The Institutional Animal Care and Use Committee approved all animal handling procedures (case number 038/2020), and the experiment was conducted according to the experimental protocol for the use of live birds from the Brazilian College of Animal Experimentation.

The experiment was conducted in Viçosa, MG, Brazil (20°45'57.19" S, 42°51'35.42" W, and 682 m altitude). The male broiler chickens (Cobb 500) used in the experiment were obtained from a commercial hatchery (Rivelli Alimentos SA, Matheus Leme, MG, Brazil). Chicks were vaccinated against bursal disease and Marek's disease (Serotype 3, Live Marek's Disease Vector, Merial Inc., Athens, GA). From one day old until the beginning of the experiment, the birds were reared in a masonry house divided into protected circular pens containing tube feeders, manual drinkers, and a litter of wood shavings. They had free access to water and were fed ad libitum with a corn/soybean meal-based mash diet formulated to meet their nutritional requirements according to Rostagno et al. (2017).

At 22 days old, 280 broiler chickens (983±38 g) were distributed based on body weight in a completely randomized design with four treatments, ten repetitions, and seven birds per experimental unit. They were housed in 40 wire floor cages (1,008 cm²/bird) in a four-level battery equipped with a trough feeder and a nipple drinker.

Birds were subjected to daily cyclic HS in controlled chambers. A high temperature of 32±1 °C was maintained for 10 h/day (08:00–18:00 h), while the temperature was set at 22±1 °C for the remaining time. The relative humidity of the air inside the chambers was maintained at 65±5%.

Treatments were a basal diet or basal diet with the addition of 5, 10, or 15 g of purified lignin/kg of diet. The purified lignin used in this research was extracted from Eucalyptus urograndis through the kraft process, used in pulp and paper production. The corn/soybean meal basal diet was formulated to meet the nutritional recommendations given by Rostagno et al. (2017) (Table 1). Purified lignin in the basal diet was used instead of the inert. Diets were prepared in mash form. Free access to water and feed was provided throughout the experimental period (22 to 42 days old). The light program adopted for the entire experimental period was 18 h of light (4:00 to 22:00 h) and 6 h of dark.

Birds and feed leftovers were weighed at 42 days of age to calculate feed intake (FI), weight gain (WG), and feed conversion ratio (FCR). Mortalities were recorded throughout the experimental period, and the necessary corrections of performance data were carried out.

At 42 days old, three birds with weights closest to the average weight for their respective experimental unit were selected. One bird was used for blood collection. After blood collection, the bird was euthanized by cervical displacement and slaughtered. Liver samples were collected, stored individually in cryogenic tubes, and placed in liquid nitrogen. These samples were transferred to freezer storage at −80 °C until the RNA extraction process.

The two remaining birds, after 8 h of fasting, were euthanized by cervical displacement and slaughtered to measure the yield of carcass, breast, and thigh with drumstick, as well as the relative weight of the lymphoid organs (bursa and spleen), liver, intestine, and abdominal fat. Carcass yield (CY) was calculated in relation to living weight before slaughter [%CY = (carcass weight × 100)/live weight] and breast and thigh yield with drumstick, and as a function of carcass weight [%Part = (part weight × 100)/carcass weight]. The relative weights of the bursa, spleen, liver, intestine, and abdominal fat were calculated in relation to the live weight of the birds before slaughter.

The collected blood was used to analyze serum levels of glucose, uric acid, triglycerides (Cobas c 311; Roche Diagnostics GmbH, Basel, Switzerland), and the hormones triiodothyronine (T3) and tetraiodothyronine (T4; Atellica IM, Siemens Healthcare Diagnostics Inc, New York, USA), following the manufacturer's instructions. To measure MDA, 2.5 mL of 20% trichloroacetic acid and 1.0 mL of 0.67% thiobarbituric acid were added to 0.5 mL of serum; the mixture was then heated for 30 min in boiling water. The resulting chromogen was extracted with 4.0 mL of n-butyl alcohol. The absorbance of the organic phase was determined at a wavelength of 530 nm.

Total RNA was extracted from 50 mg of powdered liver samples, using TRIzol^® (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. The resulting precipitate was rehydrated with 25 μL of UltraPure DNase/RNase-Free water. The RNA concentration was estimated using a NanoDrop^TM Lite Spectrophotometer (ThermoFisher Scientific, Beverly, MA, USA). RNA integrity was determined in 1.0% agarose gel. The first cDNA strand was synthesized using High-Capacity cDNA Reverse Transcription Kits (Applied Biosystems, Thermo Fisher Scientific, Beverly, MA, USA). The primer sets used are shown in Table 2; β-actin (β-ACT) was used as the reference gene for data normalization. The following target genes were assessed: heat shock protein 70 (HSP70), nuclear factor-κB (NF-κB), glutathione peroxidase (GPX), and Cu,Zn-superoxide dismutase (SOD1).

The RT-qPCR analyses were performed in duplicate with an Applied Biosystems^TM QuantStudio Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific, Beverly, MA, USA), using the Relative Quantification method and applying the SYBR^® Green system (Applied Biosystems, Foster City, CA, USA) and GoTaq^® qPCR Master Mix kit (Promega Corporation, Madison, WI, USA). PCR reactions were subjected to the cycles protocol according to the program: 95 °C for 2 min, 40 cycles of 95 °C for 15 s, and 60 °C for 1 min. Threshold cycle (Ct) values obtained were later normalized (ΔCt) based on Ct values of the endogenous control gene β-ACT. The calculation of gene expression levels was performed according to the 2^-ΔCt method, as described by Livak and Schmittgen (2001).

Cage averages were considered as an experimental unit for statistical analysis of growth performance parameters. For analyses of yield of carcass, breast, and thigh with drumstick, as well as the relative weights of the lymphoid organs, liver, intestine, and abdominal fat, the average of two birds per replicate was considered as the experimental unit. For serum and gene expression analyses, one bird per replicate was considered as the experimental unit. Data were analyzed via one-way ANOVA, according to the following general model:

in which Y_ij is the measured dependent variable, μ is the overall mean, α_i is the effect of treatments, and ε_ij is the random error.

Analyses were carried out using the GLM procedure of SAS (Statistical Analysis System, version 9.4). Comparison between treatment averages was performed using Tukey's test. A significance level of 0.05 was applied.

There was no treatment effect (P>0.05) on performance and carcass yield (Tables 3 and 4).

There was no significant treatment effect (P>0.05) on the relative weights of the bursa, spleen, liver, intestine, and abdominal fat (Table 5).

Purified lignin supplementation did not influence (P>0.05) serum levels of glucose, triglycerides, uric acid, MDA, or the T3 and T4 hormones (Table 6).

The abundance of mRNA of NF-κB, HSP70, GPX, and SOD1 in the liver was not influenced by treatments (P>0.05; Figure 1).