Results

PD-L1 is induced in blood cells of pre-school asthmatic children with a virus-induced asthma phenotype and associated with the presence of rhinovirus in their airways

We recently described that acute in vitro infection of peripheral blood mononuclear cells from preschool children with and without asthma with rhinovirus, a single stranded RNA picornavirus, is associated with the upregulation of Interferon regulated genes like STAT1, STAT2 and Interferon Regulatory Factor (IRF) 1 (Bergauer et al., 2017b; a). Moreover, paradoxically, IFNγ upregulates also PD-L1, a factor involved in silencing/exhausting of activated T cells by ligating PD1 on the surface of T cells (Mandai et al., 2016). Consistently, we found that acute rhinovirus infection ex vivo induced PD-L1 and CTLA4 in the PBMCs of asthmatic children (Bielor et al., 2017). We thus wanted to follow up these in vitro observations in the two cohorts of our study and analyzed 21 control children and 24 children with asthma(Fig. 1a) . The clinical data of these cohorts of children were recently reported (Bergauer et al., 2017b; Bielor et al., 2017; Hentschke et al., 2017) and are summarized in Table 1 andTable 2 . By looking at the PD-L1 mRNA expression in blood, we found that PD-L1 mRNA expression was induced in children with a virus-induced asthma phenotype (in accordance to PRACTALL guidelines 2008 (Bacharier et al., 2008)) compared to healthy control children(Fig. 1b) . Children with this asthma phenotype shows symptom free periods, whereas the most common precipitating factor are colds by respiratory viruses, like humane rhinovirus (Bacharier et al., 2008).
Furthermore, by trend, we observed an induction of PD-L1 mRNA in the blood cells of asthmatic children as compared to control children(Fig. 1c) . We next analyzed PD-L1 expression after allergen and rhinovirus challenge. Considering the presence of rhinovirus (+RV) in the airways, we found that, by trend, asthmatic children with rhinovirus in the airways, have an increased PD-L1 mRNA expression in total blood cells (Fig. S1a). This is also associated with increased expression of the Low density lipoprotein receptor (LDLR) (Fig. S1b), which is one of the main receptors used by the viruses, especially for RV1b, to entering the cells.

PD-L1 is upregulated in blood cells of asthmatic children with increased bronchoconstriction

We then asked if the lung function, especially the FEV1% as well as the PEF% (peak expiratory flow, predicted), of the cohorts would correlate with increased PD-L1 expression in blood. The FEV1 (Forced expiratory volume in 1 second) / FVC (Forced vital capacity) ratio (FEV1%), is a calculated ratio used in the diagnosis of obstructive and restrictive lung disease. It represents the proportion of a person’s vital capacity that they are able to expire in the first second of forced expiration (FEV1) to the full, forced vital capacity (FVC). The result of this ratio is expressed as FEV1% (Swanney et al., 2008). Lower values of FEV1% represent airway obstruction. In our cohort of children with asthma, but not in control children, we found a PD-L1 induction in children with higher bronchoconstriction (Fig. 1d) and an inverse correlation between PD-L1 and FEV1% (Fig. 1e) , indicating that worse asthma is associated with induction of PD-L1 mRNA in blood cells of children with asthma. We then further investigated the role of another lung function parameter, the PEF% (peak expiratory flow) value (Fig. 1f, g) . The PEF% is defined as the largest expiratory flow, which is achieved with a maximum forced effort after maximum inspiration and is used as a control parameter during asthma therapy. Similarly to the FEV1% we found a significant PD-L1 induction in children with worse asthma (Fig 1f) as well as an inverse correlation between PD-L1 expression and the PEF% (Fig 1g) . We also found that increased PD-L1 mRNA expression correlated with reduced FEV1% and PEF% (Fig. 1h) (I do not see Fig 1h) , indicating that asthmatic preschool children with rhinovirus colonization in the airways have worse respiratory function associated with PD-L1 induction in their peripheral blood mononuclear cells (PBMCs). By contrast, healthy control children with and without rhinovirus in the airways as well as in asthmatic children without rhinovirus colonization in the airways no correlation between FEV1% or PEF% and PD-L1 was observed (Fig S1c, d) .

IFNβ correlated with better lung function in asthmatic children

We next reasoned that in the case of asthma induced by infections, especially rhinovirus infections, IFN-type I and specifically IFNβ might be of importance (Staples et al., 2015). Thus, we next analyzed the IFNβ level in cell culture supernatants of untreated PBMCs from healthy and asthmatic children with and without rhinovirus in the airways (Fig. 2a) as well as after a re-stimulation with RV1b in vitro (Fig. S2a, b) and correlated them with their FEV1% and PEF% (Fig. 2a,b; Fig. S2c-e and S3). Here we found that, only asthmatic children and especially asthmatic children with RV in their upper airways show a direct correlation between the IFNβ level and the FEV1% and PEF%, respectively, indicating that a subpopulation of children could respond to rhinovirus infection with IFNβ production.

PD-L1 levels correlated with IFNβ-production in healthy but not in asthmatic children

Since it is known that Interferon induces PD-L1 (Friedrich et al., 2018) we correlated the IFNβ expression in the supernatants of untreated and with RV1b re-stimulated PBMCs and the PD-L1 expression in total blood cells and found a direct correlation in control children, but not in asthmatic children (Fig. 3, S4a, b) . These data indicate that IFNβ is associated with PD-L1 in control children and that asthmatic children have a disturbed IFNβ mediated PD-L1 induction.

PD-L1 is upregulated in blood cells of asthmatic children with high C-reactive protein (CRP) serum levels and correlated with RV in the airways

We next reasoned that not only rhinovirus but also other infection or inflammatory agents could cause PD-L1 induction in asthmatic children. We thus next looked at C-reactive protein (CRP) level in serum of our cohorts of children. CRP binds to the phosphocholine expressed on the surface of dead or dying cells and some bacteria and leading to the activation of the complement system and promotion of phagocytosis by macrophages (Bray et al., 2016). Higher levels are found in inflammation, viral infections (10–40 mg/L), active bacterial infection (40–200 mg/L), severe bacterial infections and burns (>200 mg/L) (Chew, 2012). We considered high CRP levels as an indicator of ongoing infection and inflammation and found that children with asthma and a CRP value over 5 mg/l had a significantly higher PD-L1 mRNA expression in total blood cells as compared to the control children(Fig. 4a) . Moreover, in both healthy and asthmatic children, CRP was found to be associated with high PD-L1 levels in the serum(Fig. 4b; S4c) . Finally, in the presence of rhinovirus in the airways, CRP correlated with PD-L1 expression in healthy children(Fig. 4c) . Taken together, these data suggest the presence of induced PD-L1+ cells in the blood of asthmatics with worse asthma and ongoing inflammation and infection.