Age-Dependent Increase in Side Population Distribution Within Hematopoiesis: Implications for Our Understanding of the Mechanism of Aging

Authors


Abstract

It is thought that, as we age, damage to our stem cells may lead to diminished stem cell pool function and, consequently, a reduced organ regeneration potential that contributes to somatic senescence. Stem cells have evolved many antitoxicity mechanisms, and certain mechanisms may be utilized to isolate hematopoietic stem cells. One method exploits the activity of the ATP-binding cassette/G2 transporter to efflux Hoechst 33,342 and results in a stem cell population known as the side population (SP). The SP subset represents a remarkable enrichment for hematopoietic stem cells and provides an opportunity to re-evaluate age-based changes in hematopoietic stem cells. We report here that the frequency of SP cells steadily increases with age, as does the proportion of Lin/Sca-1+/c-kit+ cells that is capable of Hoechst efflux. Phenotyping, progenitor, and long-term repopulation assays have indicated that SP cells in older mice are still stem cells, albeit with a lower homing efficiency than SP cells from younger mice. Analysis of apoptosis within SP cells has revealed an apoptosis-resistant population in SP cells from old mice. Gene expression analysis has determined that SP cells from old mice have a reduced expression of apoptosis-promoting genes than SP cells from young mice. This increase in SP cells with age seems to be an intrinsic property that may be independent of the age of the microenvironment (niche), and our data might provide some clues as to how this alteration in the proportion of stem/progenitor cells occurs. A possible selection-based mechanism of stem cell pool aging is discussed.

Disclosure of potential conflicts of interest is found at the end of this article.

Introduction

As we humans are far too aware, all mammals have a finite lifespan. Apart from basic longevity, age is also a major risk factor for most cancers and diseases. There are two main theories for the mechanism of aging in mammals: evolutionary and damage-based. The former theory involves the natural selection of genes that are associated with longevity. However, evolutionary pressure is mainly exerted on genes that have an effect on or before reproduction, and genes associated with longevity would not necessarily be selected. The latter, more probable hypothesis argues that accumulative damage of key organs over the lifetime of an organism limits longevity [1].

Stem cells have evolved many mechanisms to cope with their exposure to toxins during the lifetime of an individual. Indeed, a major function of stem cells and differentiation hierarchies may be to preserve the integrity of DNA. Due to this unique role of stem cells, they seem to have evolved a different approach to damage limitation from most other cells that are ultimately destined for terminal differentiation and senescence/apoptosis, especially in rapid-turnover cell systems. It is thought that stem cells maintain “immortal strands” of DNA [2]. To maintain the integrity of their DNA, stem cells are thought to be more likely to die by apoptosis in response to DNA damage than attempt repair and risk an error [2]. Indeed, there are certain DNA repair mechanisms that are absent in somatic stem cells [3]. Since repair mechanisms may potentially make errors, it seems that it has been evolutionarily advantageous for somatic stem cells to be more capable at preventing damage than attempting repair. Cell surface transporters, which are thought to reduce the intracellular concentration of toxic compounds, are particularly active in hematopoietic stem cells [4]. Antitoxic enzymes (such as aldehyde dehydrogenase) that convert intracellular toxins to less reactive metabolites are highly expressed in hematopoietic stem cells [5]. These observations are consistent with the hypothesis of aging through stem cell damage. It may be that, as we age, damage to our stem cells accumulates, and, hence, more stem cells die by apoptosis or are removed by senescence. This may lead to diminished stem cell function and, consequently, a reduced organ regeneration potential that contributes to somatic senescence [6]. The hematopoietic system would be our choice source of stem cells to examine the effects of aging. The hematopoietic system is the best characterized of any stem cell-driven organ and, given the massive output throughout life, its degeneration is probably a major limiting factor to longevity.

The hypothesis of damage-induced stem cell aging predicts that stem cells die over the lifetime of an individual. Little work on the effects of aging on hematopoietic stem cells has been possible with human hematopoietic cells. All human hematopoietic cell studies to date have been limited to in vitro characterization and immunophenotyping. Although it has not been shown directly, there is evidence of a decline in human hematopoietic stem cell function with age. It is well known that the amount of active bone marrow decreases with age. In younger individuals, it forms the cavities of most of the long bones but is restricted to only the pelvis, sternum, and vertebrae in old age. In women heterozygous for X-linked glucose-6-phosphate dehydrogenase (G6PD), only one allele is randomly activated in each stem cell. Early in life, the distribution of each allele in the hematopoietic system is balanced, suggesting an even contribution from all stem cells. In old age, the distribution of G6PD is skewed toward one of the alleles [7]. This suggests that the number of stem cells reduces with age, or fewer cells are recruited in to cycle. Furthermore, in vitro colony formation assays have revealed that, although the absolute number of colony forming cells increases with age, their individual self-renewal potential decreases [8].

In the murine system, studies have been published that report an increase in the absolute numbers of phenotypically-defined hematopoietic stem/progenitor cells [9, 10]. It has also been reported that older animals have a reduced repopulation potential when compared to younger organisms, often with a skew away from the lymphoid lineage and toward myeloid development [9, [10], [11], [12]–13].

As mentioned previously, stem cells have evolved many mechanisms to avoid toxicity, and some of these mechanisms may be exploited to identify and isolate hematopoietic stem cells. ATP-binding cassette (ABC)/G2 is a cell surface transporter that is particularly active in hematopoietic stem cells. A flow cytometric method has been developed that exploits the unique activity of this transporter to efflux the DNA binding dye Hoechst 33342 and results in a stem cell population known as the side population (SP) [4].

We have already characterized the murine SP fraction in terms of phenotype and repopulation potential [14]. This study allowed us to conclude that the small proportion of the established, phenotypically-defined lineage negative Sca-1+, c-kit+, Thy-1+, CD135 population that is the SP contains the majority of the repopulation ability present in the murine marrow. Indeed, we observed reliable engraftment from 10 SP cells, and others have even achieved single cell engraftment [15].

Although it has been recently reported that SP cell frequency is altered with age, only a preliminary investigation into their functional potential has been described, and this was not the main focus of the study [13]. Here, we present comparative analysis of phenotype, in vivo repopulation potential, progenitor cell production, apoptosis, and gene expression profile of the SP population from mice of various ages.

Materials and Methods

Mouse Strains

Strains used in this study were C57Bl6-Thy-1.1 (Ly5.2), C57Bl6-Ly5.1, and nonobese diabetic/severe combined immunodeficient (NOD/SCID)-Ly5.1 (all originally from Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com). NOD/SCID mice were bred at Charles River before arrival in our animal facility. All C57Bl/6 colonies were maintained in our animal facility.

Cell Preparation

C57Bl/6 bone marrow was flushed from murine femurs and tibias and red blood cells were lyzed with ammonium chloride (Stem Cell Technologies, Vancouver, BC, Canada, http://www.stemcell.com). Nucleated cells were Hoechst-labeled for SP analysis as previously described [4]. Inhibitor controls involved the addition of 5 mM reserpine (in acetic acid/distilled water). Once stained, cells were maintained at 4°C during antibody labeling and until analysis/sorting.

Antibody labeling was performed in phosphate-buffered saline (PBS), 2% fetal calf serum (FCS), and 10 mM HEPES for 30 minutes against appropriate matched-isotype controls. This study used the following antibodies (all from BD Pharmingen, San Diego, http://www.bdbiosciences.com/index_us.shtml): fluorescein isothiocyanate (FITC)-conjugated anti-Thy-1, CD31, and CD34; phosphatidylethanolamine (PE)-conjugated anti-CD2, CD3, CD4, CD8, CD11b, CD135, B220, Ter-119, Gr-1, and NK1.1; biotinylated anti-Sca-1 (subsequently labeled with PerCP-conjugated streptavidin); and allophycocyanin (APC)-conjugated anti-CD117. The lineage cocktail used throughout contained PE-conjugated anti-CD5, CD11b, B220, Gr-1, and Ter-119 antibodies. All Hoechst-stained cells were resuspended in PBS, 2% FCS, 10 mM HEPES, and 2 mg/ml propidium iodide (PI) and passed through a 70-mm mesh before analysis/sorting.

Flow Cytometric Analysis

Analysis/sorting was performed on a MoFlo flow cytometer (DakoCytomation, Glostrup, Denmark, http://www.dakocytomation.com) equipped with laser lines of 350.7–356.4 (for UV), 488, and 633 nanometers. Red and blue fluorescence derived from UV excitation was separated via a 610 dichromatic longpass. Blue Hoechst fluorescence was collected with a 424/44 bandpass filter and red Hoechst/PI fluorescence via a 620 longpass filter. FITC, PE, PE-cyanine-5, and APC signals were collected via 530/40, 570/40, 675/20, and 670/40 bandpass filters. Fluorescence compensation was typically performed during acquisition. Hoechst/PI red and Hoechst blue fluorescence signals were displayed on a linear, dual fluorescence dot-plot. A rectangular gate was drawn to exclude PI+ dead cells (far right of plot) and unstained debris (origin of plot) as previously described [4]. First Sca-1+/Linlow and then CD117+ cells were readily identified as discrete populations within this gate.

Competitive Repopulation Assays

Ly5.1 C57Bl6 mice (8–12 weeks) were lethally irradiated (two doses of 500 rads given four hours apart). Sorted Ly5.2 C57Bl6 donor cells were administered in PBS via tail vein injection along with 200,000 recipient-type (Ly5.1) whole bone marrow cells. Bone marrow or peripheral blood from surviving mice was analyzed for donor cell content on the flow cytometer. Analysis involved labeling of red-cell-lyzed 4,6-diamidino-2-phenylindole (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) negative cells with anti-Ly5.1-PE and anti-Ly5.2-FITC antibodies (both from BD Pharmingen). Competitive repopulation units were calculated as (percentage engraftment due to donor cells/percentage engraftment due to recipient cells) × (2 × 105 recipient competitor cells/number of donor cells injected).

Annexin-V Labeling

Following Hoechst staining, certain experiments were examined for Annexin V expression to assess apoptotic cells. We added 100 μl of 10× Annexin V Binding Buffer (BD Pharmingen) to 900 μl of the resuspended cells and mixed. Then, 5 μl of directly conjugated Annexin V Alexa Fluor 647 (Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com) was added to the cells before incubation at 37°C for 15 minutes. PI was added to the cells, as above, prior to analysis on a BD Pharmingen LSR-2 flow cytometer.

Gene Expression Analysis

RNA was extracted using the RNeasy Micro Kit and reverse transcribed using the Sensiscript Kit according to the manufacturer's instructions (Qiagen, Hilden, Germany, http://www1.qiagen.com). Also, DNase treatment was performed to clean possible genomic DNA contamination. For quantitative real-time polymerase chain reaction (PCR), SybrGreen master mix reagent (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com) with 2–10 ng of cDNA/25 μl of reaction was used according to the manufacturer's instructions. Amplification was performed using the ABI Prism 7700 sequence detection system (Applied BioSystems). To avoid the possibility of amplifying contaminating DNA and unprocessed mRNA, most primers were designed (when possible) to anneal the end parts of two exons, and therefore each amplicon covers at least two or more distant exons. The specificity of the PCR products was verified by running a 2% agarose gel and also using the dissociation curve V1.0 software (Applied BioSystems). Primers used in this study were: p16 forward, CGAACTCTTTCGGTCGTACCC, p16 reverse, CGAATCTGCACCGTAGTTGAG; HOXA9 forward, CCGGACGGCAGGTATATGC, HOXA9 reverse, GTGAGTGTCAAGCGTGGGAC; FAS forward, ATGACGGATCTCGTAGCTGTT, FAS reverse, CGTTTCTTTGCCCACCAACTT; P21 forward, GTGGCCTTGTCGCTGTCTT, P21 reverse, GCGCTTGGAGTGATAGAAATCTG; HOXB4 forward, TCCGAGCGCCAGATCAA, HOXB4 reverse, CCGAGCGGATCTTGGTGTT; p27 forward, GTTAGCGGACGAGTGTCCAG, p27 reverse, TGTTCTGTTGGCCCTTTTGTT; ABC/G2 forward, GAACTCCAGAGCCGTTAGGAC, ABC/G2 reverse, CAGAATAGCATTAAGGCCAGG; BCL-2 forward, ATGCCTTTGTGGAACTATATGG, BCL-2 reverse, GGTATGCACCCAGAGTGATGC; BAD forward, CTCCGAAGGATGAGCGATGAG, BAD reverse, CCACCAGGACTGGATAATGCG; and Bmi forward, TTATGCAGCTCACCCGTCAG, Bmi reverse, GCTGGGCATCGTAAGTACCTTT.

Results

The Absolute Frequency of SP Cells Increases Throughout Life

Summarized in Figure 1 are the results of our analysis of the percentage of SP cells in mice from various ages. Consistent with previous reports, we found that the percentage of total nucleated cells that was capable of sufficient Hoechst efflux to be classified as a SP cell was approximately 0.07% in 8–12-week-old mice [4, 14, 16]. However, a lower SP percentage was detected in marrows from younger mice and a higher SP percentage was observed in older mice. Indeed, the percentage of total nucleated cells that is an SP cell steadily increases with age from 0.01% in young mice (49 days old) to over 0.25% in the oldest mice we have analyzed (950 days old). Although it has been reported that primitive subsets slightly increase with aging [9, 13], the SP increases more dramatically with age and may suggest that the SP, or rather how it is formed, the ABC/G2 pump, is involved in the mechanism of aging.

Figure Figure 1..

Absolute increase in SP frequency with age. All mice were assessed for Hoechst exclusion as described in Materials and Methods. The percentage of total nucleated cells that is an SP cell increases during murine life from 0.01% to more than 0.25% in the oldest mice we analyzed. Abbreviation: SP, side population.

SP Cells from Old Mice Are Enriched for Cells with the Highest Efflux Potential

It has been proposed that a hierarchy of stem cell potential within the SP subset exists. Cells that have the highest Hoechst efflux activity and, hence, appear closest to the origin of a Hoechst red versus blue dot-plot are thought to have the highest repopulation potential [4, 17]. To explore this aspect further, we geometrically divided the SP into “upper” and “lower” SP fractions. As illustrated in Figure 2, SP cells from older mice contain a significantly higher proportion of lower SP cells when compared to SP cells from younger mice (56.3% ± 3.4% vs. 35% ± 3.1% of SP cells were lower SP cells in 414-day- and 61-day-old mice, respectively; n = 6, p < .001).

Figure Figure 2..

The proportion of lower SP cells increases with age. (A): Example of SP profiles from mice aged 61 days and 414 days. In this example, only 9% of SP from young mice are “lower” SP cells, whereas SP cells from old mice contain a higher proportion of cells that are capable of the greatest Hoechst efflux (37%, in this example). (B): Scatter plot of the percentage of SP cells that falls within the lower SP region. Each dot represents the mean of at least three assessments. The proportion of SP cells that is capable of the greatest Hoechst efflux increases with age. (C): Four SP profiles that provide an example of the progressive increase with age in the proportion of SP cells that are lower SP cells. The same number of total cells is displayed on each plot. As the age of the mice increases, so does the proportion of SP cells that is in the lower SP portion. Abbreviation: SP, side population.

The Relationship Between c-kit+, Lineage, Sca-1+ Cells and SP Cells Is Altered Throughout Life

The c-kit+, lineage, Sca-1+ (KLS) cell population contains virtually all the stem and progenitor cells in the murine marrow [18]. Consistent with previous reports, the absolute number of KLS cells increases slightly with age (Fig. 3A) [9]. In adult mice (8–12 weeks), the small proportion (approximately 5%) of KLS cells that is an SP cell contains the majority of the long-term repopulating ability of the KLS population and, hence, the murine marrow [14].

Figure Figure 3..

The proportion of KLS cells that is an SP cell increases during murine life. (A): Displayed is the percentage of total live, nucleated cells that is a KLS cell. Although there is some variation within a small range, there is only a slight increase in the absolute frequency of KLS with age. (B): Displayed is the percentage of KLS cells that is an SP cell. The proportion of KLS that is an SP cell increases from approximately 1% in young mice to over 50% in the oldest mice we have analyzed so far. Abbreviations: KLS, c-kit+, lineage, Sca-1+; SP, side population.

We report here that the proportion of KLS cells that is an SP cell greatly increases with age (Fig. 3B). The proportion of KLS that is an SP cell increases from 1% in younger mice to over 50% in older mice. In younger mice, it is known that the SP fraction of KLS cells contains the majority of the repopulation population in the murine marrow [14]. To confirm that the SP in older mice correlated with HSCs in our system, we further examined the phenotype and function of SP cells from older mice.

CD48/CD150 Expression Suggests a Stem Cell Status for SP Cells from Older Mice

It has been recently reported that the CD48/CD150+ fraction of murine marrow contains the majority of the repopulation ability in murine marrow [19]. Furthermore, the stem cell status of CD48/CD150+ is conserved in older mice and throughout aging [20]. To examine the stem cell status of SP cells from various ages, we examined the relationship between SP cells, KLS, CD48/CD150+ cells, and age.

Interestingly, SP cells from both young and old mice contained a similar proportion of CD48/CD150+ cells (58.2% ± 8.8%, n = 9: 46, 78, 94, 117, 256, 416, 502, 535, and 696 days), indicating that SP cells from old mice do indeed still represent the stem cell population within the murine marrow in mice of various ages. This has implications for the relationship between the CD48/CD150+ cell population and KLS cells. The proportion of KLS cells that is CD48/CD150+ increases from approximately 10% in younger mice to more than 50% in older mice (Fig. 4). Hence, there is also an increase with age in the absolute number of CD48/CD150+ cells found in the murine marrow (data not shown).

Figure Figure 4..

Analysis of the expression of CD48/CD150+ cells. Scatter plot that displays the proportion of KLS cells that is negative for the CD48 antigen but positive for the CD150 antigen. This proportion progressively increases with age in a similar fashion to side population cells. Abbreviation: KLS, c-kit+, lineage, Sca-1+.

More Committed Cells or Progenitors Have Not Gained the SP Ability in Older Mice

As defined by phenotype, the SP compartment contains a similar proportion of committed cells in older mice when compared to SP cells from younger mice. Although the difference did not reach statistical significance, the mean proportion of SP cells that do not express lineage antigens is actually slightly lower in older mice (<350 days old = 62.2% ± 20.0% [n = 5]; 414–950 days old = 78.2% ± 17.1% [n = 8]).

To test whether more progenitors gain the ability to efflux Hoechst during murine aging, we performed colony-forming assays on SP cells from mice of various ages. Single cells were sorted into each well of a 96-well plate containing 200 μl of cytokine-supplemented methylcellulose before incubation at 37°C for 2 weeks. Although KLS cells produced colonies, SP cells from both old and young mice demonstrated a very low cloning efficiency in methylcellulose (less than 2% of wells produced colonies in three plates for each of 79- and 362-day-old mice).

SP Cells from Old Mice Are Still Capable of Long-Term Engraftment but Compete Poorly with Cells from Younger Mice

To test the stem cell activity of SP cells from both young and old mice, we performed a long-term competitive repopulation assessment. We injected 100 SP-sorted cells from young and old mice into 8–12-week-old recipients along with 2 × 105 recipient-type cells. Peripheral blood samples were taken at 4 months and assessed for donor/recipient cell content. Competitive repopulation units were calculated as described in Materials and Methods.

Consistent with previously published reports that utilized immunophenotypic cell isolations, the long-term repopulation potential of SP cells from old mice was reduced when compared to those from younger mice. We observed a 3.7-fold decrease (p = .0031) in long-term repopulating ability when we compared 61-day-old mice (mean 1,810 competitive repopulation units (CRU); n = 5) to 414-day-old mice (mean 491 CRU; n = 5).

Since it has been previously reported that there is a defect in homing in older mice, we investigated whether the lower engraftment we observe with SP cells is also due to homing. We therefore compared the engraftment (CRU) from i.v. administration to that of direct bone marrow injection.

Engraftment derived from younger mice (61 days) via direct bone marrow administration was 2,554 CRU (100 SP cells, n = 4), whereas engraftment derived from old SP cells also administered directly into the bone marrow was 1,860 CRU (100 cells, n = 4, 414 days old). However, engraftment due to i.v. administration reported above revealed a greater difference between engraftment derived from young and old mice (young = 1,810 CRU, [n = 5]; old = 491 CRU, [n = 5]). Since direct bone marrow injection circumvents the process of homing and resulted in greatly improved engraftment of cells derived from old mice, this indicates that a large proportion of the engraftment defect we report for old stem cells is due to a deficiency in homing. This observation is consistent with previous reports, but we can now confirm this with direct bone marrow injection in a long-term model [21].

SP analysis of the marrows of engrafted mice revealed that cells derived from old SP cells still possessed a higher SP frequency when compared with cells derived from young-mice SP cells. The frequency of SP cells in engrafted mice seems to have increased in a similar fashion to nontransplanted cells of the same age. For instance, engrafted marrow cells derived from SP cells from 61-day-old mice contained a proportion of SP cells that was very similar to 61 days plus the 4 months engraftment time (approximately 180 days in total), and these also made up a similar proportion of KLS to SP cells from 180-day-old mice (Fig. 5, lower row). This was also true for SP cells derived from older mice; the SP cell frequency had also increased during the 4-month-engraftment period to a similar level to mice of the same age as transplanted cells. In addition, engrafted SP cells derived from old mice also accounted for an increased proportion of KLS than when they were originally injected, similar to the level of mice of the same age as the engrafted cells (Fig. 5, upper row).

Figure Figure 5..

Phenomenon seems to be independent of the microenvironment. Example of the change in side population (SP) to c-kit+, lineage, Sca-1+ (KLS) ratio that occurs during transplantation. The upper row displays the patterns observed in transplanted cells from older mice, whereas the lower row displays the patterns seen in transplanted cells from younger mice. Plots on the far left are of CD45.1 versus CD45.2 isoform-staining that allows the discrimination of donor and recipient cells, respectively. The middle plots are Hoechst blue versus red SP profiles and demonstrate that the percentage of SP has increased in a similar fashion to nontransplanted cells. The far right plots display the KLS population from young and old mice and illustrate that these SP cells make up a larger proportion of KLS cells than when originally injected. Recipient cells displayed an SP frequency that was appropriate for the age of the mice (0.07%–0.09% of total).

This increase in SP cells with age therefore seems to be an intrinsic cellular property that is mostly independent of the microenvironment, as it occurs within the aging marrow as well as upon transplantation to a younger microenvironment. However, we irradiated the marrows of recipient mice prior to engraftment and, as such, may have masked the contribution of the age of the microenvironment.

Apoptosis Assessment of SP Cells from Mice of Various Ages

Toxicity to stem cells is thought to have an important role in aging. Since the ABC/G2 transporter that is responsible for SP formation may have an antitoxicity role and the number of SP cells, especially lower SP cells, increases with age, we wondered whether ABC/G2 activity may give us a clue as to the mechanism of SP stem cell frequency changes. It is thought that, upon damage, hematopoietic stem cells (HSCs) die by apoptosis rather than risk a replication error [2]. It may be that cells with a lower ABC/G2 activity are more susceptible to damage and, hence, more likely to die by apoptosis. Conversely, cells with a higher ABC/G2 activity may be more resistant to damage-induced apoptosis. In this way, selection may occur throughout life within the SP population for cells with a higher ABC/G2 activity (lower SP cells).

To examine this hypothesis, we analyzed the apoptosis of SP cells from young/old mice upon challenge with a toxic substrate of ABC/G2 (Hoechst 33342). To this end, we labeled freshly Hoechst-stained cells with anti-annexin-V as described in Materials and Methods. The majority of SP cells that were annexin-V+ were upper SP cells in both old and young mice. Indeed, in 59–68-day-old mice (n = 3), 85.4% ± 13.1% of annexin-V+ SP cells were upper SP cells, and in 551–553-day-old mice (n = 3), 60.3% ± 17.0% of annexin-V+ SP cells were upper SP cells. Intriguingly, the proportion of annexin-V SP cells that were lower SP cells was significantly higher in old mice when compared with young mice (70.3% ± 12.4% of annexin-V SP cells were lower SP in old mice compared with 20.7% ± 14.8% of annexin-V SP cells in young mice; p = .048, n = 6).

Gene Expression Analysis

Summarized in Figure 6 are the results of real-time quantitative reverse transcription PCR analysis of SP cells from young and old mice. Key genes were analyzed that are involved in HSC cell cycle regulation (p21 and p27), self-renewal (Bmi-1, HoxB4, HOXA9), apoptosis (Bcl-2, Bad, and Fas), and antitoxicity (ABC/G2). Values were normalized to the expression in young SP cells. There were no significant differences in the expression of ABC/G2, p21, and Bcl-2. A slightly lower expression of HoxA9 and Bmi-1 was observed in cells from older mice. There was, however, a significant reduction in the expression of p27, Bad, Fas, and HoxB4. In SP cells from young mice, p16 was not detectable, but it was present in very low amounts in SP cells from old mice (example of real-time PCR data and gel in Fig. 6B and 6C). Since Fas and Bad are proapoptotic genes, a decrease in expression may indicate a lower sensitivity to apoptosis. Such a gene expression profile is consistent with a reduction of apoptosis-sensitive cells over time and, hence, damage-based selection.

Figure Figure 6..

Analysis of expression of key genes. (A): Expression of selected genes determined by quantitative reverse transcription-polymerase chain reaction (PCR) in side population (SP) cells isolated from Y and O mice. The relative mRNA expressions were obtained after normalizing the cycle threshold (CT) values from each gene with the internal PCR control (GAPDH), then using the ΔCT values from Y SP cells as a reference. All PCRs were performed using 2 ng of equivalent cDNA. All quantitative results shown are from two independent experiments, each performed in duplicate. Expression of p16INK4a (p16) within the SP cells was restricted to cells isolated from old mice. The very low but detectable expression of p16 in O SP cells was only seen when 5 ng of equivalent cDNA was used per PCR reaction, as exemplified by the amplification curve (B). However, increasing up to 10 ng of equivalent cDNA was not sufficient to detect any expression of p16 in SP cells from young mice, as shown by the representative gel electrophoresis (B). GAPDH was used as a control. Abbreviations: GAPDH, glyceraldehyde-3-phosphate dehydrogenase; NTC, no template control; O, old; Rn, normalized reporter; Y, young.

Discussion

We report here an investigation into the relationship between SP formation and aging. Considering the larger amount of SP in embryonic tissues and their reduction in adult tissues, one could expect the SP cell frequency to continue to reduce as adulthood progresses [22]. In contrast, we observed a large increase in SP cell frequency in old mice when compared with SP cells from younger mice. This is consistent with some preliminary observations by Rossi et al., but we can now confirm this increase in SP formation in a larger group of samples that encompass a greater range of murine age [13]. Indeed, we have looked at mice up to 950 days old, whereas previous studies have examined mice up to approximately 650 days old [9, 13, 20, 21].

We have investigated age-based changes in stem cells in the C57Bl/6 strain, exclusively. Previous studies have utilized mice strains with alternative genetic backgrounds to investigate age-based changes in stem cells [23, 24]. However, models such as the Diamond-Blackfan anemia (DBA) strain that show a decline in stem cell function do not represent the wild-type situation that we are interested in. DBA is an insidious disease that features a slow, progressive decline in hematopoietic cell output during life and, since this may be corrected by a bone marrow transplant, is probably due to dysfunction of hematopoietic stem cells.

It has been reported that the KLS population contains the majority of stem and progenitor cells found in the murine marrow [18]. The KLS cell population increases approximately twofold with age, whereas the SP fraction increases over 30-fold with age. It may be that the SP, or rather how it is formed (via a high ABC/G2 activity), is more closely related to the change in primitive cell frequencies with age than non-SP KLS cells.

One remarkable aspect of the increase in SP cells we report is the dramatic increase in the proportion of cells with a high ABC/G2 activity. It may be that this increase in ABC/G2 activity is a consequence of aging of the HSC pool or, alternatively, may have a more causative role.

The age-dependent increase in SP we describe may represent the upregulation of ABC/G2 either on cells that did not previously express ABC/G2 or on cells that already express physiologically relevant levels of ABC/G2 but are not revealed by the SP technique. This would result in a higher frequency of progenitors or committed cells with SP ability. One can envisage how the need for an antitoxicity pump would spread to other cell types during aging. Alternatively, it may be that there are more stem cell SP cells in older mice and that these make up a larger proportion of stem/progenitor cells later in very late life.

We have tested both of these possibilities via analysis of the phenotype progenitor cell output and in vivo long-term repopulation. Phenotyping and progenitor assays revealed that SP cells from older mice contained a similar proportion of committed cells and progenitors, which indicates to us that more committed cells and progenitors have not gained the SP ability through upregulation of the ABC/G2 pump.

Since the stem cell status of KLS+/CD48/CD150+ cells is conserved in both young and old mice, we utilized this phenotype to confirm the stem cell status of SP cells from mice of various ages. The proportion of SP cells that is a KLS+/CD48/CD150+ cell is constant throughout aging, indicating that SP staining does indeed highlight phenotypically defined HSCs throughout aging. Consequently, the proportion of KLS that is a CD48/CD150+ cell increases with age, as does the absolute frequency of CD48/CD150+ cells.

To investigate the stem cell potential of SP cells from old mice, we compared their long-term repopulation potential in a competitive model to that of SP cells from younger mice. Consistent with previous reports, we observed a three- to fourfold reduction in long-term repopulation ability when purified HSCs from old mice were compared with those from young mice [13, 20, 21]. Consistent with recently published work, we can state that much of the deficiency in long-term engraftment may be attributed to a defect in the homing of primitive cells to the marrow [21].

When combined with evidence in the literature, our data may provide some clues as to how this alteration in the proportion of stem/progenitor cells occurs. It may be that hematopoiesis is a clonal expansion from single HSCs, and that not all HSCs are contributing to hematopoiesis at any one time [25, 26]. It has been reported that HSCs are more likely to die by apoptosis upon DNA damage than risk a replication error [2]. Indeed, various tumor suppressor mechanisms have been recently demonstrated to have a role in the mechanism of aging [27, [28], [29], [30]–31]. Interestingly, it seems that modulation of the expression of p53 and p16 can directly affect the frequency and function of stem cells [28, 31]. The differential ability of antitoxicity mechanisms that prevent DNA damage within the stem cell pool may be important in the decision as to which cell is going to undergo this tumor suppressor-mediated cell death/senescence. ABC/G2 is probably part of the extensive antitoxicity system that is present in hematopoietic stem cells. Given that there is heterogeneity in ABC/G2 activity within SP cells, there may be varying sensitivity to damage induced by ABC/G2 substrates. Specifically, it may be that cells with a lower ABC/G2 activity are more susceptible to damage and, hence, more likely to die by apoptosis. Conversely, cells with a higher ABC/G2 activity may be more resistant to damage-induced apoptosis. In this way, selection may occur throughout life within the SP population for cells with a higher ABC/G2 activity (lower SP cells). Hence, evolution may occur within the stem cell pool over time. Our annexin-V apoptosis data are consistent with this suggestion. In addition, the expression of proapoptotic genes is higher in SP cells from young mice than in cells from old mice.

There may be cells with physiologically relevant ABC/G2 activity that are not revealed by the SP technique. Some long-term repopulating cells are found in the SP/KLS+ fraction, albeit at a much lower frequency than in the SP+/KLS+ cells [14]. Natural selection within the HSC pool based on ABC/G2 activity would presumably reduce this nonresolved SP population and increase the number of cells with a higher ABC/G2 activity and, hence, increase the number of detected SP cells. Such a toxicity-resistance-based selection mechanism may account for the increase in proportion of cells with the highest ABC/G2 activity (lower SP cells). Presumably, a selection mechanism would also be independent of the microenvironment, as HSCs would continue to be assaulted by damage regardless of the age of the microenvironment.

It has been reported that the SP actually represents a quiescent population of HSCs [32]. Campisi [33] has suggested an elegant aging model of antagonistic pleiotropy in which it is proposed that mechanisms important for the prevention of carcinogenesis may actually become detrimental to the cell system later in life. Apoptosis and cellular senescence are both examples of such mechanisms, as they usually serve to clear damaged cells to prevent carcinogenesis during early adulthood but later in life may be responsible for the degeneration of stem cell-driven organs, such as hematopoiesis. Although most of the reduction in competitive repopulation potential we report here may be attributed to a homing defect, the disassociation between the SP phenotype and purity of cells capable of 4-month, long-term repopulation may also be due to the accumulation of senescent stem cells. Consistent with a senescent phenotype, we observed a slight increase in p16 expression in SP cells from older mice.

We have focused on ABC/G2, as this is the pump responsible for SP formation, but other important mechanisms are probably essential for the antitoxicity system that is active in HSCs. Heterogeneity within these mechanisms may contribute to the natural selection process. Indeed, it has been reported that there are elevated levels of expression of various pumps (ABC/A4, ABC/B1a, and ABC/C1) in stem cells from older mice when compared with young mice [13]. We are further investigating this hypothesis in a follow-up study. We intend to track the proliferative history of HSCs in a long-term repopulation model to determine if some SP cells are deleted by apoptosis or removed by senescence.

If we accept that there is a decline in stem cell function with age, then we can envisage that increasing the number and relative quality of HSCs may compensate for this intrinsic lack of function. In this way, it may be that a system that has evolved to cope with cellular damage in stem cells is also compensating for a lack of cellular potential by increasing/improving the stem cell pool.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Acknowledgements

This work was supported by Cancer Research UK. The authors acknowledge the technical assistance of all staff in the Biological Research Facility and thank Derek Davies, Malcolm King, Kirsty Allen, Lola Martinez, Carolyn Koh, and Laura Bazley (FACS Laboratory, Cancer Research UK) for their cytometric expertise and helpful discussions.

Ancillary