Heat Shock Protein 72 Expressing Stress in Sepsis: Unbridgeable Gap between Animal and Human Studies—A Hypothetical “Comparative” Study

Heat shock protein 72 (Hsp72) exhibits a protective role during times of increased risk of pathogenic challenge and/or tissue damage. The aim of the study was to ascertain Hsp72 protective effect differences between animal and human studies in sepsis using a hypothetical “comparative study” model. Forty-one in vivo (56.1%), in vitro (17.1%), or combined (26.8%) animal and 14 in vivo (2) or in vitro (12) human Hsp72 studies (P < 0.0001) were enrolled in the analysis. Of the 14 human studies, 50% showed a protective Hsp72 effect compared to 95.8% protection shown in septic animal studies (P < 0.0001). Only human studies reported Hsp72-associated mortality (21.4%) or infection (7.1%) or reported results (14.3%) to be nonprotective (P < 0.001). In animal models, any Hsp72 induction method tried increased intracellular Hsp72 (100%), compared to 57.1% of human studies (P < 0.02), reduced proinflammatory cytokines (28/29), and enhanced survival (18/18). Animal studies show a clear Hsp72 protective effect in sepsis. Human studies are inconclusive, showing either protection or a possible relation to mortality and infections. This might be due to the fact that using evermore purified target cell populations in animal models, a lot of clinical information regarding the net response that occurs in sepsis is missing.


Introduction
Sepsis is an inflammation-induced syndrome resulting from a complex interaction between host and infectious agents. It is considered severe when associated with acute organ dysfunction, which accounts for the main cause underlying sepsis-induced death. Despite increasing evidence in support of antioxidant [1], anti-inflammatory [2], or immuneenhancing [3] therapies in sepsis, recent studies failed to establish a correlation between antiseptic pathway-based therapies and improvement of sepsis [4] or septic shock [5] or among immune-competent patients [6].
Rapid expression of the survival gene heat shock protein 72 (Hsp72) was shown to be critical for mounting cytoprotection against severe cellular stress, like elevated temperature [7]. Intracellular Hsps are upregulated in cells subjected to stressful stimuli, including inflammation and oxidative stress exerting a protective effect against hypoxia, excess oxygen radicals, endotoxin, infections, and fever [8]. Recent studies imply that different biological disease processes and/or simple interventions may interfere with high temperature stress, leading to different clinical outcome in patients with and without sepsis [9]. In septic patients, administration of antipyretics independently associated with 28-day mortality, without association of fever with mortality [9]. Importantly, fever control using external cooling was safe and decreased vasopressor requirements and early mortality in septic shock [10].
Inducible Hsp72 is also found extracellularly where it exhibits a protective role by facilitating immunological responses during times of increased risk of pathogenic challenge and/or tissue damage [11]. Experimental data provide important insights into the anti-inflammatory mechanisms of stress proteins protection and may lead to the development of a novel strategy for treatment of infectious and inflammatory disorders [12]. However, although overexpression of stress proteins signals danger to inflammatory cells and aids in immune surveillance by transporting intracellular peptides to immune cells [13], it has also been linked to a deleterious role in some diseases [14]. In addition, serum Hsp72 levels were shown to be modulated according to the patient oxidant status whereas increased serum Hsp72 was associated with mortality in sepsis [15].
The purpose of this basic research-related review in critical care is to document the available evidence on the role of Hsp72 in sepsis, reporting both the state of the art and the future research directions. It might be that potential therapeutic use of stress proteins in prevention of common stress-related diseases involves achieving optimal balance between protective and immunogenic effects of these molecules [16]. In this review, we will attempt to classify experimental and clinical studies on Hsp72 in sepsis and to compare their results on inflammation, organ function, and outcome; we will also briefly discuss the mechanisms on how stress proteins might exert their protective or negative role in the disease development and highlight the potential clinic translation in the research field.

Materials and Methods
Human or animal in vivo or in vitro studies examining the beneficial effect of intra-or extracellular Hsp72 expression in sepsis were included in this study. The PRISMA [17] search method for identification of studies consisted of searches of PubMed database (1992 to September 2012) and a manual review of reference lists using the search term: "Hsp70 or 72. " The search output was limited with the search filter for any of: sepsis; severe sepsis; bacterial lipopolysaccharide (LPS); endotoxin. References in selected studies were examined also. The title and abstract of all studies identified by the above search strategy were screened, and the full text for all potentially relevant studies published in English was obtained. The full text of any potentially relevant studies was assessed by five authors (DMF, EB, IP, AK, and TT). The same authors extracted data from the published studies.

Statistical Analysis.
Proportions of methods used and results findings were compared by the 2 test. A two-sided alpha of 0.05 was used for statistical significance. The results were analyzed using SPSS software (version 20.0, SPSS, Chicago, IL, USA).

Results
Our search identified 411 PubMed titles and abstracts. After excluding duplicates, studies with no original data, or data insufficient to evaluate or those whose outcome was ischemia/reperfusion injury or others, 55 articles were finally included for analysis. The aim of this minireview was not to examine the quality of studies, but to describe induction methods and to compare in vivo and in vitro methods and results regarding a potential protective role for Hsp72 in human and animal sepsis.

Discussion
Hsps70 are emerging as powerful dichotomous immunemodulatory molecules that can have stimulatory and inhibitory effects on immune responses [63]. In our hypothetical "comparative study" model, we found that the balance between Hsp72 promotion and control of inflammatory responses and sepsis outcome differed unpredictably between human and animal studies. Clinical studies were inconclusive, showing either a low probability of protection (16.7% among septic patients) or even a possible relation to mortality and infections. In contrast, almost all (94.7%) septic animal in vivo and in vitro studies showed a biochemical, biological, and clinical protective effect for Hsp72 in sepsis. This might be due to the fact that using evermore purified target cell populations to provide insight into the direct effects of molecules on cells, a lot of clinical information regarding the net response that occurs in vivo is missing [63].

Stress Proteins Induction.
Sepsis, endotoxin tolerance, and heat shock all display downregulation of innate immunity, sharing a common immune suppressive effect, possibly through HS factor 1 (HSF1) mediated competitive inhibition of nuclear factor kappa-B (NF-B) binding [45]. It has been shown that multiple chaperones or cochaperones, including Hsp72, tend to form a complex with HSF1 monomers [64]. Once a cell is exposed to stress, these chaperones and cochaperones bind to denatured and damaged proteins, thereby "releasing" the nonactive HSF1 monomers to subsequently undergo homotrimerization [65]. However, while homotrimerization is sufficient for DNA binding and nuclear translocation, the magnitude and duration of transcriptional activity are regulated by inducible phosphorylation of specific serine residues of HSF1 by several protein kinases (Erk1/2, glycogen synthase kinase, protein kinase C) [64]. Once inside the nucleus, HSF1 binds to a heat shock element (HSE) in the promoter of Hsp genes, which is defined by a tandem repeat of the pentamer nGAAn arranged in an alternating orientation either "head to head" (e.g., 5 -nGA-AnnTTCn-3 ) or "tail to tail" (e.g., 5 -nTTCnnGAAn-3 ) [66], resulting in the upregulation of stress protein gene expression [67]. Thus, the intracellular accumulation of denatured or improperly folded proteins in response to stress is believed to be the universal signal resulting in the stressinduced gene expression of stress proteins [68,69] which is proportional to the severity of the stress [70]. Besides the innate immune response stress proteins seem to activate  (1) Hsp70 increased (1) prevented by quercetin (1) Enhanced TNF-(1) increased oxidative activity, inhibited apoptosis (1) Inconclusive (1)  Protective (2) also the adaptive immune response [71]. Thus, they have the capacity to elicit a pathogen-specific immune response [72] and to mediate the induction of peptide-specific immunity, eliciting potent T cell responses against the chaperoned peptide [73].

Experimental Hsp72
Studies. Hsp72 is the most highly induced stress protein in cells and tissues undergoing the stress response [74] and is central to the cytoprotective properties in patients with a variety of critical illnesses [52] or injuries [75]. Cell cycle components, regulatory proteins, and proteins in the mitogenic signal cascade may be protected by the molecular chaperone Hsp72 during periods of stress, by impairing proteasomal degradation of IkappaBalpha (I Ba) [47]. In addition, binding of Hsp72 to the Ser/Thr protein kinase IRE1a enhances the IRE1a/X-box binding protein XBP1 signaling at the endoplasmic reticulum and inhibits endoplasmic reticulum stress-induced apoptosis [76]. Thus, increased expression of Hsp72 by gene transfer/transfection has been demonstrated to confer protection against in vitro toxicity secondary to lethal hyperthermia [77], endotoxin [78], nitric oxide [79], hyperoxia [80], lung inflammation and injury [81], and in vivo ischemia-reperfusion injury [82]. On the contrary, microinjection of anti-Hsp72 antibody into cells impaired their ability to achieve thermotolerance [83]. We showed that in septic animal models, all reported Hsp72 induction methods increased intracellular Hsp72; this was associated with reduced proinflammatory cytokines, decreased organ damage, clinical improvement, and enhanced survival. Analysis of reviewed studies showed differed methodology approaching the Hsp72 biological and/ or genetic implication in the sepsis process.

Transgenic Animals.
When challenged with systemic endotoxin, HSF1-deficient [84] or Hsp722 −/− mice [49] had increased apoptosis and mortality compared to wild-type (WT) mice. Hsp72 expression was also required for Gln's protective effects on survival and tissue injury [34], an effect not seen in Hsp72 −/− mice [85]. On the contrary, using transgenic mice overexpressing the human Hspa12b gene, Hsp72 attenuated the endotoxin-induced cardiac dysfunction and leucocyte infiltration into the myocardium [51].

Exogenous Hsp72.
At the cellular level, Hsp72 preparations not only inhibited LPS-induced reactive oxygen species production and decreased NO expression in macrophages, but they also partially normalized the disturbed neutrophil apoptosis [37]. Prophylactic administration of exogenous human Hsp72 normalized inflammatory responses [38], limited host tissue damage [48], and reduced mortality rates [39]. Liposomal transfer of Hsp72 into the myocardium abolished LPS-induced contractile dysfunction [44], reduced mortality rates, and modified hemostasis and hemodynamics [40]. Intestinal Hsp72 overexpression reversed toll-like receptor (TLR)-4-induced cytokines and enterocyte apoptosis and prevented and treated experimental necrotizing enterocolitis [50]. Thus, mammalian Hsp72 appears to be an attractive target in therapeutic strategies designed to stimulate endogenous protective mechanisms against many deleterious consequences of septic shock by accelerating the functional recovery of susceptible organs in humans [40,90].

Glutamine.
Although Gln has little effect under basal conditions [43], endotoxin-treated animals given Gln exhibited dramatic increases in tissue Hsp72 expression [26], marked reduction of end-organ damage [28], attenuation of cytokine release [41] and peroxide biosynthesis, and improved vascular reactivity [29] associated with a significant decrease in mortality [91]. The molecular mechanism of Gln-induced Hsp72 expression appears to be mediated via enhancement of O-linked -N-acetylglucosamine modification and subsequently to increased levels of endonuclear HSF1 expression [43] and HSF1 transcription activity [42].
In a recent study, septic mice with Gln administration showed less severe damage to the kidneys and exhibited decreased HMGB1 and TLR4 in kidney tissues [35]. In Gln-treated rats, lung Hsp72 and HSF1-p expressions were enhanced [32,92], lung HMGB1 expression and NF-B DNAbinding activity were suppressed, and ARDS was attenuated and survival improved [33]. By inducing Hsp72, Gln attenuated LPS-induced cardiomyocyte damage [42] and left ventricular dysfunction [27] whereas Gln-treated sheep had a greater increase in myocardial Hsp72 immunoreactivity without aggravating the hyperdynamic circulation after endotoxemia [31]. In a rat brain model of endotoxemia, Gln upregulated the expression of Hsp72 and decreased the magnitude of apoptosis by inhibiting the translocation of NF-B from the cytoplasm to the nucleus [30].

Hyperthermic Heat Shock.
Subjected to a brief hyperthermic heat shock, Hsp72 conferred protection against sepsis-related circulatory fatality via inhibition of iNOS gene expression through prevention of NF-B activation in cellular processes that included prevention of I B kinase activation [25] and inhibition of I B degradation [20]. Also, Hsp72 induction by thermal pretreatment [21] attenuated proinflammatory cytokines [22] and improved survival in the LPSinduced systemic inflammation model, potentially involving Hsp-mediated inhibition of HMGB1 secretion [23]. A HS response induction of Hsp72 mRNA and protein expression in the lung has been shown to be associated with reduced lung injury [18], improved lung function [93], and survival [94].
Heat shock pretreatment could also attenuate the electrocortical dysfunction in rats with LPS-induced septic response, suggesting that HS induced Hsp72 might potentially be used to prevent septic encephalopathy in sepsis [24]. Similarly, HS treatment led to Hsp72 overexpression and preserved the expression of the enzyme mitochondrial cytochrome c oxidase complex associated with the minimization of ultrastructural deformities during sepsis [19]. Interestingly, Gln increased DNA binding of HSF1 in HS cells but in its absence ornithine was able to rescue the heatinduced DNA binding of HSF1 [43].

Human Studies.
Although the release of the Hsp72 in sepsis serves as a host impending danger signal to neighboring cells and might exert a cytoprotective function at low serum levels, it might also potentiate an already active host immune response leading to poor outcome once a certain critical threshold is attained. Such a sensitive balance could be an explanation of the surprising finding of this study, showing that only 16.7% of the 6 human septic studies demonstrated an Hsp72 protective effect compared to 95.8% protection shown in the 41 septic animal studies. In addition, by experimentally studying healthy individuals rather than patients in a real clinical setting, human studies mix up mild molecular reactions to stress with severe infectious systemic inflammatory response syndrome (SIRS), being thereby unconvincing and unable to verify results of experimentally controlled septic animal models.

Intracellular Hsp72: In Vitro Studies (Cell Models).
Human in vitro studies, mainly examining intracellular Hsp72 expression in hPBMC or hPMNL in patients and healthy individuals by using HS, Gln, exogenous Hsp72, and Hsp72 inhibitors or agonists, are inconclusive [57]. Thus, although Gln infusion altered neither endotoxin-induced systemic inflammation nor early expression of Hsp72 in isolated PBMCs in healthy volunteers [53], inducibility of ex vivo Hsp72 was impaired in peripheral blood lymphocytes of patients with severe sepsis [95], possibly contributing to immune dysfunction of T and B lymphocyte responses in resisting infection in severe sepsis [56].
Enhanced Hsp72 response in endurance-trained individuals, however, improved heat tolerance through both anti-inflammatory and antiapoptotic mechanisms [58]. Also, rHsp72 preconditioning ameliorated reactive oxygen species, TNF , and CD11b/CD18 adhesion receptor expression after lipoteichoic acid addition [39]. Sepsis was shown to enhance expression of iHsp72 in PBMCs correlated to plasma TNF concentrations [54] and in activated PMNLs, in which oxidative activity was increased and apoptosis was inhibited [55]. Similarly, using various Gln doses, proinflammatory cytokine release could directly be attenuated in PBMCs through enhancement of Hsp72 expression [61]. Overexpression of Hsp72 attenuated NF-B activation and proinflammatory cytokine release [88,96], inhibited LPS-mediated apoptosis, and protected lung epithelial cells [80] and pulmonary artery endothelial cells from oxidant-mediated [97] and inflammation-induced lung injury [59].

Extracellular Hsp72: In Vivo Studies (Serum Hsp).
Although PBMC Hsp72 expression was shown to be markedly decreased in critically ill septic patients [56], a significant increase in serum Hsp72 levels was reported in children with septic shock [52]. Extracellular Hsp72, reflected by increased serum levels, was also evident in children with acute lung injury [81] or following cardiopulmonary bypass [98]. Results of a recent adult study also indicated that increased serum Hsp72 is associated with mortality in sepsis [15]. Worse outcome associated with extracellular Hsp72 has also been reported in coronary artery disease [99], liver disease [90], sickle cell disease vasoocclusive crisis [100], and preeclampsia [101].
Heat shock proteins are markedly induced in response to a diverse range of cellular insults, being a reliable danger marker of cell stress [102]. Thus, extracellular Hsps act as a "danger signal, " activating immune-competent cells through LPS TLR4/CD14-dependent signaling [103]. According to the "danger hypothesis, " the release of stress proteins from severely stressed or damaged cells serves as a host impending danger signal to neighboring cells [104]. They are released in a nonspecific manner from dying, necrotic cells [105] or from viable cells release in a specific and inhibitable manner [106,107]. Using viable cell counts and lactate dehydrogenase the release of Hsp72 was shown to not be due to cellular damage [60]. Recent studies suggest that Hsp72 is actively released via an exosome-dependent nonclassical protein secretory pathway, possibly involving lysosomal lipid rafts [108]. Immune cell receptors capture Hsps released from necrotic cells or Hsp-containing exosomes [109], and receptor engagement by Hsp72 increases dendritic cell production of TNF , IL-1b, IL-6, and chemokine [110]. The host innate immune response occurs through a NF-B-dependent proinflammatory gene expression via TLR4 and TLR2 [111], similar to a LPSmediated signal transduction [112].

Factors Influencing Heat Shock Proteins Protective Role in Sepsis.
Recent work demonstrated that febrile-range temperatures achieved during sepsis and noninfectious SIRS correlated with detectable changes in stress gene expression in vivo (whole blood messenger RNA), thereby suggesting that fever can activate Hsp72 gene expression and modify innate immune responses [113]. Hsp72 serum levels may also be modulated according to the patient oxidant status [15] and prevent excessive gut apoptosis and inflammation in an age-dependent response to sepsis [49]. Importantly, Hsp72 inhibited LPS-induced NO release but only partially reduced the LPS increased expression of iNOS mRNA and exhibited LPS-induced NF-B DNA binding and LPS tolerance; in contrast, HS inhibited LPS-induced NF-B and HSF1 activity whereas HSF1 inhibited NF-B DNA binding [45]! A significant body of preexisting literature has hypothesized a relationship between Hsp72 expression and Gln's protection in both in vitro and in vivo settings [32,43,62,114,115]. Pioneer studies showed that Gln supplementation could attenuate lethal heat and oxidant injury and increase Hsp72 expression in intestinal epithelial cells [116][117][118]. Compared, however, with whey protein supplementation in a randomized, double-blinded, comparative effectiveness trial, zinc, selenium, Gln, and intravenous metoclopramide conferred no advantage in the immune-competent population [6]. In addition, we recently showed that although apparently safe in animal models (pups), premature infants, and critically ill children, glutamine supplementation did not reduce mortality or late onset sepsis [119]. Methodological problems noted in the reviewed randomized experimental and clinical trials [119] should therefore be seriously considered in any future well-designed large blinded randomized controlled trial involving glutamine supplementation in severe sepsis.
Drug interactions were also shown either to suppress Hsp72 protective effects exacerbating therefore drug-induced side effects or to induce Hsp72 beneficial effects by suppressing drug-induced exacerbations. Thus, it was recently shown that bleomycin-induced pulmonary fibrosis is mediated by suppression of pulmonary expression of Hsp72 whereas an inducer of Hsp72 expression, such as geranylgeranylacetone, could be therapeutically beneficial for the treatment of gefitinib-induced pulmonary fibrosis [120].
Finally, critically ill patients display variable physiologic responses when stressed; gene association studies have recently been employed to explain this variability. Genetic variants of Hsp72 have also been associated with the development of septic shock in patients [121,122]. Thus, the specific absence of Hsp72.1/3 gene expression can lead to increased mortality after septic insult [85].

4.5.
Limitations of the Study. The major problem that limits the comparability with human sepsis is the fact that in most cases of animal models, various forms of preconditioning were employed. This approach is nonspecific, and only a minor amount (about 10%) used genetically modified animals. Accordingly, important differences between cell and/or animal models versus clinical studies have been noted several times with various inflammatory pathways and have been written about extensively in the literature [123,124]. To the best of our knowledge, however, such discrepancies have not been summarized in detail in the context of Hsp72 and sepsis; in our opinion, these findings might be helpful for cautiously interpreting experimental data in the critical care field.

Conclusions
Heat shock proteins are molecular chaperokines that prevent the formation of nonspecific protein aggregates and exhibit sophisticated protection mechanisms. Experimental studies have repeatedly shown a strong molecular, biological, and clinical protective effect for Hsp72 in sepsis. Once again, clinical studies are inconclusive, varying from a protective in vitro effect to an in vivo Hsp72-mortality association. Possible influences by severity of disease-related factors, genetic variants, oxidant status, and unpredictable interventions such as those of temperature control, nutritional (glutamine) immune-enhancing, or drug intervening effects may unpredictably influence the Hsp72 protection efficacy in sepsis. Our "comparative" study data demonstrate that cellprotection with exogenous Hsp72, Hsp72 genes, heat stress, or glutamine is associated with induction of Hsp72 and that new Hsp72 targeted pharmaconutrition may be an approach to activating the preconditioning response in sepsis in clinical practice. However, as this hypothetical study suggests, much more work is needed to clarify the cellular and molecular mechanisms in which Hsp72 signals "danger" and regulates immune function in response to sepsis.