Trafficking of interleukin 2 and transferrin in endosomal fractions of T lymphocytes

The interaction of interleukin 2 (IL2) with high-affinity IL2 receptors on the cell surface of T lymphocytes triggers intra- cellular events that lead to T cell proliferation. One of the early events that follows the binding of IL2 to such receptors is the rapid internalization and degradation of the ligand (Robb et al., 1981; Duprez and Dautry-Varsat, 1986; Smith, 1989). The high-affinity receptor complex (K_d≅ 10-100 pM) consists of three distinct receptor components, the α chain of 50-55 kDa, the β chain of 70-75 kDa and the g chain of 65 kDa, which are associated in a non covalent manner (Waldmann, 1991; Taniguchi and Minami, 1993). The g chain is also a component of IL4 and IL7 receptors (Kondo et al., 1993; Noguchi et al., 1993; Russell et al., 1993). In addition to α, β and g, p56^lcktyrosine kinase is associated with the β chain and may partic- ipate in IL2-mediated signal transduction (Hatakeyama et al., 1991; Horak et al., 1991; Minami et al., 1993).

Following IL2 binding to surface high-affinity receptors and endocytosis, IL2 is degraded in acidic intracellular compart- ments, but the pathway it follows to its degradation is not known. Both the α and β chains, and probably g, are internal- ized together with IL2 and can be found inside the cells forming a ternary complex with IL2 (Fung et al., 1988; Duprez et al., 1992).

One consequence of IL2-receptor internalization is to decrease surface high-affinity receptor expression, and thereby to prevent continuous IL2 responsiveness in T cells (Duprez et al., 1988; Smith, 1988). Endocytosis of IL2 receptors, and maybe other growth factor receptors, may also play a role in signal transduction, although it is not well established (Duprez et al., 1991; Futter et al., 1993).

Most receptor-ligand complexes and solutes are endocy- tosed via coated-pits and coated vesicles and can later on be sorted towards degradation or recycling to the plasma membrane (Goldstein et al., 1985; Dunn et al., 1989; Hopkins et al., 1990). Internalized molecules are first observed in, so called, early endosomes from which some receptors and ligands are likely to recycle rapidly to the cell surface. It is well established in numerous cell types that transferrin and its receptors are endocytosed and rapidly recycled (Dautry-Varsat et al., 1983; Klausner et al., 1983), and this receptor-ligand complex is commonly used as a marker of early and recycling endosomes. Molecules destined to degradation are then trans- ported to late endosomes and to lysosomes. Horseradish per- oxidase (HRP), endocytosed in the fluid phase, can be used as a convenient marker of the whole endocytic pathway, allowing the definition of first early and then late endosomes and lysosomes, depending on the time of internalization (Griffiths et al., 1989; Gruenberg et al., 1989).

We have analyzed the distribution of IL2 in endocytic organelles after endocytosis and compared it to that of trans- ferrin and the fluid-phase marker HRP in the human T lym- phocytic cell line IARC 301.5. These cells are cloned, express about 3000 high-affinity IL2 receptors, internalize IL2 and pro- liferate in response to this growth factor (Duprez et al., 1985, 1988).

For that purpose, we have used a method based on endosome separation depending upon their density in a sucrose gradient (Gorvel et al., 1991).

The human tumor T cell clone IARC 301.5 (Duprez et al., 1988) was cultured in RPMI supplemented with 10% fetal calf serum, 1 mM L- glutamine, 20 mM Hepes, pH 7.2.

Pure recombinant IL2 was obtained from SANOFI (France). IL2 was radiolabelled with ¹²⁵I by the chloramine T method to a specific activity of 30-100×10⁶cpm/μg. Transferrin (Sigma) was loaded with iron and radiolabelled with ¹²⁵I by the chloramine T method (Ciechanover et al., 1983) to a specific activity of 1×10⁶cpm/μg. When cell fractionation was performed, radiolabelled molecules were used as tracers and diluted with unlabelled molecules. Protease inhibitors were from Sigma and HRP from Boehringer.

IARC 301.5 cells were washed with RPMI, 20 mM Hepes, pH 7.4. The cells were incubated at 37°C with 500 pM ¹²⁵I-IL2 or 100 nM ¹²⁵I-transferrin in RPMI, 20 mM Hepes, 1 mg/ml bovine serum albumin (BSA), pH 7.4 (2×10⁷cells/ml). At such concentrations, surface high-affinity IL2 (Duprez et al., 1988, 1991) and transferrin receptors (our unpublished data) present on IARC 301.5 cells were saturated. Non-specific binding of radiolabelled ligands was measured in the presence of a 100-fold molar excess of the same cold ligand and was below 4%. Ligand uptake was stopped by adding ice-cold RPMI, 20 mM Hepes, pH 7.4. When the ligand was subsequently chased, the cells were centrifuged and washed twice in the same medium at 4°C. A second incubation at 37°C was performed as above, except that labelled ligand was replaced by unlabelled ligand. Cell surface-associated ¹²⁵I-IL2 was removed by two successive acid pH treatments as described previously (Duprez et al., 1988). The efficiency of ¹²⁵I-IL2 removal was measured by saturating surface high-affinity receptors for 90 minutes at 4°C with ¹²⁵I-IL2 prior to acid pH treatment. In such conditions 98.5% of the bound radioactivity was removed. For ¹²⁵I-transferrin, one such treatment was performed that removed 98.8% of the surface-associated radioactivity.

The cells were washed with RPMI, 20 mM Hepes, pH 7.4, and incubated at 37°C in RPMI, 20 mM Hepes, 1 mg/ml BSA, pH 7.4, containing 5 mg/ml HRP (3×10⁷cells/ml). When HRP uptake was followed by a chase without HRP, the cells were washed in RPMI, 20 mM Hepes, 1 mg/ml BSA, pH 7.4, at 4°C, and further incubated in the same medium at 37°C. Extracellular HRP was removed by five washings and enzymatic activity present inside the cells was assayed as described (Goud et al., 1984). Internalized HRP was quantified by assaying HRP on intact cells (surface) and on cells lysed with 0.5% NP40 (surface + interior). Despite the extensive washing procedure, a small amount of HRP remained absorbed on the cell surface. After only 5 minutes uptake at 37°C, this amount represents 25% of cell-associated HRP and smaller proportions as the time of HRP internalization or chase increases. When cells were fractionated on a sucrose gradient, control cells incubated for 1 hour at 4°C with HRP were also fractionated to determine the distribution, in the gradient, of HRP associated with the plasma membrane. The amount of internalized HRP in each fraction was calculated after correction for surface-associated HRP found in the same fraction. HRP in gradient fractions was assayed in the presence of 0.5% NP40.

Cells (5-7×10⁷cells per gradient) were incubated with endocytic marker (¹²⁵I-IL2, ¹²⁵I-transferrin or HRP) and washed in homogenization buffer (250 mM sucrose, 3 mM imidazole, HCl, pH 7.4) at 4°C. The cells were then resuspended in 0.5 ml of homogenization buffer containing protease inhibitors (10 μg/ml aprotinin, 1 μg/ml pepstatin, 1 μg/ml antipain) and 1 mM EDTA. Cells were broken by 16 passages through a 26 G 1/2 needle fitted to a 1 ml plastic syringe. Gentle homogenization conditions were used to limit damage to endosomal elements (Gorvel et al., 1991). About 30% of the cells remained intact as shown by phase-contrast microscopy. The homogenate was centrifuged for 5 minutes at 1200 gand the postnuclear supernatant was further fractionated on a flotation gradient as described (Gorvel et al., 1991; Gruenberg and Gorvel, 1992). Briefly, postnuclear supernatant (0.4 ml) was brought to 42% sucrose by addition of 0.4 ml of 75% sucrose, 3 mM imidazole, HCl, pH 7.4, and loaded at the bottom of a 5 ml ultracentrifuge tube. The postnuclear supernatant was sequentially overlaid with 1.5 ml 36% sucrose, 3 mM imidazole, HCl, pH 7.4, containing O.5 mM EDTA, followed by 1 ml 25% sucrose, 3 mM imidazole, HCl, pH 7.4, containing 0.5 mM EDTA and 0.6 ml homogenization buffer. After one hour centrifugation at 125,000 gin a TST 60.4 rotor (Kontron), fractions of 0.32 ml were collected from the bottom. To control that the endocytic marker in the postnuclear supernatant was associated with intracellular vesicles, and not with soluble material, the postnuclear supernatant was centrifuged for 30 minutes at 157,000 gin a TST 60.4 rotor in a TL 100 centrifuge and the endocytic marker assayed in the pellet (vesicular material) and in the supernatant (soluble material). The sucrose density of each fraction was determined using a refractometer, and the protein concentration according to Bradford (1976). Lysosomal enzymes β-N-acetyl-glucosaminidase (EC 3.2.1.20) and cathepsin B (EC 3.4.22.1) were assayed as described (Sannes et al., 1986; Beardmore et al., 1987).

IARC 301.5 cells were incubated for 15 minutes at 37°C with a concentration of ¹²⁵I-IL2 such as to bind only to high-affinity receptors. After this incubation allowing IL2 endocytosis, the cells were washed at 4°C to stop internalization, and cell surface bound ¹²⁵I-IL2 was removed by acid wash. After this treatment, the radioactivity associated with the cells represents only internalized IL2, and no ligand remains on the plasma membrane. The cells were then broken and, after removal of nuclei, the cell extracts were fractionated with a sucrose step gradient (Materials and Methods). This method was reported to separate two classes of endosomes, so called early and late endosomes, in baby hamster kidney cells (BHK) (Gorvel et al., 1991). Transferrin has been extensively used as a marker of receptor-mediated endocytosis in numerous cell types. It is found essentially in early and recycling endocytic compartments and is one of the best markers for these compartments. The cells were treated with ¹²⁵I-transferrin as just described for ¹²⁵I-IL2. Fig. 1Bshows that both radiolabelled IL2 and transferrin are contained mostly at the 25-36% sucrose interface in one peak (called peak II) with a density between 30 and 35% in sucrose. The ¹²⁵I-IL2 represents intracellular IL2 associated with vesicular compartments, and not dissociated or degraded soluble IL2. Indeed the postnuclear supernatant contained 98% ¹²⁵I-IL2 pelleted at 157,000 g, and less than 2% soluble radioactive material. The same holds true for ¹²⁵I-transferrin. The same peak contains IL2 and transferrin and is formed at similar sucrose densities as the early endosomes isolated by Gorvel et al. in BHK cells at the 25-35% sucrose interface in H₂O (10-16% sucrose in heavy water) (Gorvel et al., 1991; Gruenberg and Gorvel, 1992).

Fluid-phase endocytosis markers are internalized in endocytic vesicles. Part of the material is then degraded and part is rapidly recycled back to the extracellular medium. Fluid-phase markers can be found all along the endocytic pathway. For instance in BHK cells, after 5 minutes endocytosis HRP is found mostly in early and not in late endosomes, while after 5 minutes endocytosis followed by 20 minutes chase, it is found in late more than in early endosomes (Gorvel et al., 1991).

Fluid-phase endocytosis of HRP was measured in IARC 301.5 cells. These cells internalize 5 ng HRP/mg protein per minute when incubated with 5 mg/ml HRP at 37°C. Fluid-phase uptake in these T lymphocytes is similar to that of BHK cells (Griffiths et al., 1989; Gruenberg et al., 1989). IARC 301.5 cells were fractionated after HRP uptake for 5 minutes at 37°C, or 5 minutes at 37°C followed by a 25 minute chase at 37°C (Fig. 2). After 5 minutes HRP uptake and 25 minutes chase, about 2-fold less HRP was found inside the cells than without chase, indicating that a good proportion of internalized HRP is recycled back to the medium. With or without chase, HRP was found in peak II between 30 and 35% sucrose, corresponding to early endosomes, in peak III between 18 and 26% sucrose, and peak I above 37% sucrose, which contains most of the lysosomal markers (Fig. 1C). The postnuclear supernatant layered on the gradient contained 75% HRP that could be pelleted at 157,000 g, indicating that it was vesicular. After gradient collection, 55 and 66% of the fractions in peak II and III were pelleted at 157,000 g. Less than 5% of peak I was pelleted indicating that it contained mostly soluble HRP. Peak III corresponds to the late endosomes recovered in the upper region of the 25% sucrose cushion in H₂O (10% sucrose in heavy water) (Gorvel et al., 1991; Gruenberg and Gorvel, 1992). In these T lymphocytes, HRP is found in low sucrose density vesicles (peak III) as early as after 5 minutes uptake. After 25 minutes of chase it is in peak II and peak III, with a larger proportion in peak II. Similar distributions of HRP between peak II and peak III were also found after a 15 minute uptake followed by 30 minutes chase and after overnight uptake of HRP. In a representative experiment, the yields of HRP after 15 minutes uptake were 24 and 3% in peak II and peak III, respectively, 21 and 4% after 15 minutes uptake and 30 minutes chase, and 23 and 5% after 17 hours uptake. Thus in IARC 301.5 cells HRP is found in peak II and peak III, contrary to IL2, which is found only in peak II.

Transferrin is internalized and recycled very efficiently to the cell surface within 10-15 minutes, and recovered as undegraded apotransferrin in the extracellular medium (Dautry-Varsat et al., 1983; Klausner et al., 1983). The internalization and recycling of transferrin in IARC 301.5 cells was studied. For that purpose, after a 5 minute incubation at 37°C with ¹²⁵I-transferrin, the radiolabelled ligand was washed away, and the cells were further incubated at 37°C for different times (Fig. 3A). Internalized ¹²⁵I-transferrin was rapidly recycled with half of the internalized transferrin getting out of the cells in less than 10 minutes, as undegraded, TCA-precipitable material.

The distribution of ¹²⁵I-transferrin was analyzed by subcellular fractionation of cells loaded with ¹²⁵I-transferrin as in Fig. 3A. ¹²⁵I-transferrin remaining on the cell surface was washed away by acid pH treatment before cell fractionation. Fig. 3Bshows the amounts of ¹²⁵I-transferrin associated with peaks I, II and III of the gradients after different times of chase at 37°C. ¹²⁵I-transferrin was never found in peak III. This shows that peak II indeed corresponds to early and recycling endocytic compartments in these T cells, as is the case in other cell types (Gorvel et al., 1991; Papini et al., 1993). Transferrin was found only in peak II, and the amount present in peak II decreased within minutes in agreement with the rapid recycling of trans-ferrin.

IL2 is internalized after binding to surface receptors but, in contrast to transferrin, it is degraded and not recycled (Duprez and Dautry-Varsat, 1986). The kinetics of endocytosis and sub-cellular localisation of IL2 were investigated in order to compare the endocytic pathway of IL2 to that of transferrin. The kinetics of IL2 internalization and degradation in IARC 301.5 cells is shown in Fig. 4A. In this experiment, after a 5 minute incubation with ¹²⁵I-IL2 at 37°C, the cells were washed and further incubated at 37°C. IL2 was internalized and was later on found in the medium as degraded, TCA soluble, material. The kinetics of IL2 entry is markedly slower than that of transferrin (Fig. 4Acompared to Fig. 3A).

The distribution of IL2 was analyzed by subcellular fractionation of cells loaded with ¹²⁵I-IL2 as in Fig. 4A, after removing plasma membrane-associated ¹²⁵I-IL2 by acid pH wash. Fig. 4Bshows the amounts of ¹²⁵I-IL2 associated with peaks I, II and III of the gradients after different times of chase at 37°C. The maximum amount of ¹²⁵I-IL2 in peak II was found at about 15 minutes of chase, a time clearly longer that for transferrin where the maximum was already reached after the 5 minute pulse, in agreement with slower kinetics of entry for IL2 than for transferrin. The amount of ¹²⁵I-IL2 in peak I increased to reach a maximum at about 30 minutes. Very little labelled material was found in peak III at any time, without any enrichment of this compartment in labelled material as chase time increased and as degraded IL2 got out to the medium. IL2 was degraded intracellularly and accumulated in the culture medium in a TCA soluble form (Fig. 4A). Therefore it appears that IL2 is first found mainly in early transferrin-containing endosomes and later on, between 30 and 60 minutes chase, IL2 is predominantly found in dense fractions containing lysosomes.

The subcellular distribution of IL2 internalized after binding to cell surface high-affinity receptors was examined in the human T cell line IARC 301.5. We used a method that has been described to separate early and late endosomes on the basis of their density in a sucrose gradient. IL2 was found in a subcellular endosomal fraction that sediments at a density of 30 to 35% sucrose (peak II). This fraction is the one that contains transferrin a marker of early and recycling endosomes. In other cells, such fractions have been shown to contain early endosomes as defined by morphology, the presence of the small GTP-binding protein rab 5 and that of the fluid-phase marker HRP after 5 minutes endocytosis (Gorvel et al., 1991; Papini et al., 1993). IL2 is endocytosed slower than transferrin in these cells and is found in this fraction later: the maximal amount is observed when the 5 minute pulse is followed by a 15 minute chase, while transferrin is maximal in this fraction before the end of the 5 minute pulse. IL2 also gets out of this fraction slower than transferrin, with a half-time of roughly 25 minutes as compared to about 10 minutes for transferrin (Fig. 4Bcompared to Fig. 3B). The fast kinetics of entry and recycling of transferrin from endosomes is in agreement with previous data on other cells (Yamashiro et al., 1984; Stoorvogel et al., 1987). IL2 is in fractions with the same density as transferrin. This suggests that IL2 and transferrin are colocalized. It has been described that when low density lipoproteins (LDL) are internalized in endosomes, as a consequence the density of these organelles diminishes (Beaumelle and Hopkins, 1989). When cells that had accumulated LDL and ¹²⁵I-IL2 (or ¹²⁵I-transferrin) for 10 minutes at 37°C were fractionated according to Beaumelle et al. (1989), the density of compartments containing both IL2 and transferrin decreased (not shown). Thus it appears that IL2, as well as transferrin, at some point during endocytosis, is in the same endosome as LDL internalized for 10 minutes.

The subcellular fractionation method described by Gorvel et al. (1991)allows us to also separate late endosomes. HRP was used to identify such compartments after fractionation of IARC 301.5 cells and was, as expected, found in a fraction that sediments at a density of 18-26% sucrose (peak III). Small amounts of cathepsin B were found in this peak while no β-N-acetyl-glucosaminidase was detected, in agreement with the reported presence of cathepsin B in late endosomes (Griffiths, 1992). This fraction has a similar sucrose density to that of late endosomes isolated in BHK cells, containing rab 7 and the cation-independent mannose 6-phosphate receptor. Transferrin was never found in peak III, showing that peak III does not contain contaminating early endosomes. In IARC 301.5 cells, HRP was found in peaks II and III as in other cells. The enrichment factor, i.e. the ratio between the specific activity in the peak and that in the homogenate, was about 7 in both peaks after 5 minutes uptake as well as after 5 minutes uptake and 25 minutes chase. The relative distribution of HRP between peak II and peak III did not show an enrichment of peak III as chase time increased (Fig. 2). Even after 1 hour or 17 hours of incubation of the cells with HRP, no relative increase of peak III and no enrichment in this peak could be detected. HRP was clearly enriched after chase, in this late endosome fraction in BHK cells, and less in Vero cells (Gorvel et al., 1991; Papini et al., 1993). Also, during the 25 minute chase that follows the 5 minute pulse of HRP, about half of HRP has recycled to the medium (Fig. 2) and this recycling may account for the lack of HRP enrichment of peak III after chase. Nevertheless, HRP is clearly found in peak III, corresponding to some late endosomes, while IL2 is not.

Several possible interpretations may explain why IL2 is detected very weakly, if at all, in peak III. The corresponding compartments may represent a minority in T lymphocytes since, even in the case of HRP, peak II contains more internalized material than peak III at any time. Also, IL2 may go through such late compartments very rapidly or mostly not go through these compartments. Finally the density of this compartment in a sucrose gradient may differ in lymphocytes and fibroblasts. At early times of endocytosis, IL2 is mainly in the early endosomes fraction, with a peak at 20 minutes. As time increases, it is mostly in dense fractions containing lysosomes before it gets degraded and is recovered in soluble form in the medium.

Intracellular trafficking of molecules during endocytosis has been extensively characterized. Sorting of recycling components (e.g. transferrin, the transferrin or LDL receptors from components that are degraded during endocytosis (e.g. LDL,α₂-macroglobulin, asialoglycoprotein or epidermal growth factor, EGF) has been studied in many cell types. This sorting is usually rapid, taking place within a few minutes of internalization, as measured both by subcellular fractionation (Stoorvogel et al., 1987; Schmid et al., 1988) and microscopy (Geuze et al., 1984; Dunn et al., 1989). Early sorting between IL2 and transferrin was hardly detected by the fractionation methods used in T lymphocytes. This is different from what has been observed with the well characterized growth factor EGF, for which sorting from transferrin appears to be efficient and rapid (Hanover et al., 1984; Gorman and Poretz, 1987).

From the data presented here and previous experiments (Duprez and Dautry-Varsat, 1986; Hémar and Dautry-Varsat, 1990; Duprez et al., 1992; Ferrer et al., 1993) the intracellular traffic of IL2 and its receptors appears as follows. IL2 binds to its high-affinity receptors on the cell surface and it is internalized bound to the α and β chains of the receptors, and probably also to the g chain (Voss et al., 1993). Internalized IL2 is located in early and recycling endosomes. During the time when it is in such organelles, IL2 can be detected by crosslinking studies still bound to the α and β chains, and maybe g, of the receptor. Later on, IL2 is degraded in lysosomal compartments. The β chain of the receptor is also degraded, while the α chain appears to be sorted to the recycling pathway (A. Hémar et al. unpublished work). Endocytosis and sorting of the different components of IL2 and its receptors is complex. Targeting of these components to different intracellular compartments may be critical for the regulation of expression of each component and for their interaction with molecules involved in signal transduction.

We thank Drs B. Beaumelle, J. C. Antoine and P. Courtoy for their advices on cell fractionation, peroxidase and lysosomal enzymes assays. We also wish to thank B. Beaumelle and A. Hémar for critically reading the manuscript. We are grateful to V. Cornet for skillful technical assistance and to K. Thébaud for typing the manuscript. This work was supported by the Agence Nationale de la Recherche sur le Sida, the Association pour la Recherche sur le Cancer, the Ligue Nationale Française contre le Cancer and EEC BIO2-CT92-0164.