Structural Features Defining NF-κB Inhibition by Lignan-Inspired Benzofurans and Benzothiophenes

A series of 2-arylbenzofurans and 2-arylbenzothiophenes was synthesized carrying three different side chains in position five. The synthesized compounds were tested for NF-κB inhibition to establish a structure activity relationship. It was found that both, the side chain in position five and the substitution pattern of the aryl moiety in position two have a significant influence on the inhibitory activity.


Introduction
Inflammation is a protective host response to infection or tissue damage (including stress or dyshomeostasis) [1,2]. Whereas an acute response resulting in the elimination of the noxious agent is beneficial, long lasting chronic inflammatory states contribute to the development of many pathologies like autoimmune, metabolic, cardiovascular and neurodegenerative diseases or cancer [2,3], PAMPs (pathogen-associated molecular patterns) as well as DAMPs (damage-associated molecular patterns) activate pattern recognition receptors (PRR) that transduce signals to NF-κB signaling pathways [2], which play a pivotal role in chronic and acute inflammation [4]. Thus, dampen NF-κB signaling will interfere with an inflammatory response.
Krameria lappacea (Dombey), Krameriaceae, is a tropical perennial shrub growing across South America. The extract of the Rhatany root was introduced into European medicine over 200 years ago as a remedy against stomach aches, diarrhea, menstrual problems, nose bleeds and oropharyngeal inflammation [5,6]. In a study on constituents of the Rhatany root, the group of Stuppner isolated eleven lignans from the dichloromethane extract of the root (Figure 1) [7].
These isolated lignan derivatives were pharmacologically characterized in topical anti-inflammatory in vivo experiments [8]. Two of the most potent compounds, 2-(2-hydroxy-4-methoxyphenyl)-5-(3hydroxypropyl)benzofuran 1 and (+)-conocarpan 3 ( Figure 1) inhibited edema development and infiltration by neutrophils time-dependently and comparably to indomethacin. In addition, all lignans were tested in vitro for their potential to inhibit the activation of the NF-κB signaling pathway and the activity of the pro-inflammatory enzymes COX-1, COX-2, 5-LO and mPGES-1. Determination of the IC 50 values for all compounds showed that inhibition of NF-κB is the most relevant mechanism likely contributing to the observed in vivo activity [8]. These isolated lignan derivatives were pharmacologically characterized in topical antiinflammatory in vivo experiments [8]. Two of the most potent compounds, 2-(2-hydroxy-4methoxyphenyl)-5-(3-hydroxypropyl)benzofuran 1 and (+)-conocarpan 3 ( Figure 1) inhibited edema development and infiltration by neutrophils time-dependently and comparably to indomethacin. In addition, all lignans were tested in vitro for their potential to inhibit the activation of the NF-κB signaling pathway and the activity of the pro-inflammatory enzymes COX-1, COX-2, 5-LO and mPGES-1. Determination of the IC50 values for all compounds showed that inhibition of NF-κB is the most relevant mechanism likely contributing to the observed in vivo activity [8].
Purification was performed on silica gel eluting with LP or LP/EtOAc mixtures (depending on the polarity of the product).

5-Chloro-2-(4-(methoxymethoxy)phenyl)benzo[b]furan (18a)
Prepared according to the general procedure A. mp 138-141 • C. 1  Procedure B: 5-chloro benzo-fused (1.0 mmol, 1.0 equiv.), allylboronic acid pinacol ester (1.5 mmol, 1.5 equiv.), cesium carbonate (486 mg, 1.5 mmol, 1.5 equiv.), Pd 2 (dba) 3 (45 mg, 0.05 mmol, 0.05 equiv.) and SPhos (41 mg, 0.1 mmol, 0.1 equiv.) was mixed in 2 mL of dried dioxane. The mixture was stirred at 100 • C for 5 h in argon atmosphere. The reaction mixture was cooled to room temperature and then diluted with 15 mL diethyl ether or ethyl acetate (depending on the polarity of the product) and filtered through a pad of celite. The organic phase was washed with saturated NH 4 Cl solution, once with brine and then dried over Na 2 SO 4 . The solvent was removed under reduced pressure. Purification was performed on silica gel eluting with LP or LP/EtOAc mixtures (depending on the polarity of the product).  Procedure C: 5-allyl benzo-fused derivatives (0.5 mmol, 1.0 equiv.) was dissolved in 0.5 mL dry THF then the solution was cooled to 0 • C. A 1M solution of BH 3 . THF (0.5 mL, 0.5 mmol, 1.0 equiv.) was added slowly. Afterwards, the reaction solution was warmed to room temperature and stirred for 24 h. On the other hand, a solution of 3M NaOH and H 2 O 2 30% was mixed in ratio of 2:3 and then cooled to 0 • C. After 24 h of reaction time, the reaction was cooled again to 0 • C and then the prepared solution of NaOH and H 2 O 2 (1.20 mL, 1.2 mmol NaOH and 7.8 mmol H 2 O 2 ) was added slowly. The reaction mixture was stirred at room temperature for 4 more hours, then diluted with 5 mL diethyl ether. The organic phase was washed with a saturated NH 4 Cl solution for 3 times, once with brine and dried over Na 2 SO 4 . The solvent was removed under reduced pressure. Purification was performed on silica gel eluting with LP/EtOAc mixtures.

General Procedure for Isomerization on Allyl Benzo-Fused Heterocycles
Procedure D: 5-allyl benzo-fused derivatives (0.5 mmol, 1.0 equiv.), Pd(dba) 2 (5.8 mg, 0.01 mmol, 0.02 equiv.), P( t Bu 3 ) . HBF 4 (5.8 mg, 0.02 mmol, 0.04 equiv.) and i PrCOCl (10 µL, 10.6 mg, 0.1 mmol, 0.2 equiv.) was mixed in 1 mL of degassed DMAc. The mixture was stirred at 100 • C for 6 h in argon atmosphere. The reaction mixture was cooled to room temperature and then diluted with 15 mL ethylacetate and filtered through a pad of celite. The organic phase was washed with a saturated NH 4 Cl solution for 3 times, once with brine and dried over Na 2 SO 4 . The solvent was removed under reduced pressure. Purification was performed on silica gel eluting with LP or LP/EtOAc mixtures (depending on the polarity of the product).

NF-κB Transactivation Activity
Measurement of the NF-κB transactivation activity was essentially performed as described previously [9]. Briefly, HEK-293 cells stably transfected with a NF-κB luciferase reporter (HEK293/NF-κB-luc cells, Panomics, RC0014) were loaded with CTG CMFDA (5-chloromethylfluorescein diacetate, Invitrogen) to stain living cells. 4 × 10 4 cells were seeded in 96-well plates and incubated at 37 • C and 5% CO 2 overnight. On the next day, the medium was exchanged with a serum-free DMEM and cells were treated with the respective test compounds dissolved in dimethyl sulfoxide (DMSO). To avoid nonspecific effects of the solvent, the final concentration of DMSO was always adjusted to 0.1%. One hour after the treatment, cells were stimulated with 2 ng/mL human recombinant TNF-α for 4 h. Then cells were lysed by a reporter lysis buffer (Promega, Madison, WI, USA). The luminescence of the firefly luciferase and the CTG fluorescence were quantified on a GeniosPro plate reader (Tecan, Grödig, Austria). The luciferase signal derived from the NF-κB reporter was normalized by the CTG-derived fluorescence to account for differences in the cell number or transfection efficiency.

Results and Discussion
Based on the results presented in the introduction, we were interested to synthesize and evaluate benzofuran-based lignans of general structure I (Figure 2). The synthetic strategy should be modular, efficient, and applicable to the synthesis of a wide range of derivatives, ideally from common intermediates. Breaking the strategic bonds as indicated in Figure 2, three fragments A, B, and C, are obtained which suggest the application of a direct arylation/cross coupling strategy, which introduces the aryl moiety in position two of general structure II via direct arylation and the alkyl or alkenyl chain in position five of the same fragment via a cross-coupling methodology, e.g., with an allyl boronic acid ester such as IV. For the direct arylation protocol we had previously developed an efficient protocol optimized for benzo-fused heterocyclic systems [10]. Hence, the main synthetic task was developing a suitable cross-coupling method introducing the desired residues in position five of building block II and subsequent elaboration of the olefin function towards the substituents identified for the naturally occurring compounds. Naturally, there are many ways to synthesize substituted benzofurans [11], and also the type of structure we are aiming for has been synthesized previously. For example, Duan et al. used a strategy in which the benzofuran core was constructed individually for each derivative synthesized [12]. This strategy is conveniently applicable for a small set of compounds, however, when a larger library of derivatives is targeted, it is lacking the modularity we were aiming for.
Consequently, fragment B was designed as a 5-halo-benzo[b]furan (general structure II, Figure 2). This building block offers two possibilities: Initial allylation at C5 and subsequent C2 arylation, or, alternatively, initial C2 arylation followed by C5-allylation. Since the direct arylation protocol uses aryl bromides as coupling partners, the allylation reaction should be carried out first to avoid homo-coupling between two 5-bromobenzofuran entities. The resulting 5-allyl benzofuran should then be arylated in position two, however it has to be considered that the terminal double bond could react in an undesired Mizoroki-Heck reaction. In case the order of events should be reversed, 5-chlorobenzofuran would need to be applied to avoid or at least suppress the aforementioned problem. To develop the most efficient protocol, it was decided to investigate both approaches. For introducing the side chain in position five, introduction of an allyl substituent is the ideal option, since the allyl substituent can be further transformed into 3-hydroxyl-propyl by a hydroboration-oxidation sequence [13][14][15] or to a 2-propenyl residue by an isomerization reaction [16][17][18][19][20][21][22][23].
Biomolecules 2020, 10, x 11 of 21 or, alternatively, initial C2 arylation followed by C5-allylation. Since the direct arylation protocol uses aryl bromides as coupling partners, the allylation reaction should be carried out first to avoid homocoupling between two 5-bromobenzofuran entities. The resulting 5-allyl benzofuran should then be arylated in position two, however it has to be considered that the terminal double bond could react in an undesired Mizoroki-Heck reaction. In case the order of events should be reversed, 5chlorobenzofuran would need to be applied to avoid or at least suppress the aforementioned problem. To develop the most efficient protocol, it was decided to investigate both approaches. For introducing the side chain in position five, introduction of an allyl substituent is the ideal option, since the allyl substituent can be further transformed into 3-hydroxyl-propyl by a hydroborationoxidation sequence [13][14][15] or to a 2-propenyl residue by an isomerization reaction [16][17][18][19][20][21][22][23]. Initially, 5-bromobenzo[b]furan 12 was tested as starting material (Scheme 1) [24][25][26][27]. The subsequent allylation reaction with 13 worked well using a Suzuki-Miyaura protocol with Pd(PPh3)2Cl2 as catalyst, K2CO3 as base and in DMAc solvent. Product 14 was obtained in 58% yield. Next, a direct arylation should introduce the aryl residue in position two, giving rise to 16. Unfortunately, the direct arylation reaction did not take place and only product 15 derived from competing Mizoroki-Heck coupling was found, instead. Initially, 5-bromobenzo[b]furan 12 was tested as starting material (Scheme 1) [24][25][26][27]. The subsequent allylation reaction with 13 worked well using a Suzuki-Miyaura protocol with Pd(PPh 3 ) 2 Cl 2 as catalyst, K 2 CO 3 as base and in DMAc solvent. Product 14 was obtained in 58% yield. Next, a direct arylation should introduce the aryl residue in position two, giving rise to 16. Unfortunately, the direct arylation reaction did not take place and only product 15 derived from competing Mizoroki-Heck coupling was found, instead. Hence, the alternative approach using 5-chlorobenzo[b]furan 17 as starting material was investigated (Scheme 2) [24,28]. In our previous publication [10], it was observed that aryl chlorides were unreactive as coupling partners under the reaction conditions and, hence, homo-coupling of the benzofuran starting material should not be an issue. Hence, the alternative approach using 5-chlorobenzo[b]furan 17 as starting material was investigated (Scheme 2) [24,28]. In our previous publication [10], it was observed that aryl chlorides were unreactive as coupling partners under the reaction conditions and, hence, homo-coupling of the benzofuran starting material should not be an issue.
were unreactive as coupling partners under the reaction conditions and, hence, homo-coupling of the benzofuran starting material should not be an issue.
Before direct arylation was tested on this substrate, the allylation in position five had to be established. Using the same allylation conditions as for 5-bromobenzofuran 12 did not lead to any conversion towards 14 and, hence, the Suzuki-Miyaura reaction had to be optimized. Several conditions for Suzuki-Miyaura reactions on chloride substrates have been reported [29][30][31][32], mainly using elaborate catalyst/ligand systems. However, the method of Thimmaiah et al. used common catalysts and ligands and was therefore very appealing for our purposes of an efficient and simple protocol [30]. After a quick optimization, 5-allylbenzo[b]furan 14 could be synthesized using Pd2 (dba)3 as catalyst, SPhos as ligand, Cs2CO3 as base in dioxane, 100 °C and for 5 h giving an isolated yield of 14 of 68% (Scheme 2).

Scheme 2. Allylation and attempted direct arylation of 5-chlorobenzo[b]furan 14.
Before direct arylation was tested on this substrate, the allylation in position five had to be established. Using the same allylation conditions as for 5-bromobenzofuran 12 did not lead to any conversion towards 14 and, hence, the Suzuki-Miyaura reaction had to be optimized. Several conditions for Suzuki-Miyaura reactions on chloride substrates have been reported [29][30][31][32], mainly using elaborate catalyst/ligand systems. However, the method of Thimmaiah et al. used common catalysts and ligands and was therefore very appealing for our purposes of an efficient and simple protocol [30]. After a quick optimization, 5-allylbenzo[b]furan 14 could be synthesized using Pd 2 (dba) 3 as catalyst, SPhos as ligand, Cs 2 CO 3 as base in dioxane, 100 • C and for 5 h giving an isolated yield of 14 of 68% (Scheme 2).
Product 16a represents a key compound for the synthetic route. From 16a, a hydroboration-oxidation sequence [13][14][15] using BH 3 . THF 1M for the first step and NaOH, H 2 O 2 in THF for the second step was utilized in situ to obtain alcohol 19a with 42% isolated yield. The MOM protecting group was cleaved in methanol with traces of concentrated HCl subsequently to obtain the final product 19b (Scheme 3). For the isomerization of 16a several different methods are reported using transition metal catalyst [16][17][18][19][20][21][22][23]. Only Gauthier et al. [23] used palladium catalyst with bulky ligand [23] to migrate the double bond into conjugation with the aryl ring with very good stereoselectivity. Using Pd(dba) 2 as catalyst and P( t Bu) 3 . HBF 4 as ligand product 20a was obtained in E configuration with 75% yield.
De-protection by concentrated HCl in methanol removed the MOM protective group to get final product 6 (Scheme 3). For structure activity relationship studies, the developed synthetic sequence should also be applied for the synthesis of bioisosteric benzo[b]thiophene derivatives. Required 5-chlorobenzo[b]thiophene 21 was synthesized according to a literature protocol [33]. The direct arylation reaction on 21 was also selective giving 22a in 58% yield. Allylation via Suzuki-Miyaura coupling on 22a worked under the identical conditions as developed for the benzofuran series, albeit in somewhat lower yield giving 23a in 46% (Scheme 4).
Product 16a represents a key compound for the synthetic route. From 16a, a hydroborationoxidation sequence [13][14][15] using BH3 . THF 1M for the first step and NaOH, H2O2 in THF for the second step was utilized in situ to obtain alcohol 19a with 42% isolated yield. The MOM protecting group was cleaved in methanol with traces of concentrated HCl subsequently to obtain the final product 19b (Scheme 3).

Scheme 3. Overview of double bond isomerization, hydroboration -oxidation, and MOMdeprotection of 16a.
For the isomerization of 16a several different methods are reported using transition metal catalyst [16][17][18][19][20][21][22][23]. Only Gauthier et al. [23] used palladium catalyst with bulky ligand [23] to migrate the double bond into conjugation with the aryl ring with very good stereoselectivity. Using Pd(dba)2 as catalyst and P( t Bu)3 . HBF4 as ligand product 20a was obtained in E configuration with 75% yield. De-protection by concentrated HCl in methanol removed the MOM protective group to get final product 6 (Scheme 3).
For structure activity relationship studies, the developed synthetic sequence should also be applied for the synthesis of bioisosteric benzo[b]thiophene derivatives. Required 5chlorobenzo[b]thiophene 21 was synthesized according to a literature protocol [33]. The direct arylation reaction on 21 was also selective giving 22a in 58% yield. Allylation via Suzuki-Miyaura coupling on 22a worked under the identical conditions as developed for the benzofuran series, albeit in somewhat lower yield giving 23a in 46% (Scheme 4).

Scheme 4. Direct arylation and allylation of 5-chlorobenzo[b]thiophene 21.
The reactions for side chain modification were carried out again under identical conditions as in the benzofuran series. The hydroboration-oxidation sequence of 23a gave 46% isolated yield of 24a while isomerization to 25a gave 71% yield. De-protection reaction on 24a and 25a proceeded in very good yield of 89% toward 24b and 90% towards 25b, respectively (Scheme 5). For the isomerization of 16a several different methods are reported using transition metal catalyst [16][17][18][19][20][21][22][23]. Only Gauthier et al. [23] used palladium catalyst with bulky ligand [23] to migrate the double bond into conjugation with the aryl ring with very good stereoselectivity. Using Pd(dba)2 as catalyst and P( t Bu)3 . HBF4 as ligand product 20a was obtained in E configuration with 75% yield. De-protection by concentrated HCl in methanol removed the MOM protective group to get final product 6 (Scheme 3).
For structure activity relationship studies, the developed synthetic sequence should also be applied for the synthesis of bioisosteric benzo[b]thiophene derivatives. Required 5chlorobenzo[b]thiophene 21 was synthesized according to a literature protocol [33]. The direct arylation reaction on 21 was also selective giving 22a in 58% yield. Allylation via Suzuki-Miyaura coupling on 22a worked under the identical conditions as developed for the benzofuran series, albeit in somewhat lower yield giving 23a in 46% (Scheme 4). The reactions for side chain modification were carried out again under identical conditions as in the benzofuran series. The hydroboration-oxidation sequence of 23a gave 46% isolated yield of 24a while isomerization to 25a gave 71% yield. De-protection reaction on 24a and 25a proceeded in very good yield of 89% toward 24b and 90% towards 25b, respectively (Scheme 5). The reactions for side chain modification were carried out again under identical conditions as in the benzofuran series. The hydroboration-oxidation sequence of 23a gave 46% isolated yield of 24a while isomerization to 25a gave 71% yield. De-protection reaction on 24a and 25a proceeded in very good yield of 89% toward 24b and 90% towards 25b, respectively (Scheme 5). Initially, a series of direct arylation reactions was conducted (Table 1)

With a practical synthesis route for benzo[b]furan and benzo[b]thiophene compounds at hand, a group of lignan-like compounds based on those heterocyclic rings was prepared to evaluate their biological properties as anti-inflammatory agents.
Initially, a series of direct arylation reactions was conducted (Table 1) with aryl moieties carrying electron donating and electron withdrawing substituents. The nature of the substituent had only a minor influence on the yield of this transformation. In the benzofuran series, aryl bromides carrying electron donating substituents such as OMOM or methoxy gave 10-15% higher yields in the arylation step (see Table 1 examples 18a, 18c and 18d) as compared to substituents with no or an electron withdrawing effect (see Table 1 examples 18e-h). In the benzothiophene series only two examples were synthesized and hence a general trend cannot be deduced. Initially, a series of direct arylation reactions was conducted (Table 1) with aryl moieties carrying electron donating and electron withdrawing substituents. The nature of the substituent had only a minor influence on the yield of this transformation. In the benzofuran series, aryl bromides carrying electron donating substituents such as OMOM or methoxy gave 10-15% higher yields in the arylation step (see Table 1 examples 18a, 18c and 18d) as compared to substituents with no or an electron withdrawing effect (see Table 1 examples 18e-h). In the benzothiophene series only two examples were synthesized and hence a general trend cannot be deduced.
The subsequent allylation reactions worked well on all benzofuran substrates giving yields in the range of 66-74%. Only the 2-Cl product 16h was obtained in somewhat lower yield of 51%. Important to note: compound 16b was obtained via MOM-deprotection of 16a in 91% yield rather than via allylation of the corresponding 4-OH-aryl precursor 18b. The direct arylation procedure turned out not to tolerate a free OH group, hence, requiring this alternate approach to 18b.
n.s. not synthesized; 1 via MOM-deprotection from 16a, the yield refers only to the deprotection step.
The subsequent allylation reactions worked well on all benzofuran substrates giving yields in the range of 66-74%. Only the 2-Cl product 16h was obtained in somewhat lower yield of 51%. Important to note: compound 16b was obtained via MOM-deprotection of 16a in 91% yield rather than via allylation of the corresponding 4-OH-aryl precursor 18b. The direct arylation procedure turned out not to tolerate a free OH group, hence, requiring this alternate approach to 18b.
The hydroboration-oxidation sequence towards the terminal alcohol products 19a-h (benzofuran series) and 24a-c (benzothiophene series) proceeded with similar efficiency (40-53% yield) independent of the substituents present on the aryl ring (Table 2, left). Again, it should be noted that the 4-OH products 19b and 24b were obtained in excellent yield via MOM-deprotection of 19a and 24a respectively. The same is true for the double bond isomerization (Table 2, right). Benzofuran products 20a-h and benzothiophene compounds 25a-c were obtained in yields between 57-77% yield. Also, in this case the 4-OH products 6 and 25b were obtained in excellent yield via MOM-deprotection of 20a and 25a, respectively.
Since the pharmacological characterization of the lignan derivatives isolated from Krameria lappaceae revealed as most relevant in vitro anti-inflammatory activity inhibition of the NF-κB signaling pathway, we decided to use again a luciferase reporter model to quantify the transactivation activity of NF-κB [8]. For this, we used HEK293 cells stably transfected with a NF-κB-driven luciferase reporter gene that were loaded with fluorescent Cell Tracker Green to allow luciferase-derived signal normalization to the amount of viable cells. Cells were then treated with test compounds at the indicated concentration or vehicle for 30 min and then stimulated with TNF-α (2 ng/mL) for four hours. Luminescence and fluorescence was quantified in cell lysates by a Genios Pro plate reader (Tecan) [34]. Table 2. Hydroboration-oxidation and isomerization.
The hydroboration-oxidation sequence towards the terminal alcohol products 19a-h (benzofuran series) and 24a-c (benzothiophene series) proceeded with similar efficiency (40-53% yield) independent of the substituents present on the aryl ring (Table 2, left). Again, it should be noted that the 4-OH products 19b and 24b were obtained in excellent yield via MOM-deprotection of 19a and 24a respectively. The same is true for the double bond isomerization(Table 2, right). Benzofuran products 20a-h and benzothiophene compounds 25a-c were obtained in yields between 57-77% yield. Also, in this case the 4-OH products 6 and 25b were obtained in excellent yield via MOMdeprotection of 20a and 25a, respectively.
Since the pharmacological characterization of the lignan derivatives isolated from Krameria lappaceae revealed as most relevant in vitro anti-inflammatory activity inhibition of the NF-κB signaling pathway, we decided to use again a luciferase reporter model to quantify the transactivation activity of NF-κB [8]. For this, we used HEK293 cells stably transfected with a NF-κBdriven luciferase reporter gene that were loaded with fluorescent Cell Tracker Green to allow luciferase-derived signal normalization to the amount of viable cells. Cells were then treated with test compounds at the indicated concentration or vehicle for 30 min and then stimulated with TNFα (2 ng/mL) for four hours. Luminescence and fluorescence was quantified in cell lysates by a Genios Pro plate reader (Tecan) [34].

Discussion
The biological data are compiled in Table 3. With the current synthetic route towards target compounds, structural diversity at position five of the benzo-heteroaromatic core could be further 1 the yield refers to the MOM-deprotection step of 19a; 2 the yield refers to the MOM-deprotection step of 24a; 3 the yield refers to the MOM-deprotection step of 20a; 4 the yield refers to the MOM-deprotection step of 25a.

Discussion
The biological data are compiled in Table 3. With the current synthetic route towards target compounds, structural diversity at position five of the benzo-heteroaromatic core could be further extended regarding the location of the olefinic system. Initially, 2-phenyl benzofurans were tested carrying an allyl group (entry 1, 16e), a 1-propenyl group (entry 2, 20e), or a 1-hydroxy-propan-3-yl group (entry 3, 19e) in position five of the benzofuran scaffold. NF-κB inhibition was measured initially at concentrations of 10 µM and/or at 20 µM. IC 50 values were determined for compounds with a significant and concentration-dependent inhibitory activity at these concentrations. Within this initial series, it was found that the propanol substituted derivative 19e (entry 3) showed the highest inhibition (0.1-fold activation at 10 µM and 0.03-fold activation at 20 µM) and already a low IC 50 value of 1.42 µM. The 5-allyl substituted derivative 16e as well as the 5-(1-propenyl) derivative 20e showed significantly lower inhibition, whereas the latter one gave the lowest activity (0.36 at 10 µM entry 1 vs. 0.66 at 10 µM entry 2). Table 3. Pharmacological data.

16e
Biomolecules 2020, 10, x 16 of 21 extended regarding the location of the olefinic system. Initially, 2-phenyl benzofurans were tested carrying an allyl group (entry 1, 16e), a 1-propenyl group (entry 2, 20e), or a 1-hydroxy-propan-3-yl group (entry 3, 19e) in position five of the benzofuran scaffold. NF-κB inhibition was measured initially at concentrations of 10 µM and/or at 20 µM. IC50 values were determined for compounds with a significant and concentration-dependent inhibitory activity at these concentrations. Within this initial series, it was found that the propanol substituted derivative 19e (entry 3) showed the highest inhibition (0.1-fold activation at 10 µM and 0.03-fold activation at 20 µM) and already a low IC50 value of 1.42 µM. The 5-allyl substituted derivative 16e as well as the 5-(1-propenyl) derivative 20e showed significantly lower inhibition, whereas the latter one gave the lowest activity (0.36 at 10 µM entry 1 vs. 0.66 at 10 µM entry 2). extended regarding the location of the olefinic system. Initially, 2-phenyl benzofurans were tested carrying an allyl group (entry 1, 16e), a 1-propenyl group (entry 2, 20e), or a 1-hydroxy-propan-3-yl group (entry 3, 19e) in position five of the benzofuran scaffold. NF-κB inhibition was measured initially at concentrations of 10 µM and/or at 20 µM. IC50 values were determined for compounds with a significant and concentration-dependent inhibitory activity at these concentrations. Within this initial series, it was found that the propanol substituted derivative 19e (entry 3) showed the highest inhibition (0.1-fold activation at 10 µM and 0.03-fold activation at 20 µM) and already a low IC50 value of 1.42 µM. The 5-allyl substituted derivative 16e as well as the 5-(1-propenyl) derivative 20e showed significantly lower inhibition, whereas the latter one gave the lowest activity (0.36 at 10 µM entry 1 vs. 0.66 at 10 µM entry 2). extended regarding the location of the olefinic system. Initially, 2-phenyl benzofurans were tested carrying an allyl group (entry 1, 16e), a 1-propenyl group (entry 2, 20e), or a 1-hydroxy-propan-3-yl group (entry 3, 19e) in position five of the benzofuran scaffold. NF-κB inhibition was measured initially at concentrations of 10 µM and/or at 20 µM. IC50 values were determined for compounds with a significant and concentration-dependent inhibitory activity at these concentrations. Within this initial series, it was found that the propanol substituted derivative 19e (entry 3) showed the highest inhibition (0.1-fold activation at 10 µM and 0.03-fold activation at 20 µM) and already a low IC50 value of 1.42 µM. The 5-allyl substituted derivative 16e as well as the 5-(1-propenyl) derivative 20e showed significantly lower inhibition, whereas the latter one gave the lowest activity (0.36 at 10 µM entry 1 vs. 0.66 at 10 µM entry 2).     Naturally occurring benzofuran lignans contain at least one oxygen functionality (OH or OMe) in the aryl ring in position two, compounds 16b, 19b, and 6 with a 4-hydroxyphenyl group in that position and the three different side chains in position five were tested (entries 4-6). Obviously, the hydroxyl group is very important for NF-κB inhibition since all three derivatives show significant inhibition at both, 10 and 20 µM concentration. Due to the small differences in the inhibitory effect, a trend between the three derivatives cannot be deduced, however, again the propanol substituted derivative 19b (entry 6) showed the lowest IC50 value with 1.24 µM, which is also the lowest value of all tested compounds.
When a methoxy group is placed in position four of the phenyl ring instead of a hydroxy group, the trend that the highest inhibition is found with 5-propanol substituted benzofurans (entry 9, compound 19c) followed by 5-allyl (entry 7, compound 16c) and 5-propenyl (entry 8, compound 20c) is clearly reestablished. Compound 19c (entry 9) shows a similar inhibitory effect as the other two propanol substituted derivatives (entries 3 and 6 compounds 19e and 19b) but with a significantly higher IC50 value of 3.82 µM (vs. 1.42 entry 3, and 1.24 entry 6). Naturally, the methoxy group in 19c is significantly larger than a proton (as in 19e) or a hydroxy group (as in 19b) and it was speculated that this size difference might have an influence on the IC50 values. Hence, in a next set of compounds, the methoxy group was substituted for a MOM group, which further increases the steric bulk.
In this set of compounds (entries 10-12, compounds 16a, 20a, and 19a respectively), significant inhibitory effects were found with the 5-allyl-and 5-propanol substituted derivatives 16a and 19a, but only at the higher concentration of 20 µM. The 0.03-fold NF-κB activation of 19a matches however the best values obtained so far. However, as hypothesized, the IC50 value is significantly higher with 9.22 µM (entry 12). Interestingly, 16a gave a very low IC50 value of 1.31 µM. This supports the argument that steric bulk in the phenyl ring influences the IC50 values.
Since naturally occurring benzofuran lignans often carry two oxygen functionalities (OH, OMe), it was tried to access such compounds synthetically. Unfortunately, our synthetic method allowed us only access to 3,5-dimethoxy substituted derivatives 16d, 20d, and 19d (entries [13][14][15]. Again, the 5propanol substituted derivative 19d (entry 15) showed highest NF-κB inhibition (0.20 at 10 µM) and the overall highest NF-κB inhibition of 0.003 at 20 µM. The IC50 value for this compound was surprisingly low (1.92 µM, entry 15), which is contradicting the steric argument previously considered. However, this compound is the only one carrying a 3,5-disubstituted phenyl ring and additional derivatives incorporation this substitution pattern would be required to further establish structure-activity relationship. The two oxygen moieties might lead to favorable interactions, which predominate over steric effects.
At this point it is safe to say that a propanol substituent in position five usually gives highest NF-κB inhibition, which is sometimes matched by 5-allyl substituted derivatives. The propenyl derivatives are largely inactive and were excluded in further biological evaluation.
So far only electron donating substituents on the phenyl ring were considered, in line with the substitution pattern of the natural products. Since fluorine substituents often beneficially influence factors such as lipophilicity and hence also biological activity, two fluorine containing derivatives were tested as well (19g entry 16 and 19f entry 17), both carrying the 5-propanol residue. NF-κB inhibition was mediocre in both compounds (entries 16 and 17) at 10 µM concentration, however at 20 µm, especially 19f showed significant inhibition. The IC50 values again showed the trend that the larger 4-CHF2 substituent gave a significantly higher IC50 value (19g, entry 16 Naturally occurring benzofuran lignans contain at least one oxygen functionality (OH or OMe) in the aryl ring in position two, compounds 16b, 19b, and 6 with a 4-hydroxyphenyl group in that position and the three different side chains in position five were tested (entries 4-6). Obviously, the hydroxyl group is very important for NF-κB inhibition since all three derivatives show significant inhibition at both, 10 and 20 µM concentration. Due to the small differences in the inhibitory effect, a trend between the three derivatives cannot be deduced, however, again the propanol substituted derivative 19b (entry 6) showed the lowest IC 50 value with 1.24 µM, which is also the lowest value of all tested compounds.
When a methoxy group is placed in position four of the phenyl ring instead of a hydroxy group, the trend that the highest inhibition is found with 5-propanol substituted benzofurans (entry 9, compound 19c) followed by 5-allyl (entry 7, compound 16c) and 5-propenyl (entry 8, compound 20c) is clearly reestablished. Compound 19c (entry 9) shows a similar inhibitory effect as the other two propanol substituted derivatives (entries 3 and 6 compounds 19e and 19b) but with a significantly higher IC 50 value of 3.82 µM (vs. 1.42 entry 3, and 1.24 entry 6). Naturally, the methoxy group in 19c is significantly larger than a proton (as in 19e) or a hydroxy group (as in 19b) and it was speculated that this size difference might have an influence on the IC 50 values. Hence, in a next set of compounds, the methoxy group was substituted for a MOM group, which further increases the steric bulk.
In this set of compounds (entries 10-12, compounds 16a, 20a, and 19a respectively), significant inhibitory effects were found with the 5-allyl-and 5-propanol substituted derivatives 16a and 19a, but only at the higher concentration of 20 µM. The 0.03-fold NF-κB activation of 19a matches however the best values obtained so far. However, as hypothesized, the IC 50 value is significantly higher with 9.22 µM (entry 12). Interestingly, 16a gave a very low IC 50 value of 1.31 µM. This supports the argument that steric bulk in the phenyl ring influences the IC 50 values.
Since naturally occurring benzofuran lignans often carry two oxygen functionalities (OH, OMe), it was tried to access such compounds synthetically. Unfortunately, our synthetic method allowed us only access to 3,5-dimethoxy substituted derivatives 16d, 20d, and 19d (entries [13][14][15]. Again, the 5-propanol substituted derivative 19d (entry 15) showed highest NF-κB inhibition (0.20 at 10 µM) and the overall highest NF-κB inhibition of 0.003 at 20 µM. The IC 50 value for this compound was surprisingly low (1.92 µM, entry 15), which is contradicting the steric argument previously considered. However, this compound is the only one carrying a 3,5-disubstituted phenyl ring and additional derivatives incorporation this substitution pattern would be required to further establish structure-activity relationship. The two oxygen moieties might lead to favorable interactions, which predominate over steric effects.
At this point it is safe to say that a propanol substituent in position five usually gives highest NF-κB inhibition, which is sometimes matched by 5-allyl substituted derivatives. The propenyl derivatives are largely inactive and were excluded in further biological evaluation.
So far only electron donating substituents on the phenyl ring were considered, in line with the substitution pattern of the natural products. Since fluorine substituents often beneficially influence factors such as lipophilicity and hence also biological activity, two fluorine containing derivatives were tested as well (19g entry 16 and 19f entry 17), both carrying the 5-propanol residue. NF-κB inhibition was mediocre in both compounds (entries 16 and 17) at 10 µM concentration, however at 20 µm, especially 19f showed significant inhibition. The IC 50 values again showed the trend that the larger 4-CHF 2 substituent gave a significantly higher IC 50 value (19g, entry 16,8.52 µM) as compared to the 4-F substituted derivative (19f, entry 17, 2.20 µM).
In one example (19h), ortho substitution in the phenyl ring was tested as well, but no significant NF-κB inhibition was obtained (entry 18).
In a next series of compounds, we tested whether the benzofuran core could be substituted by other benzothiophene. Hence, several benzothiophene derivatives were synthesized and the 4-hydroxyphenyl-(24b, entry 19), 4-methoxyphenyl-(24c, entry 20) and 4-MOM-phenyl-(24a, entry 21) derivatives were tested. In all three examples position five was substituted by the propanol side-chain. The MOM substituted derivative did not show NF-κB inhibition (entry 21), which is surprising since the corresponding benzofuran derivative (entry 9) was amongst the most active ones. The 4-OH (entry 19) and 4-MeO (entry 20) derivative showed NF-κB inhibition, especially at the higher concentration of 20 µM, however the corresponding IC 50 values were significantly higher as compared to their benzofuran counterparts (see entry 6 vs entry 19 and entry 9 vs. entry 20).
A comparison of the pharmacological data of our synthesized compounds and the benzofuran lignans isolated from krameria lappacea roots ( Figure 1, compounds 1, 2, and 5-9) shows that the IC 50 values of several compounds are in the same range (or even lower) as the most active natural product 6. For compound 6 an IC 50 of 1.4 µM was reported in literature for the natural product isolate and we measured a similar value of 2.86 µM in our assay with a synthetic sample of 6. Compound 19b, which differs from 6 only in the sidechain (19b: 3-hydroxypropyl, 6: prop-1-en-1-yl) gave an IC 50 value of 1.24 µM. The difference between synthetic and natural 6 should not be overinterpreted, since the values stem from different measurement series carried out by different researchers.

Conclusions
It was found that the propanol side chain in position five is required for good inhibitory activity, independent of the underlying scaffold (benzofuran or benzothiophene). Some 5-allyl compounds do show activity as well, however those are the exception, and 5-propenyl derivatives are basically inactive. The IC 50 values seem to correlate with steric bulk in the aryl-moiety in position two, a finding which we want to confirm further in subsequent studies. The larger the substituents get in this ring, the higher the IC 50 values become. It can be speculated that there is a certain size restriction in the active site in this position. One examples does not follow this trend, however, in comparison to other compounds, the aryl ring is disubstituted in this example: the 3,5-dimethoxy compounds shows a low IC 50 value, but other interactions induced by the two methoxy groups might be responsible for this. Here, more examples with different substitution patterns are required to complete the picture. With the established synthetic route towards this compound class, further elaboration of this scaffold has been enabled and additional studies to establish refined structure activity relationship will be conducted in our laboratories. The focus will lie on benzofurans carrying multiple oxygen-functionalities in the aromatic ring in position two. Additionally, further substitution in the benzofuran system besides a side chain in position five is not explored yet and will be investigated.