Distinct signaling mechanisms activate the target of rapamycin in

Document Sample
Distinct signaling mechanisms activate the target of rapamycin in Powered By Docstoc
					Eur. J. Immunol. 2007. 37: 2923–2936                                                                   Leukocyte signaling    2923

                                Leukocyte signaling

Distinct signaling mechanisms activate the target of
rapamycin in response to different B-cell stimuli
Amber C. Donahue and David A. Fruman

Department of Molecular Biology & Biochemistry, and Center for Immunology,
University of California-Irvine, Irvine, CA, USA

Phosphoinositide 3-kinase (PI3K) and the mammalian target of rapamycin (mTOR), a                         Received 19/3/07
downstream kinase, are both required for proliferation of splenic B cells. However, the                   Revised 31/5/07
                                                                                                         Accepted 18/7/07
functions of PI3K and mTOR in response to different stimuli and among B cell subsets
have not been fully elucidated. We used flow cytometry and magnetic cell sorting to             [DOI 10.1002/eji.200737281]
examine the requirement for PI3K and mTOR in responses of splenic B cell subsets to
BCR and LPS stimulation. BCR-mediated phosphorylation of Akt and Erk is sensitive to
the PI3K catalytic inhibitor wortmannin in both marginal zone (MZ) and follicular (FO)
cells. BCR-mediated mTOR activation in both subsets is inhibited by wortmannin,
though less strongly in MZ cells. In contrast, LPS-induced mTOR signaling is strikingly
resistant to wortmannin in both subsets. Similarly, functional responses to LPS are
partially wortmannin resistant yet sensitive to mTOR inhibition by rapamycin. We also                       Key words:
observed mitogen-independent mTOR activity that is regulated by nutrient availability,              B lymphocyte Á Cell
and is significantly elevated in MZ cells relative to FO cells. These data define both                  activation Á Cell
similarities and differences in PI3K/mTOR signaling mechanisms in MZ and FO cells,                         proliferation
and suggest that mTOR signaling can occur in the absence of PI3K activation to promote            Á Lipopolysaccharide
B cell responses to LPS.                                                                         Á Signal transduction

Introduction                                                    in part by their permanent location in the MZ
                                                                surrounding splenic follicles [1]. Previous work has
Newly formed B cells emigrate from the bone marrow to           suggested that these two mature subsets have evolved to
the spleen, where they undergo further maturation               play different roles in the adaptive immune response
through several immature phases to become one of two            [3]. In vitro, FO cells respond to B cell antigen receptor
mature subsets, follicular (FO) or marginal zone (MZ)           (BCR) and lipopolysaccharide (LPS) stimulation with
cells (reviewed in [1, 2]). FO cells recirculate through-       moderate levels of activation, proliferation, and differ-
out the blood, lymph, and lymphoid follicles, and               entiation. In contrast, BCR stimulation leads to apop-
represent the largest population found in the spleen, at        tosis of MZ cells, whereas LPS treatment induces more
about 60% of the total B cells. In contrast, MZ cells           rapid and robust levels of proliferation and differentia-
comprise only 5–10% of splenic B cells and are defined          tion of MZ cells than FO cells [3]. MZ cells possess an
                                                                antigen receptor repertoire biased towards germ-line-
Correspondence: David A. Fruman, Department of Molecular
                                                                encoded T-independent specificities that are self-reac-
Biology and Biochemistry, 3242 McGaugh Hall, University of      tive at a low level, generating cells described as "pre-
California-Irvine, Irvine, CA 92697–3900, USA                   activated". The location of MZ cells near splenic sinuses
Fax: +1-949-8248551                                             and their other unique properties suggest that this
e-mail: dfruman@uci.edu                                         subset serves as a first line of defense against blood-
Abbreviations: ASC: antibody-secreting cell Á cRPMI: complete
                                                                borne pathogens in vivo and as a link between the innate
RPMI medium Á dFBS: dialyzed fetal bovine serum Á
FO: follicular Á FOE: follicular-enriched Á mTOR: mammalian
                                                                and adaptive immune systems [3].
target of rapamycin Á MZ: marginal zone Á MZE: marginal zone-       Although there are clear differences in BCR-
enriched                                                        mediated signal transduction between immature and

f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                       www.eji-journal.eu
2924   Amber C. Donahue and David A. Fruman                                                    Eur. J. Immunol. 2007. 37: 2923–2936

       mature B cells [4, 5], it is not yet clear whether signaling   scarcity of MZ cells in wild-type (WT) mice and their
       mechanisms differ in FO vs. MZ cells. Most studies of          absence in p85a or p110d mutant mice has hindered the
       BCR- and LPS-mediated signaling to date have made use          assessment of PI3K function in this subset [13, 14]. Here,
       of unfractionated total B cells, in which the responses of     we have used flow cytometry (FACS)-based assays to
       the predominant FO cells mask the potentially hetero-          examine signaling in FO and MZ cells in mixed
       geneous responses of the other subsets. In one study           populations of WT splenocytes. We find that the two
       directly comparing purified FO and MZ cells, stimulation       subsets show similar dependence on PI3K activity for
       with matched BCR cross-linking strength yielded similar        linking BCR engagement to downstream signaling
       signal responses downstream of the BCR, including early        responses. The most striking difference between subsets
       tyrosine phosphorylation and Ca2+ mobilization [6].            is that, in the absence of mitogens, MZ cells maintain
       Various surface receptors and signaling molecules are          markedly higher nutrient-dependent mTOR activity
       differentially expressed in MZ and FO cells [7, 8].            than FO cells. In addition, we demonstrate that LPS-
       However, it has not been determined whether any of             driven B cell activation is only partly dependent on PI3K,
       these differences can account for the differential             yet strictly dependent on mTOR function [28]. The
       responses of FO and MZ cells to BCR cross-linking or           results suggest possible mechanisms for the more rapid
       LPS.                                                           proliferation and differentiation of MZ cells and help
           Phosphoinositide 3-kinase (PI3K) is a lipid kinase         resolve questions about the role of PI3K in B cell
       whose activation downstream of multiple receptors in           responses to LPS.
       B cells generates the important second messenger
       PtdIns(3,4,5)P3 [9]. The broad-spectrum PI3K inhibitors        Results
       wortmannin and LY294002 block B cell proliferation,
       with a greater effect on cells stimulated by BCR cross-        Flow cytometric discrimination of signaling states
       linking than by LPS [10, 11]. Gene disruption studies in       in splenic B cell subsets
       mice have shown that class IA PI3K is important for the
       development of B cells in the bone marrow and spleen,          For the signaling measurements described below, we
       and for B cell activation and proliferation [11–19].           stained total murine splenocytes for two or three
       Class IA PI3K predominantly exists in the cell as a            developmental surface markers and used other fluores-
       heterodimer of an 85-kD regulatory subunit and a               cence channels for measurement of intracellular phos-
       110-kD catalytic subunit, with each subunit having             phoproteins. In most experiments, we used antibodies to
       multiple isoforms [9]. B cells lacking the p85a                CD24 and CD1d [29] to differentiate between FO
       regulatory isoform exhibit developmental defects,              (CD24loCD1dlo), MZ (CD24intCD1dhi) and immature
       including near ablation of the MZ subset and the               (CD24hiCD1dint) B cells (Fig. 1A). We focused our
       appearance of a smaller pool of FO-like cells with altered     attention on the mature subsets (FO and MZ) that are
       surface phenotype [11, 14, 18, 19]. BCR-triggered signals      clearly distinguished by this staining approach, as
       are selectively impaired in p85a-deficient B cells, with       immature B cells consist of several phenotypically
       reduced Ca2+ and NF-jB activation and lesser defects in        distinct populations which can be resolved only by
       activation of Akt and the mammalian target of                  more complex staining strategies [30–33]. We chose
       rapamycin (mTOR) [15, 18, 20]. Similar defects have            CD24/CD1d staining in part to avoid ligating CD21 and
       been observed in B cells lacking p110d or treated with a       CD23; CD21 is a component of the B cell coreceptor
       selective inhibitor of p110d [12, 13, 16, 17].                 complex that regulates PI3K, and CD23 is expressed at
           Rapamycin, a highly selective inhibitor of mTOR, is        markedly different levels on FO and MZ cells [29].
       well known as an inhibitor of T cell proliferation, but also   Nevertheless, as shown in Fig. 2, signaling responses
       potently suppresses B cell proliferation and differentia-      among subsets were similar when MZ cells were
       tion [21, 22]. mTOR integrates signals from mitogens,          distinguished using antibodies to CD21/CD23/CD24
       serum factors and nutrient availability to regulate            [33]. Thus, the results described below are not likely to
       protein translation, mitochondrial function and other          be biased by responses to the staining antibodies.
       cellular processes [23, 24]. mTOR can be activated                 To compare BCR signaling responses among subsets,
       downstream of PI3K, through pathways involving the             we controlled for potential differences in signal strength
       kinases PDK1 and Akt [10, 15, 25, 26]. However, mTOR           by stimulating cells with anti-kappa antibody (F(ab')2).
       activity is PI3K/Akt independent in some human B cell          This reagent cross-links all surface Ig on 90–95% of cells
       lymphoma lines [27].                                           of all subsets regardless of heavy-chain isotype [6]. In
           PI3K and mTOR signaling mechanisms have not been           contrast, anti-IgM or anti-IgD antibodies yield differ-
       fully defined for each individual B cell subset, as most       ential cross-linking of surface Ig among subsets accord-
       data have been derived from heterogeneous populations          ing to their expression of these isotypes (FO are IgMlo/
       of total splenic B cells (CD43– or CD19+ or B220+). The        IgDhi, MZ are IgMhi/IgDlo).

       f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                       www.eji-journal.eu
Eur. J. Immunol. 2007. 37: 2923–2936                                                                        Leukocyte signaling     2925

                                                               Figure 1. Phosphorylation of Akt and Erk is PI3K dependent in
                                                               BCR-stimulated FO and MZ cells. Splenocytes were stained for
                                                               CD24 and CD1d (A), incubated in medium for 1 h, pretreated
                                                               with wortmannin (WM) or U0126 (U01) where indicated, then
                                                               left unstimulated or stimulated with anti-kappa antibody
                                                               (1 min). Fixed and permeabilized cells were then stained
                                                               intracellularly for pAkt (S473) (B, C) or pErk (T202/Y204) (D, E).
                                                               (A) A representative FACS dot plot of CD24 vs. CD1d (left panel).
                                                               FO and MZ cells were conservatively gated as shown, in this
                                                               and all subsequent experiments. Right panel shows a
                                                               representative histogram of population cell sizes as deter-
                                                               mined by forward scatter (FSC), verifying the larger average
                                                               size of MZ vs. FO cells based on the gates shown. Imm,
                                                               immature. (B) Representative histograms depicting levels of
                                                               pAkt in unstimulated (UnTx) cells or in cells stimulated with
                                                               anti-kappa antibody (anti-j). (C) Mean fold increase of pAkt
                                                               with respect to unstimulated cells treated with wortmannin
                                                               (UnTx+WM). n = 5. (D) Representative histograms of pErk in
                                                               UnTx or anti-kappa antibody-stimulated cells. (E) Mean fold
                                                               increase of pErk with respect to unstimulated cells treated with
                                                               U0126. n = 5.

                                                               studies of Akt and Erk regulation by PI3K have examined
                                                               heterogeneous populations of splenic B cells [34, 35]. To
                                                               examine the PI3K dependence of these events in MZ and
                                                               FO cells, we made use of flow cytometry-based assays
                                                               [36–40] to assess the phosphorylation states of Akt and
                                                               Erk following stimulation with anti-kappa antibody. As
                                                               the reference sample for nonspecific staining, we used
                                                               unstimulated cells treated with wortmannin for pAkt
                                                               assays, or unstimulated cells treated with the MEK
                                                               inhibitor U0126 for pErk assays. In phosflow assays,
                                                               inhibitor-treated cells are a more appropriate control
                                                               than cells stained with isotype-matched antibody [38,
                                                               41]. As shown in Fig. 1B, the basal level of pAkt (Ser473)
                                                               in both subsets increased modestly but consistently with
                                                               stimulation by anti-kappa antibody. Wortmannin was
                                                               equally effective at inhibiting pAkt in MZ and FO cells in
                                                               anti-kappa antibody-stimulated samples (Fig. 1B). When
                                                               we calculated the fold increase of pAkt with respect to
                                                               wortmannin-treated unstimulated cells across several
                                                               experiments, we saw that anti-kappa antibody-stimu-
                                                               lated FO and MZ cells showed a similar degree of
                                                               inhibition by wortmannin (Fig. 1C).
                                                                   We observed no appreciable phosphorylation of Erk
                                                               in unstimulated MZ and FO cells, whereas BCR cross-
                                                               linking with anti-kappa antibody led to clear increases in
                                                               pErk. This increase in pErk was inhibited equally in MZ
                                                               and FO cells, both by treatment with wortmannin or
Phosphorylation of Akt and Erk is PI3K dependent               with U0126 (Fig. 1D, E). These results demonstrate that
in both subsets                                                the pathways leading to phosphorylation of Akt and Erk
                                                               in FO and MZ cells are similarly PI3K dependent, as
Phosphorylation of Akt is strictly PI3K dependent, and         evidenced by their sensitivity to PI3K inhibition by
assays for Akt phosphorylation (pAkt) are often used as        wortmannin. We also observed similar wortmannin
surrogate readouts for PI3K activity in cells. Phosphor-       sensitivity of anti-kappa antibody-evoked Ca2+ mobili-
ylation of the MAP kinases Erk1 and Erk2 is also largely       zation in FO and MZ subsets (data not shown). Of note,
PI3K dependent in BCR-stimulated B cells. Previous             we have attempted repeatedly to stain B cells with an

f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                           www.eji-journal.eu
2926   Amber C. Donahue and David A. Fruman                                                        Eur. J. Immunol. 2007. 37: 2923–2936

                                                                       Figure 2. Phosphorylation of S6 can occur in the absence of
                                                                       mitogens and is partially wortmannin resistant in MZ cells
                                                                       stimulated with anti-kappa antibody. Splenocytes stained as in
                                                                       Fig. 1 (A, B), or stained for CD21/CD23/CD24 (C, D) were
                                                                       incubated in medium for 1 h, treated with wortmannin
                                                                       (WM) or rapamycin (Rap) for 15 min, left unstimulated, or
                                                                       stimulated with anti-kappa antibody. Fixed and permeabilized
                                                                       cells were then stained intracellularly for pS6 (S235/236).
                                                                       (A, D) Representative histograms of FO and MZ cells left
                                                                       untreated (UnTx; left) or stimulated with anti-kappa antibody
                                                                       (anti-j; right) for 15 min. Subsets were distinguished using
                                                                       (A) CD1d/CD24 or (D) CD21/CD23/CD24 [33]. (B) Mean fold
                                                                       increase of pS6 with respect to unstimulated cells treated
                                                                       with rapamycin in FO and MZ cells. *p <0.05; **p <0.01; n = 13
                                                                       (all using CD1d vs. CD24). For the statistical comparison of
                                                                       unstimulated/wortmannin-treated vs. unstimulated/rapamy-
                                                                       cin-treated, we used the actual mean fluorescence intensity
                                                                       data rather than comparing fold changes; this was done
                                                                       because the fold change value for the unstimulated/rapamycin
                                                                       samples was by definition set to 1.0 and invariant.
                                                                       (C) Representative FACS dot plot of CD21 vs. CD24 staining
                                                                       (left panel) and histogram of CD23 staining (right panel) within
                                                                       the CD21hiCD24hi population. Inset shows larger cell size (FSC)
                                                                       of MZ cells identified by this method.

                                                                       S6 is phosphorylated at these sites by S6K1 and S6K2,
                                                                       kinases that are activated by mTOR-dependent phos-
                                                                       phorylation [23, 24]. Previously, we used a FACS-based
                                                                       assay for pS6 to demonstrate an increase in S6
                                                                       phosphorylation following BCR engagement, gating
                                                                       on total (B220+) B cells [15]. Using the surface staining
                                                                       protocol described in Fig. 1, we examined the degree of
                                                                       S6 phosphorylation in FO and MZ cells under different
                                                                       conditions (Fig. 2A, B). Considering that rapamycin is a
                                                                       highly specific inhibitor of mTOR upstream of S6
                                                                       phosphorylation, we used rapamycin-treated samples
                                                                       as our reference for background staining.
                                                                           When splenocytes were incubated in complete
                                                                       medium alone at 37 C for 1 h without addition of
                                                                       mitogens, MZ cells showed a significantly higher degree
       antibody reported to be specific for PtdIns(3,4,5)P3 [42,       of pS6 than FO cells. A representative histogram is
       43], but have not observed wortmannin/LY294002-                 shown in Fig. 2A, and a graph of compiled experiments
       sensitive changes in fluorescence detection following           in Fig. 2B. In both subsets, mitogen-independent S6
       mature B cell stimulation [41].                                 phosphorylation was inhibited by wortmannin, but to a
                                                                       significantly lesser extent than by rapamycin (Fig. 2B).
       BCR-mediated S6 phosphorylation requires PI3K                   Note that the statistical comparison compared mean
       and other signals                                               fluorescence of wortmannin-pretreated unstimulated
                                                                       cells relative to the rapamycin-pretreated unstimulated
       The role of PI3K in activation of mTOR and its                  sample (denoted by brackets in Fig. 2B). These results
       downstream targets varies in different cell contexts            suggest that some S6 phosphorylation can occur in the
       [27]. We sought to compare the requirement for PI3K in          absence of B cell mitogens, and is regulated by both
       mTOR signaling under different stimulation conditions,          PI3K-dependent and PI3K-independent inputs.
       and among splenic B cell subsets. When we tested                    Both MZ and FO cells stimulated with anti-kappa
       various phosphospecific antibodies as phosflow read-            antibody for 15 min showed strong pS6 staining,
       outs of mTOR activation, only an antibody specific for          relative to cells pretreated with rapamycin (Fig. 2A).
       phosphorylated ribosomal protein S6 (pS6 S235/236)              Wortmannin pretreatment of anti-kappa antibody-stim-
       yielded a robust signal that was blocked by rapamycin.          ulated cells inhibited pS6 in both subsets. However, in

       f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                           www.eji-journal.eu
Eur. J. Immunol. 2007. 37: 2923–2936                                                                    Leukocyte signaling   2927

MZ cells the pS6 remained significantly higher than in         that both of these nutrient-sensing pathways contribute
rapamycin-pretreated cells (Fig. 2A, B). We obtained           to mTOR activation in cultured MZ cells. In the absence
nearly identical patterns of basal and stimulated pS6          of serum, removal of leucine or glucose reduced S6
when we stained splenocytes for CD21/CD23/CD24 to              phosphorylation to an even greater extent (Fig. 3B). We
distinguish FO and MZ cells (Fig. 2C, D). Quantification       did not observe significant wortmannin resistance in MZ
of multiple experiments using CD1d/CD24-stained                cells cultured in the absence of leucine or glucose
splenocytes showed that in anti-kappa antibody-stim-           (Fig. 3B).
ulated samples, the average percent inhibition of S6               As observed in the experiments depicted in Fig. 2, FO
phosphorylation by wortmannin was lower in MZ cells            cells cultured without mitogens exhibited very little pS6
than in FO cells (see below). Thus, the contribution of        with respect to unstimulated cells treated with rapamy-
PI3K activity to S6 phosphorylation, and presumably            cin, regardless of serum or nutrient content (Fig. 3A).
mTOR activation, is different in FO and MZ cells               Quantitation of several experiments showed that the
following BCR engagement.                                      small amount of mitogen-independent pS6 in FO cells
                                                               was regulated similarly to MZ cells, with significant
Regulation of S6 phosphorylation by nutrients                  contributions from serum, amino acids and glucose
                                                               (Fig. 3C). Collectively, the experiments analyzing media
mTOR is an evolutionarily conserved signaling protein          components indicate that serum, amino acids and
that is regulated both by growth factors and by nutrient       glucose all contribute to S6 phosphorylation in un-
availability. In mammalian cells, mTOR integrates input        stimulated mature B cells. The influence of the media
from serum components, amino acid levels, and glucose          components on S6 phosphorylation is clearly stronger in
availability [24, 44]. The contribution of these inputs        the MZ subset. We also observed that amino acids and
and their dependence on PI3K has not been extensively          glucose, but not serum components, are important for
studied in primary lymphocytes. The high degree of pS6         maximal S6 phosphorylation following BCR stimulation
observed in MZ cells that had been incubated for 1 h in        of both FO and MZ cells (data not shown).
the absence of mitogens (Fig. 2) was not observed when
splenocytes were analyzed immediately after isolation          Evidence for PI3K-independent mTOR activation
(Fig. 3A). This suggested that one or more components          by LPS
of the culture medium might promote mTOR activity
selectively in MZ cells. To investigate this possibility, we   LPS is a potent B cell mitogen that induces activation via
cultured splenocytes in a panel of different media.            a distinct set of signals and gene expression changes in
Fig. 3A shows representative histograms and Fig. 3B, C         comparison with those observed following BCR cross-
shows the fold increase of pS6 staining with respect to        linking [46, 47]. Exposure of B cells to LPS triggers
unrested cells following incubation for 1 h in our             phosphorylation of Akt and activation of S6 kinase [10,
standard medium (Std), complete RPMI medium                    46], but the magnitude and regulation of these
(cRPMI) supplemented with dialyzed fetal bovine serum          responses have not been examined in B cell subsets.
(dFBS), and cRPMI supplemented with 1% bovine                  LPS stimulation led to a modest increase in pAkt in FO
serum albumin (BSA). MZ cells rested in cRPMI                  cells with respect to wortmannin-treated unstimulated
supplemented with dFBS showed a similar level of               cells, but little if any pAkt in MZ cells (Fig. 4A). In
pS6 to MZ cells rested in standard medium (Fig. 3A, B).        contrast, S6 phosphorylation showed a more robust fold
Treatment with wortmannin still only caused partial            increase following LPS exposure in both subsets
inhibition of pS6 in MZ cells incubated in standard            (Fig. 4B). Both subsets also demonstrated a partial
medium or cRPMI/dFBS (Fig. 3A, B). Replacement of              resistance to wortmannin that was significant in MZ cells
serum with BSA led to a significant reduction in pS6           (Fig. 4B) and nearly so in FO cells (p = 0.095). The
staining, suggesting that one or more non-dialyzable           degree of wortmannin inhibition of the pS6 response
serum components contribute to the augmented S6                was consistently lower in LPS-treated cells than in BCR-
phosphorylation in MZ cells (Fig. 3A, B).                      stimulated cells, especially in the FO subset (Fig. 4C; also
    We also investigated the effects of withdrawing            compare Fig. 2C and 4B).
amino acids or glucose. The amino acid-sensing pathway             Wortmannin inhibits PI3K enzymes of all classes, as
that regulates mTOR is most sensitive to the presence or       well as some other cellular enzymes. To examine more
absence of leucine [45]. Therefore, we incubated               specifically the role of class IA PI3K in mTOR regulation,
splenocytes in RPMI lacking leucine (–Leu), or RPMI            we used spleen cells from mice lacking the predominant
lacking glucose (–Gluc). Both media were supplemented          regulatory isoform p85a [15, 19]. These mice
with dFBS, from which these small molecules had been           (Pik3r1tm1; termed here p85aKO) have fewer FO and
removed by dialysis. Removal of either leucine or              MZ cells than WT mice, but we found the residual
glucose reduced pS6 in MZ cells (Fig. 3A, B), indicating       populations to be sufficient for FACS-based signaling

f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                       www.eji-journal.eu
2928   Amber C. Donahue and David A. Fruman                                                             Eur. J. Immunol. 2007. 37: 2923–2936

       Figure 3. S6 phosphorylation is dependent in part on nutrients and serum components in FO and MZ cells. Splenocytes stained as
       in Fig. 1 were rested for 1 h in the indicated media, or kept on ice (No Rest) until fixation. Rested cells were then treated with
       wortmannin (+WM), rapamycin (+Rap), or vehicle (No Drug) for 15 min. Fixed and permeabilized cells were stained intracellularly
       for pS6 (S235/236). (A) Representative histograms of pS6-FITC in FO and MZ cells rested for 1 h in the media formulations
       indicated. Multiple panels are shown for clarity. (B) Fold increase of pS6, with respect to the No Rest sample, in MZ cells rested for
       1 h in the indicated media and drug-treated for 15 min. (C) Fold increase of pS6, with respect to the No Rest sample, in FO cells
       rested for 1 h and drug-treated for 15 min. cRPMI, complete RPMI; dFBS, dialyzed FBS; BSA, 1% BSA; –Leu, RPMI lacking leucine;
       –Gluc, RPMI lacking glucose. Brackets indicate statistical comparisons relative to the left-most data point in the bracket, with error
       bars depicting the SEM of fold change values for that sample. *p <0.05, **p <0.01, n = at least 3. In some cases, error bars were
       overlapping, yet the differences were statistically significant. This occurs when the fold change ratios among paired samples are
       highly reproducible, even when the absolute fold changes vary between experiments. For example, in (B) where we compared MZ
       cells cultured in cRPMI + BSA, wortmannin vs. rapamycin, the mean Æ SEM of the ratio of the fold changes was 1.47 Æ 0.05 (n = 4).

       f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                                 www.eji-journal.eu
Eur. J. Immunol. 2007. 37: 2923–2936                                                                      Leukocyte signaling   2929

                                                                  studies. FO and MZ cells from p85aKO mice showed a
                                                                  degree of S6 phosphorylation that was nearly equivalent
                                                                  to wortmannin-treated WT cells, either in the absence of
                                                                  mitogens or following LPS stimulation (Fig. 4D).
                                                                  Treatment of p85aKO splenocytes with wortmannin
                                                                  had little effect under these conditions. In contrast,
                                                                  p85a-deficient FO and MZ cells stimulated with anti-
                                                                  kappa antibody showed pS6 levels similar to WT without
                                                                  drug treatment, and could be significantly inhibited by
                                                                  wortmannin. These results are consistent with our
                                                                  previous studies of basal and BCR-stimulated pS6 in
                                                                  total splenic B cells from WT and p85aKO mice [15].
                                                                  Together, these findings suggest that p85a is required for
                                                                  the wortmannin-sensitive component of basal and LPS-
                                                                  stimulated mTOR activity, but dispensable for BCR-
                                                                  mediated mTOR activity.
                                                                      These FACS-based analyses of pS6 suggested a general
                                                                  difference in mTOR regulation in B cells stimulated by
                                                                  LPS compared to cells stimulated by BCR cross-linking. To
                                                                  test this hypothesis further, we performed immunoblot
                                                                  analysis on lysates of mature B cell populations
                                                                  containing both FO and MZ cells. Cells were treated
                                                                  with vehicle or with wortmannin or rapamycin, then
                                                                  activated with anti-kappa antibody or LPS for various
                                                                  times before preparation of lysates for measurement of
                                                                  phosphoproteins downstream of PI3K and mTOR
                                                                  (Fig. 5). Akt phosphorylation was completely wortman-
                                                                  nin sensitive and was not blocked by rapamycin, as
                                                                  expected. We detected a similar pattern when we probed
                                                                  for phosphorylated TSC2, an Akt substrate (data not
                                                                  shown). Rapamycin increased Akt and TSC2 phosphor-
                                                                  ylation in response to LPS but not anti-kappa antibody
                                                                  (Fig. 5, and data not shown), consistent with previous
                                                                  reports that in some situations mTOR signaling causes
                                                                  feedback inhibition of PI3K/Akt activation [48]. Analysis
Figure 4. Evidence for PI3K-independent mTOR activation by        of S6 phosphorylation revealed a basal level of pS6,
LPS. Splenocytes stained as in Fig. 1 were incubated for 1 h in
                                                                  consistent with the intracellular staining results. Both
standard medium, drug-treated with wortmannin (WM) or
rapamycin (Rap) for 15 min or left untreated, and then left       anti-kappa antibody and LPS induced further increases in
unstimulated (UnTx) or stimulated with LPS for 1 h. Fixed and     pS6 that were strongly blocked by rapamycin. Wort-
permeabilized cells were then stained intracellulary for pAkt     mannin blocked pS6 to a lesser extent than did rapamycin
(S473) or pS6 (S235/236) where indicated. (A) Mean fold           under both stimulation conditions, but wortmannin
increase of pAkt with respect to UnTx treated with wortman-
                                                                  resistance was consistently greater in the LPS-treated
nin. n = 5. (B) Mean fold increase of pS6 with respect to UnTx
treated with rapamycin. For the statistical comparison of
                                                                  samples. We also examined the phosphorylation of
unstimulated/wortmannin-treated vs. unstimulated/rapamy-          4E-BP1, a direct mTOR substrate, by distinguishing
cin-treated, we used the mean fluorescence intensity data         phosphorylated forms of 4E-BP1 based on altered
rather than comparing fold changes for the reason stated in       migration [49]. The results were consistent with pS6
Fig. 2. For the statistical comparison of LPS/wortmannin-         analyses in showing (1) an increase in basal phosphor-
treated vs. LPS/rapamycin-treated, we used fold changes. n = 3;
*p<0.05. (C) Percent inhibition of pS6 by wortmannin in cells
                                                                  ylation (slower migrating species c) in cells cultured for
that were either left unstimulated or activated with anti-kappa   1 h compared to freshly isolated samples; (2) greater
antibody or LPS; **p<0.01. (D) Comparison of pS6 in WT and        dephosphorylation (species a, b) of 4E-BP1 in cells
p85aKO (KO) B cell subsets. Representative of three indepen-      treated with rapamycin than with wortmannin, espe-
dent experiments.                                                 cially in response to LPS (Fig. 5). These data suggest that
                                                                  the PI3K dependence of mTOR activation depends on the
                                                                  stimulus, with B cells exposed to LPS showing robust
                                                                  mTOR signaling even when PI3K activity is blocked.

f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                          www.eji-journal.eu
2930   Amber C. Donahue and David A. Fruman                                                          Eur. J. Immunol. 2007. 37: 2923–2936

       LPS-mediated proliferation is rapamycin sensitive                    exhibited more extensive apoptosis (data not shown).
       but partially wortmannin resistant                                   When fractions were stimulated with LPS, we observed
                                                                            that many of the MZE cells had already entered the cell
                                                                            cycle by 28 h, and MZE cells maintained a much greater
       A hallmark of the MZ subset is the rapid and robust
                                                                            percentage of cells in cycle than FOE cells at all time
       response of these cells to LPS. For example, FACS-sorted
                                                                            points assayed (Fig. 6B). As a control for the
       MZ cells enter the cell cycle more rapidly than FO cells
                                                                            contamination of the FOE population with MZ cells,
       [50]. The increased responsiveness appears to be due in
                                                                            we also assayed B cells purified from lymph nodes (LN),
       part to higher expression of the LPS coreceptor RP105
                                                                            which consisted entirely of cells with an FO phenotype
       on MZ cells relative to FO cells [51]. In order to examine
                                                                            (Fig. 6A). These cells were able to enter the cell cycle
       the PI3K and mTOR requirements in LPS-driven
                                                                            following exposure to LPS, although the fraction of
       proliferation of B cell subsets, we developed a novel
                                                                            responding LN B cells was lower than for FOE cells
       magnetic separation protocol to enrich MZ and FO cells
                                                                            purified from spleen (data not shown). We conclude
       rapidly, without the need for multiple surface stains and
                                                                            from these data that magnetic enrichment of FO and MZ
       FACS sorting. The two-step process first removed non-B
                                                                            fractions generates populations that exhibit the general
       and immature B cells (yielding 97–99% pure mature
                                                                            pattern of mitogen responsiveness typical of these
       B cells), then separated the remaining MZ and FO cells,
                                                                            subsets, even though the fractions are not absolutely
       generating MZ-enriched (MZE) and FO-enriched (FOE)
       populations (Fig. 6A). Enrichment ranged from 65 to
                                                                                Both wortmannin and rapamycin effectively blocked
       75% for MZE, and 85 to 95% for FOE (Fig. 6A, and data
                                                                            proliferation of FOE and MZE cells stimulated with anti-
       not shown). This method is faster and gives higher cell
                                                                            IgM antibody (data not shown). For LPS-treated
       yields than cell sorting, and has the advantage of
                                                                            samples, pretreatment with wortmannin caused only a
       generating enriched MZ cells labeled with only one
                                                                            modest reduction in cell cycle entry in both populations
       antibody, while leaving the FO cells unlabeled. The
                                                                            at each time point, whereas rapamycin treatment was
       enrichment of the MZE fraction was validated function-
                                                                            much more effective at blocking cell cycle progression,
       ally in proliferation assays using anti-IgM antibody,
                                                                            causing up to 80% inhibition in FOE cells and 65% in
       which induces apoptosis rather than cell cycle progress-
                                                                            MZE cells (Fig. 6C). Thus, exposure of B cells to LPS
       ion in MZ cells [50, 52]. Consistent with effective
                                                                            favors the proliferation of MZ cells as reported [50, 52],
       separation of subsets, greater cell cycle entry was
                                                                            and the response of both subsets is more dependent on
       observed in the FOE fraction, and the MZE fraction
                                                                            mTOR function than on PI3K activity.

                                                                            Wortmannin and rapamycin sensitivity of LPS-
                                                                            mediated MZ cell differentiation

                                                                            LPS-mediated differentiation of MZ cells into antibody-
                                                                            secreting plasma cells has also been shown to proceed
                                                                            more rapidly than that of FO cells [50]. We enriched MZ
                                                                            and FO cells as described in Fig. 6 and measured the
                                                                            percentage of antibody-secreting cells (ASC) as defined
                                                                            by the phenotype B220int/Syndecan-1+ (Fig. 7A; [53]).
                                                                            By 48 h, 20% of the live cells in the MZE population
                                                                            already showed the ASC phenotype, and at each time
                                                                            point measured, MZE cells showed a much greater
                                                                            percentage of ASC than the FOE population (Fig. 7B). As
                                                                            with proliferation, differentiation into the ASC pheno-
       Figure 5. Purified mature B cells (both FO and MZ; see Fig. 6A for   type in MZE cells was reduced to a greater extent by
       purification strategy) were rested and drug-treated as de-
                                                                            rapamycin than by wortmannin (Fig. 7B). FOE popula-
       scribed in Figure 4, and stimulated with anti-kappa antibody or
       LPS for the time periods indicated, then lysed and immuno-           tions stimulated with LPS generated very few cells with
       blotted for pAkt and pS6 using the same antibodies used for          the ASC phenotype, and these cultures exhibited much
       intracellular staining. Phosphorylation levels of the mTOR           more cell death (Fig. 7B). It is likely that the few ASC
       substrate 4E-BP1 were also visualized using an antibody that         generated from FOE cultures were derived from
       detects the most highly phosphorylated c species, as well as
                                                                            contaminating MZ cells, as FO cells purified from LN
       the lower levels of phosphorylation represented by the a and
       b species. The b-actin immunoblot serves as a loading control.       showed no appreciable ASC differentiation (Fig. 7B).
       This experiment was performed twice with comparable                     The rapid differentiation of MZE cells into ASC
       results.                                                             suggested that these cells might also begin secreting

       f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                             www.eji-journal.eu
Eur. J. Immunol. 2007. 37: 2923–2936                                                                             Leukocyte signaling     2931

                                                                      Figure 7. Wortmannin and rapamycin sensitivity of LPS-
                                                                      mediated MZ differentiation. FOE, MZE, and LN FO populations
                                                                      were obtained, treated with wortmannin (WM) or rapamycin
                                                                      (Rap) and stimulated with LPS as described in Fig. 6. Cells were
                                                                      then harvested at the indicated time points and stained with
                                                                      anti-Syndecan-1 (Synd-1) and anti-B220 antibodies to identify
                                                                      ASC. B220int/Synd-1+ ASC were gated as in the representative
                                                                      dot plots shown in (A), and the percentages of ASC quantitated
                                                                      in (B). n = 3. The large amount of cell death in FOE and LN FO
                                                                      samples led to low numbers of live cell events and to
                                                                      percentages of ASC that were potentially misleading; the
                                                                      ASC percentages were therefore corrected to reflect live cells.
                                                                      (C) Supernatants were collected from LPS-stimulated MZE or
                                                                      FOE cells at the indicated time points, and the concentrations
                                                                      of secreted IgM were determined by ELISA. n = 3.
Figure 6. LPS-mediated proliferation of FO and MZ cells is
rapamycin sensitive but partially wortmannin resistant. FOE
and MZE populations were obtained using the magnetic                  soluble antibody. We collected supernatants from MZE
separation technique diagrammed in (A). Splenocytes were              cells stimulated with LPS as described and determined
labeled with anti-AA4.1-biotin antibody (AA4.1-B), then in-           the levels of secreted IgM by ELISA, in order to
cubated with anti-CD43 and anti-biotin beads. The cells were
                                                                      determine the sensitivity of this process to wortmannin
then negatively selected to remove immature B cells and non-
B cells, respectively, leaving behind FO and MZ cells (Pure           and rapamycin. Previous studies of human B cells
Mature). Purified mature cells were then incubated with anti-         concluded that these compounds inhibit differentiation
CD9-biotin antibody (anti-CD9-B) to label MZ cells, followed by       but do not inhibit antibody secretion directly [21, 54]. As
anti-biotin beads, and positively selected to yield the MZE           expected, MZE cells had already begun to secrete IgM by
fraction. Unlabeled FO cells passed through the column to yield
                                                                      48 h and production increased over time (Fig. 7C). FOE
the FOE fraction. Highly pure FO cells were generated by
purifying B cells from the LN via negative selection with anti-       cells showed a weak and delayed IgM secretion response
CD43 beads (LN FO). (B) FOE and MZE cells were pretreated with        (Fig. 7C). Wortmannin caused a partial block in the
wortmannin (WM) or rapamycin (Rap) where indicated,                   secretion of soluble IgM by MZE cells at all time points
stimulated with LPS, harvested at the times indicated, and            (Fig. 7C). Rapamycin treatment blocked all IgM
fixed, permeabilized and stained with propidium iodide for cell
                                                                      secretion at all time points (Fig. 7C). Thus, the rapid
cycle analysis. The percent of cells actively cycling (i.e. in S or
G2/M phases) is shown. (C) The percent inhibition of cell cycle
                                                                      production of IgM-secreting cells induced by LPS
entry by wortmannin and rapamycin at 34 h. **p <0.01, n = at          correlates more closely with mTOR function than with
least 3.                                                              PI3K activity.

f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                                 www.eji-journal.eu
2932   Amber C. Donahue and David A. Fruman                                                   Eur. J. Immunol. 2007. 37: 2923–2936

       Discussion                                                  acids but not serum. This suggests that the signaling
                                                                   pathways leading from serum components to mTOR are
       In this paper, we have used novel approaches to assess      PI3K dependent in B cells, whereas the pathways leading
       signaling and functional responses in FO and MZ cells,      from nutrient sensing are partially PI3K independent.
       two subsets of mature B cells that have distinct roles in   Serum contains many peptide growth factors and
       immunity. We used selective inhibitors of PI3K or mTOR      survival factors (i.e. insulin, IGF-1, PDGF) that are
       to examine the regulation and function of these             known to activate mTOR via PI3K/Akt in various cell
       signaling enzymes in FO and MZ cells under different        types [23]. One or more of these serum components
       stimulation conditions. We find both similarities and       might be responsible for basal mTOR signaling in
       differences among the subsets. Similarities include         mature B cells. Activation of mTOR by nutrients has
       comparable sensitivity to the PI3K inhibitor wortmannin     been suggested to proceed through the wortmannin-
       with respect to the degree of phosphorylation of Akt and    sensitive class III PI3K in other cell types [56, 57], but
       Erk following BCR cross-linking. The LPS-induced            might be regulated differently in mature B cells.
       increase in S6 phosphorylation, a critical signaling step       Our results demonstrate that B cell mitogens can also
       downstream of mTOR activation, is relatively resistant      induce PI3K-independent mTOR activation. In the case
       to wortmannin in both subsets. This finding led us to       of BCR cross-linking, wortmannin inhibits S6 phosphor-
       examine LPS-driven mTOR activation in more detail. We       ylation, but to a lesser extent in MZ cells than in FO cells.
       find that a direct mTOR substrate, 4E-BP1, is also          In the case of LPS treatment, wortmannin has little effect
       phosphorylated following LPS treatment in a manner          on S6 phosphorylation in either subset, nor on 4E-BP1
       mostly independent of PI3K. Supporting the functional       phosphorylation in total mature B cells. The nature of
       significance of PI3K-independent mTOR activation, we        the PI3K-independent pathways from mitogens to
       observe that LPS-triggered proliferation and differentia-   mTOR warrants further investigation. One mechanism
       tion are more strongly inhibited by rapamycin than by       might be the generation of phosphatidic acid (PA), a
       wortmannin.                                                 potent activator of mTOR [58–61]. There are several
           One difference we observed between FO and MZ cells      possible routes to PA production in cells [59, 62, 63],
       was that MZ cells exhibit far more robust S6 phosphor-      and importantly, LPS is known to promote PA produc-
       ylation in response to nutrients in the absence of          tion that is functionally important for LPS responses
       mitogens than do FO cells. It is intriguing that MZ cells   [62]. One route to PA production is through the enzyme
       maintain highly elevated mTOR signaling in the absence      lysophosphatidic acid acyltransferase (LPAAT) [62], but
       of B cell mitogens like anti-Ig antibodies or LPS.          we found that the LPAAT inhibitor lisofylline had no
       Activation of mTOR is of central importance for various     effect on LPS-mediated mTOR activation in B cells (not
       aspects of cell cycle entry including increased transla-    shown).
       tion and cell growth (size increase). MZ cells are              Several Toll-like receptors (TLR) or TLR ligands have
       significantly larger than FO cells ([55] and Fig. 1A, 2C)   been reported to activate PI3K, including LPS. However,
       and enter the cell cycle more rapidly following LPS         in macrophages and dendritic cells, PI3K appears to play
       treatment, phenomena that could be promoted by              a negative regulatory role in immune responses to TLR
       elevations in basal mTOR signaling. Most MZ cells did       ligands [64, 65]. LPS treatment of B cells induces PI3K-
       not display elevated S6 phosphorylation immediately         dependent phosphorylation of Akt and NF-jB activation
       after isolation, although a fraction did show basal pS6     ([46]; Fig. 5), but the functional significance of PI3K/
       (Fig. 3). However, it is possible that cellular phospha-    Akt activation by LPS is unclear. A role for PI3K in LPS-
       tases act during the process of organ harvest and cell      driven B cell proliferation has been inferred mainly from
       isolation, conditions of hypoxia that favor phosphatase     two types of observations. First, various mouse strains
       activity. The proximity of MZ cells to the circulation      lacking PI3K gene function have reduced B cell
       might expose them to higher concentrations of plasma        responses to LPS [11, 16, 17, 19]. However, these strains
       components and nutrients that maintain elevated mTOR        are nearly devoid of MZ cells, the subset responsible for
       signaling, priming the cells for more rapid cell growth     the majority of LPS responsiveness. Second, the PI3K
       and differentiation into ASC.                               inhibitor LY294002 reduces LPS-driven proliferation of
           Mitogen-independent S6 phosphorylation in B cells       splenic B cells by 80–90% [10, 11, 66]. However,
       is sustained by components of the serum as well as          LY294002 is a direct inhibitor of mTOR at the
       leucine and glucose. Removal of any of these compo-         concentrations used, complicating the interpretation
       nents reduces the magnitude of S6 phosphorylation in        of this result. Here, we used wortmannin to inhibit PI3K
       both subsets. The overall response is partially indepen-    at concentrations that do not strongly inhibit mTOR [67,
       dent of PI3K activity, as judged by lesser inhibition by    68], and we examined the LPS responses of isolated
       wortmannin than by rapamycin. This wortmannin-              subsets. The results show that LPS-driven proliferation
       resistant pS6 appears to require both glucose and amino     and differentiation are only partially PI3K dependent,

       f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                      www.eji-journal.eu
Eur. J. Immunol. 2007. 37: 2923–2936                                                                          Leukocyte signaling    2933

yet strongly dependent on mTOR in both FO and MZ                    magnetic separation using CD43 beads (Miltenyi Biotec,
cells. These data highlight the functional significance of          Auburn, CA) as described [11]. Purity was determined to be
PI3K-independent mTOR activation by LPS. Further-                   >97% by FACS analysis (FACSCalibur and CellQuest software,
more, the results emphasize an important difference                 BD Biosciences, Mountain View, CA; FlowJo software,
                                                                    TreeStar, San Carlos, CA). FO and MZ cells were enriched in
between the signaling mechanisms used to activate
                                                                    two steps, first generating pure mature B cells by depleting
B cells by LPS versus BCR ligands. One of the critical
                                                                    immature B cells with an anti-AA4.1-biotin antibody and anti-
PI3K-dependent signaling events downstream of the                   biotin beads (Miltenyi Biotec), in combination with non-B cell
BCR is assembly of a "signalosome" leading to increased             depletion via CD43 beads (Miltenyi Biotech). MZ and FO cells
cellular levels of diacylglycerol and Ca2+ mobilization             were then separated via positive selection of MZ cells using
[9, 69], responses that do not occur acutely following              anti-CD9-biotin antibody. Purity was determined by FACS
LPS exposure. Blockade of mitogenic signals down-                   analysis to be !94% and !65% of B cells for FOE and MZE
stream of the signalosome might explain the more                    cells, respectively, using anti-CD24-PE and anti-CD1d-biotin
potent effects of wortmannin in BCR-ligated cells.                  antibodies and streptavidin-APC. The yield of enriched MZ
    One of the challenges in comparing B cell subsets has           cells ranged from 3–6% of total splenocytes. For all assays,
been the difficulty in isolating sufficient amounts of cells        unless otherwise indicated, cells were rested or stimulated in
with adequate purity to perform biochemical and                     our standard medium, consisting of RPMI 1640 supplemented
                                                                    with 10% heat-inactivated FBS, 5 mM HEPES, 2 mM L-gluta-
functional assays. FACS-based phosphoprotein measure-
                                                                    mine, 100 U/mL penicillin, 100 lg/mL streptomycin, and
ment is a powerful tool to distinguish signaling
                                                                    50 lM 2-ME. dFBS (HyClone, Logan, UT), RPMI lacking
responses in heterogeneous populations [36–41] and                  leucine (MP Biomedicals, Solon, OH), and RPMI lacking
can be used to compare signaling in B cell subsets as               glucose (Gibco/Invitrogen, Carlsbad, CA) replaced standard
presented here. The magnetic enrichment of MZ cells we              medium components as described in the text.
have described provides a rapid alternative to FACS
sorting and yields subsets that have been either                    Inhibitors
untouched by antibodies (FO) or bound only by anti-
CD9 antibody (MZ). Highly pure FO cells can be                      The PI3K inhibitor wortmannin (100 nM), the mTOR inhibitor
obtained alternatively from LN by simple depletion with             rapamycin (10 ng/mL), and the MEK inhibitor U0126 (10 lM)
                                                                    were purchased from Calbiochem (EMD, San Diego, CA).
anti-CD43 antibody. The magnetically enriched subsets
show little contamination with immature cells and
                                                                    FACS antibodies
exhibit the expected activation properties, with anti-IgM
antibody favoring proliferation of FO cells and LPS                 Anti-Syndecan-1-biotin, anti-CD9-biotin, anti-B220-FITC, and
favoring proliferation of MZ cells. Of note, LPS                    unconjugated anti-CD21 antibodies were obtained from BD
stimulation for 3 days generated ASC from MZ cells                  Biosciences. Anti-AA4.1-biotin, anti-CD1d-biotin, anti-CD23-
but not FO cells. This could explain why LPS treatment              biotin, anti-CD24-PE antibodies, and streptavidin-PE were
of total splenic B cells for 3 days produces only a small           obtained from eBiosciences (San Diego, CA). Streptavidin-APC
percentage of ASC [53]. Considering the vastly different            was obtained from Biolegend (San Diego, CA). Anti-CD21
roles of FO and MZ cells during humoral immune                      antibody was conjugated to Alexa647 using a kit from
responses in vivo, we suggest that attempts to model                Molecular Probes (Invitrogen).
B cell differentiation in vitro should be based on isolated
                                                                    Phosflow assays for pAkt and pErk
subsets of mature B cells.
                                                                    Splenocytes were stained in HBSS with anti-CD24 and anti-
                                                                    CD1d/SA-APC antibodies as described, washed with standard
Materials and methods                                               medium, and rested for 1 h in standard medium at 37 C and
                                                                    5% CO2 to reduce basal levels of phosphoprotein. Cells were
Mice                                                                then treated for 15 min with wortmannin or U0126 prior to
                                                                    stimulation with anti-kappa antibody (10 lg/mL) for 1 min or
BALB/c mice were purchased from the Jackson Laboratory              LPS (serotype 0127:B8, 10 lg/mL; Sigma-Aldrich, St. Louis,
(Bar Harbor, ME) and bred in our colony. p85aKO (Pik3r1tm1)         MO) for 1 h. Samples were washed with and resuspended in
mice (BALB/c background) were purchased from Taconic.               1Â PBS, and fixed with MeOH-free formaldehyde (2% final;
Animal use was approved by the institutional animal care and        Polysciences Inc., Warrington, PA) for 10 min in a 37 C water
use committee.                                                      bath. After pelleting, cells were permeabilized with ice-cold
                                                                    90% MeOH for 30 min on ice, washed with and resuspended in
Cells                                                               1Â PBS/0.5% BSA and incubated at room temperature for
                                                                    10 min. Primary anti-pAkt (S473) and anti-pErk (T202/Y204)
Single-cell suspensions were obtained from the spleens or LN        rabbit mAb (Cell Signaling Techonology, Danvers, MA) were
of 6–10-wk-old mice and depleted of RBC by hypotonic lysis.         incubated with samples for 30 min at room temperature. Cells
Purified total splenic B cells or FO LN B cells were obtained via   were washed with 1Â PBS/0.5% BSA, incubated with goat

f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                              www.eji-journal.eu
2934   Amber C. Donahue and David A. Fruman                                                                 Eur. J. Immunol. 2007. 37: 2923–2936

       anti-rabbit IgG-Alexa 488 secondary antibody (Molecular           ELISA
       Probes, Invitrogen) for 30 min at room temperature, washed
       again with 1Â PBS/0.5% BSA and resuspended in PBS for             FOE and MZE cells were obtained, drug-treated and stimulated
       FACS analysis.                                                    as described above for the differentiation assays. Supernatants
                                                                         were collected at the times indicated and stored frozen. Nunc-
       Phosflow assay for pS6                                            Immuno MaxiSorp 96-well plates (NalgeNunc International,
                                                                         Rochester, NY) were coated with anti-IgM capture antibody
       Cells were stained for CD24-PE and CD1d/SA-APC or CD24-
                                                                         (3 lg/mL; BD Biosciences) in 50 mM Na2CO3 overnight at
       PE, CD23/SA-CyC and CD21-Alexa647 as described, washed
                                                                         4 C. Plates were blocked with 1Â PBS/3% BSA for 2 h at room
       and resuspended with standard medium or medium lacking
                                                                         temperature, then washed three times with 1Â PBS/1% BSA
       the indicated nutrients, and incubated for 1 h at 37 C and 5%
                                                                         and three times with 1Â PBS. IgM protein standards (South-
       CO2. Prior to stimulation with either 10 lg/mL anti-kappa
                                                                         ernBiotech) and samples were added to the plate and
       antibody (15 min) or LPS (1 h), cells were treated for 15 min
                                                                         incubated at room temperature for 2 h, and washed as
       with wortmannin or rapamycin. Samples were then fixed and
                                                                         described above. The secondary anti-IgM-biotin antibody
       permeabilized per the CytoFix/CytoPerm kit (BD Biosciences)
                                                                         (0.5 lg/mL; BD Biosciences) was incubated for 2 h at room
       as described [15]. Primary and secondary antibodies used
                                                                         temperature, then washed as described. Streptavidin-AP
       were pS6 (S235/S36) rabbit polyclonal antibody (Cell
                                                                         (4.65 lg/mL; Pierce) was incubated in the plates for 1 h at
       Signaling Technology) and goat anti-rabbit IgG-FITC (Sigma).
                                                                         room temperature, and washed as described. The samples
       Samples were resuspended in HBSS prior to FACS analysis.
                                                                         were then revealed with a Phosphatase Substrate Kit (Pierce),
                                                                         the reaction was stopped with 2 N NaOH, and the plates were
       Statistical analysis
                                                                         read at 405 nm.
       Where p values are given, data were analyzed using a paired,
       two-tailed Student's t-test. Error bars represent SEM.
                                                                         Acknowledgements: We thank Tom Rothstein, Craig
                                                                         Walsh and Aimee Edinger for helpful comments,
       Purified mature B cells (2 Â 106) were obtained from single-      Xiaocui Sun for technical assistance, and Christine
       cell spleen suspensions as described above, rested for 1 h in     McLaren for statistical consultation. This work was
       standard medium at 37 C and 5% CO2, pretreated with              supported by National Institutes of Health grants AI-
       wortmannin or rapamycin and stimulated with anti-kappa            50831 and AI-57900 (to D.A.F.) and National Institutes of
       antibody or LPS (10 lg/mL). Cells were harvested at the           Health Immunology Training Grant 1 T32 AI 60573 (to
                                                                         A.C.D.). A.C.D. was an Achievement Rewards for College
       indicated times, and lysed and subjected to SDS-PAGE as
                                                                         Scientists (ARCS) Scholar.
       described [10]. Primary antibodies were those used for
       phosflow detection of pAkt (S473) and pS6 (S235/236)
       described above, and blots were also probed for total 4E-BP1
       (Cell Signaling) and b-actin (Sigma). Bound antibodies were
       visualized by incubation with goat anti-rabbit IgG-HRP            References
       (Promega, Madison, WI), followed by ECL detection (Pierce,        1 Lopes-Carvalho, T. and Kearney, J. F., Development and selection of
                                                                           marginal zone B cells. Immunol. Rev. 2004. 197: 192–205.
       Rockford, IL).
                                                                          2 Srivastava, B., Lindsley, R. C., Nikbakht, N. and Allman, D., Models for
                                                                            peripheral B cell development and homeostasis. Semin. Immunol. 2005. 17:
       Proliferation and differentiation assays

       FOE and MZE cells were obtained via magnetic separation as         3 Lopes-Carvalho, T., Foote, J. and Kearney, J. F., Marginal zone B cells in
                                                                            lymphocyte activation and regulation. Curr. Opin. Immunol. 2005. 17:
       described above. Enriched samples were pretreated with               244–250.
       rapamycin for 15 min, while wortmannin samples were both
                                                                          4 Benschop, R. J., Brandl, E., Chan, A. C. and Cambier, J. C., Unique
       pretreated and post-treated every 9–12 h to maintain efficacy        signaling properties of B cell antigen receptor in mature and immature
       of the drug. Cells were then stimulated with LPS (10 lg/mL)          B cells: Implications for tolerance and activation. J. Immunol. 2001. 167:
       and harvested at the indicated times. Proliferation was assayed
       by entry into the S and G2/M phases of the cell cycle, as          5 Benschop, R. J., Melamed, D., Nemazee, D. and Cambier, J. C., Distinct
                                                                            signal thresholds for the unique antigen receptor-linked gene expression
       determined by propidium iodide staining as described [10].
                                                                            programs in mature and immature B cells. J. Exp. Med. 1999. 190: 749–756.
       The data were analyzed by hand using FlowJo software. The
                                                                          6 Li, X., Martin, F., Oliver, A. M., Kearney, J. F. and Carter, R. H., Antigen
       differentiation state was determined by surface expression           receptor proximal signaling in splenic B-2 cell subsets. J. Immunol. 2001.
       levels of B220 and Syndecan-1. Harvested cells were washed           166: 3122–3129.
       with and stained in HBSS with anti-B220-FITC and anti-            7 Wang, Y., Zhang, P., Li, W., Hou, L., Wang, J., Liang, Y. and Han, H.,
       Syndecan-1-biotin antibodies (BD Biosciences) and streptavi-        Mouse follicular and marginal zone B cells show differential expression of
       din-PE (eBiosciences) and analyzed by FACS.                         Dnmt3a and sensitivity to 50 -azacytidine. Immunol. Lett. 2006. 105:

                                                                          8 Zhang, P., Li, W., Wang, Y., Hou, L., Xing, Y., Qin, H., Wang, J. et al.,
                                                                            Identification of CD36 as a new surface marker of marginal zone B cells by
                                                                            transcriptomic analysis. Mol. Immunol. 2007. 44: 332–337.

       f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                                       www.eji-journal.eu
Eur. J. Immunol. 2007. 37: 2923–2936                                                                                                        Leukocyte signaling          2935

 9 Donahue, A. C. and Fruman, D. A., PI3K signaling controls cell fate at many       28 Hebeis, B. J., Vigorito, E. and Turner, M., The p110delta subunit of
   points in B lymphocyte development and activation. Semin. Cell Dev. Biol.            phosphoinositide 3-kinase is required for the lipopolysaccharide response of
   2004. 15: 183–197.                                                                   mouse B cells. Biochem. Soc. Trans. 2004. 32: 789–791.

10 Donahue, A. C. and Fruman, D. A., Proliferation and survival of activated         29 Lyubchenko, T., dal Porto, J., Cambier, J. C. and Holers, V. M., Coligation
   B cells requires sustained antigen receptor engagement and phosphoinosi-             of the B cell receptor with complement receptor type 2 (CR2/CD21) using its
   tide 3-kinase activation. J. Immunol. 2003. 170: 5851–5860.                          natural ligand C3dg: Activation without engagement of an inhibitory
                                                                                        signaling pathway. J. Immunol. 2005. 174: 3264–3272.
11 Fruman, D. A., Snapper, S. B., Yballe, C. M., Davidson, L., Yu, J. Y., Alt, F.
   W. and Cantley, L. C., Impaired B cell development and proliferation in           30 Allman, D., Lindsley, R. C., DeMuth, W., Rudd, K., Shinton, S. A. and
   absence of phosphoinositide 3-kinase p85alpha. Science 1999. 283:                    Hardy, R. R., Resolution of three nonproliferative immature splenic B cell
   393–397.                                                                             subsets reveals multiple selection points during peripheral B cell maturation.
                                                                                        J. Immunol. 2001. 167: 6834–6840.
12 Bilancio, A., Okkenhaug, K., Camps, M., Emery, J. L., Ruckle, T.,
   Rommel, C. and Vanhaesebroeck, B., Key role of the p110delta isoform of           31 Loder, F., Mutschler, B., Ray, R. J., Paige, C. J., Sideras, P., Torres, R.,
   PI3K in B-cell antigen and IL-4 receptor signaling: Comparative analysis of          Lamers, M. C. and Carsetti, R., B cell development in the spleen takes place
   genetic and pharmacologic interference with p110delta function in B cells.           in discrete steps and is determined by the quality of B cell receptor-derived
   Blood 2006. 107: 642–650.                                                            signals. J. Exp. Med. 1999. 190: 75–89.

13 Clayton, E., Bardi, G., Bell, S. E., Chantry, D., Downes, C. P., Gray, A.,        32 Srivastava, B., Quinn, W. J., 3rd, Hazard, K., Erikson, J. and Allman, D.,
   Humphries, L. A. et al., A crucial role for the p110delta subunit of                 Characterization of marginal zone B cell precursors. J. Exp. Med. 2005. 202:
   phosphatidylinositol 3-kinase in B cell development and activation. J. Exp.          1225–1234.
   Med. 2002. 196: 753–763.
                                                                                     33 Su, T. T. and Rawlings, D. J., Transitional B lymphocyte subsets operate as
14 Donahue, A. C., Hess, K. L., Ng, K. L. and Fruman, D. A., Altered splenic            distinct checkpoints in murine splenic B cell development. J. Immunol. 2002.
   B cell subset development in mice lacking phosphoinositide 3-kinase                  168: 2101–2110.
   p85alpha. Int. Immunol. 2004. 16: 1789–1798.
                                                                                     34 Aman, M. J., Lamkin, T. D., Okada, H., Kurosaki, T. and Ravichandran, K.
15 Hess, K. L., Donahue, A. C., Ng, K. L., Moore, T. I., Oak, J. and Fruman, D.         S., The inositol phosphatase SHIP inhibits Akt/PKB activation in B cells. J.
   A., Frontline: The p85alpha isoform of phosphoinositide 3-kinase is essential        Biol. Chem. 1998. 273: 33922–33928.
   for a subset of B cell receptor-initiated signaling responses. Eur. J. Immunol.
   2004. 34: 2968–2976.                                                              35 Jacob, A., Cooney, D., Pradhan, M. and Coggeshall, K. M., Convergence of
                                                                                        signaling pathways on the activation of ERK in B cells. J. Biol. Chem. 2002.
16 Jou, S. T., Carpino, N., Takahashi, Y., Piekorz, R., Chao, J. R., Carpino,           277: 23420–23426.
   N., Wang, D. and Ihle, J. N., Essential, nonredundant role for the
   phosphoinositide 3-kinase p110delta in signaling by the B-cell receptor           36 Krutzik, P. O., Clutter, M. R. and Nolan, G. P., Coordinate analysis of
   complex. Mol. Cell. Biol. 2002. 22: 8580–8591.                                       murine immune cell surface markers and intracellular phosphoproteins by
                                                                                        flow cytometry. J. Immunol. 2005. 175: 2357–2365.
17 Okkenhaug, K., Bilancio, A., Farjot, G., Priddle, H., Sancho, S., Peskett,
   E., Pearce, W. et al., Impaired B and T cell antigen receptor signaling in        37 Krutzik, P. O., Hale, M. B. and Nolan, G. P., Characterization of the murine
   p110delta PI 3-kinase mutant mice. Science 2002. 297: 1031–1034.                     immunological signaling network with phosphospecific flow cytometry. J.
                                                                                        Immunol. 2005. 175: 2366–2373.
18 Suzuki, H., Matsuda, S., Terauchi, Y., Fujiwara, M., Ohteki, T., Asano, T.,
   Behrens, T. W. et al., PI3K and Btk differentially regulate B cell antigen        38 Perez, O. D. and Nolan, G. P., Simultaneous measurement of multiple active
   receptor-mediated signal transduction. Nat. Immunol. 2003. 4: 280–286.               kinase states using polychromatic flow cytometry. Nat. Biotechnol. 2002. 20:
19 Suzuki, H., Terauchi, Y., Fujiwara, M., Aizawa, S., Yazaki, Y., Kadowaki,
   T. and Koyasu, S., Xid-like immunodeficiency in mice with disruption of the       39 Perez, O. D. and Nolan, G. P., Phospho-proteomic immune analysis by flow
   p85alpha subunit of phosphoinositide 3-kinase. Science 1999. 283: 390–392.           cytometry: From mechanism to translational medicine at the single-cell
                                                                                        level. Immunol. Rev. 2006. 210: 208–228.
20 Piatelli, M. J., Wardle, C., Blois, J., Doughty, C., Schram, B. R., Rothstein,
   T. L. and Chiles, T. C., Phosphatidylinositol 3-kinase-dependent mitogen-         40 Irish, J. M., Czerwinski, D. K., Nolan, G. P. and Levy, R., Kinetics of B cell
   activated protein/extracellular signal-regulated kinase kinase 1/2 and NF-           receptor signaling in human B cell subsets mapped by phosphospecific flow
   kappa B signaling pathways are required for B cell antigen receptor-                 cytometry. J. Immunol. 2006. 177: 1581–1589.
   mediated cyclin D2 induction in mature B cells. J. Immunol. 2004. 172:            41 Donahue, A. C. and Fruman, D. A., Measuring phosphorylated Akt and
   2753–2762.                                                                           other phosphoinositide-3-kinase-regulated phosphoproteins in primary
21 Aagaard-Tillery, K. M. and Jelinek, D. F., Inhibition of human                       lymphocytes. Methods Enzymol. 2007. 434: 131–154.
   B lymphocyte cell cycle progression and differentiation by rapamycin. Cell.       42 Anzelon, A. N., Wu, H. and Rickert, R. C., Pten inactivation alters
   Immunol. 1994. 156: 493–507.                                                         peripheral B lymphocyte fate and reconstitutes CD19 function. Nat.
22 Wicker, L. S., Boltz, R. C., Jr., Matt, V., Nichols, E. A., Peterson, L. B. and      Immunol. 2003. 4: 287–294.
   Sigal, N. H., Suppression of B cell activation by cyclosporine A, FK506 and       43 Perez, O. D., Kinoshita, S., Hitoshi, Y., Payan, D. G., Kitamura, T., Nolan,
   rapamycin. Eur. J. Immunol. 1990. 20: 2277–2283.                                     G. P. and Lorens, J. B., Activation of the PKB/AKT pathway by ICAM-2.
23 Richardson, C. J., Schalm, S. S. and Blenis, J., PI3-kinase and TOR:                 Immunity 2002. 16: 51–65.
   PIKTORing cell growth. Semin. Cell Dev. Biol. 2004. 15: 147–159.                  44 Kam, Y. and Exton, J. H., Role of phospholipase D1 in the regulation of
24 Wullschleger, S., Loewith, R. and Hall, M. N., TOR signaling in growth and           mTOR activity by lysophosphatidic acid. FASEB J. 2004. 18: 311–319.
   metabolism. Cell 2006. 124: 471–484.                                              45 Dann, S. G. and Thomas, G., The amino acid sensitive TOR pathway from
25 Alessi, D. R., Kozlowski, M. T., Weng, Q. P., Morrice, N. and Avruch, J.,            yeast to mammals. FEBS Lett. 2006. 580: 2821–2829.
   3-Phosphoinositide-dependent protein kinase 1 (PDK1) phosphorylates and           46 Bone, H. and Williams, N. A., Antigen-receptor cross-linking and
   activates the p70 S6 kinase in vivo and in vitro. Curr. Biol. 1998. 8: 69–81.        lipopolysaccharide trigger distinct phosphoinositide 3-kinase-dependent
26 Manning, B. D., Tee, A. R., Logsdon, M. N., Blenis, J. and Cantley, L. C.,           pathways to NF-kappa B activation in primary B cells. Int. Immunol. 2001.
   Identification of the tuberous sclerosis complex-2 tumor suppressor gene             13: 807–816.
   product tuberin as a target of the phosphoinositide 3-kinase/akt pathway.         47 Doyle, S., Vaidya, S., O'Connell, R., Dadgostar, H., Dempsey, P., Wu, T.,
   Mol. Cell 2002. 10: 151–162.                                                         Rao, G. et al., IRF3 mediates a TLR3/TLR4-specific antiviral gene program.
27 Wlodarski, P., Kasprzycka, M., Liu, X., Marzec, M., Robertson, E. S.,                Immunity 2002. 17: 251–263.
   Slupianek, A. and Wasik, M. A., Activation of mammalian target of                 48 Manning, B. D., Balancing Akt with S6K: Implications for both metabolic
   rapamycin in transformed B lymphocytes is nutrient dependent but                     diseases and tumorigenesis. J. Cell Biol. 2004. 167: 399–403.
   independent of Akt, mitogen-activated protein kinase/extracellular signal-
   regulated kinase kinase, insulin growth factor-I, and serum. Cancer Res.          49 Prabhu, S., Saadat, D., Zhang, M., Halbur, L., Fruehauf, J. P. and Ong, S.
   2005. 65: 7800–7808.                                                                 T., A novel mechanism for Bcr-Abl action: Bcr-Abl-mediated induction of the

f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                                                            www.eji-journal.eu
2936   Amber C. Donahue and David A. Fruman                                                                                    Eur. J. Immunol. 2007. 37: 2923–2936

           eIF4F translation initiation complex and mRNA translation. Oncogene 2007.        60 Fang, Y., Vilella-Bach, M., Bachmann, R., Flanigan, A. and Chen, J.,
           26: 1188–1200.                                                                      Phosphatidic acid-mediated mitogenic activation of mTOR signaling. Science
                                                                                               2001. 294: 1942–1945.
       50 Oliver, A. M., Martin, F. and Kearney, J. F., IgMhighCD21high lymphocytes
          enriched in the splenic marginal zone generate effector cells more rapidly        61 Tang, W., Yuan, J., Chen, X., Gu, X., Luo, K., Li, J., Wan, B. et al.,
          than the bulk of follicular B cells. J. Immunol. 1999. 162: 7198–7207.               Identification of a novel human lysophosphatidic acid acyltransferase,
                                                                                               LPAAT-theta, which activates mTOR pathway. J. Biochem. Mol. Biol. 2006.
       51 Hebeis, B., Vigorito, E., Kovesdi, D. and Turner, M., Vav proteins are               39: 626–635.
          required for B-lymphocyte responses to LPS. Blood 2005. 106: 635–640.
                                                                                            62 Rice, G. C., Brown, P. A., Nelson, R. J., Bianco, J. A., Singer, J. W. and
       52 Oliver, A. M., Martin, F., Gartland, G. L., Carter, R. H. and Kearney, J. F.,        Bursten, S., Protection from endotoxic shock in mice by pharmacologic
          Marginal zone B cells exhibit unique activation, proliferative and                   inhibition of phosphatidic acid. Proc. Natl. Acad. Sci. USA 1994. 91:
          immunoglobulin secretory responses. Eur. J. Immunol. 1997. 27: 2366–2374.            3857–3861.

       53 Omori, S. A., Cato, M. H., Anzelon-Mills, A., Puri, K. D., Shapiro-Shelef,        63 Topham, M. K., Signaling roles of diacylglycerol kinases. J. Cell. Biochem.
          M., Calame, K. and Rickert, R. C., Regulation of class-switch recombination          2006. 97: 474–484.
          and plasma cell differentiation by phosphatidylinositol 3-kinase signaling.
                                                                                            64 Fukao, T. and Koyasu, S., PI3K and negative regulation of TLR signaling.
          Immunity 2006. 25: 545–557.
                                                                                               Trends Immunol. 2003. 24: 358–363.
       54 Aagaard-Tillery, K. M. and Jelinek, D. F., Phosphatidylinositol 3-kinase          65 Yu, Y., Nagai, S., Wu, H., Neish, A. S., Koyasu, S. and Gewirtz, A. T., TLR5-
          activation in normal human B lymphocytes. J. Immunol. 1996. 156:                     mediated phosphoinositide 3-kinase activation negatively regulates flagel-
          4543–4554.                                                                           lin-induced proinflammatory gene expression. J. Immunol. 2006. 176:
       55 Kumararatne, D. S., Bazin, H. and MacLennan, I. C., Marginal zones: The
          major B cell compartment of rat spleens. Eur. J. Immunol. 1981. 11: 858–864.      66 Venkataraman, C., Shankar, G., Sen, G. and Bondada, S., Bacterial
                                                                                               lipopolysaccharide induced B cell activation is mediated via a phosphati-
       56 Byfield, M. P., Murray, J. T. and Backer, J. M., hVps34 is a nutrient-               dylinositol 3-kinase dependent signaling pathway. Immunol. Lett. 1999. 69:
          regulated lipid kinase required for activation of p70 S6 kinase. J. Biol. Chem.      233–238.
          2005. 280: 33076–33082.
                                                                                            67 Brunn, G. J., Williams, J., Sabers, C., Wiederrecht, G., Lawrence, J. C., Jr.
       57 Nobukuni, T., Joaquin, M., Roccio, M., Dann, S. G., Kim, S. Y., Gulati, P.,          and Abraham, R. T., Direct inhibition of the signaling functions of the
          Byfield, M. P. et al., Amino acids mediate mTOR/raptor signaling through             mammalian target of rapamycin by the phosphoinositide 3-kinase inhibitors,
          activation of class 3 phosphatidylinositol 3OH-kinase. Proc. Natl. Acad. Sci.        wortmannin and LY294002. EMBO J. 1996. 15: 5256–5267.
          USA 2005. 102: 14238–14243.
                                                                                            68 Knight, Z. A., Gonzalez, B., Feldman, M. E., Zunder, E. R., Goldenberg, D.
       58 Avila-Flores, A., Santos, T., Rincon, E. and Merida, I., Modulation of the           D., Williams, O., Loewith, R. et al., A pharmacological map of the PI3-K
          mammalian target of rapamycin pathway by diacylglycerol kinase-produced              family defines a role for p110alpha in insulin signaling. Cell 2006. 125:
          phosphatidic acid. J. Biol. Chem. 2005. 280: 10091–10099.                            733–747.

       59 Chen, J. and Fang, Y., A novel pathway regulating the mammalian target of         69 Fruman, D. A., Satterthwaite, A. B. and Witte, O. N., Xid-like phenotypes:
          rapamycin (mTOR) signaling. Biochem. Pharmacol. 2002. 64: 1071–1077.                 A B cell signalosome takes shape. Immunity 2000. 13: 1–3.

       f 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                                                                                           www.eji-journal.eu