Revolution drives Evolution – from measuring to understanding: Annual meeting of Swiss Society of Clinical Chemistry (SSCC) in Bern, November 15-16 2018

Collection of accepted scientific abstracts

De Gruyter | Published online: November 1, 2018

A novel LC/MS lipidomics workflow for biomarker profiling and biochemical studies

A. Othman1, T. Hornemann1, A. von Eckardstein1

1 Institute of Clinical Chemistry, University Hospital Zurich, Zurich, Switzerland

Lipids are one of the most structurally diverse classes of metabolites. Lipidomics is the emerging technology that captures this diversity by identifying and quantifying the different lipids species in any particular sample. Lipidomics has proven very powerful to identify several candidate diagnostic and prognostic biomarkers in metabolic, cardiovascular, neurodegenerative diseases and cancer. We have developed a novel liquid chromatography-mass spectrometry (LC/MS) workflow for hit enriched un-targeted lipidomics of biofluids, tissues and cell lines. Lipid extraction is achieved using one of 15 different protocols with different ratios of organic solvents such as methanol, methyl tert-butyl ether, chloroform, butanol and water. The extracted lipids are separated using C30 reverse phase chromatography using acetonitrile:water (6:4) with10mM ammonium acetate and 0.1% formic acid as mobile phase A and Isopropanol: Acetonitrile (9:1) with 10mM ammonium acetate and 0.1% formic acid as Mobile phase B. The liquid chromatography is coupled to a hybrid quadrupole-orbitrap mass spectrometer (Q-Exactive). A data-dependent acquisition ( top 10) with positive and negative polarity switching is used with iterative inclusion and exclusion lists. Identification of the lipids is achieved using two in-house lipid fragmentation libraries; an experimentally validated in-silico lipid library( >500000 lipid species) and a manually curated library for the lipids detected in biological samples ( >1500 lipid species). The lipid libraries contain all major classes of lipids; fatty acids, glycerolipids, glycerophospholipids, sphingolipids, and glycolipids. These databases are customizable and could be expanded with more lipid classes once standards for fragmentation are available. Lipid identification is performed using 1) high accuracy and resolution with an accuracy within m/z within 5 ppm shift form the predicted mass 2) Isotopic pattern fitting to expected isotopic distribution. 3) comparing the expected retention time to the curated database and 4) the specific fragments from the in-silico fragmentation library. Quantification is done using single point calibration with at least one internal standard per lipid class. Using this workflow, we can detect and quantify 200-400 lipid species. This workflow has been applied to several studies of inherited metabolic disorders in lipid metabolizing enzymes, biomarker profiling and therapeutic target identification.

Development of a LC-MS/MS method for the quantitation of sulfur-containing amino acids in human plasma and CSF

A. Rassi-Faerd1, M. Hersberger1,2, A. Cremonesi1

1 Division of Clinical Chemistry and Biochemistry, Children’s Research Center, University Children’s Hospital Zurich, Switzerland2 Zurich Centre for Integrative Human Physiology, University of Zurich, Zurich, Switzerland

Homocysteine (Hcy) and its closely related metabolite methionine (Met) play a central role is several biochemical processes. Met, via its interconversion into S-adenosylmethionine (SAM), acts as a methyl group donor for the synthesis of many compounds like creatine, choline and epinephrine as well as for the methylation of DNA, RNA and proteins. Met and Hcy are also a main source of sulfur necessary for the synthesis of sulfur-containing amino acids like cystathionine, cysteine and taurine. Thus, it is not surprising that abnormalities in the metabolism of Met and Hcy can lead to various disorders, ranging from relatively benign like cystathionuria to very debilitating diseases like cystathionine beta-synthase deficiency.

Several analytical methods exist for the measurement of Met, total homocysteine (tHcy), cystathionine and cysteine; however only a few laboratories are measuring SAM and SAH, which are abnormally increased in some rare disorders like glycine-N-methyltransferase deficiency, SAH-hydrolase deficiency or adenosine kinase deficiency. Thus, it is plausible that such disorders are currently under-diagnosed. To diagnose for such rare disorders, we established a method based on LC-MS/MS to quantify SAM, SAH, tHcy and Met in plasma and CSF. Overall, good precision was obtained for all 4 analytes (inter-day CV<10%). The sensitivity of the method enables the quantification of (patho)physiological concentrations with a LLOQ<1.5 nmol/L for SAM and SAH resp. a LLOQ<0.05 μmol/L for Met and tHcy in both plasma and CSF. The accuracies of the measurements of Met and tHcy have been assessed by measuring external quality control samples showing results within ±2SD. Lacking external quality control samples for SAM and SAH, an inter-laboratory comparison was performed, which revealed a good agreement between the laboratories for SAM (slope=0.90, r2=0.994, n=20) and SAH in plasma (slope=0.90, r2=0.982, n=20). In addition, the accuracies of Met, tHcy, SAM and SAH in plasma and CSF were assessed by measuring their recoveries after spiking, ranging from 85% to 120%. To assess pre-analytical influences, we investigated the stability of the analytes at room temperature, which revealed that tHcy, SAM and SAH are instable in whole blood and plasma over a prolonged time (>90 min.). However, when plasma samples are stored at -20 °C immediately after collection, they are stable for more than 1 year.

Measuring very low concentrations of estrogens in plasma by online SPE LC-MS/MS analysis

L. Stoob1, A. Cremonesi1, M. Hersberger1,2

1 Division of Clinical Chemistry and Biochemistry, Children’s Research Center, University Children’s Hospital Zurich, Zurich, Switzerland; 2 Zurich Centre for Integrative Human Physiology, University of Zurich, Zurich, Switzerland

Estrogens are essential compounds for the development and regulation of the female reproductive system and for the establishment of secondary sex characteristics. To date there are many genetic or acquired disorders affecting estrogen metabolism and homeostasis. It is therefore crucial to accurately measure the concentrations of both estradiol and estron in plasma. This is challenging due to their low concentrations, especially in pediatric samples, and because of the concomitant presence of other interfering steroids in the matrix. Nowadays, most clinical laboratories are using immunoassays for the routine analysis of estrogens, which are often suffering from low specificity and sometimes even low sensitivity. The presented LC-MS/MS method is based on a simple and inexpensive derivatization reaction followed by online solid-phase-extraction to improve the analytical sensitivity and specificity. The final method requires only 100 μL of plasma, it is not labor intensive and fast (total LC-MS/MS time: 7 min). The method is linear (r2>0.99), has a broad dynamic range (0.01 – 10.0 nM), is sensitive (LLOQ=15 pM for estradiol and 13 pM for estron), is precise (CV<5.8% for estradiol and<4.0% for estron) with no sample carry-over. The method accuracy was assessed by analyzing 27 external quality control samples (RfB and Chromsystems) with established consensus concentrations: 100% of the measurements were within ±2σ for estradiol. The LC-MS/MS results correlated well with those from immunoassays for estradiol (y=-0.033 + 0.996x; n=24), while the LC-MS/MS estron measurements were slightly lower (y=58 + 0.814x; n=22) suggesting a lack of specificity of the immunoassay.

Induction of aquaporin 9 in patient-derived iPSC-hepatocytes enables modeling and drug screening for urea cycle disorders

A. Laemmle1,2,3; J. Underhaug4; J. Robinson1; G. Allegri5; A. Keogh6; J.M. Nuoffer2,3; B. Thoeny5; A. Martinez4; J. Häberle5; H. Willenbring1

1 University of California, San Francisco; 2 University Children’s Hospital, Bern; 3 University Institute for Clinical Chemistry, Bern; 4 University of Bergen, Norway; 5 University Children`s Hopital, Zurich; 6 University Hospital, Bern

Background: Reprogramming of patient-derived skin fibroblasts into induced pluripotent stem cells (iPSCs) followed by directed differentiation into hepatocytes (iPSC-Heps) allows modeling of genetic liver diseases in vitro. The urea cycle disorder ornithine transcarbamoylase (OTC) deficiency (OTCD) frequently leads to liver disease. We aim to develop an iPSC-Hep-based model of this disease that could be used to investigate pathomechanisms and develop therapeutics.

Methods: We generated iPSCs from fibroblasts of several OTCD patients and controls and differentiated them into iPSC-Heps. We compared urea cycle enzyme expression and ammonia and urea metabolism between patient-derived and normal iPSC-Heps and primary human hepatocytes (PHH). We included human fetal and adult liver tissue as additional controls.

Results: Differentiation of OTCD patient-derived iPSCs into iPSC-Heps revealed a disease-specific phenotype with reduced OTC activity, ammonia metabolism and urea secretion. However, urea secretion was also low in control iPSC-Heps, which expressed all urea cycle enzymes at levels comparable to PHH, even after ammonia challenge. iPSC-Heps have previously been shown to correspond to immature fetal hepatocytes rather than mature adult hepatocytes. Therefore, we compared gene expression profiles of human fetal and adult liver tissue to identify differentially expressed genes that may be responsible for the low urea secretion in iPSC-Heps. We found that fetal liver tissue was lacking expression of aquaporin 9 (AQP9), which is required for urea secretion. AQP9 was absent in all of our iPSC-Hep lines, consistent with their fetal state of differentiation. In accord, we found higher intracellular concentrations of urea in control iPSC-Heps than in PHH.

Conclusions: These findings show that patient-derived iPSC-Heps replicate characteristic features of OTCD. A limitation of this model is low urea secretion by iPSC-Heps because of their fetal state of differentiation. We identify AQP9 as a target for overcoming this deficiency and realize the potential of iPSC-Heps for modeling of OTCD.

Integrating laboratory data into a clinical data warehouse

M. Dahlweid1,2, M. Kämpf1,2, D. Furrer1,3, A. Leichtle1,3,4

1 IDSC – Insel Data Science Center, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland; 2 Directorate of Technology and Innovation, Inselspital, Bern University Hospital, University of Bern; 3 Directorate of Teaching and Research, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland; 4 University Institute of Clinical Chemistry, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland

Background: Everyday hospitals generate huge amounts of health data: from simple epidemiological to medication, from imaging to ICU monitoring, from reimbursement data up to personal device data as e.g. handheld blood glucose monitors. Along with the rise of “data science” as an emerging discipline and to keep pace with the rapid evolution of information technologies, hospitals, funding agencies and nation-wide initiatives started to foster and to coordinate activities and to build service infrastructures to meet the future demands of clinical data scientists. In the CDWH context, lab testing information becomes especially important. As mainly structured data, importing seems trivial, and data requests frequently comprise laboratory information - as performance indicators in clinical studies, as security markers or even as quality markers in the reimbursement process. Access, assessment, and analysis of these data was a cumbersome, time-consuming, and frequently frustrating process as the source systems were neither intended nor designed to be implemented in large scale data processing systems.

Methods: Integrating laboratory data, we are now facing many challenges: different laboratories, even within one hospital use various lab information systems, master data are set up differently, no common nomenclature exists, reports are written in different languages, changes are rarely and -if at all- frequently fragmentary documented and communicated, the identification of a “valid” result status depends on a variety of factors. Laboratory data has not only to be linked to patients, but also to clinical studies, external customers or devices. Sources of laboratory can be routine labs, scientific centers, POCT devices, or external sources.

Results: With the Insel Data Science Center (IDSC), the Inselspital (Bern University Hospital) provides a service interface for the hospital (e.g. researchers, administration, controlling) to gate, facilitate, and duly document access to the resources of the CDWH. The IDSC processes the data requests, discusses together with the researchers, which information exactly is needed, and translates the request in an IT compatible query/script.

Conclusions: In this unique combination of a research-oriented service unit with an IT-powered data resource, the Inselspital creates a powerful tool for digital healthcare and research to be ready for future challenges in medicine and data science.

Genetics of metamizole-induced agranulocytosis: Development of a resequencing method for HLA genes

A.L. Cismaru1, D. Rudin2, L. Grimm1, C.R. Largiadèr1, M. Haschke3, U. Amstutz1

1 University Institute of Clinical Chemistry, Inselspital Bern University Hospital; 2 Department of Clinical Pharmacology & Toxicology, University Hospital Basel; 3 Department of Clinical Pharmacology & Toxicology, Inselspital Bern University Hospital

Background: Metamizole-induced agranulocytosis (MIA) is a rare and potentially fatal adverse drug reaction (ADR) that leads to a loss of granulocytes. As the Human Leukocyte Antigen (HLA) system has been strongly associated with agranulocytosis caused by other drugs, the aim of our case-control association study is to identify risk variants for MIA by a targeted resequencing of HLA genes using high-throughput sequencing (HTS) and a genome-wide association analysis.

Methods: HLA locus-specific primers were used to amplify full-length class I genes (HLA-A,-B and -C) and class II genes HLA-DRB1, -DQB1, -DQA1,-DPA1, and -DPB1 via long-range polymerase chain reaction (PCR). For the HLA-DQB1 and -DRB1 loci, novel modified primers were evaluated to reduce allelic imbalance. For the method validation, all eight HLA loci were analyzed by Illumina MiSeq-based multiplex paired-end sequencing in 10 previously HLA-typed individuals.

Results: The HTS method generated 3 million paired-end reads of sufficient quality and volume for all 80 amplicons. The typing of the 160 sequenced HLA alleles showed that allele calls at HLA-A, -B, -C, -DQA1, -DPA1 and -DPB1 were 100% concordant with the reference typing. For HLA-DRB1 and -DQB1, allelic dropout and co-amplified non-DRB1 genes were observed with both evaluated primer sets, but were more pronounced with the primers used initially (65% and 80% concordance with reference typing for HLA-DRB1 and -DQB1, respectively). We observed an improved typing accuracy for the HLA-DRB1 (95%) and -DQB1 (100%) loci when amplified with the novel modified primers.

Conclusions: To this date, no other studies have investigated the genetic basis of MIA using current genomic technologies. We have benchmarked an efficient 6-digit high-resolution eight-locus HTS HLA resequencing method that will enable the unmasking of potential susceptibility markers in the HLA region in our cohort and may also be useful for other pharmacogenetic studies.

Genetic case-control association study of metamizole-induced agranulocytosis

A.L. Cismaru1, D. Rudin2, C.R. Largiadèr1, M. Haschke3, U. Amstutz1

1 University Institute of Clinical Chemistry, Inselspital Bern University Hospital, 2 Department of Clinical Pharmacology & Toxicology, University Hospital Basel 3 Department of Clinical Pharmacology & Toxicology, Inselspital Bern University Hospital

Background: Genome-wide association studies (GWAS) have been a crucial determinant in identifying genetic variants that are associated with adverse drug reactions (ADRs). Metamizole-induced agranulocytosis (MIA) is a rare and potentially fatal ADR that leads to a loss of granulocytes. As previous GWAS have uncovered genetic associations for agranulocytosis caused by other drugs, the aim of our study is to identify risk variants for MIA by combining a candidate gene approach and an exome genotyping array-based GWAS.

Methods: We performed a case-control study to investigate the genetic risk of MIA using 550 601 single-nucleotide polymorphisms (SNPs) genotyped on the Infinium CoreExome Array (Illumina). Our cohort comprised of MIA patients (N=54), tolerant controls (N=39) and unexposed controls (N=162) was recruited at two sites in Switzerland (Basel and Bern). Standard data quality control (QC) procedures were applied using Illumina GenomeStudio and PLINK. After linkage-disequilibrium (LD) pruning and merging of genotypes with HapMap (Phase III) data, we used multidimensional scaling (MDS) to detect outliers of non-European ancestry and account for population stratification. Relevant candidate genes were prioritized based on findings from previous publications and mechanistic studies.

Results: After QC, 285 365 SNPs remained (>1% minor allele frequency, >98% call rate, Hardy-Weinberg Equilibrium p-value >0.001) and only 1 individual was excluded due to low genotyping quality (>2% missingness). The overall genotyping rate in the remaining individuals was 0.998. In the MDS analysis, a small number of individuals (n=9) were identified as outliers (non-Europeans). Candidate genes including NOX3, NAT2, GSTs, GPxs and NQO2 were identified based on previously published associations, whereas HMOX1 and MPO were selected based on mechanistic studies. Preliminary results from association testing of these candidate genes identified an intronic SNP (rs10041917) in the serine incorporator 5 (SERINC5) gene as suggestive of an association with MIA.

Conclusions: To date, no other studies have investigated the genetic basis of MIA using current genomic technologies. The top SNP identified in the candidate gene analysis is in LD with a variant previously associated with sulfasalazine-induced agranulocytosis. The identification of such potential susceptibility markers for MIA may provide novel insights into the underlying mechanisms and may enable the identification of at-risk patients.

Detection of glycosaminoglycans by UHPLC-MS/MS - a feasibility study

L. Arackal1, J.C. Prost1, K. Freiburghaus1, J.M. Nuoffer1

1 University Institute of Clinical Chemistry, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland

Background: Glycosaminoglycans (GAGs) display a variety of important biological roles. GAGs are used as diagnostic markers of rare inherited lysosomal storage disorders such as mucopolysaccharidoses (MPS) but also involved in processes such as inflammation and tumor metabolism. GAGs can be categorized into four main groups: heparan sulfate (HS), keratan sulfate (KS), chondroitin sulfate (CS) and dermatan sulfate (DS). GAGs are usually analyzed by different methods, which tend to be unselective or insensitive especially in mild alterations.

Aims: Evaluate a method based on ultra high performance liquid chromatography coupled to tandem mass spectrometry (UHPLC-MS/MS) to analyze DS, HS, KS, and CS and test if urine from MPS patients would be recognized.

Method: Preliminary data was generated analyzing commercially available standard urine of healthy adult donors and pathological samples with MPS type II, III and VI.

Standards of individual GAGs (50 μg/mL) and control urines were processed into disaccharides through methanolysis reaction. Measurements were performed on a Xevo TQ-S in positive electrospray ionization mode using multiple reaction monitoring. Transition used were as followed: DS/CS=m/z 426<236; HS=m/z 406<245; KS=m/z 420<258.

Results/Conclusions: UHPLC-MS/MS analysis of GAGs is feasible under physiological concentrations. The pathological samples were clearly differentiated from controls. DS and KS showed 2-5 fold changes, HS presented 20-400 fold change in absolute intensity depending on analyzed MPS types. The measured GAG patterns of MPS patients reflect similar trends as described in literature. The method will be further validated for clinical application.

High-sensitive cardiac troponin T as a marker of hemorrhagic complications in anticoagulated elderly patients for non-massive pulmonary embolism

A. Simona1, M. Mean Pascual2, A. Limacher3, P. Lescuyer1, H. Bounameaux1, D. Aujesky2, M. Righini1, N.Vuilleumier1

1 Geneva University Hospitals; 2 Bern University Hospital; 3 CTU Bern

Background: Recent data have raised concerns about the risk/benefit ratio of thrombolysis in non-high risk pulmonary embolism (PE) patients due to increased serious bleeding events. Whether cardiac biomarkers could be of help for bleeding risk stratification in these settings remains elusive.

Objectives: To determine the prognostic accuracy of hs-cTnT, NT-proBNP, RIETE and PESI score in the occurrence of clinically relevant bleeding in elderly patients under conventional anticoagulation therapy for non-massive pulmonary embolism (NMPE).

Methods: In the context of the SWITCO65+ cohort, we evaluated 230 elderly patients with available blood sample taken within one day from diagnosis. The primary study endpoint consisted in clinically relevant bleeding at 1, 3 and 24 months. Prognostic accuracies were determined using C-statistics, Cox regression and survival analysis.

Results: hs-cTnT displayed the highest prognostic accuracy (C-statistics: 0.77 - 95% CI: 0.68–0.88 at 1 month) which remained stable over time. Although AUC comparison indicated that hs-cTnT was not superior to RIETE score, risk analyses showed that association of hs-cTnT with clinically relevant bleeding was independent of RIETE score. At the standard pre-specified cut-off of 14 ng/l, negative predictive value of hs-cTnT was 96.9% (95% CI: 91.4-99.0) at 1 month and remained stable at 3 months.

Conclusions: In elderly, low hs-cTnT levels could represent a valuable tool to identify NMPE patients at low risk of bleeding potentially susceptible to benefit from thrombolysis.

Serum free light chain assay comparison for the therapeutic management of patients with monoclonal gammopathies

B. Rindlisbacher1, C. Schild1, F. Egger1, A. Leichtle1, N. Sédille-Mostafaie1

1 University Institute of Clinical Chemistry, Inselspital, Bern University Hospital, and University of Bern, Bern, Switzerland

Background: Analysis of к and λ free light chains (FLC) has largely improved the laboratory evaluation of patient samples with suspecting monoclonal gammopathy. The first available commercial assay included polyclonal reagents, while a more recently developed assay incorporated monoclonal antibodies to detect FLCs. We aimed at comparing the performance of both assays on the same analytical platform to improve the monitoring of FLC serum levels.

Methods: Analytical performance of the polyclonal Freelite (Binding Site) and the monoclonal N Latex FLC assay (Siemens) was evaluated on a BN ProSpec nephelometer (Siemens). The methods were subsequently compared using 102 clinical serum samples.

Results: Method evaluation revealed acceptable overall imprecisions (CV%) and inaccuracies (%bias) below 4.4% and 12.9%, respectively. In the 102 clinical samples, the observed concentration ranges for FLC к and λ were 4.4-6000 mg/L and 0.6-1060 mg/L, respectively. Passing-Bablok regression analysis yielded the following linear equations: FLC к Freelite (mg/L)= -3.09 + 1.24 N Latex (mg/L); FLC λ Freelite (mg/L)=0.64 + 0.63 N Latex (mg/L); and к / λ Freelite=-0.72 + 2.28 N Latex, respectively, showing significant proportional bias for FLC к, FLC λ and к / λ ratio comparisons. In addition, a significant constant bias was found for FLC к. Qualitative comparison of six samples with discrepant FLC ratio results between the assays showed that the results from the N Latex assay fit better to corresponding immunofixation results than those from the Freelite assay.

Conclusions: The nephelometric FLC к and λ N Latex and Freelite assays showed acceptable method evaluation measures. However, method comparison failed to show acceptable agreement between methods. Thus, patients should be monitored using the same method.

Combining lab developed procedures and industrial test design: Validation of a LC-MS/MS assay for immunosuppressive drug TDM

B. Näscher1, C. Timm1, D. Siefert2, M. Risch1,3, M Shipkova4, C. Seger1

1 labormedizinisches zentrum Dr Risch, Buchs SG, Switzerland 2 Zentralinstitut für Klinische Chemie und Laboratoriumsmedizin, Klinikum Stuttgart, Stuttgart, Germany 3 Kantonsspital, Chur, Switzerland 4 MVZ Leinfelden-Echterdingen GmbH, SYNLAB Holding Deutschland, Leinfelden-Echterdingen, Germany

Background: Traditionally LC-MS/MS based immunosuppressant drug TDM (ISD-TDM) in whole blood relies on dedicated assay designs and is thought to be incompatible with serum based TDM. However, since LC is an analytical science with long tradition and very well understood mechanisms of action, the transferability of a separation setup for a given set of analytes (here ISDs) to another assay format seams feasible – especially if no high-resolution LC separation is needed. Hence, the addition of whole-blood ISD-TDM to a commercial serum based TDM assay panel realization was undertaken

Methods: To realize ISD-TDM the MS/MS dimension of the IVD-CE certified ClinMass® TDM Kit (Recipe, Germany) was expanded with SRM transitions for cyclosporine A, tacrolimus, sirolimus, and everolimus (CIC, TAC, SIR, EVE, respectively). Mobile phase and stationary phase of this assay were used unchanged and an isocratic mobile phase elution regimen was optimized for ISD-TDM. The sample preparation protocol was transferred with changes from the ISD-TDM Kit System provided by the same vendor, calibration and quality control was done with IVD-CE certified materials by Recipe.

Results: The realized LC-MS/MS assay was found to provide results which were precise (inter-assay CV of QC: 2.7 – 7.0%) and true (90% of 166 tested PT samples within ±15% of target). No matrix or ion-suppression effects were observed, interference testing with several dozen drugs did not show any signs of additional signals. Sufficient sensitivity (LLOQ CIC 10 μg/l, TAC, SIR, EVE 0.5 μg/l) guaranteed measurability of clinical samples; a platform comparison study (120 randomly chosen anonymized residual patient samples from a clinical environment) showed sufficiently good results with only minor absolute mean deviations which are clinical insignificant (CIC 6.5 μg/l, TAC, SIR, EVE<±0.6 μg/l).

Conclusions: Traditionally, ISD-TDM from whole blood is seen isolated from other TDM assays, e.g. for psychoactive drugs. With our approach we did prove that ISD-TDM can be easily embedded into the setting of a multi-assay serum TDM platform devoted to offer a daily measurement service for dozens of other analytes.

Disclaimer: The presented lab developed test must be considered a research use only (RUO) application, this presentation must not be mistaken as performance evaluation study in accordance with the current IVD regulations (IVDR 2017/746).

Evaluation of an IVD-CE certified LC-MS/MS plasma metanephrine assay

V. Braun1,2, C. Timm2, M. Risch2,3, C. Seger1,2

1 Leopold-Franzens-Universität, Innsbruck, Austria 2 labormedizinisches zentrum Dr Risch, Buchs SG, Switzerland 3 Kantonsspital, Chur, Switzerland

Background: We evaluated the performance of an IVD-CE assay kit (Chromsystems, Germany) for the determination of free metanephrine congeners (metanephrine (MN), normetanephrine (NMN) and 3-methoxytyramine (3-MT)) in plasma with the aim to introduce the assay into routine operation.

Methods: All assay materials were provided by the manufacturer. Sample preparation was performed according to the SOP of the manufacturer. Assay performance claims were verified by evaluating the key parameters sensitivity (LLOQ), selectivity, trueness, and precision. The LLOQ (inter-assay precision <20%) was determined by multiple measurements of diluted QC-samples. Precision was determined by repeated analysis of internal QC samples. Accuracy was assessed by comparing analysis results with results from an established measurement service (CHUV, Switzerland) and by inter-laboratory comparison in an international proficiency testing (PT) scheme (RCPA, Australia).

Results: LLOQs were found to be 0.01 nmol/l for MN, 0.05 nmol/l for NMN and 0.03 nmol/l for 3-MT. Inter-assay precision was determined at the lowest QC-level (n=10, CV) as 5.5 % (MN), 6.0 % (NMN), and 10.8 % (3-MT); the corresponding bias was negative (better than –5.8%). All PT results (n=14) were within the acceptance limits. A positive assay bias (better than 5.0 %) to the mean PT results was found. When comparing assay results (n=40) to data generated by the CHUV laboratory relative Bland-Altman bias values of +2.5 % (MN), +11.8 % (NMN), and +4.7 % (3-MT, n=25, results below LLOQ excluded) were found. The corresponding Passing Bablok regression slopes were 1.05 (MN), 1.28 (NMN), and 1.09 (3-MT).

Conclusions: LLOQ, internal QC trueness and precision are in line with data reported by the manufacturer. PT participation proved the assay trueness to be sufficiently good although a constant positive bias of the assay was found. This finding was substantiated by results of an inter-laboratory comparison. Therefore, it can be hypothesized that a small most likely clinically insignificant (though statistical significant) calibration bias may exist between individual assay realization. We are convinced that this situation must be carefully monitored and that standardization of individual measurement services by establishing a chain of traceability to higher order reference materials and higher order measurement procedures should be envisioned as a future aim.

NMR for online metabolomics in 3D cell culture

D. Hertig1,2,3, A. Felser2, G. Diserens1, S. Kurth2, P. Vermathen1, J.M. Nuoffer2

1 Departments of Clinical Research and Radiology, University of Bern, Bern, Switzerland, 2 Institute of Clinical Chemistry, University Hospital Bern, Bern, Switzerland, 3 Graduate School for Cellular and Biomedical Sciences, University of Bern, Bern, Switzerland;

Background: NMR allows non-destructive analysis of living cells and can provide a unique molecular-window into unprecedented information on the living metabolome. NMR approaches using flow systems for real-time spectroscopy were explored as early as the 80s but did not become routine due to different technical limitations. Many of these limitations are being overcome and novel flow units for monitoring living cells are being developed. The possibility to analyse intracellular metabolome and the effect of culture media or drugs under culture condition allows NMR to become a key tool in the better understanding of cellular pathomechanisms.

Aims: A) To establish a perfused NMR bioreactor system to study short-term metabolic changes in living cells (fibroblasts (FB), 3D human brain cells). B) To investigate longitudinal and fast dynamic metabolic processes upon drug treatments, toxicological, substrate or inhibitor challenges.

Methods: 5 million FB control cell lines were measured in collagen based 3D scaffold (Matrigel™). 1H NMR experiments were performed on a 500 MHz Bruker Avance II spectrometer. HPLC pump was used for constant perfusion of substrate via perfusion apparatus InsightCell™. Atmosphere of 5% CO2 and 47% O2 was applied to the bioreactor using self-constructed gas blending system in order to maintain pH and supply sufficient oxygen. Online metabolic profiling was performed using 1D project spectra of 48 transients (ca. 7min) per spectrum. Cell viability was controlled by flow cytometry and trypan blue staining.

Results: A) Density and composition of 3D Scaffold + FB was evaluated and protocols of feasible workflow were established, to transfer 3D scaffold from cell culture into the sensitive region of the NMR spectrometer causing minimal physical damage of the 3D network. Metabolic profiling under standard cell culture condition allowed kinetic tracking of over 30 metabolites and kept stable over 12 hours. B) Serial stop-and-go flow interventions (6x) over 12 hours of measurement time showed fully reproducible metabolic response in its form and extent. Addition of rotenone during measurement showed alteration in glucose consumption and lactate production, which could be reversed after additional treatment of glycolysis inhibitor 2-deoxyglucose.

Conclusions: Preliminary experiments demonstrate the technical feasibility of metabolic monitoring in 3D culture inside NMR up to 12 hours and caption of cellular response to tested interventions.

1P-LSD: detection of a new psychoactive substance using a novel LC-HRMS method

D. Mueller1, K. Hofer-Lentner2, A. Gross3

1 Institute of Clinical Chemistry, University Hospital Zurich, 2 National Poisons Centre, Tox Info Suisse, associated Institute of the University of Zurich, 3 Spital Linth

Background: 1-propionyl-LSD (1P-LSD) is a new psychoactive substance (NPS) with LSD-like properties. Very few data of its usage in humans is known. We describe the case of a 31-year old caucasian male who developed hallucinations and undesirable effects such as palpitations, sweating, restlessness, tremor, dizziness, and dry mouth one hour after self-declared ingestion of 200 μg of 1P-LSD. He presented at the emergency department 6-7 h after intake of the substance.

Methods: We obtained serum and urine samples for analysis. Samples were analyzed using a fast and sensitive liquid chromatography - high resolution mass spectrometry (LC-HRMS) method. Analysis of 1P-LSD was verified using a 1P-LSD blotter with confirmed identity. In addition, the new method was compared to an established screening procedure using >30 real patient samples.

Results: In urine and serum samples of our case, several substances were identified in addition to LSD and metabolite. 1P-LSD itself could not be identified, nor specific metabolites of 1P-LSD, despite successful identification of 1P-LSD in the blotter with confirmed identity. The new LC-HRMS screening procedure proved its advantage over the established screening procedure by identifying considerably more substances that could be confirmed by other means, if technically possible.

Conclusions: Very little is known about the metabolism of 1P-LSD. As many other N1-substituted LSD-derivatives are metabolized by dealkylation, it is speculated that 1P-LSD may be a prodrug and undergo rapid dealkylation to LSD in vivo [1]. Our data seem to support this hypothesis, assuming that the report of the patient having consumed 1P-LSD was correct. As several other substances were present in the patient samples, symptoms of the patient could not clearly be assigned to the influence of 1P-LSD. The new LC-HRMS screening procedure proved its advantage over the established screening procedure in terms of speed and sensitivity.

Reference: [1] Brandt S. et al. Drug Test Anal. 2016; 8(9): 891–902

New LC-MS/MS method for the diagnosis of inborn errors of metabolism presenting with early onset epilepsies

D. Mathis1; K. Beese1; C. Rüegg1; M. Hersberger1

1 Division of Clinical Chemistry and Biochemistry, Children’s Research Center, University Children’s Hospital Zurich, Zurich, Switzerland

Background: Selective screening for inborn errors of metabolism (IEMs) includes several profile analyses in addition to a multitude of tests that quantify a single or several analytes. This strategy can lead to long turn-around times and is inefficient for the laboratory. Here, we present a new LC-MS/MS method that combines 5 methods for the diagnosis of IEMs presenting with early-onset epilepsies.

Methods: The new method was developed to simultaneously quantify 12 metabolites (sulfocysteine, guanidinoacetate, creatine, pipecolic acid, Δ1-piperideine-6-carboxylate (P6C), proline, Δ1-pirroline-5-carboxylate, and the B6-vitamers) enabling the diagnosis of 9 different IEMs (sulfite oxidase deficiency, molybdenum cofactor deficiency, GAMT deficiency, AGAT deficiency, creatine transporter deficiency, antiquitin deficiency, hyperprolinemia type II, PNPO deficiency and hypophosphatasia).

Plasma, urine, and cerebrospinal fluid samples were mixed with internal standards, precipitated and the supernatants were analyzed on a liquid chromatography mass spectrometry system (ThermoScientific UltiMate3000 coupled to SCIEX TripleQuadTM5500). The analytes were detected by MRM in positive ionization mode.

Results: The LC-MS/MS method was fully validated. All metabolites give quantitative results except for P6C (semi-quantitative). Determination coefficients showing linearity are above 0.99 for all quantitative metabolites. Coefficients of variation are below 15% and recovery rates range between 80-120%. Furthermore, analysis of patient samples with above-mentioned IEMs led to the correct diagnosis. Unfortunately, no samples of patients with sulfite oxidase deficiency, AGAT-deficiency or hyperprolinemia type II were available.

Conclusions: The new LC-MS/MS method improves the efficiency in the laboratory and will potentially lead to quicker diagnoses and earlier treatment avoiding irreversible damage in patients affected with IEMs.

Fatty acid profile in erythrocytes by GC-MS/MS

D. Mathis1,*; P. Srikanthan1,*; N. Perkins1; M. Hersberger1

1Division of Clinical Chemistry and Biochemistry, Children’s Research Center, University Children’s Hospital Zurich, Zurich, Switzerland; *equal contribution

Background: The fatty acid profile in erythrocytes (EC) is of use in assessing the nutritional status for the last three months, especially the supply with omega-3 and omega-6 fatty acids (FA), for therapy control during omega-3 FA supplementation, and for an estimation of the intake of trans-FA. Low levels of omega-3 FA, in particular docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) have been associated with neurological pathologies including depression and psychotic disorders, NAFLD, rheumatoid arthritis and cardiovascular diseases.

Methods: We present a GC-MS/MS method for the quantification of 27 FA, including saturated-FA, omega-3 FA, omega-6 FA and trans-FA. ECs were collected from EDTA-blood samples and the FA extracted and derivatized with acetyl chloride. These derivatives were then analyzed on a gas chromatography mass spectrometer system (Thermo Scientific Trace1310 coupled to TSQ8000 Triple Quadrupole) and detected by MRM in positive ionization mode.

Results: The new GC-MS/MS method was fully validated. Determination coefficients showing linearity are above 0.99 for all metabolites. Coefficients of variation are below 13% and recovery rates range between 83-113%, except for the low abundant trans-FA, for which the coefficients of variation are below 23% and the recovery rates between 99-122%. In addition, we established reference values from 257 EC samples from children aged 0-18 and from 152 EC samples from adults. No age dependency was observed for the fatty acids in EC.

Conclusions: The GC-MS/MS method established in our laboratory can be used to assess the nutritional supply with omega-3 FA and to control for omega-3 FA supplementation.

Characterization of uracil catabolism variability in healthy volunteers and cancer patients

D. Schärer1,2, D. Kummer1, C. Nakas3, S. Fontana4, M. Joerger5, U. Amstutz1; C.R. Largiadèr1

1 Institute of Clinical Chemistry, Inselspital, Bern; 2 Graduate School for Cellular and Biomedical Sciences, Bern; 3 Department of Agriculture Crop Production and Rural Environment, Nea Ionia; 4 Regional Blood Transfusion Service of the Swiss Red Cross, Bern; 5 Department of Medical Oncology and Hematology, Cantonal Hospital, St. Gallen

The catabolic pathway of uracil (U) mainly determines the pharmacokinetics of the anticancer drug 5-fluorouracil (5-FU). Decreased activity of dihydropyrimidine dehydrogenase (DPD), the first catabolizing enzyme, was identified as a major predictor of 5-FU toxicity. However, phenotypic variability in the catabolism downstream of DPD by dihydropyrimidinase (DHP) and β-ureidopropionase (bUP) and its potential contribution to 5-FU toxicity has only partially been investigated so far. Here, we aimed to characterize the variability of metabolites and metabolic ratios of uracil catabolism and to evaluate their association with genetic variation in the DHP and bUP genes (DPYS and UPB1). Plasma concentrations of U and its metabolites were determined by UHPLC-MS/MS and three variants in DPYS (c.1-1T>C and c.265-58T>C) and UPB1 (c.1-80C>G) previously associated with 5-FU toxicity were genotyped in 320 healthy volunteers. Similarly, endogenous and 5-FU metabolite plasma concentrations were measured by UHPLC-MS/MS in 27 cancer patients. In healthy volunteers, sex-dependent differences were observed along the entire uracil catabolism pathway. In particular, we observed lower concentrations of all metabolites (P ? 0.020) as well as lower β-ureidopropionic acid/dihydrouracil ratios (UPA/UH2; P=0.001) in women, suggesting that reduced catabolism may contribute to higher 5-FU toxicity rates in females. In cancer patients, BAL/UPA ratios were altered during 5-FU infusion (after 2h, P<0.001), indicating that bUP, besides DPD, may also play a rate-limiting role in uracil catabolism, especially in the progress of 5-FU-degradation, as bUP appears to be saturated based on the increasing FBAL plasma concentrations after 2h of 5-FU infusion (P<0.001).

Interferences in thyroid function tests due to anti-streptavidin or anti-sulfo-ruthenium antibodies

D. Werner1, G. Sykiotis1, C. Butticaz1, C. Torrent1, P. Vollenweider1, V. Mooser1

1 CHUV, Lausanne

Background: Thyroid function tests (TFT) are essential for the diagnosis of thyroid conditions and for their monitoring. Assay interferences in TFTs can have serious clinical consequences. Only anecdotal cases of TFT interferences due to antibodies directed against assay reagents have been reported.

Methods: TSH, free T4 (FT4) and free T3 (FT3) were measured by ECLIA using the Roche cobas platform, using the 2nd generation of reagents for FT4. Analytical interference was suspected in presence of high FT4/FT3 and clinically unexplained, unsuppressed TSH. Samples with suspected interferences were re-assayed: 1. on cobas after incubation in Heterophile Blocking Tubes (HBT); 2. by CMIA on the Abbott Architect platform ± HBT pretreatment; 3. when enough sample was available, by the 3rd generation of FT4 reagent; 4. by sending patient serum to Roche R & D (Penzberg/D) in order to document the cause of interference

Results: We identified 8 cases of TFT assay interferences on cobas, 7 of which were due to anti-streptavidin and 1 to anti-sulfo-ruthenium antibodies. For these 8 cases, FT4 levels (47.2 ±18.8 pmol/l, mean ± SD), were above the 97.5 percentile (21.9 pmol/l) of the CoLaus reference population (n =4462), whereas TSH levels were not suppressed accordingly. In the 7 anti-streptavidin interference cases, pretreatment with HBT led to an average 54% reduction of FT4 and an average 55% increase in TSH, consistent with the fact that interfering antibodies were from the IgM subclass. When tested (n=4), FT3 showed a similar behavior as FT4. In contrast, the anti-sulfo-ruthenium interference was from the IgG subclass with no effect of the HBT pretreatment. On the Architect, all but one measurements were in the normal range with HBT showing no effect. With the 3rd generation assay on cobas (n=5), FT4 levels were in the normal range. Interferences appear to be limited in time as, for the 4 cases for whom serial measurements were available, TFT levels returned to normal between 2 and 18 months.

Conclusions: Described here is the largest series of documented TFT interferences due to anti-reagent antibodies. TFT assay interferences due to anti-streptavidin and anti-sulfo-ruthenium antibodies are platform-specific and not excessively rare. These antibodies are most frequently of the IgM subclass and affect the measurements by ECLIA of FT4, FT3 and TSH. Interferences appear to be successfully resolved with the new FT4III assay and to be reversible over time.

Treatment of complex I deficient fibroblasts with CoQ10, idebenone and ascorbate: Effects on respiration and intracellular ROS

F. Burri1, A. Felser1, J.M. Nuoffer1

1 Institute of Clinical Chemistry, University Hospital Bern, Bern, Switzerland

Background: Mutations affecting mitochondrial oxidative phosphorylation (OXPHOS) lead to severe disorders with heterogeneous presentations. A lowered aerobic capacity of respiration and elevated reactive oxygen species (ROS) are often associated with OXPHOS defects. Treatment strategies have aimed at lowering ROS stress with antioxidants and restoring respiration with electron carriers such as coenzyme Q10 (CoQ10) and analogs like idebenone. Primary CoQ10 deficiencies generally benefit from treatments, but for complex I (CI) disorders results are contradictory.

Methods: We treated four CI, one CoQ10 and three control fibroblast cell lines with 5 μM CoQ10, idebenone or ascorbate for one week. We assessed mitochondrial respiration using Seahorse XF live cell analyser and intracellular ROS production using fluorometric DCF-DA assay.

Results: Treatment with CoQ10 increased basal respiration in primary CoQ10 deficient cells whereas idebenone and ascorbate had no effect on respiration. ROS levels were substantially increased compared to controls and were lowered with CoQ10, idebenone and ascorbate, although not reaching control levels. CI deficient cells showed a substantial limitation in maximal respiration, which could not be increased with any of the treatments, instead, idebenone lowered maximal respiration even further. ROS levels were significantly increased compared to controls and were marginally lowered with treatments.

Conclusions: The results indicate that CoQ10 substantially improves mitochondrial respiration in primary CoQ10 deficient fibroblasts, whereas CI deficient fibroblasts might not benefit from the treatments. Further investigations are necessary in order to elucidate the molecular mechanisms in more detail.

Endocannabinoids as putative biomarkers of semen quality?

F. Zufferey1, N. Donzé1, R. Rahban2, A. Senn2, E. Stettler2, S. Rudaz3, S. Nef2, M.F. Rossier1

1 Service of Clinical Chemistry & Toxicology, ICH, Hôpital du Valais; 2 Department of Genetic Medicine & Development, University of Geneva; 3 School of Pharmaceutical Sciences, Universities of Geneva and Lausanne

Background: Semen analysis is crucial in the evaluation of male fertility and remains the most common single diagnostic tool. While the methodology for analyzing sperm count, motility and morphology is well established, the biochemical composition of the seminal fluid remains poorly investigated. Endocannabinoids (ECB) have been recently considered as key players in male reproduction because of their action on human sperm and Leydig cells.

Aims & Methods: The main purpose of our project was to evaluate the relationship between specific ECB and semen parameters. To reach this goal, we investigated the correlation between semen parameters and ECB concentrations among one hundred young men randomly selected from a large cohort of Swiss conscripts. ECB’s were measured by LC/MSMS in blood serum and seminal fluid and results analyzed after stratification according to semen quality.

Results: We did not find any correlation between ECBs present in both blood and seminal fluid, neither between ECBs and semen parameters. ECB concentrations measured in seminal fluid were five to ten times higher compared to ECB concentrations found in blood serum. Palmytoylethanolamine (PEA) and oleoylethanolamide (OEA) appeared positively correlated and anandamide (AEA) levels seemed to be linked to sperm motility. Indeed, individuals with asthenozoospermia (total motility<40 %) showed higher AEA concentrations in both biological fluid as compared to samples collected from men with sperm motility higher than the threshold set by WHO for fertile men. Stratification of individuals into tertiles (T) according to sperm concentration showed a significant increase of seminal fluid PEA levels between T1 and T3 and stratification according to the morphology of spermatozoa showed a significant increase of 2-arachidonoyl-glycerol (2-AG) in their serum.

Conclusions: Our results show that this approach can lead to the identification of new putative candidates that could be later validated as biochemical markers of male fertility and become a useful tool in fertility clinics.

Evaluation of human pancreatic elastase-1 quantification in faeces extracted with a conventional method (weighing) and with the CALEX® cap using the ScheBo® elisa-kit

A. Rösel1; S. Hofmann1; L. Özdemir1; H. Hartung2; F. Frei2; C. Niederberger2; A. Pomazanovs1; R. Mauerer 1

1 SYNLAB MVZ Weiden GmbH; 2 BÜHLMANN Laboratories AG, Schönenbuch BL

Background: Human Pancreatic Elastase-1 (E1) is an important marker reflecting exocrine pancreatic function. E1 has to be extracted from fecal specimens. Due to the nature and inhomogeneity of this specimen the workload in the laboratory is high and reduction of hands-on-time through automation is needed. In this work we have tested the E1 extraction using the automatable CALEX®-System.

Methods: In our laboratory we have evaluated and compared the CALEX® Cap extraction device to the conventional method (homogenization, manual material portioning, extraction, and vortexing). In a second step we had determined the E1-concentration using an ELISA - Plate coated with monoclonal antibodies against E1 („ScheBo® Pancreatic Elastase 1 Testkit“). The sample was incubated with a monoclonal anti-elastase 1-biotin-peroxidase (POD) - streptavidin complex. The peroxidase oxidized the substrate ABTS, which was determined photometrically. 176 stool samples were collected with a stool specimen collector. A sample was homogenized, portioned and extracted with both methods for comparison. The extracts from both methods were measured with the ScheBo® Testkit.

Results: The correlation of the two extraction methods is good: correlation coefficient (r)=0,846. The concordance of diagnostic value among all patients is 90%. The 95% limits of agreement (LoA, Bland-Altman-Plot) contains 95% of the difference scores. The mean difference (bias) of the measurements between CALEX® Cap with ScheBo® ELISA and weighing in with ScheBo® ELISA did not exceed the maximum allowed difference between the methods. Both methods are in agreement and may be used interchangeably.

Conclusions: In this work we showed that both extraction methods may be used interchangeably. This makes the full-automated E1 quantification in stool using the CALEX® Cap extraction possible.

Seeking diagnostic biomarkers for acute aortic dissection

J. Gawinecka1, A. Forrer2, F. Schönrath3, V. Falk4, A. von Eckardstein5

1 Institute of Clinical Chemistry, University Hospital Zurich, Zurich, Switzerland, 2 Institute of Clinical Chemistry, University Hospital Zurich, Zurich, Switzerland, 3 Department of Cardiothoracic and Vascular Surgery, German Heart Institute Berlin, Germany, 4 Department of Cardiothoracic and Vascular Surgery, German Heart Institute Berlin, Germany, 5 Institute of Clinical Chemistry, University Hospital Zurich, Zurich, Switzerland

Background: Acute aortic dissection (AAD) is a rare but life-threatening condition with a lethality rate of 1 to 2% per hour after onset of symptoms in untreated patients. Therefore, its prompt and proper diagnosis is essential to increase a patient’s chance to survive and to prevent grievous complications. Up to now, easily accessible and cost-effective blood tests play only a minor role in the diagnostics of patients with suspected AAD. Such biomarkers may also support monitoring of patients at increased risk for AAD, e.g. patients with Marfan syndrome or with thoracic aneurysm (TAA).

Methods: To identify clinically relevant biomarker candidates for AAD, we chose targeted proteomics based on Proseek multiplex assay with matching pair of antibodies linked to unique oligonucleotides for detection and real-time PCR quantification of predefined biomarker candidates ( In details, we analyzed almost 400 mostly cardiovascular or inflammation-related proteins in plasma samples of patients with AAD, thoracic aneurysm, myocardial infarction, pulmonary embolism, and heart-healthy controls.

Results: In total, 340 proteins could be detected in more than 75% of plasma samples and concentrations of 157 from them were significantly different between patients’ groups in ANOVA test with correction for multiple testing using false discovery rate method. Through further extensive statistical analysis, five biomarker candidates were selected for a validation study using different cohorts of patients and methods for measurement. The selected biomarker candidates include interleukin 6 (IL6), interleukin 10 (IL10), interleukin-1 receptor antagonist protein (IL1ra), insulin-like growth factor-binding protein 1 (IGFBP1), and plasminogen activator inhibitor 1 (PAI1). The data of the validation study are currently validated and will be presented.

Conclusions: Using targeted proteomics, we identified several promising biomarker candidates that may become useful in the management of patients with suspected AAD.

Metabolomics by UHPLC-MS: Benefits provided by complementary use of Q-TOF and QQQ for pathway profiling

K. Freiburghaus1, Ch. Stettler2, G.M. Fiedler1, L. Bally2 and C. Bovet1

1 University Institute of Clinical Chemistry, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland; 2 Department of Diabetes, Endocrinology, Clinical Nutrition & Metabolism, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland

Background: Non-targeted metabolic profiling by UHPLC coupled to high-resolution mass spectrometry (HRMS) is often recognized as reference method for identifying regulated pathways in biological samples because of its unbiased full-scan information. However, HRMS suffers from major limitations: an ease of detector saturation, limited linear range and a complex integration pipeline. In contrast, triple quadrupoles (QQQ) – due to increased sensitivity, reproducibility and acquisition speed - may allow insight into multiple pathways by measuring hundreds of metabolites in a single run. Here, we demonstrate the complementary of the non-targeted and targeted metabolic profiling approaches using UHPLC-MS to characterize exercise-induced changes in the metabolome of patients with type 1 diabetes (T1D).

Methods: In a crossover design, 12 male adults with T1D underwent 90 min of intermittent high-intensity or iso-energetic continuous moderate intensity exercise. Metabolic profiling of serum collected before, during and after exercise was performed by UHPLC coupled to a quadrupole time-of-flight (Q-TOF, Synapt G2-S HDMS, Waters) and QQQ (Xevo TQ-S, Waters) mass spectrometers. Metabolites were separated by reversed-phase chromatography and ionized by electrospray ionization (ESI) in the positive (+) and negative (-) mode. Progenesis QI (Nonlinear Dynamics) and Skyline (MacCoss Lab Software) were used for data integration.

Results: The QQQ method covered 114 metabolites involved in multiple core energy pathways of which 97 were detected in the samples. Applying Q-TOF analysis, 1220 and 540 metabolic features were detected in ESI+ and ESI-, respectively. Amongst the 83 commonly detected metabolites, precision was increased up to 2-fold using QQQ compared with Q-TOF. Up to 61% of the detected metabolites had an excellent linearity over three orders of magnitude (r2 >0.99) with the QQQ, whereas only 29% of features reached this using the Q-TOF. Pathway coverage was improved up to 3-fold by the QQQ, which implied for the T1D cohort 62% metabolite coverage of the purine metabolism and 33% coverage of the citric acid cycle.

Conclusions: Large-scale targeted metabolomics using QQQ showed enhanced precision, linearity and pathway coverage. Thus, the combined use of QQQ and Q-TOF provided deeper insight into exercise-associated metabolism in patients with T1D and may therefore be a promising approach for enhanced metabolic profiling in clinical studies.

Work flow optimization and method comparison of urinary stone analysis by fourier-transform infrared spectroscopy

K. Höland1, R. Schär1, J. Joneli1, M. Reusser1, L. Rohner1, Y. Aebi1, N. Sédille-Mostafaie1, J.M. Nuoffer1

1 University Institute of Clinical Chemistry, Inselspital, Bern University Hospital, and University of Bern, Bern, Switzerland

Background: Urolithiasis is a frequent disease complication with increasing incidence in industrialized countries. Compositional analysis of stones is important for correct diagnosis of pathophysiological mechanisms and adequate treatment. Determination of urinary stone composition by Fourier-transform infrared (FTIR) spectrometry however greatly depends on available reference libraries.

Methods: The incidence of different stone compositions in our laboratory was determined in a retrospective data base analysis (2008–2018, n= 6774, all analyses performed with Spectrum One FTIR spectrometer with potassium bromide (KBr) pellets). For method comparison, each sample was homogenized and measured (a) directly as powder on Spectrum Two FTIR spectrometer with the universal attenuated total reflection (UATR) accessory, (b) as KBr pellet on Spectrum Two or (c) as KBr pellet on Spectrum One. Samples included 68 routine urinary stones and 32 external quality control samples. Stone composition was determined by comparing measured spectra with spectra libraries (highest search score). Further, intra- and inter-assay comparisons (n=10) were performed.

Results: Analyzed lithiases were representative of the retrospective data base cohort. Analysis of KBr pellets on Spectrum One versus Spectrum Two show good agreement, using the same spectra libraries: 99% of samples show the same major component, 85% the same components with differing %-composition, whereby these differences were considered not to be clinically relevant. UATR versus KBr pellet measurements on Spectrum Two showed the same major component in 91% of samples. In the 9% with differing major components, 87.5% were due to differing %-values of components. Overall, 33% showed the same components but with differing percentages. 43% differ due to additional detection of a minor component (mostly <20% (33/49), rarely 20–40% (7/40)). The majority of differences in the 67% of samples with differing stone type compositions can be explained by differences in the reference library (47/67). Changing from KBr pellets to the UATR accessory significantly optimizes the workflow by reduction in turnaround time of 13 min (45%).

Conclusions: In summary, replacing KBr pellet preparation with the UATR technique on FTIR spectrometers is a cost and time effective solution, still allowing acceptable and reproducible results for routine urinary stone analysis. The relevance of differing minor components has to be further evaluated.

Impact of a single oral dose of 100,000 IU vitamin D3 on profiles of serum 25(OH)D3 and its metabolites 24,25(OH)2D3, 3-epi-25(OH)D3, and 1,25(OH)2D3 in adults with vitamin D insufficiency

L. Saleh1, J. CY Tang2, J. Gawinecka1, L. Boesch3, W. D Fraser2, A. von Eckardstein1, A. Nowak3

1Institute for Clinical Chemistry, University Hospital of Zurich and University of Zurich, Zurich, Switzerland; 2 Bioanalytical Facility, Bob Champion Research and Education Building, James Watson Road, University of East Anglia, Norwich Research Park, Norwich, UK; 3 Division of Internal Medicine, University Hospital of Zurich and University of Zurich

Aims: To investigate the effect of a high dose of vitamin D3 on circulating concentrations of 25(OH)D3 and its metabolites 24,25(OH)2D3, 3-epi25(OH) D3, and 1,25(OH)2D3 in healthy individuals with vitamin D insufficiency (25(OH)D3 <50 nmol/L).

Methods: 107 healthy adults (age 20–50 years) with vitamin D insufficiency were randomized to receive a single 100,000 IU dose of vitamin D3 (n= 52) or placebo (n= 55), and vitamin D metabolite concentrations in serum were measured before, and 4 weeks after, supplementation.

Results: Overall, 52% of participants receiving vitamin D3 attained a serum 25(OH)D3 level >75 nmol/L. Among patients who received vitamin D3, there were significant increases in serum concentrations of 25(OH)D3 and its metabolites 24,25(OH)2D3, 3-epi25(OH) D3, and 1,25(OH)2D3 at 4 weeks; however, inter-individual variability in these changes was substantial. Positive correlations between serum 25(OH)D3 and 24,25(OH)2D3 and 3-epi-25(OH)D3, and a significant negative correlation between serum 1,25(OH)2D3 and 3-epi-25(OH)D3, were found 4 weeks after supplementation. The 24,25(OH)2D3/25(OH)D3 and 24,25(OH)2D3/1,25(OH)2D3 ratios were significantly increased, compared with baseline, in patients receiving vitamin D3. Baseline 25(OH)D3 concentration was the only factor predictive of the change in 25(OH)D3 after supplementation.

Conclusions: Administration of a single high dose of vitamin D3 leads to a significant increase in concentrations of 25(OH)D3, 24,25(OH)2D3, 3-epi (OH)D3 and 1,25(OH)2D3; induction of the catabolic pathway predominates over the production of 25(OH)D3. Due to the high inter-individual variation in the 25(OH)D3 response to supplementation, any given dose of vitamin D is unlikely to achieve optimal vitamin D status in all treated individuals

Development of an UHPLC-MS/MS method for quantitation of lipid mediators in biological matrices

I. Hartling1,2, A. Cremonesi1, M. Hersberger1,2

1 Division of Clinical Chemistry and Biochemistry, Children’s Research Center, University Children’s Hospital Zurich, Zurich; 2 Zurich Center for Integrative Human Physiology, University of Zurich, Zurich

Omega-3 polyunsaturated fatty acids (PUFAs) like eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) are essential fatty acids associated with many health benefits. Omega-3 PUFAs in particular have demonstrated both anti-inflammatory and antidepressant effects. In contrast, arachidonic acid (AA), an omega-6 PUFA, is the precursor of mainly pro-inflammatory lipid mediators, such as prostaglandins and leukotrienes. These pro-inflammatory lipid mediators are involved in initiating inflammation, while the omega-3 PUFA-derived pro-resolving lipid mediators, such as resolvins and maresins, are essential for the resolution of inflammation. Another class of lipid mediators are the endocannabinoids, which are metabolized from both omega-6 and omega-3 PUFAs and which have been shown to play a role in inflammation, pain and mood. To better understand the mechanism of the omega-3 health benefits, we developed an UHPLC-MS/MS method to measure 50 lipid mediators in several biological matrices including human plasma and cultured cells/media. This targeted lipidomic analysis measures the pro-inflammatory prostaglandins, leukotrienes, epoxy eicosatrienoic acids, and the pro-resolving lipoxins, resolvins, protectins, maresins, and the endocannabinoids. In particular, the method can measure arachidonoyl ethanolamide, docosahexaenoyl ethanolamide, eicosapentaenoyl ethanolamide, and 2-arachidonoyl glycerol which are all endocannabinoids with possible roles in antidepressant effects attributed to omega-3 PUFAs. This method will be particularly useful for investigating the antidepressant effects of omega-3 PUFAs, and for studying lipid mediators in inflammation and inflammatory diseases.

ELISA methods comparison for the detection of auto-antibodies against apolipoprotein A1

M.A. Frias1, J. Virzi1, J. Batuca2, S. Pagano1, N. Satta1, J. Delgado Alves2, N. Vuilleumier1

1 University Hospital of Geneva, Switzerland; 2 Universidade de Lisboa, Portugal

Background: Autoantibodies against apolipoprotein A1 (anti-apoA1 IgG) have emerged as an independent biomarker for cardiovascular disease and mortality. Across the studies, different ELISA methods and experimental protocols have been used to measure the level of circulating anti-apoA1 IgG which could lead to substantial results differences between assays.

Aims: To make a comparative study of the different available anti-apoA1 IgG detection methods and to determine whether the choice of matrix sample (serum vs plasma) could influence the results.

Methods: Blood samples were obtained from 160 healthy blood donors and collected on 4 different matrix (serum, plasma-EDTA, -citrate, -lithium-heparinate). Anti-apoA1 IgG was measured using two homemade (Geneva’s and Lisbon’s) and one commercial ELISA kits. Passing-Bablok and Bland-Altman were used to compare the results. Anti-apoA1 IgG seropositivity cut-offs were defined according to the user’s or manufacturer’s criterion.

Results: Firstly, the analysis of the results showed substantial differences between those 3 assays. The dynamic ranges were substantially different between methods, the commercial kit displaying the narrowest one. Passing-Bablok analysis demonstrated important proportional and constant biases between assays. The anti-apoA1 IgG seropositivity rate between Geneva, Lisbon and commercial assays varied between 24.5% and 1.9%. Matrix comparisons demonstrated that the matrix choice (plasma versus serum) strongly influenced anti-apoA1 IgG results as well as the seropositiviy rate in an assay-dependent manner. The coating antigen source was identified as important factor underlying results heterogeneity across assays.

Conclusions: Furthermore these results highlight the impact of existing assays on anti-apoA1 IgG results and emphasize the need of standardizing existing assays, given the important matrix influence, we suggest to use serum as matrix of choice.

Laboratory optimisation for improved management of alcohol intoxicated patients in an emergency department

O. Rossel2, C. Schild1, A.B. Leichtle1, N. Sédille-Mostafaie1

1 Center of Laboratory Medicine ZLM/University institute of Clinical Chemistry, Inselspital, Bern University Hospital, Bern, Switzerland; 2 MCL Laboratoires Médicaux, Niederwangen, Switzerland

Background: Alcohol abuse is a worldwide problematic as it may result with negative life quality outcomes. Due to increasing number of excessive alcohol consumption in Switzerland, the management of alcohol intoxicated patients is of great interest. Admission of patients in emergency department under alcohol abuse thus represents a substantial part of overall arrivals and optimal management of such patients is therefore essential. The clinical laboratory plays a central role in patients monitoring by providing quick blood ethanol level to clinicians.

Methods: We performed a method comparison between two different ethanol determination techniques in order to implement an optimized way of measuring alcohol. We compared plasma ethanol results in g/L (n=29); obtained after the manual calculation of plasma ethanol concentration from plasma osmolal gap (Measured osmolality – calculated osmolality, ethanol in g/L: osmolal gap/22), which was the current method, with a direct enzymatic measurement on Cobas 8000 c502 module (Roche). A Passing-Bablok regression was performed to assess the correlation between both methods. We also investigated the stability of ethanol in decapped tubes after two hours storage at room temperature.

Results: The method comparison showed a good correlation between both methods (Passing-Bablok: 0.15 + 0.97 * Ethanol Osmolal gap). The intra- and inter-assay of the Roche assay were both smaller than 5% (3.21% and 3.04% respectively) with an intra-assay bias of -1.2%. The stability assessment revealed a variation of 10% after two hours storage of unstoppered tubes.

Conclusions: Our results show that both methods are in good argument. However, the enzymatic method provides a direct, rapid and precise quantification of ethanol concentration while result obtained from osmolality still need to be converted in ethanol concentration by the clinicians. Moreover, ethanol concentration estimated from osmolality might vary depending on the formula used for calculation of the osmolal gap. Finally, results obtained from osmolality measurement do not provide the possibility to distinguish between different kind of toxics, such as ethanol, methanol and ethylene glycol. The enzymatic method enables clinicians an accurate diagnosis of alcohol intoxication and rapid initiating appropriate therapeutic steps, respectively. This approach supports an efficient patient management especially in a high throughput area such as emergency department.

Cerebral creatine deficiency and lower weight gain in a new KI rat model of creatine transporter deficiency

L. Duran-Trio1, M. Loup1, C. Cudalbu2, O. Braissant1

1 CHUV, Lausanne; 2 EPFL, Lausanne

Background: Creatine (Cr) is a nitrogenous organic acid essential for recycling ATP. Cr is synthesized by a 2-step pathway (AGAT and GAMT), and transported by SLC6A8. Cerebral Cr deficiency syndromes (CCDS), due to AGAT, GAMT or SLC6A8 deficiencies, are inborn errors of metabolism causing severe neurodevelopmental delays and intellectual disability, characterized by absence of brain Cr measured by magnetic resonance spectroscopy (MRS). While AGAT and GAMT deficiencies can be improved with Cr treatment, the X-linked SLC6A8 deficiency cannot. Pathological mechanisms are still largely unknown. We present the first characterization of a new rat model of SLC6A8 deficiency.

Methods: Generation of knock-in rats: Codon Tyr389 of Slc6a8 rat gene was changed to Cys (c.1166A>G) using the CRISPR/Cas9 engineering technology (Sprague-Dawley rats). Progeny were genotyped by PCR. Measurement of brain Cr by MRS: Rats were anesthetized with 1.5-4% isoflurane and 1H-MRS scans were performed on a horizontal 9.4T MRI system in different brain regions. CNS was analyzed on cryosections by immunohistochemistry for the astrocytic marker GFAP and for aquaporin 4.

Results: We have established a Slc6a8 Y389C/y rat strain based on one same missense point mutation described in human abolishing completely the Cr transporter activity. Mutant males rats showed absence of Cr peaks in CNS 1H-MRS, and a 40% decrease in body weight gain at 14-18 weeks (as compared to age-matched WT). The homozygous females had the same pattern as mutant males, while heterozygotes were indistinguishable from WT. Astrocytic fibers (GFAP) and microcapillaries (aquaporin 4) appeared disorganized, in particular in the cerebellar cortex.

Conclusions: Our first results validate this rat model as a promising tool to better understand SLC6A8 deficiency. In particular, morphological alterations of brain structures in our Slc6a8 Y389C/y rats, and their loss of weight gain, may help to comprehend and treat human pathology.

Deleterious effects of tetrahydrobiopterin on developing rat brain cell cultures

N. Remacle1, M. Gonzalez-Melo1, H.-P. Cudré-Cung1, H. Henry1, A.B. Hale3, K.M. Channon3, S. Calderon Copete4, J. Weber4, S. Pradervand4, O. Braissant1, D. Ballhausen1

1 CHUV, Lausanne; 2 Oxford University, UK; 3 UNIL, Lausanne

Background: Phenylketonuria (PKU) is an inborn error of metabolism caused by phenylalanine hydroxylase (PAH) deficiency. Most patients with mild or moderate PKU can be treated with sapropterin dihydrochloride (SD), which since 2015 is registered for PKU patients from the age of 4 months. SD is a pharmaceutical version of tetrahydrobiopterin (BH4), a cofactor of PAH. Hyperactivity has recently been reported as a post-marketing observation in PKU patients treated with SD.

Methods: 60 or 120 ng/ml sepiapterin, a stable precursor of BH4, were added every 24h over three days to culture media of 3D organotypic rat brain cell cultures at two developmental stages. BH4 and BH2 measurements, immunohistochemistry, western blotting, metabolomics and RNA sequencing were performed at different time points.

Results: We confirmed successful conversion of sepiapterin to BH4. In the earlier developmental stage, all observed effects were already present at 60 ng/ml. Interestingly, none of these effects was observed in the later developmental stage. We found swollen astrocytes, diminished astrocytic fibres, delayed differentiation of oligodendrocytes and perturbation of axonal elongation. We also discovered signs of altered GABAergic neurotransmission. RNA sequencing analyses revealed a number of significantly dysregulated genes. GO enrichment allowed identification of affected key biological processes.

Conclusions: We showed deleterious effects of BH4 on immature developing brain cells in a rat in vitro model. This observation raises the question whether the use of SD can be recommended in very young PKU patients as currently licensed. Further in vivo studies are needed to confirm our findings.

A knock-in rat model for glutaric aciduria type I confirms cerebral ammonium accumulation

M. Gonzalez-Melo1, N. Remacle1, H.-P. Cudré-Cung1, H. Henry1, C. Goepfert2, M. Costanzo3, M. Caterino3, M. Ruoppolo3, M. Barroso4, S.W. Gersting4, O. Braissant1, D. Ballhausen1

1 CHUV, Lausanne, Switzerland; 2 University of Bern, Switzerland; 3 University of Napoli, Italy; 4 University of Hamburg, Germany

Background: Glutaric aciduria type I (GA-I) is caused by deficiency of glutaryl-CoA dehydrogenase (GCDH). Most untreated patients are asymptomatic at birth and then develop encephalopathic crises most often triggered by a catabolic stress, which lead to irreversible neurological impairment. Despite numerous in vitro and in vivo studies, the pathogenesis of neurological damage in GA-I remains poorly understood.

Methods: R411W, the rat homologue of the frequent human mutation R402W, was introduced into the Gcdh gene of Sprague-Dawley rats by CRISPR/Casp9-mediated genome engineering.

Results: Homozygous Gcdh ki/ki rats showed a normal growth and were fertile. They revealed a biochemical phenotype typical for GA-I including elevations of 3-OHGA, GA and glutarylcarnitine in tissues and body fluids. Further, a significant increase of ammonium (NH4+) was found in plasma accompagnied by glutamine decrease and glutamate increase, suggesting that NH4+ was produced by the enzyme glutaminase (Gls). Histologically, Gcdh ki/ki rats developed the typical diffuse spongiform myelinopathy as known from autopsies of GA-I patients. However, Gcdh ki/ki rats did not present any signs of an encephalopathic crisis. Proteomic analyses on brain tissues revealed dysregulated genes in Gcdh ki/ki rats that are implicated in synaptic signal transmission and synapsis organization. Further, Gls was found 12-fold up-regulated in brain tissues of Gcdh ki/ki rats.

Conclusions: We successfully created the first transgenic rat model for GA-I. The characterization of this new model showed further evidence for a role of cerebral NH4+ production by Gls in the neuropathogenesis of GA-I. New therapeutic strategies targeting cerebral NH4+ accumulation have to be developed and can be tested in the Gcdh ki/ki rat.

In vivo longitudinal 1H-MRS Study of hippocampal, cerebral and striatal metabolic changes in BDL rats

D. Simicic1, K. Pierzchala1, V. Rackayova1, V. McLin2, C. Cudalbu1, O. Braissant3

1 EPFL, Lausanne; 2 HUG, Geneva; 3 CHUV, Lausanne

Background: Chronic hepatic encephalopathy (CHE) is a severe neuropsychiatric disorder associated with chronic liver disease (CLD). Previous studies showed that glutamine (Gln) synthesis and ammonia uptake rates differ in various brain regions leading to a conclusion that spectral information from more than one voxel may be useful for assessment of patients with CHE. To our knowledge, there are no published studies in animal models with CLD and no in vivo 1H-MRS longitudinal studies assessing the potential brain regional differences. The aim of this study was to investigate metabolic differences between hippocampus, cerebellum and striatum as key brain regions involved in manifestation of CHE.

Methods: Hippocampus, cerebellum and striatum of Wistar male adult rats were scanned longitudinally using in vivo 1H-MRS on 9.4 T system before and after bile duct ligation (BDL, accepted model of CHE). Scans and blood sampling (bilirubin, blood ammonia, ALAT and glucose) were performed before BDL (scan0) and every two weeks (scan2,4,6,8). MR experiments were performed using the SPECIAL sequence (TE=2.8ms). Metabolite concentrations were calculated by LCModel using water as reference.

Results: BDL rats showed increase in plasma bilirubin proving the presence of CLD. Increase of blood ammonia was observed, correlating significantly with increased Gln for the three brain regions (up to 142% increase at week8). As expected, the main brain organic osmolytes (taurine, creatine, inositol) followed a similar trend of decrease as response to Gln increase (osmoregulation) for the three brain regions, being the most significant in cerebellum.

Conclusions: This is the first study showing in vivo longitudinal analysis of metabolic response to CHE in three different brain regions (hippocampus, cerebellum and striatum). We demonstrated that cerebellum shows the strongest metabolite changes followed by hippocampus and by striatum where the changes were minor. Our results suggest that the different brain regions differentially respond to CHE.

Early and profound 1H-MRS neurometabolic changes in pup rats with cholestatic liver disease compared to changes in adults

V. Rackayova1, V. McLin2, C. Cudalbu1, O. Braissant3

1 EPFL, Lausanne; 2 HUG, Geneva; 3 CHUV, Lausanne

Background: Chronic hepatic encephalopathy (CHE) is a serious neuropsychiatric disorder due to chronic liver disease (CLD) occurring in adults and children. While in adults most of the neurological symptoms appear reversible after liver transplantation, neurocognitive deficits persist in children. Underlying mechanisms leading to CHE and differences between adults and children are not understood. We aimed to compare neurometabolic changes during progression of CHE in adult and developing brain using an animal model of CLD, bile duct ligated (BDL) rats.

Methods: 1H-MRS was performed in hippocampus using SPECIAL sequence (TE=2.8ms) at 9.4T. Rats underwent MRS and blood sampling before BDL (week0) and at post-operative weeks 2/4/6/8. Rats were BDL as adults where each animal served as its own control for 1H-MRS (pre-BDL scan/week0). Pups were BDL or sham-operated at post-natal day 21 (P21), sham animals serving as age-matched controls to take into account ongoing brain development in young rats.

Results: P21 BDL rats showed higher plasma NH4+ throughout disease. Brain Gln in P21 BDL rats increased significantly already 2 weeks post-BDL, compared to 4 weeks in adults. Steeper slopes of plasma NH4+ and hippocampal Gln increases indicate differences in glutamine synthetase kinetics between developing and adult brain. Osmotic answer to Gln increase was stronger and quicker in P21 than in adult BDL rats. Moreover, disturbances in energy metabolism and oxidative stress were stronger in young BDL rats. Finally, neurotransmitters Glu and Asp were stronger affected in young BDL rats.

Conclusions: Several striking differences were observed between BDL pups and adults: all neurometabolic changes were more pronounced in the developing brain, some changes appearing earlier (Gln, mIns, tCho, Glu); several neurometabolic changes seemed to be unique to the developing brain (Asp, Lac, PCr). We conclude that developing CNS displays increased vulnerability to CLD metabolic insults compared to adult animals. This might stand behind long-lasting cognitive and neurological problems in children even after liver transplantation.

Creatine as treatment for neurotoxicity in argininosuccinate lyase deficiency: in vitro studies in embryonic rat 3D brain organoids

C. Diez-Fernandez1, M.Loup2, D. Hertig3, G. Diserens3, J.-M. Nuoffer3, P. Vermathen3, J. Häberle1, O. Braissant2

1 Kinderspital, Zürich; 2 CHUV, Lausanne; 3 Inselspital, Bern

Background: Patients with the urea cycle disorder (UCD) argininosuccinate lyase deficiency (ASLD), even without ever suffering from hyperammonemia, often develop worse neurological outcome than other UCDs. Our aim is to understand the impact on the developing brain of argininosuccinic acid (ASA) and guanidinosuccinic acid (GSA), specific ASLD metabolites, and to test if creatine (Cr) and/or nitrite have protective effects. If confirmed, these molecules could provide a novel co-therapy for ASLD, which would help prevent neurological damage.

Methods: Rat organotypic brain cell 3D cultures were used to mimic ASLD by repeated administration of combinations of ASA, GSA, NH4+, Cr and nitrite at two time points (representing different developmental stages). After harvest and cryopreservation, metabolite analysis, immunohistochemistry, western blotting and metabolomics were performed.

Results: At 100 μM, both ASA and GSA abolished expression of several proteins, including neuronal (pNFM) and astrocytic (GFAP) markers. Cr co-treatment reverted this effect. However, when ASA and GSA were combined with NH4+, Cr was not able to restore protein expression. Interestingly, nitrite partially protected protein expression when NH4+ was present. Regarding the measured metabolites, the most notable effect was a ~150-fold increase in lactate dehydrogenase (LDH) in presence of 100 μM GSA, indicating increased cell death.

Conclusions: The 3D brain cell cultures increased their metabolic activity with time reflecting maturation. Thus, this model proved reliable to study toxicity in different stages of brain development. ASA and GSA have a toxic effect on developing brain cells, which worsens in presence of NH4+. In the model used here, Cr prevented some effects caused by ASA and GSA, but only nitrite could partially protect protein expression when NH4+ was present. We suggest further evaluation of Cr and nitrite as co-treatment for ASLD to prevent neurological damage.

Comparison of four immunoassays and one bioassay for their ability to diagnose and predict relapse in Graves’ Disease

T. Struja1, E. Mundwiler2, R. Jutzi1, N. Imahorn1, M. Kaeslin1, F. Boesiger1, A. Kutz1, A. Huber2, M. Kraenzlin3, B. Müller1, C. Meier3, L. Bernasconi2, P. Schütz1

1 Clinic for Endocrinology, Kantonsspital Aarau, 2 Department of Laboratory Medicine, Kantonsspital Aarau, 3 Endonet, Private Clinic, Basel

Aims: to assess the performance of four anti-TSH receptor antibodies immunoassays and one cAMP-bioassay for the diagnosis and relapse prediction of Graves’ Disease

Methods: in this observational study we retrospectively analyzed 96 serum samples of patients presenting with GD and 48 diseased controls (15 Hashimoto disease, 9 thyreoiditis, 13 thyroid autonomy, 10 other). Analysis was performed by BRAHMS TRAK human KRYPTOR (Thermo Fisher Scientific), Elisa RSR TRAb Fast (RSR Limited), IMMULITE 2000 TSI (Siemens Healthineers), EliA anti-TSH-R (Thermo Fisher Scientific) and stimulatory/inhibitory cAMP bioassays (RSR Limited).

Results: Diagnostic sensitivity and specificity of each assay were evaluated by ROC analysis and varied between 82.5% - 95.8% and 87.2% - 95.7% respectively. Overall 19 patients out of 96 (19.8%) experienced relapse after ATD stop. The ability of the assays to predict relapse was assessed and used for the calculation of the recently published GREAT score. The AUC for the prediction of relapse varied between 0.45 and 0.72 depending on the assay.

Conclusions: the diagnostic performance was similar between the anti-TSH receptor antibodies immunoassays. The ability to predict relapse by a single measurement of anti-TSH receptor antibodies at the beginning of ATD treatment is of limited clinical utility. The combination of anti-TSH receptor results with additional clinical and biochemical parameters, as suggested by the GREAT score, slightly improved relapse prediction ability.

Evaluation and comparison of the new free testosterone CLIA on the ids/iSYS system

L. Bernasconi1, P. Neyer1, F. Ferrara1, G. A. Tzalavras2, F. Ebrahimi3, A. Hammerer-Lercher1

1 Institut für Labormedizin, Kantonsspital, Aarau 2 Unilabs AG, Dübendorf 3 Abteilung für Endokrinologie, Universitätsspital, Basel

Background: The measurement of free testosterone (fT) is indicated when imbalance between synthesis of testosterone and its binding proteins is suspected. The goal of our study was to evaluate the analytical performance of a new fT assay.

Methods: The new test is a competitive, heterogeneous immunoassay based on chemiluminescence detection technology for the quantification of fT in human serum on the ids/iSYS system. Evaluation included intra- and inter-assay precision assessment using control materials and patient sera and method comparison with a Beckman radioimmunoassay. The fT assay was additionally validated using 70 samples of hypogonadal obese men and 30 specimens collected from young athletes.

Results: The observed imprecision (CV) ranged from 2.6% to 6.6% on replicates of quality control samples provided with the test kit (inter-assay) and from 2.7% to 6.5% on replicates of patient samples distributed over the measurement range (intra-assay). Method comparison resulted in a correlation coefficient (Pearson) of r2=0.9689, slope (Passing-Bablok) of 0.76 (95% CI, 073 – 0.80) and intercept (Passing-Bablok) of 0.15 (95% CI, -0.28 – 0.60). The median of the relative bias observed amounted to -23.5%. The mean fT value in the hypogonadal obese group (28.2 pmol/L, CI95% 26.2 – 30.2) was significantly lower (p << 0.05) than in the athletes group (35.6 pmol/L, CI95%, 32.1 –39.1). Comparison between measured and estimated fT shows a median bias of -86.7% which is in accordance with the literature.

Conclusions: Precision results have been fully satisfying and reflected the manufacturer’s declared performance. The observed bias to the current method is stable over the whole measurement range and might be explained by the use of different assay antibodies and reference material. The newly provided reference ranges take this bias into account. The fT assay shows satisfying analytical performance and could be helpful as a complementary biomarker for the diagnostic work-up of hypogonadism in combination with total testosterone and clinical assessment.

Performance evaluation of the new Abbott Alinity c system

L. Bernasconi1, N. Levy1, F. Ferrara1, P. Neyer1, A. Hammerer-Lercher1

1 Institut für Labormedizin, Kantonsspital, Aarau

Aims: To evaluate the analytical performance of Alinity c, Abbott’s next-generation clinical chemistry (CC) system in independent laboratory setting. Alinity c is based on photometric and potentiometric detection technologies for the quantitative determination of analytes in human serum, plasma, urine, cerebrospinal fluid, hemolysate and whole blood. Performance profiles including precision, limit of quantification and method comparison for selected assays (ALAT, ASAT, ALP, GGT, HDL, total bilirubin, cholesterol, creatinine, triglyceride, urea, uric acid) were assessed.

Methods: Within-run-precision was determined by measuring 20 replicates in one run using 3 levels of control material (Liquid Assayed Multiqual, Biorad; Alinity c Controls, Abbott) according to the CLSI-EP24-A3C guideline. The 5-day precision performance was evaluated for 3 levels of control material by measuring duplicates in two runs per day on five consecutive working days according to the CLSI-EP5-A2 guideline. Limits of detection and quantification were verified on 20 replicates according to the CLSI-EP17-A guideline. Correlation between the Alinity c and Dimension Vista 1500 systems was assessed by testing at least 40 samples according to the CLSI-EP9-A3 guideline.

Results/Conclusions: The performance of 11 clinical chemistry assays of the novel Alinity c system has been assessed under conditions of routine laboratory testing. Precision, bias and correlation to the current methods have been fully satisfying and reflected the manufacturer’s declared performance.

Performance on site evaluation of the new Abbott Alinity i system

L. Bernasconi1, N. Levy1, F. Ferrara1, P. Neyer1, A. Hammerer-Lercher1

1 Institut für Labormedizin, Kantonsspital, Aarau

Aims: To evaluate the analytical performance of Alinity i, Abbott’s next-generation immunoassay (IA) system in independent laboratory setting. Alinity i is based on chemiluminescent microparticle immunoassay (CMIA) detection technology for the quantitative determination of analytes in human serum, plasma, urine and cerebrospinal fluid. Performance profiles including precision, linearity, limit of quantification and method comparison for selected assays (TSH, fT3, fT4, high sensitive cardiac troponin I (hsTnI), vitamin D 25) were assessed.

Methods: Within-run-precision was determined by measuring 20 replicates in one run using 3 levels of control material (Liquid Assayed Multiqual, Biorad; Alinity i Controls, Abbott) according to the CLSI-EP24-A3C guideline. The 5-day precision performance was evaluated for 3 levels of control material by measuring duplicates in two runs per day on five consecutive working days according to the CLSI-EP5-A2 guideline. Linearity of the initial measuring range was performed using 5 concentration levels according to the CLSI-EP6-A guideline. Limits of detection and quantification were verified on 20 replicates according to the CLSI-EP17-A guideline. Correlation between the Alinity i and Dimension Vista 1500 / IDS-iSys / Architect i2000 systems was assessed by testing at least 40 patient samples according to the CLSI-EP9-A3 guideline.

Results/Conclusions: The performance of 5 immunochemistry assays of the novel Alinity i system has been assessed under conditions of routine laboratory testing. Precision, bias and correlation to the current methods have been satisfying and reflected the manufacturer’s declared performance. The fT3 assay showed a disturbed linearity. LOB, LOQ, LOD of the TSH and hsTnI assays exceeded the manufacturer declared performance showing high precision in the lower measuring range.

Detection of intrathecal immunoglobulin synthesis: a rapid and economical approach using a combination of kappa free light chain index and oligoclonal bands

E. Mundwiler-Pachlatko1, L. Bernasconi1

1 Institute of Laboratory Medicine, Kantonsspital, Aarau

Background: The gold standard for the detection of intrathecally synthesized IgG (iIgG), a diagnostic feature in multiple sclerosis, is the identification of oligoclonal IgG bands (OCB) in the cerebrospinal fluid (CSF) by isoelectric focusing. Recently, the nephelometric determination of Kappa free light chain (FLC) concentration in CSF and serum has been described to assess iIgG.

Methods: We retrospectively analyzed 599 CSF/serum pairs. We compared absolute Kappa-FLC CSF and Index values to OCB. Using a cutoff of 0.4 mg/l for absolute Kappa-FLC and 12 for Kappa-FLC-Index, the sensitivities were 94.6% (absolute) and 85.0% (Index) when compared to OCB. The specificities were 85.6% and 97.9% respectively.

Results/Conclusions: The Kappa-FLC-Index showed good performance in detecting iIgG. To discriminate multiple sclerosis patients as well as for an OCB-independent approach for the assessment of iIgG an optimal cutoff on the basis of clinical data needs to be defined. We propose a complementary use of the Kappa-FLC-Index and OCB for the analysis of iIgG. Setting the cutoff to >2.3 or >14.3 the sensitivity or, correspondingly, the specificity in relation to OCB increases to >99%. Thus, with an initial screening for Kappa-FLC-Index 67.4% of all samples (?2.3 and >14.3) could directly be reported negative or positive for iIgG based solely on the Kappa-FLC-Index. Samples with a Kappa-FLC Index between 2.3 and 14.3 (32.6%) would further be assessed by isoelectric focusing for OCB. With this approach the laborious and cost-intensive testing by isoelectric focusing can be minimized to a third of all samples analyzed for iIgG.

Pre-analytics of cell-free DNA extraction from urine

P. Aebi1, D.Schärer1, C.R. Largiadèr1, U. Amstutz1

1 University Institute of Clinical Chemistry, Inselspital Bern University Hospital, University of Bern, Switzerland

Background: In recent years the analysis of cell-free DNA (cfDNA) in blood samples and other human fluids (liquid biopsies) has been confirmed as a valid diagnostic and investigative tool. Liquid biopsies provide information regarding a tumor’s genome while being minimally invasive. In contrast to cfDNA in blood, cfDNA in urine (ucfDNA) is thought to be unstable due to the presence of DNAses. Different methods for stabilization and extraction of cfDNA from urine are available, but have not been systematically evaluated. For future research or diagnostic applications, there is thus a need to determine the optimal pre-analytical procedures for ucfDNA. This study aimed to evaluate cfDNA stabilizing agents, different extraction methods and impact of sex and time of urine collection on ucfDNA.

Methods: Three different commercially available stabilizing agents and cfDNA extraction kits were compared by total yield of DNA and the ratio of cfDNA to cell-derived DNA using urine from healthy volunteers. Furthermore different urine centrifugation protocols were evaluated for their potential to maximize cfDNA yield.

Results: Preliminary results show an increase of cellular DNA in comparison to cfDNA in stabilizing agents by Norgen Biotek Corp. and Zymo Research after seven days, but not for theagent by Streck Inc.. Furthermore, no difference in cfDNA yield has been observed so far between extraction kits by Zymo Research, Norgen Biotek Corp. and Qiagen. The different centrifugation protocols did not result in different proportions of cell-free to cellular DNA.

Conclusions: While all evaluated agents prevented degradation of cfDNA, the stabilizing agent by Streck Inc. appears to have a clear advantage to the other two stabilizers due to an inhibition cell lysis during samples storage. Based on preliminary results, centrifugation protocols do not seem to have a major impact on ucfDNA yields. Conclusions regarding the optimal ucfDNA extraction kit require the analysis of additional samples. In addition to yield, the choice of the optimal kit will also be influenced by feasibility and cost.

The effect of genetic variants in DPYD and miR-27a on uracil metabolism and liver DPD activity.

S. Hamzic1,2, D. Schärer1,2, S. Fontana3, U. Amstutz1, C.R. Largiadèr1

1 Institute of Clinical Chemistry, Inselspital, Bern University Hospital, University of Bern, Switzerland; 2 Graduate School for Cellular and Biomedical Sciences, University of Bern, Bern, Switzerland; 3 Regional Blood Transfusion Service of the Swiss Red Cross, Bern, Switzerland

Background: The enzyme dihydropyrimidine dehydrogenase (DPD) is the key metabolizing enzyme of the cancer drug 5-Fluorouracil (5-FU). During the last years, four DPD impairing variants (c.1905+1G>A, c.1679T>G, c.2846A>T and c.1129-5923C>A) in the DPD encoding gene (DPYD) have emerged as useful markers for the prediction of 5-FU related toxicities. However, other DPYD variants e.g. c.496A>G and c.1601G>A, and a variant (n.40A>G) in the DPD regulating microRNA miR-27a suspected to confer lower DPD activity require further assessment of their clinical relevance. Therefore, we assessed the impact of these polymorphisms together with the four clinically relevant DPYD variants on endogenous uracil (U), dihydrouracil (UH2) plasma levels and DPD activity in human liver.

Methods: Endogenous U and UH2 plasma levels were assessed in two cohorts of healthy volunteers (N1=320 and N2=308) and in a small cohort of 5-FU treated patients (N=24) by LC-MS/MS. DPD activity was measured in human liver S9 fractions by a LC-MS/MS assay (N=25). All samples were genotyped for the same DPYD and miR-27a variants.

Results: As expected, all clinically relevant DPYD variants were associated with lower UH2/U ratios in healthy volunteers (Pcombined< 0.0001). Homozygous or compound heterozygous carriers of DPYD c.496G and c.1601A were also associated with decreased UH2/U ratios (Pcombined_hom=0.0008). Surprisingly, miR-27a n.40G was associated with higher UH2/U ratios (Pcombined=0.0036) in the subgroup of DPYD risk variant carriers, whereas carriers of this variant showed a trend to lower UH2/U ratios in 5-FU treated patients. Additionally, homozygous n.40G carriers showed a trend to 10% lower DPD activity in human liver samples as compared to n.40A carriers.

Conclusions: Our results suggest that homozygous or compound heterozygous carriers of c.496G and c.1601A have impaired DPD activity. For the miR-27a n.40G variant, conflicting results were obtained. Interestingly, a trend in the expected direction was observed in liver tissue, a major site of DPD activity, and under 5-FU treatment, where the enzyme reaches a saturated state due to the exposure to high drug concentrations.

Translating the findings related to anti-apolipoprotein A-1 IgG in familial hypercholesterolemia towards the identification of a new endogenous disruptor of cellular cholesterol homeostasis

S. Pagano1,2, A. Magenta3, M. D’Agostino3, F. Martino4, F. Barillà5, N. Satta1,2, M. Frias2, E. Favari6, A. Ronca6, F. Mach7, B. Gencer7, N. Vuilleumier1,2

1 Division of Laboratory Medicine, Department of Genetics and Laboratory Medicine, Geneva University Hospital, Geneva, Switzerland; 2 Department of internal medicine specialities, Medical Faculty, Geneva University, Geneva, Switzerland; 3 Fondazione Luigi Maria Monti, Istituto Dermopatico dell’Immacolata-IRCCS, Experimental Immunology Laboratory, Rome, Italy; 4 Department of Pediatrics, Sapienza University of Rome, Rome, Italy; 5 Department of Cardiovascular, Respiratory, Nephrological, Anesthesiological and Geriatrical Sciences, Sapienza University of Rome, Rome, Italy 6 Department of Food and Drug, University of Parma, Parma, Italy; 7 Division of Cardiology, Geneva University Hospital, Geneva, Switzerland.

Aims: To explore i) the associations between anti-apolipoprotein (apo) A-1IgG, lipid profile, and microRNA-33 (mi-R33) in children with familial hypercholesterolemia (FH); ii) the in vitro effects of anti-apoA-1 IgG on macrophages cholesterol homeostasis (CH); and foam cell formation, the hallmark of atherogenesis.

Methods: Anti-apoA-1IgG, lipid profile and miR-33 were assessed on 29 FH children. Human monocyte-derived macrophages were used to assess the impact of anti-apoA-1IgG on key CH regulators (miR-33, SREBP-2, HMGCR, LDL-R, SCAP and ABCA1). Foam cell formation was detected by Oil-RedO staining and ABCA1 cholesterol efflux was performed on THP-1 cell.

Results: Prevalence of high anti-apoA-1IgG levels (seropositivity) was 37.9 % in FH children. Seropositive individuals had lower LDL and circulating miR-33a levels but higher ABCA1 levels than seronegative children (p=0.04, p=0.04, p=0.02, and p=0.001, respectively). On macrophages, anti-apoA-1IgG significantly induced ABCA1 level and a TLR2/4, NF-kB and AP1-dependent increase of SREBP-2, LDL-R and HMGCR, but significantly decreased mature miRNA-33a expression, independently of TLR2/4/CD14 signalling and followed by passive diffusion inhibition and ABCA1 efflux stimulation. The net effect of these anti-apoA-1 IgG-mediated effects on CH led to a significant LDL-dependent foam cell formation.

Conclusions: In FH children, the prevalence of anti-apoA-1 IgG is elevated, and surprisingly associated with a more favourable lipid profile. Our in vitro model highlights a new role for anti-apoA-1 IgG as modulator of CH through a complex interplay involving miR-33a, TLR2-4, and key proteins of lipid metabolism leading to foam cell formation, a key step of atherogenesis. These results indicate that assessing anti-apoA-1 IgG levels could represent a potential biomarker of the local propensity to atherosclerosis.

Applying LC-MS based steroid profiling in the clinical setting

S. J. Bruce1, M. Hauschild1, Z. Kolesinska1, N. Pitteloud1, P. A. Binz1

1 CHUV, Lausanne, Switzerland

Background: Within many clinical laboratories steroid hormone quantification in plasma / serum (often previously determined via immunological based methods) are being increasingly realised by LC-MS based technology. More recently due to the multiplexing nature of the LC-MS approach, more steroid hormones are being targeted and quantified in the same sample / analytical run, therefore giving rise to a larger analyte panel that is often referred to as the “steroid profile”. Within our clinical laboratory this steroid profile, which for us consists of 14 steroid hormones, contains specific progestagens, glucocorticoids, androgens and mineralocorticoids. The efficiency of such a multiplexing approach in the clinical setting is clear, particularly in terms of sample volume conservation, and such a routine approach can be a very powerful tool in supporting related clinical investigations and diagnostics.

Aims: Here we present our routine approach to clinical steroid profiling in plasma / serum by LC-MS and importantly, how these results are visualised and interpreted by clinicians within our hospital today.

Results: In our model, the LC-MS results are imported into a dynamic PDF with graphical representation of steroidogenesis. As interpretation of results depends on the clinical context of the sample workup, sample time, age, sex, treatment and other specific considerations presented by the clinician, these important details are also included in the report. Additionally, depending on the specific context, other parameters such as relevant pituitary hormone concentrations can also be included. Inconsistent results will lead to revision of analytical procedures prior to reporting, thus enhancing the overall quality of the approach.

Conclusions: The described workflow allows for a more global view of steroid hormone status, thus aiding the clinician to better understand related clinical aspects such as those that may be present in adrenal, pituitary or gonadal disturbances. This information may then guide targeted genetic analysis and therapeutic options.

Improving pre-analytical data quality with an automatized healthcare-integrated biobanking approach

T.K. Froehlich1, G.M. Fiedler1, C.R. Largiadèr1

1 University Institute of Clinical Chemistry, Bern University Hospital, University of Bern: Inselspital, Bern, Switzerland

Despite recent methodological advances in “omics-“technologies, the discovery of new biomarkers has been largely prevented by uncontrolled variability in the quality among and within existing biospecimen collections. In order to meet the quality requirements of liquid samples for high sensitive analytical technologies, such as mass spectrometry, recent efforts have mainly focused on the development of new biobanking infrastructure and on the standardization of pre-analytical protocols. With regard to the reproducibility of research results, not only the physical quality of samples but also the quality of their recorded data is crucial. Currently, pre-analytical information is often recorded manually. This type of recording is not only time consuming but also represents a considerable source of error. Here, we describe the healthcare-integrated biobank sampling process of the Liquid Biobank Bern, Switzerland, which takes advantage of multiple-interfaced IT systems and minimizes manual input of pre-analytical information. In more detail, the collection and processing of biobank samples is integrated in the automated high-throughput processing of hospital routine samples. At every processing step from the blood draw to the storage, the sample and its derivatives are identified, tracked, and directed by their barcodes, and thus, electronically monitored and documented. All essential time points within the pre-analytical pathway are recorded automatically by the processing instruments. With this high-degree of IT integration of hospital routine and biobank processes, we achieve high data quality and rapid sampling processing:>95% of samples being frozen within two hours after blood-draw and>75% even within one hour.

A simplified method for adjusting Bézier curves to ROC data to calculate likelihood ratios of quantitative test results in medical diagnosis

W Fierz1

1 labormedizinisches zentrum Dr Risch

In medical diagnostics, test results are reported more and more quantitatively in order to give the physician an accurate measure for estimating the significance of a test result for a particular diagnosis. However, it is left to the physician to assess a quantitative result as to its influence on judging the relevance for a particular disease. In principle, the physician starts with a suspicion for a particular disease that is based on the prevalence of the disease in a population that most resembles the patient and on the anamnesis and clinical examination of the patient. In order to confirm or refute the preliminary suspicion, the physician orders a specific laboratory test and a high result gives a stronger confirmation than a low result. Formally, in Bayes’ view, pretest odds multiplied by the likelihood ratio (LR) of the quantitative test result give the posttest odds. The question is how to calculate the LR. For this purpose, receiver operating characteristic (ROC) analysis provides the LR as the slope of the tangent to the ROC curve at the point corresponding to the test result. Since the exact distribution of test results in diseased and non-diseased people is often not known, the calculation of such LRs for quantitative test results is not straightforward. A simple distribution-independent method has recently been described to reach this goal using Bézier curves that are defined by tangents to a curve [1]. Here, a simplified method to adjust a Bézier curve to a ROC curve is described.

Reference: [1] Fierz W. PLoS One. 2018; 13, e0192420. doi:10.1371/journal.pone.0192420.

Published Online: 2018-11-01
Published in Print: 2018-11-27

©2018 Walter de Gruyter GmbH, Berlin/Boston