Introduction

The upper respiratory tract (URT), including the nasal, oral, and pharyngeal mucosae, serves as an important node of control for inhaled pathogens. Immune responses in the URT are critical to combat infectious agents and prevent them from spreading to fragile sites such as the lungs, an organ in which pathogen exposure may compromise gas exchange and cause life-threatening tissue pathologies.

Immunity in the upper airways is multifactorial, encompassing both innate and adaptive components. Among the adaptive components, secretory (s)IgA antibody responses induced within the upper airway lumen provide protection upon exposure to respiratory pathogens, such as influenza virus1,2. sIgA antibodies mediate their protective effects at mucosal surfaces by exclusion and neutralization of inhaled pathogens, thereby precluding the colonization or infection of local epithelial cells without eliciting overt inflammatory responses3. In addition, recent evidence suggests that sIgA may also prevent the local dissemination of viral particles4. Through these mechanisms, sIgA-mediated defense in the URT provides an initial barrier to infection, complementing IgG- and cytotoxic T cell-mediated defenses that protect the lungs against harmful infectious agents.

Cigarette smoke is a noxious inhaled stimulus that is known to compromise or alter the function of the immune system, both in the respiratory tract and at distal sites5. While it is well documented that smokers are at increased risk of various infectious diseases6, including influenza7, the mechanisms by which smoke exposure causes this predisposition have not been fully elucidated. With regards to sIgA responses in the URT, the reported effects of cigarette smoke exposure are controversial. For instance, Rylander et al. reported that smokers exhibited lower levels of IgA specific for the lipopolysaccharides (LPS) of the Gram-negative bacteria Pantoea agglomerans and Pseudomonas syringae in nasal secretions compared to non-smokers within cohorts of wood- and cotton-mill workers8. Similarly, a number of groups have reported decreased baseline sIgA levels in the saliva of cigarette smokers9,10,11,12, while still others have indicated that smoking increases salivary IgA13,14. In mouse models, cigarette smoke has been reported to inhibit IgA induction in the lungs following influenza virus infection15; however, IgA is thought to play a minimal role in directly protecting the lower respiratory tract from disease pathogenesis in this context1. Overall, to what extent cigarette smoke modulates the induction of novel sIgA responses in the upper airways, where this isotype is a potent contributor to host defense, remains unclear. This is an important knowledge gap, given that infectious diseases represent a significant source of excess mortality in the 1.1 billion individuals worldwide that continue to smoke tobacco16,17.

To address this question, we studied the effect of cigarette smoke exposure on the induction of antigen-specific sIgA responses in the URT of mice. We found that tobacco smoke inhibited IgA induction in the systemic circulation and nasal lumen, in part by attenuating the accumulation of antibody-secreting cells (ASCs) in the underlying nasal mucosa and draining lymphoid tissues. In addition to being decreased in quantity, we provide evidence that IgA generated in smoke-exposed animals possesses a lower antigen-binding avidity, but paradoxically, a greater mutational load than observed in room air controls. Furthermore, smoke exposure compromises the upregulation of molecules critical to nasal ASC homing and transepithelial IgA transport in the nasal mucosa. Overall, our findings clearly indicate that cigarette smoke compromises multiple mechanisms involved in effective sIgA induction in the URT and associated lymphoid tissues.

Results

Cigarette smoke compromises the induction of IgA responses in the upper respiratory tract following intranasal LPS/OVA administration

To investigate whether cigarette smoke compromises IgA induction in the upper airways, we exposed mice to either room air (RA) or cigarette smoke (CS) for one week, and subsequently intranasally immunized these mice with a mixture of LPS, a prototypical pathogen-associated molecular pattern, and ovalbumin (OVA), a well-described model antigen, using a multidose prime/boost schedule (Fig. 1A). Smoke exposure was continuous during this protocol. We observed consistently decreased OVA-IgA in both the nasal lavage fluid (NALF; Figs. 1B, S1A) and serum (Figs. 1C, S1A) at 3, 7, 14, and 28 days post-boost (dpb). In a similar manner, LPS-specific IgA was induced in the serum of room air-exposed, but not smoke-exposed mice (Fig. 1D). Notably, despite observing concurrent reductions in OVA-IgA in the NALF and serum at 3dpb, no significant correlation was observed between these two parameters within room air- and cigarette smoke-exposed groups (RA: r = 0.2385, p = 0.5069 and CS: r = 0.1036, p = 0.7758; Fig. 1E). Deficits in total IgA were also observed in the NALF, while no significant changes were observed in the serum (Figs. 1F, S1B/C). OVA-IgA and total IgA levels in the NALF correlated strongly within smoke-exposed mice at 3dpb (r = 0.8402, p = 0.0023), while room air-exposed mice exhibited a moderate but non-significant positive correlation (r = 0.4956, p = 0.1452; Fig. S1D). NALF and serum OVA-IgA were similarly decreased at 3dpb in mice exposed to cigarette smoke for 10 and 32 weeks prior to immunization (Fig. S1E), as well as in mice administered 1 or 7 booster doses (Fig. S1F). OVA-IgA induction was also attenuated in the vaginal lavage fluid (VALF; Fig. 1G). In contrast to IgA, OVA-specific IgM and IgG1 levels in the serum were unchanged between groups (Fig. 1H), and OVA-IgM, which can also be secreted across mucosal surfaces in a manner similar to IgA, was not detectable in the NALF (data not shown). Finally, when cigarette smoke-exposed mice were alternatively immunized with a mixture of OVA and polyI:C, a TLR-3 agonist and viral nucleic acid mimic, we found that the induction of OVA-IgA responses was similarly impaired (Fig. 1I). These findings suggest that acute and chronic cigarette smoke exposure compromises the host’s ability to induce IgA responses against antigens in the upper airways, and that such inhibition is observed in the context of diverse pathogen-associated molecular patterns.

Fig. 1: Cigarette smoke exposure compromises the induction of nasal and systemic IgA responses following intranasal LPS/OVA immunization.
figure 1

A C57BL/6 mice were exposed to room air (RA) or cigarette smoke (CS) for the indicated durations. Subsequently, three doses of LPS/OVA or PBS vehicle were administered intranasally to the upper respiratory tract, once daily. After 2 weeks of continued exposure, mice received four booster doses of LPS/OVA or vehicle. Animals were culled and samples collected at the indicated timepoints. OVA-specific or LPS-specific antibodies of the indicated isotypes were measured in the indicated samples by ELISA at (B, C) the indicated timepoints or (D, E, G–I) 3dpb. B, C Data represent n = 3–10 mice per group from three independent experiments. D, G, H Data represent n = 3–7 mice per group. E Data represent n = 10 mice per group from two independent experiments. F Total IgA was measured by ELISA at the indicated timepoints. Data represent 5–10 mice per group from two independent experiments. I Mice were exposed to room air or cigarette smoke for 1 week and immunized with polyI:C/OVA as per the schedule in panel A. Data represent n = 4–5 mice per group. E Linear regression and Pearson’s correlation, dotted lines represent 95% confidence bands. All others two-way ANOVA with Tukey’s post-hoc test. Mean ± SEM. *p < 0.05.

Full size image

Cigarette smoke exposure attenuates OVA-specific ASC accumulation in the nasal mucosa after intranasal LPS/OVA immunization

IgA antibodies secreted into the upper airway lumen are produced by antibody-secreting cells (ASCs; plasma cells) within the nasal mucosa18. To determine whether the observed reduction in nasal OVA-IgA was associated with changes in ASC populations, we next quantified the presence of OVA-specific ASCs (IgD–B220loCD138+IgG1–IgA+OVA+; Fig. 2a, FMO controls presented in Fig. S2A) in the nasal mucosa by flow cytometry. OVA-specific ASCs were quantified by intracellular staining with AlexaFluor (AF)647-conjugated OVA. We observed reductions in the number of both total and OVA-specific IgA ASCs (represented in Fig. 2b) in the nasal mucosa of cigarette smoke-exposed mice at both 3dpb and 7dpb (Figs. 2c/d, S2B/C). Notably, these differences were not caused by changes in the total cellularity of the nasal mucosa (Fig. S2D). OVA-specific ASCs comprised more than 50% of total IgA ASCs at 3dpb, with smoke-exposed mice exhibiting a small but significant reduction in this proportion compared to RA controls (Fig. 2e). Similar to antibody titres, deficits in OVA-IgA ASCs were also observed in mice exposed to cigarette smoke for either 10 or 32 weeks prior to immunization (Fig. S2C). In contrast, OVA-IgA ASCs in the bone marrow were not significantly reduced in the context of smoke exposure at either 3 or 28dpb (Figs. 2f, S2E). These data demonstrate that the deficit in nasal antigen-specific IgA observed in the context of cigarette smoke exposure occurs, at least in part, as a result of reduced ASC accumulation in the underlying mucosa.

Fig. 2: Cigarette smoke exposure attenuates OVA-specific IgA ASC accumulation in the nasal mucosa after intranasal LPS/OVA immunization.
figure 2

Mice exposed to room air (RA) or cigarette smoke (CS) for one week were intranasally immunized with LPS/OVA, and mice were analyzed at the indicated timepoints. a Nasal OVA-IgA ASC gating strategy (IgD–B220loCD138+IgG1–IgA+OVA+). b Flow cytometry plots demonstrating total ASCs (top), and OVA-IgA ASCs within IgA ASCs (bottom) in the nasal mucosa at 7dpb. Frequencies presented are mean ± SD. Absolute numbers of (c) total and (d) OVA-specific IgA ASCs were quantified in the nasal mucosa (NM) at the indicated timepoints. Data represent (c) n = 3–7 mice per group from two independent experiments, (d) n = 3–7 mice per group from three independent experiments. e Nasal OVA-IgA ASCs were quantified as a proportion of total IgA ASCs at 3dpb. Data represent n = 12 mice per group from two independent experiments. f Absolute number of OVA-IgA ASCs in the bone marrow quantified at 3dpb. Data represent n = 3–10 per group. FVD: fixable viability dye. e Student’s unpaired t-test, all others two-way ANOVA with Tukey’s post-hoc test. Mean ± SEM. *p < 0.05.

Full size image

Cigarette smoke exposure reduces nasal OVA-IgA avidity during the acute post-immunization period while augmenting nasal Igha mutational load

In addition to quantity, the antigen-binding affinity of antibodies plays an integral role in their effector capacity. IgA responses induced in upper airway-associated lymphoid tissues are able to undergo somatic hypermutation (SHM) of their variable regions in order to promote antibody affinity maturation19. In light of this, we sought to assess metrics of antibody antigen-binding capacity. Initially, we quantified the ability of intracellular IgA to retain AF647-conjugated OVA by examining the ratio of IgA and OVA median fluorescence intensities (MFIs) within nasal OVA-IgA ASCs following intracellular staining. At 3dpb, we observed that there was no significant difference in IgA MFI between OVA-IgA ASCs generated in room air- and cigarette smoke-exposed mice (Fig. 3A), suggesting that antibody production rate is likely similar between ASCs induced under either condition. In contrast, OVA MFI displayed a strong trend towards being reduced in the context of cigarette smoke. When quantified as a ratio, smoke-exposed mice demonstrated a significant reduction in OVA:IgA within ASCs compared to room air controls (Fig. 3B). A similar trend was observed at 7dpb, but by 14dpb no differences were observed. Notably, during this timeframe the MFI ratio remained stable for smoke-exposed mice but contracted for room air-exposed mice. These early differences in MFI ratio suggest that, despite the same amount of total IgA per OVA-specific IgA ASC, there is a differential ability of smoke-exposed ASCs to retain cognate antigen and thus an attenuated antigen-binding affinity.

Fig. 3: Cigarette smoke exposure reduces nasal OVA-IgA avidity during the acute post-immunization period while augmenting mutational load in the cognate nasal Igha repertoire.
figure 3

Mice exposed to room air (RA) or cigarette smoke (CS) for one week were intranasally immunized with LPS/OVA as described in Fig. 1, and metrics of antibody antigen-binding affinity and somatic hypermutation were assessed. A Median fluorescence intensity (MFI) of IgA and OVA was quantified within OVA-IgA ASCs in the nasal mucosa at 3dpb. Data represent n = 11–12 mice per group from two independent experiments. B OVA:IgA MFI ratio at various timepoints. Data represent n = 11–12 mice per group from two independent experiments (3dpb) and n = 5 mice per group (7dpb and 14dpb independently). C Avidity indices were derived for NALF OVA-IgA by chaotropic urea ELISA. Data represent 13–15 animals per group from two independent experiments at each timepoint. D Nasal Igha mRNA transcripts were analyzed at 3dpb by high-throughput immunoglobulin sequencing. Replacement mutations were quantified in CDRs 1/2 and FWRs 1/2/3 on a per-clone basis. Data represent n = 3–9 mice per group. A.U. arbitrary units. A–C Student’s unpaired t-tests, D Mann-Whitney test with Bonferroni correction. Mean ± SEM. *p < 0.05.

Full size image

Although similar measures of antigen retention by B cell receptors have been demonstrated to correlate with avidity20, we sought to confirm these findings through direct analysis of NALF OVA-IgA antibodies. To do so, we used avidity ELISAs, which feature an antibody dissociation step via the transient addition of choatropic urea after sample incubation. Using this approach, we determined that OVA-IgA in the NALF of smoke-exposed animals demonstrated a lower avidity index than that of room air controls at 3dpb (Fig. 3C). Similar to ASC MFI data, avidity indices normalized by 28dpb, suggesting that smoke exposure delays but ultimately does not permanently compromise the affinity maturation of OVA-IgA responses following immunization.

Next, we assessed whether cigarette smoke exposure alters the extent to which the IgA ASCs induced in the upper airways undergo SHM. To do so, we performed high-throughput sequencing to quantify mutational load within the heavy chain variable region of nasal IgA transcripts as described by Turchaninova et al.21. Specifically, we assessed the mean frequency of replacement mutations within combined complementarity-determining regions (CDRs 1/2), which comprise the antigen-binding paratope, and framework regions (FWRs 1/2/3; Fig. 3D), which are also reportedly also able to alter antibody effector capacity22. Through clonal clustering23, we found that room air- and smoke-exposed, immunized mice exhibited a similar predicted number of unique IgA clones per sample (Fig. S2F). In terms of mutational load, we observed a significant increase in the frequency of replacement mutations within CDR1/2 and FWR1/2/3 regions of smoke-exposed, immunized mice compared to room air controls (Fig. 3D). Overall, these data suggest that IgA from the nasal mucosa of smoke-exposed, immunized mice possesses a greater mutational burden than room air controls, despite demonstrating a reduced polyclonal avidity at this timepoint.

Cigarette smoke exposure attenuates the induction of OVA-IgA ASCs in nasal-draining lymphoid tissues

ASCs that populate the upper airways are typically induced within secondary lymphoid tissues that drain the nasal mucosa, such as the nasal-associated lymphoid tissue (NALT), cervical lymph nodes (CLNs) and spleen. To assess whether the observed reduction in nasal ASCs results from attenuated ASC induction, we next quantified OVA-IgA ASCs at these sites (IgD–B220loCD138+IgG1–IgA+OVA+; Fig. S3A). At 3dpb, a timepoint at which we observe a large attenuation of IgA induction, we also observed that cigarette smoke attenuated the expansion of both total (Fig. 4A/B) and OVA-IgA ASCs (Figs. 4C, S3G) in the NALT. In room air-exposed mice, LPS/OVA immunization caused the physical expansion of CLNs as measured by cellularity and weight (Figs. 4D, S3B). In contrast, smoke-exposed naïve mice had smaller CLNs at baseline, which did not expand to the same size as room air controls following immunization. Notably, this difference in baseline CLN weight could not be explained by differences in mouse weight, which were minimal (Fig. S3F). Within the CLNs, we observed a reduction in the absolute number of OVA-IgA ASCs, as well as the proportion of OVA-IgA ASCs within total ASCs (Figs. 4E, S3C). Similarly, although the spleen did not expand with immunization, spleens from smoke-exposed mice were significantly smaller than those of room air controls at baseline (Figs. 4F, S3D), and fewer OVA-IgA ASCs were present (Figs. 4G, S3E). Lastly, we analyzed the blood at this timepoint, and found fewer OVA-IgA ASCs in the blood of cigarette smoke-exposed mice compared to room air controls (Fig. 4H). Overall, these data suggest that cigarette smoke compromises the induction of OVA-IgA ASCs in nasal-draining lymphoid tissues, including the NALT, CLNs, and spleen, and ultimately reduces the output of these cells into the circulation.

Fig. 4: The induction of antigen-specific IgA ASC responses is diminished in the NALT, CLNs and spleen in the context of cigarette smoke exposure.
figure 4

Mice exposed to room air (RA) or cigarette smoke (CS) for one week were intranasally immunized with LPS/OVA as described in Fig. 1, and NALTs, CLNs, spleen, and blood were assessed by flow cytometry at 3dpb. Total NALT ASCs are (A) visualized as flow cytometry plots within the parent gate (IgD– live singlets) and B quantified. C NALT OVA-IgA ASCs quantified as a percentage of total ASCs. NALT data points represent the means of pooled samples (n = 5–7 mice per group per experiment) from three independent experiments. D Pooled CLN cell count per mouse. E Absolute number of CLN OVA-IgA ASCs. F Spleen cell counts per mouse. Absolute number of OVA-IgA ASCs in the (G) spleen and (H) whole blood. Data represent: (D–G) n = 3–12 mice per group from two independent experiments, (H) n = 3–10 mice per group, one of two representative experiments is shown. Total NALT GCBCs are (I) visualized as flow cytometry plots within the parent gate (CD138–B220+) and (J) quantified. K NALT OVA-IgA GCBCs quantified as a percentage of total GCBCs. NALT data points represent the means of pooled samples (n = 5–7 mice per group per experiment) from two independent experiments. B, C, J, K Student’s unpaired t-tests, D–H two-way ANOVA with Tukey’s post-hoc test. Mean ± SEM. *p < 0.05.

Full size image

In light of the observed changes in IgA antigen-binding capacity and mutational frequency in the context of smoke exposure, we also sought to assess whether smoke-exposed mice exhibited any differences in the number of germinal center B cells (GCBCs; IgD–B220+CD138–GL7+CD95+; Fig. S3A), the B cell stage in which SHM and affinity maturation occur. Interestingly, at baseline, cigarette smoke-exposed control mice exhibited an increase in the proportion of GCBCs in the NALT compared to room air controls (Fig. 4I). However, at 3dpb, smoke-exposed mice exhibited fewer total (Fig. 4I/J) and OVA-IgA GCBCs (Figs. 4K, S3H). Ultimately, these data suggest that cigarette smoke-exposed mice exhibit elevated basal NALT germinal center activity but feature reduced antigen-specific IgA germinal center induction relative to room air controls following intranasal immunization.

Cigarette smoke reduces basal CCL28 and VCAM-1 levels in the nasal mucosa and interferes with the transcriptional upregulation of VCAM-1 following intranasal immunization

Following their generation in secondary lymphoid tissues, newly-formed ASCs migrate to the nasal mucosa24. Homing of ASCs to respiratory surfaces is thought to be mediated by two molecular axes: a) the interaction between CCL28 produced at the mucosa and the CCR10 receptor on the surface of ASCs, and; b) the interaction between VCAM-1 expressed by local endothelial cells and α4β1 integrin (VLA-4) on ASCs24. To assess whether cigarette smoke compromises ASC homing, we first assessed baseline expression of tissue-derived factors CCL28 and VCAM-1. Cigarette smoke exposure reduced baseline expression of CCL28 and VCAM-1 in nasal homogenates following both 2 and 8 weeks of exposure. In contrast, no significant differences were observed in either CCL28 or VCAM-1 expression in lung homogenates or serum (Fig. 5a/b). In addition, we observed that the upregulation of VCAM-1, but not CCL28, was impaired at 3dpb in nasal homogenates of cigarette smoke-exposed mice (Fig. 5c). Vcam1 transcript upregulation was similarly diminished in smoke-exposed mice that were administered 10 μg of LPS alone for 1 h (Fig. 5d). Next, we assessed the expression of α4β1 integrin, the VCAM-1 binding partner. In this regard, we observed a marginal but significant reduction in the proportion of α4β1+ OVA-IgA ASCs within the CLNs of smoke-exposed mice, while no difference was observed on cells in the bloodstream or nasal mucosa (Fig. 5e). In the CLNs, more substantial reductions were observed in the MFIs of both α4 and β1 integrin subunits (Fig. 5f). Again, however, α4β1 expression in the bloodstream was equivalent, and no differences were observed in the effector tissue. Overall, these data suggest that cigarette smoke interferes with the transcriptional upregulation of VCAM-1 in the nasal mucosa, likely impairing the extravasation of ASCs into the submucosa following immunization.

Fig. 5: Cigarette smoke exposure attenuates the transcriptional upregulation of VCAM-1 in the nasal mucosa following immunization.
figure 5

Mice were exposed to room air (RA) or cigarette smoke (CS) for 2 or 8 weeks. Serum, NM homogenates and lung homogenates were assessed for a CCL28 and b VCAM-1 expression by ELISA. Data represent n = 5–7 mice per group from two independent experiments. c Mice were exposed to room air or cigarette smoke for 1 week and intranasally immunized with LPS/OVA as described in Fig. 1a. At 3dpb, nasal homogenates were assessed for CCL28 and VCAM-1 expression by ELISA. Data represent n = 5 mice per group. d Mice were exposed to room air or cigarette smoke for 2 weeks and intranasally administered 10 μg of LPS alone. Vcam1 gene expression was quantified in nasal lysates 1 h post-administration. Data in each group are normalized to the mean of the RA-PBS group. Data represent n = 5 mice per group. e, f Mice were treated as in (c), at which point (e) the proportion of α4+β1+ integrin OVA-IgA ASCs and (f) the MFI of each integrin was quantified. Data represent n = 5 mice per group. A.U. arbitrary units. a, b, e, f Student’s unpaired t-tests, c, d two-way ANOVA with Tukey’s post-hoc test. Mean ± SEM. *p < 0.05.

Full size image

Cigarette smoke attenuates the upregulation of pIgR in the nasal mucosa following intranasal immunization

In addition to impaired induction and homing of ASCs to the nasal mucosa, quantities of IgA in the nasal lumen may be affected by the efficiency of transepithelial transport. IgA is transported from the submucosa across epithelial barriers with the assistance of the polymeric immunoglobulin receptor (pIgR)24. Using Western blots, we found that cigarette smoke exposure attenuated the upregulation of pIgR and its post-transport cleavage product secretory component (SC) in the nasal mucosa following LPS/OVA administration (Figs. 6, S4). This impairment in pIgR upregulation may further explain the reduced OVA-IgA levels observed in the nasal lumen of smoke-exposed mice, given the non-redundant role of this receptor in IgA transcytosis.

Fig. 6: Cigarette smoke exposure attenuates the upregulation of pIgR in the nasal mucosa following immunization.
figure 6

Room air (RA) and cigarette smoke (CS)-exposed mice were immunized with LPS/OVA as per Fig. 1. At 3dpb, and nasal mucosa was extracted. a Western blots for pIgR/secretory component (SC) were performed. An uncropped version of pIgR/SC blots is presented in Fig. S4. b pIgR and SC expression was quantified by densitometry. Data represent n = 5 mice per group. A.U. arbitrary units. Two-way ANOVA with Tukey’s post-hoc test. Mean ± SEM. *p < 0.05.

Full size image

Discussion

IgA-mediated immunity has been implicated in providing protection against viral agents in the upper airways1, and is thought to be an important component of existing and novel vaccination strategies against influenza and other mucosal pathogens25,26,27,28. As a result, it is important to understand how common inhaled environmental factors such as cigarette smoke can impact IgA induction in the upper airways in order to identify circumstances in which IgA-mediated defense, elicited by either natural infection or intranasal vaccination, may be ineffective.

In our studies using LPS/OVA as a model immunogen, we observed that both acute and long-term cigarette smoke exposure attenuated the induction of antigen-specific IgA responses in the URT and systemic circulation. Nasal and serum antigen-specific IgA responses did not correlate well with one another, which support previous findings in humans2, and studies in mice suggesting differential mechanisms of mucosal and systemic IgA induction29. In contrast, OVA-IgA correlated well with total IgA in the nasal lumen, which suggests that the OVA-specific response comprised the major proportion of total nasal IgA antibodies in our model. Interestingly, cigarette smoke exposure also attenuated IgA responses at the vaginal mucosa. The common mucosal immune system is known to facilitate the dissemination of lymphocytes between distal mucosal sites30, with antigen-specific responses induced in the URT being able to provide protection at the female reproductive tract. These findings imply that tobacco smoke may compromise the efficacy of intranasal vaccination strategies intended to elicit protective IgA responses in the genital tract31,32,33. Notably, we observed an attenuation of OVA-IgA responses in the URT and circulation even when polyI:C was used as an adjuvant instead of LPS, suggesting that mechanistically, smoke exposure attenuates conserved elements of both TLR-4- and TLR-3/RIG-I-dependent IgA inductive pathways.

Cigarette smoke exposure impaired the generation of short-lived mucosal OVA-IgA ASCs in the nose, but not ASCs that populated the bone marrow. This reduction in nasal ASCs notably corresponded with a reduced number of these cells within inductive sites that drain the nasal mucosa, including the NALT, CLNs, and spleen. Evidence of reduced IgA ASCs in these secondary lymphoid organs suggests that the activation of T cell-dependent B cell responses at these sites is compromised. This may constitute deficits in the activation or polarization of nasal dendritic cells34,35, follicular helper T(FH) cells36, or B cells. Further investigation is required to understand which of these pathways may be affected. Overall, however, our findings suggest that IgA ASC expansion is differentially modulated by cigarette smoke at different effector sites, but is consistently impaired within inductive tissues.

The reduced VCAM-1 upregulation we observe is reflective of previous observations in cigarette smoke extract/TNF-α-treated human coronary artery endothelial cells37, suggesting that circulating ASCs may be unable to extravasate into the nasal mucosa though an inefficient ability to attach to local endothelial cells. In addition, our baseline CCL28 and VCAM-1 data are in line with previous clinical reports; in human smokers, CCL28 gene expression has been observed to be attenuated in nasal epithelium, but not bronchial epithelium, and VCAM-1 protein expression is unchanged in the pulmonary vasculature relative to controls38,39. Similarly, findings of impaired pIgR expression strongly indicate that IgA produced in the nasal submucosa may not be efficiently transcytosed to the nasal lumen following immunization. pIgR expression and sIgA levels have been reported to be attenuated in the airways of patients with severe chronic obstructive pulmonary disease (COPD)40,41, an effect that is thought to be caused by extensive epithelial remodeling during disease progression40. In contrast, the attenuation of nasal pIgR upregulation we observed occurs following an acute duration of smoke exposure, and thus likely occurs via an alternate mechanism. Overall, however, the data presented here demonstrate a refractory capacity of stromal cells within nasal mucosa to upregulate inflammatory molecules in response to immunogenic stimulation following cigarette smoke exposure.

Importantly, we also observed differences in antibody affinity and mutational load between room air- and smoke-exposed mice. Nasal OVA-IgA ASCs from smoke-exposed mice exhibited lower antigen retention despite expressing similar total levels of IgA during the acute post-immunization period. These findings firstly demonstrate that IgA production on a per-cell basis is similar between groups, and thus that the observed deficit in NALF OVA-IgA likely does not occur as a result of changes in antibody production rate by individual ASCs. In addition, they provide evidence that antigen-specific IgA produced in smoke-exposed mice is of lower avidity than that produced by room air controls, a finding confirmed by chaotropic ELISA analysis of NALF OVA-IgA. Notably, these data are in line with clinical observations that smokers more frequently produce low-avidity antibodies in response to human papillomavirus (HPV) vaccination compared to non-smokers42. However, to our knowledge such an observation has not been previously reported for sIgA.

Strikingly, mutational data seem at first to contradict avidity findings, in that a greater level of mutation occurred within CDRs and FWRs as a result of immunization in smoke-exposed mice compared to room air controls. However, it is possible that smoke exposure facilitates these mutations, either directly or indirectly, in a manner that precludes them from undergoing rigorous affinity-based selection against the immunization antigen. In turn, these mutations could have potentially detrimental effects on affinity, explaining the observed differences in antigen-binding capacity. Notably, one limitation of our sequencing approach is that it comprised the entire IgA repertoire rather than OVA-IgA ASCs specifically. However, OVA-specific IgA ASCs comprised more than 50% of the total IgA compartment in both room air- and smoke-exposed mice at this timepoint, suggesting that the mutational burden we observe is reflective of the antigen-specific compartment.

In assessing NALT germinal center activity, we also observed that smoke exposure upregulated the proportion of GCBCs at baseline, but attenuated the induction OVA-IgA GCBCs compared to room air controls. This suggests that the elevated mutational load observed in smoke-exposed immunized mice occurs either as a result of augmented germinal center activity at an earlier timeframe following primary immunization doses, or alternatively, does not occur as a result of events occurring after immunization. In this regard, cigarette smoke may promote SHM of B cells specific for smoke-associated antigens, which also demonstrate or develop cross-reactivity to OVA antigen, prior to immunization. Ultimately, further investigations are required to understand the manner in which cigarette smoke exposure dysregulates antigen-specific SHM and affinity maturation of sIgA responses in the URT.

In summary, our findings demonstrate for the first time that cigarette smoke exposure has the capacity to attenuate antigen-specific IgA responses induced in the upper airways of mice. Importantly, the mechanisms by which smoke exposure inhibits antigen-specific IgA induction in the upper airways are multifactorial, seemingly comprising deficits in ASC induction, homing, and transepithelial IgA transport (Fig. 7). These findings are of importance, as sIgA has been implicated in providing a first line of defense in precluding acquisition and carriage of respiratory viruses, and preventing the cellular release and local dissemination of viral particles4. Although one investigation demonstrated that intranasal live-attenuated influenza virus (LAIV) vaccination was similarly effective at generating protective serum haemagglutination-inhibiting antibodies in smoking individuals as compared to non-smokers43, it may be that cigarette smoke exposure specifically inhibits IgA-mediated defense against early asymptomatic upper airway infection, but not pulmonary defense mediated by IgG and cellular immune responses. In support of this notion, a recent study demonstrated that both cigarette smokers and e-cigarette users had a diminished IgA response to LAIV vaccination in the upper airways44, providing evidence that our findings are relevant to humans. Ultimately, understanding mechanisms by which cigarette smoke compromises antigen-specific IgA responses in the upper respiratory tract may provide valuable insight into why smokers are predisposed to infectious disease, and aid in identifying potential challenges for the design and implementation of intranasal vaccination strategies.

Fig. 7: Prospective mechanisms of cigarette smoke-mediated IgA inhibition in the upper respiratory tract.
figure 7

Mice exposed to cigarette smoke during the course of intranasal immunization with LPS/OVA demonstrated reduced induction of antigen (OVA)-specific IgA antibodies in the nasal and systemic circulation. These antibodies demonstrate reduced polyclonal avidity during the acute post-immunization period. (1) Evidence of impaired OVA-IgA ASC accumulation in the nasal mucosa, secondary lymphoid tissues, and bloodstream suggest an impairment in the activation these cells. Nasal OVA-IgA ASCs from smoke-exposed mice also demonstrate reduced OVA-binding capacity, suggesting that they possess a reduced antigen-binding affinity compared to room air controls. (2) Data also demonstrate that cigarette smoke exposure compromises the upregulation of VCAM-1 in the nasal mucosa following LPS exposure, suggesting that following generation in secondary lymphoid tissues, these cells may not be able to efficiently extravasate back into the nasal mucosa. (3) Finally, pIgR expression in the nasal mucosa was similarly impaired following immunization, strongly suggesting that IgA transepithelial transport may also be attenuated.

Full size image

Methods

Animals

6–10 week old female C57BL/6 mice were purchased from Charles River Laboratories (Montreal, QC, Canada), and housed in the McMaster Central Animal Facility. Mice were given ad libitum access to food and water, and subject to 12 h light/dark cycling. All experiments were approved by the Animal Research Ethic Board at McMaster University.

Cigarette smoke exposure

Mice were exposed to mainstream cigarette smoke of 12 3R4F research cigarettes (University of Kentucky; filters removed) over 50 min, twice per day, 5 days per week in a whole-body smoke exposure system (SIU48, Promech; Sweden). Details of the protocol were reported previously45. Duration of smoke exposure is detailed in the text. Control mice were exposed to room air only.

Intranasal Immunizations

Unanaesthetized mice were restrained, and a mixture of either purified Escherichia coli O111:B4 LPS or polyI:C (10 μg; Sigma Aldrich) and ovalbumin (OVA) (100 μg) in a combined volume of 10 μl was pipetted into the URT via the nares. This administration protocol has been shown to restrict delivery to the upper airways as opposed to the lungs46. Control mice received phosphate-buffered saline (PBS) vehicle alone. Specific experimental administration schedules are described in the text.

Sample Collection and Processing

At the timepoints described in the text, mice were anaesthetized with isoflurane and blood was collected via cardiac puncture. To collect nasal samples, mice were culled by exsanguination, and decapitated. The lower mandible was removed and nasal lavage (NAL) was performed by cannulating the nasopharynx and retrograde flushing the nasal passages with 300 μl of sterile PBS. After removing the cranial skin and the palate, the nasal-associated lymphoid tissue (NALT) was excised using a dissection microscope. Subsequently, the external nose was removed, and the skull bisected with scissors. The nasal mucosa (NM) was then excised with fine forceps. Cervical lymph nodes (CLNs; mandibular, accessory mandibular LNs) were extracted using a dissection microscope. Vaginal lavage (VAL) was performed pipetting 150 μl of PBS in and out of the vaginal opening, repeated twice.

Unheparinized blood was incubated at 37 °C for 1 h, microcentrifuged at 13,000 rpm for 10 min, and serum was collected. NAL, BAL, and VAL were microcentrifuged at 2500 rpm for 10 minutes to pellet cells. Serum, as well as nasal, bronchoalveolar, and vaginal lavage supernatant fluid samples (NALF, BALF, VALF) were stored at −80 °C for future analysis. Nasal mucosae and spleens were mechanically dissociated by abrasion against a 40μm filter. NALT and CLNs were mechanically dissociated by abrasion between glass slides. NALT cells were pooled within each group for staining and flow cytometry analysis due to low cellular yield. Bone marrow (BM) was extracted by removing the ends of one femur and flushing with sterile PBS. All samples were refiltered using 40 μm filters, centrifuged at 1400 rpm for 5 min, and resuspended in PBS. Spleen and BM cells were further treated with ACK lysis buffer. Total cells were counted for each sample using a haemocytometer or Countess™ Automated Cell Counter (Invitrogen).

ELISAs

To detect OVA-specific IgA, NUNC Maxi-Sorp 96-well plates (ThermoFisher Scientific) were coated with goat anti-mouse IgA (Southern Biotech) in carbonate-bicarbonate buffer (CBB; Sigma). Subsequently, plates were blocked with 1% bovine serum albumin (BSA; Sigma Aldrich; in PBS), and incubated with sample diluted in 1% BSA. All reagents herein were diluted in 1% BSA unless otherwise indicated. Adherent OVA-specific antibodies were detected using digoxigenated OVA (300 ng/ml), followed by anti-digoxigenin-POD (poly) Fab fragments (Roche). ELISAs were developed using tetramethylbenzidine (TMB; SurModics), and stopped with 2 N H2SO4. To detect OVA-specific IgG1 and IgM antibodies by ELISA, plates were coated with 4 μg/ml of purified OVA (Sigma Aldrich) in PBS. Following blocking and sample incubation as above, wells were incubated with biotinylated antibodies specific for the indicated isotype heavy chains (Southern Biotech) and detected using streptavidin-alkaline phosphatase (ThermoFisher) and 4-NPP (Sigma). To detect LPS-specific IgA, plates were coated with 50 μg/ml of purified Escherichia coli O111:B4 LPS in CBB, blocked and incubated with samples. Bound antibodies were detected with biotinylated anti-IgA, streptavidin-AP, and 4-NPP. Background from empty control wells was subtracted to acquire final O.D. values for all antigen-specific antibody ELISAs. To detect total IgA, plates were coated with anti-mouse IgA (Southern Biotech), blocked, and incubated with sample as above. Subsequently, wells were incubated with biotinylated anti-mouse IgA (Southern Biotech) followed by development with streptavidin-HRP and TMB. Concentrations were interpolated using a standard curve of purified mouse IgA (Southern Biotech).

To generate avidity indices, we performed avidity ELISAs47,48. ELISA plates were coated with OVA, blocked, and incubated with sample as above for 2 h. After washing, samples were treated with 50 µl of 4 M urea in PBS, or PBS alone, at room temperature under constant agitation for exactly 15 min. Urea was washed off and bound antibodies were detected using biotinylated anti-IgA, streptavidin-HRP and TMB. Avidity indices were calculated as (O.D.450 Urea / O.D.450 No Urea * 100) for each sample.

To detect CCL28 and VCAM-1, nasal mucosa and lungs were homogenized in 400 μl of RIPA buffer containing protease inhibitors. Bradford assay was used to determine the total protein content of each sample. Subsequently, samples were normalized to 0.25 μg/ml total protein, and assessed for CCL28 and VCAM-1 using commercial ELISAs (Biolegend and R&D, respectively) as per the manufacturer’s instructions.

Flow Cytometry

For flow cytometry, single-cell suspensions were transferred to a 96-well plate. Cells were stained for viability using LIVE/DEAD® Fixable Yellow Dead Cell Stain (Molecular Probes; 1:400 in PBS), and blocked with anti-CD16/32 (Biolegend; 1:100 in FACS buffer (0.5% BSA, 2 mM EDTA in PBS)). Subsequently, cells were stained with fluorophore-conjugated antibodies targeting specific antigens (Table S1) diluted in FACS buffer. Cells were next fixed and permeabilized using Cytofix/Cytoperm (BD Biosciences), and intracellular staining was performed using antibody cocktails and other stains diluted in Perm/Wash Buffer (BD Biosciences; Table S1). Fluorescence-minus-one (FMO) controls were used to determine thresholds for gating strategies. Data were collected using a BD LSRFortessa (BD Biosciences), and data analysis performed using FlowJo software (V10; TreeStar, Inc.). Absolute cell number quantification was performed by multiplying the frequency of the given population with regards to live singlet events by the total cell count for that sample.

IgA Sequencing

To quantify mutational load within the nasal IgA repertoire, high-throughput sequencing was performed. Briefly, at 3dpb the nasal mucosae were removed and homogenized in buffer RLT (Qiagen) containing β-mercaptoethanol. After centrifugation to remove cellular debris, RNA was isolated from lysates using an RNeasy kit (Qiagen) as per manufacturer’s instructions. 700 ng of total RNA per sample was used for the sequencing analysis, which was performed as per Turchaninova et al.21. Illumina adapters (NEBNext DNA Library Prep Set; NEB E6040) were added to the resulting cDNA products as per the manufacturer’s instructions. All DNA cleanup steps were performed using Monarch PCR & DNA Cleanup Kits (NEB T1030L). Asymmetric 400 + 200nt paired-end sequencing was performed on an Illumina Mi-Seq platform (v3). Sample demultiplexing and UMI-based assembly of full-length variable regions was performed using MIGEC software. V(D)J alignment against reference and clonal clustering was performed using Change-O package23 with default settings. Mutations were quantified within the full variable region using the R package ShazaM (version 0.2.0)49. Overall, following all processing steps we reconstructed a median of 69751 full length IgA reads and that were clustered into a median of 4484 IgA clones (Table S2).

RT-qPCR Analysis

To assess Vcam1 expression, the upper airways were lavaged with RLT buffer, and RNA was extracted from the resulting mucosal lysate as above. Subsequently, RNA was reverse transcribed into cDNA using Superscript II (Invitrogen), as per manufacturer’s instructions. cDNA was then quantified by qPCR using Taqman primers and probes (Vcam1 Mm01320970_m1; ThermoFisher Scientific) in a StepOnePlus real-time PCR system (Life Technologies Inc., Burlington, ON, Canada). Gene expression was determined using the ∆∆CT method. Vcam1 gene expression was normalized to that of the housekeeping gene Gapdh (Mm99999915_g1), and expressed as a fold change relative to the indicated control group.

Western Blotting

Nasal mucosae were homogenized in RIPA buffer (0.1% SDS, 0.5% sodium deoxycholate, 1% Igepal CA-630 in PBS) containing protease inhibitors (Roche, Complete Mini Protease Inhibitor Cocktail, 11836153001), and a Bradford assay was used to quantify total protein. 20 μg of protein per sample was loaded per well, and a 10% gel was used to resolve the proteins. Following gel resolution and transfer to nitrocellulose membrane, total protein was assessed using LICOR REVERT Total Protein Stain as per manufacturer’s instructions. After imaging the total protein stain, and blocking, the membrane was incubated with a primary monoclonal antibody specific for pIgR (AF2800, R&D), and, subsequently, a secondary detection antibody. Following imaging of pIgR bands, densitometry scores were determined using ImageStudio Lite. Scores from pIgR (~115 kDa) and secretory component (SC; ~95 kDa) bands were normalized to total protein stain bands.

Statistical analysis

GraphPad Prism 8 (GraphPad Software, Inc.) was used for statistical analyses. Two-group, single-variable comparisons were assessed by unpaired Student’s T-test. Mutational load data were assessed by Mann–Whitney U test with Bonferroni post-hoc analysis. Two-way ANOVAs adjusted for multiple comparisons using Tukey’s post-hoc test were used to assess all other multi-variable comparisons. Significance (*) is defined as p < 0.05.