PD-L1 is expressed on human activated naive effector CD4+ T cells. Regulation by dendritic cells and regulatory CD4+ T cells

The T cell expression of various co-signalling receptors from the CD28 immunoglobulin superfamily (Inducible T cell co-stimulator (ICOS), Programmed cell death 1(PD-1), cytotoxic T lymphocyte associated protein 4 (CTLA-4), B and T lymphocyte attenuator (BTLA) or from the tumour necrosis factor receptor superfamily (glucocorticoid-induced TNFR family related (GITR), 4-1BB, and CD27), is essential for T cell responses regulation. Other receptors (such as T cell immunoglobulin and mucin domain-containing protein 3, T cell immunoglobulin and T cell immunoglobulin and ITIM domain (TIGIT), and lymphocyte activation gene 3) are also involved in this regulation. Disturbance of the balance between activating and inhibitory signals can induce autoimmunity. We have developed an in vitro assay to simultaneously assess the function of naive CD4+ effector T cells (TEFFs), dendritic cells (DCs) and regulatory T cells (TREGs) and the expression of co-signalling receptors. By running the assay on cells from healthy adult, we investigated the regulation of activated T cell proliferation and phenotypes. We observed that TEFFs activated by DCs mainly expressed BTLA, ICOS and PD-1, whereas activated TREGs mainly expressed TIGIT, ICOS, and CD27. Strikingly, we observed that programmed death-ligand 1 (PD-L1) was significantly expressed on both activated TEFFs and TREGs. Moreover, high PD-L1 expression on activated TEFFs was correlated with a higher index of proliferation. Lastly, and in parallel to the TREG-mediated suppression of TEFF proliferation, we observed the specific modulation of the surface expression of PD-L1 (but not other markers) on activated TEFFs. Our results suggest that the regulation of T cell proliferation is correlated with the specific expression of PD-L1 on activated TEFFs.


Introduction
A large number of co-signalling molecules are involved in the production of co-stimulatory or co-inhibitory signals in the regulation of T cell activation. The expression of co-signalling molecules on effector T cells (TEFFs) initiates T cell responses, and the expression on regulatory T Funding: The study was funded by the Institut National de la Santé et de la Recherche Médicale (INSERM), a grant from the French government (managed by French National Research Agency (Agence National de la Recherche) as part of the "Investment for the Future" program (Institut Hospitalo-Universitaire Imagine, grant ANR-10-IAHU-01, Recherche Hospitalo-Universitaire, grant ANR-18-RHUS-0010)), and other grants from the Agence National de la Recherche (ANR-14-CE14-0026-01 "Lumugene"; ANR-18-CE17-0001 "Action"), the Fondation pour la Recherche Médicale (Equipe FRM EQU202103012670), the Ligue Contre le Cancer -Comité de Paris, Fondation ARC pour la recherche sur le CANCER, and the Centre de Référence Déficits Immunitaires Héréditaires (CEREDIH).

Competing interests:
The authors have declared that no competing interests exist. immunoglobulin and mucin domain-containing protein 3 (TIM3) and lymphocyte activation gene 3 (LAG3) [30]. TIM3 (a member of the TIM family of immune regulatory receptors) is expressed predominantly on Th1 cells but also on Th17 cells. It appears to be a co-inhibitory molecule for T cell function [28] but is weakly expressed by conventional CD4 + T cells and by TREGs in the peripheral blood [31]. LAG3 (CD223) bears structural homology to CD4 and is upregulated on activated CD4 + and CD8 + T cells [32]. It is also expressed by activated natural TREGs and induced CD4 + FoxP3 + TREGs, where its expression levels are higher than those observed on activated CD4 + TEFFs [33].
In the present study, we used an in vitro cell activation assay (developed previously for studies of autoimmune diseases) to investigate the surface expression of these receptors on TEFFs from healthy controls; the TEFFs were activated by a combination of DCs and staphylococcal enterotoxin E (SEE) and were regulated by TREGs [34,35]. Our results show that a change in the surface receptor profile on TEFFs is correlated with the regulation of the TEFF proliferation by TREGs.

Cell isolation
Purified peripheral blood mononuclear cells (PBMCs) from 12 healthy adult controls (8 females and 4 males; mean ± standard deviation age: 34 ± 15 (range: 18-58); French Blood Transfusion Service, Paris, France) were prepared by density gradient centrifugation on Lymphoprep (Abcyss SA) and tested in independent experiments. The donors' TEFFs, TREGs and DCs were sorted using flow cytometry, as described elsewhere [34]. All three cell preparations were at least 90% pure.

Purification of T cells from PBMCs and co-culture with sorted DCs
PBMCs were incubated for 30 min at 4˚C with specific, labelled monoclonal antibodies, washed and then sorted using a cytometer (ARIA II, BD Biosciences). Naive TEFFs were defined as CD3 + CD4 high CD25 low CD127 high CD45RA + T cells, TREGs were defined as CD3 + CD4 high CD25 + CD127 low T cells, and DCs were defined as CD3 -CD4 low CD11c + CD14 -CD16cells. To assess cell proliferation, sorted naive TEFFs and TREGs were washed and then stained with CTV and CFSE, respectively. Next, the cells were washed and incubated with DCs (8x10 3 TEFFs and 2x10 3 DCs, giving a ratio of 4:1). TREGs were added to give a TEFF:TREG ratio of 1: 0.5 or 1: 0.125. SEE (0.2 ng/ml) was then added. The cell proliferation assay was performed with autologous and heterologous samples in Panserin medium (Dutscher, Brumath, France) supplemented with 5% human AB serum in 96-well plates. After 4, 5 or 7 days of culture, the various conjugated antibodies were added to the culture for 20 min in the dark at room temperature. The percentage and phenotype of the two proliferating T cell subpopulations were measured with a MACSquant system (Miltenyi), and the data were analyzed with FlowJo software. The phenotyping results were expressed as the geometric mean and normalized against the mode. Statistical analyses were performed using GraphPad Prism software (version 6, GraphPad Software, Inc., La Jolla, CA). A nonparametric Mann-Whitney (M-W) test was used to compare the data for the various markers.
All study participants provided written informed consent. The study was performed in accordance with the 1975 Declaration of Helsinki and subsequent revisions, and was approved by the local institutional review board (CPP Ile de France II Paris, France), the French Advisory Committee on Data Processing in Medical Research (Paris, France).

PD-L1 expression on human naive CD4 + TEFFs co-cultured with DCs
We first studied the phenotype of naive human CD4 + TEFFs at different time points after activation with DCs and SEE in a previously described in vitro assay [34]. The highest TEFF proliferation index was observed on day 5 (D5) (Fig 1A). We found that this time point corresponds to a peak in the expression of the various different receptors (such as PD-1 and ICOS) known to have a regulatory role in T cell activation ( Fig 1B). Surprisingly, we also observed significant PD-L1 expression on activated TEFFs on D5 ( Fig 1B). Furthermore, a lower proliferation index on D7 was correlated with significantly lower PD-L1 expression ( Fig 1B). In contrast, the other PD-1 ligand (PD-L2) was weakly expressed (S1 Fig). We also observed that on TEFFs, (i) BTLA was strongly expressed, (ii) TIM3, 4-1BB, CTLA-4 and LAG3 were moderately expressed, and (iii) GITR was not expressed (S1 Fig). CTLA-4 expression was most easily detected on D5 (S2A Fig). By testing cells from 10 controls several times, we observed that the activated TEFFs' expression of PD-1, ICOS and PD-L1 varied from one individual to another ( Fig 1C). This finding indicates that human peripheral blood naïve CD4 + T cells stimulated with DCs and SEE express PD-L1. Moreover, the highest level of PD-L1 expression was observed on D5 and was thus correlated with the highest proliferation index.

PD-L1 expression on activated naive TEFFs is correlated with DC efficacy
We found two groups of control cells with regard to the TEFF proliferation index on D5 ( Fig  2A): one group had a high proliferation index, and the other group had a lower proliferation index. Interestingly, a comparison of PD-L1 expression in the two groups showed that PD-L1 was most strongly expressed on activated TEFFs with the highest proliferation index (Fig 2B). In contrast the PD-1 and ICOS expressions were similar in both groups of samples ( Fig 2B). This result suggests that the level of PD-L1 expression on TEFFs is linked to the proliferative response induced by DC and SEE.

PD-L1 expression on activated naive TEFFs is regulated by TREGs
We reported previously that TEFF proliferation can be significantly inhibited by TREGs when the TEFF:TREG ratio is 1:0.5 [34]. Here, we confirmed that TREGs can even inhibit TEFFs with a high proliferative index (Fig 3A and S3A Fig). We therefore looked at whether the   PD-L1 expression on activated TEFFs was affected by co-culture with TREGs. Indeed, expression levels of PD-L1 ( Fig 3B) and CD27 (S3B Fig) on TEFFs were significantly lower in the presence of TREGs. Expression levels of PD-1 and ICOS were slightly but not significantly lower (Fig 3B). In contrast, the expression levels of PD-L2, BTLA, LAG3 TIM3 and CTLA4 did not depend on the presence or absence of TREGs (S2B and S3B Figs). At a high TEFF: TREG ratio (1:0.125), there was no inhibition of proliferation (Fig 4A). Under these conditions, the PD-L1 expression on the TEFFs was similar to that observed in the absence of TREGs (Fig 4B). However, the significant inhibition of TEFF proliferation (Fig 4A) observed with a TEFF:TREG ratio of 1:0.5 was correlated with significantly lower PD-L1 expression on TEFFs (Fig 4B). This result confirms the link between PD-L1 expression on activated TEFFs and their proliferation index.

Expression of co-signalling receptors on activated TREGs
By using distinct cell tracers, we could simultaneously assess the expression of receptors on CFSE-stained TREGs and CTV-stained TEFFs co-cultured with DCs and SEE on D4, D5 and D7 (Fig 5). We observed that the TREG-induced inhibition of proliferation was greatest on D5 of the co-culture (Fig 5A and S5A Fig), which was again the best time to detect PD-L1 expression on TREGs (Fig 5B). Concerning the expression of other receptors on D5, TREGs expressed PD-1, ICOS, and CTLA-4 strongly and PD-L2, BTLA, LAG3 and TIM3 weakly at all timepoints (Fig 5B and S4 Fig). As described for PD-L1 (Fig 5B), we observed that CTL-4 expression decreased significantly on D7 of co-culture (S2C Fig). To confirm the absence of possible contamination between gated TEFFs and TREGs, we also assessed TIGIT expression because it has been reported previously that this protein is mainly expressed on suppressor TREGs [36]. Marked TIGIT expression by TREGs was observed on D5 of co-culture (S4 Fig). In contrast, TIGIT expression was never detected on TEFFs co-cultured with DCs in the absence or in the presence of TREGs-showing that the absence of TIGIT expression on TEFFs was not linked to inhibition by TREGs ( S5B Fig). Overall, these results showed for the first time that PD-L1 can also be expressed on human peripheral CD4 + TREGs co-cultured with CD4 + TEFFs activated by DCs and SEE. This observation suggests that the three types of cell interact in novel ways to regulate proliferation.

Discussion
In the present study, we assessed the surface expression of several receptors known to be involved in the regulation of T cell activation. We observed that upon activation with DCs and SEE, naive CD4 + TEFFs expressed several receptors, such as BTLA, LAG3, 4-1BB and TIM3. In contrast, the TEFFs did not express GITR-suggesting that the expression of this receptor requires longer activation. The expression of PD-1 and ICOS-two key receptors involved in the regulation of T cell activation-was also detected on TEFFs on D5 of co-culture. Interestingly, we observed significant expression of one of PD-1's ligands (PD-L1) on TEFFs under these conditions. The receptor PD-1 is reportedly expressed on antigen presenting cells, such as macrophages and DCs. Our results are consistent with a report on a mouse model in which TCR activation increased the surface expression of both PD-L1 and PD-1 on T cells from the spleen and lymph nodes [37]. In contrast, we observed that PD-L2 was not detected on activated TEFFs. Unfortunately, we could not compare the PD-L1 expression on T cells vs. DCs because the latter died after three days of co-culture (data not shown). PD-L1 expression on T cells was correlated strongly with their proliferative capacity. This might result from a DCmediated signal, through either direct contact between T cells and DCs during the first three We also observed that PD-L1 expression (but not ICOS and PD-1 expression) was correlated with a higher proliferative index in TEFFs. By performing several assays on the same samples, we observed two groups of controls (one with a high proliferative index on D5 and one with a low index D5) and thus highlighted interindividual variability. Strikingly, the TEFFs with the highest proliferative index expressed PD-L1 most strongly. In contrast, PD-1 and ICOS expression levels were not correlated with the TEFFs' proliferation index.
PD-L1 surface expression on activated CD45RA+ TEFFs was unexpected although already described on purified CD3+CD4+ T cells only after activation with anti-CD3/CD28/IgGcoated beads but not with PD-L1-Ig [38]. In addition, PD-1 expression was well induced on these purified T cells activated in both conditions. However, as mentioned above, PD-L1 mRNA has been detected on activated naive CD4 + or CD8 + T cells but not on non-activated T cell populations (https://dice-database.org). Our present results are in line with these findings. Furthermore, we showed previously that TREGs inhibit SEE/DC-induced T cell proliferation [34]. In mechanistic terms, this inhibition might involve three-way contact between TREGs, TEFFs, and targeting DCs via interactions with inhibitory receptors [39]. We therefore investigated the regulation of receptor expression on activated T cells in the presence of TREGs. We observed lower expression of PD-L1 and CD27 on TEFFs co-cultured with DCs and TREGs when the TEFF:TREG ratio efficiently inhibited TEFF proliferation. Expression levels of the other receptors were not affected by the presence of TREGs.
Importantly, we also observed PD-L1 expression on the TREGs in the co-culture. Significant PD-1 and PD-L1 expression on DC-activated TREGs (suggesting that the PD1/PD-L1 axis modulates both cells) has been reported previously [40]. Amarnath et al. suggested that PD-L1-expressing TREGs directly increased the expression of PD-L1 on DCs. We could not study the impact of PD-L1-expressing TREGs on DCs since the latter died after three days of culture, at which time PD-L1 was not expressed on the TREGs. This prevented us from determining whether the PD-L1 expression on TREGs was induced by DCs and/or activated TEFFs.
With regard to other receptors, we observed that ICOS, PD-1 and CD27 are expressed at the same time on TREGs. In contrast, PD-L2, BTLA, LAG3, TIM3 and GITR (data not shown: similar to TIM3) are poorly expressed on TREGs. Interestingly, other researchers have reported that GITR is expressed on TREGs [39] but this discrepancy might be due to a difference in activation (SEE in our study and transforming growth factor beta in the study by Lohr et al.). We also assessed the expression of CTLA-4, a key inhibitor involved in the TREGs' suppressor function. We found that activated TEFFs expressed CTLA-4 and that (in contrast to PD-L1 expression) this expression did not decrease in the presence of TREGs. Lastly, we observed CTLA-4 expression on activated TREGs, as reported previously for various disease and tumour models [41,42].
Taken as a whole, our results show that naive CD4 + T cell proliferation stimulated by DCs/ SEE and regulated by TREGs is correlated with a specific change in PD-L1 surface expression. The PD-L1 expression on TEFFs and TREGs suggests a complex interaction possibly involved in the regulation of T cells. The fact that all three cell partners express PD-L1 suggests cis-and/ or trans-interactions with either the classical PD-L1 receptor (PD-1) or another PD-L1 receptor with different affinity/avidity. We did not seek to determine the consequences of these mean fluorescence intensity on CTV-stained TEFFs in 6 to 8 independent experiments. Bars represent the means of all experiments for each proteins. The isotype control (Iso) is shown for each protein (solid grey lines). � p�0.05 in an M-W test.
https://doi.org/10.1371/journal.pone.0260206.g004 interactions, which might serve to limit TCR signalling after activation; indeed, the inhibition of proliferation was correlated with a decrease in PD-L1 expression on TEFFs. Therefore, PD-L1 expression on TEFFs might be an interesting biomarker of autoimmune disease; it might help physicians to identify and thus target the patient's most strongly proliferative T cells.