Printer Friendly

Intestinal formation of trans-crocetin from saffron extract (Crocus sativus L.) and in vitro permeation through intestinal and blood brain barrier.


Aims: Extracts of saffron (Crocus sativus L.) have traditionally been used against depressions. Recent preclinical and clinical investigations have rationalized this traditional use. Trans-crocetin, a saffron metabolite originating from the crocin apocarotenoids, has been shown to exert strong NMDA receptor affinity and is thought to be responsible for the CNS activity of saffron. Pharmacokinetic properties of the main constituents from saffron have only been described to a limited extent. Therefore the present in vitro study aimed to determine if crocin-1 and trans-crocetin are able to pass the intestinal barrier and to penetrate the blood brain barrier (BBB). Additionally, the intestinal conversion of glycosylated crocins to the lipophilic crocetin had to be investigated. Experiments with Caco-2 cells and two different porcine BBB systems were conducted. Further on, potential intestinal metabolism of saffron extract was investigated by ex vivo experiments with murine intestine.

Methodology: In vitro Caco-2 monolayer cell culture was used for investigation of intestinal permeation of crocin-1 and trans-crocetin. In vitro models of porcine brain capillary endothelial cells (BCEC) and blood cerebrospinal fluid barrier (BCSFB) were used for monitoring permeation characteristics of trans-crocetin through the blood brain barrier (BBB). Intestine tissue and feces homogenates from mice served for metabolism experiments.

Results: Crocin-1, even at high concentrations (1000 p.M) does not penetrate Caco-2 monolayers in relevant amounts. In contrast, trans-crocetin permeates in a concentration-independent manner (10-114 [micro]M) the intestinal barrier by transcellular passage with about 32% of the substrate being transported within 2 h and a permeation coefficient of [] 25.7 x [10.sup.-6] [+ or -] 6.23 x [10.sup.-6] cm/s. Trans-crocetin serves as substrate for pGP efflux pump. Trans-crocetin permeates BBB with a slow but constant velocity over a 29 h period (BCEC system: [] 1.48 x [10.sup.-6] [+ or -] 0.12 x [10.sup.-6] cm/s: BCSFB system [] 3.85 x [10.sup.-6] [+ or -] 0.21 x [10.sup.-6] cm/s). Conversion of glycosylated crocins from saffron extract to trans-crocetin occurs mainly by intestinal cells, rather than by microbiological fermentation in the colon.

Conclusion: The here described in vitro studies have shown that crocins from saffron are probably not bioavailable in the systemic compartment after oral application. On the other side the investigations clearly have pointed out that crocins get hydrolyzed in the intestine to the deglycosylated trans-crocetin, which subsequently is absorbed by passive transcellular diffusion to a high extend and within a short time interval over the intestinal barrier. Crocetin will penetrate in a quite slow process the blood brain barrier to reach the CNS. The intestinal deglycosylation of different crocins in the intestine is mainly due to enzymatic processes in the epithelial cells and only to a very minor extent due to deglycosylation by the fecal microbiome. On the other side the fecal bacteria degrade the apocarotenoid backbone to smaller alkyl units, which do not show any more the typical UV absorbance of crocins. As previous studies have shown strong NMDA receptor affinity and channel opening activity of trans-crocetin the use of saffron for CNS disorders seems to be justified from the pharmacokinetic and pharmacodynamic background.


Crocus sativus L.




Blood brain barrier



Saffron, the dried stigmata of Crocus sativus L., has been used in traditional medicine (Ferrence and Bendersky 2004) against spasms, bronchospasm, menstruation disorders, liver disease, pain, insomnia, digestive ailments and as stimulant, aphrodisiac, antidepressant and for supportive treatment of cancer (Schmidt et al. 2007). Several clinical pilot studies proved significant antidepressant effects of saffron extracts; a recent meta-analysis, including 5 randomized, controlled studies, indicated significant improvement of symptoms in patients with major depressions (Hausenblas et al. 2013). Recent studies have shown neuroprotective effects of saffron extract with trans-crocetin --the deglycosylated metabolite of crocins--being the active ingredient. This cytoprotective activity of trans-crocetin is due to the high affinity of this apocarotenoid to the phencyclidine binding site of the NMDA receptor (Lechtenberg et al. 2008). This antagonistic effect at central NMDA receptors leads to an inhibition of postsynaptic potentials and inhibition of glutamatergic transmission, as recently shown in rat cortical brain (Berger et al. 2011). Besides the antidepressive effects saffron extracts and especially its main secondary metabolite crocin has been used for the treatment of coronary diseases (for review see Kamalipour et al. 2011; Joukar et al. 2010) due to its potential antihyperlipidemic effects (Lee et al., 2005), inhibition of cardiac [Ca.sup.2+] channels (Boskabady et al. 2008) and pronounced antioxidant reactivity (Mashmoul, 2013), Main secondary metabolites of saffron extracts are about 0.5% of volatile compounds with safranal as main constituent (Schmidt et al. 2007; Lechtenberg et al. 2008). The typical spicy taste of saffron originates from picrocrocin (5-15%), the biosynthetical precursor of safranal. The main constituents of saffron constitute a series of apocarotenoids, typically carotenoid-glycosyl esters of a [C.sub.20]-dicarboxylic acid, the so-called crocins (Fig. 1). Crocins account for up to 30% of dry weight and are responsible for the intense yellow color of saffron (Schmidt et al. 2007; Lechtenberg et al. 2008). Non-glycosylated [C.sub.20]-dicarboxylic acids, with trans-crocetin as lead compound can be found only in traces in saffron of good quality (Lechtenberg et al. 2008).

From the pharmacokinetic point of view it has been assumed that crocin is excreted largely through the intestinal tract following oral administration; on the other side the intestinal tract serves as an important site for crocin hydrolysis (Xi et al. 2007) and bioavailability of isolated crocetin after oral administration has been shown in healthy volunteers (Umigai et al. 2011). Xi et al. (2007) showed that orally administed crocin is not absorbed in rats after a single dose or repeated doses and crocin is excreted largely through the intestinal tract following oral administration. Until now it is still unclear to which extent crocins are metabolized in the intestinum to the deglycosylated crocetin, in which gastrointestinal compartments this occurs and whether this is due to fecal metabolism. The following study aimed to investigate the underlying permeation mechanisms across the intestinal epithelial barrier by using Caco-2 cells (Hubatsch et al. 2007) and to clarify if trans-crocetin permeates the blood brain barrier under in vitro conditions by using two different blood brain barrier models (porcine blood capillary endothelial cells, BCEC (Franke et al. 2000) and blood cerebrospinal fluid barrier, BCSFB (Angelow et al. 2004)).

Materials and methods


Saffron, grown in Iran, purchased from the International Mumbai Saffron Trading Exchange, batch No. 7002, was identified by macroscopic and microscopic investigation against a saffron reference sample. Full analytical characterization was performed by UPLC[R] according to Lautenschlager et al. (2014). A voucher specimen (IPBP291) has been deposited in the documentation center of the Institute of Pharmaceutical Biology and Phytochemistry, Munster, Germany.

Standardized hydroethanolic saffron extract (SE), trans-crocin 1 and trans-crocetin were prepared as described recently (Lautenschlager et al. 2014).

If not stated otherwise, solvents, reagents and consumables were obtained from VWR (Darmstadt, Germany). Propranolol HCl was obtained from Acros Organics (Geel, Belgium), Lucifer Yellow CH dilithium salt, DAPI and EDTA from Sigma-Aldrich Chemie (Buchs, Switzerland), Verapamil HCl from Fluka Chemie (Buchs, Switzerland), TexasRed[R]-X Phalloidin from Molecular Biotechnology (MoBiTec, Gottingen, Germany). FBS Hyclone[R] was purchased from Thermo Fischer Scientific Inc. (Waltham, MA, USA), NEAAs were from GE Healthcare (Chalfont St Giles, Buckinghamshire, UK), Ezetimibe from Molekula GmbH (Munchen, Germany) and all other components for in vitro cell culture were from PAA Laboratories (Pasching, Austria). Tissue culture plastics were supplied by Sarstedt (Numbrecht, Germany). Transwell[R] plates (12 well plate, 12 mm diameter, 0.4 [micro]m pore size, polycarbonate membrane) were obtained from Corning Costar[R] (New York, USA). Test solutions were sterilized prior use by 0.2 [micro]m syringe filter polypropylene (VWR, Darmstadt, Germany).

Caco-2 in vitro cell culture

Cultivation of Caco-2 cell line (ECACC, European Collection of Cell Culture, Salisbury, UK, LOT 07/G/006; 6/08/07 P46), absorption studies as well as respective evaluations were performed according to Hubatsch et al. (2007) (with minor modifications by Zumdick et al. (2012)) and quantitative calculations were carried out according to Tavelin et al. (2002).

Cells from passage number 70-82 were maintained in 75 [cm.sup.2] plastic culture flasks in 10 ml of Dulbecco's Modified Eagle's Medium, supplemented with 10% FBS Hyclone[R], 1% NEAA and 1% penicillin/streptomycin (100x), cultivated in a humidified incubator at 37[degrees]C, 5% C[O.sub.2] until 70-90% confluency was achieved.

For subcultivation the cell monolayer was treated for 5 min with 0.25% trypsin/0.2% EDTA. Cell counting was performed by electric current exclusion (CASY[R]cell-counter, Scharfe Systems, Germany).

For transport experiments cells were seeded at a density of 2 x [10.sup.5] cells/ml on Transwell[R] plates and grown for 21-23 days to obtain a dense, well differentiated cell monolayer. During cultivation the culture medium was changed every 2-3 days.


Influence of test compounds on viability of Caco-2 cells was determined by MTT assay (Mosmann 1983) over 24 h incubation period. Transepithelial electrical resistance (TEER) was determined by EVOM resistance voltohm meter (World Precision Instruments Inc., Sarasota, USA) at 37[degrees]C, 2-3 times a week for monitoring and determination of the integrity of the cell layer. The background resistance values of the support filters without cells were subtracted from the resistance of the support filters containing cells. Only cell monolayers having TEER values >150 [ohm] [cm.sup.2] were used for transport experiments.

Morphology of cell monolayers was controlled by confocal laser scanning microscopy after staining f-actin with Texas Red[R]-X Phalloidin and cell nuclei with DAPI (Zippel et al. 2009).

Cytotoxicity of test compounds against Caco-2 cells

Cytotoxicity of test compounds was determined by MTT assay (Mosmann 1983) using Caco-2 cells in 96 well plates at a density of 20.000 cells per well in 200 [micro]l FBS-free medium, grown for 96 h and followed by 24 h contact time with the test compounds (100 [micro]l of serum-free media containing SE, trans-crocin-1 or trans-crocetin at different concentrations) and incubation at 37[degrees]C/5% C[O.sub.2]. The incubation solutions were aspirated, each well was washed twice with 150 [micro]l of PBS and 50 [micro]l of MTT solution were added (2.5 mg/ml in PBS). Supernatants were discarded and the formed formazan was dissolved in 50 [micro]l of DMSO. The absorption of the resulting solution was determined at [lambda] = 492 nm against reference wavelength [lambda] = 690 nm (BioTek EL 800; BioTek Instruments, Bad Friedrichshall, Germany).

Caco-2 transepithelial transport experiments

All experiments were performed at 37[degrees]C. Cultivation media and all solutions were preconditioned to 37[degrees]C prior to adding them to the cells. Before starting the transport experiment (21-23 days post-seeding) cultivation medium was aspirated from the Transwell[R] inserts. The cell monolayer was carefully flushed with [HBSS.sup.*] (*: supplemented with HEPES 5.94 g/l and pH adjusted to 7.4 with NaOH 2 M) and incubated with [HBSS.sup.*] for 30 min. Determination of TEER values was performed at 37[degrees]C directly before addition of test compounds. Test compounds, dissolved in [HBSS.sup.*], were added to the respective donor compartments (for apical (A) [??] basolateral (B) transport 0.5 ml; for B [??] A transport 1.5 ml). [HBSS.sup.*] was added to the acceptor compartment (A [??] B transport 1.5 ml; B [??] A transport 0.5 ml). The incubation time for all transport experiments was 2 h with the exception of validation studies with [sup.14]C-glucose (0.5 h permeation study). Samples (half of the volume of the acceptor compartment, 0.75 ml for A [??] B; 0.25 ml for B [??] A transport) were taken all 30 min from receiver compartments and refilled with preconditioned [HBSS.sup.*]. In case of permeation studies with crocin-1 only one sample was taken after 2 h, because of very low cellular permeability.

After the last sample had been taken (from donor and acceptor compartment) the cell monolayer was aspirated, washed with [HBSS.sup.*] and incubated for further 30 min at 37[degrees]C in [HBSS.sup.*], followed by determination of TEER values at 37[degrees]C to ensure integrity of the monolayer throughout the experiment. Inserts with TEER values <150 [ohm] cm (resistance values of empty filters were subtracted) were not included in the final evaluation. Test samples were directly analyzed by UPLC[R], stored at 15[degrees]C with light protection.

The apparent permeability coefficients [] [cm/s] for A [??] B and B [??] A experiments were calculated from the slope obtained from the linear curve-fit of the following equation: "weighted normalized cumulative amount of transported drug" (Tavelin et al. 2002)

1/A x [i.summation over (k=1)] [CR(tk) - f x CR(tk - 1)] x VR/[CD(tk - 1) + CD(tk)]/2 = P x ti

A is the surface of the filter [[cm.sup.2]], CR(tk)/CD(tk) is the receiver/donor-concentration at sampling time k [[micro]M], VR is the receiver compartment volume [[cm.sup.3]], and/is the sample replacement dilution factor (f = 1 - ([V.sub.sample]/VR)).

Preparation of trans-crocetin containing test solutions

Solving trans-crocetin in the incubation buffer (HBSS*) is problematic due to the physicochemical properties of the compound. Thus, trans-crocetin containing solutions were prepared by adding 0.8 mg of trans-crocetin to 10.0 ml of HBSS*, followed by 5 min of sonication and filtration with a 0.2 [micro]m filter. This procedure resulted in test solutions with trans-crocetin concentrations of about 160 [micro]M. Exact quantification of the content was performed by UPLC[R] analysis of the solutions (Lautenschlager et al. 2014). Subsequently, the stock solution was diluted by HBSS* to the desired concentration.

BBB transport experiments: BCEC and BCSFB models

In vitro experiments with two porcine BBB models were carried out according to Franke et al. (2000) and Angelow et al. (2004). Primary cells of brain capillary endothelial cells (BCEC) and epithelial cells of the Plexus choroideus (BCSFB) were isolated from fresh porcine brain, cultivated in cell culture flasks and seeded on Transwell[R] plates for transport experiments (BCSFB seeded directly on filters). These two in vitro models, mimicking the transition barrier into the CNS, were validated by TEER using impedance spectroscopy (CellZscope, Nanoanlytics GmbH, Munster, Germany) directly before and after the transport experiment to ensure optimal barrier properties. TEER values of all experiments matched the specification of 600-900 [ohm] cm.

Sample solutions were prepared the day before the transport experiment in serum free transport buffer. During experiments it became obvious that longer incubation times were needed to obtain valid, measurable amounts of trans-crocetin in the respective acceptor compartments. Thus, the transport experiments were extended to 29 h instead of 2 h as routinely performed. Samples were drawn from the acceptor compartments after 4, 7, 23 and 29 h (half of the compartment = 0.75 ml (BCEC), 0.25 ml (BCSFB), refilled with preconditioned buffer), stored at 20[degrees]C and analyzed by UPLC[R] the day after the experiment.

Mechanistic studies of intestinal transport: p-glycoprotein (g-pg), paracellular permeability, cholesterol transporters

Influence of p-GP on the transport of crocin-1 and trans-crocetin was investigated by addition of p-GP inhibitor verapamil (50 [micro]M) to the transport buffer. Cell monolayers were preincubated for 30 min. A [??] B transport studies were then performed following the method described above.

Influence of cholesterol transporters on the transport of transcrocetin was investigated by addition of cholesterol-transporter-inhibitor ezetimibe (15 [micro]M) to the transport buffer. Monolayers were preincubated in different experiments with ezetimibe for 30 min and for long term studies for 1 and 14 days, followed by A [??] B transport as described above.

For investigation of paracellular permeability of trans-crocetin, EDTA (3.2 mM) was added to the transport buffer. Cells were preincubated for 15 min A [??] B transport experiments in presence of EDTA were performed as described above. Transport experiments in reverse direction were conducted in order to investigate transcellular permeability.

Degradation experiments with simulated gastric/intestinal fluids

Simulated gastric fluid (SGF) and simulated intestinal fluids (SIF) for in vitro degradation experiments were prepared according to United States Pharmacopoeia 34 (USP 34, NF 29; 2011; SGF page 968, SIF page 969). Test compounds were incubated with SGF and SIF (100 mg of SE dissolved in 300 ml) in a dissolution paddle apparatus (Erweka DT600, Erweka GmbH, Heusenstamm, Germany) at 37[degrees]C and 100 rpm for 90 min. Aliquots were analyzed by UPLC[R] immediately after incubation.

Metabolism experiments with mice intestine homogenate

Use of isolated organs from mice corresponded to the German Law for Animal Testing (acceptance No. 84.- Freshly prepared mice intestine was obtained from C57B1/6N mice (Charles River, Sulzfeld, Germany), age 3 months to 1 year, fed with dry feed and water ad libitum and kept in standardized housing in 12 h night-day-cycle.

Freshly prepared and cleaned small intestine from mice was sliced to small pieces with a scalpel and homogenized with a Potter-Elvejhem homogenizer (Braun, Melsungen, Germany) on ice by 20 up and down strokes. The homogenate (0.5 g) was diluted with 10 ml of HBSS and centrifuged 500 x g, 10 min, 4[degrees]C (Centrifuge 5430 R, Eppendorf AG, Hamburg, Germany) to obtain a protein-enriched supernatant (Nevada Proteomics Center 2014).

Incubation experiments of SE with the protein-enriched supernatant were performed in incubation buffer, consisting of HBSS buffer, supplemented with D-glucose (1 mg/ml) and Carbogen gas.

2.0 ml of protein-enriched intestinal supernatant were added to 10 ml of the incubation buffer in 15 ml Falcon[R] tubes; the mixture was incubated at 37[degrees]C under gentle shaking (200 rpm) in a thermo-shaker (Infers HT Ecotron, Infers, Bottringen, Switzerland). Antibiotics, 0.24 ml of gentamicin (10 mg/ml) and 0.24 ml of penicillin/streptomycin (stock solution 100x, PAA Laboratories, Pasching, Austria) were added to 12 ml of incubation mixture and preincubated without SE for 2 h, samples without addition of antibiotics were treated the same. 12 mg of SE were added to each tube and this time point was designated as [t.sub.0]. Incubation time was 5 h and samples (0.5 ml) were taken every 60 min. Samples were mixed with 0.5 ml of acetonitrile, centrifuged (15.000 x g, 10 min, Mikro 120, Hettrich, Tuttlingen, Germany) and analyzed by UPLC[R].

Metabolism experiments with mice fecal homogenate

Appendix and colon from freshly sacrificed mice were separated from the intestine. The content (feces) was scratched out by a glass rod, firmer parts minced to small pieces and homogenized with 10 ml of HBSS to obtain a feces-enriched supernatant. Incubation with SE was performed as described in 2.9 without addition of antibiotics.

Quantification by UPLC[R] and fluorometer

UPLC[R] analysis was performed using Waters Acquity Ultra Performance LC with PDA e[lambda] Detector at 200-500 nm, Sample Manager at 15[degrees]C with light protection, Binary Solvent Manager, Software; Empower pro (Waters Corporation, Milford, MA, USA). Chromatography was performed at 70[degrees]C with a flow rate of 0.673 ml/min on a Waters BEH[R] shield RP-18, 1.7 [micro]m, 25 x 100 mm stationary phase with a binary step gradient of A (0.1% TFA in Millipore water) and B (acetonitrile): 0.14 min isocratic 90% A [right arrow] 1.82 min 50% A [right arrow] 2 min 0% A, [right arrow] 0.51 min isocratic 0% A [right arrow] 1.2 min 90% A [right arrow] equilibration 90% A for 1.33 min. The injection volume for saffron compounds was 1 [micro]l and for propranolol-containing solutions 5 [micro]l.

During the transport experiment samples were stored at 4[degrees]C, filtered through 0.2 [micro]m filter and subjected to UPLC[R] (15[degrees]C, light protection).

Quantification was performed by external standard calibration for trans-crocin-1, trans-crocetin and propranolol using the respective reference standards.

Samples containing Lucifer Yellow CH were determined by Microplate Fluorometer (Fluroscan Ascent FL, Thermo Scientific, USA) after calibration with respective external standard at [[lambda].sub.ex] 485 nm and [[lambda].sub.ex] 588 nm.

Results and discussion

Influence of SE, crocin-1 and trans-crocetin on Caco-2 cells

To ensure unchanged viability of Caco-2 cells throughout the transport experiments, cellular mitochondrial dehydrogenase activity of Caco-2 cells was measured by MTT assay (Mosmann 1983) after a 24 h incubation period with the test compounds: Hydroalcoholic saffron extract SE (0.5-1 mg/ml) and crocin-1 (250-1000 [micro]M) revealed no negative significant changes in cellular viability. Trans-crocetin at 10 [micro]M level did not change viability while higher concentrations (40-160 [micro]M) reduced significantly cellular viability (Supplementary Data, Fig. S1). Therefore, the following Caco-2 absorption experiments were performed only over 2 h incubation, not to induce negative effects on the epithelial cells, which was controlled by TEER determination after the experiment was carried out.

Validation of Caco-2 cell system

Prior to permeation studies, the Caco-2 cell system was validated by different parameters. Investigation of cell morphology by confocal laser scanning microscopy after fluorescent staining of nuclei (DAPI) and F-actin (Texas Red X[R] phalloidine) indicated optimal formation of cell monolayers, sufficient cell morphology within 3 axis and the absence of any cellular disturbance (e.g. formation of "dome-"like multilayers), which could negatively influence cell physiology and transport characteristics. This procedure was also carried out after the respective transport experiments had been performed with the cell monolayers, indicating that the handling procedure during the absorption experiments did not negatively influence the cell monolayer.

TEER values for monolayers (about 500 [ohm] [cm.sup.2] after 5 days of incubation) indicated sufficient cell integrity. Additionally TEER values had to be >150 [ohm] [cm.sup.2] after the respective transport experiments, indicating that handling during the experiment did not destroy monolayer integrity.

Cellular integrity and permeability was determined by using Lucifer Yellow CH, a hydrophilic compound only transported through the paracellular route via tight junctions and not by crossing the cell membrane by using the transcellular route. As expected and described in literature (Hidalgo et al. 1989) only low permeation coefficients of [] 1.71 [+ or -] 0.6 x [10.sup.-6] cm/s (1.7% transported throughout 2 h incubation) were found at 20 and 35 [micro]M, indicating sufficient cell integrity and correct formation of tight junctions.

Propranolol (170 [micro]M) was used as a marker compound with high [] (Artursson and Magnusson 1990). As expected [] of 7.10 [+ or -] 1.9 x [10.sup.-5] cm/s (Fig. 2) was in good accordance with published data (Yee 1997).

For validation of active transport the permeation coefficient of [sup.14]C-glucose was determined with [] of 2.5 [+ or -] 0.2 x [10.sup.-5] cm/s (Zumdick et al. 2012), a value expected for cells with sufficiently expressed glucose transporter (Hilgendorf et al. 2000).

These data show high correlation with data from literature and prove validity of the permeation system.

Apocarotenoids are known to be sensitive against oxidative stress and absorption to surfaces. Therefore, stability of crocin-1 and trans-crocetin containing test solutions was investigated by UPLC[R] over a 2 h period at 37[degrees]C (conditions similar to the cell culture absorption experiments). Data indicated less than 15% loss during the investigated time frame, mostly due to absorption to plastic material of the transportation chambers.


Absorption of crocin-1 by Caco-2 cells

Permeation experiments with crocin-1 at low concentrations of 10 and 100 [micro]M after 2 h of incubation revealed that crocin-1 had been transported into the basolateral acceptor compartment in concentrations below the limit of quantification (0.5 [micro]M, corresponding to 0.49 [micro]g/ml). At higher concentrations (1000 [micro]M) very low permeation of crocin-1 was observed (A [??] B, [] = 2.1 [+ or -] 0.8 x [10.sup.-7] cm/s), which means that only 0.34% of the applied crocin-1 did cross the Caco-2 barrier into the acceptor compartment (Fig. 2).

Similar data (not shown) were obtained when the permeation experiment was performed in reverse direction B [right arrow] A ([] = 3.3 [+ or -] 0.1 x [10.sup.-7] cm/s, n = 2).

Addition of EDTA to the A [??] B transport system slightly increased the permeation of crocin-1 ([] = 5.6 [+ or -] 0.7 x [10.sup.-7] cm/s, 0.86% crocin- 1, n = 2) (data not shown). EDTA is known to open the paracellular permeation route (Tomita et al. 1996).

Also verapamil, an inhibitor of efflux pumps, did not influence the absorption characteristics ([] = 2.0 [+ or -] 0.9 x [10.sup.-7] cm/s, 0.26% crocin-1, n = 3) (data not shown).

From these data it can be deduced that crocin-1 will not be absorbed after oral ingestion into the systemic compartment in relevant concentrations, which confirms the finding of Xi et al. (2007) that crocin, orally administered to rats, is not absorbed.

Absorption of trans-crocetin by Caco-2 cells

A [??] B permeation studies of trans-crocetin in the concentration range from 10 to 114 [micro]M revealed that the unglycosylated lipophilic compound is able to permeate the Caco-2 barrier with a high velocity ([] = 2.6 [+ or -] 0.6 x [10.sup.-5] cm/s as a mean value of 9 different concentrations), which can be correlated to a permeation rate of about 32% (Fig. 2). The permeation was not concentration-dependent in the range from 10 to 114 [micro]M (higher concentrations could not be tested because of limited solubility). Permeation of trans-crocetin was less than that of the positive control propranolol (170 [micro]M) ([] = 7.1 [+ or -] 1.9 x [10.sup.-5] cm/s, 67% permeation).


For investigation of the mode of permeation of trans-crocetin and to clarify if the absorption route is via the paracellular or transcellular pathway the following experiments were performed (Fig. 3).

Transport experiments of trans-crocetin (20, 72 [micro]M) from basolateral to apical (B [??] A) resulted in significant slower permeation ([] = 2.5 [+ or -] 0.2 x [10.sup.-5] cm/s) when compared to apical to basolateral transport (A [right arrow] B) ([] = 3.0 [+ or -] 0.2 x [10.sup.-5] cm/s). From this experiment it can be concluded that passive transcellular diffusion (reverse direction only slightly but significantly slower) is the main pathway for crocetin absorption.

Preincubation of the monolayer with EDTA (3.2 mM) revealed no increase in transport velocity of trans-crocetin ([] = 3.1 [+ or -] 0.1 x [10.sup.-5] cm/s without EDTA, compared to [] = 3.2 [+ or -] 0.8 x [10.sup.-5] cm/s with EDTA). This finding again demonstrated that trans-crocetin is not absorbed via the paracellular route.

Addition of verapamil (50 [micro]M), an inhibitor of p-glycoprotein (Pauli-Magnus et al. 2000) to the transport buffer resulted in significant higher permeation of trans-crocetin at two different concentrations (24, 62 [micro]M). Without verapamil 35% permeation with [] = 3.0 [+ or -] 0.3 x 10'5 cm/s was determined in comparison to 41% permeation and [] = 3.6 [+ or -] 0.6 x 10-5 cm/s after addition of verapamil (Fig. 3), which indicates that trans-crocetin may serve as a substrate for efflux pumps of the ABC-transporter family. On the other side this should not be overestimated because it is known that Caco-2 cells have a slight overexpression of the efflux transporters compared to primary intestinal cells.

Because it is known that carotenoid transport into enterocytes is mediated by the cholesterol transporters SR-BI, NPC1L1 and ABCA1 (Reboul and Borel 2011; During et al. 2005), permeation experiments with trans-crocetin (156 [micro]M) were performed in the presence of ezetimibe (15 [micro]M), a known inhibitor of these transporters (During et al. 2005). As shown in Fig. 3 no significant differences were observed between Caco-2 groups with and without ezetimibe addition. Also 1 and 14 days pretreatment of Caco-2 cells with ezetimibe did not change the respective [] values for frans-crocetin, which indicates that trans-crocetin is not a substrate for this class of transporters.

These findings indicate that trans-crocetin is not transported via the paracellular route but mainly permeates by passive transcellular diffusion. The fact that different concentrations of trans-crocetin exhibit no significant differences in the [] value (Fig. 2) suggests that no specific transporter is involved. On the other side it has to be taken into account that higher concentrations of trans-crocetin could not be tested because of very limited solubility of this compound, so that typical transporter saturation effects could not be determined.

Permeation of trans-crocetin through BCEC and BCSFB models

Porcine brain capillary endothelial cells (BCEC), a validated blood brain barrier model (Franke et al. 2000), was used for determination of trans-crocetin transition through the barrier system (Fig. 4A). Permeation of trans-crocetin through BCEC was monitored over a 29 h period. At 27 [micro]M level a slow, but constant transport velocity was determined (Fig. 4A). The respective permeation coefficient was calculated with [] = 1.5 [+ or -] 0.1 x [10.sup.-6] cm/s. Verapamil (50 [micro]M) as efflux pump inhibitor did not have any influence on the permeation and the respective permeation coefficient was calculated with [] 1.4 [+ or -] 0.1 x [10.sup.-6] cm/s.

A further in vitro BBB system based on porcine primary cells from the Plexus choroideus (BCSFB) (Angelow et al., 2004) was used to investigate the permeation characteristics of trans-crocetin at 50 and 100 [micro]M over a 29 h incubation period, leading to even higher [] values as found within the BCEC system (Fig. 4B). For 50 [micro]M transcrocetin a [] value of 3.9 [+ or -] 0.2 x [10.sup.-6] cm/s was determined while higher concentrated solutions at 100 [micro]M resulted in [] = 3.7 [+ or -] 0.3 x [10.sup.-6] cm/s (not significantly different).

Experiments on intestinal crocine metabolism

We systematically investigated the stability of crocins by different treatments simulating intestinal digestion in order to elucidate whether the generation of unglycosidated trans-crocetin is attributable to microbial or rather cell-dependent metabolism. Therefore, hydroethanolic saffron extract SE with defined contents of the different crocins (total amount of apocarotenoid content distributes to 65% crocin-1, 16% crocin-2, 17% crocin-4, 1% crocin-5, trace amounts of crocin-3 and trans-crocetin, percentage indicates molar distribution as deduced from peak areas at 440 nm) was incubated following different protocols.


Treatment of SE for 90 min under simulated gastric conditions (SGF, 37[degrees]C, stirring) and subsequent UPLC[R] analysis resulted in a 20%-loss of the initial contents of each crocin-1, crocin-2 and crocin-4. Formation of trans-crocetin could not be detected. Incubating SE with simulated intestinal fluids (SIF, 37[degrees]C, stirring) led to very similar results. Hence, neither the physicochemical environments nor the included enzymes of both SGF and SIF are capable of disintegrating crocins to trans-crocetin. Yet, these data suggest a total loss of about 36% of initial crocins to unknown metabolites during gastric and intestinal passage.

Further, SE was incubated with a freshly prepared tissue homogenate from purged small mouse intestine. The homogenate was pretreated with antibiotics in order to rule out interference of residual microorganisms prior to incubation. 50% of the initial total crocins disappeared while trans-crocetin was generated to account for 4.4% of total apocarotinoids after 5 h of incubation (Fig. 5). In contrast, treatment of SE with mice feces homogenate reduced the total crocin content by about 80% (Fig. 5) while only <1% of the residual apocarotinoids consisted of trans-crocetin (Fig. 6). From these data it can be concluded that the deglycosidation of crocins to trans-crocetin is mediated by enzymatic activity of the intestinal cells, which should be either esterases or [beta]-glycosidases (Day et al. 1998; Nemeth et al. 2003). On the other side crocin-1 feeding to rats indicated that this glycoside is excreted to a high extend through the intestinal tract, which indicates, that under in vivo conditions only a small amount of crcoin is metabolized to crocetin (Xi et al. 2007).

To investigate the homogenates' influence on the distribution pattern of different crocins, the incubation mixtures were analyzed by UPLC at [lambda] = 440 nm. Interestingly it was observed (Table 1A) that 5 h incubation with intestinal homogenates weakly reduced crocin-1 amount (-18%) while crocin-2 and crocin-5 disappeared dramatically (-84% and -62%, respectively). Crocin-4 content increased during the 5 h incubation time (+121%). Intestinal homogenate seems to act preferentially on substrates with single glucose units bound to alkyl chain (crocin-2 and crocin-5), while substrates that exclusively bear gentiobioside units (crocin-1 and crocin-4) are hardly affected. The completely deglycosidated trans-crocetin was found in low amounts (4.5%). These data allow concluding that the accountable hydrolytic enzymes cleave glucoside moieties faster than the disaccharidic gentiobioside moieties.



In comparison, incubation with mouse feces homogenate did hardly change the crocin pattern (Table 1C) or led to the formation of significant amounts of trans-crocetin. This means that the dramatic decline of total crocins (-80%) must be attributed to fecal metabolism towards backbone-cleaved products that are not detectable at [lambda] = 440 nm. Additional LC-MS studies did not indicate the formation of any specific degradation product (data not shown).


The here described in vitro studies have shown that crocins from saffron are probably not bioavailable in the systemic compartment after oral application. On the other side the investigations clearly have pointed out that crocins get hydrolyzed in the intestine to the deglycosylated trans-crocetin, which subsequently is absorbed by passive transcellular diffusion to a high extent and within a short time interval over the intestinal barrier. Crocetin will penetrate in a quite slow process the blood brain barrier to reach the CNS. The intestinal deglycosylation of different crocins in the intestine is mainly due to enzymatic processes in the epithelial cells and only to a very minor extent due to deglycosylation by the fecal microbiome. On the other side the fecal bacteria degrade the apocarotenoid backbone to smaller alkyl units, which do not show any more the typical UV absorbance of crocins. As previous studies have shown strong NMDA receptor affinity and channel opening activity of trans-crocetin. The use of saffron for CNS disorders seems to be justified from the pharmacokinetic and pharmacodynamic background.

Conflict of interest

The authors declare no conflict of interest.

A             apical
B             basolateral
BBB           blood-brain barrier
BCEC          brain capillary endothelial cells
BCSFB         blood cerebrospinal fluid barrier
CNS           central nervous system
DAPI          4',6-diamidin-2-phenylindol
DMSO          dimethylsulfoxide
EDTA          ethylenediaminetetraacetic acid
EVOM          epithelial voltohm meter
FBS           fetal bovine serum
HBSS          Hank's balanced salt solution
MIT           3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium
NEAA          non-essential amino acids
NMDA          N-methyl- d -aspartate
[]   apparent permeability coefficient [cm/s]
PBS           phosphate buffered saline
SE            hydroethanolic saffron extract
SGF           simulated gastric fluids
SIF           simulated intestinal fluids
TFA           trifluoroacetic acid
TEER          transepithelial electrical resistance


The study has been financed by intramural grants from University of Munster, Germany (Internal grant Institute for Pharmaceutical Biology and Phytochemistry).

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.phymed.2014.10.009.


Angelow, S., Zeni, P., Galla, H.J., 2004. Usefulness and limitation of primary cultured porcine choroid plexus epithelial cells as an in vitro model to study drug transport at the blood-CSF barrier. Adv. Drug Deliv. Rev. 56, 1859-1873.

Artursson, P., Magnusson, C., 1990. Epithelial transport of drugs in cell culture. 11: Effect of extracellular calcium concentration on the paracellular transport of drugs of different lipophilicities across monolayers of intestinal epithelial (Caco-2) cells. J. Pharm. Sci. 79, 595-600.

Berger, F., Hensel. A, Nieber, K., 2011. Saffron extract and trans-crocetin inhibit the glutamatergic synaptic transmission in rat cortical brain slices. Neuroscience 180, 238-247.

Boskabady, M.H., Shafei, M.N., Shakiha, A., Sefidi, H.S., 2008. Effect of aqueous-ethanol extract from Crocus sativus (Saffron) on guinea-pig isolated heart. Phytother. Res. 22, 330-334.

Day, A.J., DuPont, M.S., Ridley, S., Rhodes, M., Rhodes, M.J.C., Morgan, M., Williamson, G., 1998. Deglycosylation of flavonoid and isoflavonoid glycosides by human small intestine and liver [beta]-glucosidase activity. FEBS Lett. 436,71-75.

During, A., Dawson, H.D., Harrison. E.H.. 2005. Carotenoid transport is decreased and expression of the lipid transporters SR-BI, NPC1L1, and ABCA1 is downregulated in Caco-2 cells treated with ezetimibe. J. Nutr. 135, 2305-2312.

Ferrence, S.C, Bendersky, G., 2004. Therapy with saffron and the goddess at therapy. Perspect. Biol. Med. 47, 199-226.

Franke, H., Galla, H., Beuckmann, C.T., 2000. Primary cultures of brain microvessel endothelial cells: a valid and flexible model to study drug transport through the blood-brain barrier in vitro. Brain Res. Protoc. 5, 248-256.

Hausenblas, H.A., Saha, D., Dubyak, P.J., Anton, S.D., 2013. Saffron (Crocus sativus L.) and major depressive disorder: a meta-analysis of randomized clinical trials. J. Integr. Med. 11, 377-383.

Hidalgo, I.J., Raub, T.J., Borchardt, R.T., 1989. Characterization of the human colon carcinoma cell line (Caco-2) as a model system for intestinal epithelial permeability. Gastroenterology 96, 736-749.

Hilgendorf, C., Spahn-Langguth, H., Regardh, C., Lipka, E., Amidon, G., Langguth, P., 2000. Caco-2 versus Caco-2/HT29-MTX co-cultured cell lines: permeabilities via diffusion, inside- and outside-directed carrier-mediated transport. J. Pharm. Sci. 89(1), 63-75.

Hubatsch, L, Ragnarsson, E., Artursson, P., 2007. Determination of drug permeability and prediction of drug absorption in Caco-2 monolayers. Nat. Protoc. 2, 2111-2119.

Joukar, S., Najafipour, H., Khahsari, M., Sepehri, G., Shahrokhi, N., Dabiri, S., Gholamhoseinian, A, Hasanzadeh, S., 2010. The effect of saffron consumption on biochemical and histopathological heart indices of rats with myocardial infarction. Cardiovasc. Toxicol. 10, 66-71.

Kamalipour, M., Akhondzadeh. S., 2011, Cardiovascular effects of Saffron: an evidence-based review. J. Tehran Univ. Heart Cent. 6, 59-61.

Lautenschlager, M., Lechtenberg, M., Sendker, J., Hensel, A., 2014. Effective isolation protocol for secondary metabolites from Saffron: semi-preparative scale preparation of crocin-1 and trans-crocetin. Fitoterapia 92, 290-295.

Lechtenberg, M., Schepmann, D., Niehues, M., Hellenbrand, N., Wiinsch. B., Hensel, A., 2008. Quality and functionality of saffron: quality control, species assortment and affinity of extract and isolated saffron compounds to NMDA and sigma-1 receptors. Planta Med. 74, 764-772.

Lee., I.A., Lee, J.H., Baek, N.I., Kin, D.H., 2005. Antihyperlipidemic effect of crocin isolated from Gardenia jasminoids and its metabolite crocetin. Biol. Pharm. Bull. 28, 2106-2110.

Mashmoul, M., 2013. Saffron: a natural potent antioxidant as a promising anti-obesity drug. Antiox. 2, 293-308.

Mosmann, T., 1983. Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J. Immunol. Methods 65, 55-63.

Nemeth, K., Plumb, G., Berrin, J., Juge, N., Jacob, R., Naim, H., Williamson, G., Swallow, D., Kroon, P., 2003. Deglycosylation by small intestinal epithelial cell [beta]-glucosidases is a critical step in the absorption and metabolism of dietary flavonoid glycosides in humans. Eur. J. Nutr. 42, 29-42.

Nevada Proteomics Center, University of Nevada, Reno, 2014. Protocol: Preparation of soluble and membrane protein fractions. Access: 03.06. 2014.

Pauli-Magnus, C., von Richter, O., Burk, O., Ziegler, A., Mettang, T., Eichelbaum, M., Fromm, M.F., 2000. Characterization of the major metabolites of verapamil as substrates and inhibitors of P-glycoprotein. J. Pharmacol. Exp. Ther. 293, 376-382.

Reboul, E., Borel, P., 2011. Proteins involved in uptake, intracellular transport and basolateral secretion of fat-soluble vitamins and carotenoids by mammalian enterocytes. Prog. Lipid Res. 50.388-402.

Schmidt, M., Betti, G., Hensel, A., 2007. Saffron in phytotherapy: pharmacology and clinical uses. Wien. Med. Wochenschr. 157, 315-319.

Tavelin, S., Grajo, J., Taipalensuu, J., Ocklind, G., Artursson, P., 2002. Applications of epithelial cell culture in studies of drug transport. Methods Mol. Biol. 188, 233-272.

Tomita, M., Hayashi, M., Awazu. S., 1996. Absorption-enhancing mechanism of EDTA, caprate, and decanoyicarnitine in Caco-2 Cells. J. Pharm. Sci. 85, 608-611.

Umigai, N., Murakami, K, Ulit, M.V., Antonio, LS., Shirotori, M., Morikawa, H., Nakano. T., 2011. The pharmacokinetic profile of crocetin in healthy adult human volunteers after a single oral administration. Phytomedicine 18, 575-578.

Xi, L, Qian, Z, Du, P., Fu, J., 2007. Pharmacokinetic properties of crocin (crocetin digentiobiose ester) following oral administration in rats. Phytomedicine 14, 633-636.

Yee, S., 1997. In vitro permeability across Caco-2 cells (colonic) can predict in vivo (small intestinal) absorption in man--fact or myth. Pharm. Res. 14,763-766.

Zippel, J., Deters, A., Pappai, P., Hensel, A., 2009. A high molecular arabinogalactan from Ribes nigrum L.: influence on cell physiology of human skin fibroblasts and keratinocytes and internalization into cells via endosomal transport. Carbohydr. Res. 344, 1001-1008.

Zumdick, S., Deters, A., Hensel, A., 2012. In vitro transport of oligomeric procyanidins (DP2 to 4) across monolayers of Caco-2 cells. Fitotherapia 83, 1210-1217.

M. Lautenschlager (a), J. Sendker (a), S. Huwel (b), H.J. Galla (b), S. Brandt (a), M. Dufer (c), K. Riehemann (d), A. Hensel (a), *

(a) University of Munster, Institute for Pharmaceutical Biology and Phytochemistry, Corrensstrasse 48, D-48149 Munster, Germany

(b) University of Munster, Institute for Biochemistry, Wilhelm-Klemm-Str. 2, D-48149 Munster, Germany

(c) University of Munster, Institute for Pharmaceutical and Medical Chemistry, Department of Pharmacology, Corrensstrajie 48, D-48149 Munster, Germany

(d) Center for Nanotechnology/Institute of Physics, Heisenbergstrasse 11, D-48149 Munster, Germany

* Corresponding author. Tel.: +49 251 8333380: fax: +49 251 8338341.

E-mail address: (A Hensel).
Table 1
Relative amounts (%) of crocins 1,2,4,5 and trans-crocetin in
relation to their respective amount at t = 0, during 5 h treatment of
hydroethanolic saffron extract (65% crocin-1, 16% crocin-2, 17%
crocin-4, 1% crocin-5, trace amounts of trans-crocetin with tissue
homogenate from mice small intestine (A) with addition of antibiotics
(B) and by treatment with feces homogenate (C). Data are related to
the respective peak areas as determined by UHPLC[R] at [lambda] =
440 nm. Data represent 3 independent experiments with each n = 2

Compound         1 h     2 h     3 h      4 h      5 h      Composition
                                                            after exp.


Crocin-1          85.4    69.0     56.8     48.7     44.2   53.21
Crocin-2          22.1     9.0      7.3      6.0      6.2    2.51
Crocin-4         168.6   162.6    150.2    140.5    132.9   39.55
Crocin-5          48.8    28.1     25.8     26.0     24.0    0.43
trans-crocetin   589.2   866.0   1026.8   1060.3   1026.7    4.30


Crocin-1          85.2    67.9     55.8     46.2     38.8   52.83
Crocin-2          19.8     8.5      6.9      6.2      5.4    2.52
Crocin-4         173.2   162.1    145.3    129.0    115.4   39.66
Crocin-5          33.8    23.8     20.2     19.2     18.8    0.46
trans-crocetin   685.1   917.5    994.8    997.2    988.2    4.53


Crocin-1          78.3    56.1     39.0     26.9     20.0   68.54
Crocin-2          74.4    47.1     29.6     18.6     13.1   10.14
Crocin-4          86.3    64.9     45.7     31.6     23.5   19.00
Crocin-5          95.1    74.6     53.5     39.1     31.9    1.61
trans-crocetin   139.6   141.4    113.5     87.0     73.0    0.71
COPYRIGHT 2015 Urban & Fischer Verlag
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2015 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Lautenschlager, M.; Sendker, J.; Huwel, S.; Galla, H.J.; Brandt, S.; Dufer, M.; Riehemann, K.; Hense
Publication:Phytomedicine: International Journal of Phytotherapy & Phytopharmacology
Article Type:Report
Date:Jan 15, 2015
Previous Article:Mollugin from Rubea cordifolia suppresses receptor activator of nuclear factor-[kappa]B ligand-induced osteoclastogenesis and bone resorbing activity...
Next Article:Natural indole butyrylcholinesterase inhibitors from Nauclea officinalis.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters