Persisting Problem

The phenomenon of persistence was recognized in the mid-1940s in experiments in which cultures of penicillin-sensitive bacteria survived treatment with penicillin. The subpopulation of surviving bacteria has been referred to as persisters. Persister cells have been proposed as an additional innate mechanism for biofilm antibiotic resistance (Lewis 2005). In the persister theory, a small subpopulation of bacteria, whether in biofilms or planktonic culture, differentiates into dormant, spore-like cells that survive after extreme antibiotic treatment (Fig. 1). Differentiation into this dormant state has been hypothesized to be the result of phenotypic variation rather than a stable genetic change (Keren et al. 2004a).

2.3.1 Genetic Factors Influencing Persister Formation

Interestingly, the results of recent studies suggest that, while persisters may be phenotypic variants, specific genetic elements are required to form the persister state. Studies by Spoering, Vulic, and Lewis implicated altered genetic activation of the glycerol-3-phosphate regulated genes glpD, glpABC, andplsB in E. coli as a mechanism of persister development (Spoering et al. 2006). The glpD gene was initially found to be important for this developmental pathway because plasmid-driven expression of the gene could increase the formation of ampicillin-resistant persisters in the exponential phase by approximately tenfold. Mutating the glpD gene or other genes involved in glycerol-3-phosphate metabolism, including glpABC or plsB, decreased tolerance to ampicillin by greater than 100-fold, indicating a role for glycerol-3-phosphate metabolism in persister formation. However, it was not reported whether these mutations altered the growth rate of the cell or the minimum inhibitory concentration for ampicillin. Further, given glycerol-3-phosphate's central metabolic role, these mutations did not provide any direct mechanistic insight into how persisters might be generated.

One mechanism proposed to explain the ability of persisters to resist the action of antibiotics is similar to a mechanism long hypothesized for biofilm resistance, namely a slowed growth rate. Indeed, persisters exhibit slow or no growth, as observed by microscopy of E. coli in a microfluidic device (Balaban et al. 2004). This decreased growth rate may inhibit antimicrobial action, as discussed above in Sect. 2.2. However, persisters can survive even after treatment with ofloxacin, which exerts bactericidal activity against nongrowing microorganisms (Kaldalu et al. 2004; Spoering and Lewis 2001), suggesting that limited growth rate alone cannot account for the increased antibiotic resistance of persisters. Alternatively, global transcriptional profiling by microarray analysis of persister cells revealed activation of numerous stress response pathways (Kaldalu et al. 2004; Keren et al. 2004b), potentially implicating these cells as hardy, stress-resistant microorganisms.

Another major factor influencing formation of persisters appears to be chromosomal toxin/antitoxin (TA) systems (Lewis 2005), which have previously been associated with programmed cell death in bacteria. Several TA modules were upregulated by microarray analysis of persisters in E. coli, including dinJ/yafQ, relBE, and mazEF (Keren et al. 2004b; Shah et al. 2006). Overexpression of the relE toxin gene, in particular, led to tolerance of high levels of such disparate antibiotics as ofloxacin, cefotaxime, and tobramycin (Keren et al. 2004b). The hipBA TA locus has also been found to be important for formation and maintenance of persisters, and mutation of the hipA toxin gene can enrich for persisters within in an E. coli population (Harrison et al. 2005; Keren et al. 2004a; Keren et al. 2004b; Moyed and Bertrand 1983). It has been suggested that these TA modules actually induce stasis of the bacterial cell by inhibiting the activity of a particular cellular machine, such as the ribosome (Keren et al. 2004b). It was proposed that this inert state then prevents the deleterious functions induced by antibiotics. For instance, an aminoglycoside cannot induce the formation of misfolded proteins if its target ribosome has been rendered static. In this sense, persister bacteria are considered antibiotic-tolerant rather than antibiotic-resistant (Keren et al. 2004b; Lewis 2005). Evidence for this induced stasis comes from studies demonstrating that, while overexpression of the relE or chpAK toxin genes in E. coli rapidly reduced colony-forming units, subsequent transcription of the relB or chpAI antitoxins, respectively, led to a restoration of colony formation on agar plates (Pedersen et al. 2002). In other words, the toxin-expressing bacteria were nongrowing, yet non-dead, and addition of antitoxin resuscitated these cells. Thus, random fluctuations of toxin and antitoxin levels may modulate the formation and awakening of dormant persisters.

2.3.2 Persister Controversies

Intriguingly, persister research has led to several claims about biofilm antibiotic resistance in opposition to generally accepted biofilm tenets. Specifically disputed is the widely held, and well-supported, hypothesis that biofilms are more resistant to antimicrobial killing than planktonic bacteria. This argument has led some researchers to solely examine planktonic cultures for phenotypic and genotypic analysis of persisters. For instance, Spoering and Lewis concluded from their studies that stationary-phase P. aeruginosa was equally or more resistant than biofilm cultures to several antibiotics (Spoering and Lewis 2001). This effect was quantified as greater bacterial CFU after 6 h of antibiotic challenge and was hypothesized to be the result of equal or greater persister formation in the planktonic stationary-phase bacteria compared to the biofilm population. Similarly, Harrison et al. discovered that planktonic and biofilm populations of E. coli required similar levels of amikacin and ceftrioxone to effect complete eradication of the population in 2 h. However, in this latter study, E. coli biofilms were more resistant to tobramycin than planktonic phase cells at 2 h. Further, increasing the incubation time to 24 h revealed a much greater antibiotic resistance of biofilms to all three antibiotics compared to planktonic cells. In other words, planktonic bacteria were more sensitive to lower antibiotic concentrations when treated for longer periods of time. This result leads one to wonder whether increased antibiotic incubation periods could have produced a similar effect in the work by Spoering and Lewis and similar studies (Brooun et al. 2000; Spoering and Lewis 2001).

An additional concern in these studies is the variance in bacterial numbers between planktonic and biofilm populations at the start of antibiotic treatment. Thus, while Spoering and Lewis found a greater number of surviving stationary planktonic-phase bacteria compared to biofilm bacteria after antibiotic treatment, they also started with a significantly greater number of stationary planktonic phase bacteria then biofilm bacteria. In effect, in the stationary phase cultures, the units of antibiotic per bacterial cell were markedly decreased relative to biofilm bacteria, and this difference might have led to an apparent increase in antibiotic resistance. In a later study of E. coli resistance to metal oxyanions, Harrison et al. equalized planktonic and biofilm bacterial numbers before antibiotic challenge and found that this action did not significantly alter the MIC. However, in these conditions for the planktonic bacteria, they reported that "the proportion of surviving cells was smaller than the fraction of survivors recovered from biofilms" (Harrison et al. 2005). As with the increased incubation time mentioned above, it would be intriguing to determine the effect of starting with similar bacterial numbers using the system as described by Spoering and Lewis.

Based on the results of these studies, it may be misleading to consider biofilm antibiotic resistance as a stationary-phase persister phenomenon. Alternatively, perhaps planktonic persisters have differentiated into a more biofilm-like pheno-type, although there is no data to support this theory at this time. Recent microarray studies of E. coli suggested that the persister transcriptional profile represents a unique physiological state, distinct from exponential phase or stationary-phase bacteria (Shah et al. 2006). While no comparison was made to biofilms, it is intriguing to speculate that the persister phenotype is similar to the biofilm state. Indeed, the most highly expressed gene in persisters compared to nonpersisters in this microarray analysis was ygiU, which is also induced in biofilms and acts as a global regulator influencing biofilm formation (Shah et al. 2006). Further, mathematical modeling has predicted a steady accumulation of persisters as a biofilm matures and ages (Roberts and Stewart 2005). Thus, despite inconsistencies in persister literature, persister formation remains an intriguing concept as a supporting mechanism of biofilm antibiotic resistance.

In summary, innate formation of persisters might represent a common mechanism used by a wide range of bacteria during biofilm formation. Creation of this tenacious population within the biofilm may drastically inhibit the complete eradication of the biofilm during even prolonged, high-level antibiotic treatment (Fig. 1). However, at this stage, it is unclear what relationship, if any, can be drawn between planktonic persisters and biofilm resistance, and furthermore, the mechanisms(s) by which persisters form and/or confer increased antibiotic tolerance.

0 0

Post a comment