An Addiction Science Network Resource

Reprinted from R.A. Glennon and R. Young (1987), The study of structure-activity relationships using drug discrimination methodology. In M.A. Bozarth (Ed.), Methods of assessing the reinforcing properties of abused drugs (pp. 373-390). New York: Springer-Verlag.
Brain Reward System
Back to book TOC
(Search book contents) (Search entire ASNet)


Chapter 18

The Study of Structure-Activity Relationships
Using Drug Discrimination Methodology

Richard A. Glennon and Richard Young

Department of Medical Chemistry
School of Pharmacy
Medical College of Virginia
Virginia Commonwealth University
Richmond, Virginia 23298

The drug discrimination paradigm, while gaining popularity as a research tool, has been used little for systematic investigations of structure-activity relationships (SAR: the effect of chemical structure on biological activity). This paradigm has been demonstrated to be a highly sensitive and very specific "drug detection" method that provides both quantitative and qualitative provides both quantitative and qualitative results and, as such, is particularly well suited for SAR studies. Using examples from our own work, we attempt to demonstrate our approaches to the study of the SAR of tryptamine, phenylisopropylamine (amphetamine), and benzodiazepine derivatives using drug discrimination methodology


General Concepts

Various investigators have employed the drug discrimination paradigm to examine the effects of gross structural modification of drugs on discriminative responding; by and large, however, systematic and detailed investigations of structure-activity relationships (SAR) have not ordinarily been a focal point of these studies. This is rather unfortunate since the drug discrimination paradigm appears to be particularly well suited for such studies. Many techniques generating SAR data are quantitative but not necessarily qualitative; in contrast, drug discrimination is one of the few methods that can afford both quantitative and qualitative results.

Of what value are SAR data? The simplest definition of SAR is the relationship between a given activity and the nature of the molecular entities that produce that activity (i.e., the effect of structural modification on activity). Rarely is such information gathered as an end unto itself; the results of SAR studies can be used, for example, (a) to define which ss can be used, for example, (a) to define which structural features are necessary for activity, (b) to optimize a specific effect (i.e., for the eventual design of new potentially active agents or antagonists), (c) to determine the stereoselectivity or stereospecificity required for an effect, (d) to correlate several types of activity, (e) to determine if members of two or more series of agents act via a similar (i.e., common) receptor interaction, and (f) to study mechanisms of drug action. Conversely, SAR studies can be employed to learn about the sites at which drugs act (e.g., receptor mapping studies). That is, by knowing what features of a molecule are important for activity, insight is gained as to what complementary functionalities might be present at a receptor site in order to effect a productive interaction. The common goal underlying many of these SAR studies is a greater understanding of the physicochemical properties necessary for a molecule to elicit a particular biological effect.

In a specific SAR study, the effects of a series of agents (i.e., structural variants; molecular modifications) on activity are evaluated. Initially, the study might involve members of a class of agents that possess only minor structural differences; ultimately, agents with more dramatic alterations, or even agents of an entirely different chemical class of compounds, might be investigated. SAR data obtained from in vivo studies reflect all om in vivo studies reflect all of the possible consequences of absorption, distribution, and metabolism (i.e., pharmacokinetics) of a given agent. As a result, caution is advised when comparing SAR derived from whole animal studies with SAR derived from in vitro studies; in the latter the effects of absorption, distribution, and/or metabolism can be minimal, nonexistent, or altered. The results derived from in vitro studies yield incisive and succinct SAR, but they are less likely than in vivo studies to reveal the effects of drug in a "real world" (e.g., clinical) setting. Obviously, there is a need for both types of studies. The SAR generated from drug discrimination represent in vivo SAR and, as such, reflect the involvement of structure and pharmacokinetics as part of the relationship between structure and discriminative stimulus effects.

The drug discrimination paradigm can be used to generate SAR data; however, such data are only valid with respect to a particular dose of a given training drug. Because the sensitivity and duration of effect of a stimulus is related to the training dose of the training drug and because drug stimuli are time dependent, dose-response relationships represent relative, not absolute, relationships between challenge drugs and training drugs. The interested reader is referred to Colpaert and to Overton (this volume and references therein) for a more detailed discussi therein) for a more detailed discussion on the effects of these parameters.

Thus, the SAR for a series of agents is related to the training dose of the training drug and to the conditions present when the data were generated. As a consequence, the results of drug discrimination studies should only be used to formulate SAR for those members of a challenge series where there is evidence for a common effect (i.e., for those agents where stimulus generalization has occurred to a stated training dose of a particular training drug). That is, it is inappropriate to develop an SAR within a series of agents that produced only partial generalization. If a challenge drug produces anything less than stimulus generalization, it is probably safe to assume that the challenge drug is "inactive" with respect to possessing stimulus properties common to the training dose of the training drug under the temporal constraints (e.g., presession injection interval) of the assay. When stimulus generalization has occurred between a training drug and a challenge drug, ED50 values can be calculated for comparative purposes. An ED50 dose is, normally, the dose of an agent that produces a specific effect in 50% of the animals. Although this is the manner in which ED50 values have been used/calculated in some discrimination studies, this is not always the case. In tests of stimulus generalization, it is not uncommon to requialization, it is not uncommon to require, as an endpoint, that all of the animals respond on the drug-appropriate lever. In such studies, then, the ED50 dose is defined as that dose (of an agent to which stimulus generalization occurred) at which all of the animals would make 50% of their responses on the drug appropriate lever. The ED50 values used in this chapter adhere to the latter definition. Furthermore, when making potency comparisons between two agents that produce similar stimulus effects, it is most appropriate to make these comparisons on a micromole/kg basis, rather than mg/kg basis, when there is a substantial difference in molecular weights of the agents involved.

Ideally, it is desirable to evaluate doses of a challenge drug until either stimulus generalization or disruption of behavior (i.e., no responding) occurs. If, for example, the highest test dose of a challenge drug elicits 50% training-drug-appropriate responding and, for one reason or another, the evaluation of higher doses is precluded, it is incorrect to conclude that the challenge drug is half as active as the training drug (or, for that matter, that it is active at all) when, in fact, there has been no demonstration that the two agents can produce a common effect. In this situation, comparisons can only be made in a qualitative sense. That is, the only valid conclusion that can be reached is that this agent is less effective thans that this agent is less effective than the training drug in producing training-drug-like effects (or, correspondingly, that it is less effective than some other challenge drug, which at a dose below the highest test dose produced training-drug-like effects). Likewise, if two challenge drugs produce partial generalization (e.g., 40% and 60% training-drug-appropriate responding) at a given dose, it cannot be stated with certainty that the second challenge drug is more active than the first because the possibility exists that one (or both) agent(s) may not be capable of producing an effect common to that of the training drug.

Another very important consideration is the necessity of a thorough dose-response investigation. It has been our experience in tests of stimulus generalization that certain agents produce saline-like effects at particular doses and disruption of behavior at some higher dose(s). While an initial conclusion to the results of such a study might be that there is a lack of stimulus generalization, we have found in a number of instances that a careful evaluation of additional doses (i.e., doses between the highest dose that resulted in saline-like responding and the lowest dose that produced disruption of behavior) ultimately resulted in stimulus generalization. This has even been observed with agents where the difference in saline-like and disruptive doses has been quite small. Several instances have now been encounteveral instances have now been encountered where administration of a challenge drug in a logarithmic progression of doses (e.g., 0.1, 0.3, 1.0, 3.0 mg/kg) resulted in saline-like responding at the lower doses and in disruption of behavior at the highest dose, and examination of the doses between 1.0 and 3.0 mg/kg resulted, ultimately, in stimulus generalization. We believe that these results emphasize the sensitive and specific nature of the drug-induced stimulus.

Finally, it should be recognized that when challenge drugs are being examined the data obtained relate to training-dose-like effects. For example, an investigation of the effects of a series of barbiturates in diazepam-trained animals does not provide SAR of barbiturate activity; rather, the data reflect the SAR of a series of barbiturates to produce diazepam-like results. Thus, this SAR may or may not be the same as the SAR developed for the effects of this same series of barbiturates in, for example, pentobarbital-trained animals.

What follows are fragments of several examples of SAR studies that have been conducted in our laboratory; these involve the SAR of tryptamine derivatives, phenylisopropylamine derivatives (as reflecting DOM-like and amphetamine-like properties), and benzodiazepines. There will be very little discussion as to the interpretation of SAR data (except where a particular point is being emphasized); rather, an attempt is made to describe our approachtempt is made to describe our approach and to delineate the types of data that might be obtained. Additionally, some of these studies are still in progress and it would be premature to discuss detailed SAR. Nevertheless, each of these studies reveals the value and the limitations of the drug discrimination paradigm in formulating SAR.

SAR Studies On Tryptamine Derivatives

In the course of our studies on the mechanism of action of hallucinogenic agents, it became necessary to determine the in vivo SAR of series of tryptamine derivatives; the drug discrimination paradigm was employed to obtain this information. Using a standard two-lever operant procedure, rats were trained to discriminate 1.0 mg/kg of racemic 1-(2,5-dimethoxy-4-methylphenyl)-2-aminopropane hydrochloride (DOM) from saline under a variable-interval 15-second (VI-15) schedule of reinforcement for food (sweetened condensed milk) reward (Young, Glennon, & Rosecrans, 1981). Generalization studies were then conducted using various doses of a variety of tryptamine derivatives as the challenge series. In this way it was determined which members of the series produced DOM-like effects (i.e., effects similar to that produced by 1.0 mg/kg of racemic DOM when administered 15 minutes prior to a 2.5-minute extinction session). Response rates were monitored and compared with those produced by administration of the trahose produced by administration of the training dose of the training drug (i.e., DOM) in order to determine whether or not there was any disruption of behavior. The response rates following DOM were not significantly different from those seen after administration of 1.0 ml/kg of saline.

The study focused on the effects of structural modification in each of three different areas of the tryptamine molecule: the terminal amine group, the aromatic nucleus, and the alkyl side chain (see Figure 1). The first region to be examined was the terminal amine. Administration of tryptamine, up to doses of 25 mg/kg, resulted in saline-like responding, while administration of the N,N-dimethyl derivative of tryptamine (i.e., N,N-dimethyltryptamine: DMT) resulted in DOM-stimulus generalization (ED50 = 5.8 mg/kg). Because tryptamine did not result in stimulus generalization at the doses evaluated but because it might have at higher doses, it cannot be concluded that tryptamine is inactive. Nevertheless, DMT is more effective than tryptamine in producing DOM-like effects. This finding is consistent with the results of other pharmacological studies that have concluded that certain primary indolealkylamines, such as tryptamine, penetrate the blood-brain barrier only with difficulty and/or are rapidly metabolized in vivo by oxidative deamination. Dimethylation of tryptamine to produce DMT is known to enhance its ability to penetrate the bloode its ability to penetrate the blood-brain barrier and to offer some protection from oxidative deamination; whether the N-methyl groups also contribute to a specific drug-receptor interaction is as yet unknown. (See Glennon, Young, & Jacyno, 1983c, for further discussion of this topic.) Homologation or extension of the terminal amine alkyl groups of DMT to ethyl, propyl, and isopropyl groups (i.e., DET, DPT, and DIPT, respectively; see Table 1) yields derivatives that result in DOM-stimulus generalization when administered to DOM-trained animals. On this basis the relative order of potency is DPT > DET > DIPT > DMT. Thus, the homologs of DMT are not only active but apparently are more potent than DMT in producing DOM-like effects (Glennon, Young, Jacyno, Slusher, & Rosecrans, 1983d).

Figure 1: The structure of tryptamine showing the aromatic nucleus (A), the alkyl side chain (B), and the terminal amine (C).
Structure of tryptamine

Table 1
Structures of Tryptamine Derivatives

Structures of tryptamine derivatives

Agent R’ R" Ra n
Tryptamine H H H 1
DET C2H5 H H 1
DPT nC3H7 H H 1
DIPT iC3H7 H H 1
4-OMe DMT CH3 H 4-OCH3 1
5-OMe DMT CH3 H 1
4-OMe DMT CH3 H 4-OCH3 1
5-OMe DMT CH3 H 5-OCH3 1
6-OMe DMT CH3 H 6-OCH3 1
5-OMe DET C2H5 H 5-OCH3 1
5-OMe DIPT iC3H7 H 5-OCH3 1
5-OMe Gramine CH3 H 5-OCH3 0
a-MeT H CH3 H 1
5-OMe a-MeT H CH3 5-OCH3 1
a-EtT H C2H5 H 1

Note: a, Location of substituents is indicated by the position number, except for unsubstituted rings, where R = H.

Next to be explored was the effect of aromatic subsfor unsubstituted rings, where R = H.  

Next to be explored was the effect of aromatic substitution (see Figure 1). While a number of different substituents were examined, only the results of methoxy substitution will be discussed here. The 4-methoxylation of DMT (i.e., introduction of a methoxy group at the 4-position of DMT [4-OMe DMT]; see Table 1) somewhat enhanced activity while 5-OMe DMT (ED50 = 1.2 mg/kg) was found to be approximately 5 times more potent than DMT in producing DOM-like effects. On the other hand, administration of doses of 6-OMe DMT up to 10 mg/kg resulted in saline-appropriate responding; additional doses were not examined because only limited supplies of this compound were available. Nevertheless, with respect to effectiveness 5-OMe DMT > 4-OMe DMT > DMT > 6-OMe DMT (Glennon et al., 1983d); 5-methoxylation was also found to enhance the potency of DET and DIPT, and 5-OMe DET was more potent than 5-OMe DIPT.

The final tryptamine region to be explored was the alkyl side chain (see Figure 1). All of the tryptamine derivatives discussed to this point possess an unbranched alkyl side chain of two methylene units that separates the aromatic nucleus from the terminal amine; shortening this side chain to one methylene unit results in a series of agents called gramines. For example, 5-OMe DMT might be considered a homolog of its one methylene unit counterpart 5-OMe gramine (see Table 1). Administration of doses of 1.0 to 6.0 mg/kg of 5-OMe gramine to the DOMto 6.0 mg/kg of 5-OMe gramine to the DOM-trained animals resulted in a saline-like responding; solubility problems preclude administration of higher doses. However, the highest dose evaluated of 5-OMe gramine was 5 times the ED50 dose of 5-OMe DMT; thus, it was concluded that 5-OMe gramine was less effective than the latter agent in producing DOM-like effects (Glennon et al., 1983d).

The effect on activity of a-alkylation of the side chain was also investigated. Unlike tryptamine, both the a-methyl and a-ethyl derivatives of tryptamine (i.e., a-methyltryptamine [a-MeT] and a-ethyltryptamine [a-EtT], respectively; see Table 1) produced DOM-like effects with a-MeT (ED50 = 3.13 mg/kg) being more potent than a-EtT. Both of these a-alkyltryptamines are primary amines, and yet both produce DOM-like effects at lower doses than those where tryptamine still produces saline-like responding. Further, a-MeT is approximately twice as potent as DMT. These results suggest that the N-methyl groups of DMT do not participate in a specific drug-receptor interaction (and, in fact, that they may actually detract from such an interaction) and detract from such an interaction) and that the a-alkyl groups may act either to enhance blood-brain barrier permeability and/or to protect against oxidative deamination of the terminal amine.

This examination can be carried one step further. a-MeT possesses an asymmetric center and, thus, exists as optical isomers (see Figure 2). S(+)-a-MeT was found to be about twice as potent as racemic a-MeT, while R(-)-a-Met produced partial generalization (i.e., 61% DOM-appropriate responding) at 3.25 mg/kg and disruption of behavior at higher doses (Glennon et al., 1983c). Thus, if the role of the a-methyl group was to simply enhance blood-brain barrier permeability, a potency difference between the isomers might not be expected; the observed enantiomeric difference may be explained in one of several ways. It might be speculated that the methyl group of S(+)-a-MeT contributes to an enhanced drug-receptor interaction; alternatively, there may be differences in the metabolism of the two isomers. It is difficult to interpret the results, however, because only one isomer produced DOM-stimulus generalization. Another a-alkyltryptamine examined was 5-OMe a-MeT; as might be expected, 5-methoxylation of a-MeT enhanced its potency (ED50 = 0.52 mg/kg).

Structure of R(-)- and S(+)-alpha-methyltryptamine
Figure 2: The structure of R(-)-a-methyl-tryptamine (A) and S(+)-a-methyltryptamine (B).

However, a strict SAR comparison cannot be made here because, although a 15-minute presession injection interval was employed for most of the studies, a 90-minute presession injection interval was determined to be optimal for 5-OMe a-MeT. Nevertheless, the effect of 5-methoxylation of a-MeT paralleled the effect observed with 5-methoxylation of DMT (Glennon, Jacyno, & Young, 1983a). Furthermore, the (+) isomer was found to be approximately twice as potent as racemic 5-OMe a-MeT; however, in the case of this compound, (-)-5-OMe a-MeT also produced DOM-like effects although it was several-fold less potent than its enantiomer. Although differences in the metabolism of the two isomers cannot be eliminated at this point, it does appear that two isomers cannot be eliminated at this point, it does appear that the methyl group of (+)-5-OMe a-MeT contributes in a positive manner to a possible drug-receptor interaction.

Recently, a series of isotryptamine derivatives was synthesized; these are analogs of tryptamine where the alkyl side chain and pendant terminal amine are now attached to the indole nitrogen atom. Derivatives of the isotryptamines, such as 5-methoxy-N,N-dimethylisotryptamine (5-OMe isoDMT) and 6-methoxy-N,N-dimethylisotryptamine (6-OMe isoDMT; see Figure 3), may be viewed either as tryptamine derivatives where the side chain has been relocated or as derivatives where the indole nitrogen atom has been moved. In other words, it is unclear whether 5-OMe isoDMT would mimic 5-OMe DMT or 6-OMe DMT in biological situations.

One way to address this problem would be to compare the SAR of the two series of agents; this was the approach that was taken and the drug discrimination procedure was used, in part, to obtain some of this information (Glennon, Jacyno, Young, McKenney, & Nelson, 1984a). For example, administration of 5-OMe DMT to DOM-trained animals results in stimulus generalization, while 6-OMe DMT elicits saline-appropriate responding at doses up to 10 mg/kg (vide supra). If the isoDMT derivatives mimic DMT derivatives, one of the two methoxy derivatives might be anticipated to be more potent than the other; 5-OMe isoDMT produced saline-like responOMe isoDMT produced saline-like responding at doses to 16 mg/kg, while the DOM-stimulus generalized to 6-OMe isoDMT (ED50 = 7.1 mg/kg. Furthermore, the time course of effects of 5-OMe DMT and 6-OMe isoDMT was found to be identical. Thus, because 6-OMe isoDMT more closely mimics the effects of 5-OMe DMT than does its 5-OMe counterpart, it appears that the isotryptamines may be viewed as tryptamine analogs with the indole nitrogen replacing the 3-position carbon atom at the aromatic nucleus. Based on the results of this and other studies, isoDMT derivatives are less potent bioisosteres of DMT derivatives.

Structural relationships between 5-OMe DMT and 6-OMe DMT
Figure 3: Structural relationships between (a) 5-OMe DMT, (b) 6-OMe DMT, and, each drawn in two different orientations, (c,d) 5-OMe isoDMT and (e,f) 6-OMe isoDMT; Z = CH2CH2N(CH3)2.

SAR Studies on Phenylisopropylamine Derivatives

Phenylisopropylamine or 1-phenyl-2-aminopropane is known as amphetamine; this agent is a CNS stimulant in humans and in animals. Su amphetamine; this agent is a CNS stimulant in humans and in animals. Substitution on the aromatic nucleus of amphetamine affords a variety of derivatives that may possess similar or different pharmacological properties. One such derivative is 1-(2,5-dimethoxy-4-methylphenyl)-2-aminopropane or DOM; this agent is known to be hallucinogenic in man. Thus, if given a series of 1-phenyl-2-aminopropane derivatives, it should be possible to develop two independent SARs—an SAR for amphetamine-like activity and an SAR for DOM-like activity. Using the drug discrimination paradigm, it might also be possible to determine if there are any 1-phenyl-2-aminopropane derivatives possessing both amphetamine-like and DOM-like properties.


Silverman and Ho (1978) were the first to demonstrate that racemic DOM would serve as a discriminative stimulus in rats. Since then, a large number of rats have been trained in our laboratory to discriminate DOM from saline. The training and testing procedures were identical to those discussed for the study of the tryptamine derivatives (for greater detail see Young et al., 1981). Using these DOM-trained animals, we have formulated SAR based on the results of well over one hundred stimulus generalization studies. Silverman and Ho (1978) demonstrated that the DOM-stimulus does not generalize to amphetamine and that an amphetamine-stimulus does not generalize to DOM; results obtained in our laboratory confirm ults obtained in our laboratory confirm these findings that DOM and amphetamine produce dissimilar discriminative cues.

The general phenylisopropylamine skeletal structure common to DOM and amphetamine can be subdivided into three areas for examination: the terminal amine, the alkyl side-chain, and the aromatic nucleus (see Figure 4). Both monomethylation of DOM (i.e., N-Me DOM) and N,N-demethylation of the a-demethyl derivative of DOM (i.e., DD-DOM) reduce potency. In general, the primary amine derivatives (i.e., derivatives that are unsubstituted on the terminal amine) are more potent than their mono or dimethyl counterparts. Substitution on the aromatic nucleus is also an important factor for DOM-like activity; for example, neither amphetamine nor any of its monomethoxy derivatives (i.e., 2-OMe PIA, 3-OMe PIA, PMA) produce DOM-appropriate responding (see Table 2 for structures).

Within the dimethoxy (DMA) series the DOM-stimulus generalizes only to 2,4-DMA and 2,5-DMA but not to 2,3-DMA, 2,6-DMA, 3,4-TMA or 3,5-DMA. Of the trimethoxyphenylisopropylamines (TMA derivatives), complete DOM-stimulus generalization occurred to all five compounds although 2,4,5-TMA and 2,4,6-TMA were the only derivatives found to be more potent than either 2,4-DMA or 2,5-DMA (Glennon & Young, 1982b; Glennon, Young, Benington, & Morin, 1982b; Glennon et al., 1983c). The 2,4- and 2,5-dimethoxy substitution patterns app,5-dimethoxy substitution patterns appear to be important features for DOM-stimulus generalization; as such, additional derivatives of 2,4-DMA and 2,5-DMA were selected for more extensive evaluation.

Without belaboring the detailed SAR of these agents, it might simply be concluded that many derivatives of 2,4-DMA and 2,5-DMA produce DOM-like responding and also constitute some of the more potent agents investigated. For example, substitution of the 4-position of 2,5-DMA by small alkyl groups (e.g., DOET, DOPR) or by halogen (e.g., DOB, DOI) all result in agents that are at least as potent, if not more potent, than DOM itself (Glennon et al., 1982d; Glennon, Young, & Rosecrans, 1982e). However, the 2,5-dimethoxy substitution pattern appears quite sensitive to the presence of substituents at the 3-position. Moving the 4-position methyl or bromo group of DOM or DOB, respectively, to the 3-position (i.e., 3-Me, 2,5-DMA and 3-Br 2,5-DMA) results in agents that produce saline-like responding even at doses of 10 to 20 times the ED50 doses of their parent compounds (Glennon et al., 1982d; Glennon, Young, & Rosecrans, 1982f).

Finally, there is the effect of alteration of alkyl side chain (see Figure 4). Removal of the a-methyl groups of those agents that produce DOM-like effects seems to have a common effect—to reduce potency. For example, mescaline is approximately half as potent as its a-methyl analog, 3,4,5-TMA (Glennon & Young, 1982b). Again, it might be speculated that the lack of this methyl function either decreases the ability of a compound to penetrate the blood-brain barrier and/or makes it a better substrate for oxidative deamination. The presence of an a-methyl group, as in the tryptamine series, appears to be optimal; homologation of the a-methyl to an a-ethyl group reduces potency (Glennon et al., 1983c). The phenylisopropylamines, by virtue of the presence of the a-methyl group, are also optically active. The optical isomers of several of these agents that are capable of producing DOM-like effects have been examined (e.g., DOM, DOB) and in every case the R(-)-isomers have been found to be more potent than either their S(+)-enantiomers and/or racemates (Glennon et al., 1982d).

Figure 4: The structure of 1-phenyl-2-aminopropane (phenylisopropylamine) showing the aromatic nucleus (A), the alkyl side chain (B), and the terminal amine (C).
Structure of 1-phenyl-2-aminopropane

Table 2
Structures of 1-phenyl-2-aminopropane (Phenylisopropylamine) Derivatives

Structures of 1-phenyl-2-aminopropane

Agent  R’ R2 R3 R4 R5 R6
Amphetamine H H H H H H
2,5-DMA H OCH3 H H H H
2,3,4-TMA H OCH3 OCH3 OCH3 H H
2,3,5-TMA OCH3 OCH3 H H
2,3,5-TMA H OCH3 OCH3 H OCH3 H
2,4,5-TMA H OCH3 H OCH3 OCH3 H
2,4,6-TMA H OCH3 H OCH3 H OCH3
3,4,5-TMA H H OCH3 OCH3 OCH3 H
3-Me 2,5-DMA H OCH3 Me H OCH3 H
3-Br 2,5-DMA H OCH3 Br H OCH3 H
3,4-MDA H H    -OCH2O- H H

 Thus, in this way it has been possible to determine which phenylisopropylamines produce DOM-like effects and then to develop an SAR based on these results. The interested reader is referred to a review by Glennon, Rosecrans, and Young (1983b) for a more detailed discussion of these SAR.

One of the uses of SAR data is to compare two types of activity. If the SAR for a given series of agents is similar for different measures of activity, this suggests that the two types of activity may be related or may be mediated via a similar mechanism. A comparison of the SAR developed for DOM-like effects reveals similarities with the SAR otwo types of activity may be related or may be mediated via a similar mechanism. A comparison of the SAR developed for DOM-like effects reveals similarities with the SAR of these same phenylisopropylamines for human hallucinogenic activity. In fact, there is a significant correlation (r = 0.96; p < 0.001; N = 22) between the ED50 values of those phenyl-isopropylamines with DOM-stimulus generalization and their human hallucinogenic potencies (Glennon, Rosecrans, & Young, 1982a; Glennon et al., 1982d).

Amphetamine-Like SAR

Although our work on the amphetamine-like SAR of these agents is still in progress, sufficient data have been collected to demonstrate that this SAR is quite different from the DOM-like SAR. Various investigators have studied the discriminative stimulus properties of amphetamine and several reviews are available (e.g., Young & Glennon, 1986; Silverman & Ho, 1977).

We have trained rats to discriminate 1.0 mg/kg of (+)-amphetamine sulfate from saline using a VI-15 schedule of reinforcement for food reward. Some of our initial findings are that monomethylation of the terminal amine of amphetamine has little or no effect on amphetamine-appropriate responding and that the S-isomer of amphetamine is more potent than its racemic mixture or corresponding R-enantiomer. These findings are in direct opposition to the DOM-SAR. In general, those agents that elicit DOM-appropriate responding do not elicit amphetamine-appropriate responding and vice versa. For example, agents such as 2-OMe PIA and 3 example, agents such as 2-OMe PIA and 3-OMe PIA, that do not result in DOM-stimulus generalization, do produce amphetamine-like effects. Again, this work is currently in progress and detailed SAR must await the results of additional studies. However, there appears to be at least one agent, 3,4-methylenedioxyamphetamine (3,4-MDA) that is capable of producing both DOM-like and amphetamine-like effects. We now take the opportunity to discuss this compound in somewhat greater detail.


A discriminative stimulus is generally regarded as being quite specific; DOM-stimulus generalization does not occur to amphetamine, for example, nor does amphetamine-stimulus generalization occur to DOM. However, both the amphetamine and DOM stimulus generalize to 3,4-MDA; thus, this compound appears capable of producing both amphetamine-like and DOM-like effects (Glennon, Young, Anderson, & Rosecrans, 1982c; Glennon & Young, 1984c).

Animals were trained to discriminate 1.5 mg/kg of racemic 3,4-MDA hydrochloride from saline employing a VI-15 schedule of reinforcement (Glennon & Young, 1984c). Interestingly, the MDA-stimulus generalized to both amphetamine and DOM. Initial thinking was that 3,4-MDA might represent an agent with a new spectrum of effects and that an SAR might be developed for these MDA-like effects. However, realizing that 3,4-MDA is optically active and drawing upon the above SAR ( active and drawing upon the above SAR (where those agents producing DOM-like effects have more potent R-isomers than S-enantiomers while the opposite is true for those agents that produce amphetamine-like effects), it was entirely possible that the individual isomers of 3,4-MDA might possess dissimilar properties. In other words, 3,4-MDA might represent a transition agent in a continuum of amphetamine-like to DOM-like agents with the individual isomers being capable of displaying one type of activity or the other.

Using the above mentioned DOM-trained and amphetamine-trained animals, generalization studies were conducted on the individual optical isomers of 3,4-MDA. Substantiation for the presented hypothesis was obtained when it was found that the DOM-stimulus generalized to R(-)-3,4-MDA but not to S(+)-3,4-MDA, while the amphetamine-stimulus generalized to S(+)-3,4-MDA but not to R(-)-3,4-MDA (see Figure 5). Furthermore, in each case the isomer of 3,4-MDA which generalized to the respective training-drug was more potent than racemic 3,4-MDA (Glennon & Young, 1984b). Thus, although an MDA-like SAR was not developed, this series of studies exemplifies the use of previously generated SAR to study a novel agent using drug discrimination methodology.

Structures of various isomers
Figure 5: Structures of (a) the more isomer of DOM in DOM-trained animals, (b) the more active isomer of amphetamine in amphetamine-trained animals, (c) the isomer of 3,4-MDA active in DOM-trained animals, and (d) the isomer of 3,4-MDA active in amphetamine-trained animals.

SAR Studies on Benzodiazepine Derivatives

The benzodiazepines constitute the most widely prescribed class of compounds in current clinical use; the number of prescriptions for benzodiazepines written annually in the United States alone is thought to exceed 100 million (Tallman, Paul, Skolnick, & Gallager, 1980). The popularity of this class of agents is probably the result of its four major pharmacological actions: anxiolytic, muscle relaxant, sedative, and anticonvulsant. A number of comprehensive reviews have appeared dealing with the behavioral and pharmacological effects of the benzodiazepines (e.g., Haefely, Pieri, Polc, & Schaffner, 1981; Skolnick & Paul, 1982).

Our studies with benzodiazepines were not aimed at delineating structure-activity relationships per se but rather were an attempt to determine if a relationship exists between human potencies and discrimination-derived diazepam-like potencies in animals. A complete SAR erived diazepam-like potencies in animals. A complete SAR study on benzodiazepine derivatives using the drug discrimination paradigm would be an ambitious undertaking. The results of various human and animal studies have already identified several important positions on the benzodiazepine nucleus where the presence of substituents can have a dramatic effect on activity/potency; thus, a comprehensive SAR study might require the investigation of hundreds of compounds. Our studies were limited to a small number of benzodiazepine derivatives (Some are shown in Table 3.) which for the most part are clinically available. As a consequence all of the compounds are active in the drug discrimination paradigm and the overall range of potencies is only two orders of magnitude. Nevertheless, because the SAR for the benzodiazepines is already relatively well established (e.g., Sternbach, 1973), it should be possible to make use of these SARs to determine if the SAR generated from the drug discrimination studies adhere to these same generalities.

Using a two-lever operant choice task, rats were trained to discriminate 3.0 mg/kg of diazepam from saline under a fixed ratio-10 (FR-10) schedule of reinforcement for food (sweetened condensed milk) reward. Once the discrimination was learned, generalization studies were conducted using various doses of a variety of 1,4-benzodiazepine derivatives. Although our goal was to determine which members of the series would produce d members of the series would produce diazepam-like effects (i.e., effects similar to that produced by 3.0 mg/kg of diazepam administered 10 minutes prior to the test session), one of the by-products of the investigation was an accumulation of data that could be useful for the formulation of limited SAR. As was the case with the phenylisopropylamine derivatives, there was a significant correlation (r = 0.94; p < 0.001, N = 15) between the ED50 values of those agents to which the benzodiazepam-stimulus generalized and their previously reported "drug effect" potencies in humans (Zbinden & Randall, 1967).

In describing the SAR of the benzodiazepines, it should be realized that substituents at a number of positions around the benzodiazepine nucleus can influence activity as well as potency. The basic benzodiazepine skeleton is shown in Figure 6 and can be divided into the fused aromatic ring, the 1,4-diazepine ring, and the phenyl ring at the benzodiazepine 5-position; the ring numbering system is also shown. Some of the previously reported SAR generalities will be briefly described; comparisons will then be made between SAR and the results of our discrimination studies in order to determine if similar structure-activity relationships are apparent.

Removal of existing alkyl substituents from the 1-position of the benzodiazepines generally results in a slight decrease in potency (Sternbach, 1973); this same effect is(Sternbach, 1973); this same effect is evident upon comparing diazepam (ED50 = 1.16 mg/kg) with desmethyldiazepam (ED50 = 2.33 mg/kg), although the difference in potency between temazepam (ED50 = 1.36 mg/kg) and its demethyl counterpart oxazepam (ED50 = 1.35) is minimal. Ordinarily, removal of the 2-position carbonyl group of the benzodiazepines reduces potency somewhat, although the presence of this carbonyl function is not necessary for activity (Sternbach, 1973); a comparison of diazepam with medazepam (ED50 = 2.10 mg/kg) reveals a small drop in potency.

Table 3
Structure of Benzodiazepine Derivatives

Structures of benzodiazepine derivatives

Substituents at position:

Agent 1 2 1 2 3 7 X
Diazepam CH3 O H Cl CH
Oxazepam H O OH Cl CH
RO 5-2904 H O H CF3 CH
Temazepam CH3 O OH Cl CH
RO 5-3027 H O H OH Cl CH
RO 5-3027 H O H Cl C-Cl
Desmethyldiazepam H O H Cl CH
RO 5-4528 CH3 O H CN CH
Flunitrazepam CH3 O H NO2 C-F
3-Methylflunitrazepam CH3 O CH3 NO2 C-F
Lorazepam H O OH Cl C-F
Lorazepam H O OH Cl C-Cl
Medazepam CH3 H H Cl CH
RO 5-3590 H O H NO2 C-CF3
Clonazepam H O H NO2 C-Cl
Nitrazepam H O H NO2 CH
Bromazepam H O H Br N
Bromazepam H O H Br N
Flurazepam a O H Cl C-F

Note: a, Substituent at 1-position = CH2CH2N(C2H5)2.

Figure 6: The basic benzodiazepine skeleton showing the fused aromatic ring (A), the 1,4-diazepine ring (B), and the phenylthe fused aromatic ring (A), the 1,4-diazepine ring (B), and the phenyl ring (C) at the 5-position.
Basic benzodiazepine skeleton

Alkyl substitution at the 3-position normally reduces potency while the incorporation of a hydroxyl group at this position has relatively little effect on potency. Flunitrazepam (ED50 = 0.05 mg/kg) is clearly 10 times more potent than its alkyl derivative 3-methylflunitrazepam (ED50 = 0.65 mg/kg). Comparing diazepam with temazepam, desmethyldiazepam with oxazepam, or RO 5-3027 (ED50 = 0.13 mg/kg) with lorazepam (ED5013 mg/kg) with lorazepam (ED50 = 0.35 mg/kg) reveals that 3-hydroxylation has relatively little effect (with the greatest effect being a 3-fold increase in potency observed for hydroxylation of RO 5-3027) on potency. The presence of substituents at the 3-position creates a chiral center and hence the possibility of optical stereoisomers (e.g., see Figure 7). Evaluation of the S(+)-isomer of 3-methylflunitrazepam (ED50 = 0.34 mg/kg) revealed it to be twice as potent as its racemate; this suggests that it is the S(+)-isomer that is largely responsible for the activity of the racemic mixture.

The benzodiazepines normally possess an aromatic requirement (i.e., a C ring) at the 5-position of the heterocyclic skeleton; ordinarily aleton; ordinarily a phenyl, substituted phenyl, or 3-pyridyl ring is required for optimal activity. Compounds such as diazepam, temazepam, and desmethyldiazepam possess an unsubstituted phenyl ring, while bromazepam (ED50 = 0.71 mg/kg) is an example of 2-pyridyl derivative. Substitution on the 5-position phenyl ring at the 3’- or 4’-position is not well tolerated, whereas a 2’-chloro or 2’-fluoro substituent ordinarily enhances potency. Introduction of a 2’-chloro substituent to oxazepam (i.e., lorazepam), to nitrazepam (ED50 = 0.13 mg/kg (i.e., clonazepam, ED50 = 0.06 mg/kg), and to desmethyldiazepam (i.e., RO 5-3027), for example, resulted in an increase in potency in each case. Good activity was also shown with 2’-fluoro-substituted derivativestituted derivatives such as flunitrazepam and flurazepam (ED50 = 5.1 mg/kg). Substituents at the 2’-position other than chloro or fluoro are reported to cause no significant increase and can even decrease potency (Sternbach, 1973). In agreement with this generality, we found that incorporation of a trifluoromethyl group at the 2’-position of nitrazepam, to yield RO 5-3590 (ED50 = 0.22 mg/kg), actually decreased potency somewhat.

Optical isomers of 3-methylflunitrazepam
Figure 7: The structures of the optical isomers of 3-methylflunitrazepam—theal isomers of 3-methylflunitrazepam—the R-isomer (A) and its S-enantiomer (B).

Substitutions at positions 6, 8, or 9 of the A ring do not lead to effective compounds; however, the presence of an electron withdrawing group at the 7-position seems essential for optimal activity. We investigated several such derivatives. RO 5-4528 (ED50 = 1.18 mg/kg), which possesses a cyano group at the 7-position, appears to be equi-active with diazepam. In the 1-demethyl series a direct comparison can be made between desmethyldiazepam and its trifluoromethyl and nitro derivatives. Both the trifluoromethyl and nitro groups are more electron withdrawing than a chloro group, and,awing than a chloro group, and, if electron withdrawing effects are important to activity, it might be anticipated that the chloro derivative of desmethyldiazepam would be the least potent of the three. This was found to be the case. Desmethyldiazepam (ED50 = 2.33 mg/kg) was less active than either its trifluoromethyl (RO 5-2904, 0.46 mg/kg) or nitro (nitrazepam, 0.13 mg/kg) analogs.

In the formulation of structure-activity relationships, it is desirable to examine as many compounds as possible. Evaluation of large numbers of compounds will provide the most valid SAR and will also serve to identify more readily those agents that may be "exceptions." Although the goal of this present study was not to formulate SAR, the limited SAR that was generated was demonstrated to be consistent with known SAR generalities for the benzodiazepines. This only strengthens the above mentioned correlation between human potencies and discrimination-derived ED50 values for benzodiazepine derivatives; if the correlation is valid, structural variation should have similar (barring species differences) effects.


The drug discrimination paradigm is rapidly gaining popularity as a research tool; however, the use of this paradigm in medicinal chemistry—and in particular its application to investigations of structure-activity relationships—is still in its infancy. A relationships—is still in its infancy. Although the potential benefits of this procedure are obvious, additional studies will be required to explore the use of this method as well as to identify its possible shortcomings and limitations for SAR studies. Several factors that can influence the results of such studies have already been described. Perhaps one of the most important factors to be considered is the choice of doses to be examined for a particular challenge compound under investigation; the need for thorough dose-response investigations cannot be over-emphasized. In addition, the formulation of SAR for a series of agents should be reserved only for those members of the series showing generalization to the training-drug stimulus and, even then, only where identical testing conditions (e.g., presession injection interval, route of administration) have been employed. Because the drug discrimination paradigm is an in vivo procedure, SAR data obtained with this method will reflect the pharmacokinetics of the challenge drugs; variation in testing conditions from one challenge drug to another might serve to over-emphasize pharmacokinetic differences. In this respect drug discrimination is no different than any other in vivo method. On the other hand, the sensitive and specific nature of this procedure should make it an excellent tool for studying structure-activity relationships.


Much of the work from our laboratory was supported by Public Health Service grant DA-01642.


Glennon, R. A., Jacyno, J. M., & Young, R. (1983a). A comparison of the behavioral properties of (±)-, (-)-, and (+)-5-methoxy-a-methyltryptamine. Journal of Biological Psychiatry, 18, 493-498.

Glennon, R. A., Jacyno, J. M., Young, R., & Nelson, D. (1984). Synthesis and evaluation of a novel series of N,N-dimethylisotryptamines. Journal of Medicinal Chemistry, 27, 41-45.

Glennon, R. A., Rosecrans, J. A., & Young, R. (1982a). The use of the drug discrimination paradigm for studying hallucinogenic agents. A review. In F. C. Colpaert & J. L. Slangen (Eds.), Drug discrimination: Applications in CNS pharmacology (pp. 69-96). Amsterdam: Elsevier/North Holland Biomedical Press.

Glennon, R. A., Rosecrans, J. A., & Young, R. (1983b). Drug-induced discrimination: A description of the paradigm and a review of its specific application to the study of hallucinogenic agents. Medicinal Research Reviews, 3, 289-340.

Glennon, R. A., & Young, R. (1982). A comparison of behavioral properties of di- and tri-methoxyphenylisopropylamines. Pharmacology Biochemistry & Behavior, 17, 603-607. , 17, 603-607.

Glennon, R. A., & Young, R. (1984a). Further investigation of the discriminative stimulus properties of MDA. Pharmacology Biochemistry & Behavior, 20, 501-505.

Glennon, R. A., & Young, R. (1984b). MDA: A psychoactive agent with dual stimulus effects. Life Sciences, 34, 379-383.

Glennon, R. A., Young, R., Anderson, G. M., & Rosecrans, J. A. (1982b). Discriminative stimulus properties of MDA analogs. Journal of Biological Psychiatry, 17, 807-814.

Glennon, R. A., Young, R., Benington, F., & Morin, R. D. (1982c). Behavioral and serotonin receptor properties of 4-substituted derivatives of the hallucinogenic 1-(2,5-dimethyloxyphenyl)-2-aminopropane. Journal of Medical Chemistry, 25, 1163-1168.

Glennon, R. A., Young, R., & Jacyno, J. M. (1983c). Indolealkylamine and phenalkylamine hallucinogens: Effect of a-methyl and N-methyl substituents on behavioral activity. Biochemical Pharmacology, 32, 1267-1273.

Glennon, R. A., Young, R., Jacyno, J. M., Slusher, M., & Rosecrans, J. A. (1983d). DOM-stimulus generalization to LSD and other hallucinogenic indolealkylamines. European Journal of Pharmacology, 86, 453-459.

Glennon, R. A., Young, R., & Rosecrans, J. A. (1982d). A comparison of the behavioral effects of DOM homologs. Pharmacology Biochemcts of DOM homologs. Pharmacology Biochemistry & Behavior, 16, 557-559.

Glennon, R. A., Young, R., & Rosecrans, J. A. (1982e). Discriminative stimulus properties of DOM and several molecular modifications. Pharmacology Biochemistry & Behavior, 16, 553-556.

Haefely, W., Pieri, L., Polc, P., & Schaffner, R. (1981). General pharmacology and neuropharmacology of benzodiazepine derivatives. In F. Hoffmeister & G. Stille (Eds.), Psychotropic agents (Part II, pp. 13-262). New York: Springer-Verlag.

Silverman, P. B., & Ho, B. T. (1977). Characterization of discriminative response control by psychomotor stimulants. In H. Lal (Ed.), Discriminative properties of drugs (pp. 107-119). New York: Plenum Press.

Silverman, P. B., & Ho, B. T. (1978). Stimulus properties of DOM: Commonality with other hallucinogens. In F. C. Colpaert & J. A. Rosecrans (Eds.), Stimulus properties of drugs: Ten years of progress (pp. 189-198). Amsterdam: Elsevier/North Holland Biomedical Press.

Skolnick, P., & Paul, S. M. (1982). Benzodiazepine receptors in the central nervous system. International Review of Neurobiology, 23, 103-140.

Sternbach, L. H. (1973). Chemistry of 1,4-benzodiazepines and some aspects of the structure-activity relationships. In S. Garattini, E. Mussini, &nships. In S. Garattini, E. Mussini, & L. O. Randall (Eds.), The benzodiazepines (pp. 1-26). New York: Raven Press.

Tallman, J. F., Paul, S. M., Skolnick, P., & Gallager, D. W. (1980). Receptors for the age of anxiety: Pharmacology of the benzodiazepines. Science, 207, 274-281.

Young, R., & Glennon, R. A. (1986) Discriminative stimulus properties of amphetamine and structurally related phenalkylamines. Medicinal Research Reviews, 6, 99-130.

Young, R., Glennon, R. A., & Rosecrans, J. A. (1981). Discriminative stimulus properties of the hallucinogenic agent DOM. Communication in Psychopharmacology, 4, 501-506.

Zbinden, G., & Randall, L. O. (1967). Pharmacology of benzodiazepines: Laboratory and clinical correlations. In S. Garattini & P. A. Shore (Eds.), Advances in pharmacology (Vol. 5, pp. 213-291). New York: Academic Press.

©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 19:44 EDT.
Send comments to:
Report technical problems to: