HgCl2-induced Interleukin-4 Gene Expression in T Cells Involves a Protein Kinase C-dependent Calcium Influx through L-type Calcium Channels*

Mercuric chloride (HgCl2) induces T helper 2 (Th2) autoreactive anti-class II T cells in Brown Norway rats. These cells produce interleukin (IL)-4 and induce a B cell polyclonal activation that is responsible for autoimmune disease. In Brown Norway rats, HgCl2 triggers early IL-4 mRNA expression both in vivo and in vitro by T cells, which may explain why autoreactive anti-class II T cells acquire a Th2 phenotype. The aim of this study was to explore the transduction pathways by which this chemical operates. By using two murine T cell hybridomas that express IL-4 mRNA upon stimulation with HgCl2, we demonstrate that: 1) HgCl2 acts at the transcriptional level without requiring de novo protein synthesis; 2) HgCl2 induces a protein kinase C-dependent Ca2+ influx through L-type calcium channels; 3) calcium/calcineurin-dependent pathway and protein kinase C activation are both implicated in HgCl2-induced IL-4 gene expression; and 4) HgCl2 can activate directly protein kinase C, which might be one of the main intracellular target for HgCl2. These data are in agreement with an effect of HgCl2 which is independent of antigen-specific recognition. It may explain the T cell polyclonal activation in the mercury model and the expansion of pathogenic autoreactive anti-class II Th2 cells in this context.

HgCl 2 and gold salts induce in Brown Norway (BN) 1 rats and in susceptible mice a T helper 2 (Th2) cell-dependent B cell polyclonal activation responsible for an increase in serum IgE concentration and for the production of various autoantibodies (1,2). Anti-laminin autoantibodies are associated in BN rats with the development of a glomerulopathy (1) which resembles the one observed in some patients exposed to mercurials or gold salts. Autoreactive anti-class II T cell lines have been derived from diseased BN rats. These T cell lines produce interleukin (IL)-4 and may transfer autoimmunity in CD8 ϩ cell-depleted BN rats (3) by stimulating B cells polyclonally.
CD4 ϩ T cells are divided into at least two subsets, Th1 and Th2, which differ by their functions and the profile of cytokines they produce (4). Th1 cells produce IL-2 and interferon-␥ and are responsible for cell-mediated immune reactions; Th2 cells produce IL-4, IL-5, IL-6, IL-10, and IL-13 and are implicated mainly in B cell help for IgG1 and IgE production. In addition, each cell subset antagonizes the other.
It is well known that IL-4 is crucial for the differentiation of naive T cells into Th2 cells. Th2 cells, once activated, produce IL-4. However, the nature of the cell that initially produces IL-4 and allows the differentiation into Th2 cells is a matter of debate. Candidates include natural killer 1.1 ϩ T cells (5), mast cells, basophils, and eosinophils (6). It is also possible that IL-6, which is produced by antigen-presenting cells, initiates IL-4 production by naive CD4 ϩ T cells (7).
Our previous results (8) demonstrate that CD4 ϩ T cells from BN rats produce in vitro IL-4 when cultured in the presence of HgCl 2 . This suggests that some stimuli may induce an early production of IL-4 by CD4 ϩ T cells and that T cells themselves may condition the differentiation of Th2 cells. Interestingly, HgCl 2 also induces IL-4 gene expression in BN mast cells (9).
It is currently admitted that ligation of the extracellular domains of the TCR activates a cascade of protein tyrosine kinases including p56 lck , p59 fyn , and ZAP-70, which leads to phosphorylation and activation of the ␥ isoform of phospholipase C (10 -12). Inositol 1,4,5-trisphosphate and diacylglycerol are produced, resulting in release of stored intracellular Ca 2ϩ and protein kinase C (PKC) activation, respectively. Both IL-2 and IL-4 promoters bind NF-AT and AP-1 nuclear factors, which integrate Ca 2ϩ -and PKC-dependent signaling pathways in T cells (13)(14)(15). In fact, Ca 2ϩ -and PKC-dependent pathways have been explored mainly in the context of TCR-dependent IL-2 production, and much less is known about IL-4 production. It has been put forward that the variation of [Ca 2ϩ ] i required to induce IL-4 in Th2 cells is lower than the one required to induce IL-2 in Th1 cells (16,17). It has also been proposed that IL-4 production does not involve the classical protein tyrosine kinases associated with the TCR or phospholipase C activation (18,19).
The aim of this work was to understand the mechanisms of HgCl 2 -induced IL-4 mRNA expression by T cells. Using two murine T cell hybridomas that express IL-4 mRNA upon stimulation with HgCl 2 , we show that HgCl 2 induces a PKC-dependent calcium influx through L-type calcium channels and that the Ca 2ϩ -dependent pathway and PKC activation are both required for HgCl 2 -induced IL-4 gene expression. PKC might be therefore one of the main target of HgCl 2 in this cell model because chemical can activate PKC in a cell-free system.
Stimulation Assay-Four protocols were used.
In the first protocol, T cell hybridomas were cultured in RPMI plus 10% FCS in the absence or in the presence of HgCl 2 or ionomycin. HgCl 2 (10 Ϫ2 M; Sigma) was prepared as a stock solution in 0.9% NaCl, and ionomycin (1 M; Sigma) was initially dissolved in dimethyl sulfoxide at a concentration of 2 mM. Further dilutions were done in FCS-free medium.
In the third protocol, T cell hybridomas were preincubated with the intracellular calcium chelator BAPTA/AM (30 M; Calbiochem) or with vehicle for 1 h. Cells were washed twice and were then treated with HgCl 2 or medium for 4 h before mRNA extraction. BAPTA/AM (10 Ϫ2 M) was dissolved initially in dimethyl sulfoxide, and further dilutions were done in FCS-free culture medium.
In the fourth protocol, spleen cells from normal 8-week-old male BN rats (Center d'Elevage Janvier, Le Genest-Saint-Isle, France) were prepared and incubated or not with HgCl 2 (20 M) for 4 h before RNA extraction. The effects of cyclosporin A and Ro 31-8220 were tested as described in the second protocol. A putative cytotoxic effect of HgCl 2 and of the different inhibitors was tested by trypan blue exclusion; none of them was cytotoxic at the concentrations used; the viability was Ͼ 85% and did not differ from that in controls.
IL-4 mRNA Detection and IL-4 Enzyme-linked Immunosorbent Assay-RNA extraction was done after 4 h of culture by using the TRIzol procedure (Life Technologies, Inc.). Semiquantitative reverse transcriptase polymerase chain reaction (PCR) was performed as described already (26). Briefly, RNA was reverse transcribed to cDNA using poly(dT) as primer and Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.) in a final volume of 40 l (26). The following primers were used: ␤-actin sense, 5Ј-TGG AAT CCT GTG GCA TCC ATG AAA C-3Ј; ␤-actin antisense, 5Ј-TAA AAC GCA GCT CAG TAA CAG TCC G-3Ј; mouse IL-4 sense, 5Ј-AAC ACC ACA GAG AGT GAG CTC GTC T-3Ј; mouse IL-4 antisense, 5Ј-TGG ACT CAT TCA TGG TGC AGC TTA T-3Ј; rat IL-4 sense, 5Ј-TGA TGG GTC TCA GCC CCC ACC TTG C-3Ј; rat IL-4 antisense, 5Ј-CTT TCA GTG TTG TGA GCG TGG ACT C-3Ј. Primers were designed to amplify cDNA fragments representing mature 348-base pair mRNA transcripts for ␤-actin, 178 base pairs for mouse IL-4, and 378 base pairs for rat IL-4. cDNAs were amplified in a 50-l reaction volume containing a 0.6 mM concentration of each dNTP (dATP, dTTP, dGTP, and dCTP) (Pharmacia, Uppsala, Sweden), 1 g/ml of each oligonucleotide primer, 2.5 mM MgCl 2 , 1 unit of Taq-DNA polymerase (Boehringer Mannheim, Meylan, France), and 5 l of PCR buffer (10 ϫ) (Boehringer). Reactions were performed in a DNA thermal cycler (Perkin-Elmer) for 20 cycles (␤-actin), 30 cycles (mouse IL-4), or 35 cycles (rat IL-4): 45 s at 94°C, 45 s at 60°C, and 2 min 30 s at 72°C preceded by an initial denaturation step (1 min at 93°C). Each PCR was performed at least twice. In some experiments, serial dilutions of cDNA were amplified. Aliquots of the PCR products were analyzed by electrophoresis on a 2% agarose gel in Tris borate EDTA buffer plus ethidium bromide. Photographs of gels were numbered, and densitometric analysis of the bands was performed by using the Gel Analyst program (ICONIX, Greystone). Results are expressed in arbitrary units and represent the ratio of the intensity of the band for IL-4 to the intensity of the band for ␤-actin ϫ 10, except when mentioned otherwise.
IL-4 production was quantified by using two-site sandwich enzymelinked immunosorbent assay with paired monoclonal antibody purchased from Pharmingen (27). Briefly, 11B11 anti-IL-4 monoclonal antibody was used for capture. After three washes with phosphatebuffered saline (PBS) containing 0.1% Tween 20 (PBS-Tween), undiluted culture supernatants (100 l/well) were incubated overnight at 4°C. Plates were then washed three times and incubated with biotinylated BVD6-24G2 anti-IL-4 monoclonal antibody in PBS-Tween containing 1% bovine serum albumin (PBS-Tween-bovine serum albumin). After washing, the bound biotinylated monoclonal antibody was revealed by an additional 30-min incubation with alkaline phosphataseconjugated streptavidin (Jackson, Immunoresearch Laboratories, Avondale, PA) diluted 1/5,000 in PBS-Tween-bovine serum albumin. After washing, the plates were incubated with the substrate p-nitrophenyl phosphate disodium (Sigma) in diethanolamine buffer, pH 9.6. The reaction was stopped by adding 3 N NaOH, and the absorbance was read at 405 nm. Cytokines were quantified from a standard curve generated by using various concentrations of recombinant mouse IL-4 diluted in PBS containing 1% FCS and 0.1% phenol. The detection limit was 15 pg/ml.
PCR-based Nuclear Run-on Assay-1H11.3 cells were incubated with HgCl 2 (20 M) for 1 h 30 min or 3 h. Cells were harvested, and nuclei were prepared as described by Rolfe and Sewell (28). Nuclei were split into two aliquots of 100 l and incubated for 30 min at 30°C in 20% glycerol, 30 mM Tris-HCl, pH 8.0, 2.5 mM MgCl 2 150 mM KCl, 1 mM dithiothreitol, and 40 units of RNasin. A 0.5 mM concentration of each rNTP (rATP, rCTP, rGTP, and rUTP) was added to one aliquot. No rNTPs were added to the second aliquot. After 30 min, nuclei were lysed. RNA extraction, reverse transcription, and PCR were performed as described above. Results were expressed by using the IL-4:␤-actin ratio as described above.
Protein Kinase C (PKC) Assay-Purified rat brain PKC (enriched in ␣, ␤, and ␥ isoforms; 5 ng/well; Calbiochem) was incubated or not with HgCl 2 (5, 10, 25, or 100 M) in the presence of EGTA (20 mM). We also tested the effect of Ca 2ϩ (10 or 100 M) in the presence or in the absence of EGTA. The capacity of PKC to phosphorylate a specific substrate was assessed by using a specific PKC kit assay (Calbiochem). This assay was performed in the presence of phosphatidylserine and ATP but in the absence of diacylglycerol.
Analysis of Intracellular Ca 2ϩ Concentration-Measurement of [Ca 2ϩ ] i was performed by emission microspectrofluorometry as described previously (29). Cells were incubated with 5 M fluo3/AM (Molecular Probes) for 30 min at 37°C. [Ca 2ϩ ] i was measured in T cells stimulated by HgCl 2 (15 or 20 M), ionomycin (1 M), phorbol 12myristate 13-acetate (PMA; 10 ng/ml; Sigma), and S(Ϫ)-Bay K8644 as an agonist of the L-type calcium channel (6 M). Ionomycin, PMA, and S(Ϫ)-Bay K8644 were dissolved initially in dimethyl sulfoxide at concentrations of 2 mM, 0.2 mg/ml, and 5 mM, respectively. We also checked the effect of Ro 31-8220. In some experiments, cells were stimulated by HgCl 2 in Hanks' balanced saline solution without calcium and magnesium and without phenol red (Life Technologies, Inc.) supplemented with EGTA (20 M). Cell preparation was then placed on the stage of an inverted microscope (Diaphot, Nikon) and observed with an objective (ϫ 40). Excitation light was 490 nm with a 525 nm barrier filter. Fluorescence was detected by a CCD camera intensified (Hamamatsu C2400-80). With the magnification used (ϫ 40) a field of 200 ϫ 200 m was recorded by the camera. Three to five fields were observed for each type of experiment, and in each field 12 windows (9 m) were distributed on different hybridoma cells and analyzed for fluorescence. Images were captured at intervals of 10 s and processed with the Argus 50 processing image system (Hamamatsu Photonics, Hamamatsu, Japan). Time courses of Ca 2ϩ signals in cells were analyzed with the Argus 50 software. Data are presented as the ratio of fluorescence (F) in stimulated cells to fluorescence (F 0 ) at the resting level. Cells were scored as positive if the fluorescence intensity variation was 5% above the resting level.

Effect of HgCl 2 on the Induction of the IL-4 Gene in Mouse T
Cell Hybridomas-Two T cell hybridomas (SM1.27.9 and 1H11.3) out of the five tested were selected because they expressed IL-4 mRNA upon stimulation with nontoxic amounts of HgCl 2 . The experiments herein reported have been performed on both hybridomas. Results were similar whatever the hybridoma tested. HgCl 2 induced IL-4 mRNA expression in a dosedependent manner in both hybridomas ( Fig. 1A and not shown). The effect was optimal when 1H11.3 T cells and SM1.27.9 T cells were incubated with 20 and 15 M HgCl 2 , respectively. In these conditions and in six independent experiments the IL-4:␤-actin ratio was 9.2 Ϯ 4.5 in SM1.27.9 T cells stimulated with HgCl 2 versus 1.2 Ϯ 2 in unstimulated cells and was 14.2 Ϯ 8.7 in 1H.11.3 T cells stimulated with HgCl 2 versus 1.8 Ϯ 1.5 in unstimulated cells. A semiquantitative assay in which serial dilutions of cDNA were performed confirms that HgCl 2 induces IL-4 mRNA expression as ionomycin does ( Fig.  1B and not shown). As soon as 2 h after stimulation with HgCl 2 , IL-4 mRNA was observed, with a peak at 4 h and a decline at 6 h (not shown). Actinomycin D, an inhibitor of transcription, abolished HgCl 2 -induced IL-4 gene expression in SM1.27.9 (Fig. 1C) and 1H11.3 (not shown) T cell hybridomas, whereas cycloheximide, an inhibitor of protein synthesis, had no effect (Fig. 1D and not shown).
To demonstrate that HgCl 2 actually induced IL-4 gene tran-scription, a PCR-based run-on assay was performed because it has indeed been shown previously that a classical run-on assay may be not sensitive enough to detect cytokine gene transcription (28). As shown in Fig. 2, no expression of IL-4 mRNA was observed after a 1-h 30-min stimulation with HgCl 2 in the absence of rNTPs, whereas the addition of rNTPs to the isolated nuclei allowed detection of IL-4 messenger. In contrast, 3 h after stimulation with HgCl 2 , IL-4 mRNA was detected in nuclei whether rNTPs were added or not. Altogether, these results show that HgCl 2 induces IL-4 gene and that this effect does not require de novo protein synthesis. IL-4 was not detected by enzyme-linked immunosorbent assay when 5 ϫ 10 5 cells/ml were cultured in the presence of HgCl 2 because the IL-4 assay is not sensitive enough. Indeed, when the cell density was increased (5 ϫ 10 6 /ml), although mortality was high (around 40% in both stimulated and unstimulated cultures), IL-4 was detected (144 Ϯ 41 pg/ml, n ϭ 4 in HgCl 2 -stimulated cells versus Ͻ15 pg/ml in control cultures). This shows that HgCl 2 induces not only IL-4 mRNA expression but also IL-4 production.
HgCl 2 Induces a PKC-dependent Influx of Ca 2ϩ in T Cell Hybridomas-It is well known that the Ca 2ϩ -dependent pathway is important for IL-4 gene induction. Therefore, we first checked whether HgCl 2 was able to increase [Ca 2ϩ ] i in 1H11.3 T cell hybridoma. As shown by microspectrofluorometry, in 58 out of 60 cells HgCl 2 induced a transient increase in fluores- cence which returned to resting level within 15 min (Fig. 3A), showing that this increase was not the result of a toxic effect of HgCl 2 . This increase was no longer observed when a Ca 2ϩ -free medium supplemented with EGTA (20 M) was used (Fig. 3B), demonstrating that the fluorescence increase was dependent on Ca 2ϩ entry from the extracellular medium (Figs. 3B and 4). Interestingly, the Ro 31-8220 PKC inhibitor (24,25) consistently and markedly suppressed this HgCl 2 -(Figs. 3C and 4) but not the ionomycin-induced increase in [Ca 2ϩ ] i (not shown). To demonstrate whether direct PKC activation may be responsible for a Ca 2ϩ influx, the effect of PMA, a well known PKC activator, was tested; PMA (10 ng/ml) actually increased [Ca 2ϩ ] i (Fig. 4) in 44 out of 48 cells. L-type calcium channels contain PKC consensus sites (30). To see whether 1H11.3 T cells express L-type calcium channels, we have tested the effect of an agonist of these channels. The S(Ϫ)-Bay K8344 agonist induced an increase in [Ca 2ϩ ] i in 49 out of 69 1H11.3 T cells (Fig. 4).

HgCl 2 -induced IL-4 Gene Expression Requires Both PKC Activation and an Increase in Intracellular Ca 2ϩ Concentration-
Because HgCl 2 induced a PKC-dependent Ca 2ϩ influx, we tested the effects of an inhibitor of PKC and of an intracellular Ca 2ϩ chelator on HgCl 2 -induced IL-4 gene expression. The specific PKC inhibitor (Ro 31-8220) suppressed, in a dose-dependent manner, HgCl 2 -induced IL-4 gene expression (Fig.  5A). BAPTA/AM, a chelator of intracellular Ca 2ϩ , abolished the effect of HgCl 2 on IL-4 gene expression (Fig. 5B). Cyclosporin A, which inhibits calcineurin phosphatase (31), also inhibited HgCl 2 -induced IL-4 gene transcription (Fig. 5C), supporting a role for the calcium/calmodulin/calcineurin-dependent pathway. Because there is some evidence that 1H11.3 T cells express L-type calcium channels we tested the effect of an L-type calcium channel blocker. R(ϩ)-Bay K8344 abolished HgCl 2induced IL-4 gene induction (Fig. 5D), suggesting that HgCl 2 induces a Ca 2ϩ influx through L-type calcium channels which leads to IL-4 gene transcription.
It has been shown recently that an increase in extracellular Ca 2ϩ concentration amplified calcium-dependent pathways including NF-AT nuclear translocation (32). In our system, increasing medium Ca 2ϩ concentration from 1 to 10 mM resulted in a marked increase in IL-4 gene expression (Fig. 6A) and IL-4 production (Fig. 6B). These results reinforce the role of a calcium influx in HgCl 2 -mediated IL-4 induction.
HgCl 2 Activates PKC-To assess whether PKC activation implies phospholipase C-mediated pathway, we used U-73122 as an inhibitor of phospholipase C. By itself this agent increased IL-4 gene expression in a 1H11.3 T cell hybridoma (Fig.  7A). Moreover, not only it did not decrease but it enhanced the effect of HgCl 2 on IL-4 gene expression (Fig. 7A). Similar results were observed with herbimycin A, a protein tyrosine kinase inhibitor (Fig. 7B).
Because PKC activation plays a major role in HgCl 2 -induced IL-4 gene expression and because upstream activation pathways for PKC activation did not seem to be involved, the question was addressed as to whether HgCl 2 could activate PKC directly. As shown in Fig. 8, HgCl 2 activated rat brain PKC in a dose-dependent manner even in the virtual absence of Ca 2ϩ , as the experiments were done in the presence of 20 mM EGTA, a concentration that abolished Ca 2ϩ -dependent PKC activation.

HgCl 2 Triggers IL-4 mRNA Expression in a Ca 2ϩ -and PKCdependent Manner in BN Spleen
Cells-To demonstrate that the Ca 2ϩ -and PKC-dependent pathways were also implicated in the effect of HgCl 2 on IL-4 expression in BN T cells, we tested the effects of cyclosporin A and of the inhibitor of PKC (Ro 31-8220). As shown in Fig. 9, HgCl 2 induced IL-4 mRNA The presence of IL-4 was assessed by enzyme-linked immunosorbent assay. One experiment among four is presented. expression in BN spleen cells, and this effect was abolished both by cyclosporin A and by the PKC inhibitor. DISCUSSION Previous results indicated that the ability of HgCl 2 to induce IL-4 gene expression and IL-4 production by rat T cells directly (8) was of major importance in an understanding of why BN rats develop Th2-mediated autoimmunity after HgCl 2 exposure. The mechanisms of HgCl 2 -induced IL-4 gene expression were explored by using two mouse T cell hybridomas that responded to HgCl 2 . Using these T cell hybridomas, we showed that 1) HgCl 2 induced IL-4 gene transcription without de novo protein synthesis; 2) HgCl 2 induced a PKC-dependent Ca 2ϩ influx through L-type calcium channels; 3) PKC activation and Ca 2ϩ -dependent pathways were both required for HgCl 2 -in-duced IL-4 gene expression; and that 4) HgCl 2 was able to activate PKC directly in a cell-free system.
Others have shown that HgCl 2 induces expression of the mercuric ion reductase gene (33) and metallothionein gene (34) by interacting directly with DNA or transcription factors, respectively. The fact that a chelator of intracellular calcium and inhibitors of PKC both abolished IL-4 gene expression ruled out a role for an interaction of HgCl 2 with DNA or transcription factors. This indicates that HgCl 2 , which has the ability to enter the cell (35), has different intracellular targets.
PKC and Ca 2ϩ -dependent pathways were implicated in HgCl 2 -induced IL-4 gene transcription since an inhibitor of PKC (Ro 31-8220), the chelator of intracellular calcium BAPTA/AM and cyclosporin A abolished the induction of IL-4 gene by HgCl 2 . H-7, another inhibitor of PKC which is different chemically from Ro 31-8220, also abolished HgCl 2 -induced IL-4 gene expression (not shown), confirming that PKC was implicated. An elevation in the extracellular calcium concentration, which increases calcium-dependent NF-AT translocation in the nucleus (32), enhanced the effect of HgCl 2 on IL-4 mRNA expression and increased the production of IL-4, thus confirming the importance of the Ca 2ϩ -dependent pathway.
Our results are in agreement with numerous reports that demonstrate the involvement of both PKC and calcium-dependent pathways in IL-4 production upon stimulation through TCR or after ionomycin plus PMA treatment (15,36). However, whether Ca 2ϩ influx is inositol 1,4,5-trisphosphate-dependent or not is a matter of debate. On the one hand, it is generally accepted that T cell activation through TCR induces inositol 1,4,5-trisphosphate-dependent mobilization of calcium stores and secondarily an influx of Ca 2ϩ from the external medium (for review, see Ref. 37). On the other hand, depletion of extracellular Ca 2ϩ inhibited both the initial and the sustained Ca 2ϩ elevation induced by TCR-mediated stimulation in some T cells, which rules out a role for an initial Ca 2ϩ mobilization from the stores (38). Gajewski et al. (17) also reported that signaling in IL-4-producing Th2 clones was associated with weak variations in [Ca 2ϩ ] i in the absence of inositol 1,4,5trisphosphate production. The fact that HgCl 2 induced a cal- cium influx in the absence of an initial mobilization of Ca 2ϩ stores could be related to the fact that the T cell hybridomas used in this study resemble Th2 cells. Indeed, they produced IL-4 but no interferon-␥ upon stimulation via TCR (not shown). Alternatively, HgCl 2 is able to induce Ca 2ϩ entry in brain and renal cells without mobilization of Ca 2ϩ stores (39), and it might have the same effect on T cells.
We then investigated the relationship that could exist between PKC and the Ca 2ϩ -dependent pathway. Ro 31-8220, a PKC inhibitor, suppressed HgCl 2 -induced Ca 2ϩ entry, but it had no effect on an ionomycin-induced [Ca 2ϩ ] i increase (not shown). In addition, PMA, an activator of PKC, also triggered an entry of calcium in 1H11.3 T cell hybridoma. Dihydropyridine-sensitive L-type Ca 2ϩ channels, known to be a target for PKC (30), have already been described in T cells (40,41). (SϪ)-Bay K8344, an agonist of L-type Ca 2ϩ channels, induced an entry of Ca 2ϩ , indicating that these channels were expressed by 1H11.3 T cell hybridoma. These channels are likely to be implicated in HgCl 2 -induced IL-4 gene expression because an L-type Ca 2ϩ channel blocker, R(ϩ)-Bay K8344, abolished the induction of IL-4 gene by HgCl 2 . Thus, our experiments are in agreement with a pathway in which PKC activation is responsible for an influx of calcium through L-type Ca 2ϩ channels. An ability of PKC to activate L-type or other Ca 2ϩ channels has not been described in T cells to the best of our knowledge, but it is widely accepted in other cell types (29,42,43).
To answer the question of how HgCl 2 may activate PKC, we used in a cell-free system rat brain PKC that contains predominantly Ca 2ϩ -dependent ␣, ␤, and ␥ isoforms (44). In this system we show that Hg was as efficient as Ca 2ϩ in activating rat brain PKC. That relatively high concentrations of Hg 2ϩ and Ca 2ϩ were required for PKC activation is probably because the test was performed in the absence of diacylglycerol, which dramatically increases the affinity of the enzyme for Ca 2ϩ (45). Interestingly, lead also activates rat brain PKC (46) and favors IL-4 production in vivo and in vitro (47). Which PKC isoform(s) is(are) activated by HgCl 2 as well as the site of interaction of HgCl 2 with PKC are under investigation. Other authors have reported that HgCl 2 (50 -500 M) activates p56 lck among other src protein tyrosine kinases in mouse T cells and fibroblasts (48,49). However, the concentrations of HgCl 2 used in this study were in a toxic range, the cell types used were different, and activation of p56 lck was not correlated with any biological effect. Another effect of protein tyrosine kinase and phospholipase C inhibitors was that they enhanced IL-4 mRNA expression by themselves. It has been reported previously that inhibition of phospholipase C␥ or inhibition of protein tyrosine kinases does not induce IL-4 production (18,19), but the effect on IL-4 mRNA has not been studied. In addition, inhibition of protein tyrosine kinase and phospholipase C␥ enhanced the effect of HgCl 2 on IL-4 gene expression, suggesting that this pathway may have an inhibitory effect on IL-4 gene transcription. Ca 2ϩ -and PKC-dependent pathways were also implicated in HgCl 2 -induced IL-4 gene expression in BN rat T cells (Fig.  9). Whether direct PKC activation is also responsible for Ca 2ϩ influx in these cells remains to be determined. If confirmed, the effect of HgCl 2 on PKC activation might explain not only IL-4 production but also the T cell polyclonal activation described recently in BN rats treated with HgCl 2 (50).
To conclude, we propose that HgCl 2 activates PKC, which is responsible for an influx of Ca 2ϩ through L-type channels; Ca 2ϩ -and PKC-dependent pathways amplify each other, leading to IL-4 gene induction in cells that are engaged in IL-4 production. These findings may be of major importance in understanding how T cells differentiate into a Th2 subtype in the mercury model. They may also be relevant in other situations such as Leishmania major infection in BALB/c mice in which CD4 ϩ T cells are thought to be the initial source of IL-4 (51).