Internalization of Human Interleukin 4 and Transient Down-regulation of Its Receptor in the CD23-inducible Jijoye Cells*

Human interleukin 4 (IL-4) specifically induces the low affinity receptor for IgE (FctR2/CD23) on the surface of the Burkitt lymphoma cell line Jijoye. At 4 "C 12'I-IL-4 specifically binds to high affinity receptors Following a rapid temperature shift from 4 to 37 "C, 80% of the receptor-bound 1261-LL-4 disappeared from the cell surface within 20 min (tY = 8.9 min). For every two internalized molecules of IL-4 (ts = 13 min), one molecule of IL-4 dissociated from the cell surface (tH = 25 min). More than 90% of the internalized IL-4 was released in a degraded form into the medium following first order kinetics (ts = 68 min). Internalization was inhibited by cytoskeletal disrupting lysosomo- tropic

Jijoye, through specific high affinity receptors (12). The IL-4 concentrations that bind with high affinity to IL-4 receptors (IL-4R) are the same as the concentrations of IL-4 that induce FccR2/CD23 expression in uitro, that is the equilibrium dissociation constant (&) is the same as the median effective concentration.
In this paper, we have extended our studies to a determination of the fate of the receptor-ligand complex in the Jijoye cell line. Here it is shown that at 37 "C, the majority of receptor-occupied 12'I-IL-4 is internalized, then degraded. Meanwhile the IL-4R disappears from the cell surface before being slowly reexpressed in spite of the presence of an excess of IL-4. The reexpression of the IL-4R may be an important step necessary for IL-4 induction of the FccR2/CD23.
Each cell expressed an average of 600-1200 binding sites. "' I-  bound to cells at 4 "C was localized on the cell surface since 90-95% of the radioactivity could be eluted by acid treatment (not shown). Here, experiments were designed to study the fate of cell surface receptor-bound "'I-IL-4 following a temperature shift from 4 to 37 "C ( Fig. 1). Within 20 min of incubation at 37 "C, approximately 80% of the cell surface receptor-bound lz5I-IL-4 disappeared (t% = 8.6 min). By 90 min no cell surface '251-IL-4 was detectable, and concomitantly, there was an increase of internalized '251-IL-4 which reached a maximum at 30 min (60-65% of the initial cell surface receptor-bound lZ5I-IL-4) but then declined to near background levels by 180 min. Intact dissociated "'1-IL-4 also appeared in cell supernatant reaching a maximum of 35-40% at 60 min. A significant amount of degraded lZ5I-IL-4 was observed in supernatants by 30 min (7% of the initially bound , and this level increased steadily, as concomitantly the level of internalized 1251-IL-4 declined. Fig. 2 shows that sodium azide, cytochalasin B, and lysosomotropic agents (NHICl, methylamine, monensin, and chloroquine) were able to inhibit 12'I-IL-4 degradation by 80-100%. Since Jijoye cellconditioned medium is not able to degrade "'1-IL-4 over a 24h incubation time (Fig. 2, supernatant), the results demonstrate that the degradation of '"I-IL-4 is a consequence of the ligand internalization. The data for the dissociation (P) and internalization ( I ) of lz5I-IL-4 can be treated as a system * Portions of this paper (including "Experimental Procedures" and Figs. 2-9) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

6984
Fate of IL-4 a n d Its Rec of simultaneous first order reactions competing for a common reactant (P c S + I) and calculated accordingly ((16) see "Experimental Procedures"). Fig. 3A shows that 80% of receptor-bound '251-IL-4 disappeared from the surface with first order kinetics and a rate constant, ks = 1.3 X s-' (tlh = 8.9 min). Moreover, a plot of dissociated intact lZ5I-IL-4 versus 1e-koffxt ( Fig. 3B) is linear with an intercept close to 0. This indicates that the dissociation of intact '251-IL-4 is also a first order process with an estimated kff = 4.7 X s-' (tH = 25 min). The rate constant, kin (rate of lZ51-IL-4 internalization), was calculated to be 8.8 X W 4 s-' (tlh = 13 min). kin versus koff can also be established by calculating the ratio I/P = 1.7-1.6 which is close to the ratio kin,koff = 1.9. These results indicate that two competitive first order processes exist: (a) internalization and ( b ) dissociation of the lymphokine from receptor. Finally, release of degraded IZ51-IL-4 was also found to follow first order kinetics with a tlh = 68 min ( k h = 1.7 X s-') ( Fig.  3C).
Modulation of the Cell Surface IL-4R Expression by IL-4-Having demonstrated that IL-4 was internalized by Jijoye cells, we wondered whether a subsequent modulation of the expression of the cell surface IL-4R occurred. To this end, Jijoye cells were cultured at 37 "C with an excess of unlabeled IL-4. Residual IL-4R were measured by binding of lz5I-IL-4 after acid elution of cell surface-bound unlabeled IL-4. Addition of unlabeled IL-4 (1 nM) results in a rapid decrease of the IL-4R levels (Fig. 4A). A nadir is reached after 2 h when 75% of the cell surface receptors have disappeared. This decrease of IL-4R did not occur when cells were incubated at 4 "C with 1 nM unlabeled IL-4, demonstrating that downregulation of the IL-4R is temperature-dependent. Scatchard analysis of equilibrium saturation binding data at times 0 and 2 h revealed that the decrease of cell surface-bound lZ5I-IL-4 koff kin :eptor after Endocytosis 6985 was due to a decrease of the number of binding sites/cell from 960 (time 0 ) to 330 (time 2 h) and not to a change in the affinity constant K d = 45 PM (time 0), K d = 59 pM (time 2 h) ( Fig 4B). Interestingly, following a 2-h incubation designed to downregulate the receptor, we found that the IL-4R was reexpressed, in spite of the presence of an excess of IL-4. Jijoye cells cultured for 0 or 24 h with IL-4 or 24 h without IL-4 displayed IL-4R with a similar K d of 45, 31 (Fig. 4B), and 50 PM (not shown), respectively, and also a similar number of binding sites/cell of 960, 820 (Fig. 4B), and 900 (not shown). The expression of IL-4R after 24 h of culture was not due to a decreased ability of the remaining IL-4 to induce receptor down-regulation or internalization: (i) after 24 h of incubation with the cells, the IL-4-containing supernatant was still able to down-regulate (75% at 2 h) IL-4R on another cell batch (not shown); (ii) since the intracellular pool of lZ51-IL-4 remains constant between 13 and 24 h, the rate of '261-IL-4 internalization between 13 and 24 h is equal to the rate of degradation and represents 0.73 molecules/s (2.5% of '251-IL-4 is degraded; Fig. 5). This value is comparable to the rate of internalization calculated when cells encountered IL-4 for the first time: Vi, = kin X S = 8.8 X X 969 = 0.84 molecules/ Reexpression of IL-4R Requires DNA Transcription and Protein Synthesis-The reexpression of cell surface IL-4R was further investigated by culturing Jijoye cells with 1 nM IL-4 together with agents able to block either protein synthesis, DNA transcription, or receptor recycling. The reappearance of IL-4R was inhibited by cycloheximide (CHX) which blocks protein synthesis and actinomycin D which blocks DNA transcription (Fig. 6). CHX was a strong inhibitor since cells cultured for 1-2 h with IL-4 and CHX expressed less IL-4R than cells cultured with IL-4 alone, and under these conditions reappearance of IL-4R after 2 h could not be seen. These data indicate that protein synthesis is required very early in the reexpression of IL-4R. Chloroquine which is known to block receptor recycling (25) delays the reappearance of the 1L-4R in IL-4-treated cultures (Fig. 6).
Removal of IL-4 from Jijoye Cell Cultures Results in a Transient Up-regulation of the IL-4R"Since the rate of internalization of IL-4 over the 13-and 24-h culture period was similar to the internalization rate measured at the origin, we speculated that the reexpression of the IL-4R was due to an increased receptor reinsertion into the plasma membrane. Thus, cells cultured for 24 h in the presence of IL-4 were washed free of IL-4 and then recultured without IL-4. At various time points, aliquots of cells were harvested, acidtreated, and incubated with lz5I-IL-4 to determine the total number of IL-4R. Cells precultured for 24 h with IL-4 display an increased amount (+50%) of IL-4R 2-4 h after the removal of IL-4 ( Fig. 7). This increase of IL-4R is only transient since 6 h after the removal of IL-4, cells express approximately the same number of receptors as control cells. The increased binding of IL-4 reflects an increased number of binding sites as determined by Scatchard analysis 2 h after removal of IL-4 (not shown). Removal of IL-4 from cells which had been cultured with IL-4 for 2 h (a time at which cells express the lowest numbers of receptors) also results in a rapid increase of cell surface IL-4R 2 h later, but in these conditions, cells never expressed more receptors than control cells.
Biochemical Analysis of the IL-4R: Transient Down-regulation by IL-4-Cross-linking of bound lZ5I-IL-4 to Jijoye cells revealed the presence of three labeled polypeptides with M, 130,000,80,000, and 70,000 (Fig. 8, lane I). Formation of these complexes was prevented by inclusion of a 100-fold excess of

9.
Fate of IL-4 and Its Receptor after Endocytosis unlabeled IL-4, indicating a specific saturable interaction between IL-4 and the binding molecules (Fig. 8, lane 2). Preincubation of cells at 37 "C for 20 min ( l a n e 4) or 90 min (lane 5) with 1 nM unlabeled IL-4, followed by acid treatment and binding/cross-linking of '"1-IL-4, demonstrated a strong decrease in the radioactivity of the three bands. Cells preincubated for 24 h with unlabeled IL-4 (lune 6) display the same complexes as cells cultured without IL-4 (lane 7). These results show that the three cell surface IL-4 binding proteins disappear and reappear in a coordinated fashion in response to IL-4 exposure.
Induction of FccR2/CD23 Requires Prolonged Exposure to Il-~--Incubation of Jijoye cells in the presence of IL-4 for 48 h results in the cell surface expression of FccR2/CD23. To determine whether this phenomenon required a short or a prolonged exposure of cells to IL-4, cells were cultured for varying periods of time with IL-4, then recultured without IL-4 so as to reach a total incubation time of 48 h. IL-4 was removed either by centrifugation/washes or by addition of a neutralizing anti-IL-4 rabbit antiserum. Fig. 9 shows that exposure of Jijoye cells to IL-4 for 2-8 h is not enough to induce FctR2/CD23 48 h later. Exposure of cells to IL-4 for 20 h resulted in a near-maximal induction of FccR2/CD23. Thus, induction of FccR2/CD23 on Jijoye cells in response to IL-4 requires a prolonged exposure of cells to the ligand.

DISCUSSION
The present study elucidates the fate of IL-4 and of the IL-4R under "physiological conditions" where IL-4 delivers its biological effects (in this case, induction of FccRZ/CD23). It was shown that a fraction of the cell surface receptor-bound lZ5I-IL-4 internalized while another fraction dissociated into the culture medium. Internalization and dissociation of surface receptor-bound '"I-IL-4 were found to be competitive first order events. The rate of internalization (k, = 8.8 X s-', tJh = 13 min) was twice as fast as the rate of dissociation (brr = 4.7 x s-l, tu = 25 min). Thus, for every three molecules of IL-4 associated to the cell surface receptor, two were internalized while one dissociated. Internalization is likely due to receptor-mediated endocytosis rather than fluid phase pinocytosis since the internalization half-time is of the order of a few min (tnh = 8.9 min) and sodium azide inhibits the appearance of acid-insensitive radioactivity (not shown). Once internalized, lZ5I-IL-4 is degraded (tlh = 68 min). Inhibition of 1251-IL-4 degradation by lysosomotropic agents suggests that 1251-1L-4 degradation occurs in lysosomes. Thus IL-4 seems to be processed like EGF (22), insulin (231, colony stimulating factor 1 (16), IL-2 (26), and [27][28][29]. Internalization of EGF (30) or insulin (31) is accompanied by a down-regulation of their cell surface receptors, whereas internalization of IFN-7 (28) or bombesin-like peptides (32) is not accompanied by a down-regulation of their respective receptors. Here, it is demonstrated that the IL-4R is rapidly downregulated from the cell surface with a nadir reached after 2 h. However, upon prolonged incubation of the cells with IL-4, it was observed that the IL-4R were reexpressed on the cell surface. Since the internalization rate of the occupied receptor, measured between 13 and 24 h (0.73 molecule/s), was found to be close to the internalization rate of the occupied receptor at the origin (0.84 molecule/s), it is concluded that the IL-4R reexpression seen at 24 h is a consequence of an enhanced reinsertion of the unoccupied receptors into the plasma membrane. This is substantiated by the demonstration that removal of IL-4 from the culture medium at 24 h, which decreases internalization, results in a rapid but transient up-regulation of cell surface receptors (Fig. 7).
The increased reinsertion of the receptor into the plasma membrane may come from three sources: recycling, mobilization of an intracellular pool, and/or accelerated synthesis of the receptors. Addition of chloroquine to cells cultured in the presence of IL-4 delayed the reexpression of the IL-4R thus indicating that receptor recycling may be partially involved in the reexpression of the receptor (25). Preliminary experiments showed that the total number of IL-4R measured after solubilization of the cells was equivalent to the number of receptors measured on the surface of intact cells? This observation suggests that the reexpression of cell surface receptors cannot be accounted for by the presence of a secondary readily accessible intracellular pool of receptors. Finally, the addition of cycloheximide prevented the reexpression of IL-4R, demonstrating that protein synthesis is required and that recycling alone is insufficient to replenish the cell surface pool of IL-4R. DNA transcription seems also to be necessary for the reexpression of the IL-4R since treatment of cells with actinomycin D blocks receptor reexpression. Whether CHX and actinomycin D inhibit the transcription and synthesis of a protein required for the transport of the newly synthesized IL-4R to the cell surface cannot be established yet and will require tools which are not presently available (e.g. antibodies against the IL-4R).
Regulation of cell surface IL-4R has also been investigated at the molecular level of cross-linking experiments. '"1-IL-4 was specifically cross-linked to three polypeptides of M, 130,000,80,000, and 70,000. The M , 130,000 protein probably corresponds to the previously identified M, 139,000 protein (33); however, the 80,000 and 70,000 proteins were not described in this earlier study. These two proteins may either represent degradation products of the M, 130,000 polypeptide or may represent polypeptides in an IL-4R complex. Whatever the relationship between these three proteins, the experiments described here clearly show that they are tightly associated since down-regulation and reexpression of cell surface IL-4R correlated with the coordinated disappearance and reappearance of these three proteins on the cell surface.
Earlier we described that the biological effects of IL-4 on normal human B cells required a prolonged contact between IL-4 and the target cells (e.g. induction of proliferation (31, induction of IgG and IgM production (9), inhibition of IL-2dependent B cell proliferation (6)). It is comparable to the 8-12-h EGF exposure time of cells requisite for cell commitment and entry into S phase (34), the continuous presence of IFNa for the inhibition of Daudi cell growth (35), or the minimum 4-h contact necessary for IFN--, to induce tumoricidal activity of murine macrophages (27). On the other hand, the Fcy-and Ia-inducing activity of IFN--, (29) on macrophages was found to require only short periods of incubation (5-30 min) with the ligand. However, the IL-4-mediated induction of FccR2/ CD23 required a prolonged exposure of the cells to saturating concentrations of IL-4 since 8 h of incubation were totally ineffective. Maximal induction of FctR2/CD23 required cells to be in contact with IL-4 for 24 h and was thus comparable, in this respect, to the other biological effects of IL-4. We calculate that during this period of time, lo6 cells (960 sites/ cells) have internalized and degraded (3.5% degraded over 24h incubation, see Fig. 5) 44-fold more IL-4 than the original number of IL-4R expressed at the cell surface. Whether the induction of FccR2/CD23 is directly linked to the amount of internalized and processed IL-4 or to an increase of receptor turnover during the culture period remains to be elucidated.
Biochemical analysis of the reexpressed receptor uersus expressed receptor did not reveal any differences in the struc-J. P. Galizzi, manuscript in preparation. different polypeptide components or may display a very high affinity (Kd > 4 X IO-" M or undissociable lZ5I-IL-4 binding), as was described with EGF (36) and platelet-derived growth factor (37). Additional studies will be required to elucidate whether the reexpressed IL-4R is really identical to the expressed receptor (Le. delivery of different biochemical signals).
As a conclusion, the study established that internalization of IL-4 results in the rapid down-regulation of the IL-4R, followed by a reexpression of the receptors on the cell surface. This latter phenomenon may represent an important step for the delivery of IL-4 biological activity. 14.