Introduction

Urinary tract infections (UTI) result in significant morbidity in young women, as about half have at least one cystitis episode by the age of 35 years [1] and a similar proportion experience recurrent episodes [2].

Cranberry juice has traditionally been used to reduce recurrent UTIs. In two randomized trials, regular consumption of cranberry or cranberry-lingonberry juice (CLJ) reduced the occurrence of UTIs by 30–40% in adult women [3–5]. The mechanism of action is not fully understood but it has been suggested cranberry products confer anti-adhesive properties on urine with respect to UPEC (uropathogenic Escherichia coli) strains and uroepithelial cells [6].

It has been suggested recently that recurrent UTI may be caused by Escherichia coli that reside in a dormant, biofilm-like state in intracellular pods in the bladder [7, 8]. Biofilm formation by uropathogenic Escherichia coli (UPEC) in vitro has been shown to correlate with the presence of virulence factors of UPEC strains such as papC and PapG alleles and sfa/focDE [9]. The ability of UPEC strains to form biofilm has been associated with the clinical presentation of UTI, as such strains isolated from patients with pyelonephritis proved to be strong biofilm producers in our previous study [10].

We now used a similar CLJ dose to that which showed good efficacy in preventing UTIs in our earlier clinical trial among young adult women and tested whether urine obtained after consumption of the juice reduces the biofilm formation and gene expression of virulence factors of UPEC strains in vitro [3].

Methods

Subjects, juice and urine collection

Twenty healthy adult women employed in Oulu University Hospital gave informed consent and volunteered to drink CLJ and donate urine for laboratory studies. The Ethical Committee of Oulu University Hospital was consulted about the study protocol. They were asked to drink 100 ml of CLJ at the same time every morning for two weeks, including the day when urine was collected. During a one-week control period the volunteers were asked to avoid any products containing cranberry or lingonberry or other Scandinavian berries such as bilberries or raspberries. Otherwise they were to follow their normal diet during both the juice period and the control period. The CLJ contained 8.19 g of cranberry concentrate and 7.0 g of lingonberry concentrate in 100 ml (manufactured and donated by the same manufacturer as for the earlier clinical trial, Eckes-Granini, Finland, previously known as Marli). The total carbohydrate content of the juice was 4.3 g per 100 ml. The absolute amount of cranberry concentrate administered daily was close to the single daily dose used in our earlier clinical trial (7.5 g per day in the clinical trial) and the amount of lingonberry concentrate was greater (1.7 g per day in the clinical trial) [3]. Urine samples were collected 2–4 hours after the last dose of the juice. Control samples were collected at the same time in the morning after the control period in order to make samples comparable for their other properties such as urine concentration. The time frame for collecting these samples was chosen on the basis of previous reports showing that anti-adhesive properties are likely to be present in urine within 2–4 hours after drinking cranberry juice [6]. The urine samples were immediately put in a refrigerator (+4°C) and were sterilized using a Stericup® filter unit (pore size 0.22 μm, Millipore, Espoo, Finland) later the same day. They were then stored at −20°C until used.

UPEC strains

Twenty consecutive clinical Escherichia coli strains isolated from the urine of young adult women (20–40 years of age) in the Microbiology Laboratory of Oulu University Hospital were used for the laboratory studies. Seventeen of these were UPEC strains, as they were obtained during an acute cystitis episode with typical clinical symptoms, pyuria and the growth of a single pathogen to more than 105 CFU /ml, and the other three were asymptomatic bacteriuria strains, as they were isolated from control urine samples obtained from patients without clear acute symptoms. Seven strains had been obtained from patients with recurrent UTIs within the last 3 months. Sixteen strains were susceptible to all the antimicrobials tested. ATCC 35218 was chosen as a reference strain as it had previously been used in a study showing a decrease in visible p-fimbriae after growth in media spiked with cranberry juice or proanthocyanidins (PACs) [11].

For the biofilm assay we had 20 clinical strains, each of which was tested by using urine pairs collected from all 20 healthy volunteers during the control period and after CLJ consumption. As one urine sample gave unexpectedly high optical density values (>1.0) in negative control wells without inoculated bacteria during the experiments, the pair concerned was not included in the final data analyses. For the gene expression studies we used five clinical strains (1–5 in Table 1) and ATCC 35218, tested using urine pairs from six volunteer subjects.

Table 1 Primer combinations used in the QPCR analyses
Full size table

Biofilm formation assay

The UPEC strains were revived on blood agar and grown overnight in Luria broth at 37°C. Their ability to form biofilm was tested by simultaneous subculture of 5 μl of an overnight Luria broth in 200 μl of urine obtained after juice consumption and 5 μl in 200 μl of control urine in triplicate on 96-well flat-bottom polystyrene microtitre plates (Nunc®). The bacteria were incubated for 72 hours under anaerobic conditions at 37°C, after which the urine was discarded and the wells were stained with 50 μl of 0.4% (w/v) crystal violet (FF Chemicals, Haukipudas, Finland) for 15 minutes, rinsed twice with sterile water and, after the addition of 200 μl of 96% ethanol to each well, measured for optical density (OD) at 570 nm (Multiscan MCC/340P). The negative control wells without inoculated bacteria were handled otherwise in a similar manner for each urine sample. The mean optical density value of the negative control wells was 0.414 (standard deviation [SD] 0.169). The OD measurements for assessing biofilm formation by the UPEC strains had been validated in our previous study by comparing them with the imaging findings obtained by SEM (scanning electron microscopy) and CSLM (confocal laser scanning microscopy) with LIVE/DEAD Baclight® staining [10].

Gene expression studies

The genes selected for the expression studies were pga (biofilm adhesion poly-β-1,6-N-Acetyl-D-glucosamine) and cpxA (sensor for a two-component signalling system that responds to envelope stress), as they are involved in the biofilm formation and envelope stress associated with Escherichia coli, and fimA (type-1 fimbrial protein A) and papF (minor pilin subunit papF), as they are involved in the uropathogenicity of Escherichia coli [12–15].

The E. coli samples were dissolved in 100 μl of RNAprotect Bacteria Reagent (Qiagen, Valencia, CA, USA), and total RNA was isolated with the RNeasy Mini kit (Qiagen). The quantity and quality of the isolated RNA was verified by measuring its absorbance spectrum with a NanoDrop N-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). The RNA samples were reverse transcribed to cDNA with a DyNAmo cDNA synthesis kit (Finnzymes Oy, Espoo, Finland). The sample cDNAs were purified from genomic DNA using the method described by Jaakola et al. [16].

Transcript accumulation of E. coli biofilm formation-related genes was detected using a LightCycler® SYBR Green I Master qPCR kit (Roche Molecular Biochemicals, Germany), and quantitative real-time PCR analyses were performed with a LightCycler 480 instrument and software (Roche Molecular Biochemicals, Germany). The PCR conditions were as follows: initial denaturation at 95°C for 10 min, followed by 45 cycles at 95°C for 10 sec (ramp rate 4.4°C/sec), 60°C for 20 sec (ramp rate 2.2°C/sec) and 72°C for 10 sec (ramp rate 4.4°C/sec). The melting curve was measured at 95°C for 0.5 sec (ramp rate 4.4°C/sec), 57°C for 15 sec (ramp rate 2.2°C/sec) and 98°C for 0 sec (ramp rate 0.11°C/sec). The primers presented in Table 1 were used for amplifying the pga, cpxA , fimA and papF fragments. The PCR products were quantified using a 16S rRNA primer pair as the control gene [17]. The results were calculated with the LightCycler®480 (Roche) software (release 1.5.0 SP1), using the calibrator-normalized PCR efficiency-corrected method (technical Note No. LC 13/2001, Roche Applied Science). The specificities of the PCR products were verified by melting curve analysis.

Statistical methods

The differences in optical density values between the control and cranberry consumption urine samples were first analyzed by a two-way analysis of the variance in density values between the strains. As there were statistically significant differences in the cranberry effect between the UPEC strains in a two-way analysis of variance, the results were not summarized statistically and the strains were analyzed by comparing the results obtained for the control and cranberry consumption urine samples from the 19 healthy subjects using a paired t-test for each UPEC strain separately. The gene expression levels were analysed by comparing the values for the control and cranberry consumption urine samples from the six healthy subjects using a non-parametric test for related samples (Wilcoxon signed rank test) separately for each UPEC strain in the biofilm and then in planktonic bacteria. Logarithmic transformation of the absolute values was used to visualize the changes in levels of gene expression. The statistical analyses were performed using the PASW Statistics 18 software.

Results

The clinical strains varied in their ability to produce biofilm in human urine on microtitre plates. Two strains (Fig. 1a) had high optical density values throughout the experiments. There were statistically significant differences in the cranberry effect between the UPEC strains (P-value 0.045 in two-way analysis of variance for differences between strains). Four of the 20 UPEC strains produced more biofilm in the urine collected after cranberry-lingonberry consumption than in the control urine, and one produced less (Fig. 1b). Consumption of CLJ did not affect biofilm formation by the other strains studied (Fig. 1b).

Fig. 1
figure 1

a Biofilm formation measured by mean optical density values on polystyrene plates of 17 clinical Escherichia coli strains isolated from patients with cystitis and three strains from subjects with asymptomatic bacteruria (ABU) in the urine of 19 healthy volunteers after the control period and after daily consumption of cranberry-lingonberry juice. The error bar indicates standard deviation (SD). b The mean difference of optical density values in paired urine samples. The error bar indicates 95% confidence interval. Paired t-test was used to compare the difference for each strain. An asterisk indicates the statistically significant difference (P < 0.05)

Full size image

The expression levels of the pga and cpxA genes, those related to biofilm formation by Escherichia coli, did not differ between the control and CLJ consumption urine samples when tested for both bacteria residing in biofilm and planktonic bacteria (P > 0.05 for all comparisons, except for P = 0.04 for the pga gene, expression of which was reduced in one strain, data not shown). The expression levels of the fimA and papF genes, those involved in the uropathogenicity of Escherichia coli, varied among volunteers within each strain but did not differ between the control and CLJ consumption urine samples in any of the strains tested for both bacteria residing in biofilm and planktonic bacteria (P > 0.05 for all comparisons, Fig. 2a and b).

Fig. 2
figure 2figure 2

Calibrator-normalized fim A (a) and pap F (b) gene expression levels after 72 hours of incubation in control urine or urine passed after drinking CLJ from six healthy women. Graphs A–E represent five clinical UPEC strains and graph F represents the ATCC 35218 strain. Each pair of dots indicates log-transformed data for one person. Data for bacteria in biofilm and planktonic bacteria are combined in the same graph for each strain. Statistical comparisons were performed between the control and after juice samples with respect to biofilm and planktonic bacteria for each strain separately. None of the differences was statistically significant (P > 0.05)

Full size image

Discussion

The consumption of cranberry juice and other cranberry products has been shown to increase the anti-adhesive properties of human urine in several studies (Table 3) [6, 18–26]. In the majority of cases it has been demonstrated that the adherence of uropathogenic Escherichia coli, and particularly its P-fimbriated strains, to uroepithelial cells, human red blood cells or other epithelial cells decreases. Cranberry has been shown to reduce biofilm formation by oral streptococci, but only a few small studies have been carried out on the effect of cranberry consumption on biofilm formation in urine [22, 24, 27]. The CLJ that we used here was manufactured by the same company and administered in a similar daily dose to that used in our previous randomized clinical trial, which showed the efficacy of cranberry juice in reducing the absolute risk of recurrent UTIs in healthy young women by 20% [3], and we now found that biofilm formation in the urine was increased in four of the 20 clinical Escherichia coli strains and reduced in one strain after CLJ consumption.

It has been suggested that cranberry products can reduce the virulence of Escherichia coli, since the number of visible P-pili and the expression of genes encoding virulence factors is lower in growth media spiked with PACs or cranberry juice [11]. In addition, Escherichia coli bacteria grown in urine obtained after cranberry consumption seemed to cause less severe disease in a worm infection model [25]. We could not, however, observe any decrease in the expression of the fimA gene, which is the major subunit of the type 1 pilus rod, or in that of the papF gene, which is a distal protein of P-pilus fibre in human urine.

Both cranberries and lingonberries contain high levels of phenolic compounds, specifically rare A-type proanthocyanidins [28, 29]. Growth medium or urine has been spiked directly with juice or PACs in several earlier experiments, since it has been suggested that A-type PACs may explain the anti-adherence activity of cranberry juice [30–38]. Most studies, however, have failed to confirm the presence of A-type proanthocyanidins in human urine after the consumption of cranberries [6, 18–23, 25, 26]. PACs from cranberries are generally poorly absorbed from the gut per se and are rarely found in human urine [24, 39]. On the other hand, the concentration of flavonoid metabolites such as hippuric acid has been shown to increase in urine after cranberry consumption, as is also observed after drinking black or green tea [24, 40, 41]. Accordingly, many Japanese volunteers already had anti-adhesive properties in their urine before any cranberry consumption [26]. Our negative findings in human urine in this study in spite of a positive finding in our earlier clinical trial after using a similar small daily dose of CLJ could easily be explained if CLJ already alters the properties of UPEC strains in the gut. The anti-adhesive effects in urine reported in several earlier studies (Table 2) might partly be caused by unspecific metabolites of cranberry such as hippuric acid remaining after large doses of cranberry products, whereas our small, yet clinically efficacious dose might not be sufficient to produce such anti-adhesive effects in the urine.

Table 2 Effect of cranberry consumption on the growth, adherence and biofilm formation of Escherichia coli in human urine
Full size table

In the pathogenesis of UTI the interaction between Escherichia coli and the host cell is mediated by adhesins that are incorporated into extracellular fibres called pili or fimbriae [42, 43], and it has been suggested that the anti-adhesive properties of cranberry may be mediated in both a non-specific and a specific way: by fructose, which inhibits type 1 pili-mediated adherence; and by PACs, which inhibit P pili-mediated adherence [44]. We did not select the UPEC strains to be studied here on the basis of specific virulence factors, as we wanted to have a clinical sample of UPEC strains from young women comparable to that used in our earlier trial.

Biofilm formation by Escherichia coli in vitro is prone to variation deriving from different strains, growth media and methods [45]. Even though the virulence of UPEC strains seems to be related to their ability to form biofilm under selected circumstances [9, 10], the strains causing asymptomatic bacteruria are more capable of producing biofilm than UPEC strains when human urine is used for the experiments [46]. We chose to use urine as the growth medium in our experiments as it has been suggested in many reports that it may have anti-adhesive properties after cranberry consumption (Table 2).

In conclusion, given a comparable product and similar daily doses to those used in our previous clinical trial, which showed the efficacy of cranberry for preventing UTIs, we were able to demonstrate increased biofilm formation of clinical UPEC strains in urine in vitro after consumption of CLJ in the case of 20% of the strains. Expression of the fimA and papF genes of the UPEC strains in urine did not decrease after CLJ consumption. It thus appears that the effect of CLJ in preventing UTIs cannot be explained by mechanisms that reduce biofilm formation on abiotic surfaces or the uropathogenicity of UPEC strains in human urine after the consumption of CLJ.