|Home | About | Journals | Submit | Contact Us | Français|
Cellular and molecular mechanisms involved in aging are notoriously complex. Aging-related immune decline of T lymphocyte function is partly caused by attrition of thymic T cell development, which involves programmed creation and repair of DNA breaks for generating T cell receptors. Aging also leads to significant alterations in the cellular DNA repair ability. We show that higher levels of gamma-phosphorylated H2AX (pH2AX), which marks DNA double-stranded breaks (DSBs), were detectable in early thymocyte subsets of aged as compared to young mice. Also, while only 1–2 foci of nuclear accumulation of pH2AX were detectable in these early thymocytes from young mice, cells from aged mice showed higher numbers of pH2AX foci. In CD4-CD8- double-negative (DN) thymocytes of aged mice, which showed the highest levels of DSBs, there was a modest increase in levels of the DNA repair protein MRE11, but not of either Ku70, another DNA repair protein, or the cell cycle checkpoint protein p53. Thus, immature thymocytes in aged mice show a marked increase in DNA DSBs with only a modest enhancement of repair processes, and the resultant cell cycle block could contribute to aging-related defects of T cell development.
Aging at the cellular and molecular level has been shown to involve a decline in the immune system in several different ways. Although a decrease in immune function is seen with age (Refaeli et al., 1998; Jiang and Chess, 2004; Gupta, 2005; Min et al., 2005; Vasto et al., 2006), no mechanism has yet been elucidated to explain all the deleterious effects of aging on the immune response. A wide array of defects in the aged immune system has been identified including problems in T cell development such as thymic involution, decreased TCR variability (Goronzy and Weyand, 2005) and a decrease in the output of circulating naïve T cells (Haynes and Eaton, 2005; Min et al., 2005). A part of the thymic defect in aging has been attributed to lack of the essential growth factor IL-7 from thymic stromal cells (Andrew and Aspinall, 2001; Aspinall and Andrew, 2001; Min et al., 2005; Taub and Longo, 2005). However, the cellular and molecular attributes of developmental dysfunction that lead to cell loss in aging are not yet clearly understood (Gupta, 2005; Min et al., 2005; Taub and Longo, 2005; Vasto et al., 2006).
Several hypothesizes exist to explain aging as a function of overall failure of cellular and molecular processes. The accumulation of somatic DNA damage, possibly related to free radical-mediated effects (Hasty, Campisi et al. 2003), as well as a diminished repair capacity of mitochondrial and nuclear DNA damage has been shown to contribute to aging in a wide variety of species (Wojda and Witt, 2003). Double stranded DNA breaks (DSBs) are lethal DNA lesions if not rapidly repaired (Michel et al., 1997; van Gent et al., 2001; Kuhne et al., 2003; Sedelnikova et al., 2003; Ting and Lee, 2004). One of the earliest molecular events occurring at the location of a DSB is the recruitment of gamma-H2AX histone which is phosphorylated. H2AX phosphorylation is an early ATM kinase-dependant response to DSBs and it is speculated that the pH2AX may help keep broken DNA ends together (Sprent and Surh, 2002; Franco et al., 2006). Previous evidence indicates that one nuclear pH2AX focus is equivalent to one DSB and about 200 H2AX molecules are phosphorylated per DSB (Sprent and Surh, 2002). H2AX functions to recruit DSB repair proteins and to stabilize the broken ends of a DSB (Ting and Lee, 2004; Franco et al., 2006) , and is necessary for the G2/M cell cycle checkpoint competency (Fernandez-Capetillo et al., 2002). Failure to resolve DSBs results in apoptotic cell death (Ting and Lee, 2004).
Normal thymic development involves the controlled creation of DSBs at T cell receptor (TCR) loci. The most immature subset of thymic T lineage cells are CD4-CD8- (double-negative; DN), and development proceeds to the CD4+CD8+ (double-positive; DP) stage and then onwards to becoming either CD4+CD8− or CD8+CD4− (single-positive; SP) mature T cells (Sebzda et al., 1999). The locus coding for the heavy chain of the TCR, the TCR-beta locus, undergoes controlled induction of DSBs. Regulated formation of DSBs is the initial step of the VDJ recombination process by which the coding exon of the variable region of TCR-beta is formed at the DN stage, followed by the TCR-alpha locus at the DP stage (Sebzda et al., 1999).
In this context, we have examined the extent of DSBs detectable in thymocytes from aged mice, and show an enhanced persistence of DSBs in them, a finding relevant for understanding the deleterious effects of aging on the immune system.
C57Bl/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME) or Harlan Laboratories (Indianapolis, IN) or bred in our colony at the University of Arkansas Central Laboratory Animal Facility. Mice of 2–4 months of age were considered young adults and mice greater than 16 months of age were considered aged. While many studies on mouse models of aging take mice at 18 months or more as ‘aged’, thymic involution has been reported to be progressive and notable even in 12 month-old mice (Aspinall and Andrew, 2000; Sempowski et al., 2002). We therefore used mice older than 16 months as our ‘aged’ group. Data are representative of findings in 16 month-old as well as older mice. Either male or female mice were used, animals were sex-matched in individual experiments. All experiments were done with the approval of the University of Arkansas Institutional Animal Care and Use Committee.
The reagents used were: TO-PRO-3 (Invitrogen Carlsbad, CA), propidium iodide (PI; BD, San Jose, CA), mouse monoclonal antibody against pH2A.X (Invitrogen, Carslbad, CA), MRE11 (BD Biosciences San Jose, CA) and Ku70 (BD Biosciences San Jose, CA) followed by a goat Alexa 488 conjugated antibody against mouse IgGs (Invitrogen Carslbad,CA). The anti-pH2A.X antibody recognizes only the gamma-phosphorylated form of H2A.X and has been previously used for detecting this form of H2A.X recruited to DSBs (Rogakou et al., 1998; Bonner et al., 2008) . Antibodies against CD4 conjugated to either APC or FITC (BD Biosciences San Jose, CA), against CD8 conjugated to either PE or FITC (BD Biosciences San Jose, CA), and p53 conjugated to PE (BD Biosciences San Jose, CA).
Single-cell suspensions from thymus or spleen were made using sterile glass slides (Erie Scientific, Portsmouth, NH) or nylon mesh (Bally Ribbon Mills, Bally, PA) in sterile PBS. Cells were stained at 1 × 106 cells/mL using stains optimally titrated previously. Cell viability was routinely tested prior to staining by trypan blue dye exclusion, and was >95%. Cells were stained on ice using 30-min incubations, followed by washes with PBS. Non-binding immunoglobulins were used as appropriate to account for non-specific binding. Where intracellular staining was required cells were fixed with paraformaldehyde (Sigma-Aldrich, St. Louis, MO) in PBS to a final concentration of .05% for 1 h and washed, before being treated with primary and secondary antibodies diluted in 0.5% saponin (Sigma-Aldrich, St. Louis, MO). At least 100,000 cells were analyzed for each sample using a FACSort flow cytometer (Becton Dickinson, Lincoln Park, NJ), and no individual sample analyzed contained less than ~4,000 DN cells. Data analysis, including the calculation of MFI, was done with the FlowJo (Tree Star, San Carlos, CA) software package. Residual dead cells were eliminated from the analysis by using scatter gating on the forward scatter (FSC) parameter.
Cells were fixed and stained as described above for flow cytometry. Samples were mounted in ProLong Gold anti-fade reagent (Molecular Probes, Eugene, OR), covered with 1.5 mm coverslips and analyzed on a Leica TCS SP2 AOBS laser-confocal microscope (Leica Microsystems, Heidelberg, Germany). Light detection was optimized for the specific fluorescent probes used. Foci of pH2A.X were counted from at least 100 cells from each thymocyte subpopulation, while excluding apoptotic cells. Images were obtained and analyzed using Leica confocal software. For three-color analyses, spectral overlap was manually corrected using single-color stained samples and optimization of laser intensities and detectors to minimize crosstalk between channels and sequential scanning of each individual channel, and residual overlap was subtracted using the Leica confocal software.
In some cases, magnetic beads and columns were used to separate subpopulations of cells before experimentation. Separations were completed using MS MACS magnetic separation columns (Miltenyi Biotec Auburn, CA) per the manufacturer’s described protocol. Cells were labeled with the appropriate antibodies, including biotin conjugated antibodies against B220, Mac-1, CD4, CD8, and/or CD3 (all from BD Biosciences San Jose, CA) and then incubated with streptavidin conjugated magnetic beads (Dynabeads M-280). The MS column was then placed in a magnetic field and cell suspensions were run through. Both the positive and negative fractions were collected. Cell separation was used in order isolate relatively low-frequency subpopulations as indicated.
Mean, standard deviation and standard error were calculated for all experimental populations. Significance was tested by Student’s ‘t’ test comparisons between young and aged populations. Mean fluorescence intensity was calculated using FlowJo software (Tree Star). The analysis software uses statistical calculations to determine fluorescence intensity of the data obtained by each individual detector. This determination takes into account a variety of factors including logarithmic or linear scale and other instrument settings. Once this fluorescence intensity for the data of interest relative fluorescence intensities was determined as follows
RFI = Fluorescence Intensity of experimental sample - Fluorescence Intensity of control. An arithmetic mean of the fluorescence intensities (MFI) was calculated by combining multiple experiments.
In keeping with previous reports, the thymus in aged C57BL/6 mice was hypocellular (Figure 1A). Three-color flow cytometry (with gates as shown in Fig. 1B) was used to estimate the extent of detectable DSBs in various thymocyte subpopulations defined by expression of CD4 and CD8. Anti-gamma-phospho-H2AX antibody was used for this purpose, since pH2AX levels correlate well with the presence of DSBs as described earlier (Pilch et al., 2003; Sedelnikova et al., 2003). Upon phosphorylation, H2AX is recruited to DSB sites(Rogakou et al., 1998; Bonner et al., 2008) and flow cytometric detection of pH2AX correlates with increased numbers of microscopy-detectable foci and with DSBs (Sedelnikova et al., 2002; Rothkamm and Lobrich, 2003; Qvarnstrom et al., 2004; Sedelnikova et al., 2004; Lobrich et al., 2005). On this background, we have used pgH2AX levels per cell as a marker for DSBs, as have other studies (Bonner et al., 2008). In young mice, DN thymocytes showed higher intensity of staining for pH2AX than in DP thymocytes (Fig. 2A). Curiously, while CD4 single-positive (SP) thymocytes did not show any detectable pH2AX, CD8 SP thymocytes did (Fig. 2A). There were no major differences in the forward scatter profiles for each population between aged and young mice (Fig. 2B), ruling out the likelihood of higher pH2AX levels per cell simply as a consequence of larger cell volumes.
However, all four thymocyte subsets in aged mice showed notably higher levels of pH2AX (Fig. 2A). The analysis of fluorescent intensities showed significantly higher pH2AX levels in aged mouse DN and DP thymocytes (Fig. 2C). These data indicated that immature thymocytes from aged mice were likely to have more DSBs than their counterparts from young mice.
Since pH2AX is reported to accumulate at DSB sites (Pilch et al., 2003; Sedelnikova et al., 2003), we next examined the numbers of foci of pH2AX accumulation per cell in thymocytes at various stages from young and aged mice. For this, thymocyte populations were stained in a three-color analysis for stage identification and for pH2AX and examined by fluorescence microscopy. For DP, CD4SP or CDSP cells, we stained whole, unfractionated thymocyte populations. These cells were stained for CD4 versus CD8 versus pH2AX using three colors, and DP, CD4SP and CD8SP cells were identified by the CD4 and CD8 staining as shown, for counting pH2AX foci. For DN thymocytes, we used thymic cells DN-enriched by MACS using negative selection for the B220, Mac-1, CD4, CD8, and CD3 markers. These DN-enriched cells were stained for Thy-1 versus CD4+CD8 (in the same color) versus pH2AX. The Thy-1 staining identified individual cells as T lineage cells since most DN stage thymocytes do express Thy-1 while non-T lineage cells in the thumus do not do so (data not shown; and Shortman and Wu, 1996). The lack of CD4 and/or CD8 staining on these cells excluded the possibility that they were contaminating DP or CD4/8SP cells allowing a reasonable assumption that they were genuine DN T lineage cells in which pH2AX foci could be counted. This allowed us to phenotype each cell in which pH2AX foci were being counted.
At least 100 cells from each subpopulation were counted and the average numbers of pH2AX foci per cell determined. The numbers of pH2AX foci per cell were substantially higher in all subsets of thymocytes from aged mice than in those from young mice (Fig. 3) with the largest difference apparent in the DN population. We examined the distribution of the number of breaks per cell in various thymocyte populations by plotting the numbers of cells with a given number of DSB foci (Fig. 4). As can be seen, the data show a broadly unimodal distribution, with many aged cells, particularly in the DN group, having very high numbers of DSB foci. Also, the distribution moves leftwards as the maturation progresses in both aged and young cells, indicating that the major difference in DSB frequency in the DN population may be an important factor in the thymic dysfunction of aged mice.
We similarly evaluated mature splenic CD4 and CD8 T cells for pH2AX levels. Mature T cells from young mice did not show any detectable pH2AX (Fig. 5A), in contrast to SP thymocytes. While both CD4 and CD8 T cells from aged mice showed some increase in their pH2AX levels as compared to the corresponding cells from young mice, this difference was not statistically significant (p>0.5) (Fig. 5A and 5C). There were no differences in the forward scatter profiles for each splenic T cell population between aged and young mice (Fig. 5B), ruling out the likelihood of higher pH2AX levels per cell simply as a consequence of larger cell volumes. When examined by fluorescence microscopy as above, the numbers of pH2AX foci per cell were also marginally higher in both CD4 and CD8 T cells from aged mice than in those from young mice (Fig. 6). Since these differences were statistically significant (although the levels of pH2AX did not show any differences), it is possible that a low level of DSBs may be a persistent feature of the T cell lineage in aged mice.
Since the highest levels of DSBs were seen in DN thymocytes of aged mice and since the major block in T cell development in aged mice has been reported to be at the DN stage (Hsu et al., 2002; Min et al., 2005; Taub and Longo, 2005), we examined if DN thymocytes from aged mice showed any alteration in the levels of DNA repair proteins, using flow cytometry. We examined the levels of MRE11 and Ku70, DNA repair proteins involved in homologous repair (HR) (van Gent et al., 2001; Bassing and Alt, 2004) and in non-homologous end-joining (NHEJ) repair particularly in the lymphocyte lineages (Kuhne et al., 2003; Bassing and Alt, 2004). MRE11 levels were notably higher in DN thymocytes from aged mice (Fig. 7), while Ku70 levels, while somewhat higher, were not significantly different (p=0.75) (Fig. 8).
We further examined the levels of the cell cycle checkpoint protein p53 in DN thymocytes from aged mice in light of the increased presence of DSBs in them. Again, while DN cells from aged mice showed modestly higher levels of p53 than DN cells from young mice, the differences were not significant (p=0.11) (Fig. 9). Elevated p53 levels in response to DNA damage by gamma-irradiation in vitro were easily detected in cells from both young and aged mice (data not shown), indicating that the lack of enhancement of p53 levels in aged mouse DN thymocytes ex vivo was not an artifact or limitation of the sensitivity of the staining protocols used.
We have attempted to examine the well-defined failure of thymopoiesis in aged animals (Andrew and Aspinall, 2001; Aspinall and Andrew, 2001; Min et al., 2005; Taub and Longo, 2005) in the context of the reported decline in DNA damage repair capacity during aging (Wojda and Witt, 2003). For this, we have compared thymocytes from young and aged mice using a robust marker, pH2AX (Ting and Lee, 2004; Franco et al., 2006), for DNA double-stranded breaks (DSBs) to assess DNA damage, the cellular levels of MRE11 and Ku70 proteins as indicators of homologous and non-homologous repair respectively (Ting and Lee, 2004; Franco et al., 2006), and p53 (Ting and Lee, 2004; Franco et al., 2006) as a marker significant for a G1/S phase checkpoint. We find a notable enhancement of DSBs in aged thymocytes, while the enhancement of repair and checkpoint mechanisms is only modest.
Normal thymocyte development involves the controlled creation of DSBs at TCR-beta loci in the most immature DN subset, and at TCR-alpha loci in the succeeding DP subset (Sebzda et al., 1999). These two subsets showed higher levels of pH2AX than the succeeding SP stages, as shown earlier (Ting and Lee, 2004; Franco et al., 2006). Further, the levels of pH2AX were higher in aged thymocyte subsets, particularly in the DN stage. These data suggested that thymocytes from aged mice may have more DSBs present than those from young mice. Confocal microscopic data confirmed the presence of higher DSB levels in aged thymocytes, both in the number of cells having at least one DSB and in the numbers of breaks found in individual cells.
An examination of the distribution of the number of breaks per cell in various thymocyte sub-populations showed a broadly unimodal distribution, indicating that this is not likely to be a situation where a discrete sub-population of cells is behaving differently, but an effect on the entire population. Also, the distribution moves leftwards as the maturation progresses in both aged and young cells, indicating DSBs are most frequent in DN thymocytes in aged mice. Since explanations of thymic atrophy during aging revolve around the DN stage, it is possible that the difference in DSB frequency in the DN population may be an important factor in the thymic dysfunction in aged mice. The frequency of CD3-bright CD4− CD8− DN T cells, which are gamma/delta TCR-bearing cells, was not significantly different between the thymocytes from young (~20%) and aged (~22%) mice (data not shown), ruling out the possibility that the differences in pH2AX levels could be related to differences between DN cells of the alpha/beta lineage versus gamma/delta TCR-bearing cells.
Although splenic T cells from both young and aged mice showed relatively low levels of pH2AX, those from aged mice continued to show some increase in the number of DSBs. This increase in DSB levels may be the result of poor repair of normal V(D)J recombination breaks, and/or of more frequently occurring of DSBs at other sites. This is a major issue that our data have not resolved, and it is possible that both mechanisms may contribute to the aging phenotype. The number of DSB/pH2AX foci is highest in the DN thymocytes of aged mice, and steadily declines even in them with the lowest frequencies being found in the periphery. Persistent TCR-beta VDJ recombination-induced breaks in DN cells may interfere with the proliferation during beta-selection, although checkpoint mechanisms would need to be robust for this. Again, DSBs may well be generated during DNA replication associated with proliferation, especially if checkpoint mechanisms are inefficient. Persistent TCR-alpha VJ recombination-induced breaks, and/or breaks caused during beta-selection proliferation, may well persist in naïve T cells, and may be lost only during immune responses in the periphery. Further, DSBs may be re-induced during peripheral proliferation of aged T cells. Also, this tendency of DN thymocytes to acquire persistent DSBs may be a consequence of either an intrinsic dysfunction in the thymic progenitors populating the thymus in aged mice, and/or a consequence of the altered thymic microenvironment in aged animals. Clearly, the genesis of these persistent DSBs is likely to be complex and will require extensive analyses.
Our data indicated that high DSB levels are most prominently seen in the DN thymocyte subpopulation of aged mice. We therefore evaluated a few DNA repair and checkpoint protein levels in this thymocyte subpopulation. To investigate the major pathways of DSB repair, we looked at the expression of two proteins involved in the HR and NHEJ pathways. Endogenous DSBs created by V(D)J recombination are normally repaired by NHEJ and therefore it was relevant to estimate Ku70 levels in these DN cells. However, despite the persistence of DSBs, DN thymocytes from aged mice showed no significant increase in Ku70 levels. In human senescent cell lines, Ku70 is not only poorly expressed but shows poor translocation, indicating that senescent cells are unable to respond properly to DNA damage (Seluanov et al., 2007). While Ku70 levels are similar between DN thymocytes from young and aged mice, it is possible that the enhanced DSB levels in the latter cells may also reflect a decrease in the effectiveness of NHEJ due to poor translocation. NHEJ predominates in G1 and early S and a lack of NHEJ repair would result in an increase in cell cycle arrest and possible cell death.
Notably, MRE11 levels were substantially higher in DN thymocytes from aged than from young mice. MRE11 is involved in both HR and NHEJ pathways (van Gent et al., 2001; Kuhne et al., 2003; Bassing and Alt, 2004; Helmink et al., 2009). Together, these data may be indicative of a tendency of DN thymocytes in aged mice to be susceptible to random DNA breaks, rather than simply poor NHEJ repair of V(D)J breaks. However, once the MRN (MRE11, Nbs1,Rad50) complex, involved in the initial recognition and processing of DSBs (Uziel et al., 2003), is activated, ATM is activated and phosporylates a wide variety of downstream DNA damage response proteins and kinases (Uziel et al., 2003; Kurz and Lees-Miller, 2004). The downstream targets of ATM regulate a wide diversity of events from cell cycle checkpoints to induction of apoptosis. Therefore, the increase in MRE11 expression seen in the aged DN thymocytes could be an indication of increases in apoptosis, cell cycle arrest, as well as DNA repair. Further examination of other components and functionality of the DNA repair pathways is necessary to resolve this issue.
The p53 protein can regulate cell fate at the G1/S cell cycle checkpoint, permitting progression of cell cycle or triggering apoptotic cell death. DN thymocytes from aged mice showed no significant increase in the level of p53 expression over the DN cells from young mice. An increase in unresolved DSB levels would normally be expected to trigger p53 and lead to cell cycle arrest at the G1/S checkpoint. Thus, it is possible that, rather than the DN thymocytes from aged mice undergoing cell cycle arrest, they would continue through the cycle with unresolved DSBs and die at later points.
Regardless of the precise pathway and cell cycle stage of cell death, a decrease in the number of viable DN thymocytes could lead to the thymic involution, the decrease in TCR variability, and the overall decrease in naïve T cells exported from the thymus so characteristic of aged animals. Together, our data show that thymocytes from aged mice have more double-stranded DNA damage on the one hand, and little increase in either the cell cycle checkpoint protein p53 or the DNA repair protein Ku70 on the other. Another DNA repair pathway protein, MRE11, does show an increase. Thus, it is possible that there is a greater tendency to random DSB generation in aged DN thymocytes, and that despite the increased expression of MRE11, repair is not efficient. However, the failure of p53 level enhancement may lead to damaged cells progressing into the cell cycle before dying at later points. There are three potential outcomes when DSBs are not repaired (Sprent and Surh, 2002), permanent cell cycle arrest, cellular induction of apoptosis, and mitotic cell death caused by loss of genomic material (van Gent et al., 2001). Loss of thymocytes by any of these pathways in aged mice could thus contribute to thymic involution and reduction in naïve T cell frequencies observed in the aged immune system.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.