Discovery of novel paeonol-based derivatives against skin inflammation in vitro and in vivo

Abstract T-LAK-cell-originated protein kinase (TOPK), a novel member of the mitogen-activated protein kinase family, is considered an effective therapeutic target for skin inflammation. In this study, a series (A − D) of paeonol derivatives was designed and synthesised using a fragment growing approach, and their anti-inflammatory activities against lipopolysaccharide (LPS)-induced nitric oxide production in RAW264.7 cells were tested. Among them, compound B12 yielded the best results (IC50 = 2.14 μM) with low toxicity (IC50 > 50 µM). Preliminary mechanistic studies indicated that this compound could inhibit the TOPK-p38/JNK signalling pathway and phosphorylate downstream related proteins. A murine psoriasis-like skin inflammation model was used to determine its therapeutic effect.

Paeonol isolated from traditional Chinese herbal medicines and used as a topical drug in China 15,16 exhibits a wide range of biological effects, including anti-inflammatory, immune regulatory, anti-tumour, and anti-oxidative effects [18][19][20] . Studies have indicated that paeonol could inhibit the expression of TOPK with antiinflammatory activity at 100 mM 16 and a microthermophoresis (MST) assay confirmed the affinity of paeonol for TOPK. Phosphorylation levels of p38, JNKs, MSK1 and histone H2AX were suppressed by paeonol by inhibiting TOPK activity in a time-and dose-dependent manner in vitro and in vivo 17 . Although paeonol has multiple sites for structural modification, modification of the hydroxyl group is a general means. Among them, compounds 1 and 2, reported by Huang, showed strong anti-inflammatory activity and were obtained by introducing bromine and a long alkyl chain 21 . Compound 3 showed suitable anti-inflammatory activity and was modified at its hydroxyl 22 (Figure 1(A)). In general, the derivatives of paeonol by modification of the 2-OH or 1-acetyl groups have shown anti-inflammatory activities. However, their SARs and specific target characteristics for anti-inflammation remain unclear. Additionally, modification of the substituents at the 5-position of paeonol and evaluating their anti-inflammatory properties have rarely been reported, posing a challenging project due to their unknown anti-inflammatory activities.
Herein, we present the design and synthesis of paeonol derivatives by the introduction of a urea linker at the 5-position of paeonol so as to discover potent compounds that regulate the TOPK pathway. The urea moiety is inherent in a various clinically approved drugs such as sorafenib, celiprolol and boceprevir ( Figure 1(B)) because of their strong ability to form multiple stable hydrogen bonds with protein or receptor targets 23 . Thus, a urea linker, which could modulate drug potency and selectivity and improving drug properties, was introduced at the ortho-position of the methoxy group. Meanwhile, to balance the activity and toxicity and explore the structure-activity relationship, the linker was substituted with an array of amines, including alkyl amines, heterocyclic/cyclic amines, and aniline ( Figure 1(B)).

Materials and methods
All commercial reagents and solvents were purchased from commercial suppliers and used without further purification. Reactions were monitored by thin-layer chromatography (TLC) and visualised under UV light at 254 and 365 nm. The chromatograms were conducted on silica gel (3 0 0 $ 400 mesh) using combiflash chromatography systems (Teledyne ISCO R f 200). 1 H and 13 C NMR spectral data were recorded with a Bruker or Agilent, 400, 600 MHz spectrometer in CDCl 3 or DMSO-d 6 using tetramethylsilane (TMS) as internal standard (IS). High-resolution mass spectrometry (HRMS) was recorded on an Agilent Technologies LC-TOF instrument.
1-(5-amino-2-hydroxy-4-methoxyphenyl) ethan-1-one (2). Compound 1 (2 g, 10 mmol) was dissolved in methanol (40 ml) and 10% Pd/C (4% mmol) was added. The suspension was acidified to pH 3 $ 4 by addition of 1 M HCl solution. The mixture was vigorously stirred under hydrogen gas at atmospheric pressure about 4 h. Upon completion of the reaction (as determined by TLC), the catalyst was removed by filtration and the filtrate was concentrated under reduced pressure. The residue was recrystallised with ethanol to give compound 2 (1.7 g, 95% yield) as light yellow crystal. 1   3H). 13  Griess reagent were mixed at a ratio of 1:1, and reacted at room temperature for 15 min. Measurement of absorbance at 450 nm with enzyme reader. At least three independent experiments were performed for the determination of IC 50 values, which were shown as mean ± SD. All experimental data were analysed by using GraphPad Prism software, version 8.0 (GraphPad Inc.).

Cell viability assay (MTT)
RAW264.7 cells (1 Â 10 4 cells/well) were seeded into 96-well plate and used for experiments after 24 h. Compounds (4 mM, 20 mM and 100 mM) treated cell for 24 h. Forty microliters of MTT solution (5 mg/mL, Sigma-Aldrich) was then added and incubated for an additional 4 h. After 4 h, cell culture supernatants were removed and DMSO (150 mL) was added into per well for dissolving the resulting crystals. Shaking about 1 0 $ 15 min and the absorbance at 492 nm was measured by a microplate reader (MQX200, Bio-Tek, Winooski, VT). At least three independent experiments were performed for the determination of IC 50 values, which were shown as mean ± SD. All experimental data were analysed by using GraphPad Prism software, version 8.0 (GraphPad Inc.).

Hyperproliferation assay
HaCaT cells were seeded in 96-well plates and cultured at 37 C in a humidified 5% CO 2 atmosphere. When the cell confluency reached 60-70%, HaCaT cells were then starved in serum-free DMEM for 12 h before stimulated by 100 ng/mL LPS for 24 h to mimic hyperproliferative psoriatic keratinocytes. The cell viability was measured by MTT method. Results are expressed as the percentage survival of cells in the cell treated group and the untreated control group (without the tested compound and LPS).

Western blotting
RAW264.7 cells were seeded into 6-well plate with 2 Â 10 6 cells per well and maintained about 24 h, and then pre-treated with compound B12 (2.5, 5, 10 mM) for 1 h, co-treated with LPS (0.5 mg/mL) for 0.5 h or 24 h. Western blotting assay was performed as the method described previously 24 . Briefly, the cells were lysed in 300 ml RIPA cell lysis buffer (Contains PMSF and phosphatase inhibitors, Beyotime China) and incubated on ice for 30 min. Collecting supernatant by centrifugation. The same amount of protein was separated by SDS-PAGE and transferred to PVDF membrane (GE Healthcare, Amersham, UK). The blotted membrane incubated with the primary antibodies and allowed to react for an additional 16 h at 4 C. All antibodies obtain from Cell Signalling Technology, Boston, MA. The membranes incubated with a 1:5000 dilutions of HRP-conjugated secondary antibody (Beyotime Biotech, Nantong, China) for 1 h at room temperature. Signals were visualised using ECL system (Thermo Fisher Scientific, Waltham, MA).
HaCaT cells were seeded into 6-well plate with 2 Â 10 6 cells per well. After 24 h culture and 12 h starvation, they were stimulated with SUV irradiation. The cells were irritated with 40 KJ/m 2 doses of SUV at 30 min. In order to study the effect of compound B12, the cells were starved in serum-free medium for 12 h, and then pre-treated with compound B12 (2.5, 5, 10 lM) for 6 h before SUV irradiation. Follow the same steps as above to transfer the electrophoretic band to PVDF membrane. All steps were performed in the same way as above except that the types of primary and secondary antibodies were different.

Animal study
Inbred 6-8-week-old female BALB/c mice were obtained from Animal Department of Anhui Medical University (China). All mice were allowed to acclimatise in the specific pathogen free (SPF) conditions for 1 week before any experiments were started. Mice were raised in a 12 h light/dark cycle with humidity (55 ± 5%) and temperature (22 ± 1 C). All animal protocols were approved by the Ethics Committee in Animal Experimentation at Anhui Medical University (Hefei, China) following the guidelines for Care and Use of Laboratory Animals. To induce the experimental psoriatic skin inflammation, mice were randomly divided into 5 groups (n ¼ 6): Normal, Vehicle (Imiquimod (IMQ), Sichuan Mingxin Pharmaceutical, Sichuan, China applied only), Dexamethasone treatment (IMQ with 20 mg/kg Dex), Compound B12 L (IMQ with 20mg/kg Compound B12), and Compound B12 H (IMQ with 40 mg/kg Compound B12). The psoriatic skin inflammation was established by topically applying 62.5 mg of IMQ cream on the shaved 2 cm Â 3 cm back skins. All rats were given intragastric administration once a day. During treatment for 7 days, to score the severity of inflammation of mouse skin, an objective scoring system was developed based on the clinical psoriasis area and severity index. Erythema, scaling, and thickening were scored independently from 0 to 4 as follows: 0, none; 1, slight; 2, moderate; 3, marked; 4, very marked and weight loss. The cumulative score (erythema plus scaling plus thickening) served to indicate the severity of inflammation (scale 0 $ 12). On the 8th day, the mice were taken eyeball blood and back skin. One-half of the samples were immediately fixed in 4% paraformaldehyde and for haematoxylin and eosin (H&E) staining. The other samples were put in a À80 C freezer. Before used, they were placed at room temperature for 30 min. After that, they were added 1 Â PBS proportionally, homogenised and centrifuged. The supernatant was collected and used for ELISA assay and Western blot assay. All animal studies were conducted according to the guidelines approved by the Laboratory Animal Centre of Anhui University. The histological image was obtained using 3DHISTECH's Slide Converter (3DHISTECH, Hungary).   3 Results and discussion

Chemistry
Compounds A1ÀA14, B1ÀB13, C1ÀC7 and D1ÀD4 were synthesised from paeonol as described in Scheme 1. First, nitration was achieved by treatment of paeonol with nitric acid in acetic acid, giving the nitro paeonol 1 at a 57% yield 25 . Subsequently, a selective reduction of 1 was performed by hydrogenation under 1 bar of hydrogen pressure using Pd/C to provide amino 2 at a 95% yield. Then 5-Amino-paeonol 2 was reacted with triphosgene in DCM to give crude 5-isocyanato-paeonol, directly reacted with various amines to give the desired compounds A1 À 14, B1 À 13, C1 À 7, and D1 À 4. Following the procedure, compounds A1ÀA14 substituted with alkyl amines were obtained in moderate yields; compounds B1ÀB13 substituted with aniline derivatives were produced in yields of 15 À 55%. The detailed synthesis procedures are described in the experimental section. The chemical structures of the compounds were confirmed by 1 H NMR, 13 C NMR and HRMS.

Inhibiting of NO release
Before determining the anti-inflammatory activity of the title compounds, cytotoxic activities were evaluated by an MTT assay. As shown in Figure 2(A-C), none of compounds showed significant intrinsic cytotoxicity against RAW264.7, LO2 or HaCat cells at 100 mM. Excessive NO production plays an important role in many inflammatory diseases. Subsequently, the anti-inflammatory activity of compounds was determined in LPS-stimulated RAW264.7 cells according to NO release 24,26,27 . As shown in Figure 2(D), most of compounds reduced LPS-induced NO secretion at 10 mM. And some compounds exhibited stronger inhibitory activity compared than celecoxib and paeonol. Compounds A1ÀA8 showed weak potencies as well as paeonol, while compounds A9ÀA14 showed a slight increase in potency. Based on this structural optimisation, it was found that alkyl or heterocyclic/cyclic amine-modified paeonol showed unsatisfactory results.
To improve anti-inflammatory activity and clearly explore structure-activity relationships, phenylurea, benzylurea and phenethylurea were also introduced at the 5-positon of paeonol. Compounds B1, C2, and D1, substituted with aniline, benzylamine, and phenethylamine, respectively, were synthesised and dramatically different inhibitory activities were observed (IR ¼ 51%, 7.86% and 15.26%, respectively). Aniline-substituted B1 proved to be a privileged scaffold. Subsequently, a range of Nphenyl groups was introduced. Ortho-substituted N-phenyl compounds B3, B7 and B9 showed weak to moderate inhibitory activity. Para-substituted N-phenyl compounds B4, B6 and B10 showed moderate inhibitory activity. However, compounds B5, B8 and B12, having substituents at the meta-position of N-phenyl, showed dramatically improved activities (IR ¼ 62.98%, 74.67%, and 95.46%, respectively). Among them, compound B12 gave the best results, with an IC 50 against the NO production was 2.14 lM, which may be due to lipid solubility and strong electron-withdrawing properties of the trifluoromethyl moiety 28 . Therefore, compound B12 was selected as the lead compound for further evaluation (Table 1).

Admet prediction
The absorption, distribution, metabolism, excretion, and toxicity (ADMET) of all compounds were predicted using the Discovery Studio 2018 (DS 2018) software 29,30 . The results are depicted in the 2 D graphs of ADMET PSA 2 D and ADMET AlogP98. Compound B12 exhibited suitable solubility at 25 C (solubility level of 2). All compounds were within the 95% and 99% confidence intervals of blood-brain barrier (BBB) penetration and had suitable HIA (absorption level of 0). Moreover, compound B12 was likely to bind to plasma proteins (PPB# prediction was true). The results showed that compound B12 was a non-cytochrome P450 2D6 inhibitor (CYP2D6# prediction was false). The drugability and safety of the compounds were initially evaluated using the ADMET predictions (Supplementary Information). Anti-inflammatory activity of compound B12 was studied in the following steps.

Suppression LPS-induced iNOS and COX-2 activation
Compound B12 was selected as the title compound for the comprehensive consideration of potency, toxicity and drug-like properties. Notably, compound B12 showed the highest potency activity against NO production (IC 50 ¼ 2.14 mM) and lower cytotoxicity    (IC 50 > 100 mM) in RAW264.7 cells. Therefore, inhibition of compound B12 on the pro-inflammatory mediators iNOS and COX-2 was analysed by western blotting. As shown in Figure 3, the expression of iNOS and COX-2 was inhibited by compound B12 in a dose-dependent manner. Thus, compound B12 could prevent LPS-induced expression of inflammatory mediators.

Effect of compound B12 on LPS-induced HaCaT hyperproliferation
The changes in the cell function of keratinocytes are closely related to the occurrence and development of dermatitis. HaCaT cells are immortalised human keratinocytes with normal mutations, which is similar to primary human epidermal keratinocytes (KCs) 31 . The LPS-induced keratinocyte line (HaCaT cell) hyperproliferation model was used to simulate the hyperproliferation of dermatitis keratinocyte epidermis in this study. Therefore, title compounds against HaCaT cells hyperproliferation were tested, with results showing that compound B12 inhibited excessive proliferation of LPS-induced HaCaT cells in a dose dependent manner, while paeonol did not have this effect (Figure 4).

Interaction of compound B12 and TOPK through CETSA
Based on the ligand-induced thermal stabilisation of target proteins, the cellular thermal shift assay (CETSA) assay is likely to validate drug binding, off-target effects or drug resistance in cell lines. To further detect the direct binding of compound B12 with TOPK protein, CETSA was used in this study 32 . As shown in Figure 5, TOPK could be stabilised in presence of compound B12, especially at 57 C, and noticeable results could be observed, which indicating that compound B12 should be bound to TOPK to avoid its degradation.

Compound B12 suppressed LPS-induced TOPK/NF-jB/p38/ JNK activation
In RAW264.7 cell, TOPK is activated by LPS and activates the downstream protein pathways MAPK and NF-j 33,34 . Some of inflammatory   cytokines, such as LPS, IL-6 and TNF-a can activate IjB kinase (IKK), allowing the liberated NF-jB to translocate into the nucleus and bind to target promoters to turn on transcriptions. These results indicated that compound B12 could affect TOPK phosphorylation ( Figure 6(A)). Meanwhile, compound B12 inhibited P38/JNK protein phosphorylation and NF-jB p65 translocated into the nucleus (Figure 6(B,C)). Therefore, we believe that compound B12 may affect the phosphorylation of TOPK by targeting TOPK, and the conduction of the cell signal NF-jB/MAPK in vitro.

Down-regulated SUV-induced TOPK signalling pathway
Similarly, TOPK can be activated by the SUV. Histone H2AX is a substrate of TOPK and can be phosphorylated at Ser-139 (c-H2AX) by TOPK 35 . TOPK is also an upstream kinase of p38. And STAT3 is a downstream kinase of p38. The levels of phosphorylated c-H2AX and STAT3 gradually decreased after pre-treatment with compound B12 from 2.5 lM to 10lM for 6 h in HaCaT cells before 40 KJ/m 2 SUV irradiation (Figure 7(A,B)). Results showed that compound B12 could inhibit SUV-induced DNA damage through the regulation of TOPK.

Anti-inflammatory activity of compound B12 in vivo
The anti-inflammatory activity of compound B12 was further evaluated in vivo. An IMQ-induced psoriasis mouse model was used to mimic the clinical manifestations of psoriasis patients 36,37 . As shown in Figure 8(A), severe skin lesions were occurred on the back skin, including scales, thickness and erythema during consecutive application of IMQ cream on BALB/c mice for 7 days in model the groups. Compound B12 significantly attenuated experimental symptoms and PASI score in a dose-dependent manner during treatment, as well as in the dexamethasone-treated group (Figure 8(B,C)). HE analysis of skin tissue was performed to assess the levels of inflammation and tissue changes. Representative histological images of the tissue sections are shown in Figure  8(D). Mice induced by IMQ showed pathological psoriatic lesions, including loss of granular layer, epidermal hyperplasia, thickening of the acanthosis cell layer, parakeratosis, and inflammatory cell infiltration. IMQ-mice treated with compound B12 (20 mg/kg) showed moderate pathological reduction, while IMQ-mice treated with compound B12 (40 mg/kg) showed high pathological reduction as well as the positive drug dexamethasone. These results indicated that compound B12 successfully reduced the scales, thickness and erythema in psoriasis-like mice, histopathologically alleviated hyperkeratosis, acanthocyte proliferation and inflammatory cell infiltration. As shown in Figure 9, the expression of related proteins in mouse skin tissues was decreased significantly after treatment with compound B12 in a dosedependent manner. It is worth noting that compound B12 inhibits the expression of proliferation-related protein STAT3 and precancerous factor PCNA in skin tissues.
Collectively, these results demonstrate that compound B12 could improve psoriasis-like skin inflammation in vivo.

Metabolic stability assay in liver microsomes
With the above information, the metabolic stability of compound B12 was further evaluated in human liver microsomes in vitro. As shown in Table 2, compound B12 showed acceptable stability against the metabolism of human liver microsomes with a half-life (t 1/2 ) of 111.1 min. This result indicated that compound B12 might be selected as a promising candidate for the treatment of psoriasis-like skin inflammation.

Conclusions
In summary, 38 new paeonol derivatives were synthesised and their anti-inflammatory activities were evaluated. Title compound B12 was found to be the most potent compound without obvious cytotoxicity. The preliminary mechanism indicated that compound B12 could interact with TOPK, to regulating its downstream pathways MAPK and NF-jB and inhibit the expression of DNA damage-related protein H2AX and proliferation-related protein STAT3 ( Figure 10). Anti-inflammatory studies revealed that this compound could effectively relieve histological changes in IMQinduced murine psoriasis-like skin inflammation.

Disclosure statement
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.