Abstract

   SMYD3 (SET and MYND domain‐containing 3) is a histone lysine
   methyltransferase highly expressed in different types of cancer(s) and
   is a promising epigenetic target for developing novel antitumor
   therapeutics. No selective inhibitors for this protein have been
   developed for cancer treatment. Therefore, the current study describes
   developing and characterizing a novel small molecule ZYZ384 screened
   and synthesized based on SMYD3 structure. Virtual screening was
   initially used to identify a lead compound and followed up by
   modification to get the novel molecules. Several technologies were used
   to facilitate compound screening about these novel molecules' binding
   affinities and inhibition activities with SMYD3 protein; the antitumor
   activity has been assessed in vitro using various cancer cell lines. In
   addition, a tumor‐bearing nude mice model was established, and the
   activity of the selected molecule was determined in vivo. Both RNA‐seq
   and chip‐seq were performed to explore the antitumor mechanism. This
   work identified a novel small molecule ZYZ384 targeting SMYD3 with
   antitumor activity and impaired hepatocellular carcinoma tumor growth
   by reducing H3K4 trimethylation of the Rac1 promoter triggering the
   tumor cell cycle arrest through the AKT pathway.

   Keywords: antitumor, H3K4me3, Rac1 promoter, small molecule, SMYD3,
   ZYZ384
     __________________________________________________________________

   The current study describes developing and characterizing a novel small
   molecule ZYZ384 screened and synthesized based on SMYD3 structure.
   ZYZ384 targeted SMYD3 with antitumor activity and impaired
   hepatocellular carcinoma tumor growth by reducing H3K4 trimethylation
   of the Rac1 promoter triggering the tumor cell cycle arrest through the
   AKT pathway.

   graphic file with name MCO2-5-e711-g005.jpg

1. INTRODUCTION

   Cancer(s) are still the leading cause of disease‐related death
   worldwide, and epigenetic systems have become important targets for
   antitumor research.[44] ^1 , [45]^2 Epigenetics is the study of
   heritable changes in gene expression that occur without alterations to
   the underlying DNA sequence. However, these modifications are mainly
   reversible. The epigenetic modification mechanisms include DNA
   methylation, chromatin remodeling, histone modifications, and RNA
   interference.[46] ^3 These post‐translational changes are intimately
   linked to tumorigenesis and cellular development.

   In recent times, it has been reported that the histone
   methyltransferase SMYD3 (SET and MYND domain‐containing protein 3) is
   highly expressed in various human tumors and that overexpression
   enhances K‐Ras‐associated signaling in cancer cells.[47] ^4 , [48]^5 To
   date, SMYD3 overexpression in cancerous tissues has been reported to
   occur in liver cancer, colon cancer,[49] ^6 breast cancer,[50] ^7
   pancreatic cancer,[51] ^8 prostate cancer,[52] ^9 bladder cancer,[53]
   ^10 gastric cancer,[54] ^11 malignant glioma,[55] ^12 esophageal
   squamous cell carcinoma,[56] ^13 chronic lymph cell leukemia,[57] ^14
   and cervical cancer.[58] ^15 This information points to SMYD3 as a
   cancer drug discovery and development candidate. At present, there are
   few descriptions of molecules that can inhibit SMYD3 in the literature:
   several have issues linked to low inhibition efficiency,[59] ^16 or are
   not applied to in vivo study.[60] ^17 Although there is a small
   molecule SMYD3 inhibitor that is considered to have good
   bioavailability,[61] ^18 or others can inhibit cancer cell
   proliferation,[62] ^19 , [63]^20 currently, no molecules targeting
   SMYD3 have been reported to be used in clinical cancer treatment.

   Our current discovery of the novel small molecule targeting SMYD3 may
   open a new avenue in future cancer treatment.

   Opportunities for using technologies by researchers in drug discovery
   are becoming more widely available and facilitating the identification
   of novel drug molecules. Indeed, computer‐assisted drug virtual
   screening models have allowed for the assessment of drug screening
   through computer simulation technology and can be used in the
   evaluation of docking small molecules to protein targets.[64] ^21
   Similarly, biolayer interferometry (BLI), a label‐free technology for
   detecting biomolecular interactions, can provide high‐throughput
   biomolecular interaction information in real time and offers kinetic
   and affinity information. Studies assessing the interaction between
   inhibitors and proteins can often be challenging, so molecular dynamics
   (MD) simulations can be helpful in evaluating the stable binding
   pattern of ligands to proteins. Moreover, MD simulations enable us to
   derive trajectories containing all information about the stability and
   relationships of molecular interactions between ligand–protein
   complexes.[65] ^22

   In the current research, we have characterized drug‐like molecules
   initially identified via a virtual screening of small molecular weight
   compound libraries. This research led to the identification of a series
   of lead compounds that were assessed based on their inhibitory
   potential of SMYD3 (please refer to the [66]Supporting Information for
   details). The lead compound was then used to design and synthesize a
   series of novel small molecules. These molecules were tested in vitro
   and in vivo models for SMYD3 inhibitory properties and anticancer
   activities.

2. RESULTS

2.1. Design and synthesis of novel small molecules targeting SMYD3

   In order to get novel molecules targeting SMYD3, we first conducted
   virtual screening (Figure [67]S1) to identify lead compound 1 (Table
   [68]S1) and modified its structure. Two series of novel small molecules
   were obtained from the modification and synthesis of lead compound 1
   (Figure [69]1). For synthesis details please refer to Section 4.
   Nuclear magnetic resonance (NMR) verified the structure of each
   compound. The NMR and high‐resolution mass spectrometry (HRMS) of each
   novel small molecule, please refer to the Figures [70]S2–S37 for
   details. Two series of novel small molecules were successfully
   synthesized for subsequent experiments.

FIGURE 1.

   FIGURE 1
   [71]Open in a new tab

   Design and synthesis of novel small molecules targeting SMYD3. Reagents
   and conditions: (A) NH[2]NH[2]·H[2]O, methanol, reflux, 3 h; (B) AcOH,
   methanol, reflux, 8 h; (C) CH[3]I, potassium fluoride, acetonitrile,
   70°C, 24 h; (D) p‐toluenesulfonic acid, toluene, reflux, 8 h.

2.2. ZYZ384 (SMI1) has a good binding affinity to SMYD3

   Next, to test the binding affinity of novel small molecules to SMYD3,
   BLI analyses were performed to determine the binding affinity of the
   two series of novel small molecules targeting SMYD3. SMYD3‐glutathione
   S‐transferase (GST)‐tagged protein was immobilized onto GST biosensors
   and used to determine the binding affinity of all respective small
   molecules and labeled probes. The binding and dissociation curves were
   generated in real time by applying the SMYD3‐GST label proteins onto
   the biosensor surfaces, followed by monitoring small molecules
   interaction across several concentrations. The results for the SM and
   SMI series of compounds are shown in Table [72]1. SMI1 shows the good
   binding affinity of all compounds tested (R ^2 > 0.95, K [D] < 10
   E‐05). The binding curves for this interaction suggest a simple 1:1
   binding pattern (R ^2 = 0.953), the equilibrium dissociation constant
   (K [D]) of the interaction between SMYD3 and SMI1 was determined to be
   78 µM (steady‐state analysis, lower panel), better than two commercial
   inhibitors EPZ031686 (K [D]: 195 µM) and BCI‐121 (K [D]: 114 µM). These
   results show that SMI1 interacts and binds to SMYD3 protein and has a
   good binding affinity to SMYD3. SMI1 has been renamed ZYZ384 and
   further tested for anticancer activities using both in vitro and in
   vivo models.

TABLE 1.

   The binding affinity of small molecules to the target protein SMYD3.
   Compd     R ^2   K [D] (M)
   SMI1/8a   0.953  7.80 E‐05
   SMI3/8b   0.7514 4.47E‐05
   SMI11/7a  0.9085 1.15 E‐04
   SMI12/7c  0.5939 6.99 E‐06
   SMI13/7b  0.7926 9.57 E‐05
   SM20/4a   0.933  2.16E‐05
   SM17/4b   0.4431 4.67E‐06
   SM28/4c   0.8129 2.76E‐06
   SM31/4d   0.8931 2.60E‐04
   SM23/4e   0.9218 2.10E‐05
   EPZ031686 0.9745 1.95E‐04
   BCI‐121   0.9493 1.14E‐04
   [73]Open in a new tab

2.3. ZYZ384 binding with SMYD3 and inhibits SMYD3 activity

   The previous data show that ZYZ384 and SMYD3 have good binding
   affinity. Next, we used computer simulation methods to predict possible
   binding sites and modes. Molecular docking was assessed to predict the
   binding site interaction of ZYZ384 with SMYD3 protein. The binding
   energy of ZYZ384 and SMYD3 was calculated to be −8.73 kcal/mol with a
   reasonable combining pocket; this pocket contributes to maintaining a
   stable conformation of ZYZ384 and SMYD3 (Figure [74]2A: Best
   combination mode; D: Active binding pocket). The binding region
   includes multiple amino acid residues, namely, SER‐182, PHE‐183,
   CYS‐186, MET‐190, ILE‐214, TYR‐257, LYS‐297, ASP‐332, HIS‐366, and
   VAL‐368 (Figure [75]2B: two‐dimensional [2D] interaction of ZYZ384 and
   SMYD3). The interaction included the formation of hydrogen bonds, van
   der Waals forces, Pi–Pi conjugation, and other interactions. Indeed,
   the methoxy group on ZYZ384 forms hydrogen bond interactions with
   TYR‐257, SER‐182, ASP‐332, and LYS‐297 of SMYD3 protein with an average
   hydrogen bond distance of 3.125 Å. PHE‐183, ILE‐214, HIS‐366, VAL‐368,
   and MET‐190 of SMYD3 (Figure [76]2C: three‐dimensional [3D] interaction
   of ZYZ384 and SMYD3). The reported interactions are in maintaining the
   stability of the ZYZ384 and SMYD3 complexes.

FIGURE 2.

   FIGURE 2
   [77]Open in a new tab

   ZYZ384 binds and inhibits SMYD3 activity. (A) The best combination
   mode; (B) the two‐dimensional (2D) interaction of ZYZ384 and SMYD3);
   (C) three‐dimensional interaction of ZYZ384 and SMYD3; (D) the active
   binding pocket; (E) the root mean square displacement during the
   molecular dynamics (MD) simulation; (F) the root mean square
   fluctuation during the MD simulation; (G) the energetic decomposition
   of main individual residues; (H) the energetic contribution of main
   individual residues; (I) the SMYD3 Homogeneous Assay Kit was used to
   detect the SMYD3 activity exposed to different concentrations of
   ZYZ384. The alpha counts were recorded and analyzed with GraphPad 8.0.

   The root mean square displacement (RMSD) was used to measure the
   average distance between atoms and reveal any structural changes
   between SMYD3 and target compounds over time. High RMSD values indicate
   the target protein undergoing structural changes via conformational
   rearrangement induced by the ligand to fit binding pockets. In
   contrast, lower deviations indicate complex stability. According to the
   RMSD analyses, the complex system trajectories are comparatively stable
   from 0 ns, with RMSD values fluctuating within 0.1 nm (Figure [78]2E).
   This result indicates that the systems had reached convergence. Toward
   the end of the simulation (80‐100 ns), the complex showed a low RMSD
   (about 0.3 nm) with a smaller conformational change. This indicates a
   higher interaction efficiency and good stability of the ZYZ384, SMYD3
   interaction.

   The root mean square fluctuation (RMSF) was calculated from the
   movement of each residue along the trajectory around the mean position,
   revealing the flexibility of specific protein regions during MD
   simulations. The larger RMSF values correspond to regions with greater
   freedom that is potentially located in the active site at residues
   SER‐182, PHE‐183, CYS‐186, MET‐190, ILE‐214, TYR‐257, LYS‐297, HIS‐366,
   and VAL‐368. All values for RMSF were consistent with the molecular
   docking results (Figure [79]2F). The residues above 400 have the
   largest fluctuation, and these residues are known to have special
   functional correlations with RMSF values of more than 0.4 nm.

   The binding free energies of ZYZ384 and SMYD3 were calculated to be
   −94.405 ± 9.50 kcal/mol, respectively. These free energies are
   representative of van der Waals energy (−173.890 ± 9.798 kcal/mol),
   electrostatic energy (−13.585 ± 6.080 kcal/mol), polar solvation energy
   (113.27 ± 10.691 kcal/mol), and Solvent Accessible Surface Area (SASA)
   nonpolar solvation energy (−20.200 ± 1.042 kcal/mol). We subsequently
   calculated the energetic contribution of the potential key residues
   involved in the interactions of ZYZ384 and SMYD3. The results indicated
   that the following residues are important for the inhibition of SMYD3:
   SER‐182 (−0.1828 kcal/mol), PHE‐183 (−0.1828 kcal/mol), CYS‐186
   (−2.5796 kcal/mol), MET‐190 (−4.0405 kcal/mol), ILE‐214
   (−3.3782 kcal/mol), TYR‐257 (−0.0391 kcal/Mol), LYS‐297
   (1.5286 kcal/mol), ASP‐332 (3.0213 kcal/mol), HIS‐366
   (−0.4137 kcal/mol), and VAL‐368 (−6.1474 kcal/mol). VAL‐368 contributed
   to the total binding free energies among the identified residues.
   Interactions with LYS‐297 and ASP‐332 seem unfavorable in this model
   (Figure [80]2G). In addition, the energy decomposition of potential
   thermal residues was determined to identify which interactions dominate
   the binding free energy (Figure [81]2H). Discrete energy terms that
   contribute to binding free energy indicate that Molecular Mechanics
   (MM) energy (van der Waal energy and electrostatic energy) is the
   driving component of the binding in the docking complex. Polar
   solvation energy did not contribute favorably to the total interaction.
   Finally, the contribution of the nonpolar solvation energy of SASA to
   binding free energy was similar. The decomposition energy per residue
   of surrounding moieties indicates that the main beneficial energy
   contribution of ZYZ384 binding comes from residues PHE‐183, CYS‐186,
   MET‐190, and VAL‐368. These appear to be critical sites for ZYZ384 and
   would be important to produce any potential physiological effects
   following the interaction of this molecule with SMYD3.

   The homogeneous assay kit was used to evaluate ZYZ384 inhibition
   activity on SMYD3. As we know, SMYD3 can catalyze the di‐ and
   tri‐methylation status of lysine4 of histone H3. This protein is also
   reported to drive the methylation of a lysine residue present in the
   protein MAPK/ERK Kinase 2 (MAP3K2). Therefore, a series of different
   concentrations of ZYZ384 were incubated with SMYD3, and the activity
   was determined using the kit. The alpha counts were calculated and
   analyzed and showed that ZYZ384 inhibits SMYD3 in a
   concentration‐dependent manner.

   The above data revealed ZYZ384 inhibits SMYD3 activity and predict the
   possible binding site and mode with SMYD3.

2.4. ZYZ384 inhibits different cancer cell proliferation and suppresses the
overexpression of SMYD3

   The cytotoxicity of ZYZ384 was determined across several different
   cancer cell lines. 3‐(4,5‐Dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium
   bromide (MTT) assay was used to assess cell viability. The results
   showed that ZYZ384 (Figure [82]3A) significantly decreased cancer cell
   viability in a concentration‐dependent manner after 24 h incubation
   (Figure [83]3B–G). The cancer cell lines tested include human liver
   carcinoma cell HepG2, adenocarcinoma human alveolar basal epithelial
   cell A549, human colon carcinoma cell HTC116, human breast
   adenocarcinoma cell MDA‐MB‐231, and human pancreatic adenocarcinoma
   cell Miapaca2. ZYZ384 demonstrated the most effective inhibitory
   activity in human hepatoma HepG2 cells and could reduce cell
   proliferation at low concentrations, namely, IC[50] = 5.23 µM. In
   addition, ZYZ384 had low toxicity when tested in normal liver cells
   (LO2; Figure [84]3H). SMYD3 overexpression was conducted on LO2 cells
   by transfection and subsequently suppressed by ZYZ384 (Figure [85]3I).
   These results indicated that ZYZ384 Inhibits different cancer cell
   proliferation and suppress the overexpression of SMYD3.

FIGURE 3.

   FIGURE 3
   [86]Open in a new tab

   The cytotoxicity of ZYZ384 on different cell lines. (A) The structure
   of ZYZ384, (B) human liver cancer cell HepG2, (C) human liver cancer
   cell Huh7, (D) human normal liver cells LO2, (E) human pancreas
   adenocarcinoma Miapaca2, (F) human colon cancer cell HTC116, (G) human
   breast cancer cell MDA‐MB‐231, (H) human non–small‐cell lung cancer
   cell A549, (I) SMYD3 overexpression and ZYZ384 treatment on LO2. Data
   were expressed as mean ± SEM. ***: p < 0.001; **: p < 0.01 versus
   control, one‐way ANOVA analysis.

2.5. ZYZ384 is safe with potential druggability

   The toxicity of ZYZ384 (Figure [87]4A) was determined using C57BL/6
   mice in pilot studies to assess the safety and the maximum‐tolerable
   dose (MTD) determined (as shown in Figure [88]4B). No toxicity was
   reported in C57BL/6 mice orally administrated with 2 g/kg of ZYZ384 one
   time, and a survival rate of 100% was reached after 2 weeks of
   treatment. No abnormalities in food and water consumption are reported
   during this time, nor any unusual spontaneous activities; a reported
   LD5 for ZYZ384 >2 g/kg is noted for the current work. As shown in
   Figure [89]4C, the Absorption, Distribution, Metabolism, Excretion,
   Toxicity (ADMET) plot is a 2D plot of ADMET_PSA_2D versus
   ADMET_AlogP98. This diagram shows 95% and 99% confidence intervals for
   the BBB model and 95% and 99% confidence intervals for the HIA model.
   As indicated, ZYZ384 has salient ADMET properties and shows potential
   druggability (Figure [90]4C).

FIGURE 4.

   FIGURE 4
   [91]Open in a new tab

   In vivo study and druggability prediction of ZYZ384. (A) The structure
   of ZYZ384. (B) Safety and attempted evaluation of lethal dose for
   ZYZ384. C57BL/6 mice were orally administrated with 2 g/kg (n = 10) of
   ZYZ384 for 14 days, and the survival of the mice was monitored and
   recorded. (C) ADMET analysis of novel small molecules. Discovery Studio
   software (2020) was used to calculate ADMET properties. A, Absorption;
   D, Distribution; M, Metabolism; E, Excretion; T, Toxicity, the 99% and
   95% confidence intervals of the blood–brain barrier permeability (BBB)
   model, and the 99% and 95% confidence intervals of the human intestinal
   absorption (HIA) model. (D) The establishment of tumor‐bearing nude
   mice model and treatment method. (E) Tumor volume monitoring during
   treatment. (F) Body weight monitoring during treatment. (G) Tumor
   volume of each group. (H) Comparison of tumor volume in each group.
   Data were expressed as mean ± SEM. ADMET, Absorption, Distribution,
   Metabolism, Excretion, Toxicity. #: p < 0.05 versus positive control;
   ***: p < 0.001; *: p < 0.05 versus control, one‐way ANOVA analysis.

2.6. Growth suppression in HepG2 tumors of nude mice by ZYZ384

   The in vivo antitumor effect of ZYZ384 was assessed using a HepG2
   tumor‐bearing nude mouse model (cell‐derived xenograft, CDX;
   Figure [92]4D). The results show that the 100 mg/kg ZYZ384 promotes
   growth suppression in HepG2 tumors of nude mice that was better than
   that of sorafenib at 100 mg/kg (Figure [93]4E,G,H). Moreover, no
   noticeable body weight changes were observed during treatment
   (Figure [94]4F). Taken together, these results show that ZYZ384 is
   effective at inhibiting the growth of liver tumors in the elected
   animal model.

2.7. ZYZ384 reduced the H3K4me3 occupancy at the transcription start site of
the Rac1 gene in tumor tissues

   In order to explore the antitumor mechanism of ZYZ384, the inhibitory
   effect of ZYZ384 on SMYD3 was determined in tumor tissues. The results
   indicated that SMYD3 expression was highly expressed in the control
   group but significantly diminished in the ZYZ384 treatment groups
   (Figure [95]5A). In parallel, the expression of H3K4me3 was also noted
   to decrease in a similar pattern to that of SMYD3 (Figure [96]5B). The
   tumor tissues from control and high‐dosage treatment group animals were
   isolated. Since SMYD3 can tri‐methylated H3K4, ChIP‐seq analysis was
   used. This method allowed for annotating H3K4me3 occupancy on genes
   (Figure [97]5C). Global H3K4me3 occupancy decreased near annotation
   genes' transcription start site (TSS; Figure [98]5D). H3K4me3 is often
   enriched at activated promoters near TSSs. When combined with Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis
   (Figure [99]5E) and Gene Ontology (GO) functional enrichment analysis
   (Figure [100]5F), attention focus on the AKT pathway, cell
   proliferation, and migration functions. Visualized reads and signal
   strength of H3K4me3 for the Rac1 gene at the promoter region decreased
   following treatment with ZYZ384 (Figure [101]5G). The H3K4me3 occupancy
   at the TSS of the Rac1 gene in tumor tissues was reduced by ZYZ384.

FIGURE 5.

   FIGURE 5
   [102]Open in a new tab

   ZYZ384 inhibition effect on SMYD3 and H3K4me3 occupancy of annotation
   genes. (A) SMYD3 expression level in tumor tissue following treatment
   with ZYZ384. (B) H3K4me3 expression level in tumor tissue after
   treatment with different dosages of ZYZ384. (C) Distribution of peaks
   at gene functional regions. (D) Relative reads of gene body
   distribution (E) KEGG enrichment analysis. (F) GO enrichment analysis.
   (G) Visualization of reads comparison. Statistical results of related
   genes data are expressed as mean ± SEM. GO, Gene Ontology; KEGG, Kyoto
   Encyclopedia of Genes and Genomes. *: p < 0.05; **: p < 0.01; one‐way
   ANOVA analysis.

2.8. ZYZ384 impairs hepatocellular carcinoma tumor growth in vivo by reducing
Rac1 and triggering cell cycle arrest via the AKT pathway

   To further investigate the mechanism, the RNA‐seq of the control and
   treatment group was also performed. RNA‐seq interrogation identified
   regulated genes (Figure [103]6A,B). The clustering heat map was used to
   show significant increases in Cdkn1a (P21; Figure [104]6C). This
   protein plays a key role in the cell cycle. The AKT pathway is likely
   involved in tumor proliferation and migration. The changes of cell
   cycle‐related genes AKT, P21, CDK1, CyclinD1, CyclinD2 in this
   signaling pathway were confirmed by q‐polymerase chain reaction (PCR)
   and Western blots analysis (Figure [105]6D–G). Both phosphorylation
   sites (473, 308) of AKT are changed (Figure [106]6G). Previous CHIP‐seq
   results revealed ZYZ384 reduced the H3K4me3 occupancy at the TSS of the
   AKT upstream Rac1 gene, the transcription and expression of the Rac1
   gene are shown in Figure [107]6E,G. These results indicated that ZYZ384
   impairs hepatocellular carcinoma tumor growth in vivo by reducing Rac1
   and triggering cell cycle arrest via the AKT pathway.

FIGURE 6.

   FIGURE 6
   [108]Open in a new tab

   Molecular mechanism of ZYZ384 antitumor activity. (A) Differential gene
   volcano plot. (B) Venn diagram of gene coexpression. (C) Differentially
   gene expressed as presented as a clustering heat map. (D) P21
   expression. (E) The mRNA level of related genes was assessed using
   quantitative real‐time reserve transcription polymerase chain reaction
   followed by (F, G) Western blot protein expression analysis.
   Statistical results of related genes data are expressed as mean ± SEM.
   *: p < 0.05; **: p < 0.01; ***: p < 0.001; one‐way ANOVA analysis.

3. DISCUSSION

   The global rates of cancer incidence in the general population remain
   high, and there is a clear need to identify new therapeutic targets and
   drugs.[109] ^23 , [110]^24 The histone methyltransferase, SMYD3, is
   overexpressed in many types of cancer(s) yet is often underexpressed in
   corresponding normal healthy tissues.[111] ^4 , [112]^5 Moreover, SMYD3
   is considered a biomarker for the poor prognosis of liver cancer,[113]
   ^25 , [114]^26 and these traits make SMYD3 an ideal target for drug
   discovery initiatives. Therefore, the current research focuses on SMYD3
   and uses a combination of bioinformatics, classical pharmacology, and
   screening to identify novel SMYD3 therapeutics.

   Virtual screening and structural‐based drug design technologies,
   namely, various bioinformatics tools, including 3D structure
   information, has allowed researchers to study protein interactions with
   small molecular weight molecules.[115] ^17 , [116]^27 , [117]^28 ,
   [118]^29 Consequently, in the current work, computational docking, BLI
   analysis, and MD approaches were utilized in the current research. This
   integrated approach has allowed for the rapid analysis of small
   molecular compounds and the interrogation of interactions of candidate
   molecules with the SMYD3 protein. In addition, docking studies assisted
   in the rapid analysis of how selected compounds interacted with the
   SMYD3 protein at the amino acid residue level, allowing for the
   calculation of binding energies in protein–compound complexes.

   Using these approaches, we developed and designed several novel
   compounds; one molecule, ZYZ384, has SMYD3 inhibition effect. Molecular
   docking and MD simulation showed that ZYZ384 binds strongly to SMYD3 at
   regions, SER‐182, PHE‐183, CYS‐186, MET‐190, ILE‐214, TYR‐257, LYS‐297,
   ASP‐332, HIS‐366, and VAL‐368, respectively. This interaction is
   mediated by extensive hydrogen bond formation between amino acid
   residues and ZYZ384, Van der Waals force interactions, and Pi–Pi
   conjugation. PHE‐183, CYS‐186, MET‐190, and VAL‐368 are critical
   residues for ZYZ384 action and activity. In vitro assessment indicated
   that ZYZ384 binds to the SMYD3 protein in a concentration‐dependent
   manner, reducing the activity of the SMYD3 enzyme. Moreover, when
   applied to various cancer cell lines, ZYZ384 significantly reduced the
   proliferation of cells. Among the cells tested, ZYZ384 showed a
   significant inhibitory effect on the proliferation of human hepatoma
   HepG2 cells, and these results informed the in vivo mouse model
   studies.

   In mice, the maximum dose of ZYZ384 was administered by gavage
   (2 g/kg), and toxicology impacts were determined postadministration at
   14 days. No animals died within 14 days, and no detrimental impacts of
   ZYZ384 treatment, namely, feeding or drinking responses, impacts on
   body weight, or behavioral changes, were reported. In the tumor‐bearing
   animals, ZYZ384 was found to inhibit the growth of implanted HepG2
   tumors in nude mice correlating with the inhibition of SMYD3. RNA‐seq
   and ChIP‐seq were performed to explore the in‐depth antitumor
   mechanisms linked to ZYZ384. In epigenetic studies, H3K4me3 is enriched
   at promoters of activated genes near TSSs.[119] ^30 , [120]^31 ,
   [121]^32 ChIP‐seq revealed the H3K4me3 decreased at the promoter of the
   Rac1 gene following treatment with ZYZ384, which led to Rac1gene
   transcriptional repression. Rac1 is a member of the Rho GTPase family,
   which typically controls the assembly and disassembly of cytoskeletal
   elements.[122] ^33 In addition, Rac1 activity is critical for many
   cellular activities, including phagocytosis, migration, adhesion, and
   differentiation of multiple cell types,[123] ^34 and it also plays an
   important role in cell growth and cell cycle regulation.[124] ^35
   Importantly, the overactivation and overexpression of Rac1 are closely
   associated with aggressive tumor growth.[125] ^33 Similarly, RNA‐seq
   revealed parallel changes in the AKT pathway. In many forms of human
   cancers, genes associated with the AKT pathway are most frequently
   altered. As a result, the aberrant activation of this pathway is
   associated with cellular transformation, tumorigenesis, cancer
   progression, and drug resistance.[126] ^36 RNA‐seq indicated a
   significant increase of Cdkn1a (P21) after treatment of ZYZ384, Cell
   cycle‐related gene expressions have been evaluated and verified. GO
   functional annotation after treatment is also mainly enriched in the
   cell migration function, but the nude mice tumor‐bearing model cannot
   well simulate the invasion and migration phenotype of actual tumors, we
   will apply ZYZ384 on more alternative cancer models to explore its
   mechanism of action in the future.

   We were granted a Chinese patent (CN114246851A) to protect the
   structure and activity of ZYZ384, the effectiveness of ZYZ384 at high
   concentrations presents a challenge for future clinical use. This is
   due to the poor water solubility of ZYZ384. Currently, we are trying to
   transform it into hydrochloride or culture cocrystallization to improve
   its solubility, bioavailability, and druggability. Drug resistance and
   combination therapy of ZYZ384 on other models will be gradually carried
   out in the follow‐up work. Although we developed a small molecule
   targeting SMYD3, we still need to figure out whether it is a selective
   inhibitor, and related research is to be continued.

   Taken together, our study developed a novel molecule ZYZ384 targeting
   SMYD3 for hepatocellular carcinoma via reducing H3K4 trimethylation of
   the Rac1 promoter. ZYZ384 is a promising candidate drug that could be
   further developed to treat on cancer(s).

4. MATERIALS AND METHODS

4.1. Materials

   Reagents were obtained from the following: lead compounds were
   purchased from Guangzhou BaoYan Technology Co., Ltd. Company. A series
   of novel small molecules were synthesized by our group at the State Key
   Laboratory (SKL) Macau and verified by NMR. Dulbecco's modified eagle
   medium (DMEM), penicillin/streptomycin, 0.25% Trypsin‐EDTA, and fetal
   bovine serum (FBS) were all purchased from Gibco. The SMYD3 Homogeneous
   Assay Kit and purified SMYD3 N‐terminal GST tag protein was purchased
   from BPS Bioscience. MTT
   (3‐(4,5‐Dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazoliumromide) powder was
   obtained from Thermo Fisher. Sorafenib was got from MedChemExpres.
   SMYD3, H3, H3K4me3, P21, CyclinD1, CyclinD2, CDK1, Rac1, AKT, p‐AKT
   (308, 473), and Glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH)
   primary antibodies were all purchased from Abcam or Santa Cruz.

4.2. Structure‐based virtual screening

   The structure of SMYD3 was derived from the Protein DataBank (PDB ID:
   5CCL) and prepared with the Protein Preparation Wizard in Maestro
   (Schrodinger 2015) including removal of water and ions, protonation,
   addition of missing atoms, completion of missing groups, and use of
   OPLS3e force field for constraint minimization.[127] ^37 For the
   ligands, a total of 211,838 compounds from Specs databases were
   prepared using the property calculator module in the Schrödinger
   Maestro software, including calculating the physical and chemical
   properties of the small molecule compound library data and filtering by
   the Lipinski rules, protonation, and energy minimization. Before the
   virtual screening, it was necessary to dock the original ligand in the
   complex to investigate the docking method's reliability and then select
   the protein's original ligand as the centroid of the 10 Å box. Standard
   precision (SP) and extra precision (XP) methods were used to screen
   compound databases and construct the compound ranking score.

4.3. Structure modification based on lead compound 1

   In the structural design, based on the indolin‐2‐one structure of the
   lead compound 1, the Schiff base was added by Mannich reaction to
   obtain compounds. The Schiff hydrazone (C = N–) framework is an
   exceptionally adaptable drug‐like moiety that has recently been
   utilized in the development of cancer(s) treatments or cellular
   apoptosis.[128] ^38 The structures containing this bioactive moiety
   showed remarkable anticancer activities via the restraint of numerous
   kinds of enzymes, proteins, and/or receptors that play essential roles
   in cell growth and survival.[129] ^39 Consequently, Schiff‐fused
   chemical compounds, because of their wide scope of biological
   activities and synthetic applications, have been developed as a target
   heterocyclic framework in the research and development of medicinal
   chemistry. The SM series synthesis method (general procedure) is as
   follows:

   A mixture of isatin (2, 0.1 mM) and hydrazine hydrate (1.2 Equiv.) in
   methanol was refluxed for 3 h and cooled to room temperature (r.t.).
   The precipitate was filtered, washed with cold methanol, and dried at
   r.t. in the open air to give the hydrazine‐derivative compound 3 with a
   96% yield. A mixture of compound 3 (1.61 g, 10 mM) and the appropriate
   aromatic aldehyde (10 mM) in methanol (80 mL) and acetic acid (0.5 mL)
   was refluxed for 10 h. After standing at r.t. for 1 h, the product
   obtained was filtered and crystallized from methanol to generate 4a–e.

   1H‐indene‐1,3(2H)‐dione was used as raw material in the design and
   synthesis of Schiff base analogs via the Mannich reaction. These
   compounds are named the SMI series. The general procedure for synthesis
   is as follows:

   A mixture of 1,3‐Indanedione (5, 2 mM), methyl iodide (6 mM), and
   potassium fluoride (10 mM) in anhydrous acetonitrile (2 mL) was stirred
   at 70°C under nitrogen gas for 24 h. After the reaction was completed,
   the solution was cooled to r.t., and the mixture was extracted with
   dichloromethane. The organic layers were then pooled and dried over
   anhydrous Na[2]SO[4] before being concentrated under reduced pressure.
   The residue was purified by silica gel column chromatography to give
   the 2,2‐dimethylindan‐1,3‐dione (6, 84%). Subsequently, compound 6
   (174.1 mg, 1 mM) was added to a solution of appropriate amines (2.5 mM)
   and potassium fluoride (0.24 mL,10 mM) in anhydrous toluene (2 mL). Two
   drops of p‐toluenesulfonic acid were added to the solution with
   constant stirring. Then the reaction mixture allowed to proceed under
   reflux conditions for 8 h. After the reaction was completed, the
   mixture was cooled to r.t. via the addition of water and extraction
   with dichloromethane. The combined organics layers were dried over
   anhydrous Na[2]SO[4] and concentrated in vacuo. The residue was
   purified using silica column chromatography to obtain molecules.

4.4. Biolayer interferometry (BLI) assay

   The SMYD3‐GST tag protein (BPS Bioscience) was immobilized onto a GST
   biosensor (Fortebio). ZYZ384 was tested across the concentrations range
   of 6.25–200 µM. After baseline setup using Phosphate buffer saline
   (PBS), the biosensor tips were immersed in wells containing ZYZ384
   continuous diluent for 180 s to combine, followed by a 180‐s
   dissociation step. They calculated K [D] value with a 1:1 model in data
   analysis software 9.0 (Fortebio).

4.5. Molecular docking studies

   To determine molecule selectivity, the optimal binding mode and
   interaction of ZYZ384 with SMYD3 protein were assessed by molecular
   docking techniques using the CDOCKER module of Discovery Studio version
   4.5 (Dassault Systemes BIOVIA) software.[130] ^40 The structure of
   ZYZ384 was produced using Chemdraw software, and docking runs was
   performed and analyzed. A ranked list of the docked molecule to protein
   target was assessed. The ZYZ384 protein complex that best represented
   the docked complex was utilized for further analysis using the PyMol
   (Version 2.5.2) molecular visualization platform.[131] ^41

4.6. Molecular dynamic simulation study

   The stability of the ZYZ384 interaction and binding to the SMYD3 active
   site was determined by MD simulation with the 2019.3 package GROMACS
   version.[132] ^42 The amber99sb‐ildn force field was applied to protein
   parameters. Multiwfn was used to calculate ZYZ384 Electrostatic
   Potential[133] ^43 and parameterized with the General Amber Force Field
   (GAFF)[134] ^44 to allocate atom types and bonding parameters. Two
   sodium ions were added to neutralize the charge of the system, and the
   long‐range electrostatic interactions were calculated using the
   Particle Mesh Ewald (PME) method.[135] ^45 Proteins were solvated in a
   12‐layer cubic box with periodic boundary conditions using the TIP3P
   water model with a minimum distance of 1.0 Å from solute to box
   boundary. Prior to simulation, the two systems were first balanced
   through 0.1 ns in the normal volume and temperature (NVT), with normal
   pressure and temperature (NPT) gathering at 300 K and 1 atmosphere,
   using periodic boundary conditions to apply position constraints on
   proteins and small molecules. Subsequently, a 100 ns MD production run
   was performed in the NPT assembly using a Berendsen
   thermostat–barostat. The periodic boundary condition is used in the
   system, the motion equation is used together with the leap‐frog
   algorithm, and the time step is 2 fs. Various geometric properties of
   the system, such as RMSD and RMSF, were determined and graphed using
   the Xmgrace tool. The gmmpbsa method was applied to calculate binding
   free energy and energy decomposition.[136] ^46

4.7. SMYD3 homogeneous assay

   The commercial SMYD3 homogenous assay kit is designed to detect SMYD3
   activity for screening and analytical applications. The SMYD3
   homogenous assay kit is available in a convenient AlphaELISA format
   with GST‐labeled SMYD3 substrate, primary antibody, methylation
   detection buffer, and purification of SMYD3 for 384 enzyme reactions.
   The key to the SMYD3 homogenous assay is a particular antibody that
   recognizes methylated substrates. Using the kit, methyltransferase can
   be detected in three simple steps on a microtitration plate. First,
   samples containing the SMYD3 enzyme and ZYZ384 in different dosages
   were incubated in substrates for 3 h. Next, the receptor bead and
   primary antibody are added, the donor bead is added, and the Alpha
   counts are recorded.

4.8. Cell culture

   Human HepG2, A549, HCT116, MDA‐MB‐231, and Miapaca2 cells were
   purchased from American Type Culture Collection (ATCC) and cultured in
   DMEM medium supplemented with 1% penicillin/streptomycin and 10% FBS at
   37°C under 5% CO[2] atmosphere and 95% air. Trypsin EDTA was used for
   cell passage.

4.9. Cytotoxicity analysis

   Cell viability was assessed in several cancer cell lines and determined
   using the MTT assay in a 96‐well plate format. In brief, cells from
   100 mL medium (1 × 10^4/well) were inoculated in 96‐well plates and
   treated with ZYZ384. After 24 h of treatment, cell viability was
   determined by adding 1 mg/mL MTT‐containing medium for 4 h, followed by
   100 mL Dimethyl sulfoxide (DMSO) to dissolve formazan. The absorbance
   at 570 nm was recorded using a microplate ultraviolet/visible (UV/VIS)
   spectrophotometer (Tecan).

4.10. Cell transfection

   SMYD3 overexpression (SMYD3 OE) plasmid (Sino Biological) was mixed
   with lipofectamine 2000 for plasmid transfection. After transfection
   for 6 h, cells were exposed to the indicated ZYZ384 treatments for
   another 24 h.

4.11. Animal experiment

4.11.1. Animal care

   Six‐week‐old male nude mice and C57BL/6 mice were obtained from
   Chongqing Tengxin Biotechnology Co. LTD and free access to food and
   water in specific pathogen‐free (SPF) rooms with 25°C, 50% relative
   humidity, and 12‐h light /12‐h dark cycle.

4.11.2. Short‐term toxicity study of C57BL/6 mice

   C57BL/6 mice were randomly divided into two groups (n = 10) to
   determine the short‐term toxicity of ZYZ384. Mice were given the
   maximum dose of 2 g/kg orally administered ZYZ384 once, while the
   control group was treated with the same volume of sterilized water. The
   health conditions of all mice were monitored for 2 weeks.

4.11.3. The establishment and treatment of the tumor‐bearing HepG2 nude mice

   The tumor‐bearing nude mice model was established by direct injection
   of hepG2 cells into the left armpit of nude mice. Animals were
   subsequently divided into five groups (n = 7). All compounds were
   applied via intragastric administration, and sorafenib was used as the
   positive control. Other groups, except the control group, were treated
   with different concentrations of ZYZ384. Body weight and tumor volume
   were monitored and recorded throughout the treatment period.

4.12. RNA‐seq and ChIP‐seq analysis

4.12.1. RNA‐seq

   Sample preparation and sequencing: Control and ZYZ384 (100 mg/kg)
   groups of tumor tissue (n = 5) RNA integrities were assessed using the
   RNA Nano 6000 Bioanalyzer 2100 system (Agilent Technologies), Library
   preparation for transcriptome sequencing, Clustering, and sequencing on
   a cBot Cluster Generation System using TruSeq PE Cluster Kit v3‐cBot‐HS
   (Illumina).

   Data analysis: Quality control, the reference genome reads mapping,
   novel transcripts prediction, quantification of gene expression level,
   differential expression analysis, GO and KEGG enrichment analysis.

4.12.2. ChIP‐seq (H3K4me3)

   Sample preparation and sequencing: Control and ZYZ384 (100 mg/kg)
   groups of tumor tissue (n = 3). ChIP DNA degradation and contamination
   were monitored on agarose gels. DNA purity was checked using the
   NanoPhotometer® and spectrophotometer (IMPLEN). DNA concentration was
   measured using Qubit® DNA Assay Kit in Qubit® 3.0 Fluorometer (Life
   Technologies), and the purified DNA was used for ChIP‐seq library
   preparation. The library was constructed by Novogene Corporation.
   Subsequently, pair‐end sequencing of the sample was performed on the
   Illumina platform (Illumina).

   Data analysis: Quality control reads mapping to the reference genome,
   fragment size estimation, peak detection, motif analysis, peak
   annotation, and different peak analyses.

4.13. Quantitative real‐time polymerase chain reaction (qRT‐PCR) analysis

   Total tumor tissue mRNA was extracted using TRIzol reagent
   (Invitrogen), and NanoDrop™ spectrophotometer (Thermo Scientific)
   quantification. A Biotool cDNA Synthesis Kit (Biotool) produced cDNA.
   Target gene expression was measured by semiquantitative q‐PCR using
   iTaqTM Universal SYBR Green Supermix (Bio‐Rad) on ViiA™ 7 Real‐Time PCR
   System (Applied Biosystems). β‐Actin served as the reference gene. The
   2 − △△Ct method was used to analyze the relative target gene
   expression. The gene‐specific primer sequences used in this work are
   listed in Table [137]S2.

4.14. Western blot analysis

   Isolated protein (30 µg) was obtained from tissues, separated on a
   sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE)
   gel, and transferred to a polyvinylidene fluoride membrane. Membranes
   were then incubated in 5% Bovine Serum Albumin (BSA). The membranes
   were then probed with antibodies against SMYD3, H3, H3K4me3, P21,
   CyclinD1, CyclinD2, CDK1, Rac1, AKT, p‐AKT (308, 473), and GAPDH and
   incubated with horseradish peroxidase‐conjugated goat antirabbit or
   antimouse antibodies. The immunoreactive protein is visualized by
   enhanced chemiluminescence, and the signal intensity was detected and
   quantified using ImageJ.

   The antibodies Anti‐p21(ab109520), Anti‐Cyclin D1 (ab134175),
   Anti‐Cyclin D2(ab207604), Anti‐CDK1(ab131450), H3 (ab267372), H3k4me3
   (ab213224), Anti‐pan‐AKT (phospho T308)(ab8933), Anti‐AKT (phospho
   S473) (ab285140), pan‐AKT (ab18785), Anti‐Rac1(ab155938), and
   Anti‐SMYD3 (ab228015) are purchased from Abcam; SMYD3(C3): sc‐398085
   and GAPDH sc‐32233 are purchased from Santa Cruz.

4.15. Statistical analysis

   All experimental results except high‐throughput sequence data were
   expressed as mean ± SEM. All data were analyzed using GraphPad Prism
   software 8.0. One‐way ANOVA determined statistical significance between
   multiple groups. p value <0.05 was considered statistically
   significant.

AUTHOR CONTRIBUTIONS

   Qian Ding: Designed experiments. Qian Ding; Jianghong Cai; Li Jin; and
   Wu: Song performed experiments. Wei Hu and Wei Lei: Gave valuable
   suggestions on the research. Qian Ding: Analyzed data. Qian Ding and
   Jianghong Cai: Wrote the manuscript. Peter Rose: Checked the
   manuscript. Zhiyuan Tang; Yangyang Zhan; and Leilei Bao: Provided
   valuable opinions when revising the article and helped process the
   figures. Yi Zhun Zhu: Supervised this work. All the authors have read
   and approved the final manuscript.

CONFLICT OF INTEREST STATEMENT

   The authors declare no conflicts of interest.

ETHICS STATEMENT

   Special Approval number: MUST‐SA‐20240221001.

   Approved by Medical Ethics Committee Macau University of Science and
   Technology.

   Statements:

   In the work entitled “A novel small molecule ZYZ384 targeting SMYD3 for
   hepatocellular carcinoma via reducing H3K4 trimethylation of the Rac1
   promoter.” These carcinoma xenograft mice were originally planned to be
   treated with ZYZ384 for 30 days. During the experiment, we observed and
   supposed that the tumor volume may be too large on the 30th day, so we
   terminated the experiment on the 24th day of treatment in advance.
   After removing and measuring the tumors, we found some exceeded the
   limit. However, considering that ZYZ384 is a brand‐new structure that
   can reduce the growth of solid tumors, it is of great significance for
   future cancer treatment, to avoid repeated animal experiments causing
   more animal suffering, we apply special ethical approval for this
   experiment.

Supporting information

   Supporting Information
   [138]MCO2-5-e711-s001.docx^ (998.7KB, docx)

ACKNOWLEDGMENTS