Activation of Interferon-γ Inducing Factor Mediated by Interleukin-1β Converting Enzyme

See allHide authors and affiliations

Science  10 Jan 1997:
Vol. 275, Issue 5297, pp. 206-209
DOI: 10.1126/science.275.5297.206


The interleukin-1β (IL-1β) converting enzyme (ICE) processes the inactive IL-1β precursor to the proinflammatory cytokine. ICE was also shown to cleave the precursor of interferon-γ inducing factor (IGIF) at the authentic processing site with high efficiency, thereby activating IGIF and facilitating its export. Lipopolysaccharide-activated ICE-deficient (ICE−/−) Kupffer cells synthesized the IGIF precursor but failed to process it into the active form. Interferon-γ and IGIF were diminished in the sera of ICE−/− mice exposed to Propionibacterium acnes and lipopolysaccharide. The lack of multiple proinflammatory cytokines in ICE−/− mice may account for their protection from septic shock.

ICE is a member of the growing family of ICE-like cysteine proteases (caspases) with a substrate specificity for aspartate (1). ICE (caspase-1) was identified on the basis of its proteolytic activity for cleaving the inactive IL-1β precursor into the 17-kD mature cytokine (2). ICE-deficient mice are impaired in their production of mature IL-1β (3), which establishes the physiological role of ICE in the processing and export of IL-1β. In contrast to IL-1β-deficient mice (4), ICE−/− mice also have less IL-1α, tumor necrosis factor-α (TNF-α), and IL-6 and are resistant to septic shock induced by endotoxin (3), which suggests that ICE may have additional functions in the regulation of the immune system.

IGIF, an ∼18-kD polypeptide that stimulates production of interferon-γ (IFN-γ) by T cells (5), is synthesized as a polypeptide precursor (proIGIF) devoid of a conventional signal sequence (6). The precursor of IGIF is cleaved after Asp35 (6), which suggests that an aspartate-specific protease may be involved. Two families of proteases with substrate specificity for aspartate have been identified; these include the ICE family of cysteine proteases and granzyme B, a serine protease involved in cytotoxic lymphocyte-mediated cell killing and activation of ICE-like cysteine proteases (7, 8). Therefore, we investigated whether one or more of the ICE-family proteases or granzyme B may be involved in the processing of proIGIF and investigated the role that such a cleavage may have in the function of IGIF.

We first used transient coexpression in COS cells (9) to determine whether proIGIF could be processed by some of the known ICE-family proteases (Fig. 1A). Coexpression of proIGIF with ICE or its homolog TX (caspase-4) (10) resulted in the cleavage of proIGIF into a polypeptide similar in size to the naturally occurring 18-kD IGIF. Single point mutations of the catalytic cysteine residues that inactivate ICE and TX (11) blocked cleavage. Coexpression with CPP32 (caspase-3), a protease involved in programmed cell death (apoptosis) (12), resulted in the cleavage of proIGIF into a ∼14-kD polypeptide, whereas CMH-1 (caspase-7), a homolog of CPP32 (13), did not appreciably cleave proIGIF. Thus, ICE and TX could cleave proIGIF into a polypeptide similar to the naturally occurring IGIF.

Fig. 1.

ICE cleaves and activates proIGIF. (A) ICE cleaves proIGIF in COS cells. COS cells were transfected with an expression plasmid for proIGIF alone (lane 2) or in combination with the indicated expression plasmids encoding wild-type or inactive mutants of proteases of the ICE family (lanes 3 to 10). Cell lysates were prepared and analyzed for the presence of IGIF protein by immunoblotting with antiserum to IGIF (9). Lane 1 contained lysates from mock transfected cells. Mobilities of proIGIF and the 18-kD recombinant mature IGIF (6) are indicated on the right; molecular mass markers (in kilodaltons) are shown on the left. C, cysteine; S, serine in mutant designations. (B) ICE cleaves proIGIF in vitro. Purified recombinant (His)6-tagged proIGIF (2 μg) was incubated with the indicated proteases in the presence or absence of ICE or CPP32 inhibitors (lanes 3 to 9) (14). The cleavage products were analyzed by SDS-PAGE and Coomassie blue staining. Proteases and inhibitors used: lane 3, 1 nM ICE; lanes 4 and 5, 1 nM ICE with 10 nM Cbz-Val-Ala-Asp-[(2,6-dichlorobenzoyl)oxy]methyl ketone and 100 nM Ac-Tyr-Val-Ala-Asp-aldehyde, respectively; lanes 6 and 7, 15 nM CPP32 with and without 400 nM Ac-Asp-Glu-Val-Asp-aldehyde (12), respectively; lane 8, 100 nM CMH-1; lane 9, granzyme B (10 U/ml). Lanes 1 and 2 contained proIGIF and recombinant mature IGIF (6), respectively; lane M, molecular mass markers. NH2-terminal amino acid sequencing indicated that ICE cleaved proIGIF at the authentic processing site Asp35-Asn36, whereas CPP32 cleavage occurred at Asp69-Ile70. (C) ICE cleavage activates proIGIF. Uncleaved or ICE- or CPP32-cleaved products of proIGIF, or recombinant mature IGIF (rIGIF), were added to A. E7 cell cultures to a final concentration of 12 or 120 ng/ml; 18 hours later, IFN-γ in the culture medium was determined by ELISA (17). The data represent the average of three determinations.

We examined the cleavage of proIGIF by these proteases in vitro with the use of purified recombinant (His)6-tagged proIGIF as a substrate (14). ICE cleaved the 24-kD proIGIF into two polypeptides of ∼18 and ∼6 kD (Fig. 1B). The 18-kD polypeptide comigrated with recombinant mature IGIF upon SDS-polyacrylamide gel electrophoresis (PAGE) and contained the same amino acid residues (Asn-Phe-Gly-Arg-Leu) at its NH2-terminus as did the naturally occurring murine IGIF, indicating that ICE cleaved proIGIF at the authentic processing site (Asp35-Asn36) (6). This cleavage was specific with a catalytic efficiency (kcat/Km, where Km is the Michaelis constant) of 1.4 × 107 M−1 s−1 (Km = 0.6 ± 0.1 μM; kcat = 8.6 ± 0.3 s−1) (15) and was inhibited by the specific ICE inhibitors Ac-Tyr-Val-Ala-Asp-aldehyde (2) and Cbz-Val-Ala-Asp-[(2,6-dichlorobenzoyl)oxy]methyl ketone (16). Recombinant (His)6-tagged human proIGIF was also cleaved by ICE with a similar specificity. Although proIGIF had no detectable IFN-γ-inducing activity, ICE-cleaved proIGIF was active in inducing IFN-γ production in T helper type 1 (TH1) cells (Fig. 1C) (17). TX also cleaved proIGIF into polypeptides of similar size; however, its catalytic efficiency was about two orders of magnitude lower than that of ICE. In a manner consistent with the observation from the COS cell experiments, CPP32 cleaved proIGIF at a different site (Asp69-Ile70) and the resulting polypeptides had little IFN-γ-inducing activity, whereas CMH-1 and granzyme B did not cleave proIGIF. Thus, both in COS cells and in vitro, ICE can process the inactive IGIF precursor at the authentic maturation site to generate the biologically active form of IGIF.

IGIF is produced by activated Kupffer cells and macrophages in vivo and is exported from the cells upon stimulation by endotoxin (5, 6). We used the COS cell coexpression system to investigate whether the cleavage of proIGIF by ICE would facilitate the export of mature IGIF, as in the case of IL-1β (2). COS cells coexpressing proIGIF and ICE were labeled with [35S]methionine (18). COS cell lysates and conditioned medium were immunoprecipitated with an antiserum to IGIF that recognizes both the precursor and the mature form (6) (Fig. 2A). An 18-kD polypeptide corresponding to the mature IGIF was detected in the conditioned medium of COS cells coexpressing proIGIF and ICE, whereas COS cells expressing proIGIF alone or with the inactive ICE mutant exported only a very small amount of proIGIF. We estimated by PhosphorImager analysis that ∼10% of the mature IGIF was exported from transfected cells, whereas <1% of proIGIF was exported. We also measured the presence of IFN-γ-inducing activity in cell lysates and in the conditioned media of transfected cells (19). IFN-γ-inducing activity was detected in both cell lysates and conditioned medium of COS cells coexpressing proIGIF and ICE, but not those of cells expressing proIGIF or ICE alone (Fig. 2B). The relative amounts of mature IGIF in the medium and in cell lysates (19) indicated that the secreted IGIF was at least as active as the cytosolic mature IGIF. Thus, ICE cleavage of proIGIF can facilitate the export of mature and active IGIF from cells.

Fig. 2.

Processing of proIGIF by ICE facilitates the export of IGIF. (A) COS cells transfected with an expression plasmid for proIGIF alone (lanes 2 and 6) or in combination with an expression plasmid encoding wild-type (lanes 3 and 7) or inactive mutant (lanes 4 and 8) ICE were metabolically labeled with [35S]methionine. Cell lysates (left) and conditioned media (right) were immunoprecipitated with antiserum to IGIF (18). The immunoprecipitated proteins were analyzed by SDS-PAGE and fluorography. Fluorograms were exposed for 10 hours (left) and 3 days (right), respectively. Mobilities of proIGIF and the 18-kD mature IGIF are indicated on the right; molecular masses are shown on the left. Quantitative PhosphorImager analysis indicated that ∼10% of mature IGIF is exported out of the cells, whereas only 1% of proIGIF is exported. (B) Cell lysates and conditioned medium from similarly transfected but unlabeled COS cells were assayed for IFN-γ-inducing activity (19) as in Fig. 1C.

To study the role of ICE in the activation and export of IGIF under physiological conditions, we examined the processing and export of IGIF from lipopolysaccharide (LPS)-stimulated Kupffer cells isolated from Propionibacterium acnes-elicited wild-type and ICE−/− mice (20). Although lysates of Kupffer cells from wild-type and ICE−/− mice contained similar amounts of IGIF [as determined by an enzyme-linked immunosorbent assay (ELISA) that recognized both proIGIF and mature IGIF], IGIF was detected in the conditioned medium of wild-type cells but not in that of ICE−/− cells (Fig. 3A). Metabolic labeling and immunoprecipitation experiments confirmed the presence of unprocessed proIGIF in both wild-type and ICE−/− Kupffer cell lysates. However, the 18-kD mature IGIF was present only in the conditioned medium of wild-type Kupffer cell cultures and not in that of ICE−/− cultures (Fig. 3B). Similarly, the conditioned medium of LPS-stimulated wild-type adherent splenocytes contained IFN-γ-inducing activity that was sensitive to a neutralizing antibody to IGIF (anti-IGIF); this activity was reduced in the medium of adherent splenocytes of ICE−/− mice (Fig. 3C). The absence of IGIF in the conditioned medium of ICE−/− Kupffer cells and adherent splenocytes established that the processing of proIGIF by ICE is required for the export of IGIF.

Fig. 3.

IGIF is a physiological substrate of ICE. (A) Kupffer cells from ICE−/− mice are defective in the export of IGIF. Wild-type (ICE+/+) and ICE−/− mice were primed with heat-inactivated P. acnes. Kupffer cells were isolated from these mice after 7 days and were exposed to LPS (1 μg/ml) for 3 hours. The amounts of IGIF in the conditioned media were measured by ELISA (20). ND, not detectable (<0.1 ng/ml). (B) Kupffer cells isolated from ICE+/+ and ICE−/− mice as above were metabolically labeled with [35S]methionine in the presence of LPS (1 μg/ml). Cell lysates and conditioned media were immunoprecipitated and analyzed by SDS-PAGE and fluorography. Mobilities of proIGIF and the 18-kD mature IGIF are indicated on the right; molecular mass markers are shown on the left. (C) ICE−/− splenocytes are deficient in IGIF production. Conditioned media from LPS-stimulated ICE+/+ or ICE−/− adherent splenocytes were added to wild-type nonadherent splenic T cell cultures in the presence or absence of neutralizing anti-IGIF. IFN-γ concentrations in the medium were determined 20 hours later.

The sera of ICE−/− mice stimulated by P. acnes and LPS (21) also contained reduced amounts of IGIF (Fig. 4A). This finding may account for the lower concentrations of IFN-γ in the sera of treated ICE−/− mice (Fig. 4B) (22) because we observed no difference between wild-type and ICE−/− mice in the production of IL-12, the other cytokine known to induce IFN-γ (23). Nonadherent splenocytes from wild-type and ICE−/− mice produced similar amounts of IFN-γ when stimulated with IGIF in vitro. Administration of recombinant mature IGIF (6) into ICE−/− mice restored IFN-γ production in these animals (Fig. 4B), which indicated that the impaired production of IFN-γ was not the result of a defect in the T cells of ICE−/− mice. Moreover, injection of neutralizing anti-IGIF suppressed IFN-γ production in wild-type animals stimulated by P. acnes and LPS (Fig. 4C). The defect in IFN-γ production in ICE−/− mice was comparable in magnitude to the defect in IL-1β release, whereas only slight reductions were observed for TNF-α or IL-6 (3). Thus, ICE is necessary for processing of the IGIF precursor and export of active IGIF.

Fig. 4.

ICE-deficient mice have reduced serum concentrations of IGIF and IFN-γ. (A and B) Wild-type (ICE+/+) and ICE−/− mice (n = 3) primed with heat-inactivated P. acnes were exposed to LPS or LPS plus recombinant IGIF (21), and the concentrations of IGIF (A) and IFN-γ (B) in the sera were measured by ELISA 3 hours after LPS exposure. (C) Anti-IGIF blocks IFN-γ production in wild-type animals. Wild-type mice (n = 3) primed with P. acnes were injected with LPS plus anti-IGIF. Serum concentrations of IFN-γ were determined 3 hours later. IgG, immunoglobulin G.

IFN-γ and IL-1β are pleiotropic cytokines that contribute to the pathology associated with a variety of infectious, inflammatory, and autoimmune diseases. IFN-γ promotes the activation of macrophages and natural killer cells and contributes to the regulation of T helper cell immune responses, whereas IL-1β stimulates proinflammatory responses in neutrophils, endothelial cells, synovial cells, osteoclasts, and other cell types (24). The processing of proIGIF by ICE establishes a link in the regulation of IL-1β and IFN-γ production with implications for monocyte- or macrophage-mediated and T cell-mediated immune functions. IFN-γ can increase the expression of ICE in monocytic cells (25), which suggests a positive-feedback regulation between ICE and IFN-γ that may further enhance the production of IGIF and IL-1β. However, IFN-γ production by antigen-specific T cells may not be dependent on the ICE-IGIF pathway, because mitogen (concanavalin A) or antigen stimulation of splenic T cells from ICE−/− mice elicited release of normal amounts of IFN-γ (26). T cell proliferation and delayed-type hypersensitivity responses are normal in ICE−/− mice after a secondary exposure to Listeria monocytogenes (22). Thus, the ICE-IGIF pathway of IFN-γ production may be more relevant in vivo to monocyte- or macrophage-mediated inflammatory insults, as opposed to T cell-dependent immune responses.

ICE processing of proIGIF and IFN-γ production may be central events in the pathogenesis of sepsis. Mice lacking IFN-γ or its receptor are resistant to endotoxic shock (27), and neutralizing anti-IGIF prevents LPS-induced hepatic injury in P. acnes-primed mice (6). These observations suggest that the reduced concentrations of IL-1β, IGIF, and IFN-γ in LPS-exposed ICE−/− mice (3, 22) account for their increased resistance to LPS-induced sepsis relative to mice lacking a functional IL-1β gene (4), which have a normal septic response. The involvement of ICE in the regulation of these multiple proinflammatory cytokines should be considered in future evaluations of the therapeutic effects of ICE inhibition.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 12.
  14. 13.
  15. 14.
  16. 15.
  17. 16.
  18. 17.
  19. 18.
  20. 19.
  21. 20.
  22. 21.
  23. 22.
  24. 23.
  25. 24.
  26. 25.
  27. 26.
  28. 27.
  29. 28.
  30. 29.
  31. 30.
  32. 31.
View Abstract

Navigate This Article