Dual repair modulation reverses Temozolomide resistance in vitro

Temozolomide is an alkylating agent that mediates its cytotoxic effects via O-methylguanine (O-meG) adducts in DNA and their recognition and processing by the postreplication mismatch repair system (MMR). O-meG adducts can be repaired by the DNA repair protein Oalkylguanine-DNA-alkyltransferase (MGMT), which therefore constitutes a major resistance mechanism to the drug. Resistance to Temozolomide can also be mediated by loss of MMR, which is frequently mediated by methylation of the hMLH1 gene promoter. Methylation of hMLH1 can be reversed by treatment of cells with 5-aza-2V-deoxycytidine, while the MGMT pseudosubstrate O-(4-bromothenyl)guanine (PaTrin-2) can deplete MGMT activity. Using a drug-resistant cell line which expresses MGMT and has methylated hMLH1, we show that while either of these treatments can individually sensitize cells to Temozolomide, the combined treatment leads to substantially greater sensitization. The increased sensitization is not observed in matched MMR proficient cells. [Mol Cancer Ther. 2004;3(2):123–127]


Introduction
Temozolomide is an alkylating agent that mediates its cytotoxic effect by forming O 6 -methylguanine (O 6 -meG) DNA adducts which, during DNA replication, pair preferentially with thymidine.These O 6 -meG:T mispairs result in a G to A point mutation during a subsequent round of DNA replication (1).O 6 -mG:T mispairs are also recognized by the postreplication mismatch repair system (MMR), which leads the cell to death.Consistent with a role of MMR in engaging cell death, loss of MMR leads to resistance to Temozolomide (2).MMR is a group of proteins involved in the repair of base-base mispairs and deletion-insertion loops (3).Mutations of some of the MMR genes (hMLH1 and/or hMSH2) are known to play a major role in the etiology of hereditary non-polyposis colorectal cancer (4).MMR is not functional in many tumors, either as a result of a mutation in one of the MMR genes, or because of an epigenetic silencing due to methylation of CpGisland, in the hMLH1 gene promoter.This is a common event in tumor tissues (5), and in the case of hMLH1, it has been observed that as much as 44% of colon tumor tissues contain hMLH1 promoter methylation (6).
5-Aza-2V-deoxycytidine (5-aza-CdR) is a pyrimidine analogue that has been extensively studied since its synthesis in 1964 (7).In the cell, 5-aza-CdR is phosphorylated by deoxycytidine kinase, then rapidly converted to the triphosphate, 5-aza-dCTP, by other kinases.5-aza-CTP is a substrate for DNA polymerase a and is incorporated into DNA.The presence of 5-aza-CdR in DNA results in the inactivation of DNA cytosine methyltransferase by the irreversible formation of a covalent bond between this enzyme and the analogue (8).The inhibition of DNA methylation using 5-aza-CdR became a research focus when some reports suggested that such analogues could reactivate previously silenced DNA repair genes in mammalian cells (9,10).More recently, it has been observed that resistance to Temozolomide (and a variety of other clinically important cytotoxic drugs) in ovarian and colon tumor xenografts could be reversed by 5-aza-CdR, which induces demethylation of the hMLH1 promoter and therefore restoration of a functional MMR (11).
O 6 -meG DNA adducts can be repaired by the DNA repair protein O 6 -alkylguanine-DNA-alkyltransferase (MGMT).MGMT removes adducts from the O 6 position of guanine by accepting them onto a cysteine residue within its active site.The protective role of MGMT against the cytotoxic effect of Temozolomide has been shown in human cell lines and human xenograft models (12).Fibroblasts and bone marrow cells of MGMT knockout mice are more sensitive to Temozolomide (13) and the important role played by both MGMT and MMR (hMLH1) in the cytotoxic effect of alkylating agents has been confirmed in a double knockout mouse model (14).
MGMT can be successfully inactivated by free guanine base derivatives, with alkyl groups at the O 6 position, which act as ''pseudosubstrates.''O 6 -benzylguanine (15) and O 6 -(4-bromothenyl)guanine (PaTrin-2) (16) have been identified as the most promising MGMT inactivators.Compared to Temozolomide used as a single agent, the combination PaTrin-2/Temozolomide has been shown to significantly increase tumor growth inhibition in human melanoma xenografts (17); PaTrin-2 and O 6 -benzylguanine have recently entered phase II clinical trials (12).
Given the role of MMR in O 6 -meG toxicity, MGMT inactivators would not be expected to be effective in increasing the killing effects of Temozolomide in MMR defective cells.Conversely, MMR reactivation by 5-aza-CdR might not significantly increase killing in a cell that was MGMT proficient.We therefore set out to examine if reactivation of MMR in a silenced but MGMT proficient cell line would increase the toxicity of Temozolomide more extensively if combined with an MGMT inactivator.
A2780-Cp70 is a human ovarian tumor cell line which lacks functional MMR, compared to the parental cell line A2780, because of hMLH1 gene silencing by promoter methylation (18).It has been observed that the sensitivity of A2780-Cp70 xenografts to the cytotoxic effect of Temozolomide can be increased by 5-aza-CdR (11).We now show that MGMT, which is expressed in both cell lines, can be successfully depleted by PaTrin-2, and that the hMLH1 gene is re-expressed in the A2780-Cp70 cell line following exposure to 5-aza-CdR.Furthermore, we demonstrate that by combining 5-aza-CdR with PaTrin-2, a further sensitization of the A2780-Cp70 cell line to the cytotoxic effect of Temozolomide can be achieved.
A2780 and A2780-Cp70 cells were screened on a weekly basis throughout the experiment for Mycoplasma by PCR (VenorGeM, Cambridge, United Kingdom, PCR-Mycoplasma Detection Kit; protocol detailed in the user manual, pp. 9 -12) and always proved to be contamination-free.

MGMT Assay
The preparation of cell extracts for MGMT assay, the MGMT assay itself, and protein estimations were performed following the protocols described by McElhinney et al. (16).Briefly, the MGMT assay involved measuring [ 3 H]methyl group transfer to MGMT protein: high specific radioactivity [ 3 H]-methylated DNA substrate was incubated with extract under protein-limiting conditions until the transfer reaction was complete.Excess substrate DNA was hydrolyzed to acid solubility and radioactivity in the residual protein was measured by liquid scintillation counting.

Depletion of MGMT by PaTrin-2
The MGMT activity of A2780 and A2780-Cp70 cells and the ability of PaTrin-2 to deplete this activity were initially determined.On day 1, equal number (2 Â 10 6 ) of A2780 and A2780-Cp70 cells were seeded in a series of T75 flasks containing 10 ml of RPMI 1640, supplemented with 10% FCS and 200 mM L-glutamine, and 10 AM PaTrin-2.After 2, 24, 48, 72, 96, and 120 h, cells were rinsed twice with 10 ml EDTA (200 mg/l H 2 O; Sigma), then harvested for MGMT assay.On day 5, the medium was changed to one not containing PaTrin-2.Cells were harvested for MGMT assay on days 6, 7, 8, and 9.
Effect of 5-Aza-CdR on hMLH1 Expression in A2780-Cp70 Equal number (0.5 Â 10 6 ) of A2780 and A2780-Cp70 cells were seeded in T75 flasks containing 10 ml of RPMI 1640, supplemented with 10% FCS and 200 mM L-glutamine on day 1.On day 2, the medium was changed to one containing 10 AM 5-aza-CdR (Sigma), or the equivalent concentration of the vehicle (0.16 mM acetic acid) as negative control.On day 12, the culture medium was removed from the flasks containing vehicle-treated or 5-aza-CdR-treated A2780 and A2780-Cp70 cells.Cells were rinsed twice with 10 ml EDTA, trypsinized (Sigma) and removed from the culture flasks on dilution with 10 volumes of medium.The resulting cell suspension was centrifuged at 150 Â g for 5 min and the pellet was then washed once in culture medium to remove excess trypsin and recentrifuged.Culture medium was removed from the final pellet and cells were fixed by the addition of 70% ethanol, and left overnight at room temperature.Ethanol was then removed, the samples were embedded in paraffin wax, and 4-AM sections were cut and mounted on slides.
Immunocytochemistry Fixed cells were immunostained for hMLH1 using an immunoperoxidase procedure as previously described (19).Briefly, monoclonal antibodies against human hMLH1 (BD PharMingen, Franklin Lakes, NJ) were used as the primary antibody.The secondary antibody was a biotinylated rabbit anti-mouse (DAKO, Glostrup, Denmark), which was then followed by streptavidin-horseradish peroxidase incubation (StreptABComplex/horseradish peroxidase; DAKO).Finally, the samples were counterstained with haematoxylin, dehydrated, and mounted.
Cell Growth Assays MTT Assay.Cells were seeded in T75 flasks and treated with 5-aza-CdR or its vehicle as described above.On day 12, an equal number (500) of vehicle-treated or 5-aza-CdRtreated A2780 and A2780-Cp70 cells were seeded in 96-well plates (Falcon), with 100 Al of medium containing 10 AM PaTrin-2 or the equivalent concentration of the vehicle (0.05% DMSO; Sigma) as negative control.On day 13, 100 Al of medium containing Temozolomide at increasing concentrations (0 -100 Ag/ml of DMSO) were added to each well.On day 19 (or when cells were subconfluent), 50 Al of MTT (3 mg/ml H 2 O, Sigma) were added to the cells, and incubated at 37jC for 3 h.Each well was then emptied.Two hundred microliters of DMSO were added to each well, and after 10 min, the plates were read at 540 and 690 nm on a multiscan plate reader.
Clonogenic Assay.Cells were seeded in T75 flasks and treated with 5-aza-CdR or its vehicle as described above.On day 12, an equal number (8 Â 10 4 ) of vehicle-treated or 5-aza-CdR-treated A2780 and A2780-Cp70 cells were seeded in T25 flasks (Falcon), with 5 ml of medium containing 10 AM PaTrin-2 or the equivalent concentration of the vehicle as negative control.On day 14, the medium was replaced by one containing 10 AM PaTrin-2 or the equivalent concentration of the vehicle, and Temozolomide at increasing concentrations (0 -100 Ag/ml of DMSO).On day 16, cells were rinsed twice with 10 ml EDTA, harvested, and counted.Equal number (10 3 ) of cells were seeded in Petri dishes (Falcon) (three dishes per experiment, i.e., 196 dishes in total) with 5 ml of medium (no drug addition).On day 23, the medium was removed and 5 ml of gentian violet [10% (w/v) methyl violet 2B (Sigma-Aldrich Co. Ltd., Poole, Dorset, United Kingdom) in 70% ethanol (Prolabo, Manchester, United Kingdom)] was added to each dish for 5 min (cell staining).The dye was then poured off, the dish rinsed with tap water, and the colonies counted after the dishes were inverted and allowed to dry.

Statistical Analysis
In each experiment, and for each concentration of Temozolomide, a series of nine absorbance data measurements (MTT assay) or cell colonies count (clonogenic assay) was obtained.The mean was taken and expressed as a percentage of the appropriate control.Each set of experimental data was compared using Student's t test.

MGMT Activity of A2780 and A2780-Cp70 Cells and MGMT Depletion by PaTrin-2
The MGMT activity of A2780 and A2780-Cp70 cells were 447.2 F 45.3 and 473.5 F 53.3 fmol/mg protein, A2780-Cp70 cells (n).MGMT activity was measured before the addition of 10 AM PaTrin-2, then 2, 24, 48, 72, 96, and 120 h after the addition of 10 AM PaTrin-2.Cell culture medium was changed on day 5 for one not containing PaTrin-2; MGMT activity was measured on days 6, 7, 8, and 9.  respectively.The MGMT activity became undetectable 2 h after the addition of PaTrin-2 and remained undetectable while PaTrin-2 was present in the cell culture medium.
When the medium was replaced by one not containing PaTrin-2, the MGMT activity remained undetectable for a further 48 h.Four days after medium replacement, the MGMT activities of A2780 and A2780-Cp70 cells had recovered but only to 35% and 33%, respectively, of their normal values (Fig. 1).Effect of 5-Aza-CdR on hMLH1 Expression in A2780-Cp70 Cells The immunocytochemical analysis showed an absent expression of hMLH1 in A2780-Cp70 cells before treatment with 5-aza-CdR.In contrast, after 5-aza-CdR treatment, expression of hMLH1 in A2780-Cp70 cells was abundant and similar to that observed in the parental A2780 cells (treated or not with 5-aza-CdR) (Fig. 2).
When treated with the demethylating agent 5-aza-CdR (Table 1), the sensitivity of A2780-Cp70 cells to Temozolomide was significantly increased (P = 0.0003), and became similar to that of A2780 cells.The combination of 5-aza-CdR and PaTrin-2 further enhanced the sensitivity of A2780-Cp70 cells to Temozolomide (P = 0.0003), the overall sensitivity being closely similar to that of similarly treated A2780 cells (Table 1 and Fig. 4).

Discussion
Our hypothesis was that MGMT inactivation would increase Temozolomide sensitivity in MMR-reactivated cells.To test this, we first measured MGMT expression in the two cell lines and found them to be closely similar.
Next, we showed that PaTrin-2 would inactivate MGMT, and determined that the level of inactivation would be adequate to prevent resynthesis during the course of the experiment.As anticipated, MGMT inactivation had only a marginal effect on the sensitivity of the A2780-Cp70 cell line to Temozolomide toxicity, but markedly increased the sensitivity of A2780 cells to Temozolomide.Also, 5-aza-CdR increased the sensitivity of A2780-Cp70 cells to Temozolomide, but had no effect on the sensitivity of A2780 cells to the methylating agent.Finally, the combination of 5-aza-CdR and PaTrin-2 resulted in increasing the Temozolomide sensitivity of A2780-Cp70 cells to be almost identical to that of PaTrin-2-treated A2780 cells.
It therefore appears that this combination of agents may be appropriate for clinical studies specifically of tumors with promoter methylation-mediated down-regulation of MMR and expressing MGMT.Both of these conditions could be determined in biopsies of the tumors, so that patients with the most appropriate tumors could be preselected.However, another question to be addressed when considering the use of combined therapies in clinical trials is the additional toxicity.PaTrin-2 depletes ATase not only in tumors, but also in other tissues such as the bone marrow.The cytotoxic effect of Temozolomide, combined with PaTrin-2, is therefore enhanced in tumors, but myelosuppression is also increased, forcing a modest reduction of the dose of the alkylating agent, although some strategies are now being implemented to protect the bone marrow (20).The cytotoxic effect of 5-aza-CdR has been used in the treatment of acute myeloid (non lymphoblastic) leukaemia (21), with myelosuppression as a major side effect.However, the reactivation of previously silenced genes by 5-aza-CdR requires a significantly lower dose of the drug.Myelosuppression caused by 5-aza-CdR is therefore expected to be minimal at doses that induce demethylation: clinical trials with appropriate pharmacodynamic measures of DNA demethylation will be necessary to confirm this possibility.In addition, because normal  tissues are MMR proficient, 5-aza-CdR might not be expected to influence MMR-mediated cell killing by Temozolomide.The additional therapeutic benefit of PaTrin-2 in combination with 5-aza-CdR may thus be very substantial.
In summary, we show that a significant enhancement of sensitivity to the cytotoxic effect of Temozolomide can be achieved by combining the action of an MGMT inactivator and a demethylating agent in the A2780-Cp70 ovarian tumor cell line.This opens the possibility of further experiments involving xenograft models and future clinical trials.