The Precursor but Not the Mature Form of IL1α Blocks the Release of FGF1 in Response to Heat Shock*

Interleukin (IL)1α mediates proinflammatory events through its extracellular interaction with the IL1 type I receptor. However, IL1α does not contain a conventional signal peptide sequence that provides access to the endoplasmic reticulum-Golgi apparatus for secretion. Thus, we have studied the release of the precursor (p) and mature (m) forms of IL1α from NIH 3T3 cells. We have demonstrated that mIL1α but not pIL1α was released in response to heat shock with biochemical and pharmacological properties similar to those reported for the stress-mediated release pathway utilized by fibroblast growth factor (FGF)1. However, unlike the FGF1 release pathway, the IL1α release pathway appears to function independently of synaptotagmin (Syt)1 because the expression of a dominant-negative form of Syt1, which represses the release of FGF1, did not inhibit the release of mIL1α in response to temperature stress. Interestingly, whereas the expression of both mIL1α and FGF1 in NIH 3T3 cells did not impair the stress-induced release of either polypeptide, the expression of both pIL1α and FGF1 repressed the release of FGF1 in response to temperature stress. These data suggest that the release of mIL1α requires proteolytic processing of its precursor form and that mIL1α and FGF1 may utilize similar but distinct mechanisms for export.

precursor proteins that are cleaved by two distinct specific proteases to produce mature 17-kDa forms of the IL1 prototypes from the C-terminal end of the precursor (3). Whereas precursor (p) IL1␤ is biologically inactive until it is processed into the mature (m) form by the IL1␤-converting enzyme (ICE) (4,5), pIL1␣ is biologically active (5). Precursor IL1␣ is recognized by a calcium-dependent protease of the calpain family, and this cleavage results in the formation of the mature counterpart (6). Interestingly, the N-terminal fragment derived from pIL1␣ proteolytic processing contains a functional nuclear localization signal (7) and can be translocated to the nucleus (8). The ability of pIL1␣ to bind the IL1 type I receptors with high affinity, with subsequent activation of the signal transduction pathway, and the presence of the nuclear localization sequence anticipates the existence of a biological role for pIL1␣, independent from the activity of mIL1␣. Indeed, comparative studies using the pIL1␣ and mIL1␣ forms have suggested that pIL1␣, but not mIL1␣, is a negative regulator of cell migration (9).
Amino acid sequence (10) and x-ray crystallographic analysis (11) of the IL1 and FGF prototypes have revealed rather striking structural similarities between members of the two gene families. Like the majority of the IL1 gene family members (1,2), the FGF gene family prototypes, FGF1 and FGF2 also lack a classical signal peptide sequence to direct secretion through the ER-Golgi apparatus (12). Whereas the release of FGF1 is regulated by a variety of stress conditions (13)(14)(15)(16)(17)(18), the release of FGF2 is not regulated by stress (19), and FGF2 contains structural features that repress the stress-induced release of FGF1 (19). Because (i) the release pathways utilized by the FGF prototypes have diverged, and (ii) it is unlikely that the pathway responsible for the regulation of FGF1 would have evolved independent of the mechanisms utilized by other signal peptide-deficient gene products to gain access to the extracellular compartment, we sought to determine whether other signal peptide-deficient cytokines are also able to utilize the FGF1 release pathway for export. We focused our effort on IL1␣ because its precursor form is biologically functional (5,20), and extracellular IL1␣ is well described as an antagonist of FGFdependent biological activities (21)(22)(23)(24). We report that the release pathway utilized by IL1␣ exhibits similar biochemical, pharmacological, and biological properties to the FGF1 release pathway. In contrast, unlike the FGF1 release pathway (16,17), IL1␣ does not require the function of synaptotagmin (Syt)1 but does require the function of an intracellular protease to convert pIL1␣ to mIL1␣, because only mIL1␣ is released in response to heat shock. Lastly, the stress-induced IL1␣ and FGF1 release pathways may be convergent, because the expression of pIL1␣ acts as a dominant-negative repressor of FGF1 release.
Transfectants were grown to 70 -80% confluency and prior to the temperature stress, the cells were washed with DMEM. The heat shock was performed in DMEM at 42°C for 2 h, as previously described (12). When FGF1 transfectants were subjected to temperature stress, the heat shock medium was supplemented with 4 units/ml heparin (Upjohn Co.). The effects of Brefeldin A (Epicenter Technologies), 2-deoxyglucose (Sigma), and amlexanox (Takeda) on mIL1␣ release were evaluated as previously described (17)(18).
Processing of Cell Lysates and Conditioned Medium and Immunoblot Analysis-Prior to and after temperature stress, total cell lysates were obtained by sonication of cell pellets collected in Germino buffer DT (10 mM Tris, pH 7.6, containing 250 mM NaCl, 10 mM MgCl 2 , 1 mM EDTA, 10 mM 2-mercaptoethanol, and 0.1% (v/v) Triton X-100) as described previously (26). Conditioned medium was collected, filtered through a 0.22-m filter, and then treated with either dithiothreitol (DTT, Sigma) and (NH 4 ) 2 SO 4 (Sigma) or were left untreated. DTT treatment was performed at a concentration of 0.1% (w/v) DTT at 37°C for 2 h, and (NH 4 ) 2 SO 4 fractionation was performed at 95% (w/v) saturation as described previously (18). Cell lysates and DTT-or (NH 4 ) 2 SO 4 -treated medium from mIl1␣⅐␤-gal and pIL1␣⅐␤gal NIH 3T3 cell transfectants were further processed by affinity chromatography using a 0.5-ml paminobenzyl 1-thio-␤-D-galactopyranoside column (Sigma) previously equilibrated with Germino buffer DT. The column was washed with Germino buffer DT, followed by a wash with Germino buffer DT. The proteins were eluted with 2 ml of 0.1 M sodium borate, pH 10 and were concentrated by centrifugation using Centricon 30 concentrators (Amicon, Inc.).
Prior to immunoprecipitation, the cells were washed and scraped in cold phosphate-buffered saline, and cell pellets were obtained by centrifugation as described previously (26). The pellets were resuspended in 1 ml of cold NP lysis buffer (20 mM Tris, pH 7.5, containing 300 mM sucrose, 60 mM KCl, 15 mM NaCl, 5% (v/v) glycerol, 2 mM EDTA, 1% (v/v) Triton X-100, 2 g/ml aprotinin, 2 g/ml leupeptin). Conditioned medium with or without DTT treatment were filtered through a 0.22-m filter and concentrated using a centrifugal filter device (Ultrafree-15, Millipore) to a volume of ϳ500 l. 2 g/ml aprotinin, 2 g/ml leupeptin, and 5 l of goat anti-human IL1␣ polyclonal antibodies (1.21 mg/ml, a generous gift from Dr. R. Chizzonite, Hoffmann-La Roche Inc.) were added to either cell lysates or conditioned medium, and the samples were rotated at 4°C for 18 h. Protein A-Sepharose (Amersham Pharmacia Biotech) was added, and the samples were rotated at 4°C for an additional 2 h. All IL1␣ samples either eluted from the paminobenzyl 1-thio-␤-D-galactopyranoside column or purified by immunoprecipitation were resolved by either 15% (w/v) SDS-PAGE (native forms of pIL1␣ and mIL1␣) or 8% (w/v) SDS-PAGE (pIL1␣⅐␤-gal and IL1␣⅐␤-gal fusion proteins) and subjected to immunoblot analysis as previously described (9). Briefly, proteins were transferred to a nitrocellulose membrane (Hybond C, Amersham Pharmacia Biotech) and probed with goat anti-human IL-1␣ polyclonal antibody at a 1:1200 dilution. IL-1␣ specific bands were visualized by chemiluminescence (ECL, Amersham Pharmacia Biotech) following the manufacturer's instructions.
Analysis of the pIL1␣⅐␤-gal/FGF1 and mIL1␣⅐␤-gal/FGF1 NIH 3T3 cell cotransfectants was accomplished by dividing individual cell lysates and conditioned medium in half for the detection of FGF1 and IL1␣ immunoreactive bands. For FGF1 analysis, total cell lysates were obtained by sonication in cold lysis buffer containing 1% (v/v) Triton X-100 (Sigma) and adsorbed to a 1-ml heparin-Sepharose CL-6B column (Amersham Pharmacia Biotech), previously equilibrated with 50 mM Tris pH 7.4 containing 10 mM EDTA (TEB). The column was washed with TEB, and the proteins were eluted with 2 ml of TEB containing 1.5 M NaCl. The volume of the eluates was reduced using Centricon 10 concentrators (Amicon), and the eluates were analyzed by 15% (w/v) SDS-PAGE followed by immunoblot analysis using anti-human FGF1 polyclonal antibodies (27) as described previously (16). IL1␣ was analyzed by immunoprecipitation using anti-IL1␣ antibody and resolved by 8% (w/v) SDS-PAGE followed by immunoblot analysis, as described above. Cell lysates from mIL1␣⅐␤-gal, mIL1␣⅐␤-gal/p65 Syt1, and mIL1␣⅐␤-gal/ Syt1⌬(120 -214) mutant NIH 3T3 cell cotransfectants were obtained by sonication in NP lysis buffer and processed as described below. Conditioned medium from these cotransfectants were treated with 0.1% (w/v) DTT and divided into two equal samples. The first sample was adsorbed to heparin-Sepharose CL-6B (1 ml) and washed with TEB, and samples were eluted with 1.5 M NaCl, as reported above. Cell lysates and conditioned medium purified by heparin affinity were resolved by 10% (w/v) SDS-PAGE and subjected to immunoblot analysis using a rabbit anti-rat Syt1 polyclonal antibody as described previously (17). The second sample was processed by IL1␣ immunoprecipitation as described above. As a negative control, cell lysates were also immunoprecipitated with a control antibody (goat IgG purified immunoglobulin from pooled normal goat serum, Sigma). Cell lysates and conditioned medium purified by IL1␣ immunoprecipitation were resolved by 8% (w/v) SDS-PAGE, followed by IL1␣ immunoblot analysis, as described above. IL1␣ and Syt1 specific bands were visualized by chemiluminescence (ECL) following the manufacturer's instructions. The activity of lactate dehydrogenase in conditioned medium was utilized as an assessment of cell lysis in all experiments and was measured by a colorimetric assay using pyruvate as a substrate (Sigma). Each experiment reported was repeated at least three times with similar results in all cases, and representative data are shown in each figure.

RESULTS AND DISCUSSION
We evaluated the ability of human IL1␣ to enter the extracellular compartment using NIH 3T3 cells because these cells (i) are refractory to the activity of endogenous and exogenous IL1␣ (data not shown) and (ii) have proven to be valuable for the study of the FGF1 release pathway (16 -18). Thus, stable NIH 3T3 cell transfectants expressing either pIL1␣ or mIL1␣ with or without the ␤-gal reporter gene were obtained and were subjected to temperature stress at 42°C for 2 h. Cell lysates and conditioned medium treated with 0.1% (w/v) DTT were processed for IL1␣ immunoblot analysis. As shown in Fig. 1, A and B, mIL1␣ and mIL1␣⅐␤-gal were readily visible in medium conditioned by heat shock. In contrast, neither pIL1␣ nor pIL1␣⅐␤-gal was detected in heat shock-conditioned medium, but both forms of the polypeptide were present in cell lysates (Fig. 1, A and B). These data suggest the following: (i) the mature (residues 113 to 271) but not the precursor (residues 1 to 271) form of IL1␣ was able to enter the extracellular compartment in response to temperature stress, (ii) the release of IL1␣ does not restrict export of the reporter gene product, ␤-gal, and (iii) cell lysis does not account for the release of mIL1␣ because the absence of pIL1␣ in medium conditioned by heat shock served as a negative control.
Because FGF1 (14) but not FGF2 (19) was released from NIH 3T3 cells in response to heat shock as a DTT-sensitive latent homodimer (14), and both reducing agents and (NH 4 ) 2 SO 4 were able to activate the heparin affinity of latent FGF1(16), we questioned whether these reagents would also affect the ability of IL1␣ to be recognized by affinity reagents. Because mIL1␣⅐␤gal does not exhibit heparin-binding affinity (data not shown), we utilized ␤-gal affinity and IL1␣ immunoprecipitation to assess this issue. As shown in Fig. 1C, IL1␣ immunoblot analysis of medium conditioned by temperature stress was performed using p-aminobenzyl-1-thio-␤-D-galactopyranoside (PATG) affinity and failed to detect the presence of the IL1␣⅐␤gal fusion product. However, treatment of heat shocked-conditioned medium with either 0.1% (w/v) DTT or 95% (w/v) (NH 4 ) 2 SO 4 did resolve the presence of mIL1␣⅐␤-gal in medium conditioned by heat shock from mIL1␣⅐␤-gal NIH 3T3 cell transfectants. In contrast, the use of IL1␣ immunoprecipitation detected the presence of IL1␣⅐␤-gal in medium conditioned by heat shock, and the addition of either DTT (Fig. 1D) or (NH 4 ) 2 SO 4 (data not shown) to the medium did not alter the ability of the anti-IL1␣ antibody to recognize the fusion protein (Fig. 1D). These data suggest that whereas mIL1␣⅐␤-gal is present in the extracellular compartment in a conformation that does not efficiently recognize PATG (unlike FGF1 that utilizes Cys oxidation for export, Ref. 15), this conformation may not represent mIL1␣ but rather the conformation of ␤-gal. Thus, it is unlikely that mIL1␣ utilizes Cys oxidation for export.
We also examined the kinetics and pharmacologic properties of mIL1␣⅐␤-gal release in response to temperature stress, because the FGF1 release pathway exhibits relatively slow export kinetics (14). Moreover, FGF1 is sensitive to agents that interfere with translation (12), transcription (12), ATP biosynthesis (17), and assembly of the F-actin cytoskeleton (28) but is not sensitive to Brefeldin A, an agent that interferes with the function of the ER-Golgi apparatus (14). As previously observed with FGF1 release (14), the release of mIL1␣⅐␤-gal also required at least 90 min of temperature stress before IL1␣ immunoblot analysis was able to resolve the presence of the fusion protein in medium conditioned by heat shock (Fig. 2A). In addition, IL1␣ immunoblot analysis revealed that the release of mIL1␣⅐␤-gal was not sensitive to treatment of the mIL1␣⅐␤-gal NIH 3T3 cell transfectants with Brefeldin A (Fig.  2B) but was sensitive to the presence of 2-deoxyglucose ( Fig.  2C) as well as cycloheximide and actinomycin D (data not shown). Further, the release of mIL1␣⅐␤-gal was also sensitive to treatment with amlexanox (Fig. 2D), an agent which induces the Src-dependent disassembly of F-actin stress fibers (28) and exhibits a dose-response to amlexanox similar to that reported for the inhibition of FGF1 release in vitro (18). These data suggest that the pathway utilized by mIL1␣⅐␤-gal for entering the extracellular compartment is dependent upon transcription, translation, ATP biosynthesis, and the actin cytoskeleton but is independent of the function of the ER-Golgi apparatus.
Although the data with p65 Syt1⌬(120 -214) expression suggest that the IL1␣ and FGF1 release pathways may have diverged, we questioned whether the coexpression of IL1␣ and FGF1 in the NIH 3T3 cell could confirm this premise. Thus we obtained stable NIH 3T3 cell cotransfectants in which pIL1␣⅐␤gal and mIL1␣⅐␤-gal NIH 3T3 transfectants were cotransfected with FGF1 and examined their ability to release mIL1␣⅐␤-gal, pIL1␣⅐␤-gal, and FGF1 in response to heat shock. As shown in Fig. 4, IL1␣ and FGF1 immunoblot analysis revealed the pres-ence of both mIL1␣⅐␤-gal and FGF1 in medium conditioned by heat shock from the FGF1/mIL1␣⅐␤-gal NIH 3T3 cell cotransfectants. However, IL1␣ and FGF1 immunoblot analysis failed to detect the presence of either pIL1␣ or FGF1 from the FGF1/ pIL1␣⅐␤-gal NIH 3T3 cell cotransfectants in response to temperature stress (Fig. 4). Because the precursor form but not the mature form of IL1␣ was able to repress the release of FGF1 in response to heat shock, the precursor domain of IL1␣ may contain a structural feature that may function as a dominantnegative effector of FGF1 release. Although we do not know the element within the precursor domain of IL1␣ responsible for this event, these data do suggest that the IL1␣ and FGF1 release pathways may indeed be convergent.
Our data are consistent with the observation that human bladder carcinoma cells selectively release mIL1␣ but not pIL1␣ in vitro (40). Moreover, Siders, et al. (41) using a variety of different cell types transfected with either the precursor or mature forms of the IL1 prototypes observed that the mature forms of IL1␣ and IL1␤ were also preferred for release and that the release of mIL1␣ from pIL1␣ was dependent upon the presence of a calpain-like protease activity (41). Interestingly, whereas macrophages from the ICE-null mouse were able to release pIL1␤ but not mIL1␤ in response to lipopolysaccharide stimulation, these cells were also deficient in IL1␣ release suggesting a possible role for ICE in the release of IL1␣ (42). Because we were unable to observe the presence of mIL1␣ in cytosol derived from the pIL1␣ NIH 3T3 cell transfectants, even under conditions of apoptosis where ICE is functional (data not shown), it is unlikely that ICE expression is involved in the formation of mIL1␣ in the NIH 3T3 cell. Rather, the NIH 3T3 cells may be deficient in the expression of the appropriate calpain and as a result are not capable of processing intracellular pIL1␣. Alternatively, this may also be because of the presence of a calpain inhibitor (43) or because of the absence of calpain activation in response to heat shock. However, our data do suggest that the proteolytic conversion of pIL1␣ to mIL1␣ may be prerequisite for the appearance of IL1␣ in the extracellular compartment.
Although the ability of pIL1␣ to repress the release of FGF1 in response to heat shock suggests a connection between the IL1␣ and FGF1 release pathways, we do not know whether mIL1␣ requires the function of helper genes to gain access to the extracellular compartment. However, it is interesting to note that like FGF1 (18,44), early studies with IL1␣ using gel exclusion chromatography also noted the presence of IL1␣ in high molecular weight complexes (34,45). Whether these represent the IL1␣ equivalent of the multiprotein aggregate of FGF1(18) is not known. It is intriguing to speculate that like the FGF1 release pathway, the release of mIL1␣ in response to temperature stress may require the function of an unknown set of helper genes.