All experimental procedures involving animals were approved by the Seoul National University Institutional Animal Care and Use Committee (SNUIACUC), Republic of Korea, and conducted in accordance with the Animal Experimental Guidelines of the SNUIACUC. The study was conducted at the University Animal Farm of the College of Agriculture and Life Sciences, Pyeongchang Campus of Seoul National University, South Korea.

In the feedlot trial, 20 Korean cattle steers with an average age of 19.7±0.13 months and body weight (BW) of 550.6±9.14 kg were used. The steers had been fed commercial early fattening stage concentrate using an automatic feeding station (DeLaval Alpro System; DeLaval, Tumba, Sweden) and rice straw, following a conventional feeding program. Water was provided freely. During the 2-week adaptation period before the experiment, all animals were fed an experimental control concentrate (approximately 1.6% BW/animal) and rice straw (1 kg/d/head). Steers were assigned to one of two treatments: the control group and RPF supplementation group. The RPF is prilled form of palm oil, as described [14], and purchased from Ecolex SDN. BHD (Pulau Indah, Selangor, Malyasia). The RPF was composed of 99.63% free fatty acids, including 85.48% palmitic acid (C 16:0), 7.05% oleic acid (C 18:1), 3.45% myristic acid (C 14:0), 1.64% linoleic acid (C 18:2), and 1.04% lauric acid (C 12:0), with an energy density of 9,316 kcal/kg (Haneol Corp., Anseong, Korea). Contents of the distillers dried grains with solubles, corn flour, corn gluten feed, barley stone, and palm oil were also adjusted to make similar non-fiber carbohydrate, neutral detergent fiber (NDF), and acid detergent fiber (ADF) between two diets. Table 1 lists the formula and chemical compositions of the experimental diets. Steers were fed a concentrate diet (1.6% BW) using an automatic feeding station and a rice straw diet (1 kg/d) for 16 weeks (January 9 to February 5 [P1], February 6 to March 5 [P2], March 6 to April 3 [P3], and April 4 to May 2 [P4]). The daily feed intake of the concentrate was automatically recorded online using a computer with the DeLaval Alpro system. Equal amounts of roughage were provided twice daily (08:00 and 18:00) and residual roughage was weighed before the morning feeding. Concentrate and rice straw samples were collected weekly and stored at −20°C until analysis. BW was measured before morning feeding on the start day, at 4-week intervals thereafter.

The chemical compositions (dry matter, crude protein, ether extract, ash, calcium, and phosphorus) of the concentrate and rice straw were determined using the AOAC method [15]. The NDF and ADF contents of the rice straw were analyzed using the sequential method with an ANKOM200 Fiber Analyzer (Ankom Technology Corp., Macedon, NY, USA) and reagents, as described by Van Soest et al [16].

Blood was collected before feeding (after 9 h of fasting) at approximately 09:00 on the start date, and at 4-week intervals thereafter. Blood was collected via jugular venipuncture with both a non-heparinized vacutainer (20 mL; Becton–Dickinson, Franklin Lakes, NJ, USA) and ethylenediaminetetraacetic acid-treated vacutainer (20 mL). Serum and plasma were separated by centrifugation at 1,500×g at 4°C for 15 min and stored at −80°C until analysis.

Ambient and climate temperatures and relative humidity inside and outside the barn, were recorded in 1-h intervals using four HOBO data loggers (Onset Computer Corp., Bourne, MA, USA). Minimum, mean, and maximum temperatures and corresponding relative humidity data were chosen at every day, and monthly data were average values of 28 days per month. The experimental farm was covered with a roof and the animals were raised indoors; therefore, the animals were protected from precipitation and direct sunlight. Doors were installed on both sides of the barn, which remained open to allow exposure to cold weather. Thus, low winds may have affected the wind-chill temperature.

After blood collection, rumen fluid was collected before feeding (after 9 h of fasting) using the oral stomach tube method, as described by Shen et al [17]. Rumen fluid pH was measured immediately with a pH meter (Ohaus Corp., Parsippany, NJ, USA). For the VFA analysis, 1 mL of rumen fluid was mixed with 0.2 mL of 25% meta-phosphoric acid and stored at −20°C until analysis, and an additional 30 mL of rumen fluid was stored at −20°C for the ruminal ammonia nitrogen (NH₃-N) analysis. NH₃-N concentrations were determined using a modified colorimetric method [18]. VFA concentrations were determined by gas chromatography using an Agilent Tech 7890A (Hewlett Packard, Waldbronn, Germany) with a Supelco fused silica capillary column (30 m×0.25 mm×0.25 μm).

The analytical reagents for albumin, glucose, triglycerides (TG), high-density lipoprotein (HDL), low-density lipoprotein (LDL), cholesterol, glutamic oxaloacetic transaminase (GOT), glutamic pyruvate transaminase (GPT), calcium, magnesium, and phosphorus were purchased from JW Medical (Seoul, Korea). The analytical reagents for the NEFA analysis were purchased from Wako Pure Chemical (Osaka, Japan). These parameters were analyzed using an automated chemistry analyzer (7180; Hitachi, Tokyo, Japan). Plasma cortisol was analyzed using a Cortisol Salivary HS Enzyme-linked Immunosorbent Assay Kit (cat. no. SLV4635; DRG Instruments, Marburg, Germany). The intra- and inter-assay coefficients of variation for the cortisol kit were 4.0% and 4.6%, respectively, based on bovine plasma samples. All analytical methods were verified in our laboratory, as reported previously [19].

Differences in climate data by month were analyzed with one-way analysis of variance (ANOVA). Differences in growth performance, blood parameters, and rumen fluid parameters by month and dietary treatment were analyzed using a repeated-measures two-way ANOVA. The statistical model included month, diet, and their interaction. A p value less than 0.05 was considered to indicate significance. All statistical tests were performed using R Studio for Windows software (R Studio, Boston, MA, USA).

The mean (P1: −3.44°C, P2: −0.70°C) and minimum (P1: −9.40°C, P2: −5.5°C) indoor ambient temperatures in January (P1) and February (P2) were lower (p<0.001) than those in April (P4; mean: 11.18°C, minimum: 4.28°C; Table 2). In a previous study, cold stress was categorized as mild (0°C to −6.7°C), moderate (−7.2°C to −13.9°C), or severe (<−13.9°C) under dry-winter-coat conditions in cattle [20]. Therefore, the minimum temperatures of P1 and both P2 and P3 in this study were considered to represent moderate and mild cold-stress (CS) conditions, respectively. In addition, we measured the average temperature during blood collection (08:00 to 10:00). The temperatures were −6.42°C, −0.29°C, 6.79°C, 10.8°C, and 14.0°C on January 9, February 6, March 6, April 4, and May 2, respectively. Then sampling days in January and February were classified as mild CS conditions, while sampling days in March, April and May were considered as thermo-neutral conditions.

The BW was higher (p<0.001) during P4 than P1, reflecting animal age. Moreover, the daily concentrate intake was higher (p = 0.03) during P4 than P1 (Table 3). In this study, the daily allowance of concentrate was set at 1.6% BW and was adjusted each month based on BW. Thus, the higher concentrate intake during P4 reflected the increased concentrate allowance, corresponding to a higher BW during P4 than P1. Daily forage intake was higher in P1 than in the other months. RPF supplementation did not affect concentrate and roughage intake during all studied months. Neither month nor RPF supplementation affected the average daily gain and feed efficiency (gain-to-feed ratio; Table 3). In a previous study, dietary RPF supplementation (4.5% fatty acid calcium salts) increased the G:F ratio in beef cattle [13]. In South Korea, commercial feed for Korean cattle steers has generally higher energy levels to produce highly marbled beef [21]. Additional energy supply through RPF supplementation may not have significant impact on growth performance in mild cold conditions. Another possible explanation for no influence on growth performance by fat addition is an insufficient amount of RPF supplementation (0.5%).

Neither month nor RPF supplementation affected ruminal pH (p>0.05; Table 4). However, ruminal NH₃-N concentrations were higher (p<0.05) in cold winter (February) than spring (April and May) (Table 4). For comparison, in sheep, NH₃ production was dependent on diet and ambient temperature [22]. Moreover, cold exposure did not affect ruminal NH₃ production with barley–canola seed meal and lucerne diets, but reduced NH₃ production with a bromegrass diet. Meanwhile, cold exposure reduced the irreversible loss of both plasma urea and rumen NH₃, and the conversion of plasma urea-nitrogen into rumen NH₃ was greater in cold-exposed sheep than in warm sheep [23]. The same authors reported an increased efficiency of ruminal microbial synthesis under cold exposure in sheep. Thus, the differences in rumen NH₃ concentrations between cold winter and spring in this study may have been related to changes in the irreversible loss of NH₃, conversion of urea into NH₃, or to the efficiency of microbial synthesis upon cold exposure. Meanwhile, RPF supplementation did not affect ruminal NH₃-N concentrations.

Neither month nor RPF supplementation affected C2, C3, iso-C5, C5, or total VFA concentrations in rumen fluid (p> 0.05; Table 4). The C2:C3 ratio tended to be lower (p = 0.07) in the RPF supplementation group than in the control group. In another study, increasing levels of protected lipids linearly increased the molar proportion of C3, whereas the molar proportion of C2 remained unchanged, resulting in a linear decrease in the C2:C3 ratio [24]. RPF supplementation decreased (p = 0.02) ruminal iso-C4 concentrations, while ruminal C4 concentrations varied by season (p = 0.02; Table 4).

Plasma cortisol concentrations were lower (p<0.05) on the starting day than during the other periods, and RPF supplementation did not affect (p>0.05) cortisol concentrations (Table 5). For comparison, plasma cortisol and corticosterone concentrations were similar among newborn calves at −4.0°C and 16°C, although the concentrations varied at −12°C and −18°C in one or two animals [25]. Collectively, blood cortisol concentration does not appear to be a suitable marker of cold stress, although cortisol is commonly used as a general stress marker [26].

Serum glucose concentrations were generally higher (p< 0.001) in colder weather but were unaffected by RPF supplementation (Table 5). Young [27] suggested that increased blood glucose results from increased metabolism (e.g., metabolic rate or heart rate) under cold conditions. In humans, lipid and muscle glycogen have major roles in providing the energy required for heat production under cold exposure, whereas plasma glucose has only a minor role [28]. Meanwhile, in lactating beef cows, fat supplementation did not affect circulating glucose concentrations [29].

Month did not affect total cholesterol concentrations (p> 0.05; Table 5), and serum HDL concentrations were similar among January, February, March, and early April, albeit that the concentrations in late April were comparatively lower. RPF supplementation increased both total cholesterol (p = 0.004) and HDL concentrations (p = 0.03; Table 5). In Holstein calves, various types of fat supplementation increased plasma cholesterol and HDL concentrations, with no differences seen in TG and glucose concentrations, during the cold season [30]. Furthermore, dietary rumen-protected oleic acid increased blood HDL concentrations in cattle [31]. Collectively, the increased HDL concentrations observed in this and previous studies may be due to increased amounts of absorbed fat in the small intestine via RPF supplementation. Moreover, changes in posthepatic lipoprotein metabolism, including alterations in HDL concentrations, may contribute to increased cholesterol concentrations with RPF supplementation, as suggested by Duske et al [32].

Neither month nor RPF supplementation affected serum TG, LDL, or NEFA concentrations. The effects of RPF supplementation on blood NEFA concentrations in several studies of lactating dairy cows were inconsistent with our results [24,33], and showed an increase in blood NEFA concentrations with RPF supplementation. Overall, little information is available on the effects of RPF supplementation on NEFA concentrations in beef cattle.

Serum albumin concentrations were higher (p = 0.017) on the experimental starting day in the colder month (January) than in the other months in spring, although the concentrations were unaffected by RPF supplementation (p>0.05; Table 4). A previous study revealed higher serum albumin concentrations with changes in the osmotic pressure of body fluid during hotter summer periods than colder winter periods in dairy cows [33]. Thus, the seasonal variation in circulating albumin concentrations remains unclear.

Serum GOT and GPT concentrations were not affected by month, though GPT concentrations were lower (p<0.001) in the RPF supplementation group, which has not been reported previously (Table 5).

Month and/or RPF supplementation affected blood mineral concentrations (Table 5). Blood calcium concentrations were higher (p<0.001), while blood magnesium concentrations were lower (p<0.001), on the starting day in January compared with the spring months. In addition, the RPF supplementation group had lower (p<0.001) phosphorus and magnesium concentrations than the control group. In a previous study, serum magnesium levels were lower in hyperthyroid human patients [34]. Furthermore, thyroid hormone thyroxine increased under cold conditions in beef cattle [35]. Thus, the lower magnesium concentrations under cold conditions in this study may have been indirectly influenced by higher circulating thyroid hormone levels. However, changes in these mineral concentrations according to RPF supplementation have not been well studied.