117
`
`not hydrolyzed at measureable rates, and having a D-amino acid residue in the
`
`penultimate position greatly reduces the reaction rate. Also, compounds of the types
`Pro-X or X-Pro (Pro = proline) are very poor substrates. Compounds which act as
`inhibitors fall into three categories: those which interact with the metal ion (e.g.,
`EDTA, citrate), those which are hydrophobic and compete for the binding site (e. g. ,
`
`large aliphatics, alkanols), and those which do both of the above (e.g., L—leucine, L-
`leucinol).
`Kinetic studies on LAP have been reviewed by Smith and Hill [97] and are not par-
`ticularly illuminating. Some substrates are hydrolyzed according to first—order
`kinetics, while others follow zero—order kinetics. The enzyme is maximally active be-
`tween pH 9.0 and 9.5, depending on the substrate and on the metal ion. In general,
`Vmax for the Mn2 + -activated enzyme is greater than that for the Mgz “‘ -activated en-
`zyme, although the values of Km are comparable [I00].
`
`5.4.3. Prodrug considerations
`Since carboxy and aminopeptidase are digestive enzymes, prodrugs utilizing these en-
`zymes as reconversion sites will be restricted to oral dosage forms. Carboxypeptidase 3
`can be used for the reconversion of prodrugs having a free carboxyl group. Amino acid
`derivatives of aspirin in which the carboxyl group phenylalamine and phenyllactic acid
`ethyl esters of aspirin were targeted for sequential reconversion by a-chymotrypsin
`and carboxypeptidase have been reported [2— 4]. Choice of these esters led to high
`shelf life for aspirin and a reasonable enzymatic rate of regeneration in vitro by the
`combined effect of orchymotrypsin and carboxypeptidase A. Table 10 shows the
`
`TABLE 10
`
`Kinetic Parameters for Carboxypeptidase Hydrolysis at pH 7.5
`
`Substrate
`
`Km
`’
`(mole/liter)
`
`I
`
`'
`
`kcat
`(sec
`
`)
`
`K;
`1 /l’
`(moe iter)
`
`kcat
`F
`"1
`(moles/Second/liter)
`
`Cinnamoyl phenylalanine“
`
`6
`
`X 10”‘
`
`2.1 X 10‘2
`
`Indoleacryloyl-
`phenylalanine“
`
`5.84 x 10“4
`
`1.4 x 10-3
`
`Aspirin phenylalanineb (I)
`
`1.84 X 104
`
`2.8 X l0‘3
`
`~
`
`—
`
`—
`
`35
`
`2.39
`
`15.2
`
`Cinnamoyl phenyllactic acid
`
`1.87 X 10”‘
`
`Furylacryloyl phenyllactic
`acid
`
`1.32 x 10-4
`
`67
`
`47
`
`5.78 X 10‘5
`
`3.58 X 105
`
`~
`
`3.56 x 105
`
`Aspirin phenyllactic
`acid(Il)
`
`a. Taken from Ref. 3.
`b. Performed at pH 8.5.
`
`1
`
`><10”4
`
`25
`
`4
`
`><10“6
`
`2.5 x105
`
`
`
`116
`
`not by the second, there is insufficient data to say which is correct. In either case, the
`peptide bond is broken via acid catalysis by Tyr 248. If an acyl enzyme were formed,
`
`the subsequent attack by H20 would be aided by the proximity of the Tyr 248 con-
`jugate base.
`
`5.4.2. Leucine aminopeptidase
`Although it is thought of as the ‘classic’ N—terminal exopeptidase, there is relatively
`
`little information available on leucine aminopeptidase (LAP). This is due partly to dif-
`ficulty in identifying aminopeptidase obtained from different sources and partly to
`its size. It was originally isolated from porcine kidney [90], but many studies have been
`done on samples extracted from bovine eye lens [91, 92].
`The molecular weight of the enzyme is around 3 X 105 g/mole. The porcine kidney
`sample is thought to consist of four subunits [93], whereas the bovine eye lens sample
`consists of 10 subunits [94]. Low—angle ‘X-ray diffraction studies done on the eye lens
`sample indicate that the enzyme is shaped like a hollow cylinder with inside diameter
`of about 32 A [95]. No efforts to investigate the detailed tertiary structure have been
`made and, since efforts to determine end groups have been unsuccessful, the primary
`and secondary structures are completely unknown. However, Himmelhoch [96] has
`found that the enzyme contains approximately one atom of Zn per subunit.
`Unlike CPA, the tertiary structure of LAP is stabilized by the Zn ions: once the Zn
`has been removed, catalytic activity is lost permanently. However, the Zn can be
`replaced with other ions [97]. LAP is strongly activated by Mn“ and Mg“ , in-
`hibited by Cdz +, Cu2 + , Hgz + and Pb2 +, and unaffected by Caz + , Co“ and Ni2 + .
`The Zn (or other metal ion) has been shown to play an essential role in the catalytic
`process but to be unessential for substrate binding [98].
`
`R1
`R
`I
`I
`(9
`H3N—-C—C——N—C —c -
`H
`ll [H H
`ll
`0
`0
`
`Fig. 12. LAP catalyzes the hydrolysis of the indicated peptide bond.
`
`The primary action of LAP is indicated in Figure 12. It can also function as an
`esterase but at about 10% of the rate for the analogous peptide [97]. There are two
`absolute requirements for potential LAP substrates [97]: the N—terminal residue must
`be of the L configuration and the terminal -NH2 groups must be free (i.e., not
`acylated). In addition to peptides and esters, amides have long been known to be good
`substrates for LAP (e.g., L—leucine-amide [99]). Almost any N—terminal L-amino acid
`(or glycine) will be released at a measurable rate, although substrates with a
`hydrophobic R—group are best. Generally, substrates wherein the N—terminal residue
`is Leu, norLeu or norVal are hydrolyzed the fastest, although L—Ala-L-leucinamide is
`the best substrate known. Compounds having an N-terminal D-amino acid residue are
`
`
`_ Exhibit 1012 _ page 55
`Petitioner Amerigen Pharmaceuticals Ltd.
`
`
`Petitioner Mylan Pharmaceuticals Inc. - Exhibit 1012 - Page 65
`
`

`
`
`
`
`
`118
`
`kinetic parameters for carboxypeptidase A hydrolysis of aspirin phenyla1amine(I) and
`aspirin phenyllactic acid (11).
`
`E?
`
`ICH2
`
`E33
`
`CH21
`
`CONH-CH—CO0H
`S ‘OCOCH,
`
`: :coo-cH_cooH
`OCOCHJ
`
`1
`
`II
`
`Isolation and identification of carboxy and aminopeptidase from enterocyte brush
`
`borders and cytosol have shown different though somewhat overlapping specificities.
`It appears that these exopeptidases function at the surface or interior of the absorptive
`cell to further reduce small oligopeptides released by pancreatic protease action. Their
`
`radial distributions appear to be geared to couple small peptide and amino acid in-
`testinal transport. Tetrapeptide activity is associated exclusively with the brush border
`membrane, tripeptide activity is distributed about equally between membrane and
`cytosol, while enzymes that cleave dipeptides are more prevalent in the cytosol than
`brush border [101]. Axial distributions appear to be fairly even over the entire small
`intestinal length and less subject to species, age, and nutritional input parameters than
`is alkaline phosphatase. However, this may be attributed to the fact that isoenzymes
`
`and specific peptidases have not been well characterized with respect to their associa-
`tion with particular axial
`regions
`[l02].
`Intestinal absorption of L-lysine—p—
`nitroanilide, L-alanine-p-nitroanilide and glycine-p-nitroanilide was studied in the
`
`presence of competitive inhibitors in perfused rat intestine in order to establish the
`coupled brush border hydrolysis and transport [103]. This study has shown that the
`peptidase in the brush border region that serves as the hydrolysis site requires a free
`oz-amino group (an aminopeptidase).
`
`6. Alkaline phosphatases
`
`The mammalian alkaline phosphatases are distributed among various tissues of the
`body. Examples of tissue exhibiting alkaline phosphatase activity are: the intestinal
`mucosa, placenta, kidney, bone, liver, lung and spleen [104]. Alkaline phosphatases
`may not be as readily useful as targeted reconversion site for prodrugs as the hydrolytic
`enzymes of the GI tract. Fishman [105] has investigated the organ—specific behavior
`
`of rat alkaline phosphatases toward/a variety of compounds.
`A typical alkaline phosphatase (e.g., human placental alkaline phosphatase) has a
`molecular weight around 125,000 g/ mole and contains approximately 3 moles of
`
`
`
`119
`
`Zn“ per mole of enzyme [I06]. Limited amino acid sequencing data around the ac-
`tive site has been reported [107] and the sequence ASP-‘active Ser’-Ala has been
`observed for the active site.
`
`Alkaline phosphatases catalyze the hydrolysis of most phosphate esters. Harkness
`[108] has shown that human placental phosphatase hydrolyzes a variety of substrates
`of comparable rates. Inorganic phosphate [110] and arsenate [l08, 109] anions have
`been found to be potent competitive inhibitors. Cysteine and histidine have been
`found to act as non-competitive inhibitors by chelating the Zn2+ ion essential for
`catalytic activity [111]. L-Phenylalanine has been shown to be an uncompetitive in-
`hibitor [112, 1 13]. Kinetic studies have indicated that (1) as the substrate concentration
`is increased, the optimum pH increases from about pH 8 to about pH 10, and (2)
`substrate inhibition is significant. The data [114] in Figure 13 illustrate both these
`observations.
`
`Alkaline phosphatase activity is not distributed evenly in the axial direction of the
`vertebrate gastrointestinal tract. In many species, though not in all, enzymatic activity
`begins abruptly at the pylorus and diminishes gradually from the duodenum tothe il-
`eum. This is found in the dog, mouse and adult rat. In rats, for example, the ratio of
`duodenal to jeujunal to ileal alkaline phosphatase (per mg protein) is 16:4:l [1 15]. In
`humans, high ileal activity has been reported [116, 117]. In several species, including
`mouse and rat, axial activity distribution varies with growth and development [118,
`
`119]. For neonates of these species, alkaline phosphatase activity is distributed more
`evenly along the small intestinal length. Fasting, fat ingestion, vitamin D deficiency,
`and dietary zinc have been shown to correlate with increases in intestinal alkaline
`
`1600
`
`1400
`
`1200
`
`1000
`
`200
`
`
`
`InitialVelocity
`
`800
`
`600
`
`400
`
`7.5
`
`8
`
`8.5
`
`9
`
`9.5
`
`10
`
`10.5
`
`11
`
`Fig. 13. Hydrolysis of phenyl phosphate by calf intestinal alkaline phosphatase. The curves refer to the
`following substrate concentrations: A, 25 ;tM; B, 50 pLM; C, 100 p.M; D, 500 MM; E, 750 ;tM; F, 2.5 mM;
`G, 25 mM; and H, 75 mM. Initial velocities are expressed as }/.II10l product per mg enzyme per minute.
`
`Petitioner Amerigen Pharmaceuticals Ltd.
`_ Exhibit 1012 _ page 66
`
`
`Petitioner Mylan Pharmaceuticals Inc. - Exhibit 1012 - Page 66
`
`

`
`
`
`121
`
`1000
`
`900
`
`800
`
`700
`600
`
`500
`
`400
`
`300
`
`Phosphate
`
`Free Alcohol
`
`200
`
`
`
`120
`
`phosphatase activity, while calcium, sucrose and alcohol intake have been correlated
`with reduced activities [120, 121].
`There is also considerable evidence that intestinal alkaline phosphatases exist as
`several isoenzymes [122]. In this regard, it has been shown that the ileal enzyme is
`precipitated less easily than duodenal enzyme by mixed anti—phosphatase, L-
`phenylalanine inhibition is more marked for duodenal than jejunal enzyme, heat
`stability is greatest for duodenal alkaline phosphatase, and the phenylphosphate to
`beta~glycerophosphate activity ratio is reversed from duodenum (high ratio) to ileum.
`It is believed this enzyme functions in the small intestine for absorption of phosphate
`from dietary phosphates, but much uncertainty remains concerning its physiologic
`function.
`
`
`
`
`
`
`
`
`
`
`
`PlasmaConc.(ug/ml)
`
`
`
`
`
`
`
`
`
`
`Petitioner Amerigen Pharmaceuticals Ltd.
`
`_ Exhibit 1012 _ page 67
`
`0
`
`15
`
`30
`
`45
`
`60
`
`75
`
`90 105120135 150165180 195 210 225 240
`
`Time(Min)
`
`Fig. 14. Pooled dog plasma concentration after a 0.15 mM oral dose of the prodrugs.
`
`7. Carboxyl esterases and lipases
`
`7.1. CARBOXYL ESTERASES
`
`Carboxyl esterases are distributed widely in vertebrate tissues and blood serum, the
`
`highest mammalian activities being found in the liver, kidney, duodenum and brain
`[I24]. The physiological functions of most carboxyl esterases are not clearly
`understood. Myers and Mendel [125] were among the earliest group of scientists to
`suggest that these enzymes do not themselves participate in normal metabolic pro-
`cesses. Krisch [124], observing that carboxyl esterases hydrolyze a wide variety of
`compounds not normally found in the body, suggested that these enzymes play an im-
`portant role in the body’s detoxification system. Carboxyl esterases have been
`classified-[126] according to their behavior toward organophosphorus compounds
`such as diethyl-p—nitrophenyl phosphate. A-esterases hydrolyze organophosphorus
`compounds as substrates whereas B-esterases do not hydrolyze and are inhibited by
`such compounds. A third type, C-esterase, has been found in porcine kidney extracts,
`which neither hydrolyzes nor is inhibited by organophosphorus compounds.
`There is very little information available concerning the structure of the carbox-
`
`ylesterases, which exist in many forms. As many as 13 electrophoretically distinct
`Zieelrases-hay:bbeeridfound in the rat liver alone [I27]. Studies on esterase isozymes have
`have tr)::1:\;:undyt either and Skillen [l28]. Carboxylesterases from various sources
`b
`o ave mo ecu ar weights around 160,000 g/mol [I24]. It has also
`een shown that the equivalent weight of several esterases isapproximately half of the
`molecular weight [129 —— 132]; the dissociation of pig and ox liver esterases into active
`
`
`
`
`
`
`
`
`6.1.
`
`PRODRUG CONSIDERATIONS
`
`
`
`
`
`
`Since phosphatase activity is distributed widely in mammalian tissues, phosphate
`monoester prodrugs are not restricted to oral dosage forms. However, for planned
`and specific targeting using phosphate monoesters, alkaline phosphatase in the in-
`testine is the enzyme of choice. Since, in general, P—N bonds are not hydrolyzed, a free
`hydroxyl group in the drug moiety would be required. Prodrugs of the form
`
`‘.9
`D——O—l|°~OH
`OH
`
`
`
`
`
`
`
`would be potential phosphatase substrates.
`The concept of alkaline phosphatase hydrolysis coupled with facilitated transport
`has been studied recently [123] using the 21-phosphate ester of hydrocortisone.
`Bioavailability experiments in beagle dogs were done along with the parent drug, suc-
`cinate and lysinate esters in equivalent molar doses. The phosphate derivative, which
`was shown previously to have high rat intestinal wall permeability in the upper in-
`
`
`testine, showed high early peak plasma concentration which dropped off rapidly,
`
`
`leading, however,
`to a somewhat reduced bioavailability for this prodrug. The
`
`
`bioavailability results are shown in Figure 14. The fact that intestinal alkaline
`
`
`phosphatase distribution in dog is predominantly in the upper small intestine and that
`
`
`the ionized prodrug is not well absorbed passively accounts for the data.
`
`
`These results indicate that prodrugs targeted for intestinal alkaline phosphatase
`
`
`potentially can improve oral drug delivery. For stable phosphate derivatives, axial en-
`
`
`zyme distribution in the particular species under consideration will determine the
`
`
`relative success for these prodrugs.
`
`
`
`
`
`Petitioner Mylan Pharmaceuticals Inc. - Exhibit 1012 - Page 67
`
`

`
`
`
`
`
`TABLE 11
`Examples of Esterase Substrates
`
`l23
`
`1,
`
`Carboxyl esters
`(a)
`Esters of unsubstituted fatty acids
`Phenyl formate (acetate, propionate, butyrate, valerate)
`Ethyl formate (acetate, propionate, valerate)
`m~Carboxyphenyl esters of a homologous series of r1—fatty acids (chain length from C2 to
`C12)
`Ethyl acetate
`Glyceryl triacetate (triacetin)
`p—Nitrophenyl acetate
`0-Nitrophenyl acetate
`2,6-Dichlorobenzenone-indophenyl acetate
`Vitamin A acetate
`Methyl butyrate
`Methyl butyrate (3—methylbutyrate, pentanoate, 3- and 4-methylpentanoate, hexanoate, hep-
`tanoate)
`Glyceryl tributyrate (tributyrin)
`Ethyl butyrate
`0-, m-, and p-Nitrophenyl butyrate
`2,4-Dinitrophenyl butyrate
`m—(n—Pentanoyloxy) benzoic acid
`m-(n-Heptanoyloxy) benzoic acid
`
`(b) Esters of other acids
`fumarate,
`succinate,
`acetoacetate, diethyl
`lactate,
`Ethyl benzoate (benzenesulfonate,
`asparate, p-hydroxybenzoate, bromomalonate, terephthalate, and other ethyl esters)
`Procaine (2—diethylaminoethyl p—aminobenzoate)
`L—Tyrosine ethyl ester (and many other amino acid esters)
`
`2.
`
`Thioesters
`
`6-S- and 8-S-acetoacetyl monothioloctanoate
`8-S—Acetoacetyl, 6-ethyl monothioloctanoate
`8-S—Acetoacetyl dihydrolipoic acid
`6-S- and 8-S-Acetyl dihydrolipoic acid
`6—S—Acetoacetyldecanoate
`S—Acetyl- and S—acetoacetyl—BAL
`p—Nitrothiophenyl hippurate
`Thiophenyl acetate
`
`3.
`
`Aromatic amides
`Acetanilide
`Phenacetin
`
`Petitioner Amerigen Pharmaceuticals Ltd.
`
`- Exhibit 1012 - page 58
`
`
`
`
`
`122
`
`
`
`
`‘half—molecules’ has been observed in several laboratories (see, Refs. 132, 133). Krish
`[124] has reviewed studies of the amino acid composition of various esterases. The B-
`esterases, which have received far more attention than the other two classes, are
`phosphorylated by inhibitors such as diethyl p—nitrophenyl phosphate at a serine
`residue. This fact has enabled the determination of the amino acid sequence in the
`vicinity of the ‘active serine’ for a number of different carboxylesterases [124].
`Esterases catalyze the hydrolysis of esters according to the reaction:
`
`
`
`0
`//
`:
`ll
`R1—C:O—R2+ H,O‘¢— R;-C\
`
`0
`
`+
`OH
`
`R2—0H
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Table 11 lists a few of the many compounds known to be esterase substrates.
`Although some labile ester substrates are also hydrolyzed by several endopeptidases
`(e.g., wchymotrypsin) and other enzymes, these catalytic rates are, typically, 4 to 5
`orders of magnitude less than those of ‘true’ esterases [138, 139, 145]. For example,
`using p-nitrophenyl butyrate as the substrate, Stoops et al. [138] reported that, at pH
`8.l,Km = 2.3 X l0‘4M and kcat = 3.72 X 102 sec” for pig liver carboxylesterase.
`For the same substrate under similar conditions, Cane and Wetlaufer [146] found km
`= 7.7 X 10‘3 sec“1 for oc—chymotrypsin. Also,
`like wchymotrypsin,
`liver and
`kidney esterases have been shown to catalyze the hydrolysis of several amino acid
`esters [l32, 134, 138] at comparable rates. For the carboxylesterase- and ot-
`chymotrypsin~catalyzed hydrolysis of L-tyrosine ethyl ester, Stoops et al. [138] found
`kcat values of 71 and 39 sec‘ 1, respectively. However, in contrast with the peptidases,
`peptides are not hydrolyzed by porcine liver carboxylesterase [134], although some
`amides have been found to be poor substrates. With Km = 2.5 X 10‘3 M and kcat
`= 0.1 sec‘ 1, phenacetin was typical of the several amide substrates studied by Franz
`and Krisch [131].
`They also noted that procaine was an equally poor ester substrate of porcine
`kidney carboxylesterase, having Km and km values of 4 X 10-4 and 0.1 sec"‘,
`respectively. Levy and Ocken [135] have distinguished three groups of carboxyl ester
`substrates: (1) unsubstituted monocarboxylate esters (which are hydrolyzed with the
`highest relative velocities), (2) substituted monocarboxylate esters, and (3) dicarbox—
`ylate diesters and substituted diesters, of which only one ester group is hydrolyzed.
`Charged compounds are extremely poor substrates [134, 135]. Hofstee [l36] has
`found that the reactivity of fatty acid esters of m-hydroxybenzoic acid increases as
`the acyl chain length is increased. Chain length effects have been discussed in more
`Monoethylglycine 2,6-xylidide
`detail by Dixon and Webb [l37], who reported the reactivity and affinity (defined
`Diethylglycine 2,6-xylidide (Xylocaine; lidocaine)
`as the inverse of Km) of 33 esters (in which both the acyl and alkyl chain lengths
`N—(n—Butylamino) acetyl 2-chloro, 6-methylanilide - I-lCl
`ranged between C1 and C9) with respect to horse liver carboxylesterase. They con-
`(hostacaine: butacetoloid)
`L-Leucyl p—nitroanilide
`cluded that the alkyl and acyl chain length effects are largely independent. Both the
`L—Leucyl beta—naphthylamide
`affinity and reactivity increase as either chain is lengthened to about C4 to C6. Fur-
`2—(N4-n-Propylaminoacetyl)-sulfanilamido 4,6—dimethylpyridine HCl
`
`8. Taken from the review by Krisch [I24]. See that article for references.
`
`
`
`
`Petitioner Mylan Pharmaceuticals Inc. - Exhibit 1012 - Page 68
`
`

`
`
`
`124
`
`ther increase in the alkyl chain length results in a decrease in both the affinity and
`reactivity, suggesting that the alkyl binding site is only large enough to accomodate
`a butyl to hexyl chain. Further increase of the acyl chain length produces the same
`decline in reactivity but a sharp rise in affinity, suggesting the existence of a secon-
`dary acyl binding site which leads to non—optimal orientation at the esteratic site.
`Similarly, branched—chain substrates exhibit higher affinities but lower reactivities
`[137]. Esterases are also capable of hydrolyzing thioesters [121, 122] and aromatic
`amides, including acetanilide [l29, 132, 134] and phenacetin, at reduced rates. Last-
`ly, carboxylesterases can also act as transferases as was first observed by Bergmann
`and Wurzel [140] in 1953.
`Kinetic studies on mammalian carboxylesterases consistently report the pH op-
`timum to be in the range of 7.5—9.0 [124]. It is generally agreed that, like other
`serine hydrolases (e.g., a—CT), carboxylesterase catalysis proceeds through an acyl-
`enzyme intermediate [141]. Deviations from normal Michaelis—Menten kinetics at-
`tributable to substrate activation have been observed by several workers (see, e.g.,
`
`Refs. 138 —— 142). The details of the catalytic mechanism are still unknown, although
`the participation of a histidine residue has been discussed [137, 144, 145] but not
`verified.
`
`7.2.
`
`LIPASES
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`125
`
`molecules must be aggregated is still unknown. This lipase—esterase distinction is
`clearly illustrated in Figure 15, taken from the work of Sarda and Desnuelle [158].
`Whereas the esterase activity reached a plateau as the solution became saturated
`
`with triacetin, lipase activity on the same substrate was negligible until the solubility
`limit was passed and emulsified particles were formed. Agents which promote the
`micellization or emulsification of substrate molecules, e.g., NaCl [154] and bile salts
`
`[147], also produced the expected increase in lipase activity. Bile salts also promote
`in vivo lipolysis by forming mixed micelles with the lipolysis products (i.e., fatty
`acids, soaps) and removing them from the hydrophobic interface where the lipase
`acts [147]. Lastly, bile salts are known to shift the optimum pH of lipases from
`about 8.5 to 6, which is much closer’ to the pH of the upper intestine [I59].
`Kinetic studies on lipases are complicated by the requirement that the substrate
`molecules must be aggregated. Any factor which affects the affinity of the enzyme
`for the resulting interface or the packing and orientation of the substrate molecules
`at that interface will also affect the observed lipase activity. Also,
`the rate of
`
`lipolysis is not dependent simply on the concentration of substrate molecules but,
`rather, on the available surface area of aggregated substrate. Assuming a constant
`particle-size distribution, Sarda and Desnuelle [158] prepared triglyceride emulsions
`of various ‘concentrations’ and, using arbitrary units, demonstrated that lipolysis
`
`proceeds according to Michaelis—Menten-like kinetics.
`p—Nitrophenyl phosphate incorporated into bile salt micelles causes almost con-
`plete inactivation of pancreatic lipase [I60]. As with other esterases,
`the
`organophosphate has been shown to bind to a serine residue [161] undoubtedly in-
`
`Enzyme
`
`Units
`
`X = Lipase
`o = Esterase
`
`O
`
`.5
`
`1
`
`1.5
`Saturation
`
`2
`
`2.5
`
`3
`
`3.5
`
`Fig. 15. Hydrolysis of triacetin by a lipase and an esterase. Increasing amounts of triacetin were added
`to a fixed volume of buffer and the resulting mixtures were incubated with pure lipase (X) or purified
`horse liver esterase (0). Beyond saturation of 1.0 the solution is oversaturated and, consequently,
`emulsified particles begin to form.
`
`Petitioner Amerigen Pharmaceuticals Ltd.
`
`- Exhibit 1012 - page 59
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`of dissociated (i.e., totally solvated)/substrate molecules. Lipases generally act at
`
`hydrophobic interfaces or on emulsified particles. However, some lipases are known
`to hydrolyze micelles [156, 157] and the minimum extent
`to which substrate
`
`
`The lipases are a distinct subset of the large family of esterase enzymes. In mam-
`mals, lipases are found in the digestive tract, in such tissues as the heart, brain, mus-
`cle, arteries, kidney, adipose tissue and serum, and have also been identified in milk
`[147]. Their primary function is the metabolism of triglycerides.
`Most of the lipase found in the GI tract is produced in the pancreas. The lipase
`isolated from lyophilized aqueous extracts of fresh porcine pancreas was found to
`be almost entirely in the form of a high-molecular—weight complex of enzyme and
`lysolecithin-rich micelles [I48]. Treatment of this complex with organic solvents
`produced enzyme molecules weighing approximately 48,000 g/ mole [149]. The
`amino acid compositions of several lipases are known [149, 150], but there is no in-
`formation on their structure. The observation that the lipase activity of porcine pan-
`creatic extracts is significantly depressed upon chromatography [151] or gel filtra-
`tion [152] pointed to the existence of a pancreatic colipase, which subsequently was
`isolated and purified [153]. It was found to be a small protein (molecular weight,
`about 8,000 g/mole) which binds to the lipase in a 1:1 molar ratio [153] and in-
`creases the lipase activity when bile salts are present [l54, 155].
`Lipases are best distinguished from the other carboxylesterases by the physical
`state of the substrate. Unlike other esterases, lipases do not catalyze the hydrolysis
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Petitioner Mylan Pharmaceuticals Inc. - Exhibit 1012 - Page 69
`
`

`
`
`
`more detailed biochemical characterization of these enzyme classes is needed before
`
`more specific suggestions can be made.
`
`127
`
`8. Microfloral enzymes in the colon
`
`The glycosidase activity of the colonic microflora offers opportunity of designing
`a colon-specific drug delivery system. Co1on—specific delivery of bioactive com-
`pounds is known to occur in man. In the plant kingdom, a great many compounds
`are found as glycosides. Upon ingestion, many of these glycosides pass through the
`upper intestine and into the colon. Once there, the glycosidases of the colonic
`microflora liberate aglycones, which may then be absorbed. Certain sulfa drugs
`
`[169] are now known to be activated by the azo—reductase activity of the colonic
`microflora. A prodrug system based on polymers has been reported [170, 171] in
`which certain aromatic amines are released by reduction of an azo link between the
`
`drug and the polymeric carrier.
`Based on this concept, a ‘colon-specific’ prodrug design has been reported recent-
`
`ly [I72]. Dexamethasone—2l—beta-D-glucoside and prednisolone-21-beta-D-glucoside
`were synthesized for possible treatment of inflammatory bowel disease. Hydrolysis
`of the prodrugs by beta-glucosidase and fecal homogenates in vitro caused the
`release of the free steroids. Nearly 60% of an oral dose of glucoside of dex-
`amethasone was shown to have reached the cecum, whereas less than 15% of
`
`glucoside of prednisolone did. When free steroids were administered orally, they
`were absorbed almost exclusively in the small intestine.
`The rat model suffers from the problem of a relatively high bacterial population
`and subsequently high level of glycosidase activity present in its stomach, upper
`small intestine and lower small intestine. There are an average of 107-7, 105-9 and
`107-7 microorganisms/g wet weight in the rat stomach, upper small intestine and
`lower small intestine, respectively. In contrast, the bacterial population in man’s
`stomach and small intestine is much lower. There are only an average of 100, 102-5
`and 104-2 microorganisms/ g wet weight residing in the human stomach, upper
`small intestine and lower small intestine, respectively [173, 177]. Bacterial popula-
`tion of the large intestines of rat and man are more nearly comparable (= 103-3
`micro0rganisms/g wet weight)
`[I73]. Bacteroides and Bifidobacteria are the
`bacterial species comprising the majority of microorganisms in the gasterointestinal
`system of both the laboratory rat and man. Both species have been shown to pro-
`duce measurable quantities of beta-glucosidase in vivo [173].
`Modification of the enzymatic activity of the gut microflora to improve delivery
`is a significant feature in this system. It has been shown that certain enzymes pro-
`duced by gut bacteria are inducible with diet [l74, 175]. For instance, bean diets
`dramatically increase or-galactosidase activity in human subjects while bran diets in-
`crease [3-glucosidase activity. Manipulationof glycosidase activity by diet may be
`
`Petitioner Amerigen Pharmaceuticals Ltd.
`
`- Exhibit 1012 — Page 70
`
`
`
`126
`
`volved in the catalytic mechanism. There is also considerable evidence for the par-
`ticipation of a histidine residue in the catalytic process [162], but the detailed
`mechanism is quite unknown.
`Lipases can catalyze the hydrolysis of a variety of esters, although, owing to their
`tendency toward self-association,
`triglycerides are preferred. Only the 1- and
`3-positions of triglycerides are attacked by lipase. However, di- and monoglycerides
`undergo rapid isomerization (especially at slightly acidic pH) whereby a chain at-
`tached at the 2-position migrates to a primary (l— or 3-position) carbon [I63]. Thus,
`1,3-di- and 1-monoglycerides are found both in vitro and in vivo; in vitro lipolysis
`of triglycerides sometimes goes to completion. In contrast with the above positional
`specificity, lipases do not recognize [l64, 165] the ‘biological asymmetry’ ascribed
`to glycerol derivatives in other instances [166, 167]. The effect of acyl chain length
`on substrate reactivity has been the subject of many investigations, but it is now
`known to be less important [147] than the length of the alcohol chain. In a study
`of over 100 esters in which both the acyl and alcohol chain lengths ranged between
`C2 and C18, Mattson and Volpenhein [168] concluded that the rate of hydrolysis is
`influenced independently by the acyl and alcohol chain lengths. Esters of heptyl
`alcohol were found to be hydrolyzed faster than esters of other alcohols, and esters
`of dodecanoic (followed by butyric) acid were hydrolyzed faster than esters of other
`fatty acids. They ascribed the influence of the acyl chain length to typical enzyme-
`substrate specificity, whereas the influence of the alcohol chain length was at-
`tributed to orientiation of the substrate molecules at the substrate/water interface.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`7.3.
`
`
`
`PRODRUG CONSIDERATIONS
`
`
`Since the carboxylesterases and lipases are distributed widely in mammalian tissues,
`prodrugs designed for reconversion by these enzymes are not restricted to oral
`
`dosage forms. In view of the broad specificity and heterogeneity of these enzyme
`
`classes, many esters are likely to be at least poor substrates. In general, amides are
`much poorer substrates than the corresponding esters. Prodrug suggestions would
`include C4 — C6 esters of drugs possessing a free hydroxyl group and simple C4 ~ C6
`esters of drugs with free carboxyl groups. Glyceride esters would be potential lipase
`substrates, as would esters that are largely hydrophobic and present, predominantly,
`at the interface of emulsified particles. Suggestions would include dodecanoate and
`butyrate esters of drugs with free hydroxyl groups and heptyl esters of drugs with
`free carboxyl groups. The alcoholic, as opposed to the acyl, portion of the ester ap-
`pears to be the more important in determining lipase activity. However, since proper
`orientation of the substrate at the interface is also important for lipase activity, the
`drug portion of the prodrug may significantly affect lipase activity as well as deter-
`mining whether the lipases or esterases would be the primary reconversion site.
`Finally, it should be noted that many prodrugs currently available (steroid esters,
`for example) are probably substrates for the lipases and/or esterases. Clearly, a
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Petitioner Mylan Pharmaceuticals Inc. - Exhibit 1012 - Page 70
`
`

`
`128
`
`
`129
`
`1.‘
`1.’
`1.‘
`1?
`respectively. These en-
`D - C - NH - CI-I — COO “ and D - X - C - CH - NH;
`zymes preferentially hydrolyze prodrugs containing nonpolar R group. The physical
`properties of drugs with free —OH, -NH2 or -COOH groups may be altered to
`almost any desired direction by appropriate choice of the R group.
`The carboxyl esterases and the lipases exhibit a broad specificity towards esters.
`The prodrug must be present at a micellar or emulsified particle interface, with pro-
`per orientation in order for lipase to be active. Colipase increases the lipase activity
`in the presence of bile salts. C4-C6 esters of drugs possessing -OH or -COOH
`group may be suggested for using these enzymes as reconversion sites. The
`O
`phosphatases require prodrugs of the form D- O-1”’-(OH)2 and, consequ