Logo of plosonePLoS OneView this ArticleSubmit to PLoSGet E-mail AlertsContact UsPublic Library of Science (PLoS)
PLoS ONE. 2008; 3(4): e1969.
Published online 2008 Apr 9. doi:  10.1371/journal.pone.0001969
PMCID: PMC2279396

Strain-Specific Virulence Phenotypes of Streptococcus pneumoniae Assessed Using the Chinchilla laniger Model of Otitis Media

Niyaz Ahmed, Editor



Streptococcus pneumoniae [Sp] infection is associated with local and systemic disease. Our current understanding of the differential contributions of genetic strain variation, serotype, and host response to disease phenotype is incomplete. Using the chinchilla model of otitis media [OM] we investigated the disease phenotype generated by the laboratory strain TIGR4 and each of thirteen clinical strains (BS68-75, BS290, BS291, BS293, BS436 and BS437); eleven of the thirteen strains have been genomically sequenced.

Methodology/Principal Findings

For each strain 100 colony forming units were injected bilaterally into the tympanic bullae of 6 young adult chinchillas under general anesthesia. All animals were examined daily for local and systemic disease by a blinded observer. Pneumatic otoscopy was used to evaluate local disease, and behavioral assessments served as the measure of systemic disease. Virulence scoring was performed using a 4-point scale to assess four clinical parameters [severity and rapidity of local disease onset; and severity and rapidity of systemic disease onset] during a 10-day evaluation period. Highly significant variation was observed among the strains in their ability to cause disease and moribundity.


As expected, there was a significant correlation between the rapidity of systemic disease onset and severity of systemic disease; however, there was little correlation between the severity of otoscopic changes and severity of systemic disease. Importantly, it was observed that different strains of the same serotype produced as broad an array of disease phenotypes as did strains of different serotypes. We attribute these phenotypic differences among the strains to the high degree of genomic plasticity that we have previously documented.


Otitis media (OM) is the most common malady for which children receive medical and surgical attention in the United States and Europe [1]. Nearly all children will experience an episode of OM during the first two years of life, and approximately 2 million children receive myringotomies and ventilation tubes annually with a total treatment cost of nearly five billion dollars [2][5]. OM has a wide range of clinical presentations [1] and if untreated can, in rare cases, lead to invasive and systemic disease states that are associated with moribundity. OM pathogenesis is multifactorial with pathogen-specific virulence factors, and the host's anatomy, immune status, and genetics all playing roles in phenotypic expression [6][15]. Recent evidence has established that genetic polymorphisms in the promoter of the human IL-9 gene result in its over expression in response to endotoxin exposure resulting in increased mortality [8]. Other studies [16][19] have illuminated the complexities of pathogen-host interactions and provided evidence supporting the hypotheses that intrinsic pathogen virulence factors, as well as specific elements of the host response make independent contributions to the development of clinical phenotype providing potential targets for future therapies [20], [21], [22].

These investigations continue to yield provocative answers and questions, but the local and systemic clinical challenges associated with OM remain. A uniformly predictive way of stratifying risk in persons with OM that provides for the reliable identification of persons with self-limited disease or chronic local disease (otitis media with effusion [OME]) from those with early invasive disease has been elusive. This difficulty in differentiation exists because the initial disease presentation for both self-limited, and invasive infections, are similar.

S. pneumoniae is one of the three bacterial pathogens (together with Haemophilus influenzae and Moraxella catarrhalis ) most frequently identified by culture and molecular diagnostics from middle-ear effusions obtained from children with acute and chronic forms of OM [23]. This pathogen is also associated with a number of other important infections including pneumonia, meningitis, bacteremia and osteomyelitis [24][27]. While over 90 different serotypes of Sp are known, less than 30 are responsible for over 90% of human disease [28]. This suggests that serotype is important, but not the sole determinant of pathogenicity.

In a survey study Shen et al. [29] demonstrated that each of ten clinical pneumococcal strains varied substantially from each other in terms of gene content providing support for the distributed genome hypothesis (DGH) [30]. This study suggested that each strain contains a unique distribution of genes from a population-based supragenome [29] that is much larger than the genome of any single strain. As a direct test of the DGH, Hiller et al. showed in an exhaustive pair-wise comparison of 17 complete pneumococcal genomes that on average each clinical strain pair varied (i.e. the gene was present in one strain and absent in the other) at 18% of their genic loci [31]. Moreover, this study documented that between 21% and 33% of the genes in each strain are non-core, or distributed genes. Thus, each strain contains a unique distribution of genes, not alleles, from a population-based supragenome, which is far greater in size than the genome of any given strain. This observation of extensive genomic plasticity driven by active horizontal gene transfer mechanisms has been widely observed among other pharyngeal pathogens as well, including Haemophilus influenzae, Neisseria meningiditis, and Streptococcus agalactiae [29], [32], [33], [34].

Previous animal studies of OM using the Chinchilla laniger (chinchilla) model suggested that this model could differentiate between both different bacterial species and different clinical strains of a single species with respect to the severity of both local and systemic disease [35], [36], [37]. The gram-negative bacterium Haemophilus influenzae often yielded lethal, systemic disease characteristic of fulminant endotoxemia, whereas inoculation with Sp produced an unpredictable variety of apparently strain-specific clinical phenotypes.

The objective of this study was to determine if each of fourteen genetically characterized Sp clinical strains produced a consistent and unique disease phenotype with respect to type (local or systemic), rapidity of onset, and severity of disease following middle-ear inoculation. Future studies will examine the discrete phenotype/genotype relationships of these strains.


Study Design

We undertook a comparative evaluation of the disease phenotypes induced by 14 genotypically distinct pneumococcal strains [29], [31], including 5 strains of serotype 9 and 2 strains of serotype 6, and one unencapsulated strain that was known to cause localized disease using the chinchilla model of OM (Fig. 1). Strains will be referred to by their name followed by the serotype within parenthesis. The rapidity and severity of both otoscopic and systemic clinical signs for each strain were evaluated on a 4-point scale using a cohort of 6 animals by a blinded observer (Table 1).

Figure 1
The transbullar technique is used to induce S. pneumoniae otitis media in the chinchilla under general anesthesia.
Table 1
Scoring System to Quantify Streptococcus Pneumoniae Pathogenicity in the Chinchilla

Otoscopic Signs

Across the fourteen strains the rapidity of otoscopic changes, defined as the days to the development of unambiguous otological signs (a score of 2 or above since differences between scores of 0 and 1 were difficult to discern, Table 1), produced a mean of 1.93 days±1.35 (standard deviation, SD) (Fig. 2). Note, that animals that never developed otoscopic disease (such as all PBS controls) were excluded from this set. An ANOVA analysis of the differences among the strains produced highly significant results (p-value = 9.6e−15) as did repetition of the analysis by the non-parametric Kruskal-Wallis test (p-value = 2.4e−5) (Table 2). The means for the individual strain cohorts ranged from 1 day for strains BS291(9), BS75(19) and BS69(14) to 5.5 days for the unencapsulated strain BS293 and 4.6 for the capsulated strain BS74(6) (Fig. 2, Table 3). Statistical analysis of all the strain pairs using Tukey HSD showed that unencapsulated BS293 and BS74(6), the strains with slowest onset of disease, differed significantly from strains BS290(9), BS291(9), BS436(9), BS437(9), BS68(9), BS69(14), BS70(11), BS73(6), BS75(19), and TIGR4(4). In addition unencapsulated BS293 also differed from BS71(3). Finally BS72(23) differed significantly from BS291(9), BS68(9), BS69(14), BS73(6) and BS75(19), the strains with the fastest onset of disease.

Figure 2
Rapidity of otologic disease onset.
Table 2
Summary of All Statistical Analysis
Table 3
Averages and standard deviations observed within each strain for local and systemic disease

We also analyzed the maximal otologic score for each animal as a measure for the severity of local disease (where scores of 0 and 1 were considered to be the same since they are hard to discern) (Fig. 3). The mean score for maximal otologic disease among the fourteen strains was 2.26±0.87. Mean scores for individual strains ranged from 1.3 for strains BS71(3) and BS72(23) to 2.83 for strains BS75(19), BS437(9), BS73(6), BS69(14), and TIGR4(4) (Table 3). Statistical analysis of this data set showed a significant difference amongst the strains (ANOVA p-value = 9.56e−5, and Kruskal-Wallis p-value =  0.00118) (Table 2). Further analysis of all the strain pairs using Tukey HSD, showed that the strains with the least severe otoscopic disease BS71(3) and BS72(23) differed significantly from BS437(9), BS69(14), BS73(6), BS75-19, and TIGR4(4). There was a significant correlation between days to onset of otoscopic disease and the maximal otoscopic score (Spearman's rank correlation rho = −0.311 with p-value = 0.009).

Figure 3
Severity of otoscopic change.

Animals that were inoculated with BS71(3) and BS72(23) rapidly displayed signs of systemic moribundity, which effected our ability to gauge the development of otoscopic disease past day 2 (Fig. 4). Importantly, many strains show otoscopic changes by day 1 or 2 (Fig. 2), demonstrating that systemic moribundity by day 2 does not prohibit the development of otoscopic signs, as is most evident for strain BS68(9), which showed high otologic scores by day 2 (Fig. 3 and and4).4). Nonetheless, to account for the affect of the loss of animals infected with systemically virulent strains on the strain's otoscopic scores, we performed Kaplan-Meier Survival Probability Estimates. These results confirmed that days to onset of otoscopic disease were significantly different between the strains (the log-rank test for days to onset of moderate or worse otologic disease had a p-value = 6.446e−12, Table 2).

Figure 4
Days to moribund condition.

Systemic Signs

The rapidity of the development of systemic signs across all fourteen strains produced a mean of 2.3 days±1.6 (Fig. 5). Table 1 describes the signs of systemic disease for which the animals were monitored, scores of 0.5 and above were considered, and animals that never developed systemic signs were excluded from the analysis. The means ranged from 1.5 days for strain BS71(3) and 1.58 days for BS68(9) to 4 days for BS290(9) (Table 3). Analysis of this data using ANOVA does not show any statistical significance amongst the strains, however, repetition of the analysis by the non-parametric Kruskal-Wallis test yielded a statistically significant p-value of 0.016 (Table 2). Since even healthy animals showed slightly lethargic behavior (score 0.5), as evident by the PBS control, we also analyzed the days to onset of moderate or worse systemic disease (only scores of 2 or above; thus all animals that did not develop at least moderate symptoms were excluded from this analysis) (Fig. 6). These values range from 1.5 days for BS71(3) to 4 days for BS437(9) (Table 3). The results showed a significant difference amongst all strains (ANOVA p-value = 0.03, Kruskal-Wallis test p-value =  0.001) (Table 2).

Figure 5
Rapidity of onset of any systemic disease signs.
Figure 6
Rapidity of onset of moderate or worse systemic signs.

The maximum systemic severity score was highly significantly different amongst strains (ANOVA p-value =  2.69e−9 and Kruskal-Wallis p-value = 4.22e−6)(Table 2, Fig. 7). The mean score for all fourteen strains was 3.11±1.52. It ranged from 0.5 for the unencapsulated strain BS293 (as well as the PBS control) to 4 for strains BS291(9), BS75(19), BS71(3), BS72(23), and BS70(11), where all animals demonstrated moribundity (Table 3). Tukey-HSD analysis of all the strain pairs showed that the unencapsulated strain BS293 and TIGR4(4) differed significantly from BS291(9), BS436(9), BS68(9), BS69(14), BS70(11), BS71(3), BS72(23), BS73(6), BS75(19), in addition BS293 also differed from BS437(9).

Figure 7
Severity of systemic signs.

There was a significant correlation between severity of otoscopic disease and systemic rapidity (Spearman's rank correlation rho = 0.42 with p = 1.9e−4), and between severity and rapidity of systemic disease (Spearman's rank correlation rho = −0.30 with p = 5.95e−3); while there was no significant correlation between otoscopic and systemic severity.

Unencapsulated strains are not thought to cause systemic disease, thus we repeated the systemic score statistical analyses using only the capsulated strains (Table 2). Systemic severity scores were also highly significantly different amongst the encapsulated strains (ANOVA p-value  = 1.22e−5, and Kruskal-Wallis p-value = 0.000171). Note that the onset of moderate or worse systemic disease has the same p-value as amongst all strains, since the unencapsulated strain never reaches moderate or worse disease and thus was already excluded from this analysis.

Strain Specific Moribundity

The moribundity of the chinchillas differed significantly amongst the strains, as determined by Fisher's Exact test (p-value = 3.9e−7) (Table 2). Strains BS291(9), BS70(11), BS71(3), BS72(23), and BS75(19) were the most lethal, producing 100 % moribundity within the cohorts; in contrast, no animals inoculated with TIGR4(4) and BS293-nonencapsulated reached a state of moribundity (Table 4). Again, since unencapsulated strains are not thought to cause systemic disease the Fisher's Exact test was repeated amongst just the capsulated strains, confirming that there are major differences in moribundity among the encapsulated strains (p-value = 0.000117).

Table 4
Strain Specific Moribundity.

Intra-serotype Specific Statistical Tests

We evaluated five independent serotype 9 strains (BS68, BS290, BS291, BS436, and BS437) to determine if the entire range of clinical phenotypes observed among the 9 strains, each with a unique serotype, would be present within a single serotype as a test to determine if total genotype (as opposed to serotype alone) played an important role in disease phenotype.

Analysis of the rapidity of otoscopic changes showed a significant difference among the five serotype 9 strains (ANOVA p-value =  0.008, Kruskal-Wallis p-value = 0.01) (Table 2). The mean was 1.6 days±0.6, and the individual strain means ranged from 1 for BS291 to 2.16 for BS436 (Fig. 2, Table 3). No statistical differences were observed in otoscopic severity among the five serotype 9 strains (Table 2), which had a mean score of 2.45±0.65.

There were no significant differences amongst the serotype 9 group for rapidity of onset of systemic disease, the mean was 2.48 days±1.9 (this is not surprising since this score was also not significant amongst all strains) (Fig. 5). In contrast, the onset of moderate or worse systemic disease is significantly different amongst the serotype 9 group (ANOVA p-value = 0.001 and Kruskal-Wallis p-value = 0.0017) (Table 2, Fig. 6). Finally, the difference for systemic severity scores amongst serotype 9 strains was not significant (ANOVA p-value = 0.17 and Kruskal-Wallis p-value = 0.18). The mean score was 3.28±1.4, which ranges from 2.25 for BS290 to 4 for BS291 (Table 3 and Fig. 7).


The chinchilla OM model provides a useful and facile system to study the spectrum of disease induced by multiple clinical Sp strains of unique genotype. In the current study we used an objective clinical scoring system [38] to investigate the differences in virulence induced by a collection of Sp strains that included 8 of the serotypes most frequently associated with OM, including multiple isolates of serotype 9. These data established that individual strains, even within a single serotype, have a wide variation with regard to their ability to induce both local and systemic disease, and that there is little if any linkage between a strain's propensity to induce local and systemic signs, most evident by TIGR4(4) that induced strong local disease but very mild systemic changes. The variation amongst animals within cohorts was lower than the variation amongst all cohorts suggesting that the bacterial strain type plays a major factor in pathogenesis, however, the genetic background of the host is also clearly important as there is variability of disease within a cohort of animals infected with a single isolate. It is important to remember that the chinchillas used in these studies are of outbreed stock and are therefore genetically heterogeneous. Thus, they would not necessarily be expected to have a stereotypic host response.

An example of the Sp strain-specific effects can be seen by the large difference in local effects induced by BS71(3) and BS72(23) relative to BS75(19), BS437(9), BS73(6), BS69(14) and TIGR4(4). Similarly, the strain-specific systemic effects were highly significant when comparing strains BS70(11), BS71(3), BS72(23), BS75(19), and BS291(9) with TIGR4(4) and unencapsulated BS293. Strains BS437(9) and TIGR4(4), which caused significant local disease, were only average or mild in their systemic virulence, respectively. Similarly, some strains that produced rapid, fulminant systemic signs (strains BS70(11), BS71(3) and BS72(23)) did not cause clinically important local disease. This dichotomy may be due to one or more of several factors. However, it is likely that a major contributor to the observed differences in disease results from the unique combination of virulence genes that each strain possesses [29], [31]. The clinically different disease states presented by several strains, with the same serotype (9), strongly suggest that configuration of the capsular polysaccharides that confer serotype may not always play a major role in the disease process. Interestingly, the unencapsulated strain BS293, which causes significant localized disease, was obtained from the same patient at two separate clinic visits 23 days apart suggesting that this is a persistent unecapsulated strain (manuscript in preparation, Hiller, Hu, and Ehrlich).

We have recently completed whole genome sequencing on 11 of the 13 clinical strains evaluated in this study, which revealed very extensive genomic plasticity among them [31] and unpublished data. After annotating each genome, a conservative gene clustering algorithm was used [39] to bin together orthologous genes from the different strains to ensure that simple allelic difference did not result in our calling orthologous loci as unique. The gene clustering results were then used for the performance of global (all strains together) genome-wide comparisons to determine the number of core genes, those present in all strains, as well as the number of distributed and unique genes present in each strain and collectively within a 24 strain supragenome. These analyses were followed by the performance of an exhaustive set of pair-wise genic comparisons across all strains, which showed that on average each strain varied from every other strain by the gain or loss of nearly 400 gene clusters. Since on average each strain pair only contains 1795 gene clusters in common this means that ∼18% of each genome is unique with respect to every other genome on average. Included among the distributed and unique gene sets for each of these strains were a large number of candidate virulence genes, many of which have not been previously identified among the pneumococci [29]. One such example includes a cluster of 10 adjacent genes that are always found together, such that if a strain possesses one of them it possesses all of them suggesting that these genes may represent a previously unrecognized mobile genetic element particularly as one of the genes annotates as an integrase. This gene cluster is possessed by all of the highly virulent strains including BS70(11), BS75(19), BS72(23), BS71(3), BS291(9), and is also found in one of the intermediately virulent strains, BS68(9), as well as one of the less virulent strains BS290(9), but is absent among the remainder of the avirulent and intermediately virulent strains including TIGR4(4), BS74(18), BS69(14), and BS73(6). These genes are also present in the systemically avirulent unencapsulated strain BS293, but this strain's lack of systemic virulence is attributable to its lack of capsule. Five of these ten genes are organized into one operon and have been annotated as being involved in cellobiose PTS transport and metabolism. This is interesting because McKessar and colleagues (2007) previously showed that disruption of one of the virulence-related two-component systems in Sp, TCS08, resulted in suppression of a putative cellobiose transport system (PTS)[40]. The PTS system we have identified is only distantly related to the one described by McKessar et al. with some of the components having very limited identity (30–40%) and others which show no measurable identity whatsoever. Nonetheless, their results suggest that PTS cellobiose systems may play a role in virulence; making the PTS system described here an interesting candidate for contributing to Sp systemic virulence based upon its segregation largely within the more highly systemically virulent strains. Together these genomic and phenotypic studies suggest that the virulence factors for local and systemic disease may be unique and that some strains have accumulated more of one class than another, whereas some strains have accumulated genes from both classes. Moreover, it is clear that serotype replacement occurs amongst Sp strains, suggesting that strains that are very genetically diverse may share the same serotype, likewise strains that are highly related may have different serotypes. We have observed strain evolution in vivo that has resulted in as much as 20% genome replacement by a donor strain (Hiller et al. unpublished observations). Genome sequencing of multiple Sp strains confirms this, since strains of the same serotype are often spread apart in a phylogenic comparison based on genic differences [31]. This enormous plasticity of Sp strains makes it difficult to predict across any length of time or within different geographic regions what the pathogenicity of any specific strain or serotype will be.

The observation that there is extensive genomic plasticity among Sp strains does not preclude the contribution of unique host factors to clinical phenotype and outcome. Such host factors could include differing expression levels of certain immune system components that may affect susceptibility or resistance to certain pathogens [10], [41], [42].

Induction of OM in the chinchilla results in a reproducible model of local infection that reliably mimics the clinical disease. The dichotomy between a strain's proclivity to cause local and systemic disease is entirely consistent with the spectrum of human disease caused by Sp. The clinical challenge remains to differentiate those patient-strain combinations that will result in infectious spread by invasion or dispersion, from those that will have self-limited disease. This model provides the unique opportunity to separate pathogen weaponry from host response.

Materials and Methods

Bacterial strains and culture

Thirteen pneumococcal strains (all ampicillin sensitive) were obtained as nasal washes from symptomatic pediatric participants at Children's Hospital of Pittsburgh who were enrolled in a Fluzone vaccine trial. All clinical S. pneumoniae strains were isolated and restreaked for single colonies on Trypticase™ Soy Agar (TSA II) supplemented with 5% sheep's blood (Becton Dickinson, Sparks, MD), and then cultured in Todd Hewitt broth (Sigma, St. Louis, MO) at 37°C in a humidified 5% CO2 atmosphere for one passage followed by aliquotting and cryopreservation in 22% glycerol at –80°C. Pneumococcal isolates were typed by the Quellung reaction using polyvalent serogroup and serotype-specific antisera (Statens Seruminstitut, Copenhagen, Denmark). The eleven strains, designated as BS68-75, BS290, BS291, BS436, BS437, BS293, were serotyped as types 9V, 14, 11, 3, 23F, 6A, 6A, 19F, 9V, 9V, 9V, 9V and nontypeable, respectively. The TIGR4 (BS90, serotype 4) strain was also obtained and plated. Eleven of the thirteen (excluding BS436 and BS437) clinical strains have been completely sequenced to an average depth of coverage of >20-fold and have been demonstrated to be unique with respect to their genic content [29], [31]. Middle-ear inocula for each strain were prepared by plating one of the first passage aliquots on blood agar for overnight incubation. Chinchillas were inoculated bilaterally with ∼100 CFU/ear based on cultures with an ODA600 of 1 of having 5×108 CFU/mL. The actual inoculating titers were determined by standard dilutional plate counts.

Animal experiments and monitoring

All experiments were conducted with the approval of the local Institutional Animal Care and Use Committee (IACUC). Research grade young adult chinchillas (C. laniger, 400–600 gm; McClenahan Chinchilla Ranch, New Wilmington, PA) were obtained free of middle-ear disease. After a protocol-directed period of environmental acclimation, induction of anesthesia was attained on day 0 by intramuscular injection of 0.1 mL of a solution of ketamine hydrochloride 100 mg/mL, xylazine hydrochloride 30 mg/mL and acepromazine 5 mg/mL. After anesthesia was confirmed (abolishment of the eye-blink reflex), 0.1 mL of a 1000 CFU/mL Sp culture was injected bilaterally into the tympanic bullae using a 0.5 in, 27-gauge needle attached to a 1 ml syringe (Fig. 1). Each of the fourteen pneumococcal strains was used to infect a cohort of 6 chinchillas. BS68 was inoculated into two cohorts of 6 chinchillas to ensure experimental consistency across temporally distinct inoculation groups; the data from the two BS68 groups were combined for statistical analysis. Importantly, there was no statistically significant difference between the two BS68-inoculated sets using both parametric and non-parametric tests. The BS291 cohort was reduced to 5 due to a death unrelated to the experiment.

A multiparameter scoring system based on signs of local and systemic disease (Table 1) was applied to each chinchilla on a daily basis for ten days to assess the degree of disease induced by the individual pneumococcal strains. All clinical evaluations were performed by an otoscopist who was blinded with respect to the strains that had been used to inoculate each animal. This scale has demonstrated utility in clinical scoring of the chinchilla's response to direct bacterial inoculation of the bullae [38]. A control cohort of 6 animals was inoculated with sterile phosphate buffered saline. Moribund animals (having a systemic severity score of 4) were either treated with 150-mg/kg ampicillin daily (all Sp strains were ampicillin sensitive) or sacrificed in accordance with the IACUC protocol. Once an animal reached this state evaluation ceased, as moribundity was considered an end point.

Statistical Analysis

Raw scores for local and systemic disease ranged from 0 (clinically normal) to 4 (extreme severity) (Table 1) and moribundity was recorded (Table 4). Results are reported as mean±standard deviation (SD). One-way ANOVA and Kruskall-Wallis tests were utilized in comparisons across cohorts to determine if the differences in local and systemic diseases were statistically significant among strains. Tukey's HSD test was utilized for pairwise comparisons between strains. The Fisher Exact test was utilized to examine the statistical significance of moribundity data across all cohorts (Table 4). Kaplan Meier estimates were utilized to perform log-rank tests that measure for strain differences in local disease onset. Significance was defined as p-value ≤0.05. The open source software package R was utilized for all statistical analysis (http://www.r-project.org/).


The authors thank David P. Greenberg for the clinical pneumococcal strains, and Mary O'Toole for help in the preparation of this manuscript.


Competing Interests: The authors have declared that no competing interests exist.

Funding: This work was supported by Allegheny General Hospital/Allegheny Singer Research Institute and grants from the Health Resources and Services Administration and the NIH-NIDCD: DC02148 (GDE), DC04173 (GDE), and DC05659 (JCP). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.


1. Bluestone CD, Klein JO. Philadelphia: WB Saunders; 2000. Otitis Media in Infants and Children 3rd edition.
2. Bondy J, Berman S, Glazner J, Lezotte D. Direct expenditures related to otitis media diagnoses: extrapolations from a pediatric Medicaid cohort. Pediatrics. 2000;105(6):E72. [PubMed]
3. Elden LM, Coyte PC. Socioeconomic impact of otitis media in North America. J. Otolaryngol. 1998;27(Suppl 2):9–16. [PubMed]
4. Kaplan B, Wandstrat TL, Cunningham JR. Overall cost in the treatment of otitis media. Pediatr Infect Dis J. 1997;16(2 Suppl):S9–11. [PubMed]
5. Klein JO. Management of otitis media: 2000 and beyond. Pediatr Infect Dis J. 2000;19(4):383–387. [PubMed]
6. Adderson EE, Takahashi S, Bohnsack JF. Bacterial genetics and human immunity to group B streptococci. Mol Genet Metab. 2000;71(1–2):451–454. [PubMed]
7. Bredius RG, Derkx BH, Fijen CA, de Wit TP, De Haas M, et al. Fc gamma receptor IIa (CD32) polymorphism in fulminant meningococcal septic shock in children. J Infect Dis. 1994;170(4):848–853. [PubMed]
8. Grohmann U, Van Snick J, Campanile F, Silla S, Giampietri A, et al. IL-9 protects mice from Gram-negative bacterial shock: suppression of TNF-alpha, IL-12, and IFN-gamma, and induction of IL-10. J Immunol. 2000;164(8):4197–4203. [PubMed]
9. Kornelisse RF, Hazelzet JA, Savelkoul HF, Hop WC, Suur MH, et al. The relationship between plasminogen activator inhibitor-1 and proinflammatory and counterinflammatory mediators in children with meningococcal septic shock. J Infect Dis. 1996;173(5):1148–1156. [PubMed]
10. Mira JP, Cariou A, Grall F, Delclaux C, Losser MR, et al. Association of TNF2, a TNF-alpha promoter polymorphism, with septic shock susceptibility and mortality: a multicenter study. JAMA. 1999;11;282(6):561–568. [PubMed]
11. Stuber F, Petersen M, Bokelmann F, Schade U. A genomic polymorphism within the tumor necrosis factor locus influences plasma tumor necrosis factor-alpha concentrations and outcome of patients with severe sepsis. Crit Care Med. 1996;24(3):381–384. [PubMed]
12. Gibot S, Cariou A, Drouet L, Rossignol M, Ripoll L. Association between a genomic polymorphism within the CD14 locus and septic shock susceptibility and mortality rate. Crit Care Med. 2002;30(5):969–973. [PubMed]
13. Holmes CL, Russell JA, Walley KR. Genetic polymorphisms in sepsis and septic shock: role in prognosis and potential for therapy. Chest. 2003;124(3):1103–1115. [PubMed]
14. Schaaf BM, Boehmke F, Esnaashari H, Seitzer U, Kothe H, et al. Pneumococcal septic shock is associated with the interleukin-10-1082 gene promoter polymorphism. Am J Respir Crit Care Med. 2003;15;168(4):476–480. [PubMed]
15. Watanabe E, Hirasawa H, Oda S, Matsuda K, Hatano M, et al. Extremely high interleukin-6 blood levels and outcome in the critically ill are associated with tumor necrosis factor- and interleukin-1-related gene polymorphisms. Crit Care Med. 2005;33(1):89–97. [PubMed]
16. McCool TL, Cate TR, Moy G, Weiser JN. The immune response to pneumococcal proteins during experimental human carriage. J Exp Med. 2002;4;195(3):359–365. [PMC free article] [PubMed]
17. Ratner AJ, Lysenko ES, Paul MN, Weiser JN. Synergistic proinflammatory responses induced by polymicrobial colonization of epithelial surfaces. Proc Natl Acad Sci U S A. 2005;1;102(9):3429–3434. [PMC free article] [PubMed]
18. Wang X, Moser C, Louboutin JP, Lysenko ES, Weiner DJ, et al. Toll-like receptor 4 mediates innate immune responses to Haemophilus influenzae infection in mouse lung. J Immunol. 2002;15;168(2):810–815. [PubMed]
19. Wizemann TM, Heinrichs JH, Adamou JE, Erwin AL, Kunsch C, et al. Use of a whole genome approach to identify vaccine molecules affording protection against streptococcus pneumoniae infection. Infect Immun. 2001;69(3):1593–1598. [PMC free article] [PubMed]
20. Brogden KA, Guthmiller JM, Taylor CE. Human polymicrobial infections. Lancet. 2005;15–21;365(9455):253–255. [PubMed]
21. Bruckner R, Nuhn M, Reichmann P, Weber B, Hakenbeck R. Mosaic genes and mosaic chromosomes-genomic variation in Streptococcus pneumoniae. Int J Med Microbiol. 2004;294(2–3):157–168. [PubMed]
22. Perez-Caballero D, Garcia-Laorden I, Cortes G, Wessels MR, de Cordoba SR, et al. Interaction between complement regulators and Streptococcus pyogenes: binding of C4b-binding protein and factor H/factor H-like protein 1 to M18 strains involves two different cell surface molecules. J Immunol. 2004;173(11):6899–6904. [PubMed]
23. Post JC, Preston RA, Aul JJ, Larkins-Pettigrew M, Rudquist-White J, et al. Molecular analysis of bacterial pathogens in otitis media with effusion. JAMA. 1995;273:1598–1604. [PubMed]
24. Hoffman JA, Mason EO, Schutze GE, Tan TQ, Barson WJ, et al. Streptococcus pneumoniae infections in the neonate. Pediatrics. 2003;112(5):1095–1102. [PubMed]
25. Jacobs NM. Pneumococcal osteomyelitis and arthritis in children. A hospital series and literature review. Am J Dis Child. 1991;145(1):70–74. [PubMed]
26. Klugman KP, Feldman C. Streptococcus pneumoniae respiratory tract infections. Curr Opin Infect Dis. 2001;14(2):173–179. [PubMed]
27. Tan TQ. Pneumococcal infections in children. Pediatr Ann. 2002;31(4):241–247. [PubMed]
28. Martens P, Worm SW, Lundgren B, Konradsen HB, Benfield T. Serotype-specific mortality from invasive Streptococcus pneumoniae disease revisited. BMC Infect Dis. 2004;30;4(1):21. [PMC free article] [PubMed]
29. Shen K, Gladitz J, Antalis P, Dice B, Janto B. Characterization, distribution and expression of novel genes among eight clinical isolates of Streptoccus pneumoniae. Infect Immun. 2006;74(1):321–30. [PMC free article] [PubMed]
30. Ehrlich GD, Hu FZ, Shen K, Stoodley P, Post JC. Bacterial plurality as a general mechanism driving persistence in chronic infections. Clin Orthop Relat Res. 2005;(437):20–4. [PMC free article] [PubMed]
31. Hiller NL, Janto B, Hogg JS, Boissy R, Yu S, et al. Comparative Genomic Analyses of Seventeen Streptococcus pneumoniae Strains: Insights into the Pneumococcal Supragenome. J Bacteriol. 2007;189(22):8186–95. [PMC free article] [PubMed]
32. Shen K, Antalis P, Gladitz J, Sayeed S, Ahmed A, et al. Identification, distribution, and expression of novel genes in 10 clinical isolates of nontypeable Haemophilus influenzae. Infect Immun. 2005;73(6):3479–91. [PMC free article] [PubMed]
33. Tettelin H, Masignani V, Cieslewicz MJ, Donati C, Medini D, et al. Genome analysis of multiple pathogenic isolates of Streptococcus agalactiae:implications for the microbial “pan-genome”. Proc Natl Acad Sci U S A. 2005;102(39):13950–5. [PMC free article] [PubMed]
34. Bentley SD, Vernikos GS, Snyder LA, Churcher C, Arrowsmith C, et al. Meningococcal genetic variation mechanisms viewed through comparitive analysis of serogroup C strain FAM18. PLos Genet. 2007 Feb. 2007;16;3(2):e23. Epub 2006 Dec. 21. [PMC free article] [PubMed]
35. Aul JJ, Anderson KW, Wadowsky RM, Doyle WJ, Kingsley LA, et al. A comparative evaluation of culture and PCR for the detection and determination of persistence of bacterial strains and DNAs in the Chinchilla laniger model of otitis media. Ann Otol Rhinol Laryngol. 1998;107:508–513. [PubMed]
36. Ehrlich GD, Veeh R, Wang X, Costerton JW, Hayes JD, et al. Mucosal biofilm formation on middle-ear mucosa in the chinchilla model of otitis media. JAMA. 2002;3;287(13):1710–1715. [PubMed]
37. Post JC, Aul JJ, White GJ, Wadowsky RM, Zavoral T, et al. PCR-based detection of bacterial DNA after antimicrobial treatment is indicative of persistent, viable bacteria in the chinchilla model of otitis media. Am J Otolaryngol. 1996;17(2):106–111. [PubMed]
38. Buchinsky FJ, Forbes ML, Hayes JD, Shen K, Ezzo S, et al. Virulence phenotypes of low-passage clinical isolates of nontypeable Haemophilus influenzae assessed using the chinchilla laniger model of otitis media. BMC Microbiol. 2007;14;7:56. [PMC free article] [PubMed]
39. Hogg JS, Hu FZ, Janto B, Boissy R, Hayes J. Characterization and modeling of the Haemophilus influenzae core and supragenomes based on the complete genomic sequences of Rd and 12 clinical nontypeable strains. Genome Biol. 2007;5;8(6):R103 [Epub ahead of print]. [PMC free article] [PubMed]
40. McKessar SJ, Hakenbeck R. The two-component regulatory system TCS08 is involved in cellobiose metabolism of Streptococcus pneumoniae R6. J Bacteriol. 2007;189(4):1342–50. [PMC free article] [PubMed]
41. Dinarello CA. Proinflammatory and anti-inflammatory cytokines as mediators in the pathogenesis of septic shock. Chest. 1997;7;112(Suppl. 6):321s–329s. [PubMed]
42. Lacroix-Desmazes S, Bayry J, Kaveri SV, Hayon-Sonsino D, Thorenoor N, et al. High levels of catalytic antibodies correlate with favorable outcome in sepsis. Proc Natl Acad Sci U S A. 2005;102(11):4109–4113. [PMC free article] [PubMed]

Articles from PLoS ONE are provided here courtesy of Public Library of Science
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem chemical compound records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records. Multiple substance records may contribute to the PubChem compound record.
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.
  • Taxonomy
    Taxonomy records associated with the current articles through taxonomic information on related molecular database records (Nucleotide, Protein, Gene, SNP, Structure).
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...