Flow cytometric determination of glucocorticoid receptor (GCR) expression in lymphocyte subpopulations: lower quantity of GCR in patients with post-traumatic stress disorder (PTSD)

Abstract

Assessment of the intracellular glucocorticoid receptor (GCR) level may be useful in monitoring functional disturbances of the hypothalamic-pituitary-adrenocortical axis or effects of prolonged steroid therapy. Cytosolic ligand binding assays have recently been supplemented by flow cytometric determination of receptor expression in individual cells. A method based on multiparametric analysis of whole blood by simultaneous labelling of intracellular GCRs and surface markers of lymphocyte subsets is described. We examined 25 healthy male volunteers and 35 age- and sex-matched post-traumatic stress disorder (PTSD) patients within 8 years from traumatic event. PTSD patients had a lower relative quantity of GCR in all lymphocyte populations tested as compared with healthy volunteers. NK cells of both groups showed higher expression of GCR than other lymphocyte subsets. In PTSD patients, the expression of GCR in B lymphocytes was also higher than in T cell. Although serum cortisol level was lower in PTSD patients, there was no correlation between cortisol level and GCR expression. Multiparameter flow cytometric determination of GCR expression in lymphocyte subpopulations may provide a useful tool for monitoring immunoregulatory action of glucocorticoids.

Introduction

Glucocorticoid (GC) hormones have been recognized as potent immunomodulators with an impact on the immune cell development, traffic and functions [1–4]. The activity of GCs is mediated through their interaction with specific intracellular glucocorticoid receptors (GCR). GCRs are members of the steroid hormone receptor superfamily known as transcription factors. They are located in the cell cytoplasm in a monomeric, high affinity form associated with heat-shock proteins (hsp 90, hsp 75, hsp 56) [5]. Upon ligand binding, GCRs undergo conformational changes, dissociate from the heat-shock proteins, dimerize and, in the active form as the ligand-GCR complex, migrate to the cell nucleus where they bind to the DNA nucleotide sequence known as glucocorticoid responsive elements (GRE) of target genes [6–8]. Furthermore, the activated GCR-ligand complex has been shown to bind to other transcription factors such as AP-1 and NF-κB, down-regulating their transcriptional activity and inflammatory response [9,10]. In this way, GCs influence different parts of the genome by increasing (transactivation) or decreasing (transrepression) transcription of the respective genes into proteins.

The GCRs are ubiquitously present in almost all nucleated cells [3], and their cellular level is controlled through a negative feedback mechanism, as an elevation of GCs down-regulates GCRs in various organs, including the brain and lymphocytes [11,12]. Assessment of the level of GCRs in the cells may be important and useful in monitoring the receptor content after prolonged steroid therapy [13], or in defining functional disturbances of the hypothalamic-pituitary-adrenocortical (HPA) axis [11,12].

A method generally used for the detection of GCRs is the cytosolic radioligand binding assay. It is a quantitative but time-consuming procedure that requires a large number of cells, usually poorly characterized. Recently, a monoclonal antibody against human GCR (anti150–176 amino acid sequence in the regulatory part of GCR) has been developed, allowing for determination of GCR by flow cytometry at the single-cell level [14]. The method is semiquantitative, but importantly, this highly specific and sensitive method allows for GCR detection in well-defined cell populations. Its comparison with conventional ligand binding assay on T cell lines with a known amount of GCR revealed analogous results [1]. In this study, we used a method of GCR monitoring by multiparametric analysis of whole blood with simultaneous labelling of surface markers characteristic of lymphocyte subsets. The procedure was used in healthy subjects and in patients with post-traumatic stress disorder (PTSD). PTSD is a condition that may occur in response to extremely stressful life events, and is accompanied by psychiatric symptoms [15] and clear biological changes such as HPA axis disturbances, changes in lymphocyte GCR number [12], and changes in several immunologic parameters [16–18]. The highest GCR expression was observed in NK cells of both groups, although PTSD patients had a lower relative quantity of GCR in all three main lymphocyte subpopulations (T, B and NK cells).

Subjects And Methods

Subjects

The study included 25 healthy male volunteers (mean age 35 ± 11·7 years) and 35 age- and sex-matched PTSD patients (mean age 37 ± 8·8 years; t = 0·7, P= 0·5). The patients were Croatian combat veterans examined 2 to 8 (median = 6) years after traumatic events, while controls had no combat or any other traumatic experience. The ethics committee of the Vrapče Psychiatric Hospital approved the research and all subjects provided a written informed consent after the study design had been fully explained to them and prior to undergoing psychiatric evaluation and blood collection. The Clinician Administered PTSD Scale (CAPS) was used by experienced clinicians to make the diagnosis of PTSD [19]. All participants had been free from any psychotropic or hormonal medication, drug or alcohol abuse for at least one month, and did not suffer from any infectious, allergic or endocrine disorder.

Sample collection

Blood samples were obtained by venipuncture at 8–9 a.m. into vacuum tubes (Becton Dickinson Vacutainer System Europe, Grenoble, France). Heparinized peripheral blood for immunophenotyping was processed immediately. The sera for cortisol determination were isolated by centrifugation after clotting of unheparinized blood samples and stored at −70°C until assayed.

Simultaneous immunophenotyping and intracellular GCR determination

Whole blood was used for surface immunophenotyping and intracellular glucocorticoid receptor determination. Direct three-colour staining method [20] was modified to simultaneously label surface markers of lymphocyte subpopulations and their cytoplasmic GCRs [13,14]. The following panel of monoclonal antibodies (MoAb) were used: fluoresceinisothiocyanate (FITC) conjugated anti-GCR (IgG₁, clone no. 5E4-B1), described in [14]; phycoerythrin (PE) conjugated anti-CD3, anti-CD16, anti-CD56; peridinin chlorophyll protein (PerCP) conjugated anti-CD20 MoAbs, and mouse isotype control antibodies (Becton Dickinson, Heidelberg, Germany).

The surface staining was performed by incubating 50 µl of whole blood with 5 µl of particular MoAbs for 15 min at room temperature in the dark. Cells were washed with the staining buffer (PBS containing 1% FCS and 0·1% NaN₃) and fixed in 100 µl of 4% paraformaldehyde in PBS (fixation buffer) for 20 min at 4°C. After one more washing, the erythrocytes were lysed for 15 min with 1 ml of 10 × diluted lysing solution (Becton Dickinson) in the dark. Cells were washed again, resuspended in 50 µl of permeabilization buffer (0·1% saponin 10% FCS and 0·1% NaN₃ in PBS) containing a predetermined optimal concentration (2·66 µg/ml) of anti-GCR MoAb or 5 µl of isotype control, and incubated at 4°C for 20 min in the dark. After extensive washing with permeabilization buffer to remove any unbound antibodies from the cytoplasm, the cell pellets were resuspended in 500 µl of fixation buffer. Cell samples were run on a FACSCalibur flow cytometer (Becton Dickinson) and analysed by CELLQuest software. At least 5000 events in the light-scatter (FSC/SSC) lymphocyte region were acquired. Lymphocyte populations were identified and gated on FITC versus PE or PerCP plots. The FITC-fluorescence intensities of GCR-labelled lymphocyte populations and isotype controls were displayed and determined as mean channel values on a four-dacade log scale in histogram plots (Fig. 1). The relative quantity of GCR (mean GCR fluorescence), expressed as mean fluorescence intensity (m.f.i.), was calculated as the difference between mean values of GCR and isotype control labelled samples.

Fig. 1.

Representative flow cytometric analysis of GCR expressing CD3, CD20, and CD16,56 cells from lymphocyte region (R1 in light-scatter dot plot, not shown) in a healthy person (a) and PTSD patient (b). The regions (R2-R4) were set around double positive lymphocyte populations that were subsequently gated as cells in respective regions and R1. Fluorescence intensities of GCR in three lymphocyte subpopulations (——) and isotype control (——) are compared by overlaying FITC histograms. The highest fluorescence intensity is observed in CD16,56 cells (——). In PTSD patient a higher intensity of isotype control is observed.

The instrument calibration was performed daily by FACSComp software using CaliBRITE™ 3 beads.

Serum cortisol determination

Serum cortisol levels were determined by radioimmunoassay using commercially available AMERLEX kit (Ortho-Clinical Diagnostics, Neckargemünd, Germany). All samples were analysed in duplicate following the manufacturer's protocol. The sensitivity of the assay is 0·1 µg/100 ml, and the intra-assay and interassay coefficients of variation are less than 5% and 8%, respectively.

Statistical analysis

The normality of distribution was confirmed by the Shapiro-Wilk's W-test, and homogeneity of variances by the Levene's test. All data are presented as mean ± s.d. The difference in GCR expression in particular lymphocyte subpopulations between the control and PTSD group was analysed by one-way anova. The differences in GCR expression and fluorescence intensity of isotype control, as well as in the cortisol level and GCR expression in total lymphocytes between the two groups were analysed by manova. Post hoc testing was performed by using the Tukey's honest significance difference for unequal n. Pearson's correlations were used to assess the association between the cortisol level and GCR expression in total lymphocytes and lymphocyte subpopulations. All analyses were performed with STATISTICA v 6 (StatSoft, Inc., Tulsa, OK, USA).

Results

GCR expression

By modifying the original direct staining method for the measurement of intracellular GCR [14] to simultaneous surface immunophenotyping of whole blood, we determined relative expression of GCR in lymphocyte populations in healthy volunteers and PTSD patients. The results are summarized in Fig. 2. In both groups of subjects GCR was unevenly expressed in the three analysed populations (F_2,72 = 3·99, P < 0·05 for controls, and F_2·102 = 14·9, P < 0·0005 for patients). Post hoc testing revealed that it was higher in NK (CD16,56) cells than in the other lymphocyte populations. While in healthy volunteers the GCR expression in T (CD3) and B (CD20) lymphocytes was similar (45·3 ± 10·1 and 45·6 ± 10·6, respectively), in PTSD patients CD20 lymphocytes (32·0 ± 8·0) had a higher content of GCR than CD3 lymphocytes (26·5 ± 6·8), but still lower than NK cells (36·7 ± 8·5). Moreover, in PTSD patients the relative quantity of GCR was lower than in healthy volunteers (Wilk's Λ= 0·41, F_3,56 = 26·9, P << 0·0001). This held true for all lymphocyte subpopulations tested. For T lymphocytes it was by 40% lower, and for B and NK cells by 30% lower on an average in the patient group as compared with controls. An unexpected finding was a higher (F_1,58 = 85, P < 0·0001) fluorescence intensity of isotype control in PTSD patients (54·2 ± 9·4) than in the control group (32·2 ± 8·6), as shown in Fig. 1 for two representative cases.

Fig. 2.

Mean GCR fluorescence intensity (m.f.i.) in lymphocyte populations expressed as the difference between the mean channel values of GCR and isotype control for healthy persons (○) and PTSD patients (•). Circles indicate mean, boxes standard deviation, and whiskers represent range. The indicated significance pertain to post hoc probabilities (Tukey HSD for unequal N) following one-way anova for differences between lymphocyte populations within each group, or manova for differences between groups (*P < 0·05, **P < 0·001, ***P < 0·0005).

Serum cortisol level

Serum cortisol level was determined in a subset of 19 healthy controls and 28 PTSD patients. It was significantly (P < 0·03) higher in PTSD subjects (17·2 ± 4·2 versus 14·4 ± 3·0 µg/100 ml; Fig. 3, at the bottom of the panel). There was no correlation between the cortisol level and GCR expression in either total lymphocytes or any of the three subpopulations examined, although, when both groups were included in the analysis, there was an apparent ‘trend’ of a negative relationship between the cortisol level and lymphocyte GCR amount (r = -0.23, P= 0·12). However, at closer inspection of the categorized scatterplot (Fig. 3) it is obvious that this ‘trend’ was merely due to the prominent difference in GCR expression between the two groups (Fig. 3, at the left of the panel) and it does not represent the ‘true’ relationship between the two variables.

Fig. 3.

Relationship between plasma cortisol level and lymphocyte GCR expression. Three regression lines are fitted for: all subjects (——), control subjects (………), and PTSD subjects (——). Pearson's coefficients r and corresponding P-values are presented next to the respective line. Individual results are shown for control group (s), and for PTSD patients (•). Group means for both variables are depicted by correspondingly filled squares, and standard deviations by bars (horizontal denoting cortisol level and vertical GCR expression). The indicated significance pertains to post hoc probabilities (Tukey HSD for unequal N) following manova for differences.

Discussion

In this study, we further improved the recently described flow cytometric analysis of glucocorticoid receptors in peripheral blood mononuclear cells [14], having adjusted it to the use of whole blood and GCR detection in distinct lymphocyte subsets. Our principal finding was the higher expression of GCR in CD16,56⁺ cells than in any other lymphocyte population in healthy volunteers as well as in PTSD patients. Although the CD16,56⁺ population includes a minor subset of MHC-unrestricted cytotoxic lymphocytes (CD3⁺CD16,56⁺), these, as a subset of total CD3 lymphocytes, express a low level of GCR. So, we consider that NK cells (CD3⁻CD16,56⁺), not the former, are the subpopulation with high GCR expression. This might be the reason why NK cell activity is very sensitive to glucocorticoid changes, as observed in stress [16, 17, 21, 22, 23].

In contrast to the elevated lymphocyte GCR expression and low cortisol level found in Vietnam veterans with chronic PTSD decades after the traumatic event [24,25], our PTSD patients, in a more acute stage of the illness (up to 8 years post trauma), had elevated plasma cortisol level and lower GCR expression in each of the lymphocyte populations examined. The reason for such a discrepancy could lie in a considerably shorter duration of allostatic load caused by adaptational response to stress [26] in our PTSD patients. Still, the negative feedback mechanisms controlling cellular level of GCR is operative, as this reversal of hormone and its receptor expression takes place during chronic hyperactivity of HPA-axis. This negative feedback to GCR expression is less pronounced in the effector, NK and B, cells than in T lymphocytes, suggesting that effector cells are more sensitive to glucocorticoids than T cells.

An inverse association but not correlation between the cortisol and lymphocyte GCR levels has been repeatedly reported in various psychiatric disorders [25]. In the present study we confirmed the lack of direct correlation between the two variables. Although the ranges of serum cortisol level in both groups were wide enough (and to a large extent overlapping) to reveal any possible correlation with lymphocyte GCR level, this was not the case. So, this result supports the concept that in addition to glucocorticoids, another indirect modulation takes place in the regulation of GCR expression in lymphoid tissue [25]. The changes in cortisol level reported in PTSD, decreased [24,25] as well as increased (this report), were still within the physiological ranges. Recently, a similar profile of GCR expression in lymphocyte subpopulations was observed in renal transplant patients during and after steroid therapy (pharmacological doses) [13].

Enhanced binding of isotype control found in PTSD patients could be a consequence of higher photomultiplier tube (PMT) voltage adjustment, insufficient removal of all unbound antibodies from the cytoplasm, or higher expression of Fcγ receptors on lymphocytes of PTSD patients. PMT voltage can be disregarded, as higher fluorescence intensity of isotype control was obtained even when the patient and healthy control samples were run on the same day under the same instrument setting. Insufficient removal of unbound antibodies is also improbable, as all samples were processed under the same staining (and washing) procedure. A likely explanation could be an up-regulation of FcγRII (CD32) on B lymphocytes or/and FcγRIII (CD16) on NK cells, but this remains to be confirmed. Škarpa et al. [27] have reported on a higher level of CD16⁺ but not of CD56⁺ cells in PTSD patients, indicating up-regulation of CD16 on CD56⁺ cells.

Thus, the flow cytometric determination of glucocorticoid receptor (GCR) expression in lymphocyte subpopulations may provide a useful procedure for monitoring the immunoregulatory action of glucocorticoids in response to stress. Although it cannot replace the quantitative radioligand binding assay that gives more functional information (e.g. affinity) about the receptor, the flow cytometric method albeit semiquantitative, has an advantage of detecting total (cytoplasmic, ligand-free, and nuclear, hormone-bound) GCR in well defined cell populations.