Redistribution of Monocyte Subsets in Obstructive Sleep Apnea Syndrome Patients Leads to an Imbalanced PD-1/PD-L1 Cross-Talk with CD4/CD8 T Cells

Sleep is of fundamental importance for all living beings, although its manifold contributions to the fundamentals of life are still somewhat of a mystery. One of the vital roles of sleep is to consolidate the incredible amount of information our brains have to handle every day. Furthermore, wholesome sleep is essential to secure metabolite clearance, tissue repair, or hormone biosynthesis (1–4).

Obstructive sleep apnea syndrome (OSAS), in contrast, represents a substantial disease of recurrent breathing intermissions, resulting in hypopnea/apnea followed by reoxygenation. Up to 9% of the adult population suffer from this potentially life-threatening disorder. Besides reduced quality of life, OSAS is a major risk factor for cardiovascular diseases, metabolic disorders, and cognitive dysfunctions (5). Moreover, wholesome sleep and normal circadian patterns are highly important for a functioning immune system. Because of nocturnal oxidative stress, OSAS is considered a low-grade systemic inflammation that is initiated by intermittent hypoxia and sleep fragmentation that leads to activation of NF-κB and HIF-1α (3, 6). These promote the activation of immune cells, especially lymphocytes and monocytes that secrete high amounts of inflammatory cytokines and adhesion molecules (6, 7). Several studies demonstrated increased levels of C-reactive protein, TNF-α, IL-6, IL-8, ICAM, and VCAM in OSAS patients (6, 8, 9). Consequently, excessive adherence of immune cells and an overall prevailing proinflammatory milieu leads to vascular injury and the development of atherosclerosis and cardiovascular diseases (10–12).

OSAS-induced hypoxic stress increases the invasion of monocytes into the endothelium and lipid uptake of macrophages, which transform into foam cells and contribute to atherosclerosis (12–14). Furthermore, an increased level of TNF-α, MCP-1, and matrix-metalloproteinase-9 has been described in OSAS patients (8, 11).

Monocytes can be classified into three subsets based on their CD14 and CD16 expression (15–17). CD14⁺⁺CD16⁻ are called “classical” monocytes and comprise the biggest compartment with ∼85% of monocytes. They are professional phagocytes, have the highest capacity to migrate, and secrete the highest amounts of proinflammatory cytokines. CD14⁺CD16⁺ “intermediate” monocytes are with 5–10% of the major APCs. The remaining 5–10% constitute CD14^dim+CD16⁺ “nonclassical” monocytes, which are associated with antiviral immunity because of their higher expression of TLR7 and TLR9 (18, 19). It has been hypothesized that both CD16⁺ subsets emerge from the classical monocyte population and are therefore more differentiated cells (20–22). An increase of CD16⁺ monocytes has been described in many inflammatory conditions as sepsis (23), chronic viral infections (24), rheumatoid arthritis (25), periodontitis (26), and asthma (27). Up to now, nothing is known about the peripheral monocyte composition in OSAS patients.

T cells are crucial players for the release of proinflammatory cytokines. In recent years, it became increasingly evident that T cells are affected by OSAS-induced oxidative stress (28–30). Changes of the B, T, and NK cell composition and activation of certain lymphocyte subsets have been described (31–33). In detail, a shift to a T_H1/T_H17 dominant phenotype with a lack of regulatory T cells has been reported in OSAS patients (34, 35). Moreover, an altered phenotype and function of cytotoxic NK cells and CD8⁺ T cells has been proposed, which is characterized by increased cytotoxicity against endothelial cells (36, 37). As a consequence, vascular damage is done by activated T cells and monocytes, which leads to the development of OSAS-related cardiovascular disease.

The aim of the study was to understand the impact of OSAS on differentiation patterns of circulating monocytes with regard to the three above-described subsets. Furthermore, a detailed analysis of the CD4/CD8 T cell subset composition, especially with respect to the immune checkpoint molecules programmed death 1 (PD-1) and programmed death ligand 1 (PD-L1), was addressed because an upregulated PD-1/PD-L1 cross-talk is known to suppress T cell activation. The study aimed to increase our understanding on immunological changes in these cells and their possible underlying interplay.

All blood donors have signed an informed written consent and were informed about the aims of the study and the use of their samples. Blood samples were collected from healthy donors (n = 20) or OSAS patients (n = 48). Blood was drawn by venipuncture into a sodium citrate containing S-Monovette (Sarstedt, Nümbrecht, Germany). All OSAS patients were treatment naive and analyzed directly after first-time diagnosis. They had a mean age of 50 (±14), a body mass index (BMI) of 29.18 kg/m² (±5.16), an apnea–hypopnea index (AHI) of 17.8 events/h (±16.96), and a minimal O₂ saturation of 88.58% (±4.93). Healthy donors had a mean age of 48 (±16) and a BMI of 24.56 kg/m² (±3.60).

Within 4 h after blood collection, 20 μl of citrate blood was diluted in 80 μl PBS. Blood cells were stained with following Abs: CD45-PE, CD14-FITC, CD16-BV-510, HLA-DR-allophycocyanin-Cy7, and CD3-PerCP (all from BioLegend, San Diego, CA). After 25-min staining in the dark, 650 μl RBC Lysis Buffer (BioLegend) was added to the samples and incubated for another 20 min. Subsequently, suspension was centrifuged at 400 × g for 5 min, and supernatant was discarded. Cell pellet was resuspended in 100 μl fresh PBS and ready for FACS analysis.

PBMC were isolated from the remaining blood by density gradient centrifugation in Biocoll (Biochrom, Berlin, Germany) at 400 × g for 20 min. The upper PBS/plasma layer was removed and discarded. The PBMC layer was carefully harvested and transferred to a new 15-ml tube and washed once with PBS. The supernatant was discarded again and the PBMC pellet was resuspended in 1 ml PBS. For analysis of T cell subsets, 100 μl of the cell suspension was incubated with two different Ab mixtures: 1) CD3-PerCP, CD4-PE-Cy7, CD8-BV-510, PD-1-PE, PD-L1-allophycocyanin, CD45RA-allophycocyanin-Cy7, and CCR7-BV421 and 2) CD3-PerCP, CD4-PE-Cy7, CD8-BV-510, PD-L1-allophycocyanin, CCR6-allophycocyanin-Cy7, CXCR3-BV421, CD39-PE, and CD25-FITC. After 25-min staining in the dark, cells were washed with PBS and centrifuged at 400 × g for 5 min. Cell pellets were resuspended in 200 μl PBS and prepared for FACS analysis.

Flow cytometry was performed with a MACSQuant 10 flow cytometer (Miltenyi Biotec, Bergisch-Gladbach, Germany), and data were analyzed using the FlowJo software version 10.0 (FlowJo, Ashland, OR). Subsequent to Ab titrations and compensations, at least 100,000 CD45⁺ leukocytes were analyzed for whole blood measurements. Gating of monocyte subsets was performed as described by Marimuthu and colleagues (38) and shown in Fig. 1. In brief, CD45 was used as a pan leukocyte marker to facilitate whole blood measurement and exclusion of debris. After doublet exclusion, monocytes were first roughly gated by their forward scatter (FSC)/side scatter (SSC) characteristics and the positivity for CD14 and CD16. Neutrophil granulocytes, NK cells, and B cells were excluded with the help of HLA-DR and CD14, respectively. The remaining real monocytes were then subgated into CD14⁺⁺CD16⁻ (classical), CD14⁺⁺CD16⁺ (intermediate), and CD14^dim+CD16⁺ (nonclassical) monocytes.

For T cell analysis, 100,000 events within the PBMC gate were measured. Gating scheme of T cell subpopulations is shown in Fig. 2.

Statistical analyses were performed with GraphPad Prism Version 7.0f (GraphPad Software, San Diego, CA). The mean and SEM are presented. The differences between groups were determined after testing for Gaussian distribution (normality tests) and applying parametric (Student t test) or nonparametric one-way ANOVA with Bonferroni post hoc test. The correlation between parameters was calculated using multivariate regression with the Pearson correlation coefficient: *p < 0.05, **p < 0.01, and ***p < 0.001. Additional statistical details are given in the respective figure legends, when appropriate.

Variations in replicate numbers for certain parameters were caused by too old blood samples for proper monocyte analysis or too low PBMC numbers for reputable T cell subset characterization, respectively.

Our investigations identified strong alterations of the abundance of all three monocyte subsets in OSAS patients compared with healthy donors (Fig. 1). Classical monocytes were significantly decreased, whereas intermediate and nonclassical subsets were significantly increased (Fig. 2). Furthermore, based on the severity of monocyte alteration, OSAS patients could be subdivided into three distinct groups. Seven patients showed quite normal percentages of the monocyte subsets and no or only mild alterations in comparison with healthy donors (Fig. 3A). Most patients (n = 29) revealed a moderate decrease of classical monocytes accompanied by an increase of intermediate monocytes. Nine patients showed a severe drop of classical monocytes with a strong increase of intermediate and nonclassical monocytes (Fig. 3A, Table I). There was a positive correlation between the severity of the monocyte alterations and the BMI of the patients (p = 0.039; Fig. 4). In contrast, AHI, minimal O₂ saturation, or age was not found to be correlated with percentages of monocyte subsets. Therefore, all clinical parameters were not significantly different between the defined subgroups (Table I).

The PD-L1 expression of monocytes was also analyzed during FACS analysis of PBMCs. Monocytes were gated by their FSC/SSC characteristics and further on by the help of their low positivity for CD4 and missing CD3 expression (Fig. 2). Our data revealed a significantly higher PD-L1 expression on monocytes of OSAS patients compared with healthy donors (Fig. 3B). The PD-L1 expression level also showed a significant correlation with the severity of the monocyte alterations (p = 0.018). Furthermore, there was a positive correlation between monocytic PDL-1 expression and the AHI (p = 0.02) and the BMI (p = 0.012) of the patients (Fig. 4).

Our analyses also identified increased levels of monocyte/T cell aggregates in the peripheral blood of OSAS patients using flow cytometry. Double staining using anti-CD14 and anti-CD3 Abs represent CD14⁺CD3⁺ monocyte/T cell aggregates within the analyzed T cell fraction. The percentage of monocyte-aggregated T cells was significantly increased in OSAS patients compared with healthy donors (Fig. 3C).

Subsequent investigations of T cell subsets in OSAS patients compared with healthy donors were carried out using flow cytometry (Fig. 2). The T cell differentiation from naive to effector, effector memory, and central memory cells was analyzed for CD4⁺ and CD8⁺ T cells by specific makers. Moreover, CD4 T helper subsets T_H1, T_H2, and T_H17/22 could be distinguished. The percentages of each subset as well as the PD-1 and PD-L1 expression levels were analyzed.

Our data revealed significantly increased levels of CD4⁺ effector T cells accompanied by a decrease of effector memory and central memory T cells (Fig. 5A). CD8⁺ T cells were not significantly altered in OSAS patients (Fig. 5A). Percentages of PD-1⁺ cells as well as the PD-1 expression intensity (mean fluorescence intensity [MFI]) were significantly increased on CD4⁺ and CD8⁺ T cells in OSAS patients, indicating an activated immune phenotype (Fig. 5B). PD-L1 expression was also slightly increased on CD4⁺ and CD8⁺ T cells from OSAS patients whereby big individual differences could be observed (Fig. 5C).

We further questioned whether the grade of monocyte alteration was related to changes we found in T cell subsets. Therefore, patients were subdivided as described earlier into three groups: “mild alteration,” “moderate alteration,” and “severe alteration.” The composition of T cell subsets showed no differences between the groups (data not shown). The PD-1 expression of CD4⁺ and CD8⁺ T cells was linked to the grade of monocyte alteration. Only patients with a severe monocyte alteration revealed a significant difference compared with healthy donors in PD-1 expression on CD4⁺ and CD8⁺ T cells (Fig. 6). The PD-L1 expression on both T cell subtypes was consequently significantly correlated with the severity of monocyte alterations (p ≤ 0.001). Only OSAS patients with severe monocyte alteration revealed an increased PDL-1 expression on CD4⁺ T cells (Fig. 6). Of note, CD8⁺ T cells showed the highest increase of PD-L1 expression in the patient group with severe monocyte alterations, which was also significantly higher than in all other groups.

The present study was undertaken to investigate the impact of OSAS on percentages of monocyte subsets, monocyte–T cell complex formation, and T cell immune alterations, distinguishing between CD4⁺ and CD8⁺ T cells. Additionally, the expression levels of PD-1 and PD-L1 on the peripheral T cells of OSAS patients and healthy donors were analyzed for CD3⁺CD4⁺ and the CD3⁺CD8⁺ subsets and monocytes using flow cytometry.

Under healthy conditions, around 85% of monocytes are classical CD14⁺CD16^- monocytes, which are characterized by phagocytosis, endothelial transmigration, and production of proinflammatory cytokines. The remaining 15% are linked to more specialized Ag presentation functions, defense against viruses, and patrolling behavior (7). Under various inflammatory conditions, an increase of CD16⁺ intermediate and nonclassical subsets has been reported (15). These subsets are generally termed proinflammatory because of their high production of TNF-α and IL-1β. In our study, we showed for the first time, to our knowledge, a massive decrease of classical monocytes accompanied by an increase of CD16⁺ subsets in OSAS patients (Fig. 3A). A similar high percentage of intermediate monocytes was up to now only reported in asthma (27). A possible explanation is OSAS-related oxidative stress, followed by upregulation of proinflammatory transcription factors namely NF-κB and HIF-1α (39).

Another important factor is obesity. We found a positive correlation between the severity of monocyte alterations with the BMI of the patients. The World Health Organization reported that half of the European population is overweight, and ∼30% of those are obese. In our patient cohort, 85% were overweight, and 40% of these were obese. Of note, an increase of the intermediate and nonclassical monocyte subsets has already been shown in obesity (40–42). A coherence of obesity, OSAS, and atherosclerosis has been well described in recent years (7, 39, 43, 44). Our data corroborate this thesis as we found a positive correlation between BMI and AHI (p = 0.041) and a negative correlation between BMI and O₂ saturation by tendency (p = 0.076; Fig. 4). OSAS evokes less physical activity and metabolic dysfunction, which promotes obesity. Obesity, in contrast, is a major risk factor for OSAS (44). Both conditions promote a higher presence of proinflammatory mediators and are therefore classified as low-grade systemic inflammation. Because of mutual molecular signaling pathways, OSAS and obesity are likely to act synergistically as factors of immune disturbance. An additional therapy for body weight reduction might therefore be helpful for OSAS patients.

Furthermore, we found an increase of PD-L1 expression on monocytes (Fig. 3B), which has been described before (45). The monocytic PD-L1 expression level was the only parameter correlating with the AHI in our patient cohort (p = 0.02; Fig. 4). It was additionally correlated with the severity of monocyte alteration. The monocytic PD-L1 expression also showed a positive correlation with the PD-L1 expression on CD4⁺ and CD8⁺ T cells, which was likewise highest in patients with severe monocyte alterations (Fig. 6). Intermittent hypoxia has been linked to an upregulation of HIF-1α, which induces an increase of PD-L1 on monocytes (30) and T cells (29).

Our study revealed significantly increased T cell–monocyte aggregates in OSAS patients compared with healthy donors (Fig. 3C). A recent study proposed that these aggregates are a crucial marker for any type of immune perturbations (46). These findings could also be a hint for increased communication between these cells, causing an altered PD-1/PD-L1–mediated immune regulation. We found a significant increase of effector CD4⁺ T cells and a higher expression of PD-1 on CD4⁺ and CD8⁺ T cells (Fig. 5B). PD-1 expression is a natural consequence of T cell activation. Engagement of PD-1 on T cell surface inhibits T cell proliferation and effector functions (47, 48). PD-L1 on monocytes, in contrast, promotes a crucial mechanism for the regulation of T cell responses preventing overreaction. PD-L1⁺ T cells have also been hypothesized to perform regulatory functions by themselves (49, 50). An imbalanced activation of the PD-1/PD-L1 pathway can, however, ultimately lead to the development of chronic inflammations and related diseases. In fact, an imbalanced PD-L1 signaling has been associated with chronic viral or bacterial infections (51, 52). Impaired T cell functions caused by hypoxia-induced PD-1/PD-L1 cross-talk have been described in OSAS patients (30, 45). The authors explained thereby the higher aggressiveness of cancer in OSAS patients, and PD-L1–mediated immune suppression could also be a protective mechanism to avoid substantial damage to the vascular system and related diseases.

Several studies reported an increase of T_H1/T_H17 cells accompanied by a decrease of regulatory T cells, which leads to a shift to an overall proinflammatory milieu (33–35). Our data revealed no increase in T_H1/T_H17 cells. Our findings might be explained by the overall moderate grade of OSAS in our patient cohort as alterations were often linked to the AHI. Changes in these T cell compositions might predominantly occur in patients with severe OSAS.

In conclusion, the current study indicates that OSAS and related obesity induces alterations in the composition of peripheral monocyte subsets and increased PD-L1 expression on monocytes. Furthermore, monocyte alterations were found to be connected to increased T cell activation accompanied by an imbalanced PD-1/PD-L1 cross-talk. However, attention should be paid to the fact that our study cohort was composed mostly of patients with mild to moderate OSAS and that severe monocyte alterations were only found in 9 out of 48 patients. It is possible that other comorbidities are not negligible factors that provoke monocyte subset alterations. Nevertheless, the findings presented in this study provide new insights concerning monocyte-driven immune alterations explaining the higher incidence of infections, cardiovascular diseases, and cancer in OSAS patients.

We are grateful to all members of the Department of Otorhinolaryngology for stimulating and helpful discussions.

The authors have no financial conflicts of interest.