Antibiotic Resistance,Virulence,and Genetic Background of Community-Acquired Uropathogenic Escherichia coli from Algeria

Introduction

Escherichia coli is a harmless commensal of the digestive tract of humans and many warm-blooded animals; however, some strains are pathogenic causing intestinal (intestinal pathogenic E. coli [InPEC]) and extraintestinal (extraintestinal pathogenic E. coli [ExPEC]) infections in humans and animals. ¹³ Among ExPEC, uropathogenic E. coli (UPEC) is the most frequent causal agent in human community-acquired urinary-tract infections (CA-UTI). ²² UPEC is characterized by a particular genetic background (phylogenetic groups B2 and D) and possesses an arsenal of virulence factors involved in their pathogenicity.^11,13,59 Multidrug resistance in community-acquired UPEC (CA-UPEC) is rising worldwide, leading to antibiotic therapy failure, UTI complications, and more morbidity and mortality.^22,46 This may be linked to human and veterinary medical practices, anthropogenic environmental pressure, food chain, and others, which can impact antibiotic resistance in the community.

Antibiotic resistance in UPEC concerns the major therapeutic classes as beta-lactams, sulfamethoxazole–trimethoprim, and fluoroquinolones.^22,28 For beta-lactams, most worrying is the emergence of extended-spectrum beta-lactamases (ESBLs) and carbapenemases, particularly CTX-M-15 and NDM-1 harbored by the worldwide successful sequence type ST131.^9,10 Sulfamethoxazole–trimethoprim association was less used in different parts of the world ²⁸ because of increased resistance mediated mainly by acquired sul and dfr genes encoding drug-resistant dihydropteroate synthetase and dihydrate folate reductase, respectively.^7,54 Similarly, the emergence of fluoroquinolone-resistant E. coli is now limiting their efficiency.^14,46 High-level fluoroquinolone resistance is mainly due to chromosomal mutations in DNA gyrase and topoisomerase IV genes,^14,57 commonly found among CTX-M-15 producing UPEC isolates belonging to ST131. ⁵⁹

E. coli CA-UTI have emerged as a major public health problem worldwide, including Algeria. ⁴⁹ Data on antibiotic resistance and virulence of UPEC and their epidemiology are scarce or even nonexistent in Algeria and in North Africa. In this context, the objective of this study was to assess antimicrobial resistance, elucidate mechanisms involved, and investigate virulence and genetic background of CA-UPEC strains isolated in Algiers, Algeria. These investigations will be a valuable complement to existing international data, allowing to more globally apprehend the state and trends of antibiotic resistance, virulence, and epidemiology of UPEC.

Materials and Methods

Results

Antibiotic susceptibility testing showed resistance to all tested molecules except imipenem, ertapenem, fosfomycin, and colistin, which were active against all isolates (Table 1). The highest resistance rates were observed against amoxicillin (71.3%) and cefalotin (71.3%) followed by tetracycline (51.3%), sulfonamides (49.3%), trimethoprim (36.7%), sulfamethoxazole–trimethoprim (36.7%), nalidixic acid (34.7%), and fluoroquinolones (22.7–26%). The lowest resistance rates were noted against aminoglycosides (0.7–17.3%), third-generation cephalosporins (6–6.7%), amoxicillin–clavulanate (6%), aztreonam (6%), cefoxitin (4%), and fourth-generation cephalosporins (3.3%). Agar dilution MICs were from 32 to >128 μg/ml for cefotaxime and ceftriaxone, 32 to 64 μg/ml for ceftazidime, 16 to >128 μg/ml for cefoxitin, >128 μg/ml for sulfamethoxazole–trimethoprim, 0.5 to >128 μg/ml for ciprofloxacin, and 16 to >128 μg/ml for tetracycline. A rate of 80.7% of isolates was resistant to at least one antibiotic and a rate of 46.7% was multidrug resistant (MDR); they were distributed over 33 resistance patterns (resistance to 4–19 antibiotics). The analysis of drug resistance according to the gender and age of patients showed higher resistance rates to quinolones and fluoroquinolones in males than in females (53.8% vs. 30.6%, p=0.04 and 42.3% vs. 18.5%, p=0.02) and in patients >60 years of age than those aged 19–60 years (males: 81.8% vs. 33.3%, p=0.02 and 72.7% vs. 20%, p=0.01; females: 61.9% vs. 24.3%, p=0.001 and 38.1% vs. 14.6%, p=0.03). In addition, resistance rates to fourth-generation cephalosporins (11.5% vs. 1.6%, p=0.04) and gentamicin (19.2% vs. 5.6%, p=0.04) were more high among males. No statistical difference was observed in the rates of multidrug resistance phenotype between males and females (57.7% and 44.3%), whatever their age.

Concerning beta-lactam susceptibility, 28.7% of isolates were sensitive to all beta-lactams; the rest were distributed into the following four resistance phenotypes: (1) high-level penicillinase phenotype (HLPP, 62%), (2) inhibitor-resistant penicillinase phenotype (IRPP, 1.3%), (3) ESBL phenotype (ESBLP, 4%), and (4) high-level cephalosporinase phenotype (HLCP, 4%). Beta-lactamase genes detected are indicated in Table 2. Genes bla_TEM were present in almost all HLPP isolates (96.8%); bla_TEM-31 or bla_TEM-35 in the two IRPP isolates and bla_CTX-M-15 in the six ESBLP isolates, of which one harbored bla_TEM-4 and one bla_TEM-1+bla_SHV-2a. Genes encoding plasmid-mediated cephalosporinases were absent, while the six HLCP isolates had multiple mutations in the promoter/attenuator region of the chromosomal bla_AmpC gene. In four isolates, mutations occurred: (1) in the weak wild-type promoter at position −32 (T→A) changing the −35 box to the consensus sequence TTGACA (mutated base underlined) and between positions −13 and −14 (a thymine insertion) in the spacer region between −35 and −10 boxes resulting in a 17bp interbox distance instead of 16 bp, (2) in the attenuator region at multiple positions +22 (C→T), +26 (T→G), +27 (A→T), +32 (G→A), and +37 (G→C), and (3) outside the promoter/attenuator at positions 70 (C→T) and 81 (A→G). In the two remaining isolates, mutations occurred at positions (1) −42 (C→T) and −18 (G→A), creating an alternative promoter composed of new −35 (TTGACA, −37 to −42) and −10 (TATCGT, −14 to −19) boxes separated by 17bp, and (2) −1 (C→T), +58 (C→T), and +81 (A→G) outside the promoter/attenuator.

Non-beta-lactam resistance genes detected are indicated in Table 2. For tetracycline-resistant isolates, tetA and tetB were present at prevalences of 44.2% and 53.2%, respectively, and both genes at 2.6%. Among sulfonamide-resistant isolates, 82.4% possessed sul genes identified as sul2 (60.8%), sul1 (46%), and sul3 (6.7%); 31.1% of isolates harbored both sul1 and sul2 genes. Of trimethoprim-resistant isolates, 61.8% had dfr genes identified as dfrA14 (25.4%), dfrA1 (18.2%), dfrA12 (16.3%), and dfrA25 (5.4%); 3.6% of isolates harbored two dfr genes. Among sulfamethoxazole–trimethoprim-resistant isolates, 56.4% had at least one sul gene and one dfr gene with different gene combinations (Table 2). All isolates resistant to fluoroquinolones at a high level (ciprofloxacin MICs: 4 to 128 μg/ml) had a double mutation in the GyrA subunit gene of DNA gyrase (S83L-D87N) and a double (S80I-E84G/V) or single mutation (S80I) in the ParC subunit gene of topoisomerase IV. Isolates resistant to nalidixic acid and sensitive to ciprofloxacin had only mutation S83L in GyrA with or without mutation S80I in ParC or no mutation (Table 2). Eight (5.3%) of the total isolates had PMQR determinants identified as qnrB5 (n=3), qnrS1 (n=3), and aac(6′)-Ib-cr (n=2), of which 3 (one qnrS1-positive and aac(6′)-Ib-cr-positives) exhibited high-level resistance to fluoroquinolones; the rest were sensitive. The two aac(6′)-Ib-cr genes were found in ESBL isolates harboring bla_CTX-M-15 or bla_CTX-M-15+bla_TEM-4. The genes qnrA, qnrC, qnrD, qepA, oqxA, and oqxB were absent. In aminoglycoside-resistant isolates, aac(6′)-Ib genes encoding the most prevalent aminoglycoside-modifying enzyme were absent.

Class 1 and class 2 integrons were found in 54 (36%) and 5 (3.3%) isolates, respectively, and in both classes in 2 (1.3%) isolates. A rate of 86.9% of integron-positive isolates was MDR versus 19.1% of integron-negative isolates (p=0.0001). All sul1-positive isolates had class 1 integrons and almost all dfr-positive isolates (32/34, 94.1%) had class 1 and/or class 2 integrons.

Based on phylogenetic grouping, isolates were assigned to phylogroups B2 (n=57 isolates, 38%), A (27, 18%), B1 (27, 18%), D (18, 12%), F (9, 6%), C (7, 4.7%), and Clade I (5, 3.3%). Antimicrobial resistance distribution among phylogroups showed a low prevalence of resistance to quinolones, fluoroquinolones, and sulfamethoxazole–trimethoprim in phylogroup B2, while the prevalence of resistance to other antibiotic classes and of the multidrug resistance phenotype was important in this phylogroup. Phylogroup D showed the highest resistance rates for amoxicillin, cefalotin, quinolones, fluoroquinolones, sulfamethoxazole–trimethoprim, tetracycline, and for the multidrug resistance phenotype (Table 1). No difference was found in the distribution of phylogenetic groups according to the gender and age of patients. MLST performed on all MDR isolates of phylogroups B2 and D (35 isolates) identified 14 different STs (Table 3). For phylogroup B2, 10 isolates belonged to ST131, of which 7 were O25b serotype, 5 to ST73 (CC73), 2 to ST127, 2 to ST998, and 3 to ST12, ST636, and ST4494. For phylogroup D, five isolates belonged to ST69, three to ST354 (CC354), and five to ST117, ST394, ST405 (CC405), ST1177, and ST4529. ST4494 and ST4529 are new STs identified for the first time in this study, representing novel combinations of previously characterized alleles. They were submitted to the MLST website and approved (http://mlst.ucc.ie/mlst/dbs/Ecoli). ST131 predominates in patients older than 60 years compared to those aged 19–60 years (66.7% vs. 15.4%, p=0.007), while non-ST131 clones predominate in patients aged 19–60 years (84.6% vs. 33.3%, p=0.007). ERIC-PCR genotyping of total isolates showed a genetic diversity for 92 (61.3%) isolates, while the remaining isolates were clonally related (groups of 2 to 15 isolates), most of them belonging to phylogroups B2 and D (48, 82.7%). ERIC-PCR results of B2 and D MDR isolates matched with those of MLST; clonally related isolates (by ERIC-PCR) were included in the same ST (ST131, ST69, ST354, ST73, and ST998) (Table 3).

The results of virulence factor determination in phylogroups, B2 and D MDR isolates, are indicated in Table 4. All virulence factors were detected with variable rates; their combination allows to distinguish different virulence profiles.

Discussion

The prevalence of resistance to multiple antibiotics in CA-UPEC is high in our study. The highest resistance rates were noted for amoxicillin (71.3%) and cefalotin (71.3%). Endemic worldwide resistance to these molecules limits their efficiency in the empirical treatment of UTI.^28,46 Noteworthy was the resistance rate to tetracycline (51.3%) comparable to rates of 53% and 60.8% reported by Ho et al. ²⁹ and Martinez et al. ⁴¹ Tetracycline is used less in human medicine; the high resistance rate may be due to its common use as an animal growth promoter leading to the transmission of resistant bacteria to humans by direct contact and through the food chain. ²⁷ Resistance to sulfamethoxazole–trimethoprim and fluoroquinolones, the most common antibacterial drugs in UTI treatment, was substantial (rates >20%) and no longer allowed their empirical prescription as recommended.^14,28 As reported in the community, ³⁸ UTI in males, particularly those aged older than 60 years, were associated with more fluoroquinolone resistance. Imipenem and meropenem were active against all isolates. Carbapenems are the treatment of choice in severe ESBL E. coli infections; however, the more recently emerging resistance to these molecules in gram-negative bacilli, including E. coli, ⁵⁹ requires their use with discernment. Colistin and fosfomycin were also active on all isolates as in many studies reporting the susceptibility of over 90% of CA-UPEC.^20,29,39 The return to these old molecules is an appropriate alternative in the treatment of MDR pathogen infections; fosfomycin is currently recommended in front-line therapy of CA-UTI. ²⁸

In E. coli, beta-lactamase production remains the most important mechanism of beta-lactam resistance. ⁹ TEM beta-lactamases were the most frequent. The prevalence of ESBL is similar to reported rates in CA-UPEC (1.3–6.6%)^20,36,48; CTX-M-15 type prevail, this is the most common and prevalent ESBL in clinical Enterobacteriaceae in Algeria and worldwide.^8,43 ESBL isolates were MDR; this is worrisome because of therapeutic option limitation. The prevalence of inhibitor-resistant TEM was 1.3%, consistent with their low prevalence or absence in recent reports on E. coli.^17,40 Overexpressed chromosomal AmpC beta-lactamases were found at low prevalence as also reported in CA-UPEC in France. ²⁰ Mutations found in the bla_AmpC gene promoter region strengthen the wild-type promoter or create an alternate strong promoter, ⁶⁰ while the effect of those occurring in the attenuator and outside the promoter/attenuator is negligible.^44,60

All tetracycline-resistant isolates had tetA and/or tetB genes, which are the most common in E. coli of human and animal origin.^34,55 Resistance to sulfonamides and trimethoprim was mainly due to sul- and dfr-acquired genes, as previously described in UPEC^7,23,54 and in animal E. coli.^54,55 Resistance to sulfonamides and trimethoprim in sul- and dfr-negative isolates may be due to other mechanisms such as the permeability barrier, efflux pumps, and mutational changes in chromosomal target enzymes. As in UPEC in Europe and Canada, sul2 was the most frequent followed by sul1, while sul3 was rare. ⁷ The dfr gene distribution was different from that reported in trimethoprim-resistant CA-UPEC in the previous study where dfrA14 (the most frequent in our study) was not detected and dfrA17 (absent in our study) was the most frequent as dfrA1.^7,54 The concomitant presence of sul and dfr genes in sulfamethoxazole–trimethoprim-resistant isolates may explain the therapeutic inefficiency of this drug association. The high prevalence of class 1 and/or class 2 integrons in sulfonamide- and trimethoprim- resistant isolates is consistent with sul1 as a part of the 3′ conserved sequence of class 1 integrons and dfr as gene cassettes.^7,23,54 The high level of resistance to fluoroquinolones was correlated with the presence of multiple mutations in the QRDR region of type II topoisomerase genes gyrA and parC; mutations found in this region are the most frequently reported in E. coli.^21,30 Compared to mutations in gyrA and parC genes, which prevailed in our study, PMQR determinants were rare (5.3%). Although PMQR determinants confer low-level resistance to quinolones, they can promote emergence of high-level quinolone resistance.^30,57 The concomitant presence of aac(6′)-Ib-cr and ESBL genes is a further concern.

Integrons are genetic supports of multiple antibiotic resistance gene cassettes ²⁶ ; our study showed a high prevalence (46.7%) of the multidrug-resistant phenotype in association with a high prevalence of integrons.

The genetic background of the 150 isolates showed the prevalence of B2+D phylogroups (50%), in agreement with the belonging of ExPEC to these groups.^31,32 The consequent rate of A+B1 (36%) may be related to the recent emergence of these groups among ExPEC, as described by Ewers et al. ¹⁹ The low prevalence (13%) of newly described phylogroups is similar to that reported in human fecal isolates. ¹¹ Phylogroups B2 and D showed an important proportion of MDR isolates. The 35 MDR isolates of these phylogroups belonged to 14 STs; the most frequent were ST131, ST73, and ST69 (57.1%), which are major pandemic clonal lineages of ExPEC associated to both community and nosocomial infections.^51,63 Similar to the study of Banerjee et al., ⁴ ST131 was predominant among elderly patients and non-ST131 clones among those younger than 60 years. Recently, the worldwide increase of CTX-M-15-producing E. coli has been associated to the subclone O25b-ST131,^10,52 in agreement with our results showing CTX-M-15-producing isolates among O25b-ST131. The latter exhibited a high level of fluoroquinolone resistance related to multiple mutations in gyrA and parC genes, as commonly reported in ST131 isolates.^52,59 Due to the trend of ST131 to evolve toward more antibiotic resistance, particularly acquisition of NDM-1 carbapenemase and fosfomycin resistance, ⁹ its presence in the community context is worrying. All D-ST69 isolates were resistant to sulfamethoxazole–trimethoprim; this ST has been associated to trimethoprim–sulfamethoxazole resistance in UPEC.^51,63 The other identified STs were less frequent; among them, new ST4494 and ST4529 were described for the first time in this study and ST354, ST127, ST12, and ST405 were reported in UPEC clones, but not widely disseminated.^1,9,51,63 ST354 was reported as a successful CTX-M-14-producing clone in Spain, ⁶³ while ST354 isolates were nonproducers of ESBL in this study. ST127 has been recently reported as a newly emerged community clone in Northwest England with a higher pathogenic potential than ST131, ST73, and ST69. ¹ The D-ST405 is increasingly reported worldwide; it has been recently described as another clone involved in the dissemination of CTX-M-15 ESBL^9,63; however, the ST405 was not a producer of ESBL in our study. ST117 and ST998 are two low prevalent other STs in our study and they were reported mostly in potential UPEC from retail meat.^61,65 The presence of ERIC-PCR and MLST clonally related isolates among ST131, ST69, ST354, ST73, and ST998 in unrelated patients may suggest their acquisition from a common source; MDR UPEC can spread in the community from the environment and animals through the food chain.^22,46,61

Virulence analysis of B2 and D MDR isolates showed that type 1 fimbriae (fimH) were ubiquitous (97.1%), followed by P fimbriae (papA and papGII) in 37.1% of isolates belonging to ST131, ST73, ST69, and ST12. FimH adhesin plays an important role in bladder colonization, intracellular biofilm formation, and persistence in the bladder. ⁶⁴ PapGII adhesin is mainly associated with human pyelonephritis and bacteremia. ¹⁵ In synergy with P-fimbrial adhesin, FimH also mediate biofilm formation in the renal tubular lumen and facilitate renal colonization. ⁴² Genes of Hra, F1C, AFA/Dr, and S adhesins were present at a low frequency (25.7%, 14.3%, 14.3%, and 5.7%), mostly in B2 non-ST131 clones. F1C, AFA/Dr, and S adhesins have redundant functions with P adhesins and they facilitate bacterial dissemination and persistence within the host. ¹⁵ The heat-resistant agglutinin (Hra) gene has been identified as a putative virulence factor in UTI. ⁵⁶ Besides the heme iron acquisition ChuA gene intrinsic to phylogroups B2 and D, almost all isolates (97.1%) had genes of aerobactin and/or yersiniabactin, whereas salmochelins and enterobactin genes were less frequent (34.3% and 25.7%). Aerobactin, yersiniabactin, and heme receptor ChuA play an important role in uropathogenesis. ²⁴ Yersiniabactin also confers protection against the host antibacterial functions through reduction of reactive oxygen species formation by innate immune cells and promotes biofilm formation in human urine through upregulation of the ferric yersiniabactin outer membrane receptor FyuA. ²⁵ All isolates were resistant to the bactericidal effect of human serum and most of them harbored associated genes encoding group II or III capsular polysaccharides, which also confer protection from phagocytosis and opsonization. These polysaccharide capsules have also been recently implicated in intracellular biofilm development during cystitis. ² hlyA and cnf-1 genes were present in 34.3% and 31.4% of isolates belonging to the phylogroup B2, respectively. Toxins encoded by these genes mediate disruption of phagocyte and epithelial cell function and survival, thereby increasing amplitude and chronicity of UTI.^16,37 The simultaneous presence of hlyA, cnf-1, and hra genes in ST131, ST73, and ST998 isolates may be explained by their genetic linkage; they have been located on a pathogenicity island PAI IIJ₉₆-like domain exclusively restricted to phylogenetic group B2. ⁶ Colibactin genes were present in 28.6% of isolates belonging to phylogroup B2 non-ST131. Colibactin plays an important role in long-term persistence in the gut, an important upstream step in the uropathogenesis. ⁴⁷ Within the B2 phylogroup, colibactin is considered as a high virulence marker and its production is associated with increased likelihood of bacteremia. ³³ Autotransporter serine protease Vat and/or Pic genes were found mostly among the phylogroup B2 non-ST131 clones; they target leukocyte surface glycoproteins involved in leukocyte activation, migration, signaling, and apoptosis, resulting in impaired innate and adaptive immune responses. ⁵³ Recently, pic and vat genes have been involved in the fitness of UPEC in systemic infections. ⁵⁸ All isolates had the ability to form a biofilm allowing bacterial protection against antimicrobial agents and host immune defenses.^2,42,62,64 Overall, B2 and D MDR CA-UPEC isolates possessed virulence traits having a role in colonization, invasion, and long-term persistence in humans.

To our best knowledge, this is the first study in Algeria relating antibiotic resistance, virulence, and genetic background of CA-UPEC. It highlighted a high prevalence of multidrug resistance resulting from mutations, acquired multiple resistance genes, mobile genetic elements, and successful clones. The presence of international clones ST131, ST73, and ST69 constitutes a major threat to public health; these MLST lineages are reported for the first time in the community in Algeria. Two new sequence types ST4494 and ST4529 are also described for the first time in this study. Besides multidrug resistance, CA-UPEC isolates harbored multiple virulence factors involved in colonization, invasion, tissue injury, and complications; this increases morbidity and mortality. A constant monitoring of antimicrobial resistance is required to ensure therapeutic efficacy and prevent the emergence and spread of MDR clones. Given their inefficiency, sulfamethoxazole–trimethoprim and fluoroquinolones should be avoided in empirical therapy of UTI and fosfomycin could be an appropriate alternative. Considering evolution of antibiotic resistance and virulence in UPEC, alternative strategies for UTI prevention such as vaccination, antiadhesive approaches, and prophylactic approaches become important.