Background
Cancers frequently develop abnormal numbers of chromosomes and contain chromosomal rearrangements. This genomic instability generates daughter cells that die because of insufficient complements of chromosomes, as well as polyploid cells that acquire mutations favorable for uncontrolled proliferation. Genomic instability is less frequently observed in non-cancerous cells, which have competent surveillance mechanisms to monitor errors in DNA replication and chromosome segregation during mitosis, as well as the machinery to repair such damage. Dysregulation of these two important mechanisms leads to genomic instability, and ultimately to increased mutation rates and acquisition of the multiple mutations that lead to cancer.
Mitotic protein kinases, such as never-in-mitosis A (NIMA) in fungi and NIMA-related kinases (Neks or Nrks) [
1] in mammals, have been implicated in guarding the integrity of the genome. NIMA functions as a protein kinase, regulates G2-M phase progression, increases expression in response to DNA damage, and serves to ensure proper mitotic spindle organization and formation of the nuclear envelope [
2‐
4]. There are 11 known mammalian NEKs. NEK2 is the one best characterized to date. It has been shown to have a role in controlling orderly mitosis and in preventing chromosomal instability [
1,
5,
6]. NEK6 and NEK7 have been implicated in regulating mitotic progression [
7,
8]. Nek8, like NEK1, has been linked genetically with a form of polycystic kidney disease; it localizes to the primary cilium of each cell where it functions to anchor mitotic centrosomes [
9‐
12]. NEK11 has been linked to the CDC25A degradation in response to DNA damage and is a substrate of CHK1 [
13]. Thus, like their lower eukaryotic orthologs, the NEK family of kinases has many members. Each seems to have its unique cellular function, a function required for orderly progression through the cell division cycle.
Recently, we uncovered a role for NEK1 in DNA damage responses [
14,
15]. NEK1 is a dual serine-threonine and tyrosine kinase [
16] and its kinase activity and expression are quickly upregulated in cells treated with IR. Within minutes after exposure to IR or other genotoxic agents, a portion of NEK1 redistributes from the cytoplasm into the nucleus, where it forms discrete nuclear foci at sites of DNA damage. NEK1 colocalizes with γ-H2AX and MDC1/NFBD1, which are among the first responders to IR-induced double strand breaks (DSBs). The importance of NEK1 in the DNA damage signaling pathway was revealed by analyzing cells lacking functional NEK1. These cells fail to activate downstream checkpoint proteins, such as CHK1/CHK2, and fail to arrest at S or G2/M phase to allow for efficient DNA repair [
14,
15]. Consequently,
NEK1-deficient cells develop many more chromosome breaks than wild type cells [
14,
15].
Because
NEK1 mRNA is abundantly expressed in mouse gonads and neurons [
16], early reports suggested that NEK1 protein functions in a direct and unique way in meiosis or in regulating the cell division cycle [
17,
18]. Whether NEK1 plays a role in regulating chromosomal stability is still unknown at this time. Neither is it known whether NEK1 functions as a tumor suppressor like many checkpoint/mitotic kinases (CHK1, Mps1, and BubR1). In this report, we demonstrate that NEK1 is important for genomic and chromosome stability. Cells defective in NEK1 suffer from disordered mitosis, become aneuploid after multiple cell division cycles, and acquire transforming activity. NEK1 also seems to function as a tumor suppressor, since mice heterozygous for a
NEK1/kat2J mutation develop tumors, specifically lymphomas, with a much higher incidence compared to their wild type littermates.
Methods
Cell culture
Primary renal tubular epithelial cells (RTEs) and tail fibroblasts were obtained from
NEK1- mutated kat2J mice and their wild type littermates as previously described [
15], and cultured in the Ham's F-12/DMEM.
Antibodies
Anti-α-tubulin mAb DM1A and rabbit anti-mouse CD3 were purchased from Sigma-Aldrich (St. Louis, MO, USA), rabbit anti-CD45R antibodies from Abcam, Inc. (Cambridge, MA, USA), rabbit anti-p19ARF antibodies from Genetex (Irvine, CA, USA), fluorochrome-conjugated secondary antibodies (Alexa-Fluoro 488 for green, Alexa-Fluoro 594 for red) from Molecular Probes, Inc. (Eugene, OR, USA), and horseradish peroxidase-based secondary antibodies from Vector Technologies (Burlingame, CA, USA).
Immunocytochemistry
Cells grown on coverslips to 60% confluence were fixed in 4% formaldehyde with 0.1% Triton X-100. Fixed cells were permeabilized with 0.05% saponin and blocked with 10% normal goat serum. Primary antibodies were used at a dilution of 1:100 to 1:1,000 (3 to 0.3 μg/mL) in 10% goat serum. Secondary antibodies, including anti-rabbit or anti-mouse IgG-Alexa 594 (red) and anti-rabbit or anti-mouse IgG-Alexa 488 (green; Molecular Probes, Eugene, OR, USA) were used at a dilution of 1:3,000. Cells were mounted in Permafluor (Lipshaw-Immunon, Pittsburgh, PA, USA). Images were captured with a Ziess AxioPlan2 fluorescence microscope and digitally merged where appropriate.
Chromosome spreads
Mouse RTEs in logarithmic growth phase were treated with colchicine (1 μg/ml, from Sigma) for 30 min at 37°C. All cells, including those in the supernatants, were then collected by trypsinization and swollen in 75 mM KCl for 15 min at 37°C. Disbursed cells were then fixed with freshly prepared methanol: acetic acid (3:1). Free chromosomes were dropped onto slides and stained with Giemsa.
Flourescence-activated cell sorting
Monolayers of the same primary RTEs were trypsinized, washed with Ham's F12/DMEM containing serum to inactivate the trypsin, and then washed with PBS. Spleen and lymohoma tissues were disrupted in PBS using a rubber policeman. The tissue homogenates were passed through 70-μm filter to obtain single cell suspensions. Cells were well suspended in 1 ml of PBS before fixation with ethanol to a final concentration of 70%. The fixed cells were washed again with PBS, treated with RNase, and labeled with propidium iodide (1 μg/ml) before sorting by a Becton Dickinson instrument.
Soft agar colony formation assays were performed as previously described [
15]. Equal numbers of cells (1 × 10
5 or 2 × 10
4) from each of the indicated cell types at different passage were seeded in 0.367% agar. After 21 days of incubation at 37°C, colonies containing at least 50 cells were counted and representative colonies were photographed.
In Vivo Tumor Growth
Monolayers of the RTEs at passage 7 were trypsinized and resuspended in 1 ml of PBS at a density of 2.5 × 107/ml. 5 × 106 cells were injected subcutaneously into the flanks of NEK1 +/- kat 2J mice. Tumor formation was observed at intervals starting at 7 days later and tumors were harvested at day 10 or 21 for histological analysis.
NEK1-mutated kat2J mice and genotyping
C57BL/6J-
NEK1kat2J +/- founder mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). Genomic DNA was extracted from 1 × 10
4 cultured cells, tail fragments tissue, or blood, according to a protocol available at
http://www.jax.org/imr/tail_nonorg.html. Details of the genotyping for the kat2J mutation, a single guanine insertion at nucleotide +996 that results in a truncated, unstable protein missing the entire kinase domain [
9], and the single-strand conformational polymorphism (SSCP) analysis have already been described in detail [
14,
15]. We have since modified the published protocol only to eliminate the need to use radioactive isotope in the detection of bands on polyacrylamide gels. Gels were fixed with 10% methanol/10% glacial acetic acid, stained with 4× gel red dye, and visualized and photographed with a UVP image system.
Generation of inducible knockdown and retrovirus expression constructs
An inducible shRNAi construct to knock down NEK1 expression was designed and made as previously described [
19]. Briefly, Nek1 shRNA expression cassette was created in pTER (pTER-NEK1i). Four such cassettes from pTER-NEKi were inserted in tandem into pPUR (pPUR-4xNEK1i). U2-OS cells with inducible NEK1 shRNA [
14] were established by Lipofectamine 2000-mediated transfection of pPUR-4xNEK1i and a TetR-expressing construct, pCDNA6TR, followed by selection with 5.0 μg/ml blasticidin and 5.0 μg/ml puromycin. N27 was isolated as an inducible NEK1 knockdown clone initially, but later became a line with constitutive knockdown. Passage number 1 was defined as the cells when they were first established to knock down NEK1 expression stably. For the retrovirus expressing wild-type NEK1, GFP-tagged NEK1 was subcloned into retroviral vector pQUXIP, with a modification to replace the CMV-IE promoter with a UBC promoter. Retrovirus was produced by co-transfection of pQUXIP-GFP-Nek1 and VSV-G into 293-GP2 cells.
Histology and immunohistochemistry
Different tissues were harvested immediately after mice were euthanized, and fixed overnight in 10% neutral buffered formalin at 4°C. After progressive dehydration and embedding in paraffin, 3-μm sections were stained with Meyer's hematoxylin and eosin reagents. For immunohistochemical staining, 4-μm tissue sections on slides were deparaffinized with Histoclear (National Diagnostics, Atlanta, GA, USA) and rehydrated with graded ethanol. Primary antibodies including anti-CD3 and anti-CD45R were used to stain for markers of T or B lymphocytes. Biotinylated secondary, anti-mouse and anti-rabbit IgG antibodies and immuno-peroxidase-based ABC development kits were purchased from Vector Laboratories (Burlingame, CA, USA). Immunoperoxidase-stained sections were then counterstained with methyl green to identify nuclei.
Discussion
Our earlier studies have suggested a role for NEK1 in early DNA damage response and cell cycle checkpoint activation [
14,
15]. The data presented here demonstrate that cells deficient in NEK1 develop chromosomal abnormalities during mitosis, and become aneuploid after several division cycles. These abnormal cells gain the ability to overcome density-dependent growth inhibition, to grow in anchorage independent conditions, and to form tumors in mice, i.e., they become malignant. The studies in cultured cells were designed to represent an accelerated version, telescoped in time, of what happens to NEK1-deficient cells in vivo as they're exposed to oxidative stresses and other DNA-damaging injuries. Interestingly, lymphoid tumor cells from older
NEK1 +/- mice had distinct, polyploid DNA contents very similar to those seen in
NEK1 -/- cells passed several times in culture. Experiments using a complementary technique, FACS analysis in human U2OS tumor cells with stable knock-down of NEK1 expression, showed that NEK1 deficiency, not only the specific
NEK1/kat2J mutation, leads directly to acquisition of the polyploid phenotype (additional file
2 Fig. S2B).
In kat2J mice with only one normal allele of
NEK1, we observed an important consequence of defective DNA damage repair and chromosome instability: a very high incidence of tumors, in particular lymphomas. The cumulative incidence of lymphomas in inbred C57Bl/6 mice is reported to be 30% [
24]. We observed the same incidence in our wild type mice, but a 3-fold greater incidence in identically maintained
NEK1 +/- littermates. The lymphomas were heterogeneous, deriving from B-, T-, and unclassifiable lymphocytes, suggesting defects in repair of damaged DNA and in mitotic segregation in multiple lineages and maturation steps, not in any one particular lymphocyte differentiation pathway.
NEK1 null mice also developed lymphomas, but most of them died at early ages from other causes. Lymphomas or lymphoproferative disorders are common tumors in mammals with DNA damage sensing or repair defects [
25‐
31]. Clones of lymphoid cells have to divide multiple times, and they undergo frequent gene rearrangements as they respond to different antigen epitopes throughout the life of an animal. Many lymphocytes should also undergo apoptosis as part of normal process of clonal deletion. Lymphoid tumors therefore are especially prone to acquire mutations, and to proliferate and become malignant if they escape normal cell cycle checkpoints, DNA damage repair mechanisms, or pathways to appropriate programmed death.
The data we present here and in previous reports [
14,
15] concerning NEK1's role in DNA damage and mitotic checkpoint control is similar to what is known about fungal NIMA.
Aspergillus Nima mutants never enter mitosis. Instead, they arrest in late G2 phase with duplicated spindle pole bodies, the fungal equivalents of mammalian cell centrosomes [
32]. Overexpression of NIMA in
Aspergillus, yeast, and even in human cells, causes premature condensation of chromosomes independent of CDC2, the proximal cyclin dependent kinase controlling access to M phase [
2,
33‐
35]. If
nimA5 Aspergillus mutants are forced to bypass G2/M phase checkpoint arrest by acquisition of an additional mutation in bimE7, an anaphase-promoting complex subunit APC1 [
36], they develop grossly abnormal mitotic phenotypes: multiple spindle pole bodies, spindles with disparate sizes and shapes, disorganized microtubules arranged in multiple directions other than orthogonal ones, and defects in nuclear envelope structure and in nucleokinesis [
4]. We observed similarly aberrant mitoses in murine
NEK1 -/- cells after only a few division cycles in culture. Our observations are intriguing, for in mammalian cells NEK1 deficiency alone results in mitotic segregation errors and aneuploidy, whereas bypassing the
nimA5 mutation in
Aspergillus requires an additional mutation in
bimE7 to get the cells past a mitotic checkpoint.
Our data suggests that NEK1 is not essential for entry into mitosis, but instead that it is important for regulating the timing and fidelity of chromosome segregation via its role in centrosomes and in generating bipolar, orthogonal, mitotic spindles. Our studies highlight the multifaceted nature of NEK1 function in DNA damage checkpoint control and centrosomal function. Whether the phenotypes observed in
NEK1 -/- cells are due to centrosome maintenance directly, to checkpoint activation, or both will require further investigation. NEK1 does localize to centrosomes, where it is thought to affect the stability of primary cilia and centrosomes directly, as well as cell cycle progression indirectly through signaling and transcriptional programs [
12,
21]. Different regions within NEK1's kinase and coiled-coil domains have been shown using overexpressed, GFP-tagged, deletion constructs to affect localization of NEK1 to primary cilia and to γ-tubulin in centrosomes [
12,
21]. These experiments were performed in murine renal inner medullary collecting duct (IMCD3) cells, which express normal amounts of endogenous NEK1, so they were not be able to address differential effects of the various NEK1 deletion mutants on chromosomal maintenance. To address this issue in light of our findings of chromosome instability in NEK1-deficient cells, identification and refinement of mutants that differentially affect only NEK1 in centrosome function, in checkpoint activation, or in the DNA damage repair pathway will need to be made. These mutants would then need to be added back to
NEK1 -/- cells in order to dissect the specific, differential functions of NEK1 in chromosomal and checkpoint control phenotypes. We have started such experiments.
Cells that divide without functional NEK1 develop major defects in mitotic spindle function, thus compromising the cells' ability to segregate chromosomes faithfully to the two daughter cells, and resulting in chromosome rearrangements and aneuploidy [
37,
38].
NEK1 -/- cells frequently have aberrant mitotic spindles with lagging and misaligned chromosomes (Figure
2). Such gross chromosomal rearrangements would result in aneuploid cells that acquire copies of oncogenes or lose tumor suppressors. Only a small subset of these cells would gain a growth advantage and continue to proliferate, whereas the majority of such cells, especially those with too few chromosomes, would die by apoptosis. Polyploid cells gaining growth advantages eventually would become the survivors, and they would proliferate without proper regulation, as malignant cells do. Polyploid
NEK1 -/- cells are particularly remarkable in that they do not always have integer multiples of n chromosomes. Instead, they often have pieces of chromosomes and DNA content between 2n and 4n, or between 4n and 6n (Figure
3), as well as micronuclei, multiple nuclei of different sizes, or hollow nuclei in interphase cells (Figure
1). These unique phenotypes suggest that NEK1 deficiency affects not only cytokinesis, but also affects sister chromatid pairing and chromosomal rearrangements like translocations, losses, and losses with reduplications.
An argument can be made that the aneuploid phenotype observed in
NEK1 -/- cells could be a secondary effect, due to additional mutation in another locus. Addressing this issue has been challenging, since re-expression of wild type NEK1 in
NEK1 -/- cells induces checkpoint activation, such that those cells don't continue to proliferate (additional file
2-Fig. S2A) and such that mitotic cells are not observed. Differentiating direct from secondary events might require in vivo experiments, with transgenic mice overexpressing NEK1 crossed into the
NEK1/kat2J -/- background, where the mice could be observed for rescue of tumor formation and genomic instability phenotypes.
A growing body of evidence from mouse models has linked abnormal DNA damage repair and disturbed mitotic events to the genesis and progression of cancer (for review see [
39]). NEK1 does not appear to be absolutely required for development in mice, which seem not to acquire prominent aneuploid phenotypes during embryonic stages. Excessive apoptosis is evident in the kidneys and other organs of embryonic and newborn
NEK1/kat2J -/- mice, however (manuscript in preparation). This observation suggests that cells with abnormal chromosomes content are eliminated during development. NEK1's role in DNA damage responses may be subtle and regulatory, akin to the role associated with the key DNA damage response kinase ataxia telangiectasia mutated (ATM) [
40]. Nek1-null mice seem to be in important ways similar to Atm-deficient mice and humans with ataxia-telangiectasia, which survive embryonic and early adult stages, but which age prematurely and develop lymphomas and other tumors later in life as they're exposed to environmental insults [
26]. The ATM and Rad3-related kinase (ATR), in contrast to ATM and NEK1, is more fundamental in signaling pathways required for recognition and repair of DNA replication intermediates, and results in early embryonic lethality with fragmented chromosomes when mutated [
41]. We suggest that repeated injuries over time may need to accumulate in order to manifest gross chromosomal abnormalities and cancers late in the life of a NEK1-deficient animal, as in the life of an ATM-deficient animal. Such injuries would not occur much during embryonic development.
Our current results showing that mice heterozygous for NEK1 have a high cumulative incidence of lymphomas, derived from all types of lymphocytes, suggests that low level expression of NEK1 in cells expressing it from a single allele is not sufficient to safeguard the genome and prevent chromosome instability. Since NEK1 is important for DNA damage response/repair and centrosome maintenance, the expression of sufficient amounts of NEK1 might be required for proper mitotic checkpoint activation and for assuring precise mitotic chromosomal segregation and cellular cytokinesis. Studying the level of NEK1 expression in different human cancers will help to determine whether chromosome instability observed in these cancers can be attributed to loss of NEK1 activity, and whether NEK1 could be an important target for cancer treatment. We know of no published studies to date that have implicated NEK1 mutations in the pathogenesis of human tumors. Further studies on whether diminished NEK1 expression leads to tumor formation in humans should be explored.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
YC conceived, designed and coordinated the study, acquired, analyzed, and interpreted data, and helped draft the manuscript. CFC performed and acquired the FACS analysis data. HCC and MP performed or helped interpret the histological sections. RP performed the experiments in cultured cells. RLW helped draft the manuscript and participated in discussions. RAE and DEH participated in interpretation of the analysis of histological sections. PLC participated in the interpretation the data, discussion, and drafting of the manuscript. DJR conceived, designed, acquired, analyzed, and interpreted data, and drafted the manuscript. All authors read and approved the final manuscript.