obtained; 4 mL was diluted quickly in IS mL of Hank's Balanced. Salt Solution (HBSS) ... quots of 200 mL of the complete medium'R''"~'3 containing 50 ng/ ...... kare SB, Sachdev RK, Yuschnkoff VN, Birkett NC, Williams LR,. Satyagal VN, Tung ...
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only.
1995 85: 3538-3546
Lymphokine requirement for the generation of natural killer cells from CD34+ hematopoietic progenitor cells A Shibuya, K Nagayoshi, K Nakamura and H Nakauchi
Information about reproducing this article in parts or in its entirety may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requests Information about ordering reprints may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprints Information about subscriptions and ASH membership may be found online at: http://bloodjournal.hematologylibrary.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036. Copyright 2011 by The American Society of Hematology; all rights reserved.
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only.
Lymphokine Requirement for the Generation of Natural Killer Cells From CD34+ Hematopoietic Progenitor Cells By Akira Shibuya, Kazunari Nagayoshi, Kozo Nakamura, and Hiromitsu Nakauchi We have established a cell culture system without stromal cells that allows the CD34+ hematopoietic progenitor cells (HPC) to differentiate into natural killer (NK) cells. CD34+Lin (CD3, CD16, CD56)- cells were purified using fluorescenceactivated cell sorting from normal adult bone marrow (BM) and cultured for 28 days in medium supplemented with interleukin-2 (IL-2) and stem cell factor (SCF). NK (CD3-CD16CD56+)cells were generated in a dose-dependent manner in response to SCF. NK cells originated from CD34+CD33+Lincells, but they were barely detectable in cultures of CD34+CD33-Lin- cells. However, on addition of IL-3, an in-
duced differentiation of NK cellsfrom CD34TD33-Lin- cells was observed, although at a lower frequency.Supplementing of the cell cultureswith SCF alone or both SCF and IL-3 for the first 7 days followed by IL-2 forthe next 21 days is essential forproduction of NK cellsfrom CD34+CD33+Lincellsand from CD34TD33-Lin- cells,respectively.These data provide directevidence that NKcellsarise from CD34+HPCand show the minimum lymphokine requirement for their differentiation. 0 1995 by The American Societyof Hematology.
N
tional NK cells from CD34’ hematopoietic progenitor cells (HPC) was not caused by the action of IL-2 alone” but required the presence of stromal cells (Shibuya et al, unpublished data and Miller et a124.2s).It was also been reported that NK cells could be generated from rat BM NK progenitors only when stroma was present in the culture system.” Thus, the interaction of CD34+ HPC with stromal cells appears to be critical at least in the early stages ofNK cell differentiation. However, the culture system of NK cells with stromal cells is too complex to analyze the mechanism of NK cell differentiation, because a number of factors capable of stimulating CD34’ HPC both directly and indirectly could be produced by stromal cells. Therefore, a well-defined in vitro culture system of NK cells without stromal cells should first be established. In the present report, we describe an in vitro culture system generating NK cells at a high frequency from fluorescence-activated cell sorter (FACS)-purifiedCD34’ HPC. We show thatIL-3, stem cell factor (SCF), and IL-2 play sequential and important roles in the early differentiation ofNK cells from their BM precursors.
ATURAL KILLER (NK) cells are believed to play a role in the primary host defense against tumor cells and virus-infected cells.’,z Although itis known that NK precursor cells reside in the bone marrow (BM),’.4 their developmental sequence into mature NK cells remains largely unknown. NK and T cells have several traits in common that suggest a common developmental origin. They share cytotoxic activity and some surface antigens. NK cells are generated from immature thymocyte under appropriate in vitro and in vivo condition^.^.^ It has also been reported that functional NK cells isolated from the human fetal liver or spleen possess cytoplasmic CD3.’0”2However, NK cells appear to be thymus-independent; they arise in athymic mice’?and are present in severe combined immunodeficient (SCID) and human SCID patients.I6 It has also been proposed that NK cells are ontogenetically related to the myeloid or the monocytic lineage.14.” The clarification of the lineage relationship of NK cells with other hematopoietic cells requires an appropriate culture system of NK cells from hematopoietic BM stem cells to be established. We and others have reported that NK cells can differentiate from BM cells in a culture medium containing interleukin-2 (IL-2).’*-’’ Recently, we have shown that an IL-2-responsive NK progenitor existed among BM cells depleted of mature T and NK cells.” The phenotype of these NK progenitor cells was defined as being CD34-, CD33-, and CD25+. In addition, we found that production of funcFrom the Division of Hematology, Institute of Clinical Medicine, University of Tsukuba; and the Laboratory of Cell Growth and Differentiation, Tsukuba Life Science Center, The Institute of Physical and Chemical Research (RIKEN), Tsukuba, Ibaraki, Japan. Submitted October 12, 1994; accepted January 25, 1995. Supported by grants from the Ministry of Education, Science, and Culture of Japan, Fujisawa Pharmaceutical Company Ltd, and Nagase Science and Technology Foundation. Address reprint requests to Hiromitsu Nakauchi, MD, PhD, Institute of Basic Medical Sciences, University of Tsukuba, TsukubaCity, Ibaraki 305, Japan. The publication costs of this article were defrayedin part by page chargepayment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. section 1734 solely to indicate this fact. 0 1995 by The American Society of Hematology. 0006-4971/95/8512-0010$3.00/0
3538
MATERIALSANDMETHODS Preparation of BM cells and cell sorting.BM was aspirated from the sternum of healthy volunteers after informed consent was obtained; 4 mL was diluted quickly in I S mL of Hank’s Balanced SaltSolution(HBSS)containingheparin.BMmononuclearcells (MNC) were then separated by Ficoll-Hypaque centrifugation and were suspended in RPM1 1640 medium with 5% fetal calf serum (FCS; GIBCO, Grand Island,NY). BM MNC were incubated simulI >’: taneously with allophycocyanin (APC)-conjugated CD34 (4A IgG), phycoerythrin(PE)-conjugated CD3(Leu4;IgGI;Becton IC; IgGl; Becton Dickinson), Dickinson, San Jose, CA), CD16 (Leu1 and CD56 (NKH-I; IgG1; Coulter Clone, Hialeah, FL), and either fluorescein isothiocyanate (FITC)-conjugated CD33 (My9; IgG2b; Coulter Clone) or purified anti-c-kit antibody (IgM; Immunotech. Marseille, France) for 30 minutes at 4“C, followed by two washings. For staining with the anti-c-kit antibody, cells were then treated with FITC-conjugated rat antimouse monoclonal IgM (Zymed, San Francisco, CA). After 10 minutes of incubation, cells were washed twiceandresuspended in phosphate-bufferedsaline(PBS)containing 3% FCS and S pg/mL 7-amino-actinomycin D (7AAMD). Cell sorting windows are indicated as shown in Fig I . Computerassisted data analysis of the results was performed on a MicroVAX Corp, Maynard, MA) withFACSl computer(DigitalEquipment DESK software from the FACS development group at Stanford University. Residual erythrocytes and dead cells were gated out using Blood, Vol 85, No 12 (June 15).
1995: pp 3538-3546
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only. 3539
GENERATION OF NK CELLS FROM CD34'CELLS
I
A
V a I 2 I Lin-
4-
m
n
V
Lin(CD3,16,56)-PE
CD33-FITC
C o
Ia
b
D
c
4
I c-kit -FITC Fig 1. FACS analysis of BM MNC stained with lineage markers (CD3, CD16, and CD56). CD34, and either CD33 (A) or anti-c-kit antibody (B). Cell sorting experiments were performed as follows: for CD33 fractionation of the CD34' cells, the sorting gates were Aa (total CD34' cells), Ab (negative 30%). and Ac (brightest 50%); for c-kit fractionation of CD34' cells, the sorting gates were Ba (negative 35961, Bb (dull 50%). and BC (brightest 10%). Percentages are approximate values obtained from three independent experiments. The sorting window for each fraction is indicatedin these maps. Photomicrographs of cells sorted from CD34+CD33-Lin- (C) and theCD34+CD33' Lin- (D) fraction are also shown (original magnification x 1,000).
forward scatter, side scatter channels, and 7AAMD staining at the time of cell sorting. For investigation of CD7 expression on CD34'Lin- cells, BM MNC were stained with FITC-conjugated CD7 (Leu9; Becton Dickinson), APC-conjugated CD34, andPEconjugated lineage antigen markers (CD3, CD16, and CD56) with or without PE-conjugated CD33 (My9). Crtlrure and immunofluorescence analysis of sortedcells. Aliquots of 200 mL of the complete medium'R''"~'3containing 50 ng/ mL recombinant human SCF (Kirin Brewery CO Ltd, Maebashi, Japan), 500 UlmL recombinant human IL-2 (Shionogi Pharmaceutical, Osaka, Japan), and/or 100 UlmL recombinant human IL-3 (Genzyme, Cambridge, MA) were placed in round-bottomed 96-well microplates. Five hundred sorted cells from each fraction were plated
to each well and were cultured for 28 days. The medium was replaced with fresh medium with or without an identical dose of SCF, IL-2, and/or IL-3 on days 7, 14, and 21. In some experiments, the following cytokines were used: IL-l (200 UlmL; Sigma, St Louis, MO), IL-4 (500 UlmL; Sigma), IL-6 (100 UlmL; Chugai Pharmaceutical, Tokyo, Japan), IL-7 (200 UlmL; Sigma). IL-12 ( 1 0 0 UlmL; Genetic Institute, Cambridge, MA), interferon-y (500 UlmL: Sumitorno Pharmaceutical, Osaka, Japan), erythropoietin ( I O UlmL; Chugai Pharmaceutical), granulocyte-macrophage colony-stimulating factor (GM-CSF; I O nglmL; Shering-Plough. Osaka, Japan), and granulocyte-CSF (G-CSF; I O ng/mL; Chugai Pharmaceutical). Two-color immunofluorescence analysis of FITC-conjugated CD3 and PE-conjugated CD56 on the sorted cells during culture was
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only. 3540
SHIBUYA ET AL
performed with a FACStarPLUS. The NK (CD3- CD56') cell count was determined as the total viable cell count per well multiplied by the percentage of CD3- CD56' cells X 1/1OO. NK activity. After culture, the sorted cells were suspended in RPM1 1640 supplemented with 5% FCS. Cytotoxic activity against K562 was assessed as described In brief, K562 target cells were labeled with 50 pCi of "Cr for 2 hours at 37°C and then washed extensively before use. Target cells (S X IO' ) suspended in 10 pL of medium and effector cells suspended in 0.1 mL of medium were plated in microtiter wells at several effector/ target ratios (20:1, lO:l, 5: 1, 2.5:l). After 4 hours of incubation at 3 7 T , the supernatant was harvested using the Titerteck system (Flow Laboratories, Irvine, CA) and its radioactivity was measured with a gamma counter (Gamma 5500; Beckmann Instruments Inc, Irvine, CA). Cytotoxicity was determined from the amount of 51Crreleased by the lysed target cells. Triplicate determinations were made and "Cr release was calculated as: 100 X ([cpm in experimental wells] - [cpm in wells with target cells alone])/(cpm incorporated into target cells). One lytic unit (LU) was defined as the NK activity that caused 15% "Cr release from S X IO3 target cells, and the LU per 10' cells was determined for each sample from a regression curve of the percent cytolysis. NK activity per well was calculated as LU per 1O'cells X (1/10') X total viable cell count/well. The assay of NK activity was performed 14.21, and 28 days after starting culture of sorted cells. Limiting dilution analysis. Two-hundred microliter aliquots of the complete medium containing SO ng/mL SCF and SO0 U/mL IL2 were placed in round-bottomed 96-we[[ microplates. For limiting dilution analysis, 500, 300, 150, and 75 cells from each fraction were directly sorted into 12 microtiter wells, respectively, using FACStarPLUS and Automated Cell Deposition Unit. Twenty four microwells containing the complete medium with SCF and IL-2 were prepared as controls. On culture day 28, "Cr-labeled K562 target cells (S X lo3) were added to each well. After 4 hours of incubation, supernatants were harvested. NK activity in each well was measured and the frequency of cytolytic progenitors was calculated, as described p r e v i o ~ s l y . Otherwise, '~~~~ cells were harvested on culture day 28 and stained with FITC-conjugated CD3 and PE-conjugated CD56. Wells were determined positive for NK cells if greater than 0.1 % of CD3-CD56' cells were observed and the frequency of NK progenitors were calculated in the same way. Morphologic analysis. The sorted cells (0.5 to l X 10' in 200 pLof medium) and cultured cells ( O S to 1 X IO') were centrifuged onto microscope slides for 3 minutes at 800 rpm, using an automatic cytospinner (Autosmear CF12D; Sakura Fine Technical, Tokyo, Japan). The slides were rapidly air-dried and then stained using Wright's stain (Muto Pure Chemicals, Tokyo, Japan). RESULTS
Generation of NK cells from CD34' HPC in the presence of SCF and IL-2. SCF is a cytokine produced from stromal cells that binds to the plasma membrane receptor known as c-kit on HPC.28"' Therefore, we have investigated whether SCF could at least partially substitute to stromal cells in stimulating CD34' HPC to differentiate into NK cells. CD34' cells devoid of the T- and NK-cell lineage markers CD3, CD16, and CD56 (Lin-) were FACS-purified from normal BM (Fig IAa) and cultured for 28 days in a medium supplemented with IL-2 and SCF, but without stromal cells. Lineage marker analysis showed the presence of CD3CD56' cells and a majority of these cells were also CD16(Fig 2A). Other NK markers such as CD2 and CD8 were coexpressed with CD56 on one half and one third of the cell
population, respectively. The phenotype of the cultured NK cells was consistent with that ofNK cells generated in vitro from IL-2-responsive NK progenitors CD34-CD33CD25'Lin- cells." The number ofNK cells increased in a dose-dependent manner in response to added SCF (Fig 2B). Morphologic examination of the cells showed that most appeared to be large granular lymphocytes (LGL) admixed with other cell types, including granulocytes. macrophages, and mast cells (not shown). Generation of NK cells ,from CD34'CD33' HPC in the presence of SCF and IL-2. To define early differentiation events in the pathway leading to NK cells from HPC, CD34' cells were subfractionated by FACS on the basis of the expression of CD33. The percentages of CD34TD33"Lin cells and CD34TD33'Lin- cells were 0.5% t 0.3% and 0.7% t 0.4% in BM MNC, respectively. However, CD34' CD33-Lin- and CD34TD33'Lin- cells were sorted from the strict windows of CD33-negative 30% (corresponding to 0.3% t 0.2% of BM MNC) and CD33-brightest 50% (0.4% t 0.2%), respectively (Fig 1Ab and Ac). Morphologically, the CD34TD33-Lin- cells were small to medium-sized lymphoblastoid cells with basophilic cytoplasm (Fig 1C). By contrast, most CD34'CD33'Lin- cells appeared larger in size and had fine cytoplasmic granules (Fig 1D). These observations were consistent with the idea that CD34' CD33-Lin- cells are less differentiated than CD34'CD33' Lin- cells."CD34'CD33'Lincells divided faster than CD34TD33-Lin- cells when cultured in the presence of SCF and IL-2 (Fig 3A). As shown in Fig 3B, CD3"CD16 CD56' NK cells appeared in the cultures of CD34'CD33' Lin- cells but were barely detectable in cultures of CD34' CD33-Lin.~ cells. In addition, the cells generated from CD34TD33'Lin- cells had cytolytic activity against K562 cells (Fig 3C), indicating that CD3-CD16-CD56' cells cultured from CD34+Lin- cells were functional NK cells. Limiting dilution analysis showed that the frequency of NK progenitors was 1/275 (95% confidence range, 11295 to 1/237) among CD34TD33'Lin- cells, whereas the CD34'CD33 Lin- fraction didnot contain measurable numbers ofNK progenitors under the present culture conditions (Fig 3D). We also studied CD7 expression, which was recently reported to be an important antigen marker for NK progenitor," on CD34'Lin- cells. Three-color analysis showed that CD7 was expressed on 1.8% t 0.7% (n = 3) of CD34'Lin cells (0.023% f 0.009% ofBM MNC). When CD33, in addition to lineage antigens, was excluded from CD34' cells (ie, CD34'CD33-Lin-..), CD7 was detected on 1.9% t 0.5% (n = 3) of these cells (0.012% 2 0.004% ofBM MNC). No significant difference in CD7 expression intensity was observed between CD34'Lin- cells and CD34'CD33-Lincells. Generation of N K cells from CD34'CD33- HPC in the presence of IL-3,SCF, and IL-2. To determine whether another cytokine(s) responsive to HPC in addition to SCF and IL-2 could be required for the generation of N K cells from the less differentiated CD34TD33-Lin- cells. we added various cytokines, including IL-l, IL-3, IL-4, IL-6, IL-7, IL-12, GM-CSF, G-CSF, interferon-y, or erythropoietin, to in vitro grown CD34'CD33-Lin- cells. Of these cyto-
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only. CELLS GENERATION OF NK
FROM CD34' CELLS
3541
W
pc
W
l -
pc
m4
I
v,
v,
Control-FITC
F
v)
n
n
U
U
CD3-FITC
CD16-FITC
SCF (ng/mL) Fig 2. CD3-CD16-CD56'NKcells were cultured from CD34'Linincreased in a dose-dependent manner in response to SCF (B).
kines, only IL-3 in combination with SCF and IL-2 induced the differentiation of CD34'CD3TLin- cells into putative N K CD3-CD56' cells after 28 days of culture (Fig 4A and Table I ). Morphologic analysis of the cultured cells in the presence of IL-3 in combination with SCF and IL-2 showed a small number of lymphoid cells (< 1%) as well as nonlymphoid cells (macrophage, 39%: granulocyte, 40%; mast cell, 20%). To confirm the generation of N K cells under the culture condition, cells were sorted from CD3-CD56' cell fraction after the culture and analyzed in the morphology, showing large granular lymphocytes (Fig 49). Although the percentage of CD3 CD%' cells from these cultures was not high enough IO show functional NK activity, these phenotypic and morphological observations clearly indicated that least at a subset of N K cells was generated from CD34'CDWLin- cells when the three lymphokines SCF, IL-2. and IL-3 are present. Although N K cells were obviously generated as described above.functional limiting dilution assays could not show the frequency of N K progenitors because of low sensitivity to NK activity in each well generated from 500 cells. Be-
HPC in medium with SCF and IL-2for
28 days (A). Cultured NKcells
cause it was difficult to plate more than 500 sorted cells in each well, we developed a limiting dilution analysis using FACS. As can be seen in Fig 4C, a straight line was obtained when wells were determined positive for NK cells if greater than 0.1% of CD3-CD56' cells were observed. According to limiting dilution theory?' a P value greater than.2. as was observed in all our experiments. indicates a high probability of single-hit kinetics. The limiting dilution analysis by this method showed the frequency of N K progenitors in CD34TD33-LiI" cells to be 1/462 (1/630 to 11385). although this frequency could not be directly compared with the frequency by functional analysis. SCF ploys (111 essentiol role in the generotion of IL-2responsive N K cell progenitors. N K cells didnot arise from CD34TD33'Lin- cells in a culture medium supplemented with either SCF or IL-2 alone (Table 2). However, N K cells were generated from CD34TD33'Lin- cells when both SCF and IL-2 were present. However, the presence of the twolymphokines is not necessary during the entire length of cell culture. NK cells were generated if SCF was present during the first 7 days of culture and could never be gener-
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only. 3542
SHIBUYA ET AL
C
h
15
10
.-uh > .W
5
Y
O* 0
2
4 4 , -dI. 0
7
14
21
28
Z
7
rh
21
20
Days of Culture
Fig 3. The total cell count (AI, CD3-CD16-CD56' cell count (B), and NK activity (C) of sorted cells from CD34'CD33- Lin- (0). and CD34'CD33+ Lin- (01 cell fraction, as shown in Fig 1A. during culture in the presence of SCF and IL-2, and frequency of NK progenitors in each sorted fraction (D).
ated without SCF during this period. Furthermore, the presence of IL-2 was not necessary during these first 7 days but was compulsory later on. These results indicated that supplement of cultures with SCF for the first 7 days followed byIL-2 for the next 21 days appeared to be essential for production ofNK cells. Likewise, NK cells were also observed after CD34TD33-Lin- cells were cultured in the presence of SCF and IL-3 for the first 7 days followed by IL-2 for the next 21 days (Table 2), suggesting that SCF and IL-3 caused CD34TD33-Lin- HPC to become responsive to IL-2. However, NK cells were never detected without SCF even when cells were grown in the presence of IL-2 and IL-3. These results strongly suggest that SCF is essential for differentiation of CD34' HPC into IL-2-responsive NK progenitors. Additional evidence in support of the role of SCF in the early stages of differentiation of NK cells from CD34+Lincells, came from the analysis of the expression of the c-kit, the cell membrane receptor for SCF?8-3' Cells were sorted
from each fraction (Fig 1Ba through c) and cultured in SCF plus IL-2. Colony-forming units ( C m ) responsive to SCF were found predominantly in the c-kit weakly positive (ckit') fraction as compared with c-kit strongly positive (ckit") fraction.35This finding correlated with the observation that NK cells arose only from the c-kit'CD34'Lin" cell fraction (Table 2). These results represent a further indication that SCF, in cooperation with IL-2, is necessary for differentiation of NK cells from CD34' HPC. DISCUSSION
We report here that NK cells could be generated from CD34' HPC, in the absence of stromal cells in an in vitro culture system supplemented with IL-3 andor SCF in combination with IL-2. NK cells could never be detected after long-term culture from both CD34+CD33+Lin- cells and CD34TD33-Lin- cells in a culture medium supplemented with IL-2 alone even at a high concentration (500 U/mL). This result, which is in agreement with conclusions reached
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only. GENERATION OFNKCELLSFROM
CD34' CELLS
3543
B
A
ResponderCellsper
Well
Fig 4. The presence of CD3-CD56' cells was confirmed by FACS analysis in the culture of CD34TD33-Lin- cells after 28 days (A). The morphologic analysis of sorted cells from thefraction of CD3-CD56' cells showed large granular lymphocytes, indicating that thecells were NK cells (B).Limiting dilution analysis by FACS showed that thefrequency of NK progenitors was 1/462 among CD34TD33-Lin- cells (Cl.
by others,'4 conflicts with a recent report by Lotzova et a136 showing the generation of NK cells from CD34TD33- cells in human BM in a culture medium supplemented with only IL-2. The discrepancy can be explained if IL-2-responsive NK cells were still present in their cell preparation, although they attempted to eliminate residual mature T and N K cells
Table 1. Generation of NK Cells From CD34TD33-Lit-1- Cells ~~
Cytokine in Addition to SCF and IL-2
IL-l IL-3 IL-4 IL-6 IL-7 IL-12 GM-CSF G-CSF Interferon-y Erythropoietin
CD56' Cell Count Total Cell Count (x1O'well)
0.8 6.7 0.4 1.0 0.5 0.9 4.1 1.5 0.7 0.8
-c 0.2 t- 2.1 t 0.2 t- 0.4 t- 0.4 t 0.3 t- 1.8 t- 0.4 t- 0.3 t- 0.3
%
xlO'/well
0 0.8 t- 0.3 0 0
0 5.3 2 1.6 0
0
0 0
0
0 0
0 0
0
0 0 0
0
CD34.CD33-Lin- cells werecultured witheach cytokine in combination with SCF and IL-2 for 28 days. Total and CD56+ cell counts were determined after culture.
by treatment with anti-CD.?, anti-CDS6, and anti-CD16 monoclonal antibodies and complement. Indeed, the majority of the cultured N K cells display a CD16-CDS6' phenotype, a subset of N K cells that constitutively expresses the highaffinity IL-2 receptors" and proliferates in response to low concentrations of IL-2.3xThe absence of production of N K cells after long-term culture of the HPC in the presence of IL-2 alone in our study is strongly indicative of the complete elimination of mature N K cells from FACS-sorted cell fractions. The frequency of NK progenitors in CD34'CD33'Lincells determined by functional limiting dilution analysis under our stroma-independent culture condition was 1/275. A recent report by Miller et a1'5 showed that CD34TD7'Lincell population exhibited high cloning efficiency of N K progenitors (3.2%) under a stroma-dependent long-term culture condition. CD34'CD7'Lin- cells and CD34-CD7'CDK Lin- cells were detected in our study at 0.023% and 0.012% ofBM MNC, respectively. Therefore, the percentage of CD34'CD7'CD33'Lin- cells in BMMNC was calculated to be 0.01 l%, indicating that CD34'CD7+Lin- cell population consisted of CD33' and CD33- cells. The present study showed that CD34'CD33-Lin- cells could be distinguished from CD34TD33'Lin- cells by minimal cytokine requirement for N K cell differentiation. suggesting thatCD34'
4',
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only. SHIBUYA ET AL
3544
Table 2. Generation of NK Cells From FACS-Sorted Fractions Presence of Cytokines (culture period in days)
CD56' Cell Count Total Cell Count
Lin-
CD33'. 0-28
0-28 7-28 14-28 CD33-, CD34', Lin
0-28
SCF
IL-2
-
0-28
0-28 0-28 0-7 7-28 0-28 0-28
-
Yo
x10'iwell
0
0
0
3.4 i- 1.2 4.5 i 1.2 1.7 i- 0.6 1.0 i 0.4 2.5 2 1.0 1.8 2 0.7
0
0
45.2 10.1 2 3.2 6.2 i 2.1 0 7.9 i 2.6 0.3 i- 0.2
10.7 0 19.7 0.5
-
-
-
0-28
-
0-28
-
-
-
-
0-28
0 0 2.3 0.5 2.6 5.8 5.4 2.1
-
0-28 0-7 0-7 0-7
7-28 0-28
-
c-kit-, CD34'. 0-28Linc-kit', CD34+, 0-28Linc-kit", CD34'. 0-28 Lin
(xlO"/well)
-
-
0-28
0-28 0-28 0-28 0-7
0-28
IL-3
0-28 0-28 0-28
2 1.1 -t
0.1
i- 1.3 T 2.4 i- 5.6
i- 1.4
0 0 0 0 0 2.90.5 i- 0.2 0.2 2 0.1 0
0 0 0 0 0 1.1 0
-
0
0
0
-
4.1 2 1.6 0.6 i- 0.3
2.9 2 0.4 0
12.0 0
-
Cells sorted from each fraction (Fig 1Ab and c and 1Ba through c) were cultured in the presence or absence of cytokinek) during each culture period. Total cell and CD56' cell counts were determined after 28 days.
CD7+Lin- cells were still heterogeneous in the stage of NK cell development. Further studies are required to clarify the characteristics of CD34+CD7+CD33-or CD33+ cells as NK cell progenitors. SCF plays an essential role in the biology of CD3CD16-CD56+ NK cells. NK cells displaying this phenotype are in vivo found as a subset of peripheral NK cells.39Recently, it was reported that CD3-CD16-CD56' NK cells express the c-kit antigen and proliferate synergistically in response to SCF and IL-2.40We report here that the lymphokine is necessary for the differentiation of CD34+ HPC into NK cells. Thus, SCF is likely to play a primary role in the differentiation and proliferation of broad spectrum ofNK cells, including the differentiation of very early NK progenitors to peripheral CD16-CD56* NK cells. However, although SCF plays a central role in the early stages of differentiation of NK cells from CD34+HPC, SCF is not sufficient alone to allow the differentiation of the CD34TD33-Lin- subset to engage into the NK cell differentiation pathway. IL-3 is required for it, although it is ob-
11-2-Responslue NK Progenltor
C034+ Hematopoletic
Progenitor Cell 0
l
L
CD34+ CD33CD3CD1 CD1 6CD56+ CD56- CD56-
-
3
(
+
S
C
F
CD34+ CD33+ 6-
)
0
S
=
2
0
CD34CD33-
l
L
scure whether IL-3 has a direct effect on CD34+CD33-LinNK progenitors. Addition of IL-3 to SCF and IL-2 remarkably increased total cell number (Table l ) , most of which consisted of macrophages, granulocytes, and mast cells. It is possible that IL-3 interacts indirectly with CD34+ CD33-Lin- NK progenitors through a cytokine(s) produced from these nonlymphoid cells. However, the similar nonlymphoid cell components were observed after culture with GM-CSF in addition to SCF and IL-2, where NK cell was never generated (Table 1). Therefore, it is unlikely that a cytokine(s) produced from the nonlymphoid cells induces CD34'CD33-Lin- NK progenitor into NK cell pathway, unless IL-3 and GM-CSF each stimulate the nonlymphoid cells to produce a different cytokine(s). Based on these findings, we propose the model for NK cell differentiation shown in Fig 5 . HPC with the CD34TD33-Lin- phenotype differentiate along theNKcell pathway into CD34TD33'Lin- cells in the presence of IL-3. At this early stage, SCF is required for NK progenitors to become responsive to IL-2. Subsequently, IL-2-re-
NK Cell -
2
a
CD34CD33-
CD1 6+/CD1 6CD56C D ~ ~ ( I L - Z R ~ ) + CD25(IL-2Ra)CD1 22(IL-2Rg)+ CD1 22(IL-2Rp)-
Fig 5. Hypothetical schemadiffershowing early entiation pathway of NK cellsfrom CD%+ hematopoietic progenitor cells.
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only. GENERATION OF NK CELLS FROM CD34' CELLS
3545
cytoplasmic CD3e proteins in activated human adult natural killer sponsive NK progenitors cease to express the CD34 and (NK) cells and CD3y.6,~complexes in fetal NK cells. Implications CD33 antigenz3and then differentiate into NK cells. for the relationship of NK and T lymphocytes. J Immunol 149: 1876, Although this succession of differentiation events appears 1992 likely, the relationships between NK and other hematolym12. Hori T, Phillips JH, Duncan B, Lanier LL, Spits H: Human phoid progenitors remain unclear. It has been reported that fetal liver-drived CD7+CD2lnWCD3-CD56- clones that express functional NK cells isolated from the human fetal liver or CD3y, 6, and e and proliferate in response to interleukin-2 (IL-2), spleen have cytoplasmic CD3.""*In addition, it has also IL-3, IL-4, or IL-7: Implications for the relationship between T and been reported that triple negative fetal thymocyte can difnatural killer cells. Blood 80:1270, 1992 ferentiate along the NK-cell pathway in the presence of 13. Herberman RB, Nun ME, Holden HT, Levin D: Natural cytotoxic reactivity of mouse lymphoid cells against syngeneic and alloIL-2,5-9suggesting that NK and T cells may arise from a geneic tumors. I. Distribution of reactivity and specificity. Int J common progenitor. In the present study of adult BM, NK Cancer 16:216, 1975 cells were derived from CD34+CD33+Lin- cells among 14. Dorshkind K, Pollack SB, Bosma MJ, Phillips RA: Natural which large numbers of C m - G M are pre~ent.~'.~' The results killer (NK) cells are present in mice with severe combined immunomay suggest that adult human NK cells are closely related deficiency (SCID). J Immunol 134:3798, 1985 to the myeloid lineage. However, because of the low fre15. Hackett JJ, Bosma GC, Bosma MJ, Kumar V: Transplantable quency of NK progenitors in CD34+CD33+Lin- cells, it is progenitors of natural killer cells are distinct from those of T and B still possible that CD34+CD33+Lin- compartment includes lymphocytes. Proc Natl Acad Sci USA 83:3427, 1986 separate progenitors for myeloid and NK cells. Actually, 16. Peter HH, Friedrich W, Dopfer R, Muller W, Kortmann C, CD33 antigen is not always specific for myeloid lir~eage.~.~' Pichler WJ, Heinz F, Rieger CHL: NK cells function in severe combined immunodeficiency (SCID): Evidence of a common T and Introduction of a clonal assay system such as retroviral markNK cell defect in some but not all SCID patients. J Immunol ing in this culture system will solve this issue. ACKNOWLEDGMENT We thank Drs K. Shibuya, R.A. Shiurba, G. Gachelin, D. Tarlinton, J.H. Phillips, and L.L. Lanier for discussion; M. Isoda and Y. Inazawa for technical help; Dr Y. Matsuzaki for figures; K. Okada for secretarial assistance; and the Kirin Brewery CO Ltd (Tokyo, Japan) for providing SCF. REFERENCES 1. Herberman RB: NK Cells and Other Natural Effector Cells.
New York, NY, Academic, 1982 2. Bukowski JF. Warner JF, Dennert G, Welsh RM: Adoptive transfer studies demonstrating the antiviral effect of natural killer cells in vivo. J Exp Med 161:40, 1985 3. Hallor 0,Kiessling R, Om A, Wigzell H: Generation of natural killer cells: An autonomous function ofthebone marrow. J Exp Med 145:1411, 1977 4. Claeson MH, Olsson L: Autoreactive natural killer-like cells from agar-cloned murine bone marrow cells. Nature 283578, 1980 5. Rodewald H-R, Moingeon P, Lucich JL, Dosiou C, Lopez P, Reinherz EL: A population of early fetal thymocytes expressing FcyRII/II contains precursors of T lymphocytes and natural killer cells. Cell 69:139, 1992 6. Hori T, Spits H: Clonal analysis of human CD3-CD4-CD8thymocytes highly purified from postnatal thymus. J Immunol 146:2116, 1991 7. Matsuzaki Y, Gyotoku J, Ogawa M, Nishikawa S, Katsura Y, Gachelin G, Nakauchi H: Characterization of c-kit positive intrathymic stem cells that are restricted to lymphoid differentiation. J Exp Med 178:1283, 1993 8. Sanchez MJ, Spits H, Lainer LL, Phillips JH: Human natural killer cell committed thymocytes and their relation tothe T cell lineage. J Exp Med 178:1857, 1993 9. Sanchez MJ, Muench MO, Roncarolo MG, Lanier LL, Phillips JH: Identification of a common Thatural killer cell progenitor in human fetal thymus. J Exp Med 180569, 1994 10. Phillips JH, Hori T, Nagler A, Bhat N, Spits H, Lanier LL: Ontogeny of human natural killer (NK) cells: Fetal NK cells mediate cytotoxic function and express cytoplasmic CD3e,6 proteins. J Exp Med 1751055, 1992 11. Lanier LL, Chang C, Spits H, Phillips JH: Expression of
131:2332, 1983 17. Lohmann-Matthes M, Domizig W, Roder J: Promonocytes have the functional characteristics of natural killer cells. J Immunol 123:1883, 1979 18. Yoda Y, Kawakarni Z , Shibuya A, Abe T: Characterization of natural killer cells cultured from human bone marrow cells. Exp Hematol 16:712, 1988 19. Shibuya A, Taguchi K, Kojima H, Abe T: Inhibitory effect of granulocyte-macrophage colony-stimulating factor therapy on the generation of natural killer cells. Blood 78:3241, 1991 20. Taguchi K, Shibuya A, Inazawa Y, Abe T: Suppressive effect of granulocyte-macrophage colony-stimulating factor on the generation of natural killer cells in vitro. Blood 79:3227, 1992 21. Keever CA, Gazzola MV, Pekle K, Bourhis JH, Gillio A: Regulatory effect of recombinant cytokines on NK cell differentiationfrom early marrow precursors, in FreundM,LinkH, Welte K (eds): Cytokines in Hemopoiesis, Oncology, andAIDS. Berlin, Germany, Springer-Verlag, 1990, p 225 22. Koo GC, Manyak CL: Generation cytotoxic cells from murine bone marrow by human recombinant IL-2. J Immunol 137:1751, 1986 23. Shibuya A, Kojima H, Shibuya K, Nagayoshi K, Nagasawa T, Nakauchi H: Enrichment of interleukin-2-responsive natural killer progenitors in human bone marrow. Blood 81:1819, 1993 24. Miller JS, Verfaillie C, McGlave P: The generation of human natural killer cells from CD34+/DR- primitive progenitors in longterm bone marrow culture. Blood 80:2182, 1992 25. Miller JS, Alley KA, McGlave P: Differentiation of natural killer (NK) cells fromhuman primitive marrow progenitors in a stroma-based long-term culture system: Identification of a CD34'7' Nk progenitor. Blood 83:2594, 1994 26.Van den Brink MRM, Boggs SS, Herberman RB, Hiserodt JC: The generation of natural killer (NK) cells from NK precursor cells in rat long-term bone marrow cultures. J Exp Med 172:303, 1990 27. Gunji Y, Nakamura M, Hagiwara T, Hayakawa K, Matsushiya H, Osawa H, Nagayoshi K, Nakauchi H, Yanagisawa M, Miura Y, Suda T: Expression and function of adhesion molecules on human hematopoietic stem cells: CD34'LFA-1- cells are more primitive than CD34+LFA-l+cells. Blood 80:429, 1992 28. Williams DE, Eisenman J, BairdA,Rauch C, Ness KV, March CJ, Park LS, Martin U, Mochizuki DY, Boswen HS, Burgess
From bloodjournal.hematologylibrary.org by guest on July 13, 2011. For personal use only. 3546
GS, Cosman D, Lyman SD: Identification of a ligand for the c-kif proto-oncogene. Cell 63:167, 1990 29. Flanagan JG,Leder P The kit ligand: A cell surface molecule altered in steel mutant fibroblasts. Cell 63: 185, 1990 30. Zsebo KM, Wypych J, McNiece IK, Lu HS, Smith KA, Karkare SB, Sachdev RK, Yuschnkoff VN, Birkett NC, Williams LR, Satyagal VN, Tung W, Bosselman RA, Mendiaz EA, Langley KE: Identification, purification, and biological characterization of hematopoietic stem cell factor from buffalo rat-conditioned medium. Cell 63:213, 1990 31. Huang E, Nocka K, Beier DR, Chu T-Y, Buck J, Lahm H-W, Wellner D, Leder P, Besmer P: The hematopoietic growth factor KL is encoded by the S1 locus and is the ligand of the c-kit receptor, the gene product of the W locus. Cell 63:225, 1990 32. Ema H, Suda T, Miura Y, Nakauchi H: Colony formation of clone-sorted human hematopoietic progenitors. Blood 76: 1956, 1990 33. Taswell C: Limiting dilution assays for the determination of immunocompetent cell frequencies. I. Data analysis. J Immunol 128:1614, 1981 34. Miller RG: Frequency determination of immunocompetent cells, in Fathman CG, Fitch FW (eds): Isolation, Characterization and Utilization of T Lymphocyte Clones. San Diego, CA, Academic, 1982, p 220
SHIBUYA ET AL
35. Gunji Y, Nakamura M, Osawa H, Nagayoshi K,Nakauchi H, Miura Y, Yanagisawa M, Suda T: Human primitive hematopoietic cells are more enriched in KIT'"" cells than in KIThIeh cells. Blood 82:3283, 1993 36. Lotzova E, Savary CA, Champlin RE: Genesis of human oncolytic natural killer cells from primitive CD34+CD33- bone marrow progenitors. J Immunol 1505263, 1993 37. Nagler A, Lanier LL, Phillips JH: Constitutive expression of the high affinity interleukin 2 receptors on human CD16- natural killer cells. J Exp Med 171:1527, 1990 38. Nagler A, Lanier LL, Cwirla S, Phillips JH: Comparative studies of human FcRIII-positive and negative natural killer cells. J Immunol 143:3183, 1989 39. Matos ME, Schnier GS, Beecher MS, Caligiuri MA: Expression of functional c-kit receptor on a subset of natural killer cells. 3 Exp Med 178:1079, 1993 40. Ema H, Suda T, Nagayoshi K, Miura Y, Civin CI, Nakauchi H: Target cells for granulocyte colony-stimulating factor, interleukin-3, and interleukin-5 in differentiation pathways of neutrophils and eosinophils. Blood 75:1941, 1990 41. Nakamura Y, Noma M, Kidokoro M, Kobayashi N, Takei M, Kurashima S, Mukaiyama T, Kat0 S: Expression of CD33 antigen on normal human activated T lymphocytes. Blood 83:1442, 1994
