TSLP Production in the Human Infant Airway Epithelium and Clinical Relevance during Viral Respiratory Infections

To the Editor:

Studies using human airway epithelial cells (AECs) derived from adults with asthma have confirmed that type 2 airway epithelial cytokines, such as TSLP (thymic stromal lymphopoietin), are key mediators in the pathogenesis of the disease (1), and this fundamental knowledge has led to the discovery of novel asthma therapies (2). In contrast, the airway epithelium of human infants has been remarkably understudied despite compelling evidence that asthma often begins in early life (3). In addition, prior studies have established that airway secretion of TSLP occurs during viral respiratory infections in young children (4, 5). Thus, our goal in this human-based study was to characterize the production of TSLP in vitro in primary infant AECs and in vivo during natural viral respiratory infections in human infants. This study was approved by the Institutional Review Board of Children’s National Health System and included parental informed consent.

Methods and Results

To study early-life airway epithelial immune responses, we generated primary nasal AEC lines derived from human infants (n = 6; mean age 15 mo, range 0–24 mo) using conditional reprogramming cell methods (6). We quantified protein secretion of TSLP and other type 2 cytokines (IL-25 and IL-33), as well as IL-1β, a potent inducer of TSLP (7), using electrochemiluminescence assays (MSD). Human infant AEC lines exhibited prominent production of TSLP in response to a viral mimic (polyinosinic:polycytidylic acid [poly I:C], 10 ng/ml) (median 88.7 pg/ml at 24 h), but not at baseline. Neither IL-25 nor IL-33 was detectable at baseline or after viral mimic exposure (Figure 1A). Human infant AECs secreted IL-1β endogenously as well as after poly-I:C exposure (Figure 1B), and IL-1β induced TSLP production in a dose–response manner (Figure 1C). Because prior studies in immortalized AEC lines have shown that poly I:C and IL-1β elicit TSLP production via canonical NF-κB activation (7), we next examined this mechanism in primary infant AECs. As shown in Figures 1D–1F, we found that siRNA-mediated knockdown of the RELA/p65 subunit of NF-κB ablated TSLP production in infant AECs exposed to poly I:C or IL-1β. Taken together, our initial in vitro experiments indicated that TSLP is the predominant innate type 2 cytokine produced by the human infant airway epithelium in response to a viral stimulus (poly I:C) or IL-1β, and that this occurs via an NF-κB–dependent mechanism.

Figure 1.

(A and B) Human infant nasal airway epithelial cell cultures (six donors) were exposed to polyinosinic:polycytidylic acid (poly I:C) (10 ng/ml), and the protein levels of TSLP (thymic stromal lymphopoietin), IL-25, and IL-33 (A), and IL-1β were measured in the supernatant at the indicated time points (B). (C) TSLP production in infant airway epithelial cells after IL-1β exposure (dose response). (D–F) Western blotting showing siRNA-mediated knockdown of p65 (RELA) (D) and its effect on TSLP production after exposure to poly I:C (10 ng/ml ×  24 h) (E) and IL-1β (20 ng/ml × 24 h) (F). C–F represent data from two to three donors. P values were obtained by the Mann-Whitney U test. *P < 0.05, **P < 0.01, and ***P < 0.001. scRNA = scrambled negative control siRNA.

We next conducted an in vivo study in human infants to examine the clinical relevance of TSLP during early-life viral infections. For this purpose, we measured TSLP protein levels in nasal lavages of infants hospitalized for a PCR-confirmed viral respiratory infection (n = 111; mean age 11 mo, range 0–24 mo). We used electronic medical records to determine respiratory hospitalizations/emergency room (ER) visits before hospitalization or 12 months after discharge. Respiratory hospitalizations/ER visits were defined as those involving a primary respiratory complaint. Based on these data from electronic medical records, we divided our study population into subjects with respiratory hospitalization/ER visits before or after the index hospitalization (recurrent respiratory events group) and subjects without any other respiratory hospitalization/ER visit (single respiratory event group). As shown in Figure 2A, TSLP protein levels were significantly higher in the group of infants with recurrent respiratory events (median single 0.14 pg/ml vs. recurrent 0.29 pg/ml; P = 0.04). Using logistic regression, we found that infants with high TSLP levels (>75th percentile) had an increased probability of recurrent respiratory events independently of sex, prematurity, and respiratory syncytial virus (RSV) status (adjusted odds ratio, 4.4; P = 0.008). We did not find any differences in TSLP levels according to respiratory viruses (Table E1 in the data supplement). Infants with high TSLP levels also had higher IL-1β nasal protein levels (Figure 2B), suggesting an interaction between these cytokines during early-life viral infection.

Figure 2.

(A) TSLP protein levels were quantified in nasal lavages during a viral respiratory infection in human infants with recurrent hospitalizations (n = 62) or a single respiratory hospitalization/emergency room visit (n = 42). (B) Nasal protein levels of IL-1β according to high-TSLP-levels status (values >75th percentile). (C) Probability (proportion of subjects) of recurrent respiratory episodes according to high-TSLP-levels status and sex. P values were obtained by the Mann-Whitney U test and chi-square test. *P < 0.05 and **P < 0.01.

Discussion

Our findings provide new insights into the biology and clinical relevance of TSLP in human early life. We identified that the human infant airway epithelium responds to a viral stimulus (poly I:C) with robust TSLP secretion, but not with other type 2 epithelial cytokines (e.g., IL-25 and IL-33) as observed in adults with asthma and viral infections (8). However, we may not have been able to detect the early peak production of IL-33 as reported during allergen challenges (9, 10). Interestingly, human infants with RSV infections have elevated airway levels of group 2 innate lymphoid cells (ILC2s) that produce IL-33 and other type 2 cytokines (5). This suggests that in early life there may be a cooperative interaction between the epithelium and ILC2s to generate airway type 2 responses, as previously described in immature mice during rhinovirus infection (11). Although it is unclear whether the epithelial secretion of TSLP initiates airway type 2 immune responses in human infants, studies using TSLP-receptor knockout have demonstrated that the mechanism of ILC2 activation during RSV infection is dependent on TSLP (12). More recently, a longitudinal study in mice identified that TSLP is involved in the generation of type 2 responses after neonatal RSV infection and enhanced responses to allergy exposure, particularly in males (6). Interestingly, we also found that in human infants, the highest probability of recurrent respiratory events was observed in males with high TSLP levels (Figure 2C), which may be connected to the fact that boys have an increased likelihood of developing wheezing illnesses and childhood asthma (13).

A limitation of our study is that we used a commercially available assay that does not distinguish between the long proinflammatory TSLP isoform and the short antiinflammatory/antimicrobial isoform (14). Given that these TSLP isoforms have strikingly different biological functions, future studies are needed to fully elucidate the clinical implications of early-life TSLP production. In summary, our in vitro and in vivo data on human infants provide initial support to the growing evidence that TSLP plays an important role during viral respiratory infections in early life, as previously demonstrated in several animal models (11, 12, 15). Nonetheless, our current findings need to be confirmed in prospective human-based studies. Establishing the role of TSLP in the human infant airway is critically important because this epithelial cytokine might be a clinically relevant therapeutic target to prevent wheezing illnesses and asthma in early life. This could have a dramatic impact on the incidence of this condition because most cases of adult asthma originate during early childhood (3).