Crizotinib‐based proteolysis targeting chimera suppresses gastric cancer by promoting MET degradation

Abstract

As one of the common malignant cancer types, gastric cancer (GC) is known for late‐stage diagnosis and poor prognosis. Overexpression of the receptor tyrosine kinase MET is associated with poor prognosis among patients with advanced stage GC. However, no MET inhibitor has been used for GC treatment. Like other tyrosine kinase inhibitors that fit the “occupancy‐driven” model, current MET inhibitors are prone to acquired resistance. The emerging proteolysis targeting chimera (PROTAC) strategy could overcome such limitations through direct degradation of the target proteins. In this study, we successfully transformed the MET‐targeted inhibitor crizotinib into a series of PROTACs, recruiting cereblon/cullin 4A E3 ubiquitin ligase to degrade the MET proteins. The optimized lead PROTAC (PRO‐6 E) effectively eliminated MET proteins in vitro and in vivo, inhibiting proliferation and motility of MET‐positive GC cells. In the MKN‐45 xenograft model, PRO‐6 E showed pronounced antitumor efficacy with a well‐tolerated dosage regimen. These results validated PRO‐6 E as the first oral PROTAC for MET‐dependent GC.

Proteolysis targeting chimera degrader PRO‐6 E specifically induced degradation of targeted proteins, and led to the suppression of tumor cell proliferation, motility, and invasiveness in MET‐positive gastric cancer models.

Abbreviations

Co‐IP

co‐immunoprecipitation

CRBN

cereblon

CRO

crizotinib

D_max

maximum degradation concentration

EMT

epithelial–mesenchymal transition

GC

gastric cancer

MS/MS

tandem mass spectrometry

POI

protein of interest

POM

pomalidomide

PRO‐6 E‐Me

methylated PRO‐6 E

PROTAC

proteolysis targeting chimera

RTK

receptor tyrosine kinase

SPPIER

separation of phases‐based protein interaction reporter

UPS

ubiquitin–proteasome system

VHL

von Hippel–Lindau

1. INTRODUCTION

Gastric cancer accounts for 5.6% of malignant cancers worldwide, and is known for low overall survival rate and poor prognosis. ¹ Despite recent advancements in diagnosis and targeted therapies, there are no effective treatment options for GC because of high tumor heterogeneity, late diagnosis, and acquired resistance. ² , ³ , ⁴ Approximately 37% of GC patients are partially amenable to RTK/RAS‐targeted therapy, because alterations in RTKs is a signature feature of GC. ⁵ , ⁶

Encoded by the proto‐oncogene MET, the RTK MET (also known as hepatocyte growth factor receptor, scatter factor receptor, and c‐Met) is among the most frequently aberrantly activated RTKs, ⁷ , ⁸ playing a major role in tumorigenesis and metastasis, especially among patients with advanced stage GC. ² , ⁹ , ¹⁰ , ¹¹ , ¹² The amplification of the MET gene has been reported in approximately 4% of GC patients, ⁵ , ⁶ while high levels of MET mRNA have been detected in 20%–30% of GC patients. ¹² Over 65% of GC biopsy samples showed high levels of MET protein, which facilitates metastatic progression. ¹³ , ¹⁴ Aberrant tumorigenic activation of MET has been considered a key driver and prognostic factor for GC. Preclinical studies strongly support MET as a potential target for MET‐dependent GC treatment. ¹² However, MET‐targeting treatments showed limited activities in the past, and there are no approved GC drugs targeting MET. Therefore, a new therapeutic regimen is in urgent need. ¹⁵ , ¹⁶

Proteolysis targeting chimera has recently been successfully applied to degrade various pathogenic proteins, including kinases, transcription factors, and membrane‐bound receptors. ¹⁷ Proteolysis targeting chimera molecules consist of two active moieties, a targeting warhead for intracellular POI and an E3 ligase ligand, connected by the linker. ¹⁸ , ¹⁹ , ²⁰ These ternary chemical complexes mediate the degradation of POI by recruiting E3 ligase to the specific target, followed by protein removal through the UPS. ²¹ Due to their “event‐driven” pharmacological mode, PROTACs afford the opportunities to reduce acquired drug resistance and toxicity ²² , ²³ , ²⁴ compared with the “occupancy‐driven” mode of conventional RTK inhibitors. ²⁵ Through the degradation of specific POIs, PROTACs could yield sustained suppression of RTKs and related signaling pathways, even for the kinase‐independent functions (e.g., scaffolding). ²⁶ , ²⁷ , ²⁸ , ²⁹ The application of the PROTAC strategy for MET‐targeted degradation holds great promise for GC treatment.

In this study, we successfully designed and synthesized a series of MET‐targeted PROTACs, based on the MET inhibitor CRO. The lead compound PRO‐6 E promoted a rapid, efficient, and durable degradation of MET, following the formation of the ternary complex and the activation of UPS. In addition, PRO‐6 E significantly inhibited the proliferation, migration, and invasion of MET‐overexpressing GC cells. The therapeutic potential and mechanism(s) of action for PRO‐6 E were extensively examined, using in vitro and in vivo models for GC. Our findings afforded a new strategy for the treatment of MET‐dependent GC, further highlighting the potential of PRO‐6 E as the first oral PROTAC for MET‐dependent cancers.

2. MATERIALS AND METHODS

2.1. Antibodies

Antibodies for MET (# 8198, 1:1000), p‐MET (Tyr1234/1235, # 3077, 1:000), GAPDH (#51332, 1:1000), CRBN (#71810, 1:1000), ERK (#4695, 1:1000), p‐ERK (Thr202/Tyr204, #4370, 1:1000), E‐cadherin (#3195, 1:1000), N‐cadherin (#13116, 1:1000), vimentin (#5741, 1:1000), HRP‐conjugated anti‐rabbit IgG (#7074, 1:5000), HRP‐conjugated anti‐mouse IgG (#7076, 1:5000), and Alexa Fluor 488 conjugated‐anti‐rabbit IgG (#4340, 1:1000) were purchased from Cell Signaling Technology. The Abs for c‐Myc (Cat# ab32072, 1:1000), ubiquitin (ab134953, 1:500), IKZF1 (ab191394, 1:1000), IKZF3 (ab139408, 1:1000), and GSPT1 (ab234433, 1:1000) were obtained from Abcam.

2.2. Cell lines and cell culture

MKN‐45 cells with MET amplification and HEK293T cells were from the National Collection of Cell Bank of the Chinese Academy of Sciences. SNU‐638 and SNU‐1 cells were all purchased from the Zhong Qiao Xin Zhou Biotechnology Co., Ltd. All GC cell lines were cultured in RPMI‐1640 medium (#11875093; Gibco), and HEK293T cells in DMEM medium (#2120620). All mediums were supplemented with 10% FBS (#10099141C), 1% penicillin, and 1% streptomycin (SV30010; HyClone). All cells were maintained under 5% CO₂ air at 37°C.

2.3. Cell viability assay

MKN45, SNU‐638, and SNU‐1 cells were seeded into 96‐well plates at the density of 5 × 10 ⁴ cells/ml overnight and were exposed to drugs. At the end of treatment, CCK‐8 reagent (#MA0218; MeilunBio) was added and the cell viability was determined using the Cytation 5 Reader (Agilent BioTek).

2.4. Western blot analysis

Cells were lysed in ice‐cold RIPA buffer, supplemented with 1% phosphatase inhibitors (#MB12707; MeilunBio) and 1% cOmplete protease inhibitors (#11697498001; Sigma‐Aldrich). The concentration of proteins was determined by the BCA Assay Kit (#P0010S). Then, 10 μg proteins were separated on 7.5% SDS‐PAGE and transferred to PVDF membranes (#10600021; GE Healthcare). The membrane was then blocked and incubated with Abs. The signals were detected using the Tanon Imaging System.

2.5. Detection of ubiquitin MET protein

MKN‐45 cells (3 × 10⁵ cells/well) were grown in 6‐well plates and exposed to PRO‐6 E (1 μM) for different time periods. Cells were lysed and the total proteins were separated by 7.5% SDS‐PAGE gel (EpiZyme). The membrane was then blocked and incubated with ubiquitin Ab. The signals were detected using the Tanon Imaging System. To verify whether the polyubiquitination process induced by PRO‐6 E (1 μM) was dependent on the ubiquitin proteasome, MKN‐45 cells were pretreated with MG132 (5 μM, 2 h) following 24 h of incubation with PRO‐6 E (1 μM). The signals were detected using the Tanon Imaging System.

2.6. Immunofluorescence

MKN‐45 cells were seeded into 35‐mm dishes (BS‐20‐GJM) and treated with PRO‐6 E for 48 h. The cells were fixed with 4% paraformaldehyde and blocked with 1% BSA. The anti‐MET primary Ab was added at 4°C overnight. Alexa Fluor 488‐conjugated goat anti‐rabbit IgG was added for 1 h and the cell nuclei were stained with DAPI (ab228549). Cells were photographed using either a confocal laser scanning microscope (Leica) or high content analysis system (PerkinElmer).

2.7. Co‐immunoprecipitation

MKN‐45 cells were seeded at the density of 4 × 10⁵ cells per well and treated with PRO‐6 E (1 μM) for 48 h. Anti‐MET primary Ab (1:1000) was added and incubated at 4°C for 12 h. Protein A/G PLUS‐Agarose (#sc‐2003; Santa Cruz) (25 μl) was added to the mix and incubated at 4°C for 2 h. The samples were collected for western blot analysis. For the inverse‐phase Co‐IP experiment, we undertook Co‐IP assays immunoprecipitating with CRBN and blotting with MET.

2.8. Microscale thermophoresis assay

The labeled MET protein (#10692‐H08H; SinoBiological) was diluted with PBST (0.05% Tween‐20 in PBS), mixed with PRO‐6 E, CRO, or PRO‐6 E‐Me that have been subjected to serial dilution from a 100 μM stock solution, and measured using Monolith NT.115 (NanoTempe).

2.9. Proteomic analysis

Protein digestion and TMT labeling were carried out following the manufacturer's instruction. ³⁰ The labeled peptides were desalinated and fractionated by high pH reversed‐phase liquid chromatography using the 1290 UPLC system (Agilent). The fractions were separated by nanoscale liquid chromatography and analyzed by on‐line electrospray MS/MS. Tandem mass spectra were extracted by Proteome Discoverer software version 2.4.1.15 (Thermo Fisher Scientific). All MS/MS samples were analyzed using SEQUEST, which was set up to search the Uniprot database. The percolator algorithm was used to control peptide level false discovery rates lower than 1%. The normalization of total peptide amount was used to correct experimental deviation.

2.10. SPPIER assay for ternary complex formation

HEK293T cells grown in 35‐mm plates (J04961; Jingan) was transfected with plasmids using Neofect DNA transfection reagent (#TF20121201) for 24 h. The cells were treated with PRO‐6 E or PRO‐6 E‐Me for 48 h. Fluorescence images were captured using a high content analysis system (PerkinElmer).

2.11. Cell cycle analysis

MKN‐45 cells seeded in 6‐well plates (4 × 10⁵ cells/well) were treated with PRO‐6 E for 48 h. The cells were collected, fixed with 70% ethanol, and stained using a cell cycle detection kit (#KGA512). The cells were analyzed by flow cytometry (Beckman Coulter).

2.12. Wound healing assay

Cells (6 × 10⁵ cells/mL) were seeded into 24‐well plates that were coated with collagen from rat tail (#40125ES10; Yeasen). When cell confluence reached approximately 90%, the wells were scratched with a sterile pipet tip and treated with PRO‐6 E as specified. The migration distance was analyzed using ImageJ.

2.13. Transwell invasion assay

MKN‐45 cells were seeded into Matrigel‐coated Transwell upper chambers (#3491; Corning). After the cells were attached, the conditioned media were replaced with serum‐free media, and the chambers were placed into 24‐well plates with 600 μl media containing 20% FBS. Noninvaded cells were removed with a cotton swab, and the invaded cells were fixed with 4% paraformaldehyde and stained with crystal violet. Images were captured and analyzed by Image‐Pro plus 6.0 software (MediaCybernetics).

2.14. Xenograft tumor models

Six‐week‐old female BALB/c nude mice were purchased from Shanghai Model Organisms Center. All studies followed the Institutional Animal Care and Use Committee guidelines of Shanghai Model Organisms Center. MKN‐45 cells (5 × 10⁶ cells) were subcutaneously injected into the right flank of mice. When tumor volume reached approximately 50 mm³, animals were randomly divided into two groups (n = 5). PRO‐6 E was given at 30 mg/kg by oral gavage in 0.5% CMC‐Na (0.9% saline) daily for 15 days. Tumor volume was calculated using the formula (a × b × b)/2, with “a” representing length and “b” representing width. Finally, tumor tissues were collected for western blot, H&E, TUNEL, and Ki‐67 analysis.

2.15. Statistical analysis

Statistical analysis was carried out using a two‐tailed Student's t‐test or one‐way ANOVA followed by Tukey–Kramer multiple comparison test with GraphPad Prism 7. Data are presented as mean ± SD. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001.

3. RESULTS

3.1. Design of PROTACs

To investigate the clinical significance of altered MET expression in GC, bioinformatic analysis was undertaken using the GEPIA2 database. ³¹ The mRNA level of MET in tumor tissues was significantly higher than that in normal tissues, and the level of MET expression correlated with poor prognosis in GC (Figure 1A,B). Immunohistochemical staining of biopsy specimens showed that the content of MET in GC tumor tissues was significantly higher than that in adjacent normal tissues (Figure 1C,D). Collectively, MET represents a potential target for GC.

FIGURE 1.

Design of proteolysis targeting chimeras (PROTACs). (A) MET expression in gastric cancer (GC) using GEPIA2 database. N, normal tissue; T, tumor tissue. (B) Overall survival of GC patients with different MET levels using GEPIA2 database. TPM, transcripts per million. (C) MET expression in GC patients. (D) Analysis of (C) (n = 3). (E, F) Cytotoxic activity of PROTACs for 72 h in MKN‐45 cells. CRBN, cereblon; VHL, von Hippel–Lindau. (G) Cytotoxic activity of GC cells after PRO‐6 E treatment for 48 h. Data are shown as mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001

To design PROTACs that mediate MET degradation, a competitive MET inhibitor, CRO, was selected as the targeting warhead because of its high affinity with MET. ³² In addition, the X‐ray crystal structure of the MET/CRO complex (PDB ID: 2WGJ) identified an available site (the nitrogen atom of piperidine moiety) for extending and connecting with linker. Two members of the E3 ubiquitin ligase complex, CRBN and VHL, have been successfully used to design various PROTACs. ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ Therefore, a set of linkers with different physicochemical properties were designed and connected to the nitrogen atom of CRO piperidine moiety for recruiting CRBN or pVHL (Figure 1E). In MKN‐45 cells, compounds coupling with CRBN showed stronger antiproliferative activities than the pVHL‐based molecules (Figure 1E). In addition, PRO‐6 E (containing flexible aliphatic chain) and 22‐153 (containing rigid quaternary ring) were superior than others. Given the lower molecular weight and bioactivities, the CRBN ligand was selected for constructing MET degraders.

A set of designed PROTACs was synthesized by changing the length of alkane chain in 22‐153, and it was found that 03‐191 had the most pronounced antitumor activity (Figure S1), indicating that the coexistence of a flexible aliphatic chain and rigid four‐member ring (with a total number of nine atoms) might be more conducive to bioactivity (Figure S1). When the position of the rigid ring was changed in 03‐191, the antiproliferative activity of 03‐191 was significantly reduced in 03‐114 (Figure S1), indicating that the position of the ring should be as far away from CRO as possible.

3.2. PRO‐6 E is a potent degrader of MET protein

Based on the cell viabilities in MKN‐45 cells, PRO‐6 E and 03‐191 (IC₅₀ < 1 μM) were selected for further evaluation (Figures 1F and S2A,B). In the MKN‐45 cell‐based protein degradation study, MET protein was degraded with a D_max of 81.9% when treated with PRO‐6 E (1 μM), in comparison to 03‐191 (Figure S2C,D). Considering the structural difference between PRO‐6 E and 03‐191 (Figure S2 E,F), the rigid ring structure in 03‐191 might have inhibited formation of the ternary complex, which led to poor MET degradation. In addition, two gastric cancer cell lines with high MET expression (SNU‐638) and low MET expression (SNU‐1) were used to test whether the antiproliferative activities of PRO‐6 E are correlated with the levels of MET in GC cells (Figure S2G). The results showed that MKN‐45 cells with the highest MET expression were indeed most sensitive to PRO‐6 E (Figure 1G). Therefore, PRO‐6 E was selected as the degrader for further investigation. The scheme of PRO‐6 E synthesis is shown in Figure 2A.

FIGURE 2.

PRO‐6 E was a potent MET degrader. (A) Synthesis of PRO‐6 E. CRBN, cereblon. (B) Molecular docking results of crizotinib (CRO, blue) or PRO‐6 E (purple) binding to MET (pink, PDB ID: 2WGJ). (C, D) Microscale thermophoresis results for the binding affinity between MET and PRO‐6 E (C) or CRO (D). (E) Cell viability detection when treated with PRO‐6 E or CRO for 48 h. Data are shown as mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001. n.s., no significant difference

Molecular docking was used to determine the binding mode between PRO‐6 E and MET, based on the X‐ray crystal structure of the MET/CRO complex (PDB ID: 2WGJ). The docking study identified two key hydrogen bonds between PRO‐6 E and Pro 1158/Met 1160 in the MET kinase domain (Figure 2B), consistent with the binding between CRO and MET (PDB ID: 2WGJ). The pyrazol ring of PRO‐6 E formed an extra H‐π conjugation with Ile 1084 in MET (Figure 2B), further enhancing the binding efficiency of PRO‐6 E (K_d ≈ 5.38 μM) when compared to CRO (Figure 2C,D). This observation suggests that the construction of PROTAC did not affect the affinity between CRO and MET (K_d ≈ 10.5 μM) (Figure 2D). In the CCK‐8 assay, PRO‐6 E and CRO showed similar antiproliferative activities against MKN‐45 cells for 48 h, with IC₅₀ values of 0.127 μM and 0.106 μM, respectively (Figure 2E).

3.3. PRO‐6 E facilitates MET degradation

In MKN‐45 cells, PRO‐6 E induced MET degradation in a concentration‐dependent manner, with a half‐maximal degradation concentration of 0.203 μM and D_max of 82.8% (Figure 3A,B). PRO‐6 E‐facilitated MET degradation is also time‐dependent, starting after 12 h of exposure (Figure 3C,D). Immunofluorescence‐based investigation revealed that PRO‐6 E promoted the internalization and degradation of MET protein (Figure 3G,H), consistent with previously reported. ¹⁷ To investigate the duration of PRO‐6 E's effect on MET degradation, PRO‐6 E was added to MKN‐45 cells, washed out, and the levels of MET protein monitored over a period of 72 h. The inhibition on MET or p‐MET protein accumulation lasted for at least 72 h after PRO‐6 E removal (Figure 3E,F). To exclude the possibility that “off‐target action” is responsible for PRO‐6 E‐imposed MET inhibition, the effect of PRO‐6 E on a number of newly identified CRBN substrates was examined. ⁴⁰ , ⁴¹ , ⁴² While PRO‐6 E promoted MET degradation, it only weakly affected IKZF1, IKZF3, and GSPT1 protein levels (Figure S3A,B). Moreover, we used MKN‐45 cells with higher concentrations of PRO‐6 E (5 μM) and observed the hook effect (Figure S3C,D).

FIGURE 3.

PRO‐6 E was a MET degrader. (A) MET expression following PRO‐6 E or crizotinib treatment for 48 h. (B) Analysis of (A) (n = 3). (C) MET expression following PRO‐6 E (1 μM) treatment. (D) Analysis of (C) (n = 3). (E) MET and p‐MET expression. Following PRO‐6 E (1 μM, 48 h) treatment, the compound was removed, and the incubation continued for the specified amount of time. (F) Analysis of (E) (n = 3). (G) MET expression in PRO‐6 E treated (48 h) cells. (H) Internalization and degradation of MET following PRO‐6 E (1 μM) treatment for 48 h. ***p < 0.001. n.s., no significant difference

3.4. PRO‐6 E induces the formation of ternary complex to achieve ubiquitination degradation of MET

To further confirm the formation of a ternary complex induced by PRO‐6 E during MET degradation, we first investigated whether the effect of PRO‐6 E could be competitively antagonized by excessive CRBN ligand or targeting warhead CRO. It was found that MET depletion induced by PRO‐6 E was reversed by excess POM or CRO, whereas POM or CRO alone had no effect on MET expression (Figure 4A,B,D,E), indicating that the degradation effect of PRO‐6 E was dependent on the recruitment of CRBN and MET. Furthermore, in the presence of MG132, a protease inhibitor, the PRO‐6 E‐induced MET degradation was suppressed, indicating the proteasome‐dependent manner (Figure 4C,F). To determine PRO‐6 E's dependence on CRBN in the degradation of MET protein, we then knocked down CRBN using three CRBN siRNA (si‐CRBN) (Figure S4). Knockdown of CRBN significantly alleviated the inhibition effect of PRO‐6 E on MKN45 cells (Figure 4G), indicating the important role of CRBN in PRO‐6 E‐induced antitumor activity. In addition, the E1 inhibitor (MLN7243) or NEDD8 inhibitor (MLN4924) have also been used to further reinforce these claims. Pretreatment with MLN7243 or MLN4924 obviously alleviated the inhibition effect of PRO‐6 E (Figure 4H,I).

FIGURE 4.

PRO‐6 E induced the formation of ternary complexes. MET degradation was abolished when cotreatment with pomalidomide (POM; 15 μM) (A, D) or crizotinib (CRO; 1 μM) (B, E) with PRO‐6 E (1 μM) for 48 h. Pretreatment with MG132 (5 μM, 2 h) (C, F) abolished MET degradation induced by PRO‐6 E (1 μM, 48 h) (n = 3). Pretreatment with cereblon siRNA (si‐CRBN) (G), MLN4924 (0.2 μM, 2 h) (H), or MLN7243 (1 nM, 1 h) (I) abolished the cell viability inhibition induced by PRO‐6 E (0.3 μM, 48 h) (n = 3). (J) Formation of ternary complex following treatment with PRO‐6 E for 48 h. (K) Analysis of (J). (L, M) MKN‐45 cells were treated with PRO‐6 E (1 μM) for 12 h and the ternary complex was detected by co‐immunoprecipitation (IP). (N) Ubiquitination (Ub) of MET following treatment with PRO‐6 E (1 μM). (O) Expression of MET following treatment with PRO‐6 E (1 μM). (P) Ubiquitination of MET after pretreatment with MG132 (5 μM, 2 h) following PRO‐6 E (1 μM, 24 h) incubation. *p < 0.05, **p < 0.01, ***p < 0.001. n.s., no significant difference

Considering that the formation of a ternary complex plays a vital role in PROTAC‐induced degradation, we utilized a GFP phase transition‐based cellular assay named SPPIER, a simple signal pattern, to show in a straightforward and robust manner the protein–protein interaction induced by PRO‐6 E. ⁴³ Based on the principle of SPPIER, we generated two plasmids by genetically fusing HO‐Tag3 to CRBN and HO‐Tag6 to MET. Coexpressing those two plasmids and adding PRO‐6 E in HEK293 cells, we observed the EGFP droplet‐shaped green fluorescence in a background of homogeneous fluorescence, which was weaker at lower concentration (1 μM) and stronger at higher concentration (5 μM) of PRO‐6 E (Figure 4J,K). To facilitate the degradation mechanism of PRO‐6 E, we synthesized PRO‐6 E‐Me. ³⁴ , ⁴⁴ The microscale thermophoresis assay showed that PRO‐6 E‐Me still had comparable high affinity with MET (K_d ≈ 16.2 μM) (Figure S5C), while the western blot assays showed that PRO‐6 E‐Me had no obvious effects on MET expression when compared to PRO‐6 E (Figure S5D,E,F). The CCK‐8 assays also showed that PRO‐6 E‐Me had a weak antiproliferation effect on tumor cells compared with PRO‐6 E (Figure S5A,B). In addition, the SPPIER technique was applied to visualize PRO‐6 E‐Me‐induced protein–protein interaction. As expected, the green fluorescence droplet formation was not observed in PRO‐6 E‐Me‐treated cells (Figure S5G,H). Taken together, PRO‐6 E formed the MET‐PRO‐6 E‐CRBN ternary complex in the process of MET degradation, which was highly dependent on the E3 ubiquitin ligase CRBN, and the antitumor effect of PRO‐6 E on GC cells is MET degradation‐dependent. We also examined the ability of PRO‐6 E to form a ternary complex with endogenic MET and CRBN in MKN‐45 cells. Immunoprecipitation with anti‐MET and immunoblotting revealed a robust PRO‐6 E‐induced ternary complex formation with endogenous MET protein (Figure 4L,M).

Western blot assays were further used to determine the ubiquitination degradation of MET induced by PRO‐6 E. When the MKN‐45 cells were treated with PRO‐6 E (1 μM) for 12 h, the formation of ubiquitin markers was significantly observed (Figure 4N,O). The absence of ubiquitinated labeled proteins at 24 h could be due to the rapid degradation of ubiquitinated proteins in the proteasome, as previously reported. ⁴⁵ , ⁴⁶ Therefore, MG132 was used to observe the ubiquitin proteins. In the presence of MG132, the content of ubiquitin proteins induced by PRO‐6 E increased significantly, supporting a proteasome‐dependent manner (Figure 4P). All of these results indicated that PRO‐6 E hijacked the UPS to facilitate MET degradation. ²⁸

3.5. PRO‐6 E inhibits cell proliferation, migration, and invasion

TMT‐labeled quantitative proteomics analysis was used to determine the effect of PRO‐6 E on the biological function of MKN‐45 cells. Among the approximately 5500 quantified proteins, proteins related to the biological processes and molecular functions of intercellular adhesion and cell proliferation were significantly reduced by more than 2‐fold by PRO‐6 E (Figure 5A), indicating the potential of PRO‐6 E to effectively inhibit tumor cell proliferation and metastasis through MET degradation. Arrested cell cycle progression is an integrated result of impaired cell proliferation and metabolism. ⁴⁷ We then examined the cell cycle by flow cytometry. As shown in Figure 5B, PRO‐6 E could cause a significant increase in the percentage of G₂/M phase cells and decrease in the percentage of S phase cells.

FIGURE 5.

PRO‐6 E inhibited cell growth. (A) Multiple quantitative proteomics analysis of MKN‐45 cells following treatment with PRO‐6 E (1 μM) for 48 h. (B) Cell cycle analysis following treatment with PRO‐6 E for 48 h (n = 3). (C) Migration or invasion assays when cells were treated with PRO‐6 E. (D, E) Analyses of (C) (n = 3). (F, G) Protein expression (F) and analysis (G) when cells were treated with PRO‐6 E for 48 h (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001. n.s., no significant difference

Abnormal activation of MET is usually associated with the promotion of EMT, migration, and invasion effects. ⁴⁸ , ⁴⁹ , ⁵⁰ In the wound healing assay and Transwell invasion assay, the migration and invasion ability of MKN‐45 cells were impaired in a dose‐dependent manner with PRO‐6 E treatment (Figure 5C,E). The detection of key proteins in the EMT process showed that PRO‐6 E could significantly increase the expression of epithelial marker E‐cadherin, accompanied by the decreased mesenchymal markers N‐cadherin and vimentin (Figure S6A,B), resulting in the obvious changes of cell morphology from a dispersed state to a tightly aggregated state (Figure S6C). In addition, PRO‐6 E induced cell cycle arrest in G₂ phase of SNU‐638 cells (Figure S7A). We also found that the invasion ability and protein expression of SNU‐638 cells were also impaired with PRO‐6 E treatment (Figure S7B–D).

MET protein relies on its downstream multiple signaling pathways to play a carcinogenic role; of these, the ERK–MAPK pathway is the main mechanism responsible for tumor cell proliferation, cycle progression, and movement. Additionally, it had been proven that the activation of MET could lead to upregulated expression of c‐Myc and p‐ERK in MET‐addicted cells, promoting carcinoma proliferation and moveability. ⁴⁷ , ⁵¹ , ⁵² The previously mentioned key regulatory proteins were then examined, and treatment with PRO‐6 E could effectively degrade MET protein and reduce the level of key downstream proteins, inducing c‐Myc and p‐ERK (Figure 5F,G).

Taken together, these data indicated that PRO‐6 E was an effective MET degrader with significantly inhibitory effects on c‐Myc and p‐ERK, thus suppressing the infiltration and metastasis of tumor cells.

3.6. PRO‐6 E induces efficient inhibition of tumor growth

To evaluate the antitumor effects and future clinical translation of PRO‐6 E in MKN‐45 cell‐derived xenograft tumor models, we treated the mice with PRO‐6 E at 30 mg/kg/day for 2 weeks with oral administration of CRO (Figure 6A). ⁵³ , ⁵⁴ , ⁵⁵ PRO‐6 E could effectively inhibit the tumor growth (Figure 6C–E) without obvious toxicity (Figure 6B). Pathological and immunohistochemical analysis of tumor slices and major organ tissue slices proved that PRO‐6 E could induce tumor necrosis, decrease the ratio of Ki‐67‐positive cells and the expression of MET, and increase the ratio of TUNEL‐positive cells (Figure 6H,I), while no prominent organs damage could be observed (Figure 6J). It was also found that PRO‐6 E could significantly decrease the expression of MET and subsequently downstream proteins, such as c‐Myc and p‐ERK, achieving successful inhibition of tumor growth (Figure 6F,G).

FIGURE 6.

PEO‐6 E inhibited tumor growth in vivo. (A) Schematic diagram. (B, C) Bodyweight (B) and tumor volumes (C) changed in mice (n = 5). (D) Tumor weight (n = 5). (E) Tumor images. (F) Protein expression. (G) Analysis of (F) (n = 3). (H) Typical pictures of H&E, TUNEL, Ki‐67, and MET in tumor tissues. (I) Analysis of (H) (n = 3). (J) Pathological staining of normal tissues. ***p < 0.001

4. DISCUSSION

Dysregulation of MET is associated with tumor invasion, metastasis, recurrence, and poor survival. ⁵⁶ MET is also one of the most important mechanisms of acquired resistance in multiple RTK targeted therapies, ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ and attracted extensive attention as a proven therapeutic target. A series of Abs and small molecule inhibitors targeting MET have been reported to have excellent antitumor effects in phase I/II clinical trials. However, the evidence of successful phase III clinical trials is still lacking. ¹⁵ , ⁶¹ , ⁶² , ⁶³ The current strategy to block the phosphorylation of MET with Abs or inhibitors could not completely suppress the oncogenic function of MET due to its nonkinase functions, resulting in kinase recombination effects, reactivation, ⁶⁴ , ⁶⁵ , ⁶⁶ and subsequent acquired resistance. ⁵⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² However, the strategy of using mRNA to interfere with the transcription of MET showed successful inhibiting effects on tumor cells, ⁷³ , ⁷⁴ , ⁷⁵ but the poor drug forming properties of MET mRNA agents severely limited its clinical translation.

As a novel drug development strategy, PROTACs could comprehensively inhibit the pathogenic roles of targeted proteins through ubiquitination degradation. Recently, two leading PROTACs, ARV‐110 and ARV‐471, have completed phase II clinical trials with encouraging results. ²¹ Compared with conventional kinase inhibitors, the “event‐driven” mode and catalytic properties of PROTACs could help to resolve the therapeutic constraints imposed by target protein amplification and mutation. ²⁶ , ³⁷ , ⁷⁶ Moreover, PROTACs could induce the complete suppression of targeted protein functions through effective degradation. ¹⁷ , ⁷⁷ , ⁷⁸ , ⁷⁹ Although resistance to PROTACs has previously been shown to involve genomic alterations of E3 ligase complexes and upregulation of the drug efflux pump MDR1, ⁸⁰ , ⁸¹ the enhanced therapeutic effects of PROTACs could reduce the risks of long‐term and high‐dose induced drug resistance and toxicity. ²⁸ , ²⁹ , ⁷⁸

Foretinib‐based PROTAC has been developed as a c‐Met degrader in GC cell lines, highlighting the potential advantage of degradation. ¹⁷ However, foretinib lacked efficacy in unselected patients with metastatic GC. Additionally, foretinib has not been approved for clinical application, and no further clinical trial has started since 2014, ⁸² which raises concerns. In contrast, there is no antitumor effect reported in vivo in the previous research of foretinib‐based PROTACs. ¹⁷ Herein, we successfully developed a series of crizotinib‐based PROTACs targeting MET protein (Figure 7). Among them, PRO‐6 E showed effective and durable MET degradation, which inhibited downstream c‐Myc and p‐ERK protein expression, thus triggering a great suppression of tumor cell proliferation and motility without obvious cytotoxicity in vitro and in vivo. These findings provided important preliminary data for the further development of oral PROTACs targeting MET, thus paving the way for future clinical treatment of MET‐dependent cancer.

FIGURE 7.

Proteolysis targeting chimera (PROTAC) degrader PRO‐6 E induced degradation of targeted proteins, and led to the suppression of tumor cell proliferation, motility, and invasiveness in MET‐positive gastric cancer models. Ub, ubiquitin

AUTHOR CONTRIBUTIONS

JJC, JMJ, and WJG contributed to data curation, formal analysis, investigation, methodology, writing – original draft, and writing – review and editing. ZZ and HY contributed to data curation, methodology, and software. YDZ, DGN, QLX, and XMZ contributed to data curation and manuscript editing. QYS, YW, WDZ, and XL contributed to conceptualization, resources, supervision, funding acquisition, and writing – review and editing.

CONFLICT OF INTTEREST STATEMENT

None of the authors of this manuscript are current editors or editorial board members of Cancer Science. The authors declare no competing interests.

ETHICS STATEMENT

Approval of the research protocol by an institutional review board: N/A.

Informed consent: N/A.

Registry and registration no. of the study/trial: N/A.

Animal studies: All animal studies followed the Institutional Animal Care and Use Committee guidelines of Shanghai Model Organisms Center.

Supporting information

Figure S1 Chemical structures and cytotoxicity of proteolysis targeting chimeras designed based on 22‐153 and PRO‐6 E in MKN‐45 cells.

Figure S2 Antitumor efficacy of 03‐191 and PRO‐6 E.

Figure S3 Protein detection.

Figure S4 Protein detection of cereblon (CRBN).

Figure S5 Mechanism of action of PRO‐6 E.

Figure S6 Protein expression and cell morphology.

Figure S7 PRO‐6 E inhibited cell cycle, invasion, and protein expression.

Synthetic routes of PROTACs.

Chen J‐J, Jin J‐M, Gu W‐J, et al. Crizotinib‐based proteolysis targeting chimera suppresses gastric cancer by promoting MET degradation. Cancer Sci. 2023;114:1958‐1971. doi: 10.1111/cas.15733