Enlibio,Your biological research partner
Xinqidi Biotech Co.,Ltd,Wuhan,China 2008-2022
R&D 14th year

Targeting dual-specificity tyrosine phosphorylation-regulated kinase 2 with a highly selective inhib

Issuing time:2022-05-31 15:28

Abstract

Prostate cancer (PCa) is one of the most prevalent cancers in men worldwide, and hormonal therapy plays a key role in the treatment of PCa. However, the drug resistance of hormonal therapy makes it urgent and necessary to identify novel targets for PCa treatment. Herein, dual-specificity tyrosine phosphorylation-regulated kinase 2 (DYRK2) is found and confirmed to be highly expressed in the PCa tissues and cells, and knock-down of DYRK2 remarkably reduces PCa burden in vitro and in vivo. On the base of DYRK2 acting as a promising target, we further discover a highly selective DYRK2 inhibitor YK-2-69, which specifically interacts with Lys-231 and Lys-234 in the co-crystal structure. Especially, YK-2-69 exhibits more potent anti-PCa efficacy than the first-line drug enzalutamide in vivo. Meanwhile, YK-2-69 displays favorable safety properties with a maximal tolerable dose of more than 10,000 mg/kg and pharmacokinetic profiles with 56% bioavailability. In summary, we identify DYRK2 as a potential drug target and verify its critical roles in PCa. Meanwhile, we discover a highly selective DYRK2 inhibitor with favorable druggability for the treatment of PCa.

Introduction

Prostate cancer (PCa) is one of the most common cancers in men with an estimated incidence of 268,490 new cases in the U.S. alone, accounting for 27% of all new cancer cases in U.S. males in 20221,2. The mortality rate of PCa ranks second in U.S. male cancer and is expected to reach 11% of all male cancer deaths in 20221,2. The majority of PCa primarily relies on androgens for survival and growth, and PCa treatment mainly focused on reducing hormone levels3,4,5. Currently, hormonal therapy, including the antiandrogens6,7, gonadotropin-releasing hormone (GnRH) agonists and antagonists8,9, androgen biosynthesis inhibitor10,11, and androgen receptor inhibitor12,13, has been developed. However, hormonal therapy can only delay PCa progression but is not a curative method for PCa treatment14,15. Ultimately, drug resistance of hormonal therapy is easily generated and most PCa patients developed to a metastatic, hormone-resistant state16,17. Therefore, despite the leading role of androgen in the PCa, it is necessary to identify novel regulators of PCa and target them for PCa treatment.

Dual-specificity tyrosine phosphorylation-regulated kinases (DYRKs) are evolutionarily conserved enzymes, and the most remarkable characteristic property of them is to display both tyrosine and serine/threonine kinase activities. DYRKs can be divided into class I (DYRK1A and DYRK1B) and class II (DYRK2, DYRK3 and DYRK4) in mammals. Class I and class II DYRKs are mainly localized in the nucleus and cytoplasm, respectively. Among class II DYRKs, DYRK2 is most deeply studied, which plays important while controversial roles in human cancers18,19. Previously, DYRK2 was primarily considered as a cancer suppressor, which can promote phosphorylation of P53 to induce apoptosis20,21, facilitate degradation of c-JUN and c-MYC to inhibit the transition of the cell cycle from G1 to S phase22,23, and accelerate the degradation of snail to suppress epithelial-to-mesenchymal transition (EMT) and cell migration and so on24,25. Recently, researches on the DYRK2 have gradually revealed its distinct role as an oncogene26,27,28. In multiple myeloma (MM) and triple-negative breast cancer (TNBC), DYRK2 phosphorylates Rpt3-Thr25 of the 26S proteasome to activate the 26S proteasome, and then promotes the transition of the cell cycle from G1 to S phase29,30. The inhibition of DYRK2 impedes 26S proteasome activity and suppresses the cell cycle progression to inhibit cell proliferation31,32. Therefore, DYRK2 is a potential target for the treatment of MM and TNBC. These researches revealed the critical and multifarious roles of DYRK2 in different cancers. However, the regulation mechanism of DYRK2 in PCa is still unclear and has not been reported. Moreover, the reported small-molecule DYRK2 inhibitors lack selectivity and exhibit poor druggability. Thus, it is urgent to reveal the function of DYRK2 in PCa and develop DYRK2 inhibitors with better potency and higher selectivity to treat PCa and other cancers.

In our work, DYRK2 was identified as a potential target for PCa treatment. High expression of DYRK2 was detected and confirmed in both PCa patients and cell lines. Knock-down of DYRK2 in PCa cells suppressed cell proliferation and metastasis, promoted apoptosis, and caused a G1 arrest of the cell cycle. Furthermore, knock-down of DYRK2 significantly inhibited tumor growth of PCa in a xenograft model. Through virtual screening and structural optimization, we developed a unique DYRK2 inhibitor YK-2-69 with high selectivity over 370 kinases, and the detailed interactions between YK-2-69 and DYRK2 were further demonstrated by their co-crystal structure. Moreover, YK-2-69 displayed acceptable safety properties, favorable pharmacokinetic profiles, and stronger suppression of PCa progression than the first-line PCa drug enzalutamide in vivo. Therefore, DYRK2 was a biomarker in PCa diagnosis and a potential target to develop anti-PCa drugs. The DYRK2 inhibitor YK-2-69 with high selectivity and favorable druggability provided a potential candidate for the treatment of PCa.

Results

Highly expressed DYRK2 was a potential target in PCa

To investigate the role of DYRK2 in PCa, we first mined The Cancer Genome Atlas (TCGA) to analyze the expression of DYRK2 in normal and PCa patients. We found the higher expression of DYRK2 in PCa patients when compared with normal controls (Fig. 1a). Also, different from PCa patients with intermediate risk and below the age of 65, the expression of DYRK2 was significantly higher in high risk and above the age of 65 PCa patients, respectively (Fig. 1b, c). Importantly, the relapse-free survival (RFS) in patients with low expression of DYRK2 was remarkably better than those with high expression of DYRK2 (Fig. 1d). Therefore, DYRK2 was the potential target for PCa treatment based on analysis of TCGA. Meanwhile, other DYRK family members, DYRK1A, DYRK1B, DYRK3, and DYRK4 were not great candidate targets for anti-PCa drugs based on analysis of TCGA (Supplementary Fig. 1). Furthermore, the DYRK2 mRNA levels of malignancy PCa tissues were higher than adjacent normal prostate tissues in PCa patients (Fig. 1e). Immunohistochemistry of DYRK2 in these patient-derived PCa tumors also demonstrated that the expression of DYRK2 in tumor tissues was much higher than in normal tissues (Fig. 1f). Meanwhile, the protein levels of DYRK2 in PCa cells were further determined by western blotting analysis, and the results showed the higher expression of DYRK2 in DU145, PC-3, and 22Rv1 cells compared with prostate RWPE-1 cells (Fig. 1g). All these results indicated that DYRK2 is highly expressed in PCa, which could be a potential target to develop anti-PCa drugs.

Fig. 1: DYRK2 is highly expressed in PCa.
figure 1

ac Comparison of DYRK2 expression between tumor (red, n = 51) and normal (gray, n = 51) tissues (a), high (red, n = 345) and intermediate (blue, n = 152) risk PCa patients (b), age ≤65 (blue, n= 354) or > 65 (yellow, n = 143) PCa patients (c) in TCGA database. FRKM: fragment per killo million. The whiskers of boxplot represent the quantile percentile, from bottom to top are minima, 25%, median, 75%, and maxima respectively. Two-tailed Student’s t test was applied without adjustment for multiple comparisons (false discovery rate, FDR). d Kaplan–Meier survival plot of high (red line, n = 246) and low (blue line, n = 246) DYRK2 expression PCa patients. Log-rank test, P = 0.015. e, f Analysis of DYRK2 expression in three PCa patients. DYRK2 mRNA level (e) and immunohistochemical analysis of DYRK2 expression (f) in tumor and normal tissues. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. g DYRK2 protein levels in different PCa cell lines. Normal prostate epithelial cell RWPE-1 was used as the control. Source data are provided as a Source Data file.

Knock-down of DYRK2 remarkably reduced PCa burden

To further study the role of DYRK2 in PCa, we knocked down DYRK2 in DU145 and 22Rv1 cells using shRNAs (Fig. 2a, b and Supplementary Fig. 2a). DYRK2 depletion significantly suppressed the cell proliferation (Fig. 2c, d and Supplementary Fig. 2b), migration (Supplementary Fig. 2c) and invasion (Supplementary Fig. 2d) in PCa cells. In addition, knock-down of DYRK2 caused a G0/G1 arrest of the cell cycle (Fig. 2e, f) and induced apoptosis (Fig. 2g, h). The cell cycle-related proteins, including p-RB, CDK4, and CDK6, which promoted the cell cycle progression, were down-regulated in the DU145 shDYRK2 cell. In contrast, cyclin-dependent kinase inhibitors (CKIs) P21 and P27, which inhibited the cell cycle, were up-regulated (Fig. 2i). The cell apoptosis-related proteins P53 and cleaved PARP were up-regulated, and XIAP was down-regulated. The increased expression of E-cadherin was also detected, which demonstrated metastasis was suppressed (Fig. 2i). Furthermore, to determine the effects of DYRK2 knock-down on PCa growth in vivo, we subcutaneously implanted the DU145 shNC and shDYRK2 cells into the nude mice. The tumor growth was significantly inhibited (Fig. 2j) while the body weight of mice increased normally when compared with the shNC group (Supplementary Fig. 2e). H&E staining and Ki-67 immunohistochemical analysis of tumor tissues indicated that knock-down of DYRK2 exhibited potent efficacy of killing tumor cells and inhibiting PCa cell proliferation (Supplementary Fig. 2f, g). The WB analysis of the tumor tissues demonstrated that p-RB, CDK4, CDK6, and XIAP were down-regulated, while P27, P53, and cleaved PARP were up-regulated in vivo (Fig. 2k). Furthermore, we also inoculated subcutaneously 22Rv1 shNC and shDYRK2 cells into mice. Similar as the results in DU145 shDYRK2 studies, the body weight of mice increased normally but no visible tumors were detected in the 22Rv1 shDYRK2 group (Supplementary Fig. 2h, i). In summary, the down-regulation of DYRK2 remarkably reduced PCa tumor burden in vitro and in vivo, suggesting that DYRK2 played a critical function in regulating PCa and is a potential therapeutic target for the treatment of PCa.

Fig. 2: Knock-down of DYRK2 inhibited PCa in vitro and in vivo.
figure 2

a, b Protein level of DYRK2 in DU145 shNC/shDYRK2 (a) and 22Rv1 shNC/shDYRK2 (b) cells. c, d Cell viability of DU145 shNC/shDYRK2 cells (c) and 22Rv1 shNC/shDYRK2 cells (d) during a 5-day course. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. e, f Cell cycle phase distribution of DU145 shNC/shDYRK2 cells (e) and 22Rv1 shNC/shDYRK2 cells (f) determined by flow cytometry. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. g, h Apoptosis of DU145 shNC/shDYRK2 cells (g) and 22Rv1 shNC/shDYRK2 cells (h) determined by flow cytometry. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. i Western blotting analysis of indicated proteins in DU145 shNC and shDYRK2 cells. j, k BALB/c nude mice were implanted subcutaneously with DU145 shNC (n = 6) and shDYRK2 (n = 10) cells. Tumor volume of mice (j) was measured every two days. Error bar, mean ± SD. The shNC group was euthanatized at 29th day and shDYRK2 group was euthanatized at 49th day. Tumor tissues of mice treated with DU145 shDYRK2 and shNC cells were taken out, then the total proteins in the tumor were extracted and subjected to the western blotting analysis of indicated proteins (k). Source data are provided as a Source Data file.

YK-2-69 was discovered as a highly selective DYRK2 inhibitor

Considering the high expression level and critical regulation roles of DYRK2 in PCa, we took DYRK2 as a potential drug target and conducted a structure-based virtual screening of the Specs database and an in-house library to identify DYRK2 inhibitors (Fig. 3a, see Supplementary Methods for details). Compounds with poor drug-like properties were first filtered33,34,35,36, then the remained 195,483 compounds were subjected to structure-based virtual screening (DYRK2 PDB ID: 6K0J) via the Libdock and CDOCKER protocol of Discovery Studio 2020 (DS2020, Accelrys, CA, USA)32,37,38. 2,724 ligands were reserved and further clustered into 100 clusters, then 15 compounds were selected for DYRK2 inhibitory activity evaluation (Supplementary Fig. 3)39. Among these 15 compounds, compound 12 was identified as a top hit, which displayed potent inhibition on DYRK2 with a half maximal inhibitory concentration (IC50) value of 263 nM (Supplementary Table 1). Through systematic optimization (see Methods for details), multiple series of derivatives 16-27 (Supplementary Fig. 4) were synthesized. Among them, compound 26 (re-named as YK-2-69) exhibited the most potent DYRK2 inhibitory activity with an IC50 value of 9 nM (Fig. 3b). In particular, YK-2-69 showed selectivity to the DYRK subfamily with 60-fold selectivity over DYRK1B and more than 100-fold selectivity over DYRK1A, DYRK3, and DYRK4 (Fig. 3c). To further estimate the kinase selectivity of YK-2-69, its inhibitory activities against 370 kinases were tested. Besides DYRK2 and CDK4/640, YK-2-69 exhibited selectivity over 360+ kinases (Fig. 3d and Supplementary Fig. 5). In addition, the Kd values of lead compound 12 and YK-2-69 with DYRK2 were 4.21 µM and 92 nM, respectively (Supplementary Fig. 6a, b). Taken together, YK-2-69 was discovered as a DYRK2 ligand with potent inhibitory activity and high selectivity. The previous report reveals that 4E-binding protein 1 (4E-BP1) is a direct cellular target of DYRK2, and DYRK2 can directly phosphorylate 4E-BP141,42. Meanwhile, the phosphorylation of 4E-BP1 contributes to cell proliferation and tumor growth43,44,45. Our results revealed that p-4E-BP1 (Thr37/46) could interact with DYRK2 by the immunoprecipitation assay (Fig. 3e). Furthermore, YK-2-69 inhibited the phosphorylation of 4E-BP1 in a dose-dependent manner (Fig. 3f). These results demonstrated that YK-2-69 selectively binds to DYRK2 and inhibits its kinase activity.

Fig. 3: Discovery of the highly selective DYRK2 inhibitor YK-2-69.
figure 3

a Flowchart of virtual screening to discover the hit DYRK2 inhibitor compound 12. b Chemical structure and activities of YK-2-69. c Inhibitory activity of YK-2-69 against DYRK1A, 1B, 2, 3, and 4. dKinase selectivity experiment of YK-2-69 (1 µM) was carried out at Reaction Biology Corporation (https://www.reactionbiology.com/). The dot size indicates the inhibitory rate. A dot represents a type of kinase and the red dot represents DYRK2. e Validation of interaction of p-4E-BP1 (Thr37/46) with DYRK2 by co-immunoprecipitation assay. 1% volume of cell lysate was used as input. fWestern blotting analysis of indicated proteins after treatment with DMSO or YK-2-69 for 48 h. The special bands of 4E-BP1 and p-4E-BP1 were shown with arrows. Source data are provided as a Source Data file.

To explore the exact interaction of YK-2-69 with DYRK2 and elucidate the mechanism, we determined the co-crystal structure of YK-2-69 with DYRK2 at a high resolution of 2.5 Å (PDB ID: 7EJV, Fig. 4a and Supplementary Table 2). The co-crystal structure showed that YK-2-69 occupied the ATP-binding pocket of DYRK2, thereby preventing DYRK from exerting its enzymatic activity (Fig. 4b). The occupancy of the ATP-binding pocket is the same with all reported co-crystal structures31,32,46,47,48. The benzothiazole and pyrimidine rings were located deep into the ATP binding site. This orientation placed the pyrimidine ring adjacent to the Lys-231. Lys-231 played a critical role in forming the hydrogen bond in the previous co-crystal structures (PDB ID: 3KVW, 4AZF, 6HDR)46. The pyrimidine ring and the linked secondary amine of YK-2-69 also interacted with Lys-231 in the formation of two hydrogen bonds. The tailed piperazine ring extended out and the conjoint carbonyl formed a hydrogen bond with the amino side chain of Asn-234 (Fig. 4c).

Fig. 4: Co-crystal structure of YK-2-69 with DYRK2.
figure 4

a The FO-FC omitted map is contoured at 3.0 σ and shown as a blue mesh, which reveals the presence of YK-2-69. DYRK2 is shown as ribbons. PDB ID: 7EJV. b YK-2-69 occupies the ATP binding pocket of DYRK2. DYRK2 is shown as surface, and YK-2-69 is shown as spheres. c Detailed interactions between YK-2-69 and DYRK2 in the co-crystal structure.

YK-2-69 significantly inhibited growth and migration of PCa cells in vitro

Once we confirmed YK-2-69 as a potent and selective DYRK2 inhibitor, we further investigated its effects on PCa cells. YK-2-69 showed potent inhibitory activity against the proliferation of DU145, PC-3, and 22Rv1 cells (Supplementary Fig. 7a–c). But for DU145 and 22Rv1 shDYRK2 cells, YK-2-69 exhibited almost no inhibitory activity on the proliferation even at 80 μM (Fig. 5a), which further confirmed the selective on-target activity of YK-2-69 to DYRK2. Meanwhile, YK-2-69 significantly inhibited the cell proliferation of DU145, PC-3, and 22Rv1 cells in a dose-dependent manner (Fig. 5b–e and Supplementary Fig. 7d, e), which was similar to knocking down DYRK2 in the PCa cells. Same as depletion of DYRK2 reduced EMT, YK-2-69 also remarkably inhibited the migration and invasion of DU145, PC-3, and 22Rv1 cells in a dose-dependent manner (Fig. 5f–i and Supplementary Fig. 7f, g). All these results confirmed that YK-2-69 reduced cell proliferation and EMT through inhibiting DYRK2.

Fig. 5: YK-2-69 inhibited cell growth and metastasis and induced apoptosis in PCa cells.
figure 5

a Antiproliferative activity of YK-2-69 against DU145 and 22Rv1 shDYRK2 cells. Error bar, mean ± SD, n = 3. b, c Effects of YK-2-69 (4, 8, 12 µM or 0.5, 1, 2 µM), palbociclib (2 or 0.5 µM), or enzalutamide (40 or 20 µM) treatment on the DU145 (b) and 22Rv1 (c) cells viability during a 5-day course. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. d, e Quantification of DU145 (d) and 22Rv1 (e) cells colony numbers. Before being plated on the 24-well plate for colony formation, DU145 and 22Rv1 cells were treated with DMSO, YK-2-69 (2, 4, 8, or 0.5, 1, 2 µM), palbociclib (2 or 0.5 µM), or enzalutamide (40 or 20 µM) for 48 h. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. f, g Quantification of migration ability of DU145 (f) and 22Rv1 (g) cells after treatment with DMSO, YK-2-69 (2, 4, 8 µM or 0.5, 1, 2 µM), palbociclib (2 or 0.5 µM), or enzalutamide (40 or 20 µM) for 48 h. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. h, i Quantification of invasion ability of DU145 (h) and 22Rv1(i) cells after treatment with DMSO, YK-2-69 (2, 4, 8 or 0.5, 1, 2 µM), palbociclib (2 or 0.5 µM), or enzalutamide (40 or 20 µM) for 48 h. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. j, k Apoptosis of DU145 (j) and 22Rv1 (k) cells after treatment with DMSO, YK-2-69 (2, 4, 8 or 0.5, 1, 2 µM), palbociclib (2 or 0.5 µM), or enzalutamide (40 or 20 µM) for 48 h determined by flow cytometry. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. l, m Cell cycle phase distribution of DU145 (l) and 22Rv1 (m) cells after treatment with DMSO, YK-2-69 (2, 4, 8 or 0.5, 1, 2 µM), palbociclib (2 or 0.5 µM), or enzalutamide (40 or 20 µM) for 48 h determined by flow cytometry. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. Source data are provided as a Source Data file.

To investigate the possible role of YK-2-69 on the cell cycle and apoptosis, we conducted flow cytometry analysis. The treatment of DU145, PC-3, and 22Rv1 cells with YK-2-69 caused a significant increase of the apoptotic cell population (Fig. 5j, k and Supplementary Fig. 7h) and arrested cell cycle at the G0/G1 phase (Fig. 5l, m and Supplementary Fig. 7i) in a concentration-dependent manner. Furthermore, the DU145 cells treated with YK-2-69 decreased cell cycle-related protein levels of p-RB, CDK4 and CDK6 as well increased P21, increased apoptosis-related protein levels of P53 and cleaved PARP as well decreased XIAP, and also increased expression of E-cadherin (Supplementary Fig. 7j). In summary, knock-down of DYRK2 and small-molecule inhibitor YK-2-69 displayed the similar effects on PCa cells in vitro, which inhibited cell growth through G0/G1 arrest and apoptosis induction, and decreased the EMT activity.

DYRK2 KD targeted similar signaling pathways to YK-2-69 treatment in the proliferation inhibition of PCa cells

To investigate which signaling pathways are responsible for the anti-prostate cancer function of DYRK2 inhibitors, we performed transcriptome-wide RNA-sequencing analysis of DYRK2 KD- and YK-2-69- treated human DU145 and 22Rv1 cells as well as control cells49,50. Many signaling pathways regulated by DYRK2 KD could also be regulated by YK-2-69 treatment, especially the vast majority of pathways (28 out of 34, 82.4%; 46 out of 48, 95.8%) inhibited by DYRK2 KD were also suppressed by YK-2-69 treatment (Fig. 6a and Supplementary Fig. 8a). By independent analysis of two different comparisons, we found that both DYRK2 KD and YK-2-69 treatment induced significant inhibition of MYC targets (Fig. 6b and Supplementary Fig. 8b), which may contribute to the inhibitory effects of DYRK2 KD and YK-2-69 treatment on cell cycle and proliferation. Moreover, we made shDYRK2 and YK-2-69 as a single group and re-analyzed the sequencing data between this group and its control group (shNC group and DMSO group). Consistently, we found that YK-2-69 treatment and DYRK2 KD inhibited cell cycle and proliferation related signaling pathways MYC target V1, MYC target V2, E2F targets (Fig. 6c). DYRK2 KD and YK-2-69 treatment significantly down-regulated genes enriched in MYC target V1, MYC target V2 and Mitotic SPINDLE (Fig. 6d). Then, we also found that DYRK2 KD and YK-2-69 treatment did not affect the ANDROGEN RESPONSE signaling pathway, while enzalutamide significantly inhibited the ANDROGEN RESPONSE signaling pathway (Fig. 6e and Supplementary Fig. 8c). In summary, these data suggested that DYRK2 KD and YK-2-69 treatment played important and similar roles in cell proliferation inhibition.

Fig. 6: Transcriptome-wide RNA sequencing assays in PCa DU145 cells.
figure 6

a Transcriptome strategy of RNA-sequencing conducted on DU145 cells exposed to YK-2-69 (3 µM) for 48 h. The shNC, shDYRK2, DMSO, and YK-2-69 groups all contain two biological replicates. Venn diagram of upregulated and downregulated signaling pathways in DYRK2 KD- and YK-2-69-treated DU145 cells. The number of genes in every signaling pathway is >50. Normalized enrichment score (NES) >1 or <−1; P < 0.05; FDR < 0.25. b The signaling pathways enriched in different groups obtained through Gene Set Enrichment Analysis (GSEA). c The core-enriched decreased (blue) and increased (red) signaling pathways in shDYRK2 and YK-2-69 treatment groups when compared with shNC and DMSO groups, respectively. The signaling pathways with P < 0.05 are presented. d The relative abundance of genes involved in MYC target V1, MYC target V2 and Mitotic SPINDLE in DYRK2 KD- and YK-2-69-treated DU145 cells. n = 2. The whiskers of boxplot represent the quantile percentile, from bottom to top are minima, 25%, median, 75%, and maxima respectively. Two-tailed Student’s t test was applied without adjustment for multiple comparisons (FDR). e GSEA was used to analyze the effects of DYRK2 KD or YK-2-69 treatment on the ANDROGEN RESPONSE signaling pathway in DU145 cells. f Volcano plot of significantly affected genes (absolute fold change > 2, P < 0.05) in DU-145 shDYRK2 group relative to shNC group and YK-2-69 group relative to DMSO group. The negative binomial distribution test of DESeq2 software was used. g Effects of DYRK2 KD or YK-2-69 treatment on the RRS1, GRWD1, CCNG2, and YPEL3 mRNA levels in DU145 cells. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. h The principal component analysis was used to identify transcriptome differences between two samples. i The different signaling pathways enriched in DYRK2 KD and YK-2-69 treatment groups obtained through GSEA. Source data are provided as a Source Data file.

Through the analysis of differentially expressed genes (DEG) between shDYRK2 and YK-2-69-treated samples in transcriptome-wide RNA-sequencing (Fig. 6f and Supplementary Dataset 1) and the further experiment confirmation, we found significant expression changes in human regulator ribosome synthesis 1 (RRS1), glutamate-rich WD40 repeat containing 1 (GRWD1), cyclin G2 (CCNG2), and Yippee-like-3 (YPEL3). RRS1, an essential nuclear protein involved in ribosome biogenesis, is overexpressed in some human cancers51, and downregulation of RRS1 causes a G1 arrest of cell cycle52. GRWD1, a negative transcriptional regulator of P53, plays an oncogenic activity in human cancers53,54. Meanwhile, previous studies have shown that overexpression of CCNG2 can induce apoptosis and inhibit cell proliferation55,56. YPEL3, a P53-regulated gene, has been reported to display growth suppressive and EMT inhibitory activity57,58. The experimental results of qRT-PCR demonstrated that oncogenes RRS1and GRWD1 were downregulated and tumor suppressors CCNG2 and YPEL3 were upregulated regardless of knocking down DYRK2 or YK-2-69 treatment (Fig. 6g and Supplementary Fig. 8d). These results suggested that RRS1, GRWD1, CCNG2, and YPEL3 may play important roles in the DYRK2 regulation mechanism and YK-2-69 treatment to PCa cells.

The principal component analysis indicated that DYRK2 KD and YK-2-69 treatment also induced some different transcriptomic changes (Fig. 6h). Therefore, the further Gene Set Enrichment Analysis (GSEA) between shDYRK2 and YK-2-69 was conducted, which demonstrated that YK-2-69 treatment induced inhibition of DNA repair and TGF β signaling pathways, while DYRK2 KD had no remarkable effects on these two signaling pathways (Fig. 6i). Inhibition of DNA repair is a successful therapeutic strategy for cancers with several approved drugs in the market59,60,61. TGF-β also plays a critical role as a tumor promoter in late-stage cancer62,63, and a number of drugs for inhibiting TGF β signaling pathway have been developed and evaluated in clinical trials64. Therefore, YK-2-69 may regulate some different signaling pathways to generate potent antitumor activity when compared with DYRK2 KD.

YK-2-69 displayed favorable safety properties and pharmacokinetic profiles

To evaluate the toxic effects of DYRK2 inhibitor YK-2-69 in vivo, the ICR mice (n = 10/group) were orally administrated YK-2-69 in the single dose of 2500 mg/kg, 5000 mg/kg, and 10,000 mg/kg, respectively. No abnormality and death were observed in mice of each group in 14 days. Also, no difference was detected in the mice body weight (Supplementary Fig. 9a) and main organs, including heart, liver, spleen, lung, and kidney, between drug-treated and control groups (Supplementary Fig. 9b, c). These data confirmed the favorable safety properties of YK-2-69 in vivo.

To further explore the pharmacokinetic profiles of YK-2-69, the Sprague-Dawley (SD) rats (n = 3/group) were administrated YK-2-69 by oral and intravenous injection (Table 1). In the intravenous group, the half-life (t1/2), Cmax, and AUC0-∞ values were 3 h, 974 ng/mL, and 1503 h*ng/mL, respectively. In oral administration group, YK-2-69 displayed the pharmacokinetic parameters as follows: t1/2 = 5 h, Cmax = 674 ng/mL, and AUC0-∞ = 8384 h*ng/mL. Moreover, the oral bioavailability of YK-2-69 is 56%. In summary, these results demonstrated the favorable druggability of YK-2-69 with favorable safety properties and pharmacokinetic profiles in vivo.

Table 1 Pharmacokinetic parameters of compound YK-2-69 in SD ratsa.

YK-2-69 displayed more potent suppression on PCa than first-line drugs enzalutamide and palbociclib in vivo

To evaluate antitumor activities of YK-2-69 in vivo, the DU145 xenograft mouse model was first established. Enzalutamide, the first-line PCa drug, and palbociclib, the selective CDK4/6 inhibitor in the market, were selected as the positive controls since CDK4/6 are down-regulated in DU145 cells when treated by YK-2-69. They were administered orally once a day for seven consecutive weeks. The low dose of YK-2-69 (100 mg/kg) displayed similar antitumor activities with enzalutamide but better activities than palbociclib. While the high dose of YK-2-69 (200 mg/kg) demonstrated much better antitumor activities than both enzalutamide and palbociclib (Fig. 7a, c, d). Especially, different from enzalutamide and palbociclib which only delayed the tumor growth, the high dose of YK-2-69 not only suppressed the growth of tumor, but also decreased the volume of tumor since the 31st day (Fig. 7a). It is noteworthy that the body weight of mice also increased gradually in the high dose group (Fig. 7b). H&E staining of tumor tissues and immunohistochemical analysis of Ki-67 expression indicated that YK-2-69 exhibited potent efficacy in killing PCa cells and inhibiting cell proliferation (Fig. 7e, f). Similar to the in vitro data, the WB analysis of the tumor tissue excised from DU145 xenograft mouse also indicated that the cell proliferation and apoptosis-related proteins p-4E-BP1, p-RB, CDK4/6, RRS1, and XIAP were deregulated, and P27, P53, CCNG2, and cleaved PARP were upregulated (Fig. 7g). Meanwhile, YK-2-69 also significantly inhibited tumor growth in the PC-3 xenograft mouse model. Furthermore, same as YK-2-69 reducing tumor size in the DU145 xenograft model, the high dose of YK-2-69 could also decrease tumor volume in the PC-3 xenograft model (Supplementary Fig. 10). Taken together, YK-2-69 displayed much better antitumor activity than first-line drugs enzalutamide and palbociclib in vivo.

Fig. 7: YK-2-69 demonstrated remarkable antitumor activities in vivo.
figure 7

ad BABLc nude mice received subcutaneous injection of 1 × 107 DU145 cells in the right flank. When tumors grew ~80–100 mm3, mice (n = 10/group) were orally administrated vehicle, palbociclib (100 mg/kg), enzalutamide (100 mg/kg), and YK-2-69 (100 and 200 mg/kg) every day. Tumor volumes (a) and body weight of mice (b) were measured every 2 days. After 35 days, mice in the control group were killed. After 49 days, mice of treatment groups were killed. Tumor tissues of each group were weighed (c) and then photographed (d). Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 10. e Representative images of H&E and Ki-67 staining of paraffin section of tumor from mice treated with vehicle, palbociclib, enzalutamide, and YK-2-69. f Quantification of Ki67 positive rate of tumor from mice treated with vehicle, palbociclib, enzalutamide, and YK-2-69. Unpaired two-tailed Student’s t test. Error bar, mean ± SD, n = 3. gThe total proteins in the tumor were extracted and used in the western blotting analysis of indicated proteins. The special bands of 4E-BP1 and p-4E-BP1 were shown with arrows. h A proposed model for inhibition of DYRK2 by YK-2-69 for the treatment of PCa. Source data are provided as a Source Data file.

Based on these results, YK-2-69 selectively binds to DYRK2 and inhibits its kinase activity to suppress the phosphorylation of 4E-BP1, which results in the inhibition of cell proliferation (Fig. 7h). The inhibition of DYRK2 by YK-2-69 can also down-regulate RRS1 and further up-regulate P21 and P27 to suppress CDK4/651,52. RRS1-P21/27-CDK4/6 axes can restrain the transition from G1 to S phase of the cell cycle and eventually inhibit cell proliferation. Simultaneously, down-regulation of DYRK2 by YK-2-69 can increase the P53 and CCNG2 level, which promotes apoptosis55,56. In summary, this highlights DYRK2 inhibition by YK-2-69 as a promising combination to inhibit proliferation and promote apoptosis, which provides a “kill two birds with one stone” regimen of PCa.

Discussion

The expression level of DYRK2 widely depends on the human tumor tissues, and it plays diverse roles in the occurrence and development of various cancers, which highlights the possibility of DYRK2 as a potential target for cancer treatment. Previous reports mainly considered DYRK2 as a cancer suppressor, which induces apoptosis through promoting phosphorylation of P53 in colorectal cancer65, inhibits cell cycle from G1 to S phase via degrading c-JUN and c-MYC in breast cancer22, and suppresses EMT by accelerating SNAIL degradation in glioma66. But not all reports demonstrate that DYRK2 inhibits cancer initiation and growth as a cancer suppressor. Recent reports revealed the DYRK2 accelerated G1 to S phase transition of the cell cycle through regulating 26S proteasome activity, and was considered as an oncogene in MM and TNBC26,31. Considering the critical function of DYRK2 in cancers, we conducted a data-mining of TCGA to find the high expression of DYRK2 in PCa, which was positively correlated with clinical prognosis and mortality (Fig. 1a–d). However, the function of DYRK2 in PCa is still unclear. In our work, DYRK2 was found to be highly expressed in PCa patient samples and cells (Fig. 1e–g). The knock-down of DYRK2 in PCa cells significantly inhibited cell growth and metastasis, caused a G1 arrest of the cell cycle, and induced apoptosis (Fig. 2c–h and Supplementary Fig. 2b–d). In addition, knock-down of DYRK2 significantly inhibited tumor growth in vivo (Fig. 2j and Supplementary Fig. 2h). All these results indicate that DYRK2 is a potential therapeutic target for the treatment of PCa.

Although several DYRK2 inhibitors were discovered and exhibited anti-cancer activity in MM and TNBC31,32, their selectivity over DYRK family members and drug-like properties need to be further modified. To develop more potent and selective DYRK2 inhibitors, we performed a high-throughput virtual screening and identified a DYRK2 hit with a benzothiazole chemical scaffold, which was further modified to offer the highly selective inhibitor YK-2-69. YK-2-69 exhibited stronger inhibition to DYRK2 with an IC50 value of 9 nM and showed selectivity over the DYRK subfamily and a panel of 370 kinases (Fig. 3). To explore the clear mechanism and elucidate the exact interaction of YK-2-69 with DYRK2, we solved their co-crystal structure with a high resolution at 2.5 Å (PDB ID: 7EJV), which showed the essential interaction residues Lys-231 and Asn-234 (Fig. 4).

Similar to the knock-down of DYRK2 in the PCa cells, YK-2-69 also significantly inhibited cell growth through G1 arrest and apoptosis induction, and decreased the EMT activity (Fig. 5 and Supplementary Fig. 7). Transcriptome-wide RNA sequencing assays demonstrated that DYRK2 KD and YK-2-69 treatment played important and similar roles in cell proliferation inhibition (Fig. 6 and Supplementary Fig. 8). Importantly, YK-2-69 displayed acceptable safety properties with a maximal tolerable dose of >10,000 mg/kg (Supplementary Fig. 9) and favorable pharmacokinetic profiles with 56% bioavailability (Table 1) in vivo. Moreover, YK-2-69 exhibited much better antitumor activities than both enzalutamide and palbociclib. Especially, YK-2-69 not only suppressed the growth of tumor, but also decreased the volume of tumor, which was completely different from enzalutamide and palbociclib (Fig. 7 and Supplementary Fig. 10). These results provided us a possibility that YK-2-69 may contribute to solving the drug-resistant dilemma of enzalutamide as hormonal therapy. YK-2-69 exhibited much higher anti-PCa efficacy via synergistic regulation on a panel of pathways, including DYRK2-4E-BP1, DYRK2-RRS1-P21/27-CDK4/6, and so on, to promote apoptosis and inhibit proliferation. This might be one of the possible reasons why YK-2-69 exhibited significant anti-PCa efficacy.

The latest data from International Agency for Research on Cancer reports (World Cancer Report 2020) that prostate cancer is the second most common cancer in men worldwide, with an estimated 1.3 million new cases and 360,000 deaths in 2020. However hormonal therapy, the leading treatment of PCa, is only a remission but not a cure for PCa, and most PCa patients became resistant to hormonal therapy at last. Therefore, it is urgent and meaningful to identify novel targets and develop new drugs for PCa. Our work identified DYRK2 as a potential drug target and verified its critical roles in PCa, which offers a valuable direction for the treatment of PCa. Especially, we discovered a highly selective DYRK2 inhibitor with favorable druggability, which could be used as a small-molecule probe for biological studies and also provide a potential candidate for PCa clinical treatment. Since DYRK2 plays critical roles in various human cancers, targeting DYRK2 could also provide a promising opportunity for other patients with refractory cancers.


Article classification: Biological abstract
Share to:
Tel:+86-027-87610298
Tel:+86-027-87610297
Add:Room A11-329, 1st Floor, No.1, SBI Venture Street, Optics Valley, East Lake
New Technology Development Zone, Wuhan, China.
Certificate NO.:U18Q28010569R0S