From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
Use of a Promoter-Trap Retrovirus to Identify and Isolate Genes Involved
in Differentiation of a Myeloid Progenitor Cell Line In Vitro
By Jan-lngvar Jonsson, Qi Wu, Kenneth Nilsson, and Robert A. Phillips
Studies of gene regulation during early hematopoiesis and
of the regulatory network that controls differentiation and
lineage commitment are hampered by difficuttiesin isolating
and growing stem cells and early progenitor cells. These
difficulties preclude the application of standard molecular
genetic approaches to these problems. As an alternative approach we have introduced a lacZ-containingpromoter-trap
retrovirus into hematopoieticcells. We used the interleukin3-dependent mouse myeloid progenitor cell 32D as a model
to identify transcriptionally active genes. The frequency of
integrations that led to transcription of the lac2 gene was
estimated to be 0.5% of all integrations, of which 14% were
downregulated on differentiation of 32D cells towards neu-
trophils. Thus, one in every 1,000 to 2,000 integrationsidentified a developmentally regulated gene. Cellular DNA sequences upstream of proviral integrations were isolated by
inverse polymerasechain reaction. Five were further characterized and we confirmed by RNA expression analysis that
they were downregulatedon differentiation.Sequence analysis revealed identification of novel genes with sequence
similarity to known genes. Considering the high efficiency of
retroviral infection, our study shows the feasibility of using
promoter-trap vectors to identify and isolate developmentally regulated genes from early hematopoietic progenitors.
0 1996 by The American Society of Hematology.
T
identify such genes in a cell line. Here we report on the
usage of a promoter-trap retroviral vector containing the lucZ
gene as a reporter gene5 to identify transcriptionally active
genes at the stage of proliferation in the IL-3 dependent
myeloid progenitor cell line, 32D,l4-l6and to follow the expression and subsequent downregulation of these genes during differentiation to mature neutrophils in the presence of
granulocyte colony-stimulating factor (G-CSF). Thus, 32D
cells function as an in vitro model system to study the genetic
program involved in differentiation of hematopoietic progenitor cells towards the granulocytic lineage. We have isolated
cellular DNA sequences upstream of proviral integrations
from 32D cells with downregulated ZucZ expression during
the transition of highly proliferative cells to mature, nonproliferative, and terminally differentiated neutrophils. We have
characterized in detail five of these sequences and could
confirm by Northern blot analysis that these genes are downregulated in response to G-CSF. Comparison of the isolated
sequences with the GenBank Data Base showed three novel
genes with sequence similarities to known genes. We have
identified one additional sequence to be an endogenous retro-
HROUGHOUT adult life, multipotent stem cells and
progenitors in the bone marrow (BM) maintain a constant supply of mature cells in the hematopoietic system.
The mechanisms underlying the process by which stem cells
and early progenitors become committed towards certain
blood lineages are still unknown. Knowledge of these mechanisms are essential for understanding the differentiation
processes in this system and how errors in the genetic program can lead to leukemia. Despite advances in recent years
identifying numerous hematopoietic growth factors and
growth factor receptors, the definitive signals that determine
these events remain to be elucidated.’.’ The study of gene
regulation in hematopoietic stem cells and early progenitor
cells is mainly limited by difficulties in identifying, isolating,
and growing these cells. Consequently, it is extremely difficult to address these problems by standard molecular genetic
analysis, eg, the generation of cDNA libraries and subtractive hybridization techniques. Several groups have succeeded in purifying cell populations enriched for early hematopoietic progenitor cells able to grow in culture and
differentiate into both myeloid and lymphoid cells.’,4 These
new culture systems provide alternative approaches to analyzing the genetic regulation of early hematopoietic progenitor cells in vitro. We have developed such a strategy using
promoter-trap retroviral constructs to monitor transcriptionally active regions of the genome in progenitor^.^ In brief,
it involves the introduction of a promoterless reporter gene
into a host cell and the detection of those integrations that
lead to efficient expression of the reporter gene from an
endogenous promoter active in progenitor cells. The ZucZ
gene provides a way to easily detect transcriptionally active
promoters and a sensitive assay to follow its regulation on
differentiation. This approach has successfully been used
in the study of gene regulation during embryogenesis by
introducing enhancer-, gene-, or promoter-trap retroviral
vectors into mouse embryonic stem
Additional studies have indicated the usefulness of similar vectors to study
cell growth and differentiation in other types of
However, such vectors have not been used to identify genes
involved in early hematopoietic progenitors.
To develop this strategy to identify developmentally regulated genes in primary hematopoietic cell populations, we
first validated the efficiency of a promoter-trap retrovirus to
Blood, Vol 87, No 5 (March I ) , 1996: pp 1771-1779
From the Division of Immunology and Cancer Research, Hospital
for Sick Children, Toronto, Ontario, Canada; and the Department
of Pathology, University of Uppsala, University Hospital, Uppsala,
Sweden.
Submitted July 27, 1995; accepted October 12, 1995.
Supported by grants from the Medical Research Council and the
National Cancer Institute of Canada with funds from the Canadian
Cancer Society (R.A.P.);and from the Swedish Cancer Society, the
Children Cancer Foundation of Sweden, Hans von Kantzows and
Magnus Bergvalls Foundations (J.-LJ). J. -1.J. was supported in part
by Fellowships from The Hospital for Sick Children Foundation,
Medicinska forskningsr6det, Stockholm, Sweden, and the Swedish
Cancer Society. Q.W. was a recipient of a Fellowship from the
Medical Research Council of Canada.
Address reprint requests to Jan-lngvar Jonsson, Laboratory of
Tumor Biology, Department of Pathology, University of Uppsala,
University Hospital, S-751 85 Uppsala, Sweden.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
“advertisement” in accordance with 18 U.S.C. section 1734 solely to
indicate this fact.
0 1996 by The American Society of Hematology.
0006-4971/96/8705-003.5$3.00/0
1771
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
1772
JONSSON ET AL
virus present in the mouse genome with an apparent regulated expression on terminal differentiation of 32D cells. Our
study confirms the feasibility of using promoter-trap vectors
to identify and clone genes involved in hematopoietic differentiation.
MATERIALS AND METHODS
Promoter-trap retrovirus vll3LacZ. A packaging-defective ecotropic cell line (&) transfected with the retroviral vector
pGgTKNeoU3LacZ(-) was kindly provided by Drs Earl Ruley
(Vanderbilt University, Nashville, TN) and H. von Melchner (Universitiitsklinikum,Frankfurt, Germany). The construction of this promoter-trap virus has been described elsewhere?," Briefly, it contains
a promoterless 3.1-kb fragment of the Escherichia coli-derived lacZ
gene inserted into the 5'-end of the long terminal repeats (LTR) of
Moloney murine leukemia virus and a bacterial neo gene expressed
from the herpes simplex virus thymidine kinase (tk)promoter to
enable selection of cells with proviral integrations (see Fig 2C).
The lac2 gene extends into the U3-provirus sequence that lacks the
sequence that contains the viral transcriptional enhancer.
Cells and culture conditions. +,BG, producing recombinant retrovirus vU3Lac2, was maintained in a-modified Eagle medium
(GIBCO, Grand Island, NY) containing 10% fetal calf serum (FCS)
(Sigma, St Louis, MO) and G418 (500 pg active/mL; GIBCO). The
murine Interleukin (IL-3-dependent) 32D clone 3 cell linel4.l5was
kindly provided by Dr G. Rovera (Wistar Institute, Philadelphia,
PA) and was maintained in Iscove's modified Dulbecco's medium
(IMDM; GIBCO) supplemented with 15% FCS and 1% of an IL3-containing culture supernatant from a cell line stably transfected
with an expression vector containing the cDNA for murine IL-3 (a
gift from Dr F. Melchers, Basel Institute of Immunology, Basel,
S~itzerland).'~
Cells were cultured at 37°C in 5% CO, and were
split 1: 10 twice a week. For induction of neutrophilic differentiation,
32D cells were washed three times with IMDM and then incubated
in IMDM containing 15% FCS and 100 U/mL recombinant human
G-CSF (kindly provided by the Genetics Institute, Cambridge, MA).
Differentiation into neutrophils was verified by morphological characterization by collecting cells daily after the start of growth in the
presence of G-CSF followed by concentration on slides by cytocentrifugation. Cells were then fixed and stained with Wright stain for
microscopic inspection.
Retroviral infection and isolation of LacZ' 3 2 0 clones by fluorescence-activated cell sorter (FACS) and limiting dilution. Before
the infection experiment, the virus titer was analyzed by titration of
culture supernatant of JIZBGcells on Rat-2 fibroblast cells18 and was
estimated to 1 x IO6 (3418-resistant (G41SR)colony-forming unit/
mW107 &BG cells. One day before infection, &BG cells were
irradiated with 20 Gy and plated on a 100-mm tissue culture dish
at a concentration of 1 X lo5cellslmL in 10 mL of IMDM containing
15% FCS. After an ovemight incubation, 32D cells (4 X IO6) were
seeded onto the dish with virus producer plus 1% IL-3 supernatant
and 4 pglmL polybrene (Sigma). After 2 days of cocultivation, 32D
cells were collected, washed, and incubated for an additional 4 days
in IMDM containing 15% FCS, 1% IL-3 supernatant, and 800 pg/
mL active (3418. To enrich for cells expressing P-galactosidase, and
thus bearing a proviral integration into a transcriptionally active
gene, G418R 32D cells were washed and sorted by flow cytometry
on a FACStar cell sorter (Becton Dickinson, Mountain View, CA)
using a fluorescence assay with fluorescein di-P galactopyranoside
(FDG) (Molecular Probes, Eugene, Oregon). Conditions for staining
cells were those as described by others." Because of the low frequency of L a d + cells, the sorting parameters were set to select 2%
of the cells with the highest fluorescence. These cells were retumed
to culture for an additional 3 days. To further enrich for LacZ+ 32D
cells, the sorting procedure was repeated twice. After three rounds
of FDG-sorting, cells were cultured under limiting dilution conditions ( I celVwell in 96-well plates). Cultures were fed every third
day with fresh IMDM containing 15% FCS, 1% IL-3 supematant,
and 500 pg/mL G418.
Analysis of P-galactosidase activity in undifferentiated and differentiated 3 2 0 cells. After establishment of individual LacZ' 32D
clones, P-galactosidase activity was analyzed by staining with 5bromo 4-chloro 3-indolyl P-galactoside (X-Gal; Boehringer Mannheim, Germany) as described." Briefly, cells were washed three
times in phosphate-buffered saline (PBS) and fixed in 0.2% paraformaldehyde in PBS for 15 minutes at 4°C. They were then washed
once in PBS and stained for 4 hours to overnight with X-Gal directly
in 96-well plates at 37°C. Clones that were positive in this assay
were split into triplicate wells. One well was maintained in 1% IL3 supematant, whereas cells in the second and third well were allowed to differentiate in the presence of G-CSF for 3 and 6 days,
respectively, and at that time X-Gal staining was performed. Cells
were photographed within 24 hours.
Southern blot analysis. Genomic DNA from 32D clones with a
downregulated P-galactosidase expression activity was isolated by
a salting out method" and digested with DraI which does not cut
inside the viral sequence. Southern blot hybridization was performed
according to standard procedures."
Amplification and cloning of upstreamflanking sequences. Genomic DNA from 32D clones (3 pg) were digested with 15 U of HinfI
or Rsal in 1 X polymerase chain reaction (PCR) buffer (50 mmoV
L KCI; 10 mmol/L Tris-HCI, pH 8.3; 1.5 mmol/L MgCIZ;0.01%
gelatin) supplemented with 100 pg/mL RNAse A in a volume of
100 pL. After heat-inactivation for 20 minutes at 65"C, 1 pg was
ligated at a concentration of 10 pg/mL in 1X PCR buffer and 0.1
mmol/L adenosine triphosphate (ATP) with 1 pL T4 DNA ligase
( 5 U/pL, Boehringer). After heat-inactivation, religated DNA was
digested with 10 U of PvuII for 1 hour. One hundred nanograms of
DNA from each sample was then mixed with 50 pmol of each
primer, 200 pmol/L of each desoxyribonucleotide triphosphate, and
2.5 U of Amplitaq DNA polymerase (Perkin Elmer Cetus, Nonvalk,
CT) in I X PCR buffer in a total volume of 100 pL. Amplification
was performed using denaturation (94°C for 1 minute), annealing
(55°C for 90 seconds) and extension steps (72°C for 3 minutes)
for 35 cycles with an initial denaturation at 96°C for 5 minutes.
Oligonucleotide primers complementary to either the lad-sequence
or the U3-region of the LTRZ3were used. Samples were analyzed
and purified on 2% NuSieve agarose gels (FMC BioProducts, Rockland, ME). Gel-purified PCR products were then diluted 1:loO and
amplified for an additional 30 cycles with the identical primers. The
PCR products were identified by agarose gel electrophoresis and the
appropriate ethidium bromide stained bands were excised, the DNA
eluted, purified and ligated to Bluescript(KS-) plasmids (Stratagene,
La Jolla, CA).
Nuclear run-on assay. PCR products containing unique DNA
fragments flanking proviral integrations were cloned by inverse PCR
and separated by agarose gel electrophoresis, blotted onto nylon
membrane, and hybridized to radiolabelled nuclear run-on transcripts
from 32D cells as described previo~sly?~
Briefly, 3 X IO' cells were
washed three times with ice-cold PBS and resuspended in lysis buffer
containing 10 mmom Tris-HCI, pH 7.4, 10 mmol/L NaC1,3 mmol/L
MgCI2, 0.5% (vol/vol) NP-40 and 0.5 mmol/L dithiothreitol (DTT).
Nuclei were then isolated by 20 to 30 strokes in a Dounce homogenizer and frozen at -70°C. Radiolabelled nuclear run-on transcripts
were prepared at 37°C for 15 minutes by incubating 50 pL thawed
nuclei with 50 pL of DEPC-dH20 containing 250 pCi [(U-~~PIUTP
(3,000 Ci/mmol, Amersham, Arlington Heights, IL), 1.O mmol/L
each of ATP, cytidine (CTP) and guanosine (GTP), and 400 pg/
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
1773
GENE TRAPPING IN MYELOID PROGENITOR CELLS
mL creatinine kinase (Sigma). RNA was then isolated by phenol
extraction in the presence of guanidinium is~thiocyanate.’~
RNA extraction and Northern blot analysis. Total RNA was
prepared by phenol extraction in the presence of guanidinium isothio~yanate.~~
Poly(A)+RNA was then prepared by using a biotinylated-oligo(dT)-probe and streptavidin magnetic beads (Promega,
Madison, WI). Poly(A)’-selected RNA (2 to 3 pg) was fractionated
on 1% agarose gels containing 6% formaldehyde. After electrophoresis, RNA was transferred to Gene Screen nylon membranes (New
England Nuclear, Boston, MA) in 0.04 m o a NaPQ-buffer, pH 7.5,
and fixed in vacuum oven at 80°C for 2 hours. Membranes were
hybridized at 42°C for 16 to 24 hours in 25% to 50% formamide,
SXSSPE, 5 X Denhardt’s, 0.5% sodium dodecyl sulfate (SDS), 100
pg/mL denatured salmon sperm DNA (Sigma), 1 X lo6 cpdmL of
specific probe labelled by random oligonucleotide priming” and
washed at different stringency depending on the probe used and its
length. A 900-bp probe for the human glyceraldehyde-3-phosphate
dehydrogenase gene was used as a control to quantify the amounts
of mRNA.
Sequencing reactions and GenBank Data Base analysis. Nucleotide sequences of PCR-generated flanking DNA ligated to Bluescript
plasmids were determined by the dideoxynucleotide method, using
double-stranded plasmid templates, and a T7 polymerase sequencing
kit (Pharmacia, Uppsala, Sweden). Both strands were sequenced
across both T7 and T3 primer-binding sites. Isolated sequences were
analyzed for open-reading frames by computer analysis using the
Geneworks program. In addition, all sequences were sent by electronic mail to the GenBank Data Base at the National Center for
Biotechnology Information, National Library of Medicine (Bethesda,
MD) and compared to known genes using the BLASTN and the
BLASTX sequence analysis program for nucleotide sequences and
protein sequences, respectively.26
RESULTS
The promoter-trap retrovirus vU3LacZ identijes downregulated genes in differentiated 3 2 0 cells. To identify
genes that are downregulated in 32D cells following transfer
of cells from IL-3 to G-CSF, 32D cells growing in IL-3 were
infected by cocultivation on +,BG cells, a helper cell line
producing vU3LacZ (see Fig 2C). The retrovirus contains the
lacZ gene inserted 30 bp downstream from flanking cellular
sequences in the 5‘-end of the LTR: thus allowing the lacZ
gene to come under the control of a nearby cellular promoter
when the virus integrates into the host genome. Deletion of
the enhancer in the LTR ensures the control of the reporter
gene by the upstream cellular promoter. After 2 days of
cocultivation with virus-producing cells, 32D cells were collected and cultured for an additional 4 days in the presence
of G418 to select for cells with retroviral integrations. XGal staining of these G418R cells showed 0.5% blue cells;
ie, the frequency of integrations into actively transcribed
genes and successful transcription of the lacZ gene was 1
of 200.
To enrich for cells expressing p-galactosidase, G418R32D
cells were sorted using FDG introduced into cells by hypotonic shock.’’ After three rounds of FDG-sorting, after which
58% of the cells stained blue for X-Gal, cells were cultured
under limiting dilution conditions (1 cell/well in 96-well
plates) in the presence of IL-3 and G418. A total of 908
clones were established, of which 492 (54%) expressed pgalactosidase by X-Gal staining. Expression ranged from
very intense staining (1 1% of the clones) to intermediate
Table 1. Analysis of p-Galactosidase Expression Before and After
G-CSF Stimulation of Lad’ 32D Cell Clones After Infection With
Promoter-Trap Retrovirus vU3LacZ
Phenotype*
Factor Addedt
No. of Clones
% of Clones
LacZ’
LacZ’
LacZ-
IL-3
G-CSF
G-CSF
492
422
70
100
86
14
* A total of 492 clones were analyzed for @galactosidase expression
by X-Gal staining as described in Materials and Methods. The LacZphenotype refers to a clear and consistent downregulation and absence of 8-galactosidase expression.
t 32D clones were stimulated with either IL-3 (1% of an IL-3-containing supernatant) or rhuG-CSF (100 U/mL) for 3 days.
(51%) or low staining (37%). Only clones with intense or
intermediate staining were selected for further study.
Table 1 shows the outcome from the experiment in which
the 492 LacZ+ clones were induced to differentiate in the
presence of rhuG-CSF in 96-well plates. X-Gal staining was
analyzed on days 3 and 6. Seventy clones (14%) showed
downregulated expression of the lacZ gene and could be
divided into 3 groups according to the level of downregulation. The first group (group A) consisted of 21 clones with
intense to intermediate blue staining after growth in IL-3 but
a complete absence of p-galactosidase expression after GCSF stimulation. Group B (23 clones) showed an intermediate to low stainin;: in IL-3 that was downregulated to low
or no expression on differentiation. Typical staining of four
clones from group A, stimulated either with IL-3 (day 0) or
G-CSF (day 3), is shown in Fig 1. Group C (26 clones)
contained clones with very small and subtle differences in
p-galactosidase expression between the two stages of differentiation. None of the clones from group C were used in
later experiments. We repeated the experiment twice and
found consistent and reproducible patterns of expression and
downregulation. No clones expressed the lacZ gene on day
3 but downregulated on day 6, or vice versa. Also, we have
not observed any clones where the expression of the lacZ
gene has been upregulated in our assay.
IdentiJication of genes downregulated by G-CSF. To determine the number of proviral integrations in the 44 individual cell clones from group A (21 clones) and group B (23
clones) that showed a downregulated expression of the lac2
gene, cells were expanded in mass cultures, genomic DNA
was isolated and examined by Southern blotting. Analysis
of DNA with a neo probe revealed that the 21 cell clones
from Group A belonged to at least 8 different subgroups
(Fig 2A). Eleven clones had single integrations and seemed
to originate from four different clones. The remaining ten
clones had multiple integrations of which seven carried the
identical six integrations and must have originated from the
same parental clone. Taken together, the results identified
20 different integrations of which nine must represent unique
downregulated genes. The 23 clones from group B were
also examined by Southern blot analysis and some of them
showed multiple integrations as described for group A, ranging from two to six integrations (Fig 2B). However, most
clones contained single integrations. As was true for group
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
JONSSON ET AL
1774
cells.'J Because nuclear run-on transcripts contain both exons and introns they will hybridize to DNA probes containing either introns or exons.
We concentrated our efforts on the 20 integrations from
group A, representing the nine unique. downregulated genes.
Host DNA sequences flanking integration sites were isolated
by inverse PCR. The size of the PCR products ranged from
approximately 200 to 1,300 bp. of which 120 to 1.085 bp
contain specific host DNA sequence. Because we were not
able to amplify all twenty integrations after digestion with
Hinfl, we also amplified genomic DNA from nine cell lines
after digestion with Rsrr I. Fifteen of 21 fragments (719)
tested hybridized to nuclear run-on transcripts from 32D
cells (Fig 3A). indicating that integration had occurred to
transcribed regions of the genome. A clone containing a
mouse a-actin cDNA was used as a positive control (data
not shown). Two fragments derived from the same clone.
isolated after Hinfl and Rscr I digestion. respectively. hybridized strongly to radiolabelled total mouse genomic DNA
(Fig 3B).
A
1 3 5812
2526323334 40424449565860626367
,.------
A-'--
Fig 1. Photographs of X-Gal staining from four representative
LacZ+32D clones; P.l (A, B), P.3 (C, D), P.26 (E, F), P.32 (G, HI. Cells
were grown for 3 days in either 1% of an IL-3-containing supernatant
(Day 0) or in the presence of 100 UlmL of rhuG-CSF (Day 3). After
incubations as indicated, cells were stained with 5-bromo 4-chloro 3indolyl 0-galactoside (X-Gal) for analysis of 0-galactosidase activity.
A, some clones were identified several times. More importantly, based on the results from Southern blot analysis, only
one of the clones identified in group A was present within
group B (compare P.5 and the additional six clones with 6
integrations in Fig 2A with P.13 and P.45 in Fig 2B). The
fact that multiple isolates of clones were found in groups A
and B but that there was minimal overlap between the groups
indicates that the procedure was reproducible and that the
phenotypes distinguishing groups A and B are reproducible.
Seqirences upstream of proviral integrcrtions detect nuclear run-on transcript^. Previous reports from studies using promoter-trap retroviruses to identify active genes have
indicated that most integrations are close to exons in the
host DNA.'."'." Thus, the isolation of cellular DNA flanking
an expressed proviral integration site should automatically
lead to the isolation of parts of the coding sequence of the
cellular gene controlling the reporter gene in the virus and
should generate a probe suitable for detection of RNA transcripts by Northern blot analysis. Alternatively, integration
may have occurred in intron sequences and the probe derived
from inverse PCR may be too short to hybridize to mRNA.
We therefore decided to hybridize the PCR fragments to
radiolabeled nuclear run-on transcripts prepared from 32D
-
-..
B
C
Dra I
Dra 1
U3
////A
RU5
U3
RU5
lacZ
Fig 2. Enumeration of number of integrations by Southern blot
analysis in LacZ' 32D clones. (A) Clones with intense t o intermediate
blue staining after growth in IL-3 but a complete absence of /3-galactosidase expression after G-CSF stimulation. (B) Clones with intermediate t o low staining in IL-3 that was downregulated t o low or no
expression on differentiation. Numbers at top indicate name of the
different isolated LacZ' 32D clones. Ten micrograms of genomic DNA
was digested with 50 U of Dra Iand hybridized t o a 1.2-kb neo probe
(C).
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
GENE TRAPPING IN MYELOID PROGENITOR CELLS
1775
DNA .Seqiietice.S ide~ttjfi1 1 0 1 ~ l [ i t i d k1101~11g ~ r lrt'itli
~ . ~POteritinl i n ~ ~ ~ l i ~ c ~ in
t n emyeloid
rit
djferentiatiori. To charac-
R R H H ' R R H'R R H ' H R H ' R R ~ H ' H ' R R' I
1 2 3 4 7 8 910111315161718192021222324 25
IRR
A
B
Fig 3. Nuclear run-on analysis of host-specific upstream sequences adjacent to proviral integrations from LacZ' 32D clones.
Names of LacZ' 32D clones (top row) from which the different unique
flanking cellular DNA fragments were isolated (bottom row) are indicated. PCR-amplified upstream flanking sequences after digestion
with either Hinfl (HI or Rsa I (R) were hybridized to [3ZP1-labelednuclear transcripts preparedfrom 32D cells (A), and then the same filter
was rehybridizedto [3zPl-labeledtotal mouse genomic DNA (Bl.
Nortliern hlot cincilwis confirms the identificcition of clownregirlcitecl genes. We next examined the Ranking DNA sequences by Northern blot hybridization. Activity in this assay
would be direct evidence for integration near an exon of an
expressed gene. Poly(A)+-selected RNA was prepared from
32D cells obtained at different time points after the addition
of G-CSF. The results are summarized in Table 2. We have
screened 10 fragments that hybridized to nuclear run-on transcripts (F.I. F.2, F.3, F.lO, F.19, F.20. F.22. F.23, F.24, and
F.25). These represent eight cell clones and could therefore
contain eight of the nine downregulated genes we observed
by X-Gal staining. We also tested two fragments (F.1 1 and
F. 13) that did not hybridize to nuclear run-on transcripts and
confirmed that they did not contain expressed genes (data
not shown).
Fragments F.22, F.23, and F.25, all with single proviral
integrations, hybridized to 2.0 kb (F.23), 2.5 kb (F.22). or
7.0 kb (F.25) transcripts that appear to be downregulated
when 32D cells are induced to differentiate in G-CSF (Fig
4B, C, and E). F.23 and F.25 are only detected at low levels
in 32D cells on days 3 or 6 after stimulation with G-CSF
(Fig 4C and E); F.22 is expressed at a high level on day 3
but not on day 6 (Fig 4B). We also observed detectable
downregulation of expression for fragments F.2 (Fig 4A)
and F.24 (Fig 4D). F.2 detects a 2.2-kb transcript on all days
tested. F.24 detects three transcripts (8.0, 3.5, and 1.0 kb)
that behave differently following G-CSF (Fig 4D). None of
the transcripts change significantly by day 3 and the 3.5 kb
transcript is downregulated little, if any, by day 6. In contrast
the 8.0-kb transcript is markedly downregulated by day 6
and the 1 .O-kb transcript is undetectable by day 6.
terize the isolated DNA fragments and to identify the genes
that are downregulated on differentiation of 32D. we sequenced the PCR-products after cloning into Bluescript plasmids. Sequence analysis confirmed that all isolated DNA
fragments contained authentic junctions between viral and
cellular DNA (Fig 5). Thus each junction. as expected.
lacked the two first nucleotides (nt) of U3 in the LTR." The
sequence 5' of the provirus was unique in each clone (Fig
5 ) . thus confirming independent integration events. The five
isolated cellular DNA sequences hybridizing to RNA transcripts and four additional sequences hybridizing to nuclear
run-on transcripts were examined by computer analysis for
possible open reading frames and consensus splice junction
sequences, indicative of exons and intron-exon boundaries.
In addition. the sequences were compared to known sequences in the GenBank Data Base. using the BLASTN
and the BLASTX sequence analysis program.2hAll of the
sequences had an open reading frame (Table 3) ranging from
I 55 to 35 I nt. We have identified a complete open reading
frame with fragments F.2 and F.23 of 266 nt and 351 nt.
respectively. We have also found open reading frames within
F.22 (252 nt). F.24 (207 nt), and F.25 (219 nt). Fragments
F.22 and F.25 have consensus sequences typical for splice
junctions in the boundary between 3'-end of introns and 5'end of exons (data not shown).
BLAST analysis identified similarities to known genes in
the GenBank Data Base (Table 3). The greatest identity to a
known gene was between fragment F.24 and an endogenous
retroviral sequence in the mouse genome, called B-26, belonging to the murine leukemia virus group. In this case 404
of 426 nt (95%) were identical. Although B-26 has a striking
homology to parts of the Moloney murine leukemia virus.27
Table 2. Summary of Expression Analysis of Isolated Flanking
Cellular DNA Sequences
Flank Cell
No. Clone
F.l
F.2
F.3
F.4
F.9
F.10
F.ll
F.13
F.15
F.17
F.19
F.20
F.21
F.22
F.23
F.24
F.25
P.26
P.26
P.26
P.26
P.8
P.8
P.8
P.62
P.62
P.69
P.69
P.12
P.12
P.3
P.32
P.l
P.25
Integrations
in Parental
Cell Clone
6
6
6
6
4
4
4
3
3
2
2
2
2
1
1
1
1
Length of Proviral Hybridization to Hybridization
Flanking DNA
Nuclear Run-on
to Cellular
(nucleotides)
TranscriDts.
Transcriptst
131
266
385
685
340
440
395
1,066
225
120
380
340
300
415
351
1,085
885
+*
-*
+
+
+
-
ND
ND
ND
ND
ND
+
+
-
+++*
+
+
+
+
+
+
+
+
+
+
Abbreviation: ND, not done.
Flanking DNA sequences were separated by agarose gel electrophoresis,
blotted onto nylon membranes, and hybridized to ["PI-labeled nuclear run-on
transcripts prepared from nuclei of 32D cells.
t Expression of flanking cellular DNA sequences in uninfected 32D cells were
analyzed by Northern blot analysis.
t-, no expression; +, good expression; +++, strong expression.
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
1776
JONSSON ET AL
When fragment F.2 was analyzed. 97 nt of 266 nt were
1 2 3
64% identical to the mouse &glucuronidase gene, a primary
A
+ 2.2 kb
B
4- 2.5 kb
C
D
w
i
4-
7
+ 8.5kb
4-
a
.c.-
Y t
4-
2.0 kb
3.5 kb
3.0kb
1.0 kb
+ 7.0 kb
E
r)- 1.8 kb
F
1 2 3
Fig 4. Northern blot analysis of flanking cellular DNA sequences.
Two micrograms of polyA' RNA was selected from uninfected 320
cells after stimulation in 1% IL-3 supernatant (1). or after stimulation
in 100 UlmL G-CSF for 3 days (21 or 6 days (31, respectively, and
hybridized to isolated fragments F.2 (A), F.22 (BI, F.23 (Cl,F.24 ID),
and F.25 (E). A probe to human GADPH was used as a control to
quantify the amounts of mRNA (F).
a thorough sequence comparison showed that the homology
was not to the LTR in our provirus but to a sequence more
than 150 bp upstream of the proviral-DNA junction. Thus,
it is unlikely that the integration occurred through a recombination event between the retroviral vector and B-26.
granule marker secreted by neutrophils on stimulation by
various cytokines.2"3"The probability of similarity between
F.2 and the P-glucuronidase gene was significant with a P(N)
value of 5 lo-'.'' Fragment F.22 showed some identity (75%
of 44 nt) to PAX3, a family member of the paired box genes
in humans and mice that codes for a protein that functions
as a transcription factor and has been shown to be expressed
specifically in progenitors of the central nervous system"
and muscle cells." The sequence of fragment F.23 did not
show any significant similarity to DNA sequences in the
GenBank Data Base. However. when the amino acid sequence was compared to other known genes F.23 showed a
similarity to human proacrosin (64% identical of 31 amino
acids) expressed in spe~matocytes.~~.''
The homology was to
a proline-rich region of proacrosin that has been suggested
to contain a conserved motif responsible for interactions to
SH3 domains on other proteins.'"
DISCUSSION
In this study, we present results on the feasibility of using
promoter-trap retroviral vectors to identify developmentally
regulated genes during hematopoietic differentiation. By introducing a promoterless IacZ gene inserted in the 5'-end of
the LTR to an IL-3-dependent myeloid progenitor cell
clone, 32D, we have identified several transcriptionally active regions of the DNA. Only 0.5% of integrated proviruses
led to efficient expression of the lacZ gene. This frequency
agrees with published observations for 3T3 cells' and our
own studies with other hematopoietic cell lines (data not
shown). The estimated frequency also further underlines the
potential of using this promoter-trap retrovirus to identify
developmentally regulated genes. Three rounds of sorting
enriched the proportion of LacZ' cells to 58%. Of 492 LacZclones, 70 clones appeared to downregulate IacZ expression
in response to G-CSF. Of these, 44 clones showed a clear
and consistent downregulation and were used for further
experiments. Thus, in our in vitro culture system approximately 14% of the promoters identified by the reporter gene
were developmentally regulated (Table I ). This analysis indicates that proviral integration occurs in both regulated
genes and housekeeping genes and that every 1 of 1,000 to
1 of 2,000 integration will lead to the identification of a
developmentally regulated gene. This frequency indicates
that it will be necessary to screen large numbers of "true"
early hematopoietic progenitor cells from fresh BM to identify genes specifically downregulated during differentiation.
However, the ability to make high-titer virus stocks, and the
high efficiency of retroviral infection, makes this strategy a
feasible approach to search for new genes involved in early
hematopoietic differentiation.
Among the 70 clones showing downregulation, we focused our analysis on the 2 1 clones in group A that showed
the greatest downregulation. By Southern blot analysis (Fig
2A) these 21 clones represented nine unique clones. The fact
that the same clones were selected more than once supports
the use of promoter-trap retroviruses to identify regulated
genes during differentiation. Although infection by coculti-
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
1777
GENE TRAPPING IN MYELOID PROGENITOR CELLS
Proviral LTR
Upstream flanking host cellular DNA
U3
F.2
F.10
Fig 5. Nucleotide sequences
of approximately 60 nt from upstream flanking host cellular
DNA most adjacent to proviral
U3LacZ integrations. Deletion of
the 2 nt from the proviral LTR in
each junction is indicated.
F,ll
F.22
F.=
F,24
F.25
vation has led to multiple integration sites in some clones,
identification of the regulated site does not seem to be difficult because only 1 of 200 integrations are expressed. Results
from other investigators have shown that most integrations
giving expression of the lucZ gene are in close to or directly
in the coding ~equence.~."
Out of 21 fragments isolated from
the 9 unique clones, 15 fragments from 8 cell clones were
detected by nuclear run-on analysis.
Our data show that the isolation of flanking DNA from
proviral integrations with an actively transcribed reporter
gene by inverse PCR is an efficient and reliable method to
generate probes for subsequent analysis of the expression
pattern and actually cloning of a developmentally regulated
gene. Northern blot analysis of undifferentiated ( L 3 dependent) and differentiated (G-CSF dependent) 32D cells, respectively, using 10 PCR-generated probes showed that at
least fragments F.22, F.23, and F.25, all with single integrations, were clearly downregulated on differentiation (Fig 4B,
C, and E). Fragment F.24 is also differentially regulated
because the three transcripts are downregulated after stimulation with G-CSF (Fig 4D). One additional fragment (F.2)
detects a 2.2-kb transcript on all days tested (Fig 4A). Analysis of five fragments have failed to detect any regulated
transcripts. However, because these fragments were isolated
from cell clones with multiple integrations, negative fragments were expected.
Finally, analyses of the DNA sequences of the expressed
genes have identified several novel genes downregulated on
neutrophilic differentiation of the 32D myeloid progenitor
cell line. Despite rather short isolated flanking DNA sequences, we are encouraged by the fact that all sequences
contained complete or partial open reading frames. Moreover, among the six sequences with complete open reading
frames andor splice junctions are the five flanking sequences
that hybridized to cellular transcripts on Northern blot analysis. Comparison of isolated flanking DNA sequences to the
GenBank Data Base has for some of our cloned DNA sequences identified similarities to known genes (Table 3).
F.24 has a 95% homology (404 of 426 nt) to an endogenous
retroviral sequence in the mouse, B-26F7It has been reported
that B-26-specific transcripts of 8.4 and 3.0 kb are transcribed in different mouse tissues with strong expression in
some hematopoietic tissues, eg, thymus and spleen." Thus,
the size of the mRNAs of B-26 fits our results from the
Northern blot analysis. It remains to be seen if its expression
is developmentally regulated during hematopoietic differentiation as is the case for the human endogenous retrovirus
(ERV)37and some mouse ERV.38
Of the other DNA sequences that we have shown to be
regulated in 32D cells, a similarity was found in three cases
(Table 3). Ninety-seven nt within the 266-nt open reading
frame of F.2 showed 64% identity to the mouse P-glucuroni-
Table 3. Sequence Analysis of Cellular DNA Sequences Flanking Retroviral Integrations From 32D Cells and Their Similarities to Known
Genes in the GenBenk Data Base
Known Gene
Flank
ORF'
Descriptiont
Species
Overlap Length*
F.2
F.10
F.11
F.22
F.23
1-266
240-438
161-316
163-415
1-351
p-Glucuronidase gene, exon 3 and intron 3
Adenine phosphoribosyltransferase (APRT) gene
Atrial natriuretic factor gene (ANF), 5'-end
Paired box homeotic protein (PAXB), exon 8
Proacrosin protein
F.24
F.25
784-990
22-240
Endogenous retroviruslike DNA, 6-26
Coagulation factor X gene, 5' flank
Mouse
Mouse
Rat
Human
Human
Mouse
Human
97 nt
139 nt
136 nt
44 nt
31 aa
426 nt
65 nt
96 Identity5
~
64
77
75
75
64
95
64
*Sequence range of the longest open reading frame found in proviral flanking cellular DNA sequences.
t Name of known gene products present in the GenBank Data Base with the highest degree of similarities to the query flanking sequence.
The Genbank accession number to the different gene products are J02836 (mouse beta-glucuronidase gene), M86440 (mouse APRT), J03267
(rat ANF), U12259 (human PAX3). S23500 (human acrosin precursor), M23458 (LTR sequences of 826 provirus), and M92096 (human coagulation
factor X).
Numbers of nt within the known gene product with sequence similarity to the flanking cellular DNA sequence
§ Percentage of nt with a complete match in the overlapping sequence of the known gene product.
*
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
1778
JONSSON ET AL
dase gene, a primary granule marker secreted by neutrophils
when stimulated by cytokines.28-'0 Most granule genes are
synthesized at the progenitor stage and are then gradually
downregulated on terminal differentiation. This finding
agrees with our observation that F.2 detects transcripts on
all days tested but appears to be downregulated on day 6
(Fig 4A). F.22 and the human homeobox-containing gene
PAX3 were 75% identical in a sequence of 44 nt. PAX3 has
been specifically implicated in the development of the central
nervous system32but is also expressed in primitive muscle
progenitor cells.33However, the homology between F.22 and
PAX3 is restricted to a region not yet described as a functional domain (data not shown). Recently a second isoform
of the human PAX3 was isolated that showed expression in
most adult tissues tested.39 Further studies are necessary to
determine the role of this gene in the proliferation and differentiation of 32D cells and other myeloid progenitor cells.
When the amino acid sequence of F.23 was compared with
other known genes, a similarity to human proacrosin was
found (64% identical of 31 amino acids). Proacrosin is the
zymogen form of the serin-protease acrosin with specific
expression in spermatocytes and has been suggested to be
involved in the fertilization p r o ~ e s s . ' ~Interestingly,
,~~
a homology was found between a proline-rich region of F.23 and
the proline-rich region of proacrosin that has been suggested
to be a common motif in the ligand binding of proteins to
SH3 domains on other proteins and therefore to be involved
in signal transduction.3h
In conclusion, we have identified at least three genes apparently regulated during hematopoietic differentiation of
the myeloid lineage by the use of a promoter-trap retrovirus.
We have used a myeloid progenitor cell line in vitro to show
the feasibility of such an approach and have identified several
genes that are downregulated on the transition from growth
in IL-3 to terminal differentiation in G-CSF. We also show
how efficiently host specific DNA sequences can be isolated
by inverse PCR and be used as probes for the study of gene
expression and sequence information. Future analysis will
resolve the importance of these genes in the process of myeloid differentiation. Regardless of the outcome of such studies, the identification of these genes gives valuable new information regarding the regulation
of myeloid
differentiation. With this retroviral vector, we now have a
tool to identify and isolate genes in primary hematopoietic
progenitor cells from the adult BM.
ACKNOWLEDGMENT
We thank H. Chen, G. Knowles, and C. Spencer for excellent
technical assistance, and F. Bittkowski for expert photographic work.
We are grateful to Drs F. Melchers, G. Rovera, E. Ruley, and H.
von Melchner for kindly providing invaluable cell clones; to the
Genetics Institute (Cambridge, MA) for the generous gift of hematopoietic growth factors; and to Drs R. Bremner, L-G. Larsson, E.
Ruley, and F. Sablitzky for helpful and critical discussions.
REFERENCES
1. Phillips RA: Hematopoietic stem cells: Concepts, assays, and
controversies. Semin Immunol 3:337, 1991
2. Metcalf D: Hematopoietic regulators: Redundancy or subtlety?
Blood 82:3515. 1993
3. Cumano A, Paige CJ, Iscove NN, Brady G: Bipotential precursors of B cells and macrophages in murine fetal liver. Nature
356:612, 1992
4. Hirayama F, Shih JP, Awgulewitsch A, Wan- GW, Clark SC,
Ogawa M: Clonal proliferation of murine lymphohemopoietic progenitors in culture. Proc Natl Acad Sci USA 89:5907, 1992
5. Reddy S, DeGregori JV, von Melchner H, Ruley HE: Retrovirus promoter-trap vector to induce lacZ gene fusions in mammalian
cells. J Virol 65:1507, 1991
6. Gossler A, Joyner AL, Rossant J, Skarnes WC: Mouse embryonic stem cells and reporter constructs to detect developmentally
regulated genes. Science 244:463, 1989
7. Friedrich G, Soriano P: Promoter traps in embryonic stem cells:
A genetic screen to identify and mutate developmental genes in
mice. Genes Dev 5:1513, 1991
8. Reddy S, Raybum H, von Melchner H, Ruley HE: Fluorescence-activated sorting of totipotent embryonic stem cells expressing
developmentally regulated lacZ fusion genes. Proc Natl Acad Sci
USA 89:672 I , 1992
9. Skarnes WC, Auerbach BA, Joyner AL: A gene trap approach
in mouse embryonic stem cells: The lacZ reporter is activated by
splicing, reflects endogenous gene expression, and is mutagenic in
mice. Genes Dev 6:903. 1992
IO. von Melchner H, DeGregori JV, Rayburn H, Reddy S. Friedel
C, Ruley HE: Selective disruption of genes expressed in totipotent
embryonal stem cells. Genes Dev 6:919. 1992
I I . von Melchner H, Reddy S, Ruley HE: Isolation of cellular
promoters by using a retrovirus promoter trap. Proc Natl Acad Sci
USA 87:3733, 1990
12. Kerr WG, Nolan GP, Johnsen JB, Herzenberg LA: In situ
detection of stage-specific genes and enhancers in B cell differentiation via gene-search retroviruses. Adv Exp Med Biol 292: 187, 1991
13. Sablitzky F, Jonsson J-I, Cohen BL, Phillips RA: High frequency expression of integrated proviruses derived from enhancertrap retroviruses. Cell Growth Differen 4:451, 1993
14. Greenberger JS, Eckner RJ, Sakakeeney M, Marks P, Reid D,
Nabel G, Hapel A, Ihle JN, Humphries KC: Interleukin 3-dependent
hematopoietic progenitor cell lines. Federation Proc 42:2762, 1983
15. Valtieri M, Tweardy DJ, Caracciolo D, Johnson K, Mavilio F.
Altmann S, SantoIi D, Rovera G: Cytokine-dependent granulocytic
differentiation. Regulation of proliferative and differentiative responses in a murine progenitor cell line. J Immunol 138:3829, 1987
16. Migliaccio G, Migliaccio AR, Kreider BL, Rovera G, Adanison JW: Selection of lineage-restricted cell lines immortalized ut
different stages of hematopoietic differentiation from the murine cell
line 32D. J Cell Biol 109:833, 1989
17. Karasuyama H, Melchers F: Establishment of mouse cell lines
which constitutively secrete large quantities of interleukin 2, 3, 4 or
5, using modified cDNA expression vectors. Eur J Immunol 18:97.
1988
18. Magli MC, Dick JE, Huszar D, Bernstein A, Phillips RA:
Modulation of gene expression in multiple hematopoietic cell lineages following retroviral gene transfer. Proc Natl Acad Sci USA
84:789, 1987
19. Nolan GP, Fiering S, Nicolas J-F, Herzenberg LA: Fluorescence-activated cell analysis and sorting of viable mammalian cells
based on 0-D-galactosidase activity after transduction of Escherichia
coli IacZ. Proc Natl Acad Sci USA 85:2603, 1988
20. Price J, Turner D, Cepko C: Lineage analysis in the vertebrate
nervous system by retrovirus-mediated gene transfer. Proc Natl Acad
Sci USA 84:156, 1987
21. Miller SA, Dykes DD, Polesky HF: A simple salting out
procedure for extracting DNA from human nucleated cells. Nucleic
Acids Res 16:1215, 1988
22. Sambrook J, Fritsch EE, Maniatis T: Molecular Cloning: A
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
GENE TRAPPING IN MYELOID PROGENITOR CELLS
Laboratory Manual. Cold Spring Harbor, NY, Cold Spring Harbor
Laboratory, 1989
23. van Lohuizen M, Verbeek S, Scheijen B, Wientjens E, van
der Gulden H, Bems A: Identification of cooperating oncogenes in
E@-myctransgenic mice by provirus tagging. Cell 65:737, 1991
24. Nevins J R Isolation and analysis of nuclear RNA, in Berger
SL, Kimmel AR (eds): Methods of Enzymology, Guide to Molecular
Techniques, vol 152. Orlando, FL,Academic, 1987, p 234
25. Chomczynski P, Sacchi N: Single step method to isolate total
RNA. Anal Biochem 162:156, 1987
26. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ:
Basic local alignment search tool. J Mol Biol 215:403, 1990
27. Obata MM, Khan AS: Structure, distribution, and expression
of an ancient murine endogenous retroviruslike DNA family. J Virol
62:4381, 1988
28. Boey H, Rosenbaum R, Castracane J, Borish L: Interleukin4 is a neutrophil activator. J Allergy Clin Immunol 83:978, 1989
29. Borish L, Rosenbaum R, Albury L, Clark S: Activation of
neutrophils by recombinant interleukin 6. Cell Immunol 121:280,
1989
30. Fabian I, Kletter Y, Mor S, Geller BC, Ben YM, Volovitz B,
Golde DW. Activation of human eosinophil and neutrophil functions
by haematopoietic growth factors: comparisons of IL-1, IL-3, IL-5
and GM-CSF. Br J Haematol 80: 137, I992
31. O n SL, Hughes TP, Sawyers CL, Kat0 RM, Quan SG, Williams SP, Witte ON, Hood L: Isolation of unknown genes from
1779
human bone marrow by differential screening and single-pass cDNA
sequence determination. Proc Natl Acad Sci USA 91:11869, 1994
32. Chalepakis G, Stoykova A, Wijnholds J, Tremblay P, Gruss P.
Pax: gene regulators in the developing nervous system. J Neurobiol
2 4 1367, 1993
33. Bober E, Franz T, Arnold HH, Gruss P, Tremblay P. Pax-3
is required for the development of limb muscles: A possible role for
the migration of dermomyotomal muscle progenitor cells. Development 120:603, 1994
34. Nayemia K, Nieter S, Kremling H, Oberwinkler H, Engel
W: Functional and molecular characterization of the transcriptional
regulatory region of the proacrosin gene. J Biol Chem 269:32181,
1994
35. Raab LS, Hamilton DW, Hancock LW: Proacrosin gene expression in rat spermatogenic cells. J Androl 15:244, 1994
36. Yu H, Chen JK, Feng S, Dalgarno DC, Brauer AW, Schreiber
SL: Structural basis for the binding of prolin-rich peptides to SH3
domains. Cell 76:933, 1994
37. Larsson E, Kat0 N, Cohen M: Human endogenous proviruses,
in Current Topics in Microbiology and Immunology, vol 148. Berlin,
Germany, Springer-Verlag. 1989, p 115
38. Jenkins NA, Copeland NG: Retroviral DNA content of the
mouse genome. Birth Defects 23:109, 1987
39. Tsukamoto K, Nakamura Y, Niikawa N: Isolation of two
isoforms of the PAX3 gene transcripts and their tissue-specific alternative expression in human adult tissues. Hum Genet 93:270, 1994
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
1996 87: 1771-1779
Use of a promoter-trap retrovirus to identify and isolate genes
involved in differentiation of a myeloid progenitor cell line in vitro
JI Jonsson, Q Wu, K Nilsson and RA Phillips
Updated information and services can be found at:
http://www.bloodjournal.org/content/87/5/1771.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.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.