We have isolated CD8+ α/β T cells from the blood of atopic and healthy individuals which recognize a nonpeptide antigen present in an allergenic extract from Parietaria judaica pollen. This antigen appears to be a carbohydrate because it is resistant to proteinase K and alkaline digestion, is hydrophilic, and is sensitive to trifluoromethane-sulphonic and periodic acids. In addition, on a reverse-phase high performance liquid chromatography column the antigen recognized by CD8+ T cells separates in a fraction which contains >80% hexoses (glucose and galactose) and undetectable amounts of proteins. Presentation of this putative carbohydrate antigen (PjCHOAg) to CD8+ T cell clones is dependent on live antigen presenting cells (APCs) pulsed for >1 h at 37°C, suggesting that the antigen has to be internalized and possibly processed. Indeed, fixed APCs or APCs pulsed at 15°C were both unable to induce T cell response. Remarkably, PjCHOAg presentation is independent of the expression of classical major histocompatibility complex (MHC) molecules or CD1. CD8+ T cells stimulated by PjCHOAg-pulsed APCs undergo a sustained [Ca2+]i increase and downregulate their T cell antigen receptors (TCRs) in an antigen dose– and time-dependent fashion, similar to T cells stimulated by conventional ligands. Analysis of TCR Vβ transcripts shows that six independent PjCHOAg-specific T cell clones carry the Vβ8 segment with a conserved motif in the CDR3 region, indicating a structural requirement for recognition of this antigen. Finally, after activation, the CD8+ clones from the atopic patient express CD40L and produce high levels of interleukins 4 and 5, suggesting that the clones may have undergone a Th2-like polarization in vivo. These results reveal a new class of antigens which triggers T cells in an MHC-independent way, and these antigens appear to be carbohydrates. We suggest that this type of antigen may play a role in the immune response in vivo.

It is well known that T lymphocytes expressing α/β TCR recognize antigenic peptide bound to MHC molecules. However, there is increasing evidence that nonprotein antigens can also be recognized by T cells (for review see reference 1). A first category of nonprotein antigens is represented by lipids or glycolipids such as mycolic acid and lipoarabinomannan (24). Presentation of these antigens to α/β T cells requires processing and is carried out by CD1 molecules which display a specificity for hydrophobic antigens (1). A second class of nonprotein antigens includes low mol wt phosphorylated molecules such as isopentenyl pyrophosphate and its synthetic analogs (57). In this case presentation to γ/δ T cells is independent from uptake and processing and does not require any known antigen presenting molecule.

A long-standing question is whether carbohydrates may be recognized by T cells. It has been shown that MHC molecules can bind glycopeptides via the peptide moiety and display the carbohydrate for interaction with TCR (811). However, this type of recognition still can be regarded as a variation of conventional peptide antigen recognition as it has been shown for haptenized peptide (12). So far, there has been neither evidence that carbohydrates may bind directly to MHC or other antigen presenting molecules (13, 14) nor reports that carbohydrates as such might trigger T cells in the absence of MHC molecules.

We have observed that T cells which proliferate in vitro in response to a pollen extract from Parietaria judaica (PjE)1, a common environmental allergen of the Mediterranean area (15), comprise a sizable fraction of CD8+ α/β T cells. Here we provide evidence that these T cells recognize a putative carbohydrate antigen present in PjE. Presentation of PjCHOAg to T cells is dependent on APCs, but is not restricted by either MHC or CD1 molecules. These results reveal a new mode of antigen presentation and a new class of ligands which can trigger α/β T cells.

Materials And Methods

PjE Reactive T Cell Clones.

PBMCs from one atopic and one healthy individual were cultured in 96-well flat-bottomed plates (Costar Corp., Cambridge, MA) in RPMI 1640 supplemented with 2 mM l-glutamine, 1% nonessential amino acids, 1% pyruvate, 50 μg/ml kanamycin, 5 × 10−5 M 2-ME (GIBCO BRL, Gaithersburg, MD) and 5% human serum (Sigma Chemical Co., St. Louis, MO) in the presence of PjE (50 μg protein/ml). Recombinant IL-2 (20 U/ml; provided by Dr. A. Lanzavecchia, Basel Institute for Immunology, Basel, Switzerland) was added at day 6. After 20 d, T cell blasts were cloned by limiting dilution. T cell clones were maintained by periodic restimulation with phytohemagglutinin, irradiated allogeneic PBMCs (3,000 rad from a 60Co source), and rIL-2 as previously described (16).

Cell Lines.

Dendritic cells (DCs) were generated by culturing peripheral blood monocytes in the presence of GM-CSF and IL-4 (17). T2 (TAP-deficient), and 221 (class I–deficient) were provided by Dr. A. Lanzavecchia. SJO (class II–deficient) was a gift of Dr. J. Gorsky (Blood Research Institute, Milwaukee, WI). C1R (class I–deficient) stably transfected with the CD1a, CD1b, or CD1c genes or mock transfected were provided by Dr. M. Brenner (Brigham and Women's Hospital, Boston, MA). EBV-transformed lymphoblastoid cell lines (B-LCL) were derived from PBMCs as described (18). All cell lines were maintained in 10% FCS-RPMI (Hyclone Labs., Inc., Logan, UT).

FACS® Analysis.

Cell staining was performed using mouse mAbs followed by FITC- or PE-conjugated affinity-purified, isotype specific goat anti–mouse antibodies (Southern Biotechnology Associates, Inc., Birmingham, AL). The following antibodies were used: OKT8 (IgG2a, anti-CD8; American Tissue Culture Collection, Rockville, MD), 6D10 (IgG1, anti-CD4; provided by Dr. E. Roosnek, University of Geneva, Geneva, Switzerland); PE-Leu2a (IgG1, anti-CD8; Becton Dickinson, Mountain View, CA); FITC–anti-TCR α/β (IgG1; Becton Dickinson); B1 (IgG1, anti–pan-γ/δ, provided by Dr. G. De Libero, University of Basel, Basel, Switzerland); a panel of mAbs recognizing different Vβ gene products (Immunotech, Marseille, France); and TRAP-1 (IgG1, anti-hCD40L; provided by Dr. R. Kroczek, University of Berlin, Berlin, Germany). The samples were analyzed on a FACScan® using propidium iodide to exclude dead cells.

Antigen Presentation Assay.

T cells were cultured with APC in the presence of various concentrations of antigens in 200 μl human serum–RPMI in flat-bottomed microplates. Cultures were pulsed with 1 μCi of [3H]thymidine (2 Ci/mmol; Amersham International, Little Chalfont, UK) on day 2 and harvested 16 h later. For antibody blocking studies, cultures were carried out in the presence of different antibodies provided as 1:10 dilution of culture supernatant. The following mAbs were used: W6/32 (IgG2a, anti–HLA class I), L243 (IgG2a, anti–HLA-DR), SPVL3 (IgG2a, anti–HLA-DQ), and B7/21 (IgG1, anti–HLA-DP). APCs were either irradiated or treated with 1 mg/ml mitomycin C (Sigma Chemical Co.) for 30 min at 37°C. In some experiments the APC were fixed with 0.05% glutaraldehyde for 30 s and blocked with 0.1 M lysine in PBS for 5 min before or after antigen pulse.

Cytokine Production.

T cells were stimulated for 40 h with 10−7 M PMA (Sigma Chemical Co.) and anti-CD3 antibody (TR66, 0.2 μg/ml). Cytokine production was quantified in the supernatants by ELISA (PharMingen, San Diego, CA; reference 19).

Preparation of PjE and Removal of Proteins, Phosphates, and Carbohydrates.

PjE was prepared by extracting PjE pollen (Allergon, Angelholm, Sweden) with 0.125 M bicarbonate buffer as described (20). The extract contained ∼20% protein as assessed by the Bradford method (21). PjE was also purchased from Neo-Abello (Madrid, Spain). To remove proteins from the extract two methods were used. First, 5 mg PjE was digested in 200 μl PBS, pH 7.2, with 100 μg proteinase K (PK; Sigma Chemical Co.) in the presence of 1 mM CaCl2 for 2 h at 37°C followed by heating at 65°C for 5 min. The digestion was repeated twice. Alternatively, 5 mg PjE was incubated in 100 μl 3 M NaOH/0.1 M NaBH4 overnight at room temperature. The samples (referred to as PKPjE and NaOHPjE, respectively) were tested and compared to the same concentration of untreated or mock-treated PjE. To remove phosphates, PjE (2.5 mg) was treated with 10 U phosphatase A (Sigma Chemical Co.) in 100 μl of 50 μM Tris/HCl, pH 8.3, for 2 h at 37°C. The enzyme was inactivated by heating to 65°C for 10 min. To remove sugars two methods were used. PjE (10 mg) was treated for 1 h at 0°C with trifluoromethane-sulphonic acid (TFMS) and anisole (1:2 ratio) as described (22). After incubation, an equal volume of ice-cold aqueous pyridine (60% vol/vol) was added. Alternatively, PjE (5 mg/ml) was incubated in 10 mM acetate buffer, pH 4.7, containing 10 mM sodium periodate (SP; Sigma Chemical Co.) for 6 h at 4°C in the dark (23). The reaction was stopped with a molar excess of ethylene glycol and NaBH4 was added at 2 mg/ml. The samples (referred to as TFMSPjE and SPPjE, respectively) were tested and compared to the same concentration of untreated or mock-treated PjE.

Purification of the Antigen Recognized by CD8+ T Cell Clones.

The antigen recognized by CD8+ T cell clones was purified from crude pollen extract using two different protocols. In the first protocol, PjE was dissolved in 4% trichloroacetic acid (15 mg/ml) and precipitated proteins were removed by centrifugation. The supernatant was neutralized by 1 N NaOH and lyophilized. The sample was fractionated on a preparative reverse-phase column (RP-8; Waters, Millipore Corporation, Bedford, MA) by eluting with a linear gradient of methanol in water. Collected fractions were concentrated at reduced pressure, lyophilized, and tested for biological activity as well as for carbohydrate and protein content. In the second protocol, PjE (300 mg in 80 ml water) was extracted (80:65, vol/vol aqueous to organic) with chloroform/ methanol (2:1, vol/vol) and the two phases were separated by centrifugation. The upper layer and the interface were recovered and lyophilized, and the lower layer was dried under nitrogen. Most of the biological activity was found in the upper fraction. The water soluble material (92 mg) was dissolved in 12 ml water and separated using two cycles of centrifugal concentrators with a cutoff of 10 and 3 kD (Amicon Inc., Beverly, MA). The biological activity was recovered in the filtrate of the 10-kD concentrator and was only partially retained in the 3-kD concentrator. The 3-kD filtrate was lyophilized to give 11 mg residue. This was further fractionated on a reverse phase HPLC (rpHPLC; RP-18; Waters) by eluting with a gradient of methanol in water. The fractions were tested for biological activity as well as for carbohydrate and protein content. The sugar content in the chromatographic fractions was measured according to Ashwell (24). Briefly, 0.1 mg of the lyophilizate was dissolved in 2 ml water and treated with 0.05 ml phenol reagent (80% by weight of redistilled phenol in water) and 5 ml of concentrated sulfuric acid. The sample was kept at room temperature for 30 min and the optical density was determined at 490 nm. The amount of carbohydrate content was estimated by comparison to a standard curve of glucose. The protein content was determined by a protein assay (Bio-Rad, Hercules, CA) using BSA as standard.

Sugar Analysis on Active Antigenic Fractions.

Samples from the active chromatographic fractions were hydrolyzed with 2 M TFA followed by reduction with NaBH4 to give a mixture of alditols. Such compounds were acetylated by acetic anhydride and pyridine to yield a mixture of alditol acetates, which were analyzed by gas chromatography–mass spectrometry (GC-MS). To exclude the interference of free monosaccharides, a mock control was performed after acetylation of an untreated aliquot of the samples.

[Ca2+]i Measurement.

T cell clones were loaded with Indo-1 AM (Sigma Chemical Co.) as described (25). Cells were mixed at a 1:2 ratio with autologous B-LCL cells which had been pulsed with PjE or NaOHPjE. The cells were centrifuged at 1,500 rpm for 1 min at 4°C to allow conjugate formation, heated for 1 min at 37°C, and analyzed on a Coulter Elite Flow Cytometer (Coulter Electronics Inc., Hialeah, FL) to detect Ca2+ fluxes in conjugated T cells (25). Only live (based on forward and side scatter criteria) and Indo-1–loaded cells were included in the analysis. As positive control we recorded Ca2+ fluxes in T cells after cross-linking of CD3 with a specific mAb (clone TR66).

V β8 TCR-CDR3 Sequencing.

Three clones from the atopic patient and three clones from the normal control were further analyzed by CDR3 spectratyping as previously described (26). Once it was proved that the clones contained only one genomic sequence, direct sequencing was performed. The cDNA obtained from the relevant clones was amplified for the Vβ8 TCR family according to a published procedure (27). The Vβ8-specific PCR products were purified by means of a Microspin G50 column (Pharmacia Biotech, Uppsala, Sweden) to eliminate unincorporated primers. After purification 0.5-μl PCR products were used for cycle sequencing reaction according to instructions (Perkin-Elmer, Pasadena, CA) using sense and antisense internal primers in different PCR reactions. After amplification the samples were dried, boiled for 2 min in 4 μl loading dye, and then loaded on a fluorescence based sequencer (ABI373, Perkin-Elmer) using a denaturing 40% acrylamide gel. The samples were run for 8 h and the data were collected and analyzed using a 373 sequencing software for base calling.

Results

Isolation of CD8+ T Cell Clones Specific for PjE.

PBMCs from a PjE-sensitive patient and one nonatopic individual were stimulated in vitro with PjE. The responding cells were expanded with IL-2. After 20 d the cell lines were enriched for antigen-specific T cells as shown by a strong proliferative response to the inducing agent in the presence of autologous mononuclear cells (data not shown). Interestingly, in several independent cell lines from both individuals a sizable fraction of the T cells (20–30%) was CD8+ (Fig. 1,A). Since it is unusual to find CD8+ T cells in cultures stimulated by protein antigens, we isolated a number of CD4+ and CD8+ α/β T cell clones (Fig. 1 B) by limiting dilution and tested their capacity to recognize PjE.

As shown in Table 1, several CD8+ T cell clones isolated from the atopic patient and the healthy individual proliferated in response to PjE. Proliferation was specific since it was observed using different sources of PjE, while pollen extracts from unrelated species such as Cupressus arizonica and Olea europaea were ineffective (not shown). Interestingly, while CD8+ clones from the healthy control produced only IFN-γ, those isolated from the allergic patient produced high levels of IL-4 and IL-5 (Table 1) and expressed CD40L after stimulation (Fig. 1 C), suggesting that they may have undergone a Th2-like polarization in vivo.

α/β+ CD8+ T Cell Clones Recognize a Proteinase K–resistant, Phosphatase-resistant, Periodate-sensitive Antigen.

The nature of the antigen recognized by CD8+ PjE-specific T cell clones was investigated. The crude pollen extract was digested with proteinase K or subjected to alkaline hydrolysis. Both treatments destroyed proteins and abolished stimulation of PjE-specific CD4+ T cells, but did not affect and actually enhanced stimulation of the CD8+ clones, suggesting that CD8+ T cells recognize a nonprotein antigen (Fig. 2). Treatment with alkaline phosphatase did not decrease the capacity of PjE to stimulate CD8+ clones, indicating that the antigen does not contain critical phosphate residues (data not shown). When PjE was extracted with chloroform/methanol, the antigen recognized by CD8+ T cells separated in the aqueous phase (see below). Taken together, these properties suggested that the antigen recognized might be a carbohydrate. Therefore, we tested the sensitivity of the antigen to treatments which destroy sugars. In contrast to protease and phosphatase treatments, both periodic acid and TFMS treatment of PjE completely abolished recognition by CD8+ T cell clones, while the capacity to stimulate CD4+ T cell clones was not affected (Fig. 3). These results suggest that the antigen recognized by CD8+ T cell clones is not affected by removal of proteins or phosphate groups but is sensitive to treatment which degrades carbohydrates or related polyhydrate compounds.

Partial Purification of the Antigen Recognized by CD8+ T Cell Clones.

In a first attempt to purify the antigen recognized by CD8+ T cell clones, we separated the material remaining after protein precipitation by a semipreparative reverse-phase column. The activity was recovered in several fractions and the most active was enriched in carbohydrates (83%) but still contained ∼1% proteins (data not shown). In a second attempt, we combined phase fractionation with ultrafiltration and rpHPLC. After extraction with chloroform/methanol, the antigen recognized by CD8+ clones separated in the aqueous phase, which contains mainly polar molecules, but was absent at the interface and in the organic phase (data not shown). The hydrophilic fraction was separated using centrifugal concentrators with cutoff of 10 and 3 kD. The biological activity was recovered in the filtrate of the 10-kD concentrator and was partially retained in the 3-kD concentrator. The 3-kD filtrate was further fractionated on an rpHPLC by elution with a gradient of methanol in water (Fig. 4,A). As expected, the material with biological activity eluted very early while most of the contaminating compounds still present in the 3-kD filtrate were retained by the column and were eluted at later time points (Fig. 4,B and data not shown). The positive fraction was ∼20-fold more active on a weight basis than the starting material (Fig. 4,B, inset). When this fraction was analyzed by GC-MS, the chromatographic elution and the mass spectrum revealed the presence of glucose and galactose at a molar ratio of 47:53 (Fig. 4 C). Furthermore, the same fraction contained no detectable peptides as assayed by the Bradford method (17), direct and reverse-phase TLC of the hydrolyzed fraction (HCl 6 N, 110°C, 6 h) followed by ninidrin staining, Edman degradation, and amino acid analysis using a 494 protein sequencer (data not shown). To further exclude any peptide contamination, the active fraction was digested with carboxypeptidase Y, a protease which digests most peptides including those with blocked NH2 termini (27). This treatment did not affect the bioactivity of the purified antigen but completely abolished the ability of a tetanus toxoid peptide to stimulate a specific CD4+ T cell clone (data not shown). Taken together the above results suggest that, according to two different criteria (i.e., sensitivity to chemical treatment and purification with two methods), the antigen recognized by CD8+ clones is a low mol wt polyhydrate molecule, possibly a carbohydrate.

APC-dependent and MHC- and CD1-independent Presentation of PjCHOAg to CD8+ T Cell Clones.

The requirements for antigen recognition were studied using as readout T cell proliferation and cytokine production. The CD8+ clones proliferated in response to PjE only in the presence of APC (Fig. 5,A). Optimal presentation was provided by B-LCL and PBMCs, while, surprisingly, DCs were considerably less efficient (Fig. 5, A, and B). Several unrelated B-LCL tested were able to function as APCs for the CD8+ clones (Fig. 5,A), even if they lacked class I (221 and C1R) or class II molecules (SJO and T2) (Fig. 5,C). In addition, antibodies to class I and class II molecules did not inhibit T cell activation (Fig. 5 D), while they did inhibit antigen recognition by class I– and class II–restricted T cell clones (data not shown). A contribution of CD1 molecules was also ruled out by the finding that (a) anti-CD1 antibodies did not affect the proliferative response and (b) CD1-a, -b, and -c transfectants were as efficient as the untransfected control (data not shown). Taken together, these data suggest that none of the known antigen-presenting molecules are involved in the presentation of PjCHOAg to CD8+ α/β T cell clones.

The requirements for antigen uptake and processing were studied using [Ca2+]i increase as a sensitive and rapid assay of T cell antigen recognition. No response was observed when the CD8+ T cells were incubated with the antigen in solution, confirming the requirements for APCs (Fig. 6,D). In contrast, the clones showed a strong and sustained [Ca2+]i increase when conjugated with PjE-pulsed APC. Indeed, a time course of antigen pulsing revealed that a short incubation with antigen (1 h) was not sufficient for presentation (Fig. 6,B) and that the stimulatory capacity was progressively acquired during the next 5 h of incubation (Fig. 6,C). While APC fixed after antigen pulse retained the ability to induce [Ca2+]i increase in T cell clones (data not shown), APC fixed before a 5-h pulse at 37°C with PjE were ineffective (Fig. 6,E). In addition, live APC pulsed at 15°C failed to stimulate T cells (Fig. 6,F). Interestingly, chloroquine did not inhibit presentation of PjCHOAg (data not shown), suggesting that processing in an acidic compartment was not involved. Finally, APC retained the capacity to stimulate T cells after removal of the antigen (Fig. 6, G–I). By comparing the stimulatory capacity of APC chased for different times with that of APC pulsed with different concentrations of antigen, we estimated the half-life of antigen on pulsed APC to be ∼1 h. Taken together these results suggest that PjCHOAg has to be captured by the APC and it is displayed on their surface in a stable fashion, possibly associated with molecules preferentially expressed on B cells.

TCR-β Usage and Evidence for the Involvement of TCR in the Recognition of PjCHOAg.

All CD8+ PjCHOAg-specific T cell clones isolated from two individuals express the Vβ8 chain as detected by mAbs. This finding is compatible with stimulation by a Vβ8-specific superantigen possibly present in the allergenic extract. However, this possibility was ruled out by the finding that polyclonal Vβ8+ lines were not stimulated by PjE (data not shown). Since we could exclude a superantigen type of recognition we searched for a conserved motif in the CDR3 region (Table 2). Sequencing of the Vβ chains of six independent T cell clones revealed a remarkable conservation in positions 3 and 5 of the CDR3 region (threonine and valine, respectively), while the length differed by only one amino acid. In addition, all six clones analyzed used the Jβ1.2 segment. We also investigated the kinetics of T cell activation by PjCHOAg. As shown in Fig. 7, CD8+ T cells exposed to PjE-pulsed APC undergo a [Ca2+]i increase that is sustained for >30 min (Fig. 7,A) and downregulate their TCRs in an antigen dose– and time-dependent fashion (Fig. 7 B and data not shown). These results demonstrate that PjCHOAg triggers the TCR with the same kinetics as conventional TCR ligands such as peptide–MHC complexes.

Discussion

We have shown that CD8+ α/β T cells can specifically recognize a carbohydrate antigen present in a pollen extract. This antigen differs from other nonpeptide antigens (27, 28) because of its chemical nature and the lack of MHC or CD1 restriction.

Several lines of evidence suggest that the stimulatory ligand is a carbohydrate. First, it is resistant to treatments that destroy proteins, such as proteinase K digestion and alkaline treatment, excluding a possible contribution of a peptide. Second, it separates in the aqueous phase, ruling out the possibility that it may contain lipids. Third, it is sensitive to agents that destroy polysaccharides, such as TFMS and periodic acid. Finally, it is enriched in a fraction containing high levels of glucose and galactose. However, we cannot exclude that other polyhydrate molecules with similar sensitivity to chemical treatments may be responsible for the stimulation of PjCHOAg-specific T cell clones. Although the structure of the epitope still has to be defined, it is likely to be intracellularly generated since we have found that, in order to stimulate specific T cells, PjCHOAg has to be pulsed with live APC for >1 h. Indeed, fixed APC or APC pulsed at 15°C were both unable to induce T cell response. We have found that chloroquine does not interfere with presentation of PjCHOAg, suggesting that either processing of this antigen is not required or may involve a different intracellular compartment. More evidence is needed to elucidate these points.

A remarkable finding is that recognition of PjCHOAg, unlike recognition of other ligands by α/β T cells, does not require the expression of classical MHC class I, class II, and CD1 molecules. However, once generated the epitope remains associated with the cell surface with a half life of ∼1 h. We envisage two possibilities to explain how this epitope is displayed on APC surface for T cell recognition. First, the antigen may form a complex with a nonpolymorphic molecule which functions as classical antigen-presenting molecules, i.e., which bind antigen and interact with TCR. Second, PjCHOAg may be directly displayed on the cell membrane alone or in association with an antigen-binding molecule that serves only to concentrate the antigen but does not interact with TCR. Interestingly, DCs, which are the most efficient APC for protein antigens (29), are rather ineffective in presenting PjCHOAg, suggesting that they may lack this putative antigen-presenting/binding molecule.

Recognition of PjCHOAg by CD8+ T cell clones is mediated via TCR as shown by the kinetics of TCR triggering which is comparable to that of conventional TCR ligands such as peptide–MHC complexes (30). Moreover the presence of a conserved motif in the CDR3 region of the Vβ chain suggests the existence of a precise structural requirement for PjCHOAg recognition (30). This result, together with the fact that polyclonal Vβ8+ lines do not respond to PjCHOAg, rules out the possibility that PjCHOAg may act as a superantigen.

The finding that CD8+ T cell clones could be isolated from both atopic patients and from healthy donors is not surprising since they both are naturally exposed to this environmental allergen (31). However, it is interesting that only the clones from the atopic patient produce high levels of IL-4 and IL-5 and express CD40L , suggesting that these CD8+ cells may have undergone a Th2-like polarization in vivo (32, 33). The functional significance of these CD8+ T cells in allergy remains a matter of speculation (34). It is possible that these cells may participate in the production of IL-4 thus influencing the differentiation of naive allergen-specific T cells towards Th2 (3537). In addition, the requirement for processing and the stability of the epitope suggest the possibility that CD8+ PjCHOAg-specific T cell clones may function as typical carrier-specific T cells. Interestingly, these clones are not cytotoxic (Sallusto, F., unpublished data) and may thus provide cognate help to antigen-specific B lymphocytes. If this is a general case, the polarization of the specific CD8+ cells towards help or cytotoxicity may become a critical factor controlling the stimulation of an IgE response or its suppression (38, 39).

These results suggest that the response to carbohydrate antigens may be part of the normal response to environmental allergens and that this response can be polarized in atopic patients. Polysaccharides are generally considered typical T-independent antigens. However, T cells that modulate responses to polysaccharides have been observed in several systems (4046). Our results might explain these earlier observations and raise the possibility that T cell recognition of carbohydrates may play a role in the immune response.

Acknowledgments

We thank Adriano Mari for blood samples, Mark Dessing for calcium flux analysis, David Avila for peptide analysis, Vincenzo Barnaba, Gennaro De Libero, and Gabriella Di Felice for discussion. We are grateful to Antonio Lanzavecchia for helpful advice, discussion, and critical reading of the manuscript.

This work has been supported in part by a grant from Istituto Superiore di Sanità, Project on Tuberculosis.

References

References
1
Porcelli
SA
,
Morita
CT
,
Modlin
RL
T cell recognition of non-peptide antigens
Curr Opin Immunol
1996
8
510
516
[PubMed]
2
Porcelli
SH
,
Morita
CT
,
Brenner
MB
CD1b restricts the response of human CD4−CD8−T lymphocytes to a microbial antigen
Nature (Lond)
1992
360
593
596
[PubMed]
3
Beckman
M
,
Porcelli
SA
,
Morita
CT
,
Behar
SM
,
Furlong
T
,
Brenner
MB
Recognition of a lipid antigen by CD1-restricted αβ+T cells
Nature (Lond)
1994
372
691
694
[PubMed]
4
Sieling
PA
,
Chatterjee
D
,
Porcelli
SA
,
Prigozy
TI
,
Soriano
T
,
Brenner
MB
,
Kronenberg
M
,
Brennan
PJ
,
Modlin
RL
CD1-restricted T cell recognition of microbial lipoglycans
Science (Wash DC)
1995
269
227
230
[PubMed]
5
Constant
P
,
Davodeau
F
,
Peyrat
M-A
,
Poquet
Y
,
Puzo
G
,
Bonneville
M
,
Fournié
J-J
Stimulation of human γδ T cells by nonpeptidic mycobacterial ligands
Science (Wash DC)
1994
264
267
270
[PubMed]
6
Morita
CT
,
Beckman
EM
,
Bukowski
JF
,
Tanaka
Y
,
Band
H
,
Bloom
BR
,
Golan
DE
,
Brenner
MB
Direct presentation of nonpeptide prenyl pyrophosphate antigens to human γδ T cells
Immunity
1995
3
495
507
[PubMed]
7
Burk
MR
,
Mori
L
,
De Libero
G
Human Vγ9-Vδ2 cells are stimulated in a cross-reactive fashion by a variety of phosphorylated metabolites
Eur J Immunol
1995
25
2052
2058
[PubMed]
8
Harding
CV
,
Kihlberg
J
,
Elofsson
M
,
Magnusson
G
,
Unanue
ER
Glycopeptide bind MHC molecules and elicit specific T cell responses
J Immunol
1993
151
2419
2425
[PubMed]
9
Haurum
JS
,
Arsequell
G
,
Lellouch
AC
,
Wong
SYC
,
Dweck
RA
,
McMichael
AJ
,
Elliott
T
Recognition of carbohydrate by major histocompatibility complex class I-restricted, glycopeptide-specific cytotoxic T lymphocytes
J Exp Med
1994
180
739
744
[PubMed]
10
Michaelsson
E
,
Malmsdtrom
V
,
Reis
S
,
Engstrom
A
,
Burkhardt
H
,
Holmdahl
R
T cell recognition of carbohydrates on type II collagen
J Exp Med
1994
180
745
749
[PubMed]
11
Dudler
T
,
Altmann
F
,
Carballido
JM
,
Blaser
K
Carbohydrate-dependent, HLA class II-restricted, human T cell response to bee venom allergen phospholipase A2 in allergic patients
Eur J Immunol
1995
25
538
542
[PubMed]
12
Martin
S
,
Weltzien
HU
T cell recognition of haptens, a molecular view
Int Arch Allergy Immunol
1994
104
10
16
[PubMed]
13
Ishioka
GY
,
Lamont
AG
,
Thomson
D
,
Bulbow
N
,
Gaeta
FCA
,
Sette
A
,
Grey
HM
MHC interaction and T cell recognition of carbohydrate and glycopeptide
J Immunol
1992
148
2446
2451
[PubMed]
14
Harding
CV
,
Roof
RW
,
Allen
PM
,
Unanue
ER
Effect of pH and polysaccharides on peptide binding to class II major histocompatibility complex molecules
Proc Natl Acad Sci USA
1991
88
2740
2744
[PubMed]
15
D'Amato
G
,
Lobefalo
G
Allergenic pollens in the southern Mediterranean area
J Allergy Clin Immunol
1989
83
116
122
[PubMed]
16
Sallusto
F
,
Corinti
S
,
Pini
C
,
Biocca
MM
,
Bruno
G
,
Di Felice
G
Parietaria judaica-specific T-cell clones from atopic patients: heterogeneity in restriction, Vβ usage, and cytokine profile
J Allergy Clin Immunol
1996
97
627
637
[PubMed]
17
Sallusto
F
,
Lanzavecchia
A
Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony–stimulating factor plus interleukin 4 and downregulated by tumor necrosis factor α
J Exp Med
1994
179
1109
1118
[PubMed]
18
Lanzavecchia
A
Antigen-specific interaction between T and B cell
Nature (Lond)
1985
314
537
539
[PubMed]
19
Elsasser-Beile
U
,
von Kleist
S
,
Gallati
H
Evaluation of a test system for measuring cytokine production in human whole blood cell cultures
J Immunol Methods
1991
139
191
195
[PubMed]
20
Giuliani
A
,
Pini
C
,
Bonini
S
,
Mucci
N
,
Ferroni
L
,
Vicari
G
Isolation and purification of a major allergen from Parietaria officinalispollen
Allergy
1987
42
434
440
[PubMed]
21
Bradford
MM
A rapid and sensitive method for quantitation of microgram quantities of protein utilizing the principle of protein-dye binding
Anal Biochem
1976
72
248
254
[PubMed]
22
Edge
ASB
,
Faltinek
CR
,
Hof
L
,
Reichert
LE
Jr
,
Weber
P
Deglycosylation of glycoproteins by trifluoromethanesulfonic acid
Anal Biochem
1981
118
131
137
[PubMed]
23
Spiro
RG
Studies on the renal glomerular basement membrane. Nature of the carbohydrate units and their attachment to the peptide portion
J Biol Chem
1967
242
1943
1948
24
Ashwell, G. 1966. New colorimetric methods of sugar analysis. In Methods in Enzymology. S.P. Colowick and N.O. Kaplan, editors. Academic Press Inc., New York. VIII:93–95.
25
Valitutti
S
,
Dessing
M
,
Aktories
K
,
Gallati
H
,
Lanzavecchia
A
Sustained signaling leading to T cell activation results from prolonged T cell receptor occupancy. Role of T cell actin cytoskeleton
J Exp Med
1995
181
577
584
[PubMed]
26
De Palma
R
,
Gorsky
J
Restricted and conserved T-cell repertoire involved in allorecognition of class II major histocompatibility complex
Proc Natl Acad Sci USA
1995
92
8836
8840
[PubMed]
27
Pamer
EG
,
Wang
C-R
,
Flaherty
L
,
Fischer-Lindahl
K
,
Bevan
MJ
H-2M3 presents a Listeria monocytogenespeptide to cytotoxic T lymphocytes
Cell
1992
70
215
223
[PubMed]
28
Nataraj
C
,
Brown
ML
,
Poston
RM
,
Shawar
SM
,
Rich
RR
,
Fisher
K
,
Lindahl
,
Kurlander
RJ
H2-M3wt-restricted, Listeria monocytogenes-specific CD8 T cells recognize a novel, hydrophobic, protease-resistant, periodate-sensitive antigen
Int Immunol
1996
8
367
378
[PubMed]
29
Steinman
RM
The dendritic cell system and its role in immunogenicity
Annu Rev Immunol
1991
9
271
295
[PubMed]
30
Valitutti
S
,
Muller
S
,
Cella
M
,
Padovan
E
,
Lanzavecchia
A
Serial triggering of many T-cell receptors by a few peptide-MHC complexes
Nature (Lond)
1995
375
148
151
[PubMed]
31
Di Felice
G
,
Mari
A
,
Mucci
N
,
Afferni
C
,
Bruno
G
,
Pini
C
T cell responses to a Parietaria judaica pollen extract: comparison between Parietaria-sensitive patients, other atopics and healthy controls
Allergy
1989
44
322
329
[PubMed]
32
Salgame
P
,
Abrams
JS
,
Clayberger
C
,
Goldstein
H
,
Convit
J
,
Modlin
RL
,
Bloom
BR
Differing lymphokine profiles of functional subsets of human CD4 and CD8 T cell clones
Science (Wash DC)
1991
254
279
281
[PubMed]
33
Sad
S
,
Marcotte
R
,
Mosmann
TR
Cytokine-induced differentiation of precursor mouse CD8+ T cells into cytotoxic CD8+ T cells secreting Th1 or Th2 cytokines
Immunity
1995
2
271
279
[PubMed]
34
Romagnani
S
Lymphokine production by human T cells in disease state
Annu Rev Immunol
1994
12
227
257
[PubMed]
35
Yoshimoto
T
,
Bendelac
A
,
Hu-Li
J
,
Paul
WE
Defective IgE production by SJL mice is linked to the absence of CD4+, NK1.1+ T cells that promptly produce interleukin 4
Proc Natl Acad Sci USA
1995
92
11931
11934
[PubMed]
36
Maggi
E
,
Giudizi
MG
,
Biagiotti
R
,
Annunziato
F
,
Manetti
R
,
Piccinni
M-P
,
Parronchi
P
,
Sampognaro
S
,
Giannarini
L
,
Zuccati
G
et al
Th2-like CD8+ T cells showing B cell helper function and reduced cytolytic activity in human immunodeficiency virus type I infection
J Exp Med
1994
180
489
495
[PubMed]
37
Paganelli
R
,
Scala
E
,
Ansotegui
IJ
,
Ausiello
CM
,
Halapi
E
,
Fanales-Belasio
E
,
D'Offizi
G
,
Mezzaroma
I
,
Pandolfi
F
,
Fiorilli
M
et al
CD8+ T lymphocytes provide helper activity for IgE synthesis in human immunodeficiency virus-infected patients with hyper-IgE
J Exp Med
1995
181
423
428
[PubMed]
38
Bloom
BR
,
Modlin
RL
,
Salgame
P
Stigma variations: observations on suppressor T cells and leprosy
Annu Rev Immunol
1992
10
453
487
[PubMed]
39
Seder
RA
,
Le Gros
GG
The functional role of CD8+ T helper type 2 cells
J Exp Med
1995
181
5
7
[PubMed]
40
Baer
H
,
Chaparas
SD
Tuberculin reactivity of a carbohydrate component of unheated BCG culture filtrate
Science (Wash DC)
1964
146
245
247
[PubMed]
41
Baker
PJ
,
Stashak
PW
,
Amsbaugh
DF
,
Prescott
B
,
Barth
RF
Evidence for the existence of two functionally distinct types of cells which regulate the antibody response to type III pneumococcal polysaccharide
J Immunol
1970
105
1581
1586
[PubMed]
42
Shapiro
ME
,
Onderdonk
AB
,
Kasper
DL
,
Finberg
RW
Cellular immunity to Bacteroides fragiliscapsular polysaccharide
J Exp Med
1982
155
1188
1197
[PubMed]
43
Bellamy
A
,
Davison
AN
,
Feldmann
M
Derivation of ganglioside-specific T cell lines of suppressor or helper phenotype from cerebrospinal fluid of multiple sclerosis patients
J Neuroimmunol
1986
12
107
120
[PubMed]
44
Katsuki
M
,
Kakimoto
K
,
Kawata
S
,
Kotani
S
,
Koga
T
Induction of delayed-type hypersensitivity by the T cell line specific to bacterial peptidoglycans
J Immunol
1987
139
3570
3572
[PubMed]
45
Henningsson
CM
,
Selvaraj
S
,
MacLean
GD
,
Suresh
MR
,
Noujaim
AA
,
Longenecker
BM
T cell recognition of a tumor-associated glycoprotein and its synthetic carbohydrate epitopes: stimulation of anti-cancer T cell immunity in vivo
Cancer Immunol Immunother
1987
25
231
241
[PubMed]
46
Moll
H
,
Mitchell
GF
,
McConville
MJ
,
Handman
E
Evidence for T-cell recognition in mice of a purified lipophosphoglycan from Leishmania major.
Infect Immun
1989
57
3349
3356
[PubMed]

S. Corinti and R. De Palma contributed equally to this work.

1

Abbreviations used in this paper: B-LCL, EBV-transformed lymphoblastoid cell lines; DC, dendritic cell; GC-MS, gas chromatography–mass spectrometry; PjCHOAg, carbohydrate antigen present in PjE; PjE, Parietaria judaica pollen extract; rpHPLC, reverse-phase HPLC; SP, sodium periodate; TFMS, trifluoromethane-sulphonic acid.

Author notes

Address correspondence to Dr. Federica Sallusto, Basel Institute for Immunology, Grenzacherstrasse 487, CH-4005 Basel, Switzerland. Phone: 41-61-605-1348; FAX: 41-61-605-1364; E-mail: sallusto@bii.ch