As GalMBP is able to activate the complement pathway following binding to cancer-associated glycan structures, the ability of GalMBP to initiate the complement-dependent cell lysis of MCF7 cells was investigated. Lysis of cancer cells was measured using an enzyme-linked assay that detects the activity of lactate dehydrogenase (LDH) in the medium of cells that have previously been incubated with GalMBP and guinea pig complement (Figure 5.2). LDH was chosen due to its
0.001 C e ll L y s is ( % ) 0 20 40 60 80 100 GalMBP Concentration (µg) 0.01 0.1 1
Figure 5.1 Complement-dependent haemolysis of neuraminidase-treated erythrocytes 8
by GalMBP. Neuraminidase-treated sheep erythrocytes (1 x 10 cells) were sensitised with 1-
500 ng of GalMBP and lysed by incubation with guinea pig complement in a total volume of 1.5 ml at 37°C for one hour. The level of lysis was compared to cells incubated in buffer (0% lysis) or in water (100% lysis) for the same period.
N N + N N I - O O + N - Cl N N N N I - O O + N - Cl H + NAD +
Lactate Pyruvate NADH
+ NADH + NAD + LDH C4 C2 C5b C6 C7 C8 C9 LDH LDH Target cell
Figure 5.2 Overview of the method used for detection of GalMBP-mediated cytotoxicity.
Complement deposition on target cells leads to the activation of the terminal pathway and the formation of membrane attack complexes, a series of large pores within the target cell membrane. These pores allow movement of cytoplasmic proteins out of the cell including the
+
enzyme lactate dehydrogenase (LDH). Reduction of NAD by LDH allows the measurement of enzymatic activity by using the NADH produced to reduce a tetrazolium salt into a coloured formazan product. C5a C3b C3 C3a C3b C5
Membrane attack complexes
+ LDH LDH LDH LDH LDH LDH
cytosolic location and stability. The level of cell killing is compared to cells that have been incubated in medium with complement or a detergent-containing cell lysis solution. GalMBP was not able to induce cell killing above background levels across a broad range of concentrations and levels of complement significantly higher than that required to lyse completely an equivalent number of red blood cells (Table 5.1).
There are several mechanisms used by cells to inhibit complement-dependent cell lysis and these act at different points of the complement pathway. The proteins decay accelerating factor (DAF), membrane co-factor of proteolysis (MCP) and factor H all bind to C3b molecules and facilitate their degradation (Kim and Song, 2006), whereas CD59 acts further down the pathway, preventing formation of the membrane attack complex by inhibiting complement protein C9 binding to C5678 complexes (Kim and Song, 2006) (Figure 1.11). In order to examine whether GalMBP is able to deposit complement protein C3 on the surface of MCF7 cells, adherent cells were incubated with GalMBP and guinea pig complement, washed, and stained with an anti-guinea pig C3 antibody conjugated to horseradish peroxidase. Following further washes, horseradish peroxidase substrate was added and the level of horseradish peroxidase activity was measured. The mean absorbance values for cells incubated with GalMBP or with medium alone were identical (0.25 +/- 0.05), suggesting that any early stage complement components deposited in response to GalMBP binding are cleared efficiently by MCF7 cells.
5.4 Discussion
GalMBP has been shown to trigger complement-dependent cell lysis on erythrocytes that present cancer-associated glycan structures. This assay suggests that GalMBP could initiate a protective response against cancer cells through the activation of complement.
The in vitro assays conducted on MCF7 cells indicates that GalMBP does not deposit complement on the cell surface. This inability to deposit complement on MCF7 cells could be due to the intrinsic protective mechanisms of this cell line, a conclusion that is supported by the finding that MCF7 cells were not lysed by the complement pathway when treated with anti-MCF7 antibodies (Caragine et al., 2002). The in vitro assay system used may also have been limited by several factors. Guinea pig serine proteases are likely to have a reduced affinity for GalMBP in
Table 5.1 Complement-dependent cell lysis of MCF7 cells by GalMBP.
MCF7 cells were grown to confluence in 96-well tissue culture plates and incubated with varying concentrations of GalMBP with no, low (10%), medium (20%) or high (40%) levels of guinea pig complement. Levels of cell lysis were detected by measuring of the activity of the cytosolic enzyme lactate dehydrogenase. Cell killing was calculated by comparing values for cells incubated in medium alone (0%) or a solution containing detergent (100%).
Cell Killing (%) GalMBP
Concentration
(µg/ml) No
Complement ComplementLow ComplementMedium Complement High
0 6 2.4 1.5 5.9
10 4.5 0 3 4.1
50 5 2.9 2.3 3.6
comparison to their rat equivalents. These serine proteases could be MASPs or the classical complement pathway proteases C1r and C1s, which have also been reported to be activated by MBP (Lu et al., 1990). Analysis of human serum suggests that MASPs exist at high levels in serum, with up to 95% not complexed with MBP (Thiel et al., 2000). Conversely, only relatively small amounts of C1r and C1s exist uncomplexed with C1q (Laurell et al., 1976; Thiel et al., 2000), suggesting that GalMBP:MASP complexes are likely to be the predominant species formed in this assay. High levels of MASP1 in serum have been reported (Terai et al., 1997), which could affect the complement fixing ability of GalMBP in this assay by reducing the number of GalMBP:MASP2 complexes that are required for complement deposition.
In order to fully assess the complement-fixing activity of GalMBP in an in vitro system, complexes of rat MASP2 and GalMBP would need to be formed. This is likely to require the purification of rat MASP2 from rat serum using an anti-MASP2 antibody as a capturing reagent. As noted, MASPs have been reported to be present at higher levels than MBPs in serum, so free MASP2 should be present (Thiel et al., 2000). Purification using MASP2 antibodies would have the advantage of precisely defining the MASP associated with GalMBP and would also allow the MASPs to remain in an inactivated state.
It is likely that many of the potential clearance mechanisms used by GalMBP cannot be recreated in an in vitro system and will instead require the use of in vivo animal models. In an in vivo environment, GalMBP could interact with immune system mediators including MASPs, collectin receptors and other uncharacterised MBP-enhanced clearance mechanisms. These uncharacterised mechanisms are highlighted by studies on the therapeutic potential of human MBP to treat colon cancer (Ma et al., 1999). MBP was able to decrease tumour size and increase the survival time of mice subcutaneously injected with a colon cancer cell line. The anti- tumour activity was mediated primarily by an undefined complement-independent mechanism that has been referred to as MBP-dependent cell cytotoxicity (Ma et al., 1999). Assessment of the anti-tumour activity of GalMBP in an animal model would therefore more accurately reflect the effectiveness of retargeted mannose-binding proteins as cancer treatments.