Evaluation of two 125I-radiolabeled acridine derivatives for Auger-electron radionuclide therapy of melanoma
- First Online:
- Cite this article as:
- Gardette, M., Viallard, C., Paillas, S. et al. Invest New Drugs (2014) 32: 587. doi:10.1007/s10637-014-0086-5
We previously selected two melanin-targeting radioligands [125I]ICF01035 and [125I]ICF01040 for melanoma-targeted 125I radionuclide therapy according to their pharmacological profile in mice bearing B16F0 tumors. Here we demonstrate in vitro that these compounds present different radiotoxicities in relation to melanin and acidic vesicle contents in B16F0, B16F0 PTU and A375 cell lines. ICF01035 is effectively observed in nuclei of achromic (A375) melanoma or in melanosomes of melanized melanoma (B16F0), while ICF01040 stays in cytoplasmic vesicles in both cells. [125I]ICF01035 induced a similar survival fraction (A50) in all cell lines and led to a significant decrease in S-phase cells in amelanotic cell lines. [125I]ICF01040 induced a higher A50 in B16 cell lines compared to [125I]ICF01035 ones. [125I]ICF01040 induced a G2/M blockade in both A375 and B16F0 PTU, associated with its presence in cytoplasmic acidic vesicles. These results suggest that the radiotoxicity of [125I]ICF01035 and [125I]ICF01040 are not exclusively reliant on DNA alterations compatible with γ rays but likely result from local dose deposition (Auger electrons) leading to toxic compound leaks from acidic vesicles. In vivo, [125I]ICF01035 significantly reduced the number of B16F0 lung colonies, enabling a significant increase in survival of the treated mice. Targeting melanosomes or acidic vesicles is thus an option for future melanoma therapy.
KeywordsMelanoma Melanin Acidic vesicles Targeted radionuclide therapy Auger electron emitter Iodine 125
Targeted radionuclide therapy
N-(2-diethylaminoethyl)-9,10-dihydro-7-iodo-9-oxoacridine-4-carboxamide hydrochloride salt
N-(2-diethylaminoethyl)-5-iodoacridine-4-carboxamide dihydrochloride salt
Melanoma is the most severe type of skin cancer, with a high potential for metastatic spread. Melanoma incidence rates continue to rise worldwide . The unsatisfactory results of current targeted therapies for metastatic melanoma highlight the need to develop alternative or additional therapeutic strategies . Clinical studies targeting the MAPK pathway by inhibiting BRAF mutant V600E kinase activity underline the capacity of tumor cells to acquire resistance to this single treatment . Targeted radionuclide therapy (TRT) represents an alternative or an additional weapon to conventional chemo- and immunotherapy for melanoma treatment. TRT can deliver radiation specifically to tumor cells by using specific carriers while sparing the normal tissues and organs. Identified targets are usually membrane receptors that can be recognized by specific labeled antibodies with different isotopes for radioimmunotherapy (RIT). For example, 90Ytrium or 131Iodine anti-CD20 antibodies are used for the treatment of B-cell non-Hodgkin lymphomas . For melanoma, MC1R-specific receptors can be targeted with 188Re peptide analogs of alpha melanocyte stimulating hormone . However, this approach could be compromised by the significant renal uptake of this radiotracer family and the MC1R different expression level as its genetic variations . Melanins offer another interesting target in melanoma. Melanins are pigments produced specifically by cutaneous and ocular melanocytes, cells from the pigmented epithelium, iris and ciliary bodies in eyes. Primary melanomas (90 %) are usually pigmented  while 50 % of metastases still harbor melanins in clinical studies [8, 9]. Melanin targeting has long been a strategy, first for imaging and more recently for TRT, via two approaches: radiolabeled peptides/antibodies or small organic molecules. Monoclonal anti-melanin antibody or melanin-targeting peptide strategies are efficient in preclinical melanoma models (reviewed in ). Among the small molecules binding to melanins, various compounds have yielded promising results for TRT, including methylene blue  and more recently various arylcarboxamide analogs [12, 13, 14]. At clinical level, radioimmunotherapy mainly uses β-emitters (e.g. 90Y, 131I) that produce electrons with ranges between 0.05 and 12 mm. This kind of specific treatment carries a few adverse effects, like myelosuppression  associated with the slow distribution of 90Y-labeled antibodies or with the γ-rays associated with 131I causing non-specific irradiation. To circumvent these issues, other isotopes such as α or Auger electron-emitting radionuclides are of interest due to their high-linear energy transfer (LET) particles that deliver high radiation doses in a restricted area. Alpha emitters tend to be difficult to produce or handle safely, but some Auger electron donors are used in clinical practice, such as 123I for imaging  and 125I  for brachytherapy. High energy deposition is kept to an extremely small volume around the site of decay (nm3)  where its short radiation range induces little damage to surrounding tissues , while the close proximity of isotope to DNA facilitates radiotherapeutic effectiveness .
In order to target the DNA of pigmented melanoma cells, melanin-targeting ligands bearing acridine or acridone cores, known for their DNA intercalating properties, have been designed. Among the molecules synthesized, two compounds ICF01035 and ICF01040 were selected as candidates for therapeutic application due to their favorable pharmacokinetic profiles in vivo with high and lasting tumor uptakes in melanoma . They still exhibited DNA intercalating properties in vitro and induced a significant radiotoxicity on murine melanoma cells when radiolabeled with 125I , but [125I]ICF01040 displayed a higher radiotoxicity than [125I]ICF01035 despite similar cellular uptake. Interestingly, compound ICF01035 exhibited a stronger affinity for melanin than ICF01040, suggesting a predominant targeting to melanosome compartments, that was also assessed by SIMS analysis . Taken together, these results suggest that the radiotoxicity of these 125I-labeled compounds could be correlated to their subcellular distributions.
In malignant melanoma, two major organelles can scavenge xenobiotics: melanosomes, which can sequester compounds by a binding to melanins , and acidic organelles (i.e. lysosomes), that can concentrate organic bases due to their membrane potential . We thus ran an in vitro study on the subcellular localization of these two compounds in melanotic and amelanotic cell lines focusing on their cellular uptake and radiotoxic effects. [125I]ICF01035 and [125I]ICF01040 were also tested for their antitumor properties in vivo in the melanoma B16F0 lung colony model.
Materials and methods
Synthesis and radiolabeling of ICF01035 and ICF01040
ICF01035 and ICF01040 were synthesized as previously described . The radioiodinated compounds [125I]ICF01035 and [125I]ICF01040 were obtained following the procedure described in  at high specific activity (81.4 TBq/mmol) and with excellent radiochemical purity (>99 %) using a radioiododestannylation procedure based on the corresponding tributylstannyl derivatives.
B16F0 cell line: pigmented murine melanoma originating from a spontaneous primary tumor.
A375 cell line: amelanotic human melanoma originating from a primary tumor.
B16F0 PTU cell line: obtained in the laboratory after treating B16F0 with phenylthiourea (PTU) to inhibit melanogenesis. Melanin quantity was measured after PTU treatment in B16F0 cells (data not shown). After four passages, the cells obtained were declared amelanic.
All these cells lines were maintained as monolayers in 75-cm2 culture flasks in adapted medium. B16F0 and A375 were cultivated in DMEM medium (Sigma, Saint-Quentin-Fallavier, France), and the B16F0 PTU cell line was cultivated for at least four passages in DMEM medium supplemented with 100 μM of PTU. All these media were supplemented with 10 % fetal calf serum (Sigma, Saint-Quentin-Fallavier, Farnce) and 5 mL of a 100X solution of vitamins (Invitrogen, Cergy-Pontoise, France), 5 mL of 100 mM sodium pyruvate (Invitrogen Cergy-Pontoise, France), 5 mL of 100X non-essential amino acids (Invitrogen, Cergy-Pontoise, France) and 2 mg of gentamycin base (Invitrogen, Cergy-Pontoise, France). Cells were grown at 37 °C in a humidified incubator containing 5 % CO2.
The emission/detection wavelengths were determined for each compound on a Fluoroskan spectrofluorimeter (Fluorescent Ascent FL™, Labsystems, Gometz-le-Chatel, France) as ICF01035 λexcitation = 357 nm, λemission = 462 nm; ICF01040 λexcitation = 391 nm, λεmission = 456 nm.
For subcellular localizations, visualization was performed on a Leica TCS SP2 AOBS confocal microscope equipped with a 63 × 1.4 oil-immersion objective lens with a 2.2 zoom (all from Leica Microsystems, Rueil-Malmaison, France). Images were acquired using a dual-band 488-nm argon and 543-nm helium/neon laser and collected using Leica Confocal Software.
Nuclear and melanosome visualization
30. 103 cells were grown on a cover glass in 1 mL of complete medium for 16 h at 37 °C. The cells were treated by the appropriate compound at 4 × IC50 (i.e. 16 μM for ICF01035 and 2.4 μM for ICF01040 as previously determined ) for 2 h, fixed by 30 min incubation in AFA (Sigma, Saint-Quentin-Fallavier, France), then incubated at 37 °C for 15 min in Triton (0.2 % in PBS) (Sigma, Saint-Quentin-Fallavier, France) to allow membrane permeabilization before a 30 min saturation step in PBS/SVF (10 %). Cells were stained with a rat TRP1 (tyrosinase-related protein, marker of melanosomes) antibody (1/100 dilution) (Santa Cruz, DE) and Alexa Fluor® 647 anti-rat (1/750 dilution) (Santa Cruz, DEelaware, USA) and propidium iodide (2 μg/mL) (Santa Cruz, DE) was used to label the nucleus.
Acidic organelles labeling
10. 104 cells were grown in 6-well multidishes in 2 mL of medium for 16 h at 37 °C, then treated by the appropriate compound at 4 × IC50 for 2 h (see above). At 30 min before analysis, the cells were treated by Lyso sensor green (LSG, λexcitation = 440 nm, λemission = 505 nm, Invitrogen, Cergy-Pontoise, France) at a final concentration of 1 nM.
Determination of α parameter
To determine the α parameter , the pH of a 1-octanol-saturated PBS solution (Sigma, Saint-Quentin-Fallavier, France) was modified by adding aqueous HCl 1 N solution (Sigma, Saint-Quentin Fallavier, France), then the lipophilicity of radioiodinated compounds, i.e. log p, was determined in each pH-adjusted solution. Briefly, the radiolabeled compound was prepared in the appropriate pH-adjusted PBS solution (pH range: 2–11), 1-octanol was added and mixed, and the activity of each phase after decantation and separation was measured. p was calculated as a ratio of activities (octanol/PBS solution) and log transformed to express lipophilicity. The ratio of log p for the more acidic value to log p in the more basic conditions gives the α parameter.
Determination of acidic organelles compartment
For each cell line, the quantity of acidic organelles was determined using LSG. 50,000 cells were grown in 96-well multidishes in 150 μL of complete medium for 16 h at 37 °C. Cells were incubated for 1 h at 37 °C with 10 μM of LSG, then fluorescence at 485/530 nm was determined using the Fluoroskan spectrofluorimeter. The fluorescence measured was directly correlated to the number of acidic vesicles.
Determination of cell melanin content
Cells were grown in 10 cm-diameter petri dishes for 16 h. After recovery, cells were pelleted by centrifugation at 3,000 g for 10 min and further treated with aqueous KOH 1N solution (1.106 cells/mL). After 1 h incubation at 60 °C, melanin content was determined using a Multiskan spectrophotometer (MS, Labsystems, VA) at a wavelength of 405 nm (100 μL of each in triplicate) and comparatively to the synthetic melanin standard curve.
106 B16F0 cells were grown in 10 cm-diameter culture dishes in 10 mL of complete medium for 24 h at 37 °C. The appropriate radioiodinated compound was added (1.85 MBq/dish), and cellular uptake was determined after different incubation times. At each timepoint, cells were washed with cold PBS, scraped and counted, and radioactivity was determined on a γ-counter (Wizard 1480, Perkin Elmer, Villebon-sur-Yvette, France). Cellular uptake was calculated as percent total radioactivity added (TAA) into the dish, and expressed per million cells.
Cell cycle analysis
106 B16F0 cells were grown in 10 cm-diameter culture dishes in 10 mL of complete medium for 24 h at 37 °C. Radioiodinated ICF01035 and ICF01040 compounds were added (1.85 MBq/dish), and 24 h later the cells were trypsinized, fixed in 70 % ethanol at −20 °C for 3 h and kept in alcohol 70 % (1 million/0.5 mL) at −20 °C until FACS analysis. Cells were stained with cell cycle kit reagent from Merck Millipore (Merck Millipore, Guyancourt, France) for 30 min at room temperature in the dark before analysis using an Muse® flow cytometer (Merck Millipore, Guyancourt, France).
In vivo antitumor tests
All experiments were carried out in C57Bl6 male mice (Charles River Laboratories, L’Arbresle, France) in compliance with French laws governing animal experimentation.
B16F0 were resuspended in PBS after trypsination and injected into the tail vein (1.5 × 105 cells per mouse in 0.2 mL). Treatment was administered i.v. via a tail vein at D5 (one injection of 74 MBq) or at D5 and D14 (two injections of 74 MBq, 1.1 pmol) following cell injection. Treatment efficacy was determined by measuring median of survival and by removing the lungs and counting lung colonies at D20 post-cell inoculation. For lung colony numbering, we used 8, 9 and 10 mice in control, [125I]ICF01035 and [125I]ICF01040 groups, respectively. For survival analysis, the groups contained 10, 6 and 7 mice for control, 1 × 74 MBq [125I]ICF01035 and 2 × 74 Mbq [125I]ICF01035, respectively.
Secondary ion mass spectrometry analysis
B16F0 cells were injected intravenously into male C57BL/6J mice to obtain, within 12 day, tumor cell colonies in lungs mimicking pulmonary micrometastases. After compound administration (0.1 μmole/mouse), mice (2 per point) were sacrificed at times 24 h, 72 h, 5 day and 8 day post-administration. Typical sample preparation for SIMS analysis was previously described . Briefly, lungs were removed and small pieces of tissue including B16 melanoma colonies were isolated and fixed by slam-freezing on an LN2-precooled metal mirror. Samples were dehydrated by freeze-drying starting at −110 to −10 °C before embedding in Spurr resin. Serial 0.4 μm-thick sections were deposited on stainless steel holders for SIMS analysis or on glass slides for light microscopy observation. SIMS imaging was performed using a NanoSIMS-50™ ion microprobe (CAMECA, France) at the PICT-IBiSA facility of the Curie Institute (Orsay, France). The instrument is equipped with a magnetic mass spectrometer using a parallel detection system that allows the simultaneous detection of up to five species from the same microvolume. The primary ion beam was generated by a caesium source and accelerated to get an impact energy of 16 keV. For typical experiments, probe size was 100 nm in diameter (defined as 16–84 % rise distance of the signal intensity), with a current of 1.5 pA. The probe scanned the surface of the sample in a 256 × 256 pixel raster. Depending on the analysis, dwell time ranged from 10 to 30 ms per pixel and width of the scanned area ranged from 100 to 20 μm. The magnetic field is set to detect the heaviest mass (here, mI− = 127) on one of the largest-radius detectors. The other movable detectors were positioned to detect 12C14N−, 31P− and 32S−. However, with the magnetic field used to detect iodine, there is a limitation with spacing between adjacent detectors that precluded the simultaneous detection of 31P− and 32S− ions. Therefore, in the present determination of cellular ICF01035 distribution, images were acquired in two sequential series (12C14N−, 31P− and 127I− then 12C14N−, 32S− and 127I−). Tissue structures were identified by detecting 12C14N− ions reflecting tissue and intracellular nitrogen content or by imaging the distribution of 32S− ions. The distribution of 31P is used to study the distribution of phosphate-rich molecules and consequently to localize nuclei. The resulting map of 127I− iodine ions reflected the tissue and cellular distribution of the compounds and their iodinated metabolites. For each anion selected, the count measured using the detector is directly related to the local concentration of the element. The raw data was a 16-bit 256 × 256 pixel image resulting from the probe scanning over the surface of the sample. The intensity of each pixel corresponds to the direct local measurement of flux of the secondary ions. Image processing was performed using the ImageJ public-domain Java image processing software to obtain proper colocalization of the observed structures on the processed maps for all the ion species, and allowed further correlation with light microscope images.
In all experiments, data are reported as mean ± SEM and were analyzed using Student’s t-test or ANOVA tests with level of significance set at 0.05. For survival studies, a log Rank test was used. Analyses were performed using XLSTAT Software (Addinsoft, Paris, France).
ICF01035 and ICF01040 subcellular localization in the pigmented B16F0 cells
ICF01035 showed a cytoplasmic punctiform accumulation (Fig. 1a). Image merging revealed co-localization with TRP1 signal only. This result illustrated a melanosomal sequestration of ICF01035 as the only organelles expressing this enzyme.
ICF01040 showed a cytoplasmic punctiform accumulation (Fig. 1b). After cell treatment by ICF01040, the TRP1 signal was diffuse, so the nature of the organelles that scavenge this compound remain unknown.
Characterization of acidic organelle involvement in ICF01035 and ICF01040 subcellular localization
α parameter determination:
The α parameter reflects the lipophilicity variations for a compound between its ionized and un-ionized forms in acidic media. If α parameter value is around 1, the compound’s lipophilicity is weakly affected by pH variations. This was the case of ICF01035, with an α parameter value of 0.90 ± 0.07 (n = 3). In contrast, the lipophilicity of ICF0040 was pH-dependent, as revealed by an α parameter of 0.10 ± 0.03 (n = 3). This statistically significant difference (p = 0.004, Student test) indicated that the lipophilic properties of ICF01040 should be modified in acidic media, in contrast to ICF01035.
- b)Study in amelanotic cell lines:
Determination of acidic compartment in cell lines
LSG, a fluorophore linked to a weak base, is a pH-dependent marker of acidic organelles, as its fluorescence increases proportionally with number of acidic vesicles. Acidic organelle quantification was performed in B16F0, A375 and B16F0 PTU cell lines (Fig. 2a). Figure 2a reports the results obtained as relative fluorescence units measured in each cell lines. After labeling, measured fluorescence in B16F0 and B16F0 PTU was nearly identical (0.2 and 0.19 respectively) whereas A375 cells showed significantly less LSG accumulation (0.12). These data showed that B16F0 and B16F0 PTU possess the same bigger acidic compartment than A375.
Subcellular localization of ICF01035 and ICF01040 in amelanotic cell lines:
ICF01035 predominantly accumulated in nuclei (Fig. 2b) whereas ICF01040 was mainly sequestered in the cytoplasmic compartment. In order to identify the organelles involved in the cytoplasmic sequestration of ICF01040, acidic organelles were labeled by an LSG tracker. There was perfect co-localization between tracker and ICF01040 (Fig. 2b). In amelanic A375 cells, ICF01035 reached the nucleus whereas ICF01040 was scavenged in acidic organelles.
Cellular uptake and radiotoxic effects of [125I]ICF01035 and [125I]ICF01040
In vivo radiotherapy efficacy of [125I]ICF01035 and [125I]ICF01040 on lung B16F0 colonies
In vitro, ICF01035 was sequestered in melanosomes of pigmented cells. SIMS analysis was used to determine subcellular trafficking of ICF01035 in pigmented lung colonies. SIMS can directly identify chemical elements with high sensitivity and specificity and can be applied to visualize element distribution (chemical mapping). Figure 5c summarizes the SIMS images focused on melanoma lung colony cells at 24 h post-injection of ICF01035 at nontoxic concentrations. The distribution of cyanide ion (5Ca) revealed various tissue structures, particularly melanin polymers mainly in melanosomes. Figure 5cb shows the distribution of iodine signal, i.e. ICF01035 and/or iodinated derivatives. For ICF01035, images of iodine (5Cb) and cyanide ion (5Ca) signals exhibited similarly bright structures illustrating a clear co-localization, most likely at the melanin pigment sites in melanosomes (Fig. 5c). Images of phosphorus distribution (5Cc) clearly identified the condensed chromatin in nuclei. Phosphorus (5Cc) and iodine (5Cb) images gave a separate pattern showing no iodine signal in nuclei. Based on this evidence, ICF01035 should then be localized within cytoplasmic organelles mainly co-localized with melanin, as previously highlighted by confocal microscopy.
This in vitro study investigated the differential behavior of two melanin-targeting ligands radiolabeled with 125I (i.e. [125I]ICF01035 and [125I]ICF01040) in relation to their subcellular localization and radiocytotoxicity properties. Radiocytotoxicity was significantly higher for acridine [125I]ICF01040 than for its acridone derivative ([125I]ICF01035) in pigmented B16F0 cells . This higher efficiency was thought to be linked to a stronger melanin affinity for ICF01035 decreasing its DNA-intercalating potential .
Confocal microscopy on B16F0 found that neither compound accumulated in the nucleus yet ICF01035 was highly concentrated in melanosomes, as demonstrated by colocalization of TRP1 labeling. This was obviously supported by a massive ICF01035 concentration in amelanotic cell nuclei. [125I]ICF01035 accumulation was two-fold higher in B16F0 compared to A375 and B16F0 PTU cells, while [125I]ICF01035 A50 was similar in all three cell lines, showing that [125I]ICF01035 becomes radiotoxic when a threshold of cellular lesions is reached. Another non-exclusive hypothesis could rely on radiation multitargeting DNA and/or membranes. Indeed, xenobiotic scavenging by melanosomes has previously been demonstrated for anticancer drugs  but also for melanin-binding molecules . Interestingly, the radiotoxicity of [125I]ICF01035 could thus be tied to endogenous melanogenic cytotoxicity (EMC), as Auger electrons resulting from 125I radiation can alter membrane integrity  and lead to toxic molecule release associated with melanin production . This EMC should be higher in pigmented cell lines containing stage III and IV melanosomes than in lines with stages I and II melanosomes . However, [125I]ICF01035 only induced a decrease in S-phase in amelanotic cells, suggesting that nuclear location of Auger emitters is also necessary to decrease cell proliferation. [125I]ICF01035 radiotoxicity could then be supported by at least two mechanisms, one involving melanosoma EMC and the other cell death due to the DNA breaks observed in B16F0 cells (data not shown).
ICF01040 was mainly observed in the cytoplasmic compartment, showing that [125I]ICF01040 radiotoxicity did not rely on direct DNA breaks. [125I]ICF01040 toxicity was not melanin-dependent. Therefore, this cytoplasmic localization could be due to acidic organelles present in high numbers in all three melanoma cell lines offering a refuge to this protonable molecule by lipophilicity modification. Indeed, we observed that ICF01040 uptake and radiotoxicity were tightly correlated to amount of acidic organelles. Moreover, in A375 cells and compared to [125I]ICF01035, [125I]ICF01040 had a higher A50 (7 kBq/mL vs 11 kBq/mL, respectively). This suggested that [125I]ICF01040 was concentrated in radiosensitive organelles essential for cells. These results were supported by previous studies that have identified acidic vesicle scavenging for iodobenzamides .
In conclusion, in vitro studies found that despite different subcellular localizations, both compounds induced radiotoxic effects in melanomas, with higher performance for ICF01040 located in acidic vesicles. In vivo, lung colony count was only significantly reduced in [125I]ICF01035-treated mice, translating into a significant increase in lifespan. We posit that [125I]ICF01035 was more efficient than [125I]ICF01040 in lung colony treatment due to its pharmacological profile and notably its higher tumoral accumulation  leading to a higher biological period (269 h) than [125I]ICF01040 (43 h). SIMS analysis applied to in vivo specimens provided strong evidence that ICF01035 was delivered in B16F0 melanosomes. Despite being sequestered in the cytoplasmic compartment, [125I]ICF01035 nevertheless showed radiotherapeutic efficacy leading to cell death.
The melanoma models developed here demonstrated compelling properties of melanin-targeting ligands radiolabeled with 125I that is an Auger electron emitter but also an X- and γ-ray-emitting radionuclide. A mixed radiation effect likely supports the observed radiotoxicity. Microdosimetry should now be performed in order to decipher the precise mechanisms underpinning the radiocytotoxicity of these two 125I-radiolabeled acridine derivatives.
The French Ligue Régionale contre le Cancer provided financial sponsorship for this project. The Auvergne Regional Council and the INSERM provided funding for Maryline Gardette’s PhD work. The authors thank the PICT-IBiSA imaging facility in the Institut Curie for allowing us to use the NanoSIMS microprobe.
Conflict of interest
The authors declare that they have no conflicts of interest.