The present study investigated the antibacterial effects of NO in multidrug-resistant ESBL-producing isolates with special focus on inhibition of the NO-consuming enzyme flavohemoglobin. Measuring bacterial sensitivity to gaseous NO is difficult because NO per se is stable for only minutes under physiological conditions . Instead, NO-releasing compounds e.g., DETA/NO can be added to bacterial cultures to evaluate the growth response. In E. coli, NO acts in a bacteriostatic fashion and its targets include respiratory enzymes like cytochromes bo and bd and biosynthesis pathways of branched-chain amino acid . In our study, DETA/NO induced a temporary growth inhibition in ESBL-producing UPEC isolates but after 8 hours of DETA/NO exposure a resumed growth was found. A second dose of DETA/NO, administered after 4 hours, did not prolong the growth inhibition (data not shown), suggesting that stress-response factors and NO-defence mechanisms may have been activated by the first dose. All isolates were resistant to cefotaxime but nitrofurantoin showed a time-dependent bactericidal effect. Nitrofurantoin is an antibiotic used for treatment of uncomplicated UTIs, and is so far effective against many isolates of ESBL-producing E. coli.
Upon diffusing into the bacteria, NO may react with Fe-S clusters, undergo autoxidation or be consumed directly through enzymatic detoxification . Under aerobic conditions the vast majority of intracellular NO in E. coli is consumed through flavohemoglobin detoxification [4, 9, 10]. The flavohemoglobin enzyme is not constitutively expressed and needs to be induced by gene transcription [9, 10]. We have previously shown that uropathogenic E. coli increase gene and protein expression of flavohemoglobin after exposure to DETA/NO [6, 12]. Thus, induction of the flavohemoglobin enzyme and a fast consumption of NO to submicromolar intracellular NO concentrations may explain the temporary growth inhibition with a subsequent growth recovery in our experiments. Indeed, a mutant strain lacking the flavohemoglobin enzyme (J96Δhmp) showed prolonged inhibition of growth, but after 24 hours this mutant also showed resumed growth. It has previously been verified that the hmp-deficient mutant used in the present study does not express the flavohemoglobin gene or protein when exposed to DETA/NO . In the absence of functional flavohemoglobin NO is predicted to be metabolized mainly through autoxidation, enzymatic reduction by NorV and NrfA and by Fe-S nitrosylation . Inhibition of flavohemoglobin by gene deletion was performed in a non ESBL-producing UPEC strain (J96). Gene deletion in ESBL-producing isolates is hampered by the obvious difficulties to find selection antibiotics in these multidrug-resistant isolates. However, we have confirmed a marked increase in hmp expression in an ESBL-producing isolate by real time RT-PCR when exposed to DETA/NO (data not shown), confirming that hmp is induced by DETA/NO also in ESBL-producing isolates.
Miconazole is known to interfere with synthesis of fungal and bacterial lipid membranes as it restrains the ergosterol synthesis , but recent studies also suggest that azole antibiotics target the flavohemoglobin enzyme [14, 15]. Miconazole in combination with DETA/NO prolonged the DETA/NO-induced growth inhibition in ESBL-producing UPEC isolates. Notably, the pattern of DETA/NO-evoked growth inhibition achieved by addition of miconazole was similar to the pattern noted after hmp-deletion. Furthermore, the fact that miconazole did not increase the DETA/NO-induced growth inhibition in an hmp-mutant strain, support that inhibition of flavohemoglobin  contributes to the antibacterial effect of miconazole in our experiments. Thus, the prolonged growth inhibition evoked by DETA/NO and miconazole in combination may be a result of interactions of miconazole with flavohemoglobin, causing both inhibition of NO dioxygenase activity and oxidative stress following high levels of cytotoxic superoxide production [14, 15]. In agreement with our results, intracellular survival studies in activated NO-producing macrophages demonstrated decreased survival of miconazole-treated S. aureus compared to untreated bacteria . DETA/NO and miconazole have a synergistic antifungal effect in Candida species , and the present study demonstrates that these two compounds also caused an enhanced antibacterial effect against multidrug-resistant ESBL-producing E. coli isolates.
It is noteworthy that inhibition of flavohemoglobin activity by miconazole is more pronounced in purified enzyme than in intact E. coli, in line with the poor membrane permeability of E. coli to imidazole antibiotics . Polymyxin B antibiotics may be used to sensitize the outer membrane of gram-negative bacteria to hydrophobic antibiotics . We used polymyxin B nonapeptide (PMBN), a compound that increases the cell permeability in E. coli without affecting the bacterial viability [23, 30], to avoid that polymyxin B mask the antibacterial effects of NO. PMBN per se had no antibacterial effect, while miconazole at the concentration used showed a minor inhibitory effect on UPEC growth. An in vitro synergism of miconazole and polymyxin B has been reported in E. coli, related predominantly to the ability of polymyxin B to increase the penetration of miconazole to the intracellular space . In our experiments, miconazole and PMBN in combination caused a significant inhibition of growth compared to untreated controls. Interestingly, when DETA/NO was added to miconazole and PMBN a prolonged bacteriostatic response that persisted for 24 hours was observed. It is not likely that the underlying mechanism is a more effective inhibition of NO-detoxification by flavohemoglobin since the hmp-mutant showed recovered growth after 24 hours. However, a better access of miconazole to intracellular targets like flavohemoglobin, when combined with PMBN, may cause enhanced antibacterial activity through magnification of intracellular oxidative stress responses . Furthermore, increased formation of toxic peroxynitrite (ONOO−), a potent oxidant formed from NO and superoxide radical , could conceivably contribute to the spectrum of potential antibacterial mechanisms of the combination treatment. The bacterial cell membrane is another possible target and miconazole is known to affect the integrity of the lipid membrane [27, 32]. Morphological analysis have revealed cell membrane deteriorations with widespread structural deformations as a consequence of NO exposure in E. coli, and a synergistic effect of all three substances on the cell membrane is possible. However, the exact mechanisms for the prolonged bacteriostasis evoked by DETA/NO in combination with miconazole and PMBN need to be further studied. Interestingly, impaired adhesion to host renal epithelial cells and broken fimbriae has been reported after NO exposure in E. coli[33, 34]. This suggests that the antibacterial effects of NO may be widespread and that NO not only has growth inhibitory effects but also may affect bacterial virulence properties and host activating mechanisms.
The ESBL-producing UPEC isolates used in the present study were obtained from patients with catheter-associated UTI. Indwelling medical devices, including urinary catheters and biofilm formation increase the risk of bacterial infection and result in considerable antimicrobial use. When these infections are caused by multidrug-resistant bacteria commonly used empirical antimicrobial therapy are not effective . Administration of NO directly into the bladder through a silicone balloon catheter represents a local delivery system for NO-based therapy and has been suggested as one strategy to prevent catheter-associated infections . Urinary catheters impregnated with NO have been shown to inhibit both biofilm formation and planktonic E. coli growth . A limitation of the present study is that the number of clinical isolates used is small. However, three out of four isolates responded identical and with a prolonged bacteriostasis to the triple combination. These clinical isolates represented different CTX-M enzymes but it is, however, not possible to draw any conclusions on possible correlations between susceptibility to treatment and the CTX-M enzyme based on this material. Importantly, two of the isolates belonged to the CTX-M-15 ESBL type and the sequence type 131 (data not shown) which represent the dominating worldwide emerging CTX-M type and clone in community-acquired UTIs [37, 38].
Colistin, a polymyxin antibiotic, has regained interest for its activity against multidrug-resistant gram negative pathogens, including those harbouring carbapenamases . Thus, the emergency of multi-resistant pathogens encourages rediscovery of older antibiotics with activity against these resistant bacteria and in new combinations. Our data suggest that two existing antibiotics, an azole antifungal and a polymyxin B compound, are able to enhance the antimicrobial effects of exogenously administered NO. In contrast to antibiotics that are specific to one or few bacterial species, the antibacterial effects of NO are broad and both gram-positive and gram-negative pathogens, including antibiotic-resistant isolates show sensitivity [40, 41]. Development of resistance mechanisms to NO may be limited by its multiple biochemical targets, but metabolic adaptions to nitrosative stress including induction of flavohemoglobin  and lactate dehydrogenase in S. aureus decrease the antimicrobial action of NO. Therefore, as highlighted in the present study, a combination treatment where exogenous NO is combined with an inhibitor of NO-protective mechanisms appears to be an attractive approach to improve the antimicrobial effects of NO.