Mucin 20 modulates proteasome capacity through c‐Met signalling to increase carfilzomib sensitivity in mantle cell lymphoma

Mantle cell lymphoma (MCL) is a haematologic malignancy. The proteasome inhibitor (PI) bortezomib has been approved to treat MCL, but resistance has emerged through mechanisms that remain unclear. This study aimed to explore the mechanism of PI resistance in MCL and identify new targets for this patient subgroup. Carfilzomib‐resistant (CR) MCL cell lines and primary samples were used for both in vitro and in vivo experiments to identify gene expression and explore their related signalling pathways. We first identified mucin 20 (MUC20) suppression in carfilzomib‐resistant MCL models. MUC20 overexpression sensitized cells to carfilzomib in vitro and in vivo. MUC20 expression was inversely related to activation of c‐Met and the downstream p44/42 MAPK pathway. c‐Met activation with hepatocyte growth factor (HGF) induced PI resistance, while c‐Met inhibition restored PI sensitivity. Carfilzomib resistance and depressed MUC20 expression were associated with enhanced proteasome activity and higher expression of proteassemblin (POMP), a chaperone for catalytically active proteasome assembly. c‐Met and POMP were associated through binding and induction of MAPK‐regulated ELK1 to the POMP promoter. Our data reveal that c‐Met signalling activation enhanced proteasome capacity as a mechanism of PI resistance, and MUC20 expression may be a useful biomarker for PI therapy.

Mantle cell lymphoma (MCL) is a haematologic malignancy. The proteasome inhibitor (PI) bortezomib has been approved to treat MCL, but resistance has emerged through mechanisms that remain unclear. This study aimed to explore the mechanism of PI resistance in MCL and identify new targets for this patient subgroup. Carfilzomibresistant (CR) MCL cell lines and primary samples were used for both in vitro and in vivo experiments to identify gene expression and explore their related signalling pathways. We first identified mucin 20 (MUC20) suppression in carfilzomib-resistant MCL models. MUC20 overexpression sensitized cells to carfilzomib in vitro and in vivo.
MUC20 expression was inversely related to activation of c-Met and the downstream p44/42 MAPK pathway. c-Met activation with hepatocyte growth factor (HGF) induced PI resistance, while c-Met inhibition restored PI sensitivity. Carfilzomib resistance and depressed MUC20 expression were associated with enhanced proteasome activity and higher expression of proteassemblin (POMP), a chaperone for catalytically active proteasome assembly. c-Met and POMP were associated through binding and induction of MAPK-regulated ELK1 to the POMP promoter. Our data reveal that c-Met signalling activation enhanced proteasome capacity as a mechanism of PI resistance, and MUC20 expression may be a useful biomarker for PI therapy.

K E Y W O R D S
carfilzomib, c-Met, mantle cell lymphoma, mucin 20, resistance R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone). Bortezomib might also benefit patients with MCL as a maintenance therapy. 1 However, primary and acquired resistance to bortezomib remains a major problem. Thus, current research is focused on identifying a new generation of PIs that can overcome bortezomib resistance.
Carfilzomib is a second-generation PI that has received FDA approval for the treatment of myeloma. While bortezomib binds reversibly, carfilzomib is an epoxyketone that forms irreversible bonds, which can prolong the duration of proteasome inhibition.
Carfilzomib is both tolerable and active against relapsed and/or refractory myeloma, even in some patients who have previously received bortezomib therapy. 5,6 The effect of carfilzomib for treating MCL is currently being explored. Zhang  In this study, we aimed to explore the mechanism of PI resistance in MCL both in vivo and in vitro. Furthermore, we sought to identify a new targeted combinational therapy to improve the therapeutic effect of treatment for this subgroup of patients.

| Development of carfilzomib-resistant (CR) cell lines
Drug-naïve Granta-519 and JeKo-1 MCL cell lines were initially exposed to carfilzomib (Selleck Chemical) at 10% of the drug minimum inhibitory concentration (IC 10 ). Over eight months, drug concentrations were serially increased from 1 to 10 nmol/L following the verification of proliferative ability. Once CR Granta-519 and JeKo-1 cell lines were established, all the experiments were performed after the cells were exposed to carfilzomib-free medium for at least 7 days. All cultured cell lines had no microbial contamination.

| Patient sample collection
Bone marrow samples were obtained from MCL patients admitted to the Shengjing Hospital of China Medical University.
Informed consent was obtained from all patients, and the study was approved by the local ethics committee in accordance with the Declaration of Helsinki. Bone marrow mononuclear cells were isolated by Ficoll-Hypaque density gradient centrifugation.

| Cell viability assays
Cell viability assays were performed as described previously. 10 Carfilzomib, bortezomib and tivantinib (ARQ-197) were purchased from Selleck Chemical. The data provided were obtained from three independent experiments.

| Real-time reverse transcription-polymerase chain reaction (RT-PCR)
MUC20 assay probes (Lot.1107622) were purchased from Applied Biosystems, and RT-PCR was performed on an Applied Biosystems Step One Plus Real-Time PCR system.

| Flow cytometry
MCL CD20 + cells were analysed for MUC20 expression by multi-
Western blot analysis was performed following standard procedures. 10

| Proteasome activity assays
Chymotrypsin-like, trypsin-like and caspase-like activities were assayed using the Proteasome-Glo™ Chymotrypsin-Like, Trypsin-Like and Caspase-Like Cell-Based Assay Kits (Promega).

| Lentiviral vector technique and cell transfection
A MUC20 lentiviral expression vector was prepared using pCDH-CMV-MCS-EF1-CopGFP (System Biosciences; Mountain View, CA, USA). The open reading frame of MUC20 was PCR amplified from a cDNA clone (Origene; Rockville, MD, USA). Lentiviral particleinfected cells were sorted by flow cytometry.

| Luciferase reporter assays
The POMP promoter was amplified from genomic DNA, as follows:

| Electrophoretic mobility shift assays
The 5′ biotin-labelled wild-type (WT) probe sequence (wt-F) was as follows: GACGGACGCACTTCCGGCGGATGTGGGGGCCAGCC CTCGGAAACGGAAGTGAGCGGC. The sequence spans the POMP transcription initiation site and includes the two ELK1 sites. In the mutated ELK1 site, the sequence (mut-F) was as follows: GACGGACGCACCAGTAAGGG ATGTGG GGG CCAG CC CT TCAGA ATCTGAGTGAGCGGC. Protein-DNA binding experiments were performed using the LightShift™ Chemiluminescent EMSA Kit (Thermo Fisher Scientific).

| Chromatin immunoprecipitation
Chromatin immunoprecipitation was performed according to the manufacturer's protocol (EMD Millipore). The POMP promoter oligonucleotides used were as follows: F: CGCCGCTCACTTCCGTTT, R: TGGGTCCCTTGGGATTGC. The amplified product spanned the POMP transcription initiation site and included the ELK1 binding sites.

| Xenograft model
Granta-519/CR cells and MUC20-overexpressing Granta-519/CR cells (5 × 10 6 /mouse) were separately subcutaneously xenografted into nude mice under a protocol approved by the Institutional Animal Care and Use Facility. Mice were treated with carfilzomib (5 mg/kg) or normal saline as a vehicle via intraperitoneal injection twice weekly. Tumour volume was calculated using the formula volume = (width) 2 × length/2. The weight of the tumour was also determined. The mice were divided into the following four groups, which were injected with Granta-519/CR or Granta-519/ CR/MUC20-overexpressing (MUC20-OE) cells and treated with vehicle or CFZ.

| Statistical analysis
Statistical analyses were performed using unpaired t tests in GraphPad Prism, and p-values less than 0.05 were deemed statistically significant.

| Generation of CR MCL cells
We generated Granta-519 and JeKo-1 cell lines to explore possible mechanisms of carfilzomib resistance. These phenotypes showed stable resistance to carfilzomib ( Figure 1A). A similar pattern was also observed when these cells were exposed to bortezomib ( Figure 1B). These findings suggested that some of the mechanisms of resistance in these cells were not carfilzomib-specific and were applicable to other PI agents.

| MUC20 expression decreased in MCL cells treated with PIs
We found that MUC20 expression was suppressed in CR myeloma cell lines. The GEP data were previously deposited in GEO bank with accession ID: GSE62237. We performed RT-PCR to detect MUC20 mRNA levels in CR MCL cells. The results showed a significant reduction in  Figure 2C) and in primary MCL samples exposed to bortezomib ( Figure 2D, E). These data showed that there was varied MUC20 expression in WT and CR MCL cells, and MUC20 expression decreased when cells were treated with PIs.

| MUC20 suppression correlates with PI resistance
To determine whether MUC20 played a direct role in modulating PI sensitivity, we overexpressed MUC20 in CR cells ( Figure 3A, B).
When MUC20 was overexpressed in CR cells, sensitivity to carfilzomib ( Figure 3C) was enhanced compared to that in the control. We also sought to validate these findings in a xenograft CR MCL model.
Carfilzomib showed more activity in CR cells in the MUC20-OE xenograft MCL model compared to in the CR xenograft model and MUC20-OE xenograft MCL model treated with vehicle ( Figure 3E).
The MUC20-OE mice had significantly lower tumour weight ( Figure 3F) and volume ( Figure 3G) compared with the other mice.
MUC20 expression was not modified by CFZ therapy in MUC20overexpressing tumours. These data are consistent with the in vitro data, as presented in Figure S1. Together, these data support the notion that MUC20 expression levels are correlated with the sensitivity of MCL cells to PIs.

| MUC20 suppression activates c-Met signalling
The C-terminus of MUC20 has been reported to bind a multifunctional docking site of the MET proto-oncogene and suppress activation of some of its downstream signalling cascades. 11 We therefore used Western blot to determine whether this mechanism could also be involved in the models used in this study. Granta-519/CR and JeKo-1/CR cells showed higher levels of MET and ERK1/2 activation than control cells ( Figure 4A). Furthermore, MUC20 overexpression led to inactivity of MET and ERK1/2 ( Figure 4A), and this phenomenon was also observed in the mouse xenograft model ( Figure 3D). Since HGF is the activator of c-Met signalling ( Figure 4B), we explored whether HGF could antagonize the efficacy of carfilzomib. Indeed, HGF protected WT Granta-519 and Jeco-1 cells from the effects of carfilzomib ( Figure 4C). These findings supported the hypothesis that c-Met inhibition could restore the sensitivity of CR cells to carfilzomib. In our cell models, we found that the c-Met inhibitor ARQ-197 (tivantinib) ( Figure 4B) did not reduce the viability of WT or CR Granta-519 and Jeco-1 cells when used alone at low concentrations ( Figure 4D). However, enhanced activity was observed when cells were exposed to ARQ-197 in combination with carfilzomib. Similar results were observed in MCL primary samples from patients who were exposed to and resisted bortezomib ( Figure 4E). If the mechanism of carfilzomib resistance is through c-Met signalling, this pathway could be an attractive clinical target.

| c-Met signalling impacts proteasome catalytic activity
Carfilzomib binds constitutive proteasome β5 subunit (PSMB5), which contains the chymotrypsin-like activity (ChT-L). Thus, we first examined ChT-L activity in CR cells and found that ChT-L activity was increased in CR Granta-519 and JeKo-1 cells compared to in their WT counterparts ( Figure 5A). Moreover, the caspase-like (C-L) and trypsin-like (T-L) activities were also elevated in these two CR cell lines ( Figure 5A). MUC20 overexpression in CR cells was itself sufficient to reduce proteasome activity ( Figure 5B). Exposure of CR Granta-519 and JeKo-1 cells to HGF increased the ChT-L, C-L F I G U R E 1 Resistance of mantle cell lymphoma (MCL) cell lines to both carfilzomib and bortezomib. (A) Viability of carfilzomib(CFZ)-resistant (CR) MCL cell lines Granta-519/CR and JeKo-1/ CR compared with their wild-type (WT) counterparts. These cells were exposed to the indicated concentrations of carfilzomib or vehicle (0.1% DMSO) for 24 h, and viability was measured using the tetrazolium reagent WST-1. *p < 0.05 when comparing CR cells to WT cells for all panels. (B) These same CR cells were then exposed to the indicated concentrations of bortezomib (BTZ) or vehicle and analysed as above and T-L activity, which further supports the relationship between c-Met signalling and proteasome biology. Meanwhile, c-Met signalling blockade using ARQ-197 reduced the ChT-L, C-L and T-L activity in CR cells ( Figure 5C).

| Proteassemblin (POMP) expression is linked to c-Met activity
The finding that all three major proteasome activities were increased in CR cells suggested that the assembly of subunits into functional proteasome particles was enhanced. POMP is the key chaperone responsible for assembly of the catalytically active beta subunit rings. 12,13 According to Western blot, POMP levels were increased in the CR Granta-519 and JeKo-1 cells compared to in WT cells ( Figure 6A, B). Furthermore, HGF stimulation increased POMP expression in WT cell lines; this increase was also observed to some extent in the CR models ( Figure 6A). This finding supports the connection between POMP and the c-Met pathway. Meanwhile, pharmacologic c-Met signalling blockade with ARQ-197 reduced POMP levels in WT and CR cells ( Figure 6B).
Taken together, these data support a link between c-Met pathway signalling and assembly chaperones that control proteasome capacity.

F I G U R E 2 MUC20 expression decreased in mantle cell lymphoma (MCL) cells treated with proteasome inhibitors (PIs). (A)
Real-time polymerase chain reaction (RT-PCR) analysis of MUC20 mRNA content in carfilzomib-resistant (CR) cells compared with their wild-type (WT) counterparts. (B) Western blotting analysis of MUC20 protein levels in the WT and CR cell lines. (C) Flow cytometry analysis of membrane MUC20 expression in the WT and CR cell lines. (D, E) Flow cytometry analysis of MUC20 expression on CD20 + populations from the bone marrow (BM) of nine MCL patients (newly diagnosed and exposed to bortezomib). *p < 0.05. RR, relapse and resistance to bortezomib; ND, newly diagnosed

| ELK1 regulates POMP gene expression
To elucidate the mechanism that linked c-Met signalling and POMP, we examined transcription factor binding sites in the POMP promoter and found a consensus ELK1 binding site. This was of interest since ELK1 is a target for ERK1/2, which was found to be activated in CR cells. Western blotting demonstrated that nuclear extracts from CR cell lines were enriched for ELK1 protein ( Figure 6G). To determine whether ELK1 indeed influenced POMP expression, we performed chromatin immunoprecipitation in CR cells using an anti-ELK1 antibody, followed by PCR to detect sequences near the POMP promoter. While nonspecific IgG did not appreciably precipitate such sequences, they were comparatively enriched with anti-ELK1 antibodies, and the enrichment was even greater in CR cells. In addition, HGF increased ELK1 binding to the POMP promoter, while ARQ-197 decreased binding ( Figure 6F). Next, we used a biotin-labelled probe corresponding to the ELK1 consensus site and nuclear extracts in a gel-shift assay. We identified a protein-DNA complex in WT cells, which was present in greater abundance when extracts from CR cells were used compared to WT extracts. Furthermore, an anti-ELK1 antibody supershifted this band ( Figure 6E). Since the POMP promoter has two consensus ELK1 sites, we prepared POMP promoter reporter vectors that had one or both sites mutated ( Figure 6C

| DISCUSS ION
MUC20 is a member of the mucin gene family, and prior research has suggested that its production is correlated with the progression of IgA nephropathy and other renal injuries. 14 More recently, some studies have suggested a relationship between MUC20 and tumour pathobiology in other model systems. High MUC20 expression is considered a factor that negatively influences 5-year cancerspecific survival rates and recurrence-free survival in early-stage endometrial cancer patients. 15 Another study showed that MUC20 was overexpressed and significantly correlated with recurrence and death in colorectal cancer. 16 However, MUC20 has not yet been studied in the context of MCL, and its impact on its pathobiology and patient outcomes is unknown. In this study, we found that suppressed MUC20 expression was the most conserved and significant change in CR cells compared to in WT cells. Modulation of MUC20 expression levels using genetic approaches or pharmacologic agents targeting downstream effectors was sufficient to directly change PI sensitivity. Human MUC20 is located close to MUC4 on chromosome 3q29. MUC4 has been implicated in the pathobiology of many human cancers 17 ; furthermore, the expression and function of MUC4 in tumours vary widely and can indicate different outcomes. 18,19 Interestingly, a study found that MUC4 expression in pancreatic cancers is associated with enhanced sensitivity to bortezomib. 20 Moreover, other studies have suggested that MUC1 may play a role in bortezomib resistance, 21 and an In our study, we found that activation of c-Met with HGF-induced PI resistance, while c-Met inhibition restored PI sensitivity. We also identified for the first time that increased expression of the β subunit ring assembly chaperone, POMP, is correlated with c-Met activity.
Our results show that MUC20 expression may be used as a predictor of c-Met pathway activity and PI sensitivity in MCL. Furthermore, our data suggest that loss of MUC20 expression in the context of PI resistance could be a valuable tool to individualize MCL therapy.
MUC20 expression could be used to predicted whether patients would benefit from approaches that suppress the c-Met pathway.
In this study, we found that this decrease in MUC20 expression reduces ERK1/2 and ELK1 signalling, inhibits POMP expression and reduces proteasome capacity, thereby restoring carfilzomib sensitivity  Figure 7). This was demonstrated in cell models using drugs that suppress c-Met. More clinical trials using a combination of c-Met inhibitors with PIs may be designed to determine whether the administration of this agent to patients clinically progressing on a carfilzomib-based regimen can reverse this phenotype and restore PI drug sensitivity in MCL. Interestingly, MUC20 expression itself had no effect on Granta-915 and JeKo-1 cells in vitro ( Figure S2).
However, in the xenograft model, tumour growth was somewhat slower in MUC20-OE mice than in CR models. These data suggested that another mechanism is involved in the reversal of CFZ resistance F I G U R E 6 ELK1 regulates proteassemblin (POMP) expression and carfilzomib resistance. (A) The impact of 100 ng/mL exogenous human growth factor (HGF) for 36 h and (B) ARQ-197 at 1 µmol/L for 24 h on POMP expression levels in wild-type (WT) and carfilzomibresistant (CR) Granta-519 and JeKo-1 cells, as detected by Western blotting. (C) The sequences that were mutated to disrupt putative ELK1 binding sites in the POMP promoter. (D) Luciferase reporter assays showing the importance of the two ELK1 binding sites for POMP promoter activation in HeLa cells. These were transfected with constructs containing an empty luciferase promoter (GI3-B) or a POMP promoter luciferase reporter with intact ELK1 sites (GI3-POMP), one of the two ELK1 binding sites mutated (F-mt and H-mt), or (2 mt)." p < 0.05 vs. GI3-POMP. (E) Electrophoretic mobility shift assay (EMSA) was performed using a 5′ biotin-labelled oligonucleotide spanning the transcription initiation site of the POMP gene containing two ELK1 consensus binding sites (Biotin-POMP-P DNA). (F) Chromatin immunoprecipitation assays were performed in WT or CR cells using antibodies specific for ELK1. Nonspecific immunoglobulins (IgG) were used as a negative control, and antibodies to RNA polymerase II (RNA Pol II) were used as a positive control. The cells were pretreated with vehicle, 100 ng/mL rhHGF, or 1 μmol/L ARQ-197 for 24 h. Quantitative real-time polymerase chain reaction (PCR) analysis was used to observe the precipitation of POMP promoter sequences. The results were normalized to a 10% input of WT cells. (G) Nuclear ELK1 protein levels in CR and WT MCL cells, with Histone H3 protein as a loading control by MUC20, rather than affecting c-Met signalling, and this phenomenon mostly occurs in vivo.
Our data reveal that enhanced proteasome capacity through c-Met signalling activation is a mechanism of PI resistance.
Furthermore, our results suggest that MUC20 expression may be a useful biomarker in PI therapy.

ACK N OWLED G EM ENTS
This study was supported by the Natural Science Foundation of Liaoning Province (201800875).

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analyzed in this study.