Activation-Induced Cytidine Deaminase Deficiency
Causes Organ-Specific Autoimmune Disease
Koji Hase1, Daisuke Takahashi1,2, Masashi Ebisawa1,2, Sayaka Kawano1, Kikuji Itoh3, Hiroshi Ohno1,2*
1Research Center for Allergy and Immunology, RIKEN, Yokohama, Kanagawa, Japan, 2Supramolecular Biology, International Graduate School of Arts and Sciences,
Yokohama City University, Yokohama, Kanagawa, Japan, 3Graduate School of Agricultural and Life Sciences, The University of Tokyo, Tokyo, Japan
Activation-induced cytidine deaminase (AID) expressed by germinal center B cells is a central regulator of somatic
hypermutation (SHM) and class switch recombination (CSR). Humans with AID mutations develop not only the autosomal
recessive form of hyper-IgM syndrome (HIGM2) associated with B cell hyperplasia, but also autoimmune disorders by
unknown mechanisms. We report here that AID2/2mice spontaneously develop tertiary lymphoid organs (TLOs) in non-
lymphoid tissues including the stomach at around 6 months of age. At a later stage, AID2/2mice develop a severe gastritis
characterized by loss of gastric glands and epithelial hyperplasia. The disease development was not attenuated even under
germ-free (GF) conditions. Gastric autoantigen -specific serum IgM was elevated in AID2/2mice, and the serum levels
correlated with the gastritis pathological score. Adoptive transfer experiments suggest that autoimmune CD4+T cells
mediate gastritis development as terminal effector cells. These results suggest that abnormal B-cell expansion due to AID
deficiency can drive B-cell autoimmunity, and in turn promote TLO formation, which ultimately leads to the propagation of
organ-specific autoimmune effector CD4+T cells. Thus, AID plays an important role in the containment of autoimmune
diseases by negative regulation of autoreactive B cells.
Citation: Hase K, Takahashi D, Ebisawa M, Kawano S, Itoh K, et al. (2008) Activation-Induced Cytidine Deaminase Deficiency Causes Organ-Specific Autoimmune
Disease. PLoS ONE 3(8): e3033. doi:10.1371/journal.pone.0003033
Editor: Lena Alexopoulou, Centre d’Immunologie de Marseille-Luminy, CNRS-Inserm, France
Received June 17, 2008; Accepted July 30, 2008; Published August 21, 2008
Copyright: ? 2008 Hase et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported in part by Grants-in-Aid for Young Scientists (B) (K H), Scientific Research (B) (HO), and Scientific Research in Priority Areas (HO
and KH) from the Ministry of Education, Culture, Sports, Science and Technology of Japan, Uehara Memorial Foundation (KH) and the Sasakawa Scientific Research
Grant from The Japan Science Society (KH).
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
The targeted deamination of Ig genes by AID is a prerequisite
for Ab affinity maturation through somatic hypermutation (SHM)
and class switch recombination (CSR) . Therefore, AID
deficiency leads to a defect in these two critical events in humoral
immunity, and in humans causes hyper IgM syndrome HIGM2, a
disease characterized by recurrent bacterial infections .
Enhanced proliferation of B cells and increased repertoire diversity
were observed in aged AID2/2mice, suggesting a critical role of
AID in B-cell growth regulation . AID2/2mice also display
abnormal expansion of anaerobic commensal bacteria in the small
intestine, which induces hypertrophic enlargement of Peyer’s
patches and protrusion of isolated lymphoid follicles (ILFs). The
abnormality of intestinal flora is due to the lack of hypermutated
IgA, because reconstitution of intestinal IgA production recovered
the normal composition of gut flora. These results suggest that
AID plays a key role in homeostasis of intestinal flora.
Furthermore, a fraction of patients carrying AID mutations suffer
from various organ-specific autoimmune diseases, including
diabetes mellitus, autoimmune hepatitis and Crohn’s disease, via
unknown mechanisms .
In autoimmune-mediated tissue disorders, T cells are usually
thought to be the most important cell type for controlling
autoimmune responses. On the other hand, recent studies suggest
that interactions between B and T cells play a pivotal role in the
pathogenesis of autoimmue diseases . The B-cell receptor
(BCR) in developing B-cell precursors is produced via the
rearrangement of randomly selected V, (D) and J segments of
the Ig heavy and light chain loci. This Ig gene recombination is
crucial to increase the diversity of the B-cell repertoire however,
due to its stochastic nature, a substantial number of newly
synthesized BCRs bind autoantigens. It was recently estimated
that more than 50% of newly generated B cells are autoreactive
. Studies using transgenic mice carrying autoreactive BCR
genes indicate that autoreactive B cells are normally silenced by
immunological tolerance mechanisms including clonal deletion,
receptor editing and anergy [8,9]. However, in humans and mice
that are prone to autoimmune diseases, the B-cell tolerance
mechanisms seem to be overwhelmed by genetic or acquired
defects. This concept is underscored by the finding that
unregulated control of B-cell activation or proliferation due to
the deficiency of the inhibitory Fcc receptor (FccRIIB), the protein
phosphatase Shp1, or the protein kinase Cd causes autoimmune
diseases [10–12]. As a consequence, B-cell-targeting therapies
have become one of the most effective treatments for autoimmue
Although enhanced growth of B cells coupled with enlarged GC
has been observed in AID2/2humans and mice, the potential
contribution of AID to autoimmunity remains largely unknown.
To better define the role of AID in autoimmunity, we carefully
analyzed AID2/2mice at different ages. We found that aged mice
spontaneously developed gastritis with pathological features
similar to human type A gastritis and murine experimental
PLoS ONE | www.plosone.org1August 2008 | Volume 3 | Issue 8 | e3033
autoimmune gastritis. The disease could be reconstituted in nu/nu
mice by adoptive transfer of CD4+T cells isolated from inflamed
gastric tissue. Anti-gastric-mucosa-specific Ab were elevated in
gastritis-positive, but not in gastritis-negative AID2/2mice,
indicating that autoreactive B cells are propagated in the
gastritis-positive AID2/2mice. Importantly, gastritis development
was preceded by formation of tertiary lymphoid organs (TLO).
Self-reactive B cells, likely generated because of increased BCR
diversity in the absence of AID, in TLOs could capture large
amounts of gastric autoantigens to present to T cells, and
reciprocally receive T-cell help. Consequently, such a vicious
cycle through a B-T collaboration could facilitate the generation of
gastric antigen-specific CD4 T cells, which ultimately cause the
pathology of autoimmune gastritis. The data presented here show
that AID is necessary not only for B cell homeostasis but also for
negative regulation of autoimmune disorders.
Lymphoid neogenesis is frequently observed in AID2/2
We observed that AID2/2
pathogen-free (SPF) conditions developed a number of polyp-like
lesions in the gastric mucosa (Figure 1A, see arrows). Histological
examination revealed that these lesions resulted from the
mice housed under specific
formation of TLOs (Figure 1B, central panels). Multifocal
lymphoid structures developed in the stomachs of 70% of
AID2/2mice at 5–8 months (Figure 1C). Similar lymphoid
neogenesis was detected in other organs such as lung, salivary
gland, liver and pancreas in AID2/2mice (Figure 2), albeit at
lower frequencies. Analogous to secondary lymphoid tissues,
gastric TLOs were composed of segregated B-cell and T-cell
areas (Figure 3, upper panels). Furthermore, GCs, characterized
by CXCL13 and CR-1 expression, were observed in the B-cell
follicles of TLOs (Figure S1A, left column). Of note is that
IgMhiB220loplasma cells were scattered in the lamina propria
surrounding the TLOs (Figure S1A, right column). H&E staining
of gastric TLOs further confirmed the presence of plasma cells
with the characteristic appearance of nuclei and abundant
cytoplasm in the circumference of the TLOs (Figure S1B, see
arrows). These observations suggest the possibility that B cells
activated in the gastric TLO would differentiate and directly
migrate from the lymphoid follicles to the gastric lamina propria
without passing through the systemic circulation.
Aged AID2/2mice spontaneously develop gastritis
A reduction of parietal cells was evident around the circumfer-
ence of TLOs (Figure 1B, see arrowheads). This pathological
change was remarkably increased in older mice. The stomach of
AID2/2mice around 12 months old was greatly enlarged
Figure 1. AID2/2mice spontaneously developed gastritis. (A) Stomachs of AID+/+and AID2/2mice were cut longitudinally to expose the
gastric mucosae. Inside-out tissues were examined by stereomicroscopy. A high magnification view of the mucosal surfaces revealed the presence of
a number of nodule-like structures in the AID2/2mouse stomach (arrows). A representative sample from each group is shown. (B) Gastric tissue
sections were stained with H&E for histological examination, or with Alcian blue-hematoxylin (AB&H) for detection of mucin-producing cells (shown
as turquoise). Scale bars: 100 mm. (C) Gastric tissues of AID2/2mice at different ages were analyzed for development of ectopic lymphoid follicles and
mucosal hyperplasia. The diagnostic criteria are described in Supplemental Table S1 online. (D) Stomach weights of AID+/+and AID2/2mice were
measured and the values were normalized to the body weight of each mouse. (E) Kaplan-Meier survival curves of AID2/2(solid line, n=18) and AID+/+
(dotted line, n=19) mice revealed significant differences in survival (P,0.05).
AID Prevents Autoimmunity
PLoS ONE | www.plosone.org2 August 2008 | Volume 3 | Issue 8 | e3033
(Figure 1A and D). Histological analysis of the enlarged stomachs
revealed epithelial hyperplasia and destruction of gastric glands
(Figure 1B and C) coupled with an increase in mucus-producing
cells (Figure 1B, lower panels). Vestigial TLOs were frequently
associated with the hyperplastic gastric lesions (Figure S1C). A
massive infiltration of inflammatory cells was also detected in the
lamina propria of the gastric tissue of AID2/2mice. Immunoflu-
orescent staining revealed that these infiltrating cells were mainly
Mac1/CD11b+cells (Figure 3, lower panels). Other subsets of
immune cells, namely, B cells, CD4+and CD8+T cells, and
CD11c+cells were also found scattered in the inflamed tissue. In
contrast, neither age-matched AID+/+(Figure 1A) nor AID+/2
mice (data not shown) developed TLOs or gastritis, although
minimal lymphocyte infiltration was detected in the gastric tissues
of occasional mice in these control groups (Figure S2, and data not
shown). Perhaps associated with increasingly severe gastritis
development, more than 20% of AID2/2mice died by 12
months of age, and half of them died by 18 months, whereas
almost all AID+/+mice survived for more than 18 months
(Figure 1E, P,0.05).
Cellular phenotype of gastric mucosa-infiltrating
lymphocytes in AID2/2mice
To further characterize the immunopathology of the gastritis, we
first performed FACS analysis of gastric mucosa-infiltrating
lymphocytes in AID2/2mice. As shown in Figure 4A, gastric B
and T cells highly upregulated the activation marker CD69, in
comparison to splenic lymphocytes. The percentage of GC B cells,
characterized by Fas expression and peanut agglutinin (PNA)
binding, was much higher in the stomach (21.268.6%, n=5) than
in the spleen (2.260.9%, n=3; P,0.01) (Figure 4A). These data
indicate that active immune responses take place in the gastric
tissues of AID2/2mice. Additionally, most gastric mucosa-
infiltrating CD4+T cells had a CD44hiCD62Lloeffector/memory
phenotype (Figure 4B, left panels). The ratio of FoxP3+Treg cells to
total CD4+T cells was significantly lower in gastritis lesions
(12.661.7%, n=5) than in the spleen (19.063.8%, n=3; P,0.01).
Interestingly, the surface expression level of CD25 in gastric
FoxP3+CD4+T cells was somewhat lower than that on splenic
FoxP3+CD4+T cells (Figure 4B, right panels), and nearly half of
FoxP3+CD4+T cell were even negative for CD25 expression.
Figure 2. TLOs are formed in various organs of AID2/2mice. The tissue samples were obtained from 12-month-old AID2/2or AID+/+mice,
and were subjected to histological examination after H–E staining. Asterisks represent TLOs. Scale bars: 100 mm.
Figure 3. Massive infiltration of inflammatory cells is observed in the gastric tissue of AID2/2mice. Immunofluorescence staining for
immune cell-specific markers was performed using serial tissue sections prepared from two different gastric samples with ectopic lymphoid follicles
(upper) or mucosal hyperplasia (lower). DAPI staining (left panels) was used to identify cell nuclei. Scale bars: 100 mm.
AID Prevents Autoimmunity
PLoS ONE | www.plosone.org3August 2008 | Volume 3 | Issue 8 | e3033
Proinflammatory cytokines are upregulated in the
stomach of AID2/2mice
We subsequently measured cytokine expression profiles in the
gastric tissue. Quantitative real-time PCR (Q-PCR) analysis
demonstrated that proinflammatory cytokines TNF-a and IL-12/
23p40 were highly upregulated in gastritis-positive AID2/2mice,
but not in AID+/+or gastritis-negative AID2/2mice (Figure 5).
Additionally, Th1 (IFN-c), Th2 (IL-4) and Th17 (IL-17) cytokines
were significantly upregulated in the inflamed gastric tissue.
Likewise, intercellular FACS staining and ELISA data confirmed
that a substantial number of IL-4-, IFN-c- and IL-17-producing
CD4+T cells were present in gastric-mucosa-infiltrating cells
(Figure 4C and D).
The gastrointestinal microflora are irrelevant to the
etiology of lymphoid neogenesis and gastritis in AID2/2
The data obtained so far indicate that AID2/2
spontaneously develop lymphoid neogenesis followed by gastritis
Figure 4. Gastric mucosa-infiltrating lymphocytes in AID2/2mice have an activated phenotype. (A) Lymphocytes were prepared from
gastric mucosae (GM; upper panels) of AID2/2mice with gastritis. Since lymphocyte infiltration into gastric tissues was negligible in AID+/+mice
(Figure 1B), splenocytes (SP; lower panels) from AID2/2mice with gastritis were used as a control. (B) CD3+CD4+gastric and splenic lymphocytes were
gated and further analyzed for memory and regulatory T-cell markers. (C) Gastric CD3+CD4+cells enriched with a MACS column were treated with
PMA and Golgi-stop for 6 h and then analyzed for production of the indicated cytokines by intracellular staining. (D) Concentrations of the indicated
cytokines in culture supernatants of gastric CD4+cells stimulated with anti-CD3 and anti-CD28 mAbs for 48 h were determined by CBA and ELISA.
Figure 5. Q-PCR analysis of cytokine mRNA expression in
gastric tissue. Gastric tissue samples were obtained from 12-month-
old AID+/+mice and AID2/2mice with or without gastritis symptoms.
Data are means6S.D. of at least four different samples. *P,0.05;
AID Prevents Autoimmunity
PLoS ONE | www.plosone.org4 August 2008 | Volume 3 | Issue 8 | e3033
with mucosal hyperplasia. However, it is still unknown how AID
deficiency leads to the development of gastritis. Fagarasan et al.
previously reported that AID deficiency leads to aberrant
expansion of anaerobic bacteria in the gut, which induces
hypertrophy of Peyer’s patches and ILFs as well as activation of
the systemic immune system . This observation raised the
possibility that activation of AID2/2B cells by gastric microflora
might elicit lymphoid neogenesis followed by gastric tissue
destruction. Similar pathological mechanisms have been reported
in Helicobacter pylori-induced gastritis and MALT lymphoma,
diseases in which the eradication of H. pylori cured most cases
[14,15]. Therefore, to explore a role of gastric microflora in
gastritis development, we maintained AID2/2mice in either GF
or conventional (CV) conditions, and examined gastritis develop-
ment. If an unregulated response of AID2/2B cells to gastric
microflora was involved in the pathogenesis of gastritis, the GF
environment should prevent gastritis development and, conversely,
CV conditions might worsen the disease compared to that seen
under SPF conditions.
The hypertrophic protrusion of ILFs was not observed in
AID2/2mice maintained at GF condition (data not shown), as
expected by the previous report with antibiotic cocktail treatment
. By contrast, even under GF conditions, AID2/2mice
developed TLOs at 5 months of age, and subsequently gastritis
lesions with epithelial hyperplasia and massive inflammatory cell
infiltration at 12 months (Figure 6A). There was no difference in
stomach weight, pathological score, or cytokine expression
between AID2/2mice housed under GF and CV conditions
(Figure 6B and C), suggesting that the gastric microflora are
irrelevant to the pathogenesis of gastritis.
Gastric Ag-specific autoreactive B cells are generated in
It has been reported that AID2/2B cells are constitutively
activated without immunization, and that their germline-encoded
BCR diversity is dramatically increased compared to wild-type
mice . We therefore hypothesized that such an aberrant
diversification of the B-cell repertoire might produce abundant
autoreactive B cells that eventually would overwhelm peripheral
tolerance mechanisms. To test this hypothesis, we measured serum
levels of total and gastric mucosa-specific IgM in AID2/2and
AID+/+mice. As expected , the total IgM levels were higher in
AID2/2mice than in wild type mice whereas they were
comparable between gastritis-negative and -positive AID2/2mice
(Figure 7A). By contrast, the serum levels of gastric mucosa-specific
Ab were significantly higher in gastritis-positive than in gastritis-
negative AID2/2mice. Moreover, there was a positive correlation
between serum autoantibody levels and pathological gastritis
scores in AID2/2mice (Figure 7B). H+/K+ATPase could be one
of the self antigens involved in this response, since the serum levels
of anti-H+/K+ATPase antibodies tended to be elevated in
gastritis-positive AID2/2mice (Figure 7A). These observations
confirmed that B-cell clones specific for gastric autoantigens are
generated in AID2/2mice.
Organ-specific autoreactive CD4+T cells act as terminal
effectors in the development of gastritis
The data obtained by FACS analysis revealed that gastric
mucosa-infiltrating CD4+T cells were activated and produced
large amounts of multiple cytokines (Figure 4C). These data
Figure 6. Gastric microflora are irrelevant to gastritis development in AID2/2mice. AID2/2and AID+/+mice were raised under germ-free
(GF) or conventional (CV) conditions to examine whether the presence of gastric microflora affects the development of gastritis. (A) Histological
examination of gastric tissues from AID2/2and AID+/+mice maintained under GF condition. (B) Stomach weight and pathological score of gastritis
were compared between AID2/2mice in GF and CV conditions. (C) Cytokine expression levels in gastric tissues of AID2/2and AID+/+mice under GF
and CV conditions were measured by Q-PCR. The mRNA expression level of each cytokine was normalized to that of GAPDH.
AID Prevents Autoimmunity
PLoS ONE | www.plosone.org5 August 2008 | Volume 3 | Issue 8 | e3033
suggest that CD4+T cells may participate in gastritis development
in AID2/2mice. To test this possibility, we isolated gastric
mucosa-infiltrating CD4+and CD8+T cells from AID2/2mice
and adoptively transferred them to BALB/cnu/numice that lack T
cells. Three months later, mice that received CD4+, but not CD8+,
T cells developed gastritis characterized by a massive infiltration of
Mac1+cells, epithelial hyperplasia and reduction of gastric glands
(Figure 8A–D). This phenotype was quite similar to that observed
in the late stages of gastritis in AID2/2mice (Figure 1B and
Figure 3, lower panels), except for the absence of TLO.
Furthermore, gastric CD4+T cells recovered from BALB/cnu/nu
mice produced significant quantities of IL-17 or IFN-c but not IL-
4 (Figure 8E and F), suggesting that Th1 and/or Th17 cells
function as terminal effectors of gastritis development in AID2/2
In the present study we have shown that as they age, AID2/2
mice sequentially develop lymphoid neogenesis followed by severe
gastritis. Consequently, AID2/2mice had a significantly shorter
lifespan compared to wild type mice. Although AID deficiency was
reported to lead to aberrant expansion of anaerobic bacteria in the
gut, we found that gastritis development in AID2/2mice results
from an autoimmune reaction rather than hyper responsiveness to
gastric microflora. This view is supported by several observations.
First, even under GF conditions, both lymphoid neogenesis and
gastritis occurred in the AID2/2mice , whereas hypertrophy of
ILFs was completely prevented (Hase K., unpublished observa-
tion). This result clearly suggests that lymphoid neogenesis in the
stomach and hypertrophy of ILFs in the small intestine are
induced by the different mechanisms in AID2/2mice, although
they share a similar appearance. Second, gastric mucosa-specific
autoantibody was detected in gastritis-positive AID2/2mice, and
there was a positive correlation between serum autoantibody levels
and pathological gastritis scores. Detection of autoantibody
indicates the activation and differentiation of gastric antigen-
specific B cell clones. Third, CD4+effector T cells, which are likely
autoreactive T cells typically observed in organ-specific autoim-
mune diseases, are generated in AID2/2mice. These results imply
an important role of AID for the containment of autoimmunity.
The gastritis that developed in aged AID2/2mice was
characterized by a massive infiltration of Mac1/CD11b+cells
and the destruction of parietal cells. The gastric glands in these
mice were replaced by mucous-producing cells, resulting in
epithelial hyperplasia. All of these pathological features are
reminiscent of human type A gastritis  and its experimental
model, murine autoimmune gastritis, which are well-characterized
organ-specific autoimmune diseases. Experimental autoimmune
gastritis can be induced in BALB/c mice subjected to thymectomy
3 days after birth [17,18]. Approximately 60% of neonatally
thymecomized mice develop gastritis within 12 weeks. In this
model, CD4+T cells but not CD8+T cells prepared from the
gastric mucosa can transfer the disease to nu/nu mice, indicating
an essential role of CD4+T cell in disease progression. Similarly,
the gastritis that spontaneously develops in AID2/2mice at later
stages could be reconstituted in nu/nu mice by adoptive transfer of
CD4+T cells. This observation further confirms that AID2/2
mice spontaneously develop autoimmue gastritis through gener-
ation of gastric antigen-specific autoreactive T cells.
Of prime importance, lymphoid neogenesis precedes gastritis
development in AID2/2mice (Figure 1A–C). Furthermore,
indications of gastritis lesions, such as reduction in number of
parietal cells, were more obvious at the circumference of TLOs
both at early (Figure 1B, see arrowheads) and late stages (Figure
S1C). These observations indicate that gastritis occurs secondary
to TLO formation in AID2/2
neogenesis has been detected in various organ-specific autoim-
mune disorders, including Hashimoto’s thyroiditis, rheumatoid
arthritis, type 1 diabetes and autoimmune gastritis [19–24], the
mechanisms leading to lymphoid neogenesis as well as its
immunopathological roles in autoimmune diseases are not fully
elucidated. Given that TLO are apparently equipped with the
complete set of lymphoid structures, i.e., secondary B-cell follicles
with GC and distinct T-cell areas containing dendritic cells and
high endothelial venules , TLO may play a role in maintaining
an autoimmune response against persistent self antigens. This
concept is at least party supported by experimental studies using
the rat insulin promoter-lymphocytic choriomeningitis virus
glycoprotein (RIP-GP) transgenic mouse model, where a positive
correlation between lymphoid neogenesis and development of
autoimmune diabetes was reported . Furthermore, expression
of the recombination-activating genes RAG1 and RAG2 is evident
in ectopic GC formed in rheumatoid arthritis and autoimmune
thyroid diseases [19,26,27]. RAG genes, which are normally
expressed in primary lymphoid organs, are master regulators of
the rearrangement of V(D) J gene segments. The induction of
RAG genes in TLOs raises the possibility that secondary BCR
mice. Although lymphoid
Figure 7. Gastric mucosa-specific antibodies are elevated in AID2/2mice. (A) Serum samples were collected from AID+/+and AID2/2mice
maintained under SPF condition to measure total IgM, and gastric-mucosa- or H+/K+-ATPase-specific IgM levels by ELISA. After histological
examination, individuals with a total pathological score of 2 or more were considered to be gastritis-positive. (B) Serum levels of anti-gastric mucosa
antibodies positively correlated with total pathological score in AID2/2mice (r=0.6825; P,0.05).
AID Prevents Autoimmunity
PLoS ONE | www.plosone.org6August 2008 | Volume 3 | Issue 8 | e3033
rearrangements, termed receptor revision, occur at these sites and
may contribute to the generation of autoreactive B cells. In
parallel, a recent study indicated that B-cell specificities in
pancreatic TLO of NOD mice are skewed to enhance autoanti-
gen-binding capacity, suggesting that TLO promote selection of
autoreactive B cells . Once autoreactive B cells are selected
and enhanced in TLO, these cells could differentiate into
autoantibody-forming plasma cells (Figure S1A and B). Consistent
with this idea, there is a positive correlation between serum
autoantibody levels and numbers of TLO in several autoimmune
disorders [19,24]. However, in the case of AID2/2mice, B cells
can produce only low-affinity IgM due to the lack of SHM and
CSR. Therefore, gastric antigen-specific autoantibodies would be
expected to only minimally contribute to the pathogenesis of the
gastritis. Instead, the more important mechanism by which
autoreactive B cells promote inflammatory processes may be T-
cell activation through presentation of autoantigens and co-
stimulation. It has been suggested that autoreactive B cells have a
crucial role as APCs to capture and present autoantigens for
activating T cells, which in turn drive T-cell autoreactivity [28,29].
This scenario could indeed be the case in the pathogenesis of
autoimmune gastritis in AID2/2mice. Autoimmune B cells could
be selected and propagated in TLO formed in the stomach of
AID2/2mice in the early stages. Given that B cells in gastric TLO
Figure 8. CD4+T cells contribute to the development of gastritis in AID2/2mice. Gastric mucosa-infiltrating CD4+T cells or CD8+T cells
were isolated from 12 month old AID2/2mice and were adoptively transferred into BALB/cnu/numice. After 3 months, the stomachs of recipient mice
were removed and subjected to the following experiments. (A) Gastric tissue sections were stained with H&E for histological examination. (B)
Stomach weights were normalized to body weight of each mouse. (C) Gastric tissue sections were examined by histology and assigned a pathological
score. (D) Immunofluorescence staining revealed a massive infiltration of CD4+and Mac1+cells into the stomachs of BALB/cnu/numice that received
CD4+T cells by adoptive transfer. (E) Gastric CD3+CD4+cells enriched with a MACS column were treated with PMA and Golgi-stop for 6 hr and
analyzed for cytokine production by intracellular staining. (F) CD4+T cells recovered from gastric tissues of CD4+T-cell-transferred BALB/cnu/numice
were stimulated with anti-CD3 and anti-CD28 mAbs for 48 h. Cytokine concentrations in the culture supernatants were measured by CBA and ELISA.
Scale bars: 100 mm.
AID Prevents Autoimmunity
PLoS ONE | www.plosone.org7 August 2008 | Volume 3 | Issue 8 | e3033
are in a privileged location to capture large amount of
autoantigens and present them to T cells, this interaction could
contribute to the pathogenesis by induction of autoreactive CD4+
It should be noted that TLOs were induced not only in the
stomach but also in other organs of AID2/2mice (Figure 2).
Therefore, these organs are also considered to be potential targets
for an autoimmune response. Indeed, autoimmune diseases have
been found to develop in multiple organs in humans with AID
mutation . The precise mechanism by which AID deficiency
promotes the generation of self-reactive B cells is an open question
at this moment. A previous report revealed that the BCR diversity
is increased in AID2/2mice . Consequently, gastric Ag-specific
B cells might emerge from the expanded B-cell repertoire. Under
physiological conditions, B-cell autoimmunity is counteracted by
several tolerance mechanisms including receptor editing, anergy,
and clonal deletion. Particularly, as many as 50% of newly
produced B cells are reported to show an anergic phenotype
(CD93+B220+IgMlo), suggesting the primary importance of anergy
in peripheral tolerance . Although anergic B cells have a
shorter lifespan relative to wild-type mature B cells , anergy is
a reversible process: anergic B cells have been shown to revert to
naı ¨ve B cells upon hapten stimulation . This observation raises
the possibility that escape from anergy could result in autoimmu-
nity. Mice and humans lacking functional AID display B-cell
hyperplasia and a hyperactivated immune system [1–3]. Based on
these observations, we speculate that the aberrant activation stress
may drive B-cell clones, which are normally retained in an anergic
state, toward an autoimmune response. Intercrossing of AID2/2
mice with transgenic mice expressing a BCR prone to anergy
should help clarify this speculation.
Several mechanisms have been proposed to explain why AID
deficiency leads to dysregulated proliferation of B cells. One
possibility is that AID2/2B cells are devoid of inhibitory signals
though FccRIIB due to the lack of IgG Ab production. FccRIIB is
the only FccR that contains a cytoplasmic ITIM motif. Cross-
linking of the BCR and FccRIIB by IgG antibody-antigen
complexes leads to negative feedback regulation for B cell
activation via the phosphorylation of ITIM tyrosines and
subsequent recruitment of the inositol phosphatase SHIP to the
plasma membrane. Therefore, FccRIIB2/2mice display elevated
Ig levels and enhanced immune complex-mediated tissue injury in
response to Ag challenge . Furthermore, involvement of this
inhibitory receptor in peripheral tolerance has been supported by
the fact that FccRIIB2/2mice on a C57BL/6 background
spontaneously develop autoantibody-mediated lupus glomerulo-
nephritis . Although significantly different types of autoimmue
diseases, namely systemic versus organ-specific, are induced in
FccRIIB2/2and in AID2/2mice, respectively, this phenotypic
difference is most likely due to the inability of AID2/2mice to
produce high-affinity autoimmune IgG Abs. AID-/-mice produce
abnormally large amounts of IgM, whereas IgG is absent. IgM,
like IgG, can promote complement activation, which in turn could
support B-cell activation, because opsonization of Ags by
complement components remarkably enhances both B cell
activation (Carter and Fearon) as well as Ag uptake by B cells
. However, since IgM is unable to bind FccRIIB, it fails to elicit
negative feedback of humoral immunity. Thus, the imbalance of
IgM and IgG in AID2/2mice and humans could lead to
unregulated B-cell proliferation, leading to the loss of peripheral
tolerance. Other mechanisms, such as lack of high-affinity anti-
idiotypic antibodies that might maintain B-cell homeostasis should
also be considerable. Alternatively, insufficient elimination of
external Ags due to the lack of class-switched and hypermutated
Igs might drive B-cell activation. Elucidation of the precise
mechanisms for unregulated B-cell proliferation as a result of AID
deficiency is an important unresolved issue.
Materials and Methods
Monoclonal Abs against B220 (RA3-6B2), CD3e (145-2C11),
CD8a (53-6.7), FoxP3 (FJK-16s) (eBioscience); CD4 (GK1.5),
CD11b/Mac1 (M1/70), CD11c (HL3), CD16/CD32 (93), CD19
(1D3), CD25 (7D4), CD44 (IM7), CD62L (MEL14), CD69
18H10.1) (BD Biosciences), IFN-c (37895), IL-4 (30340) (R&D
Systems) and polyclonal Abs against IgM (Southern Biotech) and
CXCL13 (R&D Systems) were used. PNA was purchased from
Vector Laboratories. Cy3-conjugated anti-rat IgG and Alexa488-
conjugated anti-goat IgG secondary antibodies were purchased
from Jackson Immunolaboratories and Invitrogen, respectively.
CR1 (8C12), IL-17(TC11-
AID2/2mice bred on a BALB/c background were kindly
provided by Dr. T. Honjo , and maintained under SPF
condition in RIKEN animal facilities. In certain experiments, mice
were maintained under GF or CV condition at the animal facility
of the University of Tokyo as described previously [34,35]. To
examine the role of T cells in gastritis development, 2.5–46105
CD3+CD4+or CD3+CD8+cells were isolated from the stomachs
of 12-month-old AID2/2mice with FACS Vantage SE (BD
Biosciences), and were adoptively transferred to 2-month-old
BALB/cnu/numice (Clea Japan). Three months later, the stomachs
of the recipient mice were excised for histological analysis. All
animal experiments were approved by the Animal Research
Committee of RIKEN Yokohama Research Institute and of the
University of Tokyo.
Histological analysis and immunofluorescence staining
For histological examination, 4% formalin-fixed, paraffin-
embedded tissue sections were deparaffinized, rehydrated and
stained with either hematoxylin-eosin or Alcian blue-hematoxylin.
The specimens were analyzed for gastritis based on the criteria
shown in Supplemental Table S1. Gastric-mucosa-infiltrating cells
were analyzed by immunofluorescence staining. Frozen sections
(5 mm) of gastric tissue were fixed with 4% Cytofix/Cytoperm (BD
Pharmingen) and incubated with blocking buffer (Perkin Elmer)
for 30 min at room temperature, followed by primary antibodies
or isotype-matched control IgG. The binding of these antibodies
was probed by fluorescent dye-conjugated secondary antibodies.
Flow cytometric analysis
To flush out infiltrating cells, a 5% fetal bovine serum/
phosphate-buffered saline (PBS) solution was injected into the
gastric submucosa. The remaining gastric tissue was minced and
subjected to collagenase digestion as described previously .
These treatments usually gave rise to a total of 1–66107gastric
infiltrating cells per mouse, dependent on the severity of
inflammation. Splenocytes were prepared as described . After
Fc blocking, the samples were stained with fluorescent-dye- (FITC,
PE, PECy7 or APC) or biotin-conjugated primary antibodies,
followed by fluorescent-dye-conjugated streptavidin. Intracellular
staining was performed with Cytofix/Cytoperm (BD) according to
the manufacturer’s instructions. The stained cells were analyzed
with a FACS Calibur (BD).
AID Prevents Autoimmunity
PLoS ONE | www.plosone.org8August 2008 | Volume 3 | Issue 8 | e3033
Total RNA was extracted from gastric tissue using an RNeasy
kit (QIAGEN), and was reverse-transcribed using ReverTra Ace-a
(Toyobo). Q-PCR was performed to quantify cytokine mRNA
expression levels using the SYBRH-Green PCR assay and the
Thermal Cycler Dice Real Time System (Takara) essentially as
described previously. Protein concentrations of IFN-c and IL-
4 in culture supernatant of gastric CD4+T cells were measured
with a Cytometric Bead Array (CBA) kit (BD), and that of IL-17
was detected by an ELISA kit (R&D), following the manufacturers’
Detection of autoantibodies
Gastric epithelia were obtained from AID+/+mice maintained
under GF condition, as described previously. The prepared
epithelia were homogenized in lysis buffer [150 mM NaCl,
20 mM Tris (pH 7.5), 1 mM EDTA, 0.1% Triton X-100]
containing a protease inhibitor cocktail (Roche). The homogenate
was centrifuged at 12,000 rpm at 4uC for 20 min to obtain soluble
protein supernatants. To detect autoAb, 5 mg/ml of the gastric
epithelial extract or porcine H+/K+ATPase (Arotec Diagnostics)
were immobilized on Maxisorp microtiter plates (Nunc). After
blocking with 1% bovine serum albumin (BSA)/PBS, serum
samples serially diluted with 1% BSA/1% casein/0.02% Tween
20/PBS [pH 7.4] were applied to the plate. Specific autoantibody
binding was probed with a horseradish-peroxidase-conjugated
anti-mouse IgM polyclonal Ab ( Jackson ImmunoResearch Lab.),
and was visualized with 3,39,5,59-tetramethylbenzidine and H2O2
Data for all figures were collected from three or more
independent experiments. Statistical significance of two groups
was determined using the student’s t-test. For comparison of three
groups, the data were subjected to analysis of variance (ANOVA),
followed by Fisher exact tests. The survival of mice of different
genotypes was analyzed using the Kaplan-Meier method.
Differences were considered significant at P,0.05.
Found at: doi:10.1371/journal.pone.0003033.s001 (0.12 MB PPT)
AID2/2mice at 6 (A-B) or 12 (C) months of age. (A)
Immunofluorescent staining for cell surface markers was per-
formed using gastric samples from AID2/2mice that contain
TLOs. (B) High magnification of a gastric tissue section stained
with H&E demonstrating migration of plasma cells (arrows) into
the lamina propria. (C) TLOs were frequently coupled with type-A
gastritis lesions in aged AID2/2mice. Scale bars: 100 mm.
Found at: doi:10.1371/journal.pone.0003033.s002 (12.91 MB
Analysis of TLOs observed in the gastric tissue of
gastric tissue of AID2/2and AID+/+mice at different ages was
analyzed by histology and scored following the criteria listed in
Supplemental Table 1.
Found at: doi:10.1371/journal.pone.0003033.s003 (0.19 MB PPT)
Total pathological score of gastritis development. The
We thank Dr. T. Honjo (Kyoto University) for providing AID2/2mice.
We also thank Drs. S. Fagarasan, K. Suzuki and T. Dohi (International
Medical Center of Japan) for valuable discussions, and Ms. Y. Fujimura
and R. Chihara for technical support.
Conceived and designed the experiments: KH. Performed the experiments:
KH DT ME SK KI. Analyzed the data: KH. Wrote the paper: KH HO.
1. Muramatsu M, Kinoshita K, Fagarasan S, Yamada S, Shinkai Y, et al. (2000)
Class switch recombination and hypermutation require activation-induced
cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102: 553–563.
2. Revy P, Muto T, Levy Y, Geissmann F, Plebani A, et al. (2000) Activation-
induced cytidine deaminase (AID) deficiency causes the autosomal recessive
form of the Hyper-IgM syndrome (HIGM2). Cell 102: 565–575.
3. Fagarasan S, Muramatsu M, Suzuki K, Nagaoka H, Hiai H, et al. (2002) Critical
roles of activation-induced cytidine deaminase in the homeostasis of gut flora.
Science 298: 1424–1427.
4. Suzuki K, Meek B, Doi Y, Muramatsu M, Chiba T, et al. (2004) Aberrant
expansion of segmented filamentous bacteria in IgA-deficient gut. Proc Natl
Acad Sci U S A 101: 1981–1986.
5. Quartier P, Bustamante J, Sanal O, Plebani A, Debre M, et al. (2004) Clinical,
immunologic and genetic analysis of 29 patients with autosomal recessive hyper-
IgM syndrome due to Activation-Induced Cytidine Deaminase deficiency. Clin
Immunol 110: 22–29.
6. Edwards JC, Cambridge G (2006) B-cell targeting in rheumatoid arthritis and
other autoimmune diseases. Nat Rev Immunol 6: 394–403.
7. Wardemann H, Yurasov S, Schaefer A, Young JW, Meffre E, et al. (2003)
Predominant autoantibody production by early human B cell precursors.
Science 301: 1374–1377.
8. Nemazee D (1996) Antigen receptor ‘capacity’ and the sensitivity of self-
tolerance. Immunol Today 17: 25–29.
9. Shlomchik MJ (2008) Sites and stages of autoreactive B cell activation and
regulation. Immunity 28: 18–28.
10. Bolland S, Ravetch JV (2000) Spontaneous autoimmune disease in
Fc(gamma)RIIB-deficient mice results from strain-specific epistasis. Immunity
11. Pao LI, Lam KP, Henderson JM, Kutok JL, Alimzhanov M, et al. (2007) B cell-
specific deletion of protein-tyrosine phosphatase Shp1 promotes B-1a cell
development and causes systemic autoimmunity. Immunity 27: 35–48.
12. Miyamoto A, Nakayama K, Imaki H, Hirose S, Jiang Y, et al. (2002) Increased
proliferation of B cells and auto-immunity in mice lacking protein kinase Cdelta.
Nature 416: 865–869.
13. Browning JL (2006) B cells move to centre stage: novel opportunities for
autoimmune disease treatment. Nat Rev Drug Discov 5: 564–576.
14. Isaacson PG, Du MQ (2004) MALT lymphoma: from morphology to molecules.
Nat Rev Cancer 4: 644–653.
15. Wotherspoon AC, Doglioni C, Diss TC, Pan L, Moschini A, et al. (1993)
Regression of primary low-grade B-cell gastric lymphoma of mucosa-associated
lymphoid tissue type after eradication of Helicobacter pylori. Lancet 342:
16. Strickland RG, Mackay IR (1973) A reappraisal of the nature and significance of
chronic atrophic gastritis. Am J Dig Dis 18: 426–440.
17. Kojima A, Taguchi O, Nishizuka Y (1980) Experimental production of possible
autoimmune castritis followed by macrocytic anemia in athymic nude mice. Lab
Invest 42: 387–395.
18. Fukuma K, Sakaguchi S, Kuribayashi K, Chen WL, Morishita R, et al. (1988)
Immunologic and clinical studies on murine experimental autoimmune gastritis
induced by neonatal thymectomy. Gastroenterology 94: 274–283.
19. Armengol MP, Juan M, Lucas-Martin A, Fernandez-Figueras MT,
Jaraquemada D, et al. (2001) Thyroid autoimmune disease: demonstration of
thyroid antigen-specific B cells and recombination-activating gene expression in
chemokine-containing active intrathyroidal germinal centers. Am J Pathol 159:
20. Kendall PL, Yu G, Woodward EJ, Thomas JW (2007) Tertiary lymphoid
structures in the pancreas promote selection of B lymphocytes in autoimmune
diabetes. J Immunol 178: 5643–5651.
21. Ludewig B, Odermatt B, Landmann S, Hengartner H, Zinkernagel RM (1998)
Dendritic cells induce autoimmune diabetes and maintain disease via de novo
formation of local lymphoid tissue. J Exp Med 188: 1493–1501.
22. Shi K, Hayashida K, Kaneko M, Hashimoto J, Tomita T, et al. (2001)
Lymphoid chemokine B cell-attracting chemokine-1 (CXCL13) is expressed in
germinal center of ectopic lymphoid follicles within the synovium of chronic
arthritis patients. J Immunol 166: 650–655.
23. Luzina IG, Atamas SP, Storrer CE, daSilva LC, Kelsoe G, et al. (2001)
Spontaneous formation of germinal centers in autoimmune mice. J Leukoc Biol
AID Prevents Autoimmunity
PLoS ONE | www.plosone.org9 August 2008 | Volume 3 | Issue 8 | e3033
24. Katakai T, Hara T, Sugai M, Gonda H, Shimizu A (2003) Th1-biased tertiary Download full-text
lymphoid tissue supported by CXC chemokine ligand 13-producing stromal
network in chronic lesions of autoimmune gastritis. J Immunol 171: 4359–4368.
25. Aloisi F, Pujol-Borrell R (2006) Lymphoid neogenesis in chronic inflammatory
diseases. Nat Rev Immunol 6: 205–217.
26. Itoh K, Meffre E, Albesiano E, Farber A, Dines D, et al. (2000) Immunoglobulin
heavy chain variable region gene replacement As a mechanism for receptor
revision in rheumatoid arthritis synovial tissue B lymphocytes. J Exp Med 192:
27. Zhang Z, Wu X, Limbaugh BH, Bridges SL, Jr. (2001) Expression of
recombination-activating genes and terminal deoxynucleotidyl transferase and
secondary rearrangement of immunoglobulin kappa light chains in rheumatoid
arthritis synovial tissue. Arthritis Rheum 44: 2275–2284.
28. Takemura S, Klimiuk PA, Braun A, Goronzy JJ, Weyand CM (2001) T cell
activation in rheumatoid synovium is B cell dependent. J Immunol 167:
29. Yan J, Harvey BP, Gee RJ, Shlomchik MJ, Mamula MJ (2006) B cells drive early
T cell autoimmunity in vivo prior to dendritic cell-mediated autoantigen
presentation. J Immunol 177: 4481–4487.
30. Merrell KT, Benschop RJ, Gauld SB, Aviszus K, Decote-Ricardo D, et al.
(2006) Identification of anergic B cells within a wild-type repertoire. Immunity
31. Fulcher DA, Basten A (1994) Reduced life span of anergic self-reactive B cells in
a double-transgenic model. J Exp Med 179: 125–134.
32. Benschop RJ, Aviszus K, Zhang X, Manser T, Cambier JC, et al. (2001)
Activation and anergy in bone marrow B cells of a novel immunoglobulin
transgenic mouse that is both hapten specific and autoreactive. Immunity 14:
33. Takai T, Ono M, Hikida M, Ohmori H, Ravetch JV (1996) Augmented
humoral and anaphylactic responses in Fc gamma RII-deficient mice. Nature
34. Narushima S, Itoh K, Kurama K, Uchida K (1999) Caecal bile acid
compositions in gnotobiotic mice associated with human intestinal bacteria
with the ability to transform bile acids in vitro. Microb Ecol Health Dis 11: 55.
35. Hase K, Murakami T, Takatsu H, Shimaoka T, Iimura M, et al. (2006) The
membrane-bound chemokine CXCL16 expressed on follicle-associated epithe-
lium and M cells mediates lympho-epithelial interaction in GALT. J Immunol
36. Murakami T, Chen X, Hase K, Sakamoto A, Nishigaki C, et al. (2007) Splenic
CD19-CD35+B220+ cells function as an inducer of follicular dendritic cell
network formation. Blood 110: 1215–1224.
37. Hase K, Ohshima S, Kawano K, Hashimoto N, Matsumoto K, et al. (2005)
Distinct gene expression profiles characterize cellular phenotypes of follicle-
associated epithelium and M cells. DNA Res 12: 127–137.
AID Prevents Autoimmunity
PLoS ONE | www.plosone.org10 August 2008 | Volume 3 | Issue 8 | e3033