Steroid responses in sepsis: some novel thinking that may provide new insight

Glucocorticoid use in sepsis is controversial. In contrast to other extracellular signaling molecules, glucocorticoid receptors (GRs) are intra-cytoplasmic. Several GR isoforms have been identified. A study in Critical Care Forum suggests that sepsis alters the abundance of the dominant negative GRβ. Here we discuss GR isoforms and how they may affect cellular responses to glucocorticoids in sepsis.

Th e use of glucocorticoids in the management of patients with sepsis remains controversial. In this issue of Critical Care Forum, Guerrero and colleagues [1] present fi ndings that may bring some clarity to the topic.
Th e dispute regarding steroid use in sepsis touches on the most basic aspects of glucocorticoid action. Trials performed in the 1970s and 1980s indicated that pharmacologic doses were harmful [2], and glucocorticoid use was abandoned, only to resurface with the seminal article by Annane and colleagues [3] in 2002. Th ese data suggested that glucocorticoids would be of value in septic patients incapable of mounting a suffi cient endogenous response (that is, in patients with relative adrenal insuffi ciency). Th e work by Annane and colleagues was followed by the CORTICUS (Corticosteroid Th erapy of Septic Shock) trial [4], the data of which did not confi rm the fi ndings of Annane and colleagues. However, signifi cant methodo logical diff erences between the two studies leave the issue unresolved. To this day, the debate continues.
In the discrepant results of these trials, two specifi c fi nd ings in patients with sepsis are of particular importance. First, glucocorticoids induce polar opposite responses in diff erent tissues. Steroids limit activity in white blood cells and thus function as anti-infl ammatory agents. But in the liver and heart, glucocorticoids stimulate reactions that are pro-infl ammatory: expression of genes encoding acute-phase reactants and potentiation of the cardio-stimulatory actions of catecholamines. Th e second key fi nding relates to serum levels, which are often quite elevated, a fi nding that may refl ect changes in cortisol metabolism induced by critical illness [5]. Given these issues, it is not surprising that the role of glucocorticoids in sepsis is confounding.
One possible answer becomes apparent when the mechanisms underlying steroid activity are examined. Unlike protein hormone pathways, steroid pathways are wholly intracellular and do not involve cell-surface receptors. Rather, steroid receptors are intra-cytoplasmic. Several years ago, Revollo and Cidlowski [6] began a series of experiments examining the structure of the glucocorticoid receptor (GR). Th ese studies demonstrated that GRs arise from a single gene that is located on chromosome 5Rq31-32 and that contains nine exons and gives rise to three mRNAs (Figure 1), two of which are germane to this discussion. Th e fi rst, GRα, contains the entirety of all nine exons. Th e second, GRβ, is generated by an alternative splicing site that removes a long segment at the beginning of exon 9 and splices the shorter remaining portion of exon 9 onto exon 8. GRα contains several potential translational start sites, resulting in the formation of at least eight additional isoforms that diff er in the length of their N-termini (Figure 1). Although the affi nity for glucocorticoids is the same in each isoform, their ability to bind co-factors and DNA polymerase II, and thus to act as transcription factors, diff ers markedly. Gross and Cidlowski [7] suggest that this diff erential ability to induce gene expression is responsible for the tissue-specifi c actions of glucocorticoids.
In contrast to the isoforms of GRα, GRβ does not bind glucocorticoids. By means of mechanisms that are unclear, GRβ acts as a dominant negative inhibitor of glucocorticoid-responsive gene expression. Guerrero and colleagues found that the expression of GRβ in white blood cells of nine patients with septic shock was significantly higher on admission than on discharge. Serum from these patients enhanced the in vitro expression of both GRα and GRβ in cultured T and B cells but had a more profound eff ect on GRβ. Th e serum also induced glucocorticoid resistance in mononuclear cells in vitro. Th ese fi ndings suggest that septic shock inhibits the activity of glucocorticoids by enhancing expression of the dominant negative GRβ.
Variability in the abundance of GRβ might explain why exogenous glucocorticoids exert such capricious eff ects in patients with sepsis. Enhanced GRβ expression might also explain why glucocorticoid levels are so dramatically elevated in patients with sepsis: the presence of a receptor that inhibits glucocorticoid activity will induce increased release of the hormone to compensate for and stimulate the desired response. We have invoked a similar explanation for the markedly increased interleukin-6 (IL-6) levels observed in sepsis. Our work on IL-6 has demonstrated impairment of the GP130/JAKS1/STAT-3 signal transduction pathway that mediates many of that cytokine's eff ects [8]. To overcome this block, increasingly massive amounts of IL-6 are produced and released. Th e implication of these fi ndings may be that the increased abundance of glucocorticoids, IL-6, and perhaps a number of additional mediators and hormones is an adaptive response that should be aug mented. However, most trials of cytokine-based therapy attempt exactly the opposite: blocking activity. Perhaps it is time to reconsider the hypothesis underlying these interventions. Certainly, the fi ndings of Guerrero and colleagues reinforce the notion that sepsis is a complex disorder and mono-therapy is unlikely to work.