1
Supplement to the Review:Proteotoxic Stress and Circulating Cell Stress Proteins in the Cardiovascular Diseases by Brian Henderson and A. Graham Pockley
Guide to the Properties of Extracellular Cell Stress Proteins in the Cardiovascular System
Introduction
Thus supplement contains specific details of the biological actions of a range of cell stress proteins that have been found in the circulation of healthy individuals and patients with various cardiovascular pathologies and which are potentially of homeostatic importance. A number of other cell stress proteins found in the circulation are also described. These have biological actions which could be important in cardiovascular health and disease but have not yet been studied in this context.
Circulating Cell Stress Proteins with a Homeostatic Function
Protein Disulphide Isomerase (PDI): There are now at least five human PDIs (Ellgaard et al. 2005) and this protein can also be found on the outer cell surface, where it is enzymically active (Zai et al. 1999). It has also been shown to be a vascular endothelial cell hypoxic stress protein,which is significantly upregulated in cells exposed to hypoxia (Graven et al. 2002). One of the earliest reports of stress proteinrelease from cells was the release of PDI by activated platelets (Chen et al. 1992). Surprisingly, subsequent studies have concluded that PDI controls integrin-mediated platelet aggregation, adhesion and granule secretion via GPIIb/IIIa, suggesting this receptor as the target of the enzyme (Essex and Li, 1999;Jordan and Gibbins, 2006). As well as being a cell surface protein in platelets, PDI is also released by these cells and is involved in tissue factor-induced fibrin formation. Inhibition of PDI has been found to reduce fibrin generation, revealing that this protein can directly initiate blood coagulation (Reinhardt et al. 2008). In a later study it was shown that although platelets release PDI, it is the release of this protein from vascular endothelial cells which is essential for fibrin generation in the whole animal (Jasuja et al. 2010). The mechanism of action of the endothelial cell PDI is hypothesized to involve the control of surface expression of the phospholipid, phosphatidylserine (Popescu et al. 2010). So here we have a secreted stress protein which plays a homeostatic role in cardiovascular physiology. It is not known if excessive release of PDI could contribute to thromboembolic disease.
Chlamydia are hypothesised to be involved in the pathogenesis of atherosclerosis (Vainas et al. 2009). These are obligately intracellular bacteria, and it has been established that cell surface PDI in host cells is required both for bacterial attachment and for bacterial invasion (Abromaitis and Stephens, 2009).
Peroxiredoxins (Prdxs):There are six Prdx isoforms in mammals, which interact with thioredoxin and catalyze peroxide reduction of hydrogen peroxide, organic hydroperoxides and peroxynitrite (Hall et al. 2010). There are three major subclasses of the peroxiredoxins: typical 2-cysteine (2-Cys) Prdxs (Prdx1-4), atypical 2-Cys Prdx (Prdx 5) and 1-Cys Prdx (Prdx 6). The 2-Cys peroxiredoxins also have molecular chaperone activity (Kumsta and Jacob, 2009). In human erythrocytes, the peroxiredoxins undergo circadian redox cycles which are stable over prolonged times and are entrainable (i.e. can be modified by changing environmental conditions - O’Neill and Reddy, 2011). This finding may be of some cardiovascular importance as it is now realised that there is a circadian variation in the frequency of onset of various cardiovascular conditions (Shaw and Tofler, 2009; Martino and Sole, 2009).
Peroxiredoxin 1 is released from nucleated cells (Chang et al. 2006)and red blood cells secrete peroxiredoxin 2, which has been proposed to influence T cell proliferation(Antunes et al. 2011). In patients with sepsis, high serum levels of peroxiredoxin 4 are associated with greater severity of disease and a poorer outcome, suggesting this protein is a useful biomarker (Schulte et al. 2011). It is not clear if these higher levels are contributing to the septic reaction or are, inefficiently, combating it. Studies of the inactivation of the genes encoding peroxiredoxin 2 (Yang et al. 2007), peroxiredoxin 3 (Li et al. 2007) or peroxiredoxin 6 (Yang et al. 2010)in mice have revealed that a lack of these proteins exacerbates the pathology that is induced by LPS, suggesting that these peroxiredoxins have an anti-inflammatory homeostatic protective action. A similar finding has been made with bleomycin-induced pulmonary inflammation in peroxidoredoxin 1-deficient mice. The absence of this protein leads to a much aggravated lung pathology (Kikuchi et al. 2011). It is also reported that LPS enhances peroxiredoxin 1 production by mouse macrophages (Bast et al. 2010).
Evidence for a role for secreted peroxiredoxins in cardiovascular disease is, thus far, limited. Studies of leukocyte-vascular endothelial cell interactions in living peroxiredoxin 1-deficient mice revealed a significantly increased interaction of leukocytes with the vascular endothelial cell wall due to P-selectin upregulation. Furthermore, the rendering of ApoE-/- mice peroxiredoxin deficient significantly increasesthe development of atherosclerosis. These findings support the hypothesis that peroxiredoxin 1 protects the endothelium from activation and the development of atherosclerosis (Kisucka et al. 2008).
Studies of secreted levels of peroxiredoxins in patients with cardiovascular disease are just beginning. Analysis of the peroxiredoxin 1 levels in extracellular brain fluids of patients with acute stroke has found that levels of this protein increase 20-fold (p = 0.0001) suggesting that it plays some role in local pathophysiology (Dayon et al. 2011). In patients with systemic vasculitis, anti-endothelial cell antibodies (AECA) are implicated in blood vessel pathology (Alessandri et al. 2006). To determine what these circulating antibodies recognise on endothelial cells an immunoproteomic study was conducted using patient’s sera and surface proteins from human umbilical vascular endothelial cells (HUVECs). This identified 53 spots, with nine proteins being identified – one of which was peroxiredoxin 2. Measurement of circulating autoantibodies to peroxiredoxin revealed 60% of patients with vasculitis had such antibodies, compared with 0% in healthy controls. Titres of anti-peroxiredoxin 2 antibodies mirrored the disease activity during time course analysis of patients. Supporting the role of such antibodies in disease pathology was the finding that live endothelial cells express peroxiredoxin 2 on their outer cell surfaces. An antibody to peroxiredoxin 2 added to HUVECs stimulated the production of pro-inflammatory cytokines, growth factors and chemokines supporting the hypothesis that peroxiredoxin 2 is a major autoantigen in systemic vasculitis and antibodies to this protein are a useful biomarker of disease activity and progression (Karasawa et al. 2010). Another proteomic study has identified peroxiredoxin 1 as being released by the luminal areas in the intraluminal thrombi of patients with abdominal aortic aneurysm (AAA). It was further found that there were increased levels of peroxiredoxin 1 in the sera of AAA patients compared to controls and that levels of this protein correlated with AAA diameter, and with plasmin-antiplasmin and myeloperoxidase levels. In a prospective study of AAA patients, it was found that there was a positive association between serum levels of peroxiredoxin 1 and the expansion rate of the AAA and the combination of peroxiredoxin 1 levels and the size of the aneurysm show an additive predictivevalue (Martinez-Pinna et al. 2011). This suggests that peroxiredoxin 1 is a useful biomarker in AAA, although it is not clear if this protein has protective or pathological actions.
Thus the current evidence suggests that secreted peroxiredoxins play important role in controlling homeostatic tissue function in the cardiovascular tissues and associated leukocytes. Further analysis of the role of these proteins in cardiovascular pathology is urgently required. Intriguingly, it has been reported that the peroxireoxins have structural similarities to the chaperonin (Hsp)60 protein (Dekker et al. 2011) which, as will be described, has negative effects in cardiovascular disease.
Circulating Cell Stress Proteins with Cardioprotective Functions
Ubiquitin:A small intracellular protein which binds unfolded proteins in order to target them for proteosomal degradation (Lanneau et al. 2010). It is also secreted by cells and is found in serum/plasma at concentrations in the μg/ml range (Takada et al., 1997). It has a number of functions which might suggest that it has some action in controlling inflammation (Majetschak, 2010;2011). For example, exogenous ubiquitin inhibits LPS-induced pro-inflammatory cytokine production by human monocytes (Majetschak et al. 2003) and extracellular ubiquitin has been shown to bind to the lymphocyte chemokine receptor CXCR4 (Saini et al. 2010), a co-receptor for HIV. Experimental evidence suggesting a role for extracellular ubiquitin in cardiovascular disease includes the finding that the apoptosis of cardiac myocytes, induced by β-adrenergic receptor stimulation, is blocked by the ability of this agonist-receptor interaction to cause release of ubiquitin which inhibits myocyte apoptosis (Singh et al. 2010) and that ubiquitin is secreted from the arteries of hypertensive rats (Delbosc et al. 2008). Another possible function of extracellular ubiquitin, which may play a role in controlling cardiovascular pathology, is the potential of this protein to interact with and inhibit the actions of DAMPs (Majetschak et al. 2011).
Thioredoxin (Trx): This is the prototypic member of the thioredoxin superfamily, all of which have a common structure consisting of a β-sheet core surrounded with α-helices. Most of these proteins contain a canonical CXXC motif(Pan and Bardwell, 2006) showing similarity to the CXC and CX3C chemokines This small protein was initially, at least in humans, identified as a secreted cytokine (Tagaya et al. 1989). It has subsequently been established that Trx has protein folding, anti-oxidant and signal transduction properties within cells (Lillig and Holmgren, 2008). The extracellular biological actions of Trx are complex, and additional functions are constantly being found. Human Trx (now Trx-1) was first identified as a T cell growth cytokine and it has subsequently been shown that oxidative stress increases the release of Trx from lymphocytes suggesting these cells have a redox-sensing cytokine system (Kondo et al. 2004). The current appreciation of the biological role of extracellular Trx suggests that it is a natural anti-inflammatory protein which possibly acts in a manner similar to the other known extracellular anti-inflammatory cell stress proteins: Hsp10 (HSPE1), HSPB1 and HSPA5.
Thus studies in the 1990s identified Trx as a unique chemoattractant - for all leukocyte populations – which, unlike the chemokines, functions in a G protein-independent manner(Bertini et al. 1999). Increased circulating levels of this protein, caused by the administration of Trx to mice, inhibits LPS- or chemokine-induced neutrophil migration into skin air pouches, thereby demonstrating that this protein is chemotactic (Nakamura et al. 2001a). This ability of circulating Trx to inhibit leukocyte chemotaxis is important in HIV infection as the survival of HIV-infected individuals with low numbers of CD4+ T cells (<200 CD4+ T cells/µl blood) and high plasma levels of Trx is significantly poorer than in those with lower Trx levels. The explanation for this finding is that the Trx prevents the innate immune system functioning to destroy opportunistic pathogens by blocking leukocyte chemotaxis into tissues(Nakamura et al. 2001b). In addition to its chemotaxis-inhibiting actions, Trx is also able to inhibit human monocyte IL-1β expression that is induced by lipopolysaccharide (Billiet et al. 2005). Interestingly, human regulatory T lymphocytes (Tregs) produce high levels of Trx (Mougiakakos et al. 2010).
One of the key questions about cell stress proteins, which are rarely addressed, is the nature of the cell surface receptor. However, there is evidence that Trx bind to CD30, a member of the TNF receptor superfamily (Schwertassek et al. 2007) about which little is known. Potentially unrelated to inflammation, but of relevance to cardiovascular pathology, is the finding that Trx activates certain transient receptor potential channels (TRPCs), a recently discovered family of non-selective and non-voltage-gated ion channels that function as extracellular cell sensors. There is evidence that the TRPCs have a role in cardiovascular disease, with TRPC1, 3 and 6 often being found to be upregulated in models of cardiovascular disease. Inhibition of these channels often decreases the pathology associated with these models, suggesting that these ion channels may be therapeutic targets (Rowell et al. 2010). In this context, Trx has been reported to activate TRPC1 and 5 by binding and oxidising a disulphide bond (Xu et al. 2008). This finding reveals the unexpected nature of the many extracellular functions of Trx.
Thioredoxin can inhibit aspects of lipopolysaccharide-induced inflammation, but can it inhibit more complex inflammatory models? Thioredoxin has been administered in a number of experimental models of human inflammatory disease, including myosin-induced autoimmune myocarditis (Liu et al. 2004) with beneficial effects (Supplementary Table 1). These findings suggest that Trx is a systemic anti-inflammatory agent with potential therapeutic benefits in human inflammatory disease. There are a number of reviews which support the hypothesis that Trx should be used therapeutically to treat human disease, including cardiovascular disease (e.g. Billiet and Rouis, 2008; Watanabe et al. 2010).
Measurement of circulating Trx has been undertaken in patients with various forms of cardiovascular pathology (Supplementary Table 2). Circulating Trx levels generally increase in cardiovascular pathologies and appear to be inversely correlated with disease pathology. In other words, higher levels of Trx in the blood are associated with a lower level of pathology, or a better outcome.
Hsp27 (HSPB1):A member of the small heat shock protein family which has control over a number of functions including: metabolic activity, cytoskeletal structure, cell growth/differentiation/survival, cell migration and mRNA stabilisation. There is also strong evidence to suggest this molecular chaperone controls tumour progression (Kostenko and Moens, 2009). HSPB1 is detectable in serum at around 3ng/ml in a proportion of the normal population (De and Roach, 2004). Extracellular HSPB1 has the ability to induce macrophages to generate an anti-inflammatory profile of cytokines (De et al. 2000). Thus, this molecular chaperone, if released, would be expected to produce an anti-inflammatory effect.
There are now a small, but growing, number of publications suggesting that secreted HSPB1 has a role to play in cardiovascular pathology. The first evidence to support this hypothesis was a proteomic study of normal and atherosclerotic arteries, which identified HSPB1 as a protein secreted by blood vessels. Unexpectedly, normal arteries secreted much more HSPB1 than did those showing evidence of atherosclerosis. This finding was confirmed by measuring circulating immunoreactive HSPB1 levels, which revealed that healthy subjects had between 72-88ng/ml HSPB1 compared with 0.1-2.0ng/ml in patients with carotid stenosis. This suggested that circulating HSPB1 levels could provide insight into the presence of cardiovascular pathology in patients (Martin-Ventura et al. 2004). The differential levels of HSPB1 in healthy and pathological areas of atherosclerotic blood vessels were confirmed in a separate study. In this, circulating HSPB1 levels in patients with acute coronary syndrome (ACS) were analysed and found to be higher than those in the plasma of healthy controls (106±74ng/ml vs. 45.8±29.5, p<0.005).Of interest was the finding that HSPB1 concentrations in the patients with ACS significantly correlated with circulating levels of Hsp70 (Park et al. 2006).
The lowered circulating HSPB1 levels in patients with atherosclerosis supported the hypothesis that this protein is atheroprotective. To test this,HSPB1 overexpressing (HSPB1o/e) mice were crossed with ApoE-/- mice and when fed a high fat diet the HSPB1o/e mice showed a 35% reduction in atherosclerotic lesion area. Surprisingly, the decrease in atherogenesis was only found in the females and was related to the 10-fold higher serum HSPB1 levels in these animals compared to the ApoE-/-HSPB1o/e males. Circulating HSPB1 levels were significantly inversely correlated with the lesional areas of blood vessels. Potential explanations for this effect are the abilities of exogenous HSPB1 to inhibit macrophage uptake of acetylated LDL (acLDL) and of this protein to inhibit pro-inflammatory cytokine synthesis and upregulate IL-10 synthesis (Rayner et al. 2008).HSPB1 associates with oestrogen receptor beta (ERβ) in the cell and the loss of this protein from atherosclerotic arteries coincides with a loss of this receptor protein (Miller et al. 2005). In the study by Rayner and colleagues (2008)the addition of oestrogen to cultured macrophages increased the synthesis and secretion of HSPB1and this presumably accounted for the difference between male and female mice. The release of HSPB1 is by an exosomal mechanism (Rayner et al. 2009). The role of oestrogen in the atheroprotective effect of HSPB1 was further tested in ovariectomised ApoE-/-HSPB1o/e mice which showed no difference from ApoE-/- mice. Administration of an ERβ-specific antagonist similarly blocked the atheroprotective effect in the ApoE-/-HSPB1o/e mice (Rayner et al. 2009). It has recently been argued that in spite of the unfavourable risk/benefit ratio, oestrogen and related ovarian hormones have beneficial effects on the vasculature. The beneficial effects of oestrogen are probably related to its ability to increase the levels of HSPB1 in the circulation, with this protein having a significant atheroprotective profile of action (Rayner et al. 2010). HSPB1 in the blood of patients undergoing cardiopulmonary bypass is also reported to be increased acutely (Szerafin et al. 2008).
In addition to measuring HSPB1 in the circulation, a small number of papers report the presence of antibodies to this protein in the blood of patients with various cardiovascular conditions. Thus Sham et al (2008) have associated anti-HSPB1 levels in blood with age and hypertension and another group have related antibody titres to the severity of coronary artery disease (Pourghadamyari et al. 2010). HSPB1 has a fascinating profile of anti-inflammatory biological activity and may function as a natural homeostatic regulator of blood vessel function and more information on its circulating levels in cardiovascular disease is needed.
Hsp70 (HSPA) Family:Many of the investigations of circulating cell stress proteins in cardiovascular disease has focused on HSPA1A. As described, upregulation of Hsp70 (presumably HSPA1A) in animals is cardioprotective (Bielecka-Dabrowa et al. 2009;Latchman, 2001). Such increases in intracellular HSPA1A, in tissues undergoing reperfusion injury, are likely to lead to the release of this protein which may be immunogenic. HSPA1A, like HSPD1, has been shown to be an immunodominant antigen in experimental animals (Feige and van Eden, 1996) and immune responsiveness to HSPA1A has been found in cardiac allografts (Duquesnoy et al. 1999). In this context, healthy mice have low levels of IgD and IgM autoantibodies to HSPA1A in their blood (Menoret et al. 2000).Mycobacterium tuberculosis Hsp60 has been shown to profoundly influence rat and mouse T lymphocytes and induce or prevent various autoimmune models of disease (van Eden et al. 2005). An M. tuberculosis Hsp70 (DnaK) sequence has been found to be immunogenic and to generate activated T lymphocytes that cross-react with the homologous sequence in rat HSPA1A. This peptide, administered intra-nasally to rats, could inhibit the induction of adjuvant arthritis (Wendling et al. 2000). Such inhibition of arthritis is crucially dependent on the ability of the HSPA1A to induce antigen-specific IL-10 synthesis (Wieten et al. 2009). The role of T cell immunity to HSPA1A in the generation of experimental atheroma is not known. However, it has been shown that induction of oral tolerance to HSPD1 in LDLr-/- mice results in a significant diminution of plaque size (van Puijvelde et al. 2007). In different models of atherosclerosis it has been shown that HSPA1A levels increase in lesional sites (Geetanjali et al. 2002; Han et al. 2003). In a rat carotid injury model immunisation of animals with HSPA1A prior to injury resulted in increased arterial injury which, itself, was associated with increased local HSPA1A production (George et al. 2001).