Major histocompatibility complex class II fetal skin dendritic cells are potent accessory cells of polyclonal T-cell responses


: A. Elbe-Bürger, Department of Dermatology, Division of Immunology, Allergy and Infectious Diseases, University of Vienna Medical School, VIRCC, Vienna, Austria.


Whereas dendritic cells (DC) and Langerhans cells (LC) isolated from organs of adult individuals express surface major histocompatibility complex (MHC) class II antigens, DC lines generated from fetal murine skin, while capable of activating naive, allogeneic CD8+ T cells in a MHC class I-restricted fashion, do not exhibit anti-MHC class II surface reactivity and fail to stimulate the proliferation of naive, allogeneic CD4+ T cells. To test whether the CD45+ MHC class I+ CD80+ DC line 80/1 expresses incompetent, or fails to transcribe, MHC class II molecules, we performed biochemical and molecular studies using Western blot and polymerase chain reaction analysis. We found that 80/1 DC express MHC class II molecules neither at the protein nor at the transcriptional level. Ultrastructural examination of these cells revealed the presence of a LC-like morphology with indented nuclei, active cytoplasm, intermediate filaments and dendritic processes. In contrast to adult LC, no LC-specific cytoplasmic organelles (Birbeck granules) were present. Functionally, 80/1 DC in the presence, but not in the absence, of concanavalin A and anti-T-cell receptor monoclonal antibodies stimulated a vigorous proliferative response of naive CD4+ and CD8+ T cells. Furthermore, we found that the anti-CD3-induced stimulation of naive CD4+ and CD8+ T cells was critically dependent on the expression of FcγR on 80/1 DC and that the requirement for co-stimulation depends on the intensity of T-cell receptor signalling.


The essential event of a T-cell-mediated immune response is the recognition of antigen peptides bound to major histocompatibility complex (MHC) through the CD3–T-cell receptor (TCR) complex. Subsequent signal transduction, activation and differentiation lead to the generation of effector T cells. 1 Extensive research has been conducted to explore the mechanisms of this important process using anti-CD3 or anti-TCR monoclonal antibodies (mAbs) as an alternative to the physiological ligands of the CD3–TCR complex. Conflicting reports about the role of antigen-presenting cells (APC) in CD3-mediated T-cell proliferation exist. According to the ‘two-signal model’, naive T cells require one signal, provided by the interaction of antigenic peptide–MHC complex with the TCR, to give specificity to the immune system and a second, antigen-independent co-stimulatory signal delivered by APC for activation and differentiation into effector cells. 2 Triggering of CD3–TCR in the absence of accessory signals induces anergy or unresponsiveness rather than proliferation. 3 The roles of APC in anti-CD3-mediated T-cell proliferation have been suggested to include the provision of a matrix, whereby multiple interactions between surface CD3 and the stimulating anti-CD3 are promoted, the production of cytokines and the delivery of accessory signals. 4 While ligation of the mAb by APC through their FcR is generally agreed to be a requirement for peripheral T-cell proliferation, 5 the necessity and nature of the accessory signals provided by APC are still controversial. While many scientists have accepted the two-signal theory, others argue that a second signal may not be needed. 6–11 Finally, some others have suggested that whether a second signal is needed or not is dependent on factors such as the epitope specificity, isotope and concentration of the antibody. 12–15

Conflicting data also exist concerning the response of CD4+ versus CD8+ T cells. We postulate that at least one factor contributing to the discrepant results of previous studies is that different APC have been used. In view of these disputes we have used a long-term dendritic cell (DC) line established from fetal mouse skin 16 to investigate requirements for T-cell activation through the CD3 pathway and to determine responsiveness of naive CD4+ versus CD8+ T cells.

Materials and methods


C3H/HeN (C3H) and C57BL/6 mice were obtained from Charles River Wiga GmbH, Sulzfeld, Germany. At the time of the experiments all mice were 6–10 weeks old.

Antibodies and reagents

The mAbs 16-10A1 (anti-CD80), GL1 (anti-CD86), 15-5-5S (anti-H-2Dk12), 34-2-12S (anti-H-2Dd3), 145-2C11 (anti-CD3ε), H57-597 (anti-TCR αβ), 536 (anti-TCR Vγ3), IM7 (anti-CD44), MEL-14 (anti-CD62L), 23G2 (anti-CD45RB) and isotype control antibodies were purchased from PharMingen (San Diego, CA). Hybridomas 2.4G2 (anti-CD32, HB197), HO-13-4 (anti-Thy-1.2, TIB99), GK1.5 (anti-CD4, TIB207), 53-6.72 (anti-CD8, TIB105), 16-3-1N (anti-H-2Kk, HB25), 3-83P (H-2KkDk12, HB20), M5/114.15.2 (anti-I-Ab,d,q&I-Ed,k, TIB120), 10-2.16 (anti-I-Aβκ, TIB93) were purchased from the American Type Culture Collection (Rockville, MD). Purified hamster immunoglobulin G (IgG) was obtained from Cappel (West Grove, PA). Second-step reagents were fluorescein isothiocyanate (FITC)-conjugated goat anti-hamster IgG (Cappel) and FITC-labelled mouse anti-rat κ IgG2a (MAR 18.5, TIB216; American Type Culture Collection). Some of the mAb used were purified from supernatants of the corresponding hybridomas and protein concentrations were adjusted to 1 mg/ml before use. Fab fragments of mAb 16-10A1 and 536 were prepared using the ImmunoPure Fab Kit from Pierce Chemical Company (Rockford, IL) according to the manufacturer's instructions. The purity of both intact and Fab fragments was verified by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) followed by silver staining.

Immunolabelling studies

Aliquots of cells (1 × 105/sample) were incubated with non-conjugated first-step antibodies for 30 min, washed twice in cold phosphate-buffered saline (PBS)/1% fetal calf serum (FCS)/0·1%NaN3 and then exposed to an appropriately diluted FITC-labelled second-step reagent. Stained cells were analysed using a FACScan® flow cytometer equipped with lysys II software (Becton Dickinson, Mountain View, CA). Dead cells, visualized by ethidium bromide (Sigma Chemical Co., St Louis, MO) uptake, were excluded from the analysis.

Cell lines

Both 80/1 DC (established from the skin of day 15 fetal C3H mice) and DC 18 (established from the skin of day 18 fetal BALB/c mice) were expanded and maintained in culture medium (CM) supplemented with interleukin-2 (IL-2)/ concanavalin A (Con A) and granulocyte–macrophage colony-stimulating factor (GM-CSF), respectively, as previously described. 16 The B-cell line B4.1 was obtained from F. W. Alt (Howard Hughes Medical Institute, The Children's Hospital, Department of Genetics and Center for Blood Research, Harvard Medical School, Boston, MA) and is a subclone of the B-cell lymphoma M12. 17

Cultured Langerhans cells (cLC)

Single epidermal cell suspensions, prepared from C3H ear skin as described 18 were filtered through a cell strainer (70 µm, Falcon, New Jersey), washed and cultured (1·5 × 106 cells/ml) in CM for 72 hr in 75 cm2 flasks (Costar, Cambridge, MA). Thereafter, non-adherent epidermal cells were harvested and treated with anti-Thy-1.2 mAb for 30 min at 4°, followed by Low-Tox-M rabbit C′ (Cedarlane Laboratories, Hornby, Ontario, Canada) for 45 min at 37°. After removal of dead cells by density-gradient centrifugation (Lympholyte-M, Cedarlane Laboratories), approximately 80–95% of the remaining cells were dendritic in shape and I-A+, characteristic of LC.

CD4+ and CD8+ T-cell subsets

T cells from C3H or C57BL/6 mice were prepared by passing unseparated mesenteric lymph node cells through nylon wool columns. T-cell subsets were enriched using antibody- and C′-mediated lysis essentially as reported. 16 Naive CD4+ and CD8+ T cells were > 97% CD44low CD45RBhigh CD62Lhigh as determined by flow cytometry (data not shown).

T-cell-depleted spleen cells

Splenocytes from C3H mice were treated with NH4Cl to remove red blood cells and were either used for experiments or subsequently depleted of T cells by a mixture of anti-Thy-1, anti-CD4, anti-CD8 mAb and Low-Tox-M rabbit C′. The remaining population, containing DC, B cells and macrophages was used for RNA preparation. The figure legends and tables indicate the experimental design, numbers of cells per well and culture durations for each of the respective cells.

SDS–PAGE and immunoblotting

Cell lysates of each population were prepared as described 19 and submitted to electrophoresis on 10% SDS–PAGE, blotted onto nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA) which were then blocked in 5% dry milk/0·05% Tween-20 (Sigma Chemical Co.)/PBS for at least 6 hr and incubated with either biotinylated mAb 10-2.16 (anti-I-Aβκ) or a biotinylated IgG isotype control mAb (both at 200 ng/ml). Membrane-bound first-step antibodies were reacted with a streptavidin horseradish peroxidase conjugate (1 : 5000; Amersham). Membranes were incubated with ECL® developing solution (Amersham) and exposed to Kodak X-Omat S film.

Reverse transcriptase–polymerase chain reaction (RT-PCR) analysis

Total RNA of each population was isolated with the guanidinium isothiocyanate/caesium chloride method. Complementary DNA was prepared under the conditions recommended by the vendor (Superscript Preamplification System; Gibco BRL) and amplified by 30 cycles of PCR [94° for 1 min, 60° for 1 min, and 72° for 1 min (primers specific for the invariant chain (li) were used with an annealing temperature of 54°)] with pairs of the following oligonucleotide primers: I-Aακ (5′-CCTGGAGACATTGGCCAGTACAC-3′ and 5′-GGGAAAAAGCAAGTTGGGGGTC-3′), I-Aβκ (5′-GGACTGAGGGCGGAAACTCCG-3′ and 5′-CCTGTGACGGATGAAAAGGCCAAG-3′), I-Eακ (5′-CCAGGCGGAGTTCTAT CTTTTACC-3′ and 5′-GCGGAAGAGGTGATCGTCCCT C-3′), I-Eβκ (5′-GGTCAGAGACTCCAGACCATGG-3′ and 5′-CCCAACTCCACTCAACATCTTGTTC-3′), li (5′-GCT GGATGAAGCAGTGGCTC-3′ and 5′-GCAGTTATGG CGCCCGCGG-3′), H-2Mακ (5′-GCAGAAGTCAGGAG CTGTGCTG-3′ and 5′-CCGAGGCCAAAGGCCACACCAC-3′), H-2Mβ1κ (5′-GGCTCATGTGGAAAGCACGTGC-3′ and 5′-CGCTGTGCTGAACCACGCAGG-3′), H-2Mβ2κ (5′-GGTCCTCAGTCTGCACTGTATGG-3′ and 5′-GCGCCA TCTGAAGAAGCCAACAC-3′). All primers were purchased from Genset (Paris, France). Products of PCR amplification were analysed on 1·8% agarose gels and photographed using a Mitsubishi video copy processor inverting the image.

Electron microscopy

The 80/1 DC were fixed in 0·1% glutaraldehyde in cacodylate buffer for 1 hr at room temperature, post-fixed in 1% osmium tetroxide in phosphate buffer for 30 min at 4°, dehydrated and embedded in Epon. Ultrathin sections were stained with lead hydroxide and uranyl acetate and viewed with a Philips EM 410 electron microscope.

Proliferation assays

X-irradiated (1·5 Gray/min, Philips RT 305, Philips, Vienna, Austria) 80/1 DC (40 Gray) and cLC (15 Gray) were cultured with either syngeneic CD4+ (> 98%) or syngeneic and allogeneic CD8+ (> 99%) lymph node T cells, and for control purposes, alone, in the presence or absence of different concentrations of Con A (Pharmacia Biotech Inc., Uppsala, Sweden) or soluble anti-CD3ε (clone 145-2C11) and anti-TCR αβ (clone H57-597) mAb. These cultures and the allogeneic mixed lymphocyte reaction (MLR) were set in CM in 96-well round-bottom culture plates (Costar) and maintained at 37° (95% air/5% CO2). At the indicated time-points 37 KBq [3H]TdR was added to each well, and 10–12 hr later, cells were harvested and [3H]TdR incorporation was measured in a liquid scintillation counter (Packard Instruments, Meriden, CT). Data are expressed as mean counts per min (c.p.m.) ± SD of triplicate samples unless indicated otherwise. The role of MHC class I, FcγR, CD80 and CD86 molecules in lectin-and anti-CD3ε mAb-induced proliferation was examined by adding the respective mAb and the isotype controls at the indicated concentrations for the entire incubation period.


80/1 DC do not express MHC class II molecules at the protein and transcriptional level

We have previously shown that 80/1 DC fail to exhibit anti-MHC class II surface reactivity 16 but have not addressed the question whether they express MHC class II molecules at the protein and molecular level. To test this, lysates of various cell types were subjected to SDS–PAGE and were immunoblotted using the anti-I-Aβκ mAb 10-2.16 or the isotype-matched control mAb ( Fig. 1a). Whereas a specific single band could be detected in spleen cells, neither 80/1 DC nor control lysates displayed reactivity with the anti-MHC class II reagent. To glean information about the molecular expression of MHC class II and related molecules in these cells, RNA was subjected to PCR amplification. Abundant transcripts for MHC class II and -related genes (H-2M) and the invariant chain (li) were found in control samples (T-cell-depleted spleen cells and B4.1 cells) but never in 80/1 DC ( Fig. 1b). These findings demonstrate that 80/1 DC do not express MHC class II molecules either at the transcriptional or at the protein level.

Figure 1.

Fetal skin DC fail to express MHC class II molecules at the protein and transcriptional level. (a) Cell lysates of spleen cells (positive control), DC 18 (negative control) and 80/1 DC were electrophoresed and blotted onto nitrocellulose. Membranes were reacted with biotinylated I-Ak-specific mAb 10-2.16 or a biotinylated IgG control mAb. Membrane-bound first-step mAbs were detected with streptavidin peroxidase. Molecular mass calibration is shown on the left in kDa. (b) RT-PCR analysis using 10 ng plasmid DNA (ctrl.) and 1 µg total RNA of murine B cells (B4.1) and T cell-depleted spleen cells as appropriate controls versus 1 µg total RNA of 80/1 DC. Reactions were carried out as described in the Materials and Methods. Size standards (123 bp) are given to the left and to the right for each panel.

Appearance of 80/1 DC by electron microscopy

When 80/1 DC were examined by electron microscopy, they showed a polygonal shape with dendrites and veils of cytoplasm extending in several directions from the cell body, a peripherally located nucleus and many cytoplasmic organelles ( Fig. 2a). Sporadically, a coated pit could be observed ( Fig. 2b). The cells expressed features of a high metabolic activity, such as many ribosomes and Golgi bodies ( Fig. 2c,d). Centrioles were often found in the vicinity of the Golgi apparatus ( Fig. 2d). All stages of the endosomal/lysosomal system were present, from multivesicular bodies and various stages of endosomes to fully developed lysosomes ( Fig. 2e). No LC granules (i.e. Birbeck granules) could be detected.

Figure 2.

Ultrastructure of 80/1 DC. On electron micrographs the cells appeared to be of dendritic nature, with many cytoplasmic organelles (a). Occasionally, a coated pit could be observed (arrow, b). The cells expressed many ribosomes and Golgi bodies (small arrowheads, c and d). Centrioles were often found in the vicinity of the Golgi apparatus (large arrowheads, d). All stages of the endosomal/lysosomal system were present (e), from multivesicular bodies (double small arrows) and various stages of endosomes (large arrows) to fully developed lysosomes (double large arrows). (a) Bar = 2·5 µm; (b–e) bar = 0·25 µm.

80/1 DC are potent accessory cells of polyclonal T-cell responses

In the next series of experiments, we investigated whether 80/1 DC can function as stimulators in polyclonal T-cell responses. As can be seen in Fig. 3, we found that 80/1 DC act as very potent accessory cells in Con A- and anti-TCR/CD3-driven T-cell activation. The concentrations of Con A necessary to induce maximal proliferative responses in naive CD4+ and CD8+ T-cell subsets varied. While 1–2·5 µg/ml Con A was optimal for CD4+ T cells, the concentration for CD8+ T cells was higher (5 µg/ml) (data not shown). Interestingly, although using optimal concentrations for each T-cell subset, CD8+ T-cell responses to Con A and 80/1 DC were repeatedly higher than those observed with CD4+ T cells, with a peak occurring on day 2 for both T-cell subsets ( Fig. 3a,b). As shown in Fig. 3(c)–(f), CD4+ and CD8+ T cells proliferated strongly to anti-TCR αβ/CD3ε mAb in the presence of 80/1 DC but the magnitude and time–course of the responses were somewhat different. Whereas peak proliferative responses for CD4+ T cells were regularly observed between days 4 and 5 ( Fig. 3c,e), the mitotic activity of CD8+ T cells consistently peaked on day 3 of culture and then decreased ( Fig. 3d,f). Thus, 80/1 DC stimulate both CD4+ and CD8+ T-cell activation and expansion in the presence of anti-TCR αβ/CD3 mAb, but the subsets differ with respect to the development of these responses; CD4 responses are more sustained while CD8 responses are more rapid.

Figure 3.

Accessory cell function of 80/1 DC in mitogen-induced T-lymphocyte proliferation. Naive lymph node T cells (H-2k) (1 × 105/well) were cultured with irradiated, syngeneic 80/1 DC (1 × 104/well) either in the presence or absence of Con A (a, b), or 5 µg/ml soluble anti-TCR αβ (H57-597; c, d), anti-CD3ε (2C11; e, f) and isotype control mAb (c–f). T-cell proliferation was determined at the indicated time-points by [3H]TdR incorporation, and data are expressed as mean c.p.m. ± SD of triplicate cultures. Irradiated 80/1 DC did not incorporate any [3H]TdR above background levels when stimulated with any of the above mentioned stimuli and had none to marginal mitogenic effects when cultured with either highly purified CD4+ or CD8+ T cells in the absence of Con A or mAb. DNA synthesis of T cells in the presence of Con A, anti-TCR αβ, anti-CD3ε and isotype control mAb was negligible. Experiments were repeated four times, yielding essentially identical results.

MHC class I molecules have no accessory role for T-cell stimulation induced by 80/1 DC and either lectins or mAbs to the TCR complex

Next we set up experiments to examine whether the CD8+ T-cell response to 80/1 DC and Con A or anti-TCR mAb is regulated by MHC class I molecules. Therefore selected mAb directed against different MHC class I determinants were added at the initiation of the culture. As shown in Fig. 4, the continuous presence of the H-2Kk-specific mAb 16-3-1N greatly diminished the 80/1 DC-induced proliferation of allogeneic CD8+ lymphocytes ( Fig. 4a) but did not inhibit the 80/1 DC and Con A or anti-TCR αβ/CD3-activated T-cell response ( Fig. 4b–d). A similar result was obtained when the mAb 3.8P3, which is directed against H-2Kk and cross-reacts with H-2Dk determinants, was added to the cultures ( Fig. 4a–d). The addition of the H-2Dk-specific mAb 15-5-5S had no inhibitory effect in either system tested ( Fig. 4a–d). These results suggest that MHC class I molecules on stimulator cells do not influence polyclonal T-cell responses. Consistent with this was the observation that the MHC class I+ fibrosarcoma cell line L929, expressing levels of MHC class I molecules comparable with those of 80/1 DC, 16 failed to induce a significant proliferative response of CD4+ and CD8+ T cells in the presence of Con A and anti-TCR mAb (data not shown).

Figure 4.

Anti-MHC class I mAb inhibit 80/1 DC-induced proliferation of allogeneic CD8+ lymphocytes but do not influence the co-stimulatory potential of 80/1 DC for T-cell activation via the TCR–CD3 complex. Irradiated 80/1 DC (1 × 104/well) were cultured with allogeneic (H-2b) (a) and syngeneic (H-2k) (b–d) CD8+ T lymphocytes in the presence or absence of Con A (5 µg/ml), anti-TCR αβ (H57-597; 5 µg/ml) and anti-CD3ε (2C11; 5 µg/ml) mAb either with an isotype control mAb [34-2-12S (H-2Dd) ascites added at 1 : 1000] or with selected anti-MHC class I mAb [ascites of clones 16-3-1N (H-2Kk), 15-5-5S (H-2Dk) and 3.83P (H-2KkDk) were added at a dilution of 1 : 1000]. T-cell populations were depleted of detectable accessory cells, as demonstrated by the lack of a Con A or anti-TCR–CD3 response in the absence of a source of co-stimulatory activity. Background incorporation of [3H]TdR by irradiated stimulator cells in the presence of any stimulus was negligible (data not shown). Similar results were obtained in two additional experiments.

Role of FcγR and co-stimulatory molecules on 80/1 DC in polyclonal T-cell responses

Considerable evidence indicates that T-cell activation by antibodies to the TCR complex, regardless of their specificity for Vα, Vβ, or CD3ε, is a consequence of cross-linking between T lymphocytes and FcγR-bearing cells. 7,20–27 To test whether a similar mechanism is operative in our system, we first assessed the cell surface expression of CD32 molecules on 80/1 DC and, for comparison, on cLC. Whereas 80/1 DC displayed strong anti-CD32 reactivity, cLC expressed these molecules only moderately (data not shown). The decisive role of CD32 molecules in the 80/1 DC/anti-CD3-driven T-cell proliferation was tested by functional assays in the presence of the FcγR-blocking mAb 2.4G2. Because cLC act as very potent accessory cells in anti-CD3-induced assays, 20 they were included as a positive control. The continuous presence of the mAb 2.4G2 inhibited the 80/1 DC- or cLC-driven proliferative response of both T-cell subsets to 2C11 ( Table 1). While the mAb 2.4G2 essentially abrogated the cLC-induced responses of both CD4+ and CD8+ T-cell subsets at high and suboptimal concentrations of 2C11, an almost complete inhibition of the 80/1 DC-induced proliferation of both T-cell subsets was observed only with suboptimal concentrations of 2C11 ( Table 1). The specificity of the blocking is indicated by the fact that the addition of an isotype control antibody had no effect on the anti-CD3-induced T-cell response and that the Con A-induced T-cell proliferation was not affected by the 2.4G2 mAb ( Table 1). We, thus, conclude that in our system, the mitogenesis caused by anti-CD3 mAb and 80/1 DC involves FcγRII/III molecules.

Table 1.  The role of FcγR in DC and anti-CD3-induced mitogenesis
Reagents added[3H]TdR uptake (c.p.m.)
80/1 +cLC +
CD4+ T cellsCD8+ T cellsCD4+ T cellsCD8+ T cells
  1. isot. ctrl., isotype-matched control; Con A, concanavalin A.

  2. Naive lymph node T cells (H-2k) (1 × 105/well) were cultured with irradiated, syngeneic 80/1 DC or cLC (1 × 104/well) in the presence or absence of anti-CD32 (2.4G2; 10 µg/ml) and control antibody together with either Con A (2·5 µg/ml for CD4 and 5 µg/ml for CD8+ T cells, respectively) or the mAb 2C11 at the indicated concentrations. Cultures were pulsed with [3H]TdR for 10 hr on day 2–3 and proliferation was determined from triplicate microwells (mean c.p.m.) for each condition (SD were regularly < 20% and are omitted for clarity). The results have been confirmed in three independent experiments.

2C11 (10 µg/ml)73 266188 633158 762284 131
2C11 (10 µg/ml) + 2.4G260 357132 32115377152
2C11 (10 µg/ml) + isot. ctrl.93 514168 398221 918285 703
2C11 (0·01 µg/ml)172 999198 290148 098172 757
2C11 (0·01 µg/ml) + 2.4G2202513 59063992667
2C11 (0·01 µg/ml) + isot. ctrl.128 729185 824164 050175 636
Con A146 333128 772174 334137 464
Con A + 2.4G2127 015139 879141 948207 764
Con A + isot. ctrl.120 350137 575154 905227 040

Numerous studies suggest that CD28 is the principal co-stimulatory receptor for T-cell activation which binds to CD80 (B7-1) and CD86 (B7-2) molecules expressed by APC.28,29 To explore the role of B7–CD28 interaction in T-cell activation by 80/1 DC and Con A or soluble anti-TCR αβ/CD3 mAb, we first determined the cell surface expression of CD80 and CD86 molecules on 80/1 DC and, for control purposes, on cLC. While 80/1 DC and cLC displayed strong anti-CD80 reactivity, CD86 molecules could only be detected on cLC but never on 80/1 DC (data not shown). To evaluate the role of B7 molecules in triggering CD4+ and CD8+ T cells with 80/1 DC and mitogens, we used blocking anti-CD80 and anti-CD86 mAb. As illustrated in Fig. 5, the anti-CD80 mAb 16-10A1, but not its isotype control, inhibited the proliferative responses of both CD4+ and CD8+ T cells induced by co-stimulation with 80/1 DC and either Con A, anti-TCR αβ or anti-CD3ε mAb. Because whole antibody molecules might transduce a signal through the B7 receptor per se, Fab fragments were used to block CD80 co-stimulation. As shown in Table 2, anti-CD80 Fab fragments completely inhibited the 80/1 DC and Con A-activated CD4+ T-cell response at all concentrations tested. In contrast, the CD8+ T-cell response was inhibited in a dose-dependent manner, but was never complete even when high amounts of anti-CD80 Fab fragments were used ( Table 2). Similar results were obtained using the mAb 2C11 ( Table 2). The mAb against anti-CD86 failed to suppress T-cell proliferation even at high concentrations when tested in a similar experimental setting as described for anti-CD80 mAb (data not shown) implying that 80/1 DC in fact do not even express low levels of CD86 molecules which could have been missed by FACS analysis. To test whether the dependence on co-stimulation seen above would vary with the strength of TCR signalling, naive CD4+ and CD8+ T cells were stimulated with two different concentrations of anti-CD3 to mimic strong and weak TCR signalling. The anti-CD80 mAb was used to block co-stimulation that was provided by 80/1 DC. We found that the degree of inhibition caused by anti-CD80 mAb increased as the concentration of anti-CD3 decreased ( Table 3). Taken together, these results demonstrate that CD80+ CD86 80/1 DC and CD80+ CD86+ cLC were essentially equivalent in their ability to provide naive T-cell co-stimulation. Furthermore, our results suggest that T-cell activation is a balance between TCR signalling and co-stimulatory molecule signalling, with less co-stimulation required when the TCR signal is strong.

Figure 5.

An anti-CD80 mAb abrogates T-cell proliferation induced by co-stimulation with 80/1 DC and lectin or anti-TCR mAb. Irradiated 80/1 DC (1 × 104/well) were cultured with syngeneic CD4+ and CD8+ T lymphocytes in the presence or absence of Con A (1 and 5 µg/ml for CD4+ and CD8+ T cells, respectively), soluble anti-TCR αβ (H57-597; 5 µg/ml) and anti-CD3ε mAb (2C11; 5 µg/ml), either with an isotype control or with anti-CD80 mAb (16-10A1; 1 µg/ml). Both T-cell populations were depleted of detectable accessory cells, as demonstrated by the lack of a Con A or anti-TCR–CD3 response in the absence of a source of co-stimulatory activity. Background incorporation of [3H]TdR by irradiated stimulator cells in the presence of any stimulus was negligible (data not shown). Three experiments showed concordant results.

Table 2.  CD80 mAb inhibits 80/1 DC and Con A and anti-CD3-induced T cell proliferation
 [3H]TdR uptake (c.p.m.)
80/1 +
Reagents addedCD4+ T cellsCD8+ T cells
  1. isot. ctrl., isotype-matched control.

  2. Naive lymph node T cells (H-2k) (1 × 105/well) were cultured with irradiated, syngeneic 80/1 DC (1 × 104/well) in the presence or absence of Fab fragments (16-10A1 and 536) together with either Con A (2 µg/ml for CD4 and 5 µg/ml for CD8+ T cells, respectively) or the mAb 2C11 (0·01 µg/ml). Cultures were pulsed with [3H]TdR for 10 hr on days 2 (Con A) and 3 (2C11). Proliferation was determined from triplicate microwells (mean c.p.m.) for each condition (SD were regularly < 20% and are omitted for clarity). Data shown are representative of three experiments.

Con A91 494310 331
Con A + CD80 Fab (10 µg/ml)3059661
Con A + CD80 Fab (5 µg/ml)3188919
Con A + CD80 Fab (1 µg/ml)33325 836
Con A + isot. ctrl. Fab (10 µg/ml)87 657367 379
Con A + isot. ctrl. Fab (5 µg/ml)68 841334 867
Con A + isot. ctrl. Fab (1 µg/ml)72 332330 854
2C11260 393392 400
2C11 + CD80 Fab (5 µg/ml)17413305
2C11 + CD80 Fab (1 µg/ml)18684974
2C11 + CD80 Fab (0·1 µg/ml)63 106157 390
2C11 + isot. ctrl. Fab (5 µg/ml)251 623371 059
2C11 + isot. ctrl. Fab (1 µg/ml)245 885356 619
2C11 + isot. ctrl. Fab (0·1 µg/ml)241 478399 891
Table 3.  T-cell proliferation induced by 80/1 DC and anti-CD3 mAb can be completely abrogated by anti-CD80 mAb only at suboptimal levels of 2C11
 [3H]TdR uptake (c.p.m.)
80/1 +
Reagents addedCD4+ T cellsCD8+ T cells
  1. isot. ctrl., isotype-matched control.

  2. Naive lymph node T cells (H-2k) (1 × 105/well) were cultured with irradiated, syngeneic 80/1 DC (1 × 104/well) together with the mAb 2C11 in the presence or absence of the anti-CD80 mAb 16-10A1 and the respective isotype control (1 µg/ml). Cultures were pulsed with [3H]TdR for 10 hr on day 3. Proliferation was determined from triplicate microwells (mean c.p.m.) for each condition (SD were regularly < 20% and are omitted for clarity). Data shown are representative of six experiments.

2C11 (10 µg/ml)214 442340 473
2C11 (10 µg/ml) + CD8031 429137 184
2C11 (10 µg/ml) + isot. ctrl.213 679315 344
2C11 (0·01 µg/ml)215 722219 737
2C11 (0·01 µg/ml) + CD8015381159
2C11 (0·01 µg/ml) + isot. ctrl.204 442220 335


The main goal of this paper was to compare directly the activation requirements for naive CD4+ and CD8+ T-cell responses in lectin- and anti-TCR-driven systems. The MHC class I+ class II 80/1 DC were ideally suited for addressing this question because they represent a homogeneous DC population and the results obtained are comparable. With this system we show that CD80 molecules expressed by 80/1 DC can provide effective co-stimulation for both T-cell subsets. Most interestingly, we found that the level of co-stimulation required depends on the strength of TCR-mediated stimulation, with less co-stimulation necessary when TCR signalling was high.

We have demonstrated that 80/1 DC express neither invariant chain nor MHC class II and -related molecules and all our attempts to induce MHC class II expression on 80/1 DC using physiological sources of cytokines (e.g. co-culture of 80/1 DC with lymphokine-rich Con A-activated rat spleen cell supernatants, allogeneic T cells and a mixture of skin-derived cells obtained by trypsinization of H-2-incompatible skin from selected gestational ages (day 18–19) and adult skin) were ineffective (ref. 16 and A. Elbe-Bürger, unpublished observations). Our data imply that essential factors required for further differentiation of our DC line are limiting or absent in the activation conditions used or alternatively that we have succeeded in propagating a population of mature MHC class I+ class II DC.

To gain a further understanding of the phenotypic properties of 80/1 DC, they were examined by electron microscopy. We found that 80/1 DC exhibit cytological features of DC, including the presence of surface projections. The absence of LC-specific cytoplasmic organelles in 80/1 DC was not too surprising, because it has been shown that fetal mouse LC are devoid of these organelles, 30 that only a minor part of LC in adult tissues in mice contains Birbeck granules, and that these organelles are not a prerequisite for LC immune function in vitro and in vivo.31

The requirements of T-cell activation have been studied by dissecting responses to a variety of T-cell mitogens. The lectin Con A has been used extensively as a T-cell-specific mitogen. Although Con A can cross-link a number of glycosylated surface receptors, including the TCR, 32–35 Con A binding alone does not activate T cells and requires the presence of accessory cells. It has long been believed that MHC class II antigens are mandatory in T-cell proliferation by Con A because accessory function to this lectin has been shown to be efficiently mediated only by MHC class II+ cell types. 20,36–38 In this report we show that 80/1 DC are able to promote T-cell proliferation in the presence of Con A very effectively. The possibility that MHC class II recognition might have occurred during mitogen activation of T cells in our system was excluded by experiments showing that the addition of an anti-MHC class II mAb to the cultures failed to inhibit T-cell activation after stimulation with Con A and 80/1 DC (A. Elbe-Bürger, unpublished observation). In this respect our results complement numerous reports showing that MHC class I+ class II cell lines36,37,39–42 can provide accessory function for polyclonal activation of naive T cells. However these studies neither addressed the mechanism of the APC ‘help’ nor investigated whether differences exist in the responses of CD4+ and CD8+ T cells to the same APC. We have therefore assessed whether 80/1 DC are providing a separate and essential co-stimulatory signal during Con A-induced activation of CD4+ and CD8+ T cells and what specific molecules are involved. We found that CD80 molecules on 80/1 DC can initiate an effective Con A-induced activation of CD4+ and CD8+ T lymphocytes. Several lines of evidence support this conclusion. First, the anti-CD80 mAb 16-10A1, either intact or Fab, leads to complete (CD4+ T cells) and almost complete (CD8+ T cells) suppression of 80/1 DC and Con A-induced T-cell proliferation. Our observation that the 80/1 DC and Con A-induced CD8+ T-cell response could never be abrogated completely by an anti-CD80 mAb suggests that naive CD8+ T cells are less dependent on co-stimulation than are CD4+ T cells and/or that other molecules, which have been described as exhibiting co-stimulatory capacity for CD8+ T cells such as intercellular adhesion molecule (ICAM-1) 43 or 4-1BB ligand, 44 may play a role in the activation of naive CD8+ T cells and remains to be investigated in this model. Second, MHC class I+ fibroblasts failed to induce T-cell proliferation in response to Con A. While our data showing that fibroblasts fail to function as accessory cells in Con A-driven T-cell proliferation emphasize the necessity of a second signal for productive T-cell proliferation they seem to be in contrast to results of other investigators. However, all studies demonstrating that fibroblast cell lines can function as potent accessory cells in the Con A response have used L cells.42,45,46 It has recently been shown that these L cells express CD80 molecules and that their co-stimulatory activity for polyclonal T-cell activation in response to Con A could be blocked by an anti-CD80 mAb. 41 In contrast, the fibroblast cell line we employed in our study expresses neither CD80 nor CD86 molecules (A. Elbe Bürger, unpublished observation). We show here that 80/1 DC are able to provide naive T-cell co-stimulation in the presence of Con A, suggesting that CD80, in the absence of CD86, is able to deliver a strong signal for the induction of powerful CD4+ as well as CD8+ T-cell proliferation. Our findings confirm and extend previous studies demonstrating that T cells cultured with Chinese hamster ovary cells that were stably transfected with murine CD80 cDNA proliferated vigorously to Con A stimulation that could be completely abrogated by the addition of an anti-CD80 mAb.41,47 Quite unexpectedly we found that CD4+ T cells are less responsive to 80/1 DC and Con A stimulation than CD8+ T cells, suggesting that for a maximal CD4+ T-cell response either CD80 densities on 80/1 DC are suboptimal or that CD86 co-stimulation is necessary. Indeed, evidence exists that CD80 co-stimulation can only support a partial Con A response, and that CD86 is required for maximal activation. 47

Mitogenesis and T-cell activation in vitro caused by anti-CD3 mAb are dependent on interaction of these mAb with FcγR.25,48In vivo, the anti-CD3 mAb 2C11 mediates immunosuppression in mice by clearing a majority of T cells from the circulation. 49 However, before inducing immunosuppression, 2C11 triggers T-cell proliferation and a massive systemic release of cytokines, such as IL-2, IL-4, IL-6, interferon-γ, tumour necrosis factor-α and GM-CSF, resulting in severe side-effects.50,51 These effects can be reduced/inhibited by the administration of either F(ab′)2 fragments of 2C11 mAb 51 and or the FcγR mAb 2.4G2, 25, respectively, indicating that FcγR binding is also involved in T-cell activation in vivo. Consistent with this is our observation that mitogenesis caused by anti-CD3 mAb and 80/1 DC and cLC involves FcγR molecules. The following findings support this conclusion. DNA synthesis was abrogated by a mAb to the FcR. Interestingly, while at high concentrations of 2C11, the mAb 2.4G2 completely inhibited the cLC-induced T-cell proliferation, we observed a partial inhibition when 80/1 DC were used. This lack of complete inhibition can probably be ascribed to the higher expression of CD32 molecules on 80/1 DC compared to cLC. Second, soluble anti-CD3 mAb either in the absence of 80/1 DC or in the presence of FcγR cells (L929) failed to induce a proliferative T-cell response. These results are in agreement with a report showing that FcR DC do not initiate mitogenesis to anti-CD3 mAb. 52 A possible explanation for the role of FcR on 80/1 DC could be that they allow aggregation of the anti-CD3 molecules, which then more effectively stimulate the T cells. A second possibility is that FcR approximate the anti-CD3 ‘opsonized’ T cells to the DC and that subsequently other LC properties induce mitogenesis. However, observations that LC and T cells cluster with one another to comparable extents in the presence or absence of anti-CD3 imply that anti-CD3 is not simply an opsonin for 80/1 DC and LC FcR. 20 Anti-CD3 molecules on 80/1 DC and LC FcR are providing a stimulus to the T cell that is obviously independent of another mechanism that brings DC and T cells together. Our finding that resident peritoneal macrophages, expressing high levels of FcR, 20 were four times less potent to stimulate proliferative responses of naive CD4+ and CD8+ T cells to soluble anti-CD3 mAb than 80/1 DC and cLC, when compared side-by-side, indicated that the expression of FcR alone is not enough to stimulate a strong T-cell response in the presence of anti-CD3 mAb (A. Elbe-Bürger, unpublished observation). Furthermore, the fact that LC function during anti-CD3 responses increases markedly in culture even though the level of FcγR decreases suggests that some other aspect of LC function, is developing in culture and is involved in TCR–CD3-induced T-cell activation. Indeed, it was shown that LC undergo a process of maturation in culture as evidenced by changes in their phenotype (e.g. up-regulation of co-stimulatory molecules) and function. 53 Among membrane-bound co-stimulatory molecules, members of the B7-family have been demonstrated to play an important role in anti-CD3-mediated T-cell responses. 41,54–56 However, most investigators used B7-transfected cell lines as stimulators and resting CD4+ and CD8+ responder T cells that include both naive and memory populations. This may give misleading results, because memory T cells have different activation needs than do naive T cells, requiring less co-stimulation for optimum response and giving modest response in the absence of co-stimulation.57,58 In this study we have used defined DC and have isolated naive CD4+ and CD8+ T cells. Results presented here, in accord with results reported by others using coated microspheres 59 show that CD80 in the presence of anti-CD3 mAb is a strong co-stimulatory molecule for the proliferation of naive CD4+ and CD8+ T cells. We found that a mAb against CD80, either intact or Fab, blocked the accessory cell function. In contrast to the observations made with CD4+ T cells, there was usually never complete inhibition of 80/1 DC-induced CD8+ T-cell responses even though saturating levels of anti-CD80 mAb were added to the culture. This lack of complete inhibition suggests that 80/1 DC may express other molecules capable of transmitting co-stimulatory signals such as CD2, which is known to contribute to cytotoxic T lymphocyte activation. 60 Second, MHC class I+ fibroblasts failed to induce T-cell proliferation in response to anti-CD3 mAb. Our data differ from those of Abe et al. 61 who found that B-cell enriched APC resulted in the preferential activation of CD4+ T cells in the presence of anti-CD3 and results of Makrigiannis et al., 62 who reported that CD86 is the major CD28-binding ligand for CD8+ T cell in response to anti-CD3 mAb. These apparent discrepancies are probably attributable to the use of different APC sources. Therefore, the role of CD80 and CD86 in an immune response may be determined primarily by their differential expression on APC.

The magnitude of the T-cell response and the amount of co-stimulation required for this response seem to be dependent on the intensity of TCR interaction.2,56, 63,64 In favour of this idea our results show that the response of both, naive CD4+ and CD8+ T cells on co-stimulatory molecules can vary with the strength of TCR signalling. We found that less co-stimulation is required when the TCR stimulus is increased. This seems to be a density (or cross-linking) effect, since stimulation with a low and high dose of anti-CD3 requires higher and lower levels of co-stimulatory signals, respectively, whereby CD8+ T cells need less co-stimulation. This observation makes sense given the role of CD8+ CTL in infectious disease, where the ability to respond to infected cells, which are typically not professional APC, might well be an advantage.

In conclusion, the results presented in this paper show that MHC class II fetal skin DC, similar to cLC are potent accessory cells in polyclonal T-cell responses. Our results suggest that CD80 alone can provide a co-stimulatory signal that allows the efficient activation of naive CD4+ and CD8+ T cells and that activation represents an integration of TCR and co-stimulatory molecule signalling, with less co-stimulation required when the TCR signal is maximized. A better understanding of the precise requirements for naive T-cell activation is of crucial importance for the development of new therapeutic approaches aiming at specifically eradicating tumours and viral infections or curing autoimmune diseases.


We thank L. Karlsson (R.W. Johnson Pharmaceutical Research Institute, San Diego, CA) for providing the sequences of H-2 Mα, Mβ1 and Mβ2, O. Majdic (Institute of Immunology, Vienna, Austria) for the preparation of Fab fragments, and S. Olt for excellent technical assistance. This work was supported by grants from the Austrian Science Foundation [P10797-MED, P14243-MED (AEB)].