Vaccine-Induced Cellular Immune Responses Differ from Innate Responses in Susceptible and Resistant Strains of Mice Infected with Coccidioides posadasii

Abstract

Susceptibility to Coccidioides spp. varies widely in humans and other mammals and also among individuals within a species. Among strains of mice with various susceptibilities, immunohistopathology revealed that C57BL/6 mice were highly susceptible to the disease following intranasal infection, DBA/2n mice were intermediate, and Swiss-Webster mice were innately resistant. Resistant Swiss-Webster mice developed prominent perivascular/peribronchiolar lymphocytic cuffing and well-formed granulomas with few fungal elements and debris in the necrotic center, surrounded by a mantle of macrophages, lymphocytes, and fibrocytes. Susceptible C57BL/6 mice became moribund between 14 and 18 days postinfection, with overwhelming numbers of neutrophils and spherules and very few T cells, the drastic reduction of which was associated with failure and death, while intermediate DBA/2n mice controlled the fungal burden but demonstrated progressive lung inflammation with prominent suppuration, and they deteriorated clinically. Vaccinated C57BL/6 mice had an early and robust lymphocyte response, which included significantly higher Mac2⁺, CD3⁺, and CD4⁺ cell scores on day 18 than those of innately resistant SW mice and DBA/2n mice; they also had prominent perivascular/peribronchiolar lymphocytic infiltrates not present in their unvaccinated counterparts, and they appeared to be resolving lesions by day 56 compared to the other two strains, based on significantly lower disease scores and observably smaller and fewer lesions with few spherules and neutrophils.

Coccidioides spp. (C. immitis and C. posadasii) are primary pathogenic fungi that inhabit the soil of the semiarid desert regions of the southwestern United States and northern Mexico. These fungi, which appear capable of infecting nearly any mammalian species, are dimorphic, living as mycelia in their saprobic phase and as thick-walled, septating spherules in the animal host. People and animals vary in their resistance to disease after infection with Coccidioides spp. Among humans, African-Americans and Filipinos are more susceptible to severe disease (12). In animals, entire species of animals, such as domestic cattle (29), appear highly resistant to clinical disease, while the dog shares a rate and range of subclinical and clinical illness very similar to that of people (37) and some breeds of dogs appear more susceptible to severe disease (10, 30, 36). Despite these clinically observed differences, little is known about the histological correlates among susceptible and resistant host responses.

Development of vaccine candidates for coccidioidomycosis has been pursued for many years, starting with whole-killed-cell vaccines and progressing into the present with recombinant protein vaccines (1, 11, 23, 24, 26, 33, 34, 38, 42). Reasons for the ongoing pursuit of a vaccine for coccidioidomycosis include the following: (i) durable, probably lifelong immunity following recovery from infection, which is the case for the large majority of those infected, and (ii) the prolonged morbidity and sometimes fatal outcome in patients with severe pulmonary or disseminated disease (16).

Murine models assessing prolongation of survival following lethal challenges or quantitation of microbial burdens in target organs are the most common means of evaluating vaccine candidates among several infectious diseases, including coccidioidomycosis and tuberculosis, a bacterial infection with pathological lesions similar to those caused by Coccidioides spp. (3, 23, 26, 40). In recent years, the cellular and cytokine milieu of tuberculosis lesions in animal models has been described using immunohistochemistry and has been proposed to evaluate vaccine efficacy for this infection (3, 17, 20, 40). To date, only a single study performed on human coccidioidal granulomas has used immunohistochemistry to describe the lesion architecture (28).

Immunohistochemical characterization of lesions as done with the animal models of tuberculosis has not yet been done with the mouse model of coccidioidal infection. This study characterizes the in situ immune response in lethally infected susceptible mice, mice rendered resistant by a vaccine, and innately resistant mice, and it assesses immunohistochemical staining of infected murine lungs as a means of enhanced coccidioidal vaccine evaluation.

MATERIALS AND METHODS

Animals.

Female, 8-week-old C57BL/6 (B6), DBA/2n (D2), and Swiss-Webster (SW) mice were purchased from Harlan-Sprague-Dawley (Indianapolis, IN) and housed according to NIH guidelines in a biosafety level 3 animal laboratory. All procedures were approved by the Institutional Animal Care and Use Committee for the University of Arizona.

Fungal cultures and infections.

C. posadasii, strain Silveira, was grown on glucose-yeast extract (GYE) agar until arthroconidia were mature, at which time they were collected in an aqueous suspension. Arthroconidia (spores) were enumerated on a hemacytometer and viability ascertained by plating 10-fold dilutions on GYE agar and enumerating after 3 days' growth at 37°C. Infectious suspensions were adjusted to deliver a target dose of 50 viable spores to each mouse in a 30-μl volume. Mice were anesthetized with intraperitoneal ketamine (80 mg/kg of body weight) and xylazine (8 mg/kg) and infected via intranasal insufflation. The infectious dose was verified by plate counts as described above; the actual number of spores delivered is reported in Results.

Mouse studies.

For studies evaluating innate resistance of mice, four to eight mice were used per strain and time point. In the pilot studies performed to assess resistance of strains, the right lung was cultured quantitatively while the left lung was fixed for immunohistochemistry. In the corroborative study comparing all three strains together, the entire lungs were fixed for immunohistochemistry. Mice were sacrificed on days 3, 6, 12, and 18.

For the study comparing innate to vaccine-induced resistance, 16 mice per group were used per strain, vaccination state, and time point. For each of the three strains, 48 mice were vaccinated and 48 mice received nothing, except for B6 mice, for which there were only 32 unvaccinated mice. The mice that were vaccinated received 2 μg of recombinant Ag2/PRA_1-106-CSA (38), 10 μg of CpG oligodeoxynucleotide (27), and 25 μg of monophosphoryl lipid A adjuvant in 0.2 ml sterile saline for injection as described previously (38). MPL-A (Sigma) was emulsified with squalene and lecithin per the method of Baldridge and Crane (2). Mice were vaccinated twice subcutaneously 2 weeks apart. All mice were challenged 4 weeks after the booster vaccine and sacrificed on days 12, 18, and 56 postinfection. The entire lungs from eight mice in each group were quantitatively cultured to determine the fungal burden, and the entire lungs from the other 8 mice were fixed in zinc acetate for immunohistology.

For logistical reasons, the study comparing vaccinated and unvaccinated mice was performed with two subsets that included one-half of the mice from each experimental group started 2 weeks apart. No unvaccinated B6 mice were designated for 56-day sacrifice because they do not survive that long. Some D2 mice became moribund prior to scheduled sacrifice for their group and were traded with mice that appeared healthier.

Evaluation of mice and collection of tissues.

At sacrifice, each mouse was weighed with comparison to the starting weight, and a gross disease score based on the condition of the lungs was recorded (0, no gross evidence of infection; 1, 1 to 2 small granulomas; 2, 2 to 8 small granulomas or a few larger granulomas; 3, coalescing granulomas that constitute approximately 33 to 70% of the lung; 4, at least 70% of normal lung tissue replaced by lesions). Lungs and spleens were collected for processing. Mice that became moribund prior to scheduled termination were sacrificed, and tissues were handled according to their assigned group. For quantitative culture, lungs were homogenized and plated in 10-fold serial dilutions on GYE agar plates; colonies were enumerated after 3 days' incubation at 37°C as described previously (1). For histologic sections, lungs from the pilot studies were fixed with a buffered aldehyde solution (Prefer; Anatech Ltd., Battle Creek, MI), and remaining studies employed zinc acetate fixation for better preservation of antigenic epitopes for immunohistochemical staining (19).

Histological staining.

Five-micrometer sections of paraffin-embedded lung tissue were stained with a routine hematoxylin-and-eosin stain, and the extent and character of lesions were described based on these. A lung lesion score was assigned based upon the extent of lesions on a scale from 0 to 5, as follows: 0, no evidence of lesions or lymphoid aggregates; 1, <20% affected lung tissue; 2, 20 to 40% affected lung tissue; 3, 40 to 60% affected lung tissue; 4, 60 to 80% affected lung tissue; 5, >80% affected lung tissue.

Immunohistochemical staining.

Paraffin-embedded, 5-μm sections of lung and spleen were stained using an automated processor (DakoCytomation, Carpinteria, CA). The spleen acted as a positive control for Mac2, CD22, CD3, CD4, and CD8 stains on each slide. Polyclonal goat anti-Ag2/PRA Coccidioides-specific antibody (13) was a generous gift from John Galgiani and was used at a dilution of 1:750. Monoclonal antibodies against CD22 (rat immunoglobulin G1 [IgG1], 1:50; Southern Biotech), CD4 (rat IgG2a, 1:50; eBioscience), CD8 (rat IgG2a, 1:10; eBioscience), and Mac2 (rat IgG2a, 1:15,000; Cedarlane Laboratories) were incubated at listed dilutions, followed by biotin-labeled goat antirat secondary antibody at a dilution of 1:200 and horseradish peroxidase-labeled streptavidin (DakoCytomation). Positive reactions were detected with diaminobenzidine (DakoCytomation). Polyclonal anti-CD3 antibody was purchased from Cell Marque Corp, used according to the manufacturer's directions at a dilution of 1:750, and detected with antirabbit horseradish peroxidase-labeled polymer (Envision; DakoCytomation). Mac2 is a marker for cells of macrophage/monocyte lineage and was selected for its known reactivity with fixed murine tissue (25); it recognizes macrophages, monocytes, dendritic cells, and Langerhans cells and may cross-react with fibroblasts, epithelial cells, and cartilage. CD3 is a widely used pan-T-cell marker, and CD4 and CD8 are standard markers for T-cell helper and cytotoxic subsets, respectively. CD22 is a mature B-cell marker that is active in fixed tissues.

Immunohistochemically stained sections were evaluated by scoring the slide for density of the stained cells in lesional areas. Slides stained with anti-Ag2/PRA were scored from 0 to 5 (0, no spherules/endospores observed; 1, estimated <100 spherules/endospores per slide; 2, estimated >100 and <200 spherules/endospores per lesion; 3, estimated >200 large spherules per lesion and endospores [too numerous to count {TNTC}]; 4, large spherules [TNTC] in lesions but still discrete lesions; 5, large spherules [TNTC] extending to all edges of lesions and in new sites in lung) (see Fig. SA in the supplemental material). Slides stained for Mac2, CD3, CD22, CD4, and CD8 were scored from 0 to 4 (0, no lesional areas, no lymphoid aggregates, or no cells in lesions; 1, individual, scattered cells visible in one or more lesions, occasional discrete clusters of a few cells viewed at high magnification; 2 to 4, increasing gradient of cell density in lesions and/or lymphoid aggregates easily visualized at low magnification) (see Fig. SB in the supplemental material). All slides were scored by a single observer (L.F.S.), and a single section was evaluated for each antibody for each mouse.

RESULTS

Relative susceptibilities of mouse strains.

Figure 1 shows a comparison of median CFU in the right lungs of D2, B6, and SW mice in two pilot studies. Left lungs in these studies were used for histopathology. Though D2 mice appeared to be the most resistant of inbred strains previously tested using intraperitoneal infection (15), with an intranasal challenge of 50 spores they showed weight loss (average, 3.6 g; range, 2.0 to 5.0 g) and disease scores (average, 2.25; range, 2 to 3) similar to those of B6 mice (average weight loss, 3.4 g; range, 0.05 to 5.5 g; average disease score, 3.0; range, 2 to 4). Also, assessment at necropsy indicated extensive pulmonary involvement of D2 mice, similar to results for B6 mice, in spite of significantly lower numbers of CFU (P < 0.05).

FIG. 1.

Numbers of CFU in the right lung (Rt lung) of B6, D2, and SW mice after intranasal infection with Coccidioides posadasii. (A) B6 and D2 (n = 4) mice received 50 spores; numbers of CFU were significantly greater in B6 mice on days 12 and 18 (P < 0.05). (B) B6 and SW mice (n = 8) received 65 spores (target dose = 50 spores); numbers of CFU were significantly different on all days together (P < 0.05).

With a challenge dose of 65 spores (target dose = 50), the median number of CFU for SW mice was lower than that for B6 mice (P < 0.05) and most animals remained clinically healthy through day 18, while 18-day B6 mice were sacrificed as a group on day 14 postinfection due to moribundity of half the animals. The mean weight loss in SW mice was 0.6 g (median weight change, +1 g; range, +3 g to −6 g), while the mean weight loss in B6 mice was 3 g (median loss, 4.5 g; range, +3 g to −6 g).

In a corroborative study comparing all three strains, mice received 47 spores (target dose = 50) intranasally and the entire lungs and spleens were fixed for histopathological characterization. The pattern of the spherule scores (data not shown) corresponded to the quantitative cultures shown in Fig. 1 and closely resembled the data in shown Fig. 3A, below. B6 mice had significantly higher spherule scores than SW mice (P < 0.05) and D2 mice (P < 0.05) on days 12 and 18. The condition of the mice was similar to that in the initial studies. All but one of the SW mice remained clinically healthy and gained weight, while three of four B6 mice required euthanasia before day 18 and D2 mice lost weight but did not have to be euthanized before day 18. Because of the clinical and fungal burden results, as well as the histopathological features described below, we categorized SW mice as innately resistant, B6 mice as susceptible, and D2 mice as intermediately susceptible.

FIG. 3.

Mean spherule score from slides stained with the Coccidioides-specific anti-Ag2/PRA antibody (A) or gross disease score at necropsy (B) by day, vaccination status, and mouse strain. These parallel the quantitative results shown in Fig. 2 quite well with the exception of the notably high disease score in D2 mice compared to their fungal burdens on days 18 and 56. Examination of slides revealed that D2 mice have severe inflammation, though they appear to be controlling the fungal burden.

Lung fungal burdens and spherule scores in vaccinated and unvaccinated mice of three strains.

Postinfection plate cultures revealed that infecting inocula for the two subsets containing half of each experimental group for the vaccine study were 47 and 32 spores per mouse (target = 50 spores). There were no statistical differences in results between otherwise similar mice receiving one inoculum or the other, so results from the two inocula were combined for analysis.

Figure 2 shows the results of fungal culture by strain and vaccination status on days 12, 18, and 56 postinfection. There is no day-56 unvaccinated B6 group because they do not survive that long with a lethal infectious dose. Numbers of CFU were significantly reduced, by 2 to 4 logs, in vaccinated compared to unvaccinated B6 mice (P < 0.001), while vaccinated D2 and SW mice showed no statistical differences from their unvaccinated counterparts. D2 mice and unvaccinated B6 mice both had significantly higher fungal burdens than SW mice (approximately 2 logs and 3 logs, respectively; P ≤ 0.001), while the low fungal burden in vaccinated B6 mice was similar to that in SW mice. On day 56, mean numbers of CFU were actually lower in B6 mice than in SW mice, but this was not significant.

FIG. 2.

Lung fungal burdens in unvaccinated and vaccinated B6, D2, and SW mice (n = 8) at days 12, 18, and 56 postinfection. SW mice have significantly fewer fungal CFU than B6 mice (P < 0.001) and D2 mice (P = 0.001) at all time points. Vaccinated B6 mice have significantly fewer fungal CFU than unvaccinated B6 mice (P < 0.001), and there is a statistical difference between all vaccinated mice and unvaccinated mice, but strain-specific analysis reveals that all the significance comes from the B6 mice.

Among unvaccinated B6 mice scheduled for sacrifice on day 18, 10/16 were sacrificed on day 15 postinfection and the remaining 6/16 were sacrificed on day 16 due to moribundity; mean weight loss among these mice was 5.65 g (range, −1 to −9 g). All vaccinated B6 mice and all SW mice survived to 56 days in good clinical condition. Among D2 mice, 5/32 required euthanasia before day 18 and 7/32 died after day 18 and before day 56. The group mean weight change for the D2 mice that survived to 56 days was +0.5 g, with about half losing and half gaining weight.

Sections stained with a Coccidioides-specific antibody were scored for fungal elements in the lesions according to the scale described above, and the mean scores for groups are shown in Fig. 3A. As with CFU numbers, spherule scores for unvaccinated B6 mice were significantly higher than those for SW and vaccinated B6 mice (P < 0.05), while D2 mice were intermediate. The scored assessment of fungal burden from sections stained with the Ag2/PRA antibody corresponds in pattern and statistical significance between groups with the quantitative culture.

Disease scores.

The mean disease score, a gross postmortem visual assessment of the extent of lesions in the lungs, was generally higher than the spherule score for mouse groups in all studies, except for unvaccinated B6 mice, which had high fungal burdens and high disease scores (Fig. 3B). The disease score reflects the inflammatory response, which is why it is generally higher than spherule scores. Differences in disease score over time were not significant for SW mice, due primarily to the fact that they control the infection and modulate inflammation early, but D2 mice had significant increases in disease scores on days 18 and 56 compared to that on day 12 (P = 0.002 and 0.001, respectively) even though fungal burdens were not significantly increased based on both CFU and spherule scores. The vaccinated B6 group showed a decrease in the disease score between day 18 and day 56 (P = 0.033), suggesting possible resolving lesions.

Histological progression of infection in unvaccinated B6, D2, and SW mice.

No lesions were observed grossly or microscopically at day 3, but they were visible in all mice from day 6 onward. The key histopathological differences between the strains through day 18 are summarized in Table 1.

TABLE 1.

Key histopathological differences between strains of mice on days 6, 12, and 18 postinfection

Day postinfection	Histopathological finding for straina
	C57BL/6	DBA/2n	Swiss-Webster
6	Lesions of ≤1 mm	Lesions of ≤1 mm	Lesions of ≤1 mm
	Neutrophils	Neutrophils	Neutrophils
	PV/PB lymphocytic cuffing, mild	PV/PB lymphocytic cuffing, mild	PV/PB lymphocytic cuffing, mild/moderate
	CD3+ cells scattered in lesions and PV/PB cuffs	CD3+ cells less than in SW or B6 mice; PV/PB cuffs	CD3+ cells scattered in lesions and PV/PB cuffs
	Small spherules/endospores present	Small spherules/endospores present	Small spherules/endospores present
12	Lesions: many, large, often involve entire lung lobe	Lesions: many, moderate/large, some whole lung lobes involved	Lesions: few, 3-5-mm diameter, occasional involvement of entire lung lobe
	Neutrophils, macrophages: extensive	Neutrophils, macrophages: extensive	Neutrophils, macrophages, moderate no., limited extent
	PV/PB lymphocytic cuffing, mild	PV/PB lymphocytic cuffing, mild	PV/PB lymphocytic cuffing, extensive in lesions and in surrounding lung
	CD3+: reduced in all sites	CD3+: increased in PV/PB cuffs	CD3+: prominently increased
	Spherules: large with size variability, high no.	Spherules: large with size variability, moderate to high no.	Spherules, usually few (<2 dozen/lesion), small/moderate size
18	Lesions: involve whole lung lobes	Lesions: whole lobes or >5 mm in diameter	Lesions: few, cf. other strains, ≤5 mm, occasional whole lobe
	Neutrophils extensive, macrophages decreased; cf. day-12 result	Neutrophils, macrophages: extensive	Neutrophils, macrophages: moderate, organized granulomas
	PV/PB lymphocytic cuffing; mostly B cells, reduced to thin layer	PV/PB lymphocytic cuffing: same as day 12	PV/PB lymphocytic cuffing: extensive. as at day 12 or greater, both T and B cells present
	CD3+: infrequent in lesions and PV/PB cuffs	CD3+: mild increase, cf. day-12 result	CD3+: continued increase from day 12 in PV/PB cuffs, lesion margins
	Spherules: large no., various sizes, in sheets mixed with neutrophils	Spherules: various sizes, moderate to high no.	Spherules: few in most animals, <2 dozen per lesion

In all strains, mice sacrificed on day 6 demonstrated lung lesions of ≤1 mm in diameter on gross inspection. Microscopically, the lesions were microabscesses located near airways, and those in B6 and D2 mice were larger, more numerous, and more destructive than those in SW mice, with obliteration of alveolar septa. In contrast to previously reported findings (24), neutrophils were the most common cell in the microabscesses (Fig. 4A), though immunohistochemical staining also showed abundant macrophages and scattered small lymphocytes in the abscesses and lymphocytic cuffing of lesional and perilesional vessels and airways. D2 mice also exhibited perilesional edema and intralesional hemorrhage, which are evidence of more damaging inflammation than that of B6 or SW mice (Fig. 4B and C). Small spherules and endospores were readily identified in day-6 lesions from all strains, though they were not identified in lesions this early in the previous description of histopathological progression (24). It is likely the Coccidioides-specific stain made it easier to identify small spherules and endospores in these early lesions, though we also observed at least one mature spherule from all three strains at day 6 and also saw occasional spherules, neutrophils, and eosinophilic debris in airways adjacent to microabscesses.

FIG. 4.

Lesions at day 6 postinfection. (A) Neutrophils are the most abundant cells in the microabscess, with macrophages (arrowheads) and small lymphocytes seen occasionally. Magnification, ×400; hematoxylin-and-eosin stain. The microabscesses in SW and B6 mice (B) have a distinct border with the normal tissue, while lesions in D2 mice (C) exhibit hemorrhage and extensive edema. Magnification, ×40; hematoxylin-and-eosin stain.

After day 6, lesion structure varied increasingly by strain. From day 12 onward, B6 mice had large, multifocal pyogranulomas that encompassed 50 to 80% of a lung lobe. The areas of central necrosis included spherules up to 120 μm in diameter, degenerate and nondegenerate neutrophils, macrophages, pycnotic cells, and eosinophilic debris. Extensive coronas of large, densely packed spherules, including many rupturing forms, extended to the edges of the lesions adjacent to normal lung tissue. Neutrophils were the predominant inflammatory cells, followed by macrophages. Diminished lymphocytes, especially T cells, were limited to perivascular/peribronchiolar (PV/PB) infiltrates. The majority of B6 mice were euthanized due to moribundity between 14 and 16 days for the day-18 evaluations. PV/PB cuffing had become negligible at the end, accompanied by a near-disappearance of T lymphocytes, a reduction in densities of macrophages, increased neutrophils, and sheets of spherules in all stages of development (Fig. 5A and E). Multifocal microabscesses were seen outside of the main lesions, as were occasional mature spherules unassociated with inflammatory cells in the few “normal” areas of lung parenchyma still present. Spherules were seen in the spleens of all B6 mice at the day-12 and -18 time points, indicating early dissemination.

FIG. 5.

(A to D) Representative sections from day 18 postinfection. Unvaccinated B6 mice have a myriad of spherules and neutrophils with minimal perivascular cuffing (A); D2 mice have fewer obvious spherules than B6 mice, but the lesion is extensive, with hemorrhage and edema (B); both SW mice (C) and vaccinated B6 mice (D) have smaller lesions with distinct borders, few organisms, and significant PV/PB cuffing. (E to H) Day-18-postinfection sections stained with Coccidioides-specific antibody highlight the enormity of spherule burden in unvaccinated B6 mice (E), the extent of lesions despite a relatively controlled spherule burden in D2 mice (F), and the low fungal numbers with controlled lesion size in SW mice (G) and vaccinated B6 mice (H). (I to K) On day 56 postinfection, inflammation is extensive and disorganized in D2 mice (I), with multifocal neutrophilic islands surrounded by poorly organized macrophages (I, inset); SW mice (J) and vaccinated B6 mice (K) have only one or a few small, contained lesions per section and thick PV/PB infiltrates. SW mice also demonstrate lymphoid aggregates at the lesion margins. Magnification, ×40. Panels A to D and I to K show hematoxylin-and-eosin staining; panels E to H show anti-Ag2/PRA antibody; I inset shows Mac2 antibody.

D2 mice also exhibited large, multifocal pyogranulomas by day 12. Spherules up to 120 μm in diameter and endospores distributed in both necrotic centers and lesion margins were increased compared to results on day 6 but to a lesser extent than in B6 mice. Most D2 mice did not exhibit the continuous sheets of spherules seen in B6 mice. Inflammation consisted of mixed neutrophils and macrophages; lymphocytes were observed primarily in PV/PB cuffs and were only sparsely scattered in lesions on day 12, though they became increasingly prominent on days 18 and 56 in mice that did not succumb to the infection. In spite of fewer organisms being present in the D2 mouse than in the B6 mouse, inflammation was notably severe, encompassing a similar amount of lung lobe (Fig. 5B and F). Hemorrhage, edema, and suppuration remained prominent features at both days 12 and 18, with little evidence of controlling inflammation. Though spherule numbers had not undergone statistically significant increases from day 12 to day 18, scattered microabscesses disseminated within the lungs were seen in some mice on day 18. Spherules were observed in 25% of the spleens from D2 mice on days 12 and 18.

Among D2 mice that survived to 56 days, multiple lesions with distinct bands of fibrosis surrounding necrotic centers with spherules, debris, and neutrophils were present. Surrounding these granulomatous cores were extensive zones of replacement that consisted of hyperplastic type II pneumocytes, fibroplasia, macrophages, islands of neutrophils and spherules scattered throughout, and PV/PB lymphocytic cuffs surrounding airways and vessels (Fig. 5I). Overall, approximately 40 to 70% of the lung appeared abnormal.

In contrast to both B6 and D2 mouse lungs, day-12 and -18 lesions in SW mice were generally smaller and there were fewer of them in the lungs. Macrophages, fibrocytes, and scattered lymphocytes were organized surrounding areas of central necrosis that contained eosinophilic debris, degenerate and nondegenerate neutrophils, macrophages, and spherules, as is typical of granulomas (Fig. 5C and G). Neutrophils were visualized primarily near spherules and endospores in the necrotic center and were a less significant component of the fibrogranulomatous mantle. Spherules were smaller than in the more susceptible strains, with fewer endospores and rupturing forms. In addition to lymphocytes in the mantle region, SW mice also formed lymphoid aggregates of B and T lymphocytes at the lesion margin, which were not observed in the other two strains, and they exhibited robustly expanded PV/PB lymphocytic cuffs both within and peripheral to the boundaries of lesions that increased in thickness visually between days 12 and 18 (Fig. 5C), while the granulomas seemed to become more compact, with smaller but denser accumulations of macrophages. Day-56 lesions of SW mice looked relatively unchanged from those on day 18, and there was not an obvious resolution of the lesions. They still contained spherules of various sizes and neutrophils in necrotic centers, fibrogranulomatous lesion margins with lymphocytes, lymphoid aggregates near the periphery, and robust PV/PB lymphocytic infiltrates (Fig. 5J). This remained the only one of the three strains that developed perilesional lymphoid aggregates resembling those observed by Li et. al. in coccidioidal granulomas from human lungs (28).

Histopathology in vaccinated mice.

Vaccination with rAg2/PRA_1-106-CSA did not change the progression of lesions, as described above, in D2 or SW mice but dramatically altered it in B6 mice (Fig. 5D, H, and K). Whereas unvaccinated B6 mice had innumerable spherules and neutrophils, with few lymphocytes seen terminally, vaccinated B6 mice had a robust lymphocyte response at all time points evaluated (days 12, 18, and 56). The lesions were discrete and smaller and did not generally encompass entire lung lobes as they did in the unvaccinated mice, and PV/PB lymphocytic infiltrates increased in prominence at each time point, similar to results for the innately resistant SW mice. By day 56, vaccinated B6 mice had only one or two lesions in the lungs; lesions were small, with extensive fibrosis, minimal neutrophils, and few spherules or endospores seen in the necrotic centers. PV/PB infiltrates appeared at least as extensive as those in the innately resistant SW mice. Lesions in these mice looked as though they were resolving because of the small size, few organisms, few neutrophils, and predominant lymphocytes and fibroplasia.

Organization of inflammatory cell subsets (Mac2, CD3, and CD22) and relative appearance in coccidioidal lesions.

The microabscesses on day 6 in all three strains consisted most abundantly of Mac2⁺ cells, with CD22⁺ and CD3⁺ cells present in both the abscesses and the PV/PB cuffs. As described for the histopathologic progression, neutrophils comprised the vast majority of the inflammatory cells by day 18 in unvaccinated B6 mice, and the cell marker stains showed decreased CD22⁺ cells, a near-disappearance of CD3⁺ T cells with only a few remaining in the diminished PV/PB infiltrates, and a decrease in Mac2⁺ cells throughout the lesions (Fig. 6A and E). The vaccinated B6 mice, by contrast, showed robust populations of both CD3⁺ and Mac2+ cells on days 12 and 18 (Fig. 6D and H). CD22⁺ cells were also present, primarily in the PV/PB lymphocytic cuffs, and increased over time as a result of the expanded cuffing. Interestingly, at day 56, Mac2⁺ cells were decreased and also were now organized along fibrogranulomatous margins around the small lesions.

FIG. 6.

Representative images of sections stained for Mac2, CD3, and CD4 on day 18 postinfection. (A to D) B6 mice (A) have undergone regression of macrophages compared to results on day 6 (A, inset), while D2 mice (B) have expanding lesions with many macrophages. SW mice (C) and B6 mice (D) are similar in having relatively small lesions with a band of macrophages surrounding a necrotic center, though the scores for Mac2 on day 18 were statistically higher for B6 mice than for SW mice (Fig. 6). (E to H) Unvaccinated B6 mice are shown at magnification ×200 to highlight the few CD3 cells present by day 18. CD3⁺ cells in D2 mice (F) are most prominent in the PV/PB infiltrates and scattered in the lesion margins, while SW mice (G) and vaccinated B6 mice (H) have abundant CD3⁺ in the lesions. (I to L) CD4⁺ cells parallel the CD3⁺ cells in location, though fewer in number. Note the composition of the marginal lymphoid aggregate from an SW mouse (K inset) and the abundance of them in the vaccinated B6 mouse (L) compared to the SW mouse (P = 0.003). Magnification for panels A to D, ×40; magnification for panel E, ×200; magnification for panels F to L, ×100.

In D2 mice, CD3⁺ cells were relatively sparse in the lesions and abundant in the lymphocytic cuffs (Fig. 6F). We observed that in lesions which contained many spherules, CD3⁺ cells were not present, but when the spherules were very few in number, lesions contained CD3⁺ cells. CD22⁺ cells were abundant in the PV/PB infiltrates, as with the other two strains. Mac2⁺ cells were scattered across the inflamed lung, though on days 18 and 56, some mice showed evidence of attempts to encircle necrotic centers with a band of macrophages (Fig. 6B). Mac2⁺ cells did not appear to be decreased at day 56, and the overall impression that the lesions of D2 mice were disorganized and extensive was emphasized by the widespread density of Mac2⁺ cells at day 56 (Fig. 5I, inset).

The Mac2⁺ cells in SW mice were concentrated in the fibrogranulomatous margins of the lesions at all time points after day 6, though they were also abundant on days 12 and 18 in the necrotic centers (Fig. 6C). CD22⁺ and CD3⁺ cells were observed in the PV/PB infiltrates and increased as these expanded, similar to results for the vaccinated B6 mice. In addition, both kinds of lymphocytes were present in lesion margins and in perilesional lymphoid aggregates (Fig. 6G). Though the cells were not enumerated in lymphoid aggregates as Li et al. described for the human granulomas (28), there appeared to be approximately equal numbers of T and B cells in the aggregates.

Semiquantitative immunohistochemical characterization of vaccinated and unvaccinated B6 mice compared to SW and D2 mice.

Sections from the vaccine study were scored for density of CD3⁺, CD22⁺, and Mac2⁺ cells in the lesions and PV/PB infiltrates. The mean scores for vaccinated and unvaccinated B6 and SW mice are shown in Fig. 7. By day 56, statistical differences between vaccinated B6 and SW mice were no longer present. On days 12 and 18, CD22⁺, CD3⁺ (day 18; P < 0.001), and Mac2⁺ (day 18; P = 0.02) cells were all present in greater densities in the lesions of vaccinated B6 mice than in those of SW mice. Unvaccinated B6 mice had declining CD3⁺-cell scores, which were significantly lower than those for SW mice on both days 12 and 18 (P = 0.05 and P = 0.006, respectively). It is clear from these data that vaccination of B6 mice induced a more robust early lymphocyte and macrophage response than was present in the innately resistant SW mice.

FIG. 7.

Mean scores for CD3, CD22, and Mac2 in vaccinated and unvaccinated B6 and SW mice on days 12 and 18 (n = 8 for each bar). CD3⁺, CD22⁺, and Mac2+ cells are increased in vaccinated B6 mice compared to those in SW mice, with CD3⁺ and Mac2⁺ cells significantly increased on day 18 (P < 0.001 and P = 0.02, respectively). In unvaccinated B6 mice, both lymphocyte populations are smaller than those in SW mice and CD3⁺ cells are highly significantly reduced on day 18 (P = 0.006).

Unlike the response in B6 mice, vaccination did not appear to change the spherule and immune cell scores. Table 2 depicts the relationship between D2 mice and vaccinated and unvaccinated B6 mice on day 18; spherule scores were lower and CD3⁺, CD22⁺, and Mac2⁺ cell scores were significantly higher in D2 mice than unvaccinated B6 mice, while vaccinated B6 mice had comparatively more CD22⁺ and CD3⁺ cells (for CD3⁺ cells, P = 0.002). While D2 mice are innately more resistant to coccidioidal infection than B6 mice (22), vaccination of the B6 mice resulted in a significant upregulation of CD3⁺ cells and better clinical and histological control of the disease than was found for D2 mice. Significant slide score differences between the strains were not present by day 56 for any of the parameters assessed.

TABLE 2.

Score relationships for spherules, CD22, CD3, and Mac2 for B6 mice compared to those for D2 mice on day 18 postinfection^a

Finding determined by immunohistochemical staining	Score relationship for mouse group
	Unvaccinated	Vaccinated
Spherule score	B6 > D2	B6 < D2
CD22 score	B6 < D2	B6 > D2
CD3 score	B6 < D2	B6 > D2
Mac2 score	B6 < D2	B6 > D2

^aRelationships shown in boldface are statistically significant comparisons with P values of <0.05.

Vaccinated B6 mice retained a pattern of lower fungal burdens than those for D2 and SW mice at day 56 and had significantly lower disease scores than D2 mice (P = 0.002). Mac2, CD22, and CD3 scores were no longer significant among the strains, even though overall histopathological differences were observed, especially regarding lesion size.

CD4⁺ cells are upregulated in vaccinated B6 mice.

The scores for CD4⁺ cells were lower than those for CD3⁺ cells but were parallel in pattern. In all three strains, they were relatively low on day 12, spiked on day 18, and became generally lower but not statistically different than day-12 scores by day 56, except for unvaccinated B6 mice, which had lower scores than the other groups on day 12 and had very few (scores ≪ 1) terminally (Fig. 6I). Scoring of all slides (eight mice per group) and enumeration of three representative magnification-×400 fields per slide on a randomized, blinded subset of the mice (four per group) yielded the same statistical results. Vaccination significantly upregulated CD4⁺ cells in B6 mice compared to results for unvaccinated B6 mice (P < 0.001), D2 mice (P < 0.05), or SW mice (P < 0.05) on day 18. Significant differences were observed only on day 18 using these methods.

CD4 scores were highest overall on day 18 for the vaccinated B6 mice (mean score, 2.5). Histologically, they were seen in fibrogranulomatous lesion margins and PV/PB infiltrates (Fig. 6L). CD4⁺ cells were concentrated in PV/PB cuffs of D2 mice (mean score, 1.9) and were much less common in the fibroganulomatous regions (Fig. 6J), while SW mice (mean score of 1.6 from nine animals that had lesions demonstrable on slides) exhibited CD4⁺ cells in the mantle regions of the granulomas (Fig. 6K), perilesional lymphoid aggregates (Fig. 6K inset), and the PV/PB cuffs.

CD8⁺ cells were visualized at only low levels in the lesions of B6 and SW mice at all time points and were almost nonexistent on the sections from D2 mice, regardless of vaccination status or time. Scores ranged from 0 to 1, with observations for many mice recorded as <1, “rare” or “occasional.” For the purposes of statistical analysis, the mice with results recorded as “rare” were assigned a numeric score of 0.25 and the mice with results recorded as “occasional” were assigned 0.5. None of the comparisons between CD8⁺ scores in the groups of mice were significant, though this may be due to the overall paucity of cells observed.

DISCUSSION

These studies demonstrate that immunohistopathological evaluation of lungs from mice with coccidioidomycosis adds an important visual dimension to the assessment of mice with successful and unsuccessful immune responses. In addition to describing significant differences in three strains of mice ranging from innately susceptible to resistant after intranasal infection, we showed that vaccine-induced resistance appears different at the cellular level than innate resistance. This model enhances the assessment of vaccine candidates in conjunction with quantitative culture or survival, the traditional methods of determining vaccine efficacy for coccidioidal infection.

Both innate and vaccine-induced responses were assessed in situ for three strains of mice. Innately, B6 mice are highly susceptible, D2 mice are intermediate, and SW mice are resistant to moderate intranasal infections of Coccidioides spp. in the range of 50 spores, thus corroborating previous reports (14, 21, 22). Across this spectrum of susceptibility, we found distinctive immunohistochemical patterns of host responses. B6 mice were characterized by extreme neutrophilic infiltrates encompassing whole lung lobes and the near-disappearance of lymphocytes from lung lesions by the onset of moribundity (14 to 18 days postinfection), while the resistant SW mice developed classic granulomas walling off relatively low numbers of spherules and, most notably, a robust PV/PB lymphocytic cuffing that expanded both within and around the granulomas at each time point evaluated. The necrotizing granuloma, as seen in SW mice, is classically associated with control of infection in humans (12). Also, the PV/PB lymphocytic cuffs appear to be associated with the successful immune response to coccidioidal infection in SW mice and vaccinated B6 mice, while the disappearance of T cells in unvaccinated B6 mice seems to be associated with failure. In research by others, T cells are highly associated with a successful response to coccidioidal infection and adoptive immunity can be transferred by immune splenic T cells but not by B cells or immune serum (4, 5, 7, 15).

As disease progressed rapidly toward death in the B6 mice, intralesional T-cell populations virtually disappeared, B cells decreased moderately, and the PV/PB cuffing diminished to relatively thin layers of lymphocytes that consisted mostly of B cells. A study of Mycobacterium tuberculosis in the guinea pig showed a similar pattern of T-lymphocyte reduction with worsening infection (40), but this occurred much less acutely in the tuberculosis model and was accompanied by an increase in B cells that we did not see in B6 mice infected with Coccidioides spp. The reasons for the decreased T cells were not established for the M. tuberculosis-infected guinea pigs. Possible reasons with Coccidioides spp.-infected mice are that the organism itself may be immunosuppressive (8) and produces signals that prevent continued lymphocyte migration or that the high numbers of neutrophils release cytokines that inhibit the lymphocytes. A problem with the latter theory is that the intermediate D2 mice, which also have a large, persistent neutrophilic component to the inflammation through day 18, do develop increased populations of lymphocytes and a modest expansion of the PV/PB infiltrates from day 6 to day 18. A more likely reason is that the initial innate response in the B6 mouse, directing the response toward Th2, is the primary key to failure and death; Fierer et al. reported 3-logs-higher interleukin 10 production from the lungs of B6 mice than from those of D2 mice (14). However, we also observed with D2 mice that T cells were present in lesions with few organisms but were highly diminished or absent in lesions with numerous spherules. It is likely that both host and pathogen factors play roles in the severe, rapid progression in B6 mice, and it would be interesting to pursue studies to elucidate possible immunosuppressive mechanisms of Coccidioides spp.

One of the goals of this study was to compare the in situ lung immune response in mice rendered resistant by vaccination to that of innately resistant mice to determine if vaccination produced a response that looked like the innate response. Vaccine-induced responses in the B6 mice, however, were characterized by more robust early infiltrations of macrophages and both CD3⁺ and CD4⁺ cells, which were statistically significant on day 18. Also, though vaccinated B6 mice had larger lesions than SW mice on days 12 and 18 based on gross disease scores, the reduced disease and lesion scores in B6 mice on day 56 compared to those on day 18 in conjunction with histopathological findings of small, fibrogranulomatous lesions with infrequent fungal forms and dense PV/PB infiltrates give the impression of resolving inflammation. In contrast, SW mice did not have a decrease in fungal burden or disease score on day 56 compared to results on day 18, but by day 18 these innately resistant mice had formed small, controlled classical granulomas with a necrotic center and a few spherules, surrounded by a mantle of macrophages, fibrocytes, lymphocytes, and lymphoid aggregates. Both responses appear to be associated with controlled infection, but we conclude that the vaccine-induced response differs from the successful innate response in the numbers of cells recruited early and the histopathology of the lesions. Perilesional lymphoid aggregates were not seen in the B6 mice, though they occur in both SW mice and human coccidioidal granulomas (20). Whether this is an inherent difference in the inflammatory responses between the two strains of mice or a result of the vaccine response cannot be distinguished from these data.

Despite differences between the vaccinated B6 mice and innately resistant SW mice, they shared thick PV/PB lymphocyte cuffs that appeared to be a characteristic of an effective immune response. These did not form in the unvaccinated B6 mice and were slower and less robust in the intermediate D2 mice.

Intermediately susceptible D2 mice, which are typically considered resistant to coccidioidal infection (9, 14, 22, 31), appeared to be dying from inflammation based on both the clinical and histopathological conditions of mice in these studies. Clinically, many of them had weight loss and dehydration, even if they survived to day 56, and approximately 20% of the mice died before the scheduled necropsy in the vaccine study. Histopathologically, though the lung fungal burden decreased slightly from day 18 to day 56, the mice suffered an expanding pyogranulomatous inflammatory response not commensurate with the relatively controlled fungal reproduction. The genetic lack of complement component C5a in D2 mice may lead to the dysregulated inflammation seen here and make this mouse strain incapable of producing the proper cellular milieu to overcome lung coccidioidal infection in spite of the fact that they start out with strong, specific Th1-directed immunity (6, 9, 14, 18, 22, 31, 35). The vaccine used in these studies did not change the lung fungal burden and inflammatory responses but appeared to suppress dissemination, based on a lack of spherules in spleens of vaccinated D2 mice, adding further credence to the idea that D2 mice cannot control lung inflammation but do mount a Th1 response to coccidioidal infection, as others have shown. Hence, immunohistopathological evaluation of D2 mice suggests that this strain may be an inappropriate model for studying respiratory resistance to coccidioidal infection, and this would not have been determined using the traditional methods of survival or quantitation of the fungal burden.

We observed that CD4⁺ and in particular CD8⁺ cells appeared to be present at less density than we expected based on the CD3 scores from these mice. Currently we have no biological explanation for the virtual absence of CD8-staining cells in these sections, but difficulties with anti-CD8 antibodies in fixed murine tissues are documented. Clones against mouse CD8 which stain paraffin sections are few, and the stains are uneven and unpredictable even on sections from the same piece of tissue (41); we suspect that this is the reason for the paucity of CD8⁺ cells observed in these studies. Though frozen sections are recommended for immunohistochemistry of cell markers, the necessary equipment was not available at biosafety level III for these studies. In spite of this, the CD4⁺ cells paralleled the CD3⁺ quite well and showed the same statistical relationship. Though architectural information that we can glean from the tissue sections would be lost, quantitation of CD4⁺ and CD8⁺ cells would likely be better accomplished by flow cytometry, for which many of these antibodies are developed.

The third goal of these studies was to evaluate immunohistochemistry as an additional tool for the preclinical evaluation of vaccine candidates. Scoring of spherules on immunohistochemically stained slides showed concordance with CFU numbers in terms of assessing the relative fungal burden and statistical significance of one group compared to another. This provides a semiquantitative measure that can separate groups and still allow evaluation of lesion architecture without having to sacrifice tissue to quantitative culture.

The immunohistopathology of tuberculosis progression in rodents is well described and has been used to enhance tuberculosis vaccine assessment (3, 32, 40). Reduction of the mycobacterial burden in mice with tuberculosis was not necessarily predictive of the severity of lung disease (3), which we especially noted in the D2 mice in our studies, and immunohistopathology provides a more complete model of the animals' response. Also similar to observations made for vaccinated, M. tuberculosis-infected animals (3, 39), organized lesions with distinct borders in vaccinated B6 mice and innately resistant SW mice infected with Coccidioides spp. were associated with better outcomes than the progressive lesions seen in unvaccinated B6 mice. Mice with successful responses to Coccidioides species infection had T cells in the lesion margins and thick PV/PB lymphocytic cuffing, details that cannot be observed by using quantitative methods that destroy the tissue. Furthermore, the differences that were associated with a successful immune response—increased T cells and macrophages—were apparent by day 12 and statistically significant by day 18, allowing for an early assessment of vaccine candidates that does not require the 8 weeks necessary for survival. We conclude that the immunohistopathogical model demonstrated herein can provide enhanced evaluation of coccidioidal vaccine candidates, alone or in conjunction with traditional quantitative fungal cultures, survival, or flow cytometry.

Drawbacks to immunohistopathology overall are that it is labor intensive in both preparation for sectioning and assessment of individual mice and that there are relatively few complete sections from a set of mouse lungs, limiting the number of slides that can be evaluated for a single animal. As sections are cut through the lungs, especially since the tissue diminishes in size, a section may not be representative of previous sections from the same animal, as seen in the CD4 staining of the SW mice in these studies, in which only 9 of 16 mice had lesions to evaluate for CD4⁺ cells, though we know that all the mice in this study were infected since lesions were detected in other sections.

Also, even though we demonstrated statistical differences between groups, scoring of slides is a less sensitive measure of cell numbers than quantitative procedures, such as flow cytometry or counting of cells on the immunohistochemically stained sections. Flow cytometry has the additional advantage that multiple markers can be evaluated from the same lung using a sample more representative of the total and that antibodies are applied to unfixed cell populations, enormously increasing the range and sensitivity of markers that can be evaluated. As noted above, the CD8 antibody is difficult to use in fixed murine tissues, and flow cytometry would be a better way to compare relative populations of CD4 and CD8 cells. With flow cytometry, lesion architecture is sacrificed, but a combination of immunohistochemistry and flow cytometry could complement each other in future studies.

We have shown differences in the histopathological progression of coccidioidal lung infection in three strains of mice which vary in susceptibility and also that a vaccine-induced response does not appear to be the same as the immune response in innately resistant mice. This model, which includes immunohistochemical evaluation of the lungs, enhances early assessment of vaccine candidates since differences can be seen by day 18, and we plan to utilize this to test new vaccine candidates. Though this model does not replace the need for clinical trials with people, results may focus preclinical explorations to include only antigens which stimulate the most robust T-cell responses.