- Email: [email protected]
Cleavage of cellular DNA by calicheamicin γ1
DNA Repair 2 (2003) 363–374
Cleavage of cellular DNA by calicheamicin ␥1 Kecke Elmroth a , Jonas Nygren b , Susanne Mårtensson c , Ismail Hassan Ismail c , Ola Hammarsten c,∗ a
c
Deptartment of Oncology, Göteborg University, Sahlgrenska University Hospital, SE-413 45 Göteborg, Sweden b Södertörns Högskola and Karolinska Institute, CNT, Novum, SE-141 57 Huddinge, Sweden Department of Clinical Chemistry, Göteborg University, Sahlgrenska University Hospital, 41345 Göteborg, Sweden Received 16 July 2002; received in revised form 14 November 2002; accepted 18 November 2002
Abstract It is assumed that the efficient antitumor activity of calicheamicin ␥1 is mediated by its ability to introduce DNA double-strand breaks in cellular DNA. To test this assumption we have compared calicheamicin ␥1-mediated cleavage of cellular DNA and purified plasmid DNA. Cleavage of purified plasmid DNA was not inhibited by excess tRNA or protein indicating that calicheamicin ␥1 specifically targets DNA. Cleavage of plasmid DNA was not affected by incubation temperature. In contrast, cleavage of cellular DNA was 45-fold less efficient at 0 ◦ C as compared to 37◦ due to poor cell permeability at low temperatures. The ratio of DNA double-strand breaks (DSB) to single-stranded breaks (SSB) in cellular DNA was 1:3, close to the 1:2 ratio observed when calicheamicin ␥1 cleaved purified plasmid DNA. DNA strand breaks introduced by calicheamicin ␥1 were evenly distributed in the cell population as measured by the comet assay. Calicheamicin ␥1-induced DSBs were repaired slowly but completely and resulted in high levels of H2AX phosphorylation and efficient cell cycle arrest. In addition, the DSB-repair deficient cell line Mo59J was hyper sensitive to calicheamicin ␥. The data indicate that DSBs is the crucial damage after calicheamicin ␥1 and that calicheamicin ␥1-induced DSBs are recognized normally. The high DSB:SSB ratio, specificity for DNA and the even damage distribution makes calicheamicin ␥1 a superior drug for studies of the DSB-response and emphasizes its usefulness in treatment of malignant disease. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Calicheamicin ␥1; DNA double-strand break; DNA single-strand break; DNA damage; Ionizing radiation; DNA repair; Bleomycin
1. Introduction
Abbreviations: Cch1, calicheamicin ␥1; DSB, double-strand break; SSB, single-strand break; DTT, dithiothreitol; S.D., standard deviation; S.E.M., standard error of the mean; PBS, phosphate buffered saline; DNA-PK, DNA-dependent protein kinase; BSA, bovine serum albumin ∗ Corresponding author. Tel.: +46-31-3421561; fax: +46-31-828458. E-mail address: [email protected] (O. Hammarsten).
The study of DSB-specific responses is hampered by the fact that classical DNA double-strand break (DSB) inducing agents such as ionizing radiation and bleomycin mostly induce DNA single-stranded breaks (SSBs) and only a few percent of the DNA damage is DSBs [1]. Therefore, the cellular responses to these agents comprise a complex mixture of signals to SSBs, base damage, DSBs and oxidative damage to other molecules in the cell. In addition, there
1568-7864/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 1 5 6 8 - 7 8 6 4 ( 0 2 ) 0 0 2 3 5 - 5
364
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
is now a growing body of evidence that radiation therapy kills cancer cells by induction of DSBs [2] whereas the more frequent SSBs contribute little to the anti-tumor effect. To find ways to improve cancer treatment and DSB-response studies it will be important to identify agents that preferentially introduce DSBs in cells. One candidate drug in this class is the enediyne antibiotic calicheamicin ␥1 [3,4]. Several reports indicate that 50% of the strand breaks produced by calicheamicin ␥1 are DSBs when cleavage is performed on purified DNA in vitro [5,6]. The mechanism behind the efficient DSB induction is that calicheamicin ␥1 contains two radical centers that become positioned close to the backbone when the drug binds the minor groove in double-stranded DNA. Co-operative activation of the two radical centers occurs after reduction of a trisulfide in the molecule, resulting in efficient DSB induction [6–9]. A similar cleavage pattern has been assumed to occur when calicheamicin ␥1 cleaves cellular DNA. It is, however, possible that cellular DNA is cleaved differently from purified DNA in vitro. First, growth medium and serum components could affect the activity of calicheamicin ␥1. Second, passage of calicheamicin ␥1 through cellular membranes and redox-status in the cell could affect distribution and activation of calicheamicin ␥1. DNA associated proteins such as histones could result in differences in cleavage patterns from purified DNA [10,11]. In addition, DSBs induced by calicheamicin ␥1 frequently display abnormal nucleotides at both strands [12,13] that could affect how cells detect and repair these DSBs. It is, therefore, important to carefully define how calicheamicin ␥1 affects cellular DNA to validate its application in the clinic and in DSB-response studies. This information could help in the optimization of calicheamicin ␥1-linked antibody regimens [14] that have proven effective against acute myeloid leukemia [15–18], renal cell carcinoma [19] and ovarian cancer [20]. To study the cleavage of cellular DNA by calicheamicin ␥1 we have treated human primary fibroblasts with calicheamicin ␥1 under different conditions and measured DNA strand breaks and cellular responses. We found that calicheamicin ␥1 induced DSBs with high efficiency in cellular DNA eliciting a strong DSB-response and cell cycle arrest. Our findings indicate that calicheamicin ␥1 is a superior agent
for DSB-response studies and explains the extreme toxicity of this drug.
2. Materials and methods 2.1. Cells and reagents Calicheamicin ␥1 was a generous gift from George Ellestad (Wyeth-Ayers Research). The drug was dissolved at 2 mM in DMSO (Sigma) and stored at −70 ◦ C. Under these conditions, calicheamicin ␥1 showed no loss of activity after 12 months. Wortmannin (Sigma) was dissolved at 1 mg/ml in DMSO and stored at −70 ◦ C. Bleomycin (Astra Medica) was dissolved in distilled water at 15 000 U/ml and stored at −70 ◦ C. The tRNA was dissolved at 10 mg/ml in 10 mM Tris–HCl pH 7.5, 1 mM EDTA (TE-buffer). Herring sperm DNA (Sigma) was dissolved in TE-buffer and sonicated with a probe sonicator on ice at maximal output for 30 min. Agarose gel analysis indicated the medium size of the DNA fragments was 300 bases and no DNA above 2000 bp could be detected. Bovine serum albumin, fraction V (Sigma) was dissolved at 100 mg/ml in TE-buffer. The reduced form of glutathione and dithiothreitol (DTT) (Sigma) was dissolved at 0.5 M and 1 M concentration respectively in TE-buffer and stored at −20 ◦ C. Plasmid DNA (pUC13) was prepared using standard protocols. Primary human fibroblasts, Mo59K cells and Mo59J cells were grown in Dulbecco’s modified Eaagle medium with glutamax (Invitrogen) supplemented with 10% fetal calf serum, streptomycin and penicillin at 37 ◦ C in a humidified cell incubator with 5% CO2 . Cell counting was done using a Bürker chamber. Cellular DNA was labeled 4–5 days prior to each experiment by plating the fibroblasts at 30–50% confluency in growth medium supplemented with 1000 Bq/ml 2-14 C-thymidine (59.0 Ci/mmol, AmershamPharmacia). Confluent monolayers of primary human fibroblasts in 3.3 cm petri dishes were permeabilized after rinsing with PBS (100 mM NaPO4 pH 7.5, 100 mM NaCl) by incubating human primary fibroblasts for 1 h in 1 ml of permeabilizing buffer (142 mM KCl, 20 mM Tris–HCl pH 7.5, 10 mM MgCl2 , 5 mM DTT, 0.25% Triton X-100) supplemented with calicheamicin ␥1 as indicated in Fig. 5. Control human primary fibroblasts that were not permeabilized were incubated
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
in the same buffer without Triton X-100. Monolayers were irradiated on ice in PBS using a Philips RT 100 X-ray machine. An acceleration voltage of 100 kV was used for low dose irradiations (0–6 Gy) and 70 kV was used for high dose delivery (15–45 Gy).
365
by dividing the amount of linear or nicked plasmid DNA with the total amount of plasmid DNA in each lane. 2.4. Measurements of strand breaks in cellular DNA
2.2. Cleavage of cellular DNA Calicheamicin ␥1 was diluted in growth medium at 37 ◦ C and then conditioned to the appropriate temperature (0 or 37 ◦ C) for 1 h. Dilution of calicheamicin ␥1 in medium at 0 ◦ C resulted in poor reproducibility of the level of cellular DNA cleavage. Confluent monolayers of human fibroblasts in 3.3 cm petri dishes labeled with 14 C-thymidine were rinsed once with fresh growth medium and incubated at either 37 ◦ C or on ice for 10 min prior to addition of 1 ml preconditioned calicheamicin ␥1 dilutions. After incubation at the appropriate temperature for 1 h the cells were placed on ice and rinsed three times with PBS supplemented with 0.2 mg/ml sheared herring sperm DNA and then prepared for measurements of total DNA strand breaks or double-stranded DNA breaks within 1 h. Bleomycin (Astra medica, Sweden) was dissolved in water and diluted in growth media to the concentrations indicated and incubated with human primary fibroblasts at 0 ◦ C for 1 h prior to preparation for comet assay. 2.3. Cleavage and analysis of plasmid DNA Calicheamicin ␥1 cleavage of plasmid DNA (0.1 mg/ml) was performed in TE buffer supplemented with 5 mM DTT as indicated in figure legends and various concentrations of calicheamicin ␥1 as indicated in the Figs. 1 and 2. Incubation was for 30 min on ice (0 ◦ C) or 37 ◦ C as indicated. Cleaved plasmid DNA was resolved on 1% agarose gels with 0.5 × TAE (1 × TAE is 40 mM tris acetate, 1 mM EDTA supplemented with 1 mg/ml ethidium bromide) at 16 V/cm and gels were analyzed on a fluorescence scanner (Typhoon 9200, AmershamPharmacia). The relative amount of supercoiled (uncleaved), linear (one DSB per plasmid) and nicked (only SSBs) plasmid DNA was quantified from the images using the Imagequant software (AmershamPharmacia). The relative levels of DSBs (linear plasmid DNA) and SSBs (nicked plasmid DNA) were calculated
The DSB analysis was performed as described previously [21]. Briefly, the monolayer was rinsed with ice-cold PBS supplemented with 0.2 mg/ml sheared herring sperm DNA after exposure to drugs or X-rays and the cells were scraped off the plates with a rubber scrape. The suspension was mixed with agarose and transferred to a plug mold. The cells in the plug were lysed at 37 ◦ C for 48 h, washed and resolved by agarose gel electrophoresis (0.7%) in 1×TAE at 18 ◦ C for 17 h, using constant field gel electrophoresis. The gel was stained with ethidium bromide and cut into lanes and wells. The activity of 14 C in each sample was measured in a liquid scintillation counter and the fraction of activity released (FAR) from the well into the lane was calculated. In the case of rejoining experiment, treated cells were washed two times with ice-cold PBS, pre-warmed medium was added and the monolayers were incubated for different time at 37 ◦ C at 5% CO2 . Total DNA strand breaks were analyzed by using the DNA unwinding assay described by [22] with some modifications [11]. Note that the unwinding technique is also sensitive to DSB although this is normally over-looked due to low SSB:DSB ratios. Briefly, 1 ml 0.03 M NaOH in 0.15 M NaCl was added gently to 175 l of the same cell suspension prepared for DSB analysis. The unwinding procedure was performed on ice, in darkness, for 30 min. After neutralization, single and double-stranded DNA was separated on hydroxyl apatite columns. The activity of 14 C in each sample was measured by liquid scintillation. 2.5. Calculation of the SSB:DSB ratio In order to calculate the SSB:DSB ratio after treatment with calicheamicin ␥1, strand breaks/cell was calculated. The number of DSBs was calculated using the formula of Blöcher [23]. The exclusion size was estimated to 5.7 Mbp and mean chromosomal size was set to 130 Mbp. The number of total DNA strand breaks was calculated using the formula
366
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
Fig. 1. Induction of DNA strand breaks in plasmid DNA by Cch1 in vitro: (A) DNA cleavage by Cch1 in vitro is dependent on reducing agents such as DTT. Plasmid DNA was mixed with Cch1 or DTT and the combination for 30 min at 0 ◦ C; (B) DNA cleavage in vitro by Cch1 is not affected by incubation temperature. Plasmid DNA was mixed with different concentrations of Cch1 in the presence of DTT for 30 min at 0 ◦ C or 37 ◦ C; (C) stability of calicheamicin ␥1 in growth medium. Calicheamicin ␥1 (20 M) was diluted in growth medium with 10% fetal calf serum supplemented or not supplemented with 5 mM DTT as indicated and incubated at 37 ◦ C for different times. DNA cleavage capacity was measured by mixing the pre-incubated calicheamicin ␥1 with plasmid DNA in presence of 5 mM DTT on ice for 30 min. The cleaved plasmid DNA was separated on agarose gels and the amount of undamaged, linear (one DSB per plasmid) and nicked (only SSBs) plasmid DNA was monitored.
k∗ − log(ds/(ds + ss)) = total DNA strand breaks per 109 Da, where ds is the fraction of double-stranded DNA and ss is the fraction of single-stranded DNA eluted from the columns, and k is a conversion factor. The conversion factor k = −31 from [24] and the assumption that a cell contains 3.83 × 1012 Da of DNA was used. The number of SSBs was calculated by subtracting DSBs from total DNA strand breaks. Using the methods of calculation described here on
gamma irradiated samples gave results in good agreement with published data [25]. 2.6. Comet assay The comet assay was performed as described [26] with some modifications [27]. For each slide, 5000–10000 human primary fibroblasts were mixed with 150 l 0.75% low-melting agarose (Sigma type
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
367
slides were then lysed [0.03 M NaOH; 1 M NaCl; 2 mM EDTA and 0.5% NLS] for 1 h and thereafter rinsed [0.03 M NaOH; 2 mM EDTA] for 1 h and finally, electrophoresis was run at 0.8 V/cm for 15 min in the same solution. The slides were then neutralized in 0.4 M Tris–HCl, pH 7.5. All slides were then rinsed in deionized water and air-dried. They were fixed in methanol, stained with 1 g/ml ethidium bromide for 5 min and rinsed briefly in TAE buffer. Analysis of DNA that migrated from the nucleus, the tail moment, was done using an Olympus fluorescence microscope, aided by the Kinetic Imaging Komet II system. 2.7. H2AX phosphorylation (γ-H2AX)
Fig. 2. Calicheamicin ␥1 specifically cleaves DNA: (A) calicheamicin ␥1 specifically targets DNA. Plasmid DNA (0.1 mg/ml) was mixed with sheared herring sperm DNA, tRNA or BSA at the concentrations indicated in the figure before addition of calicheamicin ␥1 (1.35 M) and DTT. The amount of uncleaved, supercoiled plasmid DNA was determined as described in Section 2 and used to calculate the relative level of plasmid DNA cleavage in each sample. Error bars represent data from double samples from a single experiment; (B) calicheamicin ␥1 does not react with protein or RNA. Calicheamicin ␥1 (0.7 M) was preincubated with BSA (1.7 mg/ml), tRNA (1.7 mg/ml), sheared herring sperm DNA (1 mg/ml) or without extra addition (buffer) in TE buffer supplemented with 1 mM DTT. After preincubation for 0, 1, 3, 10 or 30 min at 37 ◦ C, plasmid DNA (0.05 mg/ml) was added and the cleavage reaction was allowed to proceed for 30 min. The amount of cleaved plasmid DNA was determined as in (A).
VII, USA), held at 37 ◦ C. The agarose-cell suspension was spread into a thin layer on clear-glass slides (Menzel Superfrost, Germany), pre-treated with a small amount of agarose, and air-dried. The preparations were left on a chilled plate for 5 min and the
Confluent monolayers of human fibroblasts in 10 cm plates were incubated with 12 or 36 nM calicheamicin ␥1 at 37 ◦ C for 1 h. After washing with cold TBS (20 mM Tris–HCl pH 7.5, 137 mM NaCl) supplemented with 5 mM EDTA, cells were scraped off the plates, and lysed in 1 ml cell lysis buffer (10 mM EDTA, 20 mM Tris–HCl pH 8.0, 1% Nonidet P-40, 10 mM NaF, 1 mM NaVO4 , 1 mM NaMo4 , protease inhibitors cocktail (Roche) and 10 mM Okadaic acid (Alexis)) and nuclear pellet collected by centrifugation. Histones were acid extracted in 3 volumes of extraction buffer (0.5 M HCl + 10% glycerol + 0.1 M mercaptoethanol), precipitated with trichloroacetic acid (25%), separated on 15% SDS-PAGE and transferred to PVDF-membrane (Biorad). The membranes were blocked with blotto (5% non-fat milk, 10 mM Tris–HCl pH 7.5, 140 mM NaCl, 0.1% Nonidet P-40) and incubated with an antibody specific for the phosphorylated form of H2AX (g-H2AX) (Upstate Biotechnology) diluted 1:2000 in blotto. After incubation with secondary antibody (AmershamPharmacia) the membrane was developed with Dura ECL kit (Pierce) and analyzed on a digital camera based system (Chemidoc, Biorad). 2.8. Colony assay Confluent monolayers of Mo59J and Mo59K cells were treated with 0, 30, 100, 300, 1000 and 3000 pM calicheamicin ␥1 in growth medium at 37 ◦ C for 1 h. Cells were rinsed three times with PBS supplemented with 5 mM DTT and 0.2 mg/ml sheared herring sperm DNA before trypsination and replating at various cell
368
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
densities in 10 cm plates. The cells were stained with Giemsa after 14 days and colonies containing over 50 cells were counted.
3. Results 3.1. Clevage of purified plasmid DNA by calicheamicin γ1 To study cleavage of cellular DNA by calicheamicin ␥1 we first wanted to establish conditions resulting in efficient cleavage of purified plasmid DNA. Calicheamicin ␥1 contains a trisulfide that must be reduced to activate the two radical centers in the molecule. Therefore, DNA-cleavage requires a reducing agent such as dithiotreitol (DTT) or glutathione [6]. To examine the dependence on reduction we mixed calicheamicin ␥1 with purified plasmid DNA in the presence or absence of DTT. DNA cleavage was monitored on agarose gels that allow separation of linear (one DSB per plasmid), nicked (only SSB) and supercoiled (undamaged) plasmid DNA. As expected, calicheamicin ␥1 failed to cleave DNA in the absence of DTT whereas efficient cleavage was observed when both calicheamicin ␥1 and DTT were included in the reaction (Fig. 1A). We also found that calicheamicin ␥1 cleavage of plasmid DNA was not affected by incubation temperature (Fig. 1B). Under all cleavage conditions tested, the ratio between nicked (SSB) and linear (DSB) plasmid DNA was 1:1.8 indicating that for each DSB, about two SSBs were produced. Identical cleavage patterns were reproduced when glutathione, the predominant thiol in cells, was used as reductant (data not shown) [28]. Calicheamicin ␥1 showed no loss of DNA-cleavage activity after 24 h preincubation at 37 ◦ C in serum containing growth medium. In contrast, when DTT was included in the preincubation medium, calicheamicin ␥1 was rapidly inactivated (Fig. 1C). This indicated that calicheamicin ␥1 was stable in cell culture media whereas the reduced form of calicheamicin ␥1 was unstable. Calicheamicin ␥1 is inactivated when it cleaves DNA and can therefore only cleave once. Therefore, if calicheamicin ␥1 were to react with proteins or RNA, a significant portion of the calicheamicin ␥1 that enters the cell might be inactivated resulting in damage
to other cellular structures and inefficient DNA cleavage. To test if calicheamicin ␥1 was able to react with proteins or RNA we first mixed plasmid DNA with tRNA or bovine serum albumin (BSA) before addition of calicheamicin ␥1 and DTT. The cleavage reaction was then allowed to proceed to completion and the extent of plasmid cleavage analyzed. As expected, addition of excess low molecular weight DNA prevented calicheamicin ␥1 from cleaving plasmid DNA (Fig. 2A). In contrast, tRNA or BSA had little effect. We also tested if calicheamicin ␥1 was inactivated faster in the presence of tRNA or BSA compared to buffer alone. We therefore preincubated calicheamicin ␥1 for various times in a DTT containing buffer supplemented with tRNA, BSA or low molecular weight DNA. Remaining DNA cleavage capacity was monitored by addition of plasmid DNA at different time points and analysis of cleaved plasmid DNA on agarose gel (Fig. 2B). Calicheamicin ␥1 was inactivated with similar kinetics in buffer alone and in presence of tRNA. Addition of BSA apparently protected calicheamicin ␥1 since inactivation was slower in the presence of this protein. These data show that calicheamicin ␥1 does not react with protein or RNA. 3.2. Cleavage of cellular DNA by calicheamicin γ1 To study calicheamicin ␥1 cleavage of cellular DNA we measured the induction of DSBs in primary human fibroblasts at 37 and 0 ◦ C using the FAR-assay [21]. We also measured the number of total DNA strand breaks (SSB + DSB) produced at 0 ◦ C in the same cells to allow calculation of the relative number of DSBs and SSBs induced at 0 ◦ C [11,22]. As a comparison, the yields of X-ray-induced SSBs and DSBs in the same human primary fibroblasts were measured. The dose-response relationship for both total DNA strand breaks and DSBs at 0 ◦ C was linear, with the yields 32.4 total DNA strand breaks/nM and cell and 8.23 DSB/nM and cell (Fig. 3A). The corresponding R2-values for the linear part of the total DNA strand break curve was 0.96. For DSBs the R2-values were 0.99 for cleavage at 0 ◦ C and 0.98 for cleavage of cellular DNA at 37 ◦ C indicating a good linear correlation between calicheamicin ␥1 concentration and induction of DNA strand-breaks (Fig. 3A and B). For calicheamicin ␥1 the ratio was 2.94, such that for
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
369
each DSB, about 3 SSBs were induced. In contrast, X-rays introduced 20 SSBs for every DSB (data not shown), similar to what have been reported previously [29]. Thus, the relative DSB/SSB ratio for cleavage of cellular DNA was close to the two-fold difference observed when calicheamicin ␥1 cleaved purified plasmid DNA. In contrast to cleavage of purified plasmid DNA by calicheamicin ␥1, cleavage of cellular DNA was strongly influenced by incubation temperature. At 0 ◦ C, calicheamicin ␥1 cleaved cellular DNA 45-fold less effective compared to cleavage at 37 ◦ C. This difference may be due to low cell permeability of calicheamicin ␥1 at low temperatures. In line with this possibility, we found that permeabilization, using the detergent Triton X-100, enhanced induction of DSBs at 0 ◦ C (Fig. 3C). In this experiment the permeabilization buffer was supplemented with 5 mM DTT to allow activation of calicheamicin ␥1. Similar stimulation by permeabilization was obtained with 5 mM glutathione, the predominant thiol in cells (data not shown). 3.3. Even distribution of DNA damage levels in cells treated with calicheamicin γ1
Fig. 3. Cleavage of cellular DNA by calicheamicin ␥1: (A) induction of total DNA strand breaks and double-strand breaks in cells treated with different concentrations of calicheamicin at 0 ◦ C for 1 h. Each point represents the mean from three separate experiments with duplicate samples and the bars represent±S.E.M. when larger than the symbol; (B) induction of double-strand breaks in cells treated with different concentrations of calicheamicin ␥1 at 37 ◦ C for 1 h. Each point represents the mean from two separate experiments with duplicate samples and the bars represent±S.E.M. when larger than the symbol; (C) the number of double-strand breaks after treatment with 0 or 3 nM calicheamicin for 1 h in intact cells or permeabilized cells. One single experiment with duplicate samples. The bars show ± S.D.
To examine if the induction of damage by calicheamicin ␥1 was evenly distributed among the cells in a population, we analyzed DNA strand-break induction on 50 individual human primary fibroblasts using the comet assay. In this assay, DNA strand breaks are monitored by electrophoresis of cells molded into agarose. The amount of DNA that migrates out of the nucleus, the tail moment, correlates with extent of DNA strand breaks in individual cells. As a control, we also analyzed human primary fibroblasts treated with bleomycin that previously have been shown to give an uneven distribution of damaged cells, and X-rays that induces an even damage distribution. As shown in Fig. 4, damage-induced by calicheamicin ␥1 followed a normal distribution similar to that after X-rays. 3.4. Repair and detection of calicheamicin γ1-induced DSBs To study repair of calicheamicin ␥1-induced DSBs, monolayers of human fibroblasts were treated with calicheamicin ␥1 or X-rays at 0 ◦ C and then allowed to rejoin at 37 ◦ C. The level of remaining DSBs was
370
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
ing of calicheamicin ␥1-induced DSB was slower, probably reflecting the more complex DSBs produced by calicheamicin ␥1 [12,13]. Despite the slow rejoining, the level of DSBs was down to background levels after 24 h. This indicates that human primary fibroblasts are fully capable of repairing calicheamicin ␥1-induced DSBs.
Fig. 4. Cumulative frequency of DNA strand breaks, measured by the comet assay and expressed as tail moment in 50 individual cells irradiated with (A) 6 Gy of X-rays (B) 1 nM calicheamicin at 0 ◦ C for 1 h and (C) 135 U bleomycin at 0 ◦ C for 1 h. The level of strand breaks, the tail moment, (black circles) was compared to the expected values, normally distributed for the same mean and S.D. (line).
analyzed at different time points. We found that rejoining of calicheamicin ␥1-induced DSBs followed biphasic kinetics similar to what was found for X-ray-induced DSBs (Fig. 5A). However, the rejoin-
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
To examine the cellular response to calicheamicin ␥1 we measured phosphorylation of H2AX in calicheamicin ␥1 treated cells [30]. The histone 2A variant, H2AX, becomes phosphorylated in response to DSBs, but not by other types of DNA-damage [30,31]. Calicheamicin ␥1 was found to efficiently induce H2AX-phosphorylation in a dose dependent manner (Fig. 5B). In addition, human primary fibroblasts displayed a permanent growth arrest after a 1 h exposure of sub-nanomolar concentrations of calicheamicin ␥1 (Fig. 5C). This indicates that the DSBs produced by calicheamicin ␥1 induce a normal and strong DSB-response. We also tested if cells with defective DSB-repair were hypersensitive to calicheamicin ␥1. The Mo59J cell line is derived from a human glioblastoma tumor and lack expression of DNA-PKcs [32,33] and is therefore sensitive to ionizing radiation and defective in DSB-repair. The matching Mo59K cell line was isolated from the same tumor but have high DNA-PKcs expression and display normal sensitivity to ionizing radiation and DSB-repair [34,35]. Mo59J and Mo59K cells were exposed to different concentrations of calicheamicin ␥1 and surviving fraction was determined by colony forming assay. The result showed that Mo59J cells were 4-fold more sensitive to calicheamicin ␥1 compared to Mo59K cells (Fig. 6) supporting the notion that DSB is the biologically important lesion induced by calicheamicin ␥1.
䉳 Fig. 5. Cellular responses to calicheamicin ␥1: (A) rejoining of double-strand breaks, expressed as percentage of initial damage, in cells treated with 100 nM calicheamicin at 0 ◦ C for 1 h and then allowed to rejoin at 37 ◦ C for the indicated times (filled circles). Each point represents the mean from three separate experiments and the bars represent ± S.E.M. The rejoining after X-irradiation of the same cell line with 45 Gy is shown for reference (empty circles); (B) phosphorylation of H2AX (␥-H2AX) in response to calicheamicin ␥1. Cells were incubated with 12 or 45 nM calicheamicin ␥1 for 1 h and H2AX phosphorylation (␥-H2AX) examined with immunoblot (upper image). The same samples were separated on a separate SDS-PGE and stained with Coomassie to control for equal loading of extracted histones (lower image); (C) cell cycle arrest after calicheamicin ␥1 treatment. Confluent monolayers (95% ␥1 cells) human primary fibroblasts were treated with 0, 40 or 400 pM calicheamicin ␥1 for 1 h at 37 ◦ C before trypsination and replating at 1:50 dilution. Cells were trypsinized and counted using a Bürker chamber at the indicated times.
371
Fig. 6. Cells defective in DSB-repair are hypersensitive to calicheamicin ␥1. Mo59K (normal DSB-repair) and Mo59J (defective DSB-repair) cells were incubated with 0, 30, 100, 300, 1000 or 3000 pM calicheamicin ␥1 at 37 ◦ C for 1 h. The fraction of cells surviving this treatment (surviving fraction) was determined with the colony assay described in Section 2.
4. Discussion The data presented in this study indicate that calicheamicin ␥1 appears to do little else than DSBs in cellular DNA. First, more than 30% of the calicheamicin ␥1-induced strand breaks in cellular DNA were DSBs. Second, calicheamicin ␥1 is likely to specifically target DNA when it enters the cell, since tRNA or protein failed to interfere with plasmid DNA cleavage. In addition, calicheamicin ␥1 induced a strong DSB-response in human fibroblasts at concentrations that would deliver less than 1000 molecules/cell. Finally, calicheamicin ␥1 is unlikely to react with extracellular structures since the drug was stable under normal cell culture conditions. It is therefore possible that calicheamicin ␥1 will be a superior DSB-inducing agent in DSB-response studies and DSB-based cancer therapies. Previous studies have shown that when calicheamicin ␥1 cleaves purified DNA, 50% of DNA strand breaks are DSBs [5]. To measure the extent of DSB and SSB in cellular DNA we separately measured DSB and total DNA strand breaks (DSB + SSB) in calicheamicin ␥1-treated cells. To calculate the DSB/ SSB ratio it was necessary to recalculate the primary
372
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
signal from the methods to the number of DNA strand breaks per cell. The formula of Blöcher [23] was used to calculate the number of DBS/cell. This formula utilizes the fraction of DNA smaller than a certain cut off size and the assumption that damage is induced in a random fashion. It is not possible to use size markers in constant field gel electrophoresis to determine the cut off size. The fraction radioactivity released (FAR) for different doses of X-ray irradiation was therefore compared with published data for the fraction of DNA smaller than certain size markers as determined by pulsed field gel electrophoresis [36,37]. It was estimated that the cut off size for the electrophoresis conditions used here was about 5.7 Mbp. The yield of DSB/Gy calculated in this way from our data was in good agreement with published data [38]. Based on these calculations we found that the DSB/SSB ratio in cellular DNA was 1:3, close to the 1:2 ratio found when calicheamicin ␥1 cleaved purified plasmid DNA. A similar DSB/SSB ratio (1:5) was found after cleavage of cellular DNA with the enediyne antibiotic neocarzinostatin [39]. However, in this study the cells were treated with neocarzinostatin at 37 ◦ C for 1 h before measurements of DSBs and SSBs. A potential problem with this protocol is that repair of SSBs is much faster compared to repair of DSBs. Therefore, the treatment with neocarzionostatin under repair proficient conditions might underestimate the number of SSB due to quick repair of this lesion during treatment. This might result in a significantly increased DSB/SSB ratio. In contrast, we have calculated the DSB/SSB ratio for calicheamicin ␥1 cleavage of cellular DNA on ice, where no repair occurs. Therefore, to be able to compare the DSB/SSB ratio for calicheamicin ␥1 and neocarzinostatin, the cleavage of cellular DNA must be done using the same protocol. One important finding was that tRNA and protein essentially failed to interfere with cleavage of purified plasmid DNA. Since calicheamicin ␥1 is inactivated after cleavage, this finding indicates that the drug does not react with RNA or protein. One previous study has found that calicheamicin ␥1 binds double-stranded RNA and is able to cleave RNA-oligonucleotides [40]. However, in the light of our results calicheamicin ␥1 is expected to target cellular DNA without interference from proteins or RNA. This selectivity decreases the extent of damage to other cellular
structures and increases the DNA-cleaving efficiency of calicheamicin ␥1. We also found that calicheamicin ␥1 produced a strong DSB-response. First, sub-nanomolar concentrations of calicheamicin ␥1 induced a permanent cell cycle arrest in primary human fibroblasts. The number of DSBs induced at this concentration, calculated from the dose-response curve, equals that of an X-ray dose resulting in similar growth arrest (3–4 Gy). Therefore, our data indicate that the toxicity of calicheamicin-induced DSBs were equal to that of X-ray-induced DSBs although our data do not validate an exact comparison. Second, calicheamicin ␥1 efficiently induced H2AX phosphorylation. H2AX is phosphorylated specifically in response to DSBs [30]. In addition, our result show that human fibroblasts were fully capable of repairing calicheamicin ␥1-induced DSBs. In line with this finding, we have previously shown that DNA-PK, a DNA activated protein kinase involved in DSB-repair, is efficiently activated by calicheamicin-induced DSBs in vitro (submitted). We also found that Mo59J cells, that are defective in DSB-repair, were hypersensitive to calicheamicin ␥1. Therefore, calicheamicin ␥1induced DSBs were recognized normally. During the course of this study we identified several parameters important for reproducible cleavage of cellular DNA by calicheamicin ␥1. Induction of DSBs in cellular DNA was 45-fold less efficient at 0 ◦ C as compared to 37 ◦ C. It is therefore important that the temperature is controlled. The poor cleavage of cellular DNA at 0 ◦ C was not due to alteration in calicheamicin ␥1 interaction with DNA, since cleavage of purified plasmid DNA was not affected by incubation temperature. However, permeabilization of the cells stimulated cleavage of cellular DNA more than one-order of magnitude at 0 ◦ C. This indicates that passage of calicheamicin ␥1 through cellular membranes is inefficient at low temperatures. Hyperthermia has been used to enhance the effects of ionizing radiation. If the temperature effect reported in this paper were linear, a 16% increase in initial damage would be obtained at 43 ◦ C, the temperature used in hyperthermia treatment regimens. It is therefore possible that localized hyperthemia could increase the selective killing of cancer cells by calicheamicin ␥1. We also found that when cells were treated with calicheamicin ␥1 at 0 ◦ C, significant cleavage of cellular DNA
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
occurred after the calicheamicin ␥1 solution was carefully washed away from the cells. This effect was alleviated by addition of sheared herring-sperm DNA in the cold PBS used to wash the cells after treatment with calicheamicin ␥1 (data not shown). We speculate that a microprecipitate of calicheamicin ␥1 was deposited on cellular membranes specifically at low temperatures that cannot be washed away completely unless DNA is included in the washing solution. Calicheamicin ␥1 was stable in normal growth medium for at least 24 h but quickly inactivated if thiols were present. Addition of DTT to the medium therefore resulted in inefficient cleavage of cellular DNA (data not shown). Importantly, addition of 5 mM DTT is an efficient way to detoxify calicheamicin ␥1 containing solutions. In summary, our data indicate that the antitumor activity of calicheamicin ␥1 is mediated by its efficient DSB-inducing activity in cells. Therefore, when used properly, calicheamicin ␥1 holds a great promise in development of cancer therapies and will allow for a more selective study of the DSB-response in human cells.
Acknowledgements We thank Henrik Zetterberg for careful reading of the manuscript. KE was supported by the King Gustav V Jubilee Clinic Cancer Research Foundation. JN was supported by Swedish Radiation Protection Agency. OH was supported by Swedish Cancer Society, the Swedish Medical Research Council, King Gustav V Jubilee Clinic Cancer Research Foundation and Sahlgren’s Hospital Research Foundation. References [1] A.H.W. Nias, Subcellular radiobiology, in: An Introduction to Radiobiology, Wiley, New York, 1998, pp. 63–79. [2] O. Tounekti, A. Kenani, N. Foray, S. Orlowski, L.M. Mir, The ratio of single- to double-strand DNA breaks and their absolute values determine cell death pathway, Br. J. Cancer 84 (2000) 1272–1279. [3] M.D. Lee, T.S. Dunne, M.M. Siegel, C.C. Chang, G.O. Morton, D.B. Borders, Calicheamicins, a novel family of antitumor antibiotics, 1 chemistry and parial strucure of calicheamicin ␥1, J. Am. Chem. Soc. 109 (1987) 3464–3466.
373
[4] M.D. Lee, G.A. Ellestad, D.B. Borders, Calicheamicins: discovery, structure, chemistry and interaction with DNA, Acc. Chem. Res. 24 (1991) 235–243. [5] P.C. Dedon, A.A. Salzberg, J. Xu, Exclusive production of bistranded DNA damage by calicheamicin, Biochemistry 13 (1993) 3617–3622. [6] N. Zein, A. Sinha, W. McGahren, G. Ellestad, Calicheamicin ␥: an antitumor antibiotic that cleaves double-stranded DNA site specifically, Science 240 (1988) 1198–1201. [7] N. Zein, M. Poncin, R. Nilakantan, G.A. Ellestad, Calicheamicin ␥1 I and DNA: molecular recognition process responsible for site-specificity, Science 244 (1989) 697–699. [8] Z. Xi, I.H. Golberg, DNA-damaging enediyne compounds, in: D. Barton, K. Nakanishi (Eds.), Comprehensive Natural Products Chemistry, Pergamon Press, Oxford, 1999. [9] D.M. Lopez-Larraza, K. Moore, P.C. Dedon, Thiols alter the partitioning of calicheamicin-induced deoxyribose 4/coxidation reactions in the absence of DNA radical repair, Chem. Res. Toxicol. 14 (2001) 528–535. [10] L. Yu, A.A. Salzberg, P.C. Dedon, New insights into calicheamicin-DNA interactions derived from a model nucleosome system, Bioorg. Med. Chem. 3 (1995) 729–741. [11] J. Nygren, M. Ljungman, G. Ahnstrom, Chromatin structure and radiation-induced DNA strand breaks in human cells: soluble scavengers and DNA-bound proteins offer a better protection against single-than double-strand breaks, Int. J. Radiat. Biol. 68 (1995) 11–18. [12] M.A. Chaudhry, P.C. Dedon, D.M. Wilson III, B. Demple, M. Weinfeld, Removal by human apurinic/apyrimidinic endonuclease 1 (Ape 1) and Escherichia coli exonuclease III of 39-phosphoglycolates from DNA treated with neocarzinostatin, calicheamicin, and ␥-radiation, Biochem. Pharmacol. 57 (1999) 531–538. [13] J.J. Hangeland, J.J. De Voss, J.A. Heath, C.A. Townsend, W.D. Ding, J.S. Ashcroft, G.A. Ellestad, Specific abstraction of the 59(S) and 49-deoxyribosyl hydrogen atoms from DNA by calicheamicin ␥1 I, J. Am. Chem. Soc. 114 (1992) 9200– 9202. [14] P.R. Hamann, L.M. Hinman, I. Hollander, C.F. Beyer, D. Lindh, R. Holcomb, W. Hallett, H.R. Tsou, J. Upeslacis, D. Shochat, A. Mountain, D.A. Flowers, I. Bernstein, Gemtuzumab ozogamicin, a potent and selective anti-CD33 antibody-calicheamicin conjugate for treatment of acute myeloid leukemia, Bioconjug. Chem. 13 (2002) 47–58. [15] P.F. Bross, J. Beitz, G. Chen, X.H. Chen, E. Duffy, L. Kieffer, S. Roy, R. Sridhara, A. Rahman, G. Williams, R. Pazdur, Approval summary: gemtuzumab ozogamicin in relapsed acute myeloid leukemia, Clin. Cancer Res. 6 (2001) 1490– 1496. [16] R. Larson, Current use and future development of gemtuzumab ozogamicin, Semin. Hematol. 38 (2001) 24–31. [17] M.L. Linenberger, T. Hong, D. Flowers, E.L. Sievers, T.A. Gooley, J.M. Bennett, M.S. Berger, L.H. Leopold, F.R. Appelbaum, I.D. Bernstein, Multidrug-resistance phenotype and clinical responses to gemtuzumab ozogamicin, Blood 98 (2001) 988–994.
374
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
[18] I.D. Bernstein, CD33 as a target for selective ablation of acute myeloid leukemia, Clin. Lymphoma 1 (Suppl.) (2002) S9–S11. [19] K. Knoll, W. Wrasidlo, J.E. Scherberich, G. Gaedicke, P. Fischer, Targeted therapy of experimental renal cell carcinoma with a novel conjugate of monoclonal antibody 138H11 and calicheamicin thetaI1, Cancer Res. 60 (2000) 6089–6094. [20] A.M. Gillespie, T.J. Broadhead, S.Y. Chan, J. Owen, A.P. Farnsworth, M. Sopwith, R.E. Coleman, Phase I open study of the effects of ascending doses of the cytotoxic immunoconjugate CMB-401 (hCTMO1-calicheamicin) in patients with epithelial ovarian cancer, Ann. Oncol. 11 (2001) 735– 741. [21] K. Elmroth, J. Nygren, L.J. Erkell, R. Hultborn, Radiationinduced double-strand breaks in mammalian DNA: influence of temperature and DMSO, Int. J. Radiat. Biol. 76 (2000) 1501–1508. [22] G. Ahnström, K. Erixon, Measurement of strand breaks by alkaline denaturation and hydroxyapatite chromatography, in: E.C. Friedberg, P.C. Hanawalt (Eds.), DNA Repair, Marcel Dekker, New York, 1980, pp. 403–418. [23] D. Blöcher, In CHEF electrophoresis a linear induction of dsb corresponds to a non-linear fraction of extracted DNA with dose, Int. J. Radiat. Biol. 57 (1990) 7–12. [24] M. Ljungman, The influence of chromatin structure on the frequency of radiation-induced DNA strand breaks: a study using nuclear and nucleoid monolayers, Radiat. Res. 126 (1991) 58–64. [25] J.F. Ward, DNA damage produced by ionizing radiation in mammalian cells: identities, mechanisms of formation, and reparability, Prog. Nucl. Acid Res. Mol. Biol. 35 (1988) 95– 125. [26] P.L. Olive, J.P. Banath, R.E. Durand, Heterogeneity in radiation-induced DNA damage and repair in tumour and normal cells measured using the comet assay, Radiat. Res. 122 (1990) 69–72. [27] Y. Yang, W.J. Griffiths, M. Nordling, J. Nygren, L. Möller, J. Bergman, E. Liepinsh, G. Otting, J.-Å. Gustafsson, J. Rafter, J. Sjövall, A major metabolic product of benzo[a]pyrene in the germfree rat is produced by ring opening, Biochemistry 39 (2000) 15585–15591. [28] P.C. Dedon, I.H. Goldberg, Influence of thiol structure on neocarzinostatin activation and expression of DNA damage, Biochemistry 31 (1992) 1909–1917.
[29] J.F. Ward, DNA damage produced by ionizing radiation in mammalian cells: identities, mechanisms of formation, and reparability, Prog. Nucl. Acid Res. Mol. 35 (1988) 95–125. [30] E.P. Rogakou, D.R. Pilch, A.H. Orr, V.S. Ivanova, W.M. Bonner, DNA double-stranded breaks induce histone H2AX phosphorylation on serine 139, J. Biol. Chem. 273 (1998) 5858–5868. [31] T.T. Paull, E.P. Rogakou, V. Yamazaki, C.U. Kirchgessner, M. Gellert, W.M. Bonner, A critical role for histone H2AX in recruitment of repair factors to nuclear foci after DNA damage, Curr. Biol. 10 (2000) 886–895. [32] C.W. Anderson, J.J. Dunn, P.I. Freimuth, A.M. Galloway, M.J. Allalunis-Turner, Frameshift mutation in PRKDC, the gene for DNA-PKcs, in the DNA repair-defective, human, glioma-derived cell line M059J, Radiat. Res. 156 (2001) 2–9. [33] B.S. Hoppe, R.B. Jensen, C.U. Kirchgessner, Complementation of the radiosensitive M059J cell line, Radiat. Res. 153 (2000) 125–130. [34] S. Lees-Miller, R. Godbout, D. Chan, M. Weinfeld, R. Day, G. Barron, J. Allalunis-Turner, Absence of p350 subunit of DNA-activated protein kinase from a radiosensitive human cell line, Science 267 (1995) 1183–1185. [35] S.J. DiBiase, Z.C. Zeng, R. Chen, T. Hyslop, W.J.J. Curran, G. Iliakis, DNA-dependent protein kinase stimulates an independently active, non-homologous, end-joining apparatus, Cancer Res. 60 (2000) 1245–1253. [36] R.S. Wong, J.R. Dynlacht, B. Cedervall, W.C. Dewey, Analysis by pulsed-field gel electrophoresis of DNA doublestrand breaks induced by heat and/or X-irradiation in bulk and replicating DNA of CHO cells, Int. J. Radiat. Biol. 68 (1995) 141–152. [37] K. Erixon, B. Cedervall, Linear induction of DNA doublestrand breakage with X-ray dose, as determined from DNA fragment size distribution, Radiat. Res. 142 (1995) 153–162. [38] E. Hoglund, E. Blomquist, J. Carlsson, B. Stenerlow, DNA damage induced by radiation of different linear energy transfer: initial fragmentation, Int. J. Radiat. Biol. 76 (2000) 539–547. [39] T. Hatayama, I.H. Goldberg, DNA damage and repair in relation to cell killing in neocarzinostatin-treated HeLa cells, Biochim. Biophys. Acta. 563 (1979) 59–71. [40] J.M. Battigello, M. Cui, S. Roshong, B.J. Carter, Enediynemediated cleavage of RNA, Bioorg. Med. Chem. 3 (1995) 839–849.
Cleavage of cellular DNA by calicheamicin ␥1 Kecke Elmroth a , Jonas Nygren b , Susanne Mårtensson c , Ismail Hassan Ismail c , Ola Hammarsten c,∗ a
c
Deptartment of Oncology, Göteborg University, Sahlgrenska University Hospital, SE-413 45 Göteborg, Sweden b Södertörns Högskola and Karolinska Institute, CNT, Novum, SE-141 57 Huddinge, Sweden Department of Clinical Chemistry, Göteborg University, Sahlgrenska University Hospital, 41345 Göteborg, Sweden Received 16 July 2002; received in revised form 14 November 2002; accepted 18 November 2002
Abstract It is assumed that the efficient antitumor activity of calicheamicin ␥1 is mediated by its ability to introduce DNA double-strand breaks in cellular DNA. To test this assumption we have compared calicheamicin ␥1-mediated cleavage of cellular DNA and purified plasmid DNA. Cleavage of purified plasmid DNA was not inhibited by excess tRNA or protein indicating that calicheamicin ␥1 specifically targets DNA. Cleavage of plasmid DNA was not affected by incubation temperature. In contrast, cleavage of cellular DNA was 45-fold less efficient at 0 ◦ C as compared to 37◦ due to poor cell permeability at low temperatures. The ratio of DNA double-strand breaks (DSB) to single-stranded breaks (SSB) in cellular DNA was 1:3, close to the 1:2 ratio observed when calicheamicin ␥1 cleaved purified plasmid DNA. DNA strand breaks introduced by calicheamicin ␥1 were evenly distributed in the cell population as measured by the comet assay. Calicheamicin ␥1-induced DSBs were repaired slowly but completely and resulted in high levels of H2AX phosphorylation and efficient cell cycle arrest. In addition, the DSB-repair deficient cell line Mo59J was hyper sensitive to calicheamicin ␥. The data indicate that DSBs is the crucial damage after calicheamicin ␥1 and that calicheamicin ␥1-induced DSBs are recognized normally. The high DSB:SSB ratio, specificity for DNA and the even damage distribution makes calicheamicin ␥1 a superior drug for studies of the DSB-response and emphasizes its usefulness in treatment of malignant disease. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Calicheamicin ␥1; DNA double-strand break; DNA single-strand break; DNA damage; Ionizing radiation; DNA repair; Bleomycin
1. Introduction
Abbreviations: Cch1, calicheamicin ␥1; DSB, double-strand break; SSB, single-strand break; DTT, dithiothreitol; S.D., standard deviation; S.E.M., standard error of the mean; PBS, phosphate buffered saline; DNA-PK, DNA-dependent protein kinase; BSA, bovine serum albumin ∗ Corresponding author. Tel.: +46-31-3421561; fax: +46-31-828458. E-mail address: [email protected] (O. Hammarsten).
The study of DSB-specific responses is hampered by the fact that classical DNA double-strand break (DSB) inducing agents such as ionizing radiation and bleomycin mostly induce DNA single-stranded breaks (SSBs) and only a few percent of the DNA damage is DSBs [1]. Therefore, the cellular responses to these agents comprise a complex mixture of signals to SSBs, base damage, DSBs and oxidative damage to other molecules in the cell. In addition, there
1568-7864/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 1 5 6 8 - 7 8 6 4 ( 0 2 ) 0 0 2 3 5 - 5
364
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
is now a growing body of evidence that radiation therapy kills cancer cells by induction of DSBs [2] whereas the more frequent SSBs contribute little to the anti-tumor effect. To find ways to improve cancer treatment and DSB-response studies it will be important to identify agents that preferentially introduce DSBs in cells. One candidate drug in this class is the enediyne antibiotic calicheamicin ␥1 [3,4]. Several reports indicate that 50% of the strand breaks produced by calicheamicin ␥1 are DSBs when cleavage is performed on purified DNA in vitro [5,6]. The mechanism behind the efficient DSB induction is that calicheamicin ␥1 contains two radical centers that become positioned close to the backbone when the drug binds the minor groove in double-stranded DNA. Co-operative activation of the two radical centers occurs after reduction of a trisulfide in the molecule, resulting in efficient DSB induction [6–9]. A similar cleavage pattern has been assumed to occur when calicheamicin ␥1 cleaves cellular DNA. It is, however, possible that cellular DNA is cleaved differently from purified DNA in vitro. First, growth medium and serum components could affect the activity of calicheamicin ␥1. Second, passage of calicheamicin ␥1 through cellular membranes and redox-status in the cell could affect distribution and activation of calicheamicin ␥1. DNA associated proteins such as histones could result in differences in cleavage patterns from purified DNA [10,11]. In addition, DSBs induced by calicheamicin ␥1 frequently display abnormal nucleotides at both strands [12,13] that could affect how cells detect and repair these DSBs. It is, therefore, important to carefully define how calicheamicin ␥1 affects cellular DNA to validate its application in the clinic and in DSB-response studies. This information could help in the optimization of calicheamicin ␥1-linked antibody regimens [14] that have proven effective against acute myeloid leukemia [15–18], renal cell carcinoma [19] and ovarian cancer [20]. To study the cleavage of cellular DNA by calicheamicin ␥1 we have treated human primary fibroblasts with calicheamicin ␥1 under different conditions and measured DNA strand breaks and cellular responses. We found that calicheamicin ␥1 induced DSBs with high efficiency in cellular DNA eliciting a strong DSB-response and cell cycle arrest. Our findings indicate that calicheamicin ␥1 is a superior agent
for DSB-response studies and explains the extreme toxicity of this drug.
2. Materials and methods 2.1. Cells and reagents Calicheamicin ␥1 was a generous gift from George Ellestad (Wyeth-Ayers Research). The drug was dissolved at 2 mM in DMSO (Sigma) and stored at −70 ◦ C. Under these conditions, calicheamicin ␥1 showed no loss of activity after 12 months. Wortmannin (Sigma) was dissolved at 1 mg/ml in DMSO and stored at −70 ◦ C. Bleomycin (Astra Medica) was dissolved in distilled water at 15 000 U/ml and stored at −70 ◦ C. The tRNA was dissolved at 10 mg/ml in 10 mM Tris–HCl pH 7.5, 1 mM EDTA (TE-buffer). Herring sperm DNA (Sigma) was dissolved in TE-buffer and sonicated with a probe sonicator on ice at maximal output for 30 min. Agarose gel analysis indicated the medium size of the DNA fragments was 300 bases and no DNA above 2000 bp could be detected. Bovine serum albumin, fraction V (Sigma) was dissolved at 100 mg/ml in TE-buffer. The reduced form of glutathione and dithiothreitol (DTT) (Sigma) was dissolved at 0.5 M and 1 M concentration respectively in TE-buffer and stored at −20 ◦ C. Plasmid DNA (pUC13) was prepared using standard protocols. Primary human fibroblasts, Mo59K cells and Mo59J cells were grown in Dulbecco’s modified Eaagle medium with glutamax (Invitrogen) supplemented with 10% fetal calf serum, streptomycin and penicillin at 37 ◦ C in a humidified cell incubator with 5% CO2 . Cell counting was done using a Bürker chamber. Cellular DNA was labeled 4–5 days prior to each experiment by plating the fibroblasts at 30–50% confluency in growth medium supplemented with 1000 Bq/ml 2-14 C-thymidine (59.0 Ci/mmol, AmershamPharmacia). Confluent monolayers of primary human fibroblasts in 3.3 cm petri dishes were permeabilized after rinsing with PBS (100 mM NaPO4 pH 7.5, 100 mM NaCl) by incubating human primary fibroblasts for 1 h in 1 ml of permeabilizing buffer (142 mM KCl, 20 mM Tris–HCl pH 7.5, 10 mM MgCl2 , 5 mM DTT, 0.25% Triton X-100) supplemented with calicheamicin ␥1 as indicated in Fig. 5. Control human primary fibroblasts that were not permeabilized were incubated
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
in the same buffer without Triton X-100. Monolayers were irradiated on ice in PBS using a Philips RT 100 X-ray machine. An acceleration voltage of 100 kV was used for low dose irradiations (0–6 Gy) and 70 kV was used for high dose delivery (15–45 Gy).
365
by dividing the amount of linear or nicked plasmid DNA with the total amount of plasmid DNA in each lane. 2.4. Measurements of strand breaks in cellular DNA
2.2. Cleavage of cellular DNA Calicheamicin ␥1 was diluted in growth medium at 37 ◦ C and then conditioned to the appropriate temperature (0 or 37 ◦ C) for 1 h. Dilution of calicheamicin ␥1 in medium at 0 ◦ C resulted in poor reproducibility of the level of cellular DNA cleavage. Confluent monolayers of human fibroblasts in 3.3 cm petri dishes labeled with 14 C-thymidine were rinsed once with fresh growth medium and incubated at either 37 ◦ C or on ice for 10 min prior to addition of 1 ml preconditioned calicheamicin ␥1 dilutions. After incubation at the appropriate temperature for 1 h the cells were placed on ice and rinsed three times with PBS supplemented with 0.2 mg/ml sheared herring sperm DNA and then prepared for measurements of total DNA strand breaks or double-stranded DNA breaks within 1 h. Bleomycin (Astra medica, Sweden) was dissolved in water and diluted in growth media to the concentrations indicated and incubated with human primary fibroblasts at 0 ◦ C for 1 h prior to preparation for comet assay. 2.3. Cleavage and analysis of plasmid DNA Calicheamicin ␥1 cleavage of plasmid DNA (0.1 mg/ml) was performed in TE buffer supplemented with 5 mM DTT as indicated in figure legends and various concentrations of calicheamicin ␥1 as indicated in the Figs. 1 and 2. Incubation was for 30 min on ice (0 ◦ C) or 37 ◦ C as indicated. Cleaved plasmid DNA was resolved on 1% agarose gels with 0.5 × TAE (1 × TAE is 40 mM tris acetate, 1 mM EDTA supplemented with 1 mg/ml ethidium bromide) at 16 V/cm and gels were analyzed on a fluorescence scanner (Typhoon 9200, AmershamPharmacia). The relative amount of supercoiled (uncleaved), linear (one DSB per plasmid) and nicked (only SSBs) plasmid DNA was quantified from the images using the Imagequant software (AmershamPharmacia). The relative levels of DSBs (linear plasmid DNA) and SSBs (nicked plasmid DNA) were calculated
The DSB analysis was performed as described previously [21]. Briefly, the monolayer was rinsed with ice-cold PBS supplemented with 0.2 mg/ml sheared herring sperm DNA after exposure to drugs or X-rays and the cells were scraped off the plates with a rubber scrape. The suspension was mixed with agarose and transferred to a plug mold. The cells in the plug were lysed at 37 ◦ C for 48 h, washed and resolved by agarose gel electrophoresis (0.7%) in 1×TAE at 18 ◦ C for 17 h, using constant field gel electrophoresis. The gel was stained with ethidium bromide and cut into lanes and wells. The activity of 14 C in each sample was measured in a liquid scintillation counter and the fraction of activity released (FAR) from the well into the lane was calculated. In the case of rejoining experiment, treated cells were washed two times with ice-cold PBS, pre-warmed medium was added and the monolayers were incubated for different time at 37 ◦ C at 5% CO2 . Total DNA strand breaks were analyzed by using the DNA unwinding assay described by [22] with some modifications [11]. Note that the unwinding technique is also sensitive to DSB although this is normally over-looked due to low SSB:DSB ratios. Briefly, 1 ml 0.03 M NaOH in 0.15 M NaCl was added gently to 175 l of the same cell suspension prepared for DSB analysis. The unwinding procedure was performed on ice, in darkness, for 30 min. After neutralization, single and double-stranded DNA was separated on hydroxyl apatite columns. The activity of 14 C in each sample was measured by liquid scintillation. 2.5. Calculation of the SSB:DSB ratio In order to calculate the SSB:DSB ratio after treatment with calicheamicin ␥1, strand breaks/cell was calculated. The number of DSBs was calculated using the formula of Blöcher [23]. The exclusion size was estimated to 5.7 Mbp and mean chromosomal size was set to 130 Mbp. The number of total DNA strand breaks was calculated using the formula
366
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
Fig. 1. Induction of DNA strand breaks in plasmid DNA by Cch1 in vitro: (A) DNA cleavage by Cch1 in vitro is dependent on reducing agents such as DTT. Plasmid DNA was mixed with Cch1 or DTT and the combination for 30 min at 0 ◦ C; (B) DNA cleavage in vitro by Cch1 is not affected by incubation temperature. Plasmid DNA was mixed with different concentrations of Cch1 in the presence of DTT for 30 min at 0 ◦ C or 37 ◦ C; (C) stability of calicheamicin ␥1 in growth medium. Calicheamicin ␥1 (20 M) was diluted in growth medium with 10% fetal calf serum supplemented or not supplemented with 5 mM DTT as indicated and incubated at 37 ◦ C for different times. DNA cleavage capacity was measured by mixing the pre-incubated calicheamicin ␥1 with plasmid DNA in presence of 5 mM DTT on ice for 30 min. The cleaved plasmid DNA was separated on agarose gels and the amount of undamaged, linear (one DSB per plasmid) and nicked (only SSBs) plasmid DNA was monitored.
k∗ − log(ds/(ds + ss)) = total DNA strand breaks per 109 Da, where ds is the fraction of double-stranded DNA and ss is the fraction of single-stranded DNA eluted from the columns, and k is a conversion factor. The conversion factor k = −31 from [24] and the assumption that a cell contains 3.83 × 1012 Da of DNA was used. The number of SSBs was calculated by subtracting DSBs from total DNA strand breaks. Using the methods of calculation described here on
gamma irradiated samples gave results in good agreement with published data [25]. 2.6. Comet assay The comet assay was performed as described [26] with some modifications [27]. For each slide, 5000–10000 human primary fibroblasts were mixed with 150 l 0.75% low-melting agarose (Sigma type
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
367
slides were then lysed [0.03 M NaOH; 1 M NaCl; 2 mM EDTA and 0.5% NLS] for 1 h and thereafter rinsed [0.03 M NaOH; 2 mM EDTA] for 1 h and finally, electrophoresis was run at 0.8 V/cm for 15 min in the same solution. The slides were then neutralized in 0.4 M Tris–HCl, pH 7.5. All slides were then rinsed in deionized water and air-dried. They were fixed in methanol, stained with 1 g/ml ethidium bromide for 5 min and rinsed briefly in TAE buffer. Analysis of DNA that migrated from the nucleus, the tail moment, was done using an Olympus fluorescence microscope, aided by the Kinetic Imaging Komet II system. 2.7. H2AX phosphorylation (γ-H2AX)
Fig. 2. Calicheamicin ␥1 specifically cleaves DNA: (A) calicheamicin ␥1 specifically targets DNA. Plasmid DNA (0.1 mg/ml) was mixed with sheared herring sperm DNA, tRNA or BSA at the concentrations indicated in the figure before addition of calicheamicin ␥1 (1.35 M) and DTT. The amount of uncleaved, supercoiled plasmid DNA was determined as described in Section 2 and used to calculate the relative level of plasmid DNA cleavage in each sample. Error bars represent data from double samples from a single experiment; (B) calicheamicin ␥1 does not react with protein or RNA. Calicheamicin ␥1 (0.7 M) was preincubated with BSA (1.7 mg/ml), tRNA (1.7 mg/ml), sheared herring sperm DNA (1 mg/ml) or without extra addition (buffer) in TE buffer supplemented with 1 mM DTT. After preincubation for 0, 1, 3, 10 or 30 min at 37 ◦ C, plasmid DNA (0.05 mg/ml) was added and the cleavage reaction was allowed to proceed for 30 min. The amount of cleaved plasmid DNA was determined as in (A).
VII, USA), held at 37 ◦ C. The agarose-cell suspension was spread into a thin layer on clear-glass slides (Menzel Superfrost, Germany), pre-treated with a small amount of agarose, and air-dried. The preparations were left on a chilled plate for 5 min and the
Confluent monolayers of human fibroblasts in 10 cm plates were incubated with 12 or 36 nM calicheamicin ␥1 at 37 ◦ C for 1 h. After washing with cold TBS (20 mM Tris–HCl pH 7.5, 137 mM NaCl) supplemented with 5 mM EDTA, cells were scraped off the plates, and lysed in 1 ml cell lysis buffer (10 mM EDTA, 20 mM Tris–HCl pH 8.0, 1% Nonidet P-40, 10 mM NaF, 1 mM NaVO4 , 1 mM NaMo4 , protease inhibitors cocktail (Roche) and 10 mM Okadaic acid (Alexis)) and nuclear pellet collected by centrifugation. Histones were acid extracted in 3 volumes of extraction buffer (0.5 M HCl + 10% glycerol + 0.1 M mercaptoethanol), precipitated with trichloroacetic acid (25%), separated on 15% SDS-PAGE and transferred to PVDF-membrane (Biorad). The membranes were blocked with blotto (5% non-fat milk, 10 mM Tris–HCl pH 7.5, 140 mM NaCl, 0.1% Nonidet P-40) and incubated with an antibody specific for the phosphorylated form of H2AX (g-H2AX) (Upstate Biotechnology) diluted 1:2000 in blotto. After incubation with secondary antibody (AmershamPharmacia) the membrane was developed with Dura ECL kit (Pierce) and analyzed on a digital camera based system (Chemidoc, Biorad). 2.8. Colony assay Confluent monolayers of Mo59J and Mo59K cells were treated with 0, 30, 100, 300, 1000 and 3000 pM calicheamicin ␥1 in growth medium at 37 ◦ C for 1 h. Cells were rinsed three times with PBS supplemented with 5 mM DTT and 0.2 mg/ml sheared herring sperm DNA before trypsination and replating at various cell
368
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
densities in 10 cm plates. The cells were stained with Giemsa after 14 days and colonies containing over 50 cells were counted.
3. Results 3.1. Clevage of purified plasmid DNA by calicheamicin γ1 To study cleavage of cellular DNA by calicheamicin ␥1 we first wanted to establish conditions resulting in efficient cleavage of purified plasmid DNA. Calicheamicin ␥1 contains a trisulfide that must be reduced to activate the two radical centers in the molecule. Therefore, DNA-cleavage requires a reducing agent such as dithiotreitol (DTT) or glutathione [6]. To examine the dependence on reduction we mixed calicheamicin ␥1 with purified plasmid DNA in the presence or absence of DTT. DNA cleavage was monitored on agarose gels that allow separation of linear (one DSB per plasmid), nicked (only SSB) and supercoiled (undamaged) plasmid DNA. As expected, calicheamicin ␥1 failed to cleave DNA in the absence of DTT whereas efficient cleavage was observed when both calicheamicin ␥1 and DTT were included in the reaction (Fig. 1A). We also found that calicheamicin ␥1 cleavage of plasmid DNA was not affected by incubation temperature (Fig. 1B). Under all cleavage conditions tested, the ratio between nicked (SSB) and linear (DSB) plasmid DNA was 1:1.8 indicating that for each DSB, about two SSBs were produced. Identical cleavage patterns were reproduced when glutathione, the predominant thiol in cells, was used as reductant (data not shown) [28]. Calicheamicin ␥1 showed no loss of DNA-cleavage activity after 24 h preincubation at 37 ◦ C in serum containing growth medium. In contrast, when DTT was included in the preincubation medium, calicheamicin ␥1 was rapidly inactivated (Fig. 1C). This indicated that calicheamicin ␥1 was stable in cell culture media whereas the reduced form of calicheamicin ␥1 was unstable. Calicheamicin ␥1 is inactivated when it cleaves DNA and can therefore only cleave once. Therefore, if calicheamicin ␥1 were to react with proteins or RNA, a significant portion of the calicheamicin ␥1 that enters the cell might be inactivated resulting in damage
to other cellular structures and inefficient DNA cleavage. To test if calicheamicin ␥1 was able to react with proteins or RNA we first mixed plasmid DNA with tRNA or bovine serum albumin (BSA) before addition of calicheamicin ␥1 and DTT. The cleavage reaction was then allowed to proceed to completion and the extent of plasmid cleavage analyzed. As expected, addition of excess low molecular weight DNA prevented calicheamicin ␥1 from cleaving plasmid DNA (Fig. 2A). In contrast, tRNA or BSA had little effect. We also tested if calicheamicin ␥1 was inactivated faster in the presence of tRNA or BSA compared to buffer alone. We therefore preincubated calicheamicin ␥1 for various times in a DTT containing buffer supplemented with tRNA, BSA or low molecular weight DNA. Remaining DNA cleavage capacity was monitored by addition of plasmid DNA at different time points and analysis of cleaved plasmid DNA on agarose gel (Fig. 2B). Calicheamicin ␥1 was inactivated with similar kinetics in buffer alone and in presence of tRNA. Addition of BSA apparently protected calicheamicin ␥1 since inactivation was slower in the presence of this protein. These data show that calicheamicin ␥1 does not react with protein or RNA. 3.2. Cleavage of cellular DNA by calicheamicin γ1 To study calicheamicin ␥1 cleavage of cellular DNA we measured the induction of DSBs in primary human fibroblasts at 37 and 0 ◦ C using the FAR-assay [21]. We also measured the number of total DNA strand breaks (SSB + DSB) produced at 0 ◦ C in the same cells to allow calculation of the relative number of DSBs and SSBs induced at 0 ◦ C [11,22]. As a comparison, the yields of X-ray-induced SSBs and DSBs in the same human primary fibroblasts were measured. The dose-response relationship for both total DNA strand breaks and DSBs at 0 ◦ C was linear, with the yields 32.4 total DNA strand breaks/nM and cell and 8.23 DSB/nM and cell (Fig. 3A). The corresponding R2-values for the linear part of the total DNA strand break curve was 0.96. For DSBs the R2-values were 0.99 for cleavage at 0 ◦ C and 0.98 for cleavage of cellular DNA at 37 ◦ C indicating a good linear correlation between calicheamicin ␥1 concentration and induction of DNA strand-breaks (Fig. 3A and B). For calicheamicin ␥1 the ratio was 2.94, such that for
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
369
each DSB, about 3 SSBs were induced. In contrast, X-rays introduced 20 SSBs for every DSB (data not shown), similar to what have been reported previously [29]. Thus, the relative DSB/SSB ratio for cleavage of cellular DNA was close to the two-fold difference observed when calicheamicin ␥1 cleaved purified plasmid DNA. In contrast to cleavage of purified plasmid DNA by calicheamicin ␥1, cleavage of cellular DNA was strongly influenced by incubation temperature. At 0 ◦ C, calicheamicin ␥1 cleaved cellular DNA 45-fold less effective compared to cleavage at 37 ◦ C. This difference may be due to low cell permeability of calicheamicin ␥1 at low temperatures. In line with this possibility, we found that permeabilization, using the detergent Triton X-100, enhanced induction of DSBs at 0 ◦ C (Fig. 3C). In this experiment the permeabilization buffer was supplemented with 5 mM DTT to allow activation of calicheamicin ␥1. Similar stimulation by permeabilization was obtained with 5 mM glutathione, the predominant thiol in cells (data not shown). 3.3. Even distribution of DNA damage levels in cells treated with calicheamicin γ1
Fig. 3. Cleavage of cellular DNA by calicheamicin ␥1: (A) induction of total DNA strand breaks and double-strand breaks in cells treated with different concentrations of calicheamicin at 0 ◦ C for 1 h. Each point represents the mean from three separate experiments with duplicate samples and the bars represent±S.E.M. when larger than the symbol; (B) induction of double-strand breaks in cells treated with different concentrations of calicheamicin ␥1 at 37 ◦ C for 1 h. Each point represents the mean from two separate experiments with duplicate samples and the bars represent±S.E.M. when larger than the symbol; (C) the number of double-strand breaks after treatment with 0 or 3 nM calicheamicin for 1 h in intact cells or permeabilized cells. One single experiment with duplicate samples. The bars show ± S.D.
To examine if the induction of damage by calicheamicin ␥1 was evenly distributed among the cells in a population, we analyzed DNA strand-break induction on 50 individual human primary fibroblasts using the comet assay. In this assay, DNA strand breaks are monitored by electrophoresis of cells molded into agarose. The amount of DNA that migrates out of the nucleus, the tail moment, correlates with extent of DNA strand breaks in individual cells. As a control, we also analyzed human primary fibroblasts treated with bleomycin that previously have been shown to give an uneven distribution of damaged cells, and X-rays that induces an even damage distribution. As shown in Fig. 4, damage-induced by calicheamicin ␥1 followed a normal distribution similar to that after X-rays. 3.4. Repair and detection of calicheamicin γ1-induced DSBs To study repair of calicheamicin ␥1-induced DSBs, monolayers of human fibroblasts were treated with calicheamicin ␥1 or X-rays at 0 ◦ C and then allowed to rejoin at 37 ◦ C. The level of remaining DSBs was
370
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
ing of calicheamicin ␥1-induced DSB was slower, probably reflecting the more complex DSBs produced by calicheamicin ␥1 [12,13]. Despite the slow rejoining, the level of DSBs was down to background levels after 24 h. This indicates that human primary fibroblasts are fully capable of repairing calicheamicin ␥1-induced DSBs.
Fig. 4. Cumulative frequency of DNA strand breaks, measured by the comet assay and expressed as tail moment in 50 individual cells irradiated with (A) 6 Gy of X-rays (B) 1 nM calicheamicin at 0 ◦ C for 1 h and (C) 135 U bleomycin at 0 ◦ C for 1 h. The level of strand breaks, the tail moment, (black circles) was compared to the expected values, normally distributed for the same mean and S.D. (line).
analyzed at different time points. We found that rejoining of calicheamicin ␥1-induced DSBs followed biphasic kinetics similar to what was found for X-ray-induced DSBs (Fig. 5A). However, the rejoin-
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
To examine the cellular response to calicheamicin ␥1 we measured phosphorylation of H2AX in calicheamicin ␥1 treated cells [30]. The histone 2A variant, H2AX, becomes phosphorylated in response to DSBs, but not by other types of DNA-damage [30,31]. Calicheamicin ␥1 was found to efficiently induce H2AX-phosphorylation in a dose dependent manner (Fig. 5B). In addition, human primary fibroblasts displayed a permanent growth arrest after a 1 h exposure of sub-nanomolar concentrations of calicheamicin ␥1 (Fig. 5C). This indicates that the DSBs produced by calicheamicin ␥1 induce a normal and strong DSB-response. We also tested if cells with defective DSB-repair were hypersensitive to calicheamicin ␥1. The Mo59J cell line is derived from a human glioblastoma tumor and lack expression of DNA-PKcs [32,33] and is therefore sensitive to ionizing radiation and defective in DSB-repair. The matching Mo59K cell line was isolated from the same tumor but have high DNA-PKcs expression and display normal sensitivity to ionizing radiation and DSB-repair [34,35]. Mo59J and Mo59K cells were exposed to different concentrations of calicheamicin ␥1 and surviving fraction was determined by colony forming assay. The result showed that Mo59J cells were 4-fold more sensitive to calicheamicin ␥1 compared to Mo59K cells (Fig. 6) supporting the notion that DSB is the biologically important lesion induced by calicheamicin ␥1.
䉳 Fig. 5. Cellular responses to calicheamicin ␥1: (A) rejoining of double-strand breaks, expressed as percentage of initial damage, in cells treated with 100 nM calicheamicin at 0 ◦ C for 1 h and then allowed to rejoin at 37 ◦ C for the indicated times (filled circles). Each point represents the mean from three separate experiments and the bars represent ± S.E.M. The rejoining after X-irradiation of the same cell line with 45 Gy is shown for reference (empty circles); (B) phosphorylation of H2AX (␥-H2AX) in response to calicheamicin ␥1. Cells were incubated with 12 or 45 nM calicheamicin ␥1 for 1 h and H2AX phosphorylation (␥-H2AX) examined with immunoblot (upper image). The same samples were separated on a separate SDS-PGE and stained with Coomassie to control for equal loading of extracted histones (lower image); (C) cell cycle arrest after calicheamicin ␥1 treatment. Confluent monolayers (95% ␥1 cells) human primary fibroblasts were treated with 0, 40 or 400 pM calicheamicin ␥1 for 1 h at 37 ◦ C before trypsination and replating at 1:50 dilution. Cells were trypsinized and counted using a Bürker chamber at the indicated times.
371
Fig. 6. Cells defective in DSB-repair are hypersensitive to calicheamicin ␥1. Mo59K (normal DSB-repair) and Mo59J (defective DSB-repair) cells were incubated with 0, 30, 100, 300, 1000 or 3000 pM calicheamicin ␥1 at 37 ◦ C for 1 h. The fraction of cells surviving this treatment (surviving fraction) was determined with the colony assay described in Section 2.
4. Discussion The data presented in this study indicate that calicheamicin ␥1 appears to do little else than DSBs in cellular DNA. First, more than 30% of the calicheamicin ␥1-induced strand breaks in cellular DNA were DSBs. Second, calicheamicin ␥1 is likely to specifically target DNA when it enters the cell, since tRNA or protein failed to interfere with plasmid DNA cleavage. In addition, calicheamicin ␥1 induced a strong DSB-response in human fibroblasts at concentrations that would deliver less than 1000 molecules/cell. Finally, calicheamicin ␥1 is unlikely to react with extracellular structures since the drug was stable under normal cell culture conditions. It is therefore possible that calicheamicin ␥1 will be a superior DSB-inducing agent in DSB-response studies and DSB-based cancer therapies. Previous studies have shown that when calicheamicin ␥1 cleaves purified DNA, 50% of DNA strand breaks are DSBs [5]. To measure the extent of DSB and SSB in cellular DNA we separately measured DSB and total DNA strand breaks (DSB + SSB) in calicheamicin ␥1-treated cells. To calculate the DSB/ SSB ratio it was necessary to recalculate the primary
372
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
signal from the methods to the number of DNA strand breaks per cell. The formula of Blöcher [23] was used to calculate the number of DBS/cell. This formula utilizes the fraction of DNA smaller than a certain cut off size and the assumption that damage is induced in a random fashion. It is not possible to use size markers in constant field gel electrophoresis to determine the cut off size. The fraction radioactivity released (FAR) for different doses of X-ray irradiation was therefore compared with published data for the fraction of DNA smaller than certain size markers as determined by pulsed field gel electrophoresis [36,37]. It was estimated that the cut off size for the electrophoresis conditions used here was about 5.7 Mbp. The yield of DSB/Gy calculated in this way from our data was in good agreement with published data [38]. Based on these calculations we found that the DSB/SSB ratio in cellular DNA was 1:3, close to the 1:2 ratio found when calicheamicin ␥1 cleaved purified plasmid DNA. A similar DSB/SSB ratio (1:5) was found after cleavage of cellular DNA with the enediyne antibiotic neocarzinostatin [39]. However, in this study the cells were treated with neocarzinostatin at 37 ◦ C for 1 h before measurements of DSBs and SSBs. A potential problem with this protocol is that repair of SSBs is much faster compared to repair of DSBs. Therefore, the treatment with neocarzionostatin under repair proficient conditions might underestimate the number of SSB due to quick repair of this lesion during treatment. This might result in a significantly increased DSB/SSB ratio. In contrast, we have calculated the DSB/SSB ratio for calicheamicin ␥1 cleavage of cellular DNA on ice, where no repair occurs. Therefore, to be able to compare the DSB/SSB ratio for calicheamicin ␥1 and neocarzinostatin, the cleavage of cellular DNA must be done using the same protocol. One important finding was that tRNA and protein essentially failed to interfere with cleavage of purified plasmid DNA. Since calicheamicin ␥1 is inactivated after cleavage, this finding indicates that the drug does not react with RNA or protein. One previous study has found that calicheamicin ␥1 binds double-stranded RNA and is able to cleave RNA-oligonucleotides [40]. However, in the light of our results calicheamicin ␥1 is expected to target cellular DNA without interference from proteins or RNA. This selectivity decreases the extent of damage to other cellular
structures and increases the DNA-cleaving efficiency of calicheamicin ␥1. We also found that calicheamicin ␥1 produced a strong DSB-response. First, sub-nanomolar concentrations of calicheamicin ␥1 induced a permanent cell cycle arrest in primary human fibroblasts. The number of DSBs induced at this concentration, calculated from the dose-response curve, equals that of an X-ray dose resulting in similar growth arrest (3–4 Gy). Therefore, our data indicate that the toxicity of calicheamicin-induced DSBs were equal to that of X-ray-induced DSBs although our data do not validate an exact comparison. Second, calicheamicin ␥1 efficiently induced H2AX phosphorylation. H2AX is phosphorylated specifically in response to DSBs [30]. In addition, our result show that human fibroblasts were fully capable of repairing calicheamicin ␥1-induced DSBs. In line with this finding, we have previously shown that DNA-PK, a DNA activated protein kinase involved in DSB-repair, is efficiently activated by calicheamicin-induced DSBs in vitro (submitted). We also found that Mo59J cells, that are defective in DSB-repair, were hypersensitive to calicheamicin ␥1. Therefore, calicheamicin ␥1induced DSBs were recognized normally. During the course of this study we identified several parameters important for reproducible cleavage of cellular DNA by calicheamicin ␥1. Induction of DSBs in cellular DNA was 45-fold less efficient at 0 ◦ C as compared to 37 ◦ C. It is therefore important that the temperature is controlled. The poor cleavage of cellular DNA at 0 ◦ C was not due to alteration in calicheamicin ␥1 interaction with DNA, since cleavage of purified plasmid DNA was not affected by incubation temperature. However, permeabilization of the cells stimulated cleavage of cellular DNA more than one-order of magnitude at 0 ◦ C. This indicates that passage of calicheamicin ␥1 through cellular membranes is inefficient at low temperatures. Hyperthermia has been used to enhance the effects of ionizing radiation. If the temperature effect reported in this paper were linear, a 16% increase in initial damage would be obtained at 43 ◦ C, the temperature used in hyperthermia treatment regimens. It is therefore possible that localized hyperthemia could increase the selective killing of cancer cells by calicheamicin ␥1. We also found that when cells were treated with calicheamicin ␥1 at 0 ◦ C, significant cleavage of cellular DNA
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
occurred after the calicheamicin ␥1 solution was carefully washed away from the cells. This effect was alleviated by addition of sheared herring-sperm DNA in the cold PBS used to wash the cells after treatment with calicheamicin ␥1 (data not shown). We speculate that a microprecipitate of calicheamicin ␥1 was deposited on cellular membranes specifically at low temperatures that cannot be washed away completely unless DNA is included in the washing solution. Calicheamicin ␥1 was stable in normal growth medium for at least 24 h but quickly inactivated if thiols were present. Addition of DTT to the medium therefore resulted in inefficient cleavage of cellular DNA (data not shown). Importantly, addition of 5 mM DTT is an efficient way to detoxify calicheamicin ␥1 containing solutions. In summary, our data indicate that the antitumor activity of calicheamicin ␥1 is mediated by its efficient DSB-inducing activity in cells. Therefore, when used properly, calicheamicin ␥1 holds a great promise in development of cancer therapies and will allow for a more selective study of the DSB-response in human cells.
Acknowledgements We thank Henrik Zetterberg for careful reading of the manuscript. KE was supported by the King Gustav V Jubilee Clinic Cancer Research Foundation. JN was supported by Swedish Radiation Protection Agency. OH was supported by Swedish Cancer Society, the Swedish Medical Research Council, King Gustav V Jubilee Clinic Cancer Research Foundation and Sahlgren’s Hospital Research Foundation. References [1] A.H.W. Nias, Subcellular radiobiology, in: An Introduction to Radiobiology, Wiley, New York, 1998, pp. 63–79. [2] O. Tounekti, A. Kenani, N. Foray, S. Orlowski, L.M. Mir, The ratio of single- to double-strand DNA breaks and their absolute values determine cell death pathway, Br. J. Cancer 84 (2000) 1272–1279. [3] M.D. Lee, T.S. Dunne, M.M. Siegel, C.C. Chang, G.O. Morton, D.B. Borders, Calicheamicins, a novel family of antitumor antibiotics, 1 chemistry and parial strucure of calicheamicin ␥1, J. Am. Chem. Soc. 109 (1987) 3464–3466.
373
[4] M.D. Lee, G.A. Ellestad, D.B. Borders, Calicheamicins: discovery, structure, chemistry and interaction with DNA, Acc. Chem. Res. 24 (1991) 235–243. [5] P.C. Dedon, A.A. Salzberg, J. Xu, Exclusive production of bistranded DNA damage by calicheamicin, Biochemistry 13 (1993) 3617–3622. [6] N. Zein, A. Sinha, W. McGahren, G. Ellestad, Calicheamicin ␥: an antitumor antibiotic that cleaves double-stranded DNA site specifically, Science 240 (1988) 1198–1201. [7] N. Zein, M. Poncin, R. Nilakantan, G.A. Ellestad, Calicheamicin ␥1 I and DNA: molecular recognition process responsible for site-specificity, Science 244 (1989) 697–699. [8] Z. Xi, I.H. Golberg, DNA-damaging enediyne compounds, in: D. Barton, K. Nakanishi (Eds.), Comprehensive Natural Products Chemistry, Pergamon Press, Oxford, 1999. [9] D.M. Lopez-Larraza, K. Moore, P.C. Dedon, Thiols alter the partitioning of calicheamicin-induced deoxyribose 4/coxidation reactions in the absence of DNA radical repair, Chem. Res. Toxicol. 14 (2001) 528–535. [10] L. Yu, A.A. Salzberg, P.C. Dedon, New insights into calicheamicin-DNA interactions derived from a model nucleosome system, Bioorg. Med. Chem. 3 (1995) 729–741. [11] J. Nygren, M. Ljungman, G. Ahnstrom, Chromatin structure and radiation-induced DNA strand breaks in human cells: soluble scavengers and DNA-bound proteins offer a better protection against single-than double-strand breaks, Int. J. Radiat. Biol. 68 (1995) 11–18. [12] M.A. Chaudhry, P.C. Dedon, D.M. Wilson III, B. Demple, M. Weinfeld, Removal by human apurinic/apyrimidinic endonuclease 1 (Ape 1) and Escherichia coli exonuclease III of 39-phosphoglycolates from DNA treated with neocarzinostatin, calicheamicin, and ␥-radiation, Biochem. Pharmacol. 57 (1999) 531–538. [13] J.J. Hangeland, J.J. De Voss, J.A. Heath, C.A. Townsend, W.D. Ding, J.S. Ashcroft, G.A. Ellestad, Specific abstraction of the 59(S) and 49-deoxyribosyl hydrogen atoms from DNA by calicheamicin ␥1 I, J. Am. Chem. Soc. 114 (1992) 9200– 9202. [14] P.R. Hamann, L.M. Hinman, I. Hollander, C.F. Beyer, D. Lindh, R. Holcomb, W. Hallett, H.R. Tsou, J. Upeslacis, D. Shochat, A. Mountain, D.A. Flowers, I. Bernstein, Gemtuzumab ozogamicin, a potent and selective anti-CD33 antibody-calicheamicin conjugate for treatment of acute myeloid leukemia, Bioconjug. Chem. 13 (2002) 47–58. [15] P.F. Bross, J. Beitz, G. Chen, X.H. Chen, E. Duffy, L. Kieffer, S. Roy, R. Sridhara, A. Rahman, G. Williams, R. Pazdur, Approval summary: gemtuzumab ozogamicin in relapsed acute myeloid leukemia, Clin. Cancer Res. 6 (2001) 1490– 1496. [16] R. Larson, Current use and future development of gemtuzumab ozogamicin, Semin. Hematol. 38 (2001) 24–31. [17] M.L. Linenberger, T. Hong, D. Flowers, E.L. Sievers, T.A. Gooley, J.M. Bennett, M.S. Berger, L.H. Leopold, F.R. Appelbaum, I.D. Bernstein, Multidrug-resistance phenotype and clinical responses to gemtuzumab ozogamicin, Blood 98 (2001) 988–994.
374
K. Elmroth et al. / DNA Repair 2 (2003) 363–374
[18] I.D. Bernstein, CD33 as a target for selective ablation of acute myeloid leukemia, Clin. Lymphoma 1 (Suppl.) (2002) S9–S11. [19] K. Knoll, W. Wrasidlo, J.E. Scherberich, G. Gaedicke, P. Fischer, Targeted therapy of experimental renal cell carcinoma with a novel conjugate of monoclonal antibody 138H11 and calicheamicin thetaI1, Cancer Res. 60 (2000) 6089–6094. [20] A.M. Gillespie, T.J. Broadhead, S.Y. Chan, J. Owen, A.P. Farnsworth, M. Sopwith, R.E. Coleman, Phase I open study of the effects of ascending doses of the cytotoxic immunoconjugate CMB-401 (hCTMO1-calicheamicin) in patients with epithelial ovarian cancer, Ann. Oncol. 11 (2001) 735– 741. [21] K. Elmroth, J. Nygren, L.J. Erkell, R. Hultborn, Radiationinduced double-strand breaks in mammalian DNA: influence of temperature and DMSO, Int. J. Radiat. Biol. 76 (2000) 1501–1508. [22] G. Ahnström, K. Erixon, Measurement of strand breaks by alkaline denaturation and hydroxyapatite chromatography, in: E.C. Friedberg, P.C. Hanawalt (Eds.), DNA Repair, Marcel Dekker, New York, 1980, pp. 403–418. [23] D. Blöcher, In CHEF electrophoresis a linear induction of dsb corresponds to a non-linear fraction of extracted DNA with dose, Int. J. Radiat. Biol. 57 (1990) 7–12. [24] M. Ljungman, The influence of chromatin structure on the frequency of radiation-induced DNA strand breaks: a study using nuclear and nucleoid monolayers, Radiat. Res. 126 (1991) 58–64. [25] J.F. Ward, DNA damage produced by ionizing radiation in mammalian cells: identities, mechanisms of formation, and reparability, Prog. Nucl. Acid Res. Mol. Biol. 35 (1988) 95– 125. [26] P.L. Olive, J.P. Banath, R.E. Durand, Heterogeneity in radiation-induced DNA damage and repair in tumour and normal cells measured using the comet assay, Radiat. Res. 122 (1990) 69–72. [27] Y. Yang, W.J. Griffiths, M. Nordling, J. Nygren, L. Möller, J. Bergman, E. Liepinsh, G. Otting, J.-Å. Gustafsson, J. Rafter, J. Sjövall, A major metabolic product of benzo[a]pyrene in the germfree rat is produced by ring opening, Biochemistry 39 (2000) 15585–15591. [28] P.C. Dedon, I.H. Goldberg, Influence of thiol structure on neocarzinostatin activation and expression of DNA damage, Biochemistry 31 (1992) 1909–1917.
[29] J.F. Ward, DNA damage produced by ionizing radiation in mammalian cells: identities, mechanisms of formation, and reparability, Prog. Nucl. Acid Res. Mol. 35 (1988) 95–125. [30] E.P. Rogakou, D.R. Pilch, A.H. Orr, V.S. Ivanova, W.M. Bonner, DNA double-stranded breaks induce histone H2AX phosphorylation on serine 139, J. Biol. Chem. 273 (1998) 5858–5868. [31] T.T. Paull, E.P. Rogakou, V. Yamazaki, C.U. Kirchgessner, M. Gellert, W.M. Bonner, A critical role for histone H2AX in recruitment of repair factors to nuclear foci after DNA damage, Curr. Biol. 10 (2000) 886–895. [32] C.W. Anderson, J.J. Dunn, P.I. Freimuth, A.M. Galloway, M.J. Allalunis-Turner, Frameshift mutation in PRKDC, the gene for DNA-PKcs, in the DNA repair-defective, human, glioma-derived cell line M059J, Radiat. Res. 156 (2001) 2–9. [33] B.S. Hoppe, R.B. Jensen, C.U. Kirchgessner, Complementation of the radiosensitive M059J cell line, Radiat. Res. 153 (2000) 125–130. [34] S. Lees-Miller, R. Godbout, D. Chan, M. Weinfeld, R. Day, G. Barron, J. Allalunis-Turner, Absence of p350 subunit of DNA-activated protein kinase from a radiosensitive human cell line, Science 267 (1995) 1183–1185. [35] S.J. DiBiase, Z.C. Zeng, R. Chen, T. Hyslop, W.J.J. Curran, G. Iliakis, DNA-dependent protein kinase stimulates an independently active, non-homologous, end-joining apparatus, Cancer Res. 60 (2000) 1245–1253. [36] R.S. Wong, J.R. Dynlacht, B. Cedervall, W.C. Dewey, Analysis by pulsed-field gel electrophoresis of DNA doublestrand breaks induced by heat and/or X-irradiation in bulk and replicating DNA of CHO cells, Int. J. Radiat. Biol. 68 (1995) 141–152. [37] K. Erixon, B. Cedervall, Linear induction of DNA doublestrand breakage with X-ray dose, as determined from DNA fragment size distribution, Radiat. Res. 142 (1995) 153–162. [38] E. Hoglund, E. Blomquist, J. Carlsson, B. Stenerlow, DNA damage induced by radiation of different linear energy transfer: initial fragmentation, Int. J. Radiat. Biol. 76 (2000) 539–547. [39] T. Hatayama, I.H. Goldberg, DNA damage and repair in relation to cell killing in neocarzinostatin-treated HeLa cells, Biochim. Biophys. Acta. 563 (1979) 59–71. [40] J.M. Battigello, M. Cui, S. Roshong, B.J. Carter, Enediynemediated cleavage of RNA, Bioorg. Med. Chem. 3 (1995) 839–849.