Mouse telomerase reverse transcriptase (mTert)expression marks slowly cycling intestinal stem cellsRobert K. Montgomerya,b,1, Diana L. Carloneb,c,1, Camilla A. Richmonda,b,c, Loredana Farillab,c,Mariette E. G. Kranendonkc, Daniel E. Hendersonc, Nana Yaa Baffour-Awuaha, Dana M. Ambruzsc,Laura K. Foglic, Selma Algraa, and David T. Breaultb,c,d,2

aDivision of Gastroenterology, Children’s Hospital Boston, Boston, MA 02115; cDivision of Endocrinology, Children’s Hospital Boston, Boston, MA 02115;bDepartment of Pediatrics, Harvard Medical School, Boston, MA 02115; and dHarvard Stem Cell Institute, Cambridge, Boston, MA 02138

Edited* by Robert A. Weinberg, Whitehead Institute for Biomedical Research, Cambridge, MA, and approved November 23, 2010 (received for reviewSeptember 7, 2010)

The intestinal epithelium is maintained by a population of rapidlycycling (Lgr5+) intestinal stem cells (ISCs). It has been postulated,however, that slowly cycling ISCs must also be present in the in-testine to protect the genome from accumulating deleteriousmutations and to allow for a response to tissue injury. Here, weidentify a subpopulation of slowly cycling ISCs marked by mousetelomerase reverse transcriptase (mTert) expression that can giverise to Lgr5+ cells. mTert-expressing cells distribute in a patternalong the crypt–villus axis similar to long-term label-retaining cells(LRCs) and are resistant to tissue injury. Lineage-tracing studiesdemonstrate that mTert+ cells give rise to all differentiated intesti-nal cell types, persist long term, and contribute to the regenerativeresponse following injury. Consistent with other highly regenera-tive tissues, our results demonstrate that a slowly cycling stem cellpopulation exists within the intestine.

Every day the small intestinal epithelium of a normal mousesheds ∼2 × 108 cells into the lumen of the gut, placing it

among the most highly regenerative mammalian tissues (1). It iswell established that the absorptive and secretory lineages of thecontinuously renewing intestinal epithelium are maintained bymultipotent intestinal stem cells (ISCs) (2), although theiridentity remained elusive until recently due to a lack of definitiveISC markers (3). Using lineage-tracing techniques, Hans Clevers’group identified Lgr5 (4) and Ascl2 (5) as markers for highlyproliferative crypt base columnar (CBC) cells, which contributeto all intestinal lineages during extended chase. Their high rateof proliferation, however, was a surprising characteristic, givenmost mammalian stem cell populations are thought to bemaintained in a slowly cycling (largely quiescent) state (6). Ad-ditional ISC markers have recently been identified, althoughtheir cell cycle status has yet to be established. For example, theCapecchi group defined Bmi-1, a member of the Polycomb groupgene family, as an ISC marker that appears to mark cells that arelargely distinct from Lgr5+ cells (7), although more recent evi-dence supports some overlap. Whereas the coexistence of qui-escent and active stem cells has been demonstrated in othermammalian tissues, the presence of quiescent ISCs remainscontroversial (8).Relative resistance to cellular senescence, despite multiple

rounds of cell division, is a common characteristic of stem cells.Telomerase is a ribonucleoprotein complex that helps maintainthe telomeric ends of chromosomes, normally shortened witheach cell division. Because loss of telomeric DNA beyond a criti-cal threshold induces senescence in most somatic cells, mainte-nance or induction of telomerase activity provides a means ofpreventing cellular senescence (9) that may be relevant for theself-renewal of tissue stem cells. Consistent with this hypothesis,loss of telomerase has been shown to result in intestinal villusatrophy, suggesting a functional requirement for telomerase ac-tivity and/or telomere maintenance in ISC function (10). In ad-dition, several reports have recently implicated mouse telomerasereverse transcriptase (mTERT) in the direct regulation of stemcell proliferation and mobilization (11, 12).

At the whole tissue level, telomerase activity and expressionhave been identified within self-renewing tissues such as testis,bone marrow, and intestine (13, 14). However, with the excep-tion of testis, telomerase is expressed at very low levels (15–19)and has been localized to single telomerase-expressing cellswithin the lower crypt (20). Previously, we generated a mTert–GFP transgenic mouse model system in which GFP expressionrecapitulates endogenous mTert expression and telomerase ac-tivity (14). Using this model, we have shown that mTert marksembryonic and adult stem cells as well as induced pluripotentstem (iPS) cells (14, 21). In the intestine, prior studies showedthat mTert–GFP marks long-term label-retaining cells (LRCs)within the intestinal crypt, suggesting it may mark quiescent ISCs(14). Here, we report that mTert expression marks a slowly cy-cling ISC population distinct from Lgr5+ cells. mTert+ cellscontribute to all differentiated intestinal cell types as well as theLgr5+ cell population, persist long-term, are resistant to injury,and contribute to the regenerative response following tissue in-jury. Thus, a slowly cycling stem cell exists within the intestinealongside and perhaps upstream of the Lgr5+ population.

ResultsmTert–GFP Expression Marks Single Cells in the Intestinal Crypt.Previously we reported a rare population of single telomerase-positive cells present within intestinal crypts occurring at an av-erage frequency of 1 GFP+ cell per 150 crypts (Fig. 1A and ref.14) consistent with its relatively low expression level (Fig. 1B).To further study these cells, we used EDTA chelation to isolateintestinal epithelial cells for flow cytometry. GFP+ cells werepresent in the main population of cells at a frequency of 1.7 ±0.3% (Fig. 1E) and were found in both proximal and distal in-testine at a ratio of ∼2:1, respectively. For comparison, twoGFP− populations were also studied, which included cells fromthe main population and from a population of smaller cells (Fig.1E), which was enriched for Lgr5+ cells (Fig. 1H). Consistentwith our prior findings (14), telomerase expression and activitywas detected within GFP+ crypt cells but was absent from thesmaller GFP− cells (Fig. 1 F and H). Whereas GFP− cells fromthe main population (Fig. 1E) were largely negative for telo-merase activity, occasionally we detected low levels of activity,which may be attributed to inefficient sorting of the somewhatdim GFP+ cells or possibly to our reporter system markinga subpopulation of mTert-expressing cells. To rule out any effect

Author contributions: R.K.M., D.L.C., C.A.R., and D.T.B. designed research; R.K.M., D.L.C.,C.A.R., L.F., M.E.G.K., D.E.H., N.Y.B.-A., D.M.A., L.K.F., S.A., and D.T.B. performed research;R.K.M., D.L.C., C.A.R., L.F., M.E.G.K., and D.T.B. analyzed data; and R.K.M., D.L.C., andD.T.B. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.1R.K.M. and D.L.C. contributed equally to this work.2To whom correspondence should be addressed. E-mail: david.breault@childrens.harvard.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1013004108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1013004108 PNAS | January 4, 2011 | vol. 108 | no. 1 | 179–184

DEV

ELOPM

ENTA

LBIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1

of genomic integration on transgene expression, two independentlines of mTert–GFP mice were studied. Both lines exhibited anoverlapping GFP expression profile, so for simplicity, resultsfrom one line are presented.

mTert–GFP Cells Are Distinct from Lgr5+ Cells. Historically, the lo-cation of a cell along the crypt axis has been used to infer itsfunction. For example, slowly cycling LRCs, presumed to bestem cells, are distributed throughout the lower crypt with a peakfrequency between positions 4 and 9, commonly referred to as“position +4” (2). Using these previously established criteria, wescored the location of GFP+ cells along the crypt axis. Themajority of cells were present between positions 5 and 8 (Fig.1G), consistent with prior colocalization of mTert–GFP andLRCs (14). A similar distribution has been reported for Bmi-1+

intestinal stem cells (7). In addition, Bmi-1+ and GFP+ cells areoccasionally present in positions 1 and 2, near the crypt bottom(Fig. 1G) (7), the region in which Lgr5-expressing CBC cells areprimarily located (4).Given the slight overlap in position between mTert–GFP+ and

Lgr5+ cells (Fig. 1G) and the stronger overlap with Bmi-1+ cells(7), we next sought to determine whether GFP+ cells coex-pressed either marker. Analysis of gene expression from FACS-isolated cells demonstrated that GFP+ cells do not express Lgr5(Fig. 1H), which was detected only within GFP− small cells (Fig. 1Hand Fig. S1). This finding was confirmed using Lgr5–GFP–IRES–CreER mice. Therefore, subsequent studies using a GFP− pop-ulation used this Lgr5-enriched population. Intriguingly, Bmi-1expression was detected in all populations (Fig. 1E and Fig. S1).In summary, mTert expression marks a population of crypt cellsdistinct from Lgr5+ CBC cells and also marks a subpopulation ofBmi-1+ cells.

mTert–GFP+ Cells Are Slowly Cycling. Telomerase expression isgenerally associated with cellular proliferation (22). Given thehighly regenerative nature of the intestinal epithelium, onewould speculate that mTert-expressing cells are rapidly dividing,similar to Lgr5+ cells. However, the fact that mTert–GFP marksLRCs (14) would suggest these cells are slowly cycling. To ad-dress this discrepancy, we determined the proliferative rate ofmTert-expressing cells. Confocal microscopic analysis of cryptsections revealed GFP+ cells were negative for Ki67 (Fig. 2A), incontrast to the adjacent transit-amplifying cells. To confirm andquantify this observation, we performed Ki67 staining in con-junction with flow cytometry and analyzed the fraction of Ki67+

cells in the GFP+ population as well as the GFP− Lgr5-enrichedsmall cell population. A significant fraction of the GFP− pop-ulation was Ki67+ (43.3 ± 3.7%), consistent with the rapidlycycling nature of Lgr5+ cells (4), whereas only a small fraction ofthe GFP+ population was Ki67+ (5.3 ± 0.9%) (Fig. 2 B and C).Similar results were obtained when Ki67 staining was performedon FACS-isolated GFP+ cells (GFP+Ki67+, 8.2 ± 1.6%) (Fig.S2). Taken together, these results establish that 90–95% ofGFP+ crypt cells are in G0 in contrast to CBC cells (4).

Coexpression ofmTert–GFP with Putative ISC Markers. To investigatethe relationship ofmTert–GFP+ cells to so-called “activated” stemcells present in position +4 we performed immunohistochemistryusing antibodies for P-β-catS552 (23) and P-PTEN (24). In-terestingly, neither P-β-catS552 nor P-PTEN colocalized withGFP(Fig. 2 D–F and Fig. S3A) despite the fact that P-PTEN+ cellsshared a similar distribution profile to GFP+ cells within the crypt(compare Fig. 1G and Fig. S3B), thereby indicating that mTert–GFP marks ISCs in the nonactivated state. Intriguingly, whereasP-β-catS552 and GFP did not colocalize within the same cell,GFP+ cells were predominantly found adjacent to the moreabundant P-β-catS552+ cells (Fig. 2F and Fig. S4). Analysis ofother proposed ISC markers revealed the majority of GFP+ cells(79.7 ± 1.9%) coexpressed β1-integrin (25), whereas only a smallpercentage coexpressed BMP-R1a (24) (6.8 ± 1.7%), Sca-1 (26)(5.9 ± 1.9%), or DCAMKL-1 (27) (18%) (Fig. S5). Taken to-gether, these results show that mTert–GFP expression partiallyoverlaps with previously proposed ISC markers, indicating it maymark a specific ISC population or a specific differentiation state.

mTert–GFP+ Cells Are Resistant to Injury. In response to injury, theintestine has a large regenerative capacity, which may result fromthe mobilization of a population of quiescent stem cells presentin position +4 (2). Paradoxically, cells in this position have beenshown to be highly sensitive to both low- and high-dose radiation(4, 28), although such studies have not used ISC-specific mark-ers, leaving open the possibility that multiple cell types may bepresent in this region. In contrast, Lgr5+ stem cells have beenshown to be resistant to low-dose (1 Gy) although sensitive tohigh-dose (10 Gy) radiation (4). To assess the sensitivity ofmTert–GFP+ cells to intestinal injury, we administered whole

A B

C D E

F G H

Fig. 1. Analysis of mTert-expressing cells in the intestine. (A) Immunohis-tochemical analysis of GFP in the intestine. Arrowheads indicate single GFP+

crypt cells. Magnification, 40×. (B) Quantitative RT-PCR analysis of mTertexpression in adult mouse small intestine, colon, bone marrow, and testis. Tonormalize the level of input RNA, 18S was used. A representative analysis isshown. Bars represent mean ± SEM, performed in duplicate. (C–E) FACSanalysis of single intestinal cells harvested from mTert–GFP mice. (C) Arepresentative forward and side-scatter plot depicts events already gated forlive intestinal cells. The small and main cell populations are circled. (D and E)Representative wild type (D) and mTert–GFP (E) forward-scatter and GFPplots are shown. Events depicted have already been gated for live, CD45−

small and main cell populations. GFP+ cells, green gate; GFP− cells from thesmall cell population, red gate; GFP− cells from the main cell population,white circle. (F) Telomerase activity for GFP+ or GFP− FACS-isolated intestinalcells from the main and small cell populations, respectively. HEK293 cellextracts, positive control; heat inactivated (HI) HEK293 cell extracts, negativecontrol. Data shown are from two to four experiments, each performed induplicate. (G) Distribution of GFP+ cells within the intestinal crypt relative tothe crypt bottom. Bars represent mean ± SEM of two separate experimentswith a total of 96 cells counted. Distribution of Lgr5+ cells (in red) as pre-viously published (4). Inset shows a schematic illustration of a crypt withLgr5+ cells in red and crypt cell numbering. (H) RT-PCR analysis for mTert,Lgr5, Bmi-1, and 18S on FACS-purified GFP− and GFP+ cells corresponding tothe gates in E.

180 | www.pnas.org/cgi/doi/10.1073/pnas.1013004108 Montgomery et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1

body gamma irradiation (1 or 10 Gy) to mTert–GFP mice andscored for cell death using the dual criteria of rounded apoptoticbody morphology and activated-caspase-3 staining (28). Despitea dose-dependent increase in the percentage of apoptotic cryptcells, analysis of ∼9,000 crypts and 44 GFP+ cells revealed noevidence for apoptotic induction within the GFP+ population(1 Gy, 0/19 cells; 10 Gy, 0/25 cells) (Fig. S6). These results in-dicate that mTert–GFP+ cells at position +4 are resistant to theeffects of ionizing radiation, in contrast to the variable sensitivityof classical position +4 and CBC cells to low- (1 Gy) and high-dose (10 Gy) radiation, respectively (4), and suggest functionalheterogeneity among cells in this position. Interestingly, theseresults are consistent with other regenerative and neoplasticsystems where slowly cycling stem cells are generally radio-resistant, whereas rapidly cycling cells are radiosensitive.

mTert–CreER Expression Marks a Subpopulation of Multipotent ISCs.To functionally establish that mTert marks ISCs, we generatedtamoxifen-inducible mTert–CreER::R26R mice (Fig. S7) andperformed short- and long-term lineage-tracing studies. Shortlyafter induction, histological analysis demonstrated single LacZ-stained cells (Fig. 3 A–C) in a distribution similar to mTert–GFP+ cells. In some cases, staining progressed to small clustersof LacZ-marked cells (Fig. 3 D–I) and ultimately to completestripes emanating from the crypt base to the villus tip (Fig. 3J–L), still detected after 2 y of chase. Interestingly, as previouslydescribed for Bmi-1–induced Cre recombination (7), we alsoobserved a gradient along the intestine with most recombinationevents restricted to the proximal half of the small bowel. Analysisof LacZ-stained villus-crypt units following long-term chaseconfirmed that mTert-expressing cells give rise to both intestinallineages (absorptive and secretory), including all four differen-tiated cell types (Fig. 3 M–R). Furthermore, quantitative analysisof LacZ-stained clones and adjacent nonstained clones showeda similar frequency of each differentiated cell type, confirmingmTert expression marks multipotent ISCs (Table S1). No LacZstaining was detected in placebo-treated control mice.To determine whether telomerase marks stem cells in related

organs, we next examined the colon and cecum of bigenic micefollowing long-term chase. Both tissues demonstrated LacZ-

stained crypts extending from the basal to the luminal surface,which included both colonocytes as well as goblet cells, con-firming permanent marking of long-lived, multipotent colonicstem cells (Fig. 3 S–W).

Majority of Cells Marked by mTert–CreER Expression Are SlowlyCycling. Given only a small fraction of mTert–GFP cells are cy-cling at any given time (Fig. 2 A–C), we speculated that themajority of mTert–CreER-expressing cells might not be activelycontributing to villus stripes under basal conditions. To in-vestigate this possibility, we developed a whole mount LacZcrypt assay, which allows for the microscopic analysis of in-dividual crypts at the single cell level. Consistent with our hy-pothesis, the majority of LacZ-marked crypts (93%) containedonly single or few (2–4) LacZ+ cells, whereas only 7% gave riseto full lineage stripes following 4 d or 1 mo of chase (Fig. 4A).These results indicate that the majority of cells marked bymTert–CreER are quiescent or slowly cycling consistent with themTert–GFP data (Fig. 2 B and C and Fig. S2). The similarity

0

5

10

15

2025

30

35

40

45

50

GFP Ki67 102 103 104

102

103

104

105

103 104 105

102

103

104

105

GFP-Negative

Control B

GFP- GFP+

Ant

i-Ki6

7

GFP+

Control

C

Ki6

7+ (%

)DAPI DAPI MergedGFP P-β-catS552

A

D E FSide Scatter

Fig. 2. Telomerase-positive cells are slowly cycling and do not express ac-tivated stem cell markers. (A) Confocal analysis of GFP and Ki67 expression inintestinal crypts. Arrowhead indicates a GFP+Ki67− cell. Magnification, 63×.Gamma correction was applied to reduce background fluorescence. (B)Representative Ki67 and side-scatter plots depict events already gated forlive, CD45−, main GFP+ (Upper) or small GFP− (Lower) cell populations. (C)Percentage of Ki67+ cells within the GFP+ and GFP− cell populations. Barsrepresent mean ± SEM of two independent experiments performed withfive replicates. (D–F) Coimmunofluorescence for GFP and P-β-catS552 in in-testinal crypts. DAPI (blue) counterstain. Magnification, 60×.

Fig. 3. Lineage-tracing in the small intestine and colon. Histological orwhole mount analysis of intestinal LacZ staining following pulse (A–F) orgreater than 1-mo chase (G–L). (M–R) Histological analysis of LacZ staining inthe various intestinal epithelial cell lineages after 1-mo chase. Colabeling ofLacZ and periodic acid-Schiff (PAS) positive cells corresponding to goblet (Mand N) and Paneth (O and P) cells. (Q and R) Colocalization of the enter-oendocrine marker, chromogranin-A (Q) and LacZ staining (R). Arrow indi-cates an enteroendocrine cell. (S–U) Analysis of long-term labeling of coloniccrypts by whole mount analysis following 1-mo (U) or 6-mo (S and T) chase.Arrows demarcate LacZ+ colonic crypts at low power (S) and crypt openingsat higher power (T and U). A crypt duplication event is shown by arrowheadand hatched box in S and at higher magnification (T, white arrow). (V andW) Colabeling of LacZ+ and PAS+ colonic goblet cells after 6-mo chase.Representative pictures are shown for all results. Magnification for histo-logical images is 60× and 2–10× for whole mount images.

Montgomery et al. PNAS | January 4, 2011 | vol. 108 | no. 1 | 181

DEV

ELOPM

ENTA

LBIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1

between the percentage of marked crypts giving rise to full lin-eage stripes and the proportion of mTert–GFP+ cells that arecycling (Fig. 2C) signifies that the cyclingmTert+ cells give rise tothe differentiated cell lineages. To investigate whether the qui-escent cell population progressively contributes to the activestem cell pool over time, we compared the proportion of LacZmarked crypts containing single cells (no stripes) vs. multiplecells (full lineage stripes) at baseline (<1 mo) and after an ex-tended period of chase (>1 y). Interestingly, with time, there wasa dramatic shift toward marked crypts giving rise to full lineagestripes (7 vs. 65%) (Fig. S8). Taken together these results in-dicated that, whereas mTert-expressing cells contribute mini-mally to normal intestinal homeostasis, over the life of theorganism they progressively contribute to the crypt/villus unitpresumably through the active stem cell pool.

mTert-Expressing Cells Contribute to the Regenerative ResponseFollowing Injury. The intestine is capable of regeneration follow-ing irradiation-induced injury (29). To investigate whethermTert+

ISCs play a role in this process,mTert–CreER mice were exposedto varying doses of ionizing radiation (0, 1, 10, and 15Gy) followedby tamoxifen administration. Analysis of crypts at the peak of theregenerative response (4 d) revealed mTert-expressing cells con-tribute to intestinal regeneration in a dose-dependent manner, asdemonstrated by an 8- to 10-fold increase in the frequency of totalLacZ-marked intestinal crypts (Fig. 4B). Whether this increase isdue to preferential survival of mTert-expressing cells or to in-duction of mTert expression remains to be determined. Further-more, the fraction of crypts containing multiple LacZ-markedcells, compared with single LacZ-marked cells, increased 12- to15-fold with high-dose radiation (Fig. 4C). These results establisha role formTert-expressing cells in the regenerative response andprovide a potential mechanism by which radiation-sensitive ISCsmight be restored following intestinal injury.

mTert-Expressing Cells Can Give Rise to Lgr5-Expressing Cells. It hasrecently been proposed that Lgr5+ cells in skin arise from a more“primitive” stem cell (30). Whether a similar hierarchy exists inthe intestine remains to be determined. Interestingly, our line-age-tracing analysis revealed LacZ+ cells at the base of the crypt(Fig. 3 H and I) where Lgr5+ CBC cells reside, suggesting theymay be derived from mTert+ cells. To investigate this possibilitywe performed short-term pulse-chase experiments with mTert–CreER::R26R mice in combination with flow cytometry. Thisapproach was chosen because adequate LGR5-specific anti-bodies are not currently available. On the basis of our lineage-tracing data (Fig. 3), we predicted that immediately followinginduction, a small number of LacZ+ cells would be present in themain population but not in the Lgr5-enriched small population(Fig. 5A). If Lgr5-expressing cells arise from mTert-expressingcells, then following a period of chase, LacZ+ Lgr5+ cells should

appear in the small population (Fig. 5A, model 1). Alternatively,if mTert-expressing cells do not contribute, then LacZ+ cellsshould be absent from the Lgr5+ small cell population (Fig. 5A,model 2). Following induction, LacZ+ cells were detected only inthe main population (Fig. 5 C and D). No LacZ+ cells weredetected in the small population. Consistent with model 1, fol-lowing a 5-d chase, LacZ+ cells were now present in both cellpopulations (Fig. 5 F and G).To confirm that Lgr5+ cells were present within the LacZ+

small population but not within the main population, we collectedLacZ+ cells from both populations by FACS and performedRT-PCR analysis. Cells isolated from the main population (Fig.5 C and F, red gates) during either the pulse or chase periodproved to beLgr5− (Fig. 5H and I), whereas LacZ+ cells collectedfrom the small population following chase (Fig. 5F, dashed gate)were Lgr5+ (Fig. 5I). These results provide important proof-of-principle evidence that Lgr5+ CBC cells can arise from mTert-expressing ISCs, suggesting a possible mechanism by which theactive stem cell population may be renewed throughout the life ofthe organism.

DiscussionThe traditional view that mammalian stem cells are slowly cy-cling has recently been challenged by the discovery of rapidlycycling (Lgr5+) stem cells in intestine (4). Here we report theidentification of a slowly cycling ISC marked by mTert expres-sion. This ISC population is long lived, multipotent, and distinctfrom rapidly cycling Lgr5+ ISCs. In addition, whereas Lgr5+ cellsand putative ISCs present at position +4 are sensitive to theeffects of ionizing radiation, mTert-expressing cells are resistantto high-dose radiation. Furthermore, our data indicate thatmTert-expressing cells contribute to the regenerative responsefollowing injury and can give rise to Lgr5+ cells, demonstratinga relationship between these populations (Fig. S9). Finally, ourresults demonstrate that both rapidly and slowly cycling stemcells can coexist in the intestine, similar to prior findings in skin(31), suggesting this may be a general feature of many organsystems (8).The coexistence of separate populations of stem cells within

mammalian tissues has raised questions about the mechanisms oftissue maintenance and repair under both homeostatic and re-generative conditions. For example, rapidly cycling stem cells arepostulated to play a major role in tissue homeostasis, whereasslowly cycling cells may function more in tissue regeneration.Consistent with this hypothesis, “dormant” stem cells have re-cently been shown to function following tissue injury, while onlyminimally contributing to tissue homeostasis. Whether mTert-expressing cells represent such a population in the intestineremains to be determined; however, their ability to give rise(albeit at low levels) to rapidly cycling Lgr5+ cells during normalintestinal homeostasis is proof of principle that they may play

*

*

*

Single Crypt Cell

10X

LacZ

+ C

rypt

s / 2

5 m

m2

0

5

10

15

20

25

30

35

40

0Gy 1Gy 10Gy 15Gy0

1

2

3

4

5

Single Cell 2-4 Cells 5 Cells0

5

10

15

20

0 Gy 1 Gy 10 Gy 15 Gy

1 Cell≥ 2 Cells

66% 27% 7%

LacZ

+ C

rypt

s / 2

5 m

m2

Fold

Cha

nge*

* *

*A B C

Fig. 4. Frequency of LacZ-marked crypts under basal and regenerative conditions. Analysis of intestinal LacZ-marked crypts using the whole mount LacZ cryptassay. (A) Analysis of crypts marked with single, few (two to four) or multiple (five or more) LacZ+ cells under basal conditions. Crypts marked with five ormore cells correspond to fully marked crypt/villus units (stripe). Inset is a representative image (arrow) and schematic illustration of a crypt containing a singleLacZ+ cell. The percentage for each group is shown. (B) Analysis of total LacZ-marked crypts following varying doses of radiation. (C) Fold increase in thefraction of crypts from Bmarked by single cells or multiple cells. Bars represent mean ± SEM obtained from two to three animals per group, ANOVA P < 0.001,(*) Bonferroni posttest.

182 | www.pnas.org/cgi/doi/10.1073/pnas.1013004108 Montgomery et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1

a role in restoring the Lgr5+ ISC population. In fact, given therelative sensitivity of Lgr5-expressing cells to ionizing radiation,compared with mTert-expressing cells, it is plausible this isa mechanism by which rapidly cycling Lgr5+ cells might bereestablished following injury. Alternatively, it is theoreticallypossible that Lgr5-expressing cells could periodically exit the cellcycle to protect themselves from injury. It will also be of interestto determine whether Lgr5-expressing cells contribute to themTert-expressing population.Despite the recent identification of additional molecular

markers for ISCs, evidence that a slowly cycling ISC populationexists has remained elusive. Whereas Bmi-1 has been proposed

to mark slowly cycling cells, our studies would indicate that it isexpressed in both rapidly and slowly cycling stem cell pop-ulations, consistent with previous findings (32). In addition,whereas DCAMKL-1 marks slowly cycling cells within the in-testine (27), definitive lineage-tracing data have yet to confirmtheir role as stem cells. The identification of a population ofmTert-expressing ISCs provides important functional evidencefor the existence of a slowly cycling ISC population within theintestinal crypt. The variable coexpression ofmTert–GFP and theproposed ISC markers BMP-R1a, Sca-1, and β1-integrin, as wellas Ki67, suggest these cells represent a heterogenous population.It will be of interest to determine at the single cell level whetheror not multiple markers are coexpressed on these cells and theircorrelation with proliferation status. Curiously, we noted thata similar percentage of mTert–GFP+ cells were positive for thecell surface marker BMP-R1a as were positive for Ki67, raisingthe possibility that BMP-R1a may be marking cycling GFP+

cells. BMP signaling is active in slowly cycling cells present inposition +4 and it has been suggested to play a role in stem cellself-renewal and the response to tissue injury (8).Asymmetric cellular division of a quiescent stem cell results in

a self-renewing daughter cell and a transit-amplifying progenitorcell (33, 34). Whether this hierarchical model applies to the in-testine has recently been studied using lineage-tracing modelstargeting the proliferative compartment (35, 36). These studiesprovide compelling evidence that rapidly cycling ISCs, which areessential for daily homeostasis, undergo symmetrical divisionsfollowing a pattern of neutral drift dynamics. As pointed out,however, these studies cannot rule out the existence of asym-metric division resulting from a slowly cycling ISC population(36). It may now be possible to determine whether asymmetricdivision, resulting from quiescent ISCs, occurs within the intes-tinal epithelium.Our observation that mTert-expression marks a slowly cycling

stem cell population stands in direct contrast to the commonlyheld view that it marks dividing cells, although the available datafor this claim are mixed (22). The Artandi lab has demonstratedthat mTERT induction is sufficient to activate quiescent epi-dermal stem cells (12) and, in the intestine, functions as a tran-scriptional modulator of the Wnt/β-catenin signaling pathway(37). While their data are highly intriguing, it is not readily ap-parent that direct comparisons can be made between their modeland ours. For example, whereas overexpression of mTERTwithin quiescent stem cells (that might not normally expresstelomerase) may cause them to be activated, it is not clear thatendogenous mTERT+ cells are similarly activated. It is alsopossible that sustained vs. transient mTERT expression couldhave differential effects on cellular proliferation. Alternatively,Wnt signaling may be suppressed in some endogenous mTERT+

cells depending on their BMP signaling status, as previouslysuggested (23), and derepression may trigger them to be acti-vated and emerge from quiescence. In addition, mTERT mayactivate only Wnt-dependent cells. Finally, given the lack of anovert phenotype in early generation mTert (−/−) knockout mice,it has been suggested that mTERT may not be essential for stemcell proliferation under basal conditions (38).In summary, we have identified a slowly cycling ISC that

coexists with rapidly cycling stem cells. These cells are primarilylocated above the Paneth cell zone, coincident with the locationof traditional LRCs. A detailed understanding of their role in theregenerative response to injury as well as the role of targetedmutations in neoplastic transformation may ultimately translateinto improved therapeutic treatment options for individuals withintestinal disease and cancer.

Materials and MethodsMice and Labeling Experiments. mTert–GFP, Lgr5–GFP–IRES–CreER (Jax008875), and Gt(ROSA)26Sortm1Sor (R26R) (Jax strain 003474) were describedpreviously (4, 14). mTert–CreER mice were generated as described in SIMaterials and Methods. Mice studied were 1–6 mo of age unless indicated.Procedures were approved by our institutional animal care and use com-

Forward Scatter

LacZ

Exp

ress

ion

(Fl

uo

resc

ein

)

% o

f To

tal L

acZ

+ C

ells

% o

f To

tal L

acZ

+ C

ells

A

B C D

E F G

H I

Fig. 5. mTert+ cells contribute to Lgr5+ cells. (A) Schematic illustration offlow cytometry plots showing the small cell (Lgr5+) and main cell (mTert+)populations at pulse and chase. Open circles, unstained cells; blue circles,LacZ+ cells. LacZ+ cells in the main population give rise to both small andmain cell populations (model 1) or only to the main cell population (model2). (B–G) Flow cytometric analysis of LacZ+ cells at pulse (B–D) or chase (E–G). Representative fluorescein and forward-scatter plots depict events al-ready gated for live, CD45− intestinal cells and indicate LacZ+ cells withinthe small (dashed gate) and main (red gate) cell populations. Control cellsfrom wild-type or placebo-treated bigenic mice used to set the fluorescentgates. (D and G) Quantitation of the percentage of LacZ+ cells from thesmall or main cell populations following pulse (D) or chase (G). A repre-sentative experiment is shown; bars represent mean ± SEM. Nine micewere assayed in duplicate or triplicate. (H and I) RT-PCR analysis of Lgr5and 18S expression in FACS-isolated small and main cell populations fol-lowing pulse (H) or chase (I). Positive (+) control corresponds to wholeintestine cDNA.

Montgomery et al. PNAS | January 4, 2011 | vol. 108 | no. 1 | 183

DEV

ELOPM

ENTA

LBIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1

mittee.mTert–CreER::R26R mice were treated with 10–20 mg of tamoxifen (T)(Sigma), 4-hydroxytamoxifen (4-OHT) (Sigma), or corn oil (per os, s.c., or i.p.)as a single dose or over the course of 3–5 d. LacZ staining was analyzed after1 d to 2 y using whole mount or histological analysis as described in SIMaterials and Methods. Irradiation was administered using a Gammacell 40irradiator at 80 cGy/min.

Flow Cytometry. Intestinal crypt cells were isolated using a chelation method(39) and confirmed to contain >99% epithelial cells. Cells were routinelystained with anti–CD45-PE or -APC (1:100; BD Pharmingen or eBioscience) toexclude all hematopoietic cells. Propidium iodide (1–2 μg/mL) was used toexclude dead cells. In some experiments, cells were stained with rat anti-β1integrin-PE (CD29; 1:400; Biolegend), rat anti–Sca-1-PE (1:100; BD Pharmi-gen), rabbit anti–BMP-R1a (1:20; Santa Cruz) with goat anti–rabbit-PE (1:200;Jackson Immunoresearch), rabbit anti-Ki67 (NeoMarker) with goat anti–rabbit-PB (1:100; Invitrogen). The FluoReporter lacZ Flow Cytometry kit(Molecular Probes) was used to detect β-galactosidase activity as described inSI Materials and Methods. BD FACSAria or MoFlo flow cytometers wereused; data were analyzed using FlowJo (Tree Star).

Immunohistochemistry. Staining was performed as described in SI Materialsand Methods using the following antibodies: rabbit anti-GFP (1:8,000;Abcam), chicken anti-GFP (1:500; Aves Labs or 1:2,000; Abcam), anti–Phospho-PTEN (1:25; Cell Signaling), anti–P-β-catS552 (1:750), anti–DCAMKL-1 (1:200),anti-Ki67 (NeoMarker), and antiactivated caspase-3 (1:100; Cell Signaling).Sections from wild-type mice were used to confirm the specificity of GFPstaining. Images were obtained using a Nikon Eclipse E800 microscope andSpot software, unless specified.

TRAP Assay. Telomerase activity was assessed using cell extracts from 1,000FACS-isolated cells and the TeloTAGGG Telomerase PCR ELISAPlus kit (Roche)in accordance with the manufacturer’s recommendations. Telomerase-expressing HEK293 cells served as a (+) control and heat inactivated (HI)HEK293 cell extracts served as a (−) control.

RT-PCR and Quantitative RT-PCR Analysis. For RT-PCR, 500–8,000 cells weresorted directly into TRIzol reagent (Sigma). For LacZ+ cells, equivalent num-bers of fluorescein-positive cells were analyzed. RNA isolated from wholeintestine was used as a positive control. Following RNA extraction, first strandcDNA synthesis was performed with the iScript Select cDNA Synthesis kit (Bio-Rad Laboratories) and one-tenth the volume was used to perform each RT-PCR analysis. For qRT-PCR analysis, total RNA was isolated from adult tissuesusing TRIzol reagent and cDNA was synthesized as described above. Relativequantification was performed using iQ SYBR green PCR master mix (Bio-Rad)and iCycler (iQ5) real-time PCR detection system (Bio-Rad).

ACKNOWLEDGMENTS. We thank Drs. Joseph Majzoub, Amy Wagers, SusanBonner-Weir, Wayne Lencer, Richard Grant, and Ramesh Shivdasani forcritical review of the manuscript. In addition, we thank Dr. Margaret Thomp-son, Joyce Levecchio, Giri Buruzula, Suzanne White, and Luke Deary for ex-pert technical assistance. This work was supported by Grants R01 DK 084056(to D.T.B.), K08 DK 066305 (to D.T.B.), R21 DK 078198 (to D.T.B.), and R37 DK32658 (to Richard Grand and R.K.M.); the Timothy Murphy Fund, Marco PoloFonds, and Stichting Groninger Universiteitsfonds, Groningen, The Nether-lands (to S.A.); Juvenile Diabetes Research Foundation International (toD.T.B.); Intellectual and Developmental Diseases Research Center Grant P30HD 18655 (to D.T.B.); Harvard Stem Cell Institute (to D.T.B.); and Harvard Di-gestive Disease Center Grant P30 DK 34854 (to D.T.B. and R.K.M.).

1. Potten CS, Loeffler M (1990) Stem cells: Attributes, cycles, spirals, pitfalls anduncertainties. Lessons for and from the crypt. Development 110:1001–1020.

2. Scoville DH, Sato T, He XC, Li L (2008) Current view: Intestinal stem cells and signaling.Gastroenterology 134:849–864.

3. Montgomery RK, Breault DT (2008) Small intestinal stem cell markers. J Anat 213:52–58.

4. Barker N, et al. (2007) Identification of stem cells in small intestine and colon bymarker gene Lgr5. Nature 449:1003–1007.

5. van der Flier LG, et al. (2009) Transcription factor achaete scute-like 2 controlsintestinal stem cell fate. Cell 136:903–912.

6. Fuchs E (2009) The tortoise and the hair: Slow-cycling cells in the stem cell race. Cell137:811–819.

7. Sangiorgi E, Capecchi MR (2008) Bmi1 is expressed in vivo in intestinal stem cells. NatGenet 40:915–920.

8. Li L, Clevers H (2010) Coexistence of quiescent and active adult stem cells in mammals.Science 327:542–545.

9. Blackburn EH (1991) Structure and function of telomeres. Nature 350:569–573.10. Rudolph KL, et al. (1999) Longevity, stress response, and cancer in aging telomerase-

deficient mice. Cell 96:701–712.11. Flores I, Cayuela ML, Blasco MA (2005) Effects of telomerase and telomere length on

epidermal stem cell behavior. Science 309:1253–1256.12. Sarin KY, et al. (2005) Conditional telomerase induction causes proliferation of hair

follicle stem cells. Nature 436:1048–1052.13. Flores I, et al. (2008) The longest telomeres: A general signature of adult stem cell

compartments. Genes Dev 22:654–667.14. Breault DT, et al. (2008) Generation of mTert-GFP mice as a model to identify and

study tissue progenitor cells. Proc Natl Acad Sci USA 105:10420–10425.15. Bekaert S, Derradji H, Baatout S (2004) Telomere biology in mammalian germ cells

and during development. Dev Biol 274:15–30.16. Kolquist KA, et al. (1998) Expression of TERT in early premalignant lesions and

a subset of cells in normal tissues. Nat Genet 19:182–186.17. Ritz JM, et al. (2005) A novel transgenic mouse model reveals humanlike regulation of

an 8-kbp human TERT gene promoter fragment in normal and tumor tissues. CancerRes 65:1187–1196.

18. Niiyama H, et al. (2001) Quantitative analysis of hTERT mRNA expression in colorectalcancer. Am J Gastroenterol 96:1895–1900.

19. Martín-Rivera L, Herrera E, Albar JP, Blasco MA (1998) Expression of mousetelomerase catalytic subunit in embryos and adult tissues. Proc Natl Acad Sci USA 95:10471–10476.

20. Booth C, Potten CS (2000) Gut instincts: Thoughts on intestinal epithelial stem cells. JClin Invest 105:1493–1499.

21. Stadtfeld M, Maherali N, Breault DT, Hochedlinger K (2008) Defining molecularcornerstones during fibroblast to iPS cell reprogramming in mouse. Cell Stem Cell 2:230–240.

22. Greider CW (1998) Telomerase activity, cell proliferation, and cancer. Proc Natl AcadSci USA 95:90–92.

23. He XC, et al. (2007) PTEN-deficient intestinal stem cells initiate intestinal polyposis.Nat Genet 39:189–198.

24. He XC, et al. (2004) BMP signaling inhibits intestinal stem cell self-renewal throughsuppression of Wnt-beta-catenin signaling. Nat Genet 36:1117–1121.

25. Fujimoto K, Beauchamp RD, Whitehead RH (2002) Identification and isolation ofcandidate human colonic clonogenic cells based on cell surface integrin expression.Gastroenterology 123:1941–1948.

26. Dekaney CM, Rodriguez JM, Graul MC, Henning SJ (2005) Isolation and character-ization of a putative intestinal stem cell fraction frommouse jejunum. Gastroenterology129:1567–1580.

27. May R, et al. (2009) Doublecortin and CaM kinase-like-1 and leucine-rich-repeat-containing G-protein-coupled receptor mark quiescent and cycling intestinal stemcells, respectively. Stem Cells 27:2571–2579.

28. Marshman E, Ottewell PD, Potten CS, Watson AJ (2001) Caspase activation duringspontaneous and radiation-induced apoptosis in the murine intestine. J Pathol 195:285–292.

29. Potten CS (2004) Radiation, the ideal cytotoxic agent for studying the cell biology oftissues such as the small intestine. Radiat Res 161:123–136.

30. Snippert HJ, et al. (2010) Lgr6 marks stem cells in the hair follicle that generate all celllineages of the skin. Science 327:1385–1389.

31. Tumbar T, et al. (2004) Defining the epithelial stem cell niche in skin. Science 303:359–363.

32. Snippert HJ, et al. (2009) Prominin-1/CD133 marks stem cells and early progenitors inmouse small intestine. Gastroenterology 136:2187–2194, e1.

33. Cowan CR, Hyman AA (2004) Asymmetric cell division in C. elegans: Cortical polarityand spindle positioning. Annu Rev Cell Dev Biol 20:427–453.

34. Lechler T, Fuchs E (2005) Asymmetric cell divisions promote stratification anddifferentiation of mammalian skin. Nature 437:275–280.

35. Lopez-Garcia C, Klein AM, Simons BD, Winton DJ (2010) Intestinal stem cellreplacement follows a pattern of neutral drift. Science 330:822–825.

36. Snippert HJ, et al. (2010) Intestinal crypt homeostasis results from neutral competitionbetween symmetrically dividing Lgr5 stem cells. Cell 143:134–144.

37. Park JI, et al. (2009) Telomerase modulates Wnt signalling by association with targetgene chromatin. Nature 460:66–72.

38. Millar SE (2009) Cell biology: The not-so-odd couple. Nature 460:44–45.39. Mariadason JM, et al. (2005) Gene expression profiling of intestinal epithelial cell

maturation along the crypt-villus axis. Gastroenterology 128:1081–1088.

184 | www.pnas.org/cgi/doi/10.1073/pnas.1013004108 Montgomery et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1