Evaluation of Acridine Orange Derivatives as DNA-Targeted Radiopharmaceuticals for Auger Therapy: Influence of the Radionuclide and Distance to DNA

A new family of 99mTc(I)- tricarbonyl complexes and 125I-heteroaromatic compounds bearing an acridine orange (AO) DNA targeting unit was evaluated for Auger therapy. Characterization of the DNA interaction, performed with the non-radioactive Re and 127I congeners, confirmed that all compounds act as DNA intercalators. Both classes of compounds induce double strand breaks (DSB) in plasmid DNA but the extent of DNA damage is strongly dependent on the linker between the Auger emitter (99mTc or 125I) and the AO moiety. The in vitro evaluation was complemented with molecular docking studies and Monte Carlo simulations of the energy deposited at the nanometric scale, which corroborated the experimental data. Two of the tested compounds, 125I-C5 and 99mTc-C3, place the corresponding radionuclide at similar distances to DNA and produce comparable DSB yields in plasmid and cellular DNA. These results provide the first evidence that 99mTc can induce DNA damage with similar efficiency to that of 125I, when both are positioned at comparable distances to the double helix. Furthermore, the high nuclear retention of 99mTc-C3 in tumoral cells suggests that 99mTc-labelled AO derivatives are more promising for the design of Auger-emitting radiopharmaceuticals than the 125I-labelled congeners.


Reagents and general procedures
All reagents were analytical grade and were used without further purification. Unless stated otherwise, the syntheses of the ligands and complexes were carried under nitrogen atmosphere, using standard Schlenk techniques and dried solvents. HPLCgrade solvents were used for HPLC purification and analysis. Sodium  125 I]iodide was obtained from Perkin Elmer, USA, as a non-carrier added solution in 0.1 M aqueous NaOH with radionuclidic purity > 99% and specific activity of 643.8 GBq/mg. Na[ 99m TcO 4 ] was eluted from a commercial 99 Mo/ 99m Tc generator (Drytec, GE Healthcare) using a 0.9% saline solution. 1 H and 13 C NMR spectra were recorded on a Varian Unity 300 MHz spectrometer, using CDCl 3 or CD 3 OD as solvents. 1 H and 13 C chemical shifts are given in ppm and were referenced to the residual solvent resonances relative to SiMe 4 . The spectra were assigned with the help of 2D NMR spectroscopy (H-H COSY, and 1H-13CHSQC).
IR spectra were recorded in the range 4000-400 cm -1 as KBr or CsI pellets on a Bruker Tensor 27 spectrometer.
Electrospray ionisation mass spectrometry (ESI-MS) was performed using a Bruker HCT electrospray ionization quadrupole ion trap mass spectrometer.
Elemental analyses were performed on a Perkin-Elmer automatic analyzer.
Analytical thin layer chromatography (TLC) was carried out on silica gel 60 F254

General procedure for the synthesis of tributyltin precursors
A mixture of AO-iodobenzamides, 127 I-C 3 , 127 I-C 5 or 127 I-C 8 , and (Sn(Bu) 3 ) 2 (2.4 eq) in dry DMF (3-4 ml) was refluxed overnight under nitrogen atmosphere in the presence of a catalytic amount of PdCl 3 (PPh 3 ) 2 (0.1 eq). The amount of reagents used and reaction yields are shown in Table S1. Solvent was removed under reduced pressure and the dry residue obtained was purified by alumina column chromatography (methanol:CH 2 Cl 2 ; 1:9) to give the tributyltin precursors (Sn-C 3 , Sn-C 5 and Sn-C 8 ) as dark red solids.

General procedure for the synthesis of the radioiodinated acridines 125 I-C 3 , 125 I-C 3 and 125 I-C 8
The radioiodinated acridines were prepared by iododestanylation using the tributyltin precursors, Sn-C 3 , Sn-C 5 and Sn-C 8  The fractions containing the product were collected, combined, diluted in 10 ml of H 2 O and subsequently loaded onto a C18 Sep-Pak cartridge (prewashed with 5 ml of methanol, followed by 2x5 ml of water). The radioiodinated product was eluted with 2 ml of ethanol. The identity of radioiodinated AO derivatives was confirmed by coelution with reference compounds (Table S2). HPLC radiochromatograms of the coelution of purified 125 I-C 5 with non-radioactive analogue 127 I-C 5 are shown in Figure   S2 as an example.

General procedure for the synthesis of Re complexes
The rhenium complexes (Re-C 3 /Re-C 5 /Re-C 8 Table S3.  The chemical identity of the 99m Tc complexes was ascertained by comparison of their HPLC profiles with those of the corresponding rhenium complexes. HPLC radiochromatograms of the co-elution of purified 99m Tc-C 5 complex with its corresponding rhenium complex, Re-C 5 are shown in Figure S4 as a representative example. HPLC purification of the AO-containing 99m Tc-complexes removed the excess of the pyrazolyl-diamine chelators, which could not be detected at the most sensitive UV detector setting. Hence, it can be considered that their specific activity is in the same range as that of the starting Na 99m TcO 4 (ca. 100,000 Ci/mmol for 99m Tc obtained from a 99 Mo/ 99m Tc generator undergoing daily elution 5 ).

In vitro stability studies
Radiochemical stability of the radioiodinated derivatives ( 125 I-C 3 , 125 I-C 5

UV-Vis absorption spectra
UV-Vis spectra were measured in all systems in order to correct the fluorescence emission spectra for reabsorption and inner-filter effects. [6][7][8] Although these spectra were not used for the determination of binding parameters, they give a clear indication of the strong interaction between all compounds and DNA, due to the occurrence of hypochromism and red shifts in the absorption bands. Figures S7-S10 show the UV-Vis absorption spectra measured for the different tested compounds in the presence of increasing amounts of CT-DNA. Table S4 summarizes the observed effects.

Fluorescence titrations
The fluorescence binding experiments with CT-DNA were performed in a quartz cuvette of 1 cm path length. Bandwidth was typically 5 nm in both excitation and emission. Fluorescence titrations were done in which increasing amounts of a CT-DNA solution (ca. 3.3mM) were added to the solution containing the fluorescent compounds (see Table S5 for conditions, such as concentrations and excitation wavelengths). The  Two types of data treatment were done. According to the Kaminoh model 9 the data was fitted to eqn. (1).
As the concentration of the fluoropohore and its emission intensity (I 0 ) are known, the value in saturation conditions, I sat , can be calculated from the representation of I vs. [DNA].
According to McGhee von Hippel model 10 the concentration of the free probe in each sample (C F ) can be calculated using eqn. (2), where C T is the total concentration of the probe and P is the ratio of the observed fluorescence intensity of the bound probe to that of the free probe.

Circular Dichroism
Circular dichroism (CD) is widely used to study the affinity and binding modes of small molecules to biomolecules, particularly DNA. 11,12 When the compounds are not chiral and thus present no CD signal, their association with the right-handed DNA helix may give rise to induced CD spectra (ICD) in the range where they absorb. Moreover, DNA is chiral due to being placed within the framework of the chiral sugar-phosphate backbone, producing a characteristic CD spectrum in the 200-300 nm range. Therefore, changes in the CD signal in this spectral range indicate DNA conformational changes.
For the compounds studied the focus was on the induced CD signal above 300 nm, although small conformational changes were also observed whenever spectra were measured below 300 nm.
The spectra measured for the different compounds are presented in Figure S16. In some cases ( 127 I-C 3 in which a cell with 1 cm of path length was used, and Re-C 3 in which the complex concentration was much lower) the ICD signal is weak but this was due to the low solubility of the compound, which precluded the use of higher concentrations that are needed for the observation of more intense ICD signals.

Molecular docking validation
The molecular docking simulation protocol were validated by first docking the bisintercalated anthracycline drug present in the NMR solution structure of the d(ACGTACGT) 2 sequence (PDB code 1AL9) 13 . The top-ranked pose revealed the anthracycline bis-intercalated in ds-DNA structure in a similar pose to that of the NMR structure solution, with a ligands all atoms root-mean-square deviation (RMSD) of 1.18 Å ( Figure S17) and demonstrating the effectiveness of the chosen methodology.

Molecular Dynamics simulations
The structures of the radioiodinated AO derivatives ( 125 I-C 3 , 125 I-C 5  The RMSD values of the acridine ligands atoms stabilize after the first 20 ns of the MD production, with values of ca. 1.0 Å for 125 I-C 3 and 125 I-C 5 and of ca. 2.5 Å for linker 125 I-C 8 (Figure S18b). The values are similar to the all atoms RMSD values of the ds-DNA structure, indicating that acridine derivatives are preserved well in the binding site of each model. This was also confirmed by the visual inspection of the MD runs.
As verified in the top-ranked molecular docking poses, during the MD simulations the acridine derivatives are stabilized by π-π stacking interactions with the nucleobases and it is worth noting that the acridine aromatic ring remains stable between GC base pairs during all simulation. Although, acridine side chains showed high flexibility and amplitude of movement and aid the complex stabilization mainly by performing Hbonding with phosphate backbone, nucleobases and/or water molecules ( Figure S19).
In Figure S18, the 125 I distance to ds-DNA helical axis is plotted along the 50 ns MD simulation. Additionally, time relevant MD snapshots are presented to depict the dynamics of the 125 I radionuclide towards the ds-DNA structure in the last 30 ns of the simulation. In the last 30 ns of the MD simulation (Figure S20), the average distances between ds-DNA helical axis and 125 I in acridine side chains range between 9.9 Å and 14.8 Å. As such the average distance between 125 I and the ds-DNA helical axis is highly dependent of the acridine side-chain length and corroborates the previous docking results. Moreover, the influence of 125 I in the ds-DNA structure is affected not only by the distance to the helix, but also by radionuclide contact time. Therefore, it is expected that acridine derivatives with side chains with higher amplitude of movement exhibit lower contact time between 125 I and the ds-DNA. In fact, as shown in Figure

Monte Carlo (MC) Simulation: geometrical setup for MC simulations
Deposited energies were calculated in a volume corresponding to the DNA segment of 10 base pairs of length and a nucleosome, both modelled as liquid water cylinders with nanometric dimensions 16 . Liquid water is the main constituent of the human body and represents a good approximation for soft biological tissue 17 .