In the summer of 2024, South Korea recorded its highest average temperature (25.6℃) since nationwide weather observation began in 1973 (KTimes, 2024). Additionally, the national average number of tropical nights reached 20.2 days, more than three times the normal level (KTimes, 2024). The highest temperature since meteorological observations began in 1907 (over 40℃) was recorded relatively recently, in 2018 (Lee et al., 2020).

Holstein-Friesian breed is the most common dairy cattle breed worldwide, but it is highly susceptible to heat, which leads to reduced milk production and decreased reproductive performance under heat stress (HS) conditions, ultimately lowering farmers’ economic income (Wolfenson and Roth, 2018; Tao et al., 2019). In the case of Holstein cows, milk production decreases by 30-40% under HS conditions compared to normal conditions (West, 1999; Baumgard et al., 2012; Pragna et al., 2017). Particularly, the conception rate of Holstein cows is 39.4 % at a temperature-humidity index (THI) below 72 (non-HS conditions), but it decreases to 31.6 % when THI exceeds 78.0 (intermediate HS) (Liu et al., 2018). In Florida, conception rates for lactating cows during the summer are particularly low, at only 13.5 % (Putney et al., 1989). This reduction in conception rates may be attributed to hormonal changes during the summer, such as decreases in E2 and LH levels (Wolfenson and Roth, 2018; Roth, 2020). Nebel et al. (1997) reported that the number of mounts per estrus also decreased by nearly half in the summer compared to the winter (Nebel et al., 1997; Hansen, 2019).

Jersey breed, the second most common dairy breed worldwide, is relatively known for its heat tolerance (Wiggans, 2000; CDCB, 2020). In South Korea, Jersey cattle embryos have been imported from Canada and the United States since 2012, with the aim of improving milk fat and protein content in dairy products, as well as addressing climate change to promote the sustainability of dairy farming (Lee et al., 2023). Notably, under HS conditions with a THI exceeding 68, one study reported that the average daily milk yield of Holstein cows decreased from 35.6 kg to 34.2 kg, whereas Jersey cows showed an increase in milk yield from 25.9 kg to 26.6 kg (Smith et al., 2013). Similarly, in Hungary, it was observed that milk production in Holstein cows decreased in summer compared to spring, while Jersey cows showed similar milk production levels between the two seasons (Jurkovich et al., 2023). Moreover, in terms of reproductive performance, the conception rate of Jersey cattle was inherently higher than that of Holstein cattle (Norman et al., 2009; Norman et al., 2020; Lee et al., 2023). Although the conception rates of both breeds declined sharply when the THI exceeded 75, the decrease was more pronounced in Holstein cows compared to Jersey cows (Lee et al., 2023).

MicroRNAs (miRNAs) are small endogenous RNA molecules, typically 18 to 22 nucleotides in length, that function as non-coding RNAs. They play a key role in RNA silencing and are involved in various developmental and pathological processes in animals by regulating numerous protein-coding transcripts (Kim and Nam, 2006; Kim et al., 2009; Ha and Kim, 2014). Recent studies have shown that miRNAs are exporting to the extracellular environment through microvesicles such as exosomes and circulate in the bloodstream, where they can be identified and quantified in biological fluids like plasma or serum (Valadi et al., 2007; Hunter et al., 2008; Turchinovich et al., 2011; Witwer, 2015). Circulating miRNAs are actively being studied and applied as biomarkers for specific traits, including various human cancers (Roth et al., 2010; Reid et al., 2011; Sørensen et al., 2014) and pregnancy diagnosis (Ioannidis and Donadeu, 2017), estrus (Ioannidis and Donadeu, 2016) and aging (Ioannidis et al., 2018) in dairy cattle.

In previous study, we identified 11 potential circulating miRNA biomarkers related to HS in lactating Holstein cows (Lee et al., 2020). This study compares miRNA profiles of lactating Jersey cows, known for superior heat tolerance, to those of Holstein cows under HS conditions.

All animal experimental protocols and procedures were reviewed and approved by the Animal Care and Ethics Committee of the National Institute of Animal Science, South Korea (approval number: NIAS-068).

Experimental animals

Before the start of the experiments, veterinarians and farmers conducted regular health assessments of the cows at the Dairy Research Center in National Institute of Animal Science, selecting four lactating Holstein-Friesian cows and two Jersey cows that were confirmed to be healthy and disease-free. Holstein cows have an average of 239.25 ± 66.95 days in milk (mean ± standard deviation) and for Jersey cows, it is 206 ± 41.01 days. Individual cow details, including age, parity, and calving date can be found in Table 1. The cows were provided with a total mixed ration (TMR), prepared as previously described (Lee et al., 2024). Briefly, the TMR was formulated following the National Research Council guidelines (National Research Council, 2001). For lactating cows, the ration was mixed twice daily and provided at 09:00 and 15:00. All animals had unrestricted access to fresh water and minerals. Lactating Holstein and Jersey cows were fed the same diet in a shared area. Following the experiment, the cows were housed for further research.

Table 1. Information on each cow used in this experiment.

Individual ID	Breed	Birth date	Parity	Previous calving date	Milking days (as of 15th Aug)	Expected calving date
08120	Holstein	8th Oct 2008	4	27th Nov 2017	262	30th Mar 2019
13003	Holstein	4th Jan 2013	1	5th Dec 2017	254	12th Feb 2019
14106	Holstein	26th Nov 2014	1	26th Mar 2018	143	1st Apr 2019
15044	Holstein	21th May 2015	1	22th Oct 2017	298	20th Feb 2019
J1404	Jersey	27th Feb 2014	3	20th Feb 2018	177	29th Jan 2019
J1405	Jersey	20th Mar 2014	2	23th Dec 2017	235	29th Dec 2018

HS indicators

1) Temperature-humidity index (THI)

To monitor the THI in the barn, a THI measuring device (Testo-174d, 5720500, Germany) was installed at three different locations within the barn. This device automatically recorded temperature and relative humidity, with measurements taken 12 times daily at 2-hour intervals. The THI was calculated using the following formula (Mader et al., 2006; Dikmen and Hansen, 2009). THI = 0.8 × temperature (℃) + [(relative humidity (%)/100) × (temperature (℃) - 14.4)] + 46.4.

2) Rectal temperature

Rectal temperature was taken concurrently with blood collection. To obtain the rectal temperature, any feces were cleared from the cow’s rectum. Using a rectal thermometer (POLYGREEN Co. Ltd., Germany), temperatures were manually recorded at 14:00 under both HS (THI: 86.29) and non-HS (THI: 60.87) conditions. For precision, each measurement was repeated three times per cow, with the thermometer inserted at a depth of over 15 cm into the rectum.

3) Milk yield measurement and body weight

Milk production and body weight were automatically recorded by a milking robot (Lely, Astronaut, Netherlands) during the milking process from June to October. The average daily gain (ADG) was calculated using the body weight data. To determine the relative average milk yield (RAMY), each cow’s milk production from June to October was compared to the average milk yield in May.

Statistical analysis

The quantitative data are presented as mean ± standard error of the mean (s.e.m). Statistical analyses were conducted using SPSS 26.0 statistical software (SPSS, Chicago, IL, USA). To account for unequal sample sizes and variances, Welch’s t-test was used for pairwise comparisons, and Welch’s ANOVA was applied for comparisons across multiple groups. A p-value < 0.05 was considered statistically significant, unless stated otherwise.

Blood collection

All cows were housed in a barn designed for natural ventilation. Daily THI values were calculated based on recorded air temperature and humidity, and the sampling date was selected when the daily minimum THI remained above 72 and the daily maximum THI stayed below 72 for over four weeks (Fig. 1). Whole blood samples were taken from the jugular vein of the same cows (Hol = 4, Jer = 2) during two distinct environmental seasons (summer and autumn) using PAXgene Blood RNA tubes (2.5 mL per cow; Qiagen, 762165, California, USA). The PAXgene Blood RNA tubes were stored at -80℃ until miRNA extraction.

Figure 1.Temperature-humidity index measured on dairy barn. The black line represents daily THI maximum and the grey line represents daily THI minimum. The tilted boxes represent THI on the day of sampling (modified from Lee et al. (2020)).

MiRNA-sequencing experiment and statistical analysis

MiRNAs were extracted from whole blood using the PAXgene Blood MicroRNA Kit (Qiagen) following the manufacturer’s protocol. The concentrations of miRNAs were measured with a NanoDrop device (Optigen NANO Q, South Korea). miRNA integrity was evaluated using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA), ensuring an RNA integrity number of 7 or higher. Library preparation involved adapter ligation, reverse transcription, PCR amplification, and pooled gel purification, performed with the Truseq Small RNA Library Prep Kit (Illumina, San Diego, USA). The library was then loaded into the Illumina Hiseq2000 sequencer, which contained millions of unique clusters. Raw sequencing reads of circulating miRNAs from all samples were pre-processed and analysed using miRDeep2 software. Adapter sequences added during the small RNA library construction were removed using Cutadapt v.1.9.1, and reads of at least 18 bp were collected to form clusters for increased accuracy. The processed and clustered reads were aligned with the Bos Taurus reference genome and further mapped to precursor and mature miRNAs from miRBase v21. miRDeep2 software was used to identify both known and novel miRNAs and estimate their expression levels. Differential gene expression analysis was conducted using the Limma voom v3.34.9 R package (Grimson et al., 2007; Robinson et al., 2010; Vejnar and Zdobnov, 2012). Prior to the analysis, genes with raw read counts were removed using the ‘filterByExpr’ function in the edgeR R package (Robinson et al., 2010). Then, TMM normalization was applied to standardize the size of each library using the ‘calcNormFactors’ function in edgeR. Subsequently, the ‘voom’ function in Limma was utilized to convert read counts into a logarithmic (base 2) scale for linear modeling. Differentially expressed (DE) genes between the two groups were identified using empirical Bayes and moderated t-test. A p-value < 0.01 and a logarithmic fold change (FC) of |logFC| > 2 were set as the criteria for significance.

Bioinformatics analysis

To predict target genes for DE miRNAs, miRmap (v1.2.0, mirmap.ezlab.org) databases were utilized (Vejnar and Zdobnov, 2012). Target genes were selected based on a miRmap Score of ≥ 80, provided by the respective programs, to ensure higher accuracy in the analysis. Gene Ontology (GO) enrichment analysis was conducted using the PANTHER Classification System (v.19.0) (Mi et al., 2013) to determine the functional significance of the gene set. Additionally, the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed by submitting the target gene list to DAVID Bioinformatics Resources (v2024q2) (Dennis et al., 2003) to identify enriched signalling pathways associated with these target genes.

Estimation of THI and blood collection

To estimate THI, we recorded the ambient temperature and relative humidity inside the barn daily (Fig. 1). Starting from the first week of July, the minimum THI exceeded 72, occasionally surpassing 80, indicating moderate to severe HS conditions (THI > 78) lasted for over one month. Both maximum and minimum THI values reached their highest around mid-August and then gradually declined; however, mild to moderate HS (THI 72-78) continued until the end of September. Whole blood samples were collected at 14:00 from both lactating Holstein and Jersey cows during the summer (HS) and autumn (non-HS). On sampling days, the THI ranged from 79.10 to 87.73 (14:00; 86.29) for HS and from 47.30 to 64.85 (14:00; 60.87) for non-HS. The minimum THI remained above 72 (the threshold for HS) for 36 consecutive days until HS sampling, while the maximum THI stayed below 72 for 28 days prior to non-HS sampling.

Effects of HS on physiological changes

We assessed physiological indicators of HS, including rectal temperature, milk yield, and ADG in lactating Holstein and Jersey cows. Rectal temperatures were recorded concurrently with blood collection. Under non-HS conditions, no significant differences were observed between Holstein (38.40℃ ± 0.07) and Jersey cows (37.80℃ ± 0.50). In contrast, during HS conditions, the rectal temperatures of Holstein cows (40.15℃ ± 0.17) were significantly higher than those of Jersey cows (38.95℃ ± 0.05, p = 0.0022, Fig. 2A).

Figure 2.Physiological indicators of heat stress recorded under environmental conditions in both lactating Holstein and Jersey cows (modified from Lee et al. (2020) for Holstein data). (A) Rectal temperature recorded on the day of sampling. (B) Relative average milk yield (RAMY) to May. (C) Average daily gain (ADG). Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001.

To assess the impact of HS on milk yield, we calculated the RAMY, defined as the ratio of the total milk yield for each month during the study period to the yield in May. Both Holstein and Jersey cows showed a gradual decline in RAMY, with a particularly significant drop observed in Holstein cows during the extreme heat of August. However, no differences in RAMY were observed across all months in Jersey cows. Recovery of RAMY was seen in September (Fig. 2B), suggesting that HS may have a more pronounced effect on the milk production of Holstein cows.

We also evaluated the influence of HS on growth performance by analysing changes in ADG for both Holstein and Jersey cows throughout the experimental period. As illustrated in Fig. 2C, ADG generally increased in both Holstein and Jersey cows, except in July, even though overall ADG during the summer was significantly lower.

Identification of DE miRNAs of HS cows

We performed small RNA sequencing on miRNAs isolated from whole blood to identify DE miRNAs between heat-stressed and non-heat-stressed Holstein and Jersey cows. In the lactating Holstein cows, 28 miRNAs showed significant differential expression (≥ 2-FC in the expression compared to non-HS controls; p < 0.05), including 10 upregulated and 18 downregulated miRNAs. In the Jersey cows, 38 DE miRNAs were identified (23 upregulated, 15 downregulated; ≥ 2-FC and p < 0.05) (Table 2).

Table 2. Differentially expressed microRNAs in Holstein and Jersey under heat stressed conditions (for the expression profile of Holstein cows, refer to the previous study (Lee et al., 2020).

Breed	Mature microRNA	RNA sequencing fold change (heat stress/non-heat stress) ≥ 2 (p < 0.05)
Holstein	FC value: up-regulation
	Bta-miR-19a	3.49
	Bta-miR-19b	3.34
	Bta-miR-20b	2.40
	Bta-miR-29d-3p	3.77
	Bta-miR-106a	2.25
	Bta-miR-378d	2.39
	Bta-miR-497	4.24
	Bta-miR-502a	2.46
	Bta-miR-2285ad	2.44
	Bta-miR-2285o	5.98
	FC value: down-regulation
	Bta-miR-30a-5p	-4.44
	Bta-miR-146b	-2.08
	Bta-miR-296-3p	-2.11
	Bta-miR-1246	-9.49
	Bta-miR-2284a	-17.52
	Bta-miR-2284aa	-2.31
	Bta-miR-2284ab	-2.40
	Bta-miR-2284b	-3.61
	Bta-miR-2284h-5p	-8.05
	Bta-miR-2284k	-7.77
	Bta-miR-2284r	-2.82
	Bta-miR-2284v	-2.84
	Bta-miR-2284w	-2.69
	Bta-miR-2284x	-10.37
	Bta-miR-2284y	-9.95
	Bta-miR-2284z	-2.19
	Bta-miR-2397-5p	-2.08
	Bta-miR-2457	-2.15
Jersey	FC value: up-regulation
	Bta-let-7a-5p	2.86
	Bta-let-7c	2.83
	Bta-let-7e	2.55
	Bta-let-7f	3.37
	Bta-let-7g	3.09
	Bta-let-7i	2.02
	Bta-miR-141	3.54
	Bta-miR-144	2.52
	Bta-miR-148a	2.25
	Bta-miR-1584-3p	4.40
	Bta-miR-196b	7.06
	Bta-miR-20b	3.63
	Bta-miR-2336	2.44
	Bta-miR-2419-3p	8.48
	Bta-miR-2450a	3.55
	Bta-miR-26b	2.03
	Bta-miR-3432a	3.77
	Bta-miR-411a	3.74
Jersey	Bta-miR-429	11.04
	Bta-miR-454	2.77
	Bta-miR-7859	15.11
	Bta-miR-9-5p	3.06
	Bta-miR-98	2.87
	FC value: down-regulation
	Bta-miR-1246	-4.43
	Bta-miR-133a	-4.06
	Bta-miR-1343-3p	-2.08
	Bta-miR-151-3p	-2.49
	Bta-miR-1843	-3.90
	Bta-miR-191	-2.19
	Bta-miR-204	-2.08
	Bta-miR-2284x	-2.11
	Bta-miR-2284y	-2.06
	Bta-miR-2332	-5.20
	Bta-miR-2389	-3.11
	Bta-miR-2394	-4.08
	Bta-miR-2419-5p	-2.15
	Bta-miR-2453	-8.03
	Bta-miR-324	-3.35

Additionally, we examined the profiles of 11 candidate DE miRNAs identified in a previous study conducted on Holstein cows to validate their expression in Jersey cows (Lee et al., 2020). The expression patterns (up / downregulation) in Jersey cows were consistent with those observed in Holstein cows; however, the expression levels (|HS/NJS|) in Jersey cows were lower, and no significant differences were found 9 of the DE miRNAs, except for bta-miR-2284x and bta-miR-2284y (p > 0.05) (Table 3).

Table 3. Comparison of expression profiles of candidate circulating microRNA biomarkers related to heat stress (identified in a previous study; Lee et al., 2020) in lactating Holstein and Jersey cows.

miRNA	Holstein			Jersey
	Fold change (heat stress/non-heat stress)	p-value		Fold change (heat stress/non-heat stress)	p-value
Bta-miR-19a	3.49	< 0.05		1.80	0.26
Bta-miR-19b	3.34	< 0.05		2.07	0.11
Bta-miR-30a-5p	-4.44	< 0.05		-1.92	0.07
Bta-miR-2284a	-17.52	< 0.05		-1.85	0.74
Bta-miR-2284b	-3.61	< 0.05		-2.74	0.08
Bta-miR-2284h-5p	-8.05	< 0.05		-2.60	0.06
Bta-miR-2284k	-7.77	< 0.05		-1.26	0.83
Bta-miR-2284v	-2.84	< 0.05		-1.92	0.10
Bta-miR-2284w	-2.69	< 0.05		-1.56	0.20
Bta-miR-2284x	-10.37	< 0.05		-2.11	0.04
Bta-miR-2284y	-9.95	< 0.05		-2.06	0.04

Putative target gene and signalling pathway analysis

In a previous study on lactating Holstein cows, we identified 2,798 potential target genes for 28 DE miRNAs using miRmap and TargetScan. In this study, we identified 6,108 potential target genes for 38 DE miRNAs in lactating Jersey cows using miRmap. Additionally, we investigated the potential target genes of 4 common DE miRNAs using miRmap and identified 500 genes. Gene set enrichment analysis (GSEA) was performed using DAVID and PANTHER. The GSEA results indicated that these predicted target genes were linked to the corpus luteum function and progesterone biosynthesis and numerous genes also involved in the immune responses (Table 4). Additionally, KEGG pathway analysis revealed 9 statistically significant pathways (Table 5). Notably, pathways such as the metabolic pathway, Fanconi anemia pathway and FoxO signaling pathway were found to be closely associated with the HS response.

Table 4. Predicted target genes related to heat stress responses of microRNAs differentially expressed in Holstein and Jersey cows.

DE miRNA	Direction of gene regulation	Target genes	Responses
Bta-miR-20b	Down	XCL1, CCL1	Improvement of corpus luteum function
Bta-miR-1246	Up	StAR	Progesterone biosynthesis
Bta-miR-2284x	Up	Numerous genes involved in the immune responses (data are not shown)
Bta-miR-2284y	Up	Numerous genes involved in the immune responses (data are not shown)

DE miRNAs, differentially expressed miRNAs..

Table 5. KEGG pathways enriched for targets of differentially expressed miRNAs in Holstein and Jersey cows.

Term	Number of genes	p-value
bta01100: Metabolic pathways	51	0.0055
bta04144: Endocytosis	12	0.019
bta03460: Fanconi anemia pathway	5	0.025
bta04068: FoxO signaling pathway	8	0.026
bta00600: Sphingolipid metabolism	5	0.036
bta05210: Colorectal cancer	6	0.048
bta05226: Gastric cancer	8	0.049
bta05167: Kaposi sarcoma-associated herpesvirus infection	10	0.049
bta04520: Adherens junction	6	0.050

The THI is commonly utilized as a measure to evaluate the intensity of HS in livestock, as an accurate and reliable estimation of heat load can help reduce or prevent economic losses, including reduced reproductive efficiency and decreased milk production in dairy cattle (Bohmanova et al., 2007; Habeeb et al., 2018). THI values are categorized into five distinct levels: no HS (THI < 72), mild HS (72 ≤ THI ≤ 78), moderate HS (78 < THI < 89), severe HS (89 ≤ THI ≤ 98), and extreme HS that can lead to mortality (THI > 98) (Armstrong, 1994; Dash et al., 2016). THI values are categorized into five distinct levels: no HS (THI < 72), mild HS (72 ≤ THI ≤ 78), moderate HS (78 < THI < 89), severe HS (89 ≤ THI ≤ 98), and extreme HS that can lead to mortality (THI > 98) (Armstrong, 1994; Dash et al., 2016). To determine if the cows were experiencing HS, we analysed the relationship between THI and physiological parameters such as milk yield and ADG. Blood samples were collected during a period when mild HS persisted for more than a month (specifically 36 days), during which both milk yield and ADG declined. Control samples (non-HS) were obtained four weeks after the minimum THI had stabilized below 72, indicating that cows subjected to HS can fully recover from prolonged HS, as evidenced by improved milk yield and ADG.

It is noteworthy that the significant reduction in milk production observed in Holstein and Jersey cows during the summer may be partly attributed to decreased feed intake. The more pronounced decline in milk yield in Holstein cows compared to Jersey cows is consistent with previous findings suggesting that Jersey cows may be more resilient to high temperatures (Roche, 2003; Smith et al., 2013; Lee et al., 2023). The feed intake was indirectly assessed using ADG. In this study, the negative ADG observed in HS cows suggests that HS may indirectly lead to reduced feed intake and appetite, ultimately causing weight loss. All lactating Holstein and Jersey cows used in this study were pregnant, and the substantial increase in ADG observed from September, after the summer period, could be attributed to compensatory placental growth driven by increased feed intake (Lee et al., 2020). Consistent with earlier research that demonstrated a positive correlation between rectal temperature and THI values (Dikmen and Hansen, 2009; Lee et al., 2020), and differences in rectal temperature according to breed (Muller and Botha, 1993), our findings also showed that Jersey cows had lower rectal temperatures compared to Holstein cows, indicating that Jersey cows are more resilient to HS. These physiological indicators confirm that Holstein cows exhibit greater sensitivity to thermal stress compared to Jersey cows.

To gain deeper insights into the relationship between these physiological markers, biological processes, and cellular responses to HS, we conducted RNA-sequencing analysis to identify DE miRNAs. An in silico method was utilized for miRNA target prediction, as circulating miRNAs in body fluids are known to influence various biological processes and could serve as potential biomarkers for the HS response. We examined 4 DE miRNAs that were commonly expressed in both Holstein and Jersey cows under HS conditions (Table 4). Interestingly, bioinformatics analysis indicated that these miRNAs are predicted to target genes involved in corpus luteum function, including progesterone biosynthesis. In our previous studies, the upregulated bta-miR-20b was found to target XCL1 (Lee et al., 2020), while the downregulated bta-miR-1246 targets StAR, a key gene involved in progesterone biosynthesis. Additionally, downregulated bta-miR-2284x and 2284y which are ruminant specific miRNAs have been reported to target numerous genes related to the immune response (Lee et al., 2020). Although we did not measure progesterone levels in this study, previous research has reported controversial findings regarding plasma progesterone levels under HS conditions (Wolfenson et al., 2000). Some studies have observed a decline in plasma progesterone levels during HS (Wise et al., 1988b; Wolfenson et al., 1988), while others have found no significant change or even an increase in progesterone levels (Thatcher and Roman-Ponce, 1981; Wise et al., 1988a). These variations can be attributed to several factors, such as the potential release of progesterone from the adrenal glands, hepatic metabolism, changes in blood volume (haemodilution or haemoconcentration), the severity and type of heat exposure (acute vs. chronic), as well as the age of cows, lactation stage, and feeding regimen, all of which influence the observed effects of HS on plasma progesterone levels (Jonsson et al., 1997; Trout et al., 1998).

Using an integrative approach with predicted target genes of DE miRNAs and DAVID, several KEGG pathways were identified, including metabolic pathways, Fanconi anemia pathway, FoxO signaling pathway, Sphingolipid metabolism associated with HS. The Fanconi anemia pathway is essential for recovering from HS by repairing DNA damage resulting from the overproduction of ROS in high-temperature environment (Yoshimoto et al., 2012; Kupfer, 2013; Li et al., 2024). The FoxO signaling pathway plays a crucial role in the transcriptional activation of heat shock proteins under HS conditions (Donovan and Marr, 2016; Farhan et al., 2017). We also found that the sphingolipid metabolism pathway is affected by HS, which can influence protein metabolism and subsequently impact growth performance in mammals (Liu et al., 2022).

Finally, when comparing the 11 potential miRNA biomarkers identified in Holstein cows from our previous study with those in Jersey cows, we found that although the expression trends (up / down) were similar to those observed in Holstein cows, the expression levels in Jersey cows were not statistically significant, and the magnitude of expression changes (|FC|) was lower compared to Holstein cows (Table 3) (Lee et al., 2020). This suggests that Jersey cows may have greater resilience to HS compared to Holstein cows, supporting findings from previous studies (Lee et al., 2023). The findings of this study are limited by the small dataset and the lack of functional validation of the roles of DE miRNAs in Holstein and Jersey cows. Therefore, further studies are needed to verify our results using a larger population and to investigate the specific roles of each DE miRNA under controlled environmental conditions.

We evaluated physiological HS indicators, including rectal temperature, milk production, and ADG in both lactating Holstein and Jersey cows, revealing that Holstein cows are more susceptible to HS conditions. RNA-sequencing-based transcriptome analysis of miRNAs identified 4 DE miRNAs (bta-miR-20b, bta-miR-1246, bta-miR-2284x and bta-miR-2284y) in both lactating Holstein and Jersey cows (|FC| ≥ 2, p < 0,05). Among these, bta-miR-20b, which is upregulated under HS, and bta-miR-1246, which is downregulated, were found to target several genes associated with corpus luteum function and progesterone biosynthesis. The remaining miRNAs, bta-miR-2284x and 2284y, were shown to be related to immune response. Furthermore, when we compared the expression levels of 11 potential HS biomarkers previously identified in lactating Holstein cows to those in lactating Jersey cows, the expression trends (up / down) were similar, but the magnitude of expression changes (|FC|) was lower in Jersey compared to Holstein cows. These findings suggest that Jersey cows may have greater resilience to HS.

None.

Conceptualization, J.L., I.C.; data curation, J.L.; formal analysis, J.L.; investigation, J.L.; methodology, J.L., B.L., I.C.; project administration, J.L., T.C., S.K.; resources, J.L., D.K., G.R., H.B., J.K., S.H.; supervision, S.L., T.C., I.C.; writing - original draft, J.L.; writing - review & editing, J.L., I.C.

This research was supported by the National Institute of Animal Science, Project (PJ015006), with the goal of improving productivity and sustainability in the dairy sector through innovative research and development efforts.

All animal experimental protocols and procedures were reviewed and approved by the Animal Care and Ethics Committee of the National Institute of Animal Science, South Korea (approval number: NIAS-068).

Not applicable.

Not applicable.

Not applicable.

No potential conflict of interest relevant to this article was reported.