BIOPHARMACEUTICS & DRUG DISPOSITION
`Biopharm. Drug Dispos. 24: 259–273 (2003)
`Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/bdd.364
`
`Presystemic Metabolism and Intestinal Absorption of
`Antipsoriatic Fumaric Acid Esters
`
`D. Werdenberga, R. Joshib, S. Wolfframc, H.P. Merklea and P. Langguthd,*
`a Department of Applied BioSciences, Swiss Federal Institute of Technology Zurich, Winterthurerstraße 190, CH-8057 Zurich, Switzerland
`b Fumapharm AG, CH-5630 Muri, Switzerland
`c Institute of Animal Nutrition, Physiology and Metabolism, Christian-Albrechts-University of Kiel, D-24098 Kiel, Germany
`d Department of Biopharmaceutics and Pharmaceutical Technology, Johannes Gutenberg- University, D-55099 Mainz, Germany
`
`ABSTRACT: Psoriasis is a chronic inflammatory skin disease. Its treatment is based on the
`inhibition of proliferation of epidermal cells and interference in the inflammatory process. A new
`systemic antipsoriasis drug, which consists of dimethylfumarate and ethylhydrogenfumarate in the
`form of their calcium, magnesium and zinc salts has been introduced in Europe with successful
`results. In the present study, a homologous series of mono- and diesters of fumaric acid has been
`studied with respect to the sites and kinetics of presystemic ester degradation using pancreas
`extract, intestinal perfusate, intestinal homogenate and liver S9 fraction. In addition, intestinal
`permeability has been determined using isolated intestinal mucosa as well as Caco-2 cell
`monolayers, in order to obtain estimates of the fraction of the dose absorbed for these compounds.
`Relationships between the physicochemical properties of the fumaric acid esters and their biological
`responses were investigated. The uncharged diester dimethylfumarate displayed a high
`presystemic metabolic lability in all metabolism models. It also showed the highest permeability
`in the Caco-2 cell model. However, in permeation experiments with intestinal mucosa in Ussing-
`type chambers, no undegraded DMF was found on the receiver side,
`indicating complete
`metabolism in the intestinal tissue. The intestinal permeability of the monoesters methyl hydrogen
`fumarate, ethyl hydrogen fumarate, n-propylhydrogen fumarate and n-pentyl hydrogen fumarate
`increased with an increase in their lipophilicity, however, their presystemic metabolism rates
`likewise increased with increasing ester chain length. It is concluded that for fumarates, an increase
`in intestinal permeability of the more lipophilic derivatives is counterbalanced by an increase in
`first-pass extraction. Copyright # 2003 John Wiley & Sons, Ltd.
`
`Key words: first-pass extraction; prodrug; permeability; degradation; bioavailability
`
`Introduction
`
`The antipsoriatic fumaric acid therapy is based
`on the peroral administration of dimethyl fuma-
`rate and various salts of ethyl hydrogen fumarate
`[1]. The diester dimethyl fumarate turned out to
`be more effective than the monoester ethyl
`hydrogen fumarate, although it was shown that
`
`*Correspondence to: Department of Biopharmaceutics and Phar-
`maceutical Technology, Johannes Gutenberg- University, D-55099
`Mainz, Germany. E-mail: langguth@mail.uni-mainz.de
`
`dimethyl fumarate is rapidly cleaved to methyl
`hydrogen fumarate by hydrolysis in the circula-
`tion [2, 3]. Successful absorption of esters into the
`systemic circulation not only requires a suffi-
`ciently high permeability of the active species
`across the intestinal mucosa but also stability
`against intestinal and hepatic hydrolysis.
`the
`The presystemic metabolic barrier of
`intestine is responsible for the rapid disappear-
`ance of the ester from the absorption site due to
`cleavage of the ester bond by esterases associated
`with the intestinal
`lumen and the mucosa.
`
`Copyright # 2003 John Wiley & Sons, Ltd.
`
`Page 1 of 16
`
`

`
`BIOPHARMACEUTICS & DRUG DISPOSITION
`Biopharm. Drug Dispos. 24: 259–273 (2003)
`Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/bdd.364
`
`Presystemic Metabolism and Intestinal Absorption of
`Antipsoriatic Fumaric Acid Esters
`
`D. Werdenberga, R. Joshib, S. Wolfframc, H.P. Merklea and P. Langguthd,*
`a Department of Applied BioSciences, Swiss Federal Institute of Technology Zurich, Winterthurerstraße 190, CH-8057 Zurich, Switzerland
`b Fumapharm AG, CH-5630 Muri, Switzerland
`c Institute of Animal Nutrition, Physiology and Metabolism, Christian-Albrechts-University of Kiel, D-24098 Kiel, Germany
`d Department of Biopharmaceutics and Pharmaceutical Technology, Johannes Gutenberg- University, D-55099 Mainz, Germany
`
`ABSTRACT: Psoriasis is a chronic inflammatory skin disease. Its treatment is based on the
`inhibition of proliferation of epidermal cells and interference in the inflammatory process. A new
`systemic antipsoriasis drug, which consists of dimethylfumarate and ethylhydrogenfumarate in the
`form of their calcium, magnesium and zinc salts has been introduced in Europe with successful
`results. In the present study, a homologous series of mono- and diesters of fumaric acid has been
`studied with respect to the sites and kinetics of presystemic ester degradation using pancreas
`extract, intestinal perfusate, intestinal homogenate and liver S9 fraction. In addition, intestinal
`permeability has been determined using isolated intestinal mucosa as well as Caco-2 cell
`monolayers, in order to obtain estimates of the fraction of the dose absorbed for these compounds.
`Relationships between the physicochemical properties of the fumaric acid esters and their biological
`responses were investigated. The uncharged diester dimethylfumarate displayed a high
`presystemic metabolic lability in all metabolism models. It also showed the highest permeability
`in the Caco-2 cell model. However, in permeation experiments with intestinal mucosa in Ussing-
`type chambers, no undegraded DMF was found on the receiver side,
`indicating complete
`metabolism in the intestinal tissue. The intestinal permeability of the monoesters methyl hydrogen
`fumarate, ethyl hydrogen fumarate, n-propylhydrogen fumarate and n-pentyl hydrogen fumarate
`increased with an increase in their lipophilicity, however, their presystemic metabolism rates
`likewise increased with increasing ester chain length. It is concluded that for fumarates, an increase
`in intestinal permeability of the more lipophilic derivatives is counterbalanced by an increase in
`first-pass extraction. Copyright # 2003 John Wiley & Sons, Ltd.
`
`Key words: first-pass extraction; prodrug; permeability; degradation; bioavailability
`
`Introduction
`
`The antipsoriatic fumaric acid therapy is based
`on the peroral administration of dimethyl fuma-
`rate and various salts of ethyl hydrogen fumarate
`[1]. The diester dimethyl fumarate turned out to
`be more effective than the monoester ethyl
`hydrogen fumarate, although it was shown that
`
`*Correspondence to: Department of Biopharmaceutics and Phar-
`maceutical Technology, Johannes Gutenberg- University, D-55099
`Mainz, Germany. E-mail: langguth@mail.uni-mainz.de
`
`dimethyl fumarate is rapidly cleaved to methyl
`hydrogen fumarate by hydrolysis in the circula-
`tion [2, 3]. Successful absorption of esters into the
`systemic circulation not only requires a suffi-
`ciently high permeability of the active species
`across the intestinal mucosa but also stability
`against intestinal and hepatic hydrolysis.
`the
`The presystemic metabolic barrier of
`intestine is responsible for the rapid disappear-
`ance of the ester from the absorption site due to
`cleavage of the ester bond by esterases associated
`with the intestinal
`lumen and the mucosa.
`
`Copyright # 2003 John Wiley & Sons, Ltd.
`
`Received 7 November 2001
`Revised 22 January 2003
`Accepted 14 March 2003
`
`Page 2 of 16
`
`

`
`260
`
`D. WERDENBERG ET AL.
`
`Considerable esterase activity is also localized in
`the hepatic system [4]. There are, however,
`indications that in the intestine dimethyl fuma-
`rate is completely biotransformed to its metabo-
`lite, methyl hydrogen fumarate, before reaching
`the liver [3].
`In the present work, the metabolic activity of
`the intestinal mucosa as a barrier
`for
`the
`absorption of various fumaric acid esters (fuma-
`rates) with different chain lengths has been
`investigated. Additionally, the effect of pancrea-
`tic and hepatic enzymes on the ester stability has
`been studied. The effect of chain length on
`metabolic stability and permeability was investi-
`gated with fumarates with different acyl chain
`lengths. In 1982, Chang and Lee showed in albino
`rabbits that the ocular hydrolysis of a series of a-
`and b-naphthyl esters increased with increasing
`chain length of the acyl moiety. Several causes
`have been suggested for this increase in hydro-
`lysis rate with chain length. An increase of the
`acyl residue is accompanied by an increase in
`lipophilicity which itself has an impact on its
`metabolic disposition [5–7], principally because
`the compound must first partition into the active
`site of
`the enzyme being considered as a
`hydrophobic pocket. On this premise, the acyl
`chain length of an ester, via its influence on
`lipophilicity, may significantly influence the rate
`at which an ester would be hydrolyzed by
`esterases. Chang and Lee as well as other
`investigators [8–11] found a parabolic relation-
`ship between the rate of hydrolysis and the acyl
`chain length. A linear increase in hydrolytic rate
`with chain length, however, has been noted by
`Hofstee [12], and a reduction was reported by
`Jordan [13].
`Esterases are classified in three groups based
`on their interaction with organophosphates: A-
`esterases hydrolyse organophosphates, B-es-
`terases are inhibited by organophosphates and
`include cholinesterase (EC 3.1.1.8) and carbox-
`ylesterase (EC 3.1.1.1) and finally, C-esterases
`which do not interact with organophosphates
`[14]. In order to classify the esterases involved in
`the metabolism of
`fumarates, studies in the
`presence of well known esterase inhibitors
`(organophosphates and others) are presented.
`The lipophilicity of a compound, in addition to
`its degradation by esterases, will also influence
`
`its intestinal permeability, the second important
`prerequisite for bioavailability by the peroral
`route of administration. In general, an increase in
`the lipophilicity of a solute is accompanied by an
`increased ability to passively enter and permeate
`cell membranes. However, there appears to be an
`upper limit of the absorption rate as a function of
`lipophilicity [15], i.e. the rate limiting step shifts
`from membrane-controlled to unstirred water
`layer-controlled permeability [16–18]. By study-
`ing acetyl-, propionyl- and butyrylsalicyclic acid,
`Kimura et al.
`[19]
`reported an increase in
`permeability with increasing acyl chain length.
`In the present study, both cell culture and
`excised tissue permeability models were used in
`order to determine the permeability coefficients
`of dimethyl, methyl hydrogen, ethyl hydrogen, n-
`propyl hydrogen and n-pentyl hydrogen fuma-
`rate. As an in vitro animal model, excised porcine
`intestinal mucosa was employed, providing full
`physiological intestinal cell heterogeneity, mucus
`layer and residuals of luminal enzymes, all of
`which may affect the transport of permeants [20].
`As a second model, the well established Caco-2
`cell culture model, derived from human colon
`adenocarcinoma and characterized by cell homo-
`geneity was used.
`In the past, only limited attention has been
`paid to the intestinal permeability of fumarates,
`although transport studies of fumarate across
`porcine intestinal brush-border membrane were
`performed by Wolffram et al. [21]. Using isolated
`brush-border membrane vesicles,
`the authors
`observed Na+-dependent saturable transport of
`fumarate consistent with secondary active Na+/
`di(tri)carboxylate co-transport. The uptake was
`reported to be mediated by a single saturable
`transport process and to a much smaller extent
`by diffusional uptake.
`
`Materials and Methods
`
`Metabolism studies
`
`Chemicals. Dimethyl
`fumarate (DMF), calcium
`methyl fumarate (CaMF), calcium ethyl fumarate
`(CaEF), calcium n-propyl fumarate (Ca n-PrPF)
`and calcium n-pentyl fumarate (Ca n-PeF) were a
`gift from Fumapharm AG (Muri, CH). Phenyl-
`
`Copyright # 2003 John Wiley & Sons, Ltd.
`
`Biopharm. Drug Dispos. 24: 259–273 (2003)
`
`Page 3 of 16
`
`

`
`ABSORPTION OF ANTIPSORIATIC FUMARIC ACID ESTERS
`
`261
`
`methylsulfonyl fluoride (PMSF), diethyl p-nitro-
`phenyl phosphate (paraoxon), bis-nitrophenol
`phosphate (BNPP) and physostigmine (eserine)
`were purchased from Sigma Chemicals (Buchs,
`CH). Mercuric chloride was obtained from Fluka
`Chemie AG (Buchs, CH) and ethylene-diamine-
`tetraacetate (ETDA) from Siegfried Handels AG
`(Zofingen, CH). All other chemicals were of
`analytical purity and purchased from Fluka
`Chemie AG (Buchs, CH) or Merck Schweiz
`(Dietikon, CH).
`
`Pancreatic enzymes. Crude porcine pancreas ex-
`1
`Interdelta, Fribourg, CH) with
`tract (Eurobiol
`standardized activities of lipase (1350 U), amy-
`lase (1350 U), trypsin (33 U) and chymotrypsin
`(450 U) at a concentration of 1 mg protein ml1
`was used [22]. The concentration of the protein
`was determined by means of the Bio-Rad protein
`assay kit (Bio-Rad Laboratories, Glattbrugg, CH)
`with bovine plasma gamma globulin as a
`1
`was suspended in pre-
`standard. Eurobiol
`warmed (378C) Krebs’ phosphate buffer contain-
`ing 120.8 mm sodium chloride, 4.8 mm potassium
`chloride, 1.2 mm magnesium sulfate, 16.5 mm
`dibasic sodium phosphate and 1.3 mm calcium
`chloride, adjusted to pH 6.5 with hydrochloric
`acid. Control samples were obtained by incubat-
`ing the esters with pancreas extract which had
`been boiled for 15 min.
`
`Intestinal metabolism. Perfusate and homogenate
`were prepared from porcine small
`intestine
`(jejunum) obtained from the local slaughterhouse
`immediately after
`slaughter. Fresh parts of
`jejunum with a length of 50 cm were perfused
`with 50 ml Krebs’ phosphate buffer pH 6.5. The
`perfusate was kept at 48C and transported to the
`laboratory within 30 min. The perfused jejunum
`was stored in 0.9% sodium chloride solution at
`48C and also transported to the laboratory for
`preparation of intestinal homogenate.
`
`Small intestine perfusate. Immediately after arrival
`in the laboratory the perfusate was centrifuged at
`1
`3000 g for 10 min at 48C (Sorvall
`RC-5B refri-
`gerated superspeed centrifuge, Du Pont Instru-
`ments, Digitana AG, Horgen, CH). Thereafter the
`amount of protein was determined (Bio-Rad
`protein assay kit, Bio-Rad Laboratories, Glatt-
`
`brugg, CH) and the perfusate was stored at
`–808C.
`
`intestine homogenate. Homogenates were
`Small
`obtained by scraping off the mucosa with a glass
`slide at 48C. The scraped material was diluted
`1:10 in ice-cold Krebs’ phosphate buffer pH 6.5
`1
`PT 3000
`and homogenized with a Polytron
`homogenizer (Kinematica AG, Luzern, CH) at
`16 000 rpm for 2–4 min. The protein content was
`determined and the homogenate was stored at
`–808C.
`
`Hepatic metabolism. Fresh porcine liver was ob-
`tained from the local slaughterhouse immedi-
`ately after slaughter and was placed into ice-cold
`1.15% KCl in 0.1m phosphate buffer pH 7.4, and
`transported to the laboratory within 30 min. All
`subsequent procedures were carried out at 48C
`according to Lake [23]. After washing the liver
`twice with fresh phosphate buffer, 3 ml of 1.15%
`KCl in 0.1m phosphate buffer pH 7.4 per g wet
`tissue was added. The liver was chopped with
`scissors and then homogenized using a Poly-
`1
`PT 3000 homogenizer (Kinematica AG,
`tron
`Luzern, CH) at 16 000 rpm for 4 min. During
`homogenization the vessel containing the homo-
`genate was kept
`in ice water to keep the
`temperature at 48C. The homogenate was cen-
`trifuged at 9000 g for 10 min in a high-speed
`1
`centrifuge at 48C (Sorvall
`RC-5B). After decant-
`ing the supernatant and determination of the
`protein concentration, the decanted supernatant
`(S9 fraction) was stored at –808C.
`
`Incubation studies. Solubilized pancreatic extract,
`jejunal perfusate, jejunal homogenate or liver S9
`fraction, each standardized to a protein concen-
`tration of 1 mg ml1, were preincubated at 378C
`for 20 min. Dimethyl fumarate and calcium salts
`of methyl, ethyl, n-propyl and n-pentyl fumarate
`were added to yield final concentrations of
`0.05 mm, 0.1 mm, 0.5 mm and 1.0 mm, respectively.
`The concentration range was similar to the
`venous plasma concentrations found in pigs after
`oral and intravenous administration of dimethyl
`fumarate and calcium salts of methyl, ethyl, n-
`propyl and n-pentyl fumarate at doses of 10 and
`20 mg kg1 bodyweight. Incubation experiments
`with liver S9 fractions were also performed in the
`
`Copyright # 2003 John Wiley & Sons, Ltd.
`
`Biopharm. Drug Dispos. 24: 259–273 (2003)
`
`Page 4 of 16
`
`

`
`262
`
`D. WERDENBERG ET AL.
`
`presence of 1.0 mm NADPH as cofactor for
`cytochrome P450 mediated metabolism. Samples
`were taken at 0, 5, 10, 15, 20, 30, 45, 60, 90 and
`120 min after the addition of the fumarate and
`diluted in an equal volume of chilled 1.0 m
`perchloric acid to stop the reaction. After main-
`taining the sample >15 min in ice-water the
`samples were centrifuged for 4 min at 10 280 g
`(Hettich EBA 12, Andreas Hettich AG, B.aach,
`CH). The supernatant was analysed by HPLC.
`
`Inhibition studies. Pancreatic extract, jejunal per-
`fusate,
`jejunal homogenate or liver S9 fraction
`(protein concentration 1 mg ml1) were preincu-
`bated at 378C with 0.2 mm PMSF (in 100%
`isopropanol, the final isopropanol concentration
`was 2% in the incubate) and 5 mm EDTA, 0.1 mm
`paraoxon, 0.1 mm BNPP, 0.1 mm eserine and
`0.1 mm mercuric chloride, respectively, for 5 min
`prior to the addition of fumarates. Final concen-
`fumarates were 0.05 mm, 0.1 mm,
`trations of
`0.5 mm or 1.0 mm,
`respectively. The rate of
`fumarate hydrolysis was measured as described
`under incubation studies.
`
`HPLC assay. Fumarate. The HPLC system used
`consisted of a L-6200A pump, an AS 2000 auto-
`sampler, a SP8773XR UV detector
`(Spectra-
`Physics, Basel, CH) and a D-2500 Chromato-
`Integrator (Merck-Hitachi, Darmstadt, D). Ana-
`lyses of dimethyl, methyl hydrogen, ethyl hydro-
`gen, n-propyl hydrogen and n-pentyl hydrogen
`fumarate in the supernatant were performed by
`using a Li-Chrospher 100 RP-8 (5 mm) column of
`25 cm length and 4 mm in diameter (Merck,
`Dietikon, CH). The mobile phase consisted of a
`mixture of 0.05 m monobasic sodium phosphate
`and acetonitrile (65:35, v/v) adjusted to pH 3.2
`with ortho-phosphoric acid (85%). The flow rate
`was 1.0 ml min1 and detection was at 230 nm.
`Calibration curves were determined in aqueous
`0.05 m monobasic sodium phosphate and covered
`the full range of the expected concentrations in
`the samples. One batch of three samples of high
`(1.0 mm) and three samples of low concentration
`(0.05 mm) was prepared and analyzed, together
`with a calibration curve, in order to determine
`precision. The limit of quantification was at
`0.1 mg l1. As expected the retention time of the
`monoesters rose with increasing chain length
`
`Table 1. Elemental compositions, molecular weights and
`retention times of fumarates
`
`Compound
`
`Formula Molecular
`weight
`
`Retention
`timea
`
`144
`C6H8O4
`Dimethyl fumarate
`130
`Methyl hydrogen fumarate C5H6O4
`144
`C6H8O4
`Ethyl hydrogen fumarate
`n-Propyl hydrogen fumarate C7H10O4 158
`n-Pentyl hydrogen fumarate C9H14O4 186
`
`6.5–6.7
`3.2–3.3
`4.2–4.6
`6.0–6.1
`15.7–16.0
`
`a HPLC assay: Li-Chrospher 100 RP-8 column (5 mm particle size;
`4  250 mm) (Merck, Dietikon, CH); mobile phase: 0.05 mm monobasic
`sodium phosphate and acetonitrile (65:35, v/v) adjusted to pH 3.2
`with ortho-phosphoric acid (85%), flow rate 1.0 ml min1, UV 230 nm.
`
`(Table 1), due to the more intense interaction
`between the more lipophilic monoesters with the
`lipophilic stationary phase.
`
`Atenolol/propranolol. The HPLC system used con-
`sisted of a L-6200A pump, an AS 4000 auto-
`sampler, a L-4250 UV-VIS detector, a D-6000A
`interface and a D-7000 HPLC System Manager
`(Merck-Hitachi, Darmstadt, D). Analyses of
`atenolol in the supernatant were performed by
`using a Li-Chrospher 100 RP-8 (5 mm) column of
`25 cm length and 4 mm in diameter (Merck,
`Dietikon, CH); analyses of propranolol were
`performed by using a Li-Chrospher 60-5 selected
`B column of 25 cm length and 4 mm in diameter
`(Macherey-Nagel AG, CH). The mobile phase A
`consisted of 0.02 m KH2PO4, pH 4.7 and the
`mobile phase B of acetonitrile. Elution of atenolol
`and propranolol was accomplished by changing
`the mobile phases, A and B, according to a
`gradient running from 10% to 20% of solution B
`in 10 min applying a flow rate of 1.4 ml min1.
`Detection was by UV absorbance at 280 nm in
`case of atenolol and at 220 nm in case of
`propranolol.
`
`Data processing. Degradation half-lives and meta-
`bolic turnover rates. Degradation half-lives were
`calculated according to the following equation:
`t1=2 ¼ 0:693
`K
`with K as the degradation rate constant.
`
`Statistical analysis. Results are presented as
`mean  SD of n=3–4 degradation determina-
`
`Copyright # 2003 John Wiley & Sons, Ltd.
`
`Biopharm. Drug Dispos. 24: 259–273 (2003)
`
`Page 5 of 16
`
`

`
`ABSORPTION OF ANTIPSORIATIC FUMARIC ACID ESTERS
`
`263
`
`tions. Student’s unpaired t-test (two-tailed) was
`used to test the significance of the difference
`between two mean values. p50.05 was consid-
`ered statistically significant.
`
`Transport studies
`
`Chemicals. The Caco-2 cell line (passage 68) was
`obtained from the Department of Physiology
`(University of Zurich, CH). Transwell polycarbo-
`nate cell culture inserts (mean pore diameter
`0.4 mm, diameter 12 mm) were supplied by Costar
`(Basel, CH). Dulbecco’s Modified Eagle’s med-
`ium (DMEM) containing 0.5% l-glutamine and
`1.0 g l1 glucose, nonessential
`amino acids
`(NEAA) solution, fetal calf serum (FCS), 0.05%
`trypsin/0.025% EDTA solution and Hank’s ba-
`lanced salt solution (HBSS) were provided by
`Life Technologies (Basel, CH). 2- morpholino-
`ethansulfonic acid (MES) was from Fluka Chemie
`AG (Buchs, CH). Radioisotopes [3H]mannitol
`(spec. act. 19.7 mCi mmol1),
`[14C]PEG 4000
`(spec. act. 11.0 mCi g1) and [3H]d-glucose (spec.
`act. 15.3 Ci mmol1) were from Du Pont de
`Nemours International S.A. (Regensdorf, CH).
`All other chemicals were of the highest purity
`available.
`
`Cell culture. Caco-2 cells were maintained at
`37 8C in DMEM supplemented with 16.5% fetal
`bovine serum and 1% (v/v) NEAA in an atmo-
`sphere of 5% CO2 and 90% relative humidity. The
`medium was changed every other day. When the
`monolayer reached 90% confluence the cells were
`subcultured for passaging at a split ratio of 1:3 by
`treatment with 0.25% trypsin in 1 mm EDTA
`solution for 10 min at 378C. For transport studies
`Caco-2 cells were seeded onto polycarbonate
`Transwell inserts at a density of 100 000 cells
`cm2. Fully confluent monolayers were from
`passage 72 and used 17–18 days after seeding.
`
`Transport studies in cell culture system. Confluence
`and integrity of monolayers used for transport
`studies were assessed by measuring the para-
`cellular flux of [3H]d-mannitol, [14C]PEG 4000 or
`[3H]d-glucose, and transepithelial electrical re-
`sistances (voltage-current clamp, VCC MC6,
`Physiologic Instruments, San Diego, USA) during
`permeation. Transport studies in Ussing-type
`
`chambers were performed according to standard
`experimental conditions as described previously
`by Anderle et al. [54]. Caco-2 monolayers were
`deemed intact when the permeability coefficients
`of the markers were less than 4  106 cm s1 for
`mannitol and less than 2  106 for PEG 4000.
`The permeability of D-glucose in the absorptive
`cm s1,
`direction averaged 46.3  1.7  106
`whereas in the opposite direction a permeability
`of 28.8  0.1  106 cm s1 was measured.
`
`Transepithelial electrical resistance. Electrophysio-
`logical measurements were made by silver/silver
`chloride reference electrodes (Physiologic Instru-
`ments, Inc., San Diego, CA, USA) consisting of a
`silver wire, contained in a glass barrel and
`terminating in a ceramic tip. Voltage across the
`tissue was continuously zero-clamped (Model
`VCC MC6 multichannel
`clamp. Physiologic
`Instruments Inc., San Diego CA, USA) by
`completing the feedback loop. Pulses of 5 mV
`were applied for 994 ms with an interval of
`142 ms (time separating the two monopolar
`pulses forming the bipolar waveform) at a
`frequency of 20 s. The transepithelial electrical
`resistance (TEER) was calculated according to
`TEER ¼PD
`Isc
`
`where PD is the transepithelial potential differ-
`ence and Isc is the short circuit current. Caco-2
`monolayers and excised intestinal mucosa were
`deemed intact when the transepithelial electrical
`resistances were 5700 O cm2 and 520 O cm2,
`respectively.
`
`Preparation of excised intestinal mucosa. Jejunal
`porcine segments were obtained at the slaughter-
`house. Immediately after killing jejunal segments
`were placed in ice-cold Krebs’ phosphate buffer,
`pH 6.5 and transported to the laboratory within
`30 min. Segments were opened along the mesen-
`teric border to expose the epithelial surface. The
`mucosa was carefully stripped off the muscular
`system. Thereafter the mucosa was sandwiched
`between the compartments of the diffusion cell
`(Precision Instruments, Costar, NL), which were
`maintained at 378C by a water-heated jacket.
`
`Copyright # 2003 John Wiley & Sons, Ltd.
`
`Biopharm. Drug Dispos. 24: 259–273 (2003)
`
`Page 6 of 16
`
`

`
`264
`
`D. WERDENBERG ET AL.
`
`Transport studies in excised intestinal mucosa. The
`compartments were filled with prewarmed
`transport buffer which was circulated by an 5%
`CO2/O2 air lift (25 ml min1). The experiment
`was started after an equilibration period of
`30 min by changing both the mucosal and the
`serosal buffer solution. Transport studies were
`performed as described under ‘transport studies
`in cell culture system’. The volume of the donor
`and receiver compartment was 1.2 and 1 ml,
`respectively. Samples taken at specified intervals
`for the determination of the drug and marker
`concentration were 0.1 ml instead of 0.5 ml. As
`markers for passive transport, [3H]mannitol and
`[14C]PEG 4000 were added to the drug solution
`(marker activity 1 mCi ml1). The permeabilities
`for mannitol were 3.6  0.9  105 cm s1 and
`4.2  0.8  105 cm s1 in the absorptive and
`secretory directions, respectively.
`
`HPLC assay. HPLC analysis was performed as
`described under metabolism studies.
`
` 
`Data processing. Effective permeability coeffi-
`cients, Peff (cm s1) were calculated according to
`V
`Peff ¼ dC
`dt
`AC0
`
`ss
`where (dC/dt)ss is the flux across the monolayer
`(mg ml1 s1) at steady state, A is the diffusion
`area (cm2), V is the volume of
`the receiver
`compartment (ml) and C0 the initial concentra-
`tion in the donor compartment (mg ml1). Data
`are presented as mean  SD of n=3–4 Caco-2
`monolayers or excised intestinal mucosa. Stu-
`dent’s unpaired t test (two-tailed) was used to
`test the significance of the difference between
`two mean values. p50.05 was considered statis-
`tically significant.
`
`Results
`
`Metabolism studies
`
`Intestinal stability of dimethyl fumarate. The degra-
`dation half-life as a function of dimethyl fuma-
`rate concentration is shown in Figure 1.
`In
`pancreatic extract as well as in intestinal perfu-
`sate and homogenate the degradation half-life of
`
`Figure 1. Intestinal stability of dimethyl fumarate in terms of
`1
`)
`degradation half-lives [min] in pancreas extract (Eurobiol
`. and intestinal homogenate & in
`, intestinal perfusate
`Krebs’ phosphate buffer (pH 6.5, 378C; protein concentration
`1 mg ml1). Mean  SD, n=3–4
`
`fumarate increased with increasing
`dimethyl
`fumarate concentration. The degradation rates
`were highest
`in the intestinal homogenate,
`followed by the perfusate and pancreas extract.
`
`inhibitors on the enzymatic activity of
`Effect of
`intestinal perfusate and homogenate. Inhibitory stu-
`dies were carried out to specify the esterases
`involved in enzymatic degradation of dimethyl
`fumarate. The classical inhibitors PMSF in com-
`bination with EDTA, paraoxon, BNPP, physos-
`tigmine (eserine) and mercuric chloride were
`used. The resulting degradation half-lives are
`given in Table 2. A significant increase in half-life
`was observed when PMSF/EDTA, paraoxon and
`BNPP, respectively, were added to intestinal
`perfusate. On the other hand, inhibition was less
`efficient when mercuric chloride and eserine
`were used. The esterases present in the intestinal
`homogenate were most affected by paraoxon and
`PMSF/EDTA, as indicated by the significantly
`higher half-lives of dimethyl fumarate. BNPP
`showed less and mercuric chloride and eserine
`had no effect on the esterase activity of intestinal
`homogenate.
`
`Intestinal stability of monoesters. The degradation
`of calcium methyl, calcium ethyl, calcium n-
`propyl and calcium n-pentyl fumarate in the
`presence of pancreas extract, intestinal perfusate
`and homogenate was determined. No significant
`decrease of
`the concentration of
`the parent
`substances was observed within the investigated
`
`Copyright # 2003 John Wiley & Sons, Ltd.
`
`Biopharm. Drug Dispos. 24: 259–273 (2003)
`
`Page 7 of 16
`
`

`
`ABSORPTION OF ANTIPSORIATIC FUMARIC ACID ESTERS
`
`265
`
`Table 2. Effect of inhibitors on esterase activity in intestinal medium, determined as degradation half-lives of dimethyl fumarate
`(DMF) at concentrations of 0.05, 0.10, 0.50 and 1.00 mm. Degradation half-lives of dimethyl fumarate (DMF) in Krebs’ phosphate
`buffer containing pancreas extract are shown as well. Mean  SD, n=3–4
`Half-lives (min) at DMF concentration (mm)
`
`Inhibitor
`
`Medium
`
`Perfusatea
`
`Homogenateb
`
`Pancreas extractc
`
`None
`PMSF/EDTA
`Paraoxon
`BNPP
`Eserine
`HgCl2
`None
`PMSF/EDTA
`Paraoxon
`BNPP
`Eserine
`HgCl2
`None
`Boiledd
`
`0.05
`
`3.4  0.2
`198  35
`72  9
`79  1
`17  1
`9  1
`1.0  0.3
`64  15
`113  3
`2.8  0.1
`0.9  0.0
`1.2  0.0
`52.5  2.0
`2075  563
`
`0.10
`
`4.3  0.2
`245  52
`108  4
`107  20
`19  3
`10  1
`1.3  0.2
`260  20
`415  130
`2.9  0.4
`1.0  0.0
`1.2  0.0
`53.4  0.9
`2872  203
`
`0.50
`
`5.4  0.1
`561  97
`362  37
`421  167
`24  6
`12  1
`1.8  0.6
`488  6
`852  88
`8.3  0.3
`1.2  0.1
`1.1  0.0
`61.9  1.6
`2921  452
`
`1.00
`
`6.4  0.3
`887  231
`828  139
`883  135
`32  2
`21  6
`2.2  0.3
`683  118
`1245  58
`11  1
`1.2  0.1
`1.1  0.0
`80.2  9.5
`3419  112
`
`a Intestinal perfusate in Krebs’ phosphate buffer (pH 6.5, 378C, protein concentration 1 mg ml1).
`b Intestinal homogenate in Krebs’ phosphate buffer (pH 6.5, 378C, protein concentration 1 mg ml1).
`1
`) in Krebs’ phosphate buffer (pH 6.5, 378C, protein concentration 1 mg ml1).
`c Pancreas extract (Eurobiol
`1
`) boiled during 20 min at 1008C.
`d Pancreas extract (Eurobiol
`
`Table 3. Effect of inhibitors on esterase activity in liver S9 fraction, determined as degradation half-lives of dimethyl- (DMF),
`methyl hydrogen- (MHF), ethyl hydrogen- (EHF), n-propyl hydrogen- (n-PrHF) and n-pentyl hydrogen fumarate (n-PeHF) at
`initial concentrations of 1.0 mm. Mean  SD, n= 3–4
`Half-lives [min] at initial fumarate concentration of 1.0 mm
`
`Inhibitor
`
`None
`NADPHa
`PMSF/EDTA
`Paraoxon
`BNPP
`Eserine
`HgCl2
`
`DMF
`
`MHF
`
`EHF
`
`464  67
`518  76
`51.0
`457  41
`532  24
`51.0
`1241  184
`1539  347
`68  2
`564  102
`3269  552
`177  19
`823  62
`1961  624
`47  8
`829  79
`872  72
`51.0
`623  150
`1144  6
`51.0
`Half-lives at initial reference concentration of 1.0 mm
`
`None
`
`Atenolol
`821  64
`
`a NADPH as cofactor for cytochrome P450-mediated metabolism.
`
`n-PeHF
`
`137  19
`136  7
`880  45
`177  19
`674  119
`312  31
`132  12
`
`n-PrHF
`
`340  56
`338  24
`1100  88
`1359  59
`1435  187
`585  95
`324  15
`
`Propranolol
`389  71
`
`time frame of 120 min (data not shown). Because
`of an overlap with the initial matrix peak fumaric
`acid was undetectable by the analytical method
`used. Therefore, it cannot be ruled out that minor
`degradation of the monoesters may have oc-
`curred.
`
`Hepatic stability of fumarates. Stability studies in
`liver S9 fraction were carried out in order to
`
`determine degradation half-lives of dimethyl-,
`calcium methyl-, calcium ethyl-, calcium n-
`propyl- and calcium n-pentyl fumarate, respec-
`tively. As controls, the degradation half-lives of
`atenolol and propranolol in the liver S9 fraction
`was studied (Table 3). The degradation half-lives
`of atenolol and propranolol were 821 and
`389 min, respectively. The degradation half-lives
`of the homologous series of fumarate monoesters
`
`Copyright # 2003 John Wiley & Sons, Ltd.
`
`Biopharm. Drug Dispos. 24: 259–273 (2003)
`
`Page 8 of 16
`
`

`
`266
`
`D. WERDENBERG ET AL.
`
`(Table 3) were found to decrease with increasing
`acyl chain length. Control studies performed in
`the presence of NADPH revealed no significant
`differences in the degradation half-lives com-
`pared with results obtained without NADPH.
`Degradation of dimethyl fumarate was the most
`rapid, with concentrations below the detection
`limit (0.1 mg l1) within 5 min after the start of
`the incubation.
`To specify the enzymes involved in the hepatic
`degradation of fumarates, inhibition studies were
`carried out
`in the presence of PMSF/EDTA,
`eserine and mercuric chloride. The degradation
`half-lives of the homo