«(Über die Bedeutung der bakteriellen Genomplastizität für die Adaptation und Evolution asymptomatischer Bakteriurie (ABU) Escherichia coli ...»
Haraoka et al., 1999). In response to pathogens, neutrophils adhere to the endothelium and transmigrate into the infected tissue, where their activation induces the production of nitric oxide (NO) and release of granular enzymes that eliminate the intruding pathogen. Nitric oxide synthesise (iNOS) from PMNs is 43-fold up-regulated in patients with urinary tract infections when compared to healthy individuals (Wheeler et al., 1997), what results in 30times higher NO concentrations in urine (Lundberg et al., 1996). Klemm and co-workers corroborated these results, finding genes involved in NO protection and metabolism to be induced in ABU strain 83972 upon intravesical growth in vivo (Roos and Klemm, 2006).
Moreover, urine contains significant amounts of nitrate (Tsikas et al., 1994) and anaerobic NO3 metabolism results in the generation of additional nitric oxide. It is well documented, that nitrosating agents produce mutagenic lesions (Weiss, 2006). Therefore, regarding the frequent occurrence of mutations in the genomes of certain ABU isolates which are not mutators by themselves, it is tempting to speculate that prolonged growth in urine and exposure to immune response promotes the mutation rate thus being a driving force for the development of the ABU lifestyle and evolution within the urinary tract.
The current approach, where the same bacterium was subjected to interaction with multiple human hosts, further corroborates this hypothesis. Numerous changes in the genome structure of consecutive re-isolates derived from the human colonisation study imply the importance of interactions with the host during bacterial evolution. Furthermore, genome structure of reisolates from the continuous flow culture, where strain 83972 was propagated without host contact, was not affected. However, two of four bacterial populations were propagated with addition of nitric oxide. Altogether, the results indicate that not only prolonged growth in urine and the presence of nitric oxide but also direct contact with the host tissues and selective pressure promote bacterial variability. Regarding the presence of nitric oxide in the continuous culture, it would need further investigation since it is difficult to stably provide the required concentration of this extremely instable compound. Moreover, the mutagenic effect of NO is rather due to the occurrence of point mutations than to bigger DNA rearrangements detectable by PFGE and final conclusions can be drawn after the analysis of the genome sequences of bacteria grown with and without contact with the human host.
Taken together, so far not well-characterized host factors play an important role in evolution of bacterial commensalism within the urinary tract. Point mutations and other DNA rearrangements, unless they are deleterious, contribute to adaptation and survival of asymptomatic E. coli in the urinary bladder environment. Without doubt, positive selection of clones which do not activate a strong immune response favours bacteria which are less aggressive to the host. Moreover, detection of the genomic regions subjected to host-driven mutagenesis might help to discover new potential drug targets.
6.5.2. Flagella expression / motility
Although E. coli strain 83972 has been characterized in many details regarding the expression of different fimbriae (Klemm et al., 2006; Roos et al., 2006a; Zdziarski et al., 2008), not much is known about flagella expression by this strain. Both animal and human in vivo transcription studies indicate that motility is down-regulated during E. coli bladder colonisation (Roos and Klemm, 2006; Snyder et al., 2004). However, flagellum-mediated motility/chemotaxis was proven to be not required but beneficial during colonization of the urinary tract by contribution to the fitness of the bacterium (Lane et al., 2005). The parent strain 83972 used for both the human colonization study and the in vitro “2000 generation experiment” is very little motile. Whereas it was found that several in vivo re-isolates were characterized by different degrees of motility suggesting that its genetic determinants is intact and is subjected to active regulation. Moreover, bacteria grown in vitro in urine exhibited flagella up-regulation in a few cases, while none of the LB-grown re-isolates exhibited this phenotype.
The flagella was shown to stimulate IL 8 production (Zhou et al., 2003). Therefore, in accordance with the model of ABU lifestyle where immune response activation is avoided, in most re-isolates some weak motility could be observed. Interestingly, close to the time points when most in vivo re-isolates with up-regulated flagella expression were found, host response was increased. Since in many cases only a few isolates from a certain time point were analysed, it is difficult to estimate how much of the bacterial population in the bladder is represented by the same phenotype. Because adhesion is not necessary for persistence of bacteria in the urinary tract (Andersson et al., 1991), other fitness factors might be more important. Heterogeneous cultures with regard to motility, like in the microfermenters, may
be advantageous for the ABU strain as urodynamic defects of colonized patients support colonisation with non-virulent strain 83972 (Wullt et al., 1998). A mixed bacterial population with a low percentage of motile individuals (minimal host response activation) could play an important role in bladder colonisation, especially at voiding time points. Because strain 83972 does not adhere to the bladder tissue, the time shortly after voiding might resemble initial steps of urinary tract colonisation, for which flagella was reported to be beneficial (Lane et al., 2005). Therefore, future studies are needed to analyze the function and importance of this still not deactivated feature of strain 83972 for the establishment of asymptomatic bacteriuria.
6.5.3. Biofilm formation
Another suggested mechanism how bacteria colonise the urinary tract is biofilm formation.
Whereas in nature bacteria often exist within biofilms (Costerton et al., 1999), formation of this structures within the urinary tract is still questionable. Although biofilm formation on abiotic surfaces is well reported and of particular interest in medical field, not much is known about biofilm structures in vivo within the bladder. Intracellular bacterial communities (IBC) are considered as one of the biofilm forms and rather restricted to pathogenic bacteria expressing type 1 fimbriae and flagellum (Anderson et al., 2004). Moreover, it has been reported that in vitro biofilm formation of a significant number of wild type E. coli isolates could not be correlated to any of the pathotypes and is dependent on the used medium (Reisner et al., 2006). On the other hand, Klemm and co-workers reported that asymptomatic bacteria isolates form significantly more biofilm than other uropathogenic E. coli strains and propose it to be the favourable strategy for successful ABU lifestyle (Ferrieres et al., 2007;
Hancock et al., 2007).
In contrast to that, our study did not reveal significant differences in biofilm formation of uropathogenic and ABU isolates. Moreover, the wild type strain 83972, being a relatively good biofilm former, did not preserve this phenotype after long term persistence within the bladder. Only one re-isolate was shown to be better and two other were nearly as good biofilm formers as their parent strain. However, this situation was observed only in human urine and already in laboratory medium M63 the differences could not be observed. This also further corroborates the results of Reisner and colleagues (2006). The transcriptome analysis of in vivo re-isolates confirmed luxS (ygaG), coding for protein involved in autoinducer 2 (AIbiosynthesis, to be up-regulated. AI-2 is a quorum sensing (QS) molecule that also
negatively controls biofilm formation (Surette et al., 1999). While it has been proposed that asymptomatic bacteriuria isolates express more biofilm than symptomatic UPEC strains which is required for biofilm formation to colonise the bladder (Hancock and Klemm, 2007), our data show a reduction in this phenotype upon prolonged contact with the host. In agreement with that, biofilm formation of re-isolates derived from the ‘host free’ experiment was at least as good as that of the parent strain, sometimes even better.
As already mentioned, the gene ygaG is involved in methionine metabolism, which was reported to be induced upon nitrosative stress (Flatley et al., 2005; Jarboe et al., 2008). This condition might be encountered by bacteria either during denitrification or NO-mediated host defence. Biofilm forming bacteria are in a very close contact with the host tissue, what increase the possibility of activation of the host immune system, the IL 8 recruit inflammatory cells to the site of infection and toxic nitric oxide is released (Hedges et al., 1995). As ABU isolates rather resemble commensal bacteria, an unnecessary immune system activation should be avoided. Interestingly, it was shown that NO causes dispersal of Pseudomons aeruginosa and Staphylococcus aureus biofilms (Barraud et al., 2006; Schlag et al., 2007).
Because NO concentrations in the bladder are rather high and further elevated when host defence is activated, it is expected that asymptomatic bacteria will not form much biofilm and rather live as planktonic cells. Furthermore, our results of the microfermenter experiments support this hypothesis, as they showed a significant reduction in biofilm formation after addition of nitric oxide to the medium.
6.5.4. Growth characteristics
Presumably, the fast growth and efficient utilization of resources available in urine belong to the most important factors enabling ABU isolates to inhabit the urinary tract. In healthy adults, normal urine production ranges from 1-2 litres per day and single micturition results in release of 200-400 ml of urine. Following micturition, about 1 ml of urine remains in the bladder and might function as a sufficient inoculum for repeated bladder colonization until the next voiding episode. When growth rate is high enough so that the number of proliferating bacteria exceeds the number of those which are lost due to micturition, surface-associated growth is not needed to persist within the bladder (Gordon and Riley, 1992). The early phase of colonisation is critical in successful establishment of bacteriuria, because at that time point the innate immune system might be able to clear the bacteria. Therefore, fast growth in the 153 Discussion early exponential phase in combination with low host defence activation might be a successful strategy to establish a permanent asymptomatic bacteriuria.
Indeed, the eleven tested ABU isolates exhibited good growth rates in vitro in human urine, however, varied from strain to strain. Moreover, the analysis of the re-isolates of strain 83972 from the human colonisation study implicated that, depending on the patient, nutrients availability and the quality of the host response, these strains might not take advantage of their growth potential. The results demonstrate, that multiple in vivo re-isolates exhibited slower growth rates than their parent strain and in many cases different re-isolates from the same patient exhibited similar growth rates. This further underlines the importance of the host background. Without doubt, the extent of the host response is one of the most important factors for the modulation of bacterial growth in the bladder. ABU isolate 83972 does also induce a human immune response to a certain extent (C. Svanborg, personal communication).
If bacteria grow too fast, their number might exceed a threshold that is considered to be dangerous and they will thus be subsequently cleared. As a consequence, the host defence, in case of asymptomatic bacteriuria, might function as a negative feedback and regulate commensalism in the urinary tract.
6.6. Metabolic activity of ABU isolates
Another critical and limiting factor during bacterial growth in the bladder is nutrient availability. The metabolic variability in an individuum and between different persons results in significant differences in the urine composition not only in between but also in the same host. Different reasons like genetic differences, age and lifestyle, nutrition, and exposure to specific chemicals result in a wide variety of urine compositions in individual patients (Rezzi et al., 2007; Stella et al., 2006). Changes in diet, day- and night-time, and different stages of hormonal cycle contribute to inter-individual fluctuations (Rezzi et al., 2007).
Bacterial metabolic networks have to be adjusted to grow fast and optimally utilize nutrients present in the urine. This is achieved in many cases by transcriptional regulation of gene expression. Stable alterations in the expression of metabolic pathways, even after in vitro cultivation of re-isolates in urine, indicate that also stable DNA modifications (e.g. point mutations, insertions, deletion) contribute to this process. Transcriptome analysis of the
chosen re-isolates uncovered significant differences not only between in vivo and in vitro growth conditions, but also among in vivo re-isolates. Bacteria in individual hosts might approach slightly different strategies how to supply enough energy for fast proliferation. This has been very well demonstrated by the comparison of the transcriptomes of re-isolates SR12 and CK12.
Naturally occurring sugar acids such as galacturonate from pectin and gluconate and ketogluconate from muscle tissues are present in the food we eat (Peekhaus and Conway, 1998). The bladder urothelium is covered by a thick layer of protective glycoprotein and is rich in N-acetylglucosoamine, N-acetylgalactosamine, sialic acid and lesser amounts of glucuronate and galacturonate. In line with that, the in vivo re-isolate SR12 possesses upregulated pathways of uptake and metabolism many of these sugars. It has been already reported that during growth in urine E. coli induces expression of genes involved in uptake and utilization of the sugar acids galacturonate, glucuronate and galactonate (Roos et al., 2006b). We found out, that the induction of these genes is not only de-regulated because of growth in urine but also depends on the patient and might be connected to the diet and physiological state.