Cerebrospinal beta-amyloid peptides(1-40) and (1-42) in severe preeclampsia and HELLP syndrome – a pilot study

During pregnancy, substantial alterations in cerebral plasticity, vascular remodeling and neuronal growth occur in the maternal brain. We investigated whether concentrations of selected neurodiagnostic biomarkers in the cerebrospinal fluid of women with preeclampsia/HELLP syndrome differ from those in healthy controls using enzyme-linked immunosorbent assay technique. We found that tau protein concentrations (p = 0.016) and phospho-tau/tau ratio (p < 0.001) in cerebrospinal fluid were significantly lower in 39 preeclamptic women compared to 44 healthy controls during third trimester of pregnancy. Beta-amyloid(1-40)/(1-42) ratio was significantly higher in HELLP syndrome than in severe preeclampsia (8.49 + 2.73 vs. 4.71 + 1.65; p = 0.007). We conclude that beta-amyloid(1-40)/(1-42) ratio in cerebrospinal fluid can discriminate severe preeclampsia and HELLP syndrome. High beta-amyloid peptide and low tau protein concentrations are associated with impaired development of the materno-feto-placental unit and correlate with placental dysfunction.


Discussion
In our study mean concentrations of beta-amyloid peptides(1-40) were higher in women with HELLP syndrome than in preeclamptic women and healthy controls. Human placenta and thrombocytes abundantly express amyloid precursor protein (APP) 5,6 . Presumably, thrombocytopenia influences metabolism of beta-amyloid  and   12 . Although the histopathologic profile and the range of placental lesions differ between PE and HELLP syndrome 13 , it remains a matter of ongoing debate whether HELLP syndrome is a severe form of PE or a separate disease. We found that beta-amyloid(1-40)/(1-42) ratio can discriminate between severe PE and HELLP syndrome. The findings of our study support the hypothesis that HELLP syndrome is a distinct disease pattern and not simply a variety of severe PE.
In our study, median CSF concentrations of tau protein and the phospho-tau/tau ratio were significantly lower in women with PE and HELLP syndrome than in healthy controls. This is in agreement with recently reported diminished CSF tau and phospho-tau-181 protein concentrations in patients with placental dysfunction 14,15 . During normal pregnancy there is an increase in tau protein concentrations 16 . Presumably, diminished CSF concentrations in PE and HELLP syndrome imply fewer maternal brain adaptations. Virtanen et al. observed an inhibitory effect on tubule formation in third trimester preeclamptic women 17 . Serum concentrations of angiogenic factors sFlt-1 levels were higher and PlGF were lower in third trimester preeclamptic women compared to healthy controls with increasing sFlt-1/PlGF ratio between the first and the third trimester 18 . Unfortunately, sFlt-1 and PlGF have limited sensitivity for stratification of women with suspected PE 19 . PE still lacks a reliable, early means of diagnosis or prediction, and a safe and effective therapy 20 , So far, the diagnosis of PE and HELLP syndrome is mostly based on clinical findings and increasing sFlt-1/PlGF ratio. In particular, the onset threshold of plasma levels of the sFlt-1/PlGF ratio proved to be a valuable screening tool for detecting the imminent onset of PE within four weeks after blood sampling during the second trimester, namely at 19 to 31 weeks 21 . Potentially, detection of amyloid-targeting fluorophores in the urine may contribute to early identification of PE patients during pregnancy 22 . Affintiy for conformational antibodies raised against aggregated beta-amyloid peptides and dysregulation in the APP proteolytic pathway may offer future diagnostic and therapeutic options 6,8 . Neuron-specific enolase and S100B were reported to be increased before clinical development of PE can be  www.nature.com/scientificreports www.nature.com/scientificreports/ verified 23 . This supports the hypothesis that altered neural remodeling in the maternal brain precedes impaired development of the placenta.
Pregnancy renders substantial changes in maternal brain, primarily reduction in gray matter volume in regions subserving social cognition 24 . Decrease in brain size begins after placental implantation 25 . It reaches maximum at term and it endures for at least 2 years post-partum 25,26 . Structural and functional changes accompany fundamental behavioural adaptations, stimulating to progress from self-involved individuality to carying motherhood 26 . Changes in the maternal brain during pregnancy cannot be explained by endocrine and environmental factors alone 27 . Van Dijk et al. reported that the PE-susceptibility gene STOX1 controls a conserved pathway shared between placenta and brain 28 . STOX1A correlated with severity of disease and was associated with increased amyloid-beta protein precursor processing in neural cells and trophoblast cells 28 . Vaiman and Miralles hypothesized that selective inhibition of one of the two isoforms STOX1A and STOX1B could possibly improve outcome in severe PE 29 .
In our study angiogenic factors in serum correlate with CSF beta-amyloid(1-40) and tau protein.
Our findings question the current perspective on etiology of PE and of HELLP syndrome and support the hypothesis that neurotrophic factors influence the development of the materno-feto-placental unit during pregnancy. However, there are some limitations. First and foremost, the number of enrolled women with PE and HELLP syndrome is small as data were only available on women who have already developed clinically evident disease and who underwent surgical or obstetrical operations in spinal anesthesia. We are aware that routine lumbar puncture for diagnosis of PE and HELLP syndrome is not practical during pregnancy. Furthermore, treatment is limited mostly to maternal stabilization and timely delivery by medically induced labor or cesarean section 2 . In women with PE and HELLP syndrome delivery by cesarean section was on average four weeks of gestation earlier than in controls, thus partly explaining the low values for placental weight, baby weight and Apgar score. Duration of storage of samples varied among women enrolled and might have biased measurement results. Sampling of patient CSF was performed mostly during the daytime. Observer bias among pediatricians and midwives might have influenced assessment of Apgar Scores and placental weight. Furthermore, rating of perceived stress and physical activity depended on subjective justification of participating women and might have been influenced by current events. Larger trials focusing on placental morphology and function are needed to draw definitive conclusions.
In conclusion we observed that beta-amyloid(1-40)/(1-42) ratio in cerebrospinal fluid can discriminate severe PE and HELLP syndrome. High beta-amyloid peptide and low tau protein concentrations are associated with impaired development of the materno-feto-placental unit and correlate with placental dysfunction.

Materials and methods
Women with PE/HELLP syndrome during pregnancy who underwent spinal puncture for regional anesthesia were consecutively enroled during normal work time in a prospective cross-sectional research at the Department of Gynecology, Innsbruck Medical University Hospital. Healthy women with uncomplicated pregnancies who underwent a cesarean section under spinal block served as controls.

Ethical approval & patients consent. The Ethics Committee of Innsbruck Medical University approved
this study (AN2017-0073 371/4.25). Written informed consent was obtained from all patients before participation with the understanding that anonymized data would be published in a scientific journal. All methods were performed in accordance with the relevant guidelines and regulations.

Study design and study population.
Anesthetists from obstetrical anesthesia screened patients by medical history before participation. Blood chemistry (blood sugar, hemoglobin, serum protein) and kidney function parameters (blood urea nitrogen, creatinine, glomerular filtration rate, and urinary protein) were obtained from medical records. Circulatory and respiratory status (systolic blood pressure, pulse rate and oxygen saturation) were recorded before and during anesthesia. Perceived stress and physical activity were inquired, as they may affect incidence and severity of PE 30 . We used a 5-point Likert scale with (1, very much; 2, much; 3, moderate; 4, poor; 5, very poor) for individual grading of conceptualization of estimated stress. Demographics, clinical characteristics and laboratory findings for participating women were recorded on a working chart.
Inclusion criteria for the study group were: pregnant women, 18 years of age or older, with evidence of PE/ HELLP syndrome verified by clinical signs of hypertension, proteinuria and edema and, if available, positive sFlt-1/PlGF ratio 31 . Exclusion criteria for the study group were: lacking written consent. Inclusion criteria for the control group were: pregnant women, 18 years of age or older, in good general health. Participants were on no chronic medication other than multivitamins, magnesium and iron supplementation. Subjects were in particular free from past or present major psychiatric disorders other than mild depression, major neurological disease, gestational diabetes, and chronic renal and hepatic disease. Exclusion criteria for the control group were: women suffering from hypertension in pregnancy, thrombocytopenia, proteinuria, evidence of placental dysfunction, IUGR, as well as women with a history of PE or HELLP syndrome, cases lacking written consent. All study findings and documents were handled in strictest confidence. Enrolment of patients, data collection and analysis followed the CONSORT 2010 checklist of information to include when reporting a randomized trial 32  www.nature.com/scientificreports www.nature.com/scientificreports/ needle, B Braun, 34209 Melsungen, Germany) was inserted through the introducer into the subarachnoid space and the trocar was removed. Before administration of the intrathecal local anesthetic one milliliter CSF was collected in a sterile polypropylene tube 34 . Immediately after transport to the psychiatric laboratory, CSF samples were frozen and stored at minus 80 °C until the assays were performed.

Definitions.
• Body mass index (BMI), calculated by dividing weight in kilograms by the square of height in meters.
• PE, systemic syndrome with hypertension exceeding 140/90 mmHg presenting beyond 20 weeks of gestation and proteinuria >300 mg protein in a 24 h urine collection or 1 + (0.3 g/L) on urine dipstick 37 . • Early-onset PE, hypertension, proteinuria, placental dysfunction due to poor placental perfusion and reduction of placental volume, IUGR and low birth weight occurring before 34 weeks of gestation 37 . • Mild PE, proteinuria and systolic blood pressure between 140 and 160 mmHg or diastolic blood pressure between 90 and 110 mmHg. • Severe PE, proteinuria and blood pressure exceeding 160 mmHg systolic or 110 mmHg diastolic, frequently combined with elevated liver enzymes 37 . • sFlt-1/PlGF ratio, calculated dividing antiangiogenic factor sFlt-1 by proangiogenic factor PlGF 33 .
Processing and analytical techniques. CSF biochemical markers were detected in cell-free samples using enzyme-linked immunosorbent assay (ELISA) assays (Fujrebio, formerly Innogenetics) that can reliably detect even low levels of CSF biomarkers based on several well-defined monoclonal tau antibodies for normal and abnormal forms of CSF tau protein. Statistical analysis. Primary study endpoint was to evaluate selected CSF neurodiagnostic biomarkers in women with PE/HELLP syndrome. Secondary study endpoint was to detect differences in CSF biomarkers between preeclamptic women and women with HELLP syndrome. Our H0 hypothesis was: there are no differences in CSF concentrations of beta-amyloid(1-42) or (1-40), total tau or phospho-tau-181 proteins between women with PE and HELLP syndrome and normal pregnancies. In order to compensate for individual variations in beta-amyloid production the beta-amyloid(1-42)/(1-40) ratio was calculated.
The sample size was calculated for an alpha error of 0.05 and a power of 80% (beta error of 0.2) to detect significant differences between CSF biomarkers of women with PE/HELLP syndrome and women with normal pregnancy using nQueryAdvisor v.7.0 software. Calculations were conducted with SPSS 25 (IBM SPSS Statistics Standard) using the T test or the Mann-Whitney U test, as indicated. For multiple pairwise comparisons between more than two subgroups, the Kruskal-Wallis test was applied with Bonferroni correction. A post hoc subgroup analysis was performed for beta-amyloid(1-40) <5000 pg/mL and for beta amyloid(1-40) >5000 pg/mL. Group comparisons of the beta-amyloid(1-42)/(1-40) ratio were assessed by analysis of variance followed by the Dunn-Bonferroni test 39 . Pearson's correlations (two-tailed, bivariate) were calculated for CSF marker levels and blood sugar, proteinuria, systolic blood pressure and body weight as expressed in BMI. Correlations with either stress or physical exercise values were established using the Spearman's rank correlation coefficient. Statistical significance was deemed when p < 0.05.

Data availability
Supplementary Information in an Additional File 1 is available in persistent web link to datasets. Further data and materials are available from the corresponding author on reasonable request.