Oxidative stress status and reproductive performance of sows during gestation and lactation under different thermal environments

Objective Two experiments were conducted using 28 healthy multiparous sows to evaluate the oxidative stress status and reproductive performance of sows during gestation and lactation under different thermal environments. Methods Fourteen multiparous sows were used in Exp. 1 under a high thermal environment, and the other 14 multiparous sows were used in Exp. 2 under a moderate thermal environment. In both experiments, reproductive performances of sows were recorded. Plasma samples were collected on d 35, 60, 90, and 109 of gestation, and d 1 and 18 of lactation for malondialdehyde, protein carbonyls, 8-hydroxy-deoxyguanosine, immunoglobulin g (IgG), and IgM analysis. Results For sows in Exp. 1, plasma malondialdehyde concentration on d 109 of gestation tended to be greater (p<0.05) than it on d 18 of lactation. Plasma concentration of protein carbonyl on d 109 of gestation was the greatest (p<0.05) compared with all the other days. Plasma concentrations of 8-hydroxy-deoxyguanosine on d 109 of gestation was greater (p< 0.05) than d 18 of lactation in Exp. 1. For sows in Exp. 2, there was no difference of malondialdehyde and protein carbonyl concentration during gestation and lactation. In both Exp. 1 and 2, litter size and litter weight were found to be negatively correlated with oxidative stress indicators. Conclusion Sows under a high thermal environment had increased oxidative stress during late gestation indicating that increased oxidative damage to lipid, protein, and DNA could be one of the contributing factors for reduced reproductive performance of sows in this environment. This study indicates the importance of providing a moderate thermal environment to gestating and lactating sows to minimize the increase of oxidative stress during late gestation which can impair reproductive outcomes.


Reactive oxygen species (ROS) including superoxide anion (O 2
•-), hydroxyl radical ( • OH), and hydrogen peroxide (H 2 O 2 ) are produced during biochemical processes within an animal's body [1]. If the balance between the production of ROS and antioxidant defense is disturbed, the animal can suffer from oxidative stress. Accumulation of ROS could react with biological macromolecules such as proteins, lipids, and nucleic acids, resulting in lipid peroxidation, protein damage, and DNA damage [2,3]. In sows, external factors such as social stress and environmental stressors can lead to increased oxidative stress [4,5]. Zhao et al [5] studied the oxidative stress status of gestating and lactating sows under different gestational housing systems and social ranks. The results showed that sows under intense social stress could increase oxidative stress level during gestation and lactation [5]. Bottje et al [4] showed that environmental stressors such as high dust and ammonia levels in poultry house environment can cause oxidative stress in lung lining fluid of broilers.
Heat stress is another important environmental stressor to animal' s reproductive performance especially during sum mer months in tropical and subtropical regions of the world. It has a significant negative impact on reproductive perfor mance of sows including impaired embryonic development [6], reduced feed intake and milk production in sows [7,8], and delayed puberty of gilts [9], which result in summer in fertility and productivity losses that amount to millions of dollars. In the last decades, genetic improvement of sows for high reproductive performance results in increased metabolic heat production which makes sows more susceptible to the high thermal environment [10,11]. Previous studies found that sows were under severe catabolic status during late ges tation and lactation [12,13], and their oxidative stress levels increased during the same period [14]. However, the nega tive impact of high thermal environment on sow's oxidative stress status have not been well studied, and it is not known if oxidative stress status is related to reproductive performance of sows under different thermal environments. Therefore, we hypothesized that oxidative stress status of sows could elevate during late gestation and lactation when sows were housed under high thermal environment, and the elevated oxidative stress could be detrimental to the reproductive performance of sows. Thus, the objective of this study was to investigate oxidative stress status and its relationship with reproductive performance of sows during gestation and lactation under moderate and high thermal environments.

Animal care
Procedures used in this study were reviewed and approved by the North Carolina State University Animal Care and Use Committee.

Animals and experimental design
Two experiments were carried out using 28  experiments were housed in the same gestation and farrowing buildings and managed with standard procedures. A negative pressure system was used to ventilate the gestation building with four 1,100rpm, 61cm wall fans. The farrowing building had an evaporative cooling system and two 1,100rpm, 61cm wall fans. Water flow through the evaporative cooling system was controlled by an onoff thermostat which was set at 25°C, and one gas heater was provided during winter season in the farrowing room. The thermal environment inside the facilities was monitored hourly using data loggers (Tinytag Plus 2, TPG 4500, Gemini Data Loggers, Chichester, UK) to measure temperature and relative humidity. Average hourly ambient temperature and temperaturehumidity index (THI) with in a day in the gestation and farrowing buildings in Exp. 1 and 2 were showed in Figure 1 and 2, respectively. In Exp. 1, sows were exposed to an ambient temperature above 25°C for an average of 17 h and 14 h per day in gestation building and farrowing building, respectively ( Figure 1A). The hourly THI was ranged from 72 to 78 in the gestation building and from 69 to 83 in the farrowing building ( Figure  1B). On the other hand, sows in Exp. 2 were kept in a moder ate thermal environment as shown in Figure 2. The hourly ambient temperature was from 12°C to 17°C in the gestation building and 19°C to 21°C in the farrowing building ( Figure  2A). The hourly THI was ranged from 54 to 61 in the ges tation building and from 65 to 67 in the farrowing building ( Figure 2B).
The estrus of weaning sows in these 2 experiments was detected daily by the immobilization response as a reaction to a heat check boar after 3 d back to the breeding barn. Sows were artificially inseminated twice (12 h apart) after estrus onset. Each insemination was conducted within 18 h of se men collection (Duroc semen). Pregnancy was detected and confirmed at d 30 postbreeding using an ultrasound scanner (VSS700 EZ Preg Checker, Veterinary Sales & Service Inc., Stuart, FL, USA). In the gestation building, sows were indi vidually housed in gestation crates (2.0×0.64 m) with semi automatic feeders, individual drinkers, and slatted flooring. They were fed 2.2 kg gestation diet daily which contained 13.3% crude protein (CP) and 3.3 Mcal metabolizable energy (ME)/kg (Table 1). Water was ad libitum during gestation and lactation. On d 108 of gestation, all of the sows were weighed and moved to farrowing crates (2.1×1.5 m) in an adjacent farrowing building. Three sows in Exp. 1 and 2 sows in Exp 2. did not maintain their pregnancy and did not farrow. Dur ing lactation, sows were fed ad libitum by electronic feeders (JYGA Technologies, SaintNicolas, QC, Canada) in the far rowing building. Lactation diet contained 15.8% CP and 3.5 Mcal ME/kg ( Table 1). The gestation and lactation diets were formulated to meet or exceed NRC nutrient requirements [15], and formulations were the same in Exp. 1 and 2. In both experiments, any feed left from the previous day was removed and weighed around 0800 h, and the difference between the amount of feed provided and remained was used as an esti mate of daily feed intake. The BW of sows was measured on d 35 and 109 of gestation, and d 1 and 18 of lactation. Backfat thickness (10th rib, 6 cm offmidline) of each sow was mea sured on d 1 and 18 of lactation using an ultrasound scanner (VSS700 EZ Preg Checker, Veterinary Sales & Service Inc., USA). After farrowing, individual weight of piglet and litter size of sows on d 1 (total number of piglets, born alive, still born, and mummy) and 18 of lactation were measured.

Blood sampling
In both Exp. 1 and 2, blood was sampled via jugular veni puncture from all of the sows restrained by snout snare between 0900 and 1000 h on d 35, 60, 90, and 109 of gesta tion in the gestation building, and d 1 and 18 of lactation in the farrowing building. Blood was collected using 9 mL ethylenediaminetetraacetic acidcoated syringes and dispos able 16gauge×0.1mm hypodermic needles (MONOVETTE, Sarstedt, Newton, NC, USA; AirTite Product Co. Inc., Vir ginia Beach, VA, USA). Plasma samples were obtained by centrifugation (5810 R, Eppendorf AG, Hamburg, Germany) at 3,000 g, 15 min, 4°C, then allocated into 1.5 mL micro centrifuge tubes, kept in liquid nitrogen for 1 h, and stored at -80°C until analysis.

Analysis of oxidative stress indicators
Plasma samples from Exp. 1 and 2 were used to measure con centrations of malondialdehyde (MDA), protein carbonyl, and 8hydroxydeoxyguanosine (8OHdG). Concentrations of MDA were measured using the thiobarbituric acid reactive substances assay kit (Cell Biolabs, San Diego, CA, USA) according to the method described by Zhao et al [5]. Plasma samples and MDA standards were incubated and reacted with thiobarbituric acid at 95°C. After centrifuge and buta nol extraction, samples and standards were read at 532 nm by a spectrophotometric plate reader (Synergy HT, BioTek Instruments, Winooski, VT, USA). Samples were quantified against the standard curve which was constructed by MDA standards. The detective limit of MDA analysis was 0.98 μM.
Protein concentration in each plasma sample was measured by a bicinchoninic acid protein assay kit (Pierce Biotech nology, Rockford, IL, USA). Then all plasma samples were diluted with bicinchoninic acid to reach protein concentra tion at 10 μg/mL before measuring protein carbonyl. Protein carbonyl concentration in each diluted plasma sample was measured via the protein carbonyl ELISA kit (Cell Biolabs, USA) according to the method described by Shen et al [16]. The protein carbonyl presented in the sample or standard was first derivatized to dinitrophenyl hydrazine and probed with an antidinitrophenyl antibody, then incubated with a secondary antibody. Finally substrates and stop solutions were added. Standards and samples were read at 450 nm by a spectrophotometric plate reader (Synergy HT, BioTek Instruments, USA). Protein carbonyl concentrations in samples were quantified against the standard curve which was drawn by protein carbonyl standards. The detective limit for protein carbonyl was 0.375 nmol/mg.
Concentrations of 8OHdG in plasma samples were measured using the oxidative DNA damage enzymelinked immuno sorbent assay (ELISA) kit (Cell Biolabs, USA) according to the method described by Weaver and Kim [17]. Briefly, plas ma samples and 8OHdG standards were first added into a 96well plate. Then an anti8OHdG monoclonal antibody was added, followed by adding a secondary antibody. A sub strate solution and stop solution were added. The absorbance of each well was read at 450 nm by a spectrophotometric plate reader (Synergy HT, BioTek Instruments, USA). Con centrations of 8OHdG in plasma samples were quantified against the standard curve which was constructed by 8OHdG standards. The detective limit for 8OHdG was 0.078 ng/mL.

Immunoglobulin evaluation
In both experiments, concentrations of immunoglobulin G (IgG) and IgM in sows' plasma on d 109 of gestation, d 1 and 18 of lactation were measured by ELISA kits (Bethyl, Mont gomery, TX, USA) according to the method described by Chaytor et al [18]. Briefly, goat antipig IgG or IgM were used as capture antibodies to coat wells. All of the samples were diluted to 1:100,000 for IgG and IgM measurement. Horse radish peroxidase goat antipig IgG or IgM was used as the detection. The plate was read at 450 nm by an ELISA plate reader (Synergy HT, BioTek Instruments, USA) and software (KC4 Data Analysis Software, BioTek Tnstrument, USA). Sample concentration was quantified against the standard curve which was drawn by standards. Detective limits were 7.8 ng/mL for IgG, and 15.6 ng/mL for IgM, respectively.

Statistical analysis
Exp. 1 and 2 were completely randomized designs. Data from oxidative stress indicators and immunoglobulin assay were compared among different gestation and lactation days in both experiments using the MIXED procedure of SAS (SAS Inst. Inc., Cary, NC, USA). Oxidative stress data from d 35 of gestation was used as a covariance when analyzing the oxida tive stress indicators. The individual sow was the experimental unit. Day was a fixed effect and the sow was a random effect. The canonical correlation between reproductive performance and oxidative stress indicators on different days in Exp. 1 and 2 were analyzed by the CANCORR procedure of SAS (SAS Inst., Inc., USA), respectively. Data was considered statisti cally different when probability values were less than 0.05, and probability less than 0.1 and equal or greater than 0.05 was considered as a trend.

Oxidative stress parameters
Plasma concentrations of MDA, protein carbonyl, and 8OHdG were compared among days during gestation and lactation in Exp. 1 and 2, respectively ( Table 2). For sows in Exp. 1, plasma MDA concentration on d 109 of gestation tended to be greater (p = 0.096) than it on d 18 of lactation (Table 2).   (Table 3).

Immunoglobulin evaluation
Plasma concentrations of IgG and IgM were compared among d 109 of gestation, and d 1 and 18 of lactation in Exp. 1 and 2, respectively (Table 4). For sows in Exp. 1, there was no dif ference for plasma concentrations of IgG and IgM on d 109 of gestation, and d 1 and 18 of lactation (Table 4). For sows in Exp. 2, plasma concentration of IgM did not differ among different days. However, plasma concentration of IgG on d 18 of lactation was greater (p<0.05) than it on d 1 of lactation (Table 4).

Canonical correlation analysis between oxidative stress and reproductive performance
Three sets of canonical relationships between oxidative stress indicators and reproductive performance of sows were ob tained in Exp. 1 ( Table 5). The first set of variables are oxidative stress indicators including concentrations of MDA, protein carbonyl, and 8OHdG on d 60 and 90 of gestation, and d18 of lactation (M d60 , P d60 , O d60 , M d90 , P d90 , O d90 , M d18 , P d18 , and O d18 ). The second set of variables are BW and backfat of sows, and piglet weight on d 3 and 18 of lactation (X 1 , X 2 , X 3 , X 4 , X 5 , and X 6 ). The correlation coefficients in the first set of ca nonical relationship reached a significant level (p<0.05), and the other two sets showed tendency (p = 0.093 and 0.059), suggesting that there is a significant canonical correlation between oxidative stress and reproductive performance of sow in Exp. 1 (Table 5). From the linear expression of canoni cal variable composition in the first set of canonical relationship, M d60 , O d60 , and X 2 were relatively large (  Two sets of canonical relationships between oxidative stress indicators and reproductive performance of sows in Exp. 2 and the composition of canonical variables are shown in Table 6. The first set of variables are oxidative stress indica tors including concentrations of MDA, protein carbonyl, and 8OHdG on d 90 of gestation, and d3 of lactation (M d90 , P d90 , O d90 , M d3 , P d3 , and O d3 ). The second set of variables are piglet weight on d 3 and 18 of lactation, litter size of born alive and wean of sows (X 5 , X 6 , X 7 , and X 8 ). The correlation coefficients in the first and second set of canonical relation ships reached a significant level (p<0.05), suggesting that there is a significant canonical correlation between oxida tive stress and reproductive performance of sow in Exp. 2 ( Table 6). From the linear expression of canonical variable composition in the first set of canonical relationship, M d3 , O d3 , and X 6 were relatively large ( Table 6), showing that MDA and 8OHdG on d 3 of lactation are negatively correlated with piglet weight on d 18 of lactation. As can be seen from the linear expression of canonical variables in the second set of canonical relationship, there was a large load on P d90 and X 8 ( Table 6), suggesting that protein carbonyl on d 90 2) Initial number of sows was 14. There were 3 sows in Exp. 1 and 2 sows in Exp. 2 did not maintain pregnancy and did not farrow which were excluded from the study. a,b Means within a row with different superscripts differ (p < 0.05). 2) The second set of variables are BW and backfat of sows, and piglet weight on d 3 and 18 of lactation (X 1 , X 2 , X 3 , X 4 , X 5 , and X 6 ). 2) The second set of variables are piglet weight on d 3 and 18 of lactation, and litter size of born alive and wean of sows (X 5 , X 6 , X 7 , and X 8 ).
of gestation is negatively correlated with litter size of sows on d 18 of lactation.

DISCUSSION
In this study we have confirmed that sows under a high ther mal environment had increased oxidative stress during late gestation, and the increased oxidative damage to lipid, pro tein, and DNA could be one of the contributing factors for reduced reproductive performance of sows under the high thermal environment. Hence, we accepted the hypothesis that oxidative stress status of sows could elevate during late gestation and lactation when sows are housed under high thermal environment, and the elevated oxidative stress could be detrimental to the reproductive performance of sows.
In this study, sows in Exp. 1 were under heat stress condi tion indicated by the observed ambient temperatures and THI. We observed that the ambient temperature in Exp. 1 were above 25°C for an average of 17 h/d in the gestation building and 14 h/d in the farrowing building. Environmental tem perature remaining above 25°C could cause heat stress to sows [19]. Meanwhile, high levels of THI (72 to 78 in the gestation building, and 69 to 83 in the farrowing building) was also observed in Exp. 1. The THI is a combination of tempera ture and humidity which has been used to indicate heat stress environment. It was reported that the THI value less than 74 was classified as a safe, whereas critical when greater than 74 in sows [20]. The observed high THI in Exp. 1 can be detri mental to sows, because the high temperature and high humidity together can limit their heat loss through evapo ration [21]. These data indicated that sows in Exp. 1 were under heat stress condition. On the other hand, based on the average daily maximum and minimum temperatures and THI recorded in Exp. 2, it seems that sows in Exp. 2 were kept within their thermosneutral environment.
Our Exp. 1 was conducted from June to August in North Carolina, which we observed the reduced reproductive per formance of sows indicated by the less number of total born, born alive, more still born piglets, and small litter size and litter weight gain. These results indicated that the reproduc tive performance of sows in Exp. 1 was negatively affected by the heat stress environment. The smaller litter size of sow in Exp. 1 contributed to reduce mobilizing of their body re serve for less milk production, which may explain why sows in Exp. 1 lost less BW and backfat during lactation com pared with sows in Exp. 2. Similar result was reported that the overall reproductive performance of sows reduced in large confinement units in North Carolina during hot months from June to October [22].These findings were also consis tent with other studies which showed that sows under high thermal environment during gestation farrowed fewer born alive with increased number of stillborn [8]. On the other hand, feed intake of lactating sows was not the main factor for the reduced milk production in Exp.1. Because feed in take in Exp. 1 was similar to that in Exp. 2, and it did not show significant reduction as some other studies have found [23]. This may be because it took a few days for sows to adjust the electronic feeders in the farrowing room.
Besides the reduced reproductive performance of sows in Exp. 1, heat stress environment also affected the oxidative stress status of sows indicated by the enhanced oxidative damage to lipid, protein, and DNA during late gestation. The plasma MDA, protein carbonyl, and 8OHdG concentra tion increased during late gestation which showing that sows could suffer severe oxidative damage during late gestation under heat stress environment. The current results indicate that oxidative stress could be one of stress responses caused by high thermal environment. On the other hand, sow in Exp. 2 did not show the similar pattern with Exp. 1. Although sows had increased oxidative DNA damage during gesta tion and early lactation compared with late lactation, other oxidative indicators for protein and lipid did not enhanced, indicating that sows in Exp. 2 did not suffer severe oxidative stress during gestation and lactation when they were housed under moderate thermal environment. Previous studies showed that sows were under severe catabolic status during late gestation and lactation which could induce oxidative damage and decrease antioxidants concentration at the same period [1214]. Studies also found that the ROS production increased under the high temperature treatment [24,25], which could explain why sows in Exp. 1 showed enhanced oxidative stress. However, the molecular mechanisms respon sible for excessive ROS production or lowered antioxidant defenses under heat stress are still not known in sows. There are some reports in poultry which found that heat stressed chickens exhibit overproduction of mitochondrial ROS in skeletal muscle, which might result from enhanced sub strate oxidation and downregulation of avian uncoupling protein [26,27].
As discussed above, heat stress environment negatively affected the reproductive performance and enhanced the oxidative damage to sows during late gestation and lactation. We hypothesized that the elevated oxidative stress could be detrimental to the reproductive performance of sows. There fore, we investigated the relationship between oxidative stress indicators and reproductive performance of sows during gestation and lactation under both moderate and high ther mal environments. The canonical correlation coefficients were close to 1 for all the sets of canonical relationships, and the linear expression of canonical variable compositions showed that there were negative correlations between oxidative stress indicators and piglet weight, BW, backfat, and litter size of sows in current studies. These results indicates a strong cor relation between oxidative stress indicators and reproductive performance of sows. Similar results were reported in pre vious study, which showed that oxidative damages to lipid, protein, and DNA were negatively correlated with repro ductive performance of gestating sows under social stress [5]. Excessive ROS could negatively affect multiple physio logical processes from oocyte maturation to fertilization, embryo development and pregnancy [28]. Study on pregnant women showed that oxidative stress can increase membrane damage which is related with fetal growth retardation and cause a higher risk of prenatal mortality [29].
In Exp. 1, plasma concentrations of IgG and IgM were not different among days, indicating that sows did not suffer im mune challenge during the study. The plasma concentration of IgG on d 1 of lactation in Exp. 2 was smaller than it on d 18 of lactation, but both of them were within the normal range [30]. The decreased plasma concentration of IgG on d 1 of lactation may be because more IgG in the blood were secreted into the colostrum.

CONCLUSION
Taken together, sows under a high thermal environment had increased oxidative stress during late gestation. These oxidative stress indicators were negatively correlated with reproductive performance of sows, suggesting that increased oxidative damage to lipid, protein, and DNA could be one of the con tributing factors for reduced reproductive performance of sows under the high thermal environment. This study in dicates the importance of providing a moderate thermal environment to gestating and lactating sows to minimize the increase of oxidative stress during late gestation which can impair reproductive outcomes. Our findings in this study also shows the importance for future studies to regulate oxi dative stress status of sow during gestation and lactation under heat stress.

CONFLICT OF INTEREST
We certify that there is no conflict of interest with any financial organization regarding the material discussed in the manu script.