| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Biochemistry, Imperial College of Science, Technology and Medicine, London, and * Nuffield Department of Clinical Medicine, John Radcliffe Hospital, Oxford, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, Vß, and Jß gene segments (AV4, BV23, and BJ2S1, respectively)
with similar third complementarity determining regions (CDR3) of the
ß-chains. Interestingly, the V and Vß gene segments employed by
an HLA-B27-restricted Hom-CTL clone are also closely related to AV4 and
BV23, indicating strong selection pressure for AV4, BV23, and related
gene products in the homocysteine-specific TCR. An arginine or lysine
residue frequently appeared at position 93 in the CDR3 of the TCR
-chains from Hom-CTL restricted by HLA-A68 or -B8. This may suggest
a potential salt bridge between the carboxyl group of homocysteine and
specific TCR. TCR usage by HLA-B27-restricted Hom-CTL from unrelated
individuals appears to be less conserved, although two T cell clones
from one individual rearranged the same V gene segments with identical
lengths of CDR3. Implications of these data for the molecular
mechanisms for homocysteine modification of HLA Ags are also discussed. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
HLA class I Ags are expressed on the surface of virtually all nucleated cells in the body, and their function is to present endogenously produced protein fragments to CD8+ CTL. We have recently shown that homocysteine can modify, through disulfide bonding, class I HLA Ags (e.g., HLA-B27, -A68, and -B8) of the cell (6). We have also identified CTL, in patients with ankylosing spondylitis (AS)4 or reactive arthritis (ReA) and also in some healthy subjects, that are capable of specifically lysing autologous B cells that had been treated with homocysteine in vitro (6). It is reasonable to suggest that homocysteine-specific CTL (Hom-CTL) may play an important part in homocysteine-mediated damage to tissues (e.g., blood vessel endothelium or synovial membrane) in vivo. Their presence may also be related to the development of autoimmune diseases. Further understanding of the nature of these CTL is therefore of significance. In this study, we have analyzed TCR usage by Hom-CTL from five patients with AS or ReA.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Five patients suffering from either AS (CR, MM, and MT) or ReA
(AB and BD) were included in this study; their tissue type for HLA-A
and -B alleles are shown in Table I
.
Isolation of Hom-CTL from PBL was as previously described (6). Briefly,
PBL from donors were cultured (2 x 106
cells/well, 24-well Costar plates) in RPMI 1640 supplemented with 10%
FCS, glutamine, penicillin, and streptomycin (Sigma Chemical, Poole,
U.K.) (R10) in the presence of DL-homocysteine (250 µM, Sigma) for 5
days. Recombinant human IL-2 (rhIL-2, a gift from EuroCetus, Amsterdam,
The Netherlands) was added to a final concentration of 10 U/ml on day
3. These cultures were restimulated twice with homocysteine-treated (5
mM, overnight at 37°C in R10), irradiated (3000 rad), autologous PBL
(2 x 106 cells) and rhIL-2 before being tested for
specific cytotoxicity. Immortal B cell lines (BCL) were developed by EB
virus transformation of fresh PBL. C1R cells, a mutant human BCL that
expresses HLA-C molecules and B35 at very low level, were transfected
with HLA-B2705, -B8, or -A68 genes as
previously described (6). CTL cultures were tested for specificity in
standard 51Cr release assays using homocysteine-treated
autologous BCL or C1R cells expressing appropriate HLA Ags as target
cells (see below).
|
Cloning of Hom-CTL
Hom-CTL clones were obtained by limiting dilution of specific CTL in 96-well round-bottom plates (Nunc) at 0.3, 1, 10, and 30 cells/well in the presence of 5 x 103 autologous B cells (treated overnight with 5 mM homocysteine), 2 x 104 allogeneic PBL from three donors, PHA (Wellcome, Dartford, U.K.) at 1 µg/ml, and rhIL-2 at 10 U/ml. The feeder cells were irradiated (3000 rad) immediately before being used. The T lymphocytes were fed twice using a similar mixture of feeder cells and rhIL-2 but in the absence of PHA. Cells that grew after two rounds of restimulation were transferred to 24-well plates (Costar) for expansion and specificity testing. The T cell clones were maintained in vitro by weekly restimulation with homocysteine-treated autologous BCL and rhIL-2 for up to 1 month.
51Cr release assays
BCL or C1R transfectant cells (2 x 106) were treated, or not treated, with DL-homocysteine overnight (5 mM) before being washed and labeled with 100 µCi 51Cr (Amersham, Little Chalfont, U.K.) for 60 min at 37°C. Standard 4-h 51Cr release assays were then performed using Hom-CTL lines or clones as effector cells and homocysteine-treated, 51Cr-labeled B cells (104 cells/well) as targets in round-bottom 96-well plates. Experimental wells and also medium alone (spontaneous release) and 1% Nonidet P-40 (maximum release)-treated target cells were dispensed in triplicates in 200 µl R10. The results were calculated as follows: percentage of specific lysis = [(experimental release - spontaneous release) x 100]/(maximum release - spontaneous release).
Anchor PCR (AnPCR)
Total cellular RNA (5 µg), extracted from approximately
2 x 106 T cells using RNAzol B (Cinna/Biotech
Labs, Liverpool U.K.), was used in standard first-strand cDNA synthesis
using avian myeloblastosis virus reverse transcriptase (Life
Technologies, Paisley, U.K.) and oligo(dT) as a primer. The
resultant cDNA was tailed with dGTP by terminal deoxy-transferase
(Boehringer Mannheim, Mannheim, Germany) before use as template in PCR
reactions. AnPCR reactions were performed using 5' poly(C)
oligonucleotide (5'-CTATCTAGAGCGGCCGCCCCCCCCCCCCCC)
containing a NotI site (underlined) and a C oligo
(5'-GATAGATCTTAGAGTCTCTCAGC), or a Cß oligo
(5'-CGCGAATTCAGATCT CTGCTTCTGATG), containing a
BglII site (underlined). Conditions for PCR amplification
were 94°C for 5 min followed by 30 cycles of 94°C 2 min, 55°C 2
min, and 72°C 2 min using a Perkin-Elmer thermocycler.
Cloning and sequencing of AnPCR products
The AnPCR products were extracted using phenol/chloroform
and digested with the restriction enzymes NotI and
BglII (Amersham) before being run on 2.5% low melting point
agarose gel in TAE buffer. DNA bands corresponding to approximately 400
base pairs were cut out and the DNA purified and cloned into an M13
mp18 vector as described by Moss et al. (7). From each Hom-CTL line, at
least 20 independent TCR - and ß-chain clones were sequenced; and
from each Hom-CTL clone, between 4 and 11 independent TCR - and
ß-chain clones were sequenced. Nomenclature and classification for
TCR V and J gene subfamilies were according to the World Health
Organization-International Union of Immunological Societies (WHO-IUIS)
Subcommittee on TCR Designation (8) and also after Moss et al. (7),
Toyonaga et al. (9), and Arden et al. (10).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Hom-CTL lines were generated by stimulating fresh PBL from donors
with homocysteine in vitro as previously described (6). All CTL lines
studied here were able to lyse specifically homocysteine-treated human
B cells expressing appropriate HLA class I Ags (Fig. 1). Hom-CTL lines from two B27-negative
patients, MM (with AS) and BD (with ReA), were restricted by HLA-A68
and -B8 (Fig. 1, A and C), respectively,
and that from two B27-positive patients, MT (with AS) and AB (with
ReA), were restricted by B27 (Fig. 1, F and
E). T cell clones were obtained, by limiting dilution
at 0.3 cell/well, from the MM and AB Hom-CTL lines. All six T cell
clones (three from each line) lysed homocysteine-treated autologous B
cells at a low E:T ratio (0.3:1) and showed no detectable lysis of
homocysteine-treated, class-I HLA-mismatched B cells at an E:T ratio of
30:1 (not shown). The original Hom-CTL line from patient CR (with AS)
was restricted by both A68 and B27, indicating a mixture of at least
two different populations of T cells in the culture (6). T cell cloning
on this line resulted in a number of oligoclonal Hom-CTL lines (from 30
cells/well), and two of these, CR-A and CR-B (restricted by A68 and
B27, respectively), were included in this study (Fig. 1, B
and D). Hom-CTL line from a healthy subject, ET, was
also confirmed to be restricted by HLA-B27 (not shown).
|
The transcribed TCR gene segments in Hom-CTLs were amplified by
AnPCR and the DNA sequences of the PCR products determined. In the
HLA-B8-restricted BD Hom-CTL line, only 1 ß-chain of BV7S2 rearranged
with BJ2S1 was identified (Table I
). A subsequently repeated experiment
gave the same result (not shown), confirming that the BV7S2/BJ2S1
ß-chain is part of the homocysteine-specific TCR. In contrast, four
different -chains were identified in this line (Table I). Two of the
four -chains used AV7S2 and made up nearly 70% of the in-frame
sequences identified. It was therefore presumed that the
AV7S2-containing -chains might make up part of the
homocysteine-specific TCR, although T cell clones would be needed for
further confirmation. For specificity control, TCR -chain
transcripts of unstimulated PBL from the same donor (BD) were also
analyzed in parallel experiments. As shown in Table II, 16 different -chains made up the
39 in-frame sequences determined. No V or J gene segments were
preferentially rearranged, and none of the 16 -chains was
significantly dominant. Similar analysis was also performed on TCR
ß-chain usage by fresh PBL from donor DB, and no selection of any
particular BV or BJ gene segment was evident (not shown). These results
suggest that the rearrangement of particular V and J gene segments by
BD Hom-CTL was clearly Ag driven.
|
The MM Hom-CTL line was highly specific and efficient in killing
homocysteine-sensitized A68-positive C1R transfectant cells (Fig. 1A). CTL of this line transcribed predominantly a
BV23S1/BJ2S1 ß-chain and two -chains of AV22S1 rearranged with
AJ17S1 and AV4S2 with AJ14S3 (Table I). As expected, all three T cell
clones (MM1, MM2, and MM4) derived from this line used the BV23S1/BJ2S1
ß-chain. They also rearranged the AV4S2/AJ14S3 -chain, although
clone MM1 had an additional -chain transcript of AV13S1/AJ13S2. DNA
sequences of the TCR - and ß-chains of the three clones, except
that of the AV13S1/AJ13S2 -chain of MM1, are identical to that of
the parent Hom-CTL line, showing that they are actually sister clones.
The AV22S1/AJ17S1 -chain accounted for >40% of the in-frame
-chain sequences from the MM Hom-CTL line (Table I), but no T cell
clones using this chain was obtained. It is not possible at the present
stage to determine whether the AV22S1/AJ17S1 was not used by Hom-CTL or
whether AV22S1/AJ17S1-expressing CTL were missed by chance during the
cloning process.
In the A68-restricted Hom-CTL line from patient CR (CR-A), four
different -chains and three different ß-chains were identified
(Table I). The AV4S1/AJ17S3 -chain and BV23S1/BJ2S1 ß-chain were
clearly dominant and therefore likely to represent the
homocysteine-specific TCR. The similarity between these TCR and that of
MM Hom-CTL was striking: they employed not only the same subfamilies of
AV, BV, and BJ gene segments but also nearly identical third
complementarity determining regions (CDR3) of the ß-chains (Tables I
and III). These results indicate strong
selection pressure for AV4 and BV23 gene segments and also conserved
CDR3 in TCR of the A68-restricted Hom-CTL.
TCR usage by HLA-B27-restricted Hom-CTL
The AB Hom-CTL line was less efficient in killing
homocysteine-treated, B27-positive BCL and also showed high background
lysis of HLA-mismatched target cells (Fig. 1E). It
was therefore not surprising that T cells of this line appeared to
express heterogeneous TCRs (Table I). Useful information on TCR usage
by AB Hom-CTL came from three T cell clones (AB1, AB5, and AB7) of this
line. One of them (AB1) rearranged the AV20S1 and BV5S1 gene segments
(Table I), which are closely related to the AV4 and BV23 genes (10)
used by the A68-restricted Hom-CTLs. The other two clones, however,
selected AV8 and BV13 with identical lengths of CDR3 of the ß-chains
(Table III, Table IV).
|
|
), suggesting that recombination of
BV6S4/BJ2S2 must have occurred in two different CTL clones. The CDR3
regions of both ß-chains are six amino acids in length and contain an
"arginine-glycine" motif (Table I). In the absence of Hom-CTL
clones using these ß-chains, however, it is unclear whether or not
they were selected for recognition of homocysteine by the Hom-CTL.
In the B27-restricted MT Hom-CTL line, an AV10S1/AJ16S1 -chain and a
BV22S1/BJ2S7 ß-chain were the only two in-frame TCR transcripts
identified, implying that the line was virtually clonal (Table I).
These TCR chains are not related to that employed by AB Hom-CTL clones
or A68- and B8-restricted Hom-CTL.
Despite their ability to specifically lyse homocysteine-treated B cells
expressing HLA-B27 (Fig. 1D), T cells of the CR-B
line expressed heterogeneous TCRs (Table I). It is difficult to
determine which of the five - and three ß-chains, as shown in
Table I, actually belong to the homocysteine-specific TCR. It should be
pointed out, however, that none of these TCR - and ß-chains is
related to that of MT or AB Hom-CTL. In conclusion, HLA-B27-restricted
Hom-CTL from different patients do not seem to use similar TCR,
although TCR of Hom-CTL clones from the same individual may show a
certain degree of conservation.
As Hom-CTL could also be identified in a small percentage of healthy
subjects, it was of interest to compare TCR usage by these T cells and
those from the patients. TCR - and ß-chain transcripts of the
B27-restricted ET Hom-CTL line were therefore analyzed. As shown in
Table I, the V gene segments employed by ET Hom-CTL are not closely
related to those of T cells from the patients (Table I). Since TCR
usage by B27-restricted Hom-CTL from the patients is not significantly
conserved, however, it is difficult to determine whether the difference
in TCR usage between Hom-CTL from the patients and from the healthy
subject is disease related.
Comparison of CDR3s of TCR from Hom-CTL
It has been shown that the VDJ junctions of TCR - and
ß-chains form CDR3s that are responsible for contacting the antigenic
peptides presented by MHC (11, 12, 13, 14, 15). Tables III and IV compare the
predicted amino acid sequences of the CDR3s of TCR - and ß-chains
of the Hom-CTL included in this study. Interestingly, the ß-chain
CDR3s of TCR from MM and CR-A Hom-CTL (restricted by HLA-A68) are
almost identical in sequence (Table III). Although the ß-chain CDR3s
of T cell clones AB5 and AB7 are identical in length, no apparent
conservation in this region is evident among B27-restricted Hom-CTL
from different individuals (Table III). Nearly all TCR -chains from
A68- and B8-restricted Hom-CTL have an arginine (R) or lysine (K)
residue at position 93, although their CDR3 lengths are different
(Table IV). In contrast, however, almost random usage of amino acid
residues in the VDJ junction is observed in the TCR -chains of
B27-restricted Hom-CTL (Table IV).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Conservation in TCR usage by Hom-CTL is best demonstrated in
HLA-A68-restricted T cells from patients MM and CR. They selected not
only V gene segments of the same subfamilies (AV4 and BV23) but also
nearly identical CDR3s of their ß-chains. Interestingly, the V gene
segments (AV20S1 and BV5S1) used by one of the B27-restricted T cell
clone, AB1, are closely related to AV4S1 and BV23S1, respectively
(AV20S1 is the closest relative of AV4 among all human TCR V
subfamilies and BV5S1 shares 70% homology with BV23S1 on protein level
(10)), suggesting strong selective pressure for AV4, BV23, and related
gene products in the homocysteine-specific TCR. It should also be
noted, however, that the AV22S1/AJ17S1 -chain accounted for >40%
of the -chain sequences from the MM Hom-CTL line (Table I),
indicating that a large percentage of T cells in this line were
expressing this -chain. Although no Hom-CTL with such an -chain
were successfully cloned, their existence cannot be ruled out. Among
all the Hom-CTL lines generated in our laboratory, MM Hom-CTL was by
far the most efficient, specific, and stable one that could be kept
growing in vitro for several months without losing its specificity
(X. M. Gao, unpublished observation). If T cells expressing the
AV22S1/AJ17S1 -chain were not specific for homocysteine-sensitized
target cells, they might have out-grown the homocysteine-specific
clones and made the MM Hom-CTL line nonspecific in several weeks. It is
interesting that the AV22S1/AJ17S1 -chain also contains an R at
position 93 (Table IV) (see below).
Another interesting observation from this study is that TCR -chains
with an R or K residue at 93 are preferably selected by Hom-CTL
restricted by A68 and B8, but not B27. This effect is possibly driven
by the complementarity of the respective receptors for antigenic
determinant, i.e., homocysteine. It has been shown that the CDR3s are
in contact with the antigenic peptide and the rest of the TCR encoded
by the V and Vß germline sequences interact with MHC (11, 12, 13, 14, 15).
Amino acid 88–103 of moth cytochrome c (MCC) is a murine T
cell epitope presented by I-Ek (20). The MCC peptide
contains a K at position 99, and the 93 of the specific TCR was
always a glutamic acid (E) or aspartic acid (D) (15, 21, 22). When
there was a K-E substitution at position 99 of the MCC peptide,
however, the E or D at 93 of the TCR was replaced by a K (21, 22),
suggesting that a salt-bridge between TCR CDR3 and the MCC peptide
was crucial for the MHC/peptide/TCR interaction. An analogy can be
drawn between this and the T cell recognition of homocysteine/HLA
complexes. When homocysteine forms a disulfide bond with peptides bound
to HLA Ags such as A68 or B8, the negatively charged carboxyl group of
homocysteine would be positioned toward TCR; it could form a
salt-bridge with the positively charged side chain of R or K at
position 93 of TCR, thus providing extra binding affinity for the
HLA/peptide/homocystine/TCR interaction.
In HLA-B27-restricted Hom-CTL, TCR usage appears to be heterogeneous
among different donors, although two B27-restricted T cell clones from
patient AB share the same V gene segments (Table I). In the direct
modification model, B27 is modified by homocysteine through its
unpaired Cys67. If so, homocysteine would be buried in the B-pocket of
B27, which is responsible for accommodating the R residue at the second
position of the bound peptides (23), and become inaccessible from the
molecular surface. The directly modified B27 would thus present a new
set of peptides (certainly no R at position 2), rather than
homocysteine, to Hom-CTL. The fact that the TCR usage by
HLA-B27-restricted Hom-CTL is relatively heterogeneous provides
supporting evidence for this view. However, concrete evidence for the
direct modification model can only come from experiments in which
direct labeling of HLA-B27 Ag with homocysteine is demonstrated.
In conclusion, the data presented in this report reveal certain
important features of Hom-CTL TCR and also provide valuable information
on how homocysteine might interact with different HLA molecules in the
cell. It should also be emphasized, however, that the number of
subjects included in this study is relatively small (five patients and
one healthy individual). It is possible that future studies involving
more subjects will reveal a pattern of modest conservation of TCR usage
by B27-restricted Hom-CTL. Furthermore, if the HLA-B27 molecule can
also be modified by homocysteine via the indirect pathway, it should be
possible to find B27-restricted Hom-CTL expressing TCR with certain
features similar to that of A68- or B8-restricted T cells (e.g., with
an R or K at position 93). Finally, it would be of importance to
determine whether an R or K at position 93 is an absolute
requirement for TCR of Hom-CTL not restricted by HLA-B27. Only analysis
of a much larger panel of Hom-CTL lines and clones from more
individuals will answer these questions in full.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Present address: Prenatal Resources, Swedish Medical Centre, Denver, CO 80110.
3 Address correspondence and reprint requests to Dr. X. M. Gao, Department of Biochemistry, Imperial College of Science, Technology and Medicine, London SW7 2AY, U.K. E-mail address:
4 Abbreviations used in this paper: AS, ankylosing spondylitis; ReA, reactive arthritis; Hom-CTL, homocysteine-specific CTL; CDR3, third complementarity determining region; AnPCR, anchor PCR; rhIL-2, recombinant human IL-2; BCL, B cell line; MCC, moth cytochrome c.
Received for publication September 3, 1997. Accepted for publication December 17, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-chain repertoire and demonstration of genetic influence on V usage. Eur. J. Immunol. 23:1153.[Medline]
ß T cell receptor structure at 25 Å and its orientation in the TCR-MHC complex. Science 274:209. domain of a T cell antigen receptor. Science 270:1812.
ß T-cell receptors "see" peptide-MHC complexes. Immunol. Today 14:597.[Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
