Main

Blood cell turnover in humans requires the production of tens of billions of cells per day with rapid increments during times of physiological stress. Maintenance of cell production requires a highly cytokine-responsive progenitor cell pool with prodigious proliferative capacity and a smaller population of stem cells intermittently feeding daughter cells into the proliferative compartment. The stem cell pool itself is relatively quiescent and cytokine resistant, a state that seems to be necessary for the prevention of premature depletion during times of stress1,2,3. The dichotomy of resistance to proliferative signals by stem cells as compared with the brisk responsiveness by progenitor cells is a central feature of the blood-forming compartment that is poorly understood.

It has recently been demonstrated that entry into the cell cycle in stem cells is governed by the CDKI, p21. In the absence of this inhibitor, the stem cell pool is larger, more actively cycling, and more sensitive to exhaustion. Dominant inhibitory control over stem cell cycling exerted by p21 dictates the size and kinetics of the stem cell compartment, and preserving the stem cell population is dependent on restricting access into the cell cycle4.

Hypothesizing that distinct molecular mechanisms may predominate in the stem cell compartment as compared to its prolific descendent progenitor pool, we evaluated other CDKI family members. The Cip/Kip family of CDKIs includes p21, p27, and p57kip2. The p27 molecule differs from p21 in its C terminus; it interacts with similar but not identical cyclin-CDKs, and it lacks p53-regulated expression5,6,7. The p27−/− mouse is also markedly different from the p21−/− mouse, with a larger body habitus, hyperplasia of most organs (including hematopoietic organs), spontaneous generation of benign pituitary tumors, and infertility8,9,10. Like p21, however, p27 is associated with postmitotic differentiation in some cell types11,12,13, and antisense p27 can suppress cell cycle arrest in mesenchymal cells14,15. Unlike p21, p27 is controlled by both translational and post-translational mechanisms16,17, and p27 has been uniquely demonstrated to be an intrinsic timer for proliferation of neural cells, restricting the number of cell divisions18,19. A role for p27 in hematopoiesis is supported directly by flow cytometric evidence for expression in primitive cells 20 and expression in more mature progenitors 21,22, and indirectly by improved retroviral transduction in the context of antisense p27 (ref. 23). We investigated the function of p27 in the hematopoietic cascade through the use of mice engineered to be deficient in p27 and a series of in vitro and in vivo measures of stem and progenitor cell function.

Stem or progenitor cell pool size

We first assessed the impact of p27 deletion on different hematopoietic cell compartments by quantifying the functional populations of progenitor cells (using methylcellulose colony-forming cell (CFC) assays) and of more primitive cells (using long-term cobblestone area-forming cell (CAFC) assays24,25. The latter assays correlate linearly with in vivo repopulating potential and were used here as a functional stem cell assay.

We observed a marked contrast between p27−/− mice and p21−/− mice. In the p21−/− animals, stem cell populations were doubled and progenitor populations unchanged in previous studies4,26, whereas p27−/− animals had an increase in progenitors but no change in stem cell numbers. Decreased numbers of CAFCs per nucleated cell in the p27−/− animals were noted compared to +/+ animals (33% reduction of CAFC frequency in −/− animals, n = 7, P = 0.0391). However, normalization of the values for the overall increase in marrow cellularity in the p27−/− mouse (4.40 vs. 2.83 × 107/femur pair; P = 0.0072; n = 6) indicated that the number of stem cells per hematopoietic organ was not significantly different from the +/+ control (P = 0.3861, n = 7) (Fig. 1a). However, the progenitor population was significantly different, with an increased CFC population in p27−/− versus +/+ animals (P = 0.0006, n = 5) (Fig. 1b). Therefore, a disproportionate increase in progenitor populations and overall cellularity diluted the stem cell fraction, but the absolute number of stem cells was unchanged from control.

Figure 1: The long-term culture and colony-forming assays demonstrate an unchanged stem cell pool size and an enlarged progenitor pool in the p27−/− mice.
figure 1

a, Comparison of CAFCs scored at week 5 between p27+/+ and −/− mice (per harvest of two femurs). Each data point was generated from three to five limiting dilutions. Each pair was pooled from two to three −/− or +/+ littermate mice in each experiment. Data were analyzed using the paired t-test (P = 0.3861, n = 7). b, Comparison of CFCs between p27+/+ and −/− mice (per harvest or two femurs). Data represent colony-forming ability at day 10. Each pair was pooled from two to three −/− or +/+ littermate mice in each experiment. Each data point was generated from four replicates, and data were analyzed using the paired t-test (P = 0.0006, n = 5). Each line shows one data pair from the same experiment, and the bold thicker line shows the average value from all the independent experiments.

Full size image

Cell cycle profile

To directly measure cell cycle parameters of primitive cell populations in the p27+/+ and −/− animals, we carried out flow cytometric analysis. Because hematopoietic stem cells have been shown to be positive for the stem cell marker Sca-1 and negative for lineage markers 27, we reasoned that lineage marker-expressing cells in the Sca-1+ population reflected a population of maturing lineage-committed progenitors. We confirmed this by testing for CAFC and detected a decrease of 9–20 times in Lin+Sca-1+ cells compared to Lin−Sca-1+ cells. Therefore, we used flow cytometry to separate the enriched stem cell (Sca-1+Lin−) and progenitor cell (Sca-1+Lin+) pools from marrow nucleated cells, and analyzed the cell cycle status (S+G2/M percentage) by simultaneously staining with the DNA dye, To-pro-3. We observed a similar S+G2/M percentage of Sca-1+Lin− in the p27+/+ and −/− animals (P = 0.3591, n = 6), but a higher S+G2/M percentage of Sca-1+Lin+ in the p27−/− animals (Fig. 2a and 2b, P = 0.0215, n = 7). To further distinguish a quiescent fraction (G0) versus G1 in the stem cell pool, we used the RNA dye, Pyronin Y, to stain the marrow nucleated cells within a stringent gate of Lin− cells in conjunction with a DNA dye, Hoechst 333421,28. No significant difference was observable between p27−/− and +/+ cells (Fig. 2c), unlike the p21−/− stem cells, in which a significantly lower fraction of quiescent cells had been previously found4. These data indicate an unperturbed cell cycle status of stem cells, but an increased fraction of progenitor cells in active cycle in the absence of p27.

Figure 2: Altered cell cycle profile of progenitor cells, but not stem cells, in the p27−/− marrow.
figure 2

Mouse bone marrow cells were stained with lineage antibodies and stem cell marker (Sca-1) to separate the enriched stem (Sca-1+Lin−) and progenitor (Sca-1+Lin+) pools (a, upper panel). Simultaneous staining with the DNA dye To-pro-3 was used to determine the percentage of S+G2/M cells in each population (a, middle and lower panels). Data from multiple experiments are summarized in b (*P = 0.0215, n = 7 in Sca-1+Lin+ cells; P = 0.3591, n = 6 in Sca-1+Lin− cells). To determine the ratio of G0 to G1 in the stem cell population, the RNA dye pyronin Y and the DNA dye Hoechst 33342 were used instead of To-pro-3, and the percentage of G0 (PYlow) was obtained in the G0/G1 fraction (HoechstlowLin–) (ref. 4,28) shown in c (P = 0.1591, n = 7). Each data point represents the mean from one to three −/− or +/+ littermate mice in each experiment. Data were analyzed using the paired t-test.

Full size image

For functional evaluation of cell cycle status we exposed animals to the cell cycle-dependent antimetabolite 5-fluorouracil (5-FU), which selectively kills cycling cells29,30. Littermate −/− or +/+ mice received injections of 200 mg/kg of 5-FU or phosphate-buffered saline alone; marrow was harvested one day later, and CAFC and CFC assays performed (Fig. 3). No difference in the yield of CAFC was noted between +/+ and −/− animals, indicating similar proliferative kinetics in the primitive or stem cell compartment (P = 0.2852, n = 6). However, we observed a significant reduction of CFC after 5-FU injection in the p27−/− group compared to the +/+ group controls (82.7% versus 50.6%, P = 0.0044, n = 5) (Fig. 3). The proliferative fraction of cells in the progenitor pool was therefore larger in those animals lacking p27, providing a basis for the expanded size of the progenitor cell population.

Figure 3: Treatment with 5-FU in vivo demonstrates more active cell cycling in the progenitor pool, but not in the stem cell pool of the p27−/− mice.
figure 3

One day after a single intravenous injection of 5-FU at the dose of 200 mg/kg, cells were obtained for long-term culture with limiting dilution and colony forming-ability were obtained. CAFCs were counted at week 5, and CFCs were counted at day 10. y-axis values = ((CAFCs or CFCs from untreated mice − CAFCs or CFCs from 5-FU-treated mice) / CAFCs or CFCs from untreated mice) × 100%. Data represent the mean from multiple independent experiments. Three littermates for each genotype were used in each experiment, and three to five limiting dilutions were used for each sample in the long-term culture. The Student's t-test was used for comparative analysis: *P = 0.0044, n = 5 for CFC and P = 0.2852, n = 6 for CAFC, comparing p27−/− (▪) with p27+/+ (□) cells.

Full size image

Apoptotic fraction

Under homeostatic conditions, an enlarged cell population in vivo may be caused by increased cell proliferation, decreased cell death, or both. To assess whether or not altered apoptosis contributed to the expanded progenitor compartment, we evaluated cells by Annexin-V staining31 and could detect no difference in apoptosis in either the Sca-1+Lin− stem cell pool or Sca-1+Lin+ progenitor cells between p27−/− and +/+ littermate control mice (mean 2.5 ± 0.8 vs. 2.5 ± 0.8 and 7.3 ± 2.5 vs. 7.2 ± 2.2, respectively; n = 4).

Transplantation analysis of stem and progenitor cells

Reasoning that steady state of the stem cell compartment may be unperturbed yet other physiological functions affected under stress, we carried out sequential bone marrow transplantation. Bone marrow from p27+/+ or −/− male animals in each genotype was transplanted into 10 lethally irradiated female mice. One to four months after engraftment, 1–2 × 106 bone marrow cells from the transplanted recipients were used as donor cells for a lethally irradiated host and the same procedure was repeated sequentially. Chimerism was determined as ∼100% donor derived after each transplant by semiquantitative Y chromosome-specific (Sry) polymerase chain reaction (PCR)32 and p27 genotyping PCR (data not shown). There was no difference in bone marrow homing among p27−/− stem cells compared to +/+ controls as assessed by carboxyfluorescein diacetate succinimidyl diester33,34 staining of Sca-1+Lin− cells (data not shown). Stem cell quantitation was performed by CAFC analysis after each transplantation. A comparable decay rate in CAFC was noted in each group, indicating that stem cell renewal was equivalent in the p27−/− and +/+ animals (Fig. 4a).

Figure 4: Serial bone marrow transplantation (BMT) demonstrates an unaltered self-renewal of hematopoietic stem cells and an enhanced activity of progenitor cells in the p27−/− transplanted mice.
figure 4

a, CAFC decline relative to pre-BMT sample (BMTO) over the course of serial BMT. The donor cells of each transplant were subjected to long-term culture with limiting dilution to quantify the frequencies of stem cells. Normal, not-transplanted marrow was used as a control to assure the quality of the stroma and the comparability of the experiments at different times. There was no significant difference between the p27−/− (●) and p27+/+ (○) groups. b, CFC activities during serial BMT indicate expansion of progenitor pools in the p27−/− transplanted mice. The Student's t-test was used for analysis: *P = 0.001 in the third BMT and P > 0.05 in other BMTs between p27−/− (▪) and p27+/+ (□) cells. c Short-term radiation protection of the marrow from the 4th transplant. 105 cells from the fourth BMT mice were transplanted into the lethally irradiated recipients, and survival data were analyzed using a log-rank nonparametric test (P = 0.036, n = 10 in the p27−/− group (●) or n = 9 in the p27+/+ group (○) and expressed as Kaplan–Meier survival curves.

Full size image

Interestingly, however, the progenitor cell pool from the p27−/− animals was capable of expansion and relative regeneration after serial transplantation when wild-type progenitors were markedly depleted (Fig. 4b). Furthermore, the functional capacity of these cells was evident in improved animal survival in a short-term radiation-protection assay (Fig. 4c), even after the fourth serial transplant when stem cells were no longer detectable. These data demonstrate markedly altered cell kinetics among progenitors, but not stem cells, in the absence of p27. This contrasts dramatically with the increased stem cell pool and unaffected progenitor population in the p21−/− setting4.

Competitive repopution

We next tested the role of p27 and thereby the role of progenitor cell cycle inhibition in the context of long-term engraftment. We performed a competitive transplantation in which −/− and +/+ bone marrow nucleated cells were admixed 1:1 and transplanted into an irradiated wild-type recipient. It should again be noted that the representation of stem cells in the −/− nucleated cell preparations is proportionately lower than in +/+ controls. The admixture of the stem cell population is therefore uneven, with 40% derived from −/− marrow. After transplantation, semiquantitative PCR of p27 was used to monitor each genotype in populations of bone marrow and blood cells over a one-year interval. It has been shown in other settings that the proportion of stem cells from a normal host is reflected in a similar proportion of total bone marrow cells and blood cells35,36. This is the basis for the competitive repopulation experiments performed in congenic mice as a tool for measuring stem cell populations. However, in the context of altered cell cycle regulation by p27 deficiency, proportionate representation in various cellular compartments was strikingly altered.

Even though the fraction of −/− stem cells transplanted was approximately 40%, after six months the fraction of p27−/− cells in the blood reached levels of ∼80% and was sustained at elevated levels (Fig. 5a). In addition, the fraction of −/− marrow nucleated cells (predominantly a progenitor cell population and its descendents) was also >80% at the time of euthanasia at 11 or 12 months (Fig. 5b). We carried out long-term culture initiating cell (LTC-IC), and CFC analyses on the marrow specimens. Individual CAFC, LTC-IC or CFC were then isolated by micropipette for PCR analysis of the p27 genotype. The data confirmed that the proportion of genotypically −/− cells in the CAFC, LTC-IC, or stem cell population was comparable to or less than what was transplanted. In contrast, the CFC or progenitor cell population demonstrated a relative overrepresentation of the −/− genotype (Fig. 5c). The fraction of stem cells transplanted thereby contributed disproportionately to the progenitor cell population, which in turn contributed disproportionately to the blood cell population in the absence of p27. Feedback governing cell pool size therefore is skewed in the absence of p27, permitting overgrowth of progenitors and their descendents in a competitive situation (Fig. 5d).

Figure 5: Competitive repopulation assay demonstrates preferential outgrowth of p27−/− progenitor and mature cells following long-term engraftment.
figure 5

Equal numbers of bone marrow nucleated cells from p27+/+ and p27−/− mice (five mice for each genotype) were mixed and transplanted into lethally irradiated recipients. Blood was collected at 6, 9 and 11 months for semiquantitative p27 genotyping PCR analysis a. At 11–12 months, mice were sacrificed and bone marrow nucleated cells were prepared for PCR analysis b, and hematopoietic cell culture. Left portion of a and b ( p27−/−%/p27+/+%) represents titration of indicated genotypes. Mouse # indicates recipient animals. Individual colonies from CFC culture or individual CAFC and LTC-ICs from different wells were harvested and analyzed by PCR for p27 to determine the distribution of p27−/− (▪) or p27+/+ (□) cells in the indicated compartment c. d, Overall observed results from this study compared with the conventionally expected result: p27−/− (▪); p27+/+ (□).

Full size image

Discussion

These data support highly differentiation stage–specific regulatory roles for distinct members of the CDKI Cip/Kip family (Table 1). The clearly delineated and apparently exclusive dominance of p27 in progenitor cells and of p21 in stem cells demarcates a molecular boundary that is unique as far as we know. Cytokine receptors30, chemokine receptors37, adhesion molecules38,39, and transcription factors40,41,42,43 have been shown to be expressed in overlapping populations of precursor populations. Whereas other regulatory molecules may contribute to the differential sensitivity of stem cells and progenitors to proliferative signals, cell cycle control is highly divergent at the level of the G1-S checkpoint. The distinction between the participating CDKIs may explain in part the highly dichotomous proliferative capability of stem cells as compared to the progenitor cells that characterize the hematopoietic and other differentiation systems. Additionally, it provides specific targets for selective manipulation of stem cell versus progenitor cell compartments. To the extent that hematopoiesis mimics other stem and progenitor populations in tissue development, this distinction may point to useful strategies for altering specific precursor pools in size and activity.

Table 1 Distinct impact of p21 and p27 on hematopoiesis
Full size table

The observation that competition between p27−/− and +/+ cells results in overrepresentation of the −/− progenitor and blood cells indicates the critical function of inhibition in dictating homeostasis in the later phases of hematopoiesis. The importance of pro-proliferative cues for hematopoiesis has been demonstrated44, but the crucial involvement of inhibitors of proliferation is demonstrated in the p27−/− mice in this study for progenitors and in p21−/− mice for stem cells elsewhere4. Where there is an inability to exert the cell cycle inhibition mediated by these molecules, disruption of normal population kinetics occurs. However, the proliferation that does occur in the p21−/− and p27 −/− mice is not as overwhelming as has been observed, for example, with disruption by an inhibitory cytokine such as TGF-β (ref. 45). Although the p27−/− animals have slightly higher blood counts than +/+ controls (white blood cells, 9.08 ± 3.12 vs. 6.89 ± 2.11 × 103/μl; erythrocytes, 8.37 ± 1.09 vs. 8.26 ± 1.41 × 106/μl; platelets, 723.10 + 172.74 vs. 635.40 ± 105.40 × 103/μl; n = 10), neither the p27- nor p21-deficient animals develop leukemia or gross polycythemia as indicated by cell counts, morphology, and phenotypic analysis by flow cytometry (data not shown). Therefore, other negative regulators must be active beyond a certain threshold of cell expansion. It is within physiological ranges of the hematopoietic compartment size that p27 and p21 appear to exert dominant roles in modulating cell dynamics.

The ability of a minority population of p27−/− stem cells to predominate in the progenitor and mature blood cell compartments indicates the potential efficacy of using p27 to enhance the efficiency of small numbers of stem cells. A controlled reduction in p27 might make it possible to effect a marked alteration in a substantially larger fraction of blood cells, particularly in the settings where small numbers of stem cells may be transduced with a therapeutic gene. The absence of an untoward effect in vivo demonstrated in the p27−/− mouse provides conceptual support, although extensive additional experimental evidence is needed. The ability of this genetic alteration to increase the size of other, nonhematopoietic tissues in vivo8,9,10 indicates that controlled manipulation of p27 may also be relevant for the expansion or possible regeneration of other tissue types.

Methods

Generation of homozygous mice.

We obtained heterozygote 129/B6 p27 +/− mice from the laboratory of Dr. Andrew Koff (Sloan-Kettering Cancer Center, New York) under the permission of the Subcommittee on Research Animal Care of the Massachusetts General Hospital (MGH). Mice were housed in sterilized microisolator cages and received autoclaved food and drinking water at the MGH animal core facility. The heterozygotes (+/−) were backcrossed into 129/sv background and bred to yield homozygous and wild-type offspring. The littermates from the same +/−parents were used in each experiment.

Mouse genotyping.

Genotyping was accomplished by DNA PCR. Briefly, genomic DNA was isolated from tail biopsy and analyzed by amplification using three primers (the sequences were provided by Dr. Andrew Koff): SW40 (5′-TCA AAC GTG AGA GTG TCT AAC GG -3′), SW41 (5′- AGG GGC TTA TGA TTC TGA AAG TCG -3′), and SW39 (5′-ATA TTG CTG AAG AGC TTG GCG G -3′). SW40 is a forward primer that binds the region nucleotides 4–26 of pLamda-KIP-34-1. Used in conjunction with SW41, a reverse primer binding to nucleotides 209–186 of pLamda-KIP-34-1, this will produce a PCR product of 206 bp from the wild-type locus. SW40 used in conjunction with SW39, a forward primer binding to nucleotides 1420–1441 of PMC1POLA, will produce a PCR product of 298 bp from the mutant locus. All three primers were used together in the same reaction to detect wild-type and mutant loci. The conditions for thermocycling were as follows: (step 1) 94 °C, 4 min; (step 2) 35 cycles of 90 °C, 30 s, 55 °C, 30 s, 72 °C, 1 min; (step 3) 72 °C, 10 min. Diagnostic mutant and wild-type amplified bands were detected on a 2.0% agarose gel post-visualization with ethidium bromide. Semiquantitative PCR was done using 1 μg of genomic DNA. The same primers and thermocycling parameters were applied as above, except that only 28 cycles were carried out for step 2. The ratios of these PCR products were compared against a proportional titration curve of mutant and wild-type amplified bands (Fig. 5a).

Bone marrow sampling.

Mouse bone marrow was obtained from eight to twelve week-old animals from each group (−/−, +/+) after euthanizing with CO2. The marrow cell suspensions were flushed from femurs and tibias, filtered with 100-mesh nylon cloth (Sefar America, Inc., Kansas City, Missouri), and stored on ice until use.

Flow cytometric analysis.

Flow cytometry was used to quantify the cell cycle status in the stem cell compartment. Bone marrow nucleated cells were labeled with anti-lineage antibodies (CD3, CD4, CD8, B220, Gr-1 and Mac-1 from Caltac (Burlingame, California); TER-119 from PharMingen (San Diego, California)) and stem cell marker (Sca-1; PharMingen, San Diego, California). An enriched stem cell phenotype (Sca-1+Lin−) and a progenitor phenotype (Sca-1+Lin+) were gated, and a DNA dye, To-pro-3, was used to stain the antibody-bound cells simultaneously to measure the cycling cell percentage in the populations. To measure stem cell quiescence, cells were stained with lineage antibodies, incubated with 1.67 μmol/L DNA dye, Hoechst 33342, and 1 μg/ml RNA dye, Pyronin Y, and the ratio of G0 vs. G1 was then measured in the Lin− population28. To detect apoptotic cells, Annexin-V (Caltac, Burlingame, California) in conjunction with the DNA dye 7-AAD (refs. 31,37), was used to stain Sca-1+, Lin−, or Lin+ cells, which were then analyzed by flow cytometry. Cells excluding 7-AAD and binding Annexin-V were considered apoptotic.

Colony-forming assay.

This assay was used to measure the progenitor cell frequency (CFC) as described in our previous publication4. Murine stem cell factor (SCF) was used in this study instead of human SCF in the previous one.

Long-term culture with limiting dilution.

To quantify the stem cells, we adapted the CAFC assay 25 with minor modifications as described in our previous publication4.

Serial bone marrow transplantation.

Serial bone marrow transplantation was used to evaluate the ability of stem cells to self-renew as described in our previous publication4. Briefly, male mice were used as marrow donors. Female recipient mice were lethally irradiated with 10 Gy whole-body irradiation at 5.96 Gy/min. Two million nucleated cells were injected intravenously into the lateral tail veins of warmed recipients. Recipient mice were monitored daily for survival for more than one month. The mice were euthanized after one to four months, and bone marrow cells were prepared from those euthanized and injected into new female irradiated recipients. This process was repeated four more times

Short-term radiation-protection assay.

105 marrow nucleated cells from the fourth transplant were transplanted into lethally irradiated female mice as described above, and animal survival frequency was plotted for each group after 30 days. Results were analyzed using a log-rank nonparametric test and expressed as Kaplan–Meier survival curves.

Competitive long-term repopulation.

Equal numbers of bone marrow nucleated cells from p27+/+ and p27−/− mice were mixed and transplanted into the lethally irradiated recipients as described in the serial transplantation section. Blood was collected at six and nine months for semiquantitative p27 PCR analysis. After 12 months, mice were euthanized and bone marrow nucleated cells were prepared for PCR analysis and hematopoietic cell culture (CFC, CAFC, and LTC-IC; see the CFC and long-term culture sections). Individual colonies from the CFC culture or individual CAFC/LTC-ICs from different wells were isolated by micropipette and analyzed by PCR for p27.