1. Introduction
In recent years, delivery of payloads such as toxins, radionuclides, and cytotoxic drugs to cancer cells by affinity proteins have proven to be an efficient route for targeted cancer therapy. Such constructs have become promising agents for the treatment of disseminated cancers [
1]. One of the well-studied receptors for targeted cancer therapy is the human epidermal growth factor receptor 2 (HER2), which belongs to the tyrosine kinase receptor family. It is overexpressed in a number of breast, ovarian, lung, and gastrointestinal cancers [
2,
3]. The differential high-level overexpression on cancer cells and internalizing character has made HER2 one of the particularly suitable receptors to target for delivery of cytotoxic drugs [
4].
Drug conjugates typically consist of a protein-based carrier with affinity for a particular molecular abnormality of a malignant cell, and a cytotoxic drug, which is connected by a linker. The most explored type of drug conjugates are antibody drug conjugates (ADCs), where the drug-carrier is a target-specific monoclonal antibody (mAb). ADCs combine several therapeutic modes of action, targeted therapy by the cytotoxic drug, sometimes direct blocking of the normal function of its target, and immunotherapy by the mAb, which invokes immunological reactions (often including an ADCC response). Over the last decade, ADCs have emerged as highly potent anti-cancer drugs for clinical use [
5].
Mertansine (DM1) is a commonly used cytotoxic drug that induces cell death by inhibiting polymerization of tubulin and is thus highly effective against rapidly proliferating cells. This is a desirable property of DM1 as it minimizes damage to normal cells with a slower rate of proliferation. Trastuzumab emtansine (T-DM1) is an ADC consisting of DM1 linked to the anti-HER2 mAb trastuzumab. T-DM1 was approved by the US Food and Drug Administration (FDA) in 2013 for the treatment of HER2 positive breast cancer. Patients with acquired resistance to trastuzumab or taxane therapy have been shown to benefit from T-DM1 treatment [
6].
Despite the benefits of ADCs in terms of safe delivery of the cytotoxic drug, there are sub-optimal properties of many ADCs in pre-clinical and clinical development, which limits their efficiency. These include a variability of the drug to antibody ratio (DAR) due to random conjugation, inefficient penetration of solid tumors due to the rather large size of the mAb, and possible alteration of the binding properties of the mAb due to random conjugation (potentially to sites involved in antigen binding) [
7]. Studies concerning site-specific attachment of drugs to mAbs have been published in recent years [
8], but these methodologies are not yet commonly used in FDA approved ADCs.
The use of small engineered scaffold proteins (ESPs) as targeting agents instead of mAbs might be a viable route to overcome some of the inherent problems of random drug attachment. For example, if the ESP does not contain cysteines in the framework, it is possible to introduce one or more cysteines at desired positions in the molecule for site-specific conjugation of the drug. Also, the small size of most ESPs should provide a better penetration of solid tumors and consequently enhance the drug conjugate’s therapeutic effect [
9,
10,
11]. In the case of HER2, clinical imaging studies have shown that different classes of ESPs (e.g., affibody molecules, ADAPTs, DARPins) specifically and efficiently accumulate in tumors expressing the receptor, which suggest that they can potentially be suitable carriers of cytotoxic drugs for cancer therapy [
6,
12,
13,
14].
Albumin binding domain (ABD)-derived affinity proteins (ADAPTs) constitute a class of ESPs which have been used for radionuclide targeting applications [
15]. ADAPTs are small (5 kDa), folded domains, derived from the G148-GA3 albumin binding domain of streptococcal protein G [
16,
17,
18]. From combinatorial libraries where surface amino acids have been randomized, binders against desired targets can be generated [
16]. ADAPT6 is a specific binder to HER2 with an equilibrium dissociation constant of 2 nM [
17]. We have previously evaluated different aspects of molecular design of ADAPT6 for radionuclide molecular imaging and have found that ADAPT6 preserves binding to the HER2 target, even after chemical modification such as conjugation with different chelators [
19,
20,
21,
22,
23,
24]. Moreover, a clinical study demonstrated that injections of
99mTc-labelled ADAPT6 were safe, and that ADAPT6 accumulated in HER2-expressing breast cancer with high specificity [
14].
Previously, we have found that it is possible to extend the plasma half-life of ESPs by fusion to an albumin binding domain (ABD), which consequently should improve its bioavailability [
25,
26,
27,
28]. Half-life extension occurs by binding of the ABD to serum albumin in the blood, leading to an increase in the overall size of the complex that exceeds the filtration cut-off of the kidneys (approximately 60 kDa). A particularly useful ABD variant is ABD
035 with femtomolar binding affinity (K
D) to human serum albumin (HSA) [
26]. Moreover, we have demonstrated that an ABD-fused ADAPT6 labeled with
177Lu can be used to efficiently target human cancer xenografts with high HER2 expression in mice [
28]. It was also found that the relative position of ABD and ADAPT6 influences the biodistribution, and that placement of the ABD at the C-terminus of ADAPT6 was particularly favorable [
28]. We have therefore used an analogous architecture for the drug conjugates in this study, with ADAPT6 followed by the ABD.
A high hepatic uptake of cytotoxic drug conjugates might cause severe drug induced liver injury (DILI), and hepatic uptake should therefore be minimized [
29]. Lipophilic patches in proteins are a main driver of hepatic uptake. Previous studies by our groups concerning ABD-fused affibody molecules conjugated with mcDM1 suggested that the use of multiple hydrophilic amino acids as a spacer between the protein part and mcDM1 reduced the lipophilicity of the drug conjugates, and consequently liver uptake, while tumor uptake and potency remained intact [
10,
11,
30].
Given the excellent properties of radiolabeled ADAPTs as probes for molecular imaging, we have in this study sought to investigate their properties to carry cytotoxic drugs to tumor cells. The HER2-targeting ADAPT6 was expressed as a fusion protein with ABD035. A C-terminal cysteine residue was incorporated into the construct, which was used to conjugate DM1 via a non-cleavable maleimidocaproyl (mc) linker, resulting in the drug conjugate ADAPT6-ABD-mcDM1. A non-toxic control protein was also created, ADAPT6-ABD-AA, where the C-terminal cysteine residue was capped with iodoacetamide. A non-targeted control drug conjugate was created, ADAPTNeg-ABD-mcDM1, which included ADAPTNeg, not interacting with any target, fused to the ABD and derivatized with mcDM1 on the C-terminal cysteine residue. In all variants, a tag was placed at the N-terminus, with the amino acid sequence His-Glu-His-Glu-His-Glu to be used as a chelator for labelling with 99mTc(CO)3. The linker connecting ADAPT6 and the ABD had the amino acid sequence (Ser-Ser-Ser-Gly)3. Since our previous studies on affibody molecules have shown that the use of multiple hydrophilic amino acids as a spacer between the protein part and mcDM1 reduced liver uptake, we introduced a Glu-Glu-Glu spacer between the ABD and the C-terminal cysteine onto which mcDM1 was conjugated. The biochemical properties and cytotoxic potential of the ADAPT-based drug conjugates was investigated. The biodistribution was further evaluated in nude mice bearing HER2-overexpressing SKOV-3 tumors.
2. Materials and Methods
2.1. General
All chemicals were obtained from Sigma-Aldrich (St. Louis, MO, USA) or Merck (Darmstadt, Germany). Restriction enzymes were purchased from New England Biolabs (Ipswitch, MA, USA). Radioactivity was measured by an automated γ-spectrometer with a NaI(Tl) detector (1480 Wizard; Wallac Oy, Turku, Finland).
2.2. Production and Purification of ADAPT Fusion Proteins
Genes encoding the HER2-binding ADAPT6-ABD and the non-targeted ADAPT
Neg-ABD fusion proteins were synthesized by Thermo Fisher Scientific (Waltham, MA, USA). The genes were PCR amplified, adding nucleotides encoding the N-terminal amino acid sequence MHEHEHEDANS and the C-terminal amino acid sequence EEEC as well as sites recognized by the restriction enzymes NcoI and AscI. The PCR-products were sub-cloned into the expression vector pET-21a(+) by restriction with NcoI and AscI followed by ligation. The proteins were expressed in
Escherichia coli BL21 (DE3) star cells at 25 °C for 24 h after induction by isopropyl β-D-1-thiogalactopyranoside to a final concentration of 1 mM. The intracellular fraction was released by sonication. The proteins were purified by affinity chromatography on a column with immobilized human serum albumin (HSA) as previously described [
28]. Briefly, the cell lysates were loaded on the column after equilibration with 1xTST buffer (25 mM Tris-HCl, 1 mM EDTA, 0.2 M NaCl, 0.05% Tween 20, pH 8.0), followed by washing with 1xTST and 5 mM NH
4Ac (pH 5.5). Bound proteins were eluted with 0.5 M acetic acid (pH 2.8) and were lyophilized.
2.3. Conjugation with mcDM1
The ADAPT fusion proteins were conjugated with mcDM1 (Levena Biopharma, San Diego, CA, USA) at a molar ratio of 3:1 (drug:protein) through coupling of the maleimide group in the mc-linker with the free thiol group of the C-terminal cysteine in the fusion proteins. Before conjugation, the lyophilized fusion proteins were dissolved in 10 mM Tris-HCl buffer (pH 7.85). Potentially oxidized thiol groups were reduced by tris (2-carboxyethyl) phosphine (TCEP) at a final concentration of 5 mM for 30 min at 37 °C. The pH was adjusted to 6.5 using HCl. The fusion proteins (5 mg/mL) were mixed with mcDM1 and were incubated at 25 °C for 1 h. The non-toxic control ADAPT6-ABD-AA was produced by alkylation of the C-terminal cysteine with 2-iodoacetamide. ADAPT6-ABD was dissolved in alkylation buffer (0.2 M NH4HCO3, pH 8.0) and was incubated with TCEP as described above, to reduce potentially oxidized cysteines. 2-iodoacetamide was added to a final concentration of 10 mM followed by incubation at room temperature for 30 min in the dark. Both ADAPT drug conjugates (ADAPT-DCs) and the non-toxic control were purified through RP-HPLC on an Agilent 1260 series Infinity II machine (Agilent, Santa Clara, CA, USA). The reaction mixtures after conjugation were diluted 1:1 with buffer A (0.1% trifluoroacetic acid in water) and was loaded on a Zorbax 300SB-C18 semi-preparative column (Agilent). Bound constructs were eluted using a gradient of 30–70% of buffer B (0.1% trifluoroacetic acid in acetonitrile) over 30 min using a flow rate of 3 mL/min. The fractions containing the constructs of interest were pooled, lyophilized and were stored at −80 °C until use.
2.4. Biochemical Characterization
The identity of ADAPT-DCs and the non-toxic control were analyzed by SDS-PAGE (Biorad, Hercules, CA, USA) under reducing conditions. The oligomeric state of the constructs was analyzed by size exclusion chromatography, by passage through a 5/150 column (GE Healthcare, Uppsala, Sweden), packed with Superdex-75, using a flow rate of 0.45 mL/min in PBS. The molecular weight of purified ADAPT variants was measured by liquid chromatography-electro-spray ionization-mass spectrometry (LC-ESI-MS) using an Impact II UHR QqTOF MS (Bruker Daltonics, Billerica, MA, USA). The purity of the constructs was evaluated by RP-HPLC using a Zorbax 300SB-C18 column (Agilent) using a gradient from 30–60% of buffer B over 20 min at a flow rate of 1 mL/min.
2.5. Biosensor Analysis on a Biacore Instrument
Target binding analysis was performed on a Biacore 3000 instrument (GE Healthcare) by injecting samples over a CM5 chip. The chip was prepared by immobilizing murine serum albumin (MSA) (Sigma-Aldrich) to a response level of ~500 RU in one channel, human serum albumin (HSA) (Novozymes, Bagsvaerd, Denmark) to ~400 RU in a second channel, and HER2-Fc fusion protein (R&D Systems, Minneapolis, MN, USA) to ~1000 RU in a third channel. A blank surface for normalization was also created in a fourth channel by activation and deactivation. The experiments were preformed essentially as described previously [
31]. Briefly, a series of five different concentrations of each analyte in 1xPBST (PBS supplemented with 0.05% Tween 20, pH 7.4) was sequentially injected over the flow-cells. The injections were carried out with a flow rate of 30 μL/min, for 300 s for binding, followed by 1200 s for dissociation of the analytes.
2.6. Cell Culture
The cancer cell lines, AU565, SKBR3, SKOV-3, A549, and MCF7 were purchased from American Type Culture Collection (ATCC, via LGC Promochem, Borås, Sweden) and were grown in RPMI-1640 (SKOV-3, SKBR3, AU565), or Dulbecco’s modified Eagle medium (A549, MCF7) (Cytiva Hyclone, Uppsala, Sweden) supplemented with 10% FBS (Sigma-Aldrich) in a humidified incubator at 37 °C in a 5% CO2 atmosphere.
2.7. In Vitro Cytotoxicity Analysis
AU565, SKBR3 A549, and MCF7 (5000 cells/well) or SKOV-3 (2000 cells/well) were seeded in 96-well plates and cultured overnight. On the next morning, the medium was aspirated and increasing concentrations of ADAPT-DCs or controls were prepared in four replicates, by dilution in a fresh media, and were added to the wells. The plates were incubated for 72 h in a humidified incubator at 37 °C in a 5% CO2 atmosphere. Cell viability was determined using a Cell Counting Kit-8 (CCK-8, Sigma-Aldrich) according to the manufacturer’s protocol. The obtained data were analyzed using nonlinear regression in Prism (GraphPad Software, San Diego, CA, USA).
2.8. Radiolabeling
99mTcO4− was obtained by elution from a 99Mo/99mTc generator (Mallinckrodt, Petten, The Netherlands) with sterile 0.9% NaCl.
Site-specific radiolabeling of ADAPT-DCs and ADAPT6-ABD-AA with [
99mTc(CO)
3]
+ was performed as described earlier [
21,
32,
33]. A CRS kit (PSI, Villigen, Switzerland) was used to produce [
99mTc(CO)
3(H
2O)
3]
+ from
99mTcO
4− for labeling of the N-terminal His-Glu-His-Glu-His-Glu sequence in the constructs with
99mTc. The labeled compounds were purified using NAP-5 columns (GE Healthcare). The radiochemical yield and radiochemical purity were determined by radio iTLC using iTLC silica gel strips (Varian, Lake Forest, CA, USA) with subsequent measurement using a Cyclone phosphor system (PerkinElmer, Waltham, MA, USA). Additionally, the radio-iTLC data was cross-validated by high performance liquid chromatography (HPLC) analysis on a Hitachi Chromaster HPLC system with radioactivity detector and a Luna RP C18 column at room temperature. The samples were diluted in buffer A (0.1% trifluoroacetic acid in water) and loaded onto the column. The samples were eluted using a gradient of 5–80% of buffer B (0.1% trifluoroacetic acid in acetonitrile), using a flow rate of 1 mL/min over 20 min.
2.9. In Vitro Characterization of the 99mTc-Labeled Constructs
In vitro specificity tests were performed according to a method described by Wållberg et al. [
31]. Briefly, 1 × 10
6 SKOV-3 or AU565 cells were seeded in six-well plates on the day before the experiment. The dishes were divided into two sets. In the binding sets, radiolabeled constructs were added to reach a concentration of 50 nM. In the control sets, HER2 receptors on the cells were saturated with 50 mM of non-radiolabeled proteins 15 min before adding the labeled compounds. After incubation for 1 h at 37 °C, cells were washed, and the cell-associated radioactivity was measured using a γ-spectrometer.
Analysis of cellular processing was performed according to a method described by Wållberg et al. [
31]. In brief, 1 × 10
6 SKOV-3 and AU565 cells were seeded in six-well plates on the day before the experiment. Cells were incubated at 37 °C with 50 nM of radiolabeled constructs. At 1, 2, 4, 8, and 24 h after addition of the constructs, media was aspirated and the membrane-bound fraction was released by acid wash using 4 M urea in 0.2 M glycine buffer, pH 2.0. The internalized fraction was released by incubating with 1 M NaOH. The radioactivity of the fractions was measured by a γ-spectrometer. Prism (version 8.00 for Windows, GraphPad Software) was used to analyze cellular processing using an unpaired two-tailed
t-test.
The binding affinity of the radiolabeled constructs to HER2-expressing SKOV-3 cells was measured using a LigandTracer biosensor (Ridgeview Instruments, Uppsala, Sweden) based on a method described previously [
34].
2.10. Biodistribution in Tumor Bearing Mice
Animal studies were planned in agreement with EU Directive 2010/63/EU for animal experiments and Swedish national legislation concerning the protection of laboratory animals and were approved by the Ethics Committee for Animal Research in Uppsala, Sweden (Permit Number: C4/2016).
Biodistribution and targeting properties of the labeled compounds were evaluated in female BALB/c nu/nu mice bearing HER2-positive SKOV-3 xenografts. The xenografts were established by subcutaneously implanting 1 × 107 SKOV-3 cells in a hind leg, three weeks before the experiment.
The mice were randomized into six groups with four mice per group. Each mouse received a tail vein injection of 6 μg of 99mTc-ADAPT6-ABD-mcDM1 in 100 μL PBS containing 1% BSA and the biodistribution was measured 1, 4, 24, and 48 h after injection. The injected activities were 60 kBq/mouse for mice dissected 1 and 4 h after injection, 640 kBq/mouse for mice dissected 24 h after injection, and 10.2 MBq for mice dissected 48 h after injection. After exsanguination under anesthesia (25 μL/g body weight; ketamine 10 mg/mL; Rompun 1 mg/mL), the organs and tissues of interest were excised, weighed, and their activity was measured using an automated γ-spectrometer.
To confirm the specificity of 99mTc-ADAPT6-ABD-mcDM1 accumulation in the tumors, two control experiments were performed. One group of four mice was intravenously injected with 99mTc-ADAPTNeg-ABD-mcDM1 (6 µg, 640 kBq), and the biodistribution was measured 24 h after injection. Another group of four mice was subcutaneously injected with trastuzumab (8.4 mg/per mouse) to block HER2 receptors. 24 h after injection of trastuzumab, the mice were injected with 99mTc-ADAPT6-ABD-mcDM1 (6 µg, 640 kBq), and the biodistribution was measured an additional 24 h later.
For confirmation of the biodistribution results obtained by the ex vivo measurements, a SPECT/CT imaging was performed. Two mice were injected with
99mTc-ADAPT
6-ABD-mcDM1 (6 µg, 30 MBq). One mouse was pre-injected 24 h before injection of
99mTc-ADAPT6-ABD-mcDM1 with 8.4 mg trastuzumab. Imaging of the mice was performed 24 h after injection of radiolabeled ADAPT6-ABD-mcDM1 using a nanoScan SPECT/CT scanner (Mediso Medical Imaging Systems, Budapest, Hungary). Immediately before being placed in the camera, the mice were euthanized by CO
2 asphyxiation. The acquisition time was 60 min. The camera settings were as described previously [
35].
4. Discussion
Engineered scaffold proteins (ESPs) capable of specific delivery of cytotoxic compounds to cancer cells have the potential to become agents for targeted cancer therapy in the future. However, different scaffold proteins have different structures and surface amino acids, which influence their distribution properties and off-target interactions. This necessitates a careful evaluation of every type of ESP for their suitability as targeting agent. The results of this study suggest that the relatively unexplored ADAPT-class of ESPs might be suitable for targeted delivery of anti-cancer drugs.
Recombinant expression and purification of ADAPT6-ABD, as well as conjugation of mcDM1, was straightforward. After successful purification and conjugation, the drug conjugate could specifically bind to, get internalized, and efficiently kill cells with high HER2 expression (
Figure 3 and
Figure 4). A difference in cytotoxic potency was noted among the high-HER2 expressing cell lines. ADAPT6-ABD-mcDM1 showed efficient killing of the breast cancer cell lines SKBR3 and AU565, but the ovarian cancer cell line SKOV-3 was considerably more resistant to DM1 treatment. Previous studies of ADCs have demonstrated different responses in different cell lines with a high HER2 expression level [
36,
37,
38]. It was suggested that the differences might be related to differences in the internalization rate of the receptor, expression level of multi-drug resistance proteins, and differences in the efficiency of lysosomal degradation. In our previous studies with anti-HER2 affibody molecules, conjugated with drugs and toxins, the SKOV-3 cell line was also more resistant to targeted treatment in vitro than other cell lines with a similarly high HER2 expression level [
11,
39].
ADAPT6-ABD-mcDM1, the non-toxic control ADAPT6-ABD-AA, and the non-target control ADAPT
Neg-ABD-mcDM1 could be stably radiolabeled with
99mTc (
Table S2). The ADAPT6-containing constructs showed high-affinity binding to living cells with high HER2 expression when analyzed with a LigandTracer biosensor, however the binding was characterized by two interactions, one with stronger and one with weaker affinity (
Figure 6). The association rates of both interactions were similar, but the dissociation rates differed. Such types of multiple interactions could be due to the fact that HER2 is present on the membranes of living cells in different states; as a monomer or in a homo- or heterodimeric form. Monomers and dimers may have slightly different conformations, which in turn may affect the interaction. Previous studies have also reported this phenomenon for different binders to HER2 [
31,
40,
41] and other HER family receptors [
42]. The equilibrium dissociation constant for the strong interaction with SKOV-3 cells in the LigandTracer instrument was similar to the equilibrium dissociation constant for the interaction with the extracellular domain of HER2 in the Biacore experiment (
Table 1).
The internalization rate of
99mTc-ADAPT6-ABD-mcDM1 and
99mTc-ADAPT6-ABD-AA was 20–38% of the cell-associated radioactivity and data were similar or slightly lower compared to those previously reported for the similar, affibody-based, drug conjugate
99mTc-Z
HER2:2891-ABD-mcDM1 (30–40%) [
11]. Since internalization is critical for drug action, therapy using ADAPT6 conjugate might have a similar efficacy compared to the affibody variant.
The radiolabeled
99mTc-ADAPT6-ABD-mcDM1 drug conjugate was further evaluated in nude mice bearing SKOV-3 xenografts. The fusion with ABD had appreciably expanded the residence of the construct in circulation, with a blood concentration of 11 ± 2% ID/g at 4 h after injection. It was significantly longer compared to the blood concentration of non-ABD fused ADAPT6 (
99mTc-(HE)
3-ADAPT6), which was only 0.31 ± 0.05% ID/g at the same time point in a previous study [
21]. The targeting specificity of
99mTc-ADAPT6-ABD-mcDM1 was confirmed in two ways, by saturation of binding sites on HER2 using trastuzumab, which appreciably lowered tumor uptake, and by comparison with the tumor uptake of the non-targeted control,
99mTc-ADAPT
Neg-ABD-mcDM1 (
Figure 7), which was appreciably lower than the uptake of
99mTc-ADAPT6-ABD-mcDM1. The results of these tests clearly demonstrated HER2-specific accumulation of ADAPT6-ABD-mcDM1 in the tumors. The biodistribution results showed that the uptake in the tumor increased up to 24 h p.i. where it reached a plateau of 5% ID/g. This was lower compared to the homologous ADAPT6-ABD labeled at the C-terminus with
177Lu,
177Lu-DOTA-ABD
035-ADAPT6, which has been evaluated in a similar mouse model with SKOV-3 tumor xenografts [
28]. In that study, the tumor uptake increased up to 26 ± 4% ID/g at 24 h. The higher tumor uptake of
177Lu-labeled construct is likely a consequence of its higher blood radioactivity, 17 ± 2% IA/g at 24 h, and thus higher bioavailability compared to
99mTc-ADAPT6-ABD-mcDM1, with a blood radioactivity of 5.0% ID/g at 24 h. At 24 h, the uptake in liver of
177Lu-DOTA-ABD
035-ADAPT6 (4.8 ± 0.3% ID/g) and technetium-99 labeled ADAPT6-ABD-mcDM1 (4.83 ± 0.27% ID/g) were similar, but the uptake in the kidneys was significantly lower for
177Lu-DOTA-ABD
035-ADAPT6 (10.9 ± 0.7% ID/g) than for ADAPT6-ABD-mcDM1 (83 ± 4% ID/g). This strongly suggests a low and similar clearance rate of both constructs through the liver, but a more rapid clearance of
99mTc-ADAPT6-ABD-mcDM1 through the kidneys. The Biacore experiments with ADAPT6-ABD-mcDM1 confirmed reasonably strong binding to serum albumins. However, DM1 still appears to influence the binding properties of ABD, since the blood clearance is faster and kidney uptake is higher compared to the
177Lu-labeled construct. It would be interesting to investigate this phenomenon further in future studies, and possibly to compare the ABD-technology with other methods of plasma half-life extension.
Another class of ESPs are the affibody molecules (58 amino acids), which are slightly larger than ADAPTs (46 amino acids), and similarly folded into three-helix bundle domains. An affibody molecule with strong and specific affinity for HER2 has previously been expressed as a fusion to ABD
035 and derivatized with mcDM1 to Z
HER2-ABD-mcDM1 [
11], a conjugate analogous to ADAPT6-ABD-mcDM1. In the present study, the cytotoxic potential (IC
50 value) of ADAPT6-ABD-mcDM1 were 5 nM, 7 nM, and 80 nM for SKBR3, AU565 and SKOV-3 cells, respectively. In a similarly performed experiment, the IC
50 values for Z
HER2-ABD-mcDM1 was 0.6 nM, 1 nM, and 33 nM towards the same cell lines. A partial explanation for the weaker cytotoxic potential of ADAPT6-ABD-mcDM is most likely its weaker affinity for HER2 (6 nM) compared to the affinity of Z
HER2-ABD-mcDM1 for HER2 (0.7 nM) [
11]. Further, the cytotoxic potential of T-DM1 towards the same cell lines has previously been determined to 0.2 nM, 0.2 nM, and 0.5 nM, respectively [
10], which is stronger than ADAPT6-ABD-mcDM1, particularly for the SKOV3 cell line. In this case, the cytotoxic potential appears to be cell line dependent since the relative difference between ADAPT6-ABD-mcDM1 and T-DM1 is not the same for the three cell lines. It should also be noted that an antibody and an ADAPT are quite different in terms of biochemical behavior, and the linker connecting DM1 differs, which makes it difficult to draw conclusions concerning the reason for the difference in cytotoxic potential between the two.
The biodistribution of technetium-99 labeled ZHER2-ABD-mcDM1 showed a similar pattern to 99mTc-ADAPT6-ABD-mcDM1, which was characterized by a low unspecific uptake in normal organs except for the kidneys. 99mTc-ZHER2-ABD-mcDM1 had a slightly higher blood retention at 24 h (6.7 ± 0.3 versus 3.9 ± 0.5% ID/g), and a slightly longer plasma half-life of 12.5 h compared to 9.0 h for 99mTc-ADAPT6-ABD-mcDM1. The uptake of both constructs in the tumor was similar, 6.7 ± 0.3% ID/g for 99mTc-ZHER2-ABD-mcDM1 compared to 5.5 ± 1.8% ID/g for 99mTc-ADAPT6-ABD-mcDM1. Interestingly, 99mTc-ZHER2-ABD-mcDM1 also had high renal uptake. However, it could successfully be used for experimental therapy of SKOV-3 xenografts in mice, and a pathology investigation did not reveal any renal damage. Most likely, the explanation can be attributed to the mode of action of DM1, which is inhibition of tubulin polymerization and it should thus be more toxic for rapidly dividing cells. Apparently, the proliferation rate of proximal tubuli cells in the kidneys is slower, which permits accumulation of an appreciable amount of DM1 without any toxic effect. The same effect could, hopefully, be expected for ADAPT6-ABD-mcDM1.
In conclusion, we have shown that the relatively unexplored class of engineered scaffold proteins, the ADAPTs, can be fused with an ABD for extension of residence in circulation and site-specifically conjugated with the highly cytotoxic microtubulin inhibitor DM1. The resulting drug conjugate retained high affinity to HER2 and albumin. ADAPT6-ABD-mcDM1 was highly potent towards cells with high HER2 expression in vitro. 99mTc-ADAPT6-ABD-mcDM1 accumulated specifically in HER2-expressing human xenografts in vivo. Its biodistribution in mice was characterized by low uptake in normal organs, except the kidneys. Taken together, the results show that ADAPTs are potentially suitable carriers of cytotoxic drugs to malignant tumors.