Introduction

Homeorhetic regulation encompasses the coordinated control of metabolism essential for maintaining the physiological state of pregnancy (Daniel et al. 2018). During this stage, effective nutrient partitioning becomes critical to fulfill the requirements for fetal growth and the development of maternal tissues, including the placenta and mammary glands. Studies indicate that there is improvement in the cardiovascular system to ensure an adequate supply of nutrients and oxygen to maternal tissues, particularly the uteroplacental region (Moreira et al. 2021). As gestation advances, a variety of physical and physiological changes related to feed intake and digestion may manifest. It has been reported that feeding intake gradually decreases, with a more pronounced reduction observed during the last days of gestation, which can be attributed to the limited ruminal space caused by the expansion of the gravid uterus (Rotta et al. 2015). Additionally, physical limitations may be compensated for by an increase in passage rate (Linden et al. 2014; Moyo and Nsahlai 2018), negatively impacting nutrient digestibility in pregnant cows (Ribeiro et al. 2015). The synthesis of microbial protein relies on achieving a balance between the degradation rates of degradable protein and carbohydrate sources (Hackmann and Firkins 2015). In combination, both the stage of gestation and the nutritional plan can influence microbial protein production, consequently affecting maternal conditions.

In tropical regions, beef cows often face feed shortages during the mid-to late-gestation period, as these periods align with the dry season (Barcelos et al. 2022; Costa et al. 2021a; Carvalho et al. 2022; Nascimento et al. 2022). In this context, protein supplementation during crucial stages of gestation, has been employed as a strategy to improve nutrient intake and digestibility in beef cows raised on low-quality pastures (Marquez et al. 2017). Comparing maintenance to ad libitum feeding in crossbred cows (Holstein × Gyr), Rotta et al. (2015) showed greater dry matter apparent total-tract digestibility in maintenance-fed cows during mid-to-late gestation. Moreover, the mobilization of maternal reserves is minimized by the nutrient supply offered (Lopes et al. 2020; Rodrigues et al. 2021; Meneses et al. 2022). Addressing the elevated maternal requirements during gestation not only results in well-nourished dams but also yields healthier calves with enhanced growth potential (Barcelos et al. 2022). Additionally, well-nourished dams contribute to better meat quality at harvest, as reviewed by Costa et al. (2021b) and Santos et al. (2022).

However, given that pregnancy is a physiological state that triggers notable metabolic changes, additional interactions with feeding level and digestion may occur in ruminants. Additionally, it is crucial to emphasize that most research investigating the influence of pregnancy on feed intake and digestibility in cattle has been performed in temperate regions, with Bos taurus cattle serving as the primary model species. Research investigating these variables in Zebu cattle remains limited. Hence, it is crucial to address this knowledge gap to furnish pertinent information for Zebu animals, extensively utilized as a genetic resource in tropical regions. Consequently, this experiment aimed to assess the effects of pregnancy and feeding level on intake, digestibility, and microbial N production in the rumen of pregnant and non-pregnant Zebu cows.

Material and methods

This research was conducted at the Universidade Federal de Viçosa (Viçosa, MG, Brazil), adhering to established protocols for ethical animal care and handling (protocol number: 047/2012), following the guidelines of Universidade Federal de Viçosa (Brazil).

Animals

Forty-nine multiparous Nellore cows, exhibiting an average initial body weight (BW) of 451 ± 10 kg, an age of 5.6 ± 0.5 years, and a body condition score of 4.4 ± 0.2 on a scale of 1 to 9, were employed in this investigation. From the initial pool of 49 cows, a randomized selection process led to the separation of 32 cows, which were mated with Nellore bulls to form the pregnant group. These cows were randomly hand-mated with five Nellore bulls during a 50-day breeding period. Estrous synchronization was initiated through the administration of a gonadotropin-releasing hormone (GnRH) agonist injection, followed by a prostaglandin F2α (PGF2α) injection seven days later. The day of breeding served as the starting point for the pregnancy timeline, and 28 days post-breeding, an ultrasound was utilized to confirm and validate the pregnancy. Moreover, twelve additional cows were randomly selected and allocated to the non-pregnant group, with the remaining five cows specifically assigned to the baseline group.

Diet and management

Cows were housed in pens (5–6 cows), each spanning 48 m2 and featuring a concrete floor and an additional covered area of 15 m2. Furthermore, a continuous supply of potable water was ensured for the animals throughout the experimental period. Individual feed intake was quantified using an electronic headgate system (Kloppen Soluções Tecnológicas, Pirassununga, SP, Brazil). Cows in the gestational phase were at 47 ± 3 days of pregnancy (DOP) at the initiation of the feeding trial. The diet composition is presented in Table 1. Daily records were kept for both the provided feed and the orts. Samples of corn silage and orts were collected daily, and a composite sample was assembled weekly for subsequent analysis. Samples of the ground corn and soybean meal were taken each time the concentrate portion was mixed.

Table 1 Ingredients and chemical composition of the diet
Full size table

All cows received identical total mixed rations twice daily, at 0700 and 1500 h. Cows were categorized into two groups based on feeding levels: HIGH (ad libitum; n = 16 pregnant and 5 non-pregnant) or LOW [restricted feeding at 1.2 times maintenance according to the National Research Council (NRC 2000); n = 16 pregnant and 7 non-pregnant]. Restricted feeding was calculated to support pregnancy at 1.2 times the maintenance level, while the HIGH-fed regimen facilitated the accumulation of maternal tissue. The average dry matter intake (DMI) was 16.0 ± 2.0 and 10.8 ± 1.5 g/kg of shunk body weight (SBW)/d for cows fed HIGH and LOW levels of intake, respectively.

The cows utilized in this study originated from a comparative slaughter experiment (Gionbelli et al. 2015), which is a methodology involving the collection of data by slaughtering animals at different developmental stages. This approach allows for a direct comparison between different animal groups, such as animals in different physiological stages, as evaluated in this study. As a result, the 32 pregnant cows were randomly distributed into four groups, each comprising eight cows (with four cows per feeding level). These cows were slaughtered at 136, 189, 239, and 269 days of pregnancy. To ensure a similar duration of participation in the study as the pregnant cows, the non-pregnant cows were also slaughtered at various time points throughout the experiment, ranging from 85 to 216 days of gestation of pregnant cows.

Digestibility and dry matter intake

The digestibility trials consisted of nine collection periods, each lasting for 5 consecutive days, with an interval varying from 21 to 28 days. Owing to variations in harvest periods, the intervals between collection periods were irregular, aiming to maximize the number of collections per cow. Fecal samples (approximately 60 g) were obtained directly from the rectum of each cow during all five consecutive days (at 1800, 1500, 1200, 0900, and 0600 h), pooled for each animal in each period, and stored at -20 ºC for further analyses. The in vivo apparent total tract digestibility was estimated by using the indigestible neutral detergent fiber (iNDF) as an internal marker.

The dry matter intake (DMI) was calculated every 7 days through the composite samples collected from feeds and leftovers. The DMI in proportion to BW (g/kg) for each cow was obtained at each period of fecal sampling when cows were weighed. The SBW was calculated in non-pregnant (SBWnp) and pregnant (SBWp) cows as a function of the BW according to Gionbelli et al. (2015). The equation used to estimate the SBW was 0.8084 × BW1.0303.

Chemical analyses

All samples were lyophilized. Subsequently, the samples were ground using a Willey mill (TE-650, Tecnal, Piracicaba, SP, Brazil) to pass through a 2-mm screen for the determination of iNDF using the in situ incubation method for 10 days (Casali et al. 2008). The incubation process involved three rumen-cannulated Holstein × Nellore cows, which were provided ad libitum access to the same diet employed in the experiment. To quantify dry matter (DM), ash, crude protein (CP), ether extract (EE), and ash- and protein-free neutral detergent fiber (NDFap), the samples were ground using a knife mill equipped with a 1-mm sieve. The analysis methods employed were as follows: DM (method 934.01; AOAC 2000), ash (method 942.05; AOAC 2000), CP (method 920.87; AOAC 2000), EE (method 920.39; AOAC 2000), and NDFap (method described by Van Soest et al. 1991). Non-fiber carbohydrates (NFC) were calculated according to the approach proposed by Detmann and Valadares Filho (2010), where NFC = 100 – [(% CP – % CP derived from urea + %urea) + NDFap + EE + ash]. The determination of total digestive nutrients (TDN) concentration was calculated following the method outlined by Detmann et al. (2016). Gross energy was assessed using an adiabatic bomb calorimeter (C5001, IKA-Werke Co, Staufen, Germany) (Cadenas-Soberanis et al. 2021).

Microbial nitrogen synthesis

Spot urine samples were obtained by stimulating the area below the vulva to assess the excretion of urinary nitrogenous compounds. Two spot urine samples were gathered during each collection period, specifically on the second and fifth days of fecal sampling, at 1600 and 0800 h, respectively. The estimation of urine volume utilized creatinine concentration as a marker, and the calculation was performed using the following equation (Valadares Filho et al. 2016):

$${\text{UCE}}\left({\text{mg}}/{\text{day}}\right)=37.88\times {{\text{SBW}}}^{0.9316}$$

where: UCE = urinary creatinine excretion (mg/d); and SBW = shrunk body weight. Urinary concentrations of creatinine, allantoin, and uric acid were determined by employing a high-performance liquid chromatography method, as described by George et al. (2006).

Microbial nitrogen (N) synthesis was assessed utilizing the purine derivatives in the urine technique (Chen and Gomes 1992). The excretion of purine derivatives (PD) in urine was determined by the combined excretions of allantoin and uric acid, calculated by multiplying their concentrations in urine by the daily urinary volume. The calculation of absorbed purines (AP) from the excretion of PD followed the method described by Prates et al. (2012) as follows:

$${\text{AP}}\left({\text{mmol}}/{\text{day}}\right)={\text{PD}}-\left(0.405\times {{\text{BW}}}^{0.75}\right)/0.99$$

where: 0.405 × BW0.75 value = endogenous excretion of purine derivates (mmol) in the urine per unit of metabolic body weight (BW0.75); and 0.99 = recovered of absorbed purines as purine derivates in the urine (mmol/mmol).

The computation of ruminal synthesis of nitrogen compounds (Nmic) was determined based on the AP (Prates et al. 2012), as follows:

$${\text{Nmic}}\left({\text{g}}/{\text{day}}\right)={\text{AP}}\times 70/\left(0.93\times 0.11\times 1000\right)$$

where: 70 = N content in purine derivates (mg/mmol); 0.93 = purine digestibility; and 0.11 = ratio of purine N: total N of microorganisms.

Statistical analysis

The MIXED procedure in SAS version 9.2 (SAS Inst. Inc., Cary, NC) was employed to assess the impact of feeding level and pregnancy on the variables under study. This analysis took into account the influence of collection periods, repeated measures within the same animal, and interactions among the mentioned factors. A previous evaluation of the estimated TDN of the diet based on the chemical composition (Detmann et al. 2016) revealed a significant variation between collection periods justifying the use of a period of measure as a random effect in the model.

The response variables were evaluated as follows:

$${Y}_{ijk}=\mu +{F}_{k}+{\beta }_{1}\left({G}_{l}\right)+{\beta }_{2}{\left(F\times G\right)}_{kl}+{\delta }_{kl}+{\varepsilon }_{ijkl}$$

where Yijk is the observed measurement; µ is the overall mean; β1 and β2 are the regression coefficients; Fk is the fixed effect of feeding level; Gl is the fixed effect of DOP of pregnant cows; (F × G)kl is the interaction between feeding level and DOP of pregnant cows; δkl is the random error with mean 0 and variance σ2δ. The variance between cows (subjects) within feeding level and DOP of pregnant cows is equal to the covariance between repeated measurements within cows; and εijkl is the random error with the mean 0 and variance σ2, the variance between measurements within cows.

To eliminate probable misinterpretation effects and to certify the true relationship between the voluntary feed intake and DOP a new analysis was performed removing the LOW-fed cow’s information from the database and the feeding level from the statistic model to check the quadratic and cubic effects on DM intake. Only the linear parameter was significant (P = 0.012, 0.522, and 0.739, respectively).

As the interval between collection periods exhibited variability, the response variables were regarded as irregularly measured. In such instances, a continuous-time model, as proposed by Moser (2004) to describe covariances among errors, was applied. The spatial data covariance structures available in PROC MIXED of SAS were utilized, with the spatial power serving as the chosen covariance function. The least-square means were estimated for feeding level. A significance level of 0.05 was chosen as the critical probability threshold to determine the occurrence of Type I errors.

Results

Body weight

Due to the applied feeding levels, the final SBW significantly differed (P = 0.003) between feeding groups, with an average of 564 ± 6 kg for HIGH-fed cows and 481 ± 5 kg for LOW-fed cows. The shrunk body gain (SBG, P < 0.001) among feeding level groups was 0.86 ± 0.04 kg/d for HIGH-fed cows and 0.26 ± 0.04 kg/d for LOW-fed cows (Gionbelli et al. 2015). Furthermore, within each evaluated gestational period, comparisons revealed that HIGH-fed cows consistently had greater (P < 0.001) SBG than LOW-fed cows.

Dry matter intake

Figure 1 presents a graphical representation of the voluntary dry matter intake for each pregnant cow in the HIGH-fed group. The interaction between feeding level and DOP on the DMI expressed as g/kg of SBWp and g/kg of SBWnp was significant (P ≤ 0.018, Table 2). The DMI decreased as a proportion of SBW concerning DOP. Thus, the linear relationship between the voluntary intake and DOP in Nellore cows fed ad libitum can be described as follows (based on functions shown in Table 2):

Fig. 1
figure 1

Relationship between voluntary dry matter intake and days of pregnancy. Data of sixteen pregnant cows feeding ad libitum. The dots represent the weeks of evaluation

Full size image
Table 2 Effects of feeding level and pregnancy on dry matter intake in Nellore cows
Full size table
$${\text{DMI}}\left({\text{kg}}/{\text{day}}\right)=\left(16-0.0093\times {\text{DOP}}\right)/1000\times {\text{SBWp}}$$
$${\text{DMI}}\left({\text{kg}}/{\text{day}}\right)=\left(16.4-0.0093\times {\text{DOP}}\right)/1000\times {\text{SBWnp}}$$

Digestibility

The interaction between feeding level and DOP was significant (P < 0.05) for all the components of the diet, except for the NFC (Table 3). Except for DM digestibility and TDN content, there was a decline in digestibility as gestation time increased in HIGH-fed cows, whereas digestibility in LOW-fed cows increased with the increase in DOP. The DM, OM, NDFap, and GE digestibility coefficients were higher in LOW-fed cows (P < 0.05) compared to HIGH-fed cows. The EE digestibility was higher in HIGH-fed cows than in LOW-fed cows (P = 0.015). Figure 2 illustrates the graphical representation of the association between DM digestibility and pregnancy time, incorporating the functions detailed in Table 3.

Table 3 Effects of feeding level and pregnancy on total-tract apparent digestibility of diet components and microbial N production in Nellore cows
Full size table
Fig. 2
figure 2

Relationship between dry matter digestibility and days of pregnancy. Full dots and continuous lines represent HIGH-fed cows. Empty dots and dashed lines represent LOW-fed cows. The lines fitted to the data represent the functions shown in Table 2

Full size image

Microbial N production

There were no effects (P ≥ 0.180) of DOP and interaction between feeding level and pregnancy (P ≥ 0.368) on microbial N production and efficiency (Table 3). The microbial N production (g/d) and the efficiency of microbial N production by total digestible nutrients were higher in HIGH-fed cows than in LOW-fed cows (P < 0.001 and 0.049, respectively). However, the efficiency of microbial N production by g of N intake and kg of digestible organic matter intake was higher (P = 0.021) or tended to be higher (P = 0.051) in LOW-fed cows than in HIGH-fed cows, respectively.

Discussion

Regarding energy requirements, LOW-fed cows ingested 102%, 98%, and 67% of their estimated energy requirements according to the NRC (2000) at 0, 135, and 270 DOP, respectively. In contrast, HIGH-fed cows surpassed their energy requirements, consuming 168%, 162%, and 111% during the corresponding gestational periods evaluated in the LOW group. This suggests that cows on the HIGH-fed treatment, which consistently met a larger proportion of their energy requirements throughout gestation, had a greater supply of substrates available for nutrient accretion. As a result, they were able to achieve weight gain throughout gestation and maintain maternal body reserves in comparison to LOW-fed cows, as previously described by Gionbelli et al. (2015). Consistently, findings from other studies (Lopes et al. 2020; Rodrigues et al. 2021; Barcelos et al. 2022; Meneses et al. 2022) demonstrated that increasing the proportion of nutritional requirements met has contributed to weight gain during gestation and preservation of maternal body condition.

The DM intake voluntarily decreases in pregnant cows as they progress through late gestation (Hummel et al. 2021). In the current study, was also observed a reduction in DM intake (g/kg of BW) as pregnancy advanced. This decline aligns with a reduction in both wet and dry matter content in the rumen during gestation, suggesting a diminished ruminal capacity in pregnant Zebu heifers, as reported by Moreira et al. (2023). The decrease in DMI is probably a result of limited ruminal space due to compression by the gravid uterus and visceral fat (Forbes 2007). Additionally, using the same females and experimental design, Gionbelli (2013) observed a reduction in the weight of the rumen-reticulum and omasum with advancing pregnancy. The author also noted that the 8.18% reduction in rumen weight between 136 and 239 days of pregnancy corresponds to approximately a 37% reduction in rumen volume (80.0 to 58.3 liters).

Additionally, it's worth noting that in ruminant animals, around 75% of fetal growth takes place during late gestation, which further contributes to the constraint on rumen capacity (LeMaster et al. 2017). In the present study, HIGH-fed cows produced heavier offspring compared to LOW-fed cows at 269 DOP (Gionbeli et al. 2013), which in turn, could magnify this effect. In addition to the physical control of DMI in pregnant cows, other factors could play a role, including physiological, metabolic, and behavioral aspects. Nevertheless, these intricacies, such as the impact of calf weight on the reduction of reticulum-rumen volume, hormonal regulation during pregnancy, or the homeorhetic process governing nutrient utilization, are challenging to simulate and represent the primary drivers of the variations in voluntary intake observed during this specific physiological stage in cattle.

Accompanied by the decrease in DMI, HIGH-fed cows also experienced a decrease in DM digestibility as pregnancy progressed. The magnitude of this reduction was one point percent at every 107 days (1/0.0093), corresponding to 2.7 percentile points of difference between a non-pregnant and a pregnant cow at parturition (approx. 290 days of gestation for Nellore cows) (Cavalcante et al. 2001; Rocha et al. 2005). The lower nutrient digestibility as pregnancy advances is likely due to changes in ruminal kinetics, as evidenced by downregulated markers associated with cell proliferation and apoptosis that contribute to decreased turnover of the rumen epithelium in pregnant cows (Moreira et al. 2021). The variations in the abundance of nutrient transporters and the surface area available for absorption may potentially lead to compromised absorption of volatile fatty acids in cows during late gestation (Moreira et al. 2021).

Given that an increased feed passage rate has been associated with reduced nutrient digestibility (Ribeiro et al. 2015), it is reasonable to infer that the decline in digestibility observed in HIGH-fed cows as DOP increase might be linked to an accelerated passage rate. This heightened passage rate could potentially be a compensatory response to the reduced ruminal volume, as passage rates tend to be higher in late gestation in cattle (Gionbelli et al. 2016).

In line with this, Moreira et al. (2023), in a comparison of canulated pregnant and non-pregnant cows, consistently observed that pregnancy led to an increased passage rate and a decreased degradation rate during late gestation. Consequently, this led to a shortened retention time of DM in the rumen. The greater reduction in NDFap digestibility compared to DM digestibility during pregnancy is in agreement with the previous study. The NDFap digestibility reduces by one point percent every 62 days (1/0.0161), which was almost twice that in DM digestibility. On the other hand, the enhanced digestibility observed in LOW-fed cows could be attributed to the improved efficiency of cows subjected to restricted feeding, particularly as their pregnancy advances and they continue to receive limited amounts of feed over an extended duration.

Changes in dry matter intake and nutrient digestibility during gestation as a function of the nutritional plan have the potential to affect the pattern of ruminal fermentation and the synthesis of microbial nitrogen (Hare et al. 2019). The process of microbial N production in the rumen is intricate and poses challenges for precise measurement (Dewhurst et al. 2000). It's recognized that the availability and coordination of energy and nitrogen in the rumen are crucial factors affecting microbial protein synthesis (Zhang et al. 2020). The efficiency in the synthesis of Nmic expressed as kg of digestible organic matter (DOM) increased with the DM intake (Broderick et al. 2010), and therefore, it would be expected the decrease in the efficiency as pregnancy advances.

In this study, there was no observed variation in this parameter as gestational days progressed. However, Moreira et al. (2023) reported that despite the decreasing DM intake during gestation, pregnant cows exhibited greater efficiency in microbial nitrogen production. In the current study, the variations in microbial N production during pregnancy were probably related to the effects of the feeding level. In research involving beef cows, prepartum non-supplementation was found to enhance the efficiency of Nmic synthesis (g/ kg of DOM), possibly due to more efficient nitrogen utilization (Ferreira et al. 2020). Similarly, LOW-fed cows demonstrated higher efficiency in Nmic synthesis when expressed in g/kg of DOM and g/ kg of N.

Moreover, the improved efficiency in microbial nitrogen observed in LOW-fed cows is likely attributed to their extended ruminal retention time (Bach et al. 2005). In conditions where the passage rate is high, a reduction in microbial maintenance costs is expected due to a decrease in ruminal retention time. The passage rate is influenced by the level of dry matter intake, but this assumption holds primarily when the passage rate changes in response to variations in feed quality. In other words, a positive correlation exists between the passage rate and the digestibility of the feed. In instances where the passage rate decreases, as observed in this study due to feed restriction, it is suggested that the prolonged ruminal retention time facilitates greater ruminal degradation of the feed, thereby leading to increased microbial N production per unit of feed.

In summary, this study illustrates the intricate relationship between nutrition, gestational progression, and rumen dynamics in pregnant cows. Our findings suggest that the nutritional plan during gestation directly influences the cows' ability to maintain their body condition and support fetal growth. In practical terms, these results show the importance of adopting nutritional plans for pregnant cows, especially during late gestation when fetal growth accelerates. Additionally, our findings reveal that the decrease in DM intake, linked to the pressure of the pregnant uterus and visceral fat during pregnancy in cows fed ad libitum, triggers a higher passage rate as a compensatory response to rumen compression. Nevertheless, this shorter feed retention time leads to reduced nutrient digestibility as pregnancy advances, impacting feed utilization. To provide comprehensive guidance for managing pregnant cows, future research should aim to integrate the complex interactions among feed quality and rumen dynamics. Such efforts will lead to more effective nutritional strategies benefiting both producers and the cattle industry as a whole.