EFFECTS OF LIQUID MEMBRANES ON GASTROINTESTINAL ABSORPTION OF DRUGS IN RATS

Agins, Alan Paul. M.S., University of Rhode Island, 1980. Effects of Liquid Membranes on Gastrointestinal Absorption of Drugs in Rats. Major Professor: Dr. George C. Fuller. A liquid membrane system was investigated as a potential antidotal treatment for acute drug overdose. The effectiveness of orally administered liquid membranes in simulated acute secobarbital, phencyclidine and strychnine poisoning was assessed in fasting rats. Four milliliters of a liquid membrane suspension was administered immediately following drug intubations. The effect of liquid membrane on gastrointestinal absorption of secobarbital and phencyclidine was determined by comparison of duration of action (sleeping time) and blood drug concentration over time in drug controls and treated rats. The effect on phencyclidine absorption was also determined from the total amount of drug and metabolites excreted in the urine over twenty-four hours. The effectiveness of liquid membranes and activated charcoal for reducing strychnine absorption was assessed by acute toxicity experiments. Duration of action of phencyclidine was 22 percent longer in liquid membrane treated rats while no significant difference existed between groups in secobarbital sleeping times. iii Blood level versus time analysis revealed no differences in blood levels between control and treated groups in both secobarbital and phencyclidine experiments. No difference existed in the total recovery of phencyclidine in urine in control and treated groups. Liquid membranes failed to protect rats from strychnine induced toxicity, whereas activated charcoal completely inhibited the toxic effect. These studies show that liquid membranes are relatively ineffective in reducing drug absorption in vivo and that the system probably serves as a reservoir for sustained release of the drug as it passes through the gastrointestinal tract. The lack of in vivo efficacy may be attributed to an appreciable reduction in liquid membrane stability in the presence of gastrointestinal constituents such as bile salts and pancreatic secretions.


Blood Levels of Phencyclidine in Rats 29
Gas Chromatogram for the Analysis of Phencyclidine and Metabolites in Rat Urine . .

INTRODUCTION
Human poisoning involving drug ingestion is common in the United States. Recent statistics show the number of accidental deaths from drugs increased by over SO percent between 1968(National Clearinghouse, 1978. In addition, the number of non-fatal poisonings is estimated to exceed one million per year, or about one per 200 population (Goldstein et al., 1974). following administration of a chemical to a biological speciman, the concentration of the chemical within the tissues is dependent upon the ability of biological barriers to prevent its translocation and on the chemical properties of the compound which permit or prevent its translocation in the tissue.
In addition, since translocation is a time dependent process, it may be said that the intensity of all chemicalbiological reactions is time dependent (Loomis, 1978).
The present modes of emergency treatment in acute drug overdose are thus aimed at reducing the effective concentration of the drug at the site where the interaction occurs, thereby reducing the intensity and duration of the toxic effect, Within this framework, there exist both specific and non~specific procedures directed at either decreasing the rate or amount 0£ drug absorption from the gastrointestinal tract, increasing the rate of elimination of the 3 drug from the body, or increasing the threshold for toxicity by administration of a specific pharmacological antagonist.
In the latter case, however, specific antidotes exist for relatively few drugs and therefore the major thrust of antidotal treatment lies in non-specific methodologies and must be tailored to the individual and the situation (Levine, 1978).
In the case of orally ingested compounds, removal of the ·unabsorbed chemical from the gastrointestinal tract represents the most direct and widely employed procedure for preventing further absorption of the drug. Abdallah and Tye (1967) studied the effectiveness of emetic drugs and stomach lavage in dogs fed a barium meal. Their results indicated that emesis was superior to lavage and that apomorphine was the most effective emetic agent. This study also showed that recovery of barium meal was dependent on the interval of time between ingestion and initiation of the procedure. In addition, with respect to apomorphine induced emesis, it was found that the action of this agent was completely inhibited in animals treated with hypnotic doses of thiopental. The authors concluded that following ingestion of overdoses of sedative type drugs, the production of emesis by centrally acting emetic agents cannot be relied on.
Furthermore, apomorphi, ne induced emes is can be quite severe and potentiate CNS depressant effects of sedative drugs (Arena, 1974). Syrup of Ipecac is generally considered a safer agent for the induction of emesis since it has a s direct action on the gastric mucosa and can be eliminated with the vomitus. The effectiveness of the agent, however, tends to be quite variable and often requires fifteen to thirty minutes before onset of action (Lawson and Proudfoot, 1971).
Gastric aspiration and lavage is a procedure that is controversial among physicians. Those who advocate use of the procedure contend that although the amount of drug recovered by this means is extremely variable, in some cases substantial amounts can be recovered with benefit to the patient (Matthew et al., 1966). Others contend that the lavage procedure, if improperly performed, can lead to aspiration of lavage fluids, hasten gastric emptying into the intestine and in some cases, such as strychnine poisoning, the stimulation associated with intubation may precipitate convulsions (Meyers et al., 1976).
Once actual removal of an ingested drug has been attempted or accomplished the use of adsorbants for further decontamination may be indicated. Activated charcoal has been advocated as an effective therapeutic agent in acute ingestions for many years (Holt and Holtz, 1963). In vitro studies (Decker et al., 1968) have shown that this agent is capable of adsorbing a wide variety of toxic materials from aqueous solutions, although there is considerable variability in amounts depending on the material. Gosselin and Smith (1966) have emphasized that adsorption is not the same as chemical destruction and that adsorption may be reversible, leading to release of the offending chemical as the pH of the environment changes during passage through the gastrointestinal tract. Some in vivo studies, however, have shown activated charcoal to be effective. Fiser et al. (1971) reported that in dogs receiving barbiturates and glutethimide, administration of activated charcoal one-half hour after drug intubation resulted in significantly decreased serum levels of the compounds and a concommitant reduction in CNS depression. In addition, the charcoal-drug complex appeared to be stable, in that no significant dissociation was demonstrated over the twenty-four hour monitored experiments. Levy and Tsuchiya (1972) found that activated charcoal was effective in reducing aspirin absorption in man.
They reported, however, that the adsorption on activated charcoal was partially reversible and was probably due to the higher pH of intestinal fluids and the competitive effects of constituents of these fluids. Despite the increasingly frequent recommendations in the literature that activated charcoal be used as a gastrointestinal decontaminant, the substance is rarely employed due to the lack of demonstrated efficacy in life-threatening acute intoxications under controlled conditions (Hayden and Comstock, 1975).
A variety of other adsorbants including cholestyramine (Dordoni et al., 1973), Arizona montmorillonite and evaporated milk (Chin et ~·, 1969) and "universal antidot _ e" (Picchioni et al., 1966) have been evaluated as alternatives to activated charcoal. Although these substances may be useful in isolated cases? none appear to have the general applicability of activated charcoal. In the case of "universal antidote," it has been demonstrated that the combination of tannic acid and magnesi~~ oxide with activated charcoal interfers with the adsorbant activity of the charcoal and is hence less effective than activated charcoal alone (Picchioni et al., 1966).
Cathartics such as liquid petrolatum, sodium sulfate, and magnesium hydroxide have also been implicated for preventing absorption and hastening the transit of chemical through the intestine. However, information on the effectiveness of such procedures is lacking (Loomis, 1978).

Liquid Membranes
Liquid membranes were developed at Exxon Corporation (Li, 1968) as an industrial encapsulation process to solve a variety of separation problems. Since then, liquid membranes have been implicated for several industrial applications including separation of hydrocarbons and the removal of organic contaminants such as phenol (Cahn and Li, 1974) and toxic inorganic ion such as Cr . , Hg , and Cd . (Kitagawa et al., 1977) from wastewater.
Liquid membranes aTe thin, spherical liquid shells which encapsulate microscopic droplets of one phase and separate these from a bulk external phase. Liquid membranes are formed by first making an emulsion of two immiscible phases and then dispersing the emulsion in a third phase (continuous phase). The continuous phase and the encapsulated phase are generally miscible, but they are not miscible with the membrane phase. There are presently three types of liquid membrane systems (Appendix Table A) . One is a water-in-oil-inwater emulsion, one is an oil-in-water-in-oil emulsion and the third is a gas-in-fluorocarbon-in-water system. The liquid membrane phase usually contains surfactants, additives, and a base material which is a solvent for the other ingredients.
The surfactants and additives are used to control the stability, permeability, and selectivity of the membrane.
The encapsulated phase can be formed into a "sink" for trapping certain agents, or as a reservoir for releasing substances into the external phase.
In recent years, interest has been focused on the potential application of liquid membranes in the biomedical and biochemical fields. A n~~ber of areas of research have been implicated including the encapsulation of enzyme processes, the sustained, slow release of compounds, oxygenation of blood, and removal of toxic substances from the human body.
In the encapsulation of enzymes, liquid membranes may act to immobilize or protect enzymes from deactivating substances and non-optimal environments while maintaining free access to the desired substrate. In addition, preparations may allow for the encapsulation of necessary cofactors and optimal reaction conditions. May and Li (1974)  The removal of toxic substances from the human body represents an interesting potential for liquid membranes.
Research has been divided into two areas: the treatment of chronic uremia and the emergency treatment of acute drug overdose.
J-.; Asher and coworkers (1974Asher and coworkers ( , 1976Asher and coworkers ( , 1977Asher and coworkers ( , 1978 have been developing a liquid membrane system that would serve as adjunct treatment to dialysis in chronic uremia. The system involves the use of two liquid membrane formulations for the removal of one of the uremic toxins, urea. In principle, urea diffuses from the blood into the intestine and is hydro- In addition to ion-trapping, the internal aqueous phase can be formulated into a high capacity sink using plasma proteins, activated charcoal, or specific drug antibodies that bind tenaciously to drug.
In principle, ingested liquid membranes would pass through the gastrointestinal tract, trap and unabsorbed drug present in the lumen and be eliminated via the bowel.
The effectiveness of liquid membranes has been studied in vitro. Frankenfeld and coworkers (1976) Table B). Most recently, Chilamkurti (1979) and Yang (1979)   The ntnnber of deaths and the latency to death were recorded.
Non-lethals were monitored for twenty-four hours post intubation.

Determination of Blood Secobarbital
Quantitation of secobarbital in the 'blood is determined gas chromatographically. Extraction of drug from blood is based on the procedure of Dvorchik (1975). Two hundred microliters of whole blood are placed in a glass test tube.
Twenty microliters of a 300 µg/ml hexobarbital in acetone solution is added to each tube as an internal standard. Two mls of chloroform is then added and the tubes are vortexed  (Table I).
Phencyclidine, at a dose of 50 mg/kg, produced extreme ataxia and stereotypic behavior in all rats within five minutes after intubation. The latency to loss of righting response was not significantly different between drug control and liquid membrane treated groups. All rats in both groups displayed mild clon~s intermittently throughout the in the liquid membrane treated group (Table II).

Blood Level Experiments
Under the GLC conditions described earlier, the secobarbital peak appears at 1.3 minutes and the hexobarbital internal standard peak at 1.7 minutes from the time of injection. An example of a chromatogram is shown in Figure I.
The standard curve was linear throughout the range of concentrations tested.
The peak blood concentration of secobarbital for both the drug control and liquid membrane treated groups was observed at 30 minutes post intubation and levels declined over the next 2-1/2 hours ( Figure II). There were no significant differences in mean blood levels between the two groups at all times analyzed. Analysis of area under the blood concentration versus time curves showed that bioavailabili ty of drug was within 3% for the two groups (Table III).
An example of a chromatogram for phencyclidine analysis is shown in Figure III. Under the described GLC conditions, the Ketamine internal standard peak appears at 2.0 minutes and the phencyclidine peak at 2.5 minutes after injection.
The standard curve was linear throughout the range of concentrations tested and passed through the origin. Figure IV shows the blood concentration versus time curve for drug control and liquid membrane treated groups. Blood levels   aSeconal sodium in water at a dose of 100 mg/kg was administered by oral intubation; rats received drug plus 4 mls liquid membrane suspension Code 573-31(R); controls received drug plus 4 mls water.
bN represents the number of rats in each group.
c Values represent mean ± S.E.M. dCalculated from 0-3 hours by trapezoidal rule using mean blood concentration values.
appeared to peak at one hour after intubation and rapidly declined over the following two hours . The liquid membrane treated group showed peak blood levels of drug at 30 minutes followed by a less rapid decline in levels over the next 2-1/2 hours. No statistically significant differences in blood levels existed owing in part to the exterme variab i lity in the liquid membrane treated group. Area under the curve calculations revealed a 14% increase in bioavailability of phencyclidine in the liquid membrane treated group (Table   IV).

Recovery of Phencyclidine in Urine
Analysis of urine revealed the presence of three major metabolites in addition to unchanged phencyclidine ( Figure   V). Total recovery of phencyclidine over twenty-four hours was calculated by summation of the total amount of metabolites and unchanged drug multiplied by the volume of urine collected · for eich rat. The mean recoveries for the drug control and liquid membrane treated g~oups were not significantly different. There appeared to be a slight difference in the distribution of unchanged drug and metabolites within the two groups, with liquid membrane treated group showing a higher mean recovery of unchanged phencyclidine and a lower mean recovery of metabol i tes as compared to the drug control group. These differences were not statistically significant (Table V). aPhencyclidine hydrochloride in water at a dose of SO mg/kg was administered by oral intubation; rats received drug plus 4 mls liquid membrane suspension Code 573-118-2; controls received drug plus 4 mls water.
bN represents the nwnber of rats in each group.
cValues represent mean ± S.E.M.   bN represents the number of rats in each group.

DISCUSSION
Sleeping time is a frequently employed test for the evaluation of parameters which affect the pharmacological activity of the agent producing the anesthesia. Since the intensity and duration of a pharmacological effect is proportional to the concentration of drug at an effector site, any alteration~ in the absorption, metabolism or excretion of anesthesia producing drug will alter the duration of action. Many studies have been focused on alterations in sleeping time . as a function of biological differences (Quinn et al., 1958) or chemical interactions that affect barbiturate metabo~ism (Axelrod et al., 1954). These studies have shown that there is a direct relationship between duration of action (sleeping time) and the plasma and brain tissue levels of drug. Sleeping time experiments have also been extended to studies involving alteration in gastrointestinal absorption of anesthesia producing agents.
Recently, Picchioni and Consroe (1979)   The leakage effect has been demonstrated in vitro. Chiang et al. (1978) showed that in the presence of bile salts, liquid membrane uptake of pentobarbital ceased after ten minutes, followed by increasing drug concentration in the donor solution during the next two hours of mixing.
Similar results have been shown by Asher and coworkers (1977). In an in vitro system, liquid membranes leaked more than 35% of the encapsulated aqueous phase over a two hour period when contacted with a solution containing bile and pancreatin. These findings indicated that although liquid membranes are composed of non-digestible mineral oils, they are nonetheless subject to the emulsification action of bile salts. The effect of pancreatin on liquid membrane stability has been attributed to the solid content of the crude extract and not the enzymatic activity (Asher et al., 1976).
In addition to the problem of leakage in liquid membranes, another factor governing performance is the main- The addition of methylcellulose to the suspending solution appeared to reduce both the amount of leakage and coalescence when liquid membranes were contacted with bile and pancreatin.
This phenomenon wa? attributed to the methylcellulose coating the surface of the droplets and rendering it resistant to bile. Further study indicated, however, that the methylcellulose coating could be competitively replaced by protein such as albumin and that this led to loss of resistance.
Another factor which may affect the characteristics of liquid membranes in vivo has been demonstrated by Yang (1979).
It was shown that liquid membranes, encapsulating a concentrated pH buffer, are capable of absorbing or losing water when dispersed in solutions hypotonic or hypertonic relative to the internal aqueous phase. This finding implies that variations in osmolality along the gastrointestinal tract may alter the properties of liquid membranes. The effects caused by water absorption or loss have not been studied; however it is likely that this process would not only alter the nature · of the internal pH buffer, but may lead to structural changes compromising liquid membrane stability.
In view of the data, of inadequate liquid membrane stability in in vitro systems simulating physiological con- (2) The use of liquid membranes to trap and remove toxins from the gastrointestinal tract has theoretical merit.
A more biologically stable liquid membrane formulation, subject to less rupture and coalescence would be necessary for future in vivo testing.