Hydrogen Peroxide Activated Quinone Methide Precursors with Enhanced DNA Cross-Linking Capability and Cytotoxicity towards Cancer Cells

Abstract

Quinone methide (QM) formation induced by endogenously generated H₂O₂ is attractive for biological and biomedical applications. To overcome current limitations due to low biological activity of H₂O₂-activated QM precursors, we are introducing herein several new arylboronates with electron donating substituents at different positions of benzene ring and/or different neutral leaving groups. The reaction rate of the arylboronate esters with H₂O₂ and subsequent bisquinone methides formation and DNA cross-linking was accelerated with the application of Br as a leaving group instead of acetoxy groups. Additionally, a donating group placed meta to the nascent exo-methylene group of the quinone methide greatly improves H₂O₂-induced DNA interstrand cross-link formation as well as enhances the cellular activity. Multiple donating groups decrease the stability and DNA cross-linking capability, which lead to low cellular activity. A cell-based screen demonstrated that compounds 2a and 5a with a OMe or OH group dramatically inhibited the growth of various tissue-derived cancer cells while normal cells were less affected. Induction of H2AX phosphorylation by these compounds in CLL lymphocytes provide evidence for a correlation between cell death and DNA damage. The compounds presented herein showed potent anticancer activities and selectivity, which represent a novel scaffold for anticancer drug development.

INTRODUCTION

Exploiting cellular processes to tissue-specifically generate pharmacological-active agents has found widespread applications in medicinal chemistry and biological science.[1–3] Generation of DNA alkylating agents that take advantage of the unique processes that occur in tumors can be exploited for development of new anticancer drugs and enzyme inhibitors.[1, 4–8] Tumor cells produce high levels of reactive oxygen species (ROS), which makes them distinctly different from normal cells.[9–13] Our group recently introduced a novel prodrug strategy involving H₂O₂-induced DNA cross-linking with a nitrogen mustard cytotoxin for selective destruction of tumor cells.[4, 14–16] This strategy was applied to quinone methide-based prodrugs that can be activated by H₂O₂ to release active quinone methide (QM) capable of cross-linking DNA.[1, 17–19]

Quinone methides are naturally occurring and highly reactive electrophiles with distinct biological activity.[1, 20, 21] Quinone methides occur during several biological processes such as lignification in trees, enzyme inhibition, reactions with phosphodiesters, DNA alkylation, and DNA cross-linking. They are the ultimate cytotoxins and can be found among many antitumor drugs, DNA alkylators, insecticides, and antibiotics. Some biologically inactive agents can be converted to reactive QMs via cellular processes such as enzymatic oxidation, reduction, or high level of ROS.[1, 21, 22] A variety of chemical methods have also been developed to generate QMs in situ from various precursors,[1, 20] such as photoirradiation,[23–28] thermal extrusion reactions,[1, 29] acid- or base-catalyzed reactions,[1, 30] NaIO₄ induction,[31, 32] and fluoride-[33–35] or H₂O₂-mediated reactions.[17–19] Among these methods, photogeneration and H₂O₂ induction are particular important for biological applications. Photo induction is a non-invasive method with high spatio-temporal resolution and control and offers the options of orthogonality. The research groups of Freccero and Zhou reported photo-inducible QM formation from various precursors, including biphenyl or binol quaternary ammonium salts and 2-alkynylphenols.[24, 26, 36–38] Many of these QM precursors showed potent photocytotoxicity in various cell lines.[26, 36, 37] H₂O₂-induced QM generation can occur under physiological conditions and is attractive for in vivo applications as H₂O₂ is endogenously generated.[1] Our group have developed several arylboronate analogues that can be activated by H₂O₂ to form QMs and directly alkylate DNA.[18, 19] Further investigations showed that different leaving groups and aromatic substituents strongly affect QM formation and DNA alkylation.[17, 19]

However, most existing arylboronate QM precursors did not show activity towards cancer cells.[17–19] Recently, we observed that positively charged arylboronates showed lower cellular activity than the corresponding neutral molecules, which is due to insufficient cell membrane permeability.[14, 16] In this work, we observed that compound 1a with Br as a leaving group was a potential H₂O₂-activated anticancer prodrug (Scheme 1). Thus, we modified 1a as a lead compound to develop more effective prodrugs by ring substitution and leaving group variation that has been reported to modulate their reaction rates with H₂O₂ and subsequent QM formation.[17, 19, 23, 39] In addition, it has been reported that electron rich aryl organic groups are less likely to be mutagenic, which we expect to be suitable trigger units for developing non-toxic prodrugs[40]. Thus, we modified the boron-containing ring by introducing an electron-donating group in the ortho or para position or both (2a–4a), and employing different neutral leaving groups (1a,b, 2a,b and 5a,b) (Scheme 1). We further investigated the reactivity and selectivity of these compounds toward H₂O₂ and DNA and evaluated their cytotoxicity in different tissue-derived cancer cells. The effects of different substituents and benzylic leaving group on their reactivity and cytotoxicity have been compared.

Scheme 1.

Structures of the designed compounds 1–5

RESULTS AND DISCUSSION

Based on our previous work,[41] compounds 2a–5a with Br as a leaving group were synthesized via borylation of the corresponding 1-bromobenzene 6, 10, 13, or 16 using n-butyllithium and isopropoxyboronic acid pinacol ester followed by bromination with N-bromosuccinimide (NBS) and azobisisobutyronitrile (AIBN) (Note: To be consistent, the numbering shown for 1a and 1b is used for all compounds described in this paper). Compound 6 is commercially available, while 10, 13, and 16 were synthesized starting from methoxy substituted benzenes 8, 12, and 15. As previously described,[42] compound 8 was converted to 2,5-dibromo-4,6-dimethoxy-1,3-dimethylbenzene (9), which was then debrominated using n-butyllithium yielding 10 (Scheme 2B). Similarly, compound 13 was synthesized from trimethoxybenzene 12 (Scheme 2C).[42] Protection of hydroxyl group of 15 using TBDMSCl yielded 16 (Scheme 2D).

Scheme 2.

Synthesis of 2a (A), 3a (B), 4a (C), and 5a (D)

We changed the synthetic route for 1b and 2b, because any attempts to introduce the acetate functionality at last step were not successful (Scheme 3). The boronate ester was introduced via palladium-catalyzed borylation from 2-bromobenzene analogue 21 or 24. Compound 21 was prepared from 2-bromoisophthalaldehyde (19) via reduction followed by acylation (Scheme 3A), while 24 was synthesized from 2-bromo-5-methoxy-1,3-dimethylbenzene (22) via bromination followed by nucleophilic substitution using NaOAc (Scheme 3B). Compound 5b was obtained from 5a by replacing Br with OAc using NaOAc in DMF (Scheme 3C).

Scheme 3.

Synthesis of 1b (A), 2b (B), and 5b (C)

The primary mechanism for the cytotoxicity of QM prodrugs is formation of DNA interstrand cross-links (ICLs).[1, 20, 21] ICLs are deleterious to cells, because they inhibit transcription and replication.[2, 43] Therefore, the cytotoxicity and DNA sequence selectivity of the proposed prodrugs can be predicted by their ability to induce DNA cross-linking or DNA alkylation in the presence or absence of H₂O₂. In this work, we characterized compounds 1–5 with respect to H₂O₂ sensitivity, the ability to induce DNA ICLs formation, and DNA alkylation.

The reactivity of 1–5 towards DNA was investigated by reacting them with a 49-mer DNA duplex 25 in pH 8 phosphate buffer at 37 °C for 24 h. Initially, 2 mM compounds and 2 mM H₂O₂ were used for comparison of the DNA cross-linking abilities of 1–5. ICL formation and yields were analyzed via denaturing polyacrylamide gel electrophoresis (PAGE) with phosphor image analysis (Image Quant 5.2). Among all compounds tested, 1b, 2a, 2b, 5a, and 5b did not induce DNA cross-linking in the absence of H₂O₂, indicating that they are not DNA alkylators (Figure 1, lanes 4 and 7–10). In the presence of H₂O₂, efficient ICL formation was observed (20–50%; Figure 1, lanes 12–18), while the presence of H₂O₂ alone did not lead to DNA cross-linking (Figure 1, lane 2). These data clearly demonstrate that these inactive prodrugs can be efficiently activated by H₂O₂ and act as DNA cross-linking agents. Also, compared with their parent compound 1a (22 ± 5%), 2a (36 ± 4%) and 5a (50 ± 5%) greatly improved the ICL yield.

Figure 1.

H₂O₂-induced DNA ICL formation by compounds 1–5. Lane 1: DNA only (cross-linking yield 0%); lane 2: DNA with 100 µM H₂O₂ (0%); lane 3–10 without H₂O₂: lane 3: 2 mM 1a (0%); lane 4: 2 mM 2a (0%); lane 5: 2 mM 3a (17 ± 3%); lane 6: 2 mM 4a (0%); lane 7: 2 mM 5a (0%); lane 8: 2 mM 1b (0%); lane 9: 2 mM 2b (0%); lane 10: 2 mM 5b (0%); lane 11–18 with H₂O₂: lane 11: 2 mM 1a (22 ± 5%); lane 12: 2 mM 2a (36 ± 4%); lane 13: 2 mM 3a (20 ± 2%); lane 14: 2 mM 4a (0%); lane 15: 2 mM 5a (50 ± 5%); lane 16: 2 mM 1b (20 ± 3%); lane 17: 2 mM 2b (26 ± 3%); lane 18: 2 mM 5b (32 ± 4%); [H₂O₂] = 2 mM (Reaction mixture was incubated at 37°C for 48 h).

In order to obtain more detailed information about the substituent effects on the DNA cross-linking, we took an in-depth look at the stability and reactivity of 2a–4a. Comparison of the ICL yields showed that 2a (36%) is superior to 3a (20%) and 4a (0%) as H₂O₂-activated DNA cross-linking agents. In addition, 3a was not selective and induced ICL formation even without H₂O₂ (17%), while the presence of H₂O₂ only slightly increased ICL yield of 3a (20%) (Figure 1, lane 5 and SI, Figure S4, lane 2). This might be due to the presence of two methoxy groups greatly enhancing the electrophilicity of this molecule and decreasing its stability. We observed that 3a was not stable and easily decomposed even at 0 °C. We also observed decomposition of 3a during ¹³C NMR measurement though it was stable during the short ¹H NMR measurement. Although 3,4,5-trimethoxy analogue 4a is more stable than 3a, DNA ICL products were not formed with 4a at pH 7 or higher pH (Figure 1, lane 14 and SI, Figure S7, lanes 7–9). We did, however, observe that the ICL formation induced by 4a strongly depended on the pH. Acidic conditions resulted in higher ICL yields than neutral and basic conditions (SI, Figure S7). Rokita and co-workers reported that the methoxy group at the positions-3, -4, or -5 destabilized the nucleoside-QM adducts and favored regeneration of QM that was quenched by water.[23] We propose that the ICL products formed from 4a may have decreased stability under neutral and basic conditions in comparison to acidic conditions because the presence of three methoxy groups greatly facilitate regeneration of QMs that are finally quenched by water. In order to test our hypothesis, we determined the stability of the ICL products induced by 4a under acidic and basic conditions. The DNA ICLs were generated at pH 6 with duplex 25 and 4a, then incubated at pH 6 or pH 8 at 37 °C for 24 h. We did observe that the ICL yields decreased from 16.6 ± 1.0% to 3.0 ± 0.5% at pH 8 while remained unchanged at pH 6 (Figure S7, lanes 19 and 20). These results provide evidence for that a lower ICL yield under basic condition than acidic condition is caused by the decreased stability of the ICL products formed with 4a as a result of the high reversibility of the QM-nucleoside adducts.

Among the compounds tested, 4-methoxy analogue 2a is the best H₂O₂-activated DNA cross-linking agent that is chemically stable and does not react with DNA without H₂O₂ but efficiently induced DNA ICL formation in the presence of H₂O₂. It is highly likely that the meta position in respect to the QM methylene (position-4) is an ideal site for further modification. In order to see the generality of this phenomenon, we designed and synthesized 5a with a hydroxyl group at the position-4 and investigated its reactivity toward DNA. As expected, 5a is relatively stable and inert towards DNA, but can react with H₂O₂ to form QMs that directly cross-link DNA (50 ± 5%). This provided another evidence that the electron-donating group on the position-4 favors the ICL formation. Further study showed that the cross-linking yield of 2a and 5a were affected by their concentrations, the compound/H₂O₂ ratio, and the pH of the buffer solution. The best compound:H₂O₂ ratio was 1:1 (SI, Figure S1 and S8) and the cross-linking yield increased with increasing compound concentrations (Figure SI, Figure S2 and S9). The cross-linking was more efficient under basic conditions than acidic conditions (SI, Figure S3 and S10).

In order to fully investigate the effect of the aromatic substituents on DNA ICL formation, we studied the kinetics of DNA cross-linking and determined the reaction rate of these compounds with H₂O₂ (Table 1). The ICL growth induced by 2a and 5a followed first-order kinetics, which was similar to the parent compound 1a (SI Figure S13 A and B). The rate constants for 2a and 5a with an electron donating group were 2–3 times the rate constants of 1a (Table 1). This result indicated that the electron-donating substituent facilitated DNA cross-linking as well as enhanced the cross-linking yield.

Table 1.

Kinetics of ICL formation and monomer reaction

Compds	Disappearance of starting materialsa		QM formationa		Kinetics of ICL formationb
	Time of completion (min)	kobs (10−5 s−1)	Time of completion (min)	kobs (10−5 s−1)	Time of completion (min	kobs (10−5 s−1)	ICL (%)
1a	60	39.0 ± 1.5	60	9.5 ± 0.2	600	8.8 ± 1.3	22 ± 3
2a	42	56.7 ± 4.3	50	38.3 ± 2.2	240	23.7 ± 1.4	36 ± 4
5a	40	61.7 ± 3.2	40	58.3 ± 4.3	300	20.1 ± 1.5	50 ± 5
1b	240	19.4 ± 1.2	240	n.d.c	90d	n.d.c	20 ± 3
2b	480	9.1 ± 0.6	480	n.d.c	90d	n.d.c	26 ± 3
5b	480	8.3 ± 0.5	480	n.d.c	90d	n.d.c	32 ± 4

^aThe rate constants for monomer reactions was determined by NMR analysis in a mixture of DMSO and a deuterated phosphate buffer pH 8 (3:2) (SI Figures S14–S19).

^bThe rate constants for ICL formation was determined in a phosphate buffer pH 8, where the cross-linking reaction was performed.

^cn.d.: The rate constants were not determined due to the complexity of the reactions.

^dDue to quick hydrolysis of the acetate groups at pH 8 buffer, the observed reaction time for DNA cross-linking reaction was much shorter than that of the monomer reactions.

Previous studies showed that the mechanism of DNA cross-linking by the arylboronates involved generation of the phenol intermediates followed by spontaneous release of QMs capable of cross-linking DNA.[17–19] We proposed a similar mechanism for 2a and 5a (Scheme 4A). QM-trapping experiment with a large excess of ethyl vinyl ether (EVE) as nucleophile confirmed the formation of highly active electrophile QMs which directly cross-link DNA (Scheme 4B,C) (Note: The trapping product 27 was not stable, in particular after purification and concentration. Decomposition and color change were observed during NMR measurement, but we did detect the fragment 28 in mass spectrometry (Scheme 4C and SI, Figure S63). In order to investigate the effect of electron-donating groups on both formation of the phenolic intermediates and QM generation, we determined the reaction rate of these compounds with H₂O₂ by NMR analysis (SI, Figure S14 and S18). A mixture of phosphate buffer (pH 8 in D₂O) and DMSO in a 2:3 ratio was used to ensure good solubility of these compounds and to mimic DNA cross-linking conditions. As hydrolysis of boronate esters to the corresponding boronic acids made analysis more complex, all compounds were first hydrolyzed to the corresponding boronic acids in DMSO/pH8 buffer (10:1) prior to the addition of H₂O₂ as previously described.[17] The rate for the phenolic intermediate formation (oxidation) was estimated from the disappearance of the ¹H-NMR peaks at about 5.0 ppm (peak d) corresponding to -CH₂- of the precursors. The reaction rates of these compounds with H₂O₂ are in the order of 5a≥2a>1a (Table 1). The rates of QM formation were estimated by the generation of the final products (peak f, its hydrolyzed compounds), which showed a similar trend: 5a≥2a>1a (Table 1). Both methoxy group and hydroxyl groups at the position-4 greatly increased the rate of QM formation. However, QM formation is generally slower than the generation of the corresponding phenol intermediate (A), which indicated that QM formation is the rate-determining step for DNA cross-linking. The results obtained from NMR analysis further supported that the electron-donating group favors the QM formation therefore leading to more efficient DNA cross-linking.

Scheme 4.

Mechanism of H₂O₂-induced ICL formation and QM trapping reaction

The acetate group (OAc) was used as a good leaving group in a number of inducible DNA cross-linking agents developed by research groups of Freccero and Rokita.[23, 44] We expected that the arylboronic esters with OAc as a benzylic leaving group could be efficient H₂O₂-inducible DNA ICL agents. Compounds 1b, 2b, and 5b were successfully synthesized and investigated towards H₂O₂ and DNA reactivity. As expected, these compounds can be activated by H₂O₂ to form DNA ICLs. The cross-linking reactions for 1b, 2b, and 5b were complete within 90 mins, which proceeded 2–3 times faster than 1a, 2a, and 5a (Table 1). However, their cross-linking yields (1b: 20%; 2b: 26%; and 5b: 32%) are lower than the corresponding bromides 1a, 2a, and 5a (22%, 36%, and 50%) (Table 1). Further study showed that ICL formation induced by 1b, 2b, and 5b did not follow the first order kinetics (SI, Figure S13 C–E) and the cross-linking yield did not correlate well with the reaction rates. One possibility might be the hydrolysis of the acetate under the conditions used for DNA cross-linking. In order to test our hypothesis, the NMR analysis was performed with 1b, 2b, and 5b in a mixture of phosphate buffer (pH8 in D₂O) and DMSO in a 2:3 ratio. The results showed that the acetate group was hydrolyzed prior to addition of H₂O₂ (SI, Figure S16–18 B) resulting in low ICL formation for 1b, 2b, and 5b. We also observed that the oxidative deboronation of 1b, 2b and 5b with H₂O₂ was much slower than that of 1a, 2a and 5a. The reaction of 2a and 5a with H₂O₂ was completed within 30 min for 2a and 20 min for 5a (SI, Figure S14 and 15 E, F), while the reaction of 2b or 5b with H₂O₂ were not finished within one hour (SI, Figure S17 and 18 E). However, due to the complexity of the reactions, the rate constants could not be determined. A possible reason for the slower reaction of 2b and 5b towards H₂O₂ than 2a and 5a might be the formation of the hypercoordinated complexes A and/or B caused by the interaction between the oxygen of OAc and boron. There are precedents for the formation of such complexes between boron and oxygen or nitrogen.[45] In addition, NMR analysis showed that the resonances corresponding to the pinacol ester for OAc compound 2b (δ = 1.37 ppm for (C4’,5’-H) shifted to upper field than those for bromo compound 2a (δ = 1.46 ppm for 2a) (SI Figure S24 and S26). Slightly upper field shift was also observed for ¹³C NMR for the pinacol ester and 3 ppm up field shift observed for the aromatic C1 (SI Figure S27 and S29). These data suggested that coordination of oxygen and boron increased the electron density of boron therefore decreasing the deshielding effect. All of the above observations suggested that the acetate group was not a suitable leaving group for designing novel arylboronic esters as efficient H₂O₂-inducible cross-linking agents.

Having established that these compounds could be effectively activated by H₂O₂ to induce ICL formation, their toxicity towards cancer cells was evaluated. Initially, the ability of 1a and 2a to reduce cancer cell growth was determined with 60 human cancer cells lines by National Cancer Institute DTP program. Single dose screening at 10 µM showed that 2a induced significant growth inhibition of most cancer cell lines and was more toxic than 1a (SI, Figure S20 and 21). The growth percentage of most cell lines treated with 10 µM of 2a was less than 50%. Thus, dose dependent analysis of 2a was further evaluated in all 60 human cancer cell lines exhibiting GI₅₀ values of around 2 µM in most cancer cells (SI, Table S1).

Encouraged by the NCI results, we compared the effects of these compounds in a selected number of cancer cells, such as ovarian cancer SKOV3 cell, breast cancer MDA-MB-468 cell, and seven renal cancer cell lines. The initial test with SKOV3 cells showed that 1b, 2b, and 4a were not cytotoxic, however 1a, 2a, and 5a led to significant cancer cell death with an IC₅₀ of 6.3 µM for 1a, 5.2 µM for 2a, and 3.8 µM for 5a (SI, Figure S23). Therefore, we focused on the active compounds 1a, 2a, and 5a and compared their cytotoxicity with two clinically used alkylating agents, chlorambucil and melphalan in breast cancer cells MDA-MB-468 and seven different renal cancer cell lines, UO-31, A498, SN12C, 786-0, TK-10, CAKI-1 and ACHN. In general, 1a, 2a, and 5a were more effective than chlorambucil and melphalan in these cell lines. Additionally, 2a was more potent than 1a and 5a in most cells except for CAKI-1 cells that were more sensitive toward 5a than 1a and 2a (Table 2). Among the cells, MDA-MB-468 was the most sensitive cell line in respect to H₂O₂-activated QM prodrugs (Figure 2 and Table 2). Among different renal cancer cell lines, compounds 1a, 2a, and 5a are more cytotoxic to UO-31, A498, SN12C and 786-0 (IC₅₀ of 7.7 µM–27.8 µM) than TK-10, CAKI-1 and ACHN (Table 2, SI, Figure S22).

Table 2.

The Cytotoxicity of 1a, 2a and 5a towards renal and breast cancer cell lines

Tumor type	Cell line	IC50(µM)
		1a	2a	5a	Chlorambucil	Melphalan
renal cancer	UO-31	25.3	10.8	19.5	40.7	42.9
	A498	33.3	18.1	21	280	135
	SN12C	27.0	18.6	24.4	135	71
	786-0	20.3	10.8	27.8	55.5	19.2
	TK-10	21.4	16.2	35.4	n.d.	54.5
	ACHN	26.9	20.0	35.7	133	52.1
	CAKI-1	50.3	n.d.	38.3	n.d.	n.d.
breast cancer	MDA-MB-468	12.0	9.0	11.0	34.4	48.7

Figure 2.

Comparison of inhibition activity of 1a, 2a, and 5a with that of chlorambucil and melphalan in breast cancer cell line MDA-MB-468. The Fluorescence polarization (FP) assay was conducted in 384-well microplates (Corning, #3570). Compounds transfer into 20 µl assay solution was accomplished using a stainless steel pin tool (V&P Scientific) delivering 200 nl of compound at different concentrations. Incubation time is 48 h. Inhibition of cell growth was detected by fluorescence polarization using a M1000, Tecan reader at excitation/emission wavelength of 630/685 nm. IC₅₀ value (µM) was calculated using the following non-linear regression equation: Y = Bottom + (Top-Bottom)/(1+10^((LogIC50-X)*HillSlope)).

In addition to cell lines, compounds 1a, 2a, and 5a were also tested and compared with chlorambucil and melphalan in lymphocytes obtained from chronic lymphocytic leukemia (CLL) patients (n=3). Dose dependent apoptosis was observed for CLL cells. One representative patient data is provided in Figure 3A. Compounds 1a, 2a, and 5a demonstrated potent cytotoxicity (IC₅₀ – 48.3, 28.4 and 20.8 µM, respectively) in comparison to chlorambucil and melphalan (IC₅₀ – 98 and 76.9 µM). We tested the toxicity of compounds 1a, 2a, and 5a in normal lymphocytes obtained from healthy donors (n=2). These compounds exhibited no significant toxicity to normal lymphocytes.

Figure 3.

A. CLL lymphocytes were obtained from the peripheral blood of CLL patients (n=3) and incubated with increasing concentrations of compounds 1a, 2a, 5a, chlorambucil and melphalan (1, 5, 10, 20, 50, and 100 µM). The viability of CLL cells (measure of cytotoxicity) was measured by annexin/PI staining method. IC50 data for one representative patient data is provided. B. Normal lymphocytes were obtained from corresponding normal donors (n=2) and incubated with 10 µM concentration of compounds 1a, 2a, and 5a (24 hours) and the % viable cells was measured by annexin/PI binding assay. One representative donor data is provided. NL – normal lymphocytes.

DNA damage response is associated with apoptosis in leukemia cells

With the encouraging data on H₂O₂-induced ICL formation and growth inhibition, we also tested the compounds for the induction of DNA damage response in patient cells. CLL lymphocytes were incubated with 10 µM of 1a, 2a, and 5a (n=3; 24 hours) and the H2AX phosphorylation at Ser139 was measured Figure 4A).[46] There was induction of DNA damage response exhibiting heterogeneity among samples (p values 0.45, 0.15, and 0.04 respectively). We next correlated the apoptosis with DNA damage response in the same samples. Our results demonstrated connectivity between DNA damage response and cytotoxicity in CLL cells (Figure 4B–C). Measure of viability of cells and H2AX phosphorylation in 2a- and 5a-treated cells obtained from the same patient (n=4; % control) showed a linear correlation (r² = 0.8 and 0.97 respectively) between these two end-points.

Figure 4.

A. CLL lymphocytes obtained from patients (n=4) were untreated or treated with 1a, 2a and 5a (10 µM; 24 hrs) and the DNA damage response was measured by the induction of H2AX phosphorylation. B–C. % control values of apoptosis data (measured by annexin/PI binding assay) and the H2AX phosphorylation data obtained for the same samples (n=4), treated with 2a (B) and 5a (C) were correlated. Linear regression analysis and paired student’s t-tests (two tailed) were performed by GraphPad Prism 6 software (GraphPad Software, Inc. San Diego, CA).

CONCLUSIONS

In summary, our investigation of H₂O₂-induced reactivity of a series of novel arylboronate esters with DNA revealed that the position and number of electron donating aromatic substituents significantly affected their DNA cross-linking ability and cellular activity. An electron donating group placed meta to the QM methylene greatly facilitates QM formation, which is the rate-determining step for DNA cross-linking. Additionally, leaving groups influenced the reaction rate of these boronates with H₂O₂ and subsequent QM formation and DNA cross-linking. Br was observed as a better leaving group than acetate for these arylboronate derivatives. Hydrolysis of the acetates and possible hypercoordination with boronates might be the reasons for slow reaction rates with H₂O₂ as well as QM formation. Sufficient DNA-crosslinking of compounds (2a and 5a) with a donating group at the meta position directly translated to enhanced cellular activity with significant growth inhibition of leukemia, colon cancer, melanoma, and renal cancer cells. Replicationally quiescent CLL lymphocytes also showed induction of H2AX phosphorylation which is correlated with cell death with these compounds. This is a novel example of H₂O₂-induced QM prodrugs (2a and 5a) with potent anticancer activity. These novel molecular scaffolds will be used for further drug design as well molecular probes to elucidate cellular responses of DNA-alkylation.

EXPERIMENTAL SECTION

General Methods

Unless otherwise specified, all chemicals and reagents were commercially purchased and were used as received without further purification. Oligonucleotides were synthesized via standard automated DNA synthesis techniques in a 1.0 µM scale using commercial 1000Å CPG-succinyl-nucleoside supports. Deprotection of the nucleobases and phosphate moieties as well as cleavage of the linker were carried out under mild deprotection conditions using a mixture of 40% aq. MeNH₂ and 28% aq. NH₃(1:1) at room temperature for 2 h. Oligonucleotides were purified by 20% denaturing polyacrylamide gel electrophoresis. Radiolabeling was carried out according to the standard protocols. Quantification of radiolabeled oligonucleotides was carried out using a Phosphorimage equipped with ImageQuant Version 5.2 software. For all Phosphorimage autoradiogram, the top dark spot that migrated slowest is considered as DNA cross-link product and the bottom dark spot considered as the single-stranded ODN, which were determined using molecular ladder.[15, 19] All cross-linking yields were subtracted from the control experiment with DNA and drug without addition of H₂O₂. ¹H NMR and ¹³C NMR spectra were taken on 300 MHz or 500 MHz spectrophotometer. High resolution masss pectrometry was performed at the University of California-Riverside and University of Wisconsin-Milwaukee Mass Spectrometry Lab on an atmospheric-pressure chemical ionization (APCI) TOF mass spectrometer or electron spray injection mass spectrometer (ESI). The purity was determined by RP-HPLC on a 4.6×250 mm RP-C18 column with 254 nm detection, which confirmed that all compounds had ≥95% purity.

2-(2,6-Bis(bromomethyl)-4-methoxyphenyl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (2a)

A solution of 7 (1.31 g, 5 mmol), NBS (1.87 g 10.5 mmol), and AIBN (82.1 mg 0.5 mmol) in anhydrous CCl₄ (30 mL) was stirred to reflux with a light for 2 h. The mixture was allowed to cool to room temperature. The solvent was evaporated and 50 mL CH₂Cl₂ added. The organic layer was washed with water and brine, dried over sodium sulfate, and concentrated under reduced pressure then purified through column chromatography (30% DCM/Hexane) to give 2a (0.63 g 30%) as white solid: mp 103–104 °C; ¹H NMR (300 MHz, CDCl₃): δ 6.85 (s, 2H), 4.84 (s, 4H), 3.84 (s, 3H), 1.46 (s, 12H). ¹³C NMR (125 MHz, CDCl₃): δ 160.7, 146.4, 115.7, 84.0, 55.3, 34.1, 25.1. IT-TOF-MS (APCI): m/z calcd. for C₁₅H₂₁O₃BBr₂ [M+H]⁺ 419.0026, found 419.0022.

2-(3,5-Dimethoxy-2,6-dimethylphenyl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (11)

To a stirred solution of 10 (1.88 g, 7.7 mmol) in anhydrous THF (40 mL) was added dropwise a 2.5M solution of n-BuLi (3.7 mL, 9.25 mmol) at −78°C via cannula over a 2 min period under argon. The cloudy solution was stirred at −78°C for 30 min. Isopropoxyboronic acid pinacol ester (1.89 mL, 9.25 mmol) was added at −78°C under argon via syringe. The mixture was allowed to stir at −78°C for 30 min, then warmed slowly to room temperature and stirred for 6 h. The mixture was quenched with aqueous 1M HCl solution and extracted with 3 × 30 mL EtOAc. The organic layer was washed with water and brine, dried over sodium sulfate, concentrated under reduced pressure, and purified through column chromatography (5% EtOAc/Hexane) to give 11 (1.92 g 85%) as white solid: mp 140–141°C;¹H NMR (300 MHz, CDCl₃): δ 6.46 (s, 1H), 3.81 (s, 6H), 2.21 (s, 6H), 1.42 (s, 12H). ¹³C NMR (75 MHz, CDCl₃): δ 156.1, 121.2, 97.2, 83.9, 56.0, 25.1, 14.7. HRMS (ESI): m/z calcd. for C₁₆H₂₅O₄B [M]⁺ 291.1877, found 291.1872.

2-(2,6-Bis(bromomethyl)-3,5-dimethoxyphenyl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (3a)

A solution of 11 (584 mg, 2 mmol), NBS (854 mg, 4.8 mmol), and AIBN (32.8 mg, 0.2 mmol) in anhydrous CCl₄ (10 mL) was stirred to reflux under a light for 2h. The mixture was allowed to cool to room temperature and the solvent evaporated. Then, 20 mL CH₂Cl₂ was added. The organic layer was washed with water and brine, dried over sodium sulfate, and concentrated under reduced pressure to give 3a (120 mg 13%). ¹H NMR (300 MHz, CDCl₃): δ 6.47 (s, 1H), 4.9 (s, 4H), 3.92 (s, 6H), 1.49 (s, 12H). 3a is not stable enough for ¹³C NMR and MS.

4,4,5,5-Tetramethyl-2-(3,4,5-trimethoxy-2,6-dimethylphenyl)-1,3,2-dioxaborolane (14)

To a stirred solution of 13 (3.88 g, 14.2 mmol) in anhydrous THF (40 mL) was added dropwise a 2.2 M solution of n-BuLi (7.7 mL, 17 mmol) at −78°C via cannula over a 2 min period under argon. The cloudy solution was stirred at −78°C for 30 min. Then, isopropoxyboronic acid pinacol ester (3.47 mL, 17 mmol) was added at −78°C under argon via syringe. The mixture was allowed to stir at −78°C for 30 min, warm slowly to room temperature, and stirred for 6 h. The mixture was quenched with aqueous 1M HCl solution, extracted with 3 × 30 mL EtOAc, the organic layer was washed with water and brine, dried over sodium sulfate, and concentrated under reduced pressure then purified through column chromatography (5% EtOAc/Hexane) to give 14 (3.32 g 73%) as colorless oil. ¹H NMR (300 MHz, CDCl₃): δ 3.91 (s, 3H), 3.79 (s, 6H), 2.28 (s, 6H), 1.41 (s, 12H). ¹³C NMR (75 MHz, CDCl₃): δ 149.7, 147.2, 130.3, 83.9, 60.7, 60.5, 25.0, 15.1. HRMS (ESI): m/z calcd. for C₁₇H₂₇O₅B [M]⁺ 321.1982, found 321.1983.

2-(2,6-Bis(bromomethyl)-3,4,5-trimethoxyphenyl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (4a)

A solution of 14 (1.28 g, 4 mmol), NBS (1.71 g 9.6 mmol) and AIBN (65.6 mg 0.4 mmol) in anhydrous CCl₄ (30 mL) was stirred to reflux under a light for 2h. The mixture was allowed to cool to room temperature. After evaporation of the solvent followed by addition of 20 mL CH₂Cl₂, the organic layer was washed with water and brine, dried over sodium sulfate, and concentrated under reduced pressure then purified through column chromatography (2.5% EtOAc/Hexane) to give 4a (0.53 g 28%) as white solid: mp 96–98°C;¹H NMR (300 MHz, CDCl₃) δ 4.90 (s, 4H), 3.99 (s, 6H), 3.91 (s, 3H), 1.48 (s, 12H). ¹³C NMR (75 MHz, CDCl₃): δ 152.7, 147.4, 132.8, 84.5, 61.2, 60.6, 27.6, 25.1. HRMS (ESI): m/z calcd. For C₁₇H₂₅O₅BBr₂ [M]⁺ 477.0193, found 477.0179.

(3,5-Bis(bromomethyl)-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenoxy)(tert-butyl) dimethylsilane (18)

A solution of 17 (4.1 g, 11.8 mmol), NBS (5.11 g, 28.4 mmol) and AIBN (194 mg, 0.118 mmol) in anhydrous CCl₄ (100 mL) was stirred to reflux under a light for 2h. The mixture was allowed to cool to room temperature. After evaporation of the solvent followed by addition of 100 mL CH₂Cl₂, the organic layer was washed with water and brine, dried over sodium sulfate, and concentrated under reduced pressure then purified through column chromatography (20% DCM/Hexane) to give 18 (2.67 g 44%) as white solid: mp 81–82 °C;¹H NMR (300 MHz, CDCl₃): δ 6.79 (s, 2H), 4.80 (s, 4H), 1.47 (s, 12H), 1.00 (s, 9H), 0.23 (s, 6H). ¹³C NMR (125 MHz, CDCl₃): δ 157.1, 146.3, 121.9, 84.0, 34.0, 25.7, 25.1, 18.2.IT-TOF-MS (APCI): m/z calcd. for C₂₀H₃₃O₃BSiBr₂ [M+H]⁺ 519.0736, found 519.0736.

3,5-Bis(bromomethyl)-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenol (5a)

A solution of 18 (2.6 g, 5.0 mmol) and TBATB (0.24 g, 0.5 mmol) in MeOH (50 mL) was stirred for one day, then another portion of TBATB (0.24 g, 0.5 mmol) was added to the solution. The mixture was stirred for one more day, evaporated, and purified through column chromatography (20% DCM/Hexane) to give 5a (0.72 g 36%) as white solid: mp 195–195°C;¹H NMR (300 MHz, CDCl₃): δ 6.78 (s, 2H), 4.79 (s, 4H), 1.46 (s, 12H). ¹³C NMR (125 MHz, CDCl₃): δ 156.8, 146.7, 117.1, 84.1, 33.7, 25.1. HRMS (ESI): m/z calcd. for C₁₄H₁₉O₃BBr₂ [M]⁺ 402.9825, found 402.9814.

(2-Bromo-1,3-phenylene)bis(methylene) diacetate (21)

To a solution of 20 (1.08 g, 5 mmol) in CH₂Cl₂ (25 mL) was added Et₃N (2.02 g, 20 mmol), pyridine (1.58 g, 20 mmol), and acetyl chloride (1.56 g, 20 mmol) at 0 °C. The mixture was allowed to warm up to room temperature and stirred overnight. The reaction mixture was washed with water and brine, dried over sodium sulfate, and concentrated under reduced pressure, then purified through column chromatography (2.5% EtOAc/Hexane) to give 21 (1.13 g, 75%) as white solid: mp 51–52°C;¹H NMR (300 MHz, CDCl₃) δ 7.43-7.36 (m, 3H), 5.25 (s, 4H), 2.17 (s, 6H). ¹³C NMR (75 MHz, CDCl₃): δ 170.6, 136.1, 129.4, 127.4, 124.3, 66.0, 20.9. IT-TOF-MS (APCI): m/z calcd. for C₁₂H₁₃O₄Br [M+H]⁺ 301.0070, found 301.0060.

(2-(4,4,5,5-Tetramethyl-1,3,2-dioxaborolan-2-yl)-1,3-phenylene)bis(methylene) diacetate (1b)

A mixture of 21 (301 mg, 1 mmol), bis(pinacolato)diboron (508 mg, 2 mmol), KOAc (589 mg, 6 mmol), and PdCl₂(dppf) (49 mg, 0.06 mmol) in 1,4-dioxane (20 mL) was refluxed under argon overnight and cooled to room temperature. Then, water was added and the mixture was extracted with EtOAc. The combined organic layer was washed with water and brine, dried overs odium sulfate, and concentrated under reduced pressure then purified through column chromatography (50% DCM/Hexane) to provide 1b (121.9 mg, 35%) as white solid: mp 43–44 °C;¹H NMR (300 MHz, CDCl₃): δ 7.36–7.39 (m, 3H), 5.28 (s, 4H), 2.08 (s, 6H), 1.39 (s, 12H). ¹³C NMR (75 MHz, CDCl₃): δ 170.6, 141.0, 130.0, 129.2, 84.1, 66.6, 24.9, 21.1. IT-TOF-MS (ESI): m/z calcd. for C₁₈H₂₅O₆B [M+NH₄]⁺ 366.2086, found 366.2084.

(2-Bromo-5-methoxy-1,3-phenylene)bis(methylene) diacetate (24)

A mixture of 23 (0.71 g, 1.9 mmol) and NaOAc (0.79 g, 9.6 mmol) were suspended in DMF (20 mL) and heated for 8 h at 80 °C. The mixture was allowed to cool to room temperature, diluted with EtOAc, and washed with water and brine. The organic layer was dried over sodium sulfate and concentrated under reduced pressure then purified through column chromatography (10% DCM/Hexane) to give 24 (0.51g, 81%) as white solid: mp 81–82 °C; ¹H NMR (300 MHz, CDCl₃): δ 6.96 (s, 2H), 5.21 (s, 4H), 3.84 (s, 3H), 2.18 (s, 6H). ¹³C NMR (75 MHz, CDCl₃): δ 170.5, 158.9, 137.0, 114.9, 114.2, 65.9, 55.6, 20.9.IT-TOF-MS (ESI): m/z calcd. for C₁₃H₁₅O₅Br [M+H₂O]⁺ 348.0203, found 348.0200.

(5-Methoxy-2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1,3-phenylene)bis(methylene) diacetate (2b)

A mixture of 24 (331 mg, 1 mmol), bis(pinacolato)diboron (508 mg, 2 mmol), KOAc (589 mg, 6 mmol), and PdCl₂(dppf) (49 mg, 0.06 mmol) in 1,4-dioxane (20 mL) was refluxed under argon overnight and cooled to room temperature. Water was added and the mixture was extracted with EtOAc. The combined organic layer was washed with water and brine, dried over sodium sulfate, and concentrated under reduced pressure then purified through column chromatography (50% DCM/Hexane) to provide 2b (147.5 mg, 39%) as white solid: mp 37–38 °C;¹H NMR (300 MHz, CDCl₃): δ 6.91 (s, 2H), 5.29 (s, 4H), 3.85 (s, 4H), 2.10 (s, 6H), 1.37 (s, 12H). ¹³C NMR (75 MHz, CDCl₃): δ 170.7, 160.9, 143.6, 114.5, 83.8, 66.6, 24.8, 21.1. IT-TOF-MS (ESI): m/z calcd. for C₁₉H₂₇O₇B [M+NH₄]⁺ 396.2192, found 396.2186.

(5-Hydroxy-2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1,3-phenylene)bis(methylene) diacetate (5b)

A mixture of 5a (100 mg, 0.248 mmol) and NaOAc (101.5 mg, 1.238 mmol) were suspended in DMF (10 mL) and heated for 8 h at 80 °C. The mixture was allowed to cool to room temperature, diluted with EtOAc, washed with water and brine, dried over Na₂SO₄, and concentrated under reduced pressure then purified through column chromatography (5% EtOAc/DCM) to give 5b (27 mg, 31%) as white solid: mp 178–179 °C; ¹H NMR (300 MHz, CDCl₃): δ 6.86 (s, 2H), 5.28 (s, 4H), 2.10 (s, 6H), 1.37 (s, 12H). ¹³C NMR (75 MHz, CDCl₃): δ 171.9, 157.1, 144.0, 115.7, 83.8, 66.3, 24.8, 21.1. IT-TOF-MS (ESI): m/z calcd. for C₁₈H₂₅O₇B [M+NH₄]⁺ 382.2035, found 382.2035.

QM Trapping Assay for 2a

A solution of 2a (50 mg) in a mixture of CH₃CN (3 mL) and 1 M potassium phosphate buffer (52 µL, pH 8) was incubated at 37 °C for 30 min in the presence of excess ethyl vinyl ether (EVE). Then, H₂O₂ (2 equivalent of 2a) was added to the reaction mixture that was stirred at 37 °C for 24 h, then evaporated. Water (2 mL) was added to the residue and extracted with ethyl acetate (3 × 5 mL). The organic phase was combined, dried over anhydrous Na₂SO₄, and evaporated. The crude product was purified through column chromatography with 0–50% EtOAc in hexane to provide the QM-EVE adduct 26.

(2-Ethoxy-6-methoxychroman-8-yl)methanol (26)

Colorless oil, 15% yield (4.3 mg). ¹H NMR (300 MHz, CDCl₃): δ6.76 (s, 1H), 6.62 (s, 1H), 5.34 (s, 1H), 4.59-4.49 (m, 2H), 4.05-3.95 (m, 1H), 3.77 (s, 1H), 3.73-3.68 (m, 1H), 3.04-2.93 (m, 1H), 2.67-2.59 (m, 1H), 2.08-1.97 (m, 2H), 1.22 (t, J = 7.2 Hz, 3H). ¹³C NMR (500 MHz, CDCl₃): δ 152.9, 126.0, 124.0, 115.2, 114.0, 96.8, 63.8, 55.7, 29.0, 26.4, 20.9, 15.1. IT-TOF-MS (ESI): m/z calcd. for C₁₃H₁₈O₄ [M+H]⁺ 239.1278, found 239.1285.

QM Trapping Assay for 5a

To a solution of 5a (100 mg) in CH₃CN, 83.6 mg potassium phosphate and excess EVE were added. Then, H₂O₂ (2 equivalent of 5a) was added to the reaction mixture that was stirred at 37 °C for 24 h, then evaporated. Water (2 mL) was added to the residue and extracted with ethyl acetate (3 × 5 mL). The organic phase was combined, dried over anhydrous Na₂SO₄, and evaporated. The crude product was purified through column chromatography with 10% EtOAc in hexane to provide the QM-EVE adduct 27.

8-(Bromomethyl)-2-ethoxychroman-6-ol (27)

Colorless oil, 20% yield (13.7 mg). ¹H NMR (300 MHz, DMSO-d6): δ6.63 (s, 1H), 6.46 (s, 1H), 5.31 (s, 1H), 4.59-4.50 (q, 2H, J = 9.3Hz), 3.92-3.82 (m, 1H), 3.69-3.58 (m, 1H), 2.79-2.68 (m, 1H), 2.55-2.49 (m, 1H), 1.96-1.80 (m, 2H), 1.13-1.08 (t, J = 6.9 Hz, 3H). ¹³C NMR (500 MHz, DMSO-d₆): δ 150.7, 143.1, 125.9, 123.9, 116.8, 115.8, 96.7, 63.4, 30.5, 26.4, 20.6, 15.5. Decomposition was observed during NMR measurement. We also observed that the solution changed from colorless to orange after NMR analysis. The instability also made the mass analysis of 27 C₁₂H₁₅O₃Br more difficult. The molecular ion [M]⁺ (286.0205) peak was not observed. However, the fragmentation of 28 [M-Br]⁺ 207.1021 was detected, found 207.0979.

Interstrand cross-link formation with duplex DNA 25

The ³²P-labelled oligonucleotide (0.5 µM) was annealed with 1.5 equiv of the complementary strand by heating to 65 °C for 3 min in a solution of 10 mM potassium phosphate buffer (pH 7) and 100 mM NaCl, followed by cooling slowly to room temperature overnight. The ³²P-labeled oligonucleotide duplex (2 µL, 0.5 µM) was mixed with 1 M NaCl (2 µL), 100 mM potassium phosphate (2 µL, pH 6–8), H₂O₂ (2 µL), and compounds 1–5 (concentration range: 100 µM to 2 mM in 6 µL CH₃CN). Then, the appropriate amount of autoclaved distilled water was added to give a final volume of 20 µL. The reaction was incubated at room temperature for 24 h and quenched by an equal volume of 90% formamide loading buffer, then subjected to 20% denaturing polyacrylamide gel electrophoresis.

Cell inhibition study of 1a, 2a and 5a towards tumor cells

The in vitro cancer cell screening with 60 human cancer cell lines was performed at the National Cancer Institute (NCI Developmental Therapeutics Program). The procedure details can be found in NCI website: http://dtp.nci.nih.gov/branches/btb/ivclsp.html. Methodology of the In Vitro cancer screen. The human tumor cell lines were grown in RPMI 1640 medium containing 5% fetal bovine serum and 2 mM L-glutamine. Cells were inoculated into 96 well microtiter plates in 100 µL at plating densities ranging from 5,000 to 40,000 cells/well depending on the doubling time of individual cell lines.

Cytotoxicity study of 1a, 2a, and 5a towards tumor cells

The human tumor cell lines were grown in RPMI 1640 medium containing 5% fetal bovine serum and 2 mM L-glutamine. Cells were inoculated into 384-well microtiter plates in 20 µL at plating densities ranging from 5,000 to 10,000 cells/well. After cell inoculation, the microtiter plates were incubated at 37°C, 5 % CO₂, 95% air, and 100% relative humidity for 2–3 h prior to addition of drugs.

Drugs were solubilized in DMSO at 20 mM concentration and serially diluted ten times each time 50% dilution in DMSO in a 384-well plate. Then 200 nL of the serially diluted drug solution was added to the cell plate (1:100 dilution) using FreedomEVOware two times 100 nL transfer. Following drug addition, the plates were incubated for an additional 48 h at 37°C, 5% CO₂, 95% air, and 100% relative humidity. After 48 hours, 20 µL CellTiter-Glo Luminescent solution was added to the cell plate. The plate was then incubated at room temperature for 10 mins before the luminescent was measured with Infinite M1000.

Patients and healthy donors

CLL or normal lymphocytes were separated from peripheral blood by Ficoll-hypaque gradient method and suspended in 10% autologous plasma in RPMI media. All patient and healthy donor participants had signed written informed consent forms in accordance with the Declaration of Helsinki, and the laboratory protocols were approved by the Institutional Review Board at the UT MD Anderson Cancer Center.

Measurement of cell viability

CLL cell viability (measure of apoptosis) was measured by annexin V/propidium iodide (PI) binding method.

Measurement of H2AX Phosphorylation

The lymphocytes harvested after incubation with compounds were washed with PBS once and fixed in 6–8 mL ice-cold ethanol (70%) and stored at −20°C until analysis for H2AX phosphorylation. The cells were then permeabilised and stained with FITC tagged H2AX^Ser139 antibody (Biolegend, San Diego CA) and measured in the flow cytometer according to manufacturer’s instructions.

Scheme 5.

The possible models for boronate hypercoordination

Abbreviations

AIBN

azobisisobutyronitrile

APCI

atmospheric pressure chemical ionization

CLL

chronic lymphocytic leukemia

DCM

dichloromethane

DMF

dimethylformamide

DMSO

dimethyl sulfoxide

ESI

electrospray ionization

EVE

ethyl vinyl ether

FITC

fluorescein isothiocyanate

HRMS

high resolution mass spectrometry

ICLs

interstrand cross-links

IT-TOF-MS

Ion-Trap-Time-of-Flight-Mass-Spectrometer

NBS

N-bromosuccinimide

NL

normal lymphocytes

PAGE

denaturing polyacrylamide gel electrophoresis

PBS

Phosphate-buffered saline

PI

propidium iodide

QM

quinone methide

ROS

reactive oxygen species

TBATB

tetrabutylammonium tribromide

TBDMSCl

tert-Butyldimethylsilyl chloride