*2.4. Measured Characteristics*

2.4.1. Analysis of Resveratrol, Piceid, Astringin and Piceatannol in Knotweed Biomass

Finely ground (10 mm sieve) samples (500 mg) of dry aboveground biomass were extracted with 10 mL 60% ethanol, as it was the most efficient extractant for both resveratrol and its glucosides. Prior to analysis, the samples were filtered with syringes (NY, 0.2 μm). The extracts were analysed by a validated HPLC-UV method. Instrument: Shimadzu NEXERA with PDA detection 306 nm; column: Phenomenex synergi Hydro-RP 80A, 250 mm × 4.6 mm, 4μm (30 ◦C); flow rate: 1.5 mL.min−1; mobile phases: A–10 mM ammonium acetate at pH 4.15 using acetic acid, B-acetonitrile, concentration gradient from 7 to 90%. Standards: Sigma-Aldrich, Piceid: 98%, Resveratrol: 99%, Emodin: 98%, Astringin: 99%, Piceatannol: 99%. Mixed calibration was used for quantification. Validation report called *Determination of selected biologically active substances present in knotweed by HPLC/UV method for AKP*, has been deposited in the Library of VUOS in Pardubice Rybitví as SOP 568, VU 4661, 2007.

#### 2.4.2. Blood Testing: Haematology and Biochemistry

At the beginning and end of each experiment, before the horses were released into pastures, they were weighed, and blood samples were drawn on the stud farm in the early morning for biochemical and haematological tests. Blood was drawn from the jugular vein and collected in a vacuum tube. One millilitre of blood was collected in a tube with EDTAK3 and was used for haematological tests, and 10 mL was collected in a tube with pro-coagulation gel and used for biochemical tests. The samples were kept cool during their rapid transport to the laboratory. Blood samples from Experiment 1 were analysed at the veterinary clinic in Brno, where they were analysed with a BC-2800 Vet haematological analyser (Mindray, Shenzhen, China) in horse mode. White blood cells were counted under a microscope in blood smears stained according to Pappenheim (May-Grünwald, Giemsa-Romanowski) in 200-cell subsamples. Blood samples from Experiment 2 were analysed in the Medila laboratory for biochemical parameters, and the haematological samples were analysed in the veterinary laboratory of the veterinary clinic Štrossovka s.r.o., Pardubice, using an automated haematological counter. The blood smears were simultaneously fixed and stained with Hemacolor®. The cellular elements (neutrophils, eosinophils, basophils, monocytes, lymphocytes and activated lymphocytes) were counted with the LeukoCounter system, and the individual populations of leukocytes were expressed as percentages of 200-cell subsamples. The following parameters were measured: total protein, albumin, urea, creatinine, cholesterol, glucose and triglyceride levels; AST, ALT and ALP activity; haemoglobin content (as the mean corpuscular haemoglobin (MCH), which is calculated by dividing the total weight of the haemoglobin by the total number of red blood cells in a particular volume of the blood sample and gives an average concentration of haemoglobin in each red blood cell throughout the sample); plateletcrit (PCT); and leucocyte, erythrocyte, thrombocyte, lymphocyte, neutrophil and monocyte counts. The albumin results could

not be compared between the two experiments because different methods were used to determine the levels in Experiment 1 (photometry) and Experiment 2 (electrophoresis). As all biochemical laboratories in the Czech Republic use the same methods, maximum divergence should not exceed 20% and low bias due to lab change is to be expected.

#### *2.5. Statistical Data Evaluation*

The data were statistically evaluated with ANOVA (STATISTICA v.12, Dell Inc., Tulsa, OK, USA). One-way ANOVA (factor: experimental treatment—control or knotweed-fed) or two-way ANOVA (factors: experimental treatment, horse age or horse type) was applied. The mean differences between the control and knotweed-fed horses were measured. If homoscedasticity was not met, the Kruskal-Wallis test was used. Multivariate redundancy analysis (RDA) was performed using Canoco 5.0 (Biometris, Plant Research International, Wageningen, The Netherlands) [35]. Biochemical and haematological characteristics and their differences between the beginning and the end of the experiments were used as response data. Experimental treatment (control or knotweed-fed) was used as the explanatory variable.
