Multiserotype Protection of Mice against Pneumococcal Colonization of the Nasopharynx and Middle Ear by Killed Nonencapsulated Cells Given Intranasally with a Nontoxic Adjuvant

Abstract

Intranasal challenge of C57BL/6 mice with Streptococcus pneumoniae serotypes 6B, 14, and 23F produced colonization of the middle ear and NP. Intranasal vaccination with ethanol-killed nonencapsulated cells with adjuvant protected both sites. Of four nontoxic adjuvants tested, the cholera toxin B subunit was most effective and least nonspecifically protective.

Licensed vaccines against Streptococcus pneumoniae consist of injectable polyvalent mixtures of serotype-specific capsular polysaccharides or capsular polysaccharide-protein conjugates and are therefore effective only against serotypes included. To circumvent this limitation, investigators have proposed that protective antigens common to all pneumococcal serotypes (so-called “species antigens” such as PspA, PsaA, and pneumolysin) may be useful as multiserotype vaccines (4). As an alternative, we are investigating killed nonencapsulated cells given intranasally as an inexpensive vaccine (10). The premises are that cells may present various species antigens in native configuration and, with an appropriate adjuvant, elicit both systemic and mucosal immunity. Further commending our approach is that pneumococci initiate infection from nasopharyngeal (NP) colonization (2) and are less capsulated in the NP (7). Previously, we reported that ethanol-killed nonencapsulated pneumococci (KNP) administered intranasally with cholera toxin (CT) as the adjuvant were protective against NP colonization by (encapsulated) serotype 6B; KNP administered intranasally without adjuvant were nonprotective (10).

Particularly where both vaccination and infectious challenge are intranasal, one must distinguish protection by the vaccine from nonspecific effects of adjuvants. While CT was an effective adjuvant in our model, it partially protected against colonization. Furthermore, due to toxicity, CT is not an acceptable adjuvant for humans. Therefore, to identify more appropriate adjuvants with lower nonspecific protection, we tested four candidates: mycoplasma-associated lipoprotein (MALP-2), a mucosal or systemic adjuvant in mice (11) that interacts with Toll-like receptor 2 (13); porcine lung surfactant (PLS) not containing surfactant protein A (5, 9); Bordetella pertussis monophosphoryl lipid A (MPLA), resembling preparations studied in humans (14); and cholera toxin binding subunit (CTB), used mucosally in humans (8). Accordingly, KNP plus adjuvant were compared to adjuvant alone as well as to saline. Further goals were to study additional serotypes and to determine whether intranasal challenge produced colonization of the middle ear (ME) and whether vaccination protected the ME as well as the NP.

Strain RX1AL(−), capsule negative (12) and lytA negative (3), was grown to late-log phase in Todd-Hewitt broth with yeast extract and killed in 70% ethanol to prepare KNP as described previously (10). CT was from List Biologicals (Campbell, Calif.), B. pertussis MPLA was from the Department of Parasitology, University of São Paulo; PLS (9) and CTB (1) were from the Instituto Butantan, São Paulo; and MALP-2 was from Alexis (San Diego, Calif.). Adjuvants with or without KNP (at a dose of about 5 × 10⁸ CFU) in 10 μl of saline were applied intranasally to 4- to 6-week-old C57BL/6 mice atraumatically at 1-week intervals. The adjuvant dosage was 0.5 μg of MALP-2, 60 μg of PLS, 5 μg of MPLA, or 4 μg of CTB. Each group of adjuvant with or without KNP was comprised of 8 to13 mice. Three weeks after the last immunization, mice were challenged with an intranasal administration of 2 × 10⁶ CFU in 10 μl of S. pneumoniae strain 0603 (serotype 6B, as described in reference 10) or mixtures of 2 × 10⁶ CFU each of CT882328 (serotype 14, erythromycin resistant) and TN82328 (serotype 23F, trimethoprim-sulfamethoxazole resistant) (both strains were from the Centers for Disease Control and Prevention ABC collection).

Nasopharyngeal colonization was determined 1 week postchallenge by euthanizing the animal by CO₂ inhalation, washing the NP by using a tracheal approach, and collecting lavage fluid exiting from the nares. About 100 μl of fluid was recovered, and 50 μl was plated on blood agar, which for the mixed challenge contained either erythromycin (0.3 μg/ml) or trimethoprim-sulfamethoxazole (3.2/16 μg/ml) to distinguish serotypes. Qualitatively, NP colonization was defined as ≥1 CFU in 50 μl of nasal wash. To calculate geometric means, 0 CFU per 50 μl was considered 0.5 CFU. Colonization of the ME cavities was determined by plating all fluid recovered from instillation of 20 μl of saline each into the right and left ME followed by withdrawal; the cellular response was examined by Gram staining the fluid recovered from the ME. The frequencies of positive NP or ME cultures were compared by Fisher's exact test. Also, the means of log-transformed CFU per milliliter of NP fluid were compared by the Student t test. Bonferroni correction for multiple comparisons was used where applicable. A two-tailed P value of ≤0.05 was considered significant.

Time-course studies of ME colonization in mice following NP inoculation were done with serotype 6B, using four mice per time point (uninfected mice, and mice at days 1, 2, 4, and 7 postchallenge). The density of ME colonization gradually increased over the first 4 days following challenge, proportionally with NP colonization (data not shown). There was a moderate correlation between the presence of bacteria in the ME and the number of CFU in NP fluid (r = 0.64, P = 0.005). Inflammatory cells were not present in the ME of mice prior to challenge, but their numbers gradually increased over the first 4 days, thereby implying an inflammatory response to ME colonization. These results suggest that pneumococcal ME colonization in our model is likely the consequence of sustained bacterial growth in the ME and that the inflammatory cells observed are the result of this bacterial proliferation.

The candidate adjuvants were surveyed by challenge with serotype 6B after two intranasal immunizations (Table 1). MALP-2 with KNP did not lower the density of NP colonization or ME colonization compared to MALP-2 alone. Similarly, neither PLS nor MPLA, with KNP, was significantly protective against NP or ME colonization compared to the effect of the adjuvant alone. In contrast, KNP plus CTB markedly lowered NP CFU (P < 0.001 versus CTB alone) and the incidence of ME colonization compared to CTB alone. Importantly, CTB alone was not significantly protective compared to saline with respect to NP or ME colonization.

TABLE 1.

Effects of KNP administration with several candidate adjuvants upon colonization of the NP and ME after intranasal challenge with capsulated serotype 6B pneumococci

Vaccine or control groupa	Mice colonized in NP/total	Mean log CFU/ml ± SD in NP wash	P value of mean vs that for adjuvant aloneb	Mice colonized in ME/total	P value vs that for adjuvant alonec
Saline	8/8	4.43 ± 0.67	0.028	8/8	1.0
MALP-2	12/12	3.51 ± 0.79		12/12
MALP-2 + KNP	11/12	3.01 ± 1.07	0.70	11/12	1.0
Saline	12/13	3.72 ± 1.22		12/13
PLS	10/12	3.69 ± 1.29		5/12
PLS + KNP	10/12	2.92 ± 1.21	0.72	8/12	0.41
MPLA	12/12	3.28 ± 0.70		8/12
MPLA + KNP	10/12	2.58 ± 1.08	0.09	5/12	0.41
CTB	11/12	3.72 ± 1.11		8/12
CTB + KNP	6/12	1.64 ± 1.15	<0.001	3/12	0.10

^aIntranasal immunizations at week 0 and week 1, challenge at week 4, culture at week 5.

^bStudent's t test with Bonferroni correction where applicable.

^cFisher's exact test with Bonferroni correction where applicable.

CTB was further evaluated after three immunizations (Table 2). CTB alone did not significantly reduce NP or ME colonization by serotype 6B compared to saline. KNP plus CTB significantly reduced NP and ME colonization compared to CTB alone. In the combined serotype 23F plus 14 challenge, all saline control mice were colonized by 23F in the ME as well as the NP. CTB alone reduced NP CFU of type 23F; however, KNP plus CTB almost completely eliminated NP and ME colonization and further lowered the CFU (P = 0.034 versus CTB). Only 5 of 12 of the saline controls were colonized in the NP or ME by serotype 14; so the protective trends (both incidences, 1 of 12) are insignificant. When analyzed for either serotype (not tabulated), CTB alone did not reduce the colonization in the NP or ME while KNP plus CTB reduced the incidence of both NP colonization (P = 0.003 versus CTB) and ME colonization (P = 0.003 versus CTB).

TABLE 2.

Effects of CTB administration alone or with KNP on colonization of the NP and ME of mice after intranasal challenge with capsulated pneumococci of several serotypes

Vaccine or controla	Challenge serotype	Mice colonized in NP/total	Mean log CFU/ml ± SD in NP wash	P value of mean vs that for adjuvantc	Mice colonized in ME/total	P value vs that for adjuvantd
Saline	6B	11/11	3.85 ± 1.11	0.10	11/11	0.43
CTB	6B	8/12	2.49 ± 1.44		9/12
KNP + CTB	6B	2/12	1.15 ± 1.21	0.048	2/12	0.02
Saline	23Fb	12/12	4.50 ± 1.13	≤0.001	12/12	0.08
CTB	23F	8/12	2.27 ± 1.25		7/12
KNP + CTB	23F	1/12	1.20 ± 0.69	0.033	0/12	0.009
Saline	14b	5/12	1.90 ± 1.22	0.47	5/12	0.74
CTB	14	3/12	1.38 ± .84		2/12
KNP + CTB	14b	1/12	0.61 ± 0.89	1.0	1/12	1.0

^aIntranasal immunization performed at weeks 0, 1, and 2; challenge at week 5; NP and ME cultures at week 6.

^bThe serotype 23F and 14 challenges were combined but distinguished by selective culture.

^cStudent's t test with Bonferroni correction for multiple comparisons.

^dFisher's exact test with Bonferroni correction.

Therefore, in our model, MALP-2, PLS, and MPLA at the selected concentrations did not provide significant adjuvant activity; however, dose range studies would be desirable. Notably, and similar to the findings of Wu et al. (15), CTB gave only suggestive protective effects alone but rendered KNP protective. Thus, CTB will facilitate our study of additional serotypes, antigenic specificity, and optimization of dosage and schedule of vaccination and although less potent than CT (10), may be safer for human study. CTB has been given safely by the intranasal route to humans and shown to be effective as an adjuvant by this route (8).

Others have investigated intranasal vaccination and intranasal animal challenges to evaluate pneumococcal vaccination (4, 15, 6). Killed encapsulated pneumococci given intranasally give homotypic protection (6, 15); e.g., serotype 6B cells plus CTB prevented colonization with 6B (15). Our approach differs in seeking multiserotype protection dependent upon species antigens. KNP with CTB protected against both NP and ME colonization by two (and possibly three) serotypes prevalent in human infection; additional serotypes remain to be tested. While it remains to be seen whether the ME colonization in our murine model mimics human otitis media, its prevention by intranasal KNP further commends this approach.