TEXT

Top Abstract Text References

Nontypeable Haemophilus influenzae is a common commensal organism of the human nasopharynx and an important cause of a wide range of respiratory illnesses, including acute otitis media, otitis media with effusion, sinusitis, pneumonia, and bronchitis (36). Nontypeable strains are nonencapsulated and are differentiated from strains that produce a polysaccharide capsule by the absence of agglutination with antisera directed against the six capsular serotypes (a to f) (23). Up to 80% of normal children are colonized with nontypeable strains at any given time, and longitudinal studies have demonstrated that colonization with a particular strain typically lasts from several weeks to months (11, 26).

Although the precise relationship between nasopharyngeal colonization and middle ear disease remains unclear, compared with normal children, otitis-prone subjects are usually colonized earlier, more frequently, and for longer durations with nontypeable H. influenzae and other potential respiratory pathogens, such as Streptococcus pneumoniae and Branhamella catarrhalis (6, 11, 25). Several investigators have examined paired isolates of H. influenzae recovered simultaneously from the nasopharynx and middle ear fluid of patients presenting with acute otitis media. Based on analysis of outer membrane protein profiles and genomic fingerprints, most paired isolates are isogenic, leading to the hypothesis that strains causing acute otitis media likely migrate by direct extension from sites in the nasopharynx through the eustachian tube to the middle ear, possibly facilitated by eustachian tube dysfunction (16, 20).

Successful colonization of the nasopharynx requires that the organism overcome mucociliary clearance, and adherence to respiratory epithelium is presumed to be a critical step in this process. In nontypeable H. influenzae, a number of adherence factors have been identified, including the HMW1/HMW2 family of proteins, Hia, and hemagglutinating pili (3, 10, 22, 29). The HMW1 and HMW2 adhesins mediate adherence to a wide range of human epithelial cell lines (12, 29). These proteins, which were first discovered in nontypeable H. influenzae strain 12, exhibit significant sequence homology to each other, with 71% identity and 80% similarity overall (2). Related proteins are expressed by the majority of nontypeable strains but are absent in encapsulated H. influenzae (2, 33). The hmw1 structural gene is 4.6 kb and encodes a 160-kDa protein, while the hmw2 structural gene is 4.4 kb and encodes a 155-kDa protein (2). The Hia adhesin also promotes adherence to a broad range of human epithelial cell lines (31). This protein, which was originally identified in nontypeable H. influenzae strain 11, is expressed by most nontypeable strains that lack HMW1/HMW2-like proteins (3, 33). Hia is encoded by the hia gene, which is homologous to the hsf locus that is ubiquitous among encapsulated strains of H. influenzae (31). Hemagglutinating pili are long composite fibers that mediate agglutination of AnWj-positive erythrocytes and adherence to buccal epithelial cells (10, 21, 38). These organelles are expressed by both nontypeable and serotypeable strains (8, 34). The pilus gene cluster is composed of five genes: hifA, which encodes the major structural subunit, and hifB to hifE, which encode assembly proteins and minor structural subunits (18, 19, 30, 40, 41). To further explore the relationship between these adhesins and middle ear disease, we examined matched nasopharyngeal and middle ear isolates from patients with acute otitis media. In particular, we assessed the genetic relatedness of matched isolates and the relationship between the site of isolation and the prevalence of expression of HMW1/HMW2-like proteins, Hia, and pili.

For this study, all isolates of H. influenzae were recovered from children diagnosed with acute otitis media between 1 January 1996 and 30 April 1997. Patients ranged in age between 7 months and 7 years (mean ± standard deviation, 2.05 ± 2.47 years). Nasopharyngeal culture specimens were obtained with Dacron swabs, and middle ear isolates were recovered by tympanocentesis during the same visit. Institutional review board approval was obtained for this study, and parental consent was obtained prior to collection of samples in all cases. Isolates were determined to be H. influenzae on the basis of colony morphology, appearance as gram-negative coccobacilli, and a requirement of factors X and V for aerobic growth. Strains were defined as nontypeable based on their failure to react with a polyvalent antiserum directed against the H. influenzae serotype a to f capsular polysaccharides (Difco, Detroit, Mich.). The results of serologic assays were confirmed by Southern blotting with the plasmid pUO38, which contains the cap b locus and gives a typical pattern of hybridization with serotype a to f strains (14, 15). All isolates were subcultured a maximum of two to three times on chocolate agar plates supplemented with 1% Isovitalex prior to being frozen and stored at 80°C in brain heart infusion broth with 20% glycerol (1, 27). Control strains included nontypeable H. influenzae strain 12, from which the hmw1 and hmw2 genes were originally cloned (2); nontypeable H. influenzae strain 11, from which hia was originally isolated (3); and H. influenzae type b strain Eagan, which contains an intact pilus gene cluster (18). Strain 12 lacks hia and pilus genes, strain 11 lacks hmw genes and pilus genes, and strain Eagan lacks hmw genes (3, 33).

To establish whether nasopharyngeal and middle ear isolates from a given child were isogenic, we compared the total protein profiles and genomic fingerprints of the paired specimens. To obtain total protein profiles, whole-cell lysates were prepared as previously described (24) and were resolved by electrophoresis on 10% sodium dodecyl sulfate-polyacrylamide gels. To generate genomic fingerprints, chromosomal DNA was digested to completion with EcoRI or XbaI and the resulting fragments were electrophoresed on 0.7% agarose gels, stained with ethidium bromide, and visualized on a UV transilluminator. In all 17 patients, including 11 with bilateral disease, nasopharyngeal and middle ear isolates were identical. Interestingly, isolates recovered from two neighbors (cousins) with concurrent infection were identical, suggesting acquisition from a common source or transmission from one child to the other (Fig. 1). The remaining nasopharyngeal and middle ear isolates recovered from each patient were genetically distinct from every other matched set (Fig. 1).

View larger version (80K): [in this window] [in a new window]   FIG. 1.   Analysis of paired nasopharyngeal (N/P) and middle ear (ME) isolates of nontypeable H. influenzae recovered from two neighbors with concurrent otitis media and comparison with isolates from two other patients. (A) Whole-cell lysates of the matched N/P and ME isolates were resolved on sodium dodecyl sulfate-0.7% polyacrylamide gels and stained with Coomassie brilliant blue. The positions of molecular mass standards are shown to the left of the gel. (B) Chromosomal DNA from the matched N/P and ME isolates was digested with EcoRI and resolved on a 0.7% agarose gel. A 1-kb ladder (Gibco BRL, Gaithersburg, Md.) is in the leftmost lane (St). (C) Chromosomal DNA from the matched N/P and ME isolates was digested with XbaI and then resolved on a 0.7% agarose gel. For all three gels (A, B, and C), samples were loaded in the same order. A 1-kb ladder is in the leftmost lane. Lanes: 1 to 3, samples of the N/P, LME, and RME isolates, respectively, from patient 6; 4 to 6, samples of the N/P, LME, and RME isolates, respectively, from patient 7; 7 to 9, samples of the N/P, LME, and RME isolates, respectively, from patient 8; 10 and 11, samples of the N/P and LME isolates, respectively, from patient 10. Patients 6 and 7 were neighbors (as well as cousins) and had concurrent infections.

Based on Western analysis with antiserum 25G, which was raised against recombinant HMW1 and is reactive with both HMW1 and HMW2 (2), 14 of the 17 nasopharyngeal isolates expressed one or two HMW1/HMW2-like proteins (Table 1). In all cases, matched middle ear isolates had the same pattern of reactivity as their nasopharyngeal counterparts. Southern blotting with a 3.2-kb SpeI-EcoRI fragment that corresponds to the promoter and 5' coding sequence of the HMW1 structural gene (hmw1A) of nontypeable H. influenzae strain 12 was used as a probe for hmw genes and confirmed the presence of hmw genes in all isolates that expressed HMW1/HMW2-like proteins (this probe hybridizes with both the HMW1 and HMW2 structural genes) (Table 1). In contrast, there was no evidence of hmw genetic material in isolates that failed to express HMW proteins (Table 1).

Western analysis with antiserum 36B, which was raised against the Hia protein of nontypeable strain 11 (3), demonstrated the presence of an Hia homolog in 2 of the 17 nasopharyngeal isolates, both of which failed to express HMW protein (Table 1). Again the matched middle ear isolates also expressed Hia. Southern blotting with a 1.6-kb SspI-StyI fragment that represents an intragenic piece of the nontypeable H. influenzae strain 11 hia locus was used as a probe for the hia gene and confirmed the presence of hia sequences in the two strains that expressed Hia (Table 1). Interestingly, two additional strains, including one with hmw genes, also hybridized with hia under conditions of low stringency (28) (Table 1).

To assess the presence of hemagglutinating pili, isolates were examined initially by a semiquantitative assay for the ability to agglutinate AnWj-positive human erythrocytes (22). Two nasopharyngeal isolates were hemagglutination positive, including one that lacked both HMW proteins and Hia and a second that expressed Hia (Table 1). In both cases, hemagglutination was inhibited by preincubation of the bacteria with a final concentration of 100 µg of the sialylated ganglioside GM₁ (Sigma, St. Louis, Mo.) per ml, supporting the conclusion that these isolates express hemagglutinating pili (37). Consistent with these findings, negative-staining electron microscopy, performed as previously described, revealed the presence of peritrichous fibers (data not shown) (28). In both isolates, Western analysis with antisera raised against the HifA, HifB, and HifD proteins of H. influenzae type b strain Eagan revealed reactive proteins (Table 1) (32). Interestingly, the matched middle ear isolates were found to lack pili when examined by electron microscopy and failed to react with antisera raised against HifA, HifB, and HifD (Table 1). In addition, these isolates were recalcitrant to enrichment for piliation (5), suggesting that they are not simply nonpiliated phase variants. Southern analysis with separate probes corresponding to the intact hifA, hifB, hifC, hifD, and hifE genes from strain Eagan demonstrated the presence of all five hif genes in the two nasopharyngeal isolates and in the matched middle ear organisms (Table 1). The patterns of hybridization were identical for the corresponding nasopharyngeal and middle ear isolates (Fig. 2), arguing against a major genetic rearrangement as an explanation for the lack of pili. Hybridization with the five hif probes was uniformly negative for all of the remaining strain sets (Table 1).

View larger version (41K): [in this window] [in a new window]   FIG. 2.   Southern analysis of matched nontypeable H. influenzae isolates from patients 1 and 2, probing with hifA to hifE genes from H. influenzae type b strain Eagan. Chromosomal DNA was digested with either EcoRI (hifA, -B, -D, and -E blots) or BglII (hifC blot) prior to being probed. All gels were loaded as follows: lane 1, strain Eagan (positive control); lane 2, strain 11 (negative control); lane 3, patient 1, nasopharyngeal specimen; lane 4, patient 1, left middle ear specimen; lane 5, patient 1, right middle ear specimen; lane 6, patient 2, nasopharyngeal specimen; and lane 7, patient 2, right middle ear specimen. The positions of molecular size standards are indicated on the left.

The dynamic relationship between nasopharyngeal colonization with H. influenzae and subsequent otitis media has yet to be fully characterized. In the present study, we found that in each of 17 patients with nontypeable H. influenzae acute otitis media, the same strain was recovered from both the nasopharynx and middle ear fluid. This observation is consistent with previous reports (16, 20) and supports the belief that nasopharyngeal colonization is a prerequisite for otitis media. Furthermore, it reinforces the conclusion that patients with nontypeable H. influenzae-associated acute otitis media harbor a single predominant strain in the nasopharynx (20).

Several important adherence factors have been identified in nontypeable H. influenzae, including the HMW1/HMW2-like proteins, Hia, and hemagglutinating pili (3, 10, 22, 29). In this study, we found that HMW1/HMW2-like proteins predominated among isolates associated with otitis media. In contrast, Hia and pili were uncommon and were detectable only in isolates lacking HMW proteins. One possible explanation for these observations is that strains expressing HMW proteins colonize the nasopharynx more efficiently. Alternatively, these strains may be endowed with virulence factors that enhance their spread through the eustachian tube to the middle ear. Analysis of the prevalence of HMW1/HMW2-like proteins in nasopharyngeal isolates from healthy children may help to distinguish between these possibilities.

Southern analysis revealed pilus genes in only 2 of the 17 strains that we examined. Given that our hybridization conditions were of low stringency, strains with sequences slightly divergent from those in H. influenzae strain Eagan were unlikely to be hybridization negative. Our results differ from those of Brinton and coworkers, who reported that up to 95% of middle ear isolates of H. influenzae contain pilus genes (4). One possible explanation for this discrepancy is that Brinton et al. might have used a probe containing DNA that flanks the hif genes and is thus common to almost all strains of H. influenzae. Alternatively, strain sets collected from separate geographic locations may be fundamentally different from one another. Consistent with our observations, Geluk et al. recently examined 83 strains of nontypeable H. influenzae from patients with diverse clinical syndromes and found that 18% possessed pilus genes (7). In the subset of strains from middle ear fluid, 13% had detectable pilus genes (7).

In our study, among the patients colonized with piliated organisms, middle ear isolates were uniformly nonpiliated. In this context, it is noteworthy that expression of hemagglutinating pili is subject to reversible phase variation (5, 22, 39). For example, during natural infection with H. influenzae type b, nasopharyngeal isolates are often piliated while their isogenic counterparts from systemic sites are virtually always nonpiliated (17). At the same time, systemic isolates can be enriched in vitro for piliated variants which themselves can revert to nonpiliated forms (9, 34). The mechanism underlying the change between piliated and nonpiliated forms relates to variation in the number of TA repeats in the region of overlap between the promoters of the divergently transcribed hifA and hifB genes (39). It is of note that in the two middle ear isolates that we characterized, we were unable to enrich for piliated variants. This recalcitrance was not due to a major disruption of the pilus gene cluster, since Southern blotting revealed that the nasopharyngeal and middle ear isolates had the same pattern of hybridization. Other possible explanations include a shift in TA repeat number beyond the limits associated with reversibility, as well as small deletions related to the repetitive extragenic palindromic sequences that have been identified throughout the pilus gene cluster (40). Regardless of the mechanism by which the capacity for pilus expression is eliminated, piliation may be selected against beyond the nasopharynx, perhaps because pili augment the oxidative-burst response of professional phagocytes, facilitating immunologic clearance (35).

In considering whether there is a correlation between the presence of hemagglutinating pili in nontypeable strains and their anatomic location, Brinton et al. found that in their collection of nontypeable H. influenzae strains, 35% of nasopharyngeal isolates, compared with only 5% of middle ear isolates, were piliated (4). In a similar study, Gilsdorf et al. characterized the capacity for expression of pili in a collection of 38 strains from a variety of patients. After enrichment, seven of nine isolates from the nasopharynx or throat were piliated, while only three of eight isolates from middle ear fluid were capable of expressing pili (8). Although these studies did not compare matched isolates, they reinforce our observations and suggest that pili may contribute to upper respiratory tract colonization and then become dispensable as the organism migrates through the eustachian tube.

In earlier work we demonstrated that encapsulated strains of H. influenzae possess an allelic variant of hia (31). In serotype b strains, the counterpart to hia is associated with expression of short, thin surface appendages referred to as fibrils and is called hsf (31). In a separate study involving analysis of a collection of nontypeable strains defined in terms of genetic relatedness, we found a correlation between the presence of hia and the presence of the insertion sequence IS1016 (33). Since IS1016 is associated with the capsule locus (cap) in most encapsulated strains of H. influenzae (14), we speculated that nontypeable strains containing hia evolved more recently from an encapsulated ancestor. In the collection of strains that we examined in this study, both strains with pilus genes also contained an hia gene. Furthermore, one of these strains hybridized with IS1016 (13). These observations suggest that the hia/hsf and hif genes may have coevolved and that they together define a distinct lineage within the population structure of nontypeable H. influenzae.

In conclusion, in this study of matched nasopharyngeal nontypeable H. influenzae from children with acute otitis media, we found that organisms in the nasopharynx uniformly expressed either HMW1/HMW2-like proteins, Hia, or hemagglutinating pili. The overall prevalence of these adherence factors argues for their importance in the pathogenesis of otitis media. The mechanism by which they interact with other virulence factors remains to be determined.