Sex effects in pyelonephritis

Abstract

Urinary tract infections are generally considered a disease of women. However, UTIs affect females throughout the lifespan, and certain male populations (including infants and elderly men) are also susceptible. Epidemiologically, pyelonephritis is more common in women but carries increased morbidity when it does occur in men. Among children, high-grade vesicoureteral reflux is a primary risk factor for upper-tract UTI in both sexes. However, among young infants with UTI, girls are outnumbered by boys; risk factors include posterior urethral valves and lack of circumcision. Recent advances in mouse models of UTI reveal sex differences in innate responses to UTI, which vary somewhat depending on the system used. Moreover, male mice and androgenized female mice suffer worse outcomes of experimental pyelonephritis; evidence suggests that androgen exposure may suppress innate control of infection in the urinary tract, but might have additional effects yet to be specified. These recent findings raise the hypothesis that the postnatal testosterone surge that occurs in male infants may represent an additional factor driving the higher incidence of UTI in males under 6 months of age. Ongoing work in contemporary models will further illuminate sex- and sex-hormone-specific effects on UTI pathogenesis and immune responses.

EPIDEMIOLOGY OF URINARY TRACT INFECTION

Urinary tract infections (UTIs) are among the most common bacterial infections, annually affecting 150 million people across the globe [1, 2]. Although both males and females may become infected, UTIs are traditionally considered a disease of women, among whom 50% will be affected across their lifespan [3]. In addition, approximately 25% of women presenting with a first episode of bacterial cystitis (infection of the urinary bladder) go on to suffer recurrent UTI (rUTI) within 6 months, some having 6 or more infections in the year following the initial episode [3]. Anatomic features protect otherwise healthy males at most ages and stages of life; however, males at both ends of the lifespan exhibit an increased incidence of UTI (Figure 1). Specifically, in infants < 6 months of age, boys with UTI outnumber girls [4], which may relate to circulating androgens (discussed below); meanwhile, UTIs in elderly men typically reflect urodynamic dysfunction owing to prostatic hypertrophy. In addition, uncircumcised males clearly have a higher risk for UTI than their circumcised counterparts [5–7]. In addition, various chronic conditions in males (diabetes, spinal cord injury, indwelling or intermittent bladder catheterization) also promote UTI [1]. When pyelonephritis (infection of the kidney parenchyma) does occur in males, it is associated with greater morbidity and mortality than in females [1], suggesting that differences beyond simple anatomy may influence the outcomes of these more severe infections.

Fig. 1.

Schematic representation of age-related incidence of UTI in males (blue) and females (red). In the first six months of life, boys suffer more UTIs than girls (see inset); this relationship has reversed by 12 months of age, after which females dominate UTI incidence over much of the lifespan. Incidence of UTI rises in postmenopausal women and also in older men, where it is related to prostatic enlargement and associated urodynamic abnormalities

In children, suspected or confirmed UTIs are among the most common reasons for evaluation in emergency departments; between 3–7% of febrile young girls and 1–3% of febrile young boys will prove to have UTI [8]. The majority of community-acquired UTIs are caused by uropathogenic strains of Escherichia coli (UPEC), while other bacterial species (e.g., Klebsiella, Proteus, Pseudomonas, Staphylococcus) are identified less frequently in urine cultures. In childhood, ~7% of febrile UTI result in formation of renal scars [9], which are correlated with risk for hypertension and end-stage renal disease (ESRD) later in life. Of note, 30 million US adults are estimated to have chronic kidney disease (CKD) [10]; although men are less likely to develop CKD overall, men with CKD are 50% more likely than women to develop ESRD [11].

DIAGNOSIS AND TREATMENT OF PEDIATRIC UTI

The diagnosis of UTI relies on biochemical and microscopic analysis of the urine, as well as results of urine cultures that are properly obtained (i.e., by sterile catheterization or suprapubic aspiration). The performance characteristics of these test modalities will not be reviewed here, but form the basis of current guidelines for diagnosis of UTI in children [12]. Sex per se does not strongly factor into making the diagnosis of UTI, beyond influencing a provider’s suspicion for UTI (due to epidemiologic factors, such as the higher risk for UTI in uncircumcised boys) and the logistics of urine collection. In older children and adolescents with UTI, involvement of the kidney can be signified by fever, flank pain, nausea, and vomiting [12]. In infants and preverbal children, it can be difficult to definitively assign the specific diagnosis of pyelonephritis; fever is often the presenting complaint and the reason for providers to order urinalysis and urine culture, while other symptoms can be nonspecific or absent.

As with a broad range of other medical conditions, contemporary research in UTI includes efforts to identify more sensitive and specific biomarkers, in either serum or urine. For example, a recent examination of 1305 proteins in serum and urine revealed evidence of sex-specific proteome changes in UTI patients [13]. If such studies produce candidate biomarkers for diagnosis or prognosis (e.g., risk for recurrent UTI or renal scarring), it will be important to examine the sex-specific performance and utility of these biomarkers. Additionally, biomarkers in the urine may help to distinguish lower- from upper-tract UTI; a recent study showed increased urinary levels of chemokine (C-X-C motif) ligand (CXCL)1, CXCL9, CXCL12, C-C motif chemokine ligand 2 (CCL2), interferon-γ, and interleukin (IL)-15 in children with pyelonephritis vs those with cystitis [14]. Overall, improvements in this arena may result in reduced empiric antibiotic therapy (often employed currently while awaiting urine culture results), more tailored antibiotic selection, and shorter time to definitive treatment for UTI. Importantly, early administration of effective antibiotic therapy in febrile UTI reduces risk of renal scarring [9].

Urinary tract imaging should be undertaken in UTI patients with complicated pyelonephritis, evident structural or urodynamic abnormalities, a history of UT instrumentation, or indwelling urinary catheters; in immunocompromised patients; and in any male patient with suspected pyelonephritis [15]. Ultrasonography is the most often employed modality, as it is widely available, noninvasive, relatively inexpensive, and requires no contrast administration. Magnetic resonance imaging (MRI) is a powerful tool to detect inflammation in the kidney and to assess renal vasculature [16]. Given its high sensitivity, MRI can increase clinician confidence in a negative diagnosis [17]. In selected circumstances, other imaging techniques may provide better information (e.g., computed tomography [CT] scan or intravenous pyelography in stone disease). In infants with a first febrile UTI, renal and bladder ultrasonography should be performed. There is some controversy about the utility of voiding cystourethrography (VCUG) in this patient group; current American Academy of Pediatrics practice guidelines [12] recommend that this be undertaken only if the ultrasound is abnormal, while many urologists feel that VCUG is indicated in first febrile UTI regardless of ultrasound findings. There is agreement that a second febrile UTI in an infant is a definite reason for VCUG to be performed.

As noted above, renal scars occur in a significant minority of pediatric febrile UTI; scars are not readily detected by ultrasonography and are difficult to distinguish from inflammatory changes during the acute phase of pyelonephritis. The most sensitive and specific modality for identifying and quantifying renal scar formation is ^99mTc-dimercaptosuccinic acid (DMSA) scanning. Indeed, use of this technique has provided the best data on scarring following pediatric pyelonephritis in multiple clinical study cohorts [18, 19]. DMSA scans are often abnormal in acute pyelonephritis [20–23]; however, renal ultrasound is much more readily performed in the acute phase of illness. Moreover, though detection of scar in the individual patient might help the clinician to be ‘on notice’ for later-life complications, there is little specific intervention to be undertaken; as a result, DMSA scanning is currently less useful in the clinical arena. It is hoped that future work will identify therapeutic or preventive avenues to mitigate scarring and the associated long-term risks to renal health.

GENETIC AND ANATOMIC RISKS IN PYELONEPHRITIS

In recent years, researchers have identified a number of genetic factors that increase risk for pyelonephritis, most of which relate to mediators of innate defense in the urinary tract. Toll-like receptor 4 (TLR4, which recognizes bacterial lipopolysaccharide) and CXCR1 (the IL-8 receptor) are two important receptors involved in the response to introduced pathogens. TLR4 signaling induces uroepithelial cells of the kidney and bladder to secrete cytokines, chemokines and antimicrobial peptides [24–28]. Single-nucleotide polymorphisms that cause missense variants in TLR4 may increase the risk of developing UTI [29]. Additional TLR family members may also play a role in innate defense of UTI; specifically, TLR5 and TLR1 variants have been shown to alter risk for recurrent UTI and pyelonephritis [29]. After TLR signaling has been initiated, CXCR1 detects the initial IL-8 (CXCL1) signal used to recruit neutrophils and other immune cells in the early host response to infection. Polymorphisms and expression changes in CXCR1 are also associated with acute pyelonephritis and renal scarring in certain patient populations [30]. These human variants and phenotypes highlight the importance of a rapid and potent innate response to UPEC introduction into the urinary tract.

Vesicoureteral reflux

In pediatric populations, urodevelopmental abnormalities like vesicoureteral reflux (VUR; retrograde flow of urine from the bladder toward the kidneys) and posterior urethral valves (PUV) substantially increase risk for pyelonephritis. Indeed, VUR represents the primary risk factor for febrile UTI in children [31]. One study identified VUR (by VCUG) in 0.8% of neonates, with an incidence ratio of 2:1 in males vs females [32]. VUR is a partially heritable condition, as one third of siblings of children with demonstrated VUR will also have the condition [33]. However, the genetic lesions responsible for nonsyndromic VUR development remain incompletely defined [33–35]. Between 30–40% of children presenting with febrile UTI are found to have VUR [1, 12, 36]. Moreover, 68% of patients with UTI and VUR will have detectable scarring later in life [37]. The anatomic predisposition to UTI associated with higher grades of VUR may be compounded by other genetic variation. For instance, DEFA1A3 (encoding α-defensins that exhibit antimicrobial activity against many uropathogens) is present at a lower copy number in children with VUR and UTI compared with healthy controls. Children with VUR who do have higher DEFA1A3 copy numbers are relatively protected from breakthrough UTI [38]. Meanwhile, a study in Iran reported an association between certain IL-10 and IL-12 alleles and the presence of VUR [39].

Posterior urethral valves

PUV occur only in male infants (1 per 5,000–8,000 live male births [40]) and result from the formation of a thickened membrane (arising from Wolffian duct tissue) that obstructs the urethra. A recent analysis of Pediatric Health Information Systems (PHIS) data reported a 5% risk of death in PUV patients, though the strongest predictor of mortality was pulmonary hypoplasia (not renal dysplasia) [41]. A recent single-center retrospective study of 64 PUV patients (with a median follow-up of 70 months) found that 22% experienced recurrent UTI, while 35% had CKD at various stages [42]. The urologic injury associated with PUV may begin in fetal life, as evidenced by a study of 24 fetuses with antenatal ultrasound evidence of PUV who underwent sampling of bladder urine at 22 ± 4 weeks of gestation. Compared to controls, bladder urine from PUV fetuses showed an intense inflammatory profile, with significantly increased urinary levels of IL-2, IL-4, IL-6, TNFα, IFN-γ, MCP-1 (CCL2), and other mediators. On the basis of these data, the authors posited a pathophysiologic role in PUV for local inflammation just after embryological formation of the urethral membrane [43].

SEX INFLUENCES ON UTI PATHOGENESIS

Sex differences influence the initiation of immune defenses, with females showing enhanced responses to a wide array of pathogenic microbes and vaccine antigens (reviewed recently in [44]. The incidence of bacterial infection and sepsis is higher in men [45, 46], while women are more likely to develop autoimmune diseases and exhibit a lower risk of mortality in sepsis [47, 48]. Some have speculated that this difference has evolved in relation to the female’s needs for tolerating fetal antigens while offering heightened innate protection for the developing fetus against infectious agents. In a variety of infectious disease models, the amplitude of innate responses relates directly to the severity of infection [12, 49], and sex and sex hormones influence the behavior of innate leukocytes, though results vary across cell types and experimental systems (see [50]. Resident peritoneal macrophages in female mice were found to express more TLR4 and exhibit more efficient phagocytosis and bacterial killing, while also not excessively secreting pro-inflammatory cytokines [51]. Accordingly, androgen treatment of macrophages reduces TLR4 expression [52], and immortalized macrophage-like cell lines produce less TNF-α and inducible NO while producing more IL-10, consistent with testosterone having a net anti-inflammatory effect [53]. By some contrast, in a murine atherosclerosis model, androgen receptor signaling in macrophages upregulated TNFα and other molecules involved in inflammation-related atherosclerosis [54]. In addition, a recent study of peripheral blood monocytes from human volunteers studied ex vivo reported that cells from men exhibited higher pro-inflammatory cytokine secretion upon LPS stimulation [55]. In short, further work is needed across an array of disease-specific in vitro and in vivo models and human populations to further reveal the complicated interactions between sex, sex hormones, and inflammatory responses.

Modeling sex differences in UTI

The molecular pathogenesis of bacterial cystitis has been extensively studied and defined using mouse models and correlation with human samples; meanwhile, comparatively less is known about the molecular pathogenesis of infection in the kidney. In traditional mouse models, severe kidney infection (including renal abscess formation) is uncommon, hampering the study of this entity. Attenuation in mouse kidney infection has been reported with UPEC mutants lacking specific virulence factors, such as type 1 pili, P pili, flagella, α-hemolysin, and cytotoxic necrotizing factor 1 (CNF1) [56–60].

Historically, nearly all cystitis and pyelonephritis studies have been performed in female mice, as the male mouse bladder is not easily accessed by catheter. Of note, instillation of uropathogens into the urethra of male mice elicits prostatic infection [61–64]. In a recently reported new model of UTI, a small abdominal incision permits inoculation of the bladder by needle under direct visualization in both mouse sexes [64]. This inoculation method recapitulates many features of experimental UTI established in studies with catheter-infected females. Interestingly, once anatomic barriers were bypassed in this way, male mice experienced more severe infection than females, mirroring epidemiologic data observed clinically in men. Specifically, male C3H mice uniformly developed severe pyelonephritis and renal abscesses (Figure 2) that are seen much less frequently in female mice [64]. More recently, Ingersoll and colleagues reported studies of sex differences in bladder innate responses after catheter inoculation of both males and females (see below) [65, 66]. These new models open doors to study sex differences in UTI pathogenesis and host response, as well as sequelae of severe pyelonephritis and abscess formation; these latter phenotypes are relevant to febrile UTI in children, following which renal scarring is a common complication.

Fig. 2.

Photomicrograph exemplifying the severe pyelonephritis observed in C3H male mice, 14 days post inoculation of the bladder with UPEC. Findings include tubular necrosis, extensive inflammation, and micro- and gross abscess formation (hematoxylin and eosin staining; scale bar, 200 μm)

Estrogens and UTI

The role of estrogens in susceptibility to, or protection from, UTI is a matter of ongoing debate, with potentially opposing activities demonstrated in a variety of studies. Luthje et al. found that estradiol facilitated bacterial invasion into the bladder epithelium in mice. Of note, this same study reported that 75% of tested postmenopausal women, when supplemented with estradiol, exhibited enhanced expression of antimicrobial peptides such as human beta-defensin 1 (HBD-1), HBD-2, HBD-3, psoriasin, and RNase 7 [67]. In another study, HBD-2 production by cultured vaginal epithelial cells was amplified by estradiol and suppressed by progesterone [68]. Antimicrobial peptides whose expression is altered by these exposures, including systemic and local hormone therapies, may influence UTI risk in a variety of ways (e.g., by changing the vaginal microbiota, halting perineal transit of uropathogens, or changing the inflammatory milieu within the UT itself) [64, 69]. Mysorekar et al. mimicked menopause with ovariectomy in adult C57BL/6 female mice and found that pro-inflammatory responses to UPEC cystitis were heightened in comparison to sham-operated females; estrogen supplementation blunted these differences [70]. The influence of estrogens on immune responses to pyelonephritis is less well studied; a recent paper found that an estrogen-receptor agonist modestly reduced kidney bacterial loads in mice given UTI [71]. Our work in the mini-surgical model of pyelonephritis did not reveal any change in renal phenotype with ovariectomy in C3H females [72].

Clinical studies of the effects of estrogen on UTI risk have yielded mixed results; a 2008 Cochrane review concluded that estrogens may have a beneficial effect in reducing recurrent UTI in postmenopausal women, depending on the form and route of estrogen administration [73]. In a frequently cited randomized controlled study of 93 postmenopausal women with rUTI, intravaginal estriol supplementation lowered vaginal pH, reduced vaginal colonization with Enterobacteriaceae while promoting lactobacilli, and cut rUTI rates by 10-fold [74]. A more recent study of estrogens delivered by contemporary methods (vaginal ring or cream) in 35 postmenopausal rUTI patients confirmed a lower risk of rUTI with treatment [75]. Overall, studies of the mechanisms by which estrogen supplementation might be helpful in this population have focused on alterations in the vaginal microenvironment (e.g., moisture, pH, and microbiota) that accompany menopause. In early work, ERT lowered vaginal pH and discouraged vaginal colonization with undesired anaerobes [63]. A more recent study showed that postmenopausal women have a much lower prevalence of vaginal lactobacilli (considered a main component of a ‘healthy’ vaginal microbiota) than premenopausal women. Estrogen replacement therapy restored the ability of postmenopausal patients to harbor favorable lactobacilli (in 80% of ERT patients compared with 40% in those not receiving ERT) [76]. Other mechanisms by which estrogens might influence UTI risk (e.g., direct effects on innate immunity) have not been thoroughly evaluated in human studies.

Androgens and UTI

As described above, when anatomic protections of males are bypassed in mice or humans, severity of pyelonephritis is higher than in females [64]. Additionally, women with polycystic ovary syndrome (PCOS; a common hyperandrogenic state) exhibit increased UTI incidence, which may be normalized with antiandrogen therapy [77]. In the infant population, the higher incidence of UTI in boys < 6 months of age (compared with girls) has, in the absence of PUV or other obvious cause, traditionally been attributed nonspecifically to ‘urodynamic immaturity’ in male infants. However, the postnatal testosterone surge in male infants peaks within the first month and resolves by six months of age [78, 79], mirroring the shape and duration of their UTI incidence curve (see Figure 1 inset). This observation provokes the hypothesis that testosterone exposure contributes to UTI risk in this population.

In our work, castration of male C3H mice conferred protection against the most severe outcomes of experimental UTI, namely chronic cystitis, persistent high-titer pyelonephritis, and renal abscess. Susceptibility to these severe outcomes was restored to castrated males by provision of testosterone or its derivatives; similar treatment of C3H females induced susceptibility to these poor outcomes [64, 80]. In these experiments, male C3H mice (vs females) exhibited only minor differences in bladder and kidney pro-inflammatory cytokine content early in infection. However, Ingersoll and colleagues observed increased neutrophil infiltration and pro-inflammatory cytokines in the bladder of infected C57BL/6 females, compared with similarly infected males. Specifically, female mice produced more IL-17, which predicted effective bacterial clearance and ultimate resolution of infection [66]. Separately, in a mouse model of prostatitis, testosterone exposure resulted in more intense neutrophil recruitment and associated tissue inflammation; however, the recruited neutrophils exhibited less robust microbicidal activity [81], reminiscent of poorly effective “N2-like” neutrophils found in tumor microenvironments [82]. In total, these results suggest that androgens suppress various features of the innate response to infection in the UT, enabling increased severity in male UTI. Ongoing work by multiple groups aims to further illuminate the cellular and molecular mechanisms of androgen influence on immunity (not only control of infection, but also associated tissue damage and predisposition to recurrence) in both the bladder and kidneys.

Finally, androgens are known to promote renal scarring and fibrosis in models of noninfectious renal injury (reviewed recently in [83]). Indeed, male mice have been predominantly utilized in these models because of their more evident scarring phenotypes. Androgen exposure also augments the bladder fibrosis that accompanies experimental bladder outlet obstruction [84]. With the advent of the newer mouse models described herein, the field is poised to dissect the potential influence of androgens on the renal scarring that can follow pyelonephritis.

OUTLOOK

The success of future research in infectious diseases, including cystitis and pyelonephritis, relies heavily on including sex differences in both fundamental and clinical research. With multiple new mouse models of UTI available, the prospect of understanding sex-specific UTI pathogenesis and host responses at a molecular level is very bright. Novel approaches to recurrent and complicated UTI will be more critically needed as the management of contemporary UTI is increasingly challenged by widespread and multi-class antibiotic resistance among uropathogenic bacteria [85, 86]. For example, multiple groups have begun to target the adhesive functions of UPEC with small-molecule competitive inhibitors (d-mannose and rationally designed mannosides that bind type 1 pili with orders of magnitude higher affinity than mannose itself [68, 87–89] and with FimH-based vaccine candidates [90]. As more is learned about sex-specific effects on UTIs and their sequelae, one can imagine developing additional interventions to stimulate specific host immunity (e.g., to eradicate chronically resident bacteria from the UT), modulate testosterone signaling (not only in PCOS but in phenotypically normal women with testosterone levels in the upper range of normal), and mitigate scarring to preserve long-term renal function in patients with recurrent UTI.