The Role of Outer Membrane Proteins in UPEC Antimicrobial Resistance: A Systematic Review

Uropathogenic Escherichia coli (UPEC) are one of the most common agents of urinary tract infection. In the last decade, several UPEC strains have acquired antibiotic resistance mechanisms and some have become resistant to all classes of antibiotics. UPEC outer membrane proteins (OMPs) seem to have a decisive role not only in the processes of invasion and colonization of the bladder mucosa, but also in mechanisms of drug resistance, by which bacteria avoid killing by antimicrobial molecules. This systematic review was performed according to the PRISMA guidelines, aiming to characterize UPEC OMPs and identify their potential role in antimicrobial resistance. The search was limited to studies in English published during the last decade. Twenty-nine studies were included for revision and, among the 76 proteins identified, seven were associated with antibiotic resistance. Indeed, OmpC was associated with β-lactams resistance and OmpF with β-lactams and fluoroquinolone resistance. In turn, TolC, OmpX, YddB, TosA and murein lipoprotein (Lpp) were associated with fluoroquinolones, enrofloxacin, novobiocin, β-lactams and globomycin resistances, respectively. The clinical implications of UPEC resistance to antimicrobial agents in both veterinary and human medicine must propel the implementation of new strategies of administration of antimicrobial agents, while also promoting the development of improved antimicrobials, protective vaccines and specific inhibitors of virulence and resistance factors.


Introduction
Uropathogenic Escherichia coli (UPEC) are the most common agents of urinary tract infection (UTI), in both humans and pets [1,2]. It is estimated that UPEC are responsible for more than 80% of UTIs in humans and between 30 to 69% of UTIs in pets [1,3]. A successful combination of virulence factors confers to these strains an increased capacity of ascending through the urinary tract, colonizing, invading and disseminating in the bladder mucosa. From there on, they can continue to progress to the kidneys (pyelonephritis) as well as enter the bloodstream, causing bacteremia [4]. UPEC strains are mainly classified into virulent-phylogroup B2, possessing specific and diverse virulence factors responsible for colonization, invasion and dissemination; to a lesser extent, they are also included in phylogroup D [1,5].
The treatment of UTI relies on antibiotic therapy, such as trimethoprim-sulfamethoxazole, fluoroquinolones and cephalosporins, representing the first-line empirical antibiotics [1,6,7].

Materials and Methods
This systematic review was performed according to the relevant points of the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines [19].

Selection Strategy
In September 2022, an independent researcher (I.C.R.) searched the PubMed database without language restrictions in the past ten years (since 2012). Peer-reviewed studies describing antimicrobial susceptibility patterns in terms of proteins located in the outer membrane of UPEC isolated from human or pet samples of any age and region were included in this review. Studies published in a language other than English or Portuguese, publications comprising editorials, comments, letters to the editor, guidelines, theses, books and scientific meeting abstracts, literature reviews or case reports, studies performed specifically in ESBL and/or related to other bacteria and that did not mention uropathogenic E. coli or UTI in the title, publications without a description of outer membrane proteins, and studies published before 2012 were excluded. The author (I.C.R.) also reviewed the reference lists from the review articles reported in the PubMed and Web of Science searches to identify possible additional articles for inclusion. A combination of the following search terms was used: Uropathogenic Escherichia coli AND membrane proteins. In order to evaluate a potential inclusion in this study, titles and search results were examined.

Selection Process and Data Extraction
All search results were exported to Microsoft Office™ Excel. Results from the initial search were evaluated separately by the two review authors (I.C.R and S.C.R.) according to the inclusion criteria. First, the results were screened by reading the article titles and excluding articles that were not relevant according to the inclusion criteria. Afterwards, the study abstracts were evaluated and non-relevant articles were excluded. Finally, the fulltext articles selected by the two reviewers were collected and assessed for their relevance relative to the inclusion criteria. Any disagreements regarding the eligibility of studies were reconciled at the final step by discussion and consensus.
Upon the final consensus, the following data were extracted from each selected study and validated by the second author (S.C.R.) elaborating a systematic database: Title, authors and outer(s) membrane(s) detected; Samples origin and type; Method used for the detection of outer membrane protein; Group of outer membrane protein; Description of outer membrane protein; Function of outer membrane protein.

Quality Assessment
Risk of bias (RoB) was examined using the Newcastle-Ottawa scale (NOS) quality assessment scale for cohort studies [20]. The used scale was adapted from the NOS and the assessment was based on three criteria: sample (maximum of 3 points), comparability (maximum of 2 points) and outcome (maximum of 3 points). This was done by I.C.R. and S.C.R. and three categories were decided: low RoB (rating 7 to 8 points), moderate RoB (rating 5 to 6 points) and high RoB (rating from 0 to 4 points). No article was excluded based on this assessment.

Description of Studies
The PRISMA flow diagram summarizes the number of records screened and included ( Figure 1). The characteristics and methodological quality of the included studies are presented in Table 1. The literature search using PRISMA identified a total of 1122 studies. After removing the duplicates, 569 were screened for eligibility. After the screening of titles and abstracts, 520 studies were excluded. Full texts of the remaining 48 studies were read and 22 more studies were excluded. At the end, 29 publications were included in this systematic review.   [19]. Figure 1. PRISMA 2020 flow diagram. * Consider, if feasible to do so, reporting the number of records indenified from each database or register searched (rather than the total number across all databases/registers). Adapted from [19]. Table 1. Characteristics of included studies after full assessment. Data from the final twenty-nine studies were extracted and systematized. Author; year of publication; the species of the isolates and the type of the sample collected; and the method used for protein characterization and the OMP detected were listed. The RoB assessed through the adaptation of NOS quality scale was also included for each study.

Author
Year
All of the identified proteins have an essential function as described in Table 2. The functional diversity of the OMPs ranges from motility and adhesion (flagella, fimbriae, adhesins) to survival within the urinary tract (iron receptors, siderophores and efflux pumps) [4,5]. Their individual or joint action in the host tissue entails a high level of virulence and pathogenicity, leading to severe UTI symptoms [33,35,36,39].
According to Tavio et al. [36], the AcrA-AcrB-TolC efflux pump is the major multidrug efflux transporter in E. coli, allowing the passage of fluoroquinolones, among other antimicrobials.
Pantua et al. [34] found that the deficient production of Lpp leads to an increased OM permeability, a leakage of periplasmic components and an increased outer membrane vesicle release. The same authors also described that complete deletion of Lpp causes globomycin resistance [34].
Among the identified porins, OmpC, OmpF, OmpW and YddB were shown to be the preponderant in antimicrobial resistance strategies of UPEC strains. OmpC is involved in both the transport of small molecular weight hydrophilic materials across the outer membrane and in antimicrobial resistance (e.g., β-lactams antibiotics) [17,32]. OmpF plays an important role in antibiotic transport, such as β-lactams and fluoroquinolones [23]. In fact, a decreased ompF expression was related to an increase of resistance towards β-lactam antibiotics in E. coli [23,35], since OmpF is the main gate for enrofloxacin entrance [14]. Lastly, YddB (a putative porin protein) seemed to be the major porin involved in the passive transport of novobiocin across the outer membrane [23].

Discussion
UPEC is the most frequent uropathogen worldwide, entailing elevated healthcare costs and a high risk of treatment failure in both human and veterinary medicine due to the emergence of multidrug-resistant strains and limited therapeutic options to treat infections caused by these strains.
Just as a One Health unifying approach has the merit of reframing our understanding about the capacity of resistant bacteria to spread between different biomes [42,43], a wide overview of resistance mechanisms used by bacteria to evade antimicrobial treatment will be crucial for overcoming the problem and to find both new drug targets and new structural classes of antibiotics. Herein, among 76 OMPs described in the 29 studies included in this systematic review, seven were responsible for resistance to several antimicrobial drugs. Among them, OmpC is associated with β-lactams resistance and OmpF with β-lactams and fluoroquinolones resistance. With a prevalence of less than 75%, TolC, YddB, TosA and Lpp are also associated with resistance against fluoroquinolones, enrofloxacin, novobiocin, β-lactams and globomycin, respectively. Considering that fluoroquinolones and β-lactams are the antimicrobial classes more commonly used for the treatment of UTI in both human and animals, the spread of these OMPs between UPEC strains is of enormous clinical relevance. Indeed, both antimicrobial drugs are particularly affected by changes in the permeability of the outer membrane of UPEC strains, since β-lactams and fluoroquinolones often use water-filled diffusion channels (porins) to penetrate this external barrier [44]. Despite the predominant mechanism of resistance to β-lactams in UPEC being mainly attributed to the production of β-lactamases [45], OMPs seem to possess essential complementary "tools", as they control the entry of these molecules to the periplasmic space, allowing the production of β-lactamases in sufficient concentrations to achieve the destruction of the antibiotic molecules [46].
Porins are outer membrane β-barrel proteins that cross the membrane cell, responsible for uptake of nutrients, being essential for bacterial survival and growth [33]. On the other hand, they allow the "unwanted" entrance of several antimicrobial molecules; for example, OmpC, OmpF and OmpW are porins responsible for creating a size-selective channel, promoting the transport of small hydrophilic molecules, including β-lactams and fluoroquinolones [47]. Thus, the production of OmpW and OmpC was reduced in tetracycline and nalidixic acid resistant strains, indicating an antibiotic-specific pattern of porin expression [47,48]; and a decreased ompF expression was related to an increase of resistance towards β-lactam and fluoroquinolones antibiotics [23,35], since OmpF is the main gate for enrofloxacin entrance [14]. In addition, OmpF and OmpC production seemed to be influenced by the environment of UPEC during a UTI. Urine osmolarity appears to affect ompF and ompC expression: ompF is upregulated under a low osmolarity condition, while ompC is upregulated under high osmolarity [49]. Although it is a lesser reported porin, YddB is associated with passive transport of novobiocin across the outer membrane. Presently, novobiocin resistance is not an emerging problem; however, the mechanism by which YddB confers novobiocin resistance has not yet been characterized.
Beyond the alteration of membrane permeability, bacteria is also able to enhance its resistance by the overproduction of efflux pumps (actively expelling the antibiotic out of the cells) [50]. Efflux pumps may be the most rapid acting and effective resistance mechanism in the bacterial defense [50]. Regarding the AcrA-AcrB-TolC efflux pump, Chowdhury et al. [51] hypothesized that one of the intrinsic mechanisms of multidrugresistant UPEC strains is the overexpression of the AcrAB-TolC efflux pump, probably caused by the mutation of the acrR gene.
Among the OMPs related to antimicrobial resistance, the TosA protein, a nonfimbrial adhesin, is a virulence marker in UPEC strains, since deletion of the tosA gene affected their capacity to colonize the bladder and kidney in a murine model [52,53]. Moreover, multidrug-resistant UPEC isolates expressing the tosA gene maintained a high correlation with virulence genes presence [53]. However, additional research is required to fully understand the role of TosA during UTI.
Another OMP associated to antimicrobial resistance is Lpp, which is the most abundant lipoprotein of E. coli (more than 5 × 10 5 molecules per bacterial cell), being responsible for maintaining the stability and integrity of the outer membrane [54]. Although Pantua et al. [34] described that the complete deletion of Lpp led to globomycin resistance, other authors stated that the deletion of this lipoprotein increased the susceptibility for vancomycin, erythromycin and rifampicin antibiotics [54,55]. In fact, the reduction of Lpp stimulated vesicle formation, decreased membrane integrity and reduced cell invasiveness by turning down biofilm formation [54][55][56].
In addition to the role of the OMPs, other proteins are also involved in the antimicrobial defence of bacteria. Qnr proteins protect DNA gyrase and type IV topoisomerase IV, being responsible for quinolones resistance in UPEC isolates [35]. These proteins supplement resistance to quinolones by efflux pump activation, altered quinolone target enzymes or deficiencies in OMPs [57]. RecA protein also participates in DNA reparation and activation of the bacterial SOS system [35]. At last, Spr (peptidoglycan endopeptidase) proteins assists peptidoglycan biogenesis in UPEC and flagella expression, which are correlated to a fitness advantage during bladder colonization [4].
Given the potential zoonotic of UPEC, the origin of the isolates of each study was also assessed. Although most of the isolates included in this review were from human origin, UPEC strains are being able to circulate among human and pets populations [13]. Belanger et al. [12] corroborated the phylogenetic proximity between human isolates and those obtained from cohabitant cats and dogs. Thus, we questioned whether the OMPs of UPEC isolated from pets would be different from those isolated from human origin. Piras et al. [14] suggested that if UPEC isolated from humans are exposed to the same concentration of antibiotic that animals are exposed to, then the mechanisms involved in resistance will be similar. Moreover, it requires further characterization of OMPs in UPEC isolated from both human and pet samples, as well as the elucidation of the mechanisms involved in antimicrobial resistance.
UPEC strains have cellular membrane machinery that allows them to be very effective when infecting the urinary system. Combining the fact that E. coli possess both a great ability of adaptation to different environments and present highly conserved proteins and organelles, a rapid acquisition of the tools necessary for antimicrobial resistance is promoted. Since available antimicrobial options have become ineffective for treatment, the access to antimicrobial molecules should be rethought and empirical therapy should be avoided. Therefore, the emergence of antimicrobial resistances is an important public health problem. Without implementing more stringent measures, the emergence of "more" multi-resistant strains will occur, leading us to a point of no return.
New therapeutic alternatives need to be discovered to combat the emergence of multidrug UPEC as well as other pathogens [58,59]. Several antimicrobial molecules that interact with OMPs have been already studied, such as darobactin, Polyphor peptide 7 and MRL-494, which affect the integrity of the bacterial outer membrane by inhibition of BamA function [60]. Pilicides and curlicides compounds have the ability to inhibit the pili of UPEC [61,62]. Pre-existing molecules used for other applications may also have antibacterial potential against UPEC, such as nitazoxanide, which is used for intestinal parasitic diseases and also inhibits the function of type 1 and Pap pili [63]. Therefore, the OMPs described in this review seem to be crucial targets in the fight against UPEC and they may also be key targets for sparing other multidrug-resistant pathogens.
This systematic review has several limitations that should be considered. Firstly, the enormous variety of methodology used and often not directly detecting the protein, but rather its effect, may create some bias. Secondly, six published studies were not included as full text articles were not accessible online; thus, some important data on OMPs and antimicrobial resistance may be missing. Thirdly, the isolates included in this review were mostly from human origin; therefore, a comparison of prevalence, function and role in antimicrobial resistance of OMPs isolated from humans and pets was not possible, skewing the research towards humans irremediably. Despite limitations on reporting, the strengths of this review are the low prevalence of high RoB studies and the broader understanding of the potential antibiotic resistance for different outer membrane proteins.
This systematic review highlights the need for further investigation about the role of OMPs in antimicrobial resistance among UPEC, as well as the prudent use of antimicrobial agents in both veterinary and human medicine.

Conclusions
In conclusion, our findings showed that several OMPs are related to antimicrobial resistance. ChuA, FepA, FyuA, NmpC, OmpA, OmpC, OmpF, OmpT, OmpX and SlyB were identified in more than 80% of UPEC. However, only OmpC, OmpF, TolC, OmpX, YddB, TosA and Lpp are related with antimicrobial resistance. The classes of antibiotics most affected by antimicrobial resistance conferred through OMPs are fluoroquinolones and β-lactams. These results reflect the urgency of the implementation of new strategies of administration of antimicrobial agents in both veterinary and human medicines, in order to subvert the emergence of multidrug-resistant UPEC strains.