The immune dysregulatory compound mercuric chloride induces integrin-mediated T-lymphocyte adhesion

Abstract

Exposure of Brown Norway rats to mercuric chloride induces systemic autoimmunity, involving T- and B-lymphocyte activation, (auto-)antibody production and multiorgan inflammation. Several divalent metal ions, such as Mg²⁺ and Mn²⁺, can activate binding of integrins to their ligands, thus causing lymphocyte adhesion. To test the hypothesis that Hg²⁺ acts in a similar way, we studied the effect of HgCl₂ on integrin-mediated T-cell adhesion. HgCl₂ induced cell–cell aggregation of human T lymphoblasts. Exposure of a human T-cell clone to HgCl₂ for 1 hr enhanced, in a dose-dependent way, cell binding to fibronectin (FN) and to intercellular adhesion molecules (ICAM) -1, -2 and -3. Furthermore, HgCl₂ induced strong binding of Jurkat T cells to FN. These effects of HgCl₂ were of similar magnitude as the effects of phorbol 12-myristate 13-acetate (PMA) or MnCl₂. Studies using blocking antibodies indicated the involvement of CD11a in binding to ICAMs, and of CD49d, CD49e, and CD29 in binding to FN. Adhesion to FN induced by HgCl₂ or by PMA, but not by MnCl₂, was dependent on temperature and on extracellular Ca²⁺ or Mg²⁺. Addition of cytochalasin B enhanced synergistically the FN adhesion induced by MnCl₂, whereas the effects of PMA and HgCl₂ were not modified. These results indicate that Hg²⁺ is a potent activator of T-cell adhesion, mediated by several integrins and ligands. In contrast to the effect of MnCl₂, HgCl₂-induced cell adhesion probably involves an intracellular pathway. Activation of integrins by HgCl₂ may play an important role in activation and migration of leucocytes involved in HgCl₂-induced immune dysregulation in vivo.

Introduction

A number of drugs and environmental pollutants can induce immune dysregulation in susceptible individuals, which may for example lead to autoimmune diseases involving the kidney as a target organ.¹ Using mercuric chloride as a model compound, immune dysregulation induced by chemicals has been frequently studied in rodents. Exposure of Brown Norway (BN) rats to HgCl₂ induces a systemic lupus-like autoimmune syndrome, characterized by activation and proliferation of T and B lymphocytes, high production of (auto-) antibodies, infiltration of leucocytes in many organs, glomerulonephritis and proteinuria.^2,3 T-cell depletion experiments revealed an essential role for T lymphocytes in disease induction.⁴ Furthermore, high production of interleukin-4 (IL-4) and immunoglobulin E (IgE) in HgCl₂-exposed BN rats indicate a strong and preferential activation of T helper type 2 (Th2) cells.^5,6 The mechanism of T-cell activation by exposure to HgCl₂ has not been identified in detail. HgCl₂ may alter the structure, processing, presentation and/or recognition of (auto-)antigens, thus leading to T-cell autoreactivity.⁷ In addition, non-antigen-specific T-lymphocyte activation by HgCl₂ is likely to be involved.^8,9

We have previously shown that in vivo exposure of BN rats to HgCl₂ enhances T-lymphocyte expression of various cell adhesion and activation molecules, such as lymphocyte function-associated antigen-1 (LFA-1; CD11a/CD18), intercellular adhesion molecule-1 (ICAM-1; CD54), integrin α₄ (CD49d) and OX40 (CD134), at an early stage of disease.¹⁰ Furthermore, it has been reported by Molina et al. that in vivo treatment with a blocking monoclonal antibody (mAb) directed against integrin α₄ inhibits a major part of autoimmune manifestations induced by HgCl₂ in BN rats.¹¹ These data indicate a role for cell adhesion molecules in the immune dysregulation induced by HgCl₂. In the present study, we examined whether HgCl₂ may have a direct effect on T-cell adhesion mediated by integrins.

Integrins are heterodimeric cell membrane receptors involved in cell-to-cell and cell-to-matrix adhesion. Integrin-mediated adhesive capacity is determined by the levels of expression of the receptors and their ligands.¹² Furthermore, the receptors require activation for ligand binding.¹³ Ligand binding results in cell adhesion, as well as in signal transduction via the integrin receptor,¹⁴ and, in several cases, via the ligand.¹⁵ Thus, these processes play an essential role in numerous functions of lymphocytes.

Activation of integrins can take place either as a result of extracellular events, or via an intracellular pathway. Certain integrin-binding mAbs can stimulate ligand binding, probably by stabilizing an active conformation of the receptor.^16–18 Furthermore, several divalent metal ions that bind to extracellular domains of integrins, such as Mg²⁺ and Mn²⁺, increase integrin affinity for its ligand,^19,20 which can also be demonstrated using isolated integrins and ligands in a cell-free system, indicating its independence on cell function.²¹ In contrast, integrin activation can also result from a process named inside-out signalling. This can be achieved by ligation of cell surface molecules, such as CD3 and CD2,^22,23 as well as by agents that induce intracellular signalling events, such as phorbol esters, which activate protein kinase C, and ionomycin, which elevates the intracellular calcium level.²⁴ Evidence has been presented that integrin activation by the phorbol ester phorbol 12-myristate 13-acetate (PMA) can take place independent of changes in integrin affinity, but involves modification of the association of integrins with cytoskeletal proteins, and their mobility in the plasma membrane.^20,25–27

Induction of lymphocyte adhesion by HgCl₂, thus modifying the physiological regulation of this process, could be of direct importance for the immune dysregulatory capacity of this compound. In the present study, we investigated the effect of exposure of T lymphocytes to HgCl₂ on their binding to several ligands of integrins: fibronectin (FN), ICAM-1, ICAM-2 and ICAM-3. The effects of HgCl₂ were compared to adhesion induced by MnCl₂ and by PMA. The results show that HgCl₂ induces integrin-mediated T-cell adhesion, which involves several integrins and ligands and is probably mediated via an intracellular pathway.

Materials and methods

Cells and cell culture

Human peripheral blood obtained from healthy donors was used for isolation of peripheral blood mononuclear cells (PBMC) by density gradient centrifugation, using Ficoll-Paque (Pharmacia Biotech, Uppsala, Sweden). T lymphoblasts were generated by culturing PBMC for 5–7 days in the presence of soluble OKT3 (mAb anti-CD3) and of a cytokine mixture, that was obtained from activated PBMC, as described,²⁸ using RPMI culture medium (RPMI-1640 supplemented with 10% heat-inactivated fetal calf serum, 100 IU/ml penicillin, 100 µg/ml streptomycin, and 2 mm glutamin). The human Th0 cell clone ACC6 (kindly provided by Dr M. L. Kapsenberg, Amsterdam, the Netherlands), was stimulated every 14 days by allogeneic feeder cells and phytohaemagglutinin, as described,²⁹ and cultured in Iscove's modified Dulbecco's medium (Gibco BRL Life Technologies, Paisley, UK), supplemented with 10% heat-inactivated normal human serum, 25 U/ml recombinant IL-2 (Eurocetus, Amsterdam, the Netherlands), 100 IU/ml penicillin, 100 µg/ml streptomycin and 2 mm glutamin. ACC6 T cells were used for experiments when they were in a resting phase, at day 10 or 11 after stimulation. The T-cell leukaemia line Jurkat [obtained from the American Type Tissue Collection (ATCC), Rockville, MD] was cultured in RPMI culture medium.

Monoclonal antibodies

Mouse mAb were used, directed against CD3 (OKT3, IgG2a; supernatant from hybridoma obtained from the ATCC), CD11a (F8.8,³⁰ IgG1; kindly provided by Dr A. Bloem, Utrecht, the Netherlands, and LFA 1/2,¹⁴ IgG1; donation from Dr R. A. W. van Lier, Amsterdam, The Netherlands), CD29 (4B4,³¹ IgG1; Coulter Immunology, Hialeah, FL), CD49d (HP2/1,^11,32 IgG1; a gift from Dr F. Sánchez-Madrid, Madrid, Spain), CD49e (SAM-1, IgG2b; provided by Dr C. G. Figdor, Nijmegen, the Netherlands) and CD54 (84H10,³³ IgG1; from Dr S. Shaw, Bethesda, MD). Monoclonal mouse IgG1 (Zymed Laboratories, San Francisco, CA) was applied as a non-binding control mAb. All antibody preparations used were free of NaN₃.

Homotypic aggregation assay

Cells were cultured in flat-bottom 96-well plates (2 × 10⁵/100 µl/well) at 37°, using adhesion buffer [Hanks' balanced salt solution (HBSS) without Ca²⁺ and Mg²⁺ (Gibco BRL), supplemented with d-glucose (2 mg/ml), bovine serum albumin (BSA; 10 mg/ml), MgCl₂ (1 mm) and CaCl₂ (1 mm)], in the presence or absence of HgCl₂ (Sigma, St Louis, MO). Monoclonal antibodies were preincubated with the cells during 30 min at 4°. Cell aggregation was examined by light microscopy after 3–4 hr of culture.

Production of chimeric proteins

Four different chimeric proteins containing the extracellular domains of either ICAM-1, ICAM-2, ICAM-3, or CD8, linked to the hinge domain of human IgG1, were expressed in COS cells. The appropriate expression plasmids for ICAM-1-IgG and CD8-IgG were kindly provided by Dr B. Seed (Boston, MA) and Dr A. Aruffo (Seattle, WA), and have been described previously.^34,35 The cDNA sequences encoding the two or five amino-terminal immunoglobulin-like domains of ICAM-2,³⁶ or ICAM-3,³⁷ respectively, were amplified by polymerase chain reaction (PCR) using synthetic oligonucleotides and cDNA from U937 cells. Primer sequences were as follows: ICAM-2-sense, 5′ GCCCGGTCGACGCCGCCACCATGTCCTCTTTCGGTTACAGGACCCTG (position 43–90); and antisense, 5′ GACTATGATGGGATCCTGGCTGTCCGACACAG-GCTCATAGATCTC (position 702–746);³⁶ ICAM-3-sense, 5′ CAGGTCGACGTAGCCATCGCCACCATGGTACCATCC (position 2–22); and antisense, 5′ CACGAAGACGGGGA-GATCTTGGGAGCTCCCAGCCTC (position 1439–1474).³⁷ Specificity of amplification products was verified by restriction site analysis. Subsequently, they were cloned into the CD8-IgG1-CDM8 vector,³⁵ followed by transfection into COS7m6 cells, as described.³⁷ Seven days after transfection, the supernatants were harvested and stored at −80° until use. Immunoprecipitation and sodium dodecyl sulphate–polyacrylamide gel electrophoresis revealed production of recombinant proteins of the expected molecular weights; concentrations of chimeric proteins in culture supernatants were determined by enzyme-linked immunosorbent assay.

Cell adhesion assays

For coating of chimeric proteins, flat-bottom 96-wells plates were precoated using goat anti-human Fcγ (10 µg/ml in 50 mm Tris, pH 9·5; Jackson Immunoresearch Laboratories, West Grove, PA) for 1 hr at 37°. After blocking (4% BSA in PBS; 1 hr at 37°) and washing, diluted COS cell-supernatants containing chimeric proteins (0·2 µg/ml in PBS) were incubated for 1 hr at 37°. Human plasma fibronectin (10 µg/ml in PBS), kindly provided by Dr J. van Mourik (Amsterdam, the Netherlands), was coated for 2 hr at 37°, followed by a blocking step (4% BSA in PBS; 1 hr at 37°). All plates were washed with adhesion buffer before use.

Jurkat T cells or ACC6 T cells were washed in adhesion buffer, and added to the coated wells at 0·8 × 10⁵/well and 1·25 × 10⁵/well, respectively, followed by centrifugation for 5 min at 20 g and incubation for 1 hr at 37°, unless otherwise indicated. Subsequently, the plates were gently washed, using warm adhesion buffer, five to seven times, and stored without buffer at −80° until measurement.

Incubation of the cells was performed in the presence or absence of the following stimuli: HgCl₂ (2·5 µm), MnCl₂ (500 µm; Sigma) and PMA (50 ng/ml; Sigma); they were applied at the indicated concentrations, unless otherwise specified. In some experiments, soluble purified OKT3 was used (5 µg/ml). All mAb were preincubated with the cells for 30 min at 4°. For testing the dependence on extracellular cations, cells were washed and incubated in Ca²⁺- and Mg²⁺-free adhesion buffer containing BSA which had been previously extensively dialysed against Ca²⁺- and Mg²⁺-free HBSS. Some experiments were performed using cytochalasin B (5 µg/ml; Sigma), which was preincubated with the cells for 30 min at 37°; subsequently, it was present during the assay. Cytochalasin B was dissolved in dimethyl sulphoxide, which was added to the control cultures at equivalent concentration (≤ 0·05% v/v).

Quantification of cell adhesion

Cells which were bound to the wells were quantified by a fluorometric assay, using the DNA-binding fluorochromes Hoechst 33258 and Syto 13 (both from Molecular Probes, Leiden, the Netherlands) for Jurkat T cells and ACC6 T cells, respectively. A cell lysis procedure was performed which was previously described by Rago et al.³⁸ Briefly, the plates were thawed at room temperature, and, after addition of 100 µl distilled water per well, were incubated at 37° for 1 hr. Subsequently, the plates were again frozen at −80°. After thawing, 100 µl TNE buffer (10 mm Tris, 2 m NaCl, 1 mm ethylenediaminetetraacetic acid, pH 7·4) containing Hoechst 33258 (32 µm) or Syto 13 (10 µm) was added, and fluorescence intensity was measured by a fluorometer (Perkin Elmer LS50), using appropriate wavelengths for excitation and emission (345/460 nm for Hoechst 33258 and 488/509 nm for Syto 13). The number of cells bound to each well was calculated by interpolation in a calibration line. Binding to the ligand and to the control-coating was assessed in parallel for each stimulus: ICAM-1, ICAM-2 and ICAM-3 were compared with CD8, and FN was compared with BSA. The percentage specific binding was calculated from triplicate wells using the following formula: 100 × [(number of cells bound to the ligand−number of cells bound to control coating)/number of input cells]. Results were summarized as the mean percentage specific binding to the ligand ±standard deviation.

Results

Integrin-mediated cell adhesion induced by HgCl₂

In order to examine whether exposure to HgCl₂ modifies cell adhesion, its effect on lymphocyte aggregation was examined. Addition of 5 µm HgCl₂ to T lymphoblasts clearly increased the number and size of cell aggregates, as compared to the control culture (Fig. 1a,b). Cell aggregation in control cultures and in HgCl₂-stimulated cultures was completely inhibited when the cells were preincubated with mAb against CD11a and CD54, indicating the involvement of LFA-1 and ICAM-1 in this process (Fig. 1c,d). Exposure of the Th0 clone ACC6 to HgCl₂ also induced cell aggregation (not shown). In contrast, the human leukaemic T-cell line Jurkat did not aggregate upon exposure to HgCl₂, to PMA, or to MnCl₂.

Figure 1.

Lymphocyte aggregation induced by HgCl₂. T lymphoblasts, generated from human PBMC as explained in the Materials and Methods, were incubated in control medium (a, c), or in medium containing 5 µm HgCl₂ (b, d). Cells were preincubated in the presence (c, d) or absence (a, b) of LFA1/2 (anti-CD11a) and 84H10 (anti-CD54). Photographs were taken after 4 hr of incubation.

Cell adhesion induced by HgCl₂ was further examined in quantitative ligand binding assays, in which the number of adhered cells was measured using DNA-binding fluorochromes. This method has the convenience and accuracy of a fluorometric assay, and excludes any possible interference of the detection method in the biological process, since the detecting agent is not present during the adhesion assay. As illustrated in Fig. 2, HgCl₂ enhanced the binding of ACC6 T cells to FN, ICAM-1, ICAM-2 and ICAM-3 dose-dependently, with a maximal effect at 0·6–1·25 µm. Cell binding to ICAM-1, ICAM-2 and ICAM-3 could be completely prevented by blocking antibodies directed against CD11a, irrespective of the stimulus used (not shown). Exposure of Jurkat T cells to HgCl₂ induced strong cell binding to FN, to a similar level as that induced by PMA or by MnCl₂, as shown in Fig. 3. Fibronectin adhesion of Jurkat T cells exposed to HgCl₂ was maximal at a concentration of 2·5 µm; this concentration did not induce any cytotoxicity as assessed by trypan blue exclusion and by microscopic evaluation of cell morphology. Notably, the effect of HgCl₂ was much stronger than the effect of MnCl₂, when compared at equimolar concentrations (2·5–5 µm; not shown). Analysis of the kinetics of HgCl₂-induced FN adhesion revealed maximal binding after 1 hr of stimulation, followed by a gradual decrease (not shown). In agreement with other investigators,³⁹ as well as with the lack of inducible cellular aggregation in Jurkat T cells, these cells did not bind to ICAM-2, neither under control conditions, nor after stimulation with HgCl₂, MnCl₂, or PMA.

Figure 2.

HgCl₂ induces binding of the Th0 clone ACC6 to FN, ICAM-1, ICAM-2 and ICAM-3. Adhesion of ACC6 T cells was assessed without stimulation (control) or in the presence of HgCl₂ (concentrations as indicated), PMA (50 ng/ml) or MnCl₂ (500 µm). One of two similar experiments is shown.

Figure 3.

HgCl₂ induces adhesion of Jurkat T cells to FN. Jurkat T-cell adhesion was analysed without stimulation (control) or in the presence of HgCl₂ (concentrations as indicated), PMA (50 ng/ml) or MnCl₂ (500 µm). Data are representative for four experiments.

Inhibition studies using blocking mAb were performed to examine which integrins are involved in HgCl₂-induced binding of Jurkat T cells to FN (Fig. 4). A mAb directed against the integrin β₁ subunit (CD29) completely blocked FN adhesion under control conditions, as well as after stimulation with HgCl₂, MnCl₂, or PMA. Monoclonal antibodies directed against integrin α₄ (CD49d) or α₅ (CD49e) subunits, when applied alone, did partially and variably inhibit FN adhesion. However, when they were applied in combination, FN adhesion was completely blocked, indicating that both integrin α₄β₁ (VLA-4) and α₅β₁ (VLA-5) were able to mediate cell adhesion to FN after stimulation with HgCl₂, MnCl₂, or PMA. Flow cytometric analysis revealed membrane expression of CD29, CD49d and CD49e on Jurkat T cells at a high level (not shown). Expression was not modified after incubation for 1 hr with HgCl₂, PMA, or MnCl₂, indicating that induction of cell adhesion by these stimuli is not caused by an increase in receptor number.

Figure 4.

HgCl₂-induced FN adhesion of Jurkat T cells is mediated via VLA-4 and VLA-5. Jurkat T cells were preincubated with control IgG1, 4B4 (anti-CD29), HP2/1 (anti-CD49d), SAM-1 (anti-CD49e), or HP2/1 plus SAM-1. Subsequently, FN adhesion was assessed in the presence or absence of HgCl₂ (2·5 µm), PMA (50 ng/ml), or MnCl₂ (500 µm), as indicated. In separate experiments, optimal blocking concentrations of the mAb were determined, in which SAM-1 was titrated in the presence of HP2/1, and vice versa. Data represent three similar experiments.

HgCl₂ and MnCl₂ induce cell adhesion via different pathways

Induction of Jurkat T-cell adhesion to FN by HgCl₂, MnCl₂, or PMA, was compared with respect to its dependence on extracellular cations and on a physiological temperature, in order to study the pathways used in the adhesion process.

When the adhesion assay was performed in a buffer without Ca²⁺ or Mg²⁺, Jurkat T-cell adhesion to FN was completely blocked, unless MnCl₂ was present (Fig. 5a). Addition of 1 mm CaCl₂ (Fig. 5a) or 1 mm MgCl₂ (Fig. 5b) restored the effects of HgCl₂ and PMA but did not significantly modify the effect of MnCl₂. The presence of MgCl₂ (1 mm) also enhanced adhesion of untreated cells (Fig. 5); when MgCl₂ was applied at higher concentrations (3 mm), adhesion of unstimulated cells was further increased, and the stimulatory effects of HgCl₂, PMA and MnCl₂ were undetectable (not shown). These experiments indicate that Mn²⁺, but not Hg²⁺, bypasses the requirement of extracellular Ca²⁺ or Mg²⁺ for integrin-mediated cell adhesion.

Figure 5.

HgCl₂-induced adhesion of Jurkat T cells to FN is dependent on the presence of extracellular Ca²⁺ or Mg²⁺. The experiment was performed as indicated in the Materials and Methods; CaCl₂ or MgCl₂ were present or absent as indicated, at a concentration of 1 mm. Data shown in (a) and (b) were obtained in a parallel experiment; they are representative for at least three experiments. Data obtained in the absence of CaCl₂ were confirmed by experiments in which the cells were preincubated with 1 mm EGTA, followed by extensive washing, which procedure yielded similar results (not shown).

Cell adhesion induced by HgCl₂ or by PMA was completely prevented when the assay was performed at 4° instead of at 37° whereas the effect of MnCl₂ was unaffected by incubation at low temperature (similar data were obtained in three independent experiments; not shown). Apparently, an active cellular process is required for adhesion induced by HgCl₂ or by PMA but, as expected, not for MnCl₂-induced adhesion.

Involvement of the actin cytoskeleton in cell adhesion

The compound cytochalasin B, which disrupts the intracellular actin cytoskeleton, was used to investigate the role of the cytoskeleton in FN adhesion of Jurkat T cells, induced by different stimuli. Pre-incubation with cytochalasin B induced a dose-dependent increase of adhesion of Jurkat T cells to FN (not shown). Furthermore, treatment with cytochalasin B had a synergistic effect when combined with MnCl₂, also when MnCl₂ was applied at a dose which was in itself not effective in induction of cell adhesion (Fig. 6). However, no effect of cytochalasin B was detected when it was applied in combination with PMA or HgCl₂ (Fig. 6).

Figure 6.

Modification of FN adhesion of Jurkat T cells by cytochalasin B. Cells were preincubated with cytochalasin B (5 µg/ml), followed by stimulation of the cells as indicated. Data represent two similar experiments.

Discussion

The present study identifies the immune dysregulatory compound HgCl₂ as a potent activator of integrin-mediated T-cell adhesion. This effect of HgCl₂ is not specific for a single integrin or ligand: HgCl₂ activates cell binding via LFA-1 to ICAM-1, ICAM-2 and ICAM-3, as well as cell binding via VLA-4 and VLA-5 to FN. Furthermore, adhesion-promoting effects of HgCl₂ could be demonstrated in T lymphoblasts derived from normal human PBMC, in a human T-cell clone and in Jurkat T cells, indicating that its effect is not a unique feature of a particular cell line.

Obviously, activation of integrin-mediated ligand binding by a metal ion is not surprising.^19–21 However, comparison of the induction of Jurkat T-cell adhesion by Hg²⁺ with the effect of Mn²⁺ revealed important differences: the effect of HgCl₂, but not the effect of MnCl₂, was dependent on the presence of extracellular Ca²⁺ or Mg²⁺, and on a physiological temperature. MnCl₂ supports binding of ligands to a variety of integrins. High-affinity binding sites for Mn²⁺ have been mapped in the extracellular domain of several integrins,^21,40–42 and on some of these sites, Ca²⁺ and Mg²⁺ competed for Mn²⁺ binding.^21,41,42 Mn²⁺ is able to induce ligand binding to isolated integrin molecules or integrin domains.^21,40,43 This metal ion stabilizes a high-affinity state of the integrin that resembles the ligand-bound conformation,¹⁸ as detected by mAb which bind to specific epitopes in this conformation.^19,44 In contrast to adhesion induced by Mn²⁺, induction of integrin-mediated ligand binding by PMA requires active cell metabolism and the presence of Ca²⁺ or Mg²⁺, and does in general not induce a high-affinity state.^18,20,22,25 Since the requirements for induction of lymphocyte adhesion by HgCl₂ resemble those of PMA but not those of MnCl₂, we propose that HgCl₂ acts via an intracellular signalling pathway. Observations from our own group and from others have clearly shown that HgCl₂ is able to induce cell signalling events in T lymphocytes. Exposure to HgCl₂ can change the redox balance towards an oxidative state by decreasing the intracellular glutathione level, which can lead to signalling events such as intracellular calcium mobilization and production of oxygen radicals.^8,45 Furthermore, HgCl₂ can induce tyrosine phosphorylation in T lymphocytes and activation of the tyrosine kinases p56^lck and p60^c-src46,47 Hg²⁺ has an extreme high affinity for protein thiol groups, and can induce crosslinking of cystein residues on proteins.⁴⁸ Crosslinking of T-cell membrane proteins by HgCl₂ was proposed to trigger tyrosine phosphorylation;⁴⁶ however, these studies were performed using 10- to 1000-fold higher concentrations of HgCl₂ than required to support T-cell adhesion. We tested the involvement of tyrosine kinases in HgCl₂-induced T-cell adhesion by using the tyrosine kinase-specific inhibitor herbimycin A. This compound inhibited the induction of cell adhesion by all three stimuli used in our study (not shown) and therefore did not provide further clues about the mechanistic difference in cell adhesion induced by the stimuli examined.

Integrin activation induced by inside-out signalling has been frequently shown to require an intact cytoskeleton.^20,22 However, depending on the cell type, disruption of the actin cytoskeleton by cytochalasins can enhance LFA-1-mediated binding to ICAM-1,^27,49 probably by induction of integrin clustering on the membrane.^49,50 Kucik et al. have shown that exposure to PMA strongly increases the lateral movement of LFA-1 in the cell membrane, which may very well play a major role in increasing the cellular capacity to bind ligand.²⁷ We demonstrate here that cytochalasin B increases integrin β₁-mediated adhesion of Jurkat T cells, thus indicating that cytoskeletal anchoring regulates integrin β₁-mediated adhesion in this cell line. Furthermore, the effect of this compound was synergistically increased by MnCl₂, but not by HgCl₂ or PMA. This finding supports the hypothesis that MnCl₂ increases the affinity of integrins for their ligands, thereby synergizing with the action of cytochalasin B, whereas PMA and HgCl₂ mainly act by modification of integrin association with cytoplasmic proteins, resulting in a similar functional effect as that induced by cytochalasin B.²⁷

A recent publication describes that HgCl₂ can induce expression of an activation-dependent epitope of integrin β₁, recognized by the mAb HUTS21, on rat lymphocytes.⁵¹ Expression of the HUTS21 epitope is correlated to integrin adhesive function, and can be induced by various stimuli, including MnCl₂ and a phorbol ester.⁵² These results are in line with the functional effects of HgCl₂ on integrin-mediated adhesion described in the present study. HUTS21 expression was also demonstrated in BN rats exposed to HgCl₂,⁵¹ but the kinetics of the response suggest that lymphocyte activation secondary to treatment with HgCl₂ is at least partially responsible for the effects observed. A functional relevance of expression of activated integrins in HgCl₂-induced autoimmunity is given by an in vivo experiment in which a treatment with the HUTS21 mAb prevented renal inflammation and reduced autoantibody production induced by HgCl₂.⁵¹

Our study provides clear evidence that HgCl₂ is able to activate T lymphocytes in the absence of antigen presentation. Whether integrin-mediated ligand binding directly induced by HgCl₂ plays a role in its immune dysregulatory abilities in vivo will obviously depend on the concentration of HgCl₂ in the microenvironment of lymphocytes. The peak concentrations of mercury in the spleens of mice exposed to HgCl₂ at a dose which induces autoimmunity were reported to vary between 646 and 955 ng per gram wet tissue,^53,54 which roughly corresponds to 3·2 and 4·8 µm HgCl₂. These data suggest that integrin activation by HgCl₂ can take place in vivo, in the induction phase of the disease.

Integrin-mediated ligand binding is of great importance for many aspects of the immune system, including costimulation of T lymphocytes during antigen presentation, B-cell activation and antibody production, extravasation and migration of leucocytes, cell-mediated cytotoxicity and regulation of cell survival. Interactions mediated via integrin α₄ and β₁ are required for the full development of HgCl₂-induced disease in vivo.^11,51 Therefore, we postulate that the mechanism described in the present study plays an important role in the induction and effector phases of the HgCl₂-induced autoimmune syndrome.

Abbreviations

FN

fibronectin

ICAM

intercellular adhesion molecule

LFA

lymphocyte function-associated antigen

mAb

monoclonal antibody

VLA

very late antigen