Development and Evaluation of an Improved Mouse Model of Meningococcal Colonization

Studies of meningococcal pathogenesis have been severely restricteddue to the absence of an adequate animal model. Given the significanceof iron in meningococcal pathogenesis, we developed a modelof Neisseria meningitidis colonization in outbred adult micethat included daily administration of iron dextran. While receivingiron, the animals were inoculated intranasally with the initialdoses of bacterial suspension. Meningococci were recovered fromthe animals by nasopharyngeal washes. Approximately half ofthe animals inoculated with 10⁷ CFU remained colonized 13 daysafter the initial bacterial inoculation. The model was furtherevaluated with genetically defined isogenic serogroup B mutantstrains, and the colonization capabilities of the mutants werecompared to that of the wild-type parent. A mutant that producestruncated lipooligosaccharide (KDO₂-lipid A) and a mutant defectivein capsule transport were dramatically impaired in colonization.A mutant defective in pilus transport (pilQ) showed moderatelyimpaired colonization. The immunological aspect of the modelwas also evaluated by challenging mice after immunization withhomologous whole-cell meningococci. The immunized mice wereprotected from colonization of the homologous strain. In thismodel, long-term meningococcal colonization was maintained,allowing us to study the effects of specific genetic mutationon colonization. In addition, this model allows investigationof the role of active immune response against meningococci.

INTRODUCTION

Top

Introduction

Neisseria meningitidis is now the leading cause of serious bacterialmeningitis in the United States and is the leading cause inmany countries worldwide, despite effective antibiotics andpartially protective vaccines (3). Meningococci commonly colonizethe nasopharynx of healthy individuals and can be isolated from5 to 10% of the population in nonepidemic settings (34). Meningococcalacquisition in the nasopharynx may, in certain individuals,progress to an invasive disease resulting in fatal meningitisand/or sepsis.

Humans are the only natural host of meningococci. The bacterialvirulence and colonization strategies are highly adapted tohumans and have been difficult to study in animal models (41).Availability of iron is a major determinant of successful meningococcalinfection, because iron is sequestered by the host, and an extremelylow concentration of iron is maintained in the serum and atthe mucosal surface. The inability of meningococci to causedisease in mice and other animals has been associated with theneed of meningococci to use human transferrin or human lactoferrinas a source of iron (32). Meningococcal host specificity isdue in part to the specificity of the iron-acquisition receptorson the surface of the bacteria, because the transferrin andlactoferrin receptors are specific for the corresponding humansubstrates (32, 33). This specificity for iron acquisition hasrestricted the development of the adequate animal model forstudying the pathogenesis of meningococci. In this regard, enhancementof meningococcal virulence by the addition of iron in adultmice has been documented. Prior treatment of adult mice withhuman holo-transferrin causes bacteremia after intraperitoneal(i.p.) injection of the bacteria (14). Similarly, a parenteraldosage of iron dextran greatly enhances infection in animalschallenged with meningococci (14).

Meningococcal disease has been modeled with infant mice, sinceinfant rodents, for reasons not fully understood, are more susceptibleto meningococci (29-31). Intranasal challenge with meningococciin infant mice produces invasive infection, mimicking the courseof the meningococcal disease in humans (29). In the infant mousemodel, administration of the iron to the animals further enhancesthe infection by the meningococci (20, 29-31). However, a majordisadvantage of the infant mouse/rat model is the narrow windowof infection, because mice and rats over 10 days old do notshow signs of meningococcal disease (20, 31).

Recently, an animal model of Neisseria gonorrhoeae, a closerelative of meningococcus, was developed in female BALB/c miceas a surrogate model for human infection (15). Prolonged genitaltract infection in mice was achieved by treatment with estradioland antibiotics. In this study, we have developed a convenientand economical animal model using adult outbred mice. Utilizingthe importance of iron in the pathogenesis of meningococcus,we were able to establish colonization in the nasopharynx ofthe mouse by administering an iron source systemically and locally.Because meningococcal nasopharyngeal colonization is the firststep of host-parasite interaction in the natural meningococcalinfection, we first addressed these events by using this model.The model provides insight into understanding the early eventsof meningococcal disease.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains. The meningococcal strains used in this study are listed in Table1. A spontaneous streptomycin-resistant strain was selectedas a reference strain from a clinical isolate, with a serogroupB N. meningitidis strain, IR2781 (27). Meningococci were grownon GC medium base (Difco, Becton Dickinson, Sparks, Md.) plates(GCB plates) containing Kellog's supplements and incubated at37°C under 5% CO₂ tension or in liquid cultures in GC brothcontaining Kellog's supplements at 37°C with 5% CO₂ withagitation_. When necessary, appropriate antibiotics were addedto the medium as follows: streptomycin, 750 µg/ml; kanamycin,100 µg/ml; and spectinomycin, 70 µg/ml. Transformationwas performed as previously described (27).

View this table:
[in this window]
[in a new window]
TABLE 1. Characteristics of the strains used in this study

Mouse model. Female outbred mice [Swiss Webster, Crl:CFW(SW) BR, CharlesRiver Laboratories] were used for most of the experiments. Insome experiments, complement-deficient mice (CDM) (B10.D2-H2^dH2-T18^c Hc⁰/oSnJ; Jackson Laboratories) were used. These CDMlack the C5 component of the complement pathway. Mice received200 mg of iron dextran (Dexferrum; American Regent Laboratories)per kg of body weight i.p. in 100 µl 2 to 3 h before intranasalinoculation with meningococci. Designated numbers of CFU ofmeningococci in 20 µl of GC broth were atraumaticallyinoculated into the right nare of the animal. After 3 h of themeningococcal inoculation, 15 µl of iron dextran (50 µg/µl)was given intranasally. The i.p. iron dextran was administeredonce a day during the duration of the colonization experiment.In some instances, human holo-transferrin (Sigma) in 200 µlof phosphate-buffered saline (PBS) (pH 7.4) was injected intothe mice (800 mg per body weight) in the place of iron dextran.Human holo-transferrin was used for the subsequent intranasalinoculation (1.5 mg per mouse in a volume of 15 µl) andfor the daily dosage. Nasopharyngeal washes were carried outby nasally instilling 35 µl of GC broth. Liquid was recoveredfrom the mouth and plated on the modified Thayer-Martin mediumto monitor for meningococcus growth. The number (percentage)of mice colonized with meningococci over a period of 13 dayswas monitored.

ELISA. The enzyme-linked immunosorbent assay (ELISA) was performedas follows. Levels of immunoglobulin A (IgA) to the homologousstrain were assayed in the nasopharyngeal washes of the micecolonized with the wild-type strain for 13 days or longer. Thewells of microtiter plates (Maxisorp; Nunc) were coated overnightat 4°C with the whole-cell bacteria. Prior to coating, bacteriawere grown overnight on GCB plates and resuspended in coatingbuffer (0.1 M sodium carbonate, 0.1 M sodium bicarbonate, pH9.6) to an optical density at 600 nm (OD₆₀₀) of 0.2. The wellswere blocked with 3% bovine serum albumin (Sigma) in PBS for1 h. Nasopharyngeal washes were collected as described above,and the volumes were adjusted to 100 µl. The washes wereplaced in the coated microtitier wells and probed with horseradishperoxidase-labeled antimouse IgA (Sigma). Reactions were visualizedby applying the wells with tetramethylbenzidine with hydrogenperoxide (TMB Soluble; Calbiochem) and read at 630 nm aftera 30-min incubation.

Immunization. Mice were immunized with viable whole bacteria of N. meningitidisstrain IR2781 suspended in PBS. Bacteria were first grown onGCB plates overnight as described above. The bacterial cellswere resuspended in PBS (pH 7.4) to an OD₆₀₀ of 0.4. The micereceived 100 µl of resuspended bacterial preparation i.p.Additional i.p. immunizations were performed twice at 9-dayintervals.

Statistical analysis. All of the mouse colonization data were subjected to statisticalanalysis to determine significance. The Kaplan-Meier methodwas used to estimate cumulative clearance. Cumulative clearancebetween groups (strains, challenges, and inoculum sizes) wascompared by log-rank tests. Statistical tests were two sided.A P value of $<=$ 0.05 was considered statistically significant.The data from the ELISA reading were analyzed by using a one-sampletest (see Fig. 2).

View larger version (29K):
[in this window]
[in a new window]
FIG. 2. Levels of IgA in nasopharyngeal washes of adult mice (Swiss Webster) colonized with meningococci over 13 days. The levels of IgA to the homologous strain are presented as ELISA readings at OD₆₃₀ (x axis). The nasopharyngeal washes from colonized mice and a control mouse are indicated on the left (y axis). The ELISA readings from colonized mice (n = 6, mean = 0.065, standard deviation = 0.035, P = 0.009) were compared to the negative control value by using a one-sample test.

RESULTS

Top

Results

Model for meningococcal colonization of outbred mice. Outbred adult mice (Swiss Webster) were selected in an effortto closely reflect colonization of human populations. A dailydose of systemic iron was required to establish meningococcalcolonization of outbred adult mice. In addition, intranasallocal priming with iron dextran on day 0 was necessary to haveall of the mice colonized on day 1. In contrast to the workof Jerse with gonococcus (15), prior antibiotic treatment ofmice was not necessary for meningococcal colonization in ourexperiments (data not shown). The number (percentage) of micecolonized with meningococci over 13 days was plotted (Fig. 1).Three groups of mice intranasally received the serogroup B wild-typestrain (IR4127) at 10⁷, 10⁶, and 10⁵ CFU, respectively. Forty-twopercent of the mice inoculated with 10⁷ CFU remained colonizedwith meningococci after 13 days (Fig. 1). The cumulative clearancebetween the inoculum sizes was statistically different (P =0.001) by the log-rank test. Nasopharyngeal washes from micethat had been colonized for more than 13 days contained mouseIgA to the homologous meningococcus strain, indicating the presenceof active mucosal immune responses against colonizing meningococci(Fig. 2).

View larger version (12K):
[in this window]
[in a new window]
FIG. 1. Course of adult outbred mice (Swiss Webster) colonization with the wild-type meningococcus. Number of mice colonized in percent (y axis) was monitored at day 1 and every 2 days for 13 days (x axis). The reference strain, IR4127, was instilled at 10⁷(10-7; ${blacktriangleup}$ ), 10⁶ (10-6; ${triangleup}$ ), or 10⁵ (10-5; ${circ}$ ) CFU. The graph represents combined data from two experiments performed separately. The total number of mice used is indicated in the parenthesis on the right. The cumulative clearance was statistically different between the inoculum sizes (P = 0.001) by the log-rank test.

Colonization of adult mice with mutants. Adult mice were intranasally challenged with meningococcus mutantsto evaluate the roles of specific genes in meningococcal colonization.First, three mutants that did not express wild-type lipooligosaccharide(LOS) due to inactivation of the LOS biosynthesis genes (rfaC,lgtF, and lst) were studied. These mutants produce truncatedLOS molecules: the rfaC mutant synthesizes only KDO₂-lipid A(18), while the lgtF mutant expresses the Hep₂(GlcNAc)KDO₂-lipidA (17, 18). The LOS from the lst mutant is not sialylated dueto a defect in LOS sialyltransferase (17, 18). Figure 3 representsthe number of mice colonized with corresponding mutant strainsin three groups. The rfaC mutant (IR5389) was severely impairedin the colonization capability, while the lgtF (IR5475) andlst (IR5476) mutants showed moderately attenuated colonization.Likewise, the capsule-deficient ctrA strain (IR5390) showedseverely impaired colonization (Fig. 3). The ctrA mutant iscapsule deficient due to inability to transport the capsuleto the cell surface. As another control, a pilQ mutant thatdoes not allow pilus biogenesis was examined (8). The PilQ proteinfunctions in the terminal process of pilus expression, and inactivationof the gene affects the competence for the natural transformation(8). The pilQ mutant was moderately impaired in colonization(Fig. 3). The cumulative clearance was statistically differentbetween the wild-type and mutant strains (P < 0.001) by thelog-rank test.

View larger version (18K):
[in this window]
[in a new window]
FIG. 3. Colonization of adult mice with LOS-, capsule-, or pilus-defective mutants. The number (percentage) of mice colonized (y axis) was monitored for 13 days (x axis). The rfaC ( ${triangleup}$ ), lgtF (•), lst ( ${circ}$ ), ctrA ( ${square}$ ), and pilQ ( ${blacksquare}$ ) mutants were inoculated into the corresponding groups of mice. The data from the wild-type strain ( ${blacktriangleup}$ ) were from Fig. 1 and are included as a reference. Each inoculum was 10⁷ CFU. The graph represents combined data from two experiments performed separately. The total number of mice used is indicated in parentheses on the right. The cumulative clearance between the wild-type and mutant strains was statistically different (P < 0.001) by the log-rank test. The growth rates of the rfaC ( ${triangleup}$ ), ctrA ( ${square}$ ), pilQ ( ${blacksquare}$ ) mutants are presented, with that of wild type ( ${blacktriangleup}$ ) given in the insert. The log numbers of CFU per milliliter (y axis) are plotted over 7 h (x axis).

The in vitro growth rates of the mutants were compared withthat of wild-type parent to determine whether gene inactivationaffected the growth of the mutants, which could reduce the degreeof the nasopharyngeal colonization. The mutants grew at ratessimilar to that of the wild type, indicating that the gene inactivationof the mutants did not affect the in vitro growth of the bacteria(Fig. 3, insert).

Role of complement in the colonization of rfaC and ctrA mutant strains. Complement-mediated bactericidal activity is the major defensemechanism against meningococcal infection (7, 19, 25). Capsuleand LOS appeared to play critical roles in meningococcal colonizationin normal adult mice, because mutants devoid of capsule or expressinga truncated LOS (rfaC) failed to effectively colonize in thismodel. This led us to further investigate the role of complementin the colonization of the rfaC and ctrA mutants by using CDM.Complement was partially responsible for the failure of colonizationfor both mutants, since colonization by the mutants was partiallyrestored in the CDM (Fig. 4). The cumulative clearance betweenfour test groups was compared by the log-rank test (P < 0.001).

View larger version (12K):
[in this window]
[in a new window]
FIG. 4. Colonization of CDM with LOS-defective (rfaC) and capsule-defective (ctrA) strains. CDM were challenged with rfaC ( ${blacktriangleup}$ ) and ctrA (•) mutants. As references, normal mice were challenged with rfaC ( ${triangleup}$ ) and ctrA ( ${circ}$ ) mutants. The number (percentage) of mice colonized (y axis) was monitored every 2 days for 7 days (x axis). Each inoculum was 10⁷ CFU. The graph represents the combined data from two experiments performed separately. The total number of mice used is indicated in parentheses on the right. The cumulative clearance between four test groups was compared by the log-rank test (P < 0.001).

Effect of immunization on colonization of the mouse nasopharynx by meningococci. The effect of parenteral immunization on colonization of adultmice by nasopharyngeal administration of the homologous meningococcusstrain was evaluated. The animals were intranasally challengedafter a course of three immunizations with the whole bacterialcell (see Materials and Methods). None of the immunized micebecame colonized with the homologous strain, indicating thatthe parenteral immunization effectively protects mice from nasopharyngealcolonization by homologous meningococci (Fig. 5). ELISA detecteda high level of serum IgG to the homologous strain in sera fromall immunized mice, while IgA was not detected (data not shown).

View larger version (8K):
[in this window]
[in a new window]
FIG. 5. Effect of immunization on mouse colonization by MC. The number (percentage) of mice colonized (y axis) was monitored every 2 days for 5 days (x axis). The immunized mice ( ${triangleup}$ ) and the naive mice ( ${blacktriangleup}$ ) were challenged with the homologous wild-type strain. The total number of mice used is indicated in parentheses on the right. The cumulative clearance between immunized mice and naïve mice was statistically different (P < 0.001) by the log-rank test.

DISCUSSION

Top

Discussion

The in vivo study of meningococcal pathogenesis has been severelyrestricted by the pathogen's strict host specificity and themorbidity and mortality of human disease. An animal model ofmeningococcal colonization to study the initial course of thedisease and to facilitate the development of better vaccinesis critically needed. Studies with infant rodents had limitedsuccess in producing bacteremia by intranasal inoculation (20).The animals were 4 to $~$ 5 days old at the beginning of the experimentand maintained susceptibility for the next 5 days. This studydescribes the establishment of a model of long-term nasopharyngealmeningococcus colonization of adult mice by supplying the animals'daily iron. Both daily systemic iron dextran and intranasalpriming with iron dextran were required to initiate colonization.The importance of iron related to meningococcal virulence inmice has been studied extensively (13, 14). We tested both humanholo-transferrin and iron dextran to facilitate meningococcalcolonization. Both iron sources supported the bacterial colonizationin mice when given i.p. (human transferrin data not shown).However, iron dextran was chosen for the challenges because(i) iron dextran is relatively economical and showed very lowtoxicity (14), and (ii) the iron moiety was complexed with dextran,a low-molecular-weight polyglucose (4). Dextran is cleared rapidlyby either metabolism or excretion. Parenteral administrationof human transferrin may lead to immune responses, which shouldinfluence the model.

A previous study by Mackinnon et al. (21) showed that the nasalcolonization usually accompanied lung infection in the infantmice. In this study, animals colonized in their nasopharynxesdid not have any viable bacteria in the lung.

Curiously, reducing the amount of normal flora in the nasopharyngealcavity by antibiotic treatment did not affect the outcome ofthe colonization (data not shown). In the mouse model of gonococcalgenital tract infection, antibiotic treatment was required tosuppress inhibitory normal flora.

Binding of pilus to the host mucosal surface is an initial stepin colonization (26, 35, 36, 39). Primary adherence by pilito the epithelial cells is followed by intimate contacts withsecondary surface molecules, such as Opa/Opc proteins (6). Themouse colonization model with a pilQ mutant revealed that themutant colonized less efficiently than the parent strain. Bacterialsurface components other than pili contribute to bacterial adhesionto epithelial and endothelial cells (24, 28, 46). In the absenceof proper pilus assembly function, Opc and Opa may facilitateadhesion of meningococci to the eukaryotic cells, especiallyif capsule expression is turned off or downregulated (44, 45)

Previous studies with capsule-defective mutants in an infantrat model demonstrated the importance of capsule for bacteremia(47). We found that capsule is required for effective colonizationin our model and that complement is partially responsible forthe defect in the ability of the capsule-deficient mutant tocolonize.

The role of LOS in the pathogenesis of meningococci and gonococcihas been studied extensively. Gonococcal LOS plays a role inthe adherence of the organism to host cells in vitro (5, 23,43). In an infant rat model, the meningococcus mutant defectivein LOS sialylation did not cause bacteremia, even with a highdose of CFU (47). In our mouse model, an rfaC mutant with atruncated LOS structure (KDO₂-lipid A) did not colonize. Anearlier study with the same mutant showed its inability to causebacteremia (37). Interestingly, an lgtF mutant, which had threeadditional inner core glycosyl residues on the LOS moleculecompared with the rfaC mutant, and an lst mutant (sialylationdefective) showed only moderately attenuated colonization (Fig.3). The lgtF and lst mutants were similar in our assay despitethe differences in the length of the truncated LOS, suggestinga role for LOS core structure in colonization.

The incidence of meningococcal disease inversely correlatesto the level of bactericidal antibodies in human serum (9, 10).The role of complement-mediated bactericidal activity has beenevidenced by the susceptibility to meningococcus of individualsdeficient in the terminal component of complement. The roleof antibodies to meningococci in preventing mucosal colonizationhas been documented. The rate of carriage of serogroup C wasreduced in a group of military recruits vaccinated with thepolysaccharide capsule from serogroup C or a combination ofcapsule polysaccharides from serogroups A and C (11, 38). Thesestudies indirectly suggest a protective role of humoral antibodiesin mucosal colonization. In other vaccine trials, however, vaccinationwith serogroup A capsule did not grant a reduction in carriagerate (2, 12). More recently, a significant mucosal immune responsehas been observed in young adults after immunization with meningococcalA+C conjugate and polysaccharide vaccines (1). Vaccination withmeningococcal C conjugate polysaccharide reduced the rate ofcarriage of serogroup C meningococci by 66% among an adolescentpopulation (22). In this study, hyperimmune mice with wholemeningococci were protected from mucosal colonization by thehomologous strain. Our results suggest that parenteral immunizationcould potentially prevent colonization and thereby prevent transmission.

In summary, a model for meningococcal colonization was developedwith adult outbred mice. The colonization capabilities of severalmutants and the prevention of colonization by immunization weredemonstrated. While this model does not accurately mimic allaspects of interactions between the human nasopharyngeal tissueand meningococci, because the organism is highly adapted tohumans and its receptors (CD46, human transferrin, lactoferrin,etc.), the model will allow us to study early events of meningococcalpathogenesis.

ACKNOWLEDGMENTS

We thank Anthony Richardson for helpful suggestions.

This study was supported by an Emory University Research Committeeaward and Public Health Service grant AI472870-01A1.

FOOTNOTES

* Corresponding author. Mailing address: Department of Microbiology and Immunology, Emory University School of Medicine, 1510 Clifton Rd. NE, Atlanta, GA 30322. Phone: (404) 727-5968. Fax: (404) 727-3659. E-mail: kyi{at}emory.edu.

Editor: F. C. Fang

REFERENCES

Top