Immunity, Vol. 22, 43–57, January, 2005, Copyright ©2005 by Elsevier Inc.DOI 10.1016/j.immuni.2004.11.014
CD34 and CD43 Inhibit Mast Cell Adhesion
and Are Required for Optimal
Mast Cell Reconstitution
1995; Krause et al., 2001). Gene-targeting experiments
have further obscured this issue; deletion of the CD34-
encoding gene in mice by two separate laboratories
has revealed only subtle phenotypes, with different and
opposing observations made for the independent
strains (Cheng et al., 1996; Suzuki et al., 1996). The
discovery of two additional CD34-related genes (Podo-
calyxin and Endoglycan) with overlapping expression
patterns suggests that the minor phenotypes observed
in CD34-deficient mice may reflect functional compen-
Nielsen et al., 2002; Sassetti et al., 1998, 2000). Thus,
the use of mutant animals to delineate a function for
CD34 would require either the generation of compound
mutant mice lacking these related family members or a
focused analysis of tissues that normally express CD34
but not its homologs.
In a recent survey of the expression pattern of CD34-
related genes, we made an unexpected observation:
CD34 but not Podocalyxin or Endoglycan is expressed
at high levels by murine mast cells (Drew et al., 2002).
Despite their notoriety in the pathogenesis of diseases
noist and Mathis, 2002), mast cells are also important
players in innate and adaptive immune responses (Galli
and Nakae, 2003; Kawakami and Galli, 2002; Smith and
Weis, 1996). Murine mast cells can be broadly classified
into two distinct groups based on their morphology,
granule protease content, and tissue localization: con-
nective tissue-type mast cells (CTMC) and mucosal
mast cells (MMC) (Gurish and Boyce, 2002). CTMC re-
side predominantly in the skin and peritoneum and are
while MMC reside in the intestinal mucosa and play a
ish et al., 2001). Mast cells are known to phagocytose,
matory mediators (Mekori and Metcalfe, 2000). There-
fore, these cells are important in immune responses to
a variety of pathogens.
Mast cells arise in the bone marrow from hematopoi-
etic stem cells and are unusual in that they leave the
wakami and Galli, 2002). Little is known of the process
that leads to their maturation, but, in vivo, it is critically
dependent upon signaling through the stem cell factor
receptor (SCF), c-kit (also known as SCF-R and mast
cell growth factor receptor) (Galli, 2000). The develop-
ment and differentiation of a mast cell, including its
granule content, size, and sensitivity to stimulants is
dependent upon growth factors and cytokines present
in its local microenvironment (Gurish and Boyce, 2002;
Mekori and Metcalfe, 2000).
The W/Wvmouse possesses two mutations for c-kit
and is widely used for the study of mast cell function,
since these mice virtually lack tissue mast cells (Galli,
2000). c-kitWencodes a shortened protein lacking the
transmembrane domain, thereby ablating its expression
on the cell surface (Hayashi et al., 1991; Waskow et
Erin Drew, Jasmeen S. Merzaban, Wooseok Seo,
Hermann J. Ziltener, and Kelly M. McNagny*
The Biomedical Research Centre
University of British Columbia
Vancouver, British Columbia V6T 1Z3
CD34 is a cell-surface sialomucin expressed by hema-
topoietic stem cells (HSC), mast cells, and vascular
endothelia. Despite its popularity as an HSC marker,
the function of CD34 on hematopoietic cells remains
enigmatic. Here, we have addressed this issue by ex-
amining the behavior of mutant mast cells lacking
CD34, the related sialomucin, CD43, or both mole-
cules. Loss of these molecules leads to a gene-dose-
dependent increase in mast cell homotypic aggrega-
tion with CD34/CD43KOs ? CD43KO ? CD34KO ?
wild-type. Importantly, reexpression of CD34 or CD43
in these cells caused reversal of this phenotype.
Furthermore, we find that loss of these sialomucins
prevents mast cell repopulation and hematopoietic
precursor reconstitution in vivo. Our data provide
clear-cut evidence for a hematopoietic function for
CD34 and suggest that it acts as a negative regulator
of cell adhesion.
CD34 is a highly glycosylated transmembrane protein
that is expressed by hematopoietic stem cells (HSC),
hematopoietic progenitors, and vascular endothelia.
Since its discovery almost 20 years ago, it has become
the most widely used marker for the enrichment of hu-
man hematopoietic progenitors (reviewed in Gratama et
al., 2001; Krause et al., 1996). More recently, its use
for the enrichment of the most primitive hematopoietic
progenitors/stem cells has come into question with the
discovery that both CD34?and CD34?cells have the
capacity to reconstitute all hematopoietic lineages in
irradiated murine recipients (Bhatia et al., 1998; Ogawa,
2002). Interestingly, CD34 expression by hematopoietic
stem cells has been found to fluctuate as a function of
mammalian development and cell activation (Dao et al.,
2003; Ogawa, 2002; Zanjani et al., 2003). Thus, it is likely
that the differences in the expression of CD34 by HSC
reflect their activation or developmental status.
Given the widespread use of CD34 as a progenitor
cell marker, it is remarkable that little is known about
the function of this antigen. In the literature, CD34 has
alternatively been proposed to act as (1) an enhancer
of proliferation, (2) a blocker of differentiation, (3) a bone
marrow homing receptor, (4) a cell adhesion molecule,
and (5) a blocker of cell adhesion (Baumhueter et al.,
1993; Cheng et al., 1996; Delia et al., 1993; Fackler et al.,
Figure 1. Structure of CD34 and CD43
Schematic structure of full-length murine
CD34 (CD34FL), the naturally occurring trun-
cated form(CD34CT), and CD43 basedon pre-
dicted protein sequences. Blue boxes, mucin
domains; green boxes, cysteine-rich domain;
black circles, potential N-linked carbohy-
drates; horizontal bars, potential O-linked
carbohydrates; PKC, CK2, and TK, potential
phosphorylation sites; DTEL, potential PDZ-
domain docking motif.
al., 2004). These mice also harbor the c-kitWvallele, a
missense mutation at position 660 in the kinase domain
(threonine to methionine) (Dastych et al., 1998; Nocka
et al., 1990), resulting in weak kinase activity and poor
response to SCF (Waskow et al., 2004). W/Wvmice have
a normal number of stem cells but decreased numbers
of progenitor cells and are anemic (Migliaccio et al.,
1999). These mice can be transplanted with wild-type
bone marrow to restore mast cell populations and cure
their anemia (Williams and Galli, 2000). Since wild-type
nizing stem cell niches in W/Wvrecipients (Migliaccio
et al., 1999).
To explore the function of CD34 on hematopoietic
but not its closest homologs is expressed by murine
mast cells. We compared bone marrow-derived mast
cell (BMMC) cultures from wild-type and cd34?/?mice
and observed a clear increase in the homotypic adhe-
siveness of CD34-deficient mast cells. In an attempt to
exacerbate this phenotype, we generated compound
mutant mice lacking CD34 and the distantly related,
sialomucin, CD43 (Figure 1). CD43 is expressed by most
hematopoietic cells, and while it has previously been
shown to play a role in blocking leukocyte adhesion
(Manjunath et al., 1995; Ostberg et al., 1998), its function
has never been examined on mast cells. Our results
show that mast cells lacking CD34 and/or CD43 show
increased aggregation, with mast cells lacking both si-
increased aggregation was reversible by the ectopic
reexpression of CD34 or CD43. To assess the role of
these molecules in vivo, we measured the ability of wild-
type and mutant cells to repopulate a mast cell niche,
using two experimental model systems. These experi-
ments revealed a complete impairment of cd34?/?/
cd43?/?cells to repopulate peritoneal mast cells. Inter-
estingly, in one of these experimental models, we ob-
served a similar decreased ability of mutant cells to
reconstitute bone marrow hematopoietic progenitors.
Our results show that CD34 and CD43 play a key role
in blocking mast cell adhesion and suggest that their
function on these cells is to enhance their ability to
migrate to new microenvironments.
Loss of CD34 Leads to Increased Homotypic
Adhesion of Bone Marrow-Derived Mast Cells
To test the function of CD34 on mast cells and their
precursors, bone marrow from cd34?/?mice (Suzuki et
al., 1996) and wild-type mice was cultured in vitro under
conditions that select for the outgrowth and maturation
of mast cells. Although loss of CD34 had no effect on
the kinetics of mast cell differentiation, proliferation, or
degranulation (Drew et al., 2002) (data not shown), light
microscopic analyses of cultures from wild-type and
cd34?/?littermates revealed one consistent difference:
CD34-deficient mast cells had an increased propensity
to aggregate. Wild-type cultures consistently grew as
single-cell suspensions, while mast cells lacking the
CD34 and CD43 Block Mast Cell Adhesion
Figure 2. LossofCD34IncreasesHomotypic
Aggregation of BMMC
Bar graphs represent the percentage of cells
involved in aggregates (open bars) and the
average number of cells per aggregate (filled
bars) with standard deviation error bars. Val-
ues are indicated as significantly different
from wild-type or vector alone with (**p ?
(A) BMMC from littermate and sex-matched
wild-type andcd34?/?mice. Graphsshow de-
gree of aggregation from two independently
derived sets of littermate-matched mast
(B) cd34?/?BMMC expressing the GFP-con-
taining retrovirus pMXpie alone or pMXpie
containing CD34FLor CD34CT. Data show ag-
gregation quantification for two indepen-
CD34 antigen showed small aggregates (8 ? 2 cells per
aggregate) with 16% ? 7% of cells involved in homo-
typic adhesion (Figure 2A). Greater than 1500 cells were
counted in this and subsequent assays to ensure these
differences were consistent. Similar results were ob-
tained when cells were mechanically dissociated into
single-cell suspensions, plated at the same density, and
observed after overnight incubation.
To verify that loss of CD34 was responsible for this
observed aggregation, cd34?/?bone marrow cells were
(pMXpie) or with the same retrovirus expressing either
full-length (CD34FL) or cytoplasmically truncated CD34
(CD34CT). The latter form corresponds to a naturally oc-
curring splice variant of CD34 and results in the deletion
of the bulk of the cytoplasmic domain via a premature
stop codon (Nakamura et al., 1993) (see Figure 1). Cells
were cultured for at least 6 weeks in media containing
IL-3 and puromycin and sorted by flow cytometry for
cells for homotypic adhesion by fluorescence micros-
19% ? 7% of cells formed aggregates with 7 ? 2 cells
per clump. Strikingly, virtually no clumping was ob-
served for cells expressing either CD34FL(2% ? 5%
aggregation and 2 ? 3 cells per clump) or CD34CT
(0.1% ? 0.3% aggregates, with 1 ? 1 cell per clump).
The reexpression of recombinant CD34 by these cells
after infection was confirmed by flow cytometry (data
not shown). Our data suggest that CD34 is necessary
and sufficient to inhibit mast cell aggregation.
Loss of the Sialomucin, CD43, Also Leads
to Homotypic Aggregation/Adhesion
Our observation that CD34 confers an antiadhesive
showing that CD43, a distantly related sialomucin ex-
pressed on most hematopoietic lineage cells, also is
junath et al., 1995; Ostberg et al., 1998; Walker and
Green, 1999). Although CD43 lacks many of the bio-
chemical motifs associated with the CD34 family (Doy-
onnas et al., 2001), it shares a large extracellular mucin
domain (Figure 1). To test whether this molecule plays
a similar role in blocking mast cell aggregation, we es-
tablished mast cell cultures from the bone marrow of
cd43?/?and cd34?/?/cd43?/?double-deficient mice. Al-
though we found no difference in the ability of CD43-
deficient mast cells to differentiate (data not shown),
these cells were much more prone to aggregation than
either cd34?/?or wild-type mast cells (70% ? 20% ver-
sus 16% ? 7% or 3% ? 2%, respectively; Figure 3A).
thanthose observedin culturesof eithercd34?/?orwild-
type mast cells (30 ? 15 cells/aggregate versus 8 ? 2
and 4 ? 2, respectively). Interestingly, loss of CD34 and
CD43 had an additive effect on homotypic adhesion:
94% ? 7% of the cd34?/?/cd43?/?mast cells were in
aggregates, and the size of these aggregates was much
ing either cd43?/?or cd34?/?(94 ? 50 cells/aggregate
versus 30 ? 15 and 8 ? 2, respectively; Figure 3A).
Again, we saw no obvious defect in the proliferation or
differentiation capacity of these double-deficient cells.
To test whether the increased aggregation reflects a
general increase in nonspecific cell adhesion, we exam-
ined the ability of wild-type and cd34?/?/cd43?/?mast
cells to bind to fibronectin in vitro. Cd34?/?/cd43?/?mast
cells had a higher propensity to bind this extracellular
matrix, without stimulation, than wild-type mast cells (see
Supplemental Figure S1 at http://www.immunity.com/cgi/
content/full/22/1/43/DC1/). We conclude that CD34 and
CD43 both play a role in blocking cell aggregation and
that the loss of both molecules has an additive effect
on cell adhesion.
To test whether CD34 and CD43 are functionally re-
dundant on mast cells, cd43?/?bone marrow was in-
fected with either a control virus vector or with vectors
expressing CD34FLor CD34CT(Figure 3B). Ectopic ex-
pression of CD34FLand CD34CTled to a partial reversal
of the aggregation phenotype over controls, with a de-
crease from 76% ? 6% aggregation and 31 ? 10 cells
per aggregate (vector-alone cells) to 41% ? 19% for
CD34FL(14 ? 5 cells per aggregate) and 27% ? 2% for
CD34CT(14 ? 6 cells per aggregate). CD43 was also
introduced into cells using pMXpie and showed com-
plete reversion of aggregation (data not shown).
Overexpression of CD34FLin cd34?/?/cd43?/?mast
cells also led to a decrease in aggregation from 98% ?
1% and 45 ? 11 cells per aggregate (vector control) to
74% ? 17% and 31 ? 16 cells per aggregate (Figure
3C). CD34CTdecreased adhesion to 6% ? 4% aggrega-
tion(4 ?2cells peraggregate;Figure3C). EctopicCD43
expression led to complete disaggregation of cd34?/?/
cd43?/?mast cells to 2.3% ? 0.9% (Figure 3C). In sum-
mary, our data suggest that, on mast cells, CD34 and
CD43 have overlapping functions in blocking adhesion
and that the loss of either or both of these molecules
can be reversed by the ectopic reexpression of CD34
agents compared to controls, and aggregates failed to
accumulate during overnight incubation in the presence
sion of sialomucin-deficient mast cells is a process that
is dependent upon divalent cations and may involve
integrins. A similar effect was also observed for cd34?/?
and cd43?/?BMMC (data not shown).
ates this aggregation, cd43?/?or cd34?/?/cd43?/?cells
were disaggregated mechanically and incubated with
blocking antibodies to ?1-, ?2-, ?4-integrins, ICAM-1, or
SgIGSF (Ito et al., 2003). None of these antibodies were
able to disrupt reaggregation of these BMMC (data not
shown). This suggests that either another molecule is
receptors are involved, and this effect cannot be over-
come by blocking only one of these proteins.
To test whether this antiadhesive activity on wild-
type mast cells was cell autonomous, we labeled either
cd34?/?/cd43?/?or cd34?/?BMMC with the red fluores-
cent dye CSA-SE and mixed them with cd34?/?mast
cells infected with GFP and CD34CT. None of the GFP/
CD34-positive cells aggregated with CSA-SE-positive
cd34?/?/cd43?/?or cd34?/?cells (Figure 4B). From these
experiments, we conclude that the negative regulation
of cellular adhesion by CD34 is cell autonomous and
that it is unable to lead to the disaggregation of cells
CD34 and/or CD43 Are Required for Optimal Mast
Cell Progenitor Migration In Vivo
To determine whether the loss of CD34 and CD43 affect
the ability of mast cell precursors to migrate to periph-
eral tissues, we quantified the number of mature mast
cells within different tissues of wild-type, cd34?/?,
cd43?/?, and cd34?/?/cd43?/?mice. CTMC were enumer-
ated using toluidine blue and Giemsa stains in the skin
of the ear and back and in the peritoneal cavity, respec-
in the mucosa of the intestine. Surprisingly, we found a
similar number of connective tissue and mucosal mast
cells in each tissue examined, regardless of their cd34
or cd43 genotype (Figure 5A). Thus, our data suggest
that, despite the profound differences in mast cell adhe-
sion observed in vitro, CD34- and/or CD43-deficient
mast cell progenitors migrate to tissues during develop-
ment in sufficient numbers to result in the appropriate
frequency of mature tissue mast cells in adult mice. We
also assessed the frequency of mast cell progenitors
(MCp) per 106mononuclear cells (MNC) in adult mice
(Crapper and Schrader, 1983; Gurish et al., 2001). There
row or spleen between wild-type, cd34?/?, cd34?/?, and
cd34?/?/cd43?/?mice (Figure 5B).
In order to evaluate the kinetics of mast cell precursor
migration into adult tissues, we used intraperitoneal in-
jection of distilled water to ablate the local CTMC and
then monitored the mast cell repopulation of this com-
partment over time. This technique has been previously
shown to eradicate peritoneal mast cells (about 2% of
the resident peritoneal hematopoietic cell population)
and stimulate the influx of mast cell progenitors from
Homotypic Adhesion Is Divalent Cation Dependent
and Cell Autonomous
Cellular adhesion mediated by integrins is dependent
upon the presence of divalent cations (Leitinger et al.,
2000). In an effort to determine whether the enhanced
aggregation of mucin-deficient mast cells is cation de-
pendent, mast cells were cultured in media supple-
mented with 1 mM EDTA and/or 1 mM EGTA in vitro.
After 10 min in culture, cd34?/?/cd43?/?cells showed no
aggregation in the presence of either or both chelating
CD34 and CD43 Block Mast Cell Adhesion
Figure 3. CD34 and CD43 Both Block BMMC Aggregation
Bar graphs represent the percentage of cells involved in aggregates (open bars) and the average number of cells per aggregate (solid bars)
with standard deviation error bars. Values are indicated as significantly different from wild-type or vector alone (*p ? 0.05 or **p ? 0.01).
(A) Aggregation of cd43?/?and cd34?/?/cd43?/?BMMC. Mast cells were independently derived from two mice of each genotype.
(B) Aggregation of cd43?/?BMMC expressing the GFP-containing retrovirus pMXpie alone with pMXpie containing CD34FLor CD34CT. Data
show results from two independent infections.
(C) Aggregation of cd34?/?/cd43?/?BMMC expressing pMXpie alone or pMXpie containing CD34FL, CD34CT, or CD43.
Figure 4. Homotypic Adhesion Is Divalent
Cation Dependent and Cell Autonomous
(A) Cellular aggregation is completely lost
upon addition of 1 mM EDTA and/or EGTA.
Micrographs show double-deficient cd34?/?/
cd43?/?BMMC in the absence and presence
of these chelators.
labeled with CSA-SE and were mixed with
GFP-positive CD34CTexpressing BMMC. Cells
were allowed to aggregate overnight and
were observed under fluorescence micros-
copy to assess degree of mixed aggregation.
the bone marrow without drastically altering the total
number of peritoneal cells (Kanakura et al., 1988). Dis-
tilled water was injected into the peritoneum of wild-
type, cd34?/?, cd43?/?, and cd34?/?/cd43?/?mice, and
the frequency of resident mast cells in peritoneal lavage
(Figure 5C). Wild-type mice showed an immediate abla-
tion of mast cells after water injection and required 21
weeks to recover a near-normal frequency of mast cells.
Cd34?/?/cd43?/?mice showed a near complete absence
of mast cell recovery even after 21 weeks (p ? 0.002).
Although the recovery of mast cells in cd34?/?or cd43?/?
cd43?/?levels, it was not low enough to be considered
statistically different from wild-type mice (p ? 0.61 and
0.06, respectively; Figure 5C). We conclude that al-
though cd34?/?/cd43?/?mice show no dramatic de-
crease in the frequency of mast cells in adult tissues at
steady state, after mast cell ablation they display a clear
defect in the kinetics of mast cell progenitor migration
into the peritoneum, consistent with a requirement for
these molecules for efficient homing in vivo.
in vivo. W/Wvmice were sublethally irradiated and in-
jected with wild-type bone marrow cells mixed with
equal numbers of either cd34?/?, cd43?/?, or cd34?/?/
cd43?/?bonemarrow cells in acompetitive repopulation
assay. As a control, W/Wvmice were also injected with
cd34?/?/cd43?/?bone marrow cells alone (Figure 6A). To
distinguish between the two donor populations, some
mice were injected with CD45.1 wild-type cells and
mutant cells (CD45.2). CD45.1 is a variant of CD45, a
pan-hematopoietic marker, and is commonly used for
distinguishing between donor populations in stem cell
transplantation experiments (Hasumura et al., 2003).
Eleven to twelve weeks after injection, the frequency
based on their distinctive high side-scatter properties
(as a measure of cell granularity) and their high expres-
sion of c-kit (Figure 6B). C-kithiperitoneal cells from
confirming their identity as mast cells (data not shown).
Wild-type, cd34?/?/cd43?/?, and W/Wvmice were used
neal cells were virtually undetectable in nonreconstitu-
ted W/Wvmice (Figure 6B). Interestingly, mice reconsti-
tuted with cd34?/?/cd43?/?bone marrow alone did not
show significant peritonealmast cell reconstitution (Fig-
ure 6B), suggesting a potent defect in their ability to
populate vacant mast cell niches in the peritoneum. In
contrast, significant numbers of mast cells were found
in W/Wvmice that had been competitively reconstituted
with wild-type and cd34?/?, cd43?/?, or cd34?/?/cd43?/?
bone marrow (Figure 6B).
To determine the frequency of wild-type and mutant
mast cells in the peritoneal cavity after competitive re-
CD34 and/or CD43 Are Required for Mast Cell
Repopulation of the Peritoneal Cavity in W/WvMice
W/Wvmice, which bear mutations in c-kit, virtually lack
tissue mast cells and have been widely used as a model
system for testing the ability of mutant mast cells to
repopulate adult tissues (Galli, 2000). Here, we used
these mice as a permissive system for measuring the
ability for mutant bone marrow precursors to compete
CD34 and CD43 Block Mast Cell Adhesion
Figure 5. CD34 and CD43 Are Essential for Optimal Recovery of Peritoneal Mast Cells
(A) Mast cell numbers in vivo are similar in wild-type and cd34?/?, cd43?/?, and cd34?/?/cd43?/?mice. Ear and back skin was sectioned and
stained with toluidine blue. The number of purple, granular-positive mast cells per field is shown. Peritoneal lavage cells were cytocentrifuged
and Giemsa stained to identify mast cells. For mucosal mast cells, intestines were sectioned and stained with Alcian blue and safranin. Alcian-
blue-positive mast cells were quantified per villus for each mouse. Error bars show standard deviation values.
(B) The frequency of mast cell progenitors per 106mononuclear cells is similar in wild-type, cd34?/?, cd43?/?, and cd34?/?/cd43?/?bone marrow
and spleen. Cells from each tissue were plated in a limited dilution analysis with IL-3 and analyzed after 14 days for mast cell colony growth.
The frequency of MCp in bone marrow and spleen was measured in three independent experiments. Error bars show standard deviation values.
(C) Number of peritoneal mast cells from wild-type cd34?/?, cd43?/?, and cd34?/?/cd43?/?mice at various time points after ablation of mast
cells by intraperitoneal injection of water. Each symbol represents one mouse. Average for each genotype is shown as a horizontal line. Blue
diamonds, wild-type; green squares, cd34?/?; orange circles, cd43?/?; and red triangles, cd34?/?/cd43?/?. Values are indicated as significantly
different from wild-type (**p ? 0.01).
constitution of W/Wvmice, these cells were tested for
the presence or absence of CD34 and CD43 on their
surface. As controls for CD34 and CD43 expression,
nonreconstituted wild-type and cd34?/?/cd43?/?mice
wereanalyzed inaparallel experiment(Figure 6C).Virtu-
ally all mast cells resulting from competitive recon-
stitution by wild-type and cd34?/?/cd43?/?cells were of
peritoneal mast cells to similar frequencies as mice that
had been solely reconstituted with CD43?or CD34?
cells. (The higher level ofCD34 expression in reconstitu-
ted mice may reflect upregulation of this protein after
petitively reconstituted with wild-type CD45.1 and
cd34?/?/cd43?/?bone marrow, ?90% of the resulting
peritoneal mast cells were CD45.1?(data not shown).
CD34 and CD43 Block Mast Cell Adhesion
showing a similar frequency of CD43?mast cells to that
predicted for an equal contribution by wild-type and
cd43?/?cells (Figure 6C). Competitive reconstitution by
wild-type and cd34?/?cells showed the same number
of CD34?mast cells as reconstitution by CD34?cells
alone; however, this was not significantly different to
the predicted equal contribution (Figure 6C). Our results
suggest that, under competitive and noncompetitive
conditions, cd34?/?/cd43?/?cells are severely defective
in their ability to reconstitute peritoneal mast cells.
populated with only cd34?/?/cd43?/?bone marrow cells
(no wild-type competitors) did not have any detectable
c-kit-positive cells in the bone marrow (Figure 7A). This
suggests that CD34 and/or CD43 are essential for stem
cell engraftment in this W/Wvmodel.
To determine the donor origin of c-kit-positive cells
in competitively reconstituted mice, these cells were
analyzed for CD43 and/or CD45.1 expression. For mice
reconstituted with wild-type CD45.1?and cd34?/?
CD45.1?, suggesting that cd34?/?cells were unable to
engraft the bone marrow (Figure 7B). In contrast, mice
reconstituted with wild-type CD45.1?and cd43?/?
CD45.1?cells showed equal contribution from each
population, implying that there was no disadvantage for
cd43?/?cells (Figure 7B). In mice competitively reconsti-
tuted with wild-type and cd34?/?/cd43?/?cells, virtually
all c-kit-positive cells expressed CD45.1 and/or CD43,
indicating wild-type origins (Figure 7B). Therefore, in
this model, CD34 itself is essential for hematopoietic
CD34 Is Required for Progenitor Cell Repopulation
of the Bone Marrow in W/WvMice
Our observation that loss of CD34 and CD43 prevents
mast cell reconstitution of the peritoneal cavity of W/Wv
mice does not necessitate a mast cell precursor defect
per se, but could also be explained by a stem cell defect
of mutant cells within the bone marrow. This idea was
not supported by our observation that there was no
competitive advantage for wild-type or mutant bone
marrow cells to contribute to multilineage reconstitution
in lethally irradiated wild-type mice (data not shown).
However, since mice with mutations in c-kit have been
shown to be particularly sensitive to engraftment after
sublethal irradiation (Benveniste et al., 2003; Vecchini
competitively reconstituted W/Wvrecipients.
row (or spleen) cells from W/Wvmice (either due to
decreased number of progenitors (Migliaccio et al.,
1999) and/or decreased level of expression (Hayashi et
al., 1991; Waskow et al., 2004) (Figure 7 and data not
shown), all c-kit-positive bone marrow cells in reconsti-
tuted mice were donor derived. We could therefore enu-
merate the frequency of wild-type and mutant hemato-
poietic progenitors simply by quantifying the frequency
of c-kit?CD45.1?/CD43?and c-kit?CD45.1?/CD43?cells,
In wild-type and cd34?/?/cd43?/?mice, approximately
20% of bone marrow cells express c-kit (Figure 7A).
Although there was no expression of c-kit on bone mar-
row cells from nonreconstituted W/Wv, upon competi-
tive transplantation with wild-type and cd34?/?, cd43?/?,
or cd34?/?/cd43?/?bone marrow cells, c-kit-positive
cells were detected at the same frequency as wild-type
bone marrow (Figure 7A). Interestingly, W/Wvmice re-
Despite the widespread use of CD34 as a selective
marker of hematopoietic precursor cells over the last
two decades, its function on hematopoietic cells has
remained elusive. Here, we have taken advantage of our
observation that CD34 is a marker of murine mast cells
to address the hematopoietic function of this antigen.
Our results provide compelling evidence that CD34 as
well as the distantly related sialomucin, CD43, are nega-
tive regulators of mast cell/mast cell progenitor adhe-
sion in vitro and in vivo.
CD34 as Blocker of Differentiation, Enhancer
of Proliferation, or Homing Receptor?
The speculation in the literature that CD34, when ex-
pressed by hematopoietic cells, can act as a blocker of
differentiation or enhancer of proliferation stems from
three observations. First, CD34 is normally expressed
at the highest levels by rapidly proliferating multipotent
hematopoietic progenitors and is progressively lost as
these mature (reviewed in Krause et al., 1996). Thus, its
expression pattern correlates with a potential role in
these processes. Second, one strain of cd34 knockout
tor cells in embryonic and adult tissues, and adult-
Figure 6. CD34 and CD43 Are Required for Optimal Peritoneal Mast Cell Reconstitution of Mast Cell-Deficient Mice
(A) Schematic showing experimental design of competitive reconstitution of W/Wvmice. Sublethally irradiated mice were intravenously injected
with a 1:1 mixture of wild-type and either cd34?/?, cd43?/?, or cd34?/?/cd43?/?bone marrow cells, or cd34?/?/cd43?/?alone. Eleven to twelve
weeks later, mast cell reconstitution was assessed.
(B) Percentage of peritoneal mast cells was evaluated for nonreconstituted wild-type, cd34?/?/cd43?/?, and W/Wvmice for competitively
reconstituted W/Wvmice (with wild-type and either cd34?/?, cd43?/?, or cd34?/?/cd43?/?bone marrow) and noncompetitively reconstituted
cd34?/?/cd43?/?bone marrow. Mast cells were identified as c-kithiSSChicells. Error bars show standard deviations. Value is indicated as
significantly different from the average percentage of c-kit-positive cells resulting from competitive reconstitution (**p ? 0.01).
(C) Flow cytometric analysis of c-kithiSSChiperitoneal cells from nonreconstituted wild-type and cd34?/?/cd43?/?mice (as controls) and
competitively reconstituted W/Wvmice that were injected with a 1:1 mixture of wild-type and either cd34?/?, cd43?/?, or cd34?/?/cd43?/?bone
marrow. The bar graphs show the relative reconstitution of CD43?and CD34?mast cells for mice reconstituted with CD43?and CD34?bone
marrow alone and mice competitively reconstituted with wild-type and knockout cells. Also shown is the predicted frequency of CD43?and
CD34?cells assuming equal contribution of wild-type and cd34?/?/cd43?/?cells (gray bars). Error bars show standard deviations. Values are
indicated as significantly different from the predicted values for equal reconstitution (**p ? 0.01).
Figure 7. Loss of CD34 Prevents Bone Marrow Engraftment
(A) Graphical representation of the percentage of c-kit-positive bone marrow cells, assessed using flow cytometry, in the bone marrow of
nonreconstituted wild-type, cd34?/?/cd43?/?, and W/Wvmice. Also shown is the number of c-kit-positive cells that resulted from competitive
reconstitution of W/Wvmice by wild-type and either cd34?/?, cd43?/?, or cd34?/?/cd43?/?bone marrow or noncompetitive reconstitution with
cd34?/?/cd43?/?bone marrow alone. Error bars show standard deviations. Values are indicated as significantly different from the predicted
values for equal reconstitution (**p ? 0.01).
(B) C-kit-positive cells were gated for competitively reconstituted W/Wvmice and analyzed for expression of CD45.1 (wild-type) and/or CD43
to determine which donor these cells were derived from. Error bars show standard deviations. Values are indicated as significantly different
from the predicted values for equal reconstitution (**p ? 0.01).
(C) Proposed model for the function of CD34 and CD43 in vitro and in vivo. Loss of CD34 or CD43 results in increased adhesion of mast cells
in vitro, with the loss of both these molecules increasing this adhesion, which can be reversed by ectopic expression of CD34 or CD43. The
lack of CD34 in vivo causes decreased bone marrow engraftment, and the loss of CD34 and CD43 impairs the recovery of peritoneal mast
cells from endogenous bone marrow-derived precursors.
derived progenitor cells were somewhat defective in
their ability to proliferate in vitro (Cheng et al., 1996).
cell line that is inducible for macrophage differentiation
(M1) was shown to block terminal differentiation into
macrophages and to maintain cells in a highly prolifera-
tive “immature” state (although it was incapable of
blocking the differentiation of two other differentiation-
inducible cell lines [Fackler et al., 1995]). Although these
experiments are indirect, they are consistent with a role
for CD34 in maintaining hematopoietic progenitors in an
undifferentiated state or in enhancing their proliferation
prior to terminal differentiation. Our observation that
cd34?/?and wild-type mast cells mature and differenti-
ate at identical rates would argue strongly against a
significant, global role for CD34 in these processes
(Drew et al., 2002) (although a selective role in some
lineages is still a formalpossibility). Based on our obser-
vation that cells lacking CD34 tend to exhibit enhanced
adhesion, we propose that the previous reports of an
effect on proliferation and differentiation could reflect
downstream effects due to inappropriate cell adhesion.
More recently, it has been argued that CD34 may act
as a specific bone marrow homing receptor. This is
based on the fact that, in short-term homing assays,
cd34?/?hematopoietic progenitors migrate poorly to the
bone marrow and tend to be recovered more frequently
from the spleen than their wild-type counterparts
(Krause et al., 2001). However, the inability of cd34?/?
hematopoietic cells to home to the bone marrow in
short-term assays (Krause et al., 2001) may, in fact,
reflect increased adhesion of these cells in peripheral
tissues (spleen) rather than the loss of a specific homing
receptor. Our observation that terminally differentiated
mast cells express high levels of CD34 (Drew et al.,
2002) would argue strongly against a global role for this
molecule as a specific, bone marrow homing receptor.
In mice, mast cell precursors leave the bone marrow as
CD34 and CD43 Block Mast Cell Adhesion
in adult mouse bone marrow, the data do not support
a role for CD34 in bone marrow homing of mast cells.
However, our observation that the absence of CD34
on hematopoietic progenitors prevents bone marrow
that CD34 is important for bone marrow homing when
it is expressed on progenitor cells.
Function of CD34 and CD43 on Mast Cells
and Their Progenitors
The mechanistic details of mast cell progenitor migra-
tion remain to be clarified. However, there has been a
recent demonstration that these cells require ?4?7integ-
rin to home to the intestine and Mac-1 to home to the
peritoneum(Gurish etal.,2001;Rosenkranz etal.,1998).
Our results suggest that it is not only the presence of
adhesion molecules that is important for the proper lo-
calization of mast cell progenitors to tissues, but also
the presence of antiadhesion molecules that prevent
inappropriate adhesion during migration and/or permit
the initial release of mast cell precursors from the
Since mutant mice have normal numbers of tissue
mast cells at steady state (Figure 5A), our data may
to seed peripheral tissues during embryogenesis and
that these precursors persist in the tissues to adulthood
without a requirement for recolonization unless they are
depleted. Theability of these precursorsto seed periph-
eral tissues early in development could therefore reflect
several factors, including phenotypic difference be-
tween fetal and adult mast cell precursors (Rodewald
et al., 1996) or differences in the mechanisms that allow
colonization early in life.
It has been previously shown using chimeric mice
that, in the adult, mast cell progenitors originate in the
bone marrow, migrate to the peritoneum, and subse-
quently differentiate (Kanakura et al., 1988). Our results
show a lack of peritoneal mast cell recovery in cd34?/?/
cd43?/?mice, despite normal frequencies of bone mar-
row MCp (Figures 5B and 5C), and suggest that there
is a defect in the migration of mast cell precursors to
peripheral tissues in response to challenge. Further-
more, we have found that in vitro mutant mast cells
ity (Drew et al., 2002 and data not shown), excluding
the possibility that this observation is due to a defect
in survival or differentiation. Together, our results sup-
port the hypothesis that the impairment in peritoneal
repopulation is due to retention of mutant precursors
within the bone marrow or inappropriate adhesion to
the vasculature, thereby preventing these cells from
reaching the peritoneal cavity.
CD34 as a Proadhesion or an Antiadhesion Molecule?
Although there was initially speculation that CD34 could
act as an antiadhesion molecule, based on its similar
structure to other antiadhesion molecules, reciprocal
expression withadhesion molecules,and ultrastructural
localization on endothelial cells (Delia et al., 1993), this
hypothesis quickly lost favor with the discovery that, on
highendothelial venules(HEV),CD34 actsas aproadhe-
sive ligand for L-selectin (Baumhueter et al., 1993). Ele-
gant experiments have shown that, when expressed by
HEV, CD34 and its close relative, Podocalyxin, undergo
specific posttranslational modifications that endow
them with the ability to bind to L-selectin expressed on
the surface of migrating lymphocytes (Baumhueter et
al., 1993; Bistrup et al., 1999; Sassetti et al., 1998). This
leads to lymphocyte tethering to this specialized endo-
thelium, followed by integrin-dependent firm adhesion
Based on these data, it has been proposed that CD34
may serve a similar function in tethering hematopoietic
cells to the bone marrow microenvironment. An impor-
tant caveat to this hypothesis is that L-selectin binding
to CD34 and Podocalyxin is critically dependent upon
gens. Since these posttranslational modifications are
exquisitely tissue specific and undetectable on CD34 or
Podocalyxin expressed by the vast majority of vascular
endothelial cells or by hematopoietic cells, it is likely
that CD34’s proadhesive role on HEV is an important
exception rather than the rule.
Nevertheless, in apparent support of this “proadhe-
sion hypothesis,” several studies have shown that anti-
bodies to human CD34 (and CD43) induce antigen
capping and homotypic aggregation of CD34/CD43-
expressing progenitor cells (Majdic et al., 1994; Tada et
al., 1999). Although these results can be interpreted to
suggest an activation-dependent adhesive function for
CD34, we propose an equally plausible alternative hy-
pothesis: capping of CD34/CD43 with bivalent antibod-
ies leads to the “unmasking” of adhesion molecules,
and this results in cell aggregation. It is noteworthy that
in these antibody crosslinking studies, homotypic ag-
gregation required ?2-integrin activation and divalent
cations (Majdic et al., 1994). Thus, it is likely that the
local clearing of CD34 and CD43 in the plasma mem-
brane leads to enhanced integrin function due to the
unmasking of these molecules rather than proadhesive
function conferred directly by CD34 and CD43. We have
yet to clarify whether these molecules are responsible
for the homotypic aggregation in cd34?/?/cd43?/?
BMMC, but our observation that aggregation is divalent
tinger et al., 2000).
The Role of CD34 and CD43 in Hematopoietic
Progenitor Cell Colonization of Bone Marrow
Intriguingly, we also found that, in a competitive bone
marrowreconstitution assayof W/Wvmice, theresulting
peritoneal mast cells were derived predominantly from
wild-type donors (Figures 7A and 7B). This observation
could be due to a functional lesion at a variety of stages
in mast cell development, including (1) a bone marrow
progenitor cell defect, (2) a mast cell lineage survival
defect, or (3) a migratory defect of the mast cell progeni-
tors. Our data, showing that under competitive and non-
competitive conditions cd34?/?/cd43?/?bone marrow
cells fail to reconstitute c-kit-positive cells in the bone
marrow of W/Wvmice (Figure 7A), fit most closely with
the first hypothesis. Although we have not previously
mice, the available niches in sublethally irradiated W/Wv
mice are more limiting and therefore would cause tight
competition between the two donor populations. Our
results suggest that wild-type cells seed the bone mar-
row more efficiently in the latter, more sensitive system.
Interestingly, we found CD34 to be more important
than CD43 for bone marrow repopulation in this W/Wv
model (Figure 7B). While this explains the defect in mu-
tant mast cell colonization of the peritoneum in this
model, it also points to an important role of CD34 in
hematopoietic progenitor cell engraftment of the bone
impaired short-term homing of cd34?/?progenitors to
the bone marrow (Krause et al., 2001). It also closely
correlates with the observed upregulation of CD34 in
bone marrow reconstitution experiments (Sato et al.,
1999). The decreased ability for cd34?/?cells to engraft
could be due to (1) increased inappropriate adhesion or
(2) the requirement for CD34 binding an unknown ligand
within the bone marrow environment. However, since
the loss of CD34 does not affect the ability of progenitor
cells to bind P-/E-/or L-selectin (data not shown), this
excludes these selectins as potential ligands for CD34
in this context. In light of our in vitro observations show-
ing that CD34 and CD43 block mast cell adhesion and
we propose a block in nonspecific adhesion to be the
most likely mechanism for impaired bone marrow en-
graftment of mutant cells.
In summary, our results show that CD34 and CD43
are important in preventing adhesion of mast cells
sors to the peritoneum (Figure 7C). In addition, our re-
sults provide evidence for the importance of CD34 on
bone marrow stem cells, suggesting that, under condi-
tions of a limited number of niches, these mutant cells
are at a disadvantage to their wild-type counterparts
act as both proadhesive and antiadhesive molecules,
depending on the context of their expression and post-
translational modifications (Doyonnas et al., 2001; Niel-
sen et al., 2002). In the specialized case of HEV, we
propose that CD34 family members may have a dual
function. As has been shown previously, these mole-
cules may first act as proadhesives to tether lympho-
cytes to HEV via L-selectin recognition of HEV-specific
glycosylations decorating CD34-type proteins (Baum-
hueter et al., 1993; Sassetti et al., 1998). Subsequently,
these molecules may be induced to move to the junc-
tions between vascular endothelial cells, where they
may act as antiadhesives and aid in the disruption of
travasation. In support of this model, it has recently
been shown that overexpression of one CD34 homolog,
Podocalyxin, in adherent monolayers leads to the dis-
ruption of adherens junctions (Takeda et al., 2000). Con-
ber in mice leads to perinatal lethality due to enhanced
adherens and tight junction formation between kidney
glomerular epithelial cells and excessive adhesion be-
nas et al., 2001).
We propose that on mature hematopoietic cells and
most vascular endothelia, CD34 and/or CD43 act as
ate adhesion by blocking the binding of adhesive mole-
cules to adjacent cells via their bulky, highly negatively
charged extracellular domains. On hematopoietic pro-
genitors, these proteins may provide a means of releas-
ing cells from one niche and allowing them to migrate
to new microenvironmental niches for maturation or, in
the most extreme cases (during embryogenesis or dur-
ing mobilization of stem cells with G-CSF), endow these
cells with the ability to enter the peripheral blood and
seed new tissues. Our results support the original, and
largely abandoned, hypothesis that CD34 inhibits cellu-
forCD34 onhematopoieticcells bothinvitroand invivo.
All mice were maintained under pathogen-free conditions at the
involving mice were approved by the University of British Columbia
Animal Care Committee. C57Bl/6 and CD45.1 (Bl/6SJL) mice were
used as wild-type mice. Cd34?/?mice (on a C57Bl/6 background)
were kindly provided by Dr. T.W. Mak (Suzuki et al., 1996), and
cd43?/?mice (also on a C57Bl/6 background) were kindly provided
by Dr. Blair Ardman (Carlow et al., 2001; Manjunath et al., 1995).
eral bloodcells using anti-CD43mAb S7 and flowcytometry (Carlow
et al., 2001). Cd34?/?/cd43?/?double-deficient mice were generated
by first crossing cd34?/?and cd43?/?single-deficient mice to pro-
duce cd34?/?/cd43?/?F1 heterozygotes and subsequently crossing
siblings to generate double-homozygous double-null mice. Female
Bone Marrow-Derived Mast Cells
Bone marrow was flushed from the femurs of homozygous mutant
or wild-type mice. Red blood cells were lysed using 0.1 M NH4Cl,
with penicillin/streptomycin, sodium pyruvate, glutamine (RPMI?),
10% fetal bovine serum, and 16 U/ml IL-3 obtained from WEHI-3B
conditioned media (Tsuji et al., 1991). Cells were transferred to new
flasks periodically during culture to remove adherent cells, and they
ation (Tsuji et al., 1991).
Nucleic Acid Analyses
Genotypic analysis of wild-type and cd34?/?littermates was per-
formed using ear punches digested with proteinase K in lysis buffer
(50 mM Tris [pH 8.0], 2 mM NaCl, 10 mM EDTA 1% SDS with 1mg/
ml proteinase K) at 55?C for 40 min. Tissue samples were then
diluted with water, heated for 10 min at 100?C, then further diluted
prior to use as a template for polymerase chain reactions (PCR).
PCR was performed using the following primers (kindly provided
by Dr. Ste ´phane Corbel, Biomedical Research Centre, Vancouver,
Canada): forward 5?-CCATCTTGGGCACCACTGGTTATT-3? and re-
verse 5?-TCTTCCCAACAGCCATCAAGGTTC-3? for the wild-type
cd34 allele, and forward 5?-AGAACCTGCGTGCAATCCATC-3? and
reverse 5?-CACTGTCCTGTCTAGGTTGAACC-3? for analysis of the
neomycin cassette used for homologous recombination of the tar-
geted locus (Suzuki et al., 1996). PCR was performed for 40 cycles
of 94?C for 45 s, 58?C for 55 s, and 72?C for 2 min. Products were
resolved on a 2% agarose gel and were visualized using ethidium
bromide. Wild-type mice were identified by a 600 bp CD34 band
but no neomycin band, and cd34?/?mice were identified with no
wild-type CD34 band and a 1 kb neomycin-CD34 band.
To generate full-length mouse CD34 expression constructs
(CD34FL), CD34 cDNA (the kind gift of Dr. Mel Greaves, Chester
NotI and XhoI and inserted into pMXpie (a kind gift of Dr. Alice Mui,
Jack Bell Research Centre, Vancouver, Canada). The viral LTRs of
CD34 and CD43 Block Mast Cell Adhesion
this vector drive expression of CD34 and EGFP (enhanced green
fluorescent protein) via a bicistronic mRNA containing an IRES ele-
ment. To generate cytoplasmically truncated CD34 (CD34CT) corre-
sponding to the naturally occurring CD34 splice variant (Nakamura
et al., 1993), PCR was performed using the following primers: for-
ward 5?-ACGACTCACTATAGGGCGAAT-3? and reverse 5?-TGCT
the full-length CD34 cDNA in pBluescript as a template. Sequencing
confirmed correct amplification of a cDNA with an identical coding
sequence to this splice variant. The resulting band was excised,
digested with XhoI and XbaI, and inserted into pBluescript KS?,
which was subsequently cut with XhoI and NotI. The gene was
inserted into the retroviral expression vector, pMXpie. CD43 was
amplified from mouse kidney cDNA using the following primers:
forward 5?-GTTAAACCACAAGATGGGCTTGGCAGTTGG-3? and re-
quence followed by two glycine residues and a ClaI site was added
after the signal peptide, and the fragment was cloned into pMXpie
using BamHI and EcoRI.
was quantified. Isotype-matched hamster IgG2 (Pharmingen), rat
IgG2b, and rat IgG1 (Cedarlane) were used at the appropriate con-
centrations for controls.
Fibronectin Adhesion Assay
24-well plates (Nunc) were coated overnight with 50 ?g/ml fibronec-
tin (Sigma) at 37?C. The next day, wells were washed with PBS, cells
were washed with RPMI (no FBS), and cells were plated in triplicate
in RPMI on the coated plates with no stimulation or with 100 nM
TPA at 3 ? 105cells/ml. After 60 min, the number of suspension
cells were counted using a hemocytometer, and the percentage of
adherent cells was calculated.
To quantify the number of tissue mast cells, sections of ear and
dermal skin and the intestine were prepared as follows. Hair was
removed from the dermal skin with a depilatory cream, and skin
sections were fixed in neutral buffered formalin or 3:1 methanol:gla-
cialacetic acid.Intestinal sectionswere fixedin 3:1methanol:glacial
Tissues were embedded in paraffin, sectioned at 3 ?m, deparaffin-
ized, rehydrated, and stained. Toluidine blue or Giemsa were used
to detect CTMC as described previously (Drew et al., 2002); CTMC
stained with Alcian blue (1% w/v in 0.7 M HCl [pH 0.3]) for 60 min
and counterstained with safranin (0.5% w/v in 0.125 M HCl [pH 1.0])
for 30 s as described previously (Madden et al., 1991), and MMC
were identified as small Alcian-blue-positive cells. Sections were
observed under a Zeiss Axioplan2 microscope, and the number of
mast cells per field of view (CTMC) or per villus (MMC) was deter-
mined. Peritoneal cells were harvested by peritoneal lavage using
10 ml FACS buffer (PBS, 10% FBS, 0.05% sodium azide) injected
with a 26 gauge needle and harvested with an 18 gauge needle.
Cells were cytocentrifuged, Giemsa stained, and examined using a
Zeiss Axioplan2 microscope.
BOSC cells were transfected overnight using lipofectamine PLUS
(Invitrogen) with pclECO, encoding for retroviral packaging proteins
(Naviaux et al., 1996), and either empty pMXpie-GFP vector or
pMXpie-GFP coexpressing CD34FL, CD34CT, or CD43. Bone marrow
was isolated the next day from cd34?/?, cd43?/?, or cd34?/?/cd43?/?
mice and cultured overnight in RPMI? with 15% fetal bovine serum,
8 U/ml IL-3 (from WEHI-3B conditioned media), 10 ng/ml rmIL-6
hamster kidney cells transfected with mouse SCF, kindly provided
by Dr. Ste ´phane Corbel, Biomedical Research Centre, Vancouver,
Canada). The next day, BOSC cells were irradiated with 5000 rads,
and 7 ? 106bone marrow cells were added with 6 ?g/ml polybrene
for 2 days. Finally, nonadherent cells were removed and replated
in RPMI? with 10% fetal bovine serum, 16 U/ml IL-3 from WEHI-
3B conditioned media to induce mast cell differentiation, and puro-
mycin(0.8 ?g/ml)wasincludedto selectforinfectedcells. Toensure
ectopic gene expression, GFP-positive cells were sorted using a
microscopy (Olympus). Mast cell differentiation was confirmed by
homogenoushigh expressionofc-kit detectedusing flowcytometry
(FACSCalibur, Becton-Dickinson) and exhibition of granules de-
tected by staining cytospun cells with modified Giemsa (Diffquick).
Mast Cell Precursor Limiting Dilution Analysis
To assess the frequency of MCp, a limited dilution analysis was
performed as previously described (Crapper and Schrader, 1983;
Gurish et al., 2001). Briefly, bone marrow was flushed and spleens
were harvested using HBSS 2% FBS from wild-type, cd34?/?,
cd43?/?, and cd34?/?/cd43?/?mice. Spleens were broken up gently
between frosted glass slides and filtered. Bone marrow and spleen
cells were then resuspended in 44% Percoll (Amersham) and over-
layed on 67% Percoll, and samples were spun for 20 min at 400 ? g
at 4?C. The interface was harvested, washed, and resuspended in
media (RPMI 1640, 10% FBS, glutamine, penicillin/streptomycin,
sodium pyruvate, nonessential amino acids, 10 mM HEPES, 2-mer-
captoethanol). Cells were counted, serially diluted 2-fold, and 100
?l of each dilution (8 ? 102–1 ? 102cells per well for bone marrow
cells and 8 ? 103–1 ? 103for splenic cells) was aliquoted into 96-
well plates. Each well was supplemented with 105irradiated wild-
type helper spleen cells (3000 rads) and 10 ng/ml rmIL-3 (R&D).
Plates were incubated for 14 days and were examined under an
inverted microscope. MCp colonies were identified as nonadherent,
medium-sized cells, as previously reported (Crapper and Schrader,
1983; Gurish et al., 2001), and the number of wells that contained
one or more MCp colonies were scored as positive for each plate.
for a MCp colony was estimated using the trend function on Micro-
soft Excel from the plot of the log of the fraction of nonresponding
cultures and the cell number plated.
Homotypic Adhesion Assays
Homotypic adhesion was determined by counting the number of
single cells and the number of cells in aggregates within a micro-
scopic field of view. Percentage aggregation was calculated as the
number of cells in aggregates (number of aggregates times the
average number of cells per aggregate) over the total number of
cells counted in each assay. For all aggregation assays, at least
1500 cells were counted. For assessment of cation dependence of
cell aggregates, 1 mM EDTA (disodium ethylenediamine tetraace-
tate, Fisher) and/or EGTA (ethylenebis(oxyethylenenitrilo)-tetraace-
tic acid, Boehringer) was added to BMMC. For cell mixing experi-
ments, cd34?/?or cd34?/?/cd43?/?BMMC were stained for 7 min at
room temperature with 10 ?M CSA-SE (SNARF-1 carboxylic acid,
acetate, succinimidyl ester, Molecular Probes), washed three times
in RPMI? 10% FBS, and then passed through a 26 gauge needle
ten times to create a single cell suspension. These were then mixed
with GFP-positive cd34?/?mast cells ectopically expressing CD34CT
and were incubated overnight to allow the formation of cell aggre-
gates. Cells were examined under an inverted fluorescence micro-
scope (Olympus) using blue and green filters to examine GFP-posi-
tive and CSA-SE-stained cells, respectively. Digital photographs of
fluorescent cells were taken with a Sensys1401E camera (Roper
Scientific) using MetaVue (Universal Imaging Corporation) or
using Adobe Photoshop and Adobe Illustrator.
Blocking experiments were performedwith antibodies toward ?1-,
?2-, ?4-integrins, ICAM-1 (10 ?g/ml, Pharmingen), and SgIGSF (50
?g/ml, kindly provided by Dr. Yukihiko Kitamura). Cells were treated
with blocking antibodies overnight, after which aggregation of cells
Mast Cell Recovery
Wild-type and cd34?/?, cd43?/?, and cd34?/?/cd43?/?mice were in-
jected intraperitoneally with 3 ml sterile distilled water as previously
described (Kanakuraet al.,1988). Foreach genotypeand timepoint,
three to seven mice were evaluated (the sum of three separate
experiments). At various time points after injection, mice were sacri-
and Giemsa stained. Greater than 1500 cells were counted blindly
for each animal at each time point using a Zeiss Axioplan2 micro-
scope, and the number of mast cells per mouse was quantified.
Bone marrow was extracted from the femurs and tibias of wild-type
or CD45.1, cd34?/?, cd43?/?, and cd34?/?/cd43?/?mice with a needle
and syringe containing HBSS with 2% FBS. Red blood cells were
lysed in 0.1 M NH4Cl, and white blood cells were counted using a
hemocytometer with eosin to exclude dead cells. W/Wvmice were
sublethally irradiated with 400 rads and injected intravenously with
1 ? 107wild-type (or CD45.1) and 1 ? 107knockout bone marrow
cells mixed prior to injection or 1 ? 107cd34?/?/cd43?/?cells alone.
Mice were sacrificed 11–12 weeks after injection and were analyzed
for donor-derived hematopoiesis.
The degree of contribution by wild-type and knockout donor cells
was determined by assessing the number of c-kit-positive cells
expressing CD34, CD43, and/or CD45.1 in the peritoneum and bone
marrow. CD34?and CD43?peritoneal mast cells resulting from
competitive reconstitution were compared to mice that had been
reconstituted with only CD34?or CD43?cells to determine the rela-
from mice reconstituted with CD34?or CD43?cells. Results were
compared to the predicted equal contribution to determine if wild-
type cells contributed more than knockout cells to the resulting
mast cell population or to c-kit?bone marrow cells.
Ardman, B., Sikorski, M.A., and Staunton, D.E. (1992). CD43 inter-
feres with T-lymphocyte adhesion. Proc. Natl. Acad. Sci. USA 89,
Baecher-Allan, C.M., Kemp, J.D., Dorfman, K.S., Barth, R.K., and
Frelinger, J.G. (1993). Differential epitope expression of Ly-48
(mouse leukosialin). Immunogenetics 37, 183–192.
Baumhueter, S., Singer, M., Henzel, W., Hemmerich, S., Renz, M.,
Rosen, S., and Lasky, L.A. (1993). Binding of L-selectin to the vascu-
lar sialomucin CD34. Science 262, 436–438.
Nature 420, 875–878.
Benveniste, P., Cantin, C., Hyam, D., and Iscove, N.N. (2003). Hema-
topoietic stem cells engraft in mice with absolute efficiency. Nat.
Immunol. 4, 708–713.
Bhatia, M., Bonnet, D., Murdoch, B., Gan, O.I., and Dick, J.E. (1998).
A newly discovered class of human hematopoietic cells with SCID-
repopulating activity. Nat. Med. 4, 1038–1045.
Bistrup, A., Bhakta, S., Lee, J.K., Belov, Y.Y., Gunn, M.D., Zuo, F.R.,
Huang, C.C., Kannagi, R., Rosen, S.D., and Hemmerich, S. (1999).
Sulfotransferases of two specificities function in the reconstitution
of high endothelial cell ligands for L-selectin. J. Cell Biol. 145,
Carlow, D.A., Corbel, S.Y., and Ziltener, H.J. (2001). Absence of
CD43 fails to alter T cell development and responsiveness. J. Immu-
nol. 166, 256–261.
Cheng, J., Baumhueter, S., Cacalano, G., Carver-Moore, K., Thibo-
deaux, H., Thomas, R., Broxmeyer, H.E., Cooper, S., Hague, N.,
Moore, M., et al. (1996). Hematopoietic defects in mice lacking the
sialomucin CD34. Blood 87, 479–490.
Crapper, R.M., and Schrader, J.W. (1983). Frequency of mast cell
induces parallel increases in the frequency of P cell precursors and
mast cells. J. Immunol. 131, 923–928.
Dao, M.A., Arevalo, J., and Nolta, J.A. (2003). Reversibility of CD34
expression on human hematopoietic stem cells that retain the ca-
pacity for secondary reconstitution. Blood 101, 112–118.
Dastych, J., Taub, D., Hardison, M.C., and Metcalfe, D.D. (1998).
Tyrosine kinase-deficient Wv c-kit induces mast cell adhesion and
chemotaxis. Am. J. Physiol. 275, C1291–C1299.
Delia, D., Lampugnani, M.G., Resnati, M., Dejana, E., Aiello, A., Fon-
tanella, E., Soligo, D., Pierotti, M.A., and Greaves, M.F. (1993). CD34
expression isregulated reciprocallywith adhesionmolecules invas-
cular endothelial cells in vitro. Blood 81, 1001–1008.
Doyonnas, R., Kershaw, D.B., Duhme, C., Merkens, H., Chelliah,
S., Graf, T., and McNagny, K.M. (2001). Anuria, omphalocele, and
perinatal lethality in mice lacking the CD34-related protein podoca-
lyxin. J. Exp. Med. 194, 13–27.
Drew, E., Merkens, H., Chelliah, S., Doyonnas, R., and McNagny,
K.M. (2002). CD34 is a specific marker of mature murine mast cells.
Exp. Hematol. 30, 1211.
Fackler, M.J., Krause, D.S., Smith, O.M., Civin, C.I., and May, W.S.
(1995). Full-length but not truncated CD34 inhibits hematopoietic
cell differentiation of M1 cells. Blood 85, 3040–3047.
Galli, S.J. (2000). Mast cells and basophils. Curr. Opin. Hematol.
Galli, S.J., and Nakae, S. (2003). Mast cells to the defense. Nat.
Immunol. 4, 1160–1162.
Gratama, J.W., Sutherland, D.R., and Keeney, M. (2001). Flow cyto-
and progenitor cells. Semin. Hematol. 38, 139–147.
Gurish, M.F., and Boyce, J.A. (2002). Mast cell growth, differentia-
tion, and death. Clin. Rev. Allergy Immunol. 22, 107–118.
Gurish, M.F., Tao, H., Abonia, J.P., Arya, A., Friend, D.S., Parker,
CD49dbeta7 (alpha4beta7 integrin) for tissue-specific homing. J.
Exp. Med. 194, 1243–1252.
Hasumura, M., Imada, C., and Nawa, K. (2003). Expression change
Bone marrow and spleen cells underwent red blood cell lysis prior
to staining (see above), and all samples were Fc?RII/III blocked with
ythrin-conjugated anti-c-kit (2 ?g/ml, Pharmingen), biotinylated-
by Wooseok Seo), biotinylated-anti-CD43 (S7) (5 ?g/ml, Phar-
mingen), biotinylated-anti-CD34 (RAM34) (5 ?g/ml, Pharmingen),
and FITC-conjugated anti-CD45.1 (2.5 ?g/ml, Pharmingen). Strep-
tavidin-allophycocyanin was used to detect biotinylated antibodies
(0.25 ?g/ml, Pharmingen). To verify that c-kit-positive peritoneal
cells were mast cells, cells were double stained with phycoerythrin-
conjugated anti-c-kit, anti-dinitrophenyl-IgE (clone SPE-7) (Sigma),
and FITC-anti-mouse IgE (Pharmingen). Staining was performed on
ice in the dark for 15–30 min. Nonviable cells were gated out of
profiles using 2 ?g/ml 7-aminoactinomycin D (Molecular Probes).
Samples were analyzed using a FACSCalibur (Becton-Dickinson)
and CellQuest or FlowJo software.
Data were analyzed for averages and standard deviations using
Microsoft Excel. The measurement of statistical differences was
evaluated using a two-sample unpaired Student’s t test for unequal
variances. Results were considered to be statistically significant for
*p ? 0.05 and **p ? 0.01.
The authors wish to thank Dr. Ste ´phane Corbel for generation of
cd34?/?/cd43?/?mice, Dr. Xuecui Guo for technical assistance with
retroviral infections, Andrew Johnson for flow cytometry, and Shi-
erley Chelliah for help with genotypic analysis. We would also like
to thank Julie Chow in the Department of Pathology and Laboratory
and sectioning of tissues and Dr. Yukihiko Kitamura for generously
providing us with antibodies toward SgIGSF. Lastly, we would like
to extend our thanks to Dr. Michael Gurish for his technical advice
for the limited dilution analysis and Dr. Sebastian Furness for his
critical evaluation of the manuscript. K.M.M. is a Michael Smith
Foundation for Health Research (MSFHR) and a Canadian Institute
for Health Research (CIHR) Scholar and is a member of the Stem
Cell Network Centre of Excellence.E.D. was supported by a Michael
Smith Foundation for Health Research Trainee Scholarship and a
Heart and Stroke Foundation Doctoral Scholarship. This work was
funded by CIHR grant #117220 and #15477 and a grant in aid from
the Heart and Stroke Foundation of British Columbia and the Yukon.
Received: February 17, 2004
Revised: November 15, 2004
Accepted: November 17, 2004
Published: January 25, 2005
CD34 and CD43 Block Mast Cell Adhesion
of Flk-2/Flt-3 on murine hematopoietic stem cells in an activating
state. Exp. Hematol. 31, 1331–1337.
S. (1991). Exon skipping by mutation of an authentic splice site of
c-kit gene in W/W mouse. Nucleic Acids Res. 19, 1267–1271.
Hickey, M.J., Forster, M., Mitchell, D., Kaur, J., De Caigny, C., and
Kubes, P. (2000). L-selectin facilitates emigration and extravascular
locomotion of leukocytes during acute inflammatory responses
in vivo. J. Immunol. 165, 7164–7170.
H., Iseki, S., and Kitamura, Y. (2003). SgIGSF: a new mast-cell adhe-
sion molecule used for attachment to fibroblasts and transcription-
ally regulated by MITF. Blood 101, 2601–2608.
Kanakura, Y., Kuriu, A., Waki, N., Nakano, T., Asai, H., Yonezawa,
T., and Kitamura, Y. (1988). Changes in numbers and types of mast
cell colony-forming cells in the peritoneal cavity of mice after injec-
tion of distilled water: evidence that mast cells suppress differentia-
tion of bone marrow-derived precursors. Blood 71, 573–580.
Kawakami, T., and Galli, S.J. (2002). Regulation of mast-cell and
Krause, D.S., Fackler, M.J., Civin, C.I., and May, W.S. (1996). CD34:
structure, biology, and clinical utility. Blood 87, 1–13.
Krause, D.S., Theise, N.D., Collector, M.I., Henegariu, O., Hwang,
S., Gardner, R., Neutzel, S., and Sharkis, S.J. (2001). Multi-organ,
multi-lineage engraftment by a single bone marrow-derived stem
cell. Cell 105, 369–377.
Leitinger, B., McDowall, A., Stanley, P., and Hogg, N. (2000). The
regulation of integrin function by Ca(2?). Biochim. Biophys. Acta
Madden, K.B., Urban, J.F., Jr., Ziltener, H.J., Schrader, J.W., Finkel-
man, F.D., and Katona, I.M. (1991). Antibodies to IL-3 and IL-4 sup-
press helminth-induced intestinal mastocytosis. J. Immunol. 147,
Majdic, O., Stockl, J., Pickl, W.F., Bohuslav, J., Strobl, H.,
Scheinecker, C., Stockinger, H., and Knapp, W. (1994). Signaling
and induction of enhanced cytoadhesiveness via the hematopoietic
progenitor cell surface molecule CD34. Blood 83, 1226–1234.
tive regulation of T-cell adhesion and activation by CD43. Nature
Mekori, Y.A., and Metcalfe, D.D. (2000). Mast cells in innate immu-
nity. Immunol. Rev. 173, 131–140.
Migliaccio, A.R., Carta, C., and Migliaccio, G. (1999). In vivo expan-
sion of purified hematopoieticstem cells transplanted in nonablated
W/Wv mice. Exp. Hematol. 27, 1655–1666.
Nakamura, Y., Komano, H., and Nakauchi, H. (1993). Two alternative
forms of cDNA encoding CD34. Exp. Hematol. 21, 236–242.
Naviaux, R.K., Costanzi, E., Haas, M., and Verma, I.M. (1996). The
pCL vector system: rapid production of helper-free, high-titer, re-
combinant retroviruses. J. Virol. 70, 5701–5705.
Nielsen, J.S., Doyonnas, R., and McNagny, K.M. (2002). Avian mod-
els to study the transcriptional control of hematopoietic lineage
commitment and to identify lineage-specific genes. Cells Tissues
Organs 171, 44–63.
Nocka, K., Tan, J.C., Chiu, E., Chu, T.Y., Ray, P., Traktman, P., and
Besmer, P. (1990). Molecular bases of dominant negative and loss
of function mutations at the murine c-kit/white spotting locus: W37,
Wv, W41 and W. EMBO J. 9, 1805–1813.
Ogawa, M. (2002). Changing phenotypes of hematopoietic stem
cells. Exp. Hematol. 30, 3–6.
Ostberg, J.R., Barth, R.K., and Frelinger, J.G. (1998). The Roman
god Janus: a paradigm for the function of CD43. Immunol. Today
Rodewald, H.R., Dessing, M., Dvorak, A.M., and Galli, S.J. (1996).
Identification of a committed precursor for the mast cell lineage.
Science 271, 818–822.
Rosenkranz, A.R., Coxon, A., Maurer, M., Gurish, M.F., Austen, K.F.,
Friend, D.S., Galli, S.J., and Mayadas, T.N. (1998). Impaired mast
deficient mice. J. Immunol. 161, 6463–6467.
Sassetti, C., Tangemann, K., Singer, M.S., Kershaw, D.B., and Ro-
sen, S.D. (1998). Identification of podocalyxin-like protein as a high
endothelial venule ligand for L-selectin: parallels to CD34. J. Exp.
Med. 187, 1965–1975.
Sassetti, C., Van Zante, A., and Rosen, S.D. (2000). Identification of
endoglycan, a member of the CD34/podocalyxin family of sialomu-
cins. J. Biol. Chem. 275, 9001–9010.
Sato, T., Laver, J.H., and Ogawa, M. (1999). Reversible expression
of CD34 by murine hematopoietic stem cells. Blood 94, 2548–2554.
Smith, T.J., and Weis, J.H. (1996). Mucosal T cells and mast cells
share common adhesion receptors. Immunol. Today 17, 60–63.
Suzuki, A., Andrew, D.P., Gonzalo, J.A., Fukumoto, M., Spellberg,
J., Hashiyama, M., Takimoto, H., Gerwin, N., Webb, I., Molineux, G.,
et al. (1996). CD34-deficient mice have reduced eosinophil accumu-
lation after allergen exposure and show a novel crossreactive 90-
kD protein. Blood 87, 3550–3562.
Tada, J., Omine, M., Suda, T., and Yamaguchi, N. (1999). A common
signaling pathway via Syk and Lyn tyrosine kinases generated from
capping of the sialomucins CD34and CD43 in immature hematopoi-
etic cells. Blood 93, 3723–3735.
Takeda, T., Go, W.Y., Orlando, R.A., and Farquhar, M.G. (2000).
Expression of podocalyxin inhibits cell-cell adhesion and modifies
junctional properties in Madin-Darby canine kidney cells. Mol. Biol.
Cell 11, 3219–3232.
formation supported by IL-3, IL-4, and recombinant rat stem cell
factor, ligand for c-kit. J. Cell. Physiol. 148, 362–369.
Vecchini, F., Patrene, K.D., and Boggs, S.S. (1993). Purified murine
hematopoietic stem cells function longer on nonirradiated W41/Wv
than on ?/? irradiated stroma. Blood 81, 1489–1496.
Walker, J., and Green, J.M. (1999). Structural requirements for CD43
function. J. Immunol. 162, 4109–4114.
Waskow, C., Terszowski, G., Costa, C., Gassmann, M., and Rode-
wald, H.R. (2004). Rescue of lethal c-KitW/W mice by erythropoietin.
Blood 104, 1688–1695.
Williams, C.M., and Galli, S.J. (2000). The diverse potential effector
and immunoregulatory roles of mast cells in allergic disease. J.
Allergy Clin. Immunol. 105, 847–859.
Zanjani, E.D., Almeida-Porada, G., Livingston, A.G., Zeng, H., and
Ogawa, M. (2003). Reversible expression of CD34 by adult human
bone marrow long-term engrafting hematopoietic stem cells. Exp.
Hematol. 31, 406–412.