The Role of Endoplasmic Reticulum Aminopeptidase 1 (ERAP1 ... · The Role of Endoplasmic Reticulum...
222
The Role of Endoplasmic Reticulum Aminopeptidase 1 (ERAP1) in the Pathogenesis of Ankylosing Spondylitis by Nigil Haroon A thesis submitted in conformity with the requirements for the degree of Doctorate of Philosophy in Medical Sciences Institute of Medical Sciences University of Toronto © Copyright by Nigil Haroon 2012
Transcript of The Role of Endoplasmic Reticulum Aminopeptidase 1 (ERAP1 ... · The Role of Endoplasmic Reticulum...
The Role of Endoplasmic Reticulum Aminopeptidase 1 (ERAP1) in the Pathogenesis of Ankylosing Spondylitis
by
Nigil Haroon
A thesis submitted in conformity with the requirements for the degree of Doctorate of Philosophy in Medical Sciences
Institute of Medical Sciences University of Toronto
© Copyright by Nigil Haroon 2012
ii
The Role of Endoplasmic Reticulum Aminopeptidase 1 (ERAP1)
in the Pathogenesis of Ankylosing Spondylitis
Nigil Haroon
Doctorate of Philosophy in Medical Sciences
Institute of Medical Sciences University of Toronto
2012
Abstract
Ankylosing spondylitis (AS) is associated with HLA-B*2704 and B*2705 but not
with HLA-B*2706 and B*2709. Genome wide studies recently identified ERAP1 as an
important genetic association in AS and could be the missing link in the pathogenesis of
AS.
I studied the implications of the two known actions of ERAP1 on AS
pathogenesis. For assessing the peptide trimming function, surface HLA-B27 and MHC-I
free heavy chain (FHC) expression on peripheral blood mononuclear cells of AS patients
were studied. Subsequently, in an in vitro system of C1R cells expressing different AS-
associated and AS-neutral HLA-B27 subtypes, I studied the effect of ERAP1 suppression
on HLA-B27 and FHC expression. To assess the cytokine receptor shedding function, I
studied serum cytokine receptor level variation with ERAP1 polymorphisms and its
relationship to disease activity in AS patients. Finally, I studied the effect of variants of
ERAP1 and other members of the antigen presentation machinery on radiographic
severity in AS patients.
iii
AS patients with the major allele of the ERAP1 rs27044 polymorphism had higher
FHC expression on monocytes. In C1R cells ERAP1 suppression led to an increase in
intracellular FHC (IC-FHC) and B27-peptide complexes identified by a special MARB4
antibody, but only in C1R cells expressing the AS-associated subtypes HLA-B*2704 and
B*2705. ERAP1 variants had no effect on serum cytokine receptor levels. Baseline
radiographic severity was associated with ERAP1 polymorphism in univariate analysis
only. LMP2 variants were associated with baseline radiographic severity in multivariate
analysis.
ERAP1 affects peptide presentation and FHC formation by HLA-B27 and could be
the missing link in the pathogenesis of AS. ERAP1 through its differential HLA-B27
subtype interaction could explain why certain subtypes of HLA-B27 are associated with
AS while others are not. Larger studies are required to look closely at the effect of
ERAP1 on radiographic severity and progression in AS.
iv
Acknowledgments
v
To
Dad and Mom
Who selflessly gave up everything for me
vi
Table of Contents
Acknowledgments ........................................................................................................... iv
Table of Contents ............................................................................................................ vi
List of Figures ................................................................................................................. xi
List of Tables................................................................................................................. xiii
List of Abbreviations ...................................................................................................... xiv
Chapter 1 Introduction .................................................................................................... 1
Ankylosing Spondylitis ................................................................................................. 1 1
1.1 History ................................................................................................................... 1
1.2 Epidemiology ......................................................................................................... 2
1.3 Clinical Features ................................................................................................... 3
1.3.1 Inflammatory Back Pain (IBP)..................................................................... 3
1.3.2 Peripheral Arthritis ...................................................................................... 5
1.3.3 Extra-articular Features .............................................................................. 5
1.4 Radiographic Features .......................................................................................... 6
1.5 Classification Criteria............................................................................................. 8
1.5.1 Modified New York Criteria ......................................................................... 8
1.5.2 ASAS Criteria for Axial Spondyloarthritis .................................................... 9
1.6 Treatment of Ankylosing Spondylitis ................................................................... 12
1.6.1 Non-pharmacological treatment ............................................................... 13
1.6.2 Non-steroidal anti-inflammatory drugs ...................................................... 13
1.6.3 Disease modifying anti-rheumatic drugs (DMARDs) ................................ 14
1.6.4 Biological Therapy .................................................................................... 14
1.7 Pathogenesis ...................................................................................................... 16
1.7.1 Infections and Pathogenesis of AS ........................................................... 16
vii
1.7.2 Cytokines in AS ........................................................................................ 18
1.7.3 Genetics of ankylosing spondylitis ............................................................ 21
1.7.4 Inflammation and New Bone Formation in AS .......................................... 29
1.7.5 DKK-1 in the Pathogenesis of Ankylosing Spondylitis .............................. 31
1.8 Endoplasmic Reticulum Aminopeptidase 1 ......................................................... 35
1.8.1 Structure of ERAP1 .................................................................................. 37
1.8.2 Genetic association of ERAP1 with AS .................................................... 47
1.8.3 Normal functions of ERAP1 ...................................................................... 49
1.8.4 ERAP1 polymorphisms and AS phenotype .............................................. 59
1.8.5 ERAP1 and HLA-B27 pathogenesis ......................................................... 60
1.8.6 ERAP1 in other diseases .......................................................................... 60
1.9 Aim ...................................................................................................................... 63
1.10 Hypothesis .......................................................................................................... 64
Chapter 2 ....................................................................................................................... 65
Functional interaction of ERAP1 and HLA-B27 is subtype specific and could be 2pathogenic in ankylosing spondylitis .......................................................................... 65
2.1 Introduction ......................................................................................................... 65
2.2 Methods .............................................................................................................. 68
2.2.1 Patients .................................................................................................... 68
2.2.2 Genotyping ............................................................................................... 68
2.2.3 Cell Lines .................................................................................................. 70
2.2.4 Optimization of Nucleofection ................................................................... 70
2.2.5 ERAP1 Suppression by RNA Interference ............................................... 73
2.2.6 Western Blot Assay .................................................................................. 73
2.2.7 Fluorescence Activated Cell Sorting (FACS) ............................................ 75
2.2.8 Statistics ................................................................................................... 78
viii
2.3 Results ................................................................................................................ 79
2.3.1 Monocytes have the highest level of surface expression of HLA B27 and free heavy chains. ............................................................................. 80
2.3.2 ERAP1 rs27044 polymorphisms affect the expression of free heavy chains on the surface of monocytes ......................................................... 83
2.3.3 Surface expression of HLA B27 and FHC was not related to disease activity: ..................................................................................................... 88
2.3.5 ERAP1 suppression of HLA-B*2705 C1R cells causes changes in intracellular free heavy chains and surface MHC-peptide complexes ...... 94
2.3.6 C1R cells expressing different HLA-B*27 subtypes have variable level of total MHC-I expression ......................................................................... 94
2.3.7 C1R cells expressing AS-susceptible HLA-B*27 subtypes have high MARB4 staining. ....................................................................................... 97
2.3.8 ERAP1 suppression leads to an increase in intracellular free heavy chains only in C1R cells expressing AS associated HLA-B*27 subtypes . 97
2.3.9 ERAP1 suppression leads to increase in MARB4 staining only in C1R cells expressing AS associated HLA-B*27 subtypes .............................. 100
2.4 Discussion ......................................................................................................... 107
2.5 Conclusions: ...................................................................................................... 114
Chapter 3 ..................................................................................................................... 115
Serum Cytokine Receptors in Ankylosing Spondylitis: Relationship to 3inflammatory markers and Endoplasmic Reticulum Aminopeptidase Polymorphisms ...................................................................................................... 115
3.1 Background ....................................................................................................... 115
3.2 PATIENTS AND METHODS ............................................................................. 116
3.2.1 Patients .................................................................................................. 116
3.2.2 Genotyping ............................................................................................. 117
3.2.3 Serum cytokine and soluble cytokine receptor levels ............................. 118
3.2.4 Statistical analysis .................................................................................. 119
3.3 RESULTS .......................................................................................................... 120
ix
3.3.1 Patients .................................................................................................. 120
3.3.2 Genotyping ............................................................................................. 121
3.3.3 Soluble cytokine receptor levels ............................................................. 121
3.3.4 No significant effect of ERAP1 polymorphism on serum cytokine receptor level .......................................................................................... 124
3.3.5 Serum cytokine levels ............................................................................. 124
3.4 Discussion ......................................................................................................... 129
3.5 Conclusion ........................................................................................................ 131
Chapter 4 ..................................................................................................................... 132
Radiographic severity in ankylosing spondylitis: The effect of genetic 4polymorphisms in ERAP1 and other elements of the antigen presenting machinery ............................................................................................................... 132
4.1 Background ....................................................................................................... 132
4.2 Methods ............................................................................................................ 134
4.2.1 Patients .................................................................................................. 134
4.2.2 Radiographic Scoring ............................................................................. 135
4.2.3 Genotyping ............................................................................................. 137
4.2.4 Statistical Analysis .................................................................................. 139
4.3 Results .............................................................................................................. 142
4.3.1 Patients .................................................................................................. 142
4.3.2 Baseline radiographic severity can be influenced by polymorphisms in LMP2 ...................................................................................................... 144
4.3.3 Baseline mSASSS is the strongest predictor of progression .................. 146
4.4 Discussion ......................................................................................................... 151
4.5 Conclusions ....................................................................................................... 156
Chapter 5 ..................................................................................................................... 157
Conclusions and Future Directions ..................................................................... 157 5
5.1 Conclusions ....................................................................................................... 157
x
5.2 Future Directions ............................................................................................... 159
5.2.1 Studies in Ankylosing Spondylitis Patients ............................................. 160
5.2.2 In vitro studies ........................................................................................ 161
References.................................................................................................................. 164
xi
List of Figures
Figure 1.1 The Canonical Wnt-β Catenin Pathway ........................................................33
Figure 1.2 The Chormosome 5q15 locus with ERAP1 and other surrounding genes ....38
Figure 1.3 Nucleotide Change in the ERAP1 Polymorphisms Studied ..........................40
Figure 1.4 Residue Change in the ERAP1 Polymorphisms Studied ..............................41
Figure 1.5 ERAP1 Crystal Structure: Open Configuration .............................................43
Figure 1.6 ERAP1 Crystal Structure: Closed Configuration ...........................................45
Figure 1.7 Schematic Diagram to show the role of endoplasmic reticulum
aminopeptidases in antigen processing and presentation .............................................50
Figure 1.8 ERAP1 Peptide specificity ............................................................................53
Figure 2.1 Western Blot and analysis showing ERAP1 suppression with ERAP1 siRNA
compared to the respective negative control siRNA ......................................................74
Figure 2.2 Flow Assisted Cell Sorting of Peripheral Blood Mononuclear Cells ..............76
Figure 2.3 Total HLA-B*27 and Free Heavy Chain Expression was highest in the
monocyte population among the peripheral blood mononuclear cells. ...........................84
Figure 2.4 Flow Cytometry Analysis of MHC-I Free Heavy Chain Expression on
Peripheral Blood Mononuclear Cells from Patients with Ankylosing Spondylitis ............86
Figure 2.5 Comparing the effects of ERAP1 rs27044 SNP on free heavy chain and
Intact HLA-B27 expression on different cell populations ................................................89
Figure 2.6 Comparing the effects of ERAP1 rs30187 SNP on free heavy chain and
Intact HLA-B27 expression on different cell populations ................................................91
xii
Figure 2.7 Comparison of inflammatory markers across the different genotypes ..........93
Figure 2.8 Effect of siRNA induced ERAP1 suppression on HLA-B*27 expression in
C1R cells expressing HLA-B*2705 ................................................................................95
Figure 2.9 Western Blot and FACS analysis of C1R cells at baseline. ..........................98
Figure 2.10 The effect of ERAP1 Suppression on Intracellular Free Heavy Chain
Expression in C1R cells expressing different HLA-B*27 subtypes ...............................101
Figure 2.11 Percent of cells with high expression of intracellular free heavy chains ....103
Figure 2.12 The effect of ERAP1 suppression on HLA-B*27 complexes recognized by
MARB4 antibody on the surface of C1R cells expressing different HLA-B*27 subtypes
.....................................................................................................................................105
Figure 3.1 Correlation of TNFRI with CRP and ESR ....................................................122
Figure 3.2 No correlation of CRP, ESR and BASDAI with serum IL-6Rα and IL-1RII ..125
Figure 3.3 Serum Cytokine Receptors in patients with different ERAP1 polymorphisms
.....................................................................................................................................127
Figure 4.1 Modified Stoke’s Ankylosing Spondylitis Spine Score.................................136
Figure 4.2 Cumulative Probability Plot of patients with progression of radiographic
scores ..........................................................................................................................147
Figure 4.3 ROC Curve for regression model with baseline mSASSS predicting
radiographic progression .............................................................................................150
Figure 4.4 Hypothetical mechanism for LMP2 and ERAP1 variants affecting
radiographic progression in AS ....................................................................................155
xiii
List of Tables
Table 1.1 Modified New York criteria for ankylosing spondylitis......................................... 10
Table 1.2 ASAS classification criteria for axial Spondyloarthritis ....................................... 11
Table 1.3 Hypothesis of the role of HLA-B27 in the pathogenesis of Ankylosing
Spondylitis ................................................................................................................................... 23
Table 1.4 Single Nucleotide Polymoprhisms of ERAP1 with Nucleotide and Residue
Change ......................................................................................................................................... 41
Table 2.1 The Amaxa cell line optimization nucleofector kit ............................................... 72
Table 2.2 Clinical and laboratory features of patients with different ERAP1 rs27044
genotypes .................................................................................................................................... 81
Table 2.3 Clinical and laboratory features of patients with different ERAP1 rs30187
genotypes .................................................................................................................................... 82
Table 4.1 Genotype frequency of the tested SNPs in the study patients ........................ 138
Table 4.2 Demographic features of the two AS cohorts .................................................... 143
Table 4.3 Univariate analysis of predictors of baseline radiographic severity ................ 145
Table 4.4 Predicting patients who are likely to progress on follow up ............................. 148
xiv
List of Abbreviations
A-LAP Adipocyte-derived Leucine Aminopeptidase
AP-1 Activated Protein 1
AR Attributable Risk
ARTS-1 Aminopeptidase Regulating Tumor necrosis factor receptor I Shedding 1
AS Ankylosing Spondylitis
ASAS Assessment of Spondyloarthritis International Society
AUC Area Under Curve
BASDAI Bath Ankylosing Spondylitis Disease Activity Index
BASFI Bath Ankylosing Spondylitis Functional Index
BME Bone Marrow Edema
β2m β2-Microglobulin
CAST Calpastatin
CI Confidence Interval
CIL Corner Inflammatory Lesions
COXIBs Cycloxygenase II Inhibitors
CRP C Reactive Protein
CTX-II Cross-linking Telopeptide of type II
DDI Damage Duration Index
xv
DISH Diffuse Idiopathic Skeletal Hyperostosis
DKK1 Dickopf 1
DMARDs Disease modifying anti-rheumatic drugs
EMS Early Morning Stiffness
ERAAP Endoplasmic Reticulum Aminopeptidase Associated with Antigen
Processing
ERAP1 Endoplasmic Reticulum Aminopeptidase 1
ERAP2 Endoplasmic Reticulum Aminopeptidase 2
ESR Erythrocyte Sedimentation Rate
FBS Fetal Bovine Serum
FHC Free Heavy Chain
GFP Green Fluorescence Protein
GPR Global Posture Re-education
GSK-3β Glycogen Synthase Kinase-3β
GWAS Genome Wide Association Study
HWE Hardy Weinberg Equillibrium
HLA Human Leukocyte Antigen
HSP60 Heat Shock Protein-60kD
HUGO Human Genome Organization
IBD Inflammatory Bowel Disease
IBP Inflammatory Back Pain
xvi
ICC Inter-class Correlation Coefficients
IC-FHC Intracellular Free Heavy Chain
IFNγ Interferon Gamma
IL Interleukin
Int1 Integration 1
IQR Inter Quartile Range
JAK Janus Kinase
KIR Killer cell Immunoglobulin-like Receptors
L-Amc leucine-7-amino-4-methylcoumarin
LIR Leukocyte Immunoglobulin-like Receptors
LD Linkage Disequilibrium
LMP2 Large Multifunctional Peptidase 2
L-RAP Leukocyte-derived Arginine Aminopeptidase
LRP LDL-receptor like protein
MAF Minor Allele Frequency
MFI Mean Fluorescence Intensities
MHC Major Histocompatibility Complex
MMP3 Matrix Metalloproteinase 3
MMTV Mouse Mammary Tumor Virus
MRI Magnetic Resonance Imaging
xvii
MS Mass Spectroscopy
mSASSS Modified Stoke’s Ankylosing Spondylitis Spine Score
∆mSASSS Change in Modified Stoke’s Ankylosing Spondylitis Spine Score
mTNF Membrane-bound Tumor Necrosis Factor-α
MTX Methotrexate
NASC North American Spondylitis Consortium
NC Negative Control
NF-κB Nuclear Factor-κB
OR Odds Ratio
PASW Predictive Analytics Software
PBMC Peripheral Blood Mononuclear Cells
pGR Peptide- Glucagon Receptor
PI Propidium Iodide
PILS-AP Puromycin-Insensitive Leucyl Aminopeptidase
pLMP2 Peptide- Latent Membrane Protein 2
pVIPR Peptide- Vasoactive Intestinal Polypeptide Receptor
RCT Randomized Control Trial
R-Amc Arginine-7-amino-4-methylcoumarin
RBMX RNA-Binding Motif gene, X chromosome
ROC Receiver Operating Characteristic Curve
xviii
SDC Smallest Detectable Change
SIJ Sacroiliac Joints
siRNA Short Interfering RNA
SNP Single Nucleotide Polymorphisms
STIR Short Tau Inversion Recovery
sTNF Soluble Tumor Necrosis Factor-α
TAP Transporter associated with Antigen Processing
TASC The Australo-Anglo-American Spondylitis Consortium
TCF T Cell Factor
TIFF3 Tricorn-Interacting Factor F3
TNFα Tumor Necrosis Factor-α
UPR Unfolded Protein Responses
WB Western Blot
Wg Wingless
WTCCC Welcome Trust Case Control Consortium
1
Chapter 1
Introduction
A part of this chapter describing the biology of Endoplamic Reticulum Aminopeptidase 1 was published as a review
in Nature Review’s Rheumatology. The citation for this article is: Haroon N and Inman RD. Endoplasmic reticulum
aminopeptidases: biology and pathogenic potential. Nat Rev Rheumatol. 2010 Aug;6(8):461-7. RDI revised and
approved the draft.
Ankylosing Spondylitis 1Ankylosing spondylitis (AS) is a chronic inflammatory rheumatic disease
characterized by arthritis affecting predominantly the joints of the spine. AS classically
begins with inflammation in the sacroiliac joints but over time affects the rest of the
spine. Uncontrolled inflammation can eventuate in osteoproliferation, with pathologic
new bone formation and eventually bony ankylosis of the spine. The cause of AS is
unknown and the basic mechanisms underlying chronic inflammation remain
undefined.
Patient with AS can have significant impairment of their quality of life[1-4]. Pain
and stiffness can result in functional and work disability[5-10]. Long term sequelae
including increased risk of cardiovascular morbidity and mortality are now well
recognized in AS[11-14].
1.1 History
The existence of AS has been traced to biblical times with at least three of the
Pharaohs of Egypt having suffered from the disease[15]. In 1963 Bernard Connor, an
2
Irish Physician, described an usual skeleton from a French cemetery in which the
sacrum, ilium, lumbar and thoracic spine as well as some ribs were fused like a single
bone[16]. Remarkably by dissection he identified that ankylosis occurred along the
annulus fibrosus while the nucleus pulposus was free of such change. This is
consistent with the current concept of ankylosis in AS. AS is also known as Marie-
Strümpell disease, after the clinicians who published the first clinical reports of
AS[17,18]. Forestier demonstrated sacroiliitis and syndesmophytes formation, the two
cardinal features of AS[19,20]. Syndesmophytes refer to new bone formation seen at
the vertebral corners. These can be seen as bony spurs in x-rays developing from the
upper and lower vertebral corners, both anteriorly and posteriorly.
1.2 Epidemiology
The prevalence of AS varies from 0.2-0.8%, depending on the population
studied and carries with it health care costs which are comparable to rheumatoid
arthritis[21-23]. The peak onset of AS is in the third decade of life and imposes on
patients significant pain and disability at a key time of life. Males are affected two to
three times more frequently than females[24]. This ratio might however be an
overestimate as the degree of suspicion of AS in women is lower and they usually have
a longer delay in diagnosis compared to men[25]. With improved diagnostic criteria
and the advent of MRI, more females with AS may be identified. Females may have
more pain and radiographic changes in the neck while men have more lumbar
involvement[26,27]. This could also reduce the degree of suspicion of AS in females.
3
There is a strong association of AS with HLA-B27 with 90% of AS patients
having the gene[28]. The prevalence of AS tends to parallel the frequency of HLA-B27
in the population[29]. HLA-B27 is almost absent in the Australian aboriginal population
where AS is virtually non-existant[30]. In contrast to a low prevalence of HLA-B27 and
AS in Japan[31] and Africa[32-35], Alaskan Eskimos, Inuits and Siberian Chukotkas
have a high prevalence of both HLA-B27 and spondyloarthropathies, reaching 10% in
Haida Indians[36-41].
1.3 Clinical Features
1.3.1 Inflammatory Back Pain (IBP)
Inflammatory Back Pain (IBP) is the characteristic feature distinguishing
inflammatory arthritis of the spine from mechanical causes of back ache. Calin et al.
established criteria to identify IBP with 95% sensitivity and 85% specificity in 1977
[42]. The criteria would be satisfied if 4 out of 5 of the following features were
present: (i) age of onset less than 40 years (ii) insidious onset (iii) more than 3 months
duration (iv) improvement with exercise and (v) early morning stiffness. These criteria
were further modified by Rudwaleit et al to improve the sensitivity and specificity, with
the emphasis being on differentiating patients with AS from those with mechanical
back pain [43].
Despite these criteria for identifying IBP, the physician’s opinion in this regard is
very subjective. To overcome this hurdle, the aim of developing new criteria for IBP
was to approximate expert opinion as closely as possible. If this goal is met, early
diagnosis of axial spondyloarthritis can become a reality. The Assessment of
4
Spondyloarthritis International Society (ASAS) has now developed new criteria for IBP
with this in mind [44]. Various clinical history items were considered in developing the
new criteria, with the gold standard being expert opinion. Thirteen AS experts from 9
countries in Europe and North America evaluated 20 patients and formulated an
opinion whether IBP was present or not, with no knowledge of the final diagnosis of
the patient. All patients had chronic back pain and were undergoing diagnostic work
up at a rheumatology clinic. This kind of analysis removes the heavy bias affecting
decision-making in IBP with prior knowledge of the diagnosis.
The new ‘IBP according to experts’ criteria had the following five parameters:
(1) age at onset <40 years (2) insidious onset (3) improvement with exercise (4) no
improvement with rest and (5) pain at night with improvement on getting up [44].
Despite the sound methodology used, the specificity and sensitivity of these criteria
were not significantly different from previous criteria in a validation study on 648
patients with chronic back pain [44]. It should be noted that some cardinal parameters
were not retained after the procedure. Early morning stiffness (EMS) was not included
in the new criteria. It is known that EMS can be seen both in patients with IBP and
with mechanical pain, but if quantified as greater or less than 30 minutes in duration,
it might be a useful differentiating parameter. However in this study EMS did not stand
out in the regression analysis. Similarly pain in the second half of the night might be a
more specific feature compared to just nocturnal pain alone. This timing of night pain
was not specified in this study [44]. Unlike traditional practice of combining worsening
of pain with rest and improvement with exercise as a single parameter, here they have
been included separately as both are independently associated with IBP. Despite the
5
inherent problems of using ‘IBP according to experts’ as the gold standard, this
undertaking is a step in the right direction and continued efforts in this direction could
give us better criteria that can contribute to decreased diagnostic delays in AS.
1.3.2 Peripheral Arthritis
Peripheral arthritis is seen in up to 30% of patients with AS. Classically, it is an
asymmetrical oligoarthritis with a predilection for the lower extremities. Thus patients
usually present with pain and swelling affecting knees and ankles. Other joints can also
be affected, but typically less than 6 joints are affected at a time. Arthritis can be
associated with tenosynovitis and enthesitis. Common enthesial sites affected are
Achilles tendon insertion, plantar fasciitis, tibial tuberosity, peripatellar tendon
insertions, spinal enthesis, iliac crest and iliac spines as well as costochondral
junctions. Hip joint involvement is a marker of severity of AS. It is more common with
younger age of onset but can occur at any point in the course of AS.
1.3.3 Extra-articular Features
The common extra-articular manifestations associated with AS are inflammatory
bowel disease (IBD), uveitis and psoriasis. Acute anterior uveitis is seen in up to 40%
of AS patients [45]. Classical presentation includes unilateral blurring of vision,
photophobia, headache and redness. An acute attack of uveitis can last for a few
weeks. HLA-B27 positive patients have a higher risk of uveitis in AS [46]. Uveitis can
be recurrent and sometimes chronic with damage to the eyes, if not treated
appropriately and promptly.
6
Inflammatory bowel disease (IBD) and psoriasis occur in approximately 10% of
AS cohorts, respectively [47,48]. Asymptomatic microscopic colitis can occur in a
higher percentage (40-60%) of patients with spondyloarthritis compared to other
inflammatory arthritis [49-53]. Primary AS differs from that associated with psoriasis
and IBD with an early onset of illness and more severe spinal restriction [54]. Chronic
inflammation in the gut is associated with higher prevalence spinal restriction, bamboo
spine and sacroiliitis [55].
Other extra articular manifestations include fever, fatigue, aortic insufficiency,
cardiac conduction defects, and pulmonary fibrosis. There is a higher incidence of IgA
nephropathy in AS. Amyloidosis is rare in AS. Somewhat paradoxically in the context of
the bone forming tendency in AS, osteoporosis and osteopenia is frequent [56,57].
Males and patients with decreased functional capacity have lower bone density[58].
There is a high prevalence of vertebral fractures that are likely underdiagnosed [59].
Age, male gender, poor functional capacity, higher occiput to wall distance and higher
radiographic severity are risk factors for fracture in AS [59].
1.4 Radiographic Features
Bilateral sacroiliitis identified by x-ray is ordinarily required to make a diagnosis
of AS. Hence there can be a delay in diagnosis till radiographic changes appear.
Sacroiliitis manifests as subchondral sclerosis, joint erosions and ankylosis in
progressively increasing grades of severity. A basic radiographic assessment of AS
should include X-rays of the spine as well. Erosions of the vertebral corners can be
seen. When it is present on both anterior corners of the vertebral body, the normal
anterior concavity is lost and this is referred to as squaring. Post-inflammatory changes
7
and new bone formation is thought to mediate the sclerosis of vertebral corners and
they appear as ‘shiny corners’.
Exuberant new bone formation manifests as bony spurs developing from the
vertebral corners. These bony spurs are called syndesmophytes when they extend
from the vertebral corner in the direction of the spinal axis. They can grow and ‘bridge’
across to the adjacent vertebral corner when they are called ‘bridging
syndesmophytes’. In addition to the anterior elements of the spine, posterior elements
including facet joints can undergo ankylosis and need to be assessed. The end result
can be extensive bridging syndesmophytes from cervical to sacral vertebrae and this is
called a ‘bamboo spine’.
The introduction of Magnetic Resonance Imaging (MRI) in the diagnosis of
spondyloarthritis has greatly improved the sensitivity for diagnosing AS. In the new
ASAS criteria for axial spondylarthritis, MRI evidence of inflammation in the sacroiliiac
joints is sufficient. Such patients, unless they have sacroiliitis by X-ray imaging, cannot
be labeled as AS as defined by the modified New York criteria discussed below. The
short tau inversion recovery (STIR) sequences are special sequences which suppress
fat signals and bring out signals from free water as present in bone marrow edema
(BME) [60]. In T2 STIR sequences active inflammation can be seen as hyperintense
areas in the spine with the main areas being vertebral corners (corner inflammatory
lesions or CILs), SI joints, facet joints, costovertebral and costotransverse joints. In
T1w sequences, structural changes like erosions, ankylosis and fat replacement can
be seen. Erosions are seen in the vertebral corners as well in the sacroiliac joints.
Fatty replacement can be seen in the vertebral corners and in the periarticular areas of
8
the sacroiliac (SI) joints. When present as hyperintense lesions at the vertebral corners
they are referred to by some authors as ‘fatty Romanus lesion’ (as the location is the
same as Romanus lesions in X-ray) [61]. It should however be noted that erosions are
better seen with CT scan [62].
The ASAS/OMERACT MRI group has published consensus criteria for
diagnostic MRI in axial SpA [63]. Neither active vertebral lesions nor chronic sacroiliiac
and vertebral lesions qualify as a diagnostic MRI finding. BME should be located in the
subchondral or periarticular areas of the SI joint. At least two separate BME lesions in
one slice or one BME lesion that is visible in two consecutive slices is the minimum
intensity of lesion required [63]. MRI can also be used as a tool to judge efficacy of
anti-TNF agents as it is sensitive to change in spinal inflammation. In the ASSERT
study, 194 patients on infliximab (dose of 5mg/kg) were compared to 72 on placebo.
After 6 months of therapy, there was almost complete elimination of MRI inflammation
in the infliximab group with no significant change in the placebo group. In the CANDLE
study which used low dose infliximab (3mg/kg) [64] there was significantly more
reduction in MRI inflammation even at 12 weeks in patients on active therapy
compared to placebo (p<0.002). All patients showed a decrease in inflammation in the
Infliximab group. This effect is by no means limited to Infliximab and has been shown
with etanercept and adalimumab [65-67].
1.5 Classification Criteria
1.5.1 Modified New York Criteria
9
The modified New-York criteria (Table1.1) are the currently accepted standard for the
diagnosis and classification of AS [68]. A diagnosis of AS is made if sacroiliitis is
associated with low back pain or restriction of spinal mobility. The X-ray criterion is
required to make a diagnosis of AS. Hence, although specificity is good, the sensitivity
decreases, especially in early stages of AS. This raises the criticism that the Modified
NY criteria leads to a delay in diagnosis of AS. To meet the demand for more sensitive
criteria, the ASAS group has developed new criteria for classifying axial
spondyloarthritis.
1.5.2 ASAS Criteria for Axial Spondyloarthritis
The revised classification criteria for axial SpA (Table 1.2) was proposed and
validated by ASAS [69,70]. This effort from the ASAS group stems from an unmet need
for early diagnosis of SpA with predominant axial symptoms. To make a diagnosis of
AS according to the modified New York criteria, there should be X-ray evidence of
bilateral grade II or at least unilateral grade III sacroiliitis [68]. The delay in diagnosis of
AS can be as high as 8 to 10 years [25,71,72]. It might be many years before any
changes appear in X-rays and it is now known that MRI can precede X-ray changes by
as much as 8 years [73]. There is tremendous progress in early diagnosis of RA and
the benefits of the same are being seen in the form of better control and even healing
of erosions. There is no such data in AS and the first step in initiating a trial is to define
early SpA. The development of the ASAS criteria for axial SpA is an attempt at
resolving this gap.
10
Table 1.1 Modified New York criteria for ankylosing spondylitis
_____________________________________________________________________________
I. Clinical criteria
a) Low back pain and stiffness for more than 3 months which improves with exercise,
but is not relieved by rest.
b) Limitation of motion of the lumbar spine in both the sagittal and frontal planes.
c) Limitation of chest expansion relative to normal values corrected for age and sex.
II. Radiolographic criterion
Sacroiliitis grade 2 bilaterally or sacroiliitis grade 3-4 unilaterally.
____________________________________________________________________
A diagnosis of AS is made if the radiographic criteria is associated with one clinical criteria.
11
Table 1.2 ASAS classification criteria for axial Spondyloarthritis
Entry Criteria: Patients should be 45 years of age or younger and have back pain for
more than 3 months for this criteria to be applicable.
These patients can be classified as axial SpA if they have either A. Imaging evidence of sacroiliitis and at least 1 other SpA feature OR B. HLA B27 and at least 2 other SpA features Imaging evidence: Definite sacroiliitis according to modified New York criteria or MRI
showing definite bone marrow edema or osteitis in sacroiliac joints.
SpA features:
1. Inflammatory back pain 2. Arthritis 3. Enthesitis 4. Uveitis 5. Dactylitis 6. Psoriasis 7. Inflammatory bowel disease 8. Good NSAID response 9. Family history of SpA 10. HLA B27 11. Elevated CRP
12
For the first time, MRI has been introduced as an accepted imaging modality for
identifying sacroiliitis and a patient can be classified as axial SpA without any X-ray
changes (Table 1.2). Another novelty of this classification is a non-imaging arm,
wherein a patient with HLA-B27 and no radiographic evidence of sacroiliitis can
potentially qualify to have SpA [69,70]. There are several criticisms and the predominant
one is that it is too inclusive. A good example would be that of a person with long
standing mechanical back pain and a family history of SpA. There is a 7% chance that
this person is HLA B27 positive and he could be erroneously classified as axial SpA.
1.6 Treatment of Ankylosing Spondylitis
In AS, the goal of treatment is control of symptoms. Pain, stiffness, fatigue,
peripheral arthritis, dactylitis, enthesitis and tenosynovitis usually responds to the
current modalities of treatment available for AS patients. Unfortunately we do not have
any proven modalities for disease modification or cure. There is some evidence with
anti-inflammatory agents that treatment over prolonged periods of time may result in
some slowing of radiographic damage. Assessment of treatment response is based on
patient reported outcome measures like the Bath AS Disease Activity Index (BASDAI),
Bath AS Functional Index (BASFI) and Total Back Pain (TBP) score. BASDAI is
calculated from a questionnaire that assesses symptoms like back pain, entheseal/joint
pain, duration and severity of EMS. The score is averaged over a 0-10 scale with a
score of 4 indicating inadequate control of symptoms. When assessing treatment
response, the minimum clinically significant difference in BASDAI is estimated to be
20% or 2 units in a scale of 0-10. Thus if the BASDAI improves from 6 to 4 during
treatment, this is a clinically significant change.
13
Improvement in functional ability, or the ability to do routine and leasure
activities is captured in a questionnaire and the average score gives the BASFI.
Improvement in BASFI gives us an idea if the treatment actually changed the ability of
the patient to do physical activity.
1.6.1 Non-pharmacological treatment
There has been a surge of trials on the effect of physiotherapy in AS during the last
decade. There is strong evidence for clinically significant improvement with regular
physiotherapy and multimodal exercise therapy [74,75]. The aim of physiotherapy is to
improve and maintain global health and improve the functional status of patients.
Group therapies and supervised therapies are more effective than home based
physiotherapy [76]. Global posture re-education (GPR) refers to a specific
physiotherapy technique using gentle exercises on muscles groups aiming to correct
joint alignment, muscle strengthening, and stretching of muscles that have undergone
contracture. GPR resulted in better BASDAI and BASFI scores than routine
physiotherapy in a recently published randomized controlled trial (RCT) [77].
Occupational therapy and balneotherapy are additional adjunctive modalities of
treatment in AS [78,79].
1.6.2 Non-steroidal anti-inflammatory drugs
NSAIDs have remained the mainstay of treatment of AS for many years. NSAIDs
effectively reduce the signs and symptoms of AS [80,81]. This is sufficiently
reproducible that response to NSAIDs has been postulated as a classification criterion
14
for inflammatory back pain. Adverse effects like gastrointestinal toxicity and
nephrotoxicity remain concerns however. With the introduction of selective
cycloxygenase II inhibitors (COXIBs), the concern of serious GI side effects has
decreased [82,83]. Continuous use of NSAIDs help in abiding to a good exercise
program and indirectly improves the disease outcome. Recent data on the reduction of
radiological progression with continuous use instead of as-needed use of NSAIDs is also
very provocative [84].
1.6.3 Disease modifying anti-rheumatic drugs (DMARDs)
Sulfasalazine is effective in patients with AS and peripheral arthritis, however, in
pure axial disease, its efficacy is limited [85]. A few studies and a meta-analysis have
shown some benefit in reducing the inflammatory pain in peripheral joints and thus
helping reduce the NSAID requirement [86,87]. Methotrexate (MTX) has been used
with some benefit in patients with AS. There are conflicting reports on the role of MTX
in peripheral as well as axial disease [88-90]. The only RCT of MTX in AS showed no
benefit of AS either on axial or peripheral arthritis [91,92]. Leflunomide was not useful
in spinal symptoms in an open label and a double blind RCT in AS [93,94].
1.6.4 Biological Therapy
Biological therapy in AS primarily involves blocking tumor necrosis factor α (TNFα).
Currently there are 4 FDA-approved agents for blocking TNF: infliximab, etanercept,
adalimumab and golimumab. These agents have proved to be effective for controlling
15
inflammation in AS. The overall efficacy of all anti-TNF agents is comparable, with a
50% reduction in BASDAI in 50-80% of patients.
Infliximab (cA2, Remicade, Fa Essex/Centocor) is a chimeric monoclonal IgG1
antibody formed from human constant and murine variable regions. It binds to both
soluble (sTNF) and membrane bound forms (mTNF) of TNF, neutralizing its action.
There are conflicting reports on induction of apoptosis of T cells and macrophages after
infliximab binds to the membrane bound receptor [95-97]. Recently it was also shown
to induce regulatory macrophages [98]. It is most commonly used at a dose of 5 mg/kg
per dose on 0, 2, 6 and 14 weeks and every 8 weeks thereafter. Significant reduction in
disease activity, improvement in productivity and quality of life have been reported with
infliximab.
Etanercept (Enbrel®, Fa Lederle/Wyeth) is a recombinant human soluble TNF
receptor. It is formed by the fusion of the extracellular portion of the p75 component of
the TNF receptor and the Fc portion of IgG1. Unlike infliximab, it affects sTNF
predominantly with minimal effect on mTNF. Moreover, it does not cause complement
fixation and does not induce lysis of TNF-bearing cells such as monocytes and
macrophages. It does not decrease T-cell production of TNF [99]. The standard dose
of etanercept has been 25 mg twice weekly, but 50 mg once weekly dose was found to
be at least as effective as the twice weekly dose [100].
Adalimumab (HUMIRA®, Abbott) is a fully humanized monoclonal antibody against
TNF [101]. Using phage display technology TNFα-specific human heavy and light chain
variable regions were derived and fused to the Fc portion of human IgG1. Owing to the
16
exclusive human sequences, the chances of immunogenicity and antibody formation
are less. Golimumab is an IgG1, fully human monoclonal anti-TNF antibody
administered as subcutaneous injections once a month. It is administered at a dose of
50 mg once monthly [102].
1.7 Pathogenesis
The exact pathogenesis of ankylosing spondylitis has not been resolved. The
role of genetics is clear from the high concordance of AS in twins [103,104]. However,
even in monozygotic twins the concordance is only 40-50%[104]. Thus there are
environmental influences in the onset of AS.
1.7.1 Infections and Pathogenesis of AS
Infections are known to trigger reactive arthritis. However a link between
infections and AS is not very clear. There are several theories supporting a place for
bacterial infections in the pathogenesis of AS but the exact role is not understood.
Attempts at identifying bacterial DNA from Si joints using a nested PCR technique were
unsuccessful [105].
Among the numerous bacterial infections that have been suggested to play a
role in the pathogenesis of AS, Klebsiella pneumoniae is the most strongly implicated.
There is sequence similarity of the nitrogenase and pullulanase D proteins in Klebsiella
with HLA-B27 [106]. Thus molecular mimicry leads to cross reacting antibodies that
develop following a Klebsiella infection. Cross-reactivity with HLA-B27 can lead to
abnormal immune responses. Both CD4 and CD8 T cell responses have been shown in
17
AS patients against the 60 kD heat shock protein (HSP60) of Klebsiella [107]. Humoral
responses against HSP60 as well as Klebsiella pneumoniae nitrogenase have been
demonstrated [108,109]. Antigen processing of these Klebsiella proteins followed by
presentation on HLA-B27 is likely as chloroquine and Brefeldin A block the cellular
response to the antigens [107]. This is further strengthened by the observation that
HLA-B27 positive individuals have more humoral response to these antigens compared
to HLA-B27 negative individuals [109]. The levels of IgA and IgG antibodies against
the Klebsiella proteins have been shown to correlate with C-reactive protein (CRP), a
marker of inflammation [108]. There is also a suggestion that abnormal persistence of
the bacteria is primarily responsible rather than the immune response. By Elispot assay
it was shown that twins who have AS have reduced reactivity against K pneumoniae, S
pyogenes and C albicans compared to the healthy monozygotic twins, even though the
humoral response is similar [110]. The fact that HLA-B27 transgenic rats are normal
when raised in a germfree environment but develop a spondyloarthritis-like illness
otherwise, suggests that infection or some other second signal like Toll Like Receptor
(TLR) activation may be important in pathogenesis [111].
The above-mentioned evidence is quite compelling, yet a role for Klebsiella in
the pathogenesis of AS has not been established due to the equally prevalent negative
reports. In an extensive study looking at 4 candidate organisms- K. pneumoniae,
Salmonella typhimurium, Yersinia enterocolitica and Chlamydia trachomatis, there was
no difference in the cellular or humoral response between AS patients and their family
members [112]. There is also evidence to show that Klebsiella isolation from the gut
does not differ in AS and any other rheumatic disease [113]. The same study did not
18
find any association of the antibody levels with inflammatory markers[113]. Invoking
molecular mimicry would imply an element of autoimmunity in AS but there is no
evidence of an autoimmune process in the vast majority of these patients.
1.7.2 Cytokines in AS
Tumor necrosis factor (TNF) is a pleiotropic cytokine secreted mainly by
monocytes and macrophages. It is involved in numerous biological processes including
inflammation, apoptosis, septic shock, immunity, arthritis and other autoimmune
diseases. It stimulates prostaglandin production by fibroblasts and induces adhesion
molecules, chemokines and protease enzymes [114-116]. Acting through two
receptors, TNFR1 (p55) and TNFR2 (p75), followed by a complex interplay of
intracellular molecules, TNF activates C-Jun N-terminal kinase (JNK) and IκB Kinase
leading to activation of activated protein 1 (AP-1) and nuclear factor-κB (NF-κB)
respectively which are important transcription factors mediating inflammation
[117,118].
In the early 90s, serum TNF levels were reported to be high in AS and
subsequently TNF-producing mononuclear cells were demonstrated in the sacroiliac
joints [119,120]. It is also expressed in high concentration in the gut of both ulcerative
colitis and Crohn’s disease patients [121]. Up to 60% of patients with AS have
pathological changes in the GI tract and the lesions are strikingly similar to those in
Crohn’s disease [51,52,55,122]. TNF blockers lead to significant improvement in
inflammatory bowel disease and reduce spinal inflammation. Sensitive techniques like
MRI have shown a decrease in inflammation following the use of anti-TNF agents [64].
19
Among the several polymorphisms of the TNFα gene, those at positions -238 and -
308 are well studied [123]. These polymorphisms lead to nucleotide changes in the
putative transcription regulatory region and could possibly alter TNFα production. TNF
transgenic mice with a 3'-rnodified human TNF gene have overexpression of TNF.
These mice develop sacroiliitis like AS in addition to polyarthritis [124].
A metanalysis of 8 separate studies showed no association of AS with TNF
promoter polymorphisms [125]. This negative association was replicated by another
metanalysis of 14 studies comprising 1,766 AS patients and 2,114 controls [126]. The
recent genome-wide scans have suggested an association with TNFRSF1A [127,128].
It is already known that rheumatic disease patients with a TNFα -308 G/G genotype
respond better to anti-TNF treatment than those with A/A or A/G genotypes
independent of the diagnosis (rheumatoid arthritis, psoriatic arthritis or AS) [129]. Why
a TNF gene promoter polymorphism would affect anti-TNF response is not known.
Interestingly, the TNFα -308 G/G genotype has been shown to be associated with
decreased TNFα production [130,131]. It is possible that patients with a tendency for
lower production of TNFα respond even better with anti-TNF agents.
Other cytokines implicated in the pathogenesis of AS are IL1 and IL23. There is
increasing evidence of the involvement of the Th17 pathway in AS pathogenesis. In a
study on PBMC from AS and RA patients, IL-17 and IL-22 positive CD4 T cells were
increased with higher secretion of IL-17 compared with healthy controls [132]. The
frequency of IL-17 positive cells was higher in the facet joints of AS patients compared
to osteoarthritis patients [133]. There is exciting new evidence which shows that HLA-
20
B27 dimer expressing Antigen Presenting Cells (APC) can stimulate the proliferation of
KIR3DL2 expressing CD4 T cells. This subset of T cells is enriched for IL-17 production
[134]. The unfolded protein response (UPR) resulting from unfolding of HLA-B27 in
HLA-B27 transgenic rats induces the production of IL-23 a th17 cytokine [135].
A study of serum levels of IL-17, IL-23, IL-6, IL-12 and TGFβ did not differentiate
active and inactive AS patients [136]. The serum levels of these cytokines were,
however, higher than healthy controls. IL23R polymorphisms are associated with AS
and this has been well replicated in genome wide scans [127,128,137]. A common
association with Crohn’s disease again raises the issue of common pathogenic
mechanisms in IBD and AS. The level of IL-23p19 expression in the cecum and distal
ileum of AS patients’ is comparable to Crohn’s disease [138].
The IL1 gene cluster contains nine genes homologous to IL1A and IL1B or IL1RN.
The IL1A, IL1B and IL1RN code respectively for IL1α, IL1β and IL1Ra. There are
reports of association of AS with IL1RN and other IL1 gene cluster members [139-
142]. These findings have been corroborated by a meta-analysis [143]. The recent
genome wide scans have shown evidence of association of AS with the IL1RII [127].
In summary, cytokines are important in AS. The remarkable efficacy of anti-TNF
agents in controlling inflammation but inability to control progression points towards
an uncoupling of inflammation and new bone formation. It is not entirely clear if the
cytokine defects are the primary event or secondary to other pathogenic events.
Emerging evidence on the induction of Th17 cytokines with UPR and the KIR3DL2-B27
21
dimer interaction further shows the potential involvement of cytokines in AS
pathogenesis.
1.7.3 Genetics of ankylosing spondylitis
1.7.3.1 Lessons from Family and Twin Studies
While the pathogenesis of AS is not well understood, a strong genetic basis for
the disease is clear from the results of family and twin studies. If the risk of a disease
is higher among family members compared to the general population the possibility of
a genetic risk factor should be considered. This suspicion is strengthened if the risk
increases with the degree of familial relationship. In a disease with strong genetic risk,
the highest concordance rate will be in identical twins followed by first degree
relatives. The sibling recurrence risk (s) is calculated as the ratio of the risk of disease
developing in siblings compared to the risk in the general population.
The s in AS has been reported to be as high as 82 [144]. Thus the recurrence
risk of AS in siblings is 8.2% compared to 0.1% in the general population [144].
Similarly in twin studies, the concordance of AS in monozygotic twins is 50% compared
to 25% in dizygotic twins [103]. High rates of concordance have been reported in
other studies from Europe [104,145]. The heritability of AS is estimated to be over
95% thereby bringing the role of environmental factors to a minimum.
1.7.3.2 Role of HLA-B27 in Ankylosing Spondylitis
Tremendous hope for an imminent resolution of the pathogenesis of AS was
raised when HLA-B27 was found to be present in 96% of patients with AS, compared
22
with 4% of controls [146]. There were suspicions initially whether HLA-B27 itself was
pathogenic rather than a locus in strong linkage with the pathogenic gene. These
concerns were dispelled when the HLA-B27 transgenic rats showed features quite
similar to human spondyloarthritis [147]. Although the exact role of HLA-B27 in the
pathogenesis of AS is not clear, several hypotheses have been put forward (Table
1.3). Many of these themes will be alluded to again in the discussion on HLA-B27
transgenic rats.
In the ‘arthritogenic peptide hypothesis’ HLA-B27 in its classical role as an
MHC-I molecule presents special peptides to CD8+ T cells resulting in an arthritogenic
immune response. The presentation of distinctive peptides could be defined by the
unique peptide binding specificity of HLA-B27, which eventually results in a T-cell-
cascade, thereby leading to joint inflammation [148,149]. In a recent mass
spectrometry (MS) study, 1,268 HLA–B27 peptides derived from human proteins were
identified from HLA-B*2705 expressing C1R cells [150]. Out of these peptides 569
were identified as certain based on at least 3 to 4 MS verifications. Another 582
peptides were considered probable with 1 or 2 MS verifications. Most of the identified
peptides had 8–10 amino acids. Around 211 peptides were extended with 11–15
amino acids. Around 94 peptides were present in varying lengths and the variation was
most commonly seen at the C- terminal end. Interestingly several peptides were
derived from cartilage or bone related proteins. There were peptides with similarity to
sequences from several bacteria including Salmonella, Shigella, Campylobader and
Yersinia.
23
Table 1.3 Hypothesis of the role of HLA-B27 in the pathogenesis of Ankylosing
Spondylitis
Hypothesis Summary
Arthritogenic Peptide Hypothesis
HLA-B27 binds and presents specific peptides that can
lead to arthritis
Unfolded Protein Response
HLA-B27 has a propensity to misfold and accumulate in
the ER leading to an ER stress response and release of
pro-inflammatory cytokines
HLA-B27 dimer mediated immune response
HLA-B27 heavy chains can form dimers that can be
recognized by atypical immune receptors like KIRs and
LIRs. This can lead to abnormal immune responses
Molecular Mimicry
HLA-B27 and some bacteria share peptide sequence
similarities that can lead to molecular mimicry and
abnormal immune response to HLA-B27
Thymic Selection of T cell repertoire
The presence of HLA-B27 can affect the selection of the
T-cell repertoire and this may influence the development
of T cells that are more ‘arthritogenic’
HLA-B27 as a source of peptides
HLA-B27 itself can be degraded and form peptides that
can lead to inflammatory arthritis
ER: Endoplasmic Reticulum; KIR: Killer Cell Immunoglobulin-like Receptors; LIR:
Leukocyte Immunoglobulin-Like Receptors
24
As Jose Lopez de Castro elegantly wrote in an accompanying editorial, peptide
searches alone will not help us solve the mystery of HLA-B27 [151]. Nevertheless,
these efforts help us better understand the HLA-B27 binding motif and its variations.
We may be able to identify the target proteins from which the HLA-B27 bound peptides
are derived. Despite all the advances in technology and sincere efforts, we have no
evidence that a specific peptide bound to HLA-B27 is the cause of AS.
Crystallographic analysis of HLA-B27 revealed a typical heterodimeric structure
of MHC-I, composed of a heavy chain non-covalently linked to beta 2-microglobulin
(β2m) [152]. The heavy chain and β2 microglobulin are encoded in chromosomes 6
and 15 respectively. The peptide binding grove is formed by the α1 and α2 domains of
the heavy chain. Typically a nonamer peptide with a single main chain conformation is
bound to HLA-B27 with anchoring of the N- and C-terminal ends.
The HLA-B27 peptide binding grove has six pockets (A to F) that accommodate
the side chains of the residues of the peptides. The second amino acid is almost
always arginine with the side chain occupying Pocket B. The side chains of residues on
position P1, P4 and P8 are oriented away from the cavity and are important in T cell
responses [153]. There are unique amino acid positions in the structure of HLA-B27
including Arg2, His9, Thr24, Glu45 and Cys67 [154]. Cys67 is particularly interesting as
this amino acid can form disulphide bonds with another Cys67 leading to B27 heavy
chain dimers that can stabilize without β2 microglobulin [155]. The Cys67 residue is a
cardinal part of pocket B [156] and it is also known to affect TCR recognition [157].
HLA-B27 has a propensity to misfold during assembly in the ER, and these
misfolded B27 heavy chains may accumulate in the ER leading to an unfolded protein
25
response (UPR) [158]. The UPR theory brings a different perspective to the
pathogenesis of AS, making it an auto-inflammatory disease. HLA-B*2705 has a slow
folding rate, with misfolding and ER retention [159,160]. The resulting ER stress leads
to the release of pro-inflammatory cytokines [161]. There is a correlation between the
misfolded HC and the degree of UPR [160].
HLA-B27 has a greater tendency for surface expression of homodimers of free
heavy chains than other class I alleles [162-165]. There are opposing studies showing
that HLA-B27 dimers are recognized by KIR receptors in a peptide dependent [166]
and non-dependent manner [167]. It was not known previously that KIRs are highly
polymorphic. With this new knowledge, it is possible that different KIR subtypes could
recognize MHC in a peptide dependent manner.
A recent study by Bowness et al. indicated possible involvement of HLA-B27
dimers in the pathogenesis of AS [134]. The study showed that antigen presenting
cells expressing HLA-B27 dimers stimulate the survival and proliferation of KIR3DL2
expressing CD4+ T cells[134]. These T cells were elevated in number in AS patients
and enriched for IL-17 production. A provocative finding was that 70% of the Th17
producing cells were KIR3DL2 positive, despite being a minority among all PBMC
[134].
1.7.3.3 HLA-B27 subtypes and Ankylosing Spondylitis
There are more than 35 HLA-B27 subtypes described so far of which HLA-
B*2705 is present in almost all populations world-wide and is the most common allele
in Caucasians . HLA-B*2704 is common in Asians and HLA-B*2702 in middle-eastern
26
and North African population. HLA-B'2703 is most commonly seen in western Africans
and differs from HLA-B '2705 by a substitution of tyrosine for histidine at position 59.
The HLA-B27 subtypes most commonly associated with AS are HLA-B*2705,
B*2702, B*2704 and B*2707. Interesting HLA-B*2706 that forms half of all HLA-B27
subtypes in Indonesia, Malaysia and Thailand, has not been reported in a clinically
documented case of AS [168]. Similarly, HLA-B*2709 comprises 20% of all B27 alleles
in Sardinia and yet no single patient of AS has been identified with HLA-B*2709 here.
There are only minor differences between the AS-associated and non-associated
HLA-B27 subtypes. HLA-B*2706 differs from HLA-B*2704 at amino acids 114 and 116.
However, these amino-acids normally occupy the floor of the F pocket, within the
peptide-binding groove of HLA-B27 [168]. Similarly, HLA-B*2709 differs from B*2705
at amino acid position 116 with Asp in B*2705 and His in B*2709 [28]. The peptide
repertoire of the AS-associated and non-associated subtypes differs and this could be
due to the changes in the peptide binding groove [168]. Similarly there are differences
in the way peptides are accommodated in the groove.
HLA-B*2705, in comparison to B*2709, is known to bind at least three different
peptides (examples being derived from Vasoactive Intestinal Polypeptide Receptor
(pVIPR), Latent Membrane Protein 2 of EBV (pLMP2) and Glucagon Receptor (pGR)
respectively) in two different conformations [169-171]. In another study using the
peptide GRFAAAIAK, it was found that the bulkier histidine in position 116 induces a
movement of the peptide C-terminal Lys to form a novel salt bridge with aspartate in
position 77 [172]. His in position 116 in contrast to Asp can adopt two different
conformations and overall, the peptide biding groove of HLA-B*2709 is more flexible
27
than that of B*2705 [172]. The differences mentioned above can induce subtle
differences in the B27-peptide conformation resulting in differential immune responses
in the different HLA-B27 subtypes.
Apart from studies looking at differences in the context of the arthritogenic
peptide hypothesis in AS, the rate of folding of the different subtypes has been
studied. When the subtypes B*2702, B*2704, B*2705, B*2706, B*2707 and B*2709
were compared, B*2702, B*2704 and B*2705 folded much more slowly than the AS-
non-associated B*2706 and B*2709 subtypes [173]. However, B*2707 which is known
to be associated with AS was discordant with a high folding rate comparable to the AS-
neutral subtypes [173]. Simiarly the accumulation of Free Heavy Chain (FHC) in the ER
was discordant in only B*2707. There was increased FHC accumulation in HLA-
B*2702, B*2704 and B*2705 but not in B*2706 and B*2709.
Thus several differences exist between the AS-associated and non-associated
subtypes, but no definitive explanation is yet available as to why certain subtypes
alone increase the risk of AS.
1.7.3.4 Lessons from the HLA-B27 transgenic rats
The HLA-B27 transgenic rats was generated by Joel D. Taurog and reported
first in 1990. These rats were transgenic for human HLA-B27 and human β2-
microglobulin. More than 90% of male and female rats develop disease manifestations
similar to human spondyloarthritis including sacro-iliitis, peripheral arthritis, colitis,
psoriasiform skin lesions and nail changes. Although the incidence in both genders is
the same, male rats develop disease much earlier than females.
28
The level of expression of HLA-B27 correlated with the incidence of disease.
Thus two strains of B27 transgenic rats 33-3 and 21-4H had high copy numbers of
HLA-B27 and developed spontaneous disease. The strains 21-4L and 21-2, 25-1 and
25-6 with low B27 copy numbers did not develop any disease manifestation. The rats
with high B27 copy numbers also had higher β2m expression. Almost two-thirds of
newly formed HLA-B27 molecules are misfolded in the ER and are present as
monomers or dimers [159]. Misfolding of HLA-B27 and stimulation of the unfolded
protein response (UPR) is seen in the HLA-B27 transgenic rats [161]. The UPR
response in HLA-B27 rats was found to be associated with increased IL-23 production
linking the pathogenesis of manifestation in rats to that of human AS and the Th-17
pathway [135].
There is a possibility that disease state itself induced a higher expression of
HLA-B27. However, if inflammation is cause by inducing adjuvant induced arthritis in
early days of the B27 transgenic rats, there is no increase in B27 expression. Similarly,
in rats with both HLA-A2 and HLA-B27, only HLA-B27 expression is higher with disease
onset and HLA-A2 levels remain the same. Thus a generalized increase in MHC
expression as a response to inflammation is unlikely. A threshold effect has been
suggested to explain the requirement of higher B27 expression, such that a certain
level of B27-peptide complexes is required to induce a pathological immune response.
If unfolding of HLA-B27 is the central pathogenic event, the disease should be
preventable by interrupting B27 from unfolding. For this purpose, β2m expression was
increased in the rat but paradoxically the arthritis worsened and colitis improved [174].
This was despite the fact that HLA-B27 misfolding and UPR was less. A similar
29
discordance between arthritis and colitis is also seen with regards to peptide
specificity. A specific influenza peptide when over expressed in HLA-B27 transgenic
rats, led to more arthritis but no effect on colitis [175]. When the two findings above are
seen together there is a possibility that the pathogenesis of arthritis could be peptide
mediated while that of colitis could be due to the UPR.
Although the HLA-B27 transgenic rats appear to provide solid evidence for HLA-
B27 being sufficient in the pathogenesis of AS, complexities soon became evident.
These rats were asymptomatic if bred in a germ free state [111]. Based on the
classical function of MHC-I, HLA-B27 would be expected to present pathogenic
peptides to CD8 T cells. However the HLA B27 transgenic rats continued to have no
amelioration of the disease despite a complete absence of CD8 T cells, in the HLA-B27
CD8-/- rats [176]. This however does not exclude the possibility of a CD8 negative
MHC-I restricted T cells response as has been reported elsewhere [177]. Transfer of
CD4 cells to transgenic nude rats can lead to reconstitution of disease manifestations
of the HLA-B27 transgenic rats [178]. Another layer of confusion was added when β2
microglobulin deposition led to arthritis independent of HLA-B27 heavy chains [179].
1.7.4 Inflammation and New Bone Formation in AS
Even though there is a significant inflammatory process in AS, what
differentiates it from other forms of inflammatory arthritis is new bone formation. The
pathogenesis of new bone formation in AS is still not understood. Although it is
tempting to state that inflammation is followed by new bone formation, the evidence
for this is scanty. Due to the inaccessibility of tissue and the long duration of follow up
30
required to answer some of these compelling questions, there is still no definite
answer regarding the relationship of inflammation to bone formation. With the
introduction of highly sensitive imaging techniques like MRI it is possible to identify
inflammation non-invasively and then follow up these patients for the development of
syndesmophytes. There are at least 2 such studies done on patients treated with anti-
TNF agents that have given us some clues.
In the first study on 39 AS patients followed up for 2 years, out of 922 vertebral
edges analyzed, syndesmophytes developed in 26 [180,181]. Inflammation was
identified in 153 corners at baseline and significantly more syndesmophytes developed
in these corners than in those without inflammation (6.5% vs 2.%; p=0.002). This
would seem to be a clear indication that inflammation precedes syndesmophyte
formation. However, when we consider all syndesmophytes at follow up and look at
baseline, there were more syndesmophytes in corners which had no inflammation at
baseline compared to those with (62% vs 38%; P = 0.03). There was no clear cut
explanation for these data, but based on the evidence that corners with inflammation
prospectively led to new bone formation, Baraliakos et al. concluded that suppression
of inflammation could help prevent syndesmophyte formation.
MRI is very sensitive in picking up inflammation, but the sensitivity is not 100%.
Spondyloarthritis is well known to have a fluctuating course and an MRI evaluation at
one point may not reflect inflammation that was present in a corner on some other
time. Moreover, there are issues of inter-reader reliability. Hence it is possible that the
31
corners with syndesmophyte formation had at some point of time inflammation, but
this has not been proven.
In the second study on AS patients, syndesmophytes developed more
frequently in vertebral corners with baseline inflammation than in those without (20%
vs 5.1%)[181]. Contrary to the proposal of the previous study, syndesmophytes
developed more frequently at the corners in which inflammation resolved (14.3%)
than at those where inflammation persisted (2.9%). This finding has led the lead
author Walter Maksymowych to propose a ‘TNF-brake hypothesis’ according to which
inflammation inhibits new bone formation.
1.7.5 DKK-1 in the Pathogenesis of Ankylosing Spondylitis
The term Wnt is derived from a combination of Wg (wingless) and Int1
(Integration-1), two Wnt genes discovered initially. Mutations of a gene in Drosophila
were found in wingless flies, and it was called ‘wingless’ originally[182]. Similarly Int1
(integration 1) was activated in mice with mouse mammary tumor virus (MMTV)
induced breast tumors[183]. From the early studies, the importance of Wnt pathway in
limb development, antero-posterior axial patterning and eye formation was
established[182,184,185]. It is now known that the Wnt signaling pathway is critical in
the regulation of cell growth, differentiation and death[182,184-186].
The canonical Wnt-signalling pathway involves signaling through the surface receptors
LDL-receptor like protein (LRP) -5 and 6 along with Frizzled. In the inactive state
(Figure 1.1 A), β-catenin, the key protein that mediates the Wnt signaling pathway is
trapped in a complex and programmed for ubiquitination and degradation. When the
32
pathway is activated (Figure 1.1 B) by Wnt binding to LRP5/6 and Frizzled,
Dishevelled is activated which inhibits Glycogen Synthase Kinase-3β (GSK-3β). Axin
moves out of the complex and can secondarily attach to LRP5/6. As Axin moves away
from the complex the ability of GSK-3β to phosphorylate β-catenin decreases. β-
catenin accumulates and is transported into the nucleus where it binds with the
transcription factor T cell Factor 4 (TCF). This leads to the synthesis of chondrogenic
and osteoblastic factors.
The canonical Wnt-signaling pathway is now well recognized to play a role in
bone homeostasis[184]. Increased β-catenin is seen in committed osteoblast lineage
cells. A loss of β-catenin in osteoblast precursor cells results in reduced bone
deposition[187]. The relevance of Wnt has been noted in conditions like osteoporosis,
RA, AS, osteoarthritis, and Paget’s disease [188-195]
It is puzzling that bone resorption is the phenotype in RA while bone formation
is the hallmark of AS, even though the same cytokines appear to mediate both
diseases. This riddle seemed to be solved with the identification of divergent DKK-1
levels in the two conditions[188]. DKK-1 is a circulating inhibitor of the Wnt-signaling
pathway. It binds to the same receptors as Wnt and prevents the signalling. Thus in
AS, where there is more bone formation, DKK-1 levels are low, resulting in excess
osteoblastic factors forming. This has been replicated in Diffuse Idiopathic Skeletal
Hyperostosis (DISH), another condition characterised by excess bone formation in the
spine [196].
33
LRP: LDL-receptor Like Protein; GSK-3β: Glycogen Synthase Kinase-3β; β-Cat: β
Catenin; Dsh: Dishevelled; TCF: Trascription FactorT cell Factor 4
Figure 1.1 The Canonical Wnt-β Catenin Pathway
LRP5/6 Frizzled
34
Legend to Figure 1.1
The Canonical Wnt-β Catenin Pathway
(A) In the inactive state, β-catenin is trapped in a complex and programmed for
ubiquitination and degradation.
(B) When the pathway is activated by Wnt binding to LRP5/6 and Frizzled, Dishevelled
is activated which inhibits Glycogen Synthase Kinase-3β (GSK-3β). Axin moves out of
the complex and can secondarily attach to LRP5/6. As Axin moves away from the
complex the ability of GSK-3β to phosphorylate β-catenin decreases. β-catenin
accumulates and is transported into the nucleus where it binds to the transcription
factor T cell Factor 4 (TCF).
35
On the other hand, in RA, the DKK-1 level is higher with the balance tilting in
favour of bone resorption. Even lupus patients with erosive arthritis have elevated
DKK-1 levels[197]. DKK-1 has been shown to be low in another AS cohort [192] and
when tested in a Jurkat T-cell model, was found to be dysfunctional in AS patients at
activating the Wnt pathway [198]. Thus not only is DKK-1 lower in AS, it is also less
effective in activating the Wnt signalling pathway.
The TNF-transgenic mouse has inflammation affecting the sacroiliac joints with
no spontaneous fusion. TNF blockade reduced inflammation, erosions and osteoclasts
in the sacroiliac joints [199]. On the other hand, DKK1 blockade reduced erosions and
promoted ankylosis of the sacroiliac joints [199]. This further shows the relevance of
the Wnt-signalling pathway in AS.
1.8 Endoplasmic Reticulum Aminopeptidase 1
Ever since it was realized that HLA-B27 was only one actor in a large dramatis
personae, the hunt for other candidate genes started. It was recognized that AS is
likely to be an oligogenic disease [144] and several attempts have been made to look
for non-major histocompatibility complex (MHC) genes that might contribute to
increased susceptibility to the disease [200]. A genome wide association study
(GWAS) of nonsynonymous single nucleotide polymorphisms (SNPs) identified a novel
association of AS with ERAP1 [137].
The name ERAP1 is approved by the Human Genome Organization (HUGO)
Nomenclature Committee. ERAP1 is known by various aliases, such as endoplasmic
reticulum aminopeptidase associated with antigen processing (ERAAP), adipocyte-
36
derived leucine aminopeptidase (A-LAP), puromycin-insensitive leucyl aminopeptidase
(PILS-AP) and aminopeptidase regulating tumor necrosis factor receptor I (TNFRI)
shedding (ARTS-1). ERAP1 consists of 930 amino acids and was identified in mice
following solubilization and fractionation of microsomes derived from liver and spleen
[201]. At approximately the same time, human ERAP1, which has 941 amino acids and
86% sequence similarity to murine ERAP1 (ERAAP), was isolated from HeLa S cells and
found to be the same enzyme as A-LAP and PILS-AP [202]. The human ERAP1 gene
spans 54.61 kb of chromosome 5 from 96149849 to 96095244, on the reverse strand.
The ERAP1 gene lies between the ERAP2 and Calpastatin (CAST) genes (Figure 1.2).
The human ERAP1 mRNA has 2826 nucleotides (Figure 1.3) and the protein has 941
amino acids (Figure 1.4). The two SNPs that have been reported to be strongly
associated with ERAP1 are rs30187 and rs27044 that lead to R528K and E730Q
residue changes respectively (Figures 1.3, 1.4 and Table 1.4).
The discovery of the ERAP1 association has brought new insights and renewed
excitement into the field of AS research. After HLA-B27, which has an attributable risk
of 50%, ERAP1 has the second strongest association, with an attributable risk of 26%
[137]. Thus after 35 years of research since discovering the B27-AS genetic
association, for the first time a strong and replicated genetic association in AS has
been identified. The known functions of ERAP1 have further triggered the expectations
and imagination of the AS research community.
ERAP1 has two known functions. First, it aids the shedding of the membrane-
bound cytokine receptors TNFRI, IL-1 receptor II (IL-1RII) and IL-6Rα. In fact ERAP1
37
was independently discovered as a molecule that bound cytokine receptors while
searching for receptor shedding enzymes. This function of ERAP1 will be further
described and the implications to AS pathogenesis discussed later.
The second known function of ERAP1 is the trimming of peptides within the ER
for loading onto MHC-I molecules. ERAP1 has ERAP2, a structurally functionally related
ER aminopeptidase, as a close ally in this function. Delineating the functional aspects
of the AS-associated ERAP1 and ERAP2 polymorphisms has become a major focus of
research. This novel connection could be the long sought bridge between HLA-B27 and
AS.
1.8.1 Structure of ERAP1
ERAP1 and ERAP2 belong to the M1 family of zinc metallopeptidase enzymes,
which share a HEXXH(X)18E zinc-binding motif [203]. Over the last 3 years there was
a rigorous attempt to obtain a 3-dimensional model of ERAP1. Crystallographic
structures can aid in predicting the possible functional effects of genetic
polymorphisms on molecules.
The first attempt was to develop a molecular homology model of ERAP1 based
on the crystallographic structure of a closely related protein, tricorn-interacting factor
F3 (TIFF3), with which ERAP1 shares 32% amino acid sequence identity (residues 63–
665) [204]. The sequence similarity increases to 82% when considering the active
catalytic site (residues 280–486) of ERAP1, in particular the zinc-binding domain.
Baased on the crystallographic structure of TIFF3 sequence alignment was performed
and a homology model of ERAP1 was generated.
38
Figure 1.2 The Chormosome 5q15 locus with ERAP1 and other surrounding
genes
39
Legend to Figure 1.2
Chromosome 5 with ERAP1, ERAP2, LNPEP and CAST genes
The human ERAP1 gene spans 54.61 kb of chromosome 5 from from 96149849 to
96095244, on the reverse strand (note direction of arrow). The ERAP1 gene lies
between the ERAP2 and Calpastatin (CAST) genes.
40
Figure.
ERAP1 mRNA Nucleotide Sequence (2826 nt): ATGGTGTTTCTGCCCCTCAAATGGTCCCTTGCAACCATGTCATTTCTACTTTCCTCACTGTTGGCTCTCT
TAACTGTGTCCACTCCTTCATGGTGTCAGAGCACTGAAGCATCTCCAAAACGTAGTGATGGGACACCATT TCCTTGGAATAAAATACGACTTCCTGAGTACGTCATCCCAGTTCATTATGATCTCTTGATCCATGCAAAC CTTACCACGCTGACCTTCTGGGGAACCACGAAAGTAGAAATCACAGCCAGTCAGCCCACCAGCACCATCA TCCTGCATAGTCACCACCTGCAGATATCTAGGGCCACCCTCAGGAAGGGAGCTGGAGAGAGGCTATCGGA AGAACCCCTGCAGGTCCTGGAACACCCCCGTCAGGAGCAAATTGCACTGCTGGCTCCCGAGCCCCTCCTT GTCGGGCTCCCGTACACAGTTGTCATTCACTATGCTGGCAATCTTTCGGAGACTTTCCACGGATTTTACA AAAGCACCTACAGAACCAAGGAAGGGGAACTGAGGATACTAGCATCAACACAATTTGAACCCACTGCAGC TAGAATGGCCTTTCCCTGCTTTGATGAACCTGCCTTCAAAGCAAGTTTCTCAATCAAAATTAGAAGAGAG CCAAGGCACCTAGCCATCTCCAATATGCCATTGGTGAAATCTGTGACTGTTGCTGAAGGACTCATAGAAG ACCATTTTGATGTCACTGTGAAGATGAGCACCTATCTGGTGGCCTTCATCATTTCAGATTTTGAGTCTGT CAGCAAGATAACCAAGAGTGGAGTCAAGGTTTCTGTTTATGCTGTGCCAGACAAGATAAATCAAGCAGAT TATGCACTGGATGCTGCGGTGACTCTTCTAGAATTTTATGAGGATTATTTCAGCATACCGTATCCCCTAC CCAAACAAGATCTTGCTGCTATTCCCGACTTTCAGTCTGGTGCTATGGAAAACTGGGGACTGACAACATA TAGAGAATCTGCTCTGTTGTTTGATGCAGAAAAGTCTTCTGCATCAAGTAAGCTTGGCATCACAATGACT GTGGCCCATGAACTGGCTCACCAGTGGTTTGGGAACCTGGTCACTATGGAATGGTGGAATGATCTTTGGC TAAATGAAGGATTTGCCAAATTTATGGAGTTTGTGTCTGTCAGTGTGACCCATCCTGAACTGAAAGTTGG AGATTATTTCTTTGGCAAATGTTTTGACGCAATGGAGGTAGATGCTTTAAATTCCTCACACCCTGTGTCT ACACCTGTGGAAAATCCTGCTCAGATCCGGGAGATGTTTGATGATGTTTCTTATGATAAGGGAGCTTGTA TTCTGAATATGCTAAGGGAGTATCTTAGTGCTGACGCATTTAAAAGTGGTATTGTACAGTATCTCCAGAA GCATAGCTATAAAAATACAAAAAACGAGGACCTGTGGGATAGTATGGCAAGTATTTGCCCTACAGATGGT GTAAAAGGGATGGATGGCTTTTGCTCTAGAAGTCAACATTCATCTTCATCCTCACATTGGCATCAGGAAG GGGTGGATGTGAAAACCATGATGAACACTTGGACACTGCAGAAGGGTTTTCCCCTAATAACCATCACAGT
GAGGGGGAGGAATGTACACATGAAGCAAGAGCACTACATGAAGGGCTCTGACGGCGCCCCGGACACTGGG TACCTGTGGCATGTTCCATTGACATTCATCACCAGCAAATCCGACATGGTCCATCGATTTTTGCTAAAAA CAAAAACAGATGTGCTCATCCTCCCAGAAGAGGTGGAATGGATCAAATTTAATGTGGGCATGAATGGCTA TTACATTGTGCATTACGAGGATGATGGATGGGACTCTTTGACTGGCCTTTTAAAAGGAACACACACAGCA GTCAGCAGTAATGATCGGGCGAGTCTCATTAACAATGCATTTCAGCTCGTCAGCATTGGGAAGCTGTCCA TTGAAAAGGCCTTGGATTTATCCCTGTACTTGAAACATGAAACTGAAATTATGCCCGTGTTTCAAGGTTT GAATGAGCTGATTCCTATGTATAAGTTAATGGAGAAAAGAGATATGAATGAAGTGGAAACTCAATTCAAG GCCTTCCTCATCAGGCTGCTAAGGGACCTCATTGATAAGCAGACATGGACAGACGAGGGCTCAGTCTCAG AGCGAATGCTGCGGAGTCAACTACTACTCCTCGCCTGTGTGCACAACTATCAGCCGTGCGTACAGAGGGC
AGAAGGCTATTTCAGAAAGTGGAAGGAATCCAATGGAAACTTGAGCCTGCCTGTCGACGTGACCTTGGCA GTGTTTGCTGTGGGGGCCCAGAGCACAGAAGGCTGGGATTTTCTTTATAGTAAATATCAGTTTTCTTTGT CCAGTACTGAGAAAAGCCAAATTGAATTTGCCCTCTGCAGAACCCAAAATAAGGAAAAGCTTCAATGGCT ACTAGATGAAAGCTTTAAGGGAGATAAAATAAAAACTCAGGAGTTTCCACAAATTCTTACACTCATTGGC AGGAACCCAGTAGGATACCCACTGGCCTGGCAATTTCTGAGGAAAAACTGGAACAAACTTGTACAAAAGT TTGAACTTGGCTCATCTTCCATAGCCCACATGGTAATGGGTACAACAAATCAATTCTCCACAAGAACACG GCTTGAAGAGGTAAAAGGATTCTTCAGCTCTTTGAAAGAAAATGGTTCTCAGCTCCGTTGTGTCCAACAG ACAATTGAAACCATTGAAGAAAACATCGGTTGGATGGATAAGAATTTTGATAAAATCAGAGTGTGGCTGC AAAGTGAAAAGCTTGAACGTATGTAA
Nucleotides in Blue are exonic. The nucleotide change associated with rs30187 and rs27044 SNPs are highlighted in yellow and green respectively. Thus the two major polymorphisms associated with ankylosing spondylitis are exonic.
Figure 1.3 Nucleotide Change in the ERAP1 Polymorphisms Studied
41
ERAP1 Translation (941 aa): MVFLPLKWSLATMSFLLSSLLALLTVSTPSWCQSTEASPKRSDGTPFPWNKIRLPEYVIPVHYDLLIHAN LTTLTFWGTTKVEITASQPTSTIILHSHHLQISRATLRKGAGERLSEEPLQVLEHPRQEQIALLAPEPLL VGLPYTVVIHYAGNLSETFHGFYKSTYRTKEGELRILASTQFEPTAARMAFPCFDEPAFKASFSIKIRRE PRHLAISNMPLVKSVTVAEGLIEDHFDVTVKMSTYLVAFIISDFESVSKITKSGVKVSVYAVPDKINQAD YALDAAVTLLEFYEDYFSIPYPLPKQDLAAIPDFQSGAMENWGLTTYRESALLFDAEKSSASSKLGITMT VAHELAHQWFGNLVTMEWWNDLWLNEGFAKFMEFVSVSVTHPELKVGDYFFGKCFDAMEVDALNSSHPVS TPVENPAQIREMFDDVSYDKGACILNMLREYLSADAFKSGIVQYLQKHSYKNTKNEDLWDSMASICPTDG VKGMDGFCSRSQHSSSSSHWHQEGVDVKTMMNTWTLQKGFPLITITVRGRNVHMKQEHYMKGSDGAPDTG YLWHVPLTFITSKSDMVHRFLLKTKTDVLILPEEVEWIKFNVGMNGYYIVHYEDDGWDSLTGLLKGTHTA VSSNDRASLINNAFQLVSIGKLSIEKALDLSLYLKHETEIMPVFQGLNELIPMYKLMEKRDMNEVETQFK AFLIRLLRDLIDKQTWTDEGSVSERMLRSQLLLLACVHNYQPCVQRAEGYFRKWKESNGNLSLPVDVTLA VFAVGAQSTEGWDFLYSKYQFSLSSTEKSQIEFALCRTQNKEKLQWLLDESFKGDKIKTQEFPQILTLIG RNPVGYPLAWQFLRKNWNKLVQKFELGSSSIAHMVMGTTNQFSTRTRLEEVKGFFSSLKENGSQLRCVQQ TIETIEENIGWMDKNFDKIRVWLQSEKLERM
The residue changes resulting from the rs30187 and rs27044 polymorphisms are highlighted in yellow and green respectively. Thus the two major polymorphisms of ERAP1 associated with AS lead to change in protein structure.
Table 1.4 Single Nucleotide Polymoprhisms of ERAP1 with Nucleotide and
Residue Change
SNP mRNA Position Allele Change Protein Position Residue Change
rs30187 1930 AAG - AGG 528 K (Lys) – R (Arg)
rs27044 2535 CAA - GAA 730 Q (Gln) – E (Glu)
Figure 1.4 Residue Change in the ERAP1 Polymorphisms Studied
42
The model of ERAP1 contains a large cavity that could accommodate 9-mer peptides.
Based on this structure, none of the SNPs associated with AS are near the catalytic site
of the enzyme.
There is still an explanation as to how these SNPs can affect the function of
ERAP1. ERAP1 has a preference for leucine at the N-terminal end and in fact was
characterized first as a Leucine Aminopeptidase [205]. Other hydrophobic residues like
methionine, phenylalanine and alanine, when present at the N-terminal end still results
in reasonably fast peptide trimming. The enzymatic activity decreases when hydrophilic
residues like arginine, lysine, glutamic acid or threonine are present [204]. However,
this change in enzymatic activity has been shown with variations in in the C-terminal
end with a preference for hydrophobic residues [206]. It is now known that changes in
the internal sequence of substrate peptides—and not just in the N- or C-terminal end—
affect the enzymatic activity of ERAP1 [204]. Thus for internal sequences to affect the
enzymatic activity, a close interaction between the entire peptide substrate and ERAP1
has to be considered. This close association is required for efficient enzymatic activity.
Polymorphisms at sites remote from the catalytic sites, and closer to the substrate
binding areas, thus can affect this close association.
Two groups have now independently solved the crystal structure of ERAP1.
Varying conformations have been identified and clearly ERAP1 appears to take an
‘open’ (Figure 1.5) and ‘closed’ (Figure 1.6) conformation [207]. The final crystal
structure is quite similar to the original computed model with 4 domains. The N-
terminal end is formed by the first domain extending up to residue 254. This domain is
43
Figure 1.5 ERAP1 Crystal Structure: Open Configuration
44
Legend to Figure 1.5
ERAP1 Crystal Structure
Open configuration
The N-terminal end is formed by the saddle shaped first domain extending up to
residue 254, which forms the binding site for the N-terminal end of the peptide. Domain
II (residues 255–527), like all aminopeptidases, houses the catalytic domain, the
GAMEN and the HEXXHX18E motifs. Domain II is connected to the C-terminal Domain
IV by a bridging domain III (resides 528-613). In the closed state of ERAP1, domain IV
arches over domain II forming a large central cavity enclosing the catalytic site. The
zinc ion in the catalytic domain as well as the positions of the ERAP1 SNPs rs30187
(K528R) and rs27044 (Q730E) are depicted here. The surface carbohydrate moiety is
composed of Mannose and N-Acetyl Glucosamine.
Quite clearly, from the open structure, these polymorphisms are away from the
catalytic site. The figure of the crystal structure of ERAP1 [207] was generated using
the Jmol software [208].
45
Figure 1.6 ERAP1 Crystal Structure: Closed Configuration
46
Legend to Figure 1.6
The closed conformation of ERAP1
In the closed conformation of ERAP1, domain IV closes the catalytic site and forms a
large cavity that accommodates large peptides. It is evident that the K528R variation
affects the outer surface of domain II and is not near the catalytic site. The K528R
variation however could affect the conformational change from open to closed state by
altering the interaction with surrounding residues.
The Q730E variation cannot be seen here as it is located within the cavity in the closed
state. Thus there is a potential for interaction with peptides in the closed state. The
figure of the crystal structure of ERAP1 [207] was generated using the Jmol software
[208].
47
saddle shaped and consists of eight β-sheets and forms a binding site for the N-
terminal end of the peptide. Domain II (residues 255–527), like all aminopeptidases,
houses the catalytic domain, the GAMEN and the HEXXHX18E motifs [207,209].
In contrast to some other aminopeptidases, domain II is connected to the C-
terminal Domain IV by a bridging domain III (resides 528-613). As domain IV forms a
cover over the catalytic site on domain II, the presence of domain III increases the
size of the cavity. This could be the reason why ERAP1 can accommodate longer
peptides for processing. Domain IV (residues 614–941) is made of 16 alpha-helices
arranged like a cup. In the closed state of ERAP1, domain IV arches over domain II
forming a large central cavity enclosing the catalytic site [207,209].
1.8.2 Genetic association of ERAP1 with AS
The ERAP1-AS association is well established now and has been replicated in
multiple cohorts and ethnicities. The study by the Wellcome Trust Case Control
Consortium and the Australo-Anglo-American Spondylitis Consortium (WTCCC-TASC
study) was the first to establish this link. Single Nucleotide Polymorphisms (SNPs) are
changes in the DNA sequence by one nucleotide and the lowest allele frequency at
that particular locus in the population is called minor allele frequency (MAF). When the
frequency of this allele is less than 1% it is called a mutation and not as SNP. By
looking for overexpression of a particular allele, usually by comparing the MAF, in the
disease cohort vs. controls, possibly pathogenic DNA variations can be identified. The
WTCCC-TASC group used 14,500 nonsynonymous SNPs and found that, in British
Caucasian patients with AS, the minor allele frequency (MAF) of the ERAP1 SNPs
48
rs30187 and rs27044 was considerably higher than in controls [137]. The study also
showed the MAF of the SNPs rs10050860, rs17482078 and rs2287987 was lower than
in controls [137].
In a multicenter study in Canada, which included 992 cases of AS and 1,437
controls, all of Northern European Caucasian ancestry, five genes involved in the
antigen-processing pathway were assessed [210]. ERAP1 was found to be significantly
associated with AS and the haplotype rs27044/10050860/30187-CCT was strongly
associated with an increased risk of AS (odds ratio [OR] 1.8, 95% confidence interval
[CI] 1.5–2.2; P=7x10-8). The ERAP1 haplotype rs30187/26618/26653-CTG was
protective (OR 0.8, 95% CI 0.7–0.9; P=9x10-5). The association between ERAP1 and
AS has been replicated in Caucasian and non-Caucasian populations, including Han
Chinese, Korean and Portuguese individuals [211-214].
In family studies there are no confounding issues of population stratification
unlike larger population studies. We studied multiplex AS families in the North
American Spondylitis Consortium (NASC) and found a novel haplotype of ERAP1/ERAP2
to be associated with AS [215]. This study was the first to report an association of
ERAP2 with AS. This is a particularly interesting finding with implications to the
pathogenesis of AS. Unlike ERAP1, the only known function of ERAP2 is peptide
trimming for MHC class I molecules. ERAP1 and ERAP2 act in concert for efficient
antigen presentation. Hence an association with a haplotype of ERAP1 and ERAP2
would indicate that antigen presentation might be the key abnormality in the
pathogenesis of AS. This is corroborated by the recent finding that ERAP1 is associated
49
with AS only in HLA-B27 positive patients [128]. This was not the case for the other
well known AS associated genes like IL-23R and IL-12B. Similarly ERAP1 is associated
with psoriasis only in patients who are HLA-C positive [216]. Thus a strong genetic
interaction between ERAP1 and HLA-B27 could implicate functional interaction as well
considering the known functions of the two genes and the fact that the genes are not
in linkage.
1.8.3 Normal functions of ERAP1
During antigen processing, peptides generated by the proteasome are
transported into the ER with the help of the transporter associated with antigen
processing (TAP) protein. TAP can transport peptides up to 16 amino acids in length.
Hence these peptides have to be trimmed before presentation on MHC-I. Moreover.
peptides with proline in any of the first three positions of the peptide are not efficiently
transported by TAP. Analysis of MHC–peptide complexes has determined that proline is
commonly encountered in these positions [217,218]. This result suggested that
peptides are transported with N-terminal extensions into the ER and then trimmed to
appropriate lengths before MHC class I uploading. Following a search for such ER
resident enzymes, ERAP1 was discovered [202]. Thus, N-terminally extended peptides
generated by the proteasome enter the ER, assisted by TAP. These peptides are, in
turn, cleaved by ERAP1 and prepared for loading onto MHC class I molecules (Figure
1.7).
ERAP1 is peculiar among aminopeptidases in its ability to cleave long substrates
[206,219]. Although ERAP1 prefers 8-10 amino acid long peptides, it can
50
Figure 1.7 Schematic Diagram to show the role of endoplasmic reticulum
aminopeptidases in antigen processing and presentation
51
Legend to Figure 1.7
Schematic Diagram to show the role of endoplasmic reticulum aminopeptidases in
antigen processing and presentation
Proteins targeted for degradation are processed by the proteasome. The resulting
peptides are N-terminally extended, relative to the ideal length of MHC-I binding
peptides. They are transported into the Endoplasmic Reticulum (ER) by Transporter
Associated with Antigen Processing (TAP). These N-terminally extended peptides
have to be trimmed by ER-resident peptidases before they can be finally presented on
MHC-I.
Abnormal ERAP1 and ERAP2 activity can lead to the generation of abnormal peptides.
This can lead to unstable peptide-MHC complexes that can misfold. These can
accumulate in the ER or be presented as Free Heavy Chains (FHC) on the surface.
Abnormal peptide MHC complexes can also be presented on the surface with resulting
altered immune responses.
52
accommodate peptides with up to 16 amino acids [206,219]. A molecular ruler
mechanism was proposed initially to explain the length specificity of ERAP1 [206]. Not
only the N-terminal residues of the peptide that bind to the catalytic site of ERAP1, but
also the C-terminal residues and other residues along the length of the peptide can
affect the enzymatic activity of ERAP1 [204,206]. Peptides with both N- and C-terminal
ends bound respectively to the catalytic site and hydrophobic pockets are better
processed. There are some clues in the new crystal structure of ERAP1. ERAP1
assumes the closed conformation after peptide binding. ERAP1 has maximal activity in
this conformation. Thus a properly fitting, ideal peptide could promote the closed
conformation of ERAP1 and thereby improve enzymatic activity. Peptides shorter than
eight amino acids do not extend from the catalytic site up to the hydrophobic pocket of
ERAP1 (Figure 1.8) and could cause ERAP1 to remain in an open state.
Proline in the first 3 positions and hydrophilic C-terminal residues make the
peptides less favored (Figure 1.8). The inability to process peptides with N-terminal
proline is thought to be due to the absence of a free N-terminal group to be
recognized by the GAMEN motif of ERAP1 [220]. The catalytic site and the groove
extending from there are hydrophobic [209]. This could be the reason why very short
hydrophilic side chains are not preferred [221]. Valine, isoleucine and threonine with
β-branched side chains are poorly tolerated [221].
The subcellular localization of ERAP1 and ERAP2 has been a matter of debate.
In mice, ERAP1 has been shown to be present in the ER only. By contrast, the
53
Figure 1.8 ERAP1 Peptide specificity
54
Legend to Figure 1.8
Peptide binding Specificity of ERAP1
ERAP1 prefers residues between 8-10 residues but can accommodate peptides up to
16 residues due to the large size of its peptide binding groove. Longer peptides cannot
be accommodated and are poor substrates for ERAP1. Peptides shorter than eight
amino acids do not extend from the catalytic site up to the hydrophobic pocket of
ERAP1. Peptides with both N- and C-terminal ends bound respectively to the catalytic
site and hydrophobic pockets are better processed. Proline in the first 3 positions and
hydrophilic C-terminal residues make the peptides less favored.
55
presence of ERAP1 has been demonstrated on the cell surface as well as the ER in
humans. Interestingly, in human cell lines, both ERAP1 and ERAP2 are secreted when
overexpressed [203]. The postulated explanation for this phenomenon is that ERAP1
and ERAP2 are normally retained in the ER by certain binding proteins but can escape
when these retention proteins become saturated. It should, however, be borne in mind
that evidence of ERAP1 and ERAP2 secretion exists only in cell lines, and the current
understanding is that these aminopeptidases in human tissue are predominantly in the
ER but might be present on the membrane as well. ERAP1 and ERAP2 polymorphisms
associated with AS have been shown to affect the levels of expression of these genes
[222]. Whether any subsequent change in protein levels results in secreted
aminopeptidase activity needs to be resolved. Currently, the possible role of ERAP1
and ERAP2 in the pathogenesis of AS remains speculative, based on well-established
functions.
1.8.3.1 ERAP1 in peptide presentation
In HeLa cells that stably expressed an MHC class I molecule (H-2Kb) and
transfected with an H-2Kb binding epitope (MSIINFEKL), ERAP1 was suppressed by
small interfering RNA (siRNA). This led to a reduction in the processing of the
MSIINFEKL epitope, and the levels of peptide–MHC complexes decreased to 20% of
those of controls [219]. Cells from ERAP1-deficient mice show decreased levels of MHC
class I peptide expression on their surface [223]. Interestingly, a novel set of MHC–
peptide complexes were seen in the ERAP1-deficient mice, which were not present in
the wild-type mice. This observation indicates that some peptides that normally get
56
eliminated can be presented on MHC-I when ERAP1 is lacking. Thus, a change in
ERAP1 levels caused a total change in the peptide–MHC repertoire [217]. A follow-up
study showed that ERAP1 can also remove peptides that are not avidly bound to MHC
class I molecules in the ER, adding to the potential mechanisms by which ERAPs might
be associated with AS [224].
The loss of ERAP1 decreased the trimming of the aminopeptidase substrate
leucine-7-amino-4-methylcoumarin (L-Amc) but not Arginine-Amc (R-Amc), which
indicated the presence of other aminopeptidase enzymes within the ER with different
specificities.[202] Following a search for this aminopeptidase activity, ERAP2 was
cloned from a human leukocyte cDNA library (and initially known as leukocyte-derived
arginine aminopeptidase [L-RAP]) [203]. As mentioned previously, ERAP2 has not been
demonstrated in mice or rats and ERAP2 has been identified as a pseudogene by
genome sequencing [225]. The expression of ERAP1 and ERAP2 is upregulated by
interferon gamma (IFN-) in human cells; the proteins co-localize in the ER and can
form heterodimers [226]. ERAP1 and ERAP2 have unique peptide specificities, which
results in efficient and coordinated presentation of a wide variety of peptides when
these aminopeptidases act in concert. Apart from the study of multiplex AS families, no
other reports so far link ERAP2 to AS. Tsui et al. reported that polymorphisms in both
ERAP1 and ERAP2 could be involved in the pathogenesis of AS by altering peptide
processing [227].
ERAP2 was not found to be linked to AS in the initial GWAS, and there was
strong Linkage Disequilibrium (LD) between ERAP1 and ERAP2 [137]. ERAP2 is now
57
known to be associated with Crohn’s disease which has numerous similarities in
arthritis manifestations and pathogenesis to AS [228]. Individuals with the rs2248374
polymorphism that affects the 5′ canonical splice site of exon 10 of ERAP2 have
altered mRNA splicing, a truncated protein and nonsense mediated decay [229].
Almost 25% of the general population have this SNP and thus lack ERAP2 naturally
[229]. The lack of ERAP2 is associated with reduced MHC-I expression as assessed
by the W6/32 antibody [229]. Thus a closer look at this phenomenon was warranted in
AS. In a study of 470 AS patients and 420 ethnically matched normal controls, there
was no association between the ERAP2 rs2248374 SNP and AS [230].
Abnormal peptide processing could lead to unstable peptide–MHC complexes
that are prone to misfold. HLA B27 molecules that are not properly folded tend to
accumulate in the ER (see previous Figure 1.7) and cause ER stress leading to UPR, a
pro-inflammatory cellular response [135,160,161]. Alternatively, or in addition to, the
presentation on MHC class I molecules of abnormally-generated peptides as a result of
ERAP polymorphisms could lead to an abnormal immune response (see previous
Figure 1.7). It has been shown that cells from ERAP-deficient mice are immunogenic
in wild-type mice,[223] which could mean that ERAP variants can contribute to the
pathogenesis of AS by conferring anomalous peptide presentation and T-cell
recognition. Peptide-MHC instability leading to FHC formation can also cause aberrant
immune responses by binding to nontraditional immunoreceptors such as killer cell
immunoglobulin-like receptors (KIR) and leukocyte immunoglobulin-like receptors
(LIR).[231-233] The KIR receptors are found on lymphocytes and T cells while LIR are
seen predominantly in myelomonocytic lineage.[164] HLA B27 homodimers are
58
recognized by the KIR receptors KIR3DL1 and KIR3DL2 as well as by the LIR receptor
LIRB2.[163,167,234]It needs to be seen if the FHC interaction with these
immunoreceptors can play a role in the pathogenesis of AS.
1.8.3.2 ERAP1 and cytokine receptor shedding
ERAP1 has been reported to be involved in the cleavage of IL-1RII, IL-6R and
TNFRI [235-237]. In addition to the discovery of ERAP1 following a search for ER-
resident aminopeptidases, ERAP1 was independently identified through a yeast two-
hybrid approach while searching for molecules that bind to the extracellular portion of
TNFRI [237]. ERAP1 was also shown on bronchial airway cells by confocal microscopy
to be a type II transmembrane protein that co-localizes with TNFRI [237].
In the first study examining a link between ERAP1 and TNFRI shedding, an inverse
correlation between the surface expression of TNFRI and ERAP1 was shown by
immunoblots, indicating low levels of TNFRI on the surface of cells with high ERAP1
expression [237]. In the same study, culture supernatants of cells with high expression
of ERAP1 showed increased soluble TNFR-I level [237]. By site-directed mutagenesis,
it was also shown that ERAP1 does not function as the actual cleaving enzyme but,
rather, aids other metalloproteinases in TNFRI shedding [237]. It has subsequently
been demonstrated that ERAP1 must be present in a complex with nucleobindin-2 and
a 43-kDa heterogeneous nuclear ribonucleoprotein (hnRNP) G coded by the gene
RBMX (RNA-binding motif gene, X chromosome), for TNFRI shedding [238,239]. By
co-immunoprecipitation the association between ARTS-1 and RBMX was identified and
attenuating RBMX expression by RNA interference reduced the constitutive release of
59
TNFR1 exosome-like vesicles and IL-1beta-mediated cleavage of soluble TNFR1
ectodomains. Over-expression of RBMX increased TNFR1 exosome-like vesicle release
[239]. As is known for hnRNPs, hnRNP G could shuttle between the cytoplasm and
nucleus and bind to the cytoplasmic portion of the TNFRI-ERAP1 complex. ERAP1 has
also been shown to aid the cleavage of IL-1RII and IL-6R [235,236].
On the basis of this known function, I hypothesize that abnormal cytokine
receptor shedding (as a result of ERAP polymorphisms) could lead to decreased levels
of circulating cytokine receptors, which, in turn, might result in higher levels of
cytokine biological activity. Interestingly, the corresponding cytokines that bind to the
three receptors known to be affected by ERAP1 are directly or indirectly linked to the
pathogenesis of AS. The IL-1 gene complex has been found to be associated with AS
in a number of studies, and IL-6 is a cardinal cytokine involved in the generation of T
helper 17 (TH17), cells which have been found to be important in the pathogenesis of
AS [143,240].
1.8.4 ERAP1 polymorphisms and AS phenotype
There are no published reports on the effect of ERAP1 polymorphisms on the
clinical phenotype of AS. Clinical heterogeneity in AS is perplexing, as most patients
have the HLA-B27 gene and HLA-B27 cannot, therefore, be a determinant in variable
disease expression. Hence, other modifying factors have been considered, and ERAP
polymorphisms might be an important contributor in this regard.
60
1.8.5 ERAP1 and HLA-B27 pathogenesis
The hypothesis for HLA-B27 conferring susceptibility to AS has been outlined
above. It would be very interesting to see whether ERAP1 could contribute to any of
these mechanisms. As seen from the evidence so far, variations in ERAP1 can affect
the peptide trimming efficacy and support an altered peptide hypothesis. MHC-I FHC
might be formed or a UPR response evoked as a result of abnormal peptide
presentation. These FHC can exist as monomers or dimers and cause abnormal
immune interaction with KIR and LIR receptors.[134,163,233,241] Finally, abnormally
processed microbial peptides can result in molecular mimicry. Thus, almost all of the
aforementioned hypotheses could be explained by the ERAP1 association with AS.
1.8.6 ERAP1 in other diseases
Often the initial suggestions for a functional role of a novel gene come from
other conditions with which it may be associated. Both ERAP1 and ERAP2 are
associated with other diseases and thereby provide a source of clues to their biological
functions. The ERAP1 association in HLA-C positive patients with psoriasis has been
mentioned above [216]. Interestingly in Crohn’s disease, which has several common
genetic links, there was an association not with ERAP1 but with ERAP2 [228]. The
reported association was strongest with rs27524, an intronic SNP on ERAP1. These
authors did note that the ERAP1 rs30187 SNP also was significantly associated with
HLA-C positive psoriasis.
The ERAP1 K528R variant has been reported to be associated with systemic
arterial hypertension, diabetes mellitus and hemolytic uremic syndrome [242-244].
61
Investigating the functional role of ERAP1 in hypertension, the K528R variant of ERAP1
has been shown to be less active than the ancestral wild-type variant, which leads to
less bradykinin formation or lower inactivation of angiotensin II, or both [243]. ERAP1
is an enzyme that cleaves and inactivates angiotensin II[205]. ERAP1 also converts
kallidin to bradykinin[205]. Thus an efficiently functioning ERAP1 is essential for
decreasing angiotensin II (a vasopressor) and increasing bradykinin (a vasodilator).
The K528R variant would thus lead to hypertension by less vasodilator and more
vasoconstrictor level in circulation.
By site directed mutagenesis, ERAP1 variant cDNAs were synthesized and
expressed in insect Sf9 cells. ERAP1 was extracted from the supernatant of Sf9 cell
cultures and used for testing enzymatic activity. Using a fluorigenic substrate, L-leucyl-
β-naphtylamide, it was shown that replacing Leu at position 528 with any other amino
acid led to decreased enzymatic activity. The K528R variant in particular had the
lowest enzymatic activity. This was the first report to show an effect of ERAP1
polymorphisms on the catalytic function of ERAP1.
A similar hypothesis of decreased enzymatic activity has been proposed to explain the
ERAP1 association with hemolytic uremic syndrome. Abnormal cleavage can lead to
excess of the Shiga toxin, which is responsible for this disease [242].
ERAP1 and ERAP2 have been associated with pre-eclampsia.[245] Interestingly,
an exaggerated inflammatory response to the fetal allograft is a proposed mechanism
for pre-eclampsia [246]. HLA-G expression on the placenta is important for maintaining
the immune-privileged status of the fetus. Inhibitory KIR receptors bind to HLA-G
62
molecules, thereby suppressing the activity of natural killer cells [247,248]. The
expression of HLA-G can be decreased by blocking ERAP1 expression using inhibitory
RNA, but whether the association seen with ERAP1 and ERAP2 in pregnancy relates to
changes in HLA-G expression has not been resolved [249].
Variations in the antigen-presenting function of ERAP1 have been correlated
with several infectious and malignant diseases. An impaired ability to present
pathogen-derived peptides could lead to inadequate immune responses and increased
pathogenicity. ERAP1-deficient mice are not able to process the immunodominant
decapeptide HF10 of Toxoplasma gondii and die from overwhelming infection when
challenged with this pathogen [250]. An Ala146Pro mutation in the Gag protein of HIV
inhibited its ability to be cleaved by ERAP1 for MHC presentation, thereby resulting in a
decreased cytotoxic T cell response 57. This is corollary proof of the importance of
ERAP1 specificity and immune response.[250,251] Similarly, increased cervical cancer
metastasis and decreased survival have been reported to be associated with
polymorphisms in ERAP1 [252-254]. Variations in the levels of ERAP1 and ERAP2 along
with variations in the expression of surface MHC have been reported in various non-
lymphoid cancer cell lines [255,256]. ERAP polymorphisms could reduce the MHC
expression in tumor cell lines and result in a decrease in immune surveillance.
The ERAP1 K528R polymorphic variant is associated with pre-menopausal
osteoporosis in Japanese women.[257] As this variant is the same as that found to be
associated with decreased radiographic progression of AS[258], it seems that ERAP1
63
activity is important in the catabolic–anabolic balance of bone metabolism. Further
studies in this regard are warranted.
1.9 Aim
The broad aim of the study was to identify the functional implications of the newly
identified ERAP1-AS association. Specifically I had the following aims:
1. Does ERAP1 polymorphism lead to changes in the surface expression of
HLA-B27 in AS patients? Are there differences in FHC expression?
2. Can in vitro suppression of ERAP1 lead to changes in HLA-B27, surface
FHC and intracellular FHC expression? Can ERAP1 suppression lead to an
increase in HLA-B27 presenting abnormal and long peptides?
3. Are there differences in the interaction of ERAP1 with the different HLA-B27
subtypes?
4. Can ERAP1 polymorphism affect the serum cytokine receptor level in AS
patients?
5. In addition to being a susceptibility marker, is ERAP1 also a marker of
severity of disease?
64
1.10 Hypothesis
ERAP1 variants have differential peptide trimming capabilities. This leads to
altered HLA-B27 peptide loading and an abnormal immune response. Abnormal
peptide-B27 complexes can be unstable and lead to increased FHC formation that can
accumulate in the ER or be presented on the surface either as monomers or FHC
dimers. This functional interaction of ERAP1 with HLA-B27 may depend on the
subtype and can lead to pathogenic changes only in the AS-associated HLA-B2704
and B*2705 subtypes.
AS-associated ERAP1 variants can have abnormal clipping capability, resulting
in lower levels of circulating cytokine receptors in circulation. This could result in an
imbalance in the receptor-cytokine ratio with more cytokine causing inflammation.
As ERAP1 is a strong susceptibility marker it could impact the radiographic
severity of AS and the rate of radiographic progression.
65
Chapter 2
This chapter was published in the Annals of Rheumatic Disease: Haroon N, Tsui FWL, Uchanska-Ziegler B, Ziegler
A and Inman RD. Endoplasmic Reticulum Aminopeptidase 1 (ERAP1) has functionally significant interaction with
HLA B27 and dictates subtype specificity in ankylosing spondylitis. Ann Rheum Dis 2011; In Press).
FWLT aided in genotyping patients in the first part of the study. BUZ and AZ provided the MARB4 antibody. RDI
and FWT aided in planning the experiments. The draft was approved by all co-authors. I recruited patients for the
study, collected samples, separated PBMC and did flow cytometry. I did tissue culture of C1R cells, siRNA
experiments, western blot and flow cytometry. I performed all statistical analysis and prepared the draft of the
manuscript.
Functional interaction of ERAP1 and HLA-B27 is 2subtype specific and could be pathogenic in ankylosing spondylitis
2.1 Introduction
The pathogenesis of AS is not well understood. The significance of genetics is
evident from twin and family studies showing an increased risk of recurrence of AS in
first degree relatives and twins[103,104,144,145]. The first genetic risk factor
identified to be strongly associated with AS was HLA-B27 [146]. Despite thirty-five
years of research we still do not know how exactly HLA-B27 contributes to the
pathogenesis of AS. As HLA-B27 is a classical MHC-I molecule, an abnormality of
antigen presentation has been suspected for some time. Despite arduous efforts to
identify ‘arthritogenic peptides’ in AS, they have not met with great success [259].
Recurrence risk modeling studies have hinted at the possibility of AS being an
oligogenic disease with contributions from at least 6 important genes[144]. The recent
66
discovery of ERAP1 as an important genetic risk factor in AS invigorates the hope of
resolving the pathogenesis of AS. ERAP1 is an ER resident peptide cleaving enzyme
that can functionally interact with HLA-B27 by affecting antigen presentation [260].
Thus variations in ERAP1 can lead to changes in the peptide-B27 repertoire with
effects on immune response. Altered peptide-B27 complexes can have reduced
stability and lead to the formation of MHC-I free heavy chains (FHC). These FHC can
be retained in the ER leading to an ER stress response, or presented on the surface
resulting in abnormal immune interactions[158,164,167,233].
Antigens are processed and loaded on MHC-I molecules in the ER and
transported to the surface for presentation. Thus the surface expression of MHC-I is a
surrogate marker for adequacy of peptide processing. A screening approach to assess
the effects of ERAP1 variants on antigen presentation would be to study the level of
MHC-I expression on PBMC in AS patients. Considering the potential effects on the
pathogenesis of AS, it is important to look for the level of HLA-B27 expression as well
as MHC-I FHC.
The ERAP1 SNPs rs30187 and rs27044 lead to a K528R and Q730E change
respectively in the ERAP1 protein. These are the most widely replicated and strongest
ERAP1 SNPs associated with AS. It is now well established that ERAP1 variants have
altered aminopeptidase activity [207,261]. The rs27044 SNP has less enzyme activity
than the ancestral / wild type ERAP1 variant depending on the presence of other
substrates in the reaction [261].
67
As a first step, I compared the surface expression of HLA-B27 and FHC on the
peripheral blood mononuclear cells (PBMC) of AS patients with different genotypes of
these ERAP1 SNPs. Then I addressed the question whether reducing the level of
ERAP1 would have similar results in vitro. I developed an in vitro system with C1R cells
and used siRNA for ERAP1 suppression. In addition to FHC and HLA-B27 expression, I
studied the effect of ERAP1 suppression on the surface expression of HLA-B27 that
reacts with the mAb MARB4 and intracellular FHC (IC-FHC).
Taking this one step forward, I also addressed if the effects seen with ERAP1
suppression on C1R cells differ with the subtype of HLA-B27 that the cells express. I
did this by using different C1R cell lines that have been stably transfected with the
HLA-B27 subtypes: B*2704, B*2705, B*2706 and B*2709. As detailed in the
introductory chapter, HLA-B*2704 and B*2705 are associated with AS while B*2706
and B*2709 are not. Hence if the result of the ERAP1 suppression is different in the
AS-associated HLA-B27 subtypes compared to the AS-neutral subtypes, it would
further strengthen the hypothesis that ERAP1-B27 interaction is pathogenic in AS.
68
2.2 Methods
2.2.1 Patients
We enrolled Caucasian AS Patients attending the Spondylitis Clinic at Toronto Western
Hospital in the study if they met the following criteria
(a) Satisfied the modified New York criteria for the diagnosis of AS
(b) Consented to participate in the study
(c) Were HLA B27-positive
(d) Were not on any anti-TNF agents at the time of study
All patients underwent a comprehensive clinical examination. Patients were
assessed for disease activity by the Bath AS Disease Activity Index (BASDAI),
Erythrocyte Sedimentation Rate (ESR) and C Reactive Protein (CRP). Functional
capacity was evaluated using Bath AS Functional Index (BASFI). Peripheral blood was
collected at the same time as clinical assessment for flow assisted cell sorting (FACS).
2.2.2 Genotyping
DNA was prepared by Basil Chu in the Inman Lab from peripheral blood using a
standard DNA isolation kit (Gentra Systems, Minneapolis). The principle of cDNA
isolation from peripheral blood involves lysing RBCs which are non-nucleated and
lacking genomic DNA first, followed by lysing WBCs in the presence of a DNA stabilizer
to protect the DNA from the action of DNAase. Briefly, RBCs and WBCs were lysed with
69
a proporietary cell lysis solution from Gentra systems containing an anionic detergent.
Salt precipitation was used to precipitate protein and alcohol precipitation was used to
precipitate DNA from the supernatant. DNA was mixed with 100% isopropanol. The
precipitated DNA was subsequently washed with 70% ethanol and hydrated with a
DNA hydration solution consisting of 1 mM EDTA and 10 mM Tris.Cl at pH 7.5. DNA
concentrations were measured and samples aliquoted.
Genotyping was performed by Hing Wo Tsui in the Tsui Lab at the Toronto
Western Hospital [215]. Two SNPs in the ERAP1 gene known to be associated with AS
(rs27044 and rs30187) were typed by the allelic discrimination assay [262]. The
rs27044 SNP is at exon 16 of ERAP1 and leads to a C/G substitution at position
96118852 on chromosome 5. The rs30187 SNP is on exon 12 with a C/T substitution
at position 96124330 of chromosome 5. Optimized allelic discrimination assays for
SNPs were purchased from Applied Biosystems (Foster City, CA, USA). The Applied
Biosystems 7900HT fast system was used for the PCR.
Briefly, allelic discrimination assays are based on the ability to direct detect the
polymerase chain reaction (PCR) product by measuring the increase in fluorescence of
dye-labeled DNA probes. Two probes are used, one for each allele. Each probe
consists of an oligonucleotide with a 5´-reporter dye and a 3´-quencher dye. The
reporter dye is different in the two alleles while the quencher is common. Once the
PCR reaction progresses in a 5’-3’ direction, the 5’ reporter dye gets released from the
quenching effects of the quencher on the 3’ end of the probe leading to a signal that
can be detected by the PCR machine. In our system TET (6-carboxy-4,7,2´,7´-
70
tetrachlorofluorescein) is covalently linked to the 5´ end of the probe for of Allele 1.
FAM (6-carboxyfluorescein) is covalently linked to the 5´ end of the probe for Allele 2.
TAMRA (6-carboxy-N,N,N´,N´-tetramethylrhodamine) is the common quencher. Thus
if signals from TET and FAM are detected, the patient is heterozygous with both allele
present. Single signals would indicate a homozygous state for the corresponding allele.
2.2.3 Cell Lines
Hmy2.C1R (C1R) is a human lymphoid cell line with low expression of its
endogenous HLA class I molecules [263]. The cells have been stably transfected with
different HLA-B27 subtypes (B*27:04, B*27:05, B*27:06, and B*27:09) [264]. These
cells (a kind gift of Dr. José López de Castro, Madrid) were cultured in IMDM with 10%
Fetal Bovine Serum (FBS) to 70 to 90% confluence. The culture medium was changed
once every three days or earlier if required. The nucleofection protocol for C1R cells
was optimized using the cell line optimization nucleofector kit from AMAXA (Lonza, MD,
USA).
2.2.4 Optimization of Nucleofection
The Nucleofector™ Technology is a significantly better technique than
electroporation technology for gene delivery. The Amaxa Nucleofector™ Technology
involves a combination of optimized electrical parameters and specific Nucleofector™
Solutions that are ideal for different types of cells. By combining the ideal nucleofector
solution with the proper delivery pulse, delivery of the DNA of interest straight into the
71
nucleus can be achieved. As the DNA quickly reaches the nucleus, expression can
begin without delay.
The C1R cells were cultured to confluence and the cell line optimization
nucleofector kit (Amaxa) was used to identify the ideal combination of nucleofector
solution and nucleofection program that gave the best ratio of efficacy to cell death. In
brief, the cell line optimization kit has 2 nucleofection solutions L and V. A set of 7
programs were tried with both solutions bringing the total possible combinations to 14
(Table 2.1). The C1R cells were removed from culture and mixed with the
nucleofector solution and green fluorescent protein (GFP) DNA in an Amaxa-certified
cuvette. The number of cells that could be processed at one time ranged between 1 x
106 to 5 x 106 cells. The cells were mixed with 2 μg pmaxGFP and 100 μl Nucleofector
Solution. pmaxGFP encodes the GFP from copepod Pontellina p. FACS was used to
monitor the efficiency of transfection. Two control samples were performed for
assessing the effect of DNA alone and nucleofection alone on cell survival. In the first
control experiment, cells were mixed with Nucleofector Solution and pmaxGFP but no
nucleofection program was used. In the second control experiment, cells were mixed
with nucleofector solution, but without pmaxGFP and the nucleofection program was
applied.
Following nucleofection, the C1R cells were suspended in pre-warmed
supplemented IMDM media and incubated in a humidified 5% CO2 incubator at 37oC.
After 24 hours, cells were removed from culture and re-suspended in FACS buffer with
propidium iodide (PI). FACS was used to estimate the percentage of cells with GFP
72
expression and dead cells. Based on these experiments, the X-001 program with the
best efficacy:toxicity ratio was selected for further experiments.
Table 2.1 The Amaxa cell line optimization nucleofector kit
Program
Solution
A-020 L V
T-020 L V
T-030 L V
X-001 L V
X-005 L V
L-029 L V
D-023 L V
There are 14 possible combinations of nucelofection solutions and programs. These
are tried in various combinations to identify the ideal solution and program that gives
the best efficacy:toxicity ratio.
73
2.2.5 ERAP1 Suppression by RNA Interference
The Stealth RNAi™ (Invitrogen, Carlsbad, CA, USA) technology was used for the
siRNA experiments. Pools of three specific RNA duplexes for ERAP1 and a pool of non-
targeting duplex as negative control (NC) were used. The NC siRNA (Invitrogen, USA)
is a stealth control siRNA that does not induce any stress response and is not
homologous to anything in the vertebrate transcriptome. The HSS numbers of the
oligos are HSS122605, HSS122607 and HSS182221 and were delivered to the cells by
nucleofection using Amaxa® Cell Line Nucleofector Kit® V (Lonza, MD, USA). ERAP1
suppression was measured by Western blot after 96 hours. Five separate experiments
were done. Compared to the NC siRNA, ERAP1 siRNA (Figure 2.1) achieved a median
protein suppression of 74% (range: 69-79%; n=5).
2.2.6 Western Blot Assay
Western Blot (WB) was done to determine ERAP1 protein level in the C1R cells.
The level of ERAP1 after siRNA was compared to the value after NC to calculate the
level of suppression. Cell lysates were prepared from the different cell lines before and
96 hours after specific ERAP1 or negative control siRNA nucleofection. Cells were
washed and lysed with a buffered saline containing 1% Nonidet P-40 in the presence of
a mixture of protease inhibitors (Sigma-Aldrich). Following SDS-PAGE on an 8% slab
gel, the separated components were electroblotted onto a nitrocellulose membrane at
25 V overnight. The blots were incubated with peroxidase-coupled secondary Ab and
74
Legend: ERAP1 (E1) siRNA resulted in reduction of ERAP1 by 75 to 87% when
compared to the negative control (NC) siRNA. There was no difference between
the AS-associated and non-associated subtypes in level of suppression.
B*27 subtype
B*2704 B*2705 B*2706 B*2709
ERAP1 suppression
(compared to NC siRNA)
80% 80% 87% 75%
Figure 2.1 Western Blot and analysis showing ERAP1 suppression
with ERAP1 siRNA compared to the respective negative control siRNA
75
developed using an ECL detection kit. Films were analyzed with the Scion Image
Analysis software.
2.2.7 Fluorescence Activated Cell Sorting (FACS)
(1) Flow cytometry of patient samples: Peripheral blood was drawn and
mononuclear cells (PBMC) were isolated by Lympholyte® Cell Separation Media
(Cedarlane, Canada) followed by staining for flow cytometry. Intact HLA B27 was
identified by the conformation-specific mAb ME1 [265] and free MHC class I heavy
chains were identified using the HC10 mAb [266]. Secondary staining was done using
FITC-labeled goat anti-mouse IgG (Sigma-Aldrich, USA). As monocytes and B cells are
important antigen presenting cells, they were separately gated and studied using anti-
CD19 PE (BD Biosciences, USA) and anti-CD14 APC (AbD Serotec, USA) antibodies
respectively. Unstained cells and cells stained only with secondary antibodies were
used as controls for FACS. A total of 10,000 PBMCs were acquired and the mean
fluorescence intensities (MFI) of surface HLA-B27 and FHC on PBMCs as a whole and
separately on monocytes and B cells were measured (Figure 2.2).
(2) Flow Cytometry of C1R cells: C1R cell lines (transfected with B*27:04,
B*27:05, B*27:06 or B*27:09) were removed from culture 96 hours after siRNA
treatment. The cells were stained with the following antibodies: ME1, HC10 and
MARB4. The cells were also submitted to fixation and permeabilization by the BD
Cytofix/Cytoperm™ kit (BD Biosciences, USA) followed by HC10 staining for
intracellular FHC (IC-HC10). Comparison was made to flow cytometry results on cells
76
Figure 2.2 Flow Assisted Cell Sorting of Peripheral Blood
Mononuclear Cells
77
Legend to Figure 2.2
Flow Assisted Cell Sorting of Peripheral Blood Mononuclear Cells
(A) Scatter plot of peripheral blood cells with forward scatter (FSC) and side scatter
(SSC) showing the gating for PBMC.
(B) The mean fluorescence intensities (MFI) of HC10 and ME1 staining followed by
FITC tagged secondary antibody on PBMC are compared here against negative
control staining (FITC tagged secondary antibody alone)
(C) Scatter plot with FL1 against FL2 showing negative control staining (FITC
tagged secondary antibody alone)
(D) Scatter plot with single staining for HC10
(E) Scatter plot with single staining for ME1
(F) Histograms showing anti-CD14 APC positive monocyte population and (G) anti-
CD19 PE positive B cell population compared to unstained cells as controls in
the gray shaded area.
(H) Double staining of PBMC with anti-CD19 PE and anti-CD14 APC clearly
showing the distinct B cell and monocyte population.
(I) Double staining of PBMC with anti-CD14 and either ME1 or (J) HC10.
(K) Double staining of PBMC with anti-CD19 and either ME1 or (L) HC10.
78
treated similarly with control siRNA. MARB4 has been reported to bind to HLA B27 with
extended peptides [267].
2.2.8 Statistics
The MFI of HLA-B27 and FHC were compared between the genotypic groups of
the ERAP1 polymorphisms tested. In the dominant model, the MFI in the patients with
the minor allele only (homozygous for the minor allele) was compared to that in
patients with the major allele (homozygous for the major allele or heterozygous) using
the Mann-Whitney U test. For the additive model, the MFI were compared across the
three genotypes of each SNP by the Kruskal Wallis Test. The comparisons were made
in the whole PBMC population as well as separately in the monocyte and B cell
population. The Wilcoxon Matched-Pair Signed Rank test was used to compare MFI
before and after siRNA use in the B*27:05 cell lines. The Mann-Whitney U test was
used to compare the change in MFI of ME1, HC10, MARB4 and IC-HC10 between C1R
cells transfected with AS susceptible and non-susceptible HLA-B27 subtypes.
79
2.3 Results
A total of 54 patients (13 females) were enrolled in the study. These patients
were attending the spondylitis clinic in the Toronto Western Hospital with a diagnosis
of AS, based on the modified New York criteria. The modified New York criteria is
satisfied if the patient has long standing inflammatory back pain or spinal mobility
restriction with the presence of a set minimum level of damage in the sacroiliac joints.
Thus these patients had unequivocal evidence of AS. Anti-TNF medications are
prescribed for symptom control in AS patients and this can alter the immunological
profile of patients. Hence patients on anti-TNF medications were excluded. The mean
( SD) age of the patients was 41.2 ( 13.7) years and disease duration was 10 (±
9.8) years.
Standardized and widely accepted disease outcome measures were used to
assess the symptoms of patients and to compare disease activity between groups of
patients with different genotypes. This was done to ensure that changes in disease
activity did not contribute to any differences noted in HLA-B27 expression on the PBMC
of these patients. The best tool for assessing disease activity in AS is the BASDAI, a
disease activity score derived from patient reported symptom rating. A BASDAI of ≤ 4
in a scale of 0 to 10 is considered an acceptable level of control of symptoms while a
level above 4 is the threshold for starting anti-TNF medications. The mean BASDAI
values in the spondylitis clinic vary between 4 and 5 and it is higher in the anti-TNF
trials. C-Reactive Protein (CRP) and Erythrocyte Sedimentation Rate (ESR) are serum
markers of inflammation and are widely used in arthritis assessment. However, in AS
80
both CRP and ESR are neither sensitive tests for inflammation nor are they correlated
to patients symptoms [268].
The functional ability, or the ability to do routine and leisure activity is captured
in a questionnaire and the average score gives the functional index BASFI. Spinal
mobility is assessed by clinical measurements and an average of several
measurements can be represented in a 0 to 10 scale called the Bath AS Metrology
Index (BASMI). The mean BASDAI, BASFI and BASMI were 4.2 (± 2.4), 3.0 (± 2.6)
and 2.2 (± 1.5) respectively. Thus the patients included in this study had disease
activity at par with routine clinic patients. The number of patients with rs27044
genotypes CC, CG, GG were 22, 26 and 6 respectively; the number with
rs30187genotypes CC, CT and TT were 24, 23 and 7 respectively. There were no
significant differences in the age, BASDAI, BASFI, hemoglobin, total leukocyte count,
platelet count, ESR and CRP between the patients with different genotypes of ERAP1
rs27044 (Table 2.2) or rs30187 (Table 2.3).
2.3.1 Monocytes have the highest level of surface expression of HLA B27 and free heavy chains.
After gating on PBMC based on forward (FSC) and side scatter (SSC),
monocytes and B cells were separately identified by antiCD14-APC and antiCD19-PE
antibodies. The conformation specific ME1 antibody was used for estimating the intact
HLA-B27 expression and HC10 was used for FHC expression on whole PBMC and
separately on B cell and monocytes.
81
Table 2.2 Clinical and laboratory features of patients with different ERAP1
rs27044 genotypes
ERAP1 SNP
rs27044
CC CG GG
ERAP1 Variant E730 Q730
E730 Q730
N 22 26 6
Age
(years) 41.5 ± 14.1 41.96 ± 14.7 37.33
BASDAI 3.34 ± 2.3 4.30 ± 2.5 4.93
BASFI 2.35 ± 2.2 3.56 ± 2.9 2.367
Hemoglobin
(g/L) 144.3 ± 15.1 138.0 ± 15.9 147.7 ± 13.0
Leukocytes
(x 103 cells/dL) 7.39 ± 2.0 9.89 ± 11.8 6.3 ± 3.0
Platelets
(x 103 cells/dL) 284.0 ± 58.0 312.5 ±126.0 283.2 ± 42.5
ESR
(mm/Hr) 14.0 ± 27.9 12.2 ± 18.6 3.8 ± 1.5
CRP
(mg/L) 9.6 ± 13.6 10.9 ± 25.8 3.8 ± 1.5
82
Table 2.3 Clinical and laboratory features of patients with different ERAP1
rs30187 genotypes
ERAP1 SNP
rs30187
CC CT TT
ERAP1 Variant R528 K528
R528 K528
N 24 23 7
Age
(years) 41.04 ± 13.6 40.82 ± 15.2 43.50 ± 9.4
BASDAI 3.57 ± 2.3 4.5 ± 2.6 3.7 ± 1.1
BASFI 2.4 ± 2.1 3.9 ± 2.8 0.8 ± 0.5
Hemoglobin
(g/L) 145.7 ± 15.0 138.0 ± 16.7 140.5 ± 11.0
Leukocytes
(x 103 cells/dL) 7.5 ± 2.2 7.4 ± 3.1 14.7 ± 22.2
Platelets
(x 103 cells/dL) 283.6 ± 56.5 312.6 ± 132.7 293.8 ± 40.8
ESR
(mm/Hr) 13.1 ± 26.7 12.8 ± 19.4 4.5 ± 1.5
CRP
(mg/L) 9.3 ± 13.2 11.6 ± 27.3 4.0 ± 1.2
83
The median (interquartile range – IQR) values of the MFI of HLA-B27 ME1
staining were 807 (619; 1150) on monocytes, 588 (403; 775) on B cells and 445 (323;
599) on whole PBMC (Figure 2.3 A). The median (IQR) values of the mean
fluorescence intensity (MFI) of FHC HC10 staining were 57.8 (44; 124) on monocytes,
22.8 (13; 44) on B cells and 21.9 (14; 31) on whole PBMC (Figure 2.3 B).
2.3.2 ERAP1 rs27044 polymorphisms affect the expression of free heavy chains on the surface of monocytes
The monocytes, B cells and whole PBMCs were separately analyzed for total
MHC-I and FHC expression and this was compared between groups of patients with
different genotypes of the ERAP1 SNPs tested. Significant differences were noted in
the FHC expression on monocytes between groups of patients with the rs27044 SNP.
Patients with the major C allele, coding for the E730 variant of ERAP1 rs27044 had
significantly higher (92.1 vs 27.9; p<0.01) FHC on monocytes compared to those
homozygous for the minor G allele, coding for the Q730 variant (Fig. 2.4 A,B). The
Q730 variant increases the risk for AS in genetic studies. Hence patients who had only
the AS-risk allele of the ERAP1 rs27044 SNP had lower FHC on the surface of
monocytes.
The expression of FHC needs to be corrected for the total surface HLA-B27
expression and this was done by taking a ratio of the ME1 staining to the HC10
staining intensity. Even after correction for the ME1 staining level, the difference was
signifncantly different between the genetic groups with higher ME1:HC10 ratio (26.7 vs
84
A
B
Figure 2.3 Total HLA-B*27 and Free Heavy Chain Expression was
highest in the monocyte population among the peripheral blood
mononuclear cells.
85
Legend to Figure 2.3
Total HLA-B*27 and Free Heavy Chain Expression was highest in the monocyte
population among the peripheral blood mononuclear cells.
Among the PBMCs. monocytes and B cells were separately identified by
antiCD14-APC and antiCD19-PE antibodies. The conformation specific ME1 antibody
was used for estimating the intact HLA-B27 expression and HC10 was used for FHC
expression on whole PBMC and separately on B cell and monocytes.
The Mean Fluorescence Intensity (MFI) of intact surface HLA-B27 by ME1
antibody (A) and MHC-I free heavy chains by HC10 staining (B) were highest in the
monocyte population. B-cells followed by total peripheral blood mononuclear cells
including T-cells had progressively lower MFI of ME1 and HC10.
86
Figure 2.4 Flow Cytometry Analysis of MHC-I Free Heavy Chain Expression on
Peripheral Blood Mononuclear Cells from Patients with Ankylosing
Spondylitis
87
Legend to Figure 2.4
Flow Cytometry Analysis of MHC-I Free Heavy Chain Expression on Peripheral Blood
Mononuclear Cells from Patients with Ankylosing Spondylitis
Flow cytometry analysis of MHC class I free heavy chain (FHC) surface expression
(HC10+ staining) on peripheral blood mononuclear cells of patients (N=54) with
ankylosing spondylitis (AS). Shown are the differences in FHC surface expression on
monocytes (A) between patients with different genotypes of the rs27044 polymorphism
of ERAP1. A representative histogram plot (B) from patients with median HC10 MFI
from the two homozygous genotypic groups of the rs27044 ERAP1 SNP (CC vs GG)
shows the difference in FHC on monocytes. The histograms representing the CC and
CG group are labeled in the figure with the shaded area representing the GG group
and having lower FHC than the CC group. FITC-tagged secondary antibody alone was
used as negative control for staining. The difference remained statistically significant
(C) even after correcting for the expression of HLA-B27 complexes (ME1 MFI) on
monocytes.
88
12.9; p<0.01) in the group of patients with the disease susceptible Q730 variant (Fig.
2.4C).
The FHC expression changes were significant only in the monocytes (Figure
2.5 A) and not in B cells (Figure 2.5 C) or whole PBMC (Figure 2.5 E) with the
ERAP1 rs27044 SNPs. There was no significant difference in the intact HLA-B27
expression as well on any of the cell populations with the rs27044 SNP (Figure 2.5 B,
D, F). The ERAP1 SNP rs30187 did not affect HLA-B27 or FHC expression on any of
the cell populations (Figure 2.6).
2.3.3 Surface expression of HLA B27 and FHC was not related to disease activity:
BASDAI, a marker for inflammation, was comparable in patients across the
different ERAP1 genotypes (Tables 2.2, 2.3). The CRP and ESR appeared to be
higher in the group of patients with the E730 and R528 variants of ERAP1, but this was
not statistically significant. BASDAI, a patient reported form, is the gold standard for
assessing inflammation in AS while ESR and CRP are not good markers unlike in other
arthritides like RA [269]. There was no correlation of the surface expression of HLA-
B27 or FHC expression in any cell population with the inflammatory markers ESR or
CRP, nor with the clinical variables BASDAI or BASFI. The correlation between FHC
expression on monocytes and inflammatory markers (Figure2.7) BASDAI (A), ESR (B)
and CRP (C) are shown. Scatter plot with FHC MFI on y-axis and markers of
inflammation on x-axis are shown (Figure 2.7) and the linear regression lines with
95% confidence intervals are also depicted.
89
Figure 2.5 Comparing the effects of ERAP1 rs27044 SNP on free heavy
chain and Intact HLA-B27 expression on different cell populations
*
90
Legend to Figure 2.5
Comparing the effects of ERAP1 rs27044 SNP on free heavy chain and Intact B27 on different cell populations
The effect of the rs27044 polymorphism of ERAP1 on the HLA-B27 and free heavy
chain expression on all cell populations is shown here. Out of all comparisons done,
statistically significant difference (* = p<0.01) was noted in only the FHC expression on
the surface of monocytes (A, same as shown in the previous figure). There was a trend
mirroring this change with lower FHC expression in the total PBMC population (E). In
the monocytes (B) and B cells (D) the intact HLA-B27 expression, as identified by ME1
staining, was almost identical in patients with different rs27044 polymorphisms. In the
whole PBMCs, though not statistically significant, the ME1 expression showed a trend
towards lower levels in patients with the GG genotype.
91
Figure 2.6 Comparing the effects of ERAP1 rs30187 SNP on free heavy
chain and Intact HLA-B27 expression on different cell populations
92
Legend to Figure 2.6
Comparing the effects of ERAP1 rs30187 SNP on free heavy chain and Intact HLA-B27 expression on different cell populations
There was no significant effect of the ERAP1 rs30187 SNP on FHC or intact HLA-B27
expression on Monocytes (A,B), B cells (C,D) and whole PBMC (E,F). Statistical
analysis for the differences between patients with different genotypes was done using
the Kruskal Wallis One Way Anova test. For the recessive model, comparing the
patient with the major allele (CC and CT) to those patient with only the minor allele
(TT) I used the Mann-Whitney U Test. These are standard statistical analysis test
available in the statistical package of Predictive Analysis SoftWare (PASW). The
Kruskal Wallis test is used when comparing more than 2 unrelated groups and Mann-
Whitney test is used for 2 groups (CC and CT are grouped as one in the recessive
model).
93
Legend: As seen from the figures, there was no correlation of free heavy chain expression on monocytes to inflammatory markers like BASDAI (A), ESR (B) or CRP(C). Thus changes seen in FHC expression cannot be explained by change in the level of inflammation. Moreover, there was no significant difference in the inflammatory markers, between groups of patients with different ERAP1 genotypes. X-Y scatter plots are used to show the relationship between two variables. Linear fit lines are drawn to show graphically the relationship based on linear regression of the data. The fit lines were drawn by the PASW 18 program based on a linear fit method and the 95% confidence intervals are shown
Figure 2.7 Comparison of inflammatory markers across the
different genotypes
A B
C
94
2.3.5 ERAP1 suppression of HLA-B*2705 C1R cells causes changes in intracellular free heavy chains and surface MHC-peptide complexes
In this first round of investigations on C1R cells expressing HLA-B*2705, by FACS I
measured intact HLA-B27 and surface FHC as well as the level of IC-FHC and a
subpopulation of HLA-B27 reacting with the mAb MARB4 which has been reported to
bind HLA-B27 presenting unusually long peptides [267]. The C1R cells were subjected
to ERAP1 suppression by siRNA and negative controlled siRNA (non-targetting siRNA)
was used as control. Data from five siRNA experiments showed a significant increase
in the IC-HC10 (55.3 vs. 32.9; Figure 2.8 A) and MARB4 staining (8.2 vs. 6.5; Figure
2.8 C) after ERAP1 silencing. Representative histograms are shown for IC-HC10
increase (Figure 2.8 B) and MARB4 increase (Figure 2.8 D) following ERAP1 siRNA
compared to negative control siRNA. There was no significant difference in the HLA-
B27 complexes identified by ME1 (Figure 2.8 E) and surface FHC staining by HC10
antibody (Figure 2.8 F).
2.3.6 C1R cells expressing different HLA-B*27 subtypes have variable level of total MHC-I expression
To establish the baseline level of total MHC-I expression in the C1R cells with different
HLA-B27 subtypes, I did western blot analysis of the cell lysates and stained with the
HC10 antibodies that detects HLA-B and HLA-C free heavy chains. As C1R cells express
HLA-Cw4, HC10 would detect the respective HLA-B27 subtype as well as HLA-Cw4 in
the western blots. HLA-B*2705 expressing cells had the highest level of MHC-I
95
Figure 2.8 Effect of siRNA induced ERAP1 suppression on HLA-B*27
expression in C1R cells expressing HLA-B*2705
96
Legend to Figure 2.8
Effect of siRNA induced ERAP1 suppression on HLA-B*27 expression in C1R cells
expressing HLA-B*2705. Figures are representative of 5 experiments.
ERAP1 suppression led to consistent changes in IC-HC10 and MARB4 positive HLA-B27
complexes on C1R cells expressing HLA-B*2705. (A) Compared to control, non-
targettting siRNA, when ERAP1 siRNA was used to suppress ERAP1, C1R cells with
HLA-B*2705 were subjected ERAP1 suppression led to an increase in the intracellular
free heavy chain (FHC) expression as assessed by HC10 staining after cell
permeabilization (IC-HC10). (B) Representative plot showing increase in IC-HCq0
staining after ERAP1 siRNA (red) compared to control (blue). Negative control for
staining was done by secondary antibody alone (shaded area)
(C) ERAP1 suppression led to an increase in MARB4 positive surface HLA-B*2705
molecules, compared to cells treated with control siRNA. MARB4 antibody recognizes
HLA-B27 presenting long peptides. Representative plot (D) shows an increase in
MARB4 staining with ERAP1 siRNA compared to control siRNA. There was no
significant change in ME1 (E) or surface HC10 (F) staining with ERAP1 suppression.
The figures represent five experiments.
97
expression detected by HC10 while B2704, 06 and 09 expressing cells had similar
levels (Figure 2.9 A).
2.3.7 C1R cells expressing AS-susceptible HLA-B*27 subtypes have high MARB4 staining.
I tested the level of HLA-B*27 and FHC expression in the different HLA-B*27 subtype-
expressing cells. A clear distinction between the AS-associated and non-associated
subtypes was seen only with respect to HLA-B27 complexes recognized by the MARB4
antibody (Figure 2.9 E), which was higher in HLA-B*2704 and B*2705 cells compared
to those expressing B*2706 and B*2709 (p=0.004). Compared to the other HLA-B*27
subtypes, surface expression of intact HLA-B*27 (ME1) was higher in the B*2704 cells
(Figure 2.9 B) while intra-cellular FHC (HC10) was higher in B*2705 expressing cells
(Figure 2.9 D). There was no difference in the surface FHC expression between the
cell lines (Figure 2.9 C). Hence a pattern differentiating the AS-associated and AS-
neutral B*27 subtypes was not present in any other baseline characteristic.
2.3.8 ERAP1 suppression leads to an increase in intracellular free heavy chains only in C1R cells expressing AS associated HLA-B*27 subtypes
I then tested whether ERAP1 interacted differentially with distinct HLA-B27 subtypes. I
compared the changes induced by ERAP1 siRNA in C1R cells transfected either with
the AS-associated B*27:04 and B*27:05 or with the non-associated B*27:06 and
B*27:09 subtypes. After ERAP1 siRNA treatment, the IC-FHC was significantly
increased in C1R cells expressing either B*27:04 (mean 2.4 ± 1.6 fold) or B*27:05
(mean 1.8 ± 0.4 fold), but not in B*27:06 (mean 1.1 ± 0.3 fold) or B*27:09 (mean 1.1
98
Figure 2.9 Western Blot and FACS analysis of C1R cells at baseline.
Western blot analysis shows total HLA-B expression in C1R cells. FACS shows HLA-B27, surface and intracellular free heavy chains and MARB4 positive HLA-B27 levels in
C1R cells at baseline
The y-axes in all graphs are mean fluorescence intensities
of the corresponding antibody
A
B C
D E
99
Legend to Figure 2.9
Baseline expression of HLA-B*27 and MHC-I Free Heavy Chains in C1R cells with
different HLA-B27 subtypes
(A) Western blot analysis of cell lysates from C1R cells expressing the different HLA-
B27 subtypes. HC10 antibody, which binds to HLA-B and HLA-C, was used to estimate
their total baseline level in the cells. As C1R cells express HLA-Cw4, HC10 would
detect HLA-B27 and HLA-Cw4 in the western blots. HLA-B*2705 expressing cells had
the highest level of MHC-I expression detected by HC10 while B2704, 06 and 09
expressing cells had similar levels.
(B) Surface ME1 staining recognizing intact HLA-B27 was high in C1R cells expressing
HLA*B2704. The level of ME1 in HLA B*2705 expressing cells was similar to those
expressing the AS-neutral B27 subtypes.
(C) There was no pattern of surface MHC-I FHC expression (HC10), distinguishing the
AS-associated and neutral subtypes
(D) Intracellular (IC) HC10, which identifies FHC sequestered within the cells, was
higher in cells expressing HLA-B2705. But there was no pattern distinguishing AS-
associated and neutral subtypes.
(E) The C1R cells expressing AS-associated subtypes had higher MARB4 staining,
which recognizes HLA-B27 presenting distinct peptides that can be longer. However, it
must be noted that this closely parallels the total MHC-I content as seen by HC10
staining of western blots (A). Hence this difference in MARB4 positive complexes could
just be a reflection of the variation of the total MHC-I content. However, as shown such
a variation was not evident by flow cytometry using the ME1 or HC10 antibodies.
100
± 0.1 fold), resulting in a significant difference (p=0.02) between the AS-associated
and non-associated B*27 subtypes (Figure 2.10 A).
Representative plots show results of ERAP1 siRNA inhibition in C1R cells
expressing HLA-B*2704 (Figure 2.10 B), HLA-B*2705 (Figure 2.10 C), HLA-B*2706
(Figure 2.10 D) and HLA-B*2709 (Figure 2.10 E)
The percentage of cells expressing high IC-FHC levels was consistently and
similarly increased following ERAP1 suppression only in the AS-associated subtypes
(Figure 2.11 A and B) but not in the AS-neutral subtypes (Figure 2.11 C and D).
This analysis was done by gating the cells with IC-HC10 staining beyond a threshold
value (Figure 3.5 A and B) set based on negative staining with secondary antibody
alone: FITC-labeled goat anti-mouse IgG (Sigma-Aldrich, USA).
2.3.9 ERAP1 suppression leads to increase in MARB4 staining only in C1R cells expressing AS associated HLA-B*27 subtypes
We found an increase (p<0.001) in surface MARB4 staining following ERAP1
siRNA treatment in the C1R cells expressing B*27:04 (Figure 2.12 A) and B*27:05
(Figure 2.12 B), but no significant difference in the case of the AS non-associated
subtypes B*27:06 (Figure 2.12 C) and B*27:09 (Figure 2.12 D). The increase in
MARB4 MFI in the cells expressing the B*27:04 and B*27:05 subtypes was significantly
higher (1.25 ± 0.2 vs 0.9 ± 0.1; p<0.01) compared to those with B*27:06 and
B*27:09 (Figure 2.12 I). Similarly, the proportion of cells with high MARB4 MFI were
significantly increased only in the C1R transfectants expressing the AS-associated
subtypes (Figure 2.12 E, F vs G, H).
101
Figure 2.10 The effect of ERAP1 Suppression on Intracellular Free Heavy Chain
Expression in C1R cells expressing different HLA-B*27 subtypes
102
Legend to Figure 2.10
The effect of ERAP1 Suppression on Intracellular Free Heavy Chain Expression in
C1R cells expressing different HLA-B*27 subtypes.
Following ERAP1 suppression in C1R cells using siRNA, flow cytometry was used to
study changes in intracellular FHC (IC-FHC). The change in IC-FHC was significantly
higher (A) in C1R cells expressing the AS- associated subtypes (B*27:04 and B*27:05
vs B*27:06 and B*27:09).
Representative histograms show that the increase in IC-FHC is seen in C1R cells with
B*27:04 (B) and B*27:05 (C) and not in those with B*27:06 (D) and B*27:09 (E).
Histograms in blue and red indicate intracellular HC10 staining after negative control
(NC) siRNA and ERAP1 siRNA respectively. The figures are representative of 5
separate siRNA experiments.
103
Figure 2.11 Percent of cells with high expression of intracellular free heavy
chains
% o
f M
axim
um
%
of
Ma
xim
um
%
of
Ma
xim
um
%
of
Ma
xim
um
104
Legend to Figure 2.11
Histograms showing percent of cells with high expression of intracellular free heavy
chains
Based on negative control staining (shaded area in the plots) obtained after using only
secondary antibody (FITC-labeled goat anti-mouse IgG) a threshold for high positive
staining was set and the percentage of cells with high positive intracellular free heavy
chain (IC-FHC) staining was estimated.
As seen with the mean fluorescence intensities, the percentage of cells with high
positive staining for IC-FHC increased in C1R cells expressing HLA-B27 subtypes
associated with AS (B2704 in A and B2705 in B) but not in AS neutral subtypes (B2706
In C and B2709 in D). The control siRNA is a non-targeting negative control siRNA.
105
Figure 2.12 The effect of ERAP1 suppression on HLA-B*27 complexes recognized by
MARB4 antibody on the surface of C1R cells expressing different HLA-B*27 subtypes
I
Mea
n M
AR
B4
Fo
ld C
han
ge
afte
r ER
AP
1 s
iRN
A
106
Legend to Figure 3.5
The effect of ERAP1 suppression on HLA-B*27 complexes recognized by MARB4
antibody on the surface of C1R cells expressing different HLA-B*27 subtypes.
In the representative histograms shown, compared to negative control siRNA, following
ERAP1 suppression with siRNA the surface expression of MARB4-reactive HLA-B27
molecules increased in the C1R cells expressing B*27:04 (A) and B*27:05 (B) but not
in those expressing B*27:06 (C) and B*27:09 (D). Histograms in blue and red indicate
MARB4 staining after negative control (NC) siRNA and ERAP1 siRNA respectively.
There was a similar observation with the number of cells that were strongly positive for
MARB4 staining. Based on the negative control staining (solid area in histogram) with
secondary antibody alone (FITC-labeled goat anti-mouse IgG), a threshold for high
staining MARB4 positive cells was decided and the number of cells was estimated. In
the representative histograms shown with control and ERAP1 siRNA, an increase in
the number of MARB4 high positive cells after ERAP1 inhibition was seen only in C1R
cells transfected with HLA-B*2704 (E) and B*2705 (F) but not in B*2706 (G) and
B*2709 (H).
There was a significant difference (average of 5 experiments) in the change in MARB4
MFI (I) between the cell lines expressing the AS-susceptible subtypes (B*27:04 and
B*27:05) compared to those expressing the non-susceptible subtypes (B*27:06 and
B*27:09).
107
2.4 Discussion
The possibility of a distinctive interaction between peptide processing pathways
and HLA B27 has been considered for some time [270]. The recent discovery of a
genetic association of ER aminopeptidases and AS has raised expectations of finally
identifying the missing link in the pathogenesis of AS. This is the first study to show
that ERAP1 polymorphisms can affect MHC-I expression on the surface of human
peripheral blood cells. We have shown that patients with the E730 variant of ERAP1
have significantly higher FHC on the surface of monocytes. In addition, the expression
of intact HLA B27 is not different in patients with different polymorphic forms of
ERAP1.
The differences noted in the FHC expression appear to be due primarily to differences
in the genotype of rs27044 and not to disease activity variations as there was no
correlation of the surface expression of FHC and intact HLA B27 with any of the
disease activity markers. Moreover, there was no difference in the clinical variables
across the different genotypic groups. All patients included in the study were
Caucasians, and none were on any anti-TNF medications. The fact that the difference
remained significant after adjusting for the intact HLA B27 expression on the surface of
monocytes further strengthens the observation.
This novel finding is significant as it establishes a biological relevance and link
between the reported ERAP1 polymorphisms and HLA-B27. It has to be established
however, if the results would be the same in patients and controls who are HLA-B27-
108
negative. The antibody HC10 is not specific for HLA-B27 and FHC expression may be
different in patients who are HLA-B27 negative.
There was a differential effect of the rs27044 ERAP1 polymorphisms in different
peripheral blood cells. We and others [165] have observed the presence of significant
FHC expression only on the surface of monocytes and it is possible that the level was
not sufficient to show significant differences in the other cell populations. It is also
possible that peptides are handled differently or the types of peptides generated are
different in different antigen presenting cells. ERAP1 levels vary widely between
different tissues and it remains to be seen if the levels of ERAP1 vary between
different subpopulations of white blood cells. If the baseline levels are different in
monocytes and lymphocytes, a decrease in ERAP1 activity or expression might affect
only the cell population with lower levels to begin with. These questions and
possibilities need to be addressed in future studies.
With respect to possible relevance to disease pathogenesis, it is interesting that
the polymorphisms associated with decreased gene expression of ERAP1 are the same
polymorphisms strongly associated with AS [214]. Thus, if these polymorphisms lead
to a decreased level of enzyme expression, a direct effect on the catalytic site may not
be necessary to affect the overall enzymatic activity. The crystallographic structure of
ERAP1 shows that the polymorphisms reported to date are not in the catalytic site of
the enzyme [207,209]. These polymorphisms however may affect the substrate-ERAP1
interaction. A close binding of the substrate to ERAP1 might be cardinal to optimal
enzymatic activity of ERAP1. The fact that peptidase activity is affected by changing
109
the internal amino acid sequence of the peptides and not just the amino terminal end
cleaved by ERAP1 further supports this hypothesis [204].
It is also interesting that the minor allele which is seen in a higher frequency in
patients with AS results in lower FHC expression on monocytes. It is compelling to
think that, if FHC expression is important in the pathogenesis of AS, why should the
AS-associated minor allele of the tested polymorphism lead to a decrease in FHC. The
purpose of this study was to see if there is an interaction between ERAP1 and HLA B27
and that appears to be the case. It is possible that the Q730 variant of ERAP1 leads to
abnormal FHC that are retained in the ER leading to ER stress. ER retention would
result in lower surface expression of FHC with the net result that patients with the
E730 variants have higher FHC surface expression. A drawback of the studies on PBMC
is that, IC-FHC was not analyzed. However, this was done in the C1R cells and a
reduction in ERAP1 levels did result in increased IC-FHC. FHC can also be ligands for
the immunosuppressive receptors like the LIR receptors [233,241]. Thus a lower
expression of FHC could lead to less activation of the immunosuppressive receptors
and result in inflammation.
The new crystal structure shows that ERAP1 exists in both an open and closed
conformation. The closed conformation is enzymatically more active than the closed
form. This closing action happens after the peptide is accommodated in the catalytic
groove. The K528R variant is on the outer surface of domain II of ERAP1 and is
unlikely to directly affect the catalytic effect. But this variation can lead to abnormal
interactions with neighboring residues and inhibit the conformational change of ERAP1,
110
indirectly reducing enzymatic activity. The Q730E variant which we have seen to be
affecting FHC expression in our studies has a change in the inner surface of domain IV
of ERAP1. Domain 4 has a concave surface and arches over the catalytic site on
domain II forming a large cavity. As Q730E is on the inner surface, in the closed state,
this approximates the active site with the bound peptide. Thus this residue change
could impact the interaction of ERAP1 with the substrate.
It is now known that both the R528 (ERAP1 rs30187) and E730 (ERAP1
rs27044) variants have altered peptide trimming activity [261]. However, it is difficult
to predict the effect of the variation without knowing the substrate involved. There is
significant variability in the enzyme activity of the Q730E and K528R variants
depending on the substrate and its concentrations. Hence the effects of these ERAP1
variants may change with cell type and milieu. In the presence of a non-substrate
peptide activator of ERAP1 (SIINFEKL) both R528 and E730 had lower enzymatic
activity than the wild type. In the absence of the activator, the substrate reaction
consisted of leucine - amido methyl coumarin (L-AMC) alone wherein the R528 variant
had lower peptide trimming effect compared to the wild type enzyme while the E730
variant had increased peptide trimming effect. In low substrate concentration levels
the E730 variant had 2 fold lower peptide trimming action compared to the wild type
variant. Evnouchidou et al. reported that both the E730 and R528 variants decreased
HLA-B27 expression on cells with a greater effect conferred by the E730 variant [261].
Evans et al. recently reported decreased trimming by the R528 variant (rs30187)
associated with AS [128]. However, the Q730E variants were not studied.
111
Thus it is quite clear from the studies in hypertension and the recent functional
studies that the ERAP1 variants associated with AS can lead to a change of enzymatic
function. However, the studies by Evanouchido et al. outlined above clearly show the
complexity of the ERAP1 enzymatic reaction kinetics which appears to follow a
substrate inhibition model [261]. It can be appreciated that both gain and loss of
function can affect normal antigen processing. FHC can decrease on the surface as a
result of either an increase in function leading to the generation of more B27-binding
peptides. It could also reflect a loss of ERAP1 function with less degradation of the
B27-binding peptides. Hence our results cannot definitively answer the question of
whether a loss of function or gain of function underlies the observations. What we
have demonstrated is that polymorphisms of rs27044 are functionally relevant and are
associated with significant changes in the surface HC10-staining population (FHC). The
Q730 variant was associated with lower surface FHC. This by itself shows that the
polymorphism has a functional impact on MHC ± MHC-peptide expression.
Unlike the studies on PBMC, there was no significant difference in the surface
FHC expression. Probably, the peptide-HLA complexes resulting from ERAP1
suppression are so unstable after significant reduction in ERAP1 activity, that they
cannot be transported to the surface. Hence all FHC would be retained in the ER
resulting in minimal variation of the surface FHC. We did see a strong increase in the
IC-FHC in the AS-associated HLA-B27 subtypes and the relevance of this finding to AS
is further strengthened by the variation observed between AS-associated and non-
associated subtypes. There are differences between the conformation and the
flexibility of the peptide binding grooves of these HLA-B27 subtypes [259]. Changes in
112
FHC expression have been implicated in the pathogenesis of AS [163,167]. FHC dimers
can be recognized by leukocyte receptors such as Killer Cell Immunoglobulin-like
Receptors (KIR) and Leukocyte Immunoglobulin-like Receptors (LILR) leading to
abnormal immune responses [163,167]. FHC accumulation in the ER can lead to an
unfolded protein response and ER stress with production of inflammatory cytokines
[160].
The results above should be interpreted in light of the fact that cellular
activation can affect MHC-I expression. I have not corrected the FHC and HLA-B27
expression for the level of activation of the cells. However, the clinical markers of
inflammation like CRP and ESR were not different in the different genotypic groups. It
should also be pointed out that HC10 is not specific for FHC of HLA-B27 origin and
could reflect FHC from any HLA-B or HLA-C molecule. The exact specificity of MARB4
has not yet been defined.
Mass spectrometry determined that the MARB4-reactive population presented
peptides ranging from 8 to 33 amino acids [267]. It was subsequently shown to detect
HLA B27 with very low or absent β2 microglobulin but was conformation-dependent,
like mAb ME1 and unlike mAb HC10 [271]. The MARB4-positive complexes seem to be
different with respect to the peptides presented and possibly the conformation of the
B27-peptide complex. Thus changes in MARB4 MFI in the current results are relevant
for modeling pathogenesis, because this could ultimately affect immune recognition
and the response generated.
113
Dr. Adreas Ziegler and Dr. Barbara Uchanska-Ziegler from the Institut für
Immungenetik, Charité-Universitätsmedizin Berlin, Freie Universität Berlin, Berlin,
Germany, are our active collaborators and they addressed the specificity of MARB4 in
our combined manuscript [272]. Dr. Ziegler carried out binding experiments in BM36.1
cells transfected with B*27:05. MARB4 reacted only with B*27:05 presenting peptides
with Gly in the N-terminus. Up to six amino acids could be added to the N-terminal end
without abolishing MARB4 reactivity, provided the Gly-Arg motif (Gly and Arg in the A
and B pockets, respectively) was present in the peptide. These results show that
MARB4 reacts with B*27:05 molecules in a peptide-dependent fashion which is
additionally restricted to peptides that are presented such that a Gly residue occupies
the A pocket. Peptides with an N-terminal Gly and a length of nine or more residues
are known to be poor substrates for ERAP1 [273]. Reducing the activity of ERAP1
could further increase their availability for presentation by HLA-B27 molecules. This
could explain why MARB4 staining complexes increased after ERAP1 suppression.
The presence of Gly at the N-terminus leads to a conformational
rearrangement of several heavy chain residues at one end of the binding groove
(Arg62, Glu163, Trp167) that appear to be part of the MARB4 epitope on HLA-B27
molecules[274]. Hence MARB4, by detecting HLA-B27 complexes displaying a peptide
with Gly at the N-terminus (a negative for ERAP-1 activity) and Arg at P2 (ideal for
HLA-B27 binding), could be seen as a unique antibody that identifies a peptide-B27
subpopulation resulting from insufficient cleavage by an aminopeptidase which may be
of paramount importance in antigen presentation.
114
2.5 Conclusions:
ERAP1 variants can lead to changes in MHC-I FHC expression on PBMC. The
MHC-I FHC expression on monocytes is significantly higher in AS patients with the
E730 variant of ERAP1. Alteration of ERAP1 levels can affect both the conformation
and the peptide display by HLA-B27 molecules. Furthermore, these properties appear
to be differentially affected in cells expressing AS-associated as opposed to non AS-
associated HLA-B27 subtypes. The ERAP1-B*27 functional interaction could be a key
missing link in the pathogenesis of AS.
115
Chapter 3
This chapter was published in the Journal of Rheumatology: Haroon N, Tsui FWL, Chiu B, Tsui HW, Inman RD.
Serum cytokine receptors in ankylosing spondylitis: Relationship to inflammatory markers and endoplasmic
reticulum aminopeptidase polymorphisms. J Rheum 2010;37(9):1907-10.
FWLT and HWT aided in genotyping the patients. BC aided in extracting DNA from peripheral blood and storing
serum. RDI guided planning the experiments. All authors approved the draft. I sorted the serum, performed all
ELISAs, did statistical analysis and prepared the draft.
Serum Cytokine Receptors in Ankylosing 3Spondylitis: Relationship to inflammatory markers and Endoplasmic Reticulum Aminopeptidase Polymorphisms
3.1 Background
HLA-B27, the major susceptibility gene, has an estimated attributable risk of 16-
50% [28]. The role of HLA-B27 remains unknown and recently other risk factors
including non-MHC genes such as IL1R2, IL12B, TNFRSF1A and IL23R have been
identified [127,128,137,275]. As discussed above, the strongest non-MHC gene
associated with AS is ERAP1 (endoplasmic reticulum aminopeptidases 1) with a
population attributable risk of 26% [137]. ERAP1 is also known as ARTS-1
(aminopeptidase regulating tumor necrosis factor receptor [TNFR] shedding 1).
ERAP1 plays a cardinal role in peptide trimming within the ER for MHC-I antigen
presentation [260]. Mice lacking ERAP1 show reduced surface expression of MHC-I and
altered presentation of MHC-I associated antigens [223]. I have presented evidence in
the previous chapters that ERAP1 polymorphisms and varying ERAP1 levels can affect
HLA-B27 expression with increase in MARB4 positive complexes on the cell surface and
116
more IC-FHC accumulation. In this section I address another known function of
ERAP1.
As the alternate name ‘ARTS-1’ suggests, ERAP1 was first established as an
enzyme that plays a role in clipping of the membrane bound receptor of TNFα (TNFRI)
[237]. Subsequently it was found that ERAP1 also helps in clipping the receptors for
interleukin 1 receptor II (IL1RII) and IL6Rα [235,236]. Abnormal clipping could result
in lower levels of circulating cytokine receptors in circulation. This could result in an
imbalance in the receptor-cytokine ratio with more cytokine causing inflammation.
I formulated a hypothesis based on the receptor shedding function of ERAP1 as
this could provide a basis for the association of ERAP1 with AS. The AS associated
ERAP1 variants are known to have less enzymatic activity. AS patients could have less
ERAP1-mediated cytokine receptor shedding and more circulating TNFα, IL-1 and IL-6
leading to pronounced inflammation. If so, patients with evidence of active
inflammation should have lower serum cytokine receptors. Moreover, patient with
different variants of ERAP1 should have different levels of cytokine receptors in
circulation.
3.2 PATIENTS AND METHODS
3.2.1 Patients
Patients satisfying the modified New York criteria for AS, attending the
Spondylitis Clinic at Toronto Western Hospital and consenting to participate were
included in the study. Sequential patients from the clinic were invited to participate,
irrespective of disease activity at the time. Patients on anti-TNF treatments or other
117
biological agents were excluded. All patients underwent a comprehensive clinical and
radiographic examination. Patients were assessed for disease activity by BASDAI, ESR
and CRP. Functional capacity was evaluated using the BASFI.
3.2.2 Genotyping
DNA was prepared by Basil Chu in the Inman Lab, from peripheral blood using a
standard DNA isolation kit (Gentra Systems, Minneapolis) as described in the previous
chapter. Briefly, The principle of cDNA isolation from peripheral blood involves lysing
RBCs which are non-nucleated and lacking genomic DNA first, followed by lysing WBCs
in the presence of a DNA stabilizer to protect the DNA from the action of DNAase.
Briefly, RBCs and WBCs were lysed with a proporietary cell lysis solution from Gentra
systems containing an anionic detergent. Salt precipitation was used to precipitate
protein and alcohol precipitation was used to precipitate DNA from the supernatant.
DNA was mixed with 100% isopropanol. The precipitated DNA was subsequently
washed with 70% ethanol and hydrated with a DNA hydration solution consisting of 1
mM EDTA and 10 mM Tris.Cl at pH 7.5. DNA concentrations were measured and
samples aliquoted.
Two SNPs that were strongly associated with AS (rs27044 and rs30187),
included in the immunological studies described above, and one that was found to be
protective for AS in the GWAS studies (rs10050860) were selected for this study. The
ERAP1 SNP rs10050860 leads to an amino acid change D575N where aspartic acid in
position 575 is changed to asparagine. This change happens in Domain III of ERAP1
away from the catalytic site. Optimized allelic discrimination assays for SNPs from
118
Applied Biosystems (Foster City, CA, USA) and the Applied Biosystems 7900HT fast
system was used for the PCR [262].
Briefly, allelic discrimination assays are based on the ability to direct detect the
polymerase chain reaction (PCR) product by measuring the increase in fluorescence of
dye-labeled DNA probes. Two probes are used, one for each allele. Each probe
consists of an oligonucleotide with a 5´-reporter dye and a 3´-quencher dye. The
reporter dye is different in the two alleles while the quencher is common. Once the
PCR reaction progresses in a 5’-3’ direction, the 5’ reporter dye gets released from the
quenching effects of the quencher on the 3’ end of the probe leading to a signal that
can be detected by the PCR machine. In our system TET (6-carboxy-4,7,2´,7´-
tetrachlorofluorescein) is covalently linked to the 5´ end of the probe for of Allele 1.
FAM (6-carboxyfluorescein) is covalently linked to the 5´ end of the probe for Allele 2.
TAMRA (6-carboxy-N,N,N´,N´-tetramethylrhodamine) is the common quencher. Thus
if signals from TET and FAM are detected, the patient is heterozygous with both allele
present. Single signals would indicate a homozygous state for the corresponding allele.
3.2.3 Serum cytokine and soluble cytokine receptor levels
Sera were collected from AS patients at the time of clinical evaluation and
stored in aliquots at -70oC until analyzed. ELISA was used to determine the serum
levels of the cytokines IL1, IL6 and TNFα (BD, San Jose, CA) and the soluble receptors
of IL1RII (sIL1RII), IL6R (sIL6R) and TNFRI (sTNFRI) (R & D, Minneapolis, MN),
performed in duplicates according to the manufacturer’s directions. The minimum
detectable cytokine serum levels by the ELISA respectively for IL1, IL6 and TNFα were
119
0.8, 2.2 and 2 pg/ml. The minimum detectable serum cytokine receptor levels were
10, 6.5 and 0.77 pg/ml for sIL1RII, sIL6R and sTNFRI, respectively.
3.2.4 Statistical analysis
Patients were stratified according to the genotypes of the ERAP1 SNPs (rs30187
[CC, CT, and TT]; rs27044 [CC, CG, and GG]) and rs10050860 [CC, CT, and TT]). For
small sample numbers, parametric tests should be used for analysis. Comparison
across 3 or more groups can be done by the Kruskal-Wallis H test [276], while
comparison of two groups is done by the Mann-Whitney U test [277] in parametric
statistical testing. The soluble cytokine receptor levels in the different genotype
groups were compared using the Kruskal-Wallis H Test. To rule out recessive effects,
the Mann-Whitney U test was used to compare soluble receptor levels in patients with
the major allele of the three SNPs to those homozygous for the respective minor allele.
Thus comparisons were made between the following groups: patients with the Q730
variant: E730 variant alone (rs27044), K528:R528 alone (rs30187) and D575:N575
alone (rs10050860).
As inflammation could affect the cytokine receptor levels, ‘partial correlations’
was used to correlate the serum level of each cytokine receptor to other receptors
[278]. The partial correlations test allows for correction for other parameters and we
corrected for the objective markers of inflammation (ESR and CRP). The soluble
receptor level was also correlated with BASDAI, BASFI, ESR and CRP, using
Spearman’s Rank correlation.
120
3.3 RESULTS
3.3.1 Patients
Eighty AS patients (21 women and 59 men) with a mean (± SD) age of 42.3 (+
10.6) years and mean (± SD) disease duration of 18.5 (+ 11.4) years were enrolled in
the study. Sixty seven patients (83.7%) were HLA-B27-positive. The mean ESR and
CRP were 18 ± 15.6 mm/hr and 14.8 ± 18.8 mg/L, respectively. Both ESR and CRP
are measures of inflammation and are tested using blood drawn from patients. The
normal value of ESR and CRP are below 20 mm/Hr and 10 mg/L, respectively with
slight variations between labs, depending on the kit used. Although they are
recognized as very good markers of systemic inflammation it is not a particularly good
marker in AS [269]. This may be because inflammation is more localized in AS.
BASDAI, a disease activity score, on the other hand better reflects disease
activity in AS [269]. It is calculated from a patient reported questionnaire of symptoms
and a value below 4 is considered an acceptable level of disease activity. The BASFI is
a functional index and reflects any functional limitations the patients may experience.
The mean BASDAI and BASFI scores were 5.3 ± 2.4 and 4.4 ± 2.6 respectively. Thus
the patients included in the study had active disease and this is the typical range seen
in AS patients attending rheumatic disease clinics. There was no significant difference
between the patients in the different genotypic subsets with regards to age, disease
duration, BASDAI, ESR or CRP.
121
3.3.2 Genotyping
Two AS-associated and one AS-protective ERAP1 SNP was genotyped as
detailed in the methods. The AS-associated ERAP1 SNPs rs30187 and rs27044 SNPs
increase the risk of AS while rs10050860 protects against AS. The use of the terms
‘associated’ and ‘protective’ simply means that, compared to the general population,
the minor allele of the AS-associated SNPs are overrepresented while those of the
protective SNPs are underrepresented in AS patients. The allele frequencies of ERAP1
SNPs were 29, 39 and 12 (GG, AG, AA) for rs30187; 37, 39 and 4 (CC, CG and GG) for
rs27044 and 53, 25 and 2 (CC, TC, TT) for rs10050860 respectively.
3.3.3 Soluble cytokine receptor levels
The mean (±SD) serum soluble cytokine receptor levels in the AS patient cohort
were 1.38 ± 0.42 ng/ml (sTNFRI), 11.89 ± 4.3 ng/ml (sIL1RII) and 29.63 ± 11.8
ng/ml (sIL6R). X-Y scatter plots were used to show the relationship between two
variables. Linear fit lines were drawn to show graphically the relationship based on
linear regression of the data. The fit lines were drawn by the PASW 18 program based
on a linear fit method and the 95% confidence intervals are shown in blue. The linear fit
method uses the formula y = a +bx. Thus based on the data available, the
corresponding value of Y for any value of X will depend on ‘a’ (constant) and ‘b’ (the
slope of the fit line).
Contrary to the expectation that lower cytokine receptors would be associated
with higher inflammation, we observed (Figure 3.1 A,B) a significant positive (rather
than negative) correlation of sTNFRI with the markers of inflammation CRP (R=0.43;
122
A
R=0.43 p<0.001
R=0.30 p=0.01 B
Figure 3.1 Correlation of TNFRI with CRP and ESR
123
Legend to Figure 3.1
Correlation of TNFRI with CRP and ESR
The scatter plots above show TNFRI level on the y-axis and CRP (A) or ESR
(B) on the x-axis. There was significant positive correlation of serum soluble TNFRI
level with CRP (A) and ESR (B). The correlation between TNFRI and CRP was
stronger (R=0.43, p < 0.001) compared to that with ESR (R = 0.30; p = 0.01).
X-Y scatter plots are used to show the relationship between two variables.
Linear fit lines are drawn to show graphically the relationship based on linear
regression of the data. The fit lines were drawn by the PASW 18 program based on a
linear fit method and the 95% confidence intervals are shown in blue. The linear fit
method uses the formula y = a +bx. Thus based on the data available, the
corresponding value of Y for any value of X will depend on ‘a’ (constant) and ‘b’ (the
slope of the fit line). When the fit line is at 45 degrees between the X and Y axes, it
indicates perfect correlation of X and Y values with an R value (correlation co-efficient)
of 1. Correlation is significant If the p value is < 0.05. However the strength of
correlation depends on the R value. Traditionally R > 0.9 indicates very strong
correlation, 0.7-0.9 is strong, 0.4-0.7 is moderate and 0.2-0.4 is weak correlation.
124
p<0.001) and ESR (R=0.30; P=0.01). There was no correlation of sTNFRI with
BASDAI. There was also no correlation of sIL1RII and sIL6R with ESR, CRP or BASDAI
(Figure 3.2).
3.3.4 No significant effect of ERAP1 polymorphism on serum cytokine receptor level
There was no significant difference in the soluble cytokine receptor serum levels
between the different genotype groups of ERAP1 rs27044, rs30187 and rs10050860
polymorphisms (Figure 3.3). A dominant model was tested where the patients who
have the major allele of a polymorphism were grouped together and the other group
consisted of patients with only the minor allele. Thus comparisons were made between
the following groups: patients with the Q730 variant: E730 variant alone (rs27044),
K528:R528 alone (rs30187) and D575:N575 alone (rs10050860).
Thus in the Kruskal-Wallis H test, there was no difference in the serum cytokine
receptor levels across the three genotypic groups of each SNP tested. Similarly, with
the Mann-Whitney U test comparing patients with the major and minor variants
mentioned above, there was no significant influence of the ERAP1 polymorphisms on
the cytokine receptor levels.
3.3.5 Serum cytokine levels
Serum cytokine levels are difficult to measure and were detectable only in a minority
of patients that I enrolled. Serum levels were detectable in 21. 29 and 22
125
A B
C
Figure 3.2 No correlation of CRP, ESR and BASDAI with serum IL-6Rα and IL-1RII
126
Legend to Figure 3.2
Correlation of CRP, ESR and BASDAI with serum IL-6Rα and IL-1RII
In the scatter plots, IL-6R (blue) and IL1RII (gray) are shown on the y-axis with
markers of inflammation CRP (A), ESR (B) and BASDAI (C) on the x-axis. The linear fit
method was used to draw the fit lines and there was no significant correlation between
the level of soluble IL-6Rα and IL-1RII with CRP, ESR or BASDAI.
X-Y scatter plots are used to show the relationship between two variables.
Linear fit lines are drawn to show graphically the relationship based on linear
regression of the data. The linear regression fit lines are shown in blue for IL-6R and
gray for IL1RII. The linear fit method uses the formula y = a +bx.
127
Figure 3.3 Serum Cytokine Receptors in patients with different ERAP1
polymorphisms
ERAP1 rs27044 ERAP1 rs30187 ERAP1 rs10050860
A
B
C
D
E
F
G
H
I
128
Legend to Figure 4.3
Serum Cytokine Receptors in patients with different ERAP1 polymorphisms
This figure shows box plots or box and whisker plots which are ideal for showing the
distribution of the values across the groups. These plots do not assume a parametric
nature for the data. The central line stand for the median value, the upper and lower
edges of the box represent the 25th and 75th quartile values while the whiskers
represents 2 standard errors of the mean value.
As seen in the figure, the distribution of the values was similar across the different
groups and the serum cytokine receptors of TNFRI, IL-6Rα and IL-1RII, respectively
were not significantly different in patients with different genotypes of ERAP1 rs27044
(A, B, C), rs30187 (D, E, F) and rs10050860 (G, H, I).
129
patients respectively for IL1, IL6 and TNFα assays. The mean (±SD) serum levels were
28.4 ± 27.1 (IL1), 44.4 ± 71.7 (IL6) and 104.5 ± 90.2 (TNFα).
There is a possibility that cytokine receptors in circulation can affect the serum
levels of cytokines. However there was no difference in the soluble cytokine receptor
levels between patients with and without detectable corresponding cytokine levels. The
ratio of serum cytokine to the soluble receptor levels was not different in patients with
different genotypes of the ERAP1 polymorphisms tested.
3.4 Discussion
This is the first reported study of the role of the cytokine receptor shedding
function of ERAP1 in the pathogenesis of AS. I found no influence of the respective
ERAP1 SNPs (rs30187, rs27044 and rs10050860) on the serum levels of sTNFRI,
sIL1RII and sIL6R. I found a significant positive correlation of sTNFRI with the
inflammatory markers ESR and CRP. There was no correlation of serum cytokine
receptors with BASDAI.
Thus the reported ERAP1-AS association is unlikely to be related to cytokine
receptor shedding function of ERAP1. There was no difference in the serum levels of
the cytokine receptors in the different ERAP1 genotypic groups. I started with the
hypothesis that decreased receptor shedding can lead to more inflammation as a result
of less blocking receptors in circulation. On the contrary, I found that TNFRI was
higher in patients with higher CRP and ESR. Thus the original hypothesis was rejected.
This observation could be due to more receptor shedding secondary to inflammation,
130
resulting in higher sTNFRI in serum. The strong correlation seen in the current study
between TNFRI and CRP could indicate that TNF receptor shedding is more sensitive to
inflammation than IL1 and IL6 receptors.
Serum cytokine measurements are not straightforward and previous attempts
found these not to be informative in AS [279,280]. It was nevertheless important to
attempt this in the present study. My hypothesis involved an altered balance of
cytokine to its corresponding receptors that lead to inflammation. Hence I tested if the
ratio of serum TNFα, IL6 and IL1 to their corresponding receptors was affected by the
polymorphisms in ERAP1. This also was negative, corroborating my earlier findings.
Unlike humans, mouse ERAP1 has been found only in the ER and has not been
shown to be important in cytokine receptor shedding. The only known function of
ERAP1 in mice is ER peptide processing for MHC loading and as discussed before, the
ERAP1 KO mouse has an abnormal peptide-MHC repertoire at the cell surface [223]. I
have already shown that ERAP1 variants can result in changes in FHC expression in
monocytes of patients with AS. ERAP1 suppression also led to accumulation of FHC in
the intracellular compartment. Peculiar B27-peptide complexes recognized by the
MARB4 antibody were increased following ERAP1 suppression. Thus there is more
evidence that altered peptide processing can explain the AS-ERAP1 association better
than receptor shedding.
Another strong indicator of this possibility is the fact that a haplotype of
ERAP1/ERAP2 is associated with AS[215]. To date ERAP2 has not been shown to be
involved in cytokine receptor shedding. Thus the association of AS with a haplotype of
131
ERAP1/ERAP2 is a strong indicator that ERAP1 influences the pathogenesis of AS by its
peptide processing function.
3.5 Conclusion
The serum soluble TNFRI level correlates with the inflammatory markers ESR
and CRP. ERAP1 polymorphisms, reported previously to be associated with AS, do not
influence the serum cytokine receptor levels in AS patients.
132
Chapter 4
This chapter was published in Arthritis and Rheumatism: Haroon N, Maksymowych WP, Rahman P, Tsui FWL, O’shea FD and Inman RD. Radiographic severity in ankylosing spondylitis is associated with polymorphism in Large Multifunctional Peptidase 2 (LMP2) in the SPARCC cohort. Arthritis Rheum 2011 (Epub). WPM, RDI and O’Shea FD helped in recruiting patients. PR helped with genotyping the patients. RDI guided planning the study. All authors approved the draft. I recruited patients, scored the X-rays for damage assessment, performed all statistical analysis and prepared the draft manuscript.
Radiographic severity in ankylosing spondylitis: 4The effect of genetic polymorphisms in ERAP1 and other elements of the antigen presenting machinery
4.1 Background
Ankylosing Spondylitis (AS) is a chronic inflammatory joint disease
predominantly affecting the spine leading to pain and disability [281]. In addition to
inflammation of the spine, there is progressive new bone formation often leading to
ankylosis of the spine. New bone formation is a hallmark of AS but pathogenesis is not
well understood.
A disconnect between pain and radiographic manifestations is well known.
Patients with severe pain may not have any x-ray changes while those with extensive
ankylosis may be asymptomatic. Although it is proposed that inflammation sets in first
followed by new bone formation, the link between inflammation and new bone
formation is not clear[180,181,282].
Previous attempts at identifying predictors of spinal fusion and radiographic
progression have met with variable success. DKK-1 and a dysregulated Wnt signaling
pathway is considered to be key to new bone formation in AS[188,192,199,283]. Low
133
serum sclerostin levels have been reported to be linked to structural damage [284].
Urinary C-terminal cross-linking telopeptide of type II (CTX-II) is elevated in patients
with more severe radiographic damage at baseline and more progression on follow up
[285,286]. Serum matrix metalloproteinase 3 (MMP3) was found to be an independent
predictor of radiographic progression in one study, especially in patients with pre-
existing radiographic damage [287]. The best predictor of progression in AS so far is
not a biomarker but rather the presence of syndesmophytes at baseline [288].
The strong association of AS with HLA B27 was identified in 1973. Despite the
strong genetic association with AS, there is no definite evidence that HLA-B27 affects
the severity of AS as assessed by radiographic progression [146]. However, these
studies are confounded by the high prevalence of HLA-B27 in the AS population. Thus
other genetic markers are required to predict radiographic severity. Identification of
such a genetic marker would not only aid in identifying patients at higher risk of
radiographic progression, but also shed light on the pathogenesis of new bone
formation in AS. HLA B27 has an attributable risk (AR) of only 16-50% in AS and
clearly other genes are involved in the pathogenesis of the disease [144,260].
Recurrence risk modeling studies have shown that AS is probably an oligogenic disease
[144]. As discussed above, several recent genetic studies on AS have reported an
association with the gene ERAP1 and this has been replicated in Canadian AS patients
as well as in North American multiplex families [127,137,210-213,215,289]. With an
attributable risk of 26%, ERAP1 is second only to HLA B27 among genetic risk factors
reported to be associated with AS [260].
134
Polymorphisms of other genes of the antigen presentation pathway including
the proteasomes and the transporter associated with antigen processing (TAP) have
been variably reported to be associated with AS or with extra articular manifestations
of AS [290-305]. Genes strongly associated with a disease could also be a marker of
severity. This study is an attempt to look at the effect of ERAP1 and other genes
involved in antigen processing on radiographic severity and progression in AS. For
studies involving multiple genes, the size of the cohort is important. I have
collaborated with the Edmonton Spondylitis Clinic under Dr. Walter Maksymowych in
this study.
4.2 Methods
4.2.1 Patients
Caucasian AS patients, diagnosed by the modified New York criteria, attending
the Spondylitis clinics in Toronto and Edmonton were included in the study [68]. All AS
patients were followed annually with a standardized protocol and clinical information
was recorded in a common web-based database. Patients included in the progression
study should have had at least two full sets of x-rays (lumbar and cervical) at a
minimum gap of 1.5 years that could be scored for radiographic severity. Since it is
difficult to detect progression in patients with complete or near-complete fusion of the
entire spine, patients with a baseline mSASSS score of 65 or higher were excluded
from the study.
135
4.2.2 Radiographic Scoring
Radiographic severity of AS was scored using the modified Stoke’s ankylosing
spondylitis spine score (mSASSS) [306]. In brief, all patients had x-ray of the lumbar
and cervical spine in the lateral view for scoring. Each vertebral corner from C2 lower
end to T1 upper end in the cervical and T12 lower end and S1 upper end in the lumbar
spine were scored. With a total of 12 vertebral edges in the cervical and lumbar spine
each, 24 corners could be scored, each ranging from 0 to 3 points. Thus the range of
the total mSASSS score is 0 to 72 for all 24 corners. The scoring is done as given in
the box below and shown in figure 4.1.
Vertebral Corner Score
Normal 0
Squaring or Sclerosis or Erosion 1
Syndesmophyte 2
Bridging syndesmophytes 3
In each center, paired X-rays in a known temporal profile were scored by two
blinded readers, by a consensus method. The inter-class correlation coefficients (ICC)
between the centers were 0.94 and 0.57 for the status and change mSASSS scores.
The smallest detectable change (SDC) was 1.95 for Edmonton and 2.73 for Toronto.
Radiographic progression was assessed by noting the change in mSASSS scores
(∆mSASSS) between the first and last available sets of X-rays. A Damage Duration
136
The scoring method for the Modified Stoke’s Ankylosing Spondylitis Spine Score
(mSASSS) is shown with the major abnormalities that are scored. Squaring is defined
as an absence of any gap between the anterior border of the vertebral body and a line
drawn between the upper and lower anterior corners of the vertebral body. Sclerosis is
the hyperdensity seen as white patches at the vertebral corners. Erosion is a break in
the cortical bone at the corner. A syndesmophytes is a bony spur (new bone)
developing from the vertebral corner and progressing parallel to the axis of the spine.
On the contrary, an osteophyte is not scored which is also a bony spur, but
progressing at right angles to the axis of the spine. When the syndesmophytes bridges
from one vertebral corner to the next, the highest score of 3 points is given to each of
the corners involved.
Figure 4.1 Modified Stoke’s Ankylosing Spondylitis Spine Score
137
Index (DDI) was calculated as the ratio of ∆mSASSS to the gap in years between the
x-rays, which gives an average increase per year. Progressors were identified as those
patients who had an increase of at least 1 mSASSS unit per year (DDI 1) as this is
the average rate of mSASSS progression reported.
4.2.3 Genotyping
DNA was prepared from the peripheral blood of AS patients and shipped to the
Memorial University, Newfoundland, Canada. Genotyping for a panel of 13 coding-
region SNPs in the ERAP1 (N=6), Large Multifunctional Peptidase 2 (LMP2; N=1),
LMP7 (N=1), TAP2 (N=4) and TAP7 (N=1) genes (Table 4.1) was done here. In
previous studies, the SNPs rs27044, rs30187 and rs10050860 in the ERAP1 gene were
shown to be significantly associated with AS [137,289]. The SNP rs26653 was found to
be associated with AS subsequently in Canadian patients and this was not reported in
the initial genome wide studies. It was added to the analysis in this study which
included patients from two Canadian centers and was not included in the previous
studies on PBMCs or cytokine receptors. The remaining SNPs (rs26618, rs3734016,
rs241447, rs4148876, rs2228396, rs1800454, rs2071543, rs1057141, and rs17587)
were chosen based on previous data demonstrating haplotype interaction between the
ERAP1 and TAP /LMP loci [307].
Minor allele frequency (MAF) refers to the frequency of the minor allele in the
cohort and if the frequency is very low, statistical analysis is not reliable. Alleles with
MAF lower than 5% are not included in statistical analysis. The minor allele frequency
at all loci I this study was above 5%. Another quality control measure required for
138
Table 4.1 Genotype frequency of the tested SNPs in the study patients
Single Nucleotide
Polymorphism (SNP) Gene Genotype
Frequency of
Genotype
rs26618 ERAP1 T/TC/C 0.58/0.36/0.06
rs26653 ERAP1 G/GC/C 0.40/0.52/0.08
rs27044 ERAP1 C/CG/G 0.44/0.48/0.08
rs30187 ERAP1 C/CT/T 0.30/0.57/0.13
rs10050860 ERAP1 C/CT/T 0.69/0.28/0.03
rs3734016 ERAP1 G/GA 0.94/0.06
rs17587 LMP2 G/GA/A 0.36/0.57/0.07
rs2071543 LMP7 C/CA/A 0.70/0.27/0.03
rs1057141 TAP1 A/AG/G 0.72/0.27/0.01
rs1800454 TAP2 G/GA/A 0.78/0.19/0.03
rs2228396 TAP2 G/GA/A 0.86/0.13/0.01
rs4148876 TAP2 C/CT/T 0.89/0.10/0.01
rs241447 TAP2 A/AG/G 0.48/0.43/0.09
139
genetic studies is to ensure that alleles comply with the Hardy-Weinberg Equilibrium
(HWE). The Hardy-Weinberg principle states that frequencies of alleles remain
constant in the absence of major genetic changes like mutation, migration or genetic
drifts seen with epidemics. This can be tested by looking at the genotype frequency
and the allele frequency. Based on the allele frequency, we can estimate the genotype
frequency by a formula. If ‘A’ is the major allele and ‘a’ is the minor allele with
frequencies of x and y respectively, the expected genotype frequencies for [AA], [Aa]
and [aa] are x2, 2xy and y2, respectively. This expected frequency is then compared to
the observed frequency in the study cohort to see if the alleles are in HWE. No marker
in our study deviated significantly from the Hardy-Weinberg Equilibrium at the 1%
significance level. If the HWE is not satisfied, it could refer to some form of error like
systematic genotyping error or the presence of homologous regions in the genome.
The MassARRAY system (Sequenom, San Diego, CA) was used for genotyping and the
genotypes were determined using MassARRAY Typer software, version 4.0.
4.2.4 Statistical Analysis
All information on genotypes and clinical data on patients included were
compiled in Toronto. I managed this database and performed the statistical analysis.
Regression analysis is the standard statistical analysis used to predict an outcome
(dependent variable) based on several independent variables. For continuous
dependent variables like mSASSS scores, a linear regression analysis is done while a
binary logistic regression analysis is done for dichotomous dependent variables. Linear
regression analysis was done to identify the predictors of baseline mSASSS scores.
140
Binary logistic regression analysis was done for predictors of progression. Univariate
followed by multivariate regression analysis was done. This is the standard
methodology adopted for regression analysis where the interesting independent
variables are individually tested to see if they predict the outcome. Those variables
which independently predict the outcome are subsequently included in a multivariate
analysis where each variable included is corrected for the other included variables.
Thus variables that are significantly predicting outcome in a multivariate analysis would
correct for all the variables that are individually predicting outcome.
Variables significant at p ≤ 0.1 in the univariate analysis were included in the
multivariate regression model and the final model was obtained after forward method
conditioned on including variables significant at p ≤ 0.05. A forward conditional
method of multivariate regression analysis is a method in which the most significantly
associated variable is included in the regression equation first. Subsequently variables
are added to the equation only if they are significantly contributing to the prediction
equation at a significance level that we decide (0.05 or 5.00 % is standard and was
used in this case).
Apart from the SNPs tested, gender, HLA B27, mSASSS at baseline and duration
of disease were included in the univariate regression models. Considering the high
correlation between age of patients and disease duration (R=0.78; p< 1 x 10-10), age
was not included in regression models. Receiver operating characteristic (ROC) curve
and area under curve (AUC) analysis was done with the logistic regression model
obtained by entering baseline mSASSS alone.
141
ROC curves are graphical X-Y plots representing the true positive rate
(sensitivity) against the false positive rate (1 – specificity) for a binary outcome. ROC
curves can be used to identify the best cut off values to discriminate the two outcomes
by plotting the sensitivity and 1-specifity rates for the different cutoffs. Thus a curve is
generated with the points yielding highest sensitivity being higher on the Y-axis and
values with the highest specificity being on the left hand side of the X-axis. Thus the
value that lies to the left upper corner would yield the highest sensitivity and specificity
for distinguishing the binary outcome. When a statistical model is tested in this way,
the AUC can be calculated. When the values are leaning toward the left upper corner,
the AUC will be higher while models yielding curves that are shifted to the lower right
side will have lower AUC. Thus a regression model with higher AUC indicates better
prediction of the outcome (radiographic progression in this case). For interpreting the
va lue of the model, R2 is the coefficient of determination used in the context of
statistical models. The value indicates how well the prediction of future outcomes can
be done with the information included in the model. Naegelkerke R2 indicates the
proportion of the variation explained by the regression model.
In a dominant model of genetic statistical testing, the minor variant has a
dominant allele effect with the group heterozygous for the minor allele differing
significantly in the outcome from the group homozygous for the major allele. In the
recessive model only the group homozygous for the minor allele differs from the other
groups. Dominant and recessive models were initially tested and only recessive models
were significant. For the SNP rs3734016, no patients were homozygous for the minor
allele and so the dominant model was used in the initial univariate analysis.
142
Contingency 2X2 tables were analyzed with Chi-squared tests (χ2) and odds
ratio (OR) of the relevant SNPs for progression were calculated. Student’s t test was
used for comparison of means. The PASW Statistics 18 software was used for the
analysis.
4.3 Results
4.3.1 Patients
A total of 241 AS patients (81% males and 82% HLA B27-positive) were
enrolled from the two centers and were evaluated for predictors of baseline
radiographic severity. Out of this, 210 patients had follow up X-rays at a minimum gap
of 18 months and were included for studying predictors of radiographic progression.
The patients were followed up for a mean (± SD) duration of 2.4 (± 0.8) years. There
were 158 patients enrolled from Edmonton and 83 patients from Toronto (Table 4.2).
The Edmonton cohort was older (43.0 ± 12.7 vs 37.1 ± 12.5; p=0.001), with
longer duration of disease (15 ± 9.7 vs 19 ± 12; p=0.01) and the proportion of HLA
B27 positive patients was higher (87% vs 76%; p=0.04). The baseline mSASSS, rate
of progression and the proportion of progressors was not different between the
centers.
143
Table 4.2 Demographic features of the two AS cohorts
Variable Combined Cohort Toronto Edmonton p*
Patients (N) 241 83 158
Age at baseline 41.0 ± 13.0 37.1 ± 12.5 43.0 ± 12.8 0.001
Females (%) 18.7 14.5 20.9 ns
HLA B27 (%) 82.3 75.6 86.5 0.04
Baseline mSASSS 16.1 ± 17.6 13.6 ± 18.2 17.74 ± 17.2 ns
Disease duration at
baseline
17.6 ± 11.5 14.9 ± 9.7 19.1 ± 12.2 0.01
DDI 0.81 ± 1.3 0.58 ± 0.9 0.93 ± 1.5 ns
Progressors (%) 29.0 22.0 32.0 ns
Values are mean ± SD unless otherwise indicated in brackets.
mSASSS: modified Stoke’s ankylosing spondylitis spine score
DDI: damage duration index
ns: not significant.
DDI was calculated by dividing the change in mSASSS (∆mSASSS) by the gap (in years) between the x-rays, which gives an average increase in mSASSS per year.
Progressors were identified as those patients who had an increase of at least 1
mSASSS unit per year (DDI 1) as this is the average rate of mSASSS progression reported.
* P values are reported for Student’s T test in case of continuous variables and for Chi-squared test with categorical variables.
144
4.3.2 Baseline radiographic severity can be influenced by polymorphisms in LMP2
Detailed regression analysis was done to assess the predictors of baseline
radiographic severity in the cohort. Univariate followed by multivariate regression
analysis was done. As age at baseline and duration of disease are correlated highly
(R=0.78; p< 1 x 10-10), age was not included in regression models. Univariate
regression analysis showed that baseline mSASSS scores were associated with gender,
duration of disease, the ERAP1 SNP rs30187 and the LMP2 SNP rs17587 (Table 4.3).
Although, by T-test, HLA B27–positive patients had significantly higher mSASSS scores
than HLA B27-negative AS patients (20.4 ± 22.2 vs 11.9 ± 17.8; p=0.01), HLA B27
was not statistically significant in the regression model (B = 5.21;p=0.09). Baseline
mSASSS scores were higher by T-test in males than females (21.6 ± 22.69 vs 12.23 ±
16.44, p=0.002). Patients homozygous for the minor allele of the LMP2 SNP rs17587
(AA) had significantly higher mSASSS at baseline (27.2 vs 15.7; p=0.025). AS patients
with the major allele of the ERAP1 SNP rs30187 (CC/CT) had significantly higher
baseline radiographic severity (16.7 vs 9.4; p=0.04).
In multivariate analysis with the forward conditional method, duration of
disease (B=0.74; p = 1 x 10-11), gender (B = 12.1; p= 5x10-5) and the LMP2 SNP
rs17587 (B=6.2; p=0.01) were significantly associated with baseline mSASSS (Table
4.3). A forward conditional method of multivariate regression analysis is a method in
which the most significantly associated variable is included in the regression equation
first. Subsequently variables are added to the equation only if they are significantly
contributing to the prediction equation at a significance level that we decide (0.05 or
145
Table 4.3 Univariate analysis of predictors of baseline radiographic severity
Variables
Linear Regression Analysis
Univariate Analysis Multivariate Analysis*
B 95% C.I. of B Beta Sig. B 95% C.I. of B Beta Sig.
Duration of disease 0.67 0.50 to 0.85 0.45 <1x10-12
0.74 0.53 to 0.95 0.5 1x10-11
HLA B27 5.21 -0.9 to 11.4 0.12 0.09
Sex 6.28 0.4 to 12.1 0.14 0.027 12.1 6.4 to 17.8 0.3 5x10-5
LMP2: rs17587
5.74 0.7 to 10.7 0.16 0.025 6.2 1.5 to 10.8 0.2 0.01
LMP7: rs2071543
-6.5 -22.3 to 9.4 -0.06 0.42
ERAP1: rs26618
1.89 -3.1 to 6.9 0.05 0.46
rs26653 0.17 -4.4 to 4.7 0.005 0.94
rs27044 -1.5 -5.8 to 2.9 -0.04 0.53
rs30187 -3.63 -7.0 to -0.2 -0.14 0.037
rs3734016 -5.2 -16.2 to 5.9 -0.07 0.36
rs10050860 2.6 -2.0 to 7.2 0.08 0.27
TAP1: rs1057141
-1.6 -6.9 to 3.7 0.04 0.55
TAP2: rs241447
3.05 -1.2 to 7.3 0.10 0.16
rs1800454 -0.56 -14.9 to 13.8 -0.01 0.94
rs2228396 15.2 -9.9 to 40.3 0.09 0.23
rs4148876 6.72 -1.10 to 14.5 0.12 0.09
* Multivariate analysis with forward conditional method, where only significant variables in the regression model are included in the equation.
146
5.00 % is standard and was used in this case). Increased disease duration, male
gender and the presence of the major allele of the LMP2 SNP rs17587 made it more
likely to have higher baseline mSASSS scores. Disease duration alone accounted for
20% of the variability in baseline mSASSS. The sequential addition of gender and the
rs17587 SNP data independently improved the model to account for 29% and 32% of
the variability respectively.
4.3.3 Baseline mSASSS is the strongest predictor of progression
The mean increase in mSASSS scores in the entire cohort was 1.77 ± 2.9
mSASSS units. A minimum gap of 1.5 years between X rays was present in 210
patients who were included in the progression study. The mean increase in mSASSS
scores in this cohort was 1.95 ± 3.1 mSASSS units and 60 (29%) patients had
progression of mSASSS scores at a minimum rate of 1 mSASSS unit per year during
the follow up period.
The patients in the baseline and progression studies were similar in
demographics and baseline mSASSS (data not shown). As evident from the cumulative
probability plot (Figure 4.2), mSASSS change of 1, 2 and 4 mSASSS units were seen
in 98 (46.6%), 84 (40%) and 43 (20.5%) patients respectively. In univariate analysis
only baseline mSASSS (OR=1.03; p=0.001) and duration of disease (OR=1.03;
p=0.02) predicted progressors (Table 4.4). In multivariate analysis, only the baseline
mSASSS was significant in the final model (Table 4.4). The odds of progression with
every unit increase in baseline mSASSS, was 1.03 (CI: 1.01 to 1.06; p=0.003). For
147
Cumulative probability plot showing the progression of mSASSS scores in 210 patients
included in the progression study. On the y-axis, the change in x-ray score over the
follow up period is shown while the x-axis shows the percentage of patients with the
corresponding change. The dotted lines traced from the y axis correspond to
progression of 1, 2 and 4 mSASSS units. The corresponding perpendicular dashed
lines to the x-axis indicate the percentage of patients with the respective mSASSS
progression. Thus 53% of patients had no change in mSASSS scores and 80% had
less than 4 mSASSS units change. Increase in radiographic score of 1, 2 and 4
mSASSS units were seen in 98 (46.6%), 84 (40%) and 43 (20.5%) patients
respectively
Figure 4.2 Cumulative Probability Plot of patients with progression of
radiographic scores
148
Table 4.4 Predicting patients who are likely to progress on follow up
Variables
Univariate Analysis Multivariate Analysis
B Sig. Beta
95% C.I.for
Beta
B Sig. Beta
95% C.I.for
Beta
Lower Upper Lower Upper
Duration of disease 0.03 0.02 1.03 1.01 1.06
mSASSS_baseline 0.03 0.001 1.03 1.01 1.04 0.03 0.003 1.03 1.01 1.06
B is the co-efficient of the constant in the model (also known as intercept). Beta gives
the odds ratio for the variable in the model. Thus a Beta of 1.03 means for every
increase of 1 unit in the mSASSS score at baseline, the odds of progression increases
by 1.03 times. Thus compared to a patient with an mSASSS score of zero at baseline,
patients with scores of 10 and 20 respectively have a 1.3 and 1.8 times increased
chance of progressing at a rate of 1 mSASSS unit per year.
Sig. (significance) is the p value for the variable in the equation.
Only duration of disease and mSASSS at baseline were significant in the univariate
analysis at p ≤ 0.1 and included in the multivariate analysis.
149
increasing the threshold to identify progressors, they were defined variably as those
with an increase of at least 2 or 4 mSASSS units. Even with this change in definition,
baseline mSASSS continued to be the only predictor of progression in both cases.
Receiver operating characteristic (ROC) curve and area under curve (AUC)
analysis was done with the logistic regression model obtained by entering baseline
mSASSS alone. ROC curves are graphical X-Y plots representing the true positive rate
(sensitivity) against the false positive rate (1 – specificity) for a binary outcome. ROC
curves can be used to identify the best cut off values to discriminate the two outcomes
by plotting the sensitivity and 1-specifity rates for the different cutoffs. We can use the
ROC curve to test the value of our statistical model. If the AUC is high (models yielding
curves that are shifted to the left) the regression model is good at predicting the
outcome (radiographic progression in this case). Naegelkerke R2 is calculated and it
indicates the proportion of the variation explained by the regression model.
ROC curve analysis (Figure 4.3) showed that the regression model with
baseline mSASSS alone had a modest AUC of 0.68 (p=1x10-4) and the Naegelkerke R
square (R2) estimates showed that baseline mSASSS alone explained only 7% of the
variability in progression. Hence there are several important variables yet to be
identified that can influence radiographic progression. R2 is the coefficient of
determination used in the context of statistical models. The value indicates how well
the prediction of future outcomes can be done with the information included in the
model. Thus in this case, other factors that explain 93% of the variability are yet to be
identified.
150
The Receiver Operating Characteristic (ROC) curve analysis shows that the AUC for
the regression model with baseline mSASSS alone is 0.68 (C.I. 0.60-0.76; p = 1x 10-
4). As the AUC increases, the ROC curve would shift more toward the upper left
corner, thereby indicating a model with more variables that predict the outcome.
Naegelkerke R square (R2) which estimates the variability of the outcome (radiographic
progression) showed that baseline mSASSS alone explained only 7% of the variability
in progression. Hence other important predictor variables for progression are yet to be
identified.
Figure 4.3 ROC Curve for regression model with baseline
mSASSS predicting radiographic progression
151
4.4 Discussion
The identification of ERAP1 as one of the strongest genetic associations in AS,
second to only HLA B27, has triggered renewed interest in studies on the biologic basis
of AS. Although a strong genetic association with AS has been established for ERAP1
any influence on the severity of the disease is unknown. This is the first study looking
at the effect of ERAP1 and other genes involved in the antigen presentation pathway
on radiographic severity and progression in AS patients. The ERAP1 SNP rs30187
significantly predicted the baseline radiographic severity in univariate analysis, but
failed to do so in a multivariate model. I found that the LMP2 SNP rs17587, duration of
disease and gender are independently associated with baseline radiographic severity of
AS patients. The baseline mSASSS was found to be the best predictor of radiographic
progression.
It is evident from the initial analysis that the two cohorts are quite distinct. The
Edmonton cohort is larger, older and with higher HLA B27 prevalence. Though not
statistically significant, there were more progressors in the Toronto cohort and more
rapid progressors in the Edmonton cohort. Combining two cohorts with different
baseline characteristics can negate the effects of genetic variants on an outcome. It
should also be noted that the two cohorts are from two distinct geographic locations in
Canada and may be subjected to varying gene-environment interactions further
affecting the outcome. Despite these problems, the combined analysis of the cohorts
indeed showed remarkable consistency with the cohort specific analysis. The best
predictors of baseline radiographic severity in univariate analysis of the combined
152
cohorts were disease duration and gender. Indeed these same variables have
remained important in the two cohorts separately. More over the SNPs which predicted
baseline radiographic severity in the combined cohort were rs30187 and rs17587, the
same SNPs predicting progressors in the Toronto and Edmonton cohorts respectively.
HLA-B27 was the next most important variable in the regression analysis but had a p-
value of 0.09 and was included in the multivariate analysis.
For severity and long term damage assessment, mSASSS scoring for x-ray
changes is the preferred method. Baseline radiographic severity is a marker of long
term damage in patients with AS. Disease duration was the strongest predictor of
radiographic severity and by itself accounted for 20% variability in the data. It is
known that females tend to have less radiographic severity [308]. This was reflected in
this study as well. The gender of the patients explained an additional 9% variability in
mSASSS scores over and above disease duration.
After controlling for disease duration, HLA B27 status and gender, the LMP2
SNP rs17587 was independently associated with radiographic severity at baseline.
LMP2 has been variably reported to be associated with AS and/or extra articular
manifestations of AS especially uveitis [290-295,297,298]. The recent genome wide
scans on AS however failed to identify a link to LMP2.
In line with previous reports on predictors of progression, baseline radiographic
severity stands out as the best predictor of future progression. With an increase of 1
unit in the mSASSS score at baseline, the odds of progression increased by 1.03 times.
Thus compared to a patient with an mSASSS score of zero at baseline, patients with
153
scores of 10 and 20 respectively have a 1.3 and 1.8 times increased chance of
progressing at a rate of 1 mSASSS unit per year.
The rs17587 LMP2 SNP though significantly affecting baseline radiographic
severity explained only an additional 3% of variability in the mSASSS scores and it did
not significantly affect radiographic progression. It is quite evident that the effect of
individual SNPs in radiographic progression is small and a number of genetic factors
remain to be discovered [309]. The small effects also mean that larger cohorts and
longer follow up is required to see significant influences.
Changes in antigen presentation can affect immune responses and thereby
cause inflammation. Inflammation may be associated with new bone formation
[180,181,282] but the treatment of AS patients with anti-TNF agents have not
consistently shown inhibition of radiographic progression [310-312]. The SNP rs17587
results in a change in the structure of the immunoproteasome Large Multifunctonal
Peptidase 2 (LMP2) with an R60H substitution. There is no published data on the effect
of this change on function of proteasomes. The proteasome, like ERAP1 is important in
antigen presentation and is upregulated by IFN-γ [313,314]. Apart from the role in
antigen presentation and thereby on inflammation, proteasomes have a well-
established direct link to bone metabolism. LMP2 has been shown to regulate NF-kB
levels and thus can affect osteoclastic activity [315-318].
Proteasomes are important in osteoblast activity by modulating the Wnt/
Catenin pathway[319]. The Wnt/ Catenin pathway is considered to be overactive in
AS due to the low level and/or dysfunctional activity of Dicckopf 1 (DKK-1), a natural
154
inhibitor of this pathway[188,283]. Blockade of DKK-1 results in fusion of the sacroiliac
joints of TNF transgenic mice [320]. When the Wnt signaling pathway is activated,
catenin, a transcriptional co-activator, accumulates and is translocated to the
nucleus[321]. This catenin, which is normally degraded by the proteasome, can
accumulate in excess when the proteasome is abnormal resulting in excess
osteoblastic activity and new bone formation [319,321]. A hypothesis on the role of
LMP2 in radiographic progression is shown in figure 4.4. ERAP1 variants can have an
indirect effect on NFkB levels by affecting the unfolded protein response. UPR has
been shown to affect osteoclast differentiation via NFkB activation [322]. NFkB binding
regions are seen in the promoter of ERAP1 helping in regulating ERAP1 expression
[323].
In general AS is a slowly progressive disease. It is estimated that within a period
of 10 years, radiographic progression can be linear [324]. However this rate of
progression varies with follow up. It has been reported that 30% of radiographic
progression occurs during the first decade of disease followed by 40% in the next
decade and finally 35% in the 3rd decade [324]. Two things are quite evident here.
Studies of radiographic progression ideally should have long follow up, and depending
on the duration of disease, the rate of radiographic progression may vary. A larger
cohort followed over longer periods of time could help us identify additional factors
with subtler effects on radiological progression in AS. Studies including haplotype
analysis and gene-gene interactions with other AS-asociated genes like IL-23R are
warranted.
155
Abnormal proteasome LMP2 variant could affect osteoclast function by affecting NF-kB
levels. There is no published data on the effect of the LMP2 variant studied (R60H) on
the function of proteasomes. LMP2 variants can decrease degradation of -catenin in
osteoblasts and result in osteoproliferation. The increased levels of -catenin in turn
can decrease the expression of RANKL on osteoblasts and indirectly further decrease
osteoclast function.
LMP2: Large Multifunctonal Peptidase 2; NF-kB: Nuclear Factor kappa B; RANK:
Receptor activator of NF-kB; RANKL: Receptor activator of NF-kB Ligand; LRP: LDL
Receptor Protein 5
Figure 4.4 Hypothetical mechanism for LMP2 and ERAP1 variants affecting
radiographic progression in AS
156
4.5 Conclusions
This is the first study to demonstrate that LMP2 variants can affect radiographic
severity in AS. ERAP1 could affect radiographic severity, but larger studies are needed
to explore this further. Baseline mSASSS remains the strongest predictor of
radiographic progression in AS but explains only a fraction of the variability seen.
157
Chapter 5
Conclusions and Future Directions 5
5.1 Conclusions
The body of work presented here is an attempt at understanding the functional
and clinical relevance of the recently reported association between ERAP1 and
ankylosing spondylitis. The results of the functional studies are the first report of a
functional implication of ERAP1 polymorphisms in patients with AS. It is also the first
attempt at examining at the effect of ERAP1 polymorphisms on radiographic severity of
AS.
I have demonstrated that the rs27044 coding polymorphisms in ERAP1 which
leads to a Q730E amino acid change results in significant variation of the FHC
expression on the surface of monocytes. AS patients with only the minor allele of this
SNP have significantly lower surface FHC compared to those with the major allele.
There was no significant effect on the B cell FHC expression. The total HLA-B27
expression in the whole PBMC population was lower in patients with the minor allele,
but this was not statistically significant. There was no effect of the rs30187 SNP on
FHC or total HLA-B27.
The ERAP1 SNPs studied here did not show any significant effect on the serum
cytokine or cytokine receptor levels for IL-I, IL-6 and TNFα. Thus the cytokine receptor
shedding function of ERAP1 appears not to be a significant contributor in the
pathogenesis of AS.
158
Suppression of ERAP1 levels by siRNA can affect both the conformation of HLA
B27 and the peptide display on HLA-B27. Furthermore, these properties appear to be
differentially affected in cells expressing AS-associated HLA-B27 subtypes as opposed
to those not associated with AS.
The ERAP1-B*27 interaction could be the missing link in the pathogenesis of AS.
The recent evidence emerging on a genetic interaction between ERAP1 and HLA-B27 is
additional evidence of a possible functional interaction and could be pathogenic in AS.
The ERAP1 association with AS is seen only in HLA-B27 positive AS patients. Similarly,
in psoriasis ERAP1 association is seen only in patients with the HLA-C risk allele. The
association between MHC-I and ERAP1 seen in genetic studies strengthens the
evidence of functional interaction that I have shown. Patients who had the Q730
variant of ERAP1, that has been shown to have higher aminopeptidase activity than
the E730 variant, had lower FHC. Hence it seems that ERAP1 functioning results in
properly formed MHC-I molecules with less FHC formation. It is evident from the
population studies that the common variant of ERAP1 (E730) seen in the majority of
normal population and AS patients has less aminopeptidase activity than the minor
variant (Q730). A gain of function would thus lead to more MHC-peptide complexes
being formed. Putting all the observations together, it seems the pathogenic event is a
gain of function with resulting increase in MHC-peptide complexes.
The FHC hypothesis cannot be entirely excluded, although a higher FHC
expression would be expected in controls compared to AS patients. FHC can form
dimers that are recognized by inhibitory Killer Immunologublin receptors (KIR)
159
KIR3DL1 and KIR3DL2 [167]. Thus a loss of function could be protective by increased
inhibitory signals through the KIRs. The β2m overexpression study is a great example
of the complexity of AS pathogenesis. If FHC are pathogenic, by stabilizing HLA-B27,
disease should be ameliorated. HLA-B27 was stabilized by β2m overexpression in
HLA-B27 transgenic rats which paradoxically resulted in the arthritis worsening and
colitis improving [174]. Worsening of arthritis with minimal effect on colitis was seen
by another experiment that resulted in increased MHC-peptide complexes. A B27-
specific influenza peptide was overexpressed in HLA-B27 transgenic rats which led to
more arthritis but no effect on colitis [175]. This would suggest that the pathogenesis
of arthritis is peptide mediated while that of colitis could be due to the UPR.
The discovery of the ERAP1-AS association has generated a lot of interest in the
pathogenesis of AS. With high throughput studies and an explosion of molecular
biology and other translational research techniques, we are closer to the answer than
ever before.
5.2 Future Directions
The results obtained and elaborated in previous chapters raise several
questions. ERAP1 can potentially influence FHC as well as HLA-B27-peptide complexes
recognized by MARB4 antibody. It needs to be further explored as to how this
interaction could be potentially pathogenic in AS.
160
I am planning to take this study to the next step and will attempt to answer the
interesting questions raised here. I hope this attempt will help take our understanding
of the pathogenesis of AS further.
5.2.1 Studies in Ankylosing Spondylitis Patients
5.2.1.1 ERAP1 variants lead to changes in surface FHC. Does this lead to altered KIR binding?
To address this question, peripheral blood mononuclear cells isolated from B27-
positive AS patients will be stained with recombinant human Fc Chimeras (R&D
systems) of KIR3DL1 and KIR3DL2 (both known to bind B27 FHC dimers). The
secondary antibody FITC-anti human IgG (Fc gamma specific) will be used and KIR
binding assessed by mean fluorescence intensity (MFI) values in flow cytometry
(FACS). The antigen presenting cells monocytes and B cells will be separately gated
with anti-CD14 PE and anti-CD-19 APC. Patients will be genotyped using the allelic
discrimination assay for the rs30187 and rs27044 SNPs of ERAP1. Differences in KIR
binding between patients with different ERAP1 variants will be assessed.
5.2.1.2 Does ERAP1 polymorphism affect NK cell function?
Monocytes will be isolated using the antiCD-14 MACS separation system
(Miltenyl Biotech, Germany) from AS patients genotyped for rs27044 and rs30187
ERAP1 SNPs. The 7-AAD/CFSE assay (Cayman Chemicals, USA) will be used to study
NK cell cytotoxicity. Briefly, monocytes will be co-incubated with YTS cells (NK cell line;
from Dr. Deborah Burshtyn: collaboration letter appended) expressing KIR3DL1, which
are known to bind B27 dimers. Different effector:target ratios will be used with the
161
monocytes pre-incubated with CFSE. The dye 7-AAD will be used for staining the dead
target cells and quantification will be done by FACS. It is expected that the isolated
monocytes expressing B27 dimers can affect NK cell activity of the YTS cells through
the KIR receptors. In parallel, the co-incubated cells will separately be analyzed for
cytokine production by intracellular staining for IFN-, TNF-α and IL-10 followed by
FACS. Thus NK cell cytotoxicity and cytokine production will be compared between
patients with different variants of ERAP1.
5.2.2 In vitro studies
The following system will be established for the in vitro studies. C1R cells stably
transfected with HLA B*2704, 05, 06 and 09 will be used to establish cell lines with
ERAP1 stably knocked down using HIVH1 shRNA clone set targeting human ERAP1
(Puromycin selective, with fluorescent mCherry reporter and all targets at 3'UTR
region, in an HIV based lentiviral vector). Once a stable knock down of ERAP1 is
established and confirmed by western blot analysis, one of the following will be
introduced into the respective C1R cells: either wild type human ERAP1 ORF clone or
one of the human ERAP1 mutant clones, rs30187: amino acid 528 K->R (AAG->AGG)
or rs27044: 730 Q->E (CAA->GAA) in pReceiver-Lv184 (HIV-based lentiviral vector
with CMV promoter, C-3xHA tag and hygromycin selection marker). As the shRNA is
against the 3’UTR, there should not be a problem in expressing ORF cDNA based
ERAP1. The antibiotic selection is based on the fact that C1R cells have Geneticin
selection incorporated already. Negative control vector for pReceiver-Lv184 and HIVH1
shRNA scrambled control clone for psiHIV-H1 will be used as controls.
162
5.2.2.1 Does ERAP1 polymorphism lead to altered unfolded protein response?
Once the C1R cell lines with different B27 subtypes, each stably expressing the
wild type or variant ERAP1 forms are established, the UPR will be assed by a PCR
array. The Human UPR RT² Profiler™ PCR Array (SABiosciences, USA) profiles the
expression of 84 key genes recognizing and responding to misfolded protein
accumulation in the endoplasmic reticulum (ER). Significant differences noted in the
array, between cells expressing different allelic forms of ERAP1 will be confirmed by
western blot.
5.2.2.2 Does ERAP1 polymorphism lead to changes in the MHC-peptide repertoire?
Peptide repertoire variation, if any, in humans in the context of HLA B27 and
different allelic forms of ERAP1 is important considering the arthritogenic-peptide
hypothesis. Here, the MHC-peptide repertoire of HLA B27 expressing C1R cells with
different ERAP1 variants will be studied.
Briefly, the C1R cells with different B27 subtypes, each expressing the AS-
associated and wild type ERAP1 variant will be subjected to MHC-peptide isolation by
the acid elution technique. Cell suspensions will be pelletted and the supernatant
isolated for ultrafiltration. Analysis of the peptidome will be done by mass
spectrometry using a LTQ-Orbitrap mass spectrometer (Thermo Fisher Scientific).
163
5.2.2.3 Does the presence of different ERAP1 variants affect NK cell cytotoxicity and activation? Does this differ depending on the B27 subtype expressed?
Similar to aforementioned experiments in question II, the C1R cells will be co-
incubated with YTS cells in different effector:target ratios and NK cell cytotoxicity and
activation (cytokine production) will be studied. These in vitro studies will corroborate
studies in AS patients and will also help understand if the NK-cell effect varies with the
B27 subtype expressed on the C1R cell.
164
References
References
1. Ozdemir O. Quality of life in patients with ankylosing spondylitis: relationships with
spinal mobility, disease activity and functional status. Rheumatology international 2011
May;31(5):605-610.
2. Ibn Yacoub Y, Amine B, Laatiris A, Abouqal R, Hajjaj-Hassouni N. Health-related
quality of life in Moroccan patients with ankylosing spondylitis. Clinical rheumatology
2011 May;30(5):673-677.
3. Kiltz U, van der Heijde D. Health-related quality of life in patients with rheumatoid
arthritis and in patients with ankylosing spondylitis. Clinical and experimental
rheumatology 2009 Jul-Aug;27(4 Suppl 55):S108-11.
4. Younes M, Jalled A, Aydi Z, Younes K, Jguirim M, Zrour S, et al. Quality of life in
ankylosing spondylitis. La Tunisie medicale 2011 Apr;89(4):374-378.
5. Mustur D, Vesovic-Potic V, Stanisavljevic D, Ille T, Ille M. Assessment of functional
disability and quality of life in patients with ankylosing spondylitis. Srpski arhiv za
celokupno lekarstvo 2009 Sep-Oct;137(9-10):524-528.
6. O'Shea FD, Riarh R, Anton A, Inman RD. Assessing back pain: does the Oswestry
Disability Questionnaire accurately measure function in ankylosing spondylitis? The
Journal of rheumatology 2010 Jun;37(6):1211-1213.
7. Ariza-Ariza R, Hernandez-Cruz B, Collantes E, Batlle E, Fernandez-Sueiro JL,
Gratacos J, et al. Work disability in patients with ankylosing spondylitis. The Journal of
rheumatology 2009 Nov;36(11):2512-2516.
8. Bakland G, Gran JT, Becker-Merok A, Nordvag BY, Nossent JC. Work disability in
patients with ankylosing spondylitis in Norway. The Journal of rheumatology 2011
Mar;38(3):479-484.
165
9. Rohekar S, Pope J. Assessment of work disability in seronegative spondyloarthritis.
Clinical and experimental rheumatology 2010 Jan-Feb;28(1):35-40.
10. Spadaro A, Lubrano E, Massimiani MP, Gaia P, Perrotta FM, Parsons WJ, et al.
Validity, responsiveness and feasibility of an Italian version of the Canadian
Occupational Performance Measure for patients with ankylosing spondylitis. Clinical
and experimental rheumatology 2010 Mar-Apr;28(2):215-222.
11. Madsen OR, Lindhardsen J. Ankylosing spondylitis is associated with increased
cardiovascular morbidity and mortality. Ugeskrift for laeger 2011 May 23;173(21):1503-
1507.
12. Kiltz U, Sieper J, Braun J. Development of morbidity and mortality in patients with
spondyloarthritis. Zeitschrift fur Rheumatologie 2011 Aug;70(6):473-479.
13. Bakland G, Gran JT, Nossent JC. Increased mortality in ankylosing spondylitis is
related to disease activity. Annals of the Rheumatic Diseases 2011 Nov;70(11):1921-
1925.
14. McCarey D, Sturrock RD. Comparison of cardiovascular risk in ankylosing
spondylitis and rheumatoid arthritis. Clinical and experimental rheumatology 2009 Jul-
Aug;27(4 Suppl 55):S124-6.
15. Feldtkeller E, Lemmel EM, Russell AS. Ankylosing spondylitis in the pharaohs of
ancient Egypt. Rheumatology international 2003 Jan;23(1):1-5.
16. BLUMBERG BS, BLUMBERG JL. Bernard Connor (1666-1698) and his
contribution to the pathology of ankylosing spondylitis. Journal of the history of
medicine and allied sciences 1958 Jul;13(3):349-366.
17. Benoist M. Pierre Marie. Pioneer investigator in ankylosing spondylitis. Spine 1995
Apr 1;20(7):849-852.
18. Strumpell EA, Bick EM. Observations on chronic-ankylosing inflammation of the
vertebrae and hip joints. Clinical orthopaedics and related research 1971 Jan;74:4-6.
166
19. FORESTIER J, DESLOUS-PAOLI P. Radiological study of sacro-iliac joints in
ankylosing spondylitis with reference to the evolution of the disease. Annals of the
Rheumatic Diseases 1957 Mar;16(1):31-34.
20. FORESTIER J, JACQUELINE F, ROTESQUEROL J. Ankylosing spondylarthritis
with painless spinal evolution. Revue du rhumatisme et des maladies osteo-articulaires
1951 Sep;18(9):486-487.
21. Lawrence RC, Helmick CG, Arnett FC, Deyo RA, Felson DT, Giannini EH, et al.
Estimates of the prevalence of arthritis and selected musculoskeletal disorders in the
United States. Arthritis and Rheumatism 1998 May;41(5):778-799.
22. Braun J, Bollow M, Remlinger G, Eggens U, Rudwaleit M, Distler A, et al.
Prevalence of spondylarthropathies in HLA-B27 positive and negative blood donors.
Arthritis and Rheumatism 1998 Jan;41(1):58-67.
23. Bakland G, Nossent HC, Gran JT. Incidence and prevalence of ankylosing
spondylitis in Northern Norway. Arthritis and Rheumatism 2005 Dec 15;53(6):850-855.
24. Will R, Edmunds L, Elswood J, Calin A. Is there sexual inequality in ankylosing
spondylitis? A study of 498 women and 1202 men. The Journal of rheumatology 1990
Dec;17(12):1649-1652.
25. Calin A, Elswood J, Rigg S, Skevington SM. Ankylosing spondylitis--an analytical
review of 1500 patients: the changing pattern of disease. The Journal of rheumatology
1988 Aug;15(8):1234-1238.
26. Baraliakos X, Listing J, von der Recke A, Braun J. The natural course of
radiographic progression in ankylosing spondylitis: differences between genders and
appearance of characteristic radiographic features. Current rheumatology reports 2011
Oct;13(5):383-387.
27. Roussou E, Sultana S. Spondyloarthritis in women: differences in disease onset,
clinical presentation, and Bath Ankylosing Spondylitis Disease Activity and Functional
167
indices (BASDAI and BASFI) between men and women with spondyloarthritides.
Clinical rheumatology 2011 Jan;30(1):121-127.
28. Khan MA, Mathieu A, Sorrentino R, Akkoc N. The pathogenetic role of HLA-B27
and its subtypes. Autoimmunity reviews 2007 Jan;6(3):183-189.
29. Sieper J, Rudwaleit M, Khan MA, Braun J. Concepts and epidemiology of
spondyloarthritis. Best practice & research.Clinical rheumatology 2006 Jun;20(3):401-
417.
30. Roberts-Thomson RA, Roberts-Thomson PJ. Rheumatic disease and the
Australian aborigine. Annals of the Rheumatic Diseases 1999 May;58(5):266-270.
31. Hukuda S, Minami M, Saito T, Mitsui H, Matsui N, Komatsubara Y, et al.
Spondyloarthropathies in Japan: nationwide questionnaire survey performed by the
Japan Ankylosing Spondylitis Society. The Journal of rheumatology 2001
Mar;28(3):554-559.
32. Belachew DA, Sandu N, Schaller B, Guta Z. Ankylosing spondylitis in sub-Saharan
Africa. Postgraduate medical journal 2009 Jul;85(1005):353-357.
33. Blanco-Gelaz MA, Lopez-Vazquez A, Garcia-Fernandez S, Martinez-Borra J,
Gonzalez S, Lopez-Larrea C. Genetic variability, molecular evolution, and geographic
diversity of HLA-B27. Human immunology 2001 Sep;62(9):1042-1050.
34. Mbayo K, Mbuyi-Muamba JM, Lurhuma AZ, Halle L, Kaplan C, Dequeker J. Low
frequency of HLA-B27 and scarcity of ankylosing spondylitis in a Zairean Bantu
population. Clinical rheumatology 1998;17(4):309-310.
35. Mijiyawa M, Oniankitan O, Khan MA. Spondyloarthropathies in sub-Saharan Africa.
Current opinion in rheumatology 2000 Jul;12(4):281-286.
36. Peschken CA, Esdaile JM. Rheumatic diseases in North America's indigenous
peoples. Seminars in arthritis and rheumatism 1999 Jun;28(6):368-391.
168
37. Benevolenskaia LI, Boyer D, Erdes S, Templin D, Alekseeva LI, Lorence R, et al.
Comparative study of epidemiology of spondylarthropathies in indigenous population of
the Chukot peninsula and Alaska. Terapevticheskii arkhiv 1998;70(1):41-46.
38. Benevolenskaya LI, Boyer GS, Erdesz S, Templin DW, Alexeeva LI, Lawrence RC,
et al. Spondylarthropathic diseases in indigenous circumpolar populations of Russia
and Alaska. Revue du rhumatisme (English ed.) 1996 Dec;63(11):815-822.
39. Boyer GS, Templin DW, Bowler A, Lawrence RC, Heyse SP, Everett DF, et al.
Class I HLA antigens in spondyloarthropathy: observations in Alaskan Eskimo patients
and controls. The Journal of rheumatology 1997 Mar;24(3):500-506.
40. Boyer GS, Templin DW, Bowler A, Lawrence RC, Heyse SP, Everett DF, et al.
Spondyloarthropathy in the community: clinical syndromes and disease manifestations
in Alaskan Eskimo populations. The Journal of rheumatology 1999 Jul;26(7):1537-
1544.
41. Gofton JP, Chalmers A, Price GE, Reeve CE. HL-A 27 and ankylosing spondylitis
in B.C. Indians. The Journal of rheumatology 1984 Oct;11(5):572-573.
42. Calin A, Porta J, Fries JF, Schurman DJ. Clinical history as a screening test for
ankylosing spondylitis. JAMA : the journal of the American Medical Association 1977
Jun 13;237(24):2613-2614.
43. Rudwaleit M, Metter A, Listing J, Sieper J, Braun J. Inflammatory back pain in
ankylosing spondylitis: a reassessment of the clinical history for application as
classification and diagnostic criteria. Arthritis and Rheumatism 2006 Feb;54(2):569-
578.
44. Sieper J, van der Heijde D, Landewe R, Brandt J, Burgos-Vagas R, Collantes-
Estevez E, et al. New criteria for inflammatory back pain in patients with chronic back
pain: a real patient exercise by experts from the Assessment of SpondyloArthritis
international Society (ASAS). Annals of the Rheumatic Diseases 2009 Jun;68(6):784-
788.
169
45. Max R, Lorenz HM, Mackensen F. Ocular involvement in spondyloarthropathies:
HLA B27 associated uveitis. Zeitschrift fur Rheumatologie 2010 Jul;69(5):397-402.
46. Khan MA, Kushner I, Braun WE. Comparison of clinical features in HLA-B27
positive and negative patients with ankylosing spondylitis. Arthritis and Rheumatism
1977 May;20(4):909-912.
47. Rudwaleit M, Baeten D. Ankylosing spondylitis and bowel disease. Best practice &
research.Clinical rheumatology 2006 Jun;20(3):451-471.
48. Brophy S, Pavy S, Lewis P, Taylor G, Bradbury L, Robertson D, et al. Inflammatory
eye, skin, and bowel disease in spondyloarthritis: genetic, phenotypic, and
environmental factors. The Journal of rheumatology 2001 Dec;28(12):2667-2673.
49. De Vos M, Cuvelier C, Mielants H, Veys E, Barbier F, Elewaut A. Ileocolonoscopy
in seronegative spondylarthropathy. Gastroenterology 1989 Feb;96(2 Pt 1):339-344.
50. Leirisalo-Repo M, Turunen U, Stenman S, Helenius P, Seppala K. High frequency
of silent inflammatory bowel disease in spondylarthropathy. Arthritis and Rheumatism
1994 Jan;37(1):23-31.
51. Mielants H, Veys EM, Cuvelier C, De Vos M. Subclinical involvement of the gut in
undifferentiated spondylarthropathies. Clinical and experimental rheumatology 1989
Sep-Oct;7(5):499-504.
52. Mielants H, Veys EM, Cuvelier C, de Vos M. Ileocolonoscopic findings in
seronegative spondylarthropathies. British journal of rheumatology 1988;27 Suppl
2:95-105.
53. Altomonte L, Zoli A, Veneziani A, Mirone L, Santacesaria G, Chiarelli C, et al.
Clinically silent inflammatory gut lesions in undifferentiated spondyloarthropathies.
Clinical rheumatology 1994 Dec;13(4):565-570.
54. Perez Alamino R, Maldonado Cocco JA, Citera G, Arturi P, Vazquez-Mellado J,
Sampaio-Barros PD, et al. Differential features between primary ankylosing spondylitis
170
and spondylitis associated with psoriasis and inflammatory bowel disease. The Journal
of rheumatology 2011 Aug;38(8):1656-1660.
55. Mielants H, Veys EM, Goemaere S, Goethals K, Cuvelier C, De Vos M. Gut
inflammation in the spondyloarthropathies: clinical, radiologic, biologic and genetic
features in relation to the type of histology. A prospective study. The Journal of
rheumatology 1991 Oct;18(10):1542-1551.
56. Carter S, Lories RJ. Osteoporosis: a paradox in ankylosing spondylitis. Current
osteoporosis reports 2011 Sep;9(3):112-115.
57. Magrey M, Khan MA. Osteoporosis in ankylosing spondylitis. Current rheumatology
reports 2010 Oct;12(5):332-336.
58. van der Weijden MA, van Denderen JC, Lems WF, Heymans MW, Dijkmans BA,
van der Horst-Bruinsma IE. Low bone mineral density is related to male gender and
decreased functional capacity in early spondylarthropathies. Clinical rheumatology
2011 Apr;30(4):497-503.
59. Montala N, Juanola X, Collantes E, Munoz-Gomariz E, Gonzalez C, Gratacos J, et
al. Prevalence of vertebral fractures by semiautomated morphometry in patients with
ankylosing spondylitis. The Journal of rheumatology 2011 May;38(5):893-897.
60. Bollow M, Enzweiler C, Taupitz M, Golder W, Hamm B, Sieper J, et al. Use of
contrast enhanced magnetic resonance imaging to detect spinal inflammation in
patients with spondyloarthritides. Clinical and experimental rheumatology 2002 Nov-
Dec;20(6 Suppl 28):S167-74.
61. Bennett AN, Rehman A, Hensor EM, Marzo-Ortega H, Emery P, McGonagle D.
The fatty Romanus lesion: a non-inflammatory spinal MRI lesion specific for axial
spondyloarthropathy. Annals of the Rheumatic Diseases 2010 May;69(5):891-894.
62. Puhakka KB, Jurik AG, Egund N, Schiottz-Christensen B, Stengaard-Pedersen K,
van Overeem Hansen G, et al. Imaging of sacroiliitis in early seronegative
spondylarthropathy. Assessment of abnormalities by MR in comparison with
171
radiography and CT. Acta Radiologica (Stockholm, Sweden : 1987) 2003
Mar;44(2):218-229.
63. Rudwaleit M, Jurik AG, Hermann KG, Landewe R, van der Heijde D, Baraliakos X,
et al. Defining active sacroiliitis on magnetic resonance imaging (MRI) for classification
of axial spondyloarthritis: a consensual approach by the ASAS/OMERACT MRI group.
Annals of the Rheumatic Diseases 2009 Oct;68(10):1520-1527.
64. Maksymowych WP, Salonen D, Inman RD, Rahman P, Lambert RG, CANDLE
Study Group. Low-dose infliximab (3 mg/kg) significantly reduces spinal inflammation
on magnetic resonance imaging in patients with ankylosing spondylitis: a randomized
placebo-controlled study. The Journal of rheumatology 2010 Aug 1;37(8):1728-1734.
65. Baraliakos X, Davis J, Tsuji W, Braun J. Magnetic resonance imaging examinations
of the spine in patients with ankylosing spondylitis before and after therapy with the
tumor necrosis factor alpha receptor fusion protein etanercept. Arthritis and
Rheumatism 2005 Apr;52(4):1216-1223.
66. Baraliakos X, Brandt J, Listing J, Haibel H, Sorensen H, Rudwaleit M, et al.
Outcome of patients with active ankylosing spondylitis after two years of therapy with
etanercept: clinical and magnetic resonance imaging data. Arthritis and Rheumatism
2005 Dec 15;53(6):856-863.
67. Lambert RG, Salonen D, Rahman P, Inman RD, Wong RL, Einstein SG, et al.
Adalimumab significantly reduces both spinal and sacroiliac joint inflammation in
patients with ankylosing spondylitis: a multicenter, randomized, double-blind, placebo-
controlled study. Arthritis and Rheumatism 2007 Dec;56(12):4005-4014.
68. van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for
ankylosing spondylitis. A proposal for modification of the New York criteria. Arthritis
and Rheumatism 1984 Apr;27(4):361-368.
69. Rudwaleit M, Landewe R, van der Heijde D, Listing J, Brandt J, Braun J, et al. The
development of Assessment of SpondyloArthritis international Society classification
criteria for axial spondyloarthritis (part I): classification of paper patients by expert
172
opinion including uncertainty appraisal. Annals of the Rheumatic Diseases 2009
Jun;68(6):770-776.
70. Rudwaleit M, van der Heijde D, Landewe R, Listing J, Akkoc N, Brandt J, et al. The
development of Assessment of SpondyloArthritis international Society classification
criteria for axial spondyloarthritis (part II): validation and final selection. Annals of the
Rheumatic Diseases 2009 Jun;68(6):777-783.
71. Dincer U, Cakar E, Kiralp MZ, Dursun H. Diagnosis delay in patients with
ankylosing spondylitis: possible reasons and proposals for new diagnostic criteria.
Clinical rheumatology 2008 Apr;27(4):457-462.
72. Aggarwal R, Malaviya AN. Diagnosis delay in patients with ankylosing spondylitis:
factors and outcomes--an Indian perspective. Clinical rheumatology 2009
Mar;28(3):327-331.
73. Bennett AN, McGonagle D, O'Connor P, Hensor EM, Sivera F, Coates LC, et al.
Severity of baseline magnetic resonance imaging-evident sacroiliitis and HLA-B27
status in early inflammatory back pain predict radiographically evident ankylosing
spondylitis at eight years. Arthritis and Rheumatism 2008 Nov;58(11):3413-3418.
74. Ince G, Sarpel T, Durgun B, Erdogan S. Effects of a multimodal exercise program
for people with ankylosing spondylitis. Physical therapy 2006 Jul;86(7):924-935.
75. Dagfinrud H, Kvien TK, Hagen KB. Physiotherapy interventions for ankylosing
spondylitis. Cochrane database of systematic reviews (Online) 2008 Jan
23;(1)(1):CD002822.
76. Hidding A, van der Linden S, Gielen X, de Witte L, Dijkmans B, Moolenburgh D.
Continuation of group physical therapy is necessary in ankylosing spondylitis: results of
a randomized controlled trial. Arthritis Care and Research : The Official Journal of the
Arthritis Health Professions Association 1994 Jun;7(2):90-96.
77. Fernandez-de-Las-Penas C, Alonso-Blanco C, Alguacil-Diego IM, Miangolarra-
Page JC. One-year follow-up of two exercise interventions for the management of
173
patients with ankylosing spondylitis: a randomized controlled trial. American Journal of
Physical Medicine & Rehabilitation / Association of Academic Physiatrists 2006
Jul;85(7):559-567.
78. Buckelew SP, Parker JC. Coping with arthritis pain. A review of the literature.
Arthritis Care and Research : The Official Journal of the Arthritis Health Professions
Association 1989 Dec;2(4):136-145.
79. Altan L, Bingol U, Aslan M, Yurtkuran M. The effect of balneotherapy on patients
with ankylosing spondylitis. Scandinavian journal of rheumatology 2006 Jul-
Aug;35(4):283-289.
80. Dougados M, Revel M, Khan MA. Spondylarthropathy treatment: progress in
medical treatment, physical therapy and rehabilitation. Bailliere's Clinical
Rheumatology 1998 Nov;12(4):717-736.
81. Braun J, van den Berg R, Baraliakos X, Boehm H, Burgos-Vargas R, Collantes-
Estevez E, et al. 2010 update of the ASAS/EULAR recommendations for the
management of ankylosing spondylitis. Annals of the Rheumatic Diseases 2011
Jun;70(6):896-904.
82. Langman MJ, Jensen DM, Watson DJ, Harper SE, Zhao PL, Quan H, et al.
Adverse upper gastrointestinal effects of rofecoxib compared with NSAIDs. JAMA : the
journal of the American Medical Association 1999 Nov 24;282(20):1929-1933.
83. Mallen SR, Essex MN, Zhang R. Gastrointestinal tolerability of NSAIDs in elderly
patients: a pooled analysis of 21 randomized clinical trials with celecoxib and
nonselective NSAIDs. Current medical research and opinion 2011 Jul;27(7):1359-
1366.
84. Wanders A, Heijde D, Landewe R, Behier JM, Calin A, Olivieri I, et al. Nonsteroidal
antiinflammatory drugs reduce radiographic progression in patients with ankylosing
spondylitis: a randomized clinical trial. Arthritis and Rheumatism 2005 Jun;52(6):1756-
1765.
174
85. Kirwan J, Edwards A, Huitfeldt B, Thompson P, Currey H. The course of
established ankylosing spondylitis and the effects of sulphasalazine over 3 years.
British journal of rheumatology 1993 Aug;32(8):729-733.
86. Braun J, Zochling J, Baraliakos X, Alten R, Burmester G, Grasedyck K, et al.
Efficacy of sulfasalazine in patients with inflammatory back pain due to undifferentiated
spondyloarthritis and early ankylosing spondylitis: a multicentre randomised controlled
trial. Annals of the Rheumatic Diseases 2006 Sep;65(9):1147-1153.
87. Ferraz MB, Tugwell P, Goldsmith CH, Atra E. Meta-analysis of sulfasalazine in
ankylosing spondylitis. The Journal of rheumatology 1990 Nov;17(11):1482-1486.
88. Sampaio-Barros PD, Costallat LT, Bertolo MB, Neto JF, Samara AM. Methotrexate
in the treatment of ankylosing spondylitis. Scandinavian journal of rheumatology
2000;29(3):160-162.
89. Biasi D, Carletto A, Caramaschi P, Pacor ML, Maleknia T, Bambara LM. Efficacy of
methotrexate in the treatment of ankylosing spondylitis: a three-year open study.
Clinical rheumatology 2000;19(2):114-117.
90. Gonzalez-Lopez L, Garcia-Gonzalez A, Vazquez-Del-Mercado M, Munoz-Valle JF,
Gamez-Nava JI. Efficacy of methotrexate in ankylosing spondylitis: a randomized,
double blind, placebo controlled trial. The Journal of rheumatology 2004
Aug;31(8):1568-1574.
91. Haibel H, Sieper J. Use of methotrexate in patients with ankylosing spondylitis.
Clinical and experimental rheumatology 2010 Sep-Oct;28(5 Suppl 61):S128-31.
92. Ward MM, Kuzis S. Treatments used by patients with ankylosing spondylitis
comparison with the treatment preferences of rheumatologists. Journal of clinical
rheumatology : practical reports on rheumatic & musculoskeletal diseases 1999
Feb;5(1):1-8.
93. van Denderen JC, van der Paardt M, Nurmohamed MT, de Ryck YM, Dijkmans BA,
van der Horst-Bruinsma IE. Double blind, randomised, placebo controlled study of
175
leflunomide in the treatment of active ankylosing spondylitis. Annals of the Rheumatic
Diseases 2005 Dec;64(12):1761-1764.
94. Haibel H, Rudwaleit M, Braun J, Sieper J. Six months open label trial of
leflunomide in active ankylosing spondylitis. Annals of the Rheumatic Diseases 2005
Jan;64(1):124-126.
95. Ehlers S. Tumor necrosis factor and its blockade in granulomatous infections:
differential modes of action of infliximab and etanercept? Clinical infectious diseases :
an official publication of the Infectious Diseases Society of America 2005 Aug 1;41
Suppl 3:S199-203.
96. Smeets TJ, Kraan MC, van Loon ME, Tak PP. Tumor necrosis factor alpha
blockade reduces the synovial cell infiltrate early after initiation of treatment, but
apparently not by induction of apoptosis in synovial tissue. Arthritis and Rheumatism
2003 Aug;48(8):2155-2162.
97. Catrina AI, Trollmo C, af Klint E, Engstrom M, Lampa J, Hermansson Y, et al.
Evidence that anti-tumor necrosis factor therapy with both etanercept and infliximab
induces apoptosis in macrophages, but not lymphocytes, in rheumatoid arthritis joints:
extended report. Arthritis and Rheumatism 2005 Jan;52(1):61-72.
98. Vos AC, Wildenberg ME, Arijs I, Duijvestein M, Verhaar AP, de Hertogh G, et al.
Regulatory macrophages induced by infliximab are involved in healing in vivo and in
vitro. Inflammatory bowel diseases 2011 Sep 20.
99. Zou J, Rudwaleit M, Brandt J, Thiel A, Braun J, Sieper J. Up regulation of the
production of tumour necrosis factor alpha and interferon gamma by T cells in
ankylosing spondylitis during treatment with etanercept. Annals of the Rheumatic
Diseases 2003 Jun;62(6):561-564.
100. van der Heijde D, Da Silva JC, Dougados M, Geher P, van der Horst-Bruinsma I,
Juanola X, et al. Etanercept 50 mg once weekly is as effective as 25 mg twice weekly
in patients with ankylosing spondylitis. Annals of the Rheumatic Diseases 2006
Dec;65(12):1572-1577.
176
101. Poddubnyy D, Rudwaleit M. Efficacy and safety of adalimumab treatment in
patients with rheumatoid arthritis, ankylosing spondylitis and psoriatic arthritis. Expert
opinion on drug safety 2011 Jul;10(4):655-673.
102. Inman RD, Davis JC,Jr, Heijde D, Diekman L, Sieper J, Kim SI, et al. Efficacy and
safety of golimumab in patients with ankylosing spondylitis: results of a randomized,
double-blind, placebo-controlled, phase III trial. Arthritis and Rheumatism 2008
Nov;58(11):3402-3412.
103. Jarvinen P. Occurrence of ankylosing spondylitis in a nationwide series of twins.
Arthritis and Rheumatism 1995 Mar;38(3):381-383.
104. Pedersen OB, Svendsen AJ, Ejstrup L, Skytthe A, Harris JR, Junker P.
Ankylosing spondylitis in Danish and Norwegian twins: occurrence and the relative
importance of genetic vs. environmental effectors in disease causation. Scandinavian
journal of rheumatology 2008 Mar-Apr;37(2):120-126.
105. Braun J, Tuszewski M, Ehlers S, Haberle J, Bollow M, Eggens U, et al. Nested
polymerase chain reaction strategy simultaneously targeting DNA sequences of
multiple bacterial species in inflammatory joint diseases. II. Examination of sacroiliac
and knee joint biopsies of patients with spondyloarthropathies and other arthritides.
The Journal of rheumatology 1997 Jun;24(6):1101-1105.
106. Ebringer A, Ahmadi K, Fielder M, Rashid T, Tiwana H, Wilson C, et al. Molecular
mimicry: the geographical distribution of immune responses to Klebsiella in ankylosing
spondylitis and its relevance to therapy. Clinical rheumatology 1996 Jan;15 Suppl 1:57-
61.
107. Zambrano-Zaragoza F, Garcia-Latorre E, Dominguez-Lopez ML, Cancino-Diaz
ME, Burgos-Vargas R, Jimenez-Zamudio L. CD4 and CD8 T cell response to the
rHSP60 from Klebsiella pneumoniae in peripheral blood mononuclear cells from
patients with ankylosing spondylitis. Revista de investigacion clinica; organo del
Hospital de Enfermedades de la Nutricion 2005 Jul-Aug;57(4):555-562.
177
108. Ahmadi K, Wilson C, Tiwana H, Ebringer A, Shanmuganathan S, Binder A.
Antibodies to Klebsiella pneumoniae nitrogenase reductase in patients with ankylosing
spondylitis. Annals of the Rheumatic Diseases 1998 Jul;57(7):441.
109. Cancino-Diaz ME, Perez-Salazar JE, Dominguez-Lopez L, Escobar-Gutierrez A,
Granados-Arreola J, Jimenez-Zamudio L, et al. Antibody response to Klebsiella
pneumoniae 60 kDa protein in familial and sporadic ankylosing spondylitis: role of
HLA-B27 and characterization as a GroEL-like protein. The Journal of rheumatology
1998 Sep;25(9):1756-1764.
110. Hohler T, Hug R, Schneider PM, Krummenauer F, Gripenberg-Lerche C, Granfors
K, et al. Ankylosing spondylitis in monozygotic twins: studies on immunological
parameters. Annals of the Rheumatic Diseases 1999 Jul;58(7):435-440.
111. Taurog JD, Richardson JA, Croft JT, Simmons WA, Zhou M, Fernandez-Sueiro
JL, et al. The germfree state prevents development of gut and joint inflammatory
disease in HLA-B27 transgenic rats. The Journal of experimental medicine 1994 Dec
1;180(6):2359-2364.
112. Stone MA, Payne U, Schentag C, Rahman P, Pacheco-Tena C, Inman RD.
Comparative immune responses to candidate arthritogenic bacteria do not confirm a
dominant role for Klebsiella pneumonia in the pathogenesis of familial ankylosing
spondylitis. Rheumatology (Oxford, England) 2004 Feb;43(2):148-155.
113. Smith GW, Blackwell CC, Nuki G. Faecal flora in spondyloarthropathy. British
journal of rheumatology 1997 Aug;36(8):850-854.
114. Mattila P, Majuri ML, Mattila PS, Renkonen R. TNF alpha-induced expression of
endothelial adhesion molecules, ICAM-1 and VCAM-1, is linked to protein kinase C
activation. Scandinavian journal of immunology 1992 Aug;36(2):159-165.
115. Voelkel-Johnson C, Entingh AJ, Wold WS, Gooding LR, Laster SM. Activation of
intracellular proteases is an early event in TNF-induced apoptosis. Journal of
immunology (Baltimore, Md.: 1950) 1995 Feb 15;154(4):1707-1716.
178
116. Dinarello CA. Inflammatory cytokines: interleukin-1 and tumor necrosis factor as
effector molecules in autoimmune diseases. Current opinion in immunology 1991
Dec;3(6):941-948.
117. Fiers W, Beyaert R, Boone E, Cornelis S, Declercq W, Decoster E, et al. TNF-
induced intracellular signaling leading to gene induction or to cytotoxicity by necrosis or
by apoptosis. Journal of inflammation 1995 -1996;47(1-2):67-75.
118. Yoshimura A, Mori H, Ohishi M, Aki D, Hanada T. Negative regulation of cytokine
signaling influences inflammation. Current opinion in immunology 2003 Dec;15(6):704-
708.
119. Gratacos J, Collado A, Filella X, Sanmarti R, Canete J, Llena J, et al. Serum
cytokines (IL-6, TNF-alpha, IL-1 beta and IFN-gamma) in ankylosing spondylitis: a
close correlation between serum IL-6 and disease activity and severity. British journal
of rheumatology 1994 Oct;33(10):927-931.
120. Braun J, Xiang J, Brandt J, Maetzel H, Haibel H, Wu P, et al. Treatment of
spondyloarthropathies with antibodies against tumour necrosis factor alpha: first
clinical and laboratory experiences. Annals of the Rheumatic Diseases 2000 Nov;59
Suppl 1:i85-9.
121. Murch SH, Braegger CP, Walker-Smith JA, MacDonald TT. Location of tumour
necrosis factor alpha by immunohistochemistry in chronic inflammatory bowel disease.
Gut 1993 Dec;34(12):1705-1709.
122. Mielants H, Veys EM, Cuvelier C, De Vos M. Course of gut inflammation in
spondylarthropathies and therapeutic consequences. Bailliere's Clinical Rheumatology
1996 Feb;10(1):147-164.
123. Sousa E, Caetano-Lopes J, Pinto P, Pimentel F, Teles J, Canhao H, et al.
Ankylosing spondylitis susceptibility and severity--contribution of TNF gene promoter
polymorphisms at positions -238 and -308. Annals of the New York Academy of
Sciences 2009 Sep;1173:581-588.
179
124. Redlich K, Gortz B, Hayer S, Zwerina J, Kollias G, Steiner G, et al.
Overexpression of tumor necrosis factor causes bilateral sacroiliitis. Arthritis and
Rheumatism 2004 Mar;50(3):1001-1005.
125. Lee YH, Song GG. Lack of association of TNF-alpha promoter polymorphisms
with ankylosing spondylitis: a meta-analysis. Rheumatology (Oxford, England) 2009
Nov;48(11):1359-1362.
126. Li B, Wang P, Li H. The association between TNF-alpha promoter polymorphisms
and ankylosing spondylitis: a meta-analysis. Clinical rheumatology 2010
Sep;29(9):983-990.
127. Australo-Anglo-American Spondyloarthritis Consortium (TASC), Reveille JD, Sims
AM, Danoy P, Evans DM, Leo P, et al. Genome-wide association study of ankylosing
spondylitis identifies non-MHC susceptibility loci. Nature genetics 2010 Feb;42(2):123-
127.
128. The Australo-Anglo-American Spondyloarthritis Consortium (TASC), the
Wellcome Trust Case Control Consortium 2 (WTCCC2), Evans DM, Spencer CC,
Pointon JJ, Su Z, et al. Interaction between ERAP1 and HLA-B27 in ankylosing
spondylitis implicates peptide handling in the mechanism for HLA-B27 in disease
susceptibility. Nature genetics 2011 Jul 10;43(8):761-767.
129. Seitz M, Wirthmuller U, Moller B, Villiger PM. The -308 tumour necrosis factor-
alpha gene polymorphism predicts therapeutic response to TNFalpha-blockers in
rheumatoid arthritis and spondyloarthritis patients. Rheumatology (Oxford, England)
2007 Jan;46(1):93-96.
130. Lu MC, Yang KL, Tung CH, Huang KY, Yu HC, Liu SQ, et al. Higher LPS-
stimulated TNF-alpha mRNA levels in peripheral blood mononuclear cells from
Chinese ankylosing spondylitis patients with -308G/A polymorphism in promoter region
of tumor necrosis factor: association with distinct A33/B58/Cw10 haplotypes.
Rheumatology international 2008 Dec;29(2):189-195.
180
131. Gupta R, Kapuria V, Das SN. Single nucleotide polymorphisms in TNF-alpha,
TNFR2 gene and TNF-alpha production in Asian Indians. Immunological investigations
2009;38(3-4):240-254.
132. Shen H, Goodall JC, Hill Gaston JS. Frequency and phenotype of peripheral
blood Th17 cells in ankylosing spondylitis and rheumatoid arthritis. Arthritis and
Rheumatism 2009 Jun;60(6):1647-1656.
133. Appel H, Maier R, Wu P, Scheer R, Hempfing A, Kayser R, et al. Analysis of IL-
17+ cells in facet joints of patients with spondyloarthritis suggests that the innate
immune pathway might be of greater relevance than the Th17-mediated adaptive
immune response. Arthritis research & therapy 2011 Jun 6;13(3):R95.
134. Bowness P, Ridley A, Shaw J, Chan AT, Wong-Baeza I, Fleming M, et al. Th17
cells expressing KIR3DL2+ and responsive to HLA-B27 homodimers are increased in
ankylosing spondylitis. Journal of immunology (Baltimore, Md.: 1950) 2011 Feb
15;186(4):2672-2680.
135. DeLay ML, Turner MJ, Klenk EI, Smith JA, Sowders DP, Colbert RA. HLA-B27
misfolding and the unfolded protein response augment interleukin-23 production and
are associated with Th17 activation in transgenic rats. Arthritis and Rheumatism 2009
Sep;60(9):2633-2643.
136. Taylan A, Sari I, Kozaci DL, Yuksel A, Bilge S, Yildiz Y, et al. Evaluation of the T
helper 17 axis in ankylosing spondylitis. Rheumatology international 2011 Jul 16.
137. Wellcome Trust Case Control Consortium, Australo-Anglo-American Spondylitis
Consortium (TASC), Burton PR, Clayton DG, Cardon LR, Craddock N, et al.
Association scan of 14,500 nonsynonymous SNPs in four diseases identifies
autoimmunity variants. Nature genetics 2007 Nov;39(11):1329-1337.
138. Ciccia F, Bombardieri M, Principato A, Giardina A, Tripodo C, Porcasi R, et al.
Overexpression of interleukin-23, but not interleukin-17, as an immunologic signature
of subclinical intestinal inflammation in ankylosing spondylitis. Arthritis and
Rheumatism 2009 Apr;60(4):955-965.
181
139. McGarry F, Neilly J, Anderson N, Sturrock R, Field M. A polymorphism within the
interleukin 1 receptor antagonist (IL-1Ra) gene is associated with ankylosing
spondylitis. Rheumatology (Oxford, England) 2001 Dec;40(12):1359-1364.
140. Maksymowych WP, Rahman P, Reeve JP, Gladman DD, Peddle L, Inman RD.
Association of the IL1 gene cluster with susceptibility to ankylosing spondylitis: an
analysis of three Canadian populations. Arthritis and Rheumatism 2006 Mar;54(3):974-
985.
141. Kim TJ, Kim TH, Lee HJ, Peddle L, Rahman P, Hu P, et al. Interleukin 1
polymorphisms in patients with ankylosing spondylitis in Korea. The Journal of
rheumatology 2008 Aug;35(8):1603-1608.
142. Chou CT, Timms AE, Wei JC, Tsai WC, Wordsworth BP, Brown MA. Replication
of association of IL1 gene complex members with ankylosing spondylitis in Taiwanese
Chinese. Annals of the Rheumatic Diseases 2006 Aug;65(8):1106-1109.
143. Sims AM, Timms AE, Bruges-Armas J, Burgos-Vargas R, Chou CT, Doan T, et al.
Prospective meta-analysis of interleukin 1 gene complex polymorphisms confirms
associations with ankylosing spondylitis. Annals of the Rheumatic Diseases 2008
Sep;67(9):1305-1309.
144. Brown MA, Laval SH, Brophy S, Calin A. Recurrence risk modelling of the genetic
susceptibility to ankylosing spondylitis. Annals of the Rheumatic Diseases 2000
Nov;59(11):883-886.
145. Brown MA, Kennedy LG, MacGregor AJ, Darke C, Duncan E, Shatford JL, et al.
Susceptibility to ankylosing spondylitis in twins: the role of genes, HLA, and the
environment. Arthritis and Rheumatism 1997 Oct;40(10):1823-1828.
146. Brewerton DA, Hart FD, Nicholls A, Caffrey M, James DC, Sturrock RD.
Ankylosing spondylitis and HL-A 27. Lancet 1973 Apr 28;1(7809):904-907.
147. Hammer RE, Maika SD, Richardson JA, Tang JP, Taurog JD. Spontaneous
inflammatory disease in transgenic rats expressing HLA-B27 and human beta 2m: an
182
animal model of HLA-B27-associated human disorders. Cell 1990 Nov 30;63(5):1099-
1112.
148. Tanigaki N, Fruci D, Vigneti E, Starace G, Rovero P, Londei M, et al. The peptide
binding specificity of HLA-B27 subtypes. Immunogenetics 1994;40(3):192-198.
149. Fiorillo MT, Meadows L, D'Amato M, Shabanowitz J, Hunt DF, Appella E, et al.
Susceptibility to ankylosing spondylitis correlates with the C-terminal residue of
peptides presented by various HLA-B27 subtypes. European journal of immunology
1997 Feb;27(2):368-373.
150. Ben Dror L, Barnea E, Beer I, Mann M, Admon A. The HLA-B*2705 peptidome.
Arthritis and Rheumatism 2010 Feb;62(2):420-429.
151. Lopez de Castro JA. The HLA-B27 peptidome: building on the cornerstone.
Arthritis and Rheumatism 2010 Feb;62(2):316-319.
152. Madden DR, Gorga JC, Strominger JL, Wiley DC. The three-dimensional
structure of HLA-B27 at 2.1 A resolution suggests a general mechanism for tight
peptide binding to MHC. Cell 1992 Sep 18;70(6):1035-1048.
153. Bowness P, Allen RL, McMichael AJ. Identification of T cell receptor recognition
residues for a viral peptide presented by HLA B27. European journal of immunology
1994 Oct;24(10):2357-2363.
154. Bowness P, Zaccai N, Bird L, Jones EY. HLA-B27 and disease pathogenesis:
new structural and functional insights. Expert reviews in molecular medicine 1999 Oct
26;1999:1-10.
155. Allen RL, O'Callaghan CA, McMichael AJ, Bowness P. Cutting edge: HLA-B27
can form a novel beta 2-microglobulin-free heavy chain homodimer structure. Journal
of immunology (Baltimore, Md.: 1950) 1999 May 1;162(9):5045-5048.
156. Madden DR, Gorga JC, Strominger JL, Wiley DC. The structure of HLA-B27
reveals nonamer self-peptides bound in an extended conformation. Nature 1991 Sep
26;353(6342):321-325.
183
157. Appel H, Kuon W, Kuhne M, Hulsmeyer M, Kollnberger S, Kuhlmann S, et al. The
solvent-inaccessible Cys67 residue of HLA-B27 contributes to T cell recognition of
HLA-B27/peptide complexes. Journal of immunology (Baltimore, Md.: 1950) 2004 Dec
1;173(11):6564-6573.
158. Colbert RA, DeLay ML, Layh-Schmitt G, Sowders DP. HLA-B27 misfolding and
spondyloarthropathies. Advances in Experimental Medicine and Biology 2009;649:217-
234.
159. Colbert RA, DeLay ML, Klenk EI, Layh-Schmitt G. From HLA-B27 to
spondyloarthritis: a journey through the ER. Immunological reviews 2010
Jan;233(1):181-202.
160. Turner MJ, Delay ML, Bai S, Klenk E, Colbert RA. HLA-B27 up-regulation causes
accumulation of misfolded heavy chains and correlates with the magnitude of the
unfolded protein response in transgenic rats: Implications for the pathogenesis of
spondylarthritis-like disease. Arthritis and Rheumatism 2007 Jan;56(1):215-223.
161. Turner MJ, Sowders DP, DeLay ML, Mohapatra R, Bai S, Smith JA, et al. HLA-
B27 misfolding in transgenic rats is associated with activation of the unfolded protein
response. Journal of immunology (Baltimore, Md.: 1950) 2005 Aug 15;175(4):2438-
2448.
162. Khare SD, Hansen J, Luthra HS, David CS. HLA-B27 heavy chains contribute to
spontaneous inflammatory disease in B27/human beta2-microglobulin (beta2m) double
transgenic mice with disrupted mouse beta2m. The Journal of clinical investigation
1996 Dec 15;98(12):2746-2755.
163. Kollnberger S, Bird LA, Roddis M, Hacquard-Bouder C, Kubagawa H, Bodmer
HC, et al. HLA-B27 heavy chain homodimers are expressed in HLA-B27 transgenic
rodent models of spondyloarthritis and are ligands for paired Ig-like receptors. Journal
of immunology (Baltimore, Md.: 1950) 2004 Aug 1;173(3):1699-1710.
164. Raine T, Brown D, Bowness P, Hill Gaston JS, Moffett A, Trowsdale J, et al.
Consistent patterns of expression of HLA class I free heavy chains in healthy
184
individuals and raised expression in spondyloarthropathy patients point to physiological
and pathological roles. Rheumatology (Oxford, England) 2006 Nov;45(11):1338-1344.
165. Tsai WC, Chen CJ, Yen JH, Ou TT, Tsai JJ, Liu CS, et al. Free HLA class I heavy
chain-carrying monocytes--a potential role in the pathogenesis of
spondyloarthropathies. The Journal of rheumatology 2002 May;29(5):966-972.
166. Thananchai H, Gillespie G, Martin MP, Bashirova A, Yawata N, Yawata M, et al.
Cutting Edge: Allele-specific and peptide-dependent interactions between KIR3DL1
and HLA-A and HLA-B. Journal of immunology (Baltimore, Md.: 1950) 2007 Jan
1;178(1):33-37.
167. Kollnberger S, Chan A, Sun MY, Chen LY, Wright C, di Gleria K, et al. Interaction
of HLA-B27 homodimers with KIR3DL1 and KIR3DL2, unlike HLA-B27 heterotrimers,
is independent of the sequence of bound peptide. European journal of immunology
2007 May;37(5):1313-1322.
168. Reveille JD, Maganti RM. Subtypes of HLA-B27: history and implications in the
pathogenesis of ankylosing spondylitis. Advances in Experimental Medicine and
Biology 2009;649:159-176.
169. Hulsmeyer M, Fiorillo MT, Bettosini F, Sorrentino R, Saenger W, Ziegler A, et al.
Dual, HLA-B27 subtype-dependent conformation of a self-peptide. The Journal of
experimental medicine 2004 Jan 19;199(2):271-281.
170. Fiorillo MT, Ruckert C, Hulsmeyer M, Sorrentino R, Saenger W, Ziegler A, et al.
Allele-dependent similarity between viral and self-peptide presentation by HLA-B27
subtypes. The Journal of biological chemistry 2005 Jan 28;280(4):2962-2971.
171. Ruckert C, Fiorillo MT, Loll B, Moretti R, Biesiadka J, Saenger W, et al.
Conformational dimorphism of self-peptides and molecular mimicry in a disease-
associated HLA-B27 subtype. The Journal of biological chemistry 2006 Jan
27;281(4):2306-2316.
185
172. Hulsmeyer M, Hillig RC, Volz A, Ruhl M, Schroder W, Saenger W, et al. HLA-B27
subtypes differentially associated with disease exhibit subtle structural alterations. The
Journal of biological chemistry 2002 Dec 6;277(49):47844-47853.
173. Galocha B, de Castro JA. Folding of HLA-B27 subtypes is determined by the
global effect of polymorphic residues and shows incomplete correspondence to
ankylosing spondylitis. Arthritis and Rheumatism 2008 Feb;58(2):401-412.
174. Tran TM, Dorris ML, Satumtira N, Richardson JA, Hammer RE, Shang J, et al.
Additional human beta2-microglobulin curbs HLA-B27 misfolding and promotes arthritis
and spondylitis without colitis in male HLA-B27-transgenic rats. Arthritis and
Rheumatism 2006 Apr;54(4):1317-1327.
175. Zhou M, Sayad A, Simmons WA, Jones RC, Maika SD, Satumtira N, et al. The
specificity of peptides bound to human histocompatibility leukocyte antigen (HLA)-B27
influences the prevalence of arthritis in HLA-B27 transgenic rats. The Journal of
experimental medicine 1998 Sep 7;188(5):877-886.
176. Taurog JD, Dorris ML, Satumtira N, Tran TM, Sharma R, Dressel R, et al.
Spondylarthritis in HLA-B27/human beta2-microglobulin-transgenic rats is not
prevented by lack of CD8. Arthritis and Rheumatism 2009 Jul;60(7):1977-1984.
177. Andrews NP, Pack CD, Lukacher AE. Generation of antiviral major
histocompatibility complex class I-restricted T cells in the absence of CD8 coreceptors.
Journal of virology 2008 May;82(10):4697-4705.
178. Breban M, Fernandez-Sueiro JL, Richardson JA, Hadavand RR, Maika SD,
Hammer RE, et al. T cells, but not thymic exposure to HLA-B27, are required for the
inflammatory disease of HLA-B27 transgenic rats. Journal of immunology (Baltimore,
Md.: 1950) 1996 Jan 15;156(2):794-803.
179. Fukunishi S, Yoh K, Kamae S, Yoshiya S. Beta 2-microglobulin amyloid deposit in
HLA-B27 transgenic rats. Modern rheumatology / the Japan Rheumatism Association
2007;17(5):380-384.
186
180. Baraliakos X, Listing J, Rudwaleit M, Sieper J, Braun J. The relationship between
inflammation and new bone formation in patients with ankylosing spondylitis. Arthritis
research & therapy 2008;10(5):R104.
181. Maksymowych WP, Chiowchanwisawakit P, Clare T, Pedersen SJ, Ostergaard M,
Lambert RG. Inflammatory lesions of the spine on magnetic resonance imaging predict
the development of new syndesmophytes in ankylosing spondylitis: evidence of a
relationship between inflammation and new bone formation. Arthritis and Rheumatism
2009 Jan;60(1):93-102.
182. Klaus A, Birchmeier W. Wnt signalling and its impact on development and cancer.
Nature reviews.Cancer 2008 May;8(5):387-398.
183. Nusse R, van Ooyen A, Cox D, Fung YK, Varmus H. Mode of proviral activation of
a putative mammary oncogene (int-1) on mouse chromosome 15. Nature 1984 Jan 12-
18;307(5947):131-136.
184. Pinzone JJ, Hall BM, Thudi NK, Vonau M, Qiang YW, Rosol TJ, et al. The role of
Dickkopf-1 in bone development, homeostasis, and disease. Blood 2009 Jan
15;113(3):517-525.
185. Niehrs C. Function and biological roles of the Dickkopf family of Wnt modulators.
Oncogene 2006 Dec 4;25(57):7469-7481.
186. Mikheev AM, Mikheeva SA, Rostomily R, Zarbl H. Dickkopf-1 activates cell death
in MDA-MB435 melanoma cells. Biochemical and biophysical research
communications 2007 Jan 19;352(3):675-680.
187. Kim KA, Wagle M, Tran K, Zhan X, Dixon MA, Liu S, et al. R-Spondin family
members regulate the Wnt pathway by a common mechanism. Molecular biology of
the cell 2008 Jun;19(6):2588-2596.
188. Diarra D, Stolina M, Polzer K, Zwerina J, Ominsky MS, Dwyer D, et al. Dickkopf-1
is a master regulator of joint remodeling. Nature medicine 2007 Feb;13(2):156-163.
187
189. Marshall MJ, Evans SF, Sharp CA, Powell DE, McCarthy HS, Davie MW.
Increased circulating Dickkopf-1 in Paget's disease of bone. Clinical biochemistry 2009
Jul;42(10-11):965-969.
190. Rawadi G, Roman-Roman S. Wnt signalling pathway: a new target for the
treatment of osteoporosis. Expert opinion on therapeutic targets 2005 Oct;9(5):1063-
1077.
191. Hopwood B, Tsykin A, Findlay DM, Fazzalari NL. Microarray gene expression
profiling of osteoarthritic bone suggests altered bone remodelling, WNT and
transforming growth factor-beta/bone morphogenic protein signalling. Arthritis research
& therapy 2007;9(5):R100.
192. Kwon SR, Lim MJ, Suh CH, Park SG, Hong YS, Yoon BY, et al. Dickkopf-1 level
is lower in patients with ankylosing spondylitis than in healthy people and is not
influenced by anti-tumor necrosis factor therapy. Rheumatology international 2011 Jul
16.
193. Honsawek S, Tanavalee A, Yuktanandana P, Ngarmukos S, Saetan N, Tantavisut
S. Dickkopf-1 (Dkk-1) in plasma and synovial fluid is inversely correlated with
radiographic severity of knee osteoarthritis patients. BMC musculoskeletal disorders
2010 Nov 10;11:257.
194. Daoussis D, Andonopoulos AP. The Emerging Role of Dickkopf-1 in Bone
Biology: Is It the Main Switch Controlling Bone and Joint Remodeling? Seminars in
arthritis and rheumatism 2011 Mar 22.
195. Garnero P, Tabassi NC, Voorzanger-Rousselot N. Circulating dickkopf-1 and
radiological progression in patients with early rheumatoid arthritis treated with
etanercept. The Journal of rheumatology 2008 Dec;35(12):2313-2315.
196. Senolt L, Hulejova H, Krystufkova O, Forejtova S, Andres Cerezo L, Gatterova J,
et al. Low circulating Dickkopf-1 and its link with severity of spinal involvement in
diffuse idiopathic skeletal hyperostosis. Annals of the Rheumatic Diseases 2011 Sep 7.
188
197. Long L, Liu Y, Wang S, Zhao Y, Guo J, Yu P, et al. Dickkopf-1 as potential
biomarker to evaluate bone erosion in systemic lupus erythematosus. Journal of
clinical immunology 2010 Sep;30(5):669-675.
198. Daoussis D, Liossis SN, Solomou EE, Tsanaktsi A, Bounia K, Karampetsou M, et
al. Evidence that Dkk-1 is dysfunctional in ankylosing spondylitis. Arthritis and
Rheumatism 2010 Jan;62(1):150-158.
199. Uderhardt S, Diarra D, Katzenbeisser J, David JP, Zwerina J, Richards W, et al.
Blockade of Dickkopf (DKK)-1 induces fusion of sacroiliac joints. Annals of the
Rheumatic Diseases 2010 Mar;69(3):592-597.
200. Laval SH, Timms A, Edwards S, Bradbury L, Brophy S, Milicic A, et al. Whole-
genome screening in ankylosing spondylitis: evidence of non-MHC genetic-
susceptibility loci. American Journal of Human Genetics 2001 Apr;68(4):918-926.
201. Serwold T, Gonzalez F, Kim J, Jacob R, Shastri N. ERAAP customizes peptides
for MHC class I molecules in the endoplasmic reticulum. Nature 2002 Oct
3;419(6906):480-483.
202. Saric T, Chang SC, Hattori A, York IA, Markant S, Rock KL, et al. An IFN-gamma-
induced aminopeptidase in the ER, ERAP1, trims precursors to MHC class I-presented
peptides. Nature immunology 2002 Dec;3(12):1169-1176.
203. Tanioka T, Hattori A, Masuda S, Nomura Y, Nakayama H, Mizutani S, et al.
Human leukocyte-derived arginine aminopeptidase. The third member of the
oxytocinase subfamily of aminopeptidases. The Journal of biological chemistry 2003
Aug 22;278(34):32275-32283.
204. Evnouchidou I, Momburg F, Papakyriakou A, Chroni A, Leondiadis L, Chang SC,
et al. The internal sequence of the peptide-substrate determines its N-terminus
trimming by ERAP1. PloS one 2008;3(11):e3658.
205. Hattori A, Kitatani K, Matsumoto H, Miyazawa S, Rogi T, Tsuruoka N, et al.
Characterization of recombinant human adipocyte-derived leucine aminopeptidase
189
expressed in Chinese hamster ovary cells. Journal of Biochemistry 2000
Nov;128(5):755-762.
206. Chang SC, Momburg F, Bhutani N, Goldberg AL. The ER aminopeptidase,
ERAP1, trims precursors to lengths of MHC class I peptides by a "molecular ruler"
mechanism. Proceedings of the National Academy of Sciences of the United States of
America 2005 Nov 22;102(47):17107-17112.
207. Kochan G, Krojer T, Harvey D, Fischer R, Chen L, Vollmar M, et al. Crystal
structures of the endoplasmic reticulum aminopeptidase-1 (ERAP1) reveal the
molecular basis for N-terminal peptide trimming. Proceedings of the National Academy
of Sciences of the United States of America 2011 May 10;108(19):7745-7750.
208. Jmol: an open-source Java viewer for chemical structures in 3D.
http://www.jmol.org/, Jmol: an open-source Java viewer for chemical structures in 3D.
http://www.jmol.org/.
209. Nguyen TT, Chang SC, Evnouchidou I, York IA, Zikos C, Rock KL, et al.
Structural basis for antigenic peptide precursor processing by the endoplasmic
reticulum aminopeptidase ERAP1. Nature structural & molecular biology 2011
May;18(5):604-613.
210. Maksymowych WP, Inman RD, Gladman DD, Reeve JP, Pope A, Rahman P.
Association of a specific ERAP1/ARTS1 haplotype with disease susceptibility in
ankylosing spondylitis. Arthritis and Rheumatism 2009 May;60(5):1317-1323.
211. Davidson SI, Wu X, Liu Y, Wei M, Danoy PA, Thomas G, et al. Association of
ERAP1, but not IL23R, with ankylosing spondylitis in a Han Chinese population.
Arthritis and Rheumatism 2009 Nov;60(11):3263-3268.
212. Choi CB, Kim TH, Jun JB, Lee HS, Shim SC, Lee B, et al. ARTS1 polymorphisms
are associated with ankylosing pondylitis in Koreans. Annals of the Rheumatic
Diseases 2009 May 3.
190
213. Pimentel-Santos FM, Ligeiro D, Matos M, Mourao AF, Sousa E, Pinto P, et al.
Association of IL23R and ERAP1 genes with ankylosing spondylitis in a Portuguese
population. Clinical and experimental rheumatology 2009 Sep-Oct;27(5):800-806.
214. Harvey D, Pointon JJ, Evans DM, Karaderi T, Farrar C, Appleton LH, et al.
Investigating the genetic association between ERAP1 and ankylosing spondylitis.
Human molecular genetics 2009 Nov 1;18(21):4204-4212.
215. Tsui FW, Haroon N, Reveille JD, Rahman P, Chiu B, Tsui HW, et al. Association
of an ERAP1 ERAP2 haplotype with familial ankylosing spondylitis. Annals of the
Rheumatic Diseases 2010 Apr;69(4):733-736.
216. Genetic Analysis of Psoriasis Consortium & the Wellcome Trust Case Control
Consortium 2, Strange A, Capon F, Spencer CC, Knight J, Weale ME, et al. A
genome-wide association study identifies new psoriasis susceptibility loci and an
interaction between HLA-C and ERAP1. Nature genetics 2010 Nov;42(11):985-990.
217. Blanchard N, Shastri N. Coping with loss of perfection in the MHC class I peptide
repertoire. Current opinion in immunology 2008 Feb;20(1):82-88.
218. Neisig A, Roelse J, Sijts AJ, Ossendorp F, Feltkamp MC, Kast WM, et al. Major
differences in transporter associated with antigen presentation (TAP)-dependent
translocation of MHC class I-presentable peptides and the effect of flanking
sequences. Journal of immunology (Baltimore, Md.: 1950) 1995 Feb 1;154(3):1273-
1279.
219. York IA, Chang SC, Saric T, Keys JA, Favreau JM, Goldberg AL, et al. The ER
aminopeptidase ERAP1 enhances or limits antigen presentation by trimming epitopes
to 8-9 residues. Nature immunology 2002 Dec;3(12):1177-1184.
220. Tsujimoto M, Hattori A. The oxytocinase subfamily of M1 aminopeptidases.
Biochimica et biophysica acta 2005 Aug 1;1751(1):9-18.
191
221. Zervoudi E, Papakyriakou A, Georgiadou D, Evnouchidou I, Gajda A, Poreba M,
et al. Probing the S1 specificity pocket of the aminopeptidases that generate antigenic
peptides. The Biochemical journal 2011 Apr 15;435(2):411-420.
222. Dixon AL, Liang L, Moffatt MF, Chen W, Heath S, Wong KC, et al. A genome-
wide association study of global gene expression. Nature genetics 2007
Oct;39(10):1202-1207.
223. Hammer GE, Gonzalez F, James E, Nolla H, Shastri N. In the absence of
aminopeptidase ERAAP, MHC class I molecules present many unstable and highly
immunogenic peptides. Nature immunology 2007 Jan;8(1):101-108.
224. Kanaseki T, Shastri N. Endoplasmic reticulum aminopeptidase associated with
antigen processing regulates quality of processed peptides presented by MHC class I
molecules. Journal of immunology (Baltimore, Md.: 1950) 2008 Nov 1;181(9):6275-
6282.
225. Gibbs RA, Weinstock GM, Metzker ML, Muzny DM, Sodergren EJ, Scherer S, et
al. Genome sequence of the Brown Norway rat yields insights into mammalian
evolution. Nature 2004 Apr 1;428(6982):493-521.
226. Saveanu L, Carroll O, Lindo V, Del Val M, Lopez D, Lepelletier Y, et al. Concerted
peptide trimming by human ERAP1 and ERAP2 aminopeptidase complexes in the
endoplasmic reticulum. Nature immunology 2005 Jul;6(7):689-697.
227. Haroon N, Tsui FW, Chiu B, Tsui HW, O'Shea FD, Inman RD. ERAP1 Q730E
Variants Affect the HLA B27 Free Chain Expression On Monocytes of Patients with
Ankylosing Spondylitis. Arthritis Rheum. 2009;60(10):S541.
228. Franke A, McGovern DP, Barrett JC, Wang K, Radford-Smith GL, Ahmad T, et al.
Genome-wide meta-analysis increases to 71 the number of confirmed Crohn's disease
susceptibility loci. Nature genetics 2010 Dec;42(12):1118-1125.
192
229. Andres AM, Dennis MY, Kretzschmar WW, Cannons JL, Lee-Lin SQ, Hurle B, et
al. Balancing selection maintains a form of ERAP2 that undergoes nonsense-mediated
decay and affects antigen presentation. PLoS genetics 2010 Oct 14;6(10):e1001157.
230. Harvey D, Pointon JJ, Karaderi T, Appleton LH, Farrar C, Wordsworth BP. A
common functional variant of endoplasmic reticulum aminopeptidase 2 (ERAP2) that
reduces major histocompatibility complex class I expression is not associated with
ankylosing spondylitis. Rheumatology (Oxford, England) 2011 Sep;50(9):1720-1721.
231. Harris MR, Lybarger L, Myers NB, Hilbert C, Solheim JC, Hansen TH, et al.
Interactions of HLA-B27 with the peptide loading complex as revealed by heavy chain
mutations. International immunology 2001 Oct;13(10):1275-1282.
232. Fukazawa T, Hermann E, Edidin M, Wen J, Huang F, Kellner H, et al. The effect
of mutant beta 2-microglobulins on the conformation of HLA-B27 detected by antibody
and by CTL. Journal of immunology (Baltimore, Md.: 1950) 1994 Oct 15;153(8):3543-
3550.
233. Allen RL, Trowsdale J. Recognition of classical and heavy chain forms of HLA-
B27 by leukocyte receptors. Current Molecular Medicine 2004 Feb;4(1):59-65.
234. Kollnberger S, Bird L, Sun MY, Retiere C, Braud VM, McMichael A, et al. Cell-
surface expression and immune receptor recognition of HLA-B27 homodimers. Arthritis
and Rheumatism 2002 Nov;46(11):2972-2982.
235. Cui X, Rouhani FN, Hawari F, Levine SJ. Shedding of the type II IL-1 decoy
receptor requires a multifunctional aminopeptidase, aminopeptidase regulator of TNF
receptor type 1 shedding. Journal of immunology (Baltimore, Md.: 1950) 2003 Dec
15;171(12):6814-6819.
236. Cui X, Rouhani FN, Hawari F, Levine SJ. An aminopeptidase, ARTS-1, is required
for interleukin-6 receptor shedding. The Journal of biological chemistry 2003 Aug
1;278(31):28677-28685.
193
237. Cui X, Hawari F, Alsaaty S, Lawrence M, Combs CA, Geng W, et al. Identification
of ARTS-1 as a novel TNFR1-binding protein that promotes TNFR1 ectodomain
shedding. The Journal of clinical investigation 2002 Aug;110(4):515-526.
238. Islam A, Adamik B, Hawari FI, Ma G, Rouhani FN, Zhang J, et al. Extracellular
TNFR1 release requires the calcium-dependent formation of a nucleobindin 2-ARTS-1
complex. The Journal of biological chemistry 2006 Mar 10;281(10):6860-6873.
239. Adamik B, Islam A, Rouhani FN, Hawari FI, Zhang J, Levine SJ. An association
between RBMX, a heterogeneous nuclear ribonucleoprotein, and ARTS-1 regulates
extracellular TNFR1 release. Biochemical and biophysical research communications
2008 Jul 4;371(3):505-509.
240. Layh-Schmitt G, Colbert RA. The interleukin-23/interleukin-17 axis in
spondyloarthritis. Current opinion in rheumatology 2008 Jul;20(4):392-397.
241. Allen RL, Raine T, Haude A, Trowsdale J, Wilson MJ. Leukocyte receptor
complex-encoded immunomodulatory receptors show differing specificity for alternative
HLA-B27 structures. Journal of immunology (Baltimore, Md.: 1950) 2001 Nov
15;167(10):5543-5547.
242. Taranta A, Gianviti A, Palma A, De Luca V, Mannucci L, Procaccino MA, et al.
Genetic risk factors in typical haemolytic uraemic syndrome. Nephrology, dialysis,
transplantation : official publication of the European Dialysis and Transplant
Association - European Renal Association 2009 Jun;24(6):1851-1857.
243. Yamamoto N, Nakayama J, Yamakawa-Kobayashi K, Hamaguchi H, Miyazaki R,
Arinami T. Identification of 33 polymorphisms in the adipocyte-derived leucine
aminopeptidase (ALAP) gene and possible association with hypertension. Human
mutation 2002 Mar;19(3):251-257.
244. Fung EY, Smyth DJ, Howson JM, Cooper JD, Walker NM, Stevens H, et al.
Analysis of 17 autoimmune disease-associated variants in type 1 diabetes identifies
6q23/TNFAIP3 as a susceptibility locus. Genes and immunity 2009 Mar;10(2):188-191.
194
245. Johnson MP, Roten LT, Dyer TD, East CE, Forsmo S, Blangero J, et al. The
ERAP2 gene is associated with preeclampsia in Australian and Norwegian
populations. Human genetics 2009 Nov;126(5):655-666.
246. Saito S, Sakai M, Sasaki Y, Nakashima A, Shiozaki A. Inadequate tolerance
induction may induce pre-eclampsia. Journal of reproductive immunology 2007
Dec;76(1-2):30-39.
247. Ponte M, Cantoni C, Biassoni R, Tradori-Cappai A, Bentivoglio G, Vitale C, et al.
Inhibitory receptors sensing HLA-G1 molecules in pregnancy: decidua-associated
natural killer cells express LIR-1 and CD94/NKG2A and acquire p49, an HLA-G1-
specific receptor. Proceedings of the National Academy of Sciences of the United
States of America 1999 May 11;96(10):5674-5679.
248. Yan WH, Lin A, Chen BG, Zhou MY, Dai MZ, Chen XJ, et al. Possible roles of
KIR2DL4 expression on uNK cells in human pregnancy. American journal of
reproductive immunology (New York, N.Y.: 1989) 2007 Apr;57(4):233-242.
249. Shido F, Ito T, Nomura S, Yamamoto E, Sumigama S, Ino K, et al. Endoplasmic
reticulum aminopeptidase-1 mediates leukemia inhibitory factor-induced cell surface
human leukocyte antigen-G expression in JEG-3 choriocarcinoma cells. Endocrinology
2006 Apr;147(4):1780-1788.
250. Blanchard N, Gonzalez F, Schaeffer M, Joncker NT, Cheng T, Shastri AJ, et al.
Immunodominant, protective response to the parasite Toxoplasma gondii requires
antigen processing in the endoplasmic reticulum. Nature immunology 2008
Aug;9(8):937-944.
251. Draenert R, Le Gall S, Pfafferott KJ, Leslie AJ, Chetty P, Brander C, et al.
Immune selection for altered antigen processing leads to cytotoxic T lymphocyte
escape in chronic HIV-1 infection. The Journal of experimental medicine 2004 Apr
5;199(7):905-915.
252. Mehta AM, Jordanova ES, Corver WE, van Wezel T, Uh HW, Kenter GG, et al.
Single nucleotide polymorphisms in antigen processing machinery component ERAP1
195
significantly associate with clinical outcome in cervical carcinoma. Genes,
chromosomes & cancer 2009 May;48(5):410-418.
253. Mehta AM, Jordanova ES, Kenter GG, Ferrone S, Fleuren GJ. Association of
antigen processing machinery and HLA class I defects with clinicopathological
outcome in cervical carcinoma. Cancer immunology, immunotherapy : CII 2008
Feb;57(2):197-206.
254. Mehta AM, Jordanova ES, van Wezel T, Uh HW, Corver WE, Kwappenberg KM,
et al. Genetic variation of antigen processing machinery components and association
with cervical carcinoma. Genes, chromosomes & cancer 2007 Jun;46(6):577-586.
255. Fruci D, Giacomini P, Nicotra MR, Forloni M, Fraioli R, Saveanu L, et al. Altered
expression of endoplasmic reticulum aminopeptidases ERAP1 and ERAP2 in
transformed non-lymphoid human tissues. Journal of cellular physiology 2008
Sep;216(3):742-749.
256. Fruci D, Ferracuti S, Limongi MZ, Cunsolo V, Giorda E, Fraioli R, et al.
Expression of endoplasmic reticulum aminopeptidases in EBV-B cell lines from healthy
donors and in leukemia/lymphoma, carcinoma, and melanoma cell lines. Journal of
immunology (Baltimore, Md.: 1950) 2006 Apr 15;176(8):4869-4879.
257. Yamada Y, Ando F, Shimokata H. Association of candidate gene polymorphisms
with bone mineral density in community-dwelling Japanese women and men.
International journal of molecular medicine 2007 May;19(5):791-801.
258. Haroon N, O'Shea FD, Rahman P, Tsui FW, Inman RD. Inman.ERAP1 R528
Variants Influence the Radiological Progression in Ankylosing Spondylitis. Arthritis and
Rheumatism 2009;60(10):S540.
259. de Castro JA. HLA-B27-bound peptide repertoires: their nature, origin and
pathogenetic relevance. Advances in Experimental Medicine and Biology
2009;649:196-209.
196
260. Haroon N, Inman RD. Endoplasmic reticulum aminopeptidases: Biology and
pathogenic potential. Nature reviews.Rheumatology 2010 Aug;6(8):461-467.
261. Evnouchidou I, Kamal RP, Seregin SS, Goto Y, Tsujimoto M, Hattori A, et al.
Coding single nucleotide polymorphisms of endoplasmic reticulum aminopeptidase 1
can affect antigenic peptide generation in vitro by influencing basic enzymatic
properties of the enzyme. Journal of immunology (Baltimore, Md.: 1950) 2011 Feb
15;186(4):1909-1913.
262. Lee LG, Connell CR, Bloch W. Allelic discrimination by nick-translation PCR with
fluorogenic probes. Nucleic acids research 1993 Aug 11;21(16):3761-3766.
263. Zemmour J, Little AM, Schendel DJ, Parham P. The HLA-A,B "negative" mutant
cell line C1R expresses a novel HLA-B35 allele, which also has a point mutation in the
translation initiation codon. Journal of immunology (Baltimore, Md.: 1950) 1992 Mar
15;148(6):1941-1948.
264. Calvo V, Rojo S, Lopez D, Galocha B, Lopez de Castro JA. Structure and
diversity of HLA-B27-specific T cell epitopes. Analysis with site-directed mutants
mimicking HLA-B27 subtype polymorphism. Journal of immunology (Baltimore, Md.:
1950) 1990 May 15;144(10):4038-4045.
265. Ellis SA, Taylor C, McMichael A. Recognition of HLA-B27 and related antigen by
a monoclonal antibody. Human immunology 1982 Aug;5(1):49-59.
266. Stam NJ, Vroom TM, Peters PJ, Pastoors EB, Ploegh HL. HLA-A- and HLA-B-
specific monoclonal antibodies reactive with free heavy chains in western blots, in
formalin-fixed, paraffin-embedded tissue sections and in cryo-immuno-electron
microscopy. International immunology 1990;2(2):113-125.
267. Urban RG, Chicz RM, Lane WS, Strominger JL, Rehm A, Kenter MJ, et al. A
subset of HLA-B27 molecules contains peptides much longer than nonamers.
Proceedings of the National Academy of Sciences of the United States of America
1994 Feb 15;91(4):1534-1538.
197
268. Ruof J, Stucki G. Validity aspects of erythrocyte sedimentation rate and C-
reactive protein in ankylosing spondylitis: A literature review. Journal of Rheumatology
1999 APR;26(4):966-970.
269. Spoorenberg A, van der Heijde D, de Klerk E, Dougados M, de Vlam K, Mielants
H, et al. Relative value of erythrocyte sedimentation rate and C-reactive protein in
assessment of disease activity in ankylosing spondylitis. The Journal of rheumatology
1999 Apr;26(4):980-984.
270. Alvarez I, Lopez de Castro JA. HLA-B27 and immunogenetics of
spondyloarthropathies. Current opinion in rheumatology 2000 Jul;12(4):248-253.
271. Malik P, Klimovitsky P, Deng LW, Boyson JE, Strominger JL. Uniquely conformed
peptide-containing beta 2-microglobulin-free heavy chains of HLA-B2705 on the cell
surface. Journal of immunology (Baltimore, Md.: 1950) 2002 Oct 15;169(8):4379-4387.
272. Haroon N, Tsui FW, Uchanska-Ziegler B, Ziegler A, Inman RD. Endoplasmic
Reticulum Aminopeptidase 1 (ERAP1) exhibits functionally significant interaction with
HLA B27 and relates to subtype specificity in ankylosing spondylitis. Annals of the
Rheumatic Diseases 2011;(In Press).
273. Hearn A, York IA, Rock KL. The specificity of trimming of MHC class I-presented
peptides in the endoplasmic reticulum. Journal of immunology (Baltimore, Md.: 1950)
2009 Nov 1;183(9):5526-5536.
274. Hillig RC, Hulsmeyer M, Saenger W, Welfle K, Misselwitz R, Welfle H, et al.
Thermodynamic and structural analysis of peptide- and allele-dependent properties of
two HLA-B27 subtypes exhibiting differential disease association. The Journal of
biological chemistry 2004 Jan 2;279(1):652-663.
275. Sims AM, Timms AE, Bruges-Armas J, Burgos-Vargas R, Chou CT, Doan T, et al.
Prospective meta-analysis of interleukin 1 gene complex polymorphisms confirms
associations with ankylosing spondylitis. Annals of the Rheumatic Diseases 2008
Sep;67(9):1305-1309.
198
276. Chan Y, Walmsley RP. Learning and understanding the Kruskal-Wallis one-way
analysis-of-variance-by-ranks test for differences among three or more independent
groups. Physical therapy 1997 Dec;77(12):1755-1762.
277. Neely JG, Hartman JM, Forsen JW,Jr, Wallace MS. Tutorials in clinical research:
VII. Understanding comparative statistics (contrast)--part B: application of T-test,
Mann-Whitney U, and chi-square. The Laryngoscope 2003 Oct;113(10):1719-1725.
278. Freudenberg J, Wang M, Yang Y, Li W. Partial correlation analysis indicates
causal relationships between GC-content, exon density and recombination rate in the
human genome. BMC bioinformatics 2009 Jan 30;10 Suppl 1:S66.
279. Socher SH, Martinez D, Craig JB, Kuhn JG, Oliff A. Tumor necrosis factor not
detectable in patients with clinical cancer cachexia. Journal of the National Cancer
Institute 1988 Jun 15;80(8):595-598.
280. Stone MA, Payne U, Pacheco-Tena C, Inman RD. Cytokine correlates of clinical
response patterns to infliximab treatment of ankylosing spondylitis. Annals of the
Rheumatic Diseases 2004 Jan;63(1):84-87.
281. Ward MM, Reveille JD, Learch TJ, Davis JC,Jr, Weisman MH. Impact of
ankylosing spondylitis on work and family life: comparisons with the US population.
Arthritis and Rheumatism 2008 Apr 15;59(4):497-503.
282. Bennett AN, McGonagle D, O'Connor P, Hensor EM, Sivera F, Coates LC, et al.
Severity of baseline magnetic resonance imaging-evident sacroiliitis and HLA-B27
status in early inflammatory back pain predict radiographically evident ankylosing
spondylitis at eight years. Arthritis and Rheumatism 2008 Nov;58(11):3413-3418.
283. Daoussis D, Liossis SN, Solomou EE, Tsanaktsi A, Bounia K, Karampetsou M, et
al. Evidence that Dkk-1 is dysfunctional in ankylosing spondylitis. Arthritis and
Rheumatism 2010 Jan;62(1):150-158.
199
284. Appel H, Ruiz-Heiland G, Listing J, Zwerina J, Herrmann M, Mueller R, et al.
Altered skeletal expression of sclerostin and its link to radiographic progression in
ankylosing spondylitis. Arthritis and Rheumatism 2009 Nov;60(11):3257-3262.
285. Park MC, Chung SJ, Park YB, Lee SK. Bone and cartilage turnover markers,
bone mineral density, and radiographic damage in men with ankylosing spondylitis.
Yonsei medical journal 2008 Apr 30;49(2):288-294.
286. Vosse D, Landewe R, Garnero P, van der Heijde D, van der Linden S, Geusens
P. Association of markers of bone- and cartilage-degradation with radiological changes
at baseline and after 2 years follow-up in patients with ankylosing spondylitis.
Rheumatology (Oxford, England) 2008 Aug;47(8):1219-1222.
287. Maksymowych WP, Landewe R, Conner-Spady B, Dougados M, Mielants H, van
der Tempel H, et al. Serum matrix metalloproteinase 3 is an independent predictor of
structural damage progression in patients with ankylosing spondylitis. Arthritis and
Rheumatism 2007 Jun;56(6):1846-1853.
288. Baraliakos X, Listing J, von der Recke A, Braun J. The natural course of
radiographic progression in ankylosing spondylitis--evidence for major individual
variations in a large proportion of patients. The Journal of rheumatology 2009
May;36(5):997-1002.
289. Harvey D, Pointon JJ, Evans DM, Karaderi T, Farrar C, Appleton LH, et al.
Investigating the genetic association between ERAP1 and ankylosing spondylitis.
Human molecular genetics 2009 Nov 1;18(21):4204-4212.
290. Ding H, Lu H, Jiang D, Chen Z, Lin J, Wang S. The relationship between
polymorphism of LMP2 and LMP7 genes and the phenotype of ankylosing spondylitis.
Zhonghua yi xue yi chuan xue za zhi = Zhonghua yixue yichuanxue zazhi = Chinese
journal of medical genetics 1999 Aug;16(4):242-245.
291. Maksymowych WP, Adlam N, Lind D, Russell AS. Polymorphism of the LMP2
gene and disease phenotype in ankylosing spondylitis: no association with disease
severity. Clinical rheumatology 1997 Sep;16(5):461-465.
200
292. Maksymowych WP, Jhangri GS, Gorodezky C, Luong M, Wong C, Burgos-Vargas
R, et al. The LMP2 polymorphism is associated with susceptibility to acute anterior
uveitis in HLA-B27 positive juvenile and adult Mexican subjects with ankylosing
spondylitis. Annals of the Rheumatic Diseases 1997 Aug;56(8):488-492.
293. Maksymowych WP, Russell AS. Polymorphism in the LMP2 gene influences the
relative risk for acute anterior uveitis in unselected patients with ankylosing spondylitis.
Clinical and investigative medicine.Medecine clinique et experimentale 1995
Feb;18(1):42-46.
294. Maksymowych WP, Suarez-Almazor M, Chou CT, Russell AS. Polymorphism in
the LMP2 gene influences susceptibility to extraspinal disease in HLA-B27 positive
individuals with ankylosing spondylitis. Annals of the Rheumatic Diseases 1995
Apr;54(4):321-324.
295. Maksymowych WP, Tao S, Vaile J, Suarez-Almazor M, Ramos-Remus C, Russell
AS. LMP2 polymorphism is associated with extraspinal disease in HLA-B27 negative
Caucasian and Mexican Mestizo patients with ankylosing spondylitis. The Journal of
rheumatology 2000 Jan;27(1):183-189.
296. Ploski R, Flato B, Vinje O, Maksymowych W, Forre O, Thorsby E. Association to
HLA-DRB1*08, HLA-DPB1*0301 and homozygosity for an HLA-linked proteasome
gene in juvenile ankylosing spondylitis. Human immunology 1995 Oct;44(2):88-96.
297. Vargas-Alarcon G, Gamboa R, Vergara Y, Rodriguez-Zepeda JM, de la Pena A,
Izaguirre R, et al. LMP2 and LMP7 gene polymorphism in Mexican populations:
Mestizos and Amerindians. Genes and immunity 2002 Sep;3(6):373-377.
298. Vargas-Alarcon G, Gamboa R, Zuniga J, Fragoso JM, Hernandez-Pacheco G,
Londono J, et al. Association study of LMP gene polymorphisms in Mexican patients
with spondyloarthritis. Human immunology 2004 Dec;65(12):1437-1442.
299. Feng M, Yin B, Shen T, Ma Q, Liu L, Zheng J, et al. TAP1 and TAP2
polymorphisms associated with ankylosing spondylitis in genetically homogenous
Chinese Han population. Human immunology 2009 Apr;70(4):257-261.
201
300. Fraile A, Collado MD, Mataran L, Martin J, Nieto A. TAP1 and TAP2
polymorphism in Spanish patients with ankylosing spondylitis. Experimental and
clinical immunogenetics 2000;17(4):199-204.
301. Konno Y, Numaga J, Mochizuki M, Mitsui H, Hirata R, Maeda H. TAP
polymorphism is not associated with ankylosing spondylitis and complications with
acute anterior uveitis in HLA-B27-positive Japanese. Tissue antigens 1998
Nov;52(5):478-483.
302. Maksymowych WP, Tao S, Li Y, Wing M, Russell AS. Allelic variation at the TAP
1 locus influences disease phenotype in HLA-B27 positive individuals with ankylosing
spondylitis. Tissue antigens 1995 May;45(5):328-332.
303. Wang B, Su H, Gao P, Tao L, Li X, Yie D. Study on the association of transporter
associated with antigen processing gene with ankylosing spondylitis and HLA-B27 in
Han nationality. Zhonghua yi xue yi chuan xue za zhi = Zhonghua yixue yichuanxue
zazhi = Chinese journal of medical genetics 2001 Jun;18(3):209-212.
304. Westman P, Partanen J, Leirisalo-Repo M, Koskimies S. TAP1 and TAP2
polymorphism in HLA-B27-positive subpopulations: no allelic differences in ankylosing
spondylitis and reactive arthritis. Human immunology 1995 Dec;44(4):236-242.
305. Niu Z, Lei R, Shi J, Wang D, Shou W, Wang Z, et al. A polymorphism rs17336700
in the PSMD7 gene is associated with ankylosing spondylitis in Chinese subjects.
Annals of the Rheumatic Diseases 2010 Jul 19.
306. Creemers MC, Franssen MJ, van't Hof MA, Gribnau FW, van de Putte LB, van
Riel PL. Assessment of outcome in ankylosing spondylitis: an extended radiographic
scoring system. Annals of the Rheumatic Diseases 2005 Jan;64(1):127-129.
307. Mehta AM, Jordanova ES, van Wezel T, Uh HW, Corver WE, Kwappenberg KM,
et al. Genetic variation of antigen processing machinery components and association
with cervical carcinoma. Genes, chromosomes & cancer 2007 Jun;46(6):577-586.
202
308. Lee W, Reveille JD, Davis JC,Jr, Learch TJ, Ward MM, Weisman MH. Are there
gender differences in severity of ankylosing spondylitis? Results from the PSOAS
cohort. Annals of the Rheumatic Diseases 2007 May;66(5):633-638.
309. Doran MF, Brophy S, MacKay K, Taylor G, Calin A. Predictors of longterm
outcome in ankylosing spondylitis. The Journal of rheumatology 2003 Feb;30(2):316-
320.
310. van der Heijde D, Salonen D, Weissman BN, Landewe R, Maksymowych WP,
Kupper H, et al. Assessment of radiographic progression in the spines of patients with
ankylosing spondylitis treated with adalimumab for up to 2 years. Arthritis research &
therapy 2009;11(4):R127.
311. van der Heijde D, Landewe R, Baraliakos X, Houben H, van Tubergen A,
Williamson P, et al. Radiographic findings following two years of infliximab therapy in
patients with ankylosing spondylitis. Arthritis and Rheumatism 2008 Oct;58(10):3063-
3070.
312. van der Heijde D, Landewe R, Einstein S, Ory P, Vosse D, Ni L, et al.
Radiographic progression of ankylosing spondylitis after up to two years of treatment
with etanercept. Arthritis and Rheumatism 2008 May;58(5):1324-1331.
313. Fruh K, Yang Y. Antigen presentation by MHC class I and its regulation by
interferon gamma. Current opinion in immunology 1999 Feb;11(1):76-81.
314. Smith JA, Barnes MD, Hong D, DeLay ML, Inman RD, Colbert RA. Gene
expression analysis of macrophages derived from ankylosing spondylitis patients
reveals interferon-gamma dysregulation. Arthritis and Rheumatism 2008
Jun;58(6):1640-1649.
315. Hayashi T, Faustman D. Defective function of the proteasome in autoimmunity:
involvement of impaired NF-kappaB activation. Diabetes technology & therapeutics
2000 Autumn;2(3):415-428.
203
316. Hayashi T, Faustman D. A role for NF-kappaB and the proteasome in
autoimmunity. Archivum Immunologiae et Therapiae Experimentalis 2000;48(5):353-
365.
317. Hayashi T, Faustman D. Essential role of human leukocyte antigen-encoded
proteasome subunits in NF-kappaB activation and prevention of tumor necrosis factor-
alpha-induced apoptosis. The Journal of biological chemistry 2000 Feb
18;275(7):5238-5247.
318. Hayashi T, Faustman D. NOD mice are defective in proteasome production and
activation of NF-kappaB. Molecular and cellular biology 1999 Dec;19(12):8646-8659.
319. Maniatis T. A ubiquitin ligase complex essential for the NF-kappaB,
Wnt/Wingless, and Hedgehog signaling pathways. Genes & development 1999 Mar
1;13(5):505-510.
320. Uderhardt S, Diarra D, Katzenbeisser J, David JP, Zwerina J, Richards WG, et al.
Blockade of Dickkopf-1 induces fusion of sacroiliac joints. Annals of the Rheumatic
Diseases 2009 Mar 26.
321. Yavropoulou MP, Yovos JG. The role of the Wnt signaling pathway in osteoblast
commitment and differentiation. Hormones (Athens, Greece) 2007 Oct-Dec;6(4):279-
294.
322. Wang K, Niu J, Kim H, Kolattukudy PE. Osteoclast precursor differentiation by
MCPIP via oxidative stress, endoplasmic reticulum stress, and autophagy. Journal of
molecular cell biology 2011 Dec;3(6):360-368.
323. Forloni M, Albini S, Limongi MZ, Cifaldi L, Boldrini R, Nicotra MR, et al. NF-
kappaB, and not MYCN, regulates MHC class I and endoplasmic reticulum
aminopeptidases in human neuroblastoma cells. Cancer research 2010 Feb
1;70(3):916-924.
204
324. Brophy S, Mackay K, Al-Saidi A, Taylor G, Calin A. The natural history of
ankylosing spondylitis as defined by radiological progression. The Journal of
rheumatology 2002 Jun;29(6):1236-1243.
![Endoplasmic reticulum[1]](https://static.fdocuments.in/doc/165x107/58ed5fc71a28aba1678b4611/endoplasmic-reticulum1.jpg)
