Pathogenesis of gestational urinary tract infection: urinary obstruction versus immune adaptation and microbial virulence
Dr Bogdan Nowicki, Departments of Obstetrics and Gynecology and Microbiology and Immunology, Meharry Medical College, Nashville, TN 37208, USA. Email email@example.com
Please cite this paper as: Nowicki B, Sledzinska A, Samet A, Nowicki S. Pathogenesis of gestational urinary tract infection: urinary obstruction versus immune adaptation and microbial virulence. BJOG 2011;118:109–112.
The mechanism of urinary tract infections (UTIs) in non-pregnant patients involves Escherichia coli adherence to epithelial receptors, with ascending colonisation of the urogenital tract. To date, UTI in pregnant patients has been explained by obstruction of urine outflow by the gravid uterus.1 This relatively simple interpretation does not consider the receptor etiology of the ascending UTI and/or gestational adaptations of the immune system.
Gestational pyelonephritis is associated with multiple complications, such as fetal growth restriction, preterm labour, cerebral palsy, septicaemia and maternal death, and is characterised by microbial invasion of renal interstitial tissue. We have previously reported that gestational pyelonephritis is frequently caused by E. coli bearing adhesions of the Dr family (Dr E. coli), an antibiotic-resistant microorganism capable of invading renal interstitial tissue, escaping leukocytes and persisting in epithelial cells.2,3
Although reported as dogma in many textbooks, both the biological complexity and molecular epidemiology of UTI suggest that obstruction/stasis is not the only factor leading to the risk of developing gestational UTI.2 Urine stasis and obstruction should favour infection in the presence of avirulent random gram-negative and gram-positive species. Instead, 90% of gestational UTIs are associated with uropathogenic virulent E. coli strains, which (in contrast) are seldom isolated from patients with urinary tract obstructive problems. Furthermore, we previously reported data on temporary patterns showing a trimester-specific predominance of E. coli serotypes, E. coli DNA fingerprints and adherence/colonisation factors.1,3 All together, the data appear inconsistent with the interpretation that obstruction/urine stasis is the main cause of UTI in pregnant women. Furthermore, in recent years there has been an accumulation of research data on the altered immune responsiveness in pregnant animals with experimental UTI and/or with uterine infections.4,5 Here, we present an update on those immune/risk factors in pregnant patients with UTI, and further challenge the dogma that obstruction/stasis is the main cause of gestational UTI.
Epidemiology of UTI
It is intriguing that textbooks indicate that pregnant patients have an increased propensity for UTI given the fact that the frequency of UTI in pregnancy is similar to that seen in non-pregnant women.2,4 Epidemiological studies of UTIs indicate that the frequency of asymptomatic bacteriuria (ABU) is 4–10%, cystitis 1–4%, and pyelonephritis 1–2%, and that these rates are similar or identical in pregnant and non-pregnant women.2 The question therefore remains, how did the claim of an increased rate of UTI in pregnant women become established? Presumably this resulted from a combination of clinical observations and epidemiological studies on ABU. It is well recognised that in the absence of treatment 20–40% of pregnant women with ABU develop pyelonephritis.1 ABU is also associated with low birthweight and preterm labour.1 Therefore, although untreated ABU represents a higher risk in pregnancy, the question is why? Furthermore, is this the result of urinary obstruction/stasis, other factors such as an immune adaptation in pregnancy, which is required for the accommodation of the foreign antigen of the fetus, and/or the presence of gestation- specific virulent strains of E. coli?
Approximately 90% of pregnant woman develop ureter dilatation, which persists until delivery.2,6 This process begins in week 6 of pregnancy, and peaks during weeks 22–24. The decrease in ureter tone is associated with an increase in urinary stasis and/or ureterovesicle reflux. Recent studies on dilatation of the collecting system during pregnancy with ultrasound technology interpreted the findings of ureteric dilatation in the context of physiologic versus obstructive changes.7 Furthermore, women with a previous (childhood) history of vesicoureteric reflux (VUR) do have a higher incidence of urinary tract infection during pregnancy. Renal scarring was the primary risk factor for morbidity during pregnancy, and this risk was independent of the presence or absence of VUR at the time of pregnancy.8 One could argue that there are two opposing implications associated with this process: (i) physiologic adaptations that contribute to this increased risk of UTI;9 and (ii) physiologic adaptations that contribute to the well-being of both the fetoplacental unit and the mother.7,8 As an example of the latter, in contrast to ureters in non-pregnant women, which are more rigid and may kink as a result of obstruction (e.g. from the placement of sutures during surgery), perhaps pregnancy-dilated and flexible ureters allow for a more efficient urine outflow by avoiding the compression (kinking) caused by an enlarging uterus.1,7,8 Although both interpretations may be argued in the literature, the molecular bacteriological data do not strongly support the first implication, namely ‘increased obstruction risk’, as being the primary explanation for this phenomenon.2–4
Immune system in pregnancy/immune adaptations
It is well established that immune humoral mechanisms are increased in pregnancy, whereas cellular responses are down-regulated. These changes appear to protect the fetus.10 Although gestational immune adaptations are highly orchestrated processes and represent a significant component of our physiology, and probably result from adaptive or evolutionary selection processes, some of these adaptations may not always be perfect for mother and fetus. Some changes in the immune system may be exploited by adaptive pathogens.
Nitric oxide, pregnancy and infection
There are two key players of the immune system in pregnant women that alert the host to fight infection, namely nitric oxide (NO) and toll-like receptor 4 (TLR4). Among its multiple functions, a key role of NO is to provide signals to relax the uterus (to allow for a quiescent environment for the fetus), and to destroy or restrict invasive pathogens.10,11 For these reasons, the NO system functions at its highest level in the pregnant host. In other words, NO operates at a maximum or near-maximum capacity, and has a limited adaptive capacity to increase its activity under these conditions. Thus, if the host develops a benign infection of the urogenital tract, NO can most likely efficiently eliminate this less virulent pathogen. However, in the case of a severe organ-specific inflammatory process and a virulent/invasive pathogen such as Dr E. coli, NO may be far less effective.
TLR4 responsiveness and UTI
The second key player in infection processes is TLR4. TLR4 is part of a pattern recognition system, also called a ‘danger signaling’ system, that responds to endotoxin (lipopolysaccharide) and releases an inflammatory cascade of cytokines and prostaglandins, which contribute to the clearance of the infection. Unfortunately, in pregnant women, an aggressive inflammatory response may be detrimental to the fetus, and may lead to preterm labour. Hypothetically, an aggressive TLR4 response to an infection would be good: this would help resolve the infection. However, it would be contraindicated from the prospective of the foreign antigen of the fetus being rejected in the presence of a pro-inflammatory environment. The inflammatory TLR4 signalling is executed by monocytes and macrophages infiltrating the fetoplacental unit. Rather than finding a more aggressive TLR4 response, our preliminary study in pregnant patients suggested that the TLR4 response in pregnant mothers is decreased. When responsiveness to E. coli infection was tested in TLR4-negative mice, the results resembled the increased sensitivity to infection seen in experimental pregnant animals, suggesting that this reduced TLR4 adaptation in pregnancy may have an important impact on the risk for UTI.11
Gestational tissue tropism
Third-trimester pyelonephritis is dominated by Dr E. coli, an organism that is capable of invading renal interstitial tissue, escapes leukocytes, and persists in epithelial cells.2,3 The question is, why is Dr E. coli more prevalent in third trimester pyelonephritis? Based on previous studies,3 we postulated that tissue receptor CD55, also called DAF, is up-regulated by progesterone, a hormone that increases in concentration proportionally with gestational age. Progesterone increases the expression of CD55 (a complement regulator), which in the placental interface protects the fetus from maternal complement injury.2 Paradoxically, Dr E. coli, which recognises CD55, may gain an advantage in the colonisation and/or invasion of tissues, a process that is directly proportional to CD55 receptor density.2 One would assume that even in such an environment, the immune system should protect the maternal urinary tract, as there are multiple redundant systems that normally protect the host. The explanation may lie in the fact that in this environment (in which two key protective systems [NO and TLR4] have a decreased capacity to display their functions, and in which the third factor [CD55 receptor] is elevated), we observed a paradoxical increase in Dr E. coli, which in turn finds a niche of opportunity to establish a chronic and/or subclinical infection. As 90% of Dr E. coli are resistant to ampicillin, the question can be posed whether a resistant strain with the capacity to invade renal interstitium can be eliminated, or whether it will persist to cause a chronic/recurrent infection?
Chronic/recurrent UTI and Dr E. coli
We postulated that Dr E. coli can establish chronic infections, and presented experimental data that such a process is possible in TLR4-negative mice.12 However, we did not have evidence at the time that such a process occurred in humans. Interestingly, in the current issue we present the first case providing support for our hypothesis. During an autopsy we discovered that a 23-year-old pregnant woman had died from Dr E. coli septicaemia, which had spread from her kidneys. The examination of renal tissue showed the presence of chronic pyelitis and interstitial tubulonephritis, which resembled histopathology findings observed in a Dr E. coli model of chronic pyelonephritis in mice.12 Most interestingly, the chronic process in mice was directly dependent on the presence of Dr adhesins, and was prevented by an E. coli mutation in which expression of Dr adhesin was abolished. Infection with the same Dr E. coli strain in pregnant rats was lethal, and the mortality was increased by the inhibition of NO. This lethal impact was not present in non-pregnant rats. A mutation of Dr adhesin protected pregnant animals from death, and even at the highest bacterial concentrations, there was no mortality observed. Taken together, these findings support our hypothesis that Dr E. coli may have the capacity to display gestational renal tropism, and to establish a chronic infection in pregnant human subjects. Furthermore, such an infection may even be fatal.
We propose that in pregnant mothers in whom NO is operating at a maximum or near-maximum capacity, and therefore has limited capacity to increase in response to a spreading virulent infection, and in whom TLR4 is down-regulated, the simultaneous limited/diminished responsiveness of both NO and TLR4 plays a key role in a reduced gestational responsiveness to inflammation/infection. This may allow ABU to progress to UTI and urogenital infections, or even to fatal septicaemia. Extensive studies will be required to prove this model in the future. However, in the meantime, this path should be recognised as a possibility, along with the obstructive urinary changes promoted in the past decades.
Therefore, in an environment in which two key protective systems, NO and TLR4, have a limited or decreased capacity to display their functions, and in which the third factor (CD55 receptor) is elevated, we observe a paradoxical increase in Dr E. coli, which in turn finds a niche of opportunity to establish chronic and/or subclinical infections. In the context of immune adaptations, the infection process caused by Dr E. coli may appear uniquely tailored to the environment of pregnancy, with gestational tropism/virulence mediated by Dr adhesin. Therefore, the question posed in the introduction, of whether the adaptation of the immune system during pregnancy contributes to the risk of ABU/pyelonephritis, seems plausible, and deserves serious consideration.
If ABU/UTI/pyelonephritis affects a mother during gestation, this may have a lifelong impact on both the mother and the fetus. Until now aggressive screening and antibiotic therapy to prevent UTI has been our best option. However, a combination of Dr E. coli gestational tropism in the presence of NO/TLR4/CD55 adaptations and antibiotic resistance poses the question: should we apply a novel molecular approach to diagnose high-risk pregnancy-associated pathogen/virulence factors, and, in addition, specifically tailor antibiotic therapy standards to improve maternal and fetal safety?
Disclosure of interests
The authors have no conflicts of interest to disclose.
Contribution to authorship
All authors contributed equally.
Details of ethics approval
This review was supported in part by Public Health Service grant HD41687 from the National Institute of Child Health and Human Development to SN, and by DK42029 from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) to BJN.
We thank Dr Diana Marver for her editorial review.