Involvement of macroautophagy in the caspase-independent killing of Burkitt lymphoma cell lines by rituximab


The encouraging clinical impact of rituximab (RTX) upon the treatment of many B-cell lymphoproliferative disorders, including non-Hodgkin lymphoma, is well-documented, although the exact mechanism(s) by which this chimeric anti-CD20 monoclonal antibody eliminates B-cells remains unknown. Studies suggest the involvement of antibody-dependent cellular cytotoxicity, complement-dependent cytotoxicity and classical, caspase-dependent apoptosis (Smith, 2003). More recently we (Daniels et al, 2006), and others (Chan et al, 2003) have found that in addition to these systems, RTX can also kill B-cells via a novel, caspase-independent cell death pathway (CI-CD). Using Burkitt Lymphoma (BL) cell lines as a model of human disease, RTX-induced CI-CD was shown to be characterised by phosphatidylserine exposure, mitochondrial depolarisation, nuclear enlargement, cytoplasmic vacuolisation (Daniels et al, 2006) and a chronic rise in intracellular calcium (Daniels et al, 2008). A complete understanding of the mechanisms of this pathway could have important clinical implications for the treatment of chemo-refractive malignancies in which the classical, caspase-dependent pathway is disabled.

The macroautophagic destruction of macromolecules and organelles begins with the enclosure of the condemned material inside a double-membraned autophagosome that fuses with lysosomes to create an acidic autolysosome in which lysosomal enzymes can break-down and recycle the enclosed cargo (Levine & Kroemer, 2008). Like CI-CD, macroautophagy is associated with a rise in [Ca2+]i (Hoyer-Hansen et al, 2007) and can proceed independently of caspase activation (Xu et al, 2006). We therefore investigated whether RTX-induced CI-CD in BL cell lines shares any further common features with macroautophagy.

Two human BL cell lines (Bjab and MutuI c179, both gifts from Prof. C.D. Gregory, Edinburgh University, Edinburgh, UK) were studied. It has previously been shown that these cell lines undergo RTX-induced CI-CD when exposed to clinically relevant concentrations of RTX (10 μg/ml) (Berinstein et al, 1998) in an experimental system that excludes both complement and accessory cells (Daniels et al, 2006). Following stimulation of BL cells with RTX (Roche Registration Ltd, Herefordshire, UK) the accumulation of the macroautophagy marker, LC3II, was assayed by immunoblotting. Briefly, 5 μg of total cellular protein was separated on 15% wt/vol polyacrylamide gels, transferred to polyvinylidene fluoride membranes and probed for LC3II (Tanida et al, 2005). During macroautophagy the cytosolic microtubule-associated protein 1 light chain 3-I (LC3I) is lipidated to form LC3II, which is localized to and necessary for autophagosome creation. Upon formation, LC3II is broken down and recycled. The lysosomal protease inhibitors E64d and pepstatin-A prevent this rapid proteolysis of LC3II and facilitate its assay (Tanida et al, 2005). Bjab cells exposed to RTX demonstrated a clear increase in LC3II formation, compared to control cells (Fig 1A and B). This rise was completely attenuated in the presence of 0·5 mmol/l 3-MA. In sharp contrast, MutuI cells showed no such increase in LC3II (not shown).

Figure 1.

 (A) and (B) Accumulation of the macroautophagy marker, LC3II, following stimulation with RTX. Bjab cells were incubated for 60 min with 1·0 mmol/l 3-methyladenine (3-MA) or vehicle control in the presence (+, filled bar) or absence (−, open bar) of the lysosomal proteases inhibitors (E-64d and pepstatin-A, each at 10 μg/ml). In some cases RTX (10 μg/ml) was added and the cells were incubated for an additional 24 h. Proteins were separated on 15% wt/vol polyacrylamide gels, transferred to polyvinylidene fluoride membranes and probed for LC3II and β-actin. The blot shown in (A) is representative of five individual experiments. Scanning densitometry of LC3II was undertaken and the results normalised to β-actin, to compensate for any variations in protein loading (B). The results are given as the mean ± SEM of five individual experiments. (C–E) Acidic vesicular organelle staining by acridine orange (AO) in response to RTX. Bjab and MutuI cells were treated with (filled bar) or without (open bar) RTX (10 μg/ml) for 24 h, washed, incubated with AO and analysed by flow cytometry (C) and fluorescence microscopy (D and E). In (C) the results are given as the mean fluorescence ± SEM of four separate experiments, whilst (D) gives the mean number ± SEM of AO-stained Bjab cells also derived from four separate experiments. In each experiment at least 50 cells were counted. Representative images of Bjab cells treated with and without RTX are given in (E). P-values show the level of significance between the two indicated conditions (two-sided, paired t-test). CON = control.

To confirm these findings, we investigated acidic vesicular organelle (AVO) formation in response to RTX, since AVO formation is intimately linked to autophagic cell death (Xu et al, 2006). Acridine orange (AO) is an acidotropic dye that crosses the cell membrane and rapidly becomes protonated and trapped inside acidic compartments, such as the autophagosome. Fluorescence due to staining with AO is thus proportional to the number of intracellular AVOs. Cells were washed with phosphate buffered saline (PBS), re-suspended in AO (5 μg/ml in PBS), incubated at 37°C (20 min) and analysed (10 000 events) by flow cytometry. For fluorescent microscopy, cells were applied to glass slides, air dried, mounted in PBS: glycerol (50:50 v/v) and immediately visualised using an inverted fluorescent microscope (Olympus UK Ltd, Southall, Middlesex, UK) with SmartCapture VP software (Digital Scientific Ltd, Cambridge, UK). Figure 1C shows that the mean fluorescence due to AVOs increased significantly in Bjab but not MutuI cells following 24-h stimulation with RTX. Fluorescence microscopy confirmed these data, showing that incubation of Bjab with RTX increased AO stained cell number, compared with control (Fig 1D and E). In support of the data obtained for LC3II and flow cytometric analysis of AO, MutuI cells showed no significant increase in AO staining in response to RTX (data not shown).

Cell death was determined by dual staining of cells with Annexin V (AV) and propidium iodide (PI) to determine the population of cells that exposed phosphotidylserine (AV+ve), but which retained membrane integrity (PI−ve) (Daniels et al, 2006). RTX induced CI-CD in Bjab and MutuI cells, although 3-MA only protected Bjab cells from death. Interestingly, in contrast to the results reported for LC3II, 3-MA failed to completely protect Bjabs from RTX-driven CI-CD (Fig 2A and B).

Figure 2.

 Effect of the autophagy inhibitor, 3-MA, upon RTX-induced cell death. Bjab (A) and MutuI (B) cells were incubated for 60 min with 0, 0·5 or 1·0 mmol/l 3-methyladenine (3-MA). Vehicle control (open bar) or RTX (filled bar) was added and after 24 h the cells were stained with Annexin V (AV)/propidium iodide (PI) and analysed by flow cytometry (10 000 events collected). The results are given as the mean ± SEM of five individual experiments. P-values give the level of significance with respect to the RTX-treated condition (two-sided, paired t-test).

This study presents evidence to suggest that RTX can induce macroautophagy in some, but not all BL cell lines. In Bjab cells, RTX significantly upregulated the accumulation of the macroautophagy marker, LC3II. This process was completely abolished by 0·5 mmol/l 3-MA and did not occur in the MutuI cell line. Further, an elevation in the number of AVOs was witnessed in the Bjab but not the MutuI line. Our preliminary data indicated that rapamycin induces a macroautophagic response in MutuI cells, suggesting that this pathway is operational in these cells (not shown).

To exclude the possibility that the macroautophagic response observed here was a failed pro-survival defence mechanism (Eskelinen, 2005) against RTX, we studied the impact of 3-MA upon RTX-induced CI-CD. In agreement with our LC3II and AVO data, 3-MA significantly protected Bjab but not MutuI cells from RTX-induced CI-CD. This data supports a role for macroautophagy in the RTX-induced death pathway in Bjab but not MutuI cells. It is noteworthy that whilst 0·5 mmol/l 3-MA completely abolished the RTX-induced rise in LC3II, this concentration and those approaching cytotoxic levels (1·0 mmol/l), offered only partial (though statistically significant) protection from CI-CD. These results imply that even in a single cell line multiple death pathways may co-exist and be simultaneously operational. We conclude that our evidence for macroautophagy induction in the Bjab but not the MutuI line indicates that RTX induces CI-CD by different mechanisms in different BL cell lines.