Genetic Relatedness among Nontypeable Pneumococci Implicated in Sporadic Cases of Conjunctivitis

Abstract

Nontypeable Streptococcus pneumoniae is a common cause of epidemic conjunctivitis. A previous molecular fingerprinting study identified a clone of nontypeable pneumococcus that was responsible for a recent outbreak of conjunctivitis. In the present study, we examined the extent to which pneumococci that cause sporadic cases of conjunctivitis are related to this epidemic strain. Using arbitrarily primed BOX-PCR, we have determined that, of 10 nontypeable pneumococci causing sporadic conjunctivitis, 5 were clonal and closely related to a previous outbreak strain, whereas 5 others were genetically diverse.

Nontypeable Streptococcus pneumoniae is a frequent cause of epidemic bacterial conjunctivitis (9, 12). We have previously utilized BOX-PCR fingerprinting to identify clones of nontypeable pneumococcus responsible for several outbreaks of conjunctivitis (1). In this study, we applied this technique to nontypeable pneumococci implicated in sporadic cases of conjunctivitis in a single city during a 6-month period. To our knowledge, this is the first attempt to determine whether sporadic nontypeable pneumococci that cause conjunctivitis are genetically homogeneous and whether they are related to epidemic strains.

MATERIALS AND METHODS

Cases.

In order to characterize the antibiotic susceptibility of S. pneumoniae, the Clinical Microbiology Laboratory of the Louisiana State University Medical Center in Shreveport, La. (LSUMC), stored 94 consecutive isolates of this organism obtained from all culture sources between 2 February 1995 and 31 July 1995.

Eighteen of the 94 isolates were from specimens obtained by individual health care providers as part of the clinical evaluation of patients with conjunctivitis. An additional five isolates collected during this period from sites other than the conjunctivae were randomly selected for this study. A strain from a previously described outbreak of pneumococcal conjunctivitis in Illinois in 1996 (hereinafter termed the Illinois isolate) (1) had been stored in the Infectious Disease Research Laboratories, Veterans Affairs Medical Center, Houston, Tex., in tryptic soy broth containing 15% glycerol at −70°C.

Specimen collection and initial laboratory processing.

All eye specimens, a throat specimen (sample 19), and an abdominal drainage specimen (sample 20) were collected with sterile swabs. Swabs were placed in transport medium (Amies with charcoal; Remel, Inc., Lenexa, Kans.) and sent to the Clinical Microbiology Laboratory at LSUMC. Isolates were stored at −70°C in Trypticase soy broth containing 15% glycerol.

Identification and serotyping.

Isolates were identified as S. pneumoniae on the basis of typical alpha-hemolysis on Trypticase soy agar plates and by optochin sensitivity and bile solubility. For serotyping, colonies were harvested and suspended in phosphate-buffered saline with 2% formalin. Samples were tested against antiserum pools A, B, C, D, E, F, G, H, I, P, Q, R, S, and T by a method adapted (11) from that of Kronvall (8). Typeable specimens were those that agglutinated within 2 min. Typing within serogroups was performed with specific factor sera to produce the capsular swelling (Quellung) reaction. All antisera were obtained from Statens Seruminstitut (Copenhagen, Denmark). Autoagglutinated samples were noted to have formed clumps in the formalin solution, and when they were placed on glass plates for the subsequent addition of antisera, they had the appearance of a positive serotyping reaction. Consequently, serotyping could not be completed. These isolates were identified as nontypeable. The identities of these isolates were verified by detecting rRNA unique to S. pneumoniae with an Accuprobe S. pneumoniae identification kit (Gen-Probe, Inc., San Diego, Calif.).

BOX-PCR.

Strains were grown on Trypticase soy agar overnight in a candle jar at 37°C, and colonies were removed with bacterial loops, suspended in distilled water, washed, and resuspended to a standard turbidity corresponding to 10⁷ to 10⁸ CFU ml⁻¹. BOX-PCR with the BOXA1R primer (7) was then performed by a procedure modified from that of Ertugrul et al. (1). An aliquot of 7.4 μl of each PCR product was mixed with 6× loading buffer (Life Technologies, Grand Island, N.Y.) and then subjected to electrophoresis in a 1% agarose gel at 2.5 V cm⁻¹ for 3 h. Gels were stained in 5 μg of ethidium bromide ml⁻¹ and viewed and photographed under UV light. Because BOX-PCR produces bands of varied intensity, strains were considered identical if they differed only by the absence of one band. Gel banding patterns were also analyzed with RFLPscan software (Scanalytics, Billerica, Mass.) with a tolerance of 2%. A database was created by the “bin” method, which records the presence or absence of bands at given molecular weights. Dendrograms were then produced from these data with the TreeCon algorithm (Yves Van de Peer, University of Antwerp).

RESULTS

Clinical characteristics of patients and serotypes.

The 18 eye isolates available for study were obtained from unrelated patients at different clinics in the absence of an apparent outbreak of conjunctivitis. Clinical data are summarized in Table 1. One isolate (sample 6) was from a 23-year-old adult; the others were from young children (mean age, 37 months; median age, 19 months; range, 2 months to 14 years). Of the 18 strains implicated in sporadic conjunctivitis, 8 were typeable and 10 were nontypeable (autoagglutinated).

TABLE 1.

Clinical data from 18 patients with pneumococcal conjunctivitis^b

Samplea	Type	Source	Diagnosis or diagnoses	Date of isolation (mo/day/yr)	Age
1	AA	Eye	Conjunctivitis	5/17/95	14 yr
2	AA	Eye	Conjunctivitis	6/5/95	13 mo
3	AA	Eye	Conjunctivitis, otitis media	3/2/95	17 mo
4	AA	Eye	Conjunctivitis	3/17/95	29 mo
5	AA	Eye	Conjunctivitis	5/9/95	8 yr
6	AA	Eye	Conjunctivitis	3/20/95	23 yr
7	AA	Eye	Conjunctivitis	4/27/95	4 yr, 10 mo
8	AA	Eye	Conjunctivitis	3/3/95	16 mo
9	AA	Eye	Conjunctivitis, rhinitis	3/11/95	2 mo
10	AA	Eye	Conjunctivitis, URI	5/1/95	6 mo
11	3	Eye	Conjunctivitis	5/1/95	2 mo
12	6B	Eye	Conjunctivitis	2/21/95	30 mo
13	8	Eye	Conjunctivitis, otitis media	6/23/95	21 mo
14	11A	Eye	Conjunctivitis, pharyngitis, rhinitis	2/23/95	31 mo
15	23F	Eye	Conjunctivitis	2/28/95	19 mo
16	6B	Eye	Conjunctivitis, URI	3/6/95	3 mo
17	6B	Eye	Conjunctivitis	4/6/95	7 yr
18	I	Eye	Conjunctivitis, URI, thrush	5/10/95	3 mo
Illinois	AA	Eye	Epidemic conjunctivitis	?/?/96
19	3	Throat	Sinusitis, pharyngitis	6/12/95	25 yr
20	9V	Abdominal drainage	Abdominal cellulitis	4/3/95	28 yr
21	19F	Blood and CSF	Sepsis, meningitis	3/22/95	2 mo
22	23F	Blood	Septic shock (fatal)	3/10/95	52 yr
23	9V	Sputum	Burn patient	2/27/95	46 yr

^aThe Illinois sample represents the genotype identified in cases of epidemic conjunctivitis among military recruits (1). This sample and samples 19 through 23 are included for comparison. 

^bAA, an autoagglutinating strain; CSF, cerebrospinal fluid; URI, upper respiratory tract infection. 

S. pneumoniae was not the only organism grown from most clinical isolates. Many contained small amounts of commensal bacteria not implicated in conjunctivitis (Corynebacterium species, coagulase-negative staphylococcus, and Bacillus species) (12), but potentially pathogenic bacteria were occasionally isolated; four isolates contained Haemophilus influenzae (samples 9, 10, 11, and 16), three isolates contained Staphylococcus aureus (samples 10, 13, and 14), and one contained Moraxella catarrhalis (sample 18). In only two samples (samples 9 and 18) did one of these other organisms outnumber pneumococci (H. influenzae in both cases).

With 7 of the 18 patients with conjunctivitis, at least one other concurrent nonocular infection was diagnosed clinically. Two of these seven patients had otitis media, and one patient had pharyngitis. Five of the seven patients were diagnosed with concurrent upper respiratory tract infections or rhinitis. For 11 patients, conjunctivitis was the sole final diagnosis.

Five strains of nonconjunctival, typeable S. pneumoniae isolated during this time were from a variety of infections: one from blood and cerebrospinal fluid, one from blood alone, one from the tonsils, one from a case of abdominal cellulitis, and one from sputum.

PCR.

A gel was run with the 10 nontypeable isolates from LSUMC and two nontypeable isolates from the conjunctivitis outbreak in Illinois (Fig. 1). Four distinct clones were identified among the LSUMC isolates (samples 1 to 5, 6, 7, and 8 to 10); none of these clones was identical to the Illinois outbreak strain, although the clone represented by isolates 1 to 5 was very similar to it.

FIG. 1.

BOX-PCR fingerprint of 10 strains of nontypeable pneumococcus that caused sporadic conjunctivitis and of 2 isolates from an outbreak in Illinois.

In order to construct a dendrogram to depict the relative levels of similarity of the strains, two more gels were run with all 24 strains listed in Table 1. The outbreak isolate from Illinois was run on each gel to facilitate comparison, and a variety of nonconjunctival isolates were included to provide perspective for the assessment of similarity (Fig. 2). Samples 1 to 5, the most prevalent clone in this small sample, were relatively similar to the outbreak strains. However, the other five nontypeable strains were not similar to the outbreak strains. Specifically, the strain that caused an ocular infection in an adult, sample 6, was not similar to the outbreak strain, which was isolated from adult patients.

FIG. 2.

Dendrogram of the 24 pneumococcal strains. AA, autoagglutinating strains; CSF, cerebrospinal fluid.

DISCUSSION

Shayageni et al. (9) used an array of biochemical, immunologic, and microbiological tests to demonstrate that a closely related strain of nontypeable S. pneumoniae was responsible for outbreaks of conjunctivitis in New York, Illinois, and California in 1980 and 1981. Since then, advances in molecular biology have enhanced our ability to discriminate among isolates implicated in a variety of infections caused by many pathogens, including S. pneumoniae (4). Random amplification by PCR with the BOX sequence has compared well with other more established methods of molecular typing such as restriction fragment end labeling and pulsed-field gel electrophoresis (13). We (1) have previously used this technique to confirm the clonality of the strain studied by Shayageni et al. and to document its similarity to isolates of nontypeable S. pneumoniae implicated in outbreaks in Illinois in 1996.

In this study, pneumococci causing sporadic cases, even the nontypeable strains, were diverse. The preservation of an outbreak clone in the context of the diversity noted in this study is consistent with the epidemic population structure of pneumococci as defined by Hall et al. (5): a diverse collection of freely recombining strains with occasional well-conserved and virulent clones. It should be noted that previously studied outbreaks have been among adults but that our sporadic cases occurred in a much younger population. Differences in the susceptibility to infection between adults and children may explain how the pneumococcal conjunctivitis strains fit into such a population structure. Young patients are probably more susceptible to otitis media and upper respiratory tract infections due to their naive immune systems and unique anatomies. In adults, who are not as susceptible, pneumococci able to cause an outbreak would require a more specific armamentarium of virulence factors and, hence, might be clonal. It is interesting that a clone similar to the virulent outbreak strain was responsible for one-half of the nontypeable conjunctivitis seen in this study.

Nontypeable pneumococci rarely cause invasive infection (2, 3), but for unclear reasons, they are frequently implicated in cases of pneumococcal conjunctivitis (1, 9, 10, 12). While it is still uncertain whether the nontypeable strains are always unencapsulated, recent studies of phase variation in pneumococci suggest that decreased capsule production may facilitate bacterial persistence in certain tissues. Kim and Weiser and Weiser et al. (6, 15) have observed that colonies with decreased amounts of polysaccharide capsule are adept at nasopharyngeal colonization but that variants with greater amounts of polysaccharide capsule are adept at invasion. While there is no evidence that the conjunctival strains studied here undergo phase changes, these studies suggest possible tissue-specific advantages of decreased or absent capsule production. We have found (8a) that transposon-induced unencapsulated mutants of S. pneumoniae types 3 and 14 adhere far more avidly to mammalian cells in tissue culture than do encapsulated wild-type strains. While capsule is important to the ability of pneumococci to cause invasive disease (14), it may be irrelevant to, or even an impediment to, conjunctival infection.