‘ORIGINAL ARTICLES: Faculty of Pharmacy’, Ljubljana, Slovenia; Drug Transport and Delivery Research Group*, Department of Pharmacy, Faculty of Health Sciences, University of Tromso, Norway Physiological barriers to the oral delivery of curcumin K. Benainc’, J. TAONTELJ’, N. SKALXO BASNET?, A. Kristi? ‘Received August I, 2011, accepted September 12, 2011 Kauja Berginc, Faculty of Pharmacy, Atkeréeva 7, 1000 Ljubljana, Slovenia aja herginc @ fa.uni-l.i Pharmazie 67: 518-524 (2012) dois 10.1691/ph 20121112 ‘Curcumin, a principal component from Curcuma longa, with antioxidant and antiinflammatory activities ‘was proposed as a potential candidate for the preventation andlor treatment of cancer and chronic dis- ‘eases. However, curcumin could not achieve its expected therapeutic outcome in clinical trials due to its low solubility and poor bioavailability. The actual intestinal physiological barriers limiting curcumin absorp: tion after oral administration have not been fully investigated. To identify the main barriers curtaling its absorption, in vitro permeability of curcumin and flux of its glucuronide were monitored in rat jejunum and Transwell grown Caco-2 cells. Curcumin was more permeable under acidic conditions, but the permeability was substantially below the permeabilly of highly permeable standards. Its efflux could not be inhibited by specific Pgp and MRP inhibitors. BCRP was found to participate in curcumin transport, but the Organic ‘Anion Transporting Polypeptide (OATP) did not. The permeabilly of curcumin significantly increased when the structure of mucus was compromised. The inhibitor of curcumin metabolism, piperin, failed to act as @ permeability enhancer. Piperin inhibited Pap and MRP transporters and decreased the amount of glu ‘curonide transported back into the intestine. Inclusion of piperin in curcumin-containing formulations is highly recommended as to inhibit curcumin glucuronidation and to increase the transport of formed glu Ccuronides into the plasma, therefore increasing the probabilly of glucuronide distribution into target tissue ‘and inter-convertion to curcumin. It would also be beneficial, if curcumin delivery systems could reversibly compromise the mucous integrity to minimize the non-specific binding of curcumin to its constituents. 1. Introduction ‘The thizome of Curcuma longa represents the main source of turmeric, a yellow spice with along histary of use for dietary and medicinal purposes in Indian and Chinese cultures (Hatcher etal. 2008). The main constituent of turmeric extrac, curcumin, a hydrophobic polyphenole, is poorly absarbed after oral appi- cation, Low systemic availability thus mitigates against using curcumin in the prevention of conditions distal from the ga toointestinal tract (GIT) (Hatcher et al.2008; Basnet etl. 2011). Namely, therapeutic plasma concentrations requited to elicit biochemical changes germane to chemoprevention, antioxidant, antiviral and antiinflammatory effects determined in in vitro experiments are in the micromolar range (Zhang and Lim 2008; Wortelboer et al. 2003; Ireson etal. 2001), while clinical stud- ies in cancer patients, receiving 412g of pure curcumin per day determined low, nanogram steady-state levels (22-41 ng/ml ker 12g po. dase (Dhillon et al, 2008), 4 ngiml. aiter 3.62 po, dose (Sharma 2004)) of circulating curcumin, The pi dominant species of curcumin in plasma are represented by its conjugated metabolites (glucuronides and sulfates) and products of reduction reactions (tetrabydro- and hexabydrocurcumin and hhexahydrocurcuminal)(Basnetet al. 2011), Although less effet tive than the parent compound, these metabolites are believed to be responsible for systemic activity Ge. lowering plasina Ievels of glucose, cholesterol, tiglycerides, anti-obesity effect ce.) afler curcumin consumption (Hatcher etal. 2008; Basnet sis et al 2011), a promising feature rendering this polyplienolic compound worthy of further research, Poor curcumin absorption from intestine isa feature pertinent (o this phytochemical, observed across several animal species (Basnet et al. 2011). Low water solubility, decomposition at neutral or alkaline pl, photosensitivity and a coordinately egu- lated alliance between metabolizing enzymes and transporters, all actin tandem with the net result of low curcumin absorp- tion (Hatcher etal. 2008; Walang etal. 2011). fn vitro studies with Caco-2 Hatcher etal, 2008; Wablangetl.2011), MDCKTL (Worteboer etl, 2003) and L1.C-PKI (Zinang eal. 2008) cells, experiments with vesicles isolated fom Multidrug Resistance Associated Proteins | and 2 (MRP-1 and MRP-2) transfected ‘Sf cells (Wortelboer et a. 2003) and CYP3A4 studies (Zhang eal. 2008) identitied P-glycoprotein (Pep), MRP-1, MRP-2, Cytochrome P450 isoenzyme 3A4 (CYP3A4), sulfotransferase TAL and 1A3 (SULTIAL and 1A3) (reson etal, 2002), UDP- slucurony! transferases (Lreson etal. 2001; Liu et al. 2006) and nonspecific oxydoreductases (reson 2001) as the key step of enlero-enlerie and hepatic pre-systemic metabolism. However, no proper in vitro permeability studies providing deeper knowledge about barriers to its absorption have been performed by now. A detailed review of in vitro curcumin permeability studies through cell cultures reveals thatthe dter- ‘mination of curcumin permeability was either not the purpose ofthese studies (curcumin was only use to inhibit transport of Pp substrates digoxin, rhodamine (Zhang etal. 2008), MRP2 Pharmazie 67 (2012)ORIGINAL ARTICLES substrate calcein (Wortelboer 2003)) or, that the experimen- {al design of these studies was improper considering stability and solubility data oncureumin curcumin concentrations tested ‘were as high as 170 uM. the solutions for testing were prepared at pH 65 or 74) (Zhang et al. 2008; Worteboer et al. 2003; Wablang et al. 2011). Namely it is Known that curcumin has extcemely low water solubility (0.03 M) and decomposes in ‘water at pl values above 7.2 (Elatcher et al. 2008; Kamiga etal. 2003). This could significantly influence the measurement of ccurcumin’s permeability: Therefore the aim ofthis study was to cvaluate curcumin’s intestinal permeability :hrough Caco-2 cel monolayers as well as through rat jejunum and to investigate the influence of physiological bariers on its permeability (Le. ‘mucus, pHalong intestinal tact, participation of intestinal trans- porters). The potential of piperin, a known inhibitor of curcumin re-systemic glucuronidation, to act as permeability enhancer for curcumin was also investigated. In all experiments albu- smi was used in both donor and acceptor solutions to increase curcumin stability and solubility. In addition, the wansport of ‘curcumin glucuronides was aso investigated 2. Investigations and results 2.1. Curcumin stability in Ringer buffer pH 6.5 and 7-4 ‘To perform curcumin permeability experiments, the stability of curcumin in donar solutions were frst evaluated, 504M cur- cumin in Ringer butler of pH 6.5 and 7.4 was prepated from curcumin DMSO siock solution (final DMSO concentration 10.1%), similarly tothe procedure described in published papers (Wablang et al 2011; Zhang et al. 2008). The concentration of curcumin in the heated (37°C) and oxidized donor solu- tions was analyzed 25 min alter their preparation and after 5h, ‘The LC-MS/MS measurements indicated a significant decline incurcumin concentrations inthe timeframe of 25 min (less than 1$% of initial curcumin concentration was measured). After 3h, ‘curcumin concentration was below the detection limit. Because the stability of curcumin can be improved withthe addition of plasma proteins (Hatcher et al. 2008; Kaminaga et al. 2003) 2 vollvol % of human albumin was added to donor sok tions, used for permeability experiments. With albumin in the solutions, the conceatrations of curcumin remaining in the donor solutions after 3 h incubation remained unchanged (97.2-43.2% of the inital eurcumin concentration). The solu- tions with 2 vol/vol % of albumin were thus used for further permeability experiments, The mass balance and the recovery of curcumin during permeability experiments remained prac- tically unchanged and above 90%. The use of other surface acting agents was intentionally avoided, because they are known to modify membrane fuiity, ransporer-enzyme interplay and could thus lead to false permeability measurements 2.2. Cureumin permeability through Caco-2 monolayers ‘The permeability of 100 wM curcumin was determined through ‘Transwell grown Caco-2 cell monolayers with 2% of albumin added to both sides of the monolayer. Due to the presence of albumin, the concentration of the unbound curcumin in donor solutions was unknown, most probably below curcumin water solubility (0.03 uM), but constant throughout the experiment. ‘The permeability was determined inthe absorptive (AP-BL) and. the opposite (BL-AP) direction at acidic and neutral pH (.e, 6.5 and 7.4) on the apical side of the cells, wile the basolateral [pH was kept constant at pH 7.4, The pastcipation of effiux (ep, MRPs and Breast Cancer Resistance Protein (BCRP)) and absorptive (Organic Anion Transporting Polypeptide OATP) Pharmazie 67 (2012) ‘Table 1: Permeabi ‘monolayers of curcumin through Caco-2 cell onion ar osna Ret 139405 198-410" E3s(0SmM) 135415 nana Ret 61219 102408" PSCS35 (5 mM) 1oskos MKS71 (50 mM) 1o+o4 Fam (5 mM) 396438 ‘transporters was assessed by using appropriate inhibitors and pH conditions on the apical side ofthe cells ‘The results, presented in Table 1, indicated asymmettical rans- ort properties of curcumin regardless of the pH applied to the Apical side, In both cases, the BL-AP permeability was sigaif- cantly higher than the absorptive one (p <0.05). However, under slighlly acidic conditions (pH 6.5) on the apical side, the absorp- tive permeability of curcumin significanly surpassed the one, determined at iso-pH conditions (apical and basolateral pil 7). ‘Although higher, AP-BL permeability value measured at pH 6.5 ‘was not inthe range of AP-BL permeabilities determined in our laboratory for highly permeable standards through ‘Transwell _Brown Caco-2 monolayers (Yu etal, 2002), therefore curcumin could not be classified as high but as low permeable compound. ‘To idemigy twansporters involved, specific inhibitors were used (PSC833 for Pep, MKST1 for MRPs and fumitromorgin C for BCRP) at appropriate concentrations and iso-pHT conultions as indicated in Table 1. We did not observe any permeatil- ity decrease in BL-AP directions when specific Pap and MRP. inhibitors were added. For the Bist dime the participation of [BCRP to curcumin efflux from Caco-2 cells was aso assessed, We used fumitromorgin C as a specific BCRP inhibitor, which increased the BL-AP permesbility of curcumin significantly. ‘The permeability increase observed in the presence of 5 mM fumitromorgin C was not due to Caco-2 cell monolayer injures, ‘because TEER values and the permeability of paracellular ‘marker fluorescein (the marker of tight junction integrity) did ‘ot change (data nat shown) According to circumin structure and its acidic pKa values (pKa, =8.38-£0.04; pKa) =9.88 £0.02: pKa, = 10,51 £0.01) (Wablang etal 2011), the main form of euzcumin present at pl (65 and 74 is predominately unionized (Tennesen etal. 2002) Because polyphenolic compounds have been identified as OATP. substratesinhibtors Fuchikami tal. 2006), we also anticipated {hat this absorptive transporter could be involved in the absorp- sion of curcumin, Since OATPe exert their maximal transporter efficiency in acidic environment, the impact of this wansporter ‘was monitored at apical pH 6,5 and estron-3-sulfate (E38) was used for its inhibition (Nozawa et al. 2008), E38 did not affect curcumin AP-BL. permeability, which remained practically ‘identical tothe reference AP-BL curcumin permeability values 2.3. Curcumin permeability through rat jejunum To address the participation of intestine in limiting oral absorption of curcumin, Sprague-Dawley rats were used. The 519‘ORIGINAL ARTICLES: ‘The increase of eureumin permeability in ‘the absence of mucus Re: lnc of ruse pray of uri gh ua setabolism of curcumin to glucuronides and reduced curcumin species was negligible, because the quantliy of metabolites in the acceptorldonor solutions ot in the tistue extract, prepared after the permeability experiments, were below LC-MS/MS detection limit (data not shown). At apical pH 6.5 and baso- lateral pH 7-4 the M-S permeability in the absorptive direction was (14.3 + 2.5) x 10~! ems, while at apical pH'7.4 permeabil- ity was significantly lower (3.0-+ 0.2) x 10~" em, similarly to the results obtained with Caco-2 model, The S-M permesbil- ity was not significantly different from the corresponding M-S permeabilities (the $-M permeability at acidic apical pH was (1063.1) x 10, while at apical pH 7.4 the S-M values were (25:06) x 107 24, The impact of mucus on curcumin permeability ‘Todisrupt macys layer in rat jejunum in vitro, 2mMN-acetyl cysteine (NAcCys) was applied inthe incubation slines(pl16.5 and 74) on the mucosal side and incubated for 2h, Afterwards, these buffers were removed and replaced by donar solutions, containing 100 uM curcumin, 2mM NAcCys and 2% albumin and the permeability of curcumin was monitored for additional 2h, Simultaneously, reference experiments were conducted, ‘where permeability of curcumin was monitored without NAc- Cys in the donor solutions. In the absence of mucus, the M-S curcumin permeability at apical pl 74 significantly increased (dhe permeability increased 2.3-times) and the effect was ev more pronounced at acidic apical pH of 6.5, where a 5.7-time increase of curcumin absorplive permeability was noted (Fig. 1). 25. Influence of piperin on the permeability of curcumin ‘To evaluate whether piperin could be used as permeability enkiancer for curcumin, the experiments were performed at spi- eal pl 65, and the permeability of 100M curcumin was ‘monitored ¢hrough rat jejunum and Transwell grown Caco-2 cell monolayers in the presence of piperin (100 uM) on the smucosalapical side. The results are presented in Fig. 2. As shown, piperin failed to increase the absorptive permeability of curcumin in both models used (the AP-BL/M-S permeabilities determined in the presence of piperin did not change signifi- cantly compared to reference AP-BLIM-S values). Instead of increasing the amount of curcumin absorbed in the M-S/AP-BL direction, piperin significantly (p<0.01) enhanced the active effiox of curcumin from the Caco-2 cells and rat ente- rocytes even more than determined inthe reference experiments (the S-M/BL-AP permeabilities, determined inthe presence of pipetin, were significantly (p <0.01) higher than the comespond- 520 ing reference S-MUBL-AP permeabilities - see 22 and Table 1). ‘The S-M curcumin permeability increased more than 4-times rough rat jejunum, while the effect was not so pronounced in Caco-2 cells (only 1-7-time increase of BL-AP permeability) 2.6. Curcumin glucuronide transport in Transwell grown Caco-2 cells, ‘Transport properties of cureumin glucuronide were monitored in Transwell grown Cavo-2 cells (Table 2). The results mea- sured under iso-pHT conditions (Table 2, 747.4 REF) indicated thatthe glucuronides ormed inside Caco-2 cells undergo active ‘wansport into the apical and basolateral sides, but the secretion of glucuronides (BL-AP flux) back into the apical side signifi- cantly (p<0.001) surpassed their absorption (AP-BL flux) into thebasolateral compartment, With the use of specific ABC trans- porter inhibitors (Table 2; PSC833, MKS71, FumC) the efflux of curcumin glucuronides significantly declined, bat so did its flux in the absorptive direction, However, the ratio between AP-BL and BL-AP fluxes increased (Irom 0,50 far reference experi- ‘ments 00,62 for PSC and MKS71) for all inhibitors, except for FumC. The ratio between AP-BL and BL-AP fluxes in the presence of FumC decreased significantly. ‘At apical pH 6.5, which resembles physiological conditions in the proximal GIT beter, uxes of glucuronides in boh directions were significantly (p<0.01) lower than at iso-pHT conditions (Cable 2). However, the 1ux of glucuronide in the absorptive (AP-BL) direction exceeded that in the opposite -BL-AP direc- ‘Table 2: Fluxes of curcumin glucuronide through Transwel grown Caco-2 cells 144 REP 0.065 0.000 0.131-£0.026" PSCH3—0.0560,000" 0.091 0.001" MKS71 —0.0560,000"_0.090.4.0.001"* FumC —0.042£0.001 0.0964 0.003" esa REP 0.060.002 0,032. 0.002" Piperin —0.048=0,.001 0.025 + 0.006" Pharmazie 67 (2012)ORIGINAL ARTICLES tion (Table 2, 65/7 4 REF). Inthe presence of piperin (100 »M) BL-AP flux of glucuronide declined signiticanty (Table 2). 3. Discussion ‘The contribution of intestinal physiological barriers to the absorption of curcumin after oral delivery has not yet been evaluated. Namely, the use of curcumin in the in vitro per rmeabilty studies through cell cultures has been limited only to the use of curcumin as Pep/MRP inhibitor (Zhang et al. 2008; Wortelboer 2003), There is only one paper published by ‘Wablang et al. 2011), where the permeability and the impact of efflux transporter Pgp to curcumin permeability were deter ‘mined. However, very high donor concentrations (170 uM) of curcumin in HBS pIT 6.5 were applied to the cells. Tis group also reported that the recovery of curcumin after permeability experiments was very low (only 35-54%) (Wablang etal. 2008), indicating tha the experimental design used was most probably improper considering stability and solubility data on curcumin. ‘Namely, curcumin is very unstable in aqueous acidic or neutral solutions, because more than 90% of initial quantity of cur- ‘cumin prepared in water buffer pH 7.2 decomposes in 30min at 37°C (Hatcher eta. 2008), Furthermore, water solubility of ‘curcumin is less than 0.03 Mat room temperature (Hatcher cet al, 2008), The instability/poor water solubility of curcumin ‘was also demonstrated in our preliminary experiments, where the recovery of curcumin in Ringer buffer a pH 6.5 and 7.4 ‘was below the LC-MS/MS detection limit. The decline in cur- ccumin's concentrations in our solutions was thus most probably caused by the non-specific binding of curcumin to the plastic, accompanied by precipitation and decomposition. Based on this dala, we could not perform experiments with curcumin zoli- tions in Ringer buffers. However, the stability of curcumin can bbe greatly improved inthe presence of plasma proteins (Hatcher ct al. 2011; Kaminaga 2003). Namely, in the presence of 10% of FBS, only 50% of curcumin decomposed at 37°C during 8 h ‘incubation. Thus, forthe purpose of this study, human albumin (@ vol/vol %) was added into donor and acceptor Ringer solu- tions, used for permeability experiments, which significantly improved curcumin stability. With albumin inthe solutions, the mass balance andthe recovery of eurcumin during permeability experiments remained practically unchanged (> 90%), which is signiticantly greater than the recovery (« 50%) determined for in vitro permeability experiments with Caco-2 cell monolayers published by Wablang (Wablang etal. 2011). oor physio-chemical properties, instability in GIT, metabolism and active efflux of curcumin could all contribute simultaneously to curcumin low in vitro permeability through rat jejunum and. ‘Transwell grown Caco-2 monolayers, determined inthis study (Hatcher etal, 2011; Basnet et al, 2011), Curcumin has been used to inhibit secretion of Pgp and MRP? substrates (Zhang et al. 2008; Wortelboer ei al. 2003). However, the identifica. sion of transporters involved in curcumin transport through the intestine has not yet been investigated. Wablang et al (2008) ‘monitored curcumin permeability through Caco-? monolayers in both directions but found no differences between AP-BL and BL-AP permeabilities. Verapamil, a nonspecific Pgp inhibiter, hhad no impact onthe permeability either (Wablang etal. 2011). Contrary to thei results, we showed that curcumin efflux from intestinal cells could represent an important barrier to curcumin absorption, since BL-AP permeability significantly exceeded the one in AP-BL direction at both pH. However, the efflux could not be inhibited by specific Pgp (PSC883) or MRP (MK- '571) inhibitors, indicating that the binding affinity of curcumin {towards all Pgp and MRP binding sites can be much higher ‘than the one of inhibitors or on the other hand, the binding of Pharmazie 67 (2012) curcumin to these proteins was ineversible, making curcumin ‘more potent Pgp and MRP inhibitor than these currently on the ‘market. BCRP inhibitor fumitromorgin C significantly increased ‘curcumin efflux from Caco-2 cells Although monolayer vital- ity and integrity were not compromised by fumitromorgin C, the permeability increase was most probably induced by posiive- ‘cooperation between fumitromorgin C and curcumin in BCRP. ‘There aze multiple binding sites in the strcture of efflux trans- porters (Pep has four, MRP-2 and BCRP have two) (Huang ct al. 2006; Berginc etal. 2010; Sharom 2008). The substrate ceanthas occupy only one or more binding sites, depending on its affinity and concentration. It is possible thatthe binding affin- ity of curcumin towards BCRP was not as high as in the case of Pgp and MRPs, where specific transporter inhibitors failed to displace curcumin from protein and decrease its efflux. Tt is also possible, that the donor curcumin concentration was too low to occupy both BCRP binding sites simultaneously. ‘This would leave one BCRP site fre for the inhibitor or the inhibitor could successfully displace curcumin from a least one of he binding sites. Because simultaneous binding of wo dilfer cent substrates to their individual binding site, otherwise known ‘as positive-cooperation between binding sites (Bergine et al. 2010), resalis in allosteric changes and increased transporter binding affinity towards both of its substrates, their transport significantly increases (Bergine et al. 2010; Sharom 2008) Positive-cooperation between both binding sites in BCRP was also most probably the resson, why we failed to inhibit BCRP. ‘mediated curcumin wansport in the BL-AP dizection. We also confirmed that OATP wansporters donot participate in facilitated curcumin diffusion into enterocytes, although these transporters have been shown (o increase the amount of polyphenoli-type ‘of compounds (Fuchikami et al. 2006). ‘The role of the intestine in contributing to sub-therapeutic cur ‘cumin plasma concentrations after per-oral application was also ‘evaluated i vitro with rat jejunum, due to the dissimilarities inpphysiological characteristics between cancerous (i.e. Caco-2) and non-cancerous (i. ra jejunum) models (Ya etal. 2002) It is known that curcumin bioavailability ater oral adminis- tration is low regardless of the animal model (Hatcher etal. 2008) and that its metabolism in F344 rats gives the same profile of metabolites qualitatively as in humans, but aver all, the metabolism of curcumin in rats highly underestimates Duman situation (Lreson et al. 2001). However, our rat model was unsuitable for metabolism studies, because no metabolites ‘could be detected inthe incubation solutions or in tissue extract, ‘Since metabolism could not allect curcumin donor concen- ‘ration in our permeability studies, this model can be readily used to determine intestinal curcumin permeability. Surpri ingly, the permeability of curcumin through rat intestine was even lower than through Caco-2 cell monolayers. There were also no significant differences between permeabilities in M-S and S-M direction, indicating that effiux transporters did not paicipate in the intestinal absorption from rat jejunum. The ‘compound's permeability can be assessed with models of vari ‘ous degrees of complexity according to the FDA guidelines (Yu et al, 2002). Usually, 2 low permeable compound, whose per- smeabilty bas been determined with excised intestinal tissue of animals, shows even lower permeability values through Caco-2 cell monolayers, because there ae less ight junctions with nar- rower openings in Caco-2 cells than in animal intestinal issue land because the expression pattems of transporters and mem- brane fuidty significantly differ among models used (Artursson ct al. 2001). However, in the case of curcumin, the opposite ‘was observed; the permeability of low permeable compound was higher in Caco-2 model than in rat jejunum. Consider- ing the differences beoween both models, we anticipated that ‘mucus could represent an additional bari to curcumin absorp- sa‘ORIGINAL ARTICLES: tion, Namely, mucus is actively seereted tothe mucosal side of| enterocytes, but itis not present in the apical side of Caco-2 cells. Although mucus thickness varies depending on the bowl movements and content mixing along the GIT (Atuma etal 2001), its basically water Gilled layer of glycoproteins, Hipids and water, sustaining acidic microclimate pH (Legen and Krist 2003), which could bind free curcumin, lower its concentration {gradient and thus decrease its permeability through enterocytes. Disruption of mucus ahove rat jejunum by the mucalytic azent NAcCys significantly increased curcumin absorptive permeabil- tiny especially under acidic conditions. Guided by these results it is obvious that mucus and its constituents might represent an additional barrier, which non-speeificlly binds curcumin and prevents it from sufficient absorption after oral applic tion, There are contradictory reports about mucus influencing the absorption and permeability of drugs. The in vitro pecme- ability of testosterone has been shown to be greatly (up to 80%) influenced by the mucus (Karlsson etal. 1993). Permeability increase was observed also with dextropropoxyphene through Caco-2-H29GlueHl co-cutures in the pretence of mucolytic agent NAcCys (Meaney and O'Driscoll 1999). In another study ‘with pig mucus the diffusion of lipophilic drugs (hydrocort- sone, propranolol, metoprolol and testosterone) was shown to be profoundly influenced by nizcus, especially by the lipids init (OWikman Larhed and Bjork 1998), Although numerous reports can be found claiming that mucus is not a diffusion bassier to smaller molecules (molecular weight below 700 g/mol) (Lege and Krist! 2001, ic seem that generalization ofthis statement is not possible. I most probably depends on the compound and its ccharacteristis, whether the mucus can hinder is diffusion from the lumen to the absorptive surface. In the case of curcumin, also a very lipophilic compound (log P>3), mucus seems to participate in its non-speciic binding either tothe proteins orto the lipids (Legen and Krist! 2001). ecause reversible mucis removal from the surface of intes- tine in vivo presents safety issues forthe patient, the attempts to increase curcumin per-oral bioavailability have tured to for- ‘mulation strategies, Thus nanoparticles, cyclodextrines, natural viscous polymers, liposomes and mixed micelles have been used as novel drug delivery systems which appear to assure bet- ter permeability and resistance to metabolism (Hatcher et al 2008; Huang et al. 2006; Tennesen 2006). Besides formulation strategies, curcumin oral absorption can also be enhanced by the concomitant consumption of piperin, The in vivo study i humans has shown a 2000% increase in curcumin bioavailai ity alter oral application of 2g curcumin together with 20mg of piperin, which inhibited intestinal and hepatic curcumi slucurenidation, leading to a higher extent of curcumin absorp- lion without posing safety issues for the volunteers (Shoba etal. 1998), The attempt ofthis stady was thus also to eval- uuste whether piperin could be used as permeability enbancer to improve curcumin absorption and not just its metabolism. Because the absorptive curcumin permeability did not change, piperin eannot be used asa permeability enhancer for cureumin. although some studies have demonstrated its capability of mod- Uulating membrane dynamics by modifying the fluidity of the apical membrane (Khajusia et a. 1998), Instead of improving curcumin absorption, piperin significantly enhanced cure ell in both tested models. As shown for FummC and BCRP Uwansportr, the increase of curcumin S-M/BL-AP permesbil- ity with piperin in the donor solution most probably resulted from positive-cooperation triggered by piperin on one of the ‘ABC transporters (most probably BCRP), involved in curcumin ellax, Based on the in vivo data, concomitant consumption fof curcumin and piperin results in higher curcumin per-oral bioavailability, Therefore, the in vivo inhibitory effect af pip fn the intestinal glucuronidation most probably exceeds the 522 positive-cooperation and inezeased the efflux of curcumin fe enterocytes, because this latter process tends o lower the amount fof curcumin absorbed. In spite of curcumin pharmacokinetic shortcomings emanat- ing {rom its low solubility, poor absorption, rapid metabolism and elimination (Hatcher el al. 2008), which render the use of curcumin as remedy’ for per-oral application quite ineffec- tive, recent studies have suggested that curcumin metabolites could be responsible for some of its systemic effects (Hatcher etal. 2008; Basnet etal 2011), There are no available curcumin ‘metabolite standards on the market, therefore pharmacological activity of curcumin metabolites has nat been sufficiently eon ‘municated yet, However, data for texabydrocurcurin suggests that this metabolite exerts more potent antidiabetic and antiox- ‘dant activities than its parent compound and that it actively inhibits the astembly of microtubule proteins, an activity devoid from curcumin Matcher etal 2008). Furthermore, studies with radiolabelled curcumin in rats have shown distribution of cur- cumin and its metabolites into the lungs, muscles, spleen liver, kidney, intestine, blood and to a lesser extent into the brain after intra peritoneal applications (Hatcher etal. 2008). These ‘organs are thus all potential sites exposed to pharmacological activities of curcumin and/or its metabolites. Namely, although ‘metabolites ar the only species f curcumin detected in measur alle levels in the plasma, after their distebution into target sites they might inter-converse back into the curcumin hy enzymes (uch as B-glucuronidases and sulftases) residing in these ts- sues (Hatcher et al. 2008). Therefore it might be beneficial not {o monitor only curcumin plasma levels, but also the amount of curcumin metabolites absorbed from the intestine. Higher amounts of curcumin metabolites absorbed could thus inerease the probability of metabolites distributing into these target sites. (Our studies with Caco-2 cells indicated that the glucuronide formed inside the cells was actively transported to both sides of| the cells. By using specific transporter inhibitors (PSC883, MK- "we have shown that the transport of glucuronide to the apical side wae most probably mediated by Pgp and MRP?2, while MRP and MRP3 were probably responsible fr the flux ‘of metabolite to the basolateral side, Since the ratio between AP-BL and BL-AP fluxes decreased inthe case of FumC, there- fore penalizing the amount of glucuronide absorbed, BCRP was ‘most probably not involved in intestinal distibution of curcumin slucuronide. However, the lower glucuronide formation rate Jn the presence of FumC might also be a consequence of the reduced parent curcumine intracellular concentration caused by greater BCRP mediated curcumin apical excretion rate due to the already discussed activating effect of FumC on curcumine BL-AP transport (Table 1), The lux of glucuronide in the secre- tory direction could additionally be inhibited by piperin, whic ‘ost probably inhibited ABC transports (.¢. Pep, MRP2) inthe apical side of Caco-2 cell, In this manner, pipetin could con- tribute to higher amounts of glucuronides absorbed, therefore ‘increasing the probability of «higher systemic pharmacological activity, ‘Owing to stability and solubility issues, the in vitro curcumin permeability assessment could only be performed with albumin Addition to the incubations salines, used for the permeability experiments. In this manner, decomposition, precipitation and non-specific plastic binding of curcumin were avoided, enabling correct permeability measurements. The permeability of cur ccumin wae pH dependent with higher values obtained under Slighlly acidic conditions. Still, the permeability of curcumin ‘was lower than the permeabilities of highly permeable FDA standards, classifying curcumin as BCS IV compound. AL iso- pil values curcumin was subjected toa profound eux back into the intestinal lumen by BCRP and to some extent probably by Pp and MRP-2, bat the binding affinity of curcumin towards Pharmazie 67 (2012)ORIGINAL ARTICLES gp and MRP-2 was higher than that of specitic Pgp/MRP2 inhibitor. Thus, curcumin could be an even more effective PEp/MRP inhibitor. Used in vivo, it could increase the absorp- tion of drugs, transported by those two ABC proteins, but these questions need furher in vivo/in vitro conformation. Mucus was found to significantly hinder the absorption of curcumin because its main constituents, glyeoproteinsilipids, most prob- ably bind free curcumin, thue preventing adequate diffusion to the enterocytes. Piperin, a permeability enfancer and inhibitor of curcumin glucuronidation, failed to increase the absorption fof curcumin, but it increased curcumin active extrusion from cells with ABC proteins, This effect is probably of no impor tance in vivo, sce clinical studies report increased cuscumin per-oral bioavailability when it was consumed with pipesin However, piperin significantly increased the amount of cu cumin metabolite transported to the basolateral compartment, therefore making this adjuvans a useful tool for increasing the circulating levels of glucuronides, therefore increasing the prob- ability of glucuronide distribution into tissues, distal fom the GIT, and consequenly ther interconversion to curcumin, Cur- ‘cumin delivery systems should therefore pursue the option of delivering curcumin to the upper, acidic part of the GIT (i. {immediate release formulations), where curcumin permeability is higher. The per-oral curcumin bioavailability could further be increased, i these formulations would contain piperin and could reversibly Ioosen/eompromise the mucus structure. 4, Experimental Jd, Materials Sal for ineubatin suites, cuteumin,duoeseeinestone--sufate (E88), MISTI fmizemotgia C (Pum), Nacel-eysteine (NACCYS) and all the neccsuy ingiedicts for cll cultivation wete from Sign Aldcich (Chemie Deseuhofen, Genny). PSC was from Tons. Allehemicals used In is study were ofthe highest grade ealabl 42, Methods 442.1. Invitro tranaport studies across Caco? cell monolayer and at mall interne Caco.2ccis were obtaned rom Destche Sammon on Mikroongaiemen ‘nd Zlatan (DSMZ) ACC 188 lt 12 and were grown on Tanevel (Conar cular isons with «polyeatbonate membrane (@hameter 12m td pore sie 4 wm). 65,00 elefilererabranes Were ed fr seding the medium Was changed! every Wo das Atay 15, tanseptela ‘leew! esstance (TEER) as mosesed foreach ler wate Care? el Sonolaers. Ifthe THER valuee were inthe cange of 300-400 fle, the (Caco-2cellmencayer wre wed forthe ubsequet esting of prmoeaiiy stday 21 ‘The Cavo-2 cells grown on Tanswell Costar culture inserts wece carefully rinsed with Ringe buffer 1 Simi and Olof bathing sohtion (Ruger be on basolateral and ape sides ofthe Caco ell monolayer, epee ‘ively, wus meitained a 37° and cominaouay oxygenated with eatbogen {95% 0, 5% CO;) and lighy shen ding exerment 625 aM gheose ‘nd 625m0M mannitol were always aed tothe basolateral (BL) and ap {al (AP) sides, respectively, to give final 1OmM concentrations. Amin {@ volvo!) wae aed to waneport buffer on oth side to ierete ct ‘Sin stabi andto sve presipaton and Gecompostion Cucamin was Sddedo the AP ie if dying apcao-baolteral(AP-BL) anspor) ot “he BL ide staying barlatera-o-apieal (L-AP) transport). Samples (050) were wihdrsr rom the aeepir side every 20min upto 20min ‘eis inside cells fom curconin, wer also meitored by sampling donor an accep comparonens teva ine posnls- Only hose Caco ‘ell nonolayers with constant TEER value during the whole expeimeas ‘were used The expeients conducted wit rt nests tase confor o ‘heLaw forthe Protecionof Animals (Republi of Slovenia and were regis teredathe Veterinary Adaaaisteation of te Repu of Slovenia Rt sal Intestin was obained from male Sprague-Duley rats (250-3204) fasted Teh peocto th expernens, After ethnasa and lapacomy, tbe imesine vas insed wit e-old [01M glucose Ringer sltion-Tejnum, located 25-6em distally rom the pyere sphincter was used forthe experiments, Pharmazie 67 (2012) ‘The intestinal tissue was ext into 3m long segments, excluding visible Peyer's patees. These intestinal Segmets were opened along te meen fee border, selehoe onto ise with an expoted toe agen of Le" nd then placee between two compartments of EasyMosotsie-y-rige ‘iason chambers (Pyrilogc Instruments, Sen Diego, USA). The exper Snel proce entnedin he ane rater at ecb fr Cao? 422, Electrical measurements ‘The tfsion chambers were equipped with wo pars of Ag/AgCleleeodes formessring tanepthelial potential dffeeace (PD) se hon cea se ent (I) sth 4 mil channel voltagecuetelanp (model VEC MCE, Physiologie Insrumente). The vit and neg ose was checked iby of rt small ntetine wae addonally checked by recording the Incretst of le and PD after the adion of ack cote solution tthe ‘coral eompartnet athe eb f experiment (al lucseeoncetation wa 25m), Tso segments wee considered viable if the PD vale after (he addition of glucose was lower thas 1 Om and ifthe average TEER ‘ues corded dung the expen wore between 20 abd 40 Re 425. LOMSMS anaes of curcumin and curcumin metabolites Sanples obtained during the permeabiy experiments were precipitated wthice cold MeOH (3 yatNob, vigoouly vexed and left at 20°C {or 48h, Afterwards he samples were cetafuged (ISmua at 4°C and 15000. the supecnatat was asafened iat autosampler S6~well plate ey for Subsequent asalyis by LC-MS/MS. LCIMSIMS apparatus eon Sised of an Agilent 6160 ule quadruple mass specuonetr equipped ‘witha JeSteu intesfuce and connected wo aa Agileat 1290 lity UPLC (Agiea Techologis, Sane Clara, USA), For lromutograpie separa, 1 Phenomeaes Kinetex 50 201mm C-18 eolamn with 2 mm pacts ‘was used (Pheaomene, Torance, USA) The injection volume was 2 ul Sad the column temperature wat 50°C. The mobile phase contd of| ‘Tinea gradient of water (A) and actos (8), both containing 0.5 forme aie The gradient started with OT B which serene linet 0 S08 B over 2min and thea reared to orginal condivons in 5. The ow {ate was 0 631nl/min and the otal time of analysis wat 27 tin The mass {hzer 4 PST (31s 10" Pa, sheath ga empeatre: 320°C, sheath at low: 11 Lima cyplay eotrance valle: 100 V, norte vltage 1000, deta EMV. 200 VThe MRM sed forthe uanticuton of cuceuine and ‘tscursne glucuronide wer follow: Tae mull eaten enoniteing Inftranatons and cllato eneeies used fer he quabeatin feueu {be and eureume glacwonide Were 69> 77 at 16eV abd 545369 a BeNrespecsvey ‘The apparent permeability ceetcent (Py) of ewcumine and axes) af ‘cline metab wee calculated aPeordig to.) and Ba) ere dor and at nod substance/metboltes inthe aceepodonor compare per ni tise ‘der eady sate contin Vite volume ofthe seeptr compare, Ae exposed erface area (em fort eon sn 113m for Cacoe3 Mlrwards, 2d tent lest =0.08), were sed Oiherwie oe Way ‘ANOVA, followed by Bonin posthoc test were sppliee. Arsaston Palm. Lathman K (2001) Caco-? monolayers in experimental and theoretical predictions of dug anspor. Ady Drug Delv Rev 46: 45. 523‘ORIGINAL ARTICLES, ‘ums, Stross A, ol (201) The hereto Tones gel layer cna dpa sate rio, Gueite Lier Phil 280: 6522-0829. amet PSiakocBenet 2011) Careamin: An ansnlammatory ‘mol om uy spice o he pa aes eaten Moles Tes 4557-$598 Begin K, Mise 1 Kit A 200 Gace Bavonoid and ecganonls ‘compos: impact on the hpi parmacoaets of agus and ‘anor Drag Mab Phsaokin 5581-59, Dillon N. Agginval BE, Newnan R, Wolf RA, Keanomakiors AB ‘Abbe Ik, Np CS, BatnactV, KaceoekK (2008) Phase I ‘termini patieas with advanced paces cancer. Cin Cancer ck H, Sth Tajo M,Obdo S, Oa. Sada ¥ (2006) “fet of ha exacts on he fncton of hamanoran-aion an pong polypepie OATP-B. Drag Met ison 84 577-582 Iiaever Ht Pony R. Cho J, Tom FM, Tort SW” 2008) Coram ‘tom ancient meine to cet ciel wa, all Mol Lie Se 65 tiles. Hoang Wang ¥. Grin (2008) ATP-dependent trngprtfenmattin Unmembrane vse expressing Beat Cancer Resistance rte, Dig Mala Dospon 34738072 Ieson C. On 5, Jones DIL, Vascoyle Rit CK, Lao 1, Howells L PnmerS Jake, Wiliam M, Steward WP, Greer A (001 Care teationofmetablesofthe chemopreventive get cucu nan ‘dat hepatocyte and in thera i, and vation of toil nerd prowaaninprodtin Cancer Rex 1 1058107 tkcion R Tene DIL, Or, Cough MH, Boosock D, Wilms ML. Farmer PB, Stenad WR, Gesher AJ 2002) Metabolism of the en Epidemiol Binur Pre I 10S-LLL Keminaga ¥ Neato A. Akyait Sopot N, Yaar, Matans “7 Miroku H (2003) rodetin of ussutral gcondes of cueunia ‘wih drt enbaned wate sby yell opemion cle of ‘Carats onus HERS Lees 595.511316, Kecson J, Walkman 8, Arron (1595) The mines lye «baie tog shaorptin in managers of bomen esl pia TES feblet el at) Par 9: 209-218 tigre A, Zac, Rei RL. 993 Permeability cartes of ipein “ra sbortion an ative lad om peppets nds ioral shane fan J ep Bil 36: 56250, Legeal Keil A GO01 Conparaive permeability osm aeyeloi deriv ‘ie hong mtv micas an ero macn dperione Drag Dev Ind Pharm 3708-574 Lege I. Kit A 200 Factors affcting the micotinute pH of te rt "Phan Ringer carbonate ule Mil Pars Ball 96-4, Lis A Low H, Zhao L, Fan P QUOG) Vaied LMS asey for tatcnin and ay dorm na pea sd apbeaion op ‘bucdice sy of ond compl of eweunia. J Phmerat Biomed: 720-72. Meaney €, © Deicall © (1999) Macs «bait the peel ‘ypc an pope compounds ia the abecce abd presence of sodium nwocholte milla yrs eng cll cate mode. Bus} Foam Se 1-173, Nozawa rt K Neva I, Toy A, Tana (200) Panctione etc: Tran of piso cane ain anioring poeple ORTPD Inhuman J Poaneasl 308 430-5 ‘aa RA, Baden SA, Paton SL, Cooke DN, Shay A, Hewit HR. Mrcylo TH Morgan, HemingnyD, Pomme SM, Poa Geacer AJ, Stead WD (20) Pas clea lo rl uu ‘mates of ssemie sei and comane. lin Cancer Ree 0: surat Starns FF 208) ARC malig anspor: rcs fon se le inchemerelace Par 05-17 soba G Joy b, Jeph Mel M, Redan R Siinivas PSSR (198) inuece of penn on he sims of cin in sale and dna voles Phat ed ot 252336 “Toonesn HH, Mason M, Lofsion T2002 Ste of earcenin ander uniedsXXVIL Cycodxtia completion ably, cane and Phtoceniel subi Ia) Pharm 2127-135 “oeeesen HE (200) Soil te ably of cui i schaione coving gina andor Vscosty modivngmacomolrls Pha ‘illngB, Pavar VB, Baral AK (2011) Kemieton of permeaiiy nae bales rl livery of cree rng Caco? cl ode Bar $PhansBiopars 7: 275-282 ‘Wilman Lathe A, Bk (1998) The nen of its mics com ones on be dui of rugs, Phar Res 15.6671 ‘Woreier IM, Ura Man dr Velde Ak, Bossa MG, Spek Man Zande, Risjets IMO, va laden PCa SP 203) Iotrplay between MRP inhibition snd metsblim of MRP nko the ut of crs Chem Ker Tiel 112-1051 ‘WLX, Amidon GL, Poll JE, Zino H, Mek MN, Conse DE, Shab VE ‘ako El, Chan Mt, Lee VHL 2002) Biophumacstal Scan ton sates he set ass fr bowaiveretesons, Pac Re 9 Sans ng W, Lit LY (2008 ects of spice contents on Feyeepoen tethicd anor and CYPSAC mediated meabli vie.Drg Maah Dpon 36 1281290. Pharmazie 67 (2012)