CD34 Antigen and the MPL Receptor Expression Defines a Novel Class of Human Cord Blood-Derived Primitive Hematopoietic Stem Cells

Abstract

In the murine hematopoietic stem cell (HSC) compartment, thrombopoietin (THPO)/MPL (THPO receptor) signaling plays an important role in the maintenance of adult quiescent HSCs. However, the role of THPO/MPL signaling in the human primitive HSC compartment has not yet been elucidated. We have identified very primitive human cord blood (CB)-derived CD34^– severe combined immunodeficiency (SCID)-repopulating cells (SRCs) using the intra-bone marrow injection method. In this study, we investigated the roles of the MPL expression in the human primitive HSC compartment. The SRC activities of the highly purified CB-derived 18Lin^–CD34^+/–MPL^+/– cells were analyzed using NOG mice. In the primary recipient mice, nearly all mice that received CD34^+/–MPL^+/– cells were repopulated with human CD45⁺ cells. Nearly all of these mice that received CD34⁺MPL^+/– and CD34^–MPL^– cells showed a secondary repopulation. Interestingly, the secondary recipient mice that received CD34^+/–MPL^– cells showed a distinct tertiary repopulation. These results clearly indicate that the CD34^+/– SRCs not expressing MPL sustain a long-term (LT) (>1 year) human cell repopulation in NOG mice. Moreover, CD34^– SRCs generate CD34⁺CD38^–CD90⁺ SRCs in vitro and in vivo. These findings provide a new concept that CD34^–MPL^– SRCs reside at the apex of the human HSC hierarchy.

Introduction

In the murine primitive hematopoietic stem cell (HSC) compartment, thrombopoietin (THPO)/myeloproliferative leukemia virus (MPL) (THPO receptor) signaling plays an important role in the maintenance of adult quiescent HSCs in the osteoblastic niche¹ and in the posttransplantation HSC expansion². Lineage-negative, stem cells antigen-1-positive (Sca-1⁺), c-kit⁺ (LSK), CD34^–, MPL⁺ (LSKCD34^–MPL⁺), and LSKCD34⁺MPL⁺ cells were reported to be long-term (LT) repopulating HSCs¹. Additionally, LSKCD34^–MPL⁺ and LSKCD34⁺MPL⁺ cells have been regarded to be quiescent and active HSCs, respectively¹. However, the role of THPO/MPL signaling in the human primitive HSC compartment has not yet been clearly elucidated.

The findings obtained in mouse HSC biology cannot always be extrapolated to human HSC biology. For instance, the surface immunophenotype of primitive HSCs differs among species. It was previously reported that murine LT lymphohematopoietic reconstituting HSCs are Lin^–c-kit⁺Sca-1⁺CD34^lo/- (CD34^lo/- KSL) cells³. Moreover, single murine CD34^lo/- KSL cells showed a LT repopulation in recipient mice. We previously identified human cord blood (CB)-derived CD34^– severe combined immunodeficiency (SCID)-repopulating cells (CD34^– SRCs) using the intra-bone marrow injection (IBMI) method⁴. Our identified CD34^– SRCs appear to be a counterpart of murine CD34^lo/- KSL cells. A series of our previous studies demonstrated that these human CB-derived CD34^– SRCs appear to be very primitive LT-repopulating HSCs, which reside at the apex of the human HSC hierarchy^4–11. As we reported previously, the immunophenotype of human CB-derived CD34^– SRCs is Lin^–CD34^–c-kit^–FLT3 Fms-like tyrosine kinase 3 (FLT3)⁶. Our data were consistent with reported studies showing that human bone marrow (BM)- or CB-derived LT-repopulating HSCs were c-kit^low or c-kit^<low cells^12,13. Collectively, the c-kit expression pattern appears to differ between mouse and human HSCs.

From another perspective, Kiel et al. reported that the primitiveness of hematopoietic stem/progenitor cells (HSPCs) could be predicted according to the expression patterns of the signaling lymphocyte activation molecule (SLAM) family members, including CD150, CD244, and CD48¹⁴. Additionally, LT-repopulating HSCs were highly purified as CD150⁺CD244^–CD48^– cells. However, it was reported that human and rhesus macaque HSCs cannot be purified using only the SLAM family receptors¹⁵. Conversely, anti-activated leukocyte cell adhesion molecule (CD166) is a functional HSC marker that identifies both murine and human LT-repopulating cells¹⁶. Both murine LSKCD48^–CD166⁺CD150⁺ and LSKCD48^–CD166⁺CD150⁺CD9⁺ cells, as well as human Lin^–CD34⁺CD38^–CD49f⁺CD166⁺ cells, sustained significantly higher levels of chimerism in primary and secondary recipients than CD166^– cells. Collectively, these results suggest that the functional significance of receptors/molecules expressed on the surface of primitive HSCs may be valuable among species. Therefore, the reported hematopoietic properties of mouse HSCs concerning the expression of MPL receptors have been validated in the human HSC hierarchy.

In this study, we investigated the functional significance of the MPL expression in the human primitive HSC compartment. Our serial transplantation analyses clearly indicated that both the CD34^+/– SRCs not expressing MPL receptor sustained a LT (>1 year) human cell repopulation in NOG mice. Moreover, CD34^– SRCs generate CD34⁺CD38^–CD90⁺ SRCs in vitro and in vivo. Conversely, CD34^–MPL⁺ SRCs and CD34⁺MPL⁺ SRCs are short-term (ST; <6 months) and intermediate-term (IT; 6–12 months) repopulating SRCs, respectively. These findings suggest that the expression pattern of CD34 antigen and MPL receptor defines three classes of human CB-derived HSCs, including ST-, IT-, and LT-repopulating HSCs, and also that CD34^–MPL^– SRCs (HSCs) reside at the apex of the human HSC hierarchy. In addition, these findings suggest that the functional significance of the MPL expression in the human primitive HSC compartment may differ from that in murine primitive HSC compartment.

Materials and Methods

CB Samples

CB samples from normal full-term deliveries were obtained with informed consent and approved by the institutional review board (IRB) of Kansai Medical University.

Isolation of CB-Derived 18Lin^–CD34^+/–MPL^+/–Cells and the Analysis of the Expression of Flt3 on these Cells

CB samples from multiple donors were pooled, and then lineage-negative (Lin^–) cells were enriched using an EasySep^™ Human Progenitor Cell Enrichment Kit (STEM CELL Technologies, Vancouver, Canada) and RoboSep (STEMCELL Technologies), as previously reported^10,11,17. These immunomagnetically separated Lin^– cells were stained with a cocktail of fluorochrome-conjugated monoclonal antibodies (mAbs) including a mixture of anti-18 lineage-specific markers (18Lin)^9,10, including CD2 (DAKO, Kyoto, Japan), CD3 (Beckman Coulter, Fullerton, CA, USA), CD4 (eBioscience, San Diego, CA, USA), CD7 (Beckman Coulter), CD10 (Beckman Coulter), CD11b (Beckman Coulter), CD14 (eBioscience), CD16 (DAKO), CD19 (BD Biosciences, San Jose, CA, USA), CD20 (Beckman Coulter), CD24 (DAKO), CD33 (eBioscience), CD41 (Beckman Coulter), CD45RA (Southern Biotech, Cambridge, UK), CD56 (BD Biosciences), CD66c (Beckman Coulter), CD127 (eBioscience) and CD235a (DAKO), and anti-MPL (CD110, clone 1.6.1) (BD Biosciences), which could accurately recognize the MPL receptor¹⁸, anti-CD34 (BD Biosciences), and anti-CD45 mAb (BioLegend, San Diego, CA) (listed in Table 1) for 30 min on ice. The cells were washed once with phosphate-buffered saline^– (PBS^–; Nacalai Tesque, Kyoto, Japan) containing 2% fetal calf serum (FCS; PBS^–/FCS; Biofill, Elsternwick Victoria, Australia). The dead cells were stained with 7-amino-actinomycin D (7-AAD; Beckman Coulter). Eighteen Lin^–CD34^+/–MPL^+/–cells were sorted and collected by FACSAriaIII (BD Biosciences). The CD34⁺ and CD34^– fractions were defined as described in our previous report¹⁰. The MPL⁺ fractions contain cells expressing maximum phycoerythrin (PE) with fluorescence intensity (FI) of 15% of that of anti-MPL mAb. MPL^– fractions were defined based on the fluorescence minus one (FMO) control.

Table 1.

List of Antibodies Used in This Study

Antibody	Flurochrome	Distributor	Clone	Dilution
CD2	FITC	DAKO	MT910	1:25
CD3	FITC	Beckman Coulter	UCHT1	1:25
CD4	FITC	eBioscience	RPA-T4	1:25
CD4	PE	eBioscience	RPA-T4	1:100
CD7	FITC	Beckman Coulter	8H8.1	1:25
CD8a	APC	eBioscience	RPA-T8	1:100
CD10	FITC	Beckman Coulter	ALB1	1:25
CD11b	FITC	Beckman Coulter	Bear1	1:25
CD14	FITC	eBioscience	61D3	1:25
CD14	PE	Beckman Coulter	RMO2	1:25
CD16	FITC	DAKO	DJ130c	1:25
CD19	FITC	BD Biosciences	HIB19	1:25
CD20	FITC	Beckman Coulter	H299	1:25
CD24	FITC	DAKO	SN3	1:25
CD33	FITC	eBioscience	HIM3—4	1:25
CD33	PE	Beckman Coulter	D3HL60.251	1:50
CD34	APC	BD Biosciences	581	3:50
CD38	PE-Cy7	BioLegend	HIT2	1:100
CD41	FITC	Beckman Coulter	P2	1:25
CD41	PE	Beckman Coulter	P2	1:50
CD45	Pacific Blue	BioLegend	HI30	1:25
CD45RA	FITC	SouthernBiotech	MEM 56	1:50
CD56	FITC	BD Biosciences	NCAM16.2	1:25
CD66c	FITC	Beckman Coulter	KOR-SA3544	1:25
CD90	PE	BD Biosciences	5E10	1:25
CD110	PE	BD Biosciences	1.6.1	1:10
CD127	FITC	eBioscience	eBioRDR5	1:25
CD135	PE-Cy7	BioLegend	BV10A4H2	1:25
CD235a	FITC	DAKO	JC159	1:25
Anti-mouse CD45.1	PE-Cy7	Beckman Coulter	A20	1:200

In a separate experiment, the 18Lin^– cells were stained with a mixture of mAbs plus anti-Flt3 mAb (BioLegend). Then the expression pattern of Flt3 on 18Lin^–CD34^+/–MPL^+/– cells was analyzed using a FACSCantoII (BD Biosciences).

Clonal Cell Culture

Human colony-forming cells (CFCs), including megakaryocyte-containing CFCs, were assayed using our standard methylcellulose cultures, as previously reported^{4,6,9–11,19}.

Briefly, sorted 18Lin^–CD34^+/–MPL^+/– cells were plated at 200 cells per 35-mm Lux suspension culture dishes (Corning Inc., Corning, N Y, USA) in 1 ml of culture containing 1.2% of 1,500 centipoise methylcellulose (Shinetsu Chemical, Tokyo, Japan), 30% FCS (Biofill), 1% bovine serum albumin (BSA; Sigma-Aldrich, St. Louis, MO, USA), 5× 10⁻⁵ mol/L mercaptoethanol (Sigma-Aldrich), and six cytokines [THPO, stem cell factor (SCF; R&D Systems, Minneapolis, MN, USA), interleukin-3 (IL-3), granulocyte macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), and erythropoietin (Epo)]. THPO, IL-3, G-CSF, GM-CSF, and Epo were provided by Kyowa Hakko Kirin Company (Tokyo, Japan). For megakaryocyte colony formation, 10% platelet-poor plasma (PPP) was added instead of 30% FCS to the methylcellulose cultures in the presence of three cytokines (THPO, IL-3, and Epo). Dishes were incubated at 37°C in a fully humidified atmosphere flushed with a combination of 5% CO₂, 5% O₂, and 90% N₂. On day 10 or 14 of cultures, all colonies were scored under an inverted microscope according to their typical morphologic appearance as previously reported^{4,6,9–11,19}. Colony types identified in situ were CFU-GM (colony-forming unit granulocyte macrophage), BFU-E (burst-forming unit erythroid), CFU-Meg (megakaryocyte), CFU-Mix (erythrocyte-containing mixed), and CFU-EM (erythrocyte/megakaryocyte mixed).

In Vitro Lineage Differentiation Assay

Eighteen Lin^–CD34^+/–MPL^+/– cells were cocultured with human BM CD45^–Lin^–CD271⁺SSEA-4⁺ fraction-derived mesenchymal stromal cells (DP-MSCs) as previously reported¹⁷. A total of 1,000 cells were plated into each well of a 24-well culture plate (Corning Inc.). The cells were cultured in StemPro-34 serum-free medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with a cocktail of cytokines [100 ng/ml THPO, 50 ng/ml SCF, 50 ng/ml Flt-3 ligand (FL), 10 ng/ml G-CSF, 10 ng/ml IL-3, and 10 ng/ml IL-6]. After 7 days of coculture, the cells were collected by vigorous pipetting. The numbers of collected cells were counted, and then the cells were stained with mAbs including lineage markers (CD11b, CD14, CD33, and CD41), CD34, and CD45. The percentages of lineage marker-positive and CD34⁺ cells in CD45⁺ cells were analyzed by flow cytometry (FCM; FACSCantoII; BD Biosciences).

Mice

Female six-week-old NOD/Shi-scid/IL-2Rγ_c^null (NOG) mice²⁰ were purchased from the Central Institute of Experimental Animals (Kawasaki, Japan). The mice were maintained and handled under specific pathogen-free conditions. All animal experiments were approved by the Animal Care Committees of Kansai Medical University.

SRC Assay

Isolated 18Lin^–CD34^+/–MPL^+/– cells were injected into the left tibiae of sublethally irradiated (250 cGy) NOG mice using an IBMI technique⁴. The numbers of transplanted cells are described in Table 2. At weeks 5, 12, and 18, a small amount of BM cells were collected from the right tibiae of the mice using a BM aspiration technique. The collected BM cells were stained with a cocktail of mAbs [mouse CD45.1 (Beckman Coulter), CD19, CD33 (Beckman Coulter), and CD34 (BD Bioscience) and human CD45 (BioLegend)], and the human cell repopulation ability was analyzed by FCM (FACSCantoII; BD Bioscience). From weeks 22 to 24 the mice were euthanized, and the multilineage human hematopoietic cell repopulation ability in mouse BM, peripheral blood (PB), spleen, and thymus was analyzed by FCM.

Table 2.

Summary of Serial Transplantation Analyses

		Results of Transplantation
Type of Cells Transplanted		Primary Transplantation (22––24 Weeks)	Secondary Transplantation (20––23 Weeks)	Tertiary Transplantation (>18 Weeks)
	Incidence of engraftment	9/9	9/9	1/7
CD34+	% Human cells	1.8%––90.3%	0.01%––28.7%	0.01%
MPL+	(median)	(25.4%)	(1.5%)	(0.01%)
	Number of transplanted cells (SRCs)	5,000 (104)	*	*
	Incidence of engraftment	9/9	9/9	4/7
CD34+	% Human cells	0.9%––86.2%	0.01%––61.2%	0.01%––2.1%
MPL-	(median)	(40.1%)	(3.8%)	(0.03%)
	Number of transplanted cells (SRCs)	5,000 (35)	*	*
	Incidence of engraftment	12/13	1/10	0/10
CD34-	% Human cells	0.1%––62.0%	0.5%	––
MPL+	(median)	(11.6%)	(0.5%)
	Number of transplanted cells (SRCs)	2,000 (4)	*	*
	Incidence of engraftment	12/12	9/10	3/10
CD34-	% Human cells	0.9%––90.4%	0.01%––20.5%	0.01%––0.07%
MPL-	(median)	(31.0%)	(0.4%)	(0.06%)
	Number of transplanted cells (SRCs)	8,000 (5)	*	*

^*1/5 of whole BM cells recovered were transplanted to each recipient mouse.

Limiting Dilution Analyses

In order to estimate the frequencies of SRCs in the CB-derived 18Lin^–CD34^+/–MPL^+/– fractions, various numbers of 18Lin^–CD34^+/–MPL^+/– cells were transplanted into NOG mice using an IBMI technique, as previously reported^4–6,8–10. At week 12 the mice were euthanized, and the human CD45⁺ cell repopulation ability in mouse BM was analyzed by FCM. The frequency of SRC in each fraction was calculated from the dose of transplanted cells, the number of transplanted mice, and the number of human CD45⁺ cell repopulated mice using an extreme limiting dilution analysis (ELDA) software (http://bioinf.wehi.edu.au/software/elda/)²¹.

Secondary and Tertiary Transplantation

For secondary transplantation, 1/5 of BM cells obtained from pairs of femurs and tibiae of individual primary recipient mice were transplanted into secondary recipient mice, as previously reported^4,6,9,10. For tertiary transplantations, murine BM cells were obtained from the pairs of femurs and tibiae of engrafted secondary recipient NOG mice at 20 to 23 weeks after transplantation. Tertiary transplantations were then performed as described for secondary transplantation. Finally, beyond 18 weeks after transplantation, the presence of human CD45⁺ cells in the BM of tertiary recipients was analyzed by FCM, as described above.

Statistical Analysis

The differences in the mean colony numbers or repopulation levels of human CD45⁺ cells of each pair were examined by the two-tailed Student's t-test. The differences in the expression levels of lineage markers between each pair of all the means were examined by Tukey's multiple comparison procedure. All statistical analyses were performed using KaleidaGraph software version 3.6 (HULINKS Inc., Tokyo, Japan). In the limiting dilution analysis (LDA), the frequencies of SRCs were calculated using ELDA software as previously reported²¹.

Results

Purification of Human Cord Blood-Derived 18Lin^–CD34^+/–MPL^+/– Cells

A representative FACS profile of nine experiments is shown in Figure 1. First, the R1 gate was set on the blast–lymphocyte window (Fig. 1A). Next, the 7-AAD^–18Lin^– cells were gated as R2 in Figure 1B. The 18Lin^–CD45⁺CD34^+/– cells were then gated as R3 and R4 (Fig. 1C), respectively. The 18Lin^–CD45⁺CD34^+/–cells were further subdivided into four distinct populations gated as R5 (18Lin^– CD45⁺CD34⁺MPL⁺), R6 (18Lin^– CD45⁺CD34⁺ MPL^–), R7 (18Lin^–CD45⁺CD34^–MPL⁺), and R8 (18Lin^–CD45⁺CD34^–MPL^–) according to their surface MPL (CD110) expression (Fig. 1D and E). Then 18Lin^–CD34⁺ MPL^+/– (R5 and R6) and 18Lin^– CD34^– MPL^+/– (R7 and R8) fractions were sorted for the subsequent experiments.

Figure 1.

The isolation of 18Lin^–CD34^+/– myeloproliferative leukemia virus (MPL)^+/– cells from human cord blood (CB)-derived Lin^– cells. Human CB-derived Lin^– cells were stained with anti-18Lin, anti-CD34, anti-CD45, and anti-MPL monoclonal antibodies (mAbs). (A) The R1 gate was set on the blast–lymphocyte window. (B) The R2 gate was set on the 18Lin^– living cells. (C) The cells in the R2 gate were subdivided into CD34⁺ (R3) and CD34^– (R4) fractions. (D, E) The CD34^+/– cells were further subdivided into MPL^+/–(R5–R8) fractions. As shown in (D) and (E), the percentages of MPL⁺ cells in the CD34⁺ (R5) and CD34^– (R7) fractions ranged from 13.6% to 56.4% (median: 36.9%, n = 9) and 1.4% to 33.1% (median: 10.3%, n = 9), respectively. 7AAD, 7-aminoactinomycin D; FSC, forward scatter; SSC, side scatter.

Previously, we reported that human CB-derived CD34^+/– Flt3^– SRCs were LT-HSCs with a distinct secondary repopulating capacity⁶. In this study, we analyzed the expression patterns of Flt3 on the 18Lin^–CD34^+/–MPL^+/–cell populations by FCM. As shown in Figure 2, 18Lin^–CD34⁺ MPL^+/– cells contained Flt3^+/– cells. On the other hand, the 18Lin^–CD34^–MPL^+/– cells also contained Flt3^+/– cells.

Figure 2.

The expression of Fms-like tyrosine kinase 3 (Flt3) on human CB-derived 18Lin^–CD34^+/–MPL^+/– cells. Human CB-derived Lin^– cells were stained with anti-18Lin, anti-CD34, anti-CD45, anti-MPL, and anti-Flt3 mAbs. (A) The R1 gate was set on the blast– lymphocyte window. (B) The R2 gate was set on the 18Lin^– living cells. (C) The cells in the R2 gate were subdivided into CD34⁺ (R3) and CD34^– (R4) fractions. (D, E) The CD34^+/– cells were further subdivided into four cell fractions according the expression of MPL and Flt3. The percentages of each fraction of cells are depicted in the figure.

Characteristics of Hematopoietic Colony-Forming Capacity of CB-Derived 18Lin^–CD34^+/–MPL^+/– Cells

The CFC capacities of the CB-derived 18Lin^–CD34^+/–MPL^+/– cells were quite unique. In the presence of 30% FCS and six cytokines (THPO, SCF, IL-3, GM-CSF, G-CSF, and Epo) (Fig. 3A), the plating efficiencies of 18Lin^–CD34⁺MPL⁺, 18Lin^–CD34⁺MPL^–, 18Lin^–CD34^– MPL⁺, and 18Lin^–CD34^–MPL^– cells were 77%, 58%, 47%, and 24%, respectively. Interestingly, 18Lin^–CD34^–MPL^+/– cells mainly formed burst forming unit-erythroid (BFU-E; 71% and 75%) and CFU-Mix (23% and 10%), whereas they formed few CFU-GM colonies (6% and 14%). Conversely, 18Lin^–CD34⁺MPL^+/– cells formed all types of CFCs, including CFU-GM, BFU-E, and CFU-Mix.

Figure 3.

The colony-forming cell (CFC) capacities of 18Lin^–CD34^+/–MPL^+/– cells. (A) A total of 200 18Lin^–CD34^+/–MPL^+/– cells were cultured in the semisolid methylcellulose supplemented with 30% fetal calf serum (FCS) in the presence of six cytokines [thrombopoietin (THPO), stem cell factor (SCF), interleukin-3 (IL-3), granulocyte macrophage colony-stimulating factor (GM-CSF), G-CSF, and erythropoietin (Epo)] for 14 days or (B) supplemented with 10% platelet-poor plasma in the presence of three cytokines (THPO, IL-3, and Epo) for 10 days. The types of colonies were identified under inverted microscopy. The data represent the mean ± standard deviation (SD) of quadruple cultures. CFU-GM, colony forming unit-granulocyte/macrophage; BFU-E, erythroid burst-forming unit; CFU-Meg, megakaryocyte; CFU-Mix, erythrocyte-containing mixed; CFU-EM, erythrocyte/megakaryocyte mixed colony. ^*p < 0.05, ^**p < 0.01, n.s., not significant.

In the presence of 10% PPP and three cytokines (THPO, IL-3, and Epo) (Fig. 3B), the plating efficiencies of 18Lin^–CD34⁺MPL⁺, 18Lin^–CD34⁺MPL^–, 18Lin^– CD34^–MPL⁺, and 18Lin^–CD34^–MPL^– cells were 32%, 18%, 75%, and 19%, respectively. Interestingly, 18Lin^– CD34^–MPL⁺ cells formed large numbers of CFU (EM) in addition to CFU (Meg) and BFU-E. These results are consistent with our previous reports^9–11.

Coculture with Human BM-Derived Mesenchymal Stromal Cells (DP-MSCs)

As previously reported^{4–6,9–11,17}, the CD34^– SRCs could produce CD34⁺ SRCs in vitro. Therefore, 1 × 10³ 18Lin^–CD34^+/–MPL^+/– cells were cocultured with the DP-MSCs¹⁷ in the presence of six cytokines (THPO, SCF, FL, G-CSF, IL-3, and IL-6) for 1 week. The 18Lin^–CD34^+/–MPL^+/– cells actively proliferated and maintained/generated CD34⁺ cells (Fig. 4A and B). In the cocultures of 18Lin^–CD34⁺MPL^+/– cells, the total number of cells expanded by 480- to 540-fold, resulting in a significantly higher number of CD34⁺ cell recovery compared with those of 18Lin^–CD34^–MPL^+/– cells (Fig. 4A). In contrast, the total number of cells derived from 18Lin^–CD34^–MPL^+/– cells expanded by 80- to 170-fold (Fig. 4A). The 18Lin^–CD34^–MPL^+/– cells generated CD34⁺ cells; however, the overall numbers of CD34⁺ cells were significantly low (1.9 × 10⁴ cells) compared with those of 18Lin^–CD34⁺MPL^+/– cells (Fig. 4B).

Figure 4.

In vitro lineage differentiation potentials of 18Lin^–CD34^+/–MPL^+/– cells and maintenance/generation of CD34⁺ cells from 18Lin^–CD34^+/–MPL^+/– cells in the coculture with bone marrow (BM)-derived mesenchymal stem cells (DP-MSCs). A total of 1,000 18Lin^–CD34^+/–MPL^+/– cells were cocultured with DP-MSCs for 7 days. (A) The fold increase in the total number of cells. (B) The absolute numbers of CD34⁺ cells maintained/generated in cocultures with DP-MSCs are shown. The absolute numbers of (C) CD33⁺, (D) CD11b⁺, (E) CD14⁺, and (F) CD41⁺ cells generated in cocultures are shown. The data represent the mean ± SD of quadruple cocultures. ^**p < 0.01.

Next, we analyzed the generation of lineage marker-positive cells, including CD33, CD11b, CD14, and CD41 (Fig. 4C–F). The 18Lin^–CD34⁺MPL^+/– cells generated significantly higher numbers of myeloid/monocyte lineage cells compared with those of 18Lin^–CD34^–MPL^+/–cells (Fig. 4C–E). In contrast, the 18Lin^–CD34^–MPL⁺ cells generated significantly higher numbers of CD41⁺ cells compared with those of the other three cell fractions (Fig. 4F). These results are consistent with their higher CFU-Meg and CFU-EM colony-forming capacities, as shown in Figure 3.

Limiting Dilution Analysis (LDA) of CB-Derived 18Lin^–CD34^+/–MPL^+/– Cells by IBMI

We previously reported that the frequency of the CB-derived CD34^– SRCs in the 13Lin^–CD34^– cells was approximately 1/25,000^(4,5). Next, we developed a high-resolution purification method improving upon our negative selection method, resulting in an incidence of CD34^– SRCs in 18Lin^–CD34^– cells of 1/1,000⁽⁹⁾. Using this method, we performed an LDA to demonstrate the frequency of the CD34^+/– SRCs in 18Lin^–CD34^+/–MPL^+/–cells by IBMI. As shown in Figure 5, the incidences of SRCs in these four cell fractions were as follows: 1/48 in 18Lin^–CD34⁺MPL⁺ cells, 1/146 in 18Lin^–CD34⁺MPL^–cells, 1/580 in 18Lin^–CD34^–MPL⁺ cells, and 1/1,649 in 18Lin^–CD34^–MPL^– cells. As a result of these LDA data, we next analyzed the LT-repopulating capacity of these four types of SRCs.

Figure 5.

The frequency of severe combined immunodeficiency (SCID)-repopulating cells (SRCs) in 18Lin^–CD34^+/–MPL^+/– cells. Various numbers of 18Lin^–CD34⁺MPL⁺ cells (50, 100, and 200, n = 18) (A), 18Lin^–CD34⁺MPL^– cells (200, 400, and 800, n = 18) (B), 18Lin^–CD34^–MPL⁺ cells (400 and 2,000, n = 20) (C), and 18Lin^–CD34^–MPL^– cells (1,000, 1,600, and 3,000, n = 23) (D) were transplanted into NOG mice by intra-BM injection (IBMI). At week 12, the mice were euthanized, and the human CD45⁺ cell engraftment was analyzed. The solid line represents the estimated weighted mean frequency (f_WM) of SRCs, and the upper and lower dotted lines represent the 95% confidence interval of f_WM. The mice were scored as positive if more than 0.01% of the total murine BM cells were human CD45⁺ cells, as previously reported¹⁰.

Long-Term Repopulation Patterns of CD34^+/–MPL^+/–SRCs in NOG Mice and Their Multilineage Differentiation Potentials

The next approach to characterize the self-renewal potential and the LT-repopulating potential of the CD34^+/–MPL^+/– SRCs was to serially analyze the kinetics of BM engraftment for 22 to 24 weeks in NOG mice that received transplants of various numbers of 18Lin^–CD34^+/–MPL^+/– cells, including 104 CD34⁺MPL⁺ SRCs, 35 CD34⁺MPL^– SRCs, 4 CD34^–MPL⁺ SRCs, and 5 CD34^–MPL^– SRCs, as shown in Table 2. The numbers of SRCs transplanted in each recipient mouse were calculated according to the LDA analysis as shown in Figure 5.

In these experiments, nearly all primary recipient mice that received transplants of CD34^+/–MPL^+/– SRCs showed signs of human cell repopulation at 5 weeks after the transplantation (Fig. 6A and C, and Table 2). The repopulation level of all engrafted mice that received four classes of SRCs gradually increased and reached the peak levels at 12 to 18 weeks after transplantations. The levels of human cell repopulations at 22 to 24 weeks posttransplantation ranged from 1.8% to 90.3% (median: 25.4%) for 18Lin^–CD34⁺MPL⁺ cells and 0.9% to 86.2% (median: 40.1%) for 18Lin^–CD34⁺MPL^– cells, respectively. Conversely, the levels of human cell repopulations at 22 to 24 weeks posttransplantation ranged from 0.1% to 62.0% (median: 11.6%) for 18Lin^–CD34^–MPL⁺ cells and 0.9% to 90.4% (median: 31.0%) for 18Lin^–CD34^–MPL^–cells, respectively.

Figure 6.

Primary and secondary human hematopoietic cell repopulation abilities of CD34^+/-MPL^+/- SRCs. A total of 5× 10³ 18Lin-CD34⁺MPL^+/– cells (A) and 2 or 8×10³ 18Lin-CD34-MPL^+/– cells (C) were transplanted into NOG mice. The percentages of human CD45⁺ cells in the right tibiae were serially analyzed at 5, 12, 18, and 22 to 24 weeks after transplantation. The data represent the mean±SD of the results from 9 to 13 mice at each time point. (B, D) The long-term (LT) human hematopoietic cell reconstitution in secondary NOG mice was also analyzed by the BM aspiration method at 12, 18, and 20 to 23 weeks after transplantation by six-color flow cytometry (FCM). The data represent the mean±SD of the results from 9 to 10 mice at each time point. ^*p < 0.05, ^**p < 0.01.

To further evaluate the functional differences between the CD34^+/–MPL^+/– SRCs, we studied their multilineage reconstitution abilities in various organs, including the BM, PB, spleen, and thymus, in NOG mice that were transplanted with 18Lin^–CD34^+/–MPL^+/– cells by IBMI.

Analyses of the BMs, PBs, spleens, and thymi of the four representative mice that were transplanted with either (A) CD34⁺MPL⁺ SRCs, (B) CD34⁺MPL^– SRCs, (C) CD34^–MPL⁺ SRCs, or (D) CD34^–MPL^– SRCs are shown in Figure 7. In addition, the repopulation patterns of the four types of SRCs in all recipient mice are precisely shown in Table 3. The data demonstrated that all SRCs had an almost comparable in vivo differentiation capacity to generate CD34⁺ stem/progenitor cells, CD19⁺ B-lymphoid, CD33⁺ myeloid, and CD41⁺ megakaryocytic lineages at 22 to 24 weeks after transplantation. However, the mice that received CD34⁺MPL⁺ SRCs showed significantly higher levels of CD235a⁺ erythroid cell repopulation in the BMs in comparison to those that received CD34^–MPL^+/– SRCs. The CD56⁺ NK cells were also detected in the spleens in all the NOG mice. The mice that received CD34⁺MPL⁺ SRCs showed significantly higher levels of CD3⁺ T-lymphoid cell repopulation in the thymi in comparison to those that received CD34⁺ MPL^– SRCs. In addition, the CD4⁺ and CD8⁺ singlepositive and double-positive T-lymphoid cells were detected in the thymi of all the NOG mice examined. These results confirmed that all CD34^+/–MPL^+/– SRCs had multilineage differentiation potentials.

Figure 7.

LT multilineage human hematopoietic cell repopulation abilities of CD34^+/–MPL^+/– SRCs observed in the BM, peripheral blood (PB), spleen, and thymus of primary recipient mice. The human hematopoietic multilineage repopulation abilities of (A) CD34⁺MPL⁺ SRCs (upper column), (B) CD34⁺MPL^– SRCs (second column), (C) CD34^–MPL⁺ SRCs (third column), and (D) CD34^–MPL^– SRCs (bottom column) in the primary recipient mice were precisely analyzed. The surface marker expressions [CD34, CD33, CD19, and CD235a (BM); CD34, CD19, CD33, and CD41 (PB); CD34, CD19, CD33, and CD56 (spleen); and CD3, CD4, and CD8 (thymus)] in human CD45⁺ living cells (R1 gate) were analyzed. The expression of CD41 and CD235a were analyzed in human CD45^+/– living cells (dotted line). The data represent the values for four individual mice that received CD34^+/–MPL^+/– SRCs.

Table 3.

The Summary of In Vivo Lineage Differentiation Abilities of CD34^+/––MPL^+/–– SRCs in the Primary Recipient Mice

The data represent the range of percentages of marker-positive cell in each group, which contain 7 to 13 mice.

^*p < 0.05 (Tukey's multiple comparison).

Secondary and Tertiary Repopulating Potentials of CD34^+/–MPL^+/– SRCs in NOG Mice

We next performed the secondary transplantation to analyze the self-renewing capacities in greater detail. All of the secondary recipient mice (9/9) that received the transplants from the primary recipient mice, which received either 18Lin^–CD34⁺MPL^+/– cells by IBMI, showed signs of human cell repopulation at 20 to 23 weeks after transplantation (Table 2). Additionally, secondary recipient mice that received 18Lin^–CD34⁺MPL^+/– cells showed comparable median levels of human cell repopulation [0.01% to 28.7% (median: 1.5%) and 0.01% to 61.2% (median: 3.8%), respectively]. The LT repopulation patterns of these engrafted SRCs in NOG mice were nearly comparable (Fig. 6B). The repopulation levels of all engrafted mice gradually increased and reached peak levels at 12 to 18 weeks after transplantations.

Interestingly, only 1 out of 10 mice that received 18Lin^–CD34^–MPL⁺ cells was barely repopulated with human cells. Moreover, the CD45⁺ cell rates at 12 to 20 weeks after transplantation of this mouse were <1% and decreased to <0.1% at 20 weeks after the transplantation, suggesting that these CD34^–MPL⁺ SRCs were ST-repopulating SRCs (Fig. 6D). In contrast, almost all of the mice (9/10) that received 18Lin^–CD34^–MPL^– cells were engrafted with human cells. The repopulation levels of the engrafted mice were 0.01% to 20.5% (median: 0.4%). The repopulation levels of all engrafted mice gradually increased and reached the peak levels at 12 to 18 weeks after transplantations (Fig. 6D). We confirmed that all engrafted secondary recipient mice showed multilineage human cell repopulation, including CD19⁺ and CD33⁺ cells (data not shown).

Finally, we performed tertiary transplantation using all mice that lived at the time of transplantation, regardless of human cell repopulation. Very interestingly, the 4/7 secondary recipient mice that received 35 CD34⁺MPL^–SRCs and 3/10 secondary recipient mice that received 5 CD34^–MPL^– SRCs during primary transplantation showed a distinct tertiary repopulating capacity with multilineage differentiation (Table 2). None of the 10 tertiary recipient mice that received four CD34^–MPL⁺ SRCs during primary transplantation were repopulated with human cells. These results clearly indicated that CD34^+/– SRCs not expressing MPL receptor sustained a LT (at least >1 year) human cell repopulation in NOG mice, suggesting that both SRCs had significant self-renewing potentials. Conversely, only one out of seven mice that received 104 CD34⁺MPL⁺ SRCs in the primary recipient mice was barely repopulated with human cells, in which the human CD45⁺ cell rate was 0.01%, suggesting that these CD34⁺MPL⁺ SRCs were IT-repopulating SRCs.

Generation of CD34⁺CD38^–CD45RA^–CD90⁺ SRCs from CD34^– SRCs In Vivo

As reported previously, human CB-derived Lin^–CD34⁺ CD38^–CD45RA^–CD90⁺ cells contained human LT-repopulating CD34⁺CD38^–CD90⁺ SRCs^22,23. Interestingly, these Lin^–CD34⁺CD38^–CD45RA^–CD90⁺ cells were generated from the cocultures of CB-derived 18Lin^–CD34^–cells with DP-MSCs and showed a distinct SRC activity after transplantation, as we recently reported¹⁷. These results demonstrated that human CB-derived 18Lin^–CD34^– cells could generate very primitive CD34⁺ CD38^–CD45RA^–CD90⁺ SRCs in vitro. In this study, we confirmed that very primitive CD34⁺CD38^–CD45RA^–CD90⁺ SRCs could be generated in NOG mice that received a limited number of CB-derived CD34^– SRCs at 20 weeks after transplantation (Fig. 8). In contrast, CD34⁺ SRCs could not generate CD34^– SRCs in vitro and in vivo^4,9,17.

Figure 8.

The generation of CD34⁺CD38^–CD45RA^–CD90⁺ SRCs in a representative mouse that received human CB-derived 18Lin^–CD34^– cells. Freshly sorted human CB-derived 18Lin^–CD34^– cells were transplanted into NOG mice by IBMI. After 20 weeks, human CD45⁺ cell repopulation in the BM of a representative mouse was analyzed and is shown in the left column (A). Then 12Lin^–/CD45RA^–CD34⁺CD38^–CD90^+/– cells were sorted (B–D). The 12 lineages contain CD2, CD3, CD4, CD7, CD8, CD10, CD11b, CD14, CD19, CD20, CD56, and CD235a. Next, these 12Lin^–CD34⁺CD38^–CD45RA^–CD90^+/– cells (D) were transplanted into secondary NOG mice by IBMI, and the human cell repopulation was analyzed at 20 weeks after transplantation (E–G). Only the CD90⁺ cells repopulated in the secondary recipient NOG mice and showed a multilineage reconstitution. 7AAD, 7-Aminoactinomycin D.

Discussion

It is well documented that THPO is a principal cytokine that regulates megakaryocyte/platelet production, and its signals are transduced through its receptor, MPL^24–27. MPL is reported to be expressed not only on megakaryocytes/platelets but also on HSPCs^24–27. As expected, THPO and MPL knockout (KO) mice showed a significant reduction in the numbers of megakaryocytes and circulating platelets as well as HSCs^2,28–30. It was also previously reported that THPO and MPL KO mice showed increased numbers of fetal liver HSCs, but selective and progressive postnatal reduction of LT-HSCs in THPO KO mice². These results demonstrated that the THPO/MPL signaling is not required for fetal HSC expansion, but adult quiescent HSCs are highly THPO/MPL dependent². Conversely, the study and understanding of the function of the THPO/MPL signaling pathway in human primitive HSCs are far behind that of murine studies. The above-mentioned murine studies have suggested that the functional roles of THPO/MPL signaling in human BM (reflecting adult hematopoiesis) and CB (reflecting fetal hematopoiesis) may be different.

In this study, we purified human CB-derived 18Lin^–CD34^+/–MPL^+/– cells and analyzed their HSPC activities in vitro and in vivo. As shown in Figure 3, 18Lin^–CD34^+/–MPL⁺ cells formed significantly higher numbers of CFCs both in the presence of 30% FCS + six cytokines and 10% PPP + three cytokines compared with those of 18Lin^–CD34^+/–MPL^– cells. Interestingly, 18Lin^–CD34^+/–MPL⁺ cells formed significantly higher numbers of BFU-E, CFU-Mix, CFU-Meg, and CFU-EM colonies compared with their MPL^– counterparts. These results suggested that THPO/MPL signaling plays an important role at the multipotent progenitor cell (MPP) level and the erythroid and megakaryocytic lineage commitment level of CB-derived HSCs. In the cocultures of 18Lin^–CD34^+/–MPL^+/– cells with DP-MSCs¹⁷, 18Lin^–CD34⁺MPL^+/– cells maintained/generated significantly higher numbers of CD34⁺, CD33⁺, CD11b⁺, and CD14⁺ cells compared with those of 18Lin^–CD34^–MPL^+/–cells (Fig. 4). Interestingly, 18Lin^–CD34^–MPL⁺ cells generated significantly higher numbers of CD41⁺ cells compared with those of the other three populations. These results suggested that the THPO/MPL signaling plays a role in the megakaryocytic lineage commitment of CB-derived HSPCs.

Finally, we analyzed the expression patterns of CD34 and MPL on primitive human CB-derived HSCs using a sensitive SRC assay system, as reported previously^4–11. As shown in Figure 5, the LDA demonstrated that the incidence of CD34⁺ SRCs in 18Lin^–CD34⁺MPL⁺ cells was 1/48, which was higher than that in 18Lin^–CD34⁺ MPL^– cells (1/146). Conversely, the incidence of CD34^–SRCs in 18Lin^–CD34^–MPL⁺ cells was 1/580, which was higher than that in 18Lin^–CD34^–MPL^– cells (1/1,649). These results demonstrated that the expression of MPL increased the SRC incidence in the CB-derived 18Lin^–CD34^+/– cells, respectively.

We then more precisely analyzed the HSC characteristics of these four classes of CD34^+/–MPL^+/–SRCs using serial transplantation analyses. As shown in Figure 6 and Table 2, nearly all primary recipient NOG mice that received 18Lin^–CD34^+/–MPL^+/– cells were repopulated with human CD45⁺ cells. Among these four classes of SRCs, nearly all secondary recipient NOG mice that received 18Lin^–CD34⁺MPL^+/– and 18Lin^–CD34^–MPL^–cells were again repopulated with human CD45⁺ cells. Interestingly, distinct numbers of recipient NOG mice that received 18Lin^–CD34^+/–MPL^–cells were repopulated with human CD45⁺ cells even in tertiary transplantation. We transplanted only five CD34^–MPL^– SRCs into the primary recipient NOG mice; however, they sustained primary, secondary, and tertiary transplantations. Moreover, they showed nearly equivalent LT-repopulating potentials compared with those of 35 CD34⁺MPL^– SRCs transplanted into the primary recipient NOG mice. These results suggested that CD34^–MPL^– SRCs had higher proliferative and self-renewing capacities compared with those of CD34⁺MPL^–SRCs. These results were consistent with our previous observation that one CD34^– SRC could produce significantly greater numbers of CD45⁺ and CD34⁺ cells than those that were produced from one CD34⁺ SRC in vivo⁵.

It is well established that HSCs are defined as cells with self-renewal and multilineage differentiation potential. A number of studies have demonstrated that murine and human HSCs are heterogeneous populations in multiple aspects, including their degree of self-renewal^31–33, differentiation manner^11,34–37, and life span^{3,6,8,38–40}. Recently, Ema et al. proposed a model based on the reconstitution time period of murine HSCs⁴¹. The authors compared the relationship of three HSC classification models, including myeloid-biased (My-bi)/balanced (Bala)/lymphoid-biased (Ly-bi)³⁴, α-/β-/γ-δ-cells^35,36, and ST/IT/LT HSCs^{3,6,8,38–40}. Interestingly, most LT-HSCs were My-bi HSCs and classified as α-cells. IT-HSCs contained My-bi, Bala, and Ly-bi HSCs and mostly classified as α- and β-cells. Nearly all ST-HSCs were Ly-bi HSCs and classified as γ-cells. In this model, murine HSCs were reclassified as ST-, IT-, and LT-HSCs, defined as ST <6 months, IT >6 months, and LT >12 months.

According to our serial transplantation analyses, we propose that the expression pattern of CD34 antigen and MPL receptor defines/segregates human CB-derived primitive HSCs into CD34^–MPL⁺ ST-, CD34⁺MPL⁺ IT-, and CD34^+/–MPL^– LT-HSCs, as shown in Figure 9. In our model, we use the following definition: ST-HSCs only support human cell repopulation in primary recipient mice; IT-HSCs support human cell repopulation in primary and secondary recipient mice; and LT-HSCs support human cell repopulation in primary, secondary, and tertiary recipient mice. Previously, Guenechea et al. reported that the human cell repopulation in NOD/SCID mice that received Lin^– CB cells was generally oligoclonal with extensive variability in the life span and proliferative capacity of individual SRCs³¹. Our data are consistent with this previous report. Collectively, these data imply the existence of different classes of human HSCs with variable self-renewal potential and repopulation capacity in the human CB.

Figure 9.

Proposed model of the human CB-derived hematopoietic stem cell (HSC) hierarchy according to the CD34 and MPL expression pattern. According to our serial transplantation analyses, we propose that the expression pattern of CD34 antigen and MPL receptor defines/segregates human CB-derived primitive HSCs into CD34^–MPL⁺ short-term (ST)-, CD34⁺MPL⁺ intermediate-term (IT)-, and CD34^+/–MPL^– LT-HSCs, as shown in this figure. In our model, we use the following definition: ST-HSCs only support human cell repopulation in primary recipient mice; IT-HSCs support human cell repopulation in primary and secondary recipient mice; and LT-HSCs support human cell repopulation in primary, secondary, and tertiary recipient mice. The CD34^–MPL^– LT-HSCs reside at the apex of the human HSC hierarchy. It is suggested that they first commit to CD34⁺MPL^– LT-HSCs, then commit to CD34⁺MPL⁺ IT-HSCs, and finally commit to CD34^–MPL⁺ ST-HSCs.

Recently, we succeeded in highly purifying human CB-derived CD34^+/– SRCs at the 1/20 cell level using a newly identified positive marker, GPI-80⁴². CD34⁺ CD38^–GPI-80^+/– SRCs could be generated in cocultures of human CB-derived 18Lin^–CD34^–GPI-80⁺ cells with DP-MSCs¹⁷. On the other hand, CD34^–GPI-80^+/– SRCs could not be generated in the cocultures of the human CB-derived 18Lin^–CD34⁺CD38^–GPI-80⁺ cells with DP-MSCs¹⁷. According to these data, we propose a new model that these CD34^– SRCs appear to be more immature than the previously recognized most primitive CD34⁺ CD38^–CD45RA^–CD90⁺ SRCs⁴³. In this study, we demonstrated that human CB-derived CD34^+/–MPL^– SRCs could repopulate human hematopoiesis in primary, secondary, and tertiary recipient NOG mice. These results indicated that both CD34^+/–MPL^– SRCs are LT-HSCs. As mentioned above, CD34^– SRCs could generate CD34⁺ SRCs in vitro and in vivo (Fig. 8), suggesting that CD34^–MPL^– SRCs reside at the apex of the human HSC hierarchy. These CD34⁺MPL^– LT-HSCs appear to initially commit to CD34⁺MPL⁺ IT-HSCs, and then finally commit to CD34^– MPL⁺ ST-HSCs, as illustrated in Figure 9.

In conclusion, the present study demonstrated for the first time that the expression pattern of CD34 antigen and MPL receptor in the primitive human CB-derived HSCs strongly correlated with the serial transplantation analyses and resulted in the categorization of primitive human HSCs into ST-, IT-, and LT-repopulating HSCs, as shown in Figure 9. In the SRC assay system using NOG mice, the human THPO/MPL signaling pathway does not work, at least during the early phase of repopulation after transplantation, due to a lack of human active THPO in NOG mice⁴⁴. As previously reported⁴⁴, hematopoiesis in human THPO knockin RAG2^–/–γ_c^–/– mice is improved in comparison to control mice. Thus, the repopulation capacity of CD34^–MPL⁺ ST- and CD34⁺MPL⁺ IT-SRCs may be improved in the presence of human THPO/MPL signaling. However, the most primitive CD34^+/–MPL^–LT-SRCs cannot respond to human THPO. Thus, these CD34^+/–MPL^– LT-SRCs may depend on unidentified factors/signals other than THPO/MPL. From another point of view, single human CB-derived Lin^–CD34⁺CD38^–CD4 5RA^–CD90⁺Rho^lowCD49f⁺ cells transplanted directly into the mouse femur could repopulate human hematopoiesis in NOD/ShiLtSz-scid/IL2Rgnull (NSG) mice at 20 weeks after the transplantation⁴¹. These results clearly indicated that murine BM niche cells and/or some unidentified niche factors could support the homing, proliferation, and even self-renewal of human HSCs at the single-cell level. Therefore, unidentified factors/signals other than THPO/MPL may play a pivotal role in maintaining the primitive human HSCs in the BM niche of NOG/NSG mice across the species barrier. Overall, these findings suggested that the functional significance of the MPL expression in the human primitive HSC compartment differs from that in the murine primitive HSC compartment.