Glucocorticoid sensitivity in inflammatory bowel disease

Abstract

Glucocorticoid therapy is used in the treatment of moderate to severe inflammatory bowel disease (IBD). However, IBD patients display varying degrees of glucocorticoid sensitivity: some respond rapidly to the given treatment, whereas others show no response, or develop steroid therapy-related side-effects. At present, we cannot foresee whether the patient will benefit from the administered glucocorticoids or not. During the past 10 years, numerous attempts have been made to provide the means to identify and predict steroid therapy-sensitive patients in advance. This would be vital to avoid unnecessary glucocorticoid exposure in patients that do not respond to treatment with steroids. Here we provide a concise review of recent developments regarding the molecular basis of glucocorticoid sensitivity in IBD patients and the methods employed to assess it.

Key words::

• Crohn's disease
• glucocorticoid receptor
• paediatric
• steroid dependency
• steroid resistance
• ulcerative colitis

Abbreviations
CD	=	Crohn's disease
DBD	=	DNA-binding domain
ELISA	=	enzyme-linked immunosorbent assay
FOXP3	=	forkhead box P3
GC	=	glucocorticoid
G-CSF	=	granulocyte colony-stimulating factor
GITR	=	glucocorticoid-induced tumour necrosis factor inhibitor
GM-CSF	=	granulocyte-macrophage colony-stimulating factor
GR	=	glucocorticoid receptor
GRE	=	glucocorticoid response element
HR	=	hinge region
11β-HSD	=	11β-hydroxysteroid dehydrogenase
Imax	=	maximum inhibition
IBD	=	inflammatory bowel disease
IFN	=	interferon
IL	=	interleukin
IP-10	=	interferon gamma-induced protein 10
Kd	=	dissociation constant
LBD	=	ligand-binding domain
LPS	=	lipopolysaccharide
MDR1	=	multi-drug resistance 1
mRNA	=	messenger ribonucleic acid
NTD	=	N-terminal domain
PBMC	=	peripheral blood mononuclear cell
PCR	=	polymerase chain reaction
RT-PCR	=	reverse transcriptase PCR
SNP	=	single nucleotide polymorphism
TNF-α	=	tumour necrosis factor α
UC	=	ulcerative colitis

Key messages

• Glucocorticoid response is unpredictable and thus a common clinical problem encountered in the treatment of inflammatory bowel disease.

• Being able to predict the patients’ response to glucocorticoid therapy would improve the treatment of patients.

• Here we review the recent developments regarding the molecular basis of glucocorticoid sensitivity in inflammatory bowel disease (IBD) patients and the methods employed to assess it.

Introduction

Synthetic glucocorticoids are employed in the induction of remission in moderate to severe inflammatory bowel disease (IBD) (1,2). The response rates vary, and around 20%–30% of patients show resistance to the therapeutic actions of the administered steroids (1,3–5). Furthermore, glucocorticoid-related side-effects are common and range from cosmetic to life-threatening (6).

Glucocorticoids employed in the treatment of IBD include parenteral hydrocortisone and methylprednisolone and oral prednisone, prednisolone, and budesonide (Table I) (7). Hydrocortisone corresponds to the natural glucocorticoid hormone cortisol, and the other administered steroids are its synthetic analogues. Only a few studies have addressed the optimal glucocorticoid dosage and dose tapering in IBD patients (10,11). At present, the recommended treatment strategy for traditional steroids (prednisone, prednisolone) is 1 mg/kg/day of prednisone equivalents (max 60 mg/day) followed by a tapering period of 12–16 weeks (7). No direct comparisons with conventional steroids (prednisone, prednisolone) regarding the efficacy or appearance of glucocorticoid-related side-effects have been made. In severe disease, however, parenteral treatment is recommended (7). Budesonide, which has low systemic bioavailability (10%) and high affinity for glucocorticoid receptor (GR), has been studied in the treatment of Crohn's disease (CD). A recent Cochrane review, however, stated that budesonide seems to be less effective than traditional steroids in the treatment of CD, specifically if the patient has more extensive colonic involvement or severe disease (12,13).

Table I. Synthetic glucocorticoids employed in the treatment of inflammatory bowel disease, adapted from Rang et al. (8).

		Relative potency in clinical use
Compound	Relative affinity for glucocorticoid receptor^a	Anti-inflammatory	Sodium-retaining	Duration of action after oral dose (hours)
Hydrocortisone	1	1	1	short (8–12)
Prednisolone	2.2	4	0.8	short (8–12)
Methylprednisolone	11.9	5	minimal	intermediate
Dexamethasone	7.1	30	minimal	long
Budesonide^b	195	nd	nd	very short (24.5)

^aHuman foetal lung cells.

^bEdsbacker and Andersson (9).

In IBD the patient's sensitivity to glucocorticoids cannot be predicted from disease extent or severity, even though colectomy rates are higher in patients with severe colonic disease (14,15), which might imply that heavy inflammatory activity impairs glucocorticoid sensitivity in these patients and leads to a poorer response. In addition, the absorption and protein binding of glucocorticoids do not explain why some patients respond to glucocorticoid therapy and others do not (16,17). At tissue level, the interconversion between active and inert cortisol metabolites by the 11β-hydroxysteroid dehydrogenases (11β-HSDs) 1 and 2 is now considered to be a significant regulator of hormone action (18,19). In IBD-related inflammation the interconversion of cortisol metabolites in the intestine is predominantly from inactive to active form (18,19), and it is unlikely to be directly affected by the therapeutic regimens (18). Glucocorticoids enter the cells mostly via diffusion, but an active mechanism, drug-efflux pump P-glycoprotein 170, exists that transports steroids and other agents out of intestinal epithelial cells and lymphocytes. Previously, the expression of the multi-drug resistance (MDR1) gene that codes for the pump protein was found to be elevated in IBD patients who had undergone bowel resection for failed medical therapy (20). More recently, however, the impact of the MDR1 gene on glucocorticoid sensitivity in IBD patients has been questioned (21). Therefore, variable pharmacokinetics of the administered steroids does not seem to explain the variation in the glucocorticoid responsiveness.

The difference in the interindividual glucocorticoid sensitivity is a clinical problem encountered daily in the treatment of IBD. Progress in the fields of molecular and cell biology has dramatically increased our knowledge of the GR signalling pathway and has enabled us to study the structure, expression, ligand binding, and affinity of the GR. In addition, it is now possible to measure the bioactivity of the administered glucocorticoids with different methods. Here we review the recent studies on glucocorticoid sensitivity in IBD patients.

Glucocorticoid receptor structure

Glucocorticoids enter the cells via diffusion and bind to their cognate receptor GR that resides in the cytoplasm as part of a multi-protein complex. After ligand binding, the protein complex dissociates: the ligand–receptor complex dimerizes and translocates to the nucleus where it affects the transcription of glucocorticoid target genes (22).

GR belongs to the superfamily of nuclear receptors together with receptors for oestrogen, testosterone, and progesterone (23). The GR gene is localized in the long arm of chromosome 5q31-32 (24). It contains nine exons, the first of which is an untranslated region (Figure 1). The first exon has at least 13 splice variants that do not alter the amino acid sequence of the receptor (Figure 1B). However, these variants have been reported to show different tissue expression profiles and thus might be a significant control mechanism in the transcription of GR (Figure 1) (25,26).

Figure 1. The structure of the glucocorticoid receptor (GR). A: The GR gene consists of at least nine alternative exon 1s (A–J) and eight other exons (2–9). The polymorphisms of the GR are presented at their respective positions in the gene. B: The alternative mRNA transcripts of the GR gene. The 13 different exon 1 splice variants do not alter the amino acid sequence of the gene. For GRα the presence of the different translation initiation sites for A-D3 isoforms is schematically presented. Arg = arginine. C: The full-length GRα, its main functional domains, and the sites of post-translational covalent modifications: S = sumoylation; Ub = ubiquitination.

Exon 2 is the site for translation initiation. In 2005, it was reported that eight alternative translation initiation sites exist for human GRα (Figure 1B) (27). These GR isoforms differ from their amino-terminal region and have been reported to vary from one another in terms of transactivation capacity from synthetic glucocorticoid response element (GRE)-driven promoter, tissue expression, and target-gene induction profile (27). It has been hypothesized that similar isoforms could also exist for GRβ, which would further increase the capacity of the GR to regulate its signal transduction.

Exons 3 to 9 contain the sequence for DNA- and ligand-binding domains of the GR (28). Receptor isoforms that can affect the glucocorticoid sensitivity arising from this area are the GRα/β isoforms generated by the alternative splicing of the exon 9, GR-A lacking amino acids encoded by exons 5, 6, and 7, and GR-P lacking amino acids from exons 8 and 9 (Figure 1B). Alternative splicing of exon 9 produces the two receptor isoforms GRα and GRβ, of which the former is considered to be the ‘classical’ glucocorticoid receptor of 777 amino acids (28). The latter, GRβ, is 35 amino acids shorter and after its identification was thought to be a dominant negative inhibitor of GRα (29). However, recent research has revealed its potential role as an independent regulator of gene expression, even in the absence of GRα (30). GR-A and GR-P have been identified in glucocorticoid-resistant myeloma patients, but at present it is not clear whether they increase or decrease glucocorticoid sensitivity (31,32). Also, a fifth isoform GRγ with an additional arginine between exons 3 and 4 has for example been reported in childhood lymphoblastic leukaemia. This modification decreases the transactivation potential of the receptor by 50% (33–35).

In addition to the several isoforms, a number of single nucleotide polymorphisms (SNPs) have been characterized in the GR gene that can further modulate the glucocorticoid sensitivity of an individual patient. Two polymorphisms, N363S and BclI, have been associated with increased sensitivity to glucocorticoids (36). Instead, ER22/23EK and GR-9β are linked to a relative glucocorticoid resistance (36,37).

Glucocorticoid receptor in inflammatory bowel disease

Since the GR was cloned in the 1980s, it has become evident that even carrying a normal GR gene allele allows for rich variation in the GR characteristics in otherwise healthy individuals. In IBD, alterations in the number and structure of GR that may affect therapeutic response to GCs have been reported (see below).

Altered numbers of glucocorticoid receptors

Studies in patients with asthma and rheumatoid arthritis have suggested that glucocorticoids can down-regulate the expression of the GR in target tissues, which has been considered to be a protective mechanism against long-term exposure to the hormone (38,39). After glucocorticoid treatment, the receptor expression levels return to base-line, but the results regarding the association between GR numbers and glucocorticoid sensitivity have remained contradictory (38,39). In IBD patients, the number of GRs has been evaluated in peripheral blood mononuclear cells (PBMCs) and colonic mucosal cells (Table II). However, the results have been confusing, as the GR numbers in IBD patients have been reported to be lower, similar, or higher than in the control patients (40–42) (Table II). Therefore, no consensus can be drawn from these findings. This might reflect different study designs applied in these studies, different technical methods (dexamethasone-binding assay versus mRNA quantitation), and the small number of patients employed that does not allow definitive conclusions, as well as the cells’ potential to up- and down-regulate GR in different situations (43).

Table II. Studies assessing the number and affinity of glucocorticoid receptors (GR) in colonic mucosa and peripheral blood mononuclear cells (PBMC) in inflammatory bowel disease.

Ref.	Objective	Patients, disease	Method	Results
(45)	To study the number of GR and their affinity to ligand in PBMC	11 UC	Whole cell [^3H]Dex- binding assay	Steroid non-responders (n = 6) had higher number of GR binding sites than responders (n = 5) or controls and lower affinity of GR to ligand than responders.
(40)	To study GR receptor numbers in PBMC and colonic mucosa	31 CD; 18 UC	Cytosolic [^3H]Dex-radioassay	Mucosal GR levels were decreased in IBD patients when compared with controls. In steroid-treated IBD patients GR numbers were lower in PBMC.
(44)	To investigate the expression of GR and the apparent dissociation constants in PBMC of IBD patients	22 UC; 17 CD	Whole cell [^3H]Dex-binding assay	In IBD patients the Kd was constantly higher than in the healthy controls. IBD patients not on steroid therapy had higher GR levels than steroid-treated IBD patients or controls.
(41)	To study GR mRNA levels in peripheral leucocytes and the effect of low-dose Dex treatment on them	21 UC	Quantitation of GR mRNA in solution hybridization assay	GR mRNA levels were higher in IBD patients than in the controls. The mRNA levels were the same between steroid responders (11) and non-responders (10).
(42)	To study GR mRNA expression in PBMC and colonic mucosa of IBD patients and whether they correlate with disease activity or can predict GC treatment response	33 UC; 21 CD	Quantitation of GR mRNA by RT-PCR, immunohistochemistry	Systemic and local GR mRNA expression was similar to controls; however, in the mucosa of UC patients with impaired GC response the levels were down-regulated. GR immunoreactivity was found in immune and epithelial cells.

Dex = dexamethasone; UC = ulcerative colitis; CD = Crohn's disease.

Altered affinity to the ligand

Another possible explanation for the varying glucocorticoid sensitivity of IBD patients could be the altered affinity of the GR to its ligand. Indeed, IBD patients have been shown to have lower affinity to GR than normal controls (Table II) (44). In addition, in another study, patients with steroid-resistant IBD were reported to have a lower affinity to GR than glucocorticoid responders (Table II) (45). These studies, although small, do support the idea that systemic inflammation could alter the affinity of the GR to its ligand, as seen for example in asthma (39).

Glucocorticoid receptor isoforms α and β

The elevated expression of GRβ has been implicated as being one of the causes behind decreased sensitivity to glucocorticoid therapy (46). In IBD, a number of groups reported that elevated GRβ expression was detected in colonic mucosal cells and PBMCs of patients resistant to steroid treatment (Table III) (47–51). However, the largest study on this subject (86 patients) arrived at opposing results, showing no difference in the GRβ levels of glucocorticoid responders and non-responders (Table III) (52). In addition, a small study (n = 17) in patients with Crohn's disease concluded that GRα and GRβ levels were lower in PBMCs of CD patients than in healthy controls (53). Therefore, more studies would be needed to clarify the contribution of the GRβ to the glucocorticoid sensitivity of IBD patients.

Table III. Studies assessing the association between the number of glucocorticoid receptors (GR) GRα and GRβ in peripheral blood mononuclear cells (PBMC) and colonic mucosa, and the response to glucocorticoid (GC) therapy in adult patients with inflammatory bowel disease.

Ref.	Objective	Patients, disease	Method	Results
(47)	To study whether RT-PCR analysis of hGRβ mRNA in PBMC of UC patients can predict the response to GC therapy	23 UC	GRα and GRβ analysis by mRNA RT-PCR, Western blot	GRβ was positive in 83.3% of the GC non-responders (n = 12) and in 9.1% of the GC-responsive patients (n = 11).
(48)	To quantify the levels and serial changes of PBMC GR mRNA in IBD patients	34 UC; 13 CD	Real-time fluorescence-monitored PCR	GRβ mRNA expression was increased in active UC. GRβ mRNA expression was also higher in GC-resistant than GC-sensitive UC.
(53)	To quantify the levels of GR mRNA in PBMC	17 CD	RT-PCR	GRα and GRβ mRNA levels were lower in patients with CD than in the controls. The longer the disease duration, the smaller the amount of GRα mRNA in PBMCs of the patients.
(49)	To study the expression of GRα and GRβ in colonic mucosal cells and their correlation with response to GC therapy and inflammation	25 UC	Immunohistochemistry	The expression of GRα was positively associated with response to GC treatment, whereas GRβ associated negatively with the response to GCs.
(50)	To assess the expression of GRα and GRβ in PBMCs of CD patients and to study their relationship to the response to GC therapy	42 CD	RT-PCR using real-time PCR techniques; IL-18 levels were measured with ELISA	GRβ mRNA expression was significantly higher in GC non-responders with active CD. In these patients, GRβ mRNA correlated directly with IL-18 levels measured from the serum.
(52)	To study whether GRβ expression in PBMCs of IBD patients could serve as a predictor of GC response	86 IBD	RT-PCR	GRβ expression was similar between GC-treated and non-GC-treated patients and between GC-responders and non-responders.
(51)	To study the relationship between the frequency and type of infiltrating cells expressing GRα, GRβ, and FOXP3 in biopsied colonic mucosa and the GC responsiveness of UC patients	38 UC	Immunohistochemistry, RT-PCR	GC non-responders had higher GRβ^+ cell count than GC-responders. The FOXP3^+ cell count was significantly higher in GC-responders.

UC = ulcerative colitis; CD = Crohn's disease.

Glucocorticoid receptor polymorphisms

An Italian group investigated the presence of the GR polymorphisms in paediatric patients with IBD and compared the results with glucocorticoid therapy success (54). They reported that the presence of the BclI polymorphism between exons 2 and 3 associated significantly with glucocorticoid therapy responsiveness. However, another study on glucocorticoid receptor haplotypes in adult IBD patients did not find the BclI polymorphism to be significant in IBD, nor could they find any association between different GR haplotypes and steroid therapy response in IBD patients (55). Also the ER22/23EK and N363S polymorphisms have been recently studied in IBD, but these polymorphisms did not show any allele or haplotype difference between paediatric CD patients and healthy controls (56).

Glucocorticoid bioactivity

Our group has studied the possibility of using the circulating bioactivity of the administered glucocorticoids as a marker of therapy outcome or the appearance of treatment-related side-effects (Table IV) (57). Glucocorticoid bioactivity was measured with a COS-1 cell bioassay in which the cells are transfected with GR and reporter genes (58). In this technique, a 10 µL sample of patient serum is added to the cell culture medium, and after an overnight incubation the relative glucocorticoid bioactivity can be measured. Unfortunately, no associations were found between clinical response to glucocorticoids and circulating glucocorticoid bioactivity (Table IV) (57). Our findings were confirmed in another study on 50 patients receiving intravenous steroids for severe IBD (59). These findings support the hypothesis that glucocorticoid treatment response is dependent on glucocorticoid sensitivity at the target cell level and not on the circulating bioactivity of the drug. So far, there is no clear explanation why some patients failing oral glucocorticoid therapy respond to intravenous administration of steroids. In search of other biomarkers that could be associated with steroid therapy outcome, we found, however, that serum adiponectin, a hormone secreted by white adipose tissue, associated with glucocorticoid treatment-related early side-effects (Table IV) (60). If this finding is confirmed in larger studies, serum adiponectin might be used as a readily available biomarker of steroid therapy-associated adverse events.

Table IV. Other methods employed to study the glucocorticoid (GC) response in patients with inflammatory bowel disease (IBD).

Ref.	Objective	Patients, disease	Method	Results
(69)	To study the hypothesis that GC-resistant UC patients have GC-resistant T lymphocytes	18 UC	Phytohaemagglutin-stimulated peripheral blood T lymphocytes treated with dexamethasone	Treatment failures and incomplete responders had an Imax of less than 60%; all complete responders had an Imax of more than 60%.
(70)	To compare the corticosensitivity of CD patients to that of healthy subjects	19 CD	Dexamethasone inhibition of LPS-induced cytokine secretion by whole blood cell cultures	CD patients had a markedly decreased dexamethasone-mediated inhibition of TNF-α secretion.
(71)	To study the lymphocytes in the colonic mucosa	30 UC; 15 CD	Immunohistochemistry, flow cytometry	High labelling of CD19^+ Ki-67 lymphocytes was specific for active UC that was resistant to GC therapy.
(72)	To study the relationship between steroid responsiveness and SLC22A4/A5 polymorphism located within IBD 5 locus	94 UC; 94 CD	Genotyping and haplotype analysis	The G allele on –368T > G in SLC22A5 was associated with steroid resistance in CD patients. Haplotype analysis between –446C > T and –368T > G in the SLC22A5 promoter region indicated CG allele as the risk haplotype for steroid resistance in CD patients.
(57)	To assess circulating GC bioactivity in paediatric patients with IBD	18 UC; 3 CD; 1 IC	Cell culture bioassay	The circulating GC bioactivity did not associate with clinical response to steroid therapy.
(60)	To evaluate whether white adipose-derived adiponectin and leptin could associate with the response to systemic GC therapy	16 UC; 2 CD; 1 IC	ELISA	High serum adiponectin associated with early appearing glucocorticoid-related side-effects.
(61)	To analyse the effect of GC-treated patient serum on the immune activation of healthy donor PBMC	16 UC; 2 CD; 1 IC	ELISA, RT-PCR	The expression of FOXP3 and GITR and the secretion of IFN-γ mirroring effector T cells decreased during glucocorticoid therapy.

UC = ulcerative colitis; CD = Crohn's disease; IC = indeterminate colitis; PBMC = peripheral blood mononuclear cells.

Lymphocyte activation

We have also recently developed a novel bioassay in which the effect of the patient serum on healthy donor lymphocytes is measured with a panel of T cell activation markers (Table IV) (61,62). The markers analysed include the expression of FOXP3 and glucocorticoid-induced tumour necrosis factor inhibitor (GITR), reflecting the actions of regulatory T cells and the secretion of interferon (IFN) γ mirroring effector T cells. In paediatric IBD patients glucocorticoid therapy decreased the potential of the patient serum to induce expression of these activation markers in donor cells (61). In addition, in adult CD patients the immunological activation potency in patient serum before anti-tumour necrosis factor-α (TNF-α) therapy associated with endoscopic CD activity score and with anti-TNF-α treatment response (Table IV) (62). Therefore, this novel assay can be used in monitoring the individual responses to IBD therapy, including glucocorticoids. However, the assay has not yet been standardized for commercial cell lines.

Cytokines and glucocorticoid sensitivity in IBD

In IBD, a few studies have concentrated on the impact of the cytokine profile on glucocorticoid therapy response. In 1999, Ishiguro reported that the levels of TNF-α, interleukin (IL) 6, and IL-8 were elevated in steroid-resistant ulcerative colitis (UC) patients (63). Interestingly, almost 10 years later a –308A allele of the TNF-α gene was associated with steroid resistance in paediatric CD patients (64). Also, in another study several cytokines (IFN-γ, granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF), TNF-α, IL-1α, IL-10, IL-17, and IP-10) were reported to correlate with in-vitro steroid response: in this report, IL-2 seemed to convey steroid resistance, both in healthy controls and in patients with UC (65). Indeed, IL-2 has been associated with increased GRβ expression, decreased affinity of GR to its ligand, and decreased GR translocation to the cell nucleus (66). A third study on the subject, however, arrived at contradicting results and did not find any associations between cytokine mRNA expression and response to glucocorticoid therapy (67). Extracellular cytokines are able to regulate transcriptional processes via phosphorylation of transcription factors, e.g. the STAT family. At tissue level, the presence of STAT-5 has been related to steroid-resistance in UC in a preliminary study in adults (68).

Other methods employed to study glucocorticoid sensitivity in patients with inflammatory bowel disease

During the past 10 years, a number of other methods have also been employed in the study of glucocorticoid sensitivity in IBD patients (Table IV). These include studying lymphocytes and other genes and cellular markers from IBD patients in relation to glucocorticoid sensitivity, as well as assessing different methods that could be employed in the prediction of the steroid response (69–72). However, none of these has emerged as a clinically applicable means for monitoring steroid therapy. Thus, the steroid response in severe disease is currently monitored with clinical variables predicting response failures on days 3–7 of therapy to guide the need for second-line therapy or surgery (14,15,73–77).

Also other inflammatory diseases, such as asthma and rheumatoid arthritis that are treated with glucocorticoids, face the problem of glucocorticoid insensitivity (78). In asthma, genetic factors might play a role in glucocorticoid responsiveness as glucocorticoid insensitivity has been found within families and microarray studies have revealed a panel of genes that discriminate between glucocorticoid-responsive and non-responsive groups (78). Recently, a similar study was conducted in paediatric patients with severe ulcerative colitis that resulted in the characterization of 41 genes that expressed differentially between glucocorticoid responders and non-responders (79). However, so far the total number of steroid non-responsive patients studied is low, challenging the recruitment of patients to provide clear answers on the underlying mechanisms.

Conclusions

Being able to predict in advance the patients’ response to glucocorticoid therapy would be a powerful tool in the treatment of IBD. With the gradual unfolding of the molecular basis of glucocorticoid actions in different tissues, this might be possible in the future. In IBD, numerous attempts have been made to characterize patients according to their sensitivity to glucocorticoids with respect to ligand availability, GR structure, cytokine profile, and other glucocorticoid sensitivity-assessing methods and biomarkers. At present, however, the data on the glucocorticoid signalling cascade are still rather fragmentary, and clinical studies seem to reach contradictory conclusions—from bench to bedside is still a long way. Small studies with heterogeneous study designs might be one cause for the controversy, but, as recent studies have shown, the GR splice and translational isoforms identified until now can form up to 256 different homo- and heterodimers (80). These receptor complex subtypes might all express distinct patterns of intracellular localization, coactivator recruitment, and transactivation and -repression properties. Therefore, the contradictions we see in the results might only be a reflection of the complexity of the steroid receptor signalling network that has just begun to reveal itself. It is more than likely that no single cytokine, polymorphism, or isoform alone can respond to the question of whether or not a person is sensitive to glucocorticoid therapy. Rather, it is possible that the combined information from all of these variables might one day provide us with an estimate of the probable response.

Acknowledgements

This study was funded by the Emil Aaltonen Foundation, the Finnish Cultural Foundation, the Helsinki University Central Hospital Grant, the Finnish Paediatric Research Foundation, the Päivikki and Sakari Sohlberg Foundation, the Finnish Medical Society Duodecim, the Maud Kuistila Memorial Foundation, and the Biomedicum Helsinki Foundation.

Declaration of interest: The authors report no conflict of interest. The authors alone are responsible of the content and writing of the paper.