Cellular Immune Suppressor Activity Resides in Lymphocyte Cell Clusters Adjacent to Granulomata in Human Coccidioidomycosis

Abstract

The in situ immunologic response in human coccidioidomycosis remains undefined. To explore this further, pulmonary necrotizing coccidioidal granulomata were examined using immunohistochemical staining for lymphocyte subsets and for the cytokines interleukin-10 (IL-10) and gamma interferon (IFN-γ). Discrete perigranulomatous lymphocytic clusters were seen in eight of nine tissues examined. In these tissues, T lymphocytes (CD3⁺) significantly outnumbered B lymphocytes (CD20⁺) in the mantle area of the granulomata (P = 0.028), whereas the clusters were composed of roughly equal numbers of T and B lymphocytes. While the number of cells in the mantle expressing IL-10 was similar to those in the perigranulomatous clusters, there were significantly more cells expressing IFN-γ in the mantle than in the clusters (P = 0.037). Confocal microscopy revealed that CD4⁺ T lymphocytes and B lymphocytes are associated with IL-10 production. CD4⁺CD25⁺ T lymphocytes were identified in the perigranulomatous clusters but were not associated with IL-10 production. This is the first report noting perigranulomatous lymphocyte clusters and IL-10 in association with human coccidioidal granulomata and suggests that down-regulation of the cellular immune response is occurring within coccidioidal granulomata.

Coccidioidomycosis is a fungal infection of increasing importance in the southwest United States (4). In the vast majority of cases, infection is acquired by inhalation of the fungus. Two-thirds of individuals who become infected are asymptomatic, their only evidence of infection being the expression of delayed-type hypersensitivity in response to a coccidioidal skin test. A variety of studies have indicated that protective immunity in human coccidioidomycosis is associated with an appropriate cellular immune response. Smith and colleagues demonstrated that persistent expression of delayed-type hypersensitivity presaged a good outcome, while failure to develop this response was associated with disseminated disease (20). More recently, in vitro studies have shown that peripheral blood T lymphocytes from individuals with well-controlled infection produce gamma interferon (IFN-γ) in response to coccidioidal antigen, while cells from individuals with disseminated disease do not (1). Failure of T lymphocytes to react to coccidioidal antigen in vitro is associated with a poorer clinical score among those with disseminated disease (2).

The hallmark of a robust T-lymphocyte response is the granuloma (10), and the formation of necrotizing granulomata as a response to pulmonary coccidioidomycosis has long been associated with a strong cellular immune response and control of coccidioidal disease (6). However, the precise immunological events associated with this response have been largely unexplored. Modlin and colleagues, in work published in 1985 (13), described a distinct pattern of cellular response in a single human pulmonary granuloma. In this pattern, CD4⁺ lymphocytes were dispersed throughout the granuloma, while CD8⁺ lymphocytes were localized to the mantle region. This pattern was similar to that seen in tuberculoid leprosy, pulmonary tuberculosis, and sarcoidosis. In contrast, the lymphocyte pattern in biopsies from skin lesions from individuals with disseminated coccidioidomycosis revealed CD4⁺ and CD8⁺ lymphocytes distributed on the periphery of the granulomata.

Since this report, techniques have emerged that allow for further assessment of the immunological response within coccidioidal granulomata. In particular, analysis of cytokine expression in addition to lymphocyte subset categorization has become available. In the present study, we investigated the lymphocyte distribution and cytokine expression of human pulmonary necrotizing granulomata due to coccidioidomycosis.

(This work was presented in part at the 102nd General Meeting of the American Society for Microbiology, Salt Lake City, Utah, 19 to 23 May 2002.)

MATERIALS AND METHODS

Human subjects and specimens.

This study was approved by the Human Subjects Protection Committee of the University of Arizona. Formalin-fixed paraffin-embedded lung tissues obtained by open lung biopsy and demonstrating coccidioidal granulomata were studied. The identities and clinical profiles of the patients from whom tissue was obtained were unknown.

Tissue sections were cut to 4- to 5 μm and mounted on poly-l-lysine-coated glass slides for lymphocyte surface marker staining or on positively charged glass slides for cytokine staining. The slides were deparaffinized and rehydrated in xylene and graded alcohols before staining. Hematoxylin-eosin and Gomori-methenamine silver staining was performed to examine tissue architecture and for the presence of Coccidioides.

Antibodies.

A panel of murine monoclonal antibodies against human lymphocyte surface markers was purchased from Ventana Medical Systems (VMS, Tucson, AZ), which include anti-CD3 (immunoglobulin 2a [IgG2a], clone PS1), anti-CD4 (IgG1, clone 1F6), anti-CD8 (IgG1, clone 1A5), and anti-CD20 (IgG2a, clone L26). The antibodies were prediluted and optimized to use with the Ventana ES instrument (VMS). Antibodies against human IFN-γ (mouse IgG1, clone G-23; Santa Cruz Biotechnology, Santa Cruz, CA) and interleukin-10 (IL-10) (mouse IgG2b, clone 23738; R&D Systems, Minneapolis, MN) were used at 1:100 and 1:50, respectively, to determine intracellular cytokine expression.

For immunofluorescence assay (IFA), concentrated primary antibodies were used to stain cell surface markers. Anti-CD4 (Biocare Medical, Walnut Creek, CA), anti-CD8 (Vector Laboratories, Burlingame, CA), and anti-CD20 (Vector) antibodies were the same clones as stated above and were used at dilutions of 1:5, 1:10, and 1:50, respectively. Anti-CD25 (mouse IgG2b, clone 4C9) was also from Vector and was used at 1:10. Rabbit anti-human IL-10 polyclonal antibody (Abcam, Cambridge, MA) was used at 1:10. Alexa Fluor-labeled secondary antibodies (goat anti-mouse IgG and goat anti-rabbit IgG) were purchased from Molecular Probes (Eugene, OR) and used at dilutions of 1:20 for green fluorophore (Alexa 488) and 1:400 for red fluorophore (Alexa 568), respectively.

Immunohistochemical staining procedure.

Cell surface markers (CD3, CD4, CD8, and CD20) were stained on an automated Ventana ES slide staining system (17). Antigen retrieval was accomplished by the microwave technique with citrate buffer (pH 6.0). After the slides were loaded into the instrument, primary antibodies were applied, followed by biotin-conjugated goat anti-mouse IgG, avidin-conjugated horseradish peroxidase, 3,3′-diaminobenzidine tetrohydrochloride (DAB) with copper enhancement as a color substrate, and a hematoxylin counterstain. Slides were coverslipped with Permount.

The catalyzed signal amplification system (DakoCytomation, Carpinteria, CA) was used to detect the in situ cytokine expression. High-temperature-induced antigen retrieval involved immersion of tissue sections in preheated (95°C) Target retrieval solution (pH 6.0; DakoCytomation) in a water bath for 20 min and then allowing them to cool for 20 min to room temperature. Prior to staining, endogenous biotin activity was blocked using the Avidin/Biotin blocking kit (Zymed, San Francisco, CA). The staining procedure was carried out at room temperature following the manufacturer's instructions. Antibody concentrations and incubation time were experimentally determined. Following color development and hematoxylin counterstain, the slides were coverslipped with aqueous Universal Mount (Invitrogen, Calsbad, CA). Isotype-matched antibodies were included as negative controls to assess nonspecific background staining.

Slides were examined by light microscopy. Positive cells were identified by the dark brown cytoplasm and counted by a single observer. Microscopic images were acquired using a Leitz Duoplan microscope with a Hamamatsu Orca 100 camera. For lymphocyte subset analysis in the mantle region of the granuloma, the total number of positive cells per high-power field (×40) was counted. For lymphocyte subset analysis of the cell clusters, the total number of positive cells in each cluster was counted. Two to three representative fields or clusters were counted for each specimen, and average numbers were used for statistical analysis. For cytokine expression within the mantle region and in the cell clusters, 100 lymphocytes in each region were counted and the percentage of cytokine-positive cells was determined.

Confocal laser-scanning microscopy.

Single and double confocal immunofluorescence stainings were performed using the automated Universal staining system (DakoCytomation). Antigen retrieval was achieved by immersing the slides in BorgDecloaker solution (Biocare, pH 9.5) and heating in the Decloaker chamber (Biocare). The slides were then treated with cold acetone (−20°C) and 50 mM ammonium chloride for 5 min each. For IL-10-CD4, IL-10-CD8, IL-10-CD20, and IL-10-CD25 double staining, mixed primary antibodies were applied for 30 min. The slides were washed three times with phosphate-buffered saline. Alexa Fluor-conjugated secondary antibodies were then applied for 30 min. For CD4/CD25 double staining, the specimens were first stained with anti-CD25 antibody and goat anti-mouse IgG (Alexa Fluor 488 conjugated), followed by a 5-min treatment with Denaturing solution (Biocare). The anti-CD4 antibody and goat anti-mouse IgG (Alexa Fluor 568 conjugated) were then applied. Finally, specimens were washed again with phosphate-buffered saline, rinsed briefly with water, and fixed in ethanol for 3 to 5 min. The slides were mounted with Fluoromount-G (SouthernBiotech, Birmingham, AL).

Double stains were analyzed using a Leica TCS SP-Leica DM IRBE confocal laser-scanning microscope (Leica Microsystems AG, Wetzlar, Germany). An argon laser operating at 499 to 520 nm was used for visualization of Alexa Fluor 488 fluorescence. For Alexa Fluor 568 fluorescence visualization, a krypton ion laser operating at 552 to 620 nm was used. Noise levels were reduced by line averaging of the scans. Images were overlaid and processed using Adobe Photoshop 7.0 software (Adobe Systems, San Jose, CA)

Statistical analysis.

Data are expressed as the mean ± standard error of the mean. The paired t test was used to determine differences in lymphocyte subset distribution and in the percentage of cytokine-expressing cells within each sample. A P value of <0.05 was considered to be statistically significant.

RESULTS

Histological observations.

Microscopic examination of coccidioidal pulmonary tissues revealed well-formed necrotizing granulomatous inflammation (Fig. 1A). In eight out of nine specimens examined, discrete clusters of small lymphocytes were seen, in which germinal center formation was occasionally observed (Fig. 1B). The clusters were found either at the periphery of the granulomata or in the areas between the granulomata. Coccidioidal spherules and endospores were found in all tissues, predominantly within the necrotic areas of the granulomata (data not shown).

FIG. 1.

Hematoxylin-eosin staining of pulmonary coccidioidal granulomata. (A) A representative granuloma showing central necrosis surrounded by a mantle consisting of mostly lymphocytes (arrow). Several cell clusters (arrowheads) are seen at the periphery of the granuloma. Original magnification, ×4. (B) High-power view of a representative cell cluster. Original magnification, ×25.

Lymphocyte subset composition in coccidioidal granulomata.

Quantitation of lymphocyte subsets was done on the eight samples in which both necrotizing granulomata and cell clusters were observed. As shown in Table 1, within the mantle area of the granulomata of these samples, the majority are T lymphocytes, based on their expression of CD3, while significantly fewer cells expressed CD20 (P = 0.028), a marker for B lymphocytes. The mantle T lymphocytes consisted of both CD4⁺ and CD8⁺ cells in roughly equal proportion. In contrast, the cell clusters were composed of slightly fewer T lymphocytes than B lymphocytes (Fig. 2A). Lymphocytes expressing CD4 (Fig. 2B) were significantly more frequent than those expressing CD8 (Fig. 2C) in the cell clusters (P = 0.003). In addition, the CD4⁺ lymphocytes appeared to be distributed throughout the cluster, while the CD8⁺ lymphocytes were located more sparsely at the cluster periphery.

TABLE 1.

Quantification of lymphocyte subsets and cytokine expression in pulmonary coccidioidal granulomata containing perigranulomatous lymphocyte clusters

Parameter	Mantle	Cluster	P value
No. of cells in lymphocyte subseta
CD3	158 ± 54.1	306 ± 47.1	0.042b
CD4	474.6 ± 13.1	209 ± 40.2	0.011b
CD8	49.6 ± 11.2	62.1 ± 10.6c	0.288
CD20	15.5 ± 4.1d	435 ± 62.0	<0.0001b
Cytokine expression (%)e
IFN-γ	33.0 ± 8.3	4.5 ± 0.5	0.037b
IL-10	40.3 ± 9.2	24.3 ± 13.7	0.094

^aValues represent number of cells (n = 8 samples) expressed as mean ± standard error per ×40, microscopy field in the mantle region and that of the entire cell cluster associated with a single granuloma from each tissue.

^bResults significantly different comparing mantle versus cluster by paired t test.

^cP = 0.003, comparing cluster CD4 to cluster CD8.

^dP = 0.028, comparing mantle CD3 to mantle CD20.

^ePercentage of cells (n = 4 samples) positive for cytokine within the mantle or cell cluster.

FIG. 2.

Immunohistochemical staining for CD20 (A), CD4 (B), and CD8 (C) of a lymphocyte cluster in human coccidioidal granulomata. Original magnification, ×25. Note the large number of CD20⁺ lymphocytes (A) in the cluster. CD4⁺ cells (B) are distributed evenly throughout the cluster and are more frequent than CD8⁺ cells (C), which appear at the periphery of the cluster.

Lymphocytic clusters observed in human pulmonary coccidioidal granulomata were roughly equivalent in size to a ×40 microscopic field, the area used for cell counting in the mantle region. Based on this, lymphocyte subsets in an entire cluster were compared to those counted in a ×40 field in the mantle region (Table 1). Cell density was significantly higher in the cluster than in the mantle for both CD3⁺ (P = 0.042) and CD20⁺ (P < 0.001) lymphocytes. In addition, CD4⁺ lymphocytes were significantly more frequent in the clusters than in the mantle (P = 0.011), while CD8⁺ lymphocytes were not (P = 0.288).

IFN-γ and IL-10 expression within coccidioidal granulomata.

Quantitation of IFN-γ and IL-10 expression was performed in four of the available tissues (Table 1). Of the lymphocytes in the mantle, 33.0% ± 8.3% expressed IFN-γ, compared to 4.5% ± 0.5% of those in the cluster (P = 0.037). On the other hand, 40.3% ± 9.2% of lymphocytes in the mantle produced IL-10 compared to 24.3% ± 13.7% of cluster lymphocytes (P = 0.094).

Colocalization of IL-10 and cell surface markers.

Confocal laser-scanning microscopy was done on a different set of pulmonary coccidioidal tissue specimens to determine the cell types that produce IL-10. Both CD20⁺ and CD4⁺ lymphocytes express IL-10 (Fig. 3A to D), whereas CD8⁺ lymphocytes do not (Fig. 3E and F). IL-10-producing lymphocytes were found in both lymphocytic clusters as well as the mantle area. Macrophages in coccidioidal granulomata also appear to produce IL-10 (Fig. 3B). Because technical difficulties due to variability of tissue staining limited the number of available tissues for examination, we were unable to quantitate the frequency of IL-10-producing CD20⁺ and CD4⁺ lymphocytes.

FIG. 3.

Confocal laser-scanning micrograph of human pulmonary coccidioidal granulomata (scale bar as indicated; original magnification, ×40). The specimens were stained for IL-10 (red) and lymphocyte cell surface markers CD4, CD8, and CD20 (green). Arrows indicate the colocalization (overlay of red and green appears as yellow) of IL-10 and the respective surface maker, while arrowheads indicate the cytoplasmic (intracellular, red) presence of IL-10. IL-10 was produced by cells both in the clusters (A, C, E, and G) and in the mantle regions (B, D, F, and H). While both CD4⁺ (A and B) and CD20⁺ lymphocytes (C and D) express IL-10, CD8⁺ lymphocytes do not (E and F). Panel B depicts a cell expressing IL-10. The morphology and vesicular appearance suggested a macrophage. Panels G and H represent negative controls.

CD4⁺ CD25⁺ T cells in coccidioidal granulomata.

The finding of in situ IL-10 expression in the cell clusters suggested an immunosuppressive activity. Among other cells, regulatory T cells produce IL-10 (IL-10 Tregs) and down-regulate immune responsiveness (15). Naturally occurring Tregs (nTregs) display a phenotype of CD4⁺ CD25⁺ (8, 15) and therefore can be identified by staining for these surface markers. Figure 4 demonstrates the presence of CD4⁺ CD25⁺ cells in coccidioidal pulmonary granulomata. These cells appear to exist in low numbers and are principally located in the perigranulomatous cell clusters (data not shown). Fourteen distinct clusters from three separate cases were examined. The frequency of CD4⁺ CD25⁺ cells was found to be 2.93 ± 0.58, with a range from 0 to 8 cells per cluster per cross-section.

FIG. 4.

Confocal laser-scanning micrograph of lymphocyte clusters in human coccidioidal granulomata (scale bar as indicated; original magnification, ×40). Tissues were stained for CD4 (red) and CD25 (green). Representative cell clusters are shown. Panels A and B show the same cluster stained for CD4 and CD25, respectively. Panels C and D show another cluster from the same tissue. CD4⁺ CD25⁺ cells are yellow (arrow) and can be found mostly within the cell clusters. Panel D represents an enlarged portion of panel C showing cell surface colocalization (yellow outline) of the CD4 and CD25 antigens.

Comparing serial sections from the same tissues, all of the CD25⁺ lymphocytes were also found to be CD4⁺. Therefore, tissues from two cases were double stained for CD25 and IL-10 to assess if CD4⁺ CD25⁺ lymphocytes were producers of IL-10. Confocal laser-scanning microscopy revealed no instances of colocalization of these two proteins in any cells (data not shown).

DISCUSSION

Several unique observations emerged from this examination of pulmonary granulomata in human coccidioidomycosis. First, perigranulomatous lymphocyte clusters contained relatively large numbers of CD20⁺ lymphocytes and increased numbers of CD4⁺ lymphocytes. While such clusters have not been previously reported in human coccidioidomycosis, similar aggregates have been described in other diseases. Fazel and colleagues noted B-cell aggregates strikingly similar to those observed in the present study, using formalin-fixed biopsy tissue from patients with sarcoidosis (7). Similar aggregates consisting of B lymphocytes were also described in 37 of 38 cases of cryptogenic fibrosing alveolitis (21). In the latter study, the authors suggest that these aggregates could be serving as local areas of humoral immunity.

In addition, the presence of the T-lymphocyte helper type 2 (T_H2) cytokine IL-10 within the coccidioidal granulomata has not previously been reported. The finding of IL-10 in association with these necrotizing granulomata is in contrast to reports examining peripheral blood lymphocyte response to coccidioidal antigen, where IL-10 and its message have not been detected (5). In murine models, IL-10 has been shown to be critical in determining susceptibility to coccidioidal infection. Fierer et al. (9) reported that, after intraperitoneal infection with Coccidioides, susceptible C57BL/6 mice had 1,000-fold more IL-10 mRNA in the lungs than resistant DBA/2 mice. The IL-10 level correlated with the colony counts of fungus and hence the severity of infection. In addition, IL-10-knockout C57BL/10 mice possessed resistance equivalent to that of DBA/2 mice. IL-10 is known to provide feedback inhibition to limit inflammation (16). In the coccidioidal granulomata we examined, the number of cells expressing IFN-γ was roughly equivalent to the number expressing IL-10 in the mantle. However, the balance of cytokine expression in the cell clusters was shifted in favor of IL-10, suggesting that these clusters are regions where down-regulation of the cellular immune response is occurring during the coccidioidal granulomatous response.

IL-10 is produced by a broad array of cells, including Tregs, B cells, dendritic cells, and macrophages (16). Our data suggest that B cells, CD4⁺ lymphocytes, and macrophages are the major source of IL-10 in coccidioidal granulomata. While it is tempting to postulate that Tregs were a source of IL-10 in this study, our methods were inadequate to assess this. Tregs are a group of cells with variable origins and phenotypes with the capability of suppressing cellular immune responses (3, 19). Since CD4⁺ lymphocytes express CD25 upon antigen activation, coexpression of CD4 and CD25 cannot alone define a Treg. In fact, our confocal study demonstrated that CD4⁺ CD25⁺ cells were not associated with IL-10 expression in coccidioidal granulomata and may not be down-regulatory.

Because pulmonary tissue was examined, it is difficult to compare the present results to other studies of granulomatous disease, where extrapulmonary tissue was investigated (11, 12, 14, 18). In particular, Modlin and colleagues' data from humans principally consisted of skin biopsies from patients with disseminated coccidioidomycosis (13). The finding of a positive ratio of CD4⁺ to CD8⁺ lymphocytes and the production of IFN-γ in the mantle region of granulomata in the present study are consistent with a reaction favoring a strong cellular immune response. Other cytokines, especially tumor necrosis factor alpha, also play important roles in the development of the granulomatous response. Unfortunately, staining for tumor necrosis factor alpha was unsuccessful in our studies.

There were several limitations of this study. While the pulmonary tissue examined suggested a functional cellular immune response, as indicated by the necrotizing granulomata observed, clinical correlation was not available. Moreover, no tissues obtained from sites of coccidioidal dissemination or those due to other infections were studied. Finally, a vexing problem was the use of formalin-fixed tissues. In particular, staining for intracellular proteins in these specimens required harsh antigen retrieval conditions and signal amplification, which resulted in high background. There was also great variability in staining both by immunohistochemistry and by IFA, limiting our ability to analyze tissues. Future studies will attempt to focus on frozen tissue where staining is more uniform.