Immunity, Vol.

5, 207–216, September, 1996, Copyright 1996 by Cell Press

Telomerase Activity in Hematopoietic Cells

Is Associated with Self-Renewal Potential

Sean J. Morrison,*‡ Karen R. Prowse,† senescence (Allsopp et al., 1992; Vaziri et al., 1993).
Peter Ho,† and Irving L. Weissman* Harley et al. (1992) have proposed that the biological
*Departments of Pathology and Developmental Biology clock underlying the limited division potential of somatic
Stanford University School of Medicine cells is telomere length. Eukaryotic cells that have ap-
Stanford, California 94305 parently unlimited replicative potential, such as germline
† Geron Corporation cells and Tetrahymena, express telomerase (Greider and
200 Constitution Drive Blackburn, 1985; Allsopp et al., 1992; Kim et al., 1994).
Menlo Park, California 94025 Telomerase is a ribonucleoprotein complex that adds
telomere sequence during each round of DNA replica-
tion to stabilize telomere length. Telomerase is ex-
Summary pressed in 85%–90% of human tumors, but is not detect-
able in the normal tissues from which they are derived
It has been proposed that the biological clock underly- (Kim et al., 1994; Counter et al., 1994b, 1995). Mamma-
ing the limited division potential of eukaryotic cells is lian cell lines express telomerase (Morin, 1989; Prowse
telomere length. We assayed telomerase activity in et al., 1993), and spontaneous immortalization of cells
single cells of the hematopoietic and immune systems. in culture is concomitant with a gain of telomerase ex-
We examined hematopoietic stem cells at four stages pression and stabilization of telomere length (Counter
of differentiation, lineage-committed progenitors, and et al., 1994a; Prowse and Greider, 1995).
mature myeloid and lymphoid cells. The frequency of Hematopoietic stem cells (HSC) have a considerable
telomerase-expressing cells within each population but limited division potential (Siminovitch et al., 1964;
was proportional to the frequency of cells thought Ogden and Micklem, 1976; Reincke et al., 1982). As
to have self-renewal potential. Among bone marrow would be predicted (Harley et al., 1992), populations of
hematopoietic stem cells, 70% exhibited detectable human hematopoietic cells enriched for stem cell activ-
telomerase activity. The telomerase-expressing so- ity show a reduction in telomere length as donor age
matic cells observed in this study are not thought to increases (Vaziri et al., 1994). Nonetheless, low levels
be immortal, and expression was not correlated with of telomerase activity have been observed in human
cell cycle distribution or differentiation state. This leukocytes (Counter et al., 1995; Hiyama et al., 1995;
study demonstrates that the developmental charac- Bodnar et al., submitted), human hematopoietic progen-
teristic most consistently associated with telomerase itors (Chiu et al., 1996; Hiyama et al., 1995), and mouse
expression is self-renewal potential. spleen (Prowse and Greider, 1995). Mouse liver and
kidney have also been shown to express telomerase
Introduction (Prowse and Greider, 1995). In an effort to test the asso-
ciation between telomerase expression and division po-
DNA polymerase replicates DNA in a 59–39 direction and tential, we modified the telomeric repeat amplification
must start from an RNA primer that is bound to the protocol (TRAP) assay to achieve single cell sensitivity
template (Watson, 1972). After digestion of the RNA and used it to examine purified populations of mouse
primers and ligation of the Okazaki fragments, it is esti- hematopoietic cells with varied proliferative potentials.
mated that 25–100 bp are lost from the 39 ends of each
chromosome during each round of DNA replication in Results
somatic cells (Harley et al., 1990; Hastie et al., 1990;
Lindsey et al., 1991; Levy et al., 1992; Allsopp et al., Immortalized cells consistently express telomerase ac-
1992). To protect coding DNA from such loss, all verte- tivity at the population level; however, it was not known
brates have noncoding tandem TTAGGG repeats at the whether there is telomerase activity in all cells of a cell
39 ends of each chromosome, called telomeres (Moyzis line, or only in a subset of cells with particular character-
et al., 1988; Meyne et al., 1989). Shortening or eliminating istics. We developed a variation of the TRAP assay for
telomeres has been associated with cell cycle arrest, telomerase activity that has single cell sensitivity. The
chromosomal instability, and replicative senescence human embryonic kidney cell line 293 was sorted by
(Counter et al., 1992; Sandell and Zakian, 1993). fluorescence-activated cell sorting (FACS) into aliquots
Seminal in the study of the cellular basis of aging of single live cells or small numbers of cells, and these
was the description of the “Hayflick limit” (Hayflick and samples were assayed individually for telomerase activ-
Moorhead, 1961; Hayflick, 1965), which suggests an ab- ity. The results are presented in Figure 1, with results
solute limit to the number of divisions that somatic cells summarized in Table 1. Of single cell samples, 64% were
can undergo. Allsopp et al. (1992) found that the replica- telomerase positive, and all samples of more than 1 cell
tive capacity of fibroblasts correlated better with telo- were positive. It may be that only 64% of 293 cells
mere length than with donor age. Accelerated loss of express telomerase. Alternatively, there may be a false
telomere sequence is associated with premature cellular negative rate associated with the assay. The sorter may
have failed to deposit a cell into the medium within some
‡ Present address: Division of Biology 216-76, California Institute of of the Eppendorf tubes being targeted. Static charge
Technology, Pasadena, California 91125. inside the tubes or slightly off-center tubes can cause
Immunity
208

1995b). Within an immortalized cell line, every cell is

presumed to have the capacity for indefinite clonogenic-
ity, giving rise to developmentally indistinguishable
daughters by self-renewing divisions. By contrast, mito-
ses that yield only differentiated daughter cells, with
reduced developmental potentials relative to the pro-
genitor, are not self-renewing divisions. Normal somatic
progenitors often give rise only to differentiated prog-
eny, via non-self-renewing divisions. Thus, cells can pro-
liferate without self-renewal if mitoses are associated
with both daughter cells undergoing differentiation
events. We sought to examine normal somatic cells
whose developmental potential has been studied at the
clonal level to determine whether telomerase is associ-
ated principally with cycling cells, undifferentiated cells,
proliferating cells, self-renewing cells, or only with rare
immortal cells.

HSC
Mouse HSC can be purified by FACS from the fetal
liver and from the bone marrow (Spangrude et al., 1988;
Figure 1. Most Immortalized 293 Cells Express Detectable Telo- Morrison and Weissman, 1994; Morrison et al., 1995a).
merase Activity Fetal liver HSC are Thy-1loSca-11Lin2Mac-11CD42 (Mor-
The products of telomerase activity assays on immortalized 293 rison et al., 1995a). Bone marrow HSC can be separated
cells are shown. In brief, exact numbers of cells were deposited into three populations with different self-renewal po-
into Eppendorf tubes using a FACS machine operated in cloning
tentials (Morrison and Weissman, 1994). Most Thy-1lo
mode. The tubes contained a buffer in which telomerase, if present
in the deposited cells, could synthesize de novo telomere repeats
Sca-11Lin 2Mac-12CD42c-kit1 cells have long-term self-
by extending a primer provided in the buffer. De novo telomere renewal potential, and most Thy-1loSca-11Lin2Mac-
repeats were then amplified by PCR and detected on a polyacryl- 1loCD42 (Mac-1lo CD42) cells have transient self-renewal
amide gel. PCR amplification of de novo telomere repeats produces potential. The Thy-1loSca-11Mac-1loCD4lo (Mac-1lo CD4lo)
a ladder of tandem hexameres. The synthesis of the amplified re- population includes some cells with self-renewal poten-
peats by telomerase was proven by the RNase sensitivity of the
tial, but mainly cells with no detectable self-renewal
ladders. Addition of RNase during incubation greatly reduced or
eliminated ladder intensity by digesting the RNA component of the
potential by secondary transfer (Morrison and Weiss-
telomerase enzyme. Background bands created by PCR artifacts man, 1994; Morrison et al., submitted). All of these popu-
or amplification of sequences other than de novo telomeres were lations are highly proliferative, with injection of only 10
not sensitive to RNase. The first lane shows the background from cells giving rise to detectable progeny in most irradiated
primers in the absence of sorted cells. Lanes 2–5 show the products recipients, but the populations differ dramatically in self-
of assays conducted on single 293 cells; 3 of 4 lanes contain ladders
renewal potential.
diagnostic of telomerase activity. Lanes 6, 7, and 8 show the prod-
ucts of assays conducted on 5 or 10 293 cells: all were positive for
Cells of each multipotent progenitor population were
telomerase. Lane 8 shows that when RNase was added to an assay rigorously isolated by magnetic bead enrichment fol-
of 10 293 cells, no detectable products were formed. lowed by consecutive rounds of purification using flow
cytometry (Morrison and Weissman, 1994; Morrison et
al., 1995a). Single cells and small numbers of cells from
the cell to be deflected from some tubes or to be depos- each population were assayed for telomerase activity.
ited on the side of the tube above the medium. Some Figure 2 shows representative results of the assays on
cells might express a very low level of telomerase that multipotent progenitors, and all results are summarized
is not detectable. The relative contributions of these in Table 2. Among single fetal liver HSC, 65% were telo-
factors to the frequency of negative samples is un- merase positive. 85% of the single cells within this
known, but it is clear that most 293 cells express te- population are self-renewing multipotent progenitors
lomerase. (Morrison et al., 1995a). Among single long-term recon-
Self-renewal potential is defined as the ability of a stituting bone marrow HSC, 70% exhibited telomerase
progenitor cell to give rise to daughter cell(s) with devel- activity. Around 90% of single cells within this pop-
opmental potentials that are indistinguishable from the ulation are self-renewing multipotent progenitors capa-
progenitor (Siminovitch et al., 1963; Morrison et al., ble of transferring multilineage progenitor activity to

Table 1. Telomerase Activity in Single Cells of the Human Embryonic Kidney Cell Line 293
Telomerase-Positive Samples (number positive/number of samples)

1 Cell 5 Cells 10 Cells 50 Cells 100 Cells

293 cells 29 of 45 19 of 19 11 of 11 2 of 2 2 of 2

293 cells were sorted by FACS in cloning mode into microtiter format PCR strips for assay of telomerase activity.
Self-Renewing Cells Express Telomerase
209

Figure 2. Telomerase Activity in Multipotent Hematopoietic Progenitor Populations That Differ in Self-Renewal Potential
The representative results of telomerase assays conducted on fetal liver HSC (A; Morrison et al., 1995a), long-term reconstituting bone marrow
HSC (B), transiently self-renewing multipotent progenitors (C), and a transient progenitor population that has little detectable self-renewal
potential (D; Morrison and Weissman, 1994). Refer to Figure 1 for an explanation of the assay and how the results were interpreted. Refer to
the legend of Table 2 for a description of the purification of the populations.

secondary recipients (Morrison and Weissman, 1994; by a simplification of the binomial equation based on
Morrison et al., submitted). Of Mac-1loCD42 transient the frequency of negative samples (Campbell, 1989;
multipotent progenitors, 45% had telomerase activity. Smith et al., 1991): log(p) 5 nlog(q), where p is the pro-
Around 60% of single cells within this population are portion of samples that were negative for telomerase,
self-renewing multipotent progenitors, capable of trans- n is the number of cells per sample, and q is the probabil-
ferring multilineage progenitor activity to secondary re- ity of a cell being telomerase negative (which is equiva-
cipients (Morrison and Weissman, 1994; Morrison et al., lent to the proportion of telomerase-negative cells in the
submitted). The majority of long-term reconstituting population). This analysis suggests than only 10% of
HSC in the fetal liver and adult bone marrow express Mac-1loCD4lo cells exhibit telomerase activity. Only 7%
telomerase activity. The majority of transiently self- of Mac-1loCD4lo cells give rise to long-term multilineage
renewing multipotent progenitors also have telomerase reconstitution (Morrison and Weissman, 1994). Only a
activity. minority of Mac-1loCD4lo cells have detectable self-
At the 1-cell level, no Mac-1loCD4lo transient progeni- renewal potential, as most progenitors in this population
tors exhibited telomerase activity; 7 of 16 were positive cannot be detected after transfer to secondary irradi-
at the 5-cell level. The frequency of telomerase-positive ated recipients even if the transfer is attempted at 2, 4,
cells within the Mac-1loCD4lo population can be inferred or 8 weeks after reconstitution of the primary recipient
Immunity
210

Table 2. Telomerase Activity in Hematopoietic Stem Cells and Multipotent Progenitors

Telomerase-Positive Samples (number positive/number of samples)

Multipotent Progenitor
Population Assayed 1 Cell 5 Cells 10 Cells 50 Cells 100 Cells

Fetal liver HSC 13 of 20 11 of 12 4 of 4 2 of 2

Bone marrow HSC 14 of 20 11 of 12 4 of 4 2 of 2
Transient multipotents 9 of 20 12 of 12 2 of 4 2 of 2
Transient progenitors 0 of 16 7 of 16 9 of 14 10 of 10

Fetal liver HSC were purified as Thy-1.1 loSca-11Lin-Mac-11 CD42 cells from day 14.5 fetuses (Morrison et al., 1995a). Bone marrow HSC were
Thy-1.1 loSca-11Lin 2Mac-12CD42 c-kit1 long-term reconstituting multipotent progenitors (Morrison and Weissman, 1994). Transient multipotents
are Thy-1.1lo Sca-11 Lin2 Mac-1loCD42 cells from bone marrow that exhibit transient multipotent progenitor activity (Morrison and Weissman,
1994). Transient progenitors are Thy-1.1 loSca-11Mac-1lo CD4lo cells from bone marrow that contain a combination of transient multipotent
progenitors and progenitors that only exhibit detectable B lineage progenitor activity (Morrison and Weissman, 1994).

(Morrison et al., submitted). It appears that most cells Lineage-Committed Progenitors

within this population are differentiating into committed If telomerase expression is associated with self-renewal
progenitors, rather than engaging in self-renewing divi- potential, rather than proliferative potential or cells that
sions. Thus, the frequency of telomerase-expressing are relatively undifferentiated, then differentiated self-
multipotent hematopoietic progenitors appears to be renewing progenitors should also express telomerase.
closely related to the frequency of cells with self-renewal Two lineage-committed progenitors have been demon-
potential. strated to undergo self-renewing divisions: pro-B cells
An alternate possibility is that a higher proportion of (Osmond, 1990; Hardy et al., 1991) and primitive thymo-
Mac-1loCD4lo cells express telomerase, but at such a cytes (pro-T cells) (Lowenthal et al., 1988; Kraft et al.,
low level that only some multiple of cells have enough 1993). A low proportion of pro-B cells, but a high propor-
telomerase activity to be detected. We have no evidence tion of pro-T cells, appear to have self-renewal potential.
that some threshold level of telomerase activity is re- Pro-B cells, representing 0.3%–0.4% of C57BL-Thy-1.1
quired to be detectable by the assay. Indeed, a similar bone marrow, can be purified by sorting Thy-1loB2201
frequency of 293 cells, HSC, and pro-T cells (Table 3) sIgM2CD431 cells (Hardy et al., 1991; Morrison et al.,
were observed to be telomerase positive. Another varia- 1994). Most of these cells are rearranging their immuno-
tion of the TRAP assay (Kim et al., 1994; Prowse and globulin genes and therefore are undergoing progres-
Greider, 1995) that has been used to quantitate telo- sive restriction in developmental potential rather than
merase activities relative to 293 cells (Kim et al., 1994; self-renewing divisions (Hardy et al., 1991); however, the
Counter et al., 1995) using larger numbers of cells was earliest fraction of pro-B cells retain germline configura-
used to compare levels of telomerase activity among tion of their immunoglobulin genes (Hardy et al., 1991)
these three populations. It was found that bone marrow and therefore do appear to self-renew. The earliest thy-
HSC had 4.2 6 1.9% of the telomerase activity per cell mocytes are contained within the Thy-11(CD3,CD4,
as 293 cells, and pro-T cells had as low as 0.5% (0.8% CD8)2/loc-kit1HSAhi fraction, representing around 0.1%
and 5.5% in other samples) of 293 cell activity. Thus, it of C57BL-Thy-1.1 thymocytes (Wu et al., 1991; Matsu-
would appear that the single cell assay detects telo- zaki et al., 1993). These cells expand in response to
merase activities equally well despite a variation of al- interleukin-4 (IL-4) without detectable differentiation
most three orders of magnitude in the level of expression events such as T cell receptor rearrangement or up-
per cell. It is possible that certain cell types consistently regulation of CD3, CD4, or CD8 (Lowenthal et al., 1988;
express an even lower level of telomerase that is below Kraft et al., 1993). Representative results of telomerase
a detectable threshold; however, the activity in such assays on committed progenitors are shown in Figure
cells would have to be so low as to represent a real 3 and are summarized in Table 3. Although telomerase
biological difference relative to other somatic cells we activity could be detected from single pro-B cells, activ-
assay. The robustness of the assay is such that even if ity was not consistently detected until the 50-cell level.
some cells expressed telomerase below a limit of detec- Based on the binomial equation, we estimate that
tion, the association we observe between a detectable around 5% of pro-B cells have telomerase activity. The
level of telomerase and self-renewal potential would still observation of a rare single telomerase-positive cell ar-
be meaningful. gues against the possibility that only some multiple of

Table 3. Telomerase Activity in Lineage-Committed Progenitors

Telomerase-Positive Samples (number positive/number of samples)

Progenitor Population
Assayed 1 Cell 5 Cells 10 Cells 50 Cells 100 Cells 500 Cells
Pro-T cells 9 of 18 10 of 10 11 of 11 2 of 2
CD41 CD81 thymocytes 0 of 8 0 of 8 12 of 16 8 of 8 8 of 8
Pro-B cells 1 of 7 0 of 8 1 of 11 7 of 7 9 of 9
Bone marrow myeloid 0 of 4 0 of 4 0 of 12 0 of 11 0 of 8
1 2/lo 1 hi 1 1 2/lo
Pro-T cells were Thy-1 (CD3,CD4,CD8) c-kit HSA thymocytes. Small double-positive thymocytes were CD4 CD8 CD3 forward scatterlo .
Pro-B cells were Thy-1loB220 1CD431sIgM2. Bone marrow myeloid cells were Mac-11 Gr-11 B2202Ter119 2.
Self-Renewing Cells Express Telomerase
211

Figure 3. Telomerase Activity in Lineage-Committed Progenitor Populations

The representative results of telomerase assays conducted on four committed progentior populations: pro-T cells (A), small double-positive
thymocytes (B), pro-B cells (C), and bone marrow Mac-11 Gr-11 myeloid cells (D). Refer to Figure 1 for an explanation of the assay and how
the results were interpreted.

pro-B cells has detectable activity. Detectable te- 200,000 Mac-11Gr-11 cells were also unsuccessful in
lomerase activity was exhibited by 9 of 18 single pro-T detecting any telomerase activity. By May–Grunwald
cells. Depending on whether there are false negatives Giemsa staining, the Mac-11Gr-11 cells appeared to be
associated with the telomerase assay, at least half of almost entirely mature neutrophils and monocytes.
pro-T cells express telomerase. These populations contained no early (day 8) or late (day
Having consistently observed telomerase activity in 12) colony-forming unit-spleen (CFU-S) activity, which is
self-renewing populations, we sought to examine more a measure of relatively undifferentiated myeloid progeni-
non-self-renewing populations. The best candidates for tors. We assayed for more mature myeloid progenitors
non-self-renewing cells appeared to be in the myeloid by plating cells from these populations in methylcellu-
lineage, in which the half-lives of the cells are very short lose cultures (Heimfeld et al., 1991) supplemented with
(Athens et al., 1961; Dancey et al., 1976; Issekutz et al., IL-3, granulocyte colony-stimulating factor (G-CSF),
1981; Doherty et al., 1988; Ohgami et al., 1991) and there and granulocyte/macrophage colony-stimulating factor
is no evidence of self-renewing divisions (van Furth and (GM-CSF), cytokines that strongly stimulate the prolifer-
Cohn, 1968; van Furth and Diesselhoff-Den Dulk, 1970; ation of myeloid progenitors. Less than 0.1% of Mac-
van Furth et al., 1973). Using FACS, we purified Mac- 1 1Gr-11 cells had detectable proliferative capacity
11Gr-11 cells, representing around 30% of bone marrow. (formed colonies).
This population did not exhibit detectable telomerase A second population that appears to contain mainly
activity (Figure 3; Table 3). Quantitative TRAP assays non-self-renewing cells is double-positive thymocytes,
(Kim et al., 1994; Prowse and Greider, 1995) of up to which undergo positive selection. Positive selection is
Immunity
212

Table 4. Telomerase Activity in Differentiated Cells from the Spleen

Telomerase-Positive Samples (number positive/number of samples)

Splenic Population
Assayed 1 Cell 5 Cells 10 Cells 50 Cells 100 Cells 500 Cells

B cells 0 of 8 0 of 16 2 of 17 8 of 8 4 of 4
Germinal center B cells 4 of 16 8 of 8 6 of 6
T cells 0 of 8 0 of 4 8 of 16 8 of 8 11 of 11
Myeloid cells 0 of 8 0 of 4 3 of 16 3 of 8 2 of 5 9 of 9
1 2 2 1 2 2 1 2 2
B cells were sorted as B220 CD3 Mac-1 . T cells were CD3 B220 Mac-1 . Myeloid cells were Mac-1 CD3 B220 . Germinal center B cells
were B220hiPNAhiIgD 2/loIgM 1.

coupled to differentiation as failing cells die and passing al., 1987). We immunized mice with sheep red blood
cells differentiate into single-positive thymocytes by cells and then sorted B220hi Peanut agglutininhi IgM1
down-regulating one coreceptor and up-regulating T IgD2/lo B cells from the spleens 8 days later. Of these
cell receptor (Janeway, 1994). We purified CD4hiCD8hi cells, 25% were telomerase positive, and all samples
CD32/loforward scatterlo cells. The vast majority of these of greater than 1 cell were positive. Thus, telomerase
cells have failed positive selection and are destined to activity was more than 10-fold enriched in germinal cen-
die (Guidos et al., 1990). We calculate that only around ter B cells relative to all splenic B cells.
3% of small double-positive cells expressed telomerase Splenic myelomonocytic cells, Mac-11B2202CD32,
activity, as not all 50-cell samples were positive. It is expressed a low and variable level of telomerase activity.
also formally possible that double-positive cells have Using the binomial equation, we estimate that 1.2 6
such a low level of telomerase activity that activity can 0.8% of splenic myeloid cells express detectable te-
only be detected from some multiple of cells. If this were lomerase activity. By the quantitative assay, telomerase
the case, then we would underestimate the frequency activity was undetectable in tens of thousands of cells
of telomerase-positive cells, but the biological signifi- from this population, meaning that activity must be less
cance of an undetectable level of telomerase expression than 0.03% of 293 cells. Splenic myeloid cells appeared
would be unknown. to be mostly mature monocytes and neutrophils by May–
Grunwald Giemsa staining and had no detectable prolif-
Differentiated Cells erative potential in methylcellulose cultures supple-
We sought to determine whether telomerase was asso- mented with IL-3, G-CSF, and GM-CSF. However, some
ciated with differentiated cells with self-renewal poten- splenic macrophages are known to self-renew (Wijffels
tial, even though these cells have much less proliferative et al., 1994; Naito, 1993). It has been calculated that
activity. Peripheral B and T cells have been shown to around 6% of splenic Mac-11 cells are residential macro-
phages that are replenished by self-renewal (van Furth
undergo self-renewing as well as non-self-renewing divi-
and Diesselhoff-Den Dulk, 1984). Residential macro-
sions (Black et al., 1980; Sprent et al., 1991; Hilbert et
phages can be derived from the differentiation of mono-
al., 1994). We purified peripheral B and T cells by sorting
cytes or by the self-renewal of an apparently indepen-
B2201CD32Mac-12 or CD31B2202Mac-12 cells, respec-
dent lineage of residential macrophages (Wijffels et al.,
tively, from the spleens of C57BL-Thy-1.1 mice. Both of
1994; Naito, 1993; van Furth and Diesselhoff-Den Dulk,
these populations expressed low but readily detectable
1984; Volkman et al., 1983). The relative contributions
levels of telomerase activity, as shown in Figure 4 and
of the two sources is tissue specific, so the lack of
summarized in Table 4. Using the binomial equation, we
telomerase activity observed in bone marrow myeloid
estimate that around 2% of B cells and around 7% of T cells may reflect a lack of self-renewing macrophages
cells express telomerase activity. Using the quantitative
there.
TRAP assay (Kim et al., 1994), we found B cells to ex-
press 0.5% or 0.3% of the telomerase activity in 293 Discussion
cells. This is similar to the level usually observed in pro-
T cells, demonstrating that B cells as a population do Most 293 cells, fetal liver HSC, and bone marrow HSC
not express particularly low levels of telomerase. As a expressed telomerase. Most of the cells in these popula-
population, T cells expressed only 0.03% of the te- tions are known to have self-renewal potential. A some-
lomerase activity in the same number of 293 cells. Given what lower proportion of Mac-1loCD42 multipotent pro-
that B and T cells can either clonally expand or differenti- genitors have self-renewal potential, and a somewhat
ate, such as by class switching or Th1/Th2 maturation, lower proportion of cells exhibited telomerase activity.
the proportion of splenic lymphoid cells capable of self- A minority of Mac-1loCD4lo progenitors have self-renewal
renewal cannot be precisely estimated. Thus, while te- potential and a minority had telomerase activity. This
lomerase expression by B and T lymphocytes is consis- association between self-renewal potential and telo-
tent with an association with self-renewal potential, the merase expression is not specific to primitive progeni-
exact frequency of lymphocytes that should express tors. Most pro-T cells may have self-renewal potential,
telomerase according to our hypothesis is unknown. and at least half expressed telomerase. A minority of
If the telomerase-positive lymphocytes are the lym- pro-B cells have self-renewal potential, and a minority
phocytes with self-renewal potential, then germinal cen- expressed detectable levels of telomerase. Bone mar-
ter B cells should be enriched for telomerase activity as row Mac-11Gr-11 cells have not been observed to self-
they are in the process of clonal expansion (Kroese et renew and had no observable telomerase activity. Few if
Self-Renewing Cells Express Telomerase
213

Figure 4. Telomerase Activity in Differentiated Cells of the Immune System

The representative results of telomerase assays conducted on four differentiated cell populations from the spleen: B cells (A), germinal center
B cells (B), T cells (C), and myeloid cells (D). Refer to Figure 1 for an explanation of the assay and how the results were interpreted.

any small double-positive thymocytes have self-renewal with our finding that telomerase is associated with self-
potential, and only a few percent exhibited detectable renewal potential. Indeed, Sharma et al. (1995) recently
telomerase activity. Even relatively mature populations observed that when immortal cells differentiate, telo-
of splenocytes that are known to include self-renew- merase activity is lost. The authors interpreted this as
ing cells contained small populations of telomerase- a link between proliferative potential, or relatively undif-
expressing cells. Self-renewing germinal center B cells ferentiated cells, and telomerase expression. In light of
included a much higher proportion of telomerase-posi- our data, we suggest that their observed loss of telo-
tive cells. Thus, at the single cell level, telomerase ex- merase activity was due precisely to the loss of self-
pression was tightly correlated with self-renewal poten- renewal potential upon treatment with the differentia-
tial. Relatively high levels of telomerase have also been tion-inducing agents. The only documented exceptions
observed in B1a cells, which as a population are known to the apparent link between self-renewal potential and
to self-renew (R. Gerstein, J. Wilshire, and L. A. Herzen- telomerase expression that we are aware of are fibro-
berg, personal communication). We speculate that telo- blasts and endothelial cells. During early passages in
merase might be one element of a conserved genetic culture, both cell types may undergo self-renewing divi-
program that might be widely expressed by hematopoi- sions without expressing detectable telomerase activity
etic and other cells to effect self-renewing divisions. (Kim et al., 1994; Prowse and Greider, 1995); therefore,
None of the telomerase-positive hematopoietic popu- the association of self-renewal and telomerase expres-
lations observed in this study is believed to be immortal. sion may be specific to certain cell types. Furthermore,
Of course, immortal cells such as 293 cells have unlim- telomerase appears to be more frequently expressed
ited self-renewal potential; therefore, the widespread in mouse tissues than in human tissues (Prowse and
telomerase expression by immortal cells is consistent Greider, 1995), so it remains to be determined whether
Immunity
214

telomerase is similarly regulated in mouse and human Chiu et al., 1996), and we have now observed wide-
cells. spread telomerase expression by hematopoietic pro-
The frequency of telomerase-expressing cells ap- genitors and lymphocytes. It is possible that telomerase
peared unrelated to the frequency of cells in S/G2/M could be expressed at a level that decreases the rate at
phases (see also Weng et al., 1996). Most or all fetal which telomeres shorten without fully maintaining their
liver HSC and bone marrow HSC express telomerase, lengths. Second, most non-self-renewing mouse multi-
but fetal liver HSC may divide daily (Morrison et al., potent progenitors and most mouse lymphocytes did
1995a) while only 4% of bone marrow HSC were in not express detectable telomerase activity. Telomere
S/G2/M phases at any one time (Morrison and Weissman, length could decrease in the subsets of cells that do
1994). Conversely, only 7%–10% of splenic T cells and not express telomerase. Based on the frequencies of
Mac-1loCD4lo multipotent progenitors exhibited te- telomerase negative multipotent progenitors and lym-
lomerase activity, but splenic T cells are largely quies- phocytes, the mean telomere lengths of those popula-
cent and a high proportion of Mac-1loCD4lo cells are in tions would be expected to decline at the population
cycle (Morrison and Weissman, 1994). Although there level, at least in the mouse. It is not informative to assay
was no apparent correlation within a population be- for shortening of telomere length in inbred mice owing
to extremely long and variable TTAGGG repeats (Kipling
tween the frequency of telomerase-positive cells and
and Cooke, 1990; Starling et al., 1990).
the frequency of S/G2 /M phase cells, the relative distri-
Based on the persistence of long-term, transiently, or
bution of 2n cells between G1 and G0 could affect te-
non-self-renewing progenitor clones during reconstitu-
lomerase expression. It remains possible that truly qui-
tion of irradiated recipients, we have suggested that
escent G0 cells might not express telomerase. Recent
adult bone marrow multipotent progenitors are segre-
bromodeoxyuridine labeling experiments have sug-
gated into three differentiation states with discrete
gested that most bone marrow HSC might have a long (not continuous) self-renewal potentials (Morrison and
G1 phase rather than being quiescent (S. Cheshier, X. Weissman, 1995). We suggested that a biological clock
Liao, and I. L. W., unpublished data). If this were true, counts cell divisions to regulate the number of self-
the increased telomerase activity observed in germinal renewing divisions within each differentiation state, but
center B cells might result from the activation of splenic that upon transplantation to irradiated bone, the clock
B cells as well as the selection of cells with self-renewal can be reset to the maximum number of divisions per-
potential. Weng et al. (1996) have recently observed missible within the differentiation state. In keeping with
that CD3-mediated activation of human T cells induced the expanding evidence that telomere length may be the
telomerase expression concomitant with the induction biological clock (Harley et al., 1992), this study suggests
of proliferation. In addition to activating lymphocytes, that the expression of telomerase might be the mecha-
the nature of lymphocyte stimulation can also determine nism for resetting the clock. Increasing telomerase activ-
whether lymphocytes clonally expand in self-renewing ity may result in telomere lengthening. This model would
divisions or differentiate. It remains to be determined suggest that, upon transplantation into irradiated bone,
whether the telomerase induction caused by lympho- telomerase might be up-regulated in HSC, resulting in
cyte stimulation is principally associated with the induc- telomere lengthening and increased proliferative poten-
tion of self-renewal or simply with activation out of G0. tial. The development of more sensitive assays for
It remains unclear whether telomere length limits the changes in telomere length will facilitate the testing of
proliferative potential of immune system cells in mice this hypothesis.
or humans. Nonetheless, if the Hayflick limit of somatic
Experimental Procedures
cell division potential is determined by telomere length,
as has been suggested (Harley et al., 1992), the wide- Mouse Strains
spread expression of telomerase by immune system All mice used in these studies were 6- to 9-week-old C57BL/Ka-
cells suggests that the Hayflick limit might not constrain Thy1.1. They were bred and maintained at the animal care facility
the expansion of mouse immune cells. The HSC pool at the Stanford University School of Medicine. All mice were main-
tained on acidified water (pH 2.5).
has been observed to possess sufficient proliferative
capacity to support the demands of normal hematopoie- Bone Marrow Preparation and Staining
sis for multiple mouse lifespans (Harrison, 1979). Simi- Immunofluorescence staining included 19XE5 (anti-Thy1.1), 2B8
larly, memory B and T cells have sufficient capacity for (anti-c-kit), E13 (anti-Sca-1, Ly6A/E). Lineage marker antibodies in-
clonal expansion to maintain themselves in the periph- cluded KT31.1 (anti-CD3), 53-7.3 (anti-CD5), 53-6.7 (anti-CD8),
Ter119 (anti–erythrocyte-specific antigen), 6B2 (anti-B220), 8C5
ery for a lifetime and to proliferate in response to antigen (anti-Gr-1), M1/70 (anti-Mac-1), and GK1.5 (anti-CD4). Bone marrow
exposure (Hilbert et al., 1994). Proliferative capacities multipotent progenitors and fetal liver HSC were purified by FACS
in the immune system that appear to be far in excess of as described previously (Morrison and Weissman, 1994; Morrison
the Hayflick limit might confer a considerable selective et al., 1995a). All cell sorts were performed on a dual laser FACS
(Becton Dickinson), modified as described previously (Parks and
advantage.
Herzenberg, 1984). Progenitors were purified by sorting and then
Telomerase expression by a population is not incon- resorting to obtain precise numbers of cells that were essentially
sistent with a reduction in the mean telomere length of pure for the indicated surface marker phenotype. Resorts were per-
the population with age. The mean telomere lengths of formed in cloning mode directly into MicroAmp reaction tubes (Per-
human lymphocytes (Vaziri et al., 1993) and primitive kin Elmer) containing the TRAP reaction buffer.
progenitors (Vaziri et al., 1994) decrease with donor age. Telomerase Assay
Telomerase had previously been observed in these The telomerase products of the sorted cells were generated by a
populations (Counter et al., 1995; Hiyama et al., 1995; variation of the TRAP method (Kim et al., 1994; Wright et al., 1995).
Self-Renewing Cells Express Telomerase
215

Cells were sorted into 25 ml of preparatory buffer in which telomerase Campbell, R.C. (1989). Statistics for Biologists, Third Edition (Cam-
from the cell could synthesize de novo telomere repeats by ex- bridge: Cambridge University Press).
tending a TS oligonucleotide. The nascent telomere repeats were Chiu, C.-P., Dragowska, W., Kim, N.W., Vaziri, H., Thomas, T.E.,
then detected by polymerase chain reaction (PCR) amplification Harley, C.B., and Lansdorp, P.M. (1996). Differential expression of
using the TS oligonucleotide and an end-labeled RP oligonucleotide telomerase activity in hematopoietic progenitors from adult human
(modified from the CX primer used by Kim et al. [1994]; available bone marrow. Stem Cells 14, 239–248.
from Geron Corp.). When conducted with telomerase-positive cells,
Counter, C.M., Avilion, A.A., LeFeuvre, C.E., Stewart, N.G., Greider,
this procedure yields a ladder of telomere repeats that can be phos-
C.W., Harley, C.B., and Bacchetti, S. (1992). Telomere shortening
phoimaged after separation on an acrylamide gel. The preparatory
associated with chromosome instability is arrested in immortal cells
buffer contained 20 mM Tris–HCl (pH 8.3), 1.5 mM MgCl2, 63 mM
which express telomerase activity. EMBO J. 11, 1921–1929.
KCl, 1 mM EGTA, 1 mg/ml BSA, 0.5% Tween 20, 50 mM dNTPs,
100 ng of TS oligonucleotide, and DEPC-treated water. The buffer Counter, C.M., Botelho, F., Harley, C.B., and Bacchetti, S. (1994).
was placed in 8-tube microtiter format PCR strips (Robbins Scien- Stabilization of short telomeres and telomerase activity accompany
tific). The elevated concentration of Tween 20 was sufficient to immortalization of Epstein-Barr virus-transformed human B lympho-
lyse the cells, allowing telomerase, if present, to extend the TS cytes. J. Virol. 68, 3410–3414.
sequences during a 60 min incubation period at 308C. Incubation Counter, C.M., Hirte, H.W., Bacchetti, S., and Harley, C.B. (1994).
was usually in a microtiter format Perkin Elmer 9600 thermal cycler, Telomerase activity in human ovarian carcinoma. Proc. Natl. Acad.
but a water bath performed just as well. Following incubation, an Sci. USA 91, 2900–2904.
additional 25 ml of amplification buffer was added to each reaction Counter, C.M., Gupta, J., Harley, C.B., Leber, B., and Bacchetti, S.
tube. This buffer contained the same ingredients noted above ex- (1995). Telomerase activity in normal leukocytes and in hematologic
cept the TS oligonucleotide was replaced by 100 ng of g end-labeled malignancies. Blood 85, 2315–2320.
RP reverse primer and 2 U of Taq DNA polymerase (Boehringer
Dancey, J.T., Deubelbeiss, K.A., Harker, L.A., and Finch, C.A. (1976).
Mannheim). The RP oligonucleotide was labeled using T4 poly-
Neutrophil kinetics in man. J. Clin. Invest. 58, 705–715.
nucleotide kinase (New England Biolabs) and 100 mCi of 3000 Ci/
mmol [g-32P]ATP per 1 mg of RP oligonucleotide. T4 PNK was used Doherty, D.E., Downey, G.P., Worthen, S., Haslett, C., and Henson,
at 40 U per microgram of RP and incubated at 378C for 30 min, P.M. (1988). Monocyte retention and migration in pulmonary inflam-
followed by heat inactivation at 658C for 15 min. The end-labeling mation. Lab. Invest. 59, 200–213.
reaction was designed to ensure maximum efficiency of RP labeling Greider, C.W., and Blackburn, E.H. (1985). Identification of a specific
to enhance the sensitivity of detection of amplified de novo telomere telomere terminal transferase activity in Tetrahymena extracts. Cell
repeats. PCR amplification was performed in a Perkin Elmer 9600 43, 405–413.
thermal cycler for 32 rounds of 948C for 30 s and 608C for 30 s. Guidos, C.J., Danska, J.S., Fathman, G.S., and Weissman, I.L. (1990).
Products were analyzed by electrophoresis in 0.63 TBE on 15% T cell receptor-mediated negative selection of autoreactive T lym-
polyacrylamide nondenaturing gels. For additional precautions re- phocyte precursors occurs after commitment to the CD4 or CD8
garding the TRAP setup and process, see Kim et al. (1994). lineages. J. Exp. Med. 172, 835–845.
The quantitative telomerase assay was conducted on cell extracts
Hardy, R.R., Carmack, C.E., Shinton, S.A., Kemp, J.D., and Haya-
obtained by lysing pellets of thousands to hundreds of thousands
kawa, K. (1991). Resolution and characterization of pro-B and pre-
of cells of the purified cell type in a hypotonic buffer containing
pro-B cell stages in normal mouse bone marrow. J. Exp. Med. 173,
0.5% CHAPS and centrifuging the lysate at 100,000 3 g for 30 min
1213–1225.
as described previously (Kim et al., 1994; Prowse and Greider, 1995).
Harley, C.B., Futcher, A.B., and Greider, C.W. (1990). Telomeres
shorten during ageing of human fibroblasts. Nature 345, 458–460.
Acknowledgments Harley, C.B., Vaziri, H., Counter, C.M., and Allsopp, R.C. (1992). The
telomere hypothesis of cellular aging. Exp. Gerontol. 27, 375–382.
Correspondence should be addressed to S. J. M. We thank Libuse Harrison, D.E. (1979). Proliferative capacity of erythropoietic stem
Jerabek for laboratory management, Veronica Braunstein for anti- cell lines and aging: an overview. Mech. Ageing Dev. 9, 409–426.
body preparation, Lucino Hidalgo and Ricardo Salazar for animal
Hastie, N.D., Dempster, M., Dunlop, M.G., Thompson, A.M., Green,
care, and Tim Knaak for operation of the FACS machine. Thanks
go to Dan Levitt, Choy-Pik Chiu, and Calvin Harley for encourage- D.K., and Allshire, R.C. (1990). Telomere reduction in human colo-
rectal carcinoma and with ageing. Nature 346, 866–868.
ment and critical reading of the manuscript. Thanks go to Koichi
Akashi, Eric Lagasse, and Jos Domen for sharing expertise on the Hayflick, L. (1965). The limited in vitro lifetime of human diploid cell
culture and characterization of myeloid cells. S. J. M. was a Howard strains. Exp. Cell Res. 37, 614–636.
Hughes Medical Institute Predoctoral Fellow. This work was sup- Hayflick, L., and Moorhead, P.S. (1961). The serial cultivation of
ported in part by National Cancer Institute grant number CA42551 human diploid strains. Exp. Cell Res. 25, 585–621.
to I. L. W. Heimfeld, S., Hudak, S., Weissman, I., and Rennick, D. (1991). The
in vitro response of phenotypically defined mouse stem cells and
Received July 10, 1996; revised August 1, 1996. myeloerythroid progenitors to single or multiple growth factors.
Proc. Natl. Acad. Sci. USA 88, 9902–9906.
References Hilbert, D.M., Anderson, A.O., Holmes, K.L., and Rudikoff, S. (1994).
Long-term lymphoid reconstitution of SCID mice suggests self-
Allsopp, R.C., Vaziri, H., Patterson, C., Goldstein, S., Younglai, E.V., renewing B and T cell populations in peripheral and mucosal tissues.
Futcher, A.B., Greider, C.W., and Harley, C.B. (1992). Telomere Transplantation 58, 466–475.
length predicts replicative capacity of human fibroblasts. Proc. Natl. Hiyama, K., Hirai, Y., Kyoizumi, S., Akiyama, M., Hiyama, E., Piaty-
Acad. Sci. USA 89, 10114–10118. szek, M.A., Shay, J.W., Ishoika, S., and Yamakido, M. (1995). Activa-
Athens, J.W., Haab, O.P., Raab, S.O., Mauer, A.M., Ashenbrucker, tion of telomerase in human lymphocytes and hematopoietic pro-
H., Cartwright, G.E., and Wintrobe, M.M. (1961). Leukokinetic stud- genitor cells. J. Immunol. 155, 3711–3715.
ies. IV. The total blood, circulating and marginal granulocyte pools Issekutz, T.B., Issekutz, A.C., and Movat, H.Z. (1981). The in vivo
and the granulocyte turnover rate in normal subjects. J. Clin. Invest. quantitation and kinetics of monocyte migration into acute inflam-
40, 989–995. matory tissue. Am. J. Pathol. 103, 47–55.
Black, S.J., Tokuhisa, T., Herzenberg, L.A., and Herzenberg, L.A. Janeway, C.A. (1994). Thymic selection: two pathways to life and
(1980). Memory B cells at successive stages of differentiation: ex- two to death. Immunity 1, 3–6.
pression of surface IgD and capacity for self renewal. Eur. J. Immu- Kim, N.W. Piatyszek, M.A., Prowse, K.R., Harley, C.B., West, M.D.,
nol. 10, 846–851. Ho, P.L.C., Coviello, G.M., Wright, W.E., Weinrich, S.L., and Shay,
Immunity
216

J.W. (1994). Specific association of human telomerase activity with Proliferative capacity of murine hematopoietic stem cells in vitro.
immortal cells and cancer. Science 266, 2011–2015. Science 215, 1619–1622.
Kipling, D., and Cooke, H.J. (1990). Hypervariable ultra-long telo- Sandell, L.L., and Zakian, V.A. (1993). Loss of a yeast telomere:
meres in mice. Nature 347, 400–402. arrest, recovery, and chromosome loss. Cell 75, 729–739.
Kraft, D.L., Weissman, I.L., and Waller, E.K. (1993). Differentiation Sharma, H.W., Sokoloski, J.A., Perez, J.R., Maltese, J.Y., Sartorelli,
of CD324282 human fetal thymocytes in vivo: characterization of a A.C., Stein, C.A., Nichols, G., Khaled, Z., Telang, N.T., and Naraya-
CD32 41 82 intermediate. J. Exp. Med. 178, 265–277. nan, R. (1995). Differentiation of immortal cells inhibits telomerase
Kroese, F.G.M., Wubbena, A.S., Seijen, H.G., and Nieuwenhuis, P. activity. Proc. Natl. Acad. Sci. USA 92, 12343–12346.
(1987). Germinal centers develop oligoclonally. Eur. J. Immunol. 17, Siminovitch, L., McCulloch, E.A., and Till, J.E. (1963). The distribution
1069–1072. of colony-forming cells among spleen colonies. J. Cell. Comp. Phys-
Levy, M.Z., Allsopp, R.C., Futcher, A.B., Greider, C.W., and Harley, iol. 62, 327–336.
C.B. (1992). Telomere end-replication problem and cell aging. J. Siminovitch, L., Till, J.E., and McCulloch, E.A. (1964). Decline in
Mol. Biol. 225, 951–960. colony-forming ability of marrow cells subjected to serial trans-
Lindsey, J., McGill, N.I., Lindsey, L.A., Green, D.K., and Cooke, H.J. plantation into irradiated mice. J. Cell. Comp. Physiol. 64, 23–31.
(1991). In vivo loss of telomeric repeats with age in humans. Mutat. Smith, L.G., Weissman, I.L., and Heimfeld, S. (1991). Clonal analysis
Res. 256, 45–48. of hematopoietic stem-cell differentiation in vivo. Proc. Natl. Acad.
Lowenthal, J.W., Ransom, J., Howard, M., and Zlotnik, A. (1988). Sci. USA 88, 2788–2792.
Up-regulation of interleukin 4 receptor expression on immature (Lyt- Spangrude, G.J., Heimfeld, S., and Weissman, I.L. (1988). Purifica-
22 /L3T42 ) thymocytes. J. Immunol. 140, 474–478. tion and characterization of mouse hematopoietic stem cells. Sci-
Matsuzaki, Y., Gyotoku, J.-i., Ogawa, M., Nishikawa, S.-i., Katsura, ence 241, 58–62.
Y., Gachelin, G., and Nakauchi, H. (1993). Characterization of c-kit Sprent, J., Schaefer, M., Hurd, M., Surh, C.D., and Ron, Y. (1991).
positive intrathymic stem cells that are restricted to lymphoid differ- Mature murine B and T cells transferred to SCID mice can survive
entiation. J. Exp. Med. 178, 1283–1292. indefinitely and many maintain a virgin phenotype. J. Exp. Med. 174,
Meyne, J., Ratcliff, R.L., and Moyzis, R.K. (1989). Conservation of 717–728.
the human telomere sequence (TTAGGG)n among vertebrates. Proc. Starling, J.A., Maule, J., Hastie, N.D., and Allshire, R.C. (1990). Exten-
Natl. Acad. Sci. USA 89, 7049–7053. sive telomere repeat arrays in mouse are hypervariable. Nucl. Acids
Morin, G.B. (1989). The human telomere terminal transferase enzyme Res. 18, 6881–6888.
is a ribonucleoprotein that synthesizes TTAGGG repeats. Cell 59, van Furth, R., and Cohn, Z.A. (1968). The origin and kinetics of
521–529. mononuclear phagocytes. J. Exp. Med. 128, 415–433.
Morrison, S.J., and Weissman, I.L. (1994). The long-term repopulat- van Furth, R., and Diesselhoff-Den Dulk, M.M.C. (1970). The kinetics
ing subset of hematopoietic stem cells is deterministic and isolat- of promonocytes and monocytes in the bone marrow. J. Exp. Med.
able by phenotype. Immunity 1, 661–673. 132, 813–828.
Morrison, S.J., and Weissman, I.L. (1995). Heterogeneity of hemato- van Furth, R., and Diesselhoff-Den Dulk, M.M.C. (1984). Dual origin
poietic stem cells: implications for clinical applications. Proc. Am. of mouse spleen macrophages. J. Exp. Med. 160, 1273–1283.
Assoc. Phys. 107, 187–194. van Furth, R., Diesselhoff-Den Dulk, M.M.C., and Mattie, H. (1973).
Morrison, S.J., Lagasse, E., and Weissman, I.L. (1994). Demonstra- Quantitative study on the production and kinetics of mononuclear
tion that Thylo subsets of mouse bone marrow that express high phagocytes during an acute inflammatory reaction. J. Exp. Med.
levels of lineage markers are not significant hematopoietic progeni- 138, 1314–1330.
tors. Blood 83, 3480–3490. Vaziri, H., Schachter, F., Uchida, I., Wei, L., Zhu, X., Effros, R., Cohen,
Morrison, S.J., Hemmati, H., Wandycz, A.M., and Weissman, I.L. D., and Harley, C.B. (1993). Loss of telomeric DNA during aging of
(1995). The purification and characterization of fetal liver hematopoi- normal and trisomy 21 human lymphocytes. Am. J. Hum. Genet. 52,
etic stem cells. Proc. Natl. Acad. Sci. USA 92, 10302–10306. 661–667.
Morrison, S.J., Uchida, N., and Weissman, I.L. (1995). The biology Vaziri, H., Dragowska, W., Allsopp, R.C., Thomas, T.E., Harley, C.B.,
of hematopoietic stem cells. Annu. Rev. Cell Dev. Biol. 11, 35–71. and Lansdorp, P.M. (1994). Evidence for a mitotic clock in human
Moyzis, R.K., Buckingham, J.M., Cram, L.S., Dani, M., Deaven, L.L., hematopoietic stem cells: loss of telomeric DNA with age. Proc.
Jones, M.D., Meyne, J., Ratliff, R.L., and Wu, J.-R. (1988). A highly Natl. Acad. Sci. USA 91, 9857–9860.
conserved repetitive DNA sequence, (TTAGGG)n, present at the Volkman, A., Chang, N.C., Strausbauch, P.H., and Morahan, P.S.
telomeres of human chromosomes. Proc. Natl. Acad. Sci. USA 85, (1983). Differential effects of chronic monocyte depletion on macro-
6622–6626. phage populations. Lab. Invest. 49, 291–298.
Naito, M. (1993). Macrophage heterogeneity in development and Watson, J.D. (1972). Origin of concatameric T4 DNA. Nature New
differentiation. Arch. Histol. Cytol. 56, 331–351. Biol. 239, 197–201.
Ogden, D.A., and Micklem, H.S. (1976). The fate of serially trans- Weng, N.-p., Levine, B.L., June, C.H., and Hodes, R.J. (1996). Regu-
planted bone marrow cell populations from young and old donors. lated expression of telomerase activity in human T lymphocyte de-
Transplantation 22, 287–293. velopment and activation. J. Exp. Med. 183, 2471–2479.
Ohgami, M., Doerschuk, C.M., Gie, R.P., English, D., and Hogg, J.C. Wijffels, J.F.A.M., de Rover, Z., Beelen, R.H.J., Kraal, G., and van
(1991). Monocyte kinetics in rabbits. J. Appl. Physiol. 70, 152–157. Rooijen, N. (1994). Macrophage subpopulations in the mouse spleen
renewed by local proliferation. Immunobiology 191, 52–64.
Osmond, D. (1990). B cell development in the bone marrow. Semin.
Immunol. 2, 173–180. Wright, W., Shay, J., and Piatyszek, M. (1995). Modifications of a
telomeric repeat amplification protocol (TRAP) result in increased
Parks, D.R., and Herzenberg, L.A. (1984). Fluorescence activated
reliability, linearity, and sensitivity. Nucl. Acids. Res. 23, 3794–3795.
cell sorting: theory, experimental optimization, and applications in
lymphoid cell biology. Meth. Enzymol. 108, 197–241. Wu, L., Antica, M., Johnson, G.R., Scollay, R., and Shortman, K.
(1991). Developmental potential of the earliest precursor cells from
Prowse, K.R., and Greider, C.W. (1995). Developmental and tissue-
the adult mouse thymus. J. Exp. Med. 174, 1617–1627.
specific regulation of mouse telomerase and telomere length. Proc.
Natl. Acad. Sci. USA 92, 4818–4822.
Prowse, K.R., Avilion, A.A., and Greider, C.W. (1993). Identification
of a nonprocessive telomerase activity from mouse cells. Proc. Natl.
Acad. Sci. USA 90, 1493–1497.
Reincke, U., Hannon, E.C., Rosenblatt, M., and Hellman, S. (1982).