Lipoplex mediated silencing of membrane regulators (CD46, CD55 and CD59) enhances complement‐dependent anti‐tumor activity of trastuzumab and pertuzumab

Abstract

The therapeutic potential of anticancer antibodies is limited by the resistance of tumor cells to complement‐mediated attack, primarily through the over‐expression of membrane complement regulatory proteins (mCRPs: CD46, CD55 and CD59). Trastuzumab, an anti‐ HER2 monoclonal antibody, approved for the treatment of HER2‐positive breast and gastric cancers, exerts only minor complement‐mediated cytotoxicity (CDC). Pertuzumab is a novel anti‐HER2 monoclonal antibody, which blocks HER2 dimerization with other ligand‐activated HER family members. Here, we explored the complement‐mediated anti‐tumor effects of trastuzumab and pertuzumab on HER2‐positive tumor cells of various histological origins.

Delivery of chemically stabilized anti‐mCRP siRNAs using cationic lipoplexes, AtuPLEXes, to HER2‐over‐expressing BT474, SK‐BR‐3 (breast), SKOV3 (ovarian) and Calu‐3 (lung) cancer cells reduced mCRPs expression by 85–95%. Knockdown of individual complement regulators variably led to increased CDC only upon combined treatment with trastuzumab and pertuzumab. The combined down‐regulation of all the three regulators augmented CDC by 48% in BT474, 46% in SK‐BR‐3 cells, 78% in SKOV3 cells and by 30% in Calu‐3 cells and also increased complement‐induced apoptosis and caspase activity on mCRP neutralized tumor cells. In addition, antibody‐induced C3 opsonization of tumor cells was significantly enhanced after mCRP silencing and further augmented tumor cell killing by macrophages.

Our findings suggest that siRNA‐induced inhibition of complement regulator expression clearly enhances complement‐ and macrophage‐mediated anti‐tumor activity of trastuzumab and pertuzumab on HER2‐positive tumor cells. Thus – if selectively targeted to the tumor – siRNA‐induced inhibition of complement regulation may serve as an innovative strategy to potentiate the efficacy of antibody‐based immunotherapy.

Highlights

• ► Delivery of 2′‐O‐methyl modified siRNAs to CD46, CD55 and CD59 (mCRPs) using lipoplex.
• ► mCRPs inhibition sensitizes tumor cells to trastuzumab and pertuzumab induced complement attack.
• ► Enhanced C3 opsonization of HER2‐positive tumor cells upon mCRP silencing.
• ► Increased killing of opsonized tumor cells by macrophages.
• ► Enhanced anti‐tumor effect of trastuzumab and pertuzumab upon mCRP inhibition.

Abbreviations

CDC

complement-dependent cytotoxicity

CD46

membrane cofactor protein

CD55

decay accelerating factor

CD59

protectin

MAC

membrane attack complex

mCRP

membrane-bound complement regulatory protein

RNAi

RNA interference

siRNA

small interfering RNA

NHS

normal human serum

1. Introduction

Complement as an indispensable component of the innate immunity plays a major role in host defence against microbial pathogens and clearance of immune complexes. Upon complement activation, biologically active peptides are released, which mediate effector functions such as cytotoxicity, leukocyte chemotaxis, opsonization with enhanced phagocytosis and release of multiple mediators of inflammation (Walport, 2001). Host cells are protected from accidental complement attack by expressing membrane‐bound complement regulatory proteins (mCRPs), including membrane cofactor protein (CD46), decay‐accelerating factor (CD55) and protectin (CD59). CD46 and CD55 control C3/C5 convertase activation (Kojima et al., 1993; Medof et al., 1984) and CD59 blocks the terminal complement pathway, thereby preventing MAC formation (Meri et al., 1990).

The potential role of complement in the control of malignant cells has been emphasized by various studies, where complement is required for the therapeutic activity of rituximab (Golay et al., 2006; Manches et al., 2003) and ofatumumab (Teeling et al., 2004). Apart from the direct killing of tumor cells, complement can opsonize tumor cells and facilitate cellular cytotoxicity by employing complement receptor 3 (CR3, CD11b/CD18) on immune cells (Klein et al., 1990; Leidi et al., 2009; Li et al., 2006).

Over‐expression of membrane regulators has been reported in many primary cancers and tumor cell lines and appears to play an important role in tumor immune evasion (Fishelson et al., 2003; Gelderman et al., 2004; Yan et al., 2008). Lung cancer cells over‐express CD55 and CD46 and are, consequently, complement resistant relative to normal primary lung tissue (Varsano et al., 1998). In colorectal carcinoma, high expression levels of CD55 or CD59 correlated with the degree of differentiation and poor prognosis of the disease (Durrant et al., 2003; Watson et al., 2006). CD59 expression has been shown to be associated with the resistance to rituximab therapy in patients with B‐cell malignancies (Treon et al., 2001). Inhibition of CD55 and CD59 reversed resistance to rituximab‐mediated complement lysis (Macor et al., 2007). We previously reported that neutralization of membrane regulators by monoclonal antibodies or posttranscriptional gene silencing increases complement‐mediated lysis of tumor cells (Donin et al., 2003; Geis et al., 2010; Jurianz et al., 2001; Zell et al., 2007).

HER2 (Human Epidermal Growth Factor Receptor‐2, c‐erb‐B2/neu), a proto‐oncogene, encodes a M_r 185 kDa transmembrane glycoprotein that belongs to a family of receptor tyrosine kinases, which activate signal transduction cascades that regulate cell proliferation, differentiation, and apoptosis by forming homodimers and heterodimers (Yarden and Sliwkowski, 2001). HER2 overexpression has been reported in various cancers, comprising 20–25% of cases in breast cancers (Owens et al., 2004), 20–30% in ovarian cancers (Hellstrom et al., 2001), and 4–6% in lung tumors (Hirsch et al., 2002). Consequently, inhibiting HER2 mediated signaling processes provides more benefit to patients with HER2 overexpressing cancers (Shepard et al., 1991).

Trastuzumab (Herceptin) is a recombinant humanized IgG1_k monoclonal antibody directed against the extracellular domain of HER2. It exerts its anti‐tumor activity by blocking ligand‐independent HER2 signaling, inhibition of HER2 extracellular domain shedding (Molina et al., 2001), as well as the induction of antibody‐dependent cellular cytotoxicity (ADCC) (Barok et al., 2007; Clynes et al., 2000; Leidi et al., 2009). It has been approved for the treatment of HER2‐positive breast cancer in all lines of treatment and advanced metastatic gastric cancer. Pertuzumab is a new humanized IgG1 monoclonal antibody that binds to domain II of HER2. Pertuzumab inhibits the dimerization of HER2 with other HER family proteins and blocks ligand‐dependent HER2 signaling, thus inhibiting tumor growth and progression (Franklin et al., 2004). The combination of both trastuzumab and pertuzumab showed synergistic anti‐tumor activity on breast cancer cells (Nahta et al., 2004), in breast and lung cancer xenograft (Scheuer et al., 2009) as well as in ovarian cancer xenograft models (Faratian et al., 2011). A phase III trial of trastuzumab and pertuzumab combination treatment together with docetaxel in HER2‐positive metastatic breast cancer patients demonstrated highly significant improvement of the progression‐free survival, and a strong positive trend at an early interim analysis of overall survival (Baselga et al., 2012). These results led to U.S. FDA approval of pertuzumab for first‐line treatment of HER2‐positive metastatic breast cancer in combination with trastuzumab and docetaxel.

The contribution of complement to the anti‐tumor effect of trastuzumab and pertuzumab is less clear. The augmentation of immune‐mediated effector functions has been suggested as a promising approach for enhancement of the efficacy of therapeutic antibodies (Boross and Leusen, 2012; Gelderman et al., 2004).

Aim of the present investigation was to improve the anti‐tumor activity of trastuzumab and pertuzumab by overcoming the complement resistance on HER2‐positive tumor cells via silencing the expression of membrane regulatory proteins.

RNA interference (RNAi) mediated by small interfering RNA (siRNA) is the most efficient strategy for specific silencing of therapeutically relevant genes (Aagaard and Rossi, 2007). In the last years a great number of strategies has been developed for efficient delivery of siRNAs in vitro and in vivo (Burnett and Rossi, 2012). AtuPLEX is a novel cationic lipid based siRNA delivery system, consisting of the cationic non‐PEG modified lipid AtuFECT01, a neutral fusogenic co‐lipid and the corresponding siRNA. These components are forming a lipoplex with improved physico‐chemical properties, cellular uptake and improved siRNA release from endosomes after endocytosis (Santel et al., 2006a, 2006b).

Here, we designed chemically stabilized 2′‐O‐methyl modified siRNAs to specifically knockdown the expression of CD46, CD55 and CD59 on HER2‐positive tumor cells using cationic liposomes (AtuPLEXes). Upon silencing of mCRP expression, combined trastuzumab and pertuzumab treatment led to increased complement‐mediated necrosis and apoptosis as well as enhanced tumor cell opsonization of HER2/neu overexpressing BT474, SKBR‐3 (breast), SKOV3 (ovarian) and Calu‐3 (lung) cancer cells.

2. Materials and methods

2.1. Cell lines

BT474 breast carcinoma cells (ATCC number: HTB‐20) were cultured in RPMI 1640 (PAA laboratories, Pasching, Austria), SK‐BR‐3 breast adenocarcinoma cells (ATCC number: HTB‐30) in McCoy's 5a (PAN Biotech, Aidenbach, Germany), SKOV3 ovarian adenocarcinoma cells (ATCC number: HTB‐77) in DMEM (PAA laboratories) and Calu‐3 lung adenocarcinoma cells (ATCC number: HTB‐55) in EMEM (Lonza, Verviers, Belgium). For all cell lines, the medium was supplemented with 10% (v/v) fetal bovine serum (PAN Biotech) and cells were maintained at 37 °C in a humidified atmosphere with 5% CO₂.

2.2. Antibodies and other reagents

Trastuzumab and pertuzumab were obtained from Roche Diagnostics GmbH (Penzberg, Germany).

Monoclonal anti‐CD46 antibody (IgG1, clone GB24) was kindly provided by Dr. J. Atkinson (Washington University, St. Louis MO). Anti‐CD55 (IgG1, clone Bric 110), and anti‐CD59 antibodies (IgG2b, clone Bric 229) were from International Blood Group Reference Laboratory (Birmingham, UK). For the detection of surface bound complement C3, monoclonal mouse anti‐human C3d (IgG1, clone A207, Quidel, San Diego, CA, USA) was used. Goat anti‐human IgG (Fc specific)‐FITC was from Sigma Aldrich (Munich, Germany) and $\text{F}{\left(\text{ab}\right)}_2^{\prime }$ goat anti‐mouse IgG‐FITC was from Jackson ImmunoResearch Laboratories (West Grove, PA, USA).

Normal human serum (NHS) was used as a source of complement. A pool of sera from ten healthy donors was prepared and stored frozen at −70 °C. Heat‐inactivated serum (56 °C, 30 min) was used as negative control in all the experiments.

2.3. siRNA molecules and preparation of siRNA‐lipoplexes

The siRNA molecules (BioSpring, Frankfurt, Germany) used in the study were blunt end, 19‐mer double‐stranded RNA oligonucleotides chemically stabilized by alternating 2′‐O‐methyl sugar modifications on both strands (Czauderna et al., 2003) where the unmodified nucleotides face modified ones on the opposite strand. The siRNA molecules used in the study were:

• siRNA control (random sequence)

  • 5′ ugcagguuuauaguccaca 3′ (sense)
  • 5′ uguggacuauaaaccugca 3′ (antisense)
• siRNA anti‐CD46

  • 5′ gagagagcacgauuuauug 3′ (sense)
  • 5′ caauaaaucgugcucucuc 3′ (antisense)
• siRNA anti‐CD55

  • 5′ gaagaguucugcaaucgua 3′ (sense)
  • 5′ uacgauugcagaacucuuc 3′ (antisense)
• siRNA anti‐CD59

  • 5′ gcaagaaggaccuguguaa 3′ (sense)
  • 5′ uuacacagguccuucuugc 3′ (antisense)

Nucleotides with 2′‐O‐methyl modifications are underlined.

Cationic liposomes, AtuFECT01/DPhyPE, were from Silence Therapeutics AG (Berlin, Germany). siRNA‐lipoplexes (AtuPLEXes) were prepared as described previously (Santel et al., 2006b). Briefly, to generate 10 μM stock AtuPLEX, preformed liposomes (1.8 mg/ml in 270 mM sucrose solution) were mixed with an equal volume of a 0.25 mg/ml of siRNA in 270 mM sucrose solution. The resulting AtuPLEXes were used directly or stored at 4 °C.

2.4. In vitro siRNA transfection

Tumor cells were seeded 24 h before transfection and grown to 50–60% confluence. Appropriate amounts of AtuPLEX solution with siRNAs at a final concentration of 50 nM were added directly to the cells. After incubation for 24 h, cells were washed and fresh medium was replaced. Functional analysis was performed 72–96 h after siRNA transfection.

2.5. FACS analysis

Tumor cells were detached from the culture plates by Trypsin‐EDTA solution (PAA Laboratories, Pasching, Austria), washed twice and re‐suspended in FACS buffer (1% BSA, 0.1% NaN₃ in PBS). 1 × 10⁵ cells were incubated in 100 μl FACS buffer containing the first antibody or a corresponding isotype control (final concentration 10 μg/ml) and incubated for 30 min on ice. The cells were washed twice with FACS buffer, re‐suspended in 100 μl FACS buffer containing the respective FITC‐conjugated secondary antibody and incubated for 30 min on ice. Cells were washed three times in FACS buffer and fixed with 1% paraformaldehyde in PBS. Stained cells were analyzed using FACS Calibur (BD Biosciences, Heidelberg, Germany) using CellQuest Pro software (BD Biosciences).

Cell surface expression of CD46, CD55, CD59 and binding of activated C3 (C3d) was quantitatively determined by QIFIKIT (DAKO, Glostrup, Denmark) where calibration with a series of beads coated with well‐defined quantities of mouse monoclonal IgG molecules allows determining the precise number of antigenic sites present on the tumor cells.

2.6. Complement‐mediated cytotoxicity assay (CDC)

Complement‐mediated cytolysis was analyzed by ⁵¹Cr release assay. Tumor cells (1 × 10⁶) were labeled in 100 μl complete growth medium with 100 μCi ⁵¹Cr (Hartmann Analytik, Braunschweig, Germany) for 2 h at 37 °C. Labeled cells were washed three times, adjusted to 2 × 10⁵ cells/ml in assay medium (growth medium without FCS) and transferred to U‐bottom 96‐well plate. Cells (50 μl/well) were incubated for 30 min at 37 °C with trastuzumab or pertuzumab or a combination of both antibodies at concentrations of 10 μg/ml. NHS (normal human serum, diluted 1:5 in assay medium) or heat‐inactivated serum as control was added and incubated for 60 min at 37 °C. Maximal ⁵¹Cr release was determined by incubation of cells in 1% TritonX‐100 solution (Roche, Mannheim, Germany). For spontaneous release, cells were incubated in medium alone. Cells were then centrifuged for 5 min at 1200 rpm and radioactivity in supernatants was measured in a gamma counter. Specific release (%) was calculated as [(test release − spontaneous release)/(maximum release − spontaneous release)] × 100. All experiments were done in triplicates.

2.7. Analysis of complement‐mediated cell death and caspase activity

Complement‐mediated cell death of tumor cells was quantified by staining with annexin V/propidium iodide using FITC Annexin V Apoptosis Detection Kit I (BD Pharmingen, Heidelberg, Germany). Tumor cells (1 × 10⁵ cells) were incubated in combination of both trastuzumab and pertuzumab at concentrations of each 10 μg/ml in assay buffer (growth medium without FCS, 1% BSA) for 30 min. NHS or heat inactivated serum was added and incubated for 60 min at 37 °C. To terminate complement activation, tumor cells were once washed with ice‐cold EDTA‐solution (20 mM EDTA/assay buffer). Cells were then washed in annexin V binding buffer (10 mM HEPES, pH 7.4, 2.5 mM CaCl₂, 140 mM NaCl) and incubated with annexin V‐FITC (BD Pharmingen) for 15 min in the dark. Cells were washed again and resuspended in binding buffer. A quantity of 5 μg/ml propidium iodide (PI) was added to each sample prior to flow cytometric analysis (FACScan; Becton Dickinson, Heidelberg, Germany). All experiments were performed in duplicate. The obtained results were evaluated with FlowJo software 8.7.1 (Treestar, Ashland, Oregon, USA).

Activation of caspases was quantified using Cell Meter Genetic Caspase Activity Assay Kit (AAT Bioquest, Sunnyvale, CA, USA). The kit utilizes TF2‐VAD‐FMK, a green fluorescent reagent, that binds to activated caspase‐1, ‐3, ‐4, ‐5, ‐6, ‐7, ‐8 and ‐9 in apoptotic cells and can be quantified by flow cytometry. Tumor cells (1 × 10⁵ cells) were incubated with antibodies and complement as described above. After 60 min cells were washed with ice‐cold EDTA‐solution (20 mM EDTA/assay buffer) to terminate complement activation and incubated with TF2‐VAD‐FMK reagent (1:500) (AAT Bioquest) for 60 min. Cells were washed, resuspended in assay buffer and measured by flow cytometry. All experiments were performed in duplicate.

2.8. C3d deposition on tumor cells

The stable C3 split product C3d was used as a surrogate marker for opsonizing C3b and iC3b molecules as previously described (Zell et al., 2007). Briefly, tumor cells were washed and re‐suspended in VBS‐buffer (5 mM sodium barbital (pH 7.4), 0.15 mM CaCl₂, 1 mM MgCl₂, 150 mM NaCl and 0.1% BSA). Cells (1 × 10⁵) were incubated with trastuzumab (10 μg/ml), pertuzumab (10 μg/ml), or a combination of both antibodies (each 10 μg/ml) diluted in VBS‐buffer for 30 min. To avoid terminal MAC formation, C8 depleted human serum (CompTech, Tyler, Texas, USA) was added to the cells and incubated for 30 min at 37 °C and heat inactivated C8 depleted serum as control. Cells were then washed three times with FACS‐buffer and FACS analysis, as described above, was performed to quantify C3d binding (First antibody: mouse monoclonal anti‐C3d; secondary antibody: FITC goat anti‐mouse).

2.9. Preparation of monocyte‐derived macrophages

Isolation and cultivation of human monocytes/macrophages were done as described (Kzhyshkowska et al., 2006) with modifications. The cells were purified from individual buffy coats, which were diluted with Ca²⁺ and Mg²⁺ free PBS at a ratio of 1:1. Diluted buffy coat (25 ml) was layered on top of 12 ml lymphoprep (Axis Shield, Oslo, Norway). After 30 min centrifugation in a swing‐out rotor at 800 × g, peripheral blood mononuclear cells (PBMC) were collected from the lymphoprep interphase and were washed three times with PBS. For isolation of CD14+ cells, 15–20 × 10⁶ PBMC were incubated with 20 μl of anti‐CD14 microbeads (Miltenyi Biotech, Bergisch Gladbach, Germany) in 100 μl MACS buffer (PBS/2 mM EDTA/0.5% FCS). After incubation for 30 min at 4 °C, cells were washed and resuspended in 1.5 ml MACS buffer and loaded onto a LS column (Miltenyi Biotech) placed in a MidiMACS magnet. The column was washed five times with 2 ml MACS buffer. The column was removed from magnetic field and CD14+ cells were recovered from the column by flushing the column five times with 2 ml MACS buffer using a plunger. Cell purity was determined by staining with FITC‐labeled anti‐CD14 antibody (eBioscience, Frankfurt, Germany). CD14+ monocytes with purity of 90–95% or above were used for further experiments. Cells were resuspended in RPMI 1640‐supplemented medium and cultured at a final concentration of 1 × 10⁶ cells/ml. The cell suspension was supplemented with IL‐4 (10 ng/ml), IFN‐gamma (1000 U/ml), IL4+IL10 (10 ng/ml each) or M‐CSF (50 ng/ml) (all from Tebu Bio, Frankfurt, Germany), incubated at 37 °C in the presence of 5% CO₂ for 6 d and subjected to further analysis.

2.10. Complement‐dependent cellular cytotoxicity

Complement‐dependent cellular cytotoxicity of BT474 cells by human in vitro generated monocyte derived macrophages was analyzed by ⁵¹Cr release assay. Tumor cells were labeled with ⁵¹Cr and complement activation with trastuzumab and pertuzumab was performed as described above. To avoid MAC formation, C8 depleted serum (CompTech, Tyler, Texas, USA) was added to the cells and incubated for 30 min at 37 °C (with heat inactivated C8 depleted serum as control). Human monocyte‐derived macrophages (see above) were added to achieve a E:T ratio of 10:1 and cells were incubated at 37 °C in a humidified incubator at 5% CO₂ for 4 h. Cells were centrifuged for 5 min at 1200 rpm and radioactivity in supernatants was measured in a gamma counter. % Specific release was calculated as described above. All experiments were done in triplicates.

2.11. Statistical analysis

Data are presented as means ± standard deviation (SD). Statistical analysis was performed using GraphPad Prism 6.0, first applying ANOVA tests and followed by Dunnett's multiple comparisons test. Significance was accepted when p values were <0.05.

3. Results

3.1. Selection of 2′‐O‐methyl modified specific siRNAs for CD46, CD55 and CD59

To identify potent siRNA molecules with increased stability and efficient RNA interference (RNAi), we designed 2‐5 different siRNA molecules with 2′‐O‐methyl chemical modifications for each regulator that hybridizes at different target mRNA regions and the most efficient siRNA molecules in reducing each regulator mRNA were selected, i.e. anti‐CD46 siRNA (715), anti‐CD55 siRNA (336) and anti‐CD59 siRNA (306) (Figure 1).

Figure 1.

Selection of siRNAs. Schematic representation of the tested modified siRNAs (Atu‐siRNA) targeting to the different regions on CD46 (GenBank: X59410), CD55 (GenBank: M31516) and CD59 (GenBank: BC001506) mRNA. The number indicated in brackets represents the starting nucleotide of the targeted siRNA sequence of mRNA region. siRNAs sequences which showed most potent RNAi activity are denoted by circles. Abbreviations: 5′ UTR, 5′ untranslated region; signal, signal peptide, SCR, short consensus repeat; STP, serine‐threonine‐proline rich region; CTD, C‐terminal domain; UK, unknown significance area; TM, transmembrane region; Cyt, cytoplasmic domain; 3′ UTR, 3′ untranslated region.

3.2. Silencing of mCRPs expression by siRNA

Tumor cell lines of different histological origin, BT474, SK‐BR‐3 (breast cancer), SKOV3 (ovarian cancer) and Calu‐3 (lung cancer) were selected for our study. These cell lines over‐express HER2/neu, and trastuzumab and pertuzumab bind with comparable efficacy (Figure 2a). Flow cytometric analysis revealed varying complement regulators expression on tumor cells (Figure 2b).

Figure 2.

Expression of mCRPs and HER2/neu receptor. (a) Basal expression of HER2/neu in four representative tumor cell lines as assessed by FACS analysis using trastuzumab (blue line) and pertuzumab (red line). Human IgG1 was used as an isotype control (black line). (b) Basal expression of surface bound complement regulatory proteins CD46, CD55 and CD59 on SK‐BR‐3, BT474, SKOV3 and Calu‐3 cells assessed by flow cytometry. Data are represented as mean values ± SD of five independent experiments.

Tumor cells were transfected with siRNAs to CD46, CD55 and CD59 delivered by AtuPLEXes. After 72–96 h, knockdown efficiency of CD46, CD55 and CD59 was assessed by flow cytometry. CD46 expression was inhibited by 93 ± 3% in BT474, by 80 ± 2% in SK‐BR‐3, by 95 ± 1% in SKOV3 and by 90 ± 2% in Calu‐3 cells. The expression of CD55 was down regulated by 85 ± 8% in BT474, by 94 ± 4% in SK‐BR‐3, by 97 ± 2% in SKOV3 and by 91 ± 5% in Calu‐3 cells. CD59 expression was decreased by 85 ± 6% in BT474, by 83 ± 4% in SK‐BR‐3, by 94 ± 1% in SKOV3 and by 85 ± 2% in Calu‐3 cells (Figure 3a). Knockdown of each regulator in all the cell lines was presented in a homogeneous population (Figure 3b).

Figure 3.

siRNA induced knock‐down of mCRPs. (a) siRNAs anti‐CD46, anti‐CD55, anti‐CD59 were either individually or combined transfected into tumor cells using AtuPLEX. Knockdown of target protein expression was analyzed 72–96 h later by flow cytometry. The percentage relative expression of mCRP inhibition was calculated with reference to control non‐silencing siRNA (=100%). Data are given as mean values ± SD of n = 5 independent experiments; p < 0.01 (**), p < 0.001 (***). b) FACS histograms illustrating knock down of CD46, CD55 and CD59 expression on BT474, SK‐BR‐3, SKOV3 and Calu‐3 cells. 72–96 h after transfection with specific siRNA (red line) or non‐silencing control siRNA (blue line) tumor cells were stained with mCRP specific primary antibody, followed by goat anti‐mouse IgG‐FITC. Respective cells were stained with isotype control antibody (pink line).

3.3. Complement dependent cytotoxicity of tumor cells

In all the cell lines tested, increased cell lysis was observed only when tumor cells were incubated in combination with both trastuzumab and pertuzumab in contrast to minor cell lysis when targeted with the individual antibodies (Figure 4).

Figure 4.

Complement‐dependent cytotoxicity on tumor cells upon mCRPs knockdown. 72–96 h after siRNA transfection, tumor cells were incubated with trastuzumab (red bars) or pertuzumab (blue bars) or in combination of both (green bars). CDC was analyzed by 51Cr release assay. Data are presented as mean values ± SD of n = 5 independent experiments; p < 0.05 (*), p < 0.01 (**), p < 0.001 (***).

Inhibition of CD46 expression in BT474 cells resulted in 25 ± 3% increased cell lysis, but had no significant effect on other tumor cells. Downregulation of CD55 expression alone induced a slightly enhanced effect on cell lysis in SK‐BR‐3 cells but had no significant effect on other tumor cells. In contrast, inhibition of CD59 expression alone sensitized all tumor cells to complement attack, in BT474 cells by 16 ± 5%, in SK‐BR‐3 cells by 12 ± 5% and by 27 ± 12% cell lysis in SKOV3 cells. The most pronounced CDC effect was observed upon combined downregulation of all three regulators with an overall cell lysis of 48 ± 11% in BT474 cells, of 46 ± 6% in SK‐BR‐3 cells, 74 ± 13% cell lysis in SKOV3 cells and of 30 ± 6% in Calu‐3 cells. In none of the cell lines, significant cell lysis was achieved when targeted with trastuzumab or pertuzumab alone, even upon silencing the complement regulators (Figure 4). Heat inactivation of NHS completely abolished cell lysis (data not shown).

3.4. Complement‐mediated cell death and caspase activity

To analyze complement‐mediated cell death, tumor cells were treated with trastuzumab and pertuzumab, exposed to NHS and subsequently washed and stained with annexin V and propidium iodide. 71 ± 5% SKOV3 cells, 24 ± 6% BT474 cells, 22 ± 8% SK‐BR‐3 cells and 34 ± 6% Calu‐3 were stained positive for both annexin V and propidium iodide when expression of all the mCRP was inhibited. In cells transfected with control non‐silencing siRNA, only a minor proportion of cells stained positive for annexin V and propidium iodide (Figure 5a). In BT474 and SK‐BR‐3 cells, we also observed substantial proportion of cells that were annexin V positive and PI negative as compared to control non‐silencing transfected cells (Figure 5b). This effect was completely abolished when inactivated NHS was used instead of NHS (Figures 5a and b).

Figure 5.

Complement‐mediated cell death and caspase activity. Tumor cells were transfected with control siRNA or anti‐CD46, CD55 and CD59 combined siRNA or left untreated. 72–96 h after siRNA transfection, tumor cells were incubated with trastuzumab and pertuzumab antibodies, followed by the addition of NHS or heat in‐activated NHS. (a) After 60 min, cells were then stained with Annexin V‐FITC and PI, and analyzed by flow cytometry. The percentage of cells staining positive for annexin V and PI are shown. Data are presented as mean values ± SD of n = 3 independent experiments; p < 0.01 (**). (b) Fluorescence intensities were measured by flow cytometry using FL1 (Annexin V‐FITC) and FL2 (PI) channels. The values shown in the lower left, lower right and upper right quadrants of each panel represent the percentage of viable, early apoptosis and late apoptosis cells respectively. Shown is one representative of three independent experiments. (c) After 60 min, cells were incubated with TF2‐VAD‐FMK fluorescent reagent and analyzed for caspase activity by flow cytometry. The difference in mean fluorescence intensity (MFI) between the cells treated with or without trastuzumab and pertuzumab is presented. Data are presented as mean values ± SD of n = 3 independent experiments; p < 0.05 (*), p < 0.01 (**), p < 0.001 (***).

Figure 5.

(continued).

As the characterization of cell death as apoptotic or necrotic based on Annexin V/PI staining is still unclear, we then performed pan caspase activity analysis as an indicator of apoptosis. To examine the activation of caspases in complement‐mediated cell death, tumor cells were incubated with the fluorescent reagent TF2‐VAD‐FMK that binds to activated caspases (caspase‐1, ‐3, ‐4, ‐5, ‐6, ‐7, ‐8 and ‐9) in apoptotic cells. Increased fluorescence intensity was observed when tumor cells were treated with both antibodies and complement following mCRPs knockdown. Addition of inactivated NHS abolished caspase activation and fluorescence intensity was similar to that in control non‐silencing siRNA transfected cells (Figure 5c). Highest caspase activity was observed in SKOV3 cells, correlating with CDC and complement‐mediated cell death. This indicates that complement‐mediated cell death is associated with caspase‐dependent apoptosis.

Figure 5.

(continued).

3.5. Opsonisation of tumor cells

We then evaluated the efficacy of trastuzumab or pertuzumab alone or in combination to induce C3 tumor cell opsonisation. Enhanced C3d deposition was observed when tumor cells were incubated with both trastuzumab and pertuzumab antibodies in contrast to only minor C3d deposition by the individual antibodies (Figure 6a). C3 deposition was further augmented upon mCRP silencing (Figure 6b). In the presence of both trastuzumab and pertuzumab, inhibition of CD46 further increased C3d deposition when compared to control non‐silencing siRNA on BT474 cells by 48% (MFI 255 ± 51 vs. MFI 172 ± 55), on SK‐BR‐3 by 20%, (MFI 280 ± 22 vs. MFI 233 ± 34), on SKOV3 cells by 48% (MFI 244 ± 47 vs. MFI 165 ± 57), and on Calu‐3 cells by 78% (MFI 380 ± 115 vs. MFI 213 ± 60). Downregulation of CD55 resulted in significantly higher C3d deposition on BT474 cells by 82% (MFI 314 ± 72 vs. MFI 172 ± 55), on SK‐BR‐3 cells by 53% (MFI 357 ± 53 vs. MFI 233 ± 34), on SKOV3 cells by 84% (MFI 303 ± 46 vs. MFI 165 ± 57) and on Calu‐3 cells by 209% (MFI 446 ± 132 vs. MFI 213 ± 60). As expected, knock down of CD59 had no significant impact on C3d deposition on BT474 and SK‐BR‐3 cells. Surprisingly, opsonization of SKOV3 cells and Calu‐3 cells was increased by 25% (MFI 206 ± 39 vs. MFI 165 ± 57) and by 37% (MFI 291 ± 60 vs. 213 ± 60) respectively. Optimal C3d deposition was observed after the combined knockdown of all three regulators, with an increase by 253% (MFI 436 ± 87 vs. 172 ± 55) in BT474 cells, by 158% (MFI 369 ± 56 vs. MFI 233 ± 34) in SK‐BR‐3 cells, in SKOV3 cells by 223% (MFI 369 ± 100 vs. MFI 165 ± 57) and by 238% (MFI 507 ± 131 vs. 213 ± 60) in Calu‐3 cells (Figure 6b).

Figure 6.

Opsonization of tumor cells. (a) Tumor cells were incubated with trastuzumab or pertuzumab alone or with a combination of both antibodies. Deposition of activated C3 (here as C3d) on tumor cells was quantified by flow cytometry. The difference in mean fluorescence intensity (MFI) between C3d specific antibody and isotype control antibody is presented. (b) C3d deposition on tumor cells following siRNA‐mediated downregulation of CD46, CD55 and CD59. Tumor cells were incubated with both trastuzumab and pertuzumab. The difference in mean fluorescence intensity (MFI) between C3d specific antibody and isotype control antibody is presented. Data are presented as mean values ± SD of n = 5 independent experiments; p < 0.05 (*), p < 0.01 (**).

3.6. Complement‐dependent macrophage mediated cytotoxicity

BT474 cells were incubated in the presence of both trastuzumab and pertuzumab with C8 depleted human serum. The opsonized tumor cells were then exposed to macrophages that were differentiated in the presence of IFN‐γ, M‐CSF, or IL‐4, or a combination of IL‐4 + IL‐10. Knockdown of all three complement regulators led to an increased cell‐mediated cytotoxicity by 19 ± 10% by IFN‐γ stimulated macrophages, by 22 ± 16% by IL‐4 stimulated macrophages, by 18 ± 5% by IL‐4 + IL‐10 stimulated macrophages, and by 32 ± 17% by M‐CSF stimulated macrophages, but was negligible in the absence of macrophages or if cells were pretreated with heat in‐activated serum (Figure 7).

Figure 7.

Complement dependent cellular cytotoxicity of BT474 cells. CDCC was analyzed by 51Cr release assay. BT474 cells were transfected with control siRNA or combined siRNA anti‐CD46, CD55 and CD59 or left untreated. 72–96 h after siRNA transfection, tumor cells were incubated with trastuzumab and pertuzumab, followed by the addition of C8 depleted serum or heat in‐activated C8 depleted serum. After 30 min, macrophages were added to tumor cells and incubated for 4 h. Radioactivity in supernatants was measured in a gamma counter. Data are presented as mean values ± SD of n = 5 experiments with monocyte derived macrophages of 5 different donors for each analysis; p < 0.05 (*), p < 0.01 (**).

4. Discussion

In recent years, monoclonal antibodies have been designed for successful cancer immunotherapy (Weiner et al., 2010). However, due to the intrinsic resistance of cancer cells to complement‐mediated lysis, the biological activity of these antibodies has been restricted to antibody‐dependent cellular cytotoxicity (ADCC), programmed cell death (apoptosis), blocking growth factor receptor signaling and interference with angiogenesis pathways. Overexpression of complement regulatory proteins has been described in many tumors and significantly restricts the therapeutic potential of tumor‐targeted antibodies (Gancz and Fishelson, 2009). CD59 over‐expression has been shown to limit rituximab therapy in patients with B‐cell malignancies (Treon et al., 2001) and neutralization of CD55 and CD59 reversed resistance to rituximab‐mediated complement lysis (Macor et al., 2007). Recently it was shown that inhibition of CD55 and CD59 slightly sensitizes uterine serous carcinoma cells to trastuzumab‐mediated CDC and ADCC (Bellone et al., 2012). We have previously demonstrated that neutralization of cell surface complement regulators either by blocking antibodies or by down‐regulation of regulator expression with anti‐sense phosphorothioate oligonucleotides (S‐ODNs) or small interfering RNAs (siRNAs) sensitizes tumor cells to antibody‐induced complement‐dependent cytolysis (Geis et al., 2010; Jurianz et al., 1999; Zell et al., 2007).

We here demonstrate for the first time that an efficient complement‐mediated tumor cell killing by combined treatment with trastuzumab and pertuzumab is only possible if all the three membrane complement regulators CD46, CD55 and CD59 are neutralized. Targeting the tumor cells with trastuzumab or pertuzumab alone had only very minor effects on CDC. One possible explanation could be that only by the combination of both antibodies, which bind to distinct epitopes of the HER2 extracellular domain, a sufficient number of cell‐bound antibodies in critical distance could be reached to efficiently bind and activate C1q required to initiate the complement cascade reaction (Borsos and Circolo, 1983). Similar results were described where a mixture of two chimeric monoclonal antibodies to distant epitopes of the folate receptor was required to activate complement. Neutralization of CD46 and CD59 by blocking antibodies markedly enhanced CDC of ovarian tumor cells (Macor et al., 2006). Also, combinations of EGFR antibodies triggered complement activation and induced more CDC than individual EGFR antibodies. The antibody isotype and cognate non‐overlapping binding epitopes appear to be critical for CDC induction by EGFR antibody combinations (Dechant et al., 2008; Klausz et al., 2011). This is consistent with our observation that anti‐HER2 antibodies, trastuzumab and pertuzumab are of human IgG1 isotype and both bind to different epitopes on HER2 receptor, which triggers CDC in combination.

In this study we designed new siRNA molecules to CD46, CD55 and CD59 and introduced alternating 2′‐O‐methyl sugar modifications on both strands, which are significantly more resistant to plasma‐derived nucleases than unmodified siRNAs (Czauderna et al., 2003). Equally important, the non‐specific immune stimulatory activity of unmodified synthetic siRNAs was abrogated by incorporation of 2′‐O‐methyl modifications in siRNA duplex (Judge et al., 2006). Inhibition of individual complement regulators exerted only a minor effect, indicating that complement resistance can only be abolished if more than one regulator is neutralized. To some extent, inhibition of CD59 alone was sufficient to sensitize BT474, SK‐BR‐3 and SKOV3 cells to CDC. In contrast, Calu‐3 cells were less sensitized to CDC even upon silencing, which suggests that the remaining surface expression of CD59 still conferred resistance to CDC. Addition of anti‐CD59 blocking antibody further augmented tumor cell lysis (data not shown).

Cationic lipoplexes are widely used for the efficient delivery of nucleic acids to mammalian cells (Felgner et al., 1987). We used AtuPLEX (Silence Therapeutics) as a novel non‐viral siRNA carrier, which also mediates the endosomal release of siRNAs after endocytotic uptake of the siRNA complexes (Santel et al., 2006b). AtuPLEX enables a broad functional delivery of siRNA molecules to targeted diseased tissues and cells, with an increased bioavailability and intracellular uptake (Santel et al., 2006a). Atu027 is a novel AtuFECT01‐based RNAi therapeutic drug containing 2′‐O‐methyl modified PKN3 siRNA, which targets vascular endothelium and inhibits cancer progression (Aleku et al., 2008) and metastasis formation (Santel et al., 2010).

AtuPLEX assisted transfection of tumor cells with individual siRNA inhibited mCRP expression by 85–95% dependent on the cell line. A limited availability of components of the RNAi machinery or internal competition among the three different siRNAs for the incorporation into RISC could explain the reduced ability of anti‐CD46 siRNA to mediate mRNA degradation in combined transfection with all anti‐mCRP siRNAs. These results are consistent with the previously published study in which competition was observed between modified and unmodified siRNAs (Koller et al., 2006) and the efficiency of combinatorial siRNAs was reduced by selective incorporation into RISC (Castanotto et al., 2007).

Complement is involved in the induction of apoptosis (Nauta et al., 2002), also inducing DNA fragmentation and chromatin condensation characteristic of apoptosis (Cragg et al., 2000). Again, only upon silencing of all the three regulators, complement activation by combined trastuzumab and pertuzumab treatment led to tumor cell apoptosis reflected by an increased pan‐caspase activity. This effect was clearly complement dependent as it was completely abolished upon serum inactivation.

Complement exerts its biological activity not only by its direct cytotoxic action but also by opsonizing target structures. C3 fragments (iC3b/C3b) on cells interact with CR3 (CD11b/CD18) on immune cells to mediate complement‐dependent cellular cytotoxicity (Vetvicka et al., 1997). Incubation of tumor cells with trastuzumab or pertuzumab alone only led to minute C3d deposition whereas the combination of both antibodies efficiently induced C3 opsonization. As expected, inhibition of complement regulators, especially of CD46 or CD55 further augmented C3d deposition. Killing of C3 opsonized BT474 cells correlated with the CR3 levels on IFN‐γ and M‐CSF (high) or IL‐4 and IL‐4 + IL‐10 (low) macrophages (data not shown). Macrophage‐mediated tumor cell killing was only observed in combination of both antibodies, required to sufficiently activate complement.

We here present a novel strategy to significantly improve the anti‐tumor activity of trastuzumab and pertuzumab on HER2 positive tumor cells. Only in combination, trastuzumab and pertuzumab are able to employ complement with subsequent tumor cell killing, provided that complement regulator proteins are neutralized to circumvent complement resistance. The crucial aspect of this strategy is the delivery of mCRP directed siRNA to specifically target tumor cells. To avoid deleterious side effects of systemic administration, delivery of mCRP‐specific siRNA to tumor cells requires the conjugation of targeting molecules such as receptor ligands or antibodies to lipid carriers, such as AtuPLEX. The feasibility, efficacy and safety of this approach remains to be demonstrated in appropriate animal models.

Disclosure of potential conflicts of interest

The authors disclose no potential conflicts of interest.

Mamidi Srinivas, Cinci Marc, Hasmann Max, Fehring Volker, Kirschfink Michael, (2013), Lipoplex mediated silencing of membrane regulators (CD46, CD55 and CD59) enhances complement-dependent anti-tumor activity of trastuzumab and pertuzumab, Molecular Oncology, 7, doi: 10.1016/j.molonc.2013.02.011.