Anatomic localization of topically applied [14C]penicillin during experimental focal epilepsy in cat neocortex

Brain Research, 125 (1977) 293-303

293

© Elsevier/North-HollandBiomedicalPress, Amsterdam- Printed in The Netherlands

ANATOMIC LOCALIZATION OF TOPICALLY APPLIED [14C]PENICILLIN DURING EXPERIMENTAL FOCAL EPILEPSY 1N CAT NEOCORTEX

JEFFREY L. NOEBELS and TIMOTHY A. PEDLEY Department of Neurology, Stanford University Medical Center, Stanford, Calif. 94305 (U.S.A.)

(Accepted August 13th, 1976)

SUMMARY 14C-labeled penicillin was topically applied to the suprasylvian gyri of adult cats in order to study the distribution of the convulsant agent at the onset of focal epileptogenesis. Using radioassay and autoradiographic techniques, a steep gradient of penicillin was found. At the time interictal EEG spike discharges appeared, 95 ~o of the labeled drug was in the uppermost cortical layers (laminae I-III). Analysis of the concentration profiles obtained by scintillation counting showed that penetration of penicillin into brain occurs primarily by passive diffusion. An apparent diffusion coefficient for penicillin in neocortex of 1.5 sq. mm/h was calculated using modifications of standard diffusion equations. It is apparent that with a rapidly acting topical convulsant such as penicillin, the dimensions of the neuronal pool actually in contact with the drug will change significantly over time. The changing boundaries of the epileptic neuronal aggregate must be taken into account when interpreting observations made within and around such experimentally produced epileptic foci.

INTRODUCTION Topical application of penicillin is one of the methods most widely used to produce experimental focal epilepsy. While the electrophysiological events associated with this model of focal epileptogenesis have been well studied1,7,16, the convulsant action of penicillin remains poorly understood. In particular, little information is available about its pattern of spread within brain parenchyma, or its anatomical localization at the onset of surface interictal discharges. Such data are important for two reasons. First, reliable interpretation of behavioral and electrical changes requires knowledge of the size of the neuronal aggregate participating in the epileptogenic process. With a rapidly acting topical convulsant such as penicillin, the dimensions of the neuronal pool actually in contact with the drug will change over time and also vary from experiment to experiment. Thus,

294 knowledge of the linear extent of drug spread is critical to interpreting phenomena occurring outside the focus itself. Second, complete characterization of cellular mechanisms involved in focal epileptogenesis requires information about the anatomical site of drug action which can be correlated with neurophysiological observations. In this study, we have used 14C-labeled penicillin to obtain quantitative information about the kinetics of penicillin movement in cat neocortex, and to make qualitative observations about the localization of the drug at the time of onset of spontaneous interictal activity. This investigation also provides details of methodology suitable for studying diffusable substances in neocortex by autoradiographic techniques. MATERIALS AND METHODS Five adult cats were used in these experiments. Each animal was anesthetized with pentobarbital and mounted in a stereotaxic frame. The calvarium was widely removed over both cerebral hemispheres to expose a large area of cortex. Cortical electrical activity was recorded using silver ball electrodes and displayed on a Tektronix 565 oscilloscope. A rectangular gelfoam pledget (approximately 5 mm × 20 mm) was saturated with 80,000 units/ml of sodium penicillin G containing 10 #Ci/10 /zl of [14C]benzyl penicillin (Amersham/Searle). The use of this size pledget ensured relatively uniform drug application along parallel lines perpendicular to the cortical surface. Following removal of the dura, a pledget was applied bilaterally to each suprasylvian gyrus. When well-developed spontaneous interictal surface discharges were seen, the animal was killed by intravenous injection of KCI. The suprasylvian gyri were rapidly excised (within 15 sec) and instantaneously frozen at --170 °C in isopentane immersed in liquid nitrogen. In two additional cats the same techniques were employed, only the concentration of penicillin used was 5000 units/ml.

Preparation of tissue samples for liquid scintillation counting In a cryostat maintained at --30 °C, the largest possible area of the gyrus beneath the pledget presenting a relatively fiat surface was selected and trimmed to a block approximately 5 mm on a side which was then mounted on the freezing microtome. Tangential sections 10 # m thick were cut from the pial surface through the gray matter to a depth of 2.5-3 mm. Because of the natural curvature of the gyral surface, deeper tangential sections will have a larger surface area than superficial ones. In order to minimize this potential source of error, surrounding parts of the gyrus not in the initial plane of section were carefully dissected away following the first pass of the knife. Subsequent sections then contained tissue cut from an uniform column oriented perpendicular to the cortical surface (Fig. 1).

Counting and standardization of tissue samples Frozen sections 10 # m thick were collected directly from the cryotome blade. Groups of 10 sequential sections representing 100 # m of cortex were pooled, placed in frozen scintillation vials and partially solubilized by osmotic shock and agitation in

II Ill

IV V VI

Fig. 1. Diagrammatic representation of mid-suprasylvian gyrus removed and frozen at onset of.interictal activity produced by application of penicillin pledget (stippled area on surface). Tangential sections were cut from trimmed inner column of tissue for radioassay. Dry-mount autoradiography was performed on coronal section of the gyrus (cut face shown at right).

1 ml of distilled HsO. Aliquots of 300 ~1 were pipetted and frozen in a parallel series of vials for protein determinations used in standardization (see below). After adding toluene-based scintillation fluid (PPO, POPOP), containing Triton X100, each sample was counted twice for 10 min in a Packard Tricarb Liquid Scintillation Spectrometer. Quench corrections were made using tissue blanks and standards. Counting efficiency ranged from 40 “/oto 50 %. In order to permit direct comparison of radioactivity levels from one experiment to another, a simple procedure was used to standardize sample size. We chose to express measured counts per minute in terms of the total amine content in each tissue sample. For this purpose, a fluoresceamine reagent protein assay was used, slightly modified from the method of Bohlen et al .3. A 10 ~1 aliquot of the unknown protein sample was placed in a 13 x 100 mm disposable test tube, and I.5 ml NaHCOa buffer (pH 8.5) was added. During vigorous agitation, 0.5 ml fluoresceamine (FLURAM, Roche) in dimethyl sulfoxide (30 mg/lOO ml) was ejected from a syringe into the test tube. This resulted in the formation of stable fluophors of the primary amino groups with immediate hydrolysis of excess reagent. Fluorescence measurements of the samples and appropriate standards were carried out directly in the test tubes at fixed intervals using a Turner Model 430 Spectrofluorometer with an excitation wavelength at 390 nm and emission at 475 nm. An identical assay of a known volume of cortical gray matter (I 0 ~1) permitted conversion of counts/min/~l protein to units penicillin/ ml cortex. Diffusion ofpeniciliin in agar

For comparative purposes, the diffusion of penicillin into matched columns of isotonic agar was studied in five experiments. The application of [r4C]peniciIlin pledget,

296 diffusion interval, frozen sectioning, and scintillation counting techniques were identical to those used with cortical tissue specimens.

[14C]Penicillin autoradiography One suprasylvian gyrus from each cat was used for autoradiography. Transverse sections 10 # m thick were cut in the cryostat at right angles to the long axis of the gyrus (Fig. 1). To minimize spread of radioactivity by the knife blade, the tissue block was oriented so that each cut started in white matter and ended at the cortical surface, that is, from low to high tissue concentration of label. The sections were then collected on frozen glass slides coated with a paraffin film and freeze-dried over night under vacuum in Coplin jars containing a dessicant (Drierite). Under a safe light, the freezedried sections were dry-mounted in the cryostat at --30 °C onto frozen glass slides coated with Kodak NTB3 nuclear emulsion using light pressure with a teflon film. Exposure was carried out at - - 1 0 0 ° C in light-tight boxes with dessicant. Before photographic development, the slides were brought to room temperature and humidified by breathing gently directly onto the section several times to insure continued adhesion of the tissue to the emulsion. Each slide was then developed for two minutes in Kodak Dektol at 16 °C, rinsed in distilled H~O, and fixed for 5 min in 30 ~o sodium thiosulfate solution. The tissue of the labeled sections, and unlabeled positive and negative chemographic controls 17 were histologically fixed in formaldehyde vapor and post-stained with cresyl violet. RESULTS

Scintillation counting The distribution of labeled penicillin at the onset of surface spike discharges (time ~ 3 min) is shown in Fig. 2A (dotted curve). This figure represents the mean of 5 experiments using 80,000 units of penicillin. Individual concentration profiles were virtually identical from experiment to experiment. Penicillin content varied most in the initial 100/~m, since a visual estimate of the pial surface must be made for the zero point. The concentration profile shown in Fig. 2A indicates that spontaneous interictal electrical activity occurs with the drug located predominantly in the superficial cortical layers: 88 ~o of the total in the initial 500 #m and 96 ~ in the uppermost 1000/zm (cortical laminae I-IlI). Standard diffusion equations may be applied to this data in order to describe quantitatively the kinetics of penicillin movement in neocortex. In equation (Eq.) 1, measured penicillin content ([Pen]) is related to distance (x) from the pial surface; time (t) from initial application of the drug; a diffusion coefficient (D) in units of area/ time; and the probablistic mathematical function 'error function complement' (erfc): [Pen](x,t) - - [Pen] (t=o) [Pen](x=o)-- [Pen](t=o)

x erfc2~/Dt

(Eq. 1)

Since the penicillin content within the brain parenchyma is negligible at time (t) = 0, the term [Pen](t=o) may be omitted and the equation simplified to:

297 [Pen] (x,t)

x = erfc - [Pen](x=0) 2~/Dt

(Eq. 2)

The e q u a t i o n m a y then be r e a r r a n g e d to a f o r m such t h a t a plot o f penicillin c o n t e n t at each d e p t h versus the inverse erfc will be a s t r a i g h t line with a slope o f ~

1

if

d i s t r i b u t i o n o f penicillin occurs p r i m a r i l y by passive diffusion: Y = 2~/t

•

erfc_l [Pen](x.t) [Pen](x=0)

--

x.

1 ~/D

(Eq.

3)

T h e g r a p h shown in Fig. 2B was o b t a i n e d in this way. A c o m p l e t e d e s c r i p t i o n o f this m e t h o d o f analysis w i t h c o m p u t e r plots o f the e q u a t i o n s is given in F i s h e r et al. 10. Eq. 1-3 are a p p l i c a b l e to the d a t a only if a c o n s t a n t c o n c e n t r a t i o n o f penicillin exists at the cortical surface. F o r a n a l y t i c purposes, the p l e d g e t m a y be a s s u m e d to represent

[PENICILLIN] VS DEPTH A 80 L"

l

n

u

~ •

I

136

THEORETICAL OBSERVED

25

36.7

20

Y

VS

DEPTH

I

3.0

Z

o I,~C I,z ILl 0 Z 0 0

I

15 2.5

10

0 .i-

_1

- - 13.9 --

11.8

--

9.2

_ ----

5.8 4,8 3.5 2.7

--

0.1

>=

1,5 ooo

m

5 4 3 2 1 0

0.0

-

0.5

1.0

--

1,5

DEPTH (mm)

~-

2.0

1.0

000000 ooOo

0.5

• oOO go pO

0.0 0.0

I

I

1.0

2.0

3.0

DEPTH (mm)

Fig. 2. A: averaged data showing relationship between penicillin concentration and depth below cortical surface (dotted curve). Bars represent standard errors. A computer-generated theoretical curve (solid line) for penicillin diffusion with D = 1.5 sq.mm/h is shown for comparison. Note sharp deviation of theoretical curve from experimental data near the cortical surface. This is probably due to a surface barrier (e.g., pial-glial membrane) restricting entry of penicillin into the brain, which is not accounted for by the theoretical model. Refer to text for further discussion. B: transformation of the data points shown in A. See text for explanation.

298 an infinite reservoir of penicillin since the counted label remaining in the pledget at t ~- 3 min was within 1 ~ of the value present at application. It is apparent from Fig. 2B that penicillin penetration into brain is consistent with a diffusional process. Using the slope of the averaged data plotted in Fig. 2B, we calculated an apparent diffusion coefficient for penicillin in neocortex of 1.5 sq. mm/h. If a theoretical diffusional curve for 80,000 units penicillin using D = 1.5 sq. mm/h is compared to the experimental data, there is good agreement (Fig. 2A, solid line) confirming that penicillin movement in neocortex closely fits a model of passive diffusion. Preliminary experiments using lower concentrations of penicillin show similar findings. The theoretical curve deviates sharply from the observed data near the cortical surface since it does not take into account a barrier (e.g., pial membrane and associated glial connective tissue) which restricts entry of penicillin into the brain. The non-zero y-intercept of the points in Fig. 2B (arrow) is the consequence of such a barrier and the displacement along the y-axis reflects its magnitude. An extensive discussion and mathematical treatment of the surface barrier is given in Fisher et al. 10. To test the validity of our conclusions regarding the spread of penicillin in cortex, we studied penicillin movement in columns of isotonic agar using similar proceA

B C O U N T S VS D E P T H AGAR

70

I

6C

I

I

- -

THEORETICAL

•

OBSERVED Y VS D E P T H

I

50

I

3.0 40 2.5

~

-

30

~ >.

20

•

2.0

1.5

•

1.0 10

0.5

O000000Q '

0 oO0°Q

eee ee • 0.0

o.o

0.0

0.5

1.0

1.5

DEPTH (mm)

2.0

|

t

I

1.0

2.0

3.o

DEPTH (mm)

Fig. 3. A: representative depth profile for penicillin in agar (dotted curve) a n d a theoretical curve (solid line) with D -- 1.7 sq. m m / h . N o surface barrier to penicillin diffusion exists in agar so there is no discrepancy between experimental data a n d c o m p u t e r - g e n e r a t e d theoretical curve. B : t r a n s f o r m a tion o f the d a t a points s h o w n in A. See text for explanation.

299

Fig. 4. A: dry-mount autoradiogram showing distribution of [14C]labeled penicillin at onset of interictal activity. B: dark-field photomicrograph of another autoradiogram showing in more detail the penicillin gradient in cortex at onset of surface spiking. In both examples, the apparent irregularity at the top of the section, as well as some of the intense radioactivity on the surface, is due to a small piece of gelfoam pledget adhering to the pia. Roman numerals indicate approximate location of the different cortical layers determined from control sections processed specificallyfor histological detail. Magnification in both A and B, × 50. dures. The profiles obtained (Fig. 3A) closely match those found for penicillin in brain. The transformation of the data shown in Fig. 3A is illustrated in Fig. 3B. The computed D for penicillin in agar was 1.7 sq. mm/h. Note that since no surface barrier to penicillin entry exists in agar, there is no discrepancy between the theoretical and observed curves in Fig. 3A, and there is no shift along the y-axis of the points in Fig. 3B.

A utoradiography Dry mount autoradiograms showed a penicillin concentration gradient qualitatively similar to that described quantitatively by scintillation counting (Fig. 4A). A discontinuity is clearly present, with silver grain density being intense at the pial surface, and a lower, rapidly decreasing density profile into the deeper cortical layers corresponding to laminae I I I and IV. This discontinuity may be the consequence of a

300 physical barrier at the surface (see discussion above) or tissue binding to the pial-glial membrane. Dark-field photomicrography indicated that virtually no labeled penicillin was visible in the deepest layers (Fig. 4B) and adjacent white matter. Inspection of the autoradiograms at high magnification showed homogeneous distributions of grains over cellular and extracellular spaces, without any clear evidence of regions of intense membrane affinity or concentration within cellular elements. Owing to the use of a high energy radioactive source, no definite statement can be made regarding the subceUular localization of penicillin. Control slides did not indicate contamination by artifactual exposure from positive or negative chemography, or pressure. DISCUSSION A crucial issue raised by experimental studies of focal epilepsy using topical convulsants is whether abnormal firing patterns occur only in neurons actually in contact with the drug, or whether a limited group of neurons - - the 'epileptic aggregate' - - recruit nearby drug-free cells into the paroxysmal activity by alterations in normal synaptic transmission. Some observations have suggested that abnormal neuronal responses may occur in areas presumed to be remote from the direct action of the convulsant agent7,11. Dichter and Spencer 7 found paroxysmal depolarization shifts in neurons outside an area stained by a vital dye mixed with the penicillin. They concluded that the 'excitatory zone' included tissue not directly affected by drug presence. More recently, Gabor and Scobey11 have argued that ectopic spike generation ('backfiring')lz can occur, at least during ictal activity, outside a penicillin focus, the limits of which were defined by visual field stimulation. The low resolution of such indirect estimates of drug concentration, however, is insufficient to demonstrate whether the intracortical spread of epilepsy is due to drug spread or other pathophysiological mechanisms. Several studies have used radioassay techniques to localize convulsant drugs in neocortex more precisely. Stark et al. 21 and Edmonds et al. a investigated the distribution ef [~4C]penicillin in rat brain following a single intracortical injection. While they found that 'most of the radioactivity remained near the injection site' they did not study the diffusional kinetics quantitatively nor were they able to describe accurately the cortical areas affected by the drug, since no precautions were taken against diffusion of the unbound label during autoradiographic exposures. Using a different experimental model, Cornblath and Ferguson 5 reported that topically applied [ZH]acetylcholine (ACh) was found throughout the cortex and white matter at 'seizure onset'. Two methodological problems are present in this study, however, which make direct estimates of the depth profile at the onset of seizures difficult to interpret. First, the authors' assumption that the measured labeled compound represented ACh may not be valid. Even though they attempted to prevent hydrolysis of the applied ACh by pretreatment of the cortex with neostigmine, it is not clear that neostigmine and ACh will penetrate into, and diffuse through, the brain at equal rates. If ACh moves faster than neostigmine, it may well be hydrolyzed and the

301 [3H]acetyl moiety rather than the intact ACh molecule might be found in all layers of cortex. Second, neostigmine alone has a convulsant action4,9, so it is unclear whether the profiles obtained relate to epileptiform activity induced by ACh or neostigmine. Willmore et al. 23 applied ionic cobalt electrophoretically to the cortical surface of rats. At the time epileptiform activity appeared, the cobalt (stained by silver substitution method) was deposited in a cone-shaped area restricted to the supragranular layers I-Ill. The authors do not state the sensitivity of the chemical assay used in the quantitative analysis of cobalt metal in brain tissue, and without this information, a steep but nevertheless continuous drug gradient should be assumed. Our results are generally similar to those of Willmore et al. 23. We found that when interictal epileptiform activity first appeared, 95 ~ of the labeled penicillin was located in the top 1000 #m (laminae I-III) with 87 ~ in the initial 500/~m. The corresponding neuronal structures in contact with the highest concentrations of penicillin include principally the dendritic arborizations of deeper-lying large pyramidal cells, axonal terminals from callosal and association afferents and small pyramidal cells19. In addition, a small population of inhibitory interneurons has been described at these depths 22. Autoradiography using dry-mount techniques to prevent diffusion of unbound labeled drug during exposure is provided visual confirmation of the profiles obtained by scintillation counting. Analysis of the concentration profiles obtained by scintillation counting shows that penetration of penicillin occurs primarily by passive diffusion with an apparent D for penicillin in neocortex of 1.5 sq.mm/h. This value for D must be qualified, however. It is derived from measurements made at a single interval following penicillin application. It is therefore theoretically valid only for that particular time unless the assumption is made that distribution of penicillin continues to occur solely by passive diffusion. Experiments designed to look at the location of penicillin at later times might disclose deviations from a model of pure diffusion indicating tissue binding of the drug, cellular uptake, or different rates of diffusion in different cortical layers. We have no information on these speculations. Our data in cat cortex is supported by comparing the cortical drug profiles with those obtained from study of penicillin movement in agar. Similar conclusions were reached by Baleydier et al. 2 based on their autoradiographic studies of the intracerebral diffusion of penicillin away from a chronic alumina cream lesion following an intravenous injection. The ratio of diffusion coefficients Dagar/Deortex was somewhat smaller than expected. One might predict that penicillin would diffuse considerably faster thlough agar than cortex because of the effectively shorter travel path due to the absence of tortuous interstitial spaces. For example, sucrose, a compound with a molecular weight (mol.wt. 342) similar to penicillin (mol.wt. 356) is thought to move only through the extracellular space, and has diffusion coefficients of 1.3 sq. mm/h and 2.55 sq. mm/h in brain and agar respectively 14. Two considerations may be pertinent to the difference encountered with penicillin. First, penicillin anions (in contrast to the uncharged sucrose molecule) may react with the agar gel or encounter particles to which they bind resulting in apparent slower movement. Diffusion coefficients of salts in agar gels are typically 10 ~ less than those obtained in aqueous solution2°. Second, our

302 experiments in agar were carried out at room temperature (approximately 25 °C) whereas in vivo experiments were at 37 °C. Since the diffusion coefficient varies directly with temperature, the net effect would be to underestimate the value of D for penicillin in agar. Although the majority of labeled drug was confined to superficial cortical layers, a continuous gradient was evident, since penicillin was detectable in low concentrations (1-2 raM) in the deepest laminae. It is important to realize that the values shown as penicillin 'concentration' in Fig. 2A actually represent the total intracortical penicillin content per unit volume (ml) cortical tissue at each depth without regard to intraor extracellular distribution. Since it is probable that the large penicillin anion is relatively impermeant and thus restricted mainly to extracellular spaces (approximately 15-20 ~ of total brain volume), local concentrations surrounding neuronal elements may be 5-6 times higher than indicated in Fig. 2A. While the minimal concentration of penicillin necessary to alter the cellular properties of single cortical neurons is not precisely known, concentrations of 1-6 mM have been shown to depress recurrent inhibition in amphibian spinal cord 6 and reduce the amplitude of IPSPs in the isolated crayfish stretch receptor 15 and crab neuromuscular junction 13. It is entirely possible that penicillin produces very different physiological effects at the extremes of the dosage range present. Therefore it is not possible from our data to delineate any specific intracortical region that must be affected by penicillin to produce synchronous epileptic discharges. The relatively rapid diffusion of penicillin in cortex indicates that the spatial extent of the convulsant drug varies continuously with time. Electrophysiological studies which attempt to characterize abnormal cellular events within and around a topical penicillin focus, must therefore take into consideration the changing boundaries of the epileptic neuronal aggregate. Approximate concentrations of the drug at any time or distance may be calculated using the apparent diffusion coefficient for penicillin in neocortex. ACKNOWLEDGEMENTS We would like to thank Nancy Cline and Carol Ann Ineson who introduced us to the techniques of autoradiography. Dr. Robert Fisher assisted with the diffusional analysis. We had many helpful discussions with Professors David A. Prince and Robert W. P. Cutler. Ms. Geraldine Chase provided secretarial assistance. This work was supported by a fellowship from the Epilepsy Foundation of America (J.L.N.), N I H grants NSl1075 (T.A,P.) and NS06477 (D.A.P.) and the Morris Research Fund.

REFERENCES 1 Ajmone Marsan, C., Acute effects of topical epileptogenic agents. In H. H. Jasper, A. A. Ward, Jr. and A. Pope (Eds.), Basic Mechanisms of the Epilepsies, Little, Brown and Co., Boston, Mass., 1969, pp. 299-319.

303 2 Baleydier, C., Bisconte, J. C. and Quoex, C., Autoradiographic study of the blood brain barrier for penicillin in chronic experimental epilepsy, Acta Neuropath. (BerL), 24 (1973) 321-330. 3 Bohlen, P., Stein, S., Dairman, W. and Udenfriend, S., Fluorometric assay of proteins in the nanogram range, Arch. Biochem., 155 (1973) 213-220. 4 Celesia, G. G. and Jasper, H. H., Acetylcholine released from cerebral cortex in relation to state of activation, Neurology (Minneap.), 16 (1966) 1053-1064. 5 Cornblath, D. R. and Ferguson, J. H., Distribution of radioactivity from topically applied [H3]Acetylcholine in relation to seizure, Exp. Neurol., 50 (1976) 495-504. 6 Davidoff, R. A., Penicillin and pre-synaptic inhibition in the amphibian spinal cord, Brain Research, 36 (1972) 218-222. 7 Dichter, M. and Spencer, W. A., Penicillin-induced interictal discharges from the cat hippocampus. I. Characteristics and topographical features, J. NeurophysioL, 32 (1968) 649-662. 8 Edmonds, H. L., Stark, L. G. and Hollinger, M. A., The effects of diphenylhydantoin, phenobarbital, and diazepam on the penicillin-induced epileptogenic focus in the rat, Exp. NeuroL, 45 (1974) 377-386. 9 Ferguson, J. H. and Jasper, H. H., Laminar DC studies of acetylcholine-activated epileptiform discharge in cerebral cortex, Electroenceph. clin. Neurophysiol., 30 (1971) 377-390. 10 Fisher, R. S., Pedley, T. A. and Prince, D. A., Kinetics of potassium movement in normal cortex, Brain Research, 101 (1976) 223-237. 11 Gabor, A. J. and Scobey, R. P., Spatial limits of epileptogenic cortex: Its relationship to ectopic spike generation, J. NeurophysioL, 38 (1975) 395-404. 12 Gutnick, M. J. and Prince, D. A., Thalamocortical relay neurons: Antidromic invasion of spikes from a cortical epileptogenic focus, Science, 176 (1972) 424-426. 13 Hochner, B., Spira, M. E. and Werman, R., Penicillin decreases chloride conductance in crustacean muscle: A model for the epileptic neuron, Brain Research, 107 (1976) 85-103. 14 Levin, V. A., Fenstermacher, J. D. and Patlak, C. S., Sucrose and inulin space measurements of cerebral cortex in 4 mammalian species, Amer. J. PhysioL, 219 (1970) 1528-1533. 15 Meyer, H. and Prince, D. A., Convulsant actions of penicillin: effects on inhibitory mechanisms, Brain Research, 53 (1973) 477-482. 16 Prince, D. A., Topical convulsant drugs and metabolic antagonists. In D. P. Purpura, J. K. Penry, D. Tower, D. M. Woodbury and R. Walter (Eds.), Experimental Models of Epilepsy, Raven Press, New York, 1972, pp. 51-83. 17 Rogers, A. W. and John, P. N., Latent image stability in autoradiography of diffusible substances. In L. J. Roth and W. E. Stumpf (Eds.), Autoradiography of Diffusible Substances, Academic Press, New York, 1969, pp. 51-68. 18 Roth, L. J. and Stumpf, W. E. (Eds.), Autoradiography of Diffusible Substances, Academic Press, New York, 1969. 19 Sanides, F. and Hoffman, J., Cyto- and myeloarchitecture of the visual cortex of the cat and of the surrounding integration cortices, J. Hirnforsch., 11 (1969) 79-104. 20 Schantz, E. J. and Lauffer, M. A., Diffusion measurements in agar gel, Biochemistry, 1 (1962) 658-663. 21 Stark, L. G., Edmonds, H. L. and Hollinger, M. A., The distribution of 14C-penicillin in rat brain following the induction of an epileptogenic focus, Proc. Western Pharmacol. Soc., 17 (1974) 51-54. 22 Szent~lgothai, J., Architecture of the cerebral cortex. In H. H. Jasper, A. A. Ward, Jr. and A. Pope (Eds.), Basic Mechanisms of the Epilepsies, Little, Brown and Co., Boston, Mass., 1969, pp. 13-28. 23 Willmore, L. J., Fuller, P. M., Butler, A. B. and Bass, N. H., Neuronal compartmentation of ionic cobalt in rat cerebral cortex during initiation of epileptiform activity, Exp. Neurol., 47 (1975) 280-289.