Elevated ACKR2 expression is a common feature of inflammatory arthropathies

Abstract Objectives Chemokines are essential contributors to leucocyte accumulation at sites of inflammatory pathology. Interfering with chemokine or chemokine receptor function therefore represents a plausible therapeutic option. However, our currently limited understanding of chemokine orchestration of inflammatory responses means that such therapies have not yet been fully developed. We have a particular interest in the family of atypical chemokine receptors that fine-tune, or resolve, chemokine-driven responses. In particular we are interested in atypical chemokine receptor 2 (ACKR2), which is a scavenging receptor for inflammatory CC-chemokines and that therefore helps to resolve in vivo inflammatory responses. The objective of the current study was to examine ACKR2 expression in common arthropathies. Methods ACKR2 expression was measured by a combination of qPCR and immuno-histochemistry. In addition, circulating cytokine and chemokine levels in patient plasma were assessed using multiplexing approaches. Results Expression of ACKR2 was elevated on peripheral blood cells as well as on leucocytes and stromal cells in synovial tissue. Expression on peripheral blood leucocytes correlated with, and could be regulated by, circulating cytokines with particularly strong associations being seen with IL-6 and hepatocyte growth factor. In addition, expression within the synovium was coincident with aggregates of lymphocytes, potentially atopic follicles and sites of high inflammatory chemokine expression. Similarly increased levels of ACKR2 have been reported in psoriasis and SSc. Conclusion Our data clearly show increased ACKR2 in a variety of arthropathies and taking into account our, and others’, previous data we now propose that elevated ACKR2 expression is a common feature of inflammatory pathologies.


Introduction
The in vivo migration of leukocytes is regulated by proteins belonging to the chemokine family [1]. This family is defined on the basis of the presence of a conserved cysteine motif in the mature protein sequence and is further divided into CC, CXC, XC and CX3C subfamilies according to the specific nature of the cysteine motif. Chemokines can be broadly classified as being either inflammatory or homeostatic depending on the contexts in which they function [2,3]. Inflammatory chemokines are predominantly involved in recruiting inflammatory leucocytes to damaged or infected tissue sites, whereas homeostatic chemokines are involved in more precise tissue localization of leucocytes. Chemokines interact with their target cells through receptors belonging to the seven transmembrane spanning family of G-proteincoupled receptors [4]. There are 19 identified chemokine receptors and, again, these are broadly distributed into inflammatory and homeostatic subclasses. Given the essential roles played by chemokines and their receptors in regulating immune and inflammatory leucocyte migration, it is not surprising that they play essential roles in a broad range of autoimmune and inflammatory pathologies [5,6]. Disappointingly, despite two decades of study, there are currently no antagonists of chemokine receptor function licensed for use in inflammatory diseases. This represents a significant failing in the field and is likely to be related to our currently poorly developed understanding of the full complexities of the chemokine orchestration of inflammatory responses [6].
In addition to the classical signalling chemokine receptors there exists a separate subclass of receptors characterized by an inability to mount typical signalling responses to chemokine binding [7]. These receptors also possess an altered DRYLAIV motif in the second intracellular loop and have been named atypical chemokine receptors on this basis. There are currently four members of the atypical chemokine receptor family [79]: ACKR1 (formerly known as DARC), ACKR2 (formerly known as D6 or ccbp2), ACKR3 (formerly known as RDC1 or CXCR7) and ACKR4 (formerly known as CCRL1 or CCXCKR). These molecules serve essential in vivo functions in fine-tuning, or resolving, chemokine-driven responses and represent important new contributors to the overall regulation of chemokine function.
We have a particular interest in ACKR2, which is a highly promiscuous receptor for inflammatory CC-chemokines and which is expressed by lymphatic endothelial cells, syncytiotrophoblasts and some subsets of leucocytes [10, 11]. Following ligand binding, ACKR2 internalizes ligands and targets them for intracellular degradation [12,13]. It is therefore a scavenging receptor for inflammatory CC-chemokines and its broad promiscuity ensures that it can scavenge essentially all CC-chemokines involved in inflammatory responses. Work from us, and others, has demonstrated a clear role for ACKR2 in the resolution of inflammatory responses at all tissue sites at which it is expressed [1420]. In addition we recently demonstrated a role for ACKR2 in lymphatic vessel development [21]. We have also previously shown marked up-regulation of ACKR2 expression in the epidermis and peripheral blood in psoriasis [22], and whole skin and peripheral blood in systemic sclerosis [23]. Others have further shown ACKR2 overexpression in chronic obstructive pulmonary disease [24] and post-myocardial infarction tissues [25] suggesting an, as yet poorly defined, association with inflammatory disease. The purpose of the current study was to extend these analyses to incorporate inflammatory arthropathies, where ACKR2-binding inflammatory CC-chemokines such as CC chemokine ligand (CCL) 1, 2, 3 and 5 play key roles in the propagation and maintenance of synovitis [2632].

Ethics statement
This study was approved by the West of Scotland Research Ethics Committee and the University of Glasgow Medical, Veterinary and Life Sciences Ethics Committee. Written consent was obtained from all participants according to the Declaration of Helsinki.

Isolation of peripheral blood mononuclear cells and plasma
Heparinized whole blood (10 ml) was taken from healthy donors and inflammatory arthritis patients (early RA, established RA and PsA). Matched serum samples (10 ml) were taken from each patient. Heparinized blood was diluted 1:1 with wash buffer [PBS supplemented with 1% fetal calf serum (FCS) and 2 mM ethylenediaminetetraacetic acid (EDTA)] and layered onto Histopaque (cat no. 1077-1, Sigma-Aldrich, St Louis, MO, USA) according to the manufacturer's instructions. Blood was centrifuged at 400 g for 30 min (with the brake removed) and the buffy layer was extracted. Cells were washed twice with wash buffer and counted using a haemocytometer. Cells were either resuspended in RLT buffer (cat. no. 74004, Qiagen, Hilden, Germany), containing b-mercaptoethanol, and stored at À80 C for future RNA isolation or resuspended at 1 Â 10 6 /ml in complete media (RPMI + 10% FCS + 2 mM L-glutamine + penicillin/streptomycin) for future cell culture. Plasma samples were centrifuged at 500g for 10 min at 4 C and serum was stored in 500 ml aliquots at À80 C for later analysis.
Quantitative real time RT-PCR Peripheral blood mononuclear cell (PBMC) samples stored in RLT buffer at 80 C were thawed at room temperature and RNA was extracted using the RNeasy micro kit (Qiagen cat. no. 74004) according to the manufacturer's instructions. RNA concentrations and quality were measured using the Nanodrop-2000 (Thermo Fisher Scientific, Waltham, MA, USA) and stored at À80 C for future cDNA synthesis. RNA was thawed and was reverse-transcribed to cDNA (500 ng) using Affinity Script (cat. no. 600559, Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer's instructions. cDNA was stored at À20 C until used for quantitative real-time RT-PCR (qPCR) using methodology and primers as described previously [3335]. Samples were run for 40 cycles on the 7900HT Fast Real-Time PCR System (ABI) and data were analysed using ABI Prism SDS software (Thermo Fisher Scientific). The absolute number of ACKR2 copies was normalized to TATA-binding protein copy number.

Quantification of cytokines and chemokines
Plasma cytokines and chemokines were quantified using the human cytokine 30-plex luminex kit (cat. no. LHC6003, Thermo Fisher Scientific). The serum was analysed undiluted. The detection level of the luminex was 30 pg/ml; all readings lower than this were designated non-detectable.

Stimulation of PBMCs to detect ACKR2 expression
PBMCs were cultured at 1 Â 10 6 /ml in six-well plates in complete media (RPMI + 10% FCS+ 2 mM L-glutamine + penicillin/streptomycin) at 37 C for 6, 24 or 48 h with combinations of cytokines or chemokines, all from Peprotech (Rocky Hill, NJ, USA) at 100 ng/ml. Cells were harvested on ice, and processed for qPCR as described above.

Statistical analysis
For all statistical tests, non-parametric data were analysed using the MannWhitney U test and parametric data using Student's unpaired t test. To detect significant correlation between variables, Spearman's correlation coefficient was used, where r = 1 denotes a perfect positive correlation and r = À1 a perfect negative correlation. P < 0.05 denotes significant differences.

Patient cohort
A summary of the patient details is shown in supplementary Table S1, available at Rheumatology Online. Patients classified as early RA (early inflammatory disease) were defined as having a disease course of < 1 year and had been taking DMARDs for <6 months. All patients with RA and PsA had been treated with DMARDs for the majority of the duration of their disease (over 1 year). There were no significant differences in ESR or CRP between the patient groups (supplementary Fig. S1A and B, available at Rheumatology Online).
www.rheumatology.oxfordjournals.org ACKR2 expression is elevated in PBMCs from inflammatory arthritis patients To determine the expression levels of ACKR2 in PBMCs from established PsA, RA or early RA in comparison with healthy controls, we used qPCR as previously described [33,34]. As shown in Fig. 1A all three patient populations displayed significantly elevated PBMC ACKR2 expression relative to healthy controls. Specifically we found ACKR2 expression to be increased by a median of $2-fold in established PsA (P = 0.005), 5-fold in established RA (P = 0.004) and 9-fold in early RA (P = 0.0002). There was also a significant difference in ACKR2 expression between PsA and early RA patients (P = 0.04). Some patients displayed extremely elevated levels of ACKR2 expression and, overall, there was an almost five log spread in ACKR2 levels when data from healthy controls and all three arthropathy groups were pooled (Fig. 1B). To attempt to identify the peripheral blood leucocyte subpopulations responsible for the elevated ACKR2 expression we performed immunostaining of PBMC cytospins from RA patients (Fig. 1C). Quantification of the numbers of ACKR2-positive cells per field of view revealed a highly significant increase in the percentage ACKR2-positive cells in RA patients compared with healthy controls (Fig. 1D). Co-staining for ACKR2 and CD14, CD3 or CD20 indicated that ACKR2 expression was detectable in peripheral blood monocytes, T cells and B cells (Fig. 1E). In contrast, while a few isolated cells with a neutrophilic morphology were seen to be positive for ACKR2 staining in healthy controls (supplementary Fig. S2A, available at Rheumatology Online), the majority of ACKR2-positive cells in the RA patients lacked this morphology (supplementary Fig. S2B, available at Rheumatology Online) suggesting little involvement of neutrophils in contributing to the elevated ACKR2 expression. Thus elevated PBMC ACKR2 expression is seen in a variety of leucocyte subtypes in a range of inflammatory arthropathies.

ACKR2 expression levels do not correlate with age or markers of disease severity
While there were significant differences between the ages of the healthy controls and the patient groups, and also between patient groups (supplementary Fig. S1C, available at Rheumatology Online), there was no correlation between age and ACKR2 expression in PBMCs (supplementary Fig. S3A, available at Rheumatology Online). This suggests that age was not a contributing factor to the data obtained. We next tested for correlations between ACKR2 expression and acute phase response markers (ESR and CRP). This analysis showed no significant correlation between ACKR2 and ESR or CRP (supplementary Fig. S3B and C, available at Rheumatology Online). Thus PBMC ACKR2 expression levels do not correlate with age or disease severity in arthropathies.

ACKR2 expression correlates with peripheral blood cytokine and chemokine levels
To attempt to define possible regulators of ACKR2 expression in PBMCs, we measured levels of a broad range of circulating cytokines and chemokines using Luminex and tested for correlations with PBMC ACKR2 expression levels. As shown in Fig. 2, when data from all patient and control groups were combined, statistically significant correlations with ACKR2 expression levels were seen for IL-6, IL-1b, IL-2, TNF, IL-7, IL-15, IL-5 and hepatocyte growth factor (HGF). Each correlation detected was positive in the sense that increased cytokine levels correlated with increased ACKR2 levels. By far the most significant correlations were with IL-6 and HGF concentrations. In addition, for chemokines, there was a positive correlation between CCL2 levels and ACKR2 expression (Fig. 3A) and while no overall correlation was detected with CCL3 (Fig. 3B), patients expressing ACKR2 levels above the median of the whole population display significantly higher CCL3 levels than those below the median (Fig. 3C). Interestingly, when broken down to disease subtypes, the strongest correlation between IL-6 levels and ACKR2 expression was seen for early RA and between IL-1 and IL-2 and ACKR2 for PsA.
Thus ACKR2 expression in PBMCs correlates positively with circulating levels of select cytokines and chemokines.

Combinations of cytokines and chemokines can upregulate ACKR2 expression in PBMCs
To determine whether the correlating cytokines described above were functionally involved in regulating ACKR2 expression we took PBMCs from healthy individuals, with low levels of ACKR2, and tested the ability of cytokines to increase expression levels. As shown in Fig. 2, plasma IL-6 and HGF levels strongly correlated with PBMC ACKR2 levels. However as shown in Fig. 4A, IL-6 was unable to directly modulate ACKR2 expression levels in PBMCs. In a similar manner HGF was also unable to modulate expression levels (data not shown). Thus these cytokines are unable to independently induce ACKR2 expression over a 48 h time frame. Next, to test combinatorial cytokine regulation of ACKR2 expression, we combined the cytokines that correlated with ACKR2 to produce a cytomix containing IL-6, IL-1b, IL-2, TNF-a, IL-7, IL-15, IL-5 and HGF. In addition we generated a chemomix containing the chemokines CCL2 and CCL3. Healthy PBMCs were stimulated in vitro with these agents following which we detected a modest but significant and sustained increase in ACKR2 transcript levels after stimulation with the cytomix (Fig. 4B). To determine what contribution, if any, antigen activated T cells made to this increase in expression, we evaluated the impact of CD3 activation on ACKR2 expression by T cells. As shown in supplementary Fig. S4, available at Rheumatology Online, CD3 activation initially induced a modest decrease in expression that became significant by 24 h, but which was reversed at later time points when expression in resting and activating T cells was seen to be indistinguishable. In contrast to the cytomix, while the data presented in Fig. 4C reveal some modest significant differences in response to the chemomix, these were not consistent on a temporal basis and were not seen on repeat experiments. Overall therefore, we conclude that there is no increase in ACKR2 expression in PBMCs after stimulation with the chemomix. Thus PBMC expression of ACKR2 correlates with select circulating cytokines, and mixtures of these cytokines are capable of increasing ACKR2 expression, suggesting a functional association in peripheral blood.
Expression of ACKR2 within RA synovium Next we used immunohistochemistry to examine expression of ACKR2 in RA synovium. Initial staining (Fig. 5A) revealed ACKR2-positive signals associated with cells at the tissuesynovial fluid interface as well as on numerous cells internal to the synovial tissue structure. Co-staining for CD45 and ACKR2 (Fig. 5B) demonstrated that the receptor is expressed by numerous leucocytes but also by stromal cells throughout the tissue. Interestingly, www.rheumatology.oxfordjournals.org ACKR2-positive leucocytes were generally not randomly distributed within the tissue and were present predominantly within leucocyte aggregates. Notably, co-staining for ACKR2 and one of its ligands, CCL3, indicates that there is significant overlap in their expression patterns and that ACKR2 is therefore expressed at appropriate tissue positions for effective intra-synovial chemokine scavenging (Fig. 5C). We have previously reported expression of ACKR2 by T cells, B cells, monocytic cells and mast cells [33]. In agreement with this, co-staining for CD45 and markers specific to these leucocyte populations demonstrated ACKR2 expression on these cellular populations (Fig. 6AD). Particularly strong co-staining was seen with aggregates of T and B cells (Fig. 6B and C) suggestive of ACKR2 expression within ectopic follicles. Thus ACKR2 is expressed in both leucocyte, and non-leucocyte, populations within the synovium, coincident with expression of inflammatory CC-chemokines.

Discussion
ACKR2 is an important regulator of inflammation and we have previously reported marked upregulation of expression in skin and peripheral blood leucocytes in psoriasis   FIG. 2 Correlation of ACKR2 expression in peripheral blood mononuclear cells with plasma cytokines Peripheral blood mononuclear cells (PBMCs) and paired serum samples were taken from healthy controls, PsA patients, RA patients or early RA patients. Atypical chemokine receptor 2 (ACKR2) expression was measured using qPCR with absolute quantification and normalized to 10 6 copies of TATA-binding protein (TBP). Plasma was analysed by 30-plex luminex analysis. ACKR2 expression is shown against significantly correlating cytokines: IL-6, IL-1b, IL-2, TNF-a, IL-7, IL-15, IL-5 and hepatocyte growth factor (HGF). Correlation analysis was performed using Spearman's correlation coefficient. [22] and SSc [23], and others have shown increased expression in chronic obstructive pulmonary disease [24] and post-myocardial infarction tissues [25]. The purpose of the present study was to extend these observations and to examine expression of ACKR2 in arthropathies.
Here we present data demonstrating upregulation of ACKR2 in PsA, RA and early RA. We conclude, on this basis, that elevated peripheral blood ACKR2 expression is a consistent feature of inflammatory pathologies. We further demonstrate an association between select circulating cytokines and ACKR2 expression and show that mixtures of these cytokines are capable of increasing ACKR2 expression levels in healthy PBMCs. Notably, the increase in ACKR2 expression in response to cytokine treatment was modest. The fact that the cytokines were used at a high concentration, relative to their likely in vivo concentrations, and that cytokines do not demonstrate bell-shaped doseresponse curves, suggests that it is www.rheumatology.oxfordjournals.org likely that this modest increase is the maximum that can be achieved with this cytokine mix. The relevance of this increase remains to be determined, especially given that the differences detected are at the transcriptional level and not the protein level. One further consideration is whether sufficient time was given in these experiments for effective induction of ACKR2 expression by the cytokine mix. Our previous studies have shown robust induction of ACKR2 by a variety of cytokines in a variety of cell types within a period of 2448 h [22,33,34,36]. We therefore do not anticipate that prolonged incubation with the cytokine mix beyond 48 h will lead to increased ACKR2 expression. The strongest association between circulating cytokines and ACKR2 expression was seen for IL-6 and HGF although neither cytokine alone was capable of inducing ACKR2 expression. Interestingly, while a similar association with HGF was seen in psoriatic patients, these patients displayed no association between IL-6 and ACKR2 (data not shown). This contrasts with the ACKR2-inducing effects of IL-6 seen on lymphatic endothelial cells [34]. Together our observations suggest alternative molecular drivers for enhanced ACKR2 expression in different inflammatory pathologies and cell types. Surprisingly in the arthropathy groups studied (as well as in psoriatic patients) there is no apparent correlation between peripheral blood ACKR2 expression and disease severity (supplementary Fig. S3, available at Rheumatology Online). As circulating inflammatory CCchemokines will predominantly have derived from the original inflamed site, one intriguing possibility is that ACKR2 limits the activity of circulating chemokines and therefore potentially ameliorates the development of chemokineassociated comorbidities (e.g. atherosclerosis) in patients. This possibility, however, remains to be tested.
In addition we demonstrate expression of ACKR2 in both leukocytes and resident stromal populations in patient synovial biopsies. The demonstration of expression on stromal cells, especially those at the synoviumSF www.rheumatology.oxfordjournals.org interface, extends observations made in psoriatic skin that demonstrated inducible expression on keratinocytes [22]. We hypothesize that stromal cell expression serves to reduce inflammatory chemokine bio-availability, or to limit its domain of influence, and this is supported by the coincident expression of ACKR2 and CCL3 in synovial tissues. Strikingly, many of the leukocytes that are positive for ACKR2 expression are B and T cells present in large leucocyte aggregates within the synovial tissues. We propose that these correspond to ectopic follicles although the role of ACKR2 within these structures is not immediately apparent. It may be that ACKR2 is important to ensure the relative absence of inflammatory leucocytes within these ectopic lymphoid follicles [37,38].
In summary, therefore, we provide evidence demonstrating upregulation of ACKR2 in peripheral blood cells and in cells within the synovium. This, coupled with previous observations from psoriasis and SSc patients suggests that elevated ACKR2 levels are a consistent feature of many human inflammatory pathologies. Further work is required to determine the functional implications of elevated expression in these pathological contexts.