Membrane Permeability and Aqueous Stability Study of Linear and Cyclic Diarylheptanoids from Corylus maxima

Seven diarylheptanoids were isolated from Corylus maxima by flash chromatography and semipreparative high-performance liquid chromatography (HPLC) and identified by Orbitrap® mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectroscopy as linear diarylheptanoids: hirsutanonol-5-O-β-D-glucopyranoside (1), platyphyllonol-5-O-β-D-xylopyranoside (4), platyphyllenone (5); and cyclic derivatives: alnusonol-11-O-β-D-glucopyranoside (6), alnusone (7), giffonin F (8), carpinontriol B (9). Cyclic diarylheptanoids are reported in C. maxima for the first time. The aqueous stability of the isolated compounds and other characteristic constituents of C. maxima, oregonin (2), hirsutenone (3), quercitrin (10) and myricitrin (11) was evaluated at pH 1.2, 6.8 and 7.4. The passive diffusion of the constituents across biological membranes was investigated by parallel artificial membrane permeability assay for the gastrointestinal tract (PAMPA-GI) and the blood–brain barrier (PAMPA-BBB) methods. The cyclic diarylheptanoid aglycones and quercitrin were stable at all investigated pH values, while a pH-dependent degradation of the other compounds was observed. A validated ultrahigh-performance liquid chromatography-diode-array detection (UHPLC-DAD) method was utilized for the determination of compound concentrations. The structures of the degradation products were characterized by UHPLC-Orbitrap® MS. Platyphyllenone and alnusone possessed log Pe values greater than −5.0 and −6.0 in the PAMPA-GI and PAMPA-BBB studies, respectively, indicating their ability to cross the membranes via passive diffusion. However, only alnusone can be considered to have both good aqueous stability and satisfactory membrane penetration ability.


Introduction
Compounds of plant origin play an important role in therapeutic drug discovery by providing a large chemical diversity and covering an alternative chemical space compared with synthetic derivatives. In recent years, there has been an upsurge of interest in diarylheptanoids, a group of plant-derived phenolic compounds bearing a 1,7-diphenylheptan skeleton. The firstly discovered diarylheptanoid, namely curcumin, was isolated from the traditional herbal remedy and dietary spice, turmeric (Curcuma longa L.) two hundred years ago [1]. Ever since, hundreds of new compounds with 1,7-diphenylheptan structures have been identified in various plants with beneficial biological effects [2]. For instance, in our previous studies, we reported the presence of antioxidant diarylheptanoids in hazel species (common hazel, Corylus avellana L.; Turkish hazel, C. colurna L.; and the filbert, C. maxima Mill.) for the first time. The C. avellana leaves have been used in traditional medicine to treat

Plant Material
The leaves of Corylus maxima Mill. were collected in Pesthidegkút, Hungary [47.572938, 18.958063] in August, 2018. Fifty 50 g samples were collected from three trees after flowering stage. Plant samples were authenticated at the Department of Pharmacognosy, Semmelweis University, Budapest, where a voucher specimen is deposited.

Extraction and Sample Preparation
The dried and milled plant sample (150 g) was extracted with 1500 mL of ethyl acetate, which was followed by extraction with 1500 mL of methanol in an ultrasonic bath at room temperature for 8-8 h. The extracts were evaporated to dryness under reduced pressure in a rotary evaporator at 50 • C. The dried extracts (34.8 g) were dissolved in HPLC-grade methanol and filtered through Phenex-RC 15 mm, 0.2 µm syringe filters (Gen-Lab Ltd., Budapest, Hungary).

UHPLC-DAD-Orbitrap ® MS Analyses
The Corylus samples were analyzed by a Waters Acquity UHPLC system equipped with a diode array detector (Waters Corporation, Milford, MA, USA) using an Acquity BEH C18 column (2.1 × 100 mm; 1.7 µm, 25 • C) (Waters Corporation) with 0.1% formic acid in water as eluent A and acetonitrile as eluent B with a flow rate of 0.3 mL/min. For the chromatographic investigation of the extracts and flash chromatography fractions, different gradient profiles were applied, while aqueous stability and PAMPA solutions were analyzed with the following gradient: from 95:5 to 0:100 v/v in 10 min. The injection volume was 10 µL. UV spectra and chromatograms were recorded at 200-400 nm and the chromatograms acquired at the UV absorption maxima of each compound (281 nm for compounds 1, 2 and 3, 278 nm for compounds 4 and 5, 297 nm for compounds 6 and 7, 257 nm for compound 8, 248 nm for compound 9, 255 nm for compound 10 and 261 nm for compound 11) were used for the data evaluation.
The UHPLC-DAD method, utilized for the analysis of the samples acquired in the PAMPA experiments and for the determination of aqueous stability was validated for linearity, precision and accuracy. Linearity was determined by analyzing the isolated compounds or standards (1, 2, 3, 4, 5, 6, 7 and 11) at five concentrations (0.1, 0.5, 1, 50, 100 µM in DMSO), each in triplicate. The slope, intercept and correlation coefficient were determined by a least-squares weighted regression analysis. The method provided linear responses (r 2 > 0.999) for all standards within the investigated range (Table S1). The LOD parameter was determined at a 3/1 and the LOQ at a 10/1 signal-to-noise ratio (Table S1). The retention time repeatability was checked with six successive runs of the samples and was found to be suitable for all compounds: the relative standard deviation was <0.15% (n = 6). Quality control samples were prepared at three different concentrations (low, mid and high) for each standard solution, each in triplicate. These were used to determine both the intraday and interday precision and accuracy for each standard solution, which were applied for the system suitability tests: intraday and interday precision was <10% for all the analytes, while accuracy ranged from 93.69% to 116.82% (n = 9) (Table S2). Blank samples (pure solvents) were analyzed to exclude any coelution of impurities with the analytes.
For obtaining high resolution mass spectrometric data, a Dionex Ultimate 3000 UHPLC system (3000RS diode array detector, TCC-3000RS column thermostat, HPG-3400RS pump, SRD-3400 solvent rack degasser, WPS-3000TRS autosampler) was used hyphenated with an Orbitrap Q Exactive Focus Mass Spectrometer equipped with an electrospray ionization source (Thermo Fischer Scientific, Waltham, MA, USA). The same chromatographic method was applied as described above. The electrospray ionization source was operated in negative ionization mode, and operation parameters were optimized automatically using the built-in software. The working parameters were as follows: spray voltage, 2500 V; capillary temperature, 320 • C; sheath gas (N 2 ), 47.5 • C; auxiliary gas (N 2 ), 11.25 arbitrary units; and spare gas (N 2 ), 2.25 arbitrary units. The resolution of the full scan was of 70,000, and the scanning range was between 120 and 1000 m/z units. Parent ions were fragmented with a normalized collision energy of 10%, 30% and 45%.

Isolation Procedures
In the case of the Corylus ethyl acetate extract, as a starting purification step, normalphase flash chromatography (CombiFlash Nextgen 100, Teledyne Isco, Lincoln, NE, USA) was applied using two RediSep Rf Gold ® 40 g HP silica columns (Teledyne Isco, USA) attached to each other and hexane as eluent A, ethyl acetate as eluent B (from 100:0 to 0:100 v/v in 20 min) with a flow rate of 40 mL/min. Four fractions were collected (a-d).
Fractions c (540.3 mg) and d (340.1 mg) were further purified by reversed-phase flash chromatography on two RediSep Rf Gold ® 15.5 g reversed-phase C18 columns (Teledyne Isco, USA) using 0.1% formic acid in water as eluent A and methanol as eluent B with a flow rate of 10 mL/min (from 100:0 to 0:100 v/v in 20 min.). Fraction c yielded thirteen further fractions (c1-13). Fractions c3 (32.4 mg) and c8 (48.5 mg) were purified by a Waters 2690 HPLC system with a Waters 996 diode-array detector (Waters Corporation, Milford, MA, USA) equipped with a Luna C18 100 A reversed-phase column (150 × 10.0 mm, 5 µm; Phenomenex Inc; Torrance, CA, USA, 25 • C) using 0.1% formic acid in water as eluent A and acetonitrile as eluent B. The following gradient elution programs were applied: for fraction c3 from 75:25 to 70:30 v/v in 35 min and for fraction c8 from 65:35 to 60:40 v/v in 35 min with a flow rate of 4.6 mL/min. Compound 9 (5.2 mg) was isolated from fraction c3 and compound 8 (8.9 mg) from fraction c8.
Fraction d yielded six fractions (d1-6). Fraction d4 (44.3 mg) was purified by semipreparative HPLC using the same instrumentation and stationary phase as detailed above and 0.1% formic acid in water as eluent A and acetonitrile as eluent B (from 80:20 to 74:26 v/v in 30 min) with a flow rate of 4.6 mL/min. Compounds 4 (4.4 mg) and 6 (1.5 mg) were isolated from this fraction.
The methanol extract was separated by flash chromatography using the same instrumentation as above, a RediSep Rf Gold ® 150 g reversed-phase C18 column (Teledyne Isco, USA), 0.1% formic acid in water as eluent A and methanol as eluent B (from 95:5 to 0:100 v/v in 30 min). Altogether seven fractions were collected (e-k). Fractions g (850.4 mg) and j (750.7 mg) were further purified by reversed-phase flash chromatography on a RediSep Rf Gold ® 100 g reversed-phase C18 column (Teledyne Isco, USA) using the same eluents as above. The following gradient elution programs were applied: for fraction g from 70:30 to 0:100 v/v in 25 min and for fraction j from 90:10 to 40:60 v/v in 35 min.
Fraction g yielded eight further fractions (g1-8). Fraction g3 (51.0 mg) was purified by semipreparative HPLC using the same instrumentation and stationary phase as detailed above, 0.1% formic acid in water as eluent A and acetonitrile as eluent B (from 88:12 to 80:20 v/v in 30 min) with a flow rate of 4.6 mL/min to yield compound 1 (7.0 mg).
Fraction j yielded thirteen further fractions (j1-13). Fraction j12 (42.1 mg) was purified by semipreparative HPLC using 0.3% acetic acid in water as eluent A and methanol as eluent B (from 50:50 to 30:70 v/v in 45 min) with a flow rate of 4.6 mL/min to yield compound 5 (2.5 mg). Compound 7 (1.9 mg) was isolated from fraction j12 (35.2 mg) by semipreparative HPLC using 0.1% formic acid in water as eluent A and methanol as eluent B (from 70:30 to 50:50 v/v in 30 min) with a flow rate of 4.6 mL/min.

NMR Spectroscopy
NMR spectra were recorded in deuterated methanol (Methanol-d 4 ) or in deuterated dimethyl sulfoxide (DMSO-d 6 ) on a BRUKER AVANCE III HD 600 (600/150 MHz) instrument equipped with Prodigy cryoprobehead at 295 K or on a Varian DDR 600 (600/150 MHz) equipped with a 5 mm inverse-detection gradient (IDPFG) probehead, at 298 K. The pulse programs were taken from the Bruker or the Varian software library (TopSpin 3.5 or VnmrJ 3.2). Chemical shifts (δ) and coupling constants (J) are given in ppm and in Hz, respectively. 13 C and 1 H chemical shifts are given in ppm relative to the NMR solvent or relative to tetramethylsilane (TMS) when internal standard was used. The complete 1 H and 13 C assignments were achieved with widely accepted strategies based on 1 H NMR, 13 C NMR, DeptQ, 1 H-1 H COSY, 1 H-13 C edHSQC, 1 H-13 C HMBC and 1 H-1 H ROESY measurements.

Determination of Aqueous Stability
The aqueous solutions and buffers were prepared as follows: pH = 1.2, 1.0 g NaCl and 3.5 mL HCl dissolved in distilled water to achieve the final volume of 500.0 mL; pH = 6.8, 20.2 g Na 2 HPO 4 ·7H 2 O and 3.4 g NaH 2 PO 4 ·H 2 O dissolved in distilled water to achieve the final volume of 1000.0 mL, pH adjustment with 0.5 M NaOH or 0.5 M HCl; pH = 7.4, one PBS tablet (phosphate buffered saline, pH 7.4; Sigma Aldrich, St. Louis, MO, USA) dissolved in 200.0 mL of distilled water.
With the aforementioned buffers, 10 mM dimethyl sulfoxide stock solutions of the isolated compounds were diluted 100-fold. The initial peak area (AUC) recorded by UHPLC-DAD at the zero time point was compared to the area under the peak detected in the sample incubated for 4 h at 37 • C. Statistical differences between the AUC values were evaluated by a paired-sample t-test, changes were considered significant when the p-values were lower than 0.05. Compounds with no notable change in their AUC and with no degradation products detected were considered to be stable. For degradable compounds, the above-mentioned validated UHPLC-DAD method was used to examine the change in compound concentration. All experiments were performed in triplicates (n = 3).

Parallel Artificial Membrane Permeability Assay (PAMPA) Studies
The test solutions were prepared with dimethyl sulfoxide (DMSO) at the concentration of 10.0 mM. These were diluted with phosphate buffers (pH = 7.4 for the PAMPA-BBB and pH = 6.8 for the PAMPA-GI) to obtain the donor solutions (297.0 µL buffer + 3.0 µL DMSO solution) and filtered through Phenex-RC 15 mm, 0.2 µm syringe filters (Gen-Lab Ltd., Budapest, Hungary). A parallel artificial membrane permeability assay (PAMPA) system was used to determine the effective permeability (Pe) for the compounds of interest. . The donor and acceptor plates were fit, and then the sandwich system was incubated at 37 • C for 4 h in a Stat-Fax 2200. After the incubation the PAMPA plates were separated and the compound concentrations in the donor (C D (t)) and acceptor (C A (t)) solutions, as well as in the donor solution at the zero time point (C D (0)) were determined by UHPLC-DAD. The effective permeability and the membrane retention in the PAMPA-BBB (1) and the PAMPA GI (2) experiments were calculated by the following equations [12]: where Pe is the effective permeability coefficient (cm/s), A is the filter area (0.24 cm 2 ), V D and V A are the volumes in the donor (0.15 cm 3 ) and acceptor phases (0.30 cm 3 ), t is the incubation time (s), τ SS is the time (s) to reach steady-state (240 s), C D (t) is the concentration (mol/cm 3 ) of the compound in the donor phase at time t, C D (0) is the concentration (mol/cm 3 ) of the compound in the donor phase at time 0, MR is the estimated membrane retention factor (the estimated mole fraction of solute lost to the membrane) and r a is the sink asymmetry ratio (gradient-pH-induced), defined as All experiments were performed in triplicates on three consecutive days (n = 9), caffeine standard was used as positive, while rutin as negative control. ClogP values were calculated by ChemAxon Marvin 22.3.

Structure Elucidation by UHPLC-Orbitrap ® MS and NMR
The isolated compounds were identified by Orbitrap ® mass spectrometry hyphenated to UHPLC separation (Table 1)  Hz, H-4) ppm. Furthermore, four methylene groups were identified. The 13 C resonance at δ 199.5 ppm (C-3) revealed the presence of a carbonyl group. These spectroscopic data suggested a linear diarylheptanoid structure. Based on the additional 2D correlations, the unsaturation was located at the C-4-C-5 position, while the carbonyl group was located at the C-3 position. Therefore, the compound was identified as platyphyllenone. The 1 H and 13 C NMR resonances were identical to those published in a previous report [14].
The 2 ppm suggested the structure was a diarylheptanoid glycoside. The 13 C resonance at δ 210.5 ppm was evident for a carbonyl group. Furthermore, nine aliphatic resonances along with their HSQC crosspeaks recommended the presence of five methylene units in the heptane chain. The 1 H resonance at δ 4.08 ppm and its HSQC correlation to 13 C resonance at δ 74.8 ppm indicated the presence of an oxymethine function. The TOCSY spectrum confirmed that this function was located in the aliphatic chain. The 2D NMR correlations confirmed the substitution pattern of the chain: the carbonyl group was in C-9, while the oxymethine function was at the C-11 position. Based on these results, the structure of the aglycone was identified as alnusonol. The HMBC correlation between the anomeric 1 H resonance (δ 4.24 ppm) and the 13 C resonance at δ 74.8 ppm (C-11) revealed the sugar moiety was located at the C-11 position. On the basis of this information, the structure was established as alnusonol-11-O-β-D-glucopyranoside.  13 C NMR spectrum showed a single carbonyl resonance at δ 200.6 ppm (C-9). The twodimensional NMR spectra revealed that the unsaturation was at the C-10-C-11 position, while the carbonyl group was at the C-9 position. Based on these data, the compound was identified as alnusone. Early NMR report on alnusone [16] highlights only a few characteristic 1 H resonances, thus, herein, we provide a more complete NMR assignment.
The 1H, H-9) ppm suggested the occurrence of an oxymethine group in the heptane chain. The 13 C NMR spectrum displayed a carbonyl resonance at δ 215.6 (C-11) ppm. Furthermore, three methylene groups were identified. Based on the COSY correlations, the olefin function and the oxymethine group were adjacent to each other: they were located at positions C-7-8 and C-9. The correlations of the HMBC spectra revealed that the carbonyl group was at the C-11 position. Thus, the structure was determined as giffonin F. The 1 H and 13 C NMR resonances were analogous to the literature data [17].  , H-11) ppm. The 13 C resonance at δ 215.6 ppm (C-9) suggested a carbonyl group in the heptane chain. Based on the correlations of the 2D spectra, the hydroxyl groups were located at positions C-10, C-11 and C-12, while the carbonyl group was at the C-9 position. Therefore, the structure was established as carpinontriol B. The 1 H and 13 C NMR assignments were identical to those published earlier [18].

Determination of Aqueous Stability
The aqueous stability of the isolated diarylheptanoids and that of the standards oregonin (2), hirsutenone (3), quercitrin (10) and myricitrin (11) were also determined. The latter two flavonol glycosides were previously reported as the main compounds in Corylus maxima leaf extracts [5]. The presence of the linear diarylheptanoids 2 and 3 was also proven by spiking the extract with the standards in two different chromatographic systems. The aqueous stability was investigated at three physiological pH values: at pH 1.2 in simulated gastric fluid, at pH 6.8 in simulated intestinal fluid and at pH 7.4.
The cyclic diarylheptanoid aglycones, namely, alnusone (7), giffonin F (8), carpinontriol B (9) and the flavonoid quercitrin (10), were stable at all three investigated pH values ( Table 2) with no notable change in their chromatographic peak areas and with no degradation products detected by UHPLC-DAD. Table 2. Results of the aqueous stability studies: compound concentration after 4 h of incubation at 37 • C compared to the initial value (%) (n = 3) and the PAMPA experiments' log P e values (n = 9).

Aqueous Stability
log P e PAMPA-BBB (n = 9) log P e PAMPA-GI (n = 9) clog P pH = 1. Abbreviations: n.d.: not detected in the acceptor phase; PAMPA-GI: parallel artificial membrane permeability assay for the gastrointestinal tract; PAMPA-BBB: parallel artificial membrane permeability assay for the blood-brain barrier. * p < 0.05 compared with the initial solutions.
A pH-dependent degradation was observed for compounds 1, 2, 3, 4, 5, 6 and 11, thus, in their case, the UHPLC-DAD method was validated for linearity, precision and accuracy to be utilized for the determination of compound concentrations at the initial time point and after 4 h of incubation at 37 • C (Tables S1 and S2). Statistical differences between the concentrations at the initial time point and after treatment (incubation at different pH values) were evaluated by a paired-sample t-test. Decreases in compound concentrations were considered significant when p < 0.05 (results are shown in Table 2).
The structures of the degradation products detected in the stability studies were characterized by UHPLC-Orbitrap ® MS.

Characterization of the Degradation Products by UHPLC-Orbitrap ® MS
The reaction mixtures of the compounds were analyzed by UHPLC-Orbitrap ® MS in negative ionization mode ( Table 1). The proposed degradation products and the tentative characterization of the diarylheptanoid compounds are depicted in Figures 1 and 2. In the case of compound 1 (hirsutanonol-5-O-β-D-glucopyranoside), four main degradation products were observed. One of the decomposition products was detected at m/z 327.1239 (1a). The difference of 180 Da between 1a and the molecular ion of the original compound refers to the loss of a hexose moiety.   The characteristic fragment ions at m/z 205.0864 and 179.0699 indicated 1a to be hirsutenone [3]. After comparing the retention time and the mass spectral data with those of a reference standard, 1a was confirmed as hirsutenone. Two other degradation products at m/z 505.1718 (1b, 1c) were observed in the mixture. The molecular formula was C 25 H 29 O 11 in both cases, thus these peaks were assumed to be isomers. During their collision-induced dissociation, both 1b and 1c lost a sugar moiety and showed product ions at m/z 325.1080 and m/z 203.0706 in their MS/MS spectra with a difference of 2 Da compared to the fragment ions of compound 1 and the molecular and fragment ions of 1a (m/z 327.1239 and m/z 205.0864). Accordingly, we presumed that a double bond was formed in the heptane chain during the degradation of 1.
The degradation pattern of oregonin (2) was similar to that of compound 1. The aglycone hirsutenone (2a) (corresponding to 1a and 3) as well as the derivative containing a double bond in the alkyl chain (2b) were detected at m/z 327.1239 and m/z 475.1609, respectively. The aglycone of 2b was also observed at m/z 325.1083 (2c). The decomposed derivative of hirsutenone (3) was detected at m/z 325.1082 (3a). The instability of hirsutenone is well-known in the literature, however, its degradation products have not previously been identified [11].
The glycosidic bond of compound 4 was hydrolyzed producing the platyphyllenone aglycone (4a), which corresponded to the reference compound 5. Compound 5 (m/z 295.1338) was modified to form 5a with its molecular ion at m/z 313.1446. The difference of 18 Da refers to the addition of a water molecule. The product ions of 5a at m/z 163.0753 and 149.0595, produced by different neutral losses, indicated the presence of a keto and a hydroxyl group on the alkyl chain [5]. Based on the fragmentation pattern and the elemental composition (C 19 H 21 O 4 ), 5a was characterized as platyphyllone.
In the case of compound 6, only one breakdown product was detected at m/z 293.1184 (6a), where the difference was again 180 Da as compared to the molecular ion of the original compound. Thus, 6a was identified as alnusone, the aglycone of compound 6 (alnusonol-11-O-β-D-glucopyranoside). This proposed structure was further confirmed when its retention time was compared to that of the reference substance alnusone (7). The fragment ions of 6a (corresponding to 7) at m/z 265.1221, 251.1071, 224.0834 and 210.0673 were generated by a rearrangement of the deprotonated molecular ion and the subsequent opening of the diarylheptanoid cycle [19].
The flavonoid myricitrin (11) formed two isomers at m/z 925.1686 and 925.1682 (11a and 11b), and two further products at m/z 895.1583 (11c) and m/z 941.1633 (11d). Based on literature data [20], 11a and 11b were identified as dimers of myricitrin. Fragmentation of the dimers yielded ions at m/z 779.1099 [M-H-146] − and m/z 633.0499 [M-H-146-146] − , corresponding to the product ions generated by the cleavage of one and two rhamnose moieties, respectively. In addition, the dimers can undergo further oxidation, resulting in the formation of 11c and 11d [20].

Parallel Artificial Membrane Permeability Assay (PAMPA) Studies
Among the investigated compounds only platyphyllenone (5) and alnusone (7) possessed log P e values greater than −5.0 (−4.92 ± 0.07 and −4.90 ± 0.17, respectively) in the PAMPA-GI experiments, and greater than −6.0 in the PAMPA-BBB studies (−5.24 ± 0.25 and −4.66 ± 0.14, respectively) ( Table 2). Accordingly, these constituents can be considered to have good membrane penetration ability [21]; thus, it might be assumed that they are absorbed in the gastrointestinal tract and cross the blood-brain barrier via passive diffusion. However, since platyphyllenone (5) was proven to be unstable in the aqueous media of the PAMPA models, results for this compound need to be treated with caution. However, the extent of decomposition at pH 6.8 and 7.4 (94.98 ± 2.10% and 90.40 ± 1.52% final concentrations compared to the initial value, respectively) cannot be considered significant enough to notably modify the PAMPA results .  Compounds 1, 2, 3, 4, 6, 8, 10 and 11 were not detected in the acceptor phase of the PAMPA models suggesting that they were unable to cross the lipid membranes of the GIT and BBB ( Table 2). The cyclic diarylheptanoid aglycone, carpinontriol B (9) (344.1 g/mol, clog P 1.6) was detected in the acceptor phase of the PAMPA-GI model (not it the PAMPA-BBB), but the calculated log P e value (−5.49 ± 0.30) suggested a poor membrane permeability for this compound as well.
It must be pointed out that the results of the PAMPA experiments cannot be considered fully accurate in the case of the constituents proven to be chemically unstable in aqueous media at the investigated pH values (compounds 1, 2, 3, 4, 5, 6 and 11). Nevertheless, the diarylheptanoid and flavonol glycosides (compounds 1, 2, 4, 6, 10 and 11) are rather unlikely to be able to cross the model membranes by passive diffusion due to their high molar mass and polarity (Mw > 440, clog P < 2.5). In addition, the maximum decomposition rate measured was 48.44%, in the case of myricitrin (11), thus the extent of the degradation during the PAMPA experiments would not justify the complete rejection of the results. Therefore, in the case of the glycoside derivatives, the negative results can be considered valid regardless of the chemical instability. However, further studies may be needed to investigate the membrane permeability of the labile linear diarylheptanoid aglycones, platyphyllenone (5) and hirsutenone (3), after stabilization.
Our results indicated that among the investigated compounds, solely the cyclic diarylheptanoid aglycone alnusone (7) had both a good aqueous stability and satisfactory membrane penetration ability. The other cyclic diarylheptanoid aglycones were found to be stable in aqueous environment, but the ones with a larger molar mass and higher polarity, giffonin F (8) and carpinontriol B (9), were not able to cross the lipid membranes in the PAMPA-GI and PAMPA-BBB models.
All the linear diarylheptanoids (compounds 1, 2, 3, 4 and 5) and the cyclic diarylheptanoid alnusonol-11-O-β-D-glucopyranoside (6) demonstrated significant pH-dependent decomposition in aqueous media. The linear compounds 3 and 4 were labile only at the highest investigated pH value, while 1, 2, 5 and the cyclic derivative 6 were significantly decomposed both at pH 6.8 and 7.4. Among the diarylheptanoid derivatives, 1, 2 and 6 were the most unstable at pH 7.4. Based on our PAMPA results, among them, only platyphyllenone (5) can be considered to be able to cross the lipid membranes of the GIT and the BBB by passive diffusion.
The flavonoid glycosides quercitrin (10) and myricitrin (11) showed a poor membrane penetration ability in both the PAMPA models. In addition, myricitrin was found to be unstable in aqueous medium, its concentration significantly decreasing at pH 7.4.
The recent study might corroborate the therapeutic potential of naturally occurring cyclic diarylheptanoids in terms of physicochemical and pharmacokinetic properties in addition to their beneficial biological effects that have already been revealed: cyclic diarylheptanoid compounds have gained interest because of their remarkable in vitro anti-cancer, antioxidant, melanogenesis inhibitory [22] and neuroprotective activities [23], as well as in vivo antitumor [24], anti-inflammatory and antioxidant effects [25]. In order to be able to draw definite conclusions, a higher number of molecules should be evaluated. However, our results justify further investigations on naturally occurring cyclic diarylheptanoids as potential therapeutic agents bringing the ones with a lower molecular weight and higher lipophilicity into focus.