Study of Novel Furocoumarin Derivatives on Anti-Vitiligo Activity, Molecular Docking and Mechanism of Action

Vitiligo is a common chronic dermatological abnormality that afflicts tens of millions of people. Furocoumarins isolated from Uygur traditional medicinal material Psoralen corylifolia L. have been proven to be highly effective for the treatment of vitiligo. Although many furocoumarin derivatives with anti-vitiligo activity have been synthesized, their targets with respect to the disease are still ambiguous. Fortunately, the JAKs were identified as potential targets for the disease and its inhibitors have been proved to be effective in the treatment of vitiligo in many clinical trials. Thus, sixty-five benzene sulfonate and benzoate derivatives of furocoumarins (7a–7ad, 8a–8ag) with superior anti-vitiligo activity targeting JAKs were designed and synthesized based on preliminary research. The SAR was characterized after the anti-vitiligo-activity evaluation in B16 cells. Twenty-two derivatives showed more potent effects on melanin synthesis in B16 cells than the positive control (8-MOP). Among them, compounds 7y and 8 not only could increase melanin content, but they also improved the catecholase activity of tyrosinase in a concentration-dependent manner. The docking studies indicated that they were able to interact with amino acid residues in JAK1 and JAK2 via hydrogen bonds. Furthermore, candidate 8 showed a moderate inhibition of CXCL−10, which plays an important role in JAK–STAT signaling. The RT-PCR and Western blotting analyses illustrated that compounds 7y and 8 promoted melanogenesis by activating the p38 MAPK and Akt/GSK-3β/β-catenin pathways, as well as increasing the expressions of the MITF and tyrosinase-family genes. Finally, furocoumarin derivative 8 was recognized as a promising candidate for the fight against the disease and worthy of further research in vivo.


Introduction
Vitiligo is a chronic inflammatory cutaneous disorder that infects over 50 million people in the world [1]. Localized white patches and grey hair are the most common symptoms of the disease [2]. Although not fatal, it spoils the appearance and causes heavy psychological harm to patients. Furthermore, people suffering from vitiligo are vulnerable to other autoimmune diseases, such as rheumatoid arthritis, hyperthyreosis, diabetes and lupus erythematosus [3]. Many hypotheses about its mechanism have been proposed, including autoimmunity, genetics, environment, psychology, deficiency of trace elements, etc. [4]. Over the years, more and more clinical research has supported the immune theory [5][6][7]. However, the destruction of melanocytes and the obstruction of melanin synthesis directly lead to vitiligo [8,9].
3-F-4-CH 3 123 ± 6 7ad 2,5-diCH 3 125 ± 4  Subsequently, the effects of 7y and 8 on cell cytotoxicity, morphology, the catecholase activity of tyrosinase and melanin synthesis in PIG3V and B16 cells were further studied. Firstly, the CCK-8 assay was performed to determine the cytotoxicity of the two compounds in B16 cells. It was revealed, as can be seen in Figures 2C and 3C, that compounds 7y and 8 had weak toxicity at 0-10 µM and 0-50 µM compared with the control. In Figures 2B and 3B, no significant abnormalities in the morphology and number of cells were observed. Moreover, the melanosomes in the cells continued to proliferate with the increase in the concentration. Compounds 7y and 8 obviously promoted melanin production and the catecholase activity of tyrosinase in a dose-dependent manner in B16 and PIG3V melanocytes (Figures 2D-F and 3D-F). However, after treatment with 8 at 50 µM, the melanin content (315.6%) was nearly 2.5-fold compared with 8-MOP (127.3%). Considering the weak effects of many active compounds on catecholase activity explored by our group, it was gratifying that compound 8 (148.4%) was better than 8-MOP for this key enzyme in melanin biosynthesis (127.3%).

Docking Studies
Molecular docking studies were performed to analyze the interaction between active derivatives (7y and 8) and JAK kinases (JAK1 and JAK2). They interacted with JAKs by binding to the ATP sites (Figures 4-7).

Docking Studies
Molecular docking studies were performed to analyze the interaction between active derivatives (7y and 8) and JAK kinases (JAK1 and JAK2). They interacted with JAKs by binding to the ATP sites (Figures 4-7).

Docking Studies
Molecular docking studies were performed to analyze the interaction between active derivatives (7y and 8) and JAK kinases (JAK1 and JAK2). They interacted with JAKs by binding to the ATP sites (Figures 4-7).      For 7y and JAK2, the SO2 group interacted with Lys857 in the hinge region via a hydrogen bond, and another hydrogen bond was created by -C=O of Leu932 ( Figure 4A). The unsubstituted benzene, furan and pyran rings made a great number of hydrophobic interactions with Ser939, Gly935, Leu855 and Leu983. When bound to JAK1, O 16 of OSO2 and H 37 of furan interacted directly with hinge Leu959 and P-loop Gly882 through hydrogen bonds ( Figure 5A). The skeleton of furocoumarin and 5-benzene produced numerous hydrophobic contacts with Val889, Leu1010, Ala906 and Leu881. Hydrogen bonds between residues Gly887, Lys908 and F 26 on benzene were found in JAK2. The halogen-mediated hydrogen bonds between F on benzene, and Val863 and Gly884 were also observed in both JAK1 and -2, although the former was very weak, since the distance was 3.3 Å ( Figures 4A and 5A).
In the docking mode of compound 8 with JAK2, O 16 of the ester group could form hydrogen bonds with Gly856, while the other two hydrogen bonds were seen between Tyr931 and Leu932 in the hinge region, and O 13 and H 29 of the furan ring. The hydrophobic interactions with Leu855, Val863, Ala880 and Gly935 appeared to be able to stabilize the protein-small-molecule complex ( Figure 6A). As shown in Figure 7A, -C=O of the ester group in 8 had hydrogen-bond interactions with Ser963 in the hinge region of JAK1. The furan ring (O 13 and H 29 ) was involved in two hydrogen bonds, with hinge residues Leu959 and Phe958. Additionally, CH2 on pyran pointed to a P-loop, constituting a hydrogen bond with Leu881. The skeleton of compound 8 was sandwiched in the binding pocket and engaged in a large number of hydrophobic interactions with Leu881, Nal889, Ala906 and Leu1010. In addition, compounds 7y and 8 had docking poses more similar to those of control PF-06700841 with JAK1 ( Figures 5B and 7B) than to Ruxolitinib with JAK2 ( Figures 4B and 6B).

Inhibition Activity against JAK-Mediated Chemokine Production
In the present study, we performed an enzyme-linked immunosorbent assay (ELI-SA) to explore the potential involvement of derivatives 7y and 8 in the regulation of IFN-γ-induced chemokine CXCL−10 levels in the supernatants of human-skin HaCaT keratinocytes. Following the treatment of HaCaT with 10 ng/mL recombinant IFN-γ in the presence or absence of 7y and 8 (0.1, 1 or 10 μM), the levels of CXCL−10 were examined, as shown in Figure 8. The treatment of cells with a combination of both IFN-γ and For 7y and JAK2, the SO 2 group interacted with Lys857 in the hinge region via a hydrogen bond, and another hydrogen bond was created by -C=O of Leu932 ( Figure 4A). The unsubstituted benzene, furan and pyran rings made a great number of hydrophobic interactions with Ser939, Gly935, Leu855 and Leu983. When bound to JAK1, O 16 of OSO 2 and H 37 of furan interacted directly with hinge Leu959 and P-loop Gly882 through hydrogen bonds ( Figure 5A). The skeleton of furocoumarin and 5-benzene produced numerous hydrophobic contacts with Val889, Leu1010, Ala906 and Leu881. Hydrogen bonds between residues Gly887, Lys908 and F 26 on benzene were found in JAK2. The halogen-mediated hydrogen bonds between F on benzene, and Val863 and Gly884 were also observed in both JAK1 and -2, although the former was very weak, since the distance was 3.3 Å ( Figures 4A and 5A).
In the docking mode of compound 8 with JAK2, O 16 of the ester group could form hydrogen bonds with Gly856, while the other two hydrogen bonds were seen between Tyr931 and Leu932 in the hinge region, and O 13 and H 29 of the furan ring. The hydrophobic interactions with Leu855, Val863, Ala880 and Gly935 appeared to be able to stabilize the protein-small-molecule complex ( Figure 6A). As shown in Figure 7A, -C=O of the ester group in 8 had hydrogen-bond interactions with Ser963 in the hinge region of JAK1. The furan ring (O 13 and H 29 ) was involved in two hydrogen bonds, with hinge residues Leu959 and Phe958. Additionally, CH 2 on pyran pointed to a P-loop, constituting a hydrogen bond with Leu881. The skeleton of compound 8 was sandwiched in the binding pocket and engaged in a large number of hydrophobic interactions with Leu881, Nal889, Ala906 and Leu1010. In addition, compounds 7y and 8 had docking poses more similar to those of control PF-06700841 with JAK1 ( Figures 5B and 7B) than to Ruxolitinib with JAK2 ( Figures 4B and 6B).

Inhibition Activity against JAK-Mediated Chemokine Production
In the present study, we performed an enzyme-linked immunosorbent assay (ELISA) to explore the potential involvement of derivatives 7y and 8 in the regulation of IFN-γ-induced chemokine CXCL−10 levels in the supernatants of human-skin HaCaT keratinocytes. Following the treatment of HaCaT with 10 ng/mL recombinant IFN-γ in the presence or absence of 7y and 8 (0.1, 1 or 10 µM), the levels of CXCL−10 were examined, as shown in Figure 8. The treatment of cells with a combination of both IFN-γ and Ruxolitinib (positive control) resulted in a notable reduction in CXCL−10 expression (almost 35-fold) compared with IFN-γ alone. Derivative 8 suppressed the CXCL−10 levels (p < 0.001) in IFN-γ-stimulated HaCaT cells at 1 and 10 µM, which was consistent with the molecularmodeling results ( Figure 8A). In addition, when IFN-γ was not added, no obvious effects of the compounds on CXCL−10 secretion were observed, probably because JAK-STAT signaling was not activated by IFN-γ ( Figure 8B). T CXCL−10 in culture medium were determined with the ELISA kit. Each bar r SD (n = 3) (**** p < 0.0001, *** p < 0.001, ** p < 0.01 compared with IFN-γ−stimu means not significant).

Effects of 7y on the Expression of Melanogenesis-Related Genes
With real-time PCR (RT-PCR) and Western blotting analyses, studied for its promotion of the expression levels of melanogenesis demonstrated in Figure 9A-C, the protein and mRNA expressions TRP-1 increased in a concentration-dependent manner in B16 cells, ex ter treatment with 7y. These results indicate that compound 7y ma genesis by greatly upregulating the MITF and TYR genes but not TRP The secretion levels of CXCL−10 in culture medium were determined with the ELISA kit. Each bar represents the mean ± SD (n = 3) (**** p < 0.0001, *** p < 0.001, ** p < 0.01 compared with IFN-γ−stimulated group; n = 3. ns means not significant).

Effects of 7y on the Expression of Melanogenesis-Related Genes
With real-time PCR (RT-PCR) and Western blotting analyses, compound 7y was studied for its promotion of the expression levels of melanogenesis-related genes. As demonstrated in Figure 9A-C, the protein and mRNA expressions of MITF, TYR and TRP-1 increased in a concentration-dependent manner in B16 cells, except for TRP-2, after treatment with 7y. These results indicate that compound 7y may enhance melanogenesis by greatly upregulating the MITF and TYR genes but not TRP-2.

Regulation of Compound 7y of p38 MAPK Signaling Pathway
The MAPK signaling pathway is well known for its regulation of the generation of melanin. Thus, the influence of 7y on the activation of the MAPK signaling pathway was investigated. Judging from Figure 10A,B, compound 7y elevated the phosphorylation of P38 MAPK in a dose-dependent manner, but it was not beneficial for ERK and JNK phosphorylation.

Effects of 7y on the Expression of Melanogenesis-Related Genes
With real-time PCR (RT-PCR) and Western blotting analyses, compound 7y was studied for its promotion of the expression levels of melanogenesis-related genes. As demonstrated in Figure 9A-C, the protein and mRNA expressions of MITF, TYR and TRP-1 increased in a concentration-dependent manner in B16 cells, except for TRP-2, after treatment with 7y. These results indicate that compound 7y may enhance melanogenesis by greatly upregulating the MITF and TYR genes but not TRP-2.  The MAPK signaling pathway is well known for its regulation of the generation of melanin. Thus, the influence of 7y on the activation of the MAPK signaling pathway was investigated. Judging from Figure 10A,B, compound 7y elevated the phosphorylation of P38 MAPK in a dose-dependent manner, but it was not beneficial for ERK and JNK phosphorylation. To further identify the involvement of the P38 MAPK signaling pathway in compound 7y-induced melanogenesis, p38 inhibitor SB203580 was utilized to assess the catecholase activity of tyrosinase and melanin contents. It was found that the melanogenic effects (melanin content and catecholase activity) induced by 7y obviously decreased

Effects of p38 Inhibitors on Compound 7y-Induced Melanogenesis
To further identify the involvement of the P38 MAPK signaling pathway in compound 7y-induced melanogenesis, p38 inhibitor SB203580 was utilized to assess the catecholase activity of tyrosinase and melanin contents. It was found that the melanogenic effects (melanin content and catecholase activity) induced by 7y obviously decreased with SB203580 (a p38MAPK inhibitor) but were not affected by SP600125 (a JNK inhibitor) nor PD98059 (an ERK inhibitor) (Figure 11).

Effects of 8 on the Expression of Melanogenesis-Related Genes
Firstly, the effects of compound 8 on the expression of melanogenesis-related proteins were examined using Western blotting. As shown in Figure 12A,B, the protein expressions of MITF, TYR, TRP-1 and TRP-2 increased significantly in a concentration-dependent manner in B16 cells treated with compound 8, especially at 50 μM, which indicated that it could enhance melanogenesis by upregulating the MITF and TYR proteins.
After that, we studied the effects of 8 on the relative quantification of the mRNA transcriptions of the tyr, trp 1 and trp 2 genes with the aid of RT-PCR. It can be seen from Figure 12C that treatment with 8 remarkably enhanced the tyrosinase-family-protein mRNA levels, which suggests that it may also improve the transcription levels of these three melanogenesis-related genes. Firstly, the effects of compound 8 on the expression of melanogenesis-related proteins were examined using Western blotting. As shown in Figure 12A,B, the protein expressions of MITF, TYR, TRP-1 and TRP-2 increased significantly in a concentration-dependent manner in B16 cells treated with compound 8, especially at 50 µM, which indicated that it could enhance melanogenesis by upregulating the MITF and TYR proteins.
After that, we studied the effects of 8 on the relative quantification of the mRNA transcriptions of the tyr, trp 1 and trp 2 genes with the aid of RT-PCR. It can be seen from Figure 12C that treatment with 8 remarkably enhanced the tyrosinase-family-protein mRNA levels, which suggests that it may also improve the transcription levels of these three melanogenesis-related genes.

Regulation of Compound 8 of the Akt/GSK-3β/β-Catenin Signaling Pathway
In order to explore the mechanisms of 8 in promoting melanin synthesis in B16 cells, the effects of 8 on related signaling pathways were examined in vitro. The Western blot analysis was preformed to detect the expression levels of Akt, GSK-3β, β-catenin, JNK, p38, ERK and CREB. According to Figure 13A and B, the expressions of phosphorylated Akt, GSK-3β and total β-catenin increased in a dose-dependent manner after treatment with 8 at concentrations from 1 to 50 μM. However, the compound had little influence on the phosphorylation of JNK, p38, ERK or CREB.

Regulation of Compound 8 of the Akt/GSK-3β/β-Catenin Signaling Pathway
In order to explore the mechanisms of 8 in promoting melanin synthesis in B16 cells, the effects of 8 on related signaling pathways were examined in vitro. The Western blot analysis was preformed to detect the expression levels of Akt, GSK-3β, β-catenin, JNK, p38, ERK and CREB. According to Figure 13A,B, the expressions of phosphorylated Akt, GSK-3β and total β-catenin increased in a dose-dependent manner after treatment with 8 at concentrations from 1 to 50 µM. However, the compound had little influence on the phosphorylation of JNK, p38, ERK or CREB. To understand the involvement of the Akt/GSK-3β/β-catenin pathway in 8-induced melanogenesis, an Akt inhibitor was employed for the evaluation of the catecholase activity of tyrosinase and melanin content. As illustrated in Figure 14, it was obvious that

Effects of GSK-3β Inhibitor on Compound 8-Induced Melanogenesis
To understand the involvement of the Akt/GSK-3β/β-catenin pathway in 8-induced melanogenesis, an Akt inhibitor was employed for the evaluation of the catecholase activity of tyrosinase and melanin content. As illustrated in Figure 14, it was obvious that these two important biological indexes both markedly decreased when co-treated with the Akt inhibitor and compound 8 compared with the latter alone. Moreover, this results firmly confirms the participation of Akt/GSK-3β/β-catenin signaling in compound-8-induced melanogenesis. Figure 13. The effects of derivative 8 on GSK-3β/β-catenin signaling pathway analyzed using Western blot assays. (A) Effect of 8 on GSK-3β/β-catenin-signaling protein expression levels. (B) Protein-band expression levels (** p < 0.01,*** p < 0.001, **** p < 0.0001 compared with control group; n = 3).

Effects of GSK-3β Inhibitor on Compound 8-Induced Melanogenesis
To understand the involvement of the Akt/GSK-3β/β-catenin pathway in 8-induced melanogenesis, an Akt inhibitor was employed for the evaluation of the catecholase activity of tyrosinase and melanin content. As illustrated in Figure 14, it was obvious that these two important biological indexes both markedly decreased when co-treated with the Akt inhibitor and compound 8 compared with the latter alone. Moreover, this results firmly confirms the participation of Akt/GSK-3β/β-catenin signaling in compound-8-induced melanogenesis.

Discussion
Pathways such as Akt/GSK-3β/β-catenin, p38 MAPK, ERK and JNK MAPK signaling and cAMP-dependent pathways have been reported to be involved in melanogenesis via the regulation of MITF and TYR expressions. On the other hand, the MITF is regarded as the most important transcription factor for melanocyte development and the produc-

Discussion
Pathways such as Akt/GSK-3β/β-catenin, p38 MAPK, ERK and JNK MAPK signaling and cAMP-dependent pathways have been reported to be involved in melanogenesis via the regulation of MITF and TYR expressions. On the other hand, the MITF is regarded as the most important transcription factor for melanocyte development and the production of melanin, since it can regulate melanogenesis-related gene transcription, including tyrosinase, TRP-1 and TRP-2. It this research study, compound 7y could markedly ameliorate the melanogenic effect and increase the expressions of melanogenic enzymes (TYR and TRP-1) and MITF in a concentration-dependent manner in cells. Our results demonstrate the 7y-induced promotion of the phosphorylation of p38 with the exceptions of ERK and JNK, which may be blocked by p38 inhibitors. Taken together, compound 7y induced melanogenesis by activating the p38 MAPK pathway and increasing the expressions of the MITF and tyrosinase-family genes ( Figure 15).  The Western blot and RT-PCR analyses were applied to clarify that compound 8 enhanced melanogenesis by elevating the expression level of the MITF. Moreover, it was found that the expressions of phosphorylated Akt, GSK-3β and β-catenin considerably increased in cells after treatment with compound 8. Therefore, we assume that the effects The Western blot and RT-PCR analyses were applied to clarify that compound 8 enhanced melanogenesis by elevating the expression level of the MITF. Moreover, it was found that the expressions of phosphorylated Akt, GSK-3β and β-catenin considerably increased in cells after treatment with compound 8. Therefore, we assume that the effects of 8 on melanin synthesis may result from the activation of p-Akt, which promotes the phosphorylation of GSK-3β, thus increasing the accumulation of β-catenin in the cytoplasm; the accumulated β-catenin is translocated to the nucleus, where it directly binds with the MITF to stimulate its transcription and then stimulates the tyrosinase-family genes ( Figure 16). In addition, an Akt inhibitor could dramatically reduce the content of melanin and the catecholase activity of tyrosinase, meanwhile few effects of 8 on p38 MAPK, ERK or JNK MAPK signaling or cAMP-dependent pathways were observed, which implies that compound 8 improved melanogenesis by activating the Akt/GSK-3β/β-catenin pathway and increasing the expression of the MITF, as well as of tyrosinase-family genes. The Janus kinase (JAK) family, composed of JAK1, JAK2, JAK3 and TYK2, plays a crucial rule in the activation, differentiation and homeostasis of immune-associated cells [31]. Numerous JAK inhibitors have been developed due to its attractive effect on immunological regulation. In a recent study [32], IFN-γ-JAK/STAT signaling was proven to be closely related with the progression and maintenance of vitiligo. Melanocyte-reactive CD8 + T cells produce IFN-γ upon the encounter of melanocyte antigen; IFN-γ activates JAKs in keratinocytes (KCs) by binding to its receptors on KCs. After dimerization, JAKs are activated to phosphorylate their downstream STAT, and phosphorylated STAT moves into the nucleus and produces more CXCL−9 and CXCL−10, which leads to the additional recruitment of CD8 + T cells to melanocytes through the CXCR3 chemokine receptor. Treatment that inhibits this pathway, such as JAK inhibitors and monoclonal antibody agents for IFN-γ or CXCL−10, could effectively reverse the disease. Moreover, FDA-approved JAK inhibitor ruxolitinib has already been tested in a phase III trial for vitiligo (clinical trial NCT04530344). Thus, our group focused on the development of a small-molecule inhibitor of JAKs and discovered several hits from quinazoline and quinoline alkaloids with the help of virtual screening. Multi-target therapeutics that regulated multiple nodes of the disease network simultaneously showed a synergistic effect, which was unlikely to induce resistance and provided optimal clinical use. Fortunately, furocoumarin derivative 8 could inhibit CXCL−10 secretion in a concentration-dependent manner, which means that it may not only activate Akt/GSK-3β/β-catenin signaling but also block the JAK-STAT pathway, in spite of structural differences from the reported JAK inhibitors. Further structure optimization and target validation based on compound 8 are needed to explore more candidates for vitiligo. The Janus kinase (JAK) family, composed of JAK1, JAK2, JAK3 and TYK2, plays a crucial rule in the activation, differentiation and homeostasis of immune-associated cells [31]. Numerous JAK inhibitors have been developed due to its attractive effect on immunological regulation. In a recent study [32], IFN-γ-JAK/STAT signaling was proven to be closely related with the progression and maintenance of vitiligo. Melanocytereactive CD8 + T cells produce IFN-γ upon the encounter of melanocyte antigen; IFN-γ activates JAKs in keratinocytes (KCs) by binding to its receptors on KCs. After dimerization, JAKs are activated to phosphorylate their downstream STAT, and phosphorylated STAT moves into the nucleus and produces more CXCL−9 and CXCL−10, which leads to the additional recruitment of CD8 + T cells to melanocytes through the CXCR3 chemokine receptor. Treatment that inhibits this pathway, such as JAK inhibitors and monoclonal antibody agents for IFN-γ or CXCL−10, could effectively reverse the disease. Moreover, FDA-approved JAK inhibitor ruxolitinib has already been tested in a phase III trial for vitiligo (clinical trial NCT04530344). Thus, our group focused on the development of a small-molecule inhibitor of JAKs and discovered several hits from quinazoline and quinoline alkaloids with the help of virtual screening. Multi-target therapeutics that regulated multiple nodes of the disease network simultaneously showed a synergistic effect, which was unlikely to induce resistance and provided optimal clinical use. Fortunately, furocoumarin derivative 8 could inhibit CXCL−10 secretion in a concentration-dependent manner, which means that it may not only activate Akt/GSK-3β/β-catenin signaling but also block the JAK-STAT pathway, in spite of structural differences from the reported JAK inhibitors. Further structure optimization and target validation based on compound 8 are needed to explore more candidates for vitiligo.

General Methods
All the chemicals were commercially available and were used directly without purification. A thin-layer chromatography (TLC) analysis was performed on glass plates coated with silica gel (Qingdao Haiyang Chemical Co.; G60F-254) and visualized under UV light (254 nm). The products were purified using silica gel (Qingdao Haiyang Chemical Co.; 200-300 meshes) using column chromatography. The NMR spectra were recorded on Varian (400 and 600 MHz) spectrometers, referenced in CDCl 3 to tetramethylsilane (TMS). The melting points were measured on Buchi B-540. High-resolution mass spectra were acquired on an ABSciex QSTAR Elite quadrupole-time-of-flight mass spectrometer.

Synthetic Procedures Synthesis of 7-Hydroxy-4-methyl-2H-chromen-2-one (2)
To an ice-cold solution of resorcinol (2.0 g; 18.2 mol) in dioxane, conc. H 2 SO 4 (0.5 mL) was added dropwise at 20 • C. After that, ethyl acetoacetate (2.8 g; 21.8 mmol) was added, and the mixture was heated to 60 • C for 4 h. Then, the mixture was poured into cold water, and the precipitate was filtered and dried under reduced pressure. The resulting mixture was recrystallized from methanol to obtain 2 as white needle crystals. Yield of 92%, m.p. 202-204 • C.

Synthesis of 7a-7ad
To an ice-cold mixture of alcohol 6 (2.92 g; 10 mmol) and triethylamine (4.04 g; 40 mmol) in ethylacetate, diverse benzenesulfonylchloride (20 mmol) was added dropwise at 0 • C. The reaction mixture was allowed to warm to room temperature, stirred until the reaction was complete and then diluted with water (20 mL). The organic layer was dried over MgSO 4 and filtered. The residue was concentrated and purified using flash chromatography on silica gel to obtain benzene sulfonate 7a-7ad. The NMR and spectrum can be found in Supplementary Materials. Synthesis of (7-Oxo-3-phenyl-7H-furo[3,2-g]chromen-5-yl)methyl Acetate (8) To an ice-cold mixture of alcohol 6 (2.92 g; 10 mmol) and triethylamine (4.04 g; 40 mmol) in ethylacetate, acetyl chloride (1.58 g; 20 mmol) was added dropwise at 0 • C. The reaction mixture was allowed to warm to room temperature, stirred for 2 h and diluted with water (20 mL

Synthesis of 8a-8ag
To an ice-cold mixture of alcohol 6 (2.92 g; 10 mmol) and triethylamine (4.04 g; 40 mmol) in ethylacetate, diverse benzoylchloride (20 mmol) was added dropwise at 0 • C. The reaction mixture was allowed to warm to room temperature, stirred until the reaction was complete and then diluted with water (20 mL). The organic layer was dried over MgSO 4 and filtered. The residue was concentrated and purified using flash chromatography on silica gel to obtain benzene sulfonate 8a-8ag. The NMR and spectrum can be found in Supplementary Materials.

Cell-Viability Measurement
Cell viability was assayed by adding CCK-8 solution. Generally speaking, B16 cells were seeded in 96-well plates at a density of 8 × 10 3 cells per well and were allowed to adhere for 24 h. The medium was replaced with medium containing samples diluted to the appropriate concentrations. The control cells were treated with DMSO at the final concentration of 0.1%. After 24 h, the culture medium of the cells was discarded; a volume of 10 µL of CCK-8 solution was added into each well, and cells were incubated at 37 • C for another 2 h. The absorbance was measured at 450 nm using Spectra Max M5 (Molecular Devices, San Diego, CA, USA). All assays were performed in triplicate. Absorbance of cells without treatment was regarded as 100% of cell survival. Cell viability was calculated using the following formula: cell viability (%) = (A sample /A control ) × 100%.

Melanin Measurement
B16 cells were seeded at a density of 2 × 10 5 cells/well in a 6-well plate. After overnight incubation, the test samples were added to individual wells, and cells were incubated for 48 h and washed twice with ice-cold PBS. After cells lysed, the harvested cells were centrifuged, and the pellet was dissolved by adding 1 M NaOH, followed by incubation at 80 • C for 1 h. Each lysate (150 µL) was put in a 96-well microplate and measured spectro-photometrically at 405 nm with a multi-plate reader. The protein concentration of each sample was determined with BCA Protein Assay Kit (Biomed, Beijing, China). Intracellular melanin amounts were expressed as abs/µg protein and were shown as percentage values. The percentage values of the sample-treated cells were calculated with respect to the untreated cells.

Assay of Catecholase Activity of Tyrosinase
The assay of catecholase activity was carried out as previously described, with a slight modification. B16 cells were seeded in a 6-well plate at a density of 2 × 10 5 cells per well and were allowed to attach for 24 h. Test samples were then added to individual wells. After 24 h of incubation, cells were washed twice with ice-cold PBS and lysed with 1% Triton X-100 solution containing 1% sodium deoxycholate for 30 min at −80 • C; then, each lysate was centrifuged at 12,000× g RPM for 15 min to obtain the supernatant. After protein quantification and adjustment, 90 µL of the supernatant was incubated in duplicate with 10 µL of freshly prepared substrate solution (10 mM L-DOPA) in a well of a 96-well plate. Then, the cells were incubated at 37 • C in the dark for 60 min. The absorbance was measured at 490 nm, and the values of the sample-treated cells were presented as percentages against the untreated cells.

Measurement of CXCL−10 Release by HaCaT Keratinocytes
HaCaT cells were plated at a density of 4 × 10 5 cells per well in a 6-well plate and treated with the indicated concentrations of the compounds with Human IFN-γ (PeproTech, Cranbury, NJ, USA) at 10 ng/mL for 24 h and 3 µM Ruxolitinib as the positive control. To remove cell debris, the cell-culture supernatant was centrifuged at 1000× g RPM for 10 min. The release levels of CXCL−10 were detected with a Human CXCL−10/IP-10 enzyme-linked immunosorbent assay (ELISA) kit, which was purchased from Absin (Shanghai, China), based on the manufacturer's instructions. The absorbance was measured at 450 nm using a Spectra Max M5 microplate reader (Molecular Devices company, San Diego, CA, USA). 4.2.6. Western Blot Analysis B16 cells were treated with different concentrations of the tested compounds in a 6-well plate for 48 h. Cells were then lysed in cold RIPA (radio immunoprecipitation assay) lysis buffer (pH 7.4) containing protease and protease inhibitor cocktail (1 M 4-nitrophenyl phosphate disodium salt hexahydrate (PNPP), 1 M sodium fluoride (NaF), 10 mM phenylmethanesulfonylfluoride (PMSF), 100 mM benzamidine, 100 mM DL-Dithiothreitol (DTT), 200 mM sodium orthovanadate (OV)) for 30 min on ice. The lysates were centrifuged at 12,000× g rpm for 20 min at 4 • C before the supernatant was collected. The samples' protein concentrations were measured with BCA Protein Assay Kit (Biomed, Beijing, China) and were separated by 10% SDS polyacrylamide gels and transferred onto polyvinylidene fluoride (PVDF) membranes (Merck Millipore Ltd., Billerica, MA, USA). Membranes were incubated with the primary antibodies at 4 • C overnight and then with horseradishperoxidase-conjugated secondary antibodies for 1 h at room temperature. The targeted proteins were detected using ECL Western blotting detection reagents (GE Healthcare, Beijing, China) and were visualized using a ChemiDoc MP Imaging system (Bio-Rad Laboratories, Inc., Berkeley, CA, USA). All Western blot assays were performed in triplicate.

Primer Sequence in Quantitative Real-Time PCR
Total cellular RNA was prepared from B16 cells treated with the tested compounds and isolated with TRIzol reagent in accordance with the manufacturer's instructions. Quantitative PCR was performed to determine the expressions of target genes. The primers were as follows: forward 5 -GTCGTCACCCTGAAAATCCTAACT-3 and reverse 5 -CATCGCATAAAACCTGATGGC-3 for Tyr (111 bp); forward 5 -ACCCATTTGTCTCCCAATGA-3 and reverse 5 -GTCCAATAGGTGCGTTTTCC-3 for TRP-1 (130 bp); forward 5 -TACCATCTGTTGTGGCTGGA-3 and reverse 5 -TGGGTCATCTTGCTG-3 for TRP-2 (147 bp); forward 5 -AGTACAGGAGCTGGAGATG-3 and reverse 5 -GTGAGATCCAGAGTTGTCGT-3 for MITF (181 bp). β-Actin was used as an internal control in all cases, and its primer sequence was as follows: forward 5 -TCAAGA TCATTGCTCCTCCTG-3 and reverse 5 -CTGCTTGCT GATCCA-CATCTG-3 (59 bp). All determinations were performed three times. The reaction parameters were 95 • C for 10 min, followed by 40 cycles of 15 s at 95 • C for melting and 1 min at 60 • C for annealing. Real-time PCR was performed using an Applied Biosystems 7300 PCR machine (Applied Bioscience, Foster City, CA, USA). The results were normalized to the controls.

Statistical Analysis
All results were expressed as mean ± SD, and statistical analyses were performed with one-way ANOVAs, followed by Tukey's multiple comparisons tests. Statistical analyses were performed using GraphPad Prism 9 (La Jolla, CA, USA). p-values < 0.05 were considered to be statistically significant.

Molecular Docking
Human JAK1 (PDB: 6DBN) [33] and JAK2 (PDB: 6VGL) [34] were downloaded from the Protein Data Bank (http://www.rcsb.org accessed on 9 November 2021). They were prepared by removing water molecules and adding hydrogen atoms using protein preparation wizard in Discovery Studio 2016. After energy minimization, the docking studies of compounds 7y and 8 with JAK1 and JAK2 were performed via CDOCKER. Before docking, a sphere (with a 10 Å radius) around the template molecule was established as the binding site. For the simulated annealing, all the parameters were set as default. After the docking procedure, ten top-ranked ligand-receptor conformations were obtained, and the binding patterns of the docked molecules were visualized and analyzed according to the receptor-ligand interactions in Discovery Studio.

Conclusions
Furocoumarins have been used to treat vitiligo for a long time. Moreover, in this work, sixty-five novel ester derivatives of furocoumarin were designed, synthesized and evaluated for their anti-vitiligo activity. Further pharmacological studies let us evaluate two candidates, 7y and 8 that were more potent than the positive control (8-MOP) on both melanin synthesis and the catecholase activity of tyrosinase. The possible binding mode of the compounds with JAKs were identified with molecular docking, together with the inhibitory effect on IFN-γ-induced CXCL−10 production observed in the HaCaT-cell experiment.
Western blotting and RT-PCR proved that 7y could promote the phosphorylation of p38, but not ERK and JNK, which demonstrated that the activation of p38 MAPK may be its underlying mechanism. On the other hand, compound 8 could activate p-Akt, which increased the phosphorylation level of GSK-3β, thus inhibiting the degradation of β-catenin. The accumulated β-catenin was translocated to the nucleus and combined with the MITF to strengthen its transcription; then, it stimulated the tyrosinase-family genes. Further studies of 8 in animal models of vitiligo are required to evaluate its safety and efficacy.