The human protein Nα-terminal acetyltransferase A complex (hNatA), composed of the catalytic hNaa10p (hArd1) and auxiliary hNaa15p (hNat1/NATH/Tubedown) subunits, was reported to be important for cell survival and growth of various types of cancer. However, little is known about the mechanisms mediating growth inhibition and apoptosis following loss of hNatA function. Here, we have screened 11 different thyroid cell lines for hNAA10 RNAi phenotypes and observed mostly growth inhibition, which was independent of TP53 functional status and developed by several different mechanisms involving (i) downregulation of cyclin D1, (ii) increase in p27/Kip1 and (iii) inactivation of Rb/E2F pathway. hNatA depletion in aggressive thyroid cancer cell lines (8305C, CAL-62 and FTC-133) with mutated TP53 increased sensitivity to drug-induced cytotoxicity, but in a cell type specific manner: 8305C (TRAIL), CAL-62 (daunorubicin) and FTC-133 (troglitazone). Cells harboring wild-type TP53 were also prone to apoptosis via the p53 pathway after hNatA downregulation. Importantly, in hNatA-depleted cells DNA-damage signaling was activated in the absence of exogenous DNA damage independent on TP53 status. Our findings indicate that several mechanisms of growth inhibition and apoptosis may be induced by hNatA knockdown and that hNatA knockdown could be exploited for use in combinatorial chemotherapy.
Protein acetylation is a widespread and important protein modification linked to regulation of cell differentiation, apoptosis, cell cycle and cancer.1–4 Human Nα-terminal acetyltransferase A (hNatA), composed of the catalytic subunit hNaa10p (hArd1) and auxiliary subunit hNaa15p (NATH), is the major Nα-acetyltransferase in humans cotranslationally acetylating nascent polypeptides with Ala-, Ser-, Thr-, Val- or Gly- N-termini after cleavage of the initiator Met.5–7
An increasing interest to the hNatA complex results from recent findings demonstrating correlation between expression of the hNatA subunits and tumor development. hNAA15 overexpression was found characteristic for gastric cancer,8 thyroid neoplasms9, 10 and neuroblastomas,11 while upregulation of hNAA10 (hNaa10p) was found in hepatocellular carcinoma,12 colon cancer13, 14 and breast cancer.15 Several studies indicate a drop in hNAA10/hNAA15 expression levels during differentiation.11, 16 Functional studies of hNatA revealed that its subunits are essential for the maintenance of growth and survival of several cancer cell types.16–18 Taken together, hNatA may be a potential target for cancer therapy.19 However, none of the studies elaborates the exact mechanisms through which hNatA knockdown phenotypes are mediated.
In the present communication, we investigate the effects of hNatA depletion in a variety of thyroid cancer cell lines with respect to changes in signaling pathways and sensitization to drug treatment. Disruption of hNatA function through knockdown of the catalytic subunit, hNaa10p, leads to p53-dependent apoptosis in cells with functional p53 or growth defects in cells with nonfunctional p53. In addition, we reveal early DNA-damage signaling after hNatA knockdown independent on p53 functional status.
Material and Methods
Human thyroid cell lines derived from anaplastic carcinoma (CAL-62, 8305C), papillary carcinoma (B-CPAP, BHT-101, ONCO-DG1), follicular carcinoma (FTC-133, CGTH-W1), differentiated carcinoma (K1) and immortalized normal thyroid follicular cells (Nthy-ori 3.1) were purchased from DSMZ (Braunschweig, Germany) or from European Collection of Cell Cultures (Salisbury, UK). Anaplastic thyroid carcinoma (ARO) and immortalized fetal normal thyroid follicular cells (TAD-2) were kindly provided by Profs. M. Santoro and A. Fusco, University of Naples, Italy and Prof. T.F. Davies, Mount Sinai School of Medicine, NY, respectively. All cell lines were grown according to originator's recommendations either in DMEM (CAL-62, ARO, BHT-101), in RPMI 1640 (CGTH-W-1, B-CPAP, 8305C, ONCO-DG1, Nthy-ori-3.1), or in DMEM:Ham-F12 (1:1) (K1 and FTC-133) with 2 mM L-glutamine, 50 μg/ml gentamicin and 10% NCS (Lonza Group, Basel, Switzerland). HCT116 (human colon carcinoma) p53+/+ and parental p53−/− were a generous gift from Dr. Fred Bunz, Bert Vogelstein & Kenneth W. Kinzler (John Hopkins University and Howard Hughes Medical Institute, MD), and were routinely maintained in McCoy's 5A medium with 10% NCS.
siRNA mediated knockdown was performed using DharmaFECT1 transfection reagent and siRNA duplexes from ThermoFisher Scientific (Lafayette, CO). The sequences targeting hNAA10 (NM_003491) were: sihNAA10-1 (CUACCAGAUGAAAUACUAC), sihNAA10-2 (CUUUCAGAUCAGUGAAGUG) and sihNAA10-3 (ON-TARGETplus SMARTpool targeting siRNA (Dharmacon # L-009606-00). All siRNAs were used in a final concentration 50 nM and ON-TARGETplus SMARTpool nontargeting siRNA (Dharmacon # L-D-001810-10) was a negative control.
Daunorubicin and troglitazone were obtained from Sigma (St. Louis, MO). KillerTRAIL was from Axxora (Bingham, UK).
Rabbit polyclonal antibodies against hNaa10p (anti-hArd1) and hNaa15p (anti-NATH) were previously described.9, 20 Antibodies against Bcl-2 (clone 7) and Bcl-xL (rabbit polyclonal) were from BD/Pharmingen (Franklin Lakes, NJ). Antibodies against phospho-p53(Ser15), phospho-Chk2(Thr68) and phospho-H2A.X (Ser139) were part of a DNA-damage sampler kit; antibodies against Rb (clone 4H1), phospho-Rb (Ser780) and phospho-Rb (Ser807/811) were from a Rb sampler kit; antibodies against PARP-cleaved (Asp214) (19F4), α-Fodrin cleaved (Asp1185) and rabbit polyclonal anti-acetyl-lysine antibodies were from Cell Signaling Technology (Danvers, MA). Antibodies against cyclin A (BF683), cyclin B1 (GNS1), cyclin E (M-20), p27 (sx53G8.5), p21 (sx118), E2F-1 (KH129), β-catenin (E-5) and actin (c-11) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies against cyclin D1 (DSC-6), p53 (pAb1801) and β-tubulin (2-28-33) were from Sigma. Anti-MDM2 (Ab-1) mouse mAb (IF2) and anti-p14ARF (ab3642) antibodies were obtained respectively from Calbiochem/EMD Chemicals (Merck KGaA, Darmstadt, Germany) and Abcam (Abcam plc, Cambridge, UK). The horseradish peroxidase-conjugated secondary antibodies were from Cell Signaling.
Cell monolayer was trypsinized, resuspended in a fresh medium and counted with Coulter (Beckman Coulter, Fullerton, CA). Equal amount of cells was seeded in 12-well plates. Four hours later the cells were transfected with correspondent siRNAs. Counting of cells was carried out every day until control cells reached confluence. On the last day of counting, inhibitory effect on cell growth was determined as strong (>40% growth inhibition), moderate (20–40%) or weak (10–20%).
Bromodeoxyuridine (BrdU) incorporation assay
Cells growing in a 96-well plate were transfected with correspondent siRNAs. After 48 hr, the proliferation rate/mitotic activity was determined by measuring the amount of bromodeoxyuridine incorporated into nuclear DNA using the chemiluminescent BrdU incorporation assay, Cell Proliferation ELISA (Roche, Mannheim, Germany) according to manufacturer's instructions.
HCT116 p53+/+ cells were seeded at 4 × 103 cells per well in 96-well plates and after 8 hr were transfected with 50 nM siRNA. Seventy-two hours after transfection, water-soluble tetrazolium 1, WST-1 (Roche) was added to each well (1:10) for 1 hr and absorbance at 450 nm was measured. Reduction in cell viability was expressed in percentage as the ratio of absorbance (sample) versus absorbance (control) multiplied by 100.
Cell cycle analysis
Approximately 1 × 106 cells were harvested by Trypsin-EDTA treatment, and rapidly fixated in ice-cold 70% ethanol for at least 2 hr at 4°C. Cells were then washed once, resuspended in 800 μl phosphate-buffered saline (PBS), Rnase A (1 mg/ml) and propidium iodide (PI, 40 μg/ml) before incubation at 37°C for 30 min. Samples were immediately analyzed by flow cytometry (Becton Dickinson FACS Calibur) with Cell Quest Lysis II software (Becton Dickinson, Franklin Lakes, NJ). Thirty thousand events were collected and a cytogram based on forward light scatter versus side light scatter was used to eliminate aggregates, debris and dead cells. Red fluorescence (FL3) in linear mode from gated cells (R1) were detected and analyzed by FlowJo (Tree Star, Ashland, OR), cell cycle module (Watson model). Results are representative of 3–4 independent experiments.
Lysates were prepared with 1× SDS Sample buffer [62.5 mM Tris-HCl (pH 6.8), 1% w/v SDS, 10% glycerol, 50 mM β-ME], then centrifuged at >10,000g and the protein concentration was determined using micro-BCA kit (Thermo Fisher Scientific, Rockford, IL). Equal amounts of total protein were analyzed by SDS-PAGE, blotted and detected with corresponding antibodies.
RNA purification and cDNA synthesis
Total RNA was extracted with TRIzol reagent (Invitrogen, San Diego, CA) according to manufacturer's instructions. cDNA was synthesized from 1 μg total RNA using Transcriptor Reverse Transcriptase (Roche) and oligo(dT)15 primer.
Amplification and sequencing of TP53 cDNA
Amplification of the coding region of TP53 was performed by specific nested PCR using outer primer set (forward primer: GACACTTTGCGTTCGGGC; reverse primer: CTTGTTCAGT GGAGCCCCG); and further inner primer set (forward primer: GACACGCTTCCCTGGATTGG; reverse primer: CGCACACCTATTGCAAGCAAGGG).21 PCR amplification was carried out using Pfu DNA polymerase (Fermentas, Burlington, Canada) in a 50 μl reaction mix containing 1× Pfu buffer containing MgSO4, 0.2 mM of each deoxynucleotide triphosphate, 0.5 μM of each primer, 1.25 U Pfu DNA polymerase and DNA template (1 μl cDNA (1:1) or 1 μl nested PCR product (1:1). The PCR conditions were an initial denaturation step of 3 min at 95°C followed by 40 cycles of 30 sec at 95°C, 30 sec at 51°C and 2 min at 73°C, and a final elongation step of 5 min at 73°C. DNA sequencing was performed using forward primer: TGGCCCCT CCTCAGCATCTTA and reverse primer: GTACAGTCA GAGCCAACCTC. Capillary electrophoresis, data collection and sequence analysis were performed on an automated DNA sequencer (ABI 3700).
Real-time quantitative PCR
Relative gene expression levels of p53-downstream genes (BAX, PUMA, NOXA, FAS and KILLER/DR5) in hNatA-depleted cells were determined by real-time quantitative PCR (RT-qPCR). The sequences of primers are listed in Supporting Information Table II. Ribosomal protein large P2, RPLP2, was used as a reference gene. Two microliters of cDNA (1:5 diluted) and 0.5 μM primers were mixed with components from the LightCycler 480 SYBR Green I Master mix kit (Roche) in a final volume 10 μl. Reactions in triplicate were carried out in the LightCycler 480 real-time PCR machine (Roche) under the following conditions: initial denaturation at 95°C for 5 min and then 40 cycles of denaturation at 95°C for 10 sec, annealing at 54°C for 10 sec and extension at 72°C for 10 sec. Melting curves were obtained to examine the purity of amplified products. Absolute quantitative data and CP values were obtained by analysis with LightCycler 480 Software 1.25 by second derivative method. Normalization of obtained data was done as follows: [copy numbers (BAX, NOXA, PUMA, FAS, KILLER/DR5 or hNAA10) in a sample/copy numbers of reference gene RPLP2 in a sample)]/[copy numbers (BAX, NOXA, PUMA, KILLER/DR5, FAS or hNAA10) in control)/copy numbers of reference gene RPLP2 in the control]. The p-values were obtained by independent t-test.
Terminal deoxyribonucleotidyl transferase-mediated dUTP nick end labeling (TUNEL)
Cells growing on coverslips in 24-well plates were transfected by 50 nM siRNA targeting hNAA10. After 72 hr, the cells were stained with Hoechst 33342 (1 μg/ml) for 15 min at 37°C, rinsed with PBS, fixed for 30 min at a room temperature in 3.7% formaldehyde. Cells were permeabilized for 2 min on ice in 0.1% Triton X-100 in sodium citrate and then incubated with TUNEL reaction mixture (In Situ Cell Death Detection Kit, TMRred, Roche) for 1 hr at 37°C. Positive and negative controls for each sample were done by treating samples with DNaseI or incubation in TUNEL Label reaction without enzyme, respectively. After staining, the cells were washed 4 times with PBS, and the coverslips were subsequently mounted onto slides. Images were taken using Leica DMI 6000B microscope (Leica Microsystems), equipped with a 20× objective (numerical aperture 0.7) and Leica DFC350 FX camera. TUNEL-positive cells were counted as a fraction of the total number of counted cells and expressed in percentage as mean ± SD for 3 independent experiments. Minimum total 1,000 cells were counted per each sample.
Growth inhibition, cell cycle arrest and cell death induced by depletion of hNatA
In a screen for phenotypes induced by hNatA depletion, we used cell lines derived from different types of thyroid cancer (anaplastic, follicular and papillary). Transfected cells were routinely checked for efficient knockdown of hNAA10 by qRT-PCR (Supporting Information Figure 1) and phenotypes were evaluated based on morphological changes, growth curves, cell cycle changes and verification of apoptosis. In Table 1, we present the diversity of phenotypes in hNatA-depleted cells: growth inhibition, G1 cell cycle arrest, apoptosis or absence of any phenotype. Interestingly, downregulation of hNatA caused mainly growth inhibition in all screened thyroid cell lines except for anaplastic carcinoma, ARO and follicular thyroid carcinoma, CGTH-W1, cell lines. Since the most severe changes after hNatA depletion were observed in 2 morphologically different anaplastic thyroid cell lines, 8305C and CAL-62, we used these cell lines for in-depth studies. As is observed from Table 1, the cell death phenotype mainly correlated with wild-type TP53, although an “apoptosis-like programmed cell death” also occurred in anaplastic cell lines 8305C and CAL62 with mutated TP53. The TP53 mutational status within its coding sequence was determined in all studied thyroid cell lines and identified mutations their localization and also predicted functional status of p53, are presented in Supporting Information Table I. Notably, the growth inhibitory phenotype was independent of TP53 mutational status, whereas apoptosis was linked to functional p53 (Table 1).
Table 1. hNatA knockdown phenotypes in human thyroid cell lines
Mechanisms of growth inhibition mediated by hNatA depletion in anaplastic thyroid carcinoma CAL-62
To investigate the mechanisms of growth inhibition and cell cycle arrest induced after loss of hNatA function, anaplastic carcinoma CAL-62 cells were used. Growth curves of hNatA knockdown cells demonstrated a strong negative effect on cell growth (Fig. 1a). There were also changes in cell morphology: round cell shape, blebs and ruffles formation at the cell surface and decrease in cell–cell contacts (Fig. 1b). This may indicate cytoskeleton reorganization, which in turn affects the cell surface. Furthermore, fluorescence-activated cell sorting (FACS) cell cycle analysis was performed on hNatA-depleted cells and revealed an increase in G0/G1 fraction, decrease in the S- and G2/M fractions, indicative of G1-arrest (Fig. 1c). Biochemical changes detected in hNatA-depleted cells by Western blotting suggested G1-arrest due to reduction of G1-cyclins (cyclin D1 and E), pointed at hypophosphorylation of Rb, and thus repression of E2F-mediated genes (cyclin A, B1, E, E2F-1 and Rb) as an underlying cause. Moreover, cyclin-dependent kinase inhibitor p27/Kip1 was upregulated in hNatA-depleted cells potentially contributing to the observed G1-arrest (Fig. 1d).
Mechanisms of growth inhibition mediated by hNatA depletion in anaplastic thyroid carcinoma 8305C
To elucidate the generality of the observed mechanisms of growth inhibition and cell cycle arrest induced by lack of hNatA function, the anaplastic thyroid carcinoma cell line 8305C, known for different cell morphology, growth and proliferation properties as compared to the CAL-62 cell line, was used. Growth curves of 8305C cells exhibited increasing inhibition of growth with time after hNatA knockdown (Fig. 2a). Inspection of hNatA-depleted 8305C cell cultures by phase contrast microscopy revealed changes in cell morphology leading to formation of round-shaped cells, refinement of cellular protrusions, loss of cell to cell contacts and increased detachment from growth support (Fig. 2b). Cell cycle analysis by FACS revealed changes in the histogram for 8305C hNatA-knockdown cells (affected bi-modal peak in G1-phase) indicating on events in G1 (Fig. 2c). However, hypertriploid karyotype (≈22% polyploidy) of 8305C cells prevented gating of cell cycle phases (for this reason percent distribution of cell cycle phases was not shown on histograms). Western blotting analysis demonstrated strong decrease in cyclin D1 and increase in p27/Kip1 possibly contributing to G1-arrest (Fig. 2d). These findings resembled those obtained with hNatA-depleted CAL-62 cells. Conversely, there were no changes in Rb phosphorylation and, thus, the effect on transcriptional repression of E2F-mediated genes cyclin A, B1, E, E2F-1 and Rb was absent (Fig. 2d).
Apoptosis-like programmed cell death associated with hNatA depletion
Since 8305C and CAL-62 hNatA-depleted cells were only weakly TUNEL-positive, although demonstrated morphological signs of cell death with late onset, another assay for a classic biomarker of apoptosis, Annexin V-labeling, was used. The assay was positive for both CAL-62 and 8305C hNatA-knockdown cells, showing presence of cells in early and late stages of apoptosis (Fig. 3a). Interestingly, biochemical markers of cell death (cleaved PARP, α-Fodrin) were also feebly marked both for 8305C and CAL62 (Fig. 3b). This may indicate an undergoing “apoptosis-like programmed cell death” due to signal pathway crosstalk.22
Sensitization of thyroid cancer cells to drug treatment by hNatA depletion
To elaborate the features of the thyroid cancer cells with respect to drug sensitization by hNatA depletion, we treated 2 anaplastic thyroid cell lines (8305C and CAL-62) and a follicular thyroid cell line (FTC-133) with 3 drugs (daunorubicin, KillerTRAIL and troglitazone) known for different mechanisms of drug action in combination with hNatA knockdown.23–27 Notably, each cell line following hNatA knockdown could be sensitized to only 1 of 3 drugs in a cell-type specific manner: 8305C responded to KillerTRAIL, CAL-62 to daunorubicin and FTC-133 to troglitazone (Figs. 4a–c). Synergistic effect of hNatA depletion with drug treatment was shown for CAL-62 cells in the daunorubicin dose range of 0.2–3.0 μM; and also for 8305C cells in the KillerTRAIL dose range of 2–100 ng/ml. hNatA depletion from FTC-133 cells demonstrated additive effect with troglitazone treatment in the dose range 2.8–22.5 μM.
hNatA depletion induces p53-dependent cell death in human cells with wild-type TP53
Since screening for hNatA-knockdown phenotypes was done among cancer thyroid cell lines harboring mutated TP53 except for thyroid carcinoma K1 and immortalized normal thyroid follicular cells without fully functional p53, another model system was used to study the apoptotic phenotype that was previously reported in hNatA-depleted cells16 and also weakly observed in the TP53 WT cells (Table 1). To shed light on the possible role of TP53 in apoptosis induction by hNatA depletion, human colon carcinoma HCT116 p53+/+ and p53−/− were used as cellular models. Comparative phenotypic observations of hNatA knockdown in HCT116 cell lines revealed induction of apoptosis in HCT116 p53+/+ and growth inhibition in HCT116 p53−/−. Indeed, FACS analysis of cell cycle changes demonstrated an increase of the sub-G0/G1 fraction (dead cells) only in HCT116 p53+/+ and not in HCT116 p53−/− cells (Fig. 5a). Furthermore, apoptosis in HCT116 p53+/+ hNatA-knockdown cells was verified qualitatively and quantitatively by a positive TUNEL assay (Figs. 5b and 5c). Western blotting analysis of HCT116 p53+/+ and p53−/− pointed at caspase-dependent cell death in p53+/+ cells (based on appearance of cleaved PARP and cleaved α-Fodrin, which both are caspase substrates) and indicated also an activation and stabilization of p53 in hNatA-depleted cells. Moreover, expression of p53 protein correlated with increased MDM2 and p14/ARF protein levels (Fig. 5d). Interestingly, we detected activation of p53 by phosphorylation at Ser15. To elaborate on the role of p53 in apoptosis induction, it was examined whether p53 was transcriptionally activated. Indeed, qRT-PCR data showed increased transcription of the p53-dependent proapoptotic genes: BAX, NOXA, PUMA, FAS and KILLER (Fig. 5e). Cotransfection of sihNAA10 and plasmid expressing hNaa10p rescued HCT116 p53+/+ cells from sihNAA10 induced apoptosis demonstrating the specificity of the observed phenotype (Fig. 5f).
Loss of hNatA function activates DNA-damage signaling
We then addressed the issue concerning p53-upstream events: whether checkpoint signaling is activated after hNatA depletion in HCT116 p53+/+ and p53−/− cells. Western-blot data demonstrated consistently the appearance of γH2A.X (activated H2A.X by phosphorylation at Ser139) and activated Chk2 (phosphorylated at Thr68) as a consequence of hNatA depletion regardless of cellular TP53 status (Fig. 6a). The γH2A.X/Chk2 module is usually activated after double-strand DNA breaks. To draw a direct or indirect link between Chk2 and activation of p53, phosphorylation of p53 at Ser15, Ser20 and Ser37 was analyzed. Increased p53 phosphorylation at Ser15 and to a lesser extent at Ser37 was observed. No change in p53 phosphorylation at Ser20 was detected as compared to positive control treated with the DNA-damaging drug daunorubicin (Fig. 6b). Nonetheless, our data indicate an activation of DNA-damage dependent signaling in hNatA-depleted cells and suggest that loss of hNatA activity generates events (like double-strand DNA breaks) resulting in DNA-damage signaling through γH2A.X and Chk2 and then through p53 at Ser15.
About 80% of soluble human proteins are N-terminally acetylated by 1 of 3 major Nα-terminal acetyltransferase complexes, hNatA, hNatB and hNatC, which differ in their subunit composition and substrate specificity. hNatA substrates represent the largest group, comprising proteins with Ala-, Ser-, Thr-, Gly-, Val- N termini.6, 20, 28, 29 Knockdown of the human Nat complexes induces several phenotypes: apoptosis and cell cycle arrest for hNatA16–18; cell cycle arrest for hNatB28, 30; apoptosis for hNatC.29 The apoptosis induced by hNatC knockdown was also p53-dependent similar to what we here found for hNatA knockdown, but in the case of hNatC knockdown, p53 was phosphorylated at Ser37 while hNatA knockdown stimulated phosphorylation of p53 at Ser15.29 This differential signaling probably reflects the involvement of specific functionally important substrates of the 2 distinct Nat complexes. hNatE (hNAA50) depletion in humans and fruitfly induces defects in sister chromatid cohesion and chromatin condensation.31–33
In search for hNatA knockdown phenotypes, we here depleted the catalytic subunit hNaa10p and investigated phenotypes in a variety of thyroid cancer cell lines representing different morphological features and grades of dedifferentiation (Table 1, Supporting Information Figure 2). Interestingly, 9 of 11 thyroid cancer cell lines demonstrated growth defects and 3 of 11 showed induction of apoptosis after loss of hNatA. Since growth defects and apoptosis may be directly linked to functional p53,34, 35 we sequenced the coding sequence of TP53 from all studied thyroid cell lines and checked functional status of mutated TP53 from IARC TP53 database R13 release.36 Analysis of TP53 mutations showed that all detected mutations led to loss of p53 transactivation. Indeed, the growth inhibitory phenotype was observed among thyroid cell lines independent of TP53 mutational status (Table 1).
Further exploration of mechanisms causing growth inhibition and cell cycle arrest on hNatA knockdown revealed several pathways involving (i) cyclin D1, (ii) Rb/E2F and (iii) p27/Kip1. All 3 pathways were affected in anaplastic thyroid carcinoma CAL-62 while only the cyclin D1 and p27/Kip1 pathways were altered in anaplastic thyroid carcinoma 8305C (Figs. 1 and 2). Also, Lim and colleagues found that hNatA RNAi in human lung cancer cells induced growth inhibition and G1-arrest, mediated by suppression of cyclin D1. Since the promoter activity of the CCND1 gene was repressed, and binding of the β-catenin/TCF4 transcription factor complex to the CCND1 promoter was negatively regulated in hNatA-depleted cells, they suggested that growth defects could be explained by an affected β-catenin pathway via the direct Nε-acetylation of β-catenin by hNaa10p.17 From our present data, we believe that growth inhibition and cell cycle arrest caused by a decrease in the cyclin D1 level may also be independent on β-catenin regulation. Analysis of endogenous β-catenin from CAL-62 and 8305C hNatA-depleted cells did not show any changes in β-catenin acetylation (Supporting Information Figure 3).
Another finding in CAL-62 cells with depleted hNatA was an affected Rb/E2F pathway resulting in decrease of E2F transcriptional activity. Interestingly, the Rb pathway is often inactivated in human tumors, leading to deregulated E2F activity and uncontrolled cell cycle progression.37 Conversely, hNatA depletion in CAL62 cells results in hypophosphorylation of Rb, and, thus, securing Rb-restricted control of cell cycle, mainly by regulating members of the E2F-transcription-factor family. The underlying mechanism of how hNatA depletion leads to dephosphorylation of Rb remains to be investigated, and may involve 1 or more hNatA substrates directly or indirectly acting upstream of Rb.
Accumulation of p27/Kip1 levels in hNatA-depleted cells is also in support of growth inhibition and cell cycle arrest. Increased p27/Kip1 levels prevent a passage through a restriction point in the G1-phase. Interestingly, studies in several tumor types indicate that p27/Kip1 expression levels have both prognostic and therapeutic implications,38 and elevated levels of p27/Kip1 correlate with apoptosis.39 However, the mechanism of p27/Kip1 upregulation in hNatA-depleted cells remains unknown and occurs, most likely, independently of Rb/E2F pathway.
Drug treatment of 3 selected thyroid cell lines with depleted hNatA revealed that sensitization to drugs is highly dependent on the cellular context. The 8305C cell line was specifically sensitized to KillerTRAIL, CAL-62 to daunorubicin and FTC-133 to troglitazone (Fig. 4). The mechanisms of action of daunorubicin, KillerTRAIL and troglitazone are very different from each other, although effects of treatment are connected with execution of apoptosis and/or growth inhibition. Daunorubicin possesses antimitotic and cytotoxic activity through several mechanisms: intercalation into DNA, generation of free radicals, direct membrane effects, inhibition of topoisomerase II and polymerase activity. As a result of induced DNA damage, p53 is activated and guides cells through the intrinsic apoptotic pathway.25KillerTRAIL triggers apoptosis in cells through the extrinsic apoptotic pathway, regardless of TP53 status. By coupling with death receptors DR4/DR5, KillerTRAIL leads to formation of death-inducing signal complex (DISC) and autoactivation of caspase 8. Depending on cell type, caspase 8 activation is either sufficient to induce apoptosis through activating the downstream effector caspases, or the amplification loop through the mitochondrial pathway is needed.26, 27 Troglitazone, a ligand to peroxisome proliferator-activated receptor-gamma (PPAR-γ), inhibits cell growth and induces apoptosis by several mechanisms including direct activation of PPAR-γ pathway, interconnected activation of ERK and p38, and by the reduction of SAPK/JNK activity.23, 24 Our present data on drug sensitization indicate that lack of Nα-acetylation by hNatA causes complex changes in interconnected pathways, which can be intensified either by DNA damage, by death receptors or by nuclear hormone receptors signaling.
Thyroid anaplastic carcinomas 8305C and CAL-62, having the strongest growth defects after hNatA depletion, demonstrated additionally apoptosis-like programmed cell death with late onset. Based on weakly positive TUNEL data (display less compact chromatin condensation), Annexin V positive staining (detection of externalized phosphatidylserine) and cleavage of caspase substrates (cleaved PARP and α-Fodrin), we considered the cell death occurring in CAL-62 and 8305C as a result of interconnection of signaling pathways.40
To elaborate whether TP53 has a direct link to classic apoptosis induced after loss of hNatA, we used a nonthyroid cancer model: the 2 human colon carcinoma cell lines HCT116 p53+/+ and HCT116 p53−/−. hNatA depletion in HCT116 p53+/+ cells leads to cell death characterized by increase in sub-G0/G1 fraction representing nuclear fragmentation, double-stranded DNA breaks and cleavage of PARP and α-Fodrin (Fig. 5), in agreement with findings in HeLa cells16 and HEp-2 cells.41 Oppositely, growth inhibition rather than apoptosis was induced in HCT116 p53−/− cells after hNatA knockdown, indicating that functional p53 is essential for induction of apoptosis. Moreover, our data (Figs. 5 and 6) suggest that this effect is mediated through the transcriptional activity of p53,42–44 since hNatA knockdown stabilized the protein level of p53, increased phosphorylation of p53 at Ser15, least at Ser37, and induced the transcription of p53-dependent apoptosis effector genes.45 In addition, in its active state p53 has an enhanced transactivation of Mdm2 and correlates positively with p14/ARF expression indicating on an active p53-MDM2-p14 feedback loop in hNatA-depleted HCT116 p53+/+ cells.
Since hNatA-depleted cells are prone either to growth inhibition, cell cycle arrest or apoptosis, which are the effector mechanisms of the DNA-damage response network, the next issue was to examine DNA-damage sensing and signaling pathway. Indeed, lack of acetylation by hNatA activated H2A.X and Chk2 in both HCT116 cell lines independent of TP53 status (Fig. 6). The exact mechanisms activating the DNA-damage response network in hNatA-depleted cells even without exogenous DNA damage remains unclear. However, we may speculate based on data available from yeast studies: Orc1p and Sir3p, verified substrates of yNatA46, 47 are required for telomeric silencing. Disruption of yNatA-mediated acetylation diminishes Orc1p and Sir3p acetylation, which in turn abrogates telomeric silencing, and thus, telomere uncapping may be recognized as double strand breaks and further activation of DNA-damage response network may occur.48 Moreover, the growth defects on hydroxyurea, methylsulfonate, o-dinitrobenzene and 4-nitroquinoline containing media of yNatA mutant strains may indicate on DNA repair or DNA replication deficiency when lacking yNatA.6 Since DNA-damage sensing and signaling pathways are tightly tied to cell proliferation, cell cycle arrest, cellular senescence and apoptosis, these 2 models represent potential mechanisms of the hNatA depletion phenotype. Recently, it was reported that hNaa10p was a caspase modulator with a proapoptotic function in DNA-damage induced apoptosis. HeLa cells were less prone to apoptosis under doxorubicin treatment when hNAA10 was depleted.49 In contrast, in our current and previous experiments, various cell types were always sensitized or alternatively were not affected by hNatA knockdown when combined with different drugs.16, 41 This may indicate an ability of hNatA to direct different cellular fates depending on cellular context.
In conclusion, we have provided evidence for several distinct phenotypes associated with hNatA knockdown in human cancer cell lines (Fig. 7). This complexity between cell lines and within the same cell type probably reflects the great number of hNatA substrates6 in human cells of which several are likely to be functionally important. Other important factors appear to be differences in cellular states and genetic background between the cell lines. Growth inhibition is linked to several p53-independent pathways including cyclin D1 inhibition, dephosphorylation of Rb and accumulation of p27/Kip1. Induction of apoptosis depends on functional p53 and probably occurs via activation of a γH2A.X-Chk2-p53 signaling pathway (Fig. 7). An important future task will be to elucidate the exact pathways including the identification of the key hNatA substrates involved.