Increase in negative charge of 68Ga/chelator complex reduces unspecific hepatic uptake but does not improve imaging properties of HER3-targeting affibody molecules

Upregulation of the human epidermal growth factor receptor type 3 (HER3) is a common mechanism to bypass HER-targeted cancer therapy. Affibody-based molecular imaging has the potential for detecting and monitoring HER3 expression during treatment. In this study, we compared the imaging properties of newly generated 68Ga-labeled anti-HER3 affibody molecules (HE)3-ZHER3-DOTA and (HE)3-ZHER3-DOTAGA with previously reported [68Ga]Ga-(HE)3-ZHER3-NODAGA. We hypothesized that increasing the negative charge of the gallium-68/chelator complex would reduce hepatic uptake, which could lead to improved contrast of anti-HER3 affibody-based PET-imaging of HER3 expression. (HE)3-ZHER3-X (X = DOTA, DOTAGA) were produced and labeled with gallium-68. Binding of the new conjugates was specific in HER3 expressing BxPC-3 and DU145 human cancer cells. Biodistribution and in vivo specificity was studied in BxPC-3 xenograft bearing Balb/c nu/nu mice 3 h pi. DOTA- and DOTAGA-containing conjugates had significantly higher concentration in blood than [68Ga]Ga-(HE)3-ZHER3-NODAGA. Presence of the negatively charged 68Ga-DOTAGA complex reduced the unspecific hepatic uptake, but did not improve overall biodistribution of the conjugate. [68Ga]Ga-(HE)3-ZHER3-DOTAGA and [68Ga]Ga-(HE)3-ZHER3-NODAGA had similar tumor-to-liver ratios, but [68Ga]Ga-(HE)3-ZHER3-NODAGA had the highest tumor uptake and tumor-to-blood ratio among the tested conjugates. In conclusion, [68Ga]Ga-(HE)3-ZHER3-NODAGA remains the favorable variant for PET imaging of HER3 expression.

www.nature.com/scientificreports www.nature.com/scientificreports/ For effective treatment, it is essential to reliably detect upregulation and monitor the status of HER3 expression. Radionuclide-based molecular imaging with positron emission tomography (PET) and single-photon emission tomography (SPECT) is a promising approach for non-invasive and repeatable assessment of target expression. The possibility for repetitive scans and in the case of PET also quantification is of particular interest for HER3 expression because of its dynamic oncogenic expression.
High contrast imaging of HER3 expression in cancer lesions is generally challenging. Overexpression in the malignant tissue is typically below 50.000 receptors/cells 16 and healthy organs, such as liver, endogenously express HER3. Several studies have reported monoclonal antibodies labeled with 89 Zr for imaging of HER3 expression and monitoring of HER3 status during treatment [17][18][19] . However, antibodies clear slowly from blood and therefore typically only provide suitable imaging contrast days after administration, which also prevents serial imaging within short time intervals. Furthermore, the enhanced permeability and retention effect, and hepatobiliary excretion limit the contrast in HER3 positive lesions and therefore clinical application 17 . The use of smaller imaging agents with faster kinetics and better tumor penetration, such as F(ab') 2 fragments, nanobodies, peptides or engineered scaffold proteins might be preferable 15,20,21 .
Affibody molecules are a class of engineered scaffold proteins with small size (7)(8), fast kinetics, extravasation and clearance, and good tumor penetration 22,23 . This makes them suitable candidates for molecular imaging. For example, 68 Ga-labeled HER2 (Human Epidermal Growth Factor Receptor 2) detecting affibody molecules have shown promising results in clinical trials 24 . Capable of visualizing both high and low HER2 expressing lesions, they could successfully identify the HER2-status of breast cancer patients 24 .
We have earlier reported on HER3-targeting affibody molecules radiolabeled with both PET and SPECT-isotopes as a promising alternative for imaging of HER3 expression [25][26][27][28] . Radiolabeled affibody molecules for imaging could be a suitable theranostic companion for future antibody or affibody based therapeutic agents against HER3 [29][30][31] . So far, the results are encouraging, but HER3 expression in healthy organs together with comparably low expression in tumors still pose challenges in achieving high imaging contrast. Generally, the imaging contrast could be improved by increasing tumor uptake or by decreasing uptake in healthy tissue (or both).
Hepatic metastases are common in many cancers and our recent efforts have focused on improving the imaging contrast by reducing the hepatic uptake. Liver uptake is mediated by the natural expression of HER3 but also partially related to unspecific or "off-target"-interactions which can be influenced by local charge of the targeting molecule and its hydrophilicity/lipophilicity [32][33][34][35] . We have previously demonstrated that co-injection of unlabeled trivalent affibody can block endogenous HER3 receptors in the liver to a greater extent than in the tumor, which consequently dramatically increased the tumor-to-liver contrast 36 .
Another approach is to focus on the molecular design of the affibody molecules to reduce unspecific uptake. Differences in structure of the metal/chelator-complex, surface exposure of functional groups and local distribution of charge, have shown to influence blood clearance, renal, hepatic and tumor uptake of anti-HER2 affibody molecules [37][38][39] . Particularly, presence of negative charged complexes can result in reduction of non-specific hepatic uptake. We also demonstrated that this is applicable to anti-HER3 affibody molecules 28,40,41 . Comparison of different indium-111/chelator complexes ( 111 In-NOTA, 111 In-NODAGA, 111 In-DOTA, 111 In-DOTAGA) conjugated to the C-terminus of Z HER3 showed that exchanging the positively charged 111 In-NOTA-complex with a negatively charged 111 In-DOTAGA resulted in a two-fold reduction in hepatic uptake and clearly improved the tumor-to-liver contrast 41 . Recently, we investigated the influence of a hydrophilic N-terminal (HE) 3 -tag on the 68 Ga-labeled anti-HER3 affibody molecules Z HER3 40 . The study included a head-to-head comparison of (HE) 3 -Z HER3 with Z HER3 labeled with 68 Ga via a NOTA (1,4,7-triazacyclononane-N,N0,N0 0-triacetic acid) or NODAGA (1-(1,3-carboxypropyl)-4,7-carboxymethyl-1,4,7-triazacyclononane) chelator. Presence of the hydrophilic (HE) 3 -tag increased the clearance rate from blood and reduced activity uptake in normal tissue, including liver, which resulted in increased tumor-to-liver ratios. In the same study, we found that NODAGA, which is neutrally charged when loaded with trivalent metals, was favorable compared to positively charged NOTA.

Results
production. The (HE) 3 -tagged HER3-binding affibody, (HE) 3 -Z HER3 , was produced in E. coli and purified by IMAC, followed by coupling to maleimide derivatives of DOTA and DOTAGA. The purity, determined with RP-HPLC, exceeded 95% for all conjugates (Fig. S1). The experimental molecular mass of each conjugate was in perfect agreement with the theoretical mass (Fig. S2). Notably, the mass determination revealed non-processed N-terminal methionine for all conjugates, due to the presence of the (HE) 3 -tag at the N-terminus. The alpha-helical content, thermal stability, refolding of the conjugates and melting temperatures were investigated by circular dichroism spectroscopy (Fig. S3, Table S1). Binding affinities were measured with surface plasmon resonance (SPR) analysis and K D values are presented in Table 1 as the average from duplicate injections. K D values refer to the monovalent affinity for human HER3 according to a Langmuir 1:1 model. Sensorgrams with fitted curves are shown in Fig. S4.
Following purification, both conjugates showed no major release of the radiolabel when incubated in phosphate-buffered saline (PBS). After incubation in human serum [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA showed a somewhat higher fraction of free gallium than [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA. Labeling of (HE) 3 -Z HER3 -NODAGA resulted in almost quantitative yields and purity > 99% matching previously reported results 40 . The results of the binding specificity experiment are illustrated in Fig. 1. Cells were incubated with 0.1 nM of [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA or [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA for 1 hour. In the blocked groups, HER3 receptors were pre-saturated by addition of 50 nM unlabeled Z HER3 , resulting in a significant decrease of activity uptake. Thus, binding of [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA and [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA was HER3-mediated. Overall uptake of the conjugates in DU145 cells was lower than in BxPC-3 cells.

In vitro
Cellular processing was studied by continuously incubating BxPC-3 and DU145 cells with 0.1 nM of the radiolabeled conjugates for up to 4 hours. At preselected time points, the membrane bound activity and internalized fractions were collected for BxPC-3 cells. For DU145 cells, only the total cell associated activity was studied, because of low signal due to the low level of HER3 expression. Figure 2 shows the uptake pattern of the activity, normalized to the maximum cell associated activity in BxPC-3 cells. Data for DU145 cells can be found in the Supplementary Material (Fig. S5). The binding of both conjugates to the cells was quick and increased in BxPC-3 cells over time. After 4 h the fraction of internalized activity was 23 ± 8% for [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA and 24 ± 8% for [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA. Uptake in DU145 cells was lower compared to uptake in BxPC-3 cells. The conjugates also associated quickly, but uptake did not increase over time. For in vivo specificity test, the amount of injected protein was adjusted to 70 µg using unlabeled anti-HER3 affibody.
Injection of excess amount of unlabeled HER3-targeting affibody molecules resulted in significant decrease of uptake in mErbB3 expressing organs and in the tumors, confirming HER3-mediated uptake of the molecules (Table 3). Still, even with co-injection of excess amount of protein, activity concentration in the liver was more than 2-fold higher for [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA than for [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA.

Discussion
Oncogenic expression of HER3 is dynamic, heterogeneous and a common cause for therapy resistance 11 . Radionuclide-based molecular imaging of HER3 expression could be valuable in evaluation of the status of HER3 expression in cancer patients, in pre-and post-treatment analysis as well as monitoring of treatment progress. Radiolabeled HER3-targeting affibody molecules can visualize and discriminate between different levels of HER3 expression in pre-clinical models 26,27 . We have previously shown that an increase in hydrophilicity using a N-terminal (HE) 3 -tag improves the biodistribution and tumor-to-liver contrast of [ 68 Ga] Ga-(HE) 3 -Z HER3 . Furthermore, we found that, thus far, [ 68 Ga]Ga-(HE) 3 -Z HER3 -NODAGA was the favorable Z HER3 -variant for PET-imaging with gallium-68 40 . In the present study, our aim was to investigate whether the C-terminal conjugation of tetraaza-chelators DOTA and DOTAGA for labeling of (HE) 3 -Z HER3 with gallium-68 would further improve the imaging properties. We therefore compared the newly produced Z HER3 -variants [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA and [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA with the previously selected [ 68 Ga] Ga-(HE) 3 -Z HER3 -NODAGA (Fig. 4).  www.nature.com/scientificreports www.nature.com/scientificreports/ All conjugates could be labeled with gallium-68 (     The labeled conjugates bound specifically to HER3 receptors in vitro and in vivo (Fig. 1, Table 3). In BxPC-3 cells, total uptake and the internalized activity increased with time (Fig. 2). Internalization rates for both conjugates were higher compared with our previously reported rates for [ 68 Ga]Ga-(HE) 3 -Z HER3 -NODAGA 40 . The influence of chelators on internalization properties of anti-HER2 and anti-HER3 affibody molecules has previously been reported 37,41 . Within the HER-family, internalization is dependent on receptor expression level and availability of dimerization partners. It is possible that the composition of the radiometal-chelator complex interferes with the heterodimerization of HER3. Overall, uptake was lower in DU145 cells due to lower receptor density.
The overall biodistribution correlated well with previously published data 27,40 . Typically for affibody molecules in general and HER3-targeting specifically, all conjugates cleared quickly from the blood via the renal pathway and accumulated in organs with endogenous expression of mErbB3. The uptake in the HER3-expressing xenografts was in the range of 3 to 4%ID/g. The different C-terminal compositions did not have a significant effect on the tumor uptake, but influenced the uptake in normal organs such as liver, blood, spleen, and small intestine.
Both tetraaza-conjugated variants showed slightly but significantly higher activity concentration in blood compared to [ 68 Ga]Ga-(HE) 3 3 -Z HER3 -DOTA had three-fold higher uptake in spleen and significantly higher hepatic uptake than [ 68 Ga]Ga-(HE) 3 -Z HER3 -NODAGA and -DOTAGA. Elevated uptake in these organs can indicate lower stability of the radiolabel. Release of 68 Ga from the chelator could result in trans-chelation to transferrin or the formation of gallium-hydroxide colloids, the latter tend to accumulate in the spleen and to some extent in the liver 33,43 . This in vivo finding corroborates with literature data and the observed lower in vitro stability of [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA 44 .
It was previously observed that increasing the negative charge of the radiometal/chelator complex can alter the biodistribution of HER3-targeting affibody molecules and particularly reduce unspecific uptake in the liver 28,33,40,41 . Our data aligned with this observation. [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA with a negatively charged 68 Ga/chelator-complex reduced the activity uptake in liver and small intestine compared to NODAGA-conjugated variant with neutral complex charge. However, the tumor-to-liver-ratio, which is an indicator for imaging contrast, was not significantly higher and the tumor-to-blood ratio was 3-fold lower for the DOTAGA-containing conjugate than for [ 68 Ga]Ga-(HE) 3 -Z HER3 -NODAGA. This may indicate that the variant with the most negative charge of the radiometal/chelator complex is not always the most favorable one overall. This is not concurrent with our findings with indium-111, where the 111 In-DOTAGA complex with the most negative charge also provided the most favorable biodistribution 41 . Indium and gallium isotopes differ in size, coordination number, and complex geometry with DOTA and its derivative DOTAGA 45 . Because of a smaller ionic radius, gallium-68 tends to prefer triaza-chelators, such as NOTA and NODAGA, whereas indium may prefer tetraaza-ligands 44,46,47 . Hence, conclusions from indium-labeled targeting molecules may not always be directly translatable to gallium-labeled variants.
Comparing the molecules included in this study with other small HER3-targeting molecules, both [ 68 Ga] Ga-(HE) 3 -Z HER3 -DOTA and [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA showed similar or higher tumor-to-liver ratios and higher tumor-to-blood ratios than a 89 Zr-labeled nanobody 21 and 68 Ga-labeled HER3 peptide 20 . Reported tumor uptake of the 89 Zr-labeled nanobody in xenografted mice was comparable at 3 h pi, and increased at 24 h pi 21 . However, the higher concentration of the nanobody in blood at 3 h pi compared to (HE) 3 -Z HER3 is a drawback for high contrast imaging within this timeframe. The reported uptake of the HER3-targeting peptide did not exceed 1%ID/g in the xenografts 20 .
In conclusion, the imaging properties of radiolabeled affibody molecules are determined by many different parameters. In the present study, we hypothesized that an increase in negative charge of the gallium-68/ chelator complex would decrease hepatic uptake and increase the contrast of HER3 PET-imaging. The results demonstrated that the negatively charged 68 Ga-DOTAGA complex indeed reduced the hepatic uptake, but did not improve the overall imaging properties. We therefore conclude that parameters influencing the imaging properties of affibody molecules should be studied for each molecule/isotope individually and that [ 68 Ga] Ga-(HE) 3 -Z HER3 -NODAGA remains the most promising variant for PET imaging of HER3 expression. www.nature.com/scientificreports www.nature.com/scientificreports/

Materials and Methods
General. Human cancer cell lines BxPC-3 and DU145 were purchased from American type tissue culture collection (ATTC via LGC Promochem, Borås, Sweden).
Metal contaminants were removed from buffers with Chelex100 Resin (Sigma Aldrich, St.Louis, MO, USA). ITLC was used to measure the distribution of activity to determine labeling yield and stability of the radiolabeled compounds. For analysis, 1 µl of the sample was added to strips made of silica gel-impregnated glass microfiber chromatography paper (Agilent Technologies, Santa Clara, CA, USA). Citric acid (0.2 M) was used for elution. With this method, free gallium-68 will move to the front of the strips and the radiolabeled affibody will stay at the application point. Distribution of activity was measured in the Cyclone Storage Phosphor System and analyzed with OptiQuant image analysis software (Perkin Elmer, Waltham, MA, USA).
Activity in cells and organs samples was measured with an automated gamma counter with a 3-inch NaI(Tl) detector (1480 Wizard; Wallac Oy, Turku, Finland). Raw data were corrected for decay.
Statistical significance for in vitro and in vivo specificity was tested with two-tailed, unpaired t-test. Comparison of the different groups in the biodistribution was done with 1-way ANOVA and post-hoc t-test corrected for multiple comparisons with the Bonferroni method.
Cells were lysed with French press and the supernatant was heated to 90 °C for 10 min followed by incubation on ice for 20 minutes and the aggregates were spun down for bulk removal of unwanted proteins. Thereafter, (HE) 3 -Z HER3 was purified on an ÄKTAexplorer (GE Healthcare, Uppsala, Sweden) using a 3 ml Ni Sepharose 6 Fast Flow column (GE Healthcare). Finally, the buffer of the eluate was changed to 20 mM NH 4 Ac (pH 5.5) and the proteins were freeze-dried.
(HE) 3 -Z HER3 was dissolved in 20 mM NH 4 Ac (pH 5.5) and reduced with a molar concentration of tris(2-carboxyethyl)phosphine (TCEP) equal to the protein concentration for 30 min at 37 °C. The proteins were incubated at 37 °C for 90 min with a ten-fold molar excess of maleimide derivatives of DOTA and DOTAGA (CheMatech) for site-specific conjugation to a C-terminal cysteine on the affibody. Metal ion contaminations were removed from all buffers used with Chelex 100 resin (Bio-Rad Laboratories).
For purification, reverse-phase high performance liquid chromatography (RP-HPLC) on a 1200 series HPLC system using a Zorbax 300SB-C18 semi-preparative column (Agilent Technologies, Santa Clara, CA) was used. Water with 0.1% trifluoroacetic acid was used as running buffer and an acetonitrile gradient was used for elution.
The molecular mass of the conjugates was confirmed with electrospray ionization mass spectrometry (ESI-MS) using a 6520 Accurate-Mass Q-TOF LC/MS (Agilent Technologies).
Characterization. Characterization of affibody conjugates was done as previously described 40 .
The purity of the conjugates was determined with RP-HPLC using an analytical Zorbax 300SB-C18 column (Agilent Technologies).
Alpha-helical content, thermal stability and refolding capacity of all conjugates were analyzed by circular dichroism spectroscopy (Chirascan spectropolarimeter Applied Photophysics, United Kingdom) with an optical path length of 1 mm at a concentration of 0.25 mg/ml.
The thermal stability was evaluated by measuring the change in ellipticity at 221 nm during heating (5 °C/min) from 20 to 90 °C. The melting temperatures (T m ) were estimated from the data acquired from variable temperature measurements (VTM) by curve fitting using a Boltzmann Sigmoidal model (GraphPad Prism, version 7). Spectra obtained from measurements at wavelengths in the range 195-260 nm at 20 °C, before and after thermal denaturation, were used to study the refolding capacity of the conjugates.
Binding affinity towards human HER3 was investigated using surface plasmon resonance (SPR) on a Biacore T200 system (GE Healthcare). The analysis was performed using single-cycle kinetics on a CM5 sensor chip with immobilized human HER3-Fc (Sino Biological). Five concentrations (3.125, 6.25, 12.5, 25 and 50 nM) of each conjugate were sequentially injected in a single cycle with a contact time of 150 seconds for each concentration.
To test stability, 1 µg of [ 68 Ga]Ga-(HE) 3 -Z HER3 -X was incubated for 1 hour in PBS at room temperature or in human serum at 37 °C. After incubation, the activity distribution was measured by ITLC.
In vitro characterization of [ 68 Ga]Ga-(HE) 3 3 -Z HER3 -DOTAGA towards HER3, HER3 receptors were blocked by addition of 50 nM unlabeled Z HER3 . After 10 minutes incubation at room temperature, 0.1 nM of [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA or [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA was added and samples were incubated for 1 hour at 37 °C. Thereafter, the cell samples were measured for activity content in the automated gamma counter.
To study the internalization of the compounds, BxPC-3 and DU145 cells were continuously incubated with 0.1 nM of either [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA or [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA for up to 4 hours. At selected time points, the membrane-bound activity fraction was collected after 5 minutes incubation with 0.2 M glycine buffer (0.15M NaCl, 4 M Urea, pH 2) on ice. The remaining activity was considered internalized and collected after incubating the cells with 1 M NaOH for 30 minutes at 37 °C.

In vivo experiments.
In vivo experiments were carried out as described previously 40 in compliance with national legislation on protection of laboratory animals and permission from the Ethics Committee for Animal Research in Uppsala, Sweden (approval number C5/16 from 26-02-2016).
Female Balc/c nu/nu mice were implanted subcutaneously with 5 × 10 6 cells/animal 20 days prior to the experiment. At the time of the experiment, the average tumor weight was 0.06 ± 0.03 g. Average mouse weight was 19 ± 1 g.
Mice were pre-injected with Ketalar-Rompun solution 10 mg/mL Ketalar and 1 mg/mL Rompun; 20 µL solution/gram of body weight) and sacrificed 3 h pi. Tumors and samples from blood, mErbB3 expressing organs (salivary gland, lung, liver, stomach, small intestine), spleen, kidney muscle and bone were collected. Samples were weighed and measured for activity content in the automated gamma counter. Uptake is presented as %ID/g. GI and carcass were collected and activity was measured. Uptake was presented as %ID.
To confirm binding specificity of [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTA and [ 68 Ga]Ga-(HE) 3 -Z HER3 -DOTAGA the injected protein dose was adjusted to 70 µg. A biodistribution experiment was done according to the protocol described above.