Skip to main content
PMC home page
British Journal of Pharmacology logoLink to British Journal of Pharmacology
. 2020 Apr 5;177(13):3009–3023. doi: 10.1111/bph.15030

A low MW inhibitor of CD44 dimerization for the treatment of glioblastoma

Chongwu Wang 1,✉, Zhaotao Wang 2, Chen Chen 1, Xiaojun Fu 3, Ji Wang 4, Xiaowei Fei 1, Xiaojing Yan 5, Ruxiang Xu 1,4,6,✉
PMCID: PMC7280016  PMID: 32080830

Abstract

Background and Purpose

As a hallmark of glioblastoma multiforme (GBM), CD44 plays a crucial role in promoting glioblastoma stem cell (GSC) stemness phenotypes and multiple drug resistance. The therapeutic potential of CD44 has been validated by the clinical successes of several CD44 inhibitors, including antibodies and hyaluronan‐related drugs.

Experimental Approach

We used systemsDock software to predict verbascoside as a candidate CD44 inhibitor. Microscale thermophoresis was used to confirm the interaction between CD44 and verbascoside. Four glioblastoma cell lines and a patient‐derived glioblastoma cell line were used to test the influences of verbascoside on glioblastoma. CD44‐overexpressing and CD44‐knockout cell lines were also used. Real‐time quantitative PCR and western blot analyses were performed. A xenograft mouse model was used to test verbascoside.

Key Results

Verbascoside bound to CD44 and suppressed its dimerization. By inhibiting CD44 dimerization, verbascoside decreased the release of the CD44 intracellular domain (CD44ICD) and suppressed the expression of CD44 downstream genes. Verbascoside treatment suppressed the stemness phenotypes of cells with high CD44 expression. In a mouse model of glioma, verbascoside treatment highly reduced the growth of intracranial tumours and inhibited CD44ICD release. Both stem cell marker and mesenchymal GBM subtype marker genes were down‐regulated in verbascoside‐treated mice.

Conclusion and Implications

Verbascoside suppressed growth of glioblastoma cells by inhibiting CD44 dimerization. Stem cell‐like cell properties and tumour cell growth were also suppressed by verbascoside, both in vitro and in vivo. Verbascoside significantly prolonged survival of xenografted mice.

Abbreviations

BiFC

bimolecular fluorescence complementation

GBM

glioblastoma

MST

microscale thermophoresis

PD‐GBMC

patient‐derived glioblastoma cell

What is already known

  • CD44 is related to cell stemness and up‐regulated in many kinds of cancer.

  • Verbascoside shows anti‐tumour activity in breast cancer.

What this study adds

  • CD44 is a direct target of verbascoside.

  • Verbascoside suppresses CD44 dimerization.

What is the clinical significance

  • Verbascoside suppresses growth of glioblastoma xenografts in mice and in vitro.

  • At effective doses, verbascoside showed no side effects in the mouse model.

1. INTRODUCTION

Usually, glioma begins in glial cells and includes grade II and grade III tumours and grade IV glioblastoma (GBM). As the most lethal primary brain tumour, GBM has a 12‐ to 15‐month survival length after diagnosis and a 3% to 5% 5‐year survival rate (Di Carlo, Cagnazzo, Benedetto, Morganti, & Perrini, 2017; Schapira, 2007). Even after complete resection, most patients will develop recurrent tumours either near the original site or at a distant location within the brain. The general treatment protocol for GBM involves radiotherapy and/or chemotherapy after surgery (Gallego, 2015). However, eventually, over 98% of patients will suffer from GBM recurrence and radiotherapy and chemotherapy tolerance (IARC, 2014). Although they are highly heterogeneous, a group of stem cell‐like cells, usually called glioblastoma stem cells (GSCs), are thought to be mainly responsible for recurrence and chemotherapy/radiation resistance. GSCs can be enriched by the stem cell markers CD133 and CD44(Anido et al., 2010; Pietras, 2011; Singh et al., 2003). GSCs are also able to exclude Hoechst 33342 in the side population and are related to chemotherapy tolerance, as are other cancer stem cells (Shimoda, Ota, & Okada, 2018; Was et al., 2018).

The transmembrane glycoprotein CD44 is a marker of stem cells and is highly expressed in many tumour cells (Curtarelli et al., 2018; Zoller, 2011), including the mesenchymal subtype of GBM (Bhattacharya, Mitra, Ray Chaudhuri, & Roy, 2018; Brown et al., 2015; Rimkus et al., 2018). CD44 has many ligands, such as hyaluronan, hepatocyte growth factor (HGF) and osteopontin (Bennett et al., 1995; Couchman, 2010; Pietras et al., 2014; van der Voort et al., 1999). Linked by disulfide bonds, dimerization is crucial for the functions of CD44 (Hartmann et al., 2015; Liu & Sy, 1997). When bound to hyaluronan, CD44 triggers inhibition of proliferation, while its interaction with osteopontin triggers an increase in proliferation (Godar et al., 2008; Tian et al., 2013; Todaro et al., 2014). In addition, both of these interactions need CD44 dimerization. Activated by osteopontin, the CD44 intracellular domain (CD44ICD) is released by γ‐secretase and localizes in both the cytoplasm and nucleus (Barat et al., 2017; Bourguignon, Zhu, Shao, & Chen, 2001; Nagano & Saya, 2004). Together with other transcription factors, CCND1, HIF2α, and other genes are up‐regulated and enhance cell stemness (Fan et al., 2018; Jafari, Joshaghani, Panjehpour, Aghaei, & Zargar Balajam, 2017; Johansson et al., 2017; Kaul et al., 2012; Khan, Adesida, & Hardingham, 2007; Kim et al., 2018). CD44 is therefore considered an important target for tumour therapy.

Verbascoside is an active compound of traditional herbal medicine that is mainly found in species in all the families of the Lamiales order. It has been widely used in traditional Chinese medicine for a very long time. However, the underlying molecular mechanisms have not been fully elucidated. Here, we show that verbascoside can bind to both the monomer and dimer of CD44. After binding with CD44 monomer, verbascoside suppresses CD44 dimerization in vivo. Following culture with verbascoside, the stemness of glioblastoma cells was suppressed. An in situ xenograft mouse tumour model also confirmed the anti‐tumour activity of verbascoside through CD44. Taken together, our results demonstrated that verbascoside had potential as an anti‐glioma antigen for clinical application, through its inhibition of the CD44 signalling pathway.

2. METHODS

2.1. Cell lines and culturing conditions

The human glioblastoma cell lines U87MG, T98G, U118, and U251MG were purchased from the Chinese Academy of Medical Sciences (Beijing, China). The U87MG, T98G, U118, and U251MG are cultured in GSC medium (high‐glucose DMEM/NeuroBasal medium = 50/50, 10 ng·ml−1 of EGF, 10 ng·ml−1 of FGF, 1× B27 supplement, 1× N2 supplement). The patient‐derived glioblastoma cell line (PD‐GBMC) was a kind gift from Dr Quanhong Ma from SooChow University and was cultured in GSC medium. The HEK293 cell line was cultured in high‐glucose DMEM supplemented with 10% FBS. All cells were incubated at 37°C in a humidified atmosphere of 5% CO2.

2.2. Plasmids and sequencing

All plasmids were purchased from GeneCopoeia. Homo sapiens CD44 transcript variant 4 (NM_001001391.2) was used for all plasmid construction. For sequence alignment, Homo sapiens CD44 antigen isoform 4 (NP_001001391.1) and Mus musculus CD44 antigen isoform a (NP_001034239.1) were used.

2.3. Western blotting

Samples were prepared by adding 2× reducing loading buffer or 2× non‐reducing loading buffer to harvested cells or xenograft glioblastoma tissue, followed by homogenization using electric tissue grinders. Homogenates then were boiled at 100°C for 10 min. Equal amounts of total protein were separated by 10% SDS‐PAGE gel electrophoresis in Tris‐MOPS buffer (Tris 6.06 g, MOPS 10.46 g, SDS 1 g, EDTA 0.3 g in 1,000 ml deionized water) and then transferred to polyvinylidene difluoride membranes (Merck, KGaA, Darmstadt, Germany), which were blocked with 5% skim milk in TBST for 1 hr at room temperature before being incubated with specific primary antibodies overnight at 4°C. The membranes were then incubated with the appropriate HRP‐conjugated secondary antibodies for 1 hr at room temperature. The resulting signals were obtained using Super Signal ECL (Pierce, Rockford, IL, USA). The immuno‐related procedures used comply with the recommendations made by the British Journal of Pharmacology.

2.4. Quantitative real‐time PCR

RNA was isolated using TRIzol™ Reagent (Invitrogen, 15596018), and cDNA was synthesized using HiScript II 1st Strand cDNA Synthesis Kit with gDNA wiper (Vazyme, R212‐02) according to the manufacturer's protocol. Quantitative PCR was performed using a 7900HT Fast Real Time PCR system and ChamQ SYBR Color qPCR Master Mix with low ROX Premixed (Vazyme, Q431‐03) using primers listed in Table S1. Relative gene expression level was normalized to the expression of housekeeping genes (actin) using the comparative Ct method (Vandesompele et al., 2002).

2.5. Cell proliferation assay

The cells were seeded in a 96‐well plate at a density of 2.5 × 103 cells per well for 24 hr and then treated as indicated. For testing, the supernatant was discarded, CCK‐8 added (10 μl of CCK‐8 solution (Dojindo, CK04) in 100‐μl high‐glucose DMEM per well), and then incubated at 37°C for the required time. Each test was repeated four times. The absorbance of the reaction mixture was subsequently measured by a microplate reader at 450 nm.

2.6. Microscale thermophoresis

Microscale thermophoresis (MST) was conducted using an NT.115 MST instrument (Nano Temper Technologies GmbH) equipped with green and blue filter sets. Both CD44ECD‐His and CD44ECD‐Fc peptides (20 μM in PBS buffer) were labelled by vendor‐supplied kit (Nano Temper) and buffer exchanged to the assay buffer (25‐mM Tris (pH 7.5), 17‐mM NaCl, 0.05%, Tween‐20, and 5‐mM β‐mercaptoethanol). The mixture was incubated at 37°C in the dark for 30 min. Free dye was separated from the protein–dye conjugate using vendor‐supplied kit (Nano Temper). Both verbascoside (500 μM), osteopontin (400 μM), and hyaluronic acid (2000 mM, without biotin) were serially diluted with 1:1 ratio of 16 gradients. Then the labelled peptide and ligands were mixed (1:1 ratio) and incubated at room temperature in the dark for 10 min. Then the mixtures were loaded and tested. Data were acquired using 40% MST and 20% LED settings. The temperature‐jump methodology was used to analyse data. Data were analysed in NTAnalysis‐1.5 software. MST figures were rendered using Origin9.

2.7. Bimolecular fluorescence complementation assay

Sixteen hours after transfection, cells were seeded on coverslips placed at the bottom of 12‐well plates. Then cells were treated with either control (normal saline) conditions or under serial concentrations of verbascoside (120 and 240 μM) for 8 hr. After treatments, cells were digested and washed three times with 1× PBS and fixed with 4% PFA in PBS at room temperature for 20 min. The cells were washed three times with 1× PBS; then the coverslips were mounted in DAPI Fluoromount‐G® (SouthernBiotechnology, 0100‐20) for microscopic analysis. Fluorescent photographs were generated with a Leica TCS SP5 II confocal microscope, and the results were analysed by the confocal acquisition software.

2.8. Flow cytometry

For quantitation of bimolecular fluorescence complementation (BiFC) assays, a mCherry encoding plasmid was cotransfected together with BiFC constructs (as described above). BiFC plasmids were added in excess to mCherry encoding plasmid (DNA ratio 3:1); thus, the mCherry positive cells should also express BiFC constructs; 16 hr after transfection, cells were seeded on 12‐well plates. Then cells were treated with either control (normal saline) conditions or under serial concentrations of verbascoside (120 and 240 μM) for 8 hr. After treatments, cells were digested and washed three times with 1× PBS and fixed with 4% PFA in PBS at room temperature for 20 min. The percentage of BiFC‐positive cells was counted among all mCherry positive cells by flow cytometry. For Cd44 knockout cell line identification, cells were digested and washed three times with PBS and fixed with 4% PFA in PBS at room temperature for 20 min. After washing three times with PBS, cells were incubated with PBS containing 0.1% TritonX‐100, APC anti‐human CD44 antibody, and PE anti‐human SOX2 antibody for 30 min at room temperature. Flow cytometry was performed with a BD Accuri C6 (BD Biosciences). Data were analysed with Accuri C6 software (BD Biosciences).

2.9. Treatment of the U251MG xenograft mouse model with verbascoside

All animal care and experimental procedures were in accordance with a standard protocol approved by the Institutional Animal Care Committee of PLA General Hospital. All procedures performed in the studies involving the animals were compliant with the ethical standards of the institution or practice at which the studies were conducted. Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny, Browne, Cuthill, Emerson, & Altman, 2010; McGrath & Lilley, 2015) and with the recommendations made by the British Journal of Pharmacology.

Female 6‐week‐old BALB/c‐Nud mice were purchased from Beijing HFK Bioscience Co., Ltd (Beijing, China). All mice were housed in a specific pathogen free facility with six mice per cage under a 12‐hr light/12‐hr dark cycle at a constant temperature of 24°C and fed a standard rodent diet in the laboratory animal research centre of PLA General Hospital. The animals were acclimatized to the laboratory for at least 1 week prior to the start of the experiments. All surgical procedures were performed under anaesthesia using isoflurane. At the end of the experiments, animals were humanely killed by CO2 asphyxiation. In animal studies, both the carer of the animals and the assessor of the results were not aware of the treatments given.

The mice were anaesthetized with 2% isoflurane (flow rate 200 ml·min−1 ) for induction and then 1.5% (200 ml·min−1) for maintenance. Then each mouse was intracranially injected with of 4 μl of cultured, luciferase‐transfected, U251MG cells (1 × 106 cells per mouse) at a rate of 0.5 μl·min−1 using a Micro 4 Microsyringe Pump Controller (World Precision Instruments, Sarasota, FL) attached to a Hamilton syringe with a 33‐gauge needle (Hamilton, Reno, NV). The injection was performed in the mid‐right striatum at the following coordinates (in millimetre from the bregma): +0.5 anterior–posterior, +2.0 medio‐lateral, and −2.8 dorsoventral. After surgery, 5 mg·kg−1 of ketoprofen were used for analgesia. After 7 days of cell transplantation, the tumour‐bearing mice were distributed into two groups (n = 15 each) and injected i.p. with verbascoside (100 mg·kg−1·day−1) or normal saline. Tumour sizes and body weights were measured once every 7 days. At the end of the experiments, the mice were killed, and the tumours were resected for further analysis.

2.10. Sphere formation assays

U251MG cells or PD‐GBMCs were seeded at 250 cells per well in DMEM/F12 medium (1% B27, 10 ng·ml−1 of EGF, 10 ng·ml−1 of FGF, and 1% N2) with the indicated amounts of verbascoside or PBS in a 24‐well plate with an Ultra‐Low Attachment surface. After 10 days of incubation, spheres with diameters over 50 μm were counted, and representative fields were microphotographed. Each treatment was performed in six replicates.

2.11. Bioluminescence imaging

d‐luciferin was purchased from Promega (P1043) and was resuspended in normal saline at a concentration of 100 mg·ml−1. Mice were injected (i.p., 150 mg·kg−1) with d‐luciferin and imaged 10 min thereafter using an IVIS® Spectrum optical imaging system fitted with an XGI‐8 Gas Anesthesia System (Caliper Life Sciences, Hopkinton, MA). Bioluminescent images were acquired using the auto‐exposure function. Analysis of the signal intensities and image comparisons were performed using Living Image® Software (Caliper Life Sciences).

2.12. Competitive elisa

The binding of verbascoside to CD44 was analysed by a competitive elisa in which the binding of biotin‐labelled hyaluronan (1,200 Da, TianjinshilanBio, China, H00134) (Biotin‐HA) to CD44 was in competition with verbacoside and vehicle (PBS); 96 well plates were coated with goat Anti‐Human IgG Fc Antibody (R&D Systems #G‐102‐C) in PBS at 5 μg·ml−1, 100 μl per well. Plates were incubated at room temperature overnight and then washed three times with ELISA wash buffer (PBS + 0.05% Tween 20, pH 7.2–7.4). Plates were blocked with 300 μl per well of assay buffer (PBS + 0.5% BSA, pH 7.2–7.4) at room temperature for 1 hr. The blocking solution was removed and recombinant Human CD44 Fc Chimera Protein, CF (R&D Systems #3660‐CD), was added (100 μl per well of 0.1 μg·ml−1 in assay buffer) and incubated at room temperature for 2 hr. The assay plate was washed three times with ELISA wash buffer. In a dilution plate, threefold serial dilutions of Biotin‐HA with assay buffer were prepared, starting at 20 μg·ml−1, 120 μl per well (final first well concentration of Biotin‐HA is a minimum of 6.67 μg·ml−1), with or without verbascoside (10 μM). Then 100 μl per well were transferred from the dilution plate to the assay plate followed by incubation at room temperature for 2 hr. Then the assay plate was washed three times, streptavidin‐HRP (Abcam, ab7403, 1:10000) is added (100 μl per well) and incubated at room temperature for 45 min. After washing three times with ELISA wash buffer, the colour reagents A and B (Solarbio, China, PR1210‐2 * 50 ml, mixed just prior to use) were added (100 μl per well) and incubated at room temperature for 20–30 min. Then stop solution (Solarbio, China, C1058) was added (50 μl per well). Plates were read at 450 nm with 540 nm as the correction wavelength. Plot dose curve uses four‐parameter fit.

2.13. Data and statistical analysis

In each experiment, the same cell was used to evaluate the effects of verbascoside and its related control. Mice used in this study were randomly allocated to cages by vivarium staff and randomized into vehicle‐treated or verbascoside‐treated group before the treatment. For blinding, we had different people conducting experiments (operator) and analysing data (analyst).

The data are presented as the mean ± SD from five or more (shown in figure legend) independent experiments. Simple comparisons between two groups were analysed using independent t tests, and multiple comparisons between the groups were assessed with one‐way ANOVA, followed by post hoc analyses, which were performed with the LSD test or Dunnett's T3 test. All analyses were performed using Origin9 software. P < .05 was considered statistically significant. The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology.

2.14Materials

Verbascoside was purchased from TianJin ShiLan (TianJin, China). Hyaluronan biotin sodium salt (Sigma‐Aldrich, B1557‐5MG) and hyaluronic acid sodium salt from Streptococcus equi. with MW of 1,200 (Sigma‐Aldrich, 49775) was purchased from Sigma‐Aldrich. DMEM (Catalog no. 10566016) were purchased from Gibco (Grand Island, USA). The following antibodies were used: CD44(150‐250) (Abcam, ab51037) (RRID:AB_868936); CD44(692‐742) (Abcam, ab157107); GAPDH (Abcam, ab8245) (RRID:AB_2107448); GFP (Abcam, ab13970) (RRID:AB_300798); goat anti‐mouse IgG (HRP) (Abcam, ab6789) (RRID:AB_955439); goat anti‐rabbit IgG (HRP) (Acam, ab6721) (RRID:AB_955447); goat anti‐chicken IgY (HRP) (Abcam, ab6877) (RRID:AB_955465); APC anti‐human CD44 antibody (Biolegend, 338806) (RRID:AB_1501195); PE anti‐human SOX2 antibody (Biolegend, 656104) (RRID:AB_2562853); and streptavidin (HRP) (Abcam, ab7403). The following proteins were used: recombinant Human CD44 Fc Chimera Protein, CF (R&D Systems, 3660‐CD); goat anti‐human IgG Fc Antibody (R&D Systems, G‐102‐C) (RRID:AB_573135); BSA, protease free (Sigma‐Aldrich, A7030); recombinant human osteopontin protein (His Tag) (Sino Biological, 10352‐H08H); and recombinant human osteopontin (Peprotech, 120‐35). All lentivirus and plasmids were purchased from Shanghai OBio Technology.

2.14. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2019/20 (Alexander et al., 2019).

3. RESULTS

3.1. Discovery of a CD44 inhibitor by machine learning

systemsDock has been reported as a well‐developed machine learning‐based ligand‐protein binding prediction and analysis web server tool (Hsin et al., 2016; Hsin, Ghosh, & Kitano, 2013). To develop a CD44 inhibitor, we used it to analyse several 2‐hydroxytetrahydro‐pyran‐containing compounds that are crucial for the interaction between CD44 and hyaluronic acid (Figure S1A). A mouse CD44 structure (2JQC) was used (Banerji et al., 2007). In those compounds, verbascoside showed a higher potential interaction capacity with CD44 than its native ligand, and EGFR was used as a negative control (Figure S1B,C). Visualization of the protein–ligand interaction shows that CD44 interacts with its native ligand through Ala103 and Tyr109 (Banerji et al., 2007) (Figure S1D,F). CD44 interacts with verbascoside through Ala103, Tyr109, Leu111, and Val112 (Figure S1E,G). The binding interaction between CD44 residues and verbascoside is shown in Figure S1F. The native ligand of CD44 is shown as a control (Figure S1E). On comparison with human CD44, we found that these amino acid resides are highly conserved between humans and mice (Figure S1H).

3.2. Verbascoside binds to CD44 in vitro

An MST assay was performed to measure the half‐maximal effective concentration (EC50) of the interaction between CD44 and verbascoside. Osteopontin and hyaluronan were used as controls. Here, two kinds of CD44 proteins were used: CD44ECD‐His, which contains the CD44 exocellular domain (CD44ECD) and the carboxyl terminal His tag, and CD44ECD‐Fc, which contains CD44ECD and the carboxyl terminal human Fc tag (Figure 1a). Two Fc peptides formed dimers by disulfide bonds and simulated CD44 dimerization in vitro, while CD44ECD‐His formed monomers in vitro (Figure 1c). The CD44–osteopontin interaction was also tested for comparison with the CD44–verbascoside interaction.

FIGURE 1.

FIGURE 1

Open in a new tab

MST assays quantify the interaction between CD44ECD proteins with osteopontin, hyaluronan, and verbascoside. (a) Schematic overview of CD44‐related proteins used in the MST assay. (b) The binding of CD44ECD‐His with osteopontin is undetectable, while the binding of CD44ECD‐Fc (EC50 = 94.1 μM) with osteopontin induces a relatively large change in thermophoretic mobility. (c) Western blot analysis detected CD44‐Fc and CD44‐His on non‐reducing and reducing SDS‐PAGE gels. CD44‐Fc‐cl indicates CD44‐Fc cross‐linked by paraformaldehyde. (d) The binding of CD44ECD‐His with hyaluronan (molecular mass 1200 Da) is undetectable, while the binding of CD44ECD‐Fc with osteopontin shows biphasic binding, with a high binding phase (EC50 = 0.36 μM;; left) and a low binding phase (EC50 = 56.7 μM; right). (e) The binding of CD44ECD‐His with verbascoside yielded an EC50 value of 90.4 μM, while the binding of CD44ECD‐Fc with osteopontin showed triphasic binding, with the first binding phase (EC50 = 0.93 μM; left), the second binding phase (EC50 = 6.23 μM; middle), and the third binding phase (EC50 = 154 μM; left). (f) Binding of different CD44 mutants with verbascoside. (g) Competitive ELISA shows that CD44 binds to Biotin‐HA (1200 Da) in the presence of vehicle (PBS) or verbascoside. Mean OD of each group under each concentration were calculated. K D of CD44 and Biotin‐HA under PBS is 150, and K D of CD44 and Biotin‐HA under verbascoside (10 μM) is 7069. The data are presented as mean ± SD from five independent experiments for (a)–(f) (n = 5) and eight independent experiments for (g) (n = 8), * P < .05, significantly different as indicated; Student's t test. MST assays were carried out under medium MST power at 25°C. The LED power was set to 40%. The resulting dose–response curves were fitted to a multisite binding model to extract the EC50 values; the SD was calculated using the EC50 values from each independent experiment. Fnorm = normalized fluorescence. ΔFnorm = (Fnormn − Fnorm0)/Fnorm0

Interactions between CD44ECD‐His and osteopontin, CD44ECD‐His and hyaluronan (molecular mass 1,200 Da) were undetectable under the experimental conditions (Figure 1b,d). These results are consistent with those from previous reports, which have shown that the interaction between CD44 and some of its native ligands depends on CD44 dimerization. The interaction between CD44ECD‐Fc and hyaluronan (molecular mass 1200 Da) shows biphasic binding (Figure 1d), while the interaction between CD44ECD‐Fc and verbascoside shows triphasic binding (Figure 1e). The high affinity binding is superimposed by a second binding event at higher verbascoside/osteopontin concentrations. Verbascoside shows higher binding capacity than osteopontin. As shown in Figure S1H, four amino residues are predicted to be crucial for the interaction between CD44 and verbascoside. Interactions between CD44A99D, CD44Y105D, CD44L107D, CD44L108D, and CD44A99DY105DL107DL108D (CD444M) and verbascoside were tested with MST assays (Figure 1f). A99 showed the greatest influence, while Y105 showed the smallest influence.

To further confirm the interaction between CD44 and verbascoside, we performed a competitive ELISA (Figure 1g). An interaction between CD44 and biotinylated hyaluronan was detected in the presence of PBS or verbascoside. The mean OD in the presence of verbascoside was lower than that in PBS, which indicates that verbascoside alters the interaction between CD44 and hyaluronan.

Together, these data confirm that CD44 interacts with verbascoside in vitro.

3.3. Verbascoside has antiproliferative activity in human GBMCs

Given the variety of functions of CD44, we hypothesized that the interaction between CD44 and verbascoside may influence the proliferative activity of cancer cells. The expression level of CD44 was tested by western blot and real‐time quantitative PCR (Figure S2). The cell lines showed different expression levels of CD44. U251MG cells showed the highest expression level of CD44, while its expression in PD‐GBMCs was barely detected (Figure S3).

We assessed the anti‐proliferative effect of verbascoside on a series of glioblastoma cell lines, including U87MG, T98G, U118, U251MG, and a PD‐GBMC line. The response of those cell lines to verbascoside under different concentrations and different treatment times is shown in Figure S4. Table S2 shows the EC50 values of each cell line. At each time point, compared with the solvent, with increasing concentrations of verbascoside and increasing treatment times, the cell proliferation rates decreased significantly in U87MG, T98G, U118, and U251MG cells but not in PD‐GBMCs. Compared with the solvent, even after 72 hr of treatment with 240‐μM verbascoside, the growth rate of PD‐GBMCs decreased only 27%. Verbascoside treatment induced death in all the cell lines tested, with variable suppression levels ranging from 27% to 88% under 240 μM after 72 hr and EC50 values ranging from 29.98 to 46.23 μM. This range of EC50 values is within the value of the first combination of CD44ECD‐Fc/verbascoside (0.94 μM) and CD44ECD‐His/verbascoside (90.4 μM). This result suggests that verbascoside influences both the CD44 monomer and dimer in vivo simultaneously.

The proliferative activity of verbascoside showed a negative correlation (R 2 = .9688) with CD44 expression levels across all tumour cell lines (Figure 2a).

FIGURE 2.

FIGURE 2

Open in a new tab

Verbascoside suppresses cell proliferation through CD44. (a) The suppression of cell proliferation (24 hr, 120‐μM verbascoside) negatively correlated with the CD44 level. See also Figures S3 and S4. (b) Proliferative activity of U251MG cells under different concentrations of CD44ECD‐His proteins. BSA was used as a negative control. (c) Relative growth of CD44‐knockout cell lines under a range of concentrations of verbascoside (VB). U251MGCd44KO‐814, U251MGCd44KO‐815, and U251MGCd44KO‐818 cells are CD44‐knockout U251MG cell lines generated by CRISPR‐CAS9 technology. U251MG cells were used as a positive control. (d) Schematic overview of the GFP‐fused full‐length gene construct and the truncated CD44 gene construct. ECD, extracellular domain; ICD, intracellular domain; SP, signal peptide; TM, transmembrane domain. (e) Response to verbascoside (VB) in CD44‐GFP, CD44ECD‐GFP, CD44ICD‐GFP, and GFP transgenic U251MG cell lines. The proliferation activity of each group in (b)–(d) was tested by CCK‐8 assays after 24 hr of treatment. (f) Expression of GFP‐fused full‐length CD44 and truncated CD44 in U251MG cells detected by western blot with an anti‐GFP antibody. GAPDH was used as a loading control. (g) Proliferation activity of different CD44 mutant transgenic U251MGCd44KO814 cells under verbascoside treatment. The data are presented as mean ± SD from five independent experiments (n = 5), * P < .05, significantly different as indicated; Student's t test

3.4. Verbascoside suppresses the proliferation of human GBMCs in a CD44‐dependent manner

To further determine whether verbascoside influenced GBMCs through CD44, we added a series of concentrations of CD44ECD‐Fc to U251MG cell culture medium containing verbascoside and tested the proliferative activity of each cell line. Figure 2b shows that the relative growth of U251MG cells was increased along with increasing concentrations of CD44ECD‐Fc.

CD44‐knockout cells (generated with the CRISPR‐CAS9 system) were used to test the hypothesis that the lack of CD44 confers U251MG cell insensitivity to verbascoside. Three individual cell lines that contain different CD44 mutations were used ( Figure S4). Viability experiments comparing verbascoside‐treated wild‐type (WT) and CD44‐knockout U251MG cells indicated that CD44‐knockout U251MG cells were more resistant to verbascoside treatment than WT U251MG cells (approximately threefold decrease in sensitivity) (Figure 2c).

Moreover, it has been reported that after the dimerization of CD44, the CD44ICD is released by γ‐secretase, transported into the nucleus, and regulates the expression of downstream genes (Hartmann et al., 2015). We generated lentivirus‐mediated CD44‐GFP, CD44ECD‐GFP, GFP‐CD44ICD, and GFP transgenic HEK293 cell lines (Figure 2d,f). Overexpressing CD44ICD can mimic osteopontin‐induced CD44‐dependent stem‐like cell‐like phenotypes. Consistent with this finding, U251MG cells with CD44ICD‐GFP showed resistance to verbascoside. On the other hand, CD44‐GFP and CD44ECD‐GFP transgenic cell lines were insensitive to lower concentrations of verbascoside (30 and 60 μM) but sensitive to higher concentrations of verbascoside (90 and 120 μM) (Figure 2e). We hypothesized that this shift was caused by CD44 and CD44ECD saturation with a high concentration of verbascoside. Furthermore, we tested the proliferation activity of CD44A99D, CD44Y105D, CD44L107D, CD44L108D, and CD444M mutant cells following verbascoside treatment (Figure 2g). Consistent with the in vitro data, CD444M cells showed the most resistance to verbascoside, and CD44Y105D and CD44L108D cells showed no significant difference.

Collectively, these findings support the idea that verbascoside exerts its effects on cell viability by inhibiting CD44.

3.5. Verbascoside inhibits CD44 dimerization

As shown in Figure 1, both the EC50 value between CD44‐His and verbascoside and the EC50 value of the high affinity phase between CD44‐Fc and verbascoside are lower than the EC50 value between osteopontin and verbascoside. This result suggests that verbascoside may compete with osteopontin and influence CD44 dimerization in vivo. The CD44 dimer is linked by disulfide bonds.

We performed a BiFC assay to detect CD44 dimerization in vivo. HEK293 cells‐transfected CD44‐YNE/CD44‐YCE‐complemented fluorescence plasmids (Figure 3c) were treated with different concentrations of verbascoside. We found that CD44‐YNE/CD44‐YCE‐complemented fluorescence decreased under verbascoside treatment both at the plasma membrane and cytoplasm (Figure 3a,c).

FIGURE 3.

FIGURE 3

Open in a new tab

Verbascoside suppresses CD44 dimerization. (a) CD44 dimerization. CD44 homodimerization in living cells was determined with BiFC assays (HEK293 cells) under different concentrations of verbascoside. Plasmids encoding human CD44 proteins were fused to nonfluorescent N‐terminal and C‐terminal fragments of the Venus protein. (b) BiFC‐positive cells were analysed by flow cytometry. The MCherry coding plasmid was cotransfected with BiFC plasmids as a transfection control. The percentage of BiFC‐positive cells among mCherry‐positive cells was calculated. (c) Schematic overview of CD44‐related proteins used in the BiFC assay. (d) Western blot detected CD44 on a nonreducing SDS‐PAGE gel. GAPDH was used as a loading control. (e) Western blot detected CD44 in CD44 and CD44C286AC29A mutant transgenic U251MGCd44KO814 cells on a non‐reducing SDS‐PAGE gel. (f) Quantification of CD44 signals from Figure 3e across five independent experiments. VB, verbascoside. (g) Proliferative activity of CD44 and CD44C286AC29A mutant transgenic U251MGCd44KO814 cells under verbascoside treatment. The data are presented as mean ± SD from five independent experiments (n = 5), * P < .05, significantly different as indicated; Student's t test

For quantitation of the BiFC assay, an mCherry coding plasmid was cotransfected into HEK293 cells with the CD44 BiFC system. Flow cytometry was performed to detect BiFC‐positive cells among mCherry‐positive cells. The amount of BiFC‐positive cells decreased with increasing concentrations of verbascoside in the culture medium (Figure 3b,c).

By performing non‐reducing SDS gel electrophoresis and western blot, we determined that the amounts of CD44 dimer in U251MG cells were decreased with increasing concentrations of verbascoside in the culture medium (Figure 3d,f). CD44, C286, and C295 are crucial for CD44 dimerization (Hartmann et al., 2015; Liu & Sy, 1997). As shown in Figure 3e, CD44C286AC295A mutant cells show a great reduction in the CD44 dimer. Cell proliferation assays also showed that the CD44C286AC295A transgenic cell line tolerated verbascoside compared with the CD44 WT cell line (Figure 3g). By combining these data, we could confirm that verbascoside suppressed CD44 dimerization.

3.6. Verbascoside inhibits the release of the CD44ICD and down‐regulates CD44 downstream genes

Activated CD44 is cleaved in a two‐step process that depends on γ‐secretase, resulting in the release of CD44ICD. It has been reported that CD44ICD release is induced by osteopontin and involves CD44 dimerization in brain tumours. We hypothesized that the inhibition of CD44 dimerization by verbascoside may inhibit the release of CD44ICD. To test this hypothesis, we treated U251MG cells with 60‐μM verbascoside. Western blot analysis showed that the amount of free CD44ICD decreased under verbascoside treatment (Figure 4a,b).

FIGURE 4.

FIGURE 4

Open in a new tab

Verbascoside suppresses the release of CD44ICD and the expression of CD44 downstream genes. (a) Western blot analysis shows CD44ICD in U251MG cells treated with verbascoside (60 μM) at the indicated times. Two CD44 antibodies were used. CD44ECD (ab51037, Abcam) targets the CD44 amino acid 150‐250 region. CD44ICD (ab157107, Abcam) targets the CD44 amino acid 692‐742 region. (b) Quantification of CD44ICD signals from Figure 4a from five independent experiments. Chemiluminescence signals were calculated. * P < .05, significantly different from initial values (0 hr). (c) Real‐time quantitative PCR analysis of CD44, CCND1, and HIF2α in U251MG cells and PD‐GBMCs treated with verbascoside (60 μM) at the indicated times. The data are presented as mean ± SD from five independent experiments (n = 5), * P < .05, significantly different as indicated; Student's t test

CD44 activation can induce the expression of CD44, CCND1, and HIF2α. Real‐time quantitative PCR was performed to test the expression of these genes under verbascoside treatment. The expression levels of CD44, CCND1, and HIF2α in U251MG cells, but not in PD‐GBMCs, were decreased under verbascoside treatment over time (Figure 4c).

3.7. Verbascoside suppresses the stem cell‐like phenotype of GBMCs

It has been reported that osteopontin induces a cancer stem cell‐like phenotype by facilitating CD44ICD release18,(Cho et al., 2015; Nagano & Saya, 2004). We asked whether the inhibition of CD44 by verbascoside could suppress the stem cell‐like phenotype. Throughout the study, we performed colony formation assays and examined the expression of a panel of well‐established stem cell markers (NONOG, SOX2, OCT4, Nestin, and ID1) for a broad and relevant measure of stemness.

Figure 5b shows that U251MG cells treated with verbascoside formed fewer colonies than did control cells (Figure 5a) in a concentration‐dependent manner, while PD‐GBMCs, with decreased CD44 expression, showed less affected colonies (Figure 5b). Furthermore, the stem cell markers NANOG, SOX2, OCT4, Nestin, and ID1 were all down‐regulated in verbascoside‐treated U251MG cells but not in PD‐GBMCs, as shown by quantitative real‐time quantitation PCR (Figure 5c). High expression levels of CD44 are also associated with a mesenchymal phenotype in GBMCs with aggressive glioma cell growth. Here, we tested several marker genes of the mesenchymal subtype. Figure S5 shows that the expression levels of STAT3, VEGF, RUNX1, and BHLHB2 were down‐regulated under verbascoside treatment.

FIGURE 5.

FIGURE 5

Open in a new tab

Verbascoside suppresses the stem‐like phenotype in glioma cells expressing high levels of CD44. Effects of verbascoside on the colony formation ability in U251MG cells (a) and PD‐GBMCs (b). Colonies were detected by crystal violet staining after 12 days of culture with verbascoside as indicated for the colony forming assays. The histogram shows the mean colony values of U251MG cells and PD‐GBMCs. Colony numbers were calculated. * P < .05, significantly different from control (0 μM). (c) The expression levels of SOX2, OCT4, NANOG, Nestin, and ID1 in U251MG cells and PD‐GBMCs under verbascoside treatment were tested by real‐time quantitative PC. The data are presented as mean ± SD from five independent experiments (n = 5), * P < .05, significantly different as indicated; Student's t test. (d) Neurospheres of U251MG cells and PD‐GBMCs under verbascoside treatment. (e) Side population of U251MG cells and PD‐GBMCs under verbascoside treatment

Neurospheres and side populations are characteristics of neural stem cells. Both the neurosphere (Figure 5d) and side population (Figure 5e) of U251MG cells decreased under verbascoside treatment.

Together, these data suggest that verbascoside suppresses glioma cell stem cell‐like phenotypes in a CD44‐dependent manner.

3.8. In vivo therapeutic efficacy of verbascoside in a mouse model of GBM

Accumulating evidence suggests that verbascoside penetrates the brain. Therefore, we examined the therapeutic effects of verbascoside on mice bearing intracranial tumours derived from U251MG cells. However, there are many sites of action by verbascoside in vivo. To exclude potential biotoxicity, we applied double doses and double treatment times of verbascoside in the mouse experiments. Haematoxylin and eosin staining showed no significant differences in the heart, kidney, liver, lung, or spleen between the treatment group and the control group (Figure S6). Bioluminescence imaging confirmed the anti‐tumour effects of verbascoside (Figure 6a). Furthermore, we quantified the residual number of each fluorescently labelled tumour cell population after therapy and confirmed the activity of verbascoside (Figure 6b). These in vivo results translated into greater survival with treatment (Figure 6c,e). Based on these data, verbascoside showed a reduction of tumour growth and a prolongation of survival in the xenograft mouse model, which indicate that verbascoside was an effective treatment for GBM.

FIGURE 6.

FIGURE 6

Open in a new tab

Verbascoside suppresses xenografted U251MG tumour growth by affecting CD44 and stem cell‐related genes. (a) Bioluminescence images of mice bearing xenografted tumours derived from luciferase‐expressing U251MG cells showing the effect of treatment with 100 mg·kg−1 verbascoside once per day on tumour growth. Time points indicate weeks after the intracranial injection of U251MG cells. (b) Bioluminescence was measured in photons·s−1·cm−2·sr−1. Quantification of bioluminescence signals during 5 weeks of treatment in mice implanted with luciferase‐expressing U251MG cells. Signals were normalized to those on Day 10 for each mouse. * P < .05, significantly different as indicated; one‐way ANOVA with Tukey's method for multiple comparisons. (c) Kaplan–Meier survival curves for mice bearing U251MG orthotopic tumours following treatment with 100 mg·kg−1 of verbascoside. * P < .05, significantly different as indicated; log‐rank test. (d) Western blot analysis shows decreased CD44ICD levels in orthotopic tumours following verbascoside treatment. Four individual treatment groups (nos. 2, 3, 5, and 9) and control groups (nos. 15,16, 18, and 21) are shown. (e) Body weights of mice bearing U251MG orthotopic tumours following treatment with 100 mg·kg−1 of verbascoside. * P < .05, significantly different as indicated; log‐rank test. (f) The expression levels of CD44, CCND1, and HIF2α in U251MG‐derived xenografted tumours under verbascoside treatment were tested by real‐time quantitative PCR. (g) The expression levels of SOX2, OCT4, NANOG, Nestin, and ID1 in U251MG‐derived xenografted tumours under verbascoside treatment were tested by real‐time quantitative PCR. The data are presented as mean ± SD from 15 independent experiments for animal experiments (n = 15) and five independent experiments (n = 5) for others, * P < .05, significantly different as indicated

To further determine the mechanism by which verbascoside suppresses GBM, the cleavage of CD44ICD in mouse intracranial tumours was assessed. Western blot showed that CD44ICD cleavage was suppressed by verbascoside (Figure 6d). Furthermore, CD44‐regulating genes were tested. The expression level of HIF2α showed a significant reduction (Figure 6f), which is consistent with the experiment in U251MG cells. On the other hand, the expression of CD44 and CCND1 was not affected (Figure 6f), suggesting that there may be additional regulatory pathways for these genes. Consistent with the results in U251MG cells, both stem‐related genes (Figure 6g) and hallmark genes (Figure S7) were down‐regulated in verbascoside‐treated mice. Taken together, these data indicate that verbascoside is specific for CD44 and has efficacy in glioma tumour models of human cancers.

4. DISCUSSION

Although CD44 is a mesenchymal GBM marker that is highly related to the stem cell‐like phenotype (Phillips et al., 2006), to radiation and chemotherapy resistance and to recurrence, few inhibitors of CD44 have entered human clinical studies. Here, we report the identification of a low MW inhibitor of CD44 (verbascoside) that inhibits CD44 dimerization and activation, with consequent loss of cellular stem cell‐like phenotypes and the suppression of proliferation. This compound was identified by machine‐learning‐based screens, using the interaction between CD44 and hyaluronic acid, which provided many candidates. Those with high scores were subsequently tested with MST assays (data not shown).

The MST assay showed that osteopontin binds to CD44 dimers but not monomers, which is consistent with a previous report. Likewise, hyaluronan (molecular mass 1,200 Da) also binds to CD44 dimers but not monomers. The binding between hyaluronan and CD44ECD‐Fc shows biphasic binding (Seidel et al., 2013). Because hyaluronan is a high MW polymer, we hypothesized that the high affinity binding is caused by both CD44ECD‐Fc dimer binding to one hyaluronan molecule. The low affinity binding is caused by one CD44ECD‐Fc dimer binding to two hyaluronan molecules. This hypothesis is also supported by the MST calculated binding ratio: 1.3 (data not shown). From the MST assays, binding between verbascoside and CD44 shows a more complicated triphasic style. However, we believe that the reason for this finding is relatively simple. The first binding phase is caused by one verbascoside molecule binding to one CD44ECD‐Fc of the dimer, and another verbascoside molecule binds to the other CD44ECD‐Fc in the same dimer, causing the second binding event. The third binding phase may be caused by verbascoside binding to the CD44ECD‐Fc monomer. The EC50 value between CD44 and osteopontin is 94.1 μM. The binding affinity between CD44 dimer and verbascoside is extremely high, with EC50 values of 0.9 and 6.23 μM. Even the binding between the CD44 monomer and verbascoside (EC50 = 90.4 μM) is stronger than that with osteopontin. This result suggests that verbascoside can both compete with osteopontin in unbound CD44 and release osteopontin from the CD44–osteopontin complex. We hypothesize that even binding to one CD44 molecule of the dimer will disrupt the CD44–osteopontin complex. Because CD44 has many native ligands in vivo, this property may be very useful in clinical applications.

The in vivo data showed that the suppression exerted by verbascoside on the tested cell lines was negatively related to the CD44 expression levels. After knocking out CD44 with the CRISPR‐CAS9 system, U251MG cells showed increased tolerance to verbascoside. The relative growth of CD44‐knockout U251MG cells was close to that of PD‐GBMCs. The CD44ECD‐His protein can restore the suppression of proliferation in U251MG cells by verbascoside, in a dose‐dependent manner. By combining in vitro and in vivo data, we can conclude that CD44 is the target of verbascoside.

CD44 activation followed by dimerization and the release of CD44ICD by γ‐secretase is well established. Our data show that verbascoside treatment suppressed the generation of CD44ICD. The expression of CD44 downstream genes, such as CD44, CCND1, and HIF2α, was also significantly suppressed. CD44 participates in a wide variety of functions, including cell adhesion, cell–cell interactions, microfilament regulation, and stemness maintenance. CD44 binds to osteopontin and FGF2 and participates in the HGF and VEGF signalling pathways. Holland et al recently reported that osteopontin plays a crucial role in CD44‐dependent stem‐like phenotype maintenance through HIF2α. The inhibition of γ‐secretase or the knockdown of HIF2α can abolish CD44ICD‐induced stem cell marker gene expression. Our results show that by adding verbascoside, the colony forming ability is weakened in the CD44‐positive cell line U251MG but not in the CD44‐negative cell line PD‐GBMC. In addition, the neurosphere and side population assays showed stemness defects under verbascoside treatment. The expression of the stem cell marker genes SOX2, NANOG, Nestin, OCT4, and ID1 was significantly suppressed. Although there is still debate about the existence of tumour stem cells, the stem cell‐like phenotype of tumour cells is mainly responsible for radiotherapy and chemotherapy resistance, recurrence, and a poor prognosis. CD133 and CD44 are mainly markers of GSC‐like cells. By targeting CD44, verbascoside suppresses the stem cell‐like phenotype of U251MG cells, which may inhibit tumour growth and migration and reduce the risk of recurrence.

The main problem associated with drugs aimed at tumours in the CNS is the ability to penetrate the blood–brain barrier. However, there are many clinical studies on CD44‐targeted drugs, such as antibodies (Avin, Haimovich, & Hollander, 2004; Jin, Hope, Zhai, Smadja‐Joffe, & Dick, 2006), competitive peptides (Hibino et al., 2005), low‐MW HA (Slomiany et al., 2009) and HA‐linked drugs (De Stefano et al., 2011), in acute lymphocytic leukaemia, chronic lymphocytic leukaemia, multiple myeloma, and breast cancer. The application of these drugs to CNS tumours is restricted by their high molecular masses and poor penetration of the blood–brain barrier. By using an in situ U251MG xenograft model, we showed that verbascoside, injected i.p., significantly suppressed tumour growth tumour in the mouse brain and prolonged survival. The reduced concentration of CD44ICD confirmed the CD44 targeting ability of verbascoside.

CD44 is expressed by glioblastoma stem cell‐like cells and many other stem cells, such as adipose‐derived stem cells and haematopoietic stem cells. Stemness phenotype suppression through CD44 may not interfere with CD44‐positive highly differentiated cells but indicates a risk to CD44‐positive stem cells. We assessed this risk in our model by giving four months of treatment with a double concentration of verbascoside. Most organs showed no detectable damage and the weights of the treatment group were not significantly different from those of the control group.

Traditional methods for discovering new drugs have been developed to test a very large number of candidates. This process usually requires much time and money. Here, by combining well‐established machine learning tools and widely used herbal medicine components, we have demonstrated a high‐efficiency and low‐cost method for drug development. The discovery of verbascoside provides not only a valuable tool for the study of CD44 functions and its role in tumour cell stemness phenotype maintenance but also a promising drug for clinical therapy.

In conclusion, this is the first report of a low MW inhibitor of CD44 dimerization. Although CD44 is a very important stem cell‐related protein, there are very few inhibitors of its biological actions. The metalloprotease inhibitor batimastat, the MMP inhibitor GM‐6001 and the γ‐secretase inhibitor DAPT are usually used in CD44‐related research. Although a metalloprotease and γ‐secretase are needed for the activation of CD44, they are also involved in many other signalling pathways. By inhibiting CD44 dimerization specifically, verbascoside can significantly promote relevant research. Previously reported CD44‐targeted drugs are all macromolecules, which restricts their application to the peripheral tissues. Verbascoside is able to cross the blood–brain barrier, which is very important for clinical research. Altogether, these results validate verbascoside as a promising drug for both research and clinical therapy.

AUTHORS' CONTRIBUTIONS

C.W. and Z.W. performed the experiments and wrote the paper, and C.W., Z.T., X.F., J.W., C.C., X.F., and X.Y. collected the data. C.W. and R.X. wrote the paper and analysed the data. All authors read and approved the final manuscript.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design & Analysis, Immunoblotting and Immunochemistry and Animal Experimentation, and as recommended by funding agencies, publishers and other organizations engaged with supporting research.

Supporting information

Table S1: Primers Sequences for qPCR

Click here for additional data file. (15.9KB, docx)

Table S2. Anti‐proliferative EC50 values of verbascoside in cell lines

Click here for additional data file. (14.7KB, docx)

Figure S1: Supporting Information

Click here for additional data file. (4.2MB, tif)

Figure S2: Supporting Information

Click here for additional data file. (1MB, tif)

Figure S3: Supporting Information

Click here for additional data file. (448.7KB, tif)

Figure S4: Supporting Information

Click here for additional data file. (2.4MB, tif)

Figure S5: Supporting Information

Click here for additional data file. (358.3KB, tif)

Figure S6: Supporting Information

Click here for additional data file. (23.6MB, tif)

Figure S7: Supporting Information

Click here for additional data file. (778.9KB, tif)

ACKNOWLEDGEMENTS

This work was supported by a grant from the National Natural Science Foundation of China (Grant No. 81573774) and the Military Medical Science Research Project (16CXZ001).

Wang C, Wang Z, Chen C, et al. A low MW inhibitor of CD44 dimerization for the treatment of glioblastoma. Br J Pharmacol. 2020;177:3009–3023. 10.1111/bph.15030

Chongwu Wang and Zhaotao Wang contributed equally to this work.

Contributor Information

Chongwu Wang, Email: cww_pla_agh@163.com.

Ruxiang Xu, Email: rx_pla_agh@163.com.

REFERENCES

  1. Alexander, S. P. H. , Fabbro, D. , Kelly, E. , Mathie, A. , Peters, J. A. , Veale, E. L. , … CGTP Collaborators . (2019). THE CONCISE GUIDE TO PHARMACOLOGY 2019/20: Enzymes. British Journal of Pharmacology, 176, S297–S396. 10.1111/bph.14752 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Anido, J. , Sáez‐Borderías, A. , Gonzàlez‐Juncà, A. , Rodón, L. , Folch, G. , Carmona, M. A. , … Seoane, J. (2010). TGF‐β receptor inhibitors target the CD44(high)/Id1(high) glioma‐initiating cell population in human glioblastoma. Cancer Cell, 18, 655–668. 10.1016/j.ccr.2010.10.023 [DOI] [PubMed] [Google Scholar]
  3. Avin, E. , Haimovich, J. , & Hollander, N. (2004). Anti‐idiotype x anti‐CD44 bispecific antibodies inhibit invasion of lymphoid organs by B cell lymphoma. Journal of Immunology, 173, 4736–4743. 10.4049/jimmunol.173.7.4736 [DOI] [PubMed] [Google Scholar]
  4. Banerji, S. , Wright, A. J. , Noble, M. , Mahoney, D. J. , Campbell, I. D. , Day, A. J. , & Jackson, D. G. (2007). Structures of the Cd44‐hyaluronan complex provide insight into a fundamental carbohydrate‐protein interaction. Nature Structural & Molecular Biology, 14, 234–239. 10.1038/nsmb1201 [DOI] [PubMed] [Google Scholar]
  5. Barat, S. , Chen, X. , Cuong Bui, K. , Bozko, P. , Gotze, J. , Christgen, M. , … Plentz, R. R. (2017). Gamma‐secretase inhibitor IX (GSI) impairs concomitant activation of notch and Wnt‐β‐catenin pathways in CD44+ gastric cancer stem cells. Stem Cells Translational Medicine, 6, 819–829. 10.1002/sctm.16-0335 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bennett, K. L. , Jackson, D. G. , Simon, J. C. , Tanczos, E. , Peach, R. , Modrell, B. , … Aruffo, A. (1995). CD44 isoforms containing exon V3 are responsible for the presentation of heparin‐binding growth factor. The Journal of Cell Biology, 128, 687–698. 10.1083/jcb.128.4.687 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bhattacharya, R. , Mitra, T. , Ray Chaudhuri, S. , & Roy, S. S. (2018). Mesenchymal splice isoform of CD44 (CD44s) promotes EMT/invasion and imparts stem‐like properties to ovarian cancer cells. Journal of Cellular Biochemistry, 119, 3373–3383. 10.1002/jcb.26504 [DOI] [PubMed] [Google Scholar]
  8. Bourguignon, L. Y. , Zhu, H. , Shao, L. , & Chen, Y. W. (2001). CD44 interaction with c‐Src kinase promotes cortactin‐mediated cytoskeleton function and hyaluronic acid‐dependent ovarian tumor cell migration. The Journal of Biological Chemistry, 276, 7327–7336. 10.1074/jbc.M006498200 [DOI] [PubMed] [Google Scholar]
  9. Brown, D. V. , Daniel, P. M. , D'Abaco, G. M. , Gogos, A. , Ng, W. , Morokoff, A. P. , & Mantamadiotis, T. (2015). Coexpression analysis of CD133 and CD44 identifies proneural and mesenchymal subtypes of glioblastoma multiforme. Oncotarget, 6, 6267–6280. 10.18632/oncotarget.3365 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cho, Y. , Lee, H. W. , Kang, H. G. , Kim, H. Y. , Kim, S. J. , & Chun, K. H. (2015). Cleaved CD44 intracellular domain supports activation of stemness factors and promotes tumorigenesis of breast cancer. Oncotarget, 6, 8709–8721. 10.18632/oncotarget.3325 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Couchman, J. R. (2010). Transmembrane signaling proteoglycans. Annual Review of Cell and Developmental Biology, 26, 89–114. 10.1146/annurev-cellbio-100109-104126 [DOI] [PubMed] [Google Scholar]
  12. Curtarelli, R. B. , Goncalves, J. M. , Dos Santos, L. G. P. , Savi, M. G. , Nor, J. E. , Mezzomo, L. A. M. , & Rodríguez Cordeiro, M. M. (2018). Expression of cancer stem cell biomarkers in human head and neck carcinomas: A systematic review. Stem Cell Reviews, 14, 769–784. 10.1007/s12015-018-9839-4 [DOI] [PubMed] [Google Scholar]
  13. De Stefano, I. , Battaglia, A. , Zannoni, G. F. , Prisco, M. G. , Fattorossi, A. , Travaglia, D. , … Gallo, D. (2011). Hyaluronic acid‐paclitaxel: Effects of intraperitoneal administration against CD44+ human ovarian cancer xenografts. Cancer Chemotherapy and Pharmacology, 68, 107–116. 10.1007/s00280-010-1462-2 [DOI] [PubMed] [Google Scholar]
  14. Di Carlo, D. T. , Cagnazzo, F. , Benedetto, N. , Morganti, R. , & Perrini, P. (2017). Multiple high‐grade gliomas: Epidemiology, management, and outcome. A systematic review and meta‐analysis. Neurosurgical Review, 42, 263–275. 10.1007/s10143-017-0928-7 [DOI] [PubMed] [Google Scholar]
  15. Fan, C. S. , Chen, W. S. , Chen, L. L. , Chen, C. C. , Hsu, Y. T. , Chua, K. V. , … Huang, T. S. (2018). Osteopontin‐integrin engagement induces HIF‐1α‐TCF12‐mediated endothelial‐mesenchymal transition to exacerbate colorectal cancer. Oncotarget, 9, 4998–5015. 10.18632/oncotarget.23578 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gallego, O. (2015). Nonsurgical treatment of recurrent glioblastoma. Current Oncology, 22, e273–e281. 10.3747/co.22.2436 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Godar, S. , Ince, T. A. , Bell, G. W. , Feldser, D. , Donaher, J. L. , Bergh, J. , … Weinberg, R. A. (2008). Growth‐inhibitory and tumor‐suppressive functions of p53 depend on its repression of CD44 expression. Cell, 134, 62–73. 10.1016/j.cell.2008.06.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Harding, S. D. , Sharman, J. L. , Faccenda, E. , Southan, C. , Pawson, A. J. , Ireland, S. , … NC‐IUPHAR . (2018). The IUPHAR/BPS Guide to PHARMACOLOGY in 2018: updates and expansion to encompass the new guide to IMMUNOPHARMACOLOGY. Nucleic Acids Res., 46, D1091–D1106. 10.1093/nar/gkx1121 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hartmann, M. , Parra, L. M. , Ruschel, A. , Lindner, C. , Morrison, H. , Herrlich, A. , & Herrlich, P. (2015). Inside‐out regulation of ectodomain cleavage of cluster‐of‐differentiation‐44 (CD44) and of neuregulin‐1 requires substrate dimerization. The Journal of Biological Chemistry, 290, 17041–17054. 10.1074/jbc.M114.610204 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hibino, S. , Shibuya, M. , Hoffman, M. P. , Engbring, J. A. , Hossain, R. , Mochizuki, M. , … Kleinman, H. K. (2005). Laminin α5 chain metastasis‐ and angiogenesis‐inhibiting peptide blocks fibroblast growth factor 2 activity by binding to the heparan sulfate chains of CD44. Cancer Research, 65, 10494–10501. 10.1158/0008-5472.CAN-05-0314 [DOI] [PubMed] [Google Scholar]
  21. Hsin, K. Y. , Ghosh, S. , & Kitano, H. (2013). Combining machine learning systems and multiple docking simulation packages to improve docking prediction reliability for network pharmacology. PLoS ONE, 8, e83922 10.1371/journal.pone.0083922 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hsin, K. Y. , Matsuoka, Y. , Asai, Y. , Kamiyoshi, K. , Watanabe, T. , Kawaoka, Y. , & Kitano, H. (2016). systemsDock: A web server for network pharmacology‐based prediction and analysis. Nucleic Acids Research, 44, W507–W513. 10.1093/nar/gkw335 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. IARC . (2014). World cancer report 2014. Geneva, Switzerland: World Health Organization. [Google Scholar]
  24. Jafari, S. M. , Joshaghani, H. R. , Panjehpour, M. , Aghaei, M. , & Zargar Balajam, N. (2017). Apoptosis and cell cycle regulatory effects of adenosine by modulation of GLI‐1 and ERK1/2 pathways in CD44+ and CD24− breast cancer stem cells. Cell Proliferation, 50(4), e12345 10.1111/cpr.12345 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Jin, L. , Hope, K. J. , Zhai, Q. , Smadja‐Joffe, F. , & Dick, J. E. (2006). Targeting of CD44 eradicates human acute myeloid leukemic stem cells. Nature Medicine, 12, 1167–1174. 10.1038/nm1483 [DOI] [PubMed] [Google Scholar]
  26. Johansson, E. , Grassi, E. S. , Pantazopoulou, V. , Tong, B. , Lindgren, D. , Berg, T. J. , … Pietras, A. (2017). CD44 interacts with HIF‐2α to modulate the hypoxic phenotype of perinecrotic and perivascular glioma cells. Cell Reports, 20, 1641–1653. 10.1016/j.celrep.2017.07.049 [DOI] [PubMed] [Google Scholar]
  27. Kaul, R. , Saha, P. , Saradhi, M. , Prasad, R. L. , Chatterjee, S. , Ghosh, I. , … Datta, K. (2012). Overexpression of hyaluronan‐binding protein 1 (HABP1/p32/gC1qR) in HepG2 cells leads to increased hyaluronan synthesis and cell proliferation by up‐regulation of cyclin D1 in AKT‐dependent pathway. The Journal of Biological Chemistry, 287, 19,750–19,764. 10.1074/jbc.M111.266270 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Khan, W. S. , Adesida, A. B. , & Hardingham, T. E. (2007). Hypoxic conditions increase hypoxia‐inducible transcription factor 2α and enhance chondrogenesis in stem cells from the infrapatellar fat pad of osteoarthritis patients. Arthritis Research & Therapy, 9, R55 10.1186/ar2211 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kilkenny, C. , Browne, W. , Cuthill, I. C. , Emerson, M. , & Altman, D. G. (2010). Animal research: Reporting in vivo experiments: The ARRIVE guidelines. British Journal of Pharmacology, 160, 1577–1579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kim, C. J. , Terado, T. , Tambe, Y. , Mukaisho, K. I. , Sugihara, H. , Kawauchi, A. , & Inoue, H. (2018). Anti‐oncogenic activities of cyclin D1b siRNA on human bladder cancer cells via induction of apoptosis and suppression of cancer cell stemness and invasiveness. International Journal of Oncology, 52, 231–240. 10.3892/ijo.2017.4194 [DOI] [PubMed] [Google Scholar]
  31. Liu, D. , & Sy, M. S. (1997). Phorbol myristate acetate stimulates the dimerization of CD44 involving a cysteine in the transmembrane domain. Journal of Immunology, 159, 2702–2711. [PubMed] [Google Scholar]
  32. McGrath, J. C. , & Lilley, E. (2015). Implementing guidelines on reporting research using animals (ARRIVE etc.): new requirements for publication in BJP. British Journal of Pharmacology, 172, 3189–3193. 10.1111/bph.12955 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Nagano, O. , & Saya, H. (2004). Mechanism and biological significance of CD44 cleavage. Cancer Science, 95, 930–935. 10.1111/j.1349-7006.2004.tb03179.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Phillips, H. S. , Kharbanda, S. , Chen, R. , Forrest, W. F. , Soriano, R. H. , Wu, T. D. , … Aldape, K. (2006). Molecular subclasses of high‐grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell, 9, 157–173. 10.1016/j.ccr.2006.02.019 [DOI] [PubMed] [Google Scholar]
  35. Pietras, A. (2011). Cancer stem cells in tumor heterogeneity. Advances in Cancer Research, 112, 255–281. 10.1016/B978-0-12-387688-1.00009-0 [DOI] [PubMed] [Google Scholar]
  36. Pietras, A. , Katz, A. M. , Ekstrom, E. J. , Wee, B. , Halliday, J. J. , Pitter, K. L. , … Holland, E. C. (2014). Osteopontin‐CD44 signaling in the glioma perivascular niche enhances cancer stem cell phenotypes and promotes aggressive tumor growth. Cell Stem Cell, 14, 357–369. 10.1016/j.stem.2014.01.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Rimkus, T. K. , Carpenter, R. L. , Sirkisoon, S. , Zhu, D. , Pasche, B. C. , Chan, M. D. , … Lo, H. W. (2018). Truncated glioma‐associated oncogene homolog 1 (tGLI1) mediates mesenchymal glioblastoma via transcriptional activation of CD44. Cancer Research, 78, 2589–2600. 10.1158/0008-5472.CAN-17-2933 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Schapira, A. H. V. (2007). Neurology and clinical neuroscience (Vol. 1336). Philadelphia: Mosby Elsevier. [Google Scholar]
  39. Seidel, S. A. , Dijkman, P. M. , Lea, W. A. , van den Bogaart, G. , Jerabek‐Willemsen, M. , Lazic, A. , … Duhr, S. (2013). Microscale thermophoresis quantifies biomolecular interactions under previously challenging conditions. Methods, 59, 301–315. 10.1016/j.ymeth.2012.12.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Shimoda, M. , Ota, M. , & Okada, Y. (2018). Isolation of cancer stem cells by side population method. Methods in Molecular Biology, 1692, 49–59. 10.1007/978-1-4939-7401-6_5 [DOI] [PubMed] [Google Scholar]
  41. Singh, S. K. , Clarke, I. D. , Terasaki, M. , Bonn, V. E. , Hawkins, C. , Squire, J. , & Dirks, P. B. (2003). Identification of a cancer stem cell in human brain tumors. Cancer Research, 63, 5821–5828. [PubMed] [Google Scholar]
  42. Slomiany, M. G. , Dai, L. , Tolliver, L. B. , Grass, G. D. , Zeng, Y. , & Toole, B. P. (2009). Inhibition of functional hyaluronan‐CD44 interactions in CD133‐positive primary human ovarian carcinoma cells by small hyaluronan oligosaccharides. Clinical Cancer Research, 15, 7593–7601. 10.1158/1078-0432.CCR-09-2317 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Tian, X. , Azpurua, J. , Hine, C. , Vaidya, A. , Myakishev‐Rempel, M. , Ablaeva, J. , … Seluanov, A. (2013). High‐molecular‐mass hyaluronan mediates the cancer resistance of the naked mole rat. Nature, 499, 346–349. 10.1038/nature12234 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Todaro, M. , Gaggianesi, M. , Catalano, V. , Benfante, A. , Iovino, F. , Biffoni, M. , … Stassi, G. (2014). CD44v6 is a marker of constitutive and reprogrammed cancer stem cells driving colon cancer metastasis. Cell Stem Cell, 14, 342–356. 10.1016/j.stem.2014.01.009 [DOI] [PubMed] [Google Scholar]
  45. van der Voort, R. , Taher, T. E. , Wielenga, V. J. , Spaargaren, M. , Prevo, R. , Smit, L. , … Pals, S. T. (1999). Heparan sulfate‐modified CD44 promotes hepatocyte growth factor/scatter factor‐induced signal transduction through the receptor tyrosine kinase c‐Met. The Journal of Biological Chemistry, 274, 6499–6506. 10.1074/jbc.274.10.6499 [DOI] [PubMed] [Google Scholar]
  46. Vandesompele, J. , De Preter, K. , Pattyn, F. , Poppe, B. , Van Roy, N. , De Paepe, A. , & Speleman, F. (2002). Accurate normalization of real‐time quantitative RT‐PCR data by geometric averaging of multiple internal control genes. Genome Biology, 3(7), 34.1–34.12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Was, H. , Czarnecka, J. , Kominek, A. , Barszcz, K. , Bernas, T. , Piwocka, K. , & Kaminska, B. (2018). Some chemotherapeutics‐treated colon cancer cells display a specific phenotype being a combination of stem‐like and senescent cell features. Cancer Biology & Therapy, 19, 63–75. 10.1080/15384047.2017.1385675 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Zoller, M. (2011). CD44: Can a cancer‐initiating cell profit from an abundantly expressed molecule? Nature Reviews. Cancer, 11, 254–267. 10.1038/nrc3023 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1: Primers Sequences for qPCR

Click here for additional data file. (15.9KB, docx)

Table S2. Anti‐proliferative EC50 values of verbascoside in cell lines

Click here for additional data file. (14.7KB, docx)

Figure S1: Supporting Information

Click here for additional data file. (4.2MB, tif)

Figure S2: Supporting Information

Click here for additional data file. (1MB, tif)

Figure S3: Supporting Information

Click here for additional data file. (448.7KB, tif)

Figure S4: Supporting Information

Click here for additional data file. (2.4MB, tif)

Figure S5: Supporting Information

Click here for additional data file. (358.3KB, tif)

Figure S6: Supporting Information

Click here for additional data file. (23.6MB, tif)

Figure S7: Supporting Information

Click here for additional data file. (778.9KB, tif)

Articles from British Journal of Pharmacology are provided here courtesy of The British Pharmacological Society

RESOURCES