| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
*
Department of Immunology, Institute for Cell Biology, University of Tübingen, Tübingen, Germany; and
Laboratory of Cellular Physiology and Immunology, The Rockefeller University, New York, NY 10021
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
On the one hand, alloreactive T cells recognize structural determinants of the foreign MHC molecules independent of the bound peptides (2, 3). The recognition of allelic differences in the MHC protein backbone was suggested as a general model of allorecognition by Michael Bevan in 1984 (4). He hypothesized that the strength of the alloreactive response is due to the numerous nonself MHC molecules on the cell surface that trigger a large number of high as well as low avidity T cells directed against polymorphic structures. This theory was supported by the finding of an alloreactive CTL line that could be stimulated by immobilized nonself MHC molecules stripped of bound peptide (5). Additionally, alloreactive CTL clones were found that recognized equally well H-2Kb molecules on TAP-deficient cells, T2.Kb and RMA-S, with or without external addition of peptide extracts from C57BL/6 spleen cells. Moreover, the same clones were also able to recognize acid-stripped Kb molecules on mouse M12Kb cells (6). The structural differences between self and foreign MHC molecules recognized by alloreactive T cells could be as small as one amino acid exchange (7, 8).
On the other hand, alloreactive T cells recognize nonself MHC molecules in a peptide-dependent fashion (2, 3). This mechanism was originally proposed by Matzinger and Bevan in 1977 (9) and explains the strength of the alloreactive response by the large amount of T cells stimulated by the diverse array of self peptides on the nonself MHC molecules. The theory, also known as the multiple binary hypothesis, assumes that the recognition of self peptides on nonself MHC and nonself peptides, e.g., viral peptides, on self MHC by the TCR follows the same principle. This was indeed verified for the mouse CTL clone 2C, which recognizes self H-2Kb presenting the synthetic SIYRYYGL peptide as well as nonself H-2Ld together with the self peptide QLSPFPFDL with similar efficiency (10, 11, 12). The recognition of the peptide ligand on allogeneic MHC molecules by alloreactive T cells was also shown for MHC class II molecules (13, 14). The peptides recognized by peptide-specific, allorestricted CTL can belong to rare as well as to abundant peptide species (15, 16, 17, 18). However, in most cases, different T cell clones recognize different individual peptides or different sets of peptides (3, 19, 20), which can lead to cell type-specific (21, 22) and even tumor-specific (23) recognition of the alloantigen.
Here, we analyze the repertoire of peptide-specific allorestricted human CTL semiquantitatively. We determined the frequencies of T cells reactive against self MHC plus peptide libraries and T cells reactive against foreign MHC with and without peptide. The specificities as well as the peptide affinities of the allorestricted CTL clones obtained were compared with self MHC-restricted CTL recognition. Methodological aspects of generating allorestricted human CTL as well as potential implications for adoptive tumor immunotherapy are discussed.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The human LCL T1 (24) (HLA-A2+, -Aw30+, -Bw6+, -B5+, -Cw1+), the TAP-deficient T1 variant T2 (24) (HLA-A2med, -B5low, -Cw1low), the rat TAP1 and -2a reconstituted T2 variant T3 (25), 721 (26) (HLA-A2+, -A1+, -B5+, -Cw1+), C1R-A2 (provided by Dr. J. Strominger, Cambridge, MA), the TAP-negative LCL line ST-EMO (27) (HLA-A3+, -B63+), and PHA blasts prepared from fresh or frozen PBL obtained from normal HLA-typed donors were used in 51Cr release assays or for T cell stimulation. PHA blasts were generated from HLA-A2+ or HLA-A2- donor PBL in a 5-day culture containing 1 µg/ml PHA (Boehringer Mannheim, Mannheim, Germany). The W6/32 Ab (anti-HLA class I) (28) was purified from hybridoma culture supernatants with protein A-Sepharose beads (Pharmacia, Uppsala, Sweden) by standard procedures.
Generation of RMA-S/A3/hß2m3 transfectants
RMA-S cells (107) were electroporated with 220 V at a capacity of 960 mF using an electroporator device (Bio-Rad, Munich, Germany). A pool of the following three linearized plasmids (10 mg each) was prepared for cotransfection: p44, which encodes the HLA-A3 gene (29) (provided by Dr. B. R. Jordan, Center of Immunology, Marseille-Luminy, France); pß2 m13, which contains the hß2m gene (30); and pHbAPr1-neo (31), which had been modified by replacement of the original polylinker site by the multiple cloning site of pSp72 (Promega, Heidelberg, Germany) as a plasmid contributing neomycin resistance. The latter two plasmids were gifts from Dr. F. Momburg, German Cancer Research Center (Heidelberg, Germany). After selection with 1.0 mg/ml G418 (Life Technologies, Paisley, U.K.), HLA-A3-expressing transfectants were detected with mAb B9.12 (32) followed by staining with an FITC-conjugated goat anti-mouse F(ab')2 (Dianova, Hamburg, Germany) and were separated from negative cells by flow cytometry using a FACSVantage cell sorter (Becton Dickinson, Heidelberg, Germany). After one repeated round of sorting, cells were expanded in culture medium containing 0.8 mg/ml G418.
Generation of CTL clones
PBL from healthy donors registered in the Blood Bank
(Tübingen, Germany) were isolated from buffy coats by
Ficoll-Hypaque density gradient centrifugation using FicoLite-H
(Linaris, Bettingen, Germany). PBL (106) were stimulated
with 105 irradiated T2 cells (200 Gy) that had been pulsed
with 100 µM HLA-A2 peptide libraries or peptides overnight in
FCS-free medium or with 105 irradiated ST-EMO cells (200
Gy) pulsed with 100 µM of the HLA-A3 peptide library. After 5 days of
culture, these PBL were seeded at 30 or 100 cells/well in 96-well
plates with 104 irradiated stimulators (pulsed irradiated
T2 or ST-EMO) and 105 syngeneic irradiated PBL (30 Gy) as
feeders. The cultures were restimulated weekly in the same fashion and
tested in a 51Cr release split-well assay against T2+/-
peptides or RMA-S/A3/hß2m+/- peptides after 2 wk or more
of total culture time. Wells containing CTL preferentially recognizing
peptide-loaded targets were expanded. All T cell cultures were
performed in 
-MEM (Life Technologies), 5% human serum (Diagast,
Jülich, Germany), 10 U/ml IL-2 (Lymphocult, Biotest, Dreieich,
Germany), 2 mM glutamine (BioWhittaker, Verviers, Belgium), and 50 U/ml
penicillin/50 µg/ml streptomycin solution (BioWhittaker).
51Cr release assay
Targets were labeled with 50 µCi of Na251CrO4 for 45 min at 37°C, with or without preincubation with peptide for 1–2 h at 37°C in FCS-free medium. Labeled targets were incubated for 4 h with the CTL in RPMI (Life Technologies), 10% FCS (Sigma, St. Louis, MO), 2 mM glutamine (BioWhittaker), and 50 U/ml penicillin/50 µg/ml streptomycin solution (BioWhittaker). Subsequently, 50 µl of the supernatant was harvested, and radioactivity was measured in a microplate format scintillation counter (1450 Microbeta Plus, Wallac, Turku, Finland) using solid phase scintillation (LumaPlate-96, Packard, Groningen, The Netherlands). The percent specific lysis was calculated as ((cpm experimental well - cpm spontaneous release)/(cpm maximum release - cpm spontaneous release)) x 100%. Spontaneous release was determined by incubating the labeled targets with medium. Maximum release was determined by incubating the target cells in 1% Triton X-100 solution.
Peptides
Peptides were synthesized by the F-moc solid phase method on a
peptide synthesizer (model 432 A, Applied Biosystems, Foster City, CA).
The peptides were analyzed by reverse phase HPLC (System Gold, Beckman,
Palo Alto, CA) and mass spectrometry (LD-TOF G2025A, Hewlett-Packard,
Palo Alto, CA). HLA-A2 and -A3 libraries were synthesized as shown in
Table I
. In addition, 26 sublibraries of
the original HLA-A2 library were used; in each sublibrary one of the
positions 1, 3, 4, 5, 6, 7, or 8 was fixed, while the other positions
maintained the original complexity. For some experiments, the HLA-A2
library was also fractionated by reverse phase HPLC on a System Gold
(Beckman) and concentrated in a vacuum concentrator (Bachofer)
for use in 51Cr release assays.
|
T2 or ST-EMO cells (106) were incubated with 100
µM peptide solution in FCS-free medium. The incubation with T2 was
performed overnight, whereas ST-EMO was incubated with the peptides for
4 h only. HLA surface expression was monitored after staining with
the primary Ab, W6/32, and secondary Ab, FITC-coupled goat -mouse
IgG (Dianova, Hamburg, Germany), on a FACSCalibur cytometer (Becton
Dickinson).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
PBL from an HLA-A2- donor were stimulated in bulk
culture for 5 days against T2 loaded with the HLA-A2 peptide library
and then seeded at 30 cells/well (Fig. 1a) or 100 cells/well (Fig. 1b). These numbers were chosen because previous experiments
had shown that under these conditions only 37% or less of the wells
developed killing activity and were therefore likely to contain CTL
clones (33). After two further rounds of restimulation, wells were
tested in a split-well 51Cr release assay on T2 cells or T2
cells loaded with the HLA-A2 peptide library. A representative
experiment is shown in Fig. 1. When 30 cells/well were seeded, seven of
96 wells developed a killing ability above background (20%; Fig. 1a). Four of these cultures were peptide specific. When 100
cells/well were seeded, nine of 96 wells demonstrated a killing ability
above background; five were peptide specific (Fig. 1b).
While peptide-specific killing was around 30% in most cases (except
30C11), clones that recognized T2 with and without HLA-A2 peptide
library equally well showed a characteristically increased killing
activity, which was probably due to recognition of HLA-A2 backbone
structures independent of the peptide ligand or T2-specific peptides.
The experiment in Fig. 1 is representative of six similar experiments
performed with HLA-A2- PBL from six different donors and
three experiments performed with HLA-A3- PBL from three
different donors (Table II). On the
average, the ratio of peptide-specific vs peptide nonspecific
responding cultures was 1:1.2 for HLA-A2- PBL and 1:1.5
for HLA-A3- PBL. The frequency for alloreactive CTL ranged
from 1:120 to 1:2400 for HLA-A2- PBL and from 1:160 to
1:220 for HLA-A3- PBL; the frequency of peptide-specific
allorestricted CTL ranged from 1:169 to 1:1920 for HLA-A2-
PBL and from 1:180 to 1:411 for HLA-A3- PBL. These
frequency calculations are based on cell numbers harvested after the
5-day bulk culture prestimulation. The wells labeled in Fig. 1 were
expanded for further investigation.
|
|
HLA-A3+ (Fig. 2a) and HLA-A3-
(Fig. 2b) PBL were stimulated in bulk culture with ST-EMO
loaded with the HLA-A3 peptide library. After 5 days of culture, wells
were seeded at 30 cells/well and restimulated twice. Split-well
51Cr release assays were performed using
RMA-S/A3/hß2m target cells either untreated or loaded
with the HLA-A3 library. This mouse cell line shares HLA-A3 expression
with ST-EMO, but carries none of the other HLA molecules of ST-EMO.
Thus, reactivity toward HLA-A3 could be evaluated in the absence of
alloreactivity toward the ST-EMO HLA-B and HLA-C alleles. In the HLA-A3
autologous stimulation, 17 of 96 wells showed killing activity above
the background of 15% (Fig. 2a). Four wells preferentially
lysed RMA-S/A3/hß2m cells without peptide loading, but
upon incubation with the HLA-A3 library this recognition was lost.
Thirteen cultures recognized RMA-S/A3/hß2m cells only
after loading with the HLA-A3 peptide library and are therefore
specific for peptides of the library. In the HLA-A3 allogeneic
stimulation, 25 of 96 wells showed killing activity above background
(Fig. 2b). Four wells were again specific for
RMA-S/A3/hß2m cells without the HLA-A3 peptide library
and did not recognize these targets after incubation with the library.
Thirteen wells recognized peptide-loaded and untreated
RMA-S/A3/hß2m cells equally well. Killing could be due
either to HLA-A3 backbone structure recognition or to reactivity toward
TAP-independent RMA-S/A3/hß2m peptides that could not be
replaced upon external peptide incubation. Finally, eight wells
demonstrated peptide-specific allorestricted killing of
RMA-S/A3/hß2m cells above background. Therefore, the
frequency of HLA-A3 library-specific CTL was eight in 96 wells with
allogeneic stimulation and 13 in 96 wells with autologous stimulation.
Since less than one-third of the wells showed killer activity in each
case, it could be assumed that one well contained one CTL clone (33).
The ratio of numbers of library-specific CTL in allogeneic vs
autologous stimulation was 1:1.6. Thus, the repertoire of PBL contains
approximately twice the amount of self-restricted CTL as allorestricted
CTL.
|
Five of the nine cultures labeled in Fig. 1
could be expanded to
CTL lines. Three were indeed specific for the peptide library and did
not react with any other target (Fig. 3,
a–c). Another specificity is demonstrated by the CTL line
10E11, which recognizes T2 with and without the peptide library (Fig. 3d). However, the TAP+ cell lines T1 (data not
shown), T3 and C1R-A2, as well as syngeneic or HLA-A2+
(data not shown) PHA blasts were not recognized. Still another
specificity is demonstrated by the 30F7 clone that not only recognized
T2 upon loading with the HLA-A2 library, but also recognized
TAP+ cells such as T1 (data not shown), T3, and C1R-A2. It
is therefore likely that a peptide from the library is recognized that
is identical or cross-reactive to an endogenously produced peptide on
C1R and T1 cells. Syngeneic and HLA-A2+ (data not shown)
PHA blasts were not recognized by the 30F7 clone.
|
The HLA-A2 peptide library was separated by reverse phase HPLC
chromatography (Fig. 4e).
Fractions between 12–35 min were collected in 0.5-min intervals at a
volume of 100 µl; 2.5% of these fractions were used in dilutions of
1/40, 1/400, and 1/4000 for sensitization of T2 cells and were tested
for recognition by the CTL clone 10F4 (Fig. 4, a–c). 10F4
recognized three fractions that elute in a small time window at 26 min,
indicating that peptides related in hydrophobicity are recognized. In
contrast, the CTL clone 30F4 recognized three different HPLC fractions
eluting at 14.5, 29, and 32 min (Fig. 4d).
|
, b–d, f, and
g), while for others, recognition was more or completely
degenerate (Fig. 5, a and e). Positions 2 and 9
were not tested, since these amino acids function as anchor residues
for HLA-A2 and are less likely to influence TCR recognition. Y was
preferred at position 3; G was preferred at positions 4 and 5. In
position 6, the aliphatic residues I, V, and L were recognized by 10F4
with similar efficiency. Finally, P was preferred in position 7, and E
was preferred in position 8. By aligning the most efficiently
recognized amino acids of the sublibraries, the sequence
(Y/A/I/L/S)(L/M)YGG(I/V/L)PE(V/L) was predicted to be more efficiently
recognized than the original library. After choosing the preferred
residue at each position and the strongest HLA-A2 motif anchor residues
at 2 and 9 (34), the peptide LLYGGVPEV was synthesized. This peptide,
in contrast to a control peptide (influenza MP58–66,
GILGFVFTL), was able to sensitize T2 cells for lysis by the CTL clone
10F4 (Fig. 6a).
|
|
The peptide affinity of the CTL clone 10F4 was evaluated in three
ways. 1) Peptide titration of the original HLA-A2 library was performed
(Fig. 6b). Half-maximal lysis was obtained upon incubation
with about 250 nM of the library. If only one peptide from the library
is recognized, this would indicate a peptide concentration of 7.2 pM
peptide for half-maximal lysis. 2) Titration of the recognized HPLC
fractions was performed (Fig. 4, a–c). At 1/40 and 1/400
dilutions the fractions are still recognized above background killing.
Because 50 µg of the library was used for HPLC separation, each
peptide was represented at 1.45 ng. Thus, 36 pg/single peptide (1/40),
3.6 pg/single peptide (1/400), and 0.36 pg/single peptide (1/4000) were
used in the assay (Fig. 4). This represents concentrations of 0.18 nM
for the 1/40 dilution and 18 pM for the 1/400 dilution. Therefore, if
the HPLC fraction contained only one peptide recognized by 10F4, the
peptide concentration for half-maximal lysis is around 18 pM at the
1/400 dilution. 3) The titration of the single peptide recognized by
10F4, LLYGGVPEV, gave half-maximal lysis at a concentration of 2.5 nM.
The 100-fold lower affinity for the individual peptide, LLYGGVPEV,
indicates that more than one peptide in the libraries and relevant HPLC
fraction is recognized by 10F4. However, the peptide affinity of this
allorestricted peptide-specific human CTL clone allows half-maximal
lysis upon incubation of target cells with at least 2.5 nM peptide. In
addition, the clone 30F7 recognizes T1, indicating sufficient affinity
to recognize a naturally processed peptide (Fig. 3e).
Binding properties of the peptides used
The peptides used for stimulation and 51Cr release
assays were also tested for binding to the respective HLA alleles. The
HLA-A2 binding peptides were added to T2 cells and incubated overnight.
LLYGGVPEV was found to stabilize HLA-A2 nearly as efficiently as the
influenza MP58–66 peptide, GILGFVFTL (22) (Fig. 7). Both were less efficient at
stabilizing HLA-A2 than the HLA-A2 library. The HLA-A3 library was
incubated for 4 h with ST-EMO and increased HLA-A3 surface
expression threefold. As a negative control for HLA-A2 as well as
HLA-A3 binding, the HLA-B27 binder GRLTKHTKF (60S ribosomal protein L36
(rat) 36–44) (22), was used. This peptide could stabilize neither of
the two HLA alleles used in this study.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). The frequencies for total alloreactive CTL
ranged from 1:84 to 1:1066. This is comparable to published frequencies
for alloreactive human CTL of 1:200 to 1:400 (35). The ratio of
peptide-specific allorestricted CTL to peptide-independent alloreactive
CTL was slightly smaller than 1 for both HLA-A2- PBL and
HLA-A3- PBL. Even with the larger HLA-A2 library, however,
only 45% of alloreactivity could be clearly assigned as peptide
specific. Comparing TAP-negative mutants and their mother cell lines,
the percentage of peptide-specific CTL in an alloresponse has been
reported to be 85% in humans (36) and 89% in mice (37). This
indicates either insufficient loading of the used libraries or a larger
complexity of the endogenously derived peptide pool. However, binding
to the respective HLA molecules could be shown for both libraries by
FACS staining after stabilization of HLA molecules upon incubation with
external peptide overnight (Fig. 7, d and g). The variations observed among the experiments (conducted with PBL from six unrelated A2- donors and three A3- donors) are probably due to the bias of the donor immune system for cross-reactivity toward nonself MHC class I molecules. It has been demonstrated, for example, that the environmental influence of EBV infection increases a CTL population in HLA-B8+/B*4402- patients that recognizes an EBV peptide on HLA-B8 as well as HLA-B*4402 molecules in a alloreactive fashion (38). The shaping of the T cell repertoire upon positive selection toward recognition of self MHC molecules also directs it toward recognition of allo-MHC molecules that are similar to self (7). Vice versa, negative selection probably also deletes in part allorestricted CTL-recognizing self peptides on allo-MHC molecules that are closely related to self MHC. Another factor influencing the extent of allorecognition is probably the inherited TCR variability of the donor. The TCRß rearrangement in particular is largely unaffected by thymic selection processes, since it is fixed by allelic exclusion before interaction with MHC-peptide complexes takes place (39, 40). This may explain at least in part the restricted usage of TCR receptor genes in alloreactive CTL clones (41). These considerations probably apply not only to the ability of a T cell repertoire to mount alloreactive responses, but also to the ability to recognize peptides in an allorestricted manner. The preferences reported for peptides in cytolytic responses in autologous situations (42) might also be explained by this. Biases of the individual T cell repertoires make it difficult to select peptides suitable for efficient in vitro stimulation, especially in the context of a restriction element that was not involved in positive selection. We predict that while one peptide meets a sufficient amount of specific CTL precursors, others are recognized by very few naive T cells or none at all. To circumvent such unpredictable variation among PBL from different donors, HLA-A3 and HLA-A2 peptide libraries were used under the assumption that the overall response would be similar even though T cells educated by different HLA haplotypes recognize different fractions of the library. It is therefore difficult to draw conclusions from our data to estimate frequencies for individual peptides, which may vary greatly from peptide to peptide and from donor to donor.
In our alloreactive stimulations five different CTL specificities
could be found. The peptide-independent CTL that formed 25–80% (Table II) of the reactive wells in our stimulations probably recognize MHC
backbone structure regardless of the bound peptide (5, 6). A second and
a third specificity are demonstrated by CTL that preferentially
recognize T2 either exclusively without or also with externally added
peptide (Figs. 2 and 3d). In the absence of a sufficient
peptide supply in the endoplasmic reticulum by TAP, the peptides
that reach this compartment by alternative routes, for example as
signal peptides, are loaded onto MHC class I molecules (43). The CTL
displaying specificities 2 and 3 might therefore recognize T2-specific
peptides that are either replaced by external addition of the peptide
library (specificity 2) or are not (specificity 3). It is also possible
that an HLA-A2 structure is recognized that is expressed only in the
absence of TAP (e.g., partially unfolded 1 or 2 domains) and can
be disturbed by external peptide addition or not. A fourth specificity
recognizes peptides from the peptide library (Fig. 3e).
These peptides, however, are also endogenously produced and loaded onto
HLA class I molecules in cells such as T3 and C1R-A2. Due to their
TAP-dependent transport into the ER, they are not present on T2 cells.
Finally, a fifth specificity is represented by CTL that recognize
peptides from the libraries that are not endogenously produced by
TAP+ cell lines; these CTL therefore recognized only T2
plus the HLA-A2 library (Fig. 3, a–c). Probably foreign
peptides are recognized by these CTL. In the autologous stimulation,
only specificities 2 and 5 could be found (Fig. 2a), i.e., a
smaller part of a peptide library can be recognized by self-restricted
CTL compared with the part recognized by nonself-restricted CTL. Since
only specificities 4 and 5 recognize peptides from the library and as
such are of interest for the purpose of this study, these were
investigated further.
The peptide specificity and affinity of the allorestricted CTL clone
10F4 were determined in greater detail. After HPLC separation of the
HLA-A2 library it was found to recognize three fractions eluting around
26 min (Fig. 4, a and b). By using sublibraries
of the original HLA-A2 library one ligand of 10F4 was identified as the
peptide LLYGGVPEV (Fig. 5). This peptide is also contained in the
original HLA-A2 library. In positions 3, 4, 5, 7, and 8 there were
clear-cut preferences for amino acids (Fig. 5). In the TCR-MHC class I
(HLA-A2) crystal structure reported by Garboczi et al. (44) the central
position 5 of the bound Tax peptide is in close contact with the TCR
via its CDR3 and CDR3ß loops and is therefore the ligand position
covered by the TCR most specifically. Our CTL clone 10F4 shows high
specificity for the particular amino acid glycin in positions 4 and 5
(Fig. 5, c and d). This could indicate that
positions 4 and 5 are in close contact with the CDR3 and CDR3ß
loops of TCR(10F4). The important role of P4 in the peptide ligand
resembles findings relating to the molecular modelling of the 2C TCR
crystal structure complexed with the MHC class I molecule
H-2Kb crystal structure bearing the peptide dEV8 (45). In
addition, TCR(10F4), like TCR(A6) and TCR(2C), is in contact with much
of the length of the peptide ligand, except for positions 1, 2, and 9.
Therefore, HLA-A2 recognition by the allorestricted CTL 10F4 follows
similar rules as the autologous recognition of HLA-A2.
The similarity in recognition by auto- and allorestricted CTL was
further strengthened by the avidity of the CTL clone. From titrations
of the LLYGGVPEV peptide, the peptide concentration for half-maximal
lysis of targets was determined to be 2.5 nM, while the HLA-A2 library
containing 34,560 peptides was recognized with a 100-fold lower
avidity. This indicates that more than one peptide was recognized from
the library. This is also suggested by the results with the
sublibraries. In positions 1 and 6, 10F4 shows a degenerate recognition
of the HLA-A2 peptide ligand. In addition, it can be assumed from the
binding motif of HLA-A2, in the anchor positions 2 and 9 the amino
acids used in the library mediate binding equally well. From these
considerations one can predict that at least 60 peptides (5 x
3 x 2 x 2) can be recognized by 10F4 in the library.
Nevertheless, the affinity of 10F4 toward the individual peptide
LLYGGVPEV is comparable to or even higher than the affinity of
allorestricted and xenorestricted CTL responses to self peptides on HLA
molecules (15, 17). Moreover, it is also comparable to or even higher
than affinities of self-restricted CTL responses against foreign
peptides. Van den Eynde et al., for example, reported the concentration
for half-maximal lysis of an anti-GAGE1-specific CTL clone with 100
nM (46). Similar affinities have also been described for viral
epitopes: 0.1 (47) or 30 nM (48) against influenza
MP58–66, 25–427 nM of different CTLs against different
epitopes from HIV type 1 gp160 protein (49), and around 50 nM against
the EBV EBNA-3A603–611 epitope (50). Another important
indication that we were able to elicit high avidity CTL after in vitro
priming is the fact that the CTL clone 30F7 cross-reacted on
TAP+ targets, demonstrating that the amount of endogenously
produced and MHC class I-loaded peptides was sufficient for recognition
(Fig. 3e).
In conclusion, we have shown that primary allorestricted, peptide-specific human CTL could be obtained by in vitro induction with a frequency of around 1:500. These allorestricted CTL showed reactivity toward foreign (e.g., 10F4) as well as self peptides (e.g., 30F7). In addition, the recognition of foreign MHC plus peptide was similar in specificity and avidity to conventional self MHC-restricted T cell recognition. Similar data for mice have recently been obtained (51).
Allorestricted CTL might be useful for tumor immunotherapy as has been postulated (52) and was recently demonstrated by delayed tumor growth upon adoptive transfer of allorestricted mdm2-reactive CTL in mice (23) and the existence of allorestricted cyclin D1-reactive CTL in humans (53). In cancer patients, tumor rejection by the immune system is often inefficient because of tolerance toward the tumor tissue, low immunogenicity of the tumor cells or partial destruction of the immune system by conventional antitumor chemotherapy or factors produced by the tumor cells. In these cases, adoptive transfer of allorestricted CTL created against tumor-associated peptides might present a possibility of reducing or even eliminating the tumor. Since allogeneic cells are rejected in a healthy recipient, this therapy would apply for a particular set of patients only, e.g., after immunosuppression. One field of application that seems especially suited is the allogeneic bone marrow transplantation. Cotransfusion of allogeneic lymphocytes together with the bone marrow into these immunosuppressed patients is believed to be responsible not only for the often observed graft-vs-host disease but also for the beneficial graft-vs-leukemia effect (54). In addition, delayed donor lymphocyte transfusions have been shown to be beneficial in cases of relapsing leukemia after bone marrow transplantation (55, 56). Instead of such rather crude treatment with nonseparated donor lymphocytes, the adoptive transfer of antitumor, allorestricted CTL with known specificity would allow a specific graft-vs-leukemia effect but avoid graft-vs-host disease.
|
Acknowledgments |
|---|
for expert technical assistance, and Britt Anderson for
her contribution to the earlier section of this project. We also thank
H. de la Salle for the kind gift of ST-EMO cells. |
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. H.-G. Rammensee, Department of Immunology, Institute for Cell Biology, University of Tübingen, Auf der Morgenstelle 15, 72076 Tübingen, Germany. E-mail address:
3 Abbreviation used in this paper: hß2m, human ß2-microglobulin.
Received for publication July 7, 1998. Accepted for publication August 31, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
helices. J. Immunol. 158:2805.[Abstract]
. 1997. MHC Ligands and Peptide Motifs Springer, Landes Bioscience, Austin.
. 1995. MHC ligands and peptide motifs: first listing. Immunogenetics 41:178.[Medline]
CDR1 and CDR2. Science 273:963.[Abstract]
ß T cell receptor structure at 2.5 A and its orientation in the TCR-MHC complex. Science 274:209., L. Tussey, K. Smith, S. Rowland Jones, V. Braud, A. McMichael, H. G. Rammensee. 1995. HLA-A2 subtypes are functionally distinct in peptide binding and presentation. J. Exp. Med. 182:1847., H.-G. Rammensee. 1998. Generation of allo- and self-restricted cytotoxic T lymphocytes against a peptide library: evidence for a functionally diverse allo-restricted T cell repertoire. Eur. J. Immunol. 28:2432.[Medline]
This article has been cited by other articles:
|
|
 |
|
  H. Conrad, K. Gebhard, H. Kronig, J. Neudorfer, D. H. Busch, C. Peschel, and H. Bernhard CTLs Directed against HER2 Specifically Cross-React with HER3 and HER4 J. Immunol., June 15, 2008; 180(12): 8135 - 8145. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. G. Schuster, D. H. Busch, E. Eppinger, E. Kremmer, S. Milosevic, C. Hennard, C. Kuttler, J. W. Ellwart, B. Frankenberger, E. Nossner, et al. Allorestricted T cells with specificity for the FMNL1-derived peptide PP2 have potent antitumor activity against hematologic and other malignancies Blood, October 15, 2007; 110(8): 2931 - 2939. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Landais, A. Morice, H. M. Long, T. A. Haigh, B. Charreau, M. Bonneville, G. S. Taylor, and E. Houssaint EBV-Specific CD4+ T Cell Clones Exhibit Vigorous Allogeneic Responses J. Immunol., August 1, 2006; 177(3): 1427 - 1433. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Pittet, A. Gati, F.-A. Le Gal, G. Bioley, P. Guillaume, M. de Smedt, J. Plum, D. E. Speiser, J.-C. Cerottini, P.-Y. Dietrich, et al. Ex Vivo Characterization of Allo-MHC-Restricted T Cells Specific for a Single MHC-Peptide Complex J. Immunol., February 15, 2006; 176(4): 2330 - 2336. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Geffroy-Luseau, A. Moreau-Aubry, R. Bataille, and C. Pellat-Deceunynck Allogeneic T lymphocytes as a source of peptide-dependent T cells specific for myeloma cells Int. Immunol., September 1, 2005; 17(9): 1193 - 1200. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. H.M. Heemskerk, M. Hoogeboom, R. Hagedoorn, M. G.D. Kester, R. Willemze, and J.H. F. Falkenburg Reprogramming of Virus-specific T Cells into Leukemia-reactive T Cells Using T Cell Receptor Gene Transfer J. Exp. Med., April 5, 2004; 199(7): 885 - 894. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. H. M. Heemskerk, M. Hoogeboom, R. A. de Paus, M. G. D. Kester, M. A. W. G. van der Hoorn, E. Goulmy, R. Willemze, and J. H. F. Falkenburg Redirection of antileukemic reactivity of peripheral T lymphocytes using gene transfer of minor histocompatibility antigen HA-2-specific T-cell receptor complexes expressing a conserved alpha joining region Blood, November 15, 2003; 102(10): 3530 - 3540. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T.-H. Yang, M. Lovatt, M. Merkenschlager, and H. J. Stauss Comparison of the frequency of peptide-specific cytotoxic T lymphocytes restricted by self- and allo-MHC following in vitro T cell priming Int. Immunol., November 1, 2002; 14(11): 1283 - 1290. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Mutis, E. Blokland, M. Kester, E. Schrama, and E. Goulmy Generation of minor histocompatibility antigen HA-1-specific cytotoxic T cells restricted by nonself HLA molecules: a potential strategy to treat relapsed leukemia after HLA-mismatched stem cell transplantation Blood, June 28, 2002; 100(2): 547 - 552. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. G. Cawthon, H. Lu, and M. A. Alexander-Miller Peptide Requirement for CTL Activation Reflects the Sensitivity to CD3 Engagement: Correlation with CD8{alpha}{beta} Versus CD8{alpha}{alpha} Expression J. Immunol., September 1, 2001; 167(5): 2577 - 2584. [Abstract] [Full Text] [PDF]
|
|
|
|
), T3 (
), C1R-A2 (
), and HLA-A2- autologous PHA
blasts (
) were used as targets at the indicated E:T cell
ratios.
