IFN-γ Induces Apoptosis in Developing Mast Cells¹

Mastcells are present in nearly all tissues except blood. When activated, mast cells direct inflammatory reactions by secreting vasoactive and inflammatory substances such as histamine, heparin, proteolytic enzymes, PGs, leukotrienes, and an array of cytokines (1, 2). Mast cell activation results in the recruitment of other immune cells, including Th2 lymphocytes, eosinophils, neutrophils, and basophils, which contribute to pathogen killing and immunopathology (1, 2, 3, 4). This response produces a local or systemic immediate hypersensitivity reaction best studied in atopic diseases such as asthma (1). Although best known for their role in allergic disease, mast cells have recently been implicated in the inflammatory responses associated with rheumatoid arthritis (5), multiple sclerosis (6), heart disease (7), and bacterial infection (8, 9).

Mast cells are derived from Kit⁺/CD34⁺ bone marrow progenitor cells. Their in vitro development is directed by the cytokines IL-3 and stem cell factor (SCF)³ in the mouse system (1, 2). Using this assay, we recently found that the Th2 cytokine IL-4 inhibits the survival of developing bone marrow-derived mast cells (10). These data prompted further study into how Th1 and Th2 cytokines may alter mast cell survival.

IFN-γ production is the hallmark of the Th1 response, regulating the activation, growth, and differentiation of T cells, B cells, macrophages, NK cells, endothelial cells, and fibroblasts (11, 12). Through these activities, IFN-γ plays a key role in mediating antiviral and proliferative responses (13). Like many cytokines, IFN-γ signals through the Jak-Stat pathway (14). Signal transduction involves receptor aggregation, leading to phosphorylation and activation of preassociated Janus family protein tyrosine kinases Jak1 and Jak2. Jak phosphorylates the IFN-γ receptor, creating binding sites for Stat1. Subsequent Stat1 phosphorylation results in its dimerization and nuclear translocation, whereupon it regulates the transcriptional activity of target genes (12, 14). This signaling cascade is negatively regulated in part through suppressor of cytokine signaling (SOCS)-1. The importance of this regulation is apparent in SOCS-1 knockout (KO) mice, which die perinatally of a Th1 inflammatory disease resulting in multiorgan failure (15).

The Th1 and Th2 responses have been well-described as an antagonistic system. IFN-γ is a product of Th1 cells and exerts inhibitory effects on Th2 differentiation (16). Moreover, IFN-γ treatment of a patient with mastocytosis showed symptomatic improvement that abated as the patient developed anti-IFN-γ Abs (17), implying that IFN-γ may inhibit mast cell-mediated disease in vivo. Because mast cells are now proposed to have critical roles in Th1-mediated responses (5, 6, 18), and can produce IFN-γ upon activation (19), we have sought a regulatory role for IFN-γ in mast cell biology. Importantly, IFN-γ has been shown to inhibit human and mouse mast cell development (20, 21, 22, 23), activation (24, 25, 26), and adhesion (27, 28). In contrast, human mast cells up-regulate FcγRI expression and signaling in response to IFN-γ (29). Despite these advances, mechanisms for the effects of IFN-γ on mast cells are unresolved and have not been assessed in vivo. In the current study we examine the inhibitory effects of IFN-γ on mouse mast cell survival in vitro and in vivo.

Bone marrow cells were cultured in complete RPMI 1640 (Invitrogen Life Technologies), 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 1 mM sodium pyruvate, and 1 mM HEPES (Biofluids). Bone marrow cells were harvested from femurs and tibias of B6129PF₁/J female mice (3–5 wk old) (The Jackson Laboratory) or from the indicated transgenic or gene-targeted mice and their age-matched littermates. Bone marrow cells were plated at 3 × 10⁵ cells/ml in complete RPMI 1640 containing IL-3 (5 ng/ml) plus SCF (50 ng/ml) with or without IFN-γ (50 ng/ml) for 28 days. Every 4 days nonadherent cells were transferred to new plates and fed with media and cytokines. Bax-deficient and p53-deficient mice were purchased from The Jackson Laboratory. H2K-Bcl-2 transgenic mice were the kind gift of J. Domen (Duke University, Durham, NC). SOCS-1-deficient (KO), SOCS-1/IFN-γ KO, and littermate animals were the generous gift of Dr. J. Ihle (St. Jude Children’s Research Hospital, Memphis, TN). All mice were maintained under specific pathogen-free conditions. For the analysis of peritoneal mast cell populations, gene-targeted mice and littermates were used at 3–4 wk of age.

Murine SCF, IL-3, and IFN-γ were purchased from PeproTech and R&D Systems. IgE, anti-CD16/CD32 (clone 2.4G2), PE-labeled anti-Kit, PE-labeled anti-IL-3Rβ, anti-IFN-γRα, and PE-labeled anti-Mac-1 (CD11b) were purchased from BD Pharmingen. PE-labeled rat IgG isotope control was purchased from eBioscience. FITC-labeled rat IgG isotype control, FITC-labeled, PE-labeled, and unlabeled rat anti-mouse IgE were purchased from Southern Biotechnology Associates. FITC-labeled rat anti-mouse T1/ST2 was purchased from MD Biosciences. Rat IgG2b was purchased from Southern Biotechnology Associates.

A total of 200 μl of bone marrow cells were removed from culture, washed in PBS/3% FCS/0.1% sodium azide (FACS buffer) and incubated with anti-CD16/CD32 (clone 2.4G2; 10 μg/ml) for 10 min at 4°C to block FcγR. Cells were then incubated with PE-coupled anti-Kit, FITC-coupled anti-T1/ST2, or isotype control Abs at 10 μg/ml for 30 min at 4°C. Cells were washed twice with FACS buffer and analyzed using a BD Biosciences FACScan.

The CaspaTag assay (Immunochemistry Technologies) was used to measure active caspase 3. In brief, 200 μl of cells were removed from culture and resuspended in 40 μl of fresh complete RPMI 1640 plus 10 μl of 5× FAM-DEVD-FMK for 1 h at 37°C. Cells were washed two times with 1× wash buffer and resuspended in wash buffer and analyzed by flow cytometry. DiOC₆(3) (3,3′-dihexyloxacarbocyanine iodide; Molecular Probes) was used to detect mitochondrial membrane staining. Cells were incubated with DiOC₆(3) at a final concentration of 1 nM for 15 min at 37°C. Cells were washed with PBS and analyzed by flow cytometry.

DNA fragmentation, an indication of apoptosis, was assessed by propidium iodide (PI)-DNA staining. A total of 200 μl of cells were removed from culture and centrifuged in a 96-well V-bottom plate for 5 min. Cells were then washed with PBS and fixed with 150 μl of 70% ethanol/30% PBS for 4 h at 4°C. After fixation, cells were washed with PBS and incubated with PI-DNA staining buffer (100 μg/ml RNase A, 50 μg/ml PI in PBS) for 2 h in the dark at room temperature. Flow cytometric analysis was performed to assess subdiploid DNA content, which was judged as the percentage of the population demonstrating fluorescence values lower than the major G₀/G₁ peak. Samples were assessed using a forward and side light scatter gate to exclude cellular debris for 45 s per sample (time resolution = 0.1 s).

On day 21, 200 μl of bone marrow cells were removed, washed with PBS, and transferred to glass slides by centrifugation. Slides were stained with toluidine blue (0.2% toluidine blue, 0.1 M citric acid in 50% ethanol) for 45 min at room temperature, followed by a deionized water rinse. The number of mast cells and cell morphology was assessed by light microscopy.

Results are the mean and SD for experiments with triplicate cultures. IFN-γ effects on percentage and number of mast cells were measured by comparing with t test for two data points or by ANOVA for multiple data points by SysStat9 software (SPSS). A value of p < 0.05 was considered to be significant.

To determine the effect of IFN-γ on mast cell development, mouse bone marrow cells were cultured with IL-3 plus SCF with or without IFN-γ for 28 days. Apoptosis was measured during the culture period by assessing subdiploid DNA content via PI-DNA staining, as described in Materials and Methods. IFN-γ treatment induced profound apoptosis in these cultures (Fig. 1,A). Cell death did not occur during the first week of culture, but was detected on day 14 (62%), and increased to 98% on day 21 (Fig. 1,B). The apoptotic effects of IFN-γ were concentration-dependent, with apoptosis detectable at concentrations of 0.08 ng/ml, and maximal at 50 ng/ml (Fig. 1 C).

The timing of apoptosis in these cultures coincided with commitment to the mast cell lineage, as overt granulation and FcεRI expression occur during days 7–10 of culture in IL-3 plus SCF (30). To determine whether IFN-γ could induce apoptosis in differentiated mast cells, bone marrow-derived mast cells cultured for 6 wk in IL-3 plus SCF were replated in IL-3 plus SCF with or without IFN-γ for 18 days. As shown in Fig. 1 D, IFN-γ did not affect the survival of these differentiated mast cells. Thus the death-inducing effects of IFN-γ appear to be exerted during mast cell development, with loss of apoptotic signaling after differentiation has occurred.

To determine the mechanism by which IFN-γ induced apoptosis, we measured IFN-γ-mediated effects on Stat1-deficient (Stat1 KO) bone marrow cells. Compared with wild-type cells, IFN-γ (50 ng/ml) did not alter the survival of Stat1 KO bone marrow cells (Fig. 2). The lack of apoptosis in these cultures was not a result of altered expression of the SCF, IL-3, or IFN-γ receptors, as control and littermate populations demonstrated similar flow cytometry profiles (Fig. 2 B). Thus Stat1 expression was absolutely essential for the death-inducing effects of IFN-γ.

We have recently found that the Th2 cytokine IL-4 elicits mast cell apoptosis via a mitochondrial mechanism (10). To determine the importance of the mitochondrion in IFN-γ-mediated apoptosis, we measured changes in mitochondrial membrane potential (Δψ_m) via DiOC₆(3) staining. Bone marrow cells cultured in IL-3 plus SCF with or without IFN-γ were stained on days 6, 10, 12, and 15 of culture. As shown in Fig. 3,A, IFN-γ treatment reduced DiOC₆(3) staining, consistent with loss of Δψ_m. The timing of this change correlated well with the increase in apoptosis, with onset at day 10 of culture. Moreover, reduced Δψ_m was mirrored by activation of the death effector enzyme, caspase-3 (Fig. 3 B).

To determine the importance of the mitochondrion to IFN-γ-mediated apoptosis, we studied bone marrow cells lacking proapoptotic Bax or overexpressing prosurvival Bcl-2 proteins. We also tested cells lacking p53, a transcriptional regulator of Bax. Compared with wild-type cultures, the loss of Bax or p53, as well as the overexpression of Bcl-2, completely suppressed IFN-γ-mediated apoptosis of developing mast cells (Fig. 3,C). Importantly, the lack of IFN-γ-mediated apoptosis in these cultures was not due to alterations in receptor expression, as all cultures showed similar expression levels of SCF, IL-3, or IFN-γ receptors (Fig. 3 D and data not shown). These data support the importance of a mitochondrial pathway used by IFN-γ in this system.

To determine whether IFN-γ affected mast cell differentiation as well as survival, we measured the effect of IFN-γ on expression of Kit and the IL-1R-related protein T1/ST2, both of which are highly expressed on mast cells (2, 31). Bone marrow cells were treated with IL-3 plus SCF with or without IFN-γ for 21 days. As shown in Fig. 4, IFN-γ treatment did not alter the surface expression of Kit or T1/ST2. The fraction of Kit-T1/ST2 double-positive cells in IL-3 plus SCF cultures was 88%, similar to cultures containing IL-3 plus SCF with IFN-γ (92%) (Fig. 4). Mean fluorescence intensities of these stains were also unchanged. Histochemical staining with toluidine blue demonstrated that these cells possessed a mononuclear, granulated appearance consistent with mast cell morphology that was unchanged by IFN-γ treatment (data not shown). Therefore, IFN-γ does not appear to alter mast cell differentiation, but rather reduces survival as mast cells develop.

To determine the physiological relevance of IFN-γ-mediated mast cell apoptosis, we sought a system in which IFN-γ signaling is dysregulated. SOCS-1 deficiency has been shown to result in IFN-γ overproduction and hyperresponsiveness with a Th1-type autoimmunity resulting in multiorgan failure. The symptoms of this disease are greatly improved when these animals are crossed to an IFN-γ KO background, creating SOCS-1/IFN-γ double-deficient (DKO) mice (15, 32, 33). To address the effect of SOCS-1 deficiency on mast cell development in vitro, we cultured wild-type, SOCS-1 KO, and SOCS-1/IFN-γ DKO bone marrow cells in IL-3 plus SCF. Apoptosis were measured by PI-DNA staining on day 21. SOCS-1 KO cells cultured in IL-3 plus SCF demonstrated 90% apoptosis by day 21, compared with 18% apoptosis in wild-type cultures (Fig. 5,A). This degree of cell death matched closely the effects we observed by adding IFN-γ to wild-type bone marrow (Fig. 1). Evidence that the apoptosis observed in SOCS-1 KO cultures was due to endogenous IFN-γ signaling came from cultures of SOCS-1/IFN-γ DKO cells. IFN-γ deletion completely blocked the apoptotic effects of SOCS-1 deficiency (Fig. 5 A).

To determine whether these in vitro data had an in vivo correlate, we measured the number of mast cells in the peritoneum of age-matched wild-type, SOCS-1 KO, IFN-γ KO, and SOCS-1/IFN-γ DKO mice. Mast cell numbers were reduced 67.1% in SOCS-1 KO mice compared with wild-type animals (Fig. 5 B). This change in the mast cell population was not due to alterations in total peritoneal cell numbers because SOCS-1 deficiency did not alter these values (wild-type = 1.45 × 10⁶; SOCS-1 KO = 1.68 × 10⁶ total peritoneal cells; p = 0.64 by Student’s t test). Although IFN-γ deficiency alone did not affect mast cell numbers (p = 0.18, wild-type vs IFN-γ KO by t test), mast cell numbers were consistently increased in SOCS-1/IFN-γ DKO mice. These results indicated that endogenous IFN-γ is essential for the mast cell defect in SOCS-1 KO mice, and that SOCS-1 and IFN-γ may normally combine to limit mast cell numbers in vivo.

The regulation of mast cell function and survival has long been a central issue in Th2-type diseases, specifically atopic conditions such as allergic asthma. However, the importance of mast cell activation is now appreciated in both Th1 and Th2 responses. For example, mast cell-derived TNF-α production was recently shown to elicit lymphadenopathy and to promote neutrophil chemotaxis (4, 9, 34). Further, mast cell deficiency greatly reduced the inflammatory pathology observed in animal models of arthritis, multiple sclerosis, and heart disease (5, 6, 7). These findings emphasize the importance of mast cell homeostasis in preventing the establishment of chronic inflammation.

Our laboratory has focused on the regulation of mast cell activation and survival by cytokines, focusing on IL-4 and IL-10. We have previously found that IL-4 inhibits the survival of developing mast cells (10), and that combined signaling with IL-4 and IL-10 induces apoptosis of differentiated mast cells (35, 36). Given the newly found importance of mast cells in Th1-type pathology, we extended these studies to include the prototypical Th1 cytokine IFN-γ. Several laboratories have shown inhibitory effects of IFN-γ on mast cell survival, proliferation, or function. These studies collectively demonstrated that the addition of IFN-γ to in vitro cultures of developing human or mouse mast cells reduced proliferation and development (20, 21, 22, 23), and that IFN-γ could reduce IgE-mediated degranulation and adherence to fibronectin (24, 25, 26, 27, 28). By comparison, the laboratory of Metcalfe and colleagues (29) has recently shown that IFN-γ increases FcγRI expression and signaling on human mast cells, a finding that could be critical to the role of mast cells in both Th1 and Th2 responses. It should be noted that a previous study demonstrated that IFN-γ enhanced mast cell expansion from splenocytes (37). This work used a very different starting population, and used IFN-γ in the absence of IL-3 and SCF. Hence it is difficult to compare the two studies. Because the spleen may be a source of more mature mast cell precursors, it is interesting to note that we did not find IFN-γ to inhibit the survival of differentiated mast cells (Fig. 1). On balance, it appears that the inhibitory effects of IFN-γ are exerted largely at the level of the mast cell precursor.

Despite the consensus that IFN-γ inhibits the proliferation of mast cell progenitors, the mechanism by which this occurs and its physiological significance have remained unclear. IFN-γ treatment is an area of clinical importance because it may be useful in conditions in which mast cell function or survival is pathological. In support of this, IFN-γ treatment of a patient with systemic mastocytosis reduced their eosinophilia, a symptom associated with mast cell activation. This and other symptoms returned coincidentally with the patient’s development of anti-IFN-γ Abs (17). An understanding of IFN-γ actions and their physiological relevance could further the intervention strategies for mast cell-associated diseases, which appears to be a growing list.

Our study shows that IFN-γ elicits the apoptotic death of developing mast cells. Interestingly, IFN-γ did not affect the survival of differentiated mast cells, nor did it alter mast cell surface Ag expression or granulation. Instead, IFN-γ appeared to induce apoptosis of developing mast cells near the time of lineage commitment, through a pathway involving mitochondrial damage and caspase activation, as demonstrated by loss of mitochondrial membrane potential coincidental with caspase-3 activation. These data were corroborated by the complete blockade of IFN-γ-mediated death under conditions in which Bax was deleted or Bcl-2 was overexpressed. Because Bax and Bcl-2 are known to control mitochondrial membrane potential (38), blocking IFN-γ-mediated apoptosis by altering expression of these proteins confirmed the essential role of the mitochondrion. Further, loss of p53, a known regulator of Bax expression (39, 40, 41, 42), also prevented IFN-γ-mediated. The effects of IFN-γ also showed a strict dependence on Stat1 expression, a key signal transduction protein in the IFN-γ receptor pathway. We are now investigating the mechanisms by which IFN-γ may use Stat1 to regulate p53 activation and how this contributes to the death of developing mast cells.

The physiological relevance of IFN-γ-mediated suppression of mast cell survival is an area of clinical importance, because IFN-γ therapy is an option in mast cell-associated diseases. Moreover, the etiology of conditions in which mast cell activation is chronic might be partly explained by the loss of IFN-γ-mediated mast cell homeostasis. However, there is little evidence to support suppression of mast cell survival in vivo via IFN-γ. To address this issue, we used SOCS-1-deficient mice. These animals die ∼3 wk postpartum from severe inflammatory disease that results in multiorgan failure (15). Because crossing these mice to an IFN-γ-deficient background greatly improves survival, much of the effects of SOCS-1 deficiency have been ascribed to IFN-γ overproduction and hyperresponsiveness (32, 33). In fact, we found that SOCS-1 KO mast cells could not be cultured from bone marrow progenitors, as they succumbed to apoptosis during the 21-day culture period. This effect required IFN-γ because SOCS-1/IFN-γ DKO cells survived normally in culture.

More importantly, these in vitro findings were recapitulated in vivo. SOCS-1 KO mice demonstrated a 67% reduction in peritoneal mast cell numbers. This effect was completely reversed in SOCS-1/IFN-γ DKO mice, whose mast cell numbers were greater than age-matched wild-type controls. By contrast, loss of IFN-γ alone did not affect mast cell numbers. This response implies that IFN-γ is an important feature of mast cell homeostasis in situations in which IFN-γ signaling is unbalanced. Loss of SOCS-1 expression has been shown to increase IFN-γ production (32) and would prevent the normal negative feedback signals exerted by SOCS-1 on the IFN-γ receptor pathway (32, 33). Therefore, SOCS-1 KO mice would have enhanced IFN-γ signaling that is ablated in SOCS-1/IFN-γ DKO mice. In contrast, loss of basal IFN-γ production in healthy subjects does not appear to alter mast cell homeostasis. The increase in mast cell numbers when both SOCS-1 and IFN-γ are deleted could be explained as the loss of two negative regulators. SOCS-1 is known to inhibit the signaling of many growth factor receptors (43), thus in the absence of IFN-γ, SOCS-1 deficiency may result in growth factor hyperresponsiveness and an increase in mast cell proliferation. How IFN-γ and SOCS proteins contribute to mast cell homeostasis under conditions of infection and inflammation is an area that warrants further study.

These results provide a partial mechanism for IFN-γ-mediated death of developing mast cells, with emphasis on a mitochondrial pathway. We also demonstrate that dysregulated IFN-γ signaling results in mast cell deficiency both in vitro and in vivo. Our data support the contention that IFN-γ contributes to mast cell homeostasis, and may be a useful intervention in mast cell-associated diseases. We also emphasize the need to consider defective IFN-γ signaling as a possible contributing factor to the etiology of diseases in which mast cell homeostasis is lost. Given the expanding role of the mast cell, understanding control of its survival and function may impact a broad array of immune disorders.

The authors have no financial conflict of interest.