Clinical and Experimental Immunology

OR I G INA L A RTI CLE

doi:10.1111/cei.12808

Infliximab therapy balances regulatory T cells, tumour necrosis factor receptor 2 (TNFR2) expression and soluble TNFR2 in sarcoidosis

A. Verwoerd,*1 D. Hijdra,*†1 A. D. M. Vorselaars,* H. A. Crommelin,*‡ C. H. M. van Moorsel,*§ J. C. Grutters*§ and A. M. E. Claessen† *Department of Pulmonology, Interstitial Lung Diseases Centre of Excellence, †Department of Medical Microbiology and Immunology, Department of Clinical Pharmacy, St Antonius

‡

Hospital, Nieuwegein, and §Division of Heart and Lungs, University Medical Centre Utrecht, Utrecht, the Netherlands

Accepted for publication 1 May 2016 Correspondence: A. D. M. Vorselaars, Department of Pulmonology, Interstitial Lung Diseases Centre of Excellence, St Antonius Hospital, Koekoekslaan 1, 3435 CM Nieuwegein, the Netherlands. E-mail: [email protected] 1

A. V. and D. H. contributed equally to this

study.

Summary Sarcoidosis is a systemic granulomatous disease of unknown aetiology that most commonly affects the lungs. Although elevated levels of regulatory T cells (Tregs) have been reported, the extent to which they play a role in sarcoidosis pathogenesis remains unclear. Tumour necrosis factor (TNF) is thought to be one of the driving forces behind granuloma formation, illustrated by the efficacy of infliximab in severe sarcoidosis. Tregs express TNF receptor 2 (TNFR2) highly. Here, we examined the influence of infliximab therapy on Tregs and (soluble) TNFR2 levels in sarcoidosis, and correlated these with response to therapy. We observed that relative frequencies of Tregs were significantly higher in patients (n 5 54) compared to healthy controls (n 5 26; median 6
73 versus 4
36%; P < 0
001) and decreased following therapy (4
95; P < 0
001). Baseline TNFR2 expression on Tregs was increased significantly in patients versus controls (99
4 versus 96
2%; P 5 0
031), and also in responders to therapy versus non-responders (99
6 versus 97
3%; P 5 0
012). Furthermore, baseline soluble TNFR2 (sTNFR2) was higher in responders than in non-responders (mean 174 versus 107 pg/ml; P 5 0
015). After treatment, responders showed a significant reduction in sTNFR2 levels in peripheral blood (244
7 pg/ml; P < 0
001), in contrast to non-responders (13
59 pg/ml). Our results demonstrated that Treg frequencies and TNFR2 expression on Tregs are increased in sarcoidosis, followed by a decline during infliximab therapy, suggesting a pathophysiological role of this T cell subset. Interestingly, sTNFR2 levels at baseline differed significantly between responders and non-responders, making it a potential marker in predicting which patients might benefit from infliximab. Keywords: infliximab, sarcoidosis, sTNFR2, TNFR2, Tregs

Introduction Sarcoidosis is a systemic granulomatous disease that can affect multiple organs, most commonly the lungs. In most patients it follows a self-limiting course, but a chronic course is observed in a subset of patients [1]. The latter group is eligible for third-line therapy with the anti-tumour necrosis factor (TNF) agent infliximab (Remicade, Centocor, Inc., Malvem, PA, USA) when first- and second-line therapy have failed or when contraindications are present [2]. Although the pathophysiology of the disease remains elusive, evidence suggests that the formation and maintenance of granulomas is essentially helper T cell-mediated [3]. More recently, regulatory T cells (Tregs) have received increasing

attention due to their role in dampening the release of proinflammatory cytokines from effector T cells (Teffs) and thereby their potential function in controlling and ending immune responses [4]. Paradoxically, in sarcoidosis, a relative increase in Tregs compared to the healthy population is described. In cross-over experiments, however, it was shown subsequently that these Tregs are incompetent and cannot inhibit the secretion of interferon (IFN)-g and TNF by Teffs [5,6]. Furthermore, it was shown that sarcoidosis Tregs showed impaired survival compared to healthy controls [7]. Tregs are well-studied in other immune-mediated diseases, such as Crohn’s disease and rheumatoid arthritis (RA). Although reports on Treg frequencies in the peripheral blood are contradictory in both diseases (reviewed by

C 2016 British Society for Immunology, Clinical and Experimental Immunology, 185: 263–270 V

263

A. Verwoerd et al.

Grant et al. [4]), reduced regulatory function of Tregs was also observed in RA [8]. Treatment with anti-TNF agents restored the suppressive capacity of Tregs and gave rise to a newly differentiated Treg population in RA patients [9]. Although it is known that TNF exerts its effect on Tregs by binding to TNF-receptor 2 (TNFR2), which is expressed highly on Tregs, it remains unclear whether TNF–TNFR2 interaction potentiates or down-modulates Treg function [10–14]. TNFR2 also exists in a soluble form in serum known as sTNFR2, which is produced through the proteolytical cleavage of the extracellular portions of membrane-bound TNFR2 by the TNF-a-converting enzyme (TACE). Next to shedding, sTNFR2 can also be expressed through alternative splicing of the receptor transcripts [15]. It is thought that sTNFR2 is an important homeostatic regulator of TNF activity by preventing its binding to membrane-bound TNFR2, and thereby its downstream inflammatory effects [16,17]. It was already shown that serum sTNFR2 is correlated strongly with disease activity and severity, and that patients with high sTNFR2 levels maintained a prolonged therapeutic response to anti-TNF therapy in RA [18]. The extent to which Tregs play a role in sarcoidosis pathogenesis remains unknown. Furthermore, the influence of infliximab on circulating Treg frequencies and the consequences of TNF depletion on TNFR2 expression levels has not been studied previously in sarcoidosis. Insight into these factors may provide clues as to what drives granuloma formation in sarcoidosis and how infliximab, as an anti-TNF agent, interferes in this process. Also the value of sTNFR2 as a biomarker and its relation to response to infliximab have not yet been studied in sarcoidosis. The aim of this study is therefore (1) to examine relative frequencies of Tregs and expression patterns of sTNFR2 in peripheral blood of sarcoidosis patients compared to healthy controls, (2) to evaluate the influence of infliximab on these factors and (3) to investigate possible variation in sTNFR2 and Treg presence in responders versus nonresponders to infliximab therapy.

Materials and methods Study population Fifty-six patients (20 female, 36 male; mean age 49 years, range 5 29–68 years) with proven sarcoidosis were included in the present study. This study is part of the prospective, open-label cohort study by Vorselaars et al. evaluating the efficacy of infliximab in sarcoidosis [2]. Patients received infliximab intravenously following a standardized protocol starting with 5 mg/kg at weeks 0 and 2 and then every 4 weeks for a period of 26 weeks. For a more detailed description of the cohort, we refer to our previous publication [2]. Twenty-six healthy volunteers (13 female, 13 male, mean age 45 years, range 5 22–64 years) served as controls. 264

The study was approved by the local institutional review board of St Antonius Hospital Nieuwegein, with registration number LTME/R-10.13A and acronym INFLIXIMAB. Informed consent was obtained from all participants.

Response Response to infliximab was defined as a 40% reduction in biomarker levels [sIL-2R or angiotensin-converting enzyme (ACE)] or maximum standardized uptake value (SUVmax) on 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET) at week 26 compared to baseline. Thirtyseven patients were classified as responders and 10 as nonresponders. Nine patients did not complete the study but are included in the baseline analysis.

Blood samples Blood samples were collected at weeks 0, 14 and 26, just before infliximab infusion (5 mg/kg) in sodium heparin and serum tubes. Peripheral blood cells were stained and measured on the day of collection. Serum was aliquoted after centrifugation and stored in 2808C within 2 h after withdrawal. Sera were batch-analysed after completion of the patient study. Total and differential cell counts were determined on the day before infusion. T cell subpopulations were calculated from lymphocyte cell counts.

Antibodies and flow cytometry Anti-CD3 fluorescein isothiocyanate (FITC), anti-CD4 peridinin chlorophyll (PerCP)-eFluor710, anti-CD8 allophycocyanin (APC), anti-CD25 phycoerythrin (PE) and mouse immunoglobulin (Ig)G1 PE isotype control were obtained from eBioscience (San Diego, CA, USA). AntiCD120b APC (TNFR2) was obtained from R&D Systems (Minneapolis, MN, USA). After staining, flow cytometry data were acquired on a fluorescence activated cell sorter (FACS)Calibur (BD Biosciences, San Diego, CA, USA) and analysed using FlowJo (Tree Star, Ashland, OR, USA). Lymphocytes were gated based on forward-scatter (FSC) 3 side-scatter (SSC) after a first gate of peripheral blood cells without selecting any debris. Subsequently, T helper cells were gated based on a combination of CD3 and CD4/CD8; no double-positive T cells were included. CD41 T cells were subdivided into three populations based on CD25 expression pattern (high, positive and negative; Fig. 1a); we used an isotype control as control for the CD25 staining and a fluorescence minus one (FMO) staining as control for the TNF receptor staining (Fig. 2a).

Serum markers and immunoassays Serum levels of sTNFR2 were determined by multiplex bead-based immunoassays using commercially available kits (Procartaplex Human TNFR2 Simplex; eBioscience), according to the manufacturer’s instructions. Serum samples were measured undiluted on a Bio-Plex 200 System

C 2016 British Society for Immunology, Clinical and Experimental Immunology, 185: 263–270 V

Tregs and sTNFR2 in sarcoidosis

Fig. 1. CD41CD25hi regulatory T cells (Tregs) in sarcoidosis. (a). CD41CD25hi expression of lymphocytes in human peripheral blood cells of one healthy control and one sarcoidosis patient. CD41 T cells were subdivided into three populations based on CD25 expression pattern (high, positive and negative). (b). The frequency of CD41CD25hi Tregs of sarcoidosis patients (n 5 54) at baseline was significantly higher than the frequency of healthy controls (n 5 26) and of patients after 26 weeks of therapy (both P < 0
001). There was no significant difference between responders (n 5 36) and non-responders (n 5 10).

and analysed with Bio-Plex Manager software version 5
0 (Bio-Rad, Veenendaal, the Netherlands). The lower detectable range was 0
75 pg/ml.

Statistical analysis Data are expressed as median [interquartile range (IQR)] unless stated otherwise. Statistical analyses were performed using SPSS for Windows (version 22
0; IBM, Armonk, NY, USA). The Mann–Whitney U-test, Wilcoxon’s signed-rank test and Student’s t-tests were used to evaluate the data.

Results CD41CD25hi Tregs are elevated in sarcoidosis and decrease following infliximab therapy The percentage of Tregs (CD25hi) within the T helper cell population (CD31CD41CD8–) in peripheral blood was determined in sarcoidosis patients at baseline and in controls. In patients, the relative frequency of Tregs was increased significantly [6
73% (4
74–9
38)] compared to controls [4
36% (3
80–5
55); P < 0
001] (Fig. 1a). This was not significantly different between responders [6
86% (5
56–10
90)] and non-responders [6
52% (3
95–8
73)] to

Fig. 2. Absolute numbers of CD41 T cells and CD41CD25hi regulatory T cells (Tregs). (a). The absolute number of Tregs remained stable during infliximab therapy, while the absolute numbers of total CD41 T cells increased (b). The absolute cell numbers divided into the three different CD41 T cell subsets for responders (n 5 37) and non-responders (n 5 10). C 2016 British Society for Immunology, Clinical and Experimental Immunology, 185: 263–270 V

265

A. Verwoerd et al.

Fig. 3. Tumour necrosis factor receptor 2 (TNFR2) expression on the three CD25 subsets of CD41 T cells in human peripheral blood. (a) TNFR2 expression in dot-plots from one sarcoidosis patient representative of all sarcoidosis patients. (b–d). The percentage of TNFR2 expression on CD25hi (b), CD251 (c) and CD25– (d). Relative frequencies of TNFR21 CD25hi regulatory T cells (Tregs) were significantly higher in patients (n 5 22) than in controls (n 5 5; P 5 0
031). The percentage of TNFR21CD25hi Tregs and TNFR21CD251CD41 T cells was significantly higher in responders (n 5 12) than in non-responders (n 5 6; both P 5 0
012).

infliximab therapy (P 5 0
338). Baseline data of healthy controls versus sarcoidosis patients and responders and non-responders to infliximab therapy is summarized in Table 1. After seven infusions in 26 weeks of infliximab therapy, the Treg frequency decreased significantly to 4
95% (3
44–6
20); P < 0
001, such that it resembled the level of healthy controls. The reduction in relative frequencies of Tregs during infliximab therapy was significant in both responders and non-responders (see Fig. 1b), but there was no significant difference between these groups. Regarding the absolute numbers of Tregs, the levels remained roughly stable during the course of infliximab therapy (0
022 3 109/l versus 0
025 3 109/l; P 5 0
108), although absolute numbers of total CD41 T cells increased (0
287 3 109/l versus 0
526 3 109/l; P < 0
001) (Fig. 2a). In Fig. 2b, we show these results separately for responders and non-responders, divided into the three different CD41 T cell subsets.

Higher percentages of TNFR2-positive Tregs in responders to infliximab TNFR2 expression was determined on Tregs (CD41CD25hi) and effector T cells (Teffs; CD41CD251 and CD41CD25–) (Fig. 3a). We found a significant difference in the percentage of TNFR21 Tregs between patients and controls at baseline [99
40% (97
65–99
6) versus 96
20% (93
55–97
55); P 5 0
031] (Fig. 3b). Following infliximab, the percentage of TNFR21 Tregs in patients decreased, but this was not sig266

nificant [99
40% (7
65–99
65) versus 97
40% (96
80– 98
85); P 5 0
079]. When dividing the group of sarcoidosis patients into responders and non-responders to infliximab, we found that responders showed a higher baseline percentage of TNFR2-expressing Tregs than non-responders [99
6% (98
90–100) versus 97
3% (93
90–98
43); P 5 0
012] (Fig. 3b). The relative frequency of cells expressing TNFR2 was lower in both CD251 (activated) and CD25– (non-activated) Teffs compared with Tregs, which is consistent with the literature. Also in Teffs, the relative frequency of cells expressing TNFR2 appears higher in patients than in controls, although these differences are non-significant (Fig. 3c,d). The same holds for responders versus nonresponders, where responders showed higher frequencies of TNFR21 cells than non-responders (see Table 1). After 26 weeks of infliximab, the frequency of TNFR2-expressing CD251 and CD25– cells decreased. This was significant for CD25– cells [58
95% (45
18–75
38) versus 49
0% (41
45–62
10); P 5 0
039] (Fig. 3d), but not for CD251 cells [82
35% (70
15–88
53) versus 78
10% (71
90–81
75); P 5 0
098].

Serum soluble TNFR2 at baseline might predict response to infliximab Next, the level of sTNFR2 was determined in serum. Mean sTNFR2 levels were not significantly different between

C 2016 British Society for Immunology, Clinical and Experimental Immunology, 185: 263–270 V

Tregs and sTNFR2 in sarcoidosis Table 1. Baseline values of Tregs, TNFR2 surface expression and serum sTNFR2 Sarcoidosis patients versus controls n

%

Sarc HC CD25 CD25hi (Treg) TNFR2 CD25hi (Treg) CD251 CD25– sTNFR2a

Responders versus non-responders P-value

Sarc

HC

6
73 (4
74–9
38)

4
36 (3
80–5
55)

54

26

22 22 22 53

5 99
40 (97
65–99
65) 96
20 (93
55–97
55) 5 82
35 (70
15–88
53) 69
70 (67
55–74
45) 5 58
95 (45
18–75
38) 42
90 (33
65–52
55) 14 162
41 6 78
47 129
9 6 49
43

n

%

Resp Non-resp < 0
001*

36

10

0
031* 0
070 0
081 0
147

12 12 12 37

6 6 6 10

Resp 6
86 (5
56–10
90)

P-value Non-resp 6
52 (3
95–8
73)

0
338

99
60 (98
90–100) 97
30 (93
90–98
43) 0
012* 86
30 (81
90–90
60) 74
10 (65
00–81
05) 0
012* 64
90 (48
50–75
60) 51
40 (37
28–71
68) 0
269 173
90 6 71
90 107
10 6 81
38 0
015*

Data are expressed as median (IQR); serum sTNFR2a data are expressed as mean 6 standard deviation. HC 5 healthy control; resp 5 responders; non-resp 5 non-responders; Sarc 5 sarcoidosis patients; Treg 5 regulatory T cell; TNFR2 5 tumour necrosis factor receptor 2; sTNFR2 5 soluble TNFR2.

patients and controls at baseline (162
51 6 78
47 pg/ml and 129
9 6 49
43 pg/ml, respectively, P 5 0
147) (Table 1). However, a large difference in mean baseline serum sTNFR2 levels was observed between responders (173
9 6 71
90 pg/ml) and non-responders (107
1 6 81
38 pg/ml, P 5 0
015), where levels of responders were significantly higher than levels of healthy controls (P 5 0
041) (Fig. 4a). Interestingly, levels of non-responders were below the levels of healthy controls. Furthermore, sTNFR2 levels decreased significantly after 26 weeks of infliximab, from a mean of 159
65 6 78
13 pg/

ml at week 0 to 125
26 6 62
61 pg/ml at week 26 (P < 0
001) (Fig. 4b). In addition, a significant difference in delta sTNFR2 was observed between responders and non-responders: where responders showed a mean reduction in sTNFR2 of 244
65 6 38
32 pg/ml, non-responders showed a slight increase of 13
59 6 22
67 pg/ml (P < 0
001) (Fig. 4c). Lastly, we calculated the correlation between sTNFR2 and known biomarkers of disease activity in sarcoidosis. The correlation between the change in sTNFR2 from weeks 0–26 and the changes in sIL-2R, angiotensin-converting

Fig. 4. Levels of serum tumour necrosis factor receptor 2 (sTNFR2) in sarcoidosis patients at baseline and during infliximab therapy. Serum sTNFR2 levels of 14 healthy controls and 47 sarcoidosis patients (37 responders, 10 nonresponders) before starting therapy with infliximab (a) and during therapy (b,c) were measured by multiplex immunoassay. Horizontal lines represent mean 6 standard deviation. Responders show increased levels of serum sTNFR2 compared to healthy controls (P 5 0
041) and compared to non-responders (P 5 0
015). After 26 weeks of treatment, responders show a decrease in their level of serum sTNFR2 (P < 0
001). The change in serum sTNFR2 is also different between responders (reduction) and non-responders (slight increase) (P < 0
001). C 2016 British Society for Immunology, Clinical and Experimental Immunology, 185: 263–270 V

267

A. Verwoerd et al.

enzyme (ACE) and standardized uptake values (SUVmax) were 0
56 (P < 0
001), 0
54 (P < 0
001) and 0
20 (P 5 0
18), respectively. The correlation between the change in sTNFR2- and TNFR2-positive Tregs was 0
57 (P 5 0
02). Serum levels of TNF were also measured, but these were below the detection level of 0
9 pg/ml (data not shown).

Discussion This study investigated Tregs, TNFR2 expression on T cells and serum sTNFR2 in sarcoidosis patients receiving infliximab. We showed that relative frequencies of Tregs were elevated in patients with refractory sarcoidosis and decreased during infliximab therapy. TNFR2 expression on Tregs was increased in patients and more significantly in responders to therapy. In addition, we demonstrated that serum sTNFR2 levels were significantly higher in responders than in non-responders at baseline, the former decreasing to normal levels following infliximab. Previous studies investigating the influence of TNFbinding on Tregs are inconclusive, and show that TNFbinding promotes Treg expansion [12], but also that forkhead box protein 3 (FoxP3) is dephosphorylated upon TNF binding, leading to suppression of Treg function [10]. The effect of TNF depletion, e.g. through infliximab, on Treg function thus remains elusive. We showed that relative Treg frequencies are elevated in sarcoidosis and decreased following infliximab therapy, suggesting that TNF normally has a stimulatory role. In addition to enhanced TNF–TNFR2 stimulation, which we suggest here, there might be additional mechanisms accounting for the observed elevation in Treg levels. One of these mechanisms include an apoptosis defect, as described recently by Broos et al. [7]. Furthermore, the observed elevation in Treg levels in sarcoidosis patients might be explained by the resistance of Teffs to suppression by Tregs. As a result, Teff cells continue proliferating. This view is supported by the fact that we found that only relative frequencies of Tregs were elevated in sarcoidosis patients compared to controls and decreased during therapy, while absolute numbers remained stable. Although this represents a shift in the balance between Tregs and Teffs during infliximab therapy, it also demonstrates proliferation of Teffs. Furthermore, it was shown previously that Tregs from healthy controls were not able to suppress TNF secretion by CD41CD25– Teffs from sarcoidosis patients, further supporting this hypothesis [6]. Indeed, it is thought increasingly that infliximab also affects Teffs significantly, and that changes in the functionality of Teffs are largely responsible for the efficacy of infliximab. Although the expression of TNFR2 on Teffs is much lower than on Tregs, it was shown that increased TNFR2 expression on Teffs increases its resistance to Treg-mediated inhibition in mice [19]. TNF-depletion through infliximab would then result in down-regulation of TNFR2 expression 268

on Teffs, possibly sensitizing them to suppression by Tregs. In fact, Chen et al. [20] hypothesize that in the early stage of an immune response, Teffs up-regulate their expression of TNFR2 adaptively to be able to mount an effective immune response and resist suppression by Tregs. In later stages, Tregs then outcompete Teffs for TNF and suppress Teffs, thereby restoring immune homeostasis. A disbalance in TNFR2 expression between Teffs and Tregs might then explain the ongoing immune responses in chronic inflammatory conditions such as sarcoidosis. Indeed, we found trends in increased relative frequencies of TNFR2expressing CD251 and CD25– cells in active sarcoidosis. This decreased following infliximab treatment, while TNFR2 expression on Tregs remained high, supporting the hypothesis of Chen et al. The mechanism by which resistance of Teffs to suppression works was investigated by Wehrens et al., who demonstrated that protein kinase B (PKB)/c-Akt hyperactivation plays an important role. Activation of PKB/c-Akt positively regulates the production of proinflammatory cytokines, including TNF, and TNF itself can also induce PKB/c-Akt activation. Blocking this positive feedback loop through anti-TNF therapy might thus also result in the sensitization of Teffs to suppression [14,21,22]. A disbalance in TNFR2 expression between Tregs and Teffs may thus, at least partly, explain the ongoing immune response that characterizes sarcoidosis. However, the disbalance was particularly observed in responders to therapy, whereas this effect was not shown in non-responders. Although this might be due to the small sample size of our group of non-responders, it could also suggest that responders and non-responders represent distinct subgroups of sarcoidosis, in which responders represent a strong TNF-mediated type. Next to membrane-bound TNFR2, we studied serum levels of sTNFR2, which were elevated significantly in responders. It was proposed previously that TNFR2 shedding by Tregs is one of the mechanisms by which Treg cells inhibit the action of TNF and thereby contributes to their immunosuppressive capacity [17,18,23]. Although Tregs are not the only source of shed sTNFR2, the high levels observed in responders might suggest that the body attempts to captivate TNF and thereby inhibit its proinflammatory downstream effects. With regard to the mechanism of action of infliximab, it is remarkable that, specifically, responders to infliximab have high baseline levels of sTNFR2, as infliximab and sTNFR2 both bind TNF and therefore may have similar effects. This apparent paradox might be explained by defective TNF–sTNFR2 binding in patients, resulting in the inability of sTNFR2 to modulate the proinflammatory effects of TNF. It could also merely reflect inflammation and extensive turnover of TNFR2, with higher levels of TNF-expression T cells occurring specifically in responders to therapy, underscoring the concept of sarcoidosis being a predominantly TNF-

C 2016 British Society for Immunology, Clinical and Experimental Immunology, 185: 263–270 V

Tregs and sTNFR2 in sarcoidosis

mediated disease. This would also explain the normalization of sTNFR2 following infliximab therapy when the disease is in remission. Alternatively, sTNFR2 might act as a reservoir for TNF because it binds TNF reversibly, resulting in the prolonged bioavailability of TNF and in fact prolonged continuation of inflammation [24]. In RA, it was demonstrated in turn that high baseline levels of sTNFR2 are correlated with a prolonged therapeutic response to anti-TNF therapy, and that the largest proportion of sTNFR2 is produced through alternative splicing and not through shedding [18]. Although the biological role of sTNFR2 is not entirely clear, our data demonstrate significant differences in serum sTNFR2 levels between responders and non-responders. Mean levels of non-responders were below the level of healthy controls, while levels of responders were elevated significantly before the start of treatment. Potentially, sTNFR2 could be used as a biomarker in determining which patients will benefit from infliximab treatment. The definition of response should then be based on a clinically relevant end-point, such as organ function [e.g. forced vital capacity (FVC)] in pulmonary sarcoidosis, rather than response reflected by reduction in inflammatory parameters which we have used here. Limitations of this study include small group sizes and the limited Treg characterization. The CD41CD25hi subset may be contaminated with highly activated CD25hi Teffs and thus not purely reflect Tregs. Furthermore, it would be interesting to investigate T cells and TNFR2 expression in bronchoalveolar lavage fluid, which might reflect local inflammatory processes more clearly than peripheral blood. In conclusion, increased relative frequencies of Tregs and TNFR2 expression were observed in sarcoidosis, followed by a decline during infliximab therapy. This study also shows that baseline sTNFR2 levels could be of value in selecting patients who might benefit from infliximab treatment. Responders showed both high sTNFR2 levels as well as a decrease in sTNFR2 levels during treatment. In patients not responding to treatment, sTNFR2 levels equalled those of healthy controls. Moreover, nonresponders did not show a decrease in levels during infliximab treatment, suggesting that high sTNFR2 levels and a rapid decrease might serve as a predictive tool. The next step is to validate its predictive value in a larger cohort.

Acknowledgements This study is funded by the St Antonius Hospital Innovation Fund. The authors would like to thank the technicians of the Department of Medical Microbiology and Immunology for collecting the serum samples.

Author contributions A. V. performed statistical analysis and wrote the manuscript. D. H. performed experiments, designed the research protocols and analysed data. A. D. M. V. selected patients and collected blood samples. H. A. C. collected blood samples. C. H. M. designed the research protocols. J. C. G. selected patients and designed the research protocols. A. M. E. C. designed the research protocols. All authors interpreted the data, edited the manuscript, read and approved the final manuscript.

Disclosure The authors do not have any disclosures to declare.

References 1 Valeyre D, Prasse A, Nunes H et al. Sarcoidosis. Lancet 2014; 383:1155–67. 2 Vorselaars AD, Crommelin HA, Deneer VH et al. Effectiveness of infliximab in refractory FDG PET-positive sarcoidosis. Eur Respir J 2015; 46:175–85. 3 Zissel G. Cellular activation in the immune response of sarcoidosis. Semin Respir Crit Care Med 2014; 35:307–15. 4 Grant CR, Liberal R, Mieli-Vergani G et al. Regulatory T-cells in autoimmune diseases: challenges, controversies and-yetunanswered questions. Autoimmun Rev 2015; 14:105–16. 5 Rappl G, Pabst S, Riemann D et al. Regulatory T cells with reduced repressor capacities are extensively amplified in pulmonary sarcoid lesions and sustain granuloma formation. Clin Immunol 2011; 140:71–83. 6 Miyara M, Amoura Z, Parizot C et al. The immune paradox of sarcoidosis and regulatory T cells. J Exp Med 2006; 203:359–70. 7 Broos CE, van Nimwegen M, Kleinjan A et al. Impaired survival of regulatory T cells in pulmonary sarcoidosis. Respir Res 2015; 16:108. 8 Ehrenstein MR, Evans JG, Singh A et al. Compromised function of regulatory T cells in rheumatoid arthritis and reversal by anti-TNFalpha therapy. J Exp Med 2004; 200:277–85. 9 Nadkarni S, Mauri C, Ehrenstein MR. Anti-TNF-alpha therapy induces a distinct regulatory T cell population in patients with rheumatoid arthritis via TGF-beta. J Exp Med 2007; 204:33–9. 10 Nie H, Zheng Y, Li R et al. Phosphorylation of FOXP3 controls regulatory T cell function and is inhibited by TNF-alpha in rheumatoid arthritis. Nat Med 2013; 19:322–8. 11 Valencia X, Stephens G, Goldbach-Mansky R et al. TNF downmodulates the function of human CD41CD25hi T-regulatory cells. Blood 2006; 108:253–61. 12 Chen X, Baumel M, Mannel DN et al. Interaction of TNF with TNF receptor type 2 promotes expansion and function of mouse CD41CD251 T regulatory cells. J Immunol 2007; 179: 154–61. 13 Chen X, Wu X, Zhou Q et al. TNFR2 is critical for the stabilization of the CD41Foxp31 regulatory T. cell phenotype in the inflammatory environment. J Immunol 2013; 190:1076–84. 14 Byng-Maddick R, Ehrenstein MR. The impact of biological therapy on regulatory T cells in rheumatoid arthritis. Rheumatology (Oxf) 2015; 54:768–75.

C 2016 British Society for Immunology, Clinical and Experimental Immunology, 185: 263–270 V

269

A. Verwoerd et al. 15 Lainez B, Fernandez-Real JM, Romero X et al. Identification and characterization of a novel spliced variant that encodes human soluble tumor necrosis factor receptor 2. Int Immunol 2004; 16:169–77. 16 Xanthoulea S, Pasparakis M, Kousteni S et al. Tumor necrosis factor (TNF) receptor shedding controls thresholds of innate immune activation that balance opposing TNF functions in infectious and inflammatory diseases. J Exp Med 2004; 200: 367–76. 17 Heaney ML, Golde DW. Soluble receptors in human disease. J Leukoc Biol 1998; 64:135–46. 18 Canete JD, Albaladejo C, Hernandez MV et al. Clinical significance of high levels of soluble tumour necrosis factor-alpha receptor-2 produced by alternative splicing in rheumatoid arthritis: a longitudinal prospective cohort study. Rheumatology (Oxf) 2011; 50:721–8. 19 Chen X, Hamano R, Subleski JJ et al. Expression of costimulatory TNFR2 induces resistance of CD41FoxP3- conventional T

270

20

21

22

23

24

cells to suppression by CD41FoxP31 regulatory T cells. J Immunol 2010; 185:174–82. Chen X, Oppenheim JJ. Contrasting effects of TNF and antiTNF on the activation of effector T cells and regulatory T cells in autoimmunity. FEBS Lett 2011; 585:3611–8. Wehrens EJ, Mijnheer G, Duurland CL et al. Functional human regulatory T cells fail to control autoimmune inflammation due to PKB/ c-AKT hyperactivation in effector cells. Blood 2011; 118:3538–48. Cheng J, Phong B, Wilson DC et al. Akt fine-tunes NF-kappaBdependent gene expression during T cell activation. J Biol Chem 2011; 286:36076–85. Van Zee KJ, Kohno T, Fischer E et al. Tumor necrosis factor soluble receptors circulate during experimental and clinical inflammation and can protect against excessive tumor necrosis factor alpha in vitro and in vivo. Proc Natl Acad Sci USA 1992; 89:4845–9. Aderka D. The potential biological and clinical significance of the soluble tumor necrosis factor receptors. Cytokine Growth Factor Rev 1996; 7:231–40.

C 2016 British Society for Immunology, Clinical and Experimental Immunology, 185: 263–270 V