This article is mentioned in:

Introduction

Prostate cancer (PCa) growth is primarily driven by androgens (1). Therefore, androgen deprivation therapy (ADT), accomplished by surgical or medical castration, is the first-line therapeutic approach for the treatment of oligometastatic or high-risk localized PCa. ADT generally leads to remissions that last 1.5–3 years (2). However, despite the high initial response rate (3), the cancer almost invariably relapses, progressing to a hormone refractory state manifested by increased proliferation and invasion capacity. Hormone-refractory PCa is unresponsive to further hormonal manipulation and has a high resistance to cytotoxic drugs (4,5).

Although the molecular basis for such relapse is not fully defined, several signaling pathways have been identified by which androgen receptors are activated in the absence of ligands. Understanding these pathways is highly important for the development of new strategies for second-line therapies. One known mechanism responsible for androgen-independent growth involves alterations of growth factor receptor signaling (6,7). In particular, the epidermal growth factor receptor (EGFR) is one of the frequently deregulated gene products, and EGFR overexpression is associated with advanced stage PCa, progression towards androgen-independence and a poor clinical outcome (8–10).

EGFR, also known as ErbB1 or HER1, is a member of the tyrosine kinase EGFR family of receptors that also includes HER2/c-neu, HER3 and HER4. EGFR is activated by binding of specific ligands (EGF and TGFα are the preferred ligands), and three additional ligands that it shares with HER3 (11). Recently, the role of EGFR in cancer is a subject of extensive investigations. Ligand binding to EGFR triggers its dimerization or heterodimerization with other members of the EGFR family and activates the intracellular tyrosine kinase domain. This further recruits downstream signaling proteins and triggers several signaling cascades that regulate growth, survival, proliferation, motility and differentiation (12). The understanding of these mechanisms has resulted in the development of drugs targeting EGFR and its signaling.

For patients, a confirmed EGFR overexpression in PCa could reveal new second line-treatment options involving monoclonal antibodies (mAb) such as the already established cetuximab (Erbitux®) (13,14) and small molecule tyrosine kinase inhibitors (TKIs) such as erlotinib (Tarceva®) (15,16) or gefitinib (Iressa®) (17,18). Many other new targeting drugs are also in preclinical and clinical development, either specifically targeting EGFR, or targeting several intracellular signaling pathways simultaneously (19). All these approaches are highly selective and the successful implementation of anti-EGFR therapies requires an accurate detection of EGFR expression in primary tumors and metastasis.

Currently, the most common and established method for the detection of EGFR expression involves analyses of biopsy material. However, biopsy procedures are invasive, painful, and can lead to severe complications. Repetitive biopsies and biopsies from bone metastases are difficult. Additionally, EGFR expression in PCa is a relatively late event and the target expression heterogeneity can lead to false-negative results. EGFR expression may also change during the course of the disease due to the genetic instability of the cancer or in response to treatment (20). There is a clear unmet clinical need for non-invasive robust methods to determine EGFR expression status, both initially and as a consequence of treatment.

Imaging of EGFR aberrant expression in PCa using radiolabeled EGFR-specific agents is a promising approach that allows repetitive, non-invasive investigations of receptor status in multiple lesions simultaneously. Potential radioimaging agents include the natural ligands of EGFR and the anti-EGFR mAbs. The use of radiolabeled epidermal growth factor (EGF) is a straightforward approach, but is limited by its physiological activity that may cause adverse reactions (nausea, vomiting, diarrhea, hypotension, fever and chills) (21). Radiolabeled intact anti-EGFR mAbs have been extensively investigated (22–28) demonstrating their capacity to specifically accumulate in tumors proportionally to EGFR expression. However, the sensitivity of mAb-mediated imaging is limited by the long blood residence time and slow tumor penetration, which reduces the imaging contrast, a major limitation for targets with low expression. A smaller targeting agent without physiological activity would be desirable in order to obtain a higher contrast than antibodies can provide.

Affibody molecules are a promising class of imaging agents that are characterized by small size and high affinity to the intended target. They are scaffold proteins composed of anti-parallel three-helix cysteine-free bundles, derived from the immunoglobulin-binding B domain of staphylococcal protein A. Randomization of 13 surface-exposed amino acids on helices 1 and 2 are used for generation of large combinatorial libraries from which high affinity binders can be selected. The small size (7 kDa) of affibody molecules facilitates rapid extravasation, rapid tumor penetration and fast clearance of unbound tracer from blood. This results in high contrast imaging only a few hours after injection, as seen in both preclinical and clinical studies (29–31).

Nonetheless, imaging of EGFR in PCa is challenging due to the relatively low expression in PCa (in comparison to other cancers). The feasibility of affibody-mediated imaging of EGFR expression was previously demonstrated using DOTA-ZEGFR:2377. This construct was initially radiolabeled with 111In, exhibiting favorable pharmacokinetic properties, with higher tumor-to-organ ratios than any anti-EGFR monoclonal antibody (32). In a further attempt to improve the sensitivity of EGFR detection in vivo, the use of positron-emitting labels was considered. Positron emission tomography (PET) provides better spatial resolution, higher sensitivity and more accurate quantification compared to single-photon emission computed tomography (SPECT). Several imaging probes based on ZEGFR:2377 for PET imaging of EGFR were proposed for labeling with 68Ga, 55Co, and 89Zr (33,34). Among them, the radiocobalt labeled affibody molecule had the most favorable biodistribution when evaluated on an epidermoid carcinoma model, A-431, with high EGFR-expression [2×106 receptors/cell (35)]. In this tumor model, radiocobalt labeled ZEGFR:2377 demonstrated 2-fold higher tumor-to-blood, tumor-to-bone and tumor-to-muscle ratios compared to the same construct radiolabeled with 68Ga and five-fold higher compared to 89Zr-DFO-ZEGFR:2377 at 3 h post injection (pi) (33,34).

The promising results obtained during the initial validation of radiocobalt-labeled DOTA-ZEGFR:2377 in epidermoid carcinoma, provided the incentive for its evaluation in a more challenging, low receptor-expressing PCa model. In PCa, the high tumor-to-blood, tumor-to-bone and tumor-to-muscle ratios are particularly important, since the majority of metastases occur in the abdominal lymph nodes and bones. Moreover, the significantly lower liver uptake of 57Co-DOTA-ZEGFR:2377 compared to the 68Ga-labeled analogue could be particularly relevant in a low EGFR-expressing tumors where protein dosing is crucial, as more injected compound would be available for specific binding to EGFR-expressing tumors and tissues.

The aim of the present study was to test the hypothesis that radiocobalt-labeled ZEGFR:2377 affibody molecule would enable PET imaging of EGFR receptor expression in PCa xenografts where EGFR expression is low. A long-lived PET radionuclide, 55Co has a half-life of 17.5 h, and would allow imaging the next day after injection when better contrast could be achieved (32). Cobalt-55 is not commercially available, which is an obstacle for pre-clinical investigations. However, we have recently demonstrated that data obtained for a bombesin analogue labeled with cobalt-57 (T½=272 d) were fully translatable to cobalt-55 (36). Therefore, 57Co was chosen as a surrogate for 55Co in the present study.

Materials and methods

Labeling of DOTA-ZEGFR:2377 with cobalt-57

The C-terminal cysteine-containing anti-EGFR affibody molecule Z2377, conjugated with maleimido-derivative of DOTA, was produced as previously reported (32). Buffers for labeling were purified from metal contamination using Chelex 100 resin (Bio-Rad Laboratories, Inc., Richmond, CA, USA). Labeling with radiocobalt and quality control methods were performed as previously described (33). Briefly, DOTA-ZEGFR:2377 (50 µg) in 40 µl of 0.2 M ammonium acetate buffer, pH 5.5, was mixed with 57Co stock solution (20 µl, 20–40 MBq). The reaction mixture was incubated for 30 min at 60°C. After labeling, the conjugate was purified using disposable NAP-5 size-exclusion columns pre-equilibrated with phosphate-buffered saline (PBS). The yield and radiochemical purity of the conjugate were evaluated by instant thin-layer chromatography (ITLC) using Tec-Control Chromatography 150–771 strips (Biodex Medical Systems, Shirley, NY, USA) eluted with 0.2 M citric acid, pH 2.0. The distribution of radioactivity along the ITLC strips was measured on a Cyclone Storage Phosphor System (PerkinElmer, Inc., Waltham, MA, USA).

Cell culture

Human PCa cell line DU-145 [American Type Tissue Culture Collection (ATCC) via LGC Promochem, Borås, Sweden] was used. The EGFR expression in this cell line was ~2×105 receptors/cell (37). Roswell Park Memorial Institute medium (RPMI-1640) supplemented with 10% fetal calf serum (FCS), 2 mM L-glutamine and PEST (penicillin 100 IU/ml, streptomycin 100 µg/ml) (all from Biochrom AG, Berlin, Germany) was used for cell culturing. This medium was referred to as complete medium in the text. The cells were detached using a trypsin-EDTA solution (0.05% trypsin, 0.02% EDTA; Biochrom AG) and were counted using an electronic cell counter (Beckman Coulter, Inc., Brea, CA, USA).

In vitro cell studies

The in vitro binding specificity assay was designed to determine whether the binding of 57Co-DOTA-ZEGFR:2377 to EGFR-expressing DU-145 cells was receptor-mediated. In this experiment, the cells were incubated with 1 nM 57Co-DOTA-ZEGFR:2377 for 2 h at room temperature. One set of dishes was pre-incubated with i) 0.5 µM unlabeled affibody molecule, ii) one with 0.5 µM of cetuximab, and iii) one with 0.5 µM bevacizumab for 10 min at room temperature. After incubation with 57Co-DOTA-ZEGFR:2377, the cells were washed with serum-free media and detached using 0.5 ml trypsin-EDTA solution. The cell-associated radioactivity was assessed using a γ-counter (2480 WIZARD2; PerkinElmer, Inc.) and presented as a percentage from added radioactivity.

Cellular processing was evaluated using DU-145 cells. The cells were incubated at 37°C with 1.5 nM of 57Co-DOTA-ZEGFR:2377. At predetermined time-points (1, 2, 4, 8 and 24 h after the start of incubation) the membrane-bound and internalized radioactivity fractions were collected and calculated as previously described (37).

To assess the cellular retention of radioactivity after interrupted incubation with radiocobalt-labeled affibody molecules, cultured DU-145 cells were incubated for 1 h at 4°C with 1.5 nM 57Co-DOTA-ZEGFR:2377. The cell dishes were subsequently washed with serum-free media, fresh complete media was added, and the dishes were incubated at 37°C. At predetermined time-points, membrane-bound and internalized radioactivity fractions were collected and measured using the γ-counter.

The dissociation constant (KD) of 57Co-labeled DOTA- ZEGFR:2377 binding to living DU-145 cells was measured using LigandTracer Yellow instruments (Ridgeview Instruments AB, Uppsala, Sweden) as previously described (38). Briefly, DU-145 cells were seeded in one area of the Petri dish (Nunclon™; 100-mm diameter; Thermo Fisher Scientific, Hvidovre, Denmark). After the addition of 3 ml of complete media, the dish was placed on the rotating table of the instrument. After a 15-min baseline run, 57Co-labeled DOTA-ZEGFR:2377 was added to the media to obtain a ligand concentration of 1 and then 5 nM. For each concentration, the uptake curves were measured for 120 min. Subsequently, the 57Co-labeled DOTA-ZEGFR:2377-containing medium was removed, fresh media (3 ml) was added to the dish, and the dissociation curve was recorded for 11 h. Two runs were performed at room temperature. The calculation of equilibrium dissociation constant was performed using TracerDrawer software (Ridgeview Instruments AB, Vänge, Sweden).

In vivo studies

All animal experiments were planned and performed in accordance with the Swedish national legislation on the protection of laboratory animals and the study plans were approved by the Ethics Committee for Animal Research in Uppsala. Euthanasia was performed by intraperitoneal injection of a ketamine-xylazine (Ketalar-Rompun) solution (200 mg/kg ketamine/Ketalar and 20 mg/kg xylazine/Rompun), and all efforts were made to minimize suffering. Female outbred BALB/c nu/nu mice (6 weeks of age) were purchased from Taconic M&B (Ry, Denmark). The animals (n=25) were housed at 23°C, 45% humidity, 12-h light/dark cycle, food and water ad libitum. Prior to implantation the mice were quarantined for 1 week. At the time of experiment the average animal weight was 19±1 g. EGFR-expressing xenografts were established by subcutaneous injection of ~1×107 DU-145 cells/mouse. The tumors were grown for 8 days before the experiment (in Matrigel, BD Biosciences) and the average tumor weight was 0.13±0.08 g. Before the experiments the animals were randomized into groups of four.

The mice were intravenously injected with 30 kBq of radiolabeled conjugate per mouse in 100 µl PBS with 35 µg/mouse (one group) or 10 µg/mouse (three groups). The protein dose was adjusted by the addition of non-labeled DOTA-ZEGFR:2377. The animals were sacrificed at 3 h pi (10- and 35-µg groups) and 7 and 24 h pi (10-µg groups). Blood, tumors and organ samples were collected and weighed, and their radioactivity was measured. Tissue uptake was calculated as the percent of injected dose per gram (% ID/g).

To verify specificity of in vivo targeting, a group of six mice was pre-injected with an excess amount of anti-EGFR monoclonal antibody cetuximab (5 mg/ml; cat. no. 64293; Merk KGaA, Darmstadt, Germany) 24 h prior to the injection of 10 µg (30 kBq) of 57Co-DOTA-ZEGFR:2377. The animals were sacrificed at 3 h pi. Cetuximab does not cross-react with murine EGFR, therefore only blood and tumors were collected for the blocking group.

Whole body scans of the subjects were performed using the Triumph™ Trimodality system (Gamma Medica, Inc., Salem, NH, USA), an integrated microSPECT/PET/CT platform. The mice bearing DU-145 ×enografts were euthanized by CO2 asphyxiation at 3, 7, and 24 h pi of 10 µg (4 MBq) 57Co-DOTA-ZEGFR:2377. The computed tomography (CT) acquisition was carried out at the following parameters: Field of view (FOV), 80 mm; magnification, 1.48; one projection and 512 frames for 2.13 min. SPECT acquisition was performed at the following parameters: FOV, 80 mm; 75A10 collimators (5 pinhole); 64 projections, 2 h scan time. CT raw files were reconstructed by Filter Back Projection (FBP). SPECT raw data was reconstructed by the FLEX™ SPECT software, which uses an ordered Subset Expectation Maximization (OSEM) iterative reconstruction algorithm. SPECT and CT data were fused and analyzed using PMOD v3.508 (PMOD Technologies Ltd., Zurich, Switzerland).

Statistical analysis

Data were assessed either by one-way ANOVA with Dunnett or with Bonferroni correction for multiple comparisons using GraphPad Prism 7.03 (GraphPad Software, Inc., La Jolla, CA, USA) in order to determine significant differences (P<0.05).

Results

Labeling with 57Co under the selected conditions provided an average yield of 95±5%. After purification using disposable NAP-5 size-exclusion columns, the radiochemical purity was 100%.

In vitro cell studies

The results of binding specificity tests are presented in Fig. 1. Pre-saturation of EGFR with a large molar excess of non-labeled affibody molecules or EGFR-targeting mAb cetuximab resulted in a significantly reduced cell-associated radioactivity after 2 h incubation (n=3, P<0.05). On the contrary, the non-relevant mAb bevacizumab (anti-VEGF) did not influence the binding of the radioconjugate to cells. This finding indicated EGFR-mediated binding of 57Co-DOTA-ZEGFR:2377 to DU-145 cells.

Figure 1. In vitro specificity test for 57Co-DOTA-ZEGFR:2377 on DU-145 cells. Blocked dishes were pre-saturated with molar excess of non-labeled peptide (0.5 µM), Cetuximab (0.5 µM) and Bevacizumab (0.5 µM) 10 min prior to the addition of 1 nM radiolabeled compound with 500 fold molar excess. The cell-associated radioactivity was calculated as a percentage of the total added radioactivity (mean values of three dishes ± SD). Asterisks mark significant differences (n=3, ***P<0.05). Data were assessed by one-way ANOVA with Dunnett correction for multiple comparisons.

Data concerning cellular processing of 57Co-DOTA-ZEGFR:2377 are presented in Fig. 2A. Cell-associated radioactivity increased continuously over time, while the contribution of internalized radioactivity was low, with <12% of total cell-associated radioactivity internalized after 24 h. The cellular retention of 57Co-DOTA-ZEGFR:2377 was also studied at various time-points (Fig. 2B). After an initial rapid dissociation phase (up to 2 h), the cell-associated radioactivity plateaued at 40% of initially bound radioactivity. The fraction of internalized radioactivity was low, similar to the previous experiment.

Figure 2. Cellular processing of 57Co-DOTA-ZEGFR:2377 by DU-145 cells during (A) continuous incubation with 1.5 nM of 57Co-DOTA-ZEGFR:2377 and (B) after interrupted incubation with 1.5 nM of 57Co-DOTA-ZEGFR:2377. Data are presented as the means ± SD of three culture dishes. Error bars may not be visible because they are smaller than point symbols.

The binding affinity of 57Co-DOTA-ZEGFR:2377 to DU-145 cells was assessed in real-time. The evaluation of the binding kinetics revealed that the interaction did not follow a 1:1 Langmuir adsorption model. A superior fitting was obtained for the 1:2 model, suggesting two independent interactions with different kinetic parameters (Fig. 3A). The two processes occured simultaneously and consisted of a dominating high affinity interaction (79 pM, 66%) and a minor lower affinity interaction (3.7 nM, 17%) (Fig. 3B).

Figure 3. (A) LigandTracer sensorgram and fitted binding curve (black) of 57Co-DOTA-ZEGFR:2377 interaction with EGFR-expressing DU-145 cells at room temperature. Uptake curves were recorded at 1 and 5 nM. The data were fitted to a one-to-two interaction model and represent the combination of two distinct interactions: A dominating high affinity interaction 0.08 nM (dark grey) and a minor interaction of lower affinity of 3.7 nM (light grey). (B) Interaction map: The simultaneous interactions are observed as distinct processes in the map. The grey scale reflects the contribution of each interaction (66% for 0.08 nM, 17% for 3.7 nM). Darker areas corresponded to larger contributions.

In vivo studies

Data concerning the biodistribution of 57Co-DOTA-ZEGFR:2377 in BALB/c nu/nu mice bearing DU-145 PCa xenografts are presented in Table I. In the first experiment, the biodistribution was compared at 3 h after injection of 35 and 10 µg of radiolabeled protein, respectively. A significant, dose-dependent difference in the accumulation of radioactivity in tumors and normal organs was observed. The concentration of radioactivity in blood and most normal organs (except for the spleen and muscles) was significantly higher in the 10-µg group (n=4, P<0.05). Tumor uptake was also 2-fold higher in the group of mice injected with 10 µg of 57Co-DOTA-ZEGFR:2377 (2.1±0.1% IA/g), compared to the 35-µg group (1.1±0.1% IA/g). Tumor-to-non-tumor ratios were similar in both groups. The high uptake in kidneys indicated that 57Co-DOTA-ZEGFR:2377 was cleared by glomerular filtration with subsequent tubular re-absorption.

Table I. Biodistribution of 57Co-DOTA-ZEGFR:2377 after injection in BALB/c nu/nu mice bearing DU-145 ×enografts.

Table I.

Biodistribution of 57Co-DOTA-ZEGFR:2377 after injection in BALB/c nu/nu mice bearing DU-145 ×enografts.

	35 µg	10 µg
	3 h	3 h	7 h	24 h
Blood	0.73±0.08a	1.8±0.6 (2.0±0.8)	1.3±0.2	0.30±0.04b,c
Saliv. glands	0.62±0.04	1.23±0.08	1.1±0.2	0.44±0.05b,c
Lung	0.77±0.06	1.9±0.2	1.3±0.1b	0.44±0.05b,c
Liver	4.8±0.5	7±2	5.3±1.7	3.1±0.5b
Spleen	1.1±0.2	1.0±0.1	0.69±0.05b	0.8±0.1
Small int.	0.69±0.04a	1.3±0.2	1.0±0.2	0.45±0.02b,c
Colon	0.64±0.09a	1.34±0.09	1.10±0.09b	0.48±0.09b,c
Kidney	319±17	300±30	278±27	247±29
Tumor	1.1±0.1a	2.1±0.5d (1.3±0.2)	1.8±0.3	1.0±0.2b
Muscle	0.4±0.4	0.35±0.06	0.26±0.05	0.15±0.04b
Bone	0.40±0.07a	0.7±0.2	0.57±0.09	0.26±0.03b,c

{ label (or @symbol) needed for fn[@id='tfn1-or-41-01-0534'] } Data represent the percentage of injected dose per gram tissue (% ID/g) and are expressed as the mean value ± SD (n=4). Data within parenthesis were collected from mice pre-injected with an excess amount of the anti-EGFR monoclonal antibody cetuximab. Since cetuximab does not cross-react with murine EGFR, only blood and tumor uptake were evaluated. Data were assessed either by one-way ANOVA with Dunnett (^e) or Bonferroni (^f) correction for multiple comparisons in order to determine significant differences (P<0.05).

a Value was significantly lower than for 10 µg^e

b value was significantly lower than for 3 h, 10 µg^f

c value was significantly lower than for 7 h, 10 µg^f

d value was significantly higher than for mice pre-injected with an excess amount of cetuximab^e. EGFR, epidermal growth factor receptor.

Specificity of EGFR-targeting in vivo was assessed by pre-saturation of receptors with non-labeled monoclonal antibody cetuximab. The tumor uptake in the blocked group of mice was significantly reduced (P<0.05) compared to the control group. No significant difference was found in the radioactivity uptake in blood (Table I).

Data on the biodistribution of 57Co-DOTA-ZEGFR:2377 over time are presented in Table I. In this experiment, two additional groups of mice were injected with 10 µg and the animals were sacrificed at 7 and 24 h pi. There was no significant release of radioactivity from the tumors between 3 and 7 h pi, but radioactivity uptake in tumors dropped 2-fold at 24 h pi. The clearance from blood and normal organs was also relatively slow, and with the exception of lungs, spleen and colon, there was no noticeable decrease in the radioactivity concentration between 3 and 7 h pi. However, radioactivity uptake in normal organs decreased 3-fold at 24 h pi which was rapider than in tumors. As a result, tumor-to-non-tumor ratios were not significantly improved between 3 and 7 h pi, but significantly increased at 24 h pi (n=4, P<0.05) (Fig. 4).

Figure 4. Tumor-to-organ ratios for 57Co-DOTA-ZEGFR:2377 after injection of 10 µg in BALB/C nu/nu mice bearing DU-145 ×enografts. Data are presented as the mean value ± SD. Asterisks mark organs where the tumor-to-organ ratio significantly differed between 7 and 24 h pi (n=4, P<0.05). Data were assessed by one-way ANOVA with Bonferroni correction for multiple comparisons.

MicroSPECT scans were acquired at 3, 7, and 24 h after the intravenous administration of 10 µg (4 MBq) 57Co-DOTA-ZEGFR:2377 to BALB/c nu/nu mice bearing subcutaneous DU-145 ×enografts (Fig. 5). SPECT/CT images confirmed the capacity of 57Co-DOTA-ZEGFR:2377 to visualize EGFR expression. The highest uptake of radioactivity was observed in the kidneys, followed by the liver, concordant with the biodistribution results. The microSPECT experiment confirmed that imaging of EGFR expression in PCa is possible.

Figure 5. (A) Coronal, (B) Axial, and (C) Sagittal images of EGFR expression in DU-145 ×enografted BALB/c nu/nu mice. The animals were injected with 10 µg of 57Co-DOTA-ZEGFR:2377 (4 MBq) and euthanized at 3, 7, and 24 h after injection. K, kidney; L, liver; T, tumor.

Discussion

EGFR is significantly upregulated in advanced stages of PCa and is associated with a high risk of recurrence, progression towards androgen independence and metastasis (8). Therefore, EGFR has emerged as a promising target for second-line therapy of disseminated PCa. Such therapies include neutralization of EGFR by mAb that prevent binding to its ligands and the subsequent activation of downstream signaling pathways and TKI that compete with the ATP binding site of the catalytic domain of tyrosine kinases (39,40). However, the use of anti-EGFR mAb cetuximab for treatment of unselected patients exhibited a limited success if any (41,42). Moreover, adding cetuximab to other therapeutics often aggravated the side-effects. Both studies claimed the necessity of predictive biomarkers for EGFR-targeted therapy of PCa. Fleming et al demonstrated in a post hoc analysis that appearance of a rash could be a pharmacodynamics biomarker for response enabling selection of tailored dosing of cetuximab (42). A further study demonstrated that EGFR overexpression was correlated with the response of PCa to cetuximab treatment (43). A probe enabling imaging of EGFR expression in PCa metastases before treatment and data concerning receptor occupancy during treatment would provide important predictive and pharmacodynamic information and would facilitate the use of cetuximab and TKI for the treatment of PCa.

Comparison of data from our previous studies (32–34) demonstrated that 57Co-DOTA-ZEGFR:2377 provided better contrast of EGFR imaging than 89Zr-cetuximab and ZEGFR:2377 variants labeled with 111In, 68Ga or 89Zr. Notably, 57Co-ZEGFR:2377 binds to the same epitope as cetuximab as seen in the in vitro binding specificity assay. In this experiment, the uptake of 57Co-ZEGFR:2377 on EGFR-expressing DU-145 cells was successfully blocked by the addition of excess cetuximab (Fig. 1). Binding specificity was further confirmed in vivo in mice bearing DU-145 ×enografts. The uptake of 57Co-ZEGFR:2377 in tumors was significantly lower in mice pre-injected with an excess amount of cetuximab, while the blood-born radioactivity was similar (Table I). These experiments indicated that radiocobalt-labeled ZEGFR:2377 could be used for monitoring receptor occupancy during cetuximab therapy. The significant reduction of 57Co-DOTA-ZEGFR:2377 tumor uptake in a post-treatment scan compared to a baseline image would indicate that EGFR in the lesion was blocked sufficiently.

Data concerning cellular (Fig. 2) processing of 57Co-DOTA- ZEGFR:2377 were in agreement with the data for A431 cells (37). In both cases, the internalized fraction was unexpectedly low for a radiometal label (below 15% after 24 h). For comparison, the internalized fraction reached 30% for the same conjugate radiolabeled with indium-111 (37). The study of radioactivity retention after interrupted incubation also revealed a low internalized fraction that did not increase with time (Fig. 2). Collectively, this may indicate lower residualizing properties of cobalt-containing radiocatabolites that could lead to lower retention of labels in tumors in vivo over time.

The presence of two types of binding sites characterized by different affinities was found when evaluating the binding kinetics of 57Co-DOTA-ZEGFR:2377 to EGFR-expressing DU-145 cells. The data indicated the existence of two populations of EGFR in DU-145 cells: The major one with picomolar affinity and another one with low nanomolar affinity. This finding was consistent with previous results when binding of natural ligand EGF to EGFR was studied (38) and was revealed to vary depending on culturing conditions (44,45). A possible explanation for the two types of interactions is receptor homo- and heterodimerization, which can result in different accessibility of binding sites. It should to be noted that both interactions were strong (0.08 and 3.7 nM), with the strongest interaction predominant.

One of the main obstacles for imaging of EGFR expression in tumors is an appreciable EGFR expression in a number of normal tissues, particularly in liver hepatocytes. Liver is a large well-perfused organ. The injection of an insufficient amount of e.g. anti-EGFR antibody may result in its nearly complete trapping in liver (46). Clinical studies have demonstrated, however, that it is possible to adjust the dose of the anti-EGFR antibody to saturate EGFR in liver and provide release of radiolabeled antibody into circulation (23). We have also shown in preclinical studies that it is possible to partially saturate murine EGFR in liver with 111In-DOTA-ZEGFR:2377 without saturating receptors in xenografts (32). The optimal protein dose, 35 µg, was also used in the previous study with 57Co-DOTA-ZEGFR:2377 (33).

However, these initial studies utilized the A431 cell line, having EGFR expression of ~2×106 receptors/cell. The expression level of EGFR in PCa is appreciably lower (Human Protein Atlas http://www.proteinatlas.org/). For example, the DU-145 cell line has an EGFR expression of 2×105 receptors/cell, i.e. one order of magnitude lower than A431 cells. The average tumor weight at the time of the experiment was 0.13±0.08 g corresponding to lesions of 4 mm in diameter. The spatial resolution of modern clinical PET cameras is 6–10 mm and radioactivity concentration in any objects below this size will be underestimated due to partial volume effect. Taking this in account, the tumors in our study were relatively small from a clinical perspective. The in vivo data revealed a pronounced influence of injected dose on the biodistribution profile of 57Co-DOTA-ZEGFR:2377. We compared the biodistribution and tumor uptake after the injection of 35 µg [dose found optimal for high EGFR expression xenografts A431 (32,33)] and 10 µg of radiolabeled conjugate at 3 h pi. A detailed dose-dependent uptake of anti-EGFR affibody molecules in normal organs was previously studied in detail by Tolmachev et al (32), creating the rationale for the choice of protein doses tested in the present study.

Our reasoning was that a high protein dose could lead to saturation of EGFR in tumors with low expression level in the same manner as in normal organs with EGFR expression, such as salivary glands and intestine walls. Decreasing the protein dose resulted in a significant increase of uptake in tumors as well as in most studied organs including blood. The same phenomenon was also observed for 111In-DOTA-ZEGFR:2377 and 111In-DOTA-ZEGFR:1907 studied in high EGFR-expressing A431 ×enografts (32,46). The elevated blood radioactivity may be attributed to the dissociation of non-internalized radioconjugates from EGFR-expressing organs followed by re-entrance in blood circulation. In the group of mice injected with 35 µg 57Co-DOTA-ZEGFR:2377, the binding sites were partially saturated with non-labeled compound, which lead to a reduction of uptake in tumors and EGFR-expressing organs (liver, salivary glands, lungs) and lower blood-born radioactivity. However, in the low EGFR-expressing xenografts used in the present study, the decrease in protein dose resulted in a significant 2-fold increase in radioactivity uptake in tumors (n=4, P<0.05).

The higher tumor uptake would increase the sensitivity of imaging of lesions with low receptor expression, such as EGFR in PCa. Therefore, the injected dose of 10 µg was used for measuring the biodistribution over time. The time-points assessed in the present study were chosen to be clinically relevant: 3 and 7 h pi correspond to same-day imaging while 24 h pi corresponds to next-day imaging.

Due to the slow clearance of radioactivity from blood and normal organs, the tumor-to-organ ratios were not significantly improved between 3 and 7 h pi. At the later time-point of 24 h pi, the faster clearance from normal tissues compared to tumors resulted in significantly higher tumor-to-organ ratios that enhanced imaging contrast. In contrast to the results obtained for 57Co-DOTA-ZEGFR:2377 in a high EGFR-expressing model (33), the conjugate did not provide positive contrast towards the liver. However, this is not critical for PCa where liver metastases are not common. Still, the uptake in tumors after the injection of 10 µg 57Co-DOTA-ZEGFR:2377 exceeded the uptake in all other studied organs with the exception of the liver and kidneys. For organs clinically relevant for the detection of PCa lesions, tumor-to-organ ratios at 24 h pi were 2.2±0.5 for the colon, 7±2 for muscle, and 4.0±0.7 for bones. Therefore 57Co-DOTA-ZEGFR:2377 was successful in visualizing the EGFR-expression in PCa as early as 3 h pi and imaging contrast improved with time (Fig. 4). Previously it was demonstrated that a single-dose injection of anti-EGFR affibody molecule produced no pathological evidence of toxicity in rats at a dose level of 24,490 µg/kg (corresponding to a 1,000× equivalent human microdose level) (47). Dosimetry estimated using clinical data for the anti-HER2 affibody molecule, a tracer with a comparable kidney uptake, revealed that the absorbed dose to kidneys and liver, after a 100-MBq administration (Ga-68), would be ~40 and 15 mGy, respectively, which is less than the maximum allowed absorbed dose of 50 mGy to a single organ in a healthy adult research subject. The administration of ~200 MBq would give an effective dose in the range of 5–6 mSv (lower than for typical 18F-FDG PET examinations, which often give effective doses of around 7 mSv (48). We may expect that for the anti-EGFR imaging probe the absorbed doses for kidneys should be similar and the absorbed dose for liver should be somewhat higher than that for the anti-HER2 imaging probe.

In conclusion, radiocobalt-labeled DOTA-ZEGFR:2377 allowed the successful visualization of EGFR in vivo in a PCa pre-clinical model and could be used for monitoring of receptor occupancy during cetuximab therapy. The present study also emphasized the importance of finding optimal protein dosing for tumors with low EGFR expression in order to ensure a higher sensitivity in clinical imaging.

Acknowledgements

Not applicable.

Funding

The present study was supported by the Swedish Cancer Society [grants CAN 2017/649 (JL), CAN2013/586 and CAN 2016/463 (SS), CAN2014/474, CAN 2017/425 (AO), and CAN2015/350 (VT)], the Swedish Research Council [grants 621-2012-5236 (SS), 2015-02509 (AO), and 2015-02353 (VT)], the Swedish Agency for Innovation VINNOVA [grants 2016-04060 (AO) and 2017-02015 (SS, JL)] and the Wallenberg Center for Protein Technology (SS, JL). The molecular imaging work in this publication was supported by the Wallenberg infrastructure for PET-MRI (WIPPET) at SciLifeLab Pilot Facility for Preclinical PET-MRI, a Swedish nationally available imaging platform at Uppsala University, Sweden, financed by the Knut and Alice Wallenberg Foundation (SPECT/CT).

Availability of data and materials

The datasets used during the present study are available from the corresponding author upon reasonable request.

Authors' contributions

BM participated in the labeling of proteins, the in vitro and in vivo characterization, the data acquisition, data interpretation and drafted the first version of manuscript; KGA and JG performed the production and the following biochemical, biophysical and in vitro characterization of conjugate; EL participated in the labeling of proteins, the in vitro and in vivo characterization, the data acquisition and data interpretation; MR participated in the in vivo characterization and the data acquisition; VT and AO participated in the molecular and study design, in the in vivo characterization, the data acquisition, data interpretation, drafted the relevant parts of the manuscript and critically revised the manuscript; SS and JL participated in the molecular and study design, the data acquisition, data interpretation and critically revised the manuscript. All authors read and approved the manuscript and agree to be accountable for all aspects of the research in ensuring that the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics approval and consent to participate

All animal experiments were planned and performed in accordance with the Swedish national legislation on the protection of laboratory animals and the study plans were approved by the Ethics Committee for Animal Research in Uppsala.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References