Skip to main content

Developing New Technologies to meet 21st Century Demands in Animal Forensics

Final Report Summary - DETECH21 (Developing New Technologies to meet 21st Century Demands in Animal Forensics)

Executive Summary:
The administration of both legal and illegal drugs to animals destined for food production and performance sport competition continues to be an enormous issue for regulatory authorities charged with enforcing their control as current testing to detect drug use is problematical, costly and inefficient as signified by low detection rates. DeTECH21 is a collaborative SME led and directed project seeking to apply advanced analytical techniques to develop new forensic testing services based on metabolomic profiling techniques which can identify the misuse of prohibited and licensed substances and drugs of abuse in bovine and equine animals. Research spanning a range of interdisciplinary fields including targeted mass spectrometric analysis, high resolution metabolomics profiling, and bioinformatics analysis has been applied to develop and apply new metabolomics-based screening methods to help identify animals which have been exposed to exogenous substances. This new testing relies on the use of metabolomics profiling which is the large scale and high-throughput measurement of small molecules (metabolites) present in biological samples. Such testing is unique as it can monitor for indicative biological and metabolic responses to drug administrations and is in contrast to current testing programmes which rely on the direct detection of compounds or residues. This metabolite profiling can be used as indirect evidence of inadvertent or illegal administration of compounds to equine and bovine animals helping to maintain food safety and animal welfare standards and the reputational status of the EU animal production sector. DeTECH21 has been able to establish and deliver:
• Untargeted global metabolomics profiling methods and models for predictive identification of drug use in bovine/equine animals
• Indicative panels of urinary/blood metabolite markers of animal exposure to illicit / banned substances
DeTECH21 has facilitated the transfer of methods for detection of metabolite markers from high resolution research-based mass spectrometry instruments to conventional routine instrumentation enabling high-throughput routine screening of samples at a commercial level on a cost effective basis and in a manner which can be transferable between facilities. The provision of such new and unique laboratory services together with access to panels of exposure markers will increase the competitiveness, market reach and profitability of analytical service providers and the ability to deliver quality-differentiated EU produce and animals. The impact of the outputs of DeTECH21 are focused on key sectors of economic importance to the EU which have related needs for improved analytical detection and monitoring techniques including meat produce (improved safety, integrity and quality differentiation of EU produce and equine sport and breeding (enhanced competitive integrity and reputational status).

Project Context and Objectives:
DeTECH21 is a collaborative SME led and directed project seeking to apply advanced analytical techniques to develop new forensic testing services based on metabolomic profiling techniques which can identify the misuse of prohibited and licensed substances and drugs of abuse in bovine and equine animals. Issues surrounding the administration and misuse of both legal and illegal drugs to animals destined for food production and performance sport competition continue to be an enormous challenge to authorities charged with enforcing their control. The timeliness and relevance of DeTECH21 are particularly significant as the EU attempts to move away from systems of maximising food output towards improving production efficiency of food systems and ensuring food safety and quality. Current testing regimes are problematical, costly and inefficient, traditionally targeting one single analyte class per sample. As the EU continues to tighten its regulations those carrying out illegal practices involving anabolic growth hormones have found ways of concealing their activities from regulators and avoiding detection by current testing regimes. The present testing system is compromised not only by the continued use of endogenous hormones, designer drugs and low dose chemical cocktails, but also by the limited number of samples which can be routinely analysed for drug residue presence. DeTECH21 seeks to use advanced sample profiling techniques using high level mass spectrometry and bioinformatical data analysis to develop new solutions for monitoring of drug use and abuse in food-producing and sport animals – challenges identified by SME partners as limitations to the continuing effectiveness, efficiency and market reach of their respective long-term business models. DeTECH21 project partners have come together with the collaborative aim of delivering commercially available forensic test services based on metabolomic analysis.
The metabolome is the collection of all small molecule metabolites or chemicals that can be found in a cell, organ or organism. Metabolomics focuses on large scale and high-throughput measurement of metabolites in biological matrices and its use as an indirect screening approaches to identify illegal compound use within animals. These approaches have been devised on the principles of effect-based profiling techniques whereby detection is based not on the direct analysis of chemical residues but rather on the identification of metabolite profiles representative of biological responses to exogenously administered agents. DeTECH21 aims to take proof-of-concept experimental studies forward, demonstrate suitability for wide-scale implementation and develop a commercial testing service for animal chemical forensic analysis end-users. The development of such techniques requires research spanning a range of interdisciplinary fields: including high-level profiling methodology involving:
• Targeted residue analysis
• State-of-the-art investigative metabolomics analysis
• Systems biology and bioinformatic analysis

The development and application of these techniques and methods requires key skills and expertise possessed by the RTD partners and developed by previously funded EU projects - the objective of DeTECH21 is to apply this knowledge to increase the competitiveness of indigenous EU SME companies. The immediate and short-term impact of the outputs of DeTECH21 are focused on key sectors of economic importance to the EU which have related needs for improved analytical detection and monitoring techniques – principally meat produce i.e. safety, integrity and quality differentiation of EU produce, and equine sport and breeding i.e. competitive integrity and reputational status. The overall objective of DeTECH21 will be to develop and implement metabolite-profiling techniques as new commercial diagnostic services to detect the misuse of compounds and substances in bovine and equine animals with implications for food safety and animal welfare control. Project outputs will offer an innovative and new perspective to drug detection based on the monitoring of biological and metabolic responses to confirm drug use in animals rather than the direct detection of compounds and residues which is the basis of current testing programmes. The principle scientific aims of the project are therefore to:
1: To develop and validate untargeted global metabolomic profiling techniques to screen for drug misuse in bovine/equine animals
2: To develop and validate targeted metabolite marker technique to screen for drug misuse in bovine/equine animals suitable for application as routine high sample throughput analyses

Project Results:
WP2 Technology/materials acquisition and distribution
The aim of WP2 was to establish and agree defined uniform pre-testing analytical procedures and the sourcing of appropriate standards, test materials and test matrices across the project partners to perform initial investigative studies and subsequent validation of newly developed screening approaches based on metabolomic profiling. Principle samples of interest (plasma/urine) from animal species (bovine/equine) were sourced within existing partner bio-banks (T2.1) and distributed. Test samples for WP3, WP4 and WP5 included material of known anabolic treated animals acquired previously from in-house experimental studies and field samples of unknown origin or from animals reporting positive for anabolic use through existing testing procedures. Standard procedures for the handling, processing and metabolomics analyses of test matrices were also established (T2.2). Sample preparation strategies were developed for bovine/equine blood/urine with a view to providing a broad range of metabolites representative of diverse groups of low-molecular weight compounds such as amino acids, peptides, nucleic acids, organic acids, lipids, carbohydrates and vitamins in acquired metabolomic profiles during high throughput analysis. Similar procedures were investigated for the analysis of bovine muscle / ear notch due to the potential for such matrices to be readily acquired for analysis offering advantage over biofluids such as blood and urine allowing for control of the food item itself. Mass spectrometric analysis procedures were developed for metabolomics profiling utilizing different LC conditions (RP/HILIC) and data acquisition modes (+/-). An assessment was made of the technical and analytical capabilities of project partners which would impact on implementation of project outputs which would impact the practical use of research development (T2.3). Instrumental requirements in terms of mass spectrometry and personnel skill acquisition in the use of such equipment were identified as aspects limiting successful transfer and training was offered at research provider facilities.

WP3 Quantitative residue analysis of test samples
The aim of WP3 was to develop and validate analytical methods (T3.1) for the quantification of compound residues in test matrices from selected bovine/equine animals by chromatography coupled to mass spectrometry. These methods enable accurate assessment of whether samples from animals exposed to substances of abuse/misuse actually contain such compounds or residues (T3.2) and the elimination kinetics of respective administered agents. Sample preparation, quantitative mass spectrometry procedures, and data analysis procedures for quantification of multiple compounds in urine/plasma were developed. These analytical strategies were applied to selected biological matrices to confirm administration and detection time windows of a range of drug treatments and to confirm that sourced samples were acquired from treated animals prior to the performance of metabolomics studies. As the primary focus of metabolomics investigations aimed to centre on steroid administrations in bovine/equine animals, samples of primary interest to the project focused on natural steroid administrations to animals including boldenone undecylenate and a mixture of testosterone esters (propionate, phenylpropionate, isocaproate and decanoate). Detection of steroid esters in both bovine serum and equine plasma matrices was facilitated by dedicated UPLC-MS/MS analysis based on Multiple Reaction Monitoring (MRM) techniques whilst boldenone sulphate detection was used to detect boldenone administration. However, additional samples, corresponding to other anabolic treatments were included to cover a wide range of potential anabolic practices. In this context, urine/serum samples from beta-agonist treated bovines (both calves and adult animals) and serum samples from somatotropin (peptidic growth hormone) treated cows were also characterized by targeted quantitative mass spectrometry MRM methods. Biological samples (both from treated and untreated animals) gathered during the project have been characterised according to dedicated and validated methods. Results confirm the exposure of treated animals and the status of samples collected from non-treated animals. The detection of steroid esters in blood in the period following administration attested for exposure to administered anabolic cocktails with detected levels reducing over time – similar findings were derived from the characterisation of urine samples from boldenone treated bovines based on boldenone sulphate monitoring. Analytical detection of beta-agonist compounds was applied to various biological matrices acquired from various experiments with different pharmacokinetics in urine exhibiting similar patterns, namely that identification and quantification was feasible from first day of administration and all along the treatment period with beta-agonist concentrations decreasing rapidly and reaching non-detectable levels within a few days. The specific detection of growth hormone was applied to the analysis of serum and plasma from animals treated with Lactatropin. No significant variation was observed in estimated concentrations between analysis in either serum or plasma with bovine somatotropin still detected in samples two weeks after administration.
To conclude specific targeted analytical protocols were applied to the quantitative characterisation of a large number of urine and serum samples from animals (bovine/equine) treated with different compound classes. Samples assessed included those collected from sets of animals treated with the same drug/family, those from animals treated with different anabolic groups and compliant samples from pre-treatment phases or non-treated animals. Corresponding measurements achieved through WP3 enabled verification of the treatment status of samples sourced within WP2, confirming status of exposure to anabolic substances as a pre-requisite to effect-based metabolomic analyses (WP4) and validation studies (WP5).

WP4 Untargeted global metabolomic detection method
The aim of WP4 was to focus on the development and validation of high resolution mass spectrometric methods for the untargeted global metabolomic profiling of urine/plasma (T4.1) with the objective of detecting illegal/inadvertent substance use in animals (equine/bovine) (T4.2) and relating to this the identification of specific metabolite markers of administrations of classes/groups of chemicals (T4.3). WP4 utilised extensive bioinformatical data analysis and multivariate modelling tools to process metabolomic profiles and using discriminant prediction identify samples from animals treated with anabolic agents. Standardised global metabolomics protocols were developed using liquid-chromatography coupled to high resolution mass spectrometry (LC-HRMS and SOPs of procedures generated and applied to samples selected for analysis - minimal sample preparation was conducted for both plasma and urine samples. Urine sample preparation consisted of filtration (10 kDa MWCO) and normalisation to their specific gravity through dilution or freeze-drying followed by a reconstitution to defined dry matter concentration (QUB). Plasma samples preparation protocol was based on the Bligh and Dyer sample preparation process involving liquid-liquid extraction with MeOH, water and MTBE permitting extraction of metabolomic and lipidomic fractions as required. Analyses of the final extracts were performed using LC-HRMS platforms equipped with Orbitrap or Q-ToF mass spectrometers with separation performed using both reverse phase (RP) and hydrophilic interaction (HILIC) chromatography, and full MS spectra recorded in both positive and negative electrospray ionisation (ESI) modes. Acquired MS data was extracted and processed using XCMS and Progenesis® QI software and univariate and multivariate (SIMCA-P) statistical analysis of processed data performed.
Equine samples (urine/plasma) obtained before and after the administration of a cocktail of testosterone esters were assessed by the devised metabolomics workflow and data processing and analysis enabled discrimination between samples collected before and after steroid administration. OPLS-DA score plots were built with all features acquired in either negative or positive modes and corresponding S-plots were extracted to reveal those ions responsible for discrimination with potential as candidate biomarkers representative of exogenous administrations. Further statistical analysis (both univariate and multivariate were performed), revealing features contributing significantly to discrimination between different classes of samples and demonstrating characteristic kinetic profiles (based on ion intensities) along the timeline of the animal experiment study. Following the analysis and statistical processing of the extracted data, urinary and plasmatic metabolomic models were established. Different bioinformatical models constructed (urine in both positive and negative ESI mode, plasma in positive ESI mode, and plasma lipidomics in positive and negative ESI mode) showed clear discrimination between samples collected from control horses and those collected from experimental samples following intra-muscular anabolic steroid administration. These models were subsequently statistically validated through a cross validation test, a permutation test and respecting an established threshold regarding the p-value determined by t-test. Three ions were subsequently selected as relevant and robust candidate biomarkers and combined in a weighted mathematical equation for sample class prediction (treated/non-treated) purposes. The analysis of samples and prediction based on this system from independent experiments (same drug but different animal) and control horses (n=30) was performed. All samples were correctly predicted by the established model in terms of false positive and false negative criteria confirming robustness of the proposed model.
Bovine urine and blood samples collected before and after boldenone treatment were similarly analysed according to the developed metabolomics workflow. Statistical models were built for each analytical platform (urine metabolomics, plasma polar fraction and plasma apolar lipid fractions) in both positive and negative ESI mode. The metabolomics analysis of blood samples did not establish models showing clear discrimination between control and treated animals. However, models built from urine metabolomic profiles were subsequently validated through permutation and cross validation tests. Finally, the urine models (positive and negative ESI mode acquisitions) were combined into a single efficient model presenting good descriptive and predictive capabilities for the classification of unknown samples. Independent bovine urine samples of known and unknown treatment status coming from animals varying in breed, gender, age, geographical region and diet were randomly analysed using the established model to assess the relevance and robustness of the model upon variable factors - all the samples coming from known untreated animals were classified without misallocation.
An additional and alternative profiling strategy was applied to the analysis of equine urine samples. A so-called targeted metabolomics approach which consists of the profiling of specific groups of endogenous metabolites, i.e. steroids in the present case, as potential interesting classes of compounds to discover biomarkers of exposure. Steroidomics studies are based on selective extraction of a large panel (typically between 20 and 30) of endogenous steroids from the biofluids through specific and extensive protocols involving several SPEs purification steps and liquid-liquid extraction before final derivatisation prior to GC-MS analysis. Steroidome profiling has demonstrated its relevance in different contexts related to chemical food safety assessments, and firstly as a very promising tool to screen for forbidden drugs administration. Since the administration of exogenous steroids induces temporal modifications of endogenous steroid profiles as well as related metabolite levels, monitoring for a range of modified endogenous steroid concentrations allows screening for hormone abuse in food producing animals. Successfully applied for anti-doping purposes in human, the strategy has also proven its relevance in both equine and bovine, either in urine or plasma. Application of the strategy to the analysis of the urine samples of the current study is based on a selected set of steroids, which when their measured respective intensities are combined in a weighted equation, enables efficient discrimination of sample groups.
In conclusion samples of known/unknown status using analysed using newly developed metabolomics profiling methodologies and bioinformatics processing tools confirming the predictive capability of untargeted metabolomic approaches to identify samples from treated/untreated animals. This work forms the basis for developing targeted screening methods suitable for routine high-throughput analyses.

WP5 Targeted metabolite marker detection method
The objective of WP5 was to develop quantitative mass spectrometric methods utilizing information acquired from untargeted global metabolomics profiling (WP4) to measure metabolite marker levels representative of misuse/illegal substance administration in bovine/equine animals in urine/plasma in a targeted, reliable, high-throughput and cost-effective manner. Following identification of metabolite markers through comparison of global metabolomics profiles with residue analysis data, advanced chemometric tools constructed discriminant models for data analysis and marker ions contributing to the differentiation between samples of different status were identified. Data processing and analysis performed through WP4 selected discriminative ions as candidate biomarkers which have been combined in weighted formulaic equations to predict sample status. The overarching objective of WP5 was to develop (T5.1) and validate (T5.2) analytical methods based on spectral profiles (T5.3) that can be performed on low resolution LC-MS/MS instrumentation that can form the basis of targeted metabolomic services to be provided by diagnostic testing laboratories using set SOP protocols. Sensitive, reliable and high-throughput methods based on targeted quantification of specific metabolites within metabolomic profiles was therefore developed to provide more specific detection than that offered by non-targeted metabolomic approaches. Whilst the concept of untargeted metabolomics profiles / biomarkers as performed in WP4 has been demonstrated previously, the transfer of biomarker monitoring from a research instrument (i.e. High-resolution MS) to more routine dedicated instruments (e.g. triple quadrupole MS) has never been reported. Work within WP5 has overcome a number of key challenges to deliver this objective.
Specific fragmentation patterns for biomarkers of interest were identified using HRMS to establish SRM transitions for targeted monitoring using routinely employed MS instruments. Chromatographic conditions were adapted to both allow detection of the signal for ions of interest in a known retention time range and to reduce the chromatographic separation time to fulfil high sample throughput requirements. Once targeted biomarkers could be monitored with good specificity and sensibility, attesting for efficient transfer of biomarker monitoring from an untargeted to a targeted approach, attention was paid to their kinetic profiles to ensure that correct signals were monitored and corresponded to expected profiles as observed with the untargeted HRMS workflow. Finally, monitoring of the selected putative biomarkers permitted the profiling of concentration level variation in test samples following administrations and new statistical models obtained based on outputs from low resolution MS analyses permitting threshold levels to be set. The results and predictive ability of the refined routine targeted LC-MS/MS methods and statistical models were in accordance with previous untargeted models (WP4) obtained from HRMS instruments. This facilitates the selected metabolites to be easily monitored in routine laboratory settings using LC-QqQ devices. Main challenges have been pointed out in the present project for such transferability, and solutions and strategies have been successfully proposed. For practical transfer from research to commercial testing environment in host laboratories will require only a minimal optimisation step (as would be done for any targeted method transfer) to ensure good implementation of methodologies in order to be adapted.
Both statistical and biological validations have been conducted to validate the selected markers and related models based on targeted biomarker monitoring. Statistical validation of the models included the assessment of a range of parameters: R2 and Q2 have been used to assess model performances (fitness and predictive ability, respectively), model robustness has been evaluated through p-value (CV ANOVA, T-test). Moreover, the following tests have also been implemented to achieve statistical validation of the models: permutation test, variance importance in projection (VIP), cross-validation. Biological validation consisted of assessing and challenging the markers selectivity (i.e. samples from different anabolic treatments), defining the model sensitivity (detection time window, analytical performances, etc.) to finally define the models scope and set a threshold for screening purposes. For this purpose a large number of additional samples collected from both non-treated animals representing the “normal” variability, and from animals treated with a range of anabolic substances have been analysed.
On this basis all relevant parameters have been successfully assessed enabling subsequent validation to be performed. A challenge test allowed testing samples collected from a range of non-treated animals, which enabled fixing of a threshold limit to define detection time windows which represents the test sensitivity. Such a threshold was defined to enable efficient screening with performances in accordance with screening strategies requirements in terms of false positive/ false negative scores. In the case of equine sample testing, additional “treated” samples collected from different animals which have been exposed to the same anabolic agent as the original samples used to develop models were analysed and sharing similar status with those used for generating of the models. Additionally samples collected from animals treated with various anabolic steroids have been predicted by targeted model to assess the extent to which monitored biomarkers may be used as generic ones to screen for androgen abuse.

Potential Impact:

The development of new techniques and services that can provide more effective, high-throughput and sensitive detection of chemical compound administrations in animals that are validated and which can be commercially exploited will replace existing methods – ranging from ELISA screening methods to mass spectrometry methods - which suffer from issues surrounding high cost, low throughput and overall inefficiency of testing as signified by current low detection rates. DeTECH21 outputs will meet the demand to address major challenges in detecting both inadvertent and illegal administration of compounds to equine and bovine animals which compromise food safety and animal welfare standards and the reputational status of the EU animal production sector. Key benefits arising from DeTECH21 outputs will include:
• Provision of a more comprehensive testing regime to cover a wider range of analytes within a shorter space of time both in terms of analytical sample processing and test result reporting
Greater confidence and integrity attached to results of analytical screening of samples originating from within/outside the EU due to metabolomic analysis
• Reduction of financial costs and time associated with chemical analysis through the development of methods that enable the screening for multiple compounds in a single analysis, increasing the capacity for increased sample coverage and more efficient use of existing budgets
• Increased consumer/public confidence and awareness of improved testing effectiveness relating to regulatory monitoring in food-producing and performance animals.
• Improvements in animal welfare and husbandry by reduction in illegal and inappropriate practices concerning harmful and potentially toxic chemical compounds.
• Potential for increased detection rates of prohibitive and banned substances through the use of effect-based metabolomic analyses which will act a further deterrent to drug misuse.
• Increased confidence in food integrity due to improved screening tests helping to prevent contaminated foods from animals treated with hormonal growth promoters or inadvertent veterinary drug administration, entering the food chain.
• Increased EU competitiveness through the development of advanced analytical methods for chemical contaminants. DeTECH21 will put the EU in a unique worldwide position in the area of food and feed safety control and animal drug control by being the first to deliver commercially available metabolomic-based screening techniques.
• Demonstrate the EU’s willingness and capacity to apply emerging technology and ideas to advance analytical diagnostics and to commercially exploit these new methodologies and create new opportunities for increased graduate level recruitment and employment.
• Analytical service providers will benefit from the application and commercialisation of new techniques and processes developed
• Increased skill and knowledge levels in staff of SME partners in analytical method development which will assist in obtaining accreditation and increasing reputational standing in this field.
• Increase in market reach and access to new international markets outside the EU through offering of improved analytical testing provided by newly developed methods.

Dissemination activities
RTD Performers disseminated project objectives, activities and outcomes at academic symposia and leading international meetings and conferences through both oral and poster presentations. Audiences at these meetings included personnel and representatives of food safety and doping control agencies, regulatory authorities and food producing companies. These conferences included:
• ASSET Food Integrity and Traceability Conference 2014, Queen’s University, Belfast, 8-10 April 2014
• Seventh International Symposium on Hormone and Veterinary Drug Residue Analysis, VDRA 2014, Ghent, 2-5 June 2014
• Chemical Residues Network 2014 Conference “Recent Advances In Targeted and Untargeted Analysis of Food”, Belfast, December 14th 2014
• 7th International Symposium on Recent Advances in Food Analysis, RAFA 2015, Prague, 3-6 November 2015

The research outputs are of novel and high quality and subject to assessment of IPR conflicts will be published in leading peer-reviewed journals in the fields of analytical chemistry, food safety, doping control and metabolomic research helping to increase the profile and international recognition of project researchers within these fields and reach food safety and forensic testing analytical practitioners directly.

Exploitation results
A number of key foreground assets have deriving from the DeTECH21 project including:
• Know-how for the performance of untargeted metabolomics profiling and interpretation with a view to biomarker discovery: extensive sample preparative, analytical mass spectrometry protocols, bioinformatical data analysis and multivariate modelling tools have been developed to enable successful application of metabolomics profiling methods.

• Untargeted global metabolomics profiling methods and models for predictive identification of drug misuse in bovine/equine animal samples

• Indicative panels of urinary/blood metabolite markers of animal exposure to illicit / banned substances identified by untargeted metabolomic

• Validated targeted low resolution MS metabolite marker detection methods enabling high-throughput routine screening of samples for drug misuse in bovine/equine animals

• Spectral database containing global metabolomic signature profiles of test samples and specific metabolite markers of drug misuse

The most appropriate actions required to exploit this foreground have been considered and range from public dissemination in the scientific literature to raise awareness of the potential of these new methods though to maintenance of confidentiality to maximize the economic potential attached to such knowledge.

List of Websites:

Mark Sherry
Irish Equine Centre
Johnstown, Naas
Co. Kildare

Phone: +353 (0)45866266
Mob: +353 (0)872441244