Acute pyelonephritis in pregnancy and plasma syndecan-1: evidence of glycocalyx involvement

Abstract Background Acute pyelonephritis, a risk factor for maternal sepsis, adult respiratory distress syndrome, and preterm labor, is a frequent cause of hospitalization. This condition is characterized by excessive intravascular inflammation and endothelial cell activation and dysfunction. Syndecan-1, a major component of the glycocalyx, is a gel-like layer that covers the luminal surface of healthy endothelial cells, preserving and mediating many endothelial functions. During pregnancy, there is an additional potential source of syndecan-1, the “syncytiotrophoblast glycocalyx,” which lines the intervillous space. Insults that damage the glycocalyx lead to a shedding of syndecan-1 into the circulation. Hence, syndecan-1 has been proposed as a marker of endothelial injury in conditions such as sepsis, trauma, cardiovascular disease, and diabetes mellitus. Objective The objective of this study was to determine whether the plasma syndecan-1 concentration changes in women with acute pyelonephritis in the presence or absence of bacteremia. Study design This cross-sectional study included three groups: (1) non-pregnant women (n = 25); (2) women with an uncomplicated pregnancy from whom samples were collected preterm (n = 61) or at term (n = 69); and (3) pregnant women diagnosed with acute pyelonephritis from whom samples were collected at the time of diagnosis during the second and third trimesters (n = 33). The diagnosis of acute pyelonephritis was based on clinical findings and a positive urine culture for bacteria. Blood culture results were available in 85% (28/33) of women with acute pyelonephritis. Plasma concentrations of syndecan-1 were determined by a validated immunoassay. Results (1) Women with an uncomplicated pregnancy had a higher plasma concentration of syndecan-1 than non-pregnant women. The geometric mean (95% confidence interval [CI]) of syndecan-1 concentration was 51.0 (12.1–216.1) ng/mL in non-pregnant controls; 1280 (365–4487) ng/mL in normal preterm gestations; and 1786 (546–5834) ng/mL in normal term gestations (adjusted p < .005 for all three between group comparisons); (2) plasma syndecan-1 concentrations increased with gestational age among women with a normal pregnancy (p < .001, R 2 = 0.27); (3) syndecan-1 multiple of the mean (MoM) values in pregnant patients with acute pyelonephritis were higher than those in normal pregnant women based on second- and third-trimester samples (p = .048, 1.26-fold change). The increase was driven primarily by cases with a positive blood culture (p = .009, 1.74-fold change); (4) when data from third-trimester samples were compared, overall differences in syndecan-1 MoM values between cases and controls were slightly larger (p = .03, 1.36- fold change), which were especially contributed to by cases with a positive blood culture (p = .023, fold change 1.79-fold change). Conclusions Plasma syndecan-1 concentration is higher in pregnant women and increases as a function of gestational age. Patients with acute pyelonephritis have a higher plasma concentration of syndecan-1, and this is particularly the case in the presence of bacteremia.


Introduction
Acute pyelonephritis complicates approximately 2% of pregnancies [1] and is the most common cause of nonobstetrical severe sepsis and septic shock in pregnancy [2,3]. Pregnant women with acute pyelonephritis are more likely to develop acute respiratory distress syndrome than non-pregnant patients [4][5][6], and this finding has been attributed to priming of the innate immune system during pregnancy [7]. In the context of infection, such as acute pyelonephritis and sterile inflammatory processes [8], the state of intravascular inflammation that characterizes normal pregnancy [7][8][9][10] can be exacerbated. Intravascular inflammation can disrupt the thickness of the endothelial glycocalyx [11], a gel-like layer that coats the luminal surface of endothelial cells [12] and is composed of sulfated proteoglycans and nonsulfated hyaluronic acid [13]. Glycocalyx disruption can be assessed by determining the concentrations of its components in the maternal circulation, including syndecans [14,15]. Syndecans are transmembrane proteoglycans and major components of the endothelial glycocalyx [16,17], where they play a role in multiple physiologic processes including angiogenesis [18], inflammation [19], mechanotransduction [20], and regulation of leukocyte migration [21,22]. Four syndecan molecules are present within the mammalian endothelial glycocalyx [16]. Syndecan-1 is the prototype member of the family [23], and its functions include inhibition of leukocyte adhesion, facilitation of trans-endothelial and trans-epithelial migration of leukocytes, and removal of chemokine gradients [24].
Plasma syndecan-1 is considered a marker of endothelial glycocalyx damage, and high concentrations of plasma syndecan-1 have been reported in non-pregnant individuals with a condition characterized by intravascular inflammation, such as sepsis [25,26], acute coronary syndrome [27], diabetes mellitus [28], and nephrotic syndrome [29]. Moreover, the plasma concentration of syndecan-1 appears to be related to the severity of disease in patients with septic shock [14]. Whether maternal infection changes the endothelial glycocalyx and concentrations of syndecan-1 is not known. Therefore, the purpose of this study was to characterize the temporal and comparative behavior of plasma syndecan-1 concentrations in pregnancy and in patients with acute pyelonephritis.

Study design
This retrospective cross-sectional study was designed to assess the plasma concentrations of soluble syndecan-1 in the following groups: (1) non-pregnant women (n ¼ 25); (2) uncomplicated pregnant women sampled preterm between 20 weeks and less than 37 weeks (n ¼ 61) and those sampled at term gestation 37-41.7 weeks (n ¼ 69); and (3) pregnant women diagnosed with acute pyelonephritis at 20.7-42.3 weeks (n ¼ 33). The control group was composed of women with an uncomplicated pregnancy without major medical or surgical complications, who delivered appropriate birthweight for gestational age (between the 10th and 90th percentiles [30,31]) and without major congenital anomalies. Acute pyelonephritis was defined as fever (>38 C), flank pain, costovertebral angle tenderness, laboratory findings of pyuria, and a positive urine culture >100,000 colony-forming units [32,33]. Blood cultures used to assess the presence of bacteremia in patients with acute pyelonephritis were available in 85% (28/33)

Sample collection and syndecan-1 immunoassay
Blood specimens were collected in tubes containing ethylene-diamine-tetra-acetic acid (EDTA) at the time of diagnosis of acute pyelonephritis during hospitalization or at prenatal visits for women in the control group. Plasma was obtained by centrifugation of blood samples at 1300Âg for 10 min at 4 C and then stored at À70 C. Maternal plasma syndecan-1 concentrations were determined with a human syndecan-1 enzyme linked immunoassay kit (Cell Sciences, Canton, MA, USA), according to the manufacturer's instructions. The sensitivity of the assay for syndecan-1 was 4.47 ng/mL, and the intra-and inter-assay coefficients of variation were 6.5% and 6.7%, respectively.

Statistical analysis
Demographic data were analyzed by using IBM SPSS version 19.0 (IBM Corporation., Armonk, NY). Chisquare tests were utilized to compare proportions and t-tests were applied for continuous data. All other analyses were conducted with the R statistical language version 3.6.1. Given that the data of plasma syndecan-1 concentrations were not normally distributed, they were log (base 2) transformed before analysis and compared among non-pregnant women, pregnant controls sampled preterm (from 20-36 þ6 weeks of gestation), and pregnant controls sampled at term (!37 weeks of gestation). Significance of pairwise differences were assessed by using two-tailed unequal variance t-tests followed by false discovery rate correction of p-values.
To assess the relationship between the plasma log (base 2) syndecan-1 concentrations as a function of gestational age in normal pregnant women, regression analysis was conducted using polynomial functions of gestational age and the best model fit was selected using F-tests. The expected value (mean) of log (base 2) plasma syndecan-1 concentration for each gestational age in normal pregnancy was estimated using a quadratic function of gestational age at venipuncture.
Due to the gestational-age differences between women with acute pyelonephritis and those with normal pregnancy, the multiple of the mean (MoM) value for gestational age (observed value divided by the expected mean value) was calculated for each case and control. The MoM values were then compared between pregnant women with acute pyelonephritis and controls, using linear models with adjustment for gestational age.

Demographic and clinical data
The demographic data and clinical characteristics of the non-pregnant and normal pregnancy groups are shown in Table 1. Nulliparity differed between groups with a higher rate in non-pregnant women and in control pregnancies sampled preterm than in those sampled at term (p ¼ .01). Gestational age at venipuncture was lower, by design, in the preterm group. The demographics and clinical characteristics of patients with a normal pregnancy and those with acute pyelonephritis are shown in Table 2. Gestational age at venipuncture was lower in pregnant patients with acute pyelonephritis compared to those without this condition. Among patients with acute pyelonephritis, blood cultures were positive in 8/28 (28.6%) cases and negative in 20/28 (71.4%) cases; cultures were not performed in 5 cases.

Plasma syndecan-1 concentrations in normal pregnancy
Women with an uncomplicated pregnancy had a higher plasma concentration of syndecan-1 than nonpregnant women [geometric mean (95% confidence interval [CI]) of syndecan-1 concentration was 51.0 (12.1-216.1) ng/mL in non-pregnant controls; 1280 (365-4487) ng/mL in normal preterm gestations; and 1786 (546-5834) ng/mL in normal term gestations (adjusted p < .005 for all between-group comparisons)] ( Figure 1). Plasma syndecan-1 concentrations increased with gestational age among women with a normal pregnancy (p < .001, R 2 ¼ 0.27) (Figure 2). The non-linear (quadratic) trend of log transformed syndecan-1 concentrations as a function of gestational age had a significantly better fit compared to a linear trend (F-test, p ¼ .0018).  Continuous variables expressed as means ± standard deviation; proportions shown as numbers (n-values) and percentages in parentheses.

Plasma syndecan-1 concentrations in pregnant women with acute pyelonephritis
After converting plasma syndecan-1 concentrations in pregnant women into MoM values for gestational age, the mean syndecan-1 MoM value was higher in women with acute pyelonephritis than in normal pregnant women based on second-and third-trimester samples (p ¼ .048, 1.26-fold change) (Figure 3). The increase was driven primarily by cases with a positive blood culture (p ¼ .009, 1.74-fold change). When only data from third-trimester samples (>28 weeks) were assessed in the analysis, overall differences in the mean syndecan-1 MoM value between cases and controls were slightly greater (p ¼ .03, 1.36-fold change) ( Figure 4). Cases with a positive blood culture contributed to these differences (p ¼ .023, 1.79-fold change), yet in the third trimester, there was also a trend of increase in syndecan-1 MoM values even in cases with a negative blood culture (p ¼ .06, 1.38-fold change). These results suggest that the increase of syndecan-1 concentration in acute pyelonephritis is more pronounced later in gestation and in the presence of a positive blood culture. An alternative analysis, in which log 2 syndecan-1 concentration data were fit directly as a function of the patient group, quadratic terms of gestational age, and their interactions, also suggested that syndecan-1 concentration is higher in cases than in controls, especially later in gestation (p < .05 for each week after 32 weeks, Figure S1).

Discussion
Principal findings of the study (1) Women with a normal pregnancy had a higher plasma syndecan-1 concentration than non-pregnant women; (2) the plasma syndecan-1 concentration increased with gestational age in normal pregnancy; and (3) pregnant women with acute pyelonephritis had a higher plasma syndecan-1 concentration than those with an uncomplicated pregnancy. Collectively, these results suggest that degradation of the glycocalyx, and subsequent release of syndecan-1 into the circulation, is a feature of acute pyelonephritis in pregnancy.

Plasma syndecan-1 concentration is higher in pregnancy and increases with gestational age
The observation that the plasma concentration of syndecan-1 was higher in pregnant women than in nonpregnant women is consistent with the findings of other investigators [1,34] and may represent evidence of glycocalyx breakdown in pregnancy. The most likely explanation for the increased concentration is the release of syndecan-1 by the syncytiotrophoblast glycocalyx [35][36][37]. Indeed, glycocalyx lines the intervillous space of the placenta from which syndecan-1 is shed into the maternal circulation [35,[38][39][40][41]. The second contributor to the increased concentration of syndecan-1 in pregnancy is likely to be the physiologic intravascular inflammatory state observed in normal gestation [7][8][9][10]. Stimulation of the innate immune system may lead to endothelial cell activation and a physiologic change in the endothelial glycocalyx reflected in the increased concentration of syndecan-1 during uncomplicated pregnancy [34,42].

High plasma concentration of syndecan-1 in acute pyelonephritis during pregnancy
In pregnancies complicated by acute pyelonephritis, we found an increase in plasma syndecan-1 concentration. This finding supports the hypothesis that acute pyelonephritis is an inflammatory state characterized by endothelial glycocalyx damage similar to other conditions in which there is intravascular inflammation [43][44][45][46]. Previous studies conducted in non-pregnant patients reported that plasma syndecan-1 concentration is higher in patients with septic shock [14,47] and that the severity of sepsis, mortality rate [14,48], and the Sequential Organ Failure Assessment score [47,48] correlated with plasma syndecan-1 concentration.
In the present study, the increase in plasma syndecan-1 concentration was driven primarily by cases with a positive blood culture, indicating that bacteremia leads to greater degradation of the glycocalyx. The data reported herein suggest that plasma syndecan-1 concentration may be a marker for disease severity, progression, and the need for intensive care in the setting of acute pyelonephritis in pregnancy. However, since the observed differences of plasma syndecan-1 concentration between patients with and without bacteremia or between those with acute pyelonephritis and normal pregnancy were modest, plasma syndecan-1 concentration, if utilized alone, is unlikely to be a good discriminator for bacteremia or for the diagnosis of acute pyelonephritis.
Acute pyelonephritis, preeclampsia, plasma syndecan-1 concentration, and endothelial dysfunction Two important complications of pregnancy-preeclampsia and acute pyelonephritis-are characterized by endothelial cell activation/dysfunction [49]. Indeed, different markers of endothelial cell activation/dysfunction are increased in both conditions, and these include soluble E-selectin [49] and vascular cell adhesion molecule-1 (sVCAM-1) [49,50] (Table 3). One may expect the behavior of plasma syndecan-1 would be similar. However, plasma syndecan-1 concentration is significantly lower in patients with preeclampsia [35,37,[51][52][53] while the concentration of syndecan-1 is higher in women with acute pyelonephritis than in normal pregnant women. The findings in women with preeclampsia were derived from several retrospective case-control studies. The concentrations of syndecan-1 were determined in either serum or plasma samples, using immunoassays similar to that of our study. We propose three potential explanations for these observations. First, the difference may be attributed to decreased release of syndecan-1 into the maternal circulation from syncytiotrophoblasts in preeclampsia. Szabo et al. examined placental syndecan-1 expression in patients with preeclampsia and HELLP syndrome and proposed that an alteration in the cytoskeleton actin network within syncytiotrophoblasts may lead to an accumulation of these proteoglycans within the cytoplasm, resulting in decreased plasma concentrations of syndecan-1 [35]. However, Jokimaa et al. observed weaker syndecan-1 expression on the villous syncytiotrophoblast, using immunohistochemistry, in placentas of women with preeclampsia [40]. Second, the degree of damage of the endothelial glycocalyx may be greater in patients with acute pyelonephritis, especially in those with bacteremia, than in patients with preeclampsia, which may explain the greater release of syndecan-1 in acute pyelonephritis. Third, it is also possible that the endothelial cell dysfunction of acute pyelonephritis is different than that of preeclampsia.

Strengths and limitations
The prevalence of bacteremia in our study population was 28.6%, which is higher than previously reported as 10-20% [1,54,55]. This could be attributed to two reasons: first, as a sole referral center in the region, our institution received more severe cases, thus more blood cultures performed than at other institutions; second, the inclusion of only patients with a positive urine culture and not just suspicious cases of acute pyelonephritis into this study. Moreover, the cross-sectional study design precludes the attribution of causal association between acute pyelonephritis, plasma syndecan-1 concentration, and the inferred glycocalyx injury. Future cohort studies guided by the results reported herein would help to establish causal relationships.  Table 3. Pattern of response in markers of endothelial cell dysfunction.

Conclusion
Acute pyelonephritis is associated with an elevation in maternal plasma syndecan-1 concentration. We propose that the different profiles of syndecan-1 in maternal plasma in preeclampsia versus acute pyelonephritis may reflect changes in the villous trophoblast glycocalyx, or different types of endothelial cell dysfunction phenotypes, or both. Future studies are needed to elucidate whether plasma syndecan-1 concentration can differentiate the nature of systemic inflammation between mothers with acute infectious disease and those with preeclampsia.

Disclosure statement
No potential conflict of interest was reported by the author(s).