Gene amplification is one of the most frequent manifestations of genomic instability in human tumors and plays an important role in tumor progression and acquisition of drug resistance. To better understand the factors involved in acquired resistance to cytotoxic drugs via gene amplification, we have analyzed the structure and dynamics of dihydrofolate reductase (DHFR) gene amplification in HT29 cells treated with methotrexate (MTX). Analysis of the DHFR gene amplification process shows that the amplicon exhibits a complex structure that is consistently reproduced in independent treatments. The cytogenetic manifestation of the amplification in advanced stages of the treatment may be in the form of double minutes or as a homogeneously stained region. To get insights into the mechanisms of resistance, we have also investigated the sensitization to MTX of MTX-resistant cells after drug withdrawal and reexposure to MTX. Passive loss of the DHFR amplicon by withdrawal of the drug results in MTX-sensitive cells exhibiting a substantial reduction of their capacity or even an incapacity to generate resistance when submitted to a second cycle of MTX treatment. On a second round of drug administration, the resistant cells generate a different amplicon structure, suggesting that the formation of the amplicon as in the first cycle of treatment is not feasible. These results indicate that DHFR gene amplification is a “wear and tear” process in HT29 cells and that MTX-resistant cells may become responsive to a second round of treatment if left untreated during a sufficient period of time. [Mol Cancer Ther 2009;8(2):424–32]
Tumor cells arise from normal cells through the accumulation of multiple genetic and epigenetic alterations that positively and negatively regulate aspects of cell proliferation, apoptosis, genome stability, angiogenesis, invasion, and metastasis (reviewed in ref. 1). Genetic alterations include point mutations, deletions, inversions, translocations, and DNA sequence amplifications and result in the activation or inactivation of proto-oncogenes or tumor suppressor genes, respectively (1). Gene amplification is one of the most frequent manifestations of genomic instability in human tumors and plays an important role in tumor progression and in the acquisition of drug resistance. Amplified chromosomal regions (amplicons) contain multiple copies of one or more genes and result in their overexpression (2, 3). Cytogenetic manifestations of gene amplification are mainly of two types: double minutes (DM) and homogeneously staining regions (HSR; ref. 2, 4). DMs are small, spherical, acentric, atelomeric, usually paired, and autonomously replicating extrachromosomal structures. HSRs appear as light disruption regions in a normal pattern of alternating dark and light bands in Giemsa staining of metaphase chromosomes. HSRs may be located at the same locus of the original gene or translocated to another chromosome. The relationship between these two types of amplification has been widely discussed but remains uncertain and controversial (2). Different mechanisms have been proposed to explain the dynamics of gene amplification, including breakage-fusion-bridge cycles, unequal sister chromatid exchange, unscheduled rounds of DNA replication during cell division, or replication fork arrest followed by recombination (3, 5). All of them have been associated with chromosome double-strand breaks (6) or activation of fragile sites (7).
Methotrexate (MTX) and its polyglutamate forms are potent competitive inhibitors of the dihydrofolate reductase (DHFR) enzyme, which plays a key role in intracellular folate metabolism and is essential for DNA synthesis and cell growth (8). Combination therapies including MTX are currently used to treat several tumor types, including breast, bladder, and head and neck cancers, osteogenic sarcoma, and leukemias. MTX is also notable for its use in inflammatory disease, rheumatoid arthritis, and dermatologic disorders (9). However, the usefulness of treatment with MTX is limited by the development of drug resistance as occurs in most colorectal cancers. Resistance to MTX may be acquired through different ways, including increased expression of the target gene DHFR via gene amplification (10–13), down-regulation of the reduced folate carrier (SLC19A1; refs. 14, 15), inefficient polyglutamylation of MTX because of decreased activity of folylpolyglutamate synthase (16), up-regulation of glutamate hydrolase (17), mutation of DHFR gene resulting in a decreased affinity for MTX (18, 19), or by epigenetic modulation of genes involved in drug resistance (20).
In colon cancer cells, multiple mechanisms operate in MTX resistance. The capacity to develop resistance and the mechanism(s) involved are associated with the genetic features of the tumor, namely, the karyotype and the presence of a microsatellite mutator phenotype (21–23). Our objective was to identify the appropriate chemotherapeutic strategy that may improve the efficacy of the treatment for each case. We have focused in MTX resistance through DHFR gene amplification because it is the most common mechanism of resistance (10) and can be easily detected at molecular level. In a first setting, we have analyzed the structure and dynamics of DHFR amplification in HT29 MTX-treated cells. Afterwards, we have investigated the sensitization to MTX of MTX-resistant cells after drug withdrawal and reexposure to MTX (Fig. 1). These studies shed more light on the mechanisms involved in the generation of drug resistance by gene amplification and indicate that chemoresistance may be spontaneously reversed and partially impaired, opening new perspectives in pharmacologic regimens.
Materials and Methods
Cell Lines and MTX Treatment
HT29 colon cancer cell line was obtained from the American Type Culture Collection. Three HT29 10−6 mol/L MTX-resistant clones (clones 2, 3, and 6) were generated in a previous study (22). Cells were propagated and maintained in T25 flasks (25 cm2) with DMEM-F12 (parental cells) and DMEM (clones) culture medium (Life Technologies Ltd.) supplemented with 10% fetal bovine serum (Life Technologies) at 37°C in an atmosphere of 5% CO2. MTX treatment was done in triplicate experiments by exposure to stepwise increasing concentrations of MTX (Supplementary Table S1).5
Supplementary material for this article is available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).
For sensitization studies, MTX-resistant cells (clone 2) were passaged 35 times in the absence of MTX. Afterwards, clones were obtained by limiting dilution (clones E, M, N, and P), grown independently, and passaged six more times in the absence of the cytotoxic drug. For the second cycle of treatment, 105 cells of each clone and the parental HT29 cells (as a control) were seeded in 25 cm2 T-flasks and a second cycle of MTX treatment was applied. This treatment was done in triplicate as a stepwise treatment with increasing concentrations of MTX from 10−8 mol/L to 10−7 mol/L in identical conditions to that of the first round of treatment.
Fluorescence In situ Hybridization
Chromosomes 8q22qter, 5p12, and 5q13-14 probes used for fluorescence in situ hybridization (FISH; Supplementary Table S2)5 were obtained from RPCI-11 human male bacterial artificial chromosome (BAC) libraries [Wellcome Trust Sanger Institute (Hinxton, United Kingdom) and the Children's Hospital (Oakland, CA)] and BAC clone collections of Invitrogen Life Technologies. FISH was done as described (24). A brief description is provided as Supplementary Data.
DNA and RNA Extraction
DNA was obtained using proteinase K/RNase treatment and phenol-chloroform extraction according to standard procedures (25). Total RNA was obtained by using the guanidinium-isothiocyanate method followed by phenol-chloroform extraction and isopropanol precipitation (26). RNA was reeluted in water, quantified spectrophotometrically, and stored at −80°C.
Real-time PCR Quantification of Gene Copy Number
Gene copy number at the amplicon boundaries was determined by quantitative PCR using a LightCycler (Roche Diagnostics). Primer sequences (Supplementary Table S3)5 were designed using PrimerSelect 3.1 program (DNASTAR, Inc.). PCR was done in duplicate and using 4 ng DNA, 2.5 mmol/L MgCl2, 0.25 μmol/L forward and reverse primer, and the LightCycler Mix (FastStart DNA Master SYBR Green I; Roche Diagnostics) as recommended by the manufacturer. Standard PCR conditions were used (Supplementary Table S3).5 B2M was used as control gene and the copy number was calculated using a standard curve derived from serial dilutions of normal human control DNA (10−2, 10−3, 10−4, 10−5, and 10−6) and normalized as described (27).
Real-time PCR Quantification of Gene Expression
Gene expression of genes DHFR, hMSH3, and RASGRF2 included in the amplicon was determined by quantitative real-time PCR. 18S and cyclophilin 1 (PPIA) were used as control genes. cDNAs were prepared by reverse transcription with Moloney murine leukemia virus (Invitrogen) using random hexamer primer (Amersham) according to standard procedures. Primer sequences (Supplementary Table S4)5 were designed using PrimerSelect 3.1 program. Duplicate PCR was done in a LightCycler as described above and using 62 ng cDNA (except for 18S and PPIA that were 0.2 and 7.8 ng, respectively).
Fine Characterization of the Breakage Site
PCR amplification was done using 20 ng DNA in a reaction volume of 25 μL containing 1 μmol/L of each primer, 1 unit Taq DNA polymerase (Roche Diagnostics), 1 mmol/L of deoxynucleotide triphosphate mix, and the PCR buffer recommended by the Taq manufacturer. Primer sequence and PCR conditions are detailed in Supplementary Table S5.5 PCR products were resolved on a 6% polyacrylamide denaturing gel and silver stained, purified using a Concert Rapid PCR Purification System (Life Technologies), and sequenced using a BigDye version 3.1 sequence kit (Applied Biosystems) according to the manufacturer's instructions. Sequence homologies were searched using the Blat engine.6
Characterization of a Preexisting Amplicon
Previous studies have shown the presence of an amplification at 8q24 involving MYC oncogene in HT29 cells (28). We have confirmed the karyotypic features of this cell line by using conventional comparative genomic hybridization (CGH) and FISH (22) and CGH array (data not shown). A high-level gain affecting a 25-Mb region at 8q23qter was observed (Supplementary Fig. S1; Supplementary Table S6).5 FISH analysis revealed MYC interspersed signals in both arms of a metacentric marker chromosome 8 with four to five copies per arm (Fig. 2A and B). Amplicon repeats seem to have head-tail orientation (Fig. 2A and B). MYC overexpression (as compared with normal colonic mucosa) was detected by real-time PCR (data not shown). Interestingly, the amplicon also included other genes that may be involved in proliferation and tumorigenesis (EIF3S3, RAD21, TAF2, MTBP, MLZE, and WISP1).
Dynamics of Amplicon Formation and Resistance to MTX
The DHFR genomic region was characterized at cytogenetic and molecular level in HT29 cells treated with increasing doses of MTX. To determine the dynamics of the process, the intermediate stages were investigated (Fig. 1). HSRs were the main form of DHFR amplification in response to doses of MTX ranging from 3 × 10−8 mol/L to 10−6 mol/L as illustrated in FISH experiments (Fig. 2C and D). HSRs appeared as different combinations of asymmetric chromosome 5, some of them with more copies of DHFR in one chromatid than in the other (Fig. 2C and D). DMs were only prevalent after cells were maintained in the higher dose of drug for 3 months (Supplementary Table S1).5 On the other hand, one of the clones (clone 1) exhibited DMs containing the DHFR amplicon at relatively low doses of MTX (10−7 mol/L), which extended to near 100% of the cells when 10−6 mol/L MTX doses were reached (Supplementary Table S1).5
At the highest MTX dose (10−6 mol/L), the DHFR gene copy number amplification was 4-fold (clone 4) in the form of HSR, whereas the appearance of DMs resulted in a sudden increase of the DHFR copy number that reached up to 80-fold (clone 1). DHFR gene expression levels paralleled the degree of amplification reaching up to 100-fold in one of the clones (Supplementary Table S1).5
Characterization of 5q14 Amplification and Its Boundaries
Analysis of HT29 MTX-resistant clone 2 with Spectral Genomics BAC array (1-Mb resolution) showed significant gains in two BACs (Supplementary Fig. S2),5 indicating that the size of the amplicon was below 2 Mb. The low-resolution CGH array platform (472-clone genomic array from Hutchinson/MRC Research Center) was not informative at this level of resolution (1 Mb). FISH analysis with a collection of 11 BACs (Supplementary Table S2)5 was done to get a more accurate definition of the involved region. The amplified region was common to all three MTX-resistant clones (2, 3, and 6), spanned ∼1.078 Mb (Supplementary Fig. S2),5 and included BACs CTD-2327L5, RP11-241J12, and RP11-79I10 (Fig. 2E and F). The amplicon included the genes ZFYVE16, DP58, DHFR, MSH3, RASGRF2, CKMT2, CACH-1, ZCCHC9, UNQ9217, and SSBP2 (Supplementary Fig. S2).5 Real-time reverse transcription-PCR showed the concomitant overexpression of DHFR, MSH3, and RASGRF2 genes in all the three clones (Supplementary Table S7).5
A more precise characterization of the amplicon boundaries was done by quantitative real-time PCR of sequences at the amplicon limits (BACs CTD-2327L5 and RP11-79I10 that are amplified and CTC-512J14 and CTD-2113H21 that show no coamplification with the DHFR locus; Supplementary Fig. S2).5 A stepwise targeting strategy was used to set the limits of the amplicon. These corresponded to two regions of 1,130 and 353 bp representing the putative centromeric and telomeric break points, respectively (Supplementary Table S8).5 Analysis of the features of these two genomic regions revealed the presence of different repetitive elements in both. The centromeric side contained an AluSx element flanked by two LINE sequences. In the telomeric side, two inverted Alu repeats of the Jb and Sx families were separated by 75 bp (Fig. 3).
In an attempt to characterize the fusion points of the amplicon, different PCRs were done using primers designed to amplify putative circularized or inverted structures (this is by using one primer on each side of the amplicon but oriented outwards; Supplementary Table S5;5 Fig. 3). Only in one case (P0 PCR; Supplementary Table S5),5 a specific PCR product was obtained allowing us to identify an inverted reorganization at the telomeric breakage site of the amplicon, near the two inverted Alu repeats (Fig. 3). The presence of this reorganized sequence was confirmed with a second set of primers (PCR P4; Supplementary Table S5).5 The resulting fusion point between the two reorganized regions is a duplication of the sequence ATTTTAC present in both regions.
Sensitization to MTX of HT29 MTX-Resistant Clones
Because gene amplification involves reorganization of the affected chromosome, we wondered if this de novo structure will be stable in cells after withdrawal of the drug, or rather, the copy number of the DHFR gene would be restored to normal levels. And if this were the case, would resulting cells be more or less resistant to a second cycle of treatment? As explained above, the different MTX-resistant clones exhibited a high level of DHFR amplification in the form of DM structures (Table 1). Moreover, two cell populations could be observed with regard to chromosome 5: type A, containing one normal and two reorganized chromosome 5 (Fig. 4A), and type B, displaying two normal and one reorganized chromosome 5 (Fig. 4B). Because all MTX-resistant clones exhibited the same amplicon structure, similar levels of amplification, and MTX resistance, the sensitization studies were done in clone 2 cells.
|.||DHFR copy number at the initiation of the second treatment .||DHFR copy number after the second treatment* .||Resistant foci† .|
|Clone E||0.62||Not resistant||0|
|Clone N||1.18||Not resistant||0|
|.||DHFR copy number at the initiation of the second treatment .||DHFR copy number after the second treatment* .||Resistant foci† .|
|Clone E||0.62||Not resistant||0|
|Clone N||1.18||Not resistant||0|
NOTE: Relative values respect peripheral blood DNA (1 corresponds to a diploid DHFR copy number). No correction for the ploidy of the cell has been applied.
Abbreviation: ND, not determined.
Increasing concentrations of MTX up to 10−7 mol/L as described in Materials and Methods.
Number of colonies in a T25 flask seeded with 105 cells after the treatment with 10−7 mol/L MTX.
MTX-resistant clone maintained in MTX.
Single-cell clones were obtained from MTX-resistant cells by limiting dilution, and those exhibiting the type A karyotype were cultured for 41 passages in the absence of MTX. FISH analysis using RP11-241J12 BAC showed that all clones contained a single complete chromosome 5 and had lost all DM chromosomes (Fig. 4C). Quantitative PCR also confirmed that DHFR copy number was restored to levels similar to untreated parental cells (Table 1). To ascertain the sensitivity of these cells to MTX, a single dose (3 × 10−7 mol/L) was administered and all cells died after 2 weeks, indicating a loss of the MTX resistance capacity. Four different clones were submitted to a new MTX treatment that was applied in triplicate experiments and in identical conditions than the first treatment. Two of the clones (E and N) were unable to develop resistance and all cells died after 48 and 28 days, respectively. An average of seven to eight resistant colonies was observed in the other two MTX-treated clones (M and P). Parental HT29 cells submitted to the same treatment presented a higher survival rate with an average of 42 resistant clones per flask (Table 1).
As expected, quantitative PCR analysis of resistant clones confirmed an increased DHFR copy number in response to MTX (Table 1). G-banding and FISH analysis of resistant M and P clones showed a reorganized chromosome 5 [del(5)(q11.2)], one normal chromosome 5 with its corresponding DHFR signal, and DHFR amplification as DMs (Fig. 4D and E) or as a HSR in 14.3% and 21.4% of clones M and P metaphases, respectively (Fig. 4F). HSR mapped in chromosome 5. Moreover, DHFR signals at chromosomes other than chromosome 5 were observed along with chromosome 5 without a DHFR signal, suggesting the reorganization of chromosome 5 (data not shown).
Next, we wondered if the amplicon generated during the second MTX treatment had the same structure that the one produced in the first treatment. Initially, we investigated if clones M and P (second MTX treatment) contained the fusion point present in the amplicon of all clones analyzed in the first MTX treatment. PCR P4 that specifically amplifies the fused sequence (see above) failed to produce a product from clones E and M (Fig. 3), indicating the de novo nature of the amplicon resulting from the second cycle of MTX treatment. These results suggest the involvement of a different mechanism or different sequences in the reorganization process. By quantitative real-time PCR of multiple sequences distributed along 6 Mb around the DHFR gene, we showed that MTX-resistant clones M and P had a larger amplicon than clones generated in the first treatment (Fig. 5). Moreover, unlike resistant clones obtained in the first treatment that exhibited identical amplicon extension, clones M and P displayed different amplicon size (Fig. 5).
Gene amplification as a mechanism of drug resistance is a stepwise selection process in which cells become resistant through repeated cycles of cell death and proliferation accompanied by genomic instabilization and successive cycles of gene copy number gains (29). Most chemotherapeutic failures are due to the development of drug resistance by the tumor cells (3), which usually leads to giving up the chemotherapy regimen. To design alternative strategies, it is necessary to investigate the mechanisms and dynamics of gene amplification, together with its repeatability and reversibility.
DHFR gene amplification occurs through a preferential process in HT29 cells as shown by the common structure of the amplicon generated in independent treatments done with parental or derived clones. Interestingly, besides this common structure at genomic level, the cytogenetic manifestation may be different and HSR and DM may appear in different phases of the treatment. In any case, DMs are associated with maintained resistance to high doses of MTX. Although the concurrence of HSR and DMs was observed in some cells, the alternative prevalence of one or the other form of gene amplification was the most common situation. Our data indicate that the formation of DMs probably occurs following HSR breakdown, but the presence of DMs at intermediate doses of MTX and a low frequency of previous HSRs suggests that DMs do not necessarily represent a later stage of the amplification process.
HSRs formed at the intermediate steps of MTX treatment often seemed associated with different chromosome 5 with asymmetric signals for DHFR at the two chromatids (Fig. 2D). This fact points to an unequal sister chromatid homologue recombination as the initial amplification mechanism (30, 31). It is possible that the inverted Alus found near the telomeric breakage point confer instability to this region, promoting complementary interactions in the same strand and facilitating a hairpin or a secondary structure formation leading to a unequal recombination between chromatids (Fig. 2C and D; refs. 32–34). The small asymmetric intrachromosomal repeats would ultimately split out by recombination mechanisms (homologous recombination or nonhomologous end joining) to smaller extrachromosomal DMs, which suggests the participation of different mechanisms during the amplification process (35, 36). The telomeric boundary of the amplicon involved a direct repeat sequence, which suggests an inverted recombination at the breakage point (36, 37). The instability observed in the 5q14.1 region was independent of the DHFR copy number or the MTX dose, as the proportion of chromosome 5 with asymmetric signal of DHFR with respect to the number of chromosome 5 with normal copies of DHFR was maintained at all stages of the process (data not shown). The complexity of the amplification process is also revealed by the structure of the amplicon at 8q23.3-q24.3, which formation involves secondary rearrangements (this study and ref. 38).
The number of intact chromosome 5 bearing the amplification target gene (DHFR) may play an important role in early stages of drug resistance. As reported in a previous study, SW480 and SK-CO-1 colon cancer cell lines that show amplification capability are unable to survive MTX treatment even at low doses of the cytotoxic drug (22). It is of note that they only contained a single intact chromosome 5, whereas the untreated HT29 cells had three intact copies. It has been hypothesized that at least two homologue chromosomes are needed in the amplification event (4, 7, 22, 35). One chromosome 5 would be reorganized, whereas the other will remain stable (39). Analysis of HT29 MTX-resistant clones 2, 3, 6, 1, and 4 showed that the generation of the amplicon involved the reorganization as marker chromosomes of one or (more often) two of the intact chromosome 5. As expected, MTX-resistant clones deprived of MTX reverted to a low copy number of the DHFR gene and maintained a unique intact chromosome 5. Interestingly, these cells showed a decreased ability to generate resistance in a second cycle of treatment: two clones died and the other two generated a fifth of resistant colonies compared with HT29 parental cells undergoing a first cycle of treatment (Fig. 1). Beyond the decreased efficiency, the structure of the amplicon was also different (Figs. 3 and 5), suggesting that retreated cells cannot generate the amplicon structure borne during the first round of MTX. Alternatively, some of the cells develop resistance through the generation of a different structure but at a lower rate than previously untreated cells.
As a whole, our data show that, besides the complexity of the mechanisms involved in the generation of amplicons, HT29 cells develop resistance to the cytotoxic drug MTX through a consistent process resulting in a common amplicon structure (Fig. 5). Moreover, after passive loss of the amplicon by withdrawal of the drug, cells become MTX sensitive and exhibit a substantial reduction of their capacity to generate de novo resistance in a second cycle of MTX treatment. When they reacquire resistance is through the generation of an amplicon with a different structure, suggesting that the formation of the first amplicon is a “wear and tear” process. Obviously, we cannot make a direct extrapolation of the conclusions of this study to clinical settings. Nevertheless, our results prompt us to postulate that a second round of treatment may be considered even in patients that have developed chemoresistance by gene amplification, provided a recess in the therapeutic regimen is feasible to allow the resensitization of the cells. In vivo investigations are needed to validate such strategies.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant support: Spanish Ministry of Science and Innovation (SAF2008/1409, SAF2006/0351, and Consolider-Ingenio 2010 CSD2006/49). C. Morales was a fellow of Institut d'Investigació Biomèdica de Bellvitge.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Current address for C. Morales: Biotech Research and Innovation Centre, University of Copenhagen, Ole Maaløes Vej 5, DK-2200 Copenhagen, Denmark. Current address for M.J. García: Centro Nacional de Investigaciones Oncológicas, Human Cancer Genetics Programme, C/Melchor Fernández Almagro, 3, E-28029 Madrid, Spain.
We thank Gemma Aiza for technical support.