.

Effects of Ascorbic Acid

Control basal levels of Glu in the dialysate were 250 80 nM (n = 7). One hour

after infusion of 500 pM AA the extracellular concentration of Glu in dialysate was 910

140 nM (n = 4), a 360 % increase over control (p < 0.001). It must be noted that the










effect of AA on extracellular Glu was not an immediate effect but increased over

approximately 40 min. While this effect was observed there were no attempts made to

monitor the time course of the change. The addition of 500 mM IAA to the perfusion

medium resulted in no change in basal Glu (n = 3).

The application of an electrical stimulus to the prefrontal cortex resulted in an

immediate increase in extracellular Glu and return to basal upon termination of stimulus

within the resolution of the instrument and can be seen in Figure 3-2. The peak increase

in Glu was 76 13 % (n = 17) over basal and returned to within 10% of original basal

levels within 40 s after termination of stimulus. The total mass of Glu collected over

basal level was 328 31 finol (n = 17). In the presence of 500 pM AA, the peak increase

of Glu was 201 31 % (n = 11) and returned to within 10 % of basal 160 s after cessation

of stimulus. The total mass collected over the elevated basal level was 1650 144 frnol

(n = 11), or a 503 % increase over control (p < 0.0001). In contrast 500 PM IAA added

to the perfusion medium resulted in a 92 21 % (n = 7) increase over basal and returned

to within 10 % of basal levels 60 s after cessation of stimulus. The total mass of Glu

collected over basal was 355 37 finol (n = 7). Neither the maximum stimulated

increase nor total mass collected was statistically different from control.

To quantify the change in overall clearance rate due to addition of 500 pM AA to

the perfusion medium, the concentration dependent decay rates of the stimulation curves

were determined (see Figure 3-3). The slope of the line indicates the overall clearance

rate, including uptake, diffusion, and recovery by the probe. Assuming that diffusion and

probe recovery are constant, the change in slope will be due to the change in uptake. The

decay rates are: aCSF only = 0.143 0.010 s'1, IAA = 0.136 0.010 s', and AA = 0.109








41

0.010 s' (different from aCSF p < 0.01, from IAA p < 0.05). The difference between

aCSF and IAA was not statistically significant.



Discussion

As seen in Figure 3-2, addition of AA to the perfusion medium resulted in both a

360% increase in basal levels and 500% increase in stimulated overflow of Glu in the

striatum. The addition of IAA to the perfusion medium lead to no change in basal Glu

and a small but statistically insignificant increase in stimulated overflow. The

combination of these two pieces of data is consistent with Glu and AA heteroexchange.

The small change with IAA may indicate a minor role for oxidative loss of Glu from the

extracellular space. But, because this change is not significant no reliable conclusions

can be made. Higher concentrations of IAA should be infused to determine any role of

oxidative loss of extracellular Glu.

The gradual increase in basal Glu does not immediately suggest a direct action of

AA on the transporters when compared to the action of drugs that directly inhibit the Glu

transporter. The direct infusion of an uptake inhibitor such as L-trans-pyrrolidine-2,4-

dicarboxylic acid (PDC) results in an immediate increase in Glu basal levels. However,

because there already is a high level of AA in the extracellular fluid and the added AA is

not much higher than the natural level it may take quite some time for the added AA to

have an effect. It has been previously suggested that the Glu-AA interaction is not direct

(GRU93) and may be associated with carrier proteins not involved with Glu transport

(FIL86) although no such proteins have been found.

This relatively small addition of AA leads to an interesting possibility. When









aCSF without AA is perfused, extracellular AA will actually be depleted due to reverse

dialysis. Therefore, control experiments may not be representative of true physiological

conditions. It may be that in this case, the amount of AA added only returns the

concentration to near normal levels. If this holds true, it would appear that uptake in the

brain is not as strong as previously believed. It is not an uncommon practice to add AA

to the perfusion medium but is most often used to protect against oxidation of the analyte,

not to specifically add it back to the brain.

The results from the work here demonstrate for the first time direct evidence for

the modulation of Glu by AA which is consistent with heteroexchange. However, this is

not a complete picture, additional studies to confirm these results should be carried out.

These studies include concentration dependent studies as well as a rigorous study of the

time course of the effect of AA on both basal and stimulated levels of Glu.



















PEA
Arg



Gly

GSH

|\ Glu

LSOP j Asp


0 1 2 3 4 5
Time (s)
Figure 3-1. Basal electropherogram. The peaks are identified as follows:
LSOP = o-phosphoserine (overlapped from previous run), Arg = arginine,
Gly = glycine (partially resolved), GSH = glutathione, PEA = o-phospho-
ethanolamine, Glu = glutamate, Asp = aspartate. The remaining peaks were
not conclusively identified. See text for conditions.














600-

500 loo100s

400
0 300
200-
100
U U
ioo^Joo vl
0- aCSF 500 gM 500 pM
0 I-AA AA
Figure 3-2. Evidence for ascorbic acid heteroexchange with Glu.
500 pM of AA added to the perfusion medium resulted in an increase
in basal and stimulated overflow of Glu by 360% and 500%
respectively. IAA had no effect on basal or stimulated overflow. See
text for separation conditions. Error bars represent one S. E. M.









0.03 0.03 0.05
aCSF IAA AA



0.02- 0.02- 0.04 /





0.01- 0.01 0.03





0.00 -i 0.00 0.02-
0.5 0.7 0.9 0.5 0.7 0.9 2.8 3 3.2

Concentration (pM)
Figure 3-3. Caculated decay rates for aCSF, IAA, and AA. Points were taken from data in Figure 3-2. Slopes:
aCSF = 0.143 +/- 0.010, IAA = 0.136 +/- 0.010, AA = 0.1093 +/- 0.010.
















CHAPTER 4
CHARACTERIZATION OF METABOTROPIC RECEPTOR CONTROL AND
UPTAKE OF GLUTAMATE AND ASPARTATE IN THE RAT HIPPOCAMPUS


Introduction

Glutamate has been well established as the major excitatory neurotransmitter in

the CNS. Despite overwhelming evidence for the role of glutamate as a neurotransmitter,

direct measurement of its extracellular level in vivo by microdialysis sampling or sensors

has not provided convincing evidence of a neuronal origin for glutamate in the

extracellular space. Rather, extracellular glutamate level appears to have an atypical

regulation that suggests a more complex level of control than that encountered with

amine neurotransmitters or acetylcholine.

Particular problems with glutamate measurements include that the basal levels are

not sensitive to Na+ channel blockage by local administration of TTX or to local

depletion of Ca2 as summarized in several recent reviews and papers (FIL95, TIM97,

HER96). TTX-sensitivity and Ca2-dependency are taken as minimal requirements for

considering extracellular measurements to reflect neuronal activity. Thus, the exact

source of the basal levels of glutamate in the extracellular space sampled by sensors and

microdialysis is unclear. Several hypotheses have been advanced to explain the source of

the basal glutamate level detected by microdialysis including carrier mediated release

(LEV93) or transport from the plasma (FIL95).










Additionally, physiological events such as stress (MOG93) or seizure generation

(MIL93) have been shown to increase Glu in a partially TTX-dependent fashion. These

changes have been interpreted as illustration of measurement of synaptic overflow of Glu

released from neurons. An alternative interpretation however is the increased level, while

dependent upon neuronal activity is actually a secondary change (TIM97) perhaps

reflective of changes in metabolism. This view acknowledges the fact that inclusion of

TTX will stop all neuronal firing, including non-glutamatergic neurons thus causing

cessation of any process that depends upon neuronal electrical activity. Because of these

limitations, it seems useful to utilize selective stimulus ofglutamatergic cell bodies

combined with measurement in the terminal fields of these neurons in order to attempt to

characterize neuronal release of glutamate in vivo. One approach to avoiding the

confusion over the basal levels of glutamate is to characterize the extracellular level

changes during stimuli as has been shown in the previous chapter.

In the hippocampus, stimulation of the perforant pathway has allowed detection of

glutamate increases in two different reports. In one report, an amperometric glutamate

sensor placed in the dentate gyrus detected increases over -45 s that peaked at a 21 pM

increase over basal following a 5 s stimulation (HU94). Such an observation is consistent

with the notion of detection of electrically stimulated overflow; however, the TTX-

sensitivity and Ca2+ dependency of this change was not studied. In another report, 20 s

stimulations allowed detection of 20% increases over basal glutamate over a 15 s period

(WAL95). Again however, the TTX sensitivity and Ca2-dependency were not assessed.

In this chapter, the influence of EAA transporters and presynaptic metabotropic

Glu autoreceptors (effects of utilized drugs are summarized in Table 4-1) on extracellular










levels of Glu and Asp in rat hippocampus under basal (steady-state) conditions as well as

under dynamic conditions associated with perforant pathway stimulation are evaluated.

Our results provide novel evidence that synaptic overflow of neuronally-released Glu

occurs in vivo within hippocampal circuits following afferent stimulation.



Experimental

Drugs and Reagents

All chemicals were purchased from Sigma Chemical (St. Louis, MO, USA) unless

otherwise indicated. All drugs were administered by reverse microdialysis. Fused silica

capillary was purchased from Polymicro Technologies (Phoenix, AZ, USA) and was 360

pm o.d. unless otherwise stated.

Capillary Electrophoresis and Laser-Induced Fluorescence Detection

The instrument is the same as was described in Chapter 2. aCSF is perfused (0.60

gl/min) through the microdialysis probe and mixed with 10 mM o-phthaldialdehyde/40

mM 03-mercaptoethanol in 35 mM borate at pH 10.5 (0.60 gl/min) in a zero dead volume

tee described in chapter 4. The mixture reacts for 50 s in a 16 cm x 75 Pm i.d. fused

silica capillary and is carried to a flow-gated interface. Cross flow is 340 gl/min and the

distance between reaction and separation capillaries is 30 gim. For injection, the cross-

flow is stopped for 40 ms via a pneumatically actuated valve with a high-speed switching

assembly. For better injection reproducibility (1-2% improvement) 1-2 mm of polyimide

was removed from the tip of each capillary. The separation capillary was 5 cm x 10 gmrn

i.d. x 360 gm o.d. and connected to voltage via a 20 cm x 150 pm i.d. x 360 4m o.d.

linker capillary. A 5 mm detection window was burned by electric arc 3.5 cm from the








injection end. The separation buffer was 40 mM carbonate adjusted to pH 9.5 with

saturated NaOH. The applied potential was 13,000 V with the interface held at ground.

Injections were made every 5 s.

Fluorescence detection was performed by excitation with the 354 nm line of a He-

Cd laser (Liconix, Carlsbad, CA, USA). Emission over 400 nm was detected with a PMT

and associated components (CRG Instruments, Austin, TX, USA) and low pass filtered at

200 Hz. Data was collected via a personal computer (Gateway, Sioux City, SD, USA)

and data acquisition board (AT-MIO-16F-5, National Instruments, Austin,TX, USA) at

1000 Hz. All software was written in house in Lab Windows (National Instruments).

Microdialysis and Electrical Stimulation

Side by side microdialysis probes were constructed as described in Appendix 2.

Probes consisted of 9 cm of 20 i.d. and 100 o.d. fused silica capillary as the inlet and 7

cm of 20 i.d. and 100 o.d. as the outlet. The dialysis membrane was regenerated cellulose

(Spectrum, Houston, TX, USA) with 18 kD cutoff and the tip sealed with polyimide

sealing resin (Alltech Associates, Deerfield, IL, USA). Active areas were 200 pm o.d.

and 2 mm in length. Before use, all probes were perfused with 70% ethanol in water for

30 min at 1.2 WL/min followed by perfusion with artificial cerebrospinal fluid (aCSF) that

consisted of 145 mM NaCI, 2.68 mM KCI, 1.01 mM MgSO4, and 1.22 mM CaCI2, for a

minimum of 30 min. Unless otherwise stated, perfusion flow rate was 0.60 UL/min and

was driven by a syringe pump (Harvard apparatus, Holliston, MA, USA) for all

experiments. Calibration and in vitro recovery was performed before implantation by

placing the probe in Glu and Asp standards dissolved in aCSF held at 37 C.

Probes were stereotaxicly implanted in the dentate gyrus of chloral hydrate










anesthetized rat (coordinates: AP 3.8 mm, ML -3.0 mm, DV = -4.7 mm (from dura)

from bregma). Bipolar stimulating electrodes (Rhodes Medical Equipment, Woodland

Hills, CA, USA) were stereotaxicly implanted in the lateral perforant path (coordinates:

AP: 0.0 mm, ML -4.4 mm, DV -3.5 mm (from dura) from lambda). Both probe and

electrode were implanted over a period of 15 to 20 min in order to minimize tissue

damage. No experiments were performed for at least 1 hr post implantation to allow

stabilization of basal levels.

The perforant path was stimulated by application of a 20 s train of 0.1 ms, 20 Hz,

10 V square wave pulses (Grass Instruments, Quincy, MA, USA). Stimulations were

applied at least 20 min apart. Four stimulations were performed for each treatment per

rat and only one drug treatment was given per rat.

Pharmacological Treatments

All treatments were administered by reverse dialysis with at least 30 min allowed

before stimulation. Treatments were as follows: Na+ channel block Ige, 2 pM

tetrodotoxin (TTX); Ca2+ blockage, 2 mM EGTA in Ca2+ free CSF (same as regular CSF,

without CaCI2 added); uptake inhibition, 200 pM L-trans-pyrrolidine-2,4-dicarboxylic

acid (PDC); mGluR blockage, 200 uM (RS)-a-methyl-4-carboxyphenylglycine (MCPG)

(Tocris Cookson, Ballwin, MO, USA); mGluR activation, 200 PM (1S,3R)-1-

aminocyclopentane-trans-l,3-dicarboxylic acid (ACPD) (Tocris Cookson) or 20 JM o-

phosphoserine (LSOP).



Results

Basal levels of Glu and Asp in dialysates obtained from the hippocampus were







51

310 40 nM and 87 11 nM (n = 21) respectively. Average in vitro relative recoveries

at 37 C were 36.5 + 1.2 % and 39.4 1.6 % for Glu and Asp respectively.

As shown in Figure 4-1, we observed that with the onset of electrical pulses, Glu

level rose instantaneously, within the limits of our temporal resolution, and continued to

increase during the entire stimulation. As soon as the stimulus was ended, Glu level

began to decline and returned to pre-stimulus basal levels within 20 s. Interestingly, the

level after stimulus was often slightly lower than the prestimulus level. The amount of

Glu collected by the dialysis probe over basal level as a result of the stimulation averaged

66.5 7.3 fminol (n = 80 stimulations in total of 21 rats). The maximal level achieved

during the stimulation was 112 3% of the basal level. The response of Glu to

successive stimulations was highly reproducible as shown by the traces in Figure 4-1 A

and C with no significant difference in maximum Glu increase, duration, or peak area

after 4 stimulations. The high reproducibility of successive stimulations allowed changes

in stimulated Glu levels to be precisely monitored by this approach. Asp showed a small

change correlated with the stimulation that is only apparent after averaging all

stimulations together (Figure 4-1B). In view of the difficulty of measuring changes in

Asp level, it was not quantified in further studies.

Effect of Na+-Channel Blockade with TTX and Ca2' Depletion

To ascertain if the observed Glu concentration changes were related to neuronal

events, we tested the TTX-sensitivity and Ca2+-dependency of the signals. Infusion of 2

pM TTX through the probe in the hippocampus did not significantly affect basal Glu or

Asp levels (summary of basal data are given in Table 4-2). TTX infusion completely

eliminated electrically-stimulated increases in Glu concentration in the hippocampus, as








seen in Figures 4-2A and Table 4-3. Removal of Ca2+ from the dialysis medium with

addition of 2.0 mM of the Ca2-chelator EGTA, resulted in no change in the basal level

for Glu or Asp in the hippocampus as shown in Table 4-2; however, this manipulation

significantly (p < 0.005) reduced the stimulated overflow of Glu in the hippocampus to

17 4% of the stimulated overflow in control conditions (n = 12 stimulations in 3 rats) as

shown in Figures 4-2B and Table 4-3. The observation that EGTA did not completely

halt electrical stimulation is likely due to ineffective removal of extracellular Ca2.

mGluR Regulation of Glu

mGluR autoreceptors are reported to have significant effects on glutamate release

in the hippocampus. We therefore examined the effect of infusion of different mGluR

antagonists and agonists on the basal and stimulated level of Glu in the hippocampus.

Infusion with 200 VM MCPG, a broad spectrum mGluR antagonist, increased the basal

level of Glu to 116 5% (p < 0.05) of the control value and increased the stimulated

overflow (measured as the area under the peak) to 190 11% of the control (p < 0.02, n =

12 stimulations in 4 rats) as shown in Figures 4-3 and Table 4-2. This result indicates

significant autoreceptor control over release during electrical stimulations. Asp basal

levels increased dramatically with MCPG (Table 4-2), but substantial stimulated release

was not observed. Addition of the group I and II agonist ACPD (200 4M) to the dialysis

perfusion medium decreased the basal Glu level to 87 5% of control (p < 0.05) but did

not significantly affect the stimulated overflow (overflow was 102 21% of control, n =

16 stimulations in 4 rats) as shown in Figure 4-4A. As ACPD is only effective at group I

and II mGluR, we also examined the effect of L-SOP, a group LII agonist. At 20 PM L-

SOP did not significantly affect either the basal (Figure 4-4B and Table 4-2) or










stimulated overflow of Glu in the hippocampus (Table 4-3). These results indicated that

the blockade of autoinhibition apparent with MCPG was achieved at basal concentrations

of Glu and further addition of agonist had no additional inhibitory effect. Interestingly,

when a higher concentration of LSOP (200 pM) was added, the opposite effect as

expected was seen. 200 pM LSOP increased Glu to 166 +/- 190% and Asp to 152 +/-10%

above basal (Figure 4-4C and Table 4-2). Stimulated overflow of Glu was also increased

by 162 +/- 21% of control values (Table 4-3).

Effect of Uptake Inhibition

Infusion of the transportable Glu uptake inhibitor PDC (200 PM) into the

hippocampus caused a dramatic increase in both Glu and Asp to 614 81% and 540 +

74% (n = 4) of basal levels which may reflect both decreased uptake and increased efflux

of these amino acids by reverse transport. (see Figure 4-5 and Table 4-2) This result is in

agreement with previous studies on the effect of PDC on basal Glu level in the

hippocampus (ZUI96, OBR96). PDC also caused a dramatic increase in the electrically

stimulated overflow of Glu to 987 146% of control stimulations (p < 0.0001, n = 16

stimulations in 4 rats). (Figure 4-5 and Table 4-3) In the presence of PDC, changes in

Asp were still not observed aside from small change after the end of the stimulus. Since

PDC raised basal levels and affected the endogenous agonist of the mGluR, the effect of

MCPG was re-tested under these elevated basal conditions. Addition of 200 JM MCPG

to the perfusion fluid after 40 min of PDC treatment reduced the basal level of Glu and

Asp to 431 90% and 397 65% (n = 4) of control basal levels. This was a statistically

significant decrease from the PDC induced basal level (p < 0.05). Electrically stimulated

overflow of Glu was reduced to 583 146% of control (n = 14 stimulations in 4 rats)








54

which was significantly smaller (p < 0.05) than the overflow that occurred with only PDC

present. Traces illustrating the effect of PDC and PDC with MCPG on the stimulated

overflow are shown in Figure 4-5 and a summary the effects on basal and stimulated

overflow are shown in Tables 4-2 and 4-3 respectively. Reversal of the order of addition

of MCPG and PDC did not affect the final result (data not shown).



Discussion

The measurement of electrical stimulation of the perforant path proved to be an

analytically challenging task. Because the average change in Glu was only 12% over

basal, the instrument reproducibility became the limiting factor in being able to measure a

change. It was necessary to use great care in maintaining and setting up the instrument.

For example, mechanical factors such as keeping the syringe pumps well oiled and the

injection valve and rotor clean were crucial in minimizing variability. Additionally any

misalignment in the flow gated interface contributed significantly to injection

irreproducibility.

Another factor to consider in the measurement of such small changes is the

sensitivity of the detector. An average change of 12% in dialysate of Glu corresponds to

a concentration change of 36 nM. For Asp, this change would be only 10 nM. It is likely

that this small a change in Asp was below the sensitivity of the instrument, supported by

the fact that a change in Asp could only be seen after averaging all stimulations. The

response of Asp is discussed below.

Evidence for Synaptic Glu Overflow in Hippocampus Following Perforant Path

Stimulation.









In these experiments we evaluated glutamate level under basal, i.e., steady state,

conditions and dynamically during electrical stimulation. In the absence of an applied

stimulus, basal levels of Glu and Asp are insensitive to either pharmacological blockade

of voltage-activated sodium channels or a marked reduction in the concentration of

extracellular free calcium ions (Table 4-2). The insensitivity of extracellular EAA levels

to these experimental manipulations have been reported widely by other investigators and

have been used as supportive evidence for the hypothesis that extracellular EAA levels in

vivo do not provide an accurate measure of neurogenically-derived EAA release (FIL95;

TIM97; HER96). In this regard, the fast (5 s) sampling intervals with on-line

measurement of EAA level used in this study have provided data that parallel results

from studies that employ much longer (up to 15 min) sampling intervals. However, the

clear advantage of fast on-line measurements is apparent when this approach is used to

monitor rapid changes in extracellular EAA levels initiated by a brief electrical stimulus.

Similar to previous reports that utilized sensors (HU94; WAL95), extracellular Glu

concentration in the hippocampal formation exhibits close temporal coupling with

perforant path stimulation insofar as Glu level undergoes a sharp rise and fall that

coincides closely with the onset and termination, respectively, of a 20-s stimulus period

(Fig. 4-1). In view of the modest size and duration of the stimulated rise in hippocampal

extracellular Glu, it is likely that conventional microdialysis sampling methods would not

detect such brief elevations of extracellular Glu owing to the extensive loss of signal

caused by protracted temporal averaging. For the first time, we demonstrate that these

changes exhibit complete TTX-sensitivity as well as a nearly total reliance upon

extracellular free Ca2+ (Fig. 4-2). From the temporal relationship, TTX-sensitivity, and








Ca2+-dependence, it appears that impulse-dependent neuronal release is the major

contributor for this increased extracellular Glu level following perforant path stimulation.

In light of these observations, we conclude that, following exocytotic release from

hippocampal nerve terminals, a significant amount of Glu can diffuse from synapses into

the surrounding extracellular fluid. As a consequence of this synaptic overflow, it is

likely that a rapidly dissipating wave of Glu diffuses into the surrounding tissues wherein

it could elicit non-synaptic excitatory effects.

The detected overflow was greatly increased in the presence of PDC, a

competitive and transportable inhibitor of glutamate uptake. These results support the

hypothesis that uptake plays a strong role in preventing spread of the glutamate signal

and cross-talk among neurons in the hippocampus. This idea has previously been

advanced based on theoretical and electrophysiological studies of Glu spread in the

hippocampus (RUS98, ASZ97, CLE92).

mGluR Control

While uptake plays an obviously strong role in regulating the spread of a Glu

signal, the role of mGluR is less clear from our results. Metabotropic glutamate receptors

have been implicated as autoreceptors in regulation of Glu release with both facilatory

and inhibitory effects having been observed (LIU95, SCA97). In our experiments we

observed a fairly weak autoreceptor regulation of stimulated overflow. For example,

addition ofACPD (agonist for group I and II mGluR) or 20 pM L-SOP (agonist for

group III mGluR) had no effect on stimulated release and MCPG (broad spectrum

antagonist) caused a small increase in stimulated overflow. These results are consistent

with the notion of autoinhibition of release mediated by mGluR that is maximally










activated under basal conditions so that addition of agonist does not inhibit release but

antagonist can block the endogenous ligand and give rise to enhanced release.

Interestingly, in the presence of PDC, where endogenous ligand level was enhanced

914%, MCPG actually had the opposite effect on overflow and substantially reduced it as

well as the basal level. These results are consistent with a feed-forward mechanism of

glutamate release mediated by mGluR under conditions of high agonist availability

(HER92, TIZ95, LIU95). Thus, it appears that when Glu levels are low, autoreceptors

are active that mediate an inhibition of release so that antagonists increase the stimulated

release of Glu; however, when glutamate levels are high, mGluR mediate positive

feedback thus giving rise to a decrease in release with antagonist. A previous in vivo

microdialysis study of the hippocampus of freely moving rats revealed a similar

regulation of Glu level in that mGluR mediated tonic inhibitory and phasic excitatory

control over Glu release (LIU95). A detailed explanation of this observation would

require considerably further study. It should be recognized that the dialysis probes are

relatively large and sample Glu release from multiple sub-regions of the hippocampus

and many neurons and glia. Since multiple sub-types of mGluR with opposite effects are

co-localized and heterogeneously distributed throughout the hippocampus (SHI97), this

lack of spatial resolution available with microdialysis makes interpretation of the results

difficult. Probes with better spatial resolution and more selective drugs would be

required to better address this unusual effect.

Regulation of Extracellular Asp

While the major focus of this investigation has centered on changes in

hippocampal extracellular Glu levels during perforant path stimulation, it remains unclear








58

whether or to what degree Asp may contribute to the synaptic or extra-synaptic effects of

EAA within the hippocampal formation. Our previous study of stimulated changes in

striatal EAA levels revealed large parallel increases in both Glu and Asp following

electrical stimulation of the prefrontal cortex (LAD98). Results from that study provided

compelling and novel evidence for exocytotic Asp release insofar as the stimulated rise in

Asp occurred simultaneously with the rise in extracellular Glu and was completely

attenuated by calcium depletion or TTX treatment. In light of these earlier results, we

believe that Asp could exhibit a similar profile in hippocampus and potentially play a role

in synaptic or non-synaptic excitatory communication. Indeed, following perforant path

stimulation a small increase in hippocampal extracellular Asp levels is detectable (Fig. 4-

1B). The low signal to noise ratio of the Asp signal required extensive signal averaging

to detect this change; therefore further characterization of the Asp signal was not

performed; however, in view of the basal levels of Glu (310 nM) and Asp (87 nM) in

dialysate samples, the average peak rise in Asp following perforant path stimulation

represents approximately 20 percent of the net stimulated rise in extracellular Glu

concentration. This result agrees closely with studies in hippocampal tissues in vivo

wherein depolarizing stimuli have been reported to release Glu and Asp in a 5:1 ratio

(SZE87, BUR88, ZH095). In view of this data and the close temporal coupling between

perforant path stimulation and the rise in extracellular Asp, we suggest that the stimulated

increase in Asp level may derive from neuronal exocytotic release. While it remains to

be determined whether Asp exhibits the expected requirement for calcium and TTX

sensitivity, several lines of evidence support our premise that Asp undergoes exocytotic

release in the hippocampus. Studies in vivo reveal that a variety of depolarizing stimuli









can elicit calcium-dependent release of Glu and Asp from hippocampal synaptosomes

(ZH095) as well as hippocampal slices (SZE87; BUR88, NAD90, ROI91, KLA92). In

addition, a recent report by Gundersen and coworkers (1998) indicates that Glu and Asp

are co-stored in hippocampal synaptic vesicles and that both amino acids are depleted

from vesicles through a tetanus toxin-sensitive process following depolarization of

hippocampal slices (GUN98). Taken together, these results provide compelling evidence

for exocytotic Asp release from hippocampal neurons. In view of our in vivo evidence

for exocytotic Asp release in striatum (LAD98), the approach of Gundersen and

coworkers (1998) may prove useful in order to determine whether Asp and Glu are co-

stored in synaptic vesicles of the striatum or other forebrain regions. Although evidence

regarding vesicular localization and calcium-dependent release are just two of several

criteria that must be satisfied in order to validate Asp as an excitatory neurotransmitter, it

is important to note that this issue remains unresolved. Previous failures to detect Asp

within synaptic vesicles have been used to discount a possible transmitter role for Asp

within forebrain pathways (FIL95). However, we have noted that nearly all studies that

have reported a lack of vesicular Asp uptake/storage have been carried out in

homogenates or purified subcellular fractions from the cerebral cortex. (BUR89, CAR89,

VIL90, TAB92). In view of the convincing results of Gundersen and coworkers (1998)

in hippocampal slices, it seems plausible that vesicular pools of Asp may be extremely

labile or factors) that are required for vesicular Asp uptake may be depleted during

subcellular fractionation procedures. Alternatively, neurons within the cerebral cortex

may differ from excitatory neurons in other brain regions with respect to the capacity for

Asp uptake or storage. In any event, we conclude that current evidence regarding a








60

possible role of Asp as a neurotransmitter or neuromodulator in the hippocampal

formation is inconclusive and warrants more careful consideration.

In conclusion, electrical stimulation of the perforant yields an increase in

extracellular Glu that is susceptible to Na+ channel blockage and Ca2 depletion,

indicating a neuronal origin. Additionally, application ofmGluR agonists and

antagonists indicate that the receptors are activated under normal conditions.

























Table 4-1. Additives Used for Glu Characterization.


Additive Abbreviation Action Reference

(1 S,3R)- 1 -aminocyclopentane- ACPD Group I and II KAH93
trans-1,3-dicarboxylic acid metabotropic receptor BR095
agonist PIN95
l-serine-o-phosphate LSOP Group III metabotropic NAK92
receptor agonist SAU94
(RS)-a-methyl-4- MCPG Group I, II, and Il1 CH096
carboxyphenylglycine metabotropic receptor
antagonist
L-trans-pyrrolidine-2,4- PDC Glu uptake inhibitor BRI91
dicarboxylic acid GRI94
Bis-(aminoethyl)glycoether- EGTA Ca2+ chelator, disrupts IMP84
N,N,N',N'-tetraacetic acid Ca2+ channels
Tetrodotoxin TTX Na+ channel blocker WES87

















Table 4-2. Effects of Chemical Additives on Basal Levels of Glutamate and Aspartate


Treatment Glu (% of Control) Asp (% of Control)

Control 100 +/-12 100 +/-8

2 iM TTX 97.8+/-1.2 100+/-4

2 mM EGTA/Ca2+ Free 96.7 +/- 11.5 98.7+/-9.5

200 uM ACPD 87.4 +/- 5.5* 73.5 +/- 16.1*

20 iM LSOP 103 +/- 3 86.2 +/- 7.4*

200 jiM LSOP 166+/- 19 152 +/- 10*

200 gM MCPG 116 +/- 5* 328 +/- 130*

200 uM PDC 614 +/- 81* 540 +/- 117*

200 uiM PDC/MCPG 431 +/- 90.1 397 +/- 65*


Basal Levels of glutamate and aspartate were 310 +/- 40 and 87 +/- 11 nM respectively
(n = 21). Values in the table are expressed as percent of control levels, mean +/- SEM
following a 30 min infusion of the drug (n = 4 for additives).
* indicates significantly different from control (p < 0.01)
















Table 4-3. Effects of Chemical Additives on Electrically Stimulated Levels Glutamate


Treatment


Giun (% of Control)


Control


2 gM TTX


2 mM EGTA/Ca2+ Free


200 pM ACPD

20 gM LSOP

200 WM LSOP

20 uiM LSOP

200 pM MCPG

200 uM PDC


100+/-11

3.0 +/- 1.0*

17.1 +/- 4.3*

95.8 +/- 21.5

103 +/- 17

162+/-21

103 +/- 17


190 +/-9


987 +/- 146*

583 +/- 146*


200 pM PDC/MCPG


Data for electrically stimulated increases in glutamate were determined as net changes
over baseline values. Control evoked response was 66.5 +/- 7.3 fminol (n = 80
stimulations, 21 rats). Values in the table are expressed as percent of the control levels,
mean +/- SEM. For additives, n = 16, 4 rats.
* indicates significantly different from control (p < 0.02)


.........................................................................................................................................................................................................................................













A


m


130

S10-0
110
g o-
e 90

70

130-

" 110-

e 90-

70-


80"1 C


70-
60-


U----


1 2


3 4


Ave.


Stimulation Number


Figure 4-1. Electrical stimulation of the perforant path. Traces represent four
consecutive stimulations spaced 20 min apart followed by the averages trace +/-
SEM. A) Glu, B) Asp. C) Bar graph representing the average Glu overflow for
consecutive stimulations followed by the overall average overflow (n = 20). Bars
in A and B denote stimulation and represent 20 s.


B



A^44V. s


504--

















2 pM TTX


Ca2+ Free


Figure 4-2. Evidence for neuronal origin of electrically stimulated Glu. A)
Blocking sodium channels with 2 gM TTX or B) Removing calcium with 2 mM
EGTA in calcium free aCSF obolishes stimulated overflow but does not alter
basal levels. Bars indicate stimulation and correspond to 20 s. (4 rats, n = 16)


aCSF


aCSF


120
115
110
105
100
95
90
85
80


130

120
w:110


100
90

80


















130-
125
120
115 aCSF
S110 '.
105
100 i 200 pM MCPG
95 "
90

Figure 4-3. Effect of metabotropic receptor blockage on electrically
stimulated Glu. Addition of 200 iM MCPG to the perfusion medium
results in a 116% increase in basal levels and 190% increase in
stimulated overflow of Glu. Bars indicate stimulation and correspond
to 20 s.













120 A
- 110}
S 200 pM ACPD
100-
80-
80 aCSF L

130 B
S120
110
100 R"
90
aCSF 20 pM LSOP
200 C

Ix160 .
0 aCSF 200 pM LSOP
0, 120-

80-
Figure 4-4. Effect of metabotropic receptor activation on electrically
stimulated Glu A) Activation by 200 pM ACPD. Basal levels
increased to 87 % of control while stimulated overflow did not change
B) Activation by 20 pM ACPD. elicited no change in either basal or
stimulated overflow of Glu. C) Activation by 200 pM LSOP. caused
an increase in basal and stimulated overflow by 166% and 162%,
respectively. Bars indicated stimulation and correspond to 20 s.









190 -

170-

150

130-

110-

90-


aCSF


700

680

660

640

620

600

590

570

550

530

510

490


PDC


510

490

470

450

430

410

520

500

480

460

440

420


Figure 4-5. Effect of uptake inhibition on electrically stimulated Glu and Asp. A and D) control stimulation. B and E) Inhibition
of uptake with 200 pM PDC added to the perfusion medium results in a 614% increase in basal levels of Glu and a 987% increase
in stimulated overflow. Asp basal levels increases by 510%. C and F) 200 pM PDC and 200 pM MCPG decreases Glu and Asp
relative to PDC only to 430% and 450% of original basal. Stimulated overflow of Glu decreases to 580% of control. Bars indicate
stimulation and correspond to 20 s.


PDC/MCPG


180

160

140

120

100

80
















CHAPTER 5
REGIONAL DIFFERENCES BETWEEN STRIATUM AND HIPPOCAMPUS:
EVIDENCE FOR VOLUME TRANSMISSION


Introduction

There are two types of synaptic transmission, volume or wire transmission (see

ZOL96, ZOL98, ZOL99 for reviews). In the first, a neurotransmitter can diffuse out of

the synaptic cleft and interact with neighboring cells. Dopamine is considered the model

transmitter for this type of communication (VIZ91, BUN99). In wire transmission a

synapse is completely insulated and once a neurotransmitter is released it cannot escape

the synaptic cleft to interact with other neurons. The glutamatergic system is thought to

use this type of transmission because Glu neurons are typically ensheathed in glial cells

(CLE92, MEN95). This line of evidence has been used to discount the ability of

microdialysis to make a meaningful measurement of Glu (TIM97). However, recent

studies have demonstrated the possibility that Glu can escape from the cleft and perhaps

participate in limited volume transmission (RUS98, MIT00, KUL96, KUL98).

Additionally, in results presented in the previous chapter and elsewhere (LAD98), we

have demonstrated that Glu and Asp can diffuse into the extracellular space under

electrically stimulated conditions in both the hippocampus and striatum.

In view of these observations, a further examination of the previous data is

warranted. In this chapter the characteristics of electrically stimulated release and their

modulation by mGluR active drugs in both regions are compared with respect to volume










vs. wire transmission. To further elucidate the transmission properties in each region

addition experiments were conducted to determine the extent of extracellular transport

and diffusion.



Experimental

Chemicals and Reagents

All chemicals were purchased from Sigma Chemical (St. Louis, MO USA). All

fused silica capillary tubing was from Polymicro Technologies (Phoenix, AZ, USA) and

had a 360 gm outer diameter (o.d.) unless noted otherwise.

Capillary Electrophoresis and Laser-Induced Fluorescence Detection

The CE-LEF system for monitoring dialysate was the same as described in

Chapter 2. Dialysate was derivatized on-line by pumping 10 mM o-phthaldialdehyde/40

mM 13-mercaptoethanol (OPA/BME) in 35 mM borate at pH 10.5 at 0.60 lL/min via a

syringe pump (Harvard apparatus, Holliston, MA, USA) into a tee attached to the outlet

of the dialysis probe. The dialysate and OPA/BME were allowed to mix and react for 50

s as they flowed out of the tee and through a 16 cm long by 75 pm i.d. fused silica reactor

capillary. Derivatized dialysate was pumped from the reactor capillary into a flow-gated

interface which allowed automatic injection onto the electrophoresis capillary. Dialysate

was electokinetically injected for 40 ms at the separation voltage every 3.5 s.

Electrophoresis was carried out inside a 5 cm long (3.5 cm from injection point to

detector) by 10 gm i.d. capillary. Voltage was applied by grounding the flow-gated

interface and applying -14.5 kV to a buffer reservoir that was connected to the outlet of

the separation capillary by 20 cm long by 150 gm i.d. linker capillary. The








electrophoresis buffer was 40 mM carbonate adjusted to pH 9.5 with NaOH.

Fluorescence detection was performed on the electrophoresis column with an

epillumination fluorescence microscope (Carl Zeiss, Hanover, MD, USA). Fluorescence

was excited with 1.3 mW of the 354 nm line ofa He-Cd laser (Liconix, Carlsbad CA,

USA). Emission over 400 nm was collected with a 40x objective and detected by a PMT

(CRG Instruments, Austin, TX, USA). Data was collected via a personal computer

(Gateway, Sioux City, SD, USA) and data acquisition board (AT-MIO-16F-5, National

Instruments, Austin, TX, USA) at 1 kHz and low-pass filtered at 200 Hz. All collection

and analysis software was written in house in Lab Windows (National Instruments).

Microdialysis

Side by side microdialysis probes were constructed as described in Appendix 2.

Probes consisted of 9 cm of 20 gxm i.d. x 100 4.m o.d. fused silica capillary as the inlet

and 7 cm of 20 pim i.d. and 100 gtm o.d. as the outlet. The dialysis membrane was

regenerated cellulose (Spectrum, Houston, TX, USA) with 18 kD cutoff and the tip

sealed with polyimide sealing resin (Alltech Associates, Deerfield, IL, USA). Active

areas were 200 p4m o.d. and 2 mm in length. Before use, all probes were perfused with

70% ethanol in water for 30 min at 1.2 gL/min followed by perfusion with artificial

cerebrospinal fluid (aCSF) that consisted of 145 mM NaCI, 2.68 mM KCI, 1.01 mM

MgSO4, and 1.22 mM CaCl2, for a minimum of 30 min. Unless otherwise stated,

perfusion flow rate was 0.60 gL/min and was driven by a syringe pump (Harvard

apparatus, Holliston, MA, USA) for all experiments. Calibration and in vitro recovery

was performed before implantation by placing the probe in Glu and Asp standards

dissolved in aCSF held at 37 C. For calibration and in vivo experiments, the outlet









capillary of the probe was connected to the reactor tee described above for on-line CE-

LIF analysis.

Surgical Procedures

Male Sprague-Dawley rats (250-350 g) were anesthetized with 100 mg/ml in H20

chloral hydrate by an initial subcutaneous injection of 400 mg/kg followed by booster

injections of 100 mg/kg at 30 min intervals until the rat no longer exhibited a withdrawal

reflex to limb pinch. Additional injections were given as needed. Microdialysis probes

were stereotaxically implanted in the dentate gymrus at AP 3.8 mm, ML -3.0 mm, DV =

-4.7 mm (from dura) from bregma. Concentric bipolar stimulating electrodes (Rhodes

Medical Equipment, Woodland Hills, CA, USA) were stereotaxically implanted in the

lateral perforant path at AP 0.0 mm, ML -4.4 mm, DV -3.5 mm (from dura) from

lambda. Both probe and electrode were implanted over a period of 15 to 20 min in order

to minimize tissue damage. Recordings were made at least 1 h post implantation and

after basal levels were found to stabilize.

All experimental uses of laboratory rats in the present study were reviewed and

approved by the University of Florida Institutional Animal Care and Use Committee and

conform with policies and procedures set forth by U.S. Public Health Service Policy on

Humane Care and Use of Laboratory Animals.

Infusion of D-Asp

For comparison of EAA clearance between the hippocampus and striatum,

exogenous D-aspartate was microinjected into the brain region of interest. For these

experiments, a 75 gtm i.d. fused silica capillary was co-implanted with the dialysis probe.

The microinjection capillary was clamped to the same holder as the dialysis probe so that









its outlet was at the dorsal-ventral center of the 2 mm dialysis probe active area and

approximately 350 pLm medial of the dialysis probe, see Figure 5-1. The outlet tip of the

microinjection capillary was ground to a fine point with a Dremel tool so that its o.d. was

approximately 90 gtm. The inlet of the injector capillary was connected to a 10 pL glass

microsyringe (Hamilton). For infusions, 200 nL of 1 mM D-Asp in aCSF was manually

injected as recordings were made with the CE-LIF system. To minimize inter-animal

variability measurements were performed in the hippocampus and striatum in each rat

when possible.

Data Analysis

Electropherograms were analyzed using a statistical moments program written in

house (Lab Windows). Glu and Asp peaks were marked manually and dialysate

concentration determined by in vitro calibration. The 100 % basal level was measured as

the average concentration of neurotransmitter (NT) in 10 electropherograms immediately

prior to stimulation. Elevated or reduced concentrations due to chemical additives were

determined by comparison to the NT concentration before application of the additive.

Stimulated overflow was calculated as the total mass of NT collected by the

microdialysis probe above the average pre-stimulus level. A two-tailed student's t-test

was used to determine all statistical differences.



Results

Comparison of Electrically Stimulated Release in the Striatum and Hippocampus

Striatal stimulations from Chapter 3 and hippocampal stimulations from Chapter 4

are replotted in Figure 5-2. In contrast to the results in the hippocampus, we observed










large changes in Glu in the striatum flowing electrical stimulation of the PFC. Like the

results in the hippocampus, the increase began immediately with the stimulation and

began to decrease following end of the stimulation. The levels returned to baseline

within 40 s. We have also previously characterized the effects of TTX, Ca2+ removal,

ACPD, MCPG, PDC, and PDC with MCPG on this overflow and basal levels (LAD98).

The results from this prior study are summarized in Figures 5-3 and 5-4 for efficient

comparison to the data from the hippocampus from Chapter 4.

A striking difference between the electrically stimulated overflow in the

hippocampus and the striatum is the magnitude of the responses. Using similar dialysis

probes, the average amount collected during a 10 s stimulation in the striatum was 328

31 finol (n = 17 stimulations from 7 rats) which is 490% of the hippocampus signal

(difference significant to p < 0.001). Furthermore, the maximal increase was 176 13%

for the striatum and 112 3 % in the hippocampus (p < 0.0001). The net amount of

stimulated Glu overflow is a reflection of Glu release, both synaptic and extrasynaptic,

and Glu elimination by active (Glu transport) and passive (diffusion) processes; therefore,

it is not possible to discern whether the differences in the electrically stimulated release in

the two brain regions are mediated by differences in release or elimination.

Microinjection of D-Asp

To better distinguish the differences in elimination in the two brain regions, we

carried out complimentary studies in which a fixed amount (200 pmol) of D-Asp was

rapidly infused from a micropipette that was positioned 350 pm from the microdialysis

probe and followed the kinetics of D-Asp recovery. D-Asp was selected as the analyte

for these experiments in view of the following: (1) it is a chemically stable compound








75

that is resistant to enzyme-catalyzed degradation and exhibits little or no pharmacological

activity at the concentrations used here; (2) D-Asp interacts with EAA transporters in a

manner similar to L-Glu insofar as it binds to and is transported by all known classes of

EAA transporters with low micromolar affinity; and (3) D-Asp possesses charge

characteristics and diffusion properties that are similar to L-Glu.

Dialysate basal levels in the hippocampus were 690 100 nM for Glu and 260 +

60 nM for Asp and in the striatum 330 120 nM for Glu and 120 40 nM for Asp. This

difference in concentrations is consistent with other results (TOS86). Relative recoveries

were 37.9 2.2 % for Glu and 42.8 1.8 % for Asp. Infusion results are shown in

Figure 5-5. In the striatum 7.4 1.2 pmol (n = 20) was recovered in the dialysis probe

from an infusion which was significantly more (p < 0.0001) than the 2.2 0.5 pmol (n =

14) recovered in the hippocampus. In the presence of PDC, the amount collected in the

striatum was increased to 11.8 1.8 pmol (160% of the control, p < 0.05). In the

hippocampus, the PDC had a larger effect increasing the amount collected to 4.8 1.3

pmol (218% increase, p <0.02). To further characterize the tissue clearance of D-asp in

the two brain regions, we analyzed the decay rate of D-Asp level in the dialysate

following the direct infusion as shown in Figure 5-6. Linear decay rates were seen for all

conditions in both brain regions. In striatum, clearance rate was 0.189 0.001 s'1 and

0.140 0.007 s' for control and PDC conditions respectively while in the hippocampus

the clearance rates were 0.146 0.01 s'" and 0.089 0.003 s4. Thus, the clearance rate

in the striatum was significantly higher for the striatum than for the hippocampus (p <

0.005) and addition of PDC significantly decreased the clearance rate in both brain

regions (p < 0.001). The data also indicate however that PDC had a greater effect on the










decay rate in the hippocampus than in the striatum.

Discussion

Comparison of Electrically Stimulated Overflow in Hippocampus and Striatum.

The electrically stimulated overflow of Glu in the hippocampus provides an

interesting comparison to results for overflow in the striatum following electrical

stimulation of prefrontal cortex. As shown in Figure 5-2, in both cases a rapid rise in Glu

was detected that was well correlated with the stimulus in that the rise began immediately

after the stimulus and began to decrease at the termination of the stimulus. As in the

hippocampus, the stimulated overflow of Glu in striatum was TTX-sensitive and Ca2-

dependent. These results suggest a neuronal origin for the electrically stimulated signal

in striatum as well as the hippocampus. In contrast to the rather weak and complex

mGluR effects observed in the hippocampus, we have previously observed a strong

autoinhibition of electrically stimulated Glu overflow in the striatum mediated by mGluR

(LAD98). As shown in Figure 5-4, ACPD could drastically inhibit release, an effect that

was antagonized by MCPG. When glutamate levels were elevated by PDC, overflow

was not increased because release was strongly autoinhibited as evidenced by the large

increase in overflow by addition of MCPG (see Figure 5-4). These results led to the

conclusion that increasing agonist availability activated mGluR to suppress Glu release in

a classical autoinhibitory fashion. We also did not observe any switching of the effect of

mGluR at different agonist concentrations in the striatum unlike in the hippocampus.

Thus, from the electrical stimulation results it is apparent that overflow is regulated

differently in the striatum and hippocampus. In the hippocampus, overflow is dominantly

regulated by uptake as indicated by the strong effect of PDC. mGluR autoreceptors do










not strongly restrict overflow and in fact, at high agonist availability actually seems to

promote release. In the corticostriatal pathway, transporters and autoreceptors appear to

work together to restrict and regulate overflow. The importance of autoreceptors can be

seen in that with uptake inhibited, no strong overflow occurs because of activation of

autoinhibition.

Roles of Diffusion and EAA Transporters in Shaping Glu Overflow

An obvious difference in the stimulated overflow is that hippocampal responses

are considerably smaller and of shorter duration than the response evoked in striatum

(Figure 5-2). To discern whether the differences in overflow magnitude are mediated by

differences in release or elimination we examined the time course of microinjections of

fixed amounts of D-Asp in the two brain regions. This experiment allows the elimination

and transport processes to be studied in isolation of release processes. As shown in Fig.

5-5, recovery of D-Asp in a microdialysis probe positioned 350 pm from the injection

site was reduced by more than 70% in hippocampus relative to striatum. Factors that

could give rise to the marked reduction in hippocampal peak amplitude and duration

include more efficient transport of D-Asp into hippocampal neurons and glia or a

decreased rate of D-Asp diffusion in hippocampal tissues. Studies with the competitive

EAA transport inhibitor PDC lead us to conclude that both factors contribute to the

diminished recovery of D-Asp in hippocampus relative to striatum. While pretreatment

with a high concentration of PDC (200 iM) enhances D-Asp recovery by the probe in

both regions, the impact of EAA transport inhibition is two-fold greater in the

hippocampus (118 36% vs. 59 15% increase for hippocampus and striatum,

respectively) (Fig. 5-5B). This in vivo evidence for greater EAA transporter activity









within the hippocampus is consistent with results from in vitro studies, including

evidence for a greater density of sodium-dependent high-affinity D-[3fH]Asp binding

sites, a useful marker for EAA transporters, throughout the hippocampal formation

relative to striatum (AND90, GRE90, SEI91).

Further examination of the data suggests differences in the diffusional properties

of D-Asp between hippocampus and striatum as well. Plots of concentration decay rates

in both tissues are shown in Fig. 5-6. Inspection of these data reveals linear rates of

decay as a function of concentration in hippocampus and striatum in both the absence and

presence of PDC. The signal decays because of both reuptake by the tissue and diffusion

away from the probe. The response does not show saturation, which would be indicated

by non-linearities at higher concentrations, that would be expected for Michaelis-Menton

kinetic control of the decay as expected for reuptake. This may be because diffusion,

which would be linear under any condition, is participating in the decay or because the

concentrations necessary for saturation were not reached by this experiment.

Nevertheless, the results show that the decay rate is significantly lower in the

hippocampus than in the striatum, even in the presence of uptake inhibitors. A slower

decay rate is due to less reuptake or slower diffusion away from the probe. Since the

curve areas suggest greater reuptake in the hippocampus, the lower decay rate must be

due to more restricted diffusion. This conclusion is consistent with the expected effective

diffusion coefficients (Dff) in the striatum and hippocampus. Deff = D/I2 where D is the

diffusion coefficient in buffer and X is the tortuosity factor of the tissue (HAR49). In the

striatum X is 1.54 (RIC91) leading to an expected Deff of 4.8 x 106 cm2/s. In the

hippocampus, the tissue is anisotropic with an average = 1.61 (MAZ98) giving rise to a








Deff of 4.4 x 10-6 cm2/s. (D for glutamate is 1.14 x 10-5 cm2/s which was used for D-Asp

in this estimation) (NIC81).

To summarize, direct infusion of D-Asp into the hippocampus and striatum

reveals greater reuptake and slower diffusion for EAAs in the hippocampus than in the

striatum. This result indicates a greater restriction on the spread of an extracellular

concentration pulse of Glu in the hippocampus than in the striatum. This observation

helps explain the large difference in stimulated overflow that was observed in the two

brain regions. Even with the magnitude of these effects taken into account, it appears

that the electrically stimulated hippocampal Glu signals are weaker than those observed

in the striatum suggesting less release as well. It is difficult to conclude anything

significant from this observation however because the amount released could be affected

by the number of neurons in the tract, their relative density at the site of measurement,

and the effectiveness of the stimulation.

Implications for EAA Mediated Volume Transmission

Glutamate is generally considered to be a point-to-point neurotransmitter

(HIL92); however, many recent studies have evaluated the possibility of synaptic cross-

talk or volume transmission in the hippocampus by modeling the mass transport of Glu,

observing the architecture of the synaptic region, and local electrophysiological events

(RUS98; ASZ97, BAR97, KUL98). These reports have supported the notion of cross-

talk that is limited to short distances and nearest neighbor synapses. An interesting

observation is that frequency dependent Glu overflow can occur under conditions of

high-frequency stimulation to activate receptors outside the synapse using

electrophysiological techniques in slice preparations (SCA97). Our in vivo results would









be consistent with this observation in that with high frequency stimulation, it is possible

to detect synaptic overflow of Glu that is tonically inhibited by autoreceptors; however,

the striatum appears to be more suited for volume transmission involving Glu. As shown

above, Glu signals spread easier in the striatum than in the hippocampus because of freer

diffusion and less effective uptake. These differences are manifested in larger signals in

the striatum due to electrical stimulation or direct infusion. In contrast, electrically

stimulated overflow is strongly inhibited by mGluR autoreceptors in the striatum whereas

this control is relatively weak, even becoming ineffective (or facilatory) at high agonist

levels, in the hippocampus. In this case, the autoreceptor is viewed as necessary in the

striatum to sense the spread of glutamate and control further release. While volume

transmission of glutamate in the striatum has not been evaluated before, the observation

that Glu receptors are found on cells which do not have synaptic connections to Glu

neurons is supportive of a role for glutamate volume transmission (OLE98). Indeed,

mismatch of receptors and release sites is often given as an important criteria for volume

transmission (ZOL98).

In conclusion, we have demonstrated that high frequency stimuli can cause

synaptic overflow of glutamate in both the hippocampus and striatum. In hippocampus,

such overflow is restricted by uptake and diffusion through the tissue suggesting a

relatively small circle of influence of glutamate. Such a result is consistent with point-to-

point transmission or restricted volume transmission. These results contrast sharply with

prior results from the striatum which revealed large overflows that were regulated

primarily by autoreceptors and relatively less restricted by diffusion and uptake. These

results suggest the potential for more significant volume transmission in the striatum.













Puffer -P


NV


350 ..m
Ik


Figure 5-1. Puffer and probe geometry. Puffer is constructed from a 75 gm i.d.,
360 pm o.d. Fused silica capillary ground to a point with a diameter at the tip of
approximately 90 gm.


*- Probe



















200


180


160


w140


120


100


80


200-


180-


160-


140 -


120-


100-


80-


Figure 5-2. Comparison of Glu response to electrical stimulation in different
brain regions. A) Striatum, stimulation parameters, 10 s, 20 Hz, 0.5 ms pulse,
80 V (7 rats, n = 17). B) Hippocampus, stimlation parameters, 20 s, 20 Hz,
0.1 ms pulse, 10 V (20 rats, n = 80). Bars indicate stimulation.


w^\mlwfm

























Hippocampus Glu ________*

o Striatum Glu

N Hippocampus Asp **
-- Striatum Asp *
1 *


pa *\9m^''*
_____________ __ $LIt


r* *


- - '


lnmi


800

700

600

500

400

300

200

100

0


Control TTX


EGTA ACPD MCPG


PDC PDC +
MCPG


Figure 5-3. Comparison of chemical additives on hippocampus and striatumrn extracellular
concentrations. Basal Levels of glutamate and aspartate were 310 +/- 40 and 87 +/- 11 nM
respectively (n = 21) in hippocampus. Values are expressed as percent of control levels, mean +/-
SEM following a 30 min infusion of the drug (n = 4 for additives). Striatum data is taken from
LAD98


I1 r ,


i


.l 1 i .*.'.t 1 [//J i lA


I


*




















1200a
1100-- U Hippocampus Glu
E1000 Striatum Glu
900
800
700 T -
S600
S500 -
400
300
200
100' *
0'
Control TTX EGTA ACPD MCPG PDC PDC +
MCPG

Figure 5-4. Comparison of chemical additives on hippocampus and striatum Electrically Etimulated
Overflow. Control evoked response was 66.5 +/- 7.3 fmoil (n = 80 stimulations, 21 rats) in
hippocampus. Values are expressed as percent of control levels, mean +/- SEM (n = 16,4 rats).
Striatum data is taken from LAD98.














60 60s
A
50
40
S 30 aCSF PDC
20
'UL
0

60 60s
S50 B
S40
130-
20- aCSF PDC

0


Figure 5-5. Introduction of exogenous 200 pmol D-Asp. 200 nl of 1 mM D-Asp
was microinjected approximately 350 ptm medial from the probe. A) Striatum. 7.4
+/- 1.2 pmol were collected in the control puff, 11.8 +/- 1.8 pmol collected in the
presence of 200 pM PDC via reverse dialysis. B) Hippocampus 2.2 +/- 0.5 pmol
collected in control, 4.8 +/- 1.3 pmol in the presence of 200 pM PDC.

















3.0 3.0
A B
2.5 2.5-

2.0 2.0-

g 1.5 / lS /

S1.0 1.0

0.5 0.5

0.0 0.0
0 10 20 0 10 20
Dialysate (pM) Dialysate (IM)

Figure 5-6 Clearance rates for exogenous D-Asp. A) Striatum. calculated slopes
are: control (diamond) = 0.189 +/- 0.001 s'; PDC (square) = 0.140 +/- 0.007 s'".
B) Hippocampus. calculated slopes are: control = 0.146+/- 0.010 s'1; PDC = 0.089
+/- 0.003 s". Values for concentration and rate are taken from the averaged puffs
in Figure 5-5.














CHAPTER 6
THE ROLE OF GLUTAMATE IN THE PILOCARPINE MODEL OF SEIZURE


Introduction

Cholinomimetics are a class of compounds that affect cholinergic processes in the

brain often resulting in an epileptic-like seizure and death. Because of this effect,

cholinomimetics are used as chemical warfare agents, so called weapons of mass

destruction. Most of the chemicals used in this manner, sarin, soman, and VX (TAY96),

are acetylcholinesterase (AChE) inhibitors. AChE is the enzyme in the brain that removes

acetylcholine from the synaptic cleft by conversion to choline. Through an as yet not

fully understood process, a rise in acetylcholine eventually results in seizures. There is

evidence, however, that these agents also have direct effects on other systems in brain

such as GABAergic and glutamatergic neurons (CHE99, IDR86, MCD97).

A different, less potent cholinomimetic, pilocarpine (PILO), a muscarinic

cholinergic agonist, produces similar seizures to AChE inhibitors and is used as a

substitute for warfare agents in the laboratory for safety reasons. It is also used to

produce model seizures for the study of epilepsy (for review see TUR89). PILO induced

seizures, via intramuscular or intravenous injection, in awake rats are behaviorally

characterized by tremors and gustatory and olfactory automatisms followed by limbic

seizures and status epilepticus. The seizures result in altered pH, blood gas, and glucose

concentrations in the brain (FIJ86, CLI87). Neurological effects can be seen

electroencephalographically a few minutes after injection and include high voltage










spiking and low voltage fast activity. These effects are first seen in the hippocampus,

followed by the amygdala and cortex (TUR83). Brain damage associated with seizure is

greatest in the forebrain and hippocampus and resembles damage observed in the

excitotoxic reaction to Glu (OLN86). Lethality can be as high as 85%, usually occurring

in the first 24 hr during an episode of status epilepticus (JOP86). Spontaneous recurrent

seizures occur for the remainder of the animal's life (TUR83A).

With a pretreatment of lithium, usually between 2 to 24 hr prior to PILO

administration (CHA99), the seizure threshold can be greatly reduced. A 10 fold lower

dose of PILO after Li pretreatment produces identical pathologies to high PILO

treatments only (CLI87). It is suggested that the mechanism for this effect is inhibition of

inositol phosphatase activity which in turn depletes the supply of myo-inositol, a second

messenger system that is utilized by the glutamatergic system in the brain (TRI91).

Even though PILO directly acts on acetylcholine receptors, most of the evidence

pertaining to seizure generation does not point to a cholinergic mechanism for generation.

Most notably, anticholinergic agents are not effective anticonvulsants once a seizure has

been initiated (MAY75). In addition to its direct effect on acetylcholine, PILO has been

shown to modulate both levels of GABA (FRI99, HOU96, TRE92) dopamine (GEO97,

ALA96, STA96), and Glu (HAL96, MIL93, MEL94, SMO97) and their receptors, any of

which may have a role in either seizure generation or maintenance.

Glu has long been believed to be responsible, at least in part, for the generation

and long term effects of seizures mostly because of its excitotoxic effects. Glu's role in

seizures is supported by several lines of evidence (see SHE99, CHAOO, URB98 for recent

reviews) including an increase in Glu and Asp in humans or animals with chronic










seizures. Perhaps most studied as of late, is the activity of Glu receptors before, during,

and after seizure.

Of particular interest here is the effect of metabotropic Glu receptors on seizure

activity. Several studies have been performed to understand how each group of mGluR

participates in seizures. The application of a group III agonist had no effect on kindled

seizures in the rat dentate gyrus and therefore it was concluded that this group has no

effect on seizure generation (FRI99). In a similar experiment it was found that a group II

mGluR selective agonist attenuates kindled seizures, so activation of mGluR II inhibits

seizure activity (KEE99). It has been shown in several studies that group I receptors have

pro-convulsant actions (BOR99, MER97). In particular the mGluR agonist ACPD has

been found to induce seizure activity measured by electrographic activity in hippocampal

slices (TEB97) and in vivo (SAC92).

Unfortunately, the effect of PILO on extracellular Glu does not appear to be as

clear-cut as Glu receptor activity seems to be. There are conflicting published reports on

its effects including, no change, an increase, a decrease, or even a temporary decrease

followed by an increase. Most of these differences may arise from different experimental

protocols, such as introduction method, injection or direct infusion and whether the

animal is awake or anesthetized.

In the awake rat, via intravenous injection, PILO was shown to have no effect on

extracellular Glu in the hippocampus via microdialysis with LC/EC detection (MIL93).

By intramuscular injection, Glu decreased during status epilepticus, remained low for 24

h, then increased and maintained an elevated level for at least 45 days (CAV94). In

another report with intramuscular injection, Glu increased after status epilepticus









(LIU97). Direct infusion through a microdialysis probe immediately decreased Glu in

both striatum (SM097) and hippocampus (KHA99). In the striatum, once PILO was

removed Glu returned to basal levels while in the hippocampus, Glu increased to an

elevated level approximately 200% of the original pre-PILO level. Only one paper has

reported PILO's effect on Glu in anesthetized rats. In this case Glu increased immediately

with PILO infusion via a microdialysis probe in the hippocampus and then returned to

basal 20 min after PILO was removed (MIL93). Behavioral or physical effects of PILO

were seen in all experiments with awake animals while there was no mention of any

effects in anesthetized rats.

In the work presented earlier (Chapters 4 and 5) we demonstrated the ability to

measure neuronal Glu and Asp in the hippocampus (the region suspected to generate

seizures (TUR89)) via microdialysis with on-line capillary electrophoresis. The

regulation of both Glu and Asp via mGluR under normal conditions was demonstrated by

the application of receptor agonists and antagonists. In this chapter, we will apply these

methods to determine the effects of pilocarpine on seizure generation and extracellular

Glu and Asp in the hippocampus of anesthetized rats. Seizure generation was monitored

by physiological changes as well as by electroencephalographic recording at the site of

microdialysis.

Additionally, because electrical stimulation elicits a response in extracellular Glu

and Asp that is completely neuronal in origin it can be used to explicitly study the effect

of PILO on glutamatergic neurons in the hippocampus. In anesthetized rats, stimulation

of the perforant path, similar to the stimulation used in Chapter 4, was shown to evoke

seizure-like high voltage spikes in the dentate gyms (WAL94). In awake animals,