Discovery of the First Potent DYRK2 Proteolysis Targeting Chimera Degraders

Abstract

Dual-specificity tyrosine phosphorylation-regulated kinase 2 (DYRK2) has been identified as a promising oncogenic driver of several types of cancer and is considered to be a critical cancer therapeutic target. Several inhibitors of DYRK2 have been reported, but no degraders have been found yet. In this work, we designed and synthesized the first series of proteolysis-targeting chimeras (PROTACs) using curcumin and its analogs as warheads to target and degrade DYRK2. The results of degradation assays showed that the compound CP134 could effectively downregulate the intracellular DYRK2 level (DC₅₀ = 1.607 μM). Further mechanism of action experiments revealed that CP134 induced DYRK2 degradation through the ubiquitin–proteasome system. Altogether, CP134 disclosed in this study is the first potent DYRK2 degrader, which could serve as a valuable chemical tool for further evaluation of its therapeutic potential, and our results broaden the substrate spectrum of PROTAC-based degraders for further therapeutic applications.

Protein kinase DYRK2 is a member of the dual-specificity tyrosine phosphorylation-regulated kinase (DYRK) family, which in turn belongs to the cyclin-dependent kinases, mitogen-activated protein kinases, glycogen synthase kinases, and CDC-like kinases (CMGC) superfamily.¹ 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.² DYRK2 plays different roles in cell growth, proliferation, and development. For example, it can promote phosphorylation of p53 to induce apoptosis.³ Meanwhile, DYRK2 has also been reported as a potential target for several human cancers. In triple-negative breast cancer (TNBC) cells, DYRK2 phosphorylates HSF1, promoting its nuclear stability and transcriptional activity.⁴ DYRK2 phosphorylates the Rpt3 subunit in the regulatory particle of the proteasome at Thr25, leading to the upregulation of proteasome activity.⁵ Both multiple myeloma (MM) and TNBC are extremely dependent on the 26S proteasome function for survival and progression of disease,^6,7 and DYRK2 is overexpressed in these tumors. DYRK2 has also been reported as a potential target for prostate cancer (PCa). Knockdown of DYRK2 in PCa cells suppressed cell proliferation and metastasis, promoted apoptosis, and caused a G1 arrest of the cell cycle.^8,9 Thus, DYRK2 is believed to be a promising target for cancer drug discovery.

Targeted protein degradation (TPD) is a rapidly growing strategy in drug discovery, which could overcome the challenges encountered by traditional chemical inhibition of a target of interest, such as drug resistance, toxic side effects, and off-target effects.¹⁰ Proteolysis-targeting chimeras (PROTACs) are heterobifunctional molecules consisting of an E3 ligase binder and a substrate-targeting ligand that exploits the cellular protein degradation machinery to selectively degrade the protein of interest (POI).¹¹ PROTAC molecules can polyubiquitinate the POI, which is subsequently degraded via the proteasome.¹² Unlike occupancy-based inhibition, PROTACs act catalytically,^13,14 enabling them to target previously intractable proteins and to be effective even when resistance to inhibitors develops.

Because DYRK2 plays important while controversial roles in human cancers, it is the most deeply studied among the class II DYRK protein subfamily.^15,16 Prior to this, several teams have reported various potent selective DYRK2 inhibitors like YK-2-69, LDN192960, curcumin, and harmine.^8,17−21 However, no DYRK2 degraders have been reported. Though these inhibitors have demonstrated potent inhibitory effects, they still confront limitations as traditional inhibitors such as drug resistance and off-target effects. It is crucial to discover new degraders, such as PROTACs, that can treat diseases effectively by overcoming the shortcomings of traditional drugs. Curcumin, the active ingredient in Curcuma longa, has been found to exhibit anticancer and antiproliferative properties through multiple pathways.^22,23 However, its application has been reported to be limited by its poor bioavailability,²⁴ often leading to a lack of in vivo activity. In 2018, Guo’s group discovered that curcumin had potent and selective inhibition against DYRK2, and the cocrystal structure of the curcumin-DYRK2 complex (PDB: 5ZTN) revealed that curcumin occupied the ATP-binding pocket of DYRK2 directly.²⁰ Hence, developing new PROTACs using curcumin as a warhead could be possible, and it also would be quite a challenge since curcumin is generally considered a pan-assay interference (PAINS) compound.²⁵ In this study, we selected curcumin and its optimized analogs A1–A3²⁶ as a POI binding moiety via a virtual-screening-based workflow, and the first series of small-molecule degraders of DYRK2 were subsequently designed. These compounds were synthesized and characterized, and their degradation effects on DYRK2 were evaluated using a variety of bioassays. The corresponding mechanism of action (MOA) has also been explored. Overall, our study will expand a new target for PROTACs, prove that DYRK2 can also be targeted for degradation, and present a set of novel chemical tools for the research community to further explore the therapeutic potential of DYRK2.

The heterobifunctional PROTAC molecules are known to consist of three elements: a ligand for the POI, a linker, and a ligand for recruiting an E3 ligase.¹⁸ The design of new PROTACs was started from the previously mentioned DYRK2 inhibitor curcumin, which could be bound tightly to DYRK2. To rationally design PROTAC degraders, molecular docking studies of curcumin and its analogs A1–A3 with DYRK2 provided a structural basis. As shown in Figure 1a, molecular docking simulation of DYRK2–curcumin revealed that curcumin bound to the DYRK2 ATP-binding pocket directly. One of the 4-hydroxy-3-methoxyphenyl groups and two carbonyl groups of curcumin formed hydrogen bonds with DYRK2’s Leu231, Lys178, and Asp295, respectively (Figure 1a), that anchored curcumin deep within the ATP-binding pocket of DYRK2. This result was similar to the cocrystal structure (PDB entry 5ZTN). The molecular docking of analogue A3 with DYRK2 showed that it also occupied the ATP-binding pocket, but the anchoring mode was different from that of curcumin. The imine in A3 formed a hydrogen bond with Ile155 (Figure 1b), and the surrounding amino acid residues Lys165 and Ile294 participated in the hydrophobic interaction. In addition, the alignment of curcumin with A3 in the crystal structure revealed that they occupied the same position of DYRK2 (Figure S1d). A1 and A2 were similar to A3, occupying the same binding pocket as curcumin (Figure S1a–d).

Figure 1.

Design of small-molecule degraders based on PROTAC. (a) Molecular docking simulation of CUR in complex with DYRK2 (PDB entry 5ZTN). (b) Molecular docking simulation of CUR analog A3 in complex with DYRK2 (PDB entry 5ZTN). (c, d) SPR results of CUR and CUR analog A3 with DYRK2. The unit is μM. (e) Enzyme inhibition assay of CUR and A1–A3 on purified DYRK2 in 20 μM ATP concentrations. (f) Chemical structures of the designed compounds. Quantitative data are represented as the mean ± SEM of three independent replicates.

Surface plasmon resonance (SPR) results further proved that curcumin and its analogs A1–A3 had a binding affinity with DYRK2 (Figures 1c,d and S1e,f). The dissociation constant (K_d) values of A1–A3 and curcumin with DYRK2 were 5.0, 4.7, 7.9, and 14.5 μM, respectively. In addition, in vitro enzyme inhibition assays were performed, and the results showed that A1–A3 and curcumin inhibited DYRK2 with half-maximal inhibitory concentration (IC₅₀) values of 28, 19, 24, and 10 nM, respectively (Figure 1e). Curcumin and A1–A3 were employed as the ligand for DYRK2, and poly(ethylene glycol) (PEG)-based chains and carbon chains were used as linker moieties which could maximize the chance of the POI associating with the ubiquitin ligase complex.²⁷ For the E3 ligase ligand, widely used Von Hippel–Lindau (VHL) and CRBN ligands were explored. VH032 and pomalidomide (POM) were selected as the E3 ligands. Eventually, two sets of heterobifunctional compounds (UR and CP PROTACs) were thus designed (Figure 1f).

The synthetic routes for the two series of compounds are illustrated in Scheme 1. Curcumin condensed with the corresponding prefunctionalized linkers 2 and 5 to obtain intermediates 3 and 6, respectively. The series UR compounds were prepared through amide condensations between curcumin–linker conjugates (3 or 6) and VH032 (4) or prefunctionalized POM (7).

Scheme 1. Synthesis of Compounds.

Reagents and conditions: (a) DCC, DMAP, acetone, r.t.; (b) HATU, DIEA, DMF, r.t.; (c) TFA, DCO, r.t.; (d) K₂CO₃, DIEA, DMF, r.t.

In addition, intermediate 8 was obtained through nucleophilic substitution between curcumin analogs A1–A3 and n-butyl bromoacetate, and intermediate 11 was obtained by aromatic nucleophilic substitution of raw materials 9 and 10 followed by deprotection. Finally, intermediates 8 and 10 were condensed to obtain the CP series compounds. All synthesized new degrader compounds were characterized by ¹H NMR, ¹³C NMR, MS, and HRMS.

To evaluate the suitability of five curcumin-based PROTACs and six analog-based PROTACs for chemically inducing DYRK2 degradation, we determined the cellular levels of DYRK2 in human cervical carcinoma cells HeLa and human breast cancer cells MDA-MB-231 (MM231) through Western blot analysis. As shown in Figure 2a,b, the curcumin-based UR PROTACs had less efficient degradation than the analog-based CP PROTACs. UR113 and UR114, UR PROTACs based on the CRBN E3 ligase ligand pomalidomide and a PEG linker, significantly decreased the DYRK2 level. CP PROTACs with four PEG chains exhibited more effective degradation than others. It could be seen that analogs may be more suitable for the warhead of PROTAC than curcumin, and PROTACs that contained ligand POM and four PEG chains performed better than others. Thus, CP134 was selected for follow-up experiments.

Figure 2.

DYRK2 degradation effects of compounds. (a) Western blot analyses of DYRK2 in HeLa cells treated with UR PROTACs (10 and 20 μM) or CP PROTACs (5 and 10 μM) for 24 h. (b) Quantitative analysis of (a). (c) Western blot analyses of DYRK2 in MM231 cells after treatments with different concentrations of CP134 for 24 h. (d) Quantitative analysis of (c). (e) Western blot analyses of DYRK2 in HeLa cells after treatments with different concentrations of CP134 for 24 h. (f) Quantitative analysis of (e). (g) Western blot analyses of DYRK2 in HeLa cells after treatment with 5 μM CP134 for different periods. (h) Quantitative analysis of (g). (i) Representative confocal LSM image and quantitative analysis of HeLa cells after treatments with different concentrations of CP134 for 12 h. (j) FCM histograms showing the fluorescence instensities of CP134-treated MM231 cells. GAPDH was selected as a loading reference. Quantitative data are represented as the mean ± SEM of three independent replicates, and statistical significance was assessed by one-way ANOVA (n.s., not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.005).

Next, a gradient degradation experiment was performed on HeLa and MM231 cells. Immunoblot assays and quantitative analysis showed that CP134 exhibited a potent degradation effect (Figures 2c–f and S2a,b). The half-maximal degradation concentration (DC₅₀) of CP134 for DYRK2 was 1.607 μM in MM231 and 3.265 μM in HeLa, respectively. Further exploration revealed that CP134 led to downregulation of DYRK2 protein levels from 3 to 24 h, as illustrated in Figure 2g,h. Notably, a significant decrease in DYRK2 levels was observed after the ninth hour of treatment. These results indicated that CP134 reduced the intracellular DYRK2 level in a concentration- and time-dependent manner.

To further confirm that CP134 could effectively degrade DYRK2 in vitro, immunofluorescence experiments using confocal laser scanning microscopy (LSM) and flow cytometry (FCM) were performed. HeLa and MM231 cells were treated at doses of 5 and 10 μM CP134 for 12 h. The fluorescence intensity in the Cy3 and FITC-A channels served as a measure of the cellular DYRK2 protein level. The confocal LSM examination showed that fluorescence emitted by the treatment group was much weaker compared to the control group, and a similar result was also observed through a flow cytometry assay (Figure 2i,j). These results proved that CP134 could effectively reduce the DYRK2 level in vitro.

DYRK2 has been identified as capable of directly phosphorylating p53 at Ser46³ and c-Myc at Ser62.²⁸ Thus, p53 and c-Myc were selected as downstream substrates of DYRK2. MM231 cells were treated with CP134 from 1.25 to 20 μM, and then Western blot analyses were carried out. As shown in Figure 3a,b, after 12 h of treatment with 5 μM CP134 or more, the levels of phosphorylated p53 (Ser46) and c-Myc (Ser62) were significantly downregulated in MM231 cells. The degradation of DYRK2 by CP134 caused a downregulation of its downstream substrate levels. Phosphorylated p53 and c-Myc were compared under the same concentration of CP134 and warhead. It is worth noting that CP134 has a similar effect on phosphorylated p53 and c-Myc as its warhead A3 (Figure S2c,d).

Figure 3.

CP134 can downregulate the level of DYRK2 substrate. (a) Western blot analyses of phosphorylated p53 (Ser46) and phosphorylated c-Myc (Ser62) in MM231 cells after treatments with different concentrations of CP134 for 12 h. (b) Quantitative analysis of (a). GAPDH was selected as a loading reference. Quantitative data are represented as the mean ± SEM of three independent replicates, and statistical significance was assessed by one-way ANOVA (n.s., not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.005).

The SPR assay further confirmed the interaction between DYRK2 and CP134, with a K_d value of 29.7 μM (Figure S2e). To assess the importance of the CRBN-binding moiety in compound CP134, n-CP134, a negative control compound, was synthesized by installing a methyl group in the CRBN ligand portion of CP134 (Figure 4a). HeLa cells were treated with compounds CP134 and n-CP134 under the same condition. Compound n-CP134 caused no significant change in the level of DYRK2 compared with CP134 (Figure 4b,c), highlighting the importance of E3 ligase binding for the degrader molecules to achieve effective DYRK2 degradation. In the presence of A3, the degradation of DYRK2 by CP134 was reduced in a dose-dependent manner (Figure 4d,e). This proved that CP134 targets DYRK2 through its inhibitor moiety.

Figure 4.

Mechanistic evaluation of CP134-induced degradation. (a) Chemical structure of n-CP134. (b) Western blot analyses of DYRK2 in CP134- or n-CP134-treated (12 h) MM231 cells. (c) Quantitative analysis of (b). (d) Western blot analyses of DYRK2 in CP134-treated (12 h) MM231 cells in the absence or presence of A3. (e) Quantitative analysis of (d). (f) Western blot analyses of DYRK2 in CP134-treated (12 h) MM231 cells in the absence or presence of a UPS inhibitor. (g) Quantitative analysis of (f). (h) Co-IP analysis of ubiquitin and DYRK2 in CP134-treated (12 h) MM231 cells. The cells were all pretreated with 10 μM MG132 for 2 h before the administration of CP134. GAPDH was selected as a loading reference. Quantitative data are represented as the mean ± SEM of three independent replicates, and statistical significance was assessed by one-way ANOVA (n.s., not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.005).

Investigations to further explore whether the degradation of DYRK2 by CP134 relied on the proteasome were performed. The dependence of CP134’s degradation on proteasome activity was assessed using the proteasome inhibitor MG132, which is widely used to inhibit proteasome activity. Blocking the proteasome with MG132 remarkably inhibited the degradation of DYRK2 (Figure 4f,g). Moreover, coimmunoprecipitation (co-IP) analysis showed that the interaction between ubiquitination and DYRK2 was enhanced after the treatment with CP134 (Figure 4h), which indicated that CP134 promoted the ubiquitination level of the POI. As shown in Figure 4h, the ubiquitin was pulled down with DYRK2 together after the treatment with CP134, which meant that CP134 could promote the ubiquitination level of DYRK2. Collectively, these experiments suggested that CP134 induces DYRK2 degradation mainly through the Ub-dependent proteasome system.

As mentioned before, DYRK2 plays a significant and complex role in cancer cell growth and proliferation.³⁻⁹ Cytotoxicity assays of CP134 and warheads were performed in the MM231 cell line (Figure S2f). The results showed that CP134 and A1–A3 effectively inhibited cell proliferation with IC₅₀ values of 9.1, 5.6, 1.6, and 4.2 μM respectively, while curcumin (IC₅₀ > 20 μM) did not have an obvious effect on cell proliferation.

In summary, we discovered the first series of PROTACs using curcumin or its analogs as warheads, and the activity of these compounds for degradation of DYRK2 protein was evaluated in HeLa and MM231 cell lines. Compound CP134, an E3 ligase CRBN-recruiting heterobifunctional small molecule, could effectively reduce the DYRK2 protein level in a concentration-, time-, CRBN-, and UPS-dependent manner. Although the degradation efficiency did not reach the nanomolar level, CP134 expanded a new target for PROTACs and demonstrated that DYRK2 could be targeted by the TPD technology. In addition, as the first potent degrader for DYRK2, CP134 could be used as an effective tool to further evaluate its therapeutic potential and deserves further investigation.

Glossary

Abbreviations

DC₅₀

half-maximal degradation concentration

K_d

dissociation constant

UPS

ubiquitin–proteasome system

IC₅₀

half-maximal inhibitory concentration

PROTAC

proteolysis-targeting chimera

CRBN

Cereblon

MS

low-resolution mass spectrometry

HPLC

high-performance liquid chromatography

DMSO

dimethyl sulfoxide

NMR

nuclear magnetic resonance

co-IP

coimmunoprecipitation

FCM

flow cytometry

TPD

targeted protein degradation

VHL

von Hippel–Lindau protein

LSM

laser scanning microscopy

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

DYRK2

dual-specificity tyrosine phosphorylation-regulated kinase 2

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.4c00065.

• Molecular docking simulation of CUR analogs A1 and A2 in complex with DYRK2 and SPR result of CUR analogs A1 and A2 with DYRK2 (Figure S1); Western blot analyses of DYRK2, phosphorylated-p53 (Ser46), and phosphorylated-c-Myc (Ser62) in MM231 cells, SPR result of CP134 with DYRK2, induction of cytotoxicity in the MM231 cell line by CP134, CUR, and A1–A3, and proteasome activity in total cell lysates from MM231 cells (Figure S2); Experimental details about the synthesis procedure and biological experiments; Appendix A: ¹H and ¹³C NMR spectra of synthesized compounds (Appendix A); and HRMS spectra of synthesized compounds (Appendix B) (PDF)

• Docking results (ZIP)

Author Contributions

^§ J. Chen and W. Zhu are co-first authors and contributed equally to this work. The manuscript was written through contributions of all authors. All of the authors approved the final version of the manuscript.

Safety statement: No unexpected or unusually high hazards were encountered.

The authors declare no competing financial interest.