Synthesis of Novel Spiro-Tetrahydroquinoline Derivatives and Evaluation of Their Pharmacological Effects on Wound Healing

A highly diastereoselective method for the synthesis of novel spiro-tetrahydroquinoline derivatives is reported here, using a one-pot reaction method. All compounds were characterized by 1H-nuclear magnetic resonance (NMR) and mass spectroscopy, and their stereo configurations were confirmed by X-ray analysis. These activities of these derivatives were then tested in human keratocyte cells. The responses of cells to treatment with selected compounds were studied using scratch analysis, and the compounds were tested in a mouse excision wound model. Three of the derivatives demonstrated significant wound-healing activities.


Introduction
Skin is the largest organ and the first defense to protect the human body [1,2]. It can be damaged by trauma, burns, skin diseases and so on. Severe skin trauma can impose physical, psychological, and economic burdens on patients. The wound healing process involves the coordination of many distinct but overlapping physiological spaces, comprising hemostasis, inflammation, epithelial cell proliferation, and tissue remodeling [3][4][5]. In elderly or diabetic patients, the risk of wound infection increased due to vascular aging and the lower tissue repair capability, which may eventually lead to chronic wounds [6]. Therefore, wound healing is one of the hot topics in skin surgery. At present, few drugs have been found with substantial abilities of promoting wound healing [7]. Actually, the quality of wound regeneration mainly depends on the efficiency of wound care [8]. In this study, we synthesized a series of novel spiro-tetrahydroquinoline derivatives and compared their effects in wound healing in human epidermal cells and animal models.
Quinoline derivatives have attracted both synthetic and biological chemists because of their diverse chemical and pharmacological properties, such as anticancer, antimycobacterial, anticonvulsant, anti-inflammatory and anti-cardiovascular diseases [9]. Some quinoline derivatives were found to have wound healing activity [10][11][12][13]. Tetrahydroquinoline derivatives, which are the reduced form of quinolines, were used as antibacterials, antitumor and anti-HIV agents [14]. However, there are very few reports about tetrahydroquinoline on wound healing even though their structures are so similar to each other [15]. It

Results
The ortho-N-sulfonated aminophenyl α,β-unsaturated ketone 1a and 2-benzylidene-1,3indandione 2a were used as model substrates for the synthesis of spiro-tetrahydroqunioline derivative 3a via an aza-Michael/Michael reaction (Table 1). Initially, the reaction of 1a with 2a in the presence of DABCO furnished the desired product 3a in 71% yield at 30 • C in p-xylene (Table 1, entry 1). Examination of different tertiary amines such as DMAP, Et 3 N and DIPEA as catalysts did not improve the yield of the product 3a (Table 1, entries [2][3][4]. It is worthy to note that all the catalysts furnished the product 3a in excellent diastereomeric ratio (>20:1). Furthermore, various solvents like toluene, CH 2 Cl 2 , THF, EtOAc and Et 2 O were screened to optimize the reaction conditions (Table 1, entries 5-10). The results proved that aprotic solvents are the best for the reaction conditions such as toluene and CH 2 Cl 2 ( Table 1, entries 5 and 6). Considering the solubility of the starting materials 1 and 2, we chose to test the CH 2 Cl 2 as optimized solvent for further screening of reaction. When the reaction was carried out at 0 • C, the yield of the desired product 3a was improved to 97%, but the reaction required longer times (38 h) compared to 30 • C (Table 1, entries 6 and 11). The reaction conditions were further evaluated by the catalyst loading and concentration of the solvent, which indicated that 5 mol% of DABCO was sufficient for the completion of the reaction in CH 2 Cl 2 (1 mL) ( Table 1, entries 12 and 13). The reaction conditions in Table 1, entry 13 were selected as the optimal conditions for further studies. The products were purified by the recrystallization, and identified by 1 H-NMR, 13 C-NMR and HRMS. The stereo configuration of the compound 3a was determined by single crystal X-ray diffraction analysis as shown in Figure 1 [31].  , and catalyst (20 mol%) in 0.5 mL solvent. b Yield and diastereomeric ratio were determined by 400 MHz NMR with Ph3CH as an internal standard. c Diastereomeric ratio was >20:1. d 1.0 mL solvent was used. e 10 mol% of catalyst was used. f 5 mol% of catalyst was used. With the optimal condition in hand, the substrate scope was further investigated. In general, all the substrates 1 and 2 with different electron-withdrawing and electron-donating R 1 and R 2 substituents were provided the desired spiro-tetrahydroquinolines in good to excellent yields (Table 2). At first, different indandione derivatives of R 2 substituents were tested with 1a. We noticed significant steric and electronic effects of the R 2 substituent in the reaction outcome. Substrates with meta-and para-bromo groups as R 2 substituents reacted well with 1a to afford the spiro-tetrahydroquinoline products 3c and 3d in up to 83% yields within 3 h, whereas the substrate with R 2 as ortho-bromo substituent was less reactive and the desired product 3b was obtained in 71% yield in longer reaction times (12 h). It clearly indicates that the rate of the reaction is reduced by the sterically hindered substituents. We also found that the  , and catalyst (20 mol%) in 0.5 mL solvent. b Yield and diastereomeric ratio were determined by 400 MHz NMR with Ph3CH as an internal standard. c Diastereomeric ratio was >20:1. d 1.0 mL solvent was used. e 10 mol% of catalyst was used. f 5 mol% of catalyst was used. With the optimal condition in hand, the substrate scope was further investigated. In general, all the substrates 1 and 2 with different electron-withdrawing and electron-donating R 1 and R 2 substituents were provided the desired spiro-tetrahydroquinolines in good to excellent yields ( Table 2). At first, different indandione derivatives of R 2 substituents were tested with 1a. We noticed significant steric and electronic effects of the R 2 substituent in the reaction outcome. Substrates with meta-and para-bromo groups as R 2 substituents reacted well with 1a to afford the spiro-tetrahydroquinoline products 3c and 3d in up to 83% yields within 3 h, whereas the substrate with R 2 as ortho-bromo substituent was less reactive and the desired product 3b was obtained in 71% yield in longer reaction times (12 h). It clearly indicates that the rate of the reaction is reduced by the sterically hindered substituents. We also found that the With the optimal condition in hand, the substrate scope was further investigated. In general, all the substrates 1 and 2 with different electron-withdrawing and electrondonating R 1 and R 2 substituents were provided the desired spiro-tetrahydroquinolines in good to excellent yields ( Table 2). At first, different indandione derivatives of R 2 substituents were tested with 1a. We noticed significant steric and electronic effects of the R 2 substituent in the reaction outcome. Substrates with metaand para-bromo groups as R 2 substituents reacted well with 1a to afford the spiro-tetrahydroquinoline products 3c and 3d in up to 83% yields within 3 h, whereas the substrate with R 2 as ortho-bromo substituent was less reactive and the desired product 3b was obtained in 71% yield in longer reaction times (12 h). It clearly indicates that the rate of the reaction is reduced by the sterically hindered substituents. We also found that the reaction rate also depends on the electronic properties of the R 2 substituents. For example, the substrates bearing electronwithdrawing groups 2d-2g subjected with 1a, furnished the corresponding products 3d-3g in high yields (81-85%) within 3 h (Table 1, entries 4-7), but the substrate bearing an electron-donating group such as 4-OMePh as R 2 substituent 2h resulted in the desired product 3h in only moderate yield (68%), when prolonging the reaction time up to 29 h ( Table 1, entry 8). In contrast, when substrates with heteroaryl (furyl or thienyl) groups 2i and 2j were employed, the reaction could not proceed to provide the products even after 24 h (Table 1, entries 9 and 10). reaction rate also depends on the electronic properties of the R 2 substituents. For example, the substrates bearing electron-withdrawing groups 2d-2g subjected with 1a, furnished the corresponding products 3d-3g in high yields (81-85%) within 3 h (Table 1, entries 4-7), but the substrate bearing an electron-donating group such as 4-OMePh as R 2 substituent 2h resulted in the desired product 3h in only moderate yield (68%), when prolonging the reaction time up to 29 h (Table 1, entry 8). In contrast, when substrates with heteroaryl (furyl or thienyl) groups 2i and 2j were employed, the reaction could not proceed to provide the products even after 24 h (Table 1, entries 9 and 10). Ph OEt (1f) Ph (2a) 24 N.R. 16 2-Naph (1g) Ph (2a) 6 85 (3m) a Unless noted, all reactions were carried out with 1 (0.2 mmol), 2 (0.2 mmol) and CH2Cl2 (1 mL) was used. b Yield of the product 3 was recrystallized from the ethanol and hexane, dr > 20:1 (3a-3m).
Furthermore, different R 1 substituents of 1 were also tested in the reaction conditions to prepare the desired spiro-tetrahydroquinolines 3. Delightfully, the substrate having aldehyde (R 1 = H) group reacted well with 2a to afford the corresponding product 3i in 95% yield within 3 h (Table 1, entry 11). As similar as R 2 , substrates bearing electron-withdrawing R 1 groups (1c and 1d) more efficiently furnished the desired products 3j and 3k compared to the electron-donating group such as 1e (Table 1, entries [12][13][14]. When substrate with ester group 1f was employed as the reactant, the corresponding product could not be found in the reaction (Table 1, entry 15). It could be understood that the second Michael addition was not efficient when the electron-rich ester is present rather than an aldehyde or ketone. In addition, 2-naphthyl group of 1g also furnished the product 3m in 85% yield in 6 h ( Table 1, entry 16).
Furthermore, different R 1 substituents of 1 were also tested in the reaction conditions to prepare the desired spiro-tetrahydroquinolines 3. Delightfully, the substrate having aldehyde (R 1 = H) group reacted well with 2a to afford the corresponding product 3i in 95% yield within 3 h (Table 1, entry 11). As similar as R 2 , substrates bearing electronwithdrawing R 1 groups (1c and 1d) more efficiently furnished the desired products 3j and 3k compared to the electron-donating group such as 1e (Table 1, entries [12][13][14]. When substrate with ester group 1f was employed as the reactant, the corresponding product could not be found in the reaction (Table 1, entry 15). It could be understood that the second Michael addition was not efficient when the electron-rich ester is present rather than an aldehyde or ketone. In addition, 2-naphthyl group of 1g also furnished the product 3m in 85% yield in 6 h ( Table 1, entry 16).
The results of the MTT test revealed the cell viability of human keratinocyte cells (HaCaT) treated with the 13 derived compounds as shown in Table 3. The survival rates of cells treated with 3b were approximately 80% at all treatment concentrations (Supplementary Figure S1A). Cells treated with 3c displayed a survival rate of 80% at concentrations of 6.25 and 12.5 µM, but the growth was inhibited at concentrations equal to or greater than 25 µM (Supplementary Figure S1B). Cells treated with 3i displayed low survival rates at concentrations greater than 25 µM (Supplementary Figure S1C). Cells treated with 3l displayed a survival rate of approximately 80% at all concentrations (Supplementary Figure S1D). The survival rates of cells treated with 3m at concentrations of 6.25 and 12.5 µM were 80% and 60%, respectively. Compared with the other tested compounds, the five compounds described above had little effect on HaCaT cell growth (Supplementary Figure S1), whereas the remaining compounds displayed the strong inhibition of HaCaT cell growth. The results of the remaining samples can be found in the supplementary materials. Therefore, we selected the five compounds with limited effects on cell viability for use in subsequent experiments. HaCaT are human keratinocyte cell line.
Based on the results of the cell viability experiments, the activities of the five selected compounds were evaluated in the scratch assay. The results shown in Figure 2 indicated that the wounds healed gradually healed within 15-18 h when the tested drugs were added to the culture medium. 3b had a significant effect on wound healing at the concentration of 25 µM, but no significant effect was observed when the concentration was increased to 50 µM ( Figure 3A). 3c had an effect on wound healing at concentrations of 12.5 and 25 µM ( Figure 3B). 3i had a significant effect on wound healing at concentrations of 12.5 and 25 µM. Compared with the control groups, no significant effects on wound healing were observed for low concentration (6.25 µM) or high concentration (50 µM) treatment with either 3c or 3i ( Figure 3C). 3l and 3m displayed no significant effects on the promotion of wound healing at any of the tested concentrations and were found to have inhibitory effects ( Figure 3D,E). The optimal concentration data based on these results is shown in Figure 3F, and the compounds 3b, 3c, and 3i were selected for further study in an animal model of wound closure.
Based on the results obtained from the scratch assay, wound-healing tests in mice were performed to examine the effects of wound treatment using 3b, 3c, and 3i ( Figure 4A). The control group was observed to shed the scab on the 9th day, and the wound repair was completed on the 13th day. In mice treated with 3b at a concentration of 50 µM, the scab fell off on the 12th day, and the wound repair was completed on the 13th day ( Figure 4B). Treatment with 6.25 µM of 3c resulted in the shedding of the pupae on the 9th day, but the wound was not completely repaired by the 13th day ( Figure 4C). When a 25 µM concentration of 3i was applied, the pupae were shed on the 11th day, and the wound was repaired. The wound healing was completed by the 11th day, and the pupae were shed on the 7th day when the concentration of 3i was increased to 50 µM ( Figure 4D). We compared the optimal concentrations of 3b, 3c, and 3i treatment based on the results of the mouse wound healing model. 3i was demonstrated to have the best effect on wound healing at a concentration of 50 µM ( Figure 4E).  Based on the results obtained from the scratch assay, wound-healing tests in mice were performed to examine the effects of wound treatment using 3b, 3c, and 3i ( Figure  4A). The control group was observed to shed the scab on the 9th day, and the wound repair was completed on the 13th day. In mice treated with 3b at a concentration of 50 µM, the scab fell off on the 12th day, and the wound repair was completed on the 13th day ( Figure 4B). Treatment with 6.25 µM of 3c resulted in the shedding of the pupae on the 9th day, but the wound was not completely repaired by the 13th day ( Figure 4C).   concentrations of 3b, 3c, and 3i on  wound healing. 3b, 3c, and 3i are arranged in accordance with the specific concentration of each compound that was used in a gel applied to wounds in mice once per day for a total of 13 days.

Discussion
The plausible mechanism of the reaction is depicted in Scheme 1. Initially, the chalcone derivative 1 is deprotonated by tertiary amine, yielding a nitrogen-nucleophile I which would attack 2-arylidene-1,3-indandiones 2 to generate aza-Michael adduct II, through a first Michael addition (Scheme 1). The second intramolecular Michael addition upon II to provide III with both R 1 and R 2 in the same side and subsequent protonation and enolization would result in the spiro-tetrahydroquinolines 3 in high yields. (E) Comparison of the effects of the optimal concentrations of 3b, 3c, and 3i on wound healing. 3b, 3c, and 3i are arranged in accordance with the specific concentration of each compound that was used in a gel applied to wounds in mice once per day for a total of 13 days.

Discussion
The plausible mechanism of the reaction is depicted in Scheme 1. Initially, the chalcone derivative 1 is deprotonated by tertiary amine, yielding a nitrogen-nucleophile I which would attack 2-arylidene-1,3-indandiones 2 to generate aza-Michael adduct II, through a first Michael addition (Scheme 1). The second intramolecular Michael addition upon II to provide III with both R 1 and R 2 in the same side and subsequent protonation and enolization would result in the spiro-tetrahydroquinolines 3 in high yields.
As shown in Scheme 1, the transition states have two isomers (syn-and anti-isomers) that can interchange with each other. In the piperidine ring, if the green hydrogen atom heads upward (syn-isomer), the steric effect of the sulfonamide and enone moiety was less impact, and compound 3 was more easily formed. On the contrary, green enone moiety heads upward in the anti-isomer which shows more steric effect and makes it more difficult to form 3′.
In previous studies that have examined wound healing, most researchers have focused on natural products, such as extracts and extensions of Chinese herbal medicines or marine natural products. Only a few studies have examined the effects of synthetic chemicals on wound-healing outcomes. In this study, we explored the efficacy of wound healing by studying synthetic compounds and their related skeletal extensions. All compounds were preliminarily studied in human epidermal cells (HaCaT). Compound 3i was selected because, at low concentrations, the survival rate of cells was over 80%. Compounds 3b, 3c, 3k, and 3m were shown to have less effect on the growth of HaCaT cells, based on the results of the MTT assay (Supplementary Figure  S1). We performed a scratch analysis and observed that compound 3i has positive effects on wound healing. In the MTT assay, 3i displayed the significant inhibition of cell growth at the 25 µM concentration. Due to the different outcomes observed for these two experiments, we considered whether 3i could be effective in animal models. Compared with 3b and 3c, 3i was demonstrated to represent the best compound for healing wounds on mouse skin. The wound was completely healed by the 11th day in mice treated with 3i. Based on the results of the MTT and wound-healing assay, 3i is thought to promote Scheme 1. Proposed mechanism of synthesizing spiro-tetrahydroquinolines by one-pot reaction.
As shown in Scheme 1, the transition states have two isomers (syn-and anti-isomers) that can interchange with each other. In the piperidine ring, if the green hydrogen atom heads upward (syn-isomer), the steric effect of the sulfonamide and enone moiety was less impact, and compound 3 was more easily formed. On the contrary, green enone moiety heads upward in the anti-isomer which shows more steric effect and makes it more difficult to form 3 .
In previous studies that have examined wound healing, most researchers have focused on natural products, such as extracts and extensions of Chinese herbal medicines or marine natural products. Only a few studies have examined the effects of synthetic chemicals on wound-healing outcomes. In this study, we explored the efficacy of wound healing by studying synthetic compounds and their related skeletal extensions. All compounds were preliminarily studied in human epidermal cells (HaCaT). Compound 3i was selected because, at low concentrations, the survival rate of cells was over 80%. Compounds 3b, 3c, 3k, and 3m were shown to have less effect on the growth of HaCaT cells, based on the results of the MTT assay (Supplementary Figure S1). We performed a scratch analysis and observed that compound 3i has positive effects on wound healing. In the MTT assay, 3i displayed the significant inhibition of cell growth at the 25 µM concentration. Due to the different outcomes observed for these two experiments, we considered whether 3i could be effective in animal models. Compared with 3b and 3c, 3i was demonstrated to represent the best compound for healing wounds on mouse skin. The wound was completely healed by the 11th day in mice treated with 3i. Based on the results of the MTT and wound-healing assay, 3i is thought to promote wound healing by promoting cell proliferation. In summary, we hope that the synthesis of spiro-tetrahydroqunioline derivatives might provide a new method for identifying chemicals for application in future wound-healing research.

Materials and Methods
All chemicals were analytical grade, purchased from Sigma-Aldrich ( (1) 2-aminobenzyl alcohol (20 mmol, 2.46 g) was dissolved in 100 mL DCM, and benzenesulfonyl chloride (22 mmol, 4.18 g) and 1 mL pyridine were added. The mixture was stirred for 12 h at r.t. The solvent was removed under reduced pressure. The residue 4 was not purified and dissolved in 50 mL DCM. Pyridinium chlorochromate (30 mmol, 6.46 g) was added and the resulting solution was stirred at r.t. for 4 h. The reaction mixture was filtered by Celite 545 and washed by DCM. After solvent was removed under reduced pressure, the residue was purified by flash chromatography (DCM: Hexanes = 2:1) to give 5, yield 97%. 5 (3 mmol, 0.83 g) was dissolved in 15 mL toluene, 6 (3.3 mmol) was added. The mixture was stirred at 80 • C for 12 h. The solvent was removed under reduced pressure, residue was purified by flash chromatography to yield 1 as shown in Scheme 2.  132.9, 131.1, 131.0, 129.0, 128.63, 128.60, 127.9, 127.3, 127.2, 127.1, 124

Spiro-Tetrahydroquinoline (3)
Compounds 1 (0.2 mmol) and 2 (0.2 mmol) were dissolved in 1 mL DCM, 1.2 mg DABCO (5 mol%) was added as catalysis. The reaction was sealed and stirred at 30 • C. After the reaction completed, the mixture was quenched with 1 N hydrochloric acid aqueous solution and extracted with DCM. The organic layer was washed with water, dried over MgSO 4 , filtered, re-crystallized in ethanol and hexane.

Splints
The outer diameter, inner diameter, and thickness of round silicone splints were 26, 16, and 500-600 mm, respectively. All procedures and instruments were conducted and prepared, respectively, under aseptic conditions, which were maintained using autoclaves, ethylene oxide gas, 70% ethanol, and povidone-iodine. Mice were anesthetized with pentobarbital. An electric razor was used to remove the back hair, and a circular mark (1.0 cm diameter) was placed on the center of the lumbar area; this section of skin was totally excised using scissors. A splint was inserted beneath the skin near the wound defect and attached to the fascia with 6-stitch ligations. The splint was then fixed to the skin with surgical silk thread (6 stitches at regular intervals) [36,37].

Preparation of Sample-Containing PEG-Based Ointment
PEG 400 Da (100 mg) and PEG 4 kDa (20 mg) were mixed at a ratio of 5:1. The mixture was then heated to above 85 • C until it became a clear solution. The prepared compound (1 mg) was added to the solution before the solution was allowed to cool to room temperature, forming a gel [38].

Animal Experiments
Compounds were tested in 8-week-old wild-type male C57BL/6 mice (BioLASCO Taiwan Co., Ltd., Taipei, Taiwan). The mice were artificially wounded using splints. The control gel or the compound-containing gel was applied daily and photographed to measure the wound area for 13 days. The mice were sacrificed by CO 2 exposure, and the tissues at the location of the silicone ring were removed for subsequent analysis.

Statistical Analysis
Analyses were performed in triplicate, and the results were expressed as mean ± SD. Analysis of variance (ANOVA) was conducted, followed by Dunnett's post hoc test, to determine significant (p < 0.05) differences. Statistical analyses were performed using GraphPad Prism v8.0 (GraphPad Software, San Diego, CA, USA).

Conclusions
In this article, we developed a one-pot reaction method that requires only a small amount of catalyst to obtain nonselective isomers with high selectivity. In addition, the experimental procedure is simple and only requires extraction and recrystallization. The reaction was performed at room temperature and had a good yield. This reaction will be studied in our further research by using different catalysts. We hope that the synthesis strategy described in this article can be widely applied to various starting materials with diverse substituents. We also hope that the synthesis of spiro-tetrahydroqunioline derivatives could provide a new method for identifying chemicals that can be applied to wound-healing research. Further structural modifications and biological evaluations are ongoing, and the results will be reported in due course.