A Conjugate Based on Anti-HER2 Diaffibody and Auristatin E Targets HER2-Positive Cancer Cells

Antibody-drug conjugates (ADCs) have recently emerged as efficient and selective cancer treatment therapeutics. Currently, alternative forms of drug carriers that can replace monoclonal antibodies are under intensive investigation. Here, a cytotoxic conjugate of an anti-HER2 (Human Epidermal Growth Factor Receptor 2) diaffibody with monomethyl-auristatin E (MMAE) is proposed as a potential anticancer therapeutic. The anti-HER2 diaffibody was based on the ZHER2:4 affibody amino acid sequence. The anti-HER2 diaffibody has been expressed as a His-tagged protein in E. coli and purified by Ni-nitrilotriacetyl (Ni-NTA) agarose chromatography. The molecule was properly folded, and the high affinity and specificity of its interaction with HER2 was confirmed by surface plasmon resonance (SPR) and flow cytometry, respectively. The (ZHER2:4)2DCS-MMAE conjugate was obtained by coupling the maleimide group linked with MMAE to cysteines, which were introduced in a drug conjugation sequence (DCS). Cytotoxicity of the conjugate was evaluated using the 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide MTT assay and the xCELLigence Real-Time Cell Analyzer. Our experiments demonstrated that the conjugate delivered auristatin E specifically to HER2-positive tumor cells, which finally led to their death. These results indicate that the cytotoxic diaffibody conjugate is a highly potent molecule for the treatment of various types of cancer overexpressing HER2 receptors.


Introduction
Antibody-drug conjugates (ADCs) are one of the most advanced approaches in the targeted treatment of cancers and are the leading cause of death in developed countries [1]. This promising strategy is based on the conjugation of monoclonal antibodies with cytotoxic drugs [2][3][4]. ADCs allow one to use the high selectivity of antibodies for the targeted delivery of cargo molecules directly to the tumor tissue [5,6]. Upon binding to tumor specific markers, ADCs undergo internalization followed by drug release in lysosomes, which eventually leads to cell death [7].
Currently, two ADCs, brentuximab vedotin (Adcetris, Seattle Genetics, Bothell, WA, USA) [8,9] and trastuzumab emtansine (Kadcyla, Genentech, South San Francisco, CA, USA) [10,11], are approved by the US Food and Drug Administration (FDA). A significant effort has been made to improve ADCs The final yield of the diaffibody was 20-25 mg from one litre of bacterial culture. The calculated molecular weight of the construct, 17,882 kDa, was confirmed by SDS-PAGE and mass spectrometry.

Structure and Thermal Stability of the (ZHER2:4)2DCS Diaffibody
The secondary structure of (ZHER2:4)2DCS was analyzed by circular dichroism (CD). The CD spectra were acquired in the range of 260 to 200 nm at 21 °C using 1 μM protein concentration and a 1 cm path length quartz cuvette. The CD spectrum was averaged over three scans ( Figure 2). Analysis of the secondary structure content in the diaffibody showed that it represents a folded protein of αhelical structure. Quantitative analysis was performed using the DichoroWeb server, with the use of SELCON3 [38] and K2D algorithms, and CDpro software [39] using CDSSTR, SELCON3, and CONTIN/LL algorithms with SP43, SDP48, and SMP56 reference sets. Our results indicate that the (ZHER2:4)2DCS diaffibody contains more than 80% of α-helical structures. This is in accordance with the nuclear magnetic resonance (NMR) structure of a diaffibody protein that adopts a classical 'up-down' three-helical bundle fold [40]. To determine the stability of the designed protein, we performed thermal denaturation experiments ( Figure 3). The denaturation process of (ZHER2:4)2DCS was monitored by circular dichroism (CD) in phosphate buffer, pH 7.4, at 222 nm. Thermodynamic parameters were calculated assuming a two-state reversible equilibrium transition. The denaturation temperature and van't Hoff enthalpy are 57 °C and 46 kcal/mol, respectively.

Structure and Thermal Stability of the (Z HER2:4 ) 2 DCS Diaffibody
The secondary structure of (Z HER2:4 ) 2 DCS was analyzed by circular dichroism (CD). The CD spectra were acquired in the range of 260 to 200 nm at 21 • C using 1 µM protein concentration and a 1 cm path length quartz cuvette. The CD spectrum was averaged over three scans ( Figure 2). Analysis of the secondary structure content in the diaffibody showed that it represents a folded protein of α-helical structure. Quantitative analysis was performed using the DichoroWeb server, with the use of SELCON3 [38] and K2D algorithms, and CDpro software [39] using CDSSTR, SELCON3, and CONTIN/LL algorithms with SP43, SDP48, and SMP56 reference sets. Our results indicate that the (Z HER2:4 ) 2 DCS diaffibody contains more than 80% of α-helical structures. This is in accordance with the nuclear magnetic resonance (NMR) structure of a diaffibody protein that adopts a classical 'up-down' three-helical bundle fold [40]. To determine the stability of the designed protein, we performed thermal denaturation experiments ( Figure 3). The denaturation process of (Z HER2:4 ) 2 DCS was monitored by circular dichroism (CD) in phosphate buffer, pH 7.4, at 222 nm. Thermodynamic parameters were calculated assuming a two-state reversible equilibrium transition. The denaturation temperature and van't Hoff enthalpy are 57 • C and 46 kcal/mol, respectively.

Structure and Thermal Stability of the (ZHER2:4)2DCS Diaffibody
The secondary structure of (ZHER2:4)2DCS was analyzed by circular dichroism (CD). The CD spectra were acquired in the range of 260 to 200 nm at 21 °C using 1 μM protein concentration and a 1 cm path length quartz cuvette. The CD spectrum was averaged over three scans ( Figure 2). Analysis of the secondary structure content in the diaffibody showed that it represents a folded protein of αhelical structure. Quantitative analysis was performed using the DichoroWeb server, with the use of SELCON3 [38] and K2D algorithms, and CDpro software [39] using CDSSTR, SELCON3, and CONTIN/LL algorithms with SP43, SDP48, and SMP56 reference sets. Our results indicate that the (ZHER2:4)2DCS diaffibody contains more than 80% of α-helical structures. This is in accordance with the nuclear magnetic resonance (NMR) structure of a diaffibody protein that adopts a classical 'up-down' three-helical bundle fold [40]. To determine the stability of the designed protein, we performed thermal denaturation experiments ( Figure 3). The denaturation process of (ZHER2:4)2DCS was monitored by circular dichroism (CD) in phosphate buffer, pH 7.4, at 222 nm. Thermodynamic parameters were calculated assuming a two-state reversible equilibrium transition. The denaturation temperature and van't Hoff enthalpy are 57 °C and 46 kcal/mol, respectively.

Structure and Thermal Stability of the (ZHER2:4)2DCS Diaffibody
The secondary structure of (ZHER2:4)2DCS was analyzed by circular dichroism (CD). The CD spectra were acquired in the range of 260 to 200 nm at 21 °C using 1 μM protein concentration and a 1 cm path length quartz cuvette. The CD spectrum was averaged over three scans ( Figure 2). Analysis of the secondary structure content in the diaffibody showed that it represents a folded protein of αhelical structure. Quantitative analysis was performed using the DichoroWeb server, with the use of SELCON3 [38] and K2D algorithms, and CDpro software [39] using CDSSTR, SELCON3, and CONTIN/LL algorithms with SP43, SDP48, and SMP56 reference sets. Our results indicate that the (ZHER2:4)2DCS diaffibody contains more than 80% of α-helical structures. This is in accordance with the nuclear magnetic resonance (NMR) structure of a diaffibody protein that adopts a classical 'up-down' three-helical bundle fold [40]. To determine the stability of the designed protein, we performed thermal denaturation experiments ( Figure 3). The denaturation process of (ZHER2:4)2DCS was monitored by circular dichroism (CD) in phosphate buffer, pH 7.4, at 222 nm. Thermodynamic parameters were calculated assuming a two-state reversible equilibrium transition. The denaturation temperature and van't Hoff enthalpy are 57 °C and 46 kcal/mol, respectively.

Specificity of the Dimeric Anti-HER2 Affibody
In order to analyze by flow cytometry the specificity of the anti-HER2 diaffibody binding to HER2 present on cancer cells, (Z HER2:4 ) 2 DCS was fluorescently labeled with fluorescein isothiocyanate (FITC). Labeling was confirmed by mass spectrometry that showed traces of the unmodified diaffibody as well as the diaffibody labeled with one, two or three fluorescein molecules. The fluorescently labeled anti-HER2 diaffibody was used to stain the SK-BR-3 cells, which strongly overexpress HER2, and the control U-87 MG cells, which have physiological levels of HER2. The HER2 status of these cell lines was previously confirmed by SDS-PAGE analysis [41]. A similar experiment was also performed with commercially available anti-HER2 mouse monoclonal antibodies, followed by donkey anti-mouse polyclonal antibodies conjugated with FITC. Analysis of the histograms confirmed that diaffibodies bind to the HER2-positive cells in a concentration-dependent manner (Figure 4b) similar to the anti-HER2 monoclonal antibody ( Figure 4a). As expected, the HER2-negative cells were not stained with either (Z HER2:4 ) 2 DCS-FITC or the anti-HER2 monoclonal antibody (Figure 4c).

Specificity of the Dimeric Anti-HER2 Affibody
In order to analyze by flow cytometry the specificity of the anti-HER2 diaffibody binding to HER2 present on cancer cells, (ZHER2:4)2DCS was fluorescently labeled with fluorescein isothiocyanate (FITC). Labeling was confirmed by mass spectrometry that showed traces of the unmodified diaffibody as well as the diaffibody labeled with one, two or three fluorescein molecules. The fluorescently labeled anti-HER2 diaffibody was used to stain the SK-BR-3 cells, which strongly overexpress HER2, and the control U-87 MG cells, which have physiological levels of HER2. The HER2 status of these cell lines was previously confirmed by SDS-PAGE analysis [41]. A similar experiment was also performed with commercially available anti-HER2 mouse monoclonal antibodies, followed by donkey anti-mouse polyclonal antibodies conjugated with FITC. Analysis of the histograms confirmed that diaffibodies bind to the HER2-positive cells in a concentrationdependent manner (Figure 4b) similar to the anti-HER2 monoclonal antibody ( Figure 4a). As expected, the HER2-negative cells were not stained with either (ZHER2:4)2DCS-FITC or the anti-HER2 monoclonal antibody (Figure 4c).

(ZHER2:4)2DCS-MMAE Preparation
MC-Val-Cit-PABC-MMAE (referred to as vcMMAE), which was used in this study, is composed of a maleimide attachment group (MC) that allows conjugation with the target protein via thiol groups, followed by a valine-citrulline (vc) linker and monomethyl auristatin E (MMAE). The linker is cleaved by cathepsins inside the endosomes of target cells. The MMAE molecule is separated from the cathepsin recognition site with a self-immolative p-aminobenzoic acid (PABC) spacer ( Figure 5a). Prior to conjugation, (ZHER2:4)2DCS was reduced with 1 μM tris(2-carboxyethyl)phosphine (TCEP) for 1 h at room temperature. Following reduction, we used a 10-fold molar excess of vcMMAE and incubated the mixture for another 2 h. The conjugation occurred with about 70% efficiency. We attempted to separate the conjugation products using hydrophobic interaction chromatography (HIC), but we were unable to achieve a good separation. Therefore, the mixture was simply dialyzed against phosphate buffer saline (PBS). The mass spectrometry (MS) and SDS-PAGE analyses of the conjugation mixture showed that none, one, two, or three molecules of MMAE were conjugated to the diaffibody molecule due to the three cysteine residues inserted at the C-terminus of the diaffibody (Figure 5b,c). The conjugation product is referred to as (ZHER2:4)2DCS-MMAE.

(Z HER2:4 ) 2 DCS-MMAE Preparation
MC-Val-Cit-PABC-MMAE (referred to as vcMMAE), which was used in this study, is composed of a maleimide attachment group (MC) that allows conjugation with the target protein via thiol groups, followed by a valine-citrulline (vc) linker and monomethyl auristatin E (MMAE). The linker is cleaved by cathepsins inside the endosomes of target cells. The MMAE molecule is separated from the cathepsin recognition site with a self-immolative p-aminobenzoic acid (PABC) spacer ( Figure 5a). Prior to conjugation, (Z HER2:4 ) 2 DCS was reduced with 1 µM tris(2-carboxyethyl)phosphine (TCEP) for 1 h at room temperature. Following reduction, we used a 10-fold molar excess of vcMMAE and incubated the mixture for another 2 h. The conjugation occurred with about 70% efficiency. We attempted to separate the conjugation products using hydrophobic interaction chromatography (HIC), but we were unable to achieve a good separation. Therefore, the mixture was simply dialyzed against phosphate buffer saline (PBS). The mass spectrometry (MS) and SDS-PAGE analyses of the conjugation mixture showed that none, one, two, or three molecules of MMAE were conjugated to the diaffibody molecule due to the three cysteine residues inserted at the C-terminus of the diaffibody (Figure 5b,c). The conjugation product is referred to as (Z HER2:4 ) 2 DCS-MMAE. The affinity of the monomeric ZHER2:4 binding to HER2 was previously assessed for ~50 nM [28], while the dimeric construct bound HER2 with a higher affinity of ~3 nM [36]. To verify whether (ZHER2:4)2DCS retains the high affinity for HER2, we performed surface plasmon resonance (SPR) analysis. Recombinant HER2 protein was immobilized on the sensor chip and the diaffibody was injected at three concentrations; 0.01, 0.1, and 1 μM. The binding parameters were calculated using the simple 1:1 Langmuir binding model. The obtained data indicate that both (ZHER2:4)2DCS and (ZHER2:4)2DCS-MMAE bind to HER2 with high affinity, since the apparent dissociation equilibrium constant (KD) was about 18 nM for the diaffibody and for its conjugate ( Figure 6  The affinity of the monomeric Z HER2:4 binding to HER2 was previously assessed for~50 nM [28], while the dimeric construct bound HER2 with a higher affinity of~3 nM [36]. To verify whether (Z HER2:4 ) 2 DCS retains the high affinity for HER2, we performed surface plasmon resonance (SPR) analysis. Recombinant HER2 protein was immobilized on the sensor chip and the diaffibody was injected at three concentrations; 0.01, 0.1, and 1 µM. The binding parameters were calculated using the simple 1:1 Langmuir binding model. The obtained data indicate that both (Z HER2:4 ) 2 DCS and (Z HER2:4 ) 2 DCS-MMAE bind to HER2 with high affinity, since the apparent dissociation equilibrium constant (K D ) was about 18 nM for the diaffibody and for its conjugate ( Figure 6). However, (Z HER2:4 ) 2 DCS and (Z HER2:4 ) 2 DCS-MMAE exhibited different association (k on ) and dissociation (k off ) rate constants. The k on and k off determined for (Z HER2:4 ) 2 DCS were 8.7 × 10 4 M −1 ·s −1 (standard error (SE) = 1.3 × 10 4 ) and 1.6 × 10 −3 s −1 (SE = 1 × 10 −4 ), respectively, whereas the conjugate was characterized by k on of 6.4 × 10 4 M −1 ·s −1 (SE = 3.5 × 10 2 ) and k off of 1.2 × 10 −3 s −1 (SE = 4.5 × 10 −5 ).

Cytotoxicity of the (ZHER2:4)2DCS-MMAE Conjugate
Prior to quantitative cytotoxicity assays on several cell lines, we decided to investigate whether the (ZHER2  The SK-BR-3 cells showed no morphological changes upon treatment with PBS and the unconjugated (ZHER2:4)2DCS molecule. In contrast, the cells treated for 72 h with 500 nM (ZHER2:4)2DCS-MMAE exhibited a round shape. We could also observe a significantly lower number of the cells that were attached to the culture plate surface, which indicated that the conjugate killed a vast majority of SK-BR-3 cells (Figure 7). Additionally, we monitored SK-BR-3 cells treated with different concentrations of our conjugate in real time using the xCELLigence Real-Time Cell Analyzer (ACEA Biosciences, San Diego, CA, USA) ( Figure 8). The xCELLigence instrument measures change in electrical impedance, which depends on cell attachment to the culture plate with built-in microelectrodes. This allows for non-invasive, label-free cell death estimation. The obtained results confirmed that SK-BR-3 cells are sensitive to all the doses of (ZHER2:4)2DCS-MMAE used in this experiment.

Cytotoxicity of the (ZHER2:4)2DCS-MMAE Conjugate
Prior to quantitative cytotoxicity assays on several cell lines, we decided to investigate whether the (  The SK-BR-3 cells showed no morphological changes upon treatment with PBS and the unconjugated (ZHER2:4)2DCS molecule. In contrast, the cells treated for 72 h with 500 nM (ZHER2:4)2DCS-MMAE exhibited a round shape. We could also observe a significantly lower number of the cells that were attached to the culture plate surface, which indicated that the conjugate killed a vast majority of SK-BR-3 cells (Figure 7). Additionally, we monitored SK-BR-3 cells treated with different concentrations of our conjugate in real time using the xCELLigence Real-Time Cell Analyzer (ACEA Biosciences, San Diego, CA, USA) ( Figure 8). The xCELLigence instrument measures change in electrical impedance, which depends on cell attachment to the culture plate with built-in microelectrodes. This allows for non-invasive, label-free cell death estimation. The obtained results confirmed that SK-BR-3 cells are sensitive to all the doses of (ZHER2:4)2DCS-MMAE used in this experiment. The SK-BR-3 cells showed no morphological changes upon treatment with PBS and the unconjugated (Z HER2:4 ) 2 DCS molecule. In contrast, the cells treated for 72 h with 500 nM (Z HER2:4 ) 2 DCS-MMAE exhibited a round shape. We could also observe a significantly lower number of the cells that were attached to the culture plate surface, which indicated that the conjugate killed a vast majority of SK-BR-3 cells (Figure 7). Additionally, we monitored SK-BR-3 cells treated with different concentrations of our conjugate in real time using the xCELLigence Real-Time Cell Analyzer (ACEA Biosciences, San Diego, CA, USA) ( Figure 8). The xCELLigence instrument measures change in electrical impedance, which depends on cell attachment to the culture plate with built-in microelectrodes. This allows for non-invasive, label-free cell death estimation. The obtained results confirmed that SK-BR-3 cells are sensitive to all the doses of (Z HER2:4 ) 2 DCS-MMAE used in this experiment.   [41,42]. The above-mentioned cells were treated with increasing doses of the conjugate, and their viability was monitored after 72, 96, and 120 h. According to the obtained results, 0.8 nM (Z HER2:4 ) 2 DCS-MMAE reduced the SK-BR-3 cells' viability to 15% within 120 h (Figure 9a). This concentration of the conjugate had a minor effect on the viability of MDA-MB-453 cells, which express lower levels of HER2 than SK-BR-3 cells [41,42], and no effect on the viability of T-47D cells (Figure 9). A 5-fold increase of (Z HER2:4 ) 2 DCS-MMAE concentration caused a decrease of MDA-MB-453 viability to 15% within 120 h (Figure 9b). To achieve a similar reduction of T-47D viability, we had to apply 100 nM (Z HER2:4 ) 2 DCS-MMAE for 120 h (Figure 9c). The half maximal inhibitory concentration (IC 50 ) for SK-BB-3 cells treated with (Z HER2:4 ) 2 DCS-MMAE for 96 h was lower than 0.5 nM. This parameter was slightly elevated for MDA-MB-453 cells (IC 50 = 1.9 nM) and T-47D cells (IC 50 = 5.5 nM). The obtained results demonstrated that the effect of (Z HER2:4 ) 2 DCS-MMAE depends on the level of HER2 expression, since the (Z HER2:4 ) 2 DCS-MMAE potency increased along with the amount of HER2 expressed on the cell surface.
The MTT proliferation assay was also performed for the U-87 MG (Figure 10a), SK-MES-1 (Figure 10b

Discussion
Recently, targeted anticancer therapies have become a powerful tool in medical practice due to increased efficacy and decreased side effects. Antibody-drug conjugates (ADCs) are effective in destroying cancer cells and are highly specific to selected targets. However, they are burdened with some disadvantages, including their large size that impairs tumor tissue penetration, a complicated

Discussion
Recently, targeted anticancer therapies have become a powerful tool in medical practice due to increased efficacy and decreased side effects. Antibody-drug conjugates (ADCs) are effective in destroying cancer cells and are highly specific to selected targets. However, they are burdened with some disadvantages, including their large size that impairs tumor tissue penetration, a complicated structure that hampers their production in bacterial cells, and an extensive protection of intellectual property [13,14]. A promising approach to achieve a comparably selective treatment is targeted drug therapy based on the conjugates of an anticancer drug with non-antibody scaffolds [15,43].
Affibody is an example of a potential non-antibody drug carrier [41,[44][45][46]. This small protein has been engineered to bind to a large number of target proteins or peptides with high affinity [29]. It is approximately 20 times smaller than antibodies and has a simple structure composed of three helices, which makes its production straightforward. Phage display has been used for the selection of affibodies interacting with particular molecular targets. Moreover, according to clinical studies, affibodies are non-immunogenic and well tolerated by patients [47,48]. One of the most commonly investigated affibodies is the Z HER2:4 affibody, which binds HER2. This protein kinase receptor is overexpressed in certain aggressive types of breast cancer and is one of the most common targets in cancer therapies [49,50].
We employed the (Z HER2:4 ) 2 DCS diaffibody, which is composed of two identical affibody sequences separated by a glutamic acid residue, the N-terminal His-tag, and the C-terminal drug conjugation sequence (DCS) as a drug-targeting molecule (Figure 1). Notably, a very similar construct, referred to as (Z HER2:4 ) 2 , was previously generated in order to improve the affinity and pharmacokinetics of the monovalent affibody [36,37]. (Z HER2:4 ) 2 DCS was overexpressed in the bacterial E. coli Bl21CodonPlus (DE3)-RIL strain and purified using an Ni-NTA column. The characterization of this protein by circular dichroism showed that the (Z HER2:4 ) 2 DCS diaffibody retains the α-helical structure of the parental protein ( Figure 2). According to the denaturation experiments, the dimeric anti-HER2 affibody is a stable protein with a thermal denaturation temperature of 57 • C in a phosphate buffer (Figure 3).
The specificity of (Z HER2:4 ) 2 DCS binding to HER2 present on cancer cells was determined by flow cytometry (Figure 4). This method confirmed that (Z HER2:4 ) 2 DCS labeled with FITC specifically recognizes the SK-BR-3 HER2-positive cells. As expected, the HER2-negative U-87 MG cells were not detected by the construct. (Z HER2:4 ) 2 DCS was conjugated with monomethyl auristatin E (vcMMAE), resulting in the (Z HER2:4 ) 2 DCS-MMAE conjugate, which was confirmed by MS analysis and SDS-PAGE ( Figure 5). The affinity of (Z HER2:4 ) 2 DCS and (Z HER2:4 ) 2 DCS-MMAE for HER2 was analyzed using surface plasmon resonance (SPR) (Figure 6a,b). According to our analysis, both (Z HER2:4 ) 2 DCS and (Z HER2:4 ) 2 DCS-MMAE bind HER2 with an apparent affinity of about 18 nM. This indicates that the auristatin conjugation to the C-terminal drug conjugation sequence does not influence the strength of diaffibody binding to HER2. However, the K D obtained for our construct is six times higher than the K D calculated for (Z HER2:4 ) 2 (3 nM) [36]. This discrepancy may be caused by the presence of the drug conjugation sequence that enabled us to conjugate MMAE to (Z HER2:4 ) 2 .
Our initial microscopic observation indicated that (Z HER2:4 ) 2 DCS-MMAE destroys SK-BR-3 cells that strongly overexpress HER2 on their surface (Figure 7). This result was also confirmed by the real-time cell viability analysis that demonstrated that the response of SK-BR-3 cells to (Z HER2:4 ) 2 DCS-MMAE is dose and time dependent (Figure 8).
To demonstrate that the conjugate can selectively target HER2 overexpressing cancer cells, we decided to use a broad spectrum of cell lines, including MDA-MB-453 cells expressing high levels of HER2 (HER2+), T-47D expressing slightly elevated levels of HER2 (HER2+/−), and cell lines expressing low levels of HER2 (HER2−), including U-87 MG, SK-MES-1, and MDA-MB-231. Importantly, we previously reported that all these cell lines are similarly sensitive to free MMAE [41,51]. The cytotoxicity of (Z HER2:4 ) 2 DCS-MMAE was measured by the MTT proliferation assay (Figures 9 and 10). The MTT assay showed that the unconjugated (Z HER2:4 ) 2 DCS diaffibody had a minor negative effect on the viability of SK-BR-3 cells. This is consistent with the previous report by Ekerljung et al. that demonstrated that (Z HER2:4 ) 2 inhibits proliferation of SK-BR-3 cells [52]. In the case of the SK-BR-3 and MDA-MB-453 cells that overexpress HER2, we observed a cytotoxic effect of the (Z HER2:4 ) 2 DCS-MMAE conjugate, causing a 90% decrease in cell growth after 120 h of the experiment in the concentration range from 100 nM to 4 nM (Figure 9). The IC 50 values for the tested HER2-positive cells were within the low nanomolar range (from 0.5 nM for SK-BR-3 cells to 5.5 nM for T-47D cells). In contrast, the cytotoxicity of (Z HER2:4 ) 2 DCS-MMAE towards the HER2-negative cells was much lower than observed for the HER2-positive cells ( Figure 10). This effect may be caused by a HER2-mediated endocytosis of the diaffibody-auristatin conjugate followed by auristatin release in endosomes, which leads to cell death. However, we cannot rule out the possibility that extracellular or cell surface associated proteases cleave the valine-citrulline linker and the released drug penetrates into cells. Therefore, the stability of the diaffibody-auristatin conjugate should be further addressed.
Importantly, the potency of the presented conjugate towards HER2-positive cells was comparable with ADC that has been successfully developed [53,54], and it can be further improved by obtaining homogeneous species of (Z HER2:4 ) 2 DCS loaded with three MMAE molecules. Overall, the cytotoxic conjugate based on the anti-HER2 diaffibody and MMAE efficiently destroys HER2-positive cancer cells in vitro.

Cell Lines
The

Protein Expression and Purification
The anti-HER2 diaffibody was designed on the basis of the anti-HER2 affibody Z HER2:4 sequence [28]. The amino acid sequence of Z HER2:4 affibody was duplicated in a head-to-tail configuration and separated with a glutamic acid residue linker. An N-terminal hexahistidyl (6× His) tag was added, allowing purification by immobilized metal ion affinity chromatography (IMAC), as well as the C-terminal drug conjugation site sequence (DCS), containing three cysteine residues (CAACAAAC). The constructed gene was cloned into the pET-30a vector (Novagen, Merck, Darmstadt, Germany). The protein was expressed in the E. coli Bl21CodonPlus (DE3)-RIL strain and purified using IMAC chromatography on Ni-NTA agarose (Qiagen, Hilden, Germany). The cell lysate was loaded onto an Ni-NTA agarose column equilibrated with 50 mM NaHPO 4 , 300 mM NaCl, and 10 mM imidazole. The column was then extensively washed with 50 mM NaHPO 4 , 300 mM NaCl, and 20 mM imidazole buffer, and the bound proteins were eluted with 50 mM NaH 2 PO 4 , 300 mM NaCl, and 250 mM imidazole buffer. Fractions containing diaffibody were pooled together and dialyzed overnight against PBS with 0.1 mM sucrose, 10% glycerol, and 0.1 mM TCEP. The purity of the diaffibody sample was analyzed by SDS-PAGE in reducing conditions. Protein bands were visualized by Instant Blue staining. Protein concentration was calculated from absorbance measurements at 280 nm, using the appropriate extinction coefficient (30,940 M −1 ·cm −1 ) and the molecular mass was verified by mass spectrometry analysis (MALDI TOF/TOF 4800, Applied Biosystems, Waltham, MA, USA).

Circular Dichroism
To verify the native conformation of the purified diaffibody, circular dichroism measurements were performed (J-715 spectropolarimeter, Jasco, Tokyo, Japan). Far-UV CD spectra in the range of 260-200 nm were acquired at 21 • C using 1 µM protein concentration at 68.5 mM NaCl, 1.35 mM KCl, 5 mM Na 2 HPO 4 , 0.9 mM KH 2 PO 4 , pH 7.4 buffer, and a 1 cm path length quartz cuvette. CD spectra were averaged over three scans and then converted to mean residue ellipticity [θ]. The secondary structure content was analyzed using tools available online; DichroWeb Server [55] with SELCON3 [35] and K2D algorithms and CDpro (Colorado State University, Fort Collins, CO, USA) software [36] using CDSSTR, SELCON3, and CONTIN/LL algorithms with SP43, SDP48, and SMP56 reference sets. Thermal denaturation experiments were conducted by monitoring changes in the CD signal at 222 nm between 20 • C and 80 • C at a constant rate of 1 • C/min during the denaturation and renaturation processes. Denaturation data were analyzed using PeakFit software (Systat Software, San Jose, CA, USA), assuming a two-state reversible equilibrium transition as described previously [56].

Affinity Measurements
The recombinant extracellular domain of HER2 (Sino Biological, Beijing, China) was diluted in acetate buffer with pH 6.0 and immobilized (1000 resonance units (RU)) onto the surface of a CM5 sensor chip by standard amine coupling. The binding analysis was carried out with 0.01, 0.1, and 1 µM (Z HER2:4 ) 2 DCS and (Z HER2:4 ) 2 DCS-MMAE diluted in 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.02% sodium azide, with pH 7.4, using a BIAcore 3000 instrument (GE Healthcare, Chicago, IL, USA). The running conditions were 10 µL/min flow rate, 25 • C, a 3 min association time, and a 2 min dissociation time. Following dissociation, the chip was regenerated with 50 mM NaOH as the regeneration buffer. All the buffers were filtered and degassed prior to each experiment. The dissociation equilibrium constant (K D ) was determined using BIA evaluation 3.2 software (Biacore, GE Healthcare, Chicago, IL, USA), assuming one-to-one binding.

Fluorescein Labeling
Prior to coupling to 5-iodoacetamidofluorescein (5-IAF) (Thermo Fisher Scientific, Waltham, MA, USA), (Z HER2:4 ) 2 DCS was reduced with 1 µM TCEP for 1 h at room temperature. 5-IAF was added to a final concentration of 10 times molar excess over each of three cysteines in the protein. The sample was incubated for 2 h with shaking. After incubation, the sample was dialyzed against PBS overnight. The efficiency of the reaction was confirmed by mass spectrometry analysis. The conjugation product is referred to as (Z HER2:4 ) 2 DCS-FITC.

Flow Cytometry
The U-87 MG and SK-BR-3 cells were trypsinized, washed three times with PBS, and incubated for 1 h with (Z HER2:4 ) 2 DCS-FITC at the concentrations of 0.03, 0.3, and 3.0 µM. As a control, the cells were incubated for 1 h with mouse monoclonal anti HER-2 antibody (AbD Serotec, Bio-Rad, Hercules, CA, USA), washed three times with PBS, and again incubated for 1 h with FITC-conjugated donkey anti-mouse antibody (Jackson Immuno Research, West Grove, PA, USA). Cells were washed three times with PBS and fixed with 1% paraformaldehyde, following their analysis on a FACSDiva instrument (BD Biosciences, Franklin Lakes, NJ, USA). The obtained results were analyzed using WinMidi software (Purdue University, West Lafayette, IN, USA).
The sample was incubated for 2 h with shaking. After incubation, the sample was dialyzed against PBS and the product of the reaction was analyzed by mass spectrometry.

Cytotoxicity Measurements
The cytotoxicity of the (Z HER2:4 ) 2 DCS-MMAE conjugate was measured using the MTT proliferation assay and xCELLigence impedance-based, label-free, real-time cell analyzer (ACEA Biosciences, San Diego, CA, USA). For the MTT assay, cells were seeded on 96-well plates at 7500 cells per well in 100 µL of medium and were cultured for 24 h. (Z HER2:4 ) 2 DCS-MMAE at final concentrations of 100, 20, 4, and 0.8 nM was added to the cell cultures and incubated for 72, 96, and 120 h. After incubation, 20 µL of the MTT reagent was added to each well and the plates were incubated at 37 • C. Following 4 h incubation, the cells were lysed by the addition of 80 µL of lysis buffer (45% dimethylformamide (DMF), 13.5% sodium dodecyl sulfate (SDS)). The fluorescence was measured at 590 nm with a reference filter of 620 nm using the Envision Multimode Plate Reader (PerkinElmer, Waltham, MA, USA). For the xCELLigence assay, the cells were seeded in plates at 10,000 cells per well in the appropriate medium and cultured for 24 h in standard conditions. The (Z HER2:4 ) 2 DCS-MMAE conjugate at final concentrations of 100, 20, 4, and 0.8 nM was added to the cell cultures and cell viability was monitored for 120 h with 1 h intervals. Cells were also analyzed using a light microscope.