In Vitro Proliferation Potential of AC133 Positive Cells in Peripheral Blood

Authors


Abstract

AC133 antigen is a novel marker for human hematopoietic stem/progenitor cells. In this study, we examined the expression and proliferation potential of AC133+ cells obtained from steady-state peripheral blood (PB). The proportion of AC133+ cells in the CD34+ subpopulation of steady-state PB was significantly lower than that of cord blood (CB), although that of cytokine-mobilized PB was higher than that of CB. The proliferation potential of AC133+CD34+ and AC133CD34+ cells was examined by colony-forming analysis and analysis of long-term culture-initiating cells (LTC-IC). Although the total number of colony-forming cells was essentially the same in the AC133+CD34+ fraction as in the AC133CD34+ fraction, the proportion of LTC-IC was much higher in the AC133+CD34+ fraction. Virtually no LTC-IC were detected in the AC133CD34+ fraction. In addition, the features of the colonies grown from these two fractions were quite different. Approximately 70% of the colonies derived from the AC133+CD34+ fraction were granulocyte-macrophage colonies, whereas more than 90% of the colonies derived from the AC133CD34+ fraction were erythroid colonies. Furthermore, an ex vivo expansion study observed expansion of colony-forming cells only in the AC133+CD34+ population, and not in the AC133CD34+ population. These findings suggest that to isolate primitive hematopoietic cells from steady-state PB, selection by AC133 expression is better than selection by CD34 expression.

Introduction

Effective selection of hematopoietic stem and progenitor cells is important, both for clinical transplantation therapy and basic research. Currently, monoclonal antibodies to the cell surface antigens on hematopoietic cells are routinely used for identification and subsequent isolation of subsets of these cells. Over the past decade, antibodies to CD34 have been commonly used as a marker to select hematopoietic stem and progenitor cells [1, 2]. However, several reports have raised serious questions about CD34 expression by hematopoietic stem cells in the mouse [3-5]. After three years of controversy, Ogawa's group recently showed that CD34 expression in murine hematopoietic stem cells reflects the activation state and this may be so in humans [6]. In addition, there is increasing evidence that human CD34 hematopoietic stem cells may exist and that such cells may be more immature than CD34+ stem cells [3, 7, 8]. Consequently, alternative selection markers to CD34 antigen would be very useful. A novel antigen, AC133 [9-12], is a strong candidate for this purpose.

AC133 antigen is modeled as a 5-transmembrane molecule, a structure that is unique among known cell surface molecules [10]. AC133 antibody was originally identified and isolated from mouse footpads immunized with human fetal liver CD34+ cells. CD34 and AC133 antigens are coexpressed on primitive hematopoietic progenitors and some leukemic cells [12]. AC133 antigen is expressed only in the CD34bright subset of human hematopoietic progenitors, although AC133 expression is not always associated with CD34 expression on leukemia cells [13-15]. A transplantation experiment using fetal sheep recipients demonstrated that AC133+ human bone marrow (BM) cells engraft and home to the fetal sheep BM, and cells harvested from primary sheep recipients successfully engraft secondary recipients [12]. Moreover, SCID-repopulating cells are found exclusively in the AC133+CD34+ fraction of human cord blood (CB) cells [16]. These findings suggest that AC133 antigen selection might overcome the possible problems associated with CD34 antigen expression on immature primitive hematopoietic cells.

AC133 expression analysis has been performed on various hematopoietic cell sources such as fetal liver, BM, CB, and cytokine-mobilized peripheral blood (mPB) cells. The proportion of AC133+ cells varied among these sources of hematopoietic material. Wynter et al. [16] reported that the percentage of AC133+ cells was highest in mPB CD34+ cells, and speculated that AC133 antigen might be expressed on a greater portion of circulating CD34+ cells. However, there is no information on AC133 antigen expression in steady-state peripheral blood (PB). Presumably, the percentage of CD34+ cells in nonmobilized, circulating, steady-state PB is relatively low compared to that in other sources. In this study, we analyzed the biological properties of AC133+ cells obtained from steady-state PB. We show that AC133CD34+ cells consist mainly of erythroid-committed progenitors, and virtually no long-term culture-initiating cells (LTC-ICs) were detected in this fraction. In addition, these AC133CD34+ cells decreased the colony-forming cell (CFC) expansion of CD34+ cells. These results strongly suggest that to isolate HSPC—from steady-state PB at least—AC133+ selection is superior to CD34+ selection.

Materials and Methods

Cell Sources

Buffy coat PB cells were obtained from volunteer blood donors. CB was collected following normal pregnancies and deliveries, after obtaining informed consent from the mothers. All samples were processed within 24 h of collection.

Cell Purification by Magnetic Activated Cell Sorting

Mononuclear cells (MNCs) were isolated from PB and CB using Ficoll-Paque PLUS (Amersham Pharmacia Biotech AB; Uppsala, Sweden; http://www.apbiotech.com) density centrifugation. After washing three times (twice with phosphate-buffered saline [PBS] and once with PBS-bovine serum albumin [BSA] [PBS/0.5% BSA/5 mM EDTA]), the MNCs were resuspended in PBS-BSA at a concentration of 108 cells per 300 μl.

For AC133+ selection, MNCs were subjected to immunomagnetic separation using a magnetic activated cell sorting (MACS) AC133 Cell Isolation Kit (Miltenyi Biotech; Auburn CA; http://www.miltenyibiotec.com), following the manufacturer's recommendations. Briefly, MNCs were incubated for 30 min at 6°C with FcR blocking reagent (human immunoglobulin G [IgG]) and AC133 MicroBeads (microbeads conjugated to monoclonal mouse anti-human AC133 antibody). After washing with PBS-BSA, labeled cells were filtered through a 30-μm nylon mesh and loaded onto a column installed in a magnetic field. Trapped cells were eluted after the column was removed from the magnet. The collected cells were applied to a second column and the purification step was repeated.

For CD34+ selection, MNCs were subjected to immunomagnetic separation using a MACS CD34 Progenitor Cell Isolation Kit (Miltenyi Biotech), following the manufacturer's recommendations. Briefly, MNCs were incubated for 15 min at 6°C with QBEND10 antibody (mouse antihuman CD34+) (Miltenyi Biotech) and human IgG to prevent nonspecific binding. After washing with PBS-BSA, the cells were incubated for 15 min at 6°C with an antimouse antibody coupled to MACS microbeads. Column separation was performed in the same manner as for AC133+ selection, described above. The purity of isolated CD34+ cells was generally greater than 90%, as evaluated by flow cytometry using a FACSCalibur (Becton Dickinson; San Jose, CA).

For AC133CD34+ selection, MNCs were first subjected to immunomagnetic separation using the MACS AC133 Cell Isolation Kit. Cells in the flow-through fraction were collected and applied to a second column after incubation with AC133 MicroBeads once more. Cells in the second flow-through fraction (AC133 cells) were collected and subjected to immunomagnetic separation using the MACS CD34 Progenitor Cell Isolation Kit. Column separation was performed following the procedure described above, and cells (AC133CD34+ cells) trapped in the second MACS CD34 Progenitor Cell Isolation Kit column were collected.

Cell Sorting and Analysis by Fluorescence-Activated Cell Sorter

To isolate the AC133+CD34+ and AC133+CD34+ subsets, MACS-purified CD34 cells were stained with fluorescein isothiocyanate (FITC)-conjugated anti-CD34 (HPCA-2-FITC; Becton Dickinson) and phycoerythrin (PE)-conjugated anti-AC133 for 60 min on ice in the dark. After incubation, the cells were washed once in PBS. Cell sorting was performed on a FACSVantage (Becton Dickinson) equipped with an argon laser tuned to 488 nm. The setting of the sort windows produced the results shown in Figure 2.

Figure Figure 2..

FACS profiles of AC133+and AC133populations from MACS-purified PB CD34+cells.MACS-purified PB CD34+cells were stained with FITC-conjugated anti-CD34 and PE-conjugated anti-AC133 antibodies as described inMaterials and Methods. The AC133+CD34+and AC133CD34+populations are seen in regions AC133+and AC133, respectively.

The expression of AC133 antigen on CD34+ cells from PB or CB was investigated by dual-color immunolabeling with FITC-labeled anti-CD34 (HPCA2-FITC; Becton Dickinson) and PE-labeled anti-AC133 (Miltenyi Biotech) for 60 min on ice in the dark. The cells were then washed with PBS and analyzed on a FACSCalibur flow cytometer (Becton Dickinson) using CellQuest software (Becton Dickinson). The kinetics of AC133 expression on AC133+ cells and MACS-purified AC133CD34+ cells were investigated using PE-labeled anti-AC133/2 (Miltenyi Biotech), which recognizes a different epitope of AC133 antigen from that recognized by anti-AC133.

To detect apoptosis, cells were tested using an Early Apoptosis Detection Kit-FITC (Kamiya Biochemical Company; Seattle WA; http://www.kamiyabiomedical.com) according to the manufacturer's directions. Briefly, 105 ∼ 106 cells were suspended per ml of ice-cold PBS/0.5% Tween 20. The cell suspension was incubated with annexin V-FITC [17] (final concentration 0.25 μg/ml) and propidium iodide (final concentration 2.5 μg/ml) on ice for 10 min in the dark. Stained cells were analyzed on a FACSCalibur flow cytometer using a single laser emitting extinction light at 488 nm.

CFC Assay

Test cells were suspended in Iscove's modified Dulbecco's medium (IMDM) containing 2% fetal bovine serum (FBS) at a concentration of 1-5 × 103 cells per ml. The cell suspensions were mixed with 10 volumes of methylcellulose-based semisolid culture medium (MethoCult GF H4434V; StemCell Technologies Inc.; Vancouver, British Columbia, Canada; http://www.stemcell.com). The medium contained 30% FBS, 1% BSA, 10–4 M 2-mercaptoethanol, 2 mM L-glutamine, 3 U/ml recombinant human (rh) erythropoietin, optimized concentration of rh stem cell factor, rhGM-CSF, rh interleukin 3 (IL-3), and rhG-CSF, and 0.9% methylcellulose (4,000 cps) in IMDM. Aliquots of the mixtures (1.1 ml, 100-500 cells) were plated in duplicate in 35-mm dishes and incubated for 14 days in a humidified atmosphere with 5% CO2 at 37°C. Each plate was scored for erythroid (BFU-E), myeloid (colony-forming units-granulocyte-macrophage [CFU-GM]), and multi-lineage colonies (CFU-mixture [CFU-Mix]).

LTC-IC Assay

The frequency of LTC-IC at week 5 was determined using a limiting dilution assay. Test cells were cultured with stromal layers of MS-5 cells (established by Dr. K. J. Mori, Niigata University, Japan) [18] in 96-well plates at six different concentrations with 10 replicates per dilution. The cells were cultured in 0.2 ml of MyeloCult H5100 medium (StemCell Technologies, Inc.) for five weeks. Cultures were maintained in a humidified atmosphere with 5% CO2 at 37°C, with weekly changes of half the medium. At the end of the five-week LTC-IC assay period, the medium was replaced by 0.1 ml of semisolid CFC assay medium (MethoCult GF H4434V; StemCell Technologies, Inc.). After 14-day incubation, the wells were scored as positive or negative for the presence of colonies. The LTC-IC frequencies were calculated using Poisson statistics.

Ex Vivo Expansion

Cell cultures were initiated with 1 or 2 × 104 cells in nontreated 24-well plates in 1 ml of StemPro-34 serum-free medium (GIBCO/BRL; Grand Island, NY) supplemented with 2 mM L-glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin, 50 ng/ml Flt3 ligand (FL) (DIACLONE Research; Besançon Cedex, France), and 10 ng/ml thrombopoietin (TPO) (Kirin Brewery Co. Ltd.; Maebashi, Japan; http://www1.kirin.co.jp/english/r_d/pha/index.html). Cultures were maintained in a humidified atmosphere at 37°C and 5% CO2. The two cytokines were added to each series of microwells twice a week, and half the medium with the growth factors was changed on day 7. On day 14, cells were collected by pipetting, and viable cells were counted by trypan blue exclusion. Suitable aliquots of the cell suspension were assayed for CFC and LTC-IC.

Statistical Analysis

All the results were expressed as the mean of data obtained from three or more separate experiments ± 1 standard error (SE). Significance levels were determined using a two-sided Student's t-test.

Results

Hematopoietic Activity of AC133+ Cells and CD34+ Cells in Steady-State PB

CD34+ cells and AC133+ cells were isolated from PB mononuclear cells (PBMNC), using MACS. The average percentages of recovered CD34+ cells and AC133+ cells per total PBMNC were 0.057 ± 0.029% (0.031 ∼ 0.133%, n = 13) and 0.035 ± 0.011% (0.023 ∼ 0.057%, n = 6), respectively, all significantly different by the t-test (p < 0.05). To analyze the hematopoietic activity of these two cell populations, the clonogenic capacity, frequency of LTC-IC, and ex vivo expansion ability were examined.

The results of the colony-forming capacities of CD34+ cells and AC133+ cells are summarized in Table 1. The frequency of total CFC was almost the same in both fractions. However, the percentage of BFU-E was significantly higher in the CD34+ population than in the AC133+ population (p < 0.001). Reciprocally, the incidence of CFU-GM was higher in the AC133+ population than in the CD34+ population (p < 0.001). Although no significant difference was observed in the incidence of CFU-Mix, it seemed slightly higher in CD34+ cells.

Table Table 1.. Clonogenic potential and LTC-IC frequency of the subpopulations of steady-state PB
  LTC-IC
  Frequency
Fractionated populationTotalnof coloniesBFU-ECFU-MixCFU-GM 
  1. a

    In the CFC assay, 100 fractionated cells were cultured, as described in the Materials and Methods section. In the LTC-IC assay, 320, 160, 80, 40, 20, and 10 AC133+ cells and AC133+CD34+ cells/well, 640, 320, 160, 80, 40, and 20 CD34+ cells/well and 1,280, 640, 320, 160, 80, and 40 AC133CD34+ cells/well were cultured, as described in the Materials and Methods section.

  2. b

    For each assay on MACS-purified subpopulations, the number of experiments was as follows: CFC/AC133+ (n = 7), LTC-IC/AC133+ (n = 5), CFC/CD34+ (n = 13), LTC-IC/CD34+ (n = 9). In the case of FACS-sorted subpopulations, all data are shown as the mean ± SD for four experiments.

  3. c

    LTC-IC = long-term culture-initiating cells; PB = peripheral blood; CFC = colony-forming cells; CFU-Mix = colony-forming units-mixture; CFU-GM = colony-forming units-granulocyte-macrophage; MACS = magnetic activated cell sorting; FACS = fluorescence-activated cell sorting; NS = not significant (i.e., p > 0.05). The p value was calculated using Student's t-test.

MACS-purified     
AC133+45 ± 1736 ± 52 ± 262 ± 61/160 ± 43
CD34+49 ± 1662 ± 74 ± 234 ± 81/300 ± 100
p valueNS<0.001NS<0.001<0.01
AC133+CD34+53 ± 1427 ± 44 ± 269 ± 31/210 ± 54
AC133CD34+79 ± 1294 ± 31 ± 15 ± 21/>1,280
p valueNS<0.001NS<0.001

Next, to quantify the frequency of more primitive progenitor cells, a limiting dilution analysis by LTC-IC assay was performed on the CD34+ and AC133+ cells. As shown in Table 1, primitive hematopoietic cells were detected more frequently by LTC-IC at week 5 in AC133+ cells (1/160) than in CD34+ cells (1/300) (p < 0.01).

Finally, we compared the ex vivo expansion ability of CD34+ cells and AC133+ cells. Both cell populations were incubated in STEMPRO-34 serum-free medium. We supplied FL and TPO to the medium as growth factors to facilitate the proliferation of immature hematopoietic cells as reported by Piacibello et al. [19]. After two weeks of culture, the numbers of CFC and viable cells were determined and compared with those before culture (Table 2). The total number of cells increased 14-fold and 23-fold, in the CD34+ and AC133+ cell populations, respectively, but the difference between these two populations was not significant. However, the expansion of the CFC was significantly more pronounced in the AC133+ population (7.1-fold) than in the CD34+ population (2.9-fold) (p < 0.01).

Table Table 2.. Total cells and CFC increase of the subpopulations of steady-state PB
  
Fractionated populationViable cellsCFCnof experiments
  1. a

    Data are given as the mean relative increase ± SD for a 14-day culture.

  2. b

    CFC = colony-forming cells; MACS = magnetic-activated cell sorting; FACS = fluorescence-activated cell sorting; NS = not significant (i.e., p > 0.05).

MACS-purified   
AC133+23 ± 137.1 ± 2.4 7
CD34+14 ± 72.9 ± 1.213
p value by t-testNS<0.01 
FACS-sorted   
AC133+CD34+15.0 ± 4.73.3 ± 1.1 3
AC133CD34+7.9 ± 3.6<0.1 3
p value by t-test<0.05<0.05 

Expression of AC133 Antigen on Steady-State PB Cells

We analyzed AC133 antigen expression on steady-state PB cells and compared it with the expression on CB cells. We measured AC133 antigen expression after purifying CD34+ cells from PB and CB to ensure accurate results, since CD34+ hematopoietic stem/progenitor cells selectively express AC133 antigen, and the percentage of CD34+ cells in steady-state PB has been estimated to be very low [20]. Representative results of the two-color analysis are shown in Figure 1A. The expression of AC133 antigen on CD34+ cells ranged from 41% to 73% on cells from steady-state PB, whereas on cells from CB it ranged from 68% to 95%. The proportion of AC133+ cells in the CD34+ cells was lower in steady-state PB (54.6 ± 8.6%) than in CB (82.6 ± 8.6%), both significantly different by the t-test (p < 0.001) (Fig. 1B).

Figure Figure 1..

FACS analysis of AC133 antigen expression on CD34+cells from PB and CB.(A) Representative FACS profiles of AC133 antigen expression on CD34+cells from PB and CB. MACS-purified CD34+cells from PB (upper) and CB (lower) were stained with FITC-conjugated anti-CD34 and PE-conjugated anti-AC133 antibodies as described inMaterials and Methods. Quadrant markers are defined by FITC and PE fluorescence of isotype controls. (B) The percentages of CD34+cells from PB (left, n = 12) and CB (right, n = 18) expressing AC133 antigen. The horizontal lines indicate mean AC133 antigen expression.

Hematopoietic Activity of AC133+ and AC133 Cell Populations in PB CD34+ Cells

We compare the hematopoietic activities of AC133+ and CD34+ cells in Tables 1 and 2, Table 2.. About half (54.6 ± 8.6%) the CD34+ cells express AC133 antigen in steady-state PB (Fig. 1). Therefore, to more precisely elucidate the characteristics of AC133+ cells, it is better to compare the hematopoietic properties of AC133+CD34+ and AC133CD34+ cells. The AC133+ and AC133 cells in the CD34+ population were sorted as shown in Figure 2, and their hematopoietic capacity was analyzed. Hereafter, most data compare AC133+CD34+ and AC133CD34+ cells. In the experiments shown in Figures 3 and 4, Figure 4., we utilized MACS-purified AC133+ cells instead of AC133+CD34+ cells, because the numbers of cells and fluorescence available for FACS purification were limited. However, we consider that the kinetics of AC133+ cells must be similar to those of AC133+CD34+ cells, because the vast majority (>99%) of PB AC133+ cells express CD34 antigen.

Figure Figure 3..

Kinetics of the apoptosis of AC133+cells and AC133CD34+cells from steady-state peripheral blood (PB) during ex vivo culture.Magnetic-activated cell sorting (MACS)-purified PB AC133+(upper) and AC133CD34+(lower) cells were cultured in StemPro-34 serum-free medium supplement with FL and TPO. Apoptotic cell death was evaluated by comparing annexin V and propidium iodine staining after 0, 3, 6, 9, and 12 days of incubation. The ratio of apoptotic cells at each time is given in the middle of the right-hand side of each panel.

Figure Figure 4..

Kinetics of AC133 and CD34 antigen expression of AC133+cells and AC133CD34+cells from steady-state PB during ex vivo culture.MACS-purified PB AC133+(upper) and AC133CD34+(lower) cells were cultured in StemPro-34 serum-free medium supplement with FL and TPO. The expression of AC133 and CD34 antigen was evaluated by dual color immunolabeling with FITC-labeled anti-CD34 (HPCA2-FITC) and PE-labeled anti-AC133/2 after 0, 3, and 6 days of incubation. Quadrant makers are defined by FITC and PE fluorescence of isotype controls.

The numbers of CFC and the frequencies of LTC-IC are shown in Table 1 (lower column). The total numbers of CFC from 100 sorted cells tended to be higher in the AC133CD34+ population than in the AC133+CD34+ population, although the difference was not significant. The percentages of BFU-E, CFU-Mix, and CFU-GM in AC133+CD34+ and AC133CD34+ cells are compatible with those of AC133+ and CD34+ cells shown in Table 1 (upper column). More than 90% of the colonies formed by the AC133CD34+ population were erythroid colonies. The percentage of GM colonies was significantly higher in the AC133+CD34+ population than in the AC133CD34+ population (p < 0.001). Notably, significant numbers of LTC-IC were detected in the AC133+CD34+ population, while cells were not detected in the AC133CD34+ population.

Next, we compared the ex vivo expansion ability of AC133+CD34+ and AC133CD34+ cells. They were cultured for two weeks using the same procedure used for the AC133+ and CD34+ cells in Table 2 (upper column). In the first five days of the culture, about twice as many AC133CD34+ cells proliferated as AC133+CD34+ cells (data not shown). However, a considerable portion of these cells died by apoptosis after six days of culture, as shown in Figure 3. The percentage of apoptotic cells indicated by the doubly stained fraction was significantly higher when AC133CD34+ cells were used (Fig. 3). The percentage of apoptotic cells in the AC133CD34+ population was twice as high as that in the AC133+CD34+ population at day 12. In all, the number of AC133CD34+ cells increased 7.9-fold, which was only about half the increase in AC133+CD34+ cells, as shown in Table 2 (lower column). In contrast, the number of CFC from AC133CD34+ cells was greatly diminished during the culture, whereas the number of CFC from the AC133+CD34+ population significantly increased. Thus, in steady-state PB, the AC133+CD34+ population contained a higher percentage of primitive hematopoietic cells than the AC133CD34+ population, which is similar to the higher percentage of primitive hematopoietic cells in the AC133+ population compared to the CD34+ population, as described above (Table 1).

Kinetics of AC133 Antigen Expression on Steady-State PB During Ex Vivo Culture

We examined the kinetics of both AC133 and CD34 antigen expression in ex vivo cultures from MACS-purified AC133+ and AC133CD34+ PB cells. As shown in Figure 4, more than half of the AC133+ cells lost their AC133 antigen as early as day 3 of the culture. However, no changes were observed in CD34 antigen expression and a significant decrease in CD34 antigen was not observed until day 6. Thus, it is clear that the loss of AC133 antigen preceded that of CD34 antigen in the AC133+ PB cell culture. Meanwhile, PB AC133CD34+ cells stopped expressing CD34 antigen before AC133+ cells. Taken together, it is reasonable to consider the AC133+ population more primitive than the AC133CD34+ population in steady-state PB.

Discussion

An average of 3.5 × 105 CD34+ cells were purified from a single bag of buffy coat (400 ml donation) (n = 13). During the two-week-culture, the CFC from these CD34+ cells increased up to 2.9-fold ex vivo, as shown in Table 2. Although an average of 1.9 × 105 purified AC133+ cells were obtained per bag of buffy coat (n = 6), their CFC increased up to 7.1-fold under the same culture conditions. That is, in terms of the absolute number of CFC, more CFC emerged from the AC133+ population than from the CD34+ population after ex vivo expansion. At day 3 of the culture of AC133CD34+ cells, a transient increase of the AC133 antigen expression occurred and the apoptosis of the culture of AC133CD34+ cells was lower than the culture of AC133+CD34+. However, still a significant degree of apoptotic cell death was observed in the culture of AC133CD34+ population after six days (Figs. 3 and 4, Figure 4.). Since the AC133CD34+ cells did not contribute to CFC expansion (Table 2), some factor(s) produced by these apoptotic cells exerted a deleterious influence on the expansion of the other hematopoietic cells. LTC-ICs, the most primitive human hematopoietic cells that can be assayed in vitro, were highly enriched in the AC133+CD34+ population but could not be detected in AC133CD34+ population (Table 1). The kinetics of AC133 antigen expression also support the notion that in steady-state PB the AC133+ population is more primitive than the AC133CD34+ population (Fig. 4). Further, FACS analysis revealed that CD38 cells and Thy 1 weak-positive cells were observed almost exclusively in the AC133+ population (data not shown).

In this report, we clearly show that the vast majority of AC133CD34+ PB cells are erythroid-committed progenitors, as more than 90% of the colonies produced from AC133CD34+ PB cells were pure erythroid colonies (Table 1). This tendency was reported previously in a study of CB by Yin et al. [12]. However, the proportion of erythroid colonies produced by AC133CD34+ CB cells was much lower (51%) than in our results. Wynter et al. [16] reported that the proportion of erythroid colonies produced by AC133CD34+ CB cells was 63%. They could detect the enrichment of erythroid precursors (BFU-E) in this cell population in a serum-free assay but not in their standard clonogenic assays containing 30% (v/v) fetal calf serum. We believe it unlikely that these discrepancies are derived from differences in the hematopoietic cell sources. Instead, we attribute them to the methods of cell sorting or the clonogenic assay used, for the following reasons: The fraction of erythroid colonies produced in a clonogenic analysis of AC133CD34+ CB cells (average 96%, n = 4) was almost the same as that produced from PB cells in our study. That is, more than 90% of the colonies developing from the AC133CD34+ population of both cell sources were erythroid colonies. However, the frequency of total CFC was much lower in previous reports than in our study. Yin et al. [12] reported that the frequencies of clonogenic cells in AC133+CD34+ CB cells and AC133CD34+ CB cells were 25% (124/500) and 12% (59/500), respectively, while Wynter et al. [16] obtained values of 9.9% (149/1,500) and 4.9% (73/1,500), respectively, whereas the frequencies in AC133+CD34+ PB cells and AC133CD34+ PB cells were as high as 53% and 79%, respectively, in our study (Table 1). Furthermore, the frequencies of clonogenic cells in AC133+CD34+ CB and AC133CD34+ CB cells were as high as 57% and 61% (n = 4), respectively, in our clonogenic assay system. We are confident that our results are more reliable than those of previous studies as higher clonogenicity generally reflects a better experimental system.

Two groups [12, 16] have also reported data on AC133 antigen expression by CD34+ cells from various hematopoietic sources, and their results are summarized in Table 3, together with our data. All three studies analyzed AC133 antigen expression on CD34+ cells from CB and mPB; however, there are large discrepancies in the data. It is possible that the discrepancies arise from differences in cell preparation methods and sample numbers, as higher percentages of AC133+ cells tend to be obtained from highly purified samples, and it is thought that the more samples analyzed, the more accurate the results. In any case, the proportion of AC133+ cells in the CD34+ population in the steady-state PB seems to be significantly lower than in other sources. In other words, the proportion of AC133CD34+ cells that is highly enriched with erythroid progenitors is higher in steady-state PB than in other hematopoietic sources. Consequently, we feel that the points raised by Wynter et al. [16] should be reconsidered. They argued that AC133 antigen expression is closely connected to the circulation of hematopoietic cells. As they pointed out, AC133 expression on CD34+ cells of mPB, which were mobilized from BM by administration of G-CSF, was significantly higher than on cells from BM. However, the AC133 expression on PB cells raised by G-CSF mobilization presumably declined gradually to a steady state after administration of G-CSF. We speculate that AC133 expression correlates with the long-term proliferative activity of the hematopoietic cells. This speculation is very consistent with our data and with the differences in AC133 antigen expression between BM and mPB or between mPB and steady-state PB. Elucidating the functional characteristics of AC133 antigen will provide important clues to help solve this problem. Conversely, a detailed investigation of the changes in AC133 antigen expression should open the way to determining the function of the AC133 antigen.

Table Table 3.. Comparison of data for AC133 antigen expression on CD34+ cells
ReportsCell sourcesBMmPBCBssPB
  1. a

    The numbers of samples are shown in parentheses.

  2. b

    BM = bone marrow; mPB = mobilized peripheral blood; CB = cord blood; ssPB = steady-state peripheral blood; NT = not tested.

Yin et al.MNCs70% (2)55% (3)35% (2)NT
Wynter et al.Unfractionated36% (7)75% (13)51% (9)NT
Matsumoto et al.CD34+ cells48% (2)89% (4)83% (18)55% (12)

Acknowledgements

We would like to thank K. Nakamura (Osaka University) for operating the FACS, T. Kimura (Kyoto Prefectural University of Medicine) for advice on LTC-IC assay protocols, C. Terada for assistance with the statistical analysis, K. J. Mori (Niigata University) for the generous gift of MS-5 cells, and Kirin Brewery for providing TPO.

Ancillary