Design, Synthesis and Biological Evaluation of Novel and Potent Protein Arginine Methyltransferases 5 Inhibitors for Cancer Therapy

Protein arginine methyltransferases 5 (PRMT5) is a clinically promising epigenetic target that is upregulated in a variety of tumors. Currently, there are several PRMT5 inhibitors under preclinical or clinical development, however the established clinical inhibitors show favorable toxicity. Thus, it remains an unmet need to discover novel and structurally diverse PRMT5 inhibitors with characterized therapeutic utility. Herein, a series of tetrahydroisoquinoline (THIQ) derivatives were designed and synthesized as PRMT5 inhibitors using GSK-3326595 as the lead compound. Among them, compound 20 (IC50: 4.2 nM) exhibits more potent PRMT5 inhibitory activity than GSK-3326595 (IC50: 9.2 nM). In addition, compound 20 shows high anti-proliferative effects on MV-4-11 and MDA-MB-468 tumor cells and low cytotoxicity on AML-12 hepatocytes. Furthermore, compound 20 possesses acceptable pharmacokinetic profiles and displays considerable in vivo antitumor efficacy in a MV-4-11 xenograft model. Taken together, compound 20 is an antitumor compound worthy of further study.

PRMT5 belongs to the predominant type II PRMTs, which is found in nearly all eukaryotic species [6]. The N-terminal domain of PRMT5 adopts a triosephosphate isomerase barrel structure, which forms a stable complex with methylosome protein 50 (MEP50). MEP50, as the most important interacting partner of PRMT5, is essential for the catalytic activity of PRMT5 [7]. A variety of intracellular substances was methylated by PRMT5, including histone H4 residue Arg3 (H4R3) and H3 residue Arg8(H3R8), which mediated various cellular processes such as transcriptional repression [8,9]. PRMT5 overexpression is involved in the proliferation and survival of numerous different cancers, including orectal, lung, ovarian, prostate, and pancreatic cancers, and lymphoma, leukemia, and glioblastoma [10][11][12][13][14][15]. Interestingly, it is reported that shPRMT5 or PRMT5 inhibitors motivated cGas/ STING and NLRC5 pathway and repressed MYC downstream genes, which provided possibility to the combination therapy of PD-1 and PRMT5 inhibitors [16][17][18]. Consequently, targeting PRMT5 is being pursued as a new cancer therapeutic strategy.
In the past few years, tremendous efforts from academia and pharmaceutical companies have been made to discover and develop PRMT5 inhibitors, and several PRMT5 inhibitors are currently in clinical trials ( Figure 1). As the active site of the enzyme consists of an SAM-binding pocket and a substrate-binding pocket, the mechanism of action of the clinical PRMT5 inhibitors can be classified into SAM-uncompetitive (GSK-3326595 [19], EPZ015666 [20]) and SAM-competitive (JNJ64619178 [21], PF06939999 [22]) inhibitors. In addition, there are some reports about PRMT5 allosteric inhibition, covalent inhibition and PROTACS [23][24][25]. GSK-3326595 from GlaxoSmithKline is an SAM-uncompetitive PRMT5 inhibitor, which has good brain permeability and anti-tumor effect, and is currently in clinical phase I/II (NCT04676516). Due to the lack of selectivity for tumor cells and normal cells, however, GSK-3326595 caused grade 3 or 4 adverse events in clinical studies. Therefore, the search for safe and effective PRMT5 inhibitors remains a pressing task.
In the present study, we present the design, synthesis and biological evaluation of a series of PRMT5 inhibitors based on the co-crystal structure of EPZ015666 and PRMT5. Among them, compound 20 displays high inhibitory activity both in enzymatic and cellular assays, which is better than GSK-3326595. Furthermore, compound 20 demonstrates potent antitumor in vivo efficacy, making further progress as a potent small molecule inhibitor targeting PRMT5.

Structure-based Design of Novel PRMT5 Inhibitors
Our approach to design PRMT5 inhibitors took inspiration from the co-crystal structure of EPZ015666 and PRMT5 protein (PDB code: 4X61) ( Figure 2). The computer modeling displays a key cation-π interaction between the benzene ring of THIQ and the cofactor SAM, helping to maintain high binding affinity and selectivity for PRMT5. Notably, In the present study, we present the design, synthesis and biological evaluation of a series of PRMT5 inhibitors based on the co-crystal structure of EPZ015666 and PRMT5. Among them, compound 20 displays high inhibitory activity both in enzymatic and cellular assays, which is better than GSK-3326595. Furthermore, compound 20 demonstrates potent antitumor in vivo efficacy, making further progress as a potent small molecule inhibitor targeting PRMT5.
THIQ also forms a potential π-π stacking interaction with Phe327 and the tertiary nitrogen atom of the THIQ ring system forms a key hydrogen bond, with Glu435 and Leu437 mediated by water molecule. We also noticed that the aminopyrimidine region is situated at the hydrophobic site of the substrate-binding pocket, surrounded by Tyr304, Val326, Phe327, Phe577 and Phe580. Therefore, increasing the volume of the aromatic ring in this area might be advantageous to strengthening the π-π or/and hydrophobic interaction with residues in the active site, thereby enhancing the binding affinity between the compound and PRMT5.

Evaluation of PRMT5 Enzymatic Activities
We initially attempted replacing 1-(4-(pyrimidin-4-ylamino) piperidin-1-yl) ethan-1one with the carbazole ring. Using a radioactivity-based assay monitoring the transfer of the methyl group from 3 H-SAM to peptide substrate, we found that compound 1 displays similar PRMT5 inhibitory activity compared with GSK-3326595. However, the structureactivity study of R 1 and R 2 did not significantly improve the activity (Table 1). It is noteworthy that the inhibitory activity could be maintained when the carbazole moiety was replaced by less rigid tetrahydrocarbazole ring, with an IC50 of 18.6 ± 0.9 nM (compound 7). No dramatic change is observed in inhibitory activity upon increasing or decreasing the size of ring A. The substituent R 1 on the ring A were further investigated, and unfortunately, when R 1 is methyl or ethyl, the PRMT5 inhibitory activity of the compound 11 and 12 decreases slightly (Table 2).

Evaluation of PRMT5 Enzymatic Activities
We initially attempted replacing 1-(4-(pyrimidin-4-ylamino) piperidin-1-yl) ethan-1one with the carbazole ring. Using a radioactivity-based assay monitoring the transfer of the methyl group from 3 H-SAM to peptide substrate, we found that compound 1 displays similar PRMT5 inhibitory activity compared with GSK-3326595. However, the structureactivity study of R 1 and R 2 did not significantly improve the activity (Table 1). THIQ also forms a potential π-π stacking interaction with Phe327 and the tertiary nitrogen atom of the THIQ ring system forms a key hydrogen bond, with Glu435 and Leu437 mediated by water molecule. We also noticed that the aminopyrimidine region is situated at the hydrophobic site of the substrate-binding pocket, surrounded by Tyr304, Val326, Phe327, Phe577 and Phe580. Therefore, increasing the volume of the aromatic ring in this area might be advantageous to strengthening the π-π or/and hydrophobic interaction with residues in the active site, thereby enhancing the binding affinity between the compound and PRMT5.

Evaluation of PRMT5 Enzymatic Activities
We initially attempted replacing 1-(4-(pyrimidin-4-ylamino) piperidin-1-yl) ethan-1one with the carbazole ring. Using a radioactivity-based assay monitoring the transfer of the methyl group from 3 H-SAM to peptide substrate, we found that compound 1 displays similar PRMT5 inhibitory activity compared with GSK-3326595. However, the structureactivity study of R 1 and R 2 did not significantly improve the activity (Table 1). It is noteworthy that the inhibitory activity could be maintained when the carbazole moiety was replaced by less rigid tetrahydrocarbazole ring, with an IC50 of 18.6 ± 0.9 nM (compound 7). No dramatic change is observed in inhibitory activity upon increasing or decreasing the size of ring A. The substituent R 1 on the ring A were further investigated, and unfortunately, when R 1 is methyl or ethyl, the PRMT5 inhibitory activity of the compound 11 and 12 decreases slightly ( It is noteworthy that the inhibitory activity could be maintained when the carbazole moiety was replaced by less rigid tetrahydrocarbazole ring, with an IC 50 of 18.6 ± 0.9 nM (compound 7). No dramatic change is observed in inhibitory activity upon increasing or decreasing the size of ring A. The substituent R 1 on the ring A were further investigated, and unfortunately, when R 1 is methyl or ethyl, the PRMT5 inhibitory activity of the compound 11 and 12 decreases slightly (Table 2).  Then we explored the effects of substituent R 2 on PRMT5 inhibition and expected to enhance the interaction between the pocket defined by Thr323, Val326 and Phe327 and target compounds by introducing different hydrophobic groups (13)(14)(15)(16)(17)(18)(19)(20). Surprisingly, compounds with the bulkier n-propyl, cyclobutyl and oxetane groups yield excellent PRMT5 inhibitory activities, and compounds 15 and 20 display the most potent inhibitory activity on PRMT5 with IC50 of 4.2 nM. Modeling of the binding pose of 20 was performed by manual ligand building from the crystal structure of the complex of PRMT5 and EPZ-015666 (PDB code: 4X61)( Figure 3). The result shows that the 2,3,4,9-tetrahydro-1H-carbazole fragment, which mimics the N-(oxetan-3-yl) pyrimidin-4-amine moiety in EPZ015666, is situated at the hydrophobic site of the substrate-binding pocket, and the oxetan-3-ylmethyl moiety is extended to the small hydrophobic region and undergoes a hydrogen-bonding interaction with Thr323. Finally, we investigated the effects of the tetrahydropyrido [4,3-b] indole moiety and its substituent R 1 on PRMT5 inhibition. However, whether the tetrahydrocarbazole moiety is replaced by tetrahydropyrido [4,3-b]indole or substituent R 1 is replaced by alkyl, the outcome is unfavorable for PRMT5 inhibitory activity.  Then we explored the effects of substituent R 2 on PRMT5 inhibition and expected to enhance the interaction between the pocket defined by Thr323, Val326 and Phe327 and target compounds by introducing different hydrophobic groups (13)(14)(15)(16)(17)(18)(19)(20). Surprisingly, compounds with the bulkier n-propyl, cyclobutyl and oxetane groups yield excellent PRMT5 inhibitory activities, and compounds 15 and 20 display the most potent inhibitory activity on PRMT5 with IC50 of 4.2 nM. Modeling of the binding pose of 20 was performed by manual ligand building from the crystal structure of the complex of PRMT5 and EPZ-015666 (PDB code: 4X61)( Figure 3). The result shows that the 2,3,4,9-tetrahydro-1H-carbazole fragment, which mimics the N-(oxetan-3-yl) pyrimidin-4-amine moiety in EPZ015666, is situated at the hydrophobic site of the substrate-binding pocket, and the oxetan-3-ylmethyl moiety is extended to the small hydrophobic region and undergoes a hydrogen-bonding interaction with Thr323. Finally, we investigated the effects of the tetrahydropyrido [4,3-b] indole moiety and its substituent R 1 on PRMT5 inhibition. However, whether the tetrahydrocarbazole moiety is replaced by tetrahydropyrido [4,3-b]indole or substituent R 1 is replaced by alkyl, the outcome is unfavorable for PRMT5 inhibitory activity.  Then we explored the effects of substituent R 2 on PRMT5 inhibition and expected to enhance the interaction between the pocket defined by Thr323, Val326 and Phe327 and target compounds by introducing different hydrophobic groups (13)(14)(15)(16)(17)(18)(19)(20). Surprisingly, compounds with the bulkier n-propyl, cyclobutyl and oxetane groups yield excellent PRMT5 inhibitory activities, and compounds 15 and 20 display the most potent inhibitory activity on PRMT5 with IC50 of 4.2 nM. Modeling of the binding pose of 20 was performed by manual ligand building from the crystal structure of the complex of PRMT5 and EPZ-015666 (PDB code: 4X61)( Figure 3). The result shows that the 2,3,4,9-tetrahydro-1H-carbazole fragment, which mimics the N-(oxetan-3-yl) pyrimidin-4-amine moiety in EPZ015666, is situated at the hydrophobic site of the substrate-binding pocket, and the oxetan-3-ylmethyl moiety is extended to the small hydrophobic region and undergoes a hydrogen-bonding interaction with Thr323. Finally, we investigated the effects of the tetrahydropyrido [4,3-b] indole moiety and its substituent R 1 on PRMT5 inhibition. However, whether the tetrahydrocarbazole moiety is replaced by tetrahydropyrido [4,3-b]indole or substituent R 1 is replaced by alkyl, the outcome is unfavorable for PRMT5 inhibitory activity. Then we explored the effects of substituent R 2 on PRMT5 inhibition and expected to enhance the interaction between the pocket defined by Thr323, Val326 and Phe327 and target compounds by introducing different hydrophobic groups (13)(14)(15)(16)(17)(18)(19)(20). Surprisingly, compounds with the bulkier n-propyl, cyclobutyl and oxetane groups yield excellent PRMT5 inhibitory activities, and compounds 15 and 20 display the most potent inhibitory activity on PRMT5 with IC50 of 4.2 nM. Modeling of the binding pose of 20 was performed by manual ligand building from the crystal structure of the complex of PRMT5 and EPZ-015666 (PDB code: 4X61)( Figure 3). The result shows that the 2,3,4,9-tetrahydro-1H-carbazole fragment, which mimics the N-(oxetan-3-yl) pyrimidin-4-amine moiety in EPZ015666, is situated at the hydrophobic site of the substrate-binding pocket, and the oxetan-3-ylmethyl moiety is extended to the small hydrophobic region and undergoes a hydrogen-bonding interaction with Thr323. Finally, we investigated the effects of the tetrahydropyrido [4,3-b] indole moiety and its substituent R 1 on PRMT5 inhibition. However, whether the tetrahydrocarbazole moiety is replaced by tetrahydropyrido [4,3-b]indole or substituent R 1 is replaced by alkyl, the outcome is unfavorable for PRMT5 inhibitory activity. Then we explored the effects of substituent R 2 on PRMT5 inhibition and expected to enhance the interaction between the pocket defined by Thr323, Val326 and Phe327 and target compounds by introducing different hydrophobic groups (13)(14)(15)(16)(17)(18)(19)(20). Surprisingly, compounds with the bulkier n-propyl, cyclobutyl and oxetane groups yield excellent PRMT5 inhibitory activities, and compounds 15 and 20 display the most potent inhibitory activity on PRMT5 with IC50 of 4.2 nM. Modeling of the binding pose of 20 was performed by manual ligand building from the crystal structure of the complex of PRMT5 and EPZ-015666 (PDB code: 4X61)( Figure 3). The result shows that the 2,3,4,9-tetrahydro-1H-carbazole fragment, which mimics the N-(oxetan-3-yl) pyrimidin-4-amine moiety in EPZ015666, is situated at the hydrophobic site of the substrate-binding pocket, and the oxetan-3-ylmethyl moiety is extended to the small hydrophobic region and undergoes a hydrogen-bonding interaction with Thr323. Finally, we investigated the effects of the tetrahydropyrido [4,3-b] indole moiety and its substituent R 1 on PRMT5 inhibition. However, whether the tetrahydrocarbazole moiety is replaced by tetrahydropyrido [4,3-b]indole or substituent R 1 is replaced by alkyl, the outcome is unfavorable for PRMT5 inhibitory activity.
Then we explored the effects of substituent R 2 on PRMT5 inhibition and expected to enhance the interaction between the pocket defined by Thr323, Val326 and Phe327 and target compounds by introducing different hydrophobic groups (13)(14)(15)(16)(17)(18)(19)(20). Surprisingly, compounds with the bulkier n-propyl, cyclobutyl and oxetane groups yield excellent PRMT5 inhibitory activities, and compounds 15 and 20 display the most potent inhibitory activity on PRMT5 with IC5 0 of 4.2 nM. Modeling of the binding pose of 20 was performed by manual ligand building from the crystal structure of the complex of PRMT5 and EPZ-015666 (PDB code: 4X61) (Figure 3). The result shows that the 2,3,4,9-tetrahydro-1H-carbazole fragment, which mimics the N-(oxetan-3-yl) pyrimidin-4-amine moiety in EPZ015666, is situated at the hydrophobic site of the substrate-binding pocket, and the oxetan-3-ylmethyl moiety is extended to the small hydrophobic region and undergoes a hydrogen-bonding interaction with Thr323. Finally, we investigated the effects of the tetrahydropyrido [4,3-b] indole moiety and its substituent R 1 on PRMT5 inhibition. However, whether the tetrahydrocarbazole moiety is replaced by tetrahydropyrido [4,3-b]indole or substituent R 1 is replaced by alkyl, the outcome is unfavorable for PRMT5 inhibitory activity.

In Vitro Anti-Proliferative Activity Evaluation
MV-4-11 and MDA-MB-468 cell lines were used to evaluate the anti-proliferative activities of the selected target compounds (CellTiter-Glo assay), and AML-12 hepatocytes were selected to evaluate the toxicity of the compounds. As shown in Table 3, compound 20 exhibits stronger anti-proliferative activities than GSK-3326595 in MV-4-11 and MDA-MB-468 cell lines. however, it has almost no toxicity compared to normal AML-12 hepatocytes.

Cellular Thermal Shift Assay (CETSA)
To confirm the target engagement in cells, compound 20 and GSK-3326595 were selected and evaluated using an MV-4-11 cell line by CETSA [26,27]. As showed in Figure 4, we measured the PRMT5 melting curve and facilitated the temperature selection, as the PRMT5 melting point of GSK-3326595 and compound 20 increased by 5.5°C and 7.2°C compared to the DMSO vehicle, indicating the thermal stabilization of proteins upon ligand binding. Moreover, the binding of 20 also inhibits the degradation of PRMT5 at 58°C in a dose-dependent manner. All the results demonstrate better binding potency in a cellular context.

In Vitro Anti-Proliferative Activity Evaluation
MV-4-11 and MDA-MB-468 cell lines were used to evaluate the anti-proliferative activities of the selected target compounds (CellTiter-Glo assay), and AML-12 hepatocytes were selected to evaluate the toxicity of the compounds. As shown in Table 3, compound 20 exhibits stronger anti-proliferative activities than GSK-3326595 in MV-4-11 and MDA-MB-468 cell lines. however, it has almost no toxicity compared to normal AML-12 hepatocytes.

Cellular Thermal Shift Assay (CETSA)
To confirm the target engagement in cells, compound 20 and GSK-3326595 were selected and evaluated using an MV-4-11 cell line by CETSA [26,27]. As showed in Figure 4, we measured the PRMT5 melting curve and facilitated the temperature selection, as the PRMT5 melting point of GSK-3326595 and compound 20 increased by 5.5 • C and 7.2 • C compared to the DMSO vehicle, indicating the thermal stabilization of proteins upon ligand binding. Moreover, the binding of 20 also inhibits the degradation of PRMT5 at 58 • C in a dose-dependent manner. All the results demonstrate better binding potency in a cellular context.

Characterization of Cell Methylation and Proliferation
To further evaluate if compound 20 exhibits anti-proliferative activity depend PRMT5 inhibition, we analyzed the effect of compound 20 on PRMT5 substrate protein expression in MV-4-11 cell, with high expression in PRMT5. As shown in 5, the treatment of compound 20 leads to a concentration-dependent decrease in s indicating its better PRMT5 inhibition than GSK-3326595. All the cellular data su that compound 20 inhibits cell proliferation through inhibiting PRMT5 activity.

Characterization of Cell Methylation and Proliferation
To further evaluate if compound 20 exhibits anti-proliferative activity depending on PRMT5 inhibition, we analyzed the effect of compound 20 on PRMT5 substrate sDMA protein expression in MV-4-11 cell, with high expression in PRMT5. As shown in Figure 5, the treatment of compound 20 leads to a concentration-dependent decrease in sDMA, indicating its better PRMT5 inhibition than GSK-3326595. All the cellular data suggests that compound 20 inhibits cell proliferation through inhibiting PRMT5 activity.

Characterization of Cell Methylation and Proliferation
To further evaluate if compound 20 exhibits anti-proliferative activity depending on PRMT5 inhibition, we analyzed the effect of compound 20 on PRMT5 substrate sDMA protein expression in MV-4-11 cell, with high expression in PRMT5. As shown in Figure  5, the treatment of compound 20 leads to a concentration-dependent decrease in sDMA, indicating its better PRMT5 inhibition than GSK-3326595. All the cellular data suggests that compound 20 inhibits cell proliferation through inhibiting PRMT5 activity.

In Vivo Pharmacokinetic Study
Considering the desirable enzymatic potency, as well as the potent anti-proliferative activity in several cancer cell lines, compound 20 was selected and evaluated for its pharmacokinetics (PK) data in male ICR mice at 2 mg/kg and 10 mg/kg for intravenous (iv) and oral (po) administration, respectively. The resulting data (Tables S1 and S2) are summarized in Table 4. It reveals that compound 20 exhibits an acceptable PK profile with a T 1/2 of 6.06 h, a modest drug exposure in blood (AUC 0−∞ values after oral treatment of 10 mg/kg are 747 h·ng/mL) and an acceptable oral bioavailability of 14.5%. * n = 3. The compound was formulated as solution in 20% HP-β-CD in saline. Dosage: 2 mg/kg for intravenous (iv) and 10 mg/kg for oral (po)administration. ICR mouse were used. The maximum drug concentration (C max ) was observed at t = 5 min, the first sampling time point after iv administration.

Antitumor Effect of 20 In Vivo
Based on the excellent enzymatic and anti-proliferative activities of compound 20 in vitro, we then evaluated antitumor activities in vivo in an MV-4-11 xenograft mouse model. GSK-3326595 and compound 20 were administered by intraperitoneal injection twice daily (BID) for 28 consecutive days. As shown in Figure 6, compound 20 demonstrates its potent antitumor efficacy at a dose of 10 mg·kg −1 , and the tumor suppression effect of 20 (TGI: 47.6%) is more effective than that of the positive control GSK-3326595 (TGI: 39.3%). It is noteworthy that no significant weight fluctuations were observed, and compound 20 was well-tolerated with no mortality.

In Vivo Pharmacokinetic Study
Considering the desirable enzymatic potency, as well as the potent anti-proliferative activity in several cancer cell lines, compound 20 was selected and evaluated for its pharmacokinetics (PK) data in male ICR mice at 2 mg/kg and 10 mg/kg for intravenous (iv) and oral (po) administration, respectively. The resulting data (Tables S1 and S2) are summarized in Table 4. It reveals that compound 20 exhibits an acceptable PK profile with a T1/2 of 6.06 h, a modest drug exposure in blood (AUC0−∞ values after oral treatment of 10 mg/kg are 747 h.ng/mL) and an acceptable oral bioavailability of 14.5%. Table 4. Pharmacokinetics (PK) of Compound 20*.

Antitumor Effect of 20 In Vivo
Based on the excellent enzymatic and anti-proliferative activities of compound 20 in vitro, we then evaluated antitumor activities in vivo in an MV-4-11 xenograft mouse model. GSK-3326595 and compound 20 were administered by intraperitoneal injection twice daily (BID) for 28 consecutive days. As shown in Figure 6, compound 20 demonstrates its potent antitumor efficacy at a dose of 10 mg·kg⁻¹, and the tumor suppression effect of 20 (TGI: 47.6%) is more effective than that of the positive control GSK-3326595 (TGI: 39.3%). It is noteworthy that no significant weight fluctuations were observed, and compound 20 was well-tolerated with no mortality.  The expression of Ki-67 is positively correlated to the malignancy of tumors [28]. H&E staining shows the malignancy of the tumor and drug prognostic effect [29]. Therefore, H&E staining and Ki-67 immunohistochemistry were further used to gain insight into the changes in the structure and appearance of cancerous cellular structures after compound 20 administration. As shown in Figure 7, all administration groups cause a decrease in the population of tumor cells, and morphological features including cell shrinkage, condensation and margination of nuclear chromatin are observed by microscopy. Meanwhile, the administration group also down-regulates the expression of Ki-67, indicating the decrease in cell proliferation. The expression of Ki-67 is positively correlated to the malignancy of tumors [28]. H&E staining shows the malignancy of the tumor and drug prognostic effect [29]. Therefore, H&E staining and Ki-67 immunohistochemistry were further used to gain insight into the changes in the structure and appearance of cancerous cellular structures after compound 20 administration. As shown in Figure 7, all administration groups cause a decrease in the population of tumor cells, and morphological features including cell shrinkage, condensation and margination of nuclear chromatin are observed by microscopy. Meanwhile, the administration group also down-regulates the expression of Ki-67, indicating the decrease in cell proliferation.

Expression of sDMA In Vivo
Subsequently, we investigated the effects of compound 20 on the expression of the PRMT5 substrate sDMA in vivo (Figure 8). Similarly, the expression of sDMA was decreased after the treatment of compound 20, which was in line with its antitumor activity in vivo.

Expression of sDMA In Vivo
Subsequently, we investigated the effects of compound 20 on the expression of the PRMT5 substrate sDMA in vivo (Figure 8). Similarly, the expression of sDMA was decreased after the treatment of compound 20, which was in line with its antitumor activity in vivo. The expression of Ki-67 is positively correlated to the malignancy of tumors [28]. H&E staining shows the malignancy of the tumor and drug prognostic effect [29]. Therefore, H&E staining and Ki-67 immunohistochemistry were further used to gain insight into the changes in the structure and appearance of cancerous cellular structures after compound 20 administration. As shown in Figure 7, all administration groups cause a decrease in the population of tumor cells, and morphological features including cell shrinkage, condensation and margination of nuclear chromatin are observed by microscopy. Meanwhile, the administration group also down-regulates the expression of Ki-67, indicating the decrease in cell proliferation.

Expression of sDMA In Vivo
Subsequently, we investigated the effects of compound 20 on the expression of the PRMT5 substrate sDMA in vivo (Figure 8). Similarly, the expression of sDMA was decreased after the treatment of compound 20, which was in line with its antitumor activity in vivo.  The data are shown as the mean ± SD of three independent experiments. * p < 0.05 vs. control.
sDMA as indicated in MV-4-11 xenograft tumors treated with the indicated compounds for 28 days (B) The relative strength of sDMA signal on the Western blots was measured by densitometry. The data are shown as the mean ± SD of three independent experiments. * p < 0.05 vs. control.

Chemistry
The preparation of compounds 1-6 was described in Scheme 1. Initially, the commercially available 1,2,3,4-tetrahydroisoquinoline was substituted by (R) epichlorohydrin to afford compound 28, which was ring-opened by ammonolysis o ammonia water to give compound 29. Meanwhile, compounds 30a-30e were reacted with ethyl 4-bromobenzoate to afford compounds 31a-31e, which were catalyzed by Pd(OAc) to afford compounds 32a-32e. 32a reacted with MeI by nucleophilic substitution to afford compound 33. Subsequently, 32a-32e and 33 were hydrolyzed by KOH to obtain compounds 34a-34f. As a result, compounds 1-6 were synthesized from carboxylic acid 34a-34f and the intermediate 29. The synthetic routes of compounds 7-26 were shown in Scheme 2. 35a-35l were re acted with p-carboxyphenylhydrazine to afford compounds 36a-36l, which were subse quently reacted with MeOH in the presence of H2SO4 to give compound 37. R 2 was intro duced by nucleophilic substitution using halogenated alkanes to yield compounds 38m-38t, which were hydrolyzed by LiOH to afford compounds 39m-39t. Finally, the targe compounds 7-26 were prepared from the corresponding carboxylic acid and the key in termediates 29. The synthetic routes of compounds 7-26 were shown in Scheme 2. 35a-35l were reacted with p-carboxyphenylhydrazine to afford compounds 36a-36l, which were subsequently reacted with MeOH in the presence of H 2 SO 4 to give compound 37. R 2 was introduced by nucleophilic substitution using halogenated alkanes to yield compounds 38m-38t, which were hydrolyzed by LiOH to afford compounds 39m-39t. Finally, the target compounds 7-26 were prepared from the corresponding carboxylic acid and the key intermediates 29.

General Chemistry
All solvents and reagents were purchased from commercial vendors and used without further purification. The progress of the reaction was monitored by TLC on pre-coated silica gel plates (silica gel 60 F254), and spots were observed by UV light (λmax = 254 or 365 nm) or other suitable stain. 1 H (300 or 400 MHz) and 13 C NMR (75 MHz or 101 MHz) spectra were recorded by Bruker spectrometer with TMS as internal standard. Coupling constants (J) are expressed in hertz (Hz). Ordinary and high-resolution mass spectra were obtained by ESI-MS. The purification of compounds was conducted by flash column chromatography (silica gel 200-300 mesh). The purity of all target compounds was > 95%, as determined by HPLC (BDS Hypersil C18, λ = 254 nm).

General Chemistry
All solvents and reagents were purchased from commercial vendors and used without further purification. The progress of the reaction was monitored by TLC on pre-coated silica gel plates (silica gel 60 F254), and spots were observed by UV light (λ max = 254 or 365 nm) or other suitable stain. 1 H (300 or 400 MHz) and 13 C NMR (75 MHz or 101 MHz) spectra were recorded by Bruker spectrometer with TMS as internal standard. Coupling constants (J) are expressed in hertz (Hz). Ordinary and high-resolution mass spectra were obtained by ESI-MS. The purification of compounds was conducted by flash column chromatography (silica gel 200-300 mesh). The purity of all target compounds was > 95%, as determined by HPLC (BDS Hypersil C18, λ = 254 nm).  (29). A mixture of 28 (1.00 g, 26.0 mmol) in NH 4 OH (10 mL) was refluxed at 80 • C for 3 h. The reaction mixture was allowed to cool to 25 • C and concentrated under reduced pressure. The crude product was purified using column chromatography (DCM/MeOH: 15/1) to obtain the titled compound (0.37 g, 34.0%) as a pale-yellow oil; 1  A mixture of 30a-e (2.80 mmol, 1 eq), ethyl 4-bromobenzoate (2.80 mmol, 1 eq), Pd(OAc)2 (0.14 mmol, 0.05 eq), BINAP (0.14 mmol, 0.05 eq) and K 2 CO 3 (3.0 mmol, 1.07 eq) in toluene (10 mL) was refluxed under an atmosphere of N 2 for 6 h. The reaction mixture was allowed to cool to rt and concentrated under reduced pressure. The crude product was purified using a column chromatography cation to obtain compound 31a-e.

General Procedure for the Synthesis of Compound 32a-e
A mixture of 31a-e (1.24 mmol, 1 eq), Pd(OAc) 2 (1.37 mmol, 1 eq) in AcOH (8 mL) was refluxed under an atmosphere of N 2 for 1 h. The reaction mixture was allowed to cool to rt and concentrated under reduced pressure. The crude product was purified using a column chromatography cation to obtain compound 32a-e.

General Procedure for the Synthesis of Compounds 34a-f
A mixture of 32a-e or 33 (0.38 mmol, 1 eq), KOH (1.13 mmol, 3 eq) in EtOH (5 mL) and H 2 O (1 mL) was refluxed under an atmosphere of N 2 for 3 h. The reaction mixture was allowed to cool to rt and concentrated under reduced pressure. 5 mL H 2 O was added, 1 mL 2N (HCL) was added to adjust the pH to about 3, and suction filtration to obtain the compounds 34a-f. A mixture of 35a-l (6.49 mmol, 1.3 eq), 4-hydrazinylbenzoic acid (5.00 mmol, 1 eq) in 10% H 2 SO 4 (20 mL) was refluxed under an atmosphere of N 2 for 3 h. The reaction mixture was allowed to cool to rt the solid was precipitated and filtrated, and the filter cake with 10 mL of H 2 O, and dried to obtain the compound 36a-l.

In Vivo Xenograft Research
Male BALB/c nude mice (6-8 weeks old) were purchased from Hangzhou Medical College. All animals were housed in a specific pathogen-free facility. MV-4-11 cells (5 × 10 6 ) suspended in 0.2 mL of PBS were inoculated subcutaneously on the right flank of each BALB/c nude mouse. Mice were divided randomly (5 mice for each group) into two treatment groups and a control group when the size of the tumors reached 150 mm 3 . GSK-3326595 (10 mg/kg, dissolved in DMSO/10% PEG300/20% β-cyclodextrin/7%), compound 20 (10 mg/kg, DMSO/10% PEG300/20% β-cyclodextrin/7%) and vehicle (DMSO/10% PEG300/20% β-cyclodextrin/7%) were administered twice per day for 28 days by intraperitoneal administration. The tumor volume and body weight were measured every 2 days. The tumor volume was determined with Vernier calipers and calculated as follows: tumor volume = a × b 2 /2, where a and b stand for the longest and shortest diameters measured by the vernier caliper, respectively. All animal experiments have been approved by Pharmaceutical Experiment Center of China Pharmaceutical University.

H&E Staining
The tumor tissue samples of xenograft model mice were fixed with 4% paraformaldehyde, dehydrated with ethanol, immersed in xylene, embedded in paraffin and cut into 4.0 µm longitudinal sections. The paraffin-embedded sections were stained with H&E according to the manufacturer's instructions (Beyotime, Shanghai, China). Each group of samples was observed with a DM6B-positive fluorescence microscope (Leica, Frankfurt, Germany). Five images were randomly captured per slide.

Immunohistochemical Staining
Tumors of xenograft model mice were embedded in paraffin, incubated with 0.3% hydrogen peroxide to block endogenous peroxidase, blocked with 1.0% BSA, stained with the primary antibody, incubated with secondary antibody and counterstained with hematoxylin. Each group was examined using the DM6B-positive fluorescence microscope (Leica, Germany). Five images were randomly captured per slide. The percentage of stained dots were analyzed by Image J software.

Statistical Analysis
Data were analyzed using Prism 8.0. Results were expressed as means ± SD. Differences between treatment regimens were analyzed by one way ANOVA. p < 0.05 was considered statistical significance.

Molecular Modelling
The protein crystallized was downloaded from the PDB (PDB code: 4X61). Protein Preparation Wizard of the Schrodinger Suite was used to prepare protein structure to ensure that the downloaded X-ray structure was reliable and qualitatively considerable for further in silico studies. Small molecules were prepared using LigPrep [30] prior to docking simulation to obtain the accessible least-energy ionized conformer and tautomeric states of the ligands. The binding site was defined by a box centered on the centroid of the crystal ligand and with similar size to the ligand. The Glide implemented in Schrodinger 2009 was used for molecular docking [31]. The standard precision (SP) mode was used for the docking and scoring. All other parameters were kept default. The best pose was output on the basis of Glide score and the protein-ligand interactions. PyMOL software was employed to depict structural representations [32].

Conclusions
In this work, we designed and synthesized a series of THIQ derivatives. Among them, compound 20 has potent PRMT5 inhibitory activity, with the IC 50 value of 4.2 nM. In addition, compound 20 also dramatically suppresses the growth of MV-4-11 and MDA-MB-468 cell lines with no significant cytotoxicity in normal hepatocyte AML-12 cell lines. In MV-4-11 cell-derived xenograft model, compound 20 (TGI = 47.6%) at a dose of 10 mg/kg (iv) displays more effective antitumor effects than GSK-3326595 (TGI = 39.3%) at the same dose (iv). The preliminary mechanism studies in vivo confirm that 20 exerts antitumor effect via decreasing the expression of sDMA in tumor tissues. Encouraging to researchers owing to the excellent properties in vitro and in vivo, 20 as a novel PRMT5 inhibitor is worthy of further study.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27196637/s1. Figure S1: Plasma concentration curve after administration of 20 (2 mg/kg, iv). The value at each time point represents. Figure S2: Plasma concentration curve after administration of 20 (10 mg/kg, ig). The value at each time point represents. Table S1: intravenous injection of 2 mg/kg compound 20 in male mice pharmacokinetic parameters. Table S2: gavage of 10 mg/kg compound 20 in male mice pharmacokinetic parameters.