Patrick 4
`
`
`
`IUCrenem@outeng
`
`ee
`
`|.
`
`J
`r,ii
`
`w
`
`9 P
`
`I
`
`4
`
`fj If
`
`> ;.l ■'
`
`'*1
`
`nm
`
`■-.■S
`.^■
`
`\
`
`□
`
`l--r
`(y
`
`■T
`
`SAWAI EX. 1021
`Page 1 of 31
`
`

`

`An Introduction to
`
`Medicinal Chemistry
`
`GRAHAM L. PATRICK
`Department of Chemistry,
`Paisley University
`
`>:■
`
`f'mI
`!;
`
`I I
`
`I II I
`
`g
`
`Oxford New York Tokyo
`OXFORD UNIVERSITY PRESS
`1995
`
`SAWAI EX. 1021
`Page 2 of 31
`
`

`

`is
`
`...
`
`7 ■ Drug development
`
`For several thousand
`years, man has used herbs and potions as medicines, but it is
`,
`.
`only since the m,d-„n.eteenth centuo- that serious efforts were made to isAte and
`purify the active principles of these remedies. Since then, a large variety of biologically
`active compounds have been obtained and their structures determined (e.g. morohiHl
`TTikT™!’
`<!“«“ from the bark of the cinchona tree)
`These natural products became the lead compounds for a major synthetic effort
`where chemists made literally diousands of analogues in an attempt to impmf“
`what Nature had provided. The vast majority of this work was carried out wi* no real
`design or reason but out of the results came an appreciation of certain tacfc wUch
`generally worked. A pattern for drug development evolved. This chapter attempts to
`show what that pattern is and the useful tactics which can be employed for dev2“ng
`
`foltolg;"”*""''"'
`
`““““ follow the
`
`» Screening of natural compounds for biological activity.
`® Isolation and purification of the active principle.
`® Determination of structure.
`® Structure-activity relationships (SARs).
`® Synthesis of analogues.
`® Receptor theories.
`@ Design and synthesis of novel drug
`
`structures.
`
`7.1 Screening of natural products
`
`” natural products became highly popular following the discovery of
`mould. Plants, fungi, and bacterial strains were collected from all
`penici in
`Tjv” 1 \ ^
`metabolites with useful biological activities
`IS e m particu ar to an impressive arsenal of antibacterial agents (Chapter 9')
`Screening of natural products from plant and microbial
`'
`sources continues today in
`
`the never-endi
`marine source:
`this is a field 1
`
`7.2 Isolatioi
`The ease with
`much on the s
`Penicillin pi
`recognized the
`man, he disreg
`it in solution.
`Now that we
`purification pr
`ment of a new
`was achieved.
`Other adva:
`particular in tl
`techniques av<
`
`7.3 Structu
`In the past, d
`overcome. It i
`structure dete:
`structure whic
`or three decac
`was carried ou
`fully establish
`Structures 1
`to recognizabl
`proposed, bu
`structures anc
`natural compc
`Today, stru
`when the nat
`required to es
`In cases wl
`spectroscopy >
`
`SAWAI EX. 1021
`Page 3 of 31
`
`

`

`I;
`
`ii
`
`i■
`
`$
`
`tK
`>0 i
`?:f-' 'A
`
`1:
`
`-’.f-
`
`'Ilf-
`ill':
`
`i
`
`i;
`
`{
`
`I1!
`!
`
`ill- 'if
`
`■■
`
`never-ending quest to find new lead compounds. In recent years, organisms from
`m
`marine sources have given novel compounds with interesting biological activity and
`^tiis is a field likely to expand.
`
`Structure determination
`
`83
`
`|,2 Isolation and purification
`|nhe ease with which the active principle can be isolated and purified depends very
`the structure, stability, and quantity of the compound.
`on
`Penicillin proved a difficult compound to isolate and purify. Although Fleming
`recognized the antibiotic qualities of penicillin and its remarkable non-toxic nature to
`man, he disregarded it as a useful drug since it appeared too unstable. He could isolate
`it in solution, but whenever he tried to remove the solvent, the drug was destroyed.
`Now that we know the structure of penicillin (Chapter 9), its instability under the
`purification procedures of the day is understandable and it was not until the develop­
`ment of a new procedure called freeze-drying that a successful isolation of penicillin
`was achieved.
`Other advances in isolation techniques have occurred since those days and in
`are now a variety of chromatographic
`particular in the field of chromatography. There
`techniques available to help in the isolation and purification of a natural product.
`
`7.3 Structure determination
`In the past, determining the structure of a new compound was a major hurdle to
`overcome. It is sometimes hard for present-day chemists to appreciate how difficult
`structure determinations were before the days of NMR and IR spectroscopy. A novel
`structure which may now take a week’s work to determine would have provided two
`or three decades of work in the past. For example, the microanalysis of cholesterol
`was carried out in 1888 to get its molecular formula, but its chemical structure was not
`fully estabhshed until an X-ray crystallographic study was carried out in 1932.
`Structures had to be degraded to simpler compounds, which were further degraded
`to recognizable fragments. From these scraps of evidence, possible structures were
`proposed, but the only sure way of proving the theory was to synthesize these
`structures and to compare their chemical and physical properties with those of the
`natural compound or its degradation products.
`Today, structure determination is a relatively straightforward process and it is only
`when the natural product is obtained in minute quantities that a full synthesis is
`required to establish its structure.
`In cases where there is not enough sample for an IR or NMR analysis, mass
`spectroscopy can be helpful. The fragmentation pattern can give useful clues about
`
`is
`
`Mg
`i
`
`■ H
`M if
`
`'St
`
`. ....
`
`SAWAI EX. 1021
`Page 4 of 31
`
`

`

`1
`
`{
`
`j
`
`84
`
`Drug development
`
`N
`
`OH
`
`a
`
`H
`
`the structure, but it does not, how­
`ever, prove the structure. A full synth­
`esis is still required as final proof.
`Vinblastine (Fig. 7.1), an alkaloid
`used against advanced teratomas and
`lymphomas, is an example of how com­
`plex the structures of natural products
`N I
`can be. However, analytical skills and
`H
`Me H(
`instruments have advanced to such an
`extent that even this structure is rel­
`Fig. 7.1 Vinblastine.
`atively simple compared to some of the natural product structures being studied
`today.
`
`O.
`'•El
`
`^OCOMe
`COgMe
`
`MeOjC
`
`MeO'
`
`NH
`
`f*
`
`7.4 Structure-activity relationships
`Once the structure of a biologically active compound is known, the medicinal chemist
`is ready to move on to study the structure-activity relationships of the compound.
`The aim of such a study is to discover which parts of the molecule are important to
`biological activity and which are not. The chemist makes a selected number of
`compounds, which vary slightly from the original molecule, and studies what effect
`that has on the biological activity.
`One could imagine the drug as a chemical knight entering the depths of a forest (the
`body) in order to make battle with an unseen dragon (the body’s affliction) (Fig. 7.2).
`The knight (Sir Drugalot) is armed with a large variety of weapons and armour, but
`since his battle with the dragon goes unseen, it is impossible to tell which weapon he
`uses or whether his armour is essential to his survival. We only know of his success if
`he returns unscathed with the dragon slain. If the knight declines to reveal how he
`slew the dragon, then the only way to find out how he did it would be to remove some
`of his weapons and armour and to send him in against other dragons to see if he can
`still succeed.
`As far as a drug is concerned, the weapons and armour are the various chemical
`functional groups present in the structure, which can bind to the receptor or enzyme.
`We have to be able to recognize these functional groups and determine which ones are
`important.
`Let us imagine that we have isolated a natural product with the structure shown in
`Fig. 7.3. We shall name it Glipine. There are a variety of groups present in the
`structure and the diagram shows the potential bonding interactions which are possible
`with a receptor.
`It is unlikely that all of these interactions take place, so we have to identify those
`which do. By synthesizing compounds (such as the examples shown in Fig. 7.4) where
`
`SAWAI EX. 1021
`Page 5 of 31
`
`

`

`I
`
`il;
`
`\fi
`
`Ai
`
`1=
`
`5?;'t
`
`i i
`
`■
`
`i'i
`‘5
`im?!
`
`■Si
`
`iI
`
`I II1
`
`II
`■I I
`;g
`
`n
`
`Structure-activity relationships
`
`SAWAI EX. 1021
`Page 6 of 31
`
`

`

`I I Potential Ionic Binding Sites
`Potential Van Der Waals Binding
`Sites
`O Potential H-Bonding Binding Sites
`
`m
`
`I
`
`I I'
`
`I
`
`j
`
`I
`
`I
`1
`
`iOMe
`
`©Me
`iS:;
`
`,NMe
`
`Fig. 7.3 Glipine.
`
`OH
`
`OH
`
`CH3
`
`CH3
`
`CH3
`
`CH
`
`Fig. 7.4 Modifications of gfipine.
`
`86 Drug development
`
`CH3
`
`CH
`
`CH3
`
`CH3'
`
`OH
`
`OH
`
`OH
`
`OH
`
`NMe
`
`particular group of the molecule is removed or altered, it is possible to find out
`one
`which groups are essential and which are not.
`The ease with which this task can be carried out depends on how easily we can carry
`out the necessary chemical transformations.to remove or alter the relevant group. For j
`example, the importance or otherwise of an amine group is relatively easy to establish,
`whereas the importance of an aromatic ring might be more difficult. Hydr^xyLiEOUPS?
`amino groups, and aromatic rings are particularly common binding groups in medicinal
`chemistry, so let us consider what analogues could be synthesized to establish whether
`they are involved or not.
`7.4.1 The binding role of hydroxyl groups
`Hydroxyl groups are commonly involved in hydrogen bonding. Converting such a s
`group to a methyl ether or an ester is straightforward (Fig. 7.5) and will usually
`destroy or weaken such a bond.
`
`J
`
`}1
`
`R —OH
`
`CH3I
`
`R—OMe
`
`CH3COCI
`
`R—OH
`
`CH3SO2CI
`
`R—OH
`
`0
`
`R~0 C—CH3
`
`0
`
`R—O—S—CH3
`
`0
`
`■f
`
`LiAlH4
`
`R—H
`
`Rg. 7.5 Conversions of hydroxyl
`groups.
`
`i
`
`I
`
`I'
`
`is
`
`I;
`
`:t:
`
`.f.
`
`;d
`
`I
`
`11
`
`|i;
`i41
`
`I
`
`;r
`
`y
`
`SAWAI EX. 1021
`Page 7 of 31
`
`

`

`I®
`
`Receptor
`
`O
`
`Structure-activity relationships 87
`
`No Inleraction
`Me
`
`o'
`
`iQ
`
`DRUG
`
`DRUG
`(Methylated)
`Fig. 7.6 Possible hydrogen bond interactions.
`
`\l
`
`'■<>*$ i
`
`'oci; p
`
`)1
`
`m1
`
`s
`
`ti
`
`f-
`I
`
`sS
`
`I
`
`11
`
`i.. -i
`• j- >'■
`
`•''•5
`
`f j
`
`4
`
`•I-
`
`There are several possible explanations for this. The obvious explanation is that the
`iproton of the hydroxyl group is involved in the hydrogen bond to the receptor and if it
`removed, the hydrogen bond is lost (Fig. 7.6). However, suppose it is the oxygen
`IS
`atom which is hydrogen bonding to a suitable amino acid residue?
`The oxygen is still present in the ether or the ester analogue, so could we really
`expect there to be any effect on hydrogen bonding? Well, yes we could. The hydrogen
`bonding may not be completely destroyed, but we could reasonably expect it to be
`weakened, especially in the case of an ester.
`The reason is straightforward. When we consider the electronic properties of an
`ester compared to an alcohol, then we observe an important difference. The carboxyl
`group can ‘pull’ electrons from the neighbouring oxygen to give the resonance struc­
`ture shown in Fig. 7.7 Since the lone pair is involved in such an interaction, it cannot
`' take part so effectively in a hydrogen bond.
`Steric factors also count against the hydrogen bond. The extra bulk of the acyl
`group will hinder the close approach which was previously attainable.
`This steric hindrance also explains how a methyl ether could disrupt hydrogen
`bonding.
`If there is still some doubt over whether a hydroxyl group is involved in hydrogen
`
`|7V
`
`I
`
`^ lb
`
`O—C—CH3
`
`oO
`I
`0“ C—CH3
`e
`
`STERIC FACTOR
`ELECTRONIC FACTOR
`Fig. 7.7 Factors by which an ester group can disrupt hydrogen bonding.
`
`SAWAI EX. 1021
`Page 8 of 31
`
`

`

`I;
`F '
`
`It
`
`I
`
`iBi
`-B*'
`
`:SH
`
`!f F
`
`if' ■
`
`1)
`
`I;%
`
`88
`
`Drug development
`
`bonding or not, it could be replaced with an isosteric group such as methyl (see later).
`This would be more conclusive, but synthesis is more difficult.
`Another possibility is to react the hydroxyl group with methanesulfonyl chloride
`followed by lithium aluminum hydride (Fig. 7.5). This would replace the hydroxyl
`with a proton, but any group which is prone to reduction would have to be protected
`first.
`
`7.4.2 The binding role of amino groups
`Amines may be involved in hydrogen bonding or ionic bonding, but the latter is more
`common. The same strategy used for hydroxyl groups works here too. Converting the
`amine to an amide will prevent the nitrogen’s lone pair taking part in hydrogen
`bonding or taking up a proton to form an ion.
`Tertiary amines have to be dealkylated first, before the amide can be made.
`Dealkylation is normally carried out with cyanogen bromide or a chloroformate such
`as vinyloxycarbonyl chloride (Fig. 7.8).
`
`Fig. 7.8 Dealkylation of tertiary
`amines.
`
`R
`
`R'
`
`\
`/
`
`NMe
`
`CNBr
`
`R
`
`\
`
`orVOC-Cl R
`
`NH
`
`CHoCOCl
`
`o
`C—CHg
`
`R
`
`\
`
`N
`
`R'
`
`7.4.3 The binding role of aromatic rings
`Aromatic rings are commonly involved in van der Waals interactions with flat hydro-
`phobic regions of the binding site. If the ring is hydrogenated to a cyclohexane ring,
`the structure is no longer flat and interacts far less efficiently with the binding site
`(Fig. 7.9).
`
`Fig. 7.9 Reduction in the
`binding efficiency of aromatic
`rings by hydrogenation.
`
`Good
`Interaction
`
`H2
`
`H
`
`H
`
`Poor
`Interaction
`
`Flat Hydrophobic
`Region
`
`Flat Hydrophobic
`Region
`
`However, carrying out the reduction may well cause problems elsewhere in the
`structure, since aromatic rings are difficult to reduce and need forcing conditions.
`Replacing the ring altogether with a bulky alkyl group could reduce van der Waals
`bonding for the same reason given above, but obtaining such compounds could
`involve a major synthetic effort.
`
`SAWAI EX. 1021
`Page 9 of 31
`
`

`

`Synthetic analogues
`
`89
`
`r?
`
`i
`
`■r<
`
`ltd
`;t-i
`
`■1 i
`'J'
`
`15
`
`1^!
`
`i
`
`U'-iin
`•urn
`
`L''|l
`.4'i :■
`^ Is
`
`%
`
`7.4.4 The binding role of double bonds
`Unlike aromatic rings, double bonds are easy to reduce and this has a significant effect
`on the shape of that part of the molecule. The planar double bond is converted into a
`bulky alkyl group.
`If the original alkene was involved in van der Waals bonding with a flat surface on
`the receptor, reduction should weaken that interaction, since the bulkier product is
`less able to approach the receptor surface (Fig. 7.10).
`
`Flat
`
`H2
`Pd/C
`
`Bulky
`
`H
`
`H
`
`Fig. 7.10 The binding role
`of double bonds.
`
`Once it is established which groups are important for a drug’s activity, the medicinal
`chemist can move on to the next stage—the synthesis of analogues which still contain
`these essential features.
`
`7.5 Synthetic analogues
`Why is this stage necessary? If a natural compound such as our hypothetical Glipine
`has useful biological activity, why bother making synthetic analogues? The answer is
`that very few drugs are ideal. Many have serious side-effects and there is a great
`advantage in finding analogues which lack them. In general, the medicinal chemist is
`developing drugs with three objectives in mind:
`® to increase activity
`® to reduce side-effects
`® to provide easy and efficient administration to the patient
`Drug development in the past has mostly been a hit or a miss affair with a large
`number of compounds being synthesized at random. Luck has played a great part,
`but we can now recognize strategies which have evolved over the years:
`® variation of substituents
`® extension of the structure
`® chain extensions/contractions
`® ring expansions/contractions
`® ring variations
`® isosteres
`® simplification of the structure
`® rigidification of the structure
`
`1 I!f
`
`'
`
`SAWAI EX. 1021
`Page 10 of 31
`
`

`

`90 Drug development
`
`7.5.1 Variation of substituents
`Once the essential groups for biological activity have been identified, substituents are
`varied since this is usually quite easy to do synthetically. The aim here is to fine tune
`the molecule and to optimize its activity. Biological activity may depend not only on
`how well the compound interacts with its receptor, but also on a whole range of
`physical features such as basicity, lipophilicity, electronic distribution, and size (see
`Chapter 8). The idea of varying substituents is to attach a series of substituents such
`that these physical features are varied one by one. In reality, it is rarely possible to
`change one physical feature without affecting another. For example, replacing a
`methyl group on a nitrogen with an ethyl group could affect the basicity of the
`nitrogen atom, but the size of the molecule is also increased. Either of these changes
`might have an effect on the activity of a drug and it would be difficult to know which
`was more important without more results.
`The following are routine variations which can be carried out.
`
`Alkyl substituents
`If the molecule has an easily accessible functional group such as an alcohol, phenol, or
`amino group, then alkyl chains of various lengths and bulks such as methyl, ethyl,
`propyl, butyl, isopropyl, isobutyl or tert-hutyl can be attached.
`Different alkyl groups on a nitrogen atom may. alter the basicity and/or lipophilicity
`of the drug and thus affect how strongly the drug binds to its binding site or how
`easily the compound crosses membrane barriers (see Chapter 8).
`Larger alkyl groups, however, increase the bulk of the compound and this may
`confer selectivity on the drug. For example, in the case of a compound which interacts
`with two different receptors, a bulkier alkyl substituent may prevent the drug from
`binding to one of those receptors and so cut down side-effects (Fig. 7.11).
`
`RECEPTOR 1
`
`RECEPTOR 2
`Binding Site for N
`Fig. 7.11 Use of a larger alkyl group to confer selectivity on a drug.
`
`w
`
`1
`
`II
`I
`
`ii-i,
`
`I
`
`I
`
`I f
`
`f Si
`j;
`
`I;
`I
`
`II
`
`j
`
`I'
`r-
`f:
`if
`
`h
`
`II, Vi
`
`P
`
`SAWAI EX. 1021
`Page 11 of 31
`
`

`

`tli sj «I
`t v '
`; >« i
`
`(!■
`
`■i
`
`if ■ f
`
`t
`
`5
`
`in
`
`IC-
`
`S''"
`
`J
`
`I I
`
`Synthetic analogues
`
`91
`
`Aromatic substitutions
`A favourite approach for aromatic compounds is to vary the substitution pattern. This
`may give increased activity if the relevant binding groups are not already in the ideal
`positions for bonding (Fig. 7.12).
`
`Weak
`H-Bond'-
`
`O'
`
`Strong
`H-Bond'
`(increased
`activity)
`
`Receptor
`Surface
`
`Binding Site
`(H-Bond)
`Binding Site
`(forY)
`
`Receptor
`Surface /
`
`Y
`\j
`para Substitution
`
`meta Substitution
`Fig. 7.12 Aromatic substitutions.
`
`Electronic effects may also be involved. For example, an electron withdrawing nitro
`group will affect the basicity of an aromatic amine more significantly if it is in the para
`position rather than the meta position (Fig. 7.13). We have noted already that varying
`the basicity of a nitrogen atom may have a biological effect.
`If the substitution pattern is ideal, then we can try varying the substituents
`themselves. Substituents of different sizes and electronic properties are usually tried
`to see if steric and electronic factors have any effect on activity. It may be that activity
`is improved by having a more electron withdrawing substituent, in which case a nitro
`substituent might be tried in place of a chloro substituent.
`The chemistry involved in these procedures is usually straightforward and so these
`analogues are made as a matter of course whenever a novel drug structure is dis-
`
`NHj
`
`NHg
`
`oo
`
`O
`META (Inductive electron
`withdrawing effect)
`
`I®
`
`N
`
`o
`
`..©
`NHa
`
`©N
`
`O
`O
`
`o
`0
`
`PARA (Electron Withdrawing Effect
`due to Resonance and Inductive
`Effects leading to a Weaker Base)
`Fig. 7.13 Electronic effects of aromatic substitutions.
`
`SAWAI EX. 1021
`Page 12 of 31
`
`

`

`92
`
`Drug development
`
`covered or developed. Furthermore, the variation of aromatic or aliphatic substituents
`is open to quantitative structure-activity studies (QSARs) as described in Chapter 9.
`
`7.5.2 Extension of the structure
`This strategy has been used successfully on natural products such as morphine. It
`might seem strange that a natural product which is produced in a plant or a fungus
`should have important biological effects in the human body. One possible explanation
`for this could be that the natural product is present in the body as well. However, this
`seems unlikely. Therefore, we have to conclude that it is a happy coincidence that
`morphine has the right shape and binding groups to interact with a painkilling
`receptor in the body. This leads to some interesting conclusions.
`Since there is a painkilling receptor in the body we have to accept that there is a
`neurotransmitter (or hormone) which switches on that receptor. We already know
`that it cannot be morphine, so the painkilling molecule has to have a different shape
`and possibly different binding groups. Assuming that the body’s own painkiller is the
`ideal molecule for its receptor, then we must also conclude that morphine is not the
`ideal molecule. For example, it is perfectly possible that the natural painkiller has four
`important binding interactions with its receptor, whereas morphine has only three
`(Fig. 7.14). Therefore, why not add binding groups to the morphine skeleton
`to
`search for that fourth binding site? This tactic has been employed successfully to
`produce compounds such as the phenethyl analogue of morphine which has 14 times
`greater activity. Such a result suggests that the extra binding site on the receptor is
`hydrophobic, interacting with the aromatic ring by van der Waals interactions.
`Frequently, this extension tactic has resulted in a compound which acts as an
`antagonist rather than as an agonist. In this case, the extra binding site is not one used
`
`MORPHINE
`
`MORPHINE
`
`RECEPTOR
`
`Extension
`
`RECEPTOR
`
`HO.
`
`o
`
`N-i-CHg—CHg
`
`\VJf
`
`Extra
`Binding
`Group
`
`HO
`Fig. 7.14 Extension of morphine to provide a fourth binding group.
`
`li
`
`r. ■
`li - '
`
`:'r
`
`i :
`ll-
`'I-'.
`'i-
`
`I
`
`i\
`U '
`if;-
`
`i:
`
`|v
`
`s;
`
`s>I I
`I,
`1;.S-
`
`1--S:-
`m
`
`SAWAI EX. 1021
`Page 13 of 31
`
`

`

`Synthetic analogues 93
`
`V ■:
`
`by the natural agonist or substrate. The binding interaction is different and no
`biological response results.
`
`7.5.3 Chain extensions/contractions
`Some drugs have two important binding groups linked together by a chain. Many of
`the natural neurotransmitters are'like this. It is possible that the chain length is not
`ideal for the best interaction. Therefore, shortening or lengthening the chain length is
`a useful tactic to try (Fig. 7.15).
`
`Weak
`Interaction
`
`Chain
`Extension
`
`A
`
`]j
`
`Strong
`Interaction
`
`I receptorI
`
`[receptor!
`
`1
`
`BINDING SITES
`^ BINDING GROUPS
`B
`
`Fig. 7.15 Chain extension.
`
`7.5.4 Ring expansions/contractions
`If a drug has a ring, it is generally worth synthesizing analogues where one of these
`rings is expanded or contracted by one unit. The principle behind this approach is
`much the same as varying the substitution pattern of an aromatic ring. Expanding or
`contracting the ring puts the binding groups in different positions relative to each
`other and, with luck, may lead to better interactions with the binding site (Fig. 7.16).
`
`Receptor
`Surface^\
`
`Hydrophobic Regions
`
`6,6,6 Ring System has a good
`interaction with both hydrophobic
`Regions
`
`6,7,6 Ring System has the optimum
`interaction with both hydrophobic
`regions
`Fig. 7.16 Ring expansion.
`
`^■5.5 Ring variations
`A further popular approach for compounds containing an aromatic ring is to try
`replacing the aromatic ring with a range of heteroaromatic rings of different ring size
`^nd heteroatom positions. Admittedly, a lot of these changes are merely ways of
`
`i,fo;':
`
`s
`
`5;
`
`I,'
`I,
`
`7:
`
`Bi
`
`fi4 I
`c
`
`ii
`
`[](
`I-
`(•f '
`
`f5
`
`f
`:: 'I
`■ .•
`
`4—-
`
`!■
`
`3
`
`>. A
`
`M
`
`J
`
`SAWAI EX. 1021
`Page 14 of 31
`
`

`

`94
`
`Drug development
`restrictions and do not result in significant improvements, but some-
`avoiding patent
`times there are significant advantages in changing a ring system.
`One of the major advances in the development of the selective beta blockers was the
`replacement of the aromatic ring in adrenaline with a naphthalene ring systern
`(pronethalol) (Fig. 7.17). This resulted in a compound which was able to distmguis
`between two very similar receptors, the alpha and beta receptors for adrenaline One
`possible explanation for this could be that the beta receptor has a larger van der Waals
`binding area for the aromatic system than the alpha receptor and can interact more
`strongly with pronethalol than with adrenaline. Another possible explanation is that
`the naphthalene ring system is sterically too big for the alpha receptor but is )ust right
`for the beta receptor.
`
`HO
`
`H OH
`\ / "
`
`HO
`
`CHj—NHR
`
`OH
`‘O-
`
`H
`
`H
`N
`
`Me
`
`Me
`
`R = Me ADRENALINE
`R = H NORADRENALINE
`Fig. 7.17 Ring variation of adrenaline.
`
`PRONETHALOL
`
`7.5.6 Isosteres
`of atoms which have the same valency (or number of
`Isosteres are atoms or groups
`isosteres of OH, while S,
`shell electrons). For example, SH, NH2, and CH3 are
`outer
`NH and CH2 are isosteres of O. Isosteres have often been used to design an inhibitor
`or to increase metabolic stabUity. The idea is to alter the character of the molecule m
`as subtle a way as possible. Repla.cmg_.Q with CH„ for example, will make little
`difference to the size of the analogue, but will have a marked effect on its polarity,
`electronic distribution, and bonding. Replacing OH with the larger SH may not have
`. the electronic character, but steric factors become more significant,
`such an influence on
`could be used to determine whether a particular group is involved
`Isosteric groups l
`. ,
`,
`, j
`in hydrogen bonding. For example, replacing OH and CH3 would completely destroy
`hydrogen bonding, whereas replacing OH with NH2 would not.
`The beta blocker propranolol has an ether linkage (Fig. 7.18). Replacement of the
`CH=CH, SCH2, or CH2CH2 eliminates activity.
`OCH2 segment with the isosteres
`whereas replacement with NHCH2 retains activity (though reduced). These results
`show that the ether oxygen is important to the activity of the drug and suggests that it
`is involved in hydrogen bonding with -the receptor.
`with NH2 has been a useful tactic in
`Replacing the methyl of a methyl ester group
`_
`stabilizing esters which are susceptible to enzymatic hydrolysis <Fig. 7.19). 1 he NH2
`
`SAWAI EX. 1021
`Page 15 of 31
`
`

`

`II
`
`1-cti a
`
`{;•
`
`4 4I
`
`'4
`
`4 1
`
`:!
`
`l
`
`I
`
`. S'I'f
`a
`
`: 4
`"T
`I
`
`ill
`
`Synthetic anaiogues
`
`95
`
`O
`
`HgC—C
`
`O—R
`
`O C
`
`~0—R
`
`ISOSTERE
`
`Me
`
`O'
`
`NH
`
`Me
`
`H
`
`OH
`
`Fig. 7.18 Propranolol.
`
`Fig. 7.19 Isosteric replacement of a methyl with
`an amino group.
`
`group is the same size as the methyl and therefore has no steric effect. However, it has
`totally different electronic properties, and as such can feed electrons into the carboxyl
`and stabilize it from hydrolysis (see Chapter 11).
`group
`Although fluorine does not have the same valency as hydrogen, it is often con­
`sidered an isostere of that atom since it is virtually the same size. Replacement of a
`hydrogen atom with a fluorine atom will therefore have little steric effect, but since
`the fluorine is strongly electronegative,jhe.ekclronic_.effect may be dramatic.
`The use of fluorine as an isostere for hydrogen has been highly successful in recent
`One example is the antitumour drug 5-fluorouracil described in Section 4.5.3.
`years.
`The drug is accepted by the target enzyme since it appears little different from the
`normal substrate (uracil). However, the mechanism of the enzyme-catalysed reaction
`totally disrupted. Fluorine has replaced a hydrogen atom which must be lost as a
`IS
`during the mechanism. Themismoxhance of fluorine departing as a positively
`proton
`charged species.
`7.5.7 Simplification of the structure
`If the essential groups of a drug have been identified, then by implication, it might be
`possible to discard non-essential parts of the structure without losing activity. The
`advantage would be in gaining a far simpler compound which would be much easier
`and cheaper to synthesize in the laboratory. For example, let us consider our hypo­
`thetical natural product Glipine (Fig. 7.20). The essential groups have been
`marked and so we might aim to synthesize compounds such as those shown in
`Fig. 7.20. These have simpler structures, but still retain the groups which we have
`identified as being essential.
`This tactic was used successfully with the alkaloid cocaine (Fig. 7.21). It was well
`known that cocaine had local anaesthetic properties and it was hoped to develop a
`simplified structure of cocaine which could be easily
`focal anaesthetic based on a
`synthesized in the laboratory. Success resulted with the discovery of procaine (or
`Novocaine) in 1909.
`However, there is a trade-off involved when simplifying molecules. The advantage
`io obtaining simpler compounds may be outweighed by the disadvantage of increased
`
`1)
`
`1;
`
`f;-
`
`W
`
`tv.
`
`S'.
`
`M.
`
`SAWAI EX. 1021
`Page 16 of 31
`
`

`

`iS
`MiJ
`
`I
`
`J
`
`I
`
`I i
`
`j
`
`96
`
`Drug development
`
`OH
`
`OH
`
`OH
`
`OH
`
`GLIPINE
`
`A
`
`B
`Fig. 7.20 Glipine analogues.
`
`3
`
`c
`
`CH3
`
`CHg
`
`D
`
`COCAlNt
`
`EtgNCHgCHg—O—C
`li^ /
`o
`PROCAINE
`
`NH2
`
`Fig. 7.21 Cocaine and procaine.
`
`side-effects and reduced selectivity. We shall see below how these undesirable prop­
`erties can creep in with simpler molecules and why the opposite tactic of rigidification
`can be just as useful as that of simplification.
`
`7.5.8 Rigidification of the structure
`Rigidification has been a popular tactic used to increase the activity of a drug or to
`reduce its side-effects. In order to understand why, let us consider again our hypo­
`thetical neurotransmitter from Chapter 5 (Fig. 7.22). This is quite a simple molecule
`and highly flexible. Bond rotation can lead to a large number of conformations or
`shapes. However, as seen from the receptor/messenger interaction, conformation I is
`the conformation accepted by the receptor. Other conformations such as II have the
`ionized amino group too far away from the anionic centre to interact efficiently and so
`this is an inactive conformation for our model receptor site. However, it is quite
`possible that there exists a different receptor which is capable of binding conformation
`II. If this is the case, then our model neurotransmitter could switch on two different
`receptors and give two different biological responses.
`The body’s own neurotransmitters are highly flexible molecules (Chapter 5), but
`fortunately the body is quite efficient at releasing them close to their target receptors,
`then quickly inactivating them so that they do not make the journey to other recep­
`tors. However, this is not the case-for drugs. They have to be sturdy enough to travel
`through the body and consequently will interact with all the receptors which are
`prepared to accept them. The more flexible a drug molecule is, the more likely it will
`
`II'' ' '
`i,^
`
`G'
`
`♦li ■
`
`4
`
`s
`
`ft
`
`'M
`
`I?
`
`-
`
`T
`
`I)
`
`I
`
`i
`
`IIIIII
`
`•15
`
`1If:
`
`li
`
`SAWAI EX. 1021
`Page 17 of 31
`
`

`

`ii' ■* n 1
`i'-i .
`•! .iv t
`
`i
`
`•M ■I
`
`!
`
`f ■:
`
`i .
`
`t
`
`i
`
`s
`i /
`i iC “
`i
`; m+f]
`ii
`
`Receptor theories
`
`97
`
`BOND
`ROTATION
`
`H
`
`+N
`
`HjMeA/s -Q^Q
`
`I
`
`H/
`
`RECEPTOR 1
`
`RECEPTOR 2
`
`Fig. 7.22 Two conformations of a neurotransmitter which are capable of binding with different
`receptors.
`
`interact with more than one receptor and produce other biological responses (side-
`effects).
`The strategy of rigidification is to ‘lock’ the drug molecule into a more rigid
`conformation such that it cannot take up these other shapes or conformations.
`Consequently, other receptor interactions and side-effects are eliminated. This same
`strategy should also increase activity since, by locking the drug into the active
`conformation, the drug is ready to fit its target receptor site more readily and does not
`need to ‘find’ the correct conformation. Incorporating the skeleton of a flexible drug
`com-
`into a ring is the usual way of ‘locking’ a conformation and so for our model
`pound the analogue shown in Fig. 7.23 would be suitably rigid.
`The sedative etorphine (Fig. 7.24) was designed by this approach (Chapter 12).
`
`7.6 Receptor theories
`The synthesis of a large number of analogues not only gives compounds with im­
`proved activity and reduced side-effects, but can also give information about the
`protein with which the drugs interact. Clearly, if a drug has an important binding
`
`r(
`
`i
`
`!
`
`I,
`[
`
`1 1i
`
`!■
`
`f
`
`5
`
`SAWAI EX. 1021
`Page 18 of 31
`
`

`

`H-'-
`
`98
`
`Drug deveiopmeut
`
`H
`
`H
`
`HO.
`
`O
`
`+
`NHMe
`
`N—CH3
`
`FLEXIBLE
`MESSENGER
`
`RIGID MESSENGER
`
`MeO
`
`H
`
`Me-
`
`•OH
`CH2CH2CH3
`
`Fig. 7.23 Rigidification of the hypothetical neurotransmitter.
`
`Fig. 7.24 Etorphine.
`
`there must be a complementary binding group present in the binding site of
`group,
`the receptor or enzyme.
`A 3D model of the binding site containing these complementary groups could then
`be built. With such a model, it would be possible to predict whether new analogues
`would have activity. Before the age of computers, this was not an easy task, and the
`best one could do was to build models of the drugs themselves and to match them up
`to see how similar they were. This is clearly unsatisfactory, since it is impossible to
`superimpose one solid object on another.
`The introduction of computer graphics changed all that and revolutionized the field of
`medicinal chemistry such that the goal of rational, scientific drug design is now feasible.
`At the simplest level, the computer can be used to compare drugs and to see how
`similar they are. The st