| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
The Hospital For Sick Children,
St. Michael’s Hospital, University of Toronto,
Sunnybrook and Women’s College Health Sciences Center, University of Toronto, Ontario, Canada; and
§
Children’s Hospital of Pittsburgh, Pittsburgh, PA 15219
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
transcripts in the brain. Diabetes and MS appear more
closely related than previously perceived. NOD mouse-specific,
autoimmune encephalitis provides a new MS model to identify factors
that determine alternative disease outcomes in hosts with similar
autoreactive T cell repertoires. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The major predisposing genes for MS and diabetes map to generally different class II histocompatibility molecules that are common in the general population; their risk association emphasizes the major role of T lymphocytes in both diseases. In addition, gene products from nearly 20 predisposing genomic regions conspire toward development of autoimmunity and overt disease, with overlap among diabetes, MS, and several other autoimmune disorders (9, 10).
Diabetic autoimmunity is characterized by self-reactive B and T
lymphocytes that target a set of proteins expressed in pancreatic 
cells. Proinsulin (PI), IA2, GAD65 and 67, and islet cell autoantigen
of 69 kDa (ICA69) are the major examples (11). These
target self-Ags are not islet cell specific, and neither is diabetic
autoimmunity; signs of celiac and thyroid autoimmunity are fairly
common in patients (12, 13), and diabetes-prone NOD mice
develop signs of thyroid and Sjögren’s disease
(14). Occasional islet-reactive T cells are found in
almost 10% of the general population (15), but <0.5% of
these subjects are likely to develop overt diabetes. Although it is
uncertain what expands autoimmune T cell pools and what determines
their tissue-destructive potential, access to islet target tissue has
been suggested as a critical element in diabetes-prone hosts, despite
availability of most relevant autoantigens in other tissues
(16).
Myelin basic protein (MBP) is the classical autoantigen in MS patients, but major additional target proteins such as proteolipid protein (PLP) and others have been identified (17). T cell autoreactivity to myelin components is common in MS patients (18, 19), but CNS-reactive T cell lines can also be grown from blood samples of healthy donors (20, 21). Self-tolerance to myelin Ags is incomplete, and mammals readily develop experimental autoimmune encephalitis (EAE) following immunization with CNS Ags emulsified in CFA plus transient impairment of the blood-brain barrier (BBB) with pertussis toxin (PT) (22). However, incomplete self-tolerance to CNS Ags is without clinical consequences in the normal host, and it is unclear what drives the development of and allows intermittent CNS invasion by tissue-destructive, myelin-specific T cell pools in the MS-prone human.
The classical concept of organ-selective autoimmune diseases holds that each is characterized by a requisite set of autoreactive T cells that mediate the disease-specific tissue damage in a susceptible host (23). Here we report a mutual lack of disease fidelity among autoreactive T cells from diabetes and MS patients. T cells reactive with islet autoantigens were common in MS patients, whereas myelin-reactive T cells were found in many diabetes patients and in first degree relatives (FDRs) with high risk to develop overt diabetes. To test whether the development of CNS autoreactivity was a typical part of diabetic autoimmunity, we examined diabetes-prone NOD mice. These animals were found to spontaneously generate similar CNS autoreactivity during prediabetes. Animals develop acute monophasic and delayed-onset relapsing-remitting autoimmune encephalitis following injection of PT, without any immunizations or adjuvants. Here we characterize this new model for quasi spontaneous autoimmune encephalitis. These animals offer the unique opportunity to examine the elements that determine organ-selective disease outcomes in the face of nonselective autoimmunity.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Currently untreated patients (n = 38) with
stable or relapsing MS who visited the Toronto MS clinic (St.
Michael’s Hospital, Toronto, Ontario, Canada) for routine follow-up,
and 34 healthy, adult volunteers provided blood samples for this study.
Three study cohorts were recruited from the population-based,
prospective follow-up program in diabetes kindreds from Allegheny
County (PA) (24). Type I diabetes patients, newly
diagnosed at the Children’s Hospital of Pittsburgh (n
= 54) were recruited with informed consent. Twenty-seven siblings to
index cases, participating in the follow-up study (15),
were recruited here because they had a high risk to develop overt
diabetes (70%/15 yr), as determined by the presence of islet cell
autoantibodies and risk-associated HLA-DQ genotypes (24).
FDRs (n = 78) without autoantibodies were recruited as
controls. Previous studies have identified a relatively low cumulative
diabetes risk (
5%) in such FDRs, although two-thirds have
diabetes-associated risk alleles (25).
Animal experiments
NOD/Lt, NOD.scid, SJL/J, NZB, NZB/w, BALB/c, and C57BL/6 mice were bred and housed in our vivarium, where the incidence of diabetes is 83% in NOD females <8 mo of age. NOD mouse-specific, autoimmune encephalitis (AENOD) was induced in NOD mice through i.v. injection of PT (200 ng, twice) 48 h apart. In adoptive transfer experiments, 1 x 107 pooled spleen cells from prediabetic (12 wk old), diabetic, or young (4 wk old) NOD females were injected i.v. into irradiated (650 rad) NOD males or into 8-wk-old NOD.scid mice. If not indicated otherwise, PT (200 ng i.v.) was injected on days 1 and 3 of a given experiment. Diabetes was determined by blood glucose measurements as described (26). Mice were frequently weighed and monitored daily for clinical signs of CNS disease, with at least one of two to three observers blinded to the protocol. The standard EAE grading (22) was modified because the clinical picture differed: 0, normal; 1, drowsy, ruffled fur, weight loss of at least 10%, flaccid tails; 2, 1 plus abnormal righting reflex; 3, impaired balance and/or ataxia, involuntary movements, weakness of limbs, usually asymmetric; 4, paralysis and/or persistent tremors, possible incontinence; 5, moribund or death. The observer scores were averaged. Animals with severe disease were sacrificed, and spleens were removed before perfusion with 40 ml PBS, followed by 10% buffered formalin. Tissues were prepared for light microscopic interpretation by staining with Luxol Fast Blue and hematoxylin and eosin. Slides were analyzed by a veterinary pathologist without prior knowledge of experimental regimen or clinical course. Final conclusions were blindly confirmed by three independent pathologists. Artifact-free video electroencephalogram recordings were performed in freely moving animals using a 1–5 monopolar electrode head connector, FET, preamplifiers and batteries, a signal conditioning device (Axon Instruments, Foster City, CA), A/D converter (MP100; Biopac Systems, Santa Barbara, CA), and a video/PC-PC video computer board for electroencephalogram and image data. The system was developed with untreated and genetically modified NOD and control mice, none of which showed spontaneous seizures.
BBB integrity
Evans Blue (2% w/v, 100 µl) was injected i.p. into NOD, NOD.scid, or encephalitis-prone SJL mice 7–10 wk of age. After 1 h, mice were sacrificed and perfused. Brains were washed in PBS. Dye was extracted (72 h) in formamide and measured (OD 620 nm).
T cell assays
PBMC from the study subjects were purified on Ficoll-Hypaque gradients and cultured (105/well) for 1 wk in protein-free Hybrimax 2897 medium (Sigma, St. Louis, MS) supplemented with human IL-2 (10 U/well) and 0.01–10 µg of a given test Ag (15). Proliferative responses ([3H]thymidine incorporation) were expressed as stimulation index (SI, experimental/control cpm, background counts, x ± SD 1280 ± 210, range: 1000–2000 cpm). Positive responses were greater than the mean SI in OVA-stimulated cultures plus 4 SDs (15). This assay gave satisfactory results in a large, blinded study of diabetes families and in the first international T cell workshop of the Immunology of Diabetes Society (15, 27). The means and range of absolute background counts or of positive responses from diabetes or MS patients were not statistically different between the present and the previous study (15). The assay of murine T cell responses was similar except for the use of AIM-V serum-free medium (Life Technologies, Mississauga, Ontario, Canada), 72 h of culture, and a higher cell input (2 x 105/well). Conclusions from human and mouse experiments were similar when data were calculated as net cpm (experimental minus background cpm), but variations and confidence intervals were predictably larger.
Reagents
The diabetes-relevant test Ags have been described
(15). Recombinant human baculo-ICA69 was a gift from J.
Ilonen (University of Turku, Turku, Finland). Escherichia
coli-derived human GAD65 was a gift from B. Singh (University of
Western Ontario, London, Ontario, Canada). PI was a gift from Lilly
(Indianapolis, IN), IA-2 (ICA512) was a gift from Bayer (Elkhart, IN).
PT was a gift from Pasteur-Merrieux-Connaught Canada (Toronto, Ontario,
Canada), and the enzymatically inactive PT B subunit was purchased from
Calbiochem (La Jolla, CA). MBP was isolated from human or calf brain
white matter as described (28). The components of human
MBP were separated by CM-52 chromatography and HPLC. Unmodified MBP
(component 1, MBP-c1) has a net positive charge of 21, component 8
(MBP-c8) a charge of 15, due to the natural conversion of arginine to
citrulline (28). Water-soluble PLP was provided by D. Wood
(University of Toronto, Toronto, Canada). OVA, BSA, human hemoglobin
(Hb), cytochrome C, and actin, as well as mouse collagen IV and muscle
extract were purchased. These were nontoxic in cultures costimulated
with PI. Peptides: T cell epitope peptide (Tep69) from ICA69, residues
36–47 (ICA69–36: AFIKATGKKEDE), BSA150 peptide from BSA residues
150–164 (ABBOS; FKADEKKFWGKYLYE), as well ICA69–350 (EEGACLGPVA)
were highly purified (>95%) and confirmed by mass spectroscopy. Ab
reagents for cytokine ELISAs were purchased from BD PharMingen (San
Diego, CA), and the assays followed the manufacturer’s instructions.
Standard template-calibrated RT-PCR (29) used the
following primers: IFN- (449-bp amplicon):
5'-ACACTGCATCTTGGCTTTGC-3', 5'-CGACTCCTTTTCCGCTTCCT-3', internal:
5'-CCTTCTTCAGCAACAGCAAGGC-3'; IL-4 (422-bp amplicon): 5'-ATGG
GTCTCAACCCCCAGCTA-3', 5'-CTACGAGTAATCCATTTGCAT-3',internal:
5'-GTAGGGCTTCCAAGGTGCTTCGC-3'; CD3 (269-bp amplicon):
5'-TACTGGAGCAAGAATAGGAAG-3', 5'-AGTCTGCAGTCTGT CCAG-3', internal:
5'-GGCCAGCGGGACCTGTATTCTG-3'; -glucuronidase (GUS) (321 bp):
5'-GTGATGTGGTCTGTGGCCAA-3', 5'-TCTGCTCCATACTCGCTCTG-3', internal:
5'-GCCAGTTTGAGAACTGGTA TAAGAC-3'. Mice were anesthetized and perfused
with 40 ml of PBS; brain tissue was immediately lysed in Trizol (Life
Technologies) and extracted, and 10 µg of total RNA was treated with
DNase I (Life Technologies), reverse-transcribed, and calibrated as
reported (29).
Statistics
Mann-Whitney U tests were used to compare numeric results. Significance was set at 5%; all p values were two-tailed. Fischer’s exact test was used to analyze tables, with Katz’ approximation to calculate relative risks.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MS patients have islet-reactive T cells
In our assay system, developed for the identification of
disease-associated T cell autoreactivities (15), 90%
of MS patients (Fig. 1
A), but
few healthy controls (Fig. 1B) had T cells that recognized
MBP component 1 (MBP-c1) or the citrullinated MBP-c8 (28).
These T cells are thus MS associated (p <
0.0001 patients vs controls). Bovine MBP was recognized less frequently
(20%). Nearly all subjects responded to tetanus toxoid, but not to
other self-Ags (Hb, actin, cytochrome-C) or to a control
nonself-protein (OVA).
|
A). Responses to
PI and IA-2 were about as common as in diabetics (Figs. 1C
vs 2C), had the same magnitude
as MBP responses (p = 0.2, Fig. 1A),
and were rare in controls (Fig. 1B). T cell autoreactivity
to other diabetes Ags was also common; one MS patient had no
autoreactive T cells, another recognized only PI but not MBP, and only
one had MBP-specific T cells but recognized none of the diabetes
autoantigens tested. The presence of PI- and/or IA2-reactive T cells
was significantly associated with MS (p =
0.0002) and thus provided an unexpected marker for MS
(p < 0.0001).
One-third of MS patients had ICA69-reactive T cells, but none
recognized the major diabetes epitope, T cell epitope peptide (Tep69)
from ICA69, residues 36–47 (ICA69–36, Fig. 1A) (15, 30). Four of the seventeen patients tested with the ICA69–350
peptide had positive responses, suggesting that other epitopes are
targeted in MS (Fig. 1C). We included in these experiments
BSA because this dietary protein is abnormally targeted by diabetic T
cells in patients (Fig. 2C) and NOD mice (26).
BSA (but not egg albumin, OVA) was recognized by T cells from most MS
patients, but few healthy controls (Fig. 1C). However, MS T
cells did not recognize the diabetes-associated BSA150 peptide from BSA
residues 150–164 (ABBOS) (15, 31). Although
disease-associated T cells target similar proteins in MS and diabetes,
their epitope specificities and thus their TCRs are likely
different.
|
A, open vs solid circles). Thus
the development of these T cell pools appears to reflect MS
autoimmunity in general, rather than clinical course. The presence of
islet-reactive T cells was unlikely reflective of islet destruction
because no MS patient had diabetes-associated islet cell autoantibodies
(ICAs and insulin, GAD65, and IA-2 autoantibodies, data not shown). In
diabetes-prone subjects, such Abs accumulate in mid- to late-stage
prediabetes, providing strong markers of islet destruction (32, 33). We conclude that self-reactive T cells in MS patients show
no disease fidelity and commonly target CNS as well as proteins
specifically associated with diabetic autoimmunity. It is unclear what
triggers and sustains these T cell pools and what precludes their
cell pathogenicity in most MS patients (5, 6). Diabetes patients have CNS-reactive T cells
A similar set of test Ags was used to analyze recently diagnosed,
type I diabetic children and two cohorts of FDRs with high or low
diabetes risk. Most patients had T cells targeting multiple diabetes
autoantigens (a representative subset of 12 test Ags/peptides is shown
here, Fig. 2) (15). Five patients (9%) had no or only one
such T cell response. Hb, actin, or cytochrome C evoked little T cell
response.
Twenty-nine diabetic children (53%) had proliferative responses to
MBP-c1, 45% had responses to PLP (Fig. 2A), and 68% of
those tested with both CNS Ags responded to at least one. Response
amplitudes to CNS and diabetes Ags were similar
(p > 0.2).
Multiple T cell autoreactivities are associated with high diabetes risk
(15, 34, 35). Twenty-seven siblings of diabetes patients
had a high disease risk (70% within 15 years, see Materials and
Methods for definition), and most had multiple T cell
autoreactivities to islet cell Ags (Fig. 2B). MBP and/or PLP
responses were equally common in diabetics and the high risk subjects
(p > 0.3, Fig. 2B), but
significantly less common in low risk relatives (n =
78, Fig. 2C, p < 0.0001). The presence of
MBP-reactive T cells predicted the presence of multiple T cell
autoreactivities to islet Ags in patients (p <
0.0001) and high risk relatives (p = 0.0076,
Fig. 2D), whereas only 6% of low risk relatives had MBP
responses, and only 2% had more than one T cell autoreactivity (Fig. 2C). Thus the single measurement of MBP reactivity provides
an unexpected marker of diabetes risk.
We concluded that diabetic patients as well as FDRs with signs of progressive prediabetes commonly generate autoimmunity to classical MS autoantigens. Thus autoimmunity in diabetes and MS targets a similar set of self-proteins, with neither disease nor tissue selectivity, although epitopes and T cell specificities appear to differ.
Myelin-reactive T cells in NOD mice
To test the preceding conclusions and ask whether the generation
of CNS autoreactivity could be part of diabetic autoimmunity in
general, we examined T cell autoreactivity to CNS Ags in untreated,
diabetes-prone NOD and other strains of mice of different ages (Fig. 3). CNS-reactive T cells were detected in
some NOD mice by 5–7 wk of age, they were routinely observed in NOD
mice older than 12 wk, and maintained in 4- to 7-mo-old diabetic
animals. Thus, NOD mice spontaneously develop T cells recognizing MBP,
MBP exon-2, and/or PLP, all major autoimmune targets in MS
(36). As expected (37), T cell responses to
diabetes-associated Ags began to appear by 5 wk of age (data not
shown). There were no T cell responses against the murine self-Ags,
type IV collagen, or semipurified myoglobin.
|
Autoimmune encephalitis in NOD mice
We next asked whether CNS-reactive NOD T cells had
pathogenic potential or whether epitopes targeted in NOD mice
(39) would preclude their pathogenicity. Here we
demonstrate that i.v. injection of PT (200 ng twice), without
administering Freund’s adjuvant or any CNS Ag, leads to the
development of an age-dependent autoimmune encephalitis with a
gender-independent acute disease phase and a variable, low incidence
secondary disease exclusively in females, reminiscent of primary
progressive and relapsing-remitting MS (Table I).
|
). Few
mice 
6 wk old showed any symptoms, whereas most animals >8–10 wk
old developed acute, PT-induced malaise. Most of these animals
recovered completely within 4–7 days. However, 12 of 78 adult females
and 4 of 30 males in this series died. In addition, five females
developed obvious, asymmetric paralysis and/or ataxia and were
sacrificed (grade 3–4, Table I). These 21 animals with primary
progressive, monophasic disease were diagnosed to have acute
lymphocytic encephalitis upon necropsy (see histological examples
below, Fig. 4). The diagnosis of MS-like
disease was confirmed by three experienced pathologists blinded to the
experimental protocols.
|
). Four of these relapsing mice were unremarkable until 2–3 mo
after the initial malaise, and then developed a progressively worsening
tendency to move in tight circles, reminiscent of temporal lobe
seizures (40). A typical baseline electroencephalographic
(video) recording from one of these animals is shown in Fig. 4
(top). There was asymmetric frontal activity (Fig. 4A) with clear, absence-like seizure events (Fig. 4B). A -hydroxybutyrate test (41) confirmed
the integrity of the thalamocortical circuitry in this animal (Fig. 4C). The acute and relapsing disease required facilitation
by PT because we have observed no sign of this disease in over 200
untreated NOD females monitored over the past 18 mo. The incidence
(Table I, p = 0.5) and clinical picture of acute
post-PT encephalitis in adult NOD males and females was similar.
However, no male developed signs of secondary relapses over 3–4 mo of
follow-up. Thus the secondary relapsing-remitting NOD encephalitis
follows the female gender bias of diabetes development.
NOD.scid mice were entirely free of CNS disease signs throughout 2–4
mo of observation after PT injection (Table I), even if they received
600 ng three times (data not shown). This suggested that PT-facilitated
CNS disease in wild-type NOD mice depended on the presence of
functional lymphocytes. Consistent with a lack of relevant observations
in the literature, C57BL/6, BALB/c, and SJL mice showed no signs of
disease following PT injection, suggesting that the lymphocytes
required for disease had to arise in and were typical of spontaneously
autoimmune NODs. The absence of clinical signs, including weight loss
in NOD.scid and other control animals ruled out merely toxic effects
of PT.
Histopathological analysis was performed in over 50 animals. Typical
findings in brain sections from some of the 63 female and 21 male NOD
mice with acute malaise 5–6 days post-PT (Table I) were characterized
primarily by small, multifocal, perivascular cuffs of mononuclear cells
in the cortex and brainstem (Fig. 4, D and E),
and mixed inflammatory cell infiltrates in the ventricular system (Fig. 4F). Monophasic, acutely progressive disease, 7–10 days
post-PT (Table I), was associated with more marked, multifocal,
anisomorphic gliosis and mononuclear, perivascular cuffing in
hippocampus, thalamus, cerebellum (Fig. 4G), and brain stem.
Mononuclear cell infiltration was more pronounced, and often observed
in the leptomeninges (Fig. 4H) and ventricular system (Fig. 4I). Multifocal plaques of vacuolar
degeneration/demyelination were observed in the deep cerebellar white
matter (Fig. 4J) and brain stem (Fig. 4K) of
symptomatic, but not asymptomatic, PT-treated NOD mice (Fig. 4L).
NOD females with secondary relapses (Table I) showed lymphocytic
infiltrates similar to animals with acute disease and more prominent
perivascular cuffing (Fig. 4M). No histological changes were
detected in the brains of PT-treated NOD.scid, C57BL/6, or SJL mice
(Fig. 4N). Lymphocytic infiltration and other pathological
abnormalities were rare in spinal cords from PT-treated mice (Fig. 4O). Collectively, these observations characterize
PT-facilitated AENOD as a condition distinct from
classical EAE and resemblance to human MS. In these animals (and the
adoptive transfer recipients below) we did not observe evidence for
frank autoimmunity in other tissues, but we did not perform
quantitative analyses that would rule out acceleration/exacerbation of
sialitis. In terms of major tissue destructive and pathogenic
autoimmunity, NOD mice are prone to diabetes and
AENOD.
Adoptive transfer of AENOD
Spleen cells from 12-wk-old females
(107/recipient) were adoptively transferred into
8-wk-old NOD.scid mice (Fig. 5A) or diabetic spleen cells
were transferred to sublethally irradiated, 8-wk-old NOD males (Fig. 5B). Other NOD.scid mice received spleen cell grafts from
4-wk-old NOD females (Fig. 5C) or highly purified splenic T
cells from diabetic donors (Fig. 5D). Recipients received
two PT injections as indicated, or two injections of the enzymatically
inactive B subunit instead of holotoxin (Fig. 5E). In
another series of experiments, adoptive grafts were pretreated for 2
days in vitro with 100 ng/ml PT before adoptive transfer into untreated
NOD.scid (Fig. 5F), or NOD.scid recipients were pretreated
with PT 2 days before adoptive transfer of diabetic splenocytes (Fig. 5G). For analysis of these 58 adoptive transfer recipients
(Fig. 5) and 37 additional controls of adoptively transferred, but not
PT-treated recipients (data not shown), we used a disease grading
analogous to EAE (see Materials and Methods).
|
A), PT injection generated an
acute disease similar to PT-treated wild-type NOD mice in all but two
animals (9 and 10). Two animals progressed rapidly toward death (1 and
2), three progressed to grade 3 clinical disease over the following
week and were sacrificed for histopathology (3, 4, 5), and three recovered
completely (6, 7, 8). Three of the five then apparently healthy animals
(6, 7, 8, 9, 10), developed early (6 and 7) or late (8) relapse, and two (9 and
10) remained symptom free over a 62-day follow-up period.
A comparable course of events characterized the adoptively transferred,
irradiated NOD males (Fig. 5B). Two animals quickly
developed primary progressive, monophasic disease with asymmetric
neurological symptoms, ataxia, weight loss, and death. Four animals
showed intermediate disease progression. All animals were sacrificed
when they developed diabetes.
NOD autoimmune encephalitis is age dependent, because spleen cells from
young NOD females (4 wk old) were unable to transfer disease,
consistent with the absence of detectable CNS-reactive T cells at that
age (Fig. 5C). AENOD is mediated by T
lineage cells that spontaneously arise in the course of prediabetes
because the disease could be transferred with purified
CD3+ cells from newly diabetic donors (Fig. 5D).
The morphology and distribution of perivascular inflammatory cell cuffs
and lymphocytic meningitis in the brain were indistinguishable from
those described above for PT-treated wild-type NOD mice (Fig. 4P). Histological examination of sections of cervical,
thoracic, and lumbar spinal cord from most symptomatic, adoptively
transferred NOD.scid mice were unremarkable, but in a few of these
animals we observed multifocal plaques of demyelination of the dorsal,
lateral, and ventral funiculi (Fig. 4, Q and
R).
Thirty-seven irradiated adoptive transfer recipients of diabetic spleen cells but without PT injections were closely monitored. None developed neurological signs, malaise, or weight loss until they developed diabetes as expected, and no abnormalities in CNS histopathology were observed in the subset of mice analyzed (data not shown). A trend toward less or slower diabetes development in PT-pretreated NOD.scid mice requires further study (42).
When PT was replaced with the enzymatically inactive pertussis B
subunit, no signs of CNS disease were observed (Fig. 5E).
This suggested that AENOD required PT-mediated
ADP-ribosylation of G(i) proteins (43) either in adoptive
transfer recipients, in the adoptive transfer grafts, or in both. PT
pretreatment of NOD.scid recipients 2 days before adoptive transfer was
sufficient for AENOD development, although the
engrafted cells would have little PT exposure (Fig. 5G). In
vitro PT pretreatment of the adoptive transfer graft was insufficient
for AENOD development, but these pretreated
grafts were able to mediate diabetes development (Fig. 5F).
These observation implied that AENOD requires PT
targeting of a somatic NOD.scid tissue, but not of the lymphoid and
APCs in the adoptive transfer graft.
Collectively, these data demonstrate that NOD mice spontaneously generate a CNS-pathogenic T cell repertoire that can mediate a heterogeneous autoimmune encephalitis reminiscent of the clinical, histopathological, and electrophysiological spectrum of MS. Identical adoptive transfer grafts produce this heterogeneity in identical, littermate recipients, suggesting that disease course depends on stochastic elements, amenable to further study in this model. Adoptive transfer of spleen cells from diabetic animals provides a model for accelerated, high-incidence AENOD.
AENOD involves cytokine bias
We used template-calibrated RT-PCR to amplify IFN-, IL-4, CD3,
and GUS (control) messages in brain tissue from PT-treated NOD mice
(Fig. 6A, lanes 1–3) and from
adoptively transferred NOD.scid mice with neurological disease
(lane 4 provides an example). Brain tissue from mice
with neurological symptoms contained CD3 message, locating T cells to
the CNS of symptomatic animals. Most of these mice had IFN- messages
in the brain, whereas IL-4 transcripts were barely detectable. However,
in 4 of 22 symptomatic mice tested, IL-4 transcripts were prominent,
with fewer IFN- transcripts; one example is shown in lane
3. This suggested that CNS-invasive T cell pools acquire a
cytokine bias, usually, but not categorically, for the Th1 cell
sublineage.
|
, and IL-4 transcripts were both easily detected in
spleen cell RNA from PT-treated, symptomatic (e.g., Fig. 6A, lane
5), and untreated animals (data not shown). As a control, we
analyzed SJL-strain mice with MBP-induced severe EAE (Fig. 6,
lane 6) and found a considerable Th1 bias among spinal cord
cytokine messages. Diabetic mice not treated with PT (Fig. 6A,
lanes 7 and 8) but with CNS-reactive T cells in the
spleen (see Fig. 3) lacked CD3 and cytokine transcripts in the brain,
consistent with absence of CNS disease in untreated NOD mice. Activation and expansion of CNS-reactive T cells during AENOD
When spleen cells from PT-treated, encephalitic mice were
stimulated in vitro with CNS Ags, there was a robust, proliferative
response (Fig. 6B, filled column), significantly higher than
those in untreated, age-matched NOD mice (Fig. 6B, open
column, p < 0.001). In addition, spleen cells from
PT-treated, symptomatic mice showed an Ag-dependent release of IFN-,
but little IL-4 (Fig. 6C). Polyclonal activation with the
mitogen Con A generated balanced cytokine production. In contrast,
healthy, PT-treated NOD mice (or untreated diabetic animals, data not
shown) had proliferative responses to CNS Ags similar to those shown in
Fig. 3, but they produced at best borderline amounts of IFN- and
IL-4 (Fig. 6C).
Collectively, our data have shown that islet-specific as well as CNS-reactive T cells arise spontaneously during human and NOD prediabetes and at some point during development of CNS disease in patients with MS. In NOD mice, ignorant CNS-reactive T cells develop cytokine responses and biases during AENOD. In adoptive transfer recipients, PT seems to target G(i) proteins in host tissue rather than the graft.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We were surprised by the lack of disease fidelity among autoreactive T cells in diabetes and MS patients. The finding that CNS autoreactivity emerged as a marker of diabetes risk in families of children with type I diabetes associated the development of these T cell pools and progressive prediabetes with strong statistical confidence. This conclusion applies to NOD mice as well, where protracted relapsing-remitting CNS disease was only observed in NOD females, prone to develop progressive prediabetes and overt disease.
The observation that PI autoreactivity showed a strong association with MS was equally unexpected, as this molecule is perhaps the most prominent marker of diabetic autoimmunity (15, 44). Although it remains obscure what initiates, expands, and sustains these T cell pools in diabetes- and MS-prone hosts, the common denominator may well be autoimmunity that breaks self-tolerance to some, but not many other proteins, of which we measured but a small subset. It should be instructive to study patients with other organ-selective autoimmune disorders and map T cell disease fidelity.
The measurement of autoreactive T cells is difficult, and we spent considerable efforts to develop and validate our serum-free assay system in diabetes families (15, 27) and in NOD mice (26, 30, 31). Although these responses have small amplitudes, they provided the impetus for NOD mouse experiments that, in turn, delineated similar CNS autoreactivities as well as functional confirmation for the presence of CNS-reactive T cells in these animals. Small proliferative T cell responses in our assay system are likely of biological relevance in unimmunized autoimmune humans and mice. We failed to detect many CNS-reactive T cells in healthy humans, unlike observations reported in many MS studies (e.g., Refs. 18, 19). Our serum-free culture conditions appear to be insensitive for the routine detection of autoreactive T cells from normal hosts, cells that may differ functionally from those in patients (45).
Studies of NOD mice confirmed our hypothesis that CNS autoreactivity is a routine part of prediabetes. Nevertheless, we were surprised by the results because MBP was routinely targeted by NOD T cells, although it is poorly encephalitogenic in the nonautoimmune NOD cousins, IAg7-bearing Biozzi mice (39). To analyze an extended range of CNS autoantigens and map NOD target epitopes should be rewarding, because we do not know whether few or many autoreactive T cell pools participate in AENOD, and whether these T cells were measured in our in vitro experiments. Autoreactivities underlying the disease now seem likely to arise spontaneously, and if so, they may be susceptible to early experimental treatments with immunotherapeutic peptides identified in such mapping studies.
The finding that PT injection at moderate doses directs progression to overt autoimmune disease in the CNS provided indirect support for the conclusions of in vitro human and mouse T cell studies, i.e., that islet and CNS self-Ags are targeted in parallel. However, it is unclear what protects the diabetes-prone host from spontaneous CNS disease, what triggers the pathogenicity of these T cells well after their initial appearance, and, by inference, what protects the islet in MS. These are not all new questions. The development of islet-reactive T cells in humans precedes disease onset by years (15). T cells of prediabetic NOD mice infiltrate islets months before disease onset (46), and disease transfer proceeds at a slow rate unless progression checkpoints are overcome, e.g., through transfer into NOD.scid (47).
The study of AENOD may shed new light on these questions. AENOD requires T cells probably arising spontaneously in young adult mice. A role for T cells was corroborated by the presence of lymphoid cells in pathological CNS lesions and results of RT-PCR studies that demonstrated CD3 and cytokine transcripts in the brain of symptomatic animals. T cells are prerequisite for diabetes and for AENOD, and both processes have critical checkpoints that protect the islet (temporarily) and the CNS (completely until PT injection). The identity of these checkpoints is unknown, but the one that governs AENOD appears to be controlled through G(i) signaling pathways, because only PT was effective in disease facilitation. This may open opportunities for studies of small molecule pharmaceuticals that modulate such pathways.
CNS-reactive T cells in prediabetic NOD mice were able to proliferate following cognate activation in vitro, but they produced very little cytokine, a property they acquired rapidly following disease onset. Our adoptive transfer experiments mapped the PT target to host tissue. Transient breaches of the BBB are thought to be an important part of the pleiotropic PT actions (48). If PT treatment promoted CNS tissue access, this would be a good candidate event associated with loss of T cell ignorance and transition to tissue-destructive autoimmunity and overt disease.
It will be interesting to determine whether the converse is true,
i.e., that CNS-autoimmune NOD mice might be protected from CNS disease
by a tight BBB. To begin addressing this possibility, we compared the
integrity of BBBs in untreated NOD and encephalitis-prone SJL mice
(Fig. 6D) using a standard dye extravasation assay
(49). Following i.v. injection of Evan’s Blue, NOD and
NOD.scid mice showed significantly less dye extravasation than
encephalitis-prone SJL mice.
The acute disease phase of NOD encephalitis is reminiscent of the
probably TNF-
-mediated initial malaise in EAE (50). It
occurs at high incidence, and its lack of gender bias separates it from
the secondary relapsing-remitting disease of NOD females. This
secondary, low-incidence CNS disease spans a spectrum of variable
courses and outcomes. Stochastic, rather than genetic or environmental
elements are likely to determine varied disease courses in littermates.
Nonimmune elements may contribute to this variability, perhaps
involving varied tissue access in animals receiving identical adoptive
grafts. The considerable similarities between
AENOD and MS imply that mechanistic studies of
disease heterogeneity in NOD mice might become relevant for
understanding MS heterogeneity.
Our findings contrast with the view that diabetes and MS reflect the presence of mutually exclusive T cell pools, which target tissue-selective autoantigens and, in a functional sense, would define each disease. The limited study of T cell epitopes conducted so far suggested that although the same proteins are targeted in diabetes and MS, the epitopes may differ. Mechanisms for the selection of proteins and epitopes require further study. Once somehow singled out for autoimmune targeting, these proteins may drive selection of disease-specific epitopes within these proteins, likely governed by diabetes- and MS risk-associated MHC alleles. However, it remains to be determined if and how these class II alleles affect disease outcome; NOD IAg7 appears fully permissive for the development of pathogenic autoimmunity in both islets and CNS.
The autoimmunity measured in MS and in diabetes patients and functionally analyzed in NOD mice was unexpectedly similar. To learn what (usually) protects diabetics from MS, and MS patients from diabetes will be an important goal that could have therapeutic ramifications. Our observations describe a new model of quasi spontaneous, not Ag-induced autoimmune encephalitis as a platform to decipher genetic and/or environmental factors that decide on progression to diabetes or MS. Alternative roles for adhesion molecules (51), tight BBBs (52), possible abnormalities in the cbl-vav/CD28 pathways of T cell activation (53, 54), and crosses of NOD with other inbred strains (10, 55) may be among the avenues to pursue.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 S.W., I.A., and R.K.C. contributed equally to this work.
3 Address correspondence and reprint requests to Dr. H.-Michael Dosch, The Hospital For Sick Children, 555 University Avenue, Toronto, Ontario, Canada M5G 1X8.
4 Abbreviations used in this paper: MS, multiple sclerosis; AENOD, nonobese diabetic mouse-specific, autoimmune encephalitis; ICA69, islet cell autoantigen of 69 kDa; PI, proinsulin; PT, pertussis holotoxin; MBP, myelin basic protein; PLP, proteolipid protein; EAE, experimental autoimmune encephalitis; FDRs, first-degree relatives; SI, stimulation index; Hb, human hemoglobin; GUS, -glucuronidase; BBB, blood-brain barrier.
Received for publication August 17, 2000. Accepted for publication December 6, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells drive T cell-immune responses in the nonobese diabetic mouse model. J. Exp. Med. 181:1635. 4-integrins in the development of insulin-dependent diabetes mellitus. Curr. Top. Microbiol. Immunol. 231:65.[Medline]
This article has been cited by other articles:
|
|
 |
|
  H.-J. Kim, C.-G. Jung, M. A. Jensen, D. Dukala, and B. Soliven Targeting of Myelin Protein Zero in a Spontaneous Autoimmune Polyneuropathy J. Immunol., December 15, 2008; 181(12): 8753 - 8760. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Tsui, Y. Chan, L. Tang, S. Winer, R. K. Cheung, G. Paltser, T. Selvanantham, A. R. Elford, J. R. Ellis, D. J. Becker, et al. Targeting of Pancreatic Glia in Type 1 Diabetes Diabetes, April 1, 2008; 57(4): 918 - 928. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Suri, J. J. Walters, H. W. Rohrs, M. L. Gross, and E. R. Unanue First Signature of Islet {beta}-Cell-Derived Naturally Processed Peptides Selected by Diabetogenic Class II MHC Molecules J. Immunol., March 15, 2008; 180(6): 3849 - 3856. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-U. Lee, A. Grigorian, J. Pawling, I-J. Chen, G. Gao, T. Mozaffar, C. McKerlie, and M. Demetriou N-Glycan Processing Deficiency Promotes Spontaneous Inflammatory Demyelination and Neurodegeneration J. Biol. Chem., November 16, 2007; 282(46): 33725 - 33734. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Seyfert-Margolis, T. D. Gisler, A. L. Asare, R. S. Wang, H. M. Dosch, B. Brooks-Worrell, G. S. Eisenbarth, J. P. Palmer, C. J. Greenbaum, S. E. Gitelman, et al. Analysis of T-Cell Assays to Measure Autoimmune Responses in Subjects With Type 1 Diabetes: Results of a Blinded Controlled Study Diabetes, September 1, 2006; 55(9): 2588 - 2594. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. M. Nielsen, T. Westergaard, M. Frisch, K. Rostgaard, J. Wohlfahrt, N. Koch-Henriksen, M. Melbye, and H. Hjalgrim Type 1 diabetes and multiple sclerosis: a danish population-based cohort study. Arch Neurol, July 1, 2006; 63(7): 1001 - 1004. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Bour-Jordan, H. L. Thompson, and J. A. Bluestone Distinct Effector Mechanisms in the Development of Autoimmune Neuropathy versus Diabetes in Nonobese Diabetic Mice J. Immunol., November 1, 2005; 175(9): 5649 - 5655. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. A. Ivakine, C. J. Fox, A. D. Paterson, S. M. Mortin-Toth, A. Canty, D. S. Walton, K. Aleksa, S. Ito, and J. S. Danska Sex-Specific Effect of Insulin-Dependent Diabetes 4 on Regulation of Diabetes Pathogenesis in the Nonobese Diabetic Mouse J. Immunol., June 1, 2005; 174(11): 7129 - 7140. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Setoguchi, S. Hori, T. Takahashi, and S. Sakaguchi Homeostatic maintenance of natural Foxp3+ CD25+ CD4+ regulatory T cells by interleukin (IL)-2 and induction of autoimmune disease by IL-2 neutralization J. Exp. Med., March 7, 2005; 201(5): 723 - 735. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Carrillo, M. C. Puertas, A. Alba, R. M. Ampudia, X. Pastor, R. Planas, N. Riutort, N. Alonso, R. Pujol-Borrell, P. Santamaria, et al. Islet-infiltrating B-Cells in Nonobese Diabetic Mice Predominantly Target Nervous System Elements Diabetes, January 1, 2005; 54(1): 69 - 77. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. G. Marrosu, C. Motzo, R. Murru, R. Lampis, G. Costa, P. Zavattari, D. Contu, E. Fadda, E. Cocco, and F. Cucca The co-inheritance of type 1 diabetes and multiple sclerosis in Sardinia cannot be explained by genotype variation in the HLA region alone Hum. Mol. Genet., December 1, 2004; 13(23): 2919 - 2924. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. M. Libman, M. Pietropaolo, S. A. Arslanian, R. E. LaPorte, and D. J. Becker Evidence for Heterogeneous Pathogenesis of Insulin-Treated Diabetes in Black and White Children Diabetes Care, October 1, 2003; 26(10): 2876 - 2882. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. V. Serreze, M. A. Pierce, C. M. Post, H. D. Chapman, H. Savage, R. T. Bronson, P. B. Rothman, and G. A. Cox Paralytic Autoimmune Myositis Develops in Nonobese Diabetic Mice Made Th1 Cytokine-Deficient by Expression of an IFN-{gamma} Receptor {beta}-Chain Transgene J. Immunol., March 1, 2003; 170(5): 2742 - 2749. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. P. L. Chiu, E. Ivakine, S. Mortin-Toth, and J. S. Danska Susceptibility to Lymphoid Neoplasia in Immunodeficient Strains of Nonobese Diabetic Mice Cancer Res., October 15, 2002; 62(20): 5828 - 5834. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Lassmann, C. Kincaid, V. C. Asensio, and I. L. Campbell Induction of Type 1 Immune Pathology in the Brain Following Immunization Without Central Nervous System Autoantigen in Transgenic Mice With Astrocyte-Targeted Expression of IL-12 J. Immunol., November 1, 2001; 167(9): 5485 - 5493. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Salomon, L. Rhee, H. Bour-Jordan, H. Hsin, A. Montag, B. Soliven, J. Arcella, A. M. Girvin, S. D. Miller, and J. A. Bluestone Development of Spontaneous Autoimmune Peripheral Polyneuropathy in B7-2-deficient NOD Mice J. Exp. Med., September 4, 2001; 194(5): 677 - 684. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Winer, I. Astsaturov, R. K. Cheung, K. Schrade, L. Gunaratnam, D. D. Wood, M. A. Moscarello, P. O'Connor, C. McKerlie, D. J. Becker, et al. T Cells of Multiple Sclerosis Patients Target a Common Environmental Peptide that Causes Encephalitis in Mice J. Immunol., April 1, 2001; 166(7): 4751 - 4756. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and relapsing
(n = 18, •) MS patients are expressed as SI
values. Background counts were similar in all cohorts (1000–2000 cpm).
B, Individual response amplitudes in healthy controls
(n = 34). C, Prevalence of positive
responses (%) from A and B.
and 2
mm lateral to midline. B, Absence-like seizures in a
PT-treated NOD mouse with late relapse of neurological symptoms.
Spontaneous bilaterally synchronous spike-and-wave discharges (SWD) at
6–7 cycles/s were associated with intermittent immobility, vibrissal
twitching, and absence-like seizures. C, Subcutaneous
injection of 
