Cycling Status of CD34⁺ Cells Mobilized Into Peripheral Blood of Healthy Donors by Recombinant Human Granulocyte Colony-Stimulating Factor

SUCCESSFUL TRANSPLANTATION of allogeneic recombinant human granulocyte colony-stimulating factor (rhG-CSF )–mobilized peripheral blood stem cells (PBSC) has recently been reported by several groups.1-3 As with autografting, the most striking finding of PBSC transplantation was the faster reconstitution of hematopoiesis after myeloablative conditioning regimen as compared with bone marrow (BM)-derived stem cells.1 The results of these clinical trials raise the question as to whether circulating progenitor cells may differ from their BM counterparts with respect to cell-cycle characteristics, immunophenotype, frequency of both committed and very primitive precursors, and their proliferative response upon stimulation with cytokines. Few reports so far have addressed these issues in the setting of allogeneic transplantation.4-7 One early report4 showed a high expression of myeloid antigens on PB CD34⁺ cells at the expense of B-lineage–associated antigens (ie, CD10, CD19, and CD20) coupled with a high colony-forming capacity of G-CSF–stimulated apheresis products. More recent investigations5,7 have shown that only a small minority of murine and human circulating progenitor cells (either in steady-state or after priming with G-CSF ) is in the S-phase of the cell cycle as compared with BM cells. Thus, it was concluded that progenitors mobilized into PB after cytokine treatment are noncycling, whereas hematopoietic precursors in BM are actively proliferating. However, these seminal studies did not analyze either the cell-cycle distribution of hematopoietic stem cells or the activity of G-CSF on CD34⁺ cells programmed cell death. In addition, it is still unclear whether the most immature subsets within mobilized PBSC show the same functional and kinetic properties as BM cells. It is, in fact, very important to quantify and characterize the cellular populations that are believed to ensure permanent engraftment after PBSC allografting, along with the more committed progenitor cells responsible for short-term BM reconstitution. In this regard, the long-term culture initiating cells (LTC-IC) are the most reliable approximation to true stem cells in humans.8 

To further elucidate the functional and kinetic characteristics of G-CSF–mobilized hematopoietic progenitor cells, we have analyzed highly purified CD34⁺ cells from the apheresis products of 16 normal subjects undergoing PBSC collection for allogeneic transplantation. The results were then compared with those obtained on CD34⁺ cells enriched from the BM of the same donors under steady-state conditions and during G-CSF administration on the same day of PBSC harvest. The results presented here indicate that PB CD34⁺ cells show the same frequency of colony-forming units-cells (CFU-C) and LTC-IC as BM samples. The percentage of both committed and very primitive hematopoietic precursors in the S/G₂M-phase is significantly lower than that observed in the BM at baseline or after G-CSF administration. However, the vast majority of PB CD34⁺ cells, including the more immature LTC-IC, are cycling and myeloid precursors show increased responsiveness to additional cytokines in vitro. Moreover, G-CSF treatment significantly protects PB CD34⁺ cells from apoptosis.

PBSC donors and mobilization protocol.PB and BM samples were obtained from 16 healthy adults (6 men and 10 women, 18 to 55 years old). The donors received glycosylated rhG-CSF (Lenograstim; Rhone-Poulenc Rorer, Milan, Italy) administered subcutaneously at 12 μg/kg/d for 5 to 6 days for mobilization and collection of PBSC before allogeneic transplantation. All donors were related and had HLA-full match with the recipient. Inclusion criteria were good general health (Eastern Cooperative Oncology Group performance status = 0), normal full blood examination, normal coagulation profile, and normal renal and liver functions. Normal subjects were excluded from the study in case of pregnancy or lactation, positive serology for hepatitis C and B virus, human immunodeficiency virus positivity, or any major organ or system dysfunction. Hematologic indices and the number of CD34⁺ cells in PB were recorded at baseline and daily until the completion of PBSC collection. Leukaphereses were performed on days 5 and 6 using either a Fenwal CS3000 continuous flow blood cell separator (Baxter Healthcare Corp, Deerfield, IL), as previously reported,9 or a Cobe Spectra separator (Cobe BCT, Inc, Lakewood, CO). Apheresis products were then cryopreserved in liquid nitrogen until transplantation.9 No donor required central line placement to undergo apheresis. All normal subjects successfully completed mobilization and collection of PBSC. Hematopoietic CD34⁺ cell purification and subsequent studies were always performed using day-5 collections at the peak time of PB CD34⁺ cells. BM specimens were obtained before G-CSF mobilization, in steady-state conditions, and (in selected cases) during G-CSF treatment on the same day as PBSC harvest. To minimize PB cell contamination, no more than 2 to 3 mL of BM was collected from each aspiration. The protocol was approved by the ethical commitee of the University Hospital and each normal donor gave written informed consent.

Cell preparation and CD34⁺ cell purification.Mononuclear cells were obtained by gradient centrifugation (Lymphoprep; 1.077 g/mL; Nycomed Pharma, Oslo, Norway). Light-density cells were washed twice in phosphate-buffered saline (PBS) with 1% bovine serum albumin (BSA; Sigma Chemical Co, St Louis, MO) and CD34⁺ cells were highly purified by MiniMacs high-gradient magnetic separation column (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's instructions.10 To assess the percentage of CD34⁺ elements, aliquots of the CD34⁺ target cells were restained with an antibody (HPCA-2; IgG1a-fluorescein isothiocyanate [FITC]; Becton Dickinson, San Jose, CA) directed at a different epitope of CD34 antigen than that (QBend10) used with the MiniMacs system. Briefly, CD34⁺ cells were incubated for 30 minutes in the dark at 4°C with HPCA-2-FITC. Propidium iodide (2 μg/mL) was added for the detection of nonviable cells, which were excluded from analysis. After two washes in PBS/BSA, flow cytometric analysis was performed on a gated population set on scatter properties by using FACScan equipment (Becton Dickinson). A minimum of 10,000 events was collected in list mode on FACScan software.

Colony assays.Normal BM and PB cells were cultured in semisolid medium as previously described.11 Briefly, 1,000 to 5,000 CD34⁺ cells were plated in duplicate in culture medium consisting of 1 mL of Iscove's modified Dulbecco's medium (IMDM) supplemented with 24% fetal calf serum (Sera Lab, Crawley Down, Sussex, UK), 0.8% BSA (Sigma), 10⁻⁴ mol/L 2-mercaptoethanol (Sigma), 2 U recombinant human erythropoietin (Epo; Dompè Biotec, Milan, Italy), and 0.2 mmol/L bovine hemin. To measure the optimum clonogenic efficiency, 10% (vol/vol) of a selected lot of phytohemagglutinin–lymphocyte-conditioned medium (PHA-LCM) was added. The final concentration of methylcellulose was 1.1%. Granulocyte-macrophage CFU (CFU-GM), erythroid progenitors (burst-forming unit-erythroid [BFU-E]), and mixed colonies (CFU-granulocyte, erythroid, monocyte, megakaryocyte [CFU-GEMM]) were scored after 14 days of incubation at 37°C in a fully humidified 5% CO₂ atmosphere. When indicated, PHA-LCM was replaced by combinations of cytokines including the following recombinant human CSFs: G-CSF (1,000 U/mL), interleukin-3 (IL-3; 50 ng/mL; Genetics Institute, Cambridge, MA), and stem cell factor (SCF; 50 ng/mL; Amgen, Thousand Oaks, CA). All cultures were performed in the presence of 2 U/mL of Epo.

Megakaryocyte progenitor cells (CFU-megakaryocyte [CFU-MK]) derived from highly purified CD34⁺ cells were grown in plasma clot cultures as reported elsewhere.11 Recombinant human cytokines were added at the following concentrations: IL-3 at 10 ng/mL, Epo at 2 U/mL, granulocyte-macrophage colony-stimulating factor (GM-CSF; Amgen) at 2 ng/mL, and IL-6 (Genzyme) at 100 ng/mL. After 12 to 14 days of incubation, plasma clots were fixed with methanol-acetone (1:3) for 20 minutes, washed with PBS, and air dried. Fixed dishes were stored at −20°C until immunofluorescence staining was performed. CFU-MK colonies were recorded as strongly positive aggregates after staining with a monoclonal antibody directed to the glycoprotein complex IIb-IIIa (CD41; Dako, Glostrup, Denmark). Binding was assessed by FITC-goat antimouse IgG (Ortho, Milan, Italy).11 

LTC-IC assay.The number of LTC-IC was determined from the leukapheresis products and the BM, as previously described,12 with some modifications. Briefly, 10,000 to 20,000 highly purified CD34⁺ cells/mL of medium were seeded onto preestablished irradiated genetically engineered murine stromal cells.13 The murine marrow-derived stromal cell line M2-10B4 (kind gift of Dr C. Carlo-Stella, University of Parma, Parma, Italy) was genetically engineered by retroviral-mediated gene transfer to produce G-CSF and IL-3 and was used in all experiments to optimize and standardize the clonogenic output per LTC-IC. The growth medium consisted of 12.5% horse serum (Sera Lab), 12.5% fetal calf serum, and 10⁻⁶ mol/L hydrocortisone sodium succinate (Sigma) in IMDM supplemented with 1% antibiotics and glutamine. Cultures were incubated at 37°C in a fully humidified 5% CO₂ atmosphere with weekly half-medium change. After 5 weeks in culture, nonadherent and adherent cells were pooled, washed, and plated together in methylcellulose for the determination of total progenitor cell content (CFU-GM, BFU-E, and CFU-GEMM), and the number of LTC-IC was calculated as earlier reported.12 

Cytosine arabinoside (Ara-C) suicide assay for S-phase analysis.To evaluate the proportion of hematopoietic progenitors in S-phase, the Ara-C suicide test was performed by incubating 1 × 10⁵ cells/mL with 2 × 10⁻⁶ mol/L Ara-C (Aracytin; Upjohn, Kalamazoo, MI) for 1 hour at 37°C, as previously reported.11 The cycling status of LTC-IC was determined by incubating CD34⁺ cells with Ara-C for 16 to 18 hours in the presence of 20 ng/mL of SCF, IL-3, and G-CSF¹⁴ in serum-containing medium or in serum-free conditions.15 In selected experiments, G-CSF was replaced by 1,000 U/mL of IL-6, and the results were found to be superimposable. After two washes, CD34⁺ cells were plated in semisolid medium (see clonogenic assays) or grown in liquid culture (see LTC-IC assay), and the difference in the number of colonies between treated and untreated cultures, representing the percentage of BM and PB precursors undergoing DNA synthesis, was calculated after 2 and 5 weeks, respectively.

Cell-cycle studies.To evaluate the cell-cycle distribution of CD34⁺ cells, the acridine orange (AO) flow cytometric technique was used. Cellular DNA and RNA content (percentage of cells in G₀ , G₁ , S, and G₂M; mean RNA content of cells in each phase of the cell cycle) was measured as previously reported.15 Discrimination between G₀ and G₁-phase of cell cycle was made on the basis of RNA cellular content. G₀ cells were defined as cells with an RNA content equal or lower than that of G₁ cells and equal to that of control lymphocytes. The RNA content of G₁ cells was expressed as the RNA index (RNA-I G₁ ), which was determined as the ratio of the mean RNA content of G₁ cells of the samples times 10, divided by the median RNA content of control lymphocytes.16,17 Moreover, we used the AO assay to evaluate the number of apoptotic cells as a sub-G₁ peak on the DNA frequency histograms, because these cells can be recognized by their diminished stainability with DNA-specific fluorochromes.15,18 In fact, AO staining allows discrimination between necrotic and apoptotic cells because of decreased stainability of apoptotic elements in DNA green fluorescence coupled with a higher red fluorescence (which is common to chromatin condensation and higher content of single-stranded DNA).18 Cell debris were excluded from analysis on the basis of their forward light scatter properties. Modified FACScan equipment (Becton Dickinson) was used to measure fluorescence upon excitation at 488 nm. Five thousand cells were measured for each analysis at separate wavelength bands for green/DNA and red/RNA. Samples were analyzed using a Hewlett Packard microcomputer and Becton Dickinson software including Consort 32, Cellfit, and Lysis II.

Statistical analysis.The results are expressed as the mean ± the mean of the standard error (SEM) or the standard deviation (SD), as indicated. Statistical analysis was performed by mean of the nonparametric paired Wilcoxon rank-sum test.

Collection and enrichment of CD34⁺ cells.Sixteen normal donors underwent G-CSF treatment and PBSC collection for allogeneic transplantation. The number of CD34⁺ cells peaked in all but 1 healthy subject at day 5 (data not shown), when PBSC harvests were initiated. Positive selection of CD34⁺ cells was always performed using PBSC samples recovered at day 5. Ten donors agreed to donate BM cells for purification and analysis of hematopoietic progenitors cells before and during G-CSF administration (D0 and D5, respectively). The percentage of CD34⁺ cells in BM (D0 and D5) and PB samples was 2.5% ± 1% (SD), 2.1% ± 1% (SD), and 0.7% ± 0.4% (SD) of the mononuclear cell fraction, respectively. After magnetic separation, the percentage of CD34⁺ in BM was 95% ± 6% (SD) and 96% ± 4% (SD) at D0 and D5, respectively, and in PB was 98.9% ± 0.8% (SD). The mean enrichment factors were 57-fold in BM and 137-fold in PB. The mean overall recovery of CD34⁺ cells was greater than 87% (Table 1). When we analyzed the coexpression on CD34⁺ cells of lineage-associated antigens as well as selected activation molecules, we found that greater than 90% of hematopoietic cells in both compartments coexpressed the CD38 and HLA-DR antigens, whereas there was a significant increase in the proportion of PB myeloid cells (ie, CD34⁺/CD33⁺ and CD34⁺/CD13⁺) at the expense of B-lymphocyte precursors (CD34⁺/CD19⁺), as already reported4 (data not shown).

Frequency and proliferative response of CFU-C to rhCSFs.All of the tested CD34⁺ cell fractions showed significant colony formation when stimulated with PHA-LCM and Epo. The clonogenic efficiency of CFU-GM and BFU-E was 1.4% ± 0.3% (SEM) and 2.5% ± 0.5% (SEM) for PB-derived samples and 1.4% ± 0.4% (SEM) and 2.2% ± 0.3% (SEM) for BM cells, respectively. Thus, no significant difference was noted in the frequency (number of colonies formed per number of cells plated ) of CFU-C in primed PB versus steady-state BM when a conditioned medium was used for colony growth stimulation. We then assessed the trilineage proliferation of PB and baseline BM CD34⁺ cells in response to selected CSFs (Figs 1 and 2). Comparative experiments on colony formation of megakaryocyte precursors (CFU-MK) in the presence of IL-3 and GM-CSF or IL-3, IL-6, and Epo did not show any significant difference between the two compartments (Fig 1). Similar data were obtained when the more immature BFU-MK were assayed (data not shown). Conversely, the results presented in Fig 2 show that priming with G-CSF significantly increased the number of PB CFU-GM stimulated both by G-CSF/SCF and IL-3/SCF (P < .05) and BFU-E cultured with IL-3 and CSF combinations in the presence of Epo (P < .05). The more pronounced effect was seen on IL-3/SCF–responsive CFU-C (1.5- ± 1-fold increase in PB v BM CD34⁺ cells; P < .02). CD34⁺ cells enriched from steady-state BM generated more CFU-GM colonies than did their circulating counterparts when stimulated with G-CSF alone (P < .04). Notably, the addition of SCF to culture media increased additively or synergistically the number and size of CFU-C induced by G-CSF or IL-3 (Fig 2). Taken together, these results suggest a higher responsiveness of mobilized CD34⁺ cells to selected growth factors as compared with steady-state BM cells. However, previous studies from our laboratory showed that pretreatment with G-CSF also modifies the response of BM hematopoietic progenitor cells to subsequent stimulation with additional growth factors.11 Thus, G-CSF seems to exert its priming effect in a similar way on both BM and PB CD34⁺ cells.

Assessment of LTC-IC.The number of LTC-IC at week 5 of culture was 48.2 ± 35 (SEM) and 62.5 ± 54 (SEM) per 10⁴ CD34⁺ cells in PB and BM, respectively (Table 2). Although there was a trend toward higher numbers of LTC-IC in BM samples, this observation did not achieve statistical significance. Overall, we harvested from 16 donors a mean number of 39,883 LTC-IC/kg of recipient body weight (range, 1,800 to 232,058) collected over two daily procedures.

Ara-C assays.At day 5 of G-CSF treatment, the proportion of BM progenitor cells in S-phase was 34.6% ± 11% (SD), as compared with a baseline value of 25.5% ± 12% (SD; P < .05; Table 3). As expected, G-CSF exerted its effect mainly on CFU-GM (52% ± 8% [SD] S-phase cells v 39.8% ± 10% [SD]; P < .05) rather than on BFU-E (Table 3). In sharp contrast, the number of circulating clonogenic precursors in S-phase from the same G-CSF–treated individuals was remarkably lower (4% ± 5% [SD]). This result was highly significant as compared with BM samples before and after G-CSF treatment (P < .001).

Short-term incubation with Ara-C was also used to determine the proportion of very primitive LTC-IC in S-phase. As in more mature progenitors, very few if any (1% ± 3% [SEM]) LTC-IC were in active synthesis of DNA, as compared with 21% ± 8% (SEM) in steady-state BM (P < .001; Table 2). Long-term exposure to Ara-C in the presence of growth factors was then used to evaluate the cycling status of LTC-IC.14 We found that 67% ± 10% (SEM) of very primitive progenitor cells are recruited in cell cycle, with no statistical difference with BM samples (Table 2).

Cell-cycle studies.The experiments reported above show that the great majority of PB CD34⁺ cells are not in S-phase. However, they readily respond to CSFs and progress to mitosis, giving rise to a large number of colonies. Moreover, PB LTC-IC are inhibited by long-term incubation with Ara-C, suggesting that they progress through the cell cycle. Thus, we posed the question as to whether, rather than being deeply quiescent in G₀-phase, circulating CD34⁺ cells might actually be cycling in G₁-phase. To test this hypothesis, we determined the cell-cycle characteristics of BM and PB CD34⁺ cells by the AO flow cytometric technique. The mean results are shown in Table 4 and confirm the negligible proportion of PB CD34⁺ cells in S- and G₂M-phase as compared with their BM counterparts. By contrast, the majority of PB hematopoietic progenitor cells was found in G₁-phase and their percentage was significantly higher than that reported in BM samples before and after G-CSF administration (80% ± 5% [SEM] v 61.9% ± 6% [SEM] and 48% ± 4% [SEM], respectively; P < .05). Notably, the number of PB CD34⁺ cells in G₀-phase was similar to that of BM samples. Moreover, combined DNA/RNA analysis showed that PB CD34⁺ cells have a significantly lower content of RNA than BM cells (Table 4). A representative example of simultaneous DNA/RNA analysis of PB and BM cells from the same healthy donor is presented in Fig 3.

We also used the AO staining to evaluate cellular apoptosis in the study samples (Table 4). A small but consistent and statistically significant difference between BM and PB cells was found (2.4% ± 0.5% [SEM] of circulating CD34⁺ cells were apoptotic compared with 3.7% ± 1% [SEM] in steady-state BM; P < .05). BM CD34⁺ cells assayed at day 5 of G-CSF administration showed the same percentage of apoptotic CD34⁺ cells as mobilized cells, suggesting that the effect on programmed cell death is due to G-CSF by itself rather than to a change of compartment (ie, mobilization into PB).

G-CSF has been shown to mobilize into PB cells with the capacity to reconstitute allogeneic hematopoiesis after a myeloablative conditioning regimen in both a murine model19 and, more recently, in humans.1-3 Transplantation of allogeneic PBSC has also resulted in a more rapid hematopoietic recovery as compared with marrow grafts.1 Whether these effects are solely explained by quantitative changes in the numbers of different types of transplanted progenitors or whether the mobilization procedures induce relevant qualitative changes in the proliferative and differentiative behavior of both committed and primitive precursors is still a matter of investigation. Previous reports have suggested that circulating progenitor cells might be phenotypically and functionally distinct from those in the BM.4,5,20 For instance, a somewhat surprising finding was that few committed PB precursors were in S-phase, whereas their BM counterparts were actively proliferating upon G-CSF priming.5 20 It is a well-recognized fact that transplantation of mobilized stem cells results in a more rapid recovery of hematopoiesis than does reinfusion of BM cells. Moreover, mobilized CD34⁺ cells have generally shown a higher efficiency of retroviral vectors infection, requiring cell cycling for integration, as compared with marrow cells.

To better define the functional and kinetic characteristics of the human stem progenitor cells pool mobilized into PB by G-CSF, we determined (1) the frequency of CFU-C among mobilized CD34⁺ cells and their trilineage proliferative response to rhCSFs, (2) the content of very primitive LTC-IC, and (3) the cell-cycle distribution of PB CD34⁺ cells, including both clonogenic cells and LTC-IC, and the effects of G-CSF treatment on prevention or induction of apoptosis on hematopoietic stem cells. The results were then compared with those observed on BM CD34⁺ cells assessed before and during G-CSF administration. Notably, we studied cell populations that were virtually pure in CD34⁺ cells (mean value, >95%). In addition, the overall recovery of CD34⁺ cells after positive selection was always greater than 85%; thus, the results reported here may be truly representative of the whole progenitor stem cells compartment. It should also be pointed out that, in contrast to previous studies, we did not compare our PBSC donors with random BM donors because of potential bias deriving from donor-to-donor variability with respect to progenitor cell clonogenic efficiency, LTC-IC content, and in vitro response to CSFs. Therefore, we studied highly purified BM and PB CD34⁺ cells from the same healthy individuals.

In our study, G-CSF–primed CD34⁺ cells showed the same clonogenic efficiency as their steady-state BM counterparts when stimulated with a conditioned medium. This is in agreement with previous data generated in a murine model5 but not in humans.4 To further investigate the proliferative potential of mobilized CD34⁺ cells, we assessed their trilineage colony-forming capacity in response to single growth factors or combinations of CSF selected for their ability to stimulate mature and more primitive hematopoietic progenitor cells, respectively.21,22 Colony formation of CD34⁺ megakaryocyte progenitors (CFU-MK) collected from PB was not significantly different from that of BM CD34⁺ cells upon IL-3/GM-CSF and IL-3/IL-6/Epo stimulation. Circulating myeloid and erythroid CFU-C derived from CD34⁺ cells showed the same pattern of response to CSFs as BM cells, in line with reports by others (reviewed in Metcalf21 and Ogawa22 ) and by our own group.23 An early acting growth factor (ie, SCF ) augmented both the size (data not shown) and the number of colonies stimulated by IL-3 and G-CSF (Fig 2). However, mobilized CD34⁺ cells showed a significantly higher clonogenic efficiency in response to cytokines other than G-CSF (ie, IL-3 and SCF ) used alone and in combination. This finding may not be specific to mobilized progenitor cells, because we have already shown that priming with G-CSF increases the proliferative response of BM CD34⁺ cells to additional CSFs, including IL-3 and GM-CSF.11 Therefore, this property of hematopoietic cells seems to be directly due to G-CSF treatment rather than to mobilization into PB.

We then performed comparative studies to further address the issue of kinetic changes of PB and BM CD34⁺ cells upon G-CSF treatment. In line with the findings of Roberts and Metcalf,5 we documented the low percentage of mobilized CD34⁺ cells in S- and G₂M-phase through the use of two different assays. However, studies of cell-cycle distribution showed that the vast majority of circulating CD34⁺ cells are actually cycling, being in G₁-phase. Taken together, these data extend to normal PBSC donors previous observations from nonhuman primates on cell-cycle status and response to CSF of cytokine-mobilized CD34⁺ cells.20 Similar to the present study, the vast majority of circulating CD34⁺ cells were not found to be in active DNA synthesis. However, a greater proportion of PB hematopoietic cells entered the S/G₂M-phase within 72 hours after IL-6 and SCF stimulation.20 Therefore, our results support the view that, rather than being deeply quiescent (G₀ ), as previously suggested by the low CD71 expression and the Rhodamine 123^dull status,24 many of the circulating progenitors are in G₁ phase and might readily enter S-phase, particulary if exposed to certain growth factors.20,25 The PB G₁ cells had a lower RNA content than BM cells and promptly proliferated in response to CSFs. This may indicate that the critical level of chromatin decondensation and nuclei accumulation of nonhistone proteins and RNA before DNA synthesis initiation is reached at a subthreshold level of RNA and protein content, thus resulting in a faster progression through cell cycle. The activity of G-CSF in the competence progression framework of the cell cycle may be similar to that of other early acting growth factors (ie, IL-11 and SCF ) that have recently been shown to shorten G₁-phase.26 The cycling status and the high response in vitro to growth factors may also explain why mobilized PB CD34⁺ cells represent an optimal target for efficient retroviral infection.27 In addition, through kinetic analysis we have been able to show that G-CSF significantly protects circulating CD34⁺ cells from apoptosis, although this effect was not restricted to mobilized hematopoietic progenitor cells and was actually also observed on BM CD34⁺ cells. Thus, at least part of the response to G-CSF can be said to have been due to the increased survival of mobilized precursor populations.28 

A further set of experiments was then designed to investigate the frequency and the cycling status of the most primitive subset of hematopoietic progenitor cells (LTC-IC) functionally identified for their capacity to generate CFU-C after 5 weeks of liquid culture. LTC-IC have recently been proposed as candidate long-term marrow repopulating cells,8 and their presence in PB has been reported after mobilization protocols in cancer patients.12 However, neither the level of LTC-IC nor their kinetic status has been evaluated in PBSC collections from normal donors intended for allogeneic transplantation. The assessment of cell-cycle characteristics of putative hematopoietic stem cells is important in the setting of allograft. In fact, competitive transplant studies in irradiated host mice have recently shown a defective long-term repopulating activity of early BM cells induced to S-phase by cytokines.29 Moreover, there is increasing interest in determining the kinetic status of very primitive subfractions of CD34⁺ cells to improve our capacity to expand hematopoietic stem cells and use viral vectors. The results presented here do not show a significant difference in the frequency of LTC-IC between steady-state BM and primed PB CD34⁺ cells derived from the same normal subjects. Similar observations have been reported in cancer patients treated with G-CSF for 5 to 7 days for mobilization of autologous PBSC,30 and this represents an approximately 100-fold increase compared with steady-state PB.31 We observed a substantial interindividual variation with no relationship to age, sex, and white blood cell count (data not shown). Because PBSC grafts have been shown to contain 3³² to 5 (R.M.L., unpublished observations) more CD34⁺ cells than BM harvests, a similar increase may be assumed for LTC-IC. So far, 10 of our 16 patients have successfully received transplants after thawing PBSC collections, with a mean number of 26,723 PB LTC-IC/kg of recipient body weight (range, 1,800 to 113,739; collected over 2 daily procedures). These figures compare favorably with the number of LTC-IC per kilogram reported for autologous PBSC collections after G-CSF treatment (670 ± 300 [SEM] per single leukapheresis) and allogeneic BM grafts.12 As with the PB committed progenitor cells, very few if any primitive hematopoietic cells were shown to be in S-phase, whereas 21% ± 0.8 % (SEM) of BM LTC-IC were inhibited by Ara-C (P < .02). However, by exposing CD34⁺ cells to Ara-C for 16 to 18 hours in the presence of appropriate CSFs,14 we showed that about two-thirds of LTC-IC are in cell-cycle with no difference with baseline BM samples. Recently, Ponchio et al14 have found that more than 80% of PB LTC-IC are quiescent under steady-state conditions, whereas 85% ± 5% (SEM) of BM LTC-IC are actively cycling. Thus, G-CSF seems to induce the mobilization of a large number of very primitive hematopoietic progenitor cells that share the same kinetic profile of BM cells.

In summary, our results indicate that CD34⁺ cells mobilized into PB of normal donors by glycosylated G-CSF (lenograstim) can be considered equivalent to their BM counterparts, on a cell per cell basis, in their content of both committed and very primitive hematopoietic progenitors capable of repopulating the hematopoietic tissue. Therefore, given their higher cellularity, PBSC collections should contain a substantially greater number of early hematopoietic cells than BM harvests. Kinetic and functional analysis of circulating CD34⁺ cells, studied at various stages of differentiation, shows that the majority of these cell populations are actually cycling, contain a very low number of apoptotic elements, and show an increased proliferative response to CSF stimulation as compared with steady-state BM precursors. However, comparative studies on BM cells before and during G-CSF treatment showed that some of these differences appear to be related directly to G-CSF administration rather than to change of compartment. These studies may have a major clinical significance in view of the increasing evidence that transplantation of allogeneic PBSC represents a feasible alternative to conventional BM graft.