Higher serum levels of interleukin 10 occur at onset of acute otitis media caused by Streptococcus pneumoniae compared to Haemophilus influenzae and Moraxella catarrhalis

Abstract

Objective

Acute otitis media (AOM) involves an inflammatory response to microbes in the middle ear that facilitates clearance of otopathogens. Clinically, Streptococcus pneumoniae (Spn) infections of the respiratory tract are characterized by greater inflammatory responses than nontypeable Haemophilus influenzae (NTHi) and Moraxella catarrhalis (Mcat). Interleukin 10 (IL-10) plays an important role in down-regulating the inflammatory response. We compared serum IL-10 levels in children before onset, at onset and after recovery from AOM caused by Spn, NTHi, and Mcat. We sought to determine if IL-10 could serve as a biomarker to distinguish AOM caused by Spn versus NTHi and Mcat.

Study Design

Prospective, longitudinal study in a primary care pediatric practice in Rochester, NY.

Methods

Participants were 54 children 6 to 30 months of age. Outcomes measured were serum IL-10 levels when healthy, at onset of AOM and after recovery from AOM.

Results

Serum IL-10 was elevated when children developed AOM (p=0.013) due to infections caused by Spn (p= 0.011) but not AOM caused by NTHi or Mcat. Middle ear fluid levels of IL-10 mirrored those seen in serum but were 10-fold higher (p=0.02). Other effector cytokines in serum: IL-4, IFN-γ and TNF-alpha, did not show the same increases as IL-10 at onset of AOM.

Conclusion

Our study indicates that AOM caused by Spn elicits a significantly higher IL-10 response compared to NTHi and Mcat and may prove to be a biomarker of AOM infections by Spn.

Introduction

Acute otitis media (AOM) is characterized by a local inflammation in the middle ear, nearly always in association with an upper respiratory viral infection (URI). The most common bacterial pathogens isolated from middle ear fluid (MEF) are Streptococcus pneumoniae (Spn) (25–50%), non-typeable Haemophilus influenzae (NTHi) (25–50%) and Moraxella catarrhalis (Mcat) (5–15%) ¹. Genes encoding a wide array of cytokines are differentially regulated in cells in the blood, and in cells that migrate from blood to the middle ear at the onset of AOM ². We have recently described variations in serum intercellular adhesion molecule 1 (sICAM-1) and S100A12 that correlated with the presence of Spn or NTHi in the middle ear, but not with Mcat or upper respiratory viral infection in children with AOM ^3,4

Interleukin 10 (IL-10) plays an important role in down-regulating the inflammatory response by inhibiting the release of immune mediators, phagocytosis, and the presentation of antigens to T-cells and macrophages ^5–7. In our recent evaluation of the transcriptome of peripheral blood mononuclear cells (PBMCs) in children infected with Spn and NTHi at the onset of AOM ², we found IL-10 was significantly up-regulated by Spn but not NTHi. The differential regulation of genes controlling cytokines, except for IL-10, were not significantly different when comparing Spn and NTHi infected children. The observation led us to hypothesize that the need to down-regulate the inflammatory response may be greater with Spn infections because Spn produces a greater inflammatory response compared to NTHi and Mcat.^8,9.

In this paper we sought to determine if IL-10 levels in serum might serve as a biomarker for AOM caused by Spn and differentiate such infections from those caused by NTHi or Mcat. We describe our assessment of IL-10 concentrations compared to IL-4, IFN-γ and TNF-α concentrations in serum of young children with AOM to determine (i) if IL-10 serum levels are elevated to greater levels when AOM is caused by Spn, the otopathogen that produces the greatest inflammatory response in the middle ear; (ii) if recovery from AOM is associated with a return of IL-10 levels to pre-infection levels in children infected with Spn; (iii) whether otitis prone children (who have impaired immune responses to otopathogen colonization and infection ^10–13 and other predisposing factors, such as presence of young siblings and out-of-home daycare) show similar or divergent patterns of IL-10 response as children with infrequent AOM; and (iv) levels of IL-10 in middle ear fluid of children with AOM caused by Spn and NTHi mirror the response seen in blood.

Methods

Subjects and sample collection

The 54 subjects were selected as a convenience sample from children 6 to 30 months of age who participated in an ongoing prospective study funded by National Institute of Deafness and Communication Disorders. The sample collection from children was approved by the IRB at Rochester General Hospital. At 6, 9, 12, 15, 18, 24, and 30 months of age the children had serum collected; these samples allowed assessment of levels of cytokines when children were healthy, prior to onset of AOM infection and at times when they had a viral URI without concurrent AOM. Whenever a child developed AOM, serum was obtained and a tympanocentesis performed to determine the etiology of AOM. The general results of tympanocentesis cultures from the overall study population has been previously published. ¹ Three weeks after an AOM episode, children returned for a convalescent blood sample.

The inclusion and exclusion criteria for study participation have been previously described. ¹ AOM was diagnosed using pneumatic otoscopy by validated otoscopists ¹⁴ according to the recommendations by the AAP ¹⁵ but with a requirement of a bulging tympanic membrane. Tympanocentesis was done to confirm bacterial AOM caused by Spn, NTHi or Mcat. The children were all treated with antibiotics directed to eradicate the otopathogens and when the convalescent blood sample was taken, all children were deemed to have recovered because they had no symptoms or signs of AOM although persistent middle ear effusion was considered normal. If the child experienced AOM three times within 6 months or 4 times within 12 months then they were classified as otitis prone. Children with immunodeficiency or chronic disease, or other infectious diseases were excluded from the study.

Clinical viral upper respiratory infection (URI) symptoms were assessed by a physician investigator. As described by Kalu et al ¹⁶, we defined a viral URI clinically by the occurrence of acute onset of rhinorrhea, cough, decreased appetite, and malaise with or without fever. The general results of viral cultures in the study population have been previously published. ¹⁷ All children in the study had a concurrent viral URI at onset of AOM.

Serum

Four milliliters of heparinized peripheral venous blood were collected from each child during every visit. The samples were centrifuged at 2,000 RPM for 10 minutes at room temperature. Peripheral blood mononuclear cells (PBMC) were isolated and serum was prepared and stored at −80°C until it was assayed.

MEF

MEF was collected from all children by tympanocentesis as previously described. ¹. Samples were processed within 3 hours of collection. Trizol (Sigma) was added to stabilize an aliquot of the MEF samples for RNA and DNA detection.

Bacteriology

Bacterial cultures were processed as previously described. ¹

Virology

Viral identification was done by a one-step RT-PCR reaction for RSV-A, RSV-B, Influenza-A, Influenza-B, and PIV-3 (Primerdesign, UK) using an iCycler (Biorad, CA) as previously described.¹⁷

MAP assay

IL-10, IL-4, TNF-alpha and IFN-γ levels in the serum were analyzed simultaneously by using a Bio-Plex Pro Assay (Bio-Rad Laboratories, Hercules, California, USA) which is a specific sandwich immunoassay formatted on magnetic beads. Capture antibodies directed against the cytokines were covalently coupled to beads. Coupled beads reacted with the sample containing the corresponding cytokine. After a series of washes to remove unbound protein, a biotinylated detection antibody was added to create a sandwich complex. The final detection complex was formed with the addition of streptavidin-phycoerythrin (SA-PE) conjugate. The protocol was performed according to the instruction of the manufacturer. Briefly, filter plates were wetted with 100 μl of assay buffer, and then 50 μl of 1x beads was added to the assay plate. After washing 2 times with 100 μl wash buffer, 50 μl samples diluted in 1:4 in buffer were added and incubated in the dark at RT with shaking at 300 RPM for 30 min. After washing 3 times with 100 μl wash buffer, 25 μl of detection antibody was added, incubated, and washed as above. Then 50 μl of streptavidin-PE was added and plates incubated for 10 min. After washing, the beads were resuspended in 125 μl of assay buffer and read with a Luminex instrument. The cytokine concentration was calculated based on a standard curve. MEF samples were analyzed in the same manner but only for IL-10 levels.

Statistical Analysis

Each subject in this study had cytokines measured under more than one clinical condition, e.g. at onset of AOM, during a well visit, in a comparison between otitis prone and non-otitis prone (Table 1). Therefore the observations were related (perhaps correlated) and could not be treated as if they were independent, precluding parametric tests such as t-tests. As recommended in this situation¹⁸ we used generalized estimating equations (GEE), which most resemble linear regression modeling, for the main analysis to calculate probability (p) values. The two largest values (> 20) were imputed to be the highest value less than 20 (= 5.5), after which approximate normality held. The Spn, NTHi and Mcat comparisons were fully paired by subject, while each remaining comparison contained multiple measurements from at least some subjects. Analysis of the relationship between serum and middle ear fluid levels of IL-10 were calculated using a Mann-Whitney test. For all tests p<0.05 was considered significant.

Table 1.

Subjects and Samples Tested for IL-10

	N Subjects	N Data
All	54	336
Spn+NTHi+Mcat AOM	44	47
Healthy	45	57
Spn AOM	21	21
Healthy	21	21
NTHi AOM	16	16
Healthy	16	16
M cat AOM	8	8
Healthy	8	8
Pre-infection	24	27
Convalescence	25	27
Otitis prone	7	7
Non-otitis prone	44	47
Virus and AOM	11	11
Non virus infected	33	35

Results

Serum IL-10 levels are significantly elevated in children with AOM

54 children and 336 samples were included in the study for serum IL-10 testing simultaneously with 3 other cytokines (IL-4, TNF-α and IFN-γ) in one or more of the analyses (Table 1). To determine levels when children were healthy compared to at onset of AOM, 47 samples from 44 healthy non-otitis prone children were compared with 57 samples from 45 non-otitis prone children with AOM due to Spn, NTHi or Mcat. IL-10 was detectable in all the samples. The level of IL-10 in children with AOM was significantly (p=0.013) higher than the level in children prior to AOM (Figure 1). In contrast, IL-4 could not be detected in 91% of samples, IFN-γ could not be detected in 74% of the samples and TNF-α could not be detected in 33% of the samples. Although TNF-α was detected in 67% of the samples, the levels were not significantly different at onset of AOM compared to healthy children (31.1 pg/ml, and 28.7 pg/ml, respectively). There was no correlation among concentrations of IL-10 and IL-4, TNF-alpha or IFN-γ (data not shown). Since IL-10 was the only cytokine significantly different in the serum of AOM patients among the cytokines surveyed, our subsequent analysis focused on IL-10.

Fig. 1.

Serum IL-10 levels in children are plotted as boxplots showing the median value, the 25^th percentile (box), 75^th percentile (bars) and outliers (dots) for each analysis. The vertical axis is the IL-10 measurements in picograms/milliliter.

Comparison at onset of AOM caused by Spn or NTHi or Mcat compared to the pre-infection healthy state. (n=39 and 45, respectively) p=0.013.

Comparison in children at onset of AOM caused by Spn versus the pre-infection healthy state. (n=15 and 15 respectively) p=0.011.

Comparison in children at onset of AOM caused by NTHi versus the pre-infection healthy state. (n=8 and 8, respectively) p=0.29

Comparison in children at onset of AOM caused by Mcat versus the pre-infection healthy state. (n= 8 and 8, respectively) p=0.85

Serum IL-10 levels are significantly increased in children with AOM due to Streptococcus pneumoniae

We had paired pre-infection and onset of infection samples from 21 non-otitis prone children who had AOM caused by Spn. We found a significant increase (p=0.011) in serum IL-10 levels from prior to infection compared to onset of infection (Fig. 1).

Serum IL-10 levels do not significantly change in children with AOM due to NTHi or Mcat

We had paired pre-infection and onset of infection samples from 16 non-otitis prone children who had AOM caused by NTHi and 8 children with AOM caused by Mcat. We found no significant increase in IL-10 levels at onset of infection compared to samples taken pre-infection (Figure 1).

Serum IL-10 levels return to pre-infection concentrations after recovery from infection

The serum level of IL-10 in 24 non-otitis prone children prior to AOM returned to pre-infection levels when measured 3 weeks after AOM onset in 25 children (Figure 2).

Fig. 2.

Serum IL-10 levels in children are plotted as boxplots showing the median value, the 25^th percentile (box), 75^th percentile (bars) and outliers (dots) for each analysis. The vertical axis is the IL-10 measurements in picograms/milliliter.

A IL-10 serum concentration prior to AOM infection compared to after recovery from infection (n=24 and 25, respectively) p=0.2

IL-10 serum concentrations in otitis prone (n=7) and non-otitis prone (n=44) children. P=0.8

Concurrent upper respiratory virus impact on IL-10 levels

All the children in the study had a concurrent clinical viral URI at onset of AOM. Therefore the difference in levels of IL-10 between Spn versus NTHi and Mcat were likely to be attributable to the bacteria causing AOM. From 11 of 44 tested NP samples non-otitis prone children with AOM, we detected a virus (5 cases of respiratory syncytial, 3 of influenza A, 1 of influenza B and 2 of parainfluenza virus), too few positive samples to make distinctions among viral species. Our methods did not detect other respiratory viruses including rhinovirus, metapneumovirus or bocavirus.

No significant difference of serum IL-10 levels in OP and NOP children

We compared IL-10 levels in 7 otitis prone children with 44 non-otitis prone children at the time of onset of AOM. There was an insignificantly lower level of IL-10 among otitis prone children (p=0.35) (Fig. 2).

Concurrence of serum and MEF IL-10 patterns

In children with AOM caused by Spn or NTHi, serum and MEF levels of IL-10 were compared. The mean IL-10 level in MEF for 4 samples taken from non-otitis prone children that had Spn in the MEF was 314 pg/mL (S.E. = 195 pg/mL) whereas the serum level was 4.8 pg/mL (S.E.= 0.9 pg/mL). (p=0.02) The mean IL-10 level in MEF for 4 samples taken from non-otitis prone children that had NTHi in the MEF was 178 pg/mL (S.E. = 95 pg/mL) whereas the serum level was 3.5 pg/mL (S.E.= 0.7 pg/mL). (p=0.02) (Fig. 3).

Fig. 3.

IL-10 MEF and serum levels in children with AOM caused by Spn (n=4) or NTHi (n=4). MEF levels for IL-10 were significantly higher in MEF than serum for both Spn and NTHi, (p=0.02 for both).

Discussion

In this assessment of IL-10 concentrations compared to IL-4, IFN-γ and TNF-α concentrations in serum of young children with AOM we provide data to support several conclusions. (1) IL-10 but not IL-4, IFN-γ or TNF-α serum levels are elevated at onset of AOM. (2) The elevation of IL-10 in serum at onset of AOM is statistically associated with infections caused by Spn, the otopathogen that produces the greatest inflammatory response in the middle ear. Similar elevations are not seen in infections caused by NTHi or Mcat. (3) Recovery from AOM caused by Spn is associated with a return of IL-10 levels to pre-infection levels. (4) Children prone to recurrent AOM infections show similar patterns of IL-10 response as children with infrequent AOM. (5) Levels of IL-10 in middle ear fluid of children with AOM caused by Spn and NTHi mirror the response seen in blood.

Differences in elevation of IL-10 in children at onset of AOM caused by Spn infection are consistent with the known role of IL-10 in dampening the innate and adaptive immune response so as to avoid inflammatory damage to tissues. Spn is more invasive and causes systemic infection unlike NTHi and Mcat. Spn is known to cause more severe symptoms and signs of inflammation than other otopathogens in AOM. Van der Poll et al ¹⁹ found that administration of Spn intranasally in mice caused increased production of IL-10 in the lungs during pneumococcal pneumonia. IL-10 levels reached 25.8 ± 3.0 ng/ml in the mice compared to pre-infection levels of 1.2 ± 0.2 ng/ml. Concentrations of IL-10 in middle ear effusions and plasma of adults with otitis media showed a strong correlation with the duration of effusion²⁰.

The role of increased IL-10 in bacterial infection is sometimes complex, varying with the pathogen, the host and the interaction with other anti-inflammatory and pro-inflammatory cytokines. Lee et al ²¹ found that IL-10 injected into the middle ear cleft of mice inhibited AOM induced by lipopolysaccharide (LPS) by reducing the mucosal and sub-mucosal infiltration of inflammatory cells. Oral et al ²² described that IL-10 could inhibit the CD28 signaling pathway and protect the middle ear mucosa from damage by limiting the activation of T-cells and natural killer cells and the general inflammation response. Diab et al ²³ demonstrated elevated IL-10 in Haemophilus influenzae b-infected rat brains at 8 hour, but at 18h of infection IL-10 mRNA expression declined to levels below that of uninfected control rats whereas in Spn inoculated rats, IL-10 mRNA induction was elevated at both 8h and 18h of infection.

Emonts et al ²⁴ and Ilia et al ²⁵ showed that children with lower IL-10 levels due to an IL-10-1082 A/A polymorphism were less susceptible to recurrent episodes of AOM, leading to the notion that higher IL-10 levels signify a greater likelihood for chronic and AOM prone conditions. Our results do not support those observations since we found no significant difference in serum IL-10 levels in otitis prone and non-otitis prone children.

We did not assess the effect of viral infections on IL-10 levels at onset of infection because all of the children we studied had a concurrent viral URI clinically diagnosed and a limitation of our study is the absence of such an assessment and reliance on a clinical diagnosis for definition of viral URI. In a subset of children the presence of a virus was detected. It is known that the most common viruses to cause upper respiratory infections are respiratory syncytial virus (RSV) and human rhinovirus (HRV) ²⁶. Previous studies show that RSV and HRV are able to infect human immune cells and trigger them to release IL-10 inhibiting the host immune response and resulting in a longer and more severe respiratory infection.²⁷ Skovbjerg et al. ²⁸ found that IL-10, TNF-alpha and IL-8 were unrelated to the presence or absence of viruses. Another limitation of our study is the small number of subjects used to compare levels of IL-10 in serum versus middle ear fluid.

Previous studies showed variable detection of IFN-γ, IL-4 and TNF-alpha in children with AOM. Kele ²⁹ reported IFN-γ levels were significantly lower and IL-4 levels significantly higher in plasma samples from AOM patients compared to controls. Johnston et al ³⁰ found no significant difference in plasma IL-4 but significantly elevated levels IFN-γ comparing AOM cases to controls in older children (mean = 4.5 years old). The difference in results from those studies compared to ours may be due to age of the child, certainty of AOM diagnosis (since those studies did not confirm AOM with tympanocentesis) and methods applied. With xMAP technology, we found IFN-γ and IL-4 were largely undetectable whereas TNF-alpha was detected in 67% of serum samples but no difference was found between the AOM and healthy state. Willett et al ³¹ detected TNF-alpha by ELISA in 40% of 65 middle ear effusions from 41 children with chronic otitis media with effusion; the mean concentration was 1.24 +/- 3.1 pg/mg. Skotnicka et al ³² detected TNF-alpha in 47.4% of middle ear effusions tested.

Conclusion

In conclusion, at onset of AOM a significant rise in serum IL-10 levels occurs when the infection is caused by Spn but not NTHi or Mcat. When children recover from AOM caused by Spn, IL-10 serum levels return to pre-infection values. This result suggests that IL-10 may be a useful biomarker to assist in diagnosis of AOM caused by Spn in young children (6–30 months of age). Serum IL-6 may be a similar cytokine biomarker for AOM caused by Spn ³³. The relatively high production of IL-10 at onset of AOM caused by Spn (Gram positive infections) but not by NTHi and Mcat (Gram negative pathogens) is consistent with the clinical observation of a greater inflammatory response that requires dampening in association with Spn infection compared to NTHi and Mcat infection.