Age-related differences in IL-1 signaling and capsule serotype affect persistence of Streptococcus pneumoniae colonization

Young age is a risk factor for prolonged colonization by common pathogens residing in their upper respiratory tract (URT). Why children present with more persistent colonization is unknown and there is relatively little insight into the host-pathogen interactions that contribute to persistent colonization. To identify factors permissive for persistent colonization during infancy, we utilized an infant mouse model of Streptococcus pneumoniae colonization in which clearance from the mucosal surface of the URT requires many weeks to months. Loss of a single bacterial factor, the pore-forming toxin pneumolysin (Ply), and loss of a single host factor, IL-1α, led to more persistent colonization. Exogenous administration of Ply promoted IL-1 responses and clearance, and intranasal treatment with IL-1α was sufficient to reduce colonization density. Major factors known to affect the duration of natural colonization include host age and pneumococcal capsular serotype. qRT-PCR analysis of the uninfected URT mucosa showed reduced baseline expression of genes involved in IL-1 signaling in infant compared to adult mice. In line with this observation, IL-1 signaling was important in initiating clearance in adult mice but had no effect on early colonization of infant mice. In contrast to the effect of age, isogenic constructs of different capsular serotype showed differences in colonization persistence but induced similar IL-1 responses. Altogether, this work underscores the importance of toxin-induced IL-1α responses in determining the outcome of colonization, clearance versus persistence. Our findings about IL-1 signaling as a function of host age may provide an explanation for the increased susceptibility and more prolonged colonization during early childhood.

Introduction demonstrate that Ply-mediated mucosal IL-1 signaling via the release of IL-1α is critical for clearance of otherwise persistent colonization and that infants are deficient in IL-1 signaling compared to adults. Our observations of reduced IL-1 signaling early in life provides mechanistic insight into the altered dynamics of pneumococcal colonization with age.

Ethics statement
All animal studies were performed in compliance with the federal regulations set forth in the Animal Welfare Act (AWA), the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health, and the guidelines of the New York University School of Medicine and the University of Pennsylvania Institutional Animal Use and Care Committee. All protocols used in this study were approved by the Institutional Animal Care and Use Committee at the New York University School of Medicine (protocol #160622 and #161219) and the University of Pennsylvania (protocol #804928).

Bacterial stains and culture
S. pneumoniae isolates serotypes 4 and 23F and previously described isogenic Ply mutants were used throughout the study [18,20,21]. In the serotype 4 background, these mutants included an unmarked, inframe deletion of ply, a point mutant expressing Ply W433F , which is deficient at oligomerization to form pores after membrane insertion, and a corrected mutant where the intact gene was restored (ply+). In the serotype 23F background, we used a mutant expressing Ply TL!AA , which fails to bind to cholesterol and insert into membranes [17,18]. The genotype of mutants was confirmed by sequence analysis of PCR products. The phenotype of all strains was confirmed using a previously described horse erythrocyte lysis assay [17,18]. The hemolytic activity of Ply-deficient and point mutants was less than 1% of WT levels. The serotype 4 and 23F parent strains showed no differences from one another in hemolytic activity or Ply protein levels as measured in Western blots. Type 4 and 23F isogenic capsule-switch strains 23F 4 (23F genetic background expressing the serotype 4 capsule) and 4 23F (4 genetic background expressing the serotype 23F capsule) were used to study serotype-dependent effects and were previously described [22]. All pneumococcal strains were grown in tryptic soy (TS) broth (BD) at 37˚C to an optical density of 1.0 at 620 nm (OD 620 ). Alternatively, pneumococci were incubated on TS agar plates supplemented with 100 μl of catalase (30,000 U/ml; Worthington Biomedical) and streptomycin (200μg/ml) at 37˚in 5% CO 2 overnight.
Pups at day 4 of life were infected with 10 3 −10 4 CFU of S. pneumoniae in 3 μl PBS by intranasal (i.n.) instillation divided over both nares without the use of anesthesia. At the time points indicated following challenge, mice were euthanized by CO 2 asphyxiation followed by cardiac puncture. The trachea was lavaged with 200 μl sterile PBS collected from the nares to determine URT colonization density of S. pneumoniae. A second URT lavage with 600 μL RLT lysis buffer (Qiagen) + 1% β-mercaptoethanol was performed to obtain host RNA from the URT epithelia for gene expression analyses.

RNA-sequencing
Uninfected infants, serotype 23F wild-type and ply-mutant infected pups (aged 1 week) were euthanized 3 days post-challenge and a URT lavage with RLT lysis buffer was obtained. RNA was extracted according to manufacturer's directions (RNeasy kit, Qiagen) and five replicates per age group were subjected to RNA sequencing (RNA-seq) using Hi-seq and the raw fastq reads were aligned to mm10 mouse reference genome using STAR aligner [24]. Fastq Screen was used to check for any contaminations in the samples and Picard RNA-seqMetrics was used to obtain the metrics of all aligned RNA-Seq reads. featureCounts [25] was used to quantify the gene expression levels. All alignments and read metrics are summarized in the supplementary data. The raw gene counts data were used for further differential expression analysis. To identify the differentially expressed genes (DEGs), DESeq2 R package [26] was used and results were subsequently analyzed using the online annotation tool DAVID Bioinformatics Resources [27]. The resulting genes with FDR < 0.05 were considered significant. Heatmaps were generated using pheatmap R package. RNAseq data are made available in the GEO repository under project accession number GSE116604.

ELISA
Anti-pneumococcal IgG titers were assessed in serum from uninfected and infected mice using whole bacterial cells as the capture antigen, as described previously with some adjustments [28]. Type 23F strain was grown to an OD 620 of 1 and washed with PBS. Pneumococcal cells were diluted to final OD 620 of 0.1 in coating buffer (0.015 M Na 2 CO 3 , 0.035 M NaHCO 3 ). Microtiter plates (2HB, Immunolon) were coated with 100 μL suspension/well at 4˚C overnight. The next day, the plates were blocked with 1% bovine serum albumin (BSA) in PBS at room temperature for 1 hour, after which the plates were incubated with serial serum dilutions in PBS at 4˚C overnight. Peroxidase-conjugated goat anti-mouse IgG (Jackson Immuno Research Laboratories) was applied and plates were incubated at 37˚C for 1.5h. Between incubation steps, plates were washed three times with 0.05% Brij-35 (Thermo Fisher Scientific) in PBS. Plates were developed using 100 μL/well substrate o-phenylenediamine dihydroxychloride (OPD, Thermo Fisher Scientific, 1 tablet in 7.5 ml H 2 O with 7.5 μL 30% H 2 O 2 ) and incubated at room temperature for 30 minutes in the dark. Reactions were stopped with 50 μl/well of 2M H 2 SO 4 and the absorbance was measured at 492 nm. Serum IgG titers were determined by calculating the dilution at which the absorbance was equal to an OD 492 of 0.1.

Quantitative RT-PCR
RNA extraction (RNeasy, Qiagen) from nasal RLT lavages and subsequent cDNA generation using High Capacity cDNA Reverse Transcriptase Kit (Applied Biosystems, Thermo Fisher Scientific) were performed as per manufacturer's instructions. Reaction samples contained 10ng cDNA and 0.5 μM primers in Power SYBR 1 Green PCR Master Mix (Applied Biosystems, Thermo Fisher Scientific) and samples were tested in duplicate. qRT-PCR reactions were run in a 384 well plate (Bio-Rad) using CFX384™ Real-Time System (Bio-Rad). Expression of Gapdh was an internal control and fold change in gene expression was quantified according to the ΔΔCt method [29]. The primer sequences used in this study are indicated in S1 Table.

Recombinant pneumolysin purification and treatment
As previously described, recombinant pneumolysin (PLY) and pneumolysoid (PdB), bearing the Ply W433F point mutation, were expressed in E. coli, after which bacterial cells were lysed using sonication [17,30]. The His-tagged proteins from cell suspensions were purified on a HisTrap column (GE Healthcare) by FPLC using Ä kta Start (GE Healthcare) using binding buffer (20 mM NaH 2 PO 4 , 500 mM NaCl, 20 mM Imidazole, pH 7.4) and elution buffer (20 mM NaH 2 PO 4 , 500 mM NaCl, 500 mM Imidazole, pH 7.4). Cell lysates were loaded on the column after which washing was performed to eliminate binding by non-specific proteins. His-tagged protein was recovered from the column by running a 70-100% gradient of elution buffer over the column after which the protein fractions were collected. Protein desalting was done using Amicon 1 Ultra-15 30 Kb centrifuge filters (Millipore). Protein concentration was assessed by Bradford protein assay (Bio-rad) and hemolytic activity was confirmed by horse erythrocyte lysis assay.
Pups infected with the serotype 4 ply-mutant at age 4 days received a first dose of PLY or PdB 1 day after challenge. Infants were treated with 100 ng protein/dose in PBS given by i.n. administration and control pups received PBS alone. Pups received daily treatment for 4 consecutive days and 5 days post-challenge pups were euthanized to determine pneumococcal colonization density and gene expression.

Treatment with IL-1 recombinant protein
Recombinant IL-1α (Peprotech) and IL-1β (eBioscience) protein was administered daily via i.n. treatment to 4-day old pups infected with the serotype 4 ply-mutant strain from day 1 through 3 post-challenge. Pups received 10 or 100 ng protein in 3 μl PBS, whereas PBS alone was administered to control animals. At day 4 post-challenge, infant mice were euthanized to obtain URT lavages to assess colonization density and for gene expression analyses.

Statistical analyses
All statistical analyses were performed using GraphPad Prism 7.0 (GraphPad Software Inc., SanDiego, CA). Colonization density values were transformed into logarithmic values. Unless otherwise specified, differences in colonization and gene expression were determined using ttest or one-way Anova with Sidak's multiple comparison test.

Lack of pneumolysin toxin allows persistent colonization of S. pneumoniae
Infant mice have an increased density and duration of pneumococcal carriage compared to adult mice, recapitulating the effect of age during natural colonization [31]. To determine whether loss of pneumolysin (ply-) prolongs pneumococcal colonization in infant mice, pups were challenged i.n. at day 4 of age with a serotype 23F isolate or ply-deficient mutant. Clearance of the wild-type strain required approximately 9 weeks by which time the pups had become adults (Fig 1A). In contrast, at this time point colonization of the ply-strain remained at >10 4 CFU/ml. To determine whether the effect of pneumolysin on colonization was independent of serotype, we infected pups i.n. at day 4 of age with a serotype 4 isolate or ply-deficient mutant. Although the wild-type serotype 4 isolate was cleared earlier than the wild-type 23F strain, at approximately 6 weeks post-challenge, clearance of the serotype 4 ply-deficient strain was significantly delayed compared to the wild-type serotype 4 strain ( Fig 1B). Correction of the ply deletion (ply+) restored clearance to wild-type levels confirming the contribution of pneumolysin in attenuating colonization duration ( Fig 1C). Together these results indicate that expression of pneumolysin significantly decreases duration of pneumococcal colonization. At 6 weeks post-challenge, most mice remained colonized with the non-hemolytic point mutant (W 433 F), suggesting that the pore forming activity of pneumolysin contributes to clearance of pneumococcus ( Fig 1C). The requirement of toxin-host cell interaction was further supported by persistent colonization of the ply TL!AA mutant, which expresses a toxin deficient in a prior step, attachment to cholesterol, in the serotype 23F background ( Fig 1D). Additionally, i.n. dosing of recombinant pneumolysin (PLY), but not the toxoid containing the W 433 F modification (PdB), accelerated clearance of the ply-deficient mutant ( Fig 1E). We concluded that pneumolysin was both necessary and sufficient for pneumococcal clearance in infant mice.

Host responses involved in pneumolysin-mediated clearance
Clearance of pneumococcal colonization is generally thought to require adaptive immune responses consisting of specific immunoglobulin or T H 17 immunity [32][33][34][35]. Colonization increased levels of pneumococcal-specific IgG, however these did not differ between wild-type Clearance of the serotype (A) 23F and (B) 4 was monitored until 9 or 6 weeks post-challenge, respectively. Pneumococcal colonization density of (C) serotype 4 strains: wild-type, ply-, ply W433F (deficient pore formation), and corrected mutant ply+ at 6 weeks post-infection, and of (D) serotype 23F wild-type, ply-and ply TL!AA (deficient cholesterol binding) at 9 weeks after infection. (E) Pups infected with the serotype 4 ply-received daily treatments with 100 ng recombinant PLY or PdB (ply W433F ) starting at 24 hours until 4 days after infection. Control animals received PBS dosing. Colonization density of S. pneumoniae was determined 5 days post-challenge. Groups represent n = 7-11 animals with mean ±SEM. Dotted line represents the lower limit of detection. Significance is indicated by � , P < 0.05; �� , P < 0.01, ��� , P < 0.001. and ply-deficient mutant colonization (Fig 2A). Moreover, mice deficient in the generation of specific antibody (muMT -/-) cleared colonization of wild-type and ply-strain similar to C57BL6 wild-type mice ( Fig 2B). Il17a expression was increased during colonization, but expression did not differ between WT and the ply-deficient mutant at day 7 and 21 post- Fold-change was calculated from age-corrected mock-infected controls. Groups represent n = 5-11 animals with mean ±SEM. Dotted line represents the lower limit of detection. Significance is indicated by �� , P < 0.01; ��� , P < 0.001. (D) At day 21 after infection, five wild-type, ply-, and mock challenged pups received RLT lavages to obtain RNA for RNA-seq analysis. Heat-map of top 100 differentially regulated genes comparing wildtype to ply-demonstrating relative gene expression in log2 foldchange with increased expression in red and decreased expression in blue.  2C). These observations suggested that the increased persistence in infants may largely depend on innate immune responses.
To more broadly explore how pneumolysin mediates clearance of otherwise persistent colonization, we performed an RNA-Seq screen on RNA isolated from URT lavages of infant mice infected with the wild-type or ply-deficient mutant at 21 days post-challenge, the time point at which clearance of the serotype 23F isolate is initiated. Large numbers of host genes were significantly affected by pneumolysin expression. For the 100 most differentially expressed genes (Fig 2D), pneumococcal colonization without pneumolysin resembled mock-infected controls indicating pneumolysin affects critical host responses that promote clearance.

IL-1 signaling is essential for clearance of persistent colonization
We then investigated the role of innate immune signaling through TLR2, NOD2, IFNAR and IL-1R. These sensors of pneumococcal PAMPs or cytokines are expressed at significantly higher levels in wild-type versus mock infected mice in the RNA-Seq analyses and were previously shown to contribute to pneumococcal clearance in adult mouse models [13,14,19,36]. Colonization of the wild-type and ply-mutant was assessed at 9 weeks post-challenge in wildtype versus knockout mice (Fig 3A-3C and 3E). Only for Il1r -/pups colonization of the wildtype strain persisted and was indistinguishable from the ply-mutant.
IL-1α and IL-1β are two cytokines released under different circumstances that exert identical biological effects downstream of the IL-1 receptor binding [37]. Previous in vitro studies showed that pneumolysin promotes release of both IL-1α and IL-1β by S. pneumoniae infected macrophages [19,38]. Daily i.n. dosing of recombinant IL-1α or IL-1β accelerated clearance of ply-colonized pups (Fig 3F and 3G). Quantitative RT-PCR (qRT-PCR) of RNA isolated from URT lavages showed that pneumococcal colonization minimally upregulated Il1a expression, whereas the difference in Il1a expression between wild-type and ply-strains did not reach significance (Fig 3H). This was not unexpected, since unlike IL1β, the damage associated molecular pattern (DAMP) IL-1α is constitutively expressed in epithelial, endothelial and hematopoietic cells in the airway epithelium, although its transcription can be upregulated by strong inflammatory stimuli [39,40] In contrast, colonization of the wildtype, but not the ply-mutant, significantly increased expression of Il1b in the URT (Fig 3I). Although IL-1α levels in URT lavages were detectable by ELISA, there was no detectable difference between infected and uninfected pups. For IL-1β, its levels in the highly diluted URT lavages were below the level of detection in all treatment groups despite using a sensitive ELISA. IL-1β is generated from pro-IL-1β as a result of caspase 1 activation by the inflammasome. In vitro and in vivo models using S. pneumoniae or pneumolysin support a role of the NLRP3 inflammasome in this process [41][42][43][44]. However, persistence of the wild-type strain was unaffected in Nlrp3 -/pups (Fig 3J), suggesting a role for another inflammasome or noncanonical pathway for IL-1β processing or redundant role for IL-1β [37]. To determine the roles of individual IL-1 cytokines in clearance, we infected IL-1α and IL-1β deficient pups with the wild-type S. pneumoniae or the ply-deficient mutant (Fig 3K). Loss of IL-1α, but not IL-1β, resulted in impaired clearance of the wild-type strain. Colonization density at 9 weeks post-challenge of the wild-type strain was slightly less (<1 log) compared to its ply-mutant in Il1a -/infant mice (Fig 3J), suggesting that factors other than IL-1α play a minor role in Plymediated clearance. We concluded that IL-1 signaling is necessary and sufficient to prevent persistent colonization of wild-type pneumococci in infant mice, with a dominant role for IL-1α.

Pathways involved in IL-1-mediated clearance driven by pneumolysin
The RNA-Seq screen on RNA isolated from infant URT lavages at 21 days post-challenge confirmed the importance of IL-1 signaling in clearance of persistent colonization. Pathway analyses demonstrated increased expression of gene cluster Signal Transduction Through IL-1 Receptor following colonization with S. pneumoniae, while these gene transcripts were less stimulated in absence of pneumolysin (Fig 4A). This was confirmed by qRT-PCR showing decreased Ccl2 and IL1rn expression during ply-compared to wild-type colonization at 21 days (D21) post-challenge (Fig 4B and 4C). Furthermore, the RNA-Seq data revealed two major pathway upregulated during colonization potentially contributing to clearance of pneumococcal colonization, Chemokine-mediated signaling pathway and Phagocytosis (Fig 4D  and 4E). These pathways were also less activated during colonization by the ply-deficient strain. Clearance of pneumococcal colonization requires chemokine-dependent recruitment of professional phagocytes into the lumen of the URT and the expression of multiple chemokines were upregulated by colonization [19,34]. There was, however, no non-redundant role for CCR2 in clearance in infant mice, although this chemokine receptor was previously shown to be important in clearance in adult mice (Fig 3D) [14]. By comparing transcript levels using qRT-PCR on RNA isolated from URT lavages, we confirmed that representative genes revealed in these pathways analyses involved in phagocyte recruitment (chemokine Cxcl2), activation (leukocyte integrin CD11b), activity (nitric oxide synthase, Nos2), and regulation (Fc gamma receptor, FcγR3) were all increased in expression during wild-type compared to ply-colonization (Fig 4F-4I). Additionally, i.n. administration of recombinant IL-1α and IL-1β were both sufficient to enhance transcription of FcγR3 ( Fig 4J). As was the case for ply-infection, transcription of FcγR3 following wild-type colonization was also impaired in Il1r -/mice ( Fig 4K).
These observations were consistent with a clearance mechanism dependent on professional phagocytes that requires IL-1 signaling.
Intranasal administration of recombinant Ply upregulated expression of Il1b, but not Il1a, (Fig 5A and 5B) and markers of phagocyte recruitment or activity, including Cxcl2, Nos2, Fcγr3 (Fig 5C-5F). These increases in gene expression were not observed for toxoid PdB, deficient in pore-formation. These findings confirm the critical role of pneumolysin cytotoxicity in inducing IL-1 responses that promote clearance of otherwise persistent pneumococcal colonization.

Capsule serotype affects persistence of S. pneumoniae colonization
In Fig 1, we showed differences in the duration of colonization for two isolates differing in CPS (serotypes 4 and 23F) (Relevant data juxtaposed in Fig 6A). To determine whether this difference was due to IL-1 signaling, we compared the IL-1 stimulating capacity of the two isolates. We found that serotype 4 and 23F isolates induced similar expression of Il1a, Il1b, and IL-1 signaling related transcripts at 7 and 21 days post-challenge (Fig 6B and 6C, and S2 Fig), a result that could be due to similar levels of pneumolysin expression and hemolytic activity in these two isolates. To determine whether CPS type or bacterial genetic background is the determinant of isolate-specific persistence, we made use of capsule-switch strains of these two isolates [30,45]. We infected pups at day 4 of life with the isogenic strains 23F 4 (23F isolate expressing 4 CPS) and 4 23F (4 isolate expressing the 23F CPS). The capsule-switch strains colonized equivalently during early infection (Fig 6D). However, at 6 weeks post-challenge, the 4 23F strain persisted whereas the 23F 4 strain was cleared (Fig 6E), demonstrating that strains carrying the 23F CPS are more persistent, regardless of genetic background. Thus, serotypedependent differences in clearance appear to act through processes downstream or independent of IL-1 signaling.

Age-related IL-1 responses during early pneumococcal colonization
Given the importance of IL-1 in clearance of otherwise persistent colonization, we questioned whether differences in IL-1 signaling underlie the age-dependent susceptibility to S. pneumoniae colonization. We previously showed that Il1r -/adult mice have reduced numbers of neutrophils during early colonization, fewer macrophages later in carriage, and prolonged bacterial colonization [19]. We used qRT-PCR to transcriptionally profile the IL-1 signaling pathway in the URT of uninfected infant (1 week of age) and adult (8 weeks of age) mice. As hypothesized, qRT-PCR confirmed significantly decreased expression of multiple IL-1-related signaling transcripts in uninfected infant compared to adult mice (Fig 7A-7J). Although colonization increased its expression, the level of Il1b expression in colonized infant mice only reached that of uncolonized adult mice (Fig 7K). In adult mice with increased IL-1 responses at baseline, colonization of Il1r -/mice at 3 days post-challenge resulted in impaired initial clearance, consistent Chemokine-mediated signaling and (E) Phagocytosis that are significantly upregulated following pneumococcal colonization. qRT-PCR was used to measure expression of (F) Cxcl2, (G) Cd11b, (H), Nos2, and (I) Fcyr3 at 21 days post-inoculation. Fold-change compared to mock-challenged age-controlled animals. (J) Fold change in Fcyr3 expression in serotype 4 ply-colonized pups following IL-1α and IL-1β dosing as compared to PBS-treated control pups at 4 days postchallenge. (K) Fold change in Fcyr3 expression in 23F colonized C57BL6 and Il1r -/pups at 21 days post-inoculation. Groups represent n = 7-11 animals with mean ±SEM. Significance is indicated by � , P < 0.05; �� , P < 0.01. https://doi.org/10.1371/journal.ppat.1007396.g004 with results of prior studies showing the importance of IL-1 signaling in adults at 14 days postinoculation [19] (Fig 7L). Loss of IL-1α, but not IL-1β, led to impaired initial clearance in adult mice (Fig 7M). Thus, the importance of IL-1 signaling, and in particular, the role of IL-1α could account for why clearance in infancy is delayed until reaching adulthood. Accordingly, for infant mice with dampened IL-1 signaling at baseline there was no effect of the absence of the IL-1R at 3 days post-challenge ( Fig 7N). As shown in Fig 3, stimulation of robust responses with recombinant IL-1 cytokines was sufficient to reduce colonization density in infants during early colonization, suggesting the defect in infants is not due to an inability to respond to IL-1 cytokines (Fig 3E and 3F). Together these results support that diminished IL-1 signaling during infancy enhances susceptibility to S. pneumoniae colonization.

Discussion
Although epidemiological studies have shown that carriage of S. pneumoniae is significantly prolonged in young children, indicative of a commensal lifestyle, the mechanisms underlying IL-1 responses in pneumococcal persistence pneumococcal persistence remain unknown. Here we showed that IL-1 signaling was required for efficient clearance but is relatively deficient in infants compared to adults. Prolonged colonization of S. pneumoniae was shown to be mediated by either loss of its sole toxin, pneumolysin, or through repression of the IL-1 pathway. While studies in vitro have associated pneumolysin expression with induction of IL-1 responses, our results provide in vivo evidence that pneumolysin drives IL-1 signaling through IL-1α to promote clearance of S. pneumoniae in the URT. Consequently, absence of pneumolysin or loss of the IL-1 signaling pathway, either by genetic ablation of the IL-1 receptor, IL-1α or young age, was sufficient to promote persistent pneumococcal colonization of the URT. Prompt clearance, therefore, does not appear to be initiated until IL-1 responses mature (>21 days of age) and is dependent on the presence of pneumolysin.
Several aspects of the effect of pneumolysin merit further comment. Pneumococcal colonization with the Ply point mutant deficient in pore-formation showed a clearance defect. However, a greater effect was observed for the Ply mutant unable to bind membrane cholesterol, an earlier step in its activity. Apparently, binding of Ply to membrane cholesterol also contributes to perturbation of cell homeostasis as proposed from in vitro studies and this impacts toxinmediated clearance [46]. Without the toxin, there appears to be little stimulation of host responses that drive eventual clearance. Our findings are consistent with prior reports that the gradual process of Ply-mediated clearance requires a mild sustained phagocyte influx for clearance [19,34]. This raises the question of why the organism expresses a toxin that precipitates its own clearance. Although the expression of Ply may be disadvantageous for S. pneumoniae colonization duration in the current host, inflammation induced by the toxin is necessary for pneumococcal shedding at levels sufficient for transmission to a new, susceptible host [18]. Some naturally circulating pneumococcal strains lack ply or express low or non-cytolytic Ply variants [47,48]. These might have a deficit in transmission, but an advantage in carriage duration, although natural colonization dynamics of these strains have not yet been assessed. In Fig 3K  we show that in absence of IL-1α the wild-type strain colonizes at lower levels than the plymutant, indeed suggesting immune mechanisms other than IL-1 also contribute to clearance.
A caveat of this study is that we were unable to measure colonization-induced differences in IL-1 protein in the URT, perhaps because of the mild, gradual nature of the inflammatory process requiring many weeks to months for clearance to complete. Our colonization model contrasts with more dramatic infections where such events transpire over days. Pulse dosing with recombinant Ply, however, was sufficient to facilitate more rapid clearance. In vitro studies have reported that during pneumococcal infection Ply can regulate IL-1 responses by inducing rapid cell death through necroptotic or pyroptotic pathways due to disruption of the cell membrane [49][50][51]. Recently, pore-forming toxins including Ply were found to trigger necroptosis, the major cell death pathway in respiratory epithelial cells, in mice and nonhuman primates during bacterial lung infection [50]. In in vitro macrophage models, Ply poreformation also causes a passive release of alarmin IL-1α following rapid cell death [38,52]. Our RNA-seq analyses show more functional clustering of cell death pathways, including necroptosis and apoptosis, following pneumococcal colonization. The IL-1α precursor protein, which is constitutively expressed in many cell types at the mucosal barrier, including both epithelial and myeloid lineages [38-40, 52, 53], does not require cleavage for binding to the IL-1 receptor [39,54], thus Ply-mediated passive release of IL-1α precursor could directly activate IL-1 signaling. In contrast, IL-1β precursor expression is induced in response to TLR stimulation, TNFα, and IL-1 itself, and requires active processing in order to bind the IL-1 receptor [37]. Canonical IL-1β processing and release is mediated by caspase 1 following activation of an inflammasome [37,55,56]. In contrast, cell death induced by bacterial virulence factors that result in the release of IL-1α protein does not require the inflammasome but may depend on caspase-11 [57]. Despite the extensive evidence of Ply and necroptosis in activating the NLRP3 inflammasome, we observed that Ply-mediated clearance of pneumococcal colonization was NLRP3 independent [42-44, 51, 58]. This result, along with the lack of a contribution of TLR2 upstream of IL-1β is consistent with a dominant role for IL-1α, which does not require increased expression or the inflammasome for its activity [19]. Similar to IL-1β, signaling downstream of cytosolic sensing (via NOD2 or Type I Interferons) affects clearance in adult mice but showed no contribution in infant mice, which exhibit more persistent colonization [14,36]. Ply-dependent cytosolic sensing may only occur following uptake by professional phagocytes and, as noted above, the recruitment of these is attenuated in infant compared to adult mice [17,31]. By dosing with recombinant pneumolysin the lack of a secretion mechanism for the toxin would have been bypassed potentially allowing for direct effects on nonprofessional phagocytes, including epithelial cells.
Our study raises the question of how signaling downstream of IL-1α affects carriage. IL-1-dependent activation of chemokines from neighboring nonhematopoietic cells or tissue-resident macrophages triggers the recruitment and activation of inflammatory hematopoetic cells to the site of damage [39]. This in turns initiates a positive feedback loop whereby the recruited myeloid cells respond to the inflammatory process by the release of more IL-1 cytokines. This loop could explain why we observed increased transcription of IL-1β, a cytokine transcribed and released predominantly by cells of hematopoetic origin. We were unable to detect significant differences in numbers of neutrophils or monocytes/macrophages in URT lavages by flow cytometry, perhaps due to the relatively mild and prolonged nature of the inflammatory process. Additionally, IL-1 signaling downstream of IL-1α has potent effects on phagocytic cells. Transcriptional profiling during persistent pneumococcal colonization demonstrated pneumolyin-dependent upregulation of factors involved in both phagocyte recruitment and function. IL-1 signaling also contributes to the differentiation of IL-17 + T cells and could impact eventual Th17-mediated pneumococcal clearance when these responses mature [59].
Colonization persistence was also found to depend on serotype, which is in agreement with clinical observations of serotype-dependent colonization duration in young children [5]. Serotype-specific differences in colonization duration in our model were not attributable to differences in IL-1 signaling. Isolates of serotype 23F and 4, which are cleared in 9 and 6 weeks, respectively, that express equivalent amounts of Ply generated similar IL-1 responses. Differences in physical properties of the capsule type or its amount may affect processes downstream of IL-1 signaling, such as the deposition of opsonins (complement and antibody) or uptake by professional phagocytes as previously documented [4,37,60,61].
The attenuation of IL-1 signaling during infancy could account for why initiation of clearance is delayed until mice are approximately 25 days of age. The association of young age with impaired IL-1 responses observed from our mouse model provides a possible mechanism for enhanced susceptibility to S. pneumoniae during early childhood. However, factors regulating IL-1 signaling during early life remain unknown. We have shown previously that infant mice have an elevated URT mucosal inflammatory profile that depends on the presence of a microbiota [31]. The high mucosal expression of chemokines in infants decreases the gradient for phagocyte recruitment to the airway lumen and delays clearance of colonizing pneumococci. The situation with IL-1 signaling is different as we measured reduced expression of genes in the IL-1 pathway at baseline as compared to adult mice. One possibility is that infants have blunted IL-1 responses to allow for acquisition and establishment of a stable URT microbiota. Alternatively, it seems plausible that the imbalanced infant microbiota could underlie the repressed URT expression of IL-1 signaling genes, or blunted IL-1α responses, and is responsible for increased susceptibility of infants to URT pathogens. Studies in adult mice have shown that the microbiota affects expression of IL-1β precursor protein and is important for regulation of mucosal defense; whereas the impact of the microbiota on IL-1α-dependent effects have yet to be explored [62,63]. Alternatively, metabolic and nutritional differences with age could affect mucosal inflammatory responses directly or indirectly through temporal changes in the microbiota [64].
The importance of IL-1 signaling in clearance of otherwise persistent colonization may have broad relevance to other systems [65][66][67][68]. IL-1α and IL-1β were found to have overlapping and non-redundant roles in bacterial clearance during lung infection with Legionella pneumophila [57]. Lack of IL-1 signaling, specifically IL-1α, during mycobacterial lung infection led to an inability to control bacterial replication and earlier mortality [66]. Additionally, both IL-1α and IL-1β were necessary for phagocyte recruitment and function to control lung infection of Aspergillus fumigatus [68]. The importance of intact and robust IL-1 responses is further reflected by the increase in severe bacterial infections seen in individuals with inborn errors involving IL-1 signaling or in patients receiving anti-IL-1 treatment for inflammatory diseases [69][70][71][72][73]. In particular, deficiencies in IL-1 signaling, including in IRAK4, Myd88 and NEMO, are associated with susceptibility to severe and recurrent pneumococcal infections during childhood [69,70]. The elderly also suffer from frequent pneumococcal disease. Pneumococcal colonization is prolonged in aged mice [72][73][74][75], and in vitro and in vivo studies demonstrate reduced IL-1 protein release by aged human and mouse monocytes, and in the aged lung, following infection with S. pneumoniae [76,77]. Differences in IL-1 signaling, therefore, may be relevant to a number of vulnerable populations.