Cationic antiseptics: diversity of action under a common epithet

Authors


Peter Gilbert, School of Pharmacy and Pharmaceutical Sciences, University of Manchester, Manchester M13 9PL, UK (e-mail: peter.gilbert@man.ac.uk).

1. Summary

Cationic antimicrobials have been in general use within clinical and domestic settings for over half a century. Recently, the use of antiseptics and disinfectants has been questioned in such settings because of the possibility that chronic exposure of the environment to such agents might select for less susceptible strains towards these agents and towards third party antibiotics. Whilst no supportive evidence has emerged from retrospective field studies of high use environments such debate has tempered new applications for these molecules. In the clinic, use of antiseptics, together with products, such as dressings, catheters and sutures, which are impregnated with biocides has increased. Prominent amongst these biocides are the cationics. Much of the research pertaining to the mechanisms of action of cationic antibacterials was conducted in the 1960s and 1970s and has not been subject to extensive review. Analysis of available publications suggest that monoquaternary ammonium compounds (QAC, cetrimide, benzalkonium chloride), biquaternaries and bisbiguanides (Chlorhexidine, Barquat), and polymeric biguanides (Vantocil, Cosmocil) whilst having similarities in action mechanism, differ substantially in the nature of their interaction with cell envelopes. This has profound implications in terms of cross-resistance where changes in susceptibility towards QAC is not reflected in changes towards other cationics. This review examines action mechanisms for these agents and highlights key differences that render them distinct categories of antibacterial agent.

2. Perspectives

In recent years there has been some questioning of the potential of antiseptic residues, accumulating within high use environments such as clinics and hospitals, to select for bacteria that are altered with respect to their susceptibility not only towards other antiseptics but also towards third party agents such as antibiotics. The utility of cationic antibacterials to combat cross infection is undeniable, as is the overall contribution of antisepsis to the reduction of hospital-acquired infection. Indeed, reductions in antibiotic use brought about by such policies can be argued to have had a positive impact upon antibiotic resistance development. Nevertheless, re-examination of the action and resistance mechanisms associated with all antibacterial molecules destined for use in a clinical and domestic setting is appropriate. Cationic antimicrobials have been widely deployed in antisepsis for well over half a century without any apparent reduction in their effectiveness. They remain the mainstay of routine chemical antisepsis and disinfection. Amongst the commonly deployed cationic antimicrobials are the quaternary ammonium compounds (QAC; cetrimide, benzalkonium chloride), bisbiguanides (chlorhexidine, hibitane) and polymeric biguanides (vantocil). All of these positively charged molecules bind strongly to the cell walls and membranes of bacteria because of their opposite, negative charge. Disruption of the target cell is brought about by perturbation of these sites. The nature of the interaction, following binding, determines activity and the potential for resistance development. The purpose of this article is to consider the mechanism of action of cationic antimicrobials, possible resistance mechanisms and their use in infection control highlighting key differences that make each group of molecules distinct.

3. Background

Problems associated with the development and spread of antibiotic resistance have been increasing since the early 1960s and, in the clinic, is currently viewed as a major threat to the treatment of hospital and community-acquired infection (McMurray 1992; Kaatz et al. 1993) with as many as one-third of nosocomial infections believed to be preventable (Senior 2001). Such resistance has generally been associated with the overuse and abuse of therapeutic agents, and with the acquisition and fusion of genetic elements encoded within plasmids. It is widely accepted that the main cause of this problem has been, and still is widespread inappropriate use and over-prescribing of antibiotics in clinical medicine, animal husbandry and veterinary practice (Rao 1998; House of Lords Select Committee on Science and Technology 1998; Feinman 1999; Georgala 1999; Magee et al. 1999; Dixon 2000). Concerns about bacterial resistance have led to calls for increased education, of both public and professionals, on the correct use of antibiotics. Additionally, more stringent infection control measures have been advocated in order to reduce the transmission of infection (Anon 1997, 1999a,b, 1999c; Hart 1998; Cristino 1999; Dwyer 1999; Smith et al. 1999; Waldvogel 1999; Dixon 2000). These measures recognize the tremendous contributions that antisepsis has made, over the last century, towards our current advanced state of public health. Indeed, if reductions in the number of infections requiring antibiotic treatment can be achieved through effective hygiene, including the use of antiseptic products, then this will delay increases in the incidence of antibiotic resistance. Accordingly, it is important to ensure that the use of antiseptic products is not discouraged in situations where it is part of good hygienic practice and where there may be tangible reductions in the transmission of infection. With respect to the management of postoperative wounds then, whilst multiple factors related to the nature of the surgical procedure can influence the risk of wound infection, their approximate incidence, for clean procedures, is between 2 and 3% (Futoryan and Grand 1995). The normal flora of the skin is an important source of serious postoperational infections with the involvement of skin organisms such as Staphylococcus aureus and Staphylococcus epidermidis being widely acknowledged. Furthermore, antibiotic resistant coryneform bacteria have been isolated from the skin of both hospitalized patients and control groups (Larson et al. 1986). Systemic infections caused by such bacteria are often associated with concurrent use of indwelling medical devices such as central venous lines or catheters (Passerini et al. 1992), but systemic treatment with vancomycin neither eliminates nor prevents colonization of the device (Larson et al. 1986). Topical application of broad-spectrum antimicrobial agents such as QACs, biguanides, halogen release agents and triclosan remain safe and effective preventative, and treatment, measures. In this respect, a chlorhexidine containing medication was the only formulation to thoroughly eliminate both 1-day and 3-day-old Enterococcus faecalis biofilms (Lima et al. 2001). Whilst such use has generally been confined to medicated soaps, handwashes and bathing formulae, incorporation of these and other antimicrobial agents within the polymer materials and coatings that comprise indwelling medical devices and dressings have demonstrated significant applications in the localized prevention of infection (Storch et al. 2002).

For over a century cationic antimicrobials have been used both in infection control and within many consumer products and have often been assumed to possess a single, generic mechanism of action directed towards biological membranes. Cationic antimicrobials that have been in use for over 40 years include a variety of quaternary ammonium-based molecules (cetrimide, benzalkonium chloride), bisbiguanides (chlorhexidine) and polymeric biguanides (VantocilTM; Arch Chemicals, Blackley, UK). Sadly, in spite of their long and widespread use, assumptions relating to such agents are compounded by a general lack of experimental evidence surrounding their biological mechanisms. Recently the use of such antimicrobial agents has been questioned in many application areas. This questioning is based on an association between trace levels of antimicrobial residue and their implied potential to select for less susceptible bacteria that are co-incidentally resistant to third party antibiotics (Gilbert and McBain 2002, 2003). QACs are used extensively in the food processing industry to prevent the persistence of pathogens such as Escherichia coli and Listeria monocytogenes in the environmental microflora (Holah et al. 2002). Biofilm formation is thought to play an important role in the survival of virulent strains of food-related staphylococci. Staphylococci isolated from the food industry were found to vary greatly in their ability to form biofilms, but biofilm formation was positively correlated with resistance to QACs (Moretro et al. 2003). Chlorhexidine is also a well-known anti-plaque biocide that plays a crucial role in the reduction of supragingival plaque and treatment of gingivitis. Large numbers of common oral bacteria isolated from patients using chlorhexidine indicate no increase in microbial resistance to chlorhexidine, or to commonly used antibiotics (Sreenivasan and Gaffar 2002). An effective strategy for the reduction of legionella in water cooling systems is regular disinfection with polyhexamethylene biguanide (PHMB) (Kusnetsov et al. 1997).

Although a number of laboratory studies have demonstrated links between the exposure of pure cultures to sublethal concentrations of biocides and changes in antibiotic and antimicrobial susceptibility (Braoudaki and Hilton 2004; Joynson et al. 2002; Tattawasart et al. 1999), there is little or no evidence suggesting that this is a significant factor in the development of antibiotic resistance in clinical practice (Loughlin et al. 2002; Gilbert and McBain 2003). The current view is that if biocides do play a role, then it is likely to be a very minor one but it is a subject that requires constant review (Russell 2003). Biocides have played a major part in reducing the number of nosocomial infections through effective hygiene, so it is important to ensure that biocide use is not discouraged in situations where there is real benefit. Conversely, it is also necessary to assess the possibility that widespread and inappropriate use of biocides could compromise their in-use effectiveness (Bloomfield 2002; Braoudaki and Hilton 2004).

A number of mechanisms account for the wide range of sensitivity noted for the antibacterial action of antibiotics and biocides (Heinzel 1998). Some organisms and genera, by virtue of the absence of critical targets sites or an inability of the agents to accumulate at those targets, are intrinsically resistant to particular groups of agent under all growth conditions (Hancock 1998). Other groups of organism may undergo phenotypic changes in susceptibility that reflect the conditions under which they were cultivated or exposed, the temporary expression of efflux pumps or synthesis and export of protective enzymes (inductive change), or mutations in the genes encoding or regulating a sensitive target site (chromosomal change). Growth as a biofilm also reduces the susceptibility profile and is probably caused by a variety of factors including nutrient depletion within the biofilm, reduced access of the biocide to cells in the biofilm, chemical interaction between the biocide and the biofilm, and the production of degradative enzymes and neutralizing chemicals (Gilbert et al. 1990, 2002; Brown and Gilbert 1993; McDonnell and Russell 1999; Campanac et al. 2002).

Polyhexamethylene biguanide is a polymeric cationic antimicrobial agent that has been deployed in consumer applications for over 40 years. Recently it has been used in the treatment of Acanthamoeba keratitis and as an adjunct in various wound dressing materials. Whilst PHMB shares many of the attributes of other, simpler, cationic agents, it has additional actions that render it unique amongst this generic class of antimicrobials.

The purpose of the current article is to consider the mechanism of action of the family of cationic antimicrobials, possible resistance mechanisms and the potential impact of their use on resistance development. The molecular basis of antimicrobial action will therefore be considered from the most simple (monoquaternary ammonium compounds), through bisbiguanides to PHMB. At each stage the potential for resistance development will be considered against a background of published susceptibility surveillance articles (Gilbert and McBain 2003).

4. Cationic antimicrobial agents

The outermost surface of bacterial cells universally carries a net negative charge, often stabilized by the presence of divalent cations such as Mg2+ and Ca2+. This is associated with the teichoic acid and polysaccharide elements of Gram-positive bacteria, the lipopolysaccharide of Gram-negative bacteria, and the cytoplasmic membrane itself. It is not therefore surprising that many antimicrobial agents are cationic and have a high binding affinity for bacterial cells. Often, cationic antimicrobials require only a strong positive charge together with a hydrophobic region in order to interact with the cell surface and integrate into the cytoplasmic membrane. Such integration into the membrane is sufficient to perturb growth and at the treatment levels associated with antiseptic formulations is sufficient to cause the membrane to lose fluidity and for the cell to die. Indeed, for many decades such compounds have been loosely designated as ‘membrane active agents’ or as ‘biological detergents’ broadly recognizing their lack of specificity in mechanism of action. It is however worthwhile considering for a moment the general characteristics of microbial membranes in order to fully understand the mode of action of cationic antimicrobials. The membranes are composed primarily of proteins, embedded within a lipid matrix and approximating to a bilayer (Singer and Nicolson 1972). The proteins either fully traverse the two sides of the bilayer (integral proteins) or are peripheral and associate with one specific side. Many of these membrane proteins are required in order to maintain the structural integrity of the membrane, whilst others are functional and associated with catabolism, cellular transport and the biosynthesis of wall and extracellular products (toxins, virulence factors, etc.). The hydrophobic environments of the neighbouring phospholipids moderate the functionalities of these proteins. Thus, each protein is surrounded by particular phospholipids that interact with the protein. In many cases the precise nature of the phospholipid head group assists maintenance of an active configuration for the enzyme. The lipid bilayer is further stabilized by divalent cations such as Ca2+. Cationic antimicrobials are relatively hydrophobic but interact initially with the wall and membrane by displacing these divalent cations. Such action is shared with simple chelating agents such as EDTA and EGTA that perturb membrane structure solely through the sequestration of stabilizing metal cations. Subsequent interactions of the cationic biocides with membrane proteins and lipid bilayer depend upon the specific nature of the biocide. Many of the cationic antimicrobials have been deployed as surface and topical antimicrobials, in the clinic and in general hygiene delivery, for more than half a century. Notable amongst these agents are the QAC (benzalkonium chloride, cetrimide, Barquat), bisbiguanides (chlorhexidine) and polymeric biguanides (VantocilTM, CosmocilTM; Arch Chemicals) together with antibiotics such as Polymyxin and Tyrocidin. Of these the quaternary ammonium group of compounds and the polymeric biguanides are mixtures of compounds that share a common generic structure. Even within groups such as the monoquaternaries the use of simple vegetable oils as the synthetic starting material for their chemical synthesis leads to the marketed molecules being mixtures with very diverse compositions with respect to the composition of different alkyl chain substituents (Daoud et al. 1983; Gilbert and Al-Taae 1985). Such chemical diversity broaden the spectrum of activity but makes standardization between batches and different manufacturers difficult.

In considering the action and utility of the cationic biocides, together with their potential for resistance development, it is worthwhile characterizing them according to the number of cationic groupings per molecule. Thus the QACs are generally monocationic, whilst bisbiguanides (chlorhexidine) carry two cationic groups separated by a hydrophobic bridging structure (hexamethylene), and polymeric biguanides are polycationic linear polymers comprising a hydrophobic backbone with multiple cationic groupings separated by hexamethylene chains.

4.1 Quaternary ammonium compounds

The QACs are amphoteric surfactants, generally containing one quaternary nitrogen associated with at least one major hydrophobic substituent (Fig. 1). Cetrimide USP is tetradecyltrimethylammonium bromide whereas the generic term Cetrimide relates to mixtures of n-alkyltrimethyl ammonium bromides where the n-alkyl group (the hydrophobic moiety) is between eight and 18 carbons long. Benzalkonium chlorides are always mixtures of n-alkyldimethylbenzyl ammonium chlorides where the n-alkyl groups can be of variable length within a specified range. In addition to these, various dialkylmethyl ammonium halides and dialkylbenzyl ammonium halides also have antimicrobial activity and are variously deployed as biocides, preservatives and antiseptics (Brannon 1997). All share a common mechanism of action. The raw materials providing the alkyl group of these synthetic compounds are often natural oils such as coconut or soya bean oil. These natural oils are heterogenous mixtures of fatty acids with carbon chain lengths of between 6 and 22. Whilst 10, 12, 14 and 16 carbon fatty acids dominate their relative abundance varies from batch to batch. Commercially available QACs that utilize natural oils as the source of the alkyl chain substituents will therefore be highly diversified not only in their fatty-acyl chain length distributions but also in the degree of C–C saturation. Each of these factors will significantly affect antimicrobial activity.

Figure 1.

Structure of tetradecyldimethylbenzyl ammonium chloride (one of the benzalkonium chlorides where the major n-alkyl group is C14 (a), and tetradecyltrimethylammonium bromide (Cetrimide USP) (b). Counter-ions are not shown. Positively charged head groups are shaded blue. Major hydrophobic chains are shaded purple

The activity of quaternary ammonium biocides is an approximate parabolic function of the compounds lipophilicity (n-alkyl chain length) (Daoud et al. 1983; Gilbert and Al-Taae 1985). For Gram-positive bacteria and yeast, such activity maximizes with chain lengths of n = 12–14, whilst for Gram-negative bacteria, optimal activity is achieved for compounds with a chain length of n = 14–16. Compounds with n-alkyl chain lengths of <n = 4 or >n = 18 are virtually inactive. As the antimicrobial activity of QACs towards specific species of bacteria is dependent upon the hydrophobicity of the n-alkyl chain then, given the raw material source of the n-alkyl substituent, the overall activity of individual commercial products can be highly variable (Daoud et al. 1983; Gilbert and Al-Taae 1985). Many quaternary ammonium mixtures are however blended so as to optimize activities against specific groups of bacteria, or to gain as broad a spectrum of activity as is possible. Pharmacopoeial standards for such molecules define the mixture chain lengths, but such standards differ between regulatory bodies.

The mode of action of QAC against bacterial cells is thought to involve a general perturbation of lipid bilayer membranes as found to constitute the bacterial cytoplasmic membrane and the outer-membrane of Gram-negative bacteria. Such action leads to a generalized and progressive leakage of cytoplasmic materials to the environment. Low concentrations of QAC bind firmly to anionic sites found on the membrane surface, cause cells both to lose osmoregulatory capability and to leak potassium ions and protons (Lambert and Hammond 1973). Intermediate levels perturb membrane-located physiologies such as respiration, solute transport, and cell wall biosynthesis (Salt and Wiseman 1970). The high concentrations used in many biocidal formulations however, kill cells by solubilization of the membranes to release all of the cells contents, hence their designation as biological detergents (Salton 1951, 1968). Indeed, the surfactant properties of QACs are often used to good advantage in disinfectant cleansing formulations (Hugo 1956). At a molecular level, action involves an association of the positively charged quaternary nitrogen with the head groups of acidic-phospholipids within the membrane (Fig. 2b). The hydrophobic tail then interdigitates into the hydrophobic membrane core (Fig. 2b,c). Thus, at low concentration (approximately minimum growth inhibitory concentrations, MIC) such interaction increases the surface pressure in the exposed leaflet of the membrane to decrease membrane fluidity and phase transition temperature. The membrane undergoes a transition from fluid to liquid crystalline state and loses many of its osmoregulatory and physiological functions (Fig. 2d). The membrane core decreases in hydrophobicity and phospholipids tend towards a stable hexagonal arrangement. At use-concentrations, solutions of QACs form mixed micellar aggregates that solubilize hydrophobic membrane components (i.e. lipid A, phospholipids etc. see Fig. 2e,f).

Figure 2.

Cartoon showing the mechanism of action for quaternary ammonium biocides. The segments (a–f) show progressive adsorption of the quaternary headgroup to acidic phospholipids in the membrane with increasing QAC exposure/concentration. This leads to decreased fluidity of the bilayers and the creation of hydrophilic voids in the membrane. Protein function is perturbed with an eventual lysis of the cell, and solubilization of phospholipids and proteins into QAC/phospholipid micelles. Inset micrograph shows vesicle formation from outer membrane caused by QAC treatment

The QACs have been actively deployed since the 1930s with no apparent reduction in their effectiveness. Nevertheless there are numerous reports of apparent resistance towards QAC. Such resistance invariably refers to changes in the MIC and do not affect the activity at use-levels (Gilbert and McBain 2003). The latter are often 100–1000× higher than the MIC. Where changes in QAC MIC have been demonstrated then these have either been relatively minor (2–3 fold) and associated with changes in the acidic phospholipid content of the membrane (Wright and Gilbert 1987a) or they have been associated with the acquisition, or hyperexpression of multi-drug efflux pumps (i.e. qac genes) (Heir et al. 1999). Such efflux pumps can actively remove QAC from the membrane core and thereby reduce effectiveness at sub-MIC. Acquisition or hyperexpression of multidrug efflux pumps has however been associated with changes in MIC of therapeutically important third party antibiotics which co-incidentally serve as substrates to those pumps (Poole 2003). Some species of bacteria, notably Pseudomonas aeruginosa, are relatively insensitive to QAC biocides. This is thought to relate to a failure of the compounds to penetrate the outer-membrane and to access the cytoplasmic membrane. Such insensitivity can often be overcome by formulating in solutions of EDTA and EGTA. These sequester divalent cations from the outer and cytoplasmic membrane and thereby aid interaction with QAC.

Order of magnitude increases are noted in the antimicrobial activities of n-alkyl-QACs when the n-alkyl chain length is increased beyond 10. This is related to a concentration independent dimerization of the molecules in solution. Above these critical chain lengths attraction between the adjacent hydrophobic chains exceeds the electrostatic repulsion of their charged nitrogen head groups. QAC dimers are formed that bear bi-polar positive charges in conjunction with interstitial hydrophobic regions. Such dimers both interact more strongly with the cytoplasmic membrane, than the monomeric form of QAC, and are able to more easily solubilize within it (Daoud et al. 1983; Gilbert and Al-Taae 1985). Bisbiguanides, such as chlorhexidine, provide similar bi-polar configurations of cationic and hydrophobic domains within a single molecule and are potent antibacterial agents.

4.2 Biguanides

Biguanides were first synthesized in the early part of the 20th century and have since provided a variety of drugs with a broad range of pharmacological activities (anti-malarial, blood-sugar lowering, antiseptic, anti-protozoal). Marked antibacterial activity was noted for mixtures of PHMB salts produced by the reaction of hexamethylene bis-dicyanodiamide and hexamethylene diamine (Rose and Swain 1956). PHMB (Fig. 3) produced by this route contains polydisperse oligomers with molecular weights of between 500 and 6000, in which n varies from 2 to 20. The mixtures have a weight average of n of 5·5 with the tetramer being the dominant species. Due to the method of synthesis each member of the series might have either amine or cyanoguanidine groups at either end position. Attempts to rationalize the PHMB mixtures were unsuccessful at that time, and precluded their use in pharmaceutical products. Further synthesis led to the development of the closely related bisbiguanides. In this series of molecules, optimal antibacterial activity was exhibited by the bisbiguanide 1,6-bis(4′-chlorophenylbiguanide) hexane (Davies et al. 1954) (Fig. 4a). This molecule became marketed as chlorhexidine. Alexidine (Fig. 4b), a related molecule with ethylhexyl end-groups replacing the 4-chlorophenol endgroups was developed for its activity against plaque-forming organisms (Spolsky and Forsythe 1977).

Figure 3.

Generalized structure for polyhexamethylene biguanide (PHMB) chloride counter-ion not shown. N average chain length is 5·5 with the tetramer dominating in the mixture. The end groups are randomly dispersed

Figure 4.

Chemical structures of the bisbiguanides chlorhexidine (a) and alexidine (b). Cationic phospholipid binding sites are indicated by blue shading. Hydrophobic hexamethylene group indicated by purple shading

4.2.1 The bisbiguanides chlorhexidine and alexidine.  Chlorhexidine, is active towards a wide range of Gram-positive and Gram-negative bacteria and is compatible with a variety of commonly used antibiotics. Whilst the molecule had little systemic activity in mice, it was found to be highly effective against wounds infected with haemolytic streptococci (Rose and Swain 1956).

Chlorhexidine has since been widely deployed in surgical handwashes, as an antiseptic and in various topical treatments for wound sepsis. Chlorhexidine has also been marketed extensively within various oral hygiene products as an anti-plaque agent, and within topical slow release vehicles for the treatment of periodontal disease (Hope and Wilson 2004). There have been few if any reports of chlorhexidine resistance at use concentrations, in spite of its widespread use for almost 50 years in clinical and domestic settings, but small changes (c. fivefold) in MIC have been noted (Thomas and Stickler 1979; Kropinski et al. 1982). The latter is thought to relate to changes in envelope composition particularly with regard to anionic targets and cation binding rather than to target modification and or efflux (Kropinski et al. 1982; Wright and Gilbert 1987b). Notable, in the spectrum of activity of bisbiguanides is however their ineffectiveness against some Gram-negative bacteria particularly Pseudomonadaceae and Providentia spp. (Thomas and Stickler 1979). As with the QAC biocides such insensitivity can often be overcome by formulating together with a chelating agent such as EDTA.

Bisbiguanides antiseptics have a very similar mechanism of action to the QAC biocides in that the biguanide groupings associate strongly with exposed anionic sites on the cell membrane and cell wall, particularly acidic phospholipids and proteins (Chawner and Gilbert 1989b,c). Binding to such sites is stronger than that of the QAC's and can causes displacement of wall and membrane associated divalent cations (Mg2+, Ca2+) (Davies 1973; Jensen 1975). A major difference between the bisbiguanides and QAC biocides is that the hydrophobic regions of the QAC biocides become solubilized within the hydrophobic core of the cell membrane whilst those of chlorhexidine do not. Being six carbons long, rather than 12–16 carbons, the hydrophobic region of chlorhexidine is somewhat inflexible and incapable of folding sufficiently to interdigitate into the bilayer. Chlorhexidine therefore bridges between pairs of adjacent phospholipid headgroups each being bound to a biguanide moiety and displaces the associated divalent cations (Davies 1973; Fig. 5).

Figure 5.

Diagramatic representation of the interaction of chlorhexidine with the bacterial cytomplasmic membrane. Diagram shows progressive decreases in fluidity of the outer leaflet with increasing exposure to the bisbiguanide

Interestingly the distance between phospholipid headgroups in a closely packed monolayer is roughly equivalent to the length of a hexamethylene grouping. A bisbiguanide would therefore be capable of binding to two adjacent phospholipid headgroups. Such binding is critical for the bisbiguanides as activity is reduced significantly if the polymethylene bridge is made longer or shorter than six carbons (Davies et al. 1954). An interaction with the cell membrane, such as this, will reduce membrane fluidity at low concentrations and affect the osmoregulatory and metabolic capability of the cell membrane and its contained enzymes (Hugo and Longworth 1966). These effects have been variously reported as cellular leakage of potassium ions and protons (Hugo and Longworth 1964, 1965, 1966; Rye and Wiseman 1968; Elferink and Booij 1974), and inhibition of respiration and solute transport. At higher, in-use, concentrations, the interactions are more severe and cause the membrane to adopt a liquid crystalline state, lose its structural integrity and allow catastrophic leakage of cellular materials (Longworth 1971; Chawner and Gilbert 1989a,b, 1989c). Whilst the action of multi-drug efflux pumps is able to moderate the action of QAC's at low concentrations they have no effect upon the action of bisbiguanides. This is presumably because the bisbiguanides do not become solubilized within the membrane core.

4.2.2 Polyhexamethylene biguanides.  As antibacterials, PHMB (Fig. 3) was recognized as possessing superior antimicrobial effect to other cationic biocides, but it could only be poorly defined chemically. Early attempts to rationalize the PHMB mixtures were unsuccessful and precluded their use in pharmaceutical products. Nevertheless, PHMB was marketed as a broad-spectrum antimicrobial agent in a number of diverse applications. These included their use as swimming pool sanitizers (BaquacilTM; Arch Chemicals) and as preservatives of plasticized PVC (VanquishTM; Arch Chemicals), as well as well as general-purpose environmental biocides and antiseptics. The antimicrobial activity of PHMB is superior to that of the bisbiguanide molecules subsequently derived from it. Analysis of the antimicrobial activity of PHMB reveals an enigma because, when mixtures were purified with respect to polymer chain length, it was observed that antimicrobial activity increased dramatically, on a mass basis (i.e. μg ml−1), with increasing polymer chain length (Broxton et al. 1983). Thus, the amine-ended dimers, corresponded crudely to the bisbiguanides, but were only poorly active, whilst high molecular weight materials with n > 10 were highly effective. Clearly whilst there were broad similarities between the actions of chlorhexidine and PHMB (Davies et al. 1968; Davies and Field 1969) the latter had additional properties that rendered it a superior biocide. This included a general lack of toxicology rendering the molecule suitable for use in diverse applications such as swimming pools, beer glass sanitizer together with a lack of colour, taste and surfactancy.

A schematic representation of the interactions of PHMB with the cell membrane is given in Fig. 6. As with the bisbiguanides, PHMB was shown to bind rapidly to the envelope of both Gram-positive and Gram-negative bacteria and in doing so displaces the otherwise stabilizing presence of Ca2+ (Broxton et al. 1984a). This binding is to the cytoplasmic membrane itself, and also to lipopolysaccharide and peptidoglycan components of the cells wall. The hexamethylene bridging groups of the polymer, as with the bisbiguanides, are hydrophobic yet sufficiently inflexible that they cannot interdigitate into the hydrophobic core of the cell membrane. Once again, therefore, a bridging of adjacent acidic phospholipids is brought about by interaction of the antimicrobial with the cell membrane (Broxton et al. 1984c; Ikeda et al. 1985a,b). One additional feature of this interaction is that it will tend to become concentrated around any points of maximum charge density within the membrane (Ikeda et al. 1984a,b). It has been shown that integral proteins constitute such sites. Thus, the initial interactions of PHMB and the membrane will be concentrated around such proteins, leading to a loss of their function by inflicting changes in their boundary phospholipid environment. This manifests itself, as with the bisbiguanides, as a loss of transport, biosynthetic and catabolic capability. The unique polymeric nature of PHMB means that, unlike for the bisbiguanide, such bridging is not restricted to pairs of adjacent phospholipids. Rather adsorption to the cell membrane will lead to a sequestration of common acidic phospholipids into domains comprised of single rather than mixtures of the phospholipids (Broxton et al. 1984a,b, 1984c; Ikeda et al. 1985a,b). Thus, in the presence of PHMB the homogeneous distribution of phospholipids normally associated with biological membranes is transformed into a mosaic of individual phospholipid domains. Each of these will have different phase transition properties causing the membrane to fragment into fluid and liquid crystalline regions. As with other cationic biocides this is manifested as a generalized cellular leakage, first of small cationic materials such as potassium ions and later by losses of intracellular pool materials (Broxton et al. 1983). A secondary consequence of domain formation is that the separated phospholipid types will assume their energetically favoured position of a hexagonal phase leading to a total loss of the membrane permeability barrier (Ikeda et al. 1985b). The ability of PHMB to create single lipid domains within heterogeneous lipid-bilayers is clearly a function of polymer chain length with longer polymers being able to form the larger domains and hence the greater perturbation of membrane function (Ikeda 1991).

Figure 6.

Diagrammatic representation of the mechanism of action of PHMB directed against a bacterial cytoplasmic membrane. The diagram illustrates a progressive interaction of PHMB with the acidic membrane components leading to a loss of fluidity and eventual phase separation of the individual lipids. Individual domains then undergo a transition to the more stable hexagonal arrangement, leading to membrane dissolution

The PHMB therefore not only embodies the attributes of the bisbiguanides in terms of antimicrobial action but possesses additional molecular activity. As with chlorhexidine there is no evidence that PHMB susceptibility is affected by the induction or hyperexpression of multi-drug efflux pumps, neither have there been any reports of acquired resistance towards these agents. Rather, as with all membrane active antimicrobials, small changes in MIC have been reported that correlate with alterations in envelope lipid composition and cation binding (Broxton et al. 1984c; Das et al. 1998).

5. Conclusions

The PHMB was the forerunner of the highly successful antiseptic agent chlorhexidine, and falls into a general category of antibacterial agents that are cationic, displace divalent cations from the wall and membrane of bacteria and bring about a disruption of the lipid bilayer. The biguanides and polymeric biguanides differ from other cationic biocides in that they interact only superficially with the lipid bilayer altering fluidity through cation displacement and head-group bridging. QACs on the contrary interact fully with the membrane and are therefore susceptible to resistance mechanisms mediated through multidrug efflux pumps. The activity of biguanides and bisbiguanides is unaffected by such hyperexpression of efflux. Their deployment in the clinic, or indeed within consumer products, cannot therefore have implications towards the selection of multi-resistant organisms through this mechanism (Gilbert and McBain 2003). The multiplicity of critical lethal targets affected by the interaction of biguanides with the membrane dictate against singular mutations leading to changes in susceptibility (Gilbert and McBain 2002, 2003). This is borne out by a lack of evidence to suggest that the use of either compound over 40 years has affected their activity in the field.

The toxicity profile with regard to skin irritancy and hypersensitivity of both the biguanides and the polymeric biguanides is excellent at typical in-use levels. Whilst the use of PHMB has until recently been restricted to use as a nonspecific disinfectant (Swimming pool sanitizer, beer glass sanitizer, antimicrobial fabric conditioner), such restriction relates to difficulties in standardization of PHMB formulations rather than to toxicity issues. Second generation PHMB formulations such as Cosmocil CQTM give a much better definition of polymer dispersity than was previously possible with PHMB. Such PHMB formulations are now widely deployed in clinical applications such as the treatment of Acanthamoeba keratinitis (Larkin et al. 1992) and have been included within certain contact lens cleansing formulations. PHMB has also been used as an antiseptic for various applications in medicine and wound care (Larkin et al. 1992; Wagner 1995; Willenegger et al. 1995; Kramer and Behrens-Baumann 1997).

With respect to the deployment of PHMB as part of a wound care system there is little or no evidence to suggest that this would lead to the emergence of PHMB resistant nor antibiotic resistant strains of nosocomial pathogen. Use of the agent within a barrier wound dressing such as Kerlix AMD would impair the growth and penetration through the dressing of adventitious pathogens both from the environment to the dressed wound and from the wound to health-care workers and other human contacts. Such action can only contribute to breaking the cycle of nosocomial infection and will inevitably reduce the usage of antibiotics currently used in the care of such infections.

Ancillary