Cutting Edge: Impaired Mast Cell Development and Innate Immunity in Mac-1 (CD11b/CD18, CR3)-Deficient Mice¹

Mast cells (MCs)⁴ are bone marrow-derived cells that play an important role in normal host defense, as well as in allergic diseases such as asthma (1, 2, 3). MC recruitment into tissues is thought to occur by the release of MC precursors from the bone marrow into the peripheral blood, followed by the migration of these precursors into mucosal and connective tissues and their differentiation into mature MCs (1, 2, 3, 4, 5). The phenotype of the mature MCs is regulated by the microenvironment of the tissue (1, 2, 3, 4, 5, 6). At least some of the MCs present in interstitial tissues can be induced to proliferate (1, 3, 4, 5). Consequently, the proliferation of differentiated MCs in tissues, as well as the recruitment and maturation of MC precursors, can contribute to the local expansion of MC populations (1, 2, 3, 4, 5). MCs participate in immunological and inflammatory responses, including innate immunity (7, 8): For example, peritoneal mast cells (PMCs) are a central component of the host’s defense against bacterial infection. Genetically MC-deficient Kit^W/Kit^W-v mutant mice exhibit increased mortality following cecal ligation and puncture (CLP), a model of acute bacterial peritonitis (7).

The mechanism(s) whereby immature MCs leave the bloodstream during normal development, or in the context of inflammatory responses, remain poorly understood. It appears likely, based upon studies of other leukocytes, that the emigration of MC precursors into the tissue requires their interaction with the blood vessel wall, a process that in leukocytes is mediated by leukocyte adhesion receptors such as the β₂ integrins. The β₂ integrin Mac-1 (CD11b/CD18, CR3) is present on neutrophils, monocytes, NK cells, eosinophils, and basophils and has also been shown to be weakly expressed on a human leukemic MC line HMC-1 (9), and in a small fraction (<5%) of in vitro-derived human MCs (10). Although Mac-1 has not been detected in mature MCs isolated from various human tissues (11), it is possible that the techniques for isolation of such cells which require enzymatic digestion, may have altered surface Mac-1 expression. Alternatively, Mac-1 may not be expressed in the mature MCs of the tissues examined. Thus, the expression of Mac-1 on non-neoplastic, in vivo-derived tissue MCs, and its relevance to MC development and/or function, has not been previously reported. Mac-1 mediates many leukocyte functions including leukocyte adhesion to the endothelium (12). Mice deficient in this receptor have defects in several neutrophil functions (13, 14, 15). In this study, we show that Mac-1 is expressed on normal PMCs and that it is important both in establishing resident MC populations and in the expression of MC-dependent innate immunity.

Mice which lack Mac-1 (Mac-1^−/−), generated by gene targeting (13), and their littermate wild-type controls were bred and maintained in a virus Ab-free facility at the Longwood Medical Research Center of the Harvard Medical School (Boston, MA). The Mac-1^−/− and wild-type colonies are of a mixed C57Bl/129Sv background that is maintained by ongoing heterozygous breedings. Mice from these litters are then used to set up wild-type and homozygous matings. To prevent the two genotypes from straying in their background genes, new breedings with mice derived from heterozygous breedings are set up routinely, avoiding matings between littermates. Animals used for experiments were age- and sex-matched for each experiment described.

The peritoneum of mice was lavaged as previously described (16). For flow cytometric sorting, cells were stained with a phycoerythrin (PE)-conjugated mAb to CD117 (c-kit) (clone 2B8; PharMingen, San Diego, CA) and sorted on a flow cytometer (EPICS Elite; Coulter, Miami, FL). In select experiments, PMCs were purified using a 22.5% metrizamide gradient as previously described (17).

BMCMCs were obtained by culturing bone marrow cells from 4- to 6-wk-old wild-type and Mac-1^−/− mice in medium containing 100 U/ml of murine rIL-3 for 7–8 wk as previously described (18). BMCMC populations were >90% c-kit positive. In some experiments cells were incubated with LPS from Escherichia coli 0.26:B6 (Difco, Detroit, MI) for 24 h.

c-kit-PE labeled PMCs, recovered by flow cytometry, were transferred to a slide, air dried, fixed in acetone, washed in sterile PBS, and incubated with a FITC-conjugated IgG2b mAb specific for murine Mac-1 (clone M1/70; PharMingen) a FITC-conjugated mAb to CD18 (C71/16) or a FITC-conjugated IgG isotype control. BMCMCs were stained with both anti-c-kit PE and anti-Mac-1-FITC mAbs or a directly conjugated IgG isotype control as described (9), and analyzed on a FACSCalibur (Becton Dickinson, San Jose, CA).

PMCs were stained and counted with Kimura stain (19). The numbers of MCs in tissue was counted by light microscopy (×400) of Alcian blue-stained sections of Carnoy’s fixed, paraffin-embedded tissues (4). The data were expressed as the number of MCs per mm² of tissue, except in the jejunal mucosa, where it was MC per villus crypt unit. The histological analysis was performed by an observer who was blinded to the origin of the tissue sections.

Mice were subjected to CLP, as has been previously described (7). Briefly, 20 wild-type and 20 Mac-1^−/−, 8- to 12-wk-old male mice were anesthetized and subjected to ligation of ∼50% of the distal end of the cecum followed by a single puncture with a 0.7-mm needle. After the procedure, animals were observed for mortality over a period of 14 days. Lavage fluid was harvested from select mice before and 3 h following CLP: The total number of cells present was assessed on a hematocytometer and a differential count on cytospinned samples was performed as previously described (13). Cell associated and spontaneously released histamine in peritoneal fluid was measured by enzyme-immunoassay (Immunotech, Westbrook, ME), and TNF-α levels in peritoneal fluid were detected by ELISA (R&D Systems, Minneapolis, MN) using manufacturers protocols.

Five wild-type and five Mac-1^−/− mice were subjected to CLP and, at 1 or 3 h after CLP, peritoneal lavage fluid was collected and pooled for each genotype. CFU of bacteria were determined by overnight culture of serial dilutions of peritoneal lavage fluid as previously described (20).

Data are presented as average ± SEM. Statistical significance was assessed by the Mantel-Cox Logrank test for the survival rate in the CLP experiment, and by unpaired Student’s t test for the remainder of the experiments.

MCs represent only 1–3% of the cells recovered from the peritoneal lavage of mice, while the remainder of the cells are predominantly Mac-1 positive macrophages that aggregate with MCs upon addition of Ab (data not shown). Therefore, a pure population of PMCs was first obtained from wild-type and Mac-1^−/− mice by FACs sort of c-kit-positive cells before subjecting samples to immunofluorescence staining with Mac-1 Ab and IgG isotype control. Among mature hematopoietic cells, c-kit is only significantly expressed on MCs (21). In the wild-type MC population, ∼30–40% of the c-kit-positive MCs were positive for Mac-1, whereas all c-kit-positive cells from Mac-1^−/− mice incubated with the anti-Mac-1 Ab revealed staining that was similar to that observed in the IgG control. c-kit-positive peritoneal MCs from wild-type and Mac-1-deficient mice were also positive for CD18. The percentages of wild-type or Mac-1-deficient MCs that were positive for CD18 was similar to that seen in wild-type MC populations stained with Ab to Mac-1 alone (Fig. 1). Similarly, Mac-1 was expressed at a frequency of ∼30% by wild-type PMCs purified by density centrifugation (data not shown). Mac-1 expression was also determined in MCs that were differentiated in vitro by maintaining bone marrow cells derived from wild-type and Mac-1^−/− mice in medium containing rIL-3. IL-3 differentiates MC progenitors into immature MCs that are c-kit positive (>95%) (1, 2, 3); these cells are referred to herein as BMCMCs. c-kit-positive wild-type BMCMCs had weak, if any, surface Mac-1 expression as assessed by FACs analysis, and c-kit-positive BMCMCs from Mac-1^−/− mice were completely negative (Fig. 1,B). However, Mac-1 expression could be up-regulated on wild-type BMCMCs upon their incubation with LPS from E. coli for 24 h (Fig. 1 B).

As part of an effort to assess the physiological relevance of Mac-1 on MCs, we quantified the numbers of MCs resident in several tissues harvested from wild-type and Mac-1^−/− mice. Significant reductions in the numbers of MCs were observed in the peritoneal cavity and in the peritoneal wall of Mac-1^−/− mice (Table I), whereas the total number of cells in the peritoneal cavity of wild-type and Mac-1^−/− mice as well as the number of peritoneal macrophages were similar in mice of both genotypes (data not shown) (13). All of the PMCs in Mac-1^−/− and wild-type mice stained positive for Alcian blue and safranin, as would be expected for mature serosal-type MCs (data not shown). The number of MCs in the dorsal skin dermis, but not in the ear skin dermis of Mac-1^−/− animals, was also reduced compared with their wild-type littermates. In the jejunal and cecal mucosa and spleen, Mac-1^−/− and wild-type mice had comparable numbers of MCs (Table I).

The role of Mac-1 in MC-dependent innate immunity was examined by subjecting Mac-1^−/− and wild-type mice to CLP, an acute septic peritonitis model in which enteric bacteria are released into the peritoneum. This model is both MC (7) and complement (20) dependent. Mac-1^−/− mice subjected to CLP had significantly increased mortality (50%) compared with wild-type mice (10%, p = 0.009) by 48 h after CLP, a difference which persisted throughout the 14-day period of observation (Fig. 2 A). This increased susceptibility of Mac-1^−/− mice to sepsis is probably due to a defect in bacterial clearance: The number of CFU in peritoneal lavage fluid of Mac-1^−/− mice was nearly threefold higher than in wild-type mice after 1 h (wild-type: 0.5 × 10⁴ CFU; Mac-1^−/−: 1.4 × 10⁴ CFU), and was still 1.5 fold higher at 3 h after CLP (wild-type: 90 × 10⁴ CFU; Mac-1^−/−: 142 × 10⁴ CFU).

The protective role of MCs in innate immunity is mediated, at least in part, through the elaboration of mediators which can promote neutrophil recruitment (7, 8, 20). Neutrophil accumulation in the peritoneum at 3 h after CLP was diminished by 36% in Mac-1^−/− compared with wild-type mice (Fig. 2,B). The actual difference may even be underestimated since we have previously shown that neutrophil apoptosis is delayed in Mac-1^−/− mice, thus leading to enhanced neutrophil accumulation in the peritoneum of Mac-1^−/− mice during chemically induced peritonitis (13). Histamine was diminished by 84% in Mac-1^−/− mice 3 h after CLP (Fig. 2 C). A decrease in total histamine content at 3 h following CLP vs baseline in both genotypes is probably the result of degranulation of MCs and subsequent metabolism of this MC marker. Preliminary experiments have revealed no difference in TNF-α in wild-type vs Mac-1^−/− mice (data not shown).

In these studies we demonstrate that Mac-1 is expressed on peritoneal MCs, and, using Mac-1^−/− mice, we determine that Mac-1 plays an important role both in establishing physiological levels of resident MCs in certain normal tissues and in maintaining normal levels of host defense in a MC- and complement-dependent model of acute bacterial peritonitis. Furthermore, we show that Mac-1 expression on MCs is induced by LPS from E. coli, which can play a key role in innate immunity.

In the present study, Mac-1 was shown on MCs harvested from the peritoneal cavity of mice, which to our knowledge is the first report of Mac-1 on non-neoplastic in vivo-derived MCs. However, only ∼30–40% of the total population of PMCs were reproducibly Mac-1 positive. This suggests that surface expression of Mac-1 in PMCs may be subject to regulation by microenvironmental signals. Indeed, MCs can express substantial heterogeneity with respect to multiple characteristics, including their content of MC proteases and other mediators (1, 2, 3, 6) as well as their proliferative potential. For example, only 25% of morphologically mature PMCs proliferate in methycellulose culture in vitro (22). We show that BMCMCs can be induced to express Mac-1 upon stimulation with endotoxin. In mice, changes in the phenotypic characteristics of MC populations, in response to alterations in the microenvironment, may allow the MCs to respond appropriately to local changes associated with diseases or immunologic responses (1, 2, 3).

Mac-1^−/− mice had significant reductions in the numbers of MCs resident in certain anatomical sites. The mechanisms that influence the establishment of resident MC populations may differ for different organs at least in adult mice (23, 24) and remain poorly understood. How Mac-1 affects the establishment of resident MC populations is not yet clear, but our study is the first demonstration that an adhesion receptor can significantly influence this process. Populations of resident MCs can be maintained by migration of MC progenitors, as well as by the proliferation and survival of differentiated MCs (3, 4). It is possible that Mac-1 plays a role in one or more of these processes. A committed MC precursor cell (the “pro-mastocyte”) has been identified in the circulation of fetal mice (5), but the mechanisms that regulate the migration and distribution of immature and mature MCs are largely unclarified. Given that Mac-1 on neutrophils and monocytes plays a prominent role in leukocyte trafficking through interaction of these leukocytes with the endothelium (12), it is possible that Mac-1 may play a similar role in recruitment of MC progenitors by facilitating their interaction with endothelial cells. Our finding that Mac-1 is important in establishing resident MCs in some tissues, such as the peritoneum and skin, but not in others further highlights the heterogeneity in the mechanisms which can regulate the baseline levels of MCs in different organs as has been suggested by previous studies (23, 24). It is possible that in tissues where Mac-1^−/− mice had no reduction of MCs that Mac-1 is not expressed on these MCs and/or their progenitors, and therefore is not important in the biology of these MC populations, or that these cells do express Mac-1, but that Mac-1 is not critical in establishing or maintaining baseline populations of MCs at these sites. A parallel situation occurs in the biology of neutrophils, in that Mac-1 is expressed in all neutrophils, yet is not important in the recruitment of neutrophils in all inflammatory conditions (25).

The relevance of Mac-1 on MCs in innate immunity was investigated using a MC-dependent model of bacterial infection in the peritoneum (CLP), a site where Mac-1 is expressed on PMCs and the number of PMCs in Mac-1^−/− mice is significantly less than in wild-type mice. Following CLP, the Mac-1^−/− mice exhibited significant mortality compared with wild-type mice, which was associated with an increased bacterial burden. The defect in bacterial clearance in Mac-1^−/− mice may be due to one or more of the many consequences of the animals’ Mac-1 deficiency. For example, the defect may be related to the observed reduction in neutrophil accumulation into the peritoneum, perhaps reflecting decreased MC-derived mediators in the peritoneum of Mac-1^−/− mice. The defect may also reflect abnormalities in other processes which have been shown to be Mac-1 dependent. These include the inability of Mac-1^−/− neutrophils to adhere normally to the endothelium, and/or to bind complement and phagocytose bacteria (26).

Our studies provide evidence that an adhesion molecule can regulate MC development and/or survival in vivo, and can importantly influence the expression of a MC-dependent innate immune response in vivo.