Experimental location and animals

The study was conducted at the Unidad academica de investigación en producción lechera of the Instituto Tecnologico de Sonora (ITSON), located in block 910 (N: 27°21’ O: 109°54’) of the Yaqui Valley, Sonora, Mexico, during June and July, 2011. Twenty-eight (28) multiparous Holstein cows with 150 ± 25.9 d in milk and 3.5 ± 0.35 of body condition score (BCS) were selected for this study. Cows were maintained in shaded holding pens with free access to water, and they were fed twice a day (i.e., 0800 to 0900 h and 1700 to 1800 h) with a ration made up with 48.8 % corn silage, 16.2 % alfalfa hay, and 35 % commercial energy supplement added with vitamins and minerals premix. All cows received bovine somatotropin (BST; 500 mg S.C.) each 14 d during lactation, from wk 9 postpartum until 15 d before to start the dry period.

Treatments

Cows were randomly assigned in two treatments: cooling (C) before milking (two times per day, 0700 and 1700 h) and additional cooling (AC) each 2 h (six times per day, from 0700 to 1700 h). Therefore, cows from treatments C and AC entered daily into the holding pen of the milking parlor, and they received 2 and 6 cooling periods that were equivalent to 36 and 108 min/d of cumulative cooling per treatment, respectively. Each cooling period included a sequence of three soaking cycles alternated with three ventilation cycles.

Cooling system

Each cooling cycle included 3 min of water aspersion (~ 15 L per cow/cycle) with a system made up by 16 sprayers (located at 2.73 m from the bottom) which allowed soaking the back and side of each cow, alternated with 3 min of forced ventilation with three ½ HP electric fans located at 2.73 m from the bottom. All cows received 8.5 m² of shadow inside the holding pens.

Climatic data

Ambient temperature (AT) and relative humidity (RH) data were collected from a nearby (~500 m) and automatic climatic station. Such information was used to calculate the temperature-humidity index (THI), that estimates the level of heat stress at which the cows were exposed, using the formula⁷: THI= (0.8*AT)+[(RH/ 100)*(AT”14.4)]+46.4. In order to obtain accuracy climatic values during the sampling period, climatic data were registered each hour from 0900 through 1700 h, which were averaged and combined to information from the last 2 d before each sampling date⁴.

Physiological markers of heat stress

Respiratory rate (RR; breaths/min) and rectal temperature (RT; °C) were collected in both groups at 1200 h twice per week during the study. The RR was measured by visual counting of intercostal movements¹, while RT was collected using a digital thermometer (TES-1310®) with a contact sounding line (type K; 9-cm large), that was introduced into the animal to touch the rectal mucosa²⁰.

Dynamic of blood hormone levels

Blood samples (i.e., 8 ml) were collected weekly from the coccygeal vein using Vacutainer tubes without anticoagulant. Once the samples were collected (i.e., from 1000 to 1100), each tube was labeled and transported (5 °C) to the Reproductive Biotechnology Lab from the Department of Veterinary and Agronomic Sciences at ITSON (5 °C) and centrifuged at 2,500 rpm/30 min to obtain the blood serum. Then, serum samples were transported to the Endocrinology Lab of New Mexico State University, USA, in order to measure concentrations (ng/ml) of prolactin (PRL) and insulin-like growth factor 1 (IGF1), using radioimmunoassay (Packard Instrument Company Inc., Meriden, CT, USA) of double antibody^11,12 under a coefficient of variation of 9.4 and 13.4 % for PRL and IGF1, respectively.

Milk yield

Cows were milked twice per day (0700 and 1700) and milk production (MY; kg) was recorded daily using an electronic system (Metraton 21™, GEA. Westfalia-Surge Farm Technologies, Siemensstraße, Bönen Germany).

Statistical analyses

The procedures PROC UNIVARIATE and PROC GLM (Levene’s test) were used to test normal distribution and equality of variances, respectively. Repeated and random statements from the MIXED procedure were processed using the SAS software V9.4²¹, in order to analyze repeated measures per week of RR, RT, PRL, IGF1 and MY. The mixed model included treatment, time, treatment*time interaction, and lactation number as fixed effects, and the cow as random effect. Days in milk was included in the model as covariant, which resulted as significant source of variation (P<0.05). According to Bayesian and Akaike information criterion, the auto-regressive covariance resulted as the best fitted structure for repeated measures in the mixed model. If the term time resulted as significant source of variation (P<0.05), then a pairwise comparison of the minimum square means was executed using the PDIFF option in SAS.

Climatic data

Ambient temperature showed a variation from 30.4 to 36.6 °C (Table 1). The maximum thermic threshold for Holstein cattle is 25 °C⁷, and thermic fluctuations from 21 to 32 °C lead to a reduction in milk yield²². The RH level fluctuated during the day from 18.1 y 32.8 % through the experiment. Effect of humidity is non-significant at 24 °C; however, an important negative impact appears when daily temperature increases above 34 °C²³, and evaporative functions (i.e., panting and sweating) of thermic regulation will be limited when RH is higher than 45 %. Although in this study HR was lower, it tended to intensify early in the summer which represents the beginning of regular conditions of heat stress in this region. The average THI during the day was 79.1 units that is considered as moderate heat stress¹⁵, and it became more adverse after wk 4. Collier et al²⁵ showed that milk yield in high producer cows starts to decline when the THI reaches 68 units.

Table 1 Climatic conditions and temperature-humidity index during the study 

ET= Environmental temperature; RH=Relative humidity; THI= Temperature-humidity index. All climatic variables correspond to the hourly average (0900 to 1700 h) of -2 d before sampling.

Heat stress markers

Treatment by time interaction was significant (P<0.001) for RR and RT (Table 2), which are considered as indicators of heat stress. It was expected that an increase in the number of showers improves RR and RT; however, the effect of the treatment on these variables was only detected at the end of the study. It coincided with the beginning of the summer (wk 6 and 7); therefore, the favorable effect after to increase the number of showers depended of the variable time/week which involves the climatic variables. Berman²⁰ tested different cooling levels at 38 °C and 47 % of humidity including until seven cycles of shower and ventilation to reduce rectal temperature in Holstein cows, but it did not improve after increase cycles from eight to ten. Spiers et al⁹ reported a group of cows which under thermo neutral conditions (AT of 19 °C and RH ~55 %) showed 59.6 breaths/min of RR and 39 °C of RT; however, once the cows were heat-stressed (AT of 29 °C and RH ~50 %), their RR and RT increased 32.8 and 2.5 %, respectively. In this study, RR was similar between treatments (P>0.05) during the first 5 wk (Figure 1), but it was lower in AC group in comparison to C group (P<0.05) in wk 6 (78.5 vs 82.9 breaths/ min) and 7 (69.5 vs 69.9 breaths/min), which coincided with the time when climatic conditions became more severe. Likewise, the RT in AC cows was lower (P<0.05) than C cows only in wk 6 (38.5 vs 39.0 °C) and 7 (38.6 vs 39.1 °C).

Table 2 Least squares means and probability for treatment, time and its interaction effects in Holstein cows subjected to different frequencies of cooling 

RR= Respiratory rate (breaths per minute); RT= Rectal temperature (°C); PRL= Prolactin levels (ng/ml); IGF1= Insulin like growth factor-1(ng/ml); MY= Milk yield (kilograms per day).

* Difference (P<0.05) of the treatment*time fixed effect. RT= Rectal temperature; RR= Respiratory rate; C= Cooling; AC= Additional cooling.

Figure 1 Respiratory rate (RR) and rectal temperature (RT) in Holstein cows subjected to different frequencies of cooling at the summer onset 

During the study, cows from C group increased (P<0.001) 32.0 and 3.4 % their RR and RT, respectively, while cows from AC group only increased 11.8 and 2.3 %, respectively. Differences between treatments are lower because the two groups received cooling. However, heat stress conditions start once the cow exceeds a RT of 39.2 °C and RR of 70 breaths/min. Under these conditions, the objective of the artificial cooling is to maintain RT and RR closer to a physiological threshold. At the end of this study, both groups maintained a RT <39.2 °C, even though RR in C cows was higher than 70 breaths/min.

Dynamic of blood hormone levels

Serum levels of PRL differed by treatment (P=0.004) and sampling time (P<0.001; Table 2). During the experiment, PRL average values were higher for AC treatment (35.1 ± 1.3) in comparison with C treatment (29.4 ± 1.1 ng/ ml). In contrast, Igono et al²² reported a reduction in PRL levels from cows that received cooling and shadow, in comparison to those cows that only received shadow, and they concluded that serum PRL declined in response to heat stress reduction and the improvement of the environmental comfort of the cows. However, the higher values of PRL found in this study in cows that received additional cooling could be associated to lactation process, because the blood samples were collected in the morning after milking. Therefore, we assume that this early sampling masked the effect of heat stress on PRL levels in cows without additional cooling.

Regarding to the time of sampling, the Figure 2 depicts a gradual reduction of serum PRL (P<0.05) on wk 4 and 5, which become steady on wk 6 and 7. In contrast to such results, Scharf et al¹⁴ have reported that PRL concentrations increase in warm climates, because it is an homoeothermic hormone associated to heat stress events^26,27. Likewise, PRL stimulates the expression of heat shock proteins (HSP)²⁸, which avoid protein denaturalization in bovine body cells threatened by environmental heat stress²⁹. The unexpected PRL response found in middle July was probably due to the short experimental time, because the negative effect of the weather in this area is more severe after July¹⁹ due to the accumulation of heat load in cows; then, activation of mechanisms involving PRL would be more evident at the end of the summer. Moreover, the covariant of days in milk could influence PRL levels at the end of the study, because this variable resulted as significant source of variation (P<0.05).

As effect of time [_*] indicate statistical difference (P<0.05) in relation to the previous week on average level per week (i.e., including both treatments; ng/ml) of PRL and IGF1 during the study.

Figure 2 Dynamic of the prolactin (PRL) and insulin like growth factor-1 (IGF1) levels (ng/ml) during the study 

The serum concentrations of IGF1 were only effected by time (P=0.0143; Table 2), and tended to increase at the end of the study (P<0.05; Figure 2). Rhoads et al¹⁰ have reported a small IGF1 reduction in cows exposed to warm climate conditions. Growth hormone (GH) and IGF1 have a close relationship on metabolic processes associated to milk yield¹¹, then GH measurements would help to explain the environmental effect on IGF1 levels and its relationships with milk production. It is important to mention that cows from this study received a subcutaneous application of bovine somatotropin (STB) as part of the herd management. The STB is a synthetic hormone with similar function to GH in bovine cattle, and its exogenous application is a potential enhancer of heat load in lactating cows under heat stress^13,30,31, as well as blood concentrations of IGF1^32-35. Therefore, administration of STB in this study could induce a direct effect on IGF1 levels, and consequently on PRL levels.

Although levels of IGF1 were constant during the study, it showed a significant increasing (P<0.05) until to reach its higher concentration on wk 7, which coincided with the hottest period of the experiment.

Growth hormone has alternative pathways to mobilize energy from other tissues in bovine cattle exposed to a high energetic demand needed for maintenance (i.e., thermoregulation) and lacto genesis in milking cows¹⁰. Metabolic regulation of GH and IGF1 allows to prioritize some emerging functions such as the adaptive response to a warm climate in transition for lactation maintenance of milking cattle. The PRL activates milk synthesis in ruminants, however, it is not essential for lactation maintenance³⁶, while the IGF1 pathway stimulates galactophoyesis and interacts by separate with PRL and STB³⁷. It explains the reduction in PRL levels and the increasing in IGF1 levels during the warmest period of this study.

Milk yield

The treatment by time interaction was significant (P=0.002) for milk yield (Table 2). A higher production was observed in cows from AC group in comparison with cows from C group on wk 6 (17.9 vs 21.0 kg) and 7 (18.0 vs 21.1 kg), when the heat stress represented by the ITH was higher (Figure 3). Direct cooling (shower) and forced ventilation under warm and moist climates is able to reduce negative effects from a summer heat season on cows physiology^16,18, which contributes to hormonal status¹² and help to correct energy metabolism¹³ in order to improve their productive performance¹. A 9.3 % improvement in milk production has been reported after to increase cooling series from 5 to 8 in the waiting parlor area¹⁷. Low pressure cooling in this area before milking can avoid a milk reduction of 0.76 kg/d per cow, when ambient temperature fluctuates from 27 y 46 °C¹⁵.

MY= Milk yield. [_*] In means of columns indicate the effect of treatment by week (P<0.05).THI= Temperature-Humidity Index.The treatments C= Cooling and AC= Additional cooling.

Figure 3 Effect of additional cooling on milk yield (MY) performance of Holstein cows 

In this experiment, the reduction in MY started at the third week when the cows were exposed to an THI level of 76.6 units (Figure 3). Collier et al²⁵ demonstrated that cows producing >35 kg/d need artificial cooling once the THI reaches 68 units, because at this threshold dairy cows reduce milk production 2.2 kg/d. Leyva et al¹⁹ found in Sonora a milk yield reduction of 0.65 kg/d per cow, once the threshold for THI reached 71.2 units.

While the negative effect of summer was approaching, higher conditions of heat stress where mitigated with the rise of cooling sessions in AC cows during the day.

This represents an advantage if we consider that all physiological and metabolic changes that the cows experienced trying to regulate body temperature by themselves could be diminished using additional cooling. Serum PRL levels and milk yield had the same trend in both treatments (Figures 2 and 3), then we assume that PRL was associated to milking instead of thermoregulation. However, both the increase in IGF1 levels and the additional cooling probably allowed the cow to mobilize energy to maintain lactation as the cows experienced a higher energy requirement to support thermoregulation the last 2 wk of this study.