Download Bone resorption correlates with the frequency of CD5+ B cells in the

RHEUMATOLOGY
Rheumatology 2015;54:545–553
doi:10.1093/rheumatology/keu351
Advance Access publication 5 September 2014
Original article
Bone resorption correlates with the frequency of
CD5+ B cells in the blood of patients with rheumatoid
arthritis
Robby Engelmann1, Ni Wang1,2, Christian Kneitz3 and Brigitte Müller-Hilke1
Abstract
Objective. The prevention of bone resorption and subsequent joint destruction is one of the main challenges in the treatment of patients suffering from RA. Various mechanisms have previously been described
that contribute to bone resorption in tightly defined cohorts. Here we analysed a cross-sectional cohort of
RA patients and searched for humoral and cellular markers in the peripheral blood associated with bone
resorption.
Methods. We enrolled 61 consecutive RA patients positive for ACPA. Blood was analysed by flow
cytometry to determine the percentages of regulatory T cells and B cell subpopulations. Cytokine
(TNF-a, IL-6, IL-10) and ACPA levels as well as the bone resorption marker CTX-1 were determined
from the patients’ sera. Standard clinical disease parameters were included.
Results. Multivariate analyses showed that the percentages of CD5+ B cells were positively correlated
with CTX-1 serum levels. However, neither low-avidity ACPA nor serum IL-6 levels, both known to be
produced by CD5+ cells, were associated with CTX-1 in patients’ sera. There was no correlation between
CTX-1 levels and clinical parameters or ACPA levels.
Key words: bone resorption, CTX-1, CD5, rheumatoid arthritis, biomarker.
Introduction
The hallmark of RA is chronic joint inflammation. Cells
of the innate and adaptive immune systems infiltrate the
synovial membrane and lead to the formation of the highly
aggressive pannus tissue that is held to be responsible
for joint destruction and local bone erosion [1, 2]. Bone
erosion results from the uncoupling of bone formation and
bone resorption and several attempts have been made to
link chronic inflammation with elevated osteoclast activity
[1, 3] and reduced bone formation [4] leading to the focal
1
Institute of Immunology, Rostock University Medical Center, Rostock,
Germany, 2Institute of Blood Research, Dalian Blood Center, Liaoning
Province, China and 3Klinik für Innere Medizin II, Klinikum Südstadt
Rostock, Rostock, Germany.
Submitted 31 December 2013; revised version accepted 3 July 2014.
Correspondence to: Robby Engelmann, Institute of Immunology,
Rostock University Medical Center, Schillingallee 68, 18057 Rostock,
Germany. E-mail: [email protected]
bone loss characteristic of RA. Indeed, the close phylogenetic relatedness between bone and the immune
system allows for an intimate communication that makes
use of numerous cellular and humoral factors [5, 6].
Among the cellular factors, the receptor activator of
nuclear factor kB ligand (RANKL) is expressed on osteoblasts and Th17 cells alike and initiates osteoclast maturation [7–9]. Humoral factors include cytokines as well as
chemokines and growth factors. Thus pro-inflammatory
cytokines such as TNF-a, IL-6 and IL-17, which are predominantly produced by activated macrophages, B cells
and T cells, respectively, have been shown to stimulate
osteoclast differentiation and thereby enhance bone
resorption [7, 10]. TGF-b, which is produced by regulatory
T cells (Tregs), Th cells and B cells alike, plays an important and dual role. On the one hand, it impacts directly
on the osteoblast to produce osteoprotegerin, which
prevents osteoclast maturation [11]. On the other hand,
TGF-b balances Th17 and Tregs [6]: while low levels of
! The Author 2014. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: [email protected]
BASIC
SCIENCE
Conclusion. In summary, we found that the CD5+ B cell population is associated with bone resorption as
measured via serum CTX-1 levels in a cross-sectional cohort of RA patients. However, a possible functional link between CD5+ B cells and bone resorption still needs to be defined.
Robby Engelmann et al.
TGF-b in combination with IL-6 promote differentiation of
inflammatory Th17 cells, high levels induce the generation
of Tregs that secrete TGF-b and IL-10 and thereby inhibit
osteoclastogenesis [12]. The contribution of B cells to
bone erosion in RA is still a matter of debate. Some
authors have demonstrated that B cells enhance osteoclast maturation via expression of RANKL [13] or via an
elevated population of CD5+ B cells that, within the synovium of RA patients [14], produce considerable amounts
of IL-6 [15]. Conversely, others have shown that B cells
inhibit osteoclastogenesis by secretion of TGF-b [16] or by
regulatory B cells characterized by the expression of IL-10
[17]. Moreover, B cells in RA produce not only cytokines
but also autoantibodies like RF and ACPA. The latter
are highly specific for RA and have recently been shown
to induce osteoclast differentiation by direct binding to
osteoclast precursors [18].
We set out to investigate which immune cells and
which humoral factors impact on bone resorption in RA.
We focused on CTX-1 as a serum marker for ongoing
bone turnover as opposed to the accumulated bone erosion assessable on radiographs. CTX-1 is the carboxyterminal cross-linked telopeptide generated during the
degradation of collagen type I via cathepsin K and thus
directly mirrors osteoclast activity as well as predicts
future radiological progression [19–22]. Performing a
cross-sectional study including serum parameters and
cellular composition, we identified the CD5+ B cell population as the major factor associated with bone resorption
in RA.
Methods
Patients
Sixty-one consecutive ACPA-positive RA patients were
enrolled at the Clinics Südstadt Rostock, Germany.
Our patient cohort consisted of 72% females and had a
mean age of 63.1 years (range 30–87). The patients
showed a mean disease duration of 13.7 years (range
0.4–40.1), a mean treatment duration of 4.7 years
(range 0–25.1), a mean CRP of 14.7 mg/l (range 0–144;
normal range 0–3) and a mean 28-joint DAS (DAS28)
of 3.1 (range 0.9–7.1). Treatment regimes were heterogeneous, including DMARDs alone (n = 34), anti-B cell therapy (n = 2), anti-TNF-a therapy (n = 12), anti-IL6 receptor
(n = 7), modulation of T cell co-stimulation (n = 4) and no
therapy at all (n = 2). The majority of patients (82%)
included in the present study were treated with vitamin
D supplementation, but none with bisphosphonates or
other known bone-modifying treatments.
This study was approved by the ethics committee of
the Rostock University Medical Center (A2011-134). All
patients fulfilled the 1987 ACR classification criteria for
RA and gave informed written consent prior to blood
sampling according to the Declaration of Helsinki.
Serum samples were collected between 8:30 and
11:00 a.m. Food intake prior to serum sampling was not
specifically controlled for.
546
Flow cytometry
Peripheral blood mononuclear cells (PBMCs) were isolated from 7.5 ml of EDTA blood by Ficoll density gradient
centrifugation. First, aqua fluorescent reactive dye
(Invitrogen, Darmstadt, Germany) was used for live/dead
discrimination according to the manufacturer’s instructions. Next, cells were stained for surface markers in
ice-cold PBS pH 7.4, 0.5% BSA and 0.1% sodium
azide. Appropriate isotype controls were used to quantify
CD157, glucocorticoid-induced TNF receptor–related protein, CD80 and CD86. All cells were fixed after the surface
staining with either 4% paraformaldehyde for 10 min at
room temperature or prior intracellular staining for Foxp3
and helios using the FoxP3 Staining Buffer Set (Miltenyi,
Bergisch Gladbach, Germany) according to manufacturer’s instructions. Data were acquired on a FACS Aria
II machine (BD, Heidelberg, Germany) and analysed using
FACS Diva Software (BD). For normal control, B cell
values refer to previously published data [23, 24]. The
MiFlowCyt-conform documentation of our flow cytometry
experiments, including, for example. the gating strategies,
can be found in supplementary 1-MiFlowCyt documentation S1, available at Rheumatology Online.
ACPA ELISA
IgG1 and IgG4 antibody levels for CCP and mutated citrullinated vimentin (MCV) were determined by a direct ELISA
as previously described [25]. Briefly, we combined CCPcoated (Euroimmun, Lübeck, Germany) and MCV-coated
(Orgentec, Mainz, Germany) ELISA plates with detection
antibodies specific for IgG1 (Binding Site, Birmingham,
UK) or IgG4 (AbD Serotec, Puchheim, Germany). In the
first step, the sera were applied in dilutions of 1:400 and
1:20 for IgG1 and IgG4, respectively. Thereafter the
plates were incubated with IgG1-specific (1:15 000) or
IgG4-specific (1:25 000) horseradish peroxidase (HRP)coupled detection antibodies. Finally, colour reaction
was performed using 3,30 ,5,50 -tetramethylbenzidine
(TMB) substrate (BioLegend, Fell, Germany) and the
optical density (OD) was determined by an automated
plate reader (Millenia Kinetic Analyser; DPC, Los
Angeles, CA, USA).
IgM MCV serum levels were determined after depletion
of IgG in order to remove potential background originating
from IgG-specific IgM RF. IgG was depleted using
protein-G agarose (Sigma Aldrich, Munich, Germany).
The resin was washed five times with 1% BSA in PBS
pH 7.4 prior to incubation with 1:62.5 diluted serum in
an oscillatory shaker for 1 h. The resin was cleaned after
use by washing five times with 100 mM glycine-HCl
buffer pH 2.7 and five times with PBS pH 7.4 containing
0.05% sodium azide for storage at 4 C. Depleted samples (100 ml/well) were applied to a MCV-coated plate
(Orgentec) for 1.5 h. After a washing step, plate-bound
IgG and IgM were detected using 100 ml of HRP-coupled
anti-human IgM (1:3000; AbD Serotec) or anti-human IgG
antibodies (1:5000; Biomol, Hamburg, Germany). Finally,
the plates were developed with TMB substrate and
www.rheumatology.oxfordjournals.org
Bone resorption and CD5+ B cells
ODs were determined by an automated plate reader
(Millenia Kinetic Analyser; DPC).
Avidity index of ACPA
The avidity of ACPA was determined using an elution
ELISA system as previously described [26]. In brief, sera
were first titrated (1:25–1:3200) to determine the dilution
showing a midrange OD for the CCP2 assay in order
to allow the binding of low-avidity antibodies, if present.
To determine the avidity, diluted patient serum was
applied to a CCP2 plate and the initial OD (without elution)
was measured as well as the OD of the remaining platebound antibodies after elution with 2 200 ml glycine-HCl
pH 2.7. An HRP-coupled anti-human IgG antibody
(1:5000; Biomol) was used for detection. The plates
were developed with TMB substrate and the ODs were
determined by an automated plate reader (Millenia
Kinetic Analyser; DPC). An avidity index (AI) was calculated as described elsewhere [27].
CTX-1 and serum cytokine levels
Serum CTX-1 was determined by using a competitive
ELISA (USCN Life Science, Houston, TX, USA) and
serum TNF-a by using a sandwich ELISA (Diaclone SAS,
Besancon Cedex, France) according to the manufacturer’s instructions. The CTX-1 ELISA has an intra-assay
coefficient of variation (CV) of 19.7% (our own measurements performed in duplicate) and an inter-assay CV of
<12% (according to the manufacturer’s information), the
minimal detectable dose is 46.7 pg/ml and the detection
range is given as 0.1235–10 ng/ml (according to the
manufacturer’s information). To minimize interassay variation, we ran the CTX-1 ELISA on three consecutive days
using one plate each day.
IL-6 and IL-10 serum levels were determined using
DuoSet ELISA antibody pairs (R&D Systems,
Minneapolis, MN, USA) according to the manufacturer’s
instructions. In brief, Medisorb ELISA plates (Thermo
Scientific, Waltham, MA, USA) were coated with 2 mg/ml
capture antibody in carbonate–bicarbonate buffer pH
9.4 (Thermo Scientific) at 4 C overnight. After blocking
with 1% BSA in PBS pH 7.4 (Thermo Scientific), serial
dilution of standard protein (500 pg/ml to 4.7 pg/ml) and
undiluted sera were applied for 2 h. Biotinylated detection
antibodies were incubated at concentrations of 50
and 150 ng/ml for IL-6 and IL-10, respectively.
Streptavidin–HRP included within the kit was diluted
1:200 and incubated for 30 min. The subsequent colour
reaction was performed using TMB solution (BioLegend)
and the OD was determined by an automated plate reader
(Millenia Kinetic Analyser; DPC).
Statistics
Statistical analyses were performed in R (version 3.0.0; R
Project for Statistical Computing, Vienna, Austria). The
mean (S.D.) is given for all cell populations. Differences
between two groups were tested by Mann–Whitney
U-test and correlations were calculated by Spearman’s
rank correlation. To proceed with a hypothesis-free
www.rheumatology.oxfordjournals.org
approach, we utilized importance scores from a random
forest analysis. Prior to random forest model fitting, missing values (1.1% of all values) were imputed using the
pcaMethods package. For factor analysis we first calculated a correlation matrix using Spearman’s rank correlation. The R Sweave documentation (pdf) of our statistical
analyses can be found in supplementary 2-Rsweave statistical documentation S2, available at Rheumatology
Online. Original fcs files are available from the corresponding author upon request.
Results
The bone resorption marker CTX-1 does not correlate
with clinical parameters
To screen for clinical parameters and possible confounders impacting on bone resorption in RA patients, we performed correlation analyses between serum CTX-1 levels
and various patient data. We could thus show that serum
CTX-1 titres correlate with neither age (P = 0.95, R = 0) nor
sex (P = 0.44) and therefore rule out postmenopausal
osteoporosis in the elderly female patient as a confounder
for bone resorption in RA (Fig. 1A and B). CTX-1 levels
also did not correlate with different treatment regimens
(Fig. 1C), DAS28 (P = 0.84, R = 0.03) or CRP (P = 0.38,
R = 0.12) (Fig. 1D and 1E). These results suggest that
bone resorption in RA is not merely a function of disease
activity or inflammation.
We next investigated whether bone resorption in RA
patients correlates with serum ACPA levels, and in particular with the IgG1, IgG4 and IgM isotypes. However, we
did not find any correlation between the CTX-1 levels and
IgG1 CCP (P = 0.78, R = 0.03), IgG1 MCV (P = 0.58,
R = 0.07), IgG4 CCP (P = 0.89, R = 0.02) or IgG4 MCV
(P = 0.97, R = 0) in our cohort (Fig. 1F and G). Moreover,
we did not see any differences in IgM MCV levels comparing patients with high CTX-1 and patients with low CTX-1
levels (P = 0.87; Fig. 1H).
Adaptive immune cells in the peripheral blood of
RA patients significantly correlate with CTX-1
We continued to analyse whether peripheral blood cells
that represent the adaptive immune system were
somehow linked to bone resorption in RA patients. To
that extent we determined the percentages of Tregs
[CD25+Foxp3+helios+CD127 among live CD4+; 2.49%
(S.D. 2.64)], transitional B cells [CD24highCD38high among
live CD19+; 4.58% (S.D. 4.74)], naive B cells [CD24+CD38+
among live CD19+; 46.37% (S.D. 12.81)], memory B cells
[CD24+CD38 among live CD19+; 39.4% (S.D. 13.89)] and
CD5+ B cells [CD5+ among live CD19+; 18.1% (S.D. 11.32)]
by flow cytometry and correlated these data with the
respective CTX-1 levels in the patients’ sera (Fig. 2; for
reasons of clarity, the scaling is provide in supplementary
Fig. S2, available at Rheumatology Online). We found significant correlations between CTX-1 levels and all five
cell populations. The strongest associations were found
for the percentages of transitional B cells (P = 0.0004,
R = 0.44), memory B cells (P = 0.0008, R = 0.42) and
547
Robby Engelmann et al.
FIG. 1 Clinical parameters are not associated with serum CTX-1 level
Serum CTX-1 does not correlate with (A) the age or (B) sex of RA patients. (C) CTX-1 levels do not vary between patients
receiving different treatments. CTX-1 serum levels are independent of (D) the DAS28 and (E) CRP level. Serum ACPA
levels [(F) CCP IgG1 and (G) IgG4] do not correlate with CTX-1. (H) No differences in IgM mutated citrullinated vimentin
levels were detected when comparing patients with high and low CTX-1. Each dot represents one patient. Horizontal
lines indicate medians. Differences between groups were tested by the Mann–Whitney U-test and correlations were
calculated using Spearman’s test. DAS28: 28-joint DAS.
CD5+ B cells (P = 0.002, R = 0.41). Importantly, the cell
populations showed strong correlations between each
other. For example, the percentages of memory and
naive B cells were strongly correlated (P < 0.001,
R = 0.90), as were the percentages of CD5+ B cells and
transitional B cells (P < 0.001, R = 0.61). We therefore
performed multivariate analyses to define the relevant
predictors for CTX-1 levels.
548
CD5+ B cells are most intriguingly associated with
CTX-1 serum levels in RA patients
We performed a random forest analysis to calculate importance scores for each of the variables (i.e. cell populations). These scores reflect the gain of information each
predictor added to the modelling of CTX-1 serum levels as
the outcome. Interestingly, we found the percentages of
CD5+ B cells, the CD86 expression level on transitional B
www.rheumatology.oxfordjournals.org
Bone resorption and CD5+ B cells
FIG. 2 Serum bone resorption marker CTX-1 correlates with B cell populations in RA patients
The correlation matrix provides P- and R-values (upper right) and original data (lower left) for the correlation of CTX-1
serum levels with the percentages of regulatory T cells (CD25+Foxp3+helios+CD127 cells among live CD4+), transitional
B cells (CD24highCD38high among live CD19+), naive B cells (CD24+CD38+ among live CD19+), memory B cells
(CD24+CD38 among live CD19+) and CD5+ cells among live CD19+ B cells. Correlations were tested using Spearman’s
test. Each dot represents one patient. Dashed lines indicate correlations. *P < 0.05, **P < 0.01, ***P < 0.001. For detailed
scaling of the graphs see supplementary Fig. S2, available at Rheumatology Online.
cells and the percentages of CD5+ cells among the
transitional B cell subsets most strongly associated with
CTX-1 serum levels (Fig. 3A). Likewise, we performed
factor analyses whereby serum CTX-1 levels, the percentages of transitional B cells as well as the percentage
of CD5+ B cells aggregated into one factor with
loadings of 0.48, 0.65 and 0.96, respectively. We therefore
concluded that bone resorption in RA patients—as measured via CTX-1 serum levels—is associated with CD5+
B cells.
To narrow down the specific CD5+ B cell subset correlating with CTX-1 levels, we determined the percentages
of CD5+ cells among the transitional, naive and memory B
www.rheumatology.oxfordjournals.org
cell compartments. We thus found significant correlations
between CTX-1 levels and the percentages of CD5+ cells
among naive (P = 0.014, R = 0.34) and transitional B cell
populations (P = 0.003, R = 0.39), but not among memory
B cells (P = 0.19, R = 0.18).
Some authors consider CD5 expression a marker for B
cell activation [28]. However, our data demonstrated that
the expression of the co-stimulatory molecules CD80
(P = 0.18, median MFI CD5+ vs all B cells: 64 vs 73) and
CD86 (P = 0.22, median MFI CD5+ vs all B cells: 89 vs 84)
were not significantly elevated on CD5+ B cells compared
with all B cells. We therefore concluded that in RA patients
CD5 expression is not merely a result of B cell activation.
549
Robby Engelmann et al.
FIG. 3 Multivariate analyses show an association between CD5+ B cells and serum CTX-1 in RA patients
(A) Random forest importance scores indicate an association of CD5+ B cells with CTX-1 serum levels. Nine parameters
featuring the highest importance scores are shown. The percentage of CD5+ B cells showed the highest importance
score and was set to 100%. (B) The avidity index is similar between patients with high and low CTX-1 levels and/or the
percentage of CD5+ B cells. (C) Plots show no significant correlation between serum CTX-1 and IL-6 or IL-10 level. Each
dot represents one patient. Horizontal lines indicate medians.
The strong association with CD5+ transitional B cells led
us to investigate CD1d expression on B cell populations
as an additional marker for regulatory B cells. Our data
showed that the expression of CD1d was independent of
CD5 and comparable between naive and memory B cells.
However, CD1d expression was significantly elevated in
the transitional B cell population (P < 0.001) and did not
correlate with CTX-1 levels. In conclusion, our analyses
suggest that CD5+ but not CD1dhigh B cells are positively
associated with serum CTX-1 level in RA patients.
How could CD5+ B cells promote CTX-1 levels?
Finally, we addressed two CD5+ B cell-associated mechanisms for their potential impact on bone resorption as
measured via CTX-1 levels in the serum. First, CD5+ B1
cells undergo very limited affinity maturation and therefore
produce low-affinity IgG antibodies, among them selfreactive ones [18]. Interestingly, Suwannalai et al. [29] proposed a population of low-avidity ACPA associated with
radiographic progression in RA. These authors defined
550
an AI reflecting the ratio of ACPA bound to a CCP2coated plate after elution with a chaotropic reagent to
the amount of ACPA initially bound to the plates. We similarly determined the AI for 22 RA patients, who featured
either high or low CTX-1 levels and covered a range of
CD5+ percentages. Again, we did not find any correlation
between CTX-1 levels and ACPA avidity, suggesting that
in our cohort the avidity of self-reactive antibodies does
not contribute to CTX-1 levels (Fig. 3B).
Secondly, CD5+ B cells produce IL-6, which in turn promotes osteoclast maturation and bone resorption.
Moreover, CD5+ B cells are the main producers of B
cell-derived IL-10, which drives autoimmunity via numerous pleiotropic effects [30]. We therefore determined IL-6
and IL-10 levels in the serum of our RA patients and correlated them with both serum CTX-1 levels and percentages of CD5+ B cells. Unfortunately we did not find any
association between CTX-1 and IL-6 (P = 0.33), CTX-1 and
IL-10 (P = 0.56) (Fig. 3C) or between the percentages
of CD5+ B cells and IL-6 (P = 0.64) or IL-10 (P = 0.25).
www.rheumatology.oxfordjournals.org
Bone resorption and CD5+ B cells
Factor analysis showed that IL-6 (loading 0.3) as a major
pro-inflammatory cytokine aggregated into one factor with
CRP, DAS28 and steroid treatment dosage (loadings 0.87,
0.65 and 0.74, respectively). Moreover, IL-6 correlated
with therapy duration (P = 0.003, R = 0.38). We therefore
concluded that CD5+ B cells contribute to elevated CTX-1
serum levels in RA by an as yet unknown mechanism.
Discussion
The present study investigates reasons for bone resorption in RA patients and focuses on clinical parameters,
serum markers and immune cell populations. We aimed
to evaluate ongoing bone resorption and erosive progression instead of established erosions and we therefore
assessed CTX-1 levels in the serum [22]. Indeed, serum
CTX-1 levels have been shown to be elevated in RA
patients and to predict subsequent radiographic progression [21]. We here confirm these findings as our cohort
exhibited elevated CTX-1 levels [1.8 ng/ml (S.D. 0.81)]
compared with a healthy control group [1.13 ng/ml (S.D.
0.81)] whose serum was analysed using exactly the
same assay [31]. We refrained from determining our own
reference range in a healthy control cohort since our focus
was on correlations between bone erosion and peripheral
blood parameters among RA patients.
Bone resorption results from the uncoupling of osteoblastic bone formation and osteoclastic bone resorption
[32]. In RA, this uncoupling can occur in three ways: directly
via RANKL expressed on Th17 cells; indirectly via soluble
TNF-a and IL-6, which impact on the osteoblasts to induce
further osteoclast maturation; or via ACPA [18, 33, 34]. It is
therefore fair to assume that patients with high disease
activity will show elevated CTX-1 levels. This, however,
was not the case in our study (Fig. 1C). The more effective
therapies for the prevention of bone erosion combine
DMARDs—to contain inflammation—with biologics neutralizing TNF-a, IL-6 or B cells, which produce autoantibodies
[32]. And indeed, our patients with a high DAS28 received
anti-TNF-a, anti-IL-6R or anti-CD19 so that immune cellmediated bone resorption was possibly compensated for
by the neutralization of osteoblasts stimulating cytokines
and autoantibodies [35, 36].
Interestingly, we could not confirm dose dependency
between serum ACPA levels and bone resorption as
previously described [18]. However, we argue that by
enrolling consecutive ACPA-positive patients into a crosssectional study we accepted the influence of treatment and
disease activity on bone resorption. This is in contrast to
the previous selection of newly diagnosed and untreated
patients matched for age, sex and disease activity [18].
We found peripheral CD5+ B cells associated with bone
resorption in RA patients. The origin and the exact immunological features of CD5+ B cells are still a matter
of debate. Some authors have suggested that the expression of CD5 is merely a marker for activation [28]. Our
finding that co-stimulatory molecules like CD80 and
CD86 are expressed independently of CD5 argues
against this notion. Furthermore, CD5+ B cells were formerly thought to be a population prone to produce
www.rheumatology.oxfordjournals.org
autoantibodies and to thus promote autoimmunity. And
indeed, CD5+ B cells have been described to produce
RF [37, 38] and to be elevated in RA patients [39]. Yet it
became clear that CD5+ B cells are not the main producers of autoantibodies [40]. This is in line with our study,
as we did not see a correlation between ACPA and the
percentage of CD5+ B cells. More recently, CD5 has
been suggested to be a negative regulator of B cell
receptor activation and thus may serve to control
autoimmunity [41]. Along the same lines, a population
of regulatory B cells characterized by the marker combination CD19+CD24highCD38highCD5+CD1dhigh has been
described [42]. These regulatory B cells are capable
of producing high levels of IL-10 and suppress Th1 and
Th17 differentiation of naive Th cells dependent on CD80
and CD86 co-stimulation [43]. However, the CD5+ B cell
population in our study that correlates with bone resorption is unlikely to belong to the regulatory B cells. In our
study the regulatory B cell-associated CD1d expression is
only elevated on CD19+CD24highCD38high transitional B
cells. Most importantly, these transitional B cells showed
lower importance scores and loadings than the CD5+ B
cells in our multivariate analyses.
CD5+ B cells are known to produce IL-6 [44, 45] and
IL-6 in turn supports osteoclast differentiation [33, 34],
leading to an increase in bone resorption. We therefore
analysed IL-6 serum levels yet did not find a correlation
between serum IL-6 level and CTX-1. However, we cannot
exclude that CD5+ B cells contribute locally to elevated
IL-6 levels within the joints.
Moreover, CD5+ B cells undergo limited germinal centre
reactions and therefore lack somatic hypermutations,
which in turn results in low-affinity antibodies [46].
Suwannalai et al. [29] recently demonstrated that lowavidity ACPA is somehow associated with more severe
radiological progression in RA. We therefore determined
the amount of low-avidity ACPA, but again were unable
to find any differences comparing patients with either high
or low CTX-1 and CD5+ cells.
In summary, we identified the CD5+ B cell population
to be significantly associated with bone resorption in
a cross-sectional cohort of RA patients. However, a possible functional link between CD5+ cells and bone resorption in RA still needs to be determined.
Rheumatology key messages
The frequency of CD5+ B cells is associated with
bone resorption in RA patients.
. Neither systemic IL-6 nor low-affinity ACPA is associated with bone resorption in RA patients.
+
. A possible link between CD5 cells and bone
resorption in RA patients needs to be determined.
.
Acknowledgements
We thank Dr Carsten Wiethe for his help in setting up
the panels for the flow cytometry and Birgit Ahrens for
collecting the patient data.
551
Robby Engelmann et al.
Funding: This study was funded by the German Research
Foundation (DFG; Mu 844/13-1), an intramural grant
(FORUN), as well as an Investigator Initiated Research
Grant from Pfizer (WI170970).
Disclosure statement: B.M.-H. received grant support
from Pfizer. C.K. has provided consultancy services for
AbbVie, Chugai, Pfizer and Roche; has received research
support from Pfizer and has served on speakers’ bureaus
for AbbVie, Berlin Chemie, Chugai, MSD, Pfizer, Roche
and UCB. All other authors have declared no conflicts of
interest.
13 Manabe N, Kawaguchi H, Chikuda H et al. Connection
between B lymphocyte and osteoclast differentiation
pathways. J Immunol 2001;167:2625–31.
14 Sowden JA, Roberts-Thomson PJ, Zola H. Evaluation of
CD5-positive B cells in blood and synovial fluid of patients
with rheumatic diseases. Rheumatol Int 1987;7:255–9.
15 Sigal LH. Basic science for the clinician 54: CD5. J Clin
Rheumatol 2012;18:83–8.
16 Weitzmann MN, Cenci S, Haug J et al. B lymphocytes
inhibit human osteoclastogenesis by secretion of TGFb.
J Cell Biochem 2000;78:318–24.
17 Mauri C, Bosma A. Immune regulatory function of B cells.
Annu Rev Immunol 2012;30:221–41.
Supplementary data
Supplementary data are available at Rheumatology
Online.
References
1 Gravallese EM, Manning C, Tsay A et al. Synovial tissue
in rheumatoid arthritis is a source of osteoclast differentiation factor. Arthritis Rheum 2000;43:250–58.
2 Pettit AR, Walsh NC, Manning C, Goldring SR,
Gravallese EM. RANKL protein is expressed at the
pannus-bone interface at sites of articular bone erosion
in rheumatoid arthritis. Rheumatology 2006;45:1068–76.
3 Redlich K, Hayer S, Ricci R et al. Osteoclasts are essential
for TNF-a-mediated joint destruction. J Clin Invest 2002;
110:1419–27.
4 Jimenez-Boj E, Redlich K, Türk B et al. Interaction
between synovial inflammatory tissue and bone marrow
in rheumatoid arthritis. J Immunol 2005;175:2579–2588.
5 Müller B. Cytokine imbalance in non-immunological
chronic disease. Cytokine 2002;18:334–9.
6 Okamoto K, Takayanagi H. Regulation of bone by the
adaptive immune system in arthritis. Arthritis Res Ther
2011;13:219.
7 Kotake S, Udagawa N, Takahashi N et al. IL-17 in synovial
fluids from patients with rheumatoid arthritis is a potent
stimulator of osteoclastogenesis. J Clin Invest 1999;103:
1345–52.
8 Sato K, Suematsu A, Okamoto K et al. Th17 functions as
an osteoclastogenic helper T cell subset that links T cell
activation and bone destruction. J Exp Med 2006;203:
2673–82.
9 Lacey DL, Boyle WJ, Simonet WS et al. Bench to bedside:
elucidation of the OPG–RANK–RANKL pathway and the
development of denosumab. Nat Rev Drug Discov 2012;
11:401–19.
10 Romas E, Gillespie MT, Martin TJ. Involvement of receptor
activator of NFkappaB ligand and tumor necrosis factoralpha in bone destruction in rheumatoid arthritis. Bone
2002;30:340–6.
11 Kasagi S, Chen W. TGF-beta1 on osteoimmunology and
the bone component cells. Cell Biosci 2013;3:4.
12 Luo CY, Wang L, Sun C, Li DJ. Estrogen enhances the
functions of CD4(+)CD25(+)Foxp3(+) regulatory T cells that
suppress osteoclast differentiation and bone resorption
in vitro. Cell Mol Immunol 2011;8:50–8.
552
18 Harre U, Georgess D, Bang H et al. Induction of osteoclastogenesis and bone loss by human autoantibodies
against citrullinated vimentin. J Clin Invest 2012;122:
1791–802.
19 Garnero P, Ferreras M, Karsdal M et al. The type I collagen
fragments ICTP and CTX reveal distinct enzymatic pathways of bone collagen degradation. J Bone Miner Res
2003;18:859–67.
20 Garnero P, Geusens P, Landewé R. Biochemical markers
of joint tissue turnover in early rheumatoid arthritis.
Clin Exp Rheumatol 2003;21(Suppl 31):S54–8.
21 Syversen SW, Goll GL, van der Heijde D et al. Cartilage
and bone biomarkers in rheumatoid arthritis: prediction
of 10-year radiographic progression. J Rheumatol 2009;
36:266–72.
22 Garnero P, Thompson E, Woodworth T, Smolen JS.
Rapid and sustained improvement in bone and cartilage
turnover markers with the anti-interleukin-6 receptor
inhibitor tocilizumab plus methotrexate in rheumatoid
arthritis patients with an inadequate response to methotrexate: results from a substudy of the multicenter doubleblind, placebo-controlled trial of tocilizumab in inadequate
responders to methotrexate alone. Arthritis Rheum 2010;
62:33–43.
23 Kaminski DA, Wei C, Qian Y, Rosenberg AF, Sanz I.
Advances in human B cell phenotypic profiling. Front
Immunol 2012;3:302.
24 Moura RA, Weinmann P, Pereira PA et al. Alterations
on peripheral blood B-cell subpopulations in very early
arthritis patients. Rheumatology 2010;49:1082–92.
25 Engelmann R, Brandt J, Eggert M et al. IgG1 and IgG4
are the predominant subclasses among auto-antibodies
against two citrullinated antigens in RA. Rheumatology
2008;47:1489–92.
26 Engelmann R, Brandt J, Eggert M et al. The anti-mutated
citrullinated vimentin response classifies patients with
rheumatoid arthritis into broad and narrow responders.
J Rheumatol 2009;36:2670–4.
27 Suwannalai P, Scherer HU, van der Woude D et al.
Anti-citrullinated protein antibodies have a low avidity
compared with antibodies against recall antigens.
Ann Rheum Dis 2011;70:373–9.
28 Wortis HH, Teutsch M, Higer M, Zheng J, Parker DC.
B-cell activation by crosslinking of surface IgM or ligation
of CD40 involves alternative signal pathways and results
in different B-cell phenotypes. Proc Natl Acad Sci USA
1995;92:3348–52.
www.rheumatology.oxfordjournals.org
Bone resorption and CD5+ B cells
29 Suwannalai P, Britsemmer K, Knevel R et al. Low-avidity
anticitrullinated protein antibodies (ACPA) are associated
with a higher rate of joint destruction in rheumatoid
arthritis. Ann Rheum Dis 2014;73:270–6.
38 Mantovani L, Wilder RL, Casali P. Human rheumatoid
B-1a (CD5+ B) cells make somatically hypermutated
high affinity IgM rheumatoid factors. J Immunol 1993;151:
473–88.
30 Youinou P, Jamin C, Saraux A. B-cell: a logical target for
treatment of rheumatoid arthritis. Clin Exp Rheumatol
2007;25:318–28.
39 Youinou P, Mackenzie L, Jouquan J, Le Goff P,
Lydyard PM. CD5 positive B cells in patients with
rheumatoid arthritis: phorbol ester mediated
enhancement of detection. Ann Rheum Dis 1987;46:
17–22.
31 Fernandez-Real JM, Bullo M, Moreno-Navarrete JM et al.
A Mediterranean diet enriched with olive oil is associated
with higher serum total osteocalcin levels in elderly men at
high cardiovascular risk. J Clin Endocrinol Metab 2012;97:
3792–8.
32 Villeneuve E, Haraoui B. Uncoupling of disease activity
and structural damage. Does it matter clinically?
Ann Rheum Dis 2013;72:1–2.
33 Kotake S, Sato K, Kim KJ. Interleukin-6 and soluble
interleukin-6 receptors in the synovial fluids from
rheumatoid arthritis patients are responsible for osteoclast-like cell formation. J Bone Miner Res 1996;11:88–95.
34 Hashizume M, Mihara M. The roles of interleukin-6 in the
pathogenesis of rheumatoid arthritis. Arthritis 2011;2011.
35 Smolen JS, Han C, van der Heijde DMFM. Radiographic
changes in rheumatoid arthritis patients attaining different
disease activity states with methotrexate monotherapy
and infliximab plus methotrexate: the impacts of remission
and tumour necrosis factor blockade. Ann Rheum Dis
2009;68:823–7.
36 Landewé R, van der Heijde D, Klareskog L, van
Vollenhoven R, Fatenejad S. Disconnect between inflammation and joint destruction after treatment with etanercept plus methotrexate: results from the trial of etanercept
and methotrexate with radiographic and patient outcomes. Arthritis Rheum 2006;54:3119–25.
37 Burastero SE, Casali P, Wilder RL, Notkins AL.
Monoreactive high affinity and polyreactive low affinity
rheumatoid factors are produced by CD5+ B cells from
patients with rheumatoid arthritis. J Exp Med 1988;168:
1979–92.
www.rheumatology.oxfordjournals.org
40 Reap EA, Sobel ES, Cohen PL, Eisenberg RA.
Conventional B cells, not B-1 cells, are responsible for
producing autoantibodies in lpr mice. J Exp Med 1993;
177:69–78.
41 Hippen KL, Tze LE, Behrens TW. Cd5 maintains
tolerance in anergic B Cells. J Exp Med 2000;191:
883–90.
42 Blair PA, Noreña LY, Flores-Borja F et al.
CD19+CD24hiCD38hi B cells exhibit regulatory capacity
in healthy individuals but are functionally impaired in systemic lupus erythematosus patients. Immunity 2010;32:
129–40.
43 Lemoine S, Morva A, Youinou P, Jamin C. Human T cells
induce their own regulation through activation of B cells.
J Autoimmun 2011;36:228–38.
44 Cantaert T, Doorenspleet ME, Francosalinas G et al.
Increased numbers of CD5+ B lymphocytes with a regulatory phenotype in spondylarthritis. Arthritis Rheum 2012;
64:1859–68.
45 Spencer NF, Daynes RA. IL-12 directly stimulates
expression of IL-10 by CD5+ B cells and IL-6 by both
CD5+ and CD5- B cells: possible involvement in ageassociated cytokine dysregulation. Int Immunol 1997;9:
745–54.
46 Fischer M, Klein U, Küppers R. Molecular single-cell
analysis reveals that CD5-positive peripheral blood B
cells in healthy humans are characterized by rearranged
Vkappa genes lacking somatic mutation. J Clin Invest
1997;100:1667–76.
553