In vivo efficacy of melanoma internal radionuclide therapy with a 131I-labelled melanin-targeting heteroarylcarboxamide molecule


Correspondence to: Mathilde Bonnet, IUT, Université d'Auvergne, Aubière F-63172, France, Tel.:+334–7317-8376, E-mail:


The development of alternative therapies for melanoma treatment is of great interest as long-term tumour regression is not achieved with new targeted chemotherapies on selected patients. We previously demonstrated that radioiodinated heteroarylcarboxamide ([131I]ICF01012) induced a strong anti-tumoural effect by inhibiting both primary tumour growth and dissemination process in a B16BL6 melanoma model. In our study, we show that a single injection of [131I]ICF01012 (ranging from 14.8 to 22.2 MBq) was effective and associated with low and transient haematological toxicity. Concerning pigmented organs, cutaneous melanocytes and skin were undamaged. In 30% of treated animals, no histological alteration of retina was observed, and in the remaining 70%, damages were restricted to the optic nerve area. Using the Medical Internal Radiation Dose methodology, we determined that the absorbed dose in major organs is very low (<4 Gy) and that a delivery of 30 Gy to the tumour is sufficient for an effective anti-tumoural response. Molecular analyses of treated tumours showed a strong radiobiological effect with a decrease in proliferation, survival and pro-angiogenic-related markers and an increase in tumour suppressor gene expression, melanogenesis and anti-angiogenic markers. All these features are in accordance with a tumour cell death mechanism that mainly occurs by mitotic catastrophe and provide a better understanding of in vivo anti-tumoural effects of [131I] radionuclide. Our findings raise [131I]ICF01012 a good candidate for disseminated melanoma treatment and strongly support transfer of [131I]ICF01012 to clinical trial.

Metastatic melanoma has a poor prognosis with an estimated death rate ranging from 1.8 to 3.5 per 100,000 cases worldwide.[1, 2] As with other cancers, chemotherapy (dacarbazine) remained the conventional treatment with poor benefit for over half a century. Two strategies have improved melanoma treatment, one using the tumour immune response by blocking cytotoxic T-lymphocyte activation (antibody anti-CTL4)[3] and the other targeting BRAF.[4] A recent multicentre phase 1 clinical trial testing a mutated BRAF (V600E) inhibitor showed complete or partial tumour regression in the majority of patients.[5] However, BRAF mutations are detected in only 60% of melanoma cases, and the therapy involves selection of patients with tumours that carry the V600E BRAF mutation.[5, 6] Furthermore, most treated patients acquire resistance to this inhibitor after initial clinical response.[7] Metastatic melanoma is always described as a refractory disease, and new observations suggest that combined therapies will have a greater impact on melanoma residual activity. As the mechanisms of resistance to treatment rely on tumour adaptation by somatic mutations,[8] other non-protein targets such as melanin could be considered for future therapies.

Melanin pigment is detected in more than 90% of primary melanoma cases, and thus, a strategy targeting this pigment could be used to treat a large number of patients.[9] Melanin-targeted radionuclide therapy (TRT) could be an effective alternative to conventional therapeutic agents. Although melanoma is described as a radioresistant tumour, radiations have proven effective in clinical applications such as brachytherapy for choroidal melanoma treatment[10] and palliation with external beam radiation for patients with metastatic melanoma.[11, 12] In addition, preclinical studies have demonstrated significant anti-tumoural effect of internal delivery of radionuclide. Peptide analogues of α-melanocyte-stimulating hormone targeting melanocortin-1 receptor have been developed, and promising experimental results have been reported.[13, 14] Radioimmunotherapy using monoclonal antibody against melanin induces significant anti-tumoural effects.[15, 16] However, considerable renal uptake occurs with these strategies and may result in a renal toxicity. Methylene blue, which interacts with melanin, was labelled with 131I or 211At and gave some promising results but without further clinical development.[17] Recently, strong anti-tumoural effects were obtained with a 131I-labelled benzamide derivative ([131I]MIP-1145) on nude mice bearing xenograft tumours.[18] Using murine and human xenograft melanoma models, we selected an iodinated quinoxaline derivative molecule, N-(2-diethylaminoethyl)−6-iodoquinoxaline-2-carboxamide dihydrochloride salt (ICF01012) for its high anti-tumoural efficacy after labelling with 131I radionuclide.[19-21] The aim of the our work was to study the toxicity of the [131I]ICF01012 TRT focusing on blood and normal pigmented organ analysis in a syngenic pigmented preclinical model bearing B16BL6 tumours. In parallel, anti-tumoural efficacy was evaluated, and dosimetric parameters were extrapolated in tumours and in the major organs from biodistribution parameters of the tracer, 131I radionuclide properties and size of the treated tumours. Because the mechanisms by which 131I-irradiation slow down solid tumour growth are not fully understood, we considered some molecular tumoural events that initiate anti-tumoural efficacy with transcriptomic, proteomic and metabolomic analyses.

Material and Methods

Preparation of radioiodinated [131I]ICF01012

N-(2-Diethylaminoethyl)−6-iodoquinoxaline-2-carboxamide dihydrochloride salt (ICF01012) and tributylstannyl precursor N-(2-diethylaminoethyl)−6-(tributylstannyl)quinoxaline-2-carboxamide were prepared according to a previously described method.[21] The compound [131I]ICF01012 was labelled at a high specific activity using radioiododestannylation reaction with [131I]NaI (1.49 GBq; Perkin Elmer, Courtaboeuf, France) as previously reported.[20]

Cell culture

The transplantable B16BL6 melanoma cells derived from C57BL/6J mice were kindly provided by Dr. Fidler's laboratory (Texas University, Houston, TX). Cells were maintained as monolayers using culture medium consisting of Eagle's MEM-glutaMAX medium (Invitrogen, Cergy Pontoise, France) supplemented with 10% foetal calf serum (Biowest, Nuaillé, France), 1% vitamins, 1 mM non-essential amino acids, 1 mM sodium pyruvate and 4 µg µl−1 gentamycin (Invitrogen) at 37°C in a humidified incubator containing 5% CO2.


All experiments were carried out in accordance with the institutional recommendations based on the Guide for the Care and Use of Laboratory Animals (European directive 86/609/EEC). For all experiments, we used 6- to 8-week-old C57BL/6J male mice (Charles River Laboratories, L'Abresle, France).

Therapy experiment on C57BL/6J mice bearing B16BL6 tumours

For radionuclide therapy experiments, 6-to 8-week old C57BL/6J male mice anaesthetised by isoflurane inhalation were inoculated with 3 × 105 melanoma B16BL6 cells by dorsal subcutaneous injection at Day 0 of the experiment. Various protocols were tested involving a total of 111 mice. Different series of experiments were performed testing one intravenous injection with [131I]ICF01012 at Day 6 with low (11 ± 2 MBq; n = 15) or moderate (20 ± 1 MBq; n = 14) activity. Other animals (n = 18) were treated with two intravenous injections at Days 6 and 10 with a high activity (37 ± 6 MBq). In the same conditions, untreated groups were used as controls. Pre-treatment with lugol's iodine solution was added in alimentation of all group mice to block thyroid function. To monitor potential toxicity, body weight was measured twice a week. To monitor tumoural growth, tumour volume in millimetre cubed was calculated twice a week from the measurement of two perpendicular diameters using a calliper according to the formula L × S2/2, where L and S are the largest and smallest diameters in millimetres, respectively. For each animal, tumour doubling time (DT) was determined to evaluate tumoural growth. We defined the effective anti-tumoural response threshold corresponding to untreated DT mean plus two standard deviations.

The mice were sacrificed at Day 20 of the experiment (14 days post-treatment). The blood of each animal was removed and processed on a Melet Schloesing MS9-5 Hematology Analyzer to determine the different haematological parameters (IBiSA INTRAGENE platform, CNRS UMR 6218, Orléans, France). Tumours were removed and prepared for histology and molecular biology experiments. To evaluate the number of animals with dissemination, metastases (pigmented cells) were sought on each lung after macroscopic observations. In addition, metastases were counted on each lung. Eyes and skin fragments (remote or near the tumour) were collected, fixed and prepared for histology. For toxicological analysis on normal pigmented tissues (skin and retina) and on blood, an additional series of mice (n = 22) was sacrificed at Day 27 (21 days post-treatment). For mechanistic studies, an additional series of mice (n = 25) was sacrificed at 24 and 72 hr post-injection (p.i.) with [131I]ICF01012 (moderated activity).

Calculation of absorbed dose to the tumour and organs

Dose calculations based on the Medical Internal Radiation Dose (MIRD) methodology were obtained from [125I]ICF01012 kinetic data previously published[21] and were extrapolated to [131I]ICF01012. The absorbed dose was calculated for each organ using the cumulative activity A, the equilibrium dose constant of iodine 131, the absorbed dose fractions calculated according to the nature and energy of its emissions, the geometric volumes of the organs and tumour and considering tissue densities equal to 1 g cm−3. To take into account the inhibition of tumoural growth using [131I]ICF01012 compared to [125I]ICF01012, we modified the calculation of the biological period for the tumour correcting the biological decay constant math formula by the differential growth constant math formula. The values of equilibrium dose constants were extracted from the MIRD Pamphlet No. 11.[22] The tumour and organs were modelled using spherical volumes with an average diameter of 7 mm ± 5 mm. The uvea was defined by the envelope of 3.5-mm-diameter sphere with a thickness of 0.5 mm ± 0.25 mm. Self-absorption was calculated taking into account the energy deposited in the volumes of interest coming from electron collisions and photon attenuation. Inter-organ radiations were considered as negligible. The dose D for each organ (Gy) was calculated using the following equation:

display math

where A is the cumulative activity per mass (Bq s kg−1), math formula is the dose constant and math formula is the absorbed energy fraction (J Bq−1 s−1) for each i emission.

Histological analyses

Normal skin and melanoma tumour pieces were fixed in formol solution. Paraffin-embedded sections were cut into 4-µm slices, and tissue sections were prepared for haematoxylin–phloxin saffron staining and routine pathological analysis. An anti-PS100 antibody (Dako, Les Ulis, France) was used to stain melanocytes. The secondary horseradish peroxidase antibody was revealed with diaminobenzidine. The number of melanocytes was quantified as the number of PS100-positive cells per field. At least three fields per slide were counted by a pathologist.

Eyes were placed in fixative solution and embedded in paraffin. Sections of 5 µm were cut and prepared for haematoxylin–eosin saffron staining. The thicknesses of the retina pigment epithelium (RPE) and that of photoreceptor (PR) were measured both near to and remote from the optic nerve using metamorph software from CICS platform (Clermont-Ferrand, France).[23]

Flow cytometry DNA analysis

B16BL6 tumour samples were mechanically disaggregated in phosphate buffered saline by fine mincing with 26-G needles and filtered through a 200-µm nylon filter. The suspensions of cells were spun (400 g, 8 min, 4°C), and the dry pellets obtained were stored in liquid nitrogen. After thawing, the tumour cell extracts were resuspended in 500 µl of ribonuclease A (1 mg ml−1; Sigma-Aldrich, France), and then 500 µl of propidium iodide (0.1 mg ml−1; Sigma-Aldrich) was added. After 30 min at 4°C in the dark, the cell cycle was analysed using a flow cytometer (CoulterEpics XL, Coulter, Hialeah, FL) at wavelengths of 488 nm (excitation) and 620 nm (emission).

RNA extraction, reverse transcription and TaqMan low-density arrays

Total RNAs were extracted from B16BL6 tumours using the RNeasy Fibrous Tissue Mini Kit (Qiagen, Courtaboeuf, France) according to the manufacturer's instructions. RNA samples were subjected to reverse transcription using the High-Capacity cDNA Reverse Transcription Kit and non-specific random hexamer primers (Applied Biosystems, Courtaboeuf, France) at 37°C for 120 min. Pre-designed TaqMan probe and primer sets for 45 target genes were chosen from an online catalogue (Applied Biosystems). Gene families were cell cycle/apoptosis (n = 14), transcription factor and cell signalling-related genes (n = 13), oxidative stress (n = 9), melanogenesis-related genes (n = 2), others (n = 5) and housekeeping genes (n = 3) (see Supporting Information Table). Once selected, the sets were factory loaded into the 384 wells of TaqMan low-density arrays (TLDAs). Samples were run and analysed using the 7900HT system with a TLDA Upgrade (Applied Biosystems) according to the manufacturer's instructions. Gene expression values were calculated by the ΔΔCt method. Expression levels relative quantities (RQ) of target genes were normalised to those of three housekeeping genes (S18 rRNA, gapdh and rpl32). For each gene, expression on treated tumours was compared to that of untreated tumours.

Western blot analysis and enzyme-linked immunosorbent assay

Proteins were extracted from 100 mg of crushed tumour tissue in lysis buffer (1 mM EDTA; 0.5% Triton X-100, 5 mM NaF, 6 M urea, 1 mM activated sodium orthovanadate and 2.5 mM sodium pryrophosphate). For Western blot analysis, 50 µg of total proteins from tumour extracts was loaded onto 12.5% acrylamide gels for SDS-PAGE. Proteins were separated, transferred to nitrocellulose membranes (Millipore, St. Quentin-en-Yvelines, France), stained with Ponceau red solution and probed with specific antibodies. The antibody–antigen binding was detected as previously described.[20] All the antibodies used were obtained from Cell Signaling Technology (Boston, MA).

Protein levels were measured by enzyme-linked immunosorbent assay (ELISA). Phospho-Akt1 (S473, Duoset ELISA; R&D Systems, Abingdon, UK) and phospho-ERK2 (T185/Y187, Duoset ELISA; R&D Systems), vascular endothelial growth factor (VEGF; Raybiotech, Tebu-bio SA, Le Perray en Yvelines, France) and interferon-γ-inducible protein 10 (IP10; Duoset ELISA; R&D systems) assays were performed from B16 tumoural protein extracts according to the manufacturer's instructions.

1H HRMAS NMR analysis of B16 tumour samples

1H NMR spectroscopy was performed on a small-bore Bruker DRX 500 magnet (Bruker, Karlsruhe, Germany) equipped with an HRMAS probe. All experiments were performed on B16BL6 tumour samples flushed for a few seconds in cooled D2O. Samples were then set into 4-mm-diameter, 50-µl free volume ZrO2 rotor tubes without upper spacer. A total of 3 µl of D2O containing 0.75% 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid (D2O+TSP; Sigma) was added to the rotor tubes to lock the spectrometer. Rotors were spun at 277 K at 4 kHz to keep the rotation sidebands out of the acquisition window. One-dimensional 1H NMR sequence was a saturation recovery sequence. Peak referencing was obtained on the signal of TSP. The resonances were assigned based on the known chemical shifts of the major structural groups using standard tissues data.[24-26] The relative content of metabolites was estimated by peak area integration as previously described[26] using the Topspin software (version 1.3; Bruker). The one-dimensional 1H NMR spectrum was immediately followed by a two-dimensional spectrum. The sequence used was a two-dimensional total correlation spectroscopy sequence with water signal suppression.[26] The two-dimensional spectrum duration was 1.5 hr, a duration in the magnet at 277 K. Percentage variations of metabolites were calculated between results from the untreated B16BL6 tumour group and the treated groups.

Statistical analyses

Data are reported as mean ± SD. Statistical analyses were performed using GraphPad Prism5 or R software ( Results were analysed by ANOVA followed by Tukey's post-test. For analysis of data with multiple comparisons, a corrected p-value with application of the Bonferroni-Holm procedure correction was calculated. The proportions of animals with lung metastasis were compared to Fisher's exact tests. Correlation studies were performed using Pearson's correlation method. We considered p values < 0.05 to indicate statistical significance.


Anti-tumoural dose effect of TRT

To determine the best treatment protocol, different series of experiments were performed comparing one intravenous injection of [131I]ICF01012 to low- or moderate-activity injection and two injections (high-activity injection). No significant change in body weight or signs of distress was observed with [131I]ICF01012 treatment (Supporting Information Fig. A). Macroscopic observation of mice showed no damage on the coat and skin and no variation in their pigmentation (Supporting Information Fig. B). The effect of the treatment on the growth rate of B16BL6 tumours in a representative experiment is shown in Figure 1a. B16BL6 tumours in the untreated group showed exponential growth with 2.3 ± 0.5 days as DT. DT was determined as 4.9 ± 2.3 and 5.7 ± 2.9 days for moderate- and high-activity injection experiments, respectively, which shows a similar significant inhibition of B16BL6 tumoural growth with both protocols. The results of all experiments (n = 111 mice) given in Figure 1b show a dose-dependent effect of [131I]ICF01012 on tumour DT of treated animals and confirm the significant efficacy of moderate-activity (p < 0.001) and high-activity (p < 0.001) injections in comparison to low injected activity, which induced no significant anti-tumoural effect. Efficacy was not dependent on initial tumour size at the beginning of treatment. (Supporting Information Fig. C). B16BL6 cells can generate spontaneous lung metastases from primary subcutaneous tumour.[27] The proportion of animals with metastases and the number of metastases per animal in untreated or treated mice were analysed (Supporting Information Fig. D). Interestingly, 54% (22/41) of the control mice bearing B16BL6 tumours had lung metastases at 14 days p.i., whereas no metastasis was observed on [131I]ICF01012 treated with both moderate (0/14) and high injected activity (0/18), showing significant effect on dissemination (p = 0.0004). In an additional study, no metastasis was observed in the lung of treated animals (moderate activity; 0/13) at 21 days p.i., whereas metastases had developed in 57% of the control group (p = 0.009; Supporting Information Fig. D).

Figure 1.

Effect of [131I]ICF01012-targeted radionuclide therapy on growth rate of B16BL6 murine melanoma tumours. (a) Representative experiment obtained with moderate (n = 8) and high (n = 7) injected activities of [131I]ICF01012. Mice were inoculated with 3 × 105 melanoma B16BL6 cells by dorsal subcutaneous injection at Day 0 of the experiment. In comparison with untreated mice (n = 10), [131I]ICF01012 treatment significantly slowed B16BL6 tumoural growth (p = 0.001). (b) Significant dose-dependent effect of [131I]ICF01012 therapy on tumoural growth. Different series of experiments were performed comparing intravenous injection to low (n = 15) and moderate (n = 14) activity injections and two injections (high activity; n = 18) for a total of 111 mice tested. For each animal, tumour volume in millimetre cubed was evaluated twice a week, and tumour DT was determined to evaluate tumoural growth. DT was significantly increased (p < 0.0001) for moderate (+100%) and high (+95%) injected activities, whereas a weak (+26%) and non-significant effect was observed for low activity. We specified the effective anti-tumoural response threshold corresponding to untreated DT mean plus two standard deviations. Significant response on DT was observed in 26 (4/15), 71 (10/14) and 67% (12/18) of treated animal with low, moderate and high injected activity, respectively. (c) Strong correlation between DT of treated animals and tumoural delivered dose. Tumoural delivered doses were evaluated on tumours with the MIRD model using complete biodistribution data of the [125I]ICF01012 tracer as previously described[21] and taking into account tumoural volume after [131I]ICF01012 treatment. The delivered dose in tumours was then extrapolated for all tested protocols with [131I]ICF01012. Correlation was observed between delivered dose and tumoural growth inhibition (Pearson's correlation factor: 0.869; p < 0.001). Under a delivered dose of 30 Gy, radiotherapy was not effective with DT under effective anti-tumoural response threshold (dotted line).

Minimal toxicity of targeted radiotherapy in non-target and normal pigmented organs

We evaluated dose distribution on tumours and non-target organs on the basis of using the MIRD committee model and a simplifying hypothesis, that is, homogeneous distribution of the dose in basic geometric volumes. We determined the biological and effective half-life of tracer from complete [125I]ICF01012 biodistribution data determined by quantitative whole-body autoradiography previously described.[21] Because [131I]ICF01012 treatment modified tumoural growth, biological half-life in the tumour was extrapolated by taking into account the tumour volume determined after treatment in these experiments. The dose delivered to the tumours and organs was then extrapolated for each animal injected with moderate activity. Delivered dose was low (<4 Gy) in all non-target organs (Table 1). Interestingly, the dose was very low (0.2 Gy) in the brain, which contains neuromelanin (Table 1). Significant dose was delivered with the treatment on pigmented tissue such as uvea (36 Gy) and tumour (54 Gy), confirming the high melanin specificity of ICF01012 tracer (Table 1). In addition, for all protocols with [131I]ICF01012, the tumoural delivered dose was calculated and correlated with anti-tumoural response (Fig. 1c). A strong correlation was obtained between DT and tumoural delivered dose (p < 0.001). Below a delivered dose of 30 Gy, radiotherapy was not effective.

Table 1. Projected organ's delivered dose extrapolated for moderate-activity injection protocol
 Effective half-life (hr)Biological half-life (hr)Delivered dose (Gy)
  1. In the major organs, we determined the biological and effective half-life of tracer from complete [125I]ICF01012 biodistribution data determined by quantitative whole-body autoradiography as previously published.[21] Because [131I]ICF01012 treatment modified tumoural growth, biological half-life in the tumours was extrapolated by taking into account tumoural volume after treatment determined on these preclinical TRT experiments. We estimated dose distribution on tumours and non-target organs using the MIRD committee model.

Blood8.28.60.7 ± 0.03
Bone marrow1.41.40.3 ± 0.01
Muscle4.44.50.2 ± 0.01
Spleen4.04.10.7 ± 0.03
Lung4.44.51.0 ± 0.04
Kidney5.05.11.9 ± 0.08
Stomach5.15.23.8 ± 0.16
Liver7.67.92.4 ± 0.10
Intestinal content7.98.33.5 ± 0.15
Pancreas4.04.10.7 ± 0.03
Brain5.75.80.2 ± 0.01
Uvea112.9273.237.6 ± 1.59
Melanoma tumour84.5150.654.0 ± 10.02

The haematological toxicity of the [131I]ICF01012 treatment was evaluated by measuring platelet and white blood cell (WBC) counts at Days 14 and 21 p.i. No variation in platelet level was observed (Fig. 2a). A significant decrease in WBC (−52%) was measured in the blood of treated mice at 14 days p.i., whereas no significant variation was observed at 21 days p.i. (Fig. 2a). No radiation damage to the melanised skin of treated mice was histologically detected (Fig. 2b). Figure 2b shows representative melanocyte-specific staining of PS100 protein on untreated or treated mice skin. No variation in melanocyte numbers in the epidermis of controls (1.1 ± 0.7 PS100-positive-cells per field) and treated mice (1.25 ± 0.6 PS100-positive-cells per field) was observed at Day 14 p.i. (p = 0.79). The same results were obtained at 24 and 72 hr and 21 days p.i. To evaluate toxicity in eyes, histological analyses were performed on pigmented areas. As no histological damage was detected on the ciliary body and choroid structures (Supporting Information Fig. E), we focused our study on the retina. The thicknesses of RPE and PR (RPE+PR) were measured and normalised by total thickness of the retina in both optic nerve and peripheral areas. No damage was observed in any of the treated animals in the peripheral area of the retina (Fig. 2c). In the optic nerve area of the same animals, we observed some histological damage with a decrease in RPE+PR thickness (−34%; Figs. 2c and 2d). Interestingly, in 30% of the treated animals, no damage was observed on both optic nerve and peripheral areas (Fig. 2c).

Figure 2.

Low toxicity of targeted radiotherapy in non-target and normal pigmented organs. (a) Haematological toxicity of the [131I]ICF01012 treatment (moderate activity) by measuring platelet (below) and white blood cell (WBC) (above) with a Melet Schloesing MS9-5 Hematology Analyzer. Platelet and WBC level were evaluated 14 and 21 days post-treatment, respectively. These counts were normalised at each time with count in blood on untreated mice bearing B16 tumours. Treatment induced no significant variation in platelet level. A significant decrease in WBC (−52%; p = 0.02) was measured in treated mice at 14 days p.i., whereas no significant variation (p = 0.230) was observed at 21 days p.i. (b) Evaluation of skin toxicity by melanocyte-specific staining of PS100 protein. Representative histological section of skin of treated (a′ and b′) and untreated (c′) black mice at 14 days p.i. (moderate activity) showed a normal morphology of epidermis and a specific staining of melanocytes. An equivalent number of melanocytes was counted on the epidermis of both untreated and treated mice at 14 days p.i. The same results were observed at 24 and 72 hr and 21 days p.i. B16 melanoma tumours were used as PS100 staining positive control (d′). (c and d) Evaluation of retina toxicity on highly pigmented eyes at 14 days p.i. (c) Representative histological images of retina on the nerve optic area of untreated or two mice treated with moderate activity at 14 days p.i. RPE: retina pigment epithelium; PR: photoreceptor. High level of melanin was observed on RPE. In 30% of treated animals, no damage was observed (left image), whereas PR thickness layer was affected in the others. (d) RPE and PR thicknesses (RPE+PR) were measured and normalised by total thickness in both the optic nerve and peripheral areas. For all treated animals, no damage was observed in the peripheral area of the retina, whereas a significant decrease was observed in the optic nerve area (−34%; p < 0.01).

In vivo anti-tumoural mechanisms of targeted radiotherapy

To assess the effects of TRT, the B16BL6 tumours from untreated and [131I]ICF01012-treated mice (moderate activity) were removed 24 and 72 hr after treatment and analysed histologically. For untreated mice, the B16BL6 tumours had a high mitotic index, a low number of infiltrating lymphocytes and a large vascularisation (more than one vessel per field with a magnification of 100; Fig. 3a-a′). In contrast, the tumours of treated mice exhibited atypical histological zones and fewer mitotic figures (Fig. 3a-b′). No difference was observed for vascularisation and pigmentation. Additional histological analyses were performed on tumours at 14 days p.i. and confirmed previously observed results.[20] Treated tumours exhibited atypical histological zones with anisocytosis, fewer mitotic figures, a decrease in micro-vessel content and an abundant extracellular melanin deposit (Fig. 3a-d′). No difference in necrosis and infiltrating lymphocyte number was observed. These findings indicate that the treated tumours acquired a loss of aggressiveness.

Figure 3.

Mechanistic studies of anti-tumoural effect of TRT. (a) Representative histological sections of B16BL6 tumours (×400). Untreated (a′ and c′) and [131I]ICF01012-treated tumours (b′ and d′) were examined at 24 hr (a′ and b′) and 14 days (c′ and d′) p.i. Tissues were stained with haematoxylin and eosin. For untreated mice, the B16BL6 tumours had a high mitotic index, a low number of infiltrating lymphocytes and significant vascularisation (a′ and c′). Twenty-four hours after [131I]ICF01012 treatment (moderate activity), tumours exhibited atypical histological zones and fewer mitotic figures (b′). No difference was observed in vascularisation and pigmentation (b′). The same results were observed at 72 hr p.i. Fourteen days after treatment, tumours exhibited atypical histological zones with anisocytosis, fewer mitotic figures, a decrease in micro-vessel content and an abundant extracellular melanin deposit (d′). No difference in necrosis and the number of infiltrating lymphocytes was observed (d′). These findings indicate that treated tumours acquired a loss of aggressiveness. (b) Cell cycle modifications were induced by [131I]ICF01012 treatment at 24 and 72 hr p.i. In untreated tumours, 28% ± 3% of cells accumulated in the S phase reflecting the high proliferative index of B16BL6 tumours. Treatment led to a significant increase in cell accumulation in the G2 phases of +400% and +280% at 24 and 72 hr p.i., respectively. A significant increase in the sub-G1 content was observed in tumours at both 24 (+72%) and 72 hr post-treatment (+133%) indicating DNA fragmentation. *p < 0.05. (c) Representative Western blot analysis of PCNA, p53 phosphorylation (S15), PTEN phosphorylation (S380/T382/383), Atg12 and PARP expressions in untreated and tumours at 72 hr p.i. A decrease in PCNA, an increase in S15p53 and PTEN, no variation of Atg12 and no cleavage of PARP were observed. (d) Quantifications of protein expression were done in comparison to the total transferred protein quantity determined on nitrocellulose membrane after Ponceau Red staining. Signal quantifications were assessed using Quantity One Version 4.4 software (BioRad, Marnes-la-Coquette, France). PCNA level expression was significantly decreased at 24 hr (−62%) and 72 hr (−36%) p.i. In untreated tumours, low expression of S15p53 was observed, whereas a strong increase was observed at both 24 and 72 hr p.i. A significant increase in PTEN phosphorylation was recorded at 72 hr p.i. (e) MAPK pathway was explored on untreated and treated tumours by the sensitive ELISA methods. A strong decrease was observed for both pAkt (−52%; p = 0.04) and pERK2 (−48%; p = 0.02) at 24 hr p.i. No significant variation was observed at 72 hr p.i. Angiogenesis-related proteins VEGF and IP10 were also assayed by ELISA on untreated and treated tumours (24 and 72 hr and 14 days p.i.). A significant decrease of VEGF was observed at 24 (−33%) and 72 hr (−46%) and 14 days p.i. (−54%). For IP10 protein, a significant increase (+100%) was detected on treated tumours at 14 days p.i. (p < 0.05).

Cell cycle modifications were evaluated on untreated and [131I]ICF01012-treated tumours. Figure 3b shows that in controls, 28% of tumoural cells accumulated in the S phase, hence reflecting a high proliferative index of B16BL6 tumours. The treatment led to a significant increase in cell accumulation in the G2 phases of +400% and +280% at 24 and 72 hr p.i., respectively. Furthermore, we observed a significant increase in the sub-G1 content in tumours at both 24 (+72%) and 72 hr (+133%) p.i.

All tumours were subjected to gene expression profile analysis by a high-throughput quantitative RT-PCR assay, with a specifically designed TLDA. In comparison to untreated tumours, 7 and 18 genes were up-regulated at 24 and 72 hr p.i., respectively (Table 2 and Supporting Information). Few genes (pcna and cdkn2a) were down-regulated after treatment. For cell cycle/apoptosis-related genes, 9/14 genes were regulated by radiotherapy. The catalytic subunit of telomerase transcript was significantly increased, and pcna transcript, a well-known proliferation marker, was significantly down-regulated. The expression of cell cycle-related transcripts was either up-regulated (cdkn1a and cyclinD1) or down-regulated (cdkn2a). Interestingly, tumour suppressor gene Rb was up-regulated at 72 hr p.i. (+80%). Apoptosis-related genes (bax and bcl2-like 1) were up-regulated, whereas bcl2, caspase3 and Fasl levels were unchanged after treatment. Transcript encoding p53 was unchanged; however, its regulator mdm2 was significantly increased. A subset of transcription factors and cell signalling-related genes (jun, relb and pten) were also observed to be up-regulated in the treated tumours at 24 and/or 72 hr p.i. Moreover, the tyrosinase gene involved in melanin biosynthesis exhibited a marked increase in transcript levels in the treated tumours at 72 hr p.i., whereas the mitf transcript was not dysregulated. Several genes encoding antioxidant defences (gpx1, hmox1 and sod3) were up-regulated after treatment. A strong increase in the expression of mgmt and plar2 genes was observed.

Table 2. Overview of significant gene expression changes after TRT treatment using TLDA transcriptomic analysis
   Untreated24h p.i72h p.i
Genes Accession numberRQ ± SDRQ ± SDFold regulationRQ ± SDFold regulation
  1. For each gene, the expression on treated tumours was compared to that of untreated tumours and was considered statistically significant with p < 0.05. NS: no significant variation.

Cell cycle/Apoptosis
Bcl2-associated X proteinBaxnm_007527.21,01 ± 0,111,32 ± 0,171,311,79 ± 0,351,78
Bcl2-like 1Bcl2l1nm_009743.40,69 ± 0,140,86 ± 0,13NS2,01 ± 0,642,93
Cyclin D1Ccnd1nm_007631.10,80 ± 0,121,03 ± 0,15NS2,79 ± 0,683,49
Cyclin-dependent kinase inhibitor 1A (P21)Cdkn1anm_007669.20,59 ± 0,132,05 ± 0,323,503,53 ± 1,076,02
Cyclin-dependent kinase inhibitor 2ACdkn2anm_009877.21,09 ± 0,171,10 ± 0,32NS0,70 ± 0,300,64
Transformed mouse 3T3 cell double minute 2Mdm2nm_010786.20,64 ± 0,071,16 ± 0,201,821,79 ± 0,452,80
Proliferating cell nuclear antigenPcnanm_011045.10,78 ± 0,110,59 ± 0,090,760,61 ± 0,090,78
Retinoblastoma 1Rb1nm_009029.10,94 ± 0,171,06 ± 0,15NS1,68 ± 0,481,79
Telomerase reverse transcriptaseTertnm_009354.10,79 ± 0,111,08 ± 0,121,361,37 ± 0,421,73
Transcription factor/Cell signalling related genes     
Jun oncogeneJunnm_010591.10,58 ± 0,100,60 ± 0,08NS1,06 ± 0,301,82
Phosphatase and tensin homologPtennm_008960.20,71  ± 0,050,87 ± 0,14NS1,20 ± 0,301,69
Avian reticuloendotheliosis viral (v-rel) oncogene related BRelbnm_009046.20,76 ± 0,100,77 ± 0,11NS1,31 ± 0,251,72
Melanogenesis related genes
TyrosinaseTyrnm_011661.10,98 ± 0,151,14 ± 0,17NS2,14 ± 0,432,17
Antioxidant defence
Glutathione peroxidase 1Gpx1nm_008160.20,70 ± 0,100,76 ± 0,08NS1,30 ± 0,261,85
Heme oxygenase (decycling) 1Hmox1nm_010442.11,06 ± 0,530,83 ± 0,16NS1,74 ± 0,081,64
Superoxide dismutase 3, extracellularSod3nm_011435.30,68 ± 0,150,91  ± 0,20NS2,45 ± 0,563,60
O−6-Methylguanine-DNA methyltransferaseMgmtnm_008598.10,94 ± 0,332,44 ± 0,942,601,95 ± 0,802,08
Phospholipase A2 receptor 1Pla2r1nm_008867.10,77 ±0,091,31 ±0,151,692,58 ±0,493,34

Protein expression levels were analysed by Western blot or ELISA methods (Figs. 3c3e). [131I]ICF01012 treatment significantly decreased PCNA protein levels in treated tumours at both 24 (−62%; p = 0.003) and 72 hr (−36%; p = 0.04) p.i. (Figs. 3c and 3d). A decrease of β-tubulin level (−65%) was observed in treated tumours compared to untreated tumours (data not shown). In untreated tumours, a low level of phosphorylation of p53 at Serine 15 (S15p53) was observed, whereas a strong increase was observed at both 24 and 72 hr post-treatment with [131I]ICF01012 (Figs. 3c and 3d). Likewise, an increase in PTEN phosphorylation (S380/T382/383) was observed at 72 hr p.i. (Figs. 3c and 3d). No variation was observed for the autophagic marker Atg12 (Fig. 3c). No cleavage of PARP was observed in treated tumours at either 24 or 72 hr p.i. (Fig. 3c). Activation of MAPK and Pi3K signalling pathways was assessed by analysing the level of AKT1 (S473) and ERK2 (T185/Y187) phosphorylations, which are major events of these pathways in B16 models. A great decrease was observed in the levels of both pAkt (−52%; p = 0.04) and pERK2 (−48%; p = 0.02) at 24 hr p.i. (Fig. 3e). A significant decrease in pro-angiogenic VEGF protein level was observed at 24 (−33%) and 72 hr (−46%) p.i. (Fig. 3e). Additional analyses were performed on tumour at Day 14 p.i. We observed a significant increase (+100%) in anti-angiogenic IP10 marker after treatment and a decrease (−54%) in VEGF level.

Metabolomic analyses were performed in tumour fragments using 1H HRMAS NMR technique (Table 3 and Supporting Information Fig. F). The glycolysis/tricarboxylic acid cycle derivatives subset (glucose and alanine) showed a transient significant increase in tumours at 24 hr after treatment and a strong decrease (glucose, lactate and alanine) at 72 hr p.i. (Table 3). Similar changes were observed for the amino acid subset (glutamate, glycine, leucine and lysine). A significant decrease in melanin precursor amino acids such as phenylalanine and tyrosine was observed. For the amino acid derivative subset, a transient increase was observed for taurine at 24 hr p.i. followed by a decrease at 72 hr p.i. A significant decrease was also measured for creatine at both 24 and 72 hr p.i. The phospholipid derivative subset showed increase in choline, glycerophosphocholine, glycerophosphoethanolamine and phosphatidylethanolamine. Finally for lipids, the spectral intensity ratio of methylene (CH2) resonance (at 1.29 ppm) to methyl (CH3) resonance (at 0.9 ppm) was significantly increased at both 24 (+23%) and 72 hr (+11%) after TRT treatment.

Table 3. Modulation of metabolite profile in treated tumours measured by 1H HRMAS NMR
CompoundAssignmentδ1H ppm24 h72 h
  1. Metabolite assignment, 1H chemical shifts and variations were observed on B16BL6 tumours after radionuclide therapy using 1H HRMAS NMR spectroscopy. Metabolites were divided into four subsets as indicated. Variations (mean ± SD) are expressed as percent of untreated values.

  2. a

    p < 0.05.

Glycolysis/TCA cycle derivatives  
1β-GlucoseC1H4.64+19 ± 5%−10 ± 5 %a
2LactateCH31.33NS−27 ± 4%a
3AlanineβCH31.47+22 ± 5%a−62 ± 15%a
6PUFCH5.33NS+53 ± 4%a
Amino acids
7GlutamateαCH3.75+21 ± 10%a−50 ± 9%a
8ClycineαCH3.55+81 ± 22%a−70 ± 15%a
9LeucineδCH30.95+36 ± 6%a−15 ± 3%a
10LysineβCH21.90NS−50 ± 15%a
11PhenylalanineβCH23.13−40 ± 10%a−20 ± 10%a
12TyrosineαCH3.93−60 ± 20%aNS
Amino acid derivatives
13Total creatineCH33.03−46 ± 3%a−22 ± 11%a
14TaurineCH23.42+35 ± 4%a−70 ± 15%a
Phospholipid derivatives
15CholineCH23.52+95 ± 20%a+50 ± 11%**
17GPChoCH23.67+81 ± 11%aNS
18PECH23.22NS+24 ± 9%a
19gPECH23.30NS+26 ± 5%a
  CH2/CH3+23 ± 5 %a+11 ± 3%a


Radiotherapy is based on traditional radiobiological models in which a cell hit by radiation sustains damage to its DNA, either directly or indirectly, which led to increase probability of cell death. In the view of the lack of effective treatment against disseminated melanoma, TRT could be an interesting therapeutic alternative. In this preclinical work, we demonstrate that melanin TRT using 131I-labelled heteroarylcarboxamide molecule could provide an effective clinical treatment of melanoma. We have previously shown that this therapy is greatly effective on human xenografts.[20] In our study, we focus on C57BL6 mice bearing the B16BL6 tumours to assess a potential deleterious effect of this radiotherapy strategy. Although B16BL6 inoculation induced murine and highly pigmented tumour, it is a syngenic model grafted on C57BL6 mice that has several advantages. First, stromal, immune and tumoural cells derived from the same species constitute a predictive model of the interactions between tumoural and surrounding cells. Second, B16BL6 cell line can generate spontaneous lung metastases from primary subcutaneous tumour and allows an assessment of the effect of the therapy on the dissemination process.[27] Finally, C57BL6 mice were pigmented, which is an essential characteristic for testing melanin-targeted radiotherapy on normally pigmented organs such as the skin and eyes. Although this highly pigmented tumour makes targeting melanin easier, it is difficult to obtain a great therapeutic effect because of its high proliferative index. In our work, we show that a single injection of moderate activity was sufficient to obtain anti-tumoural efficacy (both primary tumour growth and dissemination process inhibitions) in combination with low toxicity.

Dosimetric analyses showed that a minimal tumoural delivered dose of 30 Gy is necessary to be effective against B16 tumoural growth. This result is compatible with efficient dose described for solid tumours using internal or external radiotherapy.[28] Delivery of a higher activity using two injections did not improve therapeutic efficacy. Likewise, a single injection of [131I]-anti-CEA antibody is more effective than fractionated doses in treating colorectal tumours.[29] Furthermore, we observed a decrease of tumour vascularisation after [131I]ICF01012 treatment, which could induce a decrease in [131I]ICF01012 uptake with a second treatment and a loss of efficacy.

The mechanistic study also confirmed the strong efficacy of this TRT strategy and showed that residual-treated tumours acquired a loss of aggressiveness. We observed a decrease in mitosis numbers soon after treatment (24 hr p.i.). Likewise, the proliferation marker PCNA and the phosphorylated and activated form of the survival proteins AKT1 and ERK2 were downregulated shortly after treatment. In parallel, different tumour suppressor genes such as Rb and PTEN (phosphorylated active form) were significantly dysregulated by radiotherapy. In addition to the effect on proliferation, we observed, by histological analyses, a decrease in micro-vessel numbers on the residual tumours 14 days after a moderate-activity injection. We showed that the pro-angiogenic marker VEGF was strongly downregulated early after treatment. In parallel, an increase in the anti-angiogenic protein IP10[30] was concomitantly observed 14 days after treatment. These results suggest that TRT could not only have a direct cytotoxic effect on endothelial cells but also via regulation of the pro-angiogenic and anti-angiogenic proteins inducing a long-term anti-angiogenic effect. Finally, we observed a great increase in melanin content mainly in the extracellular space on the residual tumours, which could be due to the release of pigment during cell death. However, we showed that melanogenesis precursors such as tyrosine and phenylalanine were decreased and that the melanogenesis enzyme tyrosinase transcript level was increased. These findings suggest that treatment could induce melanogenesis, leading to the consumption of melanin precursors, the activation of key enzymes and a release overflow of pigment in extracellular space.

These mechanistic studies also provide a better understanding of the in vivo anti-tumoural effects of [131I] radionuclide. Few such studies have been published and most were in vitro performed.[29, 31, 32] We observed a significant increase in the SubG1 fraction on cell cycle analyses of treated tumours and the induction of phosphorylation of p53 protein at serine 15, suggesting induction of DNA damage as expected after radiation.[33] As previously reported in human uveal melanoma, Brantley et al.[34] observed that brachytherapy induced DNA damage, inhibited cell division and promoted cell death, at least in part, due to induction of p53. Our results suggest that p53 has a role in inducing cell death. Indeed, we observed that after treatment, many tumoural cells appear to stop at the G2 phase of the cell cycle. Bunz et al.[35] demonstrated that this arrest could be sustained only when p53 could activate transcriptionally p21 as observed in our transcriptional analyses. In the NMR experiment, we observed an increase in the CH3/CH2 ratio, which is an apoptosis marker.[36] However, this percentage of increase was low and suggested that apoptosis was induced by radiotherapy but was not the “key” mechanism of cell death after [131I]ICF01012 treatment. This hypothesis was confirmed by Western blot analysis for which no cleavage of PARP were observed despite a transient increase in bax transcripts was observed. We show that necrosis (histological evaluation) and autophagy (Atg12 measurement) do not seem to be the major mechanisms of tumoural cell death. Our data suggest that mitotic catastrophe could be the major cell death mechanism with [131I]ICF01012. We observed atypical morphology of tumoural cell nuclei on histological slides, cytoskeleton alteration, metabolite consumption and cell cycle G2/M arrest. Mitotic catastrophe is today considered to be the major cell death mechanism by which solid tumours respond to clinical external radiotherapy and is frequently observed in experimental tumours following radiotherapy and radioimmunotherapy.[37] It seems that internal radiotherapy using [131I]ICF01012 could induce an antitumoural effect via two major cell death pathways described for external radiotherapy, that is, apoptosis for a small proportion of cells and mitotic catastrophe for the rest.

Toxicological studies are essential in the anti-tumoural assessment of therapeutic strategies. We observed low toxicity of [131I]ICF01012 treatment on the highly pigmented C57BL6 model. We showed a transient haematological toxicity induced by TRT. The pharmacological profile of [131I]ICF01012 and notably the rapid clearance from non-target organs indicated that the delivered dose was very low (<4 Gy) in most organs. Concerning normally pigmented organs, a very low delivered dose of 0.2 Gy was calculated in the brain. This distribution could be due to the low proportion of pigmented tissue in the brain or/and a weak ability of [131I]ICF01012 to cross the blood–brain barrier. In a previous clinical study of the parent molecule of ICF01012 (BZA2), a melanoma brain metastasis was detected.[38] These results suggest that [131I]ICF01012 treatment does not induce damage in normal brain but it could target brain metastasis. Biodistribution on the skin was difficult to evaluate due to contamination by urine and blood. Macroscopic and histological observations of the mice's coat showed no damage to the skin and no variation in pigmentation. Finally, we observed a high delivered dose in pigmented eyes (37.6 Gy). For external radiotherapy, a dose of 45 Gy is necessary to induce retinopathy.[39] In our experiments, no histological damage was detected on pigmented structures such as ciliary body and choroid. Moreover for 30% of treated animals, the retina seemed to be unchanged. For other animals, damage was restricted to the optic nerve area. These experiments were performed on the heavily pigmented C57BL/6J mice with consequently a high uveal melanin content, which is not predictive for human status of the uvea. In the clinical imaging study with [123I]BZA2, no significant uveal uptake was visualised on the human eyes except in the case of ocular melanoma, whereas great uveal uptake was observed with BZA2 in this mouse model.[38, 40] Moreover, Joyal et al., who developed a strategy targeting melanin, reported a high uptake of the MIP-1145 tracer in the pigmented eyes of C57BL/6 mice (>30% ID g−1), whereas in cynomolgus monkeys, the highest delivered dose extrapolated from dosimetric analyses was only 6.8 Gy for a 3.7-GBq injection. In conclusion, because of murine and human differences in ocular geometry and melanin content, [131I]ICF01012 fixation on the eyes does not seem to be a major problem for clinical transfer of this radiotherapy concept.

In conclusion, our findings demonstrate strong anti-tumoural efficacy associated with low toxicity of a 131I-labelled heteroarylcarboxamide molecule for radionuclide therapy. In the view of the lack of effective treatment, they strongly support the transfer of [131I]ICF01012 to clinical trials with imaging assessment to determine dosimetric parameters in humans. The efficacy of the therapy could subsequently be assessed to develop an efficient radiopharmaceutical strategy for treating patients with advanced malignant melanoma.


The authors thank Isabelle Maillet (UMR 6128, Immunologie et Embryologie Moléculaires Orléans, France) for haematological analyses; Bruno Pereira for statistical analyses support; David De Freitas for helpful discussion on dosimetric data; Dominique Bernard Gallon and Nadège Rabiau (Laboratoire d'Oncogenétique-CBRV, Clermont-Ferrand, France) for TaqMan low-density array experiments; Amélie Montagne and Laurent Audin for technical assistance; Pierre Labarre and Aurélien Vidal for radioprotection assistance; Christelle Blavignac and Claire Szczepaniak from CICS platform (Clermont-Ferrand, France) and Hang Nguyen for discussion of the manuscript. M. Bonnet was supported by an INSERM young researcher grant.