Pediatric Nephrology

, Volume 20, Issue 3, pp 306–312

Cytokine inhibition of JAK-STAT signaling: a new mechanism of growth hormone resistance

Authors

    • Department of Cellular and Molecular PhysiologyPennsylvania State University College of Medicine
  • Ly Hong-Brown
    • Department of Cellular and Molecular PhysiologyPennsylvania State University College of Medicine
  • Robert A. Frost
    • Department of Cellular and Molecular PhysiologyPennsylvania State University College of Medicine
Review

DOI: 10.1007/s00467-004-1607-9

Cite this article as:
Lang, C.H., Hong-Brown, L. & Frost, R.A. Pediatr Nephrol (2005) 20: 306. doi:10.1007/s00467-004-1607-9

Abstract

Growth hormone (GH) and insulin-like growth factor (IGF)-I are potent regulators of muscle mass in health and disease. This somatomedin axis is markedly deranged in various catabolic conditions in which circulating and tissue levels of inflammatory cytokines are elevated. The plasma concentration of IGF-I, which is primarily determined by hepatic synthesis and secretion of the peptide hormone, is dramatically decreased during catabolic and inflammatory conditions. Moreover, many of these conditions are also associated with an inability of GH to stimulate hepatic IGF-I synthesis. This defect results from an impaired phosphorylation and activation of the traditional JAK2/STAT5 signal transduction pathway. Numerous lines of evidence support the role of tumor necrosis factor (TNF)-α as a prominent but probably not the sole mediator of the sepsis-induced impairment in basal and GH-stimulated IGF-I synthesis in liver. Additionally, catabolic conditions produce comparable alterations in skeletal muscle. However, in contrast to liver, the GH resistance in muscle is not mediated by a defect in STAT5 phosphorylation. Muscle is now recognized to respond to infectious stimuli with the production of numerous inflammatory cytokines, including TNF-α. Furthermore, myocytes cultured with TNF-α are GH resistant and this defect appears mediated via a STAT5-independent but JNK-dependent mechanism. Collectively, these changes act to limit IGF-I availability in muscle, which disturbs protein balance and results in the loss of protein stores in catabolic and inflammatory conditions.

Keywords

SepsisEndotoxinTumor necrosis factor-αInsulin-like growth factor-IMuscleJNK

Introduction

One hallmark of traumatic and inflammatory conditions (such as sepsis, endotoxemia, thermal injury, glucocorticoid excess, AIDS, chronic alcohol abuse, and uremia) is the negative nitrogen balance produced by the net catabolism of body protein, primarily derived from skeletal muscle [1]. In many studies, this catabolism results from a simultaneous increase in protein degradation and a decrease in protein synthesis [2, 3]. When prolonged, this imbalance leads to the erosion of lean body mass (LBM) and the wasting commonly observed in these patient populations. The clinical implications of protracted skeletal muscle protein loss include poor wound healing, loss of muscle strength and physical activity, increased risk of thromboembolic complications, and failure to wean patients from ventilator support. Sustained decreases in LBM are causally linked to increases in morbidity and mortality in chronic catabolic conditions [4, 5]. Thus, a more complete understanding of the factors responsible for regulating protein synthesis is of significant clinical relevance and represents an important area for improved patient care.

Our laboratory and others have reported that various inflammatory conditions, exemplified by sepsis and endotoxemia, decrease the rate of protein synthesis in skeletal muscle [6, 7, 8, 9]. This inhibition is undoubtedly multifactorial (Fig. 1). On the catabolic side of the protein balance equation, elevations in glucocorticoids and various inflammatory cytokines adversely impact the protein synthetic rate in muscle. Additionally, infection, trauma, and inflammatory cytokines also reduce the circulating concentration of or diminish the responsiveness of tissues to anabolic agents, such as insulin, amino acids, insulin-like growth factor (IGF)-I and growth hormone (GH). Several lines of evidence suggest that cytokine networks, particularly those involving the inflammatory cytokine tumor necrosis factor (TNF)-α, play a central role in the net protein catabolism observed during infection and inflammation [10]. Although the mechanism by which TNF-α and sepsis impair protein synthesis is still poorly defined, evidence outlined below indicates that these insults decrease the circulating and tissue concentrations of the anabolic hormone IGF-I under both basal conditions and in response to GH stimulation. In addition to TNF-α, sepsis and endotoxin [e.g., lipopolysaccharide (LPS)] also stimulate the synthesis of other inflammatory cytokines, including interleukin (IL)-1 and -6, whose over-expression participates in the development of the wasting syndrome [6, 11, 12, 13]. Collectively, these changes adversely impact several key regulatory steps in mRNA translation, especially peptide-chain initiation, leading to a reduction in protein synthesis [2, 7, 8, 9].
Fig

. 1 Overview of interaction between sepsis and endotoxin [lipopolysaccharide (LPS)], cytokines, insulin-like growth factor-I (IGF-I), and translation initiation in the regulation of muscle protein synthesis

Inflammation-induced decreases in circulating IGF-I

Numerous studies have reported that a variety of inflammatory mediators decrease the prevailing circulating concentration of IGF-I [14]. The reduction can occur rapidly, (e.g., within 4 h) as in the response to LPS and interleukin-1 [15, 16, 17], yet may persist for several days or weeks as seen following sepsis, trauma, and burn injury [18, 19, 20]. Under the limited conditions where it has been examined, the inflammation-induced decrease in total IGF-I is mirrored by a proportional or slightly greater decline in the circulating concentration of “free” IGF-I [21], which is believed to represent the bioactive form of the peptide. Conceptually, the reduction in plasma IGF-I may result because of a decreased rate of appearance or an increased rate of disappearance of IGF-I from the circulation. Whereas protein malnutrition has been reported to enhance the clearance of IGF-I [22], there is little available evidence supporting such a mechanism in response to sepsis and inflammation [15]. In contrast, there is overwhelming and indisputable evidence that various catabolic insults decrease the synthesis of IGF-I. This decrease can be visualized at the level of gene expression (as a decrease in the steady-state content of IGF-I mRNA) and at the level of protein content (as a decrease in the tissue IGF-I protein content) [6, 8, 16, 17, 23]. Furthermore, sepsis and LPS also inhibit the secretion of the native IGF-I peptide by the isolated perfused liver [15] as well as the net hepatic secretion of IGF-I under in vivo conditions [24]. Although not a universal finding [25], the bulk of the available data indicate that much of the sepsis-induced decrease in basal IGF-I in blood and liver appears mediated, either directly or indirectly, via the over-production of TNF-α and/or IL-1β. This modulatory role for these cytokines is illustrated by the amelioration of the sepsis-induced decrement in IGF-I by pretreatment with agents that antagonize the actions of specific cytokines [6, 16, 26].

Hepatic GH resistance

GH is a major physiological regulator of IGF-I synthesis and plays a critical role in regulating muscle protein balance. Defects in GH signaling profoundly reduce postnatal growth. Therefore, a change in the blood GH level and/or diminished GH responsiveness of peripheral tissues may mediate the inflammation-induced changes in IGF-I. Critically ill patients consistently have augmented baseline levels of GH with a concomitant reduction in IGF-I [18, 19, 20, 27, 28]. Furthermore, the injection of LPS in humans, which leads to a marked increase in circulating inflammatory cytokines, also increases GH and decreases IGF-I [20]. While these data are consistent with the presence of GH resistance, more definitive evidence has been reported by Dahn and Lange [29]. In this study, impaired GH action was evidenced by the increment in plasma IGF-I in response to a maximally stimulating dose of GH that was smaller in septic patient than in control subjects. Because the liver is the primary synthetic site for blood-borne IGF-I, collectively these data imply the presence of hepatic GH resistance.

The duration and intensity of the GH response are tightly controlled and ultimately determined by the net balance of positive and negative regulatory mechanisms [30]. GH action is initiated by binding to its cognate receptor (GHR), which results in receptor dimerization and the subsequent autophosphorylation of the non-receptor tyrosine kinases Janus kinase (JAK)2 and JAK3. Activation of JAKs stimulates the down-stream tyrosine phosphorylation of several signaling proteins, including the signal transducer and activator of transcription (STAT) proteins, STAT1, STAT3, and STAT5 [31]. Upon phosphorylation the latent STAT proteins present in the cytoplasm dimerize and translocate to the nucleus where they bind to specific cis -acting elements and activate GH-regulated genes. Hepatic GH-activated genes include the acid-labile subunit (ALS), serine protease inhibitor (Spi) 2.1, and suppressor of cytokine signaling (SOCS) 3 [30, 31]. Each tier of activation provides a potential point of regulation. The mechanism by which the STAT proteins function is becoming increasingly intricate with the discovery that they interact with various cytoplasmic and nuclear proteins capable of regulating STAT-mediated gene transcription. The STAT5 transcription factor, for example, is both necessary and sufficient for the GH-mediated increases in hepatic IGF-I gene transcription based on evidence from studies employing STAT5b null mice [32] and from studies using adenovirus-mediated gene transfer of a dominant/negative or constitutively active STAT5b [33]. Natural mutation in the human STAT5 gene also leads to decreased synthesis of IGF-I [34].

Initial studies by Mao et al. [35] demonstrated that LPS decreased the ability of GH to stimulate STAT5 phosphorylation in liver under in vivo conditions. A similar decrement in GH-stimulated hepatic STAT5 phosphorylation has also been observed during sepsis [36]. Moreover, LPS decreases the amount of phosphorylated STAT5 that translocates to the nucleus and binds to DNA [37, 38, 39]. The mechanism by which these inflammatory insults impair GH action appears mediated in part by a decreased phosphorylation of the upstream kinase JAK2 [35]. Given these results it would be reasonable to predict that sepsis and LPS would also decrease GHR phosphorylation. However, there is no direct evidence in support of this assumption. For example, while several studies report a sepsis- or LPS-induced decrease in hepatic GHR mRNA [37, 38, 40], no consistent decrease in the amount of total GHR protein was detected in liver by Western blot analysis [35, 37, 38]. Recent studies have revealed that LPS-induced down-regulation of murine GHR mRNA transcription is primarily mediated by a TNF-dependent mechanism related to an inhibition of Sp transactivator binding [41]. Finally, Defalque et al. [40] have shown that LPS decreases binding of 125I-labeled bovine GH in liver homogenates, suggesting the hepatic GH resistance originates at the level of the GHR. Regardless of the exact mechanism, it is noteworthy that the sepsis- and LPS-induced decrease in STAT5 phosphorylation results in a diminution of the hepatic synthesis and secretion of IGF-I [37, 40], a physiologically relevant endpoint of STAT5 phosphorylation.

Sepsis and LPS result in the rapid over-expression of various inflammatory cytokines. The role of TNF-α and other cytokines as mediators of the inflammation-induced decrease in the constitutive synthesis of hepatic IGF-I as well as the production of GH resistance has been studied by several investigators. Based on early studies, several lines of evidence suggest an important role for TNF-α. Firstly, the in vivo infusion of a non-lethal dose of TNF-α decreases hepatic IGF-I mRNA and the plasma protein [16, 23]. Secondly, addition of TNF-α to cultured hepatocytes decreases IGF-I secretion [37, 42]. Thirdly, sepsis and LPS upregulate hepatic mRNA and plasma TNF-α levels [43]. Finally, pretreatment of rats with the TNF-binding protein (which functions as a TNF-α antagonist) prevents the sepsis-induced decrease in basal IGF-I mRNA content [37].

TNF-α also appears central to the sepsis-induced decrease in hepatic GH sensitivity. In this regard, Yumet et al. [37] reported that TNF-α markedly blunts the ability of GH to stimulate GHR phosphorylation and the nuclear accumulation of phosphorylated STAT5b in the hepatocyte cell line CWSV-1. In addition, TNF-α also inhibits GH-induced STAT5 DNA-binding to both the rat β-casein [37] or Spi 2.1-CAT promoter [39]. Importantly, the diminution of GH action by TNF-α was also evidenced by the failure of GH to stimulate IGF-I synthesis [37]. TNF-α also suppresses GH-induced increases in IGF-I mRNA in isolated rat hepatocytes [41, 44]. Despite the extensive in vivo and in vitro evidence, it is premature to conclude that TNF-α is the sole or even primary mediator of the GH resistance. For example, IL-1β also blunts GH stimulation of IGF-I synthesis and appears more potent than TNF-α in this regard [44]. Furthermore, pretreatment of rats with an IL-1 receptor antagonist reverses both the decrease in IGF-I and the concomitant reduction in muscle protein synthesis [6]. Similarly, both IL-1β and IL-6 greatly attenuate increases in Spi 2.1 promoter activity and mRNA accumulation in isolated hepatocytes treated with GH [39]. Finally, a recent study using cytokine-specific knockout mice has reported that the hepatic GH resistance induced by LPS was primarily mediated by IL-6, not TNF-α [38]. These data suggest that there may well be considerable redundancy in terms of inflammatory cytokine regulation of hepatic IGF-I synthesis and GH sensitivity under catabolic conditions.

Inflammation-induced changes in muscle IGF-I

Cytokines modulate cellular functions either locally in an autocrine/paracrine manner or at sites distant from their origin of synthesis in a manner consistent with that of endocrine hormones. Blood-borne cytokines are primarily derived from macrophage-rich tissues (e.g., liver and spleen) and various myelomonocytic cells. However, non-immune tissues also synthesize a host of inflammatory cytokines and the increased local production of cytokines by cardiac muscle has been implicated in organ dysfunction in several pathological conditions [45]. These data suggest that the autocrine/paracrine actions of cytokines, particularly TNF-α, IL-1, and IL-6, may dramatically impact the function of striated muscle. However, despite their potential importance and the enormous contribution of skeletal muscle to total body mass (approximately 45%), little is known about the role of endogenous cytokine synthesis in muscle per se. Our studies indicate that LPS dose- and time-dependently increases the mRNA content of several early phase cytokines (e.g., TNF-α, IL-6, and IL-1β) and one late-phase cytokine (e.g., HMGB-1) in gastrocnemius muscle [43, 46]. The increased TNF-α mRNA was sustained for at least 24 h and was accompanied by an increased TNF-α peptide content in muscle. Comparable changes in muscle cytokines are observed in response to an infectious insult induced with live bacteria (C.H. Lang, unpublished observations). Furthermore, skeletal muscle also contains both toll-like receptor (TLR)-4 and TLR2, and the ability of LPS to increase the synthesis of cytokines in muscle is prevented in mice harboring a mutation in the TLR4 receptor [46]. LPS also stimulates the synthesis of cytokines in C2C12 myoblasts and myotubes, and in these cells this increase correlates with the reduction in IGF-I [46, 47]. Therefore, skeletal muscle clearly possesses both the afferent and efferent limbs of the innate immune system, including toll-like receptors and both early and late-phase cytokines.

Hypermetabolic infection, sterile peritonitis, thermal injury, LPS, and critical illness all decrease the plasma IGF-I concentration. Although the decrease in hepatic IGF-I synthesis undoubtedly accounts for the majority of the fall in circulating IGF-I, sepsis and other inflammatory conditions also decrease IGF-I mRNA and peptide content in skeletal muscle per se [6, 8, 9, 16, 17, 21]. The muscle response to sepsis and trauma is noteworthy because IGF-I clearly functions as an anabolic agent via an autocrine/paracrine mechanism [48]. This latter role is exemplified by in vitro data indicating that the incubation of myocytes with a specific IGF-I receptor antibody decreases the constitutive rate of protein synthesis [49], suggesting that endogenously produced IGF-I is necessary for maintaining basal protein balance. Our previous studies, in which rats were pretreated with various cytokine antagonists, also demonstrate that a significant portion of the sepsis- or LPS-induced decrease in IGF-I in muscle is attributable to the enhanced production of cytokines [6, 12, 17]. However, because of the redundancy and overlap of the inflammatory cytokine network, it is often difficult to determine the specific cytokine mediator under in vivo conditions. We have also reported that the content of IGF-I mRNA and peptide in skeletal muscle is strongly correlated with the in vivo rate of muscle protein synthesis as well as the rate of initiation of translation [6, 8]. Furthermore, these findings have been extended to an in vitro system, in which addition of TNF-α to cultured skeletal muscle cells was shown to impair protein synthesis [49, 50].

Cytokine-induced GH resistance in muscle

Recent studies report that inflammation also impairs the ability of GH to stimulate IGF-I synthesis in skeletal muscle [36]. However, importantly, these data indicate that the mechanism for GH resistance in muscle (e.g., gastrocnemius) differs from that seen in liver. In muscle from control rats, GH increased STAT5 phosphorylation, a response that was qualitatively similar but quantitatively smaller than that observed in liver. In contradistinction to liver, the extent of STAT5 phosphorylation observed after administering a maximally stimulating dose of GH was actually greater in muscle from septic rats than in muscle from control animals. Because comparable results were obtained in muscle from endotoxic rats stimulated with GH [36], the differential tissue response to GH is unlikely to result from differences in experimental models. In addition, sepsis did not impair the ability of GH to stimulate STAT5 phosphorylation in cardiac muscle. These data suggest that the sepsis-induced impairment in STAT5 phosphorylation may be liver specific. Finally, there was no difference in the phosphorylation state of either STAT1 or STAT3 in muscle from GH-treated control and septic rats [36]. However, despite the lack of a discernable defect in the JAK/STAT signaling pathway in muscle, the ability of GH to increase IGF-I mRNA content in muscle was completely prevented in septic rats. The exact mechanism by which IGF-I mRNA levels remain depressed despite a robust GH-inducible phosphorylation of STAT5 remains to be elucidated.

To more fully characterize the mechanisms for the sepsis-induced GH resistance in skeletal muscle additional studies were performed using cultured C2C12 myocytes. Data from recent studies indicate that LPS (used as an in vitro model of gram-negative infection) stimulates the expression of multiple inflammatory cytokines, including TNF-α, IL-1, and IL-6, in a dose- and time-dependent manner in this cell type [46]. The responsiveness of C2C12 cells to LPS is in the same range as peripheral blood mononuclear cells and cardiomyocytes. Moreover, incubation of myocytes with either TNF-α or IL-1 stimulates the synthesis of IL-6. Therefore, similar to in vivo conditions, cultured myocytes are capable of mounting an innate immune response.

Many muscle cell lines do not express endogenous GHR and, therefore, until recently it was unclear whether GH increases IGF-I in this cell type via activation of the JAK/STAT pathway. However, work by our laboratory [46] and others [51] indicates that in C2C12 myocytes GH stimulates the accumulation of IGF-I mRNA in a dose- and time-dependent manner. This increase occurs within the physiological range for GH and requires ongoing transcription and translation. Similar to hepatocytes, pulsatile administration of GH stimulates IGF-I synthesis in myocytes more effectively than continuous GH exposure. Moreover, GH also increases the phosphorylation of STAT3 and STAT5. The inhibition of JAK3, with the chemical inhibitor WHI-P154, altered GH action by decreasing the phosphorylation of STAT3 and the increase in IGF-I mRNA. Hence, these data suggest that the JAK/STAT signaling pathway plays a central role in the expression of IGF-I in response to GH in cultured myocytes, and that this cell line is an adequate model system to examine the impact of LPS and cytokines on GH action under in vitro conditions.

Subsequent studies by our laboratory investigated whether the LPS-induced increase in TNF-α in C2C12 myocytes is associated with a concomitant decline in IGF-I mRNA expression [52]. Incubation of myocytes with LPS decreased IGF-I mRNA, which was first evident at 4 h and sustained for at least 18 h. This decrease was temporally associated with an increased accumulation of TNF-α mRNA in C2C12 cells. Additional studies indicated that TNF-α produced a dose- and time-dependent decrease in IGF-I mRNA content and this response occurred in both myoblasts and differentiated myotubes. The reduction in IGF-I mRNA was not a generalized response by the cells, as evidenced by the lack of change of IGF-II mRNA. Treatment of myocytes with the protein synthesis inhibitor cycloheximide prevented TNF-α from decreasing IGF-I mRNA content. Furthermore, TNF-α was shown not to alter the stability of IGF-I mRNA as evidenced by the unchanged half-life of the message in cells treated with the transcriptional inhibitor DRB.

Incubation of C2C12 cells with GH increased IGF-I mRNA several-fold. Pretreatment of myocytes with TNF-α completely prevented the normal GH-induced increase in IGF-I mRNA [52]. However, as seen in muscle stimulated in vivo, TNF-α did not antagonize the ability of GH to stimulate the phosphorylation of either STAT3 or STAT5. Inflammatory cytokines, including TNF-α, activate multiple mitogen activate protein (MAP) kinases via their phosphorylation of serine and tyrosine residues. The ability of TNF-α to phosphorylate the MAP kinase ERK-1/2 and the JNK substrate c-Jun was assessed in myocytes, but TNF-α increased phosphorylation of only the later substrate. Additional experiments examined the ability of relatively selective MAP kinase inhibitors to prevent TNF-induced changes in GH action in C2C12 cells. Myocytes were pretreated with either PD98059 [a MAPK kinase (MEK)-1 inhibitor], SB202190 (a p38 kinase inhibitor) or SP600125 (a JNK inhibitor). Of these compounds, only the JNK inhibitor blocked the TNF-induced decrease in IGF-I mRNA expression. A number of transcription factors, including AP-1 and c-Jun, are phosphorylated by JNK. Further studies are needed to examine whether these transcription factors can either activate or inhibit IGF-I promoter activity in C2C12 cells to determine their potential role in modulating IGF-I gene transcription in skeletal muscle.

Summary

Muscle protein balance is regulated by the integrated effects of various hormones, cytokines, and growth factors. IGF-I clearly occupies a predominant role in regulating both muscle protein synthesis and degradation under basal and stress conditions. A reduction in the circulating concentration of IGF-I is a common manifestation of many catabolic conditions, including infection, trauma, AIDS, cancer, chronic alcohol abuse, and uremia. This decrease is in large part produced by a reduction in the basal rate of IGF-I synthesis and secretion by the liver. In addition, several catabolic states, typified by sepsis and endotoxemia, lead to the development of hepatic GH resistance. This resistance is evidenced by a decreased accumulation of hepatic IGF-I mRNA and a smaller increment in the prevailing plasma concentration of IGF-I in response to GH. The diminished GH action in liver results from an impairment in the classical JAK2/STAT5 signal transduction pathway that is mediated, either directly or indirectly, by the overproduction of inflammatory cytokines. It is increasingly evident that the autocrine and paracrine effects of IGF-I are important in controlling protein balance in skeletal muscle. In this regard, sepsis and LPS also markedly attenuate the ability of GH to increase IGF-I mRNA accumulation in muscle. However, in contrast to liver, this impairment is mediated independent of a defect in STAT5 phosphorylation. This STAT5-independent GH resistance can be mimicked in the C2C12 muscle cell line by the addition of TNF-α and, importantly, can be reversed by inhibition of the JNK pathway. Therefore, catabolic insults, especially those associated with an increased innate immune response, produce GH resistance in both liver and muscle via different mechanisms in response to the enhanced production of TNF-α. Because of the ability of TNF-α to stimulate the synthesis of other cytokines, particularly IL-1 and IL-6, it is not possible to yet exclude these other inflammatory cytokines as important modulators of GH resistance or the decrease in IGF-I. Furthermore, IGF-I transcription can be regulated by factors other than those described in this overview [53]. Regardless of the exact mechanism, the interplay among these various factors results in a net decrease in IGF-I availability to skeletal muscle and thereby impairs muscle protein synthesis via a reduction in translation initiation (Fig. 1). Continued research in this area is necessary to provide a better understanding of the numerous factors influencing the IGF system and cellular metabolism. This knowledge will help to realize the full potential and to avoid possible pitfalls of anabolic agents used in the management of critically ill patients.

Acknowledgements

This work was supported in part by National Institutes of Health grant GM38032. Additionally, we thank our collaborators Drs. Robert Cooney, Gladys Yumet, and Margaret Shumate.

Copyright information

© IPNA 2004