This work was supported by research grants from the Netherlands Asthma Foundation and ‘Stichting Astma Bestrijding’.
Interaction between nitric oxide and subsets of human T lymphocytes with differences in glutathione metabolism
Article first published online: 8 NOV 2002
Volume 107, Issue 3, pages 334–339, November 2002
How to Cite
Roozendaal, R., Kauffman, H. F., Dijkhuis, A.-J., Ommen, E. T. V., Postma, D. S., De Monchy, J. G. R. and Vellenga, E. (2002), Interaction between nitric oxide and subsets of human T lymphocytes with differences in glutathione metabolism. Immunology, 107: 334–339. doi: 10.1046/j.1365-2567.2002.01502.x
- Issue published online: 8 NOV 2002
- Article first published online: 8 NOV 2002
- Received 5 March 2002; revised 1 July 2002; accepted 9 July 2002.
Nitric oxide (NO) modulates human T-lymphocyte responses through several mechanisms. In the current study we show that interactions between NO and glutathione (GSH) metabolism are related to the selective persistent inhibition of interferon-γ (IFN-γ) production by NO, which we previously identified. T cells were exposed to NO using the NO-donor compound Spermine-nonoate (Sper) and activated using anti-CD3 plus anti-CD28 monoclonal antibodies. Persistent inhibition of IFN-γ by Sper was prevented by addition of the GSH precursor l-cysteine, which inhibits Sper induced GSH depletion. Subsets of cells were either susceptible (GSHlow) or resistant (GSHhigh) to NO-induced GSH depletion. The GSHlow subset was characterized by enhanced numbers of CD4+ cells, reduced numbers of activated cells as characterized by CD25 and CD69, and reduced numbers of memory (CD45RO+) cells relative to the GSHhigh population. Rather than directly affecting susceptibility to NO, these surface markers reflected different expression patterns. Particularly, the GSHlow subset was further characterized by decreased activity of the GSH synthesis related enzymes multi-drug resistance related protein (MRP)-1 and γ-glutamyltranspeptidase (γ-GT). Blocking γ-GT, using acivicin was shown to exacerbate NO-induced GSH depletion and NO-induced apoptosis. Since NO induced apoptosis selectively affects IFN-γ production these findings implicate GSH metabolism in the modulation and maintenance of the T helper (Th)1/Th2 balance.
activation induced cell death
Under resting conditions, there is at least a 100-fold difference between the intracellular levels of reduced and oxidized glutathione (GSH) in virtually all cell types.1 However, oxidized glutathione (GSSG) can rapidly be increased when cells are exposed to oxygen and nitrogen radicals during inflammation. A change in the ratio between GSH and GSSG in favour of GSSG has been described to induce programmed cell death (apoptosis).2,3 Thus, alteration in the balance during inflammation may account for the known apoptotic action of superoxide4 (O2–) and nitric oxide (NO).5,6 Consequently, the survival of cells depends in part on their ability to reconstitute their intracellular GSH/GSSG balance.
While some resynthesis of GSH may occur from precursors that are present intracellularly, GSH import via the sequential action of the enzyme γ-glutamyltranspeptidase (γ-GT) and dipeptidases is required to reconstitute intracellular GSH. The function of γ-GT in the γ-glutamyl cycle is the extracellular breakdown of GSH into cysteinylglycine and the transfer of the glutamyl moiety onto available amino acids. Besides, γ-GT can also extrude GSH itself from the cytoplasm. After further breakdown of cysteinylglycine by dipeptidases, all GSH components can readily be imported into the cell. Intracellularly, GSH is resynthesized from its imported components. High expression of γ-GT may confer resistance to apoptosis induced by oxidative and nitrosative stress, by replenishment of intracellular GSH. In accordance, high γ-GT levels and high intracellular GSH intracellular are known features of neoplastic cells7,8 and memory T cells.9 Additionally, γ-GT can break down the nitrosothiol nitrosoglutathione (GSNO), which is formed during nitrosative stress.10,11
It has previously been shown that polarized cultures of human T lymphocytes are differentially susceptible to NO-induced apoptosis.12 T helper 2 (Th2) polarized cells were shown to resist NO-induced apoptosis by means of elevated γ-GT expression. In the current study we examine whether the interactions between NO and GSH metabolism in human T lymphocytes can account for persistent interferon-γ (IFN-γ) suppression by selective induction of apoptosis in Th1-like cells. We demonstrate that NO causes GSH depletion and apoptosis in a subset of human T lymphocytes, dependent on expression of γ-GT. High γ-GT activity is found to characterize a subset of human T-lymphocytes with a reduced sensitivity to NO-induced apoptosis and GSH depletion.
Materials and methods
Spermine-nonoate (Sper) was obtained from Kordia, (Leiden, the Netherlands). The multi-drug resistance related protein (MRP)-1 inhibitor MK571 was a gift of Dr A.W. Ford-Hutchinson (Merck-Sharp, Kirkland, Quebec, Canada). Acivicin, and all other chemicals used in assays were from Sigma-Alldrich (Zwijndrecht, the Netherlands), unless stated otherwise.
Preparation of cells
Mononuclear cells were obtained from the peripheral blood of healthy volunteer platelet donors by Ficoll-Hypaque (Lymphoprep, Nycomed, Oslo, Norway) density gradient centrifugation. T lymphocytes were isolated by 2-aminoethylisothiouronium-treated sheep red blood cell (SRBC) rosetting. The SRBC were lysed with 155 mm NH4Cl, 10 mm KHCO3 and 0·1 mm ethylenediaminetetraacetic acid (EDTA). The remaining cell population contained more than 95% lymphocytes as determined by flow-cytometric analysis after staining with a monoclonal antibody (mAb) against CD2 (Becton Dickinson, Sunnyville, CA). Prior to stimulation T lymphocytes were cultured overnight, at 37° and 5% CO2 in RPMI-1640 (Flow, Rockville, MD) supplemented with 5% fetal bovine serum (HyClone, Logan, UT) and gentamicin (BioWhittaker, Verviers, Belgium).
Stimulation and staining
Cell cultures were incubated with antibodies against CD3 plus CD28 (a gift from B. J. Kroesen, Department of Immunology, University of Groningen, the Netherlands), added as 5% hybridoma supernatant for 6–48 hr. When appliccable, the exogenous NO donor compound Sper was added 15 min before stimulation with antibodies. Buthiomine sulphoximide (BSO), acivicin and MK571 were added 30 min prior to the NO-donor compound.
IFN-γ enzyme-linked immunosorbent assay (ELISA)
IFN-γ secretions were determined in the cell free supernatants using PelikineTM compact human cytokine ELISA kits (Central Laboratory of the Blood Transfusion Service, Amsterdam, the Netherlands), according to the manufacturer's instructions.
Monochlorobimane (MCB, Omnilabo, Breda, the Netherlands)-dependent fluorescence was taken as a measure of intracellular glutathione. Cells were washed, stained with 100 µm MCB on ice for 1 hr at 5 × 106 cells/ml, washed, resuspended in RPMI, and kept on ice until analysis (no more than 1 hr). All analyses were performed using an EliteTM flow cytometer (Beckman Coulter, Florida, Hialeah, FL) and WinlistTM software (Verity, Topsham, ME) for all fluorescence-activated cell sorting (FACS) analyses described. Cells were stimulated with anti-CD3/28 for 16 hr in the presence of 10 µm Sper prior to sorting of the GSH high and low populations with a MoFloTM flow cytometer (Cytomation, Fort Collins, CO). Subsets were stained with mAbs against CD4-CyQ and CD45RO-FITC (ImmunoQuality Products, Groningen, the Netherlands), CD25 and CD69 (phycoerythrin (PE), Becton Dickinson, San Jose, CA) for 30 min on ice (5 µl antibody per 1 × 106 cells/50 µl), washed and kept on ice prior to analysis.
γ-GT activity assay
γ-GT activity was determined as previously described.12 Polarized cell cultures (5 × 105 cells/ml) were incubated for 24 hr at 37° and 5% CO2 in RPMI-1640 containing 0·5 µCi/ml 35S-GSH (New England Nuclear, Zaventem, Belgium) plus 100 µm carrier GSH in the presence or absence of 100 µm acivicin. Cells were harvested and washed twice. Total cell associated radioactivity was determined by scintillation counting.
Plasma membrane asymmetry was assessed with annexin staining as previously described13 after 16 hr. Briefly, cells were washed, resuspended in 140 mm NaCl/2·5 mm CaCl2 and stained with PE-conjugated annexin V (Immuno Qualtity Products, Groningen, the Netherlands). In some experiments this was combined with MCB staining. Propidium iodide (PI) staining was used afterwards to check for plasma membrane integrity.
MRP function analysis was performed essentially as previously described, to asses the capability of cells to extrude GSH-conjugates.14 Cells were loaded with the MRP-1 substrate carboxyfluorescine-di-acetate (0·1 µm) in the presence MK571 (10 µm) for 20 min at 37°. Cells were washed to remove MK571 and incubated for 1 hr at 37° in the presence or absence of MK571. MRP-1 activity was calculated from the fluorescence ratio with/without MK571, and was called the efflux blocking factor.
P-values < 0·05 were assumed to represent significant differences. Student's t-test for paired and unpaired observations was used to calculate P-values.
NO-induced inhibition of IFN-γ is related to GSH metabolism
T cells activated with anti-CD3 plus anti-CD28 demonstrated a significant increase in intracellular GSH in time, as assessed with MCB staining, whereas cotreatment with Sper resulted in a significant decrease in GSH levels (Fig. 1a). GSH levels returned to control values within 48 hr after Sper exposure caused by decay of Sper. Therefore, subsequent figures will be displayed up to the 24-hr time point. The treatment with Sper (5 µm) also caused a marked inhibition of IFN-γ production (to 49 ± 11% of control after 24 hr stimulation, P < 0·001), which could be fully prevented with the glutathione precursor l-cysteine (Fig. 1b). Direct quenching of NO metabolites by l-cysteine could not be excluded.15–17 MCB analyses further revealed that not all cells were equally susceptible to NO-induced GSH depletion. A subpopulation of cells underwent GSH depletion (GSHlow), whereas the remainder of the cells remained largely unaffected by NO exposure (GSHhigh; Fig. 1c).
Phenotypical chacterization of GSHhigh and GSHlow subsets
To further characterize the phenotype of the GSHhigh and GSHlow subsets, the expression of the activation markers CD25 and CD69, the ‘memory’ marker CD45RO and the helper T-cell marker CD4 was determined with FACS analysis. A larger proportion of GSHhigh cells was positive for the activation marker CD69, than of the GSHlow subset (83·6 ± 7·1% versus 48·2 ± 12·4% positive cells, respectively, P < 0·01) (Fig. 2). The same type of distribution was observed for CD25 (73·5 ± 11·7% versus 30·1 ± 9·4% positive cells for GSHhigh and GSHlow, respectively, P < 0·01). CD4 positive cells were preferentially localized to the GSHlow subset (50·3 ± 7·2% versus 70·9 ± 9·7% positive cells for GSHhigh and GSHlow, respectively, P < 0·001). Finally, a larger proportion of GSHhigh cells was positive for CD45RO (51·4 ± 5·6% versus 25·4 ± 2·3% positive cells for GSHhigh and GSHlow, respectively, P < 0·01). While the surface expression of activation, helper and memory markers is correlated with susceptibility to NO-induced GSH depletion, none of these markers can in theory directly confer susceptibility or resistance, but rather reflect alternative expression patterns of other proteins.
Functional characterization of GSHhigh and GSHlow subsets
The MRP-1 antagonist MK571 was used to impede export of GSH conjugates. In addition, γ-GT function was also investigated, using the specific inhibitor acivicin. Neither of the inhibitors exhibited significant toxicity under these conditions.
MK571 largely prevented the increase in GSH in αCD3 plus αCD28 activated cells at all time points (P < 0·01 at all time points, Fig. 3a). Even though NO-induced GSH depletion was already pronounced, MK571 significantly enhanced GSH depletion (P < 0·05 at t = 24). Significant MRP-1 activity was detected in all subsets (i.e. efflux blocking factors > 1, P < 0·01 in all cases). It appears that MRP-1 function is important to maintain normal GSH levels in the face of NO-mediated oxidative stress, as the GSHlow subset was characterized by lower MRP-1 activity in both the GSHlow CD25– and the GSHlow CD25+ population (Fig. 3b, efflux blocking factor 1·4 ± 0·2 versus 1·5 ± 0·3 for GSHlow CD25– and GSHlow CD25+, respectively). In contrast, GSHhigh CD25+ is characterized by elevated MRP-1 activity relative to both the GSHhigh CD25– and the GSHlow populations (efflux blocking factor 2·1 ± 0·5, P < 0·05 in all cases).
γ-GT function did not appear to be required to support the initial elevation of intracellular GSH that occurred upon stimulation, since acivicin decreased intracellular GSH relative to control after 24 hr of stimulation (P < 0·05 at t = 24, P = 0·38 at 12 hr, up to 24% inhibition; Fig. 4a). γ-GT inhibition significantly enhanced NO-induced GSH depletion (P < 0·01 at t = 24, up to 39% inhibition). γ-GT function was also important to maintain GSH levels with NO-induced stress, since significant γ-GT activity was only detectable in the GSHhigh subset (3·5 ± 0·3, P < 0·05 relative to control; Fig. 4b).
GSH depletion is related to NO-induced apoptosis
NO-induced apoptosis exclusively occurred in the GSHlow population (Fig. 5a). None of the inhibitors induced significant apoptosis by themselves. While BSO most markedly enhanced NO-induced GSH depletion, only acivicin significantly enhanced NO-induced apoptosis (Fig. 5b and 8·3 ± 3·9% versus 11·8 ± 3·6% apoptosis, P = 0·02). Thus, decreased GSH levels per se do not appear to be related to the induction of apoptosis.
In this paper we describe the interaction of NO, generated from the nonoate compound Sper, with GSH metabolism in freshly isolated human T-lymphocytes. NO-mediated inhibition of IFN-γ production by anti-CD3/28 activated human T-lymphocytes is related to NO-induced GSH depletion. Not all T cells were equally susceptible to NO-induced GSH depletion. Resistant cells were characterized by increased expression of the activation markers CD25 and CD69 relative to cells susceptible to NO-induced GSH depletion. Furthermore, CD4 cells were more susceptible to GSH depletion than CD4-negative lymphocytes. However, the expression of none of these markers can directly account for the differences between GSH metabolism between the GSH depleted and the NO-resistant subset. Therefore, we functionally characterized the subset susceptible and resistant to NO-induced GSH depletion. It was demonstrated that the resistant subset is further characterized by elevated activity of MRP-1 and γ-GT. Blocking of either MRP-1 or γ-GT function caused enhanced NO-induced GSH depletion, while blocking of γ-GT also enhanced NO-induced apoptosis. Thus, both γ-GT and MRP-1 offer general protection against oxidative stress, whereas γ-GT also specifically protects against reactive NO intermediates.
Most cells can tolerate extensive GSH depletion without adverse effects. GSH maintenance in the current experimental set-up appeared to be important for cell homeostasis. In the GSH depleted subset, a proportion of the cells underwent programmed cell death. Although GSH depletion apparently was required for the induction of apoptosis, it was not sufficient, since the majority of GSH-depleted cells did not enter apoptosis. While inhibition of MRP-1 and γ-GT (with MK571 and acivicin, respectively) enhanced GSH depletion, GSH depletion per se did not account for the increase in apoptosis that was also observed, since only acivicin enhances apoptosis. Nevertheless, the subset of cells that was susceptible to NO-induced GSH depletion was also characterized by low MRP-1 activity, irrespective of activation status. Therefore, low MRP-1 activity of the GSHlow subset is likely to be related to a general susceptibility to oxidative stress. Low basal MRP-1 activity might result in elevated levels of GSSG when cells are exposed to reactive nitrogen intermediates. It has been described that the ratio between oxidized and reduced GSH functions as a redox sensor deciding between survival and apoptosis.2,3 Because the ratio between GSH and GSSG generally is in excess of 100,1 it is much more affected by an increase in GSSG than a decrease in GSH. Therefore, the concomitant decrease in GSH caused by the formation of GSSG is likely to be of relatively little importance for the GSH/GSSG ratio.
Our findings suggest that the expression of γ-GT and MRP-1 identifies a dichotomy in the human peripheral T-lymphocyte population with respect to sensitivity towards NO-induced GSH depletion. It has recently been shown that polarized human Th1 cells are more susceptible to NO-induced apoptosis than Th2 cells.12 In the current study γ-GT expression also appears to confer resistance to NO-induced apoptosis, whereas low MRP-1 activity identifies a subset that is susceptible to oxidative stress in general. Moreover, it has previously been shown that IFN-γ is selectively affected by NO-exposure. Therefore, the GSHlow subset that is observed after NO exposure, is likely to contain a predominance of Th1-like cells.
A dichotomy towards NO-induced apoptosis has recently been identified in Jurkat T-lymphocytes.18 The authors reported subpopulations with high and low cardiolipin content after NO exposure. While Jurkat cells clearly are of more homogeneous composition, we speculate that differences in intracellular NO metabolism may give rise to the difference in cardiolipin oxidation in Jurkat cells as well. In peripheral blood lymphocytes we do not find an increase of apoptosis when GSH is depleted using BSO (data not shown). This is probably due to the fact that these primary cells are less susceptible to activation induced death than Jurkat cells. GSH depletion in itself is not sufficient to induce apoptosis in the majority of cells, since acivicin increases apoptosis independently of a further decrease in GSH. Because NO-induced apoptosis can first be demonstrated 12 hr after exposure, this provides circumstantial evidence that longer-lived NO-metabolites are involved. An interesting candidate is GSNO, with an in vitro half-life of several hours20 although its in vivo half-life may be a lot shorter in particular because of the action of γ-GT,21 as GSNO is a substrate for γ-GT.22 Thus, the increased apoptosis upon addition of the γ-GT inhibitor acivicin appears to argue in favour of the involvement of GSNO in NO-induced apoptosis.
Interestingly, non-memory CD4+ cells are most susceptible to NO-induced GSH depletion and apoptosis. This may have important implications for diseases like asthma. Because the suppression of IFN-γ but not interleukin (IL)-4 or IL-523 is correlated with NO-induced apoptosis in activated human T lymphocytes, differential expression of MRP-1 and γ-GT could conceivably contribute to the maintenance of the Th1/Th2 balance in inflammatory diseases. MRP-1 has recently been implicated in control of dendritic cell migration, and could thus also affect the T-cell response at the level of antigen presentation.24 Research is in progress to further dissect the effector mechanisms of intracellular nitric oxide metabolism.
We thank G. Mesander and H. Moes for technical assistance, D. M. van der Kolk for helpful discussions, and G. H. Koenderink for critical reading of the manuscript.