Differential Contributions of APC Subsets to T Cell Activation in Nonobese Diabetic Mice

Abstract

Despite the pivotal role of dendritic cells (DC) in shaping immunity, little is known about their functionality in type 1 diabetes. Moreover, due to the paucity of DC in vivo, functional studies have relied largely upon in vitro-expanded cells to elucidate type 1 diabetes-associated functional abnormalities. In this study, we provide a comprehensive analysis of the functional capabilities of in vivo-derived DC subsets from NOD mice by comparing DC to other NOD APC types and to DC from autoimmune-resistant strains. NOD DC closely resemble those from nonautoimmune strains with respect to costimulation and cytokine production. The exception is the CD8alpha(+)CD11b(-)DC subset which is numerically reduced in NOD spleens, but not in the pancreatic lymph nodes, while DC from both tissues produce little IL-12 in this strain. This defect results in unusual deferral toward macrophage-derived IL-12 in NOD mice; NOD macrophages produce aberrantly high IL-12 levels that can overcompensate for the DC defect in Th1 polarization. APC subset use for autoantigen presentation also differs in NOD mice. NOD B cells overshadow DC at activating islet-reactive T cells, whereas DC and B cells in NOD-resistant mice are functionally comparable. Differential involvement of APC subsets in T cell activation and tolerance induction may prove to be a crucial factor in the selection and expansion of autoreactive T cells.

Full text

Differential Contributions of APC Subsets to T Cell Activation
in Nonobese Diabetic Mice
1
Annette M. Marleau,
2
Kelly L. Summers,
3
and Bhagirath Singh
4
Despite the pivotal role of dendritic cells (DC) in shaping immunity, little is known about their functionality in type 1 diabetes.
Moreover, due to the paucity of DC in vivo, functional studies have relied largely upon in vitro-expanded cells to elucidate type
1 diabetes-associated functional abnormalities. In this study, we provide a comprehensive analysis of the functional capabilities of
in vivo-derived DC subsets from NOD mice by comparing DC to other NOD APC types and to DC from autoimmune-resistant
strains. NOD DC closely resemble those from nonautoimmune strains with respect to costimulation and cytokine production. The
exception is the CD8
␣
ⴙ
CD11b
ⴚ
DC subset which is numerically reduced in NOD spleens, but not in the pancreatic lymph nodes,
while DC from both tissues produce little IL-12 in this strain. This defect results in unusual deferral toward macrophage-derived
IL-12 in NOD mice; NOD macrophages produce aberrantly high IL-12 levels that can overcompensate for the DC defect in Th1
polarization. APC subset use for autoantigen presentation also differs in NOD mice. NOD B cells overshadow DC at activating
islet-reactive T cells, whereas DC and B cells in NOD-resistant mice are functionally comparable. Differential involvement of APC
subsets in T cell activation and tolerance induction may prove to be a crucial factor in the selection and expansion of autoreactive
T cells. The Journal of Immunology, 2008, 180: 5235–5249.
The NOD mouse recapitulates many aspects of the patho-
genesis of type 1 diabetes (T1D)
5
in humans and is there-
fore widely applied in studies addressing the cellular
mechanisms of autoimmunity. APC, mainly dendritic cells (DC)
and macrophages, first appear at the islet periphery at ⬃4wkof
age and their appearance correlates with transient hyperinsulin-
emia and islet neogenesis (1, 2). Islet-infiltrated DC and macro-
phages produce TNF-
␣
which drives the proinflammatory re-
sponse during insulitis (3), leading to CD4
⫹
and CD8
⫹
T cell
recruitment and activation (4).
Numerous studies have investigated the role of APC, particu-
larly B cells and macrophages, in T1D pathogenesis in NOD mice.
Studies involving ablated B cell development or manipulated Ag-
presentation function have shown that B cells are essential for
disease (5–9). NOD B cells demonstrate efficient presentation of
islet Ag to T cells, which is attributable to Ag capture via Ig re-
ceptors (10). NOD B cells also exhibit a plethora of aberrant ac-
tivation characteristics, including resistance to tolerance induction
(11), increased NF-
␬
B activity (12), hyperproliferation, resistance
to apoptosis and enhanced costimulation (13, 14). Similarly, we
and others have demonstrated that NOD macrophages and bone
marrow (BM)-derived DC exhibit elevated IL-12 production and
NF-
␬
B hyperactivation (15–21). However, surprisingly little in-
formation is available concerning the functional capabilities of in
vivo-derived DC in NOD mice. Studies of NOD DC have mainly
used in vitro-generated cells, which are now recognized to be phe-
notypically and functionally distinct from the heterogeneous DC
subsets that exist in vivo (22). Specifically, the 7-day GM-CSF
plus IL-4 culture protocols for generating DC primarily expand a
myeloid population whose in vivo equivalent is not known.
In this study, we present a comprehensive functional analysis of
splenic CD8
␣
⫹
and CD8
␣
⫺
DC, the DC subsets which are prev-
alent throughout mouse lymphoid tissues, in NOD and autoim-
mune-resistant mice. In the spleen, CD8
␣
⫹
DC reside in the T cell
areas while CD8
␣
⫺
DC are localized in the marginal zones (23,
24). Both DC subsets are capable of priming naive T cells, al-
though this is accomplished using different cytokine pathways and
in response to distinct endogenous signals and microbial products
(25). This report identifies specific abnormalities in the numbers
and IL-12-producing capability of CD8
␣
⫹
DC in the NOD periph-
ery. These data also challenge several long-standing presumptions,
namely that in vitro- and in vivo-derived DC are functionally com-
parable and that an increased proclivity toward promoting type 1
cytokine responses is a generalized trait of NOD APC. Moreover,
by placing DC in the spectrum of NOD APC activity, we show
that macrophage-derived IL-12 and Ag presentation by B cells
override the functional capabilities of DC. Therefore, in terms
of mediating inappropriate T cell activation, DC appear to har-
bor minimal functional impairment in comparison to the other
NOD professional APC.
Materials and Methods
Mice
Female NOD/Lt mice were bred in the animal facility at the Robarts Re-
search Institute (London, Canada). Female C57BL/6, BALB/c, and NOD-
resistant (NOR) mice were purchased from The Jackson Laboratory. Mice
Department of Microbiology and Immunology and Robarts Research Institute, Uni-
versity of Western Ontario, London, Ontario, Canada
Received for publication July 20, 2007. Accepted for publication February 8, 2008.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
This work was supported by grants from the Canadian Institutes of Health Research.
A.M.M. was a recipient of a doctoral award from the Canadian Institutes of Health
Research.
2
Current address: Department of Immunology, The Scripps Research Institute, 10550
North Torrey Pines Road, La Jolla, CA 92037.
3
Current address: Lawson Health Research Institute, London, Ontario, Canada, N6A
4V2.
4
Address correspondence and reprint requests to Dr. Bhagirath Singh, Department of
Microbiology and Immunology, University of Western Ontario, London, Ontario,
Canada, N6A 5C1. E-mail address: bsingh@uwo.ca
5
Abbreviations used in this paper: T1D, type 1 diabetes; DC, dendritic cell; BM, bone
marrow; NOR, NOD resistant; PI, propidium iodide; PLN, pancreatic lymph node;
GAD, glutamic acid decarboxylase.
Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00
The Journal of Immunology
www.jimmunol.org

were maintained in the specific pathogen-free facility at the University of
Western Ontario (London, Canada). All experiments were performed in
accordance with institutional guidelines for animal care. Female NOD mice
were compared with age- and sex-matched diabetes-resistant control
strains. Unless otherwise specified, mice were used between 4 and 6 wk of
age and were therefore nondiabetic. Diabetes incidence was monitored by
weekly measurement of venous blood glucose concentrations in nonfasting
mice using Glucometer Elite strips (Bayer). Mice with two consecutive
blood glucose concentrations ⬎300 mg/dl were considered diabetic, which
typically occurred between 15 and 20 wk of age in our colony.
Culture medium, cytokines, TLR agonists
RPMI 1640 medium was supplemented with 2 mM L-glutamine, 0.5%
HEPES, 5
␮
g/ml penicillin, 100 U/ml streptomycin (Invitrogen Life Tech-
nologies) and 10% (v/v) FCS (HyClone Laboratories). Murine cytokines
(GM-CSF, IL-4, IFN-
␥
, and TNF-
␣
) were purchased from Cedarlane Lab-
oratories and reconstituted in sterile water. LPS (Escherichia coli serotype
055:B5) was obtained from Sigma-Aldrich. Polyinosine-polycytidylic acid
(poly(I:C)) was purchased from Sigma-Aldrich and reconstituted in
sterile PBS.
Isolation of splenic DC, macrophages, and B cells from spleens
Mice were euthanized by CO
2
narcosis and spleens were harvested. Pooled
spleens were cut into small fragments and digested for 30 min at room
temperature with gentle and continuous agitation in RPMI 1640 containing
1 mg/ml collagenase A (Roche Diagnostics) and 40
␮
g/ml DNaseI (Roche
Diagnostics). Spleen fragments were intermittently resuspended by gentle
pipetting. Cell suspensions were filtered through a sterile nylon mesh to
remove undigested material and were subsequently treated for 5 min with
PBS supplemented with 5% FCS and 5 mM EDTA (pH 7.2) to disrupt
DC/T cell complexes. Following another washing step in medium, RBC
were removed using ACK lysing buffer (BioWhittaker). For labeling and
purification of DC, the medium consisted of PBS containing 2 mM EDTA
and 0.5% BSA. FcR were blocked for 15 min with anti-mouse CD16/CD32
(2.4G2; BD Pharmingen) at 4°C. Cells were labeled with MACS CD11c
(N418) Microbeads (Miltenyi Biotec) for 15 min at 4°C. Positive selection
for CD11c
⫹
cells was done using a MidiMACS Separator and MACS
columns (Miltenyi Biotec) according to the manufacturer’s instructions.
The resulting purity of DC was consistently ⬎90% as verified by flow
cytometry.
To purify DC subsets, a two-step method was used. DC were first pre-
enriched from bulk splenocytes by negative selection using a dendritic cell
isolation kit (Miltenyi Biotec). Briefly, non-DC were labeled with an Ab
mixtureandremovedbyMACSseparation.CD11c
high
CD8
␣
⫹
andCD11c
high
CD8
␣
⫺
DC were sorted using a FACSVantage (BD Biosciences) and cell
purity was verified to be at least 97%.
To purify splenic macrophages, CD4 (GK1.5), CD8 (HO2.2), CD11c
(HL3), and B220 (RA3-3A1/6.1)-expressing cells were depleted from the
cell suspensions by complement-mediated lysis. Macrophages were then
isolated by adherence to plastic. Flow cytometry was done to verify the
purity of the cells, which were routinely ⬎92% CD11b
⫹
and ⬎65%
F4/80
⫹
.
To purify B cells, splenocytes were labeled with mouse CD19 mi-
crobeads or CD45R/B220 microbeads and positive selection was per-
formed using MACS separation columns according to the manufacturer’s
instructions (Miltenyi Biotec), which routinely yielded a population that
was ⬎80% positive for B cell surface markers.
Maturation of DC and macrophages in vitro
Splenocytes or purified APC populations were matured in vitro with com-
binations of proinflammatory cytokines and TLR agonists. For DC matu-
ration, cells were treated with LPS (100 ng/ml) plus TNF-
␣
(10 ng/ml),
GM-CSF (10 ng/ml) plus IL-4 (10 ng/ml) plus IFN-
␥
(20 ng/ml) ⫾
poly(I:C) (100 ng/ml, unless otherwise indicated) or LPS (100 ng/ml) plus
irradiated CD40L-transfected J558 cells (a gift from P. Lane, University of
Birmingham, Birmingham, U.K.). J558 cells were gamma-irradiated before
use with 2500 rad from a Cobalt 60 source. Preliminary experiments were
conducted to determine that 1 DC:1 J558 cell was an optimal ratio for
induction of IL-12p70 synthesis by DC. Macrophages were matured by
treatment with LPS (1
␮
g/ml). After 24 h of culture, unless otherwise
indicated, cells were pelleted and prepared for flow cytometry. Supernatant
samples were stored at ⫺70°C to test cytokine concentrations by ELISA.
In vivo DC activation
Four-week-old NOD and NOR mice (n⫽3/group) were injected i.v. with
PBS (vehicle control), LPS (30
␮
g/mouse), or poly(I:C) (100
␮
g/mouse).
At 6 h postinjection, splenocytes were pooled from each treatment group
and CD11c
⫹
cells were selected using the MACS separation technique. DC
(5 ⫻10
5
/ml) were plated in medium without exogenous cytokines for 30 h
and supernatant samples were collected for analysis of the IL-12p70 con-
tent by ELISA.
Flow cytometry
All staining steps were performed at 4°C in PBS. FcR were blocked with
anti-CD16/32 Ab (clone 2.4G2; BD Pharmingen) and cells were subse-
quently incubated with 0.5
␮
g/10
6
cells of the relevant Ab for 45 min. The
following anti-mouse Abs were purchased from BD Pharmingen: anti-
CD11c (HL3), anti-CD80 (16-10A1), anti-CD86 (GL1), anti-CD40
(HM40-3), anti-CD4 (L3T4), anti-CD8
␣
(53-6.7), anti-CD11b (M1/70),
anti-B220 (RA3-6B2), and anti-I-A
k
(10-3.6). Anti-mouse F4/80 Ab was
purchased from Serotec (CI:A3-1). Isotype-matched control Abs were pur-
chased from BD Pharmingen and Cedarlane Laboratories. Samples were
analyzed on a FACSCalibur flow cytometer (BD Biosciences) using
CellQuest software (BD Biosciences). Live cells were selected by forward/
side scatter gating.
Flow cytometry to identify dead and apoptotic cells was performed by
staining splenocytes with Abs against DC subset markers (CD11c and
CD8
␣
) in combination with annexin V and propidium iodide (PI) using a
kit from BD Pharmingen. Apoptotic DC were identified as annexin V-
positive, PI-negative cells.
BrdU labeling and analysis
Groups of three or four mice (NOD and C57BL/6) were given BrdU (0.8
mg/ml; Sigma-Aldrich) in sterile drinking water that was changed daily.
After 2 or 5 days, splenocytes were stained for cell surface molecules and
were subsequently stained to detect BrdU incorporation by flow cytometry
using a kit from BD Pharmingen.
Intracellular cytokine staining
Pancreatic lymph nodes (PLN) were excised from 6-wk-old NOD and
NOR mice and were pressed through a sterile nylon steel mesh to obtain
cell suspensions. Flow cytometry was performed using freshly isolated
PLN-derived cells to identify CD11c
high
CD11b
high
and CD11c
high
CD11b
low/⫺
populations. Cells obtained were cultured (2 ⫻10
6
/ml) with
GM-CSF (10 ng/ml) plus IL-4 (10 ng/ml) plus IFN-
␥
(20 ng/ml) plus
poly(I:C) (100 ng/ml) at 37°C for 10 h. Brefeldin A (5
␮
g/ml) was added
forthelast5hofculture to prevent cytokine secretion by blocking intra-
cellular transport processes. Surface staining for CD11c and CD11b was
conducted followed by fixation and permeabilization using a commercially
available kit (Cedarlane Laboratories) and staining for IL-12p40/p70 (BD
Pharmingen) for flow cytometry.
APC/T cell cocultures and T cell proliferation assays
Spleens were mechanically dissociated by passage through a nylon steel
mesh and RBC were lysed. T cell purification was performed using mouse
CD4 subset mini-column kits from R&D Systems according to the man-
ufacturer’s instructions. MACS-purified (CD11c
⫹
) or sorted DC subsets
(CD11c
⫹
CD8
␣
⫹
and CD11c
⫹
CD8
␣
⫺
cells) were cultured with syngeneic
T cells plus soluble anti-mouse CD3␧(0.5
␮
g/ml; Cedarlane Laboratories)
or with allogeneic C57BL/6 T cells. In other assays, macrophages and DC
were either plated separately or together in varying ratios (total APC equal-
ing 2 ⫻10
5
/ml) in 96-well round-bottom plates with fixed numbers of
syngeneic T cells (10
6
/ml) and soluble anti-mouse CD3␧(0.5
␮
g/ml). The
APC were freshly purified or preactivated for 18 h with LPS (1
␮
g/ml) for
macrophages and with GM-CSF (10 ng/ml) plus IL-4 (10 ng/ml) plus
IFN-
␥
(20 ng/ml) plus poly(I:C) (100 ng/ml) for DC. Mature APC were
washed three times to remove residual cytokines before coculture with T
cells. For measurement of proliferation, APC/T cell combinations were
plated in triplicate in 96-well round-bottom plates in final volumes of 200
␮
l/well. Titrations of APC were plated with fixed numbers of T cells as
indicated in the graph legends. Plates were incubated in 5% CO
2
at 37°C
for 3 or 4 days for syngeneic and allogeneic assays, respectively. Cells
were pulsed with 1
␮
Ci/well [
3
H]thymidine (NEN-DuPont) for the last
18 h of culture. [
3
H]Thymidine uptake was measured using a cell harvester
(Tomtec) and a liquid scintillation counter (Wallac). The results were ex-
pressed as the mean cpm ⫾SD from triplicate wells. To measure cytokine
concentrations, APC/T cell combinations were plated in 24-well plates at
the indicated concentrations and supernatant samples were taken from 60 h
cultures for ELISA.
5236 APCs IN NOD MICE

Assessment of glutamic acid decarboxylase (GAD)
p524-reactive T cell activation
GAD65 peptide comprised of amino acids 524 –543 (GADp524) was syn-
thesized in our laboratory using the Merrifield solid-phase technique using
an ABI Peptide Synthesizer (Applied Biosystems) as before (26). Peptides
were purified using HPLC on a C18 reverse-phase semipreparative Syn-
chropak RP-P column (Synchrom), lyophilized, and stored at ⫺20°C until
use. Purity of the peptides was verified using mass spectrometry. Peptides
were reconstituted in double-distilled H
2
O and sterilized by passage
through a 22-
␮
m filter before incorporation into cell cultures.
APC subsets were used to stimulate induced and spontaneous pro-
liferation responses to GADp524. For induced responses, 6-wk-old
NOD or NOR mice were immunized s.c. in the hind footpads with 100
␮
g of GADp524 emulsified in CFA. Popliteal lymph nodes were re-
moved 10 days later and teased through nylon meshes to obtain single-
cell suspensions. CD4
⫹
T cells were then isolated from pooled lymph
node samples using mouse CD4 subset mini column kits (R&D Sys-
tems). Splenocytes from separate, nonimmunized mice (4 –5 wk of age)
were harvested for purification of DC (CD11c
⫹
cells) by MACS sep-
aration. CD4
⫹
T cells (2 ⫻10
5
/well) together with titrations of DC
FIGURE 1. Percentages and numbers of splenic DC subsets in NOD and autoimmune-resistant control strains. A, Flow cytometry was performed
to analyze DC subsets in spleens from 8-wk-old NOD and autoimmune-resistant mice. Upper panels, Viable lymphocytes were gated based on
forward and side scatter properties and the percentages of CD11c
high
cells were analyzed. Lower panels, CD11c
high
cells were gated to identify
CD8
␣
⫹
populations. B, Summary of the percentages of CD11c
high
cells in spleens from the indicated strains/ages of mice, as determined by flow
cytometry. Diabetic mice were between 15 and 19 wk of age, whereas NOD mice in all other age categories were normoglycemic. C, The absolute
numbers of CD11c
high
cells (⫻10
6
) were calculated by multiplying total splenocyte counts by the percentages of CD11c
high
cells in B. D, Summary
of the percentages of CD8
␣
⫹
DC among gated CD11c
high
cells from NOD and control strain spleens. ⴱ, A statistically significant difference between
NOD and age-matched autoimmune resistant mice. The results are from six to eight mice per group and are expressed as the mean ⫾SD from
individual animals.
5237The Journal of Immunology

were plated in 96-well microtiter plates with GADp524 for 4 days.
GADp524 was added at the optimal concentration of 10
␮
M.
For assaying spontaneous T cell responses, CD4
⫹
T cells were purified
from the spleens of nonimmunized 8- to 10-wk-old NOD mice as described
and were cocultured with dilutions of MACS-purified B cells (B220
⫹
cells), MACS-purified DC (CD11c
⫹
cells), or FACS-sorted CD8
␣
⫹
or
CD8
␣
⫺
DC from NOD and NOR mice. Before incorporation into T cell
cocultures, B cell proliferation was inhibited by treating the cells with 50
␮
g/ml mitomycin C for 30 min at 37°C, followed by thorough washing,
counting, and resuspension in medium. Cultures were established in trip-
licate wells in 96-well flat-bottom plates with GADp524 at the indicated
concentrations. [
3
H]Thymidine incorporation assays were used to measure
T cell proliferation after 5 days. In the spontaneous assays, IFN-
␥
was also
measured by ELISA using 48-h supernatant samples and cells were taken
at 24 h for flow cytometry.
Cytokine measurements by ELISA
Cytokines in supernatant samples were measured using OptEIA ELISA
sets for IL-12p70, IL-12p40, IL-10, IFN-
␥
, IL-4, TNF-
␣
, and IL-18 (BD
Pharmingen) in accordance with the manufacturer’s instructions. All sam-
ples were analyzed in duplicate wells. Plates were read using a Bio-Rad
ELISA plate reader. All results are expressed as the mean picograms per
milliliter of cytokine ⫾SD from duplicate or triplicate wells.
Statistics
Statistical comparisons of two groups were performed using the Student t
test and multiple comparisons were conducted using one-way ANOVA
tests and Bonferroni multiple comparison tests for subsequent pairwise
comparisons where appropriate. A pvalue ⬍0.05 was considered
significant.
Results
NOD spleens have an altered DC subset composition
To assess the frequency of DC in NOD mice, comparisons were
drawn to splenic DC from autoimmune-resistant NOR, C57BL/6,
and BALB/c mice. The NOR strain shares the NOD MHC haplo-
type as well as a proportion of the diabetes susceptibility genes.
Although NOR islets exhibit APC infiltration, the subsequent re-
cruitment of T cells is minimal, suggesting that differences in APC
activity in NOD mice are responsible for the transition to overt
FIGURE 2. Turnover and survival of DC subsets in NOD and C57BL/6 spleens. Aand B, To measure the incorporation of BrdU in newly generated
DC, 6-wk-old NOD and C57BL/6 mice were given BrdU in their drinking water for 2 or 5 consecutive days. Splenocytes were harvested and stained for
CD11c and CD8
␣
, fixed/permeabilized, and stained with anti-BrdU Ab to identify proliferating cells. A, Representative histograms depict the percentages
of BrdU
⫹
cells among gated CD11C
high
, CD11c
high
CD8
␣
⫹
, and CD11c
high
CD8
␣
⫺
populations on day 2 of labeling. The bar graph depicts the mean
percentages ⫾SD (n⫽4 mice/group) on days 2 and 5 of BrdU labeling. B, The absolute numbers (⫻10
5
) of BrdU
⫹
CD11c
high
CD8
␣
⫹
and BrdU
⫹
CD11c
high
CD8
␣
⫺
cells were calculated after 5 days of BrdU labeling. C, Apoptosis of DC was assessed by staining freshly isolated NOD, NOR, and
C57BL/6 splenocytes with Abs against DC subset-specific markers, annexin V, and PI and tabulating the percentages of annexin V-positive
PI-negative cells among gated CD11c
high
CD8
␣
⫹
and CD11c
high
CD8
␣
⫺
populations (n⫽4 mice/strain). ⴱ, A statistically significant difference
between NOD and control strain DC populations.
5238 APCs IN NOD MICE

disease (27). Flow cytometry was performed to identify expression
of CD11c, a surface molecule which phenotypically identifies DC
in murine lymphoid organs (24) (Fig. 1A,upper panels, depicting
the lymphocyte gate from representative mice). We observed that
the percentages of CD11c
high
cells in NOD spleens were within a
normal range based on comparisons to age-matched control mice,
representing ⬃3% of gated lymphocytes (Fig. 1B). The absolute
numbers of CD11c
high
cells were not significantly different in
NOD and age-matched control spleens (Fig. 1C). Moreover, dia-
betic NOD mice did not display altered percentages or numbers of
CD11c
high
cells compared with age-matched normoglycemic NOD
mice or control strains (Fig. 1, Band C). In all of the strains
examined, the numbers of CD11c
high
cells increased between 4
and 8 wk of age but did not change between 8 and 16 wk of age.
Mouse spleen contains CD11c
high
DC subsets, which are com-
prised of CD8
␣
⫹
and CD8
␣
⫺
populations (28), in addition to plas-
macytoid DC having a CD11c
low
B220
⫹
Gr-1
⫹
phenotype (29, 30).
Next, we compared the composition of DC subsets in NOD and
autoimmune-resistant mice. Representative histograms depicting
the staining to identify CD8
␣
⫹
DC in 8-wk-old mice are shown in
Fig. 1A(lower panels; CD11c
high
-gated cells) and were used to
calculate the percentages of CD8
␣
⫹
cells among gated CD11c
high
cells in NOD and control strain spleens as a function of age (Fig.
1D). Although splenocytes from all the strains demonstrated a
modest age-related decline in the proportions of CD11c
high
cells
that were CD8
␣
⫹
, NOD spleens contained comparatively reduced
percentages of CD8
␣
⫹
cells at 4, 8, and 16 wk of age. Diabetic
NOD mice also presented with a reduced frequency of CD8
␣
⫹
cells in the DC compartment in comparison to age-matched control
strains; however, there were no significant differences between di-
abetic and nondiabetic NOD mice (Fig. 1D). Therefore, the im-
balance in DC subsets observed in NOD spleens is an inherent
feature of this strain that is not influenced by hyperglycemia.
DC subsets were also gated on the basis of CD8, CD11b, and
CD4 expression; the splenic CD11c
high
population consists of
CD8
␣
⫹
CD11b
low/⫺
CD4
⫺
, CD8
␣
⫺
CD11b
high
CD4
⫹
, and CD8
␣
⫺
CD11b
high
CD4
⫺
populations in autoimmune-resistant mice (28).
We observed that the NOD splenic DC compartment contained a
pronounced reduction in the percentages of CD8
␣
⫹
CD11b
low/⫺
cells and a coordinate increase in the proportions of CD8
␣
⫺
CD11b
high
cells at 8 wk of age. CD8
␣
⫹
CD11b
low/⫺
cells rep-
resented 18.0 ⫾2.4, 25.2 ⫾1.8, 27.8 ⫾3.4, and 24.4 ⫾2.7%
of gated CD11c
high
cells in NOD, NOR, C57BL/6, and BALB/c
mice, respectively. The CD8
␣
⫺
CD11b
high
population com-
prised 81.6 ⫾3.5, 74.0 ⫾2.3, 69.8 ⫾0.9, and 71.9 ⫾3.2% of
NOD, NOR, C57BL/6, and BALB/c CD11c
high
cells, respectively
(n⫽6 – 8 mice/strain, p⬍0.05 for NOD vs control strain popu-
lations). We also observed that the CD8
␣
⫺
CD11b
high
subset of
NOD DC contained increased percentages of CD4
⫹
cells com-
pared with control strains; CD4
⫹
cells comprised 70.4 ⫾3.2,
58.5 ⫾1.2, and 54.2 ⫾1.8% of gated CD11c
high
CD8
␣
⫺
CD11b
high
splenocytes in 8-wk-old NOD, NOR, and BALB/c
mice, respectively. However, because few functional distinctions
between CD8
␣
⫺
CD4
⫹
and CD8
␣
⫺
CD4
⫺
DC subsets have been
characterized to date, we chose to categorize DC on the basis of the
broader CD8
␣
⫹
and CD8
␣
⫺
subsets (as depicted in Fig. 1A)inthe
subsequent experiments.
Unlike the other DC subsets, the frequency of plasmacytoid DC
is highly variable between mouse strains (29). Our data revealed
that the percentages and numbers of CD11c
low
B220
⫹
cells in
spleens from 8-wk-old NOD mice were within the broad range
found in autoimmune-resistant mice (data not shown). Therefore,
for the purpose of this report, we have focused on the CD8
␣
⫹
and
CD8
␣
⫺
DC subsets in NOD mice, excluding the plasmacytoid DC
population.
Limited differentiation of CD8
␣
⫹
DC and skewing toward the
CD8
␣
⫺
DC subset in NOD mice
Murine DC have a short lifespan in the spleen; the CD8
␣
⫹
subset has
an ⬃3-day lifespan, whereas CD8
␣
⫺
DC are longer-lived (31, 32).
NOD and C57BL/6 mice (6 wk of age) were given the DNA precur-
sor BrdU in their drinking water for 2 or 5 consecutive days to mea-
sure the turnover of DC populations using previously described meth-
ods (31, 32). Briefly, splenocytes were surface-stained for DC subset
molecules, fixed/permeabilized, and subsequently stained with fluo-
rescent anti-BrdU or isotype control Ab for flow cytometry. Because
DC are largely nonproliferating, the BrdU-labeled population repre-
sents the cells that have entered the splenic DC compartment either
Table I. Expression of costimulatory molecules on DC subsets in NOD and autoimmune-resistant mice
a
Surface Molecules (Geo MFI)
CD80 CD86 CD40
CD8
␣
⫹
CD8
␣
⫺
CD8
␣
⫹
CD8
␣
⫺
CD8
␣
⫹
CD8
␣
⫺
Resting
NOD 93 ⫾595⫾577⫾676⫾439⫾537⫾4
NOR 95 ⫾5 104 ⫾674⫾576⫾535⫾344⫾6
BALB/c 97 ⫾6 100 ⫾785⫾589⫾641⫾440⫾3
LPS ⫹CD40L
NOD 155 ⫾10 112 ⫾9 214 ⫾15 155 ⫾14 59 ⫾669⫾6
NOR 172 ⫾10 122 ⫾9 203 ⫾14 129 ⫾11 53 ⫾560⫾6
BALB/c 180 ⫾18 177 ⫾17 245 ⫾20 177 ⫾17 54 ⫾452⫾5
LPS ⫹TNF-
␣
NOD 139 ⫾11 112 ⫾9 199 ⫾11 144 ⫾11 66 ⫾769⫾6
NOR 140 ⫾11 99 ⫾6 190 ⫾17 169 ⫾10 72 ⫾575⫾8
BALB/c 148 ⫾12 102 ⫾9 225 ⫾18 180 ⫾17 82 ⫾10 72 ⫾7
GM-CSF ⫹IL-4 ⫹IFN-
␥
⫹poly(I:C)
NOD 220 ⫾15 195 ⫾11 268 ⫾20 210 ⫾12 85 ⫾891⫾10
NOR 202 ⫾10 183 ⫾7 275 ⫾19 189 ⫾17 79 ⫾890⫾9
BALB/c 195 ⫾18 170 ⫾14 289 ⫾25 222 ⫾13 90 ⫾782⫾7
a
Splenocytes from 4- to 6-wk-old mice were given the following treatments for 18 h: LPS (100 ng/ml) plus CD40L-transfected J558 cells (1 DC:1 J558 cell), LPS (100 ng/ml)
plus TNF-
␣
(10 ng/ml), or GM-CSF (10 ng/ml) plus IL-4 (10 ng/ml) plus IFN-
␥
(20 ng/ml) plus poly(I:C) (100 ng/ml). Freshly isolated splenocytes were used as a source of
resting DC. Flow cytometry was done to analyze the geometric MFI (mean geo MFI ⫾SD) of staining for CD80, CD86, and CD40 on gated CD11c
high
CD8
␣
⫹
and
CD11c
high
CD8
␣
⫺
populations from a representative experiment where four mice per strain were examined.
5239The Journal of Immunology

through expansion of a splenic precursor or through replenishment by
precursors from the blood or BM (31). Representative histograms
depicting BrdU-stained CD11c
high
cells (total DC) and the CD8
␣
⫹
and CD8
␣
⫺
DC subsets in NOD and C57BL/6 spleens on day 2 of
labeling are provided in Fig. 2Aand are summarized for all mice on
days 2 and 5 of labeling (Fig. 2A, bar graph). The data revealed that
lower percentages of BrdU
⫹
cells were found within the NOD DC
compartment on days 2 and 5. Importantly, however, evaluation of the
BrdU labeling kinetics of individual DC subtypes showed no differ-
ences between NOD and C57BL/6 mice with respect to the percent-
ages of CD8
␣
⫹
or CD8
␣
⫺
DC that had incorporated BrdU (Fig. 2A).
This finding suggested that the difference in BrdU labeling kinetics of
total CD11c
high
cells was attributable to the differing DC subset com-
position in NOD and C57BL/6 mice, rather than due to actual differ-
ences in DC half-life. In support of this notion, the absolute numbers
of BrdU
⫹
CD11c
high
CD8
␣
⫹
cells in NOD spleens were reduced and
the numbers of BrdU
⫹
CD11c
high
CD8
␣
⫺
cells were increased in
comparison to C57BL/6 mice after 5 days of labeling (Fig. 2B).
Lastly, we also examined the steady-state survival of NOD, NOR, and
C57BL/6 DC by staining splenocytes with annexin V and PI. The
percentages of annexin V
⫹
PI
⫺
cells, which represent the cells under-
going apoptosis, did not differ in the CD8
␣
⫹
or CD8
␣
⫺
DC popu-
lations from NOD, NOR, or C57BL/6 spleens (Fig. 2C). Collectively,
these data suggest that the increased frequency of CD8
␣
⫺
DC and the
reduced differentiation of CD8
␣
⫹
DC in NOD mice are attributable to
differences in the availability of DC precursors rather than due to
altered DC turnover. These data provide valuable insight that the ac-
tivity of DC precursors is altered in NOD mice although further stud-
ies will be required to decipher whether differences in precursor mi-
gration and/or lineage commitment account for these DC subset
abnormalities.
NOD DC are effective costimulators
Because functional defects of APC have been proposed to play a
role in diabetes pathogenesis, we asked whether the immunostimu-
latory capabilities of NOD DC subsets were altered in comparison
to autoimmune-resistant strains. Resting DC subsets from NOD
mice did not display differences in expression of CD80, CD86, or
CD40 relative to DC from NOR and BALB/c spleens (Table I, geo
mean fluorescence intensity (MFI) values). To assess the pheno-
type of NOD DC during maturation, several treatments were test-
ed: LPS (100 ng/ml) plus irradiated J558 cells (a CD40L-trans-
fected cell line), LPS (100 ng/ml) plus TNF-
␣
(10 ng/ml), and
GM-CSF (10 ng/ml) plus IL-4 (10 ng/ml) plus IFN-
␥
(20 ng/ml)
plus poly(I:C) (100 ng/ml). Splenocytes were incubated with the
indicated stimuli for 18 h and flow cytometry was done to assess
costimulatory molecule expression on DC subsets. Comparable ex-
pression levels of CD80 and CD40 were found on CD8
␣
⫹
and
CD8
␣
⫺
DC from NOD and autoimmune-resistant mice under the
various maturation conditions (Table I), although NOD and NOR
CD8
␣
⫹
DC exhibited reduced CD86 expression compared with
BALB/c DC in response to proinflammatory stimuli. However, in
allogeneic MLR assays with naive C57BL/6 T cells, NOD and
NOR total DC (Fig. 3A), CD8
␣
⫹
DC (Fig. 3B), and CD8
␣
⫺
DC
(Fig. 3C) at varying concentrations demonstrated a similar allo-
stimulatory ability, proving that NOD DC are competent
costimulators.
NOD CD8
␣
⫹
DC demonstrate abnormally low IL-12
production and increased IL-10 production in response
to innate and T cell-derived signals
Numerous studies have reported that NOD macrophages and in
vitro-differentiated DC produce increased quantities of IL-12 in
response to maturation stimuli (15, 16, 18, 20, 21); therefore, we
explored whether in vivo-derived DC from NOD mice possess
similar characteristics. IL-12p70 is the biologically active het-
erodimer comprised of p40 and p35 subunits, whereas the p40
form is produced in excess as free monomers or dimers (33). IL-
12p70 synthesis by murine splenic DC requires multiple signals
which are transmitted through TLR and CD40 and/or proinflam-
matory cytokines (34, 35). To assess IL-12 production, we cultured
NOD and NOR DC (MACS-purified CD11c
⫹
cells) with a cyto-
kine mixture consisting of GM-CSF (10 ng/ml), IL-4 (10 ng/ml),
and IFN-
␥
(20 ng/ml) in combination with poly(I:C) (100 ng/ml),
a treatment which is known to induce IL-12p70 production by
CD8
␣
⫹
DC (36). Significantly, analysis of cytokine secretion
showed that NOD DC produced reduced quantities of IL-12p70
and IL-12p40 in comparison to NOR DC in response to the cyto-
kine mixture plus titrated doses of poly(I:C) (Fig. 4A). Expectedly,
FIGURE 3. Equivalent allostimulation by NOD and NOR DC popula-
tions. MACS-purified CD11c
⫹
cells (A) or FACS-sorted CD11c
high
CD8
␣
⫹
(B) and CD11c
high
CD8
␣
⫺
DC (C) from NOD and NOR spleens were cul-
tured at the indicated dilutions with allogeneic C57BL/6 T cells (10
5
cells/
well). T cells alone (T) did not proliferate. Proliferation was determined by
addition of [
3
H]thymidine for the last 18 h of the 96-h culture. Results are
presented as mean cpm ⫾SD of triplicate wells. T cell proliferation elic-
ited by NOD vs NOR DC was not statistically different.
5240 APCs IN NOD MICE

DC that were treated with cytokines in the absence of poly(I:C),
and are therefore partially mature, did not produce bioactive IL-
12p70. However, production of IL-12p40 by NOD DC was re-
duced in comparison to NOR DC (Fig. 4A). NOD and NOR DC
produced equivalent quantities of two other proinflammatory cy-
tokines, TNF-
␣
and IL-18 (Fig. 4A). As a second method of stim-
ulating IL-12 secretion, NOD and NOR DC were treated with LPS
(100 ng/ml) plus titrations of irradiated CD40L-transfected J558
cells, which also revealed significantly impaired IL-12p70 produc-
tion by NOD DC (Fig. 4B), thereby demonstrating that this feature
of NOD DC was not stimulus specific.
We also tested IL-12p70 production by DC that were activated
in vivo, where a complex array of cellular interactions can influ-
ence DC function. Four-week-old NOD and NOR mice were in-
jected i.v. with PBS (vehicle control), LPS (30
␮
g/mouse), or
poly(I:C) (100
␮
g/mouse). At 6-h postinjection, splenocytes were
pooled from each treatment group, the CD11c
⫹
cells were MACS-
purified and recultured overnight in nonsupplemented medium to
allow cytokine accumulation. The data revealed that DC from
poly(I:C)-treated NOR mice, but not from PBS or LPS treatment
groups, produced IL-12p70 (Fig. 4C). However, DC from poly(I:
C)-treated NOD mice possessed a severely diminished IL-12p70-
producing capability in comparison to their NOR counterparts.
This result confirms that the cytokine production defect of NOD
DC also occurs during immune responses in vivo.
The PLN are the primary site of autoreactive T cell priming and
tolerance induction to islet Ags in NOD mice (37). Comparison of
the DC subset composition in the PLN from 6 wk-old NOD and
FIGURE 4. Selective reduction of
IL-12 synthesis by NOD DC. A,
MACS-purified splenic DC (8 ⫻10
5
cells/ml) from NOD and NOR mice
were stimulated with GM-CSF (10 ng/
ml) ⫹IL-4 (10 ng/ml) ⫹IFN-
␥
(20
ng/ml) (designated as mature DC) ⫾
poly(I:C) at the indicated doses. Imma-
ture DC were treated with GM-CSF (10
ng/ml) alone. Supernatants were sam-
pled after 24 h for IL-12 p70, IL-12
p40, TNF-
␣
, and IL-18 measurement
by ELISA. B, Splenic DC (1 ⫻10
6
/ml)
were cultured with LPS (100 ng/ml)
plus titrations of irradiated CD40L-
transfected J558 cells (ratios ranging
from 1 DC:5 J558 cells to 10 DC:1
J558 cell) and IL-12p70 was measured
in 24-h supernatant samples. C, Assess-
ment of IL-12p70 production by DC
following in vivo activation. Four-
week-old NOD and NOR mice (n⫽
3/strain) were injected i.v. with LPS (30
␮
g/mouse), poly(I:C) (100
␮
g/mouse),
or PBS (vehicle control). Pooled
splenocytes from each strain were
taken at 6-h postinjection and CD11c
⫹
cells were selected using MACS col-
umns. DC were plated (8 ⫻10
5
cells/
ml) in nonsupplemented RPMI 1640
medium for a collection of 30-h super-
natant samples for measurement of IL-
12p70 by ELISA. D, Comparison of
IL-12 production by APC in NOD and
NOR PLN. Flow cytometry was per-
formed using PLN-derived lympho-
cytes that were treated with GM-CSF
(10 ng/ml) ⫹IL-4 (10 ng/ml) ⫹IFN-
␥
(20 ng/ml) ⫹poly(I:C) (100 ng/ml) for
10 h with inclusion of brefeldin A for
thelast5hofculture. Cells were gated
according to CD11c
high
CD11b
⫹
and
CD11c
high
CD11b
low/⫺
populations (dot
plots) and analyzed to measure intracel-
lular IL-12p40/70 expression (histo-
grams delineating the percentages of
positive cells). IL-12p40/p70 expres-
sion was not detected in nonstimulated
control DC cultures (data not shown).
ⴱ,p⬍0.05 for NOD vs NOR or
C57BL/6 DC for a particular treatment.
5241The Journal of Immunology

NOR mice revealed that the populations of CD11c
high
CD11b
high
and CD11c
high
CD11b
low/⫺
cells were equivalent between the
strains. The percentages of CD11c
high
CD11b
high
cells were 0.35 ⫾
0.10 and 0.36 ⫾0.13%, and the CD11c
high
CD11b
low/⫺
cells com-
prised 0.25 ⫾0.07 and 0.29 ⫾0.08% of NOD and NOR PLN,
respectively. Analysis of intracellular IL-12 expression was per-
formed by stimulating total PLN-derived cells with GM-CSF plus
IL-4 plus IFN-
␥
plus poly(I:C) and analyzing the DC populations
by flow cytometry. First, we checked that overnight culture of
lymphocytes did not affect the detection of DC subsets; we ob-
served no differences in the percentages of DC subsets among total
lymphocytes before vs after the 10-h culture (data not shown), nor
between poly(I:C)/cytokine-stimulated cultures from NOD and NOR
mice (Fig. 4D, dot plots). The CD11c
high
CD11b
low/⫺
population from
NOD PLN contained reduced percentages of IL-12-expressing cells
upon stimulation, and also displayed a reduced intensity of IL-12
staining compared with the corresponding NOR population (Fig. 4D,
histograms). In contrast, the NOD/NOR CD11c
high
CD11b
high
popu-
lations did not express IL-12p40/p70. Hence, unlike the splenic DC
compartment, the composition of DC subsets was similar in NOD and
NOR PLN; however, the reduced IL-12 production by NOD DC was
observed in both lymphoid organs.
To assess cytokine production by DC subsets, CD8
␣
⫹
and
CD8
␣
⫺
DC were FACS-sorted from NOD, NOR, and BALB/c
spleens and cultured for 24 h with GM-CSF plus IL-4 plus IFN-
␥
plus poly(I:C) or LPS plus CD40L-transfected J558 cells. Signif-
icantly, CD8
␣
⫹
DC from NOD mice produced reduced quantities
of IL-12p70 in comparison to NOR and BALB/c DC in both cul-
ture conditions (Fig. 5A). CD8
␣
⫹
DC were the major IL-12p70-
producing subset, whereas IL-12p70 secretion by CD8
␣
⫺
DC was
negligible in response to poly(I:C)/cytokine stimulation. Interest-
ingly, in response to LPS/CD40L stimulation, the CD8
␣
⫺
DC sub-
set produced small quantities of IL-12p70 that were consistently
reduced in NOD cultures; however, the difference from control
strain DC did not reach statistical significance (Fig. 5A). TNF-
␣
was observed to be a major cytokine product of activated CD8
␣
⫺
FIGURE 5. NOD CD8
␣
⫹
DC exhibit reduced IL-12p70 production and increased IL-10 synthesis during maturation. A–C, Comparison of cytokine
production by NOD, NOR, and BALB/c DC subsets. FACS-sorted CD8
␣
⫹
and CD8
␣
⫺
DC (5 ⫻10
5
cells/ml) were cultured for 24 h in GM-CSF
(10 ng/ml) ⫹IL-4 (10 ng/ml) ⫹IFN-
␥
(20 ng/ml) ⫹poly(I:C) (100 ng/ml) or LPS (100 ng/ml) ⫹CD40L transfected J558 cells (1 DC: 1 J558 cell).
IL-12p70 (B), TNF-
␣
(C), and IL-10 (D) concentrations were measured by ELISA. Dand E, Analysis of the kinetics of cytokine production by
MACS-purified CD11c
⫹
cells (5 ⫻10
5
cells/ml) from NOD, NOR, and BALB/c spleens. Cells were cultured in GM-CSF ⫹IL-4 ⫹IFN-
␥
⫹poly(I:C)
and supernatant samples were acquired after 14, 24, or 40 h of culture for measurement of the quantities of IL-12p70 (D) and IL-10 (E). ⴱ, A significant
difference between NOD DC vs NOR or BALB/c DC for a particular treatment.
5242 APCs IN NOD MICE

DC and was produced in comparable quantities by NOD and
control strain DC during maturation with poly(I:C)/cytokines (Fig.
5B). Therefore, the cytokine production defect of the NOD DC
compartment results from a numerical deficiency of the major IL-
12-producing CD8
␣
⫹
DC subset combined with a reduced ability
of CD8
␣
⫹
DC to produce IL-12p70.
IL-10 is a DC-derived cytokine with autocrine and paracrine
inhibitory effects on DC maturation and IL-12 synthesis (38, 39).
Interestingly, there was a trend toward increased IL-10 secretion
by NOD splenic CD8
␣
⫹
DC in response to maturation with GM-
CSF plus IL-4 plus IFN-
␥
plus poly(I:C), although the difference
from NOR and BALB/c cells was not statistically significant (Fig.
5C). However, stimulation with LPS plus CD40L transfectants re-
vealed a significant increase in IL-10 production by NOD CD8
␣
⫹
DC in comparison to the corresponding NOR and BALB/c subset
(Fig. 5C). CD8
␣
⫺
DC were poor producers of IL-10 in response to
both stimulation conditions and cells from all three strains pro-
duced similar quantities of this cytokine (Fig. 5C).
We performed a time-course analysis of IL-12p70 and IL-10
production by MACS-purified CD11c
⫹
cells in response to poly(I:
C)/cytokine stimulation. Interestingly, there was an inverse corre-
lation between the quantities of IL-12p70 (Fig. 5D) and IL-10 (Fig.
5E) in mature DC cultures; the NOR DC compartment produced
increased IL-12p70 and reduced IL-10 levels whereas NOD DC
displayed the opposite cytokine bias. The deficit in IL-12p70 pro-
duction by NOD DC was observed after 14, 24, and 40 h of poly(I:
C)/cytokine treatment, indicating that the reduced quantities of IL-
12p70 did not result from delayed kinetics of DC maturation in
NOD cultures (Fig. 5D). Once again, IL-10 synthesis by NOD DC
was noticeably increased in poly(I:C)/cytokine-stimulated cul-
tures; however, a statistically significant difference from control
strain cells was only evident after 40 h of culture (Fig. 5E). Taken
together, these data reveal an unusual cytokine profile of NOD
CD8
␣
⫹
DC in response to proinflammatory stimuli, marked by
increased IL-10 and reduced IL-12 production.
A limited ability to elicit Th1-polarized responses is associated
with reduced IL-12p70 synthesis by NOD CD8
␣
⫹
DC
To assess the role of NOD DC in Th1 cytokine polarization,
CD8
␣
⫹
and CD8
␣
⫺
DC from NOD and NOR spleens were pre-
treated with poly(I:C)/cytokines for 18 h, washed thoroughly to
remove residual cytokines, and were subsequently cultured with
syngeneic CD4
⫹
T cells and anti-CD3 to stimulate the TCRs. In
preliminary experiments, a 1:10 DC-T cell ratio induced optimal
IFN-
␥
production by T cells in 48 h cultures with 0.5
␮
g/ml anti-
CD3 Ab (in comparison to 1:1, 1:5, or 1:20 ratios; data not shown);
therefore, these cellular proportions were used in the subsequent
experiments. The cytokine and proliferation data from compari-
sons of NOD and NOR DC/T cocultures are summarized in Table
II. Control cultures containing immature (i.e., freshly purified)
CD8
␣
⫹
or CD8
␣
⫺
DC as stimulators revealed robust T cell pro-
liferation but minimal IFN-
␥
production, verifying that microbial/
cytokine stimulation is required for Th1 polarization. Notably, T
cell proliferation elicited by immature DC subsets from NOD and
NOR mice was equivalent and it was also evident that CD8
␣
⫺
DC
were more proficient than CD8
␣
⫹
DC at inducing CD4
⫹
T cell
responses, in agreement with prior studies of T cell stimulation in
vitro by DC from nonautoimmune strains (40, 41). In cultures
containing poly(I:C)/cytokine-matured CD8
␣
⫺
DC, NOD and
NOR cells demonstrated comparable T cell proliferation and
IFN-
␥
and IL-4 production. Poly(I:C)/cytokine-stimulated CD8
␣
⫺
DC from NOD and NOR mice did not produce IL-12p70, thereby
implicating alternate DC-derived cytokines in Th1 induction by
this DC subset. Significantly, CD8
␣
⫹
DC from NOD mice incited
considerably lower levels of IFN-
␥
secretion and IL-12p70 pro-
duction was coordinately reduced (⬃2.5-fold) whereas T cell pro-
liferation and IL-4 concentrations were equivalent in NOD and
NOR cultures. Addition of 2.5 ng/ml recombinant murine IL-
12p70 to NOD CD8
␣
⫹
DC/T cell cultures not only equalized the
IL-12p70 concentrations but also augmented the IFN-
␥
concentra-
tions to the approximate levels found in NOR CD8
␣
⫹
DC/T cell
cultures. These findings provide evidence for the adjuvant effect of
IL-12p70 on Th1 induction in this assay. In contrast, supplemen-
tation of NOR CD8
␣
⫹
or NOD/NOR CD8
␣
⫺
DC/T cell cultures
with 2.5 ng/ml rIL-12p70 had comparatively modest effects on
IFN-
␥
production, suggesting that IFN-
␥
production in these cul-
tures was already maximized. Altogether, these results delineate a
reduced IL-12-producing capability of NOD CD8
␣
⫹
DC which
correlates with an impaired ability to polarize T cells toward IFN-
␥
production.
Atypical reliance upon macrophages over DC for Th1
polarization in NOD mice
DC and macrophages are believed to act cooperatively in promot-
ing proinflammatory T cell responses in NOD mice (3). Given
their contrasting defects of IL-12 production in this strain, we com-
pared the contributions of these APC types to Th1 priming in NOD
and NOR mice. We isolated splenic CD11c
⫹
cells using MACS
separation and splenic macrophages by depletion of nonmacroph-
age lineage cells followed by selection based on plastic adherence
Table II. Comparison of T cell responses elicited by NOD and NOR DC subsets
a
Proliferation or
(Cytokine) Immature NOD DC
b
⫹NOD T Immature NOR DC
b
⫹NOR T Mature NOD DC
c
⫹NOD T Immature NOR DC
c
⫹NOR T
DC subset CD8
␣
⫺
CD8
␣
⫹
CD8
␣
⫺
CD8
␣
⫹
CD8
␣
⫺
CD8
␣
⫹
CD8
␣
⫺
CD8
␣
⫹
cpm 81,987 ⫾2,318 65,737 ⫾2,200 84,530 ⫾2,124 67,012 ⫾1,835 126,031 ⫾9,674 78,839 ⫾10,802 160,912 ⫾8,708 75,681 ⫾18,902
IFN-
␥
2,062 ⫾24 2,045 ⫾21 2,180 ⫾33 2,020 ⫾26 17,201 ⫾863 6,258
d
⫾262 20,295 ⫾3,451 19,053 ⫾3,297
IL-4 NS NS NS NS 1,376 1,296 1,285 1,380
IL-12 p70 ⫹
IL-12p70
e
NS NS NS NS 57 ⫾13 1,850
d
⫾51 42 ⫾7 4,637 ⫾378
IFN-
␥
ND ND ND ND 25,456 ⫾1,730 30,320 ⫾5,000 28,226 ⫾5,011 23,987 ⫾5,558
a
FACS-sorted CD11c
high
CD8
␣
⫺
and CD11c
high
CD8
␣
⫹
cells from NOD and NOR spleens were cocultured with syngeneic T cells (10
5
DC plus 10
6
T cells/ml) and soluble
anti-CD3 (0.5
␮
g/ml).
b
Control cultures containing freshly purified NOD DC with T cells/anti-CD3.
c
To generate mature DC, cells were pretreated with GM-CSF plus IL-4 plus IFN-
␥
plus poly(I:C) for 18 h prior to coculture with T cells/anti-CD3.
d
A statistically significant difference between NOD and NOR cultures. NS, Below the detection limit of the assay.
e
Cultures were supplemented with 2.5 ng/ml recombinant mouse IL-12p70. The cultures were incubated for 48 72 h for ELISA and thymidine incorporation assays,
respectively. Cytokine concentrations are represented as the mean picograms per milliliter ⫾SD from duplicate cultures, except for IL-4 where single wells were assayed.
Proliferation is shown as the mean cpm of triplicate wells ⫾SD.
5243The Journal of Immunology

(refer to Materials and Methods). DC were first matured with
poly(I:C)/cytokines (as described in the previous experiments) and
macrophages were treated with LPS (1
␮
g/ml). These distinct
stimuli were used for DC and macrophage maturation to elicit
maximal IL-12p70 production by each APC type, as determined in
preliminary experiments (data not shown). Mature APC were sub-
sequently washed and cultured with syngeneic T cells and soluble
anti-CD3. A constant ratio of APC to T cells was maintained (1
APC:10 T cells) but the composition of macrophages and DC
within the APC population was varied to assess their relative con-
tributions to T cell activation.
We first verified the cytokine production capabilities of NOD
and NOR splenic macrophages. LPS-treated NOD macrophages
produced increased quantities of IL-12p40 (Fig. 6A) whereas
IFN-
␥
production by mature NOD and NOR macrophages was low
and equivalent (Fig. 6B). As anticipated, poly(I:C)/cytokine-stim-
ulated DC demonstrated the reverse trend in IL-12p40 production
between the strains (Fig. 6A) whereas IFN-
␥
was not a product of
stimulated DC (Fig. 6B). IL-12p40 synthesis by APC was also
considerably augmented in cocultures with T cells (Fig. 6A). The
quantities of IL-12p40 and IFN-
␥
were increased in cultures of
LPS-pretreated NOD macrophages with T cells/anti-CD3, whereas
cocultures containing poly(I:C)/cytokine-stimulated NOD DC
demonstrated the opposite trend in comparison to NOR cultures
(Fig. 6, Aand B). When mature DC and macrophages (1 DC: 1
macrophage) were admixed with syngeneic T cells/anti-CD3, the
quantities of IL-12p40 and IFN-
␥
were increased in NOD vs NOR
cultures (Fig. 6, Aand B). In contrast, at 10:1 ratios of DC to
macrophages, the opposite trend was observed: NOR cells pro-
duced increased quantities of IL-12p40 and IFN-
␥
. Expectedly,
FIGURE 6. NOD macrophages
have an augmented Th1-priming abil-
ity and can compensate for the IL-12
production defect of the NOD DC
compartment. NOD and NOR splenic
DC and macrophages (MAC) were
freshly purified (immature) or were
treated with maturation stimuli indi-
vidually for 18 h before coculture
with T cells. DC were pretreated with
GM-CSF (10 ng/ml) ⫹IL-4 (10 ng/
ml) ⫹IFN-
␥
(20 ng/ml) ⫹poly(I:C)
(100 ng/ml) and macrophages were
matured with LPS (1
␮
g/ml) for 18 h
and were washed thoroughly to re-
move residual cytokines. Dilutions of
immature or mature APC (2.0 ⫻10
5
/
ml) were plated alone or with synge-
neic T cells (2.0 ⫻10
6
/ml) ⫾anti-
CD3␧(0.5
␮
g/ml) in 24-well plates.
Supernatants were sampled after 60 h
to measure the quantities of IL-12p40
(A) and IFN-
␥
(B) by ELISA. Cand
D, To assess T cell proliferation, di-
lutions of immature (C) or mature (D)
APC (admixed DC and macrophages
totaling 1.2 ⫻10
5
/ml) were plated in
the indicated combinations with their
respective syngeneic T cells (1.2 ⫻
10
6
/ml) and anti-CD3␧. Control wells
were plated lacking anti-CD3␧. Thy-
midine uptake was measured after
72 h of culture and the results were
expressed as the mean cpm from trip-
licate wells ⫾SD. ⴱ, A significant
difference between NOD and NOR
cultures.
5244 APCs IN NOD MICE

cultures of immature APC (1 DC:1 macrophage) with T cells/
anti-CD3 contained low concentrations of IL-12p40 and IFN-
␥
which did not differ in NOD vs NOR cultures. T cell prolifer-
ation did not differ between NOD and NOR cultures during T
cell activation by immature (Fig. 6C) or mature APC types (Fig.
6D) admixed at varying ratios. This finding implies that Th1
biases rather than T cell numbers confer the cytokine produc-
tion differences between NOD and NOR cultures. Collectively,
these experiments demonstrate that activated NOD
macrophages are endowed with an enhanced Th1-promoting
proclivity whereas the NOD DC compartment is impaired in its
capacity to condition CD4
⫹
T cells toward IFN-
␥
production.
FIGURE 7. Autoreactive T cell activation by NOD and NOR APC. A, Activation of in vivo-expanded GADp524-reactive T cells by NOD vs NOR DC.
GAD65 p524-specific CD4
⫹
T cells were purified from pooled popliteal lymph nodes of NOD and NOR mice that had been immunized with peptide (100
␮
g/footpad) emulsified in CFA 10 days earlier. CD4
⫹
T cells (2 ⫻10
5
/well) were cultured in crossover combinations with titrations of MACS-purified
DC from the spleens of separate, nonimmunized NOD and NOR mice and 10
␮
M GADp524 for 96 h. Control wells containing CD4
⫹
T cells plus peptide
without DC were included to measure the residual T cell expansion resulting from the in vivo priming. B–D, Activation of spontaneously primed
GADp524-reactive T cells by NOD vs NOR APC. B cells (B220) and DC (CD11c
⫹
) were MACS-purified from the spleens of 8-wk-old NOD and NOR
mice. Dilutions of DC and B cells from NOD and NOR spleens were plated with splenic CD4
⫹
T cells from unprimed NOD mice (6 ⫻10
5
/well) and
GADp524. Control wells contained APC and T cells without peptide. In B, two concentrations of APC (2.5 ⫻10
5
or 5.0 ⫻10
5
cells/well) were plated
with NOD CD4
⫹
T cells ⫾10
␮
M GADp524. In Cand D, titrations of GADp524 were plated with DC or B cells (5.0 ⫻10
5
/well) and splenic CD4
⫹
T cells (7 ⫻10
5
/well) for measurement of proliferation (Band C) and IFN-
␥
production (D) after 5 days and 72 h, respectively. E, Analysis of IFN-
␥
production by GAD65-reactive T cells stimulated by DC subsets. FACS-sorted DC subsets (CD11c
high
CD8
␣
⫹
and CD11c
high
CD8
␣
⫺
cells; 5 ⫻10
4
cells/well) from NOD and NOR spleens were cultured with unprimed CD4
⫹
T cells (2 ⫻10
5
/well) from NOD spleens plus GADp524 (5
␮
M). Supernatant
samples were collected after 48 h of culture. F, MACS-purified NOD/NOR DC or B cells (2 ⫻10
5
/well) from spleens were plated with NOD CD4
⫹
T
cells (2 ⫻10
6
/well) and GADp524 (5
␮
M) and cells were harvested after 24 h to analyze expression of I-A
g7
(MHC class II) on APC populations by flow
cytometry. [
3
H]Thymidine incorporation assays (mean cpm ⫾SD) and ELISA (mean picograms per milliliter ⫾SD) were performed to analyze prolif-
eration and cytokine concentrations, respectively. ⴱ, A significant difference for a particular APC subset from NOD vs NOR mice.
5245The Journal of Immunology

NOD macrophages can thus overcompensate for the DC defect
in Th1 activation depending on the relative abundance of APC
types. This pattern differs for the NOR APC compartment, where
the contributions of DC to Th1 differentiation outweigh those of
macrophages.
NOD DC are efficient at autoantigen presentation but are less
effective than B cells
Ag presentation is another hallmark of immunostimulatory DC;
however, in NOD mice, the functions of DC in autoantigen pre-
sentation remain undefined. T cells reactive against a peptide of
GAD65, the immunodominant 524 –543 epitope (GADp524), are
predominant at 3– 4 wk of age during disease initiation in NOD
mice (42, 43); therefore, this peptide was used for Ag-presentation
studies to assess the activation of islet-reactive T cells by NOD DC
in vitro. Ag-specific T cells from GADp524/CFA-immunized
NOD and NOR mice were used in crossover combinations with
varying numbers of NOD/NOR splenic DC (MACS-purified
CD11c
⫹
cells) for measurement of T cell proliferation in thymi-
dine incorporation assays. The results revealed that GADp524-
specific T cells from NOD mice responded more vigorously than
NOR T cells to peptide restimulation, irrespective of the DC ge-
notype and at various DC dilutions (Fig. 7A), suggesting that the
differences between the strains were T cell dependent.
Proliferation of splenocytes from young NOD mice in response
to GADp524 does not require deliberate priming because these T
cells arise spontaneously in vivo (42, 43). Next, we evaluated
whether NOD and NOR APC differ with respect to their abilities
to stimulate spontaneously arising, GADp524-reactive T cells. In
comparison to the previous experiments which tested the prolifer-
ation of in vivo-expanded GADp524-reactive T cells, the sponta-
neous proliferation assays were expected to be more sensitive to
subtleties of APC function due to much lower frequencies of
GADp524-reactive T cells. NOD B cells were used as positive
control APC due to their established efficiency at stimulating in-
duced and spontaneously primed T cell responses to
␤
cell Ags (5,
7). Notably, we found that splenic macrophages were inefficient
APC for stimulating expansion of spontaneously arising
GADp524-reactive T cells in NOD spleens (data not shown).
MACS-purified DC (CD11c
⫹
) or B cells (B220
⫹
) from NOD and
NOR spleens were cultured with NOD splenic CD4
⫹
T cells and
GADp524 for measurement of proliferation. Strikingly, NOD B
cells were significantly better than NOD DC at stimulating
GADp524-reactive T cell proliferation at two different APC con-
centrations (Fig. 7B) and at titrated doses of peptide (Fig. 7C). In
contrast, DC and B cells from NOR mice had a similar efficacy as
NOD DC in spontaneous proliferation assays (Fig. 7, Band C).
Only NOD B cells displayed a heightened ability to stimulate
GADp524-reactive T cell proliferation in comparison to the other
APC tested. Additionally, NOD B cells elicited enhanced IFN-
␥
secretion that was on par with the high T cell proliferation ob-
served at various titrations of peptide (Fig. 7D). NOD DC, NOR
DC, and NOR B cells were comparable in terms of their abilities
to elicit IFN-
␥
synthesis (Fig. 7D), whereas IL-12p70 was not
detected in the culture supernatants (data not shown). Production
of IFN-
␥
was also equivalent in comparisons of NOD and NOR
DC subsets (FACS-sorted CD11c
high
CD8
␣
⫹
and CD11c
high
CD8
␣
⫺
cells) cultured with splenic T cells and GADp524 (Fig.
7E). Interestingly, NOD/NOR CD8
␣
⫹
DC consistently elicited
lower concentrations of IFN-
␥
than the CD8
␣
⫺
subset in this as-
say. Thus, while NOD and NOR DC possess comparable Ag-pre-
sentation abilities, NOD B cells uniquely possess a heightened
ability to present autoantigen and activate GADp524-reactive T
cells. Importantly also, the percentages of MHC class II
high
(i.e.,
mature) NOD B cells as well as their MHC class II expression
levels were significantly augmented in comparison to NOR B cells
and NOD/NOR DC after 24 h of culture with T cells and
GADp524 (Fig. 7F). This observation complements prior reports
that NOD B cells are functionally hyperactive (13, 14). The hier-
archy of APC proficiency for stimulation of GADp524-reactive T
cells is therefore another parameter that distinguishes NOD mice
from diabetes-resistant strains.
Discussion
Our investigations have uncovered numerical, phenotypic, and
functional abnormalities in the APC compartment of NOD mice.
This analysis has identified several key distinguishing features of
CD8
␣
⫹
and CD8
␣
⫺
DC in NOD mice in comparison to autoim-
mune-resistant strains. First, the composition of DC subsets differs
in NOD spleens, marked by a bias toward the CD8
␣
⫺
subset and
a deficit of CD8
␣
⫹
DC, a difference which was not observed in the
PLN of NOD vs NOR mice and therefore suggests that DC num-
bers are differentially regulated in the steady-state but not during
autoreactive T cell activation. BrdU labeling and apoptosis studies
indicate that DC turnover and survival do not differ in NOD and
control strain spleens, but instead implicate altered differentiation
or recruitment of DC precursors in the bias toward the CD8
␣
⫺
DC
subset in NOD mice. Second, this study demonstrates that CD8
␣
⫹
and CD8
␣
⫺
DC in NOD mice do not harbor gross defects in mat-
uration, as have been ascribed to NOD B cells, macrophages, and
in vitro-generated DC. CD8
␣
⫹
DC from NOD mice have a di-
minished ability to synthesize IL-12 but are otherwise functionally
normal. In the context of the abundant literature documenting the
functional abnormalities of other NOD professional APC types,
the present work suggests that the NOD DC compartment exhibits
the least functional impairment. To accentuate this point, our com-
parisons between NOD professional APC types revealed that B
cells are the most efficient presenters of autoantigen and that mac-
rophages mediate aberrantly elevated Th1 priming in this strain.
This is one of only a few reports which address the functional
capabilities of naturally occurring DC in NOD mice and represents
the most comprehensive study to date. Previous studies of NOD
DC defects have used in vitro-generated cells owing to their rel-
ative homogeneity and availability in large numbers. Studies of
clinical T1D have also relied mainly upon in vitro-generated cells
and have not yet reached a consensus concerning their functional
capabilities. A recent study from our laboratory has addressed this
issue by analyzing unmanipulated DC subsets in whole blood sam-
ples from patients with T1D (44). Intriguingly, DC from patients
with diabetes exhibit impaired IFN-
␣
production and modestly re-
duced 12p70 secretion, although their costimulation and T cell
activation abilities are intact. Unlike NOD macrophages, which
reproducibly exhibit cytokine and activation abnormalities irre-
spective of their source (15–17, 20), we have observed that DC in
NOD mice are functionally heterogeneous. This raises the issue of
whether studies of in vitro-generated DC can be extrapolated as
having functional relevance in vivo. An equivalent population in
vivo may be represented by the small subset of blood-derived
monocytes that are recruited to lymphoid tissue and undergo DC
differentiation during inflammation (45). These monocyte-derived
DC codifferentiate with macrophages, possess a high phagocytic
ability, and express high levels of MHC class II and costimulatory
molecules; therefore, this population closely resembles in vitro-
generated DC, which are also monocyte-derived and exhibit GM-
CSF-dependent differentiation. In contrast, CD8
␣
⫺
and CD8
␣
⫹
DC in the spleen are derived from a resident precursor that is
5246 APCs IN NOD MICE

distinct from monocytes (46). Hence, the fact that in vitro-gener-
ated DC appear to be developmentally and functionally related to
macrophages may underlie their parallel defects in NOD mice.
Our data reveal that IL-12 production by NOD DC is impaired
as a result of a numerical deficiency of CD8
␣
⫹
DC coupled with
a diminished capacity of this subset to synthesize IL-12, as we
have demonstrated in vitro in response to two distinct stimulation
conditions, during DC-T cell interactions, and in vivo. Moreover,
the abnormally low IL-12 production by both spleen and PLN-
derived CD8
␣
⫹
DC supports the idea that this is an inherent fea-
ture of NOD CD8
␣
⫹
DC rather than an environmentally pro-
grammed one. Our data also suggest that this NOD defect is
cytokine specific, although it is presently unclear which cytokine
pathways are predominantly used by the CD8
␣
⫺
DC subset to
mediate type 1 cytokine responses. For example, a recent study has
demonstrated that LPS-activated CD8
␣
⫺
DC can direct IL-12-in-
dependent Th1 differentiation through up-regulation of Delta 4,
which signals through the Notch receptor on T cells (47). There is
presently little information available concerning the unique mo-
lecular signatures of CD8
␣
⫹
vs CD8
␣
⫺
DC in autoimmune-resis-
tant mice. Interestingly however, CD8
␣
⫹
DC have been shown to
mediate CD4
⫹
T cell apoptosis in vitro, leading to reduced T cell
survival and restricted proliferation in comparison to the CD8
␣
⫺
subset (40, 41). We also observed that NOD and control strain
CD8
␣
⫺
DC were more effective than CD8
␣
⫹
DC at eliciting T cell
proliferation during anti-CD3 and GADp524-mediated stimula-
tion, possibly due to reduced T cell survival during activation by
the latter population. The diverse activation pathways used by DC
subsets to stimulate T cells will therefore be an important area of
further study.
The fact that IL-12 up-regulation is uncoupled from the co-
stimulatory capabilities of mature NOD CD8
␣
⫹
DC suggests that
select branches of the DC maturation program are altered in NOD
mice. A signaling mediator which could be responsible for the
variable IL-12 production by NOD and control strain DC is
MyD88, an adaptor molecule upstream of NF-
␬
B which regulates
the synthesis of IL-12 family members (48). MyD88 is essential
for cytokine production by activated DC but is not required for
MHC class II or costimulatory molecule up-regulation (49). IL-10
overproduction may also contribute to the poor IL-12 responses of
NOD CD8
␣
⫹
DC. IL-10 regulates proinflammatory cytokine pro-
duction by DC (38) and is a hallmark of tolerogenic DC which
activate regulatory T cells and are associated with diabetes pro-
tection (50). However, the putative role of DC-derived IL-10 in
immune regulation in the NOD mouse requires further clarification
because the reduced IL-12 production was not always accompa-
nied by a statistically significant increase in IL-10 production dur-
ing NOD DC maturation. Aside from their immunogenic roles,
studies of the functional capabilities of NOD DC subsets in toler-
ance induction are also warranted. CD8
␣
⫹
DC are highly special-
ized for tolerance induction, having the ability to restrict IL-2 pro-
duction by CD8
⫹
T cells (51), induce apoptosis of CD4
⫹
T cells
(40, 41), mediate cross-tolerance to cell-associated Ags (52) and
catabolize the amino acid tryptophan, which mediates immune
suppression and diabetes protection (53). The reduced CD8
␣
⫹
DC
numbers in the steady-state could contribute to the loss of self-
tolerance in NOD mice. Indeed, T1D can be prevented by adoptive
transfer of DC into NOD mice (54 –56) or by therapeutic inter-
ventions which mediate increased activity of immature or semi-
mature DC with tolerogenic properties (50, 57–59). Alternatively,
the abnormal cytokine bias of activated CD8
␣
⫹
DC from NOD
mice may be involved in the diabetes-protective effects conferred
by immune stimulation following treatment with microbial prod-
ucts, such as CFA and bacillus Calmette-Gue´rin (60, 61), or TLR
agonists, such as LPS and poly(I:C) (62, 63). In this scenario, DC
maturation in NOD mice would lead to a propensity toward tolero-
genic, IL-10-driven T cell responses and protection from T1D.
Whether immunity or tolerance prevails would therefore depend
on the balance of tolerogenic DC vs proinflammatory APC during
autoreactive T cell activation.
Although this study does not define the involvement of DC sub-
sets in vivo during the course of disease, our findings are directly
relevant to the understanding of the inherent cellular defects that
are associated with T1D susceptibility. The Idd4 locus on mouse
chromosome 11 has been associated with increased IL-12 synthe-
sis (64), differential expression of genes involved in IFN response
pathways (65), and elevated GM-CSF production in NOD mice
(66). All of these functions were associated with an increased pro-
clivity toward proinflammatory responses by NOD APC. Identifi-
cation of these autoimmune-associated phenotypes was based on
characterization of macrophages and BM-derived DC; hence, our
study illuminates the fact that these abnormalities cannot be ex-
trapolated to all NOD APC. Moreover, because the presentation of
NOD DC defects is stimulus specific, our findings argue that NOD
DC are fully functional in terms of mediating efficient T cell im-
munity, with the exception of signals that elicit IL-12-dependent
type 1 cytokine polarization. For example, NOD DC are highly
effective at inducing expansion and IFN-
␥
production by sponta-
neously primed, GADp524-reactive T cells, an effect which does
not require an IL-12-inducing stimulus. It is also noteworthy that
neither IL-12 nor its IFN-
␥
-inducing effects are required for dia-
betes pathogenesis (67– 69); therefore, the reduced IL-12-produc-
ing capability of DC may have little impact on their involvement
in autoreactive T cell activation in NOD mice.
Interestingly, although NOD mice with functionally incompe-
tent B cells are diabetes-resistant, they still harbor pathogenic T
cells (6, 70). This observation indirectly implicates DC in the ini-
tial selection of diabetogenic T cells, but suggests that the sub-
sequent expansion and/or maintenance of autoreactive T cells
relies heavily on the B cell compartment. Macrophages are not
capable of naive T cell priming and most likely have a role in
the amplification of autoreactive T cell responses, particularly
given their high IL-12 production. One possible interpretation
of our findings is that B cells and macrophages possess a func-
tional advantage among NOD professional APC owing to their
aberrantly activated state in the periphery (71) and their hyperac-
tivity during maturation (13, 14), whereas DC are resting in the
steady-state and possess a largely normal ability to activate T cells.
Importantly, however, the fact that the presentation of NOD DC
defects is context-dependent suggests that the relevance of APC
functional defects to autoimmunity will depend upon the matura-
tion signals and the local APC composition. Hence, although the
importance of one particular APC subset over another is difficult to
predict in vivo, DC likely possess complex and subset-specific
roles in T cell-mediated autoimmunity.
Acknowledgments
We express gratitude to Thomas Ichim for valuable input and to Edwin
Lee-Chan for assistance in preparation of the manuscript.
Disclosures
The authors have no financial conflict of interest.
References
1. Durant, S., V. Alves, J. Coulaud, and F. Homo-Delarche. 2002. Nonobese dia-
betic (NOD) mouse dendritic cells stimulate insulin secretion by prediabetic is-
lets. Autoimmunity 35: 449 – 455.
2. Rosmalen, J. G., F. Homo-Delarche, S. Durant, M. Kap, P. J. Leenen, and
H. A. Drexhage. 2000. Islet abnormalities associated with an early influx of
5247The Journal of Immunology

dendritic cells and macrophages in NOD and NOD
scid
mice. Lab. Invest. 80:
769 –777.
3. Dahlen, E., K. Dawe, L. Ohlsson, and G. Hedlund. 1998. Dendritic cells and
macrophages are the first and major producers of TNF-
␣
in pancreatic islets in the
nonobese diabetic mouse. J. Immunol. 160: 3585–3593.
4. Yoon, J. W., H. S. Jun, and P. Santamaria. 1998. Cellular and molecular mech-
anisms for the initiation and progression of
␤
cell destruction resulting from the
collaboration between macrophages and T cells. Autoimmunity 27: 109 –122.
5. Falcone, M., J. Lee, G. Patstone, B. Yeung, and N. Sarvetnick. 1998. B lympho-
cytes are crucial antigen-presenting cells in the pathogenic autoimmune response
to GAD65 antigen in nonobese diabetic mice. J. Immunol. 161: 1163–1168.
6. Noorchashm, H., Y. K. Lieu, N. Noorchashm, S. Y. Rostami, S. A. Greeley,
A. Schlachterman, H. K. Song, L. E. Noto, A. M. Jevnikar, C. F. Barker, and
A. Naji. 1999. I-Ag7-mediated antigen presentation by B lymphocytes is critical
in overcoming a checkpoint in T cell tolerance to islet
␤
cells of nonobese dia-
betic mice. J. Immunol. 163: 743–750.
7. Serreze, D. V., S. A. Fleming, H. D. Chapman, S. D. Richard, E. H. Leiter, and
R. M. Tisch. 1998. B lymphocytes are critical antigen-presenting cells for the
initiation of T cell-mediated autoimmune diabetes in nonobese diabetic mice.
J. Immunol. 161: 3912–3918.
8. Tian, J., D. Zekzer, Y. Lu, H. Dang, and D. L. Kaufman. 2006. B cells are crucial
for determinant spreading of T cell autoimmunity among
␤
cell antigens in di-
abetes-prone nonobese diabetic mice. J. Immunol. 176: 2654 –2661.
9. Serreze, D. V., H. D. Chapman, D. S. Varnum, M. S. Hanson, P. C. Reifsnyder,
S. D. Richard, S. A. Fleming, E. H. Leiter, and L. D. Shultz. 1996. B lymphocytes
are essential for the initiation of T cell-mediated autoimmune diabetes: analysis
of a new “speed congenic” stock of NOD.Ig
␮
null mice. J. Exp. Med. 184:
2049 –2053.
10. Silveira, P. A., E. Johnson, H. D. Chapman, T. Bui, R. M. Tisch, and
D. V. Serreze. 2002. The preferential ability of B lymphocytes to act as diabe-
togenic APC in NOD mice depends on expression of self-antigen-specific im-
munoglobulin receptors. Eur. J. Immunol. 32: 3657–3666.
11. Silveira, P. A., J. Dombrowsky, E. Johnson, H. D. Chapman, D. Nemazee, and
D. V. Serreze. 2004. B cell selection defects underlie the development of diabe-
togenic APCs in nonobese diabetic mice. J. Immunol. 172: 5086 –5094.
12. Wheat, W., R. Kupfer, D. G. Gutches, G. R. Rayat, J. Beilke, R. I. Scheinman,
and D. R. Wegmann. 2004. Increased NF-
␬
B activity in B cells and bone mar-
row-derived dendritic cells from NOD mice. Eur. J. Immunol. 34: 1395–1404.
13. Hussain, S., K. V. Salojin, and T. L. Delovitch. 2004. Hyperresponsiveness,
resistance to B-cell receptor-dependent activation-induced cell death, and accu-
mulation of hyperactivated B-cells in islets is associated with the onset of insulitis
but not type 1 diabetes. Diabetes 53: 2003–2011.
14. Hussain, S., and T. L. Delovitch. 2005. Dysregulated B7-1 and B7-2 expression
on nonobese diabetic mouse B cells is associated with increased T cell costimu-
lation and the development of insulitis. J. Immunol. 174: 680 – 687.
15. Alleva, D. G., R. P. Pavlovich, C. Grant, S. B. Kaser, and D. I. Beller. 2000.
Aberrant macrophage cytokine production is a conserved feature among autoim-
mune-prone mouse strains: elevated interleukin (IL)-12 and an imbalance in tu-
mor necrosis factor-
␣
and IL-10 define a unique cytokine profile in macrophages
from young nonobese diabetic mice. Diabetes 49: 1106 –1115.
16. Liu, J., and D. Beller. 2002. Aberrant production of IL-12 by macrophages from
several autoimmune-prone mouse strains is characterized by intrinsic and unique
patterns of NF-
␬
B expression and binding to the IL-12 p40 promoter. J. Immunol.
169: 581–586.
17. Liu, J., and D. I. Beller. 2003. Distinct pathways for NF-
␬
B regulation are as-
sociated with aberrant macrophage IL-12 production in lupus- and diabetes-prone
mouse strains. J. Immunol. 170: 4489 – 4496.
18. Marleau, A. M., and B. Singh. 2002. Myeloid dendritic cells in non-obese dia-
betic mice have elevated costimulatory and T helper-1-inducing abilities.
J. Autoimmun. 19: 23–35.
19. Poligone, B., D. J. Weaver, Jr., P. Sen, A. S. Baldwin, Jr., and R. Tisch. 2002.
Elevated NF-
␬
B activation in nonobese diabetic mouse dendritic cells results in
enhanced APC function. J. Immunol. 168: 188 –196.
20. Sen, P., S. Bhattacharyya, M. Wallet, C. P. Wong, B. Poligone, M. Sen,
A. S. Baldwin, Jr., and R. Tisch. 2003. NF-
␬
B hyperactivation has differential
effects on the APC function of nonobese diabetic mouse macrophages. J. Immu-
nol. 170: 1770 –1780.
21. Weaver, D. J., Jr., B. Poligone, T. Bui, U. M. Abdel-Motal, A. S. Baldwin, Jr.,
and R. Tisch. 2001. Dendritic cells from nonobese diabetic mice exhibit a defect
in NF-
␬
B regulation due to a hyperactive I
␬
B kinase. J. Immunol. 167:
1461–1468.
22. Wu, L., and Y. J. Liu. 2007. Development of dendritic-cell lineages. Immunity 26:
741–750.
23. Pulendran, B., J. Lingappa, M. K. Kennedy, J. Smith, M. Teepe, A. Rudensky,
C. R. Maliszewski, and E. Maraskovsky. 1997. Developmental pathways of den-
dritic cells in vivo: distinct function, phenotype, and localization of dendritic cell
subsets in FLT3 ligand-treated mice. J. Immunol. 159: 2222–2231.
24. Steinman, R. M., M. Pack, and K. Inaba. 1997. Dendritic cells in the T-cell areas
of lymphoid organs. Immunol. Rev. 156: 25–37.
25. Macagno, A., G. Napolitani, A. Lanzavecchia, and F. Sallusto. 2007. Duration,
combination and timing: the signal integration model of dendritic cell activation.
Trends Immunol. 28: 227–233.
26. Zechel, M. A., J. F. Elliott, M. A. Atkinson, and B. Singh. 1998. Characterization
of novel T-cell epitopes on 65 kDa and 67 kDa glutamic acid decarboxylase
relevant in autoimmune responses in NOD mice. J. Autoimmun. 11: 83–95.
27. Fox, C. J., and J. S. Danska. 1998. Independent genetic regulation of T-cell and
antigen-presenting cell participation in autoimmune islet inflammation. Diabetes
47: 331–338.
28. Vremec, D., J. Pooley, H. Hochrein, L. Wu, and K. Shortman. 2000. CD4 and
CD8 expression by dendritic cell subtypes in mouse thymus and spleen. J. Im-
munol. 164: 2978 –2986.
29. Nakano, H., M. Yanagita, and M. D. Gunn. 2001. CD11c
⫹
B220
⫹
Gr-1
⫹
cells in
mouse lymph nodes and spleen display characteristics of plasmacytoid dendritic
cells. J. Exp. Med. 194: 1171–1178.
30. Asselin-Paturel, C., A. Boonstra, M. Dalod, I. Durand, N. Yessaad,
C. Dezutter-Dambuyant, A. Vicari, A. O’Garra, C. Biron, F. Briere, and
G. Trinchieri. 2001. Mouse type I IFN-producing cells are immature APCs with
plasmacytoid morphology. Nat. Immunol. 2: 1144 –1150.
31. Kamath, A. T., J. Pooley, M. A. O’Keeffe, D. Vremec, Y. Zhan, A. M. Lew, A.
D’Amico, L. Wu, D. F. Tough, and K. Shortman. 2000. The development, mat-
uration, and turnover rate of mouse spleen dendritic cell populations. J. Immunol.
165: 6762– 6770.
32. Kamath, A. T., S. Henri, F. Battye, D. F. Tough, and K. Shortman. 2002. De-
velopmental kinetics and lifespan of dendritic cells in mouse lymphoid organs.
Blood 100: 1734 –1741.
33. Gately, M. K., L. M. Renzetti, J. Magram, A. S. Stern, L. Adorini, U. Gubler, and
D. H. Presky. 1998. The interleukin-12/interleukin-12-receptor system: role in
normal and pathologic immune responses. Annu. Rev. Immunol. 16: 495–521.
34. Schulz, O., A. D. Edwards, M. Schito, J. Aliberti, S. Manickasingham, A. Sher,
and C. Reis e Sousa. 2000. CD40 triggering of heterodimeric IL-12 p70 produc-
tion by dendritic cells in vivo requires a microbial priming signal. Immunity 13:
453– 462.
35. Ridge, J. P., F. Di Rosa, and P. Matzinger. 1998. A conditioned dendritic cell can
be a temporal bridge between a CD4
⫹
T-helper and a T-killer cell. Nature 393:
474 – 478.
36. Hochrein, H., K. Shortman, D. Vremec, B. Scott, P. Hertzog, and M. O’Keeffe.
2001. Differential production of IL-12, IFN-
␣
, and IFN-
␥
by mouse dendritic cell
subsets. J. Immunol. 166: 5448 –5455.
37. Gagnerault, M. C., J. J. Luan, C. Lotton, and F. Lepault. 2002. Pancreatic lymph
nodes are required for priming of
␤
cell reactive T cells in NOD mice. J. Exp.
Med. 196: 369 –377.
38. Boonstra, A., R. Rajsbaum, M. Holman, R. Marques, C. Asselin-Paturel,
J. P. Pereira, E. E. Bates, S. Akira, P. Vieira, Y. J. Liu, et al. 2006. Macrophages
and myeloid dendritic cells, but not plasmacytoid dendritic cells, produce IL-10
in response to MyD88- and TRIF-dependent TLR signals, and TLR-independent
signals. J. Immunol. 177: 7551–7558.
39. Huang, L. Y., C. Reis e Sousa, Y. Itoh, J. Inman, and D. E. Scott. 2001. IL-12
induction by a TH1-inducing adjuvant in vivo: dendritic cell subsets and regu-
lation by IL-10. J. Immunol. 167: 1423–1430.
40. Suss, G., and K. Shortman. 1996. A subclass of dendritic cells kills CD4 T cells
via Fas/Fas-ligand-induced apoptosis. J. Exp. Med. 183: 1789 –1796.
41. Rizzitelli, A., E. Hawkins, H. Todd, P. D. Hodgkin, and K. Shortman. 2006. The
proliferative response of CD4 T cells to steady-state CD8
⫹
dendritic cells is
restricted by post-activation death. Int. Immunol. 18: 415– 423.
42. Kaufman, D. L., M. Clare-Salzler, J. Tian, T. Forsthuber, G. S. Ting,
P. Robinson, M. A. Atkinson, E. E. Sercarz, A. J. Tobin, and P. V. Lehmann.
1993. Spontaneous loss of T-cell tolerance to glutamic acid decarboxylase in
murine insulin-dependent diabetes. Nature 366: 69 –72.
43. Tisch, R., X. D. Yang, S. M. Singer, R. S. Liblau, L. Fugger, and H. O. McDevitt.
1993. Immune response to glutamic acid decarboxylase correlates with insulitis
in non-obese diabetic mice. Nature 366: 72–75.
44. Summers, K. L., A. M. Marleau, J. L. Mahon, R. McManus, R. Hramiak, and
B. Singh. 2006. Reduced IFN-
␣
secretion by blood dendritic cells in human
diabetes. Clin. Immunol. 121: 81–9: 226 –229.
45. Randolph, G. J., K. Inaba, D. F. Robbiani, R. M. Steinman, and W. A. Muller.
1999. Differentiation of phagocytic monocytes into lymph node dendritic cells in
vivo. Immunity 11: 753–761.
46. Naik, S. H., D. Metcalf, A. van Nieuwenhuijze, I. Wicks, L. Wu, M. O’Keeffe,
and K. Shortman. 2006. Intrasplenic steady-state dendritic cell precursors that are
distinct from monocytes. Nat. Immunol. 7: 663– 671.
47. Skokos, D., and M. C. Nussenzweig. 2007. CD8
⫺
DCs induce IL-12-independent
Th1 differentiation through Delta 4 Notch-like ligand in response to bacterial
LPS. J. Exp. Med. 204: 1525–1531.
48. Medzhitov, R., P. Preston-Hurlburt, E. Kopp, A. Stadlen, C. Chen, S. Ghosh, and
C. A. Janeway, Jr. 1998. MyD88 is an adaptor protein in the hToll/IL-1 receptor
family signaling pathways. Mol. Cell 2: 253–258.
49. Alexopoulou, L., A. C. Holt, R. Medzhitov, and R. A. Flavell. 2001. Recognition
of double-stranded RNA and activation of NF-
␬
B by Toll-like receptor 3. Nature
413: 732–738.
50. Gaudreau, S., C. Guindi, M. Menard, G. Besin, G. Dupuis, and A. Amrani. 2007.
Granulocyte-macrophage colony-stimulating factor prevents diabetes develop-
ment in NOD mice by inducing tolerogenic dendritic cells that sustain the sup-
pressive function of CD4
⫹
CD25
⫹
regulatory T cells. J. Immunol. 179:
3638 –3647.
51. Kronin, V., K. Winkel, G. Suss, A. Kelso, W. Heath, J. Kirberg, H. von Boehmer,
and K. Shortman. 1996. A subclass of dendritic cells regulates the response of
naive CD8 T cells by limiting their IL-2 production. J. Immunol. 157:
3819 –3827.
52. Belz, G. T., G. M. Behrens, C. M. Smith, J. F. Miller, C. Jones, K. Lejon,
C. G. Fathman, S. N. Mueller, K. Shortman, F. R. Carbone, and W. R. Heath.
2002. The CD8
␣
⫹
dendritic cell is responsible for inducing peripheral self-tol-
erance to tissue-associated antigens. J. Exp. Med. 196: 1099 –1104.
5248 APCs IN NOD MICE

53. Grohmann, U., F. Fallarino, R. Bianchi, C. Orabona, C. Vacca, M. C. Fioretti, and
P. Puccetti. 2003. A defect in tryptophan catabolism impairs tolerance in nono-
bese diabetic mice. J. Exp. Med. 198: 153–160.
54. Clare-Salzler, M. J., J. Brooks, A. Chai, K. Van Herle, and C. Anderson. 1992.
Prevention of diabetes in nonobese diabetic mice by dendritic cell transfer.
J. Clin. Invest. 90: 741–748.
55. Feili-Hariri, M., X. Dong, S. M. Alber, S. C. Watkins, R. D. Salter, and
P. A. Morel. 1999. Immunotherapy of NOD mice with bone marrow-derived
dendritic cells. Diabetes 48: 2300 –2308.
56. Steptoe, R. J., J. M. Ritchie, L. K. Jones, and L. C. Harrison. 2005. Autoimmune
diabetes is suppressed by transfer of proinsulin-encoding Gr-1
⫹
myeloid progen-
itor cells that differentiate in vivo into resting dendritic cells. Diabetes 54:
434 – 442.
57. Lee, L. F., B. Xu, S. A. Michie, G. F. Beilhack, T. Warganich, S. Turley, and
H. O. McDevitt. 2005. The role of TNF-
␣
in the pathogenesis of type 1 diabetes
in the nonobese diabetic mouse: analysis of dendritic cell maturation. Proc. Natl.
Acad. Sci. USA 102: 15995–16000.
58. Naumov, Y. N., K. S. Bahjat, R. Gausling, R. Abraham, M. A. Exley,
Y. Koezuka, S. B. Balk, J. L. Strominger, M. Clare-Salzer, and S. B. Wilson.
2001. Activation of CD1d-restricted T cells protects NOD mice from developing
diabetes by regulating dendritic cell subsets. Proc. Natl. Acad. Sci. USA 98:
13838 –13843.
59. Chen, Y. G., C. M. Choisy-Rossi, T. M. Holl, H. D. Chapman, G. S. Besra,
S. A. Porcelli, D. J. Shaffer, D. Roopenian, S. B. Wilson, and D. V. Serreze. 2005.
Activated NKT cells inhibit autoimmune diabetes through tolerogenic recruit-
ment of dendritic cells to pancreatic lymph nodes. J. Immunol. 174: 1196 –1204.
60. Sadelain, M. W., H. Y. Qin, J. Lauzon, and B. Singh. 1990. Prevention of type
I diabetes in NOD mice by adjuvant immunotherapy. Diabetes 39: 583–589.
61. Harada, M., Y. Kishimoto, and S. Makino. 1990. Prevention of overt diabetes and
insulitis in NOD mice by a single BCG vaccination. Diabetes Res. Clin. Pract. 8:
85– 89.
62. Iguchi, M., H. Inagawa, T. Nishizawa, T. Okutomi, A. Morikawa, G. I. Soma, and
D. Mizuno. 1992. Homeostasis as regulated by activated macrophage. V. Sup-
pression of diabetes mellitus in non-obese diabetic mice by LPSw (a lipopoly-
saccharide from wheat flour). Chem. Pharm. Bull. 40: 1004 –1006.
63. Serreze, D. V., K. Hamaguchi, and E. H. Leiter. 1989. Immunostimulation cir-
cumvents diabetes in NOD/Lt mice. J. Autoimmun. 2: 759 –776.
64. Simpson, P. B., M. S. Mistry, R. A. Maki, W. Yang, D. A. Schwarz,
E. B. Johnson, F. M. Lio, and D. G. Alleva. 2003. Cutting edge: diabetes-asso-
ciated quantitative trait locus, Idd4, is responsible for the IL-12p40 overexpres-
sion defect in nonobese diabetic (NOD) mice. J. Immunol. 171: 3333–3337.
65. Ivakine, E. A., O. M. Gulban, S. M. Mortin-Toth, E. Wankiewicz, C. Scott,
D. Spurrell, A. Canty, and J. S. Danska. 2006. Molecular genetic analysis of the
Idd4 locus implicates the IFN response in type 1 diabetes susceptibility in nono-
bese diabetic mice. J. Immunol. 176: 2976 –2990.
66. Litherland, S. A., K. M. Grebe, N. S. Belkin, E. Paek, J. Elf, M. Atkinson,
L. Morel, M. J. Clare-Salzler, and M. McDuffie. 2005. Nonobese diabetic mouse
congenic analysis reveals chromosome 11 locus contributing to diabetes suscep-
tibility, macrophage STAT5 dysfunction, and granulocyte-macrophage colony-
stimulating factor overproduction. J. Immunol. 175: 4561– 4565.
67. Trembleau, S., G. Penna, S. Gregori, H. D. Chapman, D. V. Serreze, J. Magram,
and L. Adorini. 1999. Pancreas-infiltrating Th1 cells and diabetes develop in
IL-12-deficient nonobese diabetic mice. J. Immunol. 163: 2960 –2968.
68. Trembleau, S., G. Penna, S. Gregori, N. Giarratana, and L. Adorini. 2003. IL-12
administration accelerates autoimmune diabetes in both wild-type and IFN-
␥
-
deficient nonobese diabetic mice, revealing pathogenic and protective effects of
IL-12-induced IFN-
␥
.J. Immunol. 170: 5491–5501.
69. Zhou, W., F. Zhang, and T. M. Aune. 2003. Either IL-2 or IL-12 is sufficient to
direct Th1 differentiation by nonobese diabetic T cells. J. Immunol. 170:
735–740.
70. Greeley, S. A., D. J. Moore, H. Noorchashm, L. E. Noto, S. Y. Rostami,
A. Schlachterman, H. K. Song, B. Koeberlein, C. F. Barker, and A. Naji. 2001.
Impaired activation of islet-reactive CD4 T cells in pancreatic lymph nodes of B
cell-deficient nonobese diabetic mice. J. Immunol. 167: 4351– 4357.
71. Chiu, P. P., A. M. Jevnikar, and J. S. Danska. 2001. Genetic control of T and
B lymphocyte activation in nonobese diabetic mice. J. Immunol. 167:
7169 –7179.
5249The Journal of Immunology