High thioredoxin-1 levels in rheumatoid arthritis patients diminish binding and signalling of the monoclonal antibody Tregalizumab

The humanized non-depleting anti-CD4 monoclonal antibody Tregalizumab (BT-061) is able to selectively activate the suppressive function of regulatory T cells and has been investigated up to phase IIb in clinical trials in patients suffering from rheumatoid arthritis (RA). A pharmacokinetic–pharmacodynamic model based on clinical data from RA and healthy volunteers, which used the cell surface CD4 downmodulation as marker of activity, confirmed a stronger effect in healthy volunteers compared with RA patients. We tried to understand this phenomenon and evaluated the influence of the small oxidoreductase thioredoxin-1 (Trx1). To counteract oxidative stress that is strongly associated with RA pathophysiology, the organism employs Trx1. Therefore, increased expression and secretion of Trx1 is found in the synovial fluid and plasma of RA patients. Moreover, the binding site of Tregalizumab is in close proximity to a disulphide bond in domain 2 (D2) of CD4, which is a known target for a reduction by oxidoreductase Trx1. With the experiments reported herein, we demonstrated that specific reduction of the D2 disulphide bond by Trx1 led to diminished binding of Tregalizumab to recombinant human soluble CD4 and membrane-bound CD4 on T cells. Moreover, we showed that this caused changes in the Tregalizumab-induced CD4 signalling pathway via the lymphocyte-specific protein tyrosine kinase p56Lck and CD4 downmodulation. In summary, we provide evidence that high Trx1 levels in RA patients compared with healthy subjects are a potential reason for diminished binding of Tregalizumab to CD4-positive T cells and offer an explanation for the observed decreased CD4 downmodulation in RA patients in comparison to healthy subjects.

The humanized non-depleting anti-CD4 monoclonal antibody Tregalizumab (BT-061) is able to selectively activate the suppressive function of regulatory T cells and has been investigated up to phase IIb in clinical trials in patients suffering from rheumatoid arthritis (RA). A pharmacokinetic-pharmacodynamic model based on clinical data from RA and healthy volunteers, which used the cell surface CD4 downmodulation as marker of activity, confirmed a stronger effect in healthy volunteers compared with RA patients. We tried to understand this phenomenon and evaluated the influence of the small oxidoreductase thioredoxin-1 (Trx1). To counteract oxidative stress that is strongly associated with RA pathophysiology, the organism employs Trx1. Therefore, increased expression and secretion of Trx1 is found in the synovial fluid and plasma of RA patients. Moreover, the binding site of Tregalizumab is in close proximity to a disulphide bond in domain 2 (D2) of CD4, which is a known target for a reduction by oxidoreductase Trx1. With the experiments reported herein, we demonstrated that specific reduction of the D2 disulphide bond by Trx1 led to diminished binding of Tregalizumab to recombinant human soluble CD4 and membrane-bound CD4 on T cells. Moreover, we showed that this caused changes in the Tregalizumab-induced CD4 signalling pathway via the lymphocyte-specific protein tyrosine kinase p56 Lck and CD4 downmodulation. In summary, we provide evidence that high Trx1 levels in RA patients compared with healthy subjects are a potential reason for diminished binding of Tregalizumab to CD4-positive T cells and offer an explanation for the observed decreased CD4 downmodulation in RA patients in comparison to healthy subjects. Rheumatoid arthritis (RA) is a common autoimmune disorder, which is characterized by chronic inflammation, bone and cartilage damage. 1,2 Autoimmune disorders are caused by an aberrant immunological function, including a failure of immunological self-tolerance leading to inflammations as seen in RA. 2,3 To control excessive immune responses regulatory T cells (Tregs) have an important role in immune homeostasis to maintain this selftolerance. 4,5 Tregs represent a T-cell subset, which is able to suppress pathological immune response of other surrounding cells. 6 Similar to other T cells, Tregs are initially developed in the thymus and express, among others, the interleukin-2 receptor alpha chain (CD25) and the glycoprotein CD4 molecule on their cell surface. 4,7 In the periphery, Tregs become activated upon encounter with an antigen presented on the major histocompatibility complex class II on antigen-presenting cells. They are recognized via the T-cell receptor, which associates the CD4 molecule in its complex. 7,8 CD4 consists of four extracellular domains (immunoglobulin-like (Ig) domains D1-D4), a transmembran domain and a cytoplasmic domain. 9,10 Three disulphide bridges are present in the CD4 molecule, of which the disulphide bond in D2 (between cysteines Cys155 and Cys184) is remarkable as this domain differs from the normal IgG structure. 11 In the D2 domain, an intrasheet disulphide bridge substitutes an intersheet disulphide bridge, which is usually located in IgG domains. 11,12 Owing to its special geometry, a high dihedral strain energy and low calculated enthalpy D2 can be reduced selectively by the oxidoreductase thioredoxin-1 (Trx1). 13 This reduction of the disulphide bridge has an important role in the HIV entry process, allowing the interaction with glycoprotein 120 (gp120) of HIV with the CD4 molecule. 14 Trx1, which belongs to the thioredoxin family, is a small (12 kDa) oxidoreductase secreted by activated monocytes, lymphocytes and other immune cells, 15 which exerts its function with help of thioredoxin reductase (TrxR) and NADPH in a disulphide exchange reaction 16 (for a more detailed insight, the reader is referred to Lu and Holmgren 16 ).
Many studies describe a significant association between RA and oxidative stress caused by reactive oxygen species [17][18][19] and point out that Trx1 levels are augmented in patients suffering from RA [20][21][22][23] or other diseases, such as heart, 24 lung 25 and liver complaints, 26 probably counteracting the increased oxidative stress in these diseases. Serum and plasma Trx1 levels were found to be elevated in RA compared with healthy subjects. 21,22,27 Moreover, the synovial fluid of RA patients revealed higher Trx1 concentrations than those of osteoarthritis. 22,27 In addition, it was also demonstrated that RA disease activity correlated with elevated Trx1 plasma levels 22 and Trx1 activity. 23 The high concentration of Trx1 described for RA patients in particular 21 might be associated with the hyporesponsiveness of T cells in RA patients. 20 New strategies to tackle autoimmune reactions and to re-establish a good immunological balance are the enhancement of activity of CD4 + CD25 + Tregs. Therefore, anti-CD4 antibodies were investigated emphasizing CD4 as an interesting target. 28,29 Tregalizumab (BT-061) is a non-depleting IgG1 monoclonal antibody, which binds to a unique epitope of CD4 and represents the first humanized anti-CD4 monoclonal antibody that selectively induces Treg activation. 30 As a new therapeutic approach, Tregalizumab has been investigated in eight clinical trials, including healthy subjects, RA and psoriasis patients. The antibody was investigated in RA to restore the functionality of defective Tregs and to enhance their suppressive capacity. 31 Thereby, Tregalizumab exerts its agonistic function by binding to a unique, non-linear epitope on D2 of the human CD4 molecule and induces an intracellular signalling cascade involving the protein tyrosine kinase p56 Lck . 30 After administration of Tregalizumab, downmodulation of CD4 expression levels can be determined in vitro and in vivo, representing a marker for the antibody activity. 31,32 In the latest phase IIb trial (TREAT 2b, T-cell Regulating Arthritis Trial 2b, ClinicalTrials. gov Identifier: NCT01999192 and EudraCT No. 2013-000114-38) conducted with patients suffering from RA, the primary end point was not achieved. We describe in this manuscript experiments supporting a potential link between this observation and the reaction against the oxidative stress described in RA.

RESULTS
Tregalizumab binds to CD4 in close proximity to the disulphide bridge in D2 The first step of the Tregalizumab mode-of-action is binding to the CD4 molecule via its Fab fragment followed by a cross-link with the Fcγ I receptor (CD64) on monocytes activating a signalling cascade. 33 As outcome, Tregs are activated and CD4 molecules on regulatory and effector T cells bound by Tregalizumab are internalized (Figure 1a). In particular, a signal is transduced into the T cell via the CD4-associated protein tyrosine kinase p56 Lck , leading ultimately to internalization of CD4 molecules and decline of surface expression. The declined CD4 expression refers to a decrease of CD4 on CD4-positive T cells that can be measured by flow cytometry in vivo and in vitro. 31 Indeed, it was shown that Tregalizumab induces a decrease in the CD4 surface expression in vitro and in treated patients in vivo. [30][31][32] For a better understanding of Tregalizumab's dose-response effects in patients, a pharmacokinetic-pharmacodynamic model was developed using the downmodulation of the CD4 molecule on the T-cell surface as pharmacodynamic marker (Rharbaoui et al., manuscript in preparation). 32 The pharmacokinetic-pharmacodynamic model based on clinical data from RA and healthy volunteer trials surprisingly showed higher CD4 downmodulation in healthy subjects compared with RA patients (Figure 1b). Looking for a possible explanation of this observation, we went back to the Tregalizumab-CD4 interaction. Protein crystallography revealed that Tregalizumab binds to D2 of CD4 to a conformational epitope that is not recognized by other anti-CD4 monoclonal antibodies. 30 The epitope is close to an intramolecular disulphide bridge (Figure 2a). Distances from Cys155 and Cys184 to Tregalizumab's tyrosine105 (Tyr105) carbon atoms were calculated to be 7.0-11.0 Å. As this disulphide bond between cysteines Cys155 and Cys184 is in very close proximity to the binding interface of Tregalizumab to CD4 (Figure 2b), it is likely to influence the binding characteristics between both entities. This disulphide bridge has been described to be selectively reduced by Trx1. 13 To examine whether reduction of this disulphide bridge may be responsible for higher CD4 downmodulation in healthy subjects compared with RA patients, we further examined the role of Trx1 on Tregalizumab binding to CD4 and the role of Trx1 in RA in general. Figure 1 Tregalizumab mode of action and predicted CD4 modulation: (a) Tregalizumab (BT-061) binds to the CD4 receptor of T cells and is crosslinked via its Fc part by the Fcγ I receptor (CD64) on monocytes. This leads to a signal transduction into the cell and selective activation of regulatory Tregs. In addition, the CD4 receptor is internalized and turns off signalling. This CD4 downmodulation can be measured in vitro and in vivo. (b) The theoretical downmodulation of the cell surface protein CD4 after Tregalizumab administration as predicted by a pharmacokineticpharmacodynamic model is depicted after the first SC dosing (predicted mean values ± the 95% confidence interval). The predicted CD4 downmodulation in healthy subjects (light curve) compared with patients with RA (dark curve) is shown.  Figure S1). This is in accordance with results shown by Matthias et al. 13 Treatment with Trx1 diminishes Tregalizumab binding to CD4 Next, we elaborated whether reduction of rh sCD4 by Trx1 impacts the Tregalizumab-CD4 interaction by using a well-established electrochemiluminescence enzyme-linked immunosorbent assay (MSD system). Binding of Tregalizumab to Trx1-treated rh sCD4 was compared with untreated rh sCD4 control, which was set to 100%.
As shown in Figure 3a, binding of Tregalizumab to rh sCD4 was decreased to 34.3 ± 15.2% after pretreatment with 20 μM Trx1.
Next, the influence on binding of Tregalizumab to cell surface CD4 was assessed using the CD4-positive HPB-ALL cells under physiological conditions together with the Trx1/TrxR system. HPB-ALL cells incubated with 150 nM TrxR plus 5 μM Trx1 overnight resulted in a significantly reduced binding of Tregalizumab from 661 ± 91 to 147 ± 45 mean fluorescence intensity (MFI) units of the BT-061 allophycocyanin (APC) staining (P ¼ 0.0041). This corresponds to a reduction of binding of 78% compared with the non-treated control (Figure 3b).
Incubation with 5 μM Trx1 alone without reductase displayed a less pronounced reduction of Tregalizumab binding. This is assumed as Trx1 without the reductase is not recycled and can facilitate the Trx1 impacts binding of Tregalizumab to CD4 in RA K Heim et al disulphide reduction process only once per molecule. As expected, the inactive control Trx1-SS was not able to reduce the Tregalizumab binding to CD4.
These results were confirmed using peripheral blood mononuclear cells (PBMCs) isolated from healthy subjects (Figure 3c). Pretreatment of PBMCs with Trx1/TrxR resulted in a significant (P ¼ 0.0007) decrease of 71% of Tregalizumab binding to CD4 + cells compared with untreated cells (from 479 ± 73 to 140 ± 75 MFI of the BT-061 APC staining), whereas the inactive mutant showed no significant decline of Tregalizumab binding. Preincubation of PMBC with Trx1 without reductase also revealed a less pronounced diminished binding of Tregalizumab comparable to the result observed on HPB-ALL cells.
The effect of Trx1 on binding of Tregalizumab to CD4 could also be confirmed using CD4-positive monocytes present in the PBMCs (data not shown).
Beside Trx1, other molecules are able to modulate the redox potential in the blood, such as glutathione (GSH), which represents a physiologically relevant reducing agent. 34 To investigate a potential influence of GSH on the stability of the CD4 disulphide bridges and the interaction between Tregalizumab and CD4, further binding experiments were performed. Even when using GSH concentrations up to 10 mM, no significant effect on binding of Tregalizumab to CD4 + cells from PBMCs was observed (Figure 3c).
These results demonstrate that recombinant and cellular CD4 can be selectively reduced by Trx1, and this reduction diminishes the binding of Tregalizumab to the receptor. Other physiologically relevant reducing agents such as GSH do not influence binding of Tregalizumab to CD4.
The D2 disulphide bond impacts Tregalizumab binding to CD4-positive cells In order to confirm that the disulphide bond in D2 between cysteines Cys155 and Cys184 directly impacts Tregalizumab binding to CD4, a mutant of CD4 lacking both cysteines was constructed. CD4-negative U266B1 cells were transfected with CD4 variants, which are not able to form a disulphide bridge, as the corresponding cysteines Cys155 and Cys184 were replaced by alanine. Cells which   (Figure 3d). This result further confirmed that the D2 disulphide bridge, which can be selectively reduced by Trx1, is crucial for binding of Tregalizumab to CD4.

Trx1 has no influence on CD64 expression on monocytes
To test a potential influence on CD64 binding of Tregalizumab, PBMCs encompassing monocytes were incubated with the Trx1/TrxR system and the CD64 binding on monocytes was determined by flow cytometry using a specific anti-CD64 antibody. Overall, no significant difference in CD64 expression was observed after Trx1 pretreatment or GSH preincubation, respectively (Figure 3e).

Trx1 influences Tregalizumab-induced CD4 signalling
To assess how the signal transduction induced by Tregalizumab is influenced by Trx1 and its physiological partners, phosphorylation of p56 Lck and zeta-chain-associated protein kinase 70 (ZAP-70) was analysed. In former studies, no significant differences in phosphorylation patterns of Tregs and effector T cells were observed after Tregalizumab stimulation. 30 Hence, total CD4 + T cells isolated from PBMCs were analysed. Phosphorylation of human p56 Lck after stimulation of cells with cross-linked Tregalizumab in the presence or absence of Trx1/TrxR was analysed by flow cytometry using a phosphorylation-specific antibody against p56 Lck . Preincubation of CD4 + cells with Trx1/TrxR resulted in a significant decrease of phosphorylation (1.2 ± 0.2-fold induction of phosphorylation, P ¼ 0.0057) compared with the Trx1 untreated control (2.1 ± 0.3-fold induction of phosphorylation) and the inactive Trx1-SS control with TrxR (2.1 ± 0.5-fold induction of phosphorylation; Figure 4a).
As a positive control for T-cell signalling, the CD3 antibody OKT3 was used. Stimulation of CD4 + cells with OKT3 revealed an increase in phosphorylation (3.5 ± 0.8-fold induction of phosphorylation). Preincubation with Trx1 or Trx1-SS showed no significant change on OKT3-mediated phosphorylation (3.5 ± 1.0-and 3.3 ± 0.6-fold induction of p56 Lck phosphorylation, respectively). In contrast to p56 Lck , pretreatment with active Trx1 showed no significant impact on ZAP-70 phosphorylation (Figure 4b).
Therefore, it can be concluded that reduction of the CD4 D2 disulphide bridge by Trx1 negatively impacts binding of Tregalizumab, which also affects the signalling pathway via phosphorylation of p56 Lck , whereas ZAP-70 phosphorylation after CD3 stimulation is not impacted by the small oxidoreductase.
Binding of Tregalizumab is decreased in RA patients compared with healthy subjects in clinical trials Having established the effect of reduction of the D2 disulphide bond on Tregalizumab binding to CD4 in vitro, we evaluated whether the increased levels of Trx1 in RA patients lead to the same effect in vivo. Data from two clinical studies were analysed retrospectively concerning the binding of Tregalizumab to CD4. In total, results of 127 RA patients (Biotest study 979, EudraCT: 2010-018485-24) and 35 healthy subjects (Biotest study 985, EudraCT: 2011-004956-20) were evaluated. Relative Tregalizumab binding in relation to the total CD4 molecules on T cells was analysed.
As shown in Figure 5a, relative binding of Tregalizumab (ratio of MFIs between BT-061 and the non-competing anti-CD4 antibody SK3) in RA patients is significantly (Po0.0001) lower (6.9 ± 1.6, n = 127) compared with healthy subjects (8.3 ± 0.8, n = 35). The observed difference in binding is due to the difference in mean binding of Tregalizumab to CD4 ([BT-061 + /CD4 − ]), which is significantly lower in RA patients compared with healthy subjects (Figure 5b), whereas the overall CD4 expression ([CD4 + /CD4 − ]) in general is comparable between RA patients and healthy subjects (Figure 5c).

DISCUSSION
The literature describes a strong connection between oxidative stress and RA disease, while our results demonstrated an impaired binding and signalling of Tregalizumab (BT-061) in this population. The work presented in this manuscript aimed to investigate a potential link between those observations. Oxidative stress is considered to be a possible elicitor in the development and disease activity of RA 35 and other diseases, such as metabolic disorder. 36 One possibility for the organisms to counteract high levels of oxidative metabolites is to increase the level of the oxidoreductase Trx1. It represents one of the major players in redox processes for maintaining redox homeostasis, 16 as it is able to reduce oxidized proteins. So numerous studies have investigated the relationship between Trx1 and the pathogenesis of RA. [20][21][22][23] They revealed that Trx1 protein and activity levels were augmented in patients suffering from RA. [20][21][22][23]27 The literature observations on elevated Trx1 levels in plasma of RA patients were confirmed by our own analyses (data not shown).
Starting with the unexpected differences in predicted CD4 downmodulation between RA patients and healthy subjects, we were able to demonstrate that the disulphide bridge in D2 of the CD4 molecule, which can be selectively reduced by Trx1, 13 is important for the Tregalizumab-CD4 interaction. We demonstrated that Trx1 representing a so called − RHS staple and has a major role in other physiological processes owing to its allosteric properties. 37 For example, dimerization of CD4 is important for the co-receptor function of CD4, which is induced by Trx1. 38,39 Apart from this fact, the HIV entry process is closely linked to redox changes in the D2 disulphide bond, implicating a relevant role of this molecule. 13,14,40,41 Regarding Tregalizumab, transfection experiments also emphasized the importance of this disulphide bridge in particular, as CD4 variants lacking the D2 disulphide bond showed a reduced binding of the antibody. The pharmacological effect of CD4 downmodulation can only occur when the Fc part of Tregalizumab is cross-linked with the Fc-gamma receptor 1 (FcγRI, CD64) on monocytes. 33 It could be possible that the higher CD4 downmodulation in healthy subjects compared with RA patients is caused either by a decreased binding of Tregalizumab's Fab part to CD4, by diminished expression or by modification of the FcγRI CD64. As the expression of CD64 is not influenced by Trx1 and no disulphide bridges are described for the CD64 receptor, 42 Trx1 can only influence binding of Tregalizumab to CD4 and not to the CD64 receptor on monocytes. Binding of Tregalizumab to CD4 + monocytes was assessed (data not shown) and confirmed the assumption that binding of Tregalizumab to all CD4 + cells is impacted by Trx1. Trx1 alone at a concentration of 5 μM was able to significantly decrease Tregalizumab binding, but much stronger effects were observed using the physiological Trx1 system, including TrxR and NADPH, allowing for Trx1 recycling. Although the Trx1 concentration exceeds the published levels in plasma or synovial fluid of 1-5 nM, 43 the local Trx1 concentration at the cell surface of T cells is described to be higher. This might result from the expression and secretion of Trx1 which is noticeably enhanced by some cell types such as dendritic cells and macrophages after activation. Gromer et al. 34 even reported Trx1 tissue concentrations between 1 and 20 μM and calculated intracellular levels of TrxR of 1 μM, illustrating physiological relevance of the used Trx1 concentration.
Stimulation of the CD4 molecule induces signalling events of the lymphocyte-specific protein tyrosine kinase p56 Lck . 44 Moreover, phosphorylation of ZAP-70 can be detected after activation of the T-cell receptor-CD3 complex. 44 Helling et al. 30 reported that binding of Tregalizumab to the T-cell receptor induces a unique signalling in T cells via the CD4 molecule. To assess this unique T-cell signalling pathway, we focussed on phosphorylation of p56 Lck and ZAP-70 as the first molecules of the signalling cascade triggered after activation of CD4 and CD3 by the control antibody OKT3, respectively. We demonstrated that Trx1 preincubation resulted in a significant decrease of p56 Lck phosphorylation, confirming an impact of Trx1 on Tregalizumab-induced signalling. These findings implicate that Trx1 also indirectly influences CD4-initiated signalling events, subsequent CD4 receptor internalization and probably also activation of T cells in general via influencing binding of the antibody to cell surface CD4. Figure 6 summarizes the proposed Trx1 effect on Tregalizumab-  Trx1 impacts binding of Tregalizumab to CD4 in RA K Heim et al mediated signalling. As Treg function depends on a functional T-cell signalling, addition of Trx1 might have an influence on Treg function in particular. Using in vitro assays, it was shown that Tregalizumab is able to activate Tregs selectively in contrast to other CD4 antibodies. 30 By decreasing binding of Tregalizumab to D2 of CD4 and reducing T-cell activation signal by Trx1, this unique phosphorylation signalling is affected and that is supposed to lead to a declined Treg function. However, determination of CD4 modulation as a pharmacodynamic marker confirmed activity of Tregalizumab in all clinical trials. 31 As elevated Trx1 levels impacted binding of Tregalizumab with CD4 in vitro, we determined Trx1 concentrations in RA patients retrospectively to find a correlation between Trx1 levels and efficacy of the antibody. Measurement of Trx1 levels in RA plasma samples did not reveal a significant correlation between Trx1 plasma concentration and efficacy (data not shown). However, Trx1 is secreted by different cell types such as activated monocytes or lymphocytes 15 and operates at the site of inflammation where those cells are abundant. [45][46][47] As only Trx1 plasma levels could be analysed, local Trx1 concentrations at the Tregalizumab-binding site on T cells can be much higher. Therefore, high local Trx1 levels maybe present on the cell surface and cannot be detected in the plasma. In addition, it was shown that Trx1 expression and secretion is higher in human Tregs compared with other CD4-postive T cells. 48 As the RA synovium is enriched with CD4 + CD25 + Tregs, 49 a Trx1 secretion by these cells might enhance the survival of autoreactive synovial fluid T cells owing to inhibition of apoptosis by Trx1. This potentially results in aggravation of RA. As Tregalizumab interacts with CD4 + CD25 + Tregs, 31 which secrete high amounts of Trx1, it can be speculated that Trx1 might also negatively impact the binding and signalling of Tregalizumab in the synovium of RA patients. So the Trx1 influence on the Tregalizumab activity should be taken into account in other clinical indications of this remarkable therapeutic antibody.
In summary, we hypothesized that elevated Trx1 levels in RA patients are a reason for the observed diminished CD4 downmodulation after Tregalizumab administration compared with healthy subjects. This hypothesis predicts decreased binding of Tregalizumab to CD4 in RA patients owing to higher Trx1 concentrations. And this is exactly what was observed in a retrospective analysis of Biotest clinical studies 979 and 985: relative binding of Tregalizumab to CD4 was significant lower in RA patients compared with healthy subjects, whereas the CD4 expression was nearly comparable in both populations. This finding emphasizes that Tregalizumab's binding is impaired in a particular way in RA patients. As Tregalizumab's target molecule is also a favoured target for T-cell pharmacological regulation in general, for example, by depleting CD4 + cells or downmodulating the CD4 receptor, high levels of Trx1 can potentially impact other anti-CD4 antibodies, which bind in close proximity to the D2 disulphide bridge.

Thioredoxin-1
Recombinant human thioredoxins (wild-type Trx1 and inactive mutant Trx1-SS) were provided by the Redox Regulation group at the German Cancer Research Center (DKFZ), DKFZ-ZMBH Alliance, Heidelberg, Germany.

Preparation of Trx1
Trx1 was incubated with 5 mM DTT (GE-Healthcare, Freiburg, Germany) to reduce all molecules. Prior to experiments, residual DTT was removed using Zeba Spin desalting columns (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions.

Blood and plasma
Blood was obtained from healthy blood donors from the blood donation center (Bio-Rad, Dreieich, Germany) and used within 2 h after blood collection.

Cell culture
Human T-cell leukaemia cells (HPB-ALL) (DSMZ, Braunschweig, Germany) were cultured at 37°C, in an atmosphere of 5% CO 2 in Gibco RPMI 1640 medium (Thermo Fisher Scientific) supplemented with 2 mM L-glutamine (Lonza, Basel, Switzerland) and 20% fetal calf serum (FCS, GE-Healthcare, Freiburg, Germany). Cells were spit twice a week and tested for mycoplasma.

Isolation of PBMCs
Ficoll-Hypaque density-gradient separation was used to obtain PBMCs from citrated blood from healthy subjects. Cells were incubated overnight at 37°C and 5% CO 2 in Gibco RPMI 1640 medium (supplemented with 2 mM L-glutamine and 10% FCS). Cells were harvested by centrifugation for 10 min at 220 g at room temperature.
SDS-PAGE (sodium dodecyl sulphate-polyacrylamide gel electrophoresis) and silver staining SDS-PAGE was performed using a NuPAGE X Cell Sure Lock mini-Cell system (Thermo Fisher Scientific). If possible, samples were diluted to 1 mg ml − 1 with distilled water. For analysis, non-reduced SDS-PAGE samples were prepared as follows: one part N-ethylmaleimide (400 mM) (Alfa Aesar, Karlsruhe, Germany) was mixed with one part lithium dodecyl sulfate sample buffer 4 × (Thermo Fisher Scientific) (NR mixture). The sample was diluted at a ratio of 1:2 with the NR mixture and denaturized by heating for 30 s at 70°C.
In all, 5-25 μl of samples were pipetted into a ready-to-use NuPAGE Novex 4-12% Bis-Tris gel and NuPAGE Novex 10% Bis-Tris gel (both gels were purchased from Thermo Fisher Scientific). The gel was run for 50 min at 200 V. Molecular weight was determined by SeeBlue Plus 2 Prestained Standard (Thermo Fisher Scientific).
Silver staining was performed using the Silver Stain Plus Kit (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer's instructions.
The gel was incubated for 30 min at 20 r.p.m. in fixation buffer (37% V/V methanol, 7% V/V acetic acid in distilled water) subsequently after electrophoresis to fix the proteins. The gel was washed two times with distilled water for 20 min at 20 r.p.m. on the laboratory flat shaker before staining with silver staining solution. Staining was performed for 5-15 min. The staining procedure was stopped using 5% acetic acid (Merck, Darmstadt, Germany) in distilled water and the gel was washed again for three times for 10 min with distilled water. The gel was scanned using an Epson Perfection V700 Photo scanner (SEIKO Epson CORPORATION, Suwa, Japan).
After the coating procedure, the plates were washed with a plate washer (Tecan, Crailsheim, Germany) three times using 300 μl per well washing buffer (TBST). Afterwards, plates were tapped dry.
A total of 200 μl per well blocking buffer (2% BSA (Sigma Aldrich) in Dulbecco's phosphate-buffered saline (D-PBS, Thermo Fisher Scientific)) was incubated for 75 ± 15 min at room temperature followed by three washing steps with 300 μl per well washing buffer.
Overall, 20 μM of Trx1 were added and incubated for 30 ± 15 min at room temperature.
Before addition of biotinylated Tregalizumab, the plate was washed three times (Biotinylation of Tregalizumab was performed using EZ-Link NHS-SS-Biotin (Thermo Fisher Scientific) according to the manufacturer's instructions).
Trx1 impacts binding of Tregalizumab to CD4 in RA K Heim et al After an additional washing step, 150 μl per well Read Buffer T (MSD) (diluted in a ratio 1:2 with distilled water) were added to the samples and shook for 2 min at 500-700 r.p.m. before measurement. Results were obtained using a Sector Imager 6000 (MSD).

Flow cytometry
Antibodies were titrated for optimal staining prior to use. Labelling of Tregalizumab with APC was performed using an AnaTag APC Labelling Kit (AnaSpec, Seraing, Belgium) according to the manufacturer's instructions. Fluorescein isothiocyanate (FITC)-labelling agent was purchased from Thermo Fisher Scientific, USA. The phycoerythrin (PE)-labelled anti-CD4 antibody OKT4 was purchased from BioLegend (San Diego, CA, USA; order number: 317409). Results were obtained using a four-colour FACS Calibur (BD, Heidelberg, Germany) with the BD CellQuest Software (BD, Heidelberg, Germany).

Tregalizumab binding to cellular CD4
For binding experiments of Tregalizumab to cells, HPB-ALL cells as well as PBMCs were used. PBMCs were isolated from citrated blood according to the Ficoll-Hypaque density separation protocol as described above.
In all, 1 × 10 5 cells were incubated with 1 mM NADPH (AppliChem, Darmstadt, Germany), 150 nM thioredoxin reductase (Sigma Aldrich, Buchs, Switzerland) and 5 μM Trx1 in a 96-well U-bottom plate at 37°C and 5% CO 2 for 15 ± 2 h overnight. The redox inactive mutant thioredoxin Trx1-SS served as a control. Here, the active site residues Cys 32 and Cys 35 were substituted by serine, and the additional non-active cysteines were replaced by alanine. PBMCs were incubated in RPMI 1640+10% FCS+2 mM L-glutamine; HPB-ALL cells in Dulbecco's phosphate-buffered saline (D-PBS). For assessment of GSH impact on CD4, PBMC cells were pretreated with 10 mM reduced GSH (Sigma Aldrich) and incubated overnight. Cells were centrifuged at 220 g for 10 min at room temperature and washed three times using D-PBS.
Monocytes were gated according to their size and granularity (forward scatter/side scatter (FSC/SSC)). MFI of anti-CD64 PE (order number: 558592, BD Biosciences, Heidelberg Germany) was analysed by means of flow cytometry (BD FACSCalibur).
HPB-ALL cells were stained with BT-061-APC. Mean fluorescence was determined by flow cytometry (BD FACSCalibur).

Transfection experiments
U266B1 cells were transfected via electroporation using Amaxa Cell Line optimizing Solution Box (Lonza). DNA of CD4 and mutated CD4 was generated by the GeneArt Gene Synthesis service (Thermo Fisher Scientific). In all, 675 μl SF solution was mixed with 150 μl supplement. A total of 5 × 10 5 cells were centrifuged for 10 min at 250 g and room temperature. Pellet was resuspended in 20 μl supplemented SF solution. 4 μg DNA were added to the cell suspension, which was afterwards transferred into a 16-well Nucleocuvette Strip (Lonza). An eGFP (pmaxGFP) served as positive control.
Transfection was performed by Nucleofector (Lonza) applying programme no. DY-100. Owing to regeneration reasons, cells were diluted 1:5 with prewarmed RPMI 1640 medium after the transfection procedure and incubated for 10 min at 37°C and 5% CO 2 . Afterwards, cells were transferred into a 24-well culture plate and incubated in 500 μl RPMI 1640+10% FCS+2 mM L-glutamine.
For this purpose, CD4 + cells were isolated from human PBMCs via Dynabeads Untouched Human CD4 T Cells Kit (Thermo Fisher Scientific) according to the manufacturer's instructions. A total of 1 × 10 5 cells were incubated with 5 μM Trx1, 150 nM thioredoxin reductase and 1 mM NADPH overnight for 15 ± 2 h at 37°C and 5% CO 2 . Cells were washed three times with D-PBS and resuspended in RPMI 1640 enriched with 10% FCS and 2 mM L-glutamine. Tregalizumab (1 μg ml − 1 , Biotest AG) and OKT3 (1 μg ml − 1 , order number: 317304, BioLegend), respectively were added and incubated on a plate shaker (300 r.p.m.) for 30 min at room temperature. Afterwards, cells were washed with D-PBS/3% FCS.
Cells stimulated with Tregalizumab were cross-linked with 20 μg ml − 1 polyclonal anti-human IgG (ordner number: H10300, Thermo Fisher Scientific), and OKT3-stimulated cells were cross-linked with 10 μg ml − 1 of antimurine IgG (order number: 31188, Thermo Fisher Scientific) for 10 min at 37°C and 5% CO 2 . Prewarmed 100 μl fixation buffer (BD Biosciences, Heidelberg, Germany) was added to each sample, and the cells were incubated for 10 min at 37°C and 5% CO 2 . After washing with D-PBS/3% FCS, cells were treated with cold Perm Buffer III (BD Biosciences, Heidelberg, Germany), incubated for 30 min on ice and washed again. Cells were stained for 30 min on ice. Afterwards, cells were washed and resuspended in 200 μl D-PBS/3% FCS. MFI was determined using a FACSCanto (FACSDiva Software, BD, Heidelberg, Germany).
MFI of the measured value was divided by the MFI of the untreated control referring to the fold induction.
The inactive mutant Trx1-SS served as a control. Trx1-treated cells were compared with the untreated sample.

Gating strategies
For PBMCs, the CD4 binding was determined by gating on lymphocytes according to cell size and granularity (FSC/SSC) and then within the gate on CD3 + BT-061 + cells. The monocyte population was identified by gating of cell size and granularity (FSC/SSC). Within the gate, the fluorescence of PE-labelled anti-CD64 (BD Biosciences, Heidelberg, Germany) was analysed.
For HPB-ALL cells, the CD4 binding was determined by gating on lymphocytes according to cell size and granularity (FSC/SSC) and then within the gate on APC-labelled BT-061 + cells.
CD4 binding of transfected U266B1 cells was determined by gating according to cell size and granularity (FSC/SSC) and then within the gate on CD4 + cells (APC-labelled BT-061 + cells and PE-labelled OKT4 + cells).

Evaluation of clinical data from Biotest studies 979 and 985
Binding of Tregalizumab FITC and anti-CD4 (SK3) PE to lymphocytes was All clinical studies were carried out in accordance with the international guidelines on Good Clinical Practice (ICH-GCP) and in compliance with applicable national regulations. It is confirmed that the studies were carried out and documented in accordance with the corresponding study protocol accepted by regulatory authorities and ethic committees. A signed informed consent was obtained from all subjects.

Statistical analysis
All data are described as mean ± s.d. Normal distribution of data was checked using a Kolmogorov-Smirnov test or a Shapiro-Wilk test. Accordingly, the significance was calculated using unpaired Student's t-test or Mann-Whitney test or paired t-test or Wilcoxon matched-pairs signed-rank test.