Portal de Periódicos da CAPES

BAY 43‐9006/Sorafenib blocks CSF1R activity and induces apoptosis in various classical Hodgkin lymphoma cell lines

2011; Wiley; Volume: 155; Issue: 3 Linguagem: Inglês

10.1111/j.1365-2141.2011.08685.x

1365-2141

Katrin Ullrich, Kathrin D. Wurster, Björn Lamprecht, Karl Köchert, Andreas Engert, Bernd Dörken, Martin Janz, Stephan Mathas,

Cancer Immunotherapy and Biomarkers

The malignant Hodgkin-/Reed-Sternberg (HRS) cells of classical Hodgkin lymphoma (HL) originate in most cases from B cells. However, they have almost completely lost their B cell-specific gene expression and, instead, up-regulate non-B lineage genes (Küppers, 2009). Among the latter, aberrant expression of the oncogenic tyrosine kinase receptor for CSF1, CSF1R, has been demonstrated (Paietta et al, 1990; Lamprecht et al, 2010). CSF1R expression is usually restricted to myeloid cells, but in HRS cells it supports growth and survival (Lamprecht et al, 2010). These data suggested that CSF1R inhibition could be used as treatment strategy for HL. To proceed towards its clinical use, we decided to investigate the effect of the multi-tyrosine kinase inhibitor BAY 43-9006 (Sorafenib) (Wilhelm et al, 2004) on HRS cell lines. BAY 43-9006 was chosen because it is already used for the treatment of various malignancies, is usually well tolerated, and can block CSF1R activity (Kumar et al, 2009). To investigate the ability of BAY 43-9006 to block CSF1R in human lymphoma cell lines, we examined its effect on the HRS cell line L540Cy and anaplastic large cell lymphoma (ALCL) DEL cells (Fig 1A,B). Both cell lines aberrantly express CSF1R and respond to recombinant human (rh) CSF1 (Lamprecht et al, 2010). These cell lines were preincubated with dimethyl sulfoxide (DMSO) or BAY 43-9006 (kindly provided by Bayer Healthcare Pharmaceuticals, Wayne, NJ, USA) at concentrations achievable following BAY 43-9006 administration in patients (Awada et al, 2005). Thereafter, rhCSF1 (216-MC; R&D Systems, Wiesbaden, Germany) was added and CSF1R tyrosine phosphorylation was measured (Fig 1A,B). BAY 43-9006 inhibited CSF1R phosphorylation in a dose-dependent manner, completely preventing its activation between 5 and 10 μmol/l. These data demonstrated that BAY 43-9006 interferes with CSF1R activation in human lymphoma cells and effectively prevents CSF1R activation. BAY 43-9006-mediated inhibition of CSF1R and induction of apoptosis. (A, B) L540Cy (A) and DEL (B) cells were left untreated (−) or stimulated for 5 min with rhCSF1 (+) without or after preincubation with various concentrations of BAY 43-9006 (BAY) or a dimethyl sulfoxide (DMSO) control for 30 min, as indicated. Thereafter, immunoprecipitation was performed with a CSF1R-specific antibody (MAB3291) or the respective isotype control (IC; MAB002; both R&D Systems), and CSF1R tyrosine phosphorylation was measured by immunoblotting with antibody to phospho-tyrosine (p-Tyr; upper panels; sc-7020). As a control, the membrane was reprobed with an antibody to CSF1R (lower panels; sc-692; both Santa Cruz Biotechnology, Heidelberg, Germany). (C) Various Hodgkin-/Reed-Sternberg (HRS) cell lines and the non-Hodgkin cell line Reh were treated for the indicated times with 2·5, 5 or 10 μmol/l BAY 43-9006 or DMSO control, as indicated. Apoptosis was measured by annexin V-FITC/PI staining (Bender MedSystems, Vienna, Austria) and flow cytometry. The percentage of viable, annexin V-FITC/PI-negative cells is shown. One representative out of three independent experiments is depicted. (D–G) Determination of receptor tyrosine kinase (RTK) activity in HRS cell lines under various conditions. The following cell lysates were analysed with a 'Human Phospho-RTK Array Kit' to determine the relative level of tyrosine phosphorylation of 42 different human RTKs in parallel: (D) untreated L540 cells (top panel) and L540 cells treated for 5 min with rhCSF1 (bottom panel); (E) untreated L428 cells; (F) untreated KM-H2 cells; (G) KM-H2 cells treated for 6 h and 20 h with 5 μmol/l BAY 43-9006 or DMSO control, as indicated. *, phospho-tyrosine positive controls. Note, that each RTK is spotted on the membrane as a duplicate. Next, we treated various HRS cell lines (L428, L1236, KM-H2, L591, HDLM-2, L540) and the non-Hodgkin cell line Reh with BAY 43-9006 (Fig 1C). In contrast to HRS cell lines, Reh cells lack CSF1R expression (Lamprecht et al, 2010) and were included as putative negative control. Apoptosis was measured by annexin V - fluorescein isothyocyanate/propidium iodide (annexin V-FITC/PI) staining and flow cytometry, and double-negative cells were considered as non-apoptotic, viable cells. BAY 43-9006 induced apoptosis in HRS cell lines in a time- and dose-dependent manner (Fig 1C), with kinetics similar to those observed with other CSF1R-inhibiting compounds (Lamprecht et al, 2010). In contrast, Reh cells were resistant to concentrations of 2·5 and 5 μmol/l BAY 43-9006 and showed apoptosis only at the highest concentration. Interestingly, the Epstein-Barr virus (EBV)-positive HRS cell line L591 was almost resistant to BAY 43-9006 indicating that EBV might interfere with the action of BAY 43-9006. HRS cells show activation of other receptor tyrosine kinases (RTKs) (Renne et al, 2005) and BAY 43-9006 interferes with various RTKs (Wilhelm et al, 2004; Kumar et al, 2009). To examine, which RTKs were affected by BAY 43-9006, we systematically investigated the RTK status of various HRS cell lines, and monitored changes in their phosphorylation following BAY 43-9006 treatment (Fig 1D–G) by use of a 'Human Phospho-RTK Array Kit' (Proteome ProfilerTM Array ARY001; R&D Systems, investigation of 42 RTKs) (Fig 1D–G). To investigate the efficacy of the assay, L540 cells were left untreated or stimulated with rhCSF1 (Fig 1D). Confirming our IP data (see Fig 1A and Lamprecht et al, 2010), CSF1R showed a strongly increased phosphorylation following rhCSF1 stimulation. Next, untreated L428 (Fig 1E) and KM-H2 (Fig 1F) cells were analysed. Consistent with recent data (Lamprecht et al, 2010), CSF1R phosphorylation was weak in L428 cells but prominent in KM-H2. Interestingly, while CSF1R phosphorylation was common to all HRS cell lines (Fig 1D–1G and Lamprecht et al, 2010), the cell lines varied in the activation pattern of other RTKs. Fibroblast growth factor receptor 3 (FGFR3), platelet-derived growth factor receptor-α (PDGFRΑ) and insulin-like growth factor 1 receptor (IGF1R) were activated in at least two of the cell lines, HGFR (c-MET) and c-RET in particular in L428, and MSPR in L540 cells. Overall, these data were in agreement with published data (Teofili et al, 2001; Renne et al, 2005; Khnykin et al, 2006), showed activation of several RTKs in each cell line as suggested (Renne et al, 2005), and identified c-RET, IGF1R and MSPR as possible targets for future studies. We then investigated changes of the RTK activity in KM-H2 cells following treatment with BAY 43-9006 (Fig 1G). Remarkably, following BAY 43-9006 treatment, only PDGFRΑ and CSF1R showed a rapid and pronounced decrease of activity, suggesting that inhibition of these RTKs indeed might have contributed to BAY 43-9006-induced apoptosis. Finally, we investigated whether RTK-inhibition sensitized the HRS cell lines L540 (Fig 2A,D), KM-H2 (Fig 2B) and L428 (Fig 2C) for the antitumor activity of conventional chemotherapeutics, focusing on doxorubicin (#324380; Calbiochem, Darmstadt, Germany) and vincristine (#677181; Calbiochem), which are both used for HL treatment. Cell lines were treated with the respective chemotherapeutic agents together with BAY 43-9006, at concentrations that, for each substance alone, only moderately induced apoptosis. The combination of the substances resulted in increased apoptosis induction (Fig 2). In particular, we observed a more than additive activity of BAY 43-9006 with vincristine in all cell lines. Induction of apoptosis by BAY 43-9006 in combination with conventional chemotherapeutics. (A) Representative example for detection of apoptosis by annexin V-FITC/PI staining and flow cytometry. L540 cells were treated for 72 h with dimethyl sulfoxide (DMSO) control, with 5 μmol/l BAY 43-9006, 2·5 nmol/l vincristine, or 5 μmol/l BAY 43-9006 in combination with 2·5 nmol/l vincristine, as indicated. The percentage of viable, annexin V-FITC and PI double-negative cells (lower left quadrant) is indicated. (B, C, D) KM-H2 (B), L528 (C) and L540 (D) cells were left untreated or treated with DMSO control, BAY 43-9006 (5 μmol/l), various concentrations of doxorubicin (dox) and vincristine (vin) alone, or dox and vin in combination with BAY 43-9006 (5 μmol/l), respectively. After 72 h (KM-H2, L450) or 96 h (L428), cells were analysed for apoptotic cell death as described in (A). The percentage of viable cells relative to untreated cells is indicated. One representative out of three independent experiments is shown. For statistical analyses, numbers of viable cells per sample were fitted as response to treatment variables (DMSO, BAY, dox, vin) using a linear model taking into account possible interaction effects of BAY 43-9006 with dox or vin. All analysed models explain ∼99% of variance in the experimental data; normal distribution of the residuals was tested using Shapiro-Wilk- or Kolmogorow-Smirnow-Tests. For calculation of significant differences in between treatments, linear model estimators were tested against each other using one-sided Welch′s T-test, P-values were adjusted for multiple testing using Bonferroni correction. i, significant interaction effect. n.s., not significant; *P < 0·05; **P < 0·001. These data demonstrate that BAY 43-9006 induces apoptosis in HRS cell lines, enhances the activity of conventional chemotherapeutics, and exerts its effect – at least in part – by CSF1R inhibition. A recent report investigated the effect of BAY 43-9006 on the HRS cell line L428 (Ambrosini et al, 2010). The authors observed an induction of apoptosis in vitro but no effect in vivo. However, in their study only one cell line was investigated (Ambrosini et al, 2010), and tumours in such xenotransplanation models lack the inflammatory microenvironment of HL. We believe that these data do not argue against the efficacy of BAY 43-9006 for HL treatment. Furthermore, in HL-affected lymph nodes, effects of BAY 43-9006 on the microenvironment, including on CSF1R-dependent macrophages, might potentiate an effect on HRS cells. Therefore, clinical studies of HL patients using BAY 43-9006 alone or in combination with conventional chemotherapeutics are warranted. This work was supported in part by the Deutsche Forschungsgemeinschaft. We thank Constanze Bonifer (Leeds, UK) for helpful comments and critical reading of the manuscript. KU and KDW designed and performed experiments, interpreted data and contributed to the writing of the manuscript; BL designed experiments, interpreted data and contributed to the writing of the manuscript; KK performed statistical analyses and interpreted data; AE, MJ and BD interpreted data and contributed to the writing of the manuscript; SM designed research, interpreted data, wrote the manuscript and supervised the project.