Introduction

As discussed in other chapters in this volume, heat shock proteins are traditionally viewed as protein chaperones rather than immunomodulators [1–3]. However, recent data suggest that heat shock proteins might also be ancestral danger signals which activate adaptive and innate immune responses [3]. The majority of studies examining the immunomodulatory activities of heat shock proteins have focused on the large heat shock proteins, Hsp60, Hsp70 and gp96, and these proteins have been shown to stimulate the innate immune system via binding to a variety of cellular receptors, particularly on monocytes, and to play an important role in health and disease [1–6].

This chapter focuses on the small heat shock protein, Hsp27, which, although shown to have some role in resistance to chemotherapeutic drugs, cytokine-induced cytotoxicity and to have been described as a prognostic marker in serum of breast cancer patients, has not been well characterised as an immunomodulator [7–13]. Hsp27 has been reported as present in increased amounts in the serum of patients with several human diseases, as well as being necessary for activation of the signal transduction pathway leading to monocyte production of the anti-inflammatory and immunoinhibitory cytokine IL-10 [1–3, 14]. These data led to our interest in investigating the possible immunomodulatory activity of Hsp27 on different human monocyte functions, which are pivotal in both the development of inflammatory responses as well as the triggering of lymphocyte-specific immunity.

Introduction

In the beginning, bacteria evolved chaperonins (Cpns) to help in the folding of other proteins. Then, about one and a half billion years ago, bacteria began to live with eukaryotic cells. The career of the chaperonin began to expand beyond the protein folding area, and they developed new functions in order to adapt to the eukaryotic evolutionary niche. As the eukaryotes became ever more complex, the chaperonins evolved into cell-cell signalling molecules. The sophistication of this new role has only recently begun to emerge.

The chaperonins that are the subject of this chapter belong to the 60- and 10-kDa classes and are called Cpn60 and Cpn10, respectively. The folding actions of these proteins have been described in detail in Chapter 1. If there is more than one Cpn60 or Cpn10 in any one species, they are called Cpn60.1, Cpn60.2 and so on [1]. Cpn60 proteins are also called heat shock protein (Hsp) 60 or 65, and Cpn10s are also named Hsp10. For example, the Mycobacterium tuberculosis genome contains two cpn60 genes. One of these, termed cpn60.1 [2], appears to form an operon with the cpn10 gene. This is the usual relationship in most bacteria. The second cpn60 gene encodes Cpn60.2, the well-known Hsp65 protein of M. tuberculosis, and is found elsewhere in the genome.

What is a cell-cell signalling molecule?

A cell-cell signalling molecule is one that directly communicates a message from one cell to another.

Introduction

There is a clear dichotomy between the effects of the 70-kDa heat shock protein (Hsp70) when expressed intracellularly and when released into the extracellular space. Intracellular Hsp70 is primarily implicated as a protein chaperone that transports and folds naïve, aberrantly folded, or mutated proteins, resulting in cytoprotection when cells are exposed to stressful stimuli – most notably heat shock itself. Intracellular Hsp70 also functions as a regulatory molecule and has a largely inhibitory function in cellular metabolism. In contrast, Hsp70 is implicated in immune activation, given that exposure of immunocompetent cells to exogenous Hsp70 triggers acute inflammatory responses, activates innate immunity and enhances anti-tumour surveillance. This review focuses on recent advances in understanding the contrasting roles of Hsp70 as an intracellular molecular chaperone and extracellular signalling ligand and highlights its relevance to host defence against pathogens and malignant transformation.

The formative studies of Hsp70, dating back over 30 years, suggested a strictly intracellular function for Hsp70 with the properties of a molecular chaperone – a protein that modulates the tertiary structures of other proteins and protects cells from stress [1, 2]. However, in 1989 Hightower and colleagues showed that cells in culture contain a pool of Hsp70 which is loosely associated with the cell and which could be released into the extracellular medium after heat shock or even after mild washing with tissue culture medium [3].

Introduction

Heat shock or stress proteins are important intracellular protein chaperones that control their trafficking. The function of heat shock proteins in the immunopathology of infections, tumours and autoimmune diseases has been the subject of numerous experimental and clinical investigations over the past few decades and some of their properties are summarised in Table 10.1. Because there is extensive homology between mammalian and microbial heat shock proteins, immunological cross-reactions were considered to account for a number of autoimmune diseases. However, although the biological significance of lipopolysaccharide (LPS) found in Gram-negative bacteria has been well appreciated, heat shock proteins, which are found more widely in Gram-negative and -positive bacteria, especially those in the gut, have received more limited attention. A relatively new phase in this area of biology was initiated by the discovery of specific heat shock protein receptors and by rapid advances in our understanding of the signalling pathways. This chapter will deal with the receptors used by heat shock proteins, with particular reference to Hsp70, involvement of these proteins in the stimulation of chemokine production, maturation of dendritic cells (DCs), their intrinsic adjuvanticity and capacity to enhance immunogenicity.

Structural features of Hsp70

Although the overall three-dimensional structure of Hsp70 is not known, the structures of the two domains from various members of the family have been solved separately. The crystal structures of the ATPase domains of bovine Hsc70 (heat shock constitutive protein) and human Hsp70 have been determined [1, 2].

Introduction

Heat shock proteins were initially identified as heterogeneous families of stress-induced proteins characterised by their chaperone activity [1]. Subsequently, they were identified as immunodominant antigens recognised by the host immune system following microbial infection [2] or during the course of autoimmune disease [3–6]. Recently, the role of heat shock proteins as endogenous activators of the innate and adaptive immune system has been unveiled [7]. In this chapter we discuss the relevance of heat shock proteins and their immune activities to the regulation of inflammation and autoimmune disease. We shall see that the regulatory activities of heat shock proteins on inflammation involve two types of cross-reactivity: molecular cross-reactivity exists between microbial and self-heat shock proteins and network cross-reactivity exists between different self-heat shock proteins.

Inflammation activates heat shock protein–specific T cells

Although the injection of incomplete Freund's adjuvant (IFA) to BALB/c mice induces local inflammation, Anderton and colleagues demonstrated that the injection of IFA also induces T cells reactive with the mammalian 60-kDa heat shock protein (Hsp60) [8]. These Hsp60-reactive T cells were TCRαβ+, CD4+ and major histocompatibility complex (MHC) class II-restricted [8]. Notably, Hsp60-specific cells could only be found in the local lymph nodes draining the site of IFA injection, and they were not present in distant lymph nodes. Hsp60-specific T cells are not only induced but also recruited to the site of inflammation [8].

It has been estimated that more than 80% of human cancers are caused by environmental factors, and so are potentially preventable (Higginson, 1968). Bacteria are in a unique position to mediate between the host and its external environment because they naturally colonise all external surfaces. In addition, bacterial infections expose the host to further interactions with bacteria.

The role of the flora in protecting the host against infection has been well documented. The normal bacterial flora of the mouth protects against fungal infection; the normal flora of the intestine helps to protect against enteritis; the normal skin flora helps to protect against skin infection.

Furthermore, the gut bacterial flora has an overwhelmingly beneficial role in assisting digestion and in detoxifying ingested compounds of exogenous origin (Hill, 1995). However, there are situations where bacterial infection, which can be defined as bacteria being in the wrong place, can produce carcinogens and is associated with local cancers (e.g., bladder infection, nitrosamines, and bladder cancer). In addition, there are situations where the normal flora, carrying out its normal metabolic activity, can be exposed to very high levels of a substrate resulting in toxic levels of the product (e.g., colon bacteria, bile acids, and colon cancer).

In this chapter we will first discuss the types of carcinogens/mutagens that can be produced by bacterial action, and the situations in which this production might be significant. In addition, the gut bacteria release a wide range of anti-carcinogens from the diet as well, and these are discussed.

Low molecular mass GTPases of the Rho family are master regulators of the actin cytoskeleton and are involved in various signal transduction processes. They regulate the actin cytoskeleton and control transcriptional activation, cell cycle progression, cell transformation, and apoptosis. Rho GTPases are the eukaryotic targets of a variety of bacterial protein toxins that either inhibit or activate the Rho proteins. The cytotoxic necrotizing factors CNF1 and 2 from Escherichia coli and the dermonecrotic toxin DNT from Bordetella species are Rho-activating toxins. Whereas CNFs cause deamidation of Rho proteins, DNT deamidates and/or transglutaminates the GTPases by attachment of polyamines. Both deamidation and transglutamination cause persistent activation of Rho GTPases and subsequent changes in processes governed by the GTPases.

RHO GTPASES

Rho GTPases belong to the Ras superefmily of low molecular mass GTPases and are molecular switches in various signalling pathways. Like all members of GTP-binding proteins, they are active in the GTP bound form and inactive with GDP bound. Activation occurs by GDP/GTP exchange catalysed by guanine nucleotide exchange factors (GEFs) and is inhibited by binding of guanine nucleotide dissociation inhibitors (GDIs). Inactivation by hydrolysis of the bound GTP is catalysed by GTPase activating proteins (GAPs) (Symons and Settleman, 2000).

Rho GTPases are ubiquitously expressed. The members (>15) of the Rho family of GTPases, including RhoA, B, C, D, E, and G, Cdc42, Rac1, 2, and 3, share more than 50% sequence identity. The GTPases have been described as important regulators of the actin cytoskeleton.

The concept that bacteria are rather simple organisms that interact with eukaryotic cells in a passive manner is now totally untenable, as more evidence emerges of active interaction and reaction between bacteria and host. It is also becoming clear that bacterial toxins do not merely operate as molecules of death for a cell. Toxin action may ultimately result in cell death, and indeed death of the host organism, but often the producer bacterium requires that the host cell is first organised in a particular manner. In these circumstances, the toxins sent out by the bacterium on its “behalf” have a mission to regulate the target cell in a very precise manner. Not much is known about the eukaryotic mechanisms that exist to control rather then kill bacteria, but such mechanisms must exist to regulate the homeostatic balance between a eukaryotic host and its commensal bacteria. Furthermore, a host weakened by diseases leading to immunodeficiency, such as HIV/AIDS, stress, or malnutrition, is more susceptible to infection – both by bacteria normally classified as pathogenic, but also by commensal bacteria, which in the circumstances are named as opportunistic pathogens.

Much more is known about the mechanisms that bacteria use to regulate cellular function. Many of these mechanisms have been described in detail in the chapters in this book. Several common themes are appearing. The most frequent targets for such toxins are intracellular proteins involved in the signal transduction pathways that integrate incoming signals at the cell surface.

Bacterial protein toxins are powerful biological poisons normally associated with impairment of cellular function and/or cellular death. The wide spectrum of physiological processes and cell types that are affected by bacterial products also includes bone tissue and bone cells. It has been known for many years that bacterial infection or exposure to certain toxins can lead to pathological bone disorders, most commonly, those associated with abnormal or excessive bone loss, such as periodontal disease (reviewed by Henderson and Nair, 2003). However, in most cases the bone-resorbing factors involved in these effects remain part of, or associated with, the bacterial surface. For example, the bone-resorbing effects of endotoxin, a component of lipopolysaccharide, are well established, although for the most part this action appears to be indirect, being dependent on the production of pro-inflammatory cytokines (IL-1, TNFα) from other cell types (Nair et al., 1996; Henderson and Nair 2003). In contrast, the effects of bacterial protein toxins on the cellular constituents of bone remain largely unknown. For simplicity, this review will focus only on bacterial toxins, in particular, those toxins that interfere with key signalling processes that have direct relevance to bone cell differentiation and function. However, a brief overview of the general biology of bone cells is necessary before discussing the mechanisms of toxin action and specific signal transduction pathways in bone.

BONE

Throughout life the vertebrate skeleton is in a constant state of turnover.

Helicobacter pylori, the first bacterium to be designated a Class I carcinogen, has a major aetiological role in human gastric carcinogenesis. H. pylori infection is acquired primarily in childhood and, in the majority of instances, infection and associated chronic gastritis are lifelong. A key feature of H. pylori infection of relevance to the associated increased risk of developing gastric cancer is the hyperproliferation of gastric epithelial cells induced by the bacterium. Infection is associated with increased gastric epithelial cell proliferation in both humans and in experimental animal models.

Clinically, there is a marked diversity in the outcome of H. pylori infection and only a few infected subjects will develop gastric cancer (reviewed Peek and Blaser, 2002). Recent studies in Japan show that the risk of cancer with H. pylori infection is greatest in infected subjects with nonulcer dyspesia or gastric ulceration who develop severe gastric atrophy and intestinal metaplasia (Uemura et al., 2001). Bacterial virulence factors such as the cag pathogenicity island (PAI) (Blaser et al., 1995; Kuipers et al., 1995; Webb et al., 1999) and genetic polymorphisms in the interleukin-1β and IL-1 receptor antagonist genes associated with overexpression of IL-1 and hypochlorhydria (El-Omar et al., 2000; Machado et al., 2001; Furuta et al., 2002) have each been linked to an increased risk of developing gastric atrophy and/or intestinal type gastric cancer.

H. pylori is one of several chronic infections that have recently been associated with the development of neoplasia (see Chapter 9).

During the last 10 years, information has accumulated showing that pathogenic bacteria can produce various proteins able to block the eukaryotic cell cycle or delay its progression. These observations raise the attractive hypothesis that control of cell proliferation is a real strategy of pathogenicity, giving an evolutionary advantage to bacteria, and not simply a fortuitous effect observable in cell cultures, the interest of which would eventually be confined to cellular biologists or pharmacologists. The ultimate objective of this chapter is to analyse critically the pertinence of this candidate concept within the field of cellular microbiology (Cossart et al., 1996; Henderson et al., 1998) and to propose tentative criteria to define what we suggest calling bacterial cyclostatins.

We have chosen cytolethal distending toxin (CDT) as a prototype cyclostatin. From a probable common ancestor, CDT has spread through the bacterial world and it can be found in several Gram-negative pathogenic bacteria, constituting a family of toxins sharing common molecular and biological properties in spite of a large genetic dispersion (De Rycke and Oswald, 2001). The presence of a CDT homologue in various unrelated bacterial species is peculiar and suggests that CDT confers a strong selective advantage to producing bacteria, possibly helping adaptation to the host and ecological niche, or increasing pathogenicity. Another major interest in using CDT as a prototype cyclostatin is that a consistent picture of its mode of action on mammalian cells is now emerging, as a result of intensive research effort in recent years by several teams of investigators.

The Pasteurella multocida toxin (PMT) is produced by some type A and D strains of the Gram-negative bacterium Pasteurella multocida. These bacteria cause several animal infections and can occasionally cause human disease. PMT is the major virulence factor associated with porcine atrophic rhinitis, a non-fatal respiratory infection characterised by loss of the nasal turbinate bones and a twisting or shortening of the snout. However, PMT is highly toxic to animals, being lethal to mice at similar concentrations to diphtheria toxin. Despite these toxic properties, it turns out that PMT has unexpected effects on cells in culture leading to perturbation of several signalling pathways. The consequence of this action is that PMT can affect the regulation of cell growth and differentiation.

PMT IS A MITOGEN

The cellular effects of PMT have been most widely studied on Swiss 3T3 cells, a mouse fibroblast cell line. These cells are useful for studying growth factors since they are contact-inhibited and are readily quiesced by growing to confluence and allowing the cells to deplete the medium of growth factors. Rozengurt et al. (1990) first showed that PMT caused quiescent Swiss 3T3 cells to recommence DNA synthesis. The toxin is highly potent, inducing maximal DNA synthesis at only 1.25 ng/ml (or about 2 pM). This is equivalent to the DNA synthesis induced by 10% foetal bovine serum. Thus, PMT is more mitogenic for this cell type than any known growth factor (Figure 2.1, top panel).

The concept of a bacterial protein toxin was born in the 1880s as Friedrich Loeffler in Berlin, and Émile Roux and Alexandre Yersin in Paris, puzzled over the disease diphtheria. The bacteria were localised in the throats of patients and experimental animals, yet the disease caused systemic damage throughout the body. They reasoned that the bacteria must be producing a poison that could escape from the bacteria to cause widespread damage to the host. So the toxin concept was established right at the start of Medical Microbiology (Roux and Yersin, 1888), only a decade after Robert Koch had established the first definite link between a bacterium and disease with his seminal work on anthrax. However, it was only from the mid-twentieth century onwards that the action of any toxin was understood at the molecular level. Since then progress has been rapid, not only in our appreciation of the mode of action of historically known toxins but also in the discovery of new toxins with novel means of attacking cells.

CLASSES OF BACTERIAL PROTEIN TOXINS

The first toxin to be understood at the molecular level was one from Clostridium perfringens, a bacterium notorious for causing wound infections such as gas gangrene. This toxin is a phospholipase that attacks membranes of cells and, thus, it defined one of the three main categories of toxins, i.e., those that attack membranes (MacFarlane and Knight, 1941).

Bartonella bacilliformis is the causative agent of Carrion's disease (for a recent medical review, see Maguina and Gotuzzo, 2000). B. bacilliformis was in a genus by itself until recent taxonomic revisions associated it with other bacteria. Now there are sixteen known species of Bartonella, of which seven are associated with human disease. B. bacilliformis enters the bloodstream of humans through the bite of an insect vector and targets primarily erythrocytes and endothelial cells.

For many years, it seemed as though important and interesting conclusions about the mechanism of angiogenesis might result from the study of B. bacilliformis, which can initiate and perpetuate lesions formed of proliferating endothelial cells and newly formed capillaries. Since then, the Science juggernaut has moved on, more rapidly for angiogenesis than for Bartonella, and so it now appears that we can best understand B. bacilliformis (and the closely related Bartonella henselae) by reference to the known biology of angiogenesis. However, study of B. bacilliformis has made at least one useful contribution to our knowledge of angiogenesis. Entry of B. bacilliformis by vacuole formation and macropinocytosis was shown to be accomplished by activation of Rho, Rac, and Cdc42 GTPases. This process appeared to be analogous to the angiogenic process by which endothelial cells take in large volumes of extracellular fluids to form large central vacuoles, which perhaps fuse between cells to form the nascent capillary.