Influenza A Virus Infection Predisposes Hosts to Secondary Infection with Different Streptococcus pneumoniae Serotypes with Similar Outcome but Serotype-Specific Manifestation

Abstract

Influenza A virus (IAV) and Streptococcus pneumoniae are major causes of respiratory tract infections, particularly during coinfection. The synergism between these two pathogens is characterized by a complex network of dysregulated immune responses, some of which last until recovery following IAV infection. Despite the high serotype diversity of S. pneumoniae and the serotype replacement observed since the introduction of conjugate vaccines, little is known about pneumococcal strain dependency in the enhanced susceptibility to severe secondary S. pneumoniae infection following IAV infection. Thus, we studied how preinfection with IAV alters host susceptibility to different S. pneumoniae strains with various degrees of invasiveness using a highly invasive serotype 4 strain, an invasive serotype 7F strain, and a carrier serotype 19F strain. A murine model of pneumococcal coinfection during the acute phase of IAV infection showed a significantly increased degree of pneumonia and mortality for all tested pneumococcal strains at otherwise sublethal doses. The incidence and kinetics of systemic dissemination, however, remained bacterial strain dependent. Furthermore, we observed strain-specific alterations in the pulmonary levels of alveolar macrophages, neutrophils, and inflammatory mediators ultimately affecting immunopathology. During the recovery phase following IAV infection, bacterial growth in the lungs and systemic dissemination were enhanced in a strain-dependent manner. Altogether, this study shows that acute IAV infection predisposes the host to lethal S. pneumoniae infection irrespective of the pneumococcal serotype, while the long-lasting synergism between IAV and S. pneumoniae is bacterial strain dependent. These results hold implications for developing tailored therapeutic treatment regimens for dual infections during future IAV outbreaks.

INTRODUCTION

Infection with secondary bacterial pathogens is attributed to be the major cause of excessive mortality during influenza A virus (IAV) outbreaks. This lethal synergism has been recognized as early as during the 1918-1919 IAV pandemic with an estimated global death toll of 50 to 100 million (1, 2). Retrospective studies disclosed that 71% of the fatal cases during this pandemic were positive for Streptococcus pneumoniae (also called pneumococcus), providing the first epidemiological evidence for viral-bacterial coinfections (2). A clear predisposition to bacterial disease was also evident in all of the succeeding influenza pandemics, including the more recent 2009 H1N1 outbreak, which had a 10 to 55% higher incidence of hospitalizations and mortality due to bacterial pneumonia (3). Pneumococcal colonization is transient and asymptomatic in immunocompetent individuals and most commonly occurs in early childhood (4). At the same time, however, pneumococci are able to cause a variety of diseases ranging from mild sinusitis and otitis media to more-severe infections like sepsis and meningitis. Even though the introduction of the polyvalent pneumococcal conjugate vaccines (PCV) has reduced the incidence of childhood carriage and disease for the vaccine serotypes (3–5), 1.6 million people die from pneumococcal infections annually, with the majority being children under the age of 5 (6). The development of pneumococcal disease depends on both bacterial factors, such as the pneumococcal capsular type or serotype, and the host innate immune response (7, 8). S. pneumoniae expresses a plethora of virulence factors, of which the encapsulating polysaccharide layer is the most important and best studied so far (9, 10). The capsule confers antiphagocytic and antiopsonophagocytic properties that determine the invasive pneumococcal disease (IPD) potential of different serotypes (11). IPD is defined as the recovery of S. pneumoniae from a normally sterile site such as the blood or brain (7, 12, 13). To date, 97 distinct serotypes have been described based on the unique chemical and immunogenic properties of their capsule (10, 14), and these serotypes can be divided into invasive as well as noninvasive/carrier serotypes (15, 16). Otherwise noninvasive serotypes are able to lethally infect immunocompromised patients, reflecting the impact of host immunity on the IPD potential (8, 17). Once pneumococci enter the airways, innate immune responses are initiated by lung resident alveolar macrophages (AMs) (18), which release proinflammatory cytokines and chemokines to recruit proinflammatory cells such as polymorphonuclear cells (PMNs) and mononuclear phagocytes into the lung parenchyma and alveoli to contain the infection (19).

Several studies have illustrated a multifactorial nature of IAV-S. pneumoniae copathogenesis with a plethora of underlying mechanisms (3, 20, 21). These include virus-mediated immune modulations such as aberrant inflammatory cell recruitment and function as well as increased cell death, often leading to changes in the antipneumococcal cytokine and chemokine responses (22, 23). Apart from the often devastating effect on antibacterial responses observed during acute influenza, some reports have demonstrated long-term immune defects. These include impaired neutrophil influx due to sustained desensitization of AMs and the induction of an immune-suppressive state during recovery (3, 24, 25). However, the mechanisms underlying enhanced susceptibility to S. pneumoniae following IAV infection are not fully understood and reports are at times contradictory. One major limitation of past studies is the use of singular S. pneumoniae strains despite large differences in pathogenesis. Therefore, it remains unclear if the identified mechanisms generally apply to strains of different pneumococcal serotypes. Importantly, blood cultures from living cases of the 1918 IAV pandemic revealed a higher prevalence of less invasive serotypes in secondary pneumococcal infections than the common pathogenic serotypes found at that time (2, 26). This observation implicates a preference for otherwise colonizing strains to cause severe infections in IAV-infected individuals and strongly supports the hypothesis that a preexisting IAV infection influences the pathogenic effect of these strains.

In this study, we used an S. pneumoniae-IAV coinfection mouse model to address host susceptibility to selected S. pneumoniae strains with various IPD potential during the acute and recovery phases of IAV. Our data show that acute IAV infection leads to equally fatal outcomes for all the tested S. pneumoniae strains at otherwise sublethal doses. At the same time, however, we detected S. pneumoniae strain-specific changes in the underlying innate immune responses that presumably contributed to lung immunopathology. Following recovery from IAV infection, there were no significant effects on survival following coinfection with any of the tested S. pneumoniae strains. Nevertheless, pneumococcal growth in the airways and systemic dissemination were enhanced in a strain-specific manner late after IAV infection.

MATERIALS AND METHODS

Mice.

Seven- to 8-week-old C57BL/6JOlaHsd female mice were purchased from Harlan Winkelmann (Borchen, Germany) and Harlan Laboratories (Venray, Netherlands). Mice were housed in a specific-pathogen-free environment according to the guidelines of the regional animal care committees. All the experiments were approved and conducted in accordance to the guidelines set by the local animal welfare and ethics committees for the Helmholtz Centre for Infection Research (Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit) and the Karolinska Institute (KI; Stockholms Norra Djurförsöksetiska Nämnd).

Bacterial and viral preparations.

For the S. pneumoniae infections, the serotype 4 strain TIGR4 (ATCC BAA-334), a serotype 19F strain (BHN100), and a 7F strain (BHN54) were used. All strains were obtained from the laboratory of B. Henriques-Normark (KI). The bacteria were grown to the mid-logarithmic growth phase in prewarmed Todd-Hewitt yeast (THY) medium (Todd-Hewitt broth [Sigma-Aldrich, Germany] supplemented with 1% yeast extract [Roth, Germany]) in a water bath at 37°C. Bacteria were harvested at an optical density at 620 nm of 0.35 and 0.50 (for T4 and 19F/7F, respectively) by adding 10% (vol/vol) glycerol (Roth, Germany) and were frozen at −70°C. For animal infections, the frozen stocks were thawed, centrifuged at 6,081 × g at room temperature, and washed once in 1 ml phosphate-buffered saline (PBS; Gibco, United Kingdom) before they were diluted to the desired concentration. The challenge dose was confirmed by plating 10-fold serial dilutions on blood agar plates (Columbia agar with 5% sheep blood; BD Diagnostic Systems, Germany) that were incubated overnight at 37°C and 5% CO₂. At KI, the blood agar plates and THY medium were produced by the Karolinska Microbiology Laboratory (Solna, Sweden). For viral challenges, influenza A/PR/8/34 virus (H1N1; PR8) was produced in Madin-Darby canine kidney (MDCK) cells as described previously (27). The 50% tissue culture infectious dose (TCID₅₀) of the viral stock was determined by incubating 10-fold dilutions of virus stock solution on MDCK cells cultured in Dulbecco's modified Eagle medium (DMEM; Life Technologies, Germany) supplemented with 0.0002% trypsin (Sigma-Aldrich, Germany) and 1% penicillin-streptomycin (ThermoFisher, USA). After 5 days, 0.5% chicken red blood cells were added and the TCID₅₀ was calculated from the observed agglutination using the endpoint calculation described by Reed and Muench (28).

Infection models.

Mice were weighed and anesthetized through intraperitoneal administration of a mixture of ketamine and xylazine in PBS (100 mg/kg ketamine, 10 mg/kg xylazine). For bacterial challenges, mice were placed on their back on an intubation slope and the larynx was illuminated by an external cold-light source. A dose of 1 × 10⁶ CFU S. pneumoniae in 25 μl PBS was instilled into the laryngopharynx for aspiration to the lower respiratory tract (LRT). For IAV challenges, anesthetized mice were held upright with the head tilted back slightly, and a dose of 0.31 TCID₅₀ in 25 μl of PBS was administered dropwise to each nostril using a pipette. For all survival experiments following bacterial mono- or coinfection, mice were monitored three to six times per day and scored for the following parameters: body weight, movement, posture, piloerection, respiration, eye discharge, redness of the eye conjunctiva, and response to stimulus. Moribund animals with severe symptoms of any one or a combination of the aforementioned parameters were euthanized, and the infection was considered lethal. For all IAV monoinfections, mice were monitored and weighed daily. The humane endpoint was set at 75% of the original body weight.

Assessment of the organ-wide bacterial burden.

To obtain nasopharyngeal lavage fluid, the trachea was exposed and the nasopharynx was flushed once with 1 ml PBS from the trachea toward the nasal cavity. The lavage fluid was collected from the nostrils. For bronchoalveolar lavage fluid (BALF), the lungs were flushed once with 1 ml PBS. For lung homogenates, the lungs were then aseptically excised and homogenized in 1 ml PBS through a 100-μm filter (Corning Inc., USA) using a syringe plunger. Five microliters of blood was collected from the tail vein and diluted in 45 μl PBS for plating. Bacterial loads were determined by plating serial dilutions of the samples on blood agar plates. CFU were counted manually after incubating the plates for 16 to 18 h at 37°C with 5% CO₂.

Quantitative RT-PCR for viral load.

Lungs were perfused with 10 ml PBS through the heart, excised, and stored at −70°C in RNALater solution (Ambion, USA). RNA was extracted using the RNeasy minikit (Qiagen, Germany) according to the manufacturer's protocol. One microgram of RNA was transcribed into cDNA using the SuperScript III first-strand synthesis System (Invitrogen, USA). The purity and concentration of cDNA were verified by performing a PCR for the housekeeping gene rps9. The quantitative reverse transcription (RT)-PCR performed to detect the absolute number of nucleoprotein (NP) copies in the samples was carried out on an ABI PRISM 7500 cycler (Applied Biosystems) using 35 ng of cDNA/sample. A standard curve was prepared using a reference plasmid standard with known numbers of NP copies/sample (3 × 10¹ to 3 × 10⁹). The primer sequences used for rps9 were 5′CTGGACGAGGGCAAGATGAAGC and 3′TGACGTTGGCGGATGAGCACA; those for np were 5′GAGGGGTGAGAATGGACGAAAAAC and 3′CAGGCAGGCAGGCAGGACTT (Eurofins MWG Operon, Germany).

Single-cell preparation and staining for flow cytometry.

BALF was obtained as described above, and the lungs were then perfused with PBS through the heart. The BALF was centrifuged at 4°C and 2,000 rpm in a tabletop centrifuge to pellet cells. The lungs were excised and manually minced on ice using scissors. Iscove's modified Dulbecco's medium (IMDM) with GlutaMAX-1 (Life Technologies, Germany) supplemented with 0.2 mg/ml collagenase D (Roche Diagnostics, Germany), 1 mg/ml DNase (Sigma-Aldrich, Germany), and 5% fetal bovine serum (FBS Forte; Pan Biotech, Germany) was freshly prepared for enzymatic digestion of the lung tissue. The minced lungs were suspended in 5 ml of the digestion medium for enzymatic digestion at 37°C for 45 min. The reaction was stopped by adding 5 mM EDTA (working concentration), and the cell suspension was passed through a 100-μm filter (Corning Inc., USA) followed by centrifugation at 420 × g at 4°C. Erythrocyte lysis was performed, and cells were resuspended for counting and staining. Single-cell suspensions were incubated at room temperature in the dark in a mixture of the Live/Dead fixable blue stain (ThermoFisher, USA) for dead-cell exclusion and anti-mouse CD16/CD32 antibody (clone 93, purified; BioLegend, USA) for Fc receptor blocking. Cell surface staining was then performed using antibodies specific for mouse F4/80 (clone BM8, APC; BioLegend, USA), CD11b (clone M1/70, Pacific blue; BioLegend, USA), and Ly6G (clone 1A8, PE-Cy7; BioLegend, USA). Samples were acquired on a BD LSR II Fortessa using the FACS Diva (BD) software (where FACS is fluorescence-activated cell sorter), and analysis was performed using the FlowJo software (Tree Star, USA). Tissue-resident alveolar macrophages (AMs; SSC^high FSC^high CD11b⁻ F4/80⁺ autofluorescence⁺), neutrophils (PMNs; SSC^high FSC^high CD11b⁺ Ly6G⁺ F4/80⁻) and infiltrating mononuclear phagocytes (IMPs; CD11b⁺ F4/80^−/lowLy6G⁻) were identified through gating on the respective populations.

Histopathological analysis.

Lungs were fixed in 4% formalin and routinely embedded in paraffin. Sections were cut 5 μm thick, dewaxed, and stained with hematoxylin and eosin (H&E). A blinded histological evaluation was performed by a veterinary pathologist certified by the European College of Veterinary Pathologists. The grade, extent, and pattern of pneumonia were classified into bronchointerstitial pneumonia and bronchopneumonia for IAV and S. pneumoniae infection, respectively. The scoring of lung inflammation (grade 0, not detected; grade 1, minimal; grade 2, mild; grade 3, moderate; and grade 4, severe) was based on the number, kind (neutrophils, macrophages, lymphocytes), and location (interstitial, perivascular, intra-alveolar/intrabronchial) of infiltrating cells.

Cytokine and chemokine detection.

For the detection of cytokines in lung homogenates, mouse enzyme-linked immunosorbent assay (ELISA) Max kits (BioLegend, USA) were used according to the manufacturer's protocol. For the detection of chemokines in lung homogenates, a mouse LEGENDplex proinflammatory chemokine kit (BioLegend, USA) was used according to the manufacturer's protocol. Samples were acquired on a BD LSR II Fortessa instrument and analyzed using the LEGENDplex v7.0 (Vigene Tech, USA) software.

Statistical analysis.

Graph Pad Prism 5.0 (Graph Pad software, USA) and RStudio (version 0.99.902; RStudio Inc., USA) were used to perform statistical analyses. The log rank test was applied on the Kaplan-Meier survival data. Otherwise, the one-way analysis of variance (ANOVA) with Bonferroni's multiple-comparison test was used following log transformation of the data to compare groups. The P values are indicated in figures as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

RESULTS

IAV infection establishes a persisting and unresolved pneumonia even after complete viral clearance and recovery of the original body weight.

Before establishing suitable coinfection models, we characterized the underlying sublethal PR8/34 (H1N1) IAV monoinfection. Following intranasal infection, IAV-infected mice started losing weight by day 4 to 5 postinfection with a maximum weight loss of 12 to 20% between days 7 and 9. By day 12, all mice recovered 100% of the original body weight (Fig. 1a). High viral titers on day 7 and day 9 following IAV infection were consistent with the peak of weight loss (Fig. 1b), and viral clearance was observed by day 14 following IAV infection (Fig. 1b). Pulmonary histopathology was analyzed on days 7, 14, and 21 (Fig. 1c). Seven days following IAV infection, inflammatory lesions were characterized by the accumulation of sloughed bronchial and alveolar epithelial cells and by an expansion of the alveolar septum, interstitium, and bronchial lumen by neutrophils and numerous lymphocytes. At the later time points, lesions were dominated by signs of advanced regeneration with severe hyperplasia of type II pneumocytes (Fig. 1c; see also Fig. S1b and c in the supplemental material). Surprisingly, at the same time a widespread unresolved and partially active pneumonia was established by day 14 and day 21 postinfection (Fig. 1c; see also Fig. S1a in the supplemental material). Based on these findings, in our model, the acute phase of IAV infection was represented by day 7 and the recovery phase persisted from day 14 until at least day 21 after IAV infection.

FIG 1.

Persisting unresolved pneumonia following IAV infection. Wild-type (WT) C57BL/6J mice were intranasally infected with 0.31 TCID₅₀ of the A/PR8/34 H1N1 strain of IAV or treated with PBS as control on day 0. (a) Changes in body weight are represented as percentages relative to the starting weight. Data are shown as means ± standard errors of the means (SEM). (b) Absolute quantification of viral nucleoprotein (NP) copy numbers in the lungs of PBS-treated and IAV-infected mice at the indicated number of days following IAV infection. Lines indicate the medians. Statistical analysis was performed using one-way ANOVA with Bonferroni's multiple-comparison test, with asterisks indicating significant differences between the groups: *, P < 0.05; **, P < 0.01; ***, P < 0.001. Data are compiled from at least two independent experiments with 3 to 5 mice in each group. (c) Representative example (1 of 3) of histopathological changes in the lungs of IAV-infected or PBS-treated mice at 7, 14, and 21 days postinfection analyzed following H&E staining. Lungs of PBS-treated mice were unchanged. For day 7 after IAV infection, the arrowheads indicate an attenuation of the bronchial epithelium, and the neutrophils and macrophages within the alveoli are circled; for day 14 after IAV infection, the arrowheads indicate type II hyperplasia and the neutrophils and macrophages in the bronchi (indicating bronchitis) are circled; for day 21 after IAV infection, the arrowheads indicate type II pneumocyte hyperplasia with increased numbers of lymphocytes in the surrounding interstitium (circled).

Acute IAV infection sensitizes the host to all tested S. pneumoniae serotypes.

To assess the effects of IAV infection on the susceptibility to different S. pneumoniae serotypes, we selected three strains with known differences in nasopharyngeal colonization and IPD potential. First, we characterized the infection with these strains alone as a reference for coinfection. We used the highly invasive strain TIGR4 (T4) of serotype 4 (27), an invasive strain of serotype 7F, and a carrier strain of serotype 19F (29). According to previous reports, serotype 4 and serotype 7F harbor a high invasive disease potential while serotype 19F is less invasive (30–32). To determine the clinical nature of infection with these strains in vivo in our mouse model, naive mice were infected with sublethal doses of S. pneumoniae T4, 19F, or 7F. At 18 h postinfection (hpi) with T4, nasopharyngeal colonization and pneumonia were detected in 83% and 58% of the animals, respectively (Fig. 2a). Furthermore, 21% of all the T4-infected animals showed bacteremia (Fig. 2a). In contrast, 100% of the 19F-infected mice carried bacteria in the nasopharynx and 91% showed high bacterial loads in the lungs (Fig. 2a). Importantly, despite the high frequency of pneumonia, 19F infection did not cause bacteremia in any of the mice (Fig. 2a). Following infection with the 7F strain, 96% of the animals were colonized in the nasopharynx at 18 hpi (Fig. 2a). Bacteria were detected in the lungs of 34% of the 7F-infected animals at comparably low titers but not in the blood (Fig. 2a). Taken together, results from the monoinfections with the three selected S. pneumoniae strains reflected clear differences in pathogenesis in our mouse model, since only the S. pneumoniae strain T4 caused IPD while the S. pneumoniae strains 19F and 7F were restricted to the respiratory tract at 18 hpi.

FIG 2.

Enhanced susceptibility to secondary S. pneumoniae infection during the acute phase of IAV infection. Groups of 5 or 6 WT C57BL/6J mice were oropharyngeally infected with 1 × 10⁶ CFU of S. pneumoniae strain T4, 19F, or 7F. (a) Incidence of colonization, pneumonia, and bacteremia at 18 hpi with S. pneumoniae strain T4, 19F, or 7F according to the bacterial burden detected in the nasopharynx, postlavage lung, and blood for all bacterial monoinfections performed in the study. Data are shown as means ± SEM. (b) Schematic diagram for coinfections with S. pneumoniae strain T4, 19F, or 7F following IAV infection. Mice were infected with 1 × 10⁶ CFU of S. pneumoniae strain T4, 19F, or 7F on day 7 postinfection with 0.31 TCID₅₀ of IAV or PBS treatment. (c to e) Bacterial burden in the BALF (c), postlavage lung (d), and blood (e) at 18 h after secondary infection with T4, 19F, or 7F. Lines indicate the medians. Statistical analysis was performed using one-way ANOVA with Bonferroni's multiple-comparison test with asterisks indicating significant differences: *, P < 0.05; **, P < 0.01; ***, P < 0.001. (f) Survival rates of mice infected with S. pneumoniae T4, 19F, or 7F on day 7 following IAV infection or PBS treatment. Asterisks indicate significant differences assessed by the log rank test on the Kaplan-Meier survival data for the coinfected groups compared to the S. pneumoniae-only groups: *, P < 0.05; **, P < 0.01. (g) Incidence of colonization, pneumonia, and bacteremia in the coinfected and single-S. pneumoniae-infected mice with lethal infection according to the bacterial burden detected in the nasopharynx, whole lung, and blood, respectively.

To gain insight into bacterial-strain-dependent effects on susceptibility to S. pneumoniae following IAV infection, coinfections with all three S. pneumoniae strains were performed during the acute phase of IAV infection (day 7 after IAV infection) (Fig. 2b). The progression of the bacterial disease was assessed by the quantification of the bacterial load in the respiratory tract and blood at 18 hpi as well as the approximate mortality rates over 96 hpi. A nonsignificant tendency for increased bacterial loads in the nasopharynx was evident for all three pneumococcal strains when comparing monoinfection to coinfection at 18 hpi (see Table S1 in the supplemental material). At the same time, coinfection with all three bacterial strains resulted in exceptionally high and significantly elevated bacterial loads in the BALF and lung tissue compared to the respective S. pneumoniae monoinfection (Fig. 2c and d). These results clearly demonstrated IAV-dependent and bacterial-strain-independent pneumococcal outgrowth following coinfection during acute IAV infection. Nevertheless, we observed a bacterial-strain-dependent increase in the incidence of systemic disease following coinfection compared to the respective monoinfection (Fig. 2e). Only the strains previously classified as potentially invasive strains (T4 and 7F) demonstrated systemic dissemination at 18 h after coinfection, while 19F was fully restricted to the respiratory tract at this time point (Fig. 2e). Ultimately, however, the underlying IAV infection led to significantly elevated mortality rates for all the tested S. pneumoniae strains (79%, 75%, and 63% following T4, 19F, and 7F coinfection, respectively) without significant strain-specific differences (Fig. 2f). The overall high lethality observed following coinfection correlated well with the consistently high lung bacterial loads detected at 18 hpi (Fig. 2d) and at the time point at which severely ill mice had to be euthanized according to the predefined endpoint criteria (data not shown). Interestingly, at the time of death 90% of the euthanized T4-coinfected animals showed bacteremia (median bacterial titer, 5.9 × 10⁶ CFU/ml) (Fig. 2g; see also Fig. S2 in the supplemental material). In contrast, only 33% and 55% of the animals sacrificed following 19F and 7F coinfection showed bacteremia (Fig. 2g). Additionally, these bacteremic 19F- and 7F-coinfected mice yielded lower median bacterial titers than the bacteremic T4-coinfected mice (see Fig. S2 in the supplemental material). Of note, survival studies for monoinfection with 19F or 7F using a 10-fold-higher infection dose (1 × 10⁷ CFU) than that used for the coinfections still showed 100% survival without any bacteremia for all the mice (data not shown). Taken together, these results demonstrate acute IAV infection to predispose the host to lethal secondary pneumococcal disease at low bacterial doses irrespective of the IPD potential of the S. pneumoniae serotype. However, at the same time, the manifestation of disease was S. pneumoniae strain dependent regarding the incidence and kinetics of bacteremia.

The bacterial-strain-dependent manifestation of secondary pneumococcal disease persists during the recovery phase of IAV infection.

To assess whether long-term IAV-mediated enhanced susceptibility to pneumococcal disease occurs in a strain- and time-dependent manner, coinfections were performed during the recovery phase of IAV infection either on day 14 or on day 21. Bacterial loads were determined 18 h later (Fig. 2b). In general, for coinfections performed on day 14 after IAV infection, the overall bacterial loads in BALF, lung, and blood were reduced compared to coinfections performed on day 7 after IAV infection (Fig. 2c to e and 3a to c). Regarding the bacterial load in the nasopharynx, again there were no significant differences between coinfected and S. pneumoniae-monoinfected mice irrespective of the pneumococcal strain (see Table S1 in the supplemental material). In contrast, delayed clearance was evident in the LRT for the two invasive strains T4 and 7F. Bacterial loads were higher in the BALF and lung tissue of the T4- and 7F-coinfected mice than in the respective monoinfections (significant for the 7F-coinfected lungs [Fig. 3a and b]). In line with the increased bacterial burden in the LRT, systemic dissemination occurred at a higher incidence in T4- and 7F-coinfected than monoinfected mice (significant for T4 [Fig. 3c]). Of note, the 33% mortality of T4-coinfected mice (Fig. 3d) was accompanied by strong pneumonia and bacteremia in all mice with lethal infection (data not shown). Taken together, even though there was no significant increase in mortality in any of the groups coinfected 14 days following IAV infection (Fig. 3d), bacterial clearance was markedly and significantly delayed for two of the three strains (T4 and 7F). Therefore, these results show enhanced susceptibility to IPD during recovery from IAV infection to depend on the pneumococcal strain.

FIG 3.

Strain-specific alterations in the course of secondary pneumococcal disease during recovery from IAV infection. Mice were infected with 1 × 10⁶ CFU of S. pneumoniae strain T4, 19F, or 7F on day 14 postinfection with 0.31 TCID₅₀ of IAV or PBS treatment. (a to c) Bacterial burden in the BALF (a), postlavage lung (b), and blood (c) at 18 h following secondary infection with T4, 19F, or 7F. Lines indicate the medians. Statistical analysis was performed using one-way ANOVA with Bonferroni's multiple-comparison test, with asterisks indicative of significant differences between the coinfected and S. pneumoniae-only groups: *, P < 0.05; **, P < 0.01; ***, P < 0.001. (d) Survival rates of mice coinfected on day 14 following IAV infection or infected with S. pneumoniae only.

When coinfections were performed 21 days following IAV infection, there was still a trend, which, however, was not significant, toward delayed bacterial clearance from the LRT in all the coinfected groups (see Table S2 in the supplemental material). However, at this time point after IAV infection, the pneumococcal infection remained restricted to the lungs for strain 7F, and the sporadic systemic dissemination observed only for the highly invasive strain T4 was not dependent on a previous IAV infection (see Table S2 in the supplemental material). Taken together, these results indicate that the kinetic of enhanced long-term susceptibility to S. pneumoniae was strain dependent.

IAV-mediated changes in immune cell recruitment following coinfection are bacterial strain dependent, particularly during the acute phase of IAV infection.

Of the different time points tested following IAV infection, the acute phase proved to be most detrimental to the host regarding susceptibility to secondary pneumococcal disease. The strong bacterial outgrowth in the LRT observed for all tested S. pneumoniae strains pointed at defective innate immune responses. Therefore, the recruitment of innate immune cells to the lung was assessed following coinfection with the three S. pneumoniae strains on day 7 after IAV infection. Absolute numbers (i) of alveolar macrophages (AMs), (ii) of infiltrating mononuclear phagocytes (IMPs), which are comprised of newly recruited monocytes, macrophages, and dendritic cells, as well as (iii) of polymorphonuclear neutrophils (PMNs) were determined in the lung. Strikingly, at 18 h after coinfection the T4- and 19F-coinfected groups showed significantly reduced numbers of AMs in the lung tissue compared to what was seen in the S. pneumoniae monoinfection groups (Fig. 4a). Of note, they were also marginally but not significantly reduced compared to the IAV monoinfection (Fig. 4a). In contrast, the number of IMPs was increased for all the coinfected groups compared to the respective S. pneumoniae monoinfection at 18 hpi (significantly for T4- and 7F-coinfected groups) (Fig. 4b). Importantly, the number of IMPs in all the coinfected groups remained comparable to those in the the IAV monoinfections (Fig. 4b). Interestingly, PMN recruitment to the lung following bacterial monoinfection was significantly altered between the three different S. pneumoniae strains. Here, the strongest recruitment of PMNs was detected following the infection with 19F (Fig. 4c). Following coinfection, even though there were no significant changes between the co- and monoinfected groups for any of the tested pneumococcal strains, we observed a trend toward increased PMN numbers following coinfection for strains T4 and 7F (Fig. 4c). However, there was a decrease following 19F coinfection compared to the 19F monoinfection. Importantly, mean PMN numbers in all coinfected animals exceeded those present in the lung after IAV monoinfection. Of note, a similar pattern of differences was also observed for the BAL fluid cells, except that the reduction of PMN numbers in 19F-coinfected mice was not observed (Fig. 4d, e, and f). Importantly, the elevated PMN numbers observed for the invasive S. pneumoniae strains T4 and 7F upon coinfection correlated well with the histopathological analysis performed 18 h following coinfection (Fig. 5). Here, 80% and 100% of the T4- and 7F-coinfected mice, respectively, showed the most severe grade of inflammation, while this was the case for only 57% of the 19F-coinfected mice (Fig. 5a and b). However, the percentage of the lungs affected by bacterial bronchopneumonia was comparable for all three strains after coinfection during acute IAV infection (Fig. 5c). Taken together, severe inflammatory lesions were observed for all the coinfected groups during the acute phase of IAV infection.

FIG 4.

Strain-dependent changes in the innate immune cell composition of the lungs of mice coinfected during the acute phase of IAV infection. Mice were infected with 1 × 10⁶ CFU of S. pneumoniae strain T4, 19F, or 7F or treated with PBS on day 7 postinfection with 0.31 TCID₅₀ of IAV or only PBS treatment. Absolute numbers of alveolar macrophages (# AM) (a), infiltrating mononuclear phagocytes (# IMP) (b), and polymorphonuclear cells (# PMN) (c) in postlavage lungs at 18 h after coinfection. Absolute numbers of alveolar macrophages (# AM) (d), infiltrating mononuclear phagocytes (# IMP) (e), and polymorphonuclear cells (# PMN) (f) in BALF 18 h after coinfection. Data are shown as Tukey box plots with lines indicating the medians. Data are compiled from >5 samples from at least two independent experiments with 2 or 3 WT C57BL/6J mice per group. Statistical analysis was performed using one-way ANOVA with Bonferroni's multiple-comparison test, with asterisks indicating significant differences: *, P < 0.05; **, P < 0.01.

FIG 5.

Histopathological changes in the lung tissue of S. pneumoniae-monoinfected mice and mice coinfected during the acute phase of IAV infection. Mice were infected with 1 × 10⁶ CFU of S. pneumoniae strain T4, 19F, or 7F on day 7 postinfection with 0.31 TCID₅₀ of IAV or PBS treatment. (a) Representative example (1 of 3 to 7 results) of the histopathological changes examined by H&E staining of the lungs of coinfected and S. pneumoniae-monoinfected mice at 18 h postinfection with T4, 19F, or 7F. Arrowheads indicate selected regions of bronchopneumonia. (b) Histopathological scores for inflammation observed in the lungs of coinfected and S. pneumoniae-monoinfected mice. Acute inflammation was characterized as perivascular and interstitial immune cell infiltration. (c) Percentage of the lung affected by bronchopneumonia in the coinfected and S. pneumoniae-monoinfected mice. Histological analysis was performed twice with 2 to 4 mice per group. Lines indicate the medians.

Furthermore, the recruitment of innate immune cells to the lungs was determined following bacterial coinfection during the recovery phase of IAV infection. Of note, the numbers of AMs and IMPs detected in the lungs of coinfected mice were almost equal to those in the S. pneumoniae-monoinfected groups at 18 h following coinfection on day 14 after IAV infection (Fig. 6a and b). However, there were still trends for bacterial-strain-dependent changes in the PMN numbers following coinfection (Fig. 6c). Here, the 19F-coinfected group still showed marginally, though not significantly, lower cell numbers than the 19F-monoinfected group, whereas there was no difference in PMN numbers between the coinfected and monoinfected T4 groups and a trend for increased PMN numbers in the coinfected 7F group (Fig. 6c). By day 21 following IAV infection, AM and PMN numbers were unchanged between the coinfected and monoinfected groups for all bacterial strains (Fig. 6d). In contrast, the number of IMPs remained marginally higher for the T4-coinfected group than for the bacterial monoinfection group (Fig. 6e). Altogether, these data show that the recruitment of innate immune cells in response to S. pneumoniae infection is substantially affected by a preceding IAV infection if coinfection occurs during the acute phase of the IAV infection. Most importantly, we identified both general, S. pneumoniae strain-independent changes such as the increased presence of infiltrating cells following coinfection and strain-dependent changes such as altered AM and PMN numbers.

FIG 6.

Inflammatory cell profiles in the lungs of mice monoinfected with S. pneumoniae or coinfected during the recovery phase of IAV infection. (a to c) Mice were infected with 1 × 10⁶ CFU of S. pneumoniae strain T4, 19F, or 7F or treated with PBS on day 14 postinfection with 0.31 TCID₅₀ of IAV or PBS treatment. Absolute numbers of alveolar macrophages (# AMs) (a), infiltrating mononuclear phagocytes (# IMP) (b), and polymorphonuclear cells (# PMNs) (c) in postlavage lungs 18 h after the bacterial infection or PBS treatment. (d to f) Mice were infected with 1 × 10⁶ CFU of S. pneumoniae strain T4, 19F, or 7F or treated with PBS on day 21 postinfection with 0.31 TCID₅₀ of IAV or PBS treatment. Absolute numbers of alveolar macrophages (# AMs) (d), infiltrating cells (# IMP) (e), and polymorphonuclear cells (# PMNs) (f) in postlavage lungs 18 h after the bacterial infection or PBS treatment. All data are shown as Tukey box plots with lines indicating the medians. Data are compiled from at least two independent experiments with 3 to 4 WT mice per group.

Changes in the cytokine and chemokine responses after coinfection during acute IAV infection are pneumococcal strain dependent.

The local inflammatory responses following infection are majorly orchestrated by a network of cytokines and chemokines. In order to obtain more insight into the pneumococcal-strain-dependent inflammatory processes taking place following coinfection during acute IAV infection, we characterized and compared the lung inflammatory profiles following S. pneumoniae mono- and coinfection for the three strains. Interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α) are multifunctional proinflammatory cytokines important for both local and systemic immune stimulation. For all three pneumococcal strains, coinfection during acute IAV infection led to a clear increase in the production of IL-6 compared to the respective bacterial monoinfection (Fig. 7a). Of note, this excess production of IL-6 observed following coinfection was significant only for the invasive strains T4 and 7F (Fig. 7a). Nevertheless, we detected similarly high concentrations of IL-6 in all coinfected groups at 18 hpi that demonstrated a strong inflammatory response irrespective of the coinfecting pneumococcal strain (Fig. 7a). Also for TNF-α, all the coinfected groups revealed elevated cytokine levels in the lung compared to the respective bacterial-infection-only group (Fig. 7b). Of note, this increase was again strongest and significant only for coinfections with strains T4 and 7F (Fig. 7b).

FIG 7.

Proinflammatory cytokine and chemokine profiles in the lungs of S. pneumoniae-monoinfected or coinfected mice during the acute phase of IAV infection. Mice were infected with 1 × 10⁶ CFU of S. pneumoniae strain T4, 19F, or 7F or treated with PBS on day 7 postinfection with 0.31 TCID₅₀ of IAV or PBS treatment. Protein concentrations of IL-6 (a) and TNF-α (b) in the homogenates of postlavage lungs at 18 h after the secondary bacterial infection or PBS treatment. Data are shown as means ± SEM of the number of mice/group indicated in parentheses and are compiled from two independent infection experiments with 1 to 4 mice per group. Protein concentrations of KC (c), LIX (d), MCP-1 (e), and MIP-1β (f) in the homogenates of postlavage lungs at 18 h after the secondary bacterial infection or PBS treatment. Data are shown as means ± SEM of the number of mice/group indicated in parentheses and are compiled from two independent infections with 2 to 5 mice per group. Statistical analysis was performed using one-way ANOVA with Bonferroni's multiple comparison posttest. Asterisks indicate significant differences: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Next, the lung protein concentrations of critical chemokines, such as MCP-1 (monocyte chemoattractant protein-1), MIP-1 (macrophage inflammatory protein), KC (keratinocyte chemoattractant), and LIX (lipopolysaccharide-induced CXC chemokine), all of which are implicated in driving immunopathology in IAV-infected patients, were elucidated. For the bacterial monoinfection, we show that S. pneumoniae strain 19F induced larger amounts of the neutrophil chemoattractants IL-6, KC, and LIX than did strains T4 and 7F (Fig. 7a, c, and d). The least induction of IL-6, LIX, and especially KC was detected after the 7F infection, which correlated well with the PMN counts observed in the lung following bacterial monoinfection (Fig. 7a, c, and d). Following coinfection, KC levels were significantly elevated independent of the pneumococcal strain compared to the respective bacterial monoinfection (Fig. 7c). Despite a trend for increased levels of LIX in the lungs of mice coinfected with T4 and 7F, there were no significant changes between coinfection and the bacterial monoinfection for any of the strains tested (Fig. 7d). In contrast, both MCP-1, a potent monocytic chemoattractant, and MIP-1β, which is produced by macrophages to activate granulocytes, were significantly increased following coinfection compared to the respective bacterial monoinfection for all pneumococcal strains tested (Fig. 7e and f). Of note, in most cases, i.e., for TNF-α, KC, MCP-1, and MIP-1β, lower levels were detected in the 19F-coinfected group than in the T4- and 7F-coinfected groups (Fig. 7b, c, e, and f). In conclusion, when the mono- and coinfections were compared, it was observed that coinfections with T4 and 7F especially, in contrast to coinfection with 19F, showed a strongly synergistic rather than merely an additive effect for nearly all the tested mediators. Furthermore, these data demonstrate that an underlying acute IAV infection indeed affects the local proinflammatory cytokine response toward S. pneumoniae in a pneumococcal-strain-dependent manner.

DISCUSSION

Bacterial coinfections during IAV infection remain a significant cause of hospitalizations and mortality worldwide. Moreover, as broadly used pneumococcal conjugate vaccines influence capsular switch and serotype distribution, understanding serotype-specific differences in secondary pneumococcal infection following IAV infection is of great importance. In order to elucidate pneumococcal-strain-dependent effects in the synergism between IAV and S. pneumoniae, we determined the outcome and the host immune response following coinfection with three different pneumococcal serotypes. Our model differs from previous coinfection models, as the three strains tested were exemplarily selected to represent strains of high, intermediate, and low IPD potential. The data generated from this model show clear strain-specific differences in the host immune responses and implicate the use of different treatment strategies based on the coinfecting pneumococcal strain in the future.

Interestingly, despite clear differences in the manifestation of monoinfection with the different strains, all three strains caused similarly severe pneumonia and mortality following coinfection during acute IAV infection. This clearly demonstrated that acute IAV infection has devastating effects on antipneumococcal host defenses that are independent of the IPD potential of the coinfecting pneumococcal strain. Strikingly, even the carrier strain 19F caused severe disease and mortality following coinfection at a dose that was nonlethal in the absence of IAV infection in our model as well as in previously described studies (29, 33). IAV infection has been shown to also support the development of otitis media by 19F (34, 35). The strongest colonization was observed following monoinfection with 19F compared to T4 and 7F but without any apparent disease symptoms. Importantly, these findings highlight the requirement of a fully competent innate immune system to confine this strain to its asymptomatic carrier state. In line with this, infections with 19F are most commonly found in children and immunocompromised patients (36). Furthermore, our findings are consistent with reports that bacterial-strain-specific differences are surpassed and mortality does not correlate with the incidence of bacteremia during the acute phase of IAV infection (37, 38). In fact, mortality was proportional to the uncontrolled bacterial outgrowth in the LRT and not bacteremia, as the coinfected animals that succumbed to the infection showed high variability in the extent of bacteremia. These results imply that one of the major causes for mortality was severe pneumonia. The consistently high bacterial loads detected in the lungs of mice coinfected with any of the tested S. pneumoniae strains may further increase the cytolytic activity of bacterial virulence factors such as the pore-forming pneumolysin (39), which is a potent proinflammatory signal. This in turn most likely explains the high levels of cytokines and chemokines in the lungs of all animals coinfected during acute IAV infection. In fact, computational modeling data have demonstrated that the robust local inflammatory responses induced by S. pneumoniae were responsible for rapid mortality and IPD during active IAV coinfection (37).

Next to the pneumococcal-strain-independent effects mediated by IAV infection, we observed distinct changes in the inflammatory response mounted toward S. pneumoniae coinfection that were dependent on the pneumococcal serotype. Depletion of AM during the acute phase of IAV infection has previously been shown to contribute to severe secondary pneumococcal disease (40). Strikingly, in our model we did not observe a significant reduction of AM numbers in the lung during IAV monoinfection. However, coinfection during acute IAV infection led to a reduction in the mean AM counts that was particularly clear and also significant for the strains T4 and 19F. For coinfections with T4, this decrease in AM numbers was associated with an increase in the late-apoptotic or necrotic state of AMs (data not shown). Most likely, the loss in AMs in coinfection has detrimental effects not only on bacterial clearance but also on the resolution of inflammation, as AMs are the major effector cells to clear apoptotic PMNs (41). Therefore, such a clear loss of AMs is likely to tip the balance toward a prolonged proinflammatory response culminating in immunopathology and morbidity (42, 43). For T4-coinfected mice, this scenario correlated well with the strong neutrophil influx and high levels of inflammatory mediators. In contrast, AM reduction was accompanied by a marginal decrease in lung neutrophil numbers following the 19F coinfection compared to the 19F monoinfection. Of note, also, the local concentrations of TNF-α and MIP-1β were marginally, but not significantly, lower for the 19F-coinfected mice than for the T4- and 7F-coinfected mice. This situation makes the AM decrease less consequential in the antibacterial response of 19F than in that of the T4 coinfection. In our study, neutrophils emerged as the major cell type responsible for combating 19F infection, in line with the copious PMN influx observed during 19F monoinfection that was significantly elevated compared to T4 and 7F monoinfections. This is in accordance with a recent study that demonstrated complete abrogation of bacterial clearance in 19F-precolonized animals upon neutrophil depletion (44). For coinfection with strain 7F, cell numbers in the respiratory tract represented a more balanced state and the lethal disease outcome can most likely be attributed to local immunopathology, driven by the excessive inflammatory mediators and inflammatory monocytes observed in all 7F-coinfected animals. For IAV infection alone, a strong TNF-α response has been shown to induce severe pathology (45), and thus neutralization protected against immunopathology-mediated mortality (46, 47). Even though TNF-α neutralization failed to lower the disease severity following T4 coinfection (48), it still holds potential for a treatment strategy for the 7F coinfections and should be tested in the future.

Altogether, our study has identified several characteristics in the host response toward different pneumococcal strains following coinfection during acute IAV infection. Nevertheless, the outcome was equally devastating for all tested strains. Effective treatment of secondary pneumococcal infection following IAV infection has been suggested to combine antibiotic measures with immune modulators (49, 50). Therefore, future studies will be needed to exploit our findings regarding the serotype-specific inflammatory responses following coinfection for treatment regimens tailored to the coinfecting pneumococcal strain.

Importantly, we have extended our study of the S. pneumoniae strain dependency in host susceptibility and inflammatory responses following coinfection to the recovery phase of IAV infection. Histological analyses performed until 21 days following IAV infection revealed that pneumonia was established even after complete recovery of body weight and viral clearance. In fact, the few long-term follow-up studies with small cohorts of patients that survived acute H1N1 infection have revealed that structural abnormalities in the lung parenchyma can last up to a year after respiratory disease (51, 52). When coinfections were performed on day 14 following IAV infection, we indeed still observed impaired clearance of S. pneumoniae. While survival was unaffected, we detected increased bacterial growth in the LRT and significantly increased incidence of systemic dissemination in a strain-dependent manner. In fact, IAV-induced impairment of AM sensitivity is proposed to persist until 100 days after IAV infection in mice (24). Furthermore, it has also been reported that a state of immunosuppression marked by the upregulation of IL-10 (25) is set up to promote repair and recovery after IAV infection, which may explain the marginally delayed bacterial clearance in all the coinfected animals until day 21 after IAV infection and the low levels of inflammation in the lungs compared to day 7 after IAV infection. In line with previous reports, our results show that IAV infection affects pneumococcal clearance even after elimination of the virus during recovery of the host. Importantly, we show a pneumococcal strain dependency, which implies that strains with higher IPD potential, such as T4, take advantage of the IAV-preinfected host for a longer time than do colonizers such as 19F. Despite equally low mortality following coinfection with the different pneumococcal strains late following IAV infection in our model, we believe that this finding holds implications, e.g., for the surveillance of survivors of severe IAV infections during outbreaks of pneumococcal strains with moderate-to-high IPD potential.

Altogether, the results of our study strongly highlight the need to identify the serotype infecting the patient before administering standard treatment regimens, such as, e.g., the increasingly used corticosteroids (45, 50, 53). Following coinfection during the acute phase of IAV infection, the invasive strains T4 and 7F caused stronger proinflammatory responses than those caused by the 19F strain. Therefore, immunosuppression along with antiviral and antibacterial treatment may work better for these strains. In case of coinfection with colonizing strains such as 19F, a strong bacteriostatic antibiotic along with antivirals may be more potent in preventing mortality. Certainly, these hypotheses will need to be tested in future studies, for which we believe our results provide a valuable first basis. Furthermore, our results imply the imperative need to strongly consider the pneumococcal serotype when studying the mechanisms underlying severe IAV-S. pneumoniae coinfections in animal models.