Polypeptide-polynucleotide-complex and its use in targeted effector moiety delivery

Abstract

Herein is reported a polypeptide-polynucleotide-complex as therapeutic agent and its use as tool for the targeted delivery of an effector moiety. The polynucleotide part of the complex is essentially resistant to proteolytic and enzymatic degradation in vivo. Additionally the polypeptide part specifically binds to a compound or structure such as a tissue or organ, a process or a disease. Thus, one aspect as reported herein is a polypeptide-polynucleotide-complex comprising a) a polypeptide specifically binding to a target and conjugated to a first member of a binding pair, b) a polynucleotide linker conjugated at its first terminus to the second member of the binding pair, and c) an effector moiety conjugated to a polynucleotide that is complementary to at least a part of the polynucleotide linker.

Claims

1. A method of producing a complex comprising the components a) a first polypeptide that specifically binds to a first target and that is conjugated to a first member of a first binding pair, b) a second polypeptide that specifically binds to a second target and that is conjugated to a first member of a second binding pair, and c) a polynucleotide linker conjugated to the second member of the first binding pair and conjugated to the second member of the second binding pair, wherein the polynucleotide linker comprises ss-L-DNA, comprising the steps of: i) synthesizing the first polypeptide specifically binding to the first target which is conjugated to the first member of the first binding pair, and synthesizing the second polypeptide specifically binding to the second target which is conjugated to the first member of a second binding pair, respectively, ii) synthesizing the polynucleotide linker conjugated at its first terminus to the second member of the first binding pair and conjugated at its second terminus to the second member of the second binding pair, and iii) forming the complex by hybridizing the synthesized components, wherein the first and second member of the first binding pair comprise the nucleic acid sequences of SEQ ID NO: 5 and SEQ ID NO: 8 or wherein the first and second member of the second binding pair comprise the nucleic acid sequences of SEQ ID NO: 6 and SEQ ID NO: 7.

2. The method according to claim 1, wherein the complex further comprises an effector moiety conjugated to a polynucleotide that is complementary to at least a part of the polynucleotide linker ss-L-DNA.

3. The method of claim 2, wherein the effector moiety is selected from the group consisting of a binding moiety, a labeling moiety and a biologically active moiety.

4. The method according to claim 1, wherein the first polypeptide is a monovalent antibody or monovalent antibody fragment.

5. The method of claim 1, wherein the first and second polypeptides bind to the same target and to non-overlapping epitopes on the same target.

6. The method of claim 5, wherein the polynucleotide linker has an optimal length for synergistic binding of the first and second polypeptides to the non-overlapping epitopes on the same target.

7. The method of claim 1, wherein the members of the first and second binding pairs are ss-LDNA.

8. The method of claim 1, wherein the members of the first and second binding pairs are ss-LDNA of 10 to 50 nucleotides in length.

9. The method of claim 1, wherein the polynucleotide linker has a length of at least 70 nucleotides.

10. The method of claim 1, wherein the complex is a non-covalent complex.

11. The method of claim 1, wherein the first polypeptide is a FAB′ fragment of the anti-HER2 antibody 2C4 comprising an HVR-H1 comprising the amino acid sequence of SEQ ID NO: 35, an HVR-H2 comprising the amino acid sequence of SEQ ID NO: 36, an HVR-H3 comprising the amino acid sequence of SEQ ID NO: 37, an HVR-L1 comprising the amino acid sequence of SEQ ID NO: 39, an HVR-L2 comprising the amino acid sequence of SEQ ID NO: 40, and an HVR-L3 comprising the amino acid sequence of SEQ ID NO: 41; the second polypeptide is a FAB′ fragment of the anti-HER2 antibody 4D5 comprising an HVR-H1 comprising the amino acid sequence of SEQ ID NO: 27, an HVR-H2 comprising the amino acid sequence of SEQ ID NO: 28, an HVR-H3 comprising the amino acid sequence of SEQ ID NO: 29, an HVR-L1 comprising the amino acid sequence of SEQ ID NO: 31, a HVR-L2 comprising the amino acid sequence of SEQ ID NO: 32, and an HVR-L3 comprising the amino acid sequence of SEQ ID NO: 33; the members of the first and the second binding pairs are hybridizing nucleic acids; and the ss-L-DNA-linker comprises 60 to 100 L-DNA nucleotides.

Description

FIGURES

(1) FIG. 1 Scheme of the BIAcore assay setup. ss-L-DNA-bi linkers were presented on a BIAcore SA sensor. Flow cell 1 served as a control (not shown).

(2) FIG. 2 BIAcore sensorgrams for the HER2-ECD interaction with ss-D-DNA labeled FAB fragments.

(3) FIG. 3 BIAcore sensorgrams showing concentration dependent interaction measurements of the complex as reported herein comprising a 35-mer (=35 nucleotide length) as linker polynucleotide with HER2-ECD.

(4) FIG. 4 BIAcore sensorgrams showing concentration dependent interaction measurements of the complex as reported herein comprising a 75-mer (=75 nucleotide length) as linker polynucleotide with HER2-ECD.

(5) FIG. 5 BIAcore sensorgrams showing concentration dependent interaction measurements of the complex as reported herein comprising a 95-mer (=95 nucleotide length) as linker polynucleotide with HER2-ECD.

(6) FIG. 6 Scheme of the BIAcore assay setup: polyclonal goat anti human IgG-Fc gamma antibody was presented on a BIAcore SA sensor. Flow cell 1 served as a control (not shown).

(7) FIG. 7 The BIAcore sensorgram shows an overlay plot of interaction signals upon 50 nM injections of anti-HER2 antibody 2C4-FAB′-ss-L-DNA (2C4), anti-HER2 antibody 4D5-FAB′-ss-L-DNA (4D5) and fully established complex (2C4-75mer-4D5) connected by a 75mer ss-L-DNA linker.

(8) FIG. 8 BIAcore sensorgram showing an overlay plot of concentration-dependent measurements of the fully established 75-mer complex as analyte in solution interacting with the surface presented huFc chimera HER2 ECD

(9) FIG. 9 Plot of the response levels of FIG. 8 versus the analyte concentration of the fully established complex.

(10) FIGS. 10A-10D Analytical gel filtration experiments assessing efficiency of the anti-pIGF1-R complex assembly. FIGS. 10A, 10B, and 10C show the elution profile of the individual complex components (fluorescein-ss-FAB′ 1.4.168, Cy5-ssFab′ 8.1.2 and Linker DNA (T=0)). FIG. 10D shows the elution profile after the 3 components needed to form the bivalent binding agent had been mixed in a 1:1:1 molar ratio. The thicker (bottom) curve represents absorbance measured at 280 nm indicating the presence of the ss-FAB′ proteins or the linker DNA, respectively. The thinner top curve in FIGS. 10B and 10D (absorbance at 495 nm) indicates the presence of Cy5 and the thinner top curve in a) FIG. 10A and the middle curve in FIG. 10D (absorbance at 635 nm) indicates the presence of fluorescein. Comparison of the elution volumes of the single complex components (VE.sub.ssFab′ 1.4.168˜15 ml; VE.sub.ssFab′ 8.1.2˜15 ml; VE.sub.linker˜16 ml) with the elution volume of the reaction mix (VE.sub.mix˜12 ml) demonstrates that the complex assembly reaction was successful (rate of yield: ˜90%). The major 280 nm peak that represents the eluted complex nicely overlaps with the major peaks in the 495 nm and 695 nm channel, proving the presence of both ss-FAB′ 8.1.2 and ssFab′1.4.168 in the peak representing the bivalent binding agent.

(11) FIG. 11 Scheme of the BIAcore experiment: schematically and exemplarily, two binding molecules in solution are shown. The T=0-Dig, bivalent binding agent and the T=40-Dig, bivalent binding agent. Both these bivalent binding agents only differ in their linker-length (no additional T versus 40 additional Ts, separating the two hybridizing nucleic acid sequences). Furthermore, ss-FAB′ fragments 8.1.2 and 1.4.168 were used.

(12) FIG. 12 BIAcore sensorgram with overlay plot of three kinetics showing the interaction of 100 nM bivalent binding agent (consisting of ss-FAB 8.1.2 and ss-FAB 1.4.168 hybridized on the T40-Dig ss-DNA-Linker) with the immobilized peptide pIGF-1R compared to the binding characteristics of 100 nM ss-FAB 1.4.168 or 100 nM ss-FAB 8.1.2 to the same peptide. Highest binding performance is only obtained with the complex construct, clearly showing, that the cooperative binding effect of the Complex increases affinity versus the target peptide pIGF-1R.

(13) FIG. 13 BIAcore sensorgram with overlay plot of three kinetics showing the interactions of the bivalent binding agent consisting of ss-FAB 8.1.2 and ss-FAB 1.4.168 hybridized on the T40-Dig ss-DNA-Linker with immobilized peptides pIGF-1R (phosphorylated IGF-1R), IGF-1R or IR (phosphorylated insulin receptor). Highest binding performance is obtained with the pIGF-1R peptide, clearly showing, that the cooperative binding effect of the Complex increases specificity versus the target peptide pIGF-1R as compared to e.g. the phosphorylated insulin receptor peptide (IR).

(14) FIG. 14 BIAcore sensorgram with overlay plot of two kinetics showing the interactions of 100 nM bivalent binding agent consisting of ss-FAB′ 8.1.2 and ss-FAB′ 1.4.168 hybridized on the T=40-Dig ss-DNA-Linker and a mixture of 100 nM ss-FAB′ 8.1.2 and 100 nM ss-FAB′ 1.4.168 without linker DNA. Best binding performance is only obtained with the bivalent binding agent, whereas the mixture of the ss-FAB's without linker doesn't show an observable cooperative binding effect, despite the fact that the total concentration of these ss-FAB's had been at 200 nM.

(15) FIG. 15 Schematic drawing of a BIAcore sandwich assay. This assay has been used to investigate the epitope accessibility for both antibodies on the phosphorylated IGF-1R peptide. <MIgGFcy>R presents a rabbit anti-mouse antibody used to capture the murine antibody M-1.4.168. M-1.4.168 then is used to capture the pIGF-1R peptide. M-8.1.2 finally forms the sandwich consisting of M-1.4.168, the peptide and M-8.1.2.

(16) FIG. 16 BIAcore sensorgram showing the binding signal (thick line) of the secondary antibody 8.1.2. to the pIGF-1R peptide after this was captured by antibody 1.4.168 on the BIAcore chip. The other signals (thin lines) are control signals: given are the lines for a homologous control with 500 nM 1.4.168, 500 nM target unrelated antibody <CKMM>M-33-IgG; and 500 nM target unrelated control antibody <TSH>M-1.20-IgG, respectively. No binding event could be detected in any of these controls.

(17) FIG. 17 Schematic drawing of the BIAcore assay, presenting the complex s on the sensor surface. On Flow Cell 1 (=FC1) (not shown) amino-PEO-biotin was captured. On FC2, FC3 and FC4 bivalent binding agents with increasing linker length were immobilized. Analyte 1: IGF-1R-peptide containing the M-1.4.168 ss-FAB epitope (thin line)—the M-8.1.2 ss-FAB phospho-epitope is not present, because this peptide is not phosphorylated; analyte 2: pIGF-1R peptide containing the M-8.1.2 ss-FAB phospho-epitope (P) and the M-1.4.168 ss-FAB epitope (thin line). Analyte 3: pINR peptide, containing the cross reacting M-8.1.2 ss-FAB phospho-epitope, but not the epitope for M-1.4.168.

(18) FIG. 18 Kinetic data of the Complex experiment. T40-bi complex with ss-FAB 8.1.2 and ss-FAB 1.4.168 shows a 1300-fold lower off-rate (KD=2.79E-05/s) versus pIGF-1R when compared to pINR (KD=3.70E-02).

(19) FIG. 19 BIAcore sensorgram, showing concentration dependent measurement of the T40-bi complex vs. the pIGF-1R peptide (the phosphorylated IGF-1R peptide).

(20) FIG. 20 BIAcore sensorgram, showing concentration dependent measurement of the T40-bi complex vs. the IGF-1R peptide (the non-phosphorylated IGF-1R peptide).

(21) FIG. 21 BIAcore sensorgram, showing concentration dependent measurement of the T40-bi complex vs. the pINR peptide (the phosphorylated insulin receptor peptide).

(22) FIG. 22 Staining of tumor cells with Cy5 labeled Xolair® and Herceptin®.

(23) FIGS. 23A-23B NIRF imaging of KPL-4 cells.

(24) FIG. 24 Ex vivo staining of KPL-4 xenografts.

(25) FIG. 25 Size exclusion profile of freshly prepared 4D5-95mer-2C4 complex. Upper signal: 260 nm signal, lower signal: 280 nm signal. No aggregates can be detected between start at 0.0 min and the elution peak at 5.64 min.

(26) FIG. 26 Size exclusion profile of the 4D5-95mer-2C4 complex after a freezing and thawing cycle. Upper signal: 260 nm signal, lower signal: 280 nm signal. No aggregates can be detected between start at 0.0 min and the elution peak at 5.71 min.

EXAMPLE 1

(27) Formation of FAB-ss-DNA-Conjugates

(28) Two monoclonal antibodies binding to human cardiac Troponin T at different, non-overlapping epitopes, epitope a and epitope b, respectively, were used. Both these antibodies are used in the current Roche Elecsys™ Troponin T assay, wherein Troponin T is detected in a sandwich immunoassay format.

(29) Purification of the monoclonal antibodies from culture supernatant was carried out using state of the art methods of protein chemistry.

(30) The purified monoclonal antibodies are protease digested with either pre-activated papain (anti-epitope a MAb) or pepsin (anti-epitope b MAb) yielding F(ab′)2 fragments that are subsequently reduced to FAB′-fragments with a low concentration of cysteamine at 37° C. The reaction is stopped by separating the cysteamine on a Sephadex G-25 column (GE Healthcare) from the polypeptide-containing part of the sample.

(31) The FAB′-fragments are conjugated with the below described activated ss-DNAa and ss-DNAb oligonucleotides.

(32) a) Anti-Troponin T (Epitope A) Antibody FAB-ss-DNA-Conjugate A

(33) For preparation of the anti-Troponin T <epitope a> antibody FAB-ss-DNAa-conjugate A a derivative of SEQ ID NO: 05 is used, i.e. 5′-AGT CTA TTA ATG CTT CTG C(=SEQ ID NO:5)-XXX—Y-Z-3′, wherein X=propylene-phosphate introduced via Phosphoramidite C3 (3-(4,4′-Dimethoxytrityloxy)propyl-1-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research), wherein Y=3″-Amino-Modifier C6 introduced via 3′-Amino Modifier TFA Amino C-6 lcaa CPG (ChemGenes) and wherein Z=4[N-maleinimidomethyl]cyclohexane-1-carboxy introduced via Sulfosuccinimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate (ThermoFischer).

(34) b) Anti-Troponin T (Epitope B) Antibody FAB-ss-DNA-Conjugate B

(35) For the preparation of the anti-Troponin T <epitope b> antibody FAB-ss-DNAb-conjugate B a derivative of SEQ ID NO: 06 is used, i.e. 5′-Y—Z-XXX-AGT TCT ATC GTC GTC CA-3′, wherein X=propylene-phosphate introduced via Phosphoramidite C3 (3-(4,4′-Dimethoxytrityloxy)propyl-1-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research), wherein Y=5′-Amino-Modifier C6 introduced via (6-(4-Monomethoxytritylamino)hexyl-(2-cyanoethyl)-(N,N-diisopropyl)-phosphoramidite (Glen Research), and wherein Z=4[N-maleinimidomethyl]cyclohexane-1-carboxy introduced via Sulfosuccinimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate (ThermoFischer).

(36) The oligonucleotides of SEQ ID NO: 05 or 06, respectively, have been synthesized by state of the art oligonucleotide synthesis methods. The introduction of the maleinimido group was done via reaction of the amino group of Y with the succinimidyl group of Z which was incorporated during the solid phase oligonucleotide synthesis process.

(37) The single-stranded DNA constructs shown above bear a thiol-reactive maleimido group that reacts with a cysteine of the FAB′ hinge region generated by the cysteamine treatment. In order to obtain a high percentage of single-labeled FAB′-fragments the relative molar ratio of ss-DNA to FAB′-fragment is kept low. Purification of single-labeled FAB′-fragments (ss-DNA:FAB′=1:1) occurs via anion exchange chromatography (column: MonoQ, GE Healthcare). Verification of efficient labeling and purification is achieved by analytical gel filtration chromatography and SDS-PAGE.

EXAMPLE 2

(38) Formation of Biotinylated Linker Molecules

(39) The oligonucleotides used in the ss-DNA linkers L1, L2 and L3, respectively, have been synthesized by state of the art oligonucleotide synthesis methods and employing a biotinylated phosphoramidite reagent for biotinylation.

(40) Linker 1 (=L.sub.1), a biotinylated ss-DNA linker 1 with no spacer has the following composition: 5′-GCA GAA GCA TTA ATA GAC T (Biotin-dT)-GG ACG ACG ATA GAA CT-3′. It comprises ss-DNA oligonucleotides of SEQ ID NO: 7 and 8, respectively, and was biotinylated by using Biotin-dT (5′-Dimethoxytrityloxy-5-[N-((4-t-butylbenzoyl)-biotinyl)-aminohexyl)-3-acrylimido]-2′-deoxyUridine-3′-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research).

(41) Linker 2 (=L.sub.2), a biotinylated ss-DNA linker 2 with a 10mer spacer has the following composition: 5′-GCA GAA GCA TTA ATA GAC T T5-(Biotin-dT)-T5 GG ACG ACG ATA GAA CT-3′. It comprises ss-DNA oligonucleotides of SEQ ID NO: 7 and 8, respectively, twice oligonucleotide stretches of five thymidines each and was biotinylated by using Biotin-dT (5′-Dimethoxytrityloxy-5-[N-((4-t-butylbenzoyl)-biotinyl)-aminohexyl)-3-acrylimido]-2′-deoxyUridine-3′-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research).

(42) Linker 3 (=L.sub.3), a biotinylated ss-DNA linker 3 with a 30mer spacer has the following composition: 5′-GCA GAA GCA TTA ATA GAC T T15-(Biotin-dT)-T15 GG ACG ACG ATA GAA CT-3′. It comprises ss-DNA oligonucleotides of SEQ ID NO: 7 and 8, respectively, twice oligonucleotide stretches of fifteen thymidines each and was biotinylated by using Biotin-dT (5′-Dimethoxytrityloxy-5-[N-((4-t-butylbenzoyl)-biotinyl)-aminohexyl)-3-acrylimido]-2′-deoxyUridine-3′-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research).

EXAMPLE 3

(43) Epitopes for Monovalent Troponin T Binders a and b, Respectively

(44) Synthetic peptides have been construed that individually only have a moderate affinity to the corresponding FAB′-fragment derived from the anti-Troponin T antibodies a and b, respectively.

(45) a) The Epitope “A” for Antibody a is Comprised in:

(46) TABLE-US-00007 SEQ ID NO: 9 = ERAEQQRIRAEREKEUUSLKDRIEKRRRAERAE amide, wherein U represents β-Alanine.

(47) b) The Epitope “B” for Antibody b is Comprised in:

(48) TABLE-US-00008 SEQ ID NO: 10 = SLKDRIERRRAERAEOOERAEQQRIRAEREKE amide, wherein O represents Amino-trioxa-octanoic- acid

(49) As the skilled artisan will appreciate it is possible to combine these two epitope-containing peptides two ways and both variants have been designed and prepared by linear combining the epitopes “A” and “B”. The sequences of both variants, the linear sequences of epitopes “A”-“B” (=TnT 1) and “B”-“A” (=TnT 2), respectively have been prepared by state of the art peptide synthesis methods.

(50) The sequences for epitopes “A” and “B”, respectively, had been modified compared to the original epitopes on the human cardiac Troponin T sequence (P45379/UniProtKB) in order to reduce the binding affinity for each of the FABs thereto. Under these circumstances the dynamics of the effect of hetero-bivalent binding is better visible, e.g. by analyzing binding affinity with the BIAcore Technology.

EXAMPLE 4

(51) Biomolecular Interaction Analysis

(52) For this experiment a BIAcore 3000 instrument (GE Healthcare) was used with a BIAcore SA sensor mounted into the system at T=25° C. Preconditioning was done at 100 μl/min with 3×1 min injection of 1 M NaCl in 50 mM NaOH and 1 min 10 mM HCl.

(53) HBS-ET (10 mM HEPES pH 7.4, 150 mM NaC, 1 mM EDTA, 0.05% Tween® 20 was used as system buffer. The sample buffer was identical to the system buffer.

(54) The BIAcore 3000 System was driven under the control software V1.1.1. Flow cell 1 was saturated with 7 RU D-biotin. On flow cell 2, 1063 RU biotinylated ss-DNA linker L1 was immobilized. On flow cell 3, 879 RU biotinylated ss-DNA linker L2 was immobilized. On flow cell 4, 674 RU biotinylated ss-DNA linker L3 was captured.

(55) Thereafter, FAB fragment DNA conjugate A was injected at 600 nM. FAB fragment DNA conjugate B was injected into the system at 900 nM. The conjugates were injected for 3 min at a flow rate of 2 μl/min. The conjugates were consecutively injected to monitor the respective saturation signal of each FAB fragment DNA conjugate on its respective linker. FAB combinations were driven with a single FAB fragment DNA conjugate A, a single FAB fragment DNA conjugate B and both FAB fragment DNA conjugates A and B present on the respective linker. Stable baselines were generated after the linkers have been saturated by the FAB fragment DNA conjugates, which was a prerequisite for further kinetic measurements.

(56) The peptidic analytes TnT1 and TnT2 were injected as analytes in solution into the system in order to interact with the surface presented FAB fragments.

(57) TnT1 was injected at 500 nM, TnT2 was injected at 900 nM analyte concentration. Both peptides were injected at 50 μl/min for 4 min association time. The dissociation was monitored for 5 min. Regeneration was done by a 1 min injection at 50 μl/min of 50 mM NaOH over all flow cells.

(58) Kinetic data was determined using the BIAevaluation software (V.4.1). The dissociation rate KD (1/s) of the TnT1 and TnT2 peptides from the respective surface presented FAB fragment combinations was determined according to a linear Langmuir 1:1 fitting model. The complex halftime in min were calculated according to the solution of the first order kinetic equation: ln(2)/(60*kD).

(59) Results:

(60) The experimental data given in Tables 5 and 6, respectively demonstrate an increase in complex stability between analyte (TnT1 or TnT2), respectively, and the various heterobivalent FAB-FAB conjugates A-B as compared to the monovalent dsDNA FAB A or B conjugate, respectively. This effect is seen in each Table in line 1 compared to lines 2 and 3.

(61) TABLE-US-00009 TABLE 5 Analysis data using TnT1 with linkers of various length FAB fragment FAB fragment kD t½ diss DNA conjugate A DNA conjugate B (1/s) (min) a) Linker L1 x x 6.6E−03 1.7 x — 3.2E−02 0.4 — x 1.2E−01 0.1 b) Linker L2 x x 4.85E−03  2.4 x — 2.8E−02 0.4 — x 1.3E−01 0.1 c) Linker L3 x x 2.0E−03 5.7 x — 1.57E−02  0.7 — x 1.56E−02  0.7

(62) TABLE-US-00010 TABLE 6 Analysis data using TnT2 with linkers of various length FAB fragment FAB fragment kD t½ diss DNA conjugate A DNA conjugate B (/1/s) (min) a) Linker L1 x x 1.4E−02 0.8 x — 4.3E−02 0.3 — x 1.4E−01 0.1 b) Linker L2 x x 4.9E−03 2.3 x — 3.5E−02 0.3 — x 1.3E−01 0.1 c) Linker L3 x x 8.0E−03 1.5 x — 4.9E−02 0.2 — x 3.2E−01 0.04

(63) The avidity effect is further dependent on the length of the linker. In the sub-tables shown under Table 1 the 30mer linker L.sub.3 shows the lowest dissociation rate or highest complex stability, in sub-tables shown under Table 2 the 10mer L.sub.2 linker exhibits the lowest dissociation rate or highest complex stability. These data taken together demonstrate that the flexibility in linker length as inherent to the approach given in the present invention is of great utility and advantage.

EXAMPLE 5

(64) Formation of FAB′-ss-DNA-Conjugates

(65) Two monoclonal antibodies binding to human HER2 (ErbB2 or p185.sup.neu) at different, non-overlapping epitopes A and B were used. The first antibody is anti-HER2 antibody 4D5 (huMAb4D5-8, rhuMab HER2, trastuzumab or HERCEPTIN®; see U.S. Pat. No. 5,821,337 incorporated herein by reference in its entirety). The second antibody is anti-HER2 antibody 2C4 (Pertuzumab).

(66) Purification of the monoclonal antibodies from culture supernatant can be carried out using state of the art methods of protein chemistry.

(67) The purified monoclonal antibodies are protease digested with either pre-activated papain or pepsin yielding F(ab′)2 fragments. These are subsequently reduced to FAB′-fragments with a low concentration of cysteamine at 37° C. The reaction is stopped by separating the cysteamine on a Sephadex G-25 column (GE Healthcare) from the polypeptide-containing part of the sample.

(68) The obtained FAB′-fragments are conjugated with the activated ss-DNA polynucleotides.

(69) a) Anti-HER2 Antibody 4D5 FAB′-ss-DNA-Conjugate

(70) For preparation of the anti-HER2 antibody 4D5 FAB′-ss-DNA-conjugate a derivative of SED ID NO: 05 is used, i.e. 5′-AGT CTA TTA ATG CTT CTG C(=SEQ ID NO: 05)-XXX-Y-Z-3′, wherein X=propylene-phosphate introduced via phosphoramidite C3 (3-(4,4′-dimethoxytrityloxy)propyl-1-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research), wherein Y=5′-amino-modifier C6 introduced via (6-(4-monomethoxytritylamino)hexyl-(2-cyanoethyl)-(N,N-diisopropyl)-phosphoramidite (Glen Research), and wherein Z=4[N-maleinimidomethyl]cyclohexane-1-carboxy introduced via Sulfosuccinimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate (ThermoFischer).

(71) b) Anti-HER2 Antibody 2C4 FAB′-ss-DNA-Conjugate

(72) For the preparation of the anti-HER2 antibody 2C4 FAB′-ss-DNA-conjugate B a derivative of SEQ ID NO: 06 is used, i.e. 5′-Y—Z-XXX-AGT TCT ATC GTC GTC CA-3′, wherein X=propylene-phosphate introduced via Phosphoramidite C3 (3-(4,4′-Dimethoxytrityloxy)propyl-1-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research), wherein Y=5′-Amino-Modifier C6 introduced via (6-(4-Monomethoxytritylamino)hexyl-(2-cyanoethyl)-(N,N-diisopropyl)-phosphoramidite (Glen Research), and wherein Z=4[N-maleinimidomethyl]cyclohexane-1-carboxy introduced via Sulfosuccinimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate (ThermoFischer).

(73) The polynucleotides of SEQ ID NO: 05 or SEQ ID NO: 06, respectively, have been synthesized by state of the art polynucleotide synthesis methods. The introduction of the maleinimido group was done via reaction of the amino group of Y which was incorporated during the solid phase polynucleotide synthesis process with the Sulfosuccinimidyl 4-[N-maleimidomethyl] cyclohexane-1-carboxylate (ThermoFischer).

(74) The single-stranded DNA constructs bear a thiol-reactive maleimido group that reacts with a cysteine of the FAB′ hinge region generated by the cysteamine treatment. In order to obtain a high percentage of single-labeled FAB′-fragments the relative molar ratio of ss-DNA to FAB′-fragment is kept low. Purification of single-labeled FAB′-fragments (ss-DNA:FAB′=1:1) occurs via anion exchange chromatography (column: MonoQ, GE Healthcare). Verification of efficient labeling and purification is achieved by analytical gel filtration chromatography and SDS-PAGE.

EXAMPLE 6

(75) Biomolecular Interaction Analysis

(76) For this experiment a BIAcore T100 instrument (GE Healthcare) was used with a BIAcore SA sensor mounted into the system at T=25° C. Preconditioning occurred at 100 μl/min with 3×1 min injection of 1 M NaCl in 50 mM NaOH, pH 8.0 followed by a 1 min injection of 10 mM HCl. The system buffer was HBS-EP (10 mM HEPES pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.05% P 20). The sample buffer was the system buffer supplemented with 1 mg/ml CMD (carboxymethyldextrane).

(77) Biotinylated ss-L-DNA linkers were captured on the SA surface in the respective flow cells. Flow cell 1 was saturated with amino-PEO-Biotin (PIERCE).

(78) 40 RU of the biotinylated 35mer oligonucleotide linker were captured on flow cell 2. 55 RU of the biotinylated 75mer oligonucleotide linker were captured on flow cell 3. 60 RU of biotinylated 95mer oligonucleotide linker were captured on flow cell 4.

(79) 250 nM anti-HER2 antibody 4D5-FAB′-ss-L-DNA was injected into the system for 3 min. 300 nM anti-HER2 antibody 2C4-FAB′-ss-L-DNA was injected into the system at 2 μl/min for 5 min. The DNA-labeled FAB fragments were injected alone or in combination.

(80) As a control only 250 nM anti-HER2 antibody 4D5-FAB′-ss-D-DNA and 300 nM anti-HER2 antibody 2C4-FAB′-ss-D-DNA was injected into the system. As a further control, buffer was injected instead of the DNA-labeled FAB fragments. After hybridization of the ss-L-DNA-labeled FAB fragments on the respective ss-L-DNA bi-linkers, the analyte in solution hHER2-ECD was injected at different concentration series from 24 nM, 8 nM, 3 nM, 1 nM, 0.3 nM, 0 nM into the system for 3.5 min association phase at 100 μl/min. The dissociation phase was monitored at 100 μl/min for 15 min. The system was regenerated by a 30 sec injection at 20 μl/min of 100 mM glycine buffer (Glycine pH 11, 150 mM NaCl), followed by a second 1 min injection of water at 30 μl/min.

(81) The signals were measured as analyte concentration-dependent, time resolved sensorgrams. The data was evaluated using the BIAcore BIAevaluation software 4.1. As a fitting model a standard Langmuir binary binding model was used.

(82) Results:

(83) No HER2-ECD interaction could be observed when ss-D-DNA labeled FAB fragments were injected into the system, because the ss-D-DNA-labeled FAB fragments didn't hybridize with spiegelmeric ss-L-DNA linkers presented on the sensor surface (FIG. 2).

(84) Table 7: Kinetic results of the complexation experiment. Linker: Surface presented biotinylated ss-L-DNA polynucleotide linker, Oligo_35mer-Bi, Oligo_75mer-Bi and Oligo_95mer-Bi differing in linker length. ss-L-DNA-FAB: 2C4-ss-L-DNA: anti-HER2 antibody 2C4-FAB′-ss-L-DNA labeled with 19mer-Fluorescein. 4D5-ss-L-DNA: anti-HER2 antibody 4D5-FAB′-ss-L-DNA labeled with 17mer-Fluorescein. 4D5-+2C4-ss-L-DNA: surface presented combination of both fragments. LRU: mass in response units, which is hybridized on the sensor surface. Antigen: An 87 kDa HER2-ECD was used as analyte in solution. ka: association rate in (1/Ms). kD: dissociation rate in (1/s). t1/2 diss: antigen complex halftime calculated in hours according to the solution ln(2)/kD*3600 of a first order kinetic equation. kD: affinity in molar. kD: affinity calculated in picomolar. Rmax: Maximum analyte response signal at saturation in response units (RU). MR: Molar Ratio, indicating the stoichiometry of the interaction. Chi2, U-value: quality indicator of the measurements.

(85) TABLE-US-00011 TABLE 7 k.sub.a k.sub.d t/.sub.2-diss K.sub.D K.sub.D R.sub.max Chi.sup.2 Linker ss-L-DNA-Fab LRU Antigen 1/Ms 1/s hours M pM RU MR RU.sup.2 Oligo_35mer_Bi 4D5- + 2C4-ss-L-DNA 84 Her2-ECD 5.9E+05 6.7E−05 3 1.1E−10 100 59 0.9 0.2 Oligo_35mer_Bi 4D5-ss-L-DNA 16 Her2-ECD 4.0E+05 3.4E−05 6 8.5E−11 100 29 1.2 0.1 Oligo_35mer_Bi 2C4-ss-L-DNA 31 Her2-ECD 3.3E+05 3.6E−05 5 1.1E−10 100 26 0.6 0.03 Oligo_75mer_Bi 4D5- + 2C4-ss-L-DNA 87 Her2-ECD 5.1E+05 4.6E−06 4164 9.1E−14 0.1 65 1.0 0.1 Oligo_75mer_Bi 4D5-ss-L-DNA 16 Her2-ECD 2.9E+05 6.1E−05 3 2.1E−10 200 31 1.3 0.04 Oligo_75mer_Bi 2C4-ss-L-DNA 29 Her2-ECD 3.8E+05 6.3E−05 3 1.6E−10 200 32 0.7 0.03 Oligo_95mer_Bi 4D5- + 2C4-ss-L-DNA 76 Her2-ECD 5.0E+05 4.9E−08 3942 9.9E−14 0.1 58 1.0 0.1 Oligo_95mer_Bi 4D5-ss-L-DNA 14 Her2-ECD 1.0E+05 9.5E−05 2 3.1E−10 300 28 1.3 0.03 Oligo_95mer_Bi 2C4-ss-L-DNA 28 Her2-ECD 3.8E+05 6.8E−05 3 1.8E−10 300 27 0.6 0.03

(86) The BIAcore sensorgrams show concentration dependent measurements of the 35-mer complex HER2-ECD interaction (FIG. 3). This linker is consisting of solely the hybridization motives sequences of the DNA labels. The kinetic data indicates that the fully established complex shows no improvement of the kinetic performance. This is due to the insufficient linker length and lacking flexibility of the 35-mer.

(87) The BIAcore sensorgrams showing concentration dependent measurements of the 75-mer complex HER2-ECD interaction (FIG. 4). The 75-mer linker carries poly-T to increase the linker length compared to the 35-mer linker. The kinetic data indicates that the fully established complex shows a dramatic improvement of its kinetic performance. This is due to an optimal linker length and flexibility of the 75-mer.

(88) The BIAcore sensorgrams showing concentration dependent measurements of the 95-mer complex HER2-ECD interaction (FIG. 5). The 95-mer linker carries poly-T to increase the linker length compared to the 35-mer linker. The kinetic data indicates that the fully established complex shows a dramatic improvement of its kinetic performance. This is due to increased linker length and flexibility of the 95-mer.

(89) The BIAcore assay setup comprised the following (see also FIG. 1): ss-L-DNA-bi linkers were presented on a BIAcore SA sensor. Flow cell 1 served as a control. As analyte in solution Her2-ECD was used. Anti-HER2 antibody 2C4-FAB′-ss-L-DNA and anti-HER2 antibody 2C4-FAB′-ss-L-DNA were hybridized to the surface presented linkers.

(90) Here is shown, for the first time, a fully functional cooperative binding event between Herceptin-FAB and Pertuzumab-FAB linked together via a highly flexible ss-L-DNA linker. The data in Table 3 provides evidence for the presence of a cooperative binding event. Despite the Rmax values of the fully established complex s are roughly double the signal height of the singly FAB-armed constructs, the Molar Ratio values are exactly 1 (MR=1). This is a clear evidence for the presence of a simultaneous, cooperative binding event of both FAB fragments. The complex counts as a single molecule with a 1:1 Langmuir binding stoichiometry. Despite having 2 independently binding HER2 interfaces no inter molecule binding between one complex and two HER2 domains can be detected.

(91) The avidity constants for synergizing pairs of monoclonal antibodies or for a chemically cross-linked bispecific F(ab′)2 is generally only up to 15 times greater than the affinity constants for the individual monoclonal antibodies, which is significantly less than the theoretical avidity expected for ideal combination between the reactants (Cheong, H. S., et al., Biochem. Biophys. Res. Commun. 173 (1990) 795-800). Without being bound by this theory one reason for this might be that the individual epitope/paratope interactions involved in a synergistic binding (resulting in a high avidity) must be orientated in a particular way relative to each other for optimal synergy.

(92) Furthermore, the data presented in Table 7 provides evidence, that the short 35-mer linker, which consists just from the ss-L-DNA hybridization motives doesn't show enough flexibility or/and linker length to produce the cooperative binding effect. The 35-mer linker is a rigid, double helix L-DNA construct. The hybridization generates a double L-DNA helix, which is shorter and less flexible than the ss-L-DNA sequence. The helix shows reduced degrees of freedom and can be seen as a rigid linker construct. Table 7 shows, that the 35-mer linker isn't able to generate a cooperative binding event.

(93) Extending the linker length by a highly flexible poly-T ss-L-DNA to form a 75-mer and a 95-mer, respectively, provides for an increase in affinity and especially in antigen complex stability kD (1/s).

(94) The chi2 values indicate a high quality of the measurements. All measurements show extremely small errors. The data can be fitted to a Langmuir 1:1 fitting model residuals deviate only +/−1 RU, small chi2 values and only 10 iterative calculations were necessary for obtaining the data.

(95) A cooperative binding effect works according to the physical law, in that the free binding energies ΔG1 and ΔG2 summarize. The affinities multiply: KDcoop=KD1×KD2. Furthermore, the dissociation rates also multiply: KD coop=kd1×KD 2. This is exactly observable in the 75-mer and 95-mer linker experiment. This results in very long complex half-lives of 4146 hours (173 days) and 3942 hours (164 days), respectively. The affinities are in the range of 100 fmol/l. It is obvious, that a cooperative binding event occurs.

(96) The association rates of all fully established complex s are faster, when compared to the singly hybridized constructs. Despite showing a higher molecular weight the association rate increases.

(97) Here we could show, that trastuzumab and Pertuzumab linked together in a complex as reported herein simultaneously binds to the HER-2 extracellular domain (ECD). Both FAB fragments bind to genuine epitopes on the HER2-ECD (PDB 1 S78 and PDB 1N82). Additionally both FAB fragments strongly differ in their binding angles. By using the optimal 75-mer (30 nm) ss-L-DNA linker length and its beneficial flexibility and length properties a cooperative binding event could be shown.

(98) The signals were measured as analyte concentration-dependent, time resolved sensorgrams. The data was evaluated using the BIAcore BIAevaluation software 4.1. As a fitting model a standard Langmuir binary binding model was used.

EXAMPLE 7

(99) Additional Biomolecular Interaction Analysis

(100) A BIAcore 3000 instrument was mounted with a CM-5 sensor chip. The sensor was preconditioned as recommended by the manufacturer (GE healthcare, Uppsala, Sweden). The system buffer was (10 mM HEPES pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.05% Tween® 20). The system buffer was also used as the sample buffer. The system was operated at 25° C. under the control software 4.1.

(101) 30 μg/ml polyclonal goat anti human IgG-Fc gamma antibody (Jackson Laboratories, USA) in 10 mM acetate buffer pH 4.5 were immobilized by standard NHS/EDC chemistry at 13,952 RU on flow cell 1 and 15,047 RU on flow cell 2. The system was regenerated at 20 μl/min using a 20 sec. pulse of a 10 mM glycine pH 1.5 buffer, a 1 min pulse of 10 mM glycine pH 1.7 buffer, and a 30 sec. pulse of 10 mM glycine pH 1.5 buffer. On flow cell 1, 5 nM huIgG (Bayer Healthcare) were injected for 1 min at 10 μl/min as a reference.

(102) On flow cell 2, 10 nM human HER2 extracellular receptor FC chimera (hHER2-ECDpresSFc) were injected for 1 min at 10 μl/min. Typically 100 response units of the prebuilt homodimeric hHER2-ECDpresSFc were captured via the human FC portion on flow cell 2 by a goat anti human IgG-Fc gamma antibody. Typically 130 response units of huIgG were captured via the human FC portion on flow cell 1.

(103) The signal on flow cell 2 was referenced versus flow cell 1.

(104) The ss-L-DNA labeled FAB fragments anti-HER2 antibody 4D5-FAB′-ss-L-DNA and anti-HER2 antibody 2C4-FAB′-ss-L-DNA were hybridized with the 75mer ss-L-DNA linker by a 1:1:1 molar stoichiometry. The fully established complex 2C4-75mer-4D5 was injected for three minutes at 50 nM into the system. As a control, the single FAB fragments were injected at 50 nM into the system.

(105) Immediately after injection end 250 nM streptavidin or system buffer was injected into the system for 3 min at 10 μl/min. Since the 75mer linker contains a single biotin moiety in the center of its sequence, the SA should work as a probe to recognize the biotin within the linker, but not the presence of the FAB fragments.

(106) In another experiment the fully established 4D5-75mer-2C4 complex was injected into the system at different concentration steps 0 nM, 0.6 nM, 1.9 nM, 2x 5.6 nM, 16.7 nM, 50 nM at 10 μl/min for 3 min. The concentration dependent response levels of the hHER2-ECDpresSFc analyte were monitored. The response levels were plotted over the concentration steps of the hHER2-ECDpresSFc. The data was visualized using the software Origin 7. The data was fitted using the Hill equation y=V.sub.max*x.sup.n/(k.sup.n+x.sup.n) as provided by the Origin 7 software.

(107) The BIAcore assay setup comprised the following (see also FIG. 6): A polyclonal goat anti human IgG-Fc gamma antibody was immobilized on the BIAcore CM5 sensor and serves as a capture system for the huFc chimera HER2 ECD. Anti-HER2 antibody 2C4-FAB′-ss-L-DNA (2C4 FAB), anti-HER2 antibody 4D5-FAB′-ss-L-DNA (4D5 FAB) and fully established complexes were injected, followed by the injection of streptavidin (SA). The aim of the experiment is to demonstrate the presence and accessibility of the biotin moiety within the 75mer ss-L-DNA linker.

(108) Results of the experiment are depicted in FIG. 7. The BIAcore sensorgram shows an overlay plot of interaction signals upon 50 nM injections of anti-HER2 antibody 2C4-FAB′-ss-L-DNA (2C4), anti-HER2 antibody 4D5-FAB′-ss-L-DNA (4D5) and fully established complex (2C4-75mer-4D5) connected by a 75mer ss-L-DNA linker. The overlay plot shows that due to its higher mass of 137 kDa the fully established complex binder (2C4-75mer-4D5+ buffer) generates a higher signal response level, when compared to the FAB fragment injections (2C4+ buffer, 4D5+ buffer). The FAB fragments have a calculated molecular weight of 57 kDa, each. Immediately after injection end at 420 sec, 250 nM streptavidin or system buffer was injected. The double headed arrow marks the 14 RU signal shift (ARU) induced by the 250 nM streptavidin injection (2C4-75mer-4D5+SA) compared to the buffer injection (2C4-75mer-4D5+ buffer). The FAB fragments show no signal shift upon SA injection and remain at the buffer signal level ((2C4+SA), (2C4+ buffer), (4D5+SA), (4D5+ buffer)). Streptavidin is the effector moiety. It shows the accessibility of the ss-L-DNA linker.

(109) BIAcore sensorgram showing an overlay plot of concentration-dependent measurements of the fully established 75-mer complex as analyte in solution interacting with the surface presented huFc chimera HER2 ECD is shown in FIG. 8. The black lines represent the 1:1 Langmuir fit on the data. Kinetic data, association rate ka=1.25*10.sup.5 1/Ms, dissociation rate KD=3.39*10.sup.−5 1/s, affinity constant 0.3 nM.

(110) The response levels of FIG. 8 were plotted versus the analyte concentration of the fully established complex (FIG. 9). The data was fitted according to the hill equation and the hill factor was determined (Origin 7). Equation: y=V.sub.max*x.sup.n/(k.sup.n+x.sup.n), Chi2/DoF=0.6653, R2=0.99973; n=1.00201+/−0.06143.

(111) In Table 8 the kinetic data from the BIAcore assay format as depicted in FIG. 6 is shown. The cooperative binding effect can be produced with the complex in solution. The Molar Ratios show, that exactly a single complex recognizes a single HER2-ECD chimera. Kinetic data, association rate ka 1/Ms, dissociation rate kD 1/s, affinity constant KD (M) and in (nM), maximum binding response signal (Rmax), amount of captured huFc Chim Her2ECD Ligand (RU), Complex halftime according to Langmuir t1/2 diss. Molar Ratio MR, indicating the stoichiometry of the binding events. Error chi2. 4D5-2C4-75mer is the fully established complex. 4D5-75mer and 2C4-75mer are the FAB fragments, but hybridized to the ss-L-DNA 75mer linker.

(112) TABLE-US-00012 TABLE 8 Ligand ka kd t½ diss KD KD Rmax Ligand (RU) Analyte (1/Ms) (1/s) (min) (M) (nM) (RU) MR Chi2 huFC 106 4D5-2C4-75mer 1.25E+05 3.39E−05 342 2.71E−10 0.3 83 1.1 0.28 chim. HER2 ECD 104 4D5-75mer 8.54E+04 1.45E−04 80 1.69E−09 1.7 46 1.1 0.18 103 2C4-75mer 8.87E+04 1.17E−04 99 1.32E−09 1.3 46 1.1 0.15

(113) The data presented in Table 8 demonstrate, that the fully established complex, connected via a 75mer ss-L-DNA linker shows cooperative binding. The single FAB fragments show lower affinity, when compared to the fully established complex. The signal levels at Rmax shows the increased molecular mass of the complex versus the single FAB fragments. Despite a higher signal level, the Molar Ratios are exactly at 1.1. This shows that statistically each complex binds to a single huFc chimeric HER2 ECD molecule.

(114) The amplification factor by cooperativity is not so high when compared to the previous assay format, wherein the complex was assembled on the sensor surface. KDcoop is triggered up to 6-fold. Without being bound by theory, this could be due to the nature of the homodimeric huFc chimeric HER2 ECD. Potentially the dual binder recognizes the two separated HER2 ECDs in the huFc HER2 chimera and cannot fully establish cooperativity.

(115) The efficient delivery of an effector moiety in form of a dye could be shown by the FACS analysis (see next example) sing the phycoerythrin-labeled streptavidin probe on living cells. The streptavidin labeled probe could easily access the biotin moiety in the 75mer ss-L-DNA linker construct.

(116) Data form the measurement as outlined above was used for the generation of the Hill Plot (FIG. 9). The Hill analysis of the complex shows, that the binding events of the FAB fragments are independent from each other and don't interfere with each other. No cooperative binding in terms of a structural disturbance of the HER2 molecule could be detected, the Hill coefficient (n=1.00201+/−0.06143) is exactly 1. Therefore, the linker chemistry, the nature of the ss-L-DNA linker and the oligo-labeled FAB fragment are not negatively interfering with the target molecule recognition.

EXAMPLE 8

(117) Further Complexes—Synthesis and Characterization

(118) Synthesis of Hybridizable Oligonucleotides

(119) The following amino modified precursors comprising the sequences given in SEQ ID NOs: 05 and 06, respectively, were synthesized according to standard methods. The below given oligonucleotides not only comprise the so-called aminolinker, but also a fluorescent dye. As the skilled artisan will readily appreciate, this fluorescent dye is very convenient to facilitate purification of the oligonucleotide as such, as well as of components comprising them. a) 5′-Fluorescein-AGT CTA TTA ATG CTT CTG C-(Spacer C3)3-C7Aminolinker-; b) 5′-Cy5 AGT CTA TTA ATG CTT CTG C-(Spacer C3)3-C7Aminolinker-; c) 5′-Aminolinker-(Spacer C3)3-AGT TCT ATC GTC GTC CA-Fluorescein-3′; d) 5′-Fluorescein-(beta L AGT CTA TTA ATG CTT CTG C)-(Spacer C3)3-C7Aminolinker-; (beta L indicates that this is an L-DNA oligonucleotide); and e) 5′-Aminolinker-(Spacer C3)3-(beta L-AGT TCT ATC GTC GTC CA)-Fluorescein-3′ (beta L indicates that this is an L-DNA oligonucleotide).

(120) Synthesis was performed on an ABI 394 synthesizer at a 10 μmol scale in the trityl on (for 5′ amino modification) or trityl off mode (for 3′ amino modification) using commercially available CPGs as solid supports and standard dA(bz), dT, dG (iBu) and dC(Bz) phosphoramidites (Sigma Aldrich).

(121) The following amidites, amino modifiers and CPG supports were used to introduce the C3-spacer, a dye and amino moieties, respectively, during oligonucleotide synthesis: spacer phosphoramidite C3 (3-(4,4′-Dimethoxytrityloxy) propyl-1-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research); 5′ amino modifier is introduced by using 5′-Amino-Modifier C6 (6-(4-Monomethoxytritylamino) hexyl-(2-cyanoethyl)-(N,N-diisopropyl)-phosphoramidite (Glen Research); 5′-Fluorescein Phosphoramidite 6-(3′,6′-dipivaloylfluoresceinyl-6-carboxamido)-hexyl-1-O-(2-cyanoethyl)-(N,N-diisopropyl)-phosphoramidite (Glen Research); Cy5™ Phosphoramidite 1-[3-(4-monomethoxytrityloxy) propyl]-1′-[3-[(2-cyanoethyl)-(N,N-diisopropyl phosphoramidityl] propyl]-3,3,3′,3′-tetramethylindodicarbocyanine chloride (Glen Research); LightCycler Fluorescein CPG 500 A (Roche Applied Science); and 3′-Amino Modifier TFA Amino C-6 lcaa CPG 500 A (ChemGenes).

(122) For Cy5 labeled oligonucleotides, dA(tac), dT, dG(tac), dC(tac) phosphoramidites, (Sigma Aldrich), were used and deprotection with 33% ammonia was performed for 2h at room temperature.

(123) L-DNA oligonucleotides were synthesized by using beta-L-dA(bz), dT, dG (iBu) and dC(Bz) phosphoramidites (ChemGenes)

(124) Purification of fluorescein modified hybridizable oligonucleotides was performed by a two-step procedure: First the oligonucleotides were purified on reversed-phase HPLC (Merck-Hitachi-HPLC; RP-18 column; gradient system [A: 0.1 M (Et.sub.3NH)OAc (pH 7.0)/MeCN 95:5; B: MeCN]: 3 min, 20% B in A, 12 min, 20-50% B in A and 25 min, 20% B in A with a flow rate of 1.0 ml/min, detection at 260 nm. The fractions (monitored by analytical RP HPLC) containing the desired product were combined and evaporated to dryness. (Oligonucleotides modified at the 5′ end with monomethoxytrityl protected alkylamino group are detriylated by incubating with 20% acetic acid for 20 min). The oligomers containing fluorescein as label were purified again by IEX chromatography on a HPLC [Mono Q column: Buffer A: Sodium hydroxide (10 mmol/l; pH ˜12) Buffer B 1 M Sodium chloride dissolved in Sodium hydroxide (10 mmol/l; pH ˜12) gradient: in 30 minutes from 100% buffer A to 100% buffer B flow 1 ml/min detection at 260 nm]. The product was desalted via dialysis.

(125) Cy5 labeled oligomers were used after the first purification on reversed-phase HPLC (Merck-Hitachi-HPLC; RP-18 column; gradient system [A: 0.1 M (Et.sub.3NH)OAc (pH 7.0)/MeCN 95:5; B: MeCN]: 3 min, 20% B in A, 12 min, 20-50% B in A and 25 min, 20% B in A with a flow rate of 1.0 ml/min, detection at 260 nm. The oligomers were desalted by dialysis and lyophilized on a Speed-Vac evaporator to yield solids which were frozen at −24° C.

(126) Activation of Hybridizable Oligonucleotides

(127) The amino modified oligonucleotides from Example 2 were dissolved in 0.1 M sodium borate buffer pH 8.5 buffer (c=600 μmol) and reacted with a 18-fold molar excess of Sulfo SMCC (Sulfosuccinimidyl 4-[N-maleimidomethyl] cyclohexane-1-carboxylate dissolved in DMF (c=3 mg/100 μl) from Thermo Scientific, The reaction product was thoroughly dialyzed against water in order to remove the hydrolysis product of sulfo-SMCC 4-[N-maleimidomethyl] cyclohexane-1-carboxylate.

(128) The dialysate was concentrated by evaporation and directly used for conjugation with a monovalent binder comprising a thiol group.

(129) Synthesis of Linker Oligonucleotides Comprising Hybridizable Oligonucleotides at Both Ends

(130) Oligonucleotides were synthesized by standard methods on an ABI 394 synthesizer at a 10 μmol scale in the trityl on mode using commercially available dT-CPG as solid supports and using standard dA(bz), dT, dG (iBu) and dC(Bz) phosphoramidites (Sigma Aldrich).

(131) L-DNA oligonucleotides were synthesized by using commercially available beta L-dT-CPG as solid support and beta-L-dA(bz), dT, dG (iBu) and dC(Bz) phosphoramidites (ChemGenes)

(132) Purification of the oligonucleotides was performed as described under Example 3 on a reversed-phase HPLC. The fractions (monitored by analytical RP HPLC) containing the desired product were combined and evaporated to dryness. Detritylation was performed by incubating with 80% acetic acid for 15 min). The acetic acid was removed by evaporation. The reminder was dissolved in water and lyophilized

(133) The following amidites and CPG supports were used to introduce the C18 spacer, digoxigenin and biotin group during oligonucleotide synthesis: spacer phosphoramidite 18 (18-O-Dimethoxytritylhexaethyleneglycol, 1-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research); biotin-dT (5′-Dimethoxytrityloxy-5-[N-((4-t-butylbenzoyl)-biotinyl)-aminohexyl)-3-acrylimido]-2′-deoxyUridine-3′-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite (Glen Research); biotin Phosphoramidite 1-Dimethoxytrityloxy-2-(N-biotinyl-4-aminobutyl)-propyl-3-O-(2-cyanoethyl)-(N,N-diisopropyl)-phosphoramidite and 5′-Dimethoxytrityl-5-[N-(trifluoroacetylaminohexyl)-3-acrylimido]-2′-deoxy uridine, 3′-[(2-cyanoethyl)-(N,N-diisopropyl)]-phosphoramidite for amino modification and postlabeling with Digoxigenin-N-Hydroxyl-succinimidyl ester.

(134) The Following Bridging Constructs or Linkers were Synthesized:

(135) TABLE-US-00013 Linker 1: 5′-G CAG AAG CAT TAA TAG ACT-TGG ACG ACG ATA GAA CT-3′ Linker 2: 5-G CAG AAG CAT TAA TAG ACT-(T40)-TGG ACG ACG ATA GAA CT-3′ Linker 3: 5′-[B-L] G CAG AAG CAT TAA TAG ACT-(Biotin-dT)- TGG ACG ACG ATA GAA CT-3′ Linker 4: 5′-[B-L] G CAG AAG CAT TAA TAG ACT-T5-(Biotin- dT)-T5-TGG ACG ACG ATA GAA CT-3′ Linker 5: 5′-[B-L] G CAG AAG CAT TAA TAG ACT-T20-(Biotin- dT)-T20-TGG ACG ACG ATA GAA CT-3′ Linker 6: 5′-[B-L] G CAG AAG CAT TAA TAG ACT-T30-(Biotin- dT)-T30-TGG ACG ACG ATA GAA CT-3′ Linker 7: 5′-GCA GAA GCA TTA ATA GAC T T5-(Biotin-dT)- T5 TG GAC GAC GAT AGA ACT-3′ Linker 8: 5′-GCA GAA GCA TTA ATA GAC T T10-(Biotin-dT)- T10 TGG ACG ACG ATA GAA CT-3′ Linker 9: 5′-GCA GAA GCA TTA ATA GAC T T15-(Biotin-dT)- T15 TGG ACG ACG ATA GAA CT-3′ Linker 10: 5′-GCA GAA GCA TTA ATA GAC T T20-(Biotin-dT)- T20 TGG ACG ACG ATA GAA CT-3′ Linker 11: 5-G CAG AAG CAT TAA TAG ACT-Spacer C18-(Biotin- dT)-Spacer C18-TGG ACG ACG ATA GAA CT-3′ Linker 12: 5′-G CAG AAG CAT TAA TAG ACT-(Spacer C18)2- (Biotin-dT)-(Spacer C18)2-TGG ACG ACG ATA GAA CT-3′ Linker 13: 5′-G CAG AAG CAT TAA TAG ACT-(Spacer C18)3- (Biotin-dT)-(Spacer C18)3-TGG ACG ACG ATA GAA CT-3′ Linker 14: 5′-G CAG AAG CAT TAA TAG ACT-(Spacer C18)4- (Biotin-dT)-(Spacer C18)4-TGG ACG ACG ATA GAA CT- 3′ Linker 15: 5′-G CAG AAG CAT TAA TAG ACT-T20-(Dig-dT)-T20- TGG ACG ACG ATA GAA CT-3′ Linker 16: 5′-G CAG AAG CAT TAA TAG ACT-(Dig-dT)-TGG ACG ACG ATA GAA CT-3′ Linker 17: 5′-G CAG AAG CAT TAA TAG ACT-(Biotin-dT)-TGG ACG ACG ATA GAA CT-3′

(136) The above bridging construct examples comprise at least a first hybridizable oligonucleotide and a second hybridizable oligonucleotide. Linkers 3 to 18 in addition to the hybridizable nucleic acid stretches comprise a central biotinylated or digoxigenylated thymidine, respectively, or a spacer consisting of thymidine units of the length given above.

(137) The 5′-hybridizable oligonucleotide corresponds to SEQ ID NO: 07 and the 3′-hybridizable oligonucleotide corresponds to SEQ ID NO: 08, respectively. The oligonucleotide of SEQ ID NO: 07 will readily hybridize with the oligonucleotide of SED ID NO: 06. The oligonucleotide of SEQ ID NO: 08 will readily hybridize with the oligonucleotide of SED ID NO: 05.

(138) In the above bridging construct examples [B-L] indicates that an L-DNA oligonucleotide sequence is given; spacer C 18, Biotin and Biotin dT respectively, refer to the C18 spacer, the Biotin and the Biotin-dT as derived from the above given building blocks; and T with a number indicates the number of thymidine residues incorporated into the linker at the position given.

(139) Assembly of the Complex

(140) A) Cleavage of IgGs and Labeling of FAB′ Fragments with ss-DNA

(141) Purified monoclonal antibodies were cleaved with the help of pepsin protease yielding F(ab′).sub.2 fragments that are subsequently reduced to FAB′ fragments by treatment with low concentrations of cysteamine at 37° C. The reaction is stopped via separation of cysteamine on a PD 10 column. The FAB′ fragments are labeled with an activated oligonucleotide as produced according to Example 3. This single-stranded DNA (=ss-DNA) bears a thiol-reactive maleimido group that reacts with the cysteines of the FAB′ hinge region. In order to obtain high percentages of single-labeled FAB′ fragments the relative molar ratio of ss-DNA to FAB′-fragment is kept low. Purification of single-labeled FAB′ fragments (ss-DNA: FAB′=1:1) occurs via ion exchange chromatography (column: Source 15 Q PE 4.6/100, Pharmacia/GE). Verification of efficient purification is achieved by analytical gel filtration and SDS-PAGE.

(142) B) Assembly of a Complex Comprising Two Polypeptides Specifically Binding to a Target

(143) The anti-pIGF-1R complex is based on two FAB′ fragments that target different epitopes of the intracellular domain of IGF-1R: FAB′ 8.1.2 detects a phosphorylation site (pTyr 1346) and FAB′ 1.4.168 a non-phospho site of the target protein. The FAB′ fragments have been covalently linked to single-stranded DNA (ss-DNA): FAB′ 1.4.168 to a 17mer ss-DNA comprising SEQ ID NO: 05 and containing fluorescein as a fluorescent marker and FAB′ 8.1.2 to a 19mer ss-DNA comprising SEQ ID NO: 06 and containing Cy5 as fluorescent marker. In the following, these FAB's with covalently bound 17mer or 19mer ss-DNA are named ss-FAB′ 1.4.168 and ss-FAB′ 8.1.2 respectively. Complex assembly is mediated by a linker (i.e. a bridging construct comprising two complementary ss-DNA oligonucleotides (SEQ ID NOs: 7 and 8, respectively) that hybridize to the corresponding ss-DNAs of the ss-FAB′ fragments. The distance between the two ss-FAB′ fragments of the complex can be modified by using spacers, e.g. C18-spacer or DNAs of different length, respectively.

(144) For assembly evaluation the complex components ss-FAB′ 8.1.2, ss-FAB′ 1.4.168 and the linker constructs (1) (=linker 17 of example 2.4) 5′-G CAG AAG CAT TAA TAG ACT T(-Bi)-TGG ACG ACG ATA GAA CT-3′ and (2) (=linker 10 of example 2.4) 5′-G CAG AAG CAT TAA TAG ACT-T(20)-T(-Bi)-(T20)-TGG ACG ACG ATA GAA CT-3′ were mixed in equimolar quantities at room temperature. After a 1 minute incubation step the reaction mix was analyzed on an analytical gel filtration column (Superdex™ 200, 10/300 GL, GE Healthcare). Comparison of the elution volumes (V.sub.E) of the single complex components with the V.sub.E of the reaction mix demonstrates that the complex has been formed successfully (FIG. 10).

(145) BIAcore Experiment Assessing Binding of Anti-pIGF-1R Complex to Immobilized IGF-1R and IR Peptides

(146) For this experiment a BIAcore 2000 instrument (GE Healthcare) was used with a BIAcore SA sensor mounted into the system at T=25° C. Preconditioning occurred at 100 μl/min with 3×1 min injection of 1 M NaCl in 50 mM NaOH and 1 min 10 mM HCl.

(147) HBS-ET (10 mM HEPES pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.05% Tween® 20 was used as system buffer. The sample buffer was identical with the system buffer. The BIAcore 2000 System was driven under the control software V. 1.1.

(148) Subsequently biotinylated peptides were captured on the SA surface in the respective flow cells. 16 RU of IGF-1R(1340-1366)[1346-pTyr; Glu(Bi-PEG-1340]amid (i.e. the—1346 tyrosine phosphorylated—peptide of SEQ ID NO:11 comprising a PEG-linker bound via glutamic acid corresponding to position 1340 and being biotinylated at the other end of the linker) was captured on flow cell 2. 18 RU of IGF-1R(1340-1366); Glu(Bi-PEG-1340]amid (i.e. the—1346 tyrosine non-phosphorylated—peptide of SEQ ID NO: 11 comprising a PEG-linker bound via glutamic acid corresponding to position 1340 and being biotinylated at the other end of the linker) was captured on flow cell 3. 20 RU of hInsR(1355-1382)[1361-pTyr; Glu(Bi-PEG-1355]amid (i.e. the—1361 tyrosine phosphorylated—peptide of SEQ ID NO: 12 comprising a PEG-linker bound via glutamic acid corresponding to position 1355 of human insulin receptor and being biotinylated at the other end of the linker) was captured on flow cell 4. Finally all flow cells were saturated with d-biotin.

(149) For the complex formation the assembly protocol as described above was used. When individual runs with only one of the two ss-FAB's were performed, the absence or presence of linker DNA did not affect the association or dissociation curves.

(150) 100 nM of analyte (i.e. in these experiments a bivalent dual binding agent) in solution was injected at 50 μl/min for 240 sec association time and dissociation was monitored for 500 sec. Efficient regeneration was achieved by using a 1 min injection step at 50 μl/min with 80 mM NaOH. Flow cell 1 served as a reference. A blank buffer injection was used instead of an antigen injection to double reference the data by buffer signal subtraction.

(151) In each measurement cycle one of the following analytes in solution was injected over all 4 flow cells: 100 nM ss-FAB′ 8.1.2, 100 nM ss-FAB′ 1.4.168, a mixture of 100 nM ss-FAB′ 8.1.2 and 100 nM ss-FAB′, 100 nM bivalent binding agent consisting of ss-FAB′ 8.1.2 and ss-FAB′ 1.4.168 hybridized on linker (3) (5′-G CAG AAG CAT TAA TAG ACT-T(20)-T(-Dig)-(T20)-TGG ACG ACG ATA GAA CT-3′(=linker 15)), and 100 nM bivalent binding agent consisting of ss-FAB′ 8.1.2 and ss-FAB′ 1.4.168 hybridized on linker (1) (5′-G CAG AAG CAT TAA TAG ACT-T(-Dig) -TGG ACG ACG ATA GAA CT-3′(=linker 16)), respectively.

(152) The signals were monitored as time-dependent BIAcore sensorgrams.

(153) Report points were set at the end of the analyte association phase (Binding Late, BL) and at the end of the analyte dissociation phase (Stability Late, SL) to monitor the response unit signal heights of each interaction. The dissociation rates kD (1/s) were calculated according to a linear 1:1 Langmuir fit using the BIAcore evaluation software 4.1. The complex halftimes in minutes were calculated upon the formula ln(2)/(60*kD).

(154) The sensorgrams (FIG. 11 to FIG. 14) show a gain in both specificity and complex stability in pIGF-1R binding when ss-FAB′ 1.4.168 and ss-FAB′ 1.4.168 are used in form of a complex (=bivalent binding agent), probably due to the underlying cooperative binding effect. FAB′ 1.4.168 alone shows no cross reactivity for the pIR peptide but does not discriminate between the phosphorylated and non-phosphorylated form of IGF-1R (T1/2 dis=3 min in both cases). FAB′ 8.1.2, however, binds only to the phosphorylated version of the IGF1-R peptide but exhibits some undesired cross reactivity with phosphorylated Insulin Receptor. The complex discriminates well between the pIGF-1R peptide and both other peptides (see FIG. 13) and thus helps to overcome issues of unspecific binding. Note that the gain in specificity is lost when both FAB's are applied without linker DNA (FIG. 14). The gain in affinity of the Complex towards the pIGF-1R peptide manifests in increased dissociation half times compared to individual FAB's and the FAB′ mix omitting the linker DNA (FIG. 12 and FIG. 14). Although the tested Complex s with two different DNA linker share an overall positive effect on target binding specificity and affinity, the longer linker (with T40-Dig as a spacer) (i.e. linker 15) seems to be advantageous with respect to both criteria.

(155) BIAcore Assay Sandwich of M-1.4.168-IgG and M-8.1.2

(156) A BIAcore T100 instrument (GE Healthcare) was used with a BIAcore CM5 sensor mounted into the system. The sensor was preconditioned by a 1 min injection at 100 μl/min of 0.1% SDS, 50 mM NaOH, 10 mM HCl and 100 mM H3PO4.

(157) The system buffer was HBS-ET (10 mM HEPES pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.05% Tween® 20). The sample buffer was the system buffer.

(158) The BIAcore T100 System was driven under the control software V1.1.1. Polyclonal rabbit IgG antibody <IgGFCγM>R (Jackson ImmunoResearch Laboratories Inc.) at 30 μg/ml in 10 mM Na-Acetate pH 4.5 was immobilized at 10 000 RU on the flow cells 1, 2, 3, and 4, respectively, via EDC/NHS chemistry according to the manufacturer's instructions. Finally, the sensor surface was blocked with 1M ethanolamine. The complete experiment was driven at 13° C.

(159) 500 nM primary mAb M-1.004.168-IgG was captured for 1 min at 10 μl/min on the <IgGFCγM>R surface. 3 μM of an IgG fragment mixture (of IgG classes IgG1, IgG2a, IgG2b, IgG3) containing blocking solution was injected at 30 μl/min for 5 min. The peptide IGF-1R(1340-1366)[1346-pTyr; Glu(Bi-PEG-1340]amid was injected at 300 nM for 3 min at 30 μl/min. 300 nM secondary antibody M-8.1.2-IgG was injected at 30 μl min. The sensor was regenerated using 10 mM Glycine-HCl pH 1.7 at 50 μl/min for 3 min.

(160) In FIG. 15 the assay setup is presented. In FIG. 18 the measurement results are given. The measurements clearly indicate that both monoclonal antibodies are able to simultaneously bind two distinct, unrelated epitopes on their respective target peptide. This is a prerequisite to any latter experiments with the goal to generate cooperative binding events.

(161) BIAcore Assay Complex on Sensor Surface

(162) A BIAcore 3000 instrument (GE Healthcare) was used with a BIAcore SA sensor mounted into the system at T=25° C. The system was preconditioned at 100 μl/min with 3×1 min injection of 1 M NaCl in 50 mM NaOH and 1 min 10 mM HCl.

(163) The system buffer was HBS-ET (10 mM HEPES pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.05% Tween® 20). The sample buffer was the system buffer.

(164) The BIAcore 3000 System was driven under the control software V4.1.

(165) 124 RU amino-PEO-biotin were captured on the reference flow cell 1. 1595 RU biotinylated 14.6 kDa TO-Bi 37-mer ss-DNA-Linker (1) (5′-G CAG AAG CAT TAA TAG ACT-T(-Bi)-TGG ACG ACG ATA GAA CT-3′) (=linker 17 of example 2.4) and 1042 RU biotinylated 23.7 kDa T40-Bi 77-mer ss-DNA-Linker (2) (5′-G CAG AAG CAT TAA TAG ACT-T(20)- (Biotin-dT)-(T20)-TGG ACG ACG ATA GAA CT-3′=linker 10) were captured on different flow cells. 300 nM ss-FAB 8.1.2 and 300 nM ss-FAB 1.004.168 were injected into the system at 50 μl/min for 3 min. As a control only 300 nM ss-FAB 8.1.2 or 300 nM ss-FAB 1.004.168 was injected to test the kinetic contribution of each ss-FAB. As a control, buffer was injected instead of the ss-Fabs. The peptides IGF-1R(1340-1366)[1346-pTyr]amid, INR(1355-1382)[1361-pTyr]amid IGF-1R(1340-1366)amid and were injected into system at 50 μl/min for 4 min, free in solution, in concentration steps of 0 nM, 4 nM, 11 nM, 33 nM (twice), 100 nM and 300 nM. In another embodiment to measure the affinities versus peptides IGF-1R(1340-1366)[1346-pTyr]amid the concentration steps of 0 nM, 0.4 nM, 1.1 nM, 3.3 nM (twice), 10 nM and 30 nM.

(166) The dissociation was monitored at 50 μl/min for 5.3 min. The system was regenerated after each concentration step with a 12 sec pulse of 250 mM NaOH and was reloaded with ss-FAB ligand.

(167) FIG. 17 schematically describes the assay setup on the BIAcore instrument. The tables given in FIG. 18 show the quantification results from this approach. FIG. 19, FIG. 20 and FIG. 21 depict exemplary BIAcore results from this assay setup.

(168) The table in FIG. 18 demonstrates the benefits of the complex concept. The T40 dual binding agent (a dual binding agent with linker 10 of example 2.4, i.e. a linker with a spacer of T20-Biotin-dT-T20) results in a 2-fold improved antigen complex halftime (414 min) and a 3-fold improved affinity (10 pM) as compared to the TO dual binding agent (i.e. a dual binding agent with linker 16) with 192 min and 30 pM, respectively. This underlines the necessity to optimize the linker length to generate the optimal cooperative binding effect.

(169) The T40 dual binding agent (i.e. the dual binding agent comprising the T40-Bi linker (linker 10)) exhibits a 10 pM affinity versus the phosphorylated IGF-1R peptide (table in FIG. 18, FIG. 19). This is a 2400-fold affinity improvement versus the phosphorylated insulin receptor peptide (24 nM) and a 100-fold improvement versus the non-phosphorylated IGF-1R peptide.

(170) Therefore, the goal to increase specificity and affinity by the combination of two distinct and separated binding events is achieved.

(171) The cooperative binding effect especially becomes obvious from the dissociation rates against the phosphorylated IGF-1R peptide, where the complex shows 414 min antigen complex halftime, versus 0.5 min with the monovalent binder 8.1.2 alone and versus 3 min with the monovalent binder 1.4.168 alone, respectively.

(172) Furthermore, the fully assembled construct roughly multiplies its dissociation rates kD (1/s), when compared to the singly FAB hybridized constructs (FIG. 21, FIG. 20, FIG. 21 and table in FIG. 18). Interestingly, also the association rate ka (1/Ms) slightly increases when compared to the single FAB interaction events, this may be due to an increase of the construct's molecular flexibility.

EXAMPLE 9

(173) Binding Assays—In Vitro and Ex Vivo

(174) Detection Oligonucleotide Probe-Cy5

(175) The ss-L-DNA detection oligonucleotide Probe-Cy5 5′ Cy5-Y-ATG CGA-GTA CCT TAG AGT C-Z-Cy5 3′ (SEQ ID NO: 72), has been synthesized by state of the art oligonucleotide synthesis methods. The introduction of the Cy5 dye was done via reaction of the amino groups with Cy5 monoreactive NHS ester. (GE Healthcare Lifescience, STADT, LAND). For the nucleotides L-DNA amidites (ChemGenes, STADT, LAND) were used. The 5′ and 3′ amino groups were introduced during the solid phase oligonucleotide synthesis process wherein Y=5′-Amino-Modifier C6 introduced via (6-(4-Monomethoxytritylamino)hexyl-(2-cyanoethyl)-(N,N-diisopropyl)-phosphoramidite (Glen Research), and Z=3′-Aminomodifier C6 introduced via 3′Aminomodifier TFA Amino C6 long chain aminoalkyl Controlled Pore Glass 1000 A (ChemGenes).

(176) Dual Binder Linker Oligonucleotide

(177) The ss-L-DNA oligonucleotide linker SEQ ID NO: 73 5′-G CAG AAG CAT TAA TAG ACT-T20-GAC TCT AAG GTA CTC GCA T-T20-TGG ACG ACG ATA GAA CT-3′ has been synthesized by state of the art oligonucleotide synthesis methods.

(178) Assembly of the Complex

(179) The complex was assembled by hybridizing the anti-HER2 antibody 2C4-FAB′-ss-L-DNA labeled with FITC and the anti-HER2 antibody 4D5-FAB′-ss-L-DNA labeled with FITC in equimolar stoichiometry with the ss-L-DNA linker of SEQ ID NO: 73. In order to verify the correct assembly of the complex, the complex was subjected to an SEC chromatography step and was filtered through a sterile filter.

(180) In Vitro Binding Assay

(181) Human breast cancer KPL-4 cells were seeded with a concentration of 2×10.sup.6 cells/ml in a volume of 30 μl into μ-slides VI (ibidi, Germany). Three hours thereafter, 70 μl medium (RPMI 1640, 2 mM L-glutamine, 10% FCS) was added to allow the cells to adhere.

(182) After an incubation of 24 hours at 37° C. and 5% CO.sub.2 in a water saturated atmosphere (effective for all following incubations), the supernatant was removed and cells were washed once with 100 μl PBS to remove residual medium.

(183) For the sequential application, 50 μl of the complex 4D5-2C4 as prepared above labeled with FITC solution (c=2.5 μg/ml) was added and incubated for 45 minutes, followed by one washing step with 100 μl PBS and a further incubation with 50 μl of the DNA-probe (SEQ ID NO: 72) at an equimolar amount (0.13 μg/ml).

(184) The pre-mixed procedure was performed by first mixing the complex and the detection Probe. Thereafter it was added to the cells (concentrations see above) followed by incubation for 45 minutes.

(185) Xolair®, a humanized IgG1 monoclonal antibody targeting human IgE immunoglobulin was used as a negative control and Herceptin® labeled with Cy5 targeting human HER-2 receptor was used as a positive control. Both antibodies were applied at the same concentration (2.5 μg/ml).

(186) Subsequently, the supernatant was removed and cells were washed once with 100 μl PBS. Cell nuclei were afterwards stained with DAPI by adding 50 μl of a HOECHST33342 solution (c=10 μg/ml) and incubated for 15 minutes. To remove the cell staining dye, cells were washed twice with 100 μl PBS after removal of the supernatant. Another 120 μl PBS were added to keep the cells moist to ensure viability. All dilutions were made with medium (without L-Glutamine and FCS) to ensure viability of the cells and to avoid detachment of the cells. After this procedure, slides were imaged by multispectral fluorescence analysis using the NUANCE System (CRi, Cambridge, USA). Images were normalized for comparability of the fluorescence intensities.

(187) Ex Vivo Analysis

(188) Immunodeficient SCID beige mice with established KPL-4 tumors (orthotopically implanted) were injected i.v. with 50 μg complex in 100 μl PBS and 18 hours thereafter the Cy5-labeled DNA-probe was injected at an equimolar concentration (2.63 μg per mouse). Tumors were explanted 48 hours thereafter and examined by multispectral fluorescence analysis using the MAESTRO system (CRi, Cambridge, USA).

(189) Results

(190) In Vitro Binding Assay

(191) The complex is doubly FITC labeled via each of its FAB′-ss-L-DNA components.

(192) The detection probe is a doubly Cy5 labeled ss-L-DNA 20-mer oligonucleotide probe, which can be hybridized to the 95mer ss-L-DNA linker of the complex.

(193) In contrast to Xolair-Cy5 (no fluorescence signal, negative control) Herceptin-Cy5 specifically stained the tumor cells (FIG. 22). The FITC labeled 4D5-2C4-95mer complex specifically binds to KPL-4 tumor cells as can be seen by sequential incubation with the detection Probe (measured in the Cy5 fluorescence channel) which is co-localized with the complex to the tumor cells indicating the hybridization of the detection oligonucleotide Probe-Cy5 to the complex. In the sequential incubation mode as well as in the pre-mixed setting specific staining of the tumor cells to the Her-2 antigen could be demonstrated (FIG. 2).

(194) In FIG. 22 the near infrared image of the cancer cells incubated with the complex and the detection probe is shown (NIRF imaging). In the top right of the Figure a sketch of the fully assembled 4D5-2C4-95mer complex hybridized to the Cy5 labeled detection oligonucleotide is shown. In the middle right of the Figure a cartoon of Cy5 labeled Herceptin is shown. In the bottom right of the Figure the signal intensity bar is shown.

(195) In the top left of FIG. 22 the binding of Cy5 labeled Herceptin® to the cancer cells is shown (positive control). The KPL-4 cell membranes appear as bright lighting rings surrounding the DAPI-stained cell nuclei. In the bottom left the incubation of Cy5 labeled Xolair® is shown (negative control). No membrane staining but the DAPI stain of the cell nuclei can be detected. In the bottom middle of the Figure the binding of the 4D5-2C4-FITC complex is shown. The fluorescein signal of the membrane bound complex appears as lighting rings surrounding the DAPI stained cell nuclei. In the top middle of the Figure the binding of the 4D5-2C4-FITC complex and the Cy5 labeled detection probe is shown. The detection of the complex via the Cy5 labeled ss-L-DNA detection probe, which was sequentially hybridized, can be seen. The Cy5 signal of the detection oligonucleotide appears as membrane staining, showing bright lighting rings surrounding the DAPI stained cell nuclei.

(196) In FIG. 23 the near infrared (NIRF) imaging of KPL-4 cells is shown. In FIG. 23 A the results of the sequential application of FITC labeled 4D5-2C4 complex and the Cy5 labeled detection probe is shown. In FIG. 23 B the results of the incubation of KPL-4 cells with premixed FITC labeled 4D5-2C4 complex and Cy5 labeled detection probe is shown. Both images show membrane-located signals. As a control, cells were stained with DAPI.

(197) The experiment demonstrates that the complex as reported herein can first be applied in order to specifically target HER-2 positive cells. In a second step, the labeled detection probe can be applied in order to hybridize to the target bound complex. The fluorescence labeled detection probe is thereby a proof of concept for the time delayed, sequential application and specific targeting of an oligonucleotide-based effector moiety. In this case the payload is a fluorescent dye for the purpose of in vitro cell imaging.

(198) Ex Vivo Binding Assay

(199) As depicted in FIG. 24 (left image) a strong fluorescence signal is detectable in the experimental setting where the sample was incubated first with the complex and thereafter with the Cy5 labeled detection probe. In contrast (right image), no fluorescence signal could be detected in the tumors previously injected in the KPL-4 xenograft with the Cy5 labeled detection probe alone.

(200) FIG. 24 shows explanted KPL-4 tumors subjected to NIRF Imaging. In the first image Cy5 fluorescence signals obtained from three KPL-4 tumors explanted from mice, which were sequentially treated with the first the 4D5-2C4 complex and thereafter the detection probe is shown. In the right image it is shown that no fluorescence signal was obtained from three KPL-4 tumors, when three mice where treated with detection probe alone, omitting the 4D5-2C4 complex.

EXAMPLE 10

(201) Inhibition of Cell Proliferation in MDA-MB-175 Breast Cancer Cell Line

(202) 2×10.sup.4 MDA-MB-175 breast cancer cells cultured in DMEM/F12 medium supplemented with 10% fetal calve serum, 2 mM Glutamin and Penicillin/Streptomycine were seeded in 96-well plates. Antibodies and complex, respectively, were added in the indicated concentrations the next day (40 to 0.0063 μg/ml). Alter 6 day incubation Alamar Blue was added and plates were incubated for 3-4 h in a tissue culture incubator. Fluorescence was measured (excitation 530 nm/emission 590) and percentage inhibition was calculated using untreated cells as reference.

(203) Results

(204) The anti-HER2 antibody 2C4 (Pertuzumab) showed a maximum inhibition of 44%. The anti-HER2 antibody Herceptin showed a maximum inhibition of 9%. The complex as reported herein comprising the FAB fragments of Pertuzumab and Herceptin® shows a maximum inhibition of 46%.

(205) It has to be pointed out that Petuzumab was tested as full length IgG antibody with two HER2 binding sites, whereas the complex comprises a single Pertuzumab Fab fragment with a single HER2 binding site.

EXAMPLE 11

(206) Freeze-Thaw-Stability of the Complex

(207) The complex was assembled by hybridizing the anti-HER2 antibody 2C4-FAB′-ss-L-DNA labeled with FITC and the anti-HER2 antibody 4D5-FAB′-ss-L-DNA labeled with FITC in equimolar stoichiometry with the ss-L-DNA linker of SEQ ID NO: 73. In order to verify the correct assembly of the complex, the complex was subjected to a SEC chromatography step and was filtered through a sterile filter.

(208) Fifty μl of the complex (1.5 mg/ml) were analyzed by analytical SEC using a TSK3000 column (GE). The running buffer was 0.1 M KH.sub.2PO.sub.4 pH 6.8. The flow rate was 1 ml/min. The chromatogram is shown in FIG. 25.

(209) After freezing and thawing, the complex was re-chromatographed. Fifty μl of the complex (1.5 mg/ml) were analyzed by analytical SEC using a TSK3000 column (GE). The running buffer was 0.1 M KH.sub.2PO.sub.4 pH 6.8. The flow rate was 1 ml/min. The chromatogram is shown in FIG. 26.