PD-1/PD-L1 expression on CD⁴⁺ T cells and myeloid DCs correlates with the immune pathogenesis of atrial fibrillation

Introduction

Atrial fibrillation (AF) is the most common arrhythmia, but the pathophysiological mechanism leading to this condition remains poorly understood. However, increasing evidence suggests the influence of an immune-mediated inflammatory response in AF pathogenesis 1. It has been established that serum or plasma inflammatory biomarkers including CRP and IL-6 2, 3, and autoantibodies such as anti-M2 muscarinic receptor autoantibodies 4, 5, are elevated in AF patients. Furthermore, studies indicate that activated T-lymphocytes infiltrate the endomyocardial in AF patients 6-8. Together, these observations suggest that the immune response contributes to AF pathogenesis. Nevertheless, there is insufficient knowledge to define the molecular and cellular mechanisms of the immune response in the AF process.

Programmed death-1 (PD-1, CD279) is a CD28 homolog and costimulatory molecule expressed on CD⁴⁺ and CD⁸⁺ T cells, NK cells, B cells and monocytes upon activation. It is engaged primarily by its ligands PD-L1 (B7-H1, CD274) and PD-L2 (B7-DC, CD273) 9, 10. PD-L1 is widely expressed on immunocompetent cells such as T cells, B cells, dendritic cells (DCs) and macrophages in nonlymphoid organs, such as heart, lung, placenta, kidney and liver 11, 12. However, PD-L2 expression is more restricted, and previous studies suggest that PD-L2 is inducibly expressed on DCs and macrophages 13, 14. Recently, two studies found that PD-L2 is also expressed on CD⁴⁺ and CD⁸⁺ T cells 15, 16. Increasing evidence suggests that the PD-1/PD-L pathway inhibits mainly T and B cell proliferation and cytokine production 17, 18. Owing to its critical inhibitory function in immune regulation, the PD-1/PD-L pathway has been the focus of a number of recent studies. It is noteworthy that in the heart, PD-L1 and PD-L2 are expressed at high levels 10. Meanwhile, accumulating evidence implicates PD-1/PD-L pathway in the pathogenesis of certain cardiovascular diseases, including atherosclerosis 19, 20, immune-mediated myocarditis 21, 22 and dilated cardiomyopathy (DCM) 23, 24. However, it remains unclear whether the PD-1/PD-L pathway influences immune regulation in AF.

In this study, we examined the expression profiles of PD-1 on T cells and PD-L1 and 2 on T cells, myeloid dendritic cells (mDCs) and macrophages in patients with paroxysmal and persistent AF, to assess their roles in AF pathogenesis. We found that PD-1/PD-L1 expression is substantially down-regulated on CD⁴⁺ T cells or mDCs in AF patients, and that the extent of the down-regulation is closely related with the AF burden. Additionally, PD-1/PD-L1 down-regulation significantly enhanced CD⁴⁺ and CD⁸⁺ T cell proliferation and influenced cytokines secretion in vitro. Our study indicates that the PD-L1 may be induced by IFN-α. We further up-regulated expression of PD-L1 on mDCs by IFN-α treatment and found that mDC-mediated T cell proliferation was suppressed and the pattern of cytokine secretion was converted. This suggests that PD-1/PD-L1 down-regulation on CD⁴⁺ T cells or mDCs may modulate immune regulation and play a role in AF pathogenesis.

Materials and methods

Results

Cell surface expression of PD-1 on CD⁴⁺ and CD⁸⁺ T cells

To explore the surface expression characteristics of PD-1/PD-L on peripheral blood cells, we firstly detected PD-1 expression on CD⁴⁺ and CD⁸⁺ T cells in paroxysmal AF patients, persistent AF patients and control group. Representative dot plots of PD-1+ distinct subpopulations were shown in Figure 1A and C. The results showed that PD-1+ distinct subpopulations of CD⁴⁺ T cells in AF patients was significantly decreased (P < 0.001) than healthy controls (12.04 ± 1.95). Compared the two groups, we found persistent AF patients (1.79 ± 0.32) expressed a lower level of PD-1 than paroxysmal AF patients (Fig. 1B; 8.45 ± 1.07). However, there were no statistically differences in PD-1+ distinct subpopulations of CD⁸⁺ T cells among paroxysmal AF group (3.76 ± 0.43), persistent AF group (3.61 ± 0.54) and control group (Fig. 1D; 3.73 ± 0.70).

Cell surface expression of PD-L1, and PD-L2 on mDCs, macrophages and T cells

We further detected PD-L1 and PD-L2 expression on the function cells. Representative dot plots of PD-L1+ and PD-L2+ distinct subpopulations were shown in Figures 2A, C, E, G and 3A, C, E, G. The results showed that PD-L1 expression on mDCs in AF patients was significantly down-regulated (P < 0.001) than healthy controls (0.95 ± 0.19), and persistent AF group (0.25 ± 0.08) was lower than paroxysmal AF group (Fig. 2B; 0.57 ± 0.16). We found that there were no statistically differences in PD-L1 expression on macrophages, CD⁴⁺ and CD⁸⁺ T cells among the three groups (Fig. 2D, F and H). As for the second PD-ligand, PD-L2 expression on mDCs, macrophages, CD⁴⁺ and CD⁸⁺ T cells in the three groups all had no statistically differences (Fig. 3B, D, F and H). These down-regulated expression of PD-1 on CD⁴⁺ T cells and PD-L1 on mDCs in AF patients suggested that the PD-1/PD-L1 pathway involving in T(CD⁴⁺)/mDCs interactions might participate in the pathogenesis of AF.

Results of follow-up study

We found that the frequency of AF episodes correlated with lower surface expression of PD-1/PD-L1 on peripheral blood mDC and CD⁴⁺ T cells. To further explore this phenomenon, we examined a paroxysmal AF group with a 3-month follow-up. We divided them into two subgroups (group F and group O) and measured PD-1/PD-L1 expression on CD⁴⁺ T cells and mDCs. The results showed that PD-1+ distinct subpopulations of CD⁴⁺ T cells and PD-L1+ distinct subpopulations of mDCs in group F (7.86 ± 0.77; 0.51 ± 0.09) were significantly lower than group O (9.00 ± 0.81; 0.61 ± 0.13; Fig. 4A and B). Results of the follow-up study suggested that the extent of PD-1/PD-L1 down-regulation was closely related with the AF burden. In short, the frequency of AF episodes correlated with lower expression of PD-1/PD-L1 on CD⁴⁺ T cells and mDCs.

IFN-α induce in vitro PD-L1 expression

It has been reported that IFN-α stimulates PD-L1 on mDCs, although the mechanism by which this occurs is unclear 27. To investigate IFN-α regulation of PD-L1 on mDCs, isolated mDCs were cultured with or without IFN-α stimulation. mDCs cultured in medium alone showed a slight increase in PD-L1 expression, peaking at 24 hrs before returning to background levels at 72 hrs (Fig. 5). In contrast, IFN-α (10,000 U/ml) induced PD-L1 expression on mDCs in the first 24 hrs, with expression sustained up to 72 hrs.

Influence of PD-L1 expression on mDCs on T cells proliferation

In the allogeneic mixed leukocyte reactions (MLR), we cocultured mature mDCs from persistent AF patients and healthy volunteers at different ratios with T cells from a third healthy individual and studied how the decreased PD-L1 expression on mDCs in AF patients influence the T cells proliferation. The results showed that mDCs of AF patients (n = 10) were more efficient in inducing CD⁴⁺ and CD⁸⁺ T cell proliferation at all ratios tested compared with mDCs from controls (n = 10; Fig. 6; all P < 0.05). It suggested that the T cell-stimulating capacity of mDCs from AF patients was enhanced. Then, we examined the effects of mDC-associated PD-L1 up-regulation on allogeneic T cell proliferation in the presence of IFN-α. We found that the upgrade capacity of mDCs from AF patients to stimulate CD⁴⁺ and CD⁸⁺ T cells proliferations was restored, leading to results similar to that of normal mDCs (Fig. 6; all P < 0.05). Furthermore, we examined the effects on allogeneic MLR in the presence of anti-PD-L1. As expected, we found that the restored capacity of mDCs from AF patients to stimulate CD⁴⁺ and CD⁸⁺ T cells proliferations was up-regulated again by treated with anti-PD-L1 (Fig. 6; all P < 0.05). These data suggested that the down-regulation of PD-L1 on mDCs appears to transmit stimulating signal to CD⁴⁺ and CD⁸⁺ T cells in AF patients, resulting in an increase of proliferation and PD-L1 up-regulation on mDCs can reduce the allostimulatory capacity of mDCs from AF patients.

MDC-associated PD-L1 expression is involved in cytokine production by T cells and mDCs themselves

Cytokines are the key regulatory factors that modulate the strength of the T cell response and activation 28. We therefore examined the production of type 1 effector cytokines IL-2 and IFN-γ, and the type 2 effector cytokines IL-10, in culture supernatants of allogeneic MLR to examine whether PD-L1 expression influences cytokine production. We found that the mDCs with low-level PD-L1 from AF patients were more efficient at inducing IL-2 and IFN-γ secretion by both CD4 and CD8 T cells, than were healthy controls. We then stimulated PD-L1 expression on mDCs by IFN-α treatment and found that IL-2 and IFN-γ were both significantly decreased by T cells, which was similar to that observed for mDCs from healthy volunteers. Furthermore, we found that changes in cytokine production resulting from IFN-α treatment could be reversed by anti-PD-L1 mAbs treatment. We also examined production of IL-10, an anti-inflammation cytokine antagonistic to IL-2 and IFN-γ activity 29, and found that the IL-10 secretion pattern was opposite that observed with IL-2 and IFN-γ (Fig. 7E and F). These observations suggest that the decrease of inhibitory signals by PD-L1 down-regulation on mDCs result in an increase of pro-inflammation cytokines (IL-2 and IFN-γ) secretion, and a decrease of anti-inflammation cytokines (IL-10) secretion. This is consistent with the idea that T cells proliferation seems to promote AF. These inhibitory signals can be mostly blocked by treating with IFN-α, further supporting the notion that PD-L1 down-regulation may be responsible for enhanced T cell responses and changed cytokines secretion.

Discussion

AF is the most common sustained arrhythmia and has serious health consequences owing to haemodynamic impairment and thromboembolic events 30. However, the pathophysiological nature of AF is complicated and remains largely unknown. Recently, the link between inflammation and AF has been gaining attention 1 and many studies suggest that the immune response may contribute to this inflammatory process 2-8, 31, 32. Exploring the mechanism of the immune response in AF will be critical to further elucidating the pathophysiological mechanism of AF. This will be important for developing a therapeutic regimen to improve or terminate AF.

It has been well-demonstrated that in addition to the signal from the binding between MHC/peptide complex and T cell receptor which is called the first signal, the second signal that regulates the balance between stimulatory and inhibitory signals is even more important for effective immune response 10. The PD-1/PD-L signalling pathway (belonging to the B7:CD28 family) has been highlighted recently as important regulators for maintaining this critical balance. Several studies have also shown that the PD-1/PD-L pathway plays primarily a negative regulatory role in regulating T cell activation, proliferation and cytokine production 13, 33. Additional studies further indicate that PD-1 ligation blocks TCR-mediated signalling by preventing phosphorylation and activation of protein kinase C 10, 34. Meanwhile, the PD-1/PD-L pathway has been shown to play a pathogenic role in many immunoinflammatory diseases such as chronic intestinal inflammation 35, chronic inflammatory mucocutaneous disease 36, rheumatoid arthritis 37 and helicobacter pylori infection 38. However, it is unclear whether there are important inhibitory signals that also participate in AF, which is a putative immunoinflammatory disease. It is important to note that PD-L1 and PD-L2 expression is at high levels in the heart 10. This may imply that they play an essential role in the cardiovascular system, and in fact the PD-1/PD-L pathway has been shown to play at least a partial role in cardiovascular diseases. Using a mouse model, Gotsman et al. 19 has shown that the PD-1/PD-L pathway exerts a significant regulatory influence on the immune response and is associated with development of arterial disease. Grabie et al. 22 and Lucas et al. 21 have also revealed a critical role for the PD-1/PD-L pathway in controlling autoimmune cardiac injury. Given this, it will be interesting to explore the potential relationship between the PD-1/PD-L pathway and AF pathogenesis.

In this study, we comprehensively investigated the expression of PD-1 and PD-L1 and 2 on key regulatory cells including mDCs, macrophages, and CD⁴⁺ and CD⁸⁺ T cells in AF patients. Allogeneic MLR was carried out to explore the role of the T(CD⁴⁺)/mDCs-associated PD-1/PD-L1 pathway in immune regulation of AF. Our results suggest that down-regulation of PD-1/PD-L1 on CD⁴⁺ T cells or mDCs might, at least in part, take part in AF pathogenesis. Firstly, PD-1 expression on CD⁴⁺ T cells and PD-L1 on mDCs is down-regulated in AF patients compared with controls. Secondly, down-regulation of PD-1/PD-L1 was more significant in persistent AF patients, those considered to have a more pronounced inflammatory state compared with paroxysmal AF patients. This suggests that the extent of PD-1/PD-L1 down-regulation may be closely related with AF burden, which is further supported by the subsequent follow-up investigation. Thirdly, allogeneic MLR experiments indicate that low-level mDCs PD-L1 expression promotes T cell proliferation, suggesting that the PD-L1 levels may be linked with the inflammatory state. Cytokine secretion changed with mDC PD-L1 levels, further supporting this assertion. Collectively, PD-1/PD-L1 down-regulation on CD⁴⁺ T cells or mDCs may participate in the process of immune regulation and play a role in AF pathogenesis.

Regarding the underlying mechanisms about how PD-1/PD-L1 down-regulation modulates AF pathogenesis, we speculate on the following possibilities. Firstly, PD-1/PD-L1-associated T cell proliferation may contribute to atrial electrical remodelling. T cell proliferation may also increase secretion of cytokines, such as IL-6 and MCP-1, which may affect the contractility and electrical stability of myocytes differently 39. Furthermore, studies indicate that cytokine and inflammatory responses participate in atrial structural remodelling 39-41. Additional studies found that MRL mice with a PD-1 deficiency spontaneously developed severe myocarditis, which lead to deposits in the extracellular matrix and displayed massively enlarged hearts 21, 42. Another study further demonstrated that because of cardiac troponin I autoantibodies, PD-1 deficient mice developed autoimmune DCM 23. Although these inferences are consistent with a role for the PD-1/PD-L1 pathway in AF pathogenesis, the underlying mechanisms need to be investigated further.

In this study, we investigated PD-L1 expression on several cell types, and mDCs have been proposed to be responsible for specific T cell regulation in AF. As a subpopulation of DCs, mDCs express elevated levels of MHC and a range of accessory molecules involved in cell migration, adhesion and co-signalling 43. Our study found that expression of inhibitory PD-L1 molecules on mDCs was significantly down-regulated in AF patients, whereas IFN-α expression induced PD-L1 on mDCs. We therefore used IFN-α to stimulate PD-L1 expression on mDCs isolated from AF patients in allogeneic MLR. Allogeneic MLR showed that AF mDCs with decreased PD-L1 were more effective in stimulating T cell proliferation and changing cytokine secretion, suggesting that mDCs-associated PD-L1 down-regulation tips the balance between inhibitory and stimulatory signals delivered to T cells in AF patients. Furthermore, we found that PD-L1 expression levels on AF macrophages were similar to that of controls. This might suggest that macrophages, an important immune cell, do not contribute to AF progression. Similarly, PD-L1 expression on CD⁴⁺ and CD⁸⁺ T cells in AF patients was not statistically different compared with controls, suggesting that PD-1/PD-L1 inhibitory signals between T-T cells may not participate in regulating T cells. Importantly, the expression of PD-L2 on mDCs, macrophages, and CD⁴⁺ and CD⁸⁺ T cells in AF patients was not statistically different compared with controls. Our study suggests that PD-L2 indeed is expressed on CD⁴⁺ T and CD⁸⁺ T cells, consistent with earlier studies 15, 16. Although PD-L2 has been recognized to play a pivotal role in diseases such as asthma 44 and EAE 45, it is seemingly not involved in AF pathogenesis. Collectively, these results suggest a T(CD⁴⁺)/mDCs-associated PD-1/PD-L1-independent mechanism in AF regulation.

It is important to note that several limitations that need to be addressed regarding this study. First, to supporting the suggestion that the PD-L1 levels may be linked with the inflammatory state, more cytokines except IL-2, IL-10 and IFN-γ need to be examined. In addition, TNF-α as another important inflammatory cytokines may be not relevant to PD-L1 change in MLR (data not shown). The specific mechanisms is unclear and need further research. Second, although the PD-1/PD-L1 pathway appears to be important, additional co-inhibitory pathways such as those involving CTLA-4 and B7 interactions cannot be excluded in AF pathogenesis. Further study will be needed to determine whether the PD-1/PD-L1 pathway is the only mechanism or if it is just one that regulates the immune response in AF. Third, the specific role of the PD-1/PD-L1 pathway in AF is not clear and it remains unclear whether PD-1/PD-L1 down-regulation is a cause of AF, or merely a consequence.

In summary, the present study begins to shed light on the role of immune pathogenesis in AF, showing that PD-1 is down-regulated on CD⁴⁺ T cells and PD-L1 on mDCs in AF patients. The extent of PD-1/PD-L1 down-regulation was closely related with AF burden. Furthermore, PD-1/PD-L1 down-regulation could improve, at least in part, T cell proliferation and cytokine secretion change, which reflect the inflammatory state and contribute to AF pathogenesis. Therefore, PD-1/PD-L1 down-regulation on CD⁴⁺ T cells or mDCs participates in the pathogenesis of AF and up-regulating the expression of PD-1/PD-L1 molecules may represent a new therapeutic option for AF.