A C T A O P H T H A L M O L O G I C A V O L . 5 3 1 9 7 5

The Department of Ophthalmology (Head: M . N . Luxenberg, M.D.) and the Department of Physiology (Head: R. C. Little, M.D.)

Medical College of Georgia, U.S.A., and the University Eye Department (Head: Th. L. Thomassen, M.D.) Rikshospitalet, Oslo, Norway

THE EFFECT OF BENZALKONIUM CHLORIDE ON THE ELECTROPOTENTIAL OF THE RABBIT CORNEA

BY

KEITH GREEN and ASBJ0RN M.T0NJUM

The effect of benzalkonium chloride on the electropotential of the cornea has been examined. The anterior surface of the in vivo or in vitro cornea was exposed to various concentrations of the surfactant, from 0.005 "/o to 0.02 O/o, for either 1 or 2 min. The initial effect is a hyperpolarization last- ing up to 30 sec, followed by a rapid fall in potential difference with a subsequent recovery. The degree of potential difference decrease and the recovery rate was dependent upon both the concentration of the detergent, and the exposure time. There is excellent correlation between the previous anatomical and physiological studies on tracer penetration across the in vivo and in vitro cornea and our present work. The data indicate that benzalkonium chloride acts by breaking down the physiological and ana- tomical diffusion barrier to solute and solvent which is located in the outer layer of the epithelium.

Key words: benzalkonium chloride - cationic surfactant - rabbit - cornea - electropotential.

Benzalkonium chloride (BA), a widely used preservative in ophthalmic prepara- tions, has been shown to enhance the permeability of the rabbit cornea in vitro to both fluorescein (Green & Tsnjum 1971) and a large molecule, horseradish peroxidase (Tsnjum 1975).

Received January 22, 1975.

348

Benzalkonium and Corneal Potential

The normal corneal epithelium is impermeant to horseradish peroxidase (PO) when the tracer is applied to the anterior surface. When applied posteriorly it moved between the epithelial cells to the tight junctional complexes adjacent to the tear film between the most superficial cells (Tsnjum 1974), thus demon- strating that the barrier is located at the surface. This is in agreement with Green (1969) who showed that the anterior part of the epithelium is more re- sistant to water movement than the posterior layer, and with electrophysiological studies of Ehlers (1973) and Klyce (1973).

The tracer technique with PO demonstrated that BA in vitro produced defects of the cell membranes allowing PO to pass into and through the superficial cells.

O’Brien & Swan (1942) found that benzalkonium chloride increased the in vivo effect of carbachol upon the intraocular pressure. Lund-Karlsen & Fonnum (1975) showed increased intraocular penetration of cholinesterase reactivators after topical BA application in vivo. Green & Downs (1974 a, b) demonstrated that BA enhanced the ocular penetration of prednisolone phosphate and pilo- carpine.

In light of these previous findings an investigation was made of the effect of BA on the electropotential of the cornea and the rate of possible recovery of the potential on both the in vivo and in vitro rabbit cornea.

Materials and Methods

Adult, healthy albino rabbits, 2-4 kg, and of each sex were used.

In vitro experiments

The rabbits were killed with an overdose of sodium pentobarbital administered via a marginal ear vein. An eye was proptosed, and the cornea, together with a scleral rim, removed and placed in Krebs bicarbonate Ringer with glucose added at 5 mg/ml (hereafter referred to as Ringer). The iris and lens were then gently removed and the isolated corneas mounted in chambers as described previously by Green (1966). These chambers allow a hydrostatic pressure head (15 mmHg) to be placed on the posterior surface of the cornea. The chamber was immersed in a beaker containing sufficient Ringer to cover the exposed corneal surface. All experiments were performed at room temperature.

The potential difference (PD) across the cornea was recorded with one of two systems: A Heath EU-20B recorder with a voltage clamp attachment, or a Keith- ley Model 160 Digital Multimeter. Both systems are capable of measuring PD to an accuracy of k 0.1 mV with 10 mV full scale deflection. With the Keithley

349

Keith Green and Asbjwrn M . Tsnjzim

multimeter, agar/saturated KC1 (made in PE160 tubing) connected the solutions bathing each surface of the cornea to beakers also containing saturated KCl, and the connection to the voltmeter was made via Beckman calomel electrodes. Using the Heath recorder, the agar bridges connected the bathing solutions directly to calomel electrodes, and these connected to the recording system.

Experiments were performed using normal Ringer initially on both sides of the cornea. At 15 min after mounting the tissue in the chamber all tissues ex- hibited a stable control PD. An identical Ringer solution containing either 0.005, 0.01 or 0.020/0 benzalkonium chloride (BA) was then placed in another beaker and the chamber placed in this solution for either 1 or 2 min. At the end of this exposure to BA the chamber was returned to a beaker containing Ringer alone. The PD across the cornea was measured continuously during all phases. At least four corneas were used at each concentration and exposure time.

In vivo experiments

The rabbits were anesthetized with intravenous urethane (25 O/o w/v in 0.9 O/o

NaCl). After a suitable depth of anesthesia had been induced, a corneal cup which had a height of 2 cm and a diameter of about 1 cm was placed on the cornea. The rabbit was placed on its side for this procedure and only one eye per animal was utilized to avoid problems associated with exposure or abrasion of the contralateral cornea. The cup had a rim at the bottom to fit at the corneal limbus and produce a leak-proof fit with the eye. Ringer was placed in this cup and the PD across the cornea measured between an electrode placed in this solution and an electrode inserted into a marginal ear vein. The agar/KCl electrodes used in these experiments were made of PE 50 tubing, to facilitate entry of the electrode into the vein.

After a steady PD was recorded, the Ringer in the cup was replaced by an identical solution containing BA at 0.005 or 0.1 O/O for 1 min and this was further replaced by normal Ringer. The PD was measured almost continuously during all these phases using a Heath EU-20B recorder with a voltage clamp attachment. After the first 10 min following reapplication of normal Ringer to the cornea, the PD was measured at 10 min intervals, with the corneal cup removed from the eye and the lids held together with a clamp when PD was not measured.

Results

Some results obtained from exposure of isolated cornea to benzalkonium chloride (BA) are presented in Table I, where certain trends are evident. BA usually

350

Benzalkonium and Corneal Potential

Concentration and time

Table I . Values of hyperpolarization, percentage reduction and recovery of the corneal potential difference after exposure of the anterior surface of the in vitro cornea to various concentrations of benzalkonium chloride for 1 or 2 min. The time of minimum PD is given as the time a t which the PD reached a minimum to the time a t which higher values were consistently found. The recovery of original PD is the value found 2 hours

after exposure to the surfactant

Hyper- Maximal Recovery after Time of

PD (min)

polarization reduction 2 hours oio of original oio of original o / o of original

P D PD PD

0.005 O / o 1 rnin 13 91 15 10

2 min 64 86 5 to 10

0.01 " io 1 min 50 95 5 to 20 45

2 min 81 50 10 to 60

0.02 O / o 1 min 80

2 min 91 51

63 5 to 30

18 5 to 60

caused an initial, concentration-dependent hyperpolarization of the cornea. This response was short-lasting, with the potential difference (PD) peaking between 20 and 30 sec after exposure to BA. The PD then fell rapidly in most cases, in a concentration-dependent manner, to a minimum. The decrease in PD was also greater with longer exposure times (Table I).

The PD recovered at varied rates towards normal pre-treatment levels when returned to Ringer solution. The recovery rate was also exposure-time and concentration-dependent (Table I). I t is immediately evident that a lower con- centration for 2 min is almost equivalent to a higher concentration for 1 min.

In vitro experiments

0.005 O/O B A . One minute exposure of rabbit cornea to BA resulted in a small reduction of PD from 3.97 f 0.30 (SEM) mV, with the maximum fall at 15 min, to 3.48 & 0.18 mV (P < .01) (Fig. 1 a). The average hyperpolarization for 1 and 2 rnin was 0.4 mV. The average maximum reduction in PD was 130/0. PD re- covery was 91 O/O of the control value at 2 hours.

35 1

Keith Green and Asbjern M. Tenjiini

I b

TIME Irnl""*.l

Fig. I. Effect of 0.005 O/o benzalkonium chloride on the electropotential of the isolated rabbit cornea. PD, potential difference (mV), time in minutes. 0-0, 1 min exposure to BA;

0-0, 2 min exposure to BA. Values are the mean _+ SEM.

O U - + - + - 0 10 20 30 6 0 9 0 1 2 0

TIME I r n i n u l e l l

F i g . 2. Effect of 0.01 O/o benzalkonium chloride on the electropotential of the isolated rabbit

cornea. For explanation of symbols see legend to Fig. 1 .

352

Benzalkonium and Corneal Potential

Two minute exposure to BA caused a significantly greater reduction of PD from 5.55 k 0.35 mV, with the maximum fall from 5 through 10 min, to 2.0 k 0.2 mV (P < .001) (Fig. 1 b). The average maximum reduction in PD was 640/0. PD recovery was 860/0 of the control value at 2 hours (960/0 at 3 hours). 0.01 O l o B A . One minute exposure caused a fall in PD from 2.45 k 0.35 mV, with a maximum fall from 5 through 20 min, to 1.23 k 0.26 mV (P < .001) (Fig. 2 a). The hyperpolarization for 1 and 2 min was 1 . 1 mV. The average maximum reduction in PD was 500/0. PD recovery was 950/0 of the control value at 2 hours.

Two minute exposure caused a fall in PD from 3.9 k 0.4 mV, with a maxi- mum fall from 10 through 60 min, to 0.75 & 0.15 mV (P < .001) (Fig. 2 b). The average maximum reduction in PD was 81 O/O. PD recovery was 500/0 of the control value a t 2 hours ( 7 7 O/O at 3 hours). 0.02 O/O B A . One minute exposure caused a reduction of PD from 2.35 k 0.24 mV, with a maximum fall from 5 through 50 min, to 0.45 k 0.16 mV (P < .001) (Fig. 3 a). The hyperpolarization for 1 and 2 min was 1.2 mV. The average maximum reduction in PD was 800/0. PD recovery was 630/0 of the control value at 2 hours.

Two minute exposure caused a reduction of PD from 2.45 k 0.18 mV, with a maximum fall from 5 through 60 min, to 0.25 k 0.17 mV (P < .001) (Fig. 3 b). The average maximum reduction in PD was 91 O/O. PD recovery was 180/0 of the control value at 2 hours.

: k c z z 0 10 20 30 60 90 I20

TIME I m , n u m L

F i g . S . Effect of 0.02 "/o benzalkonium chloride on the electropotential of the isolated rabbit

cornea. For explanation of symbols see legend to Fig. 1 .

353

Keith Green and Asbjorn M . Tsnjiini

0 I I , . , 6 7 8 9 1020m.060 T l Y F ,Illr4..)

Fig. 4 . Effect of benzalkonium chloride on the electropotential of the in vivo rabbit cornea. PD, potential difference (mV), time in minutes. 0-0, 0.005°/~i for 1 min: 0-0,

0.01 "/a for 1 min.

In vivo experiments

One minute exposure of the cornea to 0.005 O/O BA caused a small depression in PD from 6.9 k 0.8 to 6.5 k 0.7 mV (.2 < P < .01). The fall in PD is 5 -I- 3 O/O

(.2 < P < . l ) and is not significantly different from the control value. One minute exposure of the in vivo rabbit cornea to 0.01 O/O BA caused a fall

in PD from 5.5 k 1.1 to 3.4 k 1.0 mV (P < .01). The fall in PD is 36 t- 8 O/O

(P < .001) and is significantly different from the control value. The time of maximum decrease is at 5 min (Fig. 4), with recovery of the PD to control values at about 20 min.

Discussion

The PD of the isolated rabbit cornea shows a characteristic behaviour after exposure to BA. At high doses the P D increases within the first 30 sec and this hyperpolarization is followed by a marked fall. The magnitude of the fall is both exposure-time and concentration-dependent. Low (0.005 O/O) concentrations only elicited small hyperpolarizations, a minor fall in PD and recovery of PD is complete within 30 to 60 min. As the concentration increases, the initial hyper- polarization and the fall in PD are increased and the subsequent recovery is longer. Longer exposure times also extend the length of the maximum fall, and the extent of recovery, of the PD (Table I).

The cornea of the living rabbit also responds to BA when added to the ante- rior corneal surface (Fig. 4). The fall in PD after 0.01 O/O BA for 1 min is, how- ever, less than that elicited by the same concentration in vitro (36 O/O compared

354

Benzalkonitrm and Corneal Potential

to 50 O/O) and the effect is of shorter duration. Thus, there is a difference between the in vivo and in vitro effects of the cationic surfactant.

Anatomical evidence (Tsnjum 1974) indicates that the normal corneal epi- thelium is impermeant to horseradish peroxidase (PO). After treatment of the in vitro cornea with BA, PO penetrates into and between the surface cells of the epithelium. After treatment of the in vivo cornea with BA and PO simul- taneously there is an effect of the surfactant but, compared to the in vitro effects, the tracer penetration is not as marked (Tsnjum 1975). There is, therefore, good agreement between the anatomical and electrophysiological data, since the in vitro effects are more marked than the in vivo effects. The anatomical evidence indicates that BA acts by inducing a lysis of the membrane of the outer cell layers of the epithelium, thus removing the barrier function of these cells (Tlanjum 1975). The electrophysiological and anatomical data indicate that the PD is reduced by a breakdown of this barrier which creates a shunt pathway for the movement of solutes and solvent.

These findings are substantiated by other work on both in vitro and in vivo effects of BA. Previously, we have described the effects of BA on fluorescein permeability of the in vilro cornea where 4 min exposure of the cornea to 0.01 O/o

BA caused a 12-fold increase in fluorescein transfer (Green 8- Tenjum 1971). I n vivo studies using radioactive prednisolone sodium phosphate and pilocarpine (Green &: Downs 1974 a, b) reveal that BA, 0.01 O/O, causes only a 500/0 increase in drug penetration 15 min after the administration of one drop (50 p1) to the eye. Lund-Karlsen k Fonnum (1975) examined the entry of cholinesterase re- activators into the eye and found that BA caused a significant increase between 31 and 126010 in drug penetration.

The in vivo values indicate less enhancement in permeability than found with the same concentration of BA in vitro. Our current electrophysiological data in vivo reveal decreases in PD which are less than those found in the in vitro cornea. The recovery of the PD is much more rapid in vivo and again is presum- ably related to a quicker recovery of barrier function.

The recovery of the PD is of importance, since it reveals that the cornea is capable of undergoing regeneration of the barrier function of the membrane even in vitro. Recovery of the barrier function may be in a manner similar to that observed in the normal cornea when dead, pre-desquamating cells offer no resistance to the entry of PO across the cell boundaries. Although the cell is leaky, PO does not pass beyond this cell and the anatomical barrier is established posterior to the dead cells (Tsnjum 1974). Re-establishment of the barrier func- tion may be made posterior to the lysed superficial cells in the recovery phase of the present experiments.

The present experiments resemble those made on toad bladder using another

355

Keith Green and Asbjern M . Tsnjum

cationic surfactant, cetyl pyridinium chloride (CPC) (Saladino, Hawkins & Trump 1971). The short-circuit current of the toad bladder recovered from low doses of CPC ( l O - W ) , whereas high doses (l0-3M) caused apparently irreparable damage. Webb (1965) also found that cationic surfactants at concentrations of 10-4M caused irreversible changes in the PD across isolated frog skin. Schoffe- niels, Gilles & Dandrifosse (1962) used BA a t 50 pg/ml on isolated frog skin and the PD fell to about 150/0 of the original a t 5 min. After washing the membrane the PD recovered to about 50 to 600/0 of the original value after 6 hours (200 /0 after 2 hours).

The hyperpolarization seen after exposure of corneas to BA is of interest. Such a phase was not observed in the toad bladder short-circuit current measure- ments by Saladino et al. (1971). They did note an increased initial oxygen uptake of the tissue following detergent treatment, at a time which corresponds to the electrophysiological changes observed by ourselves. Schoeffeniels et al. (1962) and Webb (1965) observed an initial hyperpolarization of the isolated frog skin with surfactants. The hyperpolarization was thought to be caused by disruption of cellular integrity through interactions of the surfactant with calcium in the membrane. Such an interaction changes the permeability of the membrane and results in a transient efflux of potassium from the cell which results in a hyper- polarization.

The effects of BA on the electrophysiology of the rabbit cornea have been examined, and the findings have correlation with both in vitro and in viuo studies using both radioactive and large molecular weight tracers. It is evident that the primary effect of BA is to eliminate the physiological and anatomical barrier to solute and solute movement in the superficial corneal epithelium by a process of destruction of the cell membranes and lysis of the cells.

Acknowledgements

This study was supported in part by a Public Health Research Grant EY 01413 from the National Eye Institute (K.G.) and in part by the Norwegian Research Council for Science and Humanities (A.M.T.) Travel funds for K. Green were generously made available by the Norwegian Research Council for Science and Humanities.

References

Ehlers, N . (1973) Ztz vi tro studies of trans- and intraepithelial potentials of the cornea.

Green, K. (1966) Active control of corneal thickness. Li le Sci. 5, 2309-2314. Exp. Eye Res. 13. 553-565.

356

Benzalkonircm cind Corneal Potential

Green, K. (1969) Anatomic study of water movement through rabbit corneal epithelium. Amer. J. Ophthal. 67, 110-116.

Green, K. & Downs, S. J. (1974 a) Prednisolone phosphate penetration into and through the cornea. Invest. Ophthal. 13, 316-319.

Green, K. & Downs, S. J. (1974 b) Ocular penetration of pilocarpine in rabbit. Arch. Ophthal., in press.

Green, K. & Tenjum, A. M. (1971) Influence of various agents on corneal permeability. Amer. J . Ophthal. 72, 89i-905.

Klyce, S. J. (1973) Relationship of epithelial membrane potentials to corneal potential. Exp. Eye Res. 1.5, 567-575.

Lund-Karlsen, R. & Fonnum, F. (1975) The effect of locally applied cholinesterase inhibitors and oximes on the acetylcholinestrase activity in different parts of the eye. Brit. 1. Pharm., in press.

O’Brien, C. S. & Swan, K. C. (1942) Carbaminoylcholine chloride in the treatment of glaucoma simplex. Arch. Ophthal. 27, 253-263.

Saladino, A. J., Hawkins, H. K. & Trump, B. F. (1971) Ion movement in cell injury. Amer. J. Path. 64, 271-294.

Schoffeniels, E., Gilles, R. & Dandrifosse, G. (1962) Action des dktergents et du calcium sur le potential Clectrique de la peau de grenouille isoike. Arch. Int. Physiol. Pharm. 70, 335-344.

Tenjum, A. M. (1974) Permeability of horseradish peroxidase in the rabbit corneal epithelium. Acta ophthal. (Kbh.) 52, 650-658.

Tenjum, A. M. (1975) Permeability of the rabbit corneal epithelium to horseradish peroxidase after the influence of benzalkonium chloride. Acta ophthal. (Kbh.) 53, 335- 347.

Webb, G. D. (1965) The effects of surfactants on the potential, short-circuit current, and ion fluxes across the isolated frog skin. Acta physiol. scand. 63, 377-384.

Authors’ address: A. M. Tenjum, M.D., University Eye Department, Rikshospitalet, Oslo 1, Norway.

Acta ophthal. 53. 3

357