Rheum Dis Clin N Am 29 (2003) 595 – 611

Identification of HLA-B27–restricted peptides

in reactive arthritis and other
spondyloarthropathies: computer algorithms and
fluorescent activated cell sorting analysis as
tools for hunting of HLA-B27–restricted
chlamydial and autologous crossreactive
peptides involved in reactive arthritis and
ankylosing spondylitis
Wolfgang Kuon, PhD*, Joachim Sieper, MD
Section of Rheumatology, FU-Klinikum Benjamin Franklin, Berlin Hindenburgdamm 30,
12200 Berlin, Germany

Class I major histocompatibility class (MHC) protein molecules are highly

polymorphic surface glycoproteins that play a major role in protective immunity
against viruses and intracellular bacteria [1,2]. The strongest of any HLA class I
antigens associated with human disease is the HLA-B27 molecule. Seronegative
spondyloarthropathies (SpA), a group of related diseases such as ankylosing
spondylitis (AS), reactive arthritis (ReA), arthritis/spondylitis associated with
inflammatory bowel disease (IBD), some forms of psoriatic arthritis, and
undifferentiated spondyloarthropathies are clearly associated with HLA-B27.
The main function of HLA class I molecules is to present peptide antigens to
CD8+ cytotoxic T cells (CTL). Thus, it has been proposed that the antigen-
presenting properties of HLA-B27 might be crucial in the pathogenesis of SpA.
In addition, it has been suggested that interplay between certain bacteria and
HLA-B27 plays a crucial role in the pathogenesis of the SpA. Therefore, SpA-
related research has developed model character in understanding the interaction
between genetic and environmental factors in manifestation of rheumatic dis-
eases, but the pathogenic mechanisms in SpA remains unclear. In this article the

* Corresponding author.
E-mail address: kuon@medizin.fu-berlin.de (W. Kuon).

0889-857X/03/$ – see front matter D 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0889-857X(03)00050-4
596 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611

authors review recent immunologic work done in human SpA with particular
focus on a strategy to identify HLA-B27 – restricted bacterial peptides in
Chlamydia trachomatis – induced ReA.

Reactive arthritis and arthritogenic bacteria

Reactive arthritis can be regarded as a subgroup among the infection-
associated arthritides [3]. Because of the HLA-B27 association, its pattern of
joint involvement (asymmetrical, predominantly of the lower limbs), and the
possible affection of the spine, ReA is part of the spondyloarthropathies (SpA).
ReA occurs at a frequency of 2% to 4% after preceding infections of the
urogenital tract with C trachomatis, and at a frequency of 0% to 15% after
infections of the gut with enterobacteriaceae such as Yersinia, Salmonella,
Shigella, or Campyloabacter [4]. The variation is likely caused by genetic
differences of the bacteria and their strategies to infect and survive inside the
host, which might influence their arthritogenicity.
Yersinia enterocolitica, for instance, is taken up by M cells in Peyer’s
patches through an interaction between invasin, a bacterial protein, and the
host b1-intergrin. Because the microbe can use phagocytes as a carrier, Yersinia
can surmount endothelial monolayers and reach synovial tissues by way of the
bloodstream [5]. In a rat model it has been observed that live Yersinia or Yersinia-
specific DNA fragments could not be detected in the synovial tissue. Even after
several months, the lymph nodes were positive for Yersinia, which was accom-
panied by a strong and long lasting antibody response. Lymph nodes therefore
seem to be a reservoir for Yersinia [6]. In patients who have Yersinia-induced ReA,
mononuclear phagocytes are suggested to enter the peripheral bloodstream
carrying arthritogenic bacteria and to present bacterial antigenic epitope in the
synovium. Bacterial lipopolysaccharide (LPS) [7]—but no Yersinia-specific DNA
[8]—is present in synovial fluid. Furthermore, T-cell responses to HLA-B27–
restricted peptide epitopes from the 60kDa heat shock protein and the Yersinia
urease b-subunit [9] could be detected in the joint. Taken together, these results
suggest that it is likely that Yersinia can persist outside the joint in lymph nodes or
the mucosa and that bacterial antigen reaches the joint by way of monocytic cells.
In Salmonella-induced ReA infection, persistence of bacterial LPS was also
demonstrated in the joint [10], but it is difficult to detect Salmonella DNA in the
joint [11]. Lymph nodes or intestinal mucosa have been suggested as places
where Salmonella persist.
In comparison with the described enteric bacteria, C trachomatis is an obligate
intracellular pathogen that primarily resides in the epithelial cells but can infect
other cell types as well (eg, macrophages and nonprofessional phagocytic cells)
[12]. There is good evidence that Chlamydia can persist for a long time in the
body [13]. It is a common pathogen, and traces of chlamydial RNA and DNA are
detectable by polymerase chain reaction (PCR) in the joint [14]. Despite many
attempts, however, no chlamydial organisms could be cultured from the inflamed
 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611 597

joints of patients who have SpA [15]. It has been further supposed that an
imbalance of cytokines such as tumor necrosis factor (TNF)-a and interleukin
(IL)-10 support the persistence of Chlamydia and other bacteria in vivo [16].
Chlamydia has a complicated live cycle determined by an infectious (elementary
bodies) stage outside of cells and a noninfectious (reticular bodies) stage inside
cells. Chlamydia replicate inside host cells, and their antigens seem to be
processed by the class I and class II pathways to be presented to CD4+ and
CD8+ T cells of the host cells [17,18].
Taken together, these data suggest that persistence of bacteria or bacterial
antigens in patients who have ReA is likely.

HLA-B27, allotypes, and the arthritogenic peptide hypothesis

The association of HLA-B27 with AS, ReA, and other SpA is the strongest of
any HLA antigen to human disease. Because the main function of HLA class I
molecules is to present peptide antigens to cytotoxic T cells, it has been proposed
that the antigen-presenting properties of HLA-B27 could be crucial in the
pathogenesis of SpA (Fig. 1). Beside alternative mechanisms, which are dis-
cussed in this article, the favored theory for the association between HLA-B27
and SpA is the arthritogenic peptide hypothesis [19], which suggests that because
of their unique amino acid residues, some B27 subtypes bind specific arthrito-
genic peptides that are recognized by CD8+ T cells (Fig. 2). Furthermore, in

Fig. 1. Interaction of MHC class I molecule HLA-B27 with a nonamer peptide. The structure of a
theoretical nonamer peptide is bound in the pocket of the MHC groove between the a-helical
structures and in contact with the b-sheets on the bottom. The N-terminus (N) of the peptide is directed
to the left side and the carboxy terminus (C) to the right side. Two amino acid substitutions in
positions 114 and 116 on the bottom of the groove of the HLA-B27 molecule are marked. These two
amino acid substitutions differentiate the HLA-B*2705 disease-associated subtype from the
nonassociated subtype HLA-B*2709 by only one amino acid substitution (in position 116), whereas
HLA-B*2704 differs from the non – disease-associated subtype HLA-B*2706 by only two amino acid
substitutions, in positions 114 and 116.
598 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611

Fig. 2. Presentation of an arthritogenic peptide derived from proteins of C trachomatis after infection
of antigen presenting cells. The bacterial peptides are presented by the class I major histocompatibility
complex HLA-B27 on APCs to the TCR on CD8+ T lymphocytes.

response to these bacterial peptides, autoreactive T-cell –recognizing antigens

with sufficient structural similarity between bacteria and self might become
activated by self-peptides such as those in the joints and spine. This process
might be supported by integrins, which are relevant for the homing of cross-
reactive CD8+ T cells into the joint [20]. This hypothesis is supported by several
direct and indirect observations. One major support for this hypothesis came from
studies in humans showing the differential association of natural HLA-B27
allotypes to AS [21,22]. While B*2705, B*2702, B*2704, and B*2707 are
strongly associated with the disease [23], the HLA-B27 subtypes B*2709 in
Caucasians and B*2706 in Asians are not at all or only rarely present in AS
patients [24 –28]. These two subtypes differ from the disease-associated ones
only by one amino acid substitution (B*2705 to B*2709) by the exchange of an
Asp116 to His116, or two amino acid substitutions (B*2704 to B*2706) by
exchange of His114 to Asp114 and of Asp116 to Tyr116 [29], all of which are
located in the peptide-binding groove (Fig. 1). It has therefore been hypothesized
that the two disease-nonassociated B*2706 and B*2709 subtypes do not present
arthritogenic peptides in contrast with the disease-associated subtypes. The
possibly crucial role of arthritogenic peptides for the pathogenesis is supported
by experimental observations in animal models and in patients. HLA-B27
transgenic rodents develop spondyloarthropathy-like manifestations only if raised
in a nonsterile environment [30 – 32]. Zhou and colleagues have investigated the
influence of the B27-bound peptide repertoire on the development of arthritis in
transgenic rats [33]. B27 transgenic rats expressing a strong HLA-B27 –restrict-
ed, influenza-derived peptide with high affinity to HLA-B27 showed significant-
ly lower disease symptoms than rats expressing an unrelated peptide. Thus, these
results suggest that unknown arthritogenic bacterial peptides and the influenza
peptide are competing for B27 binding. At the patient level, bacteria-specific and
autoreactive CD8+ CTL have also been demonstrated in patients who have AS
[5,34,35] and ReA [36,37]. The important role for an interaction between bacteria
and B27 is supported by the finding that 20% of HLA-B27+ reactive arthritis
 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611 599

patients progress to ankylosing spondylitis after 10 to 20 years [38]. Moreover,

50% to 70% of HLA-B27+ patients who have Crohn’s disease or ulcerative
colitis [39] develop a full clinical picture of AS eventually, which suggests that
SpA-associated bacteria could cause AS, possibly through a crossreaction
between bacterial and self-antigen [40].

Molecular mimicry, T-cell oligoclonality, and spondyloarthropathies

The concept that microbial peptides with sufficient structural similarity to self-
peptides can activate potentially autoreactive T cells by crossreactivity in the pe-
riphery is thought to be a possible mechanism of triggering autoimmunity [41 – 43].
Although recognition of antigenic peptides by the T-cell receptor (TCR) is a
specific process, it has been suggested that structurally altered peptide variants or
mimicry peptides can be recognized by CD4+ or CD8+ T cells even if the linear
sequence is quite distinct from the primary sequence. In other words, it is possi-
ble that mimicry peptides require little or no homology with the original motif
[44 – 47]. Crossreactive peptides were defined by search of homologies in the data-
base or by computer algorithms considering structural requirements of the T cell
receptor to recognize MHC/peptide complexes [48]. The functional properties
of these peptides depend on their structure and their interaction with the
TCR and might lead to full or partial activation of the TCR, or even to anergy
by antagonism.
Many examples, particularly from viral experiments and from myelin basic
protein (an assumed autoantigen in multiple sclerosis) support the concept of
molecular mimicry by crossreactive peptides and their meaning for T-cell selection,
viral escape, or autoimmunity. A recent report from Bachmaier and colleagues
came from a Balb/c murine myocarditis model in which chlamydial peptides were
suggested to induce autoimmunity [49]. The crossreactivity was caused by a
sequence similarity between a peptide from the cardiac myosin heavy chain and
peptides from the 60-kDa outer membrane protein of Chlamydia pneumoniae.
Furthermore, Steinhoff and colleagues recently reported the relevance of T cell
receptor crossreactivity for CD8+-T cells in a mouse model of IBD in which
pathogenic CD8+ T-cell clones were specific for murine and mycobacterial hsp60
[50,51]. These and other examples demonstrate that molecular mimicry might be
an important mechanism for CD4+ and for CD8+ T cells leading to the develop-
ment of autoimmune diseases. The underlying mechanism is likely the structural
features of the trimolecular complex of MHC/TCR/ peptide that determines and
limits the T-cell receptor specificity for the MHC-bound peptides, which can result
in an oligoclonal expansion of CD8+ T cells specific for an immunodominant
peptide epitope.
Oligoclonal expansion of T cells has recently been demonstrated for CD4+
and CD8+ T cells in AS [52] and for CD8+ T cells in ReA [35 – 37,53]. The
synovial fluid derived from different HLA-B27+ patients suffering from ReA
triggered by different bacteria revealed an astonishingly high homology of TCRs.
600 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611

These results suggest that similar antigens are recognized by these oligoclonally
expanded CD8+ T cells, which implies the possibility that under certain con-
ditions, specific arthritogenic peptides might be produced and will be presented to
the hosts immune system.

Identification of HLA-B27– restricted arthritogenic peptides in

spondyloarthropathies

Search for bacterial peptides

In the previous sections, the importance of HLA-B27 subtypes, the possible
crossreactivity between bacterial and self-derived peptides by molecular mimicry,
and the importance of oligoclonality of the TCR in SpA were described. It was a
dogma for a long time that bacteria-derived peptides should not be presented to
MHC class I molecules because they could not gain access to the class I pathway
(with some exceptions such as Listeria). It has become clear that even peptides
from proteins that originate by way of the extracellular environment of a cell—
including the exogenous pathway by phagosomes and by way of the classically
endogenous pathway—can induce a CD8+ T-cell response [54 – 56] against
obligate intracellular bacteria such as C trachomatis with its extra- and intracel-
lular stages [57,58] or against enterobacteriaceae such as Yersinia [59] or
Salmonella [60]. Furthermore, processed bacterial peptides have been shown to
be presented by class I molecules [61], revealing that Salmonella infection can
also induce a bacteria-specific CD8+ T-cell response. It has further been reported
that Salmonella invasion can change the peptide repertoire of HLA-B27 [60].
While clinical and experimental observations suggest that autoreactive CD8+
T cells might be involved in AS and ReA, the question regarding how to search
effectively for unknown crossreactive bacterial epitopes arises. It is time
consuming to produce bacteria-specific T-cell clones or lines against those
epitopes, and because of the limitation of the TCR repertoire for peptide epitope
recognition, one might miss many of them if one uses only the procedure of T-cell
cloning. For C trachomatis, a more effective approach was described recently by
researchers screening for unknown epitopes that were recognized by CD8+ T cells.
In this study, Chlamydia-specific CD8+ T-cell lines were cultured from infected
mice then used to screen expression libraries of C trachomatis genes. An antigen
encoded within an open reading frame of the Chlamydia genome designated Cap 1
(class I accessible protein 1) could be identified, but a homology to any known
functional protein could not be determined [62].
The authors studied CD8+ HLA-B27 – restricted T-cell responses to
C trachomatis by using a completely different approach [63]. The genome of
C trachomatis recently sequenced by Stephens and colleagues [64] consists of
1,042,519 base pairs that code for the expression of 894 proteins and a 7493 base
pair plasmid (http://chlamydia-www.berkeley.edu:4231) whose open reading
frame codes for 10 proteins. The authors’ goal was to screen the entire proteome,
 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611 601

all 904 proteins from Chlamydia, for all possible peptide candidates restricted to
the class I HLA-B*2705 molecule. This is possible because for different HLA
molecules, the so-called peptide binding motifs have been defined [65]. Each
nonamer peptide that binds within the peptide-binding groove of an MHC
molecule is characterized by one or more so-called anchor residues. In the case
of HLA-B27, an arginine at position 2 of the nonamer peptide is required to bind
in the B27 MHC peptide-binding groove. The authors therefore screened the
entire proteome from C trachomatis for all likely HLA-B*2705-binding nonamer
peptide candidates with an arginine at position 2.
For further narrowing of the number of B27-restricted peptides, a combination
of two computer-based algorithms was used: the first algorithm was designed for
predicting the binding probability of nonamer peptides with an arginine at position
2 to bind HLA-B*2705. The algorithm was developed by the group of H-G
Rammensee [66] and is available on the Internet (http://www.uni-tuebingen.de/
uni/kxi). The second algorithm exploits the fact that peptides presented by MHC
class I molecules to CD8+ T cells are processed in the cytoplasm by proteasomes.

Fig. 3. (A) Peptide binding to HLA-B27, and (B) proteasomal cleavage prediction. Four nonamer
peptides (peptides 1 – 4) at different locations in the chlamydial protein papQ were selected by the
SYFPEITHI program. (From Rammensee HG, Bachmann J, Emmerich NPN, et al. SYFPEITHI:
database for MHC ligands and peptide motifs. Immunogenetics 1999;50:213; with permission.) When
the nonamer sequences were enlarged (necessary for this prediction) by 10 of their natural amino acids
at both sides at the N and C terminus of the peptide fragments, only one peptide of four (peptide 2;
indicated by shading) was predicted in this example to be cut by the proteasome.
602 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611

The algorithm predicts the cleavage sites within a peptide fragment seen by the 20S
proteasome complex (Fig. 3) [67,68]. Finally, to identify HLA-B27 –restricted
chlamydial peptides stimulatory for CD8+ T cells, T-cell responses to selected
peptides were examined by antigen-specific flow cytometry (fluorescent activated
cell sorting [FACScan]) (Fig. 4). Thus, because of the immunogenicity of the
selected chlamydial peptides, the stimulation of CD8+ T cells was measured by
determining the percentage of cells that were double-positive for the expression of
the early activation marker CD69 on the cell surface and by determining the
intracellular release of IFNg within the cells. Both proteins were stained with
different fluorescent-labeled monoclonal antibodies with specificity for the differ-
ent proteins. The percentage protein expression on the cell surface and within the
cells could be measured easily by FACScan.
The above two bioalgorithms allowed the authors to reduce the number of
HLA-B27 –restricted peptides from about 18,000 nonamers with arginine at
position 2. To exclude peptides with low binding probability, the authors set a
binding score of greater than or equal to 25 for the first program, which reduced
the number of peptide candidates to 1881. Although the binding program allowed
identification of HLA-B27 – binding peptides with high fidelity, it did not provide
an answer regarding whether or not these selected peptides are immune relevant
and will be generated by the cellular 20S proteasome. The authors then tested the
1881 theoretical peptides obtained in the first step by the proteasome cleavage
site prediction program based on the enzymatic properties of 20S proteasomes.
The authors finally came up with 199 peptides fulfilling the criteria of both
programs. Thus, combining both computer algorithms. The authors were able to
electronically screen the entire proteome of the bacterium C trachomatis without
actually eluting pathogen-derived peptides from the MHC molecules.
In the next step, the authors had to examine the selected peptides for
immunogenicity. Antigen-specific flow cytometry based on the staining of cell

Fig. 4. FACScan analysis of antigen-specific CD8+ T cells. Synovial T cells were incubated for 6 hours
with identified bacterial peptides. Stimulation was counted by quantification of CD8+ synovial fluid-
derived T cells that were double-positive (upper right quadrant) for the early activation marker CD69
(expressed on the cell surface and stained with a specific fluorescence stained monoclonal antibody)
and IFNg (secreted intracellularly and stained by a second monoclonal antibody).
 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611 603

surface expression markers and intracellular cytokine secretion [69,70] was used.
Peripheral blood mononuclear cells or synovial fluid-derived cells from patients
who have C trachomatis – induced ReA were stimulated for 6 hours with the
selected peptides. An immune response was analyzed by quantifying CD8+
T cells that were double-positive for the early activation marker CD69 and IFNg
by intracellular staining (Fig. 4). Eleven HLA-B27– restricted individual non-
amer peptides could be identified in B*2705+ patients who had Chlamydia-
induced ReA. In parallel the authors used a similar approach in HLA-B*2705
transgenic mice infected with C trachomatis. In this mouse model the authors
could identify eight peptides that were recognized by CTL, inducing cell lysis
when loaded on transporter associated with antigen processing (TAP)-defective
RMA-S-B27 [71] targets. The authors also started to construct tetramer/peptide
complexes with specificity for the selected peptides. These tetramers are useful to
track down crossreactive self-peptides. This entire screening strategy for B27-
restricted chlamydial peptides and the results are summarized in Fig. 5.

Fig. 5. Screening strategy for HLA-B*2705-binding peptides. The entire proteome of C trachomatis
was screened by two computer algorithms (first, the binding prediction SYFPEITHI program; second, a
proteasomal cutting prediction program) for B27 nonamer peptides. One thousand eight hundred
eighty-one peptides from about 18,000 potential candidate peptides were left after the first screening
step. One hundred ninety-nine peptides finally fulfilled the criteria of both algorithms after the second
screening step. These peptides were synthesized and investigated for their capacity to stimulate
synovial-derived CD8+ T cells from HLA-B27+ arthritis patients who were suffering from Chlamydia-
induced reactive arthritis. In other functional assays such as cytotoxic assays, the authors found that
most of the peptides were recognized by CTL. (From Kuon W, Holzhütter HG, Appel A, et al.
Identification of HLA-B27 – restricted peptides from the C trachomatis proteome with possible rele-
vance to HLA-B27 – associated diseases. J Immunol 2001;167:4738; with permission.)
604 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611

With the combination of the two biomathematical algorithms, one for the
peptide binding prediction and the other for examination of the 20S proteasome
cleavage requirements for peptides followed by antigen-specific flow cytometry,
powerful tools have become available to screen large numbers of bacterial or other
protein sequences for candidate peptides that are assumed to play a role in
rheumatic and other immune-mediated diseases, and to investigate their functional
relevance. With the above approach, 11 chlamydial peptides could be identified.

Search for autologous peptides

To search for B27-restricted autoantigens, two approaches can be applied.
First, the identified CD8 epitopes from C trachomatis (a nonamer peptide motif)
can be defined by systematic substitutions of amino acids that are essential for
T-cell recognition of the B27/peptide complex [45,72]. When such a motif is
detected it can be used for a databank search for autoantigens fulfilling this
motif. This work is currently in progress in the authors’ laboratory. Secondly, the
number of possible autoantigens can be limited by focusing on target antigen.
Recent work from immunohistology and MRI strongly suggest that the cartilage
macromolecules of joints and spine are the primary targets of the immune
response; therefore, T-cell responses to cartilage-derived antigens have been
investigated more thoroughly recently. It has been shown in patients that from
nineteen B27-binding peptides derived from different human procollagens
(CI, CII, and CXI), one peptide was identified to stimulate HLA-B27 –restricted
CTL; however, stimulation was only observed in one ReA patient out of three
ReA patients and 27 AS patients [73]. Another candidate protein to deliver
arthritogenic T-cell epitopes could be the G1-domain protein of the cartilage
proteoglycan aggrecan, which has been shown to induce arthritis and spondylitis
in Balb/c mice and is stimulatory to T cells in AS patients. It was therefore
regarded as an animal model for AS [74 – 76]. Furthermore, the proteoglycan-
derived cartilage link protein induced arthritis in Balb/c mice [77].
Using the algorithms described above we are currently screening 18 cartilage
proteins addressing the question whether HLA-B27 – restricted CD8+ T-cell
epitopes can be identified.
Another approach that the authors are pursuing is the construction of tetramer/
peptide complexes of the authors’ selected cartilage- or chlamydia-derived
peptides to search and enumerate for the presence of antigen-specific T cells in
the inflamed tissue [78]. The authors and others have shown that tetramers are
useful tools that can bind to CD8+ T cells specific for chlamydial peptide [63,79].
Screening of tissues with MHC class I/peptide complexes for antigen-specific
T cells is also relevant in the light of recent finding by Kuckelkorn et al [80], who
proposed that T-cell epitopes such as mycobacterial HSP-60 are produced in a
tissue-specific way after bacterial infection in the intestine. Kuckelkorn et al
suggest that this organ-specific potential to produce T-cell epitopes in a qualita-
tively and quantitatively different way is caused by the local 20S proteasomes of
the infected tissue. Their data demonstrate that 20S proteasomes are sufficient to
 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611 605

generate CD8+ T-cell epitopes of a certain protein and that 20S proteasomes from
different organs produce a distinct pattern of T-cell epitopes because of their
organ–specific subunit composition. The idea could explain elegantly the finding
of a local increase of certain self-peptides [81] and tissue- and organ-specific
immune reactions [50,51]. Thus, the generation of self-epitopes in selected tissues
that are normally tolerated or ignored by peripheral T lymphocytes might be
increased locally by, for example, microbial infections or trauma and activate
CD8+ T cells, causing tissue-specific autoimmune pathology. The authors propose
that in addition to peripheral immune regulatory mechanisms, organ-specific
epitope generation represents an important mechanism to control the reactivity of
autoreactive CD8+ T cells that escape thymic deletion.

Other hypotheses about the role of HLA-B27 in spondyloarthropathies

Despite the long-held knowledge about the association of HLA-B27 and SpA
[82,83] and intensive research, the pathogenic role of HLA-B27 remains
unclear. Other hypotheses have emerged beside the ‘‘classical’’ arthritogenic
peptide theory.
One interesting concept, the HLA-B27 misfolding hypothesis, states that
HLA-B27 itself is directly involved in the pathological process of SpA. It has
been suggested that HLA-B27 can be misfolded, which might have consequences
for the pathogenic process [84 –86]. It has been suggested that the misfolding is
caused by a particular feature of HLA-B27. Newly synthesized HLA-B*2705
seems to fold by and associate with b2-microglobuline more slowly in compar-
ison with other MHC class I molecules. Deglycosylated heavy chains (HC) were
found in the cytosol and are supposed to be dislocated from the endoplasmatic
reticulum (ER) as a consequence of misfolding [85].
Allen and coworkers reported that as a consequence of HLA-B27 misfolding,
free HLA-B27 heavy chains can form abnormal HC homodimers [87]. This
homodimer formation could be facilitated by unpaired free cystein residues at
position 67 (Cys 67) of the HLA-B27 heavy chain a1 helix. Homodimerization
was also suggested in HLA-B27 transgenic mice, but in mice there was no
evidence that HLA-B27 heavy chains could form a class II-like structure and
present peptides to CD4+ T cells [88]. In contrast, in humans, Boyle et al reported
that human CD4+ T cells would be stimulated by HLA-B27 [89].
Kollnberger and colleagues recently found cell-surface expression of HLA-
B27 homodimers in patients who have AS and formulated a different disease
hypothesis. Beside the classical view of epitope recognition by the TCR,
Kollnberger et al also investigated whether or not other molecules might be
involved in the pathomechanism of arthritis [90,91]. They reported that the
B27 HC homodimers were recognized by receptors belonging to a family called
the killer cell immunoglobulin-like receptor (KIR) family. Another kind of
receptor that recognized the B27 homodimers belonged to the immunoglobu-
lin-like transcripts (ILT) or leukocyte Ig-like receptors family. While KIR are
606 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611

expressed on certain groups of natural killer (NK) cells, T cells, and NKT cells
[92], different ILT receptors can be found on B cells, NK cells, T cells, cells of the
monocytic macrophage line, and on dendritic cells. Future experiments should
investigate whether or not patients who have AS express free HLA-B27 HC that
bind an unusual distinct and overlapping repertoire of the KIR and ILT immune
receptor families, which possibly promote the disease process
Malik and colleagues also reported the finding that HLA-B27 b2-micro-
globulin free HC forms can exist on the cell surface as homodimers and are
recognized by a specific antibody [93,94]. With the newly employed and highly
efficient perfusion affinity chromatography [95] they isolated a nonamer peptide
from the HLA-B27 HC homodimers expressed on an Epstein Barr virus (EBV)-
transformed human lymphoblastoid cell line; however, experiments by the
authors suggest that occurrence of HLA-B27 HC homodimers on, for example,
transgenic mouse splenocytes is only low. Even in the presence of the identified
peptide and antibody staining, the B27 HC formation could not be detected by
FACScan. Also, the in vitro assembly of the HLA-B2705 HC with the identified
peptide failed. Thus, the exact role of this b2-microglobulin free peptide-
containing HLA-B27 HC in SpA has to be determined.
Finally, the isolation of a new HLA-B27 self-derived dodecamer peptide from
the intracytoplasmic tail of its own molecule has been reported [96]. It was shown
that the peptide was a natural ligand of three disease-associated subtypes: HLA-
B*2702, HLA-B*2704, and HLA-B*2705, but not of the two nonassociated
subtypes, B*2706 and B*2709. This peptide showed a strikingly homology to a
peptide sequence derived from the DNA primase from the arthritogenic bacte-
rium C trachomatis. These results demonstrate that an HLA-B27 self-derived
peptide can mimic a sequence from an arthritogenic bacterium. Whether or not
this newly detected peptide that acts as a natural ligand of disease-associated B27
subtypes is crucial for the pathogenesis of SpA should be the subject for future
investigations. Finally, the breakdown of CTL tolerance to self HLA-B27 and the
influence of the interrelationship between the pathogen C trachomatis has been
discussed in a recent publication [97]. The investigators report evidence that
previously quiescent autoreactive T cells are activated by exposure to Chlamydia,
recognizing a B27 self-derived peptide even if there is no sequence homology
with the bacterial peptide. This finding is discussed in this article [43 –47].

Summary
The illustrated clinical and experimental results demonstrate the strong
relationship between the MHC class I antigen HLA-B27 and synovial CD8+
T cells with specificity for bacterial and possible self-antigen in SpA. These new
aspects obtained in recent experimental and clinical studies might also provide
clues to the pathomechanisms of joint inflammation in SpA. In particular, the
newly developed techniques will be of great relevance in the near future. New and
more precise bioalgorithms reflecting new insights in the biology and biochem-
 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611 607

istry of proteins as recently presented [98,99] can be helpful (eg, a program with
an improved prediction of the features of immunoproteasomes). Intracellular and
secreted cytokine staining by FACScan allows examination of a great number of
cells expressing certain antigens in response to certain stimuli. The analysis of
T-cell responses with tetramer/peptide complexes can be useful to screen tissue
sections for TCR, recognizing foreign or self-derived epitopes on those complexes
loaded with selected (eg, bacterial) peptides. Identification of arthritogenic
peptides and a further understanding of the immunology of the pathomechanisms
in SpA might open ways to design new peptide vaccines to prevent inflammation,
autoimmunity, and other diseases by early intervention [100].

References
[1] Kaufmann SH. Immunity to intracellular bacteria. Annu Rev Immunol 1993;11:129.
[2] Harty JT, Bevan MJ. Responses of CD8+ T cells to intracellular bacteria. Curr Opin Immunol
1999;11:89.
[3] Granfors K, Märker-Hermann E, De Keyser E, et al. The cutting edge of spondylarthropathy
research in the millennium. Arthritis Rheum 2002;46:606.
[4] Sieper J, Braun J, Kingsley GH. Report on the Fourth International Workshop on Reactive
Arthritis. Arthritis Rheum 2000;43:720.
[5] Märker-Hermann E, Höhler T. Pathogenesis of human leukocyte antigen B27-positive arthritis.
Information from clinical materials. Rheum Dis Clin N Am 1998;24:865.
[6] Zhang Y, Gripenberg-Lerche C, Soderström KO, et al. Antibiotic prophylaxis and treatment of
reactive arthritis: lessons from an animal model. Arthritis Rheum 1996;39:1238.
[7] Granfors K, Jalkanen S, von Essen R, et al. Yersinia antigens in synovial fluid cells from
patients with reactive arthritis. N Engl J Med 1989;320:216.
[8] Nikkari S, Merilaht-Palo R, Saario R, et al. Yersinia-triggered reactive arthritis. Use of polymer-
ase chain reaction and immunocytochemical staining in the detection of bacterial components
from synovial specimens. Arthritis Rheum 1992;35:682.
[9] Mertz AK, Daser A, Skurnik M, et al. The evolutionary conserved ribosomal protein L23 and
the cationic urease beta-subunit of Yersinia enterocolitica O:3 belong to the immunodominant
antigens in Yersinia-triggered reactive arthritis: implications for autoimmunity. Mol Med 1997;
1:44.
[10] Granfors K, Jalkanen S, Lindberg AA, et al. Salmonella lipopolysaccharide in synovial cells
from patients with reactive arthritis. Lancet 1990;335:685.
[11] Nikkari S, Rantakokko K, Ekman P, et al. Salmonella-triggered reactive arthritis: use of polymer-
ase chain reaction, immunocytochemical staining, and gas chromatography-mass spectrometry in
the detection of bacterial components from synovial fluid. Arthritis Rheum 1999;42:84.
[12] Inman RD, Whittum-Hudson JA, Schumacher HR, et al. Chlamydia and associated arthritis.
Curr Opin Rheumatol 2000;12:254.
[13] Bas S, Griffais R, Kvien TK, et al. Amplification of plasmid and chromosome Chlamydia DNA
in synovial fluid of patients with reactive arthritis and undifferentiated seronegative oligo-
arthropathies. Arthritis Rheum 1995;38:1005.
[14] Wilkinson NZ, Kingsley GH, Sieper J, et al. Lack of correlation between the detection of
Chlamydia trachomatis DNA in synovial fluid from patients with a range of rheumatic diseases
and the presence of an anti chlamydial immune response. Arthritis Rheum 1998;41:845.
[15] Bas S, Scieux C, Viescher TL. Different humoral immune response to Chlamydia trachomatis
major outer membrane protein variable domains I and IV in Chlamydia-infected patients with or
without reactive arthritis. Arthritis Rheum 1999;42:942.
[16] Braun J, Yin Z, Spiller I, et al. Low secretion of tumor necrosis factor alpha, but not other Th1
608 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611

or Th2 cytokines, by peripheral blood mononuclear cells correlates with chronicity in reactive
arthritis. Arthritis Rheum 1999;42:2039.
[17] Thiel A, Wu P, Lauster R, et al. Analysis of the antigen specific T cell response in reactive
arthritis by flow cytometry. Arthritis Rheum 2000;43:2834.
[18] Kuon W, Lauster R, Böttcher U, et al. Recognition of chlamydial antigen by HLA-B27-restrictd
cytotoxic T cells in HLA-B*2705 transgenic CBA (H-2k) mice. Arthritis Rheum 1997;40:945.
[19] Benjamin R, Parham P. Guilt by association: HLA-B27 und ankylosing spondylitis. Immunol
Today 1990;11:137.
[20] Lazarovits AI, Karsh J. Differential expression in rheumatoid synovium and synovial fluid of
a4b7 intergrin. A novel receptor for fibrorectin and vascular cell adhesion molecule-1. J Immunol
1993;151:6482.
[21] Khan MA. HLA-B27 polymorphism and association with disease [editorial]. J Rheumatol
2000;27:1110.
[22] Khan MA. Update on spondyloarthropathies. Annals Intern Med 2002;136:896.
[23] Gonzalez-Roces S, Alvarez MV, Gonzalez S, et al. HLA-B27 polymorphism and worldwide
susceptibility to ankylosing spondylitis. Tissue Antigens 1997;49:116.
[24] D’Amato M, Fiorillo MT, Carcassi C, et al. Relevance of residue 116 of HLA-B27 in determin-
ing susceptibility to ankylosing spondylitis. Eur J Immunol 1995;25:3199.
[25] Lopez-Larrea C, Sujirachato K, Mehra N, et al. HLA-B27 subtypes in Asian patients with
ankylosing spondylitis. Evidence for new associations. Tissue Antigens 1995;45:169.
[26] Lopez-Larrea C, Gonzales S, Martinez-Borra J. The role of HLA-B27 polymorphism and
molecular mimicry in spondyloarthropathy. Mol Med Today 1998;4:540.
[27] Ren EC, Koh WH, Sim D, et al. Possible protective role for HLA-B*2706 for ankylosing
spondylitis. Tissue Antigens 1997;49:67.
[28] Fiorillo MT, Meadows L, D’Amato M, et al. Susceptibility to ankylosing spondylitis correlates
with the C-terminal residue of peptides presented by various HLA-B27 subtypes. Eur J Immunol
1997;27:368.
[29] Fiorillo MT, Greco G, Maragno M, et al. The naturally occurring polymorphism Asp116 –
His116, differentiating the ankylosing spondylitis-associated HLA-B*2705 from the non-
associated HLA-B*2709 subtype, influences peptide-specific CD8 T cell recognition. Eur J
Immunol 1998;28:2508.
[30] Taurog JD, Richardson JA, Croft JT, et al. The germfree state prevents development of gut and
joint inflammatory disease in HLA-B27 transgenic rats. J Exp Med 1994;180:2359.
[31] Taurog JD, Maika SD, Satumtira N, et al. Inflammatory disease in HLA-B27 in transgenic rats.
Immunol Rev 1999;169:209.
[32] Khare SD, Luthra HS, David CS. HLA-B27 and other predisposing factors in spondyloarthro-
pathies. Curr Opin Rheumatol 1998;10:282.
[33] Zhou M, Sayad A, Simmons WA, et al. The specificity of peptides bound to human histo-
compatibility leukocyte antigen (HLA)-B27 influences the prevalence of arthritis in HLA-B27
transgenic rats. J Exp Med 1998;188:877.
[34] Hermann E, Yu DTY, Meyer zum Büschenfelde KH, et al. HLA-B27-restricted CD8 T cells
derived from synovial fluids of patients with reactive arthritis and ankylosing spondylitis.
Lancet 1993; 342:646.
[35] Fiorillo MT, Maragno M, Butler R, et al. CD8+ T-cell autoreactivity to an HLA-B27-restricted
self-epitope correlates with ankylosing spondylitis. J Clin Invest 2000;106:47.
[36] Duchmann R, May E, Ackermann B, et al. HLA-B27 restricted cytotoxic T lymphocyte re-
sponses to arthritogenic enterobacteria or self antigens are dominated by closely related TCR-Bv
gene segments: A study in patients with reactive arthritis. Scand J Immunol 1996;43:101.
[37] Allen RL, Gillespie GM, Hall F, et al. Multiple T cell expansions are found in the blood and
synovial fluid of patients with reactive arthritis. J Rheumatol 1997;24:1750.
[38] Dulphy N, Peyrat MA, Tieng V, et al. Common intra-articular T cell expansions in patients with
reactive arthritis: identical beta-chain junctional sequences and cytotoxicity towards HLA-B27.
J Immunol 1999;162:3830.
 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611 609

[39] Leirisalo-Repo M. Prognosis, course of disease, and treatment of the spondyloarthropathies.

Rheum Dis Clin N Am 1998;24:737.
[40] Sieper J, Braun J. Pathogenesis of spondyloarthropathies. Persistent bacterial antigen, auto-
immunity or both? Arthritis Rheum 1995;38:1547.
[41] Fujinami RS, Oldstone MB. Amino acid homology between the encephalitogenic site of myelin
basis protein and virus: mechanism for autoimmunity. Science 1985;230:1043.
[42] Oldstone MB. Molecular mimicry and autoimmune disease. Mol Cell 1987;50:819.
[43] Wucherpfennig KW. Mechanisms for the induction of autoimmunity by infectious agents. J Clin
Invest 2001;108:1097.
[44] Wucherpfennig KW, Strominger JL. Molecular mimicry in T-cell mediated autoimmunity: viral
peptides activate human T cell clones specific for myelin basis protein. Cell 1995;80:695.
[45] Evavold BD, Sloan-Lancaster J, Wilson KJ, et al. Specific T cell recognition of minimally
homologous peptides: evidence for multiple endogenous ligands. Immunity 1995;2:655.
[46] Hudrisier D, Riond J, Burlet-Schiltz O, et al. Structural and functional identification of major
histocompatibility complex class I-restricted self-peptides as naturally occurring molecular
mimics of viral antigens. Possible role in CD8 + T cell-medicated, virus-induced autoimmune
disease. J Biol Chem 2001;276:19396.
[47] Maier B, Molinger M, Cope A, et al. Multiple cross-reactive self ligands for Borrelia burg-
dorferi-specific HLA-DR-resistant T cells. Eur J Immunol 2000;30:448.
[48] Saren A, Pascolo S, Stevanovic S, et al. Identification of Chlamydia pneumoniae-derived
mouse CD8 epitopes. Infect Immun 2002;70:3336.
[49] Bachmaier K, Neu N, de la Maza LM, et al. Chlamydia infections and heart disease linked
through antigenic mimicry. Science 1999;283:1335.
[50] Steinhoff U, Brinkmann V, Klemm U, et al. Autoimmune intestinal pathology induced by
hsp60-specific CD8 T cells. Immunity 1999;11:349.
[51] Prinz I, Zerrahn J, Kaufmann SH, et al. Promiscuous peptide recognition of an autoreactive
CD8(+) cell clone is responsible for autoimmune intestinal pathology. J Autoimmun 2002;
18:281.
[52] Duchmann R, Lambert C, May E, et al. CD4+ and CD8+ clonal T cell expansions indicate a
role in antigens in ankylosing spondylitis; a study in HLA-B27+ monozygotic twins. Clin Exp
Immunol 2001;123:315.
[53] May E, Märker-Hermann E, Wittig BM, et al. Identical T cell expansion in the colon mucosa
and the synovium of a patient with enterogenic spondyloarthropathy. Gastroenterology 2000;
119:1745.
[54] Pfeifer JD, Wick MJ, Roberts RL, et al. Phagocytic processing of bacterial antigens for class I
MHC presentation to cells. Nature 1993;361:359.
[55] Harding CV, Song R. Phagocytic processing of exogenous particulate antigens by macrophages
for presentation by class I MHC molecules. J Immunol 1994;153:4925.
[56] Rock KL. A new foreign policy. MHC class I molecules monitor the outside world. Immunol
Today 1996;17:131.
[57] Kim SK, Angevine M, Demick K, et al. Induction of HLA-class I-restricted CD8 + CTL
specific for the major outer membrane protein of Chlamydia trachomatis in human genital
tract infections. J Immunol 1999;162:6855.
[58] Loomis WP, Starnbach MN. T cell responses to Chlamydia trachomatis. Curr Opin Microbiol
2002;5:87.
[59] Ugrinovic S, Mertz A, Wu P, et al. A single nonamer from the Yersinia 60-kDa heat shock
protein is the target of HLA-B27 – restricted CTL response in Yersinia-iduced reactive arthritis.
J Immunol 1997;159:5715.
[60] Maksymowych WP, Ikawa T, Yamaguchi A, et al. Invasion by Salmonella typhimurium induces
increased expression of the LMP, MECL, and PA28 proteasome genes and changes in the
peptide repertoire of HLA-B27. Infect Immun 1998;66:4624.
[61] Harding CV, Song R, Griffin J, et al. Processing of bacterial antigens for presentation to class I
and class II MHC-restricted T lymphocytes. Infect Agents Dis 1995;4:1.
610 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611

[62] Fling SP, Sutherland RA, Steele LN, et al. CD8 + T cells recognize an inclusion membrane-
associated protein from the vacuolar pathogen Chlamydia trachomatis. Proc Natl Acad Sci
USA 2001;98:1160.
[63] Kuon W, Holzhütter HG, Appel A, et al. Identification of HLA-B27-restricted peptides from the
Chlamydia trachomatis proteome with possible relevance to HLA-B27-associated diseases.
J Immunol 2001;167:4738.
[64] Stephens RS, Kalman S, Lammel C, et al. Genome sequence of an obligate intracellular
pathogen of humans: Chlamydia trachomatis. Science 1998;282:754.
[65] Rammensee HG, Friede T, Stevanovic S. MHC ligands and peptide motifs: first listing. Im-
munogenetics 1995;41:178.
[66] Rammensee HG, Bachmann J, Emmerich NP, et al. SYFPEITHI: database for MHC ligands
and peptide motifs. Immunogenetics 1999;50:213.
[67] Holzhütter HG, Froemmel C, Kloetzel PM. A theoretical approach towards the identification of
cleavage-determining amino acid motifs of the 20S proteasom. J Mol Biol 1999;286:1251.
[68] Holzhütter HG, Kloetzel PM. A kinetic model of vertebrate 20S proteasome accounting for the
generation of major proteolytic fragments from oligomeric peptide substrates. Biophys J
2000;79:1196.
[69] Kern F, Surel IP, Brock C, et al. T-cell epitope mapping by flow cytometry. Nat Med 1998;
4:975.
[70] Brosterhus H, Brings S, Leyendeckers H, et al. Enrichment and detection of live antigen-
specific CD4(+) and CD8(+) T cells based on cytokine secretion. Eur J Immunol 1999;29:4053.
[71] Ljunggren HG, Kärre K. Host resistance directed selectively against H-2-deficient lymphoma
variants: analysis of the mechanism. J Exp Med 1985;162:1745.
[72] Wucherpfennig KW. Structural basis of molecular mimicry. J Autoimmun 2001;16:293.
[73] Gao XM, Wordsworth P, McMichael A. Collagen-specific cytotoxic T lymphocyte responses in
patients with ankylosing spondylitis and reactive arthritis. Eur J Immunol 1994;24:1665.
[74] Guerassimov A, Zhang Y, Banerjee S, et al. Cellular immunity to the G1 domain of cartilage
proteoglycan aggrecan is enhanced in patients with rheumatoid arthritis but only after removal
of keratan sulfate. Arthritis Rheum 1998;41:1019.
[75] Zou J, Zhang Y, Thiel A, et al. Predominant cellular immune response to the cartilage
autoantigenic G1 aggrecan in ankylosing spondylitis and rheumatoid arthritis. Rheumalology
2003;42:1.
[76] Glant TT, Buzas EI, Finnegan A, et al. Critical roles of glycosaminoglycan side chains of
cartilage proteoglycan (aggrecan) in antigen recognition and presentation. J Immunol 1998;
160:3812.
[77] Zhang Y, Guerassimov A, Leroux JY, et al. Induction of arthritis in Balb/c mice by cartilage link
protein: involvement of distinct regions recognized by T and B lymphosytes. Am J Pathol 1998:
153:1283.
[78] Haanen JB, van Oijen MG, Tirion F, et al. In situ detection of virus-and tumor-specific T-cell
immunity. Nat Med 2000;6:1056.
[79] Kim SK, Devine L, Angevine M, et al. Direct detection and magnetic isolation of Chlamydia
trachomatis major outer membrane protein-specific CD8+ T cells with class I tetramers.
J Immunol 2000;165:728.
[80] Kuckelkorn U, Ruppert T, Strehl B, et al. Link between organ-specific antigen processing by
20S proteasomes and CD8+ T cell-mediated autoimmunity. J Exp Med 2002;195:983.
[81] Ringrose JH, Muijsers AO, Pannekoek Y, et al. Influence of infection of cells with bacteria
associated with reactive arthritis on the peptide repertoire presented by HLA-B27. J Med
Microbiol 2000;50:385.
[82] Brewerton DA, Hart FD, Nicholls A, et al. Ankylosing spondylitis and HL-A27. Lancet 1973;
1:904.
[83] Schlosstein L, Terasaki PI, Bluestone R, et al. High association of an HL-A antigen, W27, with
ankylosing spondylitis. N Engl J Med 1973;288:704.
[84] Mear JP, Schreiber KL, Munz C, et al. Misfolding of HLA-B27 as a result of its B pocket
 W. Kuon, J. Sieper / Rheum Dis Clin N Am 29 (2003) 595–611 611

suggests a novel mechanism for its role in susceptibility to spondyloarthropathies. J Immuno

1999;163:6665.
[85] Colbert RA. HLA-B27 misfolding: a solution to the spondyloarthropathy conundrum? Mol Med
Today 2000;6:224. Allan RL, O’Callaghan CA, McMichael AJ, et al. Cutting edge: HLA-B27
can form a novel b2-microglobulin-free heavy chain homodimer structure. J Immunol 1999;
162:5045.
[86] Hughes EA, Hammond C, Cresswell P. Misfolded major histocompatibility complex class I
heavy chains are translocated into the cytoplasm and degraded by the proteasome. Proc Natl
Acad Sci USA 1997;94:1896.
[87] Allen RL, O’Callaghan CA, McMichael AJ, et al. Cutting edge: HLA-B27 can form a novel
b2-microglobuline-free heavy chain homodimer structure. J Immunol 1999;162:5045.
[88] Khare SD, Bull MJ, Hanson J, et al. Spontaneous inflammatory disease in HLA-B27 trans-
genic mice is independent of MHC class II molecules: a direct role for B27 heavy chains and
not B27-derived peptides. J Immunol 1998;160:101.
[89] Boyle LH, Goodall JC, Opat SS, et al. The recognition of HLA-B27 by human CD(+)
T lymphocytes. J Immunol 2001;167:2619.
[90] Kollnberger S, Bird L, Sun MY, et al. Cell surface expression and immune receptor recognition
of HLA-B27 homodimers. Arthritis Rheum 2002;46:2972.
[91] McMichael A, Bowness P. HLA-B27: natural function and pathogenic role in spondyloarthritis.
Arthritis Res 2002;4(Suppl 3):153.
[92] Lanier LL. Follow the leader: NK cell receptors for classical and nonclassical MHC class I. Cell
1998;92:705.
[93] Malik P, Klimovitsky P, Deng LW, et al. Uniquely conformed peptide-containing b2-micro-
globulin-free heavy chains of HLA-B2705 on the cell surface. J Immunol 2002;169:4379.
[94] Urban RG, Chicz RM, Lane WS, et al. A subset of HLA-B27 molecules contains peptides
much longer than nonamers. Proc Natl Acad Sci USA 1994;91:1534.
[95] Malik P, Strominger JL. Perfusion chromatography for very rapid purification of class I and
class II MHC proteins. J Immunol Methods 2000;234:83.
[96] Ramos M, Alvarez I, Sesma L, et al. Molecular mimicry of an HLA-B27 – derived HLA-B27
ligand of arthritis-linked subtypes with chlamydial proteins. J Biol Chem 2002;277:37573.
[97] Popov I, Dela Cruz CS, Barber BH, et al. Breakdown of CTL tolerance to self HLA-B * 2705
induced by exposure to Chlamydia trachomatis. J Immunol 2002;169:4033.
[98] Hoh J, Jin S, Parrado T, et al. The p53 MH algorithm and its application in detecting
p53-responsive genes. Proc Natl Acad Sci USA 2002;99:8467.
[99] Bradley P, Kim PS, Berger B. TRILOGY: discovery of sequence-structure patterns across
divers proteins. Proc Natl Acad Sci USA 2002;99:8500.
[100] Miconnet I, Coste I, Beermann F, et al. Cancer vaccine design: a novel bacterial adjuvant for
peptide-specific CTL induction. J Immunol 2001;166:4612.