Reply to Yasuma et al.: Response of Lung Microbiota to Changes of Pulmonary Innate Immunity under Healthy Conditions

From the Authors:

We appreciate the thoughtful letter from Yasuma and colleagues regarding our recently published research letter (1).

The authors inquire about our selection of time points and speculate that had we compared lung microbiota at an earlier time point after exposure, we might have observed an effect of innate immune modulation on lung communities. We chose our time point (harvest 6 d after exposure) based on prior work characterizing the sustained effects of a single Pam2-ODN exposure on lung immunity, reflected in persistent protection from bacterial, fungal, and viral infection (2). We agree with the authors that our findings do not exclude the possibility of a transient effect and stated as much in our original manuscript: “Although we detected no appreciable effect of Pam2-ODN on lung bacterial communities 6 days after exposure, it is entirely possible that TLR agonism has a short-lived effect on lung microbiota that was resolved by the time of harvest due to the continuous exposure of the lungs to environmental microbiota” (3).

To directly address this possibility, we performed a separate experiment to determine the effect of lung innate immune modulation (via single Pam2-ODN exposure) on lung bacterial communities at 48 hours after exposure. Experimental conditions, interventions, and measurements were otherwise identical to those described in our research letter (1). As demonstrated in Figure 1, we observed no effect of Pam2-ODN exposure on lung bacterial density, diversity, or community composition at 48 hours. We believe these findings do not support the authors’ hypothesis and further confirm our interpretation that innate immune modulation (via TLR2/6 and TLR9 agonism) has no appreciable effect on the lung microbiota of healthy mice.

Figure 1.

Experimental modulation of lung innate immune tone does not influence lung microbial communities at 48 hours. Healthy, adult mice received either phosphate-buffered saline inhalation (“sham”) or synergistic TLR2/6 and TLR9 stimulation via inhaled Pam2-ODN. Lungs were harvested 48 hours after exposure, and lung microbiota were characterized via (A) droplet digital PCR or (B and C) 16S rRNA gene amplicon sequencing. TLR agonism did not influence (A) the total bacterial burden in murine lungs, (B) lung bacterial diversity, or (C) lung community composition. Horizontal lines and error bars represent median and interquartile range, respectively. Significance was determined using (A) Mann-Whitney test, (B) Student’s t test, and (C) permutational multivariate ANOVA. n = 15 mice per experimental group; one specimen per group was excluded from diversity and community composition analysis because of inadequate sequencing depth. PC = principal component; rRNA = ribosomal RNA.

We note that in the relevant study by Wu and colleagues (4) (to which our research letter was a response), a single experimental modulation of lung microbiota resulted in persistent immune effects detectable at 14 days. Thus, even if selective innate immune modulation does transiently influence lung microbiota (i.e., for hours rather than days), its effect is more self-limited than the converse interaction. As stated in our research letter, we believe important unanswered questions remain regarding 1) timing and duration of lung innate immune modulation and its potential effect on lung microbiota, 2) the effect of other specific immune-modulating exposures (e.g., TLR inhibition, agonism of other TLRs) on lung microbiota, and 3) the relative role of immune tone in shaping lung communities in disease states (as opposed to health).

Although we appreciate the authors’ summary of several mechanistic effects of synergistic TLR agonism (5, 6), we do not believe these observations alter our interpretation of our microbiome findings. Although indices of lung inflammation have been correlated with lung microbiota in both health and disease (3, 7), the causal direction underlying this correlation has not been fully elucidated. To our knowledge, no study to date has demonstrated that host-derived reactive oxygen species or cytokines exert a causal influence on lung microbial communities (as postulated by the authors). We believe that, taken with the findings of Wu and colleagues (4), our findings suggest that in health, variation in lung immunity reflects variation in lung microbiota, rather than vice versa. Yet we readily concede that considerable additional work will be necessary to definitively resolve this question.