• Open Access

Age-related impairment of mesenchymal progenitor cell function


  • Alexandra Stolzing,

    1. Centre for Biomaterials and Tissue Engineering, Department of Engineering Materials, and
    Search for more papers by this author
  • Andrew Scutt

    1. Centre for Biomaterials and Tissue Engineering, Department of Engineering Materials, and
    2. Division of Clinical Sciences South, University of Sheffield, North Campus, Sheffield S37HQ, UK
    Search for more papers by this author

A. Stolzing, University of Sheffield, North Campus, Kroto Research Institute, Broad Lane, Sheffield S3 7HQ, UK. Tel.: 01142225931; fax: 01142225945; e-mail: stolzing@gmail.com


In most mesenchymal tissues a subcompartment of multipotent progenitor cells is responsible for the maintenance and repair of the tissue following trauma. With increasing age, the ability of tissues to repair themselves is diminished, which may be due to reduced functional capacity of the progenitor cells. The purpose of this study was to investigate the effect of aging on rat mesenchymal progenitor cells. Mesenchymal progenitor cells were isolated from Wistar rats aged 3, 7, 12 and 56 weeks. Viability, capacity for differentiation and cellular aging were examined. Cells from the oldest group accumulated raised levels of oxidized proteins and lipids and showed decreased levels of antioxidative enzyme activity. This was reflected in decreased fibroblast colony-forming unit (CFU-f) numbers, increased levels of apoptosis and reduced proliferation and potential for differentiation. These data suggest that the reduced ability to maintain mesenchymal tissue homeostasis in aged mammals is not purely due to a decline in progenitor cells numbers but also to a loss of progenitor functionality due to the accumulation of oxidative damage, which may in turn be a causative factor in a number of age-related pathologies such as arthritis, tendinosis and osteoporosis.


Mesenchymal stem or progenitor cells (MPC) are thought to be involved in mesenchymal tissue maintenance and repair. Exactly how this is achieved is unclear; however, tissue resident mesenchymal progenitor cells appear to be replenished by marrow-derived progenitor cells via the circulation (Spees et al., 2003). Depletion of stem cells has been suggested to contribute to degenerative diseases affecting a number of tissues including brain, liver, skin and bone, and may also be involved in the degeneration associated with aging (Rao & Mattson, 2001). MPC reside in the bone marrow along with haematopoietic stem cells (Bentley, 1982) and can differentiate into a variety of tissues including bone, fat, cartilage and tendon in vitro and in vivo (Pittenger et al., 1999; Deans & Moseley, 2000). In addition, differentiation into muscle, endothelial cells and neurons has been shown in vivo but it is not clear if this is due to fusion events, differentiation or a combination of both (Wautier et al., 2003; Ying et al., 2002). Although present in only very small numbers in the bone marrow (Hung et al., 2002; Murphy et al., 2002), MPC are capable of extensive proliferation and expansion in culture (Colter et al., 2001). Undifferentiated MPC exhibit a fibroblast-like morphology and a characteristic pattern of cell-surface antigens (Deans & Moseley, 2000). These criteria, along with the ability to differentiate into multiple cell types, have been used to define a prototypic mesenchymal stem cell phenotype, which is consistent between a number of species (Kadiyala et al., 1997; Devine & Hoffman, 2000; Ringe et al., 2002; Quagliaro et al., 2003).

The use of mesenchymal stem cells from autologous sources is very attractive for tissue engineering and also for gene therapy. However, although MPC are capable of considerable expansion in culture, they are not immortal and cellular senescence has been reported in in vitro studies (Digirolamo et al., 1999; Banfi et al., 2002). It has been demonstrated that MPC exhibit characteristics typical of the Hayflick model of cellular senescence having a limited lifespan (Banfi et al., 2002; Stenderup et al., 2003) due to telomere shortening (Banfi et al., 2002), accumulation of senescent (β-galactosidase positive) cells, and the impairment of differentiation (Digirolamo et al., 1999).

It was shown recently that cells obtained from elderly donors exhibited decreased proliferation potential and accelerated senescence compared with cells obtained from younger donors (Stenderup et al., 2003). Despite this information, little is known about how and why MPC age in vivo. Free radical-derived reactive oxygen species (ROS) are constantly generated in most living tissue and can potentially damage DNA, proteins and lipids. ROS can be formed from extracellular sources and by the cells themselves, both actively and as a by-product of biological processes. The main intracellular sources of ROS are peroxisomes, lipoxygenases, nicotinamide adenine dinucleotide phosphate (NADPH)-oxidase complex, and mitochondrial respiratory chain reactions (Inoue et al., 2003). ROS are involved in the pathogenesis of several diseases including arthritis (Hitchon & El-Gabalawy, 2004), cancer (Pelicano et al., 2004), cardiovascular disease (Li & Shah, 2004) and aging (Mates & Sanchez-Jimenez, 1999; Huang & Manton, 2004). The free radical theory of aging proposes ROS to be one of the central causes of the aging process (Harman, 1956) causing damage to all of the macromolecules of the cell. Excessive oxidative stress caused by increased ROS levels can induce the stress-related senescence pathways regulated by p53/p19 and p16RB leading either to senescence or apoptosis (Pelicci, 2004). Cellular senescence drives cells into growth arrest (Foreman & Tang, 2003) and apoptosis can kill cells, thus avoiding an inflammatory response (Gregory & Devitt, 2004). Both are tumour-supressive mechanisms but may also contribute to organismal aging as they would exhaust the stem/progenitor cell pools normally responsible for replenishing tissues and organs (Baxter et al., 2004). Several antioxidative enzyme systems exist to prevent damage caused by ROS to tissue and cells, including catalase, superoxide dismutase (SOD) and glutathione peroxide (GPx) (De Haan et al., 2003). SOD, mainly found in the cytosol, converts superoxide anions to hydrogen peroxide and oxygen (Michiels et al., 1994), which is in turn decomposed to water and oxygen by catalase (Makino et al., 2004). Glutathione peroxidase, localized mainly in the cell membrane, can also neutralize hydrogen peroxide and in addition can inhibit lipid peroxidation (Singh & Pathak, 1990; Jung & Henke, 1996).

Recent data regarding endothelial progenitors have shown that progenitor cells expressed significantly higher antioxidative defences compared to mature endothelial cells (Dernbach et al., 2004). Because of this we investigated the role of oxidative stress in the accumulation of age-related damage and markers of aging in fibroblastic colonies derived from the clonal expansion of fibroblast colony forming units (CFU-f), which are thought to resemble closely MPC (Friedenstein, 1976). Of particular interest was the question whether such changes in MPC quality already occur during maturation and adulthood rather than in old age. We therefore included cultures of immature, mature, and adult CFU-f in our experiments.


Effect of age on fibroblast colony-forming unit (CFU-f) numbers

When bone marrow cells are plated out at low densities, individual MPC adhere and proliferate to form fibroblastic colonies. Due to the low plating density the colonies grow essentially in isolation and therefore represent the clonal expansion of a single CFU-f or MPC. Furthermore, a proportion of the colonies differentiate along the osteoblastic pathway and express a number of markers associated with bone formation including alkaline phosphatase and collagen and calcium accumulation. With increasing age, the number of fibroblastic colonies formed by rat mesenchymal progenitor cells was significantly reduced. Cultures derived from 3-week-old rats gave rise to ∼160 colonies per dish whereas those derived from 56-week-old rats produced ∼90 colonies per dish (Table 1). The proliferative potential of the MPC was estimated by calculating the mean colony size (Table 1). It was found that mean size of the colonies also decreased with age with 52-week-old rat MPC producing colonies only half the size of those produced by 3-week-old rat MPC. There was also a relationship between age and MPC differentiation potential. The numbers of alkaline phosphatase (ALP)-positive colonies were reduced fourfold from ∼85 for MPC from 3-week-old rats to ∼21 for the oldest age group. In addition, the size of the ALP-positive colonies was also decreased in the oldest age group by around 30%. Similarly, the number of calcium- and collagen-positive colonies decreased with age. Numbers of calcium-positive colonies were reduced from 47.9 to 13.2 and collagen-positive colonies from 39.2 to 21.3 in 3-week-old rats and 56-week-old rats, respectively. Mean calcium- and collagen-positive colony sizes also declined in the old group but the effect was smaller than that seen for ALP-positive colonies. In contrast, no significant age-related effect was observed for Oil red O positive colony numbers or sizes.

Table 1.  Changes in CFU-f during aging
Age of rat3 weeks7 weeks12 weeks56 weeks
  1. Whole bone marrow cells were isolated from rats of increasing age and cultured in the CFU-f assay. The number and size of the colonies were determined by image analysis as described in the methods. The number of colonies is representative of the number of progenitor cells and the size of the colonies reflects the proliferative potential of the progenitor cells. The colonies were serially stained for bone (alkaline phosphatase, calcium), cartilage (collagen) and fat (Oil red O) markers as described in methods and for the total colony numbers. The colony numbers of collagen, calcium, alkaline phosphatase or Oil red O positive colonies gives the number of progenitor cells committed to either the bone, cartilage or fat cell lineages.

  2. Isolated MPC from rats of different ages were stained for the percentage of senescent cells after 0 and 15 passages using β-galactosidase and the number of β-galactosidase-positive cells were calculated per population doublings. All experiments were performed in triplicate using MPC isolated from five rats per age group. *denotes P < 0.05, **P < 0.01, ***P < 0.005.

Colony number
 Total 160 ± 8.6 124 ± 4.3*** 121 ± 6.2***  92 ± 5.2***
 Calcium47.9 ± 2.833.8 ± 3.1***21.8 ± 5.2***13.2 ± 2.5***
 Collagen39.2 ± 5.230.4 ± 6.6*29.2 ± 5.8**21.3 ± 3.8***
 Alkaline phosphatase  85 ± 1.556.4 ± 2.8***34.9 ± 3.7***21.1 ± 2.8***
 Oil red O12.1 ± 313.6 ± 2.912.5 ± 1.712.4 ± 1.5
Mean size of colonies
 Total34.2 ± 2.532.2 ± 2.923.8 ± 3.1***17.9 ± 1.9***
 Calcium20.1 ± 2.415.3 ± 2*** 9.6 ± 0.9*** 8.8 ± 0.8***
 Collagen15.9 ± 2.313.4 ± 2.1*12.9 ± 0.8**10.5 ± 1.2***
 Alkaline phosphatase15.8 ± 3.312.5 ± 1.7*11.1 ± 0.6**10.8 ± 2.4***
 Oil red O24.6 ± 323.9 ± 4.123.7 ± 2.423.3 ± 2.8
%β-galactosidase positive colonies
 0 passage 5.6 ± 0.4 4.6 ± 0.5 6.1 ± 0.6 4.3 ± 0.2
 15 passage22.4 ± 1.619.7 ± 2.118.6 ± 0.822.6 ± 1.3
 Population doubling/passage 1 0.80.680.54
 %β-galactosidase positive cells/population doubling 1.5 ± 0.21.65 ± 0.4 1.8 ± 0.8 2.8 ± 0.2***

Growth curve

The growth kinetics of MPC derived from rats of increasing age were observed from the primary passage until cells in culture ceased to replicate for at least 4 weeks. MPC from older rats showed a much slower rate of growth and achieved fewer population doublings (PD) before reaching senescence (Fig. 1A). The morphology of the cells changed from mostly spindle-shaped fibroblastic cells to more rounded and flattened cells indicating a senescent phenotype (Fig. 1B,C). These changes were accompanied by an accumulation of actin fibres in the senescent cells. MPC lost the ability to reach confluence with age (Fig. 1D,E).

Figure 1.

Growth kinetics and morphology of rat MPC in culture. (A) MPC isolated from rats aged 3, 4, 7 and 56 weeks were cultured under standard conditions for up to 4 months and the cumulative numbers of population doublings (PD) were calculated. Each group contained five animals and all experiments were performed in triplicates. Cell morphology of representative MPC cultures from 3-week-old rat early passage (B) and late passage (C) are shown. Actin staining in representative early passage MPC cultures (D) and late passage MPC (E) are shown.

Senescence and apoptosis

It was thought that the decrease in CFU-f numbers seen in aged animals may have been the result of an increase in levels of apoptosis or senescence. Freshly isolated MPC cultivated for 2 days in 24-well plates showed essentially no signs of β-galactosidase activity at pH 6 in any of the tested age groups (Table 1, Fig. 2A,B). As a positive control, staining was also performed at pH 4 to display lysosomal β-galactosidase activity (results not shown) and positive staining was found in MPC from all age groups. Once in culture, MPC displayed increasing senescence with passage number regardless of the age of the donor animal (Table 1, Fig. 2C,D); however, when calculated as percentage of β-galactosidase-positive cells/PD then an age-related increase of senescent cells was found (Table 1).

Figure 2.

Senescence and apoptosis of MPC. MPC were stained with β-galactosidase for the determination of senescence level in early (A,B) and late (C,D) passage cultures of cells isolated from 3-week- and 56-week-old rats. Representative pictures are shown. Levels of apoptotic cells and p53 positive cells were determined in first passage cultures of MPC from 3-week- and 56-week-old rats. Experiments have been performed in triplicate and repeated independently three times. ***P < 0.001.

Levels of apoptosis were also measured in MPC from young (3 weeks) and old (56 weeks) animals. In contrast to the levels of senescence, the percentage of apoptotic MPC was increased in the oldest group by about fourfold (Fig. 2E). This was accompanied by an increase in p53 protein levels in the MPC (Fig. 2F).

Reactive oxygen species and nitric oxide generation

Because the decrease in CFU-f numbers in aged rats may be related to changes in oxidative status, we examined the effect of age on ROS and NO levels. It was found that MPC-associated extracellular ROS formation did not change with age in female rats (Fig. 3A), whereas NO production was slightly but not significantly increased (Fig. 3B). In contrast levels of intracellular ROS formation were significantly increased by approximately 20% in the 56-week-old animals compared to the 3-week-old animals (Fig. 3C).

Figure 3.

Peroxide and nitrite oxide production of MPC. Levels of peroxides were determined in first passage cultures of MPC from 3-week- and 56-week-old rats using DCF-DA (5 µm). Fluorescence intensity of DCF in the MPC supernatant was determined after 0 h and 24 h using a fluorescence plate reader. (A) NO release by MPC of rat of different ages. First passage cultures of MPC from 3-week- and 56-week-old rats were cultured for 24 h and aliquots of the supernatant were taken out and incubated with the same amount of Griess reagent to determine NO levels (B). In addition MPC were loaded with DCF-DA to measure intracellular levels of ROS (C). All experiments were performed in triplicate and repeated independently three times. **P < 0.01.

Oxidative damage to proteins and lipids

Protein carbonyl formation was increased more than fourfold in MPC from 56-week-old rats compared to 3-week-old rats. Carbonyl content also increased in brain and kidney tissue isolated from the same rats and was highest in the kidney samples (Fig. 4A). Thiobarbituric acid-reactive substances (TBARS) levels increased around threefold in MPC from 56-week-old rats compared to MPC from 3-week-old rats. A gradual increase in TBARS levels was also found in brain and kidney tissues (Fig. 4B). Similarly, levels of lipofuscin in MPC also increased with age (Fig. 6A).

Figure 4.

Oxidative damage accumulation in MPC. Protein were extracted from MPC derived from rats of different ages and from kidney and brain tissue homogenates of the same rats. Carbonyl content of these tissues was determined using ELISA (A). Oxidized lipids and aldehydes were measured in MPC, kidney and brain homogenates from rats of different age using the TBAR assay. Protein extract were analysed for TBARS and expressed as MDA equivalents (B). All experiments were performed in triplicate and repeated independently three times. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 6.

Lipofuscin and proteasome activity of MPC. Total protein was isolated from MPC, kidney and brain homogenates and the chymotrypsin-like activity of the 20S proteasome was measured using a fluoropeptide (A). First passage cultures of MPC from rats of different ages were analysed in a GUAVA personal flow cytometer for there lipofuscin content. The wavelengths at 530 nm and 630 nm were measure as equivalent to the intracellular lipofusine content. All experiments were performed in triplicate and repeated independently three times. *P < 0.05; ***P < 0.001.

Antioxidative enzyme activities

As MPC showed signs of age-related damage accumulation (oxidized proteins and lipids), we studied the activity of several antioxidative defence enzymes. SOD activity and glutathione peroxidase activity decreased steadily with age in MPC with the activity of both enzymes being reduced by about 50% in cells from 56-week-old rats compared to those from 7- and 12-week-old rats (Fig. 5A,B). Interestingly the SOD activity found in MPC was much higher than in either brain or kidney homogenates. Glutathione peroxidase activity in the youngest group was found to be nearly as high as that in kidney homogenates but by 56 weeks the activity had reduced to around 20% of that found in kidney. In contrast, 20S proteasome activity showed no correlation with the age of the rat from which the MPC were isolated (Fig. 6B). Due to the poor recovery of protein from 3- and 4-week-old rat MPC, it was not possible to perform these assays at this age stage.

Figure 5.

Antioxidative defence enzyme activities in MPC. Total protein was isolated from MPC, kidney and brain homogenates of rats of different ages. The activity of superoxide dismutase (A) and glutathione peroxide (B) was determined as described. All experiments were performed in triplicate and repeated independently three times.*P < 0.05; ***P < 0.001.


The capacity of organs to repair themselves declines with age and this may be due to a loss or failure of adult stem cells. It has also been suggested that these same adult stem cells may be used to engineer tissue constructs to replace damaged tissues or organs. MPC hold great promise for regenerative medicine (Murphy et al., 2002) as there are few ethical concerns regarding the use of these cells and few concerns about possible immunological rejection as they do not appear to elicit an immune response (Aggarwal & Pittenger, 2005). However, there may be other problems with the use of adult stem cells; in particular the age of the donor may have an impact on their vitality. Tissue stem cells or progenitor cells are exposed to various toxins and detrimental events during their lifetime in the same way as differentiated cells and the resultant accumulation of damage may be one of the causes of aging (Kirkwood & Franceschi, 1992). A better understanding of age-related changes in stem cells will help to explain the degenerative changes observed in organs during aging. Experiments with embryonic, adult neural stem cells and haematopoietic stem cells have shown that they express a certain subset of genes involved in oxidative stress resistance, detoxification and DNA repair (Ivanova et al., 2002; Ramalho-Santos et al., 2002) to increase the resistance of these cells to stress and aging. In contrast, the extent to which aging affects mesenchymal stem cells is relatively unknown.

CFU-f formation

CFU-f numbers have been shown to reflect changes in bone anabolic activity seen after age and ovariectomy (Scutt et al., 1996a), unloading (Tanaka et al., 2004) and the administration of bone anabolic drugs (Nishida et al., 1994; Erben et al., 1998; Weinreb et al., 1999). The CFU-f assay has also been shown to be more sensitive than other culture systems in terms of the response to bone active agents such as PGE2 (Still & Scutt, 2001) or PTH (Davies & Chambers, 2004). Analysis of CFU-f cultures also reveals considerable variability between colonies indicating significant heterogeneity within the MPC population (Dobson et al., 1999). In this study an age-related reduction in the number of CFU-f per culture was found. Similar results have also been found in cultures from senescence accelerated mice (SAM) compared to normal mice (Tsuboi et al., 1991) and in those from normal aged mice (C57BL) (Brockbank et al., 1983). An age-related reduction in CFU-f from rats has also been reported (Scutt et al., 1996a; Majors et al., 1997; Nishida et al., 1999; D’Ippolito et al., 2004); however, contradictory data indicate no changes in CFU-f numbers with age (Oreffo et al., 1998; Stenderup et al., 2001; Stenderup et al., 2003) and in one case an increase has been reported (Martinez et al., 1999). A variety of factors may explain these discrepancies including differences in isolation methods, counting methods, growth media, species or strain of animals as well as the age groups which were compared. The majority of studies suggest that CFU-f numbers are reduced with age. However, the effect is probably multifactorial and requires further systematic study.

Proliferative capacity

The age-related reduction in mean size of the colonies derived from old rats seen in this study perhaps reflects the functional competence of the aged stem cells and has also been seen in mice (Globerson, 1997). Colony size is a measure of the proliferative capacity of individual CFU-f ex vivo and is therefore an indicator of qualitative changes in the CFU-f. This is in contrast to colony number which simply reflects total CFU-f numbers regardless of their proliferative capacity. This is also subtly different to in vitro growth curves in that the cells are assessed directly ex vivo whereas in growth curves the cells are preselected for rapid proliferation. Despite this, the loss of proliferative potential with age during in vitro expansion seen in this study has also been found in human (Baxter et al., 2004) and murine MPC (Pendergrass et al., 1995). Loss of proliferative potential due to age has also been found in skin stem cells (Quagliaro et al., 2003), haematopoietic stem cells (Vaziri et al., 1994), muscle stem cells (Conboy et al., 2003) and myocardial stem cells (Torella et al., 2004), suggesting that stem cell aging may be a general phenomenon.

Differentiation of MPC

There was an age-related decrease in the number and size of colonies that stained positively for collagen, calcium and alkaline phosphatase. This is consistent with previous investigations showing a reduced capacity for differentiation in aged animals (Meunier et al., 1971; Nakahara et al., 1991; Bergman et al., 1996). It has, however, been shown that osteoblastic differentiation requires an initial phase of rapid proliferation, inhibition of which prevents differentiation (Stein et al., 1990). Therefore the apparent decrease in differentiation may not necessarily be intrinsic to the cells themselves but a result of their decreased proliferative capacity.


Cellular senescence is a complex phenotype that entails changes in both function and replicative capacity. Unlike apoptosis, which eliminates damaged cells from tissues, senescent cells remain alive despite changes in morphology, metabolism, and derangement of differentiated functions (Itahana et al., 2004). β-Galactosidase activity is associated with replicative senescence in vitro (Dimri et al., 1995). While we observe a general increase in β-galactosidase activity after cultivation in vitro, it might at first appear as if there were no differences between age groups which seems to contradict the observed age-related reduction in growth. However, when the number of β-galactosidase positive cells is calculated per population doubling as suggested by Stenderup et al. (2003) the age-related increase in senescence becomes apparent.

Some earlier studies on MPC evaluate senescence by testing the loss of differentiation potential or by morphology after expansion in culture. When assessing senescence in freshly isolated MPC without extensive in vitro replication we find rather low levels of β-galactosidase staining in MPC from both young and old rats. This might suggest that MPC of all ages are largely nonsenescent in vivo. The increased levels of senescent MPC seen after in vitro expansion may be an artefact due to the higher stresses experienced by the cells during in vitro cultivation (Strubing et al., 1995). Similar findings are reported with other cell types (Serra et al., 2000; Piacibello et al., 2005). In fresh ex vivo or in vivo tissue no correlation between the age of the organism and the level of senescence measured by β-galactosidase staining could be found (Severino et al., 2000; de Magalhaes, 2004). Aging in MPC only correlates with the x-gal levels after extensive proliferation in vitro.

Oxidative stress and damage accumulation

We observe a higher incidence of lipofuscin, TBARS, and carbonyls in association with aging in vivo. It is known from other cell types that lipofuscin content increases as soon as the growth rate in a cell culture declines (Gieche et al., 2001). This is consistent with the suggestion that in vivo aged MPC have a lower rate of proliferation than younger MPC and therefore begin to accumulate lipofuscin.

Progenitor cells are known to be particularly well protected against oxidative and other stresses. It has been suggested that the expression of high levels of antioxidative enzymes may be a characteristic of stem cells that is lost with differentiation (Dernbach et al., 2004). For example, it has been shown that endothelial progenitor cells express significantly higher levels of SOD and glutathione peroxidase mRNA than mature endothelial cells (He et al., 2004). We confirm this observation for MPC compared with brain and kidney cells from the same animal, but protection levels were progressively reduced in cells from aged animals.

Similarly, levels of intracellular ROS (but not extracellular ROS or NO) were increased in association with age. ROS levels are a somewhat problematic indicator as they are prone to fluctuation. However, we also observe an age-associated decrease in proteasomal activity. Declining antioxidative defence enzyme activities with no increase in 20S proteasome activity to cope with increasing levels of oxidized protein lead to a damage cascade (Merker et al., 2001; Stolzing et al., 2005). It is known that ROS and NO generation increases in other tissues, such as bone marrow, with age and this may play a contributory role in damage accumulation by MPC. These proinflammatory molecules are also involved in proliferation, differentiation, and initiation of apoptosis. Consistent with this, the differentiation of haematopoietic progenitors in vitro is inhibited by the addition of ROS-producing macrophages (Hoffman et al., 1993). Age-related increase in NO production has been found in rat bone marrow cells and mouse bone marrow-derived macrophages after challenge with lipopolysacharide (LPS) (Chen et al., 1996; Grinnemo et al., 2004). It has been hypothesized that oxidative stress increases the number of single strand breaks in the telomeres causing the cells to exit the cell cycle or enter apoptosis.


Apoptosis is an evolutionarily conserved process by which an organism can remove unwanted or damaged cells. The phospholipid-like phosphatidylserine (PS) represents a hallmark indicator of dying cells. Changes in PS asymmetry, analysed by measuring annexin-V binding, were detected before morphological changes associated with apoptosis have occurred and before membrane integrity has been lost. Late-stage apoptotic cells forming apoptotic bodies or necrotic fragments were either washed away during medium change or excluded in the flow cytometer and only cells in the same size range as unstained cells were included in the measurement. Similarly p53, a tumour-suppressor gene which mediates apoptosis primarily in connection to telomere dysfunction and DNA damage (Artandi & Attardi, 2005; Campisi, 2005), was found to be up-regulated in aged MPC. This suggests that elevated apoptosis levels in aged MPC are associated with senescence. Consistent with this, an age-related increase in apoptosis was also observed in this study.

The increased levels of apoptotic MPC in aged animals may have an impact on the relationship between aging, stem cells, and cancer. Mathematical models of a theoretical stem cell niche have suggested that stem cell death by apoptosis, rather than their entry into senescence, in aged animals may increase the possibility of developing cancer (Spees et al., 2003). It is suggested that as stem cells die by apoptosis they are replaced by increasing the rate of symmetrical stem cell self-renewal, thus increasing the likelihood of their transformation into cancer cells.


MPC display a significantly decreased functional capacity and vitality in aged animals and our data suggest that these changes are the result of age-related damage accumulation, due to a loss of SOD and glutathione peroxidase activity and decreased proteasome activity. The decreased functionality of MPC could have repercussions on physiological tissue maintenance and the repair of tissue damage, and may be a factor in organ aging. Additionally, changes in apoptosis incidence in aged stem cells may explain the increased risk of cancer in older organisms.

Our findings may have implications for the use of adult stem cells in cellular therapy and tissue engineering by indicating that MPC progressively lose their capacity to proliferate and differentiate with age, a process that begins in early maturity. This could compromise engraftment success and the lifetime of tissue engineered constructs using MPC from aged donors.

Experimental procedures


All chemicals were obtained from Sigma-Aldrich (Dorset, UK) unless stated otherwise.


Female Wistar rats were purchased from Harlan (Harlan UK Ltd) and housed two or three per cage with free access to food and water. At the end of the experiment the animals were killed by cervical dislocation. All procedures were undertaken in compliance with Home Office legislation under the Animals (Scientific Procedures) Act 1986.

Isolation of rat bone marrow stem cell

Five animals per age group were used in the experiments. Bone marrow cells were obtained by centrifugation from tibia and femurs according to Dobson et al. (1999); MPC were isolated by the method from Sekiya et al. (2002). In brief, 1 × 105 cm−2 mononuclear cells were plated out in expansion medium [DMEM-LG supplemented with 10% Serum Supreme (Cambrex Bio Science, Wokingham, UK) and antibiotics]. After 24 h, non-adherent debris was removed, and adherent cells were cultured further for 7 days. All MPC were on passage 0 until otherwise stated. Kidney and brain tissue were isolated from the same animals and homogenized in lysis buffer (250 mm saccharose, 25 mm HEPES, 10 mm MgCl2, 1 mm EDTA). Cells were lysed by repeated freezing and thawing and cell lysate was then centrifuged for 30 min at 14 000 g (at 4 °C) and the protein pellet was stored at −20 °C until needed.

Growth curve

Rat MPC were serially subcultured under standard conditions. Briefly, the cells were cultured at 5 × 103 cells cm−2 in T25 culture flasks in the above medium. When cells were 90% confluent they were trypsinized and replated into T25 culture flasks and the cultures continued until the cells stopped dividing. The number of population doublings (NPD) between subcultures was calculated according to the following equation:

NPD = Log10(N/N0) × 3.33 (where N = no. of cells in the flask at the end of growth period; N0 = no. of cells plated in the growth vessel.)

Colony-forming unit fibroblastic assay

The fibroblastic-colony forming unit (CFU-f) assay used was a modification of the technique of Kuznetsov et al. (1997). Two million mononuclear BMC were suspended in 0.5 mL media and plated on 55-cm2 petri dishes. Cells were incubated for 14 days and washed with PBS. The cells were fixed in ethanol, sequentially stained for alkaline phosphatase (APase), calcium, lipid and collagen-positive colonies and photographed as follows (Scutt & Bertram, 1995; Scutt et al., 1996b). APase was stained histochemically by incubation with naphthol phosphate (0.05 mg mL−1) in Tris buffer (0.08 m, pH 7.5) containing fast red bb (1 mg mL−1) for 30 min at 20 °C. The cultures were washed under running water, photographed and then destained with absolute ethanol. Calcium was stained histochemically with 0.5% alizarin red pH 6.2 after which the cultures were washed under running tap water. After analysis, the cultures were destained and decalcified with 5% perchloric acid. Collagen was stained with 1% sirius red F3BA in saturated picric acid for 30 min; excess sirius red was removed by washing under running tap water. After analysis, the cultures were destained with methanol/0.2 N NaOH (50 : 50). Total colonies were visualized by staining with 1% methylene blue in borate buffer (10 mm, pH 8.8) for 30 min followed by three washes with borate buffer.

Colony analysis

Colony numbers were assessed using the method of Dobson et al. (1999). Briefly, the acquired digital images were imported into Adobe Photoshop and colony irregularities were smoothed out by applying to the image a median filter with a radius of 2 pixels which evens out differences between adjacent pixels and a despeckling filter which removes single pixels whose values differ from their surroundings. Finally, as the image analysis software can only analyze greyscale images, the image was converted to an 8-bit greyscale TIFF image and saved. In order to analyse the 8-bit greyscale images, they were first imported into Bioimage ‘Intelligent Quantifier’ image analysis software (Bioimage, UK). The area of interest was then marked and analysed using the colony counting mode. Colonies can be selected according to both their size and intensity and in this study colonies of at least 20 pixels (corresponding to 1 mm) in diameter and having an intensity of at least 20 grey levels above background (corresponding to approximately 80 cells) were selected. From this data, colony number and size were calculated. By automating this process, the role of the operator in colony identification is removed, as the presence of a colony is rigidly defined in terms of intensity and diameter.


Cells were seeded on a 24-well plate and grown overnight. They were fixed in 4% paraformaldehyde, washed three times with PBS and then incubated for 10 min in 0.1% TritonX 100 (in PBS). After washing the distribution of actin filaments in the cells were analysed by incubating with 714 nM FITC-phalloidin in the dark for 20 min (RT). After washing the cells were analysed by fluorescence microscopy using an ImageXpress® 5000A automated imaging system with ×10 objective magnification (Molecular Devices Corp., Sunnyvale, CA, USA). In addition images of these cells were also taken using a conventional phase-contrast microscope using ×10 objective.

Senescence-associated β-galactosidase staining

Cells were seeded on 24-well plates and grown overnight. Senescence-associated β-galactosidase staining was performed as described previously (Dimri et al., 1995). Cells were fixed in 4% formaldehyde for 5 min, followed by three washes in PBS at room temperature. They were then incubated in freshly prepared staining solution (0.05 mg mL−1 5-bromo-4-chloro-3-indolyl-D galactoside, 40 mm citric acid/sodium phosphate, 5 mm potassium ferrocyanide, 5 mm potassium ferricyanide, 150 mm NaCl, 2 mm MgCl2, pH 6.0) at 37 °C for 24 h. Digital images were acquired and the number of stained cells in random fields of 500 cells were determined. Data are expressed as percentage positive cells.


Apoptosis was measured using the APOpercentage kit (Biocolor Ltd, Belfast, Ireland). Fifty thousand cells were seeded in 96-well plates and incubated for 24 h. The culture medium was replaced with Apomedium (9.5 mL culture medium containing 0.5 mL Apo dye), incubated for 1 h and the absorbance was measured at 550 nm in a plate reader.

Level of p53 expression

MPC pellets were incubated with an anti-p53 antibody (diluted 1 : 100; 4 °C; 30 min; Serotec, Oxford, UK) after fixation in 4% paraformaldehyde, followed by incubation with a PE-conjugated secondary antibody for 30 min at 4 °C. The cells were analysed by flow cytometry using a Guava PCS flow cytometry system (Guava Instruments, Hayward, CA, USA) and expressed as percentage positive cells.

Reactive oxygen species generation

MPC cultured in 96-well plates were washed with PBS and the MPC were supplemented with new medium. After 0 h and 24 h, medium samples were taken, mixed with 2′,7′-dichlorofluorescin diacetate (H2DCFDA) (100 µm, Molecular Probes/Invitrogen Ltd, Paisley, UK), incubated for 5 min at RT and the amount of DCF fluorescence measured in a plate reader at 388 nm/456 nm straight away. In addition MPC were cultured in the presence of DCF-DA (5 µm) at 37 °C for 30 min and then harvested. The cells were analysed in a Guava personal flow cytometry system to measure intracellular ROS production.

Nitrite assay

Nitrite concentration was measured using standard Griess reagent. Briefly, 50 µL of the supernatant was incubated with an equal volume of Griess reagent. After 5-min incubation at room temperature, the absorbance was measured at 560 nm using a plate reader.


Protein carbonyls were measured according to Buss et al. (1997) with modifications (Sitte et al., 1998). Protein extracts were normalized to 3 mg protein mL−1, derivatized with 2,4-dinitrophenylhydrazine (DNPH) and adsorbed on Maxisorb multiwell plates. Protein carbonyls were detected using an anti-DNPH primary antibody and an antirabbit-IgG peroxidase-conjugated secondary antibody. O-phenyldiamine was used to develop the plate, and the absorbance was determined using a multiwell plate reader with a detection wavelength of 492 nm (reference filter: 750 nm).

Thiobarbituric acid reactive substances

In vitro oxidative damage to lipid by copper-induced oxidation was determined by the TBARS assay. The resulting complex was measured at 535 nm spectrophotometrically and the malondialdehyde concentration per mg protein was calculated using an extinction coefficient of 1.56 × 105 M−1 cm−1 (Buege & Aust, 1978).

Lipofuscin content

The cellular lipofuscin content was measured as the autofluorescence intensity of unfixed cells by flow cytometry. Cell pellets were suspended in serum-free medium and analysed in a Guava personal cytometry system using excitation of 488 nm and emission of 530 nm (channel 1) and 630 nm (channel 2) using at least 10 000 cells for each sample.

Measurement of proteasomal activity

Cells were washed with PBS, trypsinized and the cell pellet lysed in 200 µL lysis buffer (250 mm sucrose, 25 mm HEPES pH 7.8, 10 mm MgCl2, 1 mm DTT) by repeated freeze/thaw cycles. The remaining non-lysed cells, membranes and nuclei were removed by centrifugation at 14 000 g for 30 min at 4 °C. The supernatant was incubated in a buffer consisting of 225 mm Tris (pH 7.8), 7.5 mm magnesium acetate, 7.5 mm MgCl2, 45 mm KCl and 1 mm DTT. The chymotrypsin-like activity of the proteasome was measured using 200 µm suc-LLVY MCA (Bachem Ltd, Merseyside, UK). After 60 min of incubation at 37 °C, the fluorescence of the reaction mixture was measured using 360 nm excitation/485 nm emission and free MCA as standards.

Superoxide dismutase activity and gluthatione peroxidase activity

The activity of both enzymes was evaluated using a commercially available kit (OxisResearch, Portland, OR, USA) according to the manufacturer's instructions.

Data analysis

Values were expressed as mean ± standard deviation. The statistical significance of differences among experimental groups was evaluated by one-way anova and multiple comparisons made using Tukey's test with *P < 0.05, **P < 0.01, and ***P < 0.001 being considered significantly different.


The research was supported by a grant from the BBSRC. We thank Mr Sebastian Sethe for advice in editing this manuscript.