Similarity and Divergence of Phylogenies, Antimicrobial Susceptibilities, and Virulence Factor Profiles of Escherichia coli Isolates Causing Recurrent Urinary Tract Infections That Persist or Result from Reinfection

Abstract

In order to obtain a better molecular understanding of recurrent urinary tract infection (RUTI), we collected 75 cases with repeatedly occurring uncomplicated UTI. The genetic relationships among uropathogenic Escherichia coli (UPEC) isolates were analyzed by pulsed-field gel electrophoresis. While 39 (52%) of the RUTI cases were defined as “persistence” of the same strain as the primary infecting strain, 36 (48%) were characterized by “reinfection” with a new strain that is different from the primary strain. We then examined the antimicrobial susceptibilities and phylogenetic backgrounds of 39 persistence and 86 reinfection UPEC isolates, and screened 44 virulence factor (VF) genes. We found that isolates had significant differences in the following: placement in phylogenetic group B2 (41% versus 21%; P = 0.0193) and the presence of adhesin genes iha (49% versus 28%; P = 0.0233) and papG allele I′ (51% versus 24%; P = 0.003), iron uptake genes fyuA (85% versus 58%; P = 0.0037), irp-2 (87% versus 65%; P = 0.0109), and iutA (87% versus 58%; P = 0.0014), and an aggregate VF score (median, 11 versus 9; P = 0.0030). In addition, 41% of persistence strains harbored three adhesin genes simultaneously, whereas 22% of reinfection isolates did (P = 0.0289). Moreover, 59% versus 29% (P = 0.0014) of persistence and reinfection isolates contained seven types of iron uptake genes. Taken together, the antimicrobial susceptibilities of UPEC isolates had little effect on the RUTI. Compared with reinfection strains, persistence UPEC isolates exhibited higher VF scores and carried more VF genes than may be involved in the development and progression of RUTI.

INTRODUCTION

About half of all women experience at least one urinary tract infection (UTI) during their lifetime, and UTIs are the second most common infection and account for nearly 25% of all infections in noninstitutionalized elderly populations (13). Moreover, about 25% to more than 40% of patients with UTIs experience recurrent UTI (RUTI) within weeks or months (3, 11, 13, 16, 32).

As the most predominant pathogen, uropathogenic Escherichia coli (UPEC) causes almost 75 to 95% of episodes of uncomplicated UTIs (11, 16). Some studies suggest that RUTI is primarily caused by a new strain that is quite different from the original strain (reinfection with a new strain) (4, 22). However, other studies indicate that most (50 to 78%) of the recurrent strains are indistinguishable compared with the original strains (persistence of the original strain) (12, 14, 18, 24, 37, 39). The persistence of UPEC within the bladder could be due to the fact that UPEC strains can persist in the fecal flora for years and cause recurrent ascending UTIs (16). However, RUTI cannot be prevented by daily topical antibiotic treatment of the perineum, suggesting that RUTI does not require pathogen migration from the anus to the urinary tract (17). Furthermore, the intracellular bacterial reservoirs can persist for weeks to months within the urothelium, and intracellular biofilm-like collections of bacteria have been identified in exfoliated cells in the urine of women with cystitis. These observations indicate that UPEC can invade the epithelium and develop quiescent epithelial reservoirs, and the resurgence of intracellular bacterial reservoirs can cause RUTI (1, 23, 28, 29, 36).

The genomes of the UPEC isolates contain more open reading frames and are larger than that of commensal E. coli strains (41), suggesting that UPEC isolates harbor more virulence-associated factors than commensal E. coli isolates. The virulence factors (VFs) possessed by UPEC isolates include fimbrial adhesins, toxins, hemolysin, host defense avoidance mechanisms, and multiple iron acquisition systems (19, 20, 26). Although most of the VF genes have been identified among UPEC isolates causing RUTIs that persist or result from reinfection, no single VF is sufficient for UPEC to cause UTI and no distinct virulence profile can predict RUTI (11).

In order to obtain a better molecular understanding of RUTI, we performed a study on the phylogeny, antimicrobial susceptibilities, and VF profiles of UPEC isolates causing RUTIs that persist and result from reinfection. We collected 75 RUTI cases caused by UPEC during the period from 2007 to 2011. Moreover, we performed a detailed comparison between the persistence and reinfection of UPEC isolates, and investigated their similarity and divergence.

MATERIALS AND METHODS

Definitions.

UTI is defined as a combination of the following symptoms: (i) bacteriuria with ≥10⁴ CFU/ml midstream urine, (ii) the presence of white blood cells (WBC), with ≥5 WBC per high-power field or presence of leukocyte esterase, (iii) and the presence of clinical signs or symptoms of UTI in the host, including dysuria and frequency or urgency of urination. RUTI refers to at least two episodes of repeatedly occurring UTI within an interval of more than 2 weeks. The terms “persistence” and “reinfection” used for RUTI in this study were modified from a previous report (12). “Persistence” is used to mean RUTI with the same strain as the primary infecting strain with or without an intervening negative culture separating the urine cultures. “Reinfection” is characterized by RUTI with a new strain, which is different from the primary strain when there is a negative culture or no intervening culture separating the two urine cultures.

Study setting.

E. coli-caused RUTI cases were collected during the period from 2007 to 2011. The genetic relatedness of the E. coli isolates was analyzed by pulsed-field gel electrophoresis (PFGE) typing. RUTI cases were divided into two groups (persistence and reinfection) according to the PFGE profiles of the E. coli isolates. Moreover, phylogenetic type, antimicrobial susceptibilities, and VFs were assessed. The similarity and divergence between persistence and reinfection were analyzed.

Bacterial isolates.

All clinical E. coli isolates were identified by colony morphology and Vitek 2 GN identification cards (bioMérieux, Inc., Hazelwood, MO). E. coli ATCC 25922 and ATCC 35218 were used as the quality control strains for antimicrobial susceptibility testing. Salmonella enterica serovar Braenderup strain H9812 was used as a reference standard of PFGE. No formal ethical approval was obtained to use the clinical samples, because they were collected during routine bacteriologic analyses in public hospitals, and the data were anonymously analyzed.

PFGE typing.

DNA fingerprints were obtained from the PFGE profiles of genomic DNA digested with XbaI (New England BioLabs) according to the procedures developed by the U.S. Centers for Disease Control and Prevention PulseNet program (33). Strain H9812 was used to normalize migration variation occurring across the gel and to accurately determine sample band sizes. The PFGE patterns were interpreted with BioNumerics software (Applied Maths NV) by using the Dice similarity coefficient. A tree indicating relative genetic similarity was constructed on the basis of the unweighted-pair group method of averages, with a position tolerance of 1%. Strains were considered the same clones if they possessed ≥85% genetic similarity or <4 fragment differences of PFGE profiles (12, 40).

Phylogenetic analysis.

The phylogeny of the E. coli isolates was determined by a triplex PCR method which detects the presence or absence of chuA, yjaA, and TSPE4.C2 (6).

Antimicrobial susceptibility testing.

Antimicrobial susceptibilities of the clinical isolates were determined by measuring the diameters of the zones of complete inhibition in the disk diffusion method. A double-disk synergy test was used to detect extended-spectrum β-lactamase (ESBL) production. All susceptibility results were interpreted according to the 2012 CLSI performance standards (7).

Virulence factor screening.

In total, 44 virulence factor (VF) genes, including 17 adhesin genes, 8 iron-related genes, 7 protectin genes, 3 toxin genes, 4 miscellaneous genes, 4 conserved virulence plasmidic (CVP) region genes, and 1 pathogenicity-associated island (PAI) marker gene, were detected by PCR amplification. The reactions were carried out in the following nine pools: pool 1, malX (PAI) (925 bp), papA (717 bp), fimH (508 bp), kpsMT III (392 bp), papEF (326 bp), ireA (254 bp), and ibeA (171 bp); pool 2, cnf-1 (1,105 bp), fyuA (787 bp), iroN (667 bp), bmaE (507 bp), sfa (410 bp), iutA (302 bp), and papG allele III (258 bp); pool 3, hlyD (904 bp), rfc (788 bp), ompT (559 bp), papG allele I′ (479 bp), papG allele I (461 bp), kpsMT II (272 bp), and papC (205 bp); pool 4, gafD (952 bp), cvaC (679 bp), fliC (547 bp), cdtB (430 bp), focG (364 bp), traT (290 bp), and papG allele II (190 bp); pool 5, papG allele I (1,190 bp), papG alleles II and III (1,070 bp), iha (829 bp), afa (594 bp), iss (323 bp), sfaS (244 bp), and kpsMT (K1) (153 bp); pool 6, sitA (608 bp), feoB (470 bp), and irp-2 (287 bp); pool 7, tsh (420 bp); pool 8, iucC (541 bp); and pool 9, the CVP region genes hlyF (599 bp), etsC (359 bp), cvaA (319 bp), and ompTp (189). The primer sets and PCR conditions were used as previously described (31, 34, 35). The VF score was calculated for each isolate as the sum of all VF genes for which the isolate tested as positive (8).

Statistical analysis.

Pearson's χ² test (for frequencies of higher than 5) or Fisher's exact test (for small contingency tables) was used to determine the significance of the prevalence values. The differences in the aggregate VF scores of persistence and reinfection were compared with the Wilcoxon rank sum test. For all tests, a P value of <0.05 was considered statistically significant. Statistical analysis was performed using the Chinese High Intellectualized Statistical Software (CHISS, version 2010) package.

RESULTS

Clinical information and UPEC isolates.

A total of 75 RUTI cases caused by UPEC were collected during the period from 2007 to 2011. Among those cases, 39 were persistence (52%) and 36 (48%) were reinfection. In total, 91% (68 of 75) of RUTI patients were female, 88% (66 of 75) were outpatient, and the average age was 68.35 years. The duration of persistence ranged from 15 days to 3.5 years with or without intervening negative culture separating the urine cultures, while some patients suffered reinfection caused by several UPEC isolates for more than 3 years. A total of 39 persistence isolates and 86 reinfection isolates were collected for further analysis, respectively.

Phylogenetic background.

Among these 125 detected UPEC isolates, most isolates belonged to phylogenetic group D (n = 49, 39%) and B2 (n = 34, 27%). However, the majority of RUTI persistence was caused by the UPEC strains from phylogenetic group B2 (Table 1).

TABLE 1.

Number of E. coli isolates which belong to different phylogenetic groups and are resistant to various antimicrobials

Item	No. of persistence isolates (%) (n = 39)	No. of reinfection isolates (%) (n = 86)	P valuea
Phylogenetic group
A	9 (23.08)	20 (23.26)	0.9825
B1	4 (10.26)	9 (10.47)	0.873
B2	16 (41.03)	18 (20.93)	0.0193
D	10 (25.64)	39 (45.35)	0.0365
ESBL production	22 (56.41)	43 (50)	0.5063
Antibiotics
Piperacillin	29 (74.36)	65 (75.58)	0.8834
Cefuroxime	22 (56.41)	46 (53.49)	0.7612
Cefotaxime	22 (56.41)	45 (52.32)	0.6714
Ceftazidime	9 (23.08)	16 (18.61)	0.5625
Imipenem	0	0
Amoxillin/clavulanic acid	7 (17.95)	16 (18.61)	0.9301
Trimethoprim-sulfamethoxazole	25 (64.11)	55 (63.95)	0.9872
Ciprofloxacin	32 (82.05)	63 (73.26)	0.2861
Amikacin	2 (5.13)	10 (11.63)	0.4149

^aA P value of <0.05 (bold) was considered statistically significant.

Antimicrobial susceptibilities.

Table 1 lists the resistance rate of commonly used antibiotics. We found no significant difference in the antimicrobial susceptibilities between two groups.

Prevalence of VF genes.

Table 2 lists the distribution of common VF genes among UPEC isolates analyzed in this study. Adhesin determinants (iha and papG allele I′) and iron-related genes (fyuA, irp-2, iucC, and iutA) were significantly more prevalent among persistence UPEC strains. Persistence isolates had a higher aggregate VF score (median, 11 versus 9; P = 0.0030) and carried more types of adhesin and iron-uptake genes than reinfection strains (see Fig. 2 and 3). For the majority of persistence isolates, the VF score ranged from 10 to 12, and for most reinfection strains, it ranged from 3 to 12 (Fig. 1). About 41% of persistence isolates harbored three types of adhesin genes, whereas 22% of reinfection strains did (P = 0.0289). Figure 2 shows that 18% of persistence isolates carried only one adhesin gene, whereas 36% of reinfection strains did (P = 0.0415). Moreover, about 79% (31/39) of strains possessed six of seven types of iron uptake genes, whereas 48% (41/86) of reinfection isolates did (P = 0.0009). Figure 3 shows that 11% of persistence strains carried two of five types of iron uptake genes, whereas 50% of reinfection isolates did (P = 0.0007).

TABLE 2.

Prevalence of VFs among UPEC causing RUTIs that persist and result from reinfection^a

Virulence factor	No. of persistence isolates (%) (n = 39)	No. of reinfection isolates (%) (n = 86)	P value
Adhesins
afa	4 (10)	11 (13)	0.9148
fimH	33 (86)	72 (84)	0.8994
iha	19 (49)	24 (28)	0.0233
papA	1 (3)	8 (9)	0.4316
papEF	3 (8)	2 (2)	0.3006
papG allele I′	20 (51)	21 (24)	0.0030
papG allele II	0 (0)	9 (10)	0.0848
sfa	6 (15)	5 (6)	0.1588
Iron-related
chuA	27 (69)	57 (66)	0.7447
feoB	39 (100)	81 (94)	0.2964
fyuA	33 (85)	50 (58)	0.0037
ireA	0 (0)	6 (7)	0.2153
irp-2	34 (87)	56 (65)	0.0109
iroN	6 (15)	11 (13)	0.6951
iucC	37 (95)	67 (78)	0.0187
iutA	34 (87)	50 (58)	0.0014
sitA	31 (79)	66 (77)	0.7333
Protectins
cvaC	4 (10)	6 (7)	0.7868
iss	5 (13)	14 (16)	0.6178
traT	33 (85)	69 (80)	0.5579
PAI
malX	9 (23)	17 (20)	0.6728
CVP region	4 (10)	9 (10)	0.9717

^aA P value of <0.05 (bold) was considered statistically significant. PAI, pathogenicity-associated island; CVP, conserved virulence plasmidic.

Fig 2.

Percentage of isolates carrying different numbers of adhesin genes.

Fig 3.

Percentage of isolates carrying different numbers of iron-related genes.

Fig 1.

Number of isolates with different VF scores.

DISCUSSION

In the present study, we showed that 52% (39/75) of cases were persistence during the RUTI episodes, and this result was consistent with Foxman's study showing that 50% of RUTIs exhibit persistence (14). However, our result varied from many studies on the percentage of RUTIs that persist, which reported that 65 to 78% of RUTIs exhibit persistence (12, 18, 24, 37, 39). The experimental scheme might have a significant effect on the difference, especially the definition of RUTI and subsequent collection methods of UPEC isolates causing RUTIs that persist or result from reinfection. For example, the least intervals between two UTI episodes are 7 or 8 days in some studies where the high percentages (more than 70%) of persistence are observed among RUTIs (12, 39). In our study, the samples were not collected at fixed intervals, because most of the RUTI patients were outpatient and they went to the hospital only after the symptoms of UTI occurred. Therefore, omissions of UTI episodes in some patients were inevitable, leading to some bias.

Following an initial UTI, the same-strain RUTI has been reported to persist up to 3 years (17). In this study, the longest duration of persistence was 3.5 years. It may be due to the fact that the upstream movement of bacteria may actually transport persistent UPEC isolates from the vagina, skin, or rectum into the urinary tract, resulting in a repeated accumulation of UPEC in the urinary tract and RUTI persistence (16). However, RUTI cannot be prevented by daily topical antibiotic treatment of the perineum, suggesting that RUTI does not require pathogen migration from the anus to the urinary tract (17). In addition, different factors, such as sexual intercourse, spermicide use, incontinence, and cystocele, are known to be risk factors for the onset of RUTI in women of different age groups (32, 38). Besides these host factors, the characters of the pathogens, such as the antimicrobial susceptibilities, the phylogenetic group status, and the VF profiles of UPEC, may greatly contribute to the development and progression of RUTI.

We found no significant difference of the antimicrobial susceptibilities between the two groups of UPEC isolates (Table 1). Moreover, our results were in good agreement with a multicenter study of UPEC isolates in China (5), suggesting little effect of the antimicrobial susceptibilities of UPEC isolates on the development and progression of RUTI. Since antimicrobial resistance can increase the prevalence of UPEC strains and antibiotics can facilitate the persistence formation of UPEC, further studies on the resistance characters of UPEC isolates will provide a better molecular understanding of UTIs (9, 10).

In this study, phylogenetic group D was responsible for the majority of RUTI-causing UPEC isolates (39%), followed by group B2 strains (27%). However, the primary persistence of RUTIs in our study was associated with group B2, whereas reinfections were mainly caused by group D UPEC isolates (Table 1). Another multicenter study in China revealed that 54% and 19% of UTIs are caused by group D and B2 UPEC strains, respectively (5). Therefore, in China, group D UPEC isolates account for the majority of UTI and reinfection of RUTI, whereas group B2 strains were the primary cause of persistence of RUTI. However, the predominant E. coli strains belong to phylogenetic group B2 in both UTIs and RUTIs, and even fecal E. coli strains in other areas (27, 39, 42). Further efforts are required to understand the mechanisms underlying differential prevalence. It has been proposed that phylogenetic group B2 status, but not virulence, is the key factor for persistence of E. coli strains in the intestinal microflora for two reasons. First, the extraintestinal virulence character of UPEC may be a coincidental by-product of commensalism (25), and second, phylogenetic group B2 UPEC isolates can persist in the intestinal microflora of infants, independent of carriage of virulence genes (30). Furthermore, group B2 strains tend to substitute other groups of UPEC isolates through enhanced niche fitness or inhibition of other clones (27). Therefore, phylogenetic group B2-associated characteristics other than VFs may play a significant role in the pathogenesis of RUTI persistence.

Higher aggregate virulence scores have been observed in UPEC isolates than in fecal clones (21, 27), suggesting that UPEC isolates harbor more virulence-associated factors. Although no single VF is sufficient for UPEC to cause infections, timely and stepwise expression of multiple, potentially redundant VFs can significantly contribute to UTI development (11). Adhesins and iron uptake systems are associated with the persistence of E. coli strains in the commensal intestinal microbiota (25). Therefore, UPEC strains with high adhesive and iron capture capacity might serve as the reservoirs of pathogen causing RUTI to persist, because persistence isolates exhibited a significantly higher aggregate VF score and harbored more types of adhesins and iron capture systems than reinfection strains (Fig. 1 to 3). In addition, the persistence UPEC isolates presented a higher prevalence of papG allele I′ and iha adhesins than the reinfection strains, and we found similar trends for the prevalence of fyuA, irp-2, iucC, and iutA iron capture systems (Table 1), suggesting the involvement of these genes in the pathogenesis of RUTI persistence. However, the genes coding P fimbriae (except for papG allele I′) exhibited a low prevalence (Table 1), which may be due to the fact that the proportion of strains expressing P fimbriae was high with acute pyelonephritis and low with cystitis (19). In this study, UTI cases were collected according to the diagnostic criteria for cystitis. Therefore, most UTI patients had cystitis but not pyelonephritis, leading to a low prevalence of P fimbriae. Persistence isolates exhibited significant higher carriage of papG allele I′ and iha adhesins (Table 1), and more adhesin genes were detected among persistence strains (Fig. 2). Although Iha contributes less to virulence (15), the combined effect of various adhesins may confer some persistence-causing advantages on UPEC isolates and the development of RUTI persistence. Many mechanisms can facilitate the iron acquisition in bacteria, because it is essential for bacterial growth to obtain and hoard iron in an iron-limiting environment, such as the urinary tract (2). For instance, IutA and FyuA have been shown to be the most critical iron capture functions in the urinary tract. This is particularly true for IutA, which mediates aerobactin utilization and more significantly contributes to bladder colonization than kidney infection (15). Our results showed that IutA and FyuA, combined with IRP-2 and IucC, promoted not only the bladder colonization but also UPEC persistence in the urinary tract. Moreover, two-thirds of persistence strains possessed more than seven types of iron uptake genes (Fig. 3), indicating a key role of the high carriage rate of iron capture systems in the development and progression of RUTI persistence.

The present study has several limitations. First, because this study is a retrospective analysis, limited patient information has been collected. We focused only on the phenotypic and genotypic characteristics of the UPEC isolates and not on the clinical features of patients with RUTI. Second, the urine samples were not collected at fixed intervals; therefore, some bias could result from the collection procedure. Third, the two persistence types were not distinguished from each other in this study. One is persistent UTI, where the pathogen has never been eliminated, and the other is relapsing UTI, with a negative urine culture or a culture with a strain different from the primary infecting strain separating the two urine cultures (12).