LPS regulates the expression of glucocorticoid receptor α and β isoforms and induces a selective glucocorticoid resistance in vitro

Background This study was aimed to evaluate the effect of LPS in glucocorticoid receptor (GR) isoforms expression on different cell lines and PBMC from healthy donors in vitro and glucocorticoid sensitivity of PBMC in vitro. Methods U-2 OS cell lines expressing GR isoforms, different cell lines (CEM, RAJI, K562 and HeLa) or PBMC from healthy donors, were cultured or not with LPS. The expression of GRα and GRβ was evaluated by Western blot. Glucocorticoid sensitivity was evaluated in PBMC treated with LPS, testing genes which are transactivated or transrepressed by glucocorticoid. For transactivated genes (MKP1, FKBP5) PBMC were treated with Dexamethasone 100 nM for 6 h. The mRNA expression was measured by RT-PCR. For transrepressed genes (IL-8, GM-CSF), PBMC were cultured in Dexamethasone 100 nM and LPS 10 μg/ml for 6 h and protein expression was measure by ELISA. Results GR isoforms were induced in U-2 OS cells with a greater effect on GRα expression. Both isoforms were also induced in CEM cells with a tendency to a greater effect on GRβ. LPS induced only the expression of GRα in Raji and HeLa cells, and in PBMC, with no effect in K562 cells. LPS induced a loss of glucocorticoid inhibitory effect only on the secretion of GM-CSF. Conclusion LPS in vitro differentially modulates the expression of GR isoforms in a cell specific manner. In PBMC from healthy donors LPS induces an approximately two times increase in the expression of GRα and a loss of the glucocorticoid inhibitory effect on the secretion of GM-CSF, without affecting other glucocorticoid responses evaluated. Electronic supplementary material The online version of this article (10.1186/s12950-017-0169-0) contains supplementary material, which is available to authorized users.


Background
Glucocorticoids (GC) have been used as anti-inflammatory and immunosuppressive agents [1]. However, the sensitivity to GC actions varies among individuals and even in the same patient at different time points. The mechanisms mediating changes in GC sensitivity are not completely understood [2]. A better understanding of the factors that could influence the cell response to GC is relevant to adapt GC therapy to different requirements in patients suffering autoimmune/inflammatory diseases.
The GC actions are mediated through the activation of intracellular glucocorticoid receptors (GR). The genomic structure of the single GR gene described in humans consists of 9 exons. Alternative splicing of exon 9 (exon 9α or 9β) results in the synthesis of 2 homologous mRNAs and protein isoforms, termed GRα and GRβ [3]. After GC binds GRα, it binds as a homodimer to DNA, regulating the expression of the linked gene. GRα can also modulate gene expression by interacting with other transcription factors such as AP-1 and NF-κB, exerting most of the anti-inflammatory actions by this mechanism [4][5][6][7]. GRβ differs from GRα only in its carboxi-terminus sequence. This difference renders to GRβ a role as a dominant negative inhibitor of GRα activity. Therefore, a relative over-expression of the GRβ isoform could play a role in the regulation of target cell sensitivity to GC [8,9]. Indeed, studies have demonstrated that GRβ expression is elevated in patients with chronic inflammatory diseases associated with GC resistance e.g. asthma, inflammatory bowel disease and nasal polyps [10][11][12][13][14].
It has long been proposed that infectious episodes could trigger relapses in autoimmune diseases [15,16]. Our group has demonstrated a transient increased expression of GRβ in peripheral blood mononuclear cells (PBMC) from septic patients. We also showed that septic serum influence GC sensitivity, favoring the expression of GRβ in association with a GC resistant state [17,18]. Number of factors, present in serum from septic patients, such as cytokines, bacterial products and GC, either by themselves or in combination could be responsible for these effects.
Webster et al. [19] showed that TNF-α and IL-1 induced the expression of both GR isoforms mRNA in HeLa and CEM-C7 cell lines, but the increase of the expression of GRβ was greater than GRα in those conditions. Other proinflammatory cytokines has also been reported to increase the expression of GRβ in vitro, such as IL-7, IL-18, and the combination of IL-2 and IL-4 in PBMC, and IL-8 in human neutrophils [13,20,21]. Interestingly, these cytokine treatments resulted in decreased sensitivity to GC in these cell types [13,[19][20][21]. Also, bacterial superantigens and bacterial toxins have been reported to increase the expression of GRβ [22,23].
Septic serum also contains fragments of bacteria which are recognized by innate immune system [24]. In our study, serum samples were obtained mostly from patients where a gram-negative infection was either suspected, due to the primary septic source or demonstrated by culture [18]. TLR4 recognizes lipopolysaccharides (LPS) present in gram-negative bacteria leading to the subsequent induction of the initial inflammatory response [25]. Therefore, the influence of septic serum on GR and GC cell sensitivity in our study could have been the consequence of the activation of TLR4 by LPS.
It is known that GC secreted endogenously by the adrenal gland restrict the responses to TLR [26]. Multiple mechanisms have been described for the GC interference with TLRs signaling [27,28]. However, if TLR activation influences GR expression or GC sensitivity, has been scarcely studied. Indirect evidence of an effect of TLR activation on GR isoforms expression comes from two studies were an up regulation of GRβ expression was reported in vivo after the exposure to TLR4 agonist. One, in a septic animal model induced by LPS injection [29], and the other studying children with respiratory syncytial virus (RSV) infection, another known agonist of TLR4 [30]. However, it should be noted that after the induction of sepsis, and in children with RSV infection, changes in metabolic parameters, as well as cytokines, toxins and stress hormones levels are expected. Therefore, it is unclear the cause of GR isoforms modulation in both situations. The effect of LPS has been tested in vitro in bone marrow-derived macrophages and nasal mucosa fibroblasts with contradictory results, but it has not been evaluated on PBMC or lymphocytes in vitro [31,32].
In this report, we evaluated the effect of LPS, a TLR4 agonist, on GRα and GRβ expression in different cell lines and PBMC from healthy donors in culture. We also explore the effect of LPS on GC cell sensitivity in vitro.
Cell viability was >95%, determined by trypan blue exclusion staining, in cultures in the presence or not of LPS.
U-2 OS cell lines stably expressing human GRα or GRβ isoform were generated as previously described [33].

Mononuclear cell extraction
Peripheral venous blood samples (30 ml) from healthy volunteers were collected and PBMC were isolated by density gradient centrifugation (Ficoll-Paque Plus, Amersham Bioscience). Written informed consent was previously obtained from all volunteers, according to the Declaration of Helsinki.
GC sensitivity assay PBMC treated previously with or without LPS for 24 h were washed twice with PBS 1X and then GC sensitivity was evaluated by measuring: 1) Dexamethasone (DEX) induction of FKPB5 and MKP1 mRNA expression. Briefly, PBMC were cultured in the presence or absence of DEX 100 nM for 6 h, then FKPB5 and MKP1 mRNA expression was evaluated by qPCR. 2) DEX inhibition of LPS induced IL-8 and GM-CSF release by ELISA. Briefly PBMC were re-stimulated with 10 μg/ml of LPS for 6 h in presence or absence of DEX 100 nM. After incubation, cells were centrifuged and supernatant were collected and stored until the ELISA assay.

IL-8 and GM-CSF release quantification
Immunoreactive IL-8 and GM-CSF were determined on the supernatants of PBMC from the GC sensitivity assay by ELISA according to the manufacturer's instructions (Pierce Biotechnology, Inc.). Data are presented as the average of triplicate results. The lower limit of detection was 2 pg/ml for both assays.

Statistic analysis
All data are presented as medians with interquartile range and range of at least three independent experiments. Data were analyzed using Student's t test for normally distributed data and the Mann-Whitney test for nonparametric data, considering a p-value < 0.05 statistically significant.

LPS regulates the expression of GRα and GRβ isoforms in U-2 OS transfected cells
Western blot analysis of total proteins from human osteosarcoma cell line lacking endogenous GR (parental U-OFF cells, line 1), transfected with GRα (U-2 OS alpha, line 2) or GRβ (U-2 OS beta, line 3) using different antiGR antibodies is showed in Fig. 1a. When an antibody common to both GR isoforms was used, a protein of~90-95 kDa was detected in cells stably expressing GRα or GRβ isoform, but not in parental U-OFF cells. When an antibody specific for GRβ was used, a protein of~90 kDa was identified only in U-2 OS beta cells.

LPS regulates the expression of GRα and GRβ isoforms in immune and epithelial cell lines
To assess whether LPS modulates the expression of GR isoforms in cells constitutively expressing this receptor, a monocytic cell line (K562) and two human cell lines of lymphoid origin (CEM and Raji) were used. These cells were treated or not with 10 ng/ml or 10 μg/ml of LPS for 24 h. As shown in Fig. 2, LPS treatment induced a different response in the human immune cells studied. In K562 there was no change in the expression of GR isoforms at 10 ng/ml and 10 μg/ml of LPS (Fig. 2a). In Raji cells there was a significant increase in only GRα expression at both concentrations used (LPS 10 ng/ml: median 1.57, range 1.29-1.66, p = 0.01; LPS 10 μg/ml: median 1.98, range 1.91-2.55, p = 0.03) (Fig. 2b). In CEM cells there was an increase in the expression of both GR isoforms, being only statistically significant at LPS 10 μg/ml (GRα: median 1.18, range 0.99-1.63, p = 0.002; GRβ: median 1.38, range 1.10-1.55, p = 0.002) (Fig. 2c).
In HeLa cells, an epithelial cell line, there was a significant increase only in GRα expression (median 1.86, range 1.02-2.37, p = 0.002), with no change in GRβ expression when treated with 10 μg/ml of LPS for 24 h (Additional file 1).

LPS regulates the expression of GRα in PBMC from healthy individuals
We examined whether LPS modulates de expression of GR proteins isoforms in PBMC from healthy individuals. As shown in Fig. 3a, LPS 10 μg/ml treatment significantly increased GRα protein expression (median 1.95, range 1.04-4.35, p = 0.007), while GRβ protein expression was not modified. When PBMC were stimulated with LPS 10 ng/ml, a similar result was obtained as mentioned before (Fig. 3b).

LPS modifies DEX sensitivity depending on the gene studied
To assess whether LPS modify DEX sensitivity in PBMC from healthy individuals, we evaluated genes known to be transactivated o transrepressed by GC. The expression of two genes known to be up regulated by DEX, MKP-1 and FKPB5 were first tested. PBMC were stimulated or not with LPS 10 μg/ml for 24 h, followed by 6 h of DEX 100 nM. DEX induced the expression of MKP-1 and FKPB5 mRNA in PBMC previously treated or not with LPS, but LPS pretreatment did not modify the inducing effect of DEX on MKP-1 (Control 4.82±1.1 vs LPS 3.76±1.6 fold increase, p = 0.66) (Fig. 4a) and FKPB5 (Control 16.69±7.6 vs LPS 15.62±8.5 fold increase, p = 0.95) mRNA expression (Fig. 4b).
To evaluate LPS effect in DEX sensitivity in genes known to be down-regulated by GC, we test DEX capacity to inhibit LPS induced IL-8 and GM-CSF protein expression. PBMC from healthy individuals were first stimulated or not with LPS 10 μg/ml for 24 h and then re-stimulated for 6 h with LPS 10 μg/ml and DEX 100 nM. As show in Fig. 5, DEX inhibited LPS induced expression of GM-CSF when PBMC were not pretreated with LPS (0.64 ±0.18 fold increase respect to control, p = 0.0022), while DEX in PBMC pretreated with LPS did not inhibited LPS induced expression of this  (Fig. 5b). The inhibitory effect did not change in PBMC pretreated with LPS when IL-8 was evaluated (Fig. 5a).

Discussion
Infectious episodes in patients with autoimmune diseases have been always a challenge. Not only because these patients are prone to more severe infections since they commonly are receiving immunosuppressive drugs such as GC, but also because frequently a flare is triggered. Since GC are important in the treatment of autoimmune diseases and also are a key factor in the response to infections, it is important to understand how an infection can influence the cell response to GCs.
This study has shown that LPS, present in Gamnegative bacteria cell wall, is able to differentially regulate the expression of GRα and GRβ in different cell lines and PBMC from healthy donors in culture. LPS induced both, GRα and GRβ in transfected cells (Fig. 1b  and c), but with an approximately four times greater effect on GRα expression. Some other groups have evaluated the effect of LPS on the expression of GR. Haim et al. showed that LPS on bone marrow-derived macrophages induced preferentially the expression of GRβ, resulting in a partial GC resistance [31]. However, Fernández-Bertolín et al., have shown that LPS 10 μg/ml did not modify the expression of GR on nasal mucosa fibroblasts, but it did induce a GC resistance for some but not all GC actions tested [32]. We therefore evaluated the effect of LPS on different cell lines relevant in the immune system. As is shown in Fig. 2, LPS in CEM cells (T cell line) induced both, GRα and GRβ isoforms expression at the same concentrations used by Fernandez-Bertolin et al. [32], with a tendency to a greater effect on GRβ, similar to the results of Haim et , with no effect in the expression of either isoforms in K562 cells. Therefore, our results and what has been reported previously demonstrate that LPS can affect the expression of the GR isoforms, but this is a cell type specific response.
Since in normal circumstances T cells, B cells and monocytes will coexist when exposed to LPS, influencing each other, we decided to evaluate the effect of LPS on PBMC in vitro. Figure 3 shows that in these conditions there was an almost two time increase in GRα expression with no change in GRβ. This effect is opposite to the effect of serum from septic patients, which induced preferentially GRβ expression on PBMC from healthy donors [19]. There are several possible explanations for this apparent contradiction. For instance, LPS is just one of the components of the serum of patients with septic states. In an inflammatory process induced by LPS there are many mediators such as cytokines, chemokines, lipid mediators, and reactive oxygen species among others that participate in the inflammatory process [35]. Any of these components could modulate the expression of GR isoforms, and even depending on the concentration balance of these mediators, different responses could be elicited.
The GRα commonly has been described as the active GR isoform which mediates most of known GC actions, while GRβ can exert a negative dominant action over GRα. Therefore, we expected that LPS would induced an increase sensitivity to GC actions on PBMC. Surprisingly our results showed that LPS treatment induced a loss of GC inhibitory effect on the production of GM-CSF. This resistance to GC effects was not generalized, since the inhibitory effect on the production of IL-8 or the induction of MKP-1 and FKPB-5 were maintained (Figs. 4 and 5). The effects of GC even when mediated by GR, are not only dependent on the ratio of expression of GRα and β. They can also be altered by nuclear translocation, changes in GR affinity, phosphorylation, among others [6,36]. Therefore, it is possible that one or several of this factors could be involved in the GC loss of inhibition of GM-CSF induced by LPS. Indeed, in NM fibroblast, Fernández-Bertolin et al. showed that LPS 10 μg/ml treatment lead to a reduced dexamethasoneinduced GRα nuclear translocation and a decreased GILZ expression without changing the GRα/GRβ  expression and maintaining dexamethasone's induction effect on MKP-1 and GILZ expression. Therefore, they also showed a GC resistant state limited to some but not all GC effects which was not associated to a change on GR expression [32].
It is also interesting that Armstrong et al., showed that even when most cytokines maintain a normal sensitivity to GC actions in macrophages from COPD patients a GC resistance was demonstrated for GM-CSF [37].

Conclusions
In summary our study showed that LPS in vitro modulates the expression of GRα and GRβ isoforms in a cell specific manner. In PBMC from healthy donor LPS induces the expression of GRα with no change of expression of GRβ. It also induces a loss of GC inhibitory effect on the secretion of GM-CSF without affecting GC induction of MKP1 or FKBP5.
From our results as well as others, it is important to note that GCs effects can be modulated in a cell and effect specific manner. Therefore, it is not possible when a particular change of GC effects is observed under certain condition, to extrapolate those findings to other GC actions.