An Addiction Science Network Resource

Reprinted from T. Yanagita (1987), Prediction of drug abuse liability from animal studies. In M.A. Bozarth (Ed.), Methods of assessing the reinforcing properties of abused drugs (pp. 189-198). New York: Springer-Verlag.
Brain Reward System
Back to book TOC
(Search book contents) (Search entire ASNet)


Chapter 10

Prediction of Drug Abuse Liability From Animal Studies

Tomoji Yanagita

Preclinical Research Laboratories
Central Institute for Experimental Animals
1433 Nogawa, Miyamae-ku, Kawasaki, 213 Japan

In this chapter, the pharmacological factors involved in predicting the abuse liability of drugs are discussed with major emphasis on their psychological dependence potential. For prediction of drug abuse liability, it is important to find and compare the drug’s effects against those of prototypic drugs. The important pharmacological properties include the following: central nervous system (CNS) effects, including operant behavioral effects; tolerance; physical dependence potential; and psychological dependence potential. It is also important to consider such chemical and pharmacokinetic properties as water solubility and speeds of absorption and elimination. In particular, psychological dependence potential is most relevant to abuse liability. To evaluate psychological dependence potential from animal studies, the reinforcing and discriminative stimulus properties of drugs must be assessed in comparison with those of prototypic drugs. When human data are available, it is important to verify the animal results on such points as pharmacokinetics and CNS drug susceptibility which may differ across species. 

Special consideration should also be given to the following in evaluating psychological dependence potential from animal studies: differences in reinforcing effects depending on the experimental procedure used, analysis of reinforcement data obtained in intragastric self-administration, daily dose and CNS manifestation in self-administration experiments, and influence of physical dependence. From a practical viewpoint, it is also important to predict the nature and extent of ill effects which arise as a consequence of abuse. For this, such drug properties as the CNS effects, psychological dependence potential, toxicity, tolerance, and physical dependence potential must be considered.


Abuse Liability and Pharmacological Properties of Drugs

The abuse liability of a drug refers to its likelihood of being abused by humans. The abuse liability of a drug is not solely determined by the drug’s pharmacological properties but also by many other factors such as its availability, popularity, price, ease and "fashionability" of use. A good example of this is found in the organic solvents. Until the 1960s their abuse liability was regarded to be rather low, but it is now known to have become high. The pharmacological properties of a drug are, nevertheless, the essential factors constituting its abuse liability—no past case of abuse has been reported for a drug that is free of such pharmacological properties.

The kind of pharmacological properties of a drug that have bearing in the prediction of its abuse liability are outlined below.

(1) Chemistry: chemical structure, physical state, active ingredients, water solubility, preparations, etc.

(2) General Pharmacology: general pharmacological effects, particularly central nervous system (CNS) effects, for profiling the class of the drug in comparison with those of the prototypic drugs of abuse

(3) Special Pharmacology: tolerance and physical dependence potential, reinforcing and discriminative stimulus properties, and effects on other operant behaviors

(4) Pharmacokinetics: absorption, distribution, metabolism, and excretion, including the half-life in the blood for comparison with the half-lives of the prototypic drugs of abuse

General Methods to Evaluate Psychological
Dependence Potential of Drugs from Animal Studies

As discussed in the previous section, abuse liability is constituted by many factors and cannot be predicted from the drug properties alone. Therefore, as a step towards making this prediction, what is usually discussed is the psychological dependence potential of a drug—in other words, the capacity of a drug to produce a psychic (nonphysical) state of dependence. Drug dependence in general is defined in a technical report of the World Health Organization (WHO, 1969) as follows:

A state, psychic and sometimes also physical, resulting from taking a drug, characterized by behavioral and other responses that always include a compulsion to take a drug on a continuous or periodic basis in order to experience its psychic effects, and sometimes to avoid the discomfort of its absence. (p. 6)

In this definition two points are particularly important: (a) The essential state in drug dependence is the psychic state (i.e., psychological or psychic and not physical dependence) and (b) the behavioral responses of compulsive drug-seeking and/or drug-taking form the essential constituent of psychological dependence, although this dependence may include such responses as a strong desire for a drug that is unexpressed in the behavior as well as the physical and mental states contingent upon experiencing certain of the central nervous system effects of a drug.

The first step in evaluation of the psychological dependence potential of a drug is to determine the drug’s class from its chemical and pharmacodynamic profiles. If a drug is found to bear similarity to one of the prototypic drug classes, the possibility is then suggested of the drug having psychological dependence potential similar to that of the prototypic drug. The reliability of this evaluation depends on the availability of reliable data. More reliable data usually permit better evaluation but, depending on the class, a limited evaluation may be possible from either the chemical structure or just a few critical data on general pharmacological effects. The most important animal data used in evaluating psychological dependence potential are those concerning the reinforcing and discriminative stimulus properties of a drug and other behavioral effects reviewed in other chapters of this book. The assessment of these data is the next step after evaluating the compound’s chemical and pharmacodynamic profiles. The question to be answered here is whether the drug is reinforcing or not, and if so, then to what extent. Again, comparisons with similar prototypic drugs should be made. Quite often there may be disagreement in the data on a drug or no clear answer may be obtainable, thus making prediction difficult. In addition to this, many complexities exist in the methods of testing and in the assessment of the data; some of these problems will be mentioned later. But as a rule, if the reinforcing intensity of a drug is found to be similar to the reinforcing intensity of a prototypic drug, there is a good possibility that the drug in question has similarly strong psychological dependence potential. This is particularly true when the data on the drug’s discriminative stimulus properties and other behavioral effects support the similarity. In assessment of the reinforcing properties, however, some caution must be taken; since the term reinforcement by definition refers to the increment of the response rate, it can be said that in the self-administration experiment the intensity or efficacy of the reinforcing effects of a drug will be indicated by the level of increment, but the determinants of the response rate here are not only the reinforcing properties of a drug and the experimental procedure but also the duration of the drug effects at the unit dose used in the experiment. Therefore, the reinforcing intensity or efficacy in a strict sense may not necessarily represent the intensity of the animals’ drug-seeking and drug-taking behavior.

When the reinforcing effect of a drug is found to be weak in a certain experiment, the evaluation has two possible directions: For example, although the reinforcing effects of cyclazocine and nicotine were both found to be weakly positive in an intravenous continuous self-administration experiment in rhesus monkeys (discussed in the following section), it is well known that cyclazocine has no meaningful psychological dependence potential in humans while nicotine definitely does.

In contrast, the lack of a reinforcing effect does not necessarily imply the lack of psychological dependence potential. An example of this can be seen in many of the hallucinogenic drugs such as LSD or mescaline. Although the psychological dependence potential of these drugs does not appear to be very high, it is still high enough for them to be occasionally abused by humans. Thus when a drug is found to belong to the hallucinogenic class, such data as the discriminative stimulus properties and other behavioral effects may be of great importance in making the prediction of abuse liability.

When psychological dependence potential is to be evaluated from animal data, the usual approach in the extrapolation should be applied. Namely, the animal data should be verified by the human data in such points as pharmacokinetics and drug susceptibility, particularly susceptibility to the central nervous system effects. For example, differences between the experimental animal species and humans in the plasma half-life or in the bioproduction of active metabolites may considerably change the reinforcing and other behavioral effects. Another possibility to be considered is the difference in receptor susceptibility to a particular drug. Therefore, where human data on bioavailability and drug response are available, verification of the animal data by the human data will be indispensable for reliable evaluation of the psychological dependence potential based on animal data.

Specific Problems in Evaluation of
Psychological Dependence Potential from Animal Studies

When the psychological dependence potential of a drug is evaluated from animal studies, it is necessary to be cautious about the analysis of data on animal responses that are characteristic of self-administration experiments. In this regard, the following four major points are discussed.

Influence of Differences in Experimental Procedure
on the Reinforcing Effect

In intravenous self-administration experiments in rhesus monkeys, the reinforcing effects of such drugs as cyclazocine and nicotine were not found to be demonstrable under the cross self-administration procedure but were demonstrable under the continuous self-administration procedure using a fixed ratio-1 schedule (FR-1). Tables 1 and 2 show the results of the cross and continuous self-administration experiments on cyclazocine and nicotine. In the cross-administration procedure, the period is limited to 4 hours daily for 3 days each with the reference drug first, saline next, and then a certain unit dose of the test drug; in the continuous self-administration procedure, the test drug alone is available almost continuously for several weeks. The reason for such a procedurally related difference in the demonstrability of the reinforcing effects is not clear, but it seems to occur when (a) the reinforcing effect of a test drug is relatively weak, (b) a test drug is relatively long acting, or (c) the class of the reference drug is markedly different from that of the test drug.

Table 1
Procedurally Related Differences in the Reinforcing Effect of
Intravenous (i.v.) Cyclazocine in Rhesus Monkeys
(Yanagita, unpublished data)

A. Cross self-administration experiment (i.v., FR-1)

Self-administration ratio (%)a)


Cyclazocine (mg/kg/inj.)

Unit dose 0.25 ml/kg 0.06 0.25 1 4 15

7.7 ± 3.8 5.8 ± 3.5 10.0 ± 5.8 7.7 ± 5.7 3.8 ± 2.0 2.3 ± 1.2
Note: a) Mean ± S.D. of percent ratio against lefetamine 0.1 mg/kg/inj. as 100%

B. Continuous self-administration experiment (i.v., FR-1)

Average daily number of self-administrations

Unit dose (mg/kg/inj.)
Period (weeks)
2 days
Monkey No. 669b)
No. 718b)
No. 839b)
No. 658c)
No. 858c)
a) Withdrawal observation, number of lever presses
b) Experienced monkey
c) Naive monkey

Table 2
Procedurally Related Differences in the Reinforcing Effect
of Nicotine in Rhesus Monkeys
(Adapted from Yanagita et al., 1974, 1983)

A. Cross self-administration experiment (i.v., FR-1)

Self-administration ratio (%)a)

Agent Saline

Nicotine (mg/kg/inj.)

Unit dose 0.25 ml/kg 2.5 10 40 160 640

14.3 ± 5.3 3.8 ± 2.0 7.9 ± 1.7 16.3 ± 8.8 18.5 ± 9.0 7.8 ± 2.1
Note: a) Mean ± S.D. of percent ratio against lefetamine 0.1 mg/kg/inj. as 100%

B. Continuous self-administration experiment (i.v., FR-1)

Approximate average daily
number of self-administrations

Unit dose (mg/kg/inj.)
20 50 200
Monkey Experienced (N=6) 
Naive (N=4)

Assessment of Reinforcing Properties by
Intragastric Self-Administration Experiments

Generally speaking, the intravenous route is more reinforcing than the oral route, probably because of the sharper rise and higher peak value of the blood level as well as the shorter duration of the CNS effects, which may make it easier for the animal to discriminate the drug effects and lead to more frequent responding on the lever. In animal experiments the intragastric route is mostly used when a drug is water-insoluble. Since the absorption rate of such drugs is usually lower than those in the intravenous route, relatively large unit doses are used, which results in further prolongation of the duration of the drug effects. For this reason the daily response rates tend to be low in intragastric self-administration experiments. This is particularly true with a class of drugs such as benzodiazepines. For example, the daily response rates for diazepam in rhesus monkeys under intragastric continuous self-administration were not high (see Table 3), but this should not be taken as indicating a lack of reinforcing properties with the drug, because the result obtained in the progressive-ratio experiment in the rhesus monkeys shows definite reinforcing effect. The progressive-ratio procedure presently used in our laboratory first requires monkeys to press a lever 100 times per intravenous injection (FR-100), and 24 hours later the ratio is gradually increased at each intake so that it is doubled every four intakes; when the number of lever presses in any 24-hour period drops to less than half of that achieved for the preceding period or when no intake is observable for 48 hours, the animals are said to have extinguished the self-administration behavior and the number of the lever presses for the last intake attained is regarded as the final ratio (Yanagita, 1976). This procedure is intended to assess the intensity of animals’ drug-seeking behavior as a measure of the psychological dependence potential of a drug, and the final ratios obtained with a test drug, reference drug, and saline or those obtained under physically dependent versus nondependent states are regarded to reflect the relative intensity of the drug-seeking behavior. The final ratios obtained in such experiments with diazepam were generally high as shown in Table 4, and the drug was found to have considerably high reinforcing properties depending on the procedure. This example suggests that even though the increment of the response rate in an intragastric experiment is slight, it may not be judged as negligible.

Importance of Observing the Self-Administered
Daily Dose and CNS Manifestation

For evaluation of the psychological dependence potential of a drug and prediction of the ill effects resulting from its abuse (to be discussed later), observations of the daily dose self-administered by the animals and of the behavioral manifestation resulting from self-administration of the drug are very important (Deneau, Yanagita, & Seevers, 1969; Yanagita, 1977). For example, if the self-administered dose level of one drug is higher than another in terms of their minimum effective and toxic doses, it indicates the possibility that a larger daily dose of the former drug is likely to be sought and taken by humans under the condition that the drug is relatively freely available, and thus its psychological dependence potential is evaluated to be that much higher. Similarly, where the drug effects that are sought by the animals are stronger with one drug than another, the former’s psychological dependence potential would be evaluated to be analogously higher. The continuous self-administration procedure allowing free access to a drug provides information such as the following in addition to a measure of the reinforcing effect: (a) the total ingestible dose during one day or any fixed period, (b) the property of the drug to produce overt signs of the CNS effects in animals at self-administered dose levels, (c) the long-term trends of intake of the drug, and (d) the drug-seeking behavior and withdrawal manifestation during the withdrawal period.

For observation of the overt signs of the CNS effects, the gross behavioral manifestations are usually scored using standardized observational protocols. Recently, intensive attempts have been made by several investigators to quantify behavioral toxicity in the course of testing for drug dependence potential (Brady & Griffiths, 1983).

Influence of Physical Dependence on
the Psychological Dependence Potential

Although the importance of the psychological state in drug dependence has been stressed in the previous sections, the capacity of a drug to produce physical dependence is also important if its development enhances the drug-seeking behavior. With several opioids such as morphine or heroin, it is well known that the drug-seeking behavior in humans becomes compulsive when physical dependence on these drugs has developed. In the definition of drug dependence, this phenomenon is included in the wording "in order to experience its psychic effects, and sometimes to avoid the discomfort of its absence" (WHO, 1969, p. 6). In predicting the psychological dependence potential of a drug, this factor has to be considered in case the drug has physical dependence potential. Enhancement of drug-seeking behavior under the physically dependent state is also observable in animals by the progressive-ratio procedure. An example of such an experiment on codeine and loperamide is shown in Table 5. Both drugs produced definite physical dependence in rhesus monkeys, but in the progressive-ratio experiment under physically dependent and nondependent states the enhancement of the drug-seeking behavior under the former was observed only with codeine (Yanagita, Miyasato, & Sato, 1980). Such enhancement is also observable with morphine but not with cocaine. Among CNS depressants, alcohol showed a slight enhancement (Table 6) and diazepam did not (Table 4).

Table 3
Continuous Intragastric Self-Administration of Diazepam 
in Rhesus Monkeys (Adapted from Yanagita & Kato, 1983)

Average daily number of self-administrations (per 1 or 2 wks)

0.5% CMC Diazepam (mg/kg/inj.)

CMC Diazepam CMC

1 wk
4 wk
2 wk
2 wk
4 wk

3 days
4 wk
2 wk

1 wk
No. 433 0.1 4.5
4.9 5.1 8.7
3.3 2.3
-- 3.1
No. 828 4.7  4.9
3.4 2.1 10.1
14.3 9.9
-- 12.7
No. 993 4.3 1.4
1.2 -- 1.8
1.7 1.0
0.1 0.4
No. 1062 3.3 5.9
3.1 -- 2.1
1.7 5.4
2.5 2.1

Table 4
Final Ratios in Progressive-Ratio Experiment on Diazepam in 
Rhesus Monkeys (Yanagita, unpublished data)

1 mg/kg/inj., i.v.
0.25 ml/kg/inj., i.v.
Monkey Saline-pretreateda) Diazepam-pretreateda)
No. 966 
No. 1025 
No. 1031 
No. 1057 
No. 1058

Note: a) Pretreated by programmed intravenous administration of diazepam 
(1 mg/kg/injection) or saline (0.25 ml/kg/injection) every 2 hours for 4 weeks

Table 5
Progressive-Ratio Experiment on Codeine and Loperamide
(Adapted from Yanagita et al., 1980)

0.25 ml/kg/inj.
1.0 mg/kg/inj., i.v.
0.06 mg/kg/inj., i.v.
0.25 ml/kg/inj.

No. 325 
No. 478 
No. 769

Note: a) Pretreated by programmed intravenous administration of the drug at the indicated unit doses every 20 minutes for 7 days

Table 6
Final Ratios in Progressive-Ratio Experiment on Alcohol 
in Rhesus Monkeys (Adapted from Yanagita, 1976)

Alcohol 0.8 g/kg/inj., i.v.
Monkey Saline-precededa) Alcohol-precededa)
No. 171 
No. 425 
No. 466 
No. 485

Note: a) Progressive-ratio tests were preceded by programmed administration of saline for 4 weeks or self-administration of alcohol (0.8 g/kg/injection) at FR-100 for 2 weeks. In this period they ingested 6 to 10 doses of alcohol daily

Prediction of Ill Effects

From a practical viewpoint, when the abuse liability of a drug is predicted, it is necessary to consider the nature and extent of any ill effects which may arise as a consequence of its abuse. The reinforcing and discriminative stimulus properties of a drug are also important in this respect. In the context of international control of drugs under the 1971 Convention on Psychotropic Substances (United Nations, 1977), the following criteria are used in the scheduling of substances:

If a substance
1. has the capacity to produce similar abuse and similar ill effects as a substance in Schedule I, II, III, or IV, or

2. a) has the capacity to produce a state of dependence and central nervous system stimulation or depression resulting in hallucinations or disturbances in motor functions or thinking or behavior or perception or mood, and
b) is being or is likely to be abused so as to constitute a public health and social problem warranting international control, the World Health Organization shall communicate to the [United Nations Narcotic] Commission an assessment of the substance, including the extent or likelihood of abuse, the degree of seriousness of the public health and social problem and the degree of usefulness of the substance in medical therapy, together with recommendations on control measures, if any, that would be appropriate in the light of its assessment. (p. 9)

As was true for abuse liability, the factors involved in producing such ill effects are not limited to the pharmacological and toxicological properties of a drug alone. Examples of ill effects attributable to other factors will include infection and embolism due to inappropriate handling of the preparation for injection as well as social problems such as smuggling and peddling. However, many ill effects are caused by the pharmacological and toxicological properties. These include (a) the CNS effects, particularly those on mental state and behavior, (b) the psychological dependence potential to produce compulsive drug-seeking behavior, (c) acute and chronic organ toxicity, (d) tolerance as a factor in causing acute toxic death, and (e) physical dependence potential resulting in mentally and physically hazardous withdrawal syndrome.

The general predictive methods here are the same as in the case of abuse liability, and comparative assessment on the above points should be made between the drug in question and some prototypic drugs which cause ill effects that are relatively well-known.

In this chapter the pharmacological factors involved in predicting the abuse liability of drugs were introduced and discussed with major emphasis on their psychological dependence potentials. The predictive methods described here deal for the most part with general methodology rather than actual method. Concerning the actual methods, some considerations and attempts are being made to systematize the assessment of the pharmacological factors by score/point systems and to predict the abuse liability and ill effects from these numerical values. However, at the present time, weighing the scores for each factor is in the trial and error stage because the relationships among the factors are not clear enough to validate the weighing. Progress in behavioral pharmacological research in both animals and humans on the problems of drug dependence will reveal those relationships and make it possible to properly weigh the factors and quantify them.


Brady, J. V., & Griffiths, R. R. (1983). Testing drugs for abuse liability and behavioral toxicity: Progress report from the laboratories at the Johns Hopkins University School of Medicine. In L. S. Harris (Ed.), Problems of drug dependence, 1982 (National Institute on Drug Abuse Research Monograph 43, pp. 99-124). Washington, DC: U.S. Government Printing Office.

Deneau, G., Yanagita, T., & Seevers, M. H. (1969). Self-administration of psychoactive substances by the monkey: A measure of psychological dependence. Psychopharmacologia, 16, 30-48.

United Nations (1977). Convention on psychotropic substances, 1971. New York: United Nations.

WHO Expert Committee on Drug Dependence. (1969). World Health Organization technical report series (No. 407).

Yanagita, T. (1976). Some methodological problems in assessing dependence-producing properties of drugs in animals. Pharmacological Reviews, 27, 503-509.

Yanagita, T. (1977). Brief review on the use of self-administration techniques for predicting drug dependence potential. In T. Thompson & K. R. Unna (Eds.), Predicting dependence liability of stimulant and depressant drugs (pp. 231-242). Baltimore: University Park Press.

Yanagita, T., Ando, K., Kato, S., & Takada, K. (1983). Psychopharmacological studies on nicotine and tobacco smoking in rhesus monkeys. Psychopharmacology Bulletin, 19(3), 409-412.

Yanagita, T., Ando, K., Oinuma, N., & Ishida, K. (1974). Intravenous self-administration of nicotine and an attempt to produce smoking behavior in monkeys. Committee on Problems of Drug Dependence, 567-578.

Yanagita, T., & Kato, S., (1983). Dependence studies on zopiclone. In L. S. Harris (Ed.), Problems of drug dependence, 1982 (National Institute on Drug Abuse Research Monograph 43, pp. 164-170). Washington, DC: U.S. Government Printing Office.

Yanagita, T., Miyasato, K., & Sato, J. (1980). Dependence potential of loperamide studied in rhesus monkeys. In L. S. Harris (Ed.), Problems of drug dependence, 1979 (National Institute on Drug Abuse Research Monograph 27, pp. 106-113). Washington, DC: U.S. Government Printing Office.

©1987 Springer-Verlag (printed version)
©2000-2009 Addiction Science Network (web-enhanced version)

ASNet Home

ASNet Profile

Illicit Drug Index

Virtual Lab Tour

Research Reports

Drug Classification

A Primer on Drug Addiction

Experimental Methods

Treatment Resources

Biological Basis of Addiction

Before Prohibition: Early Psychoactive Medicines

Links to Other Websites

Click here to enter the Addiction Science Network Discussion Forum

Brain Reward System©1999-2009 Addiction Science Network
This page was last revised 06 April 2009 13:10 EDT.
Send comments to:
Report technical problems to: