Modulation of B-cell endoplasmic reticulum calcium homeostasis by Epstein-Barr virus Latent Membrane Protein-1

BL-30 and BL-41 belong to a set of EBV-negative Burkitt's lymphoma cell lines that have been previously established [37], and subsequently infected with immortalizing, as well as non-immortalizing EBV strains (B95–8 and P3HR-1, respectively) in vitro. Infection led to the establishment of latently infected cell lines [38]. Whereas B95–8-infected cells express the whole set of type III latency genes including EBNA2 and LMP-1, P3HR-1-infected cells lack LMP-1 expression due to a deletion in the sequence of EBNA2 that transactivates LMP-1 [38]. In order to investigate the effect of virus on calcium pump expression, equal amounts of total protein extracts from non-infected and infected cells were analysed for SERCA2 and SERCA3 expression in a Western immunoblot format with the IID8 and PLIM430 monoclonal antibodies, respectively. As shown in Fig. 1, infection with the B95–8 virus strain led to decreased SERCA3 expression in both cell lines (Panel A), and similar results were obtained also with BL-2 and BL-31 cells (not shown). As shown in Panels B and D, SERCA3 expression levels in BL-30 and BL-41 cells decreased to 40 +/- 3% (n = 10), and 51 +/- 5% (n = 8) after infection with the B95–8 virus when compared to non-infected cells, whereas SERCA2 expression increased to 223 +/- 35% (n = 6) and 150 +/- 31% (n = 5), respectively. At the same time, infection with the P3HR-1 strain had only marginal effects on the expression of SERCA proteins. Importantly, LMP-1 expression could only be observed in cells stably infected with the B95–8 virus. This is in accordance with earlier observations indicating that LMP-1 expression is absent in P3HR-1-infected cells [38] as a result of lack of EBNA-2.

Given the association between decreased SERCA3 expression and LMP-1 expression in latently infected cells (Fig. 1), the effect of the direct expression of LMP-1 on SERCA3 expression was investigated using cells stably transfected with inducible vectors coding for LMP-1. The BJAB-tTA-LMP-1 cell line carries an LMP-1 transgene under the control of a tetracycline-repressible promoter. When the cells are grown in the presence of tetracycline, expression of LMP-1 is repressed, and tetracycline withdrawal results in LMP-1 expression at levels similar to that observed in EBV-infected cells [39]. As shown in Fig. 2, tetracycline withdrawal in the BJAB-tTA-LMP-1 cell line resulted in a marked and homogeneous induction of LMP-1 protein as detected by Western blotting (Panel A) and immunocytochemistry (Panel C). Induction of LMP-1 expression led to decreased SERCA3 levels (43 +/- 4.5%, n = 5) during five days (Panels A and B), whereas the expression of SERCA2 was not significantly affected. As a negative control, BJAB-tTA cells, devoid of the LMP-1 transgene were used, and in these cells tetracycline withdrawal had no effect on SERCA expression (not shown). Induction of LMP-1 expression led to decreased SERCA3 expression also in the BL41-MTLM-2 cell line, in which the expression of the LMP-1 cDNA is driven by a heavy metal-inducible promoter derived from the metallothionein gene [40]. When cells were grown in the presence of 100 μM Zn²⁺ for 3 days, SERCA3 expression was decreased when compared to non-induced controls, similarly to that observed in BJAB-tTA-LMP-1 cells upon tetracycline withdrawal (not shown). It is noteworthy to add, that SERCA3 expression was also investigated in BL41/K3 cells in which the expression of an EBNA2 transgene is driven by an estradiol-inducible promoter [41]. However, induction of EBNA2 expression in these cells by estradiol for three days as in [41] did not significantly modify SERCA3 expression, indicating that LMP-1 alone is sufficient for the effect.

In order to establish the effects of LMP-1 expression on B cell calcium homeostasis, resting cytosolic calcium concentration and endoplasmic reticulum calcium storage capacity were measured by Indo-I spectrofluorimetry in the BJAB-tTA-LMP-1 cell line before and after induction of LMP-1 expression. In addition, similar experiments were conducted on non-infected BL-30 and BL-41 Burkitt's-lymphoma cell lines in parallel with BL-30-B95–8 and BL-41-B95–8 cells infected with the immortalizing B95.8 EBV strain that expresses LMP-1.

As shown in Fig. 3, induction of LMP-1 expression in the BJAB-tTA-LMP-1 cells led to increased resting cytosolic calcium concentrations (Panel A). Indeed, whereas the resting cytosolic calcium concentration was 95 +/- 2 nM (n = 5) in non-induced cells, LMP-1 induction led to a gradual increase to 165 +/- 1 nM (n = 5) during a five days period. LMP-1 induction was also accompanied by increased calcium release from the ER upon full inhibition of SERCA-dependent transport by 1 μM thapsigargin. As shown in Panel B, when calcium release from the ER was induced by 1 μM thapsigargin in conditions where Ca²⁺ is trapped in the cytosol due to the inhibition of plasma membrane calcium pumps and Na⁺/Ca²⁺ exchange by 0.4 μM Hg²⁺ and replacement of extracellular Na⁺ by Li⁺, respectively, LMP-1 expressing cells displayed significantly enhanced calcium release from the ER, as reflected by a higher plateau of cytosolic Ca²⁺ levels (157 +/- 7 nM, n = 5), in LMP-1 expressing cells (Panel B), when compared to non-induced cells (129 +/- 7 nM, n = 5). When expressed as area under curve (AUC) values, LMP-1 expression under these conditions led to an approximately 24% increased calcium release when compared to non-induced cells (Panel D). Increased calcium mobilization from the ER by thapsigargin could also be seen in LMP-1-expressing cells in the presence of extracellular Na⁺ and in the absence of PMCA inhibition (Panel C), indicating that the observed LMP-1-dependent changes of cellular calcium homeostasis are also present in the presence of active plasma membrane calcium extrusion. The same experiments conducted on control BJAB-tTA cells treated identically to BJAB-tTA-LMP-1 cells, but devoid of LMP-1 expression disclosed no changes in cellular calcium homeostasis upon treatments (not shown). On the other hand, increased calcium storage in thapsigargin-sensitive intracellular calcium pools could also be observed, very similarly to LMP-1 induction, in cells infected with fully competent virus that expresses LMP-1 protein. Indeed, as shown in Fig. 4, when calcium release from the endoplasmic reticulum was induced by thapsigargin in non-infected BL-30 and BL-41 cells (black curves), and in cells infected with wild-type EBV (grey curves), an enhanced calcium release response could be observed in both cell lines in the presence of the virus. When the experiments were performed in the absence of inhibition of plasma membrane calcium pump and Na⁺/Ca²⁺ exchanger activity, this could be observed as an increased calcium peak following thapsigargin addition in BL-30 (non-infected versus infected: 35 +/- 5 nM and 61 +/- 10 nM, n = 3, p > 0.05), as well as BL-41 cells (non-infected versus infected: 68 +/- 5 nM and 82 +/- 5 nM, n = 4, p = 0.056). When the same experiments were performed in the presence of inhibition of PMCA and Na⁺/Ca²⁺ exchanger function, calcium release from the endoplasmic reticulum occurred as a sustained plateau and was markedly enhanced in infected cells. In fact, whereas in non-infected BL-30 and BL-41 cells calcium levels raised to 106 +/- 2 nM (n = 4) and 139 +/- 5 nM (n = 3), respectively, in the corresponding EBV-infected cells calcium levels attained 234 +/- 17 nM (n = 4) and 204 +/- 6 nM (n = 3), respectively, corresponding to a highly significant (p < 0.01) increase in infected cells. When this calcium release was expressed as area under curve (AUC) values of the plateau phase of the corresponding calcium fluorimetry curves of non-infected and B95–8-infected BL-30 and BL-41 cells, a highly significant increase of 110% and of 46% could be observed, respectively (Panels E and F).

EBV infection of B cells leads to the emergence of immortalized cell lines that actively proliferate and display a lymphoblastoid phenotype. In order to investigate, whether SERCA3 expression is modulated, similarly to EBV infection, also during normal B lymphocyte activation in a physiological tissue environment in situ, immunohistochemical staining for the protein was performed in normal lymph node and colic mucosa-associated lymphoid follicles. In these tissues quiescent, antigen-naïve non-stimulated small B lymphocytes reside in the mantle zone, whereas antigen-activated large centroblasts and centrocytes are located in germinal centers. Such a differential localization of resting and antigen-activated B cell populations in these tissues permits their identification on a histological basis. As shown in Fig. 5, Panel A, SERCA3 expression in a lymph node follicle displays marked differences: whereas the mantle zone that surrounds germinal centers is strongly labelled, staining in the germinal center is markedly weaker. As shown at higher magnification on Panel B, whereas small lymphocytes residing in the mantle zone are strongly labelled, SERCA3 expression in the germinal center activated lymphoblasts is markedly decreased.

In addition to a strong SERCA3 staining in colon epithelium as reported earlier [11, 14], a staining pattern very similar to that observed in lymph node follicles could be seen also in mucosa-associated lymphoid tissue in the colon (Panel C), with a marked decrease of SERCA3 expression in germinal centers when compared to small resting B cells located in a more luminal position.

These data taken together indicate that antigen-driven activation of small resting B lymphocytes located in the mantle zone into germinal center lymphoblasts is accompanied by a marked and homogeneous decrease of SERCA3 expression.

Cells were grown at 37°C in a humidified cell culture incubator in an atmosphere of air containing 5% CO₂ in RPMI-1640 medium supplemented with 2 mM glutamine, 2 mM alanyl-L-glutamine and 10% heat-inactivated fetal calf serum (complete RPMI medium). Cell culture reagents were from Lonza, Verviers, Belgium.

BJAB-tTA-LMP-1 cells were grown in complete RPMI medium supplemented with 2 mg/ml G418 and 0.5 mg/ml hygromycin B (both purchased from Sigma-Aldrich France) and 1 μg/ml tetracycline (Fluka, Germany) as described earlier [39]. Exponentially growing cells cultured in the presence of 1 μg/ml tetracycline were washed as follows: after centrifugation the cell pellet was resuspended in 10 ml complete medium containing 10% fetal calf serum without tetracycline, transferred into a new 50 ml tube, supplemented with 35 ml serum-free RPMI-1640 medium and centrifuged again. This washing step was repeated three times. Thereafter cells were resuspended in complete RPMI culture medium without tetracycline at an initial density of 2 × 10⁵ cells/ml. During the last one or two passages preceding induction of LMP-1 expression and during induction by tetracycline withdrawal, selection antibiotics were omitted.

BL41-MTLM-2 cells [40] grown in complete RPMI medium in the exponential phase of growth were centrifuged and resuspended at an initial density of 3 × 10⁵ cells/ml in complete RPMI medium. After one hour pre-incubation at 37°C in a cell culture incubator, 100 μM ZnCl₂ (Sigma-Aldrich France, Saint-Quentin Fallavier) was added from a 100-fold concentrated sterile-filtered stock solution. EBNA2 expression was induced in BL41/K3 cells that carry an estradiol-inducible EBNA2 construct as described in [41].

This was done essentially as described in [11]. Briefly, after treatments cells were harvested by centrifugation, resuspended in 1 ml ice cold NaCl (150 mM) and transferred to round-bottom 2 ml Eppendorf tubes, centrifuged and washed again by centrifugation with 1 ml ice cold NaCl solution. Cell pellet was thereafter resuspended in ice cold 5% trichloroacetic acid and kept at 4°C overnight. The protein precipitate was centrifuged at 12000 × g for 10 minutes at 4°C and supernatant was removed. The protein pellet was thereafter dissolved in lysis buffer at 30 mg TCA-precipitated protein pellet/ml lysis buffer on a horizontal shaking platform and thereafter stored at -80°C. SDS-polyacrylamide gel electrophoresis (60 μg total cellular protein per well) in 8% (for SERCA pumps) and 10% (for LMP-1) gels and transfer onto nitrocellulose membranes was done as described in [11, 14]. The presence of equal amounts of total protein per lane was controlled by Ponceau S staining and densitometry as described previously [11, 14]. SERCA2 and SERCA3 expression was detected using the IID8 and the PLIM430 antibodies, respectively, as described in detail [11, 14, 15]. Luminograms were obtained using the Enhanced Chemiluminescence (ECL) reagent kit of Amersham and were quantitated with the ScionImage software (Scion Corp. CA). Acquisition and quantitative analysis of SERCA signals by Western blotting has been described earlier in detail [11, 14, 15]. LMP-1 was detected similarly by Western blotting using an anti-LMP monoclonal antibody cocktail (Clones CS.1–4, Code N° M 0897) of DakoCytomation, Denmark.

Detection of SERCA3 protein in tissues was performed exactly as described in detail [11]. Briefly, after rehydration 5 μm thick acetone fixed frozen tissue sections were incubated with Tris-buffered saline supplemented with 5% non-fat milk powder and 0.1% Tween-20 (blocking solution) to inhibit non-specific protein binding, followed by incubation with the PLIM430 anti-SERCA3 monoclonal antibody (1 μg/ml) in the same solution. After two cycles of washing with water and blocking solution, antigen binding was detected using biotinylated anti-mouse IgG antibody and avidin-biotin-peroxydase complex (Vectastain ABC kit; Vector Laboratories, Burlingame, CA). Signal was revealed with 3,3'-diaminobenzidine as chromogen, and slides were counterstained with hematoxylin. Omission of the PLIM430 primary antibody, replacement by isotype-matched irrelevant antibody or normal mouse IgG were used as negative controls, and these gave no staining. Strong positive staining of vascular endothelial cells, present in the sections and known to express SERCA3 [21, 63‐66] served as internal positive control.

Immunocytochemistry for LMP-1 expression was performed on BJAB-tTA-LMP-1 cells before and after induction of LMP-1 expression by tetracycline withdrawal. Cells were washed with PBS by centrifugation, resuspended in PBS and cell smears were made on poly-lysine coated microscope slides and dried on air overnight. After fixation in acetone (10 min. at room temperature), slides were rehydrated and stained with a cocktail of anti-LMP monoclonal antibodies (Clones CS.1–4, Code N° M 0897, DakoCytomation, Denmark) according to the instructions of the manufacturer, using an ABC avidin-biotin-peroxydase system with 3,3'-diaminobenzidine as chromogen, and were counterstained with hematoxilin. Photographs of cells were obtained using a Leica microscope and a Nikon Coolpix 950 digital photographic camera.

Cytosolic ionized free calcium concentration ([Ca²⁺]_cyt) was measured by a fluorimetric ratio technique [67]. Cells were centrifuged and resuspended at a density of 10⁶ cells/ml in phosphate-buffered saline (PBS) supplemented with 1 mg/ml bovine serum albumin and incubated in the dark with 4 μM Indo-1 AM for one hour at room temperature under slow agitation. Cells were then centrifuged and resuspended in calcium-free Hanks' buffered saline solution (HBS; 135 mM NaCl, 5.9 mM KCl, 1.2 mM MgCl₂, 11.6 mM Hepes, 11.5 mM glucose adjusted to pH 7.3 with NaOH) prior measurements. After centrifugation, 0.5 to 1 × 10⁶ cells were suspended in 3 ml HBS in a quartz cuvette and inserted into a Shimadzu RF1501 spectrofluorimeter equipped with a stirring apparatus and a thermostatted (37°C) cuvette holder, and connected to a PC computer. Ultraviolet light at 360 nm was used for excitation of Indo-1, and emission at 405 and 480 nm was recorded in time before and after addition of reagents as indicated in figures. Background and autofluorescence of the cell suspensions were subtracted from the recordings. Maximum Indo-1 fluorescence (F_max) was obtained by adding 1 μM ionomycin to the cell suspension in the presence of 10 mM CaCl₂, and minimum fluorescence (F_min) was determined without added calcium in the presence of 5 mM EGTA. Measurements of resting cytosolic calcium concentrations in the presence of extracellular calcium were performed in HBS supplemented with 1 mM CaCl₂. Cytosolic calcium concentrations were calculated according to the equation [Ca²⁺]_cyt = Kd(R-R_min)/R_max-R), where Kd is the apparent dissociation constant of the Indo-1-calcium complex (230 nM), and R is the ratio of fluorescence values (F) measured at 405 and 480 nm (R = F₄₀₅/F₄₈₀).

In appropriate experiments, inhibition of the Na⁺/Ca²⁺ exchangers was obtained by replacement of all Na⁺ by Li⁺ in the HBS bath solution, and plasma membrane calcium pump activity was inhibited by 0.4 μM HgCl₂ [68].

Data are presented as the mean +/- SEM and correspond to at least three independent experiments; the number of experiments performed is indicated in brackets. Statistical analysis was performed using Student's paired t-test.