Telomerase is required to slow telomere shortening and extend replicative lifespan of HSCs during serial transplantation

Richard C. Allsopp, Gregg B. Morin, Ronald DePinho, Calvin B. Harley, Irving L. Weissman


Telomere shortening ultimately limits the replicative life span of cultured human somatic cells. Telomeres also shorten during replicative aging in vivo in hematopoietic cells, including early hematopoietic progenitors and hematopoietic stem cells (HSCs), from humans and mice, despite readily detectable levels of telomerase in these cells. To assess the relevance of telomerase to the long-term replicative capacity of HSCs in vivo, we serially transplanted HSCs from wild-type and telomerase-deficient mice until exhaustion and monitored telomere length in HSCs during this process. Telomerase-deficient HSCs could be serially transplanted for only 2 rounds, whereas wild-type HSCs could be serially transplanted for at least 4 rounds. Furthermore, the rate of telomere shortening was increased approximately 2-fold during serial transplantation of telomerase-deficient HSCs. These findings suggest that one role for telomerase in the HSC is to partially counter the rate of telomere shortening during division of HSCs, thereby preventing premature loss of telomere function and providing added replicative capacity.


The inability to serially transplant whole bone marrow in mice more than 4 to 6 times has led to speculation that hematopoietic stem cells (HSCs), like various somatic cell types grown in vitro, have a limited replicative life span.1,2 Telomeres, essential genetic elements that cap and protect the ends of chromosomes, have been shown to shorten during replicative aging of many different human somatic cell strains.3 Overexpression of telomerase reverse transcriptase (TERT), the catalytic component of telomerase,4 in cultured human somatic cells results in the cessation of telomere shortening and immortalization,5 demonstrating a direct causal role for telomere shortening in replicative senescence in vitro. Telomeres also shorten in human hematopoietic cells, including HSCs, during aging in vivo6,7 and during transplantation of bone marrow (BM),8 and in HSCs from mice during serial transplantation.9 However, unlike most other somatic cell types in humans, telomerase activity can be detected in the hematopoietic system, especially in early progenitors10 and HSCs.11 Thus the role of telomerase in HSCs is unclear. Here we assess the effect of telomerase deficiency on long-term transplantation capacity and telomere shortening during serial transplantation of HSCs.

Study design


The derivation of the mTR knock-out mice and mouse telomerase reverse transcriptase (mTERT) knock-out mice has been previously described.12,13 The mTR/ mice and mTERT/ mice were back-crossed 6 and 4 times, respectively, to the C57Bl/Ka-Thy1.1(Ly5.1) strain at the Stanford University animal facility prior to use in this study. In all transplantation experiments, the Thy1.1/Ly5.1 mice were used as HSC donors, and the congenic C57Bl/Ka-Thy1.2(Ly5.2) strain was used as recipients. The initial donor mice and all the recipient mice were 2 to 3 months old. All mice were bred and maintained on acidified water (pH 2.5).

HSC analysis, purification, and transplantation

Bone marrow cells were isolated and stained with fluorophore-conjugated antibodies as previously described.9 The antibodies used in the immunofluorescence staining for HSC detection are as previously described.9 The HSC population is defined as c-kithiSca-1hiThy1.1lolineageneg. Double-sorted HSCs (n = 150) were used in each round of transplantation. Every stage of serial transplantation was performed with long-term reconstituted mice as donors (≥ 4 months after transplantation). All recipient mice were lethally irradiated with a dose of 960r (137Cs). All analyses and cell sorting were performed on a dual-laser Vantage (Becton Dickinson, San Jose, CA) fluorescence-activated cell-sorter (FACS) machine.

Fluorescent in situ hybridization (FISH)

Telomeres were detected by FISH using a fluorescein isothiocyanate (FITC)–conjugated peptide nucleic acid telomeric probe ((CCCTAA)3; Applied Biosystems, Foster City, CA), as described.14 Background noise was corrected using unstained slides for each sample processed in parallel. All quantitative analyses were performed using a Zeiss confocal microscope (Jena, Germany).

Analysis of telomerase activity

The telomeric repeat amplification protocol (TRAP) assay and quantification of telomerase activity was performed using the TRAP assay kit from Intergen (Purchase, NY), as described15 but with the following modifications: the final concentration for all deoxynucleoside triphosphates was 10 mM, and 0.2 μg TS primer was used per 25 μL reaction.

Results and discussion

Successive generations of breeding mice in which the gene encoding the RNA component of telomerase (mTR) has been knocked out eventually result in the critical shortening of telomeres and telomere dysfunction16 and in attenuation of hematopoietic cell function and proliferation.17 To address the biologic significance of telomerase in HSCs in more detail, we compared the extent of telomere shortening and replicative capacity of HSCs from wild-type (mTR+/+) and generation 1 (G1) mTR/ mice during serial transplantation. We observed an abrupt decrease in the frequency of donor-derived cells in the fifth and third rounds of transplantation during serial transplantation of HSCs, the c-kithiSca-1hiThy1.1lolineageneg BM subpopulation, from mTR+/+ and mTR/ mice, respectively (Figure 1A). During serial transplantation of HSCs from mTR+/+ mice, the frequency of donor-derived cells decreased slightly after the first 2 rounds of transplantation and then stabilized at 75% to 80% until the fifth round of transplantation. However, earlier studies report a gradual decrease in frequency of donor-derived cells at each step of serial transplantation until donor-derived cells are no longer detectable.2,18 We believe this discrepancy is likely due, at least in part, to the method of serial transplantation. Notably, we find that it is necessary to avoid use of radioprotective BM and to remove any host-derived cells prior to transplantation of HSCs or whole BM cells, particularly at the third and fourth rounds of serial transplantation, in order to achieve a high frequency (ie, ≥ ∼40%) of donor-derived cells in the long-term recipient mice (not shown).

Figure 1.

Analysis of donor-derived cell frequency and telomerase activity during serial transplantation of HSCs. (A) FACS-purified HSCs (n = 150) from donor mTR–/– mice (n = 4) and mTR+/+ mice (n = 4) were serially transplanted until exhaustion of transplantation capacity. The fraction of donor-type (Ly5.1+) cells in the bone marrow of long-term (> 4 months) reconstituted mice (n ≥ 5) was determined by FACS analysis at each stage of serial transplantation. The decrease in frequency of donor-derived BM cells during the first and second rounds of serial transplantation of mTR+/+ HSCs and mTR–/– HSCs was statistically significant (P < .05; Student t test). Error bars (standard deviation) are shown. (B) HSCs (n = 250) from adult mice and recipient mice at each stage of serial transplantation were sorted directly into lysis buffer. Telomerase activity was measured for duplicate sample extracts at each stage of serial transplantation by the TRAP assay. A telomerase extract from NIH3T3 cells (n = 2500) was included as a positive control. The internal control for PCR efficiency is indicated by the arrow on the right. (C) Telomerase activity was measured for HSC samples from 3 or more mice at each stage of serial transplantation and averaged for all samples. The mean level of activity and error bars (standard deviation) are shown. We did not observe a significant change in telomerase activity level in HSCs during serial transplantation (P = .26).

To assess whether the finite reconstitutive potential of HSCs and reduction of telomeric size during serial transplantation of HSCs9 might be attributed to a concomitant decrease in telomerase expression, we assessed telomerase activity in FACS-purified HSCs from long-term reconstituted recipients at each stage of serial transplantation (Figure 1B-C). As shown in Figure 1, no significant change in telomerase activity in HSCs was observed (P = .26), indicating that this is not a factor.

During the serial transplantation of HSCs, we measured telomere length using FISH14 in mTR+/+ and mTR/ HSCs from long-term (> 4 months after transplantation) multilineage reconstituted mice at each stage of serial transplantation (Figure 2A-B). This analysis revealed a significantly reduced telomere length in HSCs from 1° (the first stage of serial transplantation) recipients and 2° recipients that were reconstituted with mTR/ HSCs compared with 1° (P = .008) and 2° (P = .006) recipient mice reconstituted with mTR+/+ HSCs. Also, the rate of telomere loss was approximately 2-fold higher during transplantation of mTR/ HSCs compared with mTR+/+ HSCs (Figure 2B). This increased rate of telomere shortening was associated with a decreased transplantation capacity for mTR/ HSCs (Figure 2C). We also observed an increased rate of telomere shortening in donor-derived BM and a reduced transplantation capacity during serial transplantation of BM cells from mTR/ mice (not shown). FISH analysis of telomere length during serial transplantation of HSCs from mTERT+/+ mice and mTERT/ mice also revealed a more rapid loss of telomere length (∼1.7-fold) and a reduced transplantation capacity for HSCs deficient in telomerase (Figure 2D-E). Furthermore, we observed both end-to-end dicentric chromosomes and an elevated frequency of signal-free chromosome ends (< 2% and 95% of metaphase spreads from mTR+/+ and mTR/ 2° HSCs, respectively; Figure 2F) in 2° mTR/ HSCs, supporting the notion that the attenuated transplantation capacity of telomerase-deficient HSCs is caused by critical shortening of telomeres.

Figure 2.

Analysis of telomere length and replicative capacity during serial transplantation of HSCs from telomerase-deficient mice. (A) At each stage of transplantation, approximately 2000 HSCs were purified via FACS from a total of 1 to 2 mice and cytospun onto glass slides. All HSC samples were collected at 15 months (mTR+/+ donor) or 7 months (mTR–/– donor) after the initial transplantation. Detection of telomeres by FISH was performed using an FITC-tagged peptide nucleic acid telomeric probe as previously described.14 Individual interphase nuclei are indicated by arrowheads and the size scale (5 μm; bottom right) is shown in the top left panel. Original magnification, × 400. (B) The fluorescent signal intensity was measured for 25 nuclei at each stage of transplantation using a Zeiss confocal microscope. During serial transplantation of HSCs from both mTR–/– and mTR+/+ mice, the change in fluorescent signal intensity was significant (P < .05) for all stages except 3° → 4° for mTR+/+ HSCs (P = .21; Student t test). (C) The average frequency of donor-derived HSCs from long-term (> 4 months) reconstituted mice (n = 5) was measured at each stage of serial transplantation. The HSC frequency is shown relative to the fraction of donor-derived cells. Error bars (standard deviation) are shown.All mTR+/+ and mTR–/– mice used in this analysis and FISH analysis of telomere length were also used in the analysis of frequency of donor-derived BM cells (Figure 1). (D-E) FISH analysis of telomere length and analysis of donor-derived HSC frequency were performed during serial transplantation of HSCs from mTERT+/+ mice and mTERT–/– mice as described above.At least 4 mice were used in the assessment of donor HSCs at each stage of transplantation. Signal intensity was significantly diminished in HSCs from 1° recipients and 2° recipients that were reconstituted with mTERT–/– HSCs compared with 1° (P = .006) and 2° (P = .009) recipient mice reconstituted with mTERT+/+ HSCs. The analysis of telomere signal intensity and HSC frequency for the 2° mTERT–/– HSC donor and successive 3° and 4° recipients in which telomere length and replicative life span were observed to increase are indicated by the open circle and dashed line. Error bars show SD. (F) HSCs from 2° recipients of either mTR+/+ or mTR–/– HSCs were FACS sorted into media containing 10% fetal bovine serum, interleukin-3 (IL-3), IL-6, IL-11, and Steel factor (all 10 ng/mL) and cultured for 5 days. Metaphase spreads were then prepared using standard procedures, and telomeres were detected using FISH. End-to-end-chromosome fusions are indicated by the arrowheads. Original magnification, × 1009,

Our finding of an accelerated loss of telomeric DNA during serial transplantation of telomerase-deficient HSCs is in contrast to the recent study by Samper et al who found little decrease in telomere fluorescence after transplantation of mTR/ BM.18 However, the mice used as donors by Samper et al were G3 mTR/ mice and therefore likely had a shorter, possibly near critical, initial telomere length, perhaps, as proposed,18 leading to the selection of HSCs or more primitive stem cells in the G3 mTR/ embryo in which alternative mechanisms19 for telomere maintenance were active.

Interestingly, in 1 of 8 2° mTERT/ HSC recipients that were used as donors for the third round of serial transplantation, we observed robust hematopoietic reconstitution and increase in telomere length in HSCs from long-term 3° recipients (Figure 2D-E). Furthermore, we have also been able to serially transplant HSCs derived from this 2° donor into 4° recipients and observe increased telomere length in 4° HSCs as well (Figure 2D-E). As expected, telomerase activity could not be detected in donor-derived cells from these 3° or 4° recipients (not shown). These data indicate that a telomerase-independent alternative mechanism19 for telomere replication has been activated in these HSCs, in agreement with a previous study on the effect of immunization on telomere length in B cells from late generation mTR/ mice.20

We find that HSC frequency remains constant during serial transplantation until donor-derived cells drop to very low levels (Figure 2C,E), in contrast to a recent study that reports a rejuvenation of the hematopoietic stem cell compartment in recipient mice back to approximately only 5% of that observed in bone marrow from resting adult mice.21 However, this discrepancy is likely accounted for by differences in both analysis of HSCs and transplantation methodology. These differences include: (i) measurement of the HSC population on the basis of cell surface phenotype, as opposed to limit dilution; (ii) calculation of HSC frequency within the donor-derived hematopoietic cell compartment, as opposed to whole bone marrow; (iii) the transplantation of only FACS-sorted HSCs or donor-derived whole bone marrow as opposed to the inclusion of competitor cells; and (iv) differences in the number of HSCs transplanted and the radiation dose of the recipient mice.

The data presented here indicate that one function for telomerase in HSCs is to provide an extended replicative capacity of HSCs via a reduction of the rate of telomere shortening during cell divison, thereby preventing a premature critical shortening of telomeres and abrogation of telomere function. The extended replicative capacity conferred by telomerase in HSCs is likely to be important throughout the organismal life span during times of high hematopoietic stress, such as infections and excessive blood loss, as evidenced by the delayed time for regeneration of peripheral blood cell count in old mTR/ mice following exposure to 5-fluorouracil.22 Moreover, the autosomal dominant form of dyskeratosis congenita, a progressive bone marrow failure syndrome, has recently been shown to be specifically associated with mutations in the human hTR gene, as well as accelerated telomere shortening.23 The ability of small numbers of telomerase-deficient murine HSCs to provide efficient hematopoietic recovery after 2 rounds of transplantation compared with the progressive BM failure observed within the lifetime of those afflicted with dykeratosis congenita may be explained by the 2 to 3 times longer telomere length in murine cells, as well as possible additional unknown deleterious consequences of expression of mutant hTR in the hematopoetic system including HSCs. The apparent reduction of replicative capacity of HSCs following extensive telomere shortening will be important to take into consideration during therapeutic transplantation of bone marrow or HSC-enriched populations in humans and any potential cancer therapies involving telomerase inhibition.


We would like to thank Lea Harrington for critical reading of the manuscript and for providing the mTERT knock-out mice. Thanks to L. Jerabek and V. Braunstein for excellent technical assistance; M. Gilbert and T. Knaak for their operations of the FACS machines; and L. Hidalgo for animal care.


  • Reprints:
    Richard Allsopp, Department of Pathology, Rm B259, Stanford University School of Medicine, Stanford University, Stanford, CA 94305; e-mail: rallsopp{at}
  • Prepublished online as Blood First Edition Paper, March 27, 2003; DOI 10.1182/blood-2002-07-2334.

  • Supported by National Institutes of Health (NIH) grants CA 42551 and DK 53074 (I.L.W.), and fellowship support from the National Research Service Award/NIH and the Irvington Institute of Immunology (R.C.A.).

  • Two of the authors (C.B.H., G.B.M.) are employed by a company (Geron Corporation) whose potential product was studied in the present work.

  • The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked ``advertisement'' in accordance with 18 U.S.C. section 1734.

  • Submitted July 5, 2002.
  • Accepted March 11, 2003.


View Abstract