Dehydroepiandrosterone and metformin regulate proliferation of murine T lymphocytes

Abstract

The aim of the present study was to assess the effect of dehydroepiandrosterone (DHEA: 10 µM) and metformin (10 µM and 100 µM) in regulating proliferation of cultured T lymphocytes. T cells were isolated from lymph nodes of prepuberal BALB/c mice. We found that DHEA, metformin and DHEA + metformin added to the incubation media diminished proliferation of T cells. The inhibition by DHEA was higher than that produced by metformin, while the combined treatment showed a synergistic action that allowed us to speculate distinct regulatory pathways. This was supported later by other findings in which the addition of DHEA to the incubation media did not modify T lymphocyte viability, while treatment with metformin and DHEA + metformin diminished cellular viability and increased both early and late apoptosis. Moreover, DHEA diminished the content of the anti-oxidant molecule glutathione (GSH), whereas M and DHEA + metformin increased GSH levels and diminished lipid peroxidation. We conclude that DHEA and metformin diminish proliferation of T cells through different pathways and that not only the increase, but also the decrease of oxidative stress inhibited proliferation of T cells, i.e. a minimal status of oxidative stress, is necessary to trigger cellular response.

Introduction

It is well known that the immune and endocrine systems are associated during both physiological and pathological processes. An example of this is that immune mechanisms regulate ovarian functions [1,2] and that T lymphocytes (LT) are associated with endocrine pathologies [3,4]. In particular, imbalances in the expression of T cell subsets and altered concentrations of cytokines have been reported in ovaries of women with polycystic ovary syndrome (PCOS) [5,6]. PCOS is a heterogeneous disease characterized by hyperandrogenaemia, hirsutism, oligo- or amenorrhoea and anovulation and is associated frequently with hyperinsulinaemia, insulin resistance syndrome, increased cardiovascular risk and diabetes mellitus [7]. Because dehydroepiandrosterone (DHEA) is one of the most abundant androgens produced by ovaries of women with PCOS, a rodent model was developed by injection of DHEA [8] and subsequent studies corroborated that the DHEA–PCOS murine model exhibits many of the salient features of human PCOS [9–15]. We have demonstrated previously that injection of BALB/c mice with DHEA produces endocrine and immune disturbances [11–15]. In that context we have reported that DHEA not only modified the LT infiltration in both the ovarian tissue and the retroperitoneal lymph nodes [11,13], but also increases serum tumour necrosis factor alpha (TNF-α) of treated BALB/c mice [13]. These results are in agreement with those reported in the literature concerning the immune functions of DHEA [16,17]. By binding to specific receptors on T cells [18], DHEA modulates the rates of cell proliferation, cell viability and cytokine production [19].

Proliferation or activation of LT is regulated by the oxidant–anti-oxidant balance. Thus, reactive oxygen species (ROS) play an important role as second messengers and can elicit either activation or apoptosis of T cells [20,21]. Nitric oxide (NO), an essential messenger for maintaining homeostasis [22], has been described as an active intermediary of cellular immune responses [23,24]. However, controversial studies have described both a role of NO in regulating T proliferation [25,26] and the influence of DHEA in the NO system [27]. Cellular redox status is maintained by intracellular anti-oxidant molecules or enzymes. Glutathione (GSH) is a tripeptide thiol that participates in cellular protection as a major anti-oxidant metabolite and disturbances on GSH metabolism are related directly to activation of T cells [28,29].

Although metformin (N,N′-dimethylbiguanide) is employed widely in the treatment of PCOS, the mechanism involved remains unknown. This biguanide, used primarily as an anti-diabetic drug, is also related to inflammatory immune processes [30] and is a scavenger of ROS [31]. In previous studies, we have reported that metformin reverses ovarian oxidative stress and modulates immune parameters (expression of ovarian and retroperitoneal LT phenotypes, serum TNF-α levels) provoked by DHEA [13]. The aim of the present study was to explore whether the previously described effect of DHEA on LT [11,13] involves the modulation of oxidative stress of T cells and the possible role of metformin treatment.

Materials and methods

Animal care

T lymphocytes were obtained from lymph nodes of female prepuberal (25-day-old) mice of the BALB/c strain. Animals were housed under controlled temperature (22°C) and illumination (14 h light : 10 h dark; lights on at 05.00 h) and were allowed free access to Purina rat chow and water. Mice were anaesthetized with ether and killed by cervical dislocation. All procedures involving animals were conducted in accordance with the Animal Care and Use Committee of the Consejo Nacional de Investigaciones Científicas y Tecnológicas (CONICET), 1996.

Cell cultures and incubations

The preparation of T cell suspensions was carried out as described previously [32]. Briefly, lymphoid organs were removed and homogenized gently in RPMI-1640 medium with a loosely fitting Teflon glass homogenizer; the cell suspension was filtered through 1-mm metal mesh and needles. Then, T cells were purified by passages through nylon-wool columns and washed three times in RPMI-1640. The T cells were harvested and dead cells were removed by centrifugation through Histopaque-1077 gradient (d = 1·077 g/ml; Sigma, St Louis, MO, USA) at 320 g for 10 min. The viable cells (> 98%) were collected and washed with RPMI-1640 complete medium (Sigma) supplemented with 10% (v/v) heat-inactivated fetal bovine serum, 2% l-glutamine and 1% penicillin–streptomycin (Sigma). The isolated cells were confirmed as T cells based on flow cytometric assay for the presence of CD3 antigen.

In order to establish the concentration of DHEA to be used in the experiments we first carried out a dose–response analysis of DHEA action on LT proliferation. T cells (3 × 10⁵ per well) were cultured in 96-well microtitre plates (Nunc Multi-dish; San Diego, CA, USA) with 0·1, 1, 10 or 15 µM of DHEA for 4 h. Different concentrations of DHEA were prepared in the culture medium from a DHEA stock solution (10 mM in dimethylsulphoxide; Sigma). We found that the DHEA vehicle did not interfere with the measurements. After incubation with DHEA for 4 h, cultured LT were added with 4 µg/ml of concanavalin A (ConA) and assays were carried out for a total of 72 h in complete RPMI-1640 in triplicate under an atmosphere of 95% O₂and 5% CO₂air at 37°C. During the last 18 h of culture, 0·5 µCi/well [³H]-thymidine ([³H]-TdR, NEN, AE 20 Ci/mmol) was added to quantify cellular proliferation. Cultures were harvested using a Harvester 96 (Tomtec, Orange, CT, USA) and the index of LT proliferation was determined by evaluating the incorporation of [³H]-thymidine to cells with a beta plate liquid scintillation counter (Wallac, Ramsey, MN, USA). Negative controls were also made in cultures without ConA. Results are expressed as counts per minute of incorporated [³H]-thymidine. Because 10 µM DHEA was the lowest dose that inhibited LT proliferation (Fig. 1a), this was used for further experiments.

Fig. 1.

Modulation of LT proliferation by dehydroepiandrosterone (DHEA) and metformin (M). (a) Lymphocytes from prepuberal BALB/c mice were incubated with 0·1, 10 and 15 µM DHEA stimulated to proliferate with concanavalin A (ConA) and added to [³H]-thymidine in RPMI-1640 complete media, as described in Materials and methods. Values represent the mean ± standard error of the mean (s.e.m.); d versus a, e versus a, P < 0·001. (b) Lymphocytes from prepuberal BALB/c mice were incubated with DHEA (10 µM) and/or metformin (M; 10 µM and 100 µM), stimulated to proliferate with ConA and added to [³H]-thymidine in RPMI-1640 complete media, as described in Materials and methods. Values represent the mean ± s.e.m.; c versus a, P < 0·05; b, d versus a, P < 0·01; e, f, versus a, P < 0·001; c versus d, P < 0·001; e versus c, f versus d, P < 0·001.

T lymphocytes were treated in vitro according to the following groups; (i) control: the cells were incubated in serum-free RPMI-1640 media; (ii) DHEA: the culture media contained 10 µM DHEA (Sigma); (iii) metformin: culture media contained either 10 or 100 µM metformin, (Montpellier, Buenos Aires, Argentina); or (iv) DHEA + metformin: the culture media contained 10 µM DHEA and either 10 or 100 µM metformin. Metformin was dissolved directly in culture medium. The incubations were carried out at 37°C in a humidified atmosphere of 95% O₂ : 5% CO₂ and after this time different assays were followed as described below. To ensure the same conditions, all the experiments were carried out with the same concentration of T cells; 3 × 10⁵ per well in 96-well plates were incubated with DHEA, metformin and its combination for 4 h except for the annexin V assay, which was also performed for 21 h as 4 h of incubation did not affect apoptosis. To carry out the different assays, the cells were recovered from the 96-well plates and used in the concentration indicated in each section.

Determination of viability, apoptosis and necrosis by annexin V assay

In order to determine cellular viability, apoptosis early and late apoptosis or necrosis (LA/N), T cells were treated with an annexin V kit (Calbiochem, Gibbstown, NJ, USA) immediately after collection. The kit contains both fluorescein isothiocyanate (FITC) conjugated to annexin V and propidium iodide (PI). Viable T cells do not bind FITC–annexin V and do not stain nuclear formation with PI, while late apoptotic or necrotic cells bind FITC–annexin V but also stain their nucleus with PI (because of the permeabilization of cellular membranes during necrosis). Early apoptotic cells bind FITC–annexin V but exclude PI. Quantification is carried out by flowcytometry, as the FITC signal can be detected by FL1 (FITC detector) at 518 nm, while PI can be detected at 620 nm and can be detected by FL2 (the phycoerythrin fluorescence detector). Viable cells are shown in the lower left-hand quadrant of the dot-plot graph, early apoptotic cells in the lower right-hand quadrant and necrotic or apoptotic cells in the upper right-hand quadrant. All procedures were carried out following the manufacturer's instructions. Briefly, cell T suspensions of each treatment (control, DHEA, metformin and DHEA + metformin) were incubated for 21 h because incubation for 4 h modified neither apoptosis nor necrosis. After incubation, 1 × 10⁶ of T cells in 1 ml of RPMI-1640 from each treatment were incubated in polypropylene tubes with 1·5 µl annexin V for 15 min at room temperature (RT) (18–24°C) in the dark. Cells were then centrifuged at 100 g for 5 min at RT, the media removed and cells resuspended in 500 µl of phosphate-buffered saline (PBS: 137 mM NaCl, 2·7 mM KCl, 4·3 mM Na₂HPO₄.7H₂O, 1·4 mM KH₂PO₄, pH 7·4). Finally, each sample was added to 10 µl PI and then placed on ice away from light. Flow cytometry assays were carried out immediately. Fluorescence analysis was evaluated with FACScan® and Winmdi 2·8 software and results are expressed as the percentage of cells at different stages.

Oxidative stress-related parameters

Lipid peroxidation

The amount of malonialdehyde (MDA) formed from the breakdown of polyunsaturated fatty acids, of which the cellular membrane is composed, is taken as an index of peroxidation. The method used in the present study was as described previously [33] and quantifies MDA as the product of lipid peroxidation that reacts with trichloracetic acid–thiobarbituric acid (TCA)–HCl [15% (w/v), 0·375% (w/v) and 0·25 M respectively] yielding a red compound that absorbs at 535 nm. Briefly, 3 × 10⁷ cells from each treatment (control, DHEA, metformin and DHEA + metformin incubated for 4 h) were lysed in 300 µl of 5% TCA and supernatants were collected to quantify MDA. Supernatants were treated with TCA–thiobarbituric acid–HCl and heated for 15 min in a boiling water-bath. After cooling, the flocculent precipitate was removed by centrifugation at 1000 g for 10 min. The absorbance of samples was determined at 535 nm. Content of thiobarbituric acid reactants is expressed as pmol MDA formed/10⁷ cells.

Glutathione content

The GSH assay was carried out as described previously [34]. The reduced form of GSH comprises the bulk of cellular protein sulphydryl groups. Thus, measurement of acid-soluble thiol is used commonly for estimation of GSH content. Briefly, 300 µl of homogenates, obtained when 3 × 10⁷ T cells from each treatment (control, DHEA, metformin and DHEA + metformin incubated for 4 h) were lysed in 5% TCA, were incubated with 1·75 M Tris buffer (pH 7·4) containing nicotinamide adenine dinucleotide phosphate and GSH reductase. The reaction involves enzymatic reduction of the oxidized form (GSSG) to GSH. When Ellman's reagent (a sulphydryl reagent, 5,5-dithiobis-2 nitrobenzoic acid; Sigma) is added to the incubation medium, the chromophoric product resulting from this reaction develops a molar absorption at 412 nm that is linear up to 6 min; after this, the reaction remains constant. Results are expressed as nmol GSH/10⁷ cells.

Nitric oxide synthase activity

Nitric oxide synthase (NOS) activity was measured by monitoring the production of [l-¹⁴C] citrulline from [l-¹⁴C] arginine, as described previously [34]. Briefly, 1 × 10⁷ cells obtained from different treatments (control, DHEA, metformin and DHEA + metformin incubated for 4 h) were incubated in 500 µl Krebs Ringer phosphate buffer with glucose (11·0 mmol/l) as external substrate (pH = 7·0) with 10 µmol/l [l-¹⁴C] arginine (0·3 µCi; 1 Ci = 37 GBq) for 1 h in a Dubnoff metabolic shaker under an atmosphere of 5% CO₂ in 95% O₂ at 37°C. After incubation, cells were disrupted by sonication (Vibra-cell, Sonics and Materials, Newtown, CT, USA) in medium containing 10 mM ethyleneglycol tetraacetic acid, 0·1 mM citrulline, 0·1 mM dithiothreitol and 20 mM HEPES, pH 7·5. After centrifugation at 5000 g for 10 min, supernatants were applied to 2 ml columns of DOWEX AG50WX-8 (Na⁺ form; BioRad, Hercules, CA, USA) resin. Radioactivity was measured by beta liquid scintillation counting. Results are expressed as nmol NO/min × 10⁷ cells.

Statistical analysis

Statistical analyses were carried out using the Instat program (GraphPAD Software, San Diego, CA, USA). Analysis of variance and Student–Newman–Keuls and Xi tests for percentages were used for comparisons between values of groups. P < 0·05 was considered significant.

Results

Dehydroepiandrosterone and metformin modulate LT proliferation

We first carried out a dose–response analysis of the action of DHEA on LT proliferation (Fig. 1a). Because 10 µM DHEA was the lowest dose that inhibited LT proliferation (d versus a, P < 0·001), this was used for further experiments. DHEA, metformin and its combination added to the incubation media significantly diminished proliferation of T cells (c versus a, P < 0·05; b, d versus a, P < 0·01; e, f versus a, P < 0·001; Fig. 1b). Inhibition by metformin in a dose-dependent fashion (c versus d, P < 0·001) and incubation with DHEA + metformin showed a synergistic inhibition of T proliferation when compared with DHEA or metformin alone (e versus c, f versus d, P < 0·001).

Dehydroepiandrosterone and metformin modulate LT viability, early apoptosis and LA/N

The annexin V–FITC assay was used to determine whether the inhibition of LT proliferation generated by DHEA and metformin involved the early apoptosis (EA) and LA/N pathways. Neither the viability, EA or LA/N were modified when treatments (DHEA, metformin and its combinations) were carried out for 4 h of incubation (data no shown). However, when incubations were carried out for 21 h (Fig. 2a: dot-plot analysis and Fig. 2b: percentages of LT) we found that DHEA did not modify LT viability and therefore did not modify EA or LA/N (Fig. 2b). On the other hand, LTs incubated with metformin (10 or 100 µM) and DHEA + metformin diminished LT viability significantly (a versus d, P < 0·001) by increasing the percentages of LT in both EA and LA/N (Fig. 2b: b versus e, c versus f, P < 0·001). Differences between percentages of LTs on EA and LA/N were significantly higher in metformin and DHEA + metformin treatments when compared with the DHEA group (Fig. 2b: b versus c, P < 0·05; e versus f, P < 0·001).

Fig. 2.

Modulation of T lymphocytes (LT) viability, early apoptosis (EA) and late apoptosis/necrosis (LA/N) by dehydroepiandrosterone (DHEA) and metformin (M). (a) Dot-plot analysis and (b) percentages of LT. Annexin V–FITC assay was used to determine whether the inhibition on LT proliferation generated by DHEA and M involved the EA or LA/N pathways, as described in Materials and methods. Values represent the mean ± standard error of the mean. a, b versus e, P < 0·001; c versus f, P < 0·001; b versus c, P < 0·05; e versus f, P < 0·001.

Dehydroepiandrosterone and metformin regulate oxidative stress of LT

We evaluated further whether the inhibition on LT proliferation and viability by DHEA and metformin involved changes in the oxidant–anti-oxidant balance of T cells. Thus, we determined lipid peroxidation as a measure of the damage on LT membrane. As seen in Fig. 3a, DHEA did not modify MDA production when compared with the control group (a versus b, not significant). However, when LTs were exposed to different doses of metformin (10 or 100 µM) and to the DHEA + metformin treatment, the concentration of MDA diminished significantly (a versus c; a versus d; e, f versus a, P <0·001 respectively). Then, we can assume that the inhibition of MDA production by the DHEA + metformin treatment is due to the action of metformin.

Fig. 3.

Modulation of oxidative stress of T lymphocytes (LT) by dehydroepiandrosterone (DHEA) and metformin (M). To evaluate oxidative stress, LT were lysed and treated as described in Material and methods. Values represent the mean ± standard error of the mean. (a) Lipid peroxidation as a measure of the damage on LT membrane; a versus c, a versus d; e, f versus a, P< 0·001 respectively. (b) Intracellular glutathione (GSH) content as an anti-oxidant metabolite; a versus b, P< 0·01; a versus c, a versus d, c versus d, P< 0·001; a versus e, f, P< 0·05. (c) Nitric oxide synthase (NOS) activity by the conversion of 14 C l-arginine into 14 C l-citrulline, as described in Materials and methods. b, c, d, e, f versus a, P < 0·001.

The intracellular content of total GSH was evaluated as a measure of the anti-oxidant defence of T cells (Fig. 3b). While DHEA diminished significantly the GSH content of T cells (a versus b, P < 0·001), metformin increased GSH content significantly in an inverse dose-dependent manner (low doses were more efficient than higher doses) (a versus c, a versus d, c versus d, P < 0·001). When LTs were exposed to DHEA + metformin treatments, the GSH content remained higher than controls (Fig. 3b; a versus e, f, P < 0·05).

As an indirect measurement of NO formation after DHEA and metformin treatments, we evaluated the NOS activity from treated T cells. Figure 3c shows that DHEA, metformin and their combinations increased NOS activity (Fig. 3c; b, c, d, e, f versus a, P < 0·001) without significant differences among treatments.

Discussion

Although the relationship between androgens and the immune system has been well established, the mechanisms involved in the immunomodulatory functions of DHEA remain unknown. The purpose of this study was to assess the direct action of DHEA on LT and some aspects involved in that possible regulation. We found that DHEA inhibited proliferation of LTs in a range of concentrations that were in agreement with those reported previously [35]. In the present investigation, the DHEA dose (10 µM) we used is undoubtedly high when compared with the normal serum levels of this hormone, but the local and intracellular DHEA levels in the steroid target tissue are uncertain. The average serum DHEA sulphatase (DHEA-S, the circulating metabolite of DHEA) increases to 20–70 µM following administration of pharmacological dosages of DHEA without any sign of systemic toxicity [36]. The circulating DHEA-S is converted to DHEA by ubiquitous steroid sulphatases. Therefore, DHEA levels in plasma may not reflect the actual concentrations of DHEA in the target tissue, which could be higher than that observed in the circulation. For all these reasons, even higher doses of DHEA than we have used are accepted in the literature for the in vitro incubation of LT [35,36].

Several studies have been carried out in order to elucidate the DHEA effects on the activation and proliferation of lymphocytes, but have resulted in controversial results [19,37]. However, Yao et al.[38] have demonstrated recently that the androgen structure is determinant in resulting in either the inhibition or the enhancement of LT proliferation. Thus, the authors have demonstrated that small structural changes of androgens (testosterone, DHEA or androstenedione) can result in markedly different biological effects. In that sense, our findings are in agreement with those authors who have reported the inhibition of LT proliferation by DHEA [19,39]. We found that metformin also inhibited proliferation of LT, inducing EA significantly higher than LA. Controversial reports have described that metformin can either prevent or enhance apoptosis [39,40]. Our findings are in agreement with those that describe inhibition of cellular growth in breast, hepatocytes and epithelial cells [40]. Interestingly, in agreement with the results of the DHEA and metformin groups on viability, apoptosis and necrosis, it seems that the inhibition of LT proliferation generated by DHEA and metformin follows different mechanisms. With regard to this point we have to consider, first, that DHEA did not modify viability (and consequently neither apoptosis or necrosis) and that incubations with metformin and DHEA + metformin did, and secondly, that there was a synergistic action of DHEA and metformin on proliferation of LT. Based on these points we could assume that DHEA and metformin inhibited LT proliferation through two different pathways.

Given that ROS act as intermediaries able to trigger either the activation or apoptosis of T cells [20,21], we also studied the action of treatments in the oxidant–anti-oxidant balance of LT. We found that DHEA diminished GSH content and for this reason acted as a pro-oxidant agent even when it had no effect on lipid peroxidation. Conversely, metformin acted as an anti-oxidant agent because it not only diminished lipid peroxidation but also increased GSH content, even above the control values. In agreement with these findings, it has been reported previously that the molecular structure of metformin gives biguanide a strong condition as a scavenger of ROS [15,31]. Therefore, and according to previous findings [20,21], high levels of oxidative stress can induce inhibition of LT proliferation as shown in the DHEA treatment. However, the diminution of basal levels of oxidative stress, as in the case of metformin, also inhibited proliferation of LT, thus suggesting that determined concentrations of ROS are necessary to induce proliferation of T cells. These data could explain the controversial results reported in relation to the effects of DHEA on LT proliferation [19,37] and are in agreement with the fact that blocks of lipid peroxidation can affect T cell activation [41]. Controversial reports have described the role of NO in regulating LT proliferation [25,26]. Our results show that the increase of NOS activity, and consequently NO levels, lead to the inhibition of T proliferation.

In the present study we found that DHEA has pro-oxidant effects on T cells, whereas metformin shows anti-oxidant actions. These data are in agreement with our previous findings in which we studied the in vivo treatment with metformin of hyperandrogenized mice [13–15]. Here we have also demonstrated that both DHEA and metformin inhibit T cell proliferation and that metformin, but not DHEA, diminishes T cell viability and increases apoptosis/necrosis. Beyond the reported effects of metformin on insulin-responsive tissues, little is known about the effects of metformin on other cell types. During recent years it has been reported that metformin is able to exert anti-oxidant actions [13–15,31] and that it can even have an anti-tumoral effect [42,43]. Controversial reports describe the relationship between metformin and apoptosis [39,44], and as we have found in T cells it has been reported previously that metformin is able to induce apoptosis in H4IIE cells [44]. We suggest that metformin decreases T cell viability and increases apoptosis/necrosis of the T cell response to the anti-inflammatory role of metformin. In this regard, it has been reported that metformin interacts with the immune system by modulating the migration of macrophages [30]. In fact, the anti-inflammatory effect of metformin has been described to prevent the progression of atherosclerosis and provides one potential mechanism underlying the beneficial effect of metformin in reducing cardiovascular morbidity and mortality [30]. We propose that the effects of metformin on T cell viability represent an anti-inflammatory action and may be a first barrier to prevent progressive T activation and consequently increased production of cytokines by T cells. However, experiments are being carried out to consider this point.