Eltrombopag binds SDC4 directly and enhances MAPK signaling and macropinocytosis in cancer cells

Chemicals

Eltrombopag (SB-497115) was purchased from SelleckChem (Shanghai, China). The FDA-approved drug library (L1021) was purchased from APExBIO (Shanghai, China).

Cell culture

Pancreatic cancer cell lines AsPC1 (TCHu8), BxPC3 (TCHu12), CFPAC1 (TCHu112), and PANC1 (TCHu98) and the colon cancer cell line HCT116 (TCHu99) were purchased from the National Collection of Authenticated Cell Cultures (Shanghai, China). AsPC1 and BxPC3 cells were cultured in RPMI 1640 medium (Thermo Fisher Scientific) and PANC1 cells were cultured in DMEM medium (Thermo Fisher Scientific). The medium was supplemented with 10% fetal bovine serum (FBS) (Gibco, Gaithersburg, MD) and 1% penicillin-streptomycin (Hyclone, Logan, UT). All cell lines were maintained at 37°C in a humidified atmosphere consisting of 5% CO₂.

Generation of SDC4 KO cells through CRISPR/Cas9-mediated gene editing

The SDC4 gene was deleted in PANC1 and HCT116 cells by CRISPR/Cas9 genome editing. A plasmid pL-CRISPR.EFS.GFP (#57818, Addgene) encoding both the Cas9 protein and the sgRNA was used. The Cas9 sequence is coupled to a P2A site and EGFP. Expression of the Cas9 protein results in simultaneous expression of EGFP, allowing for the selection of positively transfected cells. Two sgRNAs targeting exon4 and exon5 of the SDC4 gene were designed using the optimized CRISPR design online tool (http://crispr.cos.uni-heidelberg.de). The oligo sequences for the sgRNAs are listed below: sgRNA1 for exon4 forward, 5’-CACCGCACCGAACCCAAGAAACTAG-3’, reverse, 5’-AAACCTAGTTTCTTGGGTTCG-GTGC-3’; sgRNA2 for exon5 forward, 5’-CACCGTCTCTGCCTGGGCAAGAGTG-3’, reverse, 5’-AAACCACTCTTGCCCAGGCAGAGAC-3’. Plasmids were then sequenced by Sangon Biotech Company (Shanghai, China) to check the right insertion. Two CRISPR/Cas9 plasmids were co-transfected in equimolar ratio into cells using Lipofectamine3000 (L3000075, Invitrogen). After transfection for 48 h, EGFP-positive single cells were sorted into 96-well plates and the plates were kept in the incubator for 7-14 days. Genomic DNA was extracted with QuickExtract DNA Extraction Solution (QE09050, Epicenter, Madison, WI) for PCR identification followed by Sanger-sequencing. The sequences of PCR primers are shown as below: forward, 5’-GCAGCATAATTGTGGAGA-3’; reverse, 5’-CTGTGGAAATGTGCGAGA-3’.

Western blot analysis

Cells were lysed using RIPA (#9806, Cell Signaling Technology) containing protease inhibitors (#36978, Thermo Fisher Scientific) at 4°C and proteins were quantified by the BCA protein quantification kit (#23227, Thermo Fisher Scientific). Proteins (10-30 μg per lane) were separated by SDS-PAGE and then transferred onto polyvinylidene fluoride membranes (#P0807, Millipore). Subsequently, the membranes were blocked with 5% non-fat milk for an hour before incubation with primary antibodies overnight at 4°C. The antibody against SDC4 (NB110-41551) was purchased from Novus, Antibodies against SDC1 (12922), p-ERK (4370S), ERK (4695S), p-AKT (4060S), AKT (9272S), p-P38 (4511S), P38 (8690S), and GAPDH (D16H11) were purchased from Cell Signaling Technology. Finally, the membranes were incubated with HRP-goat antirabbit secondary antibodies (1:10000, 7074S, CST) for 1 hour. Enhanced chemiluminescence detection reagent (#P10300, NCM Biotech) was used to visualize the signal strength of the bands.

Macropinocytosis

The macropinocytic index was determined according to the protocol previously described [47]. Briefly, cells were seeded into 48-well plates (Costar) for 24 hours, before being serum-starved for 12 hours and subsequent incubation with 1 mg/ml TMR-dextran (#T1162, Sigma) for 35 minutes at 37°C. At the end of the incubation period, cells were rinsed five times in cold PBS and immediately fixed in 4% polyformaldehyde solution for 15 minutes. Cells were mounted with 0.1 mg/ml DAPI (#C1002, Beyotime) for nuclear staining. Images were captured with an ImageXpress Micro 4 microscope (Molecular Devices) using standard settings. Mean fluorescence intensity was determined by calculating the integrated signals from at least 10 fields that were randomly selected from different regions across the entirety of each sample.

Quantitative RT-PCR (qRT-PCR)

Total RNA was extracted with TRIzol (T9424, Sigma) according to the manufacturer’s instructions. The reverse transcription reaction was performed with the 5x Evo M-MLVRT Master Mix (AG11706, Accurate Biology). Expression of the indicated genes was assessed with the QuantStudio5 real-time PCR instrument (Applied Biosystems) using the Hieff qPCR SYBR Green Master Mix (Low Rox Plus) kit (11202ES03, Yeasen). Reaction plates were incubated in a 384-well thermal cycling plate at 95°C for 10 min and then underwent 40 cycles of 10 s at 95°C and 30 s at 60°C. All reactions were performed in triplicates. Relative quantitation was calculated using the 2^-ΔCt method, where ΔCt symbolizes the change in Ct between the sample and reference mRNA. The ologo sequences for PCR primers are: SDC4, 5’-GGCAGCTCTGATTGTGGGT-3’ (forward), 5’-CATACGGTACATGAGCAGTAGGA-3’ (reverse); GAPDH, 5’-ACAACTTTGGTATCGTGGAAGG-3’ (forward), 5’-GCCATCACGCCACAGTTTC-3’ (reverse).

Flow cytometry

Cells were resuspended in the Non-enzymatic Cell Dissociation Buffer (C5789, Sigma). The cell pellet was then resuspended in cold PBS with 1% (w/v) BSA. To measure surface populations of SDC4, resuspended cells (>10,000) were incubated with FITC-conjugated anti-SDC4 or its isotype-matched control antibody (sc-12766, Santa Cruz) for 15 minutes on ice and processed for flow cytometry analysis following the manufacturer’s instructions. To measure total SDC1, resuspended cells were immediately fixed in PBS containing 1.6% polyformaldehyde and permeabilized in 0.5% saponin before incubation with conjugated antibodies. Flow Cytometry (CytoFLEX S, Beckman) data were analyzed using the CytExpert software according to the manufacturer’s specifications. SDC4 protein expression data were obtained from two independent experiments.

Protein preparation and mass spectrometry analysis

Cell lysate was sonicated and centrifuged to pellet cellular debris. Lysate protein concentrations for all samples were determined by the Pierce BCA Assay (Thermo Fisher Scientific, Franklin, Massachusetts) per the manufacturer’s instructions. Total protein was reduced with 10 mM 1,4-dithiothreitol (DTT) at 37°C for 1 h and subsequently alkylated in 20 mM iodoacetamide (IAA) for 30 min at room temperature in the dark. Proteins were digested with trypsin (1:50, w/w) at 37°C overnight. All the tryptic peptides in the samples were desalted on a Sep-pak C18 cartridge column and then lyophilized under vacuum.

The vacuum-dried samples were resuspended in 0.1% FA for liquid chromatography (LC)-MS/MS analysis. Each sample of peptides was loaded onto a C18 trap column (75 μm ID×2 cm, 3 μm, Thermo Scientific) and then separated on a C18 analytical column (75 μm ID×50 cm, 2 μm, Thermo Scientific). Peptides were separated and analyzed on an Easy-nLC 1200 system coupled to a Q-Exactive HF-X Hybrid Quadrupole-Orbirap Mass spectrometer system (Thermo Fisher Scientific). Mobile phase A (0.1% formic acid in 2% ACN) and mobile phase B (0.1% formic acid in 98% ACN) were used to establish a 120 min gradient composed of 1 min of 5% B, 106 min of 5-28% B, 2 min of 28-38% B, 1 min 38-90% B, 10 min of 90% B at a constant flow rate of 250 nL/min at 55°C. MS data were acquired with the instrument operating in the data dependent mode. Peptides were then ionized by electrospray at 2.2 kV. Full-scan MS spectra (from m/z 375 to 1500) were acquired in the Orbitrap at a high resolution of 120,000 with an automatic gain control (AGC) of 3×10⁶ and a maximum fill time of 20 ms. The twenty most intense ions were sequentially isolated and fragmented in the HCD collision cell with normalized collision energy of 27%. Fragmentation spectra were acquired in the Orbitrap analyzer with a resolution of 15,000. Ions selected for MS/MS were dynamically excluded for a duration of 30 s.

The raw data were processed using MaxQuant with the integrated Andromeda search engine (v.1.5.4.1) against the human protein database (release 2016_07, 70630 sequences) with a common contaminant database (215 entries). Enzyme was set to trypsin allowing N-terminal cleavage to proline and two missed cleavages were allowed. Database searches were performed with the following parameters: precursor mass tolerance was up to 10 ppm and the product ion mass was up to 0.02 Da. Cysteine carbamidomethylation was set as a fixed modification and N-terminal acetylation, oxidation (M), and deamidation (NQ) were set as variable modifications. A false discovery rate (FDR) of 0.01 was required for proteins and peptides.

Microscale thermophoresis (MST)

Ten million HEK293T cells overexpressing EGFP alone or EGFP-tagged SDC1, SDC2, SDC3, or SDC4 were lysed in 0.5 mL RIPA (#9806, Cell Signaling Technology) containing protease inhibitors (#36978, Thermo Fisher Scientific). Cell lysates were diluted in PBS buffer to a final concentration at which EGFP fluorescence signals were suitable for detection on the Monolith NT.115 instrument (NanoTemper Technologies). A collection of 1,363 FDA-approved drugs was used for binding check screening. For binding affinity detection, a 10 µL protein sample was mixed with a 10 µL ligand solution at the appropriate concentrations. The purified proteins were labeled by the Red-NHS kit following the manufacturer’s protocol (MO-L011, NanoTemper Technologies). Then the mixture solutions were loaded into NT.115 standard coated capillaries or premium coated capillaries (NanoTemper Technologies). MST measurements were performed at 25°C. The fluorescence signal during thermophoresis was monitored and the change in fluorescence was analyzed by the software. Kd values were calculated by fitting a standard binding curve to the series of diluted ligands.

Recombinant protein purification

The cDNA encoding human SDC4 (P31431) was synthesized (Genscript, China) and cloned into pGEX-6P-1 to enable N-terminal glutathione S-transferase (GST)-tagging of the SDC4 protein. The construct was transformed into the expression host E. coli stain BL21 (DE3) and the cells were grown in YEP medium at 37°C until OD₆₀₀ of 0.6. The cells were then induced by adding 0.1 mM isopropyl-β-D-l-thiogalactopyranoside (IPTG) to the culture and grown at 18°C overnight. Cells were harvested by centrifugation. The cell pellet was resuspended in a buffer containing 20 mM Tris (pH 7.5), 500 mM NaCl, and 2 mM DTT and lysed by sonication. After centrifugation, the clarified cell lysate was incubated with glutathione-sepharose 4B beads. GST-tagged SDC4 was eluted with the buffer consisting of 50 mM reduced glutathione. Protein fractions were collected in buffer containing 20 mM Tris (pH 7.5), 150 mM NaCl, and 2 mM DTT and used for the indicated assays.

CCK-8 assays

The CCK-8 (Cell-Counting Kit-8) assay was used to detect the cell proliferation and ligand cytotoxicity following the manufacturer’s protocol (APExBIO). Cells were seeded into 96-well plates at a density of 5,000 cells/well overnight at 37°C. The cells were then treated with different concentrations of ligands for 24 hours and 72 hours, respectively. Then 10 μL CCK-8 reagent was added into each well and incubated for another 2 hours, OD₄₅₀ was measured by a microplate reader (Spark, TECAN).

Colony formation assays

Colony formation assays were performed according to the protocol described previously [48]. Briefly, ~200 single cells were seeded into each well in a 6-well plate. Medium was changed every 3 days for 2 weeks. Colonies that formed were fixed with PBS containing 4% formaldehyde and stained with 0.005% crystal violet.

Xenograft tumor assays

Male BALB/c nude mice (four weeks old) were purchased from Sino-British SIPPR/B&K Lab Animal Ltd (Shanghai, China) and housed under pathogen-free conditions. PANC1-SDC4-WT or PANC1-SDC4-KO cells (1×10⁷) were injected subcutaneously into the left flank of nude mice. Tumors were measured with a caliper and volume was calculated using the formula V = (width² × length)/2. Mice were sacrificed when tumor volumes reached 300 mm³.

Statistical analysis

The results were expressed as mean ± SD. Data analysis was performed using GraphPad Prism version 7.

Establishing SDC4 knockout (KO) cancer cells

SDC4 abnormal expression has been previously linked to tumor progression and poor prognosis in various malignant tumors [8,31-33]. Here, we further evaluated SDC4 expression using the Cancer Genome Atlas (TCGA) dataset. The results indicated that SDC4 was clearly highly expressed in pancreatic adenocarcinoma (PAAD) and colon adenocarcinoma (COAD) compared to normal tissues (Figure 1A, 1B). High SDC4 expression was also significantly associated with shorter overall survival (OS) of PAAD patients (Figure 1C, 1D), suggesting that SDC4 is a potential positive regulator of PAAD and COAD.

To validate our TCGA data and determine whether SDC4 is required for tumor progression, we knocked out SDC4 in cancer cells via the CRISPR-Cas9 gene-editing technology (Figure 1E). The sgRNAs were designed to target the cytoplasmic domain that is important for intracellular signal transduction (Figure 1E). We obtained two SDC4 KO pancreatic PANC1 cell clones and one SDC4 KO colorectal HCT116 cell clone. Genomic DNA sequencing confirmed the deletion of exon4 and exon5 in the SDC4 locus (Figure 1F). SDC4 KO was further confirmed by RT-PCR assays of SDC4 mRNA expression (Figure 1G, 1H), and western blotting (Figure 1I, 1J) and flow cytometry analysis of SDC4 proteins (Figure S1C). Notably, the expression level of SDC1 protein was almost unchanged in SDC4 KO cells (Figure 1I, 1J), supporting the specificity of the SDC4-targeting sgRNAs. Of note, the molecular weights of SDC4 and SDC1 were lower than expected in western blotting assays (Figure S1), indicating that SDC4 could be cleaved in the cells tested as described in previous studies [34-36].

SDC4 is required for tumor initiation and growth

SDC4 KO cells exhibited no significant differences in culture, both in their epithelial morphology and growth rates (Figure S2A, S2B). To determine whether SDC4 is required for tumorigenic activity, we performed colony-forming assays and found that SDC4 KO markedly impaired the colony-forming ability of PANC1 cells (Figure 2A), while no difference was detected in SDC4 KO HCT116 cells (Figure 2E). Furthermore, SDC4 KO markedly inhibited the growth of both PANC1 and HCT116 xenograft tumors (Figure 2B, 2C, 2F, 2G). Notably, SDC4 KO also significantly decreased the ability of tumor initiation (Figure 2D, 2H).

SDC1 has been reported to be a critical mediator of macropinocytosis, a regulated form of endocytosis for nutrient salvage to sustain uncontrolled growth [7]. And SDC4 overexpression cells exhibit a two-fold increase in peptide internalization [37]. To investigate whether SDC4 is also associated with macropinocytosis in PDAC cells, we performed the tetramethylrhodamine-labeled dextran (TMR-dextran) uptake assay in SDC4 KO PANC1 cells, and detected a clear decrease (~50%) of macropinocytosis compared with wide-type PANC1 cells (Figure 2I, 2J). Decreased macropinocytosis was also observed in SDC4 KO HCT116 cells (Figure S2C, S2D). Collectively, these findings confirm the importance of SDC4 in tumor progression.

SDC4 is associated with cell metabolism

Although SDC4 has been implicated in various signaling pathways [38], loss-of-function studies that validate such roles of SDC4 remain lacking. Taking advantage of the SDC4 KO cells, we next performed proteomic profiling analysis of wide-type and SDC4 KO PANC1 cells. A total of 6,201 proteins were detected among three cell lines with two repeat experiments. To identify significantly changed proteins, we performed a Student’s t-test and used the threshold of >1.5 fold change to filter proteins. A total of 249 and 320 differentially expressed proteins were identified in SDC4-KO1 and SDC4-KO2 cells, respectively. Specifically, 89 proteins are shared by both SDC4 KO cell clones (Figure 3A; Table S1).

To investigate the signature of the changed proteins, we performed gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis (Figure 3B, 3C). According the top 10 GO terms, most of the changed proteins are involved in biological processes including NAD biosynthetic process, oxidative stress, cell adhesion, and viral entry (Figure 3B). The major cellular localizations of these proteins are extracellular exosome, cytosol, and membrane raft (Figure 3B). Functionally, these proteins have various activities, such as nicotinate nucleotide diphophorylase activity, protein binding, and virus receptor activity (Figure 3B). The most enriched KEGG pathways are metabolic pathways, biosynthesis of amino acids, and nicotinate and nicotinamide metabolism, indicating that SDC4 mainly affects cellular metabolic pathways (Figure 3C). This is supported by the finding that the top 12 altered proteins are all involved in the metabolic pathway essential for cell energy and material metabolism (Figure 3D). Taken together, our proteomic profiling data indicate a strong association of SDC4 with metabolism.

ETBP directly targets SDC4 at the transmembrane motif

Given that SDC4 is an intrinsically disordered protein, conventional structure-based rational design is not suitable for identifying SDC4-targeting small molecules [21,22]. To this end, we took advantage of the microscale thermophoresis (MST) assay that enables the examination of direct engagement between small molecules and cellular target proteins in biological liquids [39]. We thus prepared cell lysates from HEK293T cells expressing GFP-tagged SDC4 for single-point binding-check screening (Figure 4A). Within an FDA-approved drug library, ETBP was the compound with the biggest response (R) value (Figure 4A). Notably, ETBP was originally identified as an agonist of the transmembrane protein thrombopoietin receptor (TPOR) [23]. Direct engagment of ETBP with SDC4 was confirmed by the serial dilution MST assays where a Kd value of 2.2 µM was obtained (Figure 4B). To further verify the direct interaction, we purified GST-tagged recombinant SDC4 proteins. We were able to reproduce the MST binding curve and obtain the Kd value of 1.3 µM for ETBP binding to GST-SDC4 (Figure 4C). In this experiment, no significant binding curves were detected between ETBP and the GST protein (Figure 4C).

To investigate whether ETBP could bind to other syndecan proteins, we prepared lysates from cells expressing GFP-tagged SDC1, SDC2, and SDC3, respectively. Interestingly, direct binding was only observed between ETBP and SDC1 with a Kd value of 8.87 µM (Figure 4D), indicating relative selectivity of ETBP towards syndecan proteins. To determine the ETBP binding site, we constructed a series of GFP-tagged SDC4 truncation mutants (Figure 4E), whose expression was confirmed by western blotting (Figure S3). In cellular MST assays, all truncation mutants were able to engage with ETBP where Kd values ranged from 2 µM to 6.2 µM, except for M3 and M4 that contain a common depletion of the TM motif (Figure 4F, 4G). These data indicate that the TM motif is essential for ETBP engagement. We also asked if the TM motif is essential for SDC1 interaction with ETBP. As predicted, no binding was observed between ETBP and the SDC1-ΔTM truncation mutant (Figure 4H). This is also consistent with the previous study that found the TM domain to be essential for ETBP interaction with TPOR [40].

ETBP stabilizes SDC4 and enhances SDC4-associated functions in cancer cells

To further investigate ETBP targeting of SDC4 in vivo, we first examined the effect of ETBP on SDC4 protein abundance. Surprisingly, the protein level of SDC4 in PANC1 cells was clearly enhanced upon ETBP treatment (Figure 5A). A dose-dependent increase in SDC4 abundance was also observed in the pancreatic cancer cell lines AsPC1 and CFPAC1 with ETBP (Figure 5B, 5C). Given that abnormally up-regulated SDC4 can act as a co-receptor of FGFR to augment the MAPK signaling pathway [10,12,38,41,42], we further asked whether ETBP could activate MAPK signaling by examining the level of phosphorylated AKT, ERK, and p38. Indeed, ETBP strongly enhanced the phosphorylation of AKT, ERK and p38 in a dose-dependent manner in wild-type PANC1 cells (Figure 5D). Importantly, ETBP enhancement of MAPK signaling was significantly attenuated in SDC4 KO cells (Figure S4D).

Considering that SDC4 KO decreased the level of macropinocytosis in PANC1 cells (Figure 2I, 2J), we also asked if ETBP regulated macropinocytosis in pancreatic cancer cells. As illustrated in Figure 5E-G, a strong increase in macropinocytosis was detected in all the pancreatic cancer cells examined (PANC1, BxPC3, and AsPC1) with ETBP treatment in a dose dependent manner. Unexpectedly, no clear differences in macropinocytosis were observed between wild-type and SDC4 KO PANC1 cells upon ETBP treatment (Figure S4C), possibly due to the fact that ETBP can target another macropinocytosis mediator SDC1 as well (Figure 4D). Taken together, we propose that, similar to ETBP action on its original target TPOR, ETBP is also a potential agonist of SDC4 and enhancer of cancer progression due to its ability to stabilize SDC4, and enhance MAPK signaling as well as macropinocytosis in cancer cells. Caution therefore should be taken when treating cancer patients with ETBP for chemotherapy-induced thrombocytopenia.