| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Neuropharmacology, Division of Virology, The Scripps Research Institute, La Jolla, CA 92037;
Alliance Pharmaceutical Corporation, San Diego, CA 92121;
Department of Immunology and Infectious Diseases, Harvard School of Public Health, Boston, MA 02115; and Department of Medicine, Harvard Medical School, Boston, MA 02115
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
secretion by
T lymphocytes. A shift from Th1 to Th2 T cell immunity was not
observed. Hence, our results implicate both CTL and cytokines in ß
cell destruction. Selective inhibition of the STAT4 signal transduction
pathway might constitute a novel and attractive approach to prevent
clinical insulin-dependent diabetes mellitus in prediabetic individuals
at risk. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In the present study we investigated the role of the STAT4 signaling pathway on the pathogenesis of insulin-dependent diabetes mellitus (IDDM)3 using the rat insulin promoter (RIP) lymphocytic choriomeningitis virus (LCMV) transgenic mouse model of T cell-mediated autoimmune diabetes (11, 12, 13). In this system, LCMV proteins are expressed under control of the RIP in the ß cells of the pancreas. RIP-LCMV animals do not develop spontaneous diabetes. Instead, IDDM is triggered at a high rate subsequent to infection with LCMV (11, 12). The virus initiates an immune response that leads to its clearance, but the activated antiviral/"self reactive" lymphocytes are also infiltrating into the pancreatic islets, where they initiate destruction of ß cells. CD8+ CTL are the main trigger for this autoreactive process (13). Two different transgenic RIP-LCMV lines have been established that are distinguished by the requirement of CD4+ T cell help to develop autoimmune diabetes: RIP-LCMV-GP transgenic mice expressing the LCMV glycoprotein (GP) in the pancreas develop IDDM independently of CD4+ T helper cells. In contrast, RIP-LCMV-NP mice express the LCMV nucleoprotein (NP) under control of the RIP and develop diabetes only in the presence of CD4+ T cell help (13). A reason for this CD4+ T lymphocyte dependence is the coexpression of the self/viral Ag in the thymus of RIP-LCMV-NP mice leading to deletion of high, but not low, affinity CTL that require CD4+ T cell help to cause disease (13, 14). Since human IDDM is associated with both CD4+ and CD8+ autoaggressive T cell responses, initiation and prevention of IDDM in the CD4+ T cell-dependent model is more relevant for the pathogenesis of human IDDM.
We find that absence of STAT4 signaling provides protection from
diabetes in the CD4+ T cell-dependent
RIP-LCMV-NP, but not in the CD4+ T
cell-independent RIP-LCMV-GP model for IDDM. The protection is
associated with a significant reduction of CTL precursors and IFN-
production by T cells, but not with systemic Th2 deviation. Since
antiviral immunity is retained in STAT4-deficient RIP-LCMV-NP mice and
autoreactive (ß cell-specific) CTL activity is not abrogated, our
results implicate the action of inflammatory cytokines in addition to
direct CTL killing for ß cell destruction.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
STAT4-/- and STAT6-/- mice (5, 15) (both generated on the B6 (H-2b) and BALB/c (H-2d) genetic backgrounds) were bred to RIP-LCMV-GP (H-2b) and RIP-LCMV-NP (H-2d) (11, 13) transgenic mice. The offspring of the F1 generation were intercrossed, and the resulting F2 generation was used in this study. (RIP-LCMV x STAT4) and (RIP-LCMV x STAT6) animals were genotyped either by Southern blotting (5, 15) or PCR analysis of tail DNA. For screening by PCR, two independent standard PCR reactions were performed to determine whether the animals carried wild-type and/or neomycin (neo) gene insertion (i.e., disrupted) copies of the stat4 and stat6 genes. Accordingly, (RIP-LCMV x STAT4) animals were genotyped with the primer pairs STAT4 sense (5'-CCTACTGGCAGAGAGTCTTTTCC-3')/STAT4 antisense (5'-GGTTGTAGATCAGGAAGGTAGC-3') and with primer pairs STAT4 sense/neo-sense (5'-GGATTGCACGCAGGTTCTCCG-3'). (RIP-LCMV x STAT6) mice were genotyped using the primer pairs STAT6 sense (5'-GGCCGAGGCTTCACATTTTGGC-3')/STAT6 antisense (5'-CCCGGATGACGTGTGCAATGG-3') and neo-sense/neo-antisense (5'-CCGGCCACAGTCGATGAATCC-3').
The presence of the RIP-LCMV-NP and RIP-LCMV-GP transgenes was determined by slot-blot hybridization or PCR as described previously (16).
Induction of IDDM in transgenic RIP-LCMV animals and blood glucose measurements
Virus stocks used in this study consisted of LCMV Armstrong (Clone 53b). Viruses were plaque purified three times on Vero cells and stocks were prepared by a single passage on BHK-21 cells. To induce IDDM, (RIP-LCMV-GP/RIP-LCMV-NP x STAT4/STAT6) animals were infected i.p. with 1 x 105 pfu LCMV at 5–10 wk of age. Each IDDM induction experiment was repeated 4–6 times, and animals from different experimental groups were age matched for each experiment.
After LCMV infection, mice were examined weekly for signs of IDDM by
determining the blood glucose concentrations. Blood samples were
obtained from the retroorbital plexus, and glucose levels were
determined using the ACCUCHECK III system (Boehringer Mannheim,
Indianapolis, IN) as previously described (13). Animals
that showed blood glucose levels 
300 mg/dl during two consecutive
measurements were considered diabetic.
Immunohistochemistry
Immunohistochemistry was performed on 8-µm thick cryomicrotome sections of the pancreas fixed with ice-cold 95% ethanol as described (13). Tissue sections were blocked with 2% FBS in PBS and avidin/biotin blocking solutions (Vector, Burlingame, CA). Specific cell types were detected by a complex consisting of a primary rat Ab (directed against mouse CD8 (Ly-2 and Ly-3.2; PharMingen, La Jolla, CA), mouse CD4 (L3T4; PharMingen), or mouse F4/80 (Serotec, Raleigh, NC)), a secondary biotinylated anti-rat Ab (Vector), and streptavidin-HRP (Vector). After diaminobenzidine (Zymed, San Francisco, CA) staining, slides were counterstained with hematoxylin (Sigma, St. Louis, MO) and embedded in AquaMount (Fisher) before analysis and photography. For each experimental group, we analyzed serial sections from 5–10 animals originating from at least three independent IDDM induction experiments.
Primary and secondary CTL assays
For primary (i.e., ex vivo) CTL assays, effector splenocytes were obtained from (RIP-LCMV-NP/GP x STAT4) mice 7 days after infection with 105 pfu LCMV i.p. CTL activity was determined by in vitro 51Cr release assay (13). Target cells were either MC57 (H-2b) or BALB/cl7 (H-2d) fibroblasts and were labeled for 1 h with 51Cr (Amersham, Arlington Heights, IL). Subsequently, syngeneic (MHC-matched) and allogeneic (MHC-mismatched, i.e., negative control) target cells were coated with LCMV-specific peptides. Accordingly, for assessment of CTL responses in (RIP-LCMV-NP x STAT4) (H-2d) animals, targets were coated with 10 µg/ml H-2d (Ld)-restricted LCMV peptide NP aa 118–126 (17). To determine cytolytic activity mounted by (RIP-LCMV-GP x STAT4) (H-2b (Db)-restricted) splenocytes, a mixture of the LCMV peptides GP aa 33–41 and GP aa 276–286 (each 10 µg/ml) was used for target cell coating (18).
Primary CTL assays were performed in 96-well flat-bottom plates with 1 x 104 target cells per well with E:T cell ratios of 50:1, 25:1, 12.5:1, and 6.25:1. Each sample was run in triplicate with an SE below 15%. After 5 h of incubation, 51Cr release was determined in the culture supernatant using a gamma counter. Spontaneous (no effector cells added) and total (addition of 1% Nonidet P-40 to cultures) 51Cr release was determined in parallel, allowing the assessment of cell lysis for each culture as described (11, 13). Specific 51Cr release was calculated by subtracting 51Cr release with peptide noncoated from that with peptide-coated target cells.
For secondary CTL assays, effector cells were cultured in vitro before cytolytic activity was assessed. Splenocytes were obtained from (RIP-LCMV-NP x STAT4) animals 7 days after LCMV infection (i.p. 105 pfu) and were cultured at 2 x 104/well in the presence of 2 U/ml recombinant human IL-2, 2 µg/ml LCMV-NP aa 118–126 peptide, and irradiated (2000 rad) MHC-matched spleen cells (105/well) for 5 days. Effector cells were added to 51Cr-labeled MHC-matched and MHC-mismatched targets, and cytotoxicity was assessed and evaluated as described above for primary CTL assays. E:T cell ratios were 5:1, 2.5:1, and 1:1, and the analysis was conducted in triplicate for each culture. Background values were less than 15% of total release.
Cytotoxicity against islet cells
Pancreatic islets were harvested from H-2d STAT4+/+ and STAT4-/- mice (5) 7 days postinfection with 105 pfu LCMV i.p. and isolated after fractionated collagenase P (Boehringer Mannheim) digestion as described (19). Whole islet cells were labeled in glucose-free medium for 1 h with 51Cr (Amersham) (5% CO2, 37°C, FBS in medium). After trypsinization (5 min at 37°C), dispersed islet cells (2 x 104 per well) either were coated with 1 µg/ml LCMV-NP aa 118–126 H-2d peptide or remained free of peptide. Syngeneic effector cells were obtained from the spleen of the same animals and added to the labeled islet cells at an E:T ratio of 50:1. Assays were conducted for 20 h overnight. Supernatants were harvested, and 51Cr release was determined. Three to four animals were used for each experimental group. Two independent experiments were performed.
Determination of CTL precursor frequency
Splenocytes were harvested 15 days after infection with 105 pfu LCMV i.p. as described previously (13). Dilutions of responder cells, beginning with 1 x 105 cell/well, were incubated in 96-well flat-bottom plates in the presence of 2 U/ml recombinant human IL-2, irradiated (2000 rad) MHC-matching spleen cells (105/well), and 2 µg/ml H-2d-restricted peptide LCMV-NP aa 118–126. For each dilution, 24 microcultures were set up. After 5 days, individual microcultures were assayed for cytotoxicity. MHC-matched BALB/c7 fibroblasts were used as target cells and were 51Cr labeled. Target cells (104 cells/well) were coated with 10 µg/ml LCMV-NP aa 118–126 (H-2d) peptide or remained uncoated. Twenty wells of peptide-coated target cells and four wells of uncoated target cells (used to determine the background 51Cr release) were plated for each responder dilution. Subsequently, in vitro cultured responder cells were added to target cells, and 51Cr release of targets was determined after 5 h of incubation using a gamma counter. The fraction of negative cultures (% lysis <3 SEs above background) was determined for each dilution of responder cells and plotted on a semilogarithmic scale against the number of responder cells per well. Calculation of the precursor CTL frequency by evaluation of the slope of each plot was performed as previously described (13). Two independent experiments were done with three to four animals in each group.
Viral titers
LCMV viral titers of organ homogenates and sera were determined by infection of Vero cells as described (13). Tissues (spleen, kidney, and brain) and sera were obtained from (RIP-LCMV-NP x STAT4), (RIP-LCMV-GP x STAT4), or STAT4 animals at 5, 14, and 63 days after infection with 105 LCMV i.p. Homogenates and sera were diluted serially, and viral titers were calculated from the number of counted plaques (five to six animals per group).
Influenza virus titer was assessed in lungs harvested from (RIP-LCMV-NP x STAT4+/+) and (RIP-LCMV-NP x STAT-/-) animals 7 days after intranasal infection with 104 TCID50 of A/PR/8/34 influenza virus (H1N1) (20). Mice were immunized i.p. with 103 TCID50 of A/HK/68 (H3N2) 1 mo before the H1N1 challenge. Homogenized tissues were serially diluted and incubated with MDCK cells that had been briefly washed with Trypsin-EDTA (Life Technologies, Gaithersburg, MD). After 1 h of incubation at 37°C, 150 µl/well of DMEM supplemented with 10% FCS was added to each well. Plates were incubated for 48 h. The supernatants were harvested and incubated in 96-well U-bottom plates at room temperature with a suspension of chicken RBC for 30–45 min. Hemagglutination of erythrocytes indicated presence of virus, allowing the calculation of viral titers in the lungs of infected mice. Individual measurements were conducted in triplicate, and four animals were analyzed per group.
Cytokine assays
ELISA.
Lymphocytes recovered from spleen or pancreas were stimulated in vitro
for 24 to 96 h in 96-well flat-bottom microtiter plates as
indicated in the figure legends. Culture supernatants were assayed for
cytokines (IL-4, IL-10, TNF-
, and IFN-) by ELISA assays
(21) using Ab pairs all obtained from PharMingen unless
differently noted. Splenocytes (3 x
105/well) were stimulated either with 10 µg/ml
MHC class I-restricted LCMV NP H-2d aa 118–126
peptide (RIP-LCMV-NP x STAT4) line or with the recently described
MHC class II-restricted LCMV-specific H-2b peptides GP aa
61–80 and NP aa 309–328 (22, 23, 24) (each 10 µg/ml;
(RIP-LCMV-GP x STAT4) line) in the presence or absence of
recombinant mouse IL-12 (50 ng/ml; PharMingen). Lymphocytes recovered
from pancreas (25) were polyclonally stimulated by
plate-bound anti-CD3 mAb (10 µg/ml) in the presence of
anti-CD28 mAb (2 µg/ml) and recombinant human IL-2 (10 U/ml;
PharMingen). For ELISA assays, 96-well ELISA U-bottom plates were
coated with anti-mouse IFN- (R4-6A2), IL-4 (11B11), IL-10
(JES5-2A5), or TNF- (G-281-2626) mAbs overnight at 4°C. Plates
were washed vigorously and blocked with 10% FBS. Culture supernatants
(100 µl) were incubated for 2 h at room temperature. Following
several extensive washes, specifically bound cytokines were detected by
a second biotinylated anti-mouse mAb matched with the coating
cytokine Ab (anti-IFN- (XMG1.2); anti-IL-4 (BVD6-24G2);
anti-IL-10 (JES5-16E3); and anti-TNF- (MP6-XT3)) and
streptavidin-HRP complex (Boehringer Mannheim). Plates were stained
using ABTS (2,2'-azino-di-[3-ethylbenzthiazolin-6-sulfonic acid];
Sigma) as a substrate and analyzed by an EL-800 microtiter ELISA plate
reader (Biotek, Winooski, VT) at 405 nm using the DeltaSoft 3 software
package (BioMetallics, Princeton, NJ). Serial dilutions of the
corresponding recombinant mouse cytokine served as positive controls
and standards (IFN-, IL-4, IL-10, and TNF-; all obtained from
PharMingen). Multiple negative control wells were incubated with
blocking solution to determine background levels. The detection limits
for cytokines were 5–10 pg/ml. At least three animals per group and
experiment were studied.
ELISPOT assay. Splenocytes were harvested from (RIP-LCMV-GP STAT4+/+) and (RIP-LCMV-GP STAT4-/-) mice infected with 105 pfu LCMV i.p. 8 days previously. Cells were plated at 5 x 105 cells/well and 3-fold serial dilutions were set up in 96-well nitrocellulose plates (Millipore, Bedford, MA) coated with 4 µg/ml anti-mouse IL-4 Ab (clone 11B11; PharMingen). Stimulation was provided by a mixture of LCMV-specific MHC class II-restricted GP aa 61–80 and NP aa 309–328 peptides (each 2 µg/ml) (22, 24) in the presence of 5 x 105 irradiated syngeneic feeder cells/well as well as recombinant human IL-2 (10 U/ml; PharMingen). Background controls were splenocytes originating from the same animals incubated without peptide, but in the presence of IL-2. Three mice per group were studied, and each culture was conducted in quadruplicate. After 36 h of incubation, unbound cells were washed off, and the assay was developed as described previously (24, 26). In brief, plates were incubated with biotinylated anti-mouse IL-4 Ab (clone BVD6-24G2; PharMingen) in PBS supplemented with 0.1% FBS and 0.1% Tween 20 overnight at 4°C. After washing with buffer and incubation with a streptavidin-HRP complex (Boehringer Mannheim), bound IL-4 was visualized as described using 3-amino-9-ethyl-carbozole (Sigma) as substrate. The number of spots was determined for each well, and the average for each culture dilution was calculated. The SE among the quadruplicates was less than 15%. After background subtraction (less than 25 spots among 5 x 105 splenocytes), the frequency of IL-4-producing T lymphocytes restricted by MHC-class II peptides was evaluated by interpolation.
Intracellular cytokine analysis by FACS.
Cultured lymphocytes (obtained from (RIP-LCMV-NP x STAT4) mice 7
days after infection with 105 pfu LCMV i.p.) were
phenotyped by FACS analysis utilizing color-conjugated mAbs (from
PharMingen unless stated differently) specific for the murine cell
surface markers CD4 (L3T4, Cy-chrome-conjugated) and CD8 (Ly-2,
APC-conjugated) and murine IFN- (XMG1.2, PE-conjugated).
Intracellular cytokine analysis was done as described
(21). In brief, 5 x 105
splenocytes were polyclonally stimulated for 12 h with 2 µg/ml
anti-CD28 in 96-well flat-bottom plates precoated with 10 µg/ml
anti-CD3. The Golgi-blocker brefeldin A (Sigma) was added to the
cultures at 5 µg/ml to prevent cytokine secretion. Cells were then
transferred to 96-well V-bottom plates and stained for cell surface
markers using anti-CD4 and anti-CD8 Abs in phosphate buffer
containing 1% FBS and 0.1% sodium azide. After washing, cells were
fixed and permeabilized in HBSS containing 4% paraformaldehyde, 0.1%
saponin, and 10 mM HEPES. Subsequently, cells were stained with
monoclonal anti-IFN- Ab in phosphate buffer containing 0.1%
saponin/1% FBS/0.1% sodium azide for 1 h on ice in the dark.
After washes with phosphate buffer, cells were analyzed using a
FACScalibur apparatus and CellQuest software (both Becton Dickinson,
San Jose, CA). Three different animals were tested for each study.
RNase protection assay.
Total RNA extracted from spleens (27) of (RIP-LCMV-NP
x STAT4) mice was analyzed by RNase protection assay (RPA) using the
RiboQuant system (PharMingen) according to the manufacturer’s
recommendations. In brief, cytokine-specific RNA probes were labeled in
vitro by including (-32P)UTP (ICN, Costa Mesa,
CA) in the synthesis mixture. Radioactively labeled probes were
hybridized to 20 µg total RNA overnight and digested by a mixture of
RNase A/T1. After phenol extraction, samples were run on a sequencing
gel (28), and protected RNA fragments were visualized
utilizing a Molecular Dynamics PhotoImager (Sunnyvale, CA). Levels of
cytokine RNA were quantitated using the Optiquant Image Analysis
Software (Packard, Meriden, CT). They were normalized against the
levels of L32 RNA. Controls included in each experiment were yeast tRNA
to determine background levels and (as positive standards) cytokine
RNAs obtained from the manufacturer (PharMingen). Each experimental
group consisted of three to four mice, and the experiment was
reproduced twice.
Purification of anti-mouse IL4 Ab and in vivo IL-4 neutralization
The rat anti-mouse IL-4 hybridoma (clone 11B11; Ref. 29) was obtained from American Type Culture Collection (ATCC, Manassas, VA) and expanded in culture. Supernatants harvested from confluent cultures were precipitated using ammonium sulfate according to established protocols (30). The anti-IL-4 Ab was then affinity purified using a protein G column (Pharmacia, Uppsala, Sweden) and dialyzed against phosphate buffer. The protein content of the concentrated Ab fraction was determined (Bio-Rad, Hercules, CA).
During each inoculation, animals received 200 µl sterile PBS i.v. containing 0.5 mg purified anti-IL-4 Ab. Ab injection was conducted twice per week for a total of 4 wk. Five (RIP-LCMV-NP x STAT4-/-) mice that did not develop diabetes 63 days post LCMV infection were utilized for this experiment. Animals were monitored weekly for clinical onset of IDDM for a total of 8 wk post Ab application.
Statistical analysis
Data analyses requiring Fisher’s exact test (rate of diabetes among different groups), linear regression (precursor CTL (pCTL) frequency, ELISPOT evaluation), and SE calculations were performed using InStat software (GraphPad Software, San Diego, CA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We observed a significant reduction in the incidence of diabetes
in STAT4-deficient animals in the CD4+ T
cell-dependent (13) RIP-LCMV-NP model (Fig. 1
; 85% of STAT4+/+
(n = 22) vs 30% of STAT4-/-
(n = 30) developed diabetes, p <
0.0001 according to Fisher’s exact test). Protection from diabetes in
STAT4-deficient mice was accompanied by a reduced degree of insulitis
and CD8+ T cell infiltration in (RIP-LCMV-NP
x STAT4-/-) mice compared with
STAT4+/+ littermates (Fig. 2). Likewise, the numbers of
CD4+ T lymphocytes and macrophages infiltrating
into the islets of (RIP-LCMV-NP x STAT4-deficient) animals were
decreased when compared with (RIP-LCMV-NP x STAT4-competent) mice
(data not shown).
|
|
Reduction of diabetes was specific to ablation of STAT4 function since genetic deletion of the STAT6 signaling pathway (15) did not have an effect on IDDM onset or incidence in RIP-LCMV-GP or RIP-LCMV-NP animals after challenge with LCMV.4 Unlike STAT4, which is involved in the differentiation of Th1 cells, STAT6 is specifically activated by IL-4/IL-4R interaction and promotes the generation of Th2 responses (15, 31, 32).
These results indicate that the STAT4 signaling pathway plays an important role in regulating autoimmune diabetes. The STAT4 molecule is important when low affinity CTL require the help of CD4+ T cells (13, 14) for the initiation and progression of the autoimmune process. Thus, our finding may have importance for human diabetes associated with both CD8+ and CD4+ T lymphocytes.
In the absence of STAT4 antiviral ("anti-self") CTL responses are generated that confer protection against LCMV in vivo and are able to lyse islet cells in vitro
To assess the ability of STAT4-deficient mice (33) to
overcome LCMV infection, we compared the viral titers in
(RIP-LCMV-NP x STAT4-/-) and
(RIP-LCMV-NP x STAT4+/+) animals in spleen,
kidney, and brain and in the serum at various time points after
infection (Fig. 3A). LCMV was
cleared from all virus-infected RIP-LCMV-NP animals independently of
the STAT4 genotype with similar kinetics. Complete viral clearance from
spleen, kidney, brain, and serum was also documented in
(RIP-LCMV-GP x STAT4-/-) mice or C57BL/6
STAT4-/- mice (5) with kinetics similar to
that of STAT4-competent littermates (data not shown).
|
Next we tested the ability of (RIP-LCMV-NP x STAT4-deficient)
animals to generate LCMV-specific CTL responses. In agreement with the
observed viral clearance, primary (ex vivo; Fig. 3
B) and
secondary (in vitro stimulated; Fig. 3C) CTL
activities were documented in the absence of STAT4 signaling
in RIP-LCMV-NP mice infected 7 days previously with LCMV. The levels of
LCMV-specific primary and memory CTL responses mounted by
(RIP-LCMV-NP x STAT4-/-) animals were
comparable to those of STAT4-competent littermates (Fig. 3, B and C). Furthermore, LCMV-specific CTL
generation was also assessed in (RIP-LCMV-GP x STAT4) mice 7 days
after LCMV infection (105 pfu i.p.; data not
shown). Similar to results obtained in (RIP-LCMV-NP x STAT4)
animals, RIP-LCMV-GP mice deficient in STAT4 mounted virus-specific CTL
activity indistinguishable from STAT4-competent littermates (data not
shown). This finding is in agreement with the high incidence of IDDM
observed in STAT4-deficient RIP-LCMV-GP mice.
Our results indicate that IDDM is significantly reduced through
disruption of the STAT4 signaling pathway while antiviral immune
responses are maintained. This finding is in marked contrast to earlier
observations using IFN--deficient mice crossed to RIP-LCMV-GP and
RIP-LCMV-NP transgenic mice (16). Although IDDM was
abrogated, these animals were clearly immunocompromised, and
viral doses higher than 103 pfu were lethal,
which is up to three orders of magnitude lower than the dose
STAT4-deficient, as well as STAT4-competent, mice are able to clear
(1 x 105 pfu) (35).
The high levels of anti-self (LCMV) CTL found in (RIP-LCMV-NP
x STAT4-deficient) mice prompted us to investigate whether
STAT4-deficient CTLs were also able to recognize and lyse pancreatic
islet cells. Seven days after LCMV challenge, islet cells and
splenocytes were collected from STAT4+/+ and
STAT4-/- BALB/c mice. The islets were
51Cr labeled, coated with LCMV-specific NP aa
118–126 peptide, and used as target cells in a 20-h in vitro CTL
assay. Syngeneic splenocytes obtained from the same animals were tested
for their ability to lyse the islet cells (Fig. 3D). A
similar degree of cell lysis was noted when primary
anti-"self"/anti-LCMV CTL activities were compared directed
against peptide-coated islet cells of STAT4-deficient and -competent
animals (Fig. 3D). This in vitro finding, combined with the
reduction of IDDM in STAT4-deficient RIP-LCMV-NP mice, suggests that in
vivo anti-"self"/anti-LCMV primary CTL are not capable of
directly destroying enough islet cells to cause IDDM and that other
factors are likely to be required. This is supported by our earlier
observation showing no IDDM development after adoptive transfer of high
numbers of LCMV-specific lymphocytes into RIP-LCMV-GP and RIP-LCMV-NP
recipients (36).
Development of autoimmune diabetes is associated with reduced
levels of IFN- and decreased numbers of autoreactive CTL precursors
Since islet cells are sensitive to cytokines and since several
cytokines generated during LCMV infection, such as type I and II IFNs,
are able to up-regulate MHC molecules on ß cells (16, 19), we next analyzed IFN- levels in STAT4-deficient mice in
response to IL-12 and LCMV peptides. Splenocytes recovered from
(RIP-LCMV-NP x STAT4-deficient) mice were able to generate
reduced and delayed, but still significant, amounts of IFN- after
stimulation with LCMV-NP-specific class I peptides (Fig. 4A). However, as expected,
(RIP-LCMV-NP x STAT4-/-) splenocytes were
unresponsive to recombinant mouse IL-12, regardless of whether the
cytokine was added to the cultures alone or in combination with LCMV
peptide (Fig. 4A). Similarly, MHC class II-restricted
LCMV-NP and LCMV-GP peptides were able to elicit significant production
of IFN- by splenocytes obtained from (RIP-LCMV-GP x
STAT4-deficient) animals (Fig. 4B). However, IFN-
production of MHC class II-restricted (RIP-LCMV-GP x
STAT4-/-) splenocytes was 2- to 5-fold reduced
when compared with STAT4-competent littermates. As previously noted
during the MHC class I LCMV peptide experiments, all generation of
IFN- by MHC class II-restricted lymphocytes deficient in STAT4
occurred independent of IL-12 (Fig. 4, A and B).
Supernatants of the cultures described in Fig. 4, A and
B, were also tested for TNF- and the Th2 cytokines IL-4
and IL-10. Equal quantities of TNF- were produced by class I and II
LCMV peptide-stimulated splenocytes of (RIP-LCMV-NP x
STAT4-deficient) and (RIP-LCMV-NP x STAT4-competent) littermates
(data not shown). IL-4 and IL-10 were undetectable in the supernatants
of cultures from splenocytes stimulated in the presence of MHC class I-
and II-restricted LCMV peptides (data not shown; cytokine detection
limits were approximately 5–10 pg/ml). Similar to these findings using
ELISAs, intracellular cytokine staining and FACS analysis of
polyclonally stimulated lymphocytes recovered from (RIP-LCMV-NP x
STAT4) animals (7 days post LCMV infection; 105
pfu, i.p.) showed that the overall number of IFN--producing
lymphocytes was decreased in STAT4-/- animals
(8.5% of CD8+ lymphocytes in STAT4-deficient
mice vs 29% of CD8+ lymphocytes in
STAT+/+ controls; representative of two
experiments, each with three animals).
|
, B and D), (RIP-LCMV-NP x
STAT4-/-) compared with
STAT4+/+ littermates showed a 5-fold reduction in
the precursor frequency of LCMV-specific secondary (e.g. memory) CTL 15
days after infection (Fig. 4C). The reduction in the
frequency of LCMV-specific, and hence "auto"-reactive, CTLs likely
contributes to the inability of (RIP-LCMV-NP x
STAT4-/-) mice to develop IDDM. However, it
cannot be the only factor since abrogation of disease is usually seen
only at pCTL frequencies lower than 1/5500 or less than 9000 CTLs per
spleen (37). In support of this notion, significant
numbers of CD8+ T lymphocytes were even present
in islets of nondiabetic transgenic (RIP-LCMV-NP x
STAT4-/-) mice. However, the degree of
infiltration of CD8+ T lymphocytes into the
islets was reduced when compared with (RIP-LCMV-NP x
STAT4+/+) mice (Fig. 2, B compared
with A). Similar results were obtained when pancreatic
sections were stained for CD4+ T lymphocytes, as
well as macrophages (data not shown). These findings demonstrate that
inactivation of the IL-12/STAT4 signaling pathway does not prevent
infiltration of autoreactive lymphocytes into the target organ. Indeed,
lymphocytic infiltration into the islets of nondiabetic
STAT4-/- mice is detected as late as 65 days
postinfection (Fig. 2). Hence, the STAT4 signaling pathway appears to
regulate the transition from nondestructive insulitis to complete ß
cell destruction and clinical IDDM, a process that appears to be
dependent on IFN- and probably other factors (38). Lack of Th1 to Th2 shift in the absence of STAT4 subsequent to LCMV infection
Lastly, to discriminate between a local
"bystander"-suppression mediated by Th2 cytokines (1, 39) vs a quantitative decrease of the immune response in the
absence of STAT4, we determined the expression of cytokines by
lymphocytes recovered from the pancreas of RIP-LCMV-NP mice expressing
or lacking STAT4. As shown in Fig. 4D, these cells were able
to produce significant amounts of IFN- when stimulated with
anti-CD3 and anti-CD28 Abs. However, as observed before for
splenic lymphocytes (Fig. 4, A and B), the
production of IFN- was reduced when compared with STAT4-competent
pancreatic lymphocytes.
We next tested whether the absence of STAT4 resulted in a shift toward
a Th2 response. (RIP-LCMV-NP x STAT4-competent) and
(RIP-LCMV-NP x STAT4-deficient) animals had comparable production
of IL-4 and IL-10 by pancreatic lymphocytes after polyclonal
stimulation with anti-CD3 and anti-CD28 Abs as analyzed by
ELISA (data not shown). To confirm this observed lack of Th1 to Th2
shift, we performed a more sensitive assessment of cytokine expression
by using RPAs (40) and ELISPOT (24, 41)
assays (Table I). Our findings clearly
show a significant decrease in IFN- in STAT4-deficient mice and no
alteration of Th2 (IL-4) responses.
|
300 mg/dl). However, all (5/5)
(RIP-LCMV-NP x STAT4-/-) mice receiving
anti-IL-4 Ab remained free of diabetes, showing blood glucose
values <200 mg/dl during a 2-mo observation period after initiation of
the Ab application (data not shown). Taken together, these findings
strongly suggest that a quantitative decrease of the Th1/Tc1 immune
response, rather than a qualitative shift toward Th2, accounted for the
decreased incidence of IDDM in (RIP-LCMV-NP x STAT4-deficient)
mice. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A). Virally infected APCs (43) were
therefore capable of activating primary virus-specific
CD4+ and CD8+ T
lymphocytes, promoting their differentiation toward Th1 and Tc1 without
the requirement of IL-12 signaling. The primary antiviral CTL generated
in STAT4-/- mice, in addition to their ability
to clear LCMV infection (Ref. 44 and Fig. 3A),
also exhibited in vitro a high degree of STAT4-independent cytotoxicity
directed against the islet cells targeted by the autoimmune process
(Fig. 3D). Therefore, since IDDM is significantly reduced in
(RIP-LCMV-NP x STAT4-/-) mice, other
factors acting after the primary induction of autoreactive lymphocytes
are required in vivo to manifest autoimmune diabetes. Thus, our
observations suggest a two stage process for the pathogenesis of IDDM.
During a first phase, autoreactive lymphocytes are induced and
infiltrate into the target organ. This alone is insufficient to cause
disease. The second stage is characterized by the local amplification
of autoreactive lymphocytes accompanied by local Th1/Tc1 cytokine
secretion resulting in complete ß cell destruction. It is mainly this
second stage that is affected by the absence of the STAT4 molecule. Our
findings show that the precursor frequency of CTL present in
virus-infected (RIP-LCMV-NP x STAT4-/-)
mice is decreased (Fig. 4), and IFN- production by autoreactive T
lymphocytes in response to MHC I and II self Ags is reduced (Fig. 3).
Consequently, fewer CD4+ and
CD8+ T lymphocytes and macrophages infiltrate
into the pancreatic islets of (RIP-LCMV-NP x
STAT4-/-) after LCMV challenge (Fig. 2). Thus,
although CTL trigger the development of diabetes, they are not
completely sufficient to destroy all islet cells in vivo. Islet cell
death might occur indirectly through the presence of inflammatory
cytokines as a "bystander" death. Indeed, MHC class I-restricted
CTL clones do not kill islets in vitro to the same degree to which
cytokines alone do (IFN- plus TNF- or IFN- plus IL-1ß;
M. G. von Herrath, unpublished observations). This opens up a
window for therapeutic approaches during the second phase of IDDM,
which does not need to affect the systemic generation of autoreactive
lymphocytes.
CD4+ T cell-dependent disease mediated by low
affinity CTL in RIP-LCMV-NP animals, but not CD4+
T cell-independent IDDM, was prevented in the absence of STAT4. This
finding is important because it illustrates the fine-tuned kinetic
interplay of various inflammatory factors and Th1/Tc1 lymphocytes
required for transforming autoreactivity into clinical disease. In good
agreement with this finding is a recent report demonstrating that STAT4
signaling is primarily required for IFN- production and priming of
CD4+, but not CD8+, T
lymphocytes (45). Correspondingly, when
CD4+ T cells are required to develop IDDM, STAT4
is a significant component (see Fig. 1). In contrast, STAT4 signaling
is not required for development of diabetes when only
CD8+ T cells are needed. In the absence of large
numbers of self reactive CTL, autoimmune diabetes develops only if
CD4+ T cell help and a sufficient number of
inflammatory mediators such as IFN- and/or IL-12 are present. In
this situation, a quantitative decrease of one or several inflammatory
components rather than their complete absence can have profound effects
on the autoimmune process while the overall immune competence is
maintained. Through these considerations, our approach gains
therapeutic significance.
Importantly, the local cytokine milieu in the pancreas was changed
quantitatively but not qualitatively in STAT4-deficient RIP-LCMV-GP and
RIP-LCMV-NP mice. We found less IFN- but no increased production of
IL-4 (Table I). This is partly in contrast to the previously described
increase of Th2 responses in STAT4-deficient mice (5).
However, such studies analyzed overall CD4+ T
lymphocyte function and not LCMV-specific responses or lymphocyte
activities localized to a specific compartment, such as the pancreatic
islets (5, 33). Interestingly, prevention of IDDM without
augmenting Th2 activity is clearly different from observations in
single-transgenic RIP-LCMV-NP mice (STAT4-competent) treated orally
with insulin to prevent IDDM (26). In this scenario,
secretion of IFN- in the islets is not reduced, but production of
IL-4, IL-10, and TGF-ß is significantly increased due to the
induction of "bystander suppressor" lymphocytes specific for the
insulin B chain (26). Thus, different possibilities exist
to change the local pancreatic cytokine milieu to prevent autoimmune
disease. First, if inflammatory cytokines are not decreased,
augmentation of Th2 cytokines secreted by regulatory lymphocytes is
able to prevent ß cell destruction, as demonstrated in our or
other’s oral "tolerance" studies (26). In contrast,
inflammatory cytokines such as IFN- can be reduced quantitatively,
for example, by blocking the STAT4/IL-12 pathway, as demonstrated by
the present study. In this situation, Th2 cytokines need not to be
augmented.
In summary, our results emphasize the following important points.
First, in the absence of STAT4, the degree of insulitis and incidence
of diabetes that depended on both CD4+ and
CD8+ T lymphocytes were significantly reduced.
Second, the lack of STAT4 did not preclude the generation of protective
Th1 and Tc1 responses against LCMV and influenza, a finding that is in
good agreement with a recent report (46). It is, however,
of interest that protective immunity was observed although a reduction
of the number and IFN- production of autoreactive T cells was noted.
Third, no enhancement of Th2 responses was detected. Thus, disruption
of STAT4 signaling ameliorates autoimmune disease without compromising
protective cell-mediated immune responses and could form the basis of
an attractive novel immunotherapeutic approach.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Matthias G. von Herrath, Department of Neuropharmacology, Division of Virology, IMM6, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037. E-mail address:
3 Abbreviations used in this paper: IDDM, insulin-dependent diabetes mellitus; RIP, rat insulin promoter; LCMV, lymphocytic choriomeningitis virus; GP, glycoprotein; NP, nucleoprotein; RPA, RNase protection assay; pCTL, precursor cytotoxic T lymphocyte; neo, neomycin; TCID50, 50% tissue culture-infective dose.
Received for publication June 9, 1999. Accepted for publication September 3, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
is essential for ß-cell destruction by CTL. J. Exp. Med. 185:531. during the recall response to influenza virus. J. Virol. 72:6637. to clear an acute viral infection. Virology 229:349.[Medline]
production from CD4(+) versus CD8(+) T cells. J. Exp. Med. 189:1355.This article has been cited by other articles:
|
|
 |
|
  A. K. Huber, E. M. Jacobson, K. Jazdzewski, E. S. Concepcion, and Y. Tomer Interleukin (IL)-23 Receptor Is a Major Susceptibility Gene for Graves' Ophthalmopathy: The IL-23/T-helper 17 Axis Extends to Thyroid Autoimmunity J. Clin. Endocrinol. Metab., March 1, 2008; 93(3): 1077 - 1081. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Aprahamian, R. Bonegio, J. Rizzo, H. Perlman, D. J. Lefer, I. R. Rifkin, and K. Walsh Simvastatin treatment ameliorates autoimmune disease associated with accelerated atherosclerosis in a murine lupus model. J. Immunol., September 1, 2006; 177(5): 3028 - 3034. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. J. Land, P. Gudapati, M. H. Kaplan, and G. S. Seetharamaiah Differential Requirement of Signal Transducer and Activator of Transcription-4 (Stat4) and Stat6 in a Thyrotropin Receptor-289-Adenovirus-Induced Model of Graves' Hyperthyroidism Endocrinology, January 1, 2006; 147(1): 111 - 119. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Boyton, S. Davies, C. Marden, C. Fantino, C. Reynolds, K. Portugal, H. Dewchand, and D. M. Altmann Stat4-null non-obese diabetic mice: protection from diabetes and experimental allergic encephalomyelitis, but with concomitant epitope spread Int. Immunol., September 1, 2005; 17(9): 1157 - 1165. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Land, J. S. Moll, M. H. Kaplan, and G. S. Seetharamaiah Signal Transducer and Activator of Transcription (Stat)-6-Dependent, But Not Stat4-Dependent, Immunity Is Required for the Development of Autoimmunity in Graves' Hyperthyroidism Endocrinology, August 1, 2004; 145(8): 3724 - 3730. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. E. Juedes, E. Rodrigo, L. Togher, L. H. Glimcher, and M. G. von Herrath T-bet Controls Autoaggressive CD8 Lymphocyte Responses in Type 1 Diabetes J. Exp. Med., April 19, 2004; 199(8): 1153 - 1162. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Christen, R. Darwiche, H. E. Thomas, T. Wolfe, E. Rodrigo, A. Chervonsky, R. A. Flavell, and M. G. von Herrath Virally Induced Inflammation Triggers Fratricide of Fas-Ligand-Expressing {beta}-Cells Diabetes, March 1, 2004; 53(3): 591 - 596. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Wang, N. S. Ostlie, B. M. Conti-Fine, and M. Milani The Susceptibility to Experimental Myasthenia Gravis of STAT6-/- and STAT4-/- BALB/c Mice Suggests a Pathogenic Role of Th1 Cells J. Immunol., January 1, 2004; 172(1): 97 - 103. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bot, E. Rodrigo, T. Wolfe, S. Bot, and M. G. Von Herrath Infection-Triggered Regulatory Mechanisms Override the Role of STAT 4 in Control of the Immune Response to Influenza Virus Antigens J. Virol., May 15, 2003; 77(10): 5794 - 5800. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.-C. Chang, S. Zhang, and M. H. Kaplan Neonatal Tolerance in the Absence of Stat4- and Stat6- Dependent Th Cell Differentiation J. Immunol., October 15, 2002; 169(8): 4124 - 4128. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. B. A. Oldstone, R. Race, D. Thomas, H. Lewicki, D. Homann, S. Smelt, A. Holz, P. Koni, D. Lo, B. Chesebro, et al. Lymphotoxin-{alpha}- and Lymphotoxin-{beta}-Deficient Mice Differ in Susceptibility to Scrapie: Evidence against Dendritic Cell Involvement in Neuroinvasion J. Virol., March 27, 2002; 76(9): 4357 - 4363. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Zhou, G. L. Szot, Z. Guo, O. Kim, G. He, J. Wang, M. J. Grusby, K. A. Newell, J. R. Thistlethwaite, J. A. Bluestone, et al. Role of STAT4 and STAT6 Signaling in Allograft Rejection and CTLA4-Ig-Mediated Tolerance J. Immunol., November 15, 2000; 165(10): 5580 - 5587. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Holz, T. Dyrberg, W. Hagopian, D. Homann, M. v. Herrath, and M. B. A. Oldstone Neither B Lymphocytes Nor Antibodies Directed Against Self Antigens of the Islets of Langerhans Are Required for Development of Virus-Induced Autoimmune Diabetes J. Immunol., November 15, 2000; 165(10): 5945 - 5953. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
