Effect-Directed Analysis (EDA) to Identify Chemical Mixtures in Stressed Aquatic Ecosystems

Transcription

1 Effect-Directed Analysis (EDA) to Identify Chemical Mixtures in Stressed Aquatic Ecosystems Werner Brack et al. 1 Department for Effect-Directed Analysis, Helmholtz Centre for Environmental Research UFZ, Leipzig, Germany

2 Multiple Stress on Aquatic Ecosystems multiple stressors toxicants nutrients oxygen hydromorphology temperature ph pathogens invasive species aquatic ecosystem multiple effects on different levels receptor binding biochemical response survival reproduction, fitness species interaction observed effect in situ fish decline disappearance of sensitive species regime shifts observed effect in vivo and in vitro bioassay response

3 Contamination: Complex Mixtures 5 mio known chemicals thousands of compounds in real world samples Abundance Time--> 1e (+8) priority pollutants (WFD) TIC: WER46.D How to assess complex contamination?

4 How to Assess Complex Contamination? field observation ecological relevance unresolved stressors biotesting effects, hazards mode of action toxicants as cause limited ecological relevance unknown toxicants target analysis, EQS predicted effect/hazard individual compound no hazard of mixture ignoring of most compounds target analysis, TU/msPAF predicted hazard of target mixture ignoring of most compounds non-target analysis broad array of compounds difficult identification no hazard prediction

5 How to Assess Complex Contamination? field observation ecological relevance unresolved stressors biotesting effects, hazards mode of action toxicants as cause limited ecological relevance unknown toxicants target analysis, EQS predicted effect/hazard individual compound no hazard of mixture ignoring of most compounds target analysis, TU/msPAF predicted hazard of target mixture ignoring of most compounds non-target analysis broad array of compounds difficult identification no hazard prediction None of the methods provides hazard and cause-effect relationships of complex mixtures Integrated approach!

6 How to Assess Complex Contamination? field observation ecological relevance biotesting effects, hazards mode of action toxicants as cause limited ecological relevance unknown toxicants WFD approach: target analysis, EQS predicted effect/hazard individual compound no hazard of mixture large unresolved gap between stressors ecological and chemical status ignoring of most compounds target analysis, TU/msPAF predicted hazard of target mixture ignoring of most compounds non-target analysis broad array of compounds difficult identification no hazard prediction

7 How to Assess Complex Contamination? field observation ecological relevance biotesting effects, hazards mode of action toxicants as cause limited ecological relevance unknown toxicants WFD approach: target analysis, EQS predicted effect/hazard individual compound no hazard of mixture large unresolved gap between stressors ecological and chemical status ignoring of most compounds target analysis, TU/msPAF EDA approach: may help to bridge this gap predicted hazard of target mixture ignoring of most compounds non-target analysis broad array of compounds difficult identification no hazard prediction

8 EDA Approach Site specific Endpoint specific Based on selected in vivo or in vitro biotests No a priori knowledge on or selection of compounds required Applicable to any matrix (water, sediments, biota, ) fractionation environmental contamination biological analysis biological analysis confirmation chemical analysis Basic assumption: Toxicity of complex environmental mixtures is predominated by few compounds. toxicant

9 EDA of Sediments Conventional sediment assessment: Focus on metals and nonpolar priority pollutants: PAHs, chlorinated pesticides (DDT, HCH, HCB, aldrin, dieldrin ) and PCBs and PCDD/Fs Can we confirm this selection with EDA?

10 Complex Mixture vs. key toxicants Example 1: AhR-mediated effects in sediment extracts very complex mixture strong effects Complex mixture or few outstanding key toxicants? Abundance 1e TIC: WER46.D Priority PAHs and PCDD/Fs or something new? Time-->

11 Complex Mixture vs. key toxicants Fractionation procedure open column NP-LC (Al 2 O 3 ) EXTRACT F1: non-polar aliphatic comp. NP-LC (NO-Phenyl) 2 F2: non-polar aromatic comp. F3, F4: polar fractions F2.1 PCDD/Fs PCBs, PCNs F2.2 PAHs MW F2.3 PAHs MW178 F2.4 PAHs MW 22 F2.5 PAHs MW 226/228 F2.6 PAHs MW 252 PAHs MW > 276 RP-LC (C18, endcapped, Polymer) F2.5.1 e.g. TRP F2.5.2 e.g. BaA F2.5.4 F2.5.3 e.g. CHR F2.5.5 e.g. 2MBaA F2.5.6 e.g. 1MBaA F2.5.7 e.g. 4MCHR F2.5.8 e.g. 3MCHR F2.5.9 e.g. 9MBaA e.g. 2MCHR F2.5.1 e.g. 1212DNF RP-LC (PYE) F MBaA F NBT F DNF F MCHR Brack et al. 23 J. Chrom. A

12 Complex Mixture vs. key toxicants Abundance 1e TIC: WER46.D EDA helps to dramatically reduce complexity B Time--> Extract 26 A F TIME

13 Complex Mixture vs. key toxicants IEQ B (ng SEq/L) F1 A F2 F3 F4 IEQ B (ng SEq/L) N.a.N extract B MFx PCDD/F SFx F2.1 F2.2 F2.3 F2.4 F2.5 F2.6 F2.7 N.a.N. F2 MF2.x SF2.x IEQ B (ng SEq/L) C F2.5.1 F2.5.2 F2.5.3 F2.5.4 F2.5.5 F2.5.6 F2.5.7 F2.5.8 F2.5.9 F2.5.1 N.a.N. F2.5 SF2.5.x

14 Complex Mixture vs. key toxicants IEQ B (ng SEq/L) IEQ B (ng SEq/L) F1 A B F2 PCDD/F F2.1 F2.2 F2.3 F2.4 F2.5 F3 F4 N.a.N. extract F2.6 F2.7 N.a.N. MFx F2 MF2.x SF2.x SFx Mass balances comparing original sample reconstituted mixture of fractions MFx prediction SFx (CA) Reasonable agreement considering uncertainty of IEQs and losses during fractionation

15 Complex Mixture vs. key toxicants 7 IEQ B (ng SEq/L) B PCDD/F F2.1 F2.2 F2.3 F2.4 F2.5 F2.6 F2.7 N.a.N. F2 MF2.x SF2.x ~ 7 % of IEQs caused by one PCDF isomer ~ 28 % of IEQs caused by four other PCDF isomers Cl Cl O Cl Cl Cl Cl

16 Complex Mixture vs. key toxicants 7 IEQ B (ng SEq/L) C S F2.5.1 F2.5.2 F2.5.3 F2.5.4 F2.5.5 F2.5.6 F2.5.7 F2.5.8 F2.5.9 F2.5.1 N.a.N. F2.5 SF2.5.x Excellent tool to identify new compounds to be considered in a mixture O O

17 IEQ Complex Mixture vs. key toxicants Quantitative confirmation of new compounds as significant contributors Relevant nonidentified compounds may not be excluded 2123DNF 22NBT 9MBaA 1MCHR 1212DNF

18 Complex Mixture vs. key toxicants Often (but not always) few outstanding toxicants predominating mixture effects key toxicants to be included in assessment. Abundance 1e TIC: WER46.D often include unknowns/unexpected compounds Time-->

19 Multiple Endpoints/Several Sediments Example 2: multiple toxicological endpoints, several sediments Does EDA allow for any generalisation across individual sediments and endpoints?

20 Multiple Endpoints/Several Sediments Automated NP HPLCfractionation Separation according to polarity, size of the aromatic system and planarity/degree of halogenation Cyanopropyl stationary Sample phase (CN) injection Step 1 Step 2 Step 3-5 Nitrophenyl stationary phase (NO) Nitrophenyl Nitrophenyl Nitrophenyl Porous Graphitised Carbon stationary phase (PGC) Pyrenyl Pyrenyl Pyrenyl alkanes, alkenes, sulphur, nonplanar diaromatic compounds PAHs, S-, O-heterocyclic PACs keto-, hydroxy-, nitro-pahs, N-heterocyclic PACs, quinones chlorinated diaromatic compounds (PCBs, PCDDs/Fs) Functional groups CN NO H 3 C C H 3 N O + N O - F13-F18 F6-F12 F1-5 polar fractions MODELKEY ( GOCE) PAHs is a research PCBs, PCDD/Fs project funded by increasing size PGC electron-rich surface

21 number of positive wells number of positive wells number of positive wells Multiple Endpoints/Several Sediments Ames Mutagenicity 5 In Most PAH fractions predominate. In Bitterfeld and Pardubice great mutagenicity in polar fractions unknown mutagens! Most TA98 TA98 with metabolic activation 2 1 Bitterfeld 2 1 Pardubice PCBs, PCDD/Fs PAHs polar compounds

22 P1 P2 P3 P4 P5 P6 P7 P8 P9 P1 P11 P12 P13 P14 P15 P16 P17 P18 B1 B2 B3 B4 B5 B6 B7 B8 B9 B1 B11 B12 B13 B14 B15 B16 B17 B18 % of control % of control M1 M2 M3 M4 M5 M6 M7 M8 M9 M1 M11 M12 M13 M14 M15 M16 M17 M18 % of control Multiple Endpoints/Several Sediments Inhibition of gap junctional intercellular communication Tumor promotion 12 GJIC-inhibiting activity at all sites in polar fractions Most fraction Bitterfeld Pardubice fraction fractions

23 % maximum induction % maximum induction % maximum induction Multiple Endpoints/Several Sediments ER-CALUX estrogenic activity Estrogenic activity only in polar fractions. Predominating F17 (steroids and bisphenol A) and F15 (alkylphenols) Most M13 M14 M15 M16 M17 M Bitterfeld 4 3 Pardubice B13 B14 B15 B16 B17 B18 P13 P14 P15 P16 P17 P18

24 pmol T4 eq/g dry sediment Multiple Endpoints/Several Sediments Prelouc Most Thyroid hormone disturbance only by polar fractions

25 Multiple Endpoints/Several Sediments Polar compounds play an important role for in vitro effects of sediment extracts. Similarities between activity patterns of different sediments helps to focus monitoring.

26 DHT equivalents (pmol DHT/g sed) Masking of Effects Example 3: androgenicity in sediment extracts AR-CALUX RP-NP fractionation Frequent observation for specific effects: No effects observed in complex mixture but after fractionation androgenic activity Whole sample RP fractionation NP fractionation DHT: dihydrotestosterone Weiss et al., Anal Bioanal Chem, 29

27 DHT equivalents (pmol DHT/g sed) Masking of Effects Androgenic effect in sediment extract remains non-detected due to antagonists fractionation enhances chance to detect effects in complex samples androgenic activity Whole sample RP fractionation NP fractionation anti-androgenic activity androgenic activity Weiss et al., Anal Bioanal Chem, 29 27

28 Bioaccessibility and EDA of Sediments Example 4: green algae cell multiplication, bioaccessible fraction Influence of extraction procedure on the mixture and the prioritisation of the fractions?

29 Schwab & Brack 27 J. Soils Sed. Bioaccessibility and EDA of Sediments 1 Exhaustive extraction vs. extraction of rapidly desorbable fraction with TENAX St/S S S t F rap e k rap t F desorption kinetics slow e k veryslow t [h] slow t F e k veryslow t Assumption: Only rapidly desorbable molecules are bioaccessible. Extraction of rapidly desorbing

30 n inhibition inhibition Schwab et al., ET&C, 29 inhibition.4 Bioaccessibility and EDA of Sediments pe benzanthrone O 1..8 Bitterfeld ASE TENAX bioaccessible PAHs 1. Prelouc.8 ASE TENAX pe 2 - nonbioaccessible PAHs NH pe -- Most PCBs PAHs polar ASE TENAX N-phenyl-2- naphthylamine

31 on inhibition Bioaccessibility and EDA of Sediments 1..8 Přelouč Confirmation based on CA.6 P1.4 P15.2. P8 P9 P Most EC ICQ EC original x mix x M1.6 M11

32 EDA of Sediments Consideration of exposure (e.g. desorption kinetics) may significantly impact mixture and prioritisation.

33 Partition-Based Dosing Example 4: green algae cell multiplication, partition-based dosing In sediments: pore water concentrations as a result of partitioning. Solvent dosing: attempt to dissolve whole mixture. Impact?

34 Partition-Based Dosing Partition-based dosing vs. solvent dosing (DMSO) Bandow et al., ES&T 29a PDMS rods loaded with sediment extract partitioning according to K PDMS/water simulation of partitioning in sediments

35 percentage of concentration t =h [%] Partition-Based Dosing Partition-based dosing vs. solvent dosing (DMSO) Bandow et al., ES&T 29a development of concentration in algal test PCB t [h] PDMS dosing solvent dosing PDMS rods loaded with sediment extract partitioning according to K PDMS/water simulation of partitioning in sediments constant dose by compensation of losses due to adsorption, degradation, evaporation

36 Inhibition inhibition [%] [%] Partition-Based Dosing DMSO PDMS-rods polar fractions predominant if partitioning is considered PAHs (predominant if partitioning is ignored) fraction Bandow et al., ES&T 29b

37 Inhibition inbibiton [%] [%] Partition-Based Dosing Identification 25 g SEQ/rod 12.5 gseq7rod 6.25g SEQ/rod F14 1 F14 1B triclosan F14 2 F14 3 F14 4 F14 5 F14 6 F14 7 F14 8 F14 9 F14 1 F14 11 F14 12 F14 13 F14 14 F14 15 F14 16 F14 RS F14 BW F14 rekombi F14 Original

38 Partition-Based Dosing Dramatic impact of dosing technique on prioritisation. Relevance of very lipophilic compounds e.g. PAHs overestimated. Relevance of less lipophilic compounds underestimated Again: unexpected outstanding key toxicants (triclosan)

39 EDA of Sediments Non-regulated compounds identified by EDA studies (algal toxicity, endocrine disuption, mutagenicity) O O N + O- O CH 3 O benzanthrone N + O - dinitropyrenes O O 2-methylanthraquinone O nandrolone O NH O O N-phenyl-2- naphthylamine Cl O P O O O Cl traseolide galaxolide Cl tonalide O OH tris(2-chloroisopropyl) phosphate Cl Cl triclosan Cl

40 EDA of Surface Waters Example: Mutagens in River Elbe downstream of Pardubice, Czech Republic Receiving industrial wastewater Significant mutagenicity Not related to priority pollutants

41 EDA of Surface Waters Fractionation and structure elucidation BR extract biological analysis cation exchange chemical analysis LC/MS 2 high resolution ESI & APCI ionisation acids & neutrals 2 steps fractionation anion 1 st polymeric C18 exchange 2 nd phenyl-hexyl Bases bases acids neutrals Biological analysis toxicant identification (MZmine, Molgen software & databases)

42 Number of revertents Number of Revertants per plate (max=48) Number of revertants per plate (max=48) EDA of Surface Waters MUTAGENICITY of fractions: Without S9 With S g of BR with S9 5 g of BR with S9 N2-1 N2-2 N2-3 N2-4 N2-5 N2-6 N2-7 N2-8 N2-9 N2-1 N2-11 N2-12 N2-13 N2-14 N2-15 N Acid fraction Basic fraction 1 Neutral g of BR without fraction S9 4 1 g of BR with S N1 N2 N3 N4 N5 N6 N7 N8 N9 N1 N11

43 EDA of Surface Waters BR1-N-polC18-HR-neg_ # RT: AV: 382 NL: 1.31E7 F: FTMS + p APCI corona Full ms [6.-14.] 5. BR extract masses m/z BR1-N-2-8-APCI-pos # RT: AV: 686 NL: 1.3E5 F: FTMS + p APCI corona Full ms [ ] N masses detected masses of interest m/z Comparison with blank and non mutagenic neighbouring fractions

44 EDA of Surface Waters BR1-N-polC18-HR-neg_ # RT: AV: 382 NL: 1.31E7 F: FTMS + p APCI corona Full ms [6.-14.] 5. BR extract 1 masses EDA PROCESS IS EFFICIENT IN SIMPLIFYING SAMPLE COMPLEXITY m/z BR1-N-2-8-APCI-pos # RT: AV: 686 NL: 1.3E5 F: FTMS + p APCI corona Full ms [ ] N masses detected masses of interest m/z Comparison with blank and non mutagenic neighbouring fractions

45 EDA of Surface Waters: Identification Strategy Exact mass extraction (MZmine) Target/suspect screening: Comparison with site-specific target lists of compounds Matches with the sample No match with the sample Log Kow, pka and MS 2 checking Decrease of the number of candidates (Fragmentation prediction) DataBase search (ChemSpider) (Exact mass, formula and logkow range) or Structure generation Empirical formula (isotope peaks) No match Standards for confirmation Mutagenicity prediction (toxnet) Good match

46 Conclusions Environmental samples normally contaminated with very complex mixtures of mostly unknown compounds Mixture toxicity models (CA): Valuable tool for toxicity prediction of known mixtures and supportive in EDA confirmation. However: does not solve the main problem: Uncertainty on key toxicants. EDA as an approach to simplify mixtures and to identify key toxicants significant improvement of mixture assessment. However: No guarantee to get all relevant compounds. Quantitative mass balance may be difficult. Incomplete recovery, masking effects, uncertainty in testing of complex mixtures. Exposure scenarios (how to consider bioavailability) may significantly impact identified mixtures. Polar compounds (non-accessible to GC-MS) play a significant role in water and sediment. Great challenge for isolation and identification In vitro assays are a valuable tool for EDA not for risk assessment. They should trigger in vivo studies.

47 In Vitro Triggering In Vivo Estrogenicity in Scheldt sediments in vitro assay (YES) sedimentextract Potamopyrgus antipodarum Analysis community Caging experiment in situ Schmitt et al., Chemosphere, 21

48 Special thanks to Nicole Bandow Christine Gallampois Urte Lübcke-von Varel Miroslav Machala Claudia Schmitt Katrin Schwab Jana Weiss