Synthesis and evaluation of naphthalene derivatives as potent STAT3 inhibitors and agents against triple-negative breast cancer growth and metastasis

Triple-negative breast cancer (TNBC) represents the worst prognostic subtype of breast cancer and lacks targeted therapeutic drugs. Signal transducer and activator of transcription 3 (STAT3) is overexpressed and constitutively activated in TNBCs and associated with poor patient outcomes. However, no agents targeting STAT3 have been successfully developed and marketed. Selective Estrogen Receptor Modulators (SERMs) have been reported as potential inhibitors of the IL-6/STAT3 signaling pathway. Naphthalene compounds have good pharmacological activity and significant anti-cancer activity. In this study, we synthesized a new series of naphthalene derivatives with the general structure of SERM and evaluated their effects on TNBC and STAT3 signals. A new series of compounds based on the scaffold of SERMs and an amino group were designed and screened based on the structure–activity relationship by MTT assay. The binding activity of SMY002 to STAT3 was predicted and validated by docking and SPR. The STAT3 signaling target and anti-cancer effects of SMY002 were evaluated with three TNBC cell lines and the mice transplanted tumor model. Among the compounds, SMY002 displayed the most potent activity, which could directly interact with STAT3 SH2-domain, and strongly inhibit the phosphorylation, dimerization, nuclear distribution, transcriptional activity, and target genes expression of STAT3. Furthermore, SMY002 markedly suppressed migration, invasion, survival, growth, and metastasis of TNBC cells in vitro and in vivo via down-regulating the expression of Cyclin D1 and MMP9. SMY002 can significantly inhibit the growth and metastasis of TNBC cells by targeting the STAT3 signal.


Introduction
Breast cancer (BC) has become the most prevalent cancer and the leading cause of cancer-related mortality worldwide [1]. About 10-15% of BC patients are triple-negative breast cancers (TNBCs) [2]. Due to the lack of expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2), TNBC patients are insensitive to endocrine therapy and HER2-targeted drugs [3]. Moreover, TNBCs are prone to relapse or metastasis in the early stage because of their substantial heterogeneity and invasiveness. The poor outcome highlights the necessity to exploit more efficacious agents for TNBC therapy.
Signal transducer and activator of transcription 3 (STAT3) plays essential roles in multiple signaling pathways. A prominent function of STAT3 is to transmit signals from the cell membrane to the nucleus and activate the transcription of its targeted genes. Many cytokines and growth factors can activate Janus kinases (JAKs) by binding to their receptors (such as IL-6/gp130) [4]. STAT3 can be phosphorylated at tyrosine 705 (Tyr705) by JAKs, Src, or BCR-ABL1 enzymes and form dimmers through the Src homology domain 2 (SH2) of STAT3. The STAT3 dimmers can enter the nucleus [5][6][7]. Then, STAT3 directly binds to the promoters of its target genes and regulates the transcription of these genes, such as CCND1, BCL-XL, SURVIVIN, and matrix metalloproteases (MMPs) [8][9][10][11]. High expression of these genes is associated with cancer proliferation, survival, angiogenesis, metastasis, and immune evasion. Extensive studies certified that STAT3 is overexpressed or constitutively activated in most human cancers, including TNBCs, and is closely related to malignant progression [12][13][14].
Many articles have reported on STAT3 inhibitors in cancer therapy during the past decades. A lot of natural compounds (e.g., silibinin, osthole, toosendanin, and resveratrol) [15,16] and synthetic molecules (e.g., AZD9150, BBI608, Stattic, and TTI-101) [17,18] showed good performance as potential anti-cancer agents by directly or indirectly targeting STAT3. In addition, some clinical drugs were discovered to be STAT3 signaling pathway blockers using drug repositioning, including badoxifene, bortezomib, and salinomycin [19]. It has been published that selective estrogen receptor modulators (SERMs) such as badoxifene and raloxifene exhibit anti-cancer effects by inhibiting IL-6/GP130/STAT3 signaling in TNBC cells [20,21]. However, high toxicity and low potency were the main challenges of them in translating from bench to bedside.
Naphthalene compounds have been widely developed into anti-tumor drugs for their good pharmaceutical activity, chemical stability, and cellular permeability. Amonafide and Mitonafide were the most representative naphthalene compounds, showing excellent activity in leukemia cells and solid tumors [22,23]. Wang and coworkers synthesized a novel naphthoquinone derivative, NTDMNQ, which significantly induced apoptosis by inhibiting the phosphorylation of MAPK/AKT/STAT3 signalings in gastric cancers [24]. These studies indicate that naphthylamine compounds offer the opportunity to be modified into STAT3 target anti-cancer drugs.
In this study, we synthesized and screened a new series of naphthylamine derivatives with the structure of 4-nitronaphthol and SERM as scaffolds. Among these compounds, SMY002 has the advantages of directly targeting STAT3, strong anti-cancer activity, and minimal side effects in vivo. Therefore, SMY002 holds the potential to be developed as novel STAT3 inhibitory tools or anti-TNBC treatment agents.

Computational docking
We obtained the crystal structure of the STAT3-SH2 domain from the RCSB protein data bank (PDB code: 1BG1). We prepared its form by removing substructures, repairing the 1 3 sidechains, and adding hydrogens and charges with Auto-Dock Vina. The results were obtained based on SCORE and the protein-compound complex analysis.

Cytotoxic assay
Cells were cultured in 96-well plates with a density of 5 × 10 3 per well in triplicate. The next day, SMY002 was added at a concentration range from 0.01 to 100 µM for 48 h. DMSO was utilized as the solvent. Cells exposed to SMY002-free DMSO were used as controls. The data were measured by MTT assay, compared with control cells, and converted to a percentage. The inhibitory effect was analyzed by GraphPad Prism 6.0.

Luciferase assay
Cells were cotransfected with the STAT3-Luc Reporter gene plasmid (YT455, Biolab Technology of Beijing, China) and pRL-TK vector (Promega, US) at a ratio of 1:200 (w/w) with Lipofectamine 2000 (Invitrogen, US). 48 h later, SMY002 was added to corresponding wells for another 24 h. Cells were extracted and assayed with the Dual-Luciferase Reporter Assay System (E1910, Promega, US). The luminescence values were examined by a multifunctional microplate reader (Varioskan Flash, Thermo-Fisher, US). The STAT3-dependent promoter activity was determined by the ratio of firefly luminescence value/Renilla luminescence value.

Real-time PCR
The primers used were as follows: GCT GCG AAG TGG AAA CCA TC and CCT CCT TCT GCA CAC ATT TGAA for CCND1; TGC ATT GTT CCC ATA GAG TTCCA and CCT GAA TGA CCA CCT AGA GCCTT for BCL-XL; TGT ACC GCT ATG GTT ACA CTCG and GGC AGG GAC AGT TGC TTC T for MMP9; and CAC GAT GGA GGG GCC GGA CTC ATC and TAA AGA CCT CTA TGC CAA CAC AGT for β-actin. Real-time PCR was analyzed using SYBR Green PCR Master Mix as recommended by the manufacturer (Thermo-Fisher, US). The cycle threshold (CT) values of target genes were standardized by that of β-actin. All the data were analyzed with the comparative quantitative method (2-ΔΔCt).

Zymography assay
The activity of MMP9 was analyzed with an MMP Zymography Assay Kit (P1700, Applygen, China). The cell culture supernatant was collected by pipettes and centrifuged (12,000 rpm, 15 min, 4 °C), then diluted with 2 × nonreducing PAGE loading buffer (1:1), and 20 μl samples were loaded and separated directly. An SDS-PAGE gel containing MMP9 substrate protein (substrate G) was prepared. The electrophoresis and stain of gels were performed according to the standard procedures. Gels were scanned on the Gel-Doc XR System (Bio-Rad, US).

Scratch assay
Cells were plated in a monolayer with more than 90% fusion. A pipette tip was utilized to draw straight and evenly scratches in each well. Cells were rinsed with PBS. The culture solution was replaced with DMEM containing low serum (0.5% FBS) for each well, and SMY002 was added. Wound healing was observed and photographed with an Olympus CKX41 microscope system.

Transwell
For migratory ability assay, 2 × 10 5 cells suspended with 100 µl DMEM were added to the upper chambers (8 µm The initial hit compounds of 4-nitronaphthol and SERMs. B Structural characteristics and groupings of potential TNBC inhibitors. The compounds can be divided into three structural groups. The compounds of group I (including SMBAa007) contain an amide substitution at the 4 -position (the -R group highlighted in yellow) with the highest IC50 values. The group II compounds showed the most vital activity (containing SMY006) with ether (the -R group highlighted in blue). Compounds in group III with medium activities have a carboxyl that combines the 4-nitronaphthol and SERM scaffold, such as a (SMY001) (the -R group highlighted in pink). C SMY002, a kind of hydrochloride of SMY006. D MTT assay screening of SMY002 and SMY001 against mammary epithelial cells and TNBC cells pore sizes, Corning, US) and supplemented with or without SMY002. For invasive ability assay, the upper chamber was coated with 100 µl Matrigel (1 mg/ml, BD, US) prior to seed cells. 800 µl of DMEM containing 20% FBS was added into the lower chambers. Cultured for 48 h, cells were fixed with methanol for 30 min and stained with 0.1% crystal violet for 30 min at room temperature. The inserts were immersed twice with PBS. The membranes were photographed using an Olympus CKX41 microscope system and counted by ImageJ software.

Cell cycle and apoptosis
Cells were trypsinized and rinsed with PBS. 1 × 10 6 cells were fixed in cold 70% ethanol (diluted with sterile water) for 2 h at 4 °C. The follow-up operations followed the instructions of the DNA content detection kit (Solarbio, China). The fluorescence of PI-stained cells was read by a flow cytometer (BD, US) at the wavelength 630 nm and analyzed with Cell Quest software. Apoptosis assays were performed using the Annexin V-FITC apoptosis detection kit (Solarbio, China) by flow cytometry.

Animal studies
All animal experiments were conducted following the regulations of the Laboratory Animal Welfare and Ethics Committee and approved by the local authorities. Five-week-old female BALB/c mice were obtained from Beijing Vital River Laboratory Animal Technology. 4T1 cell suspension (1 × 10 5 cells/mouse) was subcutaneously injected into the mice. All mice were randomly divided into two groups, with five in each. After seven days of implantation, the mice were administered with PBS or SMY002 (30 mg/kg in PBS as a solvent) by daily gavage. The body weight and tumor sizes of mice were detected every other day. Four weeks after implantation, all mice were euthanized with isoflurane. The transplanted tumors were dissected, weighed, imaged, and fixed in 10% neutral formalin.

Statistical analysis
All data were presented as mean ± SD. Data of two groups were evaluated by Student's unpaired t test. For comparison of one factor on each level of the other factor in three or more groups, one-way ANOVA tests were used. For multiple comparisons to explore possible interaction between different factors, two-way ANOVA tests were adopted. p < 0.05 was considered statistically significant.

Structure design, screening, and synthesis of SMY002
According to the MTT assays (SI. Fig. S1), compounds with 4-nitronaphthol and SERMs (such as Badoxifene and Raloxifene) have moderate anti-tumor activity in TNBC cells. The general structure of 4-nitronaphthol and SERMs was used as the reference scaffold of potent STAT3 inhibitors (Fig. 1A). Amino group replacing hydroxyl designation was introducing more polar interaction. The amide group substituted by the ethoxy group aimed to reduce steric hindrance between the phosphorylation binding site of the STAT3 SH2 domain and compounds. Compounds were grouped and ranked in decreasing pharmacophore similarity in MTT assays against 4T1 cells (Fig. 1B and Si. Fig. S2). The compounds in the second group showed potent inhibitory activity, while group III compounds had weak activity. The compounds in group I showed minor potency in inhibiting cell viability.
Since SMY006 in group II showed the most potent activity (Fig. 1B), it was chosen for further study. By preparing and screening various of its salt compounds, SMY002 (a kind of hydrochloride of SMY006) was produced (Fig. 1C) and showed a more selective cytotoxic effect in TNBC cells than MCF-10A (a normal mammary epithelial cell line) cells (Fig. 1D and SI. Fig. S3A-3D). In addition, SMY002 has stronger activity than the commercially available STAT3 blockers such as WP1066 and STATTIC in TNBC cells (SI. Fig. S9A-9C).
The optimized synthesis steps of SMY002 are shown in Scheme 1. Compound 2 was synthesized by prepared compound Methylparaben and 2-Piperidinoethanol in the presence of PPh3 and DIAD through a mitsunobu reaction. Compound 2 was reduced by LiAlH4 in THF under 0 °C to give compound 3. Compound 4 was obtained by Compound 3 and 1-Fluoro-4-nitrona-phthalene in the presence of NaH in THF through a nucleophilic reaction. Finally, compound 4 was reduced by Fe in 1 N. HCl and EtOH mixture solvent to give SMY002. The synthetic details and spectrographic data of the intermediates and compound SMY002 are shown in SI. Materials and methods and SI. Fig. S4-S6.

SMY002 directly interacts and binds to STAT3 with a high affinity
In predicting whether SMY002 interferes with the STAT3 protein, we obtained the crystal structure of the STAT3-SH2 domain from the RCSB protein data bank (PDB code: 1BG1) and prepared its design with Auto Dock Vina. The DFT calculations were carried out using the Gaussian 16 program package [27]. The geometry optimization of minima was carried out at the B3LYP-D3 level [28] with the 6-31G(d) basis set. According to the computational docking, SMY002 (cyan) docked in the phospho-Tyrosine-binding pocket of the STAT3-SH2 motif (light blue) ( Fig. 2A). Nine residues of Lys591, Glu594, Arg609, Ser613, Ile634, Ser636, Val637, Glu638, and Pro639 were found close to SMY002 (Fig. 2B). The model also indicates that SMY002 interacts stably with the Arg609 and Arg636 residues via two hydrogen bonds (black dash lines) (Fig. 2C). The ether bond oxygen atom in the ligand forms one hydrogen bond with the amine hydrogen atom on the branch chain of Arg609 (Fig. 2C). Meanwhile, the oxygen atom of the ether bond in the ligand forms the other hydrogen bond with the hydroxyl hydrogen atom of the Ser636 branched-chain (Fig. 2C). Thus, SMY002 might interact directly and stably with the STAT3 protein. The Surface Plasmon Resonance (SPR) experiments were carried out as described previously [25,26]. The results showed that SMY002 could compete by blocking EGFR pY1068-peptide, the ligand of STAT3, for binding to Stat3 with IC50 values of 2 µM (Fig. 2D, E). These data manifested that SMY002 has a high affinity to STAT3. The activity of SMY002 was significantly reduced in MDA-MB-231 cells knocked-down of STAT3 expression by siRNA-STAT3  Fig. S10A, 10B). Therefore, SMY002 plays anti-cancer actions mainly by targeting STAT3.

The inhibitory effects of SMY002 on STAT3 activation
Then, the effects of SMY002 on the transcriptional regulatory function of STAT3 were detected. Figure 3A-C shows that SMY002 treatment decreased the STAT3 reporter gene activity in a dose-dependent manner in TNBC cells. In addition, the downstream target genes of STAT3, such as CCND1, BCL-XL, and MMP9, were also dramatically decreased at the mRNA expression levels (Fig. 3D-F), which confirms that SMY002 inhibits the transcriptional activity of STAT3.
We further evaluated whether SMY002 suppresses the phosphorylation of STAT3 (Tyr705), which is crucial for its activation. Figure 3 G-I shows that SMY002 treatment suppressed STAT3 activation in a dose-dependent manner. Cyclin D1 expression was gradually down in corresponding with the change of p-STAT3. BCL-XL expression was also decreased following SMY002 treatment. However, SMY002 did not impair the p-JAK2, p-ERK1/2, or p-AKT expression. These results displayed that SMY002 can specifically suppress STAT3 activation, not by inhibiting its upstream kinase JAK2 or other off-target effects but by directly interfering with STAT3 phosphorylation.
The dimerization and subcellular localization of p-STAT3 (Tyr705) is crucial for STAT3 binding to specific DNA elements. After SMY002 treatment, the inactive monomers of STAT3 were increased in a dose-dependent manner, which implies STAT3 dimerization was impeded (Fig. 3J-L). The nucleoprotein separation experiments show that SMY002 considerably diminished the expression of p-STAT3 (Tyr705) in both cytoplasm and nucleus (Fig. 3M-O). To further support this finding, indirect immunofluorescence experiments were performed. As SI. Figure 8 shows, the nuclear localization of p-STAT3 (Tyr705) in 4T1 cells without SMY002 treatment was robust regardless of IL-6 stimulation. SMY002 treatment markedly reduced the nuclear staining of p-STAT3 (Tyr705). Therefore, SMY002 can impede the activation, dimerization, and nuclear localization of STAT3 by preventing the phosphorylation of Tyr705.

SMY002 inhibits the survival and cell cycle progression of TNBC
We next assessed the effect of SMY002 on survival and cycle progression of TNBC. The apoptosis ratio of MDA-MB-468 cells was increased from 1.75% to 17.54% and 54.47%, respectively, after treatment with SMY002 10 or 20 µM (Fig. 4A,B). Similar data were obtained in MDA-MB-231 cells (Fig. 4C). In addition, SMY002 induced significant G1 phase arrest and visibly shortened S phase in TNBC cells (Fig. 4D-F). These results reveal that SMY002 can induce cell apoptosis and prolongs the cell cycle.

SMY002 attenuates the motility of TNBC in vitro
One of the critical malignant phenotypes of TNBC cells is their potential for metastasis. To investigate whether SMY002 suppresses STAT3-promoted tumor cell migration, we evaluated the effect of SMY002 on TNBC cell migration by scratch assays (Fig. 5A-F). Compared with the control, SMY002 treated cells showed significantly lower migration activities, as evidenced by the delay of wound closure observed at different time points. We utilized another approach, transit cell culture, to assess the cell migration and invasion abilities. The results of Fig. 5G-L reveal that the control treatment cells displayed a remarkable capacity to pass through the membrane or invade through a reconstituted basement membrane (Matrigel) of a branchial chamber.
To explore why SMY002 inhibits cell migration and invasion, we detected the protein expression levels of MMP9, a target gene of STAT3, after treatment with SMY002 for 48 h in TNBC cells. Western blot results displayed that SMY002 repressed MMP9 expression in dose-dependently manner (Fig. 5M-O Topline). We performed a zymography assay to disclose the effect of SMY002 on MMP9 activity. As shown in the bottom line of Fig. 5M-O, the transparent protein bands on the blue background of the stained gels, which indicate the location of MMP9, were weaker with increasing treatment with SMY002, demonstrating that SMY002 reduced the activity of MMP9. The above findings suggest that SMY002 effectively suppressed the survival, cell cycle progression, mobility, and invasion of TNBC cells in vitro.

SMY002 reduced the growth and metastasis of TNBC cells
We finally evaluated the therapeutic effect of SMY002 in 4T1 cells transplanted in mice. Compared with the vehicle (PBS) treatment group, SMY002 significantly reduced the volume and weight of tumors ( Fig. 6A-C). Xenografts in the SMY002 treatment group developed less lung metastasis than those in the vehicle group (white nodule in the metastatic loci decreased from 35 to 9.8) (Fig. 6D,E). Immunohistochemical staining results meant that p-STAT3 (Tyr705) and MMP9 expression in xenograft tumors treated with SMY002 was weaker than that of the vehicle group (Fig. 6F,G). These data are compatible with the results in vitro that SMY002 blocks the growth and invasion of TNBC cells via suppressing STAT3 activation. Compared with the vehicle group, the body weight of mice in the SMY002 administration group was not significantly reduced (Fig. 6H). These data underlined that SMY002 potently suppressed the growth and dissemination of TNBC by inhibiting STAT3 activity.
In the acute toxicity test, death of SD rats occurred mainly between 6 and 24 h after administration, with the main signs of toxicity being weight loss and slow movement (SI. Fig. S11). As the dose increased, the number of rats showing signs of toxicity increased correspondingly. The mortality of rats in each group is shown in Supplementary Table 1. The LD50 of acute oral toxicity for rats was 748.199 mg/kg (equivalent to 1068.855 mg/kg in mice) by a standard procedure for the modified Spearman-Kärber method [29]. The 95% confidence interval for the LD50 of rats was 659.356-848.946 mg/kg (equivalent to 941.937-1212.780 mg/kg in mice). These results implied that the potentially toxic side effects of SMY002 may be controlled by designing a dosing regimen.