- Email: [email protected]
Depression of benzodiazepine binding and diazepam potentiation of GABA-mediated inhibition after chronic exposure of spinal cord cultures to diazepam
Brain Research, 268 (1983) 171-176
171
Elsevier Biomedical Press
Depression of benzodiazepine binding and diazepam potentiation of GABAmediated inhibition after chronic exposure of spinal cord cultures to diazepam PHYLLIS K. SHER l , ROBERT E. STUDY 2, JOANN MAZZETTA 2, JEFFERY L. BARKER 2 and PHILL1P G. NELSON l
l Laboratory of Developmental Neurobiology, National Institute of Child Health and Human Developmentand" Laboratory of Neurophysiology, National Institute of Neurologic and Communicative Disordersand Stroke, National Institutes of Health, Bethesda, MD 20205 (U.S.A.) (Accepted January 25th, 1983)
Key words: benzodiazepine - diazepam - inhibition - spinal cord cultures - GABA
Cultures of fetal mouse spinal cord were exposed to 12.6/~M (3.6/~g/ml) diazepam for 7 days. After drug removal, benzodiazepine receptor binding was assayed on intact cells and intracellular recordings of diazepam effects on GABA-mediated inhibitory responses were obtained. The biochemical and electrophysiological data revealed significant and parallel reductions in both receptor binding and pharmacological action on GABA responses which did not return to control levels until 3-4 days after removal of diazepam. The results indicate that chronic exposure of spinal cord cultures to diazepam results in a reversible down-regulation ofdiazepam binding and function.
The development of tolerance is a characteristic consequence of repeated or long-term administration of a variety of neuroactive drugs and there is some evidence to suggest that such an effect may result from a reduction in the number or affinity of the receptors for these drugs (see ref. 14 for review). The benzodiazepines (BDZs) are used in the treatment of a wide variety of disorders including spasticity secondary to spinal cord dysfunction9. Although a clinical syndrome of reduced sensitivity to some of the therapeutic effects of the BDZs is well-documented3.4-s, attempts to reveal changes in BDZ receptors after chronic exposure to these drugs have yielded inconsistent results 6,13.:,18.j9.22. Furthermore, no determinations of the effects of chronic drug exposure on the pharmacological effects of BDZs at the cellular level have been reported. We have studied the effects of chronic exposure to diazepam using dissociated mouse spinal cord (SC) cultures, which include both neurons and non-neuronal cells. This preparation offers the advantages of eliminating pharmacokinetic uncertainties inherent in whole-animal studies, thereby ensuring that only direct effects of the
drug will be measured. SC cells were obtained from 12-14-day fetal mice (C57BL6) and prepared as described previously~6. The cells were grown on collagen-coated 16 mm diameter cluster wells (24 wells/tray; Linbro) for receptor binding or on 35 mm diameter plates (Falcon) for electrophysiological recordings. The medium in which the cells were maintained was changed twice weekly. On the 23rd day after plating, diazepam (a gift from Hoffmann-LaRoche) was added to the medium to give a final concentration of 12.6 /~M (3.6 /~g/ml) (determined by Herner Analytics, Rockville, MD, by the electron capture method). The drug was added as a solution containing 360/.tg/ml diazepam, 10% ethanol, 40% propylene glycol, and 50% water. In cluster trays, diazepam was added to 6 wells while paired control wells were maintained with feeding medium to which only the drug vehicle was added. Our previous studies have indicated that the vehicle itself is not toxic to cells24, and does not affect BDZ receptor binding in the concentration used here (P.K. Sher, unpublished data). In addition, the vehicle does not alter GABA-mediated, C 1--dependent conductance responses (R.E. Study, unpublished
172 data). During the 7-day period of drug exposure, the cultures were fed as usual, with fresh diazepam (or vehicle) included with each medium change. After this period of chronic exposure, the cells were returned to medium without drug or vehicle. The medium was always changed 20 24 h prior to each assay, since we found that at this time the specific diazepam binding was highest. Assays for total, specific and clonazepam (CLO)-displaceable diazepam binding were performed on intact cells as described previously2~ at 0 h, 24 h, 72 h, 96 h and 10 days after drug removal. CLO-displaceable binding is thought to reflect a population of BDZ receptors found only on neurons 7,~:.22. Cells in each well were thoroughly washed 3 times with ice-cold, isosrnotic phosphate buffer (50 mM K:HPO4 KH2PO 4, pH 7.4, brought to 330 mOsm with NaCI) and incubated for 30 rain on ice in buffer containing 5 nM [3H]diazepam (76.8 Ci/mmol, NEN) alone, with 3/xM unlabeled diazepam, or with 0.1/~M CLO for determination of total, specific, and CLO-displaceable binding, respectively. After incubation, the cells were thoroughly washed 3 times with ice-cold buffer and removed by trituration with 0.2 N NaOH. For scintillation counting, a i ml portion was neutralized with 1 N He1. Specific binding was taken as the difference between the total binding and the binding in the presence of 3 #M diazepam. CLO-displaceable BDZ binding was calculated as the difference between specific binding and binding in the presence of 0.1 tLM CLO. All values were obtained in quadruplicate and compared to paired controls from wells in the same tray. The results of the binding studies are shown in Fig. 1. After 7 days exposure to diazepam and less than 1 h after drug removal (0 time in Fig. 1), total and specific BDZ binding were significantly reduced from control values (44% and 26% of control, respectively; P % 0.001, paired ttest), and CLO-displaceable binding was not detectable. At 24 h, CLO-displaceable binding began to recover (29% of control, P ~ 0.001), as did total and specific binding, which increased to 60% percent and 46% of control, respectively
T
80
T
,
"
0
24h
0-'4"
~!
72h
Q"
i
~
i
i il
96h
lOd
TlvlE AFTER DRUG REMOVAL
Fig. I. Recovery of BDZ receptor binding after 7 day exposure of neuron-rich spinal cord cultures to 12.6/~M diazepam. Mean control values were 319 + 15 fmol/well (n = 20) for total (T) BDZ binding, 228 _+ 9 fmol/wetl (n = 20) for specific (SP) BDZ binding; and 116 _+ 10 fmol/well for CLO-displaceable ( e L ) binding. Data shown is the mean of 4 determinations expressed as a percent of control values _+ S.E.M.
(P ~ 0.001). By 72 h, CLO-displaceable, total, and specific binding had all recovered substantially, but were still significantly different from controls (73%, P % 0.05; 75%. P % 0.01; and 69%, P % 0.001; respectively). However, by 96 h only specific binding remained significantly less than control values (66%, P % 0.001), and at 10 days all binding values had returned to control levels. In order to determine if this phenomenon depended upon the presence of neurons, similar experiments involving chronic drug exposure were performed on SC cultures virtually devoid of neurons. As anticipated, CLO-displaceable binding was not present in either control or drug-treated cultures. In a corresponding manner, absolute control values for total and specific BDZ were proportionately reduced relative to wells with neurons: total BDZ binding was 175 + 9 fmol/well (mean + S.E.M.) and specific BDZ binding was 93 _+ 6 fmol/well (n = 20) compared to total BDZ binding of 319 ___ 15 fmol/well (n = 20), specific binding of 228 _+ 9 fmol/well (n = 20) and CLO-displaceable binding of 116 _+ 10 fmol/well (n = 18) in neuronrich cultures. In cultures without neurons, significant differences between drug-treated and control cultures
173 were observed only at 0 time after drug removal, when total BDZ binding was 67 __. 7% of control (n = 8; P ( 0 . 0 1 ) and specific BDZ binding was 37 _ 7% (n = 8; P ( 0 . 0 0 1 ) . Thus, the binding of diazepam to non-neuronal cells was less affected by chronic drug exposure than was the binding to neurons. In addition, the non-neuronal component of binding recovered more rapidly after the drug was removed. The effect of chronic diazepam exposure was also assayed electrophysiologically in SC neurons from cultures derived from the same embryos and treated identically as those used for binding studies. BDZ receptor function was evaluated by comparing the increase in GABAinduced C1- ion conductance produced in the presence and absence of acutely applied diazepam tt . Membrane conductance was determined from intracellular recordings using a single microelectrode containing 3 M KC1. Before recording, the cultures were washed 5 times with a solution containing (in mM): 107 NaC1, 5 KC1, 10 MgCI2, 2 CaC12, 12 HEPES, 5 D-glucose, with 1 /zM tetrodotoxin and 25 mM tetraethylammonium chloride included to block voltage-dependent Na ÷ and K ÷ conductances. Recordings were made in this medium at 23 ___ 2 °C. The inclusion of elevated [Mg2+]o and tetrodotoxin in the recording medium attenuated synaptic and electrical activity, allowing better quantitation of conductance responses. GABA and diazepam were applied by pressure (1 2 lbs./in. 2) from individual pipettes with 3-5 ~m tip diameter, placed 10-50 ~m from the neuronal soma. The pipette position, concentration of GABA (10-20/~M), and (most frequently) duration of GABA application (20-200 ms) were varied to produce an easily quantifiable conductance response. After a stable response was achieved, diazepam (12.6 ~M) was applied for 10 s. Several seconds after the diazepam application was terminated, GABA was again applied. Since the effect of diazepam outlasts the period of drug application23, this procedure allowed reliable quantitation of responses without problems of flow deflection that would occur if the two drugs were simultaneously applied. Conductance was measured by injecting hyper-
CI
I lu~ufi~lmU~m,m,w~,-,'=q
...........
V
b
~1 o 20
see
f ...............I............JJl3lll. II. . . . . .
I
_
V
p_~
,, ~
n
~
n
Fig. 2. Diazepam effects on GABA-mediated membrane conductance in spinal cord neurons from control cultures (a) and cultures chronically treated with diazepam (b). Hyperpolarizing constant-current pulses (I) were passed through the recording electrode to assay membrane conductance. P indicates pressure application of GABA (upward deflections) or diazepam (downward deflections). Membrane potential (V) was held at - 6 0 mV before GABA application by constant hyperpolarizing current. The hyperpolarization produced by diazepam application in (b) is not an effect of the drug but reflects the flow of medium resulting from drug application. This is commonly seen in both control and treated cultures, and also occurs with the application of medium without drug. In (a) the conductance increase produced by GABA is 7.8 nS and after maximal diazepam effect, 28.2 nS (a 260% increase). In (b) the conductance increase produced by GABA is 18.3 nS and is enhanced slightly (as evidenced by the increased voltage deflection) but the enhancement is so little that it is not different from control values, when evaluated by the constant-current pulses. The increased response to 20~M GABA in (b) relative to (a) is due to the cell-to-cell variation in responsiveness. The average responsiveness of drug-treated neurons to GABA did not differ from controls (see text).
polarizing constant-current pulses (50 ms, 0.21.0 nA) through the recording electrode using a bridge circuit. All cells were held at - 6 0 _+ 3 mV (in the absence of GABA or diazepam) using steady current injection as necessary. Typical intracellular recordings obtained with control cells and after chronic diazepam exposure are shown in Fig. 2. Most voltage responses to GABA were depolarizing due to the change in the CV ion gradient induced by the use of intracellular electrodes filled with KC1. The maximal effect of di-
174 azepam on G A B A responses required about the same time to develop in both control and drugtreated cultures. The maximal effect of the drug was used for all determinations. Since the responsiveness to G A B A and to diazepam varies from cell to cell, an effort was made to obtain stable, reliable records from as m a n y cells as possible at each time point (Fig. 3). Following chronic drug exposure (0 6 h after diazepam removal) the increase in GABA-mediated conductance produced by applied diazepam was reduced to 12% of that observed in control cultures (P ~< 0.0005, unpaired t-test). At this time point, diazepam produced a 170 _+ 21% increase in G A B A - m e d i a t e d m e m b r a n e conductance in neurons from control cultures, an effect that did not change significantly at other time points tested. 48 h after drug removal, the effectiveness of diazepam in potentiating GABA responses recovered partially (to 42% of
100"
°
80-
§ 60-
0
48h
72h
96h
TIME AFTER DRUG REMOVAL
Fig. 3. Recovery of the potentiating effects of diazepam on GABA responses after chronic exposure to the drug. Responses were calculated as the increase in membrane conductance produced by GABA in the presence of diazepam relative to that produced by GABA alone. Control values were obtained from cells chronically exposed only to the drug vehicle. Mean values of responses are shown as percent of control first at the end of the chronic exposure period (0-6 h after diazepam removal, shown as 0) and then during a 4day drug-free recovery period. Results are combined from two experiments; bars indicate S.E.M. Control values were similar at each time point (see text). The number of cells studied in control and drug-treated cultures, respectively, was: 0 h, 23 and 35; 48 h, 7 and 6; 72 h, 11 and 17; and 96 h, 7 and 11.
control), but due to the smaller n u m b e r of adequate recordings (6), the results were different from controls only at the 0.1 level of significance. At 72 and 96 h after drug removal, the potentiating effects of diazepam still averaged less than that for corresponding controls (93% and 65% of control, respectively), but the differences were not statistically significant due to the scatter of the data. The results indicate that chronic exposure of cultured SC cells to a clinically relevant concentration of diazepam a,~ leads to a marked reduction both in the ability of diazepam to bind to BDZ receptors on intact cells and to potentiate m e m b r a n e responses to GABA. There are several explanations that might account for this. It is possible that the BDZ receptors have been 'down-regulated' by reducing either their affinity or number. Alternatively, it is possible that the drug was retained by the cells after chronic exposure so that the reductions in binding and potentiation of G A B A responses was due either to continued occupation of the receptors by the drug or to release of the drug from the cells into the m e d i u m followed by re-occupation of receptors. In order to examine this possibility, culture dishes which had been exposed to 12.6 ttM diazepam for 7 days were thoroughly washed 3 times at room temperature with 1 ml portions of 50 m M phosphate buffer (pH 7.4) and dried. Tissue was then scraped off the dish, suspended in this buffer, and sonicated. The resulting homogenate was centrifuged at 150,000 ,g for 10 min to pellet m e m b r a n e s and insoluble material, a procedure similar to that reported by Paul et al. ~5. Supernatant derived from the equivalent of a single treated culture well was added to each untreated well in the binding assay. Supernatant obtained using the same protocol from vehicletreated cultures was also added to untreated wells. Untreated wells without added supernarant served as additional controls. In addition, the same supernatants were applied by pressure to naive neurons to test for any effects on G A B A - m e d i a t e d responses. Neither supernatant had any effect in either the biochemical or electrophysiological assays. These results indicate that the supernatant does not contain
175 detectable drug as evaluated in these two types of biological assay. Thus, it is unlikely that the treated cells were inadequately washed or retained enough drug to account for the observed reductions in binding and function, although the possibility that diazepam remained tightly bound to the membrane fraction has not been excluded. One way that diazepam could be retained in membranes would be if the chronic exposure decreased the off-rate of the drug from BDZ receptors. The reduction in binding and pharmacologic action of diazepam might result from continued occupation of receptors, with recovery representing the off-rate. Theoretically, continued receptor occupation might produce a continued potentiation of GABA-mediated chloride conductance, which would be detectable as an enhanced response to GABA (in the absence of applied diazepam) after chronic exposure to drug. However, the mean value of conductance responses to 20/~M GABA recorded at 0 time after removal of drug was not significantly different (12.8 + 1.8 nS, n - 32) from that observed in cells chronically exposed to vehicle only (11.2 _+ 1.8 nS, n = 23). Therefore chronic activation of the BDZ receptors as measured in the electrophysiological assay appears unlikely. Our observations are consistent with either a down-regulation of BDZ receptors or a prolonged occupation of receptors by diazepam where the occupation no longer produces a response. At present we have not distinguished between the two alternatives experimentally, but studies of the effects of chronic exposure of other receptors to agonists would lead us to prefer the former ~4. The results obtained with both the binding and electrophysiological assays showed a similar pattern, although the depression of diazepam action was somewhat greater than the reduction in total and specific diazepam binding at 0 time after drug removal. This difference may be related to the fact that the electrophysiological data were obtained exclusively from neurons, while specific binding includes both neuronal and non-neuronal sites. CLO-displaceable binding, which was virtually eliminated by chronic diazepam exposure, may be the component of
binding that corresponds to BDZ receptors coupled to GABA-activated CI-- channels in neurons. This conclusion is further supported by the observation that reduction of binding was substantially less in neuron-poor cultures. In earlier studies, researchers have reported an increase6, modest decrease ~7.~ or no change2-L3.22 in the BDZ receptor after chronic treatment. However, more recent work has shown that chronic administration of clonazepam to mice induces BDZ receptor 'subsensitivity' for several days following cessation of drug treatment5, and we have found that dissociated cerebral cortical cultures from mice show a more prolonged decrease in BDZ binding than spinal cord cultures after chronic diazepam treatment ~9. In addition, a prominent withdrawal response after administration of the BDZ receptor antagonist, R015-1788, has been produced in baboons treated with diazepam for 7 days ~° suggesting that the drug exposure time in this study is sufficient to produce drug tolerance and dependence. Our data also suggest that the potentiation of GABA-mediated conductance and CLO-displaceable binding represent similar receptor populations, even though the affinities of the specific (KD ~ 11 nM) and CLO-displaceable (Kt~ ~ 5 riM) sites are much greater than the apparent affinity of the neuronal receptors coupled to GABA-mediated Ck channels (about 2-5 /~M; R.E. Study and J. L. Barker, unpublished observations). This discrepancy has not yet been resolved, but it is possible that a population of binding sites of lower affinity ( K o about 200 nM) which we have recently found in SC cultures-'° is responsible for the potentiation of GABA responses. It is important to note that the concentration of diazepam that was used for chronic exposure of the cultures (12.6 I~M) is not only high enough to produce a maximal or near maximal potentiation of GABA-mediated conductance and saturate the receptors assayed by binding, but is also in the range associated with a significant clinical response (see ref. 1 for review). The precise mechanisms of the profound alterations in specific receptor binding and poten-
176
tiation of GABA-mediated inhibition remain to be elucidated. We do not yet know if the changes in binding represent a decrease in number of receptors or their affinity, or how the reduction of the effect of diazepam in the electrophysiological assay is achieved. Even though it is not yet clear how the parameters measured in the two assays are related, each has a pharmacological profile which correlates well with the clinical efficacy of the BDZs, and thus may represent a ba-
sis for the clinically important actions of these drugs. The observation that both binding and pharmacological action are greatly reduced after long-term exposure to diazepam may therefore represent cellular correlates of the clinical observation of tolerance seen in patients and in animals chronically treated with this class of drug. The authors would like to thank Ms. Linda Bowers for her photographic assistance.
I Bowling, A. C. and DeLorenzo, R. J., Micromolar affinity benzodiazepine receptors: identification and characterization in the central nervous system, Science. 216 (1982) 1247 1250. 2 Braestrup, C., Nielsen, M. and Squires, R. F., No changes in rat benzodiazepine receptors after withdrawal from continuous treatment with lorazepam and diazepam, LifeSci., 24(1979) 347 350. 3 Browne, T. R. and Penry, J. K., Benzodiazepines in the treatment of epilepsy. A review, Epilepsia, 14 (1973) 277 310. 4 Covi, L., Lipman, R. S., Pattison, J. H., Derogatis, L. R. and Uhlenhuth, E. H., Length of treatment with anxiolytic sedatives and response to their sudden withdrawal, Actap~vchiatr. scand., 49 (1973) 51 64. 5 Crawley, J. N., Marangos, P. J., Stivers, J. and Goodwin, F. K., Chronic clonazepam administration induces benzodiazepine receptor subsensitivity, Neuropharrnacologv. 21 (1982)85 89. 6 DeStefano, P., Case, K. R., Colello, G. D. and Bosman, H. B., Increased specific binding of [3H]diazepam in rat brain following chronic diazepam administration, Cell Biol. Int. Rep.. 3 (1979) 169 167. 7 Gallager, D. W., Maltor~a, P., Oertel, L. W., Henneberry, R. and Tallman, J., [ H]Diazepam binding in mammalian central nervous system: a pharmacologic consideration, J. Neurosci., 1 (1981)218 225. 8 Greenblatt, D. J. and Shader, R. J., Dependence, tolerance, and addiction to benzodiazepines. Clinical and pharmacokinetic considerations, Drug Metab. Rev., 8 (1978) 13 28. 9 Lossius, R., Dietrichson, P. and Lunde, P. K. M., Effect of diazepam and desmethyldiazepam in spasticity and rigidity, Acta neurolscand.. 61 (1980) 378 383. 10 Lukas, S. E. and Griffiths, R. R., Precipitated withdrawal by a benzodiazepine receptor antagonist, Science, 217 (1982) 1161 1163. I1 Macdonald, R. L. and Barker, J. L., Enhancement of GABA-mediated postsynaptic in cultured mammalian spinal cord neurons: a common mode of anticonvulsant action, Brain Research, 167 (1979) 323- 336. 12 McCarthy, K. D. and Harden, T. K., Identification of two benzodiazepine binding sites on cells cultured from
rat cerebral cortex, J. Pharmacol. exp, Ther., 216 (1981) 183 191. Mohler, J., Okada, T. and Enna, S. J., Benzodiazepine and neurotransmitter receptor binding in rat brain after chronic administration of diazepam or phenobarbital, Brain Research, 156 (1978) 391 - 395. Overstreet, D. H. and Yamamura, H. l., Receptor alterations and drug tolerance, Life Sci., 25 (1979) 1865 1877. Paul, S, M., Syapin, P. J., Paugh, B. A.. Moncada, V. and Skolnick, P., Correlation between benzodiazepine receptor occupation and anticonvulsant effects of diazepam, Nature (Lond,), 281 (1979) 688 -689. Ransom, B. R., Neale, E,, Henkart, M., Bullock, P. N. and Nelson, P. G., Mouse spinal cord in cell culture. I. Morphology and intrinsic neuronal electrophysiological properties, J. Neuropt~vsiol., 40 (1977) 1132 t 150. Rosenberg, H. C. and Chiu, T. H., Regional specificity ot" benzodiazepine receptor down-regulation during chronic treatment of rats with flurazepam, Neurosei. Let& 24 (1981)49 52. Rosenberg, H. C. and Chiu, T. H., Decreased [3H]diazepam binding is a specific response to chronic benzodiazepine treatment, Life Sci.. 24 (1979) 803 808. Sher, P, K., Reduced benzodiazepine receptor binding in cerebral cortical cultures chronically exposed to diazepare, Epilepsia. in press. Sher, P, K., Development and differentiation of the benzodiazepine receptor in spinal cord cultures, Develop, Brain Res., in press. Sher, P. K., Schrier, B. K. and Van Putten, D., An in situ assay for determination of benzodiazepine binding, Develop. Neurosci.. 5 (1982) 271 277. Shibla, D. B., Gardell. M. A. and Neale, J. H., The insensitivity of developing benzodiazepine receptors to chronic treatment with diazepam, GABA, and muscimol in brain cell cultures, Brain Research, 210 ( 1981 ) 471 474. Study, R. E. and Barker, J. L., Diazepam and (-)-pentobarbital: fluctuation analysis reveals different mechanisms for potentiation of gamma-aminobutyric acid responses in cultured central neurons, Proc. nat. Acad. Sci. U.S.A.. 78 (1981) 7180- 7184. Swaiman, K. F., Neale, E. A., Schrier, B. K. and Nelson, P. G., Toxic effect of phenytoin on developing cortical neurons in culture, Ann. Neurol., 13 (1983) 48-52.
13
14 15
16
17
18 19 20 21 22
23
24
171
Elsevier Biomedical Press
Depression of benzodiazepine binding and diazepam potentiation of GABAmediated inhibition after chronic exposure of spinal cord cultures to diazepam PHYLLIS K. SHER l , ROBERT E. STUDY 2, JOANN MAZZETTA 2, JEFFERY L. BARKER 2 and PHILL1P G. NELSON l
l Laboratory of Developmental Neurobiology, National Institute of Child Health and Human Developmentand" Laboratory of Neurophysiology, National Institute of Neurologic and Communicative Disordersand Stroke, National Institutes of Health, Bethesda, MD 20205 (U.S.A.) (Accepted January 25th, 1983)
Key words: benzodiazepine - diazepam - inhibition - spinal cord cultures - GABA
Cultures of fetal mouse spinal cord were exposed to 12.6/~M (3.6/~g/ml) diazepam for 7 days. After drug removal, benzodiazepine receptor binding was assayed on intact cells and intracellular recordings of diazepam effects on GABA-mediated inhibitory responses were obtained. The biochemical and electrophysiological data revealed significant and parallel reductions in both receptor binding and pharmacological action on GABA responses which did not return to control levels until 3-4 days after removal of diazepam. The results indicate that chronic exposure of spinal cord cultures to diazepam results in a reversible down-regulation ofdiazepam binding and function.
The development of tolerance is a characteristic consequence of repeated or long-term administration of a variety of neuroactive drugs and there is some evidence to suggest that such an effect may result from a reduction in the number or affinity of the receptors for these drugs (see ref. 14 for review). The benzodiazepines (BDZs) are used in the treatment of a wide variety of disorders including spasticity secondary to spinal cord dysfunction9. Although a clinical syndrome of reduced sensitivity to some of the therapeutic effects of the BDZs is well-documented3.4-s, attempts to reveal changes in BDZ receptors after chronic exposure to these drugs have yielded inconsistent results 6,13.:,18.j9.22. Furthermore, no determinations of the effects of chronic drug exposure on the pharmacological effects of BDZs at the cellular level have been reported. We have studied the effects of chronic exposure to diazepam using dissociated mouse spinal cord (SC) cultures, which include both neurons and non-neuronal cells. This preparation offers the advantages of eliminating pharmacokinetic uncertainties inherent in whole-animal studies, thereby ensuring that only direct effects of the
drug will be measured. SC cells were obtained from 12-14-day fetal mice (C57BL6) and prepared as described previously~6. The cells were grown on collagen-coated 16 mm diameter cluster wells (24 wells/tray; Linbro) for receptor binding or on 35 mm diameter plates (Falcon) for electrophysiological recordings. The medium in which the cells were maintained was changed twice weekly. On the 23rd day after plating, diazepam (a gift from Hoffmann-LaRoche) was added to the medium to give a final concentration of 12.6 /~M (3.6 /~g/ml) (determined by Herner Analytics, Rockville, MD, by the electron capture method). The drug was added as a solution containing 360/.tg/ml diazepam, 10% ethanol, 40% propylene glycol, and 50% water. In cluster trays, diazepam was added to 6 wells while paired control wells were maintained with feeding medium to which only the drug vehicle was added. Our previous studies have indicated that the vehicle itself is not toxic to cells24, and does not affect BDZ receptor binding in the concentration used here (P.K. Sher, unpublished data). In addition, the vehicle does not alter GABA-mediated, C 1--dependent conductance responses (R.E. Study, unpublished
172 data). During the 7-day period of drug exposure, the cultures were fed as usual, with fresh diazepam (or vehicle) included with each medium change. After this period of chronic exposure, the cells were returned to medium without drug or vehicle. The medium was always changed 20 24 h prior to each assay, since we found that at this time the specific diazepam binding was highest. Assays for total, specific and clonazepam (CLO)-displaceable diazepam binding were performed on intact cells as described previously2~ at 0 h, 24 h, 72 h, 96 h and 10 days after drug removal. CLO-displaceable binding is thought to reflect a population of BDZ receptors found only on neurons 7,~:.22. Cells in each well were thoroughly washed 3 times with ice-cold, isosrnotic phosphate buffer (50 mM K:HPO4 KH2PO 4, pH 7.4, brought to 330 mOsm with NaCI) and incubated for 30 rain on ice in buffer containing 5 nM [3H]diazepam (76.8 Ci/mmol, NEN) alone, with 3/xM unlabeled diazepam, or with 0.1/~M CLO for determination of total, specific, and CLO-displaceable binding, respectively. After incubation, the cells were thoroughly washed 3 times with ice-cold buffer and removed by trituration with 0.2 N NaOH. For scintillation counting, a i ml portion was neutralized with 1 N He1. Specific binding was taken as the difference between the total binding and the binding in the presence of 3 #M diazepam. CLO-displaceable BDZ binding was calculated as the difference between specific binding and binding in the presence of 0.1 tLM CLO. All values were obtained in quadruplicate and compared to paired controls from wells in the same tray. The results of the binding studies are shown in Fig. 1. After 7 days exposure to diazepam and less than 1 h after drug removal (0 time in Fig. 1), total and specific BDZ binding were significantly reduced from control values (44% and 26% of control, respectively; P % 0.001, paired ttest), and CLO-displaceable binding was not detectable. At 24 h, CLO-displaceable binding began to recover (29% of control, P ~ 0.001), as did total and specific binding, which increased to 60% percent and 46% of control, respectively
T
80
T
,
"
0
24h
0-'4"
~!
72h
Q"
i
~
i
i il
96h
lOd
TlvlE AFTER DRUG REMOVAL
Fig. I. Recovery of BDZ receptor binding after 7 day exposure of neuron-rich spinal cord cultures to 12.6/~M diazepam. Mean control values were 319 + 15 fmol/well (n = 20) for total (T) BDZ binding, 228 _+ 9 fmol/wetl (n = 20) for specific (SP) BDZ binding; and 116 _+ 10 fmol/well for CLO-displaceable ( e L ) binding. Data shown is the mean of 4 determinations expressed as a percent of control values _+ S.E.M.
(P ~ 0.001). By 72 h, CLO-displaceable, total, and specific binding had all recovered substantially, but were still significantly different from controls (73%, P % 0.05; 75%. P % 0.01; and 69%, P % 0.001; respectively). However, by 96 h only specific binding remained significantly less than control values (66%, P % 0.001), and at 10 days all binding values had returned to control levels. In order to determine if this phenomenon depended upon the presence of neurons, similar experiments involving chronic drug exposure were performed on SC cultures virtually devoid of neurons. As anticipated, CLO-displaceable binding was not present in either control or drug-treated cultures. In a corresponding manner, absolute control values for total and specific BDZ were proportionately reduced relative to wells with neurons: total BDZ binding was 175 + 9 fmol/well (mean + S.E.M.) and specific BDZ binding was 93 _+ 6 fmol/well (n = 20) compared to total BDZ binding of 319 ___ 15 fmol/well (n = 20), specific binding of 228 _+ 9 fmol/well (n = 20) and CLO-displaceable binding of 116 _+ 10 fmol/well (n = 18) in neuronrich cultures. In cultures without neurons, significant differences between drug-treated and control cultures
173 were observed only at 0 time after drug removal, when total BDZ binding was 67 __. 7% of control (n = 8; P ( 0 . 0 1 ) and specific BDZ binding was 37 _ 7% (n = 8; P ( 0 . 0 0 1 ) . Thus, the binding of diazepam to non-neuronal cells was less affected by chronic drug exposure than was the binding to neurons. In addition, the non-neuronal component of binding recovered more rapidly after the drug was removed. The effect of chronic diazepam exposure was also assayed electrophysiologically in SC neurons from cultures derived from the same embryos and treated identically as those used for binding studies. BDZ receptor function was evaluated by comparing the increase in GABAinduced C1- ion conductance produced in the presence and absence of acutely applied diazepam tt . Membrane conductance was determined from intracellular recordings using a single microelectrode containing 3 M KC1. Before recording, the cultures were washed 5 times with a solution containing (in mM): 107 NaC1, 5 KC1, 10 MgCI2, 2 CaC12, 12 HEPES, 5 D-glucose, with 1 /zM tetrodotoxin and 25 mM tetraethylammonium chloride included to block voltage-dependent Na ÷ and K ÷ conductances. Recordings were made in this medium at 23 ___ 2 °C. The inclusion of elevated [Mg2+]o and tetrodotoxin in the recording medium attenuated synaptic and electrical activity, allowing better quantitation of conductance responses. GABA and diazepam were applied by pressure (1 2 lbs./in. 2) from individual pipettes with 3-5 ~m tip diameter, placed 10-50 ~m from the neuronal soma. The pipette position, concentration of GABA (10-20/~M), and (most frequently) duration of GABA application (20-200 ms) were varied to produce an easily quantifiable conductance response. After a stable response was achieved, diazepam (12.6 ~M) was applied for 10 s. Several seconds after the diazepam application was terminated, GABA was again applied. Since the effect of diazepam outlasts the period of drug application23, this procedure allowed reliable quantitation of responses without problems of flow deflection that would occur if the two drugs were simultaneously applied. Conductance was measured by injecting hyper-
CI
I lu~ufi~lmU~m,m,w~,-,'=q
...........
V
b
~1 o 20
see
f ...............I............JJl3lll. II. . . . . .
I
_
V
p_~
,, ~
n
~
n
Fig. 2. Diazepam effects on GABA-mediated membrane conductance in spinal cord neurons from control cultures (a) and cultures chronically treated with diazepam (b). Hyperpolarizing constant-current pulses (I) were passed through the recording electrode to assay membrane conductance. P indicates pressure application of GABA (upward deflections) or diazepam (downward deflections). Membrane potential (V) was held at - 6 0 mV before GABA application by constant hyperpolarizing current. The hyperpolarization produced by diazepam application in (b) is not an effect of the drug but reflects the flow of medium resulting from drug application. This is commonly seen in both control and treated cultures, and also occurs with the application of medium without drug. In (a) the conductance increase produced by GABA is 7.8 nS and after maximal diazepam effect, 28.2 nS (a 260% increase). In (b) the conductance increase produced by GABA is 18.3 nS and is enhanced slightly (as evidenced by the increased voltage deflection) but the enhancement is so little that it is not different from control values, when evaluated by the constant-current pulses. The increased response to 20~M GABA in (b) relative to (a) is due to the cell-to-cell variation in responsiveness. The average responsiveness of drug-treated neurons to GABA did not differ from controls (see text).
polarizing constant-current pulses (50 ms, 0.21.0 nA) through the recording electrode using a bridge circuit. All cells were held at - 6 0 _+ 3 mV (in the absence of GABA or diazepam) using steady current injection as necessary. Typical intracellular recordings obtained with control cells and after chronic diazepam exposure are shown in Fig. 2. Most voltage responses to GABA were depolarizing due to the change in the CV ion gradient induced by the use of intracellular electrodes filled with KC1. The maximal effect of di-
174 azepam on G A B A responses required about the same time to develop in both control and drugtreated cultures. The maximal effect of the drug was used for all determinations. Since the responsiveness to G A B A and to diazepam varies from cell to cell, an effort was made to obtain stable, reliable records from as m a n y cells as possible at each time point (Fig. 3). Following chronic drug exposure (0 6 h after diazepam removal) the increase in GABA-mediated conductance produced by applied diazepam was reduced to 12% of that observed in control cultures (P ~< 0.0005, unpaired t-test). At this time point, diazepam produced a 170 _+ 21% increase in G A B A - m e d i a t e d m e m b r a n e conductance in neurons from control cultures, an effect that did not change significantly at other time points tested. 48 h after drug removal, the effectiveness of diazepam in potentiating GABA responses recovered partially (to 42% of
100"
°
80-
§ 60-
0
48h
72h
96h
TIME AFTER DRUG REMOVAL
Fig. 3. Recovery of the potentiating effects of diazepam on GABA responses after chronic exposure to the drug. Responses were calculated as the increase in membrane conductance produced by GABA in the presence of diazepam relative to that produced by GABA alone. Control values were obtained from cells chronically exposed only to the drug vehicle. Mean values of responses are shown as percent of control first at the end of the chronic exposure period (0-6 h after diazepam removal, shown as 0) and then during a 4day drug-free recovery period. Results are combined from two experiments; bars indicate S.E.M. Control values were similar at each time point (see text). The number of cells studied in control and drug-treated cultures, respectively, was: 0 h, 23 and 35; 48 h, 7 and 6; 72 h, 11 and 17; and 96 h, 7 and 11.
control), but due to the smaller n u m b e r of adequate recordings (6), the results were different from controls only at the 0.1 level of significance. At 72 and 96 h after drug removal, the potentiating effects of diazepam still averaged less than that for corresponding controls (93% and 65% of control, respectively), but the differences were not statistically significant due to the scatter of the data. The results indicate that chronic exposure of cultured SC cells to a clinically relevant concentration of diazepam a,~ leads to a marked reduction both in the ability of diazepam to bind to BDZ receptors on intact cells and to potentiate m e m b r a n e responses to GABA. There are several explanations that might account for this. It is possible that the BDZ receptors have been 'down-regulated' by reducing either their affinity or number. Alternatively, it is possible that the drug was retained by the cells after chronic exposure so that the reductions in binding and potentiation of G A B A responses was due either to continued occupation of the receptors by the drug or to release of the drug from the cells into the m e d i u m followed by re-occupation of receptors. In order to examine this possibility, culture dishes which had been exposed to 12.6 ttM diazepam for 7 days were thoroughly washed 3 times at room temperature with 1 ml portions of 50 m M phosphate buffer (pH 7.4) and dried. Tissue was then scraped off the dish, suspended in this buffer, and sonicated. The resulting homogenate was centrifuged at 150,000 ,g for 10 min to pellet m e m b r a n e s and insoluble material, a procedure similar to that reported by Paul et al. ~5. Supernatant derived from the equivalent of a single treated culture well was added to each untreated well in the binding assay. Supernatant obtained using the same protocol from vehicletreated cultures was also added to untreated wells. Untreated wells without added supernarant served as additional controls. In addition, the same supernatants were applied by pressure to naive neurons to test for any effects on G A B A - m e d i a t e d responses. Neither supernatant had any effect in either the biochemical or electrophysiological assays. These results indicate that the supernatant does not contain
175 detectable drug as evaluated in these two types of biological assay. Thus, it is unlikely that the treated cells were inadequately washed or retained enough drug to account for the observed reductions in binding and function, although the possibility that diazepam remained tightly bound to the membrane fraction has not been excluded. One way that diazepam could be retained in membranes would be if the chronic exposure decreased the off-rate of the drug from BDZ receptors. The reduction in binding and pharmacologic action of diazepam might result from continued occupation of receptors, with recovery representing the off-rate. Theoretically, continued receptor occupation might produce a continued potentiation of GABA-mediated chloride conductance, which would be detectable as an enhanced response to GABA (in the absence of applied diazepam) after chronic exposure to drug. However, the mean value of conductance responses to 20/~M GABA recorded at 0 time after removal of drug was not significantly different (12.8 + 1.8 nS, n - 32) from that observed in cells chronically exposed to vehicle only (11.2 _+ 1.8 nS, n = 23). Therefore chronic activation of the BDZ receptors as measured in the electrophysiological assay appears unlikely. Our observations are consistent with either a down-regulation of BDZ receptors or a prolonged occupation of receptors by diazepam where the occupation no longer produces a response. At present we have not distinguished between the two alternatives experimentally, but studies of the effects of chronic exposure of other receptors to agonists would lead us to prefer the former ~4. The results obtained with both the binding and electrophysiological assays showed a similar pattern, although the depression of diazepam action was somewhat greater than the reduction in total and specific diazepam binding at 0 time after drug removal. This difference may be related to the fact that the electrophysiological data were obtained exclusively from neurons, while specific binding includes both neuronal and non-neuronal sites. CLO-displaceable binding, which was virtually eliminated by chronic diazepam exposure, may be the component of
binding that corresponds to BDZ receptors coupled to GABA-activated CI-- channels in neurons. This conclusion is further supported by the observation that reduction of binding was substantially less in neuron-poor cultures. In earlier studies, researchers have reported an increase6, modest decrease ~7.~ or no change2-L3.22 in the BDZ receptor after chronic treatment. However, more recent work has shown that chronic administration of clonazepam to mice induces BDZ receptor 'subsensitivity' for several days following cessation of drug treatment5, and we have found that dissociated cerebral cortical cultures from mice show a more prolonged decrease in BDZ binding than spinal cord cultures after chronic diazepam treatment ~9. In addition, a prominent withdrawal response after administration of the BDZ receptor antagonist, R015-1788, has been produced in baboons treated with diazepam for 7 days ~° suggesting that the drug exposure time in this study is sufficient to produce drug tolerance and dependence. Our data also suggest that the potentiation of GABA-mediated conductance and CLO-displaceable binding represent similar receptor populations, even though the affinities of the specific (KD ~ 11 nM) and CLO-displaceable (Kt~ ~ 5 riM) sites are much greater than the apparent affinity of the neuronal receptors coupled to GABA-mediated Ck channels (about 2-5 /~M; R.E. Study and J. L. Barker, unpublished observations). This discrepancy has not yet been resolved, but it is possible that a population of binding sites of lower affinity ( K o about 200 nM) which we have recently found in SC cultures-'° is responsible for the potentiation of GABA responses. It is important to note that the concentration of diazepam that was used for chronic exposure of the cultures (12.6 I~M) is not only high enough to produce a maximal or near maximal potentiation of GABA-mediated conductance and saturate the receptors assayed by binding, but is also in the range associated with a significant clinical response (see ref. 1 for review). The precise mechanisms of the profound alterations in specific receptor binding and poten-
176
tiation of GABA-mediated inhibition remain to be elucidated. We do not yet know if the changes in binding represent a decrease in number of receptors or their affinity, or how the reduction of the effect of diazepam in the electrophysiological assay is achieved. Even though it is not yet clear how the parameters measured in the two assays are related, each has a pharmacological profile which correlates well with the clinical efficacy of the BDZs, and thus may represent a ba-
sis for the clinically important actions of these drugs. The observation that both binding and pharmacological action are greatly reduced after long-term exposure to diazepam may therefore represent cellular correlates of the clinical observation of tolerance seen in patients and in animals chronically treated with this class of drug. The authors would like to thank Ms. Linda Bowers for her photographic assistance.
I Bowling, A. C. and DeLorenzo, R. J., Micromolar affinity benzodiazepine receptors: identification and characterization in the central nervous system, Science. 216 (1982) 1247 1250. 2 Braestrup, C., Nielsen, M. and Squires, R. F., No changes in rat benzodiazepine receptors after withdrawal from continuous treatment with lorazepam and diazepam, LifeSci., 24(1979) 347 350. 3 Browne, T. R. and Penry, J. K., Benzodiazepines in the treatment of epilepsy. A review, Epilepsia, 14 (1973) 277 310. 4 Covi, L., Lipman, R. S., Pattison, J. H., Derogatis, L. R. and Uhlenhuth, E. H., Length of treatment with anxiolytic sedatives and response to their sudden withdrawal, Actap~vchiatr. scand., 49 (1973) 51 64. 5 Crawley, J. N., Marangos, P. J., Stivers, J. and Goodwin, F. K., Chronic clonazepam administration induces benzodiazepine receptor subsensitivity, Neuropharrnacologv. 21 (1982)85 89. 6 DeStefano, P., Case, K. R., Colello, G. D. and Bosman, H. B., Increased specific binding of [3H]diazepam in rat brain following chronic diazepam administration, Cell Biol. Int. Rep.. 3 (1979) 169 167. 7 Gallager, D. W., Maltor~a, P., Oertel, L. W., Henneberry, R. and Tallman, J., [ H]Diazepam binding in mammalian central nervous system: a pharmacologic consideration, J. Neurosci., 1 (1981)218 225. 8 Greenblatt, D. J. and Shader, R. J., Dependence, tolerance, and addiction to benzodiazepines. Clinical and pharmacokinetic considerations, Drug Metab. Rev., 8 (1978) 13 28. 9 Lossius, R., Dietrichson, P. and Lunde, P. K. M., Effect of diazepam and desmethyldiazepam in spasticity and rigidity, Acta neurolscand.. 61 (1980) 378 383. 10 Lukas, S. E. and Griffiths, R. R., Precipitated withdrawal by a benzodiazepine receptor antagonist, Science, 217 (1982) 1161 1163. I1 Macdonald, R. L. and Barker, J. L., Enhancement of GABA-mediated postsynaptic in cultured mammalian spinal cord neurons: a common mode of anticonvulsant action, Brain Research, 167 (1979) 323- 336. 12 McCarthy, K. D. and Harden, T. K., Identification of two benzodiazepine binding sites on cells cultured from
rat cerebral cortex, J. Pharmacol. exp, Ther., 216 (1981) 183 191. Mohler, J., Okada, T. and Enna, S. J., Benzodiazepine and neurotransmitter receptor binding in rat brain after chronic administration of diazepam or phenobarbital, Brain Research, 156 (1978) 391 - 395. Overstreet, D. H. and Yamamura, H. l., Receptor alterations and drug tolerance, Life Sci., 25 (1979) 1865 1877. Paul, S, M., Syapin, P. J., Paugh, B. A.. Moncada, V. and Skolnick, P., Correlation between benzodiazepine receptor occupation and anticonvulsant effects of diazepam, Nature (Lond,), 281 (1979) 688 -689. Ransom, B. R., Neale, E,, Henkart, M., Bullock, P. N. and Nelson, P. G., Mouse spinal cord in cell culture. I. Morphology and intrinsic neuronal electrophysiological properties, J. Neuropt~vsiol., 40 (1977) 1132 t 150. Rosenberg, H. C. and Chiu, T. H., Regional specificity ot" benzodiazepine receptor down-regulation during chronic treatment of rats with flurazepam, Neurosei. Let& 24 (1981)49 52. Rosenberg, H. C. and Chiu, T. H., Decreased [3H]diazepam binding is a specific response to chronic benzodiazepine treatment, Life Sci.. 24 (1979) 803 808. Sher, P, K., Reduced benzodiazepine receptor binding in cerebral cortical cultures chronically exposed to diazepare, Epilepsia. in press. Sher, P, K., Development and differentiation of the benzodiazepine receptor in spinal cord cultures, Develop, Brain Res., in press. Sher, P. K., Schrier, B. K. and Van Putten, D., An in situ assay for determination of benzodiazepine binding, Develop. Neurosci.. 5 (1982) 271 277. Shibla, D. B., Gardell. M. A. and Neale, J. H., The insensitivity of developing benzodiazepine receptors to chronic treatment with diazepam, GABA, and muscimol in brain cell cultures, Brain Research, 210 ( 1981 ) 471 474. Study, R. E. and Barker, J. L., Diazepam and (-)-pentobarbital: fluctuation analysis reveals different mechanisms for potentiation of gamma-aminobutyric acid responses in cultured central neurons, Proc. nat. Acad. Sci. U.S.A.. 78 (1981) 7180- 7184. Swaiman, K. F., Neale, E. A., Schrier, B. K. and Nelson, P. G., Toxic effect of phenytoin on developing cortical neurons in culture, Ann. Neurol., 13 (1983) 48-52.
13
14 15
16
17
18 19 20 21 22
23
24