INTRODUCTION

Hemorrhagic shock is the leading cause of death and complications in combat casualties and civilian trauma. It has been shown to cause systemic inflammatory response syndrome, multiple organ dysfunction, and multiple organ failure 1. It is an old problem and to find useful remedies that are capable of reducing its casualty is a most important task in combat medicine and trauma injury 2. In this paper, the role of inducible heat shock protein 70 kD (HSP-70i) and inducible nitric oxide synthase (iNOS) in hemorrhage/resuscitation-induced injury is discussed in details.

EFFECTS OF HEMORRHAGIC SHOCK

Hemorrhage/resuscitation is a common scenario in the clinical practice of medicine and occurs in virtually all organ systems. Its hallmarks are relatively consistent across patient populations and organ systems. They lead to systemic inflammatory response syndrome, multiple organ dysfunction, and multiple organ failure 1. A variety of biomolecules is known to be involved in this response. In rodents, increases in HSP-70i by heat stress limit tissue injury such as tissue integrity, polymorphonuclear neutrophil (PMN) infiltration and leukotriene B4 (LTB4) generation caused by ischemia reperfusion 3,4, hemorrhage 5, or hemorrhage plus resuscitation 6. In transgenic mice expressing HSP-70i, protection against myocardial dysfunction after a brief ischemia was reported 7. Likewise, inhibition of nitric oxide (NO) production results in significant reduction of local tissue damage, PMN infiltration, and LTB4 generation caused by ischemia/reperfusion 8. Mice deficient in inducible NO synthase (iNOS) also demonstrate limited hemorrhage/resuscitation-induced injury 9,10. Therefore, remedies that induce HSP-70i and/or inhibit iNOS might prove very useful for reducing hemorrhage/resuscitation-induced injury in man.

The above observations are consistent with the idea that the low oxygen supply resulting from conditions such as ischemia or hemorrhage affects the expression of iNOS, which then influences the expression of other proteins that alter cell viability. It has been shown that hypoxia results in alteration of iNOS, Bcl-2, and P53 mRNA expression in cultured human intestinal epithelial T84 cells and Jurkat T cells 11,12, alterations that can be modulated by treatment with a NOS inhibitor 11. It is also found that hypoxia increases the activity of caspase-3, an aspartate-specific cysteinyl protease involved in apoptosis, an activity that is blocked by NOS inhibitors 11.

A full time-course study of the effect of hemorrhage on a series of stress-related proteins such as c-JUN, Kruppel-like factor 6 (KLF6), iNOS, HSP-70i, and hypoxia inducible factor 1α (HIF-1α) has been reported in a hemorrhage mouse model 13. Based on Western blot data obtained from mouse lung, jejunum, heart, kidney, liver, and brain, c-JUN, KLF6, iNOS, HSP-70i, and HIF-1α are up-regulated (Fig. 1). In jejunum, c-JUN protein was over-expressed within 1 h, but levels returned to baseline values 3 h later. KLF6 began to increase significantly 6 h later, reached the maximum at 24 h, and remained at that level at 48 h after hemorrhage. iNOS increased at 6 h, reached a maximum at 12 - 24 h, and returned to baseline values at 48 h. HSP-70i increased at 12 h and remained at elevated levels at 48 h. HIF-1α also increased at 12 h and continued to increase at 48 h. The sequence of protein appearance was c-JUN, KLF6, iNOS, HSP-70i, and HIF-1α. KLF4 (a repressor to iNOS, 12 and c-FOS (an AP-1 protein) were not detected, and NF-κB 65 kD was detected but not affected by hemorrhage.

Figure 1
figure 1

Time course study of stress-related proteins in hemorrhaged mouse jejunum. The albino mouse was subjected to a cardiac puncture and its 40% of total blood was removed. The mouse was then allowed for various periods of time to respond to the hemorrhage before a small portion of jejunum was dissected out for immunoblotting analysis 13. The sham-operated mice were exposed to the same procedure except blood withdrawal. *P < 0.05 vs the data observed at time = 0 h of post-hemorrahge.

Full size image

A similar time-course observation on these stress-related proteins was also obtained in mouse lung, kidney, liver, and heart. Sham-treated mouse organs displayed no changes in the basal levels of c-JUN, KLF6, and iNOS. In addition to changes in this series of stress-related proteins listed above, hemorrhage also increases cellular caspase-3 activity 14, reduces cellular ATP levels 14,15,16,17,18, and elevated mRNA of p38-MAPK, p53, and Bcl-2 18. Fig. 2 shows that 5-lipoxygenase, lipid peroxidation, cycloxygenase, cNOS, LTB4, PGE2, interleukins, tumor necrosis factor-α, caspases, and complement activation occur relatively at different time points after hemorrhage 3,4,5,6,9,10,19. The sequence of their occurrence provides the useful infor-mation for studying the mechanism(s) underlying the hemorrhage-induced injury as well as therapeutic targets to prevent or ameliorate the injury.

Figure 2
figure 2

Hemorrhage increases stress-related protein expression and activity at different time points after hemorrhage. Hemo: hemorrhage; 5-LO: 5-lipoxygenase; cNOS: constitutive nitric oxide synthase; COX: cycloxygenase; PGE2: prostaglandin E2, LTB4: leukotriene B4; iNOS: inducible nitric oxide; TNF-α: tumor necrosis factor-α.

Full size image

HEAT SHOCK PROTEINS

Heat shock proteins (HSPs) are present in most cells. They represent multi-gene families that range in molecular size from 10 to 174 kD. Based on their molecular weights, they are divided into families of HSP-10, HSP-20, HSP-40, HSP-60, HSP-70, HSP-90, and HSP-110. Tab. 1 lists the members of each family, their locations, and functions in the cell 20,21,22,23.

Table 1 Members of HSPs and their locations and functions 20,21,22,23
Full size table

Structure of HSP-70i and its functions

The family of HSP-70 contains HSP-72, HSP-73, HSP-75, and HSP-78. HSP-72 is an inducible form and also abbreviated as HSP-70i. The others are constitutive. The molecular structure of HSP-70i is very similar to that of HSP-73. They both contain a globular unit linked to a β-sheet and an α-helical tube. The globular unit is the ATPase domain (1-386 aa, 44 kD); the β-sheet and the α-helix tube (384-543 aa, 18 kD) are the peptide binding domain; and the tail (542-646 aa, 11 kD) of HSP-73 has the signal peptide of EEVD, but not the tail (542-640 aa, 10 kD) of HSP-70i. Proteins in the cytoplasm recognize the EEVD of HSP-73 as a chaperone protein, whereas HSP-70i has no such EEVD as a chaperone but instead acts as a cytoprotectant. It should be noted that there is a distinct difference between being a chaperone and a cytoprotectant. A chaperone protein is involved in bringing a protein to the targeted location for its specified purpose, whereas cytoprotection usually involves the modulation of particular protein activity or overall cell functionality thereby leading to cell protection against injury or cell death. The functions and regulations of HSP-73, -75, and -78 have been reviewed in detail 20.

HSP-70 gene regulation involves phosphorylation, heat shock transcriptional factor, heat shock elements, and ions. Each of these components is described briefly below.

Heat shock elements

Genomic footprinting of the human HSP-70 promoter has revealed that heat stress induces a quick binding of heat shock transcriptional factors (HSFs) to a region encompassing five nGAAn sequences named heat shock elements (HSEs), three of them are GAA (also called perfect HSEs ) and other two are GAC and GGG (also called imperfect HSEs). The sequences of HSEs at site 3 and site 4 are dyad symmetrical, and HSFs preferably bind to HSEs at sites 3 and 4 20. In mammalian cells, HSEs are usually bound by HSF4 to keep other HSFs away.

Heat shock transcriptional factors

Four different HSFs have been identified in vertebrate: HSF1, HSF2, HSF3, and HSF4 24,25. HSF2 has two isoforms due to splicing: HSF2A and HSF2B 26. HSFs have a binding domain, a helical trimerization surface, and a short conserved element. The binding domain is to bind protein and HSEs; the trimerization surface has leucine zipper coiled-coil motifs for trimer formation. There are two genes in mice and humans encoding HSFs. The sizes for human HSF1, HSF2, HSF4, and chicken HSF3 are 529 aa, 510 aa, 463 aa, and 467 aa, respectively. Human HSF1 can be activated by various stimuli, human HSF2 only by hemin, chicken HSF3 only by heat, while human HSF4 is not activated.

It has been reported that mouse HSF1, chicken HSF3, and mouse HSF4 bind to all 5 HSEs, whereas mouse HSF2 interacts with only 4 HSEs. Activation of HSF1, 2, or 3 up-regulates HSP-70i 20. In contrast, HSF4 is responsible for down-regulation of HSP-70i 25. During the process of HSF trimerization, homotrimers or heterotrimers have been found and influence the overall expression of HSP-70i 26.

HSP-70i Autoregulation

Normally, HSFs that reside in the cytosol of mammalian cells are bound to HSPs under unstressed conditions. Under stress conditions such as hemorrhage, HSFs are separated from HSPs. Then, these HSFs are available for phosphorylation by protein kinase C or other serine/threonine kinases. They form homotrimers 27 or heterotrimers 26. The trimers enter the nucleus, bind to HSEs located on the promoter region of HSP genes, and become further phosphorylated by HSF kinases. Transcription is then initiated and translation is resulted. Therefore, new HSP-70i is synthesized in the cytosol of cells. Soon, the elevated HSP-70i is bound by cytosolic HSFs. The complex of HSP-70i and HSFs turns off the further increase in HSP-70i. The steps of HSFs phosphorylation, trimerization, and its translocation from the cytosol to the nucleus are Ca2+-dependent 28.

Protein Regulation of HSP-70i

HSP-70 family is present in every cell type and tissue under both unstressed and stressed conditions. HSP-70i is induced by physiological stressors, pathological stressors, and environmental stressors 20. The degree of its induction depends on the level and duration of exposure to stressors. The increase usually is transient, but how long it persists is different in various cell types, ranging from hours, days, or weeks 20. It can be regulated by hormones 29, intracellular pH 30, cellular cAMP levels 31, intracellular concentrations of Ca2+32,33, and activation of protein kinase C 33,34,35,36,37, protein kinase A 33,34,35,36,37,38 and protein tyrosine kinase 4,13,39,40,41,42,43,44. Aging is also known to result in over-expression of HSP-70i 45 that in turn desensitizes Ca2+ machinery and turns off new synthesis of HSP-70i 46,47,48. Bioflavonoid such as quercetin prevents HSF1 binding to HSEs, thereby leading to attenuation of HSP-70i gene expression 49.

Anti-convulsant and mood-stabilizing drugs such as valproic acid has been shown to up-regulate HSP-70i in the cortex and striatum of both ipsilateral and contralateral sides of middle cerebral artery occusion rats and of normal rats. The increased HSP-70i reduced infarct size and caspase-3 activation 50. Lithium, a drug primarily used to treat bipolar mood disorder also has been reported to up-regulated HSP-70i by an increasing DNA binding activity of HSF1 51. Like sublethal heat stress 20, sublethal hemorrhage induced significantly the HSP-70i over-expression 52,53.

NO AND NOS

Like ischemia and reperfusion or hypoxia 11, hemorrhage increases NO production and NOS overexpression 13. The substrate of NO is L-arginine with O2 and the enzymes involved in this process are either constitutive NOS (cNOS) or iNOS.

L-arginine + O2 → L-citrulline + NO

The chemical biology of NO includes its direct effects and indirect effects 19,54. Tab. 2 shows that the direct effects, also named the primary mode of NO action, are NO reacts with heme-containing proteins. These reactions are generally rapid, require low concentrations of NO, and are the genesis of most of the physiological effects of NO such as vasodilation. In contrast, the indirect effects, namely the secondary mode of NO action, include formation of N2O3, ONOO−, NO2, and HNO that react with cellular targets and may result in a major configuration change in critical molecules. It has been shown that indirect effects require much higher concentrations of NO than direct effects. It appears that NO produced at low concentrations for short periods primarily mediates direct effects, whereas high local NO concentrations sustain over prolonged periods mediate indirect reactions and cause pathophysiological effects such as organ dysfunction and failure.

Table 2 NO chemical biology 19,54
Full size table

When NO reacts with NO2 resulting in nitrosative stress, while it does with O2− causing oxidative stress.

Nitrosative Stress —

Oxidative Stress —

NOS

NO production is mediated by NOS (Tab. 3). NOS can be divided into two major categories: constitutive form (cNOS) and inducible form (iNOS) 55,56. cNOS includes nNOS and eNOs and is Ca2+-dependent, whereas iNOS is not. Both cNOS and iNOS contain a reductase domain and an oxidase domain. However, eNOS has myristonylation and palmitonylation sites, whereas nNOS and iNOS do not. The iNOS promoter has a TATA site and differs in that from the cNOS promoters which are TATA-less promoters. The cNOS promoters are GC rich and are regulated primarily by Sp1 and other members of the Sp1-like family. In contrast, NF-κB plays a crucial role in the regulation of iNOS. In the murine iNOS promoter, the downstream NF-κB binding site (-76 to -85 bp) seems to be the most important one 57, however, the upstream NF-κB site (-974 to -960 bp) also seems to have some functionality and cooperativeness with the downstream site 58. For the human iNOS promoter, the NF-κB motif at -5.8 kb is most critical for cytokine-induced promoter activity, whereas the sites at -5.2, -5.5, -6.1, and -8.2 kb have a cooperative effect 59,60.

Table 3 Types of NOS and their characteristics 19,55,56
Full size table

cNOS is thought to generate low levels of NO at sub-micromolar range for short durations and perform physiological effect. iNOS generates NO for prolonged periods and at local concentrations as high as 1-5 μM and causes pathophysiological effects. Since the chemistry and biological outcome depend on the concentration of NO, the proximity of a biological target to the NO source becomes critical. For example, cells or tissues close to macrophages that produce high levels of NO will be subjected to direct and indirect effects due to the primary and the secondary modes of NO actions. In contrast, if they are far from the NO source, they will experience only direct effects as the primary mode of NO action.

Hemorrhage induces sequential increases in the levels of stress-related proteins c-JUN, cNOS, KLF6, iNOS and then HSP-70i and HIF-1α (Figs. 1 and 2). Both cNOS gene and iNOS gene get activated by hemorrhage and they are early response genes, though cNOS responds to hemorrhage earlier than iNOS. HSP-70i and HIF-1 are considered to be the late response genes and their overexpression probably play a role to control the post-hemorrhage tissue damage.

iNOS has previously been shown to be overexpressed after hemorrhagic shock in rodent lung 6,18,61 human liver 62 and murine tissues 13,62,63. Warke et al 12 reported that the binding of KLF6 to the iNOS promoter increased significantly in cultured cells after chemical hypoxia, heat stress, serum starvation, and phorbol 12-myristate 13-acetate and A23187 ionophore stimulation. Using GeneBank, we identified four CACCC sites on the mouse iNOS promoter (-253 to -257, -818 to -822, -856 to -860, and -1556 to -1600) that potentially bind KLF6. Furthermore, both the KLF6 promoter and iNOS promoter have AP-1 binding sites (KLF6: -364 to -371; iNOS: -644 to -650 and -1225 to -1231) that probably also bind c-JUN and/or c-FOS. In our study, no c-FOS was detected using immunoblotting analysis. Because iNOS promoters include two AP-1 sites and are involved in processes such as cell injury, wound repair, embryogenesis, tissue differentiation, and suppression of tumorigenesis, the observation that c-JUN overexpression occurs earlier than KLF6 and iNOS overexpression suggests two possibilities. 1) Hemorrhage activates c-JUN, which up-regulates KLF6 expression, leading to iNOS expression; or 2) hemorrhage activates c-JUN, which then directly binds to the iNOS promoter to induce iNOS expression.

Down-regualtion of iNOS

A series of NOS inhibitors have been designed and extensively studied 64. Among them, L-NNA (N-nitro-L-arginine), L-NMA (N-methyl-L-arginine), L-NIL [N-(1-iminoethyl)-L-lysine], and L-NIO [N-(1-iminoethyl)-L-ornithine] are shown to effectively inhibit iNOS gene expression, NO production, and caspase-3 activity 10,11.

In addition to the use of NOS inhibitors, iNOS deficiency can be caused by deletion of the iNOS gene. This has been shown to reduce or prevent hemorrhage-induced injury and death 9. This result further confirms that iNOS is responsible for the hemorrhage-induced injury. Therefore, down-regulation of iNOS by any means will be beneficial to patients suffering from hemorrhage.

In addition to treatment with iNOS inhibitors or iNOS gene deletion, compounds such as ansamycin can inhibit iNOS activity and gene expression. Both geldanamycin and 17-allylamino-17-demethoxygeldanamycin, members of the ansamycin family, have been demonstrated to reduce NO production, iNOS mRNA, and cytokine-induced activation of the iNOS gene promoter in cultured cells 49, in rats 6, and in mice 13.

INTERACTION BETWEEN HSP-70I AND INOS

Using immunoprecipitation and immunoblotting analysis, it has been found that HSP-70i forms a complex with iNOS and its transcription factor KLF6 after hemorrhage 13, probably by iNOS and KLF6 binding to the peptide binding domain of HSP-70i. No complex formation is detected between HSP-70i and p53 or Bcl-2 proteins, suggesting that HSP-70i specifically couples to iNOS and KLF6. Treatment with geldanamycin increases HSP-70i expression and decreases the hemorrhage-induced increase in iNOS expression. The complex formation between HSP-70i and iNOS is still observed. It is possible that the complex formed between HSP-70i and iNOS might decrease the enzymatic activity of iNOS and subsequently decrease NO production. As a result of a low level of NO, the morbid sequelae caused by hemorrhagic shock is markedly diminished.

Our preliminary data show that up-regulation of HSP-70i by HSP-70 gene transfection to human intestinal cells inhibited the hypoxia-induced increase in caspase-3 activity and reduced apoptosis. Likewise, treatment of human intestinal epithelial cells with iNOS inhibitors decreased caspase-3 activity and apoptosis 11. Taking these data together, one can conclude that the inhibition of HSP-70i on caspase-3 and apoptosis is via its inhibition on iNOS that leads to less NO production.

OTHER PROTECTIVE MECHANISMS

Other protective mechanisms of HSP-70i have also been proposed. For example, TNF-α overexpression caused by sepsis, endotoxemia, hemorrhagic shock, and ischemia/reperfusion 52,53,65 has been found. In vitro induction of HSP-70i in moncytes or macrophages inhibits TNF-α production following bacteria lipopolysaccharide stimulation, and in vivo induction of HSP-70i down-regulates tissue TNF-α production following an injurious insult 5,52,53,66,. The underlying mechanism of the inhibitory action of HSP-70i on TNF-α over-production induced by sepsis, endotoxemia, hemorrhagic shock, or ischemia/reperfusion is not clear. It is highly likely that over-expression of iNOS gene may be responsible for the TNF-α over-production via p38-MAPK overexpression 18. Since it is known that HSP-70i couples to iNOS and inhibits iNOS protein levels 13, a possibility of a down-regulation of iNOS resulting in decreases in TNF-α levels can not be ruled out.

Fig. 3 is a schematic representation of our model for the interaction between hemorrhage-induced changes, indicating points where HSP-70i might block these changes. Based on our data, KLF6 and HSP-70i are first normally associated with each other in some way. After hemorrhage, increases in c-JUN protein up-regulate expression of KLF6 and iNOS proteins and, somehow, caspase-3 activity and TNF-α expression. Overexpression of HSP-70i protein results in increased complex formation with KLF6 and decreased KLF expression. As a result, less KLF6 is available to bind to the iNOS promoter and less iNOS is expressed. Both caspase-3 activity and TNF-α are inhibited directly (through less availability of iNOS) or indirectly (through other intermediate proteins such as p38-MAPK 18,53. Further studies are needed to understand the underlying mechanism in detail.

Figure 3
figure 3

The Schematic representation of model for interaction between hemorrhage-induced changes and points where overexpressed HSP-70i might block changes. KLF6 and HSP-70i are normally associated with each other. Hemorrhage increases c-JUN that upregulates expression of KLF6 and iNOS proteins and subsequently probably increases caspase-3 activity and TNF-α. Overexpressed HSP-70i leads to increased complex formation between HSP-70i and both KLF6 and iNOS. Therefore, less free KLF6 available to bind iNOS promoter and less iNOS is expressed, which leads to inhibition of caspase-3 activity 11 and probably the TNF-α level. +: stimulation; −: inhibition; ↑: increase.

Full size image

RESUSCITATION FLUIDS

Currently, the major cause of death in potentially salvageable battlefield casualties is hemorrhage 67. About 20% of these deaths are preventable if the bleeding can be quickly controlled or minimized 68,69 and sufficient resuscitation fluid is administered in time to maintain critical tissue perfusion. However, it has been recognized that resuscitation fluids are not innocuous and that they may actually potentiate the cellular injury caused by hemorrhagic shock 70. It has been proposed that some resuscitation fluids may contribute to delayed multiple organ dysfunction 1.

To search for an optimal resuscitation fluid is still an ongoing research program. Questions have been raised particularly about the use of lactated Ringer fluid (LR). LR has been shown to increase neutrophil activation. Hemorrhage elevates mRNA levels of iNOS, KLF6, p-38 MAPK, p53, Bcl-2, and caspase-3 genes in rat lung and ileum. The LR resuscitation does not effectively block their increases. Instead it further increases some of them 18,71. Alam et al 72 also reported that 51 genes were altered in rats treated with hemorrhage plus LR resuscitation. Most importantly, resuscitation with LR but not whole blood is shown to restore the ATP loss and preserve the increased levels of p53 and Bcl-2 in rat lung caused by hemorrhagic shock 18, suggesting that LR may be a relatively good resuscitation fluid.

As mentioned above, resuscitation may actually potentiate the cellular injury caused by hemorrhagic shock 70 since delayed multiple organ dysfunction and failure occur and mortality results after resuscitation 1. Because iNOS is known to be responsible for the hemorrhage-induced injury and HSP-70i is shown to offer cytoprotection, it is likely that agents or remedies that can reduce iNOS expression and/or increase HSP-72 will be beneficial in conjunction with administration of resuscitation. iNOS inhibitors such as L-NIL, and L-NIO can be potentially useful for this purpose. Likewise, HSP-70i inducers such as herbimycin A 4,39,40,41,42,43,44,73, heat stress 20, ethanol 66,67,68,69, and geldanamycin 4,6,13,39,42, are also in the list for consideration.

CONCLUSIONS

Hemorrhage/reperfusion-associated pathophysiology is complicated and poorly understood. Many adverse effects of hemorrhage/resuscitation are common to ischemia/reperfusion and hypoxia/reoxygenation. Resuscitation eventually does not completely reverse the hemorrhage-induced changes. The complexity of the cellular response to hemorrhage complicates efforts to design approaches to treat or prevent injury resulting from resuscitation. Nevertheless, an additive in resuscitation fluids, which can induce HSP-70i, block iNOS, and restores ATP depletion, may be potentially therapeutic to salvageable patients.

Hemorrhage induces overexpression of iNOS and HSP-70i proteins, increases in cellular caspase-3 activity 14, and reduction in ATP 15,16,17,18. iNOS overexpression appears early and it leads to the NO production and its direct and indirect effects. HSP-70i overexpression appears 12 h after occurrence of hemorrhage 13. Because of its late appearance, it is possible that this HSP-70i is not serving its usual protective role in hemorrhage and resuscitation, but that it is rather facilitating tissue repair for salvaging the damage caused by hemorrhage and resuscitation. However, a pre-treatment to induce HSP-70i overexpression and inhibit iNOS induction prior to hemorrhage would be an ideal one to prevent the resulting injury. On the other hand, its post-treatment is still useful in order to minimize the morbid sequelae such as multiple organ dysfunction and failure.

PERSPECTIVE

Hemorrhage/resuscitation is a common scenario in the clinical practice of medicine and occurs in virtually all organ systems. Its hallmarks are relatively consistent across patient populations and organ systems. In our murine model 13,14,18, treatment with geldanamycin is known to increase HSP-70i and reduce iNOS, cellular caspase-3 activity and ATP loss. Though the underlying mechanism is not completely elucidated, geldanamycin treatment inhibits iNOS by inhibiting KLF6 and induces HSP-70i by activating HSF1. HSP-70i is shown to complex with iNOS and its transcription factor KLF6 and understanding the interaction between HSP-70i and iNOS may provide an insight for a therapeutic design to prevent the hemorrhage/resuscitation-induced injury. Taking together the data obtained from geldanamycin or agents such as ethanol 74,75,76,77, glutamine or crocetin 78, adjunct therapy with the optimal resuscitation fluid such as ATP-MgCl2 79 or ethyl pyruvate 80 may be a novel approach to address the problems raised from assaults under circumstances such as hemorrhage/resuscitation.