Immunity, Vol. 14, 603–615, May, 2001, Copyright 2001 by Cell Press
The Follicular versus Marginal Zone
B Lymphocyte Cell Fate Decision
Is Regulated by Aiolos, Btk, and CD21
1999), and the generation of these cells is markedly
Martin and Kearney, 2000b), a transmembrane protein
that associates with the CD21/CR2 coreceptor and
transduces signals downstream of the BCR (Fearon and
Carroll, 2000). MZ B cells also fail to develop in the
absence of the Pyk-2 tyrosine kinase (Guinamard et al.,
2000). We have noted the absence of these cells in
two mutant mice strains that lack specific transcription
and mice that lack Aiolos (Wang et al., 1998).
Aiolos is a zinc finger transcription factor of the Ikaros
family that is expressed in the B lineage (Morgan et al.,
1997). In the absence of Aiolos, B cell proliferation in
response to antigen receptor crosslinking is markedly
(Wang et al., 1998). Many of the phenotypic features
noted in the absence of Aiolos are consistent with a
lowering of the threshold for B cell receptor-mediated
signaling. Follicular B cells that emerge in these mice
might constitutively receive enhanced BCR-mediated
maturation signals. However, the marked reduction in
MZ B cells observed in the absence of Aiolos suggested
that enhanced BCR/Btk signaling might actually impair
the generation of this subset of peripheral B cells. Given
the requirement of CD19 for MZ B cell generation, the
precise role of BCR-derived signals in MZ B cell genera-
tion remains unclear.
In Xid mice (in which Btk is nonfunctional) and Btk?/?
mice, the survival of follicular IgDhiIgMloB cells is com-
promised (Hardy et al., 1982; Khan et al., 1995; Cariappa
et al., 1999) and peripheral B cells fail to proliferate in
response to antigen receptor ligation. While MZ B cells
are generated in Xid mice (Makowska et al., 1999; Loder
et al., 1999), the overall numbers of B cells that make
up this subset are reduced in these mice (Kraal et al.,
1988., Liu et al., 1988), suggesting that Btk might either
contribute directly to the survival or generation of MZ
B cells or that the reduced survival of follicular B cells
seen in the absence of Btk may indirectly impact upon
MZ B cell generation.
The major phenotypic characteristics and responses
of B cells that lack Btk and Aiolos suggest that the
products of these genes may functionally oppose one
another. We wished to explore whether Aiolos and Btk
interact with one another at the genetic level. We also
wished to examine whether a binary cell fate decision
might regulate whether a cell ultimately differentiates
into a MZ B cell or matures further and is maintained
as a follicular B cell. We have used mutant mice in which
the intensity of BCR signaling is altered to explore the
the antigen receptor in naive B cells may influence the
mentary studies on mice lacking CD21 support the no-
tion that antigens drive naive B cell development and
influence the follicular versus MZ B lymphocyte cell fate
Annaiah Cariappa,1Mei Tang,1
Chuenlei Parng,1Eugene Nebelitskiy,1
Michael Carroll,3Katia Georgopoulos,2
and Shiv Pillai1,4
1Cancer Center and
2Cutaneous Biology Research Center
Massachusetts General Hospital and
Harvard Medical School
Boston, Massachusetts 02129
3Center for Blood Research
Harvard Medical School
Boston, Massachusetts 02115
IgDhiIgMlofollicular lymphocytes, suggesting that mat-
uration signals delivered via the BCR are enhanced in
the absence of Aiolos. The enhanced maturation of
follicular B cells is accompanied by the absence of
MZ B lymphocytes and the downregulation of CD21
expression in follicular B cells, all of which depend on
the generation of signals via Btk, which is in epistasis
to Aiolos. The inverse relationship between the strength
of BCR signaling and MZ B cell development is sup-
ported by an examination of MZ B cells in CD21 null
mice. These data support the view that antigens (in
contrast to “tonic” signals) drive the development of
naive B cells.
maintenance of naive follicular B cells in the periphery
(Torres et al., 1996; Lam et al., 1997). Only a fraction of
the B cells generated every day in the bone marrow
emerge in the periphery (Allman et al., 1993); however,
it remains unclear whether the requirement for the BCR
during the emigration and maintenance of follicular B
cells represents a “constitutive” maturation event that
occurs in a similar manner in every developing B cell or
whether the antigen receptor mediates the selection of
a subset of B cells by endogenous antigens (Neuberger,
1997; Pillai, 1999).
Apart from follicular B cells, marginal zone (MZ) B
lymphocytes represent a distinct population of pre-
sumably naive B cells in the spleen (MacLennan et al.,
1982; Kraal, 1992; Martin and Kearney, 2000a). These
IgMhiIgDloCD21hiCD1hiB cells express higher basal and
induced levels of B7-1 and B7-2 than follicular B cells
and can be induced to rapidly differentiate into plasma
cells (Oliver et al., 1999). It is believed that MZ B cells
are derived from naive follicular B cells (Dammers et al.,
may also impair the generation or survival of MZ B cells.
In contrast, decreased signals via the BCR may reduce
follicular B cell numbers and contribute to an enhance-
ment of MZ B cells and MZ B cell precursors.
Follicular B Cell Numbers and Fraction I B Cells Are
Enhanced in the Absence of Aiolos
IgDhiIgMlo(Fraction I) B cells represent the most mature
follicular B cell subset. In contrast to the marked de-
crease in IgDhiIgMloB cells observed in Xid mice, an
increase in this fraction of follicular B lymphocytes was
noted in mice that lack Aiolos (Figure 1A). A marked
increase was observed in the ratio of IgDhifollicular B
cells (Fractions I and II)to IgMhiIgDloCD21lonewly formed
B cells in the absence of Aiolos (Figure 1B).
The Absence of MZ B Cells in Aiolos?/?Mice Is Linked
to Enhanced Signaling Downstream of the BCR
AiolosandBtk appeartoinfluencefollicular Bcelldevel-
opment in diametrically opposite ways. In an attempt
to ascertain whether Aiolos and Btk are in epistasis, we
mated Aiolos?/?and Xid mice and compared peripheral
B cells from Aiolos?/?mice with B cells from Aiolos?/?/
Xid mice. The enhancement in follicular B cell numbers
noted in Aiolos null mice vanishes in Aiolos?/?/Xid dou-
ble mutant mice, which phenotypically resemble Xid
mice (Figure 3A).
Although MZ B cells are not seen in Aiolos?/?mice,
they are easily detected in Xid splenic populations and
“reappear” in the Aiolos?/?/Xid double mutant mouse
(Figure 3B). Similarly, presumed MZ B cell precursors
in Aiolos?/?/Xid spleens (Figure 3C). We had originally
tor that is directly required for the generation of MZ B
cells. From these studies, it has also become clear that
Aiolos is not directly required for the generation of MZ
B cells; in the absence of Aiolos, enhanced signaling,
of MZ B cells to develop.
MZ B Cells and Presumed MZ B Cell Precursors
Are Diminished in the Absence of Aiolos
We have shown previously, using immunohistochemis-
try, that IgMhicells in the marginal zone are markedly
diminished in the Aiolos?/?mouse (Wang et al., 1998).
We quantitatively assessed the frequency of spleno-
cytes with a MZ B cell phenotype using flow cytometric
approaches. As seen in Figure 2A, IgMhiIgDloCD21hicells
are not found in the Aiolos?/?mouse, but a clearly de-
MZ B cells may also be identified on the basis of the
expression of high levels of CD1d (Roark et al., 1998)
on IgMhiIgDloB cells. Using these markers as well, very
few MZ B cells could be identified in the spleens of
Aiolos?/?mice (Figure 2B).
It has been suggested that MZ B cells may develop
from follicular B cells (Dammers et al., 1999), and
CD21hiIgDhifollicular B cells have been described (Amano
etal.,1998).We have noted in wild-type splenocyte pop-
ulations that CD21hifollicular B cells reside mainly in
Fraction II (made up of IgMhiIgDhicells). These cells are
not found in Aiolos?/?spleens and may represent MZ
2C). In contrast, in Xid mice more than half the follicular
cells in Fraction II are presumed MZ B cell precursors
follicular B cell subset that is probably identical to the
IgMhiIgDhiCD21hipopulation seen in Figure 2C. These
IgDhiCD43?CD21hipresumed MZ B cell precursors (seen
in the lower right quadrants in the two right-hand panels
of Figure 2D) are also markedly reduced in Aiolos?/?
The absolute numbers of follicular B lymphocytes are
numbers are dramatically reduced (Figure 6, described
below). Aiolos represents a negative regulator of follicu-
opment of MZ B cells. In Xid mice, there is a marked
tion of MZ B cells is enhanced (Makowska et al., 1999),
although we note a modest decrease in the absolute
numbersof MZB cells(data notshown). However,when
one examines the ratios of MZ B cells to follicular B
lymphocytes there is a striking decrease in this ratio in
Aiolos null mice and a corresponding enhancement in
Xid mice (Figure 2E). Taken together, the data in Figures
1 and 2 suggest that enhanced maturational signals
delivered via the BCR may not only contribute to the
increased emergence or survival of follicular B cells but
Btk Is Epistatic to Aiolos
Examination of anti-IgM-induced B cell proliferation
(Figure 3D) revealed a defect in Aiolos?/?/Xid mice that
resembles that seen in the absence of Btk signaling.
Taken together with the data in Figures 3A to 3C, these
results indicate that Btk is epistatic to Aiolos (Figure
3E). From a strict genetic standpoint, it may also be
inferred that Aiolos represents a negative regulator of
Btk (or of the BCR upstream of Btk) and that Btk in turn
is a negative regulator of MZ B cell formation. These
inferences are consistent with the view that a relative
enhancement of Btk-derived signals in naive B cells
contributes to follicular B cell maturation while de-
CD21 Levels Are Reduced in Follicular
B Cells that Lack Aiolos
The expression of high levels of the CD21/CD35 core-
ceptor represents one of the ways in which MZ B cells
are distinguished. We considered the possibility that
Aiolos is a transcriptional regulator of CD21 and that in
the absence of Aiolos, CD21 levels drop. Alternatively,
enhanced signaling via the BCR might downregulate
CD21 expression in MZ and other B cells. It is known
that the surface expression of CD21 in follicular B cells
speculated that this decrease might reflect activation
by antigen (Takahashi et al., 1997; Mandik-Nayak et al.,
1999). A similar decrease in CD21 expression has been
seen in patients with SLE (Wilson et al., 1986). Anti-IgM
crosslinking in vitro of normal human B cells has been
shown to contribute to a decrease in CD21 expression
Strength of Signal and MZ B Lymphocytes
Figure 1. IgDhiIgMloB Cells Predominate in the Absence of Aiolos
(A) In Aiolos?/?mice, most follicular B cells have an IgDhiIgMlophenotype. Dot plots show splenocytes in the lymphoid gate from wild-type,
Aiolos?/?, and Xid mice, stained with anti-IgM and anti-IgD antibodies. Gates are drawn according to Hardy et al. (1982), as Fraction III (IgDlo
IgMhi), Fraction II (IgDhiIgMhi), and Fraction I (IgDhiIgMlo). Numbers denote cells in individual fractions expressed as a percentage of all cells
within the lymphoid gate. Data shown are representative of 10 animals for wild-type and mutant mice.
(B) The ratio of follicular B cells to newly formed B cells is increased in Aiolos?/?mice. Absolute numbers of cells in individual fractions were
used to determine these ratios. All IgDhicells (Fractions I and II) were included in the follicular pool. Newly formed B cells were identified as
being CD21lo-neg(Oliver et al., 1997) in the Fraction III (IgMhiIgDlo) gate.
Figure 2. Marginal Zone B Cells and Presumed MZ B Cell Precursors Are Missing in Aiolos?/?Mice
(A and B) Splenocytes were stained with anti-IgM, anti-IgD, and anti-CD21 (A), or with anti-IgM, anti-IgD, and anti-CD1d antibodies (B). Both
CD21 and CD1d are present at high levels on MZ cells (which have an IgMhiIgDloCD21hiCD1hiphenotype). F-III (IgMhiIgDlo) cells were initially
gated on in both panels. CD21 gates were set based upon examination of wild-type mice. CD21hi, CD21int, and CD21locells denote MZ,
follicular, and newly formed B cells, respectively. Percentages of CD21hicells are shown in (A), and percentages of CD1dhicells are included
in (B). Data shown are representative of 10 sets of mice in (A) and five sets in (B).
(C) IgMhiIgDhiCD21hiB cells, presumed to be MZ B cell precursors, were examined in wild-type and mutant mice. F-II (IgMhiIgDhi) B cells were
Strength of Signal and MZ B Lymphocytes
(Boyd et al., 1985). It is unclear whether this decrease
of CD21 in activated B cells reflects the internalization
of surface CD21 or regulation of CD21 at the level of
Since Aiolos?/?B cells resemble lymphocytes that have
been constitutively activated, we sought to determine
if CD21 expression was altered in follicular B cells in
Aiolos?/?mice. As seen in Figure 4A, the level of CD21
on the surface of follicular IgDhiIgMlocells is markedly
reduced. Examination of permeabilized follicular B cells
revealed that intracellular CD21 protein levels were also
We sought to establish if this decrease in CD21 in the
absence of Aiolos occurs at the level of mRNA accumu-
lation. Although MZ B cells express high levels of CD21,
they account for only about 10% of all splenic B cells
in wild-type mice, and the bulk of the CD21 RNA from
purified splenic B cells is of follicular B cell origin. In
Aiolos null mice, a slight decrease in CD21 mRNA levels
might have been predicted because of the absence of
MZ B cells; however, as seen in Figure 4B there is a
significant decrease in the level of CD21 mRNA in
Aiolos?/?B cells, as assessed by a Northern blot. Taken
together with the data in Figure 4A, it appears that in
follicular B cellsthat lack Aiolos, theexpression of CD21
ismarkedlyreduced eitheratthelevel oftranscriptionor
message stability, possibly because these cells receive
constitutive BCR-derived signals. In keeping with the
published information on CD21 downregulation cited
above, we presume that BCR signaling can negatively
influence CD21 gene expression, perhaps as part of
a negative feedback phenomenon. These results are
consistent with our previous demonstration that MHC
class II expression is upregulated in B cells in the ab-
sence of Aiolos (Wang et al., 1998).
The abundance of MZ B cells in Aiolos?/?/Xid double
mutant mice suggested that Aiolos is not directly re-
quired for MZ B cell generation. The low levels of CD21
expression in follicular B cells seen in the absence of
Aiolos left open the possibility that Aiolos might be di-
rectly required for CD21 expression. The fact that CD21
is not downregulated in follicular B cells from Aiolos?/?/
Xid mice (Figure 4C) indicates that Aiolos is not likely
to represent a direct transcriptional regulator of CD21
gene expression. In the absence of Aiolos, Btk-depen-
dent signaling pathways are modulated, and it is this
modulation of signal transduction that contributes to
the enhanced maturation of follicular B cells, the down-
regulation of CD21 gene expression in follicular B cells,
and the block in MZ B cell generation.
We entertained the possibility that in Aiolos null mice
the reduction in CD21 levels might result in CD19 not
being expressed on the surface of B cells. In turn, this
absence of cell surface expression of CD19 might be
responsible for the absence of MZ B cells. Examination
III Aiolos?/?B cell populations revealed wild-type levels
of CD19 in all three fractions (data not shown).
CD21 Is Not Required for MZ B Cell Development
The absence of MZ B cells in Aiolos null mice was deter-
mined by identifying these cells not only on the basis
of high level expression of CD21 (Figure 2A) but also
based onimmunohistochemistry (Wang et al.,1998) and
on the expression of CD1d (Figure 2B). CD21 can bind
C3d and forms a complex with CD19 and TAPA-1
(Ahearn et al., 1996). The absence of MZ B cells in
CD19?/?mice (Makowska et al., 1999; Martin and Kear-
ney, 2000b) is consistent with a role for CD21 in MZ B
cell development. In a recent study, the localization of
a T-independent type 2 antigen, TNF-Ficoll, to the mar-
ginal zone was markedly impaired in Cr2 null mice (Gui-
namard et al., 2000), but MZ B cells were not quantita-
tively examined. Examination of frozen sections of
Cr2?/?spleens stained with anti-IgM and a monoclonal
antibody, MOMA-1, that recognizes marginal zone me-
tallophilic macrophages, revealed the presence of MZ
B cells (Figure 5A). We also used CD1d (Roark et al.,
1998) to identify MZ B cells in Cr2?/?splenocytes
(Ahearn et al., 1996). As seen in Figure 5B, CD1d-
expressing IgMhiIgDloMZ B cells are abundant in Cr2?/?
mice. This finding suggests that CD21 is not required
whether CD21 functions as a negative regulator of MZ
B cell development.
BCR Signal Strength Is Inversely Related
to MZ B Cell Development
As discussed above, Aiolos and Btk have reciprocal
influences on follicular and MZ B cell generation. The
view that relatively strong signals favor follicular B cell
generation while weaker signals are permissive for MZ
B cell generation is supported by the analysis of the
Aiolos?/?/Xid double mutant mice. Since Aiolos?/?B
cells proliferate more readily in response to BCR trig-
gering, and Aiolos?/?follicular B cells constitutively ex-
press higher levels of MHC class II molecules (Wang et
al., 1998) and lower levels of CD21 than wild-type B
cells, we wished to obtain more direct evidence for an
enhancement of BCR signal strength in B cells that lack
Aiolos. As seen in Figure 6A, there was a small increase
in the baseline calcium flux levels in Aiolos?/?IgD-
expressing follicular B cells and the amplitude of the
calcium response following BCR crosslinking with
(Fab?)2anti-IgM or with whole anti-IgM was clearly en-
hanced in these mutant B lymphocytes. Triggering with
whole anti-IgM in the absence of the 2.4G2 blocking
antibody (which blocks Fc?RIIb1 receptors) also re-
sulted in a slower decline of intracellular calcium levels
in Aiolos?/?B cells than in wild-type cells. Anti-IgM can
initiallygated on.Percentage ofCD21hicells isshown. CD21hi,CD21int,and CD21logates wereidenticalto thoseset in(A).Data arerepresentative
of 10 sets of wild-type and mutant mice.
(D) CD43?CD21hifollicular cells were analyzed in wild-type and Aiolos?/?mice. Splenocytes were stained with anti-IgD, anti-CD43, and anti-
CD21 antibodies, and IgD?cells were gated on. Percentages of CD21hiCD43?cells are shown. Data are representative of three sets of mice.
(E) The ratios of MZ to follicular B lymphocytes are markedly reduced in Aiolos null mice but are enhanced in Xid mice. Ratios were determined
from the absolute numbers of MZ and follicular B cells in 10 sets of wild-type and mutant mice.
Figure 3. Btk Is Epistatic to Aiolos
(A) Mature follicular Fraction I cells are reduced in Aiolos?/?/Xid double mutant mice to levels typically observed in Xid and Btk?/?mice. Five
sets of mutant mice were analyzed.
(B) IgMhiIgDlo/?CD21hiand IgMhiIgDlo/?CD1dhiMZ B cells reappear in Aiolos?/?/Xid double mutant mice. Five sets of mutant mice were analyzed.
(C) IgMhiIgDhiCD21hifollicular presumed precursors of MZ B cells are virtually absent in Aiolos null mice but form a large proportion of the
Fraction II cells in Aiolos?/?/Xid double mutant mice. Five sets of mutant mice were analyzed.
(D) B lymphocytes from Aiolos?/?/Xid mice do not proliferate in response to anti-IgM (Fab?)2stimulation. Each bar represents the mean
thymidine incorporation of quadruplicate wells set up with cells from each of three mice in every group.
(E) Btk is epistatic to Aiolos. Aiolos may be a negative regulator of Btk activation (left panel) or of BCR activation (right panel) itself, and
influences the fate of naive B cells.
Strength of Signal and MZ B Lymphocytes
induce positive signals via the BCR, as well as some
quenching of these signals by also engaging Fc?RIIb1
(Wilson et al., 1987). These results indicate that BCR
signal strength is enhanced in the absence of Aiolos.
If antigens, including complement coated antigens,
play a role in naive B cell development, CD21 might be
expected to positively influence follicular B cell genera-
tion and to negatively regulate MZ B cell development.
We sought to more quantitatively assess the generation
of MZ B cells in Aiolos?/?mice and in CD21?/?mice in
order to determine if the strength of BCR signaling was
indeed inversely related to MZ B cell generation.
Our examination of Aiolos null mice suggested that
the efficiency of MZ B cell formation (in contrast to
follicular B cell formation or maintenance) may be in-
versely related to the intensity of signals generated by
the BCR. As seen in Figures 6B and 6C, in the absence
of Aiolos there is a marked decrease in the absolute
numbers of MZ B cells but a proportionate and signifi-
cant increase in the absolute numbers of follicular B
cells (two tailed p ? 0.05, Mann-Whitney U test).
To further test the hypothesis of an inverse relation-
tion, we quantitatively analyzed MZ and follicular B cell
numbers in Cr2?/?and Cr2?/?mice. The absolute num-
bers of MZ B cells were significantly increased in Cr2?/?
mice as compared with Cr2?/?mice, indicating that
CD21 is a negative regulator of MZ B cell generation
(Figures 6D and 6E ; two tailed p ? 0.05, Mann-Whitney
U test). In contrast, the absolute numbers of follicular
B cells were decreased in Cr2?/?mice (p ? 0.05, Mann-
Whitney U test). These findings further support the view
that an impairment of BCR/coreceptor generated selec-
tion/maturation signals may negatively influence follicu-
lar B cell generation or survival and may simultaneously
favor the generation of MZ B cells.
the survival of IgMloIgDhimature follicular B cells, but
enhanced signals delivered via the BCR and Btk may
irrevocably commit a follicular cell to a non-MZ B cell
fate (Figure 7).
lute numbers of follicular B cells in the spleen, which
might reflect a decrease in the intensity of BCR-depen-
dent maturation signals required for the maintenance
of peripheral B cells. In contrast, there is a significant
increase in the absolute number of MZ B cells in Cr2?/?
mice. Clearly not only is CD21 not required for MZ B
cell development but, presumably by virtue of its ability
to positively regulate the intensity of BCR-dependent
maturation/selection signals in naive B cells, CD21 can
function as a negative regulator of MZ B cell devel-
While MZ B cell generation and localization does not
antigens such as TNP-Ficoll (Guinamard et al., 2000),
and these cells may be functionally compromised. The
reciprocal alterations seen in MZ and follicular B cell
numbers in CD21 null mice suggest that available
“endogenous” antigens, including complement coated
structures, influence B cell maturation.
hypothesis and indirectly implicate endogenous or self-
antigens in B cell maturation. Mutant mice have been
previously used to examine antigen receptor signal
strength particularly in the context of B lymphocyte tol-
erance (Cyster and Goodnow, 1995). Apart from the
demonstration of the requirement for the BCR (Torres
et al., 1996; Lam et al., 1997), other indirect arguments
have also been previously made to suggest a role for
antigen in B cell development (Gu et al., 1991; Levine
et al., 2000; Loder et al., 1999; Martin and Kearney,
2000a). While this report has not addressed issues that
relate to B-1 cell development, a role for antigen in B-1
cell development or maintenance is widely appreciated
(Hayakawa et al., 1999).
Peripheral lymphocytes that receive maturation and
maintenance signals are sometimes described as being
“tickled” as opposed to being “triggered” during the
course of an immune response (Neuberger, 1997; Pillai,
1999). This underscores the view that the range of sig-
sent a less intense induction of the BCR than is required
for proliferation and differentiation during B cell activa-
tion by a triggering antigen. It is within this subset of
less intense signals that a relatively strong tickle might
mediate follicular survival and a weaker signal might
be permissive for MZ B cell development. The recent
suggestion that some proliferation of follicular IgMhiIgDhi
B cells may occur during maturation (Loder et al., 1999)
suggests that stronger maturation signals that drive fol-
licular versus MZ B cell development may also induce
some proliferation (but not complete activation in the
triggering sense) of naive follicular B cells.
tor complex, it is intriguing that MZ B cells are markedly
reduced in the absence of CD19 (Makowska et al., 1999,
Martin and Kearney, 2000b). Signals initiated by recep-
tors other than the BCR and CR2 may potentially utilize
CD19 for the recruitment and activation of PI-3 kinase
lymphocyte may choose between a follicular and a MZ
B cell fate. In the absence of Aiolos, BCR signaling is
enhanced in naive follicular B cells, and peripheral B
TheseFraction IB cellsrepresent themost maturenaive
recirculating B cell population, and the accumulation of
these cells in the absence of Aiolos is consistent with
the accelerated maturation of follicular B cells when
BCR signaling is enhanced. In the absence of Aiolos,
CD21 gene expression is downregulated in follicular B
cells, and MZ B cells are virtually absent. Presumed
follicular precursors of MZ B cells are also not seen.
Btk is epistatic to Aiolos, and Btk signaling may be
mouse. Aiolos may therefore be a negative regulator of
Btk. While the biochemical basis for the relationship
between Aiolos and Btk remains to be addressed, these
results suggest that the exaggerated activation of Btk
in naive B cells may inhibit the differentiation of follicular
B cells into MZ B lymphocytes. BCR/Btk signaling may
therefore regulate a cell fate decision during peripheral
B cell development. Btk not only provides signals for
Figure 4. CD21 Expression Is Decreased in Follicular F-I (IgMloIgDhi) B Cells in Aiolos Null Mice
(A) Splenocytes were stained on the surface with anti-IgD and anti-IgM antibodies, permeabilized with 0.1% saponin, and then stained with
anti-CD21 antibody. Dot plots show IgM/IgD profiles before and after permeabilization. Histograms compare CD21 levels in F-I (IgMloIgDhi)
Strength of Signal and MZ B Lymphocytes
Figure 5. CD21 Expression Is Not Required for Marginal Zone B Lymphocyte Generation
(A) Immunofluorescence staining of splenic cryosections with anti-IgM-FITC and MOMA-1-PE revealed normal numbers of MZ B cells in the
Cr2?/?mouse (right panel). Multiple sections revealed similar results.
(B) Splenocytes stained with anti-IgD, anti-IgM (left panel), and anti-CD1d (right panel) show the presence of CD1dhicells in the F-III (IgMhiIgDlo)
gate. MZ B cells in this gate are all CD1dhi. Percentage of CD1dhicells are shown. Data are representative of four mice in each group.
(Tuveson et al., 1993) and Vav (O’Rourke et al., 1998).
CD19 may favor the generation or survival of MZ B cells
in a coreceptor-independent manner. It has been noted
that some of the phenotypic changes seen in the ab-
sence of CD19 are more severe than the changes seen
in the absence of CD21, raising the possibility that CD19
may be a receptor for unidentified ligands (Fearon and
A small proportion of MZ B cells, which cannot be
phenotypically distinguished form the rest of the MZ B
cell population, harbor point substitutions in Ig genes
Lennan, 1998). In murine MZ B cells, these point substi-
tutions are relatively rare (Makowska et al., 1999), and
it is likely that these less well defined MZ memory B
cells develop in a largely similar way to all other MZ B
follicular cells of WT and Aiolos?/?before and after permeabilization. Three sets of mice showed similar results.
(B) The decrease in CD21 expression seen in Aiolos?/?B cells occurs at the level of mRNA accumulation. Northern blot analysis shows a
marked decrease in CD21 in the absence of Aiolos. ?-actin was used as a loading control.
(C) CD21 is downregulated in Aiolos?/?follicular B cells but not in Aiolos?/?/Xid mice. F-I (IgMloIgDhi) B cells were initially gated on. CD21hi,
CD21int, and CD21logates were identical to those set in Figure 2A. Percentage of CD21intcells is shown. Data are representative of 10 sets of
wild-type and single mutant mice and five double mutant mice.
Figure 6. BCR Signal Strength Influences MZ and Follicular B Cell Generation
(A) Enhanced BCR signaling in the absence of Aiolos. Splenocytes from Aiolos?/?and wild-type mice were loaded with Indo-1 and stimulated
with (Fab?)2anti-IgM (in the presence of the 2.4G2 monoclonal antibody) or with whole anti-IgM (in the absence of 2.4G2). IgDhifollicular B
cells were gated on. The arrows mark the point of stimulation. (B) Absolute numbers of MZ B cells in Aiolos?/?and wild-type (WT) mice. (C)
Absolute numbers of follicular B cells in Aiolos?/?and wild-type mice. (D) Absolute numbers of MZ B cells in Cr2?/?and Cr2?/?mice. (E)
Absolute numbers of follicular B cells in Cr2?/?and Cr2?/?mice. Bars represent means of absolute numbers of lymphocytes in four mice per
Strength of Signal and MZ B Lymphocytes
signals from distinct receptors for differentiation to oc-
cur. Pyk-2 is a tyrosine kinase required for MZ B cell
generation (Guinamard et al., 2000) and it may be acti-
vated downstream of integrins, specific chemokines, or
receptors for TNF family members such as BAFF/BlyS,
which have been implicated in MZ B cell formation
(Mackay et al., 1999). Signaling downstream of one or
more of these receptors may activate NF?B, since it has
been established that NF?B p50 is required for MZ B
cell generation (Cariappa et al., 2000). However, given
our findings with Aiolos?/?mice, it might be worth in-
vestigating whether, at circumscribed times during B
celldevelopment, Pyk-2,CD19,andNF?B p50canactu-
ally function as negative regulators of BCR signaling.
Aiolos?/?(Wang et al., 1998) and Cr2?/?mice (Ahearn et al., 1996)
have been described previously. Xid (CBA/N) mice were obtained
from Jackson Labs. Mice were analyzed between 5–8 weeks after
birth. All animal studies were approved by institutional review
Generation of Double Mutant Mice with Aiolos and Btk Deficiency
Aiolos?/?male and Xid female mice were mated. Aiolos?/?F1 fe-
males were backcrossed to Aiolos?/?males, and F2 male progeny
were typed. Progeny were first typed by PCR as described earlier
(Wang et al., 1998), to ascertain if pups were Aiolos?/?or Aiolos?/?.
The Xid mouse carries a point mutation in exon 1 of Btk involving
codon 28 (R28C), which functionally alters the PH domain. A novel
PCR SSO approach was used to determine whether codon 28 en-
coded an R or a C. A 152 base pair fragment spanning the mutation
was obtained from genomic tail DNA using 5?-GATACTGGAGAG
CATCTTTCTGAA-3? (Thomas et al., 1993) as the forward primer and
5?-TGGGAAACTTACCCCACGTTC-3? (Rawlings et al., 1993) as the
reverse primer. The following parameters were used for forty cycles
of amplification: 94?C ? 5?, 94?C ? 1’, 60?C ? 2’, and 72?C ? 1’.
After confirming the PCR products on a 1.5% agarose gel, upto 5
?l of the product was spotted on nylon membrane filters. The filters
were prehybridized overnight at 54?C in a buffer containing 3 M
tetramethylammonium chloride, 50 mM Tris-HCl (pH 8.0), 2mM
EDTA, 5? Denhardts solution, 0.1% sodium dodecyl sulfate, and
100 ?g/ml of salmon sperm DNA. Two labeled oligonucleotide
probes, one wild-type (5?AACTTCAAGAAGCGCCTGT-3?) and the
other mutant (5?-AACTTCAAGAAGTGCCTGT-3?), which span R28C,
were used to identify the Xid allele. The SSO probes were 5? end-
labeled using [?-32P] ATP and T4 polynucleotide kinase. Hybridiza-
tion was carried out at 54?C for 2 hr. The filters were washed twice
at room temperature for 15 min each in a solution containing 2?
SSPE and 0.1% SDS followed by three washes for 10 min at 58?C
in a buffer containing 3 M tetramethylammonium chloride, 50 mM
Tris-HCl, 2 mM EDTA, and 0.1% SDS. Each filter was exposed twice
for autoradiography, once for 1–2 hr and subsequently for 14–16 hr.
The fidelity of the SSO probes was first checked by hybridizing with
control DNA obtained from known CBA/Ca and CBA/N mice.
Figure 7. A Model for MZ B Cell Generation
Enhanced BCR signaling facilitates follicular B cell maturation and
survival but impairs MZ B cell generation. Follicular B cells that do
not receive “strong” BCR signals may receive activation cues via
Pyk-2 (Guinamard et al., 2000) or NF?Bp50 (Cariappa et al., 2000)
to differentiate into MZ B cells.These cues may be initiated by BAFF
and BlyS (see text for details).
cells. There is considerable evidence in support of the
view that MZ B cells in the rat develop from recirculating
precursors of MZ B cells are follicular IgMhiIgDhiCD21hi
Fraction II B cells, which have not yet received a rela-
tively strong BCR signal that commits these cells to a
long-lived follicular fate.
It has been reported that mice expressing a VH81X Ig
heavy chain transgene exhibit a relative increase in the
number of MZ B cells inparallel with a marked reduction
in the number of follicular B cells (Oliver et al., 1999).
The decrease in follicular B cells and increase in MZ B
cells observed in mice expressing VH81X may well be
consistent with a strength of signal hypothesis for follic-
ular versus MZ B cell generation. VH81X containing anti-
gen receptors recognize phosphorylcholine with low af-
finity (Martin and Kearney, 2000a). One interpretation of
these results is that self-antigens may very weakly trig-
ger VH81X containing BCRs, thus favoring the develop-
ment of MZ B cells but providing suboptimal signals for
follicular B lymphocyte survival.
Overall, these data suggest that the strength of BCR
signaling may be inversely related to MZ B cell genera-
to give rise to MZ B cells but may require additional
Flow Cytometric Analysis
Flow cytometry was performed essentially as described in Cariappa
et al., (1999). Single cell suspensions were made from spleen follow-
ing lysis of erythrocytes. For surface staining, 1 ? 106cells were
incubated with 2.4G2 (anti-CD16/CD32 [Fc? III/II receptor], rat IgG2b,
ing antibodies: anti-IgD-biotin-clone 11-26 (Southern Biotechnol-
ogy), and anti-IgM-PE clone R6-60, anti-CD21/CD35-FITC clone
7G6, anti-CD1d-FITC clone 1B1, and anti-CD43-PE clone S7 (all
from Pharmingen). Biotinylated antibodies were revealed using
streptavidin-APC. For intracellular detection of CD21, the cells were
first surface stained with anti-IgM and anti-IgD and then permeabil-
ized by the method of Veis et al. (1993) prior to staining with anti-
Viable lymphoid cells as determined by forward and side scatter
were gated on and 30,000 events collected. F-III (IgMhiIgDlo) cells
were analyzed for the expression of CD21. Analysis was performed
on an Epics Elite ESP (Coulter Corp.) flow cytometer equipped with
an ultraviolet enhanced argon ion blue laser and a helium/neon red
laser. Negative controls were used to set voltage and single color
positive controls were used for electronic compensation. The col-
lected data were analyzed using Epics Elite analysis software and
FloJo version 3.2 (Tree Star Corp).
This work was supported by grants from the National Institutes of
Health (AI33507, CA69618, and AI42254) and the Massachusetts
Chapter of the Arthritis Foundation. We thank Iswar Hariharan and
Spyros Artavanis-Tsakonas for helpful discussions. We thank John
Daley, Suzan Kallanian, Dennis Sgroi, Taj Pathan, Michelle Connole,
and Rebbecca Levangie for their contributions.
Received October 11, 2000; revised March 22, 2001.
Immunofluorescence Staining of Tissue Sections
The method of Oliver et al. (1997) was used with minor modifications
(Cariappa et al., 2000). In brief, 5–6 ?m thick sections were cut, air-
dried, fixed in ice-cold acetone, blocked with normal horse serum,
and stained in sequence, with MOMA-1 (rat, IgG2a, tissue culture
supernatant, Serotec), biotin-SP-goat anti-rat IgG (H?L) (Jackson
Immunoresearch), andstreptavidin-PE (Pharmingen)for 30min. The
sections were then blocked with normal rat serum and stained with
anti-IgM-FITC clone R6-60.2 (Pharmingen). Sections were rinsed in
PBS (five times, 3 min each) between stainings but not after the
blocking steps. The sections were mounted in a polyvinyl alcohol-
a Zeiss Axioplan 2 fluorescence microscope equipped with appro-
priate filters. Images were acquired with a Sensys cooled CCD cam-
era (Photometrics Ltd.) and a Power Macintosh computer and pro-
cessed using Openlab version 2.0.9 (Improvision) and Adobe
Photoshop version 5.0.2 (Adobe).
Ahearn, J.M., Fischer, M.B., Croix, D., Goerg, S., Ma, M., Xia, J.,
Zhou, X., Howard, R.G., Rothstein, T.L., and Carroll, M.C. (1996).
Disruption of the Cr2 locus results in a reduction in B-1a cells and
in an impaired B cell response to T-dependent antigen. Immunity
Allman, D.M., Ferguson, S.E., Lentz, V.M., and Cancro, M.P. (1993).
Peripheral B cell maturation II. Heat stable antigen hi splenic B cells
are an immature developmental intermediate in the production of
long lived marrow derived B cells. J. Immunol. 151, 4431–4444.
Amano, M., Baumgarth, N., Dick, M.D., Brossay, L., Kronenberg,
M., Herzenberg, L., and Strober, S. (1998). CD1 expression defines
subsets of follicular and marginal zone B cells in the spleen: beta
haupt, B., Schlossman, S.F., and Nadler, L.M. (1985). Studies of in
vitro activation and differentiation of human B lymphocytes: pheno-
typic and functional characterization of the B cell population re-
sponding to anti-Ig antibody. J. Immunol. 134, 1516–1523.
Cariappa, A., Kim, T.J., and Pillai, S. (1999). Accelerated emigration
of B lymphocytes in the Xid mouse. J. Immunol. 162, 4417–4423.
Cariappa, A., Liou, H.-C., Horwitz, B.H., and Pillai, S. (2000). Nuclear
cytes. J. Exp. Med. 192, 1175–1182.
Cyster, J.G., and Goodnow, C.C. (1995). Protein tyrosine phospha-
tase 1Cnegatively regulatesantigen receptorsignaling inB lympho-
cytes and determines thresholds for negative selection. Immunity
Dammers, P.M., de Boer, N.K., Deenen, G.J., Nieuwenhuis, P., and
Kroese, F.G.M. (1999). The origin of marginal zone B cells in the rat.
Eur. J. Immunol. 29, 1522–1531.
Fearon, D.T., and Carroll, M.C. (2000). Regulation of B lymphocyte
responses to foreign and self-antigens by the CD19/CD21 complex.
Annu. Rev. Immunol. 18, 393–422.
Gu, H., Tarlinton, D., Muller, W., Rajewsky, K., and Forster, I. (1991).
Most peripheral B cells are ligand selected. J. Exp. Med. 173, 1357–
Guinamard, R., Okigaki, M., Schlessinger, J., and Ravetch, J.V.
(2000). Absence of marginal zone B cells in Pyk-2 deficient mice
defines their role in the humoral response. Nat. Immunol. 1, 31–36.
Hardy, R.R., Hayakawa, K., Haaijman, J., and Herzenberg, L.A.
(1982). B-cell subpopulations identified by two colour fluorescence
analysis. Nature 297, 589–591.
Hayakawa, K., Asano, M., Shinton,S.A., Gui, M., Allman, D., Stewart,
C.L., Silver, J., and Hardy, R.R. (1999). Positive selection of natural
autoreactive B cells. Science 285, 113–116.
Khan,W.N.,Alt, F.W.,Gerstein,R.M.,Malynn,B.A., Larsson,I.,Rath-
bun, G., Davidson, L., Muller, S., Kantor, A.B., Herzenberg, L.A., et
al. (1995). Defective B cell development and function in Btk deficient
mice. Immunity 3, 283–299.
Kraal, G. (1992). Cells in the marginal zone of the spleen. Int. Rev.
Cytol. 132, 31–74.
Kraal, G., Hoeben, K., and Janse, M. (1988). Splenic microenviron-
ment of the CBA/N mouse: immunohistochemical analysis using
monoclonal antibodies against lymphocytes and non-lymphoid
cells. Am. J. Anat. 182, 148–154.
Lam, K.-P., Kuhn, R., and Rajewsky, K. (1997). In vivo ablation of
Calcium Flux Assay
The method of Wilson et al. (1987) was essentially followed with
some modifications. Single cell suspensions of splenocytes were
prepared following erythrocyte lysis. Aliquots of 3 ? 106cells sus-
pended in Iscove’s Modified Dulbecco Medium (IMDM) containing
10% FCS were treated with 2.4G2 and surface stained with anti-
IgD-PE-clone 11-26 (Southern Biotechnology). Following a wash,
the cells were resuspended in 500 ?l of the same medium and
incubated with 2 ?M of Indo-1, AM (Molecular Probes) for 30’ at
37?C. The Indo-1 loaded cells were allowed to equilibrate for 5?,
washed at room temperature using Hanks’ Balanced Salt Solution
(HBSS) containing 2 mM Hepes, and resuspended in the same me-
dium. 1 ? 106aliquots in a volume of 500 ?l were analyzed in an
Epics Elite ESP flow cytometer (Coulter Corp.). IgD positive cells
were gated on, and a baseline of events was recorded for 30”. The
cells were then stimulated with goat anti-mouse IgM (Fab?)2(? chain
specific), 10 ?g/ml, or goat anti-mouse IgM (? chain specific), 40
?g/ml (both from Jackson Immunoresearch Labs), and events re-
corded for a further 5?. The collected data were analyzed using
Epics Elite analysis software and FloJo version 3.2 (Tree Star Corp).
CD21 mRNA Expression
A 2 kb BamH1 fragment of the CD21/CD35 cDNA was used as a
probe for CD21 expression when probing RNA samples from wild-
type and Aiolos?/?splenic B cells on a Northern blot. Splenic B
cells were isolated by T cell depletion of splenocyte suspensions
using anti-CD4 (GK1.5, rat IgG2b, ? and anti-CD8 (3.155, rat IgM),
and Scientific Corp).
Lymphocyte Proliferation Assay
5 ? 104splenocytes were cultured in quadruplicate in flat-bottomed
96-well plates with media alone or with media ? goat anti-mouse
IgM (Fab?)2(? chain specific) (10 ?g/ml, Jackson Immunoresearch
Labs) for 40 hr and then pulsed with 1 ?Ci of [3H] thymidine (New
England Nuclear) for an additional 8 hr. Cells were harvested and
incorporated radioactivity was counted in a betaplate reader.
P-values for differences between groups was determined by the
Mann-Whitney U test using StatView version 5.0.1.
Strength of Signal and MZ B Lymphocytes
surface immunoglobulin on mature B cells by inducible gene tar-
geting results in rapid death. Cell 90, 1073–1083.
Levine, M.H., Haberman, A.M., Sant’Angelo, D.B., Hannum, L.G.,
Cancro, M.P., Janeway, C.A., and Schlomchik, M.J. (2000). A B cell
receptor-specific selection step governs immature to mature B cell
differentiation. Proc. Natl. Acad. Sci. USA 97, 2713–2748.
Liu, Y.J., Lortan, J.E., Oldfield, S., and MacLennan, I.C. (1988). CBA/
ulin phenotype. Adv. Exp. Med. Biol. 237, 105–111.
Loder, F., Mutschler, B., Ray, R.J., Paige, C.J., Sideras, P., Torres,
R., Lamers, M.C., and Carsetti, R. (1999). B cell development in the
spleen takesplace indiscrete stepsand isdetermined by thequality
of B cell receptor derived signals. J. Exp. Med. 190, 75–89.
Mackay, F., Woodcock, S.A., Lawton, P., Ambrose, C., Baetscher,
M., Schneider, P., Tschopp, J., and Browning, J.L. (1999). Mice
immune manifestations. J. Exp. Med. 190, 1697–1710.
MacLennan, I.C.M. (1998). B cell receptor regulation of peripheral
B cells. Curr. Opin. Immunol. 10, 220–225.
MacLennan, I.C.M., Gray, D., Kumararatne, D.S., and Bazin, H.
(1982). The lymphocytes of splenic marginal zones: a distinct B cell
lineage. Immunol. Today 3, 305–310.
Makowska, A., Faizunnessa, N.N., Anderson, P., Midtvedt, T., and
Cardell, S. (1999). CD1high B cells: a population of mixed origin.
Eur. J. Immunol. 29, 3285–3295.
Mandik-Nayak, L., Seo, S.-j., Sokol, C., Potts, K.M., Bui, A., and
Erikson, J. (1999). MRL-lpr/lpr mice exhibit a defect in maintaining
developmental arrest and follicular exclusion of anti-double-
stranded DNA B cells. J. Exp. Med. 189, 1799–1814.
Martin, F., and Kearney, J.F. (2000a). B-cell subsets and the mature
preimmune repertoire. Marginal zone and B1 B cells as part of a
“natural immune memory”. Immun. Rev. 175, 70–79.
Martin, F., and Kearney, J.F. (2000b). Positive selection from newly
formed to marginal zone B cells depends on the rate of clonal
production, CD19, and btk. Immunity 12, 39–49.
Morgan, B., Sun, L., Avitahl, N., Andrikopulos, K., Ikeda, T., Gonza-
les, E., Wu, P., Neben, S., and Georgopoulos, K. (1997). Aiolos, a
lymphoid restricted transcription factor that interacts with Ikaros to
regulate lymphocyte differentiation. EMBO J. 16, 2004–2013.
Neuberger, M.S. (1997). Antigen receptor signaling gives lympho-
cytes a long life. Cell 90, 971–973.
O’Rourke, L.M., Tooze, R., Turner, M., Sandoval, D.M., Carter, R.H.,
Tybulevicz, V.L., and Fearon, D.T. (1998). CD19 as a membrane-
anchored adaptor protein of B lymphocytes: costimulation of lipid
and protein kinases by recruitment of Vav. Immunity 8, 635–645.
Oliver, A.M., Martin, F., Gartland, G.L., Carter, R.H., and Kearney,
J.F.(1997). MarginalzoneB cellsexhibitunique activation,prolifera-
tive and immunoglobulin secretory responses. Eur. J. Immunol. 27,
Oliver, A.M., Martin, F., and Kearney, J.F. (1999). IgMhighCD21high
cells more rapidly than the bulk of follicular B cells. J. Immunol. 162,
Pillai, S. (1999). The chosen few? Positive selection and the genera-
tion of naive B lymphocytes. Immunity 10, 493–502.
Rawlings, D.J., Saffran, D.C., Tsukada, S., Largaespada, D.A., Gri-
maldi, J.C., Cohen, L., Mohr, R.N., Bazan, J.F., Howard, M., Cope-
land, N.G., et al. (1993). Mutation of unique region of Bruton’s tyro-
sine kinase in immunodeficient Xid mice. Science 261, 358–361.
Roark, J.H., Park, S.H., Jayawardena, J., Kavita, U., Shannon, M.,
and Bendelac, A. (1998). CD1.1 expression by mouse antigen pre-
senting cellsand marginalzone Bcells. J.Immunol. 160,3121–3127.
R.J., Baron, A., and Holers, V.M. (1997). Mouse complement recep-
tors Type 1 (CR1;CD35) and Type 2 (CR2; CD21). Expression on
opment of autoimmunity in MRL/lpr mice. J. Immunol. 159, 1557–
Thomas, J.D., Sideras, P., Smith, C.I.E., Vorechovsky, I., Chapman,
V., and Paul, W.E. (1993). Co-localization of X-linked agammaglob-
ulinemia and X-linked immunodeficiency genes. Science 261,
Torres, R.M., Flaswinkel, H., Reth, M., and Rajewsky, K. (1996).
Aberrant B cell development and immune response in mice with a
compromised BCR complex. Science 272, 1804–1808.
Tuveson, D.A., Carter, R.H., Soltoff, S.P., and Fearon, D.T. (1993).
dylinositol 3-kinase. Science 260, 986–989.
Veis, D.J., Sentman, C.L., Bach, E.A., and Korsmeyer, S.J. (1993).
Expression of the Bcl-2 protein in murine and human thymocytes
and in peripheral T lymphocytes. J. Immunol. 151, 2546–2554.
Wang, J.H., Avitahl, N., Cariappa, A., Friedrich, C., Ikeda, T., Renold,
A., Andrikopoulos, K., Liang, L., Pillai, S., Morgan, B.A., et al. (1998).
Aiolos regulates B cell activation and maturation to effector state.
Immunity 9, 543–554.
Wilson, J.G., Ratnoff, W.D., Schur, P.H., and Fearon, D.T. (1986).
Decreased expression of the C3b/C4b receptor (CR1) and the C3d
receptor (CR2) on B lymphocytes and of CR1 on neutrophils of
patients with systemic lupus erythematosus. Arthritis Rheum. 29,
Wilson, H.A., Greenblatt, D., Taylor, C.W., Putney, J.W., Tsien, R.Y.,
response to anti-Ig is diminished by membrane immunoglobulin
cross-linkage to the Fcg receptor. J. Immunol. 138, 1712–1718.