Exposure to environmental heavy metals has been reported to affect the immune system. Here, we tested the hypothesis that Hg+2, acting through membrane proteins, disrupts metabolic dynamics and downstream cell functions in human neutrophils. We found that HgCl2 inhibited: (1) polarization and (2) immunoglobulin (Ig)G-mediated phagocytosis of sheep erythrocytes in a dose-dependent manner from 2.5 to 10 µm. Because these activities have been linked with pro-inflammatory signalling, we also studied the effects of HgCl2 on intracellular signalling by measuring protein tyrosine phosphorylation. HgCl2 at doses = 1 µm increased tyrosine phosphorylation. We also studied the effect of HgCl2 on neutrophil metabolism by measuring NAD(P)H autofluorescence as an indicator of intracellular NAD(P)H concentration. After HgCl2 treatment, we found that normal sinusoidal NAD(P)H oscillations became incoherent. We recently reported that the NAD(P)H oscillation frequency is affected by cell migration and activation, which can in turn be regulated by integrin-mediated signalling. Therefore, we examined the effects of HgCl2 on cell surface distribution of membrane proteins. After exposure to environmentally relevant concentrations of HgCl2 we found that CR3, but not other membrane proteins (e.g. uPAR, FcγRIIA and the formyl peptide receptor), became clustered on cell surfaces. We suggest that HgCl2 disrupts integrin signalling/functional pathways in neutrophils.
Environmental heavy-metal ions have been found to affect the immune system. High and toxic concentrations of mercury have been shown to cause aggregation of various receptors such as CD3, CD4, CD45 and Thy-1  expressed on normal human leukocytes. This aggregation is likely owing to mercury forming an S-Hg-S linkage between free sulfydryl groups on adjacent receptors . At high concentrations, chemical cross-linking by mercury has also been shown to be a potent activator of Lck, a protein kinase usually activated through receptor ligation and cross-linking . Activation of receptor-associated kinases such as Lck would likely interfere with receptor-mediated signal transduction. However, the question remains whether low and nontoxic, but environmentally relevant concentrations of mercury interfere with signal transduction in leukocytes.
Previous findings suggest that heavy metals often do impact signalling pathways [3–5]. In particular, Rosenspire et al. showed that when WEHI-231 B-cell lymphoma cells were exposed to environmentally relevant levels of HgCl2, tyrosine phosphorylation of various intracellular proteins was increased in parallel with disruption of growth control, suggesting a connection between low-level mercury exposure and disruption of protein tyrosine kinase-mediated signal transduction in B cells [6,7]. Other experiments have also shown nontoxic levels of mercury to have detrimental effects on leukocytes, inhibiting functions such as chemotaxis, phagocytosis, and superoxide production [8–13]. These leukocyte functions all have in common the fact that they take place downstream from receptor ligation and signalling.
In general membrane receptor ligation and/or aggregation have been shown to alter cellular metabolism and downstream cellular activities including cell polarization, nitric oxide (NO) production, and oxygen radical production in leukocytes [see for example 14]. We have found that in migrating neutrophils the intracellular concentration of NAD(P)H normally oscillates in a regular sinusoidal pattern, but that the frequency and amplitude of the oscillation is regulated by external ligands binding to their receptors . Furthermore, in certain genetic abnormalities, alterations in the normal sinusoidal pattern of NAD(P)H oscillations are reflective of abnormalities in energy-dependent cellular activities [16, 17]. Therefore, if low concentrations of mercury initiate inappropriate receptor ligation, we might expect to see effects of mercury on leukocyte function that are correlated with alterations in metabolism, but in the absence of overt toxicity. Consequently, in this report we directly tested the hypothesis that mercury acts directly or indirectly to cause inappropriate neutrophil membrane receptor clustering, and in so doing disregulates cellular energy metabolism, and impairs cellular function.
Materials and methods
Preparation of cells Whole blood was obtained from healthy volunteers. Neutrophils were separated by density gradient centrifugation using Histopaque 1077 and 1119 (Sigma, St. Louis, MO, USA) as described . Mercuric chloride (HgCl2; Mallinckrodt, Paris, KY) was added to the cells in Hanks' balanced salt solution (HBSS, Gibco BRL, Grand Island, NY, USA) at various concentrations for 30 min at 22 °C. The cells were washed three times in HBSS then used immediately for experiments.
Cell viability Cells were treated with various concentrations of HgCl2 for 30 min at room temperature. The cells were washed and suspended in a 0.4% trypan blue solution. The cells were observed by brightfield microscopy then scored for viability based upon trypan blue exclusion.
Cell polarization Mercury-treated human neutrophils were placed onto glass coverslips and treated with 1 × 10−8M N-Formyl-Met-Leu-Phe (fMLP, Sigma) for 20 min at 37 °C. Cells were washed once to remove nonadherent cells then observed using an axiovert 135 inverted microscope (Carl Zeiss, Thornwood, NY, USA). The cells were scored for either a round morphology or a polarized morphology.
Phagocytosis Neutrophils were placed onto glass coverslips for 20 min at 37 °C. Opsonized sheep erythrocytes (SRBC) were prepared by adding the highest subagglutinating concentration of a rabbit antisheep red blood cell IgG (ICN, Costa Mesa, CA, USA) to sheep erythrocytes (Alsevers; Rockland, Inc., Gilbertsville, PA, USA). SRBCs were added at a 10 : 1 target : effector ratio for 20 min at 37 °C. Following this incubation, cells were placed on ice to block phagocytosis. A second-step rhodamine-labelled goat antirabbit IgG (ICN) was then added to label noninternalized erythrocytes as described . The cells were imaged by brightfield and fluorescence microscopy using an axiovert 135 inverted fluorescence microscope (Carl Zeiss, Thornwood, NY, USA) coupled to a CCD-72 (Dage-MTI, Michigan City, IN, USA) camera for brightfield imaging and a Hamamatsu (Bridgewater, NJ, USA) intensified CCD camera for fluorescence imaging. Images were acquired by an image capture board (Scion Corporation, Frederick, MD, USA), and processed using Scion Image software loaded onto a Dell Precision 410 Workstation (Round Rock, TX, USA). In this assay, noninternalized erythrocytes fluoresce, whereas internalized SRBCs, which are not accessible to the fluorescent tag, do not fluoresce.
SDS-PAGE Western Blot 1 × 107 cells were prepared as described above and then treated with either 1 µm, 5 µm or 10 µm mercuric chloride for 30 min at room temperature. The cells were washed with phosphate-buffered saline (PBS) then lysed in SDS buffer. After boiling for 3 min, cellular debris was removed by centrifugation for 30 s in a microcentrifuge (1200 × g). Whole cell lysates were resolved on 10% SDS-PAGE. The gel was electrophoretically transferred to PVDF membrane (Micron Separations Inc, West Borough, MA, USA) for 1.5 h at 13 V/cm. To detect tyrosine phosphorylation, the proteins were blotted with HRP conjugated antiphosphotyrosine reagent RC-20H (Signal Transduction Labs, Lexington, KY, USA). Blots were developed using enhanced luminol-based chemiluminescence (#0987–766, Pierce, Rockford, IL, USA) and Reflection film (DuPont, Wilmington, DE, USA).
Metabolic oscillations To characterize the coherence of metabolite flux, cellular NAD(P)H autofluorescence was measured as previously described . Briefly, the NAD(P)H autofluorescence was observed using excitation at 365 nm and emission detected at wavelengths = 405 nm using a Zeiss IM-35 fluorescence microscope. The light was detected with a Hamamatsu photomultiplier tube (Bridgewater, NJ, USA) held in a Products for Research (Danvers, MA, USA) housing coupled to the microscope. The signal was digitized and recorded using a MacLab system (AD Instruments, Castle Hill, Australia) coupled to a Mac 9600 computer system. Cells were observed for up to 120 min after mercury treatment and the patterns of NAD(P)H oscillations were recorded.
Receptor labelling Normal neutrophils were treated with HgCl2 as described above. The cells were then labelled on ice for 30 min with either fluorescent-conjugated (Fab) anti-FcγRII (clone IV.3, Medarex, Annandale, NJ, USA), fluorescent-conjugated (F(ab′)2) anticomplement receptor type 3 (CR3, clone 44), fluorescent-conjugated (F(ab′)2) anti-Urokinase-type plasminogen activator receptor (antiuPAR, clone 3B10) or fluorescein-conjugated N-Formyl-Nle-Leu-Phe-Nle-Tyr-Lys (fNLPNTL, Molecular Probes, Eugene, OR, USA) to detect the formyl-peptide receptor (FPR). The cells were washed by centrifugation 400 × gthen observed with an axiovert inverted fluorescence microscope. Fluorescein emission was observed using bandpass filters with excitation at 485 nm and emission at 530 nm. Tetramethylrhodamine was observed with excitation at 530 nm and emission at 590 nm.
We first studied the effect of mercury exposure on cell viability. Normal human neutrophils were exposed to 2.5–10 µm HgCl2 in HBSS for 30 min at room temperature. The cells were washed then stained with trypan blue to assess viability. As seen in Fig. 1, concentrations up to 10 µm did not dramatically lower cell viability (80 ± 6% versus 91 ± 6% for control cells). These results show that over the time-scale of our experiments, cell viability is reduced slightly. Therefore, the dramatic effects seen in subsequent experiments are not a manifestation of cell death.
To test the hypothesis that mercury inhibits neutrophil function, we examined the effect of various concentrations of mercuric chloride on spontaneous neutrophil polarization. We began by studying the effect of mercury in concentrations of 2.5 µm, 5 µm, 7.5 µm and 10 µm compared with control cells. The cells were incubated with buffer alone or mercuric chloride at 37 °C for 30 min to allow adherence, washed, and then observed by DIC microscopy. We found that under these conditions, nearly 53% of untreated neutrophils became polarized (Fig. 2). However, as the dose of mercury increased to 10 µm, the percentage of polarized neutrophils decreased in a dose-dependent manner from 53 ± 4% for control cells to 21 ± 2% at 10 µm HgCl2 (P < 0.005).
We next studied the kinetics of fMLP-stimulated cell polarization at various timepoints after mercury exposure. Cells were treated with 10 µm HgCl2 for 30 min at 22 °C, washed, and then plated on glass coverslips in the presence of 1 × 10−8M fMLP. After exposure to 10 µm mercuric chloride, neutrophil polarization was dramatically inhibited beginning at the earliest time point tested (10 min post exposure). The inhibitory effect became greater as the time postexposure was increased up to 80 min as shown in Fig. 3.
To test the effect of mercury on antibody-dependent phagocytosis, we exposed neutrophils to various concentrations of mercury and examined their ability to phagocytose particles. After treating cells with HgCl2 and washing by centrifugation as described above, we coincubated the neutrophils with opsonized erythrocytes for 20 min at 37 °C. The cells were then placed on ice to block further phagocytic activity. A second-step fluorescent anti-IgG antibody (Ab) was added to identify SRBCs that are bound but not internalized (Fig. 4). Therefore, when viewed by fluorescence microscopy, internal SRBCs, which are not accessible to the label, do not fluoresce whereas external targets do . We found that phagocytosis of opsonized sheep red blood cells was inhibited in a dose-dependent manner similar to polarization (Fig. 5).
Similarly, the binding of opsonized targets decreased upon mercury exposure from > 1 bound target per neutrophil in untreated cells to less than one target bound per every five cells treated with 10 µm HgCl2 (P < 0.001). These results show that mercury at physiologically relevant doses can diminish Fc receptor-mediated target binding and phagocytosis by neutrophils (Fig. 5).
Cell polarization and phagocytosis depend upon intracellular signalling events. Therefore, we next studied intracellular signalling via tyrosine phosphorylation of proteins. After 30 min of mercury treatment, neutrophils were lysed. After extraction, proteins were analyzed by SDS-PAGE. We then transferred the protein to a PVDF membrane and blotted with an antiphosphotyrosine monoclonal antibody (MoAb) as described in Materials and methods. After 30 min of treatment with concentrations = 1 µm mercuric chloride, we found distinct increases in cellular tyrosine phosphorylation of a protein of 37 kDa (Fig. 6). This change was consistent up to 10 µm HgCl2, showing that the concentration ranges used in these experiments all lead to a biochemical response.
We previously reported that in migrating neutrophils, cytosolic NAD(P)H concentrations oscillate in a regular sinusoidal pattern, implying that linked metabolites such as ATP also oscillate . On theoretical as well as experimental grounds, we have suggested that oscillations in metabolic fluxes are utilized as a signal transduction conduit connecting external receptors with energy-dependent neutrophil functions . Because cell polarization and phagocytosis are energy-dependent activities that are controlled by intracellular signalling pathways, we studied the effect of HgCl2 on cellular metabolic flux.
Cytosolic NAD(P)H concentrations can be conveniently monitored on a cellular level by measuring NAD(P)H autofluorescence . Therefore, after treatment of neutrophils with mercury, we measured NAD(P)H autofluorescence using quantitative fluorescence microscopy. We found that 30 min after exposure to HgCl2, the normal sinusoidal pattern of NAD(P)H autofluorescence became incoherent (Fig. 7). This loss of coherence was not associated with cell death, as we confirmed the viability of the cells after the same time period post exposure and found no significant decrease in viability as measured by trypan blue exclusion (e.g., 66% versus 68% after 30 min).
We have previously shown that abnormal clustering of CR3 is associated with incoherent NAD(P)H oscillations and defective cell polarization in neutrophils obtained from a pyoderma gangrenosum patient . This suggested a mechanism whereby mercury-dependent CR3 clustering might lead to dysregulated NAD(P)H metabolism. Therefore, we studied the effect of various doses of HgCl2 on the spatial distribution of CR3 on neutrophil membranes. We first treated neutrophils with 5 µm mercuric chloride for 30 min at room temperature. The cells were washed then placed on ice and then labelled with fluorescent-conjugated F(ab′) 2 or Fab fragments of anti-CR3, anti-FcγRIIA, antiuPAR or fluorescently-labelled fMLP for 30 min on ice. When viewed by fluorescence microscopy, neutrophils exposed only to buffer show a uniform distribution of all receptors (Fig. 8). However, upon exposure to HgCl2 we found that greater than 84 ± 6% of cells studied displayed CR3 clustering on the cell surface. Interestingly, other membrane proteins including FcγRIIA, uPAR and FPR did not appear to cluster (data not shown).
There have been numerous reports describing the effects of mercuric chloride on immune response mechanisms in several cell types [8–12]. Much of this work has been descriptive, lacking a mechanistic interpretation of the molecular events involved. These studies have variously reported that, with respect to neutrophils, mercuric chloride may inhibit events such as adherence, polarization, chemotaxis, superoxide production, and phagocytosis [8, 9, 12]. Studies using fish leukocytes have also shown that HgCl2 alters their activation, thus showing that this is not a phenomenon unique to humans . Many of these neutrophil activities have been linked to membrane proteins as a key mediator. We and others have previously shown that CR3 and its several interacting partner receptors participate in many of these activities [14, 17–19, 22]. Thus, we tested the hypothesis that mercury alters the cellular function by affecting CR3 clustering and, in parallel, metabolic flux.
On the basis of the data presented above, a speculative mechanism of HgCl2 toxicity to leukocytes can be proposed. HgCl2 causes cross-linking of plasma membrane components, which may lead directly or indirectly to CR3 clustering. Extensive CR3 clustering may trigger inappropriate tyrosine phosphorylation-mediated signalling events in neutrophils. Aberrant tyrosine phosphorylation and excessive CR3 clustering, in turn, may lead to incoherent metabolite flux, presumably owing to aberrant kinase-phosphatase cycling . The association of CR3 clustering, dysregulated tyrosine phosphorylation, and incoherent metabolic oscillations has been previously noted in a pyoderma gangrenosum patient . Normal sinusoidal metabolic oscillations accompany normal leukocyte function. Incoherent metabolic oscillations, of a genetic or acquired origin, may thereby lead to aberrant leukocyte functions. Although this work does not prove a mechanistic cause and effect, our speculative mechanism may be an important conceptual advance.
This observation is consistent with studies showing that CR3 co-operates with FcγRIIA and FcγRIIIB in mediating an opsonin-dependent phagocytosis [19, 22]. As shown in Fig. 4, we found that exposure to mercury not only inhibited the phagocytic activity of neutrophils but also their ability to bind opsonized targets. The finding that mercury causes aggregation of CR3 suggests that this aggregation may inhibit the activity of other receptors on the cell surface. Alternatively, mercury may directly affect FcγRIIA and/or FcγRIIIB on the neutrophil surface in such a way that inhibits their normal activities.
In several systems, integrin-mediated signalling has been functionally associated with integrin-cytoskeletal linkages, and integrin β subunits have been shown to associate with microfilaments through interactions that are mediated through a variety of linker proteins (reviewed in ). We have shown in this report that Hg+2 disrupts neutrophil polarization and phagocytosis, processes that depend upon microfilaments as well as the integrin CR3. However, there is no evidence that Hg+2 directly affects the structure of F-actin. In fact, once cells have polarized, although low concentrations of Hg+2 inhibits phagocytosis, we have not observed polarized cells exposed to low concentrations of Hg+2 to depolarize, as would be expected if Hg+2 induced the deaggregation of F-actin. It is possible though that Hg+2, aside from inhibiting CR3 mediated signals responsible for actin polymerization, also affects a structural linkage of CR3 to F-actin filaments.
MacDougal et al.  and Rossi et al.  have described changes in Ca+2 signalling after mercury exposure in various cell types. Mercury exposure has also been reported to cause changes in tyrosine kinase activity [1, 3, 5, 13, 24]. Our present work helps to link alterations in neutrophil activities after mercury exposure to intracellular signalling and suggests that CR3 could be at least one proximal mediator of the harmful effects of mercury on cell function.
In these experiments we have investigated effects of ionic mercury at concentrations between 1 and 10 µm on human neutrophils. Previous studies have shown that actual mercury levels found in body fluids and tissues are quite variable . However, for about 3% of the population with no know special risk factors, blood levels of mercury exceed 0.25 µm 4% of the population exceeds 0.125 µm mercury in urine, with the highest concentrations reported to be about 1 µm. Therefore, the levels of mercury that we employed are only slightly higher than what one might expect to find in the population at large. Additionally, in our experiments cells were only exposed to mercury for short periods of time between 15 and 30 min. However, because we have found that the effects of mercury exposure seem to be exacerbated by longer exposure times, it is likely that our in vitro findings can be extrapolated to the in vivo situation where neutrophils have been chronically exposed to mercury over a protracted time.
Thus our findings may be pertinent with respect to exposure to environmentally relevant levels of mercury, for at least part of the general population. Furthermore our in vitro findings support earlier studies where neutrophil function was assayed in industrial workers who were exposed to mercury. Although greater than 90% of the subjects had urinary mercury levels below the accepted threshold level of 50 micrograms g-1 creatin, it was still found that in neutrophils isolated from exposed individuals, both chemotactic activity and the respiratory burst was significantly impaired from that of neutrophils isolated from unexposed control subjects .
A previous study has shown that mercury can act as a stimulator of neutrophil activity . Although this may seem to contradict our results, the concentrations used by Jansson and Harms-Ringdahl  were much lower than those used in our studies. In fact, at lower concentrations it is plausible that only some membrane proteins become cross-linked. This lower level of cross-linking may lead to tyrosine phosphorylation and leukocyte activation. However, once the number of membrane proteins cross-linked becomes large, the activation signal generated may become a disregulated phosphorylation signal that could lead to aberrant leukocyte functions. This may cause the cell to become inhibited in many of its activities.
Mercury may be involved in the ligation and aggregation of various receptors. Excessive aggregation of receptors may lead to competition in recruiting kinases to transmit their cellular response, as signaling can be affected by the availability of kinases . The signaling machinery involved in mercury toxicity may be shared among various receptors. Work by Metzger's group has shown that kinases must shuttle between IgE-receptor aggregates to modify the cellular response . In our study it would be plausible that CR3 is cross-linked and thus becomes a major recruiter of kinases, which causes a lack of kinase availability to other receptors. Future studies of mercury toxicity involving CR3 should focus on assessing the availability of receptor-associated kinases during mercury exposure and toxicity. We have shown in this study that mercury may alter cell functions by disruption of the cells' metabolism and signaling abilities.
This work was supported by NIH grant AI27409 (to H.R.P.) and CA39064 (to R.F.T.), and by an Interdisciplinary Research Seed Fund from Wayne State University (A.J.R). R.G.W. was supported in part by a student award program grant from the Blue Cross and Blue Shield of Michigan Foundation.