SERUM AND URINE THIOFLAVIN T (ThT) ENHANCED FLUORESCENCE IN SEVERE PREECLAMPSIA

Abstract

Common features of amyloid-like proteotoxic aggregates are the ability to bind Congo red (congophilia) and to induce fluorescence of Thioflavin-T (ThT). Based on prior discovery that women with preeclampsia exhibit urine congophilia, we proposed that amyloid-like protein aggregates present in urine also circulate in the blood stream and this feature is linked to disease severity and clinical phenotype. ThT fluorescence was investigated in 217 paired serum and urine samples from women with severe features of preeclampsia (sPE, n=101, median [IQR] gestational age (GA): 32 [29-35] weeks), mild features of preeclampsia (mPE, n=22, GA: 36 [36-37] weeks), chronic hypertension (crHTN, n=15, GA: 38 [37-39] weeks), healthy pregnant controls (n=57, GA: 39 [38-39] weeks), and non-pregnant controls (n=22). Serum and urine fluorescence attributable to advanced glycation end-products (AGE) was measured in the same samples with correction for auto-fluorescence. There were no GA-related changes in ThT fluorescence, although near-term serum ThT fluorescence increased compared to non-pregnant state. Compared to healthy pregnant controls, serum and urine ThT fluorescence was increased in sPE (P<0.001 for both) but not in mPE or crHTN. Except for crHTN, AGE-related fluorescence of serum or urine did not differ from controls. Urine congophilia had a stronger relationship with PE severity compared to either urine or serum ThT fluorescence. However, serum ThT fluorescence was independently associated with clinical features of HELLP syndrome (P=0.003). We demonstrate that ThT fluorescence, a marker of amyloid-like aggregates, is increased in serum of women with PE and likely due to their cytotoxicity associated with HELLP syndrome.

Introduction

Preeclampsia (PE), a pregnancy-specific hypertensive disorder affecting 6-8% of pregnancies, contributes considerably to maternal morbidity and mortality worldwide.^1,2 Historically, research into the pathophysiology of PE has focused on the coagulation cascade, inflammatory mediators and pro- and anti-angiogenic factors.³ Recent biomarker discovery, however, led to a novel view of PE as a pregnancy-specific protein conformational (misfolding) disorder.⁴

A common denominator of conformational disorders is accumulation of supramolecular aggregates of misfolded proteins (collectively named amyloids).⁵ Protein misfolding leads to formation of highly organized and generally intractable aggregates that are cytotoxic.⁶ Amyloidogenesis links pathogenically several human diseases, including Alzheimer’s, Parkinson’s and prion disease.^6,7 Features of protein misfolding have been recently described in atherosclerosis and cancer.^8,9

Small molecules such as Congo red (CR) and Thioflavin T (ThT) have been essential tools to study pathogenic assemblies in misfolding diseases in both tissues and biological fluids.¹⁰ The exact mechanism through which amyloidogenic proteins bind either CR (congophilia) or ThT is not well understood.¹¹ Although there is agreement that both chemicals bind through electrostatic and hydrophobic interactions, they are not identical in their affinities toward different conformations of misfolded proteins. Because ThT binds to the channels running parallel to the long axis of fibrils, fluorescence of ThT increases greatly upon interaction with fibrils, but less in the presence of prefibrillar aggregates.^{11,12,13,14,15} Conversely, CR binds in grooves formed by β-sheets of partially folded intermediates as well as of fibrils. Furthermore, due to the ability of CR to form interactions with metastable proteins, congophilia has been proposed as a measure of general protein instability.¹⁶

Amyloid fibrils are distinguished by their characteristic elongated morphology and ability to alter the spectral properties of ThT upon interaction. This property makes ThT a particularly effective tool for identifying amyloid fibrils in vivo and in vitro. Traditionally, it is believed ThT fluorescence originates only from the dye bound to amyloid fibrils. In vitro experiments showed that incubation of albumin with glucose or with methylglyoxal induces ThT fluorescence, suggesting that chemical adducts of albumin such as advanced glycation end-products (AGEs) may contribute to the variation in ThT fluorescence of human serum.¹⁷ This process could be extremely relevant in PE because AGEs are elevated in maternal circulation of women with PE and may contribute to their vascular dysfunction.^18,19

Amyloid-like proteins including Alzheimer’s β-amyloid have been demonstrated in the urine and placenta of women with PE.⁴ Likewise, urine of women with PE exhibits congophilia, a characteristic of protein misfolding, and this correlated with disease severity. Although in PE many proteins and protein fragments other than β-amyloid contribute to urine congophilia,⁴ the relationship between urine congophilia and serum and urine induction of ThT fluorescence has not been characterized. We hypothesize that serum and urine ThT fluorescence are increased in PE subjects in direct relationship with urine congophilia. The aim of this study was to perform a systematic analysis of serum and urine ThT fluorescence in relationship to urine congophilia and clinical features of PE severity.

Methods

The data that support the findings of this study and any additional methodological details that may help replicate our findings are available from the corresponding author upon reasonable request.

Subjects and sample collection

We analyzed paired serum and urine samples retrieved from 217 reproductive age women of which 195 were pregnant with singletons and 22 were non-pregnant. Among pregnant women, 138 were admitted for evaluation of PE and enrolled consecutively in the following groups: PE with severe features (sPE, n=101, median [interquartile range (IQR)] gestational age (GA): 32 [29-35] weeks), PE with mild features (mPE, n=22, GA: 36 [36-37] weeks) or chronic hypertension (crHTN, n=15, GA: 38 [37-39] weeks). Maternal blood and urine specimens from normotensive pregnant women, (P-CRL, n=57, GA: 39 [38-39] weeks) and from the non-pregnant women (NP-CRL, n=22) served as controls. All normotensive pregnant controls had uncomplicated gestations and delivered at term (>37 weeks) a healthy baby. Participants were recruited from women evaluated by or admitted to the Labor and Birth, antepartum and outpatient units at Yale New Haven Hospital (YNHH) between January 2004 to February 2009. The Yale University Human Investigation Committee approved the study protocol, and written informed consent was obtained from all participants.

GA was established based on last menstrual period and/or early ultrasound evaluations (<20 weeks of gestation). Because patient recruitment and sample collection was completed prior to publication of the 2013 Task Force recommendations for definition of PE we list below elements of the definition of the hypertensive disorders valid at the time of sample collections, specifically the American College of Obstetricians and Gynecologists (ACOG).²⁰ With the exception of proteinuria the key clinical elements responsible for the used diagnoses remained largely the same. Severe PE was defined based on: GA >20 weeks, blood pressure of 160 mm Hg systolic or higher, or 110 mm Hg diastolic or higher on two occasions at least 6 hours apart.²¹ Other elements of the clinical diagnosis of sPE included: cerebral or visual disturbances, epigastric or right upper-quadrant pain, pulmonary edema or cyanosis, oliguria (urinary output less than 500 mL/24 hr), impaired liver function, and/or thrombocytopenia (<100,000 cells/μL). HELLP syndrome was defined as follows: hemolysis [bilirubin ≥1.2 mg/dL or lactate dehydrogenase (LDH) > 600 IU/L], hepatic dysfunction defined as serum AST >2 times upper limit of normal for local laboratory (for YNHH the laboratory upper normal limits were AST >34 IU/L, ALT>34 IU/L), and thrombocytopenia documented by a platelet nadir < 100,000 cells/μL.²²

Measurement of ThT fluorescence

Serum (3 μL) and urine (30 μL) were each loaded in duplicate onto wells of a microtiter plate where they were mixed with 8 μM ThT in 100 mM sodium phosphate buffer, pH 7.4 (final volume 250 μL). Fluorescence readings were acquired in a microplate spectrofluorometer (BMG LABTECH CLARIOstar, Ortenberg, Germany) equipped with 444/485 nm excitation/emission filters. To control for possible batch effects or imperfections in the plates, fluorescence from additional duplicate wells measured in the absence of ThT was subtracted from values obtained with ThT and measured in the same plate. ThT fluorescence was expressed as arbitrary units (AU)/mL after dilutional correction. In prior experiments we determine that sample handling, including freezing and thawing did not impact on either urine congophilia or ThT fluorescence.

Advanced Glycation End-Products (AGEs)-related fluorescence

Serum and urine AGEs-related fluorescence was measured in the wells without ThT using excitation/emission wavelengths of 370/440 nm as previously described.²³

Serum and urine creatinine and protein concentrations

Serum and urinary creatinine levels were measured using a colorimetric assay (Stanbio Laboratory, Boerne, TX, USA) against standard curves derived from known creatinine concentrations. Total protein levels were measured using a bicinchoninic acid/cupric sulfate reagent assay (Pierce BCA Protein Assay Kit, Thermo Scientific, Waltham, MA, USA). Urine ThT fluorescence was expressed as AU/mg creatinine (mgc). Fractional excretion of ThT-induced fluorescence was calculated for each condition using the formula: (urine/plasma analyte concentration) ÷ (urine/plasma creatinine concentration) × 100. The fractional excretion of an analyte indicates the proportion of the analyte excreted in the urine compared with that filtered by the glomeruli. Fractional excretion is reported in relation to creatinine clearance because creatinine is neither reabsorbed nor significantly secreted. These calculations can help understand whether increased serum levels of an analyte are due to increased production or decreased excretion.

Congophilia

Urine congophilia was determined as previously described using a nitrocellulose-based array.⁴ Congophilia results are expressed as % Congo Red Retention (%CRR), which is a measure of spot redness after removing unbound Congo Red. The results of urine congophilia for 54 women have been reported previously.⁴ However, for the purpose of this study urine congophilia was re-measured for all cases from the same aliquot with the one used for ThT analysis. The calculation of %CRR was accomplished using the automated image analysis protocol implemented on a mobile phone device.²⁴ Samples measuring %CRR ≥15% in protein-normalized urine were classified as non-reassuring (NR-CRR) as previously described.⁴

Statistical analysis

As far as we know this is the first study of ThT-induced fluorescence of PE serum therefore we based our sample size calculation on the prior experiments where we found that 40 P-CRL and 40 sPE urine samples were sufficient to demonstrate statistical significance in urine congophilia and ThT fluorescence between the two groups.⁴ Data was tested for normality with the Kolmogorov-Smirnov method and reported as median and IQR. Comparisons among groups were performed using Kruskal-Wallis ANOVA on Ranks, 2-way ANOVA and Chi square tests. Correlations were perfomed using Spearman Rank Order Correlation test. Receiver Operating Characteristic (ROC) curves were created for urine %CRR, urine and serum ThT fluorescence for identification of women who had a medically indicated delivery for PE. Multiple stepwise regression analysis was used to explore concurrent relationships between variables and various clinical characteristics as independent variables. Variables were included into the model based on a P<0.05 and excluded if P>0.1. SigmaPlot v12.5 (RockWare, Golden, CO, USA) and MedCalc v16.8.4 (Broekstraat, Belgium) statistical software were used for data analysis. A P value <0.05 was considered to indicate statistical significance.

Results

The demographic, clinical and outcome characteristics of the pregnant women in the different groups are presented in Table 1. Women with crHTN were older, had a higher body weight, and higher gravidity compared to the other groups. Hypertensive women had blood pressure values significantly elevated compared to controls. Dipstick proteinuria was present and significantly higher in PE women independent of the severity of the clinical features. Women in the sPE but not mPE group had significantly higher 24h proteinuria compared to crHTN group. More women with sPE had neurological symptoms at admission. Women with sPE and mPE delivered earlier compared to either crHTN or pregnant controls. The incidence of intra-uterine growth restriction (IUGR), eclampsia was higher in sPE. In of the women in the sPE group, 25% (25/101) had one or more features of HELLP syndrome with 5 of these presenting with all three features. The subgroup of women with sPE and one or more features of HELLP had the following clinical laboratory findings: AST (median [IQR]): 69 [32-225] IU/L, ALT: 75 [24-227] IU/L, platelet count: 95,000 [80,500-162,500] cells/μL, and LDH: 467 [329-621] IU/L.

Table 1.

Group specific demographic, clinical and outcome characteristics of pregnant women enrolled in this study (n=195)

Variable	Pregnant Control n=57	Chronic Hypertension n=15	Mild Preeclampsia n=22	Severe Preeclampsia n=101	P value
Characteristics at enrolment
Age, years*	28 [23 - 32]†	35 [33 - 38]‡	29 [22 - 33]†	27 [20 - 33]†	<0.001
Gravidity*	2 [1 - 3]†	4 [3 - 5]‡	1 [1 - 2]†	1 [1 - 3]†	<0.001
Parity*	0 [0 - 1]	1 [0 - 2]	0 [0 - 1]	0 [0 - 1]	0.025
Non-Caucasian race§	25 (44)	11 (73)	10 (45)	61 (60)	0.076
Weight, kg*	80 [68 - 89]†	108 [88 - 139]‡	87 [70 - 125]	85 [72 - 100]	0.027
Nulliparity§	30 (52)	4 (26)	15 (68)	61 (60)	0.055
Gestational age, weeks*	26 [17 - 36]	33 [24 - 37]	36 [34 - 37]‡	32 [28 - 35]	<0.001
Systolic blood pressure, mmHg*	119 [110 - 120]†	152 [128 - 160]‡	147 [140 - 157]	160 [150 - 173]‡	<0.001
Diastolic blood pressure, mmHg*	70 [60 - 70]†	98 [76 - 102]‡	90 [84 - 100]‡	100 [90 - 108]‡	<0.001
Dipstick proteinuria*	0 [0 - 0]	0 [0 - 1]	2 [1 - 2]‡	3 [2 - 3]†‡	<0.001
24h-protein excretion, grams/24h*	NA	0.22 [0.09 - 0.36]	0.82 [0.42 - 1.82]	3.09 [1.49 - 4.89]†	<0.001
Neurological symptoms including headache§	0 (0)	2 (13)	7 (31)‡	54 (52)‡	0.016
Pregnancy complications and pregnancy outcome
Gestational age at delivery, weeks*	39 [38 - 39]	38 [37 - 39]	36 [36 - 37]†‡	32 [29 - 35]†‡	<0.001
Indicated delivery for preeclampsia§	0 (0)	0 (0)	22 (100)†‡	101 (100)†‡	<0.001
Delivery for preeclampsia <37 weeks§	0 (0)	0 (0)	12 (54)†‡	86 (84)†‡	<0.001
IUGR§	0 (0)	1 (6)	0 (0)	14 (13)‡	0.007
Preeclampsia with HELLP features§	0 (0)	0 (0)	0 (0)	25 (25)‡	<0.001
HELLP syndrome (all 3 features present)§	0 (0)	0 (0)	0 (0)	5 (5)	0.189
Eclampsia§	0 (0)	0 (0)	0 (0)	6 (6)	0.124

HELLP, hemolysis, elevated liver enzyme levels, and low platelet levels; IUGR, intrauterine growth restriction

^*Data presented as median [interquartile range] and analyzed by Kruskal-Wallis ANOVA on Ranks

^†P<0.05 vs. Chronic Hypertension group

^‡P<0.05 vs. Pregnant Control group

^§Data presented as n (%) and analyzed by Chi square test.

Compared to non-pregnant controls, serum induced ThT fluorescence in pregnancies with normal outcome was not impacted significantly until late gestation (≥34 weeks, Figure 1A). However, there was no significant correlation between GA and serum ThT fluorescence in women with normal pregnancy outcomes (r=0.224, P=0.164). Urine had ~50-fold less ThT fluorescence than serum; there were no differences in urine ThT fluorescence with either pregnancy or across gestation (Figure 1B). Serum fluorescence attributable to AGEs was significantly increased in early pregnancy and decreased thereafter with a significant, but inverse relationship with GA (r= −0.398, P=0.011, Figure 1C). In urine, fluorescence attributable to AGEs was increased late gestation (≥34 weeks), compared to non-pregnant state (P<0.001, Figure 1D). There was no correlation between ThT fluorescence and that attributable to AGEs (serum: r=−0.308, P=0.053; urine: r=0.224, P=0.164).

Figure 1. Serum & urine thioflavin T (ThT) and advanced glycation end-products (AGE)-related fluorescence in non-pregnant and normal healthy gestations.

A, Serum ThT fluorescence. B, Urine ThT fluorescence. C, Serum fluorescence attributable to AGEs. D, Urine fluorescence attributable to AGEs. Arbitrary units/mg creatinine (AU/mgc). Error bars show standard error. For each panel, bars with a common letter indicate groups that are not statistically different at P>0.05.

When compared to healthy controls and crHTN, sPE women displayed significantly higher serum ThT fluorescence (Figure 2A). Urine ThT fluorescence of women with sPE was significantly higher compared to mPE, crHTN and pregnant healthy controls (Figure 2B). In serum, the highest levels of AGE-related fluorescence were noted in women with crHTN (Figure 2C). Women with sPE, mPE and healthy pregnant controls had similar serum levels of AGE-related fluorescence. Remarkably, healthy pregnant controls had the highest urine AGE-related fluorescence (Figure 2D).

Figure 2. Serum & urine thioflavin T (ThT) and advanced glycation end-products (AGE)-related fluorescence in healthy pregnant controls, women with chronic hypertension (crHTN), and women with preeclampsia (PE) with mild (mPE) or severe (sPE) features.

A, Serum ThT fluorescence. B, Urine ThT fluorescence. C, Serum fluorescence attributable to AGEs. D, Urine fluorescence attributable to AGEs. Arbitrary units/mg creatinine (AU/mgc). Error bars show standard error. For each panel, bars with a common letter indicate groups that are not statistically different at P>0.05.

We identified an over 3-fold increase in fractional excretion of ThT-induced fluorescence in sPE as compared to non-pregnant and pregnant controls (Figure 3A). When we analyzed the data based on GA at diagnosis we found out that early-onset sPE was associated with increased fractional excretion of ThT-induced fluorescence as compared to their respective GA control group (P=0.001, Figure 3B). Fractional excretion for late-onset sPE was not different from their pregnant control group (P=0.545). Overall, we found no statistically significant difference in AGE-related fluorescence between non-pregnant controls, pregnant controls and women with sPE (P=0.161, Figure 3C). However, when data was separated based on GA, an increased fractional excretion in AGEs was observed for early-onset sPE that reached the physiological level expected only later in gestation (≥34 weeks). In contrast, in women with late-onset sPE the renal output of AGE-related fluorescence failed to occur, which rendered their fractional excretion significantly lower from their GA matched pregnant controls (Figure 3D).

Figure 3. Fractional excretion of thioflavin T (ThT) and advanced glycation end-products (AGEs) fluorescent substances in healthy non-pregnant and pregnant controls, and women with severe features preeclampsia (sPE).

A, Fractional excretion of ThT-induced fluorescence in women with sPE versus non-pregnant and healthy pregnant controls. B, Fractional excretion of ThT-induced fluorescence in early-onset (20–33 weeks) and late-onset sPE (≥34 weeks). C, Fractional excretion of AGEs in women with sPE versus non-pregnant and healthy pregnant controls. D, Fractional excretion of AGEs in early-onset (20–33 weeks) and late-onset sPE (≥34 weeks). Fractional excretion was calculated for each condition using the formula: (urine/plasma analyte concentration) ÷ (urine/plasma creatinine concentration) × 100. Error bars show standard error. For each panel bars with different letters are statistically different at P<0.05.

Next, we explored the relationship between serum and urine ThT fluorescence and urine congophilia. Among pregnant women, those with NR-CRR had significantly higher ThT fluorescence in both serum (Figure 4A) and urine (Figure 4B). There was no significant difference in serum (Figure 4C) or urine (Figure 4D) AGE-related fluorescence based on congophilia level.

Figure 4. Relationship of urine congophilia with serum and urine thioflavin T (ThT) and advanced glycation end-products (AGE)-related fluorescence.

A, Serum ThT fluorescence in women with and without congophilia as measured by % Congo Red Retention (%CRR) ≥15%. B, Urine ThT fluorescence in pregnant women with and without urine congophilia. C, Serum AGE-related fluorescence in women with and without urine congophilia. D, Urine AGE-related fluorescence in women with and without urine congophilia. Arbitrary units/mg creatinine (AU/mgc). Error bars show standard error. For each panel bars with different letters are statistically different at P<0.05. E, Receiver Operating Characteristic (ROC) curve for urine congophilia measured by %CRR (red line) compared to ThT fluorescence of serum (blue line) and urine (green line) for identifying women who had a medically indicated delivery for PE.

To determine whether ThT fluorescence may have diagnostic utility for PE, we created ROC curves comparing urine %CRR with serum and urine ThT fluorescence and with AGE-related fluorescence measured in the same samples. Both urine and serum ThT fluorescence showed statistical prediction for PE based on area under the ROC curve >0.5 (ROC area serum ThT: 0.753, 95%CI [0.658-0.832] and urine ThT: 0.755, 95%CI [0.660-0.834]). However, as diagnostic test for PE, urine %CRR was significantly superior compared to both urine and serum ThT fluorescence (ROC area %CRR: 0.992, 95%CI [0.950-1.000], P<0.001 for both, Figure 4E) for identification of women who had a medically indicated delivery for PE. Neither serum nor urine AGE-related fluorescence had diagnostic value for PE (ROC area serum AGE: 0.543, 95%CI [0.432-0.651] and urine ThT: 0.506, 95%CI [0.396-0.616]).

To explore concurrent relationships between urine congophilia, serum or urine ThT fluorescence as dependent variables and PE severity, we performed a multiple linear regression analysis restricted to women with a diagnosis of PE. Both systolic and diastolic blood pressures where significantly determined by %CRR (systolic r=0.329, diastolic r=0.388, P<0.001 for both) while urine and serum ThT fluorescence were eliminated from the model based on P>0.1. There were no significant relationships between the degree of proteinuria (by dipstick or the 24-hour measurement) and any of the three dependent variables. Similarly, there were no differences in congophilia or ThT fluorescence among women who either presented with IUGR or developed neurological manifestations including headache. However, among hypertensive women, serum (but not urine) ThT fluorescence emerged as significantly linked with several laboratory features of HELLP syndrome independent of congophilia and GA. Among them, serum ThT fluorescence correlated significantly with levels of liver enzymes (AST: r=0.407, ALT r=0.448, P<0.001 for both) and platelet count (r= −0.267, P=0.013) but not with LDH activity (P=0.352).

Discussion

It is well established that both ThT and Congo red bind to cross-linked β-sheet amyloid-like structures. Interaction with metastable domains of α-helical proteins such as human serum albumin has also been suggested.^14,25 In this study, we demonstrated that in pregnancies with normal outcome, serum but not urine ThT fluorescence is minimally but significantly increased at term. Recent studies showed that elevated levels of systemic inflammation and oxidative stress are most apparent in mothers who have term, rather than preterm PE.²⁶ It is possible that a progressive and heightened state of “physiologic” inflammation and oxidative stress lead toward term to an increased degree of aberrant protein misfolding, and hence increased ThT fluorescence in the absence of clinical manifestation of PE.^27,28 Because most women are induced by 41 weeks, PE will not become clinically manifest as would probably happen if the same mothers would be allowed to progress beyond 42–43 weeks. In the absence of clinical manifestation of PE, these term mothers would be classified as normal although the underlying molecular processes that are associated with PE are already ongoing. Previous studies outside pregnancy have suggested a link between serum ThT fluorescence and impaired vascular health.¹⁷ Our observation of increased serum ThT fluorescence in normal term pregnancies may explain why both maternal systolic and diastolic blood pressure are consistently increased in the third trimester above first trimester values in pregnancies with normal outcome.²⁹

Out study clearly indicated that serum, urine and fractional excretion levels of ThT-induced fluorescence were all increased in PE. Furthermore, serum and urine ThT fluorescence was significantly increased in PE women displaying congophilia. These findings strengthen our previous hypothesis that serum and urine amyloid fibrils are an integral component of the pathogenesis of PE.⁴ The possible link between ThT fluorescence and PE could be vascular dysfunction, which is known to increase the risk for future cardiovascular events in women with a history of PE.³⁰ Certainly, amyloid deposits containing apolipoproteins, β-amyloid and α1-antitrypsin have been found to be associated with atherosclerosis.³¹ It is thought that amyloid deposits in vessels decrease vascular elasticity and hence participate in hypertension. Griffin et al. measured serum ThT fluorescence in both healthy subjects and non-pregnant patients at high-risk for cardiovascular disease, demonstrating an inverse relationship between ThT fluorescence and small artery elasticity as measured by pulse-wave analysis.¹⁷ Arterial elasticity is also reduced in both currently and formerly preeclamptic women when measured objectively.^3,2,33 Similar findings have been exhibited by placental vasculature in women with PE.³⁴ Our previous histologic studies demonstrated that in PE placentas, β-amyloid staining was present and more prominent in both basal plate and villous areas compared with GA matched placentas of women with idiopathic preterm birth.⁴ Collectively, our current and previous findings imply that at least part of the increased stiffness of PE placentas is due to amyloid deposition.

Because PE often results in renal dysfunction, it is important to demonstrate that the increased serum ThT fluorescence (representing misfolded proteins) of women with PE is not a manifestation of kidney injury. The increased fractional excretion of ThT-induced fluorescence indicates that the elevated serum ThT fluorescence level in women with sPE is secondary to increased production rather than decreased excretion.

ThT fluorescence was related to clinical severity and the time of PE onset. In our study, mild PE and crHTN groups did not display increased urine or serum ThT fluorescence. In addition, serum ThT fluorescence was an independent factor related to HELLP syndrome, arguable one of the most severe clinical manifestations of PE. It remains intriguing that in women with sPE, fractional excretion of ThT-induced fluorescence was significantly increased preterm but not term. Clinical and molecular data exist to support the premise that activation of various molecular pathways could be responsible for the severity and time when the disease becomes clinically manifest. For example, asymptomatic cardiac diastolic dysfunction is evident at mid-gestation in women who subsequently develop preterm PE but not in those who develop term PE.³⁵ Higher level of placental miR-518a-5p, miR-423, miR-124 and miR-431were identified in early-onset PE (<34 weeks GA) cases implying that chemokine receptor CCR6 expression, cell growth, and alternative pre-mRNA splicing in non-neuronal cells pathways could be down-regulated when PE develops early in gestation.³⁶

Yanagisawa et al. established that measurement of AGE-specific fluorescence (i.e. AGE-peptide) might serve as a simple and useful test to assess circulating AGE levels and monitor AGE excretion.²³ In our study, serum, urine and fractional excretion levels of AGE-related fluorescence did not correlate with ThT fluorescence or with clinical measures of PE severity. This suggests AGEs alone had minimal impact on ThT fluorescence.

The observation that serum AGE-related fluorescence is increased in early gestation in pregnancies with normal outcomes is novel. One possible explanation could be the role of AGEs in limiting excessive trophoblast invasion and placentation in early pregnancy.³⁷ Increased serum AGE-related fluorescence in pregnant women with crHTN could be explained through tendency for AGEs to accumulate in tissues with aging and cross-linking of collagen.³⁸ Data from the non-pregnant population suggest elevation of the AGE prototype carboxymethyl-lysine in serum was associated with increased arterial stiffness and hypertension.^39,40

It has been suggested that detection of amyloid-like adducts in serum using ThT fluorescence could be an important test to predict cardiovascular disease risk.¹⁷ Although serum and urine ThT fluorescence did correlate with PE and its severity, the data generated from the current study has shown that assessment of urine congophilia with the CRD test performs better at predicting PE than either urine or serum ThT fluorescence. Nevertheless, the finding that among hypertensive women serum (but not urine) ThT fluorescence was significantly related to lower platelet count and elevated liver enzymes, independent of urine congophilia, could have clinical application. It is plausible that particular types of proteotoxic aggregates with affinity for ThT are less effectively cleared by the kidney and thus accumulate in maternal circulation. It is also possible that ThT fluorescence may in fact correlate with clinical features of PE such as liver inflammation, that are not assessed part of routine clinical care and will need to be more thoroughly investigated.

Perspectives

The increase in ThT fluorescence in the blood of women with PE concurs with the notion that pathophysiology of PE may include protein misfolding and aggregation. The increase in serum ThT-induced fluorescence in women with PE seems to be the result of increased production of amyloid-like substances rather than decreased excretion. However, the origin of the amyloid-like aggregates responsible for ThT-induced florescence remains unknown. The relationship between ThT-induced fluorescence, low platelet count and elevated liver function tests suggests that the circulating proteotoxic aggregate may induce cellular toxicity and liver inflammation. From a clinical perspective urine congophila remains superior at diagnosing preeclampsia compared to serum ThT fluorescence.

Novelty and significance.

What is new?

• Amyloid-like substances with affinity for Thioflavin-T (ThT) are present in serum and urine of women with severe preeclampsia.

• Serum ThT fluorescence is significantly upregulated early-mid gestation, but not at term.

• Among hypertensive women, serum ThT fluorescence related to manifestations of HELLP syndrome independent of congophilia.

What is relevant?

• The current study confirms that amyloid-like structures are circulating in the blood of women with preeclampsia displaying congophilia.

• Urine congophilia is superior to both urine and serum ThT fluorescence in predicting preeclampsia.

• Serum ThT fluorescence is associated with HELLP syndrome suggesting maternal circulating proteotoxic aggregate may induce hepatocyte and platelet toxicity.

Summary.

• Amyloid-like aggregates, as assessed by ThT fluorescence, are present in serum and urine of women with preeclampsia.

• The increased fractional excretion of ThT-induced fluorescence implies the elevated serum ThT fluorescence level of women with sPE is secondary to increased production rather than decreased excretion of amyloid-like structures.

• The observed relationship between elevated serum levels of ThT fluorescence and clinical manifestations of HELLP syndrome implies amyloid-like structures may carry hepatocyte and platelet cytotoxic effects.