Polyploid endothelial cells are found in aged and atherosclerotic arteries. However, whether increased chromosome content has an impact on endothelial cell function is unknown. We show here that human aortic endothelial cells become tetraploid as they approach replicative senescence. Furthermore, accumulation of tetraploid endothelial cells was accelerated during growth in high glucose. Interestingly, induction of polyploidy was completely prevented by modest overexpression of the NAD+ regenerating enzyme, nicotinamide phosphoribosyltransferase (Nampt). To determine the impact of polyploidy on endothelial cell function, independent of replicative senescence, we induced tetraploidy using the spindle poison, nocodazole. Global gene expression analyses of tetraploid endothelial cells revealed a dysfunctional phenotype characterized by a cell cycle arrest profile (decreased CCNE2/A2, RBL1, BUB1B; increased CDKN1A) and increased expression of genes involved in inflammation (IL32, TNFRSF21/10C, PTGS1) and extracellular matrix remodeling (COL5A1, FN1, MMP10/14). The protection from polyploidy conferred by Nampt was not associated with enhanced poly(ADP-ribose) polymerase-1 or sirtuin (SIRT) 2 activity, but with increased SIRT1 activity, which reduced cellular reactive oxygen species and the associated oxidative stress stimulus for the induction of polyploidy. We conclude that human aortic endothelial cells are prone to chromosome duplication that, in and of itself, can induce characteristics of endothelial dysfunction. Moreover, the emergence of polyploid endothelial cells during replicative aging and glucose overload can be prevented by optimizing the Nampt-SIRT1 axis.
- endothelial cells
- oxidative stress
vascular endothelial cell (EC) dysfunction is fundamental to the pathogenesis of vascular disease (25, 39). Dysfunctional ECs disrupt vascular function through multiple mechanisms, including increased expression of inflammatory molecules and coagulation factors, altered extracellular matrix deposition, and decreased expression of endothelial nitric oxide synthase (eNOS) (39). Although there may be several drivers of EC dysfunction, the intracellular accumulation of reactive oxygen species (ROS) has emerged as a key destructive force. The importance of ROS accumulation is highlighted by its relationship to risk factors for vascular disease, including aging and diabetes mellitus (12, 25).
A potential mechanistic link between ROS accumulation and EC dysfunction may involve cellular senescence. Importantly, oxidative stress appears to be an inducer of EC senescence (30, 31). Senescence in cultured ECs is accompanied by a cluster of cell biological changes, including a flattened and spread morphology, G1 arrest, senescence-associated β-galactosidase (SA β-gal) activity, accumulation of ROS, and oxidative damage (44, 47). Senescent ECs can persist in vessels, where they are proposed to contribute to vascular dysfunction through decreased nitric oxide production, decreased reparative capacity, and increased inflammatory potential (5, 10).
Another feature of ECs in aged and atherosclerotic human arteries is a predisposition to become polyploid. Studies of human aortic ECs (HAECs), examined in isolated cultures and in situ, have revealed that the occurrence of polyploid cells significantly increases with patient age and severity of atherosclerosis (1, 33, 35, 43). In fact, polyploid cells may account for 15–40% of ECs from aged and atherosclerotic arteries in humans (33, 43). However, the consequences of this increase in chromosome content are unclear. Interestingly, we recently found that replicative senescence in cultured HAECs was associated with tetraploidy (4). EC polyploidization thus appears to be a biomarker of EC senescence. However, whether the acquisition of polyploidy, per se, is functionally linked to the development of senescence, or other modes of EC dysfunction, is unknown. It is also unknown whether EC polyploidy is an irrevocable or preventable consequence of replicative aging.
NAD+ biosynthetic pathways have emerged as important regulators of cardiovascular function through their ability to modulate sirtuin (SIRT) activity (3). We recently reported that modest overexpression of nicotinamide phosphoribosyltransferase (Nampt), the rate-limiting enzyme for NAD+ salvage synthesis, enabled HAECs to resist oxidative stress, as induced by replicative aging and high glucose, in a SIRT1-dependent manner (4). However, the role of NAD+ regeneration in regulating EC ploidy, which may also be related to oxidative stress (6, 13, 24), is unknown.
Here, we report that EC polyploidy is not only a prominent consequence of replicative aging and chronic glucose exposure, but, in and of itself, drives a dysfunctional EC phenotype. We further show that this risk factor-related increase in chromosome content can be prevented by enhancing NAD+ regeneration through Nampt expression.
MATERIALS AND METHODS
Normal HAECs (Lonza, Shawinigan, QC), or HAECs stably overexpressing either enhanced green fluorescent protein (eGFP) or eGFP-Nampt [generated as previously described (4)], were grown in medium 199 (Invitrogen, Burlington, ON) supplemented with EGM-2 SingleQuots (Lonza). Total Nampt content and functional activity in eGFP-Nampt expressing ECs were increased 1.6- and 1.8-fold, respectively, compared with eGFP expressing control cells (4).
To quantify replication, cells were plated at 1,000 cells/cm2, and growth medium was changed every 2–3 days until monolayers reached 80% confluence. At each subculture, cells were harvested and counted, and the number of population doublings was calculated according to the formula log10 (number of cells harvested) − log10 (number of cells seeded)/log10(2). Growth rates were calculated by dividing the number of population doublings by the number of days in culture. Replicative lifespan was considered complete when population doublings remained at zero for a minimum of 14 days and >90% of cells were in G1 arrest.
Chemical induction of tetraploidy was achieved as previously described (20), with modifications. Briefly, young [subculture 6, approximate cumulative population doubling (cpd) of 13] HAECs were incubated for 48 h with 200 nM nocodazole, followed by 24-h recovery in fresh growth medium. For each endpoint assay involving nocodazole-treated cells, the acquisition of tetraploidy was confirmed by FACS analyses of DNA content, as described below.
Cellular DNA content was assessed in cells fixed with 50% ethanol by staining with propidium iodide/RNase staining buffer (BD Biosciences, Bedford, MA). Cell cycle and DNA ploidy were analyzed using ModFit software (Verity Software House, Topsham, ME).
Relative cellular ROS content was assessed by live cell staining using 5 μM 5-(and-6) chloromethyl-2′,7′-dichlorodihydrofluorescin diacetate, acetyl ester, according to the protocol provided by the manufacturer. Mean fluorescent signals were determined by flow cytometry (104 cells/sample) using a FACSCalibur (BD Biosciences) at the London Regional Flow Cytometry Facility (Robarts Research Institute, London, Ontario, Canada).
SA β-gal activity.
Subconfluent HAECs were fixed using 2% formaldehyde/0.2% glutaraldehyde and stained for β-gal activity at pH 6.0, as previously described (45). Photomicrographs were acquired on an Olympus IX51 inverted microscope, and SA β-gal positive (blue) cells were counted for three fields of view (×40 magnification) per well.
Endothelial tube formation.
HAECs were seeded at 75,000 cells/well in 24-well cluster plates coated with 150 μl of Matrigel basement membrane matrix (BD Biosciences, Bedford, MA). Photomicrographs were acquired on an Olympus IX51 inverted microscope and tube branch points (21) for three fields of view (×100 magnification) were counted per well.
Biotinylated complimentary RNA was prepared from 500 ng of total RNA, from three biological replicates of control and nocodazole-treated HAECs, using the Affymetrix GeneChip 3′ IVT Express kit (Affymetrix, Santa Clara, CA). Labeled complimentary RNA (10 μg) was hybridized to Human Genome 133 Plus 2.0 GeneChips (Affymetrix, Santa Clara, CA), as described in the Affymetrix Technical Analysis Manual. GeneChips were stained with streptavidin-phycoerythrin, followed by an antibody solution and a second streptavidin-phycoerythrin solution, with all liquid handling performed by a GeneChip Fluidics Station 450. GeneChips were scanned with the Affymetrix GeneChip Scanner 3000 (Affymetrix, Santa Clara, CA) controlled by Command Console software. All GeneChip processing was performed at the London Regional Genomics Centre (Robarts Research Institute, London, Ontario, Canada; http://www.lrgc.ca).
Microarray CEL file data were imported into Partek Genomics Suite (http://www.partek.com/partekgs) using GC-RMA options. Statistical significance between control and treatment samples was determined by ANOVA (P < 0.01), and data were filtered for changes in expression exceeding 2.0-fold. Gene ontology category overrepresentation analyses were performed on lists of down- and upregulated transcripts using Expression Analysis Systematic Explorer.
Differential expression of genes of interest were validated by quantitative real-time PCR using commercially available TaqMan assays, corresponding to Affymetrix probe sets and to GAPDH (Applied Biosystems, Foster City, CA).
Immunoblotting was performed after SDS-PAGE resolution of whole cell lysates prepared using standard RIPA buffer. eNOS, ICAM-1, Nampt, and p21 were detected in 40 μg of total protein using polyclonal anti-eNOS (1:1,000) (Millipore, Temecula, CA), anti-ICAM-1 (1:1,000) (Cell Signaling Technology, Danvers, MA), and anti-pre-B cell colony enhancing factor (1:5,000) (Bethyl Laboratories, Montgomery, TX), and monoclonal anti-p21 (1:2,000), respectively. α-Tubulin was probed as a loading control for all immunoblots using monoclonal antibody B-5–1-1 (Sigma, Oakville, ON).
Cellular SIRT1 and SIRT2 activities were assessed using Fluor de Lys-SIRT1 and -SIRT2 acetylated substrates (Biomol, Plymouth Meeting, PA), as previously described (7).
For all experiments, data are presented as means ± SE. Statistical differences (P < 0.05) were determined by ANOVA with post hoc t-tests and Student's t-tests, as appropriate.
EC tetraploidy increases with replicative age and glucose overload and is prevented by Nampt.
By FACS assessment of DNA content, we determined that control ECs (eGFP) became increasingly tetraploid with replicative age. In fact, after 130 days of culture, almost 100% of cells had assumed a tetraploid state, in association with the acquisition of replicative senescence (Fig. 1A). For all populations, replicative life span was considered complete when population doublings remained at zero for a minimum of 14 days, and >90% of cells were in G1 arrest.
Because polyploidization in other cell types has been observed in response to oxidative stress (6, 13, 24), we determined whether Nampt, which limits cellular ROS accumulation in ECs (4), could prevent replicative aging-induced EC tetraploidy. Using our previously established HAEC populations in which eGFP or eGFP-Nampt are stably overexpressed (4), we found that modest Nampt overexpression (1.6-fold) completely prevented the acquisition of tetraploidy (Fig. 1A). Nonetheless, eGFP-Nampt cells did achieve replicative senescence (Fig. 1). We further determined whether glucose overload, an inducer of oxidative stress relevant to vascular disease, could exacerbate polyploidization. Chronic growth in high glucose (30 mM) markedly increased the rate of EC polyploidization. Again, this effect was completely abrogated by overexpression of Nampt (Fig. 1B). All EC populations displayed replicative senescence at the final point of their respective tetraploidy curves.
Nocodazole exposure induces EC tetraploidy and some, but not all, characteristics of senescence.
To study EC polyploidization as an independent event, we developed a protocol for rapid and efficient induction of tetraploidy in early passage (subculture 6) HAECs using the microtubule disrupter nocodazole. Prolonged exposure to nocodazole (24–72 h) impairs chromosome segregation and cytokinesis, with subsequent mitotic slippage that results in hyperploid progeny that generally escape apoptosis (8). To eliminate any continued effects of nocodazole on microtubule-dependent processes, we studied cells incubated for 48 h with the drug, followed by a 24-h washout period. Nocodazole is known to be rapidly removed from cells after washout, and cells allowed to recover for as little as 30 min display a pattern of microtubules comparable to untreated cells (22). After 24-h recovery, cells were completely adherent, and we observed no increase in apoptosis.
Interestingly, ECs exposed to nocodazole developed a spread morphology, enlarged nuclei, and tetraploidy (Fig. 2, A and B), characteristics of EC senescence (44, 47). Furthermore, these cells displayed significantly decreased growth rates (Fig. 2C), consistent with the fact that 90.1 ± 2.3% of cells were in G1 of the cell cycle, and substantial accumulation of cellular ROS (Fig. 2D). Thus the induction of tetraploidy through the selective failure of cytokinesis induced several features of EC senescence. However, nocodazole-treated cells did not exhibit SA β-gal activity (Fig. 3A), and Nampt expression was unaffected (Fig. 3B). We have previously shown that Nampt protein content and activity are significantly decreased with increasing replicative age in both ECs and vascular smooth muscle cells (4, 45). These data suggest a dissociation of these particular features of senescence from other attributes. More importantly, this model of tetraploidy allowed us to further assess the effects of polyploidy on EC function and gene expression, without the confounding cumulative effects of replicative aging.
Tetraploidy impairs select EC functions.
We found that tetraploidy modestly decreased eNOS expression and increased ICAM-1 expression (Fig. 4A). Combined with the impaired proliferative capacity we observed in nocodazole-treated cells (Fig. 2C), these data suggest that induction of polyploidy impairs EC function to a certain extent. Since angiogenic activity requires both EC proliferation and the ability to form capillary tubes, we performed tube formation assays on growth factor-replete Matrigel basement membrane matrix. We found that tetraploidy induced either by replicative aging or exposure to nocodazole did not affect EC tube network formation (Fig. 4B). Consistent with this, SIRT1 activity, which is critically important for EC angiogenic function (32), was not impaired in nocodazole-treated tetraploid cells (1.00 ± 0.28 vs. 1.42 ± 0.19 relative units, P = 0.28). Thus, although acquisition of polyploidy imparted select attributes of EC dysfunction, it was not an inducer of comprehensive EC dysunction.
Tetraploidy increases the expression of genes involved in inflammation and extracellular matrix remodeling.
To further assess the effects of tetraploidy on EC gene expression, we undertook a global, unbiased comparison between control and nocodazole-exposed human ECs using Affymetrix high-density microarray analyses. Mean percentages of tetraploid cells for control and nocodazole-incubated populations from which total RNA were isolated were 20.3 ± 4.7 and 80.2 ± 2.2, respectively. Consistent with this 60% increase in tetraploidy, total RNA content per cell increased from 8.4 to 12.4 pg after nocodazole treatment. We found that 559 transcripts were affected by nocodazole-induced tetraploidy (>2.0-fold, P < 0.01), with 354 genes downregulated and 205 genes upregulated. Complete lists of the down- and upregulated genes are included in Supplemental Tables 1 and 2, respectively. (The online version of this article contains supplemental data.)
Consistent with the reduction in growth rate that we observed in tetraploid cells (Fig. 2C), gene ontology category overrepresentation analyses revealed that the most statistically significant categories corresponding to genes downregulated by nocodazole-induced tetraploidy were related to the cell cycle (Table 1). Interestingly, the most significant categories corresponding to genes upregulated in tetraploid cells were predominantly related to extracellular matrix remodeling (Table 1).
We further analyzed the differentially expressed transcripts for genes known to be associated with biological processes involved in vascular pathology. We identified 27 genes with altered expression of at least twofold, and we verified differential expression in a sample of nine of these, by real-time PCR or immunoblotting (Table 2). Several genes implicated in cell cycle control, cellular senescence, and organismal aging were differentially expressed in tetraploid cells, including BUB1B, CDKN1A (p21), and a series of E2F target genes (CCNA2, CCNE2, CDC2, RBL1) (2, 16). Aurora kinases (AURKA, AURKB) and a Polo-like kinase (PLK4), which are critical for proper chromosome segregation during mitosis (42), were also downregulated in tetraploid cells (Table 2).
Consistent with our observation of increased ICAM-1, we also found that several genes associated with inflammatory processes were upregulated in tetraploid cells (Table 2). The interferon-inducible genes, interleukin-32 (IL32) and interferon-α inducible protein 27 (IFI27), can induce monocyte-to-macrophage differentiation (27) and apoptosis (34), respectively. Several members of the tumor necrosis factor (TNF) receptor and ligand superfamilies (FAS, TNFRSF21, TNFRSF10C, TNFSF4) were also induced, as was prostaglandin G/H synthase (PTGS1), the enzyme responsible for thromboxane A2 production. Both TNF signaling and thromboxane A2 can increase the cell surface expression of adhesion molecules in ECs (26, 49). In contrast, the inhibitor of transforming growth factor type-β signaling, SMURF2, was induced, which would be expected to suppress anti-inflammatory pathways (14).
Finally, we noted that several vascular remodeling genes were significantly upregulated in tetraploid cells (Table 2). Endothelial expression of the Notch ligand Jagged 1 (JAG1) has recently been implicated in the differentiation of neighboring vascular smooth muscle cells (17) and was significantly induced in tetraploid ECs. Genes involved in the generation and turnover of extracellular matrix proteins were also significantly upregulated, including matrix metallopeptidases (MMP10, MMP14), proline 4-hydroxylase α(II), collagens (COL27A1, COL5A1, COL4A1), and fibronectin (FN1). Interestingly, the angiostatic genes thrombospondin-1 (THSD4) (41) and filamin A interacting protein 1-like (FILIP1L) (23) were also induced. These changes in gene expression are consistent with our functional data, in that tetraploid ECs retained the ability to form tubes on Matrigel (Fig. 4B), a process that requires extracellular matrix remodeling, while proliferative capacity was impaired (Fig. 2C).
Taken together, the global gene expression pattern revealed that tetraploid ECs have what can be viewed as dysfunctional phenotype, with a profile of decreased proliferative potential, proinflammatory attributes, and modified extracellular matrix remodeling.
Prevention of tetraploidy by Nampt does not depend on PARP-1 or SIRT2, but is associated with increased SIRT1 activity.
Given the shift in EC gene expression toward a dysfunctional phenotype upon polyploidization, understanding the role of Nampt in preventing tetraploidy became an important objective. Thus we investigated the mechanism whereby Nampt overexpression prevented tetraploidy during replicative aging. Our previous observations of SIRT1-dependent reductions in cellular ROS accumulation in Nampt-overexpressing ECs (4) suggested that Nampt likely prevents aging-associated polyploidization through enhanced SIRT1 activity. However, two other NAD+-dependent enzymes known to affect DNA ploidy, PARP-1 and SIRT2, could also be responsible for this striking protective effect.
PARP-1 interacts with kinetochore and spindle checkpoint proteins (36), and embryonic fibroblasts from PARP-1−/− mice exhibit increased tetraploidy (15). Thus we compared the effects of Nampt inhibition and PARP-1 inhibition on polyploidization. The Nampt inhibitor, FK-866, indirectly inhibits SIRT1 activity by virtue of the dependence of SIRT1 on a continual supply of NAD+ (18). Long-term growth of ECs in the presence of the Nampt inhibitor, FK-866, resulted in accelerated accumulation of tetraploid cells (Fig. 5A). However, consistent with previous observations in embryonic fibroblasts (38), the PARP-1 inhibitor, DPQ, did not induce tetraploidy (Fig. 5A).
The tubulin deacetylase, SIRT2 (28), has recently been described as a regulator of cell cycle progression (9, 29) and a mitotic checkpoint protein that prevents chromosomal instability (20). In fact, modest overexpression of wild-type, but not catalytically inactive, SIRT2 prevents polyploidization in response to mitotic stress (20). We found, however, that, unlike SIRT1 activity, SIRT2 activity was unchanged in Nampt-overexpressing ECs compared with control cells (Fig. 5B), suggesting that Nampt did not prevent replicative aging-induced polyploidization through globally enhanced SIRT2 activity.
Polyploid ECs have been observed in humans in association with aging and atherosclerosis (1, 33, 35, 43). EC polyploidy has also been associated with senescence (1, 4, 47); however, the functional consequences of polyploidization of ECs, if any, was unknown. Our data suggest that EC tetraploidy, although not required for EC senescence, can precede terminal growth arrest and itself imparts characteristics of EC dysfunction. By separating EC polyploidization from replicative aging, we established that chromosome duplication in ECs imparts a dysfunctional phenotype, characterized by decreased proliferative potential, proinflammatory attributes, and modified extracellular matrix remodeling. We further showed that EC polyploidization is not an irrevocable consequence of replicative aging of ECs, but can be prevented by modest overexpression of the NAD+-regenerating enzyme, Nampt.
We found that, in ECs, as has been observed in other cell types (13, 48), a predominant feature of polyploidy was reduced proliferation, whether tetraploidy was acquired through serial subculture or through treatment with nocodazole. Consistent with this, microarray analyses revealed that tetraploid ECs downregulated several genes involved in cell cycle control, including BUB1B and a series of E2F target genes, and upregulated CDKN1A (p21). These changes in gene expression have also been linked to cellular senescence (2, 16), highlighting the potential relationship between polyploidization and the onset of senescence. Importantly however, nocodazole-induced tetraploidy was independent of replicative aging, suggesting that tetraploidy, per se, can inhibit EC proliferation.
The upregulation of genes involved in inflammation and extracellular matrix remodeling observed in tetraploid ECs correlated with characteristics previously identified in ECs undergoing replicative senescence, including increased ICAM-1, fibronectin and collagen expression (37), and increased sensitivity to TNF-α (19). However, our microarray analyses revealed additional increases in potentially proinflammatory genes, including IL-32 (27), prostaglandin G/H synthase (26), and SMURF2 (14), and increases in several matrix metallopeptidases. Furthermore, we observed upregulation of the angiostatic molecules thrombospondin-1 (41), filamin A interacting protein 1-like (23), and collagen IV, a degradation fragment of which is the endogenous inhibitor of angiogenesis, tumstatin (40). For the most part, these factors restrain EC proliferation rather than migration (23, 40), which is consistent with our observations of reduced growth rates and unaltered capillary tube formation in tetraploid ECs. Previous studies have suggested that the accumulation of tetraploid cells is a protective response to genomic instability, preventing gross chromosomal aneuploidy (11). However, our data suggest that, for ECs, tetraploidy is itself deleterious, imparting diverse, harmful changes in gene expression.
Emerging evidence suggests that oxidative stress plays a key role in the development of both premature and replicative senescence in ECs (4, 30, 31, 44, 46), and in the development of polyploidy in a variety of cell types (6, 13, 24). We found that EC polyploidization preceded, but was not required for, replicative senescence and was accelerated in high glucose, a condition known to induce significant oxidative stress. Thus, as has been suggested for vascular smooth muscle cells (24, 48), excessive ROS accumulation during replicative aging and, in our system, during glucose overload, likely induces polyploidy associated with senescence. The molecular events linking oxidative stress and polyploidization are not well defined, but may involve bypass of the spindle checkpoint mechanism (6). Recent studies indicate that the NAD+-dependent enzymes PARP-1 and SIRT2 function in the mitotic checkpoint to ensure proper chromosome segregation (15, 20). However, although tetraploidy was completely prevented in Nampt-overexpressing ECs, we found no evidence to suggest that PARP-1 or SIRT2 activities were responsible for this remarkable effect. SIRT1 activity is, however, significantly increased in Nampt-overexpressing ECs, and we previously linked this increase to reduced ROS accumulation, through a metabolic shift toward energy generation through aerobic glycolysis (4). Thus it is likely that modest overexpression of Nampt reduces oxidative stress over the entire replicative lifespan of an EC, in effect eliminating a dominant stimulus for polyploidization.
In summary, human ECs are predisposed to duplicate their chromosomal content as they age and when they are subjected to glucose-mediated stress. Moreover, the acquisition of tetraploidy, independent of replicative aging, induces dysfunctional EC characteristics. Ensuring optimal expression of the NAD+-regenerating enzyme, Nampt, in HAECs preserves genomic stability, by preventing polyploidization during replicative aging and glucose overload. As previously reported, our culture system generates prolonged oxidative stress and reduced NAD+-to-NADH concentration ratios (4), conditions that approximate those observed during the stress of chronic vascular diseases. Thus we propose that maintenance of EC NAD+ levels could prevent the potentially deleterious accumulation of polyploidy ECs associated with aging-related vascular disease.
Grant support was from the Heart and Stroke Foundation of Ontario (T5675), the Canadian Institutes of Health Research (FRN11715), and the Krembil Foundation. J. G. Pickering is a Career Investigator of the Heart and Stroke Foundation of Ontario. N. M. Borradaile was funded by a Postdoctoral Fellowship Award from the Canadian Diabetes Association.
No conflicts of interest are declared by the author(s).
We thank David Carter of the London Regional Genomics Centre for assistance with statistical analyses of microarray data, and Oula Akawi for technical advice regarding quantitative real-time PCR.
- Copyright © 2010 the American Physiological Society