Mechanisms of endothelial senescence


  • Jorge D. Erusalimsky,

    1. Cardiff School of Health Sciences, University of Wales Institute, Cardiff CF5 2YB, UK and the Wolfson Institute for Biomedical Research, University College London, London WC1E 6BT, UK
    Search for more papers by this author
  • Chris Skene

    1. Cardiff School of Health Sciences, University of Wales Institute, Cardiff CF5 2YB, UK and the Wolfson Institute for Biomedical Research, University College London, London WC1E 6BT, UK
    Search for more papers by this author

Corresponding author J. D. Erusalimsky: Cardiff School of Health Sciences, University of Wales Institute Cardiff, Western Avenue, Cardiff CF5 2YB, UK. Email:


When endothelial cells from different vascular beds are grown in culture they show a limited capacity to divide, eventually entering into a permanent and phenotypically distinctive non-dividing state referred to as ‘replicative senescence’. Replicative senescence is thought to result from progressive shortening of telomeric DNA and consequent telomere dysfunction. More recently, it has been realised that senescence can also be induced by a variety of insults, including those causing intracellular oxidative stress. In this report, we review evidence for the occurrence of endothelial cell senescence in vivo. We will also examine the causes, mechanisms and regulation of this process as they emerge from our studies in cell culture, focusing in particular on the roles of oxidative stress, telomerase, growth factors and nitric oxide.


Cellular senescence was first described in the nineteen sixties in cultures of normal skin fibroblasts by Hayflick & Moorhead (1961). Contrary to the prevailing belief of the time, these investigators observed that upon serial subcultivation the cells would stop dividing after a finite number of population doublings. They also showed that cells from older donors underwent fewer divisions than those derived from younger ones (Hayflick, 1965). On the basis of these findings, they proposed that somatic cells have an intrinsic limited replicative capacity and that this property represented a manifestation of ageing at the cellular level. These initial observations gave rise to the ‘replicative senescence theory of ageing’. According to this theory, senescent cells accumulate with age in tissues in which cell proliferation continues throughout the entire lifespan of the organism and, owing to their altered phenotype, contribute to the loss of tissue homeostasis and to the development of age-associated pathologies (Campisi, 1996). A molecular mechanism that could account for the occurrence of replicative senescence emerged in the nineteen nineties, when work by a number of laboratories gave rise to the ‘telomere hypothesis of senescence’ (reviewed by Greider, 1998). Telomeres are the physical ends of chromosomes. In mammalian cells, they consist of a repeated DNA sequence (TTAGGG) that extends over a length of several thousand base pairs (bp) and associates to an array of dedicated telomere binding proteins. Synthesis of telomeric DNA requires the presence of a specialized reverse transcriptase called telomerase, which uses as a template its own RNA subunit to catalyse the addition of TTAGGG repeats to the 3′ end of the DNA chain. The telomere hypothesis states that in cells that lack telomerase, telomeric DNA shortens by 25–100 bp with each round of cell division as a consequence of the ‘end replication problem’, i.e. the inability of DNA polymerase to replicate the end of the lagging strand, with senescence occurring when telomere length reaches a critical point. A number of other studies have subsequently suggested that telomere shortening brings about the loss of telomere functional integrity and that this activates a DNA damage check point response which halts the cell cycle permanently (d'Adda di Fagagna et al. 2003; von Zglinicki et al. 2005). Likewise, senescence can result from several other forms of chromosomal disruption, such as chromatin decondensation caused by inhibition of histone deacetylases and DNA damage induced by radiation, oxidizing compounds, alkylating agents and drugs that generate double strand breaks (reviewed by von Zglinicki et al. 2005). Senescence induced by such insults occurs without the need for extensive cell proliferation, is generally telomere independent and has a much more rapid onset; hence it is often termed ‘stress-induced premature senescence’ (Toussaint et al. 2000). Telomere-independent senescence can also be induced by persistent mitogenic stimulation or by oncogene activation, highlighting the concept that senescence might have evolved as a barrier to cell transformation (Yaswen & Campisi, 2007).

Evidence that endothelial cell senescence occurs in vivo

Cell turnover in the endothelium historically has been considered to be very low (Schwartz et al. 1980). For this reason, and despite some early evidence to the contrary (Wright, 1968; Caplan & Schwartz, 1973; Repin et al. 1984; Burrig, 1991), until recently the idea that endothelial cell senescence might occur in vivo has been taken with a considerable degree of scepticism.

In 1995, the discovery of a histochemical marker of senescence, senescence-associated β-galactosidase (SA-β-gal), which detected senescent fibroblasts and keratinocytes in skin biopsies of old people (Dimri et al. 1995), revived the replicative senescence hypothesis and raised hopes of identifying senescent cells in other human tissues. Further work from our laboratory demonstrated that senescent cells become engorged with lysosomes, a feature which explains why they stain positive for β-galactosidase (Kurz et al. 2000). Using SA-β-gal as a marker, our laboratory set out to demonstrate that senescence also occurs in vascular cells in vivo. We chose balloon endothelial denudation of the rabbit carotid artery as our model, since this injury provokes endothelial and smooth muscle cell proliferation, which we speculated might result in cell senescence. Six weeks after a single denudation, we found SA-β-gal-positive cells in both the neointima and the media. A second denudation resulted in a marked acceleration in the accumulation of senescent cells, and immunohistochemical analysis identified these as endothelial and vascular smooth muscle cells (Fenton et al. 2001). In this work, we postulated that vascular cell senescence may contribute to the pathophysiological processes underlying atherogenesis and postangioplasty re-stenosis, and also that SA-β-gal could be used to investigate the presence of senescent cells in atherosclerotic lesions. Following our work, other laboratories have used SA-β-gal to demonstrate the presence of senescent endothelial cells overlying atherosclerotic plaques of human aorta and coronary arteries (Vasile et al. 2001; Minamino et al. 2002) and in the aortae of diabetic rats (Chen et al. 2002).

Other studies have shown that telomeres in the endothelium shorten with age and that this erosion is more pronounced in atherosclerosis-prone areas (Chang & Harley, 1995; Okuda et al. 2000; Aviv et al. 2001). Furthermore, a study examining the relationship between telomere length and coronary artery disease found that telomeres in endothelial cells derived from diseased portions of arteries were shorter than those from non-diseased regions (Ogami et al. 2004).

The role of oxidative stress in endothelial cell senescence

The vasculature is chronically exposed to a variety of oxidative burdens, including oxidative metabolites released from activated phagocytes, modified lipoproteins and various types of reactive oxygen species generated by vascular cells themselves (Madamanchi et al. 2005). A substantial body of evidence indicates that oxidative stress can induce or accelerate the development of cellular senescence. In some studies, the early onset of senescence has been attributed to accelerated telomere attrition, probably resulting from the generation of single strand breaks in the telomeric DNA (von Zglinicki, 2002). Other studies claim that stress-induced premature senescence may not be related to telomere damage (Chen et al. 2001). It is not clear, however, whether these differences are due to the type of oxidative insult to which cells were subjected, the degree of antioxidant protection that the cells could exert, or some other difference. In the case of endothelial cells, an association between oxidative stress, accelerated telomere shortening and senescence has been suggested by studies in which the intracellular redox environment was manipulated using a vitamin C analogue (Furumoto et al. 1998) or homocysteine (Xu et al. 2000). More recently, our studies have demonstrated the importance of the glutathione (GSH) detoxification system in the preservation of telomere integrity in these cells and also highlighted the role that chronic mild oxidative stress may play in accelerating the onset of senescence (Kurz et al. 2004).

The role of telomerase in endothelial cells

The literature is not unanimous concerning the possible expression of telomerase in adult normal endothelial cells (Hsiao et al. 1997; Yang et al. 1999; Vasa et al. 2000; Kurz et al. 2003). In our studies, we have found that telomerase activity is virtually absent in cells freshly isolated from aortic or umbilical vein quiescent endothelium, but is markedly upregulated when the same cells are stimulated to proliferate in cell culture (Kurz et al. 2003). It should be noted, however, that even in proliferating cells, levels of activity are substantially lower than those found in haematopoietic stem cells, in endothelial progenitor-derived cells or in a typical cancer cell line (Fig. 1A and B). Nevertheless, this low level of activity appears to be physiologically important, since its inhibition by stable expression of DN-hTERT, a previously described inactive mutant form of the human telomerase catalytic subunit (Masutomi et al. 2003), reduces the replicative capacity of the cells (Fig. 1C and D). These studies are consistent with the notion that introduction of exogenous telomerase into endothelial cells extends their lifespan (Yang et al. 1999) and with previous findings from our laboratory comparing the effects of different endothelial cell mitogens on telomerase activity. These studies showed that when human umbilical vein endothelial cells (HUVEC) were grown with fibroblast growth factor-2 (FGF-2), telomerase was upregulated and the cells attained a normal lifespan. In contrast, when these cells were grown with vascular endothelial growth factor-A (VEGF-A), at concentrations which showed the same mitogenic activity as FGF-2 did, telomerase remained depressed and the cells underwent senescence prematurely (Kurz et al. 2003; Trivier et al. 2004). Researchers in another laboratory, however, studying the role of telomerase in angiogenesis, have claimed that VEGF-A can activate telomerase via nitric oxide (NO) signalling (Zaccagnini et al. 2005).

Figure 1.

Telomerase in endothelial cells
Telomerase activity was determined by the telomere repeat amplification protocol as previously described (Kurz et al. 2003). A, representative images of telomerase reaction products obtained by assaying the indicated amounts of protein extracts from cultured human umbilical vein endothelial cells (HUVEC), CD34+ progenitor-derived outgrowth endothelial cells (OEC) and human umbilical cord blood-derived CD34+ stem cells. B, comparison between telomerase activity levels in endothelial cells, stem cells and HeLa tumour cells. C, downregulation of telomerase activity in HUVEC expressing a dominant-negative mutant of the human telomerase catalytic subunit (DN-hTERT). In B and C, values represent the means (±s.d.) of three determinations. D, reduction of replicative capacity in DN-hTERT-expressing HUVEC; cumulative population doublings (CPD) were determined as previously described (van der Loo et al. 1998).

The role of nitric oxide in endothelial cell senescence

It has been generally assumed that NO counteracts endothelial cell senescence, apparently by stimulating telomerase activity (Vasa et al. 2000). In contrast, our experiments using a combination of pharmacological tools and silencing RNA technology suggest that this might not be the case (Hong et al. 2007). Our findings are summarized as follows: (i) we could not detect an increase in telomerase activity by exposing HUVEC to the NO donors diethylenetriamine/NO (DETA-NO) or S-nitroso-N-acetylpenicillamine at doses (10–50 μm) that generated physiologically relevant concentrations of NO and increased intracellular cGMP levels; (ii) we observed no reduction in telomerase activity when the cells were incubated with the NO synthase inhibitor NG-monomethyl-l-arginine, although we could measure a reduction in basal cGMP levels in the same conditions; (iii) we could not unmask the putative ability of NO to stimulate telomerase by modulating the redox status of the cells; and (iv) when we downregulated endothelial nitric oxide synthase by RNA interference, causing a decrease in basal levels of cGMP, we saw no change in telomerase activity. Similarly, alterations in NO levels by any of the above means had no effect on the cellular replicative capacity or the accumulation of senescent cells.

Since excessive production of NO is associated with pathophysiological processes (Erusalimsky & Moncada, 2007), we have also investigated whether high concentrations of this molecule affected telomerase activity or cellular lifespan. Acute exposure of HUVEC to DETA-NO, at doses which generated concentrations of NO above 1 μm, inhibited telomerase in a dose-dependent manner (Fig. 2A). Furthermore, at these high doses DETA-NO induced premature senescence, an effect which was abrogated by co-incubation with N-acetyl-cysteine (NAC), but not by a guanylate cyclase inhibitor (Fig. 2B). These results are in agreement with the observation that high concentrations of NO deplete intracellular GSH (Clementi et al. 1998) and with previous findings from our laboratory and others showing that telomerase is susceptible to GSH depletion and oxidative stress (Borras et al. 2004; Kurz et al. 2004).

Figure 2.

Effect of high doses of exogenously added NO on telomerase activity and the onset of senescence
A, second passage HUVEC were treated with the indicated doses of DETA-NO for 24 h prior to measurement of telomerase activity. B, confluent HUVEC were treated for 3 days with or without 0.5 mm DETA-NO in the absence or presence of 0.3 μm 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) or 4 mmN-acetyl-cysteine (NAC). Cells were then replated and grown for a further 4 days in normal medium. The SA-β-gal activity was subsequently measured by flow cytometry as previously described (Kurz et al. 2000). Results represent the means (±s.d.) of three determinations on cells from different donors.

Concluding remarks

The occurrence of endothelial cell senescence in the vasculature is gaining increasing recognition. Cell culture studies indicate that both cell turnover and oxidative stress may contribute to this phenomenon by inducing telomere shortening. However, although the evidence for the occurrence of each of these processes in the vessel wall in vivo is compelling, the existence of a causal relationship between them and endothelial senescence awaits direct demonstration. Furthermore, while cell culture studies clearly show that endothelial cells express telomerase and that this activity is growth-regulated and sensitive to the redox environment, the relevance of these in vitro findings to human adult vascular homeostasis remains to be elucidated.



The studies from J. D. Erusalimsky's laboratory were funded by the British Heart Foundation. C. Skene was supported by a PhD fellowship (FS/04/031) awarded by the British Heart Foundation.

Author's present address

C. Skene: Royal Infirmary of Edinburgh, Little France Crescent, Edinburgh EH16 4SA, UK.