Introduction
In eukaryotes, the initiation of DNA replication involves the formation and activation of the prereplication complex (pre-RC) at the origins of replication (ORIs). The pre-RCs are formed by the sequential binding of the origin recognition complex (ORC1 to ORC6), cell division cycle 6 (Cdc6), Cdt1 and minichromosome maintenance (MCM2 to MCM7) proteins to DNA [
1]. Since loading of the MCM complex onto ORIs is the rate-limiting step in DNA replication, its recruitment to ORIs is inhibited by geminin, the only known endogenous inhibitor of DNA replication. Thus, geminin level and/or activity seem to control the assembly of pre-RCs at ORIs and to determine whether the origins are licensed [
2‐
7].
Geminin, a multifunctional small protein (about 30 kDa), was first identified in a screen for proteins degraded during mitosis using
Xenopus egg extracts [
8‐
11]. Since then, however, roles for geminin during mitosis have been described [
12‐
20], arguing against its mitotic degradation, at least in mammalian cells. More precisely, geminin silencing in human mammary epithelial (HME) cells [
12] or mouse embryos [
14], while showing minimal effect on S-phase progression, completely blocked the progress through mitosis [
12]. The HME mitosis-arrested cells (due to geminin silencing) showed increased expression and activity of cyclin B1, checkpoint protein 1 (Chk1), and Cdc7 [
12]. Surprisingly, only Cdc7 cosilencing triggered apoptosis in geminin-silenced cells [
12], implying that Cdc7 is the kinase that maintains the cytokinetic checkpoint induced by geminin silencing in HME cells [
12].
The Cdc7-Dbf4 complex is essential for ORI firing and maintenance of replication forks [
21‐
26]. Cdc7 inactivation in cancer cell lines causes growth arrest and cell death, while only arresting growth in normal cells [
27]. Although the mechanism of cancer-specific cell death is not yet defined, it is possible that insufficient levels of Cdc7 during cell division may result in stalled and incomplete replication forks, induction of genetic instability and cell death by entering aberrant mitosis in a p53-independent manner [
28‐
30].
Topoisomerases (Topo) are multifunctional enzymes that resolve topological chromosomal complexities, such as knots, tangles and catenanes, arising during DNA metabolism [
31]. Yeasts and
Drosophila cells contain a single type II Topo (TopoII), whereas mammalian cells possess two TopoII isoforms, α and β (TopoIIα and TopoIIβ). Both enzymes can facilitate transcription and replication of chromatin templates [
32,
33]. However, only TopoIIα is absolutely required for DNA decatenation and chromatid separation during anaphase [
34,
35]. During DNA decatenation, TopoIIα dimer binds a DNA helix and hydrolyzes adenosine triphosphate (ATP) to introduce a transient double-stranded break (DSB) through which it passes the other entangled intact helix. Then the DNA DSB is religated, and TopoIIα dissociates from the DNA [
32‐
35]. Furthermore, TopoIIα binding to chromosomes and its decatenation activity are modified by phosphorylation and SUMOylation [
36‐
39]. For example, casein kinase Iε (CKIε) phosphorylates TopoIIα on serine 1106 (S1106) in G
2/M cells and induces the TopoIIα chromosome localization and decatenation function as well as sensitivity to TopoIIα-targeting drugs [
40‐
42]. Moreover, the complex RAN binding protein 2/ubiquitin-conjugating enzyme 9 (RanBP2/Ubc9) SUMOylates TopoIIα and triggers its chromosome translocation and decatenation activity [
39].
TopoIIα's ability to cleave DNA in a reversible manner makes it an ideal target for agents such as doxorubicin and etoposide, which poison the enzyme via the trapping of the transient reaction intermediate composed of TopoIIα bound covalently to the 5′ end of the cleaved DNA strands (cleavable complexes), preventing religation of DNA [
43]. It thus induces DNA damage, genomic instability and cell death [
42,
44]. However, development of resistance to these agents limits their therapeutic use [
45]. Therefore, an understanding of the mechanisms that lead to the development of this resistance is essential to the improvement of the therapeutic potential of these agents.
In the present study, we show that geminin silencing induces chromosome bridge formation by inhibiting TopoIIα chromosome localization and function. Cdc7 cosilencing or CKIε overexpression in geminin-silenced cells restored TopoIIα chromosomal localization and prevented the formation of chromosome bridges. This finding suggests that CKIε is a positive regulator and Cdc7 is a negative regulator of TopoIIα chromosomal localization and function. However, these cells underwent apoptotic cell death, suggesting that they were unprepared to enter G1. Moreover, geminin and TopoIIα interact on chromosomes in G2/M/early G1 cells, and geminin overexpression prematurely releases TopoIIα from chromosomes, in part by enhancing TopoIIα deSUMOylation on chromosomes. Geminin overexpression also inhibits DNA decatenation before the religation step, leading to linearization of model entangled DNA in vitro and chromosome breakage and aneuploidy in vivo. These effects were accompanied by decreased cytotoxicity to TopoIIα inhibitors. Importantly, Cdc7 co-overexpression corrected both defects. These data represent a potential mechanism for TopoIIα drug resistance and suggest that inhibiting the activity of geminin and TopoIIα, CKIε and/or Cdc7 can be more beneficial for breast cancer patients with aggressive, drug-resistant disease.
Materials and methods
Cell culture and drug treatments
All cells were cultivated in RPMI 1640 Medium (Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (FBS) (Gemini Laboratories, Inc, West Sacramento, CA, USA) at 37°C in a 10% CO2-containing atmosphere unless otherwise mentioned, except HME cells that were maintained in growth factor-supplemented Dulbecco's modified Eagle's medium/Ham's F-12 mammary epithelium basal medium (MEBM) (Clonetics/Cambrex, Walkersville, MD, USA). For fluorescence-activated cell sorting (FACS) analysis, treated cells were fixed in 100% ethanol, stained with 2.5 μg/mL propidium iodide (PI) (Sigma, St. Louis, MO, USA), supplemented with RNase A and incubated at 37°C for one or two hours. A HME cell line that carries a pBOS-H2B plasmid (Clontech Laboratories, Mountain View, CA, USA) was generated by standard plasmid transfection, and clones were selected with blastocidin (Sigma). Etoposide, doxorubicin and IC261 were obtained from Sigma, ICRF187 and PHA767491 were purchased from Tocris Bioscienc (Ellisville, Missouri, USA) and ICRF193 was obtained from Funakoshi (Tokyo, Japan). All drugs were dissolved in dimethyl sulfoxide (DMSO).
Geminin cloning and bacterial expression
The protocol described by Nakuci
et al. [
12] was used. In brief, geminin full-length cDNA was amplified from IMR90 total RNA using the following primers cut with
BamHI/
SalI and ligated to the glutathione
S-transferase (GST) vector pGEX-4T2 cut with the same enzymes: forward 5′-CGGGATCCATGAATCCCAGTATGAAGCAGAAACAAGAA-3′ and reverse 5′-ACGCGTCGACTCATATACATGGCTTTGCATCCGTA-3′. The GST-fused geminin was expressed in competent bacteria One Shot BL21 Star (DE3)pLysS (Invitrogen, Carlsbad, CA, USA), induced with isopropyl-β-D-thiogalactoside and purified on Glutathione Sepharose™ 4B beads (GE healthcare, Piscataway, NJ, USA) and eluted from the beads using 10 mM glutathione in 50 mM Tris HCl, pH 8.0. Using a similar strategy, geminin full-length cDNA was also ligated to the retrovirus plasmid Rev-Tre (Clontech), and the retrovirus was prepared and used to infect the HME cell line expressing the inducer pTet-ON (Clontech). Geminin clones were generated by appropriate selection.
Antibodies
Mouse anti-geminin antibody (Ab) generation was described earlier [
12]. We used mAbs α-cyclin A1, anti-cyclin E, anti-cyclin B1 and anti-CKIε (610445; BD Transduction Laboratories (San Jose, CA, USA); mAb α-actin (Ab-1; Oncogene Science, Cambridge, MA, USA); rAb α-Cdk2 (Pharmingen, San Jose, CA, USA); rAb α-pChk1 (Cell Signaling Technology, Danvers, MA, USA); mAb α-cdc7 (MS-1888-P; NeoMarkers, Fremont, CA, USA); mAb α-cdc2 (B-6), anti-Chk1 (G-4), rAb α-Sp1 (H-225), anti-geminin (FL- 209) and gAb α-lamin B (C-20, sc-6216) (Santa Cruz Biotechnology, Santa Cruz, CA, USA); rAb α-H2B (ab18977) and mAb α-TopoIIα (ab52934) (Abcam, Cambridge, MA, USA); and rAb α-CKII (Millipore, Danvers, Massachusetts:, USA).
Cell synchronization and small interfering RNA transfection
HME cells were incubated in growth factor-free medium for 72 hours to produce cells in G
0/G
1 phase (> 95%) [
12]. G
0/G
1 cells were then released from arrest in medium containing growth factors, and 16 hours (S phase), 22 hours (G
2/M phase) or 24 hours (M/G
1 phase) later cells were collected and analyzed. HME cell synchronization and transfection were performed as described by ElShamy and Livingston [
46]. In brief, cells were transfected (0 hours) in serum-free medium with a relevant double-stranded RNA interference reagent by a standard method using oligofectamine. At 24 hours, the medium was changed, and growth factor-containing MEBM (Clonetics/Cambrex) was added. Small interfering RNA (siRNA) used were siGem: TGAGCTGTCCGCAGGCTTT, scrambled siGem: TGATTTGTCCGCAGCTGGC, siCdc7: TTTGTGAACACCTTTCCTGTT and siTopoIIα: sc-36695 (Santa Cruz Biotechnology). The silenced luciferase (siLuc) and silenced green fluorescence protein (siGFP) used were from previously published data.
Kinase assay
Cells were collected by trypsinization and washed twice with phosphate-buffered saline (PBS). Whole cell extract was prepared by rocking cells in EBC buffer (50 mM Tris HCl, pH 8.0, 0.5% Nonidet P-40 (NP-40) and 120 mM NaCl) at 4°C for 30 minutes and centrifuged at high speed for 15 minutes. Protein A beads were added to the supernatant along with antibodies (anti-Sp1, anti-CKIε or anti-Cdc7) for 2 to 2.5 hours. Beads were washed once with NETN lysis buffer containing 250 mM NaCl, twice with NETN containing 150 mM NaCl and once with kinase buffer (50 mM Tris HCl, pH 7.5, 10 mM MgCl2 and 1 mM dithiothreitol (DTT), added fresh). Twenty microliters of kinase buffer were added to the beads, along with 7 μL of ATP mix (5 μL from 40 μM cold ATP + 2 μL of radioactive ATP (20 μCi)) and 10 or 100 ng of purified TopoIIα (TopoGen, Columbus, OH, USA). The reaction was rocked at room temperature for 45 minutes and then stopped by adding 20 μL of sodium dodecyl sulfate (SDS) loading buffer and boiling for 10 minutes.
Cell sonication, chromatin purification, Western blot analysis and immunoprecipitation
The protocols used by Nakuci
et al. [
12] and ElShamy and Livingston [
46] were used to isolate total extracts by sonication and chromatin preparations. Briefly, cells at about 75% confluence were washed several times with PBS and trypsinized. After being washed, 1 × 10
7 cells were resuspended in 1 mL of Buffer A (110 mM KC
2H
3O
2, 15 mM NaC
2H
3O
2, 2 mM MgC
2H
3O
2, 0.5 mM ethylene glycol tetraacetic acid and 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), pH 7.3). Next, we added 2 mM DTT and 50 μg/mL digotinin to the cell suspension. The cells were agitated at 4°C for 10 minutes. Nuclei were pelleted by centrifugation in a swinging bucket rotor at 1,500 ×
g for 10 minutes. They were resuspended in hypotonic Buffer B (1 mM HEPES, pH 7.5, and 0.5 mM ethylenediaminetetraacetic acid (EDTA) supplemented with 0.5% NP-40). Typically, a nuclear pellet of about 50 μL was resuspended in 0.5 mL of Buffer B. The nuclear suspension was then agitated at 4°C for 15 minutes and layered on top of a 10-mL sucrose cushion (100 mM sucrose, 0.5 mM Tris HCl, pH 8.5), then centrifuged at 3,500 ×
g for 15 minutes at 4°C. The chromatin pallet was suspended in 0.25 mM EDTA, pH 8.0, and sonicated three times for 10 seconds each, each time using a Fisher Scientific Model 100 Sonic Dimembrator (Fisher Scientific, Pittsburgh, PA, USA). After sonication, the chromatin suspension was centrifuged twice at high speed for 10 minutes at 4°C, and the supernatants were retained. This chromatin extract was first precleared by agitation for 2 hours at 4°C in the presence of 50 μg of protein A/G Sepharose beads, followed by pelleting of the beads. The supernatant protein concentration was measured, and 500 μg of chromatin protein were routinely immunoprecipitated using 1 or 2 μg of Ab and 50 μL of protein A/G Sepharose beads in a total volume of 1 mL of NETN buffer (in which the NaCl concentration was preset at 250 to 500 mM). In some experiments, the deSUMOylation inhibitor
N-ethylmaleimide (10 nM) was added to sonicates.
Immunofluorescence analysis
Cells were seeded on slide chambers (LabTek, Rochester, NY, USA) at 25% confluence 24 hours prior to processing. Cells were fixed with 4% paraformaldehyde for 10 minutes at room temperature, permeabilized in Triton X-100 buffer (0.5% Triton X-100 in 20 mM HEPES, pH 7.4, 50 mM NaCl, 3 mM Mg2Cl and 300 mM sucrose containing 0.5% bovine serum albumin (BSA)) for 10 minutes at 4°C. Cells were then incubated for 30 minutes with 5% normal mouse or rabbit serum (MS or RS) in PBS and then incubated for 30 minutes at 37°C with primary antibody. Cells were then incubated with appropriate fluorescein isothiocyanate- or rhodamine-conjugated secondary antibodies diluted 1:5,000 to 1:10,000 in 5% MS or RS in PBS for 30 minutes at 37°C. Coverslips were then mounted in anti-fade solution (Vector Laboratories, Burlingame, CA, USA) supplemented with 4′6-diamidino-2-phenylindole (DAPI).
Comet assay
A neutral comet assay was performed to detect DSBs. After induction of geminin by 2 μg/mL doxycycline (Dox) for 72 hours, cells embedded in agarose were lysed and subjected to electrophoresis as described previously [
47]. Individual cells stained with 0.5 μg/mL DAPI were viewed using an ultraviolet (UV) light fluorescence microscope (Olympus, San-Diego, CA, USA). Quantification was achieved by analyzing
x randomly selected comets per slide with Komet 5.5 software (Kinetic Imaging, Bath, UK) using the variable Olive Tail Moment (with results measured in arbitrary units, defined as the product of the percentage of DNA in the tail multiplied by the tail length).
Trapped in agarose immunostaining assay
A 50-μL quantity of cell suspension medium warmed to 37°C was mixed with an equal volume of agarose solution (2% (wt/vol) in PBS, SeaPrep Agarose ultralow gelling; FMC BioProducts, Rockland, ME, USA), which had been melted and kept at 37°C. The mixture was immediately spread evenly across a microscope slide and quickly gelled by placing the slides onto a cold surface (0°C). Slides were lysed for 15 minutes at 20°C in a buffer containing 1% (wt/vol) SDS, 80 mM phosphate buffer, pH 6.8, 10 mM EDTA and a protease inhibitor mixture (final concentrations 2 μg/mL pepstatin A, 2 μg/mL leupeptin, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine and 1 mM DTT). Slides were next immersed in 1 M NaCl supplemented with the protease inhibitor mixture for 30 minutes at 20°C, then washed by immersion three times (5 minutes per wash) in PBS. Immunofluorescence was performed according to a protocol described earlier [
12]. Slides were counterstained with Hoechst 33258 blue (10 μM in PBS; Sigma) for 5 minutes before application of coverslips that were secured with a sealant.
TopoGen decatenation assay
TopoIIα enzymatic activity was assayed by measuring the decatenation of kinetoplast (k)-DNA (TopoGen). A standard assay carried out in a total volume of 20 μL included 50 mM Tris HCl, pH 7.9, 88 mM KCl, 10 mM MgCl2, 0.5 mM EDTA, 10 mM ATP, 10 mM DTT, 100 μg/mL BSA and 300 ng of k-DNA. The reaction mixture containing TopoIIα immunoprecipitated from control or siRNA-treated cells was incubated at 37°C, and the reaction was stopped by the addition of 5 μL of stop solution (5% SDS, 25% Ficoll and 0.05% bromophenol blue). The samples were resolved by electrophoresis at 115 V using a 1% agarose gel in Tris-acetate-EDTA buffer with 0.5 μg/mL ethidium bromide and photographed under UV illumination.
Relaxation of pBR322plasmid negative supercoiled assay
The reactions were carried out by incubating 150 ng of supercoiled pBR322 plasmid DNA at 37°C in 15 μL of reaction buffer (10 mM Tris HCl, pH 7.4, 5 mM MgCl2, 100 mM KCl and 0.5 mM ATP) and were initiated by the addition of TopoII (1 or 2 U; TopoGen) and different concentrations of GST-geminin as indicated or 100 ng of GST alone (control). Reactions were stopped by the addition of 5 μL of loading buffer, and samples were electrophoresed in 1% agarose gel in Tris-Borate-EDTA (TBE) buffer (pH 8.0) at 10 V/cm for 4 hours. The gel was then stained with TBE containing ethidium bromide (0.5 μg/mL) for 10 minutes, washed extensively and photographed.
Colcemid (100 ng/mL) was added directly to the culture dish and swirled and incubated for 1 hour. Following incubation, cells were trypsinized and washed with PBS. After the cells were washed, all excess PBS was removed and cells were gently resuspended in the residual PBS. KCl (0.075 M) was added slowly dropwise to a quantity of 10 mL to the cells resuspended in PBS. The reaction was incubated at 37°C (in a water bath) for 5 to 10 minutes. The reaction was centrifuged at 900 rpm for 5 minutes, followed by removal of as much KCl as possible, and then the cells were gently resuspended in the residual PBS. Five milliliters of freshly prepared fixative (3:1 methanol to acetic acid) were then added dropwise to the cells and carefully mixed. After centrifugation of the reaction at 900 rpm for 5 minutes and removal of the fixative solution, the whole step was repeated with 2 mL of fixative. Finally, after removing all but 300 μL of the fixative, the cell mixture was dropped from about 18 inches onto an angled, humidified microscope slide and air-dried for at least 10 minutes. Next, PI or Giemsa stain was used to stain the chromosome spread.
MTT and activated caspases 3 and 7 assays
MTT and activated caspases 3 and 7 assays were done using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (G3580; Promega, Madison, WI, USA) or the Caspase-Glo 3/7 Assay (G8091; Promega), respectively, according to the manufacturer's instructions. Measurements were obtained using optical density at 490 nm. Each experiment was done in eight samples, and the whole experiment was repeated three times.
Cell cycle analysis
Cell cycle analysis was carried out by flow cytometery after PI staining using a standard protocol.
Statistical analysis
Comparisons of treatment outcomes were tested for statistically significant differences using Student's t-test for paired data. Statistical significance was assumed at *P ≤ 0.05, **P ≤ 0.01 and ***P ≤ 0.001.
Discussion
Chromosome decatenation and/or segregation and cell division are coordinated in the cell cycle of all organisms, from bacteria to humans. In human cells, TopoIIα is involved in chromosome decatenation, condensation and segregation [
48]. Geminin's binding to TopoIIα on mitotic chromosomes and enhancing of its decatenation activity clearly show that geminin's physical and functional interaction with TopoIIα is essential to coordinate chromosome decatenation and/or segregation with cell division. Considering geminin's role in DNA replication, it is possible to suggest that geminin stimulates TopoIIα interaction and helps disentangle the freshly replicated DNA. The negative supercoiling (unwinding) generated at the initiation of replication at ORIs [
56,
57] and the positive supercoiling (overwinding) generated ahead of the replication fork during replication elongation [
58‐
61] must be resolved to facilitate strand separation. It is possible that through the interaction of geminin and TopoIIα, geminin loads onto or stabilizes TopoIIα on chromosomes and thus increases the level of DNA-bound TopoIIα and the effective rate of decatenation and relaxation of the newly made sister duplexes [
62,
63].
Since chromosome condensation, alignment at the metaphase plate and movement toward the poles [
64] occurred relatively normally in geminin-silenced cells (Figure
1 and Additional file
1), the spindle and associated molecular motors must function correctly in the absence of geminin. Likewise, the normal attachment of the chromosomes to the metaphase plate and lack of checkpoint activation that monitors spindle tension [
65] in geminin-silenced cells suggest that kinetochores are also unaffected by geminin silencing. It is thus possible to propose that the primary function for the geminin-TopoIIα complex is to resolve chromosome complexities and that, in the absence of geminin, an increase in the number and complexity of knotted replication bubbles would increase the number of nodes in the catenane [
56] that arrest segregation of the freshly replicated DNA molecules [
56], leading to chromosome bridges and mitotic arrest [
12,
66]. Our present study supports this proposition. Several proteins with wide varieties of functions (such as the bacterial condensin-like protein MukB [
67], the
Drosophila condensin protein Barren [
68] and the bacterial SeqA protein that prevents overinitiation of chromosome replication [
69]) have been shown to function in a similar manner in which they interact and/or stimulate the relaxation and decatenation activities of TopoIIα (TopoIV in bacteria). Similar to geminin silencing, mutations in these genes prevent effective separation of sister chromatids during anaphase because of the suppression of TopoIIα function. In future studies, it will be important to search for other components in this geminin-TopoIIα complex that regulate chromosome decatenation. This work will be necessary to better understand the molecular mechanism of proper decatenation/segregation during mitosis.
One of the interesting and unexpected aspects of the present study is the fact that GST-geminin could induce linearization of
k-DNA as well as
pBP322 plasmid. As mentioned above, we cannot rule out bacterial nuclease contaminant in the GST-geminin preparation as the source of the apparent DNA linearization in both assays. However, while this might be possible in the
pBP322 reaction, it is hard to imagine that this is the case in the
k-DNA reaction. The
k-DNA used consisted of interlocked minicircles (mostly 2.5 kb) that form extremely large networks of high molecular weight. Unless cut and religated specifically by TopoIIα during the decatenation process, these networks fail to enter the gel. Assuming that a protein other than geminin cleaved the DNA, it must have been pulled down specifically by GST-geminin and not by GST alone. This would make it a partner and not contaminant. However, bacteria do not express geminin. Therefore, a human homolog of this nuclease must exist and would be worth cloning in the future. Alternatively, it is possible that geminin at higher concentrations binds and masks a TopoIIα ligation-inducing domain, if it exits. This could explain the fact that only at much higher concentrations in these assays did geminin prevent religation, but not cleaving activity, of TopoIIα. Finally, it is possible that geminin itself has nuclease activity. Geminin is a coiled-coil protein [
70], and many coiled-coil proteins, such as the Werner syndrome protein WRN [
71], are known to have nuclease activity. In support of the latter assumption, the fact that incubating GST-geminin with anti-geminin antibody before the reaction restored TopoIIα's ability to decatenate the
k-DNA (Figure
4C). At the moment, we are unable to distinguish between these possibilities but have future plans to investigate which is valid.
Phosphorylation of TopoIIα on S1106 is important in TopoIIα translocation to chromosomes, DNA decatenation, formation of drug-stabilized DNA cleavable complex and modulation of drug sensitivity [
40,
41]. CKIε is the only known kinase that targets this site
in vitro and
in vivo [
40,
41]. The facts that CKIε overexpression restored chromosome decatenation and/or segregation that had stalled in the geminin-silenced cells and that geminin overexpression upregulated CKIε expression suggest a positive molecular link by which geminin controls TopoIIα chromosome localization and function. The facts that geminin overexpression decreased Cdc7 expression, that Cdc7 silencing restored stalled chromosome decatenation and/or segregation (that is, chromosome bridges) [
72] in the geminin-silenced cells, that Cdc7 overexpression reduced chromosome breakage and aneuploidy induced by geminin overexpression and that Cdc7 phosphorylated TopoIIα, at least
in vitro, suggest that Cdc7 is a negative molecular link between geminin and TopoIIα chromosome localization and function. It will be important in future studies to investigate whether Cdc7 also phosphorylates TopoIIα
in vivo and on which sites, what are the upstream kinases and/or conditions that activate Cdc7 to phosphorylate TopoIIα and what is their relation to geminin.
TopoIIα SUMOylation is inhibited and/or decreased in geminin-overexpressing cells. It is possible that geminin overexpression prevents TopoIIα SUMOylation by decreasing its binding to the SUMOylating complex RanBP2/Ubc9. Alternatively, it is possible that in normal cells, one function of geminin is to bind and/or recruit the deSUMOylating enzymes SENP1 and SENP2 to chromosomally bound TopoIIα and to facilitate its deSUMOylation and release from chromosomes after chromosome decatenation is completed. In geminin-overexpressing cells, this could be accelerated by the fact that geminin recruits more of the enzymes and/or recruits them earlier to TopoIIα, thus leading to premature deSUMOylation and release of TopoIIα from chromosomes before the ligation step. It is also possible that this is simply the result of a dominant negative effect exerted by overexpressed geminin. Whatever the reason is, this could contribute to the generation of DNA damage and low efficiency of TopoIIα-directed drugs. At present, we are investigating whether SENP1 and/or SENP2 are indeed TopoIIα deSUMOylating enzymes; whether a molecular link between geminin-induced TopoIIα phosphorylation, SUMOylation and deSUMOylation exists; and whether using inhibitors of deSUMOylating enzymes in combination with TopoIIα-directed drugs could be used to treat breast cancers with high geminin levels.
It is intriguing that geminin overexpression suppressed cell death induced by two different types of TopoIIα drugs. It has been proposed that cells with low levels of TopoIIα respond better to the types of drugs that interfere with the catalytic activity of the enzyme (for example, ICRF187 and ICRF193), while cells with high TopoIIα levels are most resistant [
73,
74]. This could explain why, compared to uninduced Gem9 cells, induced Gem9 cells were resistant to these drugs, since geminin overexpression reduced the level of TopoIIα on the chromatin. The other types of TopoIIα drugs (for example, etoposide and doxorubicin) have the potential to induce DNA DSBs by stabilizing TopoIIα on DNA and prevent its religation activity during chromosome decatenation [
73,
74]. It has also been proposed that these drugs induce a DSB for every drug-stabilized TopoIIα enzyme. Thus sensitivity to this type of drugs increases with the level of the chromosome-bound TopoIIα. Low chromosome-bound TopoIIα was detected in cells expressing endogenously (for example, MDAMB231) or exogenously (induced Gem9), so overexpression of geminin could explain resistance to this type of drug as well. This important aspect of our study implies that the efficacy of all types of TopoIIα-directed drugs should increase if combined with geminin inhibitors.
Our previously published results [
12] and those presented in this study are in principle agreement with those published recently by Zhu
et al. [
75]. Those authors claimed that selective killing of cancer cells could be achieved by inhibiting geminin activity. Whereas they claimed that normal cells depleted of geminin continue to proliferate normally [
75], we showed earlier that geminin silencing inhibited progression of immortalized HME cells from the M to G
1 phase with minimal effect on S-phase progression. Furthermore, they proposed that cancer cells depleted of geminin specifically rereplicate their genomes and that their nuclei became giant and underwent apoptosis [
75]. However, we proposed that geminin has a fundamental cytokinetic function, whereas its S phase is redundant. This discrepancy could be due to differences in the cell types and/or the techniques used. Another possible reason for this incongruity is that the system cells used in the Zhu
et al. study continued to express cyclin A [
75], while in the HME cells we observed no cyclin A expression in geminin-silenced cells [
12]. It would be interesting in future studies to determine whether, in breast cancer cell lines, for example, geminin silencing also induces rereplication as reported by Zhu
et al. [
75]. However, we doubt this will be the finding, because in a future publication (unpublished data, W. M. ElShamy) we will show that continuous geminin silencing (with three different small hairpin RNA) inhibits the proliferation of MDAMB231 cells
in vitro as well as tumor formation in a mouse xenograft model. Also in contrast to the data presented by Zhu
et al. [
75], in our work it is geminin overexpression, not its silencing in HME cells, that triggers the formation of cells containing > 4 N DNA content
in vitro. Furthermore, our work shows tetraploid and/or aneuploid karyotyping and giant nuclei both
in vitro and in a mouse xenograft model in geminin overexpressing cells. Given these apparent differences, only our overall conclusions are in accord with those of Zhu
et al. [
75]. We also propose that inhibiting geminin expression and/or activity should selectively kill cancer cells overexpressing geminin (unpublished data, W. M. ElShamy).
Decreased repair of chromosomal DSBs can lead to genome instability, including mutation, translocation and aneuploidy, all of which are hallmarks of many cancers [
76‐
79]. Interestingly, specific Chk1 and H2AX phosphatase upregulation in geminin overexpression led to their inactivation in Gem9 cells (see Results section).
Taken together, our present findings suggest that geminin overexpression induces the formation of aneuploid, aggressive and drug-resistant breast cancer cells (see model in Additional file
7). Geminin silencing prevents decatenation because it blocks TopoIIα's access to chromosomes and its function therein (see model in Additional file
7). Thus, in combination, our data provide an intriguing molecular explanation for the high percentage of patients in whom TopoIIα-directed treatment fails. We propose that etoposide, doxorubicin or any other drugs that target TopoIIα function will be more beneficial when combined with anti-geminin chemotherapeutic agents.
Competing interests
The authors declare that they have no competing interests; however, WME has submitted a patent application for the use of anti-geminin along with TopoIIalpha drugs to treat geminin overexpressing breast tumors.
Authors' contributions
LG, RM, YS and NM performed the experiments. WME designed, performed and interpreted the experiments and wrote the manuscript. LG, RM, YS, NM and WME read and approved the final manuscript.