Haptoglobin directly affects T cells and suppresses T helper cell type 2 cytokine release

Abstract

T helper cell type 1 (Th1) and type 2 (Th2) immune responses are characterized by a different pattern of cytokine expression following T-cell activation. Alterations of the ratio of Th1 to Th2 cells are important determinants of susceptibility to viral and parasitic infections, allergies, anti-tumour responses, and autoimmunity. In this work we bring new evidence for an effect of haptoglobin (Hp), a positive acute-phase protein, on T-lymphocyte functions. We show that Hp specifically interacts with both resting and activated CD4⁺ and CD8⁺ T cells. This specific binding results in a strong suppression of induced T-cell proliferation. In addition, Hp exhibits a strong in vitro inhibitory effect on Th2 cytokine release, while the production of interferon-γ (IFN-γ) and interleukin-2 (IL-2) is only slightly inhibited at high Hp doses. As a result, the presence of Hp promotes Th1 activation over Th2 activation in vivo as evidenced in Hp-deficient mice. Anti-CD3 monoclonal antibody injection indeed resulted in predominant IL-4 production in Hp^−/− mice, in contrast to predominant IFN-γ production in Hp^+/+ mice. We conclude that Hp plays a modulating role on the Th1/Th2 balance by promoting a dominant Th1 cellular response. This points to a role of acute-phase proteins in balancing immune responses.

Introduction

Different subsets of T helper lymphocytes, termed T helper 1 (Th1) and T helper 2 (Th2) cells, are responsible for induction and regulation of cellular and humoral responses, respectively. Interleukin-2 (IL-2) and interferon-γ (IFN-γ) are produced by Th1 cells and favour cell-mediated responses, whereas IL-4, IL-5, IL-10 and IL-13 are produced by Th2 cells and mediate predominantly humoral and eosinophilic responses. The Th1/Th2 concept suggests that disturbances of the Th1–Th2 balance are responsible for the development and/or severity of various immunological diseases.¹

Haptoglobin (Hp) is a positive acute-phase protein produced mainly by liver cells.² Besides its role as a haemoglobin scavenger, it has been shown to behave as an angiogenic,³ antioxidant,² and anti-inflammatory factor.^2,4 Hp and its variants have been reported to be potent immunosuppressors of lymphocyte function.^5–9 At high concentrations approximating those found in patients with cancer, Hp inhibits phytohaemagglutinin (PHA)-induced blastogenesis of lymphocytes.¹⁰ It was suggested that this inhibition contributes to the protection of tumours against immune attacks. A similar study has shown that Hp suppresses T-lymphocyte responses to PHA and concanavalin A and B-cell mitogenesis in response to lipopolysaccharides (LPS).⁵ Hp preparations from burned patients significantly suppress cell proliferation and IL-2 secretion of murine thymocytes.¹¹ Hp also interferes with Langerhans cell function by preventing them from undergoing functional transformation and from activating autologous T cells.¹² However, it is not clear whether this is an indirect effect through modulation of antigen-presenting cells functions, or whether Hp acts directly on T cells.

In the present work, we explore the binding of Hp to CD4⁺ and CD8⁺ T-cell subsets as well as the functional effects of Hp on T cells in the absence of accessory cells. T-cell proliferation and the release of Th1 and Th2 cytokines induced by polyclonal stimulation were explored for this purpose. The in vivo relevance of our findings was confirmed in Hp knockout mice.

Materials and methods

Reagents and monoclonal antibodies

PHA was purchased from Wellcome Diagnostics (Dartford, UK), Phorbol myristate acetate (PMA) and ionomycin were obtained from Sigma (St. Louis, MO). Tetanus toxoid was supplied by RIVM (Bilthoven, the Netherlands) and Apyrogen X-Vivo15 serum-free culture medium and phosphate-buffered saline (PBS) were purchased from BioWhitaker (Walkersville, MD).

Mouse anti-human CD3 monoclonal antibody [mAb; clone UCHT-1 (American Type Culture Collection (ATCC), Rockville, MD, immunoglobulin G1 (IgG1)] was purified by Protein G affinity chromatography in our laboratory. Humanized anti-Tac mAb [directed against the p55 chain of human IL-2 receptor (IL-2R)] was obtained from Hoffmann–la Roche Inc. (Nutley, NJ) and humanized Mikβ2 (directed against the p75 chain) was purchased from Pharmingen (San Diego, CA). Mouse anti-human IL-4R mAb (clone 25463.11) was purchased from R&D (Minneapolis, MN). Hamster anti-mouse CD3 mAb directed against the CD3 ɛ-chain was purified by affinity chromatography from 145 to 2C11 hybridoma culture supernatant and tested for endotoxin content as previously described.¹³ For flow cytometry, phycoerythrin or fluorescein isothiocyanate (FITC) - labelled mAbs and isotype-matched control mAbs were from Pharmingen.

Cell lines

The P815 cell line (ATCC) is a natural killer (NK)-resistant DBA/2-derived mouse mastocytoma cell line expressing FcγRII and FcγRIII. Human CD80^c (hCD80)-transfected P815 cell line was a gift from Dr L. Lanier (DNAX Research Institute for Molecular and Cellular Biology, Palo Alto, CA). The two cell lines were cultured in complete medium supplemented with gentamycin 50 μg/ml, sodium pyruvate 1 mm, non-essential amino acids 1 : 100, 2β-mercaptoethanol 50 μm, and 10% fetal calf serum. Geneticin (1 μg/ml) was added once a week to transfected cell lines to select transfectants. Cells were treated with mitomycin C (50 μg/5 × 10⁶ cells/ml) for 30 min and washed five times with PBS prior to use in cultures in combination with T lymphocytes.

Purification of Hp from human serum

Hp from healthy individuals with an Hp1-1 or Hp2-1 phenotype was purified by affinity chromatography using a rabbit anti-human Hp (Dako A/S. Glostrup, Denmark) Sepharose 4B column (Kabi Pharmacia, Uppsala, Sweden). Serum was applied to the column and eluted with 0·1 m PBS, pH 7·4, and the Hp-free serum was collected. The column was washed overnight with the same buffer. Then Hp was eluted with 0·1 m glycine–HCl, pH 2·3. The eluted Hp fractions were collected in 1 m Tris buffer to neutralize the acidic pH. The Hp-containing fractions were dialysed overnight against a 50-fold excess of 0·1 m PBS, pH 7·4. The Hp content was quantified by measuring the absorbance at 280 nm using a molar extinction coefficient of 10·2 × 10⁴. Then it was concentrated by ultrafiltration using a filter with a MW cut-off of 30 000 (Amicon, Danvers, MA), and sterilized by filtration through a 0·22-μm pore size filter (Gelman Sciences Inc., Ann Arbor, MI). The purity of obtained preparations was evaluated by either sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) or immunoelectrophoresis, a very slight contamination with IgG and IgM was detected.

Preparation of FITC-labelled Hp

Hp preparations were passed through a Protein G–Sepharose column to remove contaminating IgG. Hp (1–2 mg in 1 ml solution) was passed through an Econo 10DG column (BioRad, Richmond, CA) to exchange PBS with the labelling buffer (boric acid 0·05 m, NaCl 0·15 m, pH 9·2). Twenty microlitres of FITC solution (5 mg/ml dimethylsulphoxide) was added per mg of Hp and incubated for 2 hr at room temperature under gentle and continuous stirring. Unbound FITC was removed by gel filtration on the Econo 10 DG column previously equilibrated with PBS. The obtained preparation was further dialysed overnight against PBS. Hp–FITC fractions with an F/P (FITC/protein) ratio of 2 or higher as determined by spectrophotometry were concentrated and kept at 4° until use.

Isolation of peripheral blood mononuclear cells (PBMC)

PBMCs from the blood of healthy individuals were isolated on Lymphoprep™, Nycomed Pharma AS (Oslo, Norway). After three washings in PBS, cells were resuspended in X-VIVO 15 serum-free medium. Cell viability was always > 96% as assessed by trypan blue dye exlusion method.

Isolation of T lymphocytes and CD4⁺ and CD8⁺ T lymphocytes

For T- and NK-cell isolation, monocytes were first removed from PBMCs by cold agglutination. To this end, a 15-ml tube containing 50 × 10⁶ PBMCs in 10 ml of culture medium was slowly rotated for 30 min at 4°. Monocyte aggregates were allowed to sediment for 15 min in an ice bath, and the non-aggregated cells were aspirated. T and NK cells were then further purified using complement-fixing mAbs in Lympho-KWIK-T, One Lambda Inc. (Los Angeles, CA). Lymphocytes (20 × 10⁶) were resuspended in 200 μl of medium. Lympho-KWIK-T (0·8 ml) was added, and the mixture was incubated for 1 hr at 37°. Cells were centrifuged at 1000 g for 4 min and washed twice. To get pure T cells, NK cells were removed by treatment with complement-fixing NKH-1A anti-CD56 mAb (IgM; Coulter, Hialeah, FL) and anti-Leu11b anti-CD16 mAb (IgM; (Becton Dickinson, Erembodegem, Belgium) followed by a second Lympho-KWIK-T treatment. CD4⁺ and CD8⁺ T-lymphocyte subsets were isolated from PBMC cultures by magnetic positive immunoselection using anti-CD4- and anti-CD8-coated Dynabeads, followed by bead detachment. The procedure was carried out according to the manufacturer's instructions (Dynal, Oslo, Norway).

All cultures showed a purity exceeding 98% and a viability exceeding 96%.

Proliferation assays

To avoid the interference of serum factors with Hp, cells were cultured in serum-free medium supplemented with antibiotics.

Different mitogens have been used: PHA (5 μg/ml), immobilized anti-CD3 mAbs (2 μg/ml), PMA + ionomycin (1 ng/ml + 1 μg/ml, respectively) or tetanus toxoid (0·36 μg/ml).

Cells were incubated at 37° in a humidified atmosphere containing 5% CO₂/95% air for 3 or 6 days depending on the stimulant used. Cultures were pulsed with 0·5 μCi/well [³H]TdR for the final 6 hr. The incorporation of labelled nucleotide was determined by scintillation counting in a β-counter (Packard, Meriden, CT) after harvesting the cultures onto a glass-fibre filter paper using a semi-automated cell harvester (Skatron Instruments S.A., Lier, Norway).

In vitro and cytokine induction

T cells (0·5 × 10⁶/ml) were incubated in X-Vivo 15 apyrogene serum-free medium. The primary stimulatory signal was provided by UCHT-1 anti-CD3 mAbs at a concentration of 1 μg/ml, co-stimulatory signal was provided by CD80-transfected P815 cells at a T-cell : cell line ratio of 1 : 1. Blockade of IL-2R and IL-4R was performed using 2·5 μg/ml blocking mAbs where appropriate. Hp was added at the beginning of the culture at different doses (50, 250, 500, and 1000 μg/ml), then incubation was prolonged for 72 hr. Cell cultures were incubated in 24-well flat-bottomed plates at 37° in a 5% CO₂ humidified atmosphere.

Binding assays

CD4⁺ and CD8⁺ T cells (0·5 × 10⁶) were incubated overnight in the absence or presence of PMA + ionomycin (1 ng/ml + 1 μg/ml, respectively). After washing with PBS, cells were incubated with human serum albumin 0·1%/PBS for 15 min. Hp–FITC was added and cells were incubated for 30 min at 4°, then washed and fixed with paraformaldehyde 1% in PBS.

To show the specificity of the binding of Hp–FITC to CD4 or CD8 T cells, either resting or activated, cells were preincubated with a 25-fold excess of unlabelled Hp prior to incubation with Hp–FITC.

Mice

C57BL/6J Hp knockout mice (Hp^−/−) were generated by Dr S. K. Lim (The National University of Singapore) by homologous recombination as described previously¹⁴ and were kindly provided by Dr F. Berger (University of South Carolina, Columbia, SC). Control wild-type mice (Hp^+/+) of the same strain were obtained from The Jackson Laboratory (Bar Harbor, ME). Mice were housed at the Gasthuisberg animal facility in specific pathogen-free conditions, allowing free access to food and water. All procedures involving animals were performed according to the guidelines of the Animal Ethical Committee of the KULeuven.

Induction of cytokine release syndrome in mice

Seven to eight-week-old male Hp^+/+ and Hp^−/− C57BL/6J mice were given an intraperitoneal injection of either 10 μg of LPS (Escherichia coli, 0111:B4 serotype) in 100 μl saline or 100 μl saline alone. After 36 hr, every group was divided into two groups that were intraperitoneally administered either 10 μg of a hamster anti-mouse CD3 mAb (145-2C11) in 100 μl saline or saline alone. After 90 min, blood was withdrawn at the orbital plexus under light ether anaesthesia and the serum was harvested for assessment of IL-4 and IFN-γ release.

Assessment of cytokine production

The production of IL-2, IL-4, IL-5, IL-10, IL-13 and IFN-γ by stimulated human T cells in the presence or absence of different Hp doses was determined in cell-free supernatants by quantitative enzyme-linked immunosorbent assay (ELISA). Combinations of unlabelled and biotin-labelled mAbs were used. The coating mAbs used were 860A4B3 (Medgenix Diagnostics, Fleurus, Belgium), TRFK5 (Pharmingen), JES3-9D7 (Pharmingen), JES10-5A2 (Pharmingen), and 350B10G6 (Medgenix Diagnostics) raised against IL-4, IL-5, IL-10, IL-13, and IFN-γ, respectively. Biotinylated mAbs 860F10H12 (Medgenix Diagnostics), JES1-SA10 (Pharmingen), JES3-12G8 (Pharmingen), B69-2 (Pharmingen) and 67F12A8 (Medgenix Diagnostics) were used for detection of IL-4, IL-5, IL-10, IL-13, and IFN-γ, respectively. Streptavidin–peroxidase conjugate (Lucron, Brussels, Belgium) and 3,3′,5,5′-tetramethylbenzidine (Acros Organics, Pittsburgh, PA) were used for detection.

IL-2 production was assessed by IL-2 Duoset ELISA (Genzyme, Cambridge, MA) using the protocol provided by the manufacturer.

Human recombinant IL-2 (Pharmingen), IL-4 (Biosource, Camarillo, CA), IL-5 (Pharmingen), IL-10 (Pharmingen), IL-13 (Pharmingen), and IFN-γ (Biosource) were used as standards.

Assay sensitivity was 1·25 pg/ml for IL-2, 5 pg/ml for IL-4, and 10 pg/ml for IL-5, IL-10, IL-13, and IFN-γ.

Mouse IL-4 and IFN-γ were measured in sera using ELISA CytoSets™ kits from Biosource. The sensitivity of the assays was 2 pg/ml.

Statistical analysis

Statistical analysis was performed using an analysis of variance (anova) test followed by a Dunnett's comparison test (instat, GraphPad Software, San Diego, CA) allowing comparison between effects seen in the presence and absence of Hp. Comparison between effects in different mice groups was performed using a Student's t-test. Statistical significance was set at P < 0·05.

Results

Hp suppresses PHA-induced T-cell proliferation in a dose-dependent manner

The anti-proliferative potential of Hp has been shown in different previous reports. However, Hp used in those earlier studies was prepared by either acidic precipitation,⁷ which is known to affect protein structure, or was taken from patients with malignant tumours, who are known to produce abnormal forms of Hp.^6,8,9

In the present work, Hp was purified from the sera of healthy volunteers by affinity chromatography. The final product was very pure, as assessed by SDS–PAGE. Biological activity was confirmed by its ability to bind haemoglobin (data not shown).

When added to PBMCs concomitantly with PHA as a polyclonal stimulator, Hp at a dose of 250 μg/ml significantly inhibited thymidine uptake. The inhibitory effect became stronger with increasing Hp concentrations. In the presence of Hp at 800 μg, T-cell proliferation was abolished (Fig. 1a). Interestingly, very low doses of Hp tended to enhance cell proliferation, though this enhancement was statistically not significant.

Figure 1.

Hp inhibits T-cell proliferation. (a) Dose-dependent inhibitory effects of Hp on PHA-induced T-cell proliferation: 0·2 × 10⁶ PBMC/well were stimulated with PHA (5 μg/ml) in the presence of increasing Hp concentrations. T-cell proliferation was measured by [³H]thymidine incorporation assay at day 3. Results are shown as the mean ± SEM from three donors. (b) The inhibitory effect of Hp is compared to that of human serum albumin (HSA). One representative experiment is shown. (c) Hp1-1 and Hp2-1 similarly suppress T-cell proliferation following different stimulatory triggers; 0·5 × 10⁶ PBMC/ml were stimulated with either 5 μg PHA or 1 ng PMA/1 μg ionomycin for 3 days, or alternatively with 0·36 μg tetanus toxoid for 6 days, in the absence or presence of 500 μg of either Hp1-1 or Hp2-1. T-cell proliferation was evaluated by [³H]thymidine incorporation assay. Results represent the mean ± SEM in three independent experiments. (d) The effect of 500 μg/ml of Hp was evaluated on PHA-stimulated PBMC, T + NK, and T-cell cultures isolated from the same donors. The inhibition of T-cell proliferation was evaluated after a 72-h incubation and expressed as percentage of control cultures where no Hp was added. Results represent the mean ± SEM from three experiments. P > 0·05 for all comparisons between cultures.

Human serum albumin did not show any effect on T-cell proliferation when used at the same concentrations as Hp under the same experimental settings (Fig. 1b).

Effect of different phenotypic Hp forms on T-cell proliferation

In humans, Hp occurs in three major phenotypes (Hp1-1, Hp2-1 and Hp2-2) differing in subunit constitution and molecular weight.⁴ We stimulated PBMCs with different triggers of T-lymphocyte proliferation and the effect of two Hp phenotypes (Hp1-1 and Hp2-1) on thymidine incorporation was tested at a single dose (500 μg/ml). As shown in Fig. 1(c), both Hp forms exerted similar inhibitory effects on T-cell proliferation in response to PHA, PMA/ionomycin and tetanus toxoid.

T cells undergo a direct effect of Hp

Our group has previously identified the Mac-1 leukocyte integrin β₂ (CR3, CD11b/CD18) as a receptor for Hp on monocytes, macrophages, granulocytes, NK cells, and small subpopulations of CD8⁺ T lymphocytes and B lymphocytes.¹⁵ The CR3 receptor was shown to play a regulatory role in cell function.¹⁶ To study whether the presence of a CR3⁺ cell type was required for the anti-proliferative effects of Hp, we isolated different cell types from the blood of healthy donors (i.e. PBMC, T + NK, and T cells), and cultured these subsets with PHA in the presence or absence of 500 μg of Hp1-1. Figure 1(d) shows that Hp is similarly suppressive for T-cell proliferation in all three types of cultures, thus indicating a direct effect on T cells and discarding any contribution of monocytes and NK cells to the observed effects.

Hp binds to resting and activated CD4⁺ and CD8⁺ T cells

Binding to T cells has been reported for various acute-phase proteins such as fibrinogen, serum amyloid A, serum amyloid P, C-reactive protein, or their degradation products.^17–21 Here we addressed the question whether Hp–FITC can bind to activated T cells. We isolated CD4⁺ and CD8⁺ T lymphocytes by positive selection and stimulated them overnight with PMA + ionomycin. Cells were then washed and incubated with Hp–FITC. Most T cells appear to bind Hp–FITC at saturating doses, even in their resting state. PMA/ionomycin stimulation increased the binding of Hp–FITC per cell (Fig. 2a). Addition of labelled Hp subsequently to a 25-fold excess of unlabelled Hp strongly decreased the mean fluorescence intensity on both CD4⁺ and CD8⁺ T cells (Fig. 2b), pointing to the specificity of the binding.

Figure 2.

Hp binds to resting and activated CD4⁺ and CD8⁺ T lymphocytes. CD4⁺ and CD8⁺ T cells (0·5 × 10⁶) were incubated overnight in the absence or presence of PMA + ionomycin (1 ng/ml + 1 μg/ml, respectively). After washing with PBS, cells were incubated with human serum albumin 0·1%/PBS for 15 min and washed with PBS again. Hp–FITC was added and cells were incubated for 30 min at 4°, washed and fixed with paraformaldehyde 1% in PBS. (a) Effect of increasing doses of Hp–FITC. Hp–FITC binding was assessed by flow cytometry and results were expressed as either percentage positive cells (upper panels) or mean fluorescence intensity (MFI) (lower panels) of 10⁴ of resting (○) or activated (•) T cells. (b) Representative histograms of binding of Hp–FITC to CD4 or CD8 T cells, either resting or activated. The figure shows autofluorescence, binding of Hp–FITC (dotted line), and inhibition of Hp–FITC binding by a 25-fold excess of unlabelled Hp (solid line).

Hp differentially modulates cytokine release by stimulated human T helper lymphocytes in vitro

The influence of Hp on T helper cytokine release has not been studied before. T-cell receptor (TCR)/CD28 ligation of T cells represents an optimal trigger for both Th1 and Th2 cytokine induction. T cells were therefore stimulated with an anti-CD3 mAb presented on a cell line expressing the human CD80 co-stimulatory ligand, resulting in both TCR and CD28 ligation. Cells were incubated for 3 days and the release of IL-2, IL-5, IL-10, IL-13 and IFN-γ was evaluated. IL-4 has been reported to be internalized by T cells.²² To avoid this, we supplemented our cultures with an anti-IL-4R-blocking mAb, which led to higher amounts of IL-4 in the supernatants compared to where it was not used.²²

Under these experimental settings, Hp slightly increased the release of IL-2 when used at doses of 50 and 250 μg/ml, although this increase was not statistically significant. Hp had a slight inhibitory effect on IL-2 release when added at a dose of 1 mg/ml (P < 0·05). IFN-γ production, on the other hand, did not react to Hp (Fig. 3). Addition of increasing amounts of Hp resulted in a dose-dependent suppression of IL-4 release. This suppression became significant at an Hp dose of 500 μg/ml (P < 0·05). The release of IL-5 showed the same sensitivity to Hp as IL-4. IL-13 and IL-10, on the other hand, were more affected and were sensitive to Hp doses starting from 50 and 250 μg/ml, respectively (Fig. 3). Thus when classified according to Hp sensitivity, the order is IFN-γ < IL-2 < IL-4 = IL-5 < IL-10 < IL-13.

Figure 3.

Differential effects of Hp on T helper cytokine release; 0·5 × 10⁶ T cells in a final volume of 1 ml were cultured for 72 hr in the presence of hCD80-P815 cells (1 : 1 ratio) and 1 μg of anti-CD3 mAb. Hp was added at increasing doses (50, 250, 500 and 1000 μg). IL-2, IFN-γ, IL-4, IL-5, IL-10 and IL-13 release was measured using ELISA assays. To prevent IL-2 and IL-4 consumption, IL-2 and IL-4 receptors were blocked with mAbs (see the Materials and methods section). Results are shown as mean ± SEM from n = 4–6 experiments. *P < 0·05; **P < 0·01 versus control.

Hp deficiency in a murine model of cytokine release syndrome favours Th2 cytokine release

To confirm the relevance of our in vitro results on Th1/Th2 modulation by Hp, we used mice in which the Hp gene had been altered.¹⁴ Injection of the hamster anti-mouse CD3 mAb both intraperitoneally or intravenously led to a fast release of systemic IL-4 and IFN-γ in both Hp^+/+ and Hp^−/− mice. Interestingly, Hp^−/− mice released less IFN-γ and more IL-4 than did their Hp^+/+ counterparts (Fig. 4a). When mice were given a small dose of LPS to raise Hp (at least in wild-type mice) prior to the administration of the anti-CD3 antibody, similar differences in the cytokine profile between Hp^+/+ and Hp^−/− mice were found.

Figure 4.

Hp affects the Th1/Th2 balance in an animal model of cytokine release syndrome; 10 μg of the 145-2C11 hamster anti-mouse CD3 mAb was administered intraperitoneally to 7–8-week-old Hp^+/+ and Hp^−/− male C57BL/6J mice. Two other groups of mice received 10 μg of LPS 36 hr prior to injection of the antibody. Blood was collected after 90 min. Serum IL-4 and IFN-γ levels were determined (a) using the ELISA technique and the ratio IFN-γ/IL-4 was calculated (b). Results are reported as the mean serum cytokine level of six mice per group from one representative out of three performed experiments with similar results. ***P < 0·001 vs. Hp^−/− versus Hp^+/+.

The IFN-γ to IL-4 ratio, as shown in Fig. 4(b), highlights a remarkable shift of the immune response towards a dominant Th2 response in Hp^−/− mice (P < 0·001 for Hp^+/+ versus Hp^−/−). The shift was more pronounced when mice received endotoxins prior to anti-CD3 antibody injection.

Discussion

In this paper, we demonstrate that Hp has a direct inhibitory effect on T cells. First, Hp, in both its 1-1 and 2-1 forms, had a strong inhibitory effect on T-cell proliferation following either polyclonal (PHA or PMA + ionomycin) or antigenic stimulation (tetanus toxoid). Our findings indicate that Hp has anti-proliferative effects on T cells independently of whether activation occurs through TCR-dependent (PHA, tetanus toxoid) or TCR-independent (PMA/ionomycin) activation events, suggesting that Hp acts downstream of TCR activation.

The possibility that the decrease in thymidine uptake might be as a result of cell death is rather unlikely. T-cell viability was in fact maintained at a level higher than 96% in the presence of Hp. Inhibition of T-cell proliferation through decreased IL-2R expression was also excluded. When the IL-2R α chain (CD25) expression on PHA-stimulated T cells was determined on the last day of culture, no difference could be found between Hp-treated and untreated cells (unpublished data), which is in accordance with a previous report.⁸

Inhibition of T-cell proliferation with Hp was observed to a similar degree whether PBMC, monocyte and B-cell-depleted PBMC, or pure T cells were used. The effect of Hp is thus a direct one on T cells. We have previously identified CR3 (CD11b) as a receptor for Hp.¹⁵ Though only a small subset of T cells express CD11b, Wagner et al. have demonstrated a remarkable suppression of mitogen-induced T-cell proliferation and IL-2 release by triggering CR3 with either specific mAbs or C3bi-coated beads.¹⁶ Thus, it might be possible that Hp activates a subset of T cells which then suppresses the proliferation of the other cells. However, when we analysed Hp binding to T cells, we found that there is in fact binding to all T cells. So Hp might also exert an inhibitory effect on all T cells. In addition, we show that the fraction of cells that bind Hp–FITC is substantially higher than the subset of CD11b-expressing T cells. Although this is not yet clear cut evidence, one may conclude that the effect of Hp on T cells is exerted mainly through an as yet unidentified Hp receptor. Binding to CR3^– human mast cells was recently reported by our group.²³ Other reports mentioned unidentified Hp-binding receptors on neutrophils.^24,25

To assess further the functional consequences of Hp binding to T cells at the cytokine release level, we stimulated purified T cells in vitro and followed cytokine release in the presence of different Hp concentrations. Our results clearly show that while the effect of Hp on Th1 release was only slightly inhibitory (IL-2) or absent (IFN-γ), the effect on all tested Th2 cytokines was strongly inhibitory.

This effect was subsequently confirmed in an in vivo model. The anti-CD3-induced cytokine release syndrome provides a valuable model for the study of T-cell function in vivo. The dose of administered mAb is crucial to the onset of the syndrome, the latter is characterized by an acute reaction consisting of piloerection, diarrhoea, hypothermia, hypoglycaemia and hypomotility.¹³ The clinical syndrome as such developed similarly in both wild-type and Hp knockout mice, with a slight difference in body weight loss and piloerection, where the Hp^−/− appeared to be more affected (unpublished observation). The high dose of mAb administered (25 μg/mouse) may be too strong to be overcome by Hp. This model was used to study the effect of Hp on T-cell cytokine release, and IL-4 and IFN-γ were selected as typical representatives of Th2 and Th1 cells. IL-4 is typically expressed in Th2 cells, and also promotes the differentiation of Th2 lymphocytes, thereby playing a pivotal role in the regulation of humoral responses. Conversely, Th1-associated cytokines such as IFN-γ are central to the development of the cell-mediated, delayed type, and autoimmune responses.¹ When mice were given a low dose of anti-CD3 mAb (10 μg), Hp^+/+ mice produced high amounts of IFN-γ and low amounts of IL-4, whereas Hp^−/− mice showed an opposite profile.

In view of the differential regulatory effects exhibited by Hp on Th1 and Th2 cells both in vitro and in vivo, Hp may be an important endogenous regulator of immune reactions and a crucial molecule in the establishment of the Th1–Th2 balance. It is important to realize that anti-CD3-induced cytokine release induces the release of those cytokines (both in vitro and in vivo) for which the T cells have been programmed. The results thus suggest a direct effect of Hp on the differentiated Th2 cells, rather than on development of Th2 cells in Hp^−/− mice during post-thymic T-cell maturation. Animal models of Th1- and Th2-related diseases may bring insight into this aspect. Whether the preferential suppression by Hp of Th2 cytokine release may be regarded as a contribution of this particular protein, and probably of other acute-phase proteins, to favour a dominant Th1 response or as a mechanism of attenuation of the Th2 response, also requires further investigation.

Abbreviations

Hp

haptoglobin

Hp–FITC

fluorescein isothiocyanate-labelled haptoglobin

PBMC

peripheral blood mononuclear cells

PHA

phytohaemagglutinin

PMA

phorbol myristate acetate

TCR

T-cell receptor