Phenotypic profile of in utero heat stressed grow-finish pigs divergently selected for thermotolerance or thermosensitivity in the F2 generation

Abstract

Introduction:

This study evaluated the interaction between genomic selection for heat stress tolerance (TOL) or sensitivity (SEN) and exposure to in utero heat stress (IUHS) or in utero thermoneutral (IUTN) conditions in grow-finish pigs under postnatal cyclical heat stress (HS).

Materials and methods:

Twenty-four pigs (n = 6 per treatment; 75.8 ± 10.3 kg) were implanted with intra-abdominal core body temperature (T_C) sensors and exposed to alternating thermoneutral (TN: 22-23 °C) and cyclical HS (26-36 °C) periods. The T_C was continuously recorded at 5 min intervals. Skin temperature and respiration rate were recorded four times daily. Blood gases, average daily feed intake (ADFI), and average daily gain (ADG) were assessed within each period. Behavior was assessed on d 1 of TN1 and the first 3 d of each HS period. Pigs were euthanized and hypothalamus, pituitary, and adrenal glands were assessed forweight, mRNA abundance, and third ventricle cell counts.

Results:

Overall, T_C was greater in IUHS versus IUTN pigs, and greater in IUHS+TOL and IUTN+SEN versus IUTN+TOL pigs. Respiration rate was reduced in IUHS versus IUTN pigs. Also, ADFI was greater in TOL versus SEN pigs. During TN, ADG was greater in IUHS+TOL versus IUHS+SEN and IUTN+TOL pigs. In HS1, ADG was reduced in IUHS+TOL and IUHS+SEN versus IUTN+TOL pigs. The IUHS pigs had greater CO₂ partial pressure, reduced O₂ partial pressure, reduced brain weight, and tended tohave greater sternal laying versus IUTN pigs. The TOL pigs had greater pituitaryweight and tended to sit less than SEN pigs. Third ventricle cell counts weregreater in IUHS+TOL and IUTN+SEN versus IUHS+SEN and IUTN+TOL pigs. No other differences were detected.

Discussion:

Reduced brain and pituitary mass may indicate altered neuroendocrine development. Altered blood gases may suggest respiratory changes, while hypothalamic third ventricle cellularity may have implications for HPA-axis regulation. In summary, IUHS altered physiological, morphological, and growth responses of grow-finish pigs, while genomic selection for TOL supported greater ADFI, ADG, and altered hypothalamic and pituitary development compared with SEN pigs. Genomic selection for thermotolerance, indexed by core body temperature variability relative to ambient temperature, shows promise for mitigating IUHS effects in commercial swine under future climaticconditions.

1 Introduction

In utero heat stress (IUHS) can affect the future physiological stress response and thermoregulatory capacity of offspring (Johnson et al., 2013; Johnson et al., 2015a; Byrd et al., 2019; Maskal et al., 2021). The phenotypes associated with IUHS are likely mediated through restricted uterine blood flow (Clapp, 1980; Clapp, 2006), shortened gestation length (Kattesh et al., 1980; Monteiro et al., 2014), impaired nutrient transport to the fetus (Monteiro et al., 2014), and intrauterine growth retardation (Alvarenga et al., 2013), likely caused by insufficient placental development (Alvarenga et al., 2013; Boddicker et al., 2014), and epigenetic imprinting (Skibiel et al., 2018). While beneficial outcomes of embryonic thermal conditioning have been documented in poultry, such as enhanced stress resilience, reduced mortality, and improved thermotolerance in adulthood (Yahav et al., 2004; Piestun et al., 2008), swine appear to exhibit the opposite pattern. The divergence between pigs and egg-laying species likely reflects differences in how offspring experience gestational heat stress (HS). In oviparous species, only a thin eggshell separates the embryo from its thermal environment, allowing HS to directly recalibrate thermoregulatory set points during development. In mammals, however, placental thermoregulation buffers fetal body temperature from maternal thermal fluctuations (Laburn et al., 2002), meaning postnatal phenotypes in IUHS pigs arise not from direct thermal conditioning but from the indirect consequences of the maternal physiological HS response.

Specifically, IUHS in pigs has been linked to an elevated core body temperature set point in offspring, regardless of postnatal thermal exposure (Johnson et al., 2015a). A greater body temperature set point narrows the temperature gradient between the animal and its environment, limiting the effectiveness of heat dissipation and increasing vulnerability to thermal stress. As such, we have observed greater HS sensitivity in IUHS pigs during postnatal life (Johnson et al., 2013). This greater HS sensitivity has the potential to reduce both welfare and productivity as maintaining body temperature at a greater level is energetically expensive (Chapel et al., 2017). Furthermore, it may also predispose IUHS female offspring to increased HS sensitivity during their own pregnancies, since elevated maternal body temperature is the primary risk factor for IUHS effects (Graham et al., 1998; Ravanelli et al., 2019), thereby perpetuating a cycle of IUHS susceptibility across generations (Laporta et al., 2024). Additionally, when exposed to stressful conditions (e.g., weaning, transport, social mixing), IUHS pigs exhibit elevated cortisol production (Kattesh et al., 1980; Byrd et al., 2019), and corticotropin releasing hormone challenges result in physiological responses indicative of altered hypothalamic-pituitary-adrenal (HPA) axis function (Maskal et al., 2021), which may be programmed by greater in utero cortisol exposure (Byrd et al., 2025a). Elevated maternal cortisol during gestation can cross to the placenta, particularly when placental 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) activity is overwhelmed (Harris and Seckl, 2011; Byrd et al., 2025c) which can epigenetically silence glucocorticoid receptor gene (NR3C1) promoter regions through DNA methylation and thus reducing feedback sensitivity within the HPA axis (Hompes et al., 2013). Therefore, identifying strategies to reduce the negative effects of IUHS on developing swine offspring is paramount.

Several strategies have been explored to mitigate the effects of gestational HS in sows including dietary supplements (Cottrell et al., 2015) and the use of decision support tools to improve environmental management (McConn et al., 2022). While these approaches can help reduce the impact of HS, they primarily target short-term management. In contrast, genomic selection for HS tolerance (TOL) represents a long-term solution that promotes resilience across generations. Our research team has developed and applied a genomic selection model for climatic resilience in sows (Wen et al., 2023; Freitas et al., 2024; Byrd et al., 2025a; Byrd et al., 2025b), which is based on measuring the variability in internal body temperature relative to ambient environmental conditions. Through this approach, two genetically divergent pig populations have been established: one group selected for TOL and another for HS sensitivity (SEN). This genomic selection strategy focuses on the rate at which a pig’s internal body temperature responds to changes in environmental temperature, potentially offering a sustainable method for improving heat resilience without compromising productivity (Wen et al., 2023; Freitas et al., 2024; Byrd et al., 2025a; Byrd et al., 2025b). A lower fluctuation in internal body temperature in response to environmental changes has been associated with enhanced thermoregulatory efficiency (Gourdine et al., 2021). As a result, we have observed that lactating TOL sows (experimental F0 generation) can sustain higher metabolic activity, and by extension, greater productivity, without experiencing harmful rises in core body temperature (Byrd et al., 2025a). Additionally, gestating TOL sows (experimental F1 generation) tend to have a decrease in vaginal temperature (T_V) and larger litter sizes (Byrd et al., 2025b) when exposed to HS conditions. As a result, it is possible that selection for TOL may prove to be an effective strategy to reduce the negative impact of IUHS on developing swine offspring since the primary risk factor for developing IUHS phenotypes is a rise in maternal body temperature (Graham et al., 1998; Ravanelli et al., 2019). It remains to be known whether genomic selection for thermotolerance based on our approach (Wen et al., 2023; Freitas et al., 2024; Byrd et al., 2025a; Byrd et al., 2025b) can offset IUHS associated phenotypes in grow-finish pigs.

The study objective was to evaluate the interactions between genomic selection for TOL and SEN and exposure to IUHS or in utero thermoneutral (IUTN) conditions on the postnatal thermoregulation, physiological and behavioral stress response, and growth performance of grow-finish pigs exposed to cyclical postnatal HS. We hypothesized that (1) IUHS would result in elevated body temperature responses, a greater physiological and behavioral stress response, and reduced growth performance relative to IUTN pigs when exposed to postnatal HS; and (2) genomic selection for TOL would attenuate these effects, such that IUHS+TOL pigs would exhibit improved thermoregulation, physiological and behavioral stress response, and growth performance relative to IUHS+SEN pigs.

2 Materials and methods

All genomic selection protocols (Wen et al., 2023; Freitas et al., 2024; Byrd et al., 2025a; Byrd et al., 2025b) and in utero environment procedures (Byrd et al., 2025a; Byrd et al., 2025b) and resultant data (Wen et al., 2023; Freitas et al., 2024; Byrd et al., 2025a; Byrd et al., 2025b) have been previously reported by our group. A schematic representation of the generational intervals and the experimental design is presented in Figure 1.

Figure 1

2.1 Gestation phase

The Purdue University Animal Care and Use Committee approved all animal procedures (protocol # 2101002105A001), and animal care and use standards were based on the Guide for the Care and Use of Agricultural Animals in Research and Teaching (Federation of Animal Science Societies, 2020). In brief, F1 generation replacement gilts (N = 35; Large White × Landrace) were produced from F0 sows (Large White × Landrace) and sires (Landrace) selected for either SEN or TOL, as described previously (Freitas et al., 2023; Johnson et al., 2023; Wen et al., 2023; Byrd et al., 2025b). At 238.17 ± 25.66 d of age, all F1 generation replacement gilts were artificially inseminated with semen from individual Landrace boars genomically selected for SEN (n = 4 boars) or TOL (n = 3 boars), based on established methods (Freitas et al., 2023; Wen et al., 2023). On d 3.29 ± 0.92 post-breeding (study d -4), all gilts were transported (4 km) from the Purdue University Swine Farm to the USDA-ARS Livestock Behavior Research Units’ Food Animal Behavior and Well-Being Laboratory in West Lafayette, IN, USA. Gilts were housed in one of two environmental chambers set to either thermoneutral (TN; n = 8 SEN and 9 TOL F1 replacement gilts) or cycling HS (n = 9 SEN and 9 TOL F1 replacement gilts per room) conditions. Within each room, pregnant gilts were grouped by body weight (BW), litter of origin, genomic ranking, and servicing sire. Six data loggers (HOBO logger; model UX100-003; accuracy = 3.5%; resolution = 0.024 °C at 25.0 °C; Onset; Bourne, ME, USA) were mounted at the animals’ head height to record dry bulb temperature (T_DB) and relative humidity (RH) every 5 min daily throughout the study.

In the TN chambers, the T_DB was held constant at 21.46 ± 2.92 °C with 45.05 ± 10.24% RH from d 3.29 ± 0.92 to d 65.29 ± 0.92 of gestation. In contrast, HS gilts were kept at 18.46 ± 0.47 °C and 21.84 ± 7.21% RH from d 3.29 ± 0.92 until d 6.29 ± 0.92 post-insemination. Beginning on d 6.29 ± 0.92 through d 10.29 ± 0.92 post-insemination, HS gilts experienced a cycle of nighttime conditions of 25.34 ± 2.39 °C and 27.2 ± 8.86% RH and daytime conditions of 28.04 ± 2.77 °C and 28.46 ± 7.08% RH. From d 11.29 ± 0.92 to 65.29 ± 0.92 post-insemination, nighttime (2000 to 0800 h) and daytime (0800 to 2000 h) conditions cycled between 28.89 ± 0.77 °C and 29.75 ± 9.12% RH and 34.96 ± 1.13 °C and 31.34 ± 8.57% RH, respectively. From d 66.29 ± 0.92 post-insemination through farrowing, all gilts were placed under TN conditions at 21.40 ± 5.63 °C and 59.78 ± 8.56% RH.

Throughout gestation, gilts had unrestricted access to water and were limit-fed at a rate of 1.82 kg/d. Their diet, primarily composed of corn, soybean meal, and dried distillers grains was formulated according to or exceeding the nutrient requirements for gestating gilts (National Research Council, 2012). No pregnant gilts had any feed refusals at any time during the gestation period. During the environmental treatment phase, gestating HS gilts exhibited higher core body temperatures (T_C) than gestating TN gilts (Byrd et al., 2025b). Thus, by the accepted definition of IUHS reported by Byrd et al. (2019) and Byrd et al. (2025a), their fetuses were exposed to IUHS.

2.2 F2 generation animal selection and surgery

Piglets were weaned at d 20.61 ± 1.52 post-farrowing, and between weaning and 17 weeks of age, pigs were group-housed and fed a diet composed primarily of corn, soybean meal, and dried distillers grains formulated to meet or exceed the nutrient requirements for growing pigs (National Research Council, 2012). At 17–weeks of age, 12 barrows and 12 gilts (75.8 ± 10.3 kg BW) were selected based on in utero environment, genetic line, body weight, and equally distributed by sex and litter of origin: (n = 6 IUHS+TOL; n = 6 IUHS+SEN; n = 6 IUTN+TOL; n = 6 IUTN+SEN) for surgical implantation of temperature recording data loggers.

Automatic temperature recording data loggers (DST milli-T; manufacturer stated accuracy: ± 0.10 °C; resolution = 0.032 °C; Star Oddi, Garoabaer, Iceland) were surgically implanted into all animals following the procedure previously described by our group (Cecil et al., 2025). Briefly, pigs were restrained using a snare and then anesthetized by an intramuscular injection of a cocktail including telazol (concentration in cocktail: 100 mg/mL), ketamine (concentration in cocktail: 50 mg/mL), and xylazine (concentration in cocktail: 50 mg/mL) administered at 1 mL per 23 kg BW. Following anesthesia, pigs were injected with lidocaine HCl (Covetrus, NDC 11695-4149-1; 20 mg/mL; 100 mL vial; 4 mg/kg BW) subcutaneously at the incision site, then a 6-cm incision was made on the abdomen, 5 cm lateral to the linea alba, and sterile temperature data loggers were sutured intrabdominally to the abdominal muscle wall. After surgery, all pigs were administered an intramuscular antibiotic (Ceftiofur Hydrochloride, Zoetis, Florham Park, NJ, USA) every 3 d per manufacturer’s instructions and were allowed a 7-d recovery period before the start of the experiment.

2.3 Postnatal treatments

Following the recovery period, all pigs were transported (4 km) to the USDA-ARS Livestock Behavior Research Units’ Food Animal Behavior and Well-Being Laboratory (West Lafayette, IN, USA), 2 d prior to the start of the trial. All pigs were allotted into one of two environmental chambers based on in utero environment, genetic line, body weight, sex, and litter of origin. Six data loggers were placed at pig head height in each environmental chamber to record dry bulb temperature and RH in 5 min intervals daily. The ambient cyclic temperature pattern in the environmental chambers is presented in Figure 2. All pigs were initially maintained under similar TN environmental conditions (22.59 ± 1.82 °C; 35.94 ± 9.23% RH) and data collection began 2 d after pigs were allotted to their pens. All pigs were maintained under TN conditions from d 1 to 3 (TN1; 22.38 ± 2.18 °C and 40.51 ± 10.83% RH. From d 4 to 9, all pigs were exposed to cyclical HS conditions (HS1; 28.78 ± 3.66 °C and 33.18 ± 16.74% RH nighttime and 35.43 ± 4.84 °C and 31.14 ± 15.97% RH daytime), and then from d 10 to 11, pigs were exposed to TN conditions (TN2; 23.20 ± 1.97 °C and 32.02 ± 4.93% RH) followed by a second period of cyclical HS from d 12 to 17 (HS2; 29.13 ± 1.90 °C and 47.83 ± 9.38% RH nighttime and 35.47 ± 4.42 °C and 37.50 ± 13.50% RH daytime). All pigs were fed a standard commercial diet ad libitum, consisting of primarily corn, soybean meal, and dried distillers grains and was formulated to meet or exceed nutritional requirements (National Research Council, 2012).

Figure 2

2.4 Live phase measurements

Body weight and feed intake of all pigs were recorded at the beginning and end of each period. Respiration rate (RR) and ear (T_ear), shoulder (T_shoulder), rump (T_rump), and tail (T_tail) skin temperatures were obtained four times daily (0800, 1200, 1600, and 2000 h) throughout the study. Respiration rate was determined by counting flank movement for 15 s and multiplying by 4 to obtain breaths per minute (bpm) as previously described (Johnson et al., 2023). Skin temperature was determined on clean and dry shaved skin sites using an infrared temperature thermometer (Raytek model ST61; accuracy = ± 1%; emissivity = 0.98; resolution = 0.10 °C; Raytek Corporation, Fluke Process Instruments, Everett, WA, USA) as previously described by Byrd et al. (2025b). Core body temperature was determined every 5 min using an automatic temperature logger surgically implanted intraabdominally. All pigs were video recorded on study d 2, 4, 5, 6, 12, 13, and 14 using mounted cameras (PAR-C5BIRA2812, InVid Tech, New York, USA; 5-megapixel resolution 2592 x 1994 dpi; 20fps frame rate) attached to a digital video recorder system (PD2B-16NH, InVid Tech, New York City, NY, USA). Behavior video was recorded during the pre-maximum heating (0900 to 1000 h), maximum heating (1400 to 1500 h), and post-maximum heating (2100 to 2200 h) periods of cyclical HS (Figure 1). Video files were analyzed using continuous sampling by two trained individuals who maintained an agreement of 90% or greater. Inter-observer variability was determined by analyzing 6 h of video and comparing their observations to ensure an agreement of 90% or greater was maintained. Behaviors analyzed included pig posture (standing, sitting, lying sternal, lying lateral, or other) and consumption behavior (feeding and drinker interactions) according to an ethogram (Table 1) based on a previous report (Byrd et al., 2025b) with slight modifications.

2.5 Blood sampling and analyses

Blood samples (10 mL) were obtained from all pigs via jugular venipuncture (Vacutainers, BD, Franklin Lakes, NJ, USA; K3EDTA; lithium heparin) between 1500 and 1600 h on d 1, 3, 4, 8, 12, and 17 representing the beginning and end of each period. Blood O₂ partial pressure (PO₂) and CO₂ partial pressure (PCO₂), O₂ saturation, HCO₃ concentration, and pH were immediately determined from whole blood collected in lithium heparin tubes using a Vet iStat C68+ cartridge (Abaxis Inc., Union City, CA, USA).

2.6 Tissue collection

On d 18, all pigs were humanely euthanized using an intravenous injection of pentobarbital sodium and phenytoin sodium following veterinarian recommended dosage (390 mg pentobarbital sodium and 50 mg phenytoin sodium/mL administered at 1 mL/4.55 kg body weight; Euthasol^®; Virbac AH, Inc., Fort Worth, TX, USA) immediately followed by exsanguination. The head, liver, pancreas, adrenal gland, brain, hypothalamus, and anterior pituitary gland were collected from all pigs and weighed as previously described by Maskal et al. (2021). The hypothalamus was bisected, and the ventral half was stored in 10% neutral buffered formalin for histological analysis, whereas the dorsal half of the hypothalamus along with the pituitary gland, and adrenal cortex were immediately snap frozen in liquid nitrogen and stored at -80 °C for subsequent analyses.

2.7 Laboratory analyses

Thirty milligrams of hypothalamus, anterior pituitary, and adrenal tissue were homogenized using a TissueRuptor (Qiagen, Germantown, MD, USA) and a mix of 10 uL β-mercaptoethanol (2-Mercaptoethanol; 55 mM in D-PBS; Cat. No. 21985-023; Gibco, Invitrogen Corporation, Carlsbad, CA, USA) and 1 mL Buffer RLT from RNeasy Mini Kit (Qiagen, Germantown, MD, USA). After 3 min of incubation, the sample mix was centrifuged for 3 min at 14,000 × g at 4 °C, and then the supernatant was pipetted and added to an equal volume of 70% ethanol. The total solution was transferred to spin columns, and total RNA was purified using RNeasy Mini Kit. Eluted total RNA was analyzed for concentration and purity using spectrophotometry at 260/280 nm with a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA).

Reverse transcription was carried out with the Reverse Transcription Reagent Pack (Catalog #: N8080234, Applied Biosystems, Foster City, CA, USA). The reverse transcription reaction included 2 µL of RNase inhibitor, 2.5 µL of MultiScribe reverse transcriptase, 5 µL of random hexamers, 10 µL of TaqMan reverse transcription buffer, 20 µL of deoxynucleotides, and 22 µL of 25 mM magnesium chloride. Each sample’s final mixture totaled 61.5 µL, to which the RNA sample and RNase-free water were added for a total volume of 100 µL. A Techne TC-3000G PCR Thermal Cycler (Bibby Scientific Limited, Stone, UK) was used to convert the RNA into complementary DNA (cDNA).

Hypothalamus mRNA expression levels of glucocorticoid receptor (GR) (Assay ID #: Ss03378868_u1) and corticotropin-releasing hormone (CRH) (Assay ID #: Ss03818805_s1), adrenal gland mRNA expression levels of melanocortin 2 receptor (MC2R) (Assay ID #: Ss03373249_u1), and anterior pituitary gland mRNA expression of GR (Assay ID #: Ss03378868_u1), corticotropin-releasing hormone receptor 1 (CRHR1) (Assay ID #: Ss03373289_g1), corticotropin-releasing hormone receptor 2 (CRHR2) (Assay ID #: Ss03821383_s1), and proopiomelanocortin (POMC) (Assay ID #: Ss03381950_u1) were then measured by Real-Time Quantitative Polymerase Chain Reaction (RT-qPCR) with their respective primers and probes. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Assay ID #: Ss03375629_u1, Applied Biosystems, Foster City, CA, USA) served as the reference gene. Each PCR reaction contained 1.625 µL of TaqMan probe, 2.25 µL of each gene-specific TaqMan primer, 12.5 µL of PCR Master Mix (Catalog #: 4304437, Applied Biosystems, Foster City, CA, USA), 3.875 µL of RNase-free water, and 2.5 µL of cDNA. The thermocycling program involved an initial holding step of 50 °C for 2 min followed by 95 °C for 10 min, then 40 cycles of 95 °C for 15 s and 60 °C for 1 min. Change in cycle threshold (ΔCT) values were calculated as the difference in CT values between gene of interest and that of GAPDH on the same tissue, and ΔΔCT values were calculated as the difference between ΔCT values of each sample and the average ΔCT value of all samples. mRNA abundance was calculated as 2^-ΔΔCT of individual pigs.

2.8 Hypothalamus histology

Histological staining and processing of hypothalamic tissue from the ventral bisection was performed at the Purdue University Histology Core Facility (West Lafayette, IN, USA) following established protocols (Bitsie et al., 2021; Musa et al., 2026). Tissue samples were processed using a Sakura Tissue-Tek VIP6 automated processor (Tissue-Tek^® VIP^® 6; Sakura Finetek USA Inc., Torrance, CA, USA), including sequential dehydration in graded ethanol, clearing in xylene, and paraffin infiltration with Leica Paraplast Plus (Surgipath^® Paraplast^® Plus; Leica Biosystems; Richmond, IL, USA). Processed tissues were subsequently embedded in paraffin, sectioned at 4 µm thickness using a Thermo HM355S microtome (Microme^® HM355S; Thermo Fisher Scientific; Waltham, MA, USA), mounted on charged slides, and dried at 60 °C for 30–60 min.

Slides were deparaffinized in xylene and rehydrated through a graded ethanol series to water using a Leica Autostainer XL (Autostainer XL ST5010; Leica Biosystems; Deer Park, IL, USA). Antigen retrieval was performed in a BioCare decloaking chamber (Decloaking Chamber NxGen DC2012; Biocare Medical; Concord, CA, USA) at 95 °C for 20 min in TRIS/EDTA buffer (pH 9.0), followed by cooling at room temperature for 20 min and transfer to TBST (TRIS-buffered saline with Tween 20). Subsequent staining procedures were conducted at room temperature using a BioCare Intellipath automated stainer (intelliPATH FLX IPS0001US; Biocare Medical; Concord, CA, USA).

Endogenous peroxidase activity was quenched with 3% hydrogen peroxide for 5 min. After rinsing in TBST, non-specific binding was blocked with 2.5% normal goat serum for 20 min. Slides were then incubated with Ki67 primary antibody (Cell Marque, Rocklin, CA, USA) at a 1:100 dilution (0.364 µg/mL) for 30 min. Negative control sections were treated with rabbit IgG (Vector Laboratories, Burlingame, CA, USA) at 1:5,000 (1 µg/mL) for 30 min.

Following two TBST washes, a biotinylated goat anti-rabbit secondary antibody (Vector Laboratories) was applied for 30 min. Slides were rinsed again, and immunoreactivity was visualized using Vector ImmPACT DAB (ImmPACT DAB Cat. No. SK-4105; Vector Laboratories; Newark, CA, USA), containing avidin–peroxidase and 3,3′-diaminobenzidine chromogen, producing a brown precipitate after 5 min of incubation. Finally, slides were rinsed in water and processed on a Leica Autostainer XL for hematoxylin counterstaining, dehydration, and coverslipping. On each slide, the hypothalamic third ventricle (3V) was identified, and five non-overlapping images were captured along the length of the 3V at 400× magnification using a Leica DM750 microscope (Leica Microsystems, Wetzlar, Germany) equipped with a 4K digital camera (View 4K HD Microscope Camera; View Solutions; Rancho Cucamonga, CA, USA). Images were acquired and analyzed using InFocus_V2 software (version 1.8, 20240302), with calibration performed at each magnification using a stage micrometer.

Ependymal cells lining the 3V were quantified and classified as proliferating or non-proliferating. In addition, cells located within 120.23 ± 47 µm from the ventricular surface were included in the analysis. Each image represented an area of 69,940.97 µm², corresponding to a total analyzed area of approximately 0.35 mm² across the five images per sample.

2.9 Statistical analyses

Statistical models for all analysis are presented in Table 2. All data were analyzed using the PROC GLIMMIX procedure in SAS 9.4 (Cary, NC, USA). Individual pig was the experimental unit for all measures. The assumptions of normality of residuals, homogeneity of variance, and linearity were confirmed post-hoc. In utero environment, genetic line, and their interaction were fitted as fixed effect, and the individual pig was fitted as random effect. The means of all repeated measures collected during TN1 period were fitted as covariates for thermoregulatory, production, behavior, and blood gas measures. Study day or period was fitted as a fixed effect for all measures excluding organ weight, histology measures and 2^-ΔΔCT values. Hour was included as a fixed effect for thermoregulatory and behavior measures. Pig restraint duration (in seconds) was also fitted as a covariate in blood gas measures. For hypothalamus 3V histology, sex of pig and orientation of the hypothalamus on the slide was fitted as covariates for cellularity and proliferating KI67 cell counting. Study day was used as repeated measures, and covariance structure was selected based on goodness of fit criteria using the structure with the lowest Akaike Information Criterion value (Littell et al., 1998).

All data are presented as least squares means (LSmeans) ± standard error (SE). Statistical significance was defined as P ≤ 0.05, and a tendency was defined as 0.05 < P ≤ 0.10. Day and hour effects will only be presented or discussed when they interact with either genetic line, in utero environment, and only when differences are present within the same day or hour. If time point measures (e.g., hour, day, period) were deemed non-significant (P > 0.05) or not tendencies (P > 0.10) they were removed from the final analyses.

3 Results

3.1 Thermoregulation

Core body temperature was reduced overall (P < 0.01; -0.55 °C) in IUTN+TOL when compared to IUTN+SEN and IUHS+TOL pigs, but no differences were detected between IUTN+TOL and IUHS+SEN pigs or IUTN+SEN and IUHS+TOL pigs (Table 3). Overall, T_C was greater (P < 0.01) in IUHS pigs (39.56 ± 0.06 °C) compared to IUTN pigs (39.27 ± 0.05 °C), regardless of genetic line and period (Table 3). Moreover, IUHS pigs tended to have greater (P = 0.06; 39.66 ± 0.12 °C) T_C compared to IUTN pigs (39.07 ± 0.12 °C) during the TN period (Figure 3).

Figure 3

Respiration rate was reduced (P < 0.01) in IUHS pigs (52 ± 1 bpm) compared to IUTN pigs (56 ± 1 bpm), regardless of genetic line and period (Table 3). Maximum RR was reduced (P = 0.02) in IUHS pigs (69 ± 2 bpm) compared to IUTN pigs (74 ± 2 bpm), regardless of genetic line and period (Table 3). Minimum RR tended to be increased (P = 0.07) in TOL pigs (39 ± 1 bpm) compared to SEN pigs (37 ± 1 bpm), regardless of period and was reduced (P < 0.04) in IUHS pigs (37 ± 1 bpm) compared to IUTN pigs (39 ± 1 bpm; Table 3).

Overall, T_shoulder tended to be greater (P = 0.06) in TOL pigs (36.12 ± 0.06 °C) compared to SEN pigs (35.95 ± 0.06 °C) regardless of in utero environment (Table 3). Similarly, T_rump tended to be greater (P = 0.06) in TOL pigs (35.72 ± 0.07 °C) compared to SEN pigs (35.53 ± 0.07 °C), regardless of in utero environment (Table 3). No other thermoregulatory measure differences were detected (P > 0.10) with any comparisons (Table 3; Figure 3).

3.2 Blood biochemistry

Carbon dioxide partial pressure was greater in IUHS pigs (P = 0.05; 15.2%) compared to IUTN pigs, regardless of genetic line and period (Table 4). Oxygen partial pressure was reduced in IUHS pigs (P = 0.04; 17.7%) compared to IUTN pigs, regardless of genetic line and period (Table 4). Additionally, PO₂ tended to be reduced (P = 0.06) in the IUHS pigs compared to the IUTN pigs during the TN period (26.0%) and HS1 period (31.0%) but tended to be increased (P = 0.06) in the HS2 period (9.3%) when compared to IUTN pigs, regardless of genetic line (Figure 4A). Furthermore, O₂ saturation tended to be reduced in IUHS pigs (P = 0.06; 12.4%) compared to IUTN pigs, regardless of genetic line and period (Table 4). Moreover, O₂ saturation tended to be reduced (P = 0.06) in IUHS pigs during the TN (17.0%) and HS1 period (25.4%) but tended to be increased (P = 0.06) during the HS2 period (8.2%), regardless of genetic line (Figure 4B). No other blood biochemistry differences were detected (P > 0.10) with any comparisons (Table 4; Figure 4).

Figure 4

3.3 Growth performance

Average daily feed intake was greater (P = 0.04) in TOL pigs (3.02 ± 0.11 kg/d) when compared to SEN pigs (2.67 ± 0.10 kg/d), regardless of in utero environment and period (Table 5). Average daily gain was greater (P = 0.03) in IUHS+TOL pigs (1.31 ± 0.28 kg/d) when compared to IUHS+SEN (0.71 ± 0.22 kg/d), IUTN+SEN (0.82 ± 0.22 kg/d), and IUTN+TOL (0.34 ± 0.22 kg/d) during the TN period (Figure 5A). However, ADG was reduced (P = 0.03) in IUHS+TOL pigs (0.42 ± 0.25 kg/d), IUHS+SEN pigs (0.53 ± 0.22 kg/d), and IUTN+SEN pigs (0.44 ± 0.22 kg/d), when compared to IUTN+TOL pigs (0.60 ± 0.22 kg/d) during HS1 period (Figure 5A). Additionally, IUHS+TOL pigs had reduced ADG (P = 0.03; 0.50 ± 0.27 kg/d) when compared to IUHS+SEN pigs (0.96 ± 0.22 kg/d), IUTN+SEN pigs (0.74 ± 0.22 kg/d), and IUTN+TOL pigs (0.90 ± 0.22 kg/d) during HS2 period (Figure 5A). Furthermore, ADG: ADFI was greater (P = 0.01) in IUHS+TOL pigs (0.71 ± 0.10) compared to IUHS+SEN pigs (0.23 ± 0.09) and IUTN+TOL pigs (0.13 ± 0.09) during the TN period (Figure 5B). No other growth performance differences were detected (P > 0.10) with any comparisons (Table 5).

Figure 5

3.4 Behavior

The proportion of time spent sitting tended to be reduced (P = 0.06) in TOL pigs (1.70 ± 0.78%) compared to SEN pigs (3.79 ± 0.78%), regardless of in utero environment and period (Table 6). Additionally, the proportion of time laying sternally tended to be greater (P = 0.08) in IUHS pigs (39.18 ± 2.06%) compared to the IUTN pigs (33. 81 ± 2.07%), regardless of genetic line (Table 6). No other behavioral differences were detected (P > 0.10) with any comparison (Table 6).

3.5 Organ weight

Brain weight was reduced (P = 0.05) in IUHS pigs (73.24 ± 1.65 g) when compared to IUTN pigs (78.12 ± 1.64 g), regardless of genetic line (Figure 6A). Pituitary gland weight was greater (P = 0.05) in TOL pigs (0.36 ± 0.02 g) when compared to SEN pigs (0.32 ± 0.02 g), irrespective of in utero environment (Figure 6B). Additionally, pituitary gland weight as a percentage of the brain weight was greater (P = 0.03) in TOL pigs (0.49 ± 0.02%) when compared to SEN pigs (0.42 ± 0.02%), irrespective of in utero environment (Figure 6C). No other organ weight differences were detected (P > 0.10) with any comparison (Table 7).

Figure 6

3.6 mRNA abundance

No mRNA abundance differences were detected (P > 0.10) with any comparison (Table 8).

3.7 Hypothalamus histology

Hypothalamus 3V cellularity was greater (P = 0.05) in IUHS+TOL (1,285.00 ± 138.90) and IUTN+SEN (1,219.00 ± 129.55) compared to IUTN+TOL (1072.39 ± 104.23) and IUHS+SEN (1048.03 ± 102.38; Table 9). No other hypothalamus 3V cellularity differences were detected (P > 0.10) with any comparison (Table 9).

4 Discussion

In the present study, IUHS pigs had consistently elevated T_C across postnatal environments relative to IUTN pigs, reproducing our prior work (Johnson et al., 2013; Johnson et al., 2015a), and supporting a developmental-programming model in which IUHS raises the T_C setpoint in pigs. Prior evidence that IUHS pigs sustain greater fasting metabolic heat production (Chapel et al., 2017) and exhibit altered hypothalamic function and size (Maskal et al., 2021) offers a biologically plausible mechanism whereby central programming of a higher T_C setpoint within the hypothalamus (Clark et al., 1939) plus added endogenous heat production results in the greater T_C observed across multiple studies (Johnson et al., 2013; Johnson et al., 2015a). Interestingly, despite the increase in T_C, RR was modestly lower in IUHS versus IUTN pigs overall and skin temperatures were similar, regardless of environmental period. In prior studies, we reported a comparable increase in T_C for IUHS pigs with unchanged RR but higher skin temperatures which likely underscores that thermoregulatory responses are context-dependent (i.e., the balance between latent and sensible heat loss varies with heat-load severity/duration, airflow, RH, water availability, and measurement timing), yielding different RR and skin temperature patterns across settings (Blatteis, 1998; Renaudeau et al., 2011; Johnson et al., 2013; Johnson et al., 2015a). Taken together, the most conservative explanation is that IUHS programs a higher T_C setpoint and greater basal thermogenesis, yielding a biologically relevant upward shift in T_C, while the expression of panting and cutaneous heat loss varies with environmental and cohort context rather than a single consistent IUHS phenotype.

Genetic line further shaped the T_C phenotype, with T_C reduced in IUTN+TOL pigs compared to both IUTN+SEN and IUHS+TOL pigs, while IUHS+SEN pigs exhibited an intermediate T_C that did not clearly differ from the other groups. This pattern indicates that the lower T_C response previously reported for TOL versus SEN lines (Byrd et al., 2025b) is expressed most clearly when pigs develop under IUTN conditions, whereas IUHS appears to erode this advantage in TOL-derived offspring such that IUHS+TOL pigs shift toward the warmer T_C profile observed in IUTN+SEN pigs. Although the difference in T_C was modest in magnitude, even small, sustained reductions in T_C are biologically relevant in growing pigs because they increase the thermal margin before active heat-defense mechanisms must be recruited and reduce the cumulative time spent near upper thermal limits (Johnson et al., 2013; Johnson et al., 2015a). Additionally, RR, T_rump, and T_shoulder tended to be greater in TOL than SEN pigs, regardless of in utero environment, which is consistent with previous observations in F0 lactating sows (Byrd et al., 2025a) and suggests that TOL pigs more readily employ both latent and sensible heat-loss pathways when needed (Blatteis, 1998). Taken together, these results support a model in which TOL genetics favor a slightly cooler basal T_C and more active cooling strategies when prenatal development occurs under IUTN conditions, whereas IUHS shifts the T_C setpoint upward enough to diminish this thermal advantage in TOL offspring during subsequent heat exposure.

Response to thermal stress is also reflected in blood gas dynamics, since RR is closely linked to alveolar ventilation and, in turn, CO₂ clearance and O₂ uptake (Powers and Dhamoon, 2023). In the present study, IUHS pigs had greater overall PCO₂ and reduced PO₂ and O₂ saturation when compared to IUTN pigs, with the reductions in PO₂ and O₂ saturation most apparent during the HS1 and TN periods. When interpreted alongside the modestly lower RR reported for IUHS pigs, this pattern is consistent with mild hypoventilation and reduced gas-exchange efficiency as reported in thermally stressed rabbits exhibiting reduced alveolar ventilation (Zila et al., 2007). It is important to note that the blood gas differences were modest in magnitude and not uniformly expressed across all periods, so they should be interpreted as supportive rather than definitive evidence of altered ventilatory control in IUHS pigs. Genomic selection for TOL versus SEN did not influence any blood gas variable, which was expected given that genetic line differences in RR were limited to a tendency for higher minimum RR in TOL versus SEN pigs. This difference was likely insufficient to produce measurable changes in PCO₂, PO₂, or O₂ saturation. Overall, the blood biochemistry data are most consistent with in utero treatment, rather than genetic line, being the primary driver of the small shifts in PCO₂, PO₂, and O₂ saturation, and they fit within the broader interpretation that IUHS raises T_C and can subtly influence ventilation, while TOL and SEN primarily differ in how they deploy thermoregulatory responses as previously described (Byrd et al., 2025a; Byrd et al., 2025b) rather than in baseline pulmonary gas exchange.

Pigs generally reduce activity when exposed to thermal load, shifting from standing and walking toward more lying and sitting as a behavioral strategy to limit additional metabolic heat production (Collin et al., 2001b; Lucy and Safranski, 2017). Sitting, in particular, is often interpreted as a coping posture as it reduces muscular effort compared to standing but does not provide the full heat-dissipation advantages of lateral lying, and it has been associated with discomfort or stress in heat- and disease-challenged pigs (Collin et al., 2001b; Lucy and Safranski, 2017). Previous reports have shown that IUHS piglets display greater lying behavior and reduced activity in the first days post-weaning (Byrd et al., 2019; Merlot et al., 2019), consistent with a more passive coping style. In the present study, we observed a tendency for the IUHS pigs to spend more time lying sternally compared to the IUTN pigs. Given that increased lying behavior is linked with a passive coping response to thermal discomfort, it could be implied that the IUHS pigs were experiencing more stress and are at greater risk of poor welfare in commercial productive environment compared to the IUTN pigs (Byrd et al., 2019; Merlot et al., 2019) and this observation complements the heightened physiological and thermoregulatory stress responses in the IUHS compared to the IUTN pigs. Similarly, genetic line had clear behavioral differences with TOL pigs tending to spend less time sitting than SEN pigs. Given that increased sitting is consistent with a coping response to greater discomfort or thermal strain (Silva et al., 2025), the greater sitting time in SEN pigs suggests they were experiencing more stress than TOL pigs, which aligns with their reduced ability to maintain euthermia under higher temperatures (Piestun et al., 2008). Our previous work in F0 lactating TOL sows demonstrated enhanced behavioral thermoregulation relative to SEN animals (Byrd et al., 2025a), whereas in the current F2 grow–finish pigs the key behavioral distinction appears to be reduced time in a stress-associated posture (sitting) for TOL. A plausible explanation is that lactating sows and gestating gilts have greater metabolic heat production from milk synthesis and fetal growth (Noblet and Etienne, 1987; Johnson et al., 2019), hence could be considered more sensitive to HS exposure, leading to more pronounced and diverse behavioral adjustments in earlier generations, while in F2 finishing pigs the TOL advantage under HS is expressed more subtly through less time spent in coping postures indicative of stress.

Growth performance responses helped explain the thermoregulatory patterns observed for the genetic line by in utero environment interaction, including the higher T_C in IUHS+TOL pigs. An overall increase in ADFI was observed for TOL versus SEN pigs, suggesting a greater capacity to support growth through higher voluntary nutrient intake (Boddicker et al., 2014), whereas IUHS did not alter ADFI relative to IUTN, which is consistent with our previous work (Johnson et al., 2015a; Johnson et al., 2015b; Johnson et al., 2015c; Maskal et al., 2020). During HS1, IUTN+TOL pigs maintained the greatest ADG, with no differences detected among the other treatment combinations, indicating that in the absence of IUHS, TOL genetics can improve growth performance during HS, a positive outcome of this genomic selection strategy. During the subsequent TN period, IUHS+TOL pigs exhibited the greatest ADG and feed efficiency, suggesting that when pigs are briefly returned to TN conditions between HS bouts, TOL-selected animals were fully able to express their growth potential regardless of in utero environment. However, by the HS2 period, growth converged across treatments, potentially implying that repeated HS exposure and trial duration imposed a common ceiling on ADG that obscured further genetic line or in utero treatment differences. Taken together, the HS1 and TN responses are consistent with the heat increment of feeding, whereby higher intake and faster growth rate increase metabolic heat production that can be dissipated during TN but becomes more difficult to shed once HS is imposed (Curtis, 1983; Collin et al., 2001a; Kerr et al., 2003). In IUTN+TOL pigs, the lower T_C described previously may have preserved a wider thermal gradient during HS1, allowing these pigs to sustain both intake and growth when first exposed to HS. In contrast, IUHS+TOL pigs combined a higher T_C setpoint programmed by the in utero environment with the TOL-associated increase in feed intake and growth potential, such that the additional metabolic heat from feeding during HS1 was likely combined with greater basal thermogenesis from the higher T_C. This combination appears to have been sufficient to blunt the growth benefit of TOL under HS, even though TOL genetics have been advantageous for reproductive output and T_C control in HS-exposed lactating sows in earlier generations (Byrd et al., 2025a; Byrd et al., 2025b). Thus, the period-specific growth responses support a stage-specific interaction in which IUHS can compromise the expression of TOL-related advantages in grow–finish pigs under HS, emphasizing the need to limit IUHS in breeding females if TOL selection is to deliver its full benefits in commercial systems.

Brain weight was reduced in IUHS compared to IUTN pigs in the current study. This observation is consistent with reports on guinea pigs where IUHS resulted in reduced brain size and a compromised postnatal stress response (Jonson et al., 1976). Microcephaly has also been reported in IUHS pigs at different life stages (Cruzen et al., 2015; Johnson et al., 2015a), which is consistent with a smaller brain size because head size is a practical proxy for brain size and neurodevelopment (Frisk et al., 2002). The observed reduction in brain size in IUHS pigs may have negative implications for neurodevelopment (Graham et al., 1998) and homeothermic control, since key physiological functions including metabolism and T_C are tightly controlled by the brain (Tran et al., 2022). The reduced brain size in IUHS pigs may also relate to the greater T_C observed relative to IUTN pigs in this study as T_C is centrally regulated and brain size is an important indicator of neurodevelopment (Frisk et al., 2002), hence a reduced brain size implies altered brain and CNS function, which can reprogram T_C set point as previously hypothesized (Johnson et al., 2013; Johnson et al., 2015a). Furthermore, pituitary gland weight (both absolute and relative to the brain weight) was greater in TOL pigs compared to SEN pigs. The pituitary gland (specifically the anterior pituitary) is recognized as a central endocrine component linking the hypothalamus to the adrenal gland in the HPA axis (Smith and Vale, 2006; Maskal et al., 2021; Tran et al., 2022). The anterior pituitary gland produces the proopiomelanocortin (POMC) which is converted to adrenocorticotropic hormone (ACTH), and ACTH binds to MC2R in the adrenal cortex to stimulate glucocorticoid release into circulation (Stephens and Wand, 2012). Greater pituitary gland size (both absolute and relative to the brain) likely reflects expanded endocrine reserve, as described in physiological states with heightened tropic demands such as pregnancy (Benson et al., 2023). In our context, the combination of larger pituitary mass and improved indicators of heat dissipation observed in the TOL pigs relative to the SEN pigs support the model that genomic selection for TOL likely improves endocrine regulation of the HPA axis during HS exposure. Although the observed thermoregulatory indicators were only marginally greater, they remain consistent with improved heat dissipation in TOL relative to SEN pigs. Together, these observations suggest that genomic selection for TOL improved pigs’ adaptation to postnatal HS challenge regardless of in utero environment.

Despite differences in physiological HS responses between IUHS and IUTN pigs and between TOL and SEN pigs, mRNA abundance did not differ for any of the HPA axis–associated genes evaluated. A plausible explanation for the lack of differences in mRNA abundance between IUHS and IUTN pigs is that postnatal HS exposure altered basal transcript levels, similar to the challenge-dependent mRNA changes reported between IUHS and IUTN pigs following a CRH challenge (Maskal et al., 2021). Similarly, prenatal stress induced by repeated ACTH administration in pregnant sows did not alter fetal brain CRH mRNA abundance, yet increased cellular FBJ osteosarcoma oncogene (C-FOS) mRNA expression in the brain and decreased MC2R mRNA expression in the adrenal gland (Schwerin et al., 2005), indicating that prenatal stress can alter HPA axis activity without producing detectable changes in all axis-associated transcripts. Further studies comparing HPA axis related gene mRNA expressions between IUHS and IUTN pigs should consider non-challenged models to assess differences in gene transcripts basal levels.

The 3V is central to thermoregulation and HPA axis function, integrating thermal and osmotic signals that contribute to HS defense responses (Whyte and Johnson, 2005; Herman et al., 2016; Prevot et al., 2018). In the present study, IUHS+TOL and IUTN+SEN pigs exhibited greater periventricular cellularity than IUTN+TOL and IUHS+SEN pigs, indicating that genetic line and in utero environment interacted to influence cellular organization surrounding the 3V. Although these data do not identify the specific cell populations responsible for this effect, greater cellularity in IUHS+TOL pigs may have contributed to their ability to maintain comparable stress-related indicators despite greater T_C. However, because greater cellularity was also observed in IUTN+SEN pigs, this response may not uniformly reflect improved thermoregulatory capacity and may instead represent treatment-specific differences in hypothalamic organization. KI67-positive cells were uniformly sparse and did not differ among groups, suggesting that cellularity differences were not driven by ongoing differential proliferation at the time of sampling and may reflect previously established developmental differences. Reduced hypothalamic periventricular differentiation following cyclic HS has been reported in 3-day-old chicks (Kisliouk et al., 2014), supporting the possibility that early-life thermal conditions can alter periventricular development. A limitation of the present analysis is that hypothalamic periventricular cellularity likely reflects multiple cell populations, including tanycytes, ependymal cells, glial cells, and adjacent hypothalamic parenchymal cells. Therefore, future use of cell-associated markers, such as vimentin to help identify tanycytes (Haan et al., 2013) and S100β to help distinguish ependymal or glial populations (Goodman and Hajihosseini, 2015), would improve identification and quantification of specific cell types. Collectively, these data indicate that genetic line and in utero environment interact to influence hypothalamic periventricular cellularity, which may contribute to differences in hypothalamic regulation of thermoregulatory and HPA-axis responses during postnatal HS.

Several limitations should be considered when interpreting these observations. Although the study was powered to detect differences in core body temperature based on previous research in IUHS and IUTN pigs (Johnson et al., 2015a), several secondary outcomes, including blood gas variables, behavior, organ weights, mRNA abundance, and hypothalamic histology, should be interpreted as supportive or exploratory. Blood gas differences were not consistently expressed across all environmental periods; therefore, they should be interpreted as evidence of subtle, context-dependent alterations in respiratory physiology. Similarly, differences in brain and pituitary mass may indicate altered neuroendocrine development, but organ mass alone does not in it of itself establish functional changes in neurodevelopment, endocrine reserve, or HPA-axis responsiveness. The absence of treatment differences in HPA-axis–associated mRNA abundance may also reflect the timing of tissue collection following thermal treatments rather than a true lack of programming effects, and future studies should include non-challenged comparisons. Finally, hypothalamic third ventricle cellularity could not be attributed to specific cell populations because cell-specific markers were not used. Therefore, conclusions regarding the functional roles of these cells in thermoregulation and HPA-axis regulation remain speculative and warrant further investigation.

5 Conclusions

This study demonstrated that IUHS and genomic selection for TOL interacted to influence postnatal responses to HS in pigs. Across experimental periods, IUHS was associated with a phenotype characterized by elevated T_C, reduced RR, blood biochemistry changes consistent with reduced ventilatory compensation, and greater lying behavior, collectively suggesting a reduced capacity to cope with heat load. Because skin temperatures did not increase proportionally in IUHS pigs, these observations further indicate that prenatal heat exposure impaired effective heat dissipation. When considered alongside the reduced brain mass observed in IUHS pigs, the data support the concept that IUHS induced central thermoregulatory reprogramming that elevated body temperature and modified HS responses. In contrast, genomic selection for TOL enhanced traits associated with heat dissipation, including greater minimum RR, higher T_shoulder and T_rump, and increased pituitary gland size, suggesting potential differences in neuroendocrine regulation that may support improved thermal responsiveness. Performance responses were also period dependent, with IUHS+TOL pigs performing best under TN conditions and IUTN+TOL pigs performing best under HS conditions. Additionally, SEN pigs spent more time sitting, which may reflect greater discomfort. Collectively, these observations indicate that IUHS altered physiological and behavioral responses to HS, whereas genomic selection for TOL improved traits associated with thermal coping capacity in pigs. Implementation of genomic selection for thermotolerance in commercial production settings, using rate of change in T_C relative to T_DB as a selection criterion, has the potential to enhance both heat tolerance and productive performance during postnatal HS exposure, particularly in pigs gestated under TN conditions.

References

• 1

  AlvarengaA. L. N.Chiarini-GarciaH.CardealP. C.MoreiraL. P.FoxcroftG. R.FontesD. O.et al. (2013). Intra-uterine growth retardation affects birthweight and postnatal development in pigs, impairing muscle accretion, duodenal mucosa morphology and carcass traits. Reprod. Fertil. Dev.25, 387–395. doi: 10.1071/RD12021

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 2

  BensonJ. C.MalyukD. F.MadhavanA.GuerinJ. B.KreckeK. N.LittleJ. T.et al. (2023). Pituitary volume changes in pregnancy and the post-partum period. Neuroradiol. J. 37(1), 39–42. doi: 10.1177/19714009231196470

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  BitsieB.IsonE. K.JenkinsL. P.KloppR.McCabeC.MillsK.et al. (2021). Mammary development in gilts at one week postnatal is related to plasma lysine concentration at 24 h after birth, but not colostrum dose. Animals11, 2867. doi: 10.3390/ani11102867

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  BlatteisC. M. (1998). “ Body temperature,” in Physiology and Pathophysiology of Temperature Regulation. Ed. BlatteisC. M. ( World Scientific Publishing Co, River Edge, NJ), 13–22.

  • Google Scholar
  • View reference in article

• 5

  BoddickerR. L.SeibertJ. T.JohnsonJ. S.PearceS. C.SelsbyJ. T.GablerN. K.et al. (2014). Gestational heat stress alters postnatal offspring body composition indices and metabolic parameters in pigs. PloS One9, e110859. doi: 10.1371/journal.pone.0110859

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  ByrdC. J.AndersonN. C.LugarD. W.SafranskiT. J.LucyM. C.JohnsonJ. S. (2019). Evaluating the effects of in utero heat stress on piglet physiology and behavior following weaning and transport. Animals9, 191. doi: 10.3390/ani9040191

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  ByrdM. H.BritoL. F.MusaJ.DiggsS.SennL. K.WenH.et al. (2025b). Characterizing the effects of gestational heat stress on reproductive efficiency and thermoregulation in replacement gilts divergently selected for thermotolerance in the F1 generation. J. Anim. Sci.103, skaf353. doi: 10.1093/jas/skaf353

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  ByrdM. H.SennL. K.PasternakA.StewartK. R.JohnsonJ. S. (2025c). Early gestation heat stress influences fetal biomarkers indicative of maternal cortisol transfer in pigs. Sci. Rep.15, 26730. doi: 10.1038/s41598-025-12698-w

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  ByrdM. K. H.WenH.FreitasP. H. F.ByrdJ. L.TiezziF.MalteccaC.et al. (2025a). Phenotypic assessment of F0 generation lactating sows genomically selected for heat stress tolerance. J. Anim. Sci.103, skaf205. doi: 10.1093/jas/skaf205

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  CecilM. R.NeenoS. M.ByrdM. K. H.FieldT.RichertB. T.SchinckelA. P.et al. (2025). Determining the optimal macroenvironment temperature to house lactating sows and their litters. J. Anim. Sci.103, skaf255. doi: 10.1093/jas/skaf255

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  ChapelN. M.ByrdC. J.LugarD. W.MorelloG. M.BaumgardL. H.RossJ. W.et al. (2017). Determining the effects of early gestation in utero heat stress on postnatal fasting heat production and circulating biomarkers associated with metabolism in growing pigs. J. Anim. Sci.95, 3914. doi: 10.2527/jas2017.1730

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  ClappJ. F. (1980). Acute exercise stress in the pregnant ewe. Am. J. Obstet. Gynecol136, 489–494. doi: 10.1016/0002-9378(80)90676-6

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  ClappJ. F. (2006). Influence of endurance exercise and diet on human placental development and fetal growth. Placenta27, 527–534. doi: 10.1016/j.placenta.2005.07.010

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  ClarkG.MagounH. W.RansonS. W. (1939). Hypothalamic regulation of body temperature. J. Neurophysiol.2, 61–80. doi: 10.1152/jn.1939.2.1.61

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  CollinA.Van MilgenJ.DuboisS.NobletJ. (2001a). Effect of high temperature and feeding level on energy utilization in piglets. J. Anim. Sci.79, 1849–1857. doi: 10.2527/2001.7971849x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  CollinA.van MilgenJ.DuboisS.NobletJ. (2001b). Effect of high temperature on feeding behaviour and heat production in group-housed young pigs. Br. J. Nutr.86, 63–70. doi: 10.1079/BJN2001356

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  CottrellJ. J.LiuF.HungA. T.DiGiacomoK.ChauhanS. S.LeuryB. J.et al. (2015). Nutritional strategies to alleviate heat stress in pigs. Anim. Prod. Sci.55, 1391–1402. doi: 10.1071/AN15255

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  CruzenS. M.BoddickerR. L.GravesK. L.JohnsonT. P.ArkfeldE. K.BaumgardL. H.et al. (2015). Carcass composition of market weight pigs subjected to heat stress in utero and during finishing. J. Anim. Sci.93, 2587–2596. doi: 10.2527/jas.2014-8347

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  CurtisS. E. (1983). Environmental Management in Animal Agriculture (Ames, IA: Iowa State University Press).

  • Google Scholar
  • View reference in article

• 20

  Federation of Animal Science Societies (2020). “ Chapter 9: swine,” in Guide for the Care and Use of Agricultural Animals in Research and Teaching, 4th ed ( FASS, Champaign, IL), 127–140.

  • Google Scholar
  • View reference in article

• 21

  FreitasP. H. F.JohnsonJ. S.TiezziF.HuangY.SchinckelA. P.BritoL. F. (2024). Genomic predictions and GWAS for heat tolerance in pigs based on reaction norm models with performance records and data from public weather stations considering alternative temperature thresholds. J. Anim. Breed. Genet.141, 257–277. doi: 10.1111/jbg.12838

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  FreitasP. H. F.JohnsonJ. S.WenH.MaskalJ. M.TiezziF.MalteccaC.et al. (2023). Definition of environmental variables and critical periods to evaluate heat tolerance in large white pigs based on single-step genomic reaction norms. Genet. Sel. Evol.55, 42. doi: 10.1186/s12711-023-00842-x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  FriskV.AmselR.WhyteH. E. (2002). The importance of head growth patterns in predicting the cognitive abilities and literacy skills of small-for-gestational-age children. Dev. Neuropsychol.22, 565–593. doi: 10.1207/S15326942DN2203_2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  GoodmanT.HajihosseiniM. K. (2015). Hypothalamic tanycytes, masters and servants of metabolic, neuroendocrine, and neurogenic functions. Front. Neurosci.9, 387. doi: 10.3389/fnins.2015.00387

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  GourdineJ. L.RauwW. M.GilbertH.PoulletN. (2021). The genetics of thermoregulation in pigs: a review. Front. Vet. Sci.8, 770480. doi: 10.3389/fvets.2021.770480

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  GrahamJ. M.EdwardsM. J.EdwardsM. J. (1998). Teratogen update: gestational effects of maternal hyperthermia due to febrile illnesses and resultant patterns of defects in humans. Teratology58, 209–221. doi: 10.1002/(SICI)1096-9926(199811)58:5<209::AID-TERA8>3.0.CO;2-Q

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  HaanN.GoodmanT.Najdi-SamieiA.StratfordC. M.RiceR.El AghaE.et al. (2013). Fgf10-expressing tanycytes add new neurons to the appetite/energy-balance regulating centers of the postnatal and adult hypothalamus. J. Neurosci.33, 6170–6180. doi: 10.1523/JNEUROSCI.2437-12.2013

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 28

  HarrisA.SecklJ. (2011). Glucocorticoids, prenatal stress and the programming of disease. Horm. Behav.59, 279–289. doi: 10.1016/j.yhbeh.2010.06.007

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 29

  HermanJ. P.FigueiredoH.MuellerN. K.Ulrich-LaiY.OstranderM. M.ChoiD. C.et al. (2016). Regulation of the hypothalamic-pituitary-adrenocortical stress response. Compr. Physiol.6, 603–621. doi: 10.1002/cphy.c150015

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  HompesT.IzziB.GellensE.MorreelsM.FieuwsS.PexstersA.et al. (2013). Investigating the influence of maternal cortisol and emotional state during pregnancy on the DNA methylation status of the glucocorticoid receptor gene (NR3C1) promoter region in cord blood. J. Psychiatr. Res.47, 880–891. doi: 10.1016/j.jpsychires.2013.03.009

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  JohnsonJ. S.BoddickerR. L.Sanz-FernandezM. V.RossJ. W.SelsbyJ. T.LucyM. C.et al. (2013). Effects of mammalian in utero heat stress on adolescent body temperature. Int. J. Hyperthermia29, 696–702. doi: 10.3109/02656736.2013.843723

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  JohnsonJ. S.Sanz FernandezM. V.GutierrezN. A.PatienceJ. F.RossJ. W.GablerN. K.et al. (2015b). Effects of in utero heat stress on postnatal body composition in pigs: I. Growing phase. J. Anim. Sci.93, 71–81. doi: 10.2527/jas.2014-8354

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  JohnsonJ. S.Sanz FernandezM. V.PatienceJ. F.RossJ. W.GablerN. K.LucyM. C.et al. (2015c). Effects of in utero heat stress on postnatal body composition in pigs: II. Finishing phase. J. Anim. Sci.93, 82–92. doi: 10.2527/jas.2014-8355

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  JohnsonJ. S.Sanz FernandezM. V.SeibertJ. T.RossJ. W.LucyM. C.SafranskiT. J.et al. (2015a). In utero heat stress increases postnatal core body temperature in pigs. J. Anim. Sci.93, 4312–4322. doi: 10.2527/jas.2015-9112

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  JohnsonJ. S.WenH.FreitasP. H. F.MaskalJ. M.HartmanS. O.ByrdM. K.et al. (2023). Evaluating phenotypes associated with heat tolerance and identifying moderate and severe heat stress thresholds in lactating sows housed in mechanically or naturally ventilated barns during the summer under commercial conditions. J. Anim. Sci.101, skad129. doi: 10.1093/jas/skad129

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  JohnsonJ. S.ZhangS.MorelloG. M.MaskalJ. M.TrottierN. L. (2019). Technical note: development of an indirect calorimetry system to determine heat production in individual lactating sows. J. Anim. Sci.97, 1609–1618. doi: 10.1093/jas/skz049

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  JonsonK. M.LyleJ. G.EdwardsM. J.PennyR. H. (1976). Effect of prenatal heat stress on brain growth and serial discrimination reversal learning in the Guinea pig. Brain Res. Bull.1, 133–150. doi: 10.1016/0361-9230(76)90056-3

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  KatteshH. G.KornegayE. T.KnightJ. W.GwazdauskasF. G.ThomasH. R.NotterD. R. (1980). Glucocorticoid concentrations, corticosteroid-binding protein characteristics and reproduction performance of sows and gilts subjected to applied stress during mid-gestation. J. Anim. Sci.50, 897–905. doi: 10.2527/jas1980.505897x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  KerrB. J.YenJ. T.NienaberJ. A.EasterR. A. (2003). Influences of dietary protein level, amino acid supplementation and environmental temperature on performance, body composition, organ weights and total heat production of growing pigs. J. Anim. Sci.81, 1998–2007. doi: 10.2527/2003.8181998x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  KislioukT.CramerT.MeiriN. (2014). Heat stress attenuates new cell generation in the hypothalamus: a role for miR-138. Neuroscience277, 624–636. doi: 10.1016/j.neuroscience.2014.07.047

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  LaburnH. P.FaurieA.GoelstK.MitchellD. (2002). Effects on fetal and maternal body temperatures of exposure of pregnant ewes to heat, cold, and exercise. J. Appl. Physiol.92, 802–808. doi: 10.1152/japplphysiol.00109.2001

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  LaportaJ.KhatibH.ZachutM. (2024). Review: phenotypic and molecular evidence of inter- and trans-generational effects of heat stress in livestock mammals and humans. Animal18, 101121. doi: 10.1016/j.animal.2024.101121

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  LittellR. C.HenryP. R.AmmermanC. B. (1998). Statistical analysis of repeated measures data using SAS procedures. J. Anim. Sci.76, 1216–1231. doi: 10.2527/1998.7641216x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  LucyM. C.SafranskiT. J. (2017). Heat stress in pregnant sows: thermal responses and subsequent performance of sows and their offspring. Mol. Reprod. Dev.84, 946–956. doi: 10.1002/mrd.22844

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  MaskalJ. M.BritoL. F.DuttlingerA. W.KpodoK. R.McConnB. R.ByrdC. J.et al. (2021). Characterizing the postnatal hypothalamic–pituitary–adrenal axis response of in utero heat stressed pigs at 10 and 15 weeks of age. Sci. Rep.11, 21886. doi: 10.1038/s41598-021-01889-w

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 46

  MaskalJ. M.DuttlingerA. W.KpodoK. R.McConnB. R.ByrdC. J.RichertB. T.et al. (2020). Evaluation and mitigation of the effects of in utero heat stress on piglet growth performance, postabsorptive metabolism, and stress response following weaning and transport. J. Anim. Sci.98, skaa265. doi: 10.1093/jas/skaa265

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 47

  McConnB. R.SchinckelA. P.RobbinsL.GaskillB. N.Green-MillerA. R.LayD. C.et al. (2022). A behavior and physiology-based decision support tool to predict thermal comfort and stress in non-pregnant, mid-gestation, and late-gestation sows. J. Anim. Sci. Biotechnol.13, 115. doi: 10.1186/s40104-022-00789-x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48

  MerlotE.ConstancisC.ResmondR.ServientoA. M.RenaudeauD.PrunierA.et al. (2019). “ Heat exposure of pregnant sows modulates behaviour and corticotrope axis responsiveness of their offspring after weaning”, in: Proceedings of the 53rd Congress of the International Society for Applied Ethology (ISAE): Animal Lives Worth Living (Bergen, Norway: Wageningen Academic Publishers).

  • Google Scholar

• 49

  MonteiroA. P. A.TaoS.ThompsonI. M.DahlG. E. (2014). Effect of heat stress during late gestation on immune function and growth performance of calves: isolation of altered colostral and calf factors. J. Dairy Sci.97, 6426–6439. doi: 10.3168/jds.2013-7891

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 50

  MusaJ.CaseyT. M.FarmerC.ByrdM. K. H.BritoL. F.WenH.et al. (2026). Effect of early to mid-gestation heat stress exposure on mammary development and milk traits in F1 gilts divergently selected for thermotolerance. J. Anim. Sci.104, skag065. doi: 10.1093/jas/skag065

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 51

  National Research Council (2012). “ Composition of feed ingredients,” in Nutrient Requirements of Swine, 11th ed ( National Academies Press, Washington, DC), 239–367.

  • Google Scholar

• 52

  NobletJ.EtienneM. (1987). Metabolic utilization of energy and maintenance requirements in pregnant sows. Livest. Prod. Sci.16, 243–257. doi: 10.1016/0301-6226(87)90042-x

  • CrossRef
  • Google Scholar

• 53

  PiestunY.ShinderD.RuzalM.HalevyO.YahavS. (2008). The effect of thermal manipulations during the development of the thyroid and adrenal axes on in-hatch and post-hatch thermoregulation. J. Therm. Biol.33, 413–418. doi: 10.1016/j.jtherbio.2008.06.007

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 54

  PowersK. A.DhamoonA. S. (2023). “ Physiology, pulmonary ventilation and perfusion,” in Statpearls ( StatPearls Publishing, Treasure Island, FL).

  • Google Scholar

• 55

  PrevotV.DehouckB.SharifA.CiofiP.GiacobiniP.ClasadonteJ. (2018). The versatile tanycyte: a hypothalamic integrator of reproduction and energy metabolism. Endocr. Rev.39, 333–368. doi: 10.1210/er.2017-00235

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 56

  RavanelliN.CasasolaW.EnglishT.EdwardsK. M.JayO. (2019). Heat stress and fetal risk: environmental limits for exercise and passive heat stress during pregnancy, a systematic review with best evidence synthesis. Br. J. Sports Med.53, 799–805. doi: 10.1136/bjsports-2017-097914

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 57

  RenaudeauD.GourdineJ. L.St-PierreN. R. (2011). A meta-analysis of the effects of high ambient temperature on growth performance of growing–finishing pigs. J. Anim. Sci.89, 2220–2230. doi: 10.2527/jas.2010-3329

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 58

  SchwerinM.KanitzE.TuchschererM.BrüssowK. P.NürnbergG.OttenW. (2005). Stress-related gene expression in brain and adrenal gland of porcine fetuses and neonates. Theriogenology63, 1220–1234. doi: 10.1016/j.theriogenology.2004.06.004

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 59

  SilvaA.MaChadoN. A. F.Barbosa-FilhoJ. A. D.GadelhaC. R. F.MarquesJ. I.LeiteP. G.et al. (2025). Effect of transportation time on weaner pigs’ welfare and productive losses in a semi-arid region. Vet. Sci.12, 214. doi: 10.3390/vetsci12030214

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 60

  SkibielA. L.PeñagaricanoF.AmorínR.AhmedB. M.DahlG. E.LaportaJ. (2018). In utero heat stress alters the offspring epigenome. Sci. Rep.8, 14608. doi: 10.1038/s41598-018-32975-1

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 61

  SmithS. M.ValeW. W. (2006). The role of the hypothalamic-pituitary-adrenal axis in neuroendocrine responses to stress. Dialogues Clin. Neurosci.8, 383–395. doi: 10.31887/DCNS.2006.8.4/ssmith

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 62

  StephensM. A. C.WandG. (2012). Stress and the HPA axis: role of glucocorticoids in alcohol dependence. Alcohol. Res. Curr. Rev.34, 468–483. doi: 10.35946/arcr.v34.4.11

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 63

  TranL. T.ParkS.KimS. K.LeeJ. S.KimK. W.KwonO. (2022). Hypothalamic control of energy expenditure and thermogenesis. Exp. Mol. Med.54, 358–369. doi: 10.1038/s12276-022-00741-z

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 64

  WenH.JohnsonJ. S.FreitasP. H. F.MaskalJ. M.GloriaL. S.AraujoA. C.et al. (2023). Longitudinal genomic analyses of automatically recorded vaginal temperature in lactating sows under heat stress conditions based on random regression models. Genet. Sel. Evol.55, 68. doi: 10.1186/s12711-023-00868-1

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 65

  WhyteD. G.JohnsonA. K. (2005). Thermoregulatory role of periventricular tissue surrounding the anteroventral third ventricle (AV3V) during acute heat stress in the rat. Clin. Exp. Pharmacol. Physiol.32, 457–461. doi: 10.1111/j.1440-1681.2005.04211.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 66

  YahavS.CollinA.ShinderD.PicardM. (2004). Thermal manipulations during broiler chick embryogenesis: effects of timing and temperature. Poult. Sci.83, 1959–1963. doi: 10.1093/ps/83.12.1959

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 67

  ZilaI.BrozmanovaA.JavorkaM.CalkovskaA.JavorkaK. (2007). Effects of hypovolemia on hypercapnic ventilatory response in experimental hyperthermia. J. Physiol. Pharmacol.58, 781–792.

  • Google Scholar