• Open Access

Accelerated hematopoietic stem cell aging in a mouse model of dyskeratosis congenita responds to antioxidant treatment


  • Bai-Wei Gu,

    1. Division of Hematology, Department of Internal Medicine, Washington University School of Medicine, St Louis, MO, USA
    2. Division of Hematology, Department of Pediatrics, Abramson Research Center, The Children’s Hospital of Philadelphia, Philadelphia, PA 19104, USA
    Search for more papers by this author
  • Jian-Meng Fan,

    1. Division of Hematology, Department of Internal Medicine, Washington University School of Medicine, St Louis, MO, USA
    2. Division of Hematology, Department of Pediatrics, Abramson Research Center, The Children’s Hospital of Philadelphia, Philadelphia, PA 19104, USA
    Search for more papers by this author
  • Monica Bessler,

    1. Division of Hematology, Department of Internal Medicine, Washington University School of Medicine, St Louis, MO, USA
    2. Division of Hematology, Department of Pediatrics, Abramson Research Center, The Children’s Hospital of Philadelphia, Philadelphia, PA 19104, USA
    Search for more papers by this author
  • Philip J. Mason

    1. Division of Hematology, Department of Internal Medicine, Washington University School of Medicine, St Louis, MO, USA
    2. Division of Hematology, Department of Pediatrics, Abramson Research Center, The Children’s Hospital of Philadelphia, Philadelphia, PA 19104, USA
    Search for more papers by this author

Dr Philip J. Mason, Department of Pediatrics, Abramson Research Center, The Children’s Hospital of Philadelphia, 3615 Civic Center Boulevard, Philadelphia, PA 19104-4318, USA. Tel.: +267 426 9327; fax: +267 426 9892; e-mail: masonp@email.chop.edu


Mutations in DKC1, encoding telomerase associated protein dyskerin, cause X-linked dyskeratosis congenita (DC), a bone marrow (BM) failure, and cancer susceptibility syndrome. Decreased accumulation of telomerase RNA resulting in excessive telomere shortening and premature cellular senescence is thought to be the primary cause of disease in X-linked DC. Affected tissues are those that require constant renewal by stem cell activity. We previously showed that in Dkc1Δ15 mice, which contain a mutation that is a copy of a human mutation causing DC, mutant cells have a telomerase-dependent proliferative defect and increased accumulation of DNA damage in the first generation before the telomeres are short. We now demonstrate the presence of the growth defect in Dkc1Δ15 mouse embryonic fibroblasts in vitro and show that accumulation of DNA damage and levels of reactive oxygen species increase with increasing population doublings. Treatment with the antioxidant, N-acetyl cysteine (NAC), partially rescued the growth disadvantage of mutant cells in vitro and in vivo. Competitive BM repopulation experiments showed that the Dkc1Δ15 mutation is associated with a functional stem cell defect that becomes more severe with increasing age, consistent with accelerated senescence, a hallmark of DC hematopoiesis. This stem cell phenotype was partially corrected by NAC treatment. These results suggest that a pathogenic Dkc1 mutation accelerates stem cell aging, that increased oxidative stress might play a role in the pathogenesis of X-linked DC, and that some manifestations of DC may be prevented or delayed by antioxidant treatment.


Dyskeratosis congenita (DC) is a disorder of telomere maintenance and offers a unique opportunity to study the consequences of telomere dysfunction (Bessler et al., 2008). The disease is heterogeneous in presentation, and mutations in six different genes, all encoding components of telomerase (Vulliamy & Dokal, 2008) or shelterin, (de Lange, 2005; Savage et al., 2008) the protein complex that caps and protects telomeres, have been implicated. Experiments using telomerase null mice showed that excessively short telomeres activate a DNA damage response causing cellular senescence and cell death leading to a mouse phenotype that mimics many aspects of the clinical disease seen in patients with DC (Blasco et al., 1997; Marciniak & Guarente, 2001). However, in these mice, because of their long telomeres, a disease phenotype only becomes evident after several generations of inbreeding causing the telomeres to become very short (Blasco et al., 1997). It has therefore been postulated that excessively short telomeres are the primary cause of disease in DC (Hao et al., 2005; Mason et al., 2005; Kirwan & Dokal, 2009). In fact, in patients with DC, the tissues primarily affected are those that require constant renewal through stem cell activity, such as skin, bone marrow (BM), lung, and gut endothelium, and at the time of disease manifestation, all patients with DC have very short telomeres (Mitchell et al., 1999; Vulliamy et al., 2001; Du et al., 2009).

We have previously reported a genetic DC mouse model, in which mice (Gu et al., 2008) carry a Dkc1 exon 15 gene deletion, encoding a truncated dyskerin protein with a C-terminal deletion of 21 amino acids, previously found in a DC family (Vulliamy et al., 1999). We demonstrated that in these mice, the Dkc1Δ15 cells had a growth disadvantage compared to wild-type (WT) cells, which in heterozygous Dkc1Δ15/+ female mice, because of the localization of Dkc1 on the X-chromosome, led to progressive disparity favoring wild type over mutant cells. This was because of the growth advantage of cells expressing the wild-type Dkc1 allele over those expressing the mutant allele after random X-chromosome inactivation in early embryogenesis (Lyon, 1961), a phenotype universally seen in women who are DKC1 mutation carriers (Vulliamy et al., 1997; Gu et al., 2008). We demonstrated that the growth disadvantage of Dkc1Δ15 mutant cells in part is mediated by the p53 pathway, dependent on telomerase activity, but does not require telomere shortening (Gu et al., 2008), and that DkcΔ15 mutant cells accumulate increased levels of DNA damage. These findings challenged the current idea that DKC1 mutations solely cause disease through destabilizing the telomerase RNA (TERC) (Mitchell et al., 1999) which quantitatively reduces telomerase enzymatic activity and leads to the critically short telomeres that ultimately are responsible for disease.

We now demonstrate that the reduced growth rate of Dkc1Δ15 cells correlates with the accumulation of DNA damage foci that is dependent on and increases with proliferation in vitro and that in vivo, the Dkc1Δ15 mutation leads to an age-dependent decrease in hematopoietic stem cell (HSC) function consistent with the accelerated aging of HSC characteristic of DC. Furthermore, we find that Dkc1Δ15 mutant cells are hypersensitive to oxygen and show that the decreased growth rate is associated with an increased accumulation of reactive oxygen species (ROS). Finally, we demonstrate that the growth disadvantage of Dkc1Δ15 cells can be overcome in part by treatment with the antioxidant N-acetyl cysteine (NAC) not only in vitro but more importantly also in vivo. Thereby, our investigations in Dkc1Δ15 mice identify oxidative stress as a potential new player in the pathogenesis of DC and thus uncover a new drugable target that may be utilized to prevent or delay disease in patients with DC.


Dkc1Δ15 MEF cells have a growth disadvantage that is associated with increasing ROS levels and a proliferation-dependent enhanced accumulation of DNA damage

In heterozygous Dkc1Δ15/+ female mice, Dkc1Δ15 mutant cells have a growth disadvantage, when compared to normal cells, which is in part mediated by the p53 pathway, dependent on telomerase activity but apparently independent of telomere length (Gu et al., 2008). To investigate the growth phenotype conferred by the Dkc1Δ15 mutation in cell culture, we compared the growth rate of male Dkc1Δ15 (Δ15) and Dkc1+ (WT) mouse embryonic fibroblasts (MEFs) (Fig. 1). The growth rate of Δ15 MEF cells was lower when cultured at both ambient oxygen (21%) and low (3%) oxygen. In 21% oxygen, both types of cells stopped growing and entered senescence after 8–10 population doublings (PDs), with the Δ15 cells growing more slowly than the WT cells. In 3% oxygen, Δ15 cells grew more slowly and entered senescence earlier than WT cells (Fig. 1A,B).

Figure 1.

 Impaired proliferation and increased DNA damage in Dkc1Δ15 mouse embryonic fibroblasts (MEFs). (A) Cumulative population doubling (PD) in wild-type (WT) and Dkc1Δ15 (Δ15) MEFs growing in 21% or 3% O2. The results shown represent a single experiment but similar results were obtained in three independent experiments. (B) Senescence-associated β-galactosidase (SA-β-Gal) staining of WT and Dkc1Δ15 (Δ15) MEFs after different PDs. The blue staining indicates the senescent cells (Dimri et al., 1995). (C) Quantitation of γ-H2AX foci between WT and Dkc1Δ15 (Δ15) MEF cells after different PDs in different O2 concentrations from Fig. S2 (Supporting Information). More than 200 cells were analyzed. (D) Telomere length of male WT and Dkc1Δ15 (Δ15) MEF cells after different PDs.

Recent studies implicate ROS in mediating cell senescence and genomic instability (Colavitti & Finkel, 2005; Rassool et al., 2007). We thus measured the accumulation of ROS using the oxidation sensitive dye 5-(and -6)-carboxy-2′,7′-dichlorodihydrofluorescein diacetate (CM-H2DCFDA). We found that there was a significantly higher level of ROS accumulation in Δ15 cells and WT cells when cultured in high oxygen compared to low oxygen, and in both types of cells, ROS levels increased with increasing PDs (Fig. 2). Interestingly, Δ15 cells accumulated more ROS than WT cells and the difference became very significant with more PDs, even when cells were cultured in low oxygen.

Figure 2.

 Increased intracellular reactive oxygen species (ROS) level in Dkc1Δ15 mouse embryonic fibroblast (MEF) cells. (A) Wild-type (WT) and Dkc1Δ15 (Δ15) MEF cells grown in 3% O2 were treated with the oxidation sensitive dye, CM-H2DCFDA, and the conversion to its oxidized fluorescent derivative assessed by flow cytometry. (B) WT and Dkc1Δ15 (Δ15) MEF cells were grown in 3% or 21% O2 for eight population doublings and ROS measured as in A above. (C). Quantitation of mean CM-H2DCFDA fluorescence of A. Asterisk indicates significant difference (Student’s t-test) P < 0.01, n = 6. (D) Quantitation of mean CM-H2DCFDA fluorescence of B.

To further investigate the basis of impaired proliferation, we examined the appearance of phosphorylated histone H2AX (γ-H2AX) foci, indicators of DNA damage (Rogakou et al., 1998), in Δ15 and WT MEFs. There was a significantly higher number of foci in Δ15 cells, and the difference was greater in high oxygen (Fig. 1C and Fig. S1). γ-H2AX does not accumulate when cells are maintained in a quiescent state in a confluent culture, suggesting DNA replication is required (Fig. S2). Because no significant differences in overall telomere lengths were observed between Δ15 and WT MEFs after three and 18 PDs (Fig. 1D), these data suggest the structure of the telomeres in the mutant cells, rather than their length, leads to the DNA damage accumulation and growth impairment. An alternative explanation is that increased ROS levels may result from ribosomal stress caused by the presence of mutant dyskerin in H/ACA snoRNAs and disrupted ribosome biogenesis. These findings show that the Dkc1Δ15 mutation causes slower cell growth associated with increased DNA damage and increased levels of ROS.

Increased DNA damage in aging Dkc1Δ15 mice

The presence of DNA damage accumulation in untreated Dkc1Δ15 cells, and particularly its increase with increasing PDs in culture, led us to ask whether a significant amount of DNA damage accumulated in aging Dkc1Δ15 mice. We found that indeed, while protein extracted from the spleen of a 4-month-old mouse contained almost undetectable levels of γ-H2AX, the levels were dramatically increased in spleens from older mice (Fig. 3). In Dkc1Δ15 mice, levels of γ-H2AX were consistently higher than in WT mice in either young or old mice and in liver, spleen, and BM. An increase in the steady state levels of p53 was also detected in these tissues in Dkc1Δ15 mice, suggesting chronically increased activity of the p53 DNA damage response. In agreement with these results, accumulation of DNA damage in aging HSC (Rossi et al., 2007) and formation of γ-H2AX foci in a number of cell types in older mice (Wang et al., 2009) have been reported recently.

Figure 3.

 DNA damage in Dkc1Δ15 (Δ15) and wild-type (WT) mice. (A) A Western blot of extracted protein from mouse tissues probed with γ-H2AX, p53, and β-actin antibodies. (B) A Western blot of extracted protein from the spleens of mice of different ages probed with γ-H2AX and β-actin antibodies.

The Dkc1Δ15 mutation causes decay of stem cell function with age

In patients with X-linked DC, the clinical manifestations, including progressive BM failure, are generally not present at birth but become apparent during childhood and adolescence and affect mainly tissues with a high cell turnover. Male Dkc1Δ15 mice have no phenotype either during the first 3 months of life, but interestingly show a decrease in the proportion of B and T lymphocytes (Fig. 6C,D) with age along with a reduction in body weight (not shown). We were therefore interested in further investigating the effect of the Dkc1Δ15 mutation on hematopoiesis during aging. The number of BM cells, erythroid-committed progenitor cells, and c-Kit+, Sca-1+ and Lineage (KLS) cells (Rossi et al., 2005), a BM cell fraction enriched in functional stem cells, was not significantly different between Dkc1Δ15 and Dkc1+ mice or when comparing young and old mice (Fig. S3), showing the absence of overt BM cell depletion in these mice. To study the effect of the mutation on stem cell function during aging, we carried out competitive repopulation and serial BM transplantation studies using the CD45.1/CD45.2 congenic system (Spangrude et al., 1988) to identify the origin of BM cells (Fig. S4). In these experiments, all donors were male and in the competitive experiments the Dkc1Δ15 and Dkc1+ donors were the same age, either young (10 weeks) or old (77–88 weeks). First, we determined whether the growth disadvantage is intrinsic to BM cells and whether the growth disadvantage is similar for BM cells from young and old Dkc1Δ15 mice. For this, we performed competitive repopulation experiments with a 1:1 mixture of Dkc1Δ15 and WT BM cells derived from either young or old mice (Fig. 4). In these experiments, Dkc1Δ15 BM cells from old mice were more compromised in the competition with WT cells in primary recipients, making up only 20% of peripheral blood cells 12 weeks after transplant compared with 40% of Dkc1Δ15 peripheral blood cells in mice receiving Dkc1Δ15 and WT BM cells from young mice (Fig. 4). Twelve weeks after a second round of BM transplantation, 10–30% peripheral blood cells in the secondary recipients were derived from Dkc1Δ15 BM cells when the original donors were young. In contrast, no Dkc1Δ15 cells were detectable in recipients of BM cells from old mice (Fig. 5A). These results indicated that the growth defect is intrinsic to BM cells and is accentuated in BM cells from old vs. young Dkc1Δ15 mice.

Figure 6.

N-acetyl-l-cysteine (NAC) can partially rescue the growth disadvantage of Dkc1Δ15 (Δ15) cells at the stem cell level. (A) Western blot results of dyskerin expression level in mouse embryonic fibroblasts from heterozygous Dkc1Δ15/+ female mice after different population doublings (PDs) with or without inclusion of 100 μm NAC in the medium. The filled arrow indicates the position of wild-type (WT) dyskerin, while the open arrow indicates the Δ15 dyskerin. (B) Western blot results of dyskerin expression level in the spleen of heterozygous Dkc1Δ15/+ female mice at different ages with or without inclusion of 1 mg/mL NAC in the drinking water after weaning. Three to five mice were examined at each time point with similar results but the results from only one mouse are shown in the figure. (C) The percentage of B lymphocytes in peripheral blood of WT and Dkc1Δ15 male mice with or without NAC treatment, respectively. (D) The percentage of T lymphocytes in peripheral blood of WT and Dkc1Δ15 male mice with or without NAC treatment, respectively. (E) In the serial transplantation experiments (50/50), Dkc1Δ15 bone marrow cells (CD45.2), taken from old (52–55 weeks) Dkc1Δ15 male mice with or without NAC treatment, respectively, were mixed with CD45.1 WT cells of the same age and transplanted into a CD45.1 WT recipient. The proportion of original donor cells (Dkc1Δ15) of different lineages is plotted at 12 weeks after first and second transplantation. N = 3–5. Note that no cells from the Dkc1Δ15 mice without NAC treatment are detected in the secondary recipients.

Figure 4.

 Effect of aging on the competitive repopulation ability of Dkc1Δ15 bone marrow (BM). The percentage of total cell counts derived from Dkc1Δ15 (Δ15) mice is plotted for B lymphocytes, T lymphocytes, CD4+ lymphocytes, CD8+ lymphocytes, and granulocytes from peripheral blood of irradiated recipients 12 weeks after BM transplantation with 5 × 106Dkc1Δ15 BM cells (CD45.2) alone (100/0) or mixed with CD45.1 allotype-marked wild-type (WT) cells in 1:1 ratio (50/50). Donor ages are 10 weeks (A) or 77–88 weeks (B). In the serial transplantation experiments (100/0), Dkc1Δ15 BM cells (CD45.2) or age-matched WT cells (CD45.2) were transplanted into a CD45.1 WT recipient. In the 50/50 mixing experiments, the CD45.2 Dkc1Δ15 BM cells were mixed with CD45.1 WT cells of the same age and transplanted into a CD45.1 WT recipient. N = 3–5.

Figure 5.

 Effect of aging on the competitive repopulation ability of Dkc1Δ15 bone marrow (BM) in secondary recipient mice. (A) Secondary transplant of cells from the mice primarily transplanted with a 1:1 mixture of BM cells from Dkc1Δ15 (Δ15) and Dkc1+ [wild-type (WT)] mice. The panel on the left shows the results for the young mice and the panel on the right shows the results from the old mice, where no Δ15 cells are detected in the secondary transplant. In this experiment, the CD45.2 Dkc1Δ15 BM cells were mixed with CD45.1 WT cells of the same age and transplanted into a CD45.1 WT recipient. (B) Serial transplant of cells from the mice primarily transplanted with CD45.2 BM cells of the genotypes and ages shown (100%). The proportion of original donor cells (CD45.2) of different lineages is plotted at 12 weeks after first, second, and third transplantation. In this experiment, Dkc1Δ15 BM cells (CD45.2) or age-matched WT cells (CD45.2) were transplanted into a CD45.1 WT recipient. There were three mice in each group.

Next, we were interested in whether this age-dependent growth defect occurs during hematopoietic differentiation or occurs at the stem cell level resulting in successive depletion of functional HSCs. For this, serial BM transplant experiments were performed using BM cells from either young or old Dkc1Δ15 mice (Fig. 5). In the primary transplant experiments, both young and old cells were about equally capable of populating the recipient’s BM as suggested by the about equal proportion of Dkc1Δ15 circulating blood cells (Fig. 4). In secondary transplant recipients, young Dkc1Δ15 BM cells fully restored hematopoiesis reaching levels of 90% after 16 weeks, comparable to the results obtained with secondary transplants of WT BM cells from young and old animals. In contrast, secondary transplants of BM cells from old Dkc1Δ15 mice contributed only little to peripheral blood and their numbers decreased with time in the secondary recipients, consistent with an increasing failure of stem cell function in aging Dkc1Δ15 BM cells. Finally, BM cells from secondary recipients were transplanted a third time. In the tertiary recipients, less than 10% of peripheral blood cells were derived from the originally transplanted Dkc1Δ15 BM cells obtained from old mice, whereas in recipients receiving BM cells obtained from young Dkc1Δ15 mice, 60% were derived from the originally transplanted BM cells. In recipient mice receiving either young or old WT BM cells, about 80% of cells in tertiary recipients were derived from the original donors (Fig. 5B).

The results from competitive and serial BM transplant experiments indicate that Dkc1Δ15 BM cells have an intrinsic proliferative defect that worsens with age and are associated with an accelerated decline in stem cell function when compared with that to BM from WT mice.

Growth of Dkc1Δ15 cells is improved by antioxidant treatment in vitro and in vivo

Two findings implicated ROS in the growth defect we observed in Dkc1Δ15 cells and mice. First, Δ15 MEF cells were more sensitive than WT MEF cells to culture in 21% oxygen (Fig. 1A,C). Second, Δ15 MEF cells accumulated more ROS during normal cell growth, in 21% or 3% oxygen than did WT MEFs (Fig. 2). In addition, regulation of oxidative stress has been shown to be an important factor in HSC aging (Ito et al., 2004, 2006; Diehn et al., 2009; Abbas et al., 2010; Li & Marban, 2010). We therefore asked whether oxidative stress plays a role in the growth disadvantage of Dkc1Δ15 cells compared with WT cells by testing whether treatment with an antioxidant could rescue the growth disadvantage of Dkc1Δ15 cells. First, we tested primary MEF cells from Dkc1Δ15/+ female mice, in the presence or absence of 100 μm NAC, a clinically approved antioxidant (Fig. 6A) (Atkuri et al., 2007). MEF cells from Dkc1Δ15/+ female mice in early passages consist of 50% expressing WT and 50% expressing truncated Δ15 dyskerin, reflecting random X-chromosome inactivation. While without NAC the WT cells have almost completely outgrown the Dkc1Δ15 cells after 11 PDs, in the presence of NAC the Dkc1Δ15 cells are still clearly present after 15 PDs, suggesting that NAC at least partially rescues the growth disadvantage of dyskerin mutant cells. We next tested whether the NAC treatment would rescue the growth disadvantage of Dkc1Δ15 cells in vivo and administered NAC to Dkc1Δ15 /+ female mice from the age of 3 weeks by adding 1 mg/mL NAC into their drinking water. Analysis of the spleens from these mice compared with untreated Dkc1Δ15/+ mice showed that the Dkc1Δ15 cells persist in the NAC-treated mice for much longer than in the untreated animals, being still detectable at 24 weeks of age, while in untreated mice, the mutant protein and therefore mutant cells are barely detectable after 12 weeks (Fig. 6B), indicating that NAC treatment partially rescues the growth disadvantage of Dkc1Δ15 cells.

Decayed stem cell function rescued by antioxidant treatment

To determine whether NAC treatment could be useful therapeutically, it is important to investigate its effect on stem cell function, which is defective in DC. To address this issue, we established a cohort of mice that were given NAC in their drinking water (1 mg/mL) from 3 weeks of age and maintained on NAC for 1 year. We found that long-term NAC treatment did not show significant side effects on the mice. They had slightly increased neutrophils, but no difference in mortality and body weight compared with the untreated group (data not shown). Impressively, old (52–55 weeks) male Dkc1Δ15 mice from the NAC cohort did not show the decreased B- and T-cell proportions in peripheral blood observed in untreated Dkc1Δ15 mice (Fig. 6C,D). Competitive BM transplantation experiments were carried out in which a 1:1 mixture of BM cells from mutant and WT mice was used to repopulate lethally irradiated recipient mice. These experiments showed that, when taken from NAC treated animals, old Dkc1Δ15 BM cells could compete with age-matched WT cells with 40–45% of Dkc1Δ15 cells in primary recipients compared with only 20% for the untreated group. Moreover, after secondary transplantation, cells from the NAC treated group still represent 15–20% of Dkc1Δ15 cells in recipients while those from the untreated group could not be detected (Fig. 6E).


The data presented here show that Dkc1Δ15 mice show an accelerated aging phenotype of the HSC similar to that presumed to occur in DC in humans. In the human disease, it is generally believed that stem cell dysfunction is caused by accelerated telomere erosion, leading to critically short telomeres and cell senescence (Kirwan & Dokal, 2009). This in turn would lead to increased recruitment of stem cells and eventual exhaustion of the stem cell pool (Mason et al., 2005). The results presented here, where an accelerated rate of stem cell aging is evident in the first generation of Dkc1Δ15 mice, in the presence of long telomeres, suggest that in X-linked DC stem cell function worsens throughout life and that accumulation of DNA damage may contribute to the development of stem cell depletion.

Lesions in DNA repair genes have been found to affect HSC aging in a similar way to the Dkc1 mutation studied here (Nijnik et al., 2007; Rossi et al., 2007). Interestingly, these studies included later generation mice lacking telomerase RNA (Rossi et al., 2007). DNA damage triggered by the excessively short telomeres was thought to be the factor that caused loss of HSC function with aging. In our studies, the same effect is seen, but in contrast to the previous study, this was seen in the first generation of Dkc1Δ15 mice before deficiency in telomerase activity is expected to significantly shorten the telomeres (Blasco et al., 1997). Indeed, in-gel hybridization of DNA from MEFs did not reveal significant telomere shortening (Fig. 1). These findings infer that dyskerin may have an important role at the telomeric site of telomerase action, in agreement with the presence of dyskerin in purified active telomerase (Cohen et al., 2007). This suggests that mutant dyskerin acts in a different way to the TERT and TERC mutations, which generally act through haploinsufficiency (Vulliamy et al., 2001). The data are consistent with a model of X-linked DC whereby telomerase, with a mutated dyskerin, gains access to the telomeric DNA and extends the telomeres with reduced efficiency, rendering the telomere more susceptible than normal to DNA damage or failing to protect it from being recognized as damaged DNA. Interestingly, modulation of DNA damage was not seen in Tert−/− cells (Erdmann & Harrington, 2009) suggesting that with a reduced supply of TERT and possibly TERC, fewer molecules of otherwise normal telomerase will be present at the telomeres leading to the extension of fewer but intact telomeres without the induction of a DNA damage response. Our finding that the DNA damage response is dependent on replication is compatible with this model.

Alternatively, we cannot exclude a telomere independent role for the mutant dyskerin. Ribosomal stress caused by perturbation of ribosome biogenesis may lead to increased ROS levels which then cause DNA damage. The DNA damage, to which telomeres are particularly susceptible, may then limit stem cell function. In this model, the telomerase-dependent growth inhibition we previously observed might arise if telomeres were hypersensitive to ROS-induced damage when being extended by active telomerase. That the effect of dyskerin mutations on ribosome biogenesis may contribute to the DC phenotype is supported by the observation that another congenital BM failure syndrome, Diamond Blackfan Anemia, is caused by mutations in ribosomal proteins (Lipton & Ellis, 2010).

Our Dkc1Δ15 mouse model is distinct from a previously reported hypomorphic Dkc1 mouse model (Ruggero et al., 2003) in that our mice genetically reproduce a pathogenic DKC1 mutation previously identified in a family with X-linked DC, whereas in the previous model, a decreased level of Dkc1 mRNA levels was achieved through transcriptional interference after integration of the targeting vector downstream of the Dkc1 gene. Disease-causing DKC1 mutations cluster in the N-terminal region and the RNA-binding domain of dyskerin. Models of the three-dimensional structure have shown that these regions closely associate (Rashid et al., 2006) suggesting that the mutations have a specific effect on dyskerin which is likely to be reproduced in our Dkc1Δ15 model, but not reproduced in a model expressing low levels of otherwise normal dyskerin.

Interestingly, we identified an increased replication-dependent accumulation of ROS in Dkc1Δ15 mutant cells. ROS are being increasingly implicated in cellular aging and as mediators of cell senescence and apoptosis as well as in various normal physiological processes and signaling pathways (Bertram & Hass, 2008). In particular, control of ROS in HSC is essential to maintain their function (Gazit et al., 2008). An increased accumulation of ROS has been shown to cause a decreased hematopoietic stem cell function in mice deficient in ATM (Ito et al., 2004), a cell cycle check point regulator that is activated by DNA damage, and in mice lacking FoxO transcription factors, which are involved in the transcriptional control of genes encoding ROS-scavenging enzymes (Tothova et al., 2007). Stem cell defects in ATM-deficient mice or in mice lacking transcription factors FoxOs 1, 3, and 5 are partially corrected by antioxidant NAC treatment (Ito et al., 2004; Tothova et al., 2007). We therefore tested the effect of NAC treatment of our Dkc1Δ15 cells in vitro and in vivo and found that indeed, NAC treatment prevented the accumulation of ROS and restored the proliferation defect of Dkc1Δ15 MEF cells in vitro. Impressively, the treatment of heterozygous Dkc1Δ15/+ female mice by adding NAC to the drinking water, significantly improved the survival of cells expressing the truncated Δ15 protein, indicating that NAC improved the proliferation defect of mutant cells. More interestingly, after treatment with NAC for over 1 year, Dkc1Δ15 mice showed significant improvement in BM stem cell repopulation activity as well as a restoration of the wild-type proportion of lymphocytes in peripheral blood. These results suggest that pathogenic Dkc1 mutations may operate at least in part through the accumulation of ROS levels identifying oxidative stress as a novel factor in the pathogenesis of DC.

Telomeres are thought to be particularly vulnerable to oxidative DNA damage and oxidative stress causes telomere shortening (Passos et al., 2007; Richter & von Zglinicki, 2007). The exact mechanism whereby oxidative stress affects the growth of Dkc1Δ15 mutant cells and the aging of HSC in Dkc1Δ15 mice is not entirely clear because ROS can act as agents of DNA damage as well as being part of the response to DNA damage (Macip et al., 2002; Rai et al., 2009; Abbas et al., 2010; Li & Marban, 2010). Whether the increase in ROS that we observe accumulating in mutant cells is a consequence of the DNA damage response mediated by ATM/p53/p21 (Macip et al., 2002) or because of an unknown pathway linking mutant dyskerin, through the telomerase or ribosome biogenesis pathways, to the regulation of ROS accumulation is unknown (Passos et al., 2007) (Perez-Rivero et al., 2008) and the subject of current investigations.

Interestingly, NAC has been shown to improve pulmonary function in patients with pulmonary fibrosis (Sharma et al., 2003) and for this purpose is also empirically used in patients with DC who develop pulmonary fibrosis. Telomere shortening has been thought to be the major contributor to pulmonary fibrosis in these patients (Alder et al., 2008) but perhaps increased sensitivity to oxidative stress plays an additional role in pathogenesis. The therapeutic effect of antioxidants or NAC treatment on HSC function and whether these agents may ameliorate, delay, or prevent the occurrence of BM failure or pulmonary fibrosis in these patients if given early in the course of the disease remains to be determined.

Experimental procedures


The generation of Dkc1Δ15 mice was as previously described (Gu et al., 2008). These mice were backcrossed with wild-type C57BL/6 mice (CD45.2) at least seven times. C57BL/6 mice (CD45.1) were from the Jackson Laboratory (Bar Harbor, ME, USA).

Measurement of telomere length

Mouse embryonic fibroblasts cells were embedded in agarose plugs by using CHEF agarose plug kit according to the manufacturer’s instructions (Bio-Rad, Hercules, CA, USA). DNA embedded in the plug was extracted, digested with MboI, and electrophoresed through a 1% agarose gel for 20 h at 6 V/cm, 1–6 s switch time using CHEF DR-III pulse-field system (Bio-Rad). 32P-γ-ATP labeled (CCCTAA)4 probe was used in the in-gel hybridization procedure (Dionne & Wellinger, 1996).

Western blot analysis

Total protein from cells and mouse tissues was prepared by using RIPA lysis buffer (1× TBS, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 0.004% sodium azide, and 1× protease inhibitor cocktail). Protein concentration was measured by using the Bio-Rad protein assay (Bio-Rad).


Immunofluorescence was performed with a standard paraformaldehyde technique (slides were fixed in PBS-buffered 4% paraformaldehyde for 10 min, permeabilized with 0.5% Triton-PBS for 15 min, and blocked with 30% normal goat serum for 1 h). Primary antibody was used at 1/500 in 1.5% normal goat serum for 2 h. After washing with PBS, cells were incubated with a secondary goat anti-rabbit IgG conjugated with FITC and/or Alexa 568 at 1/1000 in 1.5% normal goat serum for 45 min. All blocking and incubation steps were carried out at room temperature. Finally, slides were counterstained with 4,6-diamidino-2-phenylindole (DAPI) and covered by mounting media. The cells were examined at 1000× magnification using a fluorescence microscope (Nikon, Melville, NY, USA). FITC, Alexa 568, and DAPI images were overlapped by using ISIS FISH imaging software (Metasystems, Waltham, MA, USA).

Measurement of intracellular ROS

After being trypsinized and washed with PBS, 5 × 105 MEF cells were resuspended in prewarmed PBS loaded with 10 μm 5,6-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (CM-H2DCFDA; Invitrogen, Carlsbad, CA, USA) in the dark for 30 min at 37 °C, 5% CO2. After one time washing with prewarmed PBS, the oxidative conversion of CM-H2DCFDA to its fluorescent product was measured immediately by Flowcytometry using 488 nm FL1 channel.


The sources of antibodies were as follows: anti-γ-H2AX-S139 (Abcam, Cambridge, UK; ab2893), anti-p53 (Abcam; ab26), and anti-dyskerin were as previously described (Mochizuki et al., 2004), and anti-β-Actin was used as total protein loading control (Abcam; ab20272).

Establishment of primary mouse embryonic fibroblasts (MEFs)

Dkc1Δ15, control WT male, and Dkc1Δ15/+ female MEF cells were prepared from a cross of Dkc1Δ15/+ females with WT male mice. Primary MEF cells were isolated from 13.5-day mouse embryos and harvested for analysis after 2–3 passages. Cells were cultured in DMEM supplemented with 10% FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin and maintained at 37 °C in a humidified atmosphere of 3% O2, 10% CO2.

MEF cells lifespan analysis

Population doublings were counted by subculturing a 90% confluent culture 1:4 for two PDs.

Senescence-associated β-galactosidase (SA-β-gal) staining

MEF cells were plated in a six-well plate, and 24 h later, the cells were washed twice in PBS for 5 min. The cells were fixed for 10 min in 4% formaldehyde and 0.2% glutaraldehyde in PBS and washed three times in PBS for 5 min each. The staining reaction was performed with 2 mL staining solution (1 mg/mL 5-bromo-4-chloro-3-indolyl β-D-galactoside (X-Gal), 40 mm citric acid/sodium phosphate, pH 6.0, 5 mm potassium ferrocyanide, 5 mm potassium ferricyanide, 150 mm NaCl and 2 mm MgCl2) at 37 °C for overnight in the dark. Cells were washed twice with PBS and photographed.

Flow cytometry

Bone marrow and blood leukocytes were processed for flow cytometry analysis as previously described (Keller et al., 2001). The following anti-mouse monoclonal antibody were used: FITC antibodies against TCRβ(H57-597), Gr1(RB6-8C5), CD45.2(104), and CD4(GK1.5); PE antibodies against B220(RA3-6B2),CD11b(M1/70), CD8(53-6.7), and CD45.1(A20). Unless otherwise indicated, all antibodies were obtained from BD Pharmingen (San Diego, CA, USA). The dates were collected with a FACScan.

To analyze KLS cells, lineage+ cells were first depleted by incubating with PE-Cy7-conjugated lineage markers: iL7, CD3, CD4, CD8, B220, CD19, Gr1, and ter-119. Following incubation with anti-PE-Cy7 colloid, lineage+ cells were depleted using the AutoMacs System (Miltenyi Biotec, Auburn, CA, USA). The lineage depleted cells were then incubated with APC anti-Sca-1(D7), APC-AlexaFluor-750 anti-c-Kit (2B8), Biotin-CD34 (RAM34). All of these antibodies were purchased from e-Bioscience (San Diego, NJ, USA). Finally, cells were sorted using a MoFlo high-speed flow cytometer (Dako Cytomation, Fort Collins, CO, USA).

Competitive bone marrow repopulation assay

CD45.1 C57/BL6 recipient mice were γ-irradiated with two 5-Gy doses 4 h apart and injected intravenously with 5 × 106 CD45.2-Dkc1Δ15 BM cells, either separately or mixed in a 1:1 ratio with CD45.1-WT cells. As a control, age-matched CD45.2-WT BM cells were injected separately. The recipients were kept on trimethoprim–sulfamethoxazole antibiotic drinking water during the first 2 weeks following transplantation. Mice were analyzed at 8 and 12 weeks after transplantation.

Serial bone marrow transplantation analysis

We collected 5 million CD45.2-WT or CD45.2-Dkc1Δ15 BM cells and transplanted them into lethally irradiated mice (CD45.1) as described for the competitive transplantation assay. Sixteen weeks after the first or second transplantation, we collected 5 million BM cells from transplanted mice and injected them into the secondary or tertiary recipient.

N-acetyl-l-cysteine (NAC) treatment

N-acetyl-l-cysteine (Sigma, St Louis, MO, USA) was dissolved in PBS buffer and was added to culture media to 100 μm for culture of MEF cells. We added NAC to mouse drinking water at 1 mg/mL concentration.


We thank William Eades, Jon Christopher Holley, and Jacqueline Hughes in the Siteman Cancer Center High Speed Sorter Core Facility for performing cell sorting segments of our experiments. The Siteman Cancer Center is supported in part by NCI Cancer Center Support Grant # P30 CA91842. We thank the NIH/NCI for financial support through grants to PJM (R01CA106995) and MB (R01CA105312). We also thank the America Society of Hematology (ASH) for financial support through a grant to B-WG (ASH Scholar Award).

Author contributions

B-WG, MB, and PM conceived the study. B-WG and J-MF did the experimental work. B-WG, MB, and PM analyzed and interpreted the data and wrote the manuscript.