Egg phosphatidylcholine (EPC) was purchased from Lipoid (Ludwigshafen, Germany), cholesterol was purchased from Calbiochem (Darmstadt, Germany). 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(7-nitro-2-1,3-benzoxadiazol-4-yl) (ammonium salt) (NBD-DOPE) and 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (ammonium salt) (rhodamine-DPPE) and all other lipids were purchased from Avanti Polar Lipids (Alabaster, AL, USA). DiI was purchased from Sigma (Taufkirchen, Germany). HRP-conjugated anti-His-tag antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA) and FITC-conjugated anti-His antibody from Dianova (Hamburg, Germany). The PE-labeled monoclonal anti-EGFR antibody (EGFR R-1) was purchased from Santa Cruz Biotechnology. The BW 431/26 (mouse anti-CEA) antibody was provided by Behringwerke (Marburg, Germany). The polyclonal anti-moIgG-FITC IgG was purchased from Sigma-Aldrich, St. Louis, USA. The human colon carcinoma cell line LS174T, the human epidermoid carcinoma cell line A431 and the human lung adenocarcinoma epithelial cell line A549 were cultured in RPMI1640 containing 5% FCS, 2 mM L-glutamine and 100 units/mL penicillin G and 100 µg/mL streptomycin. The human breast cancer cell lines SKBR3 and MCF7 and the human colon adenocarcinoma cell line LoVo were cultured in RPMI1640 containing 10% FCS, 2 mM L-glutamine and 100 units/mL penicillin G and 100 µg/mL streptomycin. Cells were cultured at 37 °C in a humidified 5% CO₂ incubator.

An anti-EGFR scFv' was generated by cloning a humanized version of scFv C225 (hu225) into vector pABC4 [19]. An anti-CEA scFv' was described previously [19]. Antibody fragments (anti-EGFR scFv' hu225, anti-CEA scFv' MFE23) were purified by immobilized metal ion affinity chromatography (IMAC) as described elsewhere [35]. Protein concentration was determined by measuring the absorbance at 280 nm. Purified scFv' were analyzed by SDS-PAGE under reducing and non-reducing conditions and stained with Coomassie brilliant blue G250 or immunoblotted with an HRP-conjugated anti-His-tag antibody.

Purified scFv' (100–150 µg) was diluted in PBS to a total volume of 1 mL and sterile filtered (0.2 µm) into a quartz cuvette. Dynamic laser light scattering intensity was measured with a Zetasizer Nano ZS (Malvern, Herrenberg, Germany) while the temperature was increased in 1 °C intervals from 30 to 70 °C with 2 min equilibration for each temperature step. The melting point was defined as the temperature at which the light scattering intensity dramatically increased.

Immunoliposomes were generated by postinsertion of scFv'-conjugated micelles into preformed PEG-liposomes [20]. Liposomes composed of EPC:Chol:mPEG₂₀₀₀-DSPE at a molar ratio of 6.75:3:0.25 were prepared by the film hydration-extrusion method as described previously [19]. All liposomes further contained 0.3 mol% fluorescent dye (DiI). Mal-PEG₂₀₀₀-DSPE micelles were generated by removal of chloroform under a stream of nitrogen at RT. The dried lipid was dissolved in double distilled water to a final concentration of 10 mg/mL and incubated for 5 min at 65 °C in a water bath to form micelles. Purified scFv were reduced in 8 mM tris(2-carboxyethyl)phosphine (TCEP) for 2 h at room temperature followed by removal of TCEP by dialysis against degassed 10 mM Na₂HPO₄/NaH₂PO₄ buffer, 0.2 mM EDTA, 30 mM NaCl (pH 6.7) overnight at 4 °C. Micellar lipid and reduced scFv' were mixed at a molar ratio of 5:1. Coupling reaction was performed at room temperature for 30 min. The reaction was quenched with 1 mM L-cysteine, 0.02 mM EDTA, pH 5.5. The scFv'-coupled micelles were inserted into preformed PEGylated liposomes by incubation at 45 °C for 2 h (anti-CEA scFv') or at 55 °C for 30 min (anti-EGFR scFv'). Dual targeted immunoliposomes were prepared by step-wise post-insertion. Firstly, anti-EGFR scFv'-coupled micelles were inserted at 55 °C for 30 min afterwards anti-CEA scFv'-coupled micelles were added to the dispersion and again incubated at 45 °C for 2 h in a water bath. Unbound scFv molecules were removed by gel-filtration using a Sepharose CL4B column (Amersham, Braunschweig, Germany). Liposome size was measured using a ZetaSizer Nano ZS (Malvern). To monitor the synchronized post-insertion of two different micelle species the Mal-PEG₂₀₀₀-DSPE-micelles were stained either with NBD-DOPE or with rhodamine-DPPE as indicated. The fluorescently-labeled phospholipids were added to the Mal-PEG₂₀₀₀-DSPE in chloroform before the organic solvent was evaporated.

For analyzing the binding of monoclonal antibodies, cells were resuspended to a concentration of 2.5 × 10⁶ cells/mL in PBA (PBS, 2% FCS, 0.02% sodium azide) containing the diluted primary antibody (anti-EGFR: 1:100, anti-CEA: 1:100). After incubation at 4 °C for 1 h in the dark, the plate was washed three times with 150 μL PBA. To detect unlabeled primary antibodies 100 μL 1:500 diluted anti-mouse IgG-FITC was applied to the respective wells and incubated at 4 °C for 45 min in the dark. Cells incubated with a fluorescently labeled antibody were resuspended in 100 μL PBA. The cells were washed twice in 150 μL PBA, resuspended in 500 μL PBA and transferred into FACS tubes. A total of 10,000 cells were analyzed using a Cytomics FC500 (Beckmann Coulter, Krefeld, Germany). Analysis of the obtained data was performed using FlowJo 7.6.1 and Microsoft Office Excel 2011. Binding of purified scFv' molecules was determined by incubating detached cells with antibody molecules in PBA buffer for 1 h at 4 °C. Cells were then washed three times with PBA and incubated with FITC-conjugated anti-His-tag antibody diluted 1:200. Binding of liposomes was analyzed by incubating cells (2.5 × 10⁵) with DiI-labeled immunoliposomes (1–1,000 nmol lipid) per 100 µL PBA for 1 h at 4 °C. After washing cells three times with PBA (4 °C) cells were resuspended in 300 µL PBA buffer and analyzed by flow cytometry. Data was evaluated with WinMDI, version 2.9. The relative mean fluorescence intensity (MFI) plotted in the binding experiment analysis was calculated according to the following formula: relative MFI = MFI_sample − (MFI_nt − MFI_cells)/MFI_cells.

Binding of the immunoliposomes to CEA and EGFR was analyzed by a sandwich ELISA. CEA was coated onto a microtiter plate at a concentration of 3 µg/mL. After blocking remaining binding sites with MPBS (PBS, 2% skimmed milk powder), liposomes were added at varying concentrations and incubated for 1 h. After washing with PBS, an EGFR-Fc fusion protein (10 µg/mL) was added and incubated for 1 h. After washing, bound EGFR-Fc was detected with an HRP-conjugated anti-human Fc antibody (Sigma-Aldrich) and subsequent addition of TMB substrate.

The flow cytometer was adjusted for counting small liposomes at a size range between 100–150 nm by modifying the voltage of the front and side scatter detectors. Furthermore, the overlapping fluorescence signals in fluorescence detector one (FL1) and detector two (FL2) of single fluorescently labeled liposomes (NBD-DOPE or rhodamine-DPPE) were compensated for the detection of double stained vesicles (NBD-DOPE and rhodamine-DPPE). All liposomes were measured at a phospholipid concentration of approximately 1 mM analyzing 10,000 vesicles.

A panel of human tumor cell lines was analyzed by flow cytometry for the expression of EGFR and CEA (Figure 1). The two human colon carcinoma cell lines LS174T and LoVo showed strong staining for EGFR and CEA. Three other cell lines, A549, MCF-7 and SKBR3 showed medium to weak expression of the two antigens. The human squamous carcinoma cell line A431 exhibited strong expression of EGFR and was negative for CEA. Thus, these cell lines express the two antigens to varying extent and should therefore be suitable for the analysis of dual targeting of immunoliposomes.

For the generation of dual-targeted immunoliposomes we applied the post-insertion of Mal-PEG₂₀₀₀-DSPE-micelles conjugated with two scFv' molecules of different specificity. Principally, this can be performed with micelles prepared by the simultaneous conjugation of both different ligands. Alternatively, the preparation and insertion of individual ligand-coupled micelles can be applied, with the insertion performed simultaneously or step-wise. In order to monitor the behavior of Mal-PEG₂₀₀₀-DSPE-micelles during the post-insertion into preformed liposomes (plain liposomes), we analyzed micelles containing either the fluorescent phospholipid NBD-DOPE (green fluorescence) or rhodamine-DPPE (red fluorescence), respectively. During insertion, these lipids will be transferred into the liposomes, thus acting as surrogate for the ligand-coupled lipids. Mal-PEG₂₀₀₀-DSPE-micelles were prepared containing between 1 to 30 mol% of the fluorescent phospholipid NBD-DOPE (Figure 2a). The size (zeta average) of Mal-PEG₂₀₀₀-DSPE micelles containing 1, 3, 10, and 30 mol% NBD-DOPE, respectively, remained stable at around 15 to 18 nm, in accordance with values described for PEG₂₀₀₀-DSPE micelles [36]. The size of the plain liposomes was 138 nm with a PDI of 0.11. After post-insertion, no micellar structures were detected by dynamic light scattering. The size of the liposomes increased slightly to 143–159 nm (Figure 2a). The poly-dispersity index (PDI) of all formulations after post-insertion was below 0.2. An increase in fluorescence intensity of the liposomes was observed after post-insertion of 1 mol% Mal-PEG₂₀₀₀-DSPE-micelles that directly correlated with the amount of dye contained within the micelles (Figure 2b,c). No remaining (free) Mal-PEG₂₀₀₀-DSPE-micelles were detected via dynamic light scattering implying a complete insertion of micellar lipids into the outer layer of the preformed liposomes.

Next, we analyzed the effects of inserting two differently labeled micelles (containing 10 mol% NBD-DOPE or rhodamine-DPPE, respectively) into preformed liposomes. The post-insertion of 0.03, 0.1, 0.3, 1 mol% Mal-PEG₂₀₀₀-DSPE-micelles containing either NBD-DOPE or rhodamine-DPPE into liposomes caused a concentration-dependent shift in fluorescence intensity (Figure 2d). Thus, insertion of NBD-DOPE-labeled micelles caused an increase in green fluorescence and insertion of rhodamine-DPPE-labeled micelles caused an increase in red fluorescence, respectively. The simultaneous post-insertion of both differently labeled Mal-PEG₂₀₀₀-DSPE-micelles into liposomes led to a similar concentration-dependent shift of green and red fluorescence intensities. No liposomes exhibiting only green or red fluorescence could be detected even at very low concentration of micellar lipid used for post-insertion (e.g., 0.03 mol% corresponding to approximately 40 to 50 micellar lipids per liposome). According to this finding, the post-insertion of differently labeled micelles results in the insertion of lipids from both kinds of micelles into one liposome, thus should be applicable for the generation of dual-targeted immunoliposomes even at low micelle to liposome ratios.

For the generation of immunoliposomes, we produced two scFv' fragments containing a C-terminal cysteine for side-directed coupling. ScFv' hu225 is directed against human EGFR and scFv' MFE23 is directed against human CEA. Both antibody fragments were produced in E. coli and purified from the periplasm with yields in the range of 0.2–0.5 mg/L. In SDS-PAGE, the molecules migrated with an apparent molecular mass of 30–35 kDa under reducing conditions. Under non-reducing conditions, an additional band of approximately 55–60 kDa was visible, indicating the formation of disulfide-linked dimers (Figure 3a). Thermal stability of the antibody fragments was determined by dynamic light scattering, which revealed a melting point of approximately 62 °C for scFv' hu225 and of 48 °C for scFv' MFE23 (Figure 3b). Both antibody fragments were capable of binding to EGFR- and CEA-expressing tumor cell lines (LS174T, A549) in a concentration-dependent manner with EC₅₀ values of 0.3–1 nM for scFv' hu225 and 3–10 nM for scFv' MFE23 (Figure 3).

In order to determine the amount of scFv-coupled micellar lipids required for a strong immunoliposomal binding to target cells, we generated immunoliposomes by inserting between 0.01 mol% and 1 mol% of scFv-MalPEG₂₀₀₀-DSPE-conjugates. Coupling efficiencies to Mal-PEG₂₀₀₀-DSPE were approximately 95% for scFv' hu225 and 50% for scFv' MFE23 as determined by SDS-PAGE (Figure 4a). Hu225-ILs were generated by post-insertion at 55 °C for 30 min and MFE23-ILs were generated performing the post-insertion step at 45 °C for approximately 2 h. Both types of immunoliposomes (hu225-IL, MFE23-IL) showed a concentration-dependent binding to EGFR- and CEA-positive tumor cell lines analyzed by flow cytometry (Figure 4b). For hu225-IL, strongest signals were observed using 0.1 to 0.3 mol% inserted lipid. MFE23-ILs showed strong binding with 0.03 to 0.1 mol% inserted lipid.

Because the two scFv' molecules showed different thermal stability, we established a two-step post-insertion protocol for the generation of dual-targeted immunoliposomes, which also allowed to use different amounts of the two antibody-coupled micellar lipids for insertion (Figure 5).

In a first step, 0.3 mol% scFv' hu225-coupled lipids were inserted into preformed liposomes at 55 °C for 30 min. In a second step, 0.03 mol% scFv' MFE23-coupled lipids were then inserted into these immunoliposomes at 45 °C for approximately 2 h. Bispecificity of the dual-targeted ILs was confirmed by ELISA. CEA was immobilized onto microtiter plates and subsequently incubated with liposomes followed by soluble EGFR-Fc fusion protein, which was then detected with an anti-Fc antibody (Figure 6). Thus, only liposomes containing anti-CEA and anti-EGFR scFvs in their lipid bilayer are capable of binding both antigens. Accordingly, anti-EGFR-ILs (hu225-IL), anti-CEA-ILs (MFE23-IL) as well as non-targeted liposomes (nt) gave no signals in ELISA, while dt-ILs caused a concentration-dependent increase in ELISA signal.

The dt-ILs as well as the monospecific ILs were then tested for binding to a panel of different tumor cell lines (Figure 7). Binding of the monospecific ILs correlated with expression of EGFR and CEA on these cell lines. Thus, MFE23-ILs showed strong binding to LS174T, LoVo and A549 cells, while weak binding was seen with MCF-7 and no binding with A431 and SKBR3 cells. Hu225-ILs showed strong binding to all cell lines except MCF-7 for which only a weak binding was observed. Interestingly, the dt-ILs combined the binding activities of the monospecific ILs. Thus, strong binding was observed with all cell lines tested. Signals obtained with the dt-ILs were also concentration-dependent and always similar or slightly better than the signals observed for either hu225-ILs or MFE23-ILs alone.