Experimental HER2-Targeted Therapy Using ADAPT6-ABD-mcDM1 in Mice Bearing SKOV3 Ovarian Cancer Xenografts: Efficacy and Selection of Companion Imaging Counterpart

Overexpression of the human epidermal growth factor receptor 2 (HER2) in breast and gastric cancer is exploited for targeted therapy using monoclonal antibodies and antibody-drug conjugates. Small engineered scaffold proteins, such as the albumin binding domain (ABD) derived affinity proteins (ADAPTs), are a promising new format of targeting probes for development of drug conjugates with well-defined structure and tunable pharmacokinetics. Radiolabeled ADAPT6 has shown excellent tumor-targeting properties in clinical trials. Recently, we developed a drug conjugate based on the HER2-targeting ADAPT6 fused to an albumin binding domain (ABD) for increased bioavailability and conjugated to DM1 for cytotoxic action, designated as ADAPT6-ABD-mcDM1. In this study, we investigated the therapeutic efficacy of this conjugate in mice bearing HER2-expressing SKOV3 ovarian cancer xenografts. A secondary aim was to evaluate several formats of imaging probes for visualization of HER2 expression in tumors. Administration of ADAPT6-ABD-mcDM1 provided a significant delay of tumor growth and increased the median survival of the mice, in comparison with both a non-targeting homologous construct (ADAPTNeg-ABD-mcDM1) and the vehicle-treated groups, without inducing toxicity to liver or kidneys. Moreover, the evaluation of imaging probes showed that small scaffold proteins, such as 99mTc(CO)3-ADAPT6 or the affibody molecule 99mTc-ZHER2:41071, are well suited as diagnostic companions for potential stratification of patients for ADAPT6-ABD-mcDM1–based therapy.


Introduction
Targeted therapy using monoclonal antibodies (mAbs) is well established and extends the survival of patients with disseminated cancer. For example, the monoclonal antibody (mAb) trastuzumab, which targets human epidermal growth factor receptor 2 (HER2), has proven efficacy against HER2-positive breast and gastric cancer and is used routinely in clinics [1]. Furthermore, the anti-epidermal growth factor receptor (EGFR) mAb cetuximab, used for treatment of metastatic colorectal cancer, improves the overall survival of patients with wild-type K-ras tumors [2,3]. Despite these successful examples, the efficacy of mAbbased therapy is limited by the mode of action. Antibodies provide both cytostatic and off-target interactions, which in turn lead to differences in the uptake in normal organs and tissues in vivo. Expanding the targeting scaffold repertoire and investigating their structure-property relationship increases the chances to develop an optimal targeting agent for any given application.
We recently developed a HER2-targeting protein based on the ADAPT scaffold [22,38,39]. The ADAPT scaffold was developed from an albumin binding domain (ABD) from streptococcal protein G and is a small (5 kDa) three-helical bundle. Randomization of amino acids located in the helices one and three in ABD resulted in combinatorial libraries permitting selection of ADAPTs that could bind selectively to different targets while their binding to human serum albumin (HSA) by helix two was preserved [40]. This strategy was used to develop ADAPTs targeting tumor necrosis factor-α (TNF-α) [40], human epidermal growth factor receptor 3 (HER3) [41] and HER2 [42]. Later, the albumin-binding ability was removed to generate ADAPT6, targeting HER2 and having a short in vivo half-life [38]. ADAPT6 binds to the same epitope on HER2 as trastuzumab (subdomain IV) [37]. It has shown excellent tumor-targeting properties and has provided high imaging contrast in vivo, which enabled its clinical translation as a HER2-imaging probe. ADAPT6 labeled with technetium-99m was evaluated in a phase I clinical trial and was well tolerated by patients without any adverse reactions [21]. It provided excellent discrimination between HER2-positive and HER2-negative tumors in patients with primary breast cancer already at 2 h after injection. These results support the use of ADAPT6 as a promising protein scaffold for development of HER2-targeting therapeutics.
In a previous study, an ADAPT6-based drug conjugate was designed for targeted delivery of a cytotoxic drug to HER2-expressing tumor cells [39]. The ADAPT6-ABD-mcDM1 conjugate consists of ADAPT6 as the HER2-targeting domain, an ABD for extension of the half-life in blood circulation for increased bioavailability, and the cytotoxic drug DM1 (Figure 1). ADAPT6-ABD-mcDM1 demonstrated nanomolar affinity to HER2, a reasonably good internalization rate by HER2-expressing cancer cells lines, a specific HER2dependent toxicity in several cancer cell lines, and the capacity for specific accumulation in HER2-expressing human tumor xenografts in mice.

General
The chemicals used in the study were purchased from Sigma-Aldrich (Sweden AB Stockholm, Sweden). Buffers were prepared using high-quality Milli-Q water.

Cell Culture
The human ovarian cancer cell line SKOV3 was purchased from American Type Cul ture Collection (ATCC via LGC Promochem, Borås, Sweden) and cultured in Roswell Par As mentioned earlier, one of the hurdles for cancer therapy is the clonal selection. In disseminated cancer with multiple metastatic sites, the HER2 expression level in the primary tumor and in the metastatic sites might be different [43]. Due to the heterogeneity of HER2 expression, it is challenging to determine its level in all metastatic sites using a conventional biopsy-based approach. Radionuclide molecular imaging might provide information about HER2 expression in all lesions and could be used to select patients with HER2-positive lesions for HER2-targeted therapy [44]. In addition, this method is non-invasive and can be performed repeatedly to monitor changes in HER2 expression over time. Several formats of imaging agents could be considered for imaging of HER2 during ADAPT6-based targeted therapy: ADAPT6-ABD-mcDM1 itself or smaller formats, such as ADAPT6 itself or an affibody molecule.
The primary goal of this study was to test the hypothesis that targeted therapy using ADAPT6-ABD-mcDM1 can significantly improve survival of mice bearing HER2overexpressing ovarian cancer xenografts. The secondary goal was to evaluate several imaging probes ( 99m Tc-labeled ADAPT6-ABD-mcDM1, 99m Tc-labeled ADAPT6 and 99m Tclabeled affibody molecules Z HER2:41071 ) for visualization of HER2 expression in tumors for potential stratification of patients for ADAPT6-ABD-mcDM1-based therapy and monitor their response to treatment.

General
The chemicals used in the study were purchased from Sigma-Aldrich (Sweden AB, Stockholm, Sweden). Buffers were prepared using high-quality Milli-Q water.

Cell Culture
The human ovarian cancer cell line SKOV3 was purchased from American Type Culture Collection (ATCC via LGC Promochem, Borås, Sweden) and cultured in Roswell Park Memorial Institute (RPMI)-1640 medium supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, penicillin 100 IU/mL and 100 µg/mL streptomycin. The cells were grown in a humidified incubator at 37 • C and 5% CO 2 . Cells were detached using a trypsin-ethylenediaminetetraacetic acid (EDTA) solution (0.25% trypsin, 0.02% EDTA in buffer).

Experimental Therapy Study
The animal experiments were performed in agreement with national legislation on laboratory animals' protection and were approved by the Ethics Committee for Animal Research in Uppsala (permit 5.8.18-11931/2020, approved 28 August 2020).
The SKOV3 xenografts were established in female BALB/c nu/nu mice by subcutaneous implantation of 1 × 10 7 SKOV3 cells in 100 µL of medium in the abdominal region. The therapy started one week after the implantation. The mice were randomized in three groups (n = 9-10). One group of mice received intravenous (i.v.) injections of 13.3 mg/kg of ADAPT6-ABD-mcDM1 in 100 µL of PBS, the second group of mice received i.v. injections of the same dose of ADAPT Neg -ABD-mcDM1 in 100 µL of PBS, and the third group of mice received i.v. injections of 100 µL of PBS. The average tumor volume at the start of treatment was 101 ± 26, 94 ± 26 and 77 ± 22 mm 3 for mice treated with ADAPT6-ABD-mcDM1, ADAPT Neg -ABD-mcDM1 and PBS, respectively. There were no significant differences between the tumor volumes in the different groups ( Figure 2A). The average body weight was 18.8 ± 1.1 g. The injections were performed once a week for four consecutive weeks. Tumor volumes and body weights were measured twice a week. The tumors were measured using a digital caliper for the largest longitudinal (length) and transverse (width) diameter, and the tumor volumes were calculated using the formula: tumor volume = 1/2 (length × width 2 ). The relative tumor volume (RTV) was calculated using the formula: RTV = (tumor volume on the day of measurement)/(tumor volume on day 0).  Figure 2. Comparison of tumor volumes in experimental therapy groups (n = 9-10) in BALB/c nu/ mice bearing SKOV3 xenografts, receiving ADAPT6-ABD-mcDM1, ADAPTNeg-ABD-mcDM1 or P (vehicle) (A) at the start of the experiment (day 0), (B) after two treatment cycles (day 9) and after four treatment cycles (day 23). Individual tumor volume growth curves for mice received ADAPT6-ABD-mcDM1, (E) ADAPTNeg-ABD-mcDM1 or (F) PBS (vehicle). The mice were thanized when the volume of the subcutaneous xenografts exceeded 1000 mm 3 (T) or ulcers on xenografts were observed (U). (G) The average tumor volume (n = 9-10) in mice receiving ADAP ABD-mcDM1, ADAPTNeg-ABD-mcDM1 or PBS (vehicle) once a week for four weeks. The data presented as an average value ± standard deviation (SD). The curves were drawn until 30-33% the mice in a group were euthanized. The arrows indicate the days when treatment was admin tered. NS corresponds to no statistically significant difference (p > 0.05); * corresponds to p < 0 (one-way ANOVA with Bonferroni correction).
The mice were euthanized at predetermined humane end-points: tumor size excee ing 1000 mm 3 , tumor ulceration, internal bleeding, and total weight loss of 15% from t treatment initiation or 10% within one week. The study was terminated 90 days after t first injection according to requirements of the ethical permit. After the mice were sac ficed, kidneys and livers were fixed in 10% formalin solution for 24 h and stored in eth nol. The tissues were embedded in paraffin according to the standard operating pro dure, sectioned (3-4 µm thickness), stained with hematoxylin-eosin (HE) and examin Figure 2. Comparison of tumor volumes in experimental therapy groups (n = 9-10) in BALB/c nu/nu mice bearing SKOV3 xenografts, receiving ADAPT6-ABD-mcDM1, ADAPT Neg -ABD-mcDM1 or PBS (vehicle) (A) at the start of the experiment (day 0), (B) after two treatment cycles (day 9) and (C) after four treatment cycles (day 23). Individual tumor volume growth curves for mice received (D) ADAPT6-ABD-mcDM1, (E) ADAPT Neg -ABD-mcDM1 or (F) PBS (vehicle). The mice were euthanized when the volume of the subcutaneous xenografts exceeded 1000 mm 3 (T) or ulcers on the xenografts were observed (U). (G) The average tumor volume (n = 9-10) in mice receiving ADAPT6-ABD-mcDM1, ADAPT Neg -ABD-mcDM1 or PBS (vehicle) once a week for four weeks. The data are presented as an average value ± standard deviation (SD). The curves were drawn until 30-33% of the mice in a group were euthanized. The arrows indicate the days when treatment was administered. NS corresponds to no statistically significant difference (p > 0.05); * corresponds to p < 0.01 (one-way ANOVA with Bonferroni correction).
The mice were euthanized at predetermined humane end-points: tumor size exceeding 1000 mm 3 , tumor ulceration, internal bleeding, and total weight loss of 15% from the treatment initiation or 10% within one week. The study was terminated 90 days after the first injection according to requirements of the ethical permit. After the mice were sacrificed, Pharmaceutics 2022, 14, 1612 6 of 15 kidneys and livers were fixed in 10% formalin solution for 24 h and stored in ethanol. The tissues were embedded in paraffin according to the standard operating procedure, sectioned (3-4 µm thickness), stained with hematoxylin-eosin (HE) and examined for histopathological changes. The pathology examination was performed by a veterinary pathologist at BioVet AB veterinary medicine laboratory (Sollentuna, Sweden).

Imaging during Experimental Therapy
The imaging experiment was performed to determine an optimal tracer for noninvasive visualization of HER2 in malignant tumors for potential stratification of patients and for monitoring response to therapy using ADAPT6-ABD-mcDM1.
The radiochemical yield and purity were measured using iTLC silica gel strips (Varian, Lake Forest, CA, USA) in PBS. The distribution of activity along the strip was measured using a Storage Phosphor System (CR35 BIO Plus, ElysiaRaytest, Bietigheim-Bissingen, Germany) and analyzed with AIDA Image Analysis software (ElysiaRaytest, Bietigheim-Bissingen, Germany).
To visualize the biodistribution of ADAPT6-ABD-mcDM1, at day 36 (two days after imaging with 99m Tc-Z HER2:41071 ) the same mice from the PBS group were imaged with 99m Tc(CO) 3 -ADAPT6-ABD-mcDM1 (6 µg, 16 MBq in 100 µL of PBS) at 4 h and 24 h p.i. The SPECT acquisition time was 10 min at 4 h p.i. and 60 min at 24 h p.i.

Statistics
The data were analyzed using GraphPad Prism (9.3.1 for Windows, GraphPad Software, La Jolla, CA, USA). One-way ANOVA with Bonferroni correction for multiple comparisons was used to compare the values between multiple groups and determine statistically significant differences (p < 0.05). The survival data were analyzed using a Mantel-Cox log-rank test. The therapy outcomes (exponential tumor growth, delayed exponential tumor growth, controlled tumor growth (tumor ulceration without tumor growth)) were analyzed by a χ-square test.

Experimental Therapy
The therapeutic potency of ADAPT6-ABD-mcDM1 was investigated in mice bearing SKOV3 xenografts. There was no statistically significant (p > 0.05) difference in tumor volume between the groups at day 0 ( Figure 2A). After two cycles of treatment, at day 9, the tumors in the ADAPT6-ABD-mcDM1 group (69 ± 23 mm 3 ) were significantly (p < 0.01) smaller than tumors in the ADAPT Neg -ABD-mcDM1 group (163 ± 61 mm 3 ) and the PBS group (152 ± 71 mm 3 ) ( Figure 2B). The average tumor size in the group treated with ADAPT6-ABD-mcDM1 remained significantly smaller compared to the other groups until day 23 (the last day when all mice in all groups were alive) ( Figure 2C).
The individual tumor growth curves are shown in Figure 2D-F, the individual tumor volumes are shown in Table S1, the average tumor volumes are shown in Figure 2G, and the relative volumes are shown in Figure S1. The tumors in the ADAPT Neg -ABD-mcDM1 group (tumor doubling time 11.5 d, 95% CI from 10 to 13.5 d) and in the PBS group (tumor doubling time 12.1 d, 95% CI from 9.6 to 15.6 d) grew rapidly. The therapy outcome of the mice in the ADAPT Neg -ABD-mcDM1 group was that 100% mice were euthanized due to tumor ulceration. In the PBS group, six out of nine (67%) mice were euthanized due to tumor ulceration and three out of nine (33%) mice were euthanized when tumors reached the size limit. The tumor growth in the ADAPT6-ABD-mcDM1 group was inhibited until day 40, followed by a slow regrowth of all tumors (tumor doubling time 31 d, 95% CI from 21 to 55 d). The therapy outcome was that six out of ten (60%) mice were euthanized due to tumor ulceration. Four out of ten (40%) mice had controlled tumor growth and survived until the study termination (day 90), which was the maximum study length according to the ethical permit.
The median survival was significantly (Mantel-Cox log-rank test, p < 0.0001) longer in the ADAPT6-ABD-mcDM1 group (80.5 days) compared to the ADAPT Neg -ABD-mcDM1 group (28.5 days) and to the PBS group (30 days) ( Figure 3A). The therapy outcome was significantly (chi-square test, p < 0.0001) better for the ADAPT6-ABD-mcDM1 group than for the ADAPT Neg -ABD-mcDM1 group or the PBS group ( Figure 3B). The therapy outcome was that six out of ten (60%) mice were euthanized due to tumor ulceration. Four out of ten (40%) mice had controlled tumor growth and survived until the study termination (day 90), which was the maximum study length according to the ethical permit. The median survival was significantly (Mantel-Cox log-rank test, p < 0.0001) longer in the ADAPT6-ABD-mcDM1 group (80.5 days) compared to the ADAPTNeg-ABD-mcDM1 group (28.5 days) and to the PBS group (30 days) ( Figure 3A). The therapy outcome was significantly (chi-square test, p < 0.0001) better for the ADAPT6-ABD-mcDM1 group than for the ADAPTNeg-ABD-mcDM1 group or the PBS group ( Figure 3B). The difference between groups was determined using chi-square test. **** corresponds to p < 0.0001.
The treatment was well tolerated, with no observable side effects. No tendency for weight loss was observed in the ADAPT6-ABD-mcDM1 group compared to the ADAPT-Neg-ABD-mcDM1 group or the PBS group. The differences in average animal weight between the groups were within one standard deviation (Figure 4). The difference between groups was determined using chi-square test. **** corresponds to p < 0.0001. The treatment was well tolerated, with no observable side effects. No tendency for weight loss was observed in the ADAPT6-ABD-mcDM1 group compared to the ADAPT Neg -ABD-mcDM1 group or the PBS group. The differences in average animal weight between the groups were within one standard deviation (Figure 4). The histopathological examination results are shown in Figures S2 and S3. amined material, no lesions that would suggest toxicity of the treatment regim found in the liver or kidneys.

HER2 Imaging
The radiochemical yields of 99m Tc(CO)3-ADAPT6-ABD-mcDM1 and 99m ADAPT6 were 87% and 98%, respectively, and the purity after size-exclusion pu was close to 100% for both compounds. The radiolabeling of 99m Tc-ZHER2:41071 p radiochemical yield of 98 ± 2% (n = 3). No purification of this compound was pe 99m Tc(CO)3-ADAPT6-ABD-mcDM1 provided a clear visualization of HER2 ing SKOV3 xenografts using SPECT/CT already at 4 h p.i. (Figure 5A). Besides t a high accumulation of the labeled compound was visualized in kidneys, and a cumulation in the blood pool in heart and liver was comparable with the accum the tumor. By 24 h p.i. (Figure 5B), the activity was cleared to a high extent fr and liver, with kidneys and tumor being the only sites with a prominent activity lation. Imaging using 99m Tc(CO)3-ADAPT6 also provided a clear visualization of t at 4 h p.i. but with a substantial accumulation of activity in kidneys. In comparis ZHER2:41071 at 4 h p.i. provided an appreciably higher uptake in the SKOV3 xenog in any other tissue, enabling high-contrast visualization of the tumors (   Figures S2 and S3. In the examined material, no lesions that would suggest toxicity of the treatment regimens were found in the liver or kidneys.
99m Tc(CO) 3 -ADAPT6-ABD-mcDM1 provided a clear visualization of HER2-expressing SKOV3 xenografts using SPECT/CT already at 4 h p.i. (Figure 5A). Besides the tumor, a high accumulation of the labeled compound was visualized in kidneys, and activity accumulation in the blood pool in heart and liver was comparable with the accumulation in the tumor. By 24 h p.i. (Figure 5B), the activity was cleared to a high extent from blood and liver, with kidneys and tumor being the only sites with a prominent activity accumulation. Imaging using 99m Tc(CO) 3 -ADAPT6 also provided a clear visualization of the tumor at 4 h p.i. but with a substantial accumulation of activity in kidneys. In comparison, 99m Tc-Z HER2:41071 at 4 h p.i. provided an appreciably higher uptake in the SKOV3 xenografts than in any other tissue, enabling high-contrast visualization of the tumors ( Figure 5C).
SPECT/CT imaging of HER2 expression using 99m Tc-Z HER2:41071 was performed in mice during therapy ( Figure 6). The imaging enabled a clear visualization of the tumors in all cases. The tumor burden observed on imaging correlated well with the caliper measurements of tumor size (Table S1).
cumulation in the blood pool in heart and liver was comparable with the accumulati the tumor. By 24 h p.i. (Figure 5B), the activity was cleared to a high extent from b and liver, with kidneys and tumor being the only sites with a prominent activity accu lation. Imaging using 99m Tc(CO)3-ADAPT6 also provided a clear visualization of the tu at 4 h p.i. but with a substantial accumulation of activity in kidneys. In comparison, 9 ZHER2:41071 at 4 h p.i. provided an appreciably higher uptake in the SKOV3 xenografts in any other tissue, enabling high-contrast visualization of the tumors ( Figure 5C).

Discussion
The development of ADCs remains to be a mainstream direction in the field of targeted drug delivery. Currently, fourteen ADCs have been approved for clinical use, and over 100 are in clinical trials [46]. Despite this great progress, ADCs still face many challenges, such as complexity of development, production and quality control, variability in pharmacokinetics, toxicities due to payload release in blood circulation, insufficient tumor targeting, as well as drug resistance.
Alternative targeting proteins based on non-immunoglobulin scaffolds, such as affibody molecules, DARPins, ADAPTs and adnectins, have the advantage of facile genetic engineering to include additional domains, e.g., an ABD for extension of the half-life in blood, and sites for precise modification, e.g., attachment of drugs, providing uniform drug conjugates with well-defined properties. Due to more favorable structural characteristics, ESPs have higher tolerance to pH and temperature changes than mAbs, resulting in better in vitro and in vivo stability of the drug conjugates. Several conjugates of affibody molecules [18,25,26], DARPins [20,[31][32][33] and adnectins [35] with drugs and toxins have been developed and have demonstrated anti-tumor activity in preclinical studies.
The current study enhances this armamentarium by investigating a drug conjugate based on the ADAPT scaffold protein. Fusion of the HER2-targeting ADAPT6 to ABD 035 with femtomolar affinity to human serum albumin was used to prolong the conjugate's half-life in circulation. Binding to albumin increases the molecular weight of the conjugate from 14 to 81 kDa, which prevents excretion via glomerular filtration in kidneys. In addition, the half-life extension by the ABD is likely mediated by indirect targeting of the neonatal Fc receptor (FcRn) [47]. We previously found that the ABD placement at C-terminus of ADAPT6 provided the most optimal pharmacokinetics and tumor-targeting properties [22]. The maytansine derivative DM1, commonly used in ADCs, was conjugated to the C-terminal cysteine residue via a non-cleavable maleimidocaproyl (mc) linker. The ADAPT6-ABD-mcDM1 drug conjugate has been extensively characterized in vitro, and its biodistribution and tumor-targeting has been evaluated in vivo [39].
In this study, we investigated the therapeutic efficacy of ADAPT6-ABD-mcDM1 in mice bearing HER2-expressing SKOV3 ovarian cancer xenografts and compared it with a non-targeting homologous construct (ADAPT Neg -ABD-mcDM1) and a vehicle control. The treatment with ADAPT6-ABD-mcDM1 resulted in a significant delay of tumor growth in comparison with both the non-targeting and the vehicle-treated control groups ( Figure 2, Table S1). The therapeutic effect was apparently caused by targeted delivery of DM1 and not by anti-proliferative effect due to the presence of the conjugate in blood circulation or its accumulation in xenografts due to the enhanced permeability and retention (EPR) effect, as the homologous ADAPT Neg -ABD-mcDM1 did not inhibit tumor growth. This resulted in a significantly improved median survival of mice treated with ADAPT6-ABD-mcDM1 ( Figure 3A). The growth of all tumors in mice treated using ADAPT6-ABD-mcDM1 was well controlled until ca. day 40, after which it increased exponentially in 40% of mice. In all mice treated using ADAPT Neg -ABD-mcDM1 or vehicle, the tumors grew rapidly and followed an exponential growth trajectory (Figures 2E and 3B). However, despite the delayed tumor growth caused by treatment with ADAPT6-ABD-mcDM1 at a 13.3 mg/kg dose, only 40% of mice survived until the end of the study while having a significant tumor burden (tumor volume over 370 mm 3 ). In comparison with a previous study using an affibody-based Z HER2 -ABD-mcDM1 conjugate at a lower dose of (10.3 mg/kg), where at the study end-point 70% of mice had tumors below 50 mm 3 and 30% of mice had tumor below 200 m 3 , the therapeutic effect of the ADAPT6-based drug conjugate was not as pronounced [18]. This indicates that a higher dose of ADAPT6-ABD-mcDM1 is needed to achieve the same therapeutic effect as for Z HER2 -ABD-mcDM1.
During the development and characterization of ADAPT6-ABD-mcDM1, one of the concerns was an elevated renal uptake, which appreciably exceeded the uptake in the tumor. The injected dose of ADAPT6-ABD-mcDM1 (13.3 mg/kg) in the current study was lower than the highest dose of the affibody-drug conjugate Z HER2 -ABD-mcDM1 (15.1 mg/kg) used in a previous study [18]. Injection of 15.1 mg/kg of Z HER2 -ABD-mcDM1 caused weight loss for some animals; however, the histopathology investigation did not reveal any pathological changes in kidneys or livers. Compared to Z HER2 -ABD-mcDM1, ADAPT6-ABD-mcDM1 had lower blood retention, higher kidney uptake and equal liver uptake. The tumor uptake of ADAPT6-ABD-mcDM1 was lower at 24 and 48 h p.i. [18,39]. This data suggested that the therapy using ADAPT6-ABD-mcDM1 might have a higher risk of hepatic and renal toxicities with lower anti-tumor effect. Still, the histology evaluation of kidneys and livers from mice treated with ADAPT6-ABD-mcDM1 did not reveal any pathological changes ( Figures S2 and S3). One of the possible explanations for the pronounced antitumor effect and the absence of toxicities to liver and kidneys despite higher uptake of ADAPT6-ABD-mcDM1 in these organs could be the preferential cytotoxicity of DM1 to rapidly dividing cancer cells in a tumor in comparison to normal cells in liver and kidneys with lower proliferation rate. This suggests that a maytansine derivative DM1 or the auristatin derivatives MMAE and MMAF, with similar mechanisms of action, offer an additional advantage of lower general toxicity, in comparison with, e.g., bacterial toxins, which are cytotoxic to all types of cells.
Experience with other scaffold proteins suggests that the biodistribution profile of ADAPT6-based drug conjugate might be improved by optimizing the molecular design of the construct [48] as well as the linkers between the domains [49] and between the protein and cytotoxic moieties [31,50]. This could further reduce the toxicity to normal organs and increase the therapeutic window. This study has provided a proof-of-principle for a targeted delivery of non-radioactive cytotoxic payload to tumors using the ADAPT scaffold. Further studies should include an optimization of format and payload. Such work could be promoted by the use of a radioactive label to find variants providing the highest accumulations in tumors and the lowest in normal tissues. The anti-tumor activity of the best variant should be verified using other HER2-expressing ovarian cancer cell lines, such as, e.g., OVCAR-2 and OVCAR-3. The possible impact of HER2 expression level on the antitumor effect should be also evaluated before proceeding to clinical development.
The imaging of HER2 using the radiolabeled drug conjugate 99m Tc(CO) 3 -ADAPT6-ABD-mcDM1 provides the advantages of both visualizing HER2 expression in tumors and reporting the receptor accessibility by the targeted drug. Despite good tumor targeting and accumulation of 99m Tc(CO) 3 -ADAPT6-ABD-mcDM1 in the tumor, a high accumulation of activity in liver was also observed ( Figure 5A). The liver is a common metastatic site for ovarian and breast cancers; thus, high liver background might reduce the sensitivity of HER2 imaging in this metastatic site. Imaging at a later time point (24 h p.i.) improved the tumor-to-liver contrast due to a reduction in liver uptake ( Figure 5B). However, to achieve an appropriate image contrast using 99m Tc(CO) 3 -ADAPT6-ABD-mcDM1 would require a 24 h delay between the injection and imaging, higher injected activity and higher dose burden for patients.
In contrast, the non-ABD fused ADAPT6 labeled with 99m Tc(CO) 3 provided a comparable imaging contrast at the day of injection, with good correlation between imaging results and therapy outcomes ( Figure 5C). The smaller affibody-based imaging probe, Z HER2:41071 , labeled with technetium-99m showed much lower accumulation of activity in normal organs, including liver and kidneys, which resulted in a high-contrast image at 4 h p.i. (Figure 5D). Dosimetry evaluation showed that for both 99m Tc(CO) 3 -ADAPT6 and 99m Tc-Z HER2:41071 , very low effective doses were reached, with the lowest dose being for 99m Tc-Z HER2:41071 (0.00066 mSv/MBq) [17] and a medium dose for 99m Tc(CO) 3 -ADAPT6 (0.010 ± 0.003 mSv/MBq) [21], in comparison to the doses of clinically used imaging probes, such as 111 In-trastuzumab (0.19 ± 0.02 mSv/MBq) [53] and 111 In-pertuzumab (0.05 mSv/MBq) [54]. Another advantage of using an anti-HER2 affibody molecule as the imaging agent is that it does not compete with ADAPT6 for the binding to HER2 [38], which means that the imaging using 99m Tc-Z HER2:41071 could not only be used to select patients with HER2-positve lesions but could also be performed repeatedly during and after the treatment to monitor the therapeutic effect and disease progression, as we demonstrated in Figure 6.

Conclusions
Treatment using the ABD-fused DM1-conjugate based on the ADAPT6 scaffold protein increased survival of mice bearing HER2-expressing ovarian cancer xenografts without inducing toxicity to liver and kidneys. Imaging using 99m Tc(CO) 3 -ADAPT6-ABD-mcDM1 might provide information about the accessibility of HER2 in tumors for the therapeutic counterpart, while small non-ABD-fused radiolabeled scaffold proteins, such as 99m Tc-Z HER2:41071 and 99m Tc(CO) 3 -ADAPT6, might be better diagnostic companions for stratification of patients and monitoring of response to ADAPT6-ABD-mcDM1-based therapy.