| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Institute of Immunology, University of Oslo, National Hospital, Oslo, Norway; and
Department of Biology, Division of Molecular Cell Biology, University of Oslo, Oslo, Norway
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2315, which is presented by
I-Ed; 110–120 hemagglutinin (I-Ed); 323–339
OVA (I-Ad); and 46–61 hen egg lysozyme (I-Ak).
We denote such APC-specific, epitope-containing Ab "Troybodies."
When mixed with APC, all four class II-specific Troybodies were
1,000 times more efficient at inducing specific T cell activation in
vitro compared with nontargeting peptide Ab. Furthermore, they were
1,000–10,000 times more efficient than synthetic peptide or native
protein. Conventional intracellular processing of the Troybodies was
required to load the epitopes onto MHC class II. Different types of
professional APC, such as purified B cells, dendritic cells, and
macrophages, were equally efficient at processing and presenting the
Troybodies. In vivo, class II-specific Troybodies were at least 100
times more efficient at targeting APC and activating TCR-transgenic T
cells than were the nontargeting peptide Ab. Furthermore, they were
100–100,000 times more efficient than synthetic peptide or native
protein. The study shows that class II-specific Troybodies can deliver
a variety of T cell epitopes to professional APC for efficient
presentation, in vitro as well as in vivo. Thus, Troybodies may be
useful as tools in vaccine development. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
New strategies have therefore been developed that are based on the finding that Ig are themselves processed intracellularly, so that Ig-derived peptides are presented on MHC class II molecules (6, 7, 8, 9). Accordingly, various T cell epitopes have been introduced into the complementarity-determining region (CDR)4 2 and 3 of Ab, which is part of the Ag binding site (10, 11). A major drawback of this strategy is loss of Ab specificity. Circumventing this problem, Baier et al. (12) fused T cell epitopes to the C termini of IgD- and MHC class II-specific Fabs. The T cell activation potential of these molecules, however, was lower than that of Ab-Ag complexes (2, 3, 4). The reduction could be due to the fact that Fabs do not allow for bivalent binding and cross-linking of target molecules. Also, peptides tailing an Ab fragment may be prone to degradation (13).
We have chosen to make rAb that have antigenic peptides integrated into their C regions in such a way that they do not disrupt the Ig structure. In addition, these Ab have been equipped with V regions specific for APC. When the rAb are internalized and degraded by the APC, the T cell epitopes can be loaded onto MHC molecules and presented to T cells. In this study, we introduce the term Troybodies for such Ab, because their effect is comparable to that of the Trojan horse: when they enter a cell (the city of Troy) by receptor-mediated endocytosis (a gate in the city wall), their T cell epitopes (soldiers) are released. We have previously made Troybodies with specificity for IgD. When compared with a peptide-containing Ab with irrelevant specificity, the IgD-specific Troybodies were far more efficient at priming B cells for T cell activation (14).
Previous studies have demonstrated that targeting of conventional Ag-Ab complexes to MHC class II has a positive effect on activation of specific CD4+ T cells (2, 3). In the present work, we describe the construction of MHC class II-specific Troybodies with either of four different model T cell epitopes embedded in their C regions. We show that targeting to APC by use of these Ab results in enhanced Ag presentation and T cell activation in vitro as well as in vivo. The current class II-specific Troybodies have the advantage that all the different professional APC expressing MHC class II molecules may be targeted. Indeed, enhanced presentation is obtained using B cells, dendritic cells (DC), and macrophages as APC in the assays.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
BALB/cABom, C.B-17/lcrBom, C57BL/6JBom, C3H/HeJBom, SCID on
C.B-17 background, and the 2315-specific
TCR-transgenic mice on a C.B-17-SCID background (15) were
all from M&B (Ry, Denmark). The 2315-specific
TCR-transgenic mice on BALB/c background (16) and the
2315-specific TCR-transgenic mice on a
Rag2-/- BALB/c background (unpublished) were
bred in our animal facility. B10.BR (express I-Ak
and I-Ek MHC class II molecules) and B10.A(4R)
(express I-Ak but lack I-E molecules) mice were
from Harlan (Oxon, U.K.).
Cell lines
The 14-4-4S hybridoma (17) was purchased from
American Type Culture Collection (ATCC, Manassas, VA), as were the NS0
cells and the P338D1 (H-2d) macrophage cell line.
The immature bone marrow-derived DC line D2SC/1
(H-2d) (18) was a kind gift from P.
Ricciardi-Castagnoli (University of Milan-Bicocca, Milan, Italy). The B
lymphoma cell A20 (H-2d) and the Fc
R-deficient
B lymphoma cell line IIA1.6 (H-2d)
(19) were kindly provided by S. Amigorena (Institut
Curier, Paris, France). The 91–101
2315-specific,
I-Ed-restricted CD4+ T cell
clone 7A10B2 has been described previously (6). The T cell
hybridoma 3A9 specific for hen egg lysozyme (HEL) 46–61
(20) was a gift from R. Germain (National Institutes of
Health, Bethesda, MD); the T cell hybridoma DO11.10 specific for OVA
323–339 (21) was a gift from P. Marrack (University of
Colorado, Denver, CO); and the hemagglutinin (HA) 110–120-specific
CD4+ T cell clone, Vir-2 (22), was a
gift from A. Rolink (Basel Institute for Immunology, Basel,
Switzerland).
Construction of MHC class II-specific Troybodies
Introduction of T cell epitopes.
The different T cell epitopes were introduced into the
CH1 domain of a human IgG3 H chain by
site-specific in vitro mutagenesis, as previously described (23, 24). The epitopes were introduced into the loop of
CH1 that corresponds to the CDR3 loop of V
domains. We have previously denoted this loop L3, but it is in this
study renamed L6, as it is the sixth loop in the domain when loops are
counted from the amino terminus of the folded polypeptide chain. Each T
cell epitope was added by replacing the 12 nt encoding the 4-aa-long
loop with nucleotides encoding the epitope. Initially, four different 5
iodo-4 hydroxy-3 nitrophenacetyl (NIP)-specific Ab were made (Table I
): L6-2315
contain the 91–101 epitope from the 2315 Ig L
chain (23), whereas the L6-HA, L6-HEL, and L6-OVA Ab
contain the 110–120 HA (25), the 46–61 HEL
(26), and the 323–339 OVA (27) T cell
epitopes, respectively (24). These nontargeting Ab with
peptide are denoted peptide Ab.
|
chain (determinant Ia.7) of the I-E
MHC class II molecule (17). Briefly, the
VL and VH genes were PCR
amplified from cDNA using a set of upstream primers that anneal to all
Ig leader sequences, in combination with a downstream primer annealing
to CH1 (28). The PCR products were
sequenced, and specific PCR primers annealing to the exact ends of the
V region genes were made. These primers were designed to include
restriction enzyme sites (underlined). The primer sequences were: 5'
VL, ggt gtg cat tcc gac att gtt ctg
aca cag tct cc; 3' VL, acg tac gtt
cta ctc acg ctt gat ttc cag ctt ggt gcc; 5' VH,
cag gtc caa ttg cag cag tct gg; 3'
VH, gac gta cga ctc acc tga gga gac
cgt gac tga ggt t. The VL and
VH gene sequences have been submitted to the
European Molecular Biology Laboratory GenBank, with accession
numbers AF292646 and AF292391, respectively. The V region genes giving
specificity for the hapten NIP have previously been described
(29).
Production and purification of Troybodies.
Vectors encoding complete Ab with different specificities and peptide
inserts were obtained by combining V region genes (giving specificity
for NIP or E) with C region genes (with or
without epitope inserts) using a strategy described previously
(14) (see Table I
). The resulting pLNOH2 and pLNO
vectors were transfected into NS0 cells by electroporation, and
transfectants were selected and cloned in medium containing 800 µg/ml
G418. Ab were affinity purified from cell supernatant by use of protein
L (Affitech, Oslo, Norway)- or protein G-Sepharose columns, and Ab
concentrations quantified by an ELISA specific for human IgG3
(23). The recombinant E-specific
Ab are denoted:
E.L6-2315,
E.L6-HA,
E.L6-HEL, and
E.L6-OVA (Table I).
Ab and flow cytometry
Ab and reagents used for flow cytometry were anti-CD19
biotin, streptavidin CyChrome, anti-Mac1 PE,
anti-CD11c FITC (all from BD PharMingen, San Diego, CA), and
anti-human IgG3 PE (Southern Biotechnology Associates, Birmingham,
AL). To detect
E.L6-2315 bound to
I-E molecules, spleen cells were double stained with biotinylated
anti-CD19 and
E.L6-2315 or
L6.2315. Streptavidin CyChrome and PE-labeled
anti-human IgG were used as secondary reagents. Ten thousand cells
were run on FACSCalibur (BD Biosciences, Mountain View, CA) and
analyzed using the winMDI software.
Sorting of B cells, DC, and macrophages from BALB/c spleens
Resting small splenic B cells were negatively sorted on a
FACSVantage (BD Biosciences), as previously described
(30). Briefly, splenocytes were stained with a mixture of
biotinylated mAb with broad non-B cell specificities detected with
streptavidin CyChrome. The nonfluorescent cells were selected in
combination with a narrow forward light scatter (FSC)/side light
scatter lymphocyte gate. To sort DC, splenocytes were stained with
FITC-labeled anti-CD11c in addition to a non-B cell mAb mixture.
CD11c+Mac1- cells were
positively sorted.
Mac1+CD11c- macrophages
were positively selected from splenocytes double stained with
PE-labeled anti-Mac1 and FITC-labeled anti-CD11c.
Short-term Th1 and Th2 cultures
2315-specific Th1- and Th2-polarized
cell lines from TCR-transgenic mice were obtained as described
(14). Briefly, TCR-transgenic lymph node cells were
cultured with irradiated BALB/c splenocytes and synthetic
2315 peptide. To obtain Th1 cultures, rIL-12
and anti-IL-4 11B11 mAb were added at initiation of cultures,
whereas IL-4 was added to induce differentiation into Th2 cells.
T cell proliferation and cytokine assays
Tissue culture medium used was RPMI 1640 with 10% FCS and
supplements, as previously described (31). T cell
proliferation assays were performed essentially as described previously
(6, 14), with some modifications to overcome the class
II-blocking activity (and thereby inhibition of T cell responses) of
the class II-specific Ab. Briefly, various types of APC were incubated
with titrated amounts of the different rAb and incubated for 4 h
at 37°C in microtiter wells. The cultures were then washed three
times before various types of responder T cells were added. When cell
lines (P388D1, IIA1.6, and D2SC/1) were used as APC, the cells were
treated with mitomycin C (Sigma-Aldrich, St. Louis, MO)
(6). Other APC populations were irradiated with 20 Gy,
except for sorted B cells, which were irradiated with only 8 Gy to
maintain their APC function (32). When T cell hybridomas
were used as responders, the APC were not irradiated. The assays were
put up as 200-µl cultures in 96-well flat-bottom microtiter plates
with the following components: 1) APC: spleen cells (5 x
105/well), sorted resting B cells (5
x 104/well), sorted
CD11c+Mac1- DC (5
x 103/well), sorted
Mac1+CD11c- macrophages
(5 x 104/well), macrophage cell line P338D1
(5 x 104/well), DC line D2SC/1 (2 x
105/well), and B lymphomas A20 and IIA1.6 (5
x 104/well). 2) Troybodies and Ags: MHC class
II-specific Troybodies, control Ab, 91–107
2315 synthetic peptide (33),
complete OVA protein (Sigma-Aldrich), 323–339 OVA peptide (gift from
B. Fleckenstein, University of Oslo, Oslo, Norway), 46–61 HEL peptide
(gift from R. Germain), or 110–120 HA peptide (obtained from S.
Degermann and K. Karjalainen, Basel Institute for Immunology,
Basel, Switzerland) were used in titrated amounts, as indicated. 3)
Responder T cells: TCR-transgenic lymph node cells (1 x
105/well, corresponding to 2 x
104 2315-specific T
cells/well), polarized 2315-specific Th1 or
Th2 cells (2 x 104/well), cloned
CD4+ 2315-specific
7A10B2 cells (2 x 104/well), anti-CD4
magnetic bead-purified (DynaBeads; Dynal Biotech, Oslo, Norway)
CD4+ T cells from spleen and lymph nodes from
2315-specific TCR-transgenic mice (2 x
104/well), cloned HA-specific
CD4+ Vir-2 T cells (2 x
104/well), HEL-specific T cell hybridoma 3A9
(5 x 104/well), or OVA-specific T cell
hybridoma DO10.11 (5 x 104/well). The
experiments with naive T cells from SCID or
Rag-/- mice were performed differently, in that
nonirradiated TCR-transgenic SCID or Rag2-/-
spleen cells (5 x 105/well) were used as
source of both APC and naive T cells. After 48 h, 50 µl of
supernatants were collected for cytokine measurements, and the cultures
were pulsed for 16–24 h with 1 µCi of
3[H]TdR (Amersham, Little Chalfont,
U.K.). The cultures were harvested, and incorporated
[3H]TdR was measured using a Matrix 96 beta
counter or a TopCount NXT scintillation counter (Packard,
Meriden, CT). IFN- concentration in the supernatant was quantified
by sandwich ELISA (34). IL-4 concentration in the
supernatant was measured by a similar sandwich ELISA, except for the
capture mAb, which was 11B11 (ATCC), and detection Ab, which was
biotinylated anti-mouse IL-4 (BD PharMingen). CTLL-2 cells were
used in a bioassay to measure IL-2 concentrations in the culture
supernatants (1/3 dilution) (34). Data in
Figs. 2–8 are
displayed as means of triplicates, and error bars
illustrate SEM.
|
|
|
|
|
|
|
Priming of APC in vivo
C.B-17 or BALB/c (IgH congenic H-2d mice) were injected i.v. in the tail vein or s.c. in the right flank region with titrated amounts of class II-specific Troybodies, nontargeting peptide Ab, synthetic peptide, or complete protein. Ninety minutes after i.v. injections, or 24–72 h after s.c.injections, the mice were killed by cervical dislocation, and the spleens or draining lymph nodes were removed. Irradiated (8 Gy) spleen or lymph node cells were cultured with responder T cells, as described above, but without addition of Ag.
In vivo blastogenesis and expansion of T cells
BALB/c mice were reconstituted with lymph node cells and
splenocytes from 2315-specific TCR-transgenic
mice (16). A total of 15.8 x
106 cells was injected i.v. into each mouse
(2.1 x 106 of these cells were
CD4+V
8.2+). On the
following day, BALB/c mice were injected s.c. in the right flank region
with the indicated amounts of MHC class II-specific Troybodies,
nontargeting peptide Ab, or synthetic peptide. All mice received 1
x 105 U of GM-CSF at the same site concurrently
with the Ag immunization. Two mice received PBS only. After 4 days,
mice were killed by cervical dislocation. Draining (inguinal) and
nondraining (mesenteric) lymph nodes were isolated and prepared for
flow cytometric staining. The cells were triple stained with
FITC-conjugated anti-V8.2 (F23.1; BD PharMingen), biotinylated
GB113 mAb clonotype specific for the transgenic TCR (35),
and APC-conjugated anti-CD4 (BD PharMingen). The biotinylated GB113
mAb was detected by streptavidin CyChrome. The cells were run on a
FACSCalibur cytometer.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This study addresses whether class II-specific Troybodies may
efficiently deliver T cell epitopes for class II presentation to
CD4+ T cells. To this end, Troybodies were
constructed that have V regions specific for class II molecules, and
various model T cell epitopes in the constant part. As model T cell
epitopes, the following were used: aa residues 91–101 from the mouse
2315 Ig L chain (33), 323–339
OVA (27), 110–120 HA (25), and 46–61 HEL
(26). The 2315 epitope is
presented on I-Ed class II molecules to
CD4+ T cells (6, 33), as is the HA
epitope (36). The HEL epitope is presented on
I-Ak class II molecules (20),
whereas OVA is presented on I-Ad
(27). These T cell epitopes have previously been
introduced into CH1 of the human 3 chain and
expressed with hapten (NIP)-specific V regions (Table I). More
specifically, such nontargeting peptide Ab were made by genetically
exchanging loop 6 (L6) of CH1 of human IgG3 with
the various T cell epitopes. L6 corresponds to the CDR3 loop in the V
region and connects -strands of the domain. Because V domains can
accommodate large sequence variation in their CDR3 loops, the L6 loop
was chosen as the site for peptide introduction. Indeed, by analogy to
CDR3 (10, 11), the L6 loop appears to accept engraftment
of new peptide sequences (14, 24).
Troybodies with MHC class II specificity were constructed by replacing
the NIP-specific V regions of the four different L6 Ab (Table I) with V
region genes cloned from the 14-4-4S hybridoma. The 14-4-4S hybridoma
produces a mAb specific for the murine E chain
(Ia.7) and binds I-E molecules of the haplotypes
H-2k, H-2d,
H-2p, and H-2r
(17). The four different Troybodies thus derived are
denoted E.L6-2315,
E.L6-HA,
E.L6-OVA, and
E.L6-HEL (Table I). The specificity of the
Troybodies was verified by flow cytometry. Fig. 1, left panel, shows that
E.L6-2315 bound to
I-E-expressing BALB/c splenic B cells (H-2d),
whereas the NIP-specific L6-2315 Ab did not.
Furthermore,
E.L6-2315 did not
bind splenic B cells from C57BL/6 mice (H-2b)
that do not express I-E molecules (Fig. 1, right panel).
Thus, the specificity of the Troybodies corresponds to that of the
14-4-4S mAb, from which the V regions are derived.
|
2315-specific
stimulation of CD4+ T cells
Class II-specific
E.L6-2315 and
nontargeting L6-2315 Ab were mixed with APC
and compared for their ability to induce specific T cell activation. As
APC, irradiated BALB/c splenocytes were used. The BALB/c strain has the
H-2d haplotype and therefore expresses
I-Ed MHC class II molecules necessary both for
targeting of
E.L6-2315 and for
presentation of the 2315 epitope to specific
CD4+ T cells.
The APC were pulsed with the various Ab for 4 h and washed to
remove excess Ab, and lymph node cells from mice transgenic for
2315-specific TCR were added as responder T
cells. The dose response curves show that the
2315 peptide was presented 1,000 times more
efficiently to TCR-transgenic lymph node cells when part of
E.L6-2315 than
when located in nontargeting L6-2315 peptide
Ab (Fig. 2A). Control Ab with
class II specificity, but without epitope insert
(E.wt), did not elicit any response.
E.L6-2315 was
10,000 times more efficient at stimulating T cell proliferation than
synthetic 2315 peptide, on a molar basis (Fig. 2B). Physical linkage of the epitope and the I-E-specific V
regions seems to be required, because simultaneous addition of both
L6-2315 and E.wt
failed to elicit an increased T cell activation (data not shown).
Possibly, the presence of
E.L6-2315 could
inhibit specific T cell activation, because they could block the same
class II molecules that present the epitope to T cells. This was
studied in an experiment in which the washing procedure performed
before T cell addition was omitted. The results in Fig. 2C
show that the presentation efficiency was dramatically reduced for
high, but not low, Ab concentrations. This indicates that high amounts
of E.L6-2315 in
the supernatant bind and block MHC class II molecules on APC, thereby
reducing T cell stimulation.
E.L6-2315 rapidly
bound APC, as cultures could be washed already after 15-min incubation
without much loss of presentation. As would be expected, no such
blocking effect was observed when nontargeting (NIP-specific) Ab was
used (data not shown).
Class II-specific Troybodies with a variety of model T cell epitopes are efficiently presented
Troybodies with the same I-E class II specificity, but with other model T cell epitopes embedded in loop 6 of their CH1 domain, were tested for their ability to activate specific T cells. The epitopes were 110–120 HA, 46–61 HEL, and 323–339 OVA.
The in vitro antigenicity of the HA-expressing Ab was assessed in dose
response T cell activation assays with cocultures of BALB/c spleen
cells as APC and HA-specific I-Ed-restricted
Vir-2 cells as responder CD4+ T cells (Fig. 3A). The results show that the
HA epitope was presented at least 100-fold more efficient when added as
part of a E.L6-HA compared with its
nontargeted counterpart. Moreover, when compared with synthetic
peptide, E.L6-HA was 100,000 times more
efficient on a molar basis.
In Fig. 3B, we compared nontargeted L6-OVA and targeted
E.L6-OVA in an in vitro T cell
proliferation assay with BALB/c spleen cells as APC and OVA-specific
I-Ad-restricted DO11.10 T cell hybridoma cells as
responders. Based on molar ratios, E.L6-OVA
was at least 1,000 times more effective than L6-OVA and 10,000 times
more effective than complete OVA.
To investigate the Troybody with the HEL epitope, spleen cells from
B10.BR mice were used as APC, because they express the targeted
I-Ek molecule as well as the
I-Ak molecules necessary for presentation of the
HEL epitope. E.L6-HEL was 10,000 times more
efficient than L6-HEL and synthetic HEL peptide on a molar basis. As
would be expected, when APC derived from B10.A(4R) mice, which express
I-Ak but no I-E molecules, were used, the
targeting effect was not observed (data not shown). This demonstrates
that the target molecule of the Troybody has to be expressed on the
surface of APC for enhanced presentation to occur.
Efficient presentation of epitopes located in Troybodies is dependent on intracellular processing
Protein Ag, including Ab, must be partially degraded
intracellularly before fragments can be loaded onto class II molecules
for transport to the APC surface. To verify that the epitopes inserted
in the class II-specific Troybodies needed processing before T cell
recognition, A20 (H-2d) B lymphoma cells were
incubated for various lengths of time with
E.L6-2315,
followed by fixation with paraformaldehyde and addition of specific T
cells. As shown in Fig. 4A,
A20 cells fixed after 120-min incubation at 37°C with
E.L6-2315 induced
T cell stimulation, while cells incubated for 30 or 60 min were only
weakly stimulatory. Furthermore, when chloroquine, an agent known to
inhibit acidification of lysosomes, was added during the 120-min
incubation period before fixation, the APC did not acquire T cell
activation capacity. Interestingly, leupeptin, a cysteine protease
inhibitor, was only weakly inhibitory. This suggests that other
proteases, some of which may not be inhibited by leupeptin, are active
in the intracellular processing liberating the
2315 epitope. Fixation of the APC did not
interfere with the presentation of the 2315
synthetic peptide, consistent with the notion that short peptides can
bind directly to class II molecules without intracellular processing
(data not shown). Similar results were obtained when A20 or IIA1.6
cells were pulsed with E.L6-OVA or
E.L6-HA, except that in these cases,
leupeptin had a much more pronounced effect (Fig. 4, B and
C). Thus, not completely overlapping sets of proteases may
be active in the release of the three different T cell epitopes.
Brefeldin A is known to inhibit egress of newly synthesized class II
molecules by fusing the endoplasmic reticulum and Golgi. Brefeldin A
inhibited T cell stimulation induced by
E.L6-OVA (Fig. 4B) and
E.L6-HA (Fig. 4C), which
indicates that newly synthesized class II molecules are required for
effective class II presentation of Troybodies to
CD4+ T cells. Monensin is another agent known to
inhibit vesicular transport, which results in accumulation of newly
synthesized proteins in the Golgi complex. Monensin blocked
presentation of E.L6-HA (Fig. 4C).
Notably, the OVA epitope is presented by I-Ad,
another class II molecule than the target for
E.L6-OVA (I-Ed) (Fig. 3B). Thus, the liberated T cell epitope is efficiently
presented by another isotype of class II molecules than the one
initially bound by E.L6-OVA on the
surface of the APC. Similarly, the HEL epitope is presented
by I-Ak molecules (Fig. 3C), whereas
E.L6-HEL is targeted to I-E molecules
(I-Ek). Nevertheless, the HEL epitope is
efficiently presented, indicating that binding of
E.L6-HEL to I-Ek
results in processing and presentation by another class II molecule,
I-Ak. Taken together, these findings are
consistent with a requirement for conventional Ag processing of
Troybodies.
B cells, DC, and macrophages efficiently present epitopes delivered by class II-specific Troybodies
Previous results have shown that IgD-specific Troybodies efficiently prime B cells (14, 24). The present class II-specific Troy-bodies should target other professional APC as well, such as DC and macrophages.
To define the role of the different APC involved, we repeated the
experiments with various in vitro cell lines as APC. In all
experiments, pure day 10 or day 20 (stimulated twice) Th2 cells were
used as responder cells. According to the dose response curve, B cell
lymphoma IIA1.6 (Fig. 5A) was
at least 1,000 times more efficient at inducing T cell proliferation
when Ag was delivered by
E.L6-2315 than
when delivered by the nontargeting L6-2315
peptide Ab. When the immature DC line D2SC/1 (Fig. 5B) and
the macrophage cell line P338D1 (Fig. 5C) were used, the
targeting effect was less pronounced (100 times).
Whereas the cultured cell lines are pure, they have a disadvantage, in
that they may differ from their in vivo counterparts. In accordance
with this, both the immature DC line and the macrophage cell line
express only low levels of MHC class II (data not shown) and,
consequently, should be poor APC compared with cells sorted directly
from the spleen. We therefore repeated the experiments with ex vivo
cells purified by cell sorting from BALB/c spleens. Negatively
selected, small B cells (Fig. 5D), positively selected DC
(Fig. 5E), and positively selected macrophages (Fig. 5F) were >1000 times more efficient at presenting T cell
epitopes delivered by
E.L6-2315,
compared with T cell epitopes located in nontargeting peptide Ab.
Similar results were observed when measuring IL-4 in the culture
supernatants (data not shown).
Class II-specific Troybodies enhance activation of naive T cells
Consistent with the results of Fig. 5, SCID spleen non-B cell APC
were found to efficiently stimulate
2315-specific Th2-type cells in the presence
of E.L6-2315
Troybodies (data not shown). However, in this and previous experiments
(
Figs. 2–5), the responder T cells were either cloned T cells, T
hybridoma cells, or lymph node cells obtained from TCR-transgenic mice.
Whereas the clones and hybridomas clearly were not naive, the
2315-specific TCR-transgenic lymph node cells
used in Fig. 2 had not been deliberately activated with Ag previous to
the experiments. Even so, we cannot exclude the possibility that
transgenic lymph node T cells may have been activated via rearranged
and expressed endogenous TCR. Thus, to test naive cells as responders,
we used recombination-deficient 2315-specific
TCR-transgenic mice on a C.B-17 SCID or BALB/c
Rag2-/- background. Nonirradiated spleen cells
served as a source of both APC (macrophages and DC) and T cells.
This approach has the advantage that T cells were not purified or
manipulated, which increases their likelihood of being naive. The dose
response curves show that
E.L6-2315 was
100–500 times more efficient than L6-2315 at
inducing proliferation (Fig. 6A) and IL-2 production (Fig. 6B) of naive T cells. Because the splenocytes were not
irradiated in these experiments (Fig. 6, A and
B), one could argue that we could be measuring proliferation
of other cells than specific naive T cells. We therefore repeated the
experiment with purified CD4+ T cells from
2315-specific TCR-transgenic SCID mice as
responders and irradiated B cell-deficient SCID splenocytes as APC.
Consistent with the results above, strong proliferation of purified
CD4+ T cells was observed; MHC class II-specific
Troybodies were 1,000–10,000 times more potent than nontargeting
peptide Ab and synthetic peptide on a molar basis (Fig. 6C).
Enhanced presentation of Ag delivered by class II-specific Troybodies in vivo
To investigate whether the class II-specific Troybodies could be delivered to splenic APC in vivo, C.B-17 or BALB/c (IgH congenic H-2d mice) were injected i.v. with various amounts of class II-specific Troybodies, the corresponding nontargeting peptide Ab, synthetic peptides, or complete proteins. The spleens were removed after 90 min, and splenocytes were used as APC in T cell activation assays.
The dose response curves in the various systems show that the class
II-specific Troybodies were at least 100 times more efficient at
loading spleen APC with peptide than were the peptide Ab (Fig. 7). Consistent results were obtained when
measuring T cell activation as proliferation of responder T cells from
TCR-transgenic lymph nodes (Fig. 7A), IFN- production
(Fig. 7, B and C), IL-2 production (Fig. 7D), and when measuring proliferation of Th1 cells derived
from TCR-transgenic mice (data not shown). Compared with synthetic
peptides, Troybodies were >100,000 times (HA system, Fig. 7C) and 100–1,000 times (OVA system, Fig. 7D)
more efficient. Compared with complete protein Ag, Troy-bodies were
>1,000 times more efficient (OVA system, Fig. 7D). Similar
results were obtained when titrated amounts of class II-specific
Troybodies (E.L6-HEL), nontargeting peptide
mAb (L6-HEL), and synthetic HEL peptide were injected s.c. in the right
flank region of C3H mice together with 1 x
104 U of GM-CSF. The draining inguinal lymph node
was taken out 24 or 72 h after injection and used as APC for
HEL-specific 3A9 T hybridoma cells (data not shown). Taken together,
the results demonstrate that class II-specific Troybodies have the
ability to reach their APC target in vivo, after both i.v. and s.c.
administration.
MHC class II-specific Troybodies injected s.c. induce clonal expansion of specific T cells
BALB/c mice were reconstituted with splenocytes and lymph node
cells from 2315-specific TCR-transgenic mice
before immunization with titrated amounts of MHC class II-specific
Troybodies
(E.L6-2315),
nontargeting peptide Ab (L6.2315), or
synthetic 2315 peptide in combination with
GM-CSF. Two mice received only PBS. The fraction of
2315-specific TCR-transgenic
CD4+ T cells among lymph node cells after
reconstitution was 1–2% of total CD4+ T cells
(Fig. 8A, upper right
contour plots). Four days postimmunization, the fraction of
TCR-transgenic CD4+ T cells was increased 4-fold
in mice receiving 200 µg of MHC class II-specific Troybodies (Fig. 8, A, upper left contour plots, and
B, left panel). As would be expected, the
expanded TCR-transgenic CD4+ T cells had
increased size (FSC), indicating blastogenesis (Fig. 8, A,
lower contour plots, and B, right).
Importantly, whereas
E.L6-2315 and
L6.2315 increased the cell size of the
2315-specific
V8.2+GB113+ populations,
they did not increase cell size of the
F23.1+GB113- population
(TCR-transgenic CD4+ T cells with endogenous
-chains and host V8.2+ T cells). Thus, only
CD4+ T cells with specificity for the
2315 peptide underwent blastogenesis. These
experiments show that MHC class II-specific Troybodies elicit specific
CD4+ T cell activation and expansion in vivo
after s.c. injections when coadministrated with GM-CSF. GM-CSF alone
apparently had no effects, as no responses above PBS were obtained with
low concentrations of Ag. MHC class II-specific Troybodies were at
least 1,000-fold more potent than synthetic peptide, whereas the
difference was 100-fold compared with peptide Ab (less for
blastogenesis).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
specificity was obtained by
cloning H and L chain V regions from an
anti-E (Ia.7) B cell hybridoma. The V
regions were combined with C regions already containing T cell epitopes
by a cloning procedure described previously (28). In
principle, the procedure should allow introduction of V region genes
from any B cell hybridoma or V gene display library, and makes possible
construction of Troybodies with any chosen specificity.
Four different model T cell epitopes (2315,
HA, OVA, and HEL) were targeted to I-E on APC, and in all cases, an
1000-fold increase in in vitro T cell activation was obtained when
compared with nontargeted peptide Ab. The need for targeting was
further demonstrated by the fact that the APC had to express the I-E
class II molecule for the effect to be seen. The magnitude of the
targeting effect was similar to that observed previously using Ag-Ab
complexes (2, 3) and IgD-specific Troybodies (14, 24). The targeting effect appears higher than that of IgD- or
class II-specific Fabs with T cell epitopes fused to their C termini
(12). Consistent with this, we have previously observed
that Fabs derived from anti-IgD Troybodies are much less efficient
than complete Ab (14). Therefore, bivalent binding or
cross-linking may be crucial. Alternatively, T cell epitopes that are
integrated as loops (10, 11, 24) could be treated
differently by APC than those attached to the carboxyl terminus
(12).
MHC class II presentation of exogenously derived peptides usually
depends on newly synthesized class II molecules (41). In
the present setting, class II molecules also serve as targets for
Troy-bodies, and the T cell epitopes derived from the Troybodies were
found to be presented on either the same (2315
and HA peptides) or on a different (OVA and HEL peptides) isotype of
the class II molecule. Internalization could occur on mature recycling
MHC molecules or on newly synthesized MHC molecules that travel via the
plasma membrane to the endosomal pathway (42). However,
newly synthesized class II molecules constitute only an extremely small
fraction of the class II molecules in the cell membrane
(42), and we therefore find it likely that most of the
class II-specific Troybodies are internalized by recycling mature class
II molecules. Recycling class II molecules are thought not to reach the
late endosomal compartments, in which efficient Ag processing and
loading onto newly synthesized MHC molecules probably occur
(43). Nevertheless, when Troybodies were internalized via
I-E, the OVA and HEL peptides generated could be presented by another
isotype of class II molecules (I-A). Moreover, the inhibition observed
for OVA and HA presentation in the presence of brefeldin A, which
blocks transport of newly synthesized class II molecules out of the
endoplasmic reticulum, suggests that newly synthesized MHC molecules
are needed for Troybody presentation. These findings suggest that
peptides liberated from Troybodies reach the conventional compartment
for class II loading. We do not know, however, where in the endosomal
pathway the partial proteolytic degradation of the Troybodies takes
place.
Because all professional APC express MHC class II molecules, class II-specific Troybodies have the potential to enroll all of these into the immune response. However, the different types of APC (B cells, DC, and macrophages) express different amounts of MHC molecules (44). In addition, they are not identical with regard to other important APC qualities, such as the mechanisms for Ag uptake and the expression of costimulatory molecules (44). We therefore used cell lines as well as ex vivo sorted B cells, DC, and macrophages to investigate the individual role of the three types of APC. Although the in vitro cell lines differed in their capability to activate T cells, ex vivo cells of the various types were remarkably similar.
When the class II-specific Troybodies were targeted to APC in vivo
before isolation of splenic APC, a targeting effect of 100- to
100,000-fold was obtained compared with nontargeting peptide Ab,
synthetic peptides, and complete proteins. This indicates that injected
class II-specific Troybodies have sufficient stability to reach and
bind their target in vivo. Furthermore, in mice adoptively transferred
with specific T cells, class II-specific Troybodies injected s.c. in
combination with GM-CSF induced expansion of specific T cells in
draining lymph nodes. In this respect, the Troybodies were >1,000
times more potent than synthetic peptide on a molar basis. They were
also more efficient than nontargeted peptide Ab, but in this case, the
difference was less (100x). This superior ability of Troybodies to
seek their targets in vivo and deliver their cargo may reflect three
important aspects of these rAb: the APC specificity of their V regions,
the localization of T cell epitopes within the intact Ig structure, and
their bivalent binding to target APC. While class II-specific
Troybodies appear efficient, other receptors on macrophages and DC,
such as those used for microbial Ag, might induce even more powerful
responses, and we are currently probing such receptors for their
potency as Troy-body targets.
For the principle to be useful in vaccine development, some
requirements have to be met. First, Ig domains must accept loading of
different T cell epitopes. To date, we have introduced five different T
cell epitopes (from 2315, OVA, HEL, HA, and
p21ras) into the L6 position in human
IgG3 (23, 24) (I. B. Rasmussen, T. F. Gregers,
and I. Sandlie, unpublished observations). These epitopes vary
in length from 11 to 25 aa and are part of very different secondary
structure motifs within the native Ag. Whereas the 91–101
(2315) epitope is found as a loop in the CDR3
region of the 2315 L chain, the OVA epitope is
mostly a -strand (45), the HEL epitope has three turns
and two -strands (46), and the
p21ras epitope has a -strand/loop/-helix
structure (47). The structure of the HA epitope is
unknown. Despite the differences in lengths and secondary structures,
the five peptides have all been successfully introduced into the L6
loop without disturbing Ab folding and secretion. In addition to using
the L6 loop, we have also loaded all 18 C domain loops of human IgG3
(M. Flobakk, I. B. Rasmussen, T. E. Michaelsen, B. Bogan, and I.
Sandlie, unpublished observations) as well as three loops in the
CH1 domain of mouse IgG2b (48) with
T cell epitopes. Taken together, it seems like the majority, but not
all C domain loops can be used for introduction of many different T
cell epitopes without severely compromising Ab folding and
secretion.
A second requirement is that the inserted T cell epitopes have to be
excised from their new positions by the Ag-processing machinery of the
APC and presented to T cells. Degradation of Ig in APC has been studied
(49), and reduction of disulfide bridges by GILT
(50) as well as proteolytic cleavage by cathepsins and
other proteases (51) are likely to be important. We have
compared rAb with the 2315 epitope in
different positions and found that most loops, but not all, are
effective in terms of peptide presentation (23, 24, 48)
(M. Flobakk, I. B. Rasmussen, T. E. Michaelsen, B. Bogan, and I.
Sandlie, unpublished observations). A likely explanation is that
the different flanking residues surrounding the epitopes in the various
loops could determine which proteolytic enzymes get involved, and the
efficiency by which they cut. In summary, it seems like some loop
positions are more suitable for peptide insertion than others, mainly
due to effects on folding and secretion, but to some extent also due to
influence of the regions flanking the epitopes.
The Troybody technology offers flexibility that could be very useful in the field of vaccine development. First, it allows design of reagents that target T cell epitopes of choice to the desired type of APC. Second, small Ig-derived constructs, such as minibodies, can be developed. Third, the strategy should allow the introduction of many different T cell epitopes into different loops of the same Ab, so that a true multivaccine can be generated.
The discussion above focuses on the vaccine aspect of Troy-bodies. However, Troybodies could also be used for induction of specific T cell tolerance to prevent or ameliorate autoimmune diseases. For example, Troybodies directed to APC in the thymus could induce central T cell tolerance to defined peptides. Similarly, Troybodies directed to appropriate APC, such as certain immature DC (52), could induce T cell anergy or regulatory, suppressive T cells.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 E.L., K.H.W., and I.B.R. contributed equally to this work
3 Address correspondence and reprint requests to Dr. Elin Lunde at the current address: Department of Biology, Division of Molecular Cell Biology, University of Oslo, P. O. Box 1050 Blindern, N-0316 Oslo, Norway. E-mail address: elunde{at}bio.uio.no
4 Abbreviations used in this paper: CDR, complementarity-determining region; DC, dendritic cell; FSC, forward light scatter; HA, hemagglutinin; HEL, hen egg lysozyme; NIP, 5 iodo-4 hydroxy-3 nitrophenacetyl.
Received for publication July 6, 2001. Accepted for publication December 18, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
