IL-2 Induces T Cell Adherence to Extracellular Matrix: Inhibition of Adherence and Migration by IL-2 Peptides Generated by Leukocyte Elastase¹

The continuous movement of leukocytes across blood vessel walls to extravascular sites and back to the bloodstream is essential for battling foreign pathogens and maintaining homeostasis (1, 2). The migration of T cells through tissues is regulated by adhesion receptors, such as integrins, and by receptors that receive signals provided by proinflammatory mediators, such as cytokines, chemokines, and extracellular matrix (ECM)³-degrading enzymes (3, 4, 5, 6). Here, we have examined the interactions between human T cells and two molecules, neutrophil elastase and IL-2, which appear to be involved in T cell activation and migration through tissues.

IL-2 is a 15.5-kDa glycoprotein that participates in the development of inflammation and in the regulation of apoptosis (7, 8). In addition to its proactivatory and proliferative roles, IL-2 also induces neutrophil adhesion to HUVEC in a CD18-mediated manner (9) as well as chemotactic responses in T cells, both directly and by regulating their expression of CC chemokine receptors (10, 11). The IL-2R consists of three distinct membrane chains: the α-, β-, and γ-chains. The ability of IL-2 to induce T cell activation, differentiation, and proliferation involves the β- and γ-chains of the IL-2R, which are coupled through their cytoplasmic domains to intracellular protein tyrosine kinases and a protein serine/threonine kinase (8, 12). X-ray crystallographic analysis and deletion experiments showed that the sites of IL-2 that bind to the β- or γ-chains of its receptor are located within the α-helical and 30-amino acid residues of the N-terminal domain of IL-2 (13). Indeed, anti-IL-2 Abs that recognize amino acid epitopes in the N-terminal region of IL-2 can inhibit IL-2-induced lymphocyte proliferation. The C-terminal portion of IL-2 and its three Cys residues seem to contribute to the folding and active conformation of IL-2 (13, 14).

Most neutrophil elastase (also termed human leukocyte elastase), which exists as either a membrane-bound or soluble moiety, is produced and released by neutrophils, although small amounts are also produced by macrophages, monocytes, and T cells (15, 16). Elastase degrades basement membrane and ECM glycoproteins, such as elastin, collagen, and fibronectin (FN), as well as molecules expressed on the surface of T cells, e.g., CD4, CD8, and CD2 (17). Recently, two novel functions of neutrophil elastase have been shown: 1) membrane-bound elastase modulates immune cell adhesiveness by interacting with the integrin α_Mβ₂ on neutrophils (18); and 2) elastase processes IL-8 and thus alters the biologic functions of this chemokine (19).

Recently, we have shown that IL-7, among other cytokines, interacts with ECM, and that both soluble and matrix-complexed IL-7 induce integrin-mediated adhesive interactions of human T cells with ECM-bound IL-7 and with purified VCAM-1 molecules (20, 21). The biologic effects of IL-7 are linked to its interaction with the γ-chain of the IL-2R, which is a tyrosine kinase signal-transducing molecule (22). Also, exposure of T cells (3 to 5 days) to IL-2 was found to induce their migration on collagen (23). Therefore, we examined whether IL-2 can induce T cell adhesion to ECM glycoproteins, whether elastase can process human rIL-2, and whether such putative IL-2 peptides can affect T cell-ECM interactions. We found that IL-2 can indeed induce T cell adhesion to ECM glycoproteins that are otherwise nonadhesive moieties. In addition, we may have identified a group of naturally occurring, elastase-generated IL-2 fractions and peptides capable of inhibiting T cell adhesion and migration.

The following reagents were obtained as indicated. Human rIL-2 (sp. act. 18 × 10⁶ U/mg; Chiron, Amsterdam, The Netherlands); human rIL-7 (sp. act. 2 × 10⁵ U/μg; Immunex, Seattle, WA); recombinant human MIP-1β (PeproTech, Rocky Hill, NJ); FN (Chemicon; Temecula, CA); BSA, laminin (LN), PMA, and TRITC-conjugated phalloidin (Sigma Chemical, St. Louis, MO); collagen type IV (CO-IV; ICN, Costa Mesa, CA); and HEPES buffer, antibiotics, heat-inactivated FCS, sodium pyruvate, and RPMI 1640 (Beit-Haemek, Israel). An anti-β₁ integrin-specific affinity-modulating mAb, 8A2 (24), was donated by Dr. J. M. Harlan (Washington University, Seattle, WA). All protected amino acids, coupling reagents, and polymers were obtained from Nova Biochemicals (Läufelfingen, Switzerland). Synthesis-grade solvents were obtained from Labscan (Dublin, Ireland). HPLC solvents and columns were obtained from Merck (Darmstadt, Germany).

Human T cells were purified from the peripheral blood of healthy donors, and T cell adhesion to immobilized protein substrates was examined as previously described (20, 21). Briefly, human leukocytes were isolated on a Ficoll gradient, washed, and incubated (2 h, 37°C, 7.5% CO₂, humidified atmosphere) on petri dishes. The nonadherent cells were then collected and incubated (1 h, 37°C, 7.5% CO₂, humidified atmosphere) on nylon wool columns (Novamed, Jerusalem, Israel). Unbound cells were eluted from the columns by extensive washings. The resulting cell population was always >92% T cells. The flat-bottom microtiter wells that had been precoated with ECM or ECM proteins (FN or LN; 1 μg/well) were blocked with 0.1% BSA. After 0.5 h at 37°C, the wells were washed and ⁵¹Cr-labeled T cells were added to the wells, 10⁵ cells/100 μl of adhesion medium (RPMI 1640 supplemented with 0.1% BSA, 1% sodium pyruvate, 1% HEPES buffer). The microtiter plates containing the cells were incubated (30 min, 37°C) in a humidified, 7.5% CO₂ atmosphere and then washed. The adherent cells were lysed, and the resulting supernatants were removed and analyzed in a gamma counter. For each experimental group, the results were expressed as the mean percentage ± SD of bound T cells from quadruplicate wells. To some wells, different concentrations of soluble IL-2 were added concomitantly with the T cells, and with others, different concentrations of elastase-degraded IL-2-derived fractions, or the corresponding synthetic peptides, were added together with stimulators (PMA (50 ng/ml), IL-7 (50 ng/ml), MIP-1β (20 ng/ml), 8A2 (1 μg/ml), mAb anti-CD3 (1 μg/ml), or IL-2 (10 U/ml)).

T cell chemotaxis was performed and analyzed as previously described (25). Briefly, the migration of human T cells (0.5 × 10⁶ cells in adhesion medium/well) was examined in a 48-well chemotaxis microchamber (NeuroProbe, Cabin John, MD). The two compartments of the microchambers were separated by a FN-coated polycarbonate filter (5-μm pore size; Osmonics Protein Products, Livemore, CA). Where indicated, MIP-1β or IL-2 was added to the lower wells, and the T cells were added to the upper chambers together with the peptides. After incubation (120 min, 37°C, in a humidified, 7.5% CO₂ atmosphere), the filters were removed, fixed, and stained with a Diff-Quik staining kit (Dade, Düdingen, Switzerland). The number of migrating T cells in five high-power fields (under 500× magnification; Wild Microscope, Heerbrugg, Switzerland) was evaluated. For each group, the results are expressed as the mean number of cells in one high-power field.

Neutrophils were isolated from the whole blood of a healthy donor by dextran sedimentation and Ficoll-Hypaque gradient centrifugation, as previously described (26). Elastase was isolated by aprotinin-Sepharose affinity chromatography, followed by carboxymethyl-cellulose ion exchange chromatography as developed by Baugh and Travis (26, 27). The purified elastase, which was lyophilized and stored at −20°C until used, was biochemically checked to be entirely free from cross-contamination with cathepsin G (not shown). IL-2 was dissolved in distilled water to yield a 1 mg/ml solution. Lyophilized neutrophil elastase (50 μg) was dissolved in 1 ml of PBS and immediately added to the IL-2 solution. The elastase-IL-2 mixture was incubated (12 h) at 37°C. Aliquots were removed and stored at −20°C until subjected to HPLC separation.

Elastase digests of IL-2 were purified with a prepacked Lichrospher-100 RP-18 column (4 × 25 mm, 5-μm bead size), using a binary gradient formed with 0.1% trifluoroacetic acid (TFA) in H₂O (solution A) and 0.1% TFA in 75% acetonitrile in H₂O (solution B; at t = 0 min, B = 3.5%; at t = 5 min, B = 3.5%; and then the concentrations began to increase: at t = 60 min, B = 100% (i.e., 75% acetonitrile)). The flow rate was constant at 0.8 ml/min. A Spectra-Physics SP8800 liquid chromatography system (Fremont, CA) equipped with an Applied Biosystems model 757 (Foster City, CA) variable wavelength absorbency detector was used. The column effluents were monitored by UV absorbency at 220 nm, and the chromatograms were recorded on a ChromeJet integrator. Fractions that were 20% or more above valley levels were pooled, rotoevaporated to a minimal volume, and diluted with HPLC grade water. The rotoevaporation and dilution with water step was performed twice to remove residual TFA and acetonitrile.

Purified peptide solutions (∼40 μg of peptide in 40 μl, with 5 μg of norvaline (an unnatural amino acid) as an internal standard) were rotoevaporated, hydrolyzed (10°C, 22 h) in 6 N HCl under vacuum, and analyzed with an amino acid analyzer (HP1090, Hewlett-Packard, Palo Alto, CA). An on-line precolumn o-phthalaldehyde/9-fluorenylmethoxycarbonyl (F-moc) derivatization, combined with reverse phase chromatography, was used to determine the amino acid composition of the peptides and the total peptide yield. Without exception, all of the peptides yielded excellent analysis ratios of corresponding amino acid deviations from expected values of <10%.

Analysis of the elastase-generated IL-2 fractions was performed using an Applied Biosystems model 470A gas phase microsequencer. Phenylthiohydantoin amino acid derivatives were separated on-line by reverse phase HPLC on a PTH C-18 column (2.1 × 220 mm) using a model 120A analyzer (Applied Biosystems).

IL-2-derived peptides were prepared by conventional solid phase peptide synthesis, using an AMS-422 automated solid phase multiple peptide synthesizer (Abimed, Langenfeld, Germany). The F-moc strategy was used for peptide chain assembly according to the commercial protocol. In each reaction vessel, we used 12.5 μmol of Wang resin containing the first covalently bound corresponding N-F-moc C-terminal amino acid (typically, polymer loadings of 0.3–0.7 mmol/g resin were used). F-moc deprotection was achieved by two consecutive treatments with 20% piperidine in dimethyl formamide, usually 10 to 15 treatments each min at 22°C, depending on the length of peptide and the F-moc-protected amino acid type. The protecting groups used for the side chain of the amino acids were tert-butyloxycarbonyl for Trp, trityl for Asn, and, tert-butylether for Thr. Coupling was usually achieved using two successive reactions (typically 20–45 min each at 22°C, depending on the length of peptide and amino acid derivative type) with 50 μmol (4 eqv) of N-F-moc-protected amino acid, 50 μmol (4 eqv) of benzotriazole-1-yl-oxy-tris-pyrolidino-phosphonium hexafluorophosphate (PyBop) reagent, and 100 μmol (8 eqv) of N-methylmorpholine were all dissolved in dimethyl formamide. The peptide was cleaved from the polymer by reacting (2 h, 22°C) the resin with TFA/H₂O/triethylsilane (90/5/5, v/v/v). The solution containing the crude unprotected peptides was then cooled down to 4°C, precipitated with ether (4°C), and centrifuged (15 min, 3000 rpm, 4°C). The pellet was washed and centrifuged (×3) with ether, dissolved in 30% acetonitrile in H₂O, and lyophilized. The lyophilized material was reconstituted in double distilled water before use; only the stock solution, not the diluted material, was stored at −20°C.

T cells were incubated (18 h, 37°C, 7.5% CO₂, humidified atmosphere) in culture medium. IL-2 was added to the cell cultures, which were then incubated for 48 h. The T cells were then washed and seeded onto FN-covered coverslips in the presence of either PMA (50 ng/ml), IL-2 (100 U/ml), or IL-2 peptides (0.1 ng/ml). After 1 h at 37°C, the adherent cells were fixed (3 min) with paraformaldehyde (3%) and Triton X-100 (0.5%), washed, and fixed (20 min) again with paraformaldehyde (3%). The fixed adherent cells were washed, treated with TRITC-phalloidin, and washed again. Photographs (×1000 magnification) were then taken.

Some cytokines induce adhesion of leukocytes to endothelial cells, the underlying basement membrane, and to the ECM (3, 4, 6, 20, 21). Therefore, we examined the ability of IL-2 to induce adhesion of human T cells to ECM, FN, LN, and CO-IV. The results indicated that IL-2 induced T cell adhesion to FN, LN, and CO-IV (Fig. 1,A), as well as to intact ECM (Fig. 1 B). Note that the adhesion of T cells to LN induced by IL-2 was lower than that induced by the other ECM glycoproteins. When T cells were activated only with PMA, 45 ± 4.4% of them adhered to immobilized ECM and ECM glycoproteins (not shown). IL-2-induced T cell adhesion to the ECM glycoproteins was inhibited by anti-human β₁ integrin mAbs (not shown), which suggests that the proadhesive effects of IL-2 were induced via cell surface-expressed integrins. However, under our experimental conditions, IL-2 did not alter the T cell surface expression of β₁ integrins (not shown). Thus, IL-2, in addition to other proinflammatory mediators, appears to regulate the adhesiveness of resting human T cells to immobilized ECM and ECM glycoproteins.

We have assumed that the degradation of IL-2 can occur in the inflamed milieu in which both cytokines, such as IL-2, and proteolytic enzymes, such as neutrophil elastase, are present. We also hypothesized that, in contrast to the intact IL-2 molecule, certain portions of IL-2 can abrogate the adhesiveness of activated T cells to ECM ligands. Hence, elastase and soluble IL-2 were incubated together at physiologic conditions. HPLC analysis of the elastase-degraded IL-2 revealed at least eight peaks of IL-2, each of which represented at least one low m.w. protein fragment (Fig. 2).

Next, we examined the ability of the HPLC-purified IL-2 fractions, generated by elastase degradation, to inhibit IL-2-induced interactions of T cells with FN. We chose to investigate the major peaks of HPLC-purified, elastase-degraded fractions of IL-2. Fractions 2, 7, and 8 inhibited the adhesion of T cells to immobilized FN in a dose-dependent and statistically significant fashion, whereas fraction 4 and the HPLC buffer did not (Fig. 3). Thus, certain IL-2 fragments, obtained by neutrophil elastase-processing of the cytokine, can inhibit IL-2-induced adhesion of T cells to FN.

Next, the primary sequence of fractions 2 and 8 were analyzed by gas phase chromatography, because these elastase-generated fractions of IL-2 appeared to contain adhesion-suppressive peptides. Our analysis revealed that fragment 2 contained an Ile-Val-Leu (IVL; IL-2_112–114) and an Arg-Met-Leu-Thr (RMLT; IL-2_58–61) peptide, whereas fragment 8 contained a Glu-Phe-Leu-Asn-Arg-Trp-Ile-Thr (EFLNRWIT; IL-2_136–143) octa-peptide. These three peptides were synthesized, and their effects on IL-2-induced T cell adhesion to FN were studied. The IVL, RMLT, and EFLNRWIT peptides, inhibited in a dose-dependent manner the adhesion of IL-2-activated T cells to FN (Fig. 4 A). Maximum inhibition was achieved with 0.1 pg/ml for both IVL, RMLT, and EFLNRWIT (0.29, 0.20, and 0.09 pM, respectively. The inhibitory dose-response curves of all peptides are similar to those of the HPLC fractions from which they were derived.

To examine the specificity, on the biologic and chemical levels, of the inhibition of T cell adhesion by the elastase-generated, synthetic IL-2 peptides, we synthesized the three IL-2 peptides in their inverse amino acid sequences, LVI, TLMR, and TIWRNLFE, and then tested their effects on IL-2-induced T cell adhesion to FN. The results, shown in Figure 4 B, indicate that none of these peptides, tested in a broad range of dosages, interfere with T cell adhesion. Thus, the anti-adhesive effects of IVL and EFLNRWIT peptides of IL-2 appear to be due to their direct biologic effect on responding lymphocytes.

Do the IL-2 peptides, IVL, RMLT, and EFLNRWIT have to be present during the entire period of the assay to exert their inhibitory effects? The results, shown in Figure 4 C, indicate that most of the anti-adhesive effects of the three peptides persevered even if these peptides (at 1 pg/ml) were removed from the T cells before their activation with IL-2 and seeding onto the FN-coated surfaces. Apparently, their prolonged inhibitory potential may involve active intracellular signaling pathways. These results suggest that the IL-2 peptides neither exert their inhibitory activities on T cell adhesion to FN via binding to the ECM protein, nor to FN-specific β₁ integrins expressed on the adhering T cells.

The next experiment was designed and performed to verify that the three IL-2-derived peptides indeed affect T cell interactions with ECM glycoproteins other than FN. T cells were pre-exposed to the three peptides (at 10 pg/ml), and then activated with IL-2. The treated cells were then added to microtiter wells coated with CO-IV, LN, and FN. The results, shown in Figure 5, indicate that both IVL, RMLT, and EFLNRWIT inhibit T cell adhesion to the three major cell-adhesive glycoproteins of the ECM (Fig. 5), suggesting that the elastase-generated IL-2 peptides exert their inhibitory effects over different subsets of β1 integrins.

Immune cell migration is the outcome of a subtle biologic equilibrium existing between adhesion and detachment events. Lymphocyte adhesion to the subendothelial ECM and subsequent migration are two active processes that can overlap, but are not mutually dependent events. Adhesion and migration may depend on the ability of the T cells to continuously integrate different pro- and anti-adhesive signals via their versatile receptors for ECM, chemokines, cytokines, and possibly, also antigenic moieties (1, 2). In fact, it has been recently shown that transient (i.e., low affinity) rather than prolonged interactions between integrins and the ECM favor IL-4-induced B cell migration, and not adhesion (28). Therefore, we next examined the effects of the IVL, RMLT, and EFLNRWIT peptides (at 1 pg/ml) on the IL-2- and MIP-1β-induced T cell chemotaxis through FN-coated polycarbonate membranes. The gradient generated by MIP-1β and IL-2, which were placed in the lower compartment of the 48-well chemotaxis apparatus, induced a marked T cell migration through FN-coated membranes, which was about 3 to 4-fold higher than the control (Fig. 6). Both IVL, RMLT, and EFLNRWIT markedly (p < 0.01) inhibited T cell migration toward IL-2, by about 30, 90, and 60%, respectively. However, although the IVL peptide, and to a lesser degree also the RMLT peptide, markedly inhibited (80% and 60%, respectively) T cell chemotaxis toward MIP-1β, the EFLNRWIT showed only a limited inhibitory effect on the chemokine-induced T cell chemotaxis. Thus, in addition to the capacity of the IVL and EFLNRWIT peptides to inhibit T cell-ECM adhesion, they seem to inhibit T cell migration through FN in response to a diffusible gradient produced by IL-2 or MIP-1β.

The preceding chemotaxis experiments indicated that the IVL and EFLNRWIT peptides can inhibit T cell adhesion and migration through FN barriers induced not only by IL-2, but also by the chemokine MIP-1β. Therefore, in an attempt to further understand the possible physiologic relevance of such phenomena, we examined the ability of these peptides to inhibit the adhesion to FN of T cells stimulated by modes other than IL-2. T cell adhesion to FN can be up-regulated by physiologic activators such as IL-7 (21) and MIP-1β (20) and via the CD3 complex, as well as by nonphysiologic stimuli of integrin avidity, such as PMA (29), and a β₁ integrin-specific activating mAb, 8A2 (24). This mAb up-regulates the endothelial cell and ECM ligand-binding activities of the α₄β₁ and α₅β₁ integrins by binding to their cell surface-expressed sites, thus converting these nonbinding integrins to their high affinity, ligand-occupying forms (24).

At 10 pg/ml, both EFLNRWIT and IVL inhibited T cell adhesion to FN that was induced by various stimulators of T cells and modulators of the β₁ integrin functions tested (Fig. 7). However, at 0.1 pg/ml, EFLNRWIT did not inhibit PMA- and 8A2-induced adhesion, and the IVL peptide did not inhibit PMA-, 8A2-, and anti-CD3 mAb-induced T cell adhesion to FN, which indicates that these modes of activation are less susceptible to IL-2-derived peptide-induced suppression than IL-2-mediated activation. In experiments similar to those shown in Figures 7, A and B, the control peptides (LVI and TIWRNLFE) did not affect T cell adhesion to FN induced by the indicated activators (data not shown). Hence, EFLNRWIT and IVL apparently inhibit T cell adhesion to the FN component of ECM via a common intracellular event that is linked to the regulation of the avidities and affinities of β₁ integrins, and therefore, to their ligand recognition and binding.

The adhesion of immune cells to ECM is dependent on the sequestering of the cytoplasmic domains of integrins in focal adhesion sites together with actin-containing microfilament bundles (30, 31, 32). Therefore, we examined the effect of EFLNRWIT on the morphologies of adherent T cells. The T cells were activated with IL-2 or PMA, treated with the IL-2 peptides, and seeded onto FN-coated coverslips. After incubation and fixation, the actin cytoskeleton of attached T cells was stained with TRITC-conjugated phalloidin. The morphologies of FN-bound IL-2- and PMA-activated T cells (Fig. 8, A and B) were markedly different from those of nonactivated lymphocytes (Fig. 8,E); the activated T cells appeared spread, and their actin cytoskeleton performed distinct structures typical of ECM-adherent cells. The EFLNRWIT peptide inhibited the redistribution of the actin skeleton in both the IL-2- (Fig. 8,C) and PMA-treated (Fig. 8 D) FN-adherent T cells. Control peptides (LVI, TLMR, and TIWRNLFE) did not inhibit the actin reorganization of the activated T cells (not shown). Hence, the adhesion-inhibiting activity of the IL-2-derived peptide EFLNRWIT, similar to the IVL peptide (data not shown), appears to involve inhibition of the redistribution of the actin cytoskeleton, and therefore, changes in cell shape and spreading.

Certain proinflammatory mediators can exert their migratory effects in solution or while bound to components of blood vessel walls (3, 6, 33). The binding of these mediators to components of tissue barriers probably ensures the persistence and effectiveness of the mediators in the vicinity of the inflammation-mediating immune cells themselves. This transient binding also creates migratory paths, which consist of cytokine gradients that affect the movements of leukocytes through vessel walls and the ECM (3, 4). IL-2 and elastase are prominent mediators of leukocyte extravasation and migration from the vasculature through the ECM to sites of inflammation. In addition to inducing activation and proliferation of T cells, IL-2 can increase the expression of the CC-CKR1 and CC-CKR2 chemokine receptors on CD45^RO T lymphocytes (11), suggesting that IL-2 may participate in the recruitment of Ag-primed T cells to sites of immune reactions. We have shown that soluble human IL-2 can induce T cell adhesion to intact ECM and several major ECM glycoproteins. IL-2 can also associate directly with ECM ligands (dissociation constant in the range of 1 μM), and this association facilitates the adhesion of resting human T cells to lower concentrations of the ECM-complexed cytokine (not shown).

We therefore hypothesized that, in contrast to the proadhesive effects of the intact IL-2 molecule, certain short IL-2-derived peptides, which may occur in vivo, can inhibit the interactions of T cells with ECM and that this interference is independent of the effects of the peptides on the activation of T cells by IL-2. We also assumed that such moieties of IL-2 can prevent the arrival of T cells at inflamed sites. Neutrophil elastase was a likely candidate for the physiologic production of these inhibitory peptides. Elastase, which is enzymatically versatile (17, 18), can be expressed and secreted by the migrating T cells themselves (16) and can transmit mitogenic stimulations from the environment into the responding T cells (15). In fact, LPS- and FMLP-activated neutrophils express catalytically active membrane-bound elastases, proteinase 3, and cathepsin G on their cell surfaces, which ensures the presence of these enzymes at the leading edge of the tissue-invading cells (34, 35). Other matrix-degrading enzymes, such as matrix metalloproteinases, cathepsin B, the urokinase-type plasminogen activator, and plasmin, can also be bound on the cell surfaces of fibroblasts and migrating cells (5, 36). An important feature of elastase is its ability to act in both soluble and immobilized forms, since a migrating immune cell that expresses immobilized elastase may encounter matrix-bound IL-2, among other cytokines. We have found that the processing of rIL-2 by elastase results in the production of at least eight different products. Three of these products (present in HPLC fractions 2, 7, and 8) inhibited IL-2-mediated T cell adhesion to FN. Amino acid composition analysis and amino acid sequencing revealed that fraction 2 contained the tripeptide IVL (IL-2_112–114) and the tetrapeptide RMLT (IL-2_58–61), and fraction 8 the octapeptide EFLNRWIT (IL-2_136–143). The RMLT peptide appeared to be located within the IL-2-binding site of the α-chain of the IL-2R, whereas the IVL and EFLNRWIT are located at sites far from the receptor-binding sites of IL-2 (7, 8). These peptides, at a picomolar range of concentrations (i.e., 0.01–1 pg/ml), inhibited the IL-2, as well as MIP-1β-induced chemotaxis of human T cells through FN-coated polycarbonate membranes.

The chemoattractive capacity of IL-2 in T cell migration studies in vitro has been shown using bare polycarbonate filters or collagen- or Matrigel-coated membranes as immobilized substrates. T cell migration in these systems was proved to be IL-2R β-chain-specific and dependent on the activities of the matrix-degrading gelatinases (10, 37, 38). Here, in addition to their antimigratory effects, both peptides of IL-2 inhibited T cell adhesion to FN induced by various physiologic and nonphysiologic stimuli. Nevertheless, neither IVL, RMLT, nor EFLNRWIT, at 1 to 100 pg/ml, interfered with either PHA- or IL-2-mediated proliferative responses of human T cells, nor did these peptides inhibit the secretion of TNF-α and IFN-γ from these proliferating cells (not shown). Moreover, the inversely synthesized molecules LVI, TLMR, and TIWRNLFE did not inhibit T cell adhesion to FN. Thus, the migration- and adhesion-suppressive capabilities of IVL, RMLT, and EFLNRWIT are specific and are not due to toxic cell death.

How do the elastase-derived IL-2 peptides exert their inhibitory functions? We have demonstrated that the three peptides do not have to be present during the entire period of the adhesion assay, since their antiadhesive effect was apparent even after their removal from the assay before T cell activation with IL-2. This finding also implies that the IVL, RMLT, and EFLNRWIT peptides do not function by interacting with putative cell-adhesive epitopes present on the tested ECM glycoproteins. The existence and activity of putative T cell surface-expressed receptors specific to the IL-2 peptides described here require additional study. However, it seems to be highly unlikely that two of these peptides (IVL and EFLNRWIT) exert their biologic functions by interacting with the IL-2R subunits or by directly binding to β₁-specific integrins. The IL-2 peptides interfered with different modes of T cell activation, leading to their adhesion to the tested matrix proteins. Moreover, both peptides, used in a picomolar range of concentrations, appear to block T cell adhesion not only to FN, mediated predominantly via the α₄β₁ and α₅β₁ integrins, but also to LN; T cell adhesion to LN was mediated primarily through the α₃β₁ and α₆β₁ integrins. Interesting, however, the RMLT peptide resides within the IL-2Rα-binding site of IL-2; two residues, Arg⁵⁸ and Phe⁶², which are present within and adjacent to this peptide, respectively, were shown to be critical for IL-2-IL-2R interactions (13, 39). Therefore, it will also be interesting to examine whether the degradation of IL-2 by elastase produces compounds that can interfere with IL-2 binding to its receptor and with the biologic outcome of such molecular interactions.

The ability of EFLNRWIT (as well as the other two IL-2 peptides, which are not shown) to inhibit PMA- and IL-2-induced T cell adhesion to FN is probably linked to its ability to block the reorganization of the intracellular actin cytoskeleton. Integrin-cytoskeleton associations can modulate cell adhesion to ECM ligands, cell spreading in areas of cell contact with the substratum, and the microclustering and redistribution of β₁ integrins on the cell surface at sites of focal adhesion located at the ends of the actin fibers (29, 31, 40). In fact, actin cytoskeleton reorganization, occurring after activation of leukocytes via a ρ-dependent activation of the ζ isoform of protein kinase C (41, 42), is linked to leukocyte emigration. Similar observations were noted for T cell activation by IL-2 (and PMA) and attachment to FN. Although the intracellular mechanisms of action of the elastase-generated fractions and peptides have not yet been determined, we postulate that these proteins effect the adhesion and migration of T cells within the ECM by active inhibition of intracellular signal transduction pathways linked to cytoskeleton organization, resulting in an inhibition of microclustering and an association of integrins with cytoskeletal elements.

Our findings imply that the tissue-invading T cells themselves can dynamically regulate their own functions. Both adhesion- and migration-promoting stimuli (i.e., intact IL-2) and suppressive by-products of inflammatory mediators can be present, although not necessarily simultaneously, within the inflammatory milieu (43). At the early stages of inflammation, both IL-2 and elastase may function concomitantly to activate T cells to penetrate tissues. Later, the degraded peptide products of IL-2, generated by elastase, may inhibit T cell migration, inhibit the costimulatory effects of IL-2 and other mediators, and probably signal the termination of the inflammatory reaction.

We thank Prof. Irun R. Cohen, Dr. Barbara Schick, and Dr. Ronen Alon for their helpful comments and discussions.