Eur. J. Immunol. 1992. 22: 2071-2076
Hichard J. Armitage., Timothy A. Sato., Brian M. Macduff., Kv N. Clifford., Alan R. Alpert., Craig A . Smith. and William C. Fanslow. Departments of ~ ~ and Bjochemistryv, I~~~~~~ Research and Development Corporation, Seattle
CD40 hgdnd
Identification of a source of biologically active CD40 ligand We have identified the murine thymoma line EL4 as a source of biologically active CD40ligand. Using a biotin-labeled soluble CD40.Fc fusion protein, consisting of the extracellular domain of human CD40 and the Fc region of human IgGl , EL4 cells cell sorting with ~ ~ were subjected ~ t o repeated ~ flow cytometric ~ l to select forcells ~ enhanced biotinylated CD40.Fc binding. After nine rounds of sorting, the number of CD40.Fc binding sitedcell had risen from 450 on the unsorted parental EL4 cells to 15 000 on EL40.9 cells (EL4 cells sorted with biotinylated CD40.Fc for nine rounds). Scatchard analysis of radiolabeled CD40.Fc binding revealed that the surface-expressed CD40 ligand on parental EL4 and EL40.9 cells bound its receptor with a single class of high-affinity sites ( K d = 0.5 nM). Supernatant (SN) from the sorted EL40.9 cells was found to contain human and murine B cell stimulatory activity which could be removed by preclearing with immobilized CD40.Fc, confirming the presence of soluble CD40 ligand in the preparations. EL40.9 supernatant enhanced soluble CD23 (sCD23) release and induced IgE secretion from interleukin 4-stimulated human B cells. In addition, EL40.9 SN contained proliferative activity for anti-IgM-activated murine B cells which could be removed by treatment with immobilized CD40.Fc. However, the same SN had no demonstrable activity on the proliferation of human B cells. The results presented here describe, for the first time, a source of membrane-bound and soluble CD40 ligand. The soluble form of this murine ligand has activity on murine and human B cells and induces some of the functional responses predicted for the ligand based on the action of stimulatory antibodies directed against the human CD40 surface molecule.
1 Introduction The Bcell surface molecule CD40 has been shown to mediate various biological effects induced by binding of specifically directed monoclonal antibodies (mAb). These include : the induction of increased cell size [I, 21; homotypic and heterotypic adhesions [2, 31; entry into cell cycle of Bcells preactivated with anti-IgM, CD20 mAb or phorbol ester either alone [4-61 or in synergistic combination with IL-4 [2, 71; secretion of IgE [S-101. IgG and IgM [lo] from IL-4-stimulated B cells in the absence of Tcells; and enhancement of levels of surface-expressed and soluble CD23 released in the presence of IL-4 [11, 12). Human B cell lines have been derived from primary B cell populations grown in the presence of CDw32+ adherent cells, IL-4 and CD40 antibody [13]. Furthermore, germinal center centrocytes can be prevented from undergoing apoptosis by activation through both CD40 and receptors tor antigen 1141, while binding of CD40 mAb to immature B cells at various stages of maturity, as well as t o activated mature cells, results in enhanced tyrosine phosphorylation [IS], demonstrating that the CD40 surface molecule is active in signal transduction at multiple stages of B cell ontogeny. [I 103261 Correspondence: Richard J. Armitage. Dcpartrnent of Immunology. Immunex Research and Development Corporation. 51 University Street, Seattle, WA 98101, USA Abbreviations: tein
2071
CD40.Fc: Soluble CD40/human IgCl fusion pro-
0 VC'H Vcrlagsgeqellschaft mbH, D-6940 Weinheim, 1992
Cross-linking of CD40 is required for optimal stimulation of B cells [6]. Fab fragments of CD40 mAb are refractory to the proliferative signal delivered by intact antibody and, on their own, impart only a weak stimulatory signal. The sensitivity with which B cells respond to nanomolar concentrations of CD40 mAb [2, 41 has led to speculation that the CD40 surface molecule may be a receptor for an as-yet-unidentified ligand. Recently Gascan and colleagues [16] have shown that human Bcells can be induced to proliferate and secrete IgE in the presence of activated CD4+ T cell clones and IL-4, and that the signal provided by T cells can be effectively replaced by CD40 mAb [ 171. The cloning of CD40 [18] revealed homology between the extracellular domains of CD40 and nerve growth factor (NGF) receptor [19, 201. Similar homology with two distinct tumor necrosis factor (TNF) receptors [21-231 formed the basis of the CD40/TNFR superfamily.The argument for a ligand receptor role for CD40 is strengthened by the fact that three members of this family are receptors for cloned ligands. We have recently reported the expression of soluble constructs of the extracellular domain of human CD40, both as a 28-kDa soluble CD40 molecule and as a 57-kDa fusion protein consisting of soluble CD40 and the Fc region of human IgGl [24]. Both forms of soluble CD40 (sCD40 and CD40.Fc) were found to inhibit IL-4-mediated release of sCD23 and secretion of IgE in the absence of added CD40 mAb, demonstrating the induction and involvement of CD40 ligand in these responses. In this report we have used biotin-labeled CD40.Fc to identifv the murine thvmoma cell line EL4 as a source of C D k ligand. Replated rounds of flow cytometric cell sorting using biotinylated 0014-2O80/92/0808-207I$3.50 + ,2510
2072
Eur. J. Immunol. 1992. 22: 2071-2076
R. J. Armitage. T. A. Sato, B. M. Macduff et a1
CD40.Fc resulted in high levels of CD40 ligand expression on the cell surface and soluble ligand in the supernatant (SN). SN from CD40.Fc-sorted cells contained stimulatory activity for both murine and human B cells which could be removed by pretreatment with immobilized CD40.Fc, but not with control human IgGl antibody, confirming the presence of biologically active CD40 ligand in these preparations.
2 Materials and methods 2.1 Cell separation
Human peripheral blood mononuclear cells (PBMC) were isolated from the blood of healthy donors by centrifugation over Histopaque (Sigma Chemical Co., St. Louis, MO). T cell-depleted preparations (E-) were obtained by removal of T cells by rosetting with 2-aminoethylisothiouronium bromide (AET)-treated sheep red blood cells (SRBC) and further centrifugation over Histopaque.Tonsil1ar tissue was gently teased and the resulting cell suspension centrifuged over Histopaque. Purification of B cells was achieved by removal of cells rosetting with AET-treated SRBC and treatment of remaining cells with B cell lympho-kwik (One Lambda Inc., Los Angeles, CA) for 1 h at 37°C to lyse contaminating non-B cells. The resultant B cell population was >98% CD20+ as determined by flow cytometry (FACScan, Becton Dickinson, San Jose, CA) of cells stained with Leu-16 mAb (Becton Dickinson). Murine splenic B cells wcre isolated using anti-T cell antibody and complement, followed by passage over Sephadex G-10 (Pharmacia, Piscataway, NJ) to remove adherent cells. Bcells were then positively selected by panning on plates coated with 5 pg/ml goat anti-mouse IgM (Southern, Birmingham, AL). Isolated cells were > 95% surface IgM+ as determined by flow cytometry.
2.2 Reagents The 57-kDa gene product of the extracellular domain of human CD40 fused to the Fc region of human IgGl (CD40.Fc) was expressed and purified as described previously [24]. CD40.Fc or control human IgGl were biotinylated as follows; 50 pg protein (200 pg/ml in 0.1 M NaHC03 p H 8.3) was incubated with 1 pg (1mg/ml in DMSO) biotin-x-nonhydroxysuccinimide (NHS, Calbiochem, La Jolla, CA) for 30 min at room temperature. At the end of the incubation period, the reaction mixture was microfuged through a 1-ml Sephadex G-25 (Pharmacia) desalting column and thc eluate adjusted to 100 yg/ml in PBS plus 0.02% NaN3. Protein concentration of biotinylated CD40.Fc and hIgG 1 was determined by micro-BCA assay (Pierce, Rockford. 1L) with ultrapure bovine serum albumin as standard. Murine and human recombinant IL-4 were purified from yeast SN as previously described [25, 261. For proliferation experiments, murine and human B cells were activated with soluble goat anti-murine IgM (Southern) and rabbit anti-human IgM bound to acrylamide beads (Bio-Rad, Richmond, CA), respectively. Affinity-purified IgGl CD40 mAb G28-5 [4] was a generous gift from Dr. E. A. Clark.
2.3 Flow cytometric cell sorting The murine thymoma cell line EL4 was obtained from American Type Culture Collection (Rockville, MD). Cells were incubated with biotinylated CD40.Fc ( 5 pg/ml plus 1 X lo7 cells) or biotinylated hIgG1 (5 pg/ml plus 1 X lo6 cells) for 30 rnin at 4°C. After washing twice in PBS, streptavidin-phycoerythrin (PE) (Becton Dickinson) :!i!uted 1: 5 in PBS was added to cells for 30 rnin at 4°C. Cells were washed twice in PBS, resuspended to 3 X 106/ml and the cells stained with biotinylated CD40.Fc were subjected to flow cytometric cell sorting on a FACStar plus (Becton Dickinson) using cells stained with biotinylated hIgGl as a control for non-specific binding. For the first four rounds of sorting, the 2% of cells exhibiting the highest intensity of staining with biotinylated CD40.Fc were collected. For subsequent rounds of sorting, the brightest staining 1%of cells were col1ected.Typically 1 X lo4 - 4 x lo4 cells were recovered after each sort. Unsorted EL4 and sorted cells were maintained in DMEM containing 5% fetal bovine serum (FBS), 50 p~ 2-mercaptoethanol, 1 mM sodium pyruvate, 2 mM L-glutamine, 20 mM Hepes, penicillin and streptomycin. Supernatant preparations were harvested from cells for use in functional assays when cell density was optimal (I 8 x 105 cells/ml). Concentration of SN was performed by centrifugation over Centriprep 10 concentrators (Amicon, Beverly, MA). In selected experiments, EL40.9 SN was pre-cleared of CD40 ligand activity by addition of biotinylated CD40.Fc ( 5 pg/ml) for 30 min at 4 "C. Streptavidin-coated Dynabeads (Dynal Inc., Great Neck, NY) were then added (50pl beaddm1 SN) for a further 30 min at 4"C, after which time the beads were removed by three cycles of placing SN in a magnetic field. In all cases control SN was prepared by pre-treatment with 5 pg/ml biotinylated hIgG1 under identical conditions. 2.4 Radiolabeling of CD40.Fc Radioiodination was effected with 5 pg of purified protein in PBS using a solid-phase chloramine T analogue (Iodogen, Pierce, St. Louis, MO), as described [23]. Radiolabeled stocks (5 X lop8 M) were stored at 4°C in RPMI 1640 containing 2% BSA, 20 mM Hepes buffer, 0.2% sodium azide, pH 7.2 (binding media). 2.5 Binding assays and data analysis Binding assays were carried out with serially diluted 1251-labeledCD40.Fc (1251-CD40.Fc)and either unsorted EL4 or sorted EL40.9 cells (3 x 106/point) in binding media at 4°C for 4 h. Aliquots were microfuged through a phthalate oil mixture to separate free and bound lZ5ICD40.Fc, and the SN and pellets counted. Nonspecific binding was measured by the inclusion of a 100-fold molar excess of unlabeled CD40.Fc. Binding data was plotted in a Scatchard format using linear least squares analysis.
2.6 Culture conditions All human cell cultures were conducted in RPMI containing 100 U/ml penicillin, 100 pg/ml streptomycin, 2 mM Lglutamine and 10% heat-inactivated FBS (Intergen, Pur-
CD40 ligand
Eur. J. Immunol. 1992. 22: 2071-2076
chase, NY) at 37°C in a humidified atmosphere of 10% COZ.For measurement of proliferation, 1 X lo5 cells/well were cultured in triplicate in flat bottomed 96-well microtiter plates (Corning, Corning, NY) for 72 h in the presence of the appropriate additives as described in Sect. 3. For determination of sCD23 and IgE production, 1 X lo5 PBM or E- cells were cultured in round-bottom 96-well microtiter plates in the presence of 5 ng/ml hIL-4 for the time indicated. Murine B cells (1 X lo5 cells/well) were cultured in flat-bottom 96-well microtiter plates in RPMl 1640 supplemented with 5% FBS (Hazelton, Lenexa, KS), 1mM sodium pyruvate, 0.1 mM non-essential amino acids, 100 U/ml penicillin, 100 pg/ml streptomycin, 2 mM L-glutamine, and 50 pM 2-mercaptoethanol. Cultures were maintained at 37°C in 10% COZ for 72 h. Human and murine cclls were pulsed with 1 pCi/well tritiated-thymidine ([’HIdThd; 25 Ci/mmol, Arnersham, Arlington Heights, IL) for the final 8 h of culture. Cells were harvested onto glass fiber disks with an automated cell harvester. Incorporated cpm were measured by liquid scintillation spectrometry. For sCD23 induction and IgE secretion, 1 X los cells/well were cultured in triplicate in round-bottom 96-well (Nunc microtiter plates (Intermountain Scientific Corp., Bountiful, UT) for the appropriate time in the presence of additives as detailed in Sect. 3.
2.7 Detection of soluble CD23 and IgE Soluble CD23 levels were determined by ELISA using a commercial sCD23 detection kit (The Binding Site, San Diego, CA). The limit of sensitivity of the sCD23 ELISA was 500 pg/ml. IgE was detected by ELISA as follows: flat-bottom 96-well microtiter plates (Corning) were coated with mouse mAb anti-human IgE (Zymed, San Francisco, CA) at 1 : 500 dilution in PBS. After washing three times, a blocking step was carried out using 5% non-fat dried milk. This was followed by the addition of a titration of IgE standard (Calbiochem, La Jolla, CA) or test supernatants. After washing three times, biotinylated goat anti-human IgE (Kirkegaard and Perry, Gaithersburg, MD) was added at 1: 500 dilution followed by further washing and addition of streptavidin-horseradish peroxidase (Zymed) 1: 500 dilution. After washing, the reaction was developed using 3,3’,5,5’-tetramethylbenzidine(TMB) substrate (Kirkegaard and Perry) and absorbance measured at 520 nm. All incubation steps were performed using a volume of 100 pl/well for 1 h at room temperature.Wash steps were carried out in PBS plus 0.05% Tween. The sensitivity of the IgE ELISA was 100 pg/ml.
2073
Fluorescence Intensity
Figure 1. Flow cytometric analysis of EL4 and EL40.9 cells. Unsorted EL4 cells (A) and sorted EL40.9 cells (B) were stained with 5 pglml biotinylated human IgG1 (-----) or biotinylated CD40.Fc (-) and streptavidin-PE. EL40.9 cells were also stained with 5 pglrnl biotinylated CD4O.Fc in the presence of a 30-fold excess of unlabeled CD40.Fc (.....) .
cells stained with biotinylated CD40.Fc were subjected to cell sorting, and the 1%-2% of cells showing highest intensity of staining were collected. Cell sorting was repeated over a period of several months with 1 X lo4 4 X lo4 cells typically recovered after each sort. After nine rounds of sorting, the intensity of staining with biotinylated CD40 had increased greatly (Fig. 1 B ) over that seen on unsorted EL4 cells. Binding of biotinylated CD40.Fc to EL40.9 cells (EL4 cells sorted for CD40.Fc binding for nine rounds) could be effectively competed by the addition of a 30-fold excess of unlabeled CD40.Fc. Quantitative analysis of 1251-CD40.Fcequilibrium binding with unsorted and sorted cells is shown in Fig. 2. While unsorted EL4 cells had approximately 500 CD40.Fc-binding sitedcell (Fig. 2 A), sorted EL40.9 cells exhibited 1.6 X lo4sitedcell (Fig. 2B).
N
0 x h
3 Results 3.1 Flow cytometric staining with biotinylated CD40.Fc The fusion protein CD40.Fc was labeled with biotin and used for flow cytometric analysis of a range of human and murine cell lines to examine cell surface expression of CD40 ligand. All human cell lines examined lacked detectable binding of biotinylated CD40.Fc when compared to cells incubated with biotinylated human IgGl as a specificity control (data not shown). The murine thymoma line EL4 was found consistently to bind biotinylated CD40.Fc with very low intensity (Fig. 1 A ) . Based on this finding, EL4
I
I
I
1
2
3
C(M) x i 0 9 Figure 2. Equilibrium binding of IZ5I-CD40.Fcto EL4 and EL40.9 cells. Serially diluted ‘2sI-CD40.Fc was incubated with either unsorted (A) or sorted EL4 cells (B) for 4 h at 4°C. Binding was assayed as described in Sect. 2.5. All binding was corrected for nonspecific binding. Inserts represent data replotted in Scatchard format.
2074
Eur. J. Immunol. 1992. 22: 2071-2076
R. J. Armitage.T. A. Sato, B. M. Macduff et al.
The two cell types bound 1251-CD40.Fcwith indistinguishable affinities (Kd = 0.44 k 0.08 n M and 0.5 k 0.10 nM, respectively). The level of CD40.Fc binding observed on cells immediately following sorting was maintained for at least 8 weeks after each sort (data not shown). The SN from EL40.9 cells was tested for the presence of other cytokines. Interleukins 1, 2, 3 , 4 , 6, 7, 9, 10, granulocyte-macrophage (GM)-CSF and mast cell growth factor were not detectable, while IL-5 and G-CSF were variably present at extremcly low (PM) levels. Murine IL-5 and G-CSF were subsequently found to have no stimulatory activity on human B cells (data not shown). 3.2 EL40.9 SN enhances sCD23 production As human CD40.Fc had been used to identify a source of murine CD40 ligand, it was of interest to determine whether EL40.9 SN contained human B cell stimulatory activity comparable to that previously shown for CD40 mAb [4-121. SN from EL40.9 cells was concentrated fivefold and was pre-cleared with biotinylated CD40.Fc (to remove CD40 binding activity) or with biotinylated hIgG1 as a specificity control, followed by streptavidin-coated magnetic beads as described in Sect. 2.3. For comparison, SN from unsorted EL4 cells were also used. The effect of
0
1
2
3
4
addition of EL4 or EL40.9 SN on IL-4-induced sCD23 release from human PBMC at day 6 is shown in Fig. 3 A. All SN preparations had an inhibitory effect at the highest concentration tested, but as EL4 SN was titrated enhanced sCD23 production was observed with a 1: 50 dilution of SN resulting in a threefold increase over that seen in the presence of TL-4 alone. EL40.9 SN induced a considerably higher level of sCD23 release than unsorted cell SN, and this activity could be greatly reduced by pretreatment of SN with biotinylated CD40.Fc. Similarly, IL-4-stimulated T-depleted E- cultures responded to the addition of EL40.9 SN by yielding greatly elevated levels of sCD23 as the SN was titrated (Fig. 3 B). Again, this activity could be removed by pretreatment of the SN with biotinylated CD40.Fc, while SN from unsorted EL4 cells contained a relatively weak stimulatory activity. 3.3 EL40.9 SN induces IgE secretion from IL-4activated B cells The effects of EL4 and EL40.9 SN on IL-4-induced IgE secretion was then examined (Table 1). In the presence of IL-4 alone, E- cells failed to secrete detectable levels of IgE. Addition of IL-4 together with a titration of fivefoldconcentrated EL4 SN resulted in low levels of IgE production at day 10 and a higher level at day 18, detectable only at the lower SN concentration, the higher concentrations being inhibitory. Concentrated EL40.9 SN was similarly titrated and induced greatly elevated secretion of IgE, again at the lower SN concentration. This activity could be largely removed by pre-clearing the SN with biotinylated CD40.Fc and streptavidin-coated Dynabeads. A similar decrease in activity could be achieved by addition of CD40.Fc directly into cultures containing IL-4 and EL40.9 SN (data not shown).
Table 1. Murine CD40 ligand induces IgE secretion from IL4-stimulated human B cells
5
dilution Stimulus
I
B
0
1
2
3
4
5
dilution F ~ ~ L L3.Y LSoluble ' murine CD40 ligand enhances IL-4-induced human sCD23 production. Soluble CD23 levels were determined in culture SN after PBMC (A) and T-depleted E- (B) cells were cultured for 7 days in the presence of 5 ng/ml IL-4 alone (A), IL-4 plus 500 ng/ml G28-5 mAb (A),or IL-4 plus a titration (threefold dilutions starting at 1: 10) of fivefold-concentrated SN from unsorted EL4 cells (W) or fivefold-concentrated SN from sorted EL40.5 cells pre-cleared with biotinylated human IgGl ( 0 )or biotinylated CD40.Fc (0). Results are expressed as the mean of triplicate cultures.
LL4 + EL40.9 SNC) pre-cleared + biotin-hIgG I IL-4 + EL40.9 SNc) pre-cleared + biotin-CD40.Fc Medium IL4 IL-4 + G28-5d)
Dilution
1: 2 1: 6 1 : 18 1: 2 1: 6 1: 18 1: 2 1: 6
1:18
IgE (ng/mO Day 10 Day 18
Hichard J. Armitage., Timothy A. Sato., Brian M. Macduff., Kv N. Clifford., Alan R. Alpert., Craig A . Smith. and William C. Fanslow. Departments of ~ ~ and Bjochemistryv, I~~~~~~ Research and Development Corporation, Seattle
CD40 hgdnd
Identification of a source of biologically active CD40 ligand We have identified the murine thymoma line EL4 as a source of biologically active CD40ligand. Using a biotin-labeled soluble CD40.Fc fusion protein, consisting of the extracellular domain of human CD40 and the Fc region of human IgGl , EL4 cells cell sorting with ~ ~ were subjected ~ t o repeated ~ flow cytometric ~ l to select forcells ~ enhanced biotinylated CD40.Fc binding. After nine rounds of sorting, the number of CD40.Fc binding sitedcell had risen from 450 on the unsorted parental EL4 cells to 15 000 on EL40.9 cells (EL4 cells sorted with biotinylated CD40.Fc for nine rounds). Scatchard analysis of radiolabeled CD40.Fc binding revealed that the surface-expressed CD40 ligand on parental EL4 and EL40.9 cells bound its receptor with a single class of high-affinity sites ( K d = 0.5 nM). Supernatant (SN) from the sorted EL40.9 cells was found to contain human and murine B cell stimulatory activity which could be removed by preclearing with immobilized CD40.Fc, confirming the presence of soluble CD40 ligand in the preparations. EL40.9 supernatant enhanced soluble CD23 (sCD23) release and induced IgE secretion from interleukin 4-stimulated human B cells. In addition, EL40.9 SN contained proliferative activity for anti-IgM-activated murine B cells which could be removed by treatment with immobilized CD40.Fc. However, the same SN had no demonstrable activity on the proliferation of human B cells. The results presented here describe, for the first time, a source of membrane-bound and soluble CD40 ligand. The soluble form of this murine ligand has activity on murine and human B cells and induces some of the functional responses predicted for the ligand based on the action of stimulatory antibodies directed against the human CD40 surface molecule.
1 Introduction The Bcell surface molecule CD40 has been shown to mediate various biological effects induced by binding of specifically directed monoclonal antibodies (mAb). These include : the induction of increased cell size [I, 21; homotypic and heterotypic adhesions [2, 31; entry into cell cycle of Bcells preactivated with anti-IgM, CD20 mAb or phorbol ester either alone [4-61 or in synergistic combination with IL-4 [2, 71; secretion of IgE [S-101. IgG and IgM [lo] from IL-4-stimulated B cells in the absence of Tcells; and enhancement of levels of surface-expressed and soluble CD23 released in the presence of IL-4 [11, 12). Human B cell lines have been derived from primary B cell populations grown in the presence of CDw32+ adherent cells, IL-4 and CD40 antibody [13]. Furthermore, germinal center centrocytes can be prevented from undergoing apoptosis by activation through both CD40 and receptors tor antigen 1141, while binding of CD40 mAb to immature B cells at various stages of maturity, as well as t o activated mature cells, results in enhanced tyrosine phosphorylation [IS], demonstrating that the CD40 surface molecule is active in signal transduction at multiple stages of B cell ontogeny. [I 103261 Correspondence: Richard J. Armitage. Dcpartrnent of Immunology. Immunex Research and Development Corporation. 51 University Street, Seattle, WA 98101, USA Abbreviations: tein
2071
CD40.Fc: Soluble CD40/human IgCl fusion pro-
0 VC'H Vcrlagsgeqellschaft mbH, D-6940 Weinheim, 1992
Cross-linking of CD40 is required for optimal stimulation of B cells [6]. Fab fragments of CD40 mAb are refractory to the proliferative signal delivered by intact antibody and, on their own, impart only a weak stimulatory signal. The sensitivity with which B cells respond to nanomolar concentrations of CD40 mAb [2, 41 has led to speculation that the CD40 surface molecule may be a receptor for an as-yet-unidentified ligand. Recently Gascan and colleagues [16] have shown that human Bcells can be induced to proliferate and secrete IgE in the presence of activated CD4+ T cell clones and IL-4, and that the signal provided by T cells can be effectively replaced by CD40 mAb [ 171. The cloning of CD40 [18] revealed homology between the extracellular domains of CD40 and nerve growth factor (NGF) receptor [19, 201. Similar homology with two distinct tumor necrosis factor (TNF) receptors [21-231 formed the basis of the CD40/TNFR superfamily.The argument for a ligand receptor role for CD40 is strengthened by the fact that three members of this family are receptors for cloned ligands. We have recently reported the expression of soluble constructs of the extracellular domain of human CD40, both as a 28-kDa soluble CD40 molecule and as a 57-kDa fusion protein consisting of soluble CD40 and the Fc region of human IgGl [24]. Both forms of soluble CD40 (sCD40 and CD40.Fc) were found to inhibit IL-4-mediated release of sCD23 and secretion of IgE in the absence of added CD40 mAb, demonstrating the induction and involvement of CD40 ligand in these responses. In this report we have used biotin-labeled CD40.Fc to identifv the murine thvmoma cell line EL4 as a source of C D k ligand. Replated rounds of flow cytometric cell sorting using biotinylated 0014-2O80/92/0808-207I$3.50 + ,2510
2072
Eur. J. Immunol. 1992. 22: 2071-2076
R. J. Armitage. T. A. Sato, B. M. Macduff et a1
CD40.Fc resulted in high levels of CD40 ligand expression on the cell surface and soluble ligand in the supernatant (SN). SN from CD40.Fc-sorted cells contained stimulatory activity for both murine and human B cells which could be removed by pretreatment with immobilized CD40.Fc, but not with control human IgGl antibody, confirming the presence of biologically active CD40 ligand in these preparations.
2 Materials and methods 2.1 Cell separation
Human peripheral blood mononuclear cells (PBMC) were isolated from the blood of healthy donors by centrifugation over Histopaque (Sigma Chemical Co., St. Louis, MO). T cell-depleted preparations (E-) were obtained by removal of T cells by rosetting with 2-aminoethylisothiouronium bromide (AET)-treated sheep red blood cells (SRBC) and further centrifugation over Histopaque.Tonsil1ar tissue was gently teased and the resulting cell suspension centrifuged over Histopaque. Purification of B cells was achieved by removal of cells rosetting with AET-treated SRBC and treatment of remaining cells with B cell lympho-kwik (One Lambda Inc., Los Angeles, CA) for 1 h at 37°C to lyse contaminating non-B cells. The resultant B cell population was >98% CD20+ as determined by flow cytometry (FACScan, Becton Dickinson, San Jose, CA) of cells stained with Leu-16 mAb (Becton Dickinson). Murine splenic B cells wcre isolated using anti-T cell antibody and complement, followed by passage over Sephadex G-10 (Pharmacia, Piscataway, NJ) to remove adherent cells. Bcells were then positively selected by panning on plates coated with 5 pg/ml goat anti-mouse IgM (Southern, Birmingham, AL). Isolated cells were > 95% surface IgM+ as determined by flow cytometry.
2.2 Reagents The 57-kDa gene product of the extracellular domain of human CD40 fused to the Fc region of human IgGl (CD40.Fc) was expressed and purified as described previously [24]. CD40.Fc or control human IgGl were biotinylated as follows; 50 pg protein (200 pg/ml in 0.1 M NaHC03 p H 8.3) was incubated with 1 pg (1mg/ml in DMSO) biotin-x-nonhydroxysuccinimide (NHS, Calbiochem, La Jolla, CA) for 30 min at room temperature. At the end of the incubation period, the reaction mixture was microfuged through a 1-ml Sephadex G-25 (Pharmacia) desalting column and thc eluate adjusted to 100 yg/ml in PBS plus 0.02% NaN3. Protein concentration of biotinylated CD40.Fc and hIgG 1 was determined by micro-BCA assay (Pierce, Rockford. 1L) with ultrapure bovine serum albumin as standard. Murine and human recombinant IL-4 were purified from yeast SN as previously described [25, 261. For proliferation experiments, murine and human B cells were activated with soluble goat anti-murine IgM (Southern) and rabbit anti-human IgM bound to acrylamide beads (Bio-Rad, Richmond, CA), respectively. Affinity-purified IgGl CD40 mAb G28-5 [4] was a generous gift from Dr. E. A. Clark.
2.3 Flow cytometric cell sorting The murine thymoma cell line EL4 was obtained from American Type Culture Collection (Rockville, MD). Cells were incubated with biotinylated CD40.Fc ( 5 pg/ml plus 1 X lo7 cells) or biotinylated hIgG1 (5 pg/ml plus 1 X lo6 cells) for 30 rnin at 4°C. After washing twice in PBS, streptavidin-phycoerythrin (PE) (Becton Dickinson) :!i!uted 1: 5 in PBS was added to cells for 30 rnin at 4°C. Cells were washed twice in PBS, resuspended to 3 X 106/ml and the cells stained with biotinylated CD40.Fc were subjected to flow cytometric cell sorting on a FACStar plus (Becton Dickinson) using cells stained with biotinylated hIgGl as a control for non-specific binding. For the first four rounds of sorting, the 2% of cells exhibiting the highest intensity of staining with biotinylated CD40.Fc were collected. For subsequent rounds of sorting, the brightest staining 1%of cells were col1ected.Typically 1 X lo4 - 4 x lo4 cells were recovered after each sort. Unsorted EL4 and sorted cells were maintained in DMEM containing 5% fetal bovine serum (FBS), 50 p~ 2-mercaptoethanol, 1 mM sodium pyruvate, 2 mM L-glutamine, 20 mM Hepes, penicillin and streptomycin. Supernatant preparations were harvested from cells for use in functional assays when cell density was optimal (I 8 x 105 cells/ml). Concentration of SN was performed by centrifugation over Centriprep 10 concentrators (Amicon, Beverly, MA). In selected experiments, EL40.9 SN was pre-cleared of CD40 ligand activity by addition of biotinylated CD40.Fc ( 5 pg/ml) for 30 min at 4 "C. Streptavidin-coated Dynabeads (Dynal Inc., Great Neck, NY) were then added (50pl beaddm1 SN) for a further 30 min at 4"C, after which time the beads were removed by three cycles of placing SN in a magnetic field. In all cases control SN was prepared by pre-treatment with 5 pg/ml biotinylated hIgG1 under identical conditions. 2.4 Radiolabeling of CD40.Fc Radioiodination was effected with 5 pg of purified protein in PBS using a solid-phase chloramine T analogue (Iodogen, Pierce, St. Louis, MO), as described [23]. Radiolabeled stocks (5 X lop8 M) were stored at 4°C in RPMI 1640 containing 2% BSA, 20 mM Hepes buffer, 0.2% sodium azide, pH 7.2 (binding media). 2.5 Binding assays and data analysis Binding assays were carried out with serially diluted 1251-labeledCD40.Fc (1251-CD40.Fc)and either unsorted EL4 or sorted EL40.9 cells (3 x 106/point) in binding media at 4°C for 4 h. Aliquots were microfuged through a phthalate oil mixture to separate free and bound lZ5ICD40.Fc, and the SN and pellets counted. Nonspecific binding was measured by the inclusion of a 100-fold molar excess of unlabeled CD40.Fc. Binding data was plotted in a Scatchard format using linear least squares analysis.
2.6 Culture conditions All human cell cultures were conducted in RPMI containing 100 U/ml penicillin, 100 pg/ml streptomycin, 2 mM Lglutamine and 10% heat-inactivated FBS (Intergen, Pur-
CD40 ligand
Eur. J. Immunol. 1992. 22: 2071-2076
chase, NY) at 37°C in a humidified atmosphere of 10% COZ.For measurement of proliferation, 1 X lo5 cells/well were cultured in triplicate in flat bottomed 96-well microtiter plates (Corning, Corning, NY) for 72 h in the presence of the appropriate additives as described in Sect. 3. For determination of sCD23 and IgE production, 1 X lo5 PBM or E- cells were cultured in round-bottom 96-well microtiter plates in the presence of 5 ng/ml hIL-4 for the time indicated. Murine B cells (1 X lo5 cells/well) were cultured in flat-bottom 96-well microtiter plates in RPMl 1640 supplemented with 5% FBS (Hazelton, Lenexa, KS), 1mM sodium pyruvate, 0.1 mM non-essential amino acids, 100 U/ml penicillin, 100 pg/ml streptomycin, 2 mM L-glutamine, and 50 pM 2-mercaptoethanol. Cultures were maintained at 37°C in 10% COZ for 72 h. Human and murine cclls were pulsed with 1 pCi/well tritiated-thymidine ([’HIdThd; 25 Ci/mmol, Arnersham, Arlington Heights, IL) for the final 8 h of culture. Cells were harvested onto glass fiber disks with an automated cell harvester. Incorporated cpm were measured by liquid scintillation spectrometry. For sCD23 induction and IgE secretion, 1 X los cells/well were cultured in triplicate in round-bottom 96-well (Nunc microtiter plates (Intermountain Scientific Corp., Bountiful, UT) for the appropriate time in the presence of additives as detailed in Sect. 3.
2.7 Detection of soluble CD23 and IgE Soluble CD23 levels were determined by ELISA using a commercial sCD23 detection kit (The Binding Site, San Diego, CA). The limit of sensitivity of the sCD23 ELISA was 500 pg/ml. IgE was detected by ELISA as follows: flat-bottom 96-well microtiter plates (Corning) were coated with mouse mAb anti-human IgE (Zymed, San Francisco, CA) at 1 : 500 dilution in PBS. After washing three times, a blocking step was carried out using 5% non-fat dried milk. This was followed by the addition of a titration of IgE standard (Calbiochem, La Jolla, CA) or test supernatants. After washing three times, biotinylated goat anti-human IgE (Kirkegaard and Perry, Gaithersburg, MD) was added at 1: 500 dilution followed by further washing and addition of streptavidin-horseradish peroxidase (Zymed) 1: 500 dilution. After washing, the reaction was developed using 3,3’,5,5’-tetramethylbenzidine(TMB) substrate (Kirkegaard and Perry) and absorbance measured at 520 nm. All incubation steps were performed using a volume of 100 pl/well for 1 h at room temperature.Wash steps were carried out in PBS plus 0.05% Tween. The sensitivity of the IgE ELISA was 100 pg/ml.
2073
Fluorescence Intensity
Figure 1. Flow cytometric analysis of EL4 and EL40.9 cells. Unsorted EL4 cells (A) and sorted EL40.9 cells (B) were stained with 5 pglml biotinylated human IgG1 (-----) or biotinylated CD40.Fc (-) and streptavidin-PE. EL40.9 cells were also stained with 5 pglrnl biotinylated CD4O.Fc in the presence of a 30-fold excess of unlabeled CD40.Fc (.....) .
cells stained with biotinylated CD40.Fc were subjected to cell sorting, and the 1%-2% of cells showing highest intensity of staining were collected. Cell sorting was repeated over a period of several months with 1 X lo4 4 X lo4 cells typically recovered after each sort. After nine rounds of sorting, the intensity of staining with biotinylated CD40 had increased greatly (Fig. 1 B ) over that seen on unsorted EL4 cells. Binding of biotinylated CD40.Fc to EL40.9 cells (EL4 cells sorted for CD40.Fc binding for nine rounds) could be effectively competed by the addition of a 30-fold excess of unlabeled CD40.Fc. Quantitative analysis of 1251-CD40.Fcequilibrium binding with unsorted and sorted cells is shown in Fig. 2. While unsorted EL4 cells had approximately 500 CD40.Fc-binding sitedcell (Fig. 2 A), sorted EL40.9 cells exhibited 1.6 X lo4sitedcell (Fig. 2B).
N
0 x h
3 Results 3.1 Flow cytometric staining with biotinylated CD40.Fc The fusion protein CD40.Fc was labeled with biotin and used for flow cytometric analysis of a range of human and murine cell lines to examine cell surface expression of CD40 ligand. All human cell lines examined lacked detectable binding of biotinylated CD40.Fc when compared to cells incubated with biotinylated human IgGl as a specificity control (data not shown). The murine thymoma line EL4 was found consistently to bind biotinylated CD40.Fc with very low intensity (Fig. 1 A ) . Based on this finding, EL4
I
I
I
1
2
3
C(M) x i 0 9 Figure 2. Equilibrium binding of IZ5I-CD40.Fcto EL4 and EL40.9 cells. Serially diluted ‘2sI-CD40.Fc was incubated with either unsorted (A) or sorted EL4 cells (B) for 4 h at 4°C. Binding was assayed as described in Sect. 2.5. All binding was corrected for nonspecific binding. Inserts represent data replotted in Scatchard format.
2074
Eur. J. Immunol. 1992. 22: 2071-2076
R. J. Armitage.T. A. Sato, B. M. Macduff et al.
The two cell types bound 1251-CD40.Fcwith indistinguishable affinities (Kd = 0.44 k 0.08 n M and 0.5 k 0.10 nM, respectively). The level of CD40.Fc binding observed on cells immediately following sorting was maintained for at least 8 weeks after each sort (data not shown). The SN from EL40.9 cells was tested for the presence of other cytokines. Interleukins 1, 2, 3 , 4 , 6, 7, 9, 10, granulocyte-macrophage (GM)-CSF and mast cell growth factor were not detectable, while IL-5 and G-CSF were variably present at extremcly low (PM) levels. Murine IL-5 and G-CSF were subsequently found to have no stimulatory activity on human B cells (data not shown). 3.2 EL40.9 SN enhances sCD23 production As human CD40.Fc had been used to identify a source of murine CD40 ligand, it was of interest to determine whether EL40.9 SN contained human B cell stimulatory activity comparable to that previously shown for CD40 mAb [4-121. SN from EL40.9 cells was concentrated fivefold and was pre-cleared with biotinylated CD40.Fc (to remove CD40 binding activity) or with biotinylated hIgG1 as a specificity control, followed by streptavidin-coated magnetic beads as described in Sect. 2.3. For comparison, SN from unsorted EL4 cells were also used. The effect of
0
1
2
3
4
addition of EL4 or EL40.9 SN on IL-4-induced sCD23 release from human PBMC at day 6 is shown in Fig. 3 A. All SN preparations had an inhibitory effect at the highest concentration tested, but as EL4 SN was titrated enhanced sCD23 production was observed with a 1: 50 dilution of SN resulting in a threefold increase over that seen in the presence of TL-4 alone. EL40.9 SN induced a considerably higher level of sCD23 release than unsorted cell SN, and this activity could be greatly reduced by pretreatment of SN with biotinylated CD40.Fc. Similarly, IL-4-stimulated T-depleted E- cultures responded to the addition of EL40.9 SN by yielding greatly elevated levels of sCD23 as the SN was titrated (Fig. 3 B). Again, this activity could be removed by pretreatment of the SN with biotinylated CD40.Fc, while SN from unsorted EL4 cells contained a relatively weak stimulatory activity. 3.3 EL40.9 SN induces IgE secretion from IL-4activated B cells The effects of EL4 and EL40.9 SN on IL-4-induced IgE secretion was then examined (Table 1). In the presence of IL-4 alone, E- cells failed to secrete detectable levels of IgE. Addition of IL-4 together with a titration of fivefoldconcentrated EL4 SN resulted in low levels of IgE production at day 10 and a higher level at day 18, detectable only at the lower SN concentration, the higher concentrations being inhibitory. Concentrated EL40.9 SN was similarly titrated and induced greatly elevated secretion of IgE, again at the lower SN concentration. This activity could be largely removed by pre-clearing the SN with biotinylated CD40.Fc and streptavidin-coated Dynabeads. A similar decrease in activity could be achieved by addition of CD40.Fc directly into cultures containing IL-4 and EL40.9 SN (data not shown).
Table 1. Murine CD40 ligand induces IgE secretion from IL4-stimulated human B cells
5
dilution Stimulus
I
B
0
1
2
3
4
5
dilution F ~ ~ L L3.Y LSoluble ' murine CD40 ligand enhances IL-4-induced human sCD23 production. Soluble CD23 levels were determined in culture SN after PBMC (A) and T-depleted E- (B) cells were cultured for 7 days in the presence of 5 ng/ml IL-4 alone (A), IL-4 plus 500 ng/ml G28-5 mAb (A),or IL-4 plus a titration (threefold dilutions starting at 1: 10) of fivefold-concentrated SN from unsorted EL4 cells (W) or fivefold-concentrated SN from sorted EL40.5 cells pre-cleared with biotinylated human IgGl ( 0 )or biotinylated CD40.Fc (0). Results are expressed as the mean of triplicate cultures.
LL4 + EL40.9 SNC) pre-cleared + biotin-hIgG I IL-4 + EL40.9 SNc) pre-cleared + biotin-CD40.Fc Medium IL4 IL-4 + G28-5d)
Dilution
1: 2 1: 6 1 : 18 1: 2 1: 6 1: 18 1: 2 1: 6
1:18
IgE (ng/mO Day 10 Day 18
