Objectives A genetic variant of the leukocyte phosphatase PTPN22 (R620W) is strongly associated with autoimmune diseases including rheumatoid arthritis (RA). Functional studies on the variant have focussed on lymphocytes, but it is most highly expressed in neutrophils. We have investigated the effects of the variant on neutrophil function in health and in patients with RA.
Methods Healthy individuals and patients with RA were genotyped for PTPN22 (R620W) and neutrophils isolated from peripheral blood. Neutrophil adhesion and migration across inflamed endothelium were measured. Calcium (Ca2+) release and reactive oxygen species (ROS) production in response to fMLP stimulation were also assessed.
Results Expression of R620W enhanced neutrophil migration through cytokine activated endothelium (non-R620W=24%, R620W=45% migrating cells, p<0.001). Following fMLP stimulation, neutrophils that were heterozygous and homozygous for R620W released significantly more Ca2+ when compared to non-R620W neutrophils, in healthy individuals and patients with RA. fMLP stimulation, after TNF-α priming, provoked more ROS from neutrophils heterozygous for R620W in patients with RA (non-R620W vs R620W=∼1.75-fold increase) and healthy individuals (non-R620W vs R620W=fourfold increase) and this increase was statistically significant in healthy individuals (p<0.001) but not in patients with RA (p<0.25).
Conclusions Expression of PTPN22 (R620W) enhanced neutrophil effector functions in health and RA, with migration, Ca2+ release and production of ROS increased. Neutrophils are found in large numbers in the RA joint, and this hyperactivity of R620W cells may directly contribute to the joint damage, as well as to the initiation and perpetuation of the chronic immune-mediated inflammatory processes driving the disease.
- Rheumatoid Arthritis
- Autoimmune Diseases
- Gene Polymorphism
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
The aetiology of rheumatoid arthritis (RA) is complex, but a strong contribution is made by polymorphisms in a number of genes, many of which are involved in the regulation of the immune response.1 One such gene is PTPN22 in which a single nucleotide polymorphism (SNP) known as R620W, is a prominent risk factor for RA and other autoimmune diseases.2 The consequences of R620W expression have been widely studied in lymphocytes3–5 where the polymorphism alters signalling through the B and T cell receptors. Thus, R620W may contribute to the survival of a broader repertoire of lymphocytes, some of which may be autoreactive.6 ,7 The PTPN22 gene has also been suggested to control other lymphocyte functions, since in its absence in mice, regulatory T cells are increased8 and their adhesion and function are enhanced.9
The protein product of the PTPN22 gene is a protein tyrosine phosphatase (Lyp) whose role is to counter the effect of tyrosine kinases and regulate signalling pathways by removing phosphorylated tyrosine residues from proteins.10 To date, many of the studies on Lyp have been conducted in lymphocytes. However, Lyp is most highly expressed by neutrophils11 where little is currently known about its function. Altered levels of tyrosine phosphorylation have been observed in neutrophils isolated from patients with RA,12 and such cells exhibit resistance to cell death and enhanced effector functions.13 Furthermore, circulating neutrophils isolated from patients with RA have been shown to release more calcium (Ca2+) in response to stimulation when compared to individuals with non-rheumatic joint conditions.14 While the genotype of the patients in these studies was not reported, these data suggest that abnormal neutrophil signalling and function is a specific feature of RA neutrophils.
Quantitatively neutrophils are a major component of the joint infiltrate, since up to 90% of leukocytes in the RA joint are neutrophils.15 This suggests they make a significant contribution to the inflammatory processes driving joint destruction, and in some animal models of arthritis, initiation of disease is completely prevented in the absence of neutrophils.16 Many of the mechanisms used by neutrophils to combat infection also cause significant collateral tissue damage when neutrophil activation is enhanced or occurs inappropriately. For example, the activation-dependent release of proteases and reactive oxygen species (ROS) produce significant damage to joint components including lipids and proteins.17 ,18 The trigger for aberrant neutrophil activation is not clear, but results from changes in neutrophil intracellular signalling. In the context of neutrophil intracellular signalling, cytosolic Ca2+ increase and release are important mediators of other neutrophil functions, such as NADPH oxidase activation and adhesion to endothelium. Additionally, Ca2+ signalling processes are highly regulated by a number of proteins, one of which could be Lyp. Given the fact Lyp is abundantly expressed in neutrophils, we investigated the effects of the genetic Lyp variant R620W on neutrophil function, to determine whether this genetic variant contributes to the altered activation and function of neutrophils in RA.
Materials and methods
All reagents were purchased from Sigma Aldrich (UK), unless otherwise stated.
Patients with RA (508) were recruited from outpatient clinics at University Hospital Birmingham NHS Foundation Trust, and Russell's Hall Hospital (Dudley), alongside healthy volunteers (107). All samples were obtained with informed consent and with approval from the Edgbaston Research Ethics Committee (12/WM/0077). All patients with RA satisfied the 1987 American College of Rheumatology (ACR) criteria for RA.
Peripheral blood (PB) was collected in vacutainers containing EDTA. DNA was extracted from whole blood using the QuickGene 810 system (Fujifilm, Japan), as described previously.19 A Roche LightCycler 480 II system (Roche Diagnostics, UK) identified the SNP in PTPN22 (rs2476601) by real-time PCR and melting curve analysis using primers and probes used previously.20
Isolation of neutrophils
PB was obtained in vacutainers containing EDTA; 2% dextran was added (1 mL/6 mL of blood) and the blood left to settle for 30 min to separate red blood cells from the buffy coat. Percoll gradients (GE Healthcare, UK) were set up by adding 5 mL of 56% Percoll to a 15 mL tube, and a Pasteur pipette was used to add 2.5 mL of 80% Percoll underneath. The buffy coat was added to Percoll gradients, and centrifuged at 190×gav for 25 min with no brake. The neutrophil layer was removed from the gradient using a Pasteur pipette. Cells were washed twice and resuspended at 1.0×106/mL in prewarmed RPMI 1640 medium supplemented with 10% heat-inactivated foetal calf serum (HIFCS), and 1% glutamine penicillin streptomycin (GPS) (complete culture medium), before use in experiments. The average purity and viability of neutrophils isolated using this method was 96% and 95%, respectively.
Adhesion and migration
Adhesion and transmigration were assessed by direct microscopic observation as previously described.21 ,22 Briefly, human umbilical endothelial cells (HUVEC) were isolated from umbilical cords and cultured. Tumour necrosis factor-α (TNF-α; 0.05 and 5 ng/mL; R&D Systems, Abingdon, UK) was added to confluent monolayers for 4 h before the adhesion assay. After washing the HUVEC, neutrophils (1×106/mL in M199BSA) were added for 6 min and then non-adherent neutrophils were removed. Digitised images of the endothelial surface were made using phase-contrast microscopy at 37°C at 2 min and 9 min from the end of the settling period and analysed offline using Image-Pro Plus software (DataCell, Finchampstead, UK) as previously described.21 Adhesion was determined as the percentage of neutrophils that adhered by counting all adherent cells per image, then calculating the number of cells per well and dividing this number by the total number of neutrophils added. Transmigration was calculated as a percentage of the adherent cells that appeared phase dark (ie, underneath the endothelial monolayer). Migration velocities (µm/minute) of phase-dark neutrophils underneath the HUVEC were determined the average distance moved by the centre of the cell per minute.
Neutrophils (2.0×106/mL in RPMI 1640) were incubated with 1 µM Indo-1 AM ester (Invitrogen, USA) for 40 min at 37°C, before washing twice in Ca2+ containing Hank's Buffered Salt Solution (HBSS) at 400×gav for 6 min. Cells (1.5×106) were placed at 37°C for 10 min prior to stimulation with 50 µM fMLP. Samples were placed in a LS50B spectrofluorimeter (Perkin Elmer, UK) and maintained in suspension at 37°C using a magnetic stirrer. Data were collected using FL Winlab V.2.1 software with a dual-emission experimental protocol for Indo-1 AM loaded cells.
Neutrophils (1.0×106/mL in RPMI 1640) were primed with 5 ng/mL TNF-α (R&D Systems, Abingdon, UK) for 15 min at 37°C and then treated with 10 µM dihydrorhodamine (DHR) for 5 min at 37°C. Samples were stimulated with 1 µM fMLP for 5 min at 37°C, washed in ice-cold phosphate buffered saline (PBS) with centrifugation at 400×gav for 6 min and fixed with 2% paraformaldehyde in PBS overnight at 4°C. Fixed cells were washed as before, resuspended in PBS, and ROS production determined by DHR fluorescence detected using a CyAn ADP flow cytometer (Beckman Coulter, UK).
Data shown are mean±SEM. Mann–Whitney U was used for comparison of groups for continuous variables, and Fisher's exact test with a two-tailed p value for the comparison of categorical variables. For paired data, Wilcoxon matched-pairs signed rank tests were performed. For data comparing more than three groups, analysis was carried out using a two-way Analysis of Variance (ANOVA) followed by Bonferonni post-test. p Values of <0.05 were considered to be statistically significant.
To allow us to test the hypothesis that expression of PTPN22 R620W alters neutrophil function, a large cohort of healthy individuals and patients with RA were recruited and genotyped for the R620W SNP. The R620W variant was found to be highly expressed in both the populations assessed (table 1). The frequency of heterozygotes (AG) among the RA population was slightly decreased when compared to previous work20 (23% vs 30%); however, the frequency of homozygotes (AA) was the same (2%). Given the paucity of homozygotes, we felt it was most appropriate to focus functional investigations on the cells heterozygous (AG) for the variant allele. To do this, neutrophils were isolated from healthy individuals (n=16) and patients with RA (n=24) closely matched for age and sex. Importantly, there were no significant differences in patient demographics, such as disease duration, drug treatments or levels of inflammation, when groups were compared according to PTPN22 genotype (table 2).
PTPN22 R620W increases neutrophil migration across inflamed endothelium
Initially, we were interested to determine whether R620W altered the ability of neutrophils to be recruited to sites of inflammation. To address this, we observed the behaviour (adhesion and migration) of neutrophils isolated from healthy individuals control (GG) and heterozygous (AG) for R620W on endothelial cells (EC) stimulated with tumour necrosis factor (TNF)-α, as previously described.22 Neutrophil adhesion increased in a TNF-α dose dependent manner, which was not affected by the PTPN22 genotype (figure 1A). We observed low levels of migration on resting EC and those stimulated with low-dose (0.05 ng/mL) TNF-α for GG and AG-expressing neutrophils (figure 1B). Significantly more AG neutrophils transmigrated through EC treated with high-dose TNF-α (5 ng/mL) at 2 min when compared to neutrophils isolated from GG individuals (GG=24±4% and AG=44±7% neutrophils migrated, defined as phase-dark cells; p<0.001, figure 1B–D). These differences were lost at 9 min, with similar levels of migration observed at this later time point (GG=38±5% and AG=43±9% neutrophils migrated, defined as phase-dark cells). These data suggest that AG neutrophils are able to migrate through inflamed endothelium more rapidly when compared to their GG counterparts. However, once the cells were underneath the endothelium, neutrophils migrated at similar velocities regardless of PTPN22 genotype (GG=8.1±1.1 μm/min and AG=7.9±0.8 μm/min).
PTPN22 R620W increases Ca2+ release by neutrophils in response to fMLP stimulation
To assess the effect of R620W on overall signalling through surface receptors, N-formyl-Met-Leu-Phe (fMLP) was added to mimic the presence of formylated peptides derived from bacteria which would be present during an infection. Formyl peptide receptor (FPR) engagement releases Ca2+,23 and thus, this was measured to give an indication of the overall signal induced by FPR activation. Neutrophils isolated from healthy individuals heterozygous (AG) and homozygous (AA) for R620W, released significantly more Ca2+ in response to fMLP when compared to age and sex-matched individuals without R620W (GG=0.24±0.02 Indo-1 ratio (I1R), AG=0.28±0.02 I1R and AA=0.43 I1R, p<0.05, figure 2A+C). Similar elevated Ca2+ release were also observed in patients with RA with the R620W variant (figure 2B+C). In fact, Ca2+ release was significantly augmented in neutrophils from patients with RA when compared with healthy individuals, in a genotype-dependent manner: GG<AG<AA (GG=0.39±0.04 I1R, AG=0.44±0.02 I1R and AA=0.56 I1R, figure 2B+C).
PTPN22 R620W increases ROS production by neutrophils following priming with TNF-α
Given that R620W was shown to increase Ca2+ release, we examined whether other Ca2+ dependent functions would also be affected. In order to investigate this, the ability of neutrophils to produce ROS in response to TNF-α priming and fMLP stimulation were tested. Unstimulated neutrophils from healthy subjects with GG or AG R620W genotype produced very little ROS at the basal level (GG=69±13, AG=155±66, n=8, figure 3A+B). Although ROS production was increased with priming, no differences were observed between the GG and AG genotypes (TNF-α: GG=110±31 vs AG=274±131, or fMLP: GG=86±18, AG=200±95, n=8). In response to priming with TNF-α and then fMLP, AG neutrophils produced fourfold more ROS when compared with age, gender-matched GG neutrophils (figure 3C). Neutrophils from patients with RA produced little ROS at baseline (GG=173±54, AG=87±28, n=4, figure 4). Similar to the healthy individuals, ROS production was significantly elevated following treatment with TNF-α and then fMLP, with AG neutrophils producing ∼1.75-fold more than GG neutrophils (figure 4).
The importance of neutrophils in the pathogenesis of RA is becoming increasingly recognised,13 and in some animal models of arthritis their absence completely prevents the initiation of disease.16 Increased oxidised lipids and proteins are present in RA joints, which may result from enhanced neutrophil activity.17 ,18 In particular, release of proteases and ROS have been implicated in joint damage via direct oxidation of joint components or oxidation of immune cell surface proteins, leading to impaired suppressor cell activity and defective immunoregulation.24 In this study, we investigated a genetic variant (R620W) associated with increased risk of RA, to determine if this contributed to altered neutrophil function. We have found that expression of R620W enhances the function of neutrophils from healthy donors, as characterised by increased transendothelial migration, increased Ca2+ release, and increased production of ROS. The results with cells for patients with RA were less clear since, although there was strong trend for enhanced responses in R620W neutrophils, this did not reach statistical significance. The greater variability in RA patient neutrophil responses may be a consequence of the complex nature and heterogeneity of the disease, and such factors as age and smoking. Given that R620W is only 1 of the 101 identified RA risk loci,25 patients are likely to express a number of these which may also influence the cellular response. In our RA cohort, there were some smokers which leads to depletion of the antioxidant glutathione,26 and we have shown that the PTP activity of Lyp is susceptible to oxidation.27 We have also shown that, in CD4T cells from the elderly, the PTP CD45 is oxidatively inactivated,28 and so it may well be that the oxidation of neutrophil Lyp as a result of smoking or ageing may have confounded our results with the RA neutrophils. However, perhaps the greatest potential source of variability in the patients would be the drugs they were taking. Neutrophil function is particularly influenced by a broad range of therapeutics including glucocorticoids,29 non-steroidals30 and disease modifying drugs,31 and so eliminating all drug effects in these cells from these patients with established disease would be ethically impossible.
Nevertheless, the results overall indicate that Lyp plays an important role in the activation and responses of neutrophils. The role of Lyp in T cell receptor (TCR) signalling is better characterised, where it acts as a negative regulator mainly via dephosphorylating activating tyrosine residues of Src family kinases.10 Neutrophils express a number of PTPs, including CD45, CD148 and Lyp.11 ,32 CD45 and CD148 have been reported to regulate neutrophil migration,33 whereas, a role for Lyp has not yet been identified. T cells expressing R620W release less Ca2+ when activated by CD3 stimulation,5 suggesting R620W results in hyporesponsive T cells. We have found the opposite in neutrophils, since R620W cells were hyper-responsive to stimulation (figure 2). These contrasting observations suggest the effects of R620W differ depending upon the leukocyte subset, and highlights the importance of studying innate and adaptive immune cells in parallel.
The signalling pathways mediating TNF-α priming of neutrophils are not well known, however, it could be that Lyp is a regulator of these events. TNF-α has been shown to increase intracellular Ca2+ in neutrophils, which is associated with increased co-localisation of gp91PHOX, a component of the NADPH oxidase enzyme necessary for ROS production.34 It has been proposed that TNF-α moves the cytosolic sub units (p40PHOX p47PHOX, and p67PHOX) of the NADPH oxidase closer to the cell membrane prior to stimulation.35 High levels of TNF-α are present in RA joints, and neutrophils from these sites display increased p47PHOX phosphorylation and ROS production,36 highlighting increased priming and activity of neutrophils. There are a number of phosphorylation sites present in the structure of these subunits, which regulate their location and activation status.37–39 To date, only serine and threonine phosphorylation sites have been documented in neutrophils, but a study in ECs has shown that p47PHOX can be subject to tyrosine phosphorylation.39 Therefore, it is possible that Lyp could dephosphorylate p47PHOX at this site, and a change in Lyp function could alter activity of the NADPH oxidase.
Ca2+ signalling is regulated by a number of proteins, including Vav, a guanine exchange factor (GEF) which acts as an adaptor molecule.40 Substrate trapping experiments using T cells found that Vav is a substrate of Lyp,41 which may also be the case in neutrophils. Vav is also a mediator of β2 integrin-dependent functions including migration,42 suggesting changes to Vav phosphorylation could mediate changes in neutrophil function induced by R620W. Another potential target for Lyp is phospholipase C (PLC) γ2, important for Fc-receptor and integrin-mediated neutrophil functions.43 PLC γ2 is activated by phosphorylation downstream of Src-family kinases43 and Lyp could be involved in dephosphorylation of PLC γ2 to reduce neutrophil activation. Lyp is known to target Src-family kinases upstream of PLC γ1 in T cells,10 which may also be similar in neutrophils. Signalling via PLC γ2 in neutrophils could be of importance in RA, as mice lacking PLC γ2 are protected from the development of arthritis.43
Our results show that the presence of R620W has no effect on neutrophil adhesion, but results in more rapid transendothelial migration (figure 1), which is dependent upon ligation of ICAM-1 by β2 integrins, and triggered by an increase in intracellular Ca2+.44 These events mediate activation of RHO GTPase and p38 MAPK, which contract the EC, allowing neutrophils to migrate.45 Lyp could regulate these signalling proteins similarly to CD45 and CD148, which have been shown to regulate PI3 K, pERK and Src-family kinases needed for migratory responses.33
In conclusion, our studies suggest that neutrophils expressing R620W are more intrinsically active and more susceptible to priming by TNF-α. Given the success of RA treatments which target TNF-α, our work suggests an additional mechanism by which anti-TNF treatment is effective is by reducing neutrophil priming and activation in patients with this genetic variant. This suggests that stratification of patients by PTPN22 genotype could be beneficial, and may help to choose the most appropriate treatment for patients positive for this genetic mutation. In particular, this work identifies a genetic mechanism by which overactivation of neutrophils could contribute to increased ROS release and subsequent damage in RA joints. Last, this study highlights that the effects of R620W are not restricted to lymphocytes, where previous studies have mainly been focussed.
For patient recruitment, we acknowledge the tremendous contribution from Dr Paresh Jobanputra, Sue Brailsford and Jacqueline Cobb from the Queen Elizabeth Hospital, University Hospitals Birmingham NHS Foundation Trust, and Chitra Ramful and Lucy Kadiki from Russell's Hall Hospital, Dudley Group of Hospitals NHS Foundation Trust.
Handling editor Tore K Kvien
Contributors RB, KAK and JPS genotyped patients and healthy individuals. RB, KAK and HMM performed the experiments. GDK and CDB recruited and diagnosed patients and acquired the clinical data. SPY conceived the project and acquired the funding. SPY and RB wrote the first draft of the paper and all authors contributed to analysis of the data and drafting of the manuscript.
Funding This work was supported by Arthritis Research UK PhD studentship Grant to SPY and CDB (19325) and Career Development Fellowship to HMM (19899).
Competing interests None.
Patient consent Obtained.
Ethics approval Edgbaston NHS Research Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.