CHAPTER 2
`Drug Discovery, Design,
`and Development
`
`5
`
`11
`
`4
`I. Drug Discovery
`A. Drug Discovery without a Lead
`1. Penicillins, 5 · 2. Librium, 6
`B. Lead Discovery
`7
`1. Random Screening, 8 · 2. Nonrandom Screening, 9 · 3. Drug Metabolism Studies,
`9 · 4. Clinical Observations, 9 · 5. Rational Approaches to Lead Discovery, 10
`II. Drug Development: Lead Modification
`11
`A. Identification of the Active Part: The Pharmacophore
`B. Functional Group Modification
`13
`C. Structure-Activity Relationships
`14
`15
`D. Structure Modifications to Increase Potency and Therapeutic Index
`1. Homologation, 16 · 2. Chain Branching, 18 · 3. Ring-Chain Transformations, 18·
`4. Bioisosterism, 19
`23
`E. Quantitative Structure-Activity Relationships
`1. Historical, 23 · 2. Physicochemical Parameters, 24 · 3. Methods Used to Correlate
`Physicochemical Parameters with Biological Activity, 35 · 4. Computer-Based
`Methods of QSAR Related to Receptor Binding, 43
`F. Molecular Graphics-Based Drug Design
`44
`G. Epilogue
`47
`References
`47
`General References
`
`50
`
`I. Drug Discovery
`
`In general, clinically used drugs are not discovered. What is more likely
`discovered is known as a lead compound. The lead is a prototype compound
`that has the desired biological or pharmacological activity, but may have
`many other undesirable characteristics, for example, high toxicity, other bio-
`logical activities, insolubility, or metabolism problems. The structure of the
`lead compound is then modified by synthesis to amplify the desired activity
`and to minimize or eliminate the unwanted properties. Prior to an elaboration
`of approaches to lead discovery and lead modification, two of the rare drugs
`discovered without a lead are discussed.
`
`4
`
`

`
`I. Drug Discovery
`
`5
`
`A Drug Discovery without a Lead
`1. Penicillins
`In 1928 Alexander Fleming noticed a green mold growing in a culture of
`Staphylococcus aureus, and where the two had converged, the bacteria were
`lysed.1 This led to the discovery of penicillin, which was produced by the
`mold. It may be thought that this observation was made by other scientists
`who just ignored it, and, therefore, Fleming was unique for following up on it.
`However, this is not the case. Fleming tried many times to rediscover this
`phenomenon without success; it was his colleague, Dr. Ronald Hare,23 who
`was able to reproduce the observation. It only occurred the first time because
`a combination of unlikely events all took place simultaneously. Hare found
`that very special conditions were required to produce the phenomenon ini-
`tially observed by Fleming. The culture dish inoculated by Fleming must have
`become accidentally and simultaneously contaminated with the mold spore.
`Instead of placing the dish in the refrigerator or incubator when he went on
`vacation as is normally done, Fleming inadvertently left it on his lab bench.
`When he returned the following month, he noticed the lysed bacteria. Ordi-
`narily, penicillin does not lyse these bacteria; it prevents them from develop-
`ing, but it has no effect if added after the bacteria have developed. However,
`while Fleming was on vacation (July to August) the weather was unseason-
`ably cold, and this provided the particular temperature required for the mold
`and the staphylococci to grow slowly and produce the lysis. Another extraor-
`dinary circumstance was that the particular strain of the mold on Fleming's
`culture was a relatively good penicillin producer, although most strains ofthat
`mold (Pénicillium) produce no penicillin at all. The mold presumably came
`from the laboratory just below Fleming's where research on molds was going
`on at the time.
`Although Fleming suggested that penicillin could be useful as a topical
`antiseptic, he was not successful in producing penicillin in a form suitable to
`treat infections. Nothing more was done until Sir Howard Florey at Oxford
`University reinvestigated the possibility of producing penicillin in a useful
`form. In 1940 he succeeded in producing penicillin that could be administered
`topically and systemically,4 but the full extent of the value of penicillin was
`not revealed until the late 1940s.5 Two reasons for the delay in the universal
`utilization of penicillin were the emergence of the sulfonamide antibacterials
`(sulfa drugs, 2.1; see Chapter 5, Section IV,B,1) in 1935 and the outbreak of
`World War II. The pharmacology, production, and clinical application of
`penicillin were not revealed until after the war so that this wonder drug would
`
`H2N—&
`
`\ - S 0 2 N HR
`
`2.1
`
`

`
`6
`
`2. Drug Discovery, Design, and Development
`
`not be used by the Germans. A team of Allied scientists who were interrogat-
`ing German scientists involved in chemotherapeutic research were told that
`the Germans thought the initial report of penicillin was made just for commer-
`cial reasons to compete with the sulfa drugs. They did not take the report
`seriously.
`The original mold was Pénicillium notatum, a strain that gave a relatively
`low yield of penicillin. It was replaced by Pénicillium chrysogenum,6 which
`had been cultured from a mold growing on a grapefruit in a market in Peoria,
`Illinois! The correct structure of penicillin (2.2) was elucidated in 1943 by Sir
`Robert Robinson (Oxford) and Karl Folkers (Merck). Several different peni-
`cillin analogs (R group varied) were isolated early on; only two of these (2.2,
`R = PhOCH2, penicillin V, and 2.2, R = CH2Ph, penicillin G) are still in use
`today.
`
`2. Librium
`The first benzodiazepine tranquilizer drug, Librium [7-chloro-2-(methyl-
`amino)-5-phenyl-3//-l,4-benzodiazepine 4-oxide, 2.3], was discovered seren-
`dipitously.7 Dr. Leo Sternbach at Roche was involved in a program to synthe-
`size a new class of tranquilizer drugs. He originally set out to prepare a series
`of benzheptoxdiazines (2.4), but when R1 was CH2NR2 and R2 was C6H5, it
`was found that the actual structure was that of a quinazoline 3-oxide (2.5).
`However, none of these compounds gave any interesting pharmacological
`results. The program was abandoned in 1955 in order for Sternbach to work
`on a different project. In 1957 during a general laboratory cleanup a vial
`containing what was thought to be 2.5 (X = 7-C1, R1 = CH2NHCH3, R2 =
`C6H5) was found and, as a last effort, was submitted for pharmacological
`testing. Unlike all the other compounds submitted, this one gave very promis-
`ing results in six different tests used for preliminary screening of tranquilizers.
`
`

`
`I. Drug Discovery
`
`7
`
`N ^ C H 2 CI
`^Γ
`
`Ό
`
`CH3NH2
`
`gCNHCH3
`P^CH2C1
`
`2.6
`
`^N^^CH 2NHCH 3
`
`NHCH,
`
`2.3
`
`Scheme 2.1. Mechanism for formation of Librium.
`Further investigation revealed that the compound was not a quinazoline
`3-oxide but, rather, was the benzodiazepine 4-oxide (2.3), presumably pro-
`duced in an unexpected reaction of the corresponding chloromethyl quinazo-
`line 3-oxide (2.6) with methylamine (Scheme 2.1). If this compound had not
`been found in the laboratory cleanup, all of the negative pharmacological
`results would have been reported for the quinazoline 3-oxide class of com-
`pounds, and benzodiazepine 4-oxides may not have been discovered for many
`years to come.
`The examples of drug discovery without a lead are quite few in number.
`The typical occurrence is that a lead compound is identified and its structure
`is modified to give, eventually, the drug that goes to the clinic.
`
`B. Lead Discovery
`
`Penicillin V and Librium are, indeed, two important drugs that were discov-
`ered without a lead. Once they were identified, however, they then became
`lead compounds for future analogs. There are now a myriad of penicillin-
`derived antibacterials that have been synthesized as the result of the structure
`elucidation of the earliest penicillins. Valium (diazepam, 2.7) was synthesized
`at Roche even before Librium was introduced on to the market; this drug was
`derived from the lead compound Librium and is almost 10 times more potent
`than the lead.
`
`

`
`8
`
`2. Drug Discovery, Design, and Development
`
`In general, the difficulty arises in the discovery of the lead compound.
`There are several approaches that can be taken to identify a lead. The first
`requirement for all of the approaches is to have a means to assay compounds
`for a particular biological activity, so that it will be known when a compound
`is active. A bioassay (or screen) is a means of determining in a biological
`system, relative to a control compound, whether a compound has the desired
`activity and, if so, what the relative potency of the compound is. Note the
`distinction between the terms activity and potency. Activity is the particular
`biological or pharmacological effect (e.g., antibacterial activity or anticon-
`vulsant activity); potency is the strength of that effect. Some bioassays (or
`screens) begin as in vitro tests, for example, the inhibition of an enzyme or
`antagonism of a receptor; others are in vivo tests, for example, the ability of
`the compound to prevent an induced seizure in a mouse. In general, the in
`vitro tests are quicker and less expensive. Once the bioassay is developed,
`there are a variety of approaches to identify a lead.
`
`1. Random Screening
`Random screening involves no intellectualization; all compounds are tested in
`the bioassay without regard to their structures. Prior to 1935 (the discovery of
`sulfa drugs), random screening was essentially the only approach; today this
`method is used to a lesser degree. However, random screening programs are
`still very important in order to discover drugs or leads that have unexpected
`and unusual structures for various targets.
`The two major classes of materials screened are synthetic chemicals and
`natural products (microbial, plant, and marine). An example of a random
`screen of synthetic and natural compounds is the "war on cancer" declared
`by Congress and the National Cancer Institute in the early 1970s. Any new
`compound submitted was screened in a mouse tumor bioassay. Few new
`anticancer drugs resulted from that screen, but many known anticancer drugs
`also did not show activity in the screen used. As a result ofthat observation,
`multiple bioassay systems are now utilized. In the 1940s and 1950s a random
`screen by various pharmaceutical companies of soil samples in search of new
`antibiotics was undertaken. In this case, however, not only were numerous
`leads uncovered, but two important antibiotics, streptomycin and the tetracy-
`clines, were found.
`
`

`
`I. Drug Discovery
`
`9
`
`2. Nonrandom Screening
`Nonrandom screening is a slightly more narrow approach than is random
`screening. In this case compounds having a vague resemblance to weakly
`active compounds uncovered in a random screen or compounds containing
`different functional groups than leads may be tested selectively. By the late
`1970s the National Cancer Institute's random screen was modified to a non-
`random screen because of budgetary and manpower restrictions. Also, the
`single tumor screen was changed to a variety of tumor screens, as it was
`realized that cancer is not just a single disease.
`
`3. Drug Metabolism Studies
`During metabolism studies drug metabolites (drug degradation products gen-
`erated in vivo) that are isolated are screened in order to determine if the
`activity observed is derived from the drug candidate or from a metabolite. For
`example, the anti-inflammatory drug sulindac (2.8) is not the active agent; the
`metabolic reduction product, 2.9, is responsible for the activity.8 A classic
`example of this approach is the discovery of the antibacterial agent sulfanil-
`amide (2.1, R = H), which was found to be a metabolite of prontosil (2.10)
`(see Chapter 5, Section IV,B,1 for details).
`
`-COOH
`
`>—CH 3
`
`^<s^-~~/
`
`JJ
`
`Fvo
`^J
`
`CH3
`
`2.8
`
`r-COOH
`F - Q A _ C „ S
`Π
`
`S o
`
`|
`
`2.9
`
`NH2
`J>-N=N^^rA-S02NH2
`
`NH2~C
`
`2.10
`
`4. Clinical Observations
`Often a drug candidate during animal testing or clinical trials will exhibit more
`than one pharmacological activity; that is, it may produce a side effect. This
`compound, then, can be used as a lead for the secondary activity. In 1947 an
`antihistamine, dimenhydrinate (2.11; Dramamine®) was tested at the allergy
`clinic at Johns Hopkins University and was found also to be effective in
`
`

`
`10
`
`2. Drug Discovery, Design, and Development
`
`relieving a patient who suffered from car sickness; a further study proved its
`effectiveness in the treatment of seasickness9 and airsickness.10 It is now the
`most widely used drug for the treatment of all forms of motion sickness.
`An antibacterial agent, carbutamide (2.12, R = NH2), was found to have an
`antidiabetic side effect. However, it could not be used as an antidiabetic drug
`because of its antibacterial activity. Carbutamide, then, was a lead for the
`discovery of tolbutamide (2.12, R = CH3), an antidiabetic drug without anti-
`bacterial activity.
`
`NMe9
`
`C H 3 ^
`
`'
`
`0K^a
`
`I
`
`2.11
`
`O
`II
`
`S02NHCNHCH2CH2CH2CH3
`
`2.12
`
`5. Rational Approaches to Lead Discovery
`None of the above approaches to lead discovery involves a major rational
`component. The lead is just found by screening techniques, as a by-product of
`drug metabolism studies, or from whole animal investigations. Is it possible to
`design a compound having a particular activity? Rational approaches to drug
`design have now become the major routes to lead discovery. The first step is
`to identify the cause for the disease state. Most diseases, or at least the
`symptoms of diseases, arise from an imbalance of particular chemicals in the
`body, from the invasion of a foreign organism, or from aberrant cell growth.
`As discussed in later chapters, the effects of the imbalance can be corrected
`by antagonism or agonism of a receptor (see Chapter 3) or by inhibition of a
`particular enzyme (see Chapter 5). Foreign organism enzyme inhibition and
`interference with DNA biosynthesis or function (see Chapter 6) are also im-
`portant approaches to treat diseases arising from microorganisms and aber-
`rant cell growth.
`Once the relevant biochemical system is identified, lead compounds then
`become the natural receptor agonists or enzyme substrates. For example, lead
`compounds for the contraceptives (+)-norgestrel (2.13) and 17a-ethynylestra-
`diol (2.14) were the steroidal hormones progesterone (2.15) and 17ß-estradiol
`
`CH3ÇH2 JJH
`
`\ """"C= CH
`
`"" CE CH
`
`Hj ft
`
`2.13
`
`2.14
`
`

`
`II. Drug Development: Lead Modification
`
`11
`
`(2.16). Whereas the steroid hormones 2.15 and 2.16 show weak and short-
`lasting effects, the oral contraceptives 2.13 and 2.14 exert strong progesta-
`tional activity of long duration.
`
`2.15
`2.16
`At Merck it was believed that serotonin (2.17) was a possible mediator of
`inflammation. Consequently, serotonin was used as a lead for anti-inflamma-
`tory agents, and from this lead the anti-inflammatory drug indomethacin (2.18)
`was developed.11
`
`OH
`
`CH.O.
`
`NH9
`
`HO,
`
`N
`H
`2.17
`2.18
`The rational approaches are directed at lead discovery. It is not possible,
`with much accuracy, to foretell toxicity and side effects, anticipate transport
`characteristics, or predict the metabolic fate of a drug. Once a lead is identi-
`fied, its structure can be modified until an effective drug is prepared.
`
`II. Drug Development: Lead Modification
`
`Once your lead compound is in hand, how do you know what to modify in
`order to improve the desired pharmacological properties?
`
`A Identification of the Active Part: The Pharmacophore
`
`Interactions of drugs with receptors are very specific (see Chapter 3). There-
`fore, only a small part of the lead compound may be involved in the appropri-
`ate interactions. The relevant groups on a molecule that interact with a recep-
`
`

`
`12
`
`2. Drug Discovery, Design, and Development
`
`tor and are responsible for the activity are collectively known as the
`pharmacophore. If the lead compound has additional groups, they may inter-
`fere with the appropriate interactions. One approach to lead modification is to
`cut away sections of the molecule in order to determine what parts are essen-
`tial and which are superfluous.
`As an example of how a molecule can be trimmed and still result in in-
`creased potency or modified activity, consider the addictive analgetics mor-
`phine (2.19, R = R' = H), codeine (2.19, R = CH3, R' = H), and heroin
`(R = R' = COCH3). The pharmacophore is darkened. If the dihydrofuran
`oxygen is excised, morphinan (2.20, R = H)12 results; the hydroxy analog
`levorphanol13 (2.20, R = OH) is 3 to 4 times more potent than morphine as an
`analgetic, but it retains the addictive properties. Removal of half of the cyclo-
`hexene ring, leaving only methyl substituents, gives benzomorphan (2.21,
`R = CH3).14 This compound shows some separation of analgetic and addic-
`tive effects; cyclazocine (2.21, R = CH2—<^ ) and pentazocine [2.21, R =
`CH2CH=C(CH3)2] are analogs with much lower addiction liabilities. Cutting
`away the cyclohexane fused ring (2.22) also has little effect on the analgetic
`activity in animal tests. Removal of all fused rings, for example, in the case of
`meperidine (2.23, Demerol®), gives an analgetic still possessing 10-12% of the
`overall potency of morphine.15 Even acyclic analogs are active. Dextropro-
`poxyphene (2.24, Darvon®) is one-half to two-thirds as potent as codeine; its
`activity can be ascribed to the fact that it can assume a conformation related
`to that of the morphine pharmacophore. Another acyclic analog is methadone
`(2.25) which is as potent an analgetic as morphine; the (-)-isomer is used in
`the treatment of opioid abstinence syndromes in heroin abusers.
`
`2.19
`
`2.20
`
`2.21
`
`2.22
`
`2.23
`
`2.24
`
`2.25
`
`

`
`II. Drug Development: Lead Modification
`
`13
`
`In some cases an increase in structural complexity and/or rigidity can lead
`to increased potency. For example, an oripavine derivative such as etorphine
`(2.26, R = CH3,R' = C3H7), which has a two-carbon bridge and a substituent
`not in morphine, is about 1000 times more potent than morphine16 and, there-
`fore, is used in veterinary medicine to immobilize large animals. The related
`analog, buprenorphine (2.26, R = CH2—<3, R' = tert-Bu, double bond re-
`duced) is 10-20 times more potent than morphine and has a very low level of
`dependence liability. Apparently, the additional rigidity of the oripavine de-
`rivatives increases the appropriate receptor interactions (see Chapter 3).
`Once the pharmacophore is identified, manipulation of functional groups
`becomes consequential.
`
`S. Functional Group Modification
`
`The importance of functional group modification was seen in Section I,B,4;
`the amino group of carbutamide (2.12, R = NH2) was replaced by a methyl
`group to give tolbutamide (2.12, R = CH3), and in so doing the antibacterial
`activity was separated away from the antidiabetic activity. In some cases an
`experienced medicinal chemist knows what functional group will elicit a par-
`ticular effect. Chlorothiazide (2.27) is an antihypertensive agent that has a
`strong diuretic (increased urine excretion) effect as well. It was known from
`sulfanilamide work that the sulfonamide side chain can give diuretic activity
`(see Section II,C). Consequently, diazoxide (2.28) was prepared as an antihy-
`pertensive drug without diuretic activity.
`There, obviously, is a relationship between the molecular structure of a
`compound and its activity. This phenomenon was first realized over 120 years
`ago.
`
`N H 2 S 0 2 / \ ^ S ^ o
`
`Cl
`
`( ^ b
`
`2.27
`
`2.28
`
`

`
`14
`
`2. Drug Discovery, Design, and Development
`
`C. Structure-Activity Relationships
`
`In 1868 Crum-Brown and Fraser,17 suspecting that the quaternary ammonium
`character of curare may be responsible for its muscular paralytic properties,
`examined the neuromuscular blocking effects of a variety of simple quater-
`nary ammonium salts and quaternized alkaloids in animals. From these stud-
`ies they concluded that the physiological action of a molecule was a function
`of its chemical constitution. Shortly thereafter, Richardson18 noted that the
`hypnotic activity of aliphatic alcohols was a function of their molecular
`weight. These observations are the basis for future structure-activity relation-
`ships (SAR).
`Drugs can be classified as being structurally specific or structurally nonspe-
`cific. Structurally specific drugs, which most drugs are, act at specific sites,
`such as a receptor or an enzyme. Their activity and potency are very suscepti-
`ble to small changes in chemical structure; molecules with similar biological
`activities tend to have common structural features. Structurally nonspecific
`drugs have no specific site of action and usually have lower potency. Similar
`biological activities may occur with a variety of structures. Examples of these
`drugs are gaseous anesthetics, sedatives and hypnotics, and many antiseptics
`and disinfectants.
`Even though only a part of the molecule may be associated with the activ-
`ity, there are a multitude of molecular modifications that could be made. Early
`SAR studies (prior to the 1960s) simply involved the syntheses of as many
`analogs as possible of the lead and their testing to determine the effect of
`structure on activity (or potency). Once enough analogs were prepared and
`sufficient data accumulated, conclusions could be made regarding structure-
`activity relationships.
`An excellent example of this approach came from the development of the
`sulfonamide antibacterial agents (sulfa drugs). After a number of analogs of
`the lead compound sulfanilamide (2.1, R = H) were prepared, it was found
`that compounds of this general structure exhibited diuretic and antidiabetic
`activities as well as antimicrobial activity. Compounds with each type of
`activity eventually were shown to possess certain structural features in com-
`mon. On the basis of the biological results of greater than 10,000 compounds,
`several SAR generalizations have been made.19 Antimicrobial agents have
`structure 2.29 (R = S02NHR' or S03H) where (1) the amino and sulfonyl
`groups on the benzene ring should be para; (2) the anilino amino group may be
`unsubstituted (as shown) or may have a substituent that is removed in vivo; (3)
`replacement of the benzene ring by other ring systems, or the introduction of
`
`NH2—ft
`
`/~R
`
`2.29
`
`

`
`II. Drug Development: Lead Modification
`
`15
`
`additional substituents on it, decreases the potency or abolishes the activity;
`(4) R may be
`
`O
`O
`O
`S 0 2 - ^ \ - N H 2 , S O - ^ ~ " V N H 2 , CNH2, CNHR, or C - / ~ \ _R
`
`but the potency is reduced in most cases; (5) N'-monosubstitution (R =
`S02NHR') results in more potent compounds, and the potency increases with
`heteroaromatic substitution; and (6) N'-disubstitution (R = S02NR2), in gen-
`eral, leads to inactive compounds.
`Antidiabetic agents are compounds with structure 2.30, where X may be O,
`S, or N incorporated into a heteroaromatic structure such as a thiadiazole or a
`pyrimidine or in an acyclic structure such as a urea or thiourea. In the case of
`ureas, the N2 should carry as a substituent a chain of at least two carbon
`atoms.20
`
`R—-^
`
`y—S02NHC—Ν^
`X\
`:
`
`2.30
`Sulfonamide diuretics are of two general structural types, hydrochlo-
`rothiazides (2.31) and the high ceiling type (2.32). The former compounds
`have 1,3-disulfamyl groups on the benzene ring, and R2 is an electronega-
`tive group such as Cl, CF3, or NHR. The high ceiling compounds contain
`l-sulfamyl-3-carboxy groups. Substituent R2 is Cl, Ph, or PhZ, where Z may
`be O, S, CO, or NH, and X can be at position 2 or 3 and is normally NHR,
`OR, or SR.21
`The sulfonamide example is strong evidence to support the notion that a
`correlation does exist between structure and activity, but how do you know
`what molecular modifications to make in order to fine-tune the lead com-
`pound?
`
`N y R 1
`
`R\
`
`NH2SOf
`
`o ^O
`
`R'NHSOf
`
`2.31
`
`2.32
`
`D. Structure Modifications to Increase Potency and Therapeutic Index
`
`In the preceding section it was made clear that structure modifications were
`the keys to activity and potency manipulations. After years of structure-
`activity relationship studies, various standard molecular modification ap-
`
`

`
`16
`
`2. Drug Discovery, Design, and Development
`
`proaches have been developed for the systematic improvement of the thera-
`peutic index (also called the therapeutic ratio), which is a measure of the ratio
`of undesirable to desirable drug effects. For in vivo systems the therapeutic
`index could be the ratio of the LD50 (the lethal dose for 50% of the test
`animals) to the ED50 (the effective dose that produces the maximum therapeu-
`tic effect in 50% of the test animals). The larger the therapeutic index, the
`greater the margin of safety of the compound. A number of these structural
`modification methodologies follow.
`
`1. Homologation
`A homologous series is a group of compounds that differ by a constant unit,
`generally a CH2 group. As will become more apparent in Section ΙΙ,Ε, biologi-
`cal properties of homologous compounds show regularities of increase and
`decrease. For many series of compounds, lengthening of a saturated carbon
`side chain from one (methyl) to five to nine atoms (pentyl to nonyl) produces
`an increase in pharmacological effects; further lengthening results in a sudden
`decrease in potency (Fig. 2.1). In Section II,E,2,b it will be shown that this
`phenomenon corresponds to increased lipophilicity of the molecule, which
`permits penetration into cell membranes' until its lowered water solubility
`becomes problematic in its transport through aqueous media. In the case of
`aliphatic amines another problem is micelle formation, which begins at about
`Ci2. This effectively removes the compound from potential interaction with
`the appropriate receptors. One of, if not the, earliest example of this potency
`versus chain length phenomenon was reported by Richardson,18 who was
`
`o
`OH
`
`- I — i — i — i — i — i — i — i — i — i — r"
`2 3 4
`5 6 7 8 9 10 11 12
`
`C Chain Length
`
`Figure 2.1. General effect of carbon chain length on drug potency.
`
`

`
`II. Drug Development: Lead Modification
`
`Table 2.1 Effect of Chain Length on Potency: Antibacterial Activity of
`4-n-Alkylresorcinols22a and Spasmolytic Activity of Mandelate Esters22b
`OH
`—
`OH
`ft V-CHCO2R
`
`#
`
`OH
`
`R
`
`methyl
`
`ethyl
`
`Λ-propyl
`
`«-butyl
`
`/i-pentyl
`
`n-hexyl
`
`n-heptyl
`
`«-octyl
`
`n-nonyl
`
`n-decyl
`
`/l-undecyl
`
`/-propyl
`
`/-butyl
`
`t'-amyl
`
`i-hexyl
`
`Phenol coefficient
`
`% Spasmolytic activity0
`
`—
`
`—
`
`5
`
`22
`
`33
`
`51
`
`30
`
`0
`
`0
`
`0
`
`0
`
`—
`
`15.2
`
`23.8
`
`27
`
`0.3
`
`0.7
`
`2.4
`
`9.8
`
`28
`
`35
`
`51
`
`130
`
`190
`
`37
`
`22
`
`0.9
`
`8.3
`
`28
`
`—
`
`a Relative to 3,3,5-trimethylcyclohexanol, set at 100%.
`
`investigating the hypnotic activity of alcohols. The maximum effect occurred
`for 1-hexanol to 1-octanol; then the potency declined upon chain lengthening
`until no activity was observed for hexadecanol.
`A study by Dohme et al.22a on 4-alkyl-substituted resorcinol derivatives
`showed that the peak antibacterial activity occurred with 4-n-hexylresorcinol
`(see Table 2.1), a compound now used as a topical anesthetic in a variety of
`throat lozenges. Funcke et alF° found that the peak spasmolytic activity of a
`series of mandelate esters occurred with the «-nonyl ester (see Table 2.1).
`
`

`
`18
`
`2. Drug Discovery, Design, and Development
`
`2. Chain Branching
`When a simple lipophilic relationship is important as described above, then
`chain branching lowers the potency of a compound. This phenomenon is
`exemplified by the lower potency of the compounds having isoalkyl chains in
`Table 2.1. Chain branching also can interfere with receptor binding. For ex-
`ample, phenethylamine (PhCH2CH2NH2) is an excellent substrate for mono-
`amine oxidase [amine oxidase (flavin-containing)], but a-methylphenethyl-
`amine (amphetamine) is a poor substrate. Primary amines often are more
`potent than secondary amines which are more potent than tertiary amines.
`For example, the antimalarial drug primaquine (2.33) is much more potent
`than its secondary or tertiary amine homologs.
`Major pharmacological changes can occur with chain branching and homol-
`ogation. Consider the 10-aminoalkylphenothiazines (2.34, X = H). When R is
`CH2CH(CH3)N(CH3)2 (promethazine) or CH2CH2N(CH3)2 (diethazine), anti-
`spasmodic and antihistaminic activities predominate. However, the homolog
`2.34 with R being CH2CH2CH2N(CH3)2 (promazine) has greatly reduced anti-
`spasmodic and antihistaminic activities, but sedative and tranquilizing activi-
`ties are greatly enhanced. In the case of the branched chain analog 2.34 with R
`equal to CH2CH(CH3)CH2N(CH3)2 (trimeprazine), the tranquilizing activity is
`reduced and antipruritic (anti-itch) activity increases.
`
`2.33
`
`2.34
`
`3. Ring-Chain Transformations
`Another modification that can be made is the transformation of alkyl substi-
`tuents into cyclic analogs. Consider the promazines again (2.34). Chlorproma-
`zine [2.34, X = Cl, R = CH2CH2CH2N(CH3)2] and 2.34 (X = Cl, R =
`]
`are equivalent as
`tranquilizers
`in animal
`tests.
`CH2CH2CH2N
`
`Trimeprazine [2.34, X = H, R = CH2CH(CH3)CH2N(CH3)2] and methdila-
`zine [2.34, X = H, R = CH2—CH—CH2
`/N—CH 3]
`CH2—CH2
`have similar antipruritic activity in man.
`
`

`
`II. Drug Development: Lead Modification
`
`19
`
`Different activities can result from a ring-chain transformation as well. For
`example, if the dimethylamino group of chlorpromazine is substituted by a
`
`methylpiperazine ring (2.34, X = Cl, R = CH2CH2CH2N
`
`NCH3; pro
`
`chlorperazine), the antiemetic (prevents nausea and vomiting) activity is
`greatly enhanced. In this case, however, an additional amino group is added.
`
`4. Bioisosterism
`Bioisosteres are substituents or groups that have chemical or physical similar-
`ities, and which produce broadly similar biological properties.23 Bioisosterism
`is a lead modification approach that has been shown to be useful to attenuate
`toxicity or to modify the activity of a lead, and it may have a significant role in
`the alteration of metabolism of a lead. There are classical isosteres2425 and
`nonclassical isosteres.2326 In 1925 Grimm27 formulated the hydride displace-
`ment law to describe similarities between groups that have the same number
`of valence electrons but may have a different number of atoms. Erlenmeyer28
`later redefined isosteres as atoms, ions, or molecules in which the peripheral
`layers of electrons can be considered to be identical. These two definitions
`describe classical isosteres', examples are shown in Table 2.2. Nonclassical
`
`Table 2.2 Classical Isosteres24·25
`
`1. Univalent atoms and groups
`a. CH3 NH2 OH F Cl
`b. Cl
`PH2 SH
`c. Br
`/-Pr
`d. I
`/-Bu
`2. Bivalent atoms and groups
`—NH—
`a. —CH2—
`b. —COCH2R —CONHR
`3. Trivalent atoms and groups
`a. —CH=
`— N=
`b. —P=
`—As=
`4. Tetravalent atoms
`
`,4-
`
`-i-
`
`—O—
`—C02R
`
`—S—
`—COSR
`
`—Se—
`
`b. = C=
`5. Ring equivalents
`a. — CH=CH—
`b. —CH=
`c. —0—
`
`+
`= N=
`
`—S—
`—N=
`—S—
`
`+
`= p=
`
`(e.g., benzene, thiophene)
`(e.g., benzene, pyridine)
`—CH2— —NH—
`(e.g., tetrahydrofuran,
`tetrahy drothiophene,
`cyclopentane, pyrrolidine)
`
`

`
`Table 2.3 Nonclassical Bioisosteres23
`
`Carbonyl group
`
`N C ^ . CN
`M
`
`II
`N
`I
`oi
`Carboxylic acid group
`
`2.
`
`- S—N-H
`
`- S—OH
`Il
`O
`
`P—OH
`I
`NH2
`
`P—OH
`I
`OEt
`
`àr
`
`N-N
`
`N
`
`3.
`
`Hydroxy group
`
`OHj —NHCR —NHS02R
`
`—CH2OH —NHCNH2
`
`—NHCN
`
`—CH(CN)2
`
`4.
`
`Catechol
`
`5d co x x
`
`Ν^,Χ
`
`X = 0,NR
`
`S.
`
`Halogen
`
`ΡΠ CF, CN N(CN)2 C(CN)3
`
`,
`
`,
`
`NC CN
`
`[X| A X
`Thiourea
`
`ÇN
`
`/Nx
`
`7.
`
`S
`
`il
`NH2
`
`-NH
`
`N02
`
`• Λ χ ιυ
`
`- N=
`
`- C=
`
`Pyridine
`
`0\ 9 9 Ç
`
`N0 2
`
`R
`
`NR3
`
`10.
`
`acer group
`Spacer group
`
`^^1 -Q-
`
`11.
`
`Hydrogen
`
`

`
`II. Drug Development: Lead Modification
`
`21
`
`bioisosteres do not have the same number of atoms and do not fit the steric
`and electronic rules of the classical isosteres, but they do produce a similarity
`in biological activity. Examples of these are shown in Table 2.3.
`Ring-chain transformations also can be considered to be isosteric inter-
`changes. There are hundreds of examples of compounds that differ by a
`bioisosteric interchange2326; some examples are shown in Table 2.4. Bioisos-
`terism also can lead to changes in activity. If the sulfur atom of the
`phenothiazine neuroleptic drugs (2.34) is replaced by the —CH=CH— or
`—CH2CH2— bioisosteres, then dibenzazepine antidepressant drugs (2.35)
`result.
`
`It is, actually, quite surprising that bioisosterism should be such a success-
`ful approach to lead modification. Perusal of Table 2.2, and especially of Table
`2.3, makes it clear that in making a bioisosteric replacement, one or more of
`the following parameters will change: size, shape, ele