| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institut National de la Santé et de la Recherche Médicale,
*
Unité 25,
Equipe 0016,
Unité 345, and
Unité 507, Faculté Necker-Enfants Malades, Université René Descartes, Paris, France; and
¶ Hôpital Européen Georges Pompidou, Paris, France
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Construction of the cDNA encoding murine sICAM-1 and an IgG1-Fc fragment was similar to TNFR-IgG heavy chain construction (4). Briefly, the CH3-CH2 domains of murine IgG1 cDNA and the extracellular domain of murine ICAM-1 cDNA were amplified using PCR. PCR primers were as follows: 5'-ggatccacgcggaaccatTGTCAGGTACACATTCCT-3' (corresponding to the 1453–1436 bp of the murine sICAM-1 cDNA and a thrombin cleavage site), 5'-gcgcatcgaTGCAATGGCTTCAACCCG-3' (coding for -4–14 bp of the murine sICAM-1 cDNA and the ClaI site), 5'-ctggttccgcgtggatccGTGCCCAGGGATTGTGGT-3' (corresponding to a thrombin cleavage site and the 5' end of the IgG1 moiety), 5'-attaagcattctagaTCATTTACCAGGAGAGTG-3' (corresponding to the 5' end of the IgG1 and a XbaI site). sICAM-1 cDNA joins IgG1-Fc cDNA through the thrombin cleavage site. Both fragments were cloned in TA cloning plasmid (Invitrogen, San Diego, CA) through a common BamHI site, digested with ClaI and XbaI sites, and ligated into Ad shuttle plasmid. To obtain a replication-deficient recombinant Ad vector, shuttle plasmid was cotransfected into 293 cells (CRL 1573; American Type Culture Collection, Manassas, VA) with pJM17 plasmid containing the Ad type 5 genome (Microbix Biosystems, Toronto, Canada). The replication-deficient Ad vector Ad.null contains an expression cassette including an Ad promoter with no exogenous gene and was used as a control Ad vector (6). Ad vectors were propagated, purified, and titered as previously described (6).
Production and purification of sICAM-1/Ig protein
HeLa cells (CCL 2; American Type Culture Collection) were incubated with Ad.sICAM-1/Ig at 40 PFU/cell. sICAM-1/Ig was purified from the cell supernatant using a goat anti-mouse IgG1-Fc (Sigma-Aldrich, St. Louis, MO). Purity of sICAM-1/Ig was confirmed by silver staining after 5% SDS-PAGE. Purified sICAM-1/Ig was quantified using the Lowry test. Western blot analyses were performed using an Ab anti-mouse ICAM-1 (3E2; BD PharMingen, San Diego, CA) after SDS-PAGE in nonreducing conditions of mice sera. sICAM-1/Ig serum levels were quantified by ELISA using two ICAM-1 mAbs: a rat Ab anti-mouse ICAM-1 (KAT-1; R&D Systems, Minneapolis, MN), and a biotin-conjugated hamster Ab anti-mouse ICAM-1 (3E2; BD PharMingen). Purified sICAM-1/Ig was used as a standard.
Mice histopathology
NOD (Kd, I-Ag7,
Db), NODscid, and congenic
C57BL/6-H2g7 mice (12 backcrosses) have been
bred in our animal facility for many years under specific
pathogen-free conditions (Institut National de la Santé et de la
Recherche Médicale, Unité 25, Paris, France). NOD females
developed insulitis by 4 wk of age and spontaneous diabetes
appeared by 14 wk of age (95% incidence at 40 wk). The transgenic NOD
BDC2.5 and NOD C
-/- mice were kindly
provided by D. Mathis and C. Benoist (Harvard Medical School, Boston,
MA) (7); the mice used in the present experiments
were backcrossed 12 times onto the NOD background. By crossing these
two latter strains, we obtained NOD BDC2.5
C-/- mice in which all T cells express a
rearranged TCR
from a CD4+ diabetogenic T
cell clone. These mice are devoid of immunoregulatory T cells; 85% of
them develop overt diabetes by 4–6 wk of age. In some
experiments, C57BL/6 mice transgenic for expression of a TCR of known
specificity recognizing PCC88-104 in association with
IEk MHC class II molecules (AND TCR), and
that were rag-2 deficient (8), were used. AND
TCR chain transgenic mice on a Rag-2 deficient background have a pure
population of naive T cells as no rearrangement of endogenous or
TCR chains could occur in these mice. These mice will be referred
to as AND TCR chain transgenic mice. Mice were monitored for glycosuria
and fasting glycemia with colorimetric strips (Glukotest and
Glucotrend; Boehringer Mannheim, Indianapolis, IN). Diabetes was
defined when a fasting glycemia >3 G/L was determined on two
consecutive occasions. Recombinant Ad vectors were administered in vivo
in NOD mice using a single i.v. injection of 2.5 x
109 PFU. Complete remission was defined as the
disappearance of glycosuria and a return to normal glycemia. For
histopathology, paraffin-embedded sections from pancreata were stained
with H&E, and the severity of insulitis was assessed by using the
following criteria: grade 0 = normal islets; grade 1 =
focal or peripheral insulitis (lymphocytes around the islet); and grade
2 = invasive destructive insulitis.
Adhesion test
Purified T cell populations were obtained from spleen and lymph nodes after B cell depletion. Purified sICAM-1/Ig was incubated at 2 µg/ml on 96-well plates overnight at 4°C, followed by saturation with 3% BSA. Purified T lymphocytes were stained with metabolic fluorescent 2'7'-bis-(2-carboxyethyl)-5-(-6-)-carboxyfluorescein (Calbiochem, La Jolla, CA), treated with PMA and incubated at 4 x 105 cells/well in sICAM-1/Ig-coated plates. Cells were pretreated with Abs against murine ICAM-1 (3E2; BD PharMingen) or against murine LFA-1 (9). In some experiments, sICAM-1/Ig-coated plates were incubated with murine serum prior to T cell plating. Assays were performed in multiple wells (4–6 per condition) and bound adherent T cells to sICAM-1/Ig were quantified using a plate reader (Millipore, Bedford, MA).
Adoptive cell transfer
Six-week-old male NODscid mice were used as recipients. Transfused cell suspensions of lymphocytes were prepared by disruption of spleen cells from either overtly diabetic or Ad.sICAM-1/Ig-treated NOD mice. Cells were injected i.v. (107 splenocytes/recipient equivalent to 4–5 x 106 T cells). For some experiments, NOD-scid recipients received a single i.v. injection of Ad.sICAM-1/Ig, Ad.null, or control buffer before or after cell injection (2.5 x 109 PFU). When needed, splenocytes were purified on the basis of the CD62L expression (10).
Skin grafting
Three days before skin graft, C57BL/6 and NOD female mice were injected with Ad.sICAM-1/Ig or Ad.null. CH3 and C57BL/6-H2g7 NOD female mice were used as skin donors, respectively. Tail skin grafts were transplanted onto the thoracic wall of the recipient. Graft survival was documented daily. In a second set of experiments, NOD mice received a skin graft from C57BL/6 mice, 25 wk after Ad.sICAM-1/Ig injection.
Abs and FACS analysis
Ab to CD3, PE-anti-CD4, PE-anti-CD8, PE-anti-CD19, biotin-anti-CD62L, and biotin-anti-CD44 were purchased from BD PharMingen. Surface markers were assessed by flow cytometry on a FACScan flow cytometer (BD Biosciences, Mountain View, CA).
T cell assays
B10BR (H-2k) mice were crossed with
C57BL/6 CD3
-deficient mice (H-2b) to
obtain H-2k CD3-deficient mice. Irradiated
peritoneal macrophages from these mice were used as APCs. T cells were
from AND TCR chain mice. 104 lymph-node T cells
were incubated with 5 x 104 APCs in the
presence of graded doses of PCC 88-104, in sICAM-1/Ig-coated plates.
Supernatants were collected 24 h later and IL-2 production was
quantified by ELISA. [3H]Thymidine was
added in the culture 24 and 48 h later, and proliferation was
measured by thymidine incorporation.
Statistics
The occurrence of diabetes in the different experimental groups was plotted using the Kaplan-Meier method, i.e., a nonparametric cumulative survival plot. The statistical comparison between the curves was performed using the logrank (Mantel-Cox). A positive correlation between sICAM-1/Ig sera levels measured at day 3 and diabetes prevention was determined by receiver operating characteristic analysis.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To overexpress murine sICAM-1 in a dimeric form, an adenovirus
vector encoding an immunoadhesin with murine sICAM-1 and IgG1-Fc cDNAs
(Ad.sICAM-1/Ig) was constructed. HeLa cells were then incubated
with Ad.sICAM-1/Ig and maintained in serum-free medium for 3 days.
Protein affinity chromatography with an Ab against the murine IgG1-Fc
domain was used to purify sICAM-1/Ig proteins from cell supernatants.
SDS-PAGE analysis showed that sICAM-1/Ig was a dimer, migrating at 200
kDa when not reduced (Fig. 1
A,
lane 1) and 100 kDa when reduced (Fig. 1A,
lane 2). These sizes were in agreement with those predicted
for sICAM-1/Ig (11).
|
B,
lane 1), Western blot analysis with an anti-mouse ICAM-1
Ab detected circulating sICAM-1/Ig in mice sera 3 days after i.v.
injection of the Ad.sICAM-1/Ig vector (Fig. 1B, lane
2). In contrast, no circulating sICAM-1/Ig was detected in mice
sera after injection of the control Ad vector Ad.null (Fig. 1B, lane 3).
To demonstrate that sICAM-1/Ig could bind its homologous ligand LFA-1,
murine T lymphocyte adhesion to immobilized sICAM-1/Ig was
evaluated (Fig. 1C). Activated T cells did not adhere
substantially to empty wells in the absence (Fig. 1C, left
empty bar; value ± SE) or presence of ICAM-1 (Fig. 1C,
middle empty bar; value ± SE) or LFA-1 Abs (Fig. 1C,
right empty bar; value ± SE). In contrast, activated T cells
bound to wells coated with sICAM-1/Ig (Fig. 1C, left crossed
bar; value ± SE). Adhesion was totally inhibited by preincubation
with an anti-LFA-1 Ab (Fig. 1C, left crossed bar;
value ± SE vs the value in the middle bar + SE, p
< 0.001), and partially inhibited by the anti-ICAM-1 Ab (Fig. 1C, left crossed bar; value ± SE vs the value in the
right crossed bar + SE; p < 0.001). This confirms that
T cells adhered to sICAM-1/Ig-coated wells through LFA-1/ICAM-1
binding.
Adenoviral-mediated sICAM-1/Ig gene transfer induced long-term remission of established diabetes
Overt diabetes in NOD females was identified by screening for
glycosuria twice a week. When glycosuric, mice were tested for
hyperglycemia. Mice showing fasting glycemia >3 G/L on two consecutive
occasions were treated with a single i.v. injection of Ad.sICAM-1/Ig or
a control Ad vector Ad.null. As expected, sICAM-1/Ig serum
levels measured by ELISA peaked 3 days after i.v. injection and were no
more detectable 35 days after injection (Fig. 2). Remission of established diabetes
(disappearance of glycosuria and a return to normal glycemia) was
observed in 70% of Ad.sICAM-1/Ig-treated mice 10 days after Ad
injection, and was still up to 50% 200 days after Ad injection (Fig. 2; p < 0.0001). In contrast, no remission was observed
in any mouse treated with Ad.null. Instead, these mice died rapidly.
Histologic examination of pancreata was performed 200 days after i.v.
Ad injection (Fig. 3A). The
vast majority of the islets were either intact (60%, grade 0), or had
an infiltrate that remained confined to the periphery of the islets
(noninvasive peri-insulitis) (16%, grade 1).
|
|
To examine whether long-lasting protection from diabetes can be
induced by transient expression of sICAM-1/Ig well after induction of
insulitis but still at a preclinical stage, nondiabetic female NOD mice
at 14 wk of age were injected i.v. with Ad.sICAM-1/Ig, Ad.null, or
control buffer (Fig. 4). No statistically
significant difference in the incidence of diabetes was observed over
time between control and Ad.null animals, with diabetes incidence of 70
and 55%, respectively, at 40 wk of age (Fig. 4). In contrast, the
transient sICAM-1/Ig expression that followed the Ad administration
induced a significant and lasting protection from diabetes, as compared
with control and Ad.null-injected mice, with the incidence of diabetes
dropping significantly to 26% in this group (p
= 0.0006; Fig. 4). Interestingly, there was a positive correlation
between sICAM-1/Ig serum levels and the incidence of diabetes.
Effective disease protection was only observed in mice showing a serum
sICAM-1/Ig concentration >10 µg/ml (p =
0.005).
|
B). Although in all experimental
groups, insulitis was observed, the topography of the infiltration
significantly differed depending on the treatment. Pancreatic islets of
control mice (buffer and Ad.null-injected) which developed a high
incidence of the disease showed severe invasive/destructive insulitis
(68 and 57%, grade 2; Fig. 3B). In clear contrast, in mice
that were protected from disease by the injection of Ad.sICAM-1/Ig, the
vast majority of islets were either normal (48%, grade 0) or showed
peri-insulitis (28%, grade 1; Fig. 3B). sICAM-1/Ig transient expression did not promote short- or long-term immunosuppression
Using FACS analysis, spleen cell phenotype was evaluated in
Ad.sICAM-1/Ig-treated, Ad.null-treated, and control mice 2
and 16 wk after Ad injection to test whether global immunosuppression
occurred in protected mice. There was no cellular depletion or
modification in Ad-treated mice using CD3, CD4, CD8, CD19, CD44, and
CD62L Abs. To determine whether Ad.sICAM-1/Ig-protected NOD mice were
globally immunocompromised, their capacity to reject a skin allograft
was evaluated. The survival was identical in fully mismatched and minor
histocompatibility Ag-mismatched allogeneic skin grafts in
Ad.sICAM-1/Ig mice and in age-matched controls, either 3 days after
Ad.sICAM-1/Ig injection (i.e., when high serum levels of sICAM-1/Ig
were detected) or 25 wk after Ad.sICAM-1/Ig injection (i.e., when no
more sICAM-1/Ig was detected) in Ad.sICAM-1/Ig-protected mice (not
shown). Finally, to evaluate the development of immunity directed
against recombinant Ad vectors, female NOD mice were injected at 12 wk
of age with Ad.sICAM-1/Ig, Ad.null, or control buffer. Four weeks
later, all mice were reinjected with Ad.sICAM-1/Ig. In contrast to
control buffer mice, no serum sICAM-1/Ig could be detected in
Ad.sICAM-1/Ig or in Ad.null-treated mice, suggesting that a strong
immunity against the Ad vector has developed. To evaluate the presence
of Abs against sICAM-1/Ig, we performed the T cell adhesion test in the
presence of both sICAM-1/Ig and serum from mice 14–32 days
after Ad.sICAM-1/Ig injection. If these mice were immunized against
sICAM-1/Ig, the Abs in their serum should bind to sICAM-1/Ig and block
sICAM-1/Ig adhesion, as shown in the presence of anti-ICAM-1 or
anti-LFA-1 Ab (Fig. 1C). Seven sera from mice injected
with Ad.null, and 20 from mice injected with Ad.sICAM-1/Ig, were
evaluated. None of these sera inhibited T cell adhesion to sICAM-1/Ig,
suggesting that these sera did not contain Ab against sICAM-1.
sICAM-1/Ig transient expression did not delete diabetogenic cells
To determine whether the protective effect of Ad.sICAM-1/Ig gene
therapy was due to a direct effect on diabetogenic T cells, adoptive
transfer studies were performed. Splenocytes from Ad.sICAM-1/Ig-treated
NOD mice showing durable protection 14 wk after Ad injection were
injected into NODscid recipients. In some
experiments, purified CD62L- splenocytes were
injected, because diabetogenic T splenocytes concentrate among
CD62L- cells (10). Unseparated or
CD62L- purified splenocytes from Ad.sICAM-1/Ig-treated mice transferred
diabetes as efficiently as splenocytes from untreated diabetic NOD mice
(Fig. 5), demonstrating that diabetogenic
cells were still present in Ad.sICAM-1/Ig-treated mice. Similarly,
BDC2.5 TCR-transgenic NOD mice, which only express a rearranged TCR
from a CD4+ diabetogenic T cell clone
(7), were injected with Ad.sICAM-1/Ig or Ad.null. No
protection against diabetes was observed in NOD BDC2.5
C-/- mice treated with Ad.sICAM-1/Ig
(n = 5, Ad.sICAM-1/Ig; n = 4,
Ad.null-treated mice; diabetes incidence 80% in each group at day 40,
100% at day 80), confirming that sICAM-1/Ig gene transfer did
not alter diabetogenic cells. Finally, cyclophosphamide (200
mg/kg) was administered to mice that had been injected with
Ad.sICAM-1/Ig at 14 wk of age and did not develop diabetes at
30–40 wk of age, and to a control group of nondiabetic female NOD mice
at 9–10 wk of age. Cyclophosphamide injection induced diabetes in
7 of 10 Ad.sICAM-1/Ig-treated mice and in 6 of 8 mice in the
control group, confirming that diabetogenic cells were still present in
Ad.sICAM-1/Ig-treated mice despite protection from diabetes.
|
Histopathological analysis of pancreata from protected NOD mice
(Fig. 3) argued against an exclusive effect of sICAM-1/Ig gene transfer
on lymphocyte homing because the insulitis, although significantly
decreased as compared with controls, was not totally prevented. To
further evaluate whether the LFA-1/ICAM-1 pathway plays a role for the
migration of islet-specific mononuclear cells, we also attempted to
block the adoptive transfer of diabetes to NODscid mice
following the injection of islet-derived mononuclear cells from newly
diabetic donors, by administering Ad.sICAM-1/Ig to the
recipients. sICAM-1/Ig serum levels in NODscid mice reached
a maximum 7 days after injection, then decreased, and remained
detectable for 8 consecutive wk (Fig. 6).
Injection of Ad.sICAM-1/Ig in NODscid mice 3 wk after
adoptive transfer did not prevent diabetes (Fig. 6). Because under
these experimental conditions, Ad.sICAM-1/Ig was injected at a time
point when cells home to the pancreas to some extent and the diabetes
process starts, experiments were also performed in which Ad.sICAM-1/Ig
injection was injected 3 days before adoptive cell transfer. Early
Ad.sICAM-1/Ig injection was also totally ineffective in preventing
diabetes (diabetes incidence in Ad.sICAM-1, Ad.null, and control groups
of >80% 7 wk after cell transfer, and of 100% 9 wk after cell
transfer). These results suggest that Ad.sICAM-1/Ig did not induce
the remission of diabetes by preventing the extravasation of
autoreactive T cells to the islets.
|
To further examine the sICAM-1/Ig effect on T cell activation, we
evaluated the sICAM-1/Ig agonist or antagonist effect on ICAM-1/LFA-1
costimulation. Lymph node naive T cells from AND TCR chain transgenic
mice were cocultured with APCs in the presence of varying
concentrations of the specific peptide, and in the presence of
sICAM-1/Ig and/or anti-LFA-1 Ab. As expected (9), the
anti-LFA-1 Ab inhibited both cell proliferation, as measured by
thymidine uptake (Fig. 7A),
and IL-2 secretion (Fig. 7B). In marked contrast, sICAM-1/Ig
induced T cell proliferation in the absence and presence of peptide at
24 (not shown) and 48 h (Fig. 7A), and increased IL-2
secretion at 24 h (Fig. 7B), demonstrating a
stimulatory effect of sICAM-1/Ig on T cell activation. The
sICAM-1/Ig stimulatory effect was totally inhibited by
anti-LFA-1 Ab (not shown), confirming that sICAM-1/Ig induced T
cell proliferation and IL-2 production through ICAM-1/LFA-1
binding.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells by autoreactive CD4+
and CD8+ cytotoxic lymphocytes (5).
Overt disease is preceded by an infiltration of the islets of
Langerhans by mononuclear cells which starts by 3 wk of age and which
initially remains confined to the periphery of the islets (i.e.,
peripheral insulitis). In quite an abrupt manner, by 12 wk of age, the
infiltrate invades the islets and actively destroys cells through
mechanisms that are not fully understood (12). Our data
demonstrate that a single injection of an Ad vector encoding an
immunoadhesin induces long-term remission of overt diabetes in NOD
mice. The Ad.sICAM-1/Ig-induced remission was durable (>6 mo) and was
maintained despite the presence of peripheral, but not
invasive/destructive, insulitis. These results are in clear contrast
with most other immunointervention strategies described in NOD mice,
which have only been shown to be effective in young animals (13, 14). In fact, only very few agents have been reported to be
similarly effective at reversing overt autoimmunity in a durable
fashion, including polyclonal anti-lymphocyte serum
(15) and anti-CD3 Ab (16). These
strategies induce a short-term immunosuppression associated with
variable lymphopenia that precedes their tolerance-promoting effect. In
marked contrast, the long-term protection observed upon the transient
expression of sICAM-1/Ig was not linked to any visible
modification in the distribution of the main lymphocyte subsets (not
shown) or any evidence of generalized immunosuppression. The fact that
skin allograft rejection was not influenced at all even in recipients
showing high levels of sICAM-1/Ig stresses that the immunoadhesin is by
no means an immunosuppressive drug. Finally, a second
injection of Ad.sICAM-1/Ig, 4 wk after Ad.null or
Ad.sICAM-1/Ig first injection, was totally inefficient in
inducing detectable sICAM-1/Ig serum levels, suggesting that a
neutralizing immune response to the Ad vector efficiently developed.
Longer expression of sICAM-1/Ig in immunocompromised NODscid
as compared with NOD mice also suggests that an immune reaction has
occurred in NOD mice, leading to rapid elimination of the Ad vector.
This again stresses the very low, if any, immunosuppressive capacity of
the sICAM-1/Ig-expressed immunoadhesin. This is a fundamental issue
because it also relates to the specificity of the observed effect.
Ad.sICAM-1/Ig-treated mice were in fact unresponsive to islet Ags while
they exhibited a normal response toward exogenous tissular or
nontissular Ags. We also verified that no Ab against sICAM-1/Ig was
raised after Ad.sICAM-1/Ig injection, suggesting that the effect of
sICAM-1/Ig was not mediated by the induction of the anti-ICAM-1 Ab.
As a whole, these data strongly argue for the capacity of Ad.sICAM-1/Ig
to restore self-tolerance in this context of established
autoimmunity.
The protection observed upon the transient expression of
sICAM-1/Ig was not linked to the elimination or deletion of
diabetogenic cells, as demonstrated by the persistence of diabetogenic
cells in transfer experiments. Furthermore, the finding that a single
injection of cyclophosphamide, an alkylating agent that has been shown
to selectively affect T cell regulation (5), rapidly and
reproducibly reversed sICAM-1/Ig-induced protection provided further
support for the fact that Ad.sICAM-1/Ig-protected hosts still harbored
significant proportions of diabetogenic effectors. This sensitivity to
cyclophosphamide also suggests that immunoregulatory T cells, similar
to T cells that have been shown to control disease onset in prediabetic
NOD mice, could still be functional in Ad.sICAM-1/Ig-protected hosts.
This is further supported by the present observation that Ad.sICAM-1/Ig
did not alter diabetes incidence in BDC
2.5/C-/- transgenic mice that are devoid of
immunoregulatory T cells and in which all T cells express a rearranged
TCR from a CD4+ diabetogenic T cell
clone (17). These results are reminiscent of those
obtained in CD3 Ab-treated mice (16). In this model also,
protected animals showed durable remission of diabetes while expressing
clear signs of overt pathogenic autoreactivity. Just as presently
described for sICAM-1/Ig-treated NOD mice, whole spleen cells from
anti-CD3-treated NOD mice transferred disease as efficiently
as splenocytes from untreated diabetic NOD.
One central question concerns the mechanism(s) allowing the metabolic
reconstitution in animals presenting with overt disease as we observed
in the present study. Although one cannot exclude that cell
regeneration may occur to some extent, there is compelling evidence
from both the experimental and clinical field to show that at the time
of diabetes onset, there is a sufficient insulin-secreting cell
mass left to allow, if the treatment is started early and if the
immunological assault is efficiently controlled, to recover an adequate
metabolic balance. Thus, Sreenan et al. (18) recently
reported that in NOD mice at the time of overt diabetes, the residual
cell mass still represented 30% of control levels. Moreover, in
these same animals insulin secretion was reduced to a greater degree
than cell mass, suggesting the presence of cell dysfunction in
addition to reduced mass. This is fully in keeping with the notion that
the physical destruction of cells is to a large extent preceded by
a phase of reversible T cell-mediated inflammation that translates into
a significant impairment in their capacity to release insulin in
response to conventional stimulations. In fact, heavily infiltrated
pancreatic islets from nondiabetic 13-wk-old female NOD mice failed to
respond with insulin secretion to high glucose concentrations when
examined immediately after the isolation. This inhibition of cell
function was fully reversible upon clearing of the immune cell
infiltrate after 7 days of in vitro culture (19). In vivo
in the mouse, agents, such as antilymphocyte serum (15),
Abs to CD3 (20), and TCR (21) that
rapidly clear the insulitis, induce complete normalization of glycemia
within no more than 24–48 h if applied within the first 7–10 days
from overt diabetes onset. This effect is independent from the
more durable tolerogenic properties some of these treatments may have.
Thus, at variance with CD3 Ab and antilymphocyte serum, Abs to
TCR do not restore self tolerance as well illustrated by the
disease relapse invariably observed within a few days from the end of
treatment (21). Similarly in the clinic, a steady and
durable increase in endogenous insulin production was observed after a
combination of insulin and cyclosporin in recently diagnosed
insulin-dependent diabetes mellitus patients (22, 23).
Adult patients were enrolled to receive cyclosporin or placebo within
1–2 mo after the first clinical symptoms had appeared. The proportion
of disease remission, as assessed by stimulated C peptide production
and the decrease of the insulin need, was significantly higher in
patients receiving the active drug as compared with placebo.
Given the well-established role of integrins in general and LFA-1 in particular in lymphocyte homing (24), and our data showing that sICAM-1/Ig binds to T cell lymphocytes and therefore may block T cell adhesion, we explored whether the disease remission observed upon Ad.sICAM-1/Ig administration could be due to a blockade of the migration of fully differentiated autoreactive effectors present in the diabetic NOD host within the target tissue. The data we obtained using diabetes transfer experiments and Ad.sICAM-1/Ig injections are strongly against this hypothesis. Moreover, the histopathological data recovered from NOD mice in diabetes remission 200 days after i.v. Ad injection, as well as in mice protected from diabetes onset, showed evidence of peripheral insulitis. These results are consistent with recent evidence demonstrating that upon Ag priming TCR-transgenic CD4+ ICAM-1-/- effector cells migrated in the pancreata but did not cause diabetes (3). Another recent study in transgenic NOD ICAM-1-/- mice showed the absence of autoimmune diabetes in these mice, confirming the important role of ICAM-1 in the development of autoimmune diabetes (25). However, in this case, no insulitis was observed in the NOD ICAM-1-/- mice, in contrast to our results.
At the present time, it is difficult to explain the difference in our
observations and studies using anti-LFA-1 and/or anti-ICAM-1,
which have only been shown to be effective in young animals (26, 27). In contrast to anti-LFA-1 Abs, Ad.sICAM-1/Ig did not
significantly alter the TH1/TH2 balance, as assessed in ribonuclease
protection assays in spleen cells from Ad.sICAM-1/Ig-treated mice (not
shown). This is again consistent with the TCR-transgenic
CD4+
ICAM-1-/- model, showing
that the imprinting of CD4+ cell effector
activity occurs independently of TH1/TH2 (3). Nonetheless,
it may be hypothesized that the Abs exclusively exert an inhibitory
effect (28), whereas sICAM-1/Ig is endowed with clear
(stimulatory) signaling potential, as we could demonstrate in the in
vitro transgenic T cell system. Results of the T cell adhesion test
(Fig. 1C) showed that sICAM-1/Ig binds T lymphocytes through
LFA-1, suggesting that T cell adhesion may be blocked by
sICAM-1/Ig. However, sICAM-1/Ig steric hindrance of LFA-1
does not imply that there is a functional blockade of the ICAM-1/LFA-1
pathway. Agonistic effects of sICAM-1 on the ICAM-1/LFA-1 pathway have
already been shown in in vitro systems (29). As expected
(29), and in contrast to CD28 binding to B7
costimulatory molecules (30), the increase in IL-2
secretion was slight, confirming a distinct effect on T cell activation
through the LFA-1/ICAM-1 pathway as compared with CD28 (31, 32).
In experiments in which Ad vectors were injected in prediabetic mice, Ad.null injection induced a short, although not statistically significant, delay in overt diabetes. However, when Ad vectors were injected in overtly diabetic mice, no remission was observed in any Ad.null mouse, suggesting that remission was not induced by Ad injection itself. Another Ad control encoding an irrelevant Ig heavy chain was not included in this study, because the Fc moiety, which provides Ab-like effector functions to sICAM-1 (complement and FcR-bearing cell activation), might participate in the therapeutic effects. The main advantage of using Ad-mediated gene transfer of sICAM-1/Ig is that a single i.v. injection of Ad.sICAM-1/Ig was sufficient to induce very high and sustained levels of circulating ICAM-1. In comparison, daily injections of recombinant sICAM-1 were necessary for 5 mo to prevent diabetes in young NOD mice (33).
Our present data suggest the induction and/or restoration of immunoregulatory or dominant tolerance immune mechanisms that closely resemble those described in young prediabetic NOD (5, 16, 34) mice. Interestingly enough, this may be achieved through the delivery of an agent, sICAM-1/Ig, that is totally deprived of any suppressive activity. Apart from its obvious potential clinical application in recently diagnosed diabetic patients, the possibility of restoring self-tolerance in overtly autoimmune animals has fundamental implications that are important for the understanding of mechanisms regulating autoimmune responses. In vivo studies with the administration of the sICAM-1/Ig protein will be rewarding.
|
Acknowledgments |
|---|
-/- mice. We are also indebted to Isabelle
Cisse for managing the specific pathogen-free mouse animal
facility. |
Footnotes |
|---|
2 Address correspondence and reprint requests to: Dr. Patricia Lemarchand, Institut National de la Santé et de la Recherche Médicale, Equipe 0016, Faculté de Médecine Necker-Enfants Malades, 156 rue de Vaugirard, F-75730 Paris cedex 15, France. E-mail address: lemarchand{at}necker.fr
3 Abbreviations used in this paper: ICAM-1, intercellular adhesion molecule-1; Ad, adenovirus; sICAM-1, soluble ICAM-1; NOD, nonobese diabetic; Ad.sICAM-1/Ig, sICAM-1/Ig adenoviral vector.
Received for publication October 9, 2001. Accepted for publication January 18, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
L2 but not of integrins 4 and/or 7 significantly prolongs intestinal allograft survival in mice. Gut 47:97.-cell proliferation and reduced mass before diabetes onset in the nonobese diabetic mouse. Diabetes 48:989.[Abstract]
-cell suppression in vitro in pancreatic islets isolated from nonobese diabetic mice during the phase preceding insulin-dependent diabetes mellitus. J. Clin. Invest. 85:1944.
T cell receptor monoclonal antibody provides an efficient therapy for autoimmune diabetes in nonobese diabetic (NOD) mice. Eur. J. Immunol. 21:1163.[Medline]
This article has been cited by other articles:
|
|
 |
|
  B. Graf, T. Bushnell, and J. Miller LFA-1-Mediated T Cell Costimulation through Increased Localization of TCR/Class II Complexes to the Central Supramolecular Activation Cluster and Exclusion of CD45 from the Immunological Synapse J. Immunol., August 1, 2007; 179(3): 1616 - 1624. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Ke, D. Sun, P. Zhang, G. Jiang, H. J. Kaplan, and H. Shao Suppression of Established Experimental Autoimmune Uveitis by Anti-LFA-1{alpha} Ab Invest. Ophthalmol. Vis. Sci., June 1, 2007; 48(6): 2667 - 2675. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
U. Sela, N. Mauermann, R. Hershkoviz, H. Zinger, M. Dayan, L. Cahalon, J. P. Liu, E. Mozes, and O. Lider The Inhibition of Autoreactive T Cell Functions by a Peptide Based on the CDR1 of an Anti-DNA Autoantibody Is via TGF-{beta}-Mediated Suppression of LFA-1 and CD44 Expression and Function J. Immunol., December 1, 2005; 175(11): 7255 - 7263. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Huang, K. Matthews, T. J. Siahaan, and C. G. Kevil {alpha}L-Integrin I domain cyclic peptide antagonist selectively inhibits T cell adhesion to pancreatic islet microvascular endothelium Am J Physiol Gastrointest Liver Physiol, January 1, 2005; 288(1): G67 - G73. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kretowski, N. Wawrusiewicz, K. Mironczuk, J. Mysliwiec, M. Kretowska, and I. Kinalska Intercellular Adhesion Molecule 1 Gene Polymorphisms in Graves' Disease J. Clin. Endocrinol. Metab., October 1, 2003; 88(10): 4945 - 4949. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or Ad.null (
). Complete remission
was defined as the disappearance of glycosuria and a return to normal
glycemia. Concurrently, sICAM-1/Ig serum levels (•) were quantified
by ELISA.
), Ad.null (
). Mice were regularly monitored for glycosuria and
glycemia. Concurrently, sICAM-1/Ig serum levels (•) were quantified
by ELISA.
) or untreated diabetic (
),
and in presence of sICAM-1/Ig (