Abstract
The innate host response system is comprised of various mechanisms for orchestrating host response to microbial infection of the oral cavity. The heterogeneity of the oral cavity and the associated microenvironments that are produced give rise to different chemistries that affect the innate defense system. One focus of this review is on how these spatial differences influence the two major defensive peroxidases of the oral cavity, salivary peroxidase (SPO) and myeloperoxidase (MPO). With hydrogen peroxide (H2O2) as an oxidant, the defensive peroxidases use inorganic ions to produce antimicrobials that are generally more effective than H2O2 itself. The concentrations of the inorganic substrates are different in saliva vs. gingival crevicular fluid (GCF). Thus, in the supragingival regime, SPO and MPO work in unison for the exclusive production of hypothiocyanite (OSCN−, a reactive inorganic species), which constantly bathes nascent plaques. In contrast, MPO is introduced to the GCF during inflammatory response, and in that environment it is capable of producing hypochlorite (OCl−), a chemically more powerful oxidant that is implicated in host tissue damage. A second focus of this review is on inter-person variation that may contribute to different peroxidase function. Many of these differences are attributed to dietary or smoking practices that alter the concentrations of relevant inorganic species in the oral cavity (e.g.: fluoride, F−; cyanide, CN−; cyanate, OCN−; thiocyanate, SCN−; and nitrate, NO3−). Because of the complexity of the host and microflora biology and the associated chemistry, it is difficult to establish the significance of the human peroxidase systems during the pathogenesis of oral diseases. The problem is particularly complex with respect to the gingival sulcus and periodontal pockets (where the very different defensive stratagems of GCF and saliva co-mingle). Despite this complexity, intriguing in vitro and in vivo studies are reviewed here that reveal the interplay between peroxidase function and associated inorganic chemistry.
INTRODUCTION
The oral cavity contains a plethora of specific and non-specific defense factors. The non-specific factors include some mucins, proline-rich proteins, salivary glycoproteins, lactoferrin, lysozyme, histatins, cystatins, and peroxidases. This review focuses on the role of peroxidases in the context of oral health and disease, with an emphasis on the relevant inorganic chemistry. Particular attention is paid to the antimicrobial properties of the inorganic chemicals of the oral cavity that are associated with the peroxidases, and to the inter-person differences in the inorganic chemistry of the oral cavity that may influence peroxidase function. For further information on the structures and origins of human oral peroxidases, the reader is referred to the recent review by Tenovuo and co-workers (Ihalin et al., 2006). There are two principal defensive peroxidase systems in the oral cavity, salivary peroxidase (SPO) and myeloperoxidase (MPO). SPO is structurally and catalytically similar to lactoperoxidase (LPO) (Ihalin et al., 2006). In vivo, the SPO and LPO systems essentially use only the pseudohalide SCN− as a substrate to produce OSCN− (Pruitt et al., 1988). Such defensive peroxidases are commonly found in regions of the human body that are controlled by the mucosa: e.g., breast milk (Shin et al., 2000), lachrymal fluid (Van Haeringen et al., 1979; Tenovuo et al., 1985), and the mucosal lining of the lungs (Gerson et al., 2000). LPO and SPO are coded for the same gene (Ueda et al., 1997). In contrast to peroxidases that essentially employ only SCN− as a substrate (e.g., LPO and SPO), the MPO system is also capable of oxidizing Cl− to produce hypochlorite (OCl−) (Arnhold et al., 2006). Hypobromite (OBr−) can also be generated by the MPO system (Thomas et al., 1995), but only in minor amounts in the oral cavity. All of the human defensive peroxidases can also utilize iodide (I−) as a substrate. However, because of sequestration in the thyroid, the environmentally rare halide I− is not abundant in most physiologic fluids, including the fluids of the oral cavity (Anttonen and Tenovuo, 1981). Accordingly, the limited bioavailability of I− precludes its significant involvement in host defense.
SPO is a normal, non-inducible component of the saliva of the parotid and submandibular glands (Riva et al., 1978), whereas MPO is an offensive mechanism of neutrophilic polymorphonuclear leukocytes (PMNs). Leukocytes are not normal components of the saliva of healthy individuals, but rather are introduced to the oral cavity by gingival crevicular fluid (GCF) during inflammatory responses (Kowolik and Grant, 1983). The leukocytes in the GCF are comprised of ca. 90% PMNs (Ebersole, 2003), and MPO accounts for about 5% of the total PMN protein (Pullar et al., 2000). PMNs degenerate in saliva due to osmotic lysis, thereby releasing the content of the azurophilic granules (including MPO). It has been estimated that ca. 75% of the peroxidase activity in mixed saliva is due to MPO, with the remaining activity attributed to SPO (Thomas et al., 1994a). Most of the SPO activity is associated with the soluble portion of the saliva, whereas most of the MPO activity is associated with the sediment (Thomas et al., 1994a). Note that, in contrast to other regions of the mucosa—for example, the lungs—eosinophils are not usually recruited into the oral cavity, although they can be introduced to saliva via the sputum of individuals who have asthma eosinophilia (Spahn, 2007) and from eosinophilic ulcers (rare lesions of the oral mucosa) (Mezei et al., 1995; Hirshberg et al., 2006). Consequently, there is no evidence that eosinophil peroxidase (another common defensive peroxidase that has properties somewhat different from those of LPO, SPO, and MPO) plays a significant role in oral fluids. Accordingly, this review focuses on SPO and MPO.
PEROXIDASES: ORAL HEALTH AND DISEASE
The two disease states of the oral cavity that we consider here are caries and periodontal diseases. The etiology of caries is clear: Acidogenic bacteria cause damage to tooth enamel in the presence of fermentable carbohydrates (e.g., sucrose, fructose, and glucose) (Featherstone, 2000). When the pH at the surface of the tooth falls below 5.5, demineralization proceeds faster than remineralization, and decay ensues. The role of inorganic chemistry in this process is multifaceted: e.g., the (de)mineralization process largely involves the inorganic mineral hydroxyapatite [ca. 96% for enamel and 70% for dentin (with some amorphous calcium phosphate)], and the aforementioned peroxidase-derived reactive inorganic species are involved in controlling the microbial growth (OSCN− in particular, although other inorganic chemical species have been proposed to be significant, vide infra).
Like caries, periodontal diseases are also caused by microbial infection (Smalley, 1994; Genco, 1996; Mombelli, 2003). Although the primary cause of periodontal diseases is the accumulation of dental plaque at the gingival margin and the consequential host response (Azuma, 2006), numerous factors affect the severity of the diseases, include smoking (Bergström, 2004), poorly controlled diabetes (Mealey and Oates, 2006), and genetic susceptibility (Baker and Roopenian, 2002; Shapira et al., 2005). Both soft tissues (gingival and periodontal ligaments) and hard tissues (alveolar bone and cementum, which are both largely hydroxyapatite) are affected, but the cause of this tissue damage is a complex and as-yet-unresolved matter. It appears likely that inflammatory agents (including OCl−) produced by the host (Pullar et al., 2000; Klebanoff, 2005) and virulence factors produced by the infectious agents (Graves et al., 2000) are both responsible for the tissue damage. The supragingival environment in which caries develops and the subgingival environment of periodontal diseases exhibit different chemistries that have a marked influence on the functions and activities of the human defensive peroxidases, which will be discussed next (Fig. 1, Table 1).
SPATIAL NATURE OF PEROXIDASE SYSTEMS AND THEIR CONSENSUS SUBSTRATES
The oxidation reactions that are catalyzed by the peroxidase systems of the oral cavity are governed by the amount of available hydrogen peroxide (H2O2), the limiting chemical reagent. A dual-oxidase system from the salivary glands is an endogenous source of H2O2 (Geiszt et al., 2003; Donko et al., 2005; Ris-Stalpers, 2006). Oral bacteria also produce H2O2 during anaerobic glycolysis (Carlsson et al., 1983). A third source of H2O2 is derived from activated neutrophils during oxidative bursts (Dahlgren and Karlsson, 1999; Quinn, 2005). The amounts of OCl− and OSCN− that are produced by the MPO system are related to the relative concentrations of Cl− and SCN− (van Dalen et al., 1997; Arnhold et al., 2006). At equal concentrations of (pseudo)halide, MPO catalyzes the oxidation of SCN− about 1000 times faster than Cl−, but Cl− is about 1000 times more abundant in most physiologic fluids [e.g., plasma and GCF (Anttonen and Tenovuo, 1981)]. Consequently, comparable amounts of OSCN− and OCl− are produced by the MPO system in such fluids. However, SCN− is essentially the only substrate of MPO in saliva, where the concentration of SCN− is higher than in most other extracelluar fluids (Tenovuo and Makinen, 1976), as a consequence of its active transport (Fragoso et al., 2004). While analysis of the data in Table 1 suggests that OCl− should also be generated in saliva, albeit in a minor amount with respect to OSCN−, it can be estimated that the half-life of the OCl− in saliva is less than 15 μsec, as a consequence of its very fast non-enzymic reaction with SCN− (Ashby et al., 2004). The reaction of OCl− and SCN− yields OSCN− (Nagy et al., 2006a). Thus, in effect, the only hypohalite of the SPO and MPO systems in saliva is expected to be OSCN−, but a continuum of products is expected at the gingival margin, where a gradient of concentration of Cl− and SCN− exists. The spatial relationship between these gradients and the areas of the oral cavity that are respectively controlled by the OCl− and OSCN− defense factors is illustrated in Fig. 1:
Extraphagosomal OCl− is cytotoxic to oral bacteria (Path A) (Briseno et al., 1992; Webb et al., 1995; Yesilsoy et al., 1995; Barnard et al., 1996; Winniczuk and Parish, 1997; Calas et al., 1998; D’Arcangelo et al., 1998; D’Arcangelo and Varvara, 1998; Huque et al., 1998; Ferreira et al., 1999; Wunder and Bowen, 1999; Spratt et al., 2001; Mikami et al., 2003; Sassone et al., 2003a,b; Moller et al., 2004; Nagayoshi et al., 2004; Radcliffe et al., 2004; Vianna et al., 2004; Carson et al., 2005; Fang et al., 2006; Sena et al., 2006; Ozok et al., 2007) and gingival tissue (Path B) (Schraufstatter et al., 1990; Vissers et al., 1999; Hidalgo and Dominguez, 2000; Pullar et al., 2000; Vile et al., 2000; Hidalgo et al., 2002). Importantly, nearly all of the investigations of the efficacy of OCl− on oral bacteria have been carried out for single species in planktonic cultures. However, a recent study has focused on the effects of OCl− on single- and dual-species biofilms of Fusobacterium nucleatum and Peptostreptococcus micros (Ozok et al., 2007). Alternatively, OCl− can react with SCN− to produce HOSCN (Path C) (Ashby et al., 2004). HOSCN is also produced by the SPO-catalyzed oxidation of SCN− by H2O2 (Ihalin et al., 2006 ; Nagy et al., 2006a). HOSCN is antimicrobial toward oral bacteria (Path D) (Clem and Klebanof, 1966; Hoogendoorn, 1976; Pruitt et al., 1979; Carlsson et al., 1983; Thomas et al., 1983, 1994b; Ellen et al., 1988; Lopatin et al., 1991; Lumikari et al., 1991; Courtois et al., 1992; Lenander-Lumikari et al., 1993, 1997; van der Hoeven and Camp, 1993; Kirstila et al., 1994; Jones et al., 1998; Fadel and Courtois, 1999, 2001; Yu et al., 2000; Ihalin et al., 2001, 2003; Korpela et al., 2002; Garcia-Graells et al., 2003; Vannini et al., 2004), but relatively non-injurious to the host (Bjoerck and Claesson, 1980; Marshall and Reiter, 1980; White et al., 1983; Carlsson et al., 1984; Carlsson, 1987). In addition to the reaction of OCl− with SCN−, it may react with other small molecules (Path E) to produce secondary antimicrobials [e.g., when Z is an amine, a cytotoxic chlora–mine is produced, vide infra (Abia et al., 1998; Hawkins and Davies, 1998; Hawkins et al., 2003; Davies, 2005)].
The corresponding relevance of OCl−vs. OSCN− in the oral cavity is related to the aforementioned spatial heterogeneity of the peroxidase defense systems and the corresponding chemistry. For example, the median concentration of OSCN− in freshly collected whole saliva is ca. 10 μM, although the concentration increases when the saliva is incubated at 37°C (Thomas et al., 1980). However, direct measurement of the concentrations of these hypohalite species is problematic, because they are chemically reactive, and consequently the abundance of free ions does not necessarily reflect their significance in vivo. The fluxes of the hypohalites (the rates at which these reactive species are produced and consumed) are difficult to define in the context of the oral cavity. An even more complex issue is the relationship between these fluxes with respect to oral health and disease. This topic will be revisited in the concluding section of this review.
ANTIMICROBIAL PROPERTIES OF INORGANIC COMPOUNDS
Chemical Basis of Cytotoxicity
In contrast to antibiotics that typically target a single chemical step in a biosynthetic pathway, inorganic antimicrobials are generally biocides (they have a propensity to cause wholesale disruption of cellular processes) (Ashby, 2007; Zhu, 2007). Accordingly, these inorganic species tend to be cytotoxic, to greater or lesser degrees, to both eukaryotes and prokaryotes. Thus, any discussion of the antimicrobial properties of inorganic compounds toward infectious agents goes hand-in-hand with a related discussion of host tissue damage. It is fascinating to the author that the human body has found generally effective ways of harnessing the potentially indiscriminant cytotoxic properties of some of these compounds for defensive purposes. The cytotoxic properties of the inorganic compounds that are discussed herein can ultimately be traced to their chemistry, which can be roughly divided into two categories: (1) compounds that engage in one-electron (radical) chemistry, and (2) compounds that engage in two-electron chemistry (generally electrophilic, with eventual oxygen atom transfer). An example of one-electron chemistry is the reduction of O2 by NADPH oxidase to give O2·− (a reaction that is carried out by PMN NADPH oxidase):
Note that NADPH is a two-electron reductant, so the chemical stoichiometry requires one NADPH to react with two oxygen molecules. However, the chemistry in fact involves one-electron steps, vis-à-vis enzyme intermediates. The reaction of HOCl with reduced glutathione (GSH) is an example of a two-electron (O-atom transfer) reaction (GSH is the principal cytoplasmic oxidative defense mechanism of eukaryotes, vide infra). The reaction occurs via a multistep mechanism, because the intermediate sulfenic acid (GSOH) is unstable (Nagy and Ashby, 2007; Nagy et al., 2007b):
Note that the first equivalent of GSH that reacts with HOCl involves a two-electron reaction (O-atom transfer, albeit probably via the hydrolysis of a sulfenyl chloride intermediate), as does the second reaction, even though the stoichiometry of the net reaction makes it appear that GSH is a one-electron reductant. In fact, the thiyl radical (GS·) is never involved in the reaction. While the distinction between one- and two-electron processes may appear to be a superfluous detail, in fact the difference distinguishes radical processes from non-radical processes. Radicals tend to target unsaturated functional groups in lipids, nucleotides, and aromatic amino acids (Buettner, 1993). In contrast, the hypohalites tend to target the nucleophiles in proteins (Hawkins et al., 2003; Davies, 2005).
Unsaturated organic compounds (e.g., aromatic amino acids and nucleotides) are particularly susceptible to derivation by high-energy radical species; hence, nuclear damage and mutagenesis are frequently the result of one-electron chemistry (Box et al., 2001; Marnett, 2002; Wang, 2008). Because radical chemistry tends to be very facile and comparatively indiscriminant from a chemical perspective, it is difficult for cells to mount an effective defense against radical species. In contrast, the reactivities of two-electron oxidants are typically related to the nucleophilicities of their reaction partners. Consequently, the chemistry of two-electron oxidants is usually well-defined (cf. radical chemistry). Cysteine (Cys) and methionine (Met) are usually the most reactive amino acid residues toward two-electron oxidants (because sulfur-containing compounds tend to be good nucleophiles) and are therefore often the first targets of two-electron oxidants (Hawkins et al., 2003). It is not a coincidence that the cytoplasms of eukaryotes (Meister, 1988; Fernandes et al., 2007) and many prokaryotes (Fahey et al., 1978; Smirnova and Oktyabrsky, 2005) contain high concentrations of GSH (a tripeptide containing Cys, which is used to combat oxidative stress by two-electron oxidants that operate by the aforementioned electrophilic mechanism) (Meister, 1988). It should be noted that glutathione also erects a significant defense against radical species (Sitte and Von Zglinicki, 2003; Djordjevic, 2004) (Table 2).
Oxygen Derivatives
Water (H2O) and molecular oxygen (O2) represent limiting extremes in the oxidation state of the element oxygen (O) in an aqueous environment. From a thermodynamic perspective, O exists as O2 in an overall aerobic (oxidative) environment, whereas it exists as H2O in an anaerobic (reductive) environment. The oral cavity contains microenvironments that represent these extremes. Molecular oxygen is itself “antimicrobial” toward strict anaerobes (which are routinely found in mature supragingival plaques and are abundant in subgingival plaques). However, ground-state O2 (3O2, triplet oxygen, a di-radical) is not generally included among the so-called “reactive oxygen species” (ROS, Table 2). The ROS of Table 2 can be divided into the radical species (O2·−, · OH) and the “closed-shell” species (1O2, H2O2, and O3). Hydroxyl radical (· OH) is not produced in large quantities by the defensive peroxidases of the oral cavity, so it will not be further discussed here, but we refer the reader to reviews of the role of ROS in periodontal tissue destruction for more information (Waddington et al., 2000; Chapple and Matthews, 2007). In contrast, O2·− is pertinent in that it is produced by NADPH oxidase during neutrophilic respiratory bursts (vide supra). Despite being a radical, O2·− is relatively chemically unreactive, and mammalian cells [and some oral bacteria (Amano et al., 1986)] contain superoxide dismutase that catalyzes the disproportionation of O2·− to give H2O2 and O2 (Packer, 2002). It is noteworthy that the conjugate acid HOO· (hydroperoxyl or perhydroxyl radical) is considerably more reactive than the conjugate base O2·−. It is conceivable that HOO· (pKa ca. 4.8) (Bielski et al., 1985) plays a role in the oral cavity (e.g., during the development of acidogenic plaques). However, since O2·− and HOO· are in rapid acid/base equilibrium, no distinction will be made between the two species in this review (cf. HOCl vs. OCl−, vide infra). Of the closed-shell species, only H2O2, 1O2, and O3 are relevant to our discussion.
Hydrogen peroxide is a powerful oxidant that is produced in the oral cavity by the aforementioned mechanisms (Duox, anaerobic glycolysis, disproportionation of O2·−, etc.). Like O2·−, H2O2 is relatively chemically inert. For example, it takes more than an hour for H2O2 to react with millimolar concentrations of Cys (Ashby and Nagy, 2006a,b). Nonetheless, H2O2 is cytotoxic toward mammalian cells (Ward, 1991) and most prokaryotes (Asad et al., 2004). Singlet oxygen has been detected in saliva (Takahama, 1993; Kou and Takahama, 1995; Sun et al., 2006). However, many of the reported measurements of 1O2 remain controversial, because the probes that are used to detect it tend to be insensitive and frequently non-specific (Martinez et al., 2000). Nonetheless, it is believed that human peroxidases produce 1O2 during their decomposition of H2O2 (Kanofsky, 1991). In addition to the use of therapeutic O3 (Azarpazhooh and Limeback, 2008), it has been suggested that 1O2 is converted to O3via an immunoglobulin-catalyzed reaction (Wentworth et al., 2000, 2002). Secretory IgA (SIgA), the most abundant immunoglobulin in saliva, is also proposed to catalyze the reaction (Uehara et al., 2006). However, the involvement of O3 remains controversial, because the probes that have been used are not specific (Kettle et al., 2004; Smith, 2004; Kettle and Winterbourn, 2005).
As a consequence of their reactive nature, all of the ROS exhibit cytotoxic properties. The relative importance of ROS as defensive agents in the oral cavity is difficult to assess, because various amounts act on microcosm plaques in diverse environments. There are likely synergisms in multi-species plaques. For example, pure cultures of oral streptococci produce H2O2 (they are catalase-negative), but H2O2 is not found in dental plaque or salivary sediment, despite streptococci being major components of their mixed bacterial populations. This is presumably due to the fast consumption of free H2O2 by catalase-positive species of bacteria (e.g., Neisseria, Haemophilus, Actinomyces, and Staphylococcus spp.) (Ryan and Kleinberg, 1995). Furthermore, the SPO and MPO systems of the oral cavity may protect H2O2-sensitive bacteria (Adamson and Carlsson, 1982). In addition to synergism between bacterial species, there are likely to be additive and/or cooperative effects between defensive agents. For example, there is some evidence that O2−· can act synergistically with OCl− (vide infra) to induce oxidative damage (Hawkins et al., 2002) (Table 3).
(Pseudo) Hypohalites
The archetypal example of a biocide is hypochlorite (OCl−, the principal component of household bleach). In sufficient concentrations, it is toxic to all life. Note that the reactive form of OCl− (and the other hypohalites as well) is the corresponding conjugate acid, hypohalous acid (HOCl, pKa = 7.4). As a neutral species, HOCl is presumably membrane-permeable, and therefore more cytotoxic. However, since the acid-base equilibrium between HOCl and OCl− is exceedingly fast, the issue of which species is actually active is academic. Under acidic conditions (and in the presence of excess Cl−), OCl− comproportionates to give Cl2, the corresponding halogen (Adam et al., 1992). However, since the equilibrium between the hypohalites and the halogens is relatively fast, they are not generally treated as different biocides. Instead, the term “total active chlorine” is often used to describe the sum amount of hypohalous acid, hypohalite, and halogen. Furthermore, HOCl/OCl− are the predominant species at physiologic pH. The relative ease with which the halides are oxidized is I− > Br− > Cl− (note that F− is never oxidized in an aqueous environment), so the trend in oxidative strengths of the hypohalous acids is HOCl > HOBr >> HOI. As mentioned before, I− is not abundant in most physiologic fluids, so only Cl− and Br− are relevant to this discussion.
HOCl and HOBr exhibit somewhat promiscuous reaction chemistry (the relative reactivities of HOCl toward proteinaceous groups are Cys ≈ Met >> cystine ≈ His ≈ α-amino > Trp > Lys >> Tyr ≈ Arg > backbone amides > Gln ≈ Asn, and a similar trend is observed for HOBr) (Pattison and Davies, 2001, 2004). However, the kinetics of some of the reactions of HOCl, and especially HOBr, approaches the diffusion limit (Nagy et al., 2006b). As a consequence of their facile reaction, HOCl and HOBr probably exhibit poor chemical selectivity in a biological setting. In addition, secondary reactive species are produced during the reactions of HOCl and HOBr that likely contribute significantly to the overall toxicity—for example, haloamines (Grisham et al., 1984; Abia et al., 1998; Hawkins and Davies, 1998, 2003; Davies, 2005). From a chemical perspective, it is appropriate to view the reactions of HOX (X = Cl, Br) as a redox cascade (thermodynamically downhill) with the eventual production of chemically stable derivatives. Because of the labile nature of many of the intermediate species in such chemical cascades, and because of the fact that many of these species exhibit similar chemistries (e.g., HOCl and chloroamines are both electrophilic chlorinating agents, albeit with different reactivities), it is difficult to chart the consequences of the damage that occurs. Accordingly, the complexity of the chemistry precludes a definitive assignment of the antimicrobial mechanism of HOX (X = Cl, Br). Nonetheless, there have been many studies of the effects of HOX (X = Cl, Br) on both eukaryotes (Hawkins and Davies, 2000; Hidalgo and Dominguez, 2000; Hawkins et al., 2001; Hidalgo et al., 2002; Soszynski et al., 2002; King et al., 2004) and prokaryotes (Albrich et al., 1981, 1986; Albrich and Hurst, 1982; Barrette et al., 1987, 1989; Hurst et al., 1991; Palazzolo et al., 2005). Although it is problematic to follow the reaction cascade that begins with the production of HOX (X = Cl, Br) in vivo, there is considerable interest in evaluation of the relevance of the HOX (X = Cl, Br) chemistry, particularly in the context of the host tissue damage that occurs with inflammatory disease. One approach to monitoring the chemistry of HOX (X = Cl, Br) in vivo is through the use of biomarkers, thermodynamically stable derivatives of HOX (X = Cl, Br). Promising biomarkers that appear to be specific for HOX (X = Cl, Br) include 3-halotyrosines (Winterbourn and Kettle, 2000) and a sulfonamide derivative of GSH (Harwood et al., 2006).
In the field of inorganic chemistry, SCN− is called a “pseudohalide”, because its reaction chemistry frequently mirrors that of the halides (Lappert and Pyszora, 1966). This is reflected in the fact that defensive peroxidases use SCN− (in addition to the halides) as a substrate. The oxidation potential of SCN− falls between those of I− and Br−. Accordingly, the chemical properties of HOSCN are most similar to those of HOI (Nagy et al., 2007a). In contrast to HOCl and HOBr (which react with a variety of functional groups), the only characterized reactions of HOSCN are with thiol moieties (which are among the most powerful nucleophiles), generally Cys and its derivatives (Ashby and Aneetha, 2004; Nimmo et al., 2007; Lemma and Ashby, 2008). Given that SH groups are apparently the targets of the HOSCN, it is important to note that roughly 40% of all enzymes are rendered ineffective by chemical agents that are reactive toward thiols (Leung-Toung et al., 2002). Thus, the destruction of functional SH moieties by HOSCN is one basis for its cytotoxicity. Importantly, the hypohalites can be interconverted through their reactions with other halides, but the process must be exothermic. For example, HOCl (Ashby et al., 2004) and HOBr (Nagy et al., 2006b) react with SCN− to give HOSCN and the corresponding halides, but not vice versa. In vivo, these reactions are important because they restrict the lifetimes of the more powerful (and less discriminant) hypohalites, thereby limiting their propensity to cause collateral host tissue damage. In addition, the antimicrobial OSCN− is produced in the reaction (Nagy et al., 2006a).
Halides
Fluoride is the only halide that is known to be antimicrobial without oxidation (Hamilton, 1990; Van Loveren, 1990; Marquis, 1995; Jenkins, 1999). Fluoride influences the metabolism of cariogenic and other bacteria via multiple mechanisms (Marquis et al., 2003): F− can bind directly to many enzymes (especially metalloenzymes) (Segal et al., 1968; Wever and Bakkenist, 1980; Zgliczynski et al., 1983; Thibodeau et al., 1985; Ferrari et al., 1997; Suzuki and Ohshima, 2003), thereby affecting their activities; catalase is inhibited by F− [thereby affecting the ability of H2O2 to kill oral bacteria (Phan et al., 2001)]; and some F− complexes of metals (e.g., AlF4− and BeF3−·H2O) can mimic phosphate, thereby affecting a variety of enzymes and regulatory phosphatases (Thongboonkerd et al., 2002). The weak-acid property of HF (pKa = 3.15), which is a transmembrane proton conductor, appears to be important for inducing the glycolytic inhibition of oral bacteria that is observed at low pH in dental plaque (Eisenberg and Marquis, 1981) (Table 4).
Nitrogen Derivatives
After many decades of angst about nitrates in our diet and their propensity to form potentially carcinogenic N-nitroso derivatives (Eichholzer and Gutzwiller, 2003), there is mounting evidence that NO3− is concentrated in saliva for beneficial purposes (McKnight et al., 1999) (although the potential deleterious properties of NO3− on systemic health should not be discounted). The concentration of NO3− in saliva is proportional to the dietary intake (Eisenbrand et al., 1980), it varies with the salivary flow rate (Granli et al., 1989), and it is influenced by smoking (Tsuchiya et al., 2002). Since NO3− is a relatively inert chemical species, the mechanism of antimicrobial action of NO3− probably involves a redox cascade. Facultative anaerobic bacteria in the oral cavity use NO3− as a terminal electron acceptor (cf. O2 for aerobic bacteria) to produce nitrite (NO2−). Acidified NO2− inhibits the growth and affects the survival of cariogenic bacteria (e.g., Streptococcus mutans, Lactobacillus casei, and Actinomyces naeslundii) (Silva Mendez et al., 1999). Acidified NO2− has a similar effect on periodontal bacteria (e.g., Fusobacterium nucleatum, Eikenella corrodens, and Porphyromonas gingivalis) (Allaker et al., 2001). Importantly, while the growth of periodontal bacteria is known to be inhibited by acid in the absence of NO2−, there is a dose-dependent decrease in these bacteria in the presence of NO2− (Allaker et al., 2001). Nitrous acid (HNO2, pKa = 3.4) is unstable and will spontaneously disproportionate: 3 HNO2 → H3O+ + NO3− + 2 ·NO. Some bacteria possess nitrite reductase (e.g., S. mutans), an enzyme that is capable of accelerating the disproportionation of NO2− (Choudhury et al., 2007). It appears that nitric oxide (·NO) is the antimicrobial component of the NO3−/NO2−/·NO redox cascade (Fang, 1997; Smith et al., 1999; Sato et al., 2008). The mechanism by which ·NO induces cell death is the subject of ongoing investigation. Alternative models include “oxidative stress” and “nitrosative stress” [e.g., nitrosylation of proteins without a major alteration in cellular redox state (Eu et al., 2000)]. Nitric oxide also reacts with O2·− to produce peroxynitrite (ONOO−), which may also contribute to collateral host tissue damage in the oral cavity (Lohinai and Szabo, 1998; Lohinai et al., 2001; Barley et al., 2004).
INTER-PERSON DIFFERENCES
There is considerable variability in the physiological concentrations of many of the chemically stable inorganic ions that have been discussed herein. In most cases, these differences can be attributed to diets or smoking. Given the influence of these ions on the activity and function of the defensive peroxidases, there has been some interest in correlating interperson differences to oral disease. As a caveat, it is important to note that many of the relevant inorganic ions are chemically reactive, and consequently the abundance of free ions may not reflect their relevance in vivo. Some ions may exist as transient species (e.g., OCl−) or as their conjugates with reaction partners (e.g., CN− and OCN−). In some cases, steady-state concentrations of reactive species may accumulate (e.g., OSCN−), but measured concentrations may not reflect the time-dependent flux of such species.
Thiocyanate
The main source of SCN−in vivo is CN−, vide infra. Cyanide is principally introduced by the digestion of glucosinolate-containing vegetables (e.g., the Brassica) (Weuffen et al., 1984). However, as a consequence of detoxification of hydrogen cyanide (HCN, pKa = 9.2, which is known to be present in microgram amounts per cigarette), the level of SCN− in smokers is considerably elevated relative to that in non-smokers. Indeed, this difference is routinely used as a biomarker for the evaluation of smoking behavior (Morabia et al., 2001). While most of the CN− that is converted to SCN−in vivo is exogenous in origin, endogenous sources contribute the sulfur vis-à-vis a multitude of reactions, some of which are enzyme-catalyzed (Wood, 1975). It has been shown that higher concentrations of SCN− in saliva can contribute to an enhancement of peroxidase activity (Tenovuo, 1976; Lamberts et al., 1984; Fonteh et al., 2005; Tahboub et al., 2005). Furthermore, as noted earlier, SCN− is a potent sequestering agent for some reactive oxidants (Ashby et al., 2004; Nagy et al., 2006b). Accordingly, one might conclude that higher SCN− (either as a consequence of diet or through smoking) should result in a suppression of oral disease. However, while smokers have elevated SCN− and OSCN− levels in their saliva, no corresponding correlation with dental caries has been observed (Lamberts et al., 1984). Interestingly, although there is not an association between smoking and caries among adults, there is a positive correlation between environmental (second-hand) tobacco smoke and primary tooth caries in children (Shenkin et al., 2004). In contrast to caries, there is a strong correlation between smoking (and consequently SCN− levels) and some periodontal diseases (Rivera-Hidalgo, 2003). Since there is also a correlation between smoking and SCN− levels in GCF, SCN− levels presumably exhibit a positive correlation with periodontal diseases. However, SCN− is only one of many inorganic and organic chemicals that are elevated by smoking. Two other inorganic ions are CN− and cyanate (OCN−), vide infra.
Cyanide
It addition to dietary sources (e.g., cyanogenic glucosides, vide supra) and tobacco smoke, other sources of CN−in vivo include micro-organisms (in particular certain pseudomonads) and industrial exposure (e.g., vis-à-vis organonitriles) (Wong-Chong et al., 2006). There appear to be no studies that have determined the normal concentration of CN− in the fluids of the oral cavity. However, there is a statistical correlation between blood and salivary SCN− levels (Tsuge et al., 2000). As noted earlier, the concentrations of SCN− in physiological fluids (including GCF and saliva) in smokers are substantially higher than those for non-smokers. Therefore, it follows that the oral cavity is exposed to higher concentrations of CN− for smokers compared with non-smokers. HCN contributes to the loss of peroxidase activity in saliva upon exposure to cigarette smoke (Klein et al., 2003). This is due to the strong complexation of CN− to the iron-active sites of the peroxidases (fundamentally the same mechanism that renders CN− toxic to the respiratory system vis-à-vis the complexation of hemoglobin and myoglobin). It is noteworthy that there is no correlation between CN− in plasma and the concentration of HCN in breath (Lundquist et al., 1988). Furthermore, the concentrations of HCN measured in breath are higher than expected for blood concentrations, which suggested a local production of HCN in the oropharynx. Under some circumstances, OSCN− can decompose to give CN− (Aune and Thomas, 1977). While the major cyanide-derived product of the decomposition of OSCN− is OCN−, we have observed the formation of substantial amounts of CN− during the hydrolysis of thiocyanogen [(SCN)2, analogous to a halogen] at neutral pH (unpublished observations). It is conceivable that the latter reaction is the source of HCN in breath.
Cyanate
Cyanate (OCN−) is produced by the oxidation of CN−. Thus, OCN− levels are higher in smokers. Under some conditions, the decomposition of OSCN− also produces OCN− (Oram and Reiter, 1966). MPO is inhibited by OCN− (Qian et al., 1997). The inhibition could be caused by heme binding of OCN− (thereby blocking the active site) or by carbamylation of the protein by OCN−. It is noteworthy that the functional impairment of proteins through carbamylation by OCN− is thought to promote human inflammatory diseases (Wang et al., 2007). However, the possible relevance of carbamylation in the oral cavity has not been investigated. There is very limited information available regarding the effect of OCN− on oral bacteria (Morita, 1977).
Fluoride
In the absence of supplementation by fluoride-containing dentrifices, the concentration of F− in saliva (and presumably GCF) is somewhat lower than the concentration in plasma (Oliveby et al., 1989). The normal concentration of F− in saliva (ca. 1 μM) can increase markedly after the ingestion of F−, and change dynamically thereafter (Dawes and Weatherell, 1990). An early study that used PMN granules (and not isolated enzymes) suggested that F− does not influence the activity of MPO (which was determined by measurement of the iodination of zymosan) (Gabler and Leong, 1980). In contrast, the same study reported that the iodination of zymosan was inhibited by F− for intact PMNs (Gabler and Leong, 1980). However, the pH-dependent competitive inhibition of isolated LPO (Segal et al., 1968; Thibodeau et al., 1985) and MPO (Zgliczynski et al., 1983; van den Abbeele et al., 1992) by F− has been demonstrated. For example, half of the activity of bovine LPO occurs for < 0.05 mM at pH 4, 0.3 mM at pH 5, 4.0 mM at pH 6, and greater than 10 mM at pH 7, as assayed with 5 mM I− and 150 μM H2O2 (Thibodeau et al., 1985). It is noteworthy that SPO was found to have lower pH optima relative to LPO, but it also was inhibited by F− at sufficiently low pH (Thibodeau et al., 1985). A similar inhibitory effect of F− on peroxidase activity has been observed in whole saliva (Thibodeau et al., 1985; Hannuksela et al., 1994; van den Abbeele et al., 1995). These observations suggest that F− in dental plaque may inhibit the peroxidase defense system. However, when F− and OSCN− are added simultaneously at pH 5.0, an additive effect of growth inhibition of S. mutans was observed (Lenander-Lumikari et al., 1997). Thus, the small inhibitory effect of F− on the defensive peroxidase systems of the oral cavity may be offset by the combinatorial antimicrobial effects of F− and OSCN−.
Nitrate and Nitrite
Basal levels of NO3− in saliva are 10–20 times those found in plasma (and presumably GCF) (Duncan et al., 1995; Benjamin and McKnight, 1999; Pannala et al., 2003; Lundberg et al., 2004). For the average diet in the US, ca. 80% of dietary NO3− is derived from vegetables (White, 1975). Cured meats, for example, typically represent a minor source of NO3− in most diets. However, urinary, plasma, and saliva NO3− concentrations increase markedly after the consumption of a high-nitrate meal (Pannala et al., 2003). The maximum concentration is achieved within a few hours following the meal, and NO3− concentrations return to basal levels within 24 hours. The amount of NO3− that is excreted in the urine following the consumption of a high-nitrate meal suggests that the majority of urinary NO3− can be accounted for in dietary sources (Pannala et al., 2003). In contrast to NO3−, an increase in NO2− is observed in saliva only after the consumption of a high-nitrate meal (Pannala et al., 2003), which is consistent with the fact that NO3− is metabolized to NO2− by bacterial flora on the posterior surface of the tongue in rat models (Duncan et al., 1995). Nitrite has been shown to enhance the reactivity of LPO (Reszka et al., 1997, 1998, 1999; van der Vliet et al., 1997; Gebicka, 1999; Bruck et al., 2001) and MPO (Burner et al., 2000). In addition to an enhancement in the rate of catalysis by NO2−, it has been suggested that NO3− reduces acidity in the oral cavity (Li et al., 2007). Thus, NO2− and NO3− may play a beneficial role in the oral cavity. We note that there has apparently been no attempt to correlate oral health with inter-person differences in nitrate-reducing capacity.
PEROXIDE TOXICITY AND PEROXIDASE FUNCTION
At sufficient concentrations, H2O2 is cytotoxic to mammalian cell lines, including human epithelial cells (Mattana et al., 1992) and gingival fibroblasts (Tenovuo and Larjava, 1984; Tipton et al., 1995a). While H2O2 is a relatively chemically inert molecule, its homolysis to give ·OH radicals is catalyzed by transition metals, particularly iron, in a reaction that is referred to as Fenton chemistry (Prousek, 2007). Much of the cytotoxicity of H2O2 is attributed to Fenton chemistry (Winterbourn, 1995). In the presence of SCN−, the LPO system protects cultured mammalian cells against H2O2 toxicity (Hänström et al., 1983). This is consistent with the observation that OSCN− is not toxic toward mammalian cells (Bjoerck and Claesson, 1980; Marshall and Reiter, 1980; White et al., 1983; Carlsson et al., 1984; Carlsson, 1987). It has been previously suggested that one of the important roles of human peroxidases is to detoxify H2O2 to prevent host tissue damage (Carlsson, 1987; Tipton et al., 1995b). Since many prokaryotes are also sensitive to H2O2, the human peroxidase systems may also protect certain bacteria by sequestering H2O2. However, the OSCN− that is produced is inhibitory toward most bacteria. Consequently, it is difficult to envisage the net effect of H2O2 sequestration vs. OSCN− on bacterial growth (Fig. 2).
THE ANTIMICROBIAL DEPLETION MODEL
The efficacy of chemically reactive antimicrobial agents can be diffusion-limited. This is one (but certainly not the only) explanation for the greater sensitivity of planktonic cultures to chemically reactive antimicrobials (Stewart, 1994; Stewart and Raquepas, 1995; Dodds et al., 2000; Stewart et al., 2001, 2004; Hunt et al., 2005). Planktonic cultures are generally agitated, and consequently fluid flows by a convection mechanism that transports solutes rapidly. In contrast, fluid flow in biofilms can be highly restricted. For thick biofilms, such as the plaques of the oral cavity, the primary mechanism of solute transport is diffusion (unagitated flow). For highly reactive molecules such as HOCl, which also exhibit promiscuous reaction chemistry, the biomass of plaques offers a formidable barrier to penetration. While HOCl is an extremely effective antimicrobial toward planktonic cell suspensions, it is a comparatively inefficient killing agent with respect to biofilms. For example, a 600-fold increase in concentration of HOCl (per cell) is required to kill certain biofilms of Staphylococcus aureus, compared with planktonic cultures of the same species (Luppens et al., 2002). More chemically selective antimicrobials, e.g., OSCN−, are more likely to penetrate thick biofilms before reacting. The relationship between chemical reactivity of the antimicrobial and biofilm penetrability forms the basis of the Antimicrobial Depletion Model (Fig. 2). There is presumably an inverse relationship between penetrability with respect to the antimicrobial reactivity. However, penetrability, of course, is not the only factor that determines the efficacy of the antimicrobial: OSCN− is generally considered to be bacteriostatic, whereas OCl− is bactericidal. But, as mentioned before, there is a tradeoff for the higher reactivity/poorer selectivity of OCl−: the potential of host tissue damage. In the oral cavity, it may be advantageous for the body to use a collection of antimicrobials that exhibit a continuum of reactivity. Thus, complementing HOCl is taurine chloroamine (TauCl), a less reactive (Carr et al., 2001) and more selective (Peskin and Winterbourn, 2006) electrophilic chlorinating agent that may play a role in host defense in the oral cavity (Woldring, 1955; Mainnemare et al., 2004). Similarly, we have discovered a corresponding less-reactive derivative of OSCN−, thiocarbamate-S-oxide [H2NC(=O)-S-O−] (Nagy et al., 2007c). Thiocarbamate-S-oxide, which is formed by the hydrolysis of OSCN−, is one of the chemical species formed during the redox cascade that results in the decomposition of OSCN− (Nagy et al., 2007c). Recently, we have learned that H2NC(=O)-S-O− reacts with thiols via a mechanism that is analogous to the one that is observed for OSCN−, albeit with slower kinetics (unpublished observations). While the effect of H2NC(=O)-S-O− on bacterial physiology remains to be investigated, its discovery demonstrates that there is much to be learned about the inorganic antimicrobials produced by the defensive peroxidases.
The Antimicrobial Depletion Model does not explicitly include the possibility that defensive peroxidases might generate hypohalites within the biofilms, thereby obviating the need for reactive species to diffuse into biofilms. We note that the production of acid in salivary sediment is more effectively inhibited when OSCN− is produced by sediment-bound peroxidases than when salivary sediment is treated with exogenous OSCN− (Tenovuo, 1979). However, salivary sediment has a larger buffer capacity and “organic load” (e.g., non-viable cells) than plaque (Singer et al., 1983), so it is not clear whether the penetration of OSCN− into plaque is comparably diffusion-limited. The production of hypohalites from within plaques requires the transport of the components of peroxidase systems into the biofilm: peroxidase, (pseudo) halide, and H2O2. Relevant to the issue of diffusion of peroxidases into plaques, it has been previously observed that LPO binds to S. mutans (LPO is a cationic protein, and the outer membranes of Gram-positive bacteria are negatively charged) (Pruitt et al., 1979). While cell-bound LPO remains catalytically active initially, it is eventually inactivated (Pruitt et al., 1979). Thus, the electrostatic attraction of the peroxidases for cells and the subsequent inactivation of the cell-bound enzyme may preclude the generation of hypohalites within a thick plaque (Pruitt et al., 1979). With regard to the availabilities of the other components of the peroxidase defense systems, it is probable that the concentrations of the (pseudo)halides in plaques reflect their concentrations in the physiologic fluids that surround them (i.e., saliva or GCF), but it was noted earlier that H2O2 has not been found in dental plaques (Ryan and Kleinberg, 1995). The issue of whether or not the human defensive peroxidase systems are active in thick plaques remains unresolved.
CONCLUSION
The distinctive chemical environments of the supragingival and subgingival regions impose restrictions on the human peroxidase defense strategies of the oral cavity (e.g., vis-à-vis substrate bioavailability). Inter-person differences may also influence the function and activity of the enzymes and the chemistry of the reactive species that the enzymes generate. The abilities of the defensive agents produced by the peroxidase systems to differentiate between host tissues and the microbiota are an unsettled issue. Oral diseases are accompanied by microbial shifts of dental plaque, so a more subtle issue is whether these agents differentiate between commensal and pathogenic microbiota. While it seems likely that ecological principles, as in the March ecological hypothesis (Marsh, 2003), can be applied to explain microbial shifts, it is not yet clear what ecological pressures are inducing these microbial shifts, nor is the root cause of host tissue damage completely clear. For dental caries in enamel, damage is clearly a direct effect of the plaque. In dentinal and root caries, some host response may be involved (as in periodontitis). The microbial shift during cariogenesis is toward acidogenic and acid-tolerant Gram-positive bacteria, which displace the acid-sensitive commensal microbiota that are associated with intact tooth tissues. Since supragingivally generated OSCN− targets Gram-positive bacteria (S. mutans may be particularly affected), it is possible that OSCN− may be important in modifying plaque cariogenicity. Other important health-maintaining functions of OSCN− may also occur, such as restricting supragingival intra-oral cross-infection by periodontal pathogens. During the development of periodontal diseases, an increase in disease-associated anaerobes occurs in conjunction with increased inflammation. The inflammatory response is primarily due to antigens that have been introduced by the bacteria. It is unclear whether the inflammation is due to changes in the antigens introduced by the microbial shift (Lamster and Novak, 1992), or whether the microbial shift has been induced by the inflammatory response, or both (Dalwai et al., 2006). Is the host tissue damage of periodontal diseases solely the consequence of host mediators (e.g., HOCl), or is there direct attack by bacterial virulence factors such as proteolytic enzymes? Definitive answers to these questions are critical for a rational approach to combating oral disease—Does one treat the inflammation, or does one target pathogenic bacteria, or is it necessary to treat both (Van Dyke, 2007)? Other risk factors being similar, is inter-individual variation in peroxidase activity a key factor in determining why some people develop oral disease and others do not? Whether inter-person differences between the inorganic chemistry of the oral cavity contribute to oral diseases is a topic that deserves further attention.
Two-electron Redox Couples for X− (Eo, pH = 7 vs. SHE), Apparent Rate Constants (k) of MPO Compound I (x 10−4M−1s−1) with X−,a Reference Range Values (RRV) of X− in Physiologic Fluids (μM or mM), and Specificities (S)b for Oxidation of X− by MPO (consensus substrates in
Major Inorganic (Reactive) Oxygen Species in the Oral Cavity
Major Inorganic (Pseudo) Hypohalites in the Oral Cavity
Major Inorganic (Reactive) Nitrogen Compounds in the Oral Cavity
Spatial relationship between the inorganic host defense factors of the oral cavity and the ion gradients that influence their relative abundance. Refer to the text for an explanation of pathways A-E and the meaning of the variable Z.
Relationship between the reactivities of inorganic defense factors of the oral cavity and maturation of plaque biofilms. The penetrability of the hypohalites and their derivatives is inversely related to their reactivities and the extent of maturation of the biofilm. See the text for a discussion of the possibility that the defensive factors could be generated within biofilms.
Footnotes
Notes
Acknowledgements
The author is supported by Public Health Service grant R21 DE016889-01A2 from the National Institute of Dental and Craniofacial Research.