Relationship between hair cortisol and plasma heat shock proteins in response to heat stress in tolerant and sensitive cattle breeds

Abstract

This study evaluated the relationship between cortisol and Heat Shock Proteins (HSP) release, and heat stress (HS) induced by high Temperature-Humidity-Index (THI) in cattle. Cows of heat-tolerant (Alentejana, ALT; Mertolenga, MERT) and heat-sensitive (Holstein-Frisian, HF) Bos taurus breeds, calving in August and January, were sampled at 39 ± 5 days postpartum at the end of Summer and Winter. For each cow, the THI of the previous 60 days (THI60) was calculated, ocular, vaginal and rectal temperatures were measured, blood was collected for BHB, total T3 and T4, HSP60, HSP70 and HSP90 assays, and neck hair was collected for measurement of cortisol concentrations (HCC). Plasma HSP60 concentrations were decreased by the independent increase in THI60 and HCC, but in ALT cows the simultaneous THI60 and HCC increase had the opposite effect. Plasma HSP70 concentrations were not related to HCC, but were increased by increased THI60 only in ALT and MERT cows. Plasma HSP90 concentrations were decreased by the independent increase in THI60 and HCC in ALT and HF cows, but in MERT cows the HSP90 decrease was only accomplished when in presence of the simultaneous increase in THI60 and HCC. In conclusion, significant breed-specific phenotypes of response to HS were identified.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-52553-0.

Introduction

The increasing demand for animal sourced foods is a worldwide growing challenge. Climate change has enhanced Heat-Stress (HS) episodes, which are responsible for productive losses in livestock¹. These losses are exacerbated in heat-sensitive selected dairy and beef cattle breeds, where the selection strategy induced a loss of hardiness and adaptive plasticity^2,3. Ability to adapt to HS depends on species, breeds and life stage, as well as the environment in which they evolved during successive generations^4,5. Evolutive adaption and acclimation determined an increased frequency of thermotolerance genes in cattle breeds (Bos taurus and Bos indicus), which confer increased ability to regulate core body temperature and maintain cellular function⁶. In this scenario, genetic selection emerges as a cornerstone strategy to mitigate the impact of HS. Experimental evidence of this strategy is illustrated by the Slick mutation⁶. A mutation in the prolactin receptor gene (PRLR), conferring a short, slick hair coat phenotype (Slick mutation), initially identified in Caribbean criollo Senepol and Carora breeds, is associated with improved body temperature regulation⁷, and has been introgressed into the Holstein breed, improving resistance to HS without an apparent negative effect on milk production⁸. Several single nucleotide polymorphisms, namely in Heat Shock Protein (HSP) genes, were associated to heat tolerance in cattle^6,9. A deletion mutation in the bovine HSP70 (HSPA1L) gene is associated with increased transcript abundance for HSPA1A/L and increased cell survival when exposed to heat shock, and polymorphisms in the HSP70A1A, HSF1, HSBP1 and HSP90AB1 genes were associated to heat tolerance in cattle, as exemplified by the HSPA1A gene in Vechur and Kasaragod cattle¹⁰.

The knowledge on strict HS adaptive phenotypes may provide the tool for the evaluation of their genomic background, which can be conveyed to genetic selection. This extends to knowledge on the relationship between these strict phenotypes and their coding genomic characterization, in order to ascertain their value for genetic selection. Indeed, the identification of mutations conferring thermotolerance at the whole animal or cellular level will depend on the collection of more phenotypes for thermotolerance traits and extensive genetic analysis to discover causative mutations^6,11. However, there is a need to deepen the current understanding of the physiologic HS adaptive mechanisms in heat-tolerant and heat-sensitive phenotypes. Additionally, phenotypic studies may enable the development of new strategies of HS control (cooling, feeding, grazing, management strategies) through different forms of heat relief.

The Iberian Peninsula is within world’s most affected places by global warming, due to increased heat waves and intensification of hot seasons¹². The native breeds Alentejana (ALT) and Mertolenga (MERT), share the same territorial origin and distribution in the Alentejo province of south Portugal. Both breeds evolved as sub-breeds of the historic Transtagana breed, being Bos primigenius their ancestor¹³. However, MERT also displays physical characteristics from Bos desertorum (desert ox)¹⁴. The ALT, formerly used for rural work, is a full-brown color heavy breed (600 kg for females, 1000 Kg for males), whereas the MERT is a less heavy (450 Kg for females, 800 Kg for males) breed with a coat color ranging from white to full brown^15,16. Both breeds are reared in a pasture-based system, under natural edaphoclimatic conditions, using extended natural mating breeding seasons^13,15,16.

Adaption to HS, involving morphological, behavioral, physiological, neuro-endocrine, blood biochemical and metabolic, and cellular and molecular responses was the object of comprehensive reviews¹¹. The known mechanisms of heat-tolerance of ALT and MERT cattle were addressed in previous studies^5,14,17. These studies documented that, when exposed to high environmental THI, ALT and MERT animals are able to control the internal temperature rise, without a significant increase in endogenous cortisol release. However, these two breeds trigger different strategies to maintain homeostasis through heat elimination. The ALT breed shows a significant decrease in endogenous metabolism, whereas the MERT breed shows a significant twofold increase in the respiratory rate. Also, the ALT breed increases the sweating rate (up to 156 g m^{− 2} h^{− 1}), whereas in the MERT breed the sweating rate is kept unchanged^14,17. This evidences that Bos taurus breeds with the same origin and adapted to the same edaphoclimatic ecosystem show different strategies of heat relief. Different strategies of HS mitigation were also developed by other cattle species and breeds from similar environments and with apparent phenotypic similarities. In dwarf cattle breeds (Bos taurus indicus), tolerance to heat is associated with small body size, low red blood cell volume and high hemoglobin concentration. However, Vechur cattle relied more on pulse rate, while Kasaragod cattle relied on the respiratory rate to maintain body temperature¹⁸. In the Senepol breed, which carries the Slick gene, the heat tolerance was once thought to be linked to its African Zebu and Bos taurus origins, but this characteristic may well seat its origin in European Bos taurus breeds¹⁹. This highlights the importance of the European, namely Iberian, cattle breeds’ genetics and phenotypes, for the study of HS responses and mechanisms.

Cortisol release is a physiological response to HS, dependent on breed, individual sensitivity, and stressor intensity⁹. Most studies established a link between high plasma cortisol concentrations and HS in cattle, evidencing increased cortisol concentrations during HS^20–22. However, as evidenced in the ALT and MERT breeds this relationship may not be present⁵. Also, Holstein cows carrying the Slick gene mutation or the wildtype gene, have the same plasma cortisol concentrations, despite their different ability to maintain internal body temperature²³. Most studies on cortisol release under HS addressed short-term acute HS, whereas chronic effects, as those originated by the whole hot season are less documented²⁴. Additionally, evaluation through plasma/serum concentrations only reflect daily fluctuations, and do not convey information on the accumulated exposure to cortisol release²⁵.

Heat shock response is a universal ancient mechanism, allowing survival when animals face sudden increased temperatures. In cattle, it includes activation of heat shock factor 1 that induces synthesis of HSP²⁶, a large family (15 kDa − 110 kDa) of heat-inducible gene products highly preserved in animals, including ruminants²⁷. Their chaperone function consists in protein trafficking and folding, degradation of misfolded proteins, disaggregation of protein complexes and solubilization of aggregated proteins, preserving cellular proteome homeostasis and functionality. Collectively, HSP constitute a ubiquitous complex and versatile network of chaperones²⁸, being the most abundant cellular protein family, representing about 5–10% of the cellular protein content²⁹. However, when the cell faces an insult, as in response to HS, this amount increases up to three folds, by up-regulation and expression of inducible forms^30,31. The relevant role of HSP during HS conducted to the concept that HSP are a primary way to assess the severity of heat-stress³².

The HSP60 is mainly found in mitochondria and, in less proportion, in the cytoplasm and cell membrane. It has no inducible isoform but is up-regulated during cell stress, although it is not clear how it appears in bloodstream^33,34. Less studied in cattle, in humans HSP60 acts as pro-inflammatory at high concentrations, and as anti-inflammatory at low concentrations, mainly interacting with macrophages and dendritic cells³⁵, showing a potential value as marker of tissue regeneration, wound healing³⁶, and cancer³⁷.

Contrary to HSP60, both HSP70 and HSP90 have stress inducible isoforms (HSPA1A and HSP90αA1, respectively) localized in the cytosol³⁰. The HSP70 is the most studied and already linked to HS in cattle^38,39. It shows a complex immunomodulatory function, being either pro- or anti-inflammatory, depending on concentration, but found in increased concentrations in diseased animals³⁷. In fact, extracellular HSP70 triggers the immune system by TNFα induction and antigen presentation to T cells⁴⁰, and suppresses inflammation via IL-10 increment, being appointed for graft tissue repair, and treatment of human colitis and arthritis⁴¹. The HSP90, mostly studied in humans and lab species, also acts as an inflammation modulator, with inhibition linked to inflammation reduction in several diseases⁴². In cattle, HSP studies reported a strong positive correlation between plasma HSP70 concentrations and THI in beef steers⁴³, and increased HSP70 (but not HSP60 and HSP90) blood expression during HS in Simmental cattle⁴⁴. Other studies reported low HSP27, HSP60 and HSP90 concentrations in endometrial tissue of dairy cows, during HS⁴⁵. Serum and endometrial HSP seasonal variations were also investigated in goats^46,47.

Most studies established higher plasma cortisol concentrations and the enhanced expression of HSP genes, namely the HSP70 gene, in cattle under HS¹¹. However, only very few studies addressed the relationship between these two parameters, and in most of these studies, not related to HS. In a study in Hanwoo steers⁴⁸, after acute HS, serum cortisol concentrations were not increased, but mRNA HSPP liver expression was increased. This gap in knowledge is explored in the present study, where in ALT and MERT breeds, exposure to increased environmental THI is not associated to an increase in endogenous cortisol release, contrary to Holstein-Frisian cattle. More research is needed to evaluate the relationship and cross-regulation between cortisol and HSP during HS, as these physiologic mechanisms may reflect biological markers of phenotypes of resistance to HS, and this information can ultimately be incorporated into genetic selection strategies to face sustainable cattle production in a climate change scenario.

Considering the above gaps in knowledge, the objectives of the present study were to evaluate: (i) the relationship between THI, cortisol and HSP release under thermoneutrality (Winter) or HS (Summer) in heat-tolerant and heat-sensitive cattle breeds; and (ii) the breed-specific associated physiologic mechanisms (body temperatures and metabolic parameters) put forward to respond to HS, in relation to cortisol and HSP endogenous release. As innovative issues, the study considered the accumulated effects (THI and cortisol) of the chronic HS resulting from the whole hot season (Summer), compared to the thermoneutral season (Winter), on the plasmatic concentrations of HSP70, and the less studied HSP60 and HSP90. The study took the advantage of evaluating two heat-tolerant native cattle breeds, with common genetic ancestry, but with different physiologic strategies to cope with HS, reared in the same herds (two herds with both breeds), in the same location, and with identical management. Two dairy herds with a heat-sensitive cattle breed, in the same location, were included in the study, to identify the potential strict phenotype of the THI-cortisol-HSP relationship. Besides the evaluation of the less documented relationship between HS, cortisol and HSP release, the rationale of the study was also to evaluate potential breed-specific strict phenotype responses to HS, as by characterizing strict phenotypes of HS adaptive mechanisms, researchers are then able to determine their genomic background, which may be used in genetic selection. New knowledge in these phenotypes may also facilitate the development of new strategies of HS control. The hypothesis to be tested were that: (i) cortisol and HSP interaction modulate the response to THI; and (ii) this modulation is breed specific.

Methods

Experimental procedures

Localization and meteorological data

This study was conducted in the Alentejo province of South Portugal (district Évora, 38.5714° N, 7.9135° W), where climate is defined by two well established seasons: Summer, which is dry and hot; and winter, which is rainy and cold⁴⁹. The location of the herds was characterized by a Csa type Mediterranean climate according to Köppen-Geiger⁴⁹. IPMA (Portuguese Institute for Sea and Atmosphere, I. P.) provided retrospective meteorological data from the nearest (< 8 Km) weather station (Table 1). Cow side meteorological data were registered with a hygro-thermometer (EXTECH-RH101, Rotterdam, Netherlands). Temperature-Humidity Index (THI) was calculated according to National Research Council⁵⁰ formula:

Table 1.

Seasonal meteorological data in the location of the herds.

Parameter	Summer			Winter
	Global (24 h)	Hottest hours (11am-4pm)	Night (9pm-6am)	Global (24 h)	Hottest hours (11am-4pm)	Night (9pm-6am)
Maximal temperature (°C)	41.6 ± 2.4	41.6 ± 2.4	31.7 ± 3.8	22.9 ± 2.7	22.7 ± 2.5	16.2 ± 1.3
Minimal temperature (°C)	9.9 ± 1.5	19.2 ± 1.7	9.9 ± 1.47	−2.7 ± 1.9	3.1 ± 3.1	−1.9 ± 1.8
Mean temperature (°C)	23.2 ± 1.4	30.2 ± 1.5	18.4 ± 1.4	10.2 ± 1.2	14.5 ± 1.9	7.8 ± 1.6
Relative humidity (%)	63.6 ± 6.1	38.9 ± 5.0	80.9 ± 6.9	81.9 ± 3.6	66.2 ± 8.6	91.0 ± 2.6
Maximal THI	85.5 ± 2.1	85.5 ± 2.1	75.3 ± 6.4	67.6 ± 2.2	67.5 ± 2.1	60.7 ± 1.8
Minimal THI	49.9 ± 2.4	65.3 ± 2.4	49.9 ± 02.4	28.9 ± 3.9	38.3 ± 5.6	29.8 ± 3.6
Mean THI	69.1 ± 1.4	76.3 ± 1.1	63.9 ± 1.9	50.7 ± 1.8	57.7 ± 2.8	46.7 ± 2.7

Where, T_db=dry-bulb air temperature (°C) an RH= relative humidity.

THI60 was calculated as the mean THI value of the previous 60 days.

Animals and management

The field study took place between January 1, 2021 and December 31, 2022, including two breeding seasons. Two beef farms with both purebred ALT and MERT cows were selected for the study. The beef cow herds followed the Portuguese traditional production system. Beef cows were grazed on natural pasture (Supplementary Table 1 A) under natural edaphoclimatic conditions, with natural shaded areas and ad libitum access to water. Hay supplementation (Supplementary Table 1B) was available ad libitum during seasonal pasture shortage (June to November). Reproductive management consisted of a natural breeding season including one bull to 30–40 cows with a calving season spanning from July to February. The ALT and MERT cows presented the full brown coat color.

Two dairy herds with Holstein-Friesian (HF) cows, in the same location of the beef herds, were selected for the study. Herd A was a 250-cow medium-yielding farm (average 9150 Kg/cow/305d) and a twice daily milking routine, whereas Herd B was a 500-cow high-yielding farm (average 13500Kg/cow/305d) with a thrice daily milking routine. In both herds, housing included stalls with free walking area, roof cover to allow continuous shade, sand individual beds, and cooling systems with mechanical ventilation over the beds, and water sprinklers over the feeding area and milking waiting area. Cows had ad libitum access to water and were fed with a TMR distributed three times a day formulated to cover maintenance and milk production (Supplementary Table 2). Reproductive management followed a year-round calving pattern, using visual heat detection and artificial insemination.

In both beef and dairy herds, exclusion criteria included cows with dystocia, stillbirth and any puerperal clinical disease, as well as cows with any treatment prior or during the study. Inclusion criteria included healthy cows that had calving in Summer (August) and Winter (January). In beef breeds, multiparous (parity range 3–5) ALT cows (n = 34) and MERT cows (n = 55) had a mean of 7 years of age (mean ± SD: 86.7 ± 26.7 months and 87 ± 40 months, respectively). In dairy herds, cows of Herd A (n = 22) had a mean age of 4 years (mean ± SD: 45.2 ± 19.7 months and cows of Herd B (n = 44) had a mean age of 3 years (mean ± SD: 38.3 ± 13.8 months). In both herds, the parity range was within 1–5 parturitions, detailed as follows, respectively in Herd A and B: Parity 1–6 and 13; Parity 2–8 and 19; Parity 3–3 and 5; Parity 4–2 and 6; Parity 5–3 and 1. After the study, cows followed their productive live within the respective herds.

Experimental design

All enrolled cows (n = 155) were sampled once at 38.7 ± 5.3 use instead 39 and 5 to stay identical to abstract(mean ± SD) days postpartum (DPP). The retrospective meteorological data considered the interval from 3 weeks before parturition (July or December) until sampling, therefore comprising 2 months (THI60). Cow-side meteorological data was registered just before sampling with a hygro-thermometer (EXTECH-RH101, Rotterdam, Netherlands). Cows were examined and sampled in the morning from 10 to 12 am, in a headlock system, in the following sequential order. Eye thermography, rectal and vaginal temperatures were immediately accomplished to avoid changes caused by handling. Blood samples were then collected from the coccygeal vein with a 18 G needle into 10 mL dry and EDTA (Vacutest KIMA, Arzegrande, Italy) tubes for measurement of serum beta-hydroxybutyrate (BHB), total T3 and T4 concentrations, and plasma HSP60, HSP70 and HSP90 concentrations. Finally, hair from the neck was collected for cortisol measurement, and body condition score (BCS) was assessed on a scale from 1 to 9⁵¹ in ALT and MERT breeds, and in a scale from 1 to 5⁵² in the HF breed.

The sampling occurred at the end of Summer (September) and Winter (February), thus the mean THI of the 60 days before sampling evidenced the effect of chronic HS or their total absence, respectively, encompassing approximately 3 weeks before and 5 weeks after parturition. This could be related to the total endogenous cortisol release, through the measurement of hair cortisol concentrations (HCC). Since the hair grows approximately 0.6 mm per day⁵³, samples collected (2–3 cm long) approximately reflected 60 days of cortisol incorporation.

The beef and dairy herds were submitted to similar climatic parameters, providing similar environmental THI for all cows. However, housing conditions differed between beef and dairy cows, and were expected to produce different cow-side THI values (lower for dairy than for beef cows). Management was different between beef and dairy herds, and the suckling versus lactating status a priori raised a difference in the metabolic status of beef and dairy cows. From the above, the study focused on the effects of THI (Summer vs. Winter) on the physiologic response parameters (body temperatures, metabolic indicators, cortisol and HSP concentrations, and their relationships) of each breed. However, a comparison between breeds was also undertaken, as this enabled the evaluation of breed specific phenotypes of response to HS. Although the comparison between beef breeds is straightforward, the comparison beef-dairy has to be taken with caution due to the intrinsic physiologic and management differences. Nevertheless, in dairy cows, as parameters were measured in both seasons, at the same postpartum timepoint, it is possible to understand the additional effect of the higher THI of Summer over the metabolic effect, on the analyzed parameters.

Ocular thermography and body temperatures

Ocular thermographs were acquired at 1 m from the cow, perpendicular to the left eyeball, with the cow placed in the shade to avoid artifacts caused by sun exposure⁵⁴. Thermographs were taken with a FLIR^®EX8 thermographer (Teledyne Flir, Oregon, USA), with an emissivity 0.98, and analyzed with software FLIR Tools^TMPC at 76.800 pixel (320 × 240) resolution and 0.05 °C thermal sensitivity (Fig. 1). For each thermograph, a circle was drawn around the eyeball, including the skin of the eye cavity and the lacrimal gland in which the maximum and minimum temperatures (OcularMax and OcularMin) were measured, respectively, and the mean ocular temperature calculated (OcularMean). The vaginal and rectal temperatures were acquired with a DIGI-VET SC12 thermometer (Kruuse^®, Langeskov, Denmark) with an accuracy of 0.1 °C.

Fig. 1.

Representative thermographic image under thermoneutrality in native Alentejana (A) and Holstein-Friesian (B) cows.

Blood metabolic parameters

After collection from the coccygeal vein, blood was centrifuged at 2000 x g for 15 min, and aliquoted and stored at −80 °C. Serum BHB was measured using a handheld meter (Freestyle Precision NEO, Abbott^®, Fremont, United States) with Optimum β-Ketone strips, as validated by others⁵⁵. Serum T3 and T4 concentrations were measured by chemiluminescence (IMMULITE 1000, Siemens) using LKT31 and LKT41 kits, respectively (Siemens Healthcare Diagnostics Products, Ltd, Gwynedd, U cows-K)⁵⁶. The intra-assay and inter-assay coefficients of variation were 3.3% and 2.1% for T3 and 5.3% and 3.0% for T4, respectively⁵⁷.

Plasma Heat Shock Proteins

Plasma HSP60, HSP70 and HSP90 were measured with the MyBiosource Bovine HSP60 ELISA kit, Abbexa Cow Heat Shock 70 kDa Protein 1 A ELISA kit and MyBiosource Bovine Heat Shock Protein 90 kDa Alpha A1 ELISA kit, respectively (MBS7606407, San Diego, USA; abx150116, Cambridge, United Kingdom; MBS45994, San Diego, USA, respectively). The intra-assay and inter-assay coefficients of variation were 6.2% and 10.4% for HSP60, 7.4% and 9.5% for HSP70 and 7.8% and 14.7% for HSP90, respectively⁵⁷.

Hair cortisol assay

Briefly, hair was shaved with an electric clipper, on the left side of the neck and close to the skin, placed into a 2mL Eppendorf, and immediately stored in a cool place protected from light. Dark hair was shaved in an area of approximately 2.5 cm² to avoid pigmentation influence in cortisol concentrations^58,59. Since the hair grows approximately 0.6 mm per day⁵³, samples (2–3 cm long) approximately reflected 60 days of cortisol incorporation.

For extraction, 200 mg of hair was washed with 3 mL 2-isopropanol (VWR, Radnor, USA), vortexed for 1 min, and the supernatant discharged (repeated three times). Samples were left to dry for 24 h at room temperature, and then manually cut into fragments up to 2 mm. Afterwards, 50 mg of hair was weighed and placed in a glass tube within 1.5 mL methanol (VWR, Radnor, USA) to be extracted for 16 h in a stirring water bath (50 °C). Upon extraction, 0.75 mL was evaporated and dried extracts were reconstituted in 0.25 mL PBS at pH8. Cortisol concentrations were measured in duplicate using a commercially available ELISA Kit for Salivary Cortisol (DRG Instruments GmbH, Marburg, Germany). Results were calculated using a 4PL curve fit. The intra-assay CV was 4.5%, and the inter-assay CV was 12.8%.

Statistical analysis

In this observational study, statistical analysis was performed using the SAS 9.4 version software (SAS Institute Inc 2024). Categorical data were analyzed by Fisher’s exact test. After testing data normal distribution by PROC UNIVARIATE, the variables ocular temperature, rectal and vaginal temperature, serum BHB, T3, T4 and HCC were normally distributed, and the variables plasma HSP60, HSP70 and HSP90 were non-normally distributed. To test the fixed effects of Breed (ALT, MERT, HF), Season (Summer – S; Winter – W) and Breed*Season, for normally distributed data, significant differences were determined using two-way ANOVA or Welch’s ANOVA when necessary, based on Levene’s test. For non-normally distributed data, the Kruskal-Wallis Test was used followed by Friedman test when necessary. Pearson correlations were calculated using PROC CORR to investigate linear relationships. Multiple regression analysis was conducted using PROC GLM and PROC Glimmix to assess the effects of Breed, THI60 and HCC and their interactions on HSP, using the following model:

Where µ is the overall intercept and ɛ is the random error or residual effect.

Since HCC had no effect in HSP70, the final model for plasma HSP70 concentrations was assessed using the following model:

Where µ is the overall intercept and ɛ is the random error or residual effect.

The best fit model (goodness of fit and complexity) was evaluated using Akaike information criterion (AICC) and coefficient of determination (R²) for normal distributed data, and Pearson Chi-Square/DF for non-normal distributed data. Linear regression graphs were performed in SPSS (IBM SPSS Statistics 27) and 3D multiple regression graphs were performed on Origin2019 (Originlab Corporation, USA) according to best fit models previously found on multiple regression analysis. Values were expressed as Mean ± SEM for normally distributed data and as Median and quartiles (25% and 75%) for non-normally distributed data and considered statistically different when p ≤ 0.05.

Results

The results of the final statistical models used in the analysis of the fixed effects Breed, Season and Breed*Season are shown in Supplementary Table 3.

Meteorological data

The environmental mean THI60 retrieved from the meteorological stations was higher (p < 0.0001) in S than in W (S – 70.0 ± 1.4 vs. W – 50.8 ± 2.0). The cow-side mean THI, retrieved from the environment at sampling was higher (p < 0.0001) in S than in W (S – 73.4 ± 5.4 vs. W – 61.6 ± 5.3). The Summer mean THI was higher (p < 0.01) in ALT and MERT than in HF (ALT – 75.7 ± 5.9 vs. MERT – 75.5 ± 5.6 vs. HF – 70.7 ± 3.7).

As expected, the THI60 calculated from data of the meteorological stations close to the beef and dairy herds was similar, as herds were in close locations within the region. As also expected, due to the different housing conditions of beef and dairy herds, the cow-side THI at sampling (location of the headlock system) was higher in the beef than in the dairy herds. However, in the three breeds the Summer THI and THI60 were significantly higher than the Winter THI and THI60, providing evidence of HS during Summer, in contrast to thermoneutrality in Winter.

Body temperatures

Vaginal temperature was affected by Breed (p < 0.0001), Season (S – 39.1 ± 0.5 °C vs. W – 38.8 ± 0.5 °C; p < 0.0001) and Breed*Season (p = 0.01) (Fig. 2A), as ALT was the only breed without a significant Summer increment. Rectal temperature followed the same pattern of vaginal temperature.

Fig. 2.

Graphical representation of the Breed*Season effect on vaginal (A) and ocular (B) temperatures of Alentejana (ALT, n = 34), Mertolenga (MERT, n = 55) and Holstein-Frisian (HF, n = 66) postpartum cows, in Summer and Winter. A – bc and dc, p < 0.05; bd, p < 0.0001. B – ab, p < 0.01; ac and bc, p < 0.0001.

Mean ocular temperature was affected by Breed (p < 0.01), Season (S – 35.6 ± 1.5 °C vs. W – 33.2 ± 2.6 °C; p < 0.0001), and Breed*Season (p < 0.0001) (Fig. 2B), as HF was the only breed without a significant increment in Summer.

These results evidenced breed differences in the body temperatures’ response to THI. Whereas ALT cows were able to control the rise in internal body temperatures (vaginal and rectal) in Summer, MERT and HF cows displayed a rise in internal body temperatures in Summer. In contrast, HF cows were able to control the rise in external body temperature (ocular) in Summer, whereas ALT and MERT cows, displayed a rise in ocular temperature in Summer.

Metabolic parameters

The total T4 serum concentrations tended (p = 0.06) to be lower in S than in W, were affected by Breed (p < 0.0001) and tended (p = 0.09) to show a Breed*Season effect (Fig. 3A). The ALT was the only breed with a significant (p < 0.05) Summer decrease in serum concentrations (S – 4.6 ± 1.0 µg/dL vs. W – 5.4 ± 1.0 µg/dL).

Fig. 3.

Graphical representation of the Breed*Season effect on serum metabolic indicators, total thyroxine (T4; A), total triiodothyronine (T3; B) and β-hydroxybutyrate (BHB; C) in Alentejana (ALT, n = 34), Mertolenga (MERT, n = 55) and Holstein Frisian (HF, n = 66) cows. A – ab, p < 0.05; ac, p < 0.0001; bc, p < 0.01. B – ab, p < 0.001; ac, p < 0.0001; cb, p < 0.01; C ab p < 0.01; ac, p < 0.0001; bc, p < 0.05.

The total T3 serum concentrations were affected by Season (p < 0.0001), by Breed (p < 0.05) and Breed*Season (p < 0.01) (Fig. 3B). The ALT was the only breed with a significant (p < 0.0001) Summer decrease in serum concentrations (S – 119.71 ± 29.40 ng/dL vs. W – 179.3 ± 34.16 ng/dL).

The serum BHB concentrations were affected by Season (p < 0.05), by Breed (p < 0.0001) and tended (p = 0.1) to be affected by Breed*Season (Fig. 3C). The HF was the only breed with a significant (p < 0.001) Summer decrease in serum concentrations (S – 0.8 ± 0.5 mmol/L vs. W – 1.3 ± 1.1 mmol/L).

These results evidenced breed differences in the regulation of the endogenous metabolic rate in response to Summer HS, as only ALT cows exhibited a significant drop in total serum T4 and T3 concentrations in Summer.

Hair cortisol concentrations (HCC)

Overall, HCC was affected by Season (p < 0.01), and by Breed (p = 0.0001), but there was no Season*Breed effect (p = 0.13) (Fig. 4). The HF was the only breed with a significant (p < 0.05) Summer increment in HCC (S – 20.2 ± 7.1 pg/mg vs. W – 13.9 ± 5.5 pg/mg).

Fig. 4.

Graphical representation of the Breed*Season effect on hair cortisol concentrations in Alentejana (ALT, n = 34), Mertolenga (MERT, n = 55) and Holstein Frisian (HF, n = 66) cows. ab and bd, p < 0.01; ad, p < 0.0001.

These results evidenced that in ALT and MERT cows, the accumulated THI of Summer did not induce an accumulated endogenous release of cortisol, whereas HF cows showed a marked increase in response to Summer THI.

Plasma HSP concentrations

Plasma HSP concentrations are presented in Table 2, showing the median values and the 0–25% and 75–100% quartiles. Plasma HSP60 concentrations were affected by Season (p < 0.0001) and by Breed (p < 0.05), as the difference between Summer and Winter was more significant in ALT. The Season*Breed effect was not significant (p = 0.43).

Table 2.

Season (Summer versus Winter) effects on plasma concentrations of HSP60, HSP70 and HSP90 in Alentejana (n = 34), Mertolenga (n = 55) and Holstein-Frisian (n = 66). Values are presented as Median, 0–25% and 75–100% quartiles. For columns within rows with different letters, p < 0.05.

Variable	Summer				Winter
	ALT	MERT		HF	ALT	MERT	HF
HSP60	0.9 bc (0.2–3.0.2.0)	2.5 c (1.4–4.8)		2.6 c (1.4–4.2)	7.4 a (1.8–46.8)	4.5 b (2.1–12.7)	2.6 b (1.2–8.2)
HSP70	6.14 ab (4.2–6.9)	8.5 a (3.6–11.4)		4.5 b (2.9–4.5)	1.8 c (1.4–2.6)	3.0 c (2.0–4.5.0.5)	4.7 b (2.5–7.6)
HSP90	76.5 b (61.4–91.9)	83.6 b (40.6–127.2.6.2)		80.1 b (50.7–113.1.7.1)	145.2 a (87.9–163.1.9.1)	98.8 b (68.23–138.4)	97.1 b (78.17–122.2)

Plasma HSP70 concentrations were affected by Season (p < 0.0001) and by Breed (p < 0.05), as the difference between Summer and Winter was more significant in ALT and MERT. The Season*Breed effect was significant (p < 0.0001), as HF was the only breed without a Summer increment [S – 4.5 (2.9–4.5) ng/mL vs. W – 4.7 (2.5–7.6) ng/mL].

Plasma HSP90 concentrations were affected by Season (p = 0.0001) and by Season*Breed (p < 0.05), as ALT was the only breed with a significant decrease in Summer [S – 76.5 (61.4–91.9) ng/mL vs. W − 145.2 (87.9–163.1.9.1) ng/mL]. The Breed effect was not significant (p = 0.31).

The above results evidenced that Summer THI significantly affected the HSP plasma concentrations. However, this effect was specific for each HSP and breed. Concentrations of HSP60 had a significant decrease in Summer in the three breeds, more evident in ALT cows. Concentrations of HSP70 had a significant Summer increase in ALT and MERT cows, but not in HF cows. Concentrations of HSP90 had a significant decrease only in ALT cows.

Relationship effects of Breed, THI60 and cortisol on HSP plasma concentrations

The results of the final models for interactions (PROC GLM and GlimmixnegBin) are presented in Supplementary Table 3.

Plasma HSP60 concentrations were affected by Breed (p < 0.05), THI60 (p < 0.001), Cortisol (p = 0.01) and by Breed*THI60*Cortisol (p < 0.05). This latter interaction only showed a significant increase in HSP60 concentrations in ALT cows, 2% for each THI60*Cortisol unit increment (p = 0.01) (Fig. 5A), but had no effect in MERT and HF cows (p = 0.95 and p = 0.5, respectively). However, Cortisol and THI60, independently, decreased HSP60 plasma concentrations in ALT cows (p = 0.01 and p = 0.001, respectively). In MERT and HF, plasma HSP60 was affected independently by THI60 and Cortisol. Plasma HSP60 concentrations were decreased (p < 0.0001) by THI60 in MERT and HF cows (5% and 1% per THI60 increment, respectively) (Fig. 5B). Plasma HSP60 concentrations were also decreased (p < 0.001) by Cortisol in MERT cows (9% per cortisol increment), but were increased (p < 0.001) (2% per cortisol increment) in HF cows (Fig. 5C). However, this Cortisol effect was not significantly different between MERT and HF breeds (p = 0.63).

Fig. 5.

Graphical illustration of models for plasma HSP60 concentrations: (A) interaction THI60*Cortisol in Alentejana (ALT, n = 34) cows; (B) independent effect of THI60 in Mertolenga (MERT, n = 55) and Holstein-Friesian (HF, n = 66) cows; (C) independent effect of Cortisol in MERT and HF cows. Values are in natural Log.

These results evidenced a relationship between THI60, HCC and HSP60 concentrations. However, breed differences emerge in this relationship. Although an independent increment in both THI60 and HCC decreased HSP60 concentrations in the three breeds, in ALT cows, a simultaneous increment in THI60 and HCC originated an increase in HSP60 concentrations, revealing an interacting effect of THI60 and HCC on HSP60 concentrations.

Plasma HSP70 concentrations were affected by Breed (p < 0.0001), THI60 (p < 0.0001) and THI60*Breed (p = 0.0001), but not by Cortisol (p = 0.63), THI60*Cortisol (p = 0.17) or Breed*THI60*Cortisol (p = 0.6). Therefore, these non-significant effects were removed from the model. The THI60*Breed effect is illustrated in Fig. 6. The THI60 increased (p < 0.0001) HSP70 plasma concentrations in ALT and MERT cows (6% and 5% per THI60 increment, respectively), however had no effect in HF cows (p = 0.63) (Fig. 6).

Fig. 6.

Linear regression for the effect of Breed*THI60 in plasma HSP70 concentrations in Alentejana (ALT, n = 34), Mertolenga (MERT, n = 55) and Holstein-Friesian (HF, n = 66) breeds. Values are in natural Log.

These results evidenced that, in contrast to HSP60, plasma concentrations of HSP70, although affected by THI60 in a breed different pattern, had no relationship with HCC.

Plasma HSP90 concentrations were affected by Breed (p < 0.05), Breed*THI60 (p < 0.01), Breed*Cortisol (p < 0.05), and Breed*THI60*Cortisol (p = 0.01). This latter interaction (Breed*THI60*Cortisol) significantly (p < 0.05) decreased HSP90 plasma concentrations in MERT cows (by 1% per THI60*Cortisol increment) (Fig. 7A), but had no effect in ALT and HF cows (p = 0.62 and p = 0.1, respectively). In these breeds, THI60 and Cortisol showed independent effects in HSP plasma concentrations. In ALT, THI60 decreased (p < 0.0001) HSP90 plasma concentrations by 4% per THI60 unit increment, whereas in HF cows only a decrease tendency (p = 0.07) was observed, resulting in a significant effect (p < 0.05) between the two breeds (Fig. 7B). Cortisol significantly decreased (p < 0.01) HSP90 plasma concentrations in ALT and HF cows, equally by 2% per HCC unit increment (Fig. 7C). Independently, THI60 and Cortisol.

Fig. 7.

Graphical illustration of models for the interaction THI60*Cortisol on plasma HSP90 concentrations. (A) Effect Breed*THI60*Cortisol in Mertolenga (MERT, n = 55); (B) THI60 effect in Alentejana (ALT, n = 34) and Holstein-Friesian (HF, n = 66); (C) Cortisol effect in ALT and HF. Values are in natural Log.

increased (p < 0.05) plasma HSP90 concentrations in MERT cows.

These results again evidenced breed differences in the relationship between THI60, HCC and HSP90 concentrations. In ALT and HF cows, HSP90 concentrations were decreased by the independent increase in THI60 or HCC. In MERT cows, the independent effect of THI60 and HCC, increased HSP90 concentrations, but the simultaneous increase of those parameters, decreased plasma HSP concentrations.

Discussion

The present study evaluated the relationship between THI, cortisol and HSP release under thermoneutrality (Winter) or HS (Summer), and the physiologic mechanisms of response to HS in heat-tolerant and heat-sensitive cattle breeds. Results evidenced that the breeds ALT, MERT and HF developed specific strategies to cope with HS, including the physiologic response of body temperatures, metabolic parameters, cortisol release and HSP production, as well as different relationships between the THI stimuli and cortisol release on HSP response.

Although all cows were submitted to similar environmental THI60, the native ALT and MERT cows were affected by higher cow-side THI at sampling during Summer, compared to the HF cows. This was expected, as all cows were located in the same region, but HF cows were housed and received permanent shade and cooling measures, whereas the native cows were under natural conditions. Therefore, ALT and MERT cows were expected to experience higher levels of HS than HF cows.

The differences in housing conditions between the native and HF breeds, potentially affected the response of body temperatures to HS, as besides the presence of cooling systems, the protection from solar radiation is an important component on the impact of HS^60,61. This effect of housing in HF cows may have been responsible for the absence of increase in external (ocular) temperature in Summer, compared to Winter. However, HF cows had increased internal (vaginal and rectal) temperature in Summer, denoting the accumulated effect of THI over the metabolic endogenous heat production due to milk yield. This increase in internal body temperatures observed in Summer, may be assumed as HS (THI) induced, as measurements obtained in both seasons were at the same postpartum time-point, thus reflecting a similar metabolic scenario, and in similar management conditions.

Contrary to HF cows, ALT and MERT cows showed an increase in ocular temperature in Summer, potentially reflecting the THI effect of pasture conditions. This effect was most noticeable in MERT cows (4.5 °C increase in Summer), and efficiently enabled heat loss by vasodilatation of superficial vessels⁶², which resulted in only a non-significant 0.3 °C increase in vaginal (and rectal) temperature in Summer, compared to Winter. Although less efficient in the ocular heat loss, ALT cows showed no significant internal temperature increment in Summer, which was related to the higher capacity in decreasing internal heat production through the metabolic rate (see below).

The metabolic response to high THI was different in the three breeds. Serum concentrations of BHB were low and had no variation between seasons in ALT and MERT cows. This was expected as, by 40 DPP, suckling does not configure a significant metabolic insult to the cows. By contrary, in HF cows, serum concentrations of BHB were much higher than in ALT and MERT cows, dealing with the metabolic challenge of lactation. In HF cows, the higher concentrations in Winter may reflect the higher milk production in Winter, as HS negative effects in milk yield are well documented⁶³. In fact, the high metabolic rate of HF cows is necessary for the production of milk, and the decrease of the metabolic rate to face HS effects also lead to the decrease in milk production.

In MERT and HF cows, serum concentrations of thyroid hormones were not different between seasons, exhibiting a non-significantly decreased in total T3 concentrations by 14% and 8%, respectively. By contrary, ALT cows had decreased serum concentrations of thyroid hormones in Summer, showing a significant decrease of 33% in total T3 concentrations. Thyroid hormones are the main players in metabolic adjustment⁶⁴, and it is well documented that an increment in environmental temperature leads to a decrease in T3 and T4 blood concentrations^65,66, which attains a notable efficiency in Bos indicus heat adaptability⁶⁶. Therefore, although evolving in the same edaphoclimatic scenario and with a presumed common ancestor, ALT and MERT, as heat-tolerant breeds, developed different physiologic strategies to mitigate HS. This breed’s strategical differences emerge as potential relevant phenotypes for the genetic selection towards resilience to climatic change.

Confirming our previous studies⁵, HCC were not different between seasons in ALT and MERT cows, indicating that cortisol is not a reliable HS maker in these breeds. In contrast, HF cows showed a significant increase in HCC in Summer compared to Winter, reinforcing that this breed is more sensitive to HS due to their higher metabolic rate^67–69.

Plasma concentrations of HSP evidenced specific trends between seasons and breeds. In the three breeds, concentrations of HSP60 were significantly lower in Summer than in Winter, but the season effect was more noticeable in ALT and MERT than in HF cows. Plasma concentrations of HSP60 increase following an injury and loss of cellular integrity⁷⁰, as this protein function as a potent chemoattractant for neutrophils, being released by injured cells³⁴. The way HSP60 reaches the bloodstream is not yet fully understood^33,34. One hypothesis put forward, is that peripheral vasodilation in response to HS leads to poor internal blood perfusion and to damage in the intestinal barrier causing leakage of some HSP to the bloodstream⁴³. Results from the present study do not support this theory, at least in MERT cows. These cows experienced a significant increase in body external temperature in Summer, and therefore an expected inherent superficial blood vasodilation and redistribution of blood volume⁶², but had lower plasma HSP60 concentrations in Summer than in Winter. Results are in accordance with studies that reported decreased endometrial HSP60 expression in healthy cows during Summer⁴⁵, and that found a decreased HSP60 expression in the afternoon, compared to early morning in Simmental cattle⁴⁴.

Contrary to HSP60, plasma concentrations of HSP70 were higher in Summer than in Winter in ALT and MERT, but not in HF cows. An elevation of HSP70 expression protects the cell from thermal injuries⁷¹. Early studies on the response to HS^72,73, showed that long term heat-acclimated rats showed 200% more HSP70 expression than rats not heat-acclimated, but this was only associated with environmental temperatures and not with the rat’s body temperature. This was also observed in Angus steers, where plasma HSP70 concentrations were related with environmental temperatures but not with body temperature⁷⁴. Other studies in calves^43,74 found increased HSP70 expression in blood mononuclear cells and in hair follicles associated with increased environmental THI. The above results indicate that HSP70 expression increases cell thermal tolerance, not related with body temperature increment. This concept was well illustrated in ALT cows, which displayed high HSP70 plasma concentrations in Summer, without a significant increment in internal body temperature. Therefore, HSP70 emerge as a potential cattle HS indicator^43,44,74, but the breed specific value must be evaluated.

The effects of HS on cattle HSP90 plasma concentrations are conflicting, either revealing a positive blood expression⁴⁴ or a negative endometrial expression⁴⁵. In the present study, the season effect was only significant in ALT, indicating a decrease in plasma concentrations in Summer.

The analysis of the interrelationships between Breed, THI60 and HCC on HSP release, again evidenced breed specific effects. The THI60 and HCC variables are well related, as the type and amount of hair used for cortisol extraction roughly entails the accumulated content of the previous 60 days, and the mean THI60 calculated at the end of Summer and Winter, reflects the accumulated effect of HS or thermoneutrality. In ALT cows, plasma HSP60 concentrations were affected by the relationship between THI60 and HCC. An increase in HSP60 concentrations were dependent on the simultaneous increase in THI60 and HCC, whereas the independent increase in THI60 or HCC decreased plasma HSP60 concentrations. By contrast, in MERT and HF cows there was no interaction between THI60 and HCC on plasma HSP60 concentrations, as their effect was independent: Increased THI60 decreased HSP60 concentrations in both breeds, whereas increased HCC decreased HSP60 concentrations in MERT, but increased concentrations in HF. In HF cows under HS, this increased HSP60 concentrations induced by cortisol, may result from a higher incidence of postpartum disease⁷⁵, which could result in the release of HSP60 from damaged endometrial cells, or from new synthesis due to immune modulation^34,70 or wound tissue repair³⁶.

Plasma HSP70 concentrations were affected by THI60, but not by HCC. Also, THI60 increased HSP70 concentrations in ALT and MERT, but not in HF cows. These results corroborate the increased HSP70 expression in response to increased ambient temperatures, but evidence that this is independent of HCC. This supports the hypothesis that environmental temperature stimuli received by subcutaneous thermoreceptors trigger the release of HSP70, which in turn lead the adjustment in animal thermoregulation⁷⁶.

Interestingly, plasma HSP90 concentrations behaved similarly to HSP60, exhibiting an interacting effect between breed, THI60 and HCC. In MERT cows, a decrease of HSP90 plasma concentrations was only observed in the presence of a concomitant increase in THI60 and HCC, whereas in ALT and HF cows, the effects of THI60 and HCC were independent, both decreasing HSP90 concentrations.

Conclusion

The present study highlights the existence of cattle breed specific phenotypes of response to HS, including the type and magnitude of the physiologic parameters put forward in the response, as well as the relationship between them. Breed specific heat loss strategies, considering body temperatures and metabolic parameters, were present even in close related native breeds managed under the same climatic stimuli. In these heat-tolerant breeds, MERT cows attained heat loss through a body external (ocular) temperature increase and associated peripheric vasodilation, and in a less degree in lowering the internal metabolic rate, whereas ALT cows displayed a high efficiency in decreasing the internal metabolic rate, through modulation of total T4 and T3 secretion. In contrast, HF cows under similar climatic stimuli had lower heat loss efficiency through the above mechanisms, turning it a more heat-sensitive breed. The breed differences in response to HS is also highlighted by the absence of effect on HCC in ALT and MERT cows, contrary to the stimulatory effect observed in HF cows.

Plasma HSP concentrations were affected by a breed specific interaction between THI60 and HCC. Plasma HSP60 concentrations were decreased by the independent increase in THI60 and HCC in the three breeds, but in ALT cows the simultaneous increase had the opposite effect. This interaction prompts plasma HSP60 concentrations as a potential marker of HS in the ALT breed. Plasma HSP90 concentrations were decreased by the independent increase in THI60 and HCC in ALT and HF cows, but in MERT cows this decrease in HSP90 concentrations was only accomplished when associated to a simultaneous increase in THI60 and HCC. This interaction prompts plasma HSP90 concentrations as a potential marker of HS in the MERT breed. Plasma HSP70 concentrations were increased by increased THI60 in ALT and MERT cows, but not in HF cows, but this was independent of cortisol release, rendering HSP70 less useful as a marker of HS.

Altogether, the results of the present study evidence that HS response strategies evolved differently in cattle breeds, even in close related breeds under the same environment, and phenotypes of heat-tolerance must be evaluated for each breed. Further studies are needed to characterize the genomics of the adaptive mechanisms governing HS tolerance, which can then be included in selection programs, to ensure the resilience of cattle under a scenario of global warming.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Abbreviations

ALT

Alentejana breed

BCS

Body Condition Score

BHB

Beta-hydroxybutyrate

DPP

Days Postpartum

HCC

Hair Cortisol Concentrations

HF

Holstein-Friesian breed

HS

Heat Stress

HSP60

Heat Shock Protein 60 kDa

HSP70

Heat Shock Protein 70 kDa

HSP90

Heat Shock Protein 90 kDa

IPMA

Portuguese Institute for Sea and Atmosphere, I. P

MERT

Mertolenga breed

OcularMax

Maximal Ocular Temperature

OcularMean

Mean Ocular Temperature

OcularMin

Minimal Ocular Temperature

S

Summer

THI60

Temperature-Humidity-Index of the previous 60 days

T3

Triiodothyronine

T4

Thyroxine

W

Winter

Author contributions

Luís G Capela – Conceptualization, Methodology, Writing, Review and Editing; Inês C Leites – Methodology, sampling; Luísa M Mateus – Methodology, hormone assays, Review; Ricardo P Romão – Methodology, sampling, Review and Editing; Alfredo Pereira – Funding acquisition; Rosa MLN Pereira – Conceptualization, Review and Editing, Funding acquisition; Luís Lopes-da-Costa - Conceptualization, Methodology, Writing, Review and Editing, Funding acquisition.

Funding

This study was funded by projects FCT PTDC/CVT-CVT/6932/2020, FCT UIDB/00276/2020 (CIISA), UIDB/05183/2020 (MED), LA/P/0059/2020 (AL4Animals AL4AnimalS) and PDR2020-101- 03112; PhD grant SFRH/BD/148804/2019 (Capela L).

Data availability

The datasets generated during the current study are not publicly available because they are still being analyzed for further work on the first author’s PhD, being confidential until the end of PhD. Data is available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval for animal use

The experiments were conducted in compliance with the Portuguese legislation for the use of animals for experimental purposes (Decreto-Lei nº129/92 and Portaria nº1005/92, DR nº245, série I-B, 4930-42) and with the European Union legislation (Directive 2010/63/UE). The research protocol was approved by the Institutional Animal Care and Use Committee (Reference CEIE nº008/2022). Animal technical procedures were performed by licensed veterinarians, with informed consent from the management of the respective associations. All procedures were performed in accordance with the ARRIVE guidelines and the institutional guidelines for animal research. The reporting of all experimental methods follows in accordance with the ARRIVE guidelines.