Genomic instability is an important factor in cancer susceptibility, but a mechanistic understanding of how it arises remains unclear. We examined hypothesized contributions of the replicative DNA polymerase δ (pol δ) subunit POLD4 to the generation of genomic instability in lung cancer. In examinations of 158 lung cancers and 5 mixtures of 10 normal lungs, cell cycle- and checkpoint-related genes generally showed mRNA expression increases in cancer, whereas POLD4 showed reduced mRNA in small cell lung cancer (SCLC). A fraction of non–small cell lung cancer patients also showed low expression comparable with that in SCLC, which was associated with poor prognosis. The lung cancer cell line ACC-LC-48 was found to have low POLD4 expression, with higher histone H3K9 methylation and lower acetylation in the POLD4 promoter, as compared with the A549 cell line with high POLD4 expression. In the absence of POLD4, pol δ exhibited impaired in vitro DNA synthesis activity. Augmenting POLD4 expression in cells where it was attenuated altered the sensitivity to the chemical carcinogen 4-nitroquinoline-1-oxide. Conversely, siRNA-mediated reduction of POLD4 in cells with abundant expression resulted in a cell cycle delay, checkpoint activation, and an elevated frequency of chromosomal gap/break formation. Overexpression of an engineered POLD4 carrying silent mutations at the siRNA target site rescued these phenotypes, firmly establishing the role of POLD4 in these effects. Furthermore, POLD4 overexpression reduced intrinsically high induction of γ-H2AX, a well-accepted marker of double-stranded DNA breaks. Together, our findings suggest that reduced expression of POLD4 plays a role in genomic instability in lung cancer. Cancer Res; 70(21); 8407–16. ©2010 AACR.

Lung cancer has become the leading cause of cancer death in many industrialized countries, with small cell lung cancer (SCLC), a very aggressive subset formed by small cells with scarce cytoplasm, molded nuclei, and neuroendocrine differentiation, accounting for about 15% to 20% of all lung cancer cases (1). Due to its highly proliferative and metastatic potential, as well as nearly certain recurrence after chemotherapy, the 5-year survival rate for SCLC is around 5% (2).

It was recently reported that highly conserved DNA damage response pathways are frequently altered at various phases in human cancers (38). Activation of DNA damage checkpoints, even in preneoplastic lesions, leads to cell cycle blockade or apoptosis (3, 6), and abrogation of that activation is thought to be critical during multistep transformation processes, whereas fully malignant cells of overt cancers frequently carry various defects in checkpoint mechanisms (4, 5, 7, 8).

In lung cancer, it is conceivable that smoking plays a major role in induction of DNA damage that may be detected by checkpoint responses, although the intrinsic driving force behind this related genomic erosion has yet to be identified (1). For many carcinogens, initiation of carcinogenesis requires DNA replication, suggesting that genetic alterations are fixed in the genome during replication of damaged DNA. For this reason, a long-term hypothesis states that errors in DNA replication and deficits in DNA repair account for multiple mutations in cancer (9, 10). Herein, we report evidence showing that polymerase δ (pol δ), which is a core DNA replication and repair protein, is frequently downregulated in the POLD4 subunit in SCLC and a fraction of non–small cell lung cancer (NSCLC).

Cell lines

Cell lines with the prefix ACC-LC- were established in our laboratories. Calu6 and SK-LC-6 cells were generously provided by L.J. Old (Memorial Sloan-Kettering Cancer Center, New York, NY) and PC-10 cells by Y. Hayata (Tokyo Medical University, Tokyo, Japan), whereas the NCI-H460, SK-MES-1, HCT116, and A549 cell lines were purchased from the American Type Culture Collection. Cells were cultured in RPMI 1640 supplemented with 5% fetal bovine serum (Sigma-Aldrich). Plat-E cells were kindly provided by Dr. T. Kitamura (University of Tokyo, Tokyo, Japan). HCT116 and Plat-E cells were cultured in DMEM supplemented with 5% fetal bovine serum.

Antibodies

A polyclonal anti-POLD4 antibody was raised against a purified protein preparation (see Fig. 3B) using a New Zealand White rabbit (Fig. 6A), whereas another polyclonal antibody was raised against a glutathione S-transferase (GST)-tagged POLD4 preparation, followed by removal of the anti-GST antibody (CycLex; Fig. 3D). Other antibodies were anti-POLD4 (POLD4/p12 subunit of pol δ) ascites (2B11, Abnova), POLD1 (p125 subunit of pol δ), Chk1 (Santa Cruz Biotechnology), NBS1-pSer343, Chk2-pTyr68, Chk1-pSer317 (Cell Signaling Technology), NBS1 (Oncogene Research/Calbiochem/Merk), ATM-pSer1981 (Rockland Immunochemicals), ATM (2C1, GeneTex, Inc.), H2AX (BL552), SMC1, SMC1-pSer966 (Bethyl laboratories), FLAG (Sigma-Aldrich), Alexa-Fluoro647–conjugated γ-H2AX, p21, p27 (BD), Chk2 (MBL), anti-trimethyl-histone H3 (Lys9), and anti-acetyl-histone H3 (Millipore/Upstate).

Patients

Prior to obtaining patient samples, requisite approval from the review board of Aichi Cancer Center, Nagoya, Japan, and written informed consent from the patients were obtained.

RNA interference

Transfection was carried out using 50 nmol/L of a siRNA (Supplementary Table S1) duplex (Sigma-Aldrich) targeting each mRNA or negative control 1 (Ambion) with Lipofectamine-2000 (Invitrogen).

Quantitative real-time reverse transcriptase-PCR

Total RNA and cDNA were prepared using an RNeasy Mini Kit (QIAGEN) and SuperScript II Reverse Transcriptase (Invitrogen), respectively. Quantitative reverse transcriptase-PCR (RT-PCR) was performed using a 7500 Fast Real-Time PCR System and Power CYBR Green PCR master mix, according to the manufacturer's instructions (Applied Biosystems). The primer sets are described in Supplementary Table S1. Ct values for POLD4 or POLD1 were normalized to those of 18S (ΔCt). The average ΔΔCt values were then calculated by normalization to the ΔCt values of A549 cells.

Chromatin immunoprecipitation

Chromatin immunoprecipitation samples were prepared from A549 or ACC-LC-48 cells logarithmically grown on 10-cm dishes, as previously described (11, 12). The genomic DNA contents were determined by quantitative RT-PCR (Supplementary Table S1). Results were normalized to the input control. SDs are also presented.

Immunoprecipitation of pol δ complexes

Cells were lysed in Nonidet P40 (NP40) lysis buffer [20 mmol/L HEPES (pH 7.8), 300 mmol/L NaCl, 1 mmol/L EDTA, and 1% NP40] supplemented with cOmplete (Roche Diagnostics Gmbh), then centrifuged at 4°C. The anti-POLD1 antibody was added to the supernatant fraction, mixed gently, and incubated overnight at 4°C. After Protein G Sepharose (GE Healthcare) was incubated with the mixture for 1 hour, it was precipitated and washed four times with NP40 lysis buffer. The pol δ complex was recovered from the Sepharose by boiling in SDS sample buffer and was analyzed using Western blotting.

Purification and characterization of pol δ

All proteins used in this study were overproduced in Escherichia coli and purified as previously described (13). The purity of the proteins, including 3- and 4-subunit pol δ, was monitored by SDS-PAGE and Coomassie brilliant blue R-250 staining. Protein concentrations were determined using Bio-Rad protein assays with bovine serum albumin (Bio-Rad) as the standard. DNA polymerase activity was measured in a reaction mixture (25 μL) containing 20 mmol/L HEPES-NaOH (pH 7.5), 50 mmol/L NaCl, 0.2 mg/mL bovine serum albumin, 1 mmol/L dithiothreitol, 10 mmol/L MgCl2, 1 mmol/L ATP, 0.1 mmol/L each of dGTP, dATP, dCTP, and [α-32P]dTTP, 33 fmol (240 pmol for nucleotides) of singly primed ss-mp18 DNA (Supplementary Table S1), 1.0 μg (9.1 pmol) of RPA, 86 ng (1.0 pmol as a trimer) of proliferating cell nuclear antigen (PCNA), 75 ng (260 fmol) of RFC, and 11 to 88 ng (46 to 372 fmol) of pol δ. After incubation at 30°C for 10 minutes, the products were electrophoresed on 0.7% alkaline-agarose gels (13).

Plasmid construction and isolation of stable clones

Using an IMAGE clone (ID: 6450262, Invitrogen) and the primer sets (Supplementary Table S1), the POLD4 open reading frame (ORF) was PCR-cloned, and inserted into the large HindIII-EcoRI fragment of pcDNA3 (FLAGPOLD4pcDNA3). ACC-LC-172 cells were transfected by FLAGPOLD4pcDNA3 and treated with G418 for establishment of stable clones. For construction of an siD4-resistant construct, silent mutations were introduced by synthetic oligomers (Supplementary Table S1). The PCR fragments were inserted into a pGFP-MSCV vector (pMSCVpold4). Plat-E cells were transfected using pMSCVpold4 and a VSV-G expression vector, and cultured for two days. The retrovirus was recovered from the culture medium and used for infection. After two days, green fluorescent protein–positive cells were isolated using FACS Vantage or Area2 (BD).

MTT assay

ACC-LC-172 cells and siRNA-treated (48 h) or non-treated cells were separately cultured in 200 μL of culture medium in 96-well plates at a density of 8,000/well. The next day, the medium was replaced with that containing 4-nitroquinoline-1-oxide (4NQO) or vinorelbine, followed by incubation for 48 hours. Viable cells were measured in triplicate using TetraColor One (Seikagaku) with reference to the viability of mock-treated cells.

Cell cycle synchronization

Calu6 was synchronized according to the method of Nakagawa et al. (8), with some modifications. Briefly, 24-hour treatment with 2 mmol/L thymidine was used to arrest exponentially proliferating cells in the S phase. The cells were then released from arrest by three washes in PBS, then siCTRL- and siD4-treated cells were grown in fresh medium for 12 and 15 hours, respectively. The release times were not the same due to a delayed cell cycle progression by siD4. Aphidicolin at 1 μmol/L was used for the second block for 24 hours, then the cells were released by three washes in PBS, incubated in normal culture medium for various times, and harvested for flow-cytometric and Western blotting analyses.

Chromosomal analysis

Colcemid at 0.1 μg/mL was added to the culture medium for 2 hours, then HCT116 cells were harvested, washed with PBS, and incubated at 37°C in 75 mmol/L KCl for 30 minutes, followed by centrifugation at 1,500 rpm for 5 minutes. The cells were further incubated in Carnoy's solution (methanol:acetate = 3:1) for 30 minutes, collected by centrifugation at 800 rpm for 5 minutes, resuspended in 0.5 mL of Carnoy's solution, dropped onto a wet (75% ethanol) slide glass from a height of 30 cm, and flamed with a burner. The slide glass was left overnight at room temperature, then dipped into 4% Giemsa solution for 15 minutes, rinsed with water from the opposite side of the sample, air-dried, and shielded by a coverslip. Mitotic chromosomes were analyzed using a Bx 60 microscope (Olympus), and the Invision V 4.07 (Biovision Technology) and Photoshop (Adobe) software packages.

Low expression of POLD4 mRNA and POLD4 protein in SCLC

Genomic instability is thought to underlie the pathogenesis of lung cancer. To elucidate the underlying mechanisms, we screened for genes involved in DNA metabolism (Dataset S1) using our previous profiling data set consisting of 149 cases of NSCLC, 9 cases of SCLC, and 5 normal lung mixture specimens (14). We found that POLD4 expression was distinctly reduced in SCLC, whereas a small fraction of NSCLC specimens also had a low expression of POLD4 that was comparable with SCLC (Fig. 1A). In contrast, other DNA pol δ subunits had expression levels in the order of SCLC > NSCLC > normal lung (Fig. 1A), whereas the majority of growth-related mRNAs had increased levels of expression in the lung cancer specimens (Supplementary Fig. S1A). In hierarchical clustering analysis, POLD4 was found in a distant small branch (Supplementary Fig. S2), suggesting a unique control mechanism of POLD4 expression. RT-PCR findings from a second panel of clinical samples confirmed our results (Fig. 1B). In addition, distinctive downregulation of POLD4 was observed in another database set (Supplementary Fig. S1B).

Figure 1.

Low expression levels of POLD4 in SCLC. A, microarray analysis of 149 NSCLC, 9 SCLC, and 5 normal lung mixture samples showed that the mRNA expression level of POLD4, but not of other pol δ subunits, was low in SCLC. Quartile data are shown in graphs. *, P < 0.05; **, P < 0.01, t-test. B, mRNA expression levels of POLD4 and POLD1 in 48 clinical tissue samples, independent of the microarray cohort, were determined by quantitative RT-PCR. The number of each sample class is in parentheses. Results were corrected with 18S expression values and are presented as relative values to that of A549, which was given a value of 1. Bars, SD. *, P < 0.05; **, P < 0.01, t-test. C, Western blotting analyses of POLD4 and POLD1 were carried out using a panel of lung cancer cell lines. D, the 149 NSCLC patients were classified into two groups according to POLD4 expression level, with the threshold set at −1 SD of the average. Kaplan-Meier survival curves for these groups showed a significant difference in both overall and relapse-free survival. P values were determined using a log-rank test.

Figure 1.

Low expression levels of POLD4 in SCLC. A, microarray analysis of 149 NSCLC, 9 SCLC, and 5 normal lung mixture samples showed that the mRNA expression level of POLD4, but not of other pol δ subunits, was low in SCLC. Quartile data are shown in graphs. *, P < 0.05; **, P < 0.01, t-test. B, mRNA expression levels of POLD4 and POLD1 in 48 clinical tissue samples, independent of the microarray cohort, were determined by quantitative RT-PCR. The number of each sample class is in parentheses. Results were corrected with 18S expression values and are presented as relative values to that of A549, which was given a value of 1. Bars, SD. *, P < 0.05; **, P < 0.01, t-test. C, Western blotting analyses of POLD4 and POLD1 were carried out using a panel of lung cancer cell lines. D, the 149 NSCLC patients were classified into two groups according to POLD4 expression level, with the threshold set at −1 SD of the average. Kaplan-Meier survival curves for these groups showed a significant difference in both overall and relapse-free survival. P values were determined using a log-rank test.

Close modal

The low level of POLD4 expression seen is unlikely attributable to the detection of splicing variants (Supplementary Fig. S3).Western blotting analysis showed that POLD4 (alternatively called p12) was detected at the level of the background in ACC-LC-76 and -172 (Fig. 1C). In addition, two other SCLC cell lines showed marginal expression levels, whereas apparent expressions were also observed in the NSCLC cell lines. In contrast, both panels of cell lines expressed a comparable amount of POLD1 (p125). It was also noted that a fraction of NSCLC specimens with low POLD4 expression levels comparable with SCLC were associated with poor prognosis (Fig. 1D). In accord with the low level of mRNA expression (Supplementary Fig. S4), POLD4 was induced by trichostatin A or valproic acid, class I and II mammalian histone deacetylase inhibitors, in ACC-LC-48 and -172 (Fig. 2A and B). ACC-LC-48 was further shown to be associated with higher histone H3K9 methylation and lower acetylation patterns in the POLD4 promoter region than was A549 (Fig. 2C). POLD4 expression in ACC-LC-76 was induced by 5 aza-2' deoxycytidine, although the promoter methylations were not dense and their contribution to POLD4 expression is yet to be studied (Supplementary Fig. S5A). On the other hand, proteasome-dependent protein degradation pathways did not seem to play a major role (Supplementary Fig. S5B).

Figure 2.

Detection of histone modifications. A, after treating cells with 0, 0.25, or 1 μg/mL of trichostatin A (TSA) for 48 (A549 and ACC-LC-48) or 16 (ACC-LC-172) hours (left), or with 0, 0.5, 1, or 2 μmol/L of valproic acid (VPA) for 16 hours (right), total RNA was extracted and quantitative RT-PCR was performed. The results were corrected by 18S expression values and are presented as relative to that without drugs, which was set at 1. Bars, SD. *, P < 0.05; **, P < 0.01, t-test. B, ACC-LC-48 and A549 cells were treated with 0, 0.25, or 1 μg/mL of TSA for 16 or 48 hours, then Western blotting analysis was performed for detection of POLD4 and POLD1, with α-tubulin used as the control (left). Western blotting analysis was also performed for POLD4 with or without 2 mmol/L VPA. C, chromatin immunoprecipitation samples were prepared as described in Materials and Methods. Input controls represent samples not subjected to immunoprecipitation (IP). In the promoter region of POLD4 of the NSCLC cell line A549, histone H3 was acetylated at a level 5.6 ± 2.4-fold greater as compared with the SCLC cell line ACC-LC-48. The trimethylation level of Lys9 of histone H3 in ACC-LC-48 was 3.6 ± 2.2-fold greater than that in A549. In a control experiment to analyze histone H3 modifications in the POLD1 promoter, similar amounts of acetylation and methylation were detected in the two cell lines.

Figure 2.

Detection of histone modifications. A, after treating cells with 0, 0.25, or 1 μg/mL of trichostatin A (TSA) for 48 (A549 and ACC-LC-48) or 16 (ACC-LC-172) hours (left), or with 0, 0.5, 1, or 2 μmol/L of valproic acid (VPA) for 16 hours (right), total RNA was extracted and quantitative RT-PCR was performed. The results were corrected by 18S expression values and are presented as relative to that without drugs, which was set at 1. Bars, SD. *, P < 0.05; **, P < 0.01, t-test. B, ACC-LC-48 and A549 cells were treated with 0, 0.25, or 1 μg/mL of TSA for 16 or 48 hours, then Western blotting analysis was performed for detection of POLD4 and POLD1, with α-tubulin used as the control (left). Western blotting analysis was also performed for POLD4 with or without 2 mmol/L VPA. C, chromatin immunoprecipitation samples were prepared as described in Materials and Methods. Input controls represent samples not subjected to immunoprecipitation (IP). In the promoter region of POLD4 of the NSCLC cell line A549, histone H3 was acetylated at a level 5.6 ± 2.4-fold greater as compared with the SCLC cell line ACC-LC-48. The trimethylation level of Lys9 of histone H3 in ACC-LC-48 was 3.6 ± 2.2-fold greater than that in A549. In a control experiment to analyze histone H3 modifications in the POLD1 promoter, similar amounts of acetylation and methylation were detected in the two cell lines.

Close modal

POLD4-deficient pol δ impairs pol and nucleotide excision DNA repair activities

Because pol δ is an essential protein complex for DNA replication and repair (15, 16), a subunit stoichiometry of the pol δ complex is required to exclude the possibility that POLD4 expression in SCLC and a fraction of NSCLC is sufficient for constituting the 4-subunit structure of the pol δ complex. POLD4 was identified in the immunoprecipitated pol δ complex using the anti-POLD1 antibody in the A549 cell line with a high expression of POLD4, but not in ACC-LC-172 cells with a low level of POLD4 (Fig. 3A), which strongly suggests that most of pol δ in ACC-LC-172 cells lack POLD4. In the presence of pol δ accessory proteins, including PCNA, RPA, and RFC, POLD4-deficient pol δ showed DNA replication activity that was lower than that of intact pol δ (Fig. 3B and C), which supports a previous finding (17, 18). We also examined the functional consequences of reduced POLD4 expression in lung cancer pathogenesis. To this end, we attempted to establish SCLC cell lines with ectopic POLD4 expression, although most of the SCLC cells did not allow POLD4 overexpression (data not shown); such intolerance has also been reported for another replicative DNA polymerase subunit, POLA1 (19). For this reason, we chose the NSCLC cell line SK-LC-6, which has a low level of POLD4 expression. Although 4NQO produces a DNA adduct that is removed through pol δ–dependent nucleotide excision DNA repair (NER), POLD4-introduced SK-LC-6 cells were more resistant to 4NQO (Fig. 3D and Supplementary Fig. S6A). Similar results were obtained with the SCLC cell line ACC-LC-172, despite the findings that this cell line is intrinsically resistant to 4NQO (Supplementary Fig. S6B). In a reciprocal experiment using the NSCLC cell lines Calu6 and PC-10, siRNA treatment against POLD4 (siD4) reduced the protein level to nearly one third of the original level, which resulted in increased sensitivity to 4NQO (Fig. 3D).

Figure 3.

Effects of POLD4 depletion on in vitro pol δ and NER activities. A, after immunoprecipitation was performed using the anti-POLD1 antibody (IP), POLD4 and POLD1 were detected by Western blotting analysis. A whole cell extract (WCE) was used as the control. B, the purity of the recombinant pol δ complex with or without the POLD4 subunit was analyzed by SDS gel electrophoresis. Each subunit is indicated on the right. C, DNA replication activity between the two structures of pol δ was compared in the presence of accessory proteins using primed M13 DNA as a template primer. The length of the synthesized DNA is indicated on the right. The amounts of pol δ complexes in the reactions varied from 0 to 88 ng, as described in Materials and Methods. D, expression levels of POLD4 in Calu6, SK-LC-6, and PC-10 cells. Depletion of POLD4 was monitored in Calu6 and PC-10 cells treated with either siCTRL or siD4. Below, results of MTT assays to determine cell growth two days after treatment with various concentrations of drugs. OD values with no drug were regarded as 100%. Clones and siRNA treatments are indicated. D4O/E and VC represent clones with POLD4 overexpression and the vector control, respectively.

Figure 3.

Effects of POLD4 depletion on in vitro pol δ and NER activities. A, after immunoprecipitation was performed using the anti-POLD1 antibody (IP), POLD4 and POLD1 were detected by Western blotting analysis. A whole cell extract (WCE) was used as the control. B, the purity of the recombinant pol δ complex with or without the POLD4 subunit was analyzed by SDS gel electrophoresis. Each subunit is indicated on the right. C, DNA replication activity between the two structures of pol δ was compared in the presence of accessory proteins using primed M13 DNA as a template primer. The length of the synthesized DNA is indicated on the right. The amounts of pol δ complexes in the reactions varied from 0 to 88 ng, as described in Materials and Methods. D, expression levels of POLD4 in Calu6, SK-LC-6, and PC-10 cells. Depletion of POLD4 was monitored in Calu6 and PC-10 cells treated with either siCTRL or siD4. Below, results of MTT assays to determine cell growth two days after treatment with various concentrations of drugs. OD values with no drug were regarded as 100%. Clones and siRNA treatments are indicated. D4O/E and VC represent clones with POLD4 overexpression and the vector control, respectively.

Close modal

To examine whether POLD4 expression has any specificity to the NER pathway, the effects of POLD4 expression on sensitivities to vinorelbine, a microtubule assembly inhibitor, were studied. For this experiment, we used PC-10 cells, as they have modest sensitivity to a variety of drugs, indicating that this cell line has a low possibility of having abnormal drug metabolic pathways (data not shown). Our results showed that siD4 did not alter the sensitivity to vinorelbine (Fig. 3D).

POLD4 required for cell cycle progression

A previous study showed that the POLD4 ortholog of Cdm1 in Schizosaccharomyces pombe (S. pombe) is a nonessential gene (20), although it may be required in mammalian cells (21, 22). To study human POLD4 functions, we depleted POLD4 using siD4 and analyzed the cell cycle progression. In the NSCLC cell line Calu6 as well as in others, siD4 altered the cell cycle of each population (Supplementary Fig. S7A and B), and induced p21 and p27 (Supplementary Fig. S7C).

We synchronized and released cells from the G1-S boundary, and chased them. A fraction of the siD4-treated cells did not progress into the S-phase, but rather stayed at G1-S (Fig. 4A, filled arrows). Again, persistent induction of p21 and p27 was detected (Fig. 4B). It was also observed that the timing of entry into the G1 phase was not the same between siCTRL and siD4 cells. In the latter, G1 entry was seen 12 hours after release, which was 2 hours later than that with the siCTRL cells (Fig. 4A, open arrows). In addition, the DNA damage checkpoint genes NBS1, ATM, CHK2, SMC1, and CHK1 were phosphorylated throughout S to G2 (Fig. 4C), suggesting that the cell cycle was delayed by activation of these checkpoint proteins. These phenotypes were rescued in stable clones that carried pold4 with silent mutations at the siD4 target site (Fig. 5). Additional knockdown of p27 resulted in efficient S-phase entry, which indicates that p27 induction inhibits cell cycle progression at the G1-S boundary in Calu6 cells (Supplementary Fig. S8A–D).

Figure 4.

POLD4 reduction induces checkpoint activation. A, Calu6 cells were synchronized at the G1-S boundary and released, then DNA contents were measured every two hours. Filled arrows, G1-S arrested population; open arrows, G1 re-entry. B, after treating Calu6 cells with siRNA for 48 hours, the cells were analyzed for the expression levels of G1-S checkpoint proteins. C, at each time point after release, Western blotting analysis was performed using the indicated antibodies.

Figure 4.

POLD4 reduction induces checkpoint activation. A, Calu6 cells were synchronized at the G1-S boundary and released, then DNA contents were measured every two hours. Filled arrows, G1-S arrested population; open arrows, G1 re-entry. B, after treating Calu6 cells with siRNA for 48 hours, the cells were analyzed for the expression levels of G1-S checkpoint proteins. C, at each time point after release, Western blotting analysis was performed using the indicated antibodies.

Close modal
Figure 5.

Ectopic overexpression of POLD4 with silent mutations rescues cell cycle phenotypes. A, Western blotting analysis using the rabbit polyclonal anti-POLD4 antibody was carried out to monitor overexpression in Calu6 pold4_1 and _2 clones (top), which were resistant to siD4 treatment (anti-FLAG antibody, bottom). B, Calu6 pold4_1 and _2 and the vector control clones of pMSCVvc_1 and _2 were synchronized and released, then monitored for DNA content. Filled arrows, G1-S arrested population; open arrows, G1 re-entry. C and D, at each time point after release, Western blotting analysis was performed using the indicated antibodies and Calu6 stable clones.

Figure 5.

Ectopic overexpression of POLD4 with silent mutations rescues cell cycle phenotypes. A, Western blotting analysis using the rabbit polyclonal anti-POLD4 antibody was carried out to monitor overexpression in Calu6 pold4_1 and _2 clones (top), which were resistant to siD4 treatment (anti-FLAG antibody, bottom). B, Calu6 pold4_1 and _2 and the vector control clones of pMSCVvc_1 and _2 were synchronized and released, then monitored for DNA content. Filled arrows, G1-S arrested population; open arrows, G1 re-entry. C and D, at each time point after release, Western blotting analysis was performed using the indicated antibodies and Calu6 stable clones.

Close modal

These results suggest that POLD4 downregulation activates checkpoint proteins, induces G1-S arrest, and delays the cell cycle from S to G2. Generation of double-stranded DNA breaks (DSB) was also assessed by a knockdown experiment of DNAPK, a nonhomologous end-joining kinase, which had a synergistic effect with siD4 on the G1-S population (Supplementary Fig. S8E and F).

Induction of chromosome breaks and gaps by POLD4 reduction

Activation of the ATM, ATR, and DNA-PK pathways suggests that both DSB and single-stranded breaks were generated by POLD4 reduction. These results are in accord with our recent observation that SCLC cells are associated with constitutive DSB (5). Notably, POLD4 overexpression reduced the γ-H2AX intensity (Fig. 6A and B). Furthermore, we investigated whether POLD4 reduction also induces any chromosome aberrations. For this experiment, most lung cancer cell lines were not applicable to this particular purpose because they exhibit aneuploidy in association with chromosomal instability (23). We used HCT116 cells where POLD4 expression levels were sufficiently high (Fig. 6A). siD4 treatment significantly increased the numbers of chromosomal gaps and breaks to twice as many as observed with siCTRL treatment, whereas this phenotype was readily suppressed in cells expressing pold4 with silent mutations at the siRNA target site (Fig. 6A, C, and D).

Figure 6.

Chromosomal gap/break formation in cells. A, Western blotting analysis was performed to confirm the overexpression of POLD4 in the stable clones (ACC-LC-172, D4 O/E1, and O/E2). Shown are representative results of Western blotting analysis using HCT116 and the monoclonal antibody (left), HCT116 stable clones and the rabbit polyclonal anti-POLD4 antibody (middle), and a mixture of 7 stable clones with pold4 overexpression and the anti-FLAG antibody (right), which were used to monitor silencing efficiency, overexpression, and resistance against siD4 treatment, respectively. B, S-phase cells from ACC-LC-172 D4 O/E and control clones were quantitated in regard to γ−H2AX induction, as indicated. C, representative photograph of mitotic chromosomes. Arrows, chromosomal breaks; bar, 10 μmol/L. D, chromosomal gaps and breaks were quantitated by counting 100 mitotic cells. The numbers of cells with ≥2 gaps/breaks were plotted for mock, vector control (VC), and mD4-overexpressed HCT116 cells (pold4). Stable VC and mD4 clones were mixtures of 7 independent clones. *, P < 0.05; **, P < 0.01, Fisher's exact test). To acquire sufficient numbers of high-quality images, the sums of aberrations from five independent experiments were used for comparison. The numbers of total gaps/breaks are also shown in Supplementary Fig. S9.

Figure 6.

Chromosomal gap/break formation in cells. A, Western blotting analysis was performed to confirm the overexpression of POLD4 in the stable clones (ACC-LC-172, D4 O/E1, and O/E2). Shown are representative results of Western blotting analysis using HCT116 and the monoclonal antibody (left), HCT116 stable clones and the rabbit polyclonal anti-POLD4 antibody (middle), and a mixture of 7 stable clones with pold4 overexpression and the anti-FLAG antibody (right), which were used to monitor silencing efficiency, overexpression, and resistance against siD4 treatment, respectively. B, S-phase cells from ACC-LC-172 D4 O/E and control clones were quantitated in regard to γ−H2AX induction, as indicated. C, representative photograph of mitotic chromosomes. Arrows, chromosomal breaks; bar, 10 μmol/L. D, chromosomal gaps and breaks were quantitated by counting 100 mitotic cells. The numbers of cells with ≥2 gaps/breaks were plotted for mock, vector control (VC), and mD4-overexpressed HCT116 cells (pold4). Stable VC and mD4 clones were mixtures of 7 independent clones. *, P < 0.05; **, P < 0.01, Fisher's exact test). To acquire sufficient numbers of high-quality images, the sums of aberrations from five independent experiments were used for comparison. The numbers of total gaps/breaks are also shown in Supplementary Fig. S9.

Close modal

To date, despite the premise that deficits in DNA replication machinery account for multiple mutations in cancer (10), and findings showing that laboratory-born animals and strains defective in DNA replication machineries exhibit genomic instability (24, 25), core DNA replication proteins have rarely been identified as a frequent target for alterations in human cancers. The present results show that the pol δ subunit POLD4 is frequently reduced in some lung cancer specimens and cell lines, whereas overexpression of POLD4 complements part of the pathologic phenotypes of lung cancer cell lines.

It is interesting to note that although the POLD4 reduction was most evident in SCLC, which is a very aggressive form of lung cancer, a small fraction of NSCLC cases also showed low expression in association with poor prognosis. In in vitro experiments, we found that POLD4 is required to suppress DSB and subsequent checkpoint activation. Together with our previous findings that clinically obtained SCLC specimens were frequently associated with discrete γ-H2AX foci formation (5), these results suggest that reduced POLD4 expression may play a role in lung cancer progression in association with the genomic instability phenotypes.

Along this line, cells with impaired activity of replicative DNA polymerases induce deletions with short nucleotides (<1 kb) with similar base substitution spectra (26, 27), whereas the results of recent next-generation sequencing studies that utilized one sample each of SCLC and NSCLC showed frequent transversions in both types of cancer, and also indicated that SCLC carried significantly more frequent deletions (53 of 65) than NSCLC (93 of 424; P = 1.28 × 10-20, Fisher's exact test, denominators include all small insertions and deletions; refs. 28, 29). Impaired activity in replicative DNA polymerases is also known to increase the possibility of large deletions and duplications of chromosomes and arms (30, 31). Using datasets made from 74 lung adenocarcinoma patients for gene expression levels as well as copy number variations (32, 33), we noted that low POLD4 expression was significantly associated with 8p, 9q, and 13q deletions, and 5p, 7p, 8q, and 14q amplifications, which are shared by both NSCLC and SCLC (refs. 34, 35; Supplementary Table S2). In addition, we observed that low POLD4 expression in primary tumors was associated with high expression of HDAC2 (P = 0.0036; see Supplementary Information), which was in accord with our in vitro data showing that POLD4 expression is regulated by histone modifications.

SCLC exhibits aggressive cell growth leading to very poor clinical prognosis, although it is also characterized by large areas of necrosis (36) and cell lines are difficult to establish (37). The putative growth disadvantage due to low POLD4 might be partially compromised by the characteristics of SCLC. First, G1-S transition may be supplemented by factors that are highly expressed in SCLC. With regard to this speculation, it has been reported that CDC6 and CDT1, which are recruited at G1-S for licensing the DNA replication, are abnormally overexpressed in the early stage of carcinogenesis (38). In our analysis of SCLC, >10-fold expressions of CDC6 and CDT1 were observed in the specimens (Supplementary Fig. S1A). The G1-S transition may be thereby facilitated by forced entry into the S phase in SCLC. Second, nearly all SCLC tumors are deficient of Rb and p53, which are critical regulators for G1-S transition (1, 23). Checkpoint impairment may benefit SCLC cell proliferation, survival, increased genomic instability, and tumor progression, as well as suppress the growth defect induction caused by POLD4 downregulation. Third, the POLD4 ortholog of Cdm1 is a nonessential gene in S. pombe (20). Although the complexity of the DNA replication mechanisms and amounts of chromosomal DNA are not the same between yeast and human cells, siD4-treated cells were able to proliferate throughout the cell cycle, as shown in Fig. 4 (22).

Tobacco smoke contains benzo(a)pyrene, one of the most potent carcinogens, whereas DNA adducts of benzo(a)pyrene are removed by pol δ–dependent NER as are those of 4NQO. Along this line, previous epidemiologic studies have indicated a strong association of lung cancer with tobacco smoking (1). Benzo(a)pyrene adducts might remain in DNA and drive genomic instability in lung cells with reduced POLD4 expression. This notion is also in line with a recent hypothesis that DNA replication stress leads to DSB, genomic instability, and selective pressure for later p53 mutations (3, 5, 6). It is important to note, however, that involvement of POLD4 as a crucial effecter of an unidentified primary target for genetic/epigenetic alterations in lung cancer cannot be formerly precluded. Future studies, such as with genetically engineered mice, are required to determine whether POLD4 downregulation is involved in the early stage of carcinogenesis or if it contributes to the late progression stage.

Our results indicate that POLD4 is required for maintenance of genomic stability of human cells. With low expression levels of POLD4, lung cancer may grow with accumulation of DNA damage. Publicly available datasets show that POLD4 expression levels vary widely in other types of cancer, although most datasets do not contain information for corresponding normal tissues. It would be interesting to study whether POLD4 reduction is specific to lung cancer or may be involved in the development of other types of cancer.

No potential conflicts of interest were disclosed.

Grant Support: A Grant-in-Aid for Scientific Research on Innovative Areas, a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science, and the Personalized Medicine Project of the Japan Science and Technology Agency.

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.

1
Osada
H
,
Takahashi
T
. 
Genetic alterations of multiple tumor suppressors and oncogenes in the carcinogenesis and progression of lung cancer
.
Oncogene
2002
;
21
:
7421
34
.
2
Worden
FP
,
Kalemkerian
GP
. 
Therapeutic advances in small cell lung cancer
.
Expert Opin Investig Drugs
2000
;
9
:
565
79
.
3
Bartkova
J
,
Horejsi
Z
,
Koed
K
, et al
. 
DNA damage response as a candidate anti-cancer barrier in early human tumorigenesis
.
Nature
2005
;
434
:
864
70
.
4
Ebi
H
,
Matsuo
K
,
Sugito
N
, et al
. 
Novel NBS1 heterozygous germ line mutation causing MRE11-binding domain loss predisposes to common types of cancer
.
Cancer Res
2007
;
67
:
11158
65
.
5
Ebi
H
,
Sato
T
,
Sugito
N
, et al
. 
Counterbalance between RB inactivation and miR-17–92 overexpression in reactive oxygen species and DNA damage induction in lung cancers
.
Oncogene
2009
;
28
:
3371
9
.
6
Gorgoulis
VG
,
Vassiliou
LV
,
Karakaidos
P
, et al
. 
Activation of the DNA damage checkpoint and genomic instability in human precancerous lesions
.
Nature
2005
;
434
:
907
13
.
7
Konishi
H
,
Nakagawa
T
,
Harano
T
, et al
. 
Identification of frequent G(2) checkpoint impairment and a homozygous deletion of 14–3-3ϵ at 17p13.3 in small cell lung cancers
.
Cancer Res
2002
;
62
:
271
6
.
8
Nakagawa
T
,
Hayashita
Y
,
Maeno
K
, et al
. 
Identification of decatenation G2 checkpoint impairment independently of DNA damage G2 checkpoint in human lung cancer cell lines
.
Cancer Res
2004
;
64
:
4826
32
.
9
Kaufmann
WK
. 
Initiating the uninitiated: replication of damaged DNA and carcinogenesis
.
Cell Cycle
2007
;
6
:
1460
7
.
10
Bielas
JH
,
Loeb
LA
. 
Mutator phenotype in cancer: timing and perspectives
.
Environ Mol Mutagen
2005
;
45
:
206
13
.
11
Osada
H
,
Tatematsu
Y
,
Masuda
A
, et al
. 
Heterogeneous transforming growth factor (TGF)-β unresponsiveness and loss of TGF-β receptor type II expression caused by histone deacetylation in lung cancer cell lines
.
Cancer Res
2001
;
61
:
8331
9
.
12
Osada
H
,
Tatematsu
Y
,
Sugito
N
,
Horio
Y
,
Takahashi
T
. 
Histone modification in the TGFβRII gene promoter and its significance for responsiveness to HDAC inhibitor in lung cancer cell lines
.
Mol Carcinog
2005
;
44
:
233
41
.
13
Masuda
Y
,
Suzuki
M
,
Piao
J
,
Gu
Y
,
Tsurimoto
T
,
Kamiya
K
. 
Dynamics of human replication factors in the elongation phase of DNA replication
.
Nucleic Acids Res
2007
;
35
:
6904
16
.
14
Takeuchi
T
,
Tomida
S
,
Yatabe
Y
, et al
. 
Expression profile-defined classification of lung adenocarcinoma shows close relationship with underlying major genetic changes and clinicopathologic behaviors
.
J Clin Oncol
2006
;
24
:
1679
88
.
15
Pavlov
YI
,
Frahm
C
,
McElhinny
SA
,
Niimi
A
,
Suzuki
M
,
Kunkel
TA
. 
Evidence that errors made by DNA polymerase α are corrected by DNA polymerase δ
.
Curr Biol
2006
;
16
:
202
7
.
16
Pavlov
YI
,
Shcherbakova
PV
,
Rogozin
IB
. 
Roles of DNA polymerases in replication, repair, and recombination in eukaryotes
.
Int Rev Cytol
2006
;
255
:
41
132
.
17
Li
H
,
Xie
B
,
Zhou
Y
, et al
. 
Functional roles of p12, the fourth subunit of human DNA polymerase δ
.
J Biol Chem
2006
;
281
:
14748
55
.
18
Podust
VN
,
Chang
LS
,
Ott
R
,
Dianov
GL
,
Fanning
E
. 
Reconstitution of human DNA polymerase δ using recombinant baculoviruses: the p12 subunit potentiates DNA polymerizing activity of the four-subunit enzyme
.
J Biol Chem
2002
;
277
:
3894
901
.
19
Tanaka
S
,
Cao
K
,
Niimi
A
, et al
. 
Functions of base selection step in human DNA polymerase α
.
DNA Repair (Amst)
2010
;
9
:
534
41
.
20
Reynolds
N
,
Watt
A
,
Fantes
PA
,
MacNeill
SA
. 
Cdm1, the smallest subunit of DNA polymerase d in the fission yeast Schizosaccharomyces pombe, is non-essential for growth and division
.
Curr Genet
1998
;
34
:
250
8
.
21
Dell'Era
P
,
Nicoli
S
,
Peri
G
,
Nieddu
M
,
Ennas
MG
,
Presta
M
. 
FGF2-induced upregulation of DNA polymerase-δ p12 subunit in endothelial cells
.
Oncogene
2005
;
24
:
1117
21
.
22
Huang
QM
,
Akashi
T
,
Masuda
Y
,
Kamiya
K
,
Takahashi
T
,
Suzuki
M
. 
Roles of POLD4, smallest subunit of DNA polymerase δ, in nuclear structures and genomic stability of human cells
.
Biochem Biophys Res Commun
2010
;
391
:
542
6
.
23
Masuda
A
,
Takahashi
T
. 
Chromosome instability in human lung cancers: possible underlying mechanisms and potential consequences in the pathogenesis
.
Oncogene
2002
;
21
:
6884
97
.
24
Niimi
A
,
Limsirichaikul
S
,
Yoshida
S
, et al
. 
Palm mutants in DNA polymerases α and η alter DNA replication fidelity and translesion activity
.
Mol Cell Biol
2004
;
24
:
2734
46
.
25
Albertson
TM
,
Ogawa
M
,
Bugni
JM
, et al
. 
DNA polymerase ϵ and δ proofreading suppress discrete mutator and cancer phenotypes in mice
.
Proc Natl Acad Sci U S A
2009
;
106
:
17101
4
.
26
Ogawa
M
,
Limsirichaikul
S
,
Niimi
A
,
Iwai
S
,
Yoshida
S
,
Suzuki
M
. 
Distinct function of conserved amino acids in the fingers of Saccharomyces cerevisiae DNA polymerase α
.
J Biol Chem
2003
;
278
:
19071
8
.
27
Kokoska
RJ
,
Stefanovic
L
,
DeMai
J
,
Petes
TD
. 
Increased rates of genomic deletions generated by mutations in the yeast gene encoding DNA polymerase δ or by decreases in the cellular levels of DNA polymerase δ
.
Mol Cell Biol
2000
;
20
:
7490
504
.
28
Pleasance
ED
,
Stephens
PJ
,
O'Meara
S
, et al
. 
A small-cell lung cancer genome with complex signatures of tobacco exposure
.
Nature
2010
;
463
:
184
90
.
29
Lee
W
,
Jiang
Z
,
Liu
J
, et al
. 
The mutation spectrum revealed by paired genome sequences from a lung cancer patient
.
Nature
2010
;
465
:
473
7
.
30
Lemoine
FJ
,
Degtyareva
NP
,
Kokoska
RJ
,
Petes
TD
. 
Reduced levels of DNA polymerase δ induce chromosome fragile site instability in yeast
.
Mol Cell Biol
2008
;
28
:
5359
68
.
31
Lemoine
FJ
,
Degtyareva
NP
,
Lobachev
K
,
Petes
TD
. 
Chromosomal translocations in yeast induced by low levels of DNA polymerase a model for chromosome fragile sites
.
Cell
2005
;
120
:
587
98
.
32
Weir
BA
,
Woo
MS
,
Getz
G
, et al
. 
Characterizing the cancer genome in lung adenocarcinoma
.
Nature
2007
;
450
:
893
8
.
33
Ding
L
,
Getz
G
,
Wheeler
DA
, et al
. 
Somatic mutations affect key pathways in lung adenocarcinoma
.
Nature
2008
;
455
:
1069
75
.
34
Merlo
A
,
Gabrielson
E
,
Mabry
M
,
Vollmer
R
,
Baylin
SB
,
Sidransky
D
. 
Homozygous deletion on chromosome 9p and loss of heterozygosity on 9q, 6p, and 6q in primary human small cell lung cancer
.
Cancer Res
1994
;
54
:
2322
6
.
35
Levin
NA
,
Brzoska
PM
,
Warnock
ML
,
Gray
JW
,
Christman
MF
. 
Identification of novel regions of altered DNA copy number in small cell lung tumors
.
Genes Chromosomes Cancer
1995
;
13
:
175
85
.
36
Pisick
E
,
Jagadeesh
S
,
Salgia
R
. 
Small cell lung cancer: from molecular biology to novel therapeutics
.
J Exp Ther Oncol
2003
;
3
:
305
18
.
37
Brouwer
M
,
de Ley
L
,
Feltkamp
CA
,
Elema
J
,
Jongsma
AP
. 
Serum-dependent “cannibalism” and autodestruction in cultures of human small cell carcinoma of the lung
.
Cancer Res
1984
;
44
:
2947
51
.
38
Liontos
M
,
Koutsami
M
,
Sideridou
M
, et al
. 
Deregulated overexpression of hCdt1 and hCdc6 promotes malignant behavior
.
Cancer Res
2007
;
67
:
10899
909
.