Apoptosis is an important cellular mechanism for controlling cell viability and proliferation. With respect to eosinophils, cytokines prolong their survival, whereas corticosteroids reduce their survival in vitro. CD30, a member of the TNFR family, is expressed on the surface of many cell types, including Hodgkin’s lymphoma cells. CD30 is capable of inducing apoptosis after Ab treatment in some cell lines. To determine whether this surface structure is involved in apoptosis of human eosinophils, we examined its expression and the effect of anti-CD30 Ab treatment on the viability of eosinophils. Purified human eosinophils expressed low, but consistently detectable, levels of CD30. Immobilized, but not soluble, forms of anti-CD30 Abs (HRS-4 and Ber-H8) or recombinant mouse CD30 ligand exhibited an extremely rapid and intense survival-reducing effect on the eosinophils in the presence of exogenous IL-5; this effect was both concentration and time dependent. Furthermore, high concentrations of IL-5 could not reverse the reduced survival rates. After treatment with anti-CD30 Ab, gel electrophoresis of DNA extracted from the eosinophils demonstrated changes consistent with apoptosis. The immobilized F(ab′)2 of the anti-CD30 Ab failed to induce eosinophil apoptosis. The addition of anti-CD18 Ab also completely abrogated the induction of eosinophil apoptosis. Further examination using specific signal transduction inhibitors suggested the involvement of p38, mitogen-activated protein kinase kinase 1/2, and specific tyrosine kinase, but not NF-κB, in the induction of CD30-mediated eosinophil apoptosis. These data demonstrate that CD30 can modify eosinophil survival by causing an extremely rapid and intense induction of apoptosis through a tightly regulated intracellular signaling pathway.
Eosinophils are known to be important effector cells in helminth infections and allergic diseases (1, 2, 3). Several cytokines, including GM-CSF, IL-3, and IL-5, are reported to regulate the development of eosinophils from hemopoietic stem cells (4, 5) and to support the survival and activation of these cells in vitro (6, 7, 8, 9). When mature peripheral blood eosinophils are cultured in the absence of these cytokines, they undergo apoptosis (10), because eosinophils at least in part substantially lack the expression of Bcl-2, an antiapoptotic protein (11). Not only deprivation of eosinophil-activating cytokines, but also some cytokines, such as IL-4 (12) and TGF-β (13), or ligation of cell surface molecules, including Fas (CD95) (14) and CD69 (15), are capable of inducing apoptosis in human eosinophils. Administration of mAbs against IL-5 (16), or against Fas (17) are reported to be of potent therapeutic value in animal models of allergic asthma.
CD30, a member of TNFR family, was originally identified as Ki-1, an Ag expressed on Reed-Sternberg cells in Hodgkin’s lymphomas and other non-Hodgkin’s lymphomas, particularly diffuse, large cell lymphoma and immunoblastic lymphoma (18). TNFR family members, including two receptors for TNF, the nerve growth factor receptor, Fas, CD27, CD40, OX40, 4-1BB, TRAIL receptors 1, 2, 3, and 4, as well as several soluble receptors of mammalian and viral origin, participate in cellular activation, induction of survival; and/or apoptosis (reviewed in Refs. 19 and 20).
To date, numerous studies have tried to elucidate the physiological and pathological roles of CD30 molecules, but the actual roles of CD30 molecules in vivo remain obscure (21). Experiments using CD30-deficient mice have indicated that this molecule is involved in the deletion of autoreactive thymocytes (22, 23), whereas a recent report denied this function (24). TNFR family members can transduce signals when surface molecules are appropriately cross-linked. However, the effects of ligation of cell surface CD30 are somewhat controversial (25, 26, 27); CD30 can induce a wide range of effects from proliferation to apoptosis in different CD30-positive cells or cell lines. In addition, such differences are reported not to be due to CD30 heterogeneity, but to the cell type (25).
To date, a wide variety of cells and cell lines have been shown to express CD30. These include activated T cells, predominantly Th2 (28, 29, 30), T cytotoxic 2 (31), CD8 (32), γδ (33), NK cells (34), and lymphocytes infected with HIV (35) or EBV (36).
The extracellular domain of CD30 can be cleaved and shed as a result of cellular activation (37). Soluble levels of CD30 are elevated and are associated with the prognosis in Hodgkin’s disease and AIDS (38). Elevated levels were also reported in patients with adult T cell leukemia and some of those with non-Hodgkin’s lymphoma, rheumatoid arthritis (39), chronic renal failure with primary growth retardation (40), atopic dermatitis (41, 42, 43), bullous pemphigoid (44), polymyositis/dermatomyositis (45), and helminth infection (46). Interestingly, some of these diseases are associated with eosinophilia (47, 48, 49, 50), and an especially positive correlation between the levels of soluble CD30 (sCD30) and eosinophilia was found in Hodgkin’s disease (51). A positive correlation between the level of sCD30 and the serum eosinophil cationic protein level was also found in patients with atopic dermatitis (41, 42).
We hypothesized that human eosinophils might express CD30, and, if present, this molecule may play a role in regulating eosinophil apoptosis, partially because inhibition of CD30-mediated signals by sCD30 (52) seems to be associated with eosinophilia.
In the present study we demonstrate that CD30 is consistently expressed on human eosinophils, although at low levels. We found that culture of eosinophils with certain immobilized anti-CD30 mAbs or recombinant mouse CD30 ligand induces extremely rapid and intense eosinophil apoptosis, even in the presence of exogenous IL-5.
Materials and Methods
The following anti-CD30 mAbs were used: clone AC10, IgG1 (Ancel, Bayport, MN); clone HRS-4, IgG1 (Immunotech, Marseilles, France); clone 1G12, IgG2a (Yealem, Rome, Italy); clone Ki-1, IgG3 (Yealem, Rome, Italy); clone Ber-H2, IgG1 (Daco, Glostrup, Denmark); and clone Ber-H8, IgG1 (BD PharMingen, San Diego, CA). Anti-Fas Ab (clone CH-11, IgM; Medical Biological Laboratories, Gunma, Japan) and an isotype-matched irrelevant negative control Abs (IgG1, MOPC-21 (Sigma-Aldrich, St. Louis, MO); IgG2a, UPC-10 (Sigma-Aldrich); IgG3, FLOPC-21 (Sigma-Aldrich)) were obtained from the indicated sources. mAbs were dialyzed against PBS to remove azide before use if necessary. The F(ab′)2 of an anti-CD30 mAb (Ber-H8) and control IgG1 were created using an F(ab′)2 preparation kit containing immobilized pepsin (Pierce, Rockford, IL). The affinity of the F(ab′)2 of anti-CD30 mAb was confirmed using a CD30-expressing cell line (AML-14). F(ab′)2 of an anti-CD32 mAb (IV.3-Fab) were purchased from Medarex (Annandale, NJ). Anti-CD18 blocking mAb (L130) was purchased from Immunotech. Mouse recombinant CD30 ligand was purchased from R&D Systems (Minneapolis, MN).
Eosinophil and neutrophil purification
Eosinophils were purified from peripheral blood of normal or slightly allergic donors (n = 21) using Percoll (Amersham Pharmacia Biotech, Uppsala, Sweden) density gradient centrifugation and negative selection with anti-CD16 Ab-coated immunomagnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany), as previously described (53). Both the purity and the viability of eosinophils, based on light microscopic examination of cytocentrifugation preparations stained with Diff-Quik (American Scientific Products, McGraw Park, IL) or trypan blue (Sigma-Aldrich) dye exclusion, respectively, always exceeded 99%. In some experiments peripheral blood neutrophils were purified from normal donors by Percoll density gradient centrifugation and hypotonic lysis of erythrocytes. The purity and viability of neutrophils always exceeded 98%. In some experiments cord blood-derived cultured eosinophils were used, as described previously (54). The purity always exceeded 90%. The contaminating cells were basophils or macrophages.
Expression of CD30 on eosinophil surface
The expression of CD30 on eosinophils was examined using indirect immunofluorescence and flow cytometry, as previously described (14). Freshly isolated eosinophils were labeled (30 min, 4°C) in Dulbecco’s PBS containing 0.1% BSA (PBS-BSA; Sigma-Aldrich) and 4 mg/ml human IgG (Sigma-Aldrich) with a saturating concentration of anti-CD30 Ab or an equivalent concentration of irrelevant isotype-matched control Ab. Cells were washed and then incubated (30 min, 4°C in PBS-BSA) with appropriate dilutions of FITC-conjugated F(ab′)2 goat anti-mouse IgG (H and L chains) Ab (BD Biosciences, Mountain View, CA). After fixation in 1% paraformaldehyde in PBS, at least 5000 cells were evaluated using a flow cytometer (FACScan analyzer; BD Biosciences).
CD30 mRNA expression
Total RNA was isolated from an Isogen (Nippon Gene, Osaka, Japan) solution in which harvested eosinophils were dissolved. The RNA was DNase-treated, converted to cDNA by RT using a kit from Invitrogen (San Diego, CA), and subjected to PCR amplification. The primer sequences for CD30 were as follows: 5′-GCC CAG GAT CAA GTC ACT CAT-3′ for the 5′ primer, and 5′-TAC ACG TCT GAA GGC CCT AGG-3′ for the 3′ primer, spanning a fragment of 501 bp. The primer sequences for GAPDH were as follows: 5′-GTC TTC ACC ACC ATG GAG AAG GCT-3′ for the 5′ primer, and 5′-CAT GCC AGT GAG CTT CCC GTT CA-3′ for the 3′ primer, spanning a fragment of 393 bp. PCR amplification was performed using a thermal cycler (GeneAmp PCR System 9700; Applied Biosystems, Foster City, CA) under the following conditions: one initial denaturation cycle for 1 min at 94°C; 37 amplification cycles for 1 min at 94°C, 1 min at 55°C, and 2 min at 72°C; and a final extension phase consisting of one cycle of 10 min at 72°C. The PCR products were visualized on a 0.8% agarose gel (Life Technologies, Gaithersburg, MD) containing 0.05 μg/ml ethidium bromide (Sigma-Aldrich).
Eosinophil culture and evaluation of postculture cell apoptosis/viability
Purified eosinophils were suspended at a cell density of 1 × 106/ml in IMEM (Life Technologies, Grand Island, NY) supplemented with 10% FCS (HyClone Laboratories, Logan, UT) and up to 10 ng/ml human recombinant IL-5 (R&D Systems). For culture experiments, 24-well microtiter plates (Costar, Cambridge, MA) were coated overnight at 4°C with 200-μl aliquots of various concentrations of anti-CD30 Ab (1–1000 μg/ml final concentration) or equivalent concentrations of isotype-matched control Ab. In some experiments 10 μg/ml recombinant mouse CD30 ligand was used to coat the plates. The wells were then blocked by incubation with 2-ml aliquots of PBS containing 1% human serum albumin (heat denatured at 65°C for 1 h) for at least 2 h at room temperature to reduce nonspecific adherence to the plastic. The wells were washed twice with prewarmed IMDM before use. In some experiments eosinophils were pretreated with the F(ab′)2 of an anti-CD32 mAb or control IgG1 mAb (10 μg/ml) for 30 min on ice before culturing in anti-CD30 mAb-coated wells to block signaling via Fc of the immobilized mAb.
To determine the signal transduction pathway upon CD30 activation, a p38 inhibitor, SB203580 (0.01–10 μM; Calbiochem, San Diego, CA); a mitogen-activated protein kinase kinase 1 (MEK1)2-specific inhibitor, PD98059 (0.01–10 μM; Calbiochem); a MEK1/2 inhibitor, U0126 (0.01–10 μM; Calbiochem); tyrosine kinase inhibitors, genistein (0.1–100 μM; Sigma-Aldrich) or tyrphostin B46 (100 μM, Sigma-Aldrich); or three inhibitors of NF-κB activation, MG-132 (100 μM; Calbiochem), gliotoxin (1 μg/ml; Sigma-Aldrich), or SN-50 (100 μg/ml; Calbiochem), were simultaneously added to the wells with the cells and allowed to remain throughout the entire survival assay.
To determine epitopes recognized by the CD30 mAbs, a soluble form of other anti-CD30 mAbs or control IgG1 (10 μg/ml) was added to the wells previously coated with 10 μg/ml anti-CD30 mAb (HRS-4 or Ber-H8), simultaneously with cells and allowed to remain throughout the entire survival assay.
Cell viability was examined with an apoptosis detection kit (MEBCYTO-Apoptosis Kit; Medical Biological Laboratories, Nagoya, Japan) using FITC-conjugated annexin V (55) and propidium iodide (PI) according to the manufacturer’s instructions. Briefly, cultured eosinophils were harvested at various time points by gentle pipetting and were washed once with PBS. The harvested eosinophils were suspended in Ca2+-containing buffer and reacted with FITC-conjugated annexin V for 15 min at room temperature. One minute before analyzing by flow cytometry, the eosinophils were also stained with 10 μg/ml PI. FITC-stained cells (annexin V-positive) and PI-stained cells were measured by flow cytometry, as described above.
DNA fragmentation assay
DNA fragmentation was determined with an Apoptosis Ladder Detection Kit (Wako Pure Chemical Industries, Osaka, Japan) according to the manufacturer’s instructions. Briefly, 5 × 105 eosinophils were lysed with DNA extraction solution containing proteinase and RNase. DNA was extracted with isopropanol (Nakalai Tesque) and 70% ethanol (Sigma-Aldrich). The extracted DNA was dissolved in 10 μl of TE buffer (10 mM Tris-Cl and 1 mM EDTA, pH 8.0; Sigma-Aldrich), analyzed on 0.8% agarose gel (Life Technologies), and visualized with CYBR Green I.
All data are presented as the mean ± SEM. Differences between groups were analyzed using Mann-Whitney’s U test or paired t test and were considered significant at p < 0.05.
Expression of CD30 on eosinophils
Indirect immunofluorescence and flow cytometry were used to examine CD30 expression on eosinophils from eight different donors. Freshly isolated peripheral blood eosinophils were found to express low, but consistently detectable, amounts of CD30 (Fig. 1,a). For all eight donors, the average mean fluorescence intensities (MFIs) for four anti-CD30 mAbs (clones AC10, HRS-4, Ber-H8, and Ber-H2) and an irrelevant IgG1 mAb were 8.4 ± 1.2, 3.9 ± 0.6, 4.6 ± 1.0, and 3.1 ± 0.5 vs 2.5 ± 0.2, respectively (p < 0.01; Fig. 1 b). In contrast to eosinophils, neutrophils did not express CD30 (MFI of Ber-H8, 2.2 ± 0.8; MFI of Ber-H2, 2.5 ± 0.3; p > 0.1)
Expression of CD30 mRNA in eosinophils
RT-PCR of the total RNA extracted from freshly isolated eosinophils consistently detected mRNA for CD30 (Fig. 2). mRNA for CD30 was also detected in human cord blood-derived cultured eosinophils.
Effect of immobilized anti-CD30 mAb and mouse recombinant CD30 ligand on eosinophil viability
In initial experiments, eosinophils were cultured in the presence of 1 ng/ml IL-5 in 24-well plates previously coated with various anti-CD30 mAbs, and the eosinophil viability was determined after 4 h. As shown in Fig. 3,a, in the presence of irrelevant control IgG1 mAb, annexin V-positive eosinophils were 5.25 ± 0.9%. Culture of eosinophils in the presence of two immobilized anti-CD30 mAbs, HRS-4 and Ber-H8, resulted in a concentration-dependent reduction in eosinophil viability that was significant at 4 h of culture, whereas four other tested mAbs (Ber-H2, AC10, 1G12, and Ki-1) or control IgG1, IgG2a, and IgG3 mAbs showed no significant effect (Fig. 3,a; 5.8 ± 1.2 and 5.2 ± 0.5% annexin V-positive eosinophils for control IgG2a and IgG3, respectively). The survival-reducing effect of the two anti-CD30 mAbs at 4 h of culture was significant when the concentration of the mAb used for coating the plate was ≥0.01 μg/ml for HRS-4 and ≥1.0 μg/ml for Ber-H8 (Fig. 3,b). When the plate was coated with 10 μg/ml mAbs (HRS-4 and Ber-H8), the survival-reducing effect was significant even at 1 h of culture, and it was significantly greater at all subsequent time points compared with those at the corresponding earlier time points (p < 0.005; Fig. 4). In addition, exogenously added high concentrations of IL-5 (up to 100 ng/ml) or pretreatment of eosinophils with 10 ng/ml IL-5 for 24 h failed to reverse the reduction in viability caused by the immobilized anti-CD30 mAbs (Ber-H8; Fig. 5 and data not shown). Culture of eosinophils with a soluble form of anti-CD30 mAb (HRS-4 or Ber-H8) did not alter eosinophil survival at 4 h (annexin V-positive eosinophils, 7.8 ± 0.6, 8.5 ± 0.8, 8.2 ± 1.3, and 8.0 ± 0.9% for HRS-4, Ber-H8, Ber-H2, and IgG1 control, respectively; p > 0.1).
To determine epitopes recognized by the CD30 mAbs, blocking experiments were performed using immobilized 10 μg/ml anti-CD30 (HRS-4 or Ber-H8) and a soluble form of other anti-CD30 mAbs or control IgG1. As a result, none of the nonsurvival-reducing mAbs (AC10, Ber-H2, or control IgG1) or the other survival-reducing mAb (HRS-4 or Ber-H8) could block the apoptosis-inducing effect of the immobilized survival-reducing mAbs, whereas the soluble form of the same mAb significantly reversed eosinophil survival (data not shown).
Culture of eosinophils in plates coated with recombinant mouse CD30 ligand significantly reduced eosinophil survival (27.5 ± 3.6 and 7.3 ± 0.4% for mouse CD30 ligand and 0.1% BSA, respectively; n = 5; p < 0.001).
In contrast to eosinophils, the viability of peripheral blood neutrophils was not altered after 4 h of culture in plates previously coated with HRS-4 or Ber-H8 compared with irrelevant IgG1 control or Ber-H2 (annexin V-positive eosinophils, 22.2 ± 0.4, 19.2 ± 0.8, 21.1 ± 0.9, and 21.8 ± 0.2% for HRS-4, Ber-H8, Ber-H2, and IgG1 control, respectively; n = 3; p > 0.1).
Examination by annexin V and PI staining, and determination of DNA fragmentation
To determine whether the observed reductions in eosinophil viability were the result of apoptosis, several criteria typically used to distinguish apoptosis from necrosis (56) were examined. Eosinophils cultured in IL-5 with anti-CD30 mAb or control IgG1 were harvested after 4 h and then stained with FITC-conjugated annexin V and PI. As a result, most eosinophils cultured in the presence of both IL-5 and anti-CD30 mAbs (HRS-4 and Ber-H8) displayed significant annexin V binding without being stained with PI. In contrast, these changes were rarely observed in cells cultured in the presence of control IgG1 or anti-CD30 clone Ber-H2 (Fig. 6,a). Another characteristic of apoptosis, internucleosomal DNA degradation (10) (DNA laddering) was detected by agarose gel electrophoresis of DNA extracted from eosinophils cultured with 1 ng/ml IL-5 and anti-CD30 mAb (HRS-4 and Ber-H8), but not from eosinophils cultured with IL-5 and irrelevant control IgG1 mAb or anti-CD30 mAb (Ber-H2; Fig. 6 b).
Involvement of the Fc portion of anti-CD30 mAb and adhesion of eosinophils in CD30-mediated eosinophil apoptosis
To test the involvement of the Fc portion of anti-CD30 mAbs in the apoptosis-inducing effect of the mAbs, an F(ab′)2 of anti-CD30 mAb (Ber-H8) was prepared using immobilized pepsin. When eosinophils were cultured in the plates previously coated with the F(ab′)2 of anti-CD30 mAb (Ber-H8), the cells did not undergo apoptosis, although the binding affinity of the F(ab′)2 was equivalent to that of the whole mAb (data not shown). In addition, pretreatment of eosinophils with the F(ab′)2 of anti-CD32 blocking mAb (IV.3-Fab) or the addition of 10 μg/ml anti-CD32 blocking mAb (KB61; Santa Cruz Biotechnology, Santa Cruz, CA) to the culture wells almost completely abrogated the apoptosis-inducing effect of the immobilized anti-CD30 mAb (data not shown).
To examine whether adhesion of the eosinophils to the plate is required for the apoptosis-inducing effect of the anti-CD30 mAb, the effect of β2 integrin blockade was tested. The addition of 10 μg/ml anti-CD18 blocking mAb completely abrogated the apoptosis of eosinophils cultured in plates previously coated with anti-CD30 mAb (Fig. 7). Interestingly, anti-CD18 mAb also inhibited mouse CD30 ligand-induced eosinophil apoptosis (data not shown)
Examination of the signal transduction pathway by which anti-CD30 mAb induces eosinophil apoptosis
To examine the signal transduction pathways through which anti-CD30 mAb-induces eosinophil apoptosis, several inhibitors of the intracellular signal transduction pathway were added to the wells simultaneously with the cells and allowed to remain throughout the entire survival assay. The results showed that the p38 inhibitor SB203580, the MEK1-specific inhibitor PD98059, and the MEK1/2 inhibitor U0126 partially inhibited the anti-CD30 mAb-induced eosinophil apoptosis in a dose-dependent manner (Fig. 8,a). In addition, the tyrosine kinase inhibitor genistein, but not tyrphostin B46, also inhibited anti-CD30 mAb-induced eosinophil apoptosis in a dose-dependent manner (Fig. 8 b). In contrast, all three inhibitors for NF-κB activation, MG-132, gliotoxin, and SN-50, did not alter the anti-CD30 mAb-induced eosinophil apoptosis (data not shown).
In the present study human peripheral blood eosinophils were found by indirect immunofluorescence and flow cytometry to express low, but significantly detectable, surface levels of CD30. In addition, CD30 mRNA was consistently detectable using RT-PCR in human peripheral blood eosinophils and cord blood-derived cultured eosinophils.
Culture with two immobilized mAbs (clone HRS-4 and Ber-H8) of six mAbs tested reduced eosinophil viability dramatically in a concentration- and time-dependent manner, even in the presence of exogenously added IL-5. Culture with a soluble form of anti-CD30 mAbs (HRS-4 or Ber-H8) failed to reduce eosinophil survival, suggesting that cross-linking is required for the survival-reducing effect of the anti-CD30 treatment. Interestingly, three mAb clones, Ber-H2, Ber-H8, and HRS-4, were reported to recognize the CD30-CD30 ligand binding site because they inhibit CD30 ligand binding to CD30 molecules (57). However, our observation that immobilized Ber-H2 failed to reduce eosinophil survival and that the exogenously added soluble form of Ber-H2 failed to inhibit the survival-reducing effect induced by immobilized Ber-H8 or HRS-4 suggests that the recognition sites for Ber-H2 and Ber-H8 and/or HRS-4 must be distinct and do not compete with each other. At low concentrations of mAbs for coating the plates, Ber-H8 seems to be approximately twice as potent as HRS-4 in reducing eosinophil survival. According to the data from flow cytometry, the survival-reducing effect of anti-CD30 mAbs was not simply associated with the affinity of the mAbs tested. Surprisingly, the effect was extremely rapid and intense. After only 4 h of culture, >50% of eosinophils exhibited positive staining by annexin V when cultured with immobilized anti-CD30 mAbs (Ber-H8 or HRS-4). Such an effect is never seen with eosinophils exposed to other conditions, including cytokine deprivation (10, 58) and anti-Fas mAb treatment (14).
Positive findings for annexin V binding by flow cytometry along with evidence of DNA fragmentation seen by gel electrophoresis were consistent with the hypothesis that the immobilized anti-CD30 mAb-induced decline in eosinophil survival was the result of apoptosis. Thus, CD30 is a functional molecule on human peripheral blood eosinophils that can regulate eosinophil apoptosis in vitro.
Various cytokines are crucial factors in granulocyte development and survival both in vitro and in vivo. For eosinophils, IL-3, IL-5, GM-CSF, and perhaps other cytokines serve this role (6, 7, 8). When mature peripheral blood eosinophils are cultured in the absence of these cytokines, they undergo apoptosis (10, 58). Corticosteroids, in addition to inhibiting the production of these cytokines, can directly reduce eosinophil survival, although this direct effect can be overcome by increasing the concentration of eosinophil-activating cytokines (58, 59, 60, 61, 62). This effect differs from the apoptosis seen with anti-CD30 mAbs, which IL-5 could not completely reverse (Fig. 5) even at the highest concentrations of IL-5 tested (100 ng/ml). This suggests that the effect of anti-CD30 mAbs was not simply to antagonize the effects of IL-5, and that the reduction in IL-5-induced eosinophil survival was at least in part mediated through pathways independent of those used by IL-5 to prolong survival.
Several members of the TNFR superfamily can induce cell death. For TNFR1, Fas/APO-1, DR3, DR6, TRAIL-R1, and TRAIL-R2, a conserved death domain in the intracellular region couples these receptors to activation of death effector molecules such as caspases (63). However, it is not yet fully understood how TNFR superfamily members lacking a death domain, such as TNFR2, CD40, lymphotoxin-βR, CD27, and CD30, execute their death-inducing capability.
With respect to CD30 molecules, CD30 signals are transmitted through adapter proteins of the TNFR-associated factor (TRAF) family, including TRAF1, -2, and -5, resulting in the activation of transcription factors such as NF-κB and Jun N-terminal kinase (64, 65, 66, 67, 68). TRAF2 is thought to be responsible for NF-κB activation and for the antiapoptotic effect mediated by CD30 (66, 68). In contrast, in certain cells, CD30 activation degrades the signaling intermediate TRAF2 (68) and disrupts the NF-κB survival pathway. It is thought that the ability of CD30 to recruit TRAFs, activate NF-κB, or degrade TRAFs to compromise NF-κB activation may depend on the specific cell type examined and could explain previous contradictory observations that CD30 mediates proapoptotic as well as antiapoptotic signals (26). It is possible that in eosinophils the latter pathway associated with degradation of TRAF2 plays a crucial role in the induction of apoptosis after ligation of CD30. However, the extremely rapid and intense induction of apoptosis in human eosinophils demonstrated in the present study could not be observed even in cell lines known to be susceptible to the induction of apoptosis by CD30. Such a difference in the magnitude of the susceptibility to the induction of apoptosis might be due to a difference in the level of Bcl-2, as it was previously reported that the effect of CD30 is regulated by the expression of Bcl-2 in the target cells (69). It should be further examined whether TRAF2 is degraded upon activation through CD30 in eosinophils.
Another possible explanation for the induction of apoptosis via CD30 is that CD30 signals induce the expression of membrane-bound TNF, thereby activating a TNF-autocrine pathway (70). However, previous reports (71, 72) as well as our unpublished data using exogenous TNF-α (up to 100 ng/ml) or exogenous TNF-α-neutralizing mAb (10 μg/ml) clearly demonstrate that even in the presence of a high concentration of TNF-α or anti-TNF-α mAb, eosinophil viability was not altered, as we have observed in the present study when using anti-CD30 mAbs (data not shown). Thus, membrane-bound TNF is unlikely to be involved in anti-CD30 mAb-mediated eosinophil apoptosis.
The facts that the immobilized F(ab′)2 of anti-CD30 mAb (Ber-H8) failed to induce eosinophil apoptosis and that anti-CD32 blocking mAb abrogated anti-CD30 mAb-mediated eosinophil apoptosis suggest that an Fc-mediated signal in addition to CD30 ligation is required for this effect. However, the mechanism by which the immobilized mCD30 ligand mediates human eosinophil apoptosis remains unclear. Interestingly, anti-β2 integrin mAb (anti-CD18 mAb) also almost completely inhibited eosinophil apoptosis induced by the immobilized anti-CD30 mAb and the mCD30 ligand. The blockade of β2 integrin is known to impair cell structural changes and diminish cell activation in eosinophils (73).
These results strongly suggest that an additional signal may critically regulate eosinophil apoptosis induced through cell surface CD30. The requirement of cross-linking of CD30 molecules and the fact that the recognized epitope is critical for the apoptosis-inducing effect of anti-CD30 mAbs strongly suggest that CD30 on eosinophils are functional, and the downstream signals after CD30 ligation seem to be tightly regulated.
With respect to the signal transduction pathways, we examined the effects of several inhibitors specific to intracellular signaling molecules on anti-CD30 mAb-mediated eosinophil apoptosis. The results clearly showed the involvement of p38 and MEK1/2 in CD30-induced eosinophil apoptosis. The inhibitory effect of U0126 seems more prominent than that of PD98059, suggesting that both MEK1 and -2 are involved in the signal transduction pathway upon CD30 activation. A tyrosine kinase inhibitor, genistein, but not tyrphostin B46, also showed a significant inhibitory effect. Genistein is reported to block β2 integrin-mediated eosinophil activation (74, 75), whereas tyrphostin B46 blocks β1 integrin-mediated eosinophil adhesion (76). These results imply the involvement of some specific tyrosine kinase in CD30-mediated apoptosis. However, whether genistein blocks the CD30-mediated signal, the β2 integrin-mediated signal, or both these signals remains uncertain. Interestingly, none of the three inhibitors of NF-κB activation (MG-132, gliotoxin, and SN-50) inhibited CD30-mediated apoptosis, even though NF-κB is known to be a major signaling molecule in the TNFR family of molecules, which includes CD30 (25). The precise mechanism of the CD30-mediated apoptosis of eosinophils should be further studied, as this mechanism may be a potent therapeutic target for the induction of eosinophil apoptosis.
The fact that human peripheral blood eosinophils express functional CD30 suggests that this molecule may be involved in allergic inflammatory diseases or diseases with eosinophilia where eosinophil activation and survival may be altered. If so, blockade of CD30-CD30 ligand signals might reflect the prolongation of eosinophil survival. Clinical observations that the level of sCD30 correlates positively with eosinophilia in Hodgkin’s disease (51) and the serum eosinophil cationic protein level in patients with atopic dermatitis (41, 42) are quite plausible if sCD30 can inhibit CD30-CD30 ligand signals in vivo (52). Needless to say, secretion of Th2-type cytokines might also participate the eosinophilia seen in these patients (77, 78). However, our observation that CD30 might regulate eosinophil survival is novel and of interest in terms of the therapeutic value of CD30 signals for eosinophil-associated disorders, particularly because such a survival-reducing effect was seen only in eosinophils, but not in neutrophils. The facts that immobilized recombinant mouse CD30 ligand also showed a significant survival-reducing effect on eosinophils, and the level of sCD30 is positively correlated with eosinophilia in Hodgkin’s disease (51) imply that CD30 on eosinophils might be involved in eosinophil apoptosis in vivo. However, the requirement for an additional signal for the induction of eosinophil apoptosis makes it difficult to conclude that this pathway is involved in eosinophil regulation in vivo. Further study is required to determine the physiological relevance of our observation.
Taken together, these data imply that interactions between CD30 and CD30 ligand in vivo may play an important role in mediating the apoptosis of eosinophils. They also indicate that the apoptotic effect of CD30 signals in eosinophils is extremely rapid and intense compared with other culture conditions and might have potent therapeutic value for diseases involving pathological eosinophil activation, particularly because such a survival-reducing effect was seen only in eosinophils, not in neutrophils.
We express our gratitude to Hisashi Tomita and Keisuke Yuki (National Research Institute for Child Health and Development, Tokyo, Japan) for their excellent technical assistance.
Abbreviations used in this paper: MEK, mitogen-activated protein kinase kinase; MFI, mean fluorescence intensity; PI, propidium iodide; sCD30, soluble CD30; TRAF, TNFR-associated factor.