High concentrations of atorvastatin reduce in‐vitro function of conventional T and regulatory T cells

Summary

Regulatory T cells (T_regs) modulate the magnitude of immune responses and possess therapeutic potential in an array of immune diseases. Statins reduce the activation and proliferation of conventional T cells (T_cons), and they seem to up‐regulate the frequency and function of T_regs. However, there is a lack of simultaneous evaluation of the in‐vitro effect of statins on the functional profile of T_regs versus T_cons. Herein, magnetically purified T_cons and T_regs were stimulated with CD3/CD28/interleukin (IL)‐2 in the presence of atorvastatin (ATV) at 1 or 10 µM. The suppressive function of T_regs, the expression of markers associated with T_reg function, activation levels, cytokine production and calcium flux in both subpopulations were assessed by flow cytometry. ATV had no cytotoxic effect on T cells at the concentrations used. Interestingly, 10 µM ATV hampered the suppressive capacity of T_regs. Moreover, this higher concentration reduced the expression of forkhead box protein 3 (FoxP3), cytotoxic T lymphocyte antigen (CTLA‐4) and programmed death 1 (PD‐1). In T_cons, ATV at 10 µM decreased PD‐1 and CD45RO expression. The expression of CD25, CD69, CD95, CD38, CD62L, CCR7 and perforin was not affected in both subpopulations or at any ATV concentrations. Remarkably, 10 µM ATV increased the percentage of tumour necrosis factor (TNF)‐α‐producing T_regs. Although there was a reduction of calcium flux in T_cons and T_regs, it was only significant in 10 µM ATV‐treated T_cons. These results suggested that 10 µM ATV affects the cellular functions of both populations; however, this concentration particularly affected several aspects of T_reg biology: its suppressive function, cytokine production and expression of T_reg‐specific markers.

Introduction

CD4⁺ regulatory T cells (T_regs) that express the forkhead box protein 3 (FoxP3) transcription factor can control the magnitude of immune responses under different clinical conditions, including autoimmune diseases, transplants, atherosclerosis, allergic diseases and colitis 1, 2, 3, 4, 5. Therefore, T_regs have become a topic of growing interest due to their therapeutic benefit in the modulation of immune responses. Their suppressive action is largely mediated by cell–cell interactions, soluble factor delivery and functions associated with cytolysis and metabolic disruption 6, 7, and these functions can be potentiated or induced by different drugs such as statins. Statins are drugs with potent immunomodulatory actions that inhibit the 3‐hydroxy‐3‐methylglutaryl‐CoA reductase (HMGCR) that converts HMGC into mevalonate, a precursor of the biosynthesis of cholesterol and non‐steroidal isoprenoid compounds. The immunomodulatory actions of statins have been associated mainly with a reduction of isoprenoid compounds that are essential for the post‐translational modifications of proteins participating in several cellular processes, such as survival, proliferation, differentiation and migration 8.

Natural and synthetic statins reduce the level of cellular activation as determined by the expression of CD69 and CD25 molecules, the proliferation of conventional T cells (T_cons) and cell cycle arrest 9, 10, 11, 12, 13, 14. Contradictory results regarding the effect of statins on the production of interferon (IFN)‐γ, tumour necrosis factor (TNF)‐α, interleukin (IL)‐2, IL‐4 and IL‐5 by T_cons have been associated with the type of statins, incubation time in culture and/or the pre‐existing inflammatory condition in which they were evaluated 10, 11, 12, 13, 14. Moreover, it is unclear if statins can modify early events downstream of T cell receptor (TCR) activation, such as calcium (Ca²⁺) mobilization 9, 13. Interestingly, statins seem to expand the T_reg population under both steady‐state and inflammatory conditions as well as modulate their phenotypes, suppressive functions and migration patterns both in vivo and in vitro 15, 16, 17, 18, 19, 20, 21. The statin‐induced T_reg expansion is primarily attributed to the conversion of forkhead box protein 3 (FoxP3)^– T cells into FoxP3⁺ T cells through several mechanisms: (i) by negatively modulating the expression of mothers against decapentaplegic homologue 6 (Smad6) and Smad7 proteins responsible for inhibiting the transforming growth factor‐β (TGF‐β) signalling pathway 22; (ii) by reducing the phosphorylation of protein kinase B (PKB), mammalian target of rapamycin (mTOR) and extracellular signal‐regulated kinases, which are involved in T_reg induction 15; (iii) by increasing the expression of TGF‐β and IL‐10 23 and reducing the IL‐6/signal transducer and activator of transcription 3 (STAT‐3) signalling pathway 24, promoting a higher T_reg/T helper type 17 (Th)17 ratio; and (iv) by inducing the transcription factor Krüppel‐like factor 2 (KLF2), which promotes T_reg accumulation in secondary lymphoid organs 25, 26.

Atorvastatin (ATV) is a second‐generation statin that reduces blood levels of low‐density lipoprotein cholesterol, even at very low therapeutic doses, compared with other available statins 27. Moreover, because ATV up‐regulates the T_reg population, we wanted to evaluate whether ATV can differentially modulate several aspects of the biology of T_regs and T_cons, such as their activation status, suppressive functions, cytokine production and Ca²⁺ mobilization in a dose‐dependent manner. Thus, we determined the effector function and Ca²⁺ efflux of T_cons and T_regs isolated by electromagnetic sorting and activated in the absence or presence of either 1 or 10 µM ATV.

Materials and methods

Isolation of T cell populations and cell culture

Peripheral blood mononuclear cells (PBMCs) were isolated by gradient centrifugation using Histopaque‐1077 (Sigma‐Aldrich, St Louis, MO, USA) from healthy individuals recruited at the Universidad de Antioquia, Colombia. Informed consent was obtained from all donors and the research was approved by the Bioethical Board for Human Research from the Universidad de Antioquia. Fresh untouched CD4⁺ T cells were purified by negative selection with a CD4⁺ T Cell Isolation Kit II (Miltenyi Biotec, Bergisch Gladbach, Germany) followed by staining with anti‐CD25‐phycoerythrin‐cyanin 5 (PE‐Cy5) (clone M‐A251; BD Biosciences, San Jose, CA, USA) and anti‐CD127‐PE (clone eBioRDR5; eBioscience, San Diego, CA, USA) monoclonal antibodies. CD25^highCD127^low/– T_reg populations and CD25^–CD127⁺ T_cons were isolated in a MOFLO XDP high‐speed cell sorter using the single mode (Beckman Coulter, Fort Collins, CO, USA). The purity of the sorted cell populations ranged from 90 to 98%. The efficiency of this procedure was approximately 90%.

The isolated T cells were seed in U‐bottomed 96‐well plates to a density of 0·1–2 × 10⁵ cells/well and cultured in RPMI‐1640 medium (Sigma‐Aldrich) supplemented with 10% heat‐inactivated fetal bovine serum (FBS). Cultures were incubated at 37˚C in 5% CO₂ and 95% relative humidity. Cells were stimulated for 48 h with 2·5 µg/ml plate‐bound anti‐CD3 monoclonal antibody (mAb) (clone OKT3; Ancell, Bayport, MN, USA) and 1·5 µg/ml soluble CD28 (clone ANC28.1/5D10; Ancell) in the presence of 20 IU/ml of human recombinant IL‐2 (National Institutes of Health AIDS Research and Reference Reagent Program, Bethesda, MD, USA). Stimulus conditions were set at 48 h as T_reg cells acquire the highest suppressive capacity at this time, as has been reported by other authors 28.

Atorvastatin treatment

ATV was provided by Biogen Laboratory, Bogotá, Colombia. It was diluted in dimethyl sulphoxide (DMSO; Carlo Erba, Rome, Italy) at a stock concentration of 100 mM. ATV was added to cultures for 48 h at different concentrations (0·5, 1, 2, 5 or 10 μM) that have been widely used to evaluate statin effects on several cellular subpopulations 12, 21. The highest DMSO concentration used in the experiments was 0·01%, which is below the concentrations causing any toxic effects on the cultures, as has been reported previously 29. Thus, the T cell subpopulations (T_con and T_reg) were cultured in the following conditions: unstimulated (or so‐called ‘basal condition’), stimulated and stimulated and treated with different ATV concentrations.

Viability assay

To assess viability, CD4⁺ T cells were harvested after 48 h incubation under the previously mentioned conditions and washed twice with phosphate buffered saline (PBS). They were then stained with 700 nM 3,3‐dihexyloxacarbocyanine iodide (DIOC‐6; Invitrogen Life Technologies, Paisley, UK) and 1 µg/mL 7‐aminoactinomycin D (7‐AAD; Invitrogen) for 20 min at room temperature in the dark. The cells were analysed on a LSR Fortessa flow cytometer (BD Biosciences).

Suppression assay

CD25^–CD127⁺ T cells were labelled with carboxyfluorescein succinimidyl ester (CFSE) at 1·25 μM (Molecular Probes, Eugene, OR, USA). CD25^highCD127^low/– T_regs were treated under the conditions mentioned above for 48 h. Cells were then washed twice with PBS and co‐cultured in 96‐well round‐bottomed plates at a 1:1 ratio in the presence of anti‐CD3/CD28 plus IL‐2 for 72 h. The proliferation of CD4⁺ T cells was detected as the dilution of CFSE by flow cytometry using FlowJo software version 9.7.6 (TreeStar, Inc., Ashland, OR, USA). The percentage of suppression was calculated as follows: [100 − (% proliferation of T_cons in the presence of T_regs/% proliferation of T_cons alone) × 100].

Flow cytometry

The phenotype, activation status and cytokine production of both T_cons and T_regs in the presence or absence of ATV was assessed by flow cytometry using anti‐human mAbs against surface and intracellular molecules (Supporting information, Table S1). In some cultures, the cells were restimulated with 50 ng/ml phorbol myristate acetate (PMA) (Sigma‐Aldrich) plus 1 μg/ml ionomycin (Sigma‐Aldrich) in the presence of 10 μg/ml brefeldin A (Sigma‐Aldrich) for 5 h to detect cytokines. Antibodies against surface molecules were incubated for 20 min at 4°C. For intracellular staining the cells were fixed/permeabilized according to the manufacturer’s instructions using the FoxP3 staining kit (eBiosciences) and incubated with specific antibodies for 30 min. The samples were acquired on an LSR Fortessa flow cytometer (Becton‐Dickinson, Franklin Lakes, NJ, USA) with FACSDiva software version 6.1.3 (BD Biosciences) and the data were analysed using FlowJo software version 9.7.6 (TreeStar). In the design of initial experiments, fluorescence minus one controls were used to distinguish positive from negative populations and set the gates. Relevant biological controls, such as unstimulated and stimulated cells, were used to determine the positive populations and the dynamic of antibody fluorescence in further experiments, as previously suggested 30. Doublets and dead cells were excluded based on the forward/side light‐scatter profiles and the absence of fluorescence in the live/dead viability staining (Thermo Fisher Scientific, Karlsruhe, Germany).

Measurement of intracellular Ca²⁺ concentration by flow cytometry

T_con and T_reg cells were stimulated with (anti‐CD3/anti‐CD28/IL‐2) in the presence or absence of ATV for 48 h. The cells were then washed with PBS and ~0·6 × 10⁶ cells were loaded with 4 μg/mL Fluo 3‐AM (Invitrogen) and incubated for 60 min at 30ºC in the presence of 0·02% pluronic F‐127 in Hanks’s buffered saline solution (HBSS) containing 1 mM CaCl₂, 1 mM MgCl₂ and 1% FBS. The loaded cells were washed and resuspended in PBS. Anti‐CD3 at 5 μg/ml was added to the loaded cells and baseline fluorescence was recorded during the first 60 s using a FACSCanto II flow cytometer; cross‐linking of CD3 was then performed by adding 10 μg/ml of mouse immunoglobulin (Ig)G and the resulting calcium release was recorded for approximately 15 min. The parameters of the kinetics of calcium fluxes were the areas under the curve (AUC), the maximum values (Max), the values of the slopes and the time required to reach the maximum value (T_max) of Ca²⁺ were calculated through FlowJo software version 9.7.6 (TreeStar).

Statistical analysis

The results are expressed as mean ± standard error of the mean (s.e.m.). Normal distribution was assessed by Shapiro–Wilk normality test and comparisons between the groups were performed using one‐way analysis of variance (anova) or general linear model (GLM) anova for repeated measurements, followed by Dunnett’s post‐hoc test. Correlations between groups were determined by Spearman’s rank correlation. Statistical analyses were performed with Prism version 8.0 (GraphPad Software, La Jolla, CA, USA). P‐values < 0·05 were considered statistically significant.

Results

High ATV concentration blocks T_reg suppressive function

To determine the effect of ATV on the suppressive function of T_regs, we evaluated the proliferation of T_cons cultured with previously preactivated T_regs using anti‐CD3/anti‐CD28/IL‐2 for 48 h in the presence or absence of different concentrations of ATV. While suppressive function was maintained at 0·5–1 μM ATV, 2–10 μM ATV blocked suppressive function (Fig. 1a,b). To rule out the possibility that the loss of suppressive activities was due to increased cell kinetics, DIOC6 or annexin‐V staining was performed. We observed that, similarly to lower ATV concentrations, more than 80% of CD4⁺ T cells remained viable at 48 and 72 h of TCR activation in the presence of the highest ATV concentration (10 µM) (Supporting information, Fig. S1a,c). Moreover, both T_cons and T_regs remained viable at 10 µM ATV as visualized by a live/dead fixable stain, suggesting that ATV concentrations used in this study were non‐toxic (Supporting information, Fig. S1b,c).

Figure 1.

Ten µM atorvastatin (ATV) reduces the suppressive function of regulatory T cells (T_regs). T_regs from healthy donors were preactivated with anti‐CD3, anti‐CD28 and interleukin (IL)‐2 in the absence or presence of the indicated ATV concentrations for 48 h. They were then washed and mixed in a 1:1 ratio with conventional T cells (T_cons) labelled with carboxyfluorescein succinimidyl ester (CFSE) and activated for 72 h. The CFSE dilution was measured by flow cytometry. Representative histograms of four independent experiments are shown. (a) Percentage suppression of responder cell proliferation is shown for each condition (n = 4); (b) untouched total CD4⁺ T cells were isolated through magnetic separation from the peripheral blood mononuclear cells (PBMCs) of healthy donors and cultured in the conditions mentioned above. Viable cells were determined as (7‐aminoactinomycin D (7‐AAD)^– and 3,3‐dihexyloxacarbocyanine iodide (DIOC)‐6⁺ cells (n = 3 donors) and (c) statistical significance determined by one‐way analysis of variance (anova) and Dunnett’s post‐hoc test (b: co‐cultures +2 μm ATV *P = 0·039, co‐cultures +5 μm ATV *P = 0·022 and co‐cultures +10 μm ATV *P = 0·029 compared with untreated co‐cultures).

High ATV concentrations decrease FoxP3, CTLA‐4 and programmed death 1 (PD‐1) expression on T_regs

We then examined if ATV could modulate the expression of molecules associated with T_reg function. ATV at the highest concentration reduced the mean fluorescence intensity (MFI) of FoxP3 and CTLA‐4 to 42 and 20%, respectively, compared with stimulated cultures without ATV treatment (P = 0·018 and P = 0·033, respectively, Fig. 2a,b). Expression of PD1 was also reduced in the presence of 10 μM ATV by 38% compared with control cultures (P = 0·018, Fig. 2c).

Figure 2.

Ten μM atorvastatin (ATV) reduces the mean fluorescence intensity (MFI) of forkhead box protein 3 (FoxP3), cytotoxic T lymphocyte antigen (CTLA‐4) and PD‐1 on regulatory T cells (T_regs). Histogram of representative flow cytometry analysis and MFI analysis of FoxP3 (n = 5 donors) (a), CTLA‐4 (n = 8 donors) (b), programmed death 1 (PD‐1) (n = 8 donors) (c) and CD25 (n = 4 donors) (d) on conventional T cells (T_cons) and T_regs in basal conditions and after activation with anti‐CD3, anti‐CD28 and interleukin (IL)‐2 in the absence or presence of the indicated ATV concentrations for 48 h. Statistical analyses were performed using the general linear model (GLM) analysis of variance (anova) and Dunnett’s post‐hoc tests. Means and standard errors of the mean (s.e.m.) are indicated. *P < 0·05, **P < 0·01 and ***P < 0·001. Each dot represents one individual in the graph.

On T_cons, ATV treatment did not change FoxP3 expression (Fig. 2a) or CTLA‐4 expression patterns compared with stimulated cultures without treatment (Fig. 2b). However, 10 µM ATV significantly reduced expression of PD‐1 by 40% compared with cultures without treatment (P = 0·025, Fig. 2c). No significant differences were observed in CD25 and perforin expression in T_regs and T_cons after ATV treatment (Fig. 2d and Supporting information, Fig. S2).

Ten μM ATV reduced CD45RO expression in T_cons but did not change homing and activation markers on cell subpopulations

As ATV can alter the activation status of different cell subpopulations, we then determined how ATV affects the expression of several surface molecules on both subpopulations. Remarkably, only the highest ATV concentration decreased the expression of CD45RO on T_cons by 17% compared with untreated stimulated cultures (P = 0·037 Fig. 3a). In contrast, CD95, CD69 and CD38 expression was not modified by any ATV concentration on T_cons or T_regs compared with control cultures (Fig. 3b–d). In addition, we wanted to determine whether ATV changed the circulation pattern of these cell subsets. However, no significant differences in CD62L, CCR7 and CCR5 expression were observed after treatment with statins in both subpopulations (Supporting information, Fig. S3a–c).

Figure 3.

Ten μM atorvastatin (ATV) reduced CD45RO expression only in conventional T cells (T_cons), but did not change activation markers on cell subpopulations. Representative flow cytometry histogram and mean fluorescence intensity (MFI) analysis of CD45RO (n = 5 donors) (a), CD95 (n = 3 donors) (b), CD69 (n = 8 donors) (c) and CD38 (n = 5 donors) molecules (d) in T_cons and regulatory T cells (T_regs) under basal conditions and after activation with anti‐CD3, anti‐CD28 and interleukin (IL)‐2 in the absence or presence of the indicated ATV concentrations for 48 h. Statistical analyses were performed using general linear model (GLM) analysis of variance (anova) and Dunnett’s post‐hoc tests. Means and standard errors of the mean (s.e.m.) are presented. *P < 0·05, **P < 0·01 and ***P < 0·001. Each dot represents one individual in the graph.

T_regs treated with 10 µM ATV produce more TNF‐α

Functional plasticity is a feature of whole T cell subpopulations, including T_regs that can gain the inflammatory functions of Th1, Th2 or Th17 cells influenced by the cytokine environment. Unlike T_cons, we observed a lower frequency of T_regs producing IL‐2, IFN‐γ and TNF‐α cytokines after anti‐CD3/anti‐CD28 stimulation. The presence of 10 μM ATV reduced the frequency of IFN‐γ‐producing T_regs to approximately 76% compared with untreated stimulated cultures (P = 0·033, Fig. 4a). Interestingly, the high concentration of ATV increased the percentage of TNF‐α‐producing T_regs from 12 to 25% (P = 0·026, Fig. 4b). Interestingly, we observed a positive correlation between TNF‐α‐producing T_regs and the MFI of CTLA‐4 on T_regs (r = 0·553, P = 0·007, Supporting information, Fig. S4a) but not with the MFIs of FoxP3 or PD‐1 on T_regs (Supporting information, Fig. S4b,c). On the contrary, ATV treatment did not modify the production of IL‐2, IL‐10 and IL‐4 in T_cons and T_regs (Fig. 4c,d and Supporting information, Fig. S5).

Figure 4.

Regulatory T cells (T_regs) treated with 10 µM atorvastatin (ATV) produce more tumour necrosis factor (TNF)‐α than untreated‐T_regs. Conventional T cells (T_cons) and T_regs were isolated by cell sorting and stimulated with anti‐CD3, anti‐CD28 and IL‐2 in the absence or presence of the indicated ATV concentrations for 48 h. Cells were restimulated with phorbol myristate acetate (PMA)/ionomycin in the presence of brefeldin A for 5 h. Then, they were stained with monoclonal antibodies to detect the percentage of cells expressing interferon (IFN)‐γ (n = 7 donors) (a), cells expressing TNF‐α (n = 7 donors) (b), cells expressing interleukin (IL)‐2 (n = 7 donors) (c) and cells expressing IL‐10 (n = 6 donors) (d) by flow cytometry. Statistical analyses of percentages were performed using general linear model (GLM) analysis of variance (anova) and Dunnett’s post‐hoc tests. Means and standard errors of the mean (s.e.m.) are indicated. *P < 0·05, **P < 0·01 and ****P < 0·0001. Left panels show representative flow cytometry histograms, and each dot on the right graphs represents one individual.

Higher ATV concentrations impaired Ca²⁺ flux in T_cons

To determine the influence of ATV on early events downstream of TCR activation, we evaluated Ca²⁺ mobilization elicited by CD3 on both cell populations by flow cytometry (Fig. 5a). In general, in response to CD3, we observed a higher Ca²⁺ mobilization in T_cons compared with T_regs, corroborating previous work 31. When the T_cons were treated with ATV, there was a marked reduction in the AUC from 161 133 (without ATV) to 135 581 (P = 0·606) and to 105 447 (P = 0·020) with 1 µM and 10 µM ATV, respectively. In the case of T_regs, the effect of ATV was less evident. ATV induced a negligible increase in the AUC from 121 108 to 124 604 and 126 099 at 1  and 10 µM ATV, respectively, but there was also a delay in the times to achieve the maximal peaks (Fig. 5b). The effect of ATV on T_cons was also evident, probably being dose‐dependent regarding the maximum values (Max), as ATV decreased these values from 381 to 260 (P = 0·750) and 175 (P = 0·011) at 1 and 10 µM ATV, respectively (Fig. 5a,b).

Figure 5.

Ten μM atorvastatin (ATV) impairs Ca²⁺ flux in conventional T cells (T_cons). Representative histograms of Ca²⁺ flux in T_cons and regulatory T cells (T_regs) isolated and activated with anti‐CD3, anti‐CD28 and interleukin (IL)‐2 in the presence or absence of different ATV concentrations for 48 h. Cross‐linking of the CD3 monoclonal antibody (mAb) was performed through immunoglobulin (Ig)G (a). Area under the curve (AUC), Max value, slope and T_max (b). Statistical analyses were performed using the general linear model (GLM) analysis of variance (anova) and Dunnett’s post‐hoc tests. Means and standard errors of the mean (s.e.m.) are indicated (n = 3 donors); *P < 0·05.

Discussion

In recent years, statins have been proposed as positive regulators of T_regs. Here we found that pretreatment of T_regs with 0·5–1 μM ATV did not modify suppressive function, and while at the highest ATV concentrations (2–10 μM), T_regs lost their suppressive function and became TNF‐α‐producing T_regs.

Early reports suggested that statins up‐regulate T_reg suppressive function 15, 23, 32; however, subsequent studies demonstrated contradictory results that could be due to differences in assays or other present technical issues. For instance, in several in‐vitro experiments, regardless of the concentration used, statins were added directly into T_reg/T_con co‐cultures 16; this could lead to an over‐estimation of the T_reg suppressive function because of statin‐induced anti‐proliferative effects on T_cons 14, an event that we also witnessed (Supporting information, Fig. S6). Although statins can also affect the proliferative response of T_regs 17, 33, further studies are necessary to determine if T_regs are less susceptible to statin‐induced anti‐proliferative activities. In other studies, human T_regs were obtained as CD4⁺CD25⁺ T cells, which might not represent true T_regs; therefore, activated T_cons could proliferate in the presence of strong stimulation, masking the suppression activity of T_regs 15, 34. This could be problematic, especially when the proliferative response is evaluated through [³H]‐thymidine incorporation, which does not particularly distinguish the proliferating cell population 35. Similarly, contradictory results could be derived from the type of statin used, considering the lipophilic or hydrophilic nature of these drugs, that could facilitate or block its spread across the cell membrane 36. Moreover, others have demonstrated higher T_reg suppressive capacity in the presence of high concentrations of statins in vitro (25 µM) 32; however, these concentrations could affect also cell viability 37. Although we used supratherapeutic ATV concentrations (1 μM), this did not affect T_reg suppressive function, which was similar to the effect of 0·001 μM ATV 37, a similar concentration to that found in individuals taking low‐dose statins 38. Similar to our findings, pretreatment with simvastatin at 2 μM or ATV or lovastatin at 5 μM impaired T_reg suppressive capability in mice, as evaluated by the [³H]‐thymidine assay 33, highlighting the relevance of the mevalonate pathway in co‐ordinating T_reg functions.

We found that sorted T_regs showed an activated phenotype, as has been reported previously, with higher expression of FoxP3, CD25, and PD‐1 compared with their T_con counterparts. However, this feature was lost when they were cultured with the highest ATV concentration as a reduction of these molecules was observed. In line with these findings, previous reports have observed a down‐regulation of functional markers such as CTLA‐4, the inducible co‐stimulator (ICOS) and the glucocorticoid‐induced TNFR‐related protein (GITR) on murine T_regs pretreated with ATV, lovastatin or simvastatin at concentrations ≥ 2 μM 33, 39 as well as FoxP3 expression on CD4⁺CD25⁺ cells treated in vitro with 10 µM ATV 37, highlighting the relevance of the mevalonate pathway co‐ordinating T_reg functions. Accordingly, the loss of FoxP3 in response to 10 μM ATV could be associated with reduced expression of CTLA‐4 and PD‐1 and with the inhibition of T_reg suppressive function, as it has been previously suggested 40, 41. Interestingly, these effects were partially observed on T_cons, as only the expression of PD‐1 along with CD45RO was reduced by 10 μΜ ATV. Moreover, activation and homing markers such as CD25, CD69, CD95, CD38, CD62L and CCR7 were not modified by ATV on T_cons or T_regs. Our results are partially supported by others; for instance, CD45RO expression has been previously shown to be reduced after ATV treatment 11; however, several authors 9, 11, 12 have demonstrated a reduction in CD25 expression after ATV treatment, whereas it remained unchanged in our experiments. Similarly, the ATV effects on CD69 expression has been controversial 9, 11, 12, possibly because to differences in stimuli, activation time and experimental conditions. In general, these findings contrasted with our previous results that demonstrated higher expression levels of functional markers on T_regs in individuals on temporary treatment with statins 17. Such a difference could be explained because therapeutic doses are not equivalent to the doses used in vitro or is probably because of transitory regulation. As our study was limited to in‐vitro assays, these results could not be directly extrapolated to individuals on statin therapy. Thus, more studies are necessary to understand the dynamics of T_con and T_reg activation through clinical trials that allow evaluating the effects of long‐term statin use on these cell subsets.

In addition to phenotypical changes, ATV induced alterations on the Ca²⁺ influx, as 10 μM ATV reduced maximal peak (Max) and total mobilized Ca²⁺ (AUC) elicited by CD3 on T_cons. Although both ATV concentrations delayed the time required to reach the peak Ca²⁺ concentration (slope and T_max) on T_cons, there was no statistical significance. Altogether, these results suggested that T_cons are more susceptible than T_regs to the effect of ATV. Probably, this lower effect on T_regs could be due to their TCR‐induced attenuated Ca²⁺ mobilization compared with T_cons 31. Furthermore, this lower Ca²⁺ mobilization on T_cons could be related to lower proliferation. It is noteworthy to mention that although 10 μM ATV reduced Ca²⁺ mobilization, it did not affect cytokine production by T_cons as Ca²⁺ activates nuclear factor of activated T cells (NFAT), which is essential to cytokine production 42. In contrast, ATV‐treated T_regs become TNF‐α‐producing cells, which could be attributed to lower FoxP3 expression that, in turn, down‐regulated CTLA‐4 expression, allowing for inflammatory phenotypes by producing Th1 cytokines 43, 44.

Taken together, our results suggested that 10 μM ATV disturbs the function of T_regs, diminishing their ‘regulator’ phenotype via decreasing the expression of surface molecules and suppressive function and inducing the acquisition of an inflammatory profile (TNF‐α‐producing T_regs). This concentration also affected T_con function by reducing their proliferation, delaying Ca²⁺ mobilization and loss of PD‐1 expression, suggesting that ATV differentially affected the function of both cell populations. Moreover, our findings may explain why some individuals taking statins over a prolonged time at high doses are at higher risk to develop autoimmune diseases 45. Although our results are interesting, the low sample size in our study does not allow us to generalize our conclusions, and thus more studies are needed to determine the effect of low concentrations of ATV that simulate the plasmatic concentrations reached in individuals taking low therapeutic doses of statins on T_reg function and the role of the mevalonate pathway on the biology of T_regs.

Disclosures

The authors declare that there are no conflicts of interest regarding the publication of this paper.

Supporting information

Fig. S1. Untouched total CD4⁺ T cells were isolated through magnetic separation from PBMCs of healthy donors and cultured during 48 h under TCR activation as it was mentioned on material and methods. Viable cells were determined as 7‐AAD‐ and DIOC‐6+ cells (n = 3 donors) (a). Representative flow cytometry plot showing live and dead isolated T cells (Tcon and Treg) by using Live/dead fixable cell stain (b). Percentage analysis of live cells in subsets of T con and Treg (c). No statistical significant differences were found (n = 4).

Fig. S2. Mean fluorescence intensity (MFI) analysis of perforin (n = 3 donors) on Tcons and Tregs in basal conditions and after activation with anti‐CD3, anti‐CD28 and IL‐2 in the absence or presence of the indicated concentrations of ATV for 48 h. No significant differences were found between groups.

Fig. S3. Representative flow cytometry histogram and MFI analysis of CD62L (n = 8 donors) (a), CCR7 molecules (n = 3 donors) (b), and CCR5 (n = 4 donors) in Tcons and Tregs in basal conditions and after activation with anti‐CD3, anti‐CD28 and IL‐2 in the absence or presence of the indicated concentrations of ATV for 48 h. Statistical analyses were performed using the GLM ANOVA, Dunnett’s post‐hoc tests. Mean and SEM. ****P < 0·0001, ***P < 0·0002. Each dot represents one individual in the graph.

Fig. S4. Correlation of TNF‐a‐producing Tregs percentage with CTLA‐4 MFI (a), FoxP3 MFI (b), and PD1 MFI (c) in Treg cells. Spearman’s rank correlation coefficients (r) and p values (P) are indicated.

Fig. S5. Tcons and Tregs were isolated by cell sorting and stimulated with anti‐CD3, anti‐CD28 and IL‐2 in the absence or presence of the indicated concentrations of ATV for 48 h. Cells were re‐stimulated with PMA ionomycin in the presence of brefeldin A for 5 h. Then, they were stained with monoclonal antibodies to detect the percentage of cells expressing IL‐4 (n = 7 donors). No significant differences were found between groups.

Fig. S6. Untouched total CD4+ T cells were isolated through magnetic separation fromÐBMCs of healthy donors and they were labeled with CFSE at 1·25 μM. Cells were washed twice with PBS and put in co‐cultured in 96‐well round‐bottom plates at a ratio 1:1 in the presence of CD3/CD28/IL‐2 for 72 hours and different atorvastatin (ATV) concentrations. Proliferation of CD4⁺ T cells was detected as the dilution of CFSE on flow cytometry and Index division (ID) was calculated using FlowJo software. (n = 3 donors). Statistical analyses were performed using One‐Way ANOVA, Dunnett's post hoc test. Mean and SEM. *P = 0·011, **P = 0·001.

Table S1. Antibodies used for flow cytometry.