TACIdeficiencyimpairssustainedBlimp ...

From www.bloodjournal.org by guest on June 23, 2016. For personal use only.
IMMUNOBIOLOGY

TACI deficiency impairs sustained Blimp-1 expression in B cells decreasing
long-lived plasma cells in the bone marrow

Shoichiro Tsuji,1 Catarina Cortesa˜ o,2 Richard J. Bram,3 Jeffrey L. Platt,1 and Marilia Cascalho1

1Departments of Surgery, Microbiology, and Immunology, University of Michigan, Ann Arbor, MI; 2Laboratory of Epigenetics and Soma, Instituto Gulbenkian de
Cieˆ ncia, Oeiras, Portugal; and 3Departments of Oncology, Pediatric Hematology/Oncology, and Pediatrics, Mayo Clinic, Rochester, MN

Deficiencies in transmembrane activator both TACI-deficient and TACI-proficient sults showing that TACI induces and
and CAML interactor (TACI) result in com- mice. We show that TACI promotes sus- maintains Blimp-1 provide, for the first
mon variable immune deficiency, a syn- tained Blimp-1 expression by B cells re- time, a unified molecular and cellular
drome marked by recurrent infections sponding to antigen, which in turn limits mechanism explaining the primary fea-
with encapsulated microorganisms, im- B-cell clonal expansion and facilitates tures of common variable immune defi-
paired production of antibodies, and differentiation of long-lived antibody- ciency, exquisite vulnerability to infection
lymphoproliferation. How TACI promotes secreting cells. Short-term IgG secretion with encapsulated organisms, lymphopro-
antibody production and inhibits lym- occurs independently of TACI as DNA liferation, and hypogammaglobulinemia.
phoproliferation is not understood. To double-strand breaks associated with iso- (Blood. 2011;118(22):5832-5839)
answer this question, we studied the gen- type class switching induce Blimp-1 tran-
eration of immunity to protein antigens in siently, independently of TACI. Our re-

Introduction

In humans, mutations in the TNFRSF13B gene that encodes liferative glomerulonephritis. To reconcile opposite manifestations
transmembrane activator and CAML interactor (TACI) is thought of TACI deficiency, lymphoproliferation, autoimmunity, and hypo-
to be the cause of 7% to 21% of cases of common variable immune gammaglobulinemia, we6 proposed that TACI has dual functions,
deficiency (CVID), the primary immunodeficiency most com- promoting B-cell differentiation on the one hand, and on the other
monly encountered in clinical practice. CVID is a heterogeneous limiting proliferation and thus decreasing the risk of autoimmunity
syndrome caused by antibody deficiency manifested in late child- and cancer. In this manuscript, we identify the molecular mecha-
hood or early to mid-adulthood. CVID patients often present with nisms responsible.
recurrent infections of the respiratory tract resulting from encapsu-
lated organisms, such as Streptococcus pneumoniae and Haemophi- Whether or not TACI promotes antibody secretion by enhancing
lus influenzae or Mycoplasma.1 Paradoxically, some CVID patients plasma cell differentiation has been the subject of controversy. We6
develop autoimmune diseases, the most common of which are found that lack of TACI on B cells impaired antibody production in
hemolytic anemia and autoimmune thrombocytopenic purpura. response to polysaccharides or lipopolysaccharide (LPS) resulting
CVID patients are at a high risk to develop neoplasms, primarily from a defect in differentiation of antibody-secreting cells (ASCs).
lymphomas.2 How TACI deficiency results in antibody deficiencies Ozcan et al7 showed that “a proliferation inducing ligand” (APRIL)
on the one hand, and lymphoproliferation, autoimmunity, and synergized with LPS in promoting plasma cell differentiation, and
malignancy on the other, is not understood. Castigli et al8 showed that APRIL stimulated immunoglobulin
production induced by anti-CD40 and anti–IL-4 by engaging
Animal models of TACI deficiency have replicated some of the TACI, suggesting that TACI promotes antibody secretion in
features of CVID associated with TNFRSF13B mutations in human response to T cell–dependent signals. In contrast, Sakurai et al9
subjects. Thus, von Bulow et al3 found that TACI-deficient animals concluded that TACI inhibited Blimp-1 expression and IgG secre-
had severely decreased antibody responses to polysaccharides and tion by B cells activated by CD40L, suggesting that TACI inhibits
enlarged spleens because of the increased number of B cells, but responses to T cell–dependent stimuli. Previous work in TACI-
they did not observe defective antibody response to protein deficient mice revealing normal serum IgG and normal antibody
antigens, autoimmune manifestations, or malignancies. Yan et al4 responses to vaccination with protein antigens3,4 challenged the
also found in independently generated TACI-deficient mice, expan- idea that TACI promotes antibody production in response to protein
sion of the mature B cell compartment, including follicular, antigens. Because the pronounced hypogammaglobulinemia ob-
marginal zone, and transitional B cells, causing enlarged B-cell served in human subjects with mutations in the gene encoding
follicle areas and deficient antibody responses to T-independent TACI suggested that TACI might promote antibody production
type II antigens. In a follow-up study of aging TACI-deficient mice, more broadly than earlier studies indicated,3,4 we undertook to
Seshasayee et al5 reported lymphomas, lymphocytic infiltration of examine the molecular mechanisms responsible for the phenotype
vital organs, and systemic lupus erythematosus with membranopro- of TACI deletion mutants. Results indicate that TACI on B cells

Submitted May 10, 2011; accepted September 30, 2011. Prepublished online The publication costs of this article were defrayed in part by page charge
as Blood First Edition paper, October 7, 2011; DOI 10.1182/blood-2011-05-353961. payment. Therefore, and solely to indicate this fact, this article is hereby
marked ‘‘advertisement’’ in accordance with 18 USC section 1734.
The online version of this article contains a data supplement.
© 2011 by The American Society of Hematology

5832 BLOOD, 24 NOVEMBER 2011 ⅐ VOLUME 118, NUMBER 22

From www.bloodjournal.org by guest on June 23, 2016. For personal use only.

BLOOD, 24 NOVEMBER 2011 ⅐ VOLUME 118, NUMBER 22 TACI INDUCES Blimp-1 5833

induces sustained Blimp-1 expression, which in turn limits clonal ELISPOT
expansion and enhances differentiation of ASCs. We found that
TACI promotes the generation and/or maintenance of long-lived ELISPOT was performed as described previously.6 The number of spots of
plasma cells. Thus, a single molecular event explains both activat- NP-specific IgM- or IgG-secreting cells was counted by ImmunoSpot
ing and inhibiting functions of TACI and provides a logical basis Professional Analyzer Version 5.0.9 (Cellular Technology) and confirmed
for the manifestations of CVID resulting from defective TACI. by direct observation.

Methods RT-PCR and quantitative PCR

Mice and immunizations Reverse transcription was prepared with SuperScript III First-Strand
Synthesis System (Invitrogen). Quantitative PCR reactions were run on
C57BL/6 mice were purchased from The Jackson Laboratory. Quasi- Mastercycler ep realplex (Eppendorf), quantified by FastStart Universal
monoclonal (QM)–, TACI-wt, and QM-TACI-knockout (ko) mice were SYBR Green Master (Roche Diagnostics) incorporation, and results were
previously described.6,10,11 Mice were housed in a specific pathogen-free expressed relative to ␤-actin. Primer sequences and amplification condi-
facility at the University of Michigan. All animal procedures were approved tions are detailed in supplemental Methods (available on the Blood Web
by the University of Michigan Committee on Use and Care of Animals. site; see the Supplemental Materials link at the top of the online article).
Mice were immunized with 100 ␮g of (4-hydroxy-3-nitrophenyl) acetyl
(NP) conjugated to ovalbumin (NP-OVA; Biosearch Technologies) in an Flow cytometry
emulsion prepared with incomplete Freund adjuvant (Difco Laboratories).
Mice were bled before immunization and every 7 days after immunization FITC-conjugated anti–mouse CD69 (H1.2F3), PE-conjugated anti–mouse
for 8 weeks. IgMa, (DS-1), and CD19 (1D3) antibodies were purchased from BD
Biosciences. Anti-idiotypic antibodies were kindly supplied by Dr Imanishi-
Adoptive transfer Kari (Tufts University, Boston, MA). Data were acquired with BD
FACSCanto II (BD Biosciences) and analyzed using FACSDiva Version 6.1.1
TACI-proficient or TACI-deficient B cells or T cells were isolated from the software (BD Biosciences).
spleens of mice, with isolation kits (Miltenyi Biotec). A total of 5 million
B cells were mixed with 5 million T cells and injected in the tail vein of Immunohistochemistry
C57BL/6-TACI-ko mice. Sera and spleens were collected 14 days after
immunization. Peyer patches removed from mice were covered with OCT (Sakura),
immediately frozen in liquid nitrogen, and stored at Ϫ80°C. Five-micron-
Cell lines and culture conditions thick sections were mounted on positively charged microscope slides
(Fisher Scientific). Sections were air-dried for 30 minutes at room
The 18.81 cells were cultured according to published methods.12 Recombi- temperature, fixed with acetone for 10 minutes at Ϫ20°C, and air-dried for
nant mouse APRIL was purchased from PeproTech. 10 minutes at room temperature. The specimens were pretreated with
0.3% hydrogen peroxide to quench the endogenous peroxidase activity,
Irradiation of cells goat anti–rat IgG (Jackson ImmunoResearch Laboratories) followed by
development using Vectastain Elite ABC Kit (Vector Laboratories). The
B cells isolated from spleens of TACI-ko mice or 18.81 cells12 were specimens were counterstained with hematoxylin, dehydrated in
irradiated in the Experimental Irradiation Core of the University of graded ethanols, cleared in xylene. Digital images were obtained by Leica
Michigan Comprehensive Cancer Center. Cells received 1 Gy or DM6000 B microscope (Leica) and with QCapture Pro Version 6.0.0
4 Gy delivered at a dose rate of 480.8 cGy/min. Irradiated cells were software (QImaging).
cultured in supplemented RPMI 1640 medium with 5 g/mL LPS (Sigma-
Aldrich; B cells) or without LPS (18.81 cells). Cells were collected to Results
extract RNA 24 hours or 48 hours after irradiation.
TACI is necessary for the production and maintenance of
DNA double-strand breaks were identified 1 hour after irradiation by serum immunoglobulin of IgM, IgG, and IgA isotypes
enumerating ␥H2AX foci identified with a rabbit anti–mouse ␥H2AX
antibody (Abcam) detected with FITC-conjugated goat anti–rabbit IgG Deficiency of TACI in human subjects causes hypogammaglobu-
antibody (Jackson ImmunoResearch Laboratories). Digital images were linemia. However, earlier studies in TACI-deficient mice3,4 failed to
obtained with a Leica DMI6000 B microscope and LAS AF6000 Core detect hypogammaglobulinemia, suggesting the possibility that
Application Software (Leica Microsystems). murine and human TACI may have distinct functions. To verify
earlier observations in mice, we quantified Ig isotypes in the serum
TACI-expression vectors of TACI-deficient and in wild-type mice, 8 to 12 weeks of age,
housed in the same specific pathogen-free room at the University of
Mouse full-length TACI fragment was amplified from cDNA clone Michigan. Our results (Figure 1A) show that, in contrast to earlier
BC141867 and cloned in pIRES2-enhanced green fluorescent protein published results,3,4 the concentrations of all Ig isotypes were
(EGFP) vector (Clontech) to originate a single bicistronic mRNA encoding reduced in mice that lack TACI. Thus, C57BL/6-TACI-ko mice had
TACI and EGFP. TACI expression on the cell surface was confirmed by concentrations of IgM in sera reduced by 12-fold, 5-fold less serum
flow cytometric analysis (data not shown). IgG1, 13-fold less IgG2b, 4-fold less IgG3, and 5-fold less IgA
compared with wild-type mice. Because antibodies in the sera of
ELISA mice maintained in specific pathogen-free facilities accumulate
slowly from birth, earlier results may have failed to detect
ELISA was performed as described previously.10,13 Plates were developed hypogammaglobulinemia3 resulting from immaturity of the im-
with ABST (Southern Biotechnology) read at 405 nm in microplate reader mune system in the very young mice analyzed. Our results showing
Synergy 2 (Biotec Laboratories) and analyzed using Gen5 Data Analysis reduction in the concentrations of IgM, IgG, and IgA serum
Software Version 1.04.5 (BioTek). isotypes suggest that TACI may promote and/or help to maintain

From www.bloodjournal.org by guest on June 23, 2016. For personal use only.

5834 TSUJI et al BLOOD, 24 NOVEMBER 2011 ⅐ VOLUME 118, NUMBER 22

Figure 1. TACI deficiency decreased serum IgM, IgG,
and IgA and impaired antigen-specific IgM produc-
tion in response to a protein antigen. (A) Sera IgM,
IgG1, IgG3, IgG2b, and IgA quantified by ELISA. Sera
were obtained from nonimmunized C57BL/6-TACI-wt
(TACI-wt) or C57BL/6-TACI-ko (TACI-ko) mice. Concen-
trations were calculated by comparison with standard
curves obtained with known concentrations of the respec-
tive isotypes. (B-C) Concentrations of NP-specific IgM
(B) or IgG (C) determined by ELISA in the sera of
C57BL/6-TACI-wt (TACI-wt) or C57BL/6-TACI-ko
(TACI-ko) mice immunized with 100 ␮g NP-OVA 21 days
earlier. TACI deficiency decreased NP-specific IgM but
not NP-specific IgG. Concentrations were calculated by
comparison with standard curves obtained with known
concentrations of NP-specific monoclonal antibodies.
(D-E) Concentrations of NP-specific IgM (D) or IgG1
(E) determined by ELISA in the sera of QM-TACI-wt or
QM-TACI-ko mice after immunization with 100 ␮g NP-
OVA. TACI deficiency decreased NP-specific IgM produc-
tion and accelerated the production of NP-specific IgG.
Concentrations were calculated by comparison with stan-
dard curves obtained with known concentrations of
NP-specific monoclonal antibodies. Each dot represents
results obtained from one mouse. Averages were com-
pared by t test. Significant differences are noted: *P Ͻ .05;
**P Ͻ .01.

antibody responses to a variety of stimuli, including both polysac- tion of serum NP-specific IgM compared with wild-type mice
charides and proteins. (12.8 ␮g/mL, on average, in QM-TACI-ko mice and 106.6 ␮g/mL,
on average, in wild-type; Figure 1D). As in mice with wild-type
To test whether TACI promotes antibody responses to protein B cell repertoire, TACI deficiency in QM mice accelerated early
antigens, we immunized TACI-ko or TACI-wt mice with a model NP-specific IgG production. By 21 days after immunization,
protein-hapten conjugate NP-OVA. We immunized mice with QM-TACI-ko mice had 253.4 ␮g/mL NP-specific IgG1, whereas
NP-OVA mixed in incomplete Freund adjuvant rather than mixed QM-TACI-wt mice had 152.7 ␮g/mL NP-specific IgG1 in sera
with complete Freund adjuvant to avoid masking the protein- (Figure 1E). Thus, TACI deficiency impairs the primary IgM
antigen specific primary responses by introducing bacterial anti- response to protein antigens and enhances early IgG production.
gens known to activate Toll-like receptors in polyclonal B cells
independently of T-cell help. We found, as Figure 1B shows, that TACI promotes differentiation of Ig-secreting cells in response
TACI deficiency severely blunts NP-specific IgM production in to protein antigens
response to immunization. In contrast, TACI deficiency hastens the
kinetics of NP-specific IgG production yielding 2.2-fold more IgG Inability to form IgM could be caused by decreased frequency of
than wild-type mice 21 days after immunization (TACI-ko mice antigen-specific B cells, defective activation in response to immu-
had 800 Ϯ 127 ␮g/mL IgG and TACI-wt mice had nization, or alternatively could reflect a selective defect in the
1708 Ϯ 198 ␮g/mL IgG, P ϭ .0051, Figure 1C). However, the differentiation of IgM B cells. To determine whether TACI
increase in the concentration of IgG in sera of TACI-deficient mice deficiency decreased the proportion of antigen-specific B cells in
was transient because by 35 days after immunization these animals the naive compartment, we detected NP-specific B cells in QM-
had significantly less NP-specific IgG (830.6 ␮g/mL, on average), TACI-wt and in QM-TACI-ko mice, 10 days after immunization.
compared with wild-type mice (1329 ␮g/mL, on average). These Figure 2A and B shows that QM-TACI-ko mice have 2.5-fold more
results suggested that, in contrast to previously published reports,8 NP-specific IgM positive B cells in the spleen than QM-TACI-wt
TACI deficiency does not impair early isotype class-switched mice, indicating that defective NP-specific IgM production is not
antibody responses to proteins. Instead, TACI deficiency impairs the result of a decreased number of antigen-specific B cells in
sustained IgG responses after immunization. TACI-deficient mice. To test whether TACI deficiency impaired
B-cell activation or clonal expansion in response to protein
To uncover the specific mechanism by which TACI promotes antigens, we measured the proportion of B cells expressing
IgM antibody production and to avoid biases resulting from CD69 (an activation marker) and detected germinal center forma-
variability in the frequency of antigen-specific B cells, we took tion in Peyer patches after immunization. We found that the
advantage of a mouse model engineered to increase the representa- proportion of activated CD69ϩ B cells was comparable in QM-
tion of a single B-cell clone in the B-cell repertoire. In QM mice, TACI-ko or in QM-TACI-wt mice (Figure 2C). We also observed
80% of the peripheral B cells express an antibody that recognizes that QM-TACI-ko mice had increased size and more germinal
the hapten NP.10 In nonimmunized QM mice, there is very little centers (GL7ϩ, Ki67ϩ) in the Peyer patches, compared with
NP-specific IgM (Ͻ 50 ␮g/mL) and NP-specific IgG is undetect- QM-TACI-wt mice (Figure 2D). These results suggested that TACI
able.10,14 This fact makes the QM mouse an ideal model to follow deficiency does not impair B-cell activation or clonal expansion
the development of NP-specific B cells in response to immuniza- after immunization with protein antigens.
tion. We immunized TACI-proficient QM mice (QM-TACI-wt) or
TACI-deficient QM mice (QM-TACI-ko) with NP-OVA, as de- To test whether TACI promotes differentiation of NP-specific
scribed. At the peak of the primary antibody response, 21 days after B cells into ASCs, we enumerated the NP-specific IgM or IgG
immunization, QM-TACI-ko mice had only 12% of the concentra-

From www.bloodjournal.org by guest on June 23, 2016. For personal use only.

BLOOD, 24 NOVEMBER 2011 ⅐ VOLUME 118, NUMBER 22 TACI INDUCES Blimp-1 5835

Figure 2. TACI deficiency does not impair B-cell activation and
germinal center formation. (A) The proportion of NP-specific
B cells (idiotype-positive) in the spleen of QM-TACI-wt or QM-
TACI-ko mice was analyzed by flow cytometry in mice immunized
10 days earlier. Splenocytes were stained with an anti–idiotypic
antibody10 and with an anti–IgMa antibody that detects the allotype
produced by the heavy chain-targeted allele. (B) Average propor-
tions (above) or average number (below) of idiotype-positive cells
obtained from 3 QM-TACI-wt or 3 QM-TACI-ko mice. Averages were
compared by t test. Significant differences: *P Ͻ .05; **P Ͻ .01.
(C) TACI deficiency does not decrease B-cell activation after
immunization. Splenocytes were obtained from QM-TACI-wt or
QM-TACI-ko mice immunized with 100 ␮g NP-OVA 10 days earlier.
Splenocytes were stained with anti-CD19 and CD69 antibodies and
analyzed by FACS. Activated B cells (CD19ϩ, CD69ϩ) were as
frequent in TACI-proficient as in TACI-deficient mice. Average
proportions of CD19ϩ, CD69ϩ cells obtained from 3 QM-TACI-wt, or
3 QM-TACI-ko mice. Averages were compared by t test. Differences
between groups were not significant. (D) TACI deficiency caused
increases in size and number of germinal centers 10 days after
immunization. Frozen sections of Peyer patches were stained with
anti–GL7 antibody (brown), which specifically labels germinal center
B cells. The figure is representative of 3 independent experiments.
Digital images were obtained by Leica DM6000 B microscope
(Leica) and with QCapture Pro Version 6.0.0 software (QImaging).

ASCs in the spleen of C57BL/6-TACI-wt or C57BL/6-TACI-ko TACI controls differentiation of Ig-secreting cells in a B
mice at the peak of the immune response, 21 days after immuniza- cell-autonomous manner
tion. TACI deficiency caused a significant decrease in the number
of NP-specific IgM ASCs in the spleen (6.250 Ϯ 1.03/106 B cells in Antibody response to proteins requires T-cell help. Because,
ko mice and 16.40 Ϯ 3.140/106 B cells in wt mice; Figure 3A). according to some studies, activated T cells express TACI,3 we
TACI deficiency also caused a reduction (albeit not statistically wondered whether defective antibody secretion was the result of
significant) in the number of NP-specific IgG ASCs in the spleen TACI deficiency on B cells, TACI deficiency on T cells, or both. To
(10.50 Ϯ 2.021/106 cells in ko and 24.60 Ϯ 6.120/106 cells in wt answer that question, we enumerated ASCs in C57BL/6-TACI-ko
mice; Figure 3B). To determine whether TACI deficiency impaired mice reconstituted with either QM-TACI-ko or QM-TACI-wt
long-lived plasma cells, we enumerated antigen-specific ASCs in B cells, and C57BL/6-TACI-ko or C57BL/6-TACI-wt T cells.
the bone marrow 56 days after immunization of QM-TACI-ko or Chimeras were immunized with 100 ␮g of NP-OVA shortly after
QM-TACI-wt mice. QM-TACI-wt mice had on average 4-fold cell transfer. TACI-wt B cells were necessary and sufficient to
more antigen-specific IgM ASCs (Figure 3C) and at least 5-fold produce NP-specific IgM and IgG secreting cells. Mice reconsti-
more IgG ASCs (Figure 3D) in the bone marrow than QM- tuted with TACI-wt B and TACI-wt T cells had on average
TACI-ko mice. These results indicate that TACI is necessary for 75.3 NP-specific IgM ASCs/million B cells, whereas mice reconsti-
development or maintenance of long-lived ASCs. tuted with TACI-ko B cells and TACI-wt T cells had only
8.7 NP-specific IgM ASCs/million B cells in the spleen, 14 days

Figure 3. TACI is required for differentiation of
B cells into ASCs. TACI-ko animals had decreased
NP-specific IgM ASCs or NP-specific IgG ASCs com-
pared with wild-type. ASCs were enumerated in B cells
isolated from the spleen (A-B) or in bone marrow cell
suspensions (C-D) of C57BL/6-wt or C57BL/6-TACI-ko
21 days after immunization, by ELISPOT. Below the
graphs, the diagrams show the ELISPOT wells from
which the ASCs were enumerated and plated with 106,
5 ϫ 105, or 105 cells, as indicated. Values were com-
pared by t test. Significant differences: *P Ͻ .05;
**P Ͻ .01. Data resulted from the analyses of 5 wild-type
mice and 4 TACI-ko mice.

From www.bloodjournal.org by guest on June 23, 2016. For personal use only.

5836 TSUJI et al BLOOD, 24 NOVEMBER 2011 ⅐ VOLUME 118, NUMBER 22

Figure 4. TACI deficiency decreases production of
ASCs in a B cell–autonomous manner. (A-B) The
number of ASCs was restored by adoptive transfer of
TACI-wt B cells but not by transfer of TACI-wt T cells
into TACI-deficient recipients. NP-specific IgM (A) or
NP-specific IgG (B) ASCs were enumerated by
ELISPOT of splenocytes 14 days after immunization,
and after reconstitution by adoptive transfer of QM-
TACI-wt or QM-TACI-ko B cells, and C57BL/6-wt or
C57BL/6-TACI-ko T cells. Representative pictures of
the ELISPOT wells are shown below each graph.
Transfer efficiency was equivalent in all experiments
and determined by enumerating IgMa-positive donor
B cells by FACS analysis in the spleen of recipients at
the time of analysis. The number of IgM-secreting cells
was significantly decreased whenever reconstitution
was done with TACI-ko B cells: *P ϭ .016 (2-way
ANOVA). The number of IgG-secreting cells in animals
reconstituted with TACI-ko B cells and TACI-wt T cells
was significantly decreased compared with animals
reconstituted with TACI-wt B and T cells: *P ϭ .041
(t test). Data resulted from analysis of 3 recipient mice
for each type of transfer.

after immunization (Figure 4A). Figure 4B shows that transfer of was motivated by the fact that the line exhibits some properties of
TACI-deficient B cells also resulted in a decreased number of germinal center B cells; 18.81 B cells proliferate very fast and
NP-specific IgG ASCs. Thus, mice reconstituted with TACI-wt undergo somatic hypermutation and Ig class switch recombina-
B and TACI-wt T cells had on average 26 NP-specific IgG tion.12 We studied 18.81 cells transfected with vectors encoding
ASC/million B cells, whereas mice reconstituted with TACI-ko TACI and GFP in a single bicistronic mRNA or with a control
B cells and TACI-wt T cells had only 5.3 NP-specific IgG vector encoding GFP. To test whether TACI expression was
ASCs/million B cells in the spleen, 14 days after immunization. sufficient to induce Blimp-1 expression, we measured Blimp-1
Because in chimeras reconstituted with TACI-proficient B cells, the RNA in sorted TACI and/or GFP-expressing cells. Figure 5C and D
number of antigen-specific IgM or IgG ASCs in the spleen did not shows that TACI expression induces Blimp-1 expression in
significantly differ in mice that had received TACI-wt 18.81 cells by Ͼ 3-fold, indicating that TACI induces Blimp-1
(75.3 NP-specific IgM or 26 NP-specific IgG ASC/million B cells, expression on its own. In addition, as TACI expression decreases
respectively) or TACI-ko T cells (97.7 NP-specific IgM or over the course of 3 days, so does Blimp-1 expression, whereas
16 NP-specific IgG ASC/million B cells), we conclude that TACI BCL-6 expression increases (Figure 5E). These results suggest that
expression by B cells determines Ig production autonomously TACI orchestrates the genetic changes necessary to initiate plasma
(Figure 4A-B). cell differentiation.

TACI induces Blimp-1 production by B cells DNA double-strand breaks induce Blimp-1 production by
TACI-deficient cells
Differentiation of germinal center B cells requires cessation of
B-cell proliferation and is at least in part dependent on Blimp-1 TACI-deficient mice have less IgG in the sera compared with
expression,15 even though initiation of B-cell differentiation may TACI-proficient mice, suggesting that TACI contributes to IgG
be Blimp-1 independent.16 Current thought supports a model in production in the mouse, thus mimicking the hypogammaglobuline-
which Blimp-1 expression enhances gene-expression changes mia observed in human subjects with deficiencies in TACI.
necessary for complete plasma cell maturation. Thus, Blimp-1, a Surprisingly, measurements of antigen-specific IgG in TACI-
transcriptional repressor, halts the cell cycle in part because of deficient animals after immunization with a T-dependent antigen
repressing PAX5, Myc, and BCL-6 and inducing X-box binding revealed hastened kinetics and at times increased concentration of
protein 1 (XBP1), which orchestrates the unfolded protein reaction IgG in the serum, compared with TACI-wt animals. Yet the number
needed to induce the secretory phenotype. We asked whether TACI of antigen-specific ASCs in the spleen 3 weeks after immunization
may promote differentiation of IgM ASCs by increasing Blimp-1 was reduced in TACI-deficient mice compared with TACI-
expression in centrocytes. To test whether TACI deficiency im- proficient mice. Because class switch recombination requires
paired Blimp-1 expression and because Blimp-1 expression is double-strand breaks, we asked whether DNA double-strand breaks
controlled at the level of transcription, we first examined Blimp-1 would induce Blimp-1 independently of TACI, thus promoting
mRNA in B cells obtained from spleens of QM-TACI-wt or differentiation of IgG-secreting plasma cells. To test whether DNA
QM-TACI-ko mice after immunization. Figure 5A and B shows double-strand breaks induce Blimp-1 expression, we irradiated
that immunization increases Blimp-1 expression by 1.8-fold in 18.81 cells or TACI-ko B cells and detected ␥H2AX foci by
TACI-proficient B cells, whereas it reduces expression of Blimp-1 immunofluorescence. Figure 6A and B shows that irradiation of
by 0.6-fold in TACI-deficient B cells, relative to expression before 18.81 (A) or TACI-ko B cells (B) resulted in ␥H2AX foci, which
immunization. The results suggested that TACI induces Blimp-1 on are induced by DNA double-strand breaks.18 Next, we determined
B-cell activation. whether irradiation induced Blimp-1 expression by quantitative
RT-PCR. Figure 6C shows that irradiation of 18.81 cells in the
To determine whether expressing TACI was sufficient on its presence of TACI-Fc, which competes for the TACI ligands,
own to induce Blimp-1, we expressed TACI in Abelson murine induced Blimp-1 expression after 24 hours. Figure 6D shows that
leukemia virus-transformed (18.81) B cells12,17 and quantified irradiation of TACI-deficient cells obtained from TACI-ko mice
Blimp-1 production in transfected cells. The choice of 18.81 B cells

From www.bloodjournal.org by guest on June 23, 2016. For personal use only.

BLOOD, 24 NOVEMBER 2011 ⅐ VOLUME 118, NUMBER 22 TACI INDUCES Blimp-1 5837

Figure 5. TACI is necessary and sufficient to induce
Blimp-1 in B cells. (A-B) TACI deficiency abrogates Blimp-1
expression by B cells 14 days after immunization. Blimp-1
mRNA expression was detected by RT-PCR (A) and quanti-
fied by real-time PCR (B). Relative expression of Blimp-1
was calculated by determining the ratio of Blimp-1/␤-actin
in immunized B cells compared with the Blimp-1/␤-actin
ratio in B cells from nonimmunized mice. Results represent
the mean Ϯ SEM of 3 measurements/mouse of each of
4 mice/genotype. Values were compared by t test and found
to be significantly different: ***P Ͻ .0001. (C-D) Expression of
TACI is sufficient to induce Blimp-1. The 18.81 B cells were
transfected with a vector encoding TACI and EGFP from a
single bicistronic mRNA or with a control vector encoding
only EGFP. RNA was isolated from sorted GFP-positive
transfected cells, 24 hours after transfection and converted
to cDNA. Blimp-1 mRNA expression was studied by RT-
PCR, after 30 PCR cycles (C) or quantitative RT-PCR (D).
Blimp-1 expression by 18.81 B cells expressing TACI was
calculated relative to the Blimp-1 expression by 18.81 cells
transfected with a GFP control vector. Results represent
mean Ϯ SEM of 3 separate measurements. Values were
compared by t test and found to be significantly different:
*P ϭ .0103. (E) Blimp-1 and TACI expression decreased
and BCL-6 expression increased with time in 18.81 B cells
transfected with a TACI expression vector. The 18.81
B cells transfected with TACI and EGFP or with EGFP
control vectors were cultured up to 3 days with or without
100 ng/mL APRIL, as indicated. TACI, BCL-6, and Blimp-1
mRNA expression was studied by RT-PCR, after 30 PCR
cycles. Cells were collected at 0, 24, 48, and 72 hours
in culture. Results are representative of 3 independent
experiments.

generated ␥H2AX foci and induced Blimp-1, demonstrating that long-lived plasma cells, thus causing hypogammaglobulinemia in
DNA double-strand breaks induce Blimp-1 independently of TACI. the long term.
Plasma cell differentiation requires sustained Blimp-1 expression,
but induction of Blimp-1 by DNA double-strand breaks may be Discussion
short lived (Figure 6C). With the caveat that the cellular conse-
quences of breaks induced by irradiation may differ substantially Here we show that hypogammaglobulinemia resulting from TACI
from those associated with class switch, our results suggest that the deficiency results from a primary defect in the differentiation
transient IgG secretion in TACI-deficient mice results from Blimp-1 and/or maintenance of long-lived ASCs. Results in this manuscript
induction by DNA breaks associated with class switch recombina- show, for the first time, that TACI deficiency impairs Blimp-1
tion. Thus, in the absence of TACI, as DNA breaks are repaired, expression in B cells responding to protein-based immunizations.
Blimp-1 expression may cease, failing to sustain differentiation of

Figure 6. Ionizing radiation induces Blimp-1 expres-
sion independently of TACI. (A-B) X-ray irradiation of
18.81 cells (A) or C57BL/6-TACI-ko cells (B) induces
␥H2AX foci, indicating accumulation of DNA double-
strand breaks. Cytospin slides of nonirradiated (0 Gy)
or irradiated (4 Gy) cells were stained with an anti-
␥H2AX antibody (top diagrams) and counterstained with
4,6-diamidino-2-phenylindole (bottom diagrams). (C) Irra-
diation of 18.81 B cells with 1 Gy or 4 Gy induces Blimp-1
expression transiently, detected by quantitative RT-PCR.
Relative expression of Blimp-1 was calculated by determin-
ing the ratio Blimp-1/␤-actin for each sample. Values
reflect 3 independent measurements. (D) Irradiation in-
duces Blimp-1 expression independently of TACI. TACI-ko
B cells, isolated from C57BL/6-TACI-ko mice, were irradi-
ated with 4 Gy and cultured with 15 ␮g/mL of LPS for
24 hours. Blimp-1 expression increased after irradiation of
TACI-deficient B cells. RNA was obtained after 24 hours
in culture and expression of Blimp-1 or ␤-actin analyzed
by quantitative RT-PCR. Relative expression of Blimp-1
was calculated by determining the ratio Blimp-1/␤-actin
for each sample. Values reflect 3 independent mea-
surements. Averages were compared by t test. Signifi-
cant differences were noted: *P Ͻ .05, **P Ͻ .01, and
***P Ͻ .001.

From www.bloodjournal.org by guest on June 23, 2016. For personal use only.

5838 TSUJI et al BLOOD, 24 NOVEMBER 2011 ⅐ VOLUME 118, NUMBER 22

Blimp-1 expression decreases clonal expansion in part by inhibit- IgG, even though IgG ASCs do not persist. In agreement with this
ing BCL-6, thus limiting the germinal center reaction. Thus, concept, we showed that DNAdouble-strand breaks induced by ionizing
TACI-deficient animals produce severely blunted IgM responses radiation caused transient expression of Blimp-1, even in the absence of
and hypertrophic germinal centers after immunization and are TACI. Because B cells undergoing class switch recombination undergo
severely hypogammaglobulinemic. Our findings help to explain the DNA double-strand breaks,21 B cells expressing class-switched isotypes
phenotype of CVID resulting from mutations in the TNFRSF13B may initiate antibody secretion as a direct consequence of persisting
gene that encodes TACI. Thus, TACI defects impair Blimp-1 DNA breaks. However, because continuous Blimp-1 expression is
expression and consequently decrease differentiation or mainte- necessary to maintain plasma cell differentiation, transient expression of
nance of long-lived plasma cells causing hypogammaglobulinemia. Blimp-1 is not compatible with long-lived plasma cell survival.15,22,23
Conversely, decreased Blimp-1 expression enhances clonal expan- Thus, in contrast to TACI-proficient mice, TACI-deficient animals had
sion in response to antigenic stimulation causing lymphoprolifera- no long-lived antigen-specific IgG 56 days after immunization.
tion which, in turn, may increase susceptibility to neoplastic
transformation and autoimmunity. Our findings indicating that TACI promotes development of
ASCs raise the question of how TACI-proficient B cells determine
We have previously shown that TACI deficiency impairs when to differentiate after activation. Two possibilities come to
antibody responses to polysaccharide antigens, explaining the mind. In one, TACI expression or function may change after B-cell
increased susceptibility to infection with encapsulated organisms activation; in the other, the availability of TACI ligands may
of subjects with CVID resulting from TACI mutations.3,6 We and increase as B cells grow. In agreement with the idea that TACI
others showed that TACI on B cells facilitates differentiation into expression changes on B-cell activation are studies by Castigli
ASCs in response to T-independent antigens.6,7 However, whether et al8 who showed that stimulation with CD40L increases TACI
or not TACI promotes differentiation of ASCs in response to expression by B cells, and studies by Benson et al24 showing that
protein-based vaccines has been a matter of controversy. Indeed, immunization with protein antigens causes a transient decrease in
some of us and others have previously found apparently normal TACI expression by germinal center B cells, thus precluding
serum Ig and antibody responses to protein antigens in TACI- premature B-cell differentiation. Regulation of TACI expression
deficient mice.3,4 Sakurai et al proposed that TACI may actually may not be the only mechanism controlling differentiation of
inhibit T cell–dependent antibody responses because TACI inhib- B cells into ASCs. Indeed, memory B cells in both humans and
ited Blimp-1 expression and IgG secretion by B cells activated by mice express high levels of TACI24 despite their quiescent state.
CD40L.9 In contrast, Castigli et al showed that APRIL stimulated Consistent with the idea that the availability of TACI ligands
immunoglobulin production induced by anti-CD40 and anti–IL-4 determines B-cell differentiation is the work by Chu et al who
by engaging TACI, suggesting that TACI promotes antibody proposed that activated B cells secrete biologically active B cell–
secretion in response to T cell–dependent signals.8 To understand activating factor of the tumor necrosis factor family (BAFF) and/or
the mechanisms by which TACI promotes immune fitness and how APRIL, either of which may activate TACI.25 According to Chu
TACI mutations associated with CVID in human subjects cause et al, germinal center B cells and follicular dendritic cells produce
hypogammaglobulinemia, we considered 2 possibilities: that TACI BAFF upon immunization.25 Increasing TACI ligands in the
induces isotype class switch and therefore production of switched germinal center may thus provide a feedback loop to control clonal
isotypes and/or that TACI promotes differentiation of ASCs. expansion by inducing Blimp-1 expression. In addition, TACI
ligands may accumulate in certain locations as part of niches
There is strong evidence to support that engagement of TACI fostering differentiation of plasma cells. Thus, Mohr et al showed
enhances Ig isotype class switch recombination in conjunction with that the greatest concentration of APRIL was restricted to the
other receptors on B cells. Thus, Castigli et al showed that ASC-associated myeloid cells colonizing the medullary cords in
TACI-deficient B cells produced hardly any class-switched Ig and lymph nodes, suggesting that the local production of these agonists
had decreased Ig␥1 and Ig⑀ germline and I␮-␥1 and I␮-⑀ tran- contributes to the early differentiation of ASCs in this location after
scripts on stimulation with APRIL and IL-4.19 He et al have immunization with alum-precipitated ovalbumin.26
recently shown that engagement of TACI triggers Ig class switch
recombination by directly activating MyD88 in conjunction with The concept that TACI activity controls B-cell differentiation in
simultaneous stimulation of coreceptors.20 However, because TACI- a dose-dependent manner predicts that clinical manifestations may
deficient B cells maintained the ability to class switch in response occur as a consequence of haploinsufficiency. Indeed, human
to stimulation with an anti-CD40 antibody and IL-4 or to LPS and subjects with CVID resulting from mutations in the TNFRSF13B
IL-4,19 the results suggested that, although TACI may enhance, it is gene express a nonmutant allele as well. And most of the known
not absolutely necessary for isotype class switch recombination. mutations do not abolish TACI expression but instead generate
Consistent with that conclusion, our results showed robust, albeit receptors that are either deficient in ligand binding and/or have
transient, IgG production in immunized TACI-deficient animals. impaired signaling.27-29 In support, Chinen et al showed that
Indeed, TACI-deficient mice produced greater amounts of IgG 21 haploinsufficiency of the TACI gene results in humoral immune
days after immunization with a protein antigen, compared with dysfunction in patients with Smith-Magenis syndrome,30 and Lee et al
TACI-proficient mice, indicating that TACI expression by B cells is showed that expression of human TACI mutant alleles in mice caused
not necessary to induce isotype class switching to IgG in vivo. defective antibody production even in the presence of wild-type
alleles.31 Which mechanism, control of TACI expression, or the
Our results indicated that TACI deficiency primarily impaired availability of TACI ligands, better explains how TACI regulates B-cell
differentiation of ASCs rather than isotype class switching. Thus, differentiation is incompletely resolved and the subject of present
we found that, despite a reduction in the number of IgG-secreting cells in inquiry.
the spleen, TACI deficiency caused a transient increase in antigen-
specific IgG relative to TACI-proficient controls in response to immuni- Our results indicating that TACI-induced Blimp-1 provides, for
zation. We propose that extensive germinal center reactions in TACI- the first time, a molecular and cellular mechanism unifying the key
deficient animals cause accumulation of DNA breaks that in turn induce features associated with CVID resulting from deficiencies in TACI.
Blimp-1 transiently. Transient Blimp-1 expression induces secretion of Deficient generation and/or maintenance of ASCs explain the

From www.bloodjournal.org by guest on June 23, 2016. For personal use only.

BLOOD, 24 NOVEMBER 2011 ⅐ VOLUME 118, NUMBER 22 TACI INDUCES Blimp-1 5839

hypogammaglobulinemia and the exquisite vulnerability to infec- Authorship
tion with encapsulated organisms associated with TACI deficiency.
Increased proliferation and accumulation of DNA breaks explain Contribution: S.T. performed research, analyzed and interpreted
both increased frequency of autoimmune manifestations in TACI- data, and wrote the manuscript; C.C. performed research; R.J.B.
deficient animals5 and in persons with CVID because of TACI contributed vital reagents and analyzed and interpreted data; J.L.P.
deficiency,32 as well as the increased frequency of neoplasia in analyzed and interpreted data and wrote the manuscript; and M.C.
some CVID patients.1 designed research, analyzed and interpreted data, and wrote the
manuscript.
Acknowledgments
Conflict-of-interest disclosure: R.J.B. is listed as a coinventor
The authors thank Ms Karen Lien who helped with immunochem- on patents regarding the TACI gene and protein. The remaining
istry, all the members of our laboratories for comments and authors declare no competing financial interests.
criticism, and Ms Jane Bugden for editorial assistance.
Correspondence: Marilia Cascalho, Transplantation Biology,
This work was supported by the National Institutes of Health University of Michigan, 109 Zina Pitcher Pl, Ann Arbor, MI 48109;
(P01 HL079067, J.L.P. and M.C.; and R37 HL052297, J.L.P). e-mail: [email protected].

References

1. Park MA, Li JT, Hagan JB, Maddox DE, lation of the T-independent humoral response by long-lived plasma cells in the bone marrow. J Exp
Abraham RS. Common variable immunodeficien- TACI. Immunity. 2001;14(5):573-582. Med. 2005;202(11):1471-1476.
cy: a new look at an old disease. Lancet. 2008;
372(9637):489-502. 12. Wabl M, Burrows PD, von Gabain A, Steinberg C. 23. Kallies A, Hasbold J, Tarlinton DM, et al. Plasma
Hypermutation at the immunoglobulin heavy cell ontogeny defined by quantitative changes in
2. Wehr C, Kivioja T, Schmitt C, et al. The EURO- chain locus in a pre-B-cell line. Proc Natl Acad blimp-1 expression. J Exp Med. 2004;200(8):967-
class trial: defining subgroups in common vari- Sci U S A. 1985;82(2):479-482. 977.
able immunodeficiency. Blood. 2008;111(1):77-
85. 13. Cascalho M, Wong J, Wabl M. VH gene replace- 24. Benson MJ, Dillon SR, Castigli E, et al. Cutting
ment in hyperselected B cells of the quasimono- edge: the dependence of plasma cells and inde-
3. von Bulow GU, Bram RJ. NF-AT activation clonal mouse. J Immunol. 1997;159(12):5795- pendence of memory B cells on BAFF and
induced by a CAML-interacting member of the 5801. APRIL. J Immunol. 2008;180(6):3655-3659.
tumor necrosis factor receptor superfamily.
Science. 1997;278(5335):138-141. 14. Cascalho M, Wong J, Steinberg C, Wabl M. Mis- 25. Chu VT, Enghard P, Riemekasten G, Berek C. In
match repair co-opted by hypermutation. Sci- vitro and in vivo activation induces BAFF and
4. Yan M, Wang H, Chan B, et al. Activation and ac- ence. 1998;279(5354):1207-1210. APRIL expression in B cells. J Immunol. 2007;
cumulation of B cells in TACI-deficient mice. Nat 179(9):5947-5957.
Immunol. 2001;2(7):638-643. 15. Shapiro-Shelef M, Lin KI, McHeyzer-Williams LJ,
Liao J, McHeyzer-Williams MG, Calame K. 26. Mohr E, Serre K, Manz RA, et al. Dendritic cells
5. Seshasayee D, Valdez P, Yan M, Dixit VM, Tumas D, Blimp-1 is required for the formation of immuno- and monocyte/macrophages that create the
Grewal IS. Loss of TACI causes fatal lymphopro- globulin secreting plasma cells and pre-plasma IL-6/APRIL-rich lymph node microenvironments
liferation and autoimmunity, establishing TACI as memory B cells. Immunity. 2003;19(4):607-620. where plasmablasts mature. J Immunol. 2009;
an inhibitory BLyS receptor. Immunity. 2003; 182(4):2113-2123.
18(2):279-288. 16. Kallies A, Hasbold J, Fairfax K, et al. Initiation of
plasma-cell differentiation is independent of the 27. Castigli E, Wilson SA, Garibyan L, et al. TACI is
6. Mantchev GT, Cortesao C, Rebrovich M, transcription factor Blimp-1. Immunity. 2007; mutant in common variable immunodeficiency
Cascalho M, Bram RJ. TACI is required for effi- 26(5):555-566. and IgA deficiency. Nat Genet. 2005;37(8):829-
cient plasma cell differentiation in response to 834.
T-independent type 2 antigens. J Immunol. 2007; 17. Rosenberg N, Baltimore D. A quantitative assay
179:2282-2288. for transformation of bone marrow cells by Abel- 28. Salzer U, Chapel HM, Webster AD, et al. Muta-
son murine leukemia virus. J Exp Med. 1976; tions in TNFRSF13B encoding TACI are associ-
7. Ozcan E, Garibyan L, Lee JJ, Bram RJ, Lam KP, 143(6):1453-1463. ated with common variable immunodeficiency in
Geha RS. Transmembrane activator, calcium humans. Nat Genet. 2005;37(8):820-828.
modulator, and cyclophilin ligand interactor drives 18. Rogakou EP, Pilch DR, Orr AH, Ivanova VS,
plasma cell differentiation in LPS-activated B Bonner WM. DNA double-stranded breaks induce 29. Lee JJ, Rauter I, Garibyan L, et al. The murine
cells. J Allergy Clin Immunol. 2009;123(6):1277- histone H2AX phosphorylation on serine 139. equivalent of the A181E TACI mutation associ-
1286. J Biol Chem. 1998;273(10):5858-5868. ated with common variable immunodeficiency
severely impairs B-cell function. Blood. 2009;
8. Castigli E, Wilson SA, Elkhal A, Ozcan E, 19. Castigli E, Wilson SA, Scott S, et al. TACI and 114(11):2254-2262.
Garibyan L, Geha RS. Transmembrane activator BAFF-R mediate isotype switching in B cells.
and calcium modulator and cyclophilin ligand in- J Exp Med. 2005;201(1):35-39. 30. Chinen J, Martinez-Gallo M, Gu W, et al. Trans-
teractor enhances CD40-driven plasma cell differ- membrane activator and CAML interactor (TACI)
entiation. J Allergy Clin Immunol. 2007;120(4): 20. He B, Santamaria R, Xu W, et al. The transmem- haploinsufficiency results in B-cell dysfunction in
885-891. brane activator TACI triggers immunoglobulin patients with Smith-Magenis syndrome. J Allergy
class switching by activating B cells through the Clin Immunol. 2011;127(6):1579-1586.
9. Sakurai D, Kanno Y, Hase H, Kojima H, Okumura K, adaptor MyD88. Nat Immunol. 2010;11(9):836-
Kobata T. TACI attenuates antibody production 845. 31. Lee JJ, Jabara HH, Garibyan L, et al. The C104R
costimulated by BAFF-R and CD40. Eur J Immu- mutant impairs the function of transmembrane
nol. 2007;37(1):110-118. 21. Petersen S, Casellas R, Reina-San-Martin B, activator and calcium modulator and cyclophilin
et al. AID is required to initiate Nbs1/gamma- ligand interactor (TACI) through haploinsuffi-
10. Cascalho M, Ma A, Lee S, Masat L, Wabl M. A H2AX focus formation and mutations at sites of ciency. J Allergy Clin Immunol. 2010;126(6):
quasi-monoclonal mouse. Science. 1996; class switching. Nature. 2001;414:660-665. 1234-1241.
272(5268):1649-1652.
22. Shapiro-Shelef M, Lin KI, Savitsky D, Liao J, 32. Conley ME, Dobbs AK, Farmer DM, et al. Primary
11. von Bulow GU, van Deursen JM, Bram RJ. Regu- Calame K. Blimp-1 is required for maintenance of B cell immunodeficiencies: comparisons and con-
trasts. Annu Rev Immunol. 2009;27:199-227.

From www.bloodjournal.org by guest on June 23, 2016. For personal use only.

2011 118: 5832-5839
doi:10.1182/blood-2011-05-353961 originally published
online October 7, 2011

TACI deficiency impairs sustained Blimp-1 expression in B cells
decreasing long-lived plasma cells in the bone marrow

Shoichiro Tsuji, Catarina Cortesão, Richard J. Bram, Jeffrey L. Platt and Marilia Cascalho

Updated information and services can be found at:
http://www.bloodjournal.org/content/118/22/5832.full.html
Articles on similar topics can be found in the following Blood collections

Immunobiology (5393 articles)
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml

Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society
of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.