Anti-Inflammatory Phenolic Acid Esters from the Roots and Rhizomes of Notopterygium incisium and Their Permeability in the Human Caco-2 Monolayer Cell Model

A new ferulic acid ester named 4-methyl-3-trans-hexenylferulate (1), together with eight known phenolic acid esters (2–9), was isolated from the methanolic extract of the roots and rhizomes of Notopterygium incisium. Their structures were elucidated by extensive spectroscopic techniques, including 2D NMR spectroscopy and mass spectrometry. 4-Methoxyphenethyl ferulate (8) NMR data is reported here for the first time. The uptake and transepithelial transport of the isolated compounds 1–9 were investigated in the human intestinal Caco-2 cell monolayer model. Compounds 2 and 6 were assigned for the well-absorbed compounds, compound 8 was assigned for the moderately absorbed compound, and compounds 1, 3, 4, 5, 7, and 9 were assigned for the poorly absorbed compounds. Moreover, all of the isolated compounds were assayed for the inhibitory effects against nitric oxide (NO) production in the lipopolysaccharide-activated RAW264.7 macrophages model and L-N6-(1-iminoethyl)-lysine (L-NIL) was used as a positive control. Compounds 1, 5, 8, and 9 exhibited potent inhibitory activity on NO production with the half maximal inhibitory concentration (IC50) values of 1.01, 4.63, 2.47, and 2.73 μM, respectively, which were more effective than L-NIL with IC50 values of 9.37 μM. These findings not only enriched the types of anti-inflammatory compounds in N. incisum but also provided some useful information for predicting their oral bioavailability and their suitability as drug leads or promising anti-inflammatory agents.


Introduction
Notopterygii Rhizoma et Radix (NRR), belonging to Apiaceae, are the dried roots and rhizomes of Notopterygium incisum Ting ex H. T. Chang and N. franchetii H. de Boiss. and is mainly grown in the Sichuan province of China. NRR is an important ingredient of traditional Chinese medicine recorded in the Pharmacopoeia of People's Republic of China [1] and has been shown to effectively treat common cold, headache, and rheumatism because of its diaphoretic, analgesic, and anti-inflammatory properties [2,3] for many thousands of years in Asia. Previous studies on the chemical constituents of NRR resulted in the elucidation of multiple components including coumarins, steroids [4,5], polyacetylenes [6], and essential oil [7]. The extent to which these components are effective in the human body depends on their bioavailability and metabolism in vivo. As is well known, intestinal permeability is a crucial factor that influences the bioavailability of drugs, especially of those being administrated orally. Therefore, explaining the intestinal permeability of these components is a critical step towards understanding their potential bioactivity. Human studies have proved that inflammation is critically involved in the disorders of bodies and even contributes to the pathogenesis of several serious diseases such as cancer [8] and neurodegenerative and cardiovascular diseases [9,10]. Hence, the suppression of inflammation in biological systems turned out to be an effective therapeutic strategy and an interesting target in the field of new drug research and development. Nitric oxide (NO) is a short-living free radical that is produced from L-arginine by constitutive nitric oxide synthase (cNOS) and inducible nitric oxide synthase (iNOS) within mammalian immune, cardiovascular, and neural systems, where it functions as a signaling or cytotoxic molecule. A low concentration of NO participates in neurotransmission and vasodilation, whereas overproduction of NO by NOS was responsible for inflammation [11]; therefore, NO inhibitor is believed to have therapeutic potential for the treatment of inflammation accompanying overproduction of NO. Our previous studies showed anti-inflammatory activities of some coumarins [12,13], sesquiterpenoids [14], polyacetylenes [15], and neolignans [16,17]. As part of an ongoing effort to search for natural anti-inflammatory agents, herein we describe the isolation and structural elucidation of one new and eight known phenolic acid esters from the dried roots and rhizomes of N. incisium, together with predicted absorbability, using a human intestinal Caco-2 cell monolayer model [18] and their inhibitory effects against NO production induced by lipopolysaccharide (LPS) in RAW 264.7 macrophage cells.

Extraction and Isolation
The dried roots and rhizomes of N. incisium (2 kg) were powdered and extracted with methanol (MeOH) (6 L × 4 times for 2 h first and 1 h each subsequent time) under reflux. The extracts were combined and then concentrated under reduced pressure to afford a residue (513.0 g, yield 25.7%). The residue was suspended in water and partitioned successively with cyclohexane (CHA), ethyl acetate (EtOAc), and normal butanol (n-BuOH) to afford corresponding extracts.
Although 4-methoxyphenethylferulate (8) derived from combinatorial enzymatic synthesis catalyzed by Candida antarctica lipase B [25,26] is a known compound, NMR data has not been reported thus far. Thorough exhaustive interpretation of the 1 H and 13 C-NMR, DEPT, 1 H-1 H COSY, HSQC, and HMBC spectra (see Supplementary data, Figures S12-S17), the complete unambiguous assignments for all the 1 H and 13 C-NMR signals (Table 1)

Transport of Phenolic Acid Esters 1-9 in the Human Intestinal Caco-2 Cell Monolayer Model
Intestinal permeability of phenolic acid esters 1-9 was evaluated by using human intestinal Caco-2 cell monolayer model [18]. The HPLC analytical methods for the phenolic acid esters had been validated (see Table S1). The bilateral (apical side (AP) → basolateral side (BL) and BL → AP) apparent permeability coefficients (Papp) values of compounds 1-9 are summarized in Table 2. The Papp AP → BL values of compounds 2 and 6 in the present study were well over 10 −5 cm/s, which were comparable to that of propranolol (2.29 × 10 −5 cm/s), a well-transported marker of the transcellular pathway [18], indicating their good absorption, whereas the Papp AP → BL magnitudes of compounds 1, 4, and 9 were below 10 −6 cm/s, which were comparable to that of atenolol (5.54 × 10 −7 cm/s), a poor-transported marker of the paracellular pathway [18], so compounds 1, 4, and 9 were assigned for the poorly absorbed compounds. Although 4-methoxyphenethylferulate (8) derived from combinatorial enzymatic synthesis catalyzed by Candida antarctica lipase B [25,26] is a known compound, NMR data has not been reported thus far. Thorough exhaustive interpretation of the 1 H and 13 C-NMR, DEPT, 1 H-1 H COSY, HSQC, and HMBC spectra (see Supplementary data, Figures S12-S17), the complete unambiguous assignments for all the 1 H and 13 C-NMR signals (Table 1) of 8 were performed for the first time.

Transport of Phenolic Acid Esters 1-9 in the Human Intestinal Caco-2 Cell Monolayer Model
Intestinal permeability of phenolic acid esters 1-9 was evaluated by using human intestinal Caco-2 cell monolayer model [18]. The HPLC analytical methods for the phenolic acid esters had been validated (see Table S1). The bilateral (apical side (AP) → basolateral side (BL) and BL → AP) apparent permeability coefficients (P app ) values of compounds 1-9 are summarized in Table 2. The P app AP→BL values of compounds 2 and 6 in the present study were well over 10 −5 cm/s, which were comparable to that of propranolol (2.29 × 10 −5 cm/s), a well-transported marker of the transcellular pathway [18], indicating their good absorption, whereas the P app AP→BL magnitudes of compounds 1, 4, and 9 were below 10 −6 cm/s, which were comparable to that of atenolol (5.54 × 10 −7 cm/s), a poor-transported marker of the paracellular pathway [18], so compounds 1, 4, and 9 were assigned for the poorly absorbed compounds. The P app AP→BL magnitude of compound 8 was a quantitative degree of 10 −6 cm/s, which fell in between propranolol and atenolol; compound 8 was thereby assigned for the moderately absorbed compound. Compounds 3, 5, and 7 were found to hardly permeate Caco-2 monolayers with P app AP→BL magnitudes <10 −7 cm/s. The efflux ratios of the above phenolic acid esters except compounds 3, 4, and 7 were within the range of 0.8-1.5. Physicochemical characters such as log D (logarithm of octanol-water partition coefficient) and MW (molecular weight) are generally utilized for the prediction of the permeability of compounds [28]. The log D values at pH 7.35 of nine phenolic acid esters, calculated with Pallas 3.3.2.6 ADME/Tox Software (CompuDrug, Bal Harbor, FL, USA), as well as their MW values are shown in Table 2. Herein, an inverse sigmoid trendline of log (P app AP→BL × MW 1/2 ) versus logD was plotted (Figure 3) with Origin Pro 7.5 SR1 (Origin Lab Corporation, Northampton, MA, USA) to elucidate the structure-permeability relationship of these phenolic acid esters. The permeability of phenolic acid esters presented a downward trend as log D values (>3) increased, indicating that a lipophilicity that is too high may result in low membrane permeability.  The Papp AP→BL magnitude of compound 8 was a quantitative degree of 10 −6 cm/s, which fell in between propranolol and atenolol; compound 8 was thereby assigned for the moderately absorbed compound. Compounds 3, 5, and 7 were found to hardly permeate Caco-2 monolayers with Papp AP→BL magnitudes <10 −7 cm/s. The efflux ratios of the above phenolic acid esters except compounds 3, 4, and 7 were within the range of 0.8-1.5. Physicochemical characters such as log D (logarithm of octanol-water partition coefficient) and MW (molecular weight) are generally utilized for the prediction of the permeability of compounds [28]. The log D values at pH 7.35 of nine phenolic acid esters, calculated with Pallas 3.3.2.6 ADME/Tox Software (CompuDrug, Bal Harbor, FL, USA), as well as their MW values are shown in Table 2. Herein, an inverse sigmoid trendline of log (Papp AP→BL × MW 1/2 ) versus logD was plotted (Figure 3) with Origin Pro 7.5 SR1 (Origin Lab Corporation, Northampton, MA, USA) to elucidate the structure-permeability relationship of these phenolic acid esters. The permeability of phenolic acid esters presented a downward trend as log D values (>3) increased, indicating that a lipophilicity that is too high may result in low membrane permeability.

Inhibitory Activity of Compounds 1-9 on NO Production
As part of our project to find natural structures with inhibitory activity on overproduction of NO, all of the isolated compounds were evaluated against NO release in LPS-activated RAW264.7 macrophage cell model [12,14]. The compounds 1-9 were initially assayed for cytotoxic effects in RAW264.7 cells by the 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay. The cell viability less than 95% of control was considered toxic. The results demonstrated that compounds 1-6, 8, and 9 showed no toxicity at any tested concentrations (0.78125-25 µM for  compounds 1 and 5, 3.125-100 µM for compounds 2, 3, and 4, and 1.5625-50 µM for compounds  6, 8, and 9), except compound 7, which had cytotoxicity at concentrations from 0.25 to 8 µM. As is shown in Table 3, compounds 1, 4, 5, 6, 8, and 9 showed potent inhibition with the half maximal inhibitory concentration (IC 50 ) values of 1.01, 11.11, 4.63, 12.62, 2.47, and 2.73 µM, respectively, comparable to the positive control L-N 6 -(1-iminoethyl)-lysine (L-NIL) with an IC 50 value of 9.37 µM. Compounds 2 and 3 showed moderate activity with IC 50 values of 53.69 and 70.50 µM, respectively, which were inferior to L-NIL. The effect on NO production of compound 7 was not further evaluated for its damage on cells at all test concentrations. Notably, NO inhibition activities of compounds 1, 5, 8, and 9 were more significant (p < 0.001, p < 0.05, p < 0.001, p < 0.001) than that of L-NIL. We also evaluated two related compounds, ferulic acid (10, Figure 1) and cinnamic acid (11, Figure 1), to discuss the initial structure-activity relationship. Ferulic acid, with an IC 50 value of 67.94 µM, exhibited higher inhibitory activity than that of cinnamic acid (IC 50 > 200 µM). Comparing compound 5 with compound 3 in structures and IC 50 values, it was found that the replacement of cinnamoyl group with feruloyl group significantly increased by an order of magnitude in the activity, suggesting a more significant effect of the feruloyl group than the cinnamoyl group on the inhibitory activity. Compared to compounds 2 and 4, compound 9 more effectively inhibited NO production, suggesting that the feruloyl group plays a more important role than do the salicyloyl or anisoyl groups in exerting the activity. The above-mentioned information confirmed that the feruloyl group was critical for maintaining or enhancing NO inhibition activities. In addition, all ferulic acid esters that were tested (compounds 1, 5, 6, 8, and 9) exhibited stronger NO inhibition than ferulic acid (10), and the activity of cinnamic acid ester (3) was also superior to that of cinnamic acid (11), suggesting that the ester moiety was a requirement in phenolic acid derivatives for the activity. For the five ester of ferulic acid, the increasing order of the inhibition activity against NO production is as follows: 6 < 5 (or 8, 9)< 1. Comparing the activity of compounds 1 and 6, which possess hydrocarbyl substitution but differ in the length of carbon chain, it was obvious that the long hydrocarbon chain was responsible for the enhanced activity. The data of compounds 8 and 9 suggested that methoxyl substitution of phenethyl moiety had little influence on the activity. Table 3. Inhibition of compounds 1-9 on NO production (n = 3). 11.11 ± 1.43 9 2.73 ± 0.58 *** L-NIL: L-N 6 -(1-iminoethyl)-lysine; * p < 0.05 vs. L-NIL; *** p < 0.001 vs. L-NIL.

No
Considering the good activity of compounds 1-6, 8, and 9, especially compounds 1, 5, 8, and 9, which were superior to L-NIL in NO production inhibition (Table 3), these phenolic acid esters can become leading candidates to research and development agents for the treatment of inflammatory disease accompanying overproduction of NO. However, given the absorption properties, only compounds 2, 6, and 8 are considered to have therapeutic potential for the treatment of inflammation accompanying overproduction of NO due to their good or moderate membrane permeability in Caco-2 cell model (Table 2). Both ferulic acid ester derivatives, 1-O-feruloyl-2-O-p-coumaroylglycerol and 1,3-O-diferuloylglycerol, significantly decreased the production of NO in LPS-stimulated mouse macrophage RAW264.7 cells in a dose-dependent manner with IC 50 values of 9.12 ± 0.72 and 12.01 ± 1.07 µM, respectively, through acting on the NF-κB/MAPKs pathway [29], which provides guidance on further research on the underlying mechanism of these phenolic acid esters from N. incisium exerting NO inhibition.

Plant Material
The roots and rhizomes of Notopterygium incisium were gathered from the Danba county of the Sichuan province of China and were identified by Dr. Shun-Yuan Jiang of Sichuan Academy of Chinese the calibration curves. Methodology was examined for precision, accuracy, recovery, and stability (see Supplementary data, Table S1) and was demonstrated to meet the requirements of determination.

Cell Culture
The human intestinal Caco-2 cell line (ATCC #HTB-37) was purchased from American Type Culture Collection (ATCC, Rockville, MD" USA). The murine macrophage cell line RAW264.7 cell (3111C0001CCC000146) was obtained from the Cell Resource Center, IBMS, CAMS/PUMC (Beijing, China). The cell culture was carried out in a Sanyo MCO-15 AC carbon dioxide (CO 2 ) incubator (Sanyo Electric Co., Ltd., Osaka, Japan). The integrity of the Caco-2 cell monolayer was examined by measuring the transepithelial electrical resistance (TEER) with an epithelial voltohmmeter (EVOM, World Precision Instrument, Sarasota, FL, USA) [18].

Caco-2 Cell Permeability
The Caco-2 cells were maintained in DMEM containing 10% FBS, 1% NEAA (100×), 100 units/mL of penicillin, and 100 µg/mL of streptomycin, in a constant humidity atmosphere of 5% CO 2 and 95% air at 37 • C. For confluence and differentiation, cells were seeded at a density of 1 × 10 5 cells/cm 2 into 12-well Transwell plates (insert diameter 12 mm, pore size 3.0 µm, membrane growth area 1.12 cm 2 , Costar ® #3402) and were allowed to grow for 21 days before the permeation experiment. On Day 21, the monolayers with TEER values >500 Ω·cm 2 were qualified for the transport experiment. The transport study was initiated by the careful removal of the culture medium from AP and BL side of the inserts. Caco-2 monolayers were rinsed twice with pre-warmed HBSS and were incubated by pre-warmed HBSS for 30 min at 37 • C. Stock solutions of test phenolic acid esters were prepared in DMSO and diluted to 50 µM with HBSS. The final DMSO concentration was less than 2%, a concentration that did not alter cell viability or permeability. The assayed solutions (50 µM) were added to the AP side (0.5 mL, for absorption transport) or BL side (1.5 mL, for efflux transport) of the inserts, while the receiving chamber contained the corresponding volume of HBSS. Incubation was performed at 37 • C for 90 min, with shaking at 50 rpm. Samples were collected from the inserts, then frozen, lyophilized, redissolved in MeOH, and injected into HPLC system for quantitative analysis.
The P app values were calculated as the following equation: where dQ/dt is the rate of the appearance of the test compound on the receiver compartment (µmol/s), C 0 is the initial test compound concentration on the donor compartment (µmol/mL), and A is the surface area of Caco-2 monolayer (cm 2 ).

NO Inhibitory Assay
The NO inhibitory assay was carried out according to the previous method [12,14]. Briefly, RAW 264.7 cells were maintained in DMEM containing 10% FBS, in a constant humidity atmosphere of 5% CO 2 and 95% air at 37 • C, seeded in 96-well culture plates (Costar ® #3599, Cambridge, MA, USA) at a density of 3 × 10 4 cells/well for 12 h, and then stimulated with LPS (1 µg/mL) and treated with various concentrations of assayed compounds for 24 h. After that, the cell culture supernatant (100 µL) was collected to react with Griess reagent (100 µL) for 15 min at room temperature. The nitrite in culture medium was measured as an indicator of NO production. NaNO 2 was used to generate a standard curve, and NO production was determined by measuring the optical density at 540 nm in comparison with the standard curve. Thermo Multiskan MK 3 Automated Microplate Reader was Thermo-Labsystems (Franklin, MA, USA). The experiments were performed in parallel three times, and L-NIL was used as a positive control. Cell viability (>95%) was assessed by using an MTT assay. The IC 50 values were calculated by the software SPSS 16.0 (SPSS Inc., Chicago, IL, USA).

Statistical Analysis
The results presented in this study were the averages of at least three replicates and were presented as means ± SD. The data were analyzed by either t-test or nonparametric test after analysis of variance using SPSS 16.0. The level of significance was set at p < 0.05.

Conclusions
In conclusion, a new ferulic acid ester (1) along with eight known phenolic acid esters (2-9) was isolated from the roots and rhizomes of N. incisium. The absorption properties of these phenolic acid esters were estimated with human intestinal Caco-2 cell monolayer model. Moreover, these phenolic acid esters except for compound 7 exhibited significantly inhibitory effects on NO production and particularly ferulic acid esters may be considered as potential therapeutic agents in inflammatory diseases associated with NO overproduction. Considering the beneficial effects of compounds 2, 6, and 8 on the inhibition of NO production and their possibility of crossing the intestinal barrier, compounds 2, 6, and 8 may be considered to contribute to the anti-inflammatory activity of the roots and rhizomes of N. incisum.
Supplementary Materials: Supplementary materials are available online: Figures S1-S17 and Table S1.