117

Epileptogenic Effects of Radiographic Contrast Agents: Experimental Study

R. Nick Bryan 1 Electrical activity in cells directly exposed to water-soluble radiographic contrast Daniel Johnston2 agents was investigated by intracellular recordings from neurons of the abdominal ganglia of Ap/ysia. Measurements of membrane resting potential, membrane conduct­ ance, synaptic and action potential configuration, and spontaneous electrical activity were performed. Test solutions included sodium diatrizoate, meglumine iothalamate, metrizamide, and control solutions. Solutions (100 and 200 mOsm) of these agents did not significantly alter resting membrane potential, membrane conductance, action potential frequency or configuration, or excitatory postsynaptic potentials. Inhibitory postsynaptic potentials were suppressed by contrast agents. These results suggest that contrast agents affect at least some neurons by disinhibition.

Im aging of the cerebrospinal fluid spaces and adjacent brain surfaces requires contrast enhancement, particularly in the posterior fossa and spine. Recent developments in water-soluble radiographic contrast med ia have permitted th eir substitution for oil y contrast material and air in many cases. Unfortunately, even the newest water-soluble contrast materi als have significant side effects, some apparently related to direct neurotoxicity. Older water-soluble contrast agents have long been known to be extremely toxic when th ey are directly applied to th e central nervous system (CNS). In fact, with older agents, the CNS was the c ritical organ in pharm acologic terms [1 , 2]. This direct neurotoxic ity is primarily reflected by abnormal electrical activity such as electroencephalographi c (EEG) spikes and grossly evidenced as motor seizures. The newer, nonionic, water-soluble contrast agents have markedly diminished epil eptogeni c effects, but motor seizures remain an occasional clinical problem [3]. Mental dysfunction reflected by short-term memory loss, confusion, and disori entation is a more frequent problem with non ioni c agents [4]. The cellular basis for the clinical symptoms and th e altered electrical activity of the brain is essentially unknown. Numerous studies show alteration of EEG

Received July 10 , 198 1; accepted after revi­ patterns with spike activity [5-7], but more specifi c neurophysiologic changes sion September 24 , 1981 . have seldom been studied. This work was supported by U.S. Public Health Spinal root electri cal activity is depressed by hypertoni c solutions of ioni c Service grant NS14907. contrast agents [8], but not by isotonic concentrati ons [9]. Spinal long-tract Presented al th e annual meeting of th e Ameri­ electrical activity is generall y depressed or unaltered [10-1 2]. Th e spinal mono­ can Society of Neu roradiology, Los Angeles, synapti c reflexes are little affected by contrast agents, but some polysynaptic March 1980. activity is markedly increased [11]. Certain spinal inhibitory pathways have been , Departm ent of Rad iology, Baylor Coll ege 0 1 Medicine, Houston. TX 77030 . Address repri nt shown to be depressed [9]. The latter study [9] suggests a more specifi c central requests to R. N. Bryan. M.S. NB703, Methodist synaptic effect of contrast agents which produce epil eptog enic activity by disin­ Hospital. 6565 Bertner, Houston, T X 77030. hibition. While these studies are more specific than EEG records, no single unit 2 Department of Neurology, Baylor Coll ege of or intracellular studies on neurons exposed to radiographic contrast agents have Medicine, Houston, TX 77030. previously been performed. To further el ucid ate the specific neurotoxic effects AJNR 3:117-120, March/ April 1982 0 195-6108/ 82/ 0302-0117 $00.00 of contrast agents, this study involves intracellular electrical recordings during © American Roentgen Ray Society exposure of neurons to various contrast agents. 118 BRYAN AND JOHNSTON AJNR:3, Marchi April 1982

Materials and Methods MEMBRANE RESTING POTENTIAL

The abdominal ganglia of Aplysia ca/ifornica, a marine mollusc , were removed and placed in a circulating seawater bath to which various solutions could be added. Selected neurons [13] were then impaled with potassium-acetate-filled glass mi c roelectrodes with v tips of less than 1 /lm . In the giant cell R2, two electrodes were _50m placed intracellularly. In other cell s, only on e electrode was used. lllll Intracellular electrical activity was monitored by a high impedence .HZO NA MEG . METRIZAMIDE METRIZAMIDE DIATRIZOATE 10THALAMATE preamplifier and fed to a differential amplifier for oscilloscope or 100 mOlm 100 mOlm 100 mOlm ZOO mOlm strip-chart dispt ay (fig. 1). Intracellular recordings allowed mea­ surements of membrane resting potential and membrane re sistance, spontaneous spike frequency and configuration, and electrical 10 mV L ..... characteri sti cs of spontaneou s postsynaptic potentials. 100 sec . In some preparations, connectives (nerves) to the ganglia were Fig . 2. -Representative recordings from cell R2 of abdominal ganglion of stimulated with a constant current suction electrode to produce Aplysia. Each recording was made 30 min after application of test solution at orthodromi c postsynaptic potentials and action potentials. This stated concentrati ons. No change in membrane re sting potential. allowed measurement of thresholds and waveform s of the synaptic and acti on potentials. Th e test solutions used to bathe the ganglia included seawater membrane resistance was obtained by comparison of the (1 ,000 mOsm) , and solutions of seawater plus 10% by volume 100 current-voltage curves of membrane resistance, and no mOsm NaCI ; 1,500 mOsm NaCI; 1,500 mOsm sucrose; 100, 200, displacement of the curves indicative of membrane resting and 300 mOsm sodium diatrizoate; meglumine iothalamate; and potential change was noted (fig . 3). metriza mide. All solutions had a final total osmolarity of 950-1 ,150 Spike configuration, both spontaneous and induced by mOsm. The contrast solutions then had iodine levels of 6-18 mg / ml. The ganglia were exposed to the test solutions for 20-60 min orthodromic stimulation, was evaluated in 12 cells. No al­ and th en rinsed with fresh seawater for a comparable period of terations in the spikes were seen as determined by mea­ time. surements of spike amplitude, duration, and overall wave­ form (fig. 4). Spontaneous spike activity was evaluated in 11 silent Results cells, mainly R2. No alterations (i.e., increase in sponta­ Membrane resting potential was evaluated in 20 cells, neous firing rate) were noted. Spontaneous firing pattern in cluding 11 silent and nine spontaneously active cells. No was also evaluated in nine intrinsically spiking cells. These chang e in membrane resting potential was produced by any cells included L 11 , R15, and those from the L2-6 complex. solution (fig. 2). This was further substantiated by the double In these cells, no consistent change was seen . In seven electrode-current injection method of evaluating membrane experiments, the spike activity was unchanged; in five ex­ resistance, performed on five cells, all R2 . No change in periments, there was a slight increase in spike rate (4 / sec to 6 / sec); and in five experiments, a mild decrease in spike rate was noted (3.5/ sec to 1.5/ sec). There was no con­ sistent relation between the minor alterations in spike activ­ Oscilloscope ity and the various test solutions, including contrast agents. Excitatory postsynaptic potentials (EPSP) were evaluated in si x cells, all R2 by stimulation of the left connective. In High impedance ~ three experiments with metrizamide, no change was seen, preamplifier while in one there was a slight (2-3 mV) depression in "0 0 amplitude. In si x experiments with meglumine iothalamate, no change was seen; in one experiment, there was slight diminution in amplitude of the EPSP (2 mV); in two other experiments, there was a slight increase in EPSP amplitude (3 mV). In no experiment was there any alteration in duration of the postsynaptic potential (fig. 5). In no experiment was current the EPSP increased to spike threshold. source Chart recorder In three cells, all L 11 , spontaneous and orthodromic Bath drain inhibitory postsynaptic potentials (IPSP) were evaluated. In all cases, there was a diminution in amplitude and frequency Fig. 1 .-Preparati on of experiment. Resected abdominal gangli a of Ap/y­ sia are placed in small chamber in which th ey are continuously bathed in of spontaneous IPSP by both ionic and nonionic agents, but various circulati ng test solutions. Specific neurons (such as R2) are then more so by the nonionic agent metrizamide (fig . 6). Likewise, impaled with one or two glass microelectrodes. One electrode is coupled by high impedence preamplifier to oscilloscope differential amplifier for display the orthodromic IPSP elicited by stimulation of the left con­ of intracell ular potentials. If second electrode is used, it is coupled to DC nective was blocked by the contrast agents; however, not current source by high impedence preampli fier. Second electrode is then by th e control solutions of NaCI or sucrose (fig . 7). This used to alter transmembrane poteOl tial by current passage, whi ch is displayed on oscilloscope or strip chart recorder. In addition, vari ous connectives alteration of inhibitory postsynaptic potentials was unasso­ (n erv es) are placed in suc tion electrodes for orthod romic sti mu lation. ci ated with alteration in membrane resting potential. AJNR :3, Marchi April 1982 EPILEPTOGENIC EFFECTS OF CONTRAST AGENTS 11 9

Fig . 3. - Passive membrane resisl­ + 10 + 10 +10 ance. Inlracell ular transmembrane po­ +:; lenlial (mV) is plotted on ordinate as -5 function of intracellular current (nA) mV mV mV passed Ihrough separate electrode, dis­ played on abscissa. Slopes of curve in­ MErRI Z AMIDE 100 mOs m dicale membrane resistance and revea l - 25 - 25 -2S no change after 30 min exposure to var­ ious contrast agents. Dual electrode re­ + 10 + 10 + 10 +S cordings from R2 . OmV = -57 mV RP . +5 0 +~~- 5 -5 -S mV - 10 mV - 10o ~ mV - IS - IS - 20 Noel - 20 MEG 10T HALAMATE METRIZ AMIOE - 2 5 100 mOsm -25 100 mOlm 200 mOsm

- so - 40 - 30 -20 - 10 0 +10 -so - 40 -30 -20 - 10 0 +10 - so · 40 - 30 -20 ' 10 0 " 0 n A nA nA

SPIKE CONFIGURATION Fig . 5. - lntracell ul ar recordin gs INHIBITORY POST of spontaneous inhibitory postsyn­ SYNAPTIC POTENTIALS aptic potentials (lPSP) from L 11 . Re­ sH2 0 MEG . 10THAL. METRIZAMIDE cordings on left are at resting poten­ tial; those on right are hyperpolar­ ized to block spontaneous spike. Meglumine ioth alamate slightly di­ minishes both frequency and ampli­ tude of IPSP, while metrizamide completely blocks IPSP. nL- :_L ---k- Meg. lolholomole 100 mOsm

~_l...--.._ Meg. lolholomole 200 mOsm

Fig . 4. - Spike configuration, recordin g from R2 . Ort hodromic spikes ~ elicited by stimU lation of left connective showed no alteration in configuration after exposure to meglumine iothalamate and metrizamide. Variations in afterpotentials not consistent, related to spontaneous baseline alterations. Melrizomi oe 100 mOsm Calibration 10 mV 110 msec.

MetrizamiDe 200 mOsm EXC ITATORY PO ST SYNAPTIC POTENTIALS

STiM. .H20 MEG 10THAL. METRIZAMIDE 5TR.(VI 100 mOsm 100 mOsm

1.0 J'-- -r- Jl c 3.0 j ...... -- ~ r-- -- ./

8.0 . ~ . .~ . ~---- 10msec Fig . 5.-lntracell ular recordings of excitatory postsynaptic potentials (EPSP) from R2 after st imulation of left connective. Stimulus strengths are at Fig . 7.-0rthodromic IP SP in cell L 11 from slimulati on of left connecti ve . 2, 5, and 15 times threshold (l ast is supramaxim al). Mild diminution in A, Control recording . B, Blockage of IPSP aft er 15 min exposure 10 100 amplitude of EPSP after exposure to meglumine iothalamate and metrizamide. mOsm metrizamide. C, Reversa l of blockage aft er 15 min of seawater rin se. Calibration 10 mV 1 10 msec. Stimulus artifact (arrows).

Discussion to the effects of radiographic contrast agents seems to be Our results indicate th at water-soluble rad iog raphic con­ in hi bitory postsynaptic potentials. This would imply th at at trast agents do not greatl y alter a vari ety of electrical ph e­ least part of the epi leptogenic properties of these drugs is nomena in neurons of the abdominal gangli a of Ap/ysia, due to disinhibition rather than a direct excitatory eff ect. except for inhibitory postsynaptic potentials, whi ch they This is compatible with previous experimental results in th e consistently depress. Th e most sensitive neuronal activity cat spinal cord [9]. Although these are limited data, th ey 120 BRYAN AND JOHNSTON AJNR:3, Marchi April 1982

suggest th at th ese contrast agents may affect neurons sim­ ments, both in this preparation and other epilepsy models, il arly to oth er known convulsant agents, such as pi crotoxin wi ll be necessary to clarify this discrepancy. which also seems to produce epileptogenic effects by dim­ While these experiments are relatively limited, they do inution of inhibitory synaptic activity [14]. In our study rela­ offer th e first information of what may be some of th e specific tively low concentrations of th e contrast agents altered cellular mechanisms of neurotoxicity of water-soluble con­ inhibitory postsynaptic activity and this effect is independent trast agents. This type of approach seems important, not of osmolarity. The normal fluid osmolarity of these animals only in determining the factors of neurotoxicity of these is about 1,000 mOsm and the total osmolarity of the test drugs, but also in evaluating future comp'ounds for improved solutions in th ese experiments was 950-1,1 50 mOsm. Con­ clinical usefulness. trol solutions of NaCI and sucrose have no significant effect on IPSP (or any of th e other parameters measured). REFERENCES These observations imply th at contrast agents directly affect specifi c synaptic activity and th at th eir neuronal ef­ 1. Gonsette RE . Th e neu rophysiolog ic acti on of contrast med ia. fects are not due to th eir hypertonicity or generalized alter­ Acta Radiol [Oiagn] (Stockh) 1972; 13: 889-904 ati on of membrane properties, such as resting potential. The 2. Melartin E, Tu ohimaa PJ , Oabb R. Neurotoxicity of iothala­ exact mechanism of the decreased inhibitory synaptic po­ mates and diatrizoates. I. Significance of concentrati on and tentials cannot be determined from th ese ex perim ents, but cati on. In vest Radio/1970;5: 13-21 3. Strother CM. Ad ve rse reactions. In : Sackett JF, Strother CM , alterations in synapti c C1 mechanisms is a possibility. Why eds. New techniques in myelography. New Yo rk: Harper & th e non ionic agents should have a greater effect th an the Row, 1979: 195-202 clinically more toxic ionic drugs is unclear, but may relate to 4. Bertoni JM , Schwartzman RJ, Van Horn G, Partin J. Asterixis an overall depressant activity of nonionic compounds re­ and encephalopathy foll owin g metrizamid e myelograph y: in­ cently demonstrated in mammali an hippocampus (N. Bryan vestigati ons into possible mechanisms and review of th e liter­ and N. Hershkowitz, unpublished data) and seen in clinical ature. Ann Neuro/1981 ;9: 366-370 EEG s. 5. Grepe A, Wid en L. Neurotoxic effect of intracranial sub­ Since th e effects of these drugs on neurons involves a arachn oid applicati on of metrizamide and meglu mine iocar­ rath er specific chemi cal activity, it seems th at future modi­ mate. An experimental in vestigati on in dogs in neurolept anal­ fications of these agents should be directed toward chemical gesia. Acta Radiol [Suppl] (Stockh) 1973;335: 1 02-118 structure relating to synaptic activity rath er than to decreas­ 6. Oltedal S, Kayed K. Epileptogenic effects of water-soluble contrast media: an experim ental in vestigation in rabbits. Acta ing osmolarity or even lipid solubility. It may, in fact, be Radiol [Suppl] (Stockh) 1973;335: 45-55 fortuitous th at the newer nonionic drugs have less neurotox­ 7. Ropper AH, Chiappa KH, Young RR. The effect of metrizami de icity than their ionic predecessors. The chemical changes on th e electroencephalog ram: a prospective study in 6 1 pa­ resulting in diminution of osmolarity and lipid solubility may ti ents. Ann Neuro/1979;6 : 222- 226 have limited their access to neuronal membrane surfaces, 8. Harvey JP Jr, Freiberger RF , Wern er G. Clinical and experi­ but more importantly, may have altered their effect on inhib­ mental observations with methiodal, an absorbable mye lo­ itory synapses. graphic contrast agent. C/in Pharmacol Th er 1961 ;2: 610-614 One mu st ad mit th at while these drugs do have what may 9. Bryan RN , Oauth GW, Gilman S, Hilal SK . Effects of radio­ be an appropriate effect (i .e., disinhibition) on the neurons graphic contrast agents on spinal cord physiology. Invest Ra­ of thi s inve rtebrate animal preparation , there is an important dio/ 1981 ; 16: 234-239 10. Allen WE III , VanGilder JC, Collins WF III. Eva luation of th e difference between th e deg ree of these effects of Aplysia neurotoxicity of water-solu ble myelographic contrast agents by neurons and neurons of mammali an systems. These drugs electrophysiolog ical monitors. Radiology 1976; 118: 89-95 produce frank " convulsive" activity in va ri ous mammali an 11 . Hilal SK, Oauth GW, Burger LC , Gilman S. Effect of isoton ic preparations including cat cortex and spinal cord, rabbit contrast agents on spin al re fl exes in th e cat. Radiology cortex, and rat cortex and hippocampus (N. Bryan and N. 1977; 122 : 149-155 Hershkowitz, unpublished data). In none of the Aplysia 12. Oltedal SI, Sawhney BB . Effect of water-solu ble contrast media neurons that we investi gated was there ever rapid repetitive on corticall y evoked potentials in th e cat. Acta Radiol [Oiagn] neuronal spiking or " burst" activity, which is generally taken (Stockh) 1973;335: 133-142 to be " epil eptogeni c ." Certainly, oth er convulsants such as 13. Fraz ier WT , Kandel ER , Kupfermann I, Waz iri R, Coggeshall picrotoxin, strychnine, and Metrazol produce more obvious RE . Morphological and functional properties of id entified neu­ neuronal spike activity in Aplysia neurons. It is unclear why rons in the abdomin al ganglion of Aplysia californica. J Neu­ rophysio/1967 ;30 : 1288-1351 contrast agents do not produce such dramatic effects. Two 14. Carpenter ~O , McCreery MJ, Woodbury CM, Yarowsky PJ . possibilities are significant species difference in neuronal Modulati on of endogenous discharge in neuron R15 through sensitivity to th e drugs and/ or a significant binding of these specific receptors for several neurotransmitters. In : Chalazon­ drugs to the membranous capsule that surrounds th e neu­ itis N, Boisson M, eds. Abnormal neuronal discharges. New rons in the abdominal gangli a of Aplysia. Further experi - Yo rk : Raven, 1978: 189- 203