99mTc-Labeled HYNIC-DAPI Causes Plasmid DNA Damage with High Efficiency

99mTc is the standard radionuclide used for nuclear medicine imaging. In addition to gamma irradiation, 99mTc emits low-energy Auger and conversion electrons that deposit their energy within nanometers of the decay site. To study the potential for DNA damage, direct DNA binding is required. Plasmid DNA enables the investigation of the unprotected interactions between molecules and DNA that result in single-strand breaks (SSBs) or double-strand breaks (DSBs); the resulting DNA fragments can be separated by gel electrophoresis and quantified by fluorescent staining. This study aimed to compare the plasmid DNA damage potential of a 99mTc-labeled HYNIC-DAPI compound with that of 99mTc pertechnetate (99mTcO4 −). pUC19 plasmid DNA was irradiated for 2 or 24 hours. Direct and radical-induced DNA damage were evaluated in the presence or absence of the radical scavenger DMSO. For both compounds, an increase in applied activity enhanced plasmid DNA damage, which was evidenced by an increase in the open circular and linear DNA fractions and a reduction in the supercoiled DNA fraction. The number of SSBs elicited by 99mTc-HYNIC-DAPI (1.03) was twice that caused by 99mTcO4 − (0.51), and the number of DSBs increased fivefold in the 99mTc-HYNIC-DAPI-treated sample compared with the 99mTcO4 − treated sample (0.02 to 0.10). In the presence of DMSO, the numbers of SSBs and DSBs decreased to 0.03 and 0.00, respectively, in the 99mTcO4 – treated samples, whereas the numbers of SSBs and DSBs were slightly reduced to 0.95 and 0.06, respectively, in the 99mTc-HYNIC-DAPI-treated samples. These results indicated that 99mTc-HYNIC-DAPI induced SSBs and DSBs via a direct interaction of the 99mTc-labeled compound with DNA. In contrast to these results, 99mTcO4 − induced SSBs via radical formation, and DSBs were formed by two nearby SSBs. The biological effectiveness of 99mTc-HYNIC-DAPI increased by approximately 4-fold in terms of inducing SSBs and by approximately 10-fold in terms of inducing DSBs.


Introduction
The study of cellular DNA damage by various chemical and physical events is hampered by the existence of effective repair systems. Moreover, cellular and nuclear membranes are barriers that are difficult to overcome. Plasmid DNA is a model that enables the investigation of unprotected interactions between molecules and DNA that result in single-strand breaks (SSBs) or double-strand breaks (DSB); the resulting fragments can be separated by gel electrophoresis and quantified by fluorescent staining. Auger electron-emitting radiotracers can be implanted directly into DNA via the incorporation of radiolabeled thymidine analogues or by indirect DNA labeling with radioactive dyes to study the influence of geometry [1,2] or to characterize different isotopes of the same element [3].
Most data have been obtained using 125 I-labeled radiotracers. Recently, 99m Tc has become an interesting candidate for imaging and radiotherapy. Haefliger et al. demonstrated the induction of DSBs by Auger electrons from 99m Tc complexes with DNAbinding ligands using gel electrophoresis; however, the damage was not quantified under consideration of the absorbed dose [4]. Santos-Cuevas et al. demonstrated by microscopy and with proliferation assays that 99m Tc-labeled bombesin effectively trafficked to the nucleus in prostate and breast cancer cells [5]. Schipper et al. reported that 99m Tc and 131 I accumulated in NISpositive neuroendo-crine tumor cells in vitro and in a preclinical model, suggesting the potential for both imaging and therapy [6]. Kriehuber et al. investigated cell survival, the induction of apoptosis and micronucleus formation in SCL-II cells after exposure to the Auger electron emitter 99m Tc [7], but cellular uptake was not sufficient to ascertain the radiobiological effectiveness. Kotzerke 188 Re. [10]. Because nuclear uptake was negligible, DNA damage by lowenergy Auger electrons could not be detected. Wunderlich et al. increased the intracellular accumulation of 99m Tc by preincubation with Sn-complexes [11]. Recently, the importance of lowenergy Auger and conversion electrons was reported by Cambien et al., who investigated 99m TcO 4 2 mediated thyroid stunning in vivo [12].
One possibility for transporting radioisotopes into the nucleus is to use radiolabeled fluorescent dyes [13,14]. DAPI (49,6-diamidino-2-phenylindole) is a DNA-specific fluorescent probe that interacts with DNA by inserting into the minor grooves of AT-rich DNA sequences [15]. In our study, DAPI was labeled with 99m Tc via a HYNIC (6-hydrazinonicotinic) linker. We characterized the damage to plasmid DNA caused by the Auger electron emitter 99m Tc using quantitative gel electrophoresis in the presence or absence of the radical scavenger DMSO and compared the radiobiological effectiveness with that of unbound 99m Tc-pertechnetate.
DNA stock solutions were diluted in TE buffer (10 mM Tris-HCl and 1 mM EDTA, pH 7.5) to a final concentration of 0.1 mg/mL and were stored at 220uC. Only samples containing . 95% closed circular DNA were used. pUC19 was digested with BamHI (Invitrogen, Karlsruhe, Germany) to obtain linear DNA.

Irradiation Procedure
The necessary samples, each containing 200 ng plasmid DNA at 0.1 mg/mL and various amounts of the radioactive solutions, were placed in microtubes (Eppendorf, Hamburg, Germany). One radiotracer-free microtube served as the control. 99m Tc-HYNIC-DAPI was added to the pUC19 plasmid solutions (0.1-25 MBq/20 mL = 5.0-1250 MBq/mL) in a total sample volume of 20 mL.
99m Tc-pertechnetate solutions were adjusted to the same radioactivity as the 99m Tc-HYNIC-DAPI solutions and were also added to the plasmid samples. In an additional test series, 0.2 M DMSO (final concentration) was added to the radiotracer solutions. Each sample was separated into two 10 mL aliquots to enable the accumulation of radiation dose for 2 h at 4uC or for 24 h at 220uC.
After irradiation, 10 mL of each irradiated DNA solution was mixed with 1.25 mL of loading buffer. The mixture was pipetted into the wells of a 1.4% agarose gel in TAE buffer. The samples were run at 4 V/cm for 120 min at 6uC. After electrophoresis, the closed circular (supercoiled), open circular and linear plasmid DNA fractions were identified based on mobility differences in the gel. The gel was stained with ethidium bromide (0.5 mg/mL) for 30 min, and then the different forms of DNA were visualized using an UV transilluminator (Diana III Digital Imaging System, Raytest, Straubenhardt, Germany). The gel was imaged using a charge-coupled device (CCD) camera. The relative amount of DNA in each conformation was quantified by integrating the corresponding intensity using the open-source platform software Fiji [16]. To quantitate the DNA-bound 99m Tc-HYNIC-DAPI activity, the DNA bands corresponding to the supercoiled, linear and open circular forms were excised, and the gel pieces were analyzed using a gamma counter (Cobra II Auto-Gamma, Perkin Elmer life sciences).

Dosimetry
The dose calculations were performed using the Geant4 Monte Carlo toolkit. Although the real shape of the liquid volume in the microtube is similar to an inverse cone with a spherical apex, the target volume was considered to be a sphere with a 10-mL volume composed of water. The decay sites were chosen to be randomly distributed inside the sphere, corresponding to a homogeneous activity distribution. The emission spectrum of 99m Tc was taken from Howell [17] (Table 1). An S value for self-irradiation was obtained by simulating 1 million trajectories and scoring the energy deposition inside the target volume. The total dose was calculated as the product of S and the time-integrated activity. It must be noted that no plasmid binding of radioactivity was considered, because that would require nanodosimetric calculations for the plasmid DNA.

Calculation of SSBs and DSBs
The mean number of single-strand breaks (N SSB ) and doublestrand breaks (N DSB ) per plasmid were calculated based on the theoretical considerations published by Cowan et al. [18]. N DSB was estimated based on the relative proportion L of linear plasmids: N SSB was calculated using the fraction SC of supercoiled plasmids:

Preparation of 99m Tc-HYNIC-DAPI
The yield and stability of 99m Tc-HYNIC-DAPI in the reaction solution (saline) were determined using reverse-phase HPLC ( Figure 2). The reaction yield of 99m Tc-HYNIC-DAPI increased to .95% in one hour. 99m Tc-HYNIC-DAPI exhibited low stability in the kit solution (50% at 6 h, ,1% after 12 h) at room temperature. Therefore, the peak at a shorter retention time (1.33 min) in the HPLC trace, representative of molecules such as pertechnetate, increased from ,5% at 1 h to approximately 100% at 12 h.

Dose calculations
The S value for the dose calculations was 2.53*10 210 Gy/(Bq s) inside a sphere with a 10-mL volume and a homogeneous activity distribution. Therefore, incubating plasmids with 1000 MBq/mL of 99m Tc in 10 mL for 2 h yields a dose of 16.3 Gy. Irradiation for 24 h produces a total dose of 74 Gy.

Strand breaks in pUC19 plasmid DNA
The fractions of the different conformations of plasmid DNA (open circular, linear and supercoiled) that were observed after 2 or 24 h of irradiation with 99m Tc-pertechnetate or 99m Tc-HYNIC-DAPI are presented in Figure 3. For both compounds, an increase in applied activity enhanced the plasmid damage, resulting in an increase in the open circular and linear fractions and a decrease in the supercoiled fraction. The DNA damage was more extensive after 24 h of irradiation than after 2 h of irradiation. Compared with 99m Tc-pertechnetate, the radiotoxicity of 99m Tc-HYNIC-DAPI was obviously higher. For example, after  (Table 2). In the example above, the number of SSBs created by 99m Tc-HYNIC-DAPI (1.03) was twice that elicited by 99m Tc-pertechnetate (0.51). The number of DSBs increased fivefold (from 0.02 to 0.10) in the 99m Tc-HYNIC-DAPI-treated samples compared with the 99m Tc-pertechnetatetreated samples. The ratio of the number of SSBs to the number of DSBs decreased from 25 for 99m Tc-pertechnetate to 10 for 99m Tc-HYNIC-DAPI.
To assess the chemotoxicity of unlabeled HYNIC-DAPI, we performed experiments with an identical setup using the same ligand concentrations as those used for 99m Tc-HYNIC-DAPI. The results revealed that supercoiled plasmid DNA remained completely intact throughout all the incubations in the presence or absence of DMSO. Thus, no chemotoxic effects were observed (data not shown).
To  circular DNA to only 3.0%. Furthermore, the formation of DSBs was completely prevented when the 99m Tc-pertechnetate irradiation occurred in the presence of DMSO, as indicated by the absence of a linear DNA fraction ( Figure 5). The dependence of SSBs and DSBs on the total number of disintegrations in the presence or absence of DMSO is illustrated in Figure 6.

Estimation of plasmid-bound activity
After quantifying the DNA damage by ethidium bromide staining and image processing, the DNA bands representing each plasmid conformation (supercoiled, open circular and linear) were excised, and the DNA-bound activity in the 99m Tc-pertechnetate and 99m Tc-HYNIC-DAPI samples was measured using a gamma counter. The bound activity was quantitated using decaycorrected values. 99m Tc-pertechnetate-irradiated plasmid DNA exhibited a low amount of bound activity (,0.005 MBq) that was independent of the administered activity. In contrast, 99m Tc-HYNIC-DAPI demonstrated a correlation between applied activity and DNA-bound activity, reaching a maximum value of 0.72 MBq in the sample dosed with 1000 MBq/mL (Figure 7). When plotting the bound activity against the fractions of open circular DNA or linear DNA (Figure 8), linear correlations were apparent. These correlations indicated that increased plasmid damage resulted from increased 99m Tc-HYNIC-DAPI accumulation.

Discussion
DAPI is a DNA-specific probe that forms a fluorescent complex by inserting into the minor groove of AT-rich DNA sequences [15]. The HYNIC linker is widely used for labeling bioactive substances with 99m Tc without a loss of avidity [19]. We found that unlabeled HYNIC-DAPI was not chemotoxic at the tested low concentrations (data not shown).
The plasmids were incubated with 99m Tc-HYNIC-DAPI in excess of unlabeled DAPI, which can be explained by stoichiometric calculations. Radiochemistry labeling was performed using 1500 MBq 99m Tc and 5 mg HYNIC-DAPI. HYNIC-DAPI has a molar mass of 943 g/mol, which corresponds to a ratio of 4.7*10 13 atoms of 99m Tc to 3.2*10 15 molecules of HYNIC-DAPI. Assuming 100% turnover, the reaction would yield 1 labeled 99m Tc-HYNIC-DAPI molecule per 67 non-radioactive HYNIC-DAPI molecules. Here, the term non-radioactive refers to HYNIC-DAPI that was unlabeled or that was labeled with 99 Tc, which is considered nonradioactive because of its long half-life (2.1*10 5 years). Two hundred nanograms of pUC19 plasmid DNA contain 6.9*10 10 plasmid molecules. When incubating 1 MBq 99m Tc-HYNIC-DAPI with 200 ng of pUC19 plasmid DNA, 0.45 99m Tc-HYNIC-DAPI molecules would bind to each plasmid molecule on average. 99m Tc-pertechnetate bound to plasmid DNA to a much lesser extent (by approximately 100-fold less, data not shown), suggesting that the binding of 99m Tc-HYNIC-DAPI to DNA was a specific event. DAPI molecules bind to double-stranded DNA in the minor groove. The distance between two minor grooves is 10 base pairs; therefore, we assumed that there were 270 binding sites per plasmid for 99m Tc-HYNIC-DAPI. Considering the competition by non-radioactive HYNIC-DAPI, a maximum of 4 99m Tc-HYNIC-DAPI molecules could bind to each pUC19 plasmid DNA on average. For our labeling procedure, 4 99m Tc-HYNIC-DAPI molecules per plasmid corresponded to the application of 10 MBq (500 MBq/mL) of 99m Tc-HYNIC-DAPI. However, our data did not demonstrate the saturation of plasmid damage at higher activities (Figure 3), indicating that not every available HYNIC-DAPI molecule bound to plasmid DNA.
The DNA damage caused by unbound 99m Tc-HYNIC-DAPI was expected to be comparable to that caused by 99m Tcpertechnetate. In total, 1000 MBq/mL of unbound, but homogeneously distributed, 99m Tc in a volume of 10 mL leads to an absorbed dose of 74 Gy within 24 h. According to Table 2, this dose produced 0.5 SSBs and 0.02 DSBs on average in the absence of DMSO (24 h, 1000 MBq/mL, 99m Tc-pertechnetate). In the presence of DMSO, the formation of SSBs and DSBs was completely prevented (0.03 SSBs, no DSBs). These data suggested that the DNA damage was radical mediated and an indirect effect of unbound 99m Tc.
Within 24 h, 500 MBq/mL of unbound 99m Tc ( Figure 3) caused 26% of the DNA to adopt an open circular configuration. The same percentage was induced by 125 MBq/mL of 99m Tc-HYNIC-DAPI, which corresponds to a 4-fold enhancement. The potential for plasmid linearization was approximately ten times higher for 99m Tc-HYNIC-DAPI because 50 MBq/mL yielded 2% linear plasmid, whereas 500 MBq/ml of 99m Tc-pertechnetate was required to elicit the same effect. A higher linear plasmid fraction could not be induced by 99m Tc-pertechnetate because of limited specific activity.
We did not perform X-ray irradiation; hence, we could not determine the relative biological effectiveness of 99m Tc-HYNIC-DAPI and 99m TcO 4 2 . Nevertheless, we compared the biological effectiveness of each compound, which we defined as the ratio of the activities of 99m Tc-HYNIC-DAPI versus 99m TcO 4 2 that caused the same effect. Based on this definition, the biological effectiveness increased by approximately 4-fold in inducing SSBs and by approximately 10-fold in inducing DSBs, which is in the range reported for Auger electrons emitted in close proximity to DNA [20].
The SSBs created by unbound 99m Tc-pertechnetate could be prevented by DMSO. In contrast, DSBs and SSBs caused by 99m Tc-HYNIC-DAPI could not be prevented by DMSO. This indicated that the impact of the Auger events was direct or indirect with an extremely high local energy transfer that could not be antagonized by DMSO. SSBs are usually assumed to be radicalmediated; however, we demonstrated that the induction of SSBs by DNA-bound 99m Tc-HYNIC-DAPI might occur via another mechanism. It is possible that SSBs were also caused by direct interactions between emitted low-energy electrons and plasmidbound 99m Tc-HYNIC-DAPI that merely damaged a single strand. On average, 4.9 electrons are emitted by the decay of 99m Tc via internal conversion and gamma emission. This event creates a dramatic charge transfer in the electron shell, which can cause a Coulomb explosion of the molecule, as discussed by Pomplun and Sutmann [21].
The absorbed dose was calculated based on the absorbed energy inside an incubation volume of 10 mL. Thus, no regional heterogeneities, such as DNA-bound activity, were considered. Therefore, the same amount of 99m Tc yielded the same absorbed dose regardless of whether 99m Tc-pertechnetate or 99m Tc-HYNIC-DAPI was used. We expected that the dose to the plasmids was higher with 99m Tc-HYNIC-DAPI due to the emission of low-energy electrons. 99m Tc emits an average of 4.9 electrons with less than 2.7 keV of energy and a maximum range of 300 nm [17]. However, the term 'absorbed dose' is a macroscopic quantity that does not account for stochastic effects. Hence, we did not calculate the absorbed dose on plasmid level because 99m Tc-HYNIC-DAPI binding is a random process, as are radioactive decay and the direction of radiation emission.
The ratio of SSBs to DSBs induced by irradiation is approximately 40:1 per gray of gamma radiation [22]; this ratio decreases with increasing dose because SSBs are converted to DSBs. High LET irradiation induces more DSBs because the high density of radicals increases the possibility of DSBs being generated by chance. We calculated the SSB and DSB yields based on the conversion of supercoiled plasmid DNA into linear DNA. The numbers of SSBs and DSBs per plasmid molecule based on the number of decays over 2 h are plotted in Figure 6   and 29.6*10 13 decays/mL for various 125 I compounds in the absence of DMSO [2]. The yield of DSBs per decay of 99m Tc was calculated as the product of the plasmid concentration and the number of DSBs per decay/mL. Although Balagurumoorthy obtained DSB yields ranging from 0.10 to 0.52 per decay of 125 I, our data demonstrated a DSB yield of 0.03 per decay of 99m Tc. Therefore, our observed number of decays/mL corresponded with the data published by Balagurumoorthy, although the DSB yield for 99m Tc appeared to be a tenth of that for 125 I because of the lower plasmid amount. We irradiated 200 ng of pUC19 plasmid DNA in a volume of 20 mL, which corresponds to 3.44*10 12 plasmid molecules/mL, whereas Balagurumoorthy used 3.06*10 13 molecules/mL.
A more reliable calculation of the DSB yield per plasmid-bound decay requires knowledge of the amount of plasmid-bound radioactivity. Furthermore, the radiochemical stability of 99m Tc-HYNIC-DAPI will introduce variability in the number of decays per plasmid. The plasmid-bound applied activity varied between 2% and 10% in different experiments. Assuming 5% plasmidbound applied activity, we calculated 0.5 to 1.5 DSBs per plasmidbound decay. Humm and Charlton calculated the probability of DSB formation due to DNA-bound Auger electron emission for 125 I, 123 I and 99m Tc to be 1.1, 0.73 and 0.43, respectively [23].
Several studies have characterized the influence of the distance between the decaying isotope and DNA on DSB yield. Balagurumoorthy et al. synthesized a series of 125 I-labeled Hoechst derivatives with a minor groove-binding motif and an increasing distance between the radioactive molecule and the binding position (ranging from 10.5 Å to 13.9 Å ) [2]. The DSB yield decreased with distance from 0.2 DSBs/decay to 0.025 DSBs/ decay, demonstrating the crucial influence of relatively small distances. Although DMSO could not prevent the formation of DSBs at the closest distance, it reduced DSBs by 10-fold at greater distances. Our data did not indicate an effect of DMSO on DNAbound 99m Tc-HYNIC-DAPI. Therefore, we assumed that the radioactive molecule was below the critical distance from the DNA helix, even if stereoisomers could exist.
Walicka et al. investigated the effect of different DMSO concentrations (0.26 to 3 M) on the radiotoxicity of the Auger electron emitter 125 I [24]. The authors concluded, that the most radiotoxic effects from DNA-incorporated 125 I were due to indirect mechanisms. In contrast to our experiments Walicka et al. performed experiments using V79 cells, whereas we used plasmids as biological model without any repair or radioprotection by other cellular structures. Additionally, the authors irradiated their cells at 2135uC what could be cryotoxic to cells. Balagurumoorthy et al. found that in the absence of DMSO, 125 IEH-induced DSBs in nicked or linear DNA were caused by both direct and indirect mechanisms, whereas in its presence, DSBs occurred predominantly by direct ionization of DNA [25].

Conclusion
The direct binding of 99m Tc-HYNIC-DAPI to plasmid DNA was demonstrated and quantified. Plasmid-bound 99m Tc-HYNIC-DAPI induced SSBs and DSBs with high efficiency. DNA damage could not be prevented by the radical scavenger DMSO, suggesting that the damage was a direct effect. This statement is true for both DSBs and SSBs. The biological effectiveness increased by approximately 4-fold for inducing SSBs and approximately 10-fold for inducing DSBs, which was within the range reported for Auger electrons emitted in close proximity to DNA.