INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: granuloma; osteopontin; macrophage

Granulomatous inflammation is a form of chronic inflammation that occurs in response to a variety of inciting antigens and in many distinct diseases (1, 2). Granulomas are pathologically defined by the presence of macrophage-derived epithelioid cells and giant cells. Structural similarity in granulomas of diverse etiology suggests that a common set of pathological signals regulates granulomatous inflammation. Although numerous chemotactic and cell-activating cytokines have been identified within granulomatous lesions, the identification of critical determinants of granuloma evolution has been elusive, and factors controlling macrophage and epithelioid cell accumulation are poorly defined (2, 3).

Osteopontin (Opn) is an arginine-glycine-aspartate (RGD)- containing adhesive glycoprotein that is abundantly expressed in many granulomatous diseases, ranging from hypersensitivity granulomas of tuberculosis and sarcoidosis to foreign body granulomas (4). Recent evidence has shown that osteopontin functions as a secreted soluble cytokine serving to modify the migration and function of various inflammatory cells, and has thus been classified as a matricellular protein (4, 8). In this regard, we and others have found that Opn regulates macrophage and T cell migration and Th1 cytokine expression (7). Opn-null (Opn/) mice have been shown to have defective macrophage accumulation at sites of renal injury and to develop abnormal cell-mediated immunity to intracellular infection (10, 12). In addition, Opn deficiency in human and murine hosts is associated with disseminated Mycobacterium (M.) bovis Bacille Calmette-Guérin (BCG) infection and poorly formed granulomas (13, 15). These data suggest that Opn may play a role in granulomatous inflammation. Although one study demonstrated reduced cutaneous granuloma-like reactions in Opn/ mice, little is known regarding the effect of Opn on typical antigen-specific hypersensitivity granulomatous inflammation, particularly involving the lung (10).

The embolization of Schistosoma mansoni (S. mansoni) eggs to the lung via tail vein injection in naive mice induces a characteristic granulomatous response (16). It is not meant to mimic schistosomiasis but rather provide a well-defined model of delayed type hypersensitivity (DTH) granuloma formation. The model involves the synchronous development of T cell-dependent and antigen-specific pulmonary granulomas and progresses through characteristic phases of initiation, maintenance, and resolution (16). The response involves coordinated recruitment of macrophages and T cells to sites of antigen challenge and is regulated by both Th1 and Th2 cytokines (19). Although certain chemokines, such as monocyte chemotactic protein (MCP) 1, are important in cellular recruitment, critical determinants of granuloma formation and structure remain undefined (18).

Using the S. mansoni egg pulmonary granuloma model in naive mice, we studied the effect of Opn absence on pulmonary granuloma structure, cellular composition, and evolution in vivo. We found that compared with control mice, the granulomatous response in Opn-null mice was delayed and granulomas lacked typical epithelioid morphology and contained markedly fewer macrophages. In contrast, T cell accumulation was normal in these mice. These results suggest that Opn may regulate macrophage accumulation in granulomatous pathology and is required for normal granuloma formation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

The generation of Opn/ (deletion of exon 4-7 on 129SVJ × Black Swiss background) was previously described and the genotype of all mice confirmed by PCR analysis (22). By RT-PCR the Opn mutant gene was not transcribed and Opn mutant mice had normal baseline immunological phenotypes with normal numbers of inflammatory cells and normal mitogenic proliferative responses (not shown). Animal studies were approved by the Boston University Institutional Animal Care and Use Committee.

Induction of Synchronous Pulmonary Granuloma

S. mansoni eggs isolated aseptically from livers of infected mice were obtained from National Institutes of Health (F. Lewis, SnailsRus). Synchronous pulmonary granulomas were induced by intravenous injection of 3,000 eggs into female naive mice aged 6-10 wk. At Days 10, 20, 30, 75, and 125 mice were sacrificed and their lungs inflated and fixed in 4% paraformaldehyde. At each time point four Opn/ mice were compared with four wild-type (Opn+/+) control mice.

Histological Analysis

Granulomatous lungs were stained with hematoxylin and eosin and analyzed for granuloma size and morphology. Granuloma area was determined using a tissue micrometer and was expressed as mean granuloma area (µm² × 10³) + standard deviation (SD) per lung. These measurements were averaged for all Opn mutant and wild-type mice per time point (mean + standard error [SEM]). To control for variation in egg size between granulomas, results were also expressed as ratio of granuloma area to egg area (granuloma area/egg area). Granulomatous reactions were analyzed for the presence of macrophages, epithelioid cells, lymphocytes, giant cells, and eosinophils. Eosinophil content was graded: 0 = no eosinophils; 1 = eosinophils detectable; 2 = eosinophil abscess. Eosinophil content was then expressed as mean eosinophil score per mouse + SD. Granulomas containing giant cells were counted and results were expressed as mean percent granulomas containing giant cells per animal + SD.

Immunohistochemistry

Immunohistochemical staining for Opn, macrophages, and T cells was performed on paraffin-embedded specimens as previously described (22, 23). Specimens were sequentially treated with H₂O₂ (0.3%), trypsin (0.25%), and 1.5% normal rabbit serum. The sections were then incubated with primary Ab at 4° C overnight. Antibodies used were Op199 (goat anti-rat Opn antibody, purified to IgG subtype; concentration 2 µg/ ml), rat anti-mouse monocyte/macrophage marker F4/80 (clone Cl:A3-1,1:5 dilution, Serotec Inc, NC), and rat anti-mouse CD3 (clone CD3-12, 1:10 dilution; Novocastra Laboratories, CA) (22). Control slides were incubated with appropriate normal serum or isotype mAb. The slides were then incubated with biotinylated secondary antibody (10 µg/ml; Vector Laboratories). Staining was detected with avidin-biotinylated horseradish peroxidase (ABC) and diaminobenzidine and graded as follows: macrophage staining: 0 = no F4/80+ cells in granuloma; 1 = less than 10 F4/80+ cells per granuloma; 2 = greater than 10 F4/80+ cells per granuloma. T cell (CD3) staining was read as present or not present.

Statistics

All analysis was performed by a pathologist blinded to the experimental conditions and at each time point at least 75 granulomas were assessed in each group (30 + 5, mean + SEM, granulomas were counted per lung). Significant differences were evaluated by paired Student's t test for granuloma size and by chi square test for granuloma morphology and grade.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Opn Is Expressed in S. mansoni Egg Pulmonary Granulomas

Granulomas were induced by tail vein embolization of S. mansoni eggs. Granuloma-associated Opn expression was confirmed by immunohistochemistry using Op199 Ab (22, 24). Figure 1a shows a lung section from Opn+/+ mice at peak granuloma formation (Day 20). Opn was expressed in a cell-associated manner within pulmonary granulomas in Opn+/+ mice (Figure 1a). Both alveolar macrophages and granuloma macrophages were immunoreactive for Opn, whereas T cells did not stain for Opn. Opn was not detected in the extracellular matrix. Similar immunoreactivity for Opn was seen in granulomas from Day 5 to Day 30 but no Opn was detectable by Day 75 (not shown). No Opn staining was seen in Opn/ mice (Figure 1b) or with isotype control antibody (not shown).

View larger version (114K): [in this window] [in a new window]   Figure 1.   Immunohistochemical analysis of Opn expression in Opn+/+ mice (a) and Opn/ mice (b). Lung granulomatous responses 20 d after egg embolization are shown. In Opn+/+ mice, Opn immunoreactivity is present in cells surrounding S. mansoni egg. No staining is seen in Opn/ specimens (b). Original magnification: ×400.

Granuloma Size Is Reduced in Opn-deficient Mice

Having shown that Opn was expressed in pulmonary granulomas induced by S. mansoni eggs, granuloma size was measured at various time points. As shown in Figure 2 and Figure 3, in wild-type Opn+/+ mice, granuloma size was typical of previous studies involving this model (18, 20). Inflammatory cell infiltration around S. mansoni eggs was seen by Day 5 and granuloma formation was present by Day 10 reaching peak size at Day 20 (Figure 2a and Figure 3). After 20 d, granuloma size gradually decreased with significant granuloma resolution occurring by Day 75 (Figure 2c).

View larger version (140K): [in this window] [in a new window]   Figure 2.   H&E stain showing granuloma size and morphology at Day 20 and Day 75 in Opn+/+ (a, c, e) and Opn/ (b, d, f  ). At Day 20 Opn+/+ (a, c) granulomas are larger than Opn/ (b, d ) granulomas, and contain abundant epithelioid cells. In contrast Opn/ (b) granulomas are composed predominantly of mononuclear lymphocytic cells. At Day 75, granuloma involution and fibrosis are occurring in Opn+/+ (e) mice but Opn/ (f  ) granulomas are cellular and lymphocytic. Original magnification: × 40 (a, b); × 100 (c-f  ).

View larger version (38K): [in this window] [in a new window]   Figure 3.   Granuloma area (µm × 103) at various time points in Opn+/+ and Opn/ mice. As is typical of this granuloma model, in Opn+/+ mice peak granuloma formation occurs at Day 20 with significant resolution by Day 75. Compared with Opn+/+ mice, Opn/ granuloma formation is delayed with reduced granuloma size at Day 20 and less reduction in granuloma size at Day 75. Peak granuloma formation occurs at Day 30 in Opn deficiency. *p < 0.001.

The pattern of granuloma evolution was different in Opn/ mice with the growth of the Opn/ granuloma lesions lagging behind those of wild-type mice. By Day 20 granulomas were significantly smaller in Opn/ mice compared with wild-type control mice (Figure 2a-2d and Figure 3; Opn+/+: 29 + 4 µm² × 10³ versus Opn/: 13.2 + 2 µm² × 10³; p < 0.001). In fact, peak granuloma size occurred at Day 30 in Opn-deficient mice, at which time granuloma size was declining in control mice (Figure 3; Opn+/+: 25.6 + 3 µm² × 10³ versus Opn/: 29.4 + 3.9 µm² × 10³; p = 0.25). Although in both Opn+/+ and Opn/ mice granuloma size was decreasing by Day 75, at this time point granuloma size had regressed to a significantly greater degree in Opn+/+ mice compared with Opn/ mice (Figure 2e and 2f and Figure 3; Opn+/+: 16.1 + 3 µm² × 10³ versus Opn/: 22.5 + 1.3 µm² × 10³; p < 0.05). By Day 125, however, few eggs or granulomas were detected in Opn+/+ and Opn/ mice, suggesting that even in the absence of Opn expression, regression of the inflammatory response had occurred at this time.

To ensure that differences in granuloma size did not reflect the size of the eggs forming the nidus of the granuloma, the data were also analyzed as the fold increase in total granuloma area compared to egg area. By this analysis, at Day 20 granuloma area was 11.6 times larger than egg area in the presence of Opn compared with 5.5 times in Opn-deficient mice. In summary, Opn deficiency was associated with a delay in granuloma formation resulting in a reduction in early granuloma size.

Epithelioid Granuloma Formation Is Reduced in Opn-deficient Mice

Pathologically, granulomas are defined as a collection of epithelioid cells surrounded by a rim of activated T cells. We wished to determine whether granuloma structure was normal in the absence of Opn expression. Cellular responses around eggs were of two distinct varieties, namely (1) epithelioid granulomas (Figure 2a and 2c; typical granulomas with a central collection of large cells with prominent cytoplasm consistent with macrophages and macrophage-derived cells such as epithelioid cells), or (2) mononuclear granulomas (Figure 2b and 2d; a collection of small mononuclear cells with scant cytoplasm typical of lymphocytes or monocytes). In the presence of Opn, the majority of granulomas were epithelioid in nature at Days 10, 20, and 30 (Table 1; Figure 2a and 2c). In the absence of Opn, mononuclear granulomas predominated at these time points (Table 1 and Figure 2b and 2d). These differences were highly significant (p < 0.001). Even at Day 75 when granulomatous inflammation was resolving in Opn+/+ mice (Table 1 and Figure 2e), a majority of Opn/ granulomas demonstrated mononuclear rather than epithelioid morphology (Table 1 and Figure 2f).

Giant cells were present in both groups of animals but at earlier time points (Days 20 and 30) giant cells were significantly more common in control animals compared to Opn null mice (Table 1, p < 0.001). By Day 75, however, significant giant cell formation was seen in Opn/ animals. As expected prominent eosinophil accumulation was detected in granulomas at Day 20 but the presence or absence of Opn did not appear to alter this aspect of the host response to S. mansoni ova (eosinophil score Opn+/+: 1.35 + 0.6 mean + SD versus Opn/: 1.12 + 0.3, mean + SD, p = 0.49). Figure E1 in the online data supplement shows a pulmonary granuloma induced in an Opn+/+ mouse stained with H&E to demonstrate abundant eosinophils in close proximity to the S. mansoni ova.

Macrophage Accumulation Is Defective in Opn-deficient Mice

Defective epithelioid granuloma formation may reflect failure to accumulate monocyte/macrophages or failure of recruited macrophages to differentiate into epithelioid cells. We assessed for the presence of monocytes and macrophages in granulomas using immunohistochemical detection of a universal monocyte/macrophage marker, F4/80 (23, 25). Macrophage content was assessed by a pathologist blinded to the experimental conditions and graded from 0 to 2 in the manner outlined in METHODS. In control mice, macrophage accumulation was evident at all time points with most granulomas containing many cells that were immunoreactive for F4/80 (histology grade 2, Figure 4a and Figure E2 in the online data supplement). In many instances, epithelioid cells also stained positively for F4/80. Conversely, granulomas in Opn/ animals contained few or no macrophages by immunohistochemistry (grade 0 or 1; Figure 4b and Figure E2 in the online data supplement). Figure 5 compares macrophage staining in granulomas from in Opn+/+ and Opn/ mice at Day 30. In Opn+/+ mice 62 + 11% (mean + SEM) of granulomas had many F4/80 positive cells (grade 2), whereas only 8.8 + 5% had no F4/80 positive cells (grade 0). Conversely, in Opn null, 71 + 6% (mean + SEM) of granulomas contained no F4/80 cells and only 2 + 1% had many positive cells. These differences were significant (p < 0.001). Even at 75 d when granulomas were involuting in control mice, Opn-deficient granulomas contained relatively few F4/80 positive cells. Opn/ macrophages did express F4/80 and could react with the anti-F4/80 antibody as determined by positive immunohistochemical staining of non-granuloma-associated lung macrophages, and by FACS analysis of peritoneal elicited macrophages and splenic monocytes in Opn/ mice (not shown). In contrast to macrophage staining, CD3-positive T cells were detected in both control and Opn-deficient mice (Figure 4c and 4d).

View larger version (148K): [in this window] [in a new window]   Figure 4.   Immunohistochemical analysis for the presence of macrophages and T cells in pulmonary granulomas in Opn+/+ and Opn/ mice. In both Opn+/+ (a, b) and Opn/ (c, d ) mice single granulomas were serially sectioned at 5 µm and were stained by immunohistochemistry for the monocyte/macrophage marker F4/80 (a, c) and the T cell antigen CD3 (b, d ). Numerous F4/80 macrophages (a) are present in Opn+/+ granulomas. In contrast, no F4/80 positive cells are detectable in Opn/ granulomas (c). In contrast, the same granulomas contained scattered CD3-positive T cells in both Opn+/+ (b) and Opn/ (d ) mice. Original magnification: ×400.

View larger version (30K): [in this window] [in a new window]   Figure 5.   Granuloma macrophage content in Opn+/+ and Opn/ mice at Day 30. Macrophage content was graded by a pathologist blinded to the experimental conditions as follows: 0 = no F4/80 positive cells per granuloma; 1 = less than 10 F4/80 positive cells per granuloma; 2 = greater than 10 F4/80 positive cells per granuloma. Results are expressed as percent cells positive for each grade. As expected for this granuloma model, most granulomas contain many macrophages (grade 2) in Opn+/+ mice. In Opn-deficient animals significantly more granulomas contain few or no F4/80 positive cells. *p < 0.001.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Opn is a novel cytokine with structural homology to noncollageneous matrix proteins (4). Recent studies have demonstrated that Opn is expressed during pulmonary granulomatous inflammation and can regulate aspects of granuloma formation such as T cell and macrophage migration, activation, and Th1/ Th2 cytokine expression (4). Opn expression has been shown to correlate with the presence of inflammatory cells within pulmonary sarcoid granulomas and with both granuloma formation and disease outcome in patients rendered susceptible to M. bovis BCG due to inherited defects of the interferon gamma receptor (7, 15). The present study is the first to provide insights into the role of Opn in pulmonary granuloma formation. We show that in the absence of Opn expression, granuloma formation was delayed and granulomas contained dramatically fewer macrophages and epithelioid cells in comparison to wild-type Opn-sufficient mice. These results suggest that Opn regulates macrophage accumulation at sites of granulomatous inflammation.

Previous studies also support a role for Opn in regulating macrophage accumulation at sites of inflammation (9, 10, 14). In one study, an IL-12-dependent cutaneous response induced by subcutaneous injection of polyvinyl pyrrolidone (PVP) was not present in Opn/ mice (10). As macrophages account for 80% of cells that accumulate at sites of PVP injection, this study also supports a role for Opn in regulating macrophage accumulation. Another study demonstrated that Opn was produced by granuloma-like lesions in the heart and lung of TO2 cardiomyopathic hamsters, and showed that transgenic expression of murine Opn in the airways of normal hamsters reproduced a similar inflammatory response (26). Renal injury after ureteral obstruction in Opn/ mice is also associated with a five-fold decrease in macrophage infiltration compared with wild-type control mice (14). Other studies have reported normal numbers of macrophages at sites of tissue injury in Opn/ mice (13, 22). In particular, we have previously reported defective killing of M. bovis BCG by Opn/ macrophages but in contrast to our findings in lung granulomas we did not detect differences in granuloma macrophage content in the livers of Opn/ mice intravenously infected with M. bovis BCG (13). This observation highlights the fact that granulomatous inflammation is dependent on organ site and the nature of the injurious agent, and suggests that Opn may be crucial to macrophage recruitment to specific organs including the lung (27).

Although decreased macrophage accumulation may be accounted for by defective macrophage differentiation or survival, it is more likely that our results reflect deficient macrophage recruitment. Opn is chemotactic to a variety of cell types including macrophages and T cells in vitro, however only macrophages accumulate at sites of subcutaneous Opn injection in vivo (7, 9, 28). As well as chemotaxis, Opn may indirectly regulate migration through its effects on intracellular mechanisms responsible for cellular migration to sites of injury or inflammation. The cutaneous inflammatory response elicited by injection of the chemotactic peptide formyl-Met-Leu-Phe (fMLP) is characterized by high macrophage expression of Opn and can be inhibited by intravenous treatment with an antibody to Opn (9). Osteoclasts, which share a common lineage with macrophages, exhibit impaired mobility in the absence of Opn, as do fibroblasts (29). Recent work has shown that intracellular Opn is an integral component of the CD44-ERM (ezrin/radixin /moesin) protein complex that mediates interactions between the plasma membrane and cortical actin filament, and thus cellular migration (29). Opn may also regulate migration by modulating the expression of certain matrix metalloproteinases (MMP) such as MMP2 (30). Although the striking defect in macrophage accumulation in Opn/ pulmonary granulomas may support a universal defect in macrophage motility, accumulation of macrophages in the M. bovis BCG-infected liver in Opn/ mice suggests that further studies are necessary to characterize Opn-dependent mechanisms of macrophage migration (13).

Whereas at earlier time points Opn-deficient granulomas contained fewer macrophages and giant cells, significantly more giant cells were present at later time points in the absence of Opn expression. These data suggest that despite the paucity of macrophages, giant cell formation was intact though delayed in Opn/ mice. Although the mechanisms of giant cell formation are poorly understood, a recent study showed that Opn and other CD44 ligands inhibited rat bone marrow-derived macrophages from fusing to form giant cell in vitro (31). Thus, we speculate that due to a CD44-dependent interaction osteopontin deficiency may be associated with increased giant cell formation in vivo.

Although our results confirm a role for Opn in regulating macrophage accumulation at sites of inflammation, T cell accumulation appeared normal. Similarly, using models of liver granuloma formation, monocytopenic and granulocyte/macrophage colony-stimulating factor-deficient mice develop delayed granuloma formation characterized by defective macrophage accumulation but intact T cell recruitment (32, 33). These studies show that T cells can accumulate in granulomatous pathology in the absence of significant macrophage recruitment. There are no studies that have previously addressed this issue in pulmonary granuloma formation. Overall, our results support other studies that show that Opn does not function as a T cell chemoattractant in vivo and suggest that T cell recruitment is regulated by a distinct set of factors at sites of pulmonary granuloma formation (9).

Granuloma formation is a tightly regulated chronic inflammatory event with characteristic pathological hallmarks. Our results show that Opn/ mice develop reduced macrophage accumulation and abnormal granuloma structure during granuloma development and evolution. Although further studies are necessary to define the mechanisms responsible for these findings, defective macrophage accumulation may contribute to the impaired ability of Opn-deficient hosts to control intracellular infections. Understanding the function of Opn in granuloma formation may ultimately provide a novel diagnostic and therapeutic target in the management of pulmonary granulomatous diseases such as sarcoidosis.