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

Abstract

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.

Figure 1

Tregalizumab mode of action and predicted CD4 modulation: (a) Tregalizumab (BT-061) binds to the CD4 receptor of T cells and is cross-linked 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 pharmacokinetic–pharmacodynamic 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 2

Tregalizumab binds to CD4 in close proximity to the disulphide bridge in D2. (a) Amino-acid sequence and domain structure of human CD4: The Tregalizumab-binding sites to CD4 are highlighted in red boxes (structure obtained by Proteros). The CD4 intramolecular disulphide bridge (Cys155–Cys184) in D2 is marked in red (yellow circle). (b) Crystal structure of the Tregalizumab–CD4 interaction. The D2 disulphide bond (marked in yellow) is in close proximity to the Tregalizumab-binding sites (Tregalizumab heavy chain in purple, light chain in cyan). Amino-acid tyrosine 105 (Tyr105) is displayed as space fill model and is important for the Tregalizumab–CD4 interaction.

Figure 3

Trx1 diminishes binding of Tregalizumab to recombinant CD4 and CD4-positive cell lines. (a) Recombinant human sCD4 was incubated with 20 μM Trx1 for 30±5 min. After three washing steps, binding of Tregalizumab to rh sCD4 was measured using an electrochemiluminescence enzyme-linked immunosorbent assay. Data represent results from three independent experiments measured in triplicate. Error bars indicate s.d. P-values were calculated by means of Wilcoxon matched-pairs signed-rank test. (b) CD4-positive T-cell leukaemia cells (HPB-ALL) were incubated with 5 μM Trx1, 150 nM thioredoxin reductase and 1 mM NADPH for 15±2 h overnight. Binding of Tregalizumab to CD4 was measured using flow cytometry. The redox inactive thioredoxin1-SS-mutant (Trx1-SS) served as negative control. MFI of BT-061 APC is shown. Data represent results from four independent experiments performed in duplicates. Error bars indicate s.d. P-values were calculated by means of a paired t-test as the data are normally distributed. (c) Impact of the Trx1 system on binding of Tregalizumab to isolated PBMCs expressing CD4. PBMCs were incubated with 5 μM Trx1, 150 nM thioredoxin reductase and 1 mM NADPH overnight. The inactive mutant Trx1-SS served as negative control. Data represent results from five independent experiments performed in duplicates. Error bars indicate s.d. P-values were calculated by means of paired t-test as the data are normally distributed. The gating strategy is furthermore illustrated: lymphocytes were selected according to cell size and granularity (FSC/SSC) and the CD4 binding of Tregalizumab was determined in the CD3⁺ CD4⁺ gated cells. A representative dot plot (lower left) and histogram (lower right) of CD3⁺ CD4⁺ (BT-061⁺) cells is depicted with (blue) and without (red) Trx1/TrxR preincubation. (d) Comparison of Tregalizumab and OKT4 according to their binding to the mutated Cys/Ala D2 CD4 variant (hCD4 mut). MFI values of OKT4 were set to 100% and binding of BT-061 was calculated. Data represent results from two independent experiments performed in triplicate. Error bars indicate s.d. P-values were calculated by means of Wilcoxon matched-pairs signed-rank test. (e) PBMCs were incubated with either 5 μM Trx/150 nM TrxR/1 mM NADPH or 10 mM GSH/1 mM NADPH for 15±2 h overnight. After three washing steps, binding of fluorochrome-labelled anti-CD64 antibody was measured via flow cytometry (n=5, duplicates). Monocytes were gated according to their size and granularity (FSC/SSC). MFI of anti-CD64 PE is shown. Error bars indicate s.d. P-values were calculated by means of paired t-test as the data are normally distributed. Significance is shown as P-value: *P<0.05, **P<0.01, ***P<0.001.

Figure 4

Impact of Trx1/TrxR on T-cell receptor signalling. A total of 10⁵ CD4⁺ cells were isolated from PBMCs and incubated overnight with 150 nM TrxR and 1 mM NADPH and either 5 μM active Trx1-CC or the inactive mutant Trx1-SS. Untreated cells were used as negative control. After the washing procedure, CD4⁺ cells were stimulated with either Tregalizumab (BT-061) or OKT3. Both antibodies were cross-linked with anti-human IgG (BT-061) or anti-murine IgG (OKT3). All data represent results from four independent experiments performed in duplicates. (a) Phosphorylation of lymphocyte-specific protein tyrosine kinase p56^Lck (pY505) was measured after intracellular staining by flow cytometry using anti-lck(pY505) antibody. (b) ZAP-70 phosphorylation was determined after intracellular staining by flow cytometry with anti-ZAP-70 (Y319)/Syk (Y352) antibody. Error bars indicate s.d. P-values were calculated by means of paired t-test as the data are normally distributed. Significance is shown as P-value: **P<0.01.

Figure 5

Tregalizumab binding to CD4 is lower in RA patients compared with healthy subjects. Data from two clinical studies (Biotest RA Study 979, EudraCT: 2010-018485-24; Biotest healthy volunteer Study 985, EudraCT: 2011-004956-20) were analysed retrospectively relating to binding of Tregalizumab to CD4. Binding of Tregalizumab to CD4-positive T cells was measured by flow cytometry using non-competing fluorochrome-labelled anti-CD4 antibodies (BT-061-FITC and SK3-PE). Thirty-five healthy subjects (Biotest study 985) and 127 RA patients (Biotest study 979) were analysed by the same central laboratory using the same assay system. The MFI values of BT-061-FITC and SK3-PE were used to calculate the relative binding of BT-061 in relation to the total CD4 expression (SK3 antibody). (a) Population-related relative binding of Tregalizumab in healthy volunteers (HV) and RA (RA) patients is depicted to show distribution of individual responses. (b) Mean relative binding of BT-061 to CD4 is depicted as a bar graph. Error bars indicate s.d. P-values were calculated by means of unpaired t-test as the data are normally distributed. (c) Mean relative binding of anti-CD4 antibody SK3 to CD4 is shown as a bar graph. Error bars indicate s.d. P-values were calculated by means of Mann–Whitney test. Significance is shown as P-value: *P<0.05, ****P<0.0001.

Figure 6

Trx1 inhibits Tregalizumab-mediated signalling via p56^Lck. Trx1 specifically reduces the CD4 disulphide bond in D2. This leads to a diminished binding of Tregalizumab to the CD4 receptor and therefore to a declined signal transduction via p56^Lck.