Biomarker-Based Evaluation of a Zearalenone-Degrading Enzyme in Broilers and Piglets Across Multiple Biological Matrices

Abstract

Zearalenone (ZEN) is an estrogenic mycotoxin that impairs animal health and productivity, necessitating effective mitigation strategies in livestock production. This study evaluated the efficacy of the ZEN lactonase ZenA, an enzyme that converts ZEN to non-estrogenic hydrolyzed ZEN (HZEN) and decarboxylated HZEN (DHZEN). Broilers were fed either uncontaminated feed, feed contaminated with 1500 µg ZEN/kg, or ZEN-contaminated feed supplemented with 20 U ZenA/kg for 35 days. Piglets received 200 µg ZEN/kg feed, with or without 10 U ZenA/kg, for 43 days. ZEN biomarkers (ZEN, α-zearalenol, β-zearalenol, HZEN, and DHZEN) were quantified in plasma, urine, feces/excreta, and gastrointestinal contents using liquid chromatography–tandem mass spectrometry. While performance parameters remained unaffected, ZenA supplementation significantly reduced ZEN concentrations (by 19.6–66.2%) in all matrices and at all time points in both species. In addition, significant formation of HZEN was observed in gastrointestinal samples. Thus, in the present study, ZenA efficiently degraded ZEN in both broilers and piglets. Biomarker analysis in multiple matrices provided complementary insights: gastrointestinal samples confirmed the enzyme’s mode of action, while plasma and urine data showed a marked reduction in systemic ZEN exposure. Finally, the results reinforce that performance parameters are insufficient for assessing the efficacy of mycotoxin-detoxifying feed additives and support biomarker-based evaluation approaches.

1. Introduction

Mycotoxins, toxic secondary metabolites produced by various molds, are common contaminants in animal feed. They are known to cause a wide range of adverse health effects, ultimately reducing animal productivity and causing economic losses. In addition, the environmental consequences of mycotoxin contamination have recently been assessed, revealing a broad environmental impact across multiple environmental footprint categories in broiler production. These effects are mainly driven by the increased feed production, as well as phosphorus and nitrogen excretion, and include, for example, an 8.5% increase in carbon footprint and significant rise in global warming potential [1].

Zearalenone (ZEN) is a mycotoxin produced by different Fusarium species. It is a frequent contaminant of cereals and cereal-derived foods products, e.g., with reported prevalence rates ranging from about 15% in Asia to more than 50% in Africa and Europe [2]. Similarly, a global 10-year feed survey found that, depending on the region, 19.6% to 52.2% of feed commodities were contaminated with ZEN [3]. Median ZEN concentrations in finished feed samples were 41 µg/kg, while maximum levels exceeded 9400 µg/kg. ZEN’s primary mode of action is activation of estrogen receptors α (full agonist) and β (partial agonist), thus inducing morphological and functional changes in the reproductive system and signs of hyperestrogenism in various species [4,5]. Beyond its role as an endocrine disruptor [6], ZEN can also affect other organ systems, such as the intestine or immune system [7,8]. Following ingestion, ZEN is rapidly absorbed, metabolized, and excreted via urine and feces after enterohepatic recycling. However, notable interspecies differences exist, including the site of metabolism (pre-absorptive in ruminants), bioavailability (low in poultry), and metabolic pathways [9,10]. Specifically, ZEN is reduced to α- and β-zearalenol (α-ZEL, β-ZEL), with α-ZEL exerting higher estrogenic potency than β-ZEL [4]. Thus, the rate of α-reduction, which is comparatively high in pigs, contributes to differences in species-dependent susceptibility to ZEN [11]. Other metabolic routes include hydroxylation and glucuronidation, which also vary by species [12]. Carry-over of ZEN from feed to food of animal origin, such as muscle, eggs or milk, is limited [9]. Hence, the contribution of ZEN residues in foodstuff of animal origin to overall consumer exposure is considered negligible [13].

Regulatory aspects regarding ZEN in feed vary globally. EU guidance values for ZEN in feed reflect species-dependent differences in susceptibility. For swine, the current recommended maximum level ranges from 100 µg ZEN/kg for piglets and gilts to 250 µg ZEN/kg for fattening pigs and sows. In contrast, there are no official guidance values for poultry finished feed. More generally, maize by-products may contain up to 3000 µg ZEN/kg, while other cereals and cereal products may contain up to 2000 µg ZEN/kg [14]. Notably, long-term exposure to this level has been shown to impair gut health and cause hepatotoxicity in laying hens [15], whereas concentrations as low as 750 µg ZEN/kg induced alterations in egg production and hormone levels [16]. In addition to ZEN limits for pig feed (100 µg/kg for sow and 150 µg/kg for piglet corn-based compound feed, respectively), China has established limits for ZEN in poultry corn-based compound feed of 500 µg/kg [17]. In contrast, the US Food & Drug Administration (FDA) has not set guidance levels for ZEN in feed [18].

Due to its frequent occurrence and negative impact on animal production, effective mitigation strategies for ZEN are essential. These include pre-harvest measures to reduce fungal growth and toxin formation, as well as post-harvest strategies to counteract residual contamination, since complete prevention is nearly impossible [19,20]. Physical or chemical post-harvest techniques, such as washing, sorting, thermal processing, or treatments with ozone, sulphur dioxide, or sodium carbonate, often face limitations related to the efficacy of reducing ZEN, impairment of feed quality, costs or applicability under field conditions [19,21]. An alternative post-harvest strategy involves feed additives that either adsorb or biotransform ZEN in the gastrointestinal (GI) tract. As the binding efficiency of adsorbents for ZEN is relatively low [22,23], microbial or enzymatic transformation of ZEN into less or non-toxic metabolites is considered a more promising approach. Several microbial strains and purified enzymes have been reported to degrade ZEN (summarized by [24]). However, progressing from an enzyme with basic in vitro activity to a functional feed additive requires fulfilling certain key criteria: the reaction must be specific and irreversible; the resulting metabolites must be significantly less toxic than the parent compound; and the enzyme must be readily active in the GI tract to degrade ZEN prior to absorption as well as to mitigate local effects. Additionally, its efficacy must be demonstrated in animal trials using relevant biomarkers, and the enzyme must remain stable during storage and feed processing [25].

The zearalenone lactonase ZenA is an enzyme that degrades ZEN by hydrolyzing the ester bond of its lactone ring, converting it to hydrolyzed zearalenone (HZEN) ([26], Figure 1). HZEN partly decarboxylates to form decarboxylated hydrolyzed zearalenone (DHZEN). Most importantly, these metabolites neither activate estrogen receptors in vitro, nor induce estrogenic symptoms in vivo [24]. In addition to their markedly reduced estrogenic activity compared with the parent toxin ZEN, HZEN and DHZEN exhibit substantially lower spermatotoxic and immunotoxic effects [27,28,29], confirming that this enzymatic conversion constitutes a detoxification reaction. ZenA, as an enzyme, is a protein that is degraded during digestion in the animal and therefore does not pose a risk to the environment after excretion. The efficacy of ZenA in degrading ZEN in the gastrointestinal tract has been demonstrated previously in dairy cows, pigs, chickens, and rainbow trout [30,31,32]. However, those studies often lacked assessment of systemic exposure, e.g., by measuring ZEN and its metabolites in blood, or did not apply enzyme concentrations reflective of practical livestock production.

Therefore, the present study aimed to expand current knowledge on the efficacy of ZenA by conducting feeding trials in two livestock species, applying economically relevant enzyme concentrations and employing biomarker analysis in multiple biological matrices, including blood plasma. This approach allowed evaluation of local versus systemic effects, as well as matrix-dependent variability in biomarker detection. The results show that ZenA effectively degrades ZEN in the gastrointestinal tract of both species, thereby reducing systemic toxin exposure. Moreover, the study provides additional insights into the toxicokinetics of ZEN in broilers and pigs.

2. Materials and Methods

2.1. Animal Experiments

2.1.1. General

The animal experiments were carried out at the Center of Applied Animal Nutrition (dsm-firmenich, Animal Nutrition and Health, Mank, Austria). All procedures for animal handling, care and treatment have been approved by the institutional ethics committee and by the Austrian Federal Ministry of Education, Science and Research (broiler trial specific authorization number: BMBWF LF1-TVG-57/038-2023; piglet trial specific authorization number: LF1-TVG-56/035-2022). The dietary ZEN concentrations were selected to remain below (broiler trial) or approximate (piglet trial) the EU guidance values for ZEN in feed [14], while ensuring measurable biomarker concentrations, particularly in blood, as further detailed in Section 3.1.1 and Section 3.2.1. ZenA inclusion levels were defined based on previous experimental data and economic feasibility, allowing evaluation of enzyme performance under practical livestock conditions.

2.1.2. Broiler Trial

1-day old, mixed sex Ross 308 broiler chickens were obtained from a local producer and randomly allocated to three experimental groups, while ensuring balanced body weight across pens (8 animals/pen) and groups (11 replicate pens/group, except for Control, which had 10 replicates). Chickens received either (i) uncontaminated basal feed (Control), (ii) feed contaminated with 1500 µg ZEN/kg (ZEN) or (iii) feed contaminated with 1500 µg ZEN/kg and supplemented with 20 units (U) zearalenone lactonase ZenA/kg (ZEN + ZenA) for 35 days. Chickens were housed on wood shavings, and environmental conditions were controlled according to the breeding company’s recommendations, encompassing an 18/6 h light/dark cycle from day two onwards. The general condition of the chickens was monitored daily, and the animals were weighed individually at trial start, day 14 and day 35.

Feed and water were provided ad libitum. Chickens were fed a starter diet (phase I feed) for the first 14 days of the trial, and a grower diet thereafter (phase II feed) (Table S1). For artificial ZEN contamination of treatment diets (ZEN, ZEN + ZenA), culture material of Fusarium verticillioides, containing 1.4 g ZEN/kg, was obtained from Romer Labs GmbH (Tulln, Austria). To ensure homogeneous distribution of ZEN within the diets, premixes were prepared using corn meal, which were added to the basal feed at an inclusion rate of 0.8% for phase I and 0.3% for phase II. Similarly, supplementation of treatment diets with ZenA (ZEN + ZenA) was achieved by preparing premixes of a ZenA preparation (ZENzyme^®, EC 3.1.1.-; BIOMIN Holding GmbH, Getzersdorf, Austria) and corn meal. Subsequently, these premixes were added to the diet at an inclusion rate of 0.8% for phase I and 0.3% for phase II. Final mycotoxin concentrations were assessed via liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis (Romer Labs GmbH, Austria; University of Natural Resources and Life Sciences, Vienna, Austria). Details on sample preparation, LC-MS/MS analysis, method validation as well as limits of detection and quantification are described in Gruber-Dorninger et al. [33]. The results showed average ZEN contamination levels of 40 µg/kg, 1560 µg/kg and 1585 µg/kg for the Control, ZEN and ZEN + ZenA group, respectively. Natural contamination of diets with other major mycotoxins was absent (ochratoxin A), marginal (aflatoxin B1 0.05 µg/kg, ergot alkaloids 0.07 µg/kg, T-2 toxin 1.3 µg/kg) or markedly below EU recommendations for complete feed in poultry (deoxynivalenol, DON, 100.9 µg/kg, fumonisin B1 503.9 µg/kg; [14]). For determination of ZenA activity in the diets, 20 g of the feed sample were extracted with 50 mL Teorell Stenhagen buffer (pH 7.5) containing 0.1 mg/mL bovine serum albumin (BSA) by shaking at room temperature for 1 h. The extract was cleared by centrifugation to obtain supernatant containing ZenA. Subsequently, the supernatant was incubated in Teorell Stenhagen buffer pH 7.5 with 0.1 mg/mL BSA and 5 mg/L ZEN at 37 °C. Samples were collected over 1.5 h, heat-inactivated at 95 °C for 5 min and analyzed for ZEN concentration by high-performance liquid chromatography with a fluorescence detector. ZenA activity was calculated from the linear decline of ZEN over time. One U of ZenA is defined as the five-fold enzymatic activity that hydrolyses 1 μmol ZEN per minute in a solution of 5 mg/L ZEN (15.71 µM) in Teorell Stenhagen buffer at pH 7.5 with 0.1 mg/mL BSA at 37 °C. The ZenA activity measured in the reaction was multiplied by the dilution factor from extraction to incubation to obtain ZenA activity in the diet (U/kg). The average ZenA concentration in broiler diets was 19.6 U/kg.

After five weeks, animals were euthanized by carbon dioxide stunning and subsequent exsanguination. Individual blood (plasma), crop and gizzard content, as well as excreta samples, were collected. Blood samples were collected from the jugular vein into lithium heparin tubes (Sarstedt, Nürnbrecht, Germany). GI content samples were collected by removing and opening crop and gizzard and transferring their contents into sampling tubes. Excreta samples were collected from containers used for individual euthanasia of chickens. Containers were disinfected after each bird. All samples were placed on dry ice until sampling was completed and then stored at −20 °C at the laboratory until analysis.

2.1.3. Piglet Trial

The trial was conducted in two separate animal rooms, with pens and treatment groups evenly distributed across both rooms. To account for logistical constraints, the trial was carried out in two consecutive rounds, staggered by 21 days. For each round, newly weaned piglets of the same age and weight distribution were used to ensure consistency across both trial periods. All treatment diets were prepared in a single batch and used for both rounds to maintain uniform feed composition.

Weaned piglets (sow: Landrace × Large White, boar: Pietrain, 27 ± 1 days old, average weight 7.1 ± 0.6 kg, mixed sex) were obtained from a local producer. Piglets were randomly allocated to two different treatment groups, while ensuring a balanced distribution of body weight and sex across pens (6 piglets/pen) and groups (12 replicate pens/group). After an acclimatization phase of six days, in which all groups received uncontaminated basal feed, piglets were either exposed to (i) feed contaminated with 200 µg ZEN/kg (ZEN) or (ii) feed contaminated with 200 µg ZEN/kg and supplemented with 10 U ZenA/kg (ZEN + ZenA) for 43 days. Piglets were housed in pens on slatted floors, with pig toys for enrichment. The general condition of the piglets was monitored daily, and the animals were weighed individually at trial start, day 14, 35 and 42, respectively.

Feed and water were provided ad libitum. Piglets were fed a phase-feeding regimen consisting of phase I (trial start to day 14), phase II (day 15 to day 35), and phase III (day 36 until trial end). The composition of individual diets is provided in Table S1. For artificial ZEN contamination of treatment diets, culture material of Fusarium verticillioides, containing 26.5 g ZEN/kg, was obtained from Romer Labs GmbH (Tulln, Austria). The culture material-corn meal-premixes were added to the basal feed at an inclusion rate of 0.2% for phase I and 0.1% for phase II and III, respectively. Supplementation of treatment diets with ZenA (ZEN + ZenA) was performed as described in Section 2.1.2, except for the inclusion rates of premixes, which were 0.2% for phase I and 0.1% for phase II and III, respectively. Final mycotoxin concentrations were assessed as described in Section 2.1.2. The results showed average ZEN contamination levels of 182 µg/kg and 230 µg/kg for the ZEN and ZEN + ZenA group, respectively. Natural contamination of diets with other major mycotoxins was absent (aflatoxin B1, ochratoxin A, ergot alkaloids), marginal (T-2 toxin 1.7 µg/kg), or markedly below EU recommendations for complete feed in piglets (DON 175.3 µg/kg, fumonisin B1 29.4 µg/kg; [14]). Analysis of ZenA concentrations in diets and the definition of ZenA U are provided in Section 2.1.2. The average ZenA concentration in piglet diets was 8.0 U/kg.

Individual blood (plasma), urine and feces samples were collected from piglets on day 7, 13, 26 and 42 in the morning. Blood samples were drawn from the jugular vein into EDTA tubes (Sarstedt, Nümbrecht, Germany). Urine samples were collected during spontaneous micturition. Due to the non-invasive sampling method, urine could not be obtained from all animals; however, a minimum of four out of six piglets per pen were successfully sampled. Feces samples were collected during defecation. If a spontaneous feces sample was not obtained by noon, rectal stimulation was performed. All samples were placed on dry ice until sampling was completed and then stored at −20 °C at the laboratory until analysis.

2.2. Analysis of ZEN Biomarkers in Animal Specimens

Analysis of ZEN biomarkers comprised LC-MS/MS based quantification of total ZEN, α-ZEL, β-ZEL, HZEN and DHZEN in plasma (piglet and broiler), urine (piglet), feces (piglet), excreta (broiler) and crop as well as gizzard contents (broiler).

2.2.1. Chemicals and Reagents

Analytical standards for ZEN, α-ZEL, β-ZEL, and ¹³C-labelled zearalenone (¹³C ZEN) were purchased from Romer Labs (Tulln, Austria). Deuterium labelled standards for α-ZEL and β-ZEL were obtained from Honeywell (Charlotte, NC, USA). HZEN and DHZEN, as well as their ¹³C-labelled analogues, were produced in-house by enzymatic degradation of ZEN or ¹³C-labelled ZEN, respectively, followed by acidified incubation at high temperatures to promote the conversion from HZEN to DHZEN. The purchased standards were stored according to manufacturers’ recommendations. HZEN and DHZEN as well as their isotope-labelled forms were stored at −20 °C. Acetonitrile (ACN) HPLC grade and acetic acid were ordered at VWR (Vienna, Austria). ACN LC-MS/MS grade for solvents and ethanol absolute (100%) were purchased from Chem-Lab (Zedelgem, Belgium). Sodium chloride (NaCl) and β-glucuronidase Type IXA from E. coli were bought from Merck (Darmstadt, Germany). PBS buffer was obtained as powder from AppliChem (Darmstadt, Germany). Ultrapure water was produced in-house using a Milli-Q IQ 7000 water purification system from Merck (Darmstadt, Germany).

2.2.2. Sample Preparation

The method for analyzing plasma samples was further developed based on the protocol described by Dänicke et al. [32]. Samples were thawed at room temperature and mixed well. Afterwards, an aliquot of 1000 µL was transferred to a 15 mL tube, to which internal standards as well as a 100 µL β-glucuronidase solution (10,000 U/mL in 1× PBS) were added. Samples were incubated overnight at 37 °C at 80 rpm. After incubation, proteins were precipitated with 4 mL ACN, mixed for 10 min on an end-over-end shaker followed by centrifugation for 5 min at 3200 rcf. The supernatants were transferred to fresh falcon tubes containing 1 g NaCl, mixed and centrifuged again for 5 min at 3200 rcf. To concentrate the samples, 2 mL of supernatants were dried under reduced pressure at 45 °C. Samples were reconstituted in 50 µL ethanol/water (50/50 v/v) for 10 min on a vortex mixer, followed by centrifugation for 5 min at 19,000 rcf. Finally, the supernatants were transferred to HPLC vials with silanized inserts and analyzed by LC-MS/MS.

The creatinine (Crea) concentration of urine samples was determined by LC-MS/MS as described by Streit et al. [34]. Sample preparation of urine samples for ZEN biomarker analysis was performed according to Dänicke et al. [32], with some modifications. First, samples were individually diluted to 0.5 mM Crea using ultrapure water. Then, internal standards and 10 µL β-glucuronidase solution (10,000 U/mL in 1× PBS) were added to 100 µL sample aliquots. Finally, the protein precipitation step after enzyme incubation was omitted. The samples were centrifuged for 5 min at 19,000 rcf, and the supernatants were transferred to HPLC vials with silanized inserts for LC-MS/MS analysis.

Freeze-dried feces samples from piglets were homogenized and extracted according to Dänicke et al. [32]. Similarly, excreta and GI contents (crop, gizzard) from broilers were processed following the same protocol with slight modifications. The sample amounts were reduced to 250 mg and the extraction solvents were adapted accordingly. Additionally, after extraction, internal standards were added to the extracts and used for quantification.

2.2.3. Chromatography and LC-MS/MS Parameters

Samples were subjected to analysis by LC-MS/MS in negative multiple reaction monitoring (MRM) mode as described by Dänicke et al. [28], with adjustments to the instruments used (Agilent 1290 Infinity II coupled to a QTRAP 6500+; Agilent Technologies, Waldbronn, Germany; AB Sciex, Singapore), injection volumes (10 µL for plasma, 20 µL for urine, and 2 µL for feces, excreta and GI contents) and extension to include isotope-labelled analogues of ZEN and its metabolites. Analytes were separated by gradient elution using water/ACN (95/5 v/v) containing 0.1% acetic acid as eluent A, and ACN/water (95/5 v/v) containing 0.1% acetic acid as eluent B. The gradient used was: 0 min–0.1 min 0% B, 0.1 min–2.2 min linear increase to 30% B, 2.2 min–5.8 min linear increase to 55% B, 5.8 min–5.9 min linear increase to 100% B, 5.9 min–6.4 min constant at 100% B, 6.4 min–6.5 min linear decrease to 0% B, 6.5 min–7.5 min constant at 0% B. The measured MRM transitions are summarized in

Table 1. Multiple reaction monitoring mode transitions for the quantification of zearalenone (ZEN), α-zearalenol (α-ZEL) and β-zearalenol (β-ZEL), hydrolyzed zearalenone (HZEN), decarboxylated hydrolyzed zearalenone (DHZEN), as well as their isotope-labelled analogues. The stereoisomers α-ZEL and β-ZEL are measured with the same transitions but differentiated by their retention times.

Analyte	Measured Form	Precursor Ion (m/z)	Quantifier Ion Qualifier Ion (m/z)	Declustering Potential (V)	Collision Energy (V)	Collision Cell Exit Potential (V)	Entrance Potential (V)	Retention Time (min)
ZEN	[M-H]−	317	131 175	−100	−42 −34	−9 −7	−10	5.43
ZELs	[M-H]−	319	160 275	−105	−42 −30	−9 −7	−10	4.19 (β-ZEL) 4.76 (α-ZEL)
HZEN	[M-H]−	335	149 161	−80	−34	−1 −9	−10	3.85
DHZEN	[M-H]−	291	123 161 41	−110 −80 −110	−18 −25 −112	−7 −8 −7	−10	3.22
13C ZEN	[M-H]−	335	184 140	−5	−34 −40	−21 −11	−10	5.43
D5 ZELs	[M-H]−	324	280 160	−175	−28 −40	−17 −19	−10	4.19 (β-ZEL) 4.76 (α-ZEL)
13C DHZEN	[M-H]−	308	158 171	−35	−26	−11 −9	−10	3.22
13C HZEN	[M-H]−	353	158 171	−165	−34 −30	−13 −7	−10	3.85

, and a chromatogram of a standard is displayed in Figure S2. Quantitation was based on standards in neat solvent using 1/x weighted linear regression. Additionally, internal standard correction was applied for all matrices except swine feces.

2.2.4. Method Validation Plasma and Urine

Method validation was performed for plasma and urine samples, as these methods were further developed. To this end, whole blood from pigs was collected in tubes containing 4.8 mg/mL EDTA at a local slaughterhouse. Plasma was separated by centrifugation for 30 min at 3000 rcf and spiked with different analyte concentrations, ranging from 0.05–10 ng/mL. Blank urine samples from a previous piglet feeding trial were used for method validation. Crea content was determined, and samples were diluted with ultrapure water to 0.5 mM Crea prior to spiking and extraction. The same spiking levels were used for plasma and urine validation.

Accuracy and precision were determined on three different days, each at least in duplicate. One day was performed in triplicate to calculate intraday precision. Acceptance criteria were set at analyte recoveries of 70–130% for accuracy and relative standard deviation (RSD) ≤ 15% for precision. All samples were measured within 24 h of extraction. The limit of quantification (LOQ) was determined by spiking experiments and is the smallest concentration where accuracy and precision acceptance criteria were met. Selectivity of the method is given by the retention time of the analyte and the specified MRM transitions. Obtained validation data are shown in

Table 2. Summary of validation results for zearalenone (ZEN), α-zearalenol (α-ZEL) and β-zearalenol (β-ZEL) (ZELs), hydrolyzed zearalenone (HZEN) and decarboxylated hydrolyzed zearalenone (DHZEN) in pig plasma.

		ZEN	α-ZEL	β-ZEL	HZEN	DHZEN
Linear range (ng/mL)		0.125–100	0.125–100	0.125–100	0.125–100	0.125–100
Concentration in plasma (ng/mL)		0.05–10	0.1–10	0.1–10	0.1–10	0.1–10
LOQ in plasma (ng/mL)		0.05	0.1	0.1	0.1	0.1
Accuracy (%)	Low 0.05 ppb	82–114	NA	NA	NA	NA
	Low 0.1 ppb	91–116	85–116	87–117	83–117	93–119
	Middle 1 ppb	91–110	84–116	85–116	88–103	92–114
	High 10 ppb	88–108	80–106	82–115	88–102	90–112
Precision (interday, %)	Low 0.05 ppb	12.1	NA	NA	NA	NA
	Low 0.1 ppb	8.9	9.8	10.0	10.0	8.5
	Middle 1 ppb	6.1	10.6	13.5	4.9	6.1
	High 10 ppb	6.1	11.1	14.3	3.7	6.4
Precision (intraday, %)	Low 0.05 ppb	6.8	NA	NA	NA	NA
	Low 0.1 ppb	5.0	1.3	5.2	3.5	3.7
	Middle 1 ppb	4.5	2.3	1.7	2.7	4.9
	High 10 ppb	4.8	5.0	3.6	4.3	6.4

NA, not applicable.

(plasma) and

Table 3. Summary of validation results for zearalenone (ZEN), α-zearalenol (α-ZEL) and β-zearalenol (β-ZEL) (ZELs), hydrolyzed zearalenone (HZEN) and decarboxylated hydrolyzed zearalenone (DHZEN), in pig urine diluted to 0.5 mM creatinine (Crea).

		ZEN	α-ZEL	β-ZEL	HZEN	DHZEN
Linear range (ng/mL)		0.05–10	0.05–10	0.05–10	0.05–10	0.05–10
Concentration in urine at 0.5 mM Crea (ng/mL)		0.05–10	0.1–10	0.1–10	0.1–10	1–10
LOQ in urine at 0.5 mM Crea (ng/mL)		0.05	0.1	0.1	0.1	1
Accuracy (%)	Low 0.05 ppb	99–123	NA	NA	NA	NA
	Low 0.1 ppb	86–123	85−116	87–113	82–119	NA
	Middle 1 ppb	106–130	78–100	83–121	100–119	98–125
	High 10 ppb	102–117	77–98	90–123	101–128	103–115
Precision (interday, %)	Low 0.05 ppb	7.8	NA	NA	NA	NA
	Low 0.1 ppb	13.9	10.0	9.8	12.9	NA
	Middle 1 ppb	7.7	7.8	13.3	6.5	7.5
	High 10 ppb	3.9	8.2	10.2	8.2	5.1
Precision (intraday, %)	Low 0.05 ppb	6.6	NA	NA	NA	NA
	Low 0.1 ppb	8.4	5.4	10.2	8.4	NA
	Middle 1 ppb	3.0	2.0	2.0	2.4	6.9
	High 10 ppb	1.4	4.0	4.7	1.2	3.1

NA, not applicable.

(urine).

2.3. Statistics

For evaluation of broiler performance data, which were adjusted for mortality, ANOVA was used when normality and variance homogeneity were met; otherwise, the Kruskal–Wallis test was used. Piglet performance data were analyzed in RStudio (v1.3.1093) using readxl, plyr, Surv, lattice, car, nlme, and emmeans. Mixed-effects models, generalized least squares and robust regression models were employed, considering any interactions between time, group and run, which were found, tested and found non-significant. Model assumptions (normality, homoscedasticity, independence) were assessed via the residual plots. Two-tailed tests were applied for both species, as no directional effect of treatments on performance was expected.

For evaluation of biomarker data, the pen was considered the experimental unit. Mean values per pen were calculated from individual animal data. For values below the LOQ, half of the LOQ was used. In cases where duplicate analyses yielded one quantifiable result and one below the LOQ, the mean was calculated using the measured value and LOQ/2. Statistical analyses were performed in R (version 4.3.2.; R Core Team, R Foundation for Statistical Computing, Vienna, Austria). The tidyverse package (v2.0.0) was used for data wrangling, car (v3.1–3) for Levene’s test of homogeneity of variances, and PMCMRplus (v1.9.12) for non-parametric multiple comparisons.

For biomarker data in broilers, Kruskal–Wallis tests followed by Bonferroni-corrected Conover post hoc tests were conducted, except when fewer than three values exceeded the LOQ across all pens. The null hypothesis assumed no differences in biomarker levels between the three test groups. As the test was non-directional, the comparisons were two-sided. For piglets, pairwise comparisons were conducted for each parameter at each sampling point except when fewer than three values exceeded the LOQ across all pens. The null hypothesis assumed no differences in biomarker levels between the ZEN and ZEN + ZenA group. Comparisons were one-sided, as the enzyme’s effect was expected to be unidirectional (i.e., ZEN decrease; HZEN increase). Prior to group comparisons, normal distribution (Shapiro–Wilk test) and variance homogeneity (Levene’s test) were assessed. Non-normally distributed data were analyzed using Wilcoxon’s rank sum test, while normally distributed data were subjected to Student’s t-test, with Welch’s correction applied if Levene’s test indicated unequal variances. Comparisons of the sum of ZEN, α-ZEL and β-ZEL across days were conducted using Kruskal–Wallis tests followed by Dunn’s allpairs rank comparison test.

Statistical significance was set at p ≤ 0.05, and figures were generated in GraphPad Prism (v10.2.2), with standard boxplots displaying the interquartile range (Q1–Q3), median (Q2) and minimum and maximum values.

3. Results and Discussion

The efficacy of ZenA in degrading ZEN in the GI tract was investigated in both broilers and piglets by analyzing ZEN and its metabolites in various biological specimens. In addition, performance parameters were recorded to assess potential effects of ZEN exposure and/or ZenA supplementation on animal performance.

3.1. Broiler Trial

3.1.1. Performance

Independent of the observation period (days 0–14, 14–35, 0–35), treatment diets had no effect on body weight, body weight gain, daily body weight gain, average daily feed intake or feed conversion ratio (FCR) of broilers (Table S2). These results correspond well with literature, suggesting poultry to be comparably tolerant to ZEN [4]. For example, dietary ZEN concentrations of up to 800,000 µg/kg fed to 6-week-old broilers for 3 weeks had no impact on body weight gain or feed consumption [35]. A more recent study found that diets contaminated with 2000 µg ZEN/kg reduced weight gain and feed conversion in broiler chickens [36]. Moreover, ZEN may adversely affect performance at even lower concentrations under practical farm conditions, considering that the potency of artificial and natural ZEN contamination may differ–as has been demonstrated for DON [37]–and that co-occurring mycotoxins can amplify its effects. Indeed, Kolawole et al. [38], who conducted a long-term feeding trial with naturally contaminated diets containing low mycotoxin levels, reported a positive correlation between ZEN (maximum concentration in feed: 241 µg/kg) and FCR in broilers. Furthermore, the occurrence of DON and fumonisins or diacetoxyscirpenol, respectively, was predicted to have negative effects on FCR, when ZEN was present at concentrations below 100 µg/kg.

3.1.2. Biomarkers

ZEN, its phase I metabolites α-ZEL and β-ZEL as well as enzymatic degradation products HZEN and DHZEN were measured in plasma, crop as well as gizzard content and excreta. In general, ZEN is described to be little bioavailable in broilers, reaching only 8.34% [10] or 29.66% [39] after oral administration of pure toxin. In addition, it is rapidly excreted, with elimination half-life values between 0.34 and 1.36 h [10,39]. Hence, sufficient dietary ZEN doses as well as sensitive analytical methods are required to enable quantification of this mycotoxin in blood. Otherwise, ZEN might remain undetectable in blood, both under experimental and field conditions [40,41].

In the present trial, the study design and applied LC-MS/MS method allowed quantification of ZEN in plasma, which ranged from 0.08–0.21 ng/mL (mean: 0.14 ng/mL) in broilers of the ZEN group (Figure 2A). Only individual birds in the ZEN group showed quantifiable α-ZEL levels, whereas pen-based concentrations remained below the LOQ. For β-ZEL, levels above the LOQ were not detected in any individual bird or pen. While in vitro studies suggest a predominant formation of β-ZEL in poultry [11], thus partly explaining the comparatively low susceptibility of this species to ZEN, in vivo results are inconsistent regarding the primary phase-I metabolite detected in broiler blood following oral ZEN administration. Devreese et al. [10] reported higher β-ZEL than α-ZEL levels, whereas Yunus et al. [42] detected only α-ZEL, with no quantifiable β-ZEL. Similarly, Buranatragool et al. [39] found α-ZEL concentrations exceeding the ones of β-ZEL at most sampling time points. Our study supports the notion of predominant α-ZEL formation in poultry. Still, further research is needed to fully elucidate ZEN metabolism in broilers, especially at low toxin levels and when using naturally contaminated feeds, and to clarify potential discrepancies between in vitro and in vivo results.

Supplementation of ZEN-contaminated diets with ZenA significantly reduced ZEN levels in plasma (range 0.08–0.14 ng/mL, mean: 0.11 ng/mL), corresponding to a 23.2% reduction in the mean value compared to the ZEN group. Statistical evaluation of α- and β-ZEL concentrations between treatments was not feasible due to insufficient number of replicate pens with concentrations above the respective LOQs. Similarly, quantifiable HZEN levels were only observed for one pen, precluding statistical analysis. The lower plasma concentrations of HZEN compared to those in the GI contents and excreta can be explained by the limited absorption of this enzymatic degradation product (factor 73.4 lower than parent compound based on in vitro absorption assay). DHZEN concentrations remained below the LOQ, not only in plasma, but in all matrices analyzed. This was expected as HZEN represents the primary metabolite formed by ZenA and undergoes only partial conversion to DHZEN.

The reduced plasma levels of ZEN in the ZEN + ZenA group are the consequence of effective degradation of ZEN in the GI tract, as evidenced by biomarker results from crop and gizzard contents as well as excreta (Figure 2B–D). A significant decrease in ZEN concentrations (crop: 66.2% reduction, gizzard: 23.6%, excreta: 28.7%) and concurrent formation of HZEN were observed in all those matrices following ZenA supplementation. α- and β-ZEL were not detected in crop and gizzard contents, which is consistent with literature suggesting limited pre-absorptive formation of these phase-I metabolites in monogastric animals [8]. In line with findings by Yunus et al. [42], absolute and relative levels of α-ZEL and β-ZEL increased from proximal to distal segments of the GI tract, reaching their highest concentrations in excreta. In this matrix, the ZEN + ZenA group showed significantly reduced β-ZEL levels than the ZEN group, while no significant differences between those treatments were observed for α-ZEL.

3.2. Piglet Trial

3.2.1. Performance

Treatment diets had no effect on body weight, body weight gain, daily body weight gain, feed intake or FCR of piglets (Table S3). As ZEN primarily exerts estrogenic effects [4], and given that the applied contamination level was below the no observed adverse effect level (NOAEL) of 220 µg ZEN/kg feed for piglets [43,44], these results were expected. They further emphasize that performance parameters alone are insufficient to evaluate the efficacy and mode of action of mycotoxin-inactivating feed additives, particularly at low and moderate dietary mycotoxin concentrations.

3.2.2. Biomarkers

ZEN, its phase I metabolites α-ZEL and β-ZEL as well as enzymatic degradation products HZEN and DHZEN were measured in piglet plasma, urine and feces on days 7, 13, 26 and 42.

Often, mycotoxin doses exceeding current legislation are applied in experiments assessing the efficacy of mycotoxin-inactivating feed additives in order to obtain detectable toxin concentrations in blood [45]. For instance, previous studies evaluating the efficacy of ZEN-inactivating feed additives by measuring exposure-based biomarkers in pig blood applied dietary ZEN concentrations of 240 µg ZEN/kg [32], 300 µg ZEN/kg [46], 970 µg ZEN/kg [47] or 75,000 µg ZEN/kg [45]. While still exceeding the EU guidance value for ZEN in pig feed [14], a comparatively low contamination level was used in the present study. Hence, the LC-MS/MS method described by Dänicke et al. [32] was further optimized in this study to enable detection of ZEN, α-ZEL, and β-ZEL, achieving sensitivity comparable to–or, in most cases, exceeding–that of previously published methods for analyzing ZEN and its metabolites in swine blood (e.g., [32,41,48,49,50,51,52,53,54]. ZEN-glucuronide has been suggested as optimal plasma biomarker for ZEN exposure in pigs [45]. Due to the lack of a commercially available standard for quantification of ZEN-glucuronide, enzymatic hydrolysis was included in the sample preparation to enable determination of total ZEN levels in blood.

On day 42, plasma concentrations of ZEN in the ZEN group ranged from 0.23 to 0.40 ng/mL, with a mean of 0.31 ng/mL (Figure 3A). As reported in previous studies [9], α-ZEL predominated over β-ZEL in blood, the latter remaining below its LOQ at all sampling time points. Unexpectedly, α-ZEL levels were lower than those of its parent toxin, which contrasts with most studies analyzing ZEN and its phase-I metabolites in pig blood [9]. Exceptions include findings by Benthem de Grave et al. [55] for piglets, but also by Catteuw et al. [56] who observed α-ZEL formation only in 8-week-old but not in 4-week-old piglets. Hence, these authors proposed an age-related reduction in 3α-hydroxysteroid dehydrogenase activity in the liver of weaned piglets. In line with this hypothesis, α-ZEL was detected only in few individual piglets at earlier sampling points but exceeded the LOQ in a relevant number of pens only by day 42. In contrast, a recent study reported α-ZEL concentrations exceeding those of ZEN in piglets of comparable age at study onset and with similar trial as in the present study [32]. These discrepancies highlight the need for further research to elucidate the factors influencing ZEN metabolism in piglets, especially under conditions reflecting practical production environments.

Supplementation of ZEN-contaminated diets with ZenA significantly reduced ZEN levels in plasma (range 0.15–0.29 ng/mL; mean: 0.22 ng/mL), corresponding to a 29.6% reduction in the mean value compared to the ZEN group. Similarly, significantly lower ZEN concentrations were detected in the urine of the ZEN + ZenA group (Figure 3B, 32.8% reduction). However, this matrix inhibited greater individual variation in biomarker levels compared to blood. On day 42, ZEN concentrations in the ZEN group exhibited approximately twice the variability in urine (31% standard deviation relative to mean) compared to plasma (16%), despite normalization to Crea. Normalization to Crea is widely used to account for varying urine volumes in metabolomics studies [57] and has been shown to minimize matrix effects in mycotoxin analysis in pig urine [58]. In the present study, initial Crea levels in samples from day 42 ranged from 2.0 to 16.5 mM (mean: 8.4 mM), underlining the high variability in urine dilution. Nonetheless, using Crea concentration as the sole normalization factor remains debated. For instance, Winkler et al. [59] reported no benefit from using urinary Crea concentrations to assess mycotoxin exposure in dairy cattle. Moreover, variability in results of LC-MS-based metabolomics analysis of urine samples has been shown to depend on the normalization approach applied, driving recommendations to use multiple normalization techniques [60,61]. To the best of our knowledge, such comparative evaluations are lacking for urinary mycotoxin biomarkers in livestock species, representing an area for future research. Despite these limitations, urine generally reflects systemic exposure to mycotoxins at higher analyte concentrations than blood [9]. In the present study, for example, this enabled quantification of β-ZEL in a sufficient number of replicates for statistical evaluation in urine (unlike in plasma), where it was significantly reduced in the ZEN + ZenA group. Such findings, together with the practical challenges of non-invasive but labor-intensive spot urine sampling, represent important considerations when selecting appropriate biological matrices for evaluating the efficacy of mycotoxin-inactivating feed additives in piglets. As observed in broilers, the reduction in systemic ZEN levels in piglets receiving ZenA supplementation was attributed to enzymatic degradation of the mycotoxin in the GI tract, as reflected by fecal biomarker data (Figure 3C). On day 42, ZEN (33.0% reduction) and α-ZEL concentrations in feces were significantly lower in the ZEN + ZenA group, while HZEN, the primary enzymatic degradation product, was significantly elevated. Overall, supplementation of feed with ZenA consistently reduced ZEN concentrations in plasma, urine and feces, not only on day 42, but across all sampling time points (days 7, 13, 26; Figure S1). When considering total exposure–defined as the molar sum of ZEN, α-ZEL, and β-ZEL–piglets in the ZEN + ZenA group showed significantly lower levels in all matrices at all time points compared to the ZEN group (

Table 4. Sum of zearalenone (ZEN), α-zearalenol and β-zearalenol in plasma (nmol/L), urine (µmol/mol crea) and feces (nmol/kg) in piglets exposed to feed contaminated with 200 µg ZEN/kg (ZEN, n = 12) or feed contaminated 200 µg ZEN/kg and supplemented with 10 U ZenA/kg (ZEN + ZenA, n = 12). Displayed values represent mean ± standard deviation, while respective median and interquartile range values are provided in Table S4. For each matrix and time point, significant differences between treatments are indicated with dissimilar superscripts ^a,b (p < 0.05). Reduction (%) describes the relative decrease in the mean concentrations of the ZEN + ZenA group compared to the ZEN group.

Matrix	Day	ZEN	ZEN + ZenA	Reduction (%)
Plasma	7	1.76 ± 0.75 a	0.98 ± 0.15 b	44.3
	13	1.77 ± 0.71 a	1.10 ± 0.36 b	37.9
	26	1.19 ± 0.21 a	1.02 ± 0.26 b	14.3
	42	1.37 ± 0.20 a	1.04 ± 0.19 b	24.1
Urine	7	63.39 ± 34.26 a	22.27 ± 4.99 b	64.9
	13	79.29 ± 36.57 a	37.58 ± 13.81 b	52.6
	26	45.22 ± 13.74 a	32.86 ± 7.24 b	27.3
	42	62.35 ± 19.20 a	42.90 ± 11.37 b	31.2
Feces	7	1147.26 ± 307.24 a	528.68 ± 65.93 b	53.9
	13	1432.93 ± 638.93 a	637.50 ± 140.26 b	55.5
	26	759.48 ± 284.27 a	443.21 ± 120.00 b	42.6
	42	837.78 ± 175.27 a	571.41 ± 126.66 b	31.8

).

Within the ZEN group, fluctuations in total biomarker levels were observed over time, with the highest concentrations typically occurring on days 7 or 13 and the lowest on day 26. Those variations reached statistical significance in urine (day 13 vs. 26) and feces (day 7 vs. 26, day 13 vs. 26, day 13 vs. 42). Differences in ZEN contamination across feed phases do not explain this pattern, as the phase II feed (fed at sampling day 26) contained slightly more ZEN (217 µg/kg) than phase I (174 µg/kg; sampling days 7 and 13) or phase III feed (154 µg/kg; sampling day 42). Thus, the highest dietary ZEN concentrations coincided with the lowest biomarker levels, suggesting that factors other than feed contamination, such as age-related metabolism or differences in intestinal absorption, may influence biomarker dynamics. These fluctuations warrant further investigation, particularly in the light of similar findings reported by Brezina et al. [62]. In that study, piglets exposed to different levels of dietary ZEN for up to 29 days showed an initial rise in plasma and urinary ZEN concentrations until day 8. Thereafter, no statistically significant time-dependent differences were observed. Yet, numerically lower ZEN concentrations were observed after 29 days of exposure compared to day 8 in piglets receiving 173 or 289 µg ZEN/kg feed. This supports the notion of a potential time-dependent reduction in systemic ZEN concentrations during extended exposure, possibly driven by physiological adaptations that modulate ZEN kinetics. At the same time, it should be noted that individual piglets might exhibit entirely different biomarker trajectories. In the present study, for instance, plasma ZEN concentrations in the same pig were up to 10-fold lower or up to 4.2-fold higher on day 42 compared to day 7. These findings highlight the need to account for both temporal and intra-individual variability when interpreting ZEN biomarker data in pigs.

3.3. Broader Context, Limitations and Outlook

Feed additives, particularly those based on biological transformation, represent a practical post-harvest strategy to mitigate the impact of ZEN in livestock production. While microorganisms may be advantageous for multi-step reactions, enzymes offer distinct benefits, including reproducible activity, homogeneous performance, and ease of handling and dosing. However, the identification, characterization, and development of mycotoxin-degrading enzymes for practical application are lengthy and effort-intensive processes [63]. Several enzyme classes, including laccases, lactone hydrolases, and oxidoreductases, have been reported to degrade ZEN [25,63], but most studies remain limited to in vitro conditions, such as testing in buffer systems or artificial gastric juice [64,65]. Yet, according to the EFSA Panel on Additives and Products or Substances used in Animal Feed, it is essential to demonstrate the efficacy of ZEN-degrading enzymes (or any other substances for the reduction in contamination of feed by ZEN) in vivo, with blood and excreta concentrations of ZEN and its metabolites α-ZEL and β-ZEL considered the most relevant endpoints [66]. While GI degradation of ZEN mediated by ZenA has previously been demonstrated in poultry and swine [30], the present study provides first evidence that this mode of action results in reduced levels of ZEN and its metabolites in broiler blood. For swine, more recent data confirmed ZenA’s efficacy in lowering ZEN levels not only in feces but also in plasma [32]. The present study corroborates these findings and further extends them by including both sexes and applying a substantially lower, economically relevant ZenA dosage. This highlights the potential of ZenA to effectively reduce the systemic ZEN burden in swine under practical feeding conditions. Beyond ZenA, to the best of our knowledge, no studies on other ZEN-degrading enzymes are available examining their effects on ZEN levels in blood or excreta. For other feed additives intended to reduce contamination of feed by ZEN, e.g., by containing binder materials, yeast cell wall components or microorganisms, a multitude of studies are available, which focus on possible benefits for health or performance parameters. This, however, does not allow direct conclusions about the extent and mode of action of ZEN inactivation. Among the few exceptions, supplementation with Bacillus spp. reduced blood ZEN levels in gilts [47], whereas other additives failed to show significant reductions in ZEN and its metabolites in blood of pigs or hens [45,67,68]. Notably, these studies were performed under experimental conditions not representative for farm exposure, for example, employing intragastric bolus application or administering high ZEN concentrations, which hampers direct comparison with the present work.

Limitations of our study include the absence of a product control group (animals receiving ZenA-supplemented feed without ZEN contamination) and a negative control group (animals receiving uncontaminated feed) in the piglet trial, as the primary focus was on evaluating the feed additive and biomarker responses across different matrices. Consequently, the study does not permit conclusions regarding the effects of the additive in the absence of mycotoxin challenge, nor does it provide data on toxin effects alone in piglets. Moreover, as our aim was to assess the efficacy of ZenA at dietary ZEN concentrations below the NOEAL for piglets [43,44], no data on reproductive performance were collected. However, previous studies conducted with higher dietary ZEN levels have demonstrated that ZenA significantly reduces estrogenic symptoms such as vulva enlargement or increased ovary weights [32,69]. For future studies, combination of biomarker assessment with evaluation of subclinical alterations in reproductive function, such as changes in adipokine profiles, would be valuable. Finally, while both trials were performed under conditions closely resembling practical livestock production, they were still conducted in a controlled research environment. The piglet trial took place at a trial facility that closely mirrored a commercial farm in terms of housing, feeding, and watering systems, whereas the broiler study was performed under more experimental conditions. Therefore, although the results are highly relevant for livestock production, further validation under fully commercial conditions would be desirable to confirm the efficacy of ZenA in routine farming practice.

Demonstrating the efficacy of mycotoxin-detoxifying feed additives at dietary toxin concentrations close to or below current regulatory guidance levels presents considerable analytical challenges. As shown in the present study, accurate quantification of ZEN and its metabolites at such low exposure levels requires highly sensitive instrumentation and extensive method validation. Achieving even higher analytical sensitivity is technically demanding with current state-of-the-art equipment. In future, potential lowering of guidance or maximum levels for mycotoxin in feed might be envisaged from the perspectives of feed/food safety and animal health. Yet, it should be recognized that such reductions may introduce additional constraints for the experimental proof of efficacy and, consequently, the authorization of otherwise effective detoxifying additives.

4. Conclusions

The present study demonstrates that the enzyme ZenA effectively degraded ZEN in the GI tract of both broilers and piglets, leading to reduced systemic ZEN exposure. Therefore, supplementation of feed with ZenA offers a promising approach to mitigate ZEN-related risks in these livestock species and production phases. Future research should address the effects of ZenA on systemic ZEN levels in additional production systems (e.g., layers), species (e.g., ruminants) and under varying experimental conditions (e.g., study duration, diet composition, feed processing). This will be essential to confirm and generalise the efficacy of the enzyme across diverse livestock settings.

Biomarker analysis across various biological matrices provided a comprehensive understanding of the enzyme’s effects and additional insights into the toxicokinetics of ZEN in broilers and piglets. When evaluating mycotoxin-inactivating feed additives, the choice of biological matrix should align with the study objective, as each matrix has specific advantages and limitations—such as relation to inner toxin exposure, invasiveness, analyte concentration, and inter-individual variability. Finally, the results underline that performance parameters alone are insufficient to assess the efficacy of such feed additives. Instead, biomarker-based approaches offer a more reliable and specific alternative, but require sensitive analytical techniques, particularly for detecting ZEN biomarkers in plasma at moderate dietary exposure levels.