A profile of immune response to herpesvirus is associated with radiographic joint damage in rheumatoid arthritis.

Abstract

Progression of joint damage despite appropriate therapy remains a significant problem for patients with rheumatoid arthritis (RA). This study was undertaken to identify profiles of immune response that correlate with radiographic joint damage as a first step toward the discovery of new pathogenic mechanisms of joint destruction in RA.
The study included 58 patients with RA and 15 healthy controls. The profiles of cytokine release from peripheral blood mononuclear cells (PBMC) in response to stimulation for 48 hours with one of six stimuli, or in media alone, were measured. Immune response profiles identified for each stimulus were correlated with radiographic joint damage as defined by the Sharp-van der Heijde score (SHS), before and after multivariable adjustment. For profiles correlated with the SHS, the distributions of individual cytokines were evaluated in patients according to the severity of joint damage and compared to healthy controls.
The immune response profile for cytomegalovirus (CMV)/Epstein-Barr virus (EBV) stimulation was correlated with both the SHS total and erosion scores (r = 0.31, P = 0.018 and r = 0.33, P = 0.011, respectively). After adjusting for age, sex, disease duration, autoantibody status, CMV/EBV serological status, current disease activity, disability and treatments, the correlation of the CMV/EBV immune response and the SHS erosion score became stronger (r = 0.43, P < 0.003). The CMV/EBV immune response correlated with CMV IgG (r = 0.44, P < 0.001), but not with EBV IgG. The most important cytokines for the CMV/EBV immune response profile were IFN-γ, IL-2, IL-4, IL-5, IL-13 and IL-17A, all of which are associated with T-cell immunity. Both the summary immune response score and the individual responses of IFN-γ and IL-13 to CMV/EBV stimulation were associated with greater joint damage.
A profile of immune response to purified CMV/EBV lysates is associated with radiographic joint damage. The correlation of this immune response to CMV serology implies possible involvement of latent CMV infection. Therefore, the findings suggest that the immune response to latent CMV infection could play a fundamental role in the progression of inflammation and structural joint damage in patients with RA.

Full text

Page 1

RESEARCH ARTICLEOpen Access
A profile of immune response to herpesvirus is
associated with radiographic joint damage in
rheumatoid arthritis
John M Davis III1*, Keith L Knutson2, John A Skinner3, Michael A Strausbauch4, Cynthia S Crowson5,
Terry M Therneau5, Peter J Wettstein4, Eric L Matteson1,6and Sherine E Gabriel1,6
Abstract
Introduction: Progression of joint damage despite appropriate therapy remains a significant problem for patients
with rheumatoid arthritis (RA). This study was undertaken to identify profiles of immune response that correlate
with radiographic joint damage as a first step toward the discovery of new pathogenic mechanisms of joint
destruction in RA.
Methods: The study included 58 patients with RA and 15 healthy controls. The profiles of cytokine release from
peripheral blood mononuclear cells (PBMC) in response to stimulation for 48 hours with one of six stimuli, or in
media alone, were measured. Immune response profiles identified for each stimulus were correlated with
radiographic joint damage as defined by the Sharp-van der Heijde score (SHS), before and after multivariable
adjustment. For profiles correlated with the SHS, the distributions of individual cytokines were evaluated in patients
according to the severity of joint damage and compared to healthy controls.
Results: The immune response profile for cytomegalovirus (CMV)/Epstein-Barr virus (EBV) stimulation was correlated
with both the SHS total and erosion scores (r = 0.31, P = 0.018 and r = 0.33, P = 0.011, respectively). After adjusting
for age, sex, disease duration, autoantibody status, CMV/EBV serological status, current disease activity, disability and
treatments, the correlation of the CMV/EBV immune response and the SHS erosion score became stronger (r =
0.43, P < 0.003). The CMV/EBV immune response correlated with CMV IgG (r = 0.44, P < 0.001), but not with EBV
IgG. The most important cytokines for the CMV/EBV immune response profile were IFN-g, IL-2, IL-4, IL-5, IL-13 and
IL-17A, all of which are associated with T-cell immunity. Both the summary immune response score and the
individual responses of IFN-g and IL-13 to CMV/EBV stimulation were associated with greater joint damage.
Conclusions: A profile of immune response to purified CMV/EBV lysates is associated with radiographic joint
damage. The correlation of this immune response to CMV serology implies possible involvement of latent CMV
infection. Therefore, the findings suggest that the immune response to latent CMV infection could play a
fundamental role in the progression of inflammation and structural joint damage in patients with RA.
Keywords: rheumatoid arthritis, RA, immune responses, cytokines, T cells, radiographic joint damage, cytomegalo-
virus, Epstein-Barr virus
* Correspondence: davis.john4@mayo.edu
1Division of Rheumatology, Department of Medicine, College of Medicine,
Mayo Clinic; 200 First Street SW, Rochester, MN 55905, USA
Full list of author information is available at the end of the article
Davis et al. Arthritis Research & Therapy 2012, 14:R24
http://arthritis-research.com/content/14/1/R24
© 2012 Davis et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.

Page 2

Introduction
Rheumatoid arthritis (RA) is a systemic autoimmune
disease that causes chronic, persistent joint inflamma-
tion, leading to irreversible structural damage. The etiol-
ogy remains unclear, but the prevailing disease model is
that multiple genes, especially HLA-DRB1 alleles, inter-
act with environmental risk factors, including tobacco
smoking, culminating in an adaptive immune response
to citrullinated autoantigens that perpetuates joint
inflammation [1]. T lymphocytes are key “conductors”
of the inflammatory process, providing help to B cells
and leading to the production of autoantibodies and the
recruitment/activation of other immune effectors [2,3].
Ultimately, the activation of plasma cells, macrophages,
fibroblasts, chondrocytes and osteoclasts, and down-
stream production of inflammatory cytokines (that is,
TNF, IL-1, and IL-6), reactive oxygen species, matrix
metalloproteinases and other toxic molecules, mediates
destruction of articular cartilage and bone [4].
Advances in therapeutics have dramatically improved
the outlook for joint structure and function in patients
with RA. Current treatment strategies retard joint
destruction in the majority. Nonetheless, between 5 to
30% of patients, depending on the treatment strategy
used, still experience rapid progression of joint damage
[5,6]. Models describing the probability of rapid radio-
graphic progression have limited accuracy and reliability
and do not fully account for joint damage [7-9]. More
patients suffer slow structural deterioration over time,
which can occur despite intensive treatment [10-12].
The most likely explanation is that inflammation can
persist despite treatment, even when the disease appears
to be in clinical remission [13-16]. Therefore, there is an
urgent need to identify new pathogenic mechanisms of
persistent, treatment-refractory inflammation, in hopes
of developing novel targeted therapies that prevent
structural deterioration over time.
We have devised an approach to discover immunolo-
gical correlates of disease phenotypes based on profiles
of ex vivo cytokine production from peripheral blood
mononuclear cells (PBMC) in response to a number of
non-specific stimuli. This approach was based on the
concept that we could sample a broad array of activa-
tion pathways and, thereby, develop a general profile of
immune responsiveness rather than define the respon-
siveness of specific innate or adaptive immune pathways.
Our previous studies demonstrated the success of this
approach by characterizing differences in the profiles of
immune response between patients with early versus
late RA and healthy controls, and also between patients
with and without myocardial dysfunction [17,18]. The
results informed our hypothesis that the broad respon-
siveness of one or more immune pathways is associated
with the potential of the inflammatory process to med-
iate joint destruction. The objective of this study was to
identify a profile of immune response based on ex vivo
cytokine production that is associated with the severity
of radiographic joint damage in patients with RA, after
accounting for known predictors of severe disease.
Materials and methods
Study design and participants
We conducted a cross-sectional correlative study in the
outpatient practice of the Division of Rheumatology at
Mayo Clinic Rochester, MN, USA. A sample of patients
referred by the 16 clinicians in the Division during the
period of May 2008 through November 2009 was
recruited. Adult patients of ≥ 18 years of age with a
diagnosis of RA were eligible to participate if they met
the following inclusion criteria: fulfillment of the 1987
American College of Rheumatology classification criteria
for RA and seropositive status for rheumatoid factor
(RF), anti-citrullinated protein antibodies (ACPA), or
both. Patients with any of the following were excluded:
clinically apparent acute infections, known chronic
infections (that is, hepatitis C), recent malignancies in
the past two years (excluding non-melanoma skin can-
cers), radiation or chemotherapy in the past two years,
history of advanced chronic kidney disease or kidney
transplantation, history of chronic liver disease, recent
major surgery, or use of prednisone for indications
other than RA. For controls, we included data on
healthy blood donors from our initial study, which used
identical methods as the present one [17]. The Mayo
Foundation institutional review board approved this
study, which was conducted according to the principles
of the Declaration of Helsinki. All patients provided
written informed consent.
Data collection
One consultant rheumatologist (JMD) evaluated tender
and swollen joints using modified 28-joint counts [19]
and the physician global assessment of disease activity
using a 0 to 100 mm visual analog scale (VAS). Patient
reported outcomes included levels of pain and fatigue (0
to 100 mm VAS), duration of morning stiffness (min-
utes), the Health Assessment Questionnaire (HAQ) dis-
ability index [20,21], and the Medical Outcomes Study
Short-Form 36 (SF-36) [22,23]. Disease activity was
defined by the Disease Activity Score in 28 joints, four-
variable version, using C-reactive protein (DAS28) [19].
Data were collected on patient demographics, disease
duration, RF and ACPA status, and current use of dis-
ease-modifying antirheumatic drugs (DMARDs), includ-
ing biologic response modifiers. C-reactive protein
(CRP) was measured by turbidometric assay (Roche,
Davis et al. Arthritis Research & Therapy 2012, 14:R24
http://arthritis-research.com/content/14/1/R24
Page 2 of 11

Page 3

Indianapolis, IN, USA). Enzyme-linked immunosorbent
assays were done to assess past CMV infection using
the VIDAS®CMV IgG (bioMerieux, Inc., Durham, NC,
USA), and multiplexed immunoassays were done to
assess past EBV infection using the BioPlex™ 2200 Sys-
tem EBV IgG and EBV IgM (Bio-Rad Laboratories, Her-
cules, CA, USA).
Radiographic scoring
Radiographs, including single views of both hands and
feet, were obtained using a standardized research proto-
col. A board-certified musculoskeletal radiologist (JAS)
interpreted all radiographs blinded to clinical and
experimental information, according to the modified
Sharp-van der Heijde score (SHS) system [24,25]. Sepa-
rate evaluations were done as planned a priori for the
total SHS as well as erosion and joint space narrowing
(JSN) scores.
Immune response assays
Our approach to quantify the broad responsiveness of ex
vivo cytokine production was previously described [17].
Functional peripheral blood mononuclear cells (PBMCs)
were procured by Ficoll density gradient centrifugation.
Within one to two hours, 4 × 105PBMCs were cultured
in 200 μL of medium (RPMI-1640 + 10% fetal bovine
serum + 1× penicillin, streptomycin, and glutamine) in
the presence of one of a panel of six stimuli, or in med-
ium alone, in a 96-well culture plate. The rationale for
the use of each stimulus was detailed in our previous
publication [17]. The stimulation panel included immo-
bilized anti-CD3 and anti-CD28 monoclonal antibodies
(anti-CD3/anti-CD28; Dynabeads Human T-Activator,
Invitrogen, Grand Island, NY, USA); bacterial CpG oli-
gonucleotides (CpG); combined lysates of purified CMV
and EBV, containing both viral peptides and DNA
(CMV/EBV; Advanced Biotechnologies, Columbia, MD,
USA); PMA with ionomycin (PMA/ionomycin; Sigma,
St. Louis, MO, USA); phytohemagglutinin (PHA;
Sigma); and staphylococcus enterotoxins A and B (SEA
and SEB; Toxin Technology, Sarasota, FL, USA). The
final concentrations of each stimulant were as follows:
anti-CD3/anti-CD28, 0.5 × 106beads per culture well
(1:1 ratio of beads to PBMCs per manufacturer’s
instructions); PHA, 5 μg/mL; Staphylococcus enterotoxin
A, 10 ng/mL, with Staphylococcus enterotoxin B, 10 ng/
mL; CMV, 1 μg/mL, with EBV, 1 μg/mL; CpG, 10 μg/
mL; PMA, 1 μg/mL, with ionomycin, 700 ng/mL. The
PBMCs were incubated at 37°C in 5% CO2for 48 hours;
the supernatants were then harvested, transferred to a
storage plate, and frozen at -80°C for subsequent
analysis.
Multiplexed cytokine analysis was performed to mea-
sure cytokine release from the PBMCs into the culture
supernatants in response to stimulation using the MSD®
96-Well MULTI-SPOT®Human Cytokine Assays tissue
culture kit (Meso Scale Discovery (MSD), Gaithersburg,
MD, USA). The cytokine panel included IL-1b, IL-2, IL-
4, IL-5, IL-6, IL-7, IL-8 (CXCL8), IL-10, IL-12, IL-13,
IL-17A, IFN-g, TNF-a, MCP-1 (CCL2), MIP-1b (CCL4),
G-CSF, and GM-CSF. Using the Sector 2400 instrument
(MSD) and manufacturer-supplied reagents, the cytokine
concentrations were determined based on a standard
curve generated for each plate. The results were ana-
lyzed using the Discovery Workbench Software v2.0
(MSD). Previous studies have demonstrated acceptable
reproducibility relative to the high degree of informative,
biological variation [17].
Statistical analysis
The patient characteristics were summarized as median
(interquartile range (IQR)) or number (%). Mixed effects
models were used to normalize the cytokine data as pre-
viously described [17]. In brief, fixed effects were
included for age, sex, cytokine-plate lot, and random
effects were included for subject and individual cytokine
plate used. This resulted in effective normalization of
assay effects and adjustment of the cytokine data for age
and sex.
The analytic approach was based on the premise that
each stimulus activates a finite number of immune
pathways. The foundational concept of principal com-
ponents analysis (PCA) is that for a large number of
independent variables, common pathways can be iden-
tified by variables that are highly correlated. We used
PCA to identify immune response profiles associated
with each stimulus based on the first and second prin-
cipal components. To create stimulus-cytokine PCA
scores for the first and second principal components
of each stimulus, all cytokine concentrations were con-
verted to Z-scores (the mean cytokine value for the
patients was subtracted from each individual cytokine
value, and this difference was divided by the standard
deviation of the cytokine value), then the Z-scores
were multiplied by the component loadings and
summed to obtain a single continuous score. For ease
of interpretation, these scores were rescaled to range
from 0 to 100.
The study objective was to determine the associations
between key immune profiles for each stimulus and
radiographic joint damage. Spearman correlations were
determined between each of the stimulus-cytokine PCA
scores and the SHS total, erosion, and JSN scores. Par-
tial Spearman correlations were also assessed after
adjusting for age, sex, disease duration, RF status, ACPA
status, CMV/EBV serological status, RA drug treatments
(methotrexate, biologic response modifiers, and predni-
sone), the HAQ disability index, and DAS28.
Davis et al. Arthritis Research & Therapy 2012, 14:R24
http://arthritis-research.com/content/14/1/R24
Page 3 of 11

Page 4

For stimulus-cytokine scores significantly correlated
with joint damage, individual cytokine values with com-
ponent loadings > 0.5 were compared between RA
patients and controls using Wilcoxon rank-sum tests.
The patients were analyzed in two groups based on a
dichotomous cutoff for the total SHS of 20 units. The
rationale for this particular cutoff was that the progres-
sion of radiographic joint damage by 20 Sharp units cor-
responds to an irreversible deterioration in physical
function as defined by the HAQ disability index of 0.2,
approximately the minimum clinically important differ-
ence for this outcome [26]. The rescaled stimulus-cyto-
kine PCA score was also analyzed among the groups.
Statistical significance was set at the 0.05 level; all tests
were two-sided.
Results
Subject characteristics
The study included 58 patients with RA (Table 1) and
15 controls. On average, the patients had relatively early
disease, with median (IQR) disease duration of 18.8
months (11.3, 35.9). According to the selection criteria,
all patients were positive for ACPA or RF; 44 (80%)
were positive for ACPA, and 53 (93%) were positive for
RF. Overall, the cohort had moderate disease activity,
with a median (IQR) DAS28 of 4.4 (3.4, 5.4), and mod-
erate disability, with a median (IQR) HAQ disability
index of 0.5 (0.1, 1.4). Overall, 56 (97%) of the patients
had erosive disease as defined by the SHS erosion score
≥ 1. The majority of patients were taking methotrexate,
and nearly half were taking prednisone; only 14 were
taking biologic response modifiers (12 patients were tak-
ing anti-TNF agents, and 2 patients had received rituxi-
mab). The control group included eight females and
seven males with a mean (SD) age of 44.6 (16.8) years.
Discovery of a correlation between CMV/EBV immunity
and radiographic joint damage using principal
components analysis
The analysis of the first and second principal compo-
nents for each of the seven stimulus-cytokine combina-
tions resulted in identification of distinct cytokine
profiles; based on the analytical premise, these were
considered to reflect activation of canonical immune
pathways (Table 2). For CMV/EBV stimulation, the sti-
mulus-cytokine combinations with component loadings
> 0.5 were IL-13, IL-2, IFN-g, IL-5, IL-4, and IL-17A (in
diminishing importance in the PCA score). This profile
clearly reflected activation of T-cell cytokine production,
across canonical T-cell subsets, in response to CMV/
EBV lysate. The first and second principal components
explained 30.2% and 20.7%, respectively, of the variation
in cytokine production from PBMC in response to
CMV/EBV stimulation.
The next step was to screen each of the immune
response profiles identified using PCA for association
with radiographic joint damage as defined by the SHS
scores. Among all of the first and second principal com-
ponents for the seven stimulation conditions, only one
had a correlation coefficient of r > 0.3, a level which
might be biologically meaningful: the first principal
component for CMV/EBV stimulation (CMV/EBV-1)
(Figure 1). Statistically significant correlations of the
CMV/EBV-1 PCA score with the total SHS (r = 0.31; P
= 0.018) and the SHS erosion score (r = 0.33; P = 0.011)
were identified (Table 3). After adjusting for potential
confounders of the relationship between viral immunity
and joint damage (Table 3 and Figure 1), including mar-
kers of RA disease activity, severity and treatment status,
the correlation of CMV/EBV-1 with both the SHS total
and erosion scores became modestly stronger (r = 0.39;
P = 0.006; see Table 3 for the list of covariates).
Additionally, a significant correlation of the PMA/
ionomycin second principal component (PMA/ionomy-
cin-2) with the erosion score (r = 0.27; P = 0.042), but
not the total SHS (r = 0.24; P = 0.07), was identified
Table 1 Characteristics of 58 patients with rheumatoid
arthritis
Variable Statistic
Age, years
Sex, female
RA disease durations, months†
Pain (0 to 100 mm)
Patient global assessment (0 to 100 mm)
Morning stiffness duration, minutes
HAQ disability index (0 to 3)
RF positive‡
ACPA positive§
C-reactive protein, mg/L
CMV IgG positive
EBV IgG positive
DAS28
Radiographic damage scores (SHS), units
Total
Erosions
Joint space narrowing
Medications
Methotrexate
Biologic response modifiers
Corticosteroids
54.0 (47.9, 63.0)
34 (59%)
18.8 (11.3, 35.9)
30 (16, 69)
28.5 (12, 63)
30 (15, 90)
0.5 (0.1, 1.4)
53 (93%)
44 (80%)
5.9 (3.0, 11.1)
28 (48%)
55 (95%)
4.4 (3.4, 5.4)
18.5 (8, 37)
11.5 (3, 30)
4.5 (3, 12)
49 (86%)
14 (24%)
27 (47%)
*Values are median (interquartile range) for continuous variables or number
(%) for categorical variables.†The duration of disease from onset of arthritis
symptoms.‡Of 57 patients tested.§Of 55 patients tested. ACPA, anti-
citrullinated protein antibodies; CMV, cytomegalovirus; DAS28, Disease Activity
Score in 28 joints, 4-variable version, using the C-reactive protein; EBV,
Epstein-Barr virus; HAQ, Health Assessment Questionnaire; IgG,
immunoglobulin G; RA, rheumatoid arthritis; RF, rheumatoid factor; SHS,
Sharp-van der Heijde score
Davis et al. Arthritis Research & Therapy 2012, 14:R24
http://arthritis-research.com/content/14/1/R24
Page 4 of 11

Page 5

(Table 3). After adjustment for covariates, the associa-
tion between PMA/ionomycin-2 and the erosion score
became non-significant (r = 0.09, P = 0.53). This asso-
ciation was confounded by the use of prednisone and
disease duration, both of which correlated with PMA/
ionomycin-2 (r = -0.26, P = 0.046 and r = 0.29, P =
0.025, respectively).
Analysis of the CMV/EBV immune response profile in
patients with RA and control subjects according to viral
serological status
The next step was to analyze the individual cytokines
comprising CMV/EBV-1 among the patient groups with
‘high’ or ‘low’ joint damage and healthy controls. Signifi-
cant differences in the production of T-cell cytokines
were observed among the groups (Figure 2A). The RA
group with high joint damage (SHS ≥ 20) had signifi-
cantly increased production of IFN-g (Th1) and IL-13
(Th2) compared to the group with low joint damage
(SHS < 20). IL-2, the other Th1 cytokine studied, had a
slightly lower median in the overall RA group relative to
controls (statistically non-significant), yet there was a
subgroup of patients with high joint damage who had
markedly increased levels compared to the low-damage
group or controls. The production of IL-4 and IL-5, the
other Th2 cytokines, in response to CMV/EBV stimula-
tion was significantly increased among the patients
Table 2 Identification of immune response profiles by means of principal components analysis*
Cytokine CMV/EBV SEA/SEB CPG PMA/ionomycin PHA CD3/CD28Media Alone
PC-1 PC-2PC-1PC-2 PC-1PC-2 PC-1 PC-2PC-1PC-2PC-1PC-2PC-1PC-2
IL-17
GM-CSF
IL-1b
IL-2
IL-4
IL-5
IL-13
IL-10
IL-12
IFN-g
MCP-1
MIP-1b
IL-6
TNF-a
IL-8
IL-7
G-CSF
0.54
0.46
0.25
0.93
0.78
0.79
0.93
0.27
0.52
0.74
0.75
0.76
0.68
0.10
0.13
0.30
0.39
0.40
0.35
0.36
0.31
0.45
0.95
-0.27
0.11
0.13
-0.13
0.38
0.73
0.76
0.17
0.58
0.36
0.24
0.34
0.92
0.27
0.39
0.25
0.29
0.59
0.43
0.20
0.20
0.73
0.77
0.79
0.87
0.41
0.25
0.63
0.29
0.37
0.20
0.49
-0.23
0.22
0.29
0.81
0.89
0.18
0.37
0.47
0.31
0.45
0.56
0.23
0.86
0.87
0.92
0.14
0.55
0.66
0.26
0.32
0.31
0.50
0.34
0.75
-0.31
0.34
0.31
0.86
0.29
0.18
0.81
0.96
0.25
0.77
0.79
0.89
0.38
0.85
0.77
0.24
0.18
0.87
0.28
0.90
0.71
0.34
0.27
0.85
0.41
-0.12
0.12
-0.17
-0.16
0.33
-0.200.13
0.56
0.48
0.85
0.48
0.29
-0.21
0.39
-0.37
-0.26
0.36
0.60
0.72
0.71
0.91
-0.77
-0.15
0.82
-0.13
0.13
0.88
0.35
0.95
0.80
-0.88
0.45
0.93
0.53
0.71
0.43
0.76
0.65
-0.70
0.40
0.85
0.43
0.94
0.67
0.93
0.87
-0.19
0.27
0.20
0.86
0.33
-0.81
0.72
0.57
0.87
0.47
-0.49
0.23
0.84
0.19
0.25
0.23
0.23
0.15
0.57
0.79
0.56
-0.25
0.29
0.15
0.20
0.23
0.67
-0.29
0.11
-0.12
0.20
0.56
0.27
-0.13-0.13
0.96
0.15-0.14-0.17 0.19
0.86
*Principal components analysis was used to identify profiles of ex vivo cytokine production by peripheral blood mononuclear cells in response to stimulation.
Shown are the component loadings of each cytokine in the first and second principal components of each stimulus. Bold values indicate loadings > 0.5. CD3/
CD28, anti-CD3/anti-CD28 monoclonal antibodies; CMV/EBV, cytomegalovirus/Epstein-Barr virus; CPG, CpG oligonucleotides; G-CSF, granulocyte colony stimulating
factor; GM-CSF, granulocyte macrophage colony stimulating factor; IFN-g, interferon gamma; IL, interleukin; MCP, monocyte chemoattractant protein; MIP,
monocyte inflammatory protein; PC, principal component; PHA, phytohemagglutinin; PMA, phorbol myristate acetate; SEA/SEB, Staphylococcal enterotoxin A/
Staphylococcal enterotoxin B; TNF, tumor necrosis factor
UnadjustedAdjusted
- 1 | 8 7 4 1
- 0 |9 4 3 -0|9 8 5 5 5 5 1 0 0
0| 4 70|1 5 7
1|71|5
2|0 2 4 2|
3
|13|9
BA
Figure 1 Results of the data reduction strategy identifying
immune response profiles correlated with radiographic joint
damage. Shown are stem-and-leaf schematics summarizing the
hierarchy of correlation coefficients for each of the stimulus-cytokine
principal component analysis scores. (A) Unadjusted Spearman
correlation coefficients. (B) Partial Spearman correlation coefficients,
adjusted for age, sex, disease duration, rheumatoid factor status, anti-
citrullinated protein antibody status, the Health Assessment
Questionnaire (HAQ) disability index, the Disease Activity Score in 28
joints (DAS28), and use of methotrexate, biologic response modifiers,
and/or prednisone. The decimal point is one digit to the left of the
vertical bar (for example, the bottom value in the left panel is r = 0.31).
The highest unadjusted and adjusted correlation coefficients (r = 0.31
and r = 0.39, respectively) correspond to the CMV/EBV first principal
component score. CMV, cytomegalovirus; EBV, Epstein Barr virus
Davis et al. Arthritis Research & Therapy 2012, 14:R24
http://arthritis-research.com/content/14/1/R24
Page 5 of 11