Monocytes show immunoregulatory capacity on CD4⁺ T cells in a human in‐vitro model of extracorporeal photopheresis

Summary

Extracorporeal photopheresis (ECP) is a widely used immunomodulatory therapy for the treatment of various T cell‐mediated disorders such as cutaneous T cell lymphoma (CTCL), graft‐versus‐host disease (GvHD) or systemic sclerosis. Although clinical benefits of ECP are already well described, the underlying mechanism of action of ECP is not yet fully understood. Knowledge on the fate of CD14⁺ monocytes in the context of ECP is particularly limited and controversial. Here, we investigated the immunoregulatory function of ECP treated monocytes on T cells in an in‐vitro ECP model. We show that ECP‐treated monocytes significantly induce proinflammatory T cell types in co‐cultured T cells, while anti‐inflammatory T cells remain unaffected. Furthermore, we found significantly reduced proliferation rates of T cells after co‐culture with ECP‐treated monocytes. Both changes in interleukin secretion and proliferation were dependent on cell‐contact between monocytes and T cells. Interestingly, blocking interactions of programmed death ligand 1 (PD‐L1) to programmed death 1 (PD‐1) in the in‐vitro model led to a significant recovery of T cell proliferation. These results set the base for further studies on the mechanism of ECP, especially the regulatory role of ECP‐treated monocytes.

Introduction

Extracorporeal photopheresis (ECP) is a widely used immunomodulatory therapy for the treatment of cutaneous T cell lymphoma 1, 2, 3, Sézary syndrome 4 and various T cell‐mediated disorders, such as graft‐versus‐host disease (GvHD) 5, 6, 7 or systemic sclerosis 8, 9.

During ECP, whole blood leucocytes are separated by apheresis, incubated with a photosensitizer (8‐methoxypsoralen, 8‐MOP) and irradiated with UVA light. Via exposure to UVA light 8‐MOP is activated leading to covalent binding to pyrimidine bases, cell surface molecules and cytoplasmic components 10, 11 and cross‐links in DNA. Upon irradiation, leucocytes are reinfused to the same patient and undergo apoptosis after 24–48 h 12. As there are only 5–10% of circulating peripheral blood mononuclear cells (PBMCs) affected during a single ECP, the removal of reactive T cells alone cannot be responsible for the therapeutic effect in patients 13, 14, 15. Indeed, clinical benefits of ECP have already been well revealed, but the underlying mechanism of action of ECP is not fully understood.

ECP is described to possess an immunoregulatory capacity especially mediated by monocyte‐derived dendritic cells (DCs). Monocytes seem to be more resistant to ECP‐induced cell damage 16, 17 than T cells, natural killer (NK) cells and B cells, and it has been shown that ECP stimulates monocytes turning into immature DCs 18. However, both, a survival of CD14⁺ antigen‐presenting cells (APCs) after ECP 14, 17, 19, as well as an induction of apoptosis in monocytes from GvHD patients 48 h after treatment 20, is reported in the literature. Therefore, knowledge on the role of monocytes in ECP is limited and controversial.

In this study, the immunoregulatory capacity of ECP‐treated monocytes on co‐cultured T cells was analysed. We therefore established an in‐vitro ECP model verified by data from ex‐vivo‐treated patients. This allows investigation of the influence of ECP‐treated monocytes independently from the barriers using patient material which is largely influenced by the underlying diagnosis, pretreatment strategies and accessibility. Specifically addressing the immunoregulatory function of monocytes identified in the present study might lead to further insights into the mode of action of ECP treatment and identification of specific targets for inflammatory diseases usually treated with ECP.

Materials and methods

Donors

Buffy coats from healthy donors were provided by the blood bank, Tuebingen, and included into the study independently from their human leucocyte antigen (HLA) typing. ECP‐treated patients were recruited from the dialysis centre of the University Hospital, Tuebingen. According to the literature, patients with different disorders were included at different time‐points of ECP schedule to investigate the generalized effects of ECP 21. Characteristics of patients are specified in Table 1. All patients gave informed consent and the independent local ethics committee approved analyses (360/2016BO1).

Table 1.

Patient characteristics

Characteristic	Number	%
Sex
Male	5	83·3
Female	1	16·7
Age (years ± s.d.)	50 ± 15
Diagnosis
Transplant rejection	2	33·3
Sézary syndrome	2	33·3
GvHD	2	33·3
Immunosuppressive therapy
Bexaroten	1	16·7
Roferon	1	16·7
Tacrolimus	4	66·7
Prednisolone	4	66·7
Mycophenolate mofetil	2	33·3
Vitamin D3	1	16·7
Azathioprin	1	16·7
Interferon	1	16·7

GvHD = graft‐versus‐host disease; s.d. = standard deviation.

Cell isolation

PBMCs were isolated via Ficoll‐Hypaque (Biochrom, Berlin, Germany) density gradient centrifugation. CD4⁺ T cells and CD14⁺CD16⁺ monocytes were isolated by magnetic cell isolation (MACS) using CD4 microbeads and the Pan Monocyte Isolation Kit, according to the manufacturer’s protocol (Miltenyi Biotec, Bergisch Gladbach, Germany). Isolation purity of CD3⁺CD4⁺ T cells and CD14⁺CD16⁺ monocytes (antibodies from BD Biosciences, Franklin Lakes, NJ, USA) was between 90 and 95%, as determined via flow cytometry.

Extracorporeal photopheresis

In‐vitro ECP was executed by PUVA Combi Light 6000 Series (Dermat BVBA, Leuven, Belgium). Isolated monocytes were seeded in six‐well plates; 200 ng/ml UVADEX^® (Therakos, West Chester, PA, USA) was added for 15 min in the dark and irradiated with 2 J/cm² UVA light (ECP⁺). Untreated cells were seeded in six‐well plates, incubated without UVADEX^® and not irradiated (ECP^–). Treated and untreated monocytes were mechanically detached from cell culture plates.

Ex‐vivo ECP was performed at the dialysis centre of the University Hospital Tuebingen via Cellex system (Therakos). Briefly, patients’ leucocytes were separated from whole blood by apheresis, treated with UVADEX^® and irradiated with UVA light prior to reinfusion to the patient. Monocytes were isolated from samples before UVADEX^® addition (ECP^–) or after UVADEX^® addition and UVA irradiation (ECP⁺).

In‐vitro ECP model

Monocytes isolated from healthy donors were in‐vitro ECP‐treated via PUVA Combi Light 6000 Series with 200 ng/ml UVADEX^® and 2 J/cm² UVA light. For ex‐vivo ECP treatment, the patient’s apheresis product was treated via Therakos CellEx 5.0, according to the manufacturer’s instructions, and monocytes were isolated directly after ECP.

Untreated and ECP treated monocytes of patients or healthy donors were co‐cultured with CD4⁺ T cells of healthy donors at a 1 : 4 monocyte : T cell ratio under stimulation with 100 ng/ml soluble anti‐CD3 Ab (BD Biosciences, San Jose, CA, USA) in RPMI‐1640 (Biochrom, Berlin, Germany) with 10% human serum (Type AB; Invent Diagnostica, Berlin, Germany), penicillin/streptomycin (100 U/ml and 100 μg/ml), 2 mM l‐glutamine (Biochrom, Holliston, MA, USA) for 5 days, as described previously 22. Therefore, the in‐vitro ECP model is an autologous setting, whereas monocytes of patients were co‐cultured with allogeneic T cells of healthy donors.

For analysis of cell‐contact dependency in the in‐vitro ECP model, monocytes were cultured for 24 h exclusively and freshly isolated, autologous CD4⁺ T cells were added at a 1 : 4 ratio with anti‐CD3 antibody (100 ng/ml) for 5 days. Furthermore, freshly isolated CD4⁺ T cells were mixed with supernatant of monocytes after 24 h with stimulation with anti‐CD3 antibody (100 ng/ml) and anti‐CD28 antibody (1 μg/ml; BD Biosciences) for 5 days to analyse the impact of cytokines, as reported earlier 22.

For characterization of monocytes after ECP, cells were cultured in RPMI‐1640 media (Biochrom) with 2% human serum (Type AB, Invent Diagnostica) and 2 mM l‐glutamine (Biochrom).

Flow cytometry

Antibodies for surface staining of T cells, monocytes, macrophages and monocytic myeloid‐derived suppressor cells (MDSCs) were purchased from BD Biosciences [CD3, clone HIT3a, CD4, clone SK3, CD14, clone M5E2, CD16, clone 3G8, CD25, clone 2A3, CD33, clone WM53, CD86, clone 2331 (FUN‐1), CD209, clone MIH18, HLA‐DR, clone G46‐6] and BioLegend (San Diego, CA, USA) [CD15, clone SSEA‐1, programmed death ligand 1 (PD‐L1), clone 29E.2A3, PD‐L2, clone MIH18]. Dead cells were stained via fixable viability dye eFluor 780 (eBioscience, Waltham, MA, USA). For intracellular cytokine staining, cells were stimulated for 5 h with phorbol 12‐myristate 13‐acetate (PMA, 50 ng/ml; Sigma‐Aldrich, Steinheim, Germany) and ionomycin (750 ng/ml; Sigma‐Aldrich). Monensin (2 μM; eBioscience) was present during the last 3 h. Antibodies for intracellular anti‐interferon (IFN)‐γ (clone 4S.B3), anti‐interleukin (IL)‐2 (clone MQ1‐17H12), anti‐IL‐4 (clone 8D4‐8), anti‐IL‐17A (clone eBio64DEC17) and intranuclear anti‐forkhead box protein 3 (FoxP3), clone PCH101] staining were purchased from eBioscience. Staining was performed according to the manufacturer’s protocol. Analyses of different stages of apoptosis were performed using calcium‐dependent detection of fluorescence‐labelled annexin V. Therefore, monocytes were stained with anti‐human annexin V antibody (clone VAA‐33) via binding buffer for annexin V and with fixable viability dye eFluor 780 (all eBioscience). Samples were analysed using LSR II flow cytometer (BD Biosciences) and FlowJo software (Tree Star, Inc., Ashland, OR, USA).

T cell proliferation assay

CD4⁺ T cells labelled with 2 µM carboxyfluorescein‐succinimidyl ester (CFSE) via CellTrace™ CFSE Cell Proliferation Kit (Thermo Fisher, Fremont, CA, USA) were co‐cultured with untreated or ECP‐treated monocytes as described previously. Percentages of proliferating T cells were determined after 5 days of co‐culture via flow cytometry. In addition, monocytes were cultured in different monocyte : T cell ratios (1 : 1; 0·5 : 1; 0·25 : 1; 0·125 : 1; 0·0625 : 1) to examine the influence of reduced monocyte proportions in the in‐vitro ECP model.

Block of PD‐L1 to programmed death 1 (PD‐1) and PD‐L2–PD‐1 interactions was performed via the addition of 25 µg/ml anti‐human PD‐L1 (clone 29E.2A3) or PD‐L2 (clone MIH18) antibody (BioLegend) or Pembrolizumab (Keytruda^®; MSD, Haar, Germany) 23 into co‐cultures for 5 days.

Statistics

Statistical significance analysing two conditions in the experiment was determined using Student’s paired t‐test. If the P‐value was  < 0·05, the difference was considered statistically significant. If the results were not significant, the P‐value was not specified.

Results

ECP‐treated monocytes induce proinflammatory T cell types in an in‐vitro setting of healthy donors

To broaden immunological investigations on the effect of ECP treatment, an in‐vitro ECP model was established and analysed regarding the induction of CD3⁺CD4⁺CD25⁺FoxP3⁺regulatory T cells (T_regs), CD3⁺CD4⁺IFN‐γ⁺ T helper type 1 (Th1) cells, CD3⁺CD4⁺IL‐17A⁺ Th17 cells, CD3⁺CD4⁺IL‐17A⁺IFN‐γ⁺ Th17/Th1 cells, CD3⁺CD4⁺IL‐2⁺ T cells and CD3⁺CD4⁺IL‐4⁺ Th2 cells 5 days after ECP.

No significant changes in percentages of T_regs (n = 8), Th17 cells (n = 27) and Th2 cells (n = 19) in co‐cultures of T cells with autologous, ECP‐treated monocytes were observed compared to co‐cultures with untreated monocytes (Fig. 1). However, a significant increase in percentages of Th1 cells (n = 27) and Th17/Th1 cells (n = 27) and IL‐2⁺ T cells (n = 19) in co‐cultures with ECP‐treated monocytes was determined.

Figure 1.

Percentages of regulatory T cells (T_regs), pro‐ and anti‐inflammatory T cell subsets after co‐culture with ECP‐treated monocytes of healthy donors. (a) In‐vitro ECP‐treated or untreated monocytes of healthy donors were co‐cultured with autologous CD4⁺ T cells with anti‐CD3 antibody (100 ng/ml) at a monocyte : T cell ratio of 1 : 4 for 5 days. Induction of CD3⁺CD4⁺CD25⁺forkhead box protein 3 (FoxP3)⁺ T_regs (n = 8), proinflammatory CD3⁺CD4⁺interferon (IFN)‐γ⁺ T helper type 1 (Th1) cells (n = 27), CD3⁺CD4⁺interleukin (IL)‐17A⁺ Th17 cells (n = 27), CD3⁺CD4⁺IL‐17A⁺IFN‐γ⁺ Th17/Th1 cells (n = 27), CD3⁺CD4⁺IL‐2⁺ T cells (n = 19) and anti‐inflammatory CD3⁺CD4⁺IL4⁺ Th2 cells (n = 18) was assessed by intracellular flow cytometry analysis. (b–f) Necessity of cell‐contact or cytokine environment for changes in T cell subsets was analysed by incubation of ECP untreated or treated monocytes for 24 h and subsequent addition of either monocytes and 100 ng/ml anti‐CD3 antibody (cell‐contact) or supernatant of monocytes, 100 ng/ml anti‐CD3 antibody and 1 μg/ml anti‐CD28 antibody (supernatant) to CD4⁺ T cells (n = 7 or n = 5). Co‐cultures were investigated after 5 days. ECP untreated or treated monocytes directly co‐cultured with anti‐CD3 antibody and CD4⁺ T cells were used as reference (reference). ****P  < 0·0001, **P  < 0·01, *P  < 0·05.

Furthermore, our data show that cell‐contact between ECP treated monocytes and CD4⁺ T cells was necessary for the significant increase of Th1 cells, Th17/Th1 cells (n = 7) and IL‐2⁺ T cells (n = 5) (Fig. 1). In contrast, transferred supernatant containing the cytokine milieu of ECP‐treated, precultured monocytes induced no differences in T cell subsets except for Th1 cells. In this case, transferred supernatant of ECP‐treated monocytes induced a significant decrease in percentages of Th1 cells (Fig. 1b).

T cell proliferation is reduced after co‐culture with in‐vitro ECP‐treated monocytes of healthy donors

For analysis of T cell proliferation rates in the in‐vitro ECP setting, ECP‐treated monocytes of healthy donors were co‐cultured with autologous, CFSE‐labelled T cells.

A significant decrease in proliferation rates of CD3⁺CD4⁺ T cells 5 days after in‐vitro ECP was observed (n = 31) (Fig. 2a,b).

Figure 2.

Influence of in‐vitro ECP‐treated monocytes on autologous T cell proliferation. ECP untreated or treated monocytes were co‐cultured with CD4⁺ T cells from the same donor with anti‐CD3 antibody (100 ng/ml) for 5 days after in‐vitro ECP. (a,b) Proliferation rates of CFSE‐labelled CD3⁺CD4⁺ T cells were assessed via flow cytometry (histogram shows one donor of 31, n = 31). (c) Necessity of cell‐contact or cytokine environment for changes in proliferation was analysed (n = 4). Therefore, ECP untreated or treated monocytes were cultured for 24 h exclusively. In cell‐contact‐dependent approaches freshly isolated CD4⁺ T cells were cultured with precultured monocytes at a monocyte : T cell ratio of 1 : 4 in the presence of 100 ng/ml anti‐CD3 antibody for 5 days (cell‐contact). For cytokine‐mediated approaches supernatant of precultured monocytes were added to freshly isolated CD4⁺ T cells with 100 ng/ml anti‐CD3 antibody and 1 μg/ml anti‐CD28 antibody (supernatant). Approaches where ECP untreated or treated monocytes were directly co‐cultured with 100 ng/ml soluble anti‐CD3 antibody and CD4⁺ T cells were used as reference (reference). Co‐cultures were investigated after 5 days. ****P  < 0·0001, **P  < 0·01.

Further investigation revealed that cell‐contact between CD4⁺ T cells and treated, precultured monocytes 24 h after ECP could reduce T cell proliferation significantly. Transferring supernatant of ECP‐treated, precultured monocytes did not decrease proliferation of CD3⁺CD4⁺ T cells and led to a reduced proliferation capacity even in the untreated approach (n = 4) (Fig. 2c).

Alterations in T cell subsets and T cell proliferation in in‐vitro ECP‐treated cells of healthy donors are comparable to an ex‐vivo and in‐vitro setting of patient samples

To confirm the established in‐vitro ECP model, data on induction of previously described T cell types and T cell proliferation after co‐culture with in‐vitro‐treated monocytes of healthy donors were compared to samples of six patients before and after ECP (ex vivo) as well as in an in‐vitro ECP setting with allogeneic T cells of healthy donors.

In ex‐vivo and in‐vitro‐treated patient samples, an increase in Th1 cells, Th17/Th1 cells (n = 6) and IL‐2⁺ T cells was observed (n = 5) (Fig. 3a,b). The induction of Th17/Th1 cells was, in fact, significant in the in‐vitro setting. No significant changes in percentages of T_regs (n = 4), Th17 cells (n = 6) and Th2 cells (n = 5) in co‐cultures of T cells with ECP‐treated monocytes were observed compared to co‐cultures with untreated monocytes (Fig. 3). Furthermore, using CFSE‐based proliferation assays, a decrease in percentages of proliferating T cells was determined after co‐culture with ex‐vivo and in‐vitro ECP‐treated monocytes (n = 3) (Fig. 3c).

Figure 3.

Comparison of T cell subsets and proliferation rates after ex‐vivo and in‐vitro ECP of patient samples. The samples were collected from the apheresis bag during ECP procedure before addition of UVADEX^® (ECP^–, ex vivo) and after ultraviolet A (UVA) irradiation (ECP⁺, ex vivo). Monocytes from patients were isolated, co‐cultured with freshly isolated CD4⁺ T cells from healthy donors at a monocyte : T cell ratio of 1 : 4 with 100 ng/ml anti‐CD3 antibody and analysed 5 days after ECP. Furthermore, monocytes from patient samples before addition of UVADEX^® were isolated and treated with the ECP in‐vitro model (ECP^– and ECP⁺, in vitro). Afterwards, monocytes were co‐cultured with CD4⁺ T cells from healthy donors and anti‐CD3 antibody (100 ng/ml) and investigated after 5 days. Induction of CD3⁺CD4⁺CD25⁺forkhead box protein 3 (FoxP3)⁺ regulatory T cells (T_regs) (n = 4*), proinflammatory CD3⁺CD4⁺interferon (IFN)‐γ⁺ T helper type 1 (Th1) cells (n = 6*), CD3⁺CD4⁺interleukin (IL)‐17A⁺ Th17 cells (n = 6*), CD3⁺CD4⁺IL‐17A⁺IFN‐γ⁺ Th17/Th1 cells (n = 6*), CD3⁺CD4⁺IL‐2⁺ T cells (n = 5*) and anti‐inflammatory CD3⁺CD4⁺IL4⁺ Th2 cells (n = 5*) in approaches with (a) ex‐vivo‐ and (b) in‐vitro‐treated monocytes was assessed by intracellular flow cytometry analysis. (c) Determination of T cell proliferation rates after co‐culture with ex‐vivo‐ and in‐vitro‐treated monocytes via CFSE labelling and flow cytometry analysis (n = 3). *One of 4 or 6 data points after ex‐vivo ECP treatment is missing due to coagulation of sample material. **P  < 0·01, *P  < 0·05.

In‐vitro ECP impairs monocyte survival and expression patterns of surface molecules

In order to investigate the influence of ECP treatment of monocytes on the proliferative capacity of co‐cultured T cells, monocyte survival and expression of several surface markers were investigated.

In‐vitro ECP of monocytes significantly decreased the absolute count of cells in monocyte cultures 2 days after ECP treatment (n = 9) (Fig. 4a) referred to cell count originally seeded after ECP at day 0. However, culturing monocytes for 2 days even without ECP led to a significant cell reduction. Monocytes treated with in‐vitro ECP do not accumulate increasingly in early or late apoptotic stage (n = 6) (Fig. 4b) compared to untreated monocytes, and proportions of living monocytes can be assigned to more than 90% of all cells.

Figure 4.

Characterization of in‐vitro ECP‐treated monocytes. ECP untreated or treated monocytes were cultured after in‐vitro ECP for 2 days and (a) absolute cell counts referred to originally seeded cell numbers directly after ECP treatment (day 0) were determined (n = 9). (b) Determination of proportions of living, early apoptotic, late apoptotic and dead cells was performed by calcium‐dependent staining of ECP untreated or treated monocytes with annexin V after 2 days and successive flow cytometry analysis (n = 6). (c) Moreover, ECP untreated or treated monocytes were co‐cultured with CFSE‐labelled CD4⁺ T cells from the same donor in different monocyte : T cell ratios with anti‐CD3 antibody (100 ng/ml) for 5 days after in‐vitro ECP and proliferation rates of CD3⁺CD4⁺ T cells were assessed via flow cytometry (n = 3–5). Induced percentages of (d) CD14⁺CD209‐human leucocyte antigen D‐related (HLA‐DR)⁺CD86⁺ monocytes and (e) CD14⁺CD209⁺HLA‐DR⁺CD86⁺ macrophages (n = 15) as well as (f) HLA‐DR‐CD33⁺CD14⁺monocytic myeloid derived suppressor cells (M‐MDSCs) (n = 16) were assessed via flow cytometry analysis 2 days after in‐vitro ECP. ****P  < 0·0001, **P  < 0·01, *P  < 0·05.

Titration of monocyte : T cell ratio from 1 : 1 to 0·0625 : 1 in CFSE‐based proliferation assays showed a significant reduction of T cell proliferation after co‐culture with ECP‐treated monocytes (Fig. 4c, except for the 0·0625 : 1 ratio) compared to untreated approaches. The proliferation rates were similar up to a titration rate of 0·25 : 1 (monocytes : T cells) in both set‐ups. A non‐significant decline of T cell proliferation was observed with lower ratios starting at 0·125 : 1 (monocytes : T cells) in the untreated and ECP‐treated groups.

Furthermore, the expression of co‐stimulatory antigens was evaluated in the untreated and ECP‐treated monocytes after 2 days in culture. Monocytes after in‐vitro ECP treatment showed significantly higher proportions of CD14⁺CD209^–HLA‐DR⁺CD86⁺ cells (n = 15) (Fig. 4e) and CD14⁺CD209⁺HLA‐DR⁺CD86⁺ macrophages (n = 15) (Fig. 4f). In contrast, HLA‐DR^–CD33⁺CD14⁺ monocytic myeloid‐derived suppressor cells (M‐MDSCs) (n = 16) were reduced after ECP treatment of monocytes (Fig. 4b).

Blockade of PD‐L1–PD‐1 interaction ameliorates T cell proliferation in the in‐vitro ECP model

Co‐stimulatory surface molecules known from the literature to be involved in cell‐contact between antigen‐presenting cells (APCs) and T cells and connected to suppression of T cell proliferation were investigated at different time‐points of culture. Therefore, untreated or in‐vitro ECP‐treated monocytes were either cultured alone or co‐cultured with autologous CD4⁺ T cells.

Proportions of CD14⁺PD‐L1⁺ monocytes showed a slight increase 16 h after in‐vitro ECP, whereas 2 days after ECP the percentages of CD14⁺PD‐L1⁺ monocytes was higher in untreated approaches of cultured monocytes (n = 4) (Fig. 5a). No CD14⁺PD‐L2⁺ monocytes could be detected at all time‐points (n = 4) (Fig. 5).

Figure 5.

Investigation of the connection between programmed cell death ligand 1 (PD‐L1) and PD‐L2 and T cell proliferation. ECP untreated or treated monocytes were cultured for 2 days and proportions of living (a) CD14⁺PD‐L1⁺ and (b) CD14⁺PD‐L2⁺ monocytes were determined via flow cytometry. Furthermore, ECP untreated or treated monocytes were co‐cultured with CD4⁺ T cells from the same donor with anti‐CD3 antibody (100 ng/ml) for 5 days after in‐vitro ECP and proportions of living (c) CD3⁺CD4⁺PD‐L1⁺ and (d) CD3⁺CD4⁺PD‐2⁺ monocytes were determined via flow cytometry. (e) To block PD‐L1‐PD‐1 and PD‐L2‐PD‐1 interactions 25 µg/ml of pembrolizumab, anti‐PD‐L1 or anti‐PD‐L2 antibody was added to co‐cultures of either untreated or ECP‐treated monocytes with CFSE‐labelled CD4⁺ T cells from the same donor with anti‐CD3 antibody (100 ng/ml) stimulation for 5 days after in‐vitro ECP. Proliferation rates of CD3⁺CD4⁺ T cells were assessed via flow cytometry. **P  < 0·01, *P  < 0·05.

In contrast, investigation of PD‐L1⁺ and PD‐L2⁺ T cells in co‐cultures with autologous, ECP‐treated monocytes displayed lower proportions of CD3⁺CD4⁺PD‐L1⁺ T cells compared to untreated approaches 16 h after ECP of monocytes (n = 4) (Fig. 5. No difference between the percentage of CD3⁺CD4⁺PD‐L1⁺ T cells in both co‐cultures with untreated or ECP‐treated monocytes could be observed 1 and 2 days after ECP. Five days after ECP no CD3⁺CD4⁺PD‐L1⁺ T cells could be detected. Proportions of CD3⁺CD4⁺PD‐L2⁺ T cells were below 4% of all living cells of the co‐cultures and showed a slight increase 1 and 2 days after co‐culturing with ECP‐treated monocytes (n = 4) (Fig. 5). In general, levels of PD‐L1⁺ T cells in co‐cultures were lower than levels of PD‐L1⁺ monocytes in monocultures.

Furthermore, PD‐L1–PD‐1 and PD‐L2–PD‐1 interactions were blocked via addition of anti‐PD‐1 (pembrolizumab), anti‐PD‐L1 or anti‐PD‐L2 antibody to co‐cultures of untreated or ECP‐treated monocytes and CD4⁺ T cells. We found a significant increase in T cell proliferation in co‐cultures with ECP‐treated monocytes after addition of pembrolizumab (n = 6) (Fig. 5) compared to unblocked approaches. Also, blockade with anti‐PD‐L1 antibody displayed a significant increase of proliferation rates in co‐cultures with ECP‐treated and untreated monocytes (n = 6). However, increase in T cell proliferation after pembrolizumab and anti‐PD‐L1 addition was always higher in ECP‐treated approaches compared to untreated approaches. Addition of anti‐PD‐L2 antibody did not change the proliferative capacity of T cells in co‐cultures (n = 6).

Discussion

Monocytes have been assigned as the most important trigger for the clinical benefit of ECP 24, which might be a deceptive statement. It is known that ECP stimulates monocytes to differentiate into DCs 18, which especially lead to an immunoregulatory function of ECP. Despite the importance of DCs in ECP, knowledge of the impact of ECP on their parental cells (monocytes) is limited and conflicting. Hence, we established an in‐vitro ECP model to further investigate the mechanism of this immunomodulating therapy by analysing the immunoregulatory function of monocytes on co‐cultured T cells with regard to the induction of intracellular cytokine production, proliferation and restoration of proliferation rates.

In‐vitro ECP‐treated monocytes induced a significant increase in proinflammatory Th1 cells, Th17/Th1 cells and IL‐2⁺ T cells. Because of the low level of Th17/Th1 cells in co‐cultures of approximately 1·3% in‐vitro ECP leads to a significant increase to 2·4%, which is almost a doubling of proportions of Th17/Th1 cells. Induction of anti‐inflammatory Th2 cells was not affected, which seemed to conflict with the anti‐inflammatory effects of ECP therapy described in the literature 25, 26. Recently, an ECP‐induced shift from inflammatory Th1 and Th17 cells 27, 28 to an anti‐inflammatory Th2 response 25, 26 and cytokine profile 25, 29, 30 was described in GvHD. In contrast, restoration of the Th1/Th2 imbalance in CTCL patients towards induction of proinflammatory Th1 cells 31 was found, which is in line with the results of the underlying in‐vitro ECP model. Human Th17 cells infiltrate tumours impairing tumour proliferation and angiogenesis in the case of head and neck squamous cell carcinoma (HNSCC) patients 32. Furthermore, ECP of hepatitis C virus‐positive patients after liver transplantation seems to reduce their immunosuppressive burden, leading to enhanced graft survival and efficacy of anti‐viral treatment 33. Th17/Th1 cells may account for this phenomenon and thereby our data underline a curative benefit for ECP patients, as induction of Th1 cells, Th17 cells and Th17/Th1 cells may indicate a remaining anti‐tumour and anti‐viral effect after ECP.

Induction of proinflammatory Th1 and Th17/Th1 cells in our in‐vitro model strongly depends upon cell‐contact between ECP‐treated monocytes and T cells. The importance of cell‐to‐cell signals between monocytes and T cells has already been described for stimulating Th17 responses 22, 34, 35.

Next, we showed a significant reduction of T cell proliferation in co‐cultures with ECP‐treated monocytes, which is also dependent upon cell‐contact rather than on cytokine environment. Earlier data have already demonstrated that T cell proliferation was reduced in both ECP‐treated PBMCs 36 and in co‐cultures with myeloid cell types. It was shown that ECP‐treated neutrophilic MDSCs suppress T cell proliferation 37. Furthermore, DCs isolated from PBMCs of photopheresis products of refractory cGvHD patients could not induce T cell proliferation 29. Consequently, our data confirm the inhibitory effect of ECP on T cell proliferation, indicating an immunosuppressive potential by limiting the amount of possibly aberrant T cells in patients.

By verification of our in‐vitro ECP model with samples from ECP patients, we found that in‐vitro ECP of monocytes of healthy donors or patients provides similar results concerning alteration in T cell subsets and T cell proliferation compared to ex‐vivo‐treated patient samples. To further confirm these data, more patient samples should be investigated; however, patient availability is limited and the patient characteristics are diverse. Therefore, our in‐vitro model enables the investigation of immunological questions concerning the mode of action of ECP without the need of samples from heavily pretreated patients.

To clarify the influence of ECP‐treated monocytes on co‐cultured T cells regarding changes in T cell subsets and T cell proliferation, we analysed monocytes 2 days after ECP. After 2 days in culture a reduction of cell count was observed to approximately 37% of the initially seeded cells in the treated monocyte fraction after ECP and to approximately 58% in the untreated monocyte group. Therefore, ECP impairs survival of monocytes, but an effect of early or late apoptotic monocytes in co‐cultures could be excluded, as none of these stages could be detected in our in‐vitro model; more than 90% of untreated or ECP‐treated monocytes were alive after 2 days of culture.

However, the loss of monocytes and hence co‐stimulatory signals by ECP cannot be held solely responsible for the observed decrease in T cell proliferation compared to the untreated approach, as comparing proliferation rates in titration experiments revealed a stable proliferation of T cells up to a titration rate of 0·25 : 1 (monocytes : T cells) in both set‐ups. Furthermore, we found a significant induction of CD14⁺CD209^–HLA‐DR⁺CD86⁺ monocytes and CD14⁺CD209⁺HLA‐DR⁺CD86⁺ macrophages by ECP treatment. No difference in CD14^–CD209⁺HLA‐DR⁺CD86⁺ mDCs or CD14^–CD209⁺HLA‐DR⁺CD86^– induced DCs (iDCs) (data not shown) and a strong but not significant reduction of HLA‐DR^–CD33⁺CD14⁺ M‐MDSCs 2 days after in‐vitro ECP treatment was observed.

To further investigate the reduction of T cell proliferation by ECP‐treated monocytes, we examined possible check‐points of immunomodulation. PD‐L1 is known to suppress T cell activity and to reduce tumour cell killing via binding of its receptor PD‐1 38, 39. PD‐L1–PD‐1 interactions further lead to a cell cycle arrest in the G₀/G₁ phase, whereas binding of PD‐L2 to PD‐1 leads to the inhibition of T cell receptor‐mediated proliferation 40. Indeed, PD‐L1 is not only described to be an inhibitor for T cell proliferation but also for Th17 differentiation 38, and PD‐L1 signalling was assigned to play an important role in down‐regulation of immune responses in an ECP model 41 of iDCs and ECP‐treated responder cells.

Interestingly, we found that ECP already slightly induced CD14⁺PD‐L1⁺ monocytes 16 h after treatment which was reversed 2 days later, whereas no CD14⁺PD‐L2⁺ monocytes were detected. Furthermore, PD‐L1 42, 43 and PD‐1 44, 45 can be up‐regulated in activated T cells, suggesting the PD‐L1–PD‐1 axis as a possible pathway in immunoregulation by ECP. We did not observe differences in CD3⁺CD4⁺PD‐L1⁺ T cells at any time‐point. Also, percentages of CD3⁺CD4⁺PD‐L2⁺ T cells were very low, with 4% of all living cells at maximum.

Blockade of PD‐L1–PD‐1 interactions via pembrolizumab and anti‐PD‐L1 antibody showed a significant increase in T cell proliferation in co‐cultures with ECP‐treated monocytes compared to unblocked approaches. Altogether, our in‐vitro model underscores earlier findings, indicating that down‐regulation of immune responses was ascribed mainly to PD‐L1 signalling 41 and therefore verifies the importance of PD‐L1–PD‐1 interactions for an immunosuppressive effect of ECP. However, T cell proliferation via blocking antibodies in our in‐vitro model was not fully restored, indicating a multi‐factorial mechanism in ECP treatment.

In conclusion, we found that in‐vitro ECP treatment of monocytes (1) significantly induced proinflammatory T cell types in co‐cultured T cells dependent on cell‐contact, whereas anti‐inflammatory T cells were unaffected, (2) significantly reduced proliferation rates of co‐cultured T cells in a cell‐contact‐dependent manner and (3) blockade of PD‐L1–PD‐1 interactions via antibodies led to a slight but significant recovery of T cell proliferation. In this study, we focused on the treatment of monocytes, as knowledge from the literature on the role of monocytes in ECP is controversial, and should be further investigated. Nevertheless, the absence of ECP treated lymphocytes in the in‐vitro model might influence the results described in the underlying study and explain varying results published with regard to the anti‐inflammatory effects of ECP. However, besides the impact of ECP on monocytes and co‐cultured CD4⁺ T cells, the in‐vitro model can be further applied to investigate the effect of ECP on other cell subsets, and can be expanded by the addition of lymphocytes and other cell subsets to the treated cell fraction. Further clarification of the mechanism of ECP treatment could help to find potential targets for treatment of several inflammatory diseases. Furthermore, the underlying in‐vitro ECP model may enable the broadening of immunological investigations independently of patient samples comprising possible bias by diversity in diagnosis and treatment.

Disclosures

There are no conflicts of interest.

Author contributions

F. W. designed and performed experiments, interpreted data and wrote the paper, K. R. contributed to data interpretation and experiment design, M. V. performed experiments, C. G. and N. K. contributed to data interpretation and expert advice to MDSCs, H. B. helped with patient recruitment and H. R. critically revised the manuscript, U. H. conceptualized the work, designed experiments, interpreted data and contributed to manuscript writing.