- Letter to the Editor
- Open Access
Shedding of bevacizumab in tumour cells-derived extracellular vesicles as a new therapeutic escape mechanism in glioblastoma
Molecular Cancervolume 17, Article number: 132 (2018)
Glioblastoma (GBM) is the most aggressive type of primary brain tumours. Anti-angiogenic therapies (AAT), such as bevacizumab, have been developed to target the tumour blood supply. However, GBM presents mechanisms of escape from AAT activity, including a speculated direct effect of AAT on GBM cells. Furthermore, bevacizumab can alter the intercellular communication of GBM cells with their direct microenvironment. Extracellular vesicles (EVs) have been recently described as main acts in the GBM microenvironment, allowing tumour and stromal cells to exchange genetic and proteomic material. Herein, we examined and described the alterations in the EVs produced by GBM cells following bevacizumab treatment. Interestingly, bevacizumab that is able to neutralise GBM cells-derived VEGF-A, was found to be directly captured by GBM cells and eventually sorted at the surface of the respective EVs. We also identified early endosomes as potential pathways involved in the bevacizumab internalisation by GBM cells. Via MS analysis, we observed that treatment with bevacizumab induces changes in the EVs proteomic content, which are associated with tumour progression and therapeutic resistance. Accordingly, inhibition of EVs production by GBM cells improved the anti-tumour effect of bevacizumab. Together, this data suggests of a potential new mechanism of GBM escape from bevacizumab activity.
GBM is amongst the most aggressive types of brain tumours for which current treatments are of limited benefit . During the past decades, AAT have provided a rationale for targeting and blocking the tumour blood supply. Unfortunately, the effects of AAT/bevacizumab, a monoclonal humanised antibody neutralising Vascular Endothelial Growth Factor-A (VEGF-A), on tumour growth are short-term and GBM patients ultimately relapse. Interestingly, since the expression of some pro-angiogenic factors and their receptors (i.e. VEGF-A/VEGF-R) has been described in tumour cells, it appears that AAT/bevacizumab also acts directly on GBM cells  that might eventually lead to therapy resistance and relapse . Recently, we identified a direct effect of bevacizumab on GBM cells and demonstrated its ability to stimulate tumour cells’ invasion in hyaluronic acid (HA) hydrogels and activate key survival signalling pathways. The intrinsic reluctance of cancer cells to AAT could also be linked to their ability of disposing the drugs . Indeed, it has been observed that cetuximab, an EGF-R monoclonal IgG1 antibody, is associated with extracellular vesicles (EVs) derived from treated cancer cells suggesting that such processes could be implicated in tumour limited response to therapy . During the last years, EVs involvement in tumour development and metastasis has been thoroughly considered . Therefore, herein we focused on the effects of bevacizumab on the production of GBM cells-derived EVs.
Bevacizumab affects the EVs proteomic content derived from GBM cells
Since VEGF-A represents the main target of bevacizumab and in order to determine the best model for our study, we examined the expression of different components of the VEGF-A signalling in three different GBM cell lines (see Additional file 1: for Materials and Methods). As LN18 and U87 secreted the highest amounts of VEGF-A, we decided to focus on the effects of bevacizumab on these cell lines (Additional file 2: Figure S1a, S1b). Although bevacizumab neutralised VEGF-A secreted by LN18 and U87 (Additional file 2: Figure S1c), cell viability and proliferation appeared to be marginally affected with clinically relevant doses (~ 0.25 mg/mL), while the only statistically significant decrease on GBM viability (~ 10%) and proliferation (~ 30%) was observed with high doses (Additional file 2: Figure S1d, S1e). Moreover, nanoparticles tracking analysis (NTA) showed no significant change in the concentration of LN18 or U87 cells-derived EVs (~ 1000 and ~ 3000 particles/mL/cell, respectively) in response to bevacizumab (Fig. 1a, b), while MS analysis showed that treatment with either bevacizumab or control IgG1 could modify the proteomic cargo of EVs derived from GBM cells (Additional file 3: Figure S2 and Additional file 4: Tables S1 and S2). Interestingly, the fact that even ‘unspecific’ IgG1 altered the EVs proteomic cargo, suggests that GBM cells respond to the immunotherapy. Furthermore, immunoglobulins peptides could be noticed in EVs derived from both IgG1- and bevacizumab-treated LN18 and U87, suggesting that the antibody used associates somehow with EVs in a non-VEGF-A specific way. Moreover, Annexin A2 expression, a well-described angiogenesis and tumour progression promoter in gliomas and breast cancer, increased in both LN18 and U87 cells-derived EVs following bevacizumab treatment (Fig. 1c). Annexin A2 has also been described as a new potential marker for GBM aggressiveness and patients’ survival . We also observed a decrease in CD44 expression in U87 cells-derived EVs following bevacizumab treatment. Yet, a decrease in CD44 along with an increase in Annexin A2 might suggest a switch in the EVs sub-populations upon bevacizumab treatment . Finally, respective patients’ gene expression in the different GBM subtypes has been obtained from The Cancer Genome Atlas (TCGA). As illustrated in Fig. 1d, Annexin A2 and CD44 are mostly over-expressed in the classical and mesenchymal subtypes. Overall, further investigation is required to decipher whether EVs derived from bevacizumab treated-GBM cells promote tumour aggressiveness through their protein content.
Bevacizumab is internalised by GBM cells
We then assessed the presence of bevacizumab in GBM cells. As previously observed , we showed a partial co-localisation of bevacizumab with EEA1 (and Rab5 in U87) in GBM cells, suggesting that the antibody uptake might involve receptor/ligand dependent endocytosis (Fig. 2a, Additional file 5: Figure S3a). Western blotting revealed a time-dependent increase of bevacizumab in treated-GBM cells, suggesting a gradual incorporation of bevacizumab over time (Fig. 2b). This mechanism appears to be quite rapid as suggested by the detection of bevacizumab in EVs following 2 h treatment, which could happen via fast recycling endosomes as recently proposed .
Bevacizumab is detectable at the surface of GBM cells-derived EVs
We hypothesised that IgG1 antibodies could end up at the surface of EVs derived from treated cells. When compared to control EVs, TEM revealed bevacizumab as aggregates on some of the EVs from treated cells, suggesting that the antibody is present at the surface of the vesicles (Fig. 2c). Accordingly, bevacizumab could be observed in the EVs produced by the respective cells following treatment (Fig. 2d). Moreover, trypsin digestion could affect IgG antibodies in EVs, suggesting that bevacizumab is predominantly bound to the surface of EVs and is not internalised in them. (Fig. 2e and Additional file 5: Figure S3b).
Finally, co-immunoprecipitation experiments revealed that EVs-associated bevacizumab could not bind to VEGF-A. As previously suggested , we believe that GBM cells derived-EVs trapping of bevacizumab reduces the available number of antibody molecules reaching their specific targets, thus decreasing its efficacy. Interestingly, fibronectin was detected in both IgG1 and bevacizumab-bound solubilised EVs (Fig. 2f). Accordingly, Deissler et al. suggest that endothelial cells could uptake IgG1 antibodies through binding to fibronectin/integrin complexes . Therefore, it is likely that IgG1 antibodies (bevacizumab) get trapped at the cell surface through binding with cell/ECM interaction components (i.e. fibronectin) . Consequently, bevacizumab shedding at the surface of GBM-derived EVs appears as a possible way for cancer cells to shield themselves against treatment.
Inhibition of EVs production increases the effects of bevacizumab on GBM cells’ viability
The effects of EVs depletion on the viability of GBM cells in combination with bevacizumab treatment have been investigated, using GW4869, a sphyngomyelinase inhibitor shown to decrease EVs biogenesis. GW4869 inhibited EVs production (Fig. 3a), while, combination of bevacizumab with GW4869 significantly affected (~ 30%) the U87 cells’ viability after 48 h (Fig. 3b), suggesting that inhibition of EVs production could contribute to the cytotoxicity of bevacizumab in vitro. Finally, U87 invasiveness was slightly increased (~ 28%) following bevacizumab treatment , as assessed by their ability to form colonies in a tri-dimensional HA hydrogel (Fig. 3c), while combination with GW4869 reduced this effect. Thus, the direct anti-tumour effect of bevacizumab might be enhanced via alteration of EVs biogenesis in GBM cells. As we did not observe similar effects on LN18 invasiveness, specific molecular pathways might be implicated in the response to both bevacizumab and GW4869, as observed before . It is still unclear if the additive effect on U87 cells is only due to the specific inhibition of the antibody shedding mechanisms. It might as well be a combinational effect of VEGF-A neutralization by bevacizumab along with blocking of pro-tumoral EVs (Additional file 6: Figure S4).
To what extent such mechanisms can affect bevacizumab efficacy in vivo is yet to be deciphered. Still, similar mechanisms have been described in different cancer cell types treated with antibodies. Hence, such shedding of IgG1 antibodies into GBM cells-derived EVs could be involved in GBM cells’ escape from bevacizumab. As EVs can be vividly uptaken by other cells, the destination of such bevacizumab-coated EVs is another outstanding question that require further in vivo studies.
Bevacizumab has a positive effect on GBM patients’ quality of life and survival, mostly through its anti-inflammatory effects. Consequently, there is an urgent need for the right therapeutic strategy that will increase its efficacy on tumour growth, thus avoiding recurrence and relapse. Here, we report that the paradoxical pro-invasive effect of bevacizumab on GBM cells might be due to alterations in the tumour cells-derived EVs, including shedding of the antibody and further modifications of the cargo, both possibly contributing to therapeutic resistance. Therefore, this study suggests the combination of bevacizumab with a local blocking of EVs-dependent intercellular communication as a potential new therapeutic strategy to improve GBM treatment.
Dulbecco’s Modified Eagle Medium
Epidermal growth factor-receptor
Fetal bovine serum
Minimal essential medium
Nanoparticles tracking analysis
Penicillin, streptomycin, glutamine
The cancer genome atlas
Transmission electron microscopy
Vascular endothelial growth factor-A
Simon T, Coquerel B, Petit A, Kassim Y, Demange E, Le Cerf D, Perrot V, Vannier J-P. Direct effect of Bevacizumab on Glioblastoma cell lines in vitro. NeuroMolecular Med. 2014;16(4):752–71.
Simon T, Gagliano T, Giamas G. Direct effects of anti-Angiogenic therapies on tumor cells: VEGF signaling. Trends Mol Med. 2017;23:282–92.
Gotink KJ, Broxterman HJ, Labots M, de Haas RR, Dekker H, Honeywell RJ, Rudek MA, Beerepoot LV, Musters RJ, Jansen G, Griffioen AW, et al. Lysosomal sequestration of sunitinib: a novel mechanism of drug resistance. Clin Cancer Res. 2011;17(23):7337–46.
van Dommelen SM, van der Meel R, van Solinge WW, Coimbra M, Vader P, Schiffelers RM. Cetuximab treatment alters the content of extracellular vesicles released from tumor cells. Nanomedicine (Lond). 2016;11:881–90.
Wendler F, Favicchio R, Simon T, Alifrangis C, Stebbing J, Giamas G. Extracellular vesicles swarm the cancer microenvironment: from tumor-stroma communication to drug intervention (in press). Oncogene. 2016;
Li S, Zou H, Shao YY, Mei Y, Cheng Y, Hu DL, Tan ZR, Zhou HH. Pseudogenes of annexin A2, novel prognosis biomarkers for diffuse gliomas. Oncotarget. 2017;8:106962–75.
Kowal J, Arras G, Colombo M, Jouve M, Morath JP, Primdal-Bengtson B, Dingli F, Loew D, Tkach M, Thery C. Proteomic comparison defines novel markers to characterize heterogeneous populations of extracellular vesicle subtypes. Proc Natl Acad Sci U S A. 2016;113:E968–77.
Deissler HL, Lang GK, Lang GE. Internalization of bevacizumab by retinal endothelial cells and its intracellular fate: evidence for an involvement of the neonatal fc receptor. Exp Eye Res. 2016;143:49–59.
Muller-Greven G, Carlin CR, Burgett ME, Ahluwalia MS, Lauko A, Nowacki AS, Herting CJ, Qadan MA, Bredel M, Toms SA, et al. Macropinocytosis of Bevacizumab by Glioblastoma cells in the perivascular niche affects their survival. Clin Cancer Res. 2017;23:7059–71.
Aung T, Chapuy B, Vogel D, Wenzel D, Oppermann M, Lahmann M, Weinhage T, Menck K, Hupfeld T, Koch R, et al. Exosomal evasion of humoral immunotherapy in aggressive B-cell lymphoma modulated by ATP-binding cassette transporter A3. Proc Natl Acad Sci U S A. 2011;108:15336–41.
This work was supported by Action Against Cancer, The Colin McDavid Family Trust, The Rothschild Foundation, The Bernard Sunley Foundation, The Searle Memorial Charitable Trust, Mr. Alessandro Dusi and Mr. Milan Markovic.
Availability of data and materials
Data generated or analysed during this study are included in this published article [and its supplementary information files]. The raw mass spectrometry datasets analysed during the current study are available from the corresponding author on reasonable request. TCGA data are available online: ‘Expression box plot (Affymetrix HT HG U133A)’ and ‘Expression box plot (Affymetrix Human Exon 1.0 ST)’ graphs on the Betastasis website (www.betastasis.com).
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Materials and methods. (DOCX 22 kb)
Supplementary Figure 1. Direct effect of bevacizumab on LN18 and U87 GBM cells. (ZIP 375 kb)
Supplementary Figure 2. MS protein hits identified in LN18 and U87 GBM cells-derived EVs following 24h treatment with 0.25 mg/mL IgG1/bevacizumab (FunRich analysis). (ZIP 842 kb)
Supplementary Table 1. Mass spectrometry analysis of EVs content following treatment with IgG1/bevacizumab. Table 2. Mass spectrometry analysis of EVs content following treatment with IgG1/bevacizumab. (ZIP 504 kb)
Supplementary Figure 3. Bevacizumab is detectable in GBM cells and on GBM cells-derived EVs following treatment. (ZIP 1930 kb)
Supplementary Figure 4. Shedding of bevacizumab in tumour cells-derived extracellular vesicles as a new therapeutic resistance mechanism in glioblastoma. (ZIP 94 kb)