Introduction
The early lactation period is a critical period for dairy cows where energy demands for milk production cannot be met by energy intake alone, resulting in a negative energy balance (NEB) (van Knegsel et al., Reference Van Knegsel, Van Den Brand, Graat, Dijkstra, Jorritsma, Decuypere, Tamminga and Kemp2007). Dairy cows in NEB mobilise body fat to compensate for this energy deficit, which leads to increased concentrations of non-esterified fatty acids (NEFA) in blood (Gross et al., Reference Gross, van Dorland, Bruckmaier and Schwarz2011). When the capacity of the liver to fully oxidise these NEFA is exceeded, remaining NEFA are stored as triglycerides in the liver or are converted into ketone bodies, such as β-hydroxybutyrate (BHB), that enter circulation (Laffel, Reference Laffel1999, Gross et al., Reference Gross, van Dorland, Bruckmaier and Schwarz2011). Elevation of BHB concentration in blood may exceed the normal blood buffering capacity, resulting in decreased blood pH, hence causing metabolic acidosis (Owens et al., Reference Owens, Secrist, Hill and Gill1998; Laffel, Reference Laffel1999). During the early lactation period, large amounts of readily fermentable carbohydrates are usually fed in an attempt to maximise energy intake (Allen et al., Reference Allen, Bradford and Oba2009). This increases the risk of high outflow of fermentable substrate, in particular starch, from the rumen into the intestines (Gressley et al., Reference Gressley, Hall and Armentano2011). Carbohydrate digestion is limited in the small intestine (Matthé et al., Reference Matthé, Lebzien, Hric, Flachowsky and Sommer2001), and thus the flow of incompletely degraded and digested substrates into the hindgut may increase, resulting in fermentation of these substrates (Van Gastelen et al., Reference Van Gastelen, Dijkstra, Gerrits, Gilbert and Bannink2023) and potentially hindgut acidosis. Accumulation of fermentation acids in the hindgut decreases digesta pH and has been related to dysbiosis in the microbial community and epithelial damage (Gressley et al., Reference Gressley, Hall and Armentano2011; Plaizier et al., Reference Plaizier, Danesh Mesgaran, Derakhshani, Golder, Khafipour, Kleen, Lean, Loor, Penner and Zebeli2018). Van Gastelen et al. (Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a) suggested that susceptibility to develop metabolic acidosis became higher when hindgut acidosis also occurs. It is interesting to determine if the reciprocal relation may exist and how the fermentation characteristics of the hindgut microbiome will be affected during incidences of hindgut acidosis and metabolic acidosis, or by the combined effects of both types of acidosis. Although fermentation in the hindgut typically provides only 5 to 10% of the dietary energy supply (Gressley et al., Reference Gressley, Hall and Armentano2011), understanding fermentation characteristics of hindgut microbiome during incidences of hindgut acidosis and metabolic acidosis is important to improve overall animal health and production.
The in vitro cumulative gas production (GP) technique is one of the techniques used to study fermentation kinetics (Yáñez-Ruiz et al., Reference Yáñez-Ruiz, Bannink, Dijkstra, Kebreab, Morgavi, O’Kiely, Reynolds, Schwarm, Shingfield, Yu and Hristov2016). Many studies have used the in vitro cumulative GP technique with rumen fluid as source of inoculum, although some studies have shown the potential of faeces as an alternative source of inoculum (Dhanoa et al., Reference Dhanoa, France, Crompton, Mauricio, Kebreab, Mills, Sanderson, Dijkstra and López2004; Zicarelli et al., Reference Zicarelli, Calabrò, Cutrignelli, Infascelli, Tudisco, Bovera and Piccolo2011). Compared with rumen fluid, fewer microorganisms may be present in faecal inocula potentially resulting in a slower rate of GP, a longer lag phase and a lower VFA production (Dhanoa et al., Reference Dhanoa, France, Crompton, Mauricio, Kebreab, Mills, Sanderson, Dijkstra and López2004; Zicarelli et al., Reference Zicarelli, Calabrò, Cutrignelli, Infascelli, Tudisco, Bovera and Piccolo2011; Chiaravalli et al., Reference Chiaravalli, Rapetti, Graziosi, Galassi, Crovetto and Colombini2019). At present, few studies conducted in an in vitro setting focused on digestion kinetics under different conditions of ruminal pH (Grant and Mertens, Reference Grant and Mertens1992; Calsamiglia et al., Reference Calsamiglia, Ferret and Devant2002; Palladino et al., Reference Palladino, Wawrzkiewicz, Danelón, Gaggiotti and Jaurena2010). To the best of our knowledge, there are no studies which investigated the effect of metabolic or hindgut acidifying treatments on hindgut microbial fermentation kinetics and end-fermentation products obtained by in vitro GP. Therefore, the aim of this study was to investigate in vitro the effect of abomasal infusion of corn starch and BHB, on hindgut microbial fermentation characteristics of early lactation dairy cows. We expected both abomasal infusion of corn starch and BHB to decrease fermentation activity of hindgut microbiome, based on observations in the study from which we used donor animals for faecal inocula, as reported by van Gastelen et al. (Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a). In that study, abomasal corn starch infusion to induce hindgut acidosis led to decreased faecal pH, decreased faecal molar proportions of propionic acid and of branched-chain fatty acids, and increased molar proportion of butyric acid. Abomasal infusion of BHB to induce compensated metabolic acidosis led to blood acid-base disturbance, increased intestinal permeability, and decreased faecal molar proportion of acetic acid but increased proportion of butyric acid.
Materials and methods
Donor animals of faecal inoculum and experimental diets
The in vitro experiment was conducted simultaneously with an in vivo trial conducted at the animal research facilities of Wageningen University & Research (Wageningen, the Netherlands) under the Dutch Law on Animal Experiments in accordance with European Union Directive 2010/63, and reported in full by van Gastelen et al. (Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a). In brief, the in vivo experiment had a 6 × 6 Latin square design with six rumen fistulated lactating Holstein-Friesian cows, second or third parity, on average 66 ± 18 days in milk at the start of the trial. Each of the six experimental periods consisted of 7 d, viz. 5 d of continuous abomasal infusion followed by 2 d of rest. Abomasal infusion treatments were 1) water only (no corn starch or BHB) as control, 2) 1.5 kg corn starch/d + 0.0 mol BHB/d, 3) 3.0 kg corn starch/d + 0.0 mol BHB/d, 4) 0.0 kg corn starch/d + 8.0 mol BHB/d, 5) 1.5 kg corn starch/d + 8.0 mol BHB/d, and 6) 3.0 kg corn starch/d + 8.0 mol BHB/d. Treatments were infused at a constant infusion rate between 0.9 and 1.0 kg of solution/h. Cows were fed a total mixed ration consisting of 350 g/kg grass silage, 374 g/kg corn silage, and 276 g/kg concentrate on dry matter (DM) basis. Cows were restricted to 90% of individual daily ad libitum intake. During the final 60 h of each 5 d infusion period, cows were fed equal portions of feed every 2 h. Water was available ad libitum. Further details of infusion treatments and procedures are reported by van Gastelen et al. (Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a).
In vitro gas production trial
Substrates
Four feedstuffs with a high content of either fibre or starch, with each carbohydrate source differing in expected fermentation rate (slow or rapid), were used as substrates in the in vitro trial. Fibre-rich sources were cellulose (expected to be slowly fermentable) and beet pulp (expected to be rapidly fermentable), and the starch-rich sources were corn grain (expected to be slowly fermentable) and pregelatinized corn flour (expected to be rapidly fermentable) (Table 1). Corn grain was obtained locally (Rijnvallei, Wageningen, the Netherlands), pregelatinized corn flour (UniFlour 32501) from Limagrain Ingredients (Rilland, The Netherlands), cellulose (Arbocel B800) from J. Rettenmaier & Soehne GmbH + Co (Zutphen, The Netherlands), and beet pulp (Pavo SpeediBeet) from Pavo (Boxmeer, The Netherlands). Corn grain and beet pulp were ground to pass through a 1 mm mesh sieve using a Wiley mill (Peppink 100 AN, Olst, The Netherlands), whereas pregelatinized corn flour and cellulose were already in meal form upon purchase.
Chemical composition (g/kg DM, unless stated otherwise) of the feedstuffs used as substrate in the in vitro gas production
DM, dry matter; CP, crude protein; NDF, neutral detergent fibre.
Collection and preparation of inocula
On d-6 of each experimental period, immediately after ending the abomasal infusion, faeces were collected directly from the rectum of each cow by grab sampling. Cows were sampled in the same order in each experimental week. Faecal samples were divided into two subsamples. One subsample was placed in a CO2 pre-flushed, pre-warmed thermos flask as inoculum for the in vitro GP, and the other subsample was placed in a plastic bottle without CO2 flushing for analysis of chemical composition, pH, VFA, and ammonia. Faecal pH was immediately measured with a calibrated pH metre (Model HI 9024; Hanna instruments, IJsselstein, The Netherlands) before further sample processing. Inocula were prepared in the same sequence as faecal sampling. An amount of 480 g of faeces was mixed with 960 ml CO2 pre-flushed and warmed buffer/mineral solution (1:2 v/v), as based on Cone et al. (Reference Cone, Van Gelder, Visscher and Oudshoorn1996). Per 960 ml buffer, 8.40 g NaHCO3, 0.96 g NH4HCO3, 1.37 g Na2HPO4, 1.49 g KH2PO4, 0.14 g MgSO4.7H2O, 0.15 g Na2S.3H2O, 0.080 g NaOH, 10.6 mg CaCl2.2H2O, 8.0 mg MnCl2.4H2O, 0.80 mg CoCl2.6H2O, 6.4 mg FeCl3.6H2O, 0.62 g hydrolyzed casein, and 0.60 mg resazurin was included. This was followed by homogenisation of the mixture for 60 s with a hand mixer. The mixture was strained using a double layer of cheesecloth. All procedures were performed under a constant CO2 flow to ensure anaerobic conditions.
In vitro fermentation, gas and methane measurement
In vitro GP was determined using fully automated equipment, as described by Cone et al. (Reference Cone, Van Gelder, Visscher and Oudshoorn1996) with modifications (Pellikaan et al., Reference Pellikaan, Hendriks, Uwimana, Bongers, Becker and Cone2011) in order to measure CH4 production. Each substrate (0.5 g) was incubated together with 60 ml of the buffer inoculum mixture in 250 ml bottles (Schott, Mainz, Germany) for 72 h. Per cow and per experimental period, each substrate was analysed in triplicate. The bottles were placed in a shaking water bath with 50 movements/min at a temperature of 39°C (Haake SWB25, Clausthal-Zellerfeld, Germany). All other procedures were conducted as described by Cone et al. (Reference Cone, Van Gelder, Visscher and Oudshoorn1996). During incubation, the headspace of each bottle was sampled (10 μl) using a gas tight syringe (Hamilton 1701 N Point Style 5 needle 51 mm) at 0, 3, 6, 9, 12, 24, 30, 36, 48, 60, and 72 h of incubation, and immediately injected into a gas chromatograph (GC; GC 8000 top CE Instruments, Milan, Italy) fitted with a flame ionisation detector to determine the CH4 concentration as described by Ellis et al. (Reference Ellis, Bannink, Hindrichsen, Kinley, Pellikaan, Milora and Dijkstra2016).
Chemical analyses
Fermentation was terminated after 72 h of incubation. Immediately, pH of the incubation fluid was measured using a calibrated pH metre (Hanna instruments, IJsselstein, The Netherlands). The incubation fluid was centrifuged at 500 × g for 10 min and then a sample (600 μl) of the supernatant was put in a 10% trichloroacetic acid solution (1:1, v/v) for ammonia analysis. Another sample (600 μl) of the supernatant was acidified with 600 μl of 0.85 M ortho-phosphoric acid containing 19.68 mM isocaproic acid as an internal standard for VFA analysis. Samples were stored at –20°C until analyses. The remaining bottle content was filtered and washed three times with warm distilled water in a pre-weighed glass filter crucible which was oven dried for 16 h at 70°C followed by additional drying at 103°C for 4 h (ISO 6496, 1999) and incineration at 530°C (ISO 5984, 2002) to determine organic matter (OM) content of the residue. In vitro OM digestibility (IVOMD) was calculated as the difference between amount of OM incubated and amount of OM in the residue, and expressed as a proportion of OM incubated. Fresh faeces as well as 72 h incubation fluid were analysed for VFA according to the method described by Hatew et al. (Reference Hatew, Cone, Pellikaan, Podesta, Bannink, Hendriks and Dijkstra2015) using gas chromatography (Fisons HRGC MEGA2, Milan, Italy). Ammonia was analysed using a spectrophotometer (Ultrospec 500 pro, Amersham-Bioscience, Barcelona, Spain) at 625 nm wavelength and indophenol colorimetric absorbance. Feed substrates and faeces samples were analysed for DM (ISO 6496, 1999), ash (ISO 5984, 2002), nitrogen (N) (ISO 5983, 2005), and starch (ISO 15914, 2004). Nitrogen was measured by the Kjeldahl method and crude protein (CP) content was calculated as N × 6.25. The neutral detergent fibre (NDF) contents of the substrates were analysed in accordance with the method described by van Soest et al. (Reference Van Soest, Robertson and Lewis1991) after pre-treatment with amylase.
Calculations and statistical analysis
Graphical representation of GP was used to identify a potential outlier between the replicates per treatment-substrate combination (n = 3). An outlier (72 out of 432 incubation bottles) was removed if the variation coefficient of the three replicates for cumulative GP at each incubation time was more than 0.30. The outlier samples were removed from the dataset for further analyses. Per run, data from bottles assigned to the same treatment and substrate (n = 3 if no outliers) were averaged. All data (chemical composition of faeces; total gas and CH4 production per individual incubation time; IVOMD, VFA and ammonia at 72 h of incubation) were analysed using Proc MIXED (SAS 9.4; SAS Institute Inc., Cary, NC). The model contained main effects and interaction effects of infusion treatment factors (corn starch and BHB) as fixed effects with experimental period and cow included as random effects. Differences were considered significant at P ≤ 0.05 and tendencies at 0.05 < P ≤ 0.10. Multiple comparisons between treatments were made using the Tukey-Kramer method when a starch effect or starch × BHB interaction was detected at P ≤ 0.05.
Results
Chemical composition of feed substrates and faeces
The NDF content of the incubated substrates varied between 18 and 449 g/kg DM and the starch content varied between 4 and 757 g/kg DM (Table 1). Chemical composition, pH, ammonia concentration, total VFA (TVFA) concentration, and the molar proportions of individual VFA of faeces used to prepare the inoculum for in vitro fermentation were not affected by a starch × BHB interaction (Table 2).
Chemical composition (g/kg DM, unless stated otherwise), pH, ammonia concentration, total volatile fatty acids (TVFA) concentration and molar proportion of TVFA, of dairy cattle faeces used to prepare inocula for the in vitro incubation trial. Cows received abomasal infusions of corn starch (0.0, 1.5 or 3.0 kg/d), β-hydroxybutyrate (BHB; 0.0 or 8.0 mol/d), or combinations of both
BHB, β-hydroxybutyrate; SEM, standard error of the mean; DM, dry matter; N, nitrogen; NDF, neutral detergent fibre; TVFA, total volatile fatty acids; VFA, volatile fatty acids.
Regardless of BHB infusion, infusion of 3.0 kg corn starch/d decreased faecal NDF content (P < 0.001) and increased faecal starch (P < 0.001) content compared with 0.0 and 1.5 kg corn starch/d. Faecal pH decreased (P < 0.001) with each level of corn starch infused. The molar proportion of propionate tended (P = 0.100) to decrease with corn starch infusion, and the molar proportions of both isobutyrate and isovalerate decreased (P ≤ 0.034) when 3.0 kg corn starch/d was infused relative to 0.0 kg corn starch/d. The molar proportions of the other VFA, as well as TVFA and ammonia concentrations were not affected by starch infusion.
Regardless of corn starch infusion, infusion of BHB resulted in a decreased NDF content in the faeces (P = 0.029). Faecal pH tended to decrease with BHB infusion (P = 0.101). Faecal ammonia and TVFA concentrations, and the molar proportions of the individual VFA were not affected by BHB infusion.
In vitro gas production
The cumulative GP of all four substrates was not affected by a starch × BHB interaction at any incubation time, with the exception of GP of cellulose at 48 h of incubation (tendency only; P = 0.097). In general, depending on the incubation time, cumulative GP of all feed substrates was affected by infusion of starch regardless of BHB infusion (Table 3; Figure 1). The cumulative GP of cellulose was greater (P ≤ 0.001) at 3, 6, 9, and 12 h of incubation and lower (P ≤ 0.016) at 48 and 72 h of incubation with infusion of 3.0 kg corn starch/d compared with both 0.0 and 1.5 kg corn starch/d. The cumulative GP of beet pulp was greater (P = 0.003) at 3 h of incubation and lower (P = 0.013) at 24 h of incubation when 3.0 kg corn starch/d was infused compared with both 0.0 and 1.5 kg corn starch/d infusion. Infusion of 1.5 and 3.0 kg corn starch/d decreased the cumulative GP of beet pulp at 9 and 12 h of incubation (P ≤ 0.017) compared with 0.0 kg corn starch/d infusion. Infusion of 3.0 kg corn starch/d increased cumulative GP of corn grain at 3 and 6 h of incubation (P ≤ 0.004) compared with 0.0 and 1.5 kg corn starch/d. Corn starch infusion did not affect cumulative GP of corn grain at later incubation times. Compared with infusion of 0.0 corn starch/d, infusion of 3.0 kg corn starch/d increased cumulative GP at 3 h (P = 0.032) and decreased cumulative GP of pregelatinized corn flour at 9 h until 48 h of incubation (P ≤ 0.057).
In vitro gas production (ml/g organic matter incubated) at different timepoints of cellulose, beet pulp, corn grain and pregelatinized corn flour incubated with faecal inoculum of dairy cows. Cows received abomasal infusions of corn starch (0.0, 1.5 or 3.0 kg/d), β-hydroxybutyrate (BHB; 0.0 or 8.0 mol/d), or combinations of both
SEM, standard error of the mean.
In vitro cumulative gas production (ml/g OM incubated) of (a) cellulose, (b) beet pulp, (c) corn grain, and (d) pregelatinized corn flour incubated with faecal inoculum of dairy cows. Cows received abomasal infusions of water (no corn starch and BHB) (
; 
), 1.5 kg starch/d and 0.0 mol BHB/d (
; 
), 3.0 kg starch/d and 0.0 mol BHB/d (
; 
), 0.0 kg starch/d and 8.0 mol BHB/d (
; 
), 1.5 kg starch/d and 8.0 mol BHB/d (
; 
), and 3.0 kg starch/d and 8.0 mol BHB/d (
; 
). Bars at the top of the figure = SEM. OM, organic matter; BHB, β-hydroxybutyrate; SEM, standard error of the mean.
Regardless of corn starch infusion, BHB infusion did not affect cumulative GP of any feed substrate. Only for cellulose, cumulative GP tended to decrease (P = 0.063) at 72 h of incubation when 8.0 mol BHB/d was infused compared with no BHB infusion.
In vitro methane production
Methane production (MP) was not affected by a starch × BHB interaction, but tended (P ≤ 0.098) to be affected by a starch × BHB interaction for beet pulp at 9 and 12 h of incubation and for corn grain at 24 h of incubation (Table 4). Regardless of BHB infusion, MP decreased or tended to decrease upon starch infusion at the later stages of incubation time for all feed substrates except cellulose. Infusion of 3.0 kg corn starch/d decreased or tended to decrease (P ≤ 0.054) MP of beet pulp at 9 and 12 h of incubation compared with 0.0 kg corn starch/d, at 24 h (P = 0.022) compared with 0.0 and 1.5 kg corn starch/d, and at 48 h (P = 0.040) compared with 1.5 kg corn starch/d. For corn grain, infusion of 3.0 kg corn starch/d decreased MP at 12 and 48 h (P ≤ 0.042) compared with 0.0 kg corn starch/d, and at 24 and 72 h (P ≤ 0.024) compared with both 0.0 and 1.5 kg corn starch/d. Pregelatinized corn flour incubated with 3.0 kg corn starch/d inoculum had lower MP from 24 h until 72 h of incubation (P ≤ 0.037) compared with 0.0 kg corn starch/d inoculum.
In vitro methane (CH4) production (ml/g of incubated OM) at different timepoints of cellulose, beet pulp, corn grain and pregelatinized corn flour incubated with faecal inoculum of dairy cows. Cows received abomasal infusions of corn starch (0.0, 1.5 or 3.0 kg/d), β-hydroxybutyrate (BHB; 0.0 or 8.0 mol/d), or combinations of both
SEM, standard error of the mean.
Regardless of corn starch infusion, BHB infusion did not affect MP for any of the substrates at any incubation time. Only with beet pulp as substrate, starch × BHB interactions (P ≤ 0.050) were observed for methane production as a fraction of total GP (fMP) at 3, 6, and 9 h of incubation (Table 5; Figure 2). At 3 h, 3.0 kg corn starch/d infusion resulted in lower fMP compared with 0.0 kg corn starch/d, but only when BHB was infused, and furthermore infusion of 3.0 kg corn starch/d in presence of BHB resulted in lower fMP compared with 1.5 kg corn starch/d in absence of BHB. At 6 and 9 h of incubation, fMP was lower with 3.0 kg corn starch/d infusion in presence of BHB compared with 1.5 kg corn starch/d infusion in absence of BHB. Regardless of BHB infusion, fMP of cellulose was lower with 3.0 kg corn starch/d infusion compared with 0.0 or 1.5 kg corn starch/d at 3 h (P = 0.006), and lower (P ≤ 0.053) compared with 1.5 kg/d corn starch at 6, 9, and 12 h of incubation. The fMP of beet pulp was lower (P ≤ 0.047) with 3.0 kg corn starch/d compared with 0.0 or 1.5 kg corn starch/d at 24, 48, and 72 h of incubation. Corn grain fMP was lower (P = 0.002) at 3 h for 1.5 and 3.0 kg corn starch/d compared with 0.0 kg corn starch/d, and decreased (P ≤ 0.048) with 3.0 kg corn starch/d at 6, 9, and 24 h compared with 0.0 kg corn starch/d. At 48 and 72 h, fMP of corn grain was lower (P ≤ 0.007) for 3.0 kg corn starch/d compared with 0.0 and 1.5 kg corn starch/d. Corn starch infusion did not affect fMP of pregelatinized corn flour, except for a tendency (P = 0.074) at 3 h of incubation. Regardless of corn starch infusion, BHB infusion did not affect fMP of any feed substrate.
Calculated methane production (fraction of total gas production) at different timepoints of cellulose, beet pulp, corn grain and pregelatinized corn flour incubated with faecal inoculum of dairy cows. Cows received abomasal infusions of corn starch (0.0, 1.5 or 3.0 kg/d), β-hydroxybutyrate (BHB; 0.0 or 8.0 mol/d), or combinations of both
SEM, standard error of the mean.
Methane production (fraction of total gas production) during in vitro fermentation of (a) cellulose, (b) beet pulp, (c) corn grain, and (d) pregelatinized corn flour incubated with faecal inoculum of dairy cows. Cows received abomasal infusions of water (no corn starch and BHB) (
; 
), 1.5 kg starch/d and 0.0 mol BHB/d (
; 
), 3.0 kg starch/d and 0.0 mol BHB/d (
; 
), 0.0 kg starch/d and 8.0 mol BHB/d (
; 
), 1.5 kg starch/d and 8.0 mol BHB/d (
; 
), and 3.0 kg starch/d and 8.0 mol BHB/d (
; 
). Bars at the top of the figure = SEM. BHB, β-hydroxybutyrate; SEM, standard error of the mean.
Organic matter digestibility, pH, volatile fatty acids, and ammonia
The IVOMD after 72 h of incubation of corn grain tended (P = 0.064) to be affected and IVOMD of beet pulp was affected (P = 0.014) by a starch × BHB interaction (Table 6). The latter interaction indicates a lower IVOMD with 3.0 kg corn starch/d compared with both 0.0 and 1.5 kg corn starch/d, but only in presence of BHB infusion. The pH after 72 h of incubation was not affected by starch × BHB interaction with any substrate, except for beet pulp (P = 0.042). With beet pulp, a more pronounced decline in pH was observed with 3.0 kg corn starch/d compared with 0.0 or 1.5 kg corn starch/d in presence of BHB infusion compared with absence of BHB infusion. Additionally, a lower pH was observed with 8.0 mol BHB/d compared with no BHB infusion, but only at the 3.0 kg corn starch/d infusion level.
In vitro organic matter digestibility (IVOMD) (fraction of incubated OM), pH, and ammonia concentration (mM) after 72 h fermentation of cellulose, beet pulp, corn grain and pregelatinized corn flour incubated with faecal inoculum of dairy cows. Cows received abomasal infusions of corn starch (0.0, 1.5 or 3.0 kg/d), β-hydroxybutyrate (BHB; 0.0 or 8.0 mol/d), or combinations of both
SEM, standard error of the mean.
Regardless of BHB infusion, infusion of 3.0 kg corn starch/d decreased (P < 0.001) IVOMD of cellulose and for corn grain (P = 0.019) compared with both 0.0 and 1.5 kg corn starch/d. With cellulose, pH after 72 h of incubation was lower (P = 0.010) for both 1.5 and 3.0 kg corn starch/d relative to 0.0 corn starch/d. Infusion of 3.0 kg corn starch/d decreased pH after 72 h of incubation for corn grain (P < 0.001) and for pregelatinized corn flour (P < 0.001) compared with both 0.0 and 1.5 kg corn starch/d. Infusion of 3.0 kg corn starch/d decreased (P ≤ 0.018) ammonia concentration of all feed substrates compared with both 0.0 and 1.5 kg corn starch/d. Regardless of corn starch infusion, BHB infusion tended to decrease ammonia concentrations for all feed substrates (P ≤ 0.103). Infusion of 8.0 mol BHB/d also tended (P ≤ 0.062) to decrease the pH after 72 h of incubation of beet pulp and corn grain compared with 0.0 mol BHB/d.
The TVFA concentration and molar proportions of individual VFA after 72 h of incubation were not affected by a starch × BHB interaction with any substrate (Table 7). Regardless of BHB infusion, the TVFA concentration with cellulose was greater (P = 0.047) with 1.5 kg corn starch/d compared with 0.0 kg corn starch/d. The TVFA concentrations with beet pulp and corn grain were greater (P = 0.001) with both 1.5 and 3.0 kg corn starch/d compared with 0.0 kg corn starch/d infusion. The molar proportion of acetate with cellulose was greater (P = 0.034) with 0.0 kg corn starch/d compared with 3.0 kg corn starch/d. For beet pulp, molar proportion of acetate was greater (P < 0.001) with 0.0 and 1.5 kg corn starch/d compared with 3.0 kg corn starch/d. Molar proportion of acetate with corn grain decreased (P < 0.001) with each level of corn starch infused. Molar proportion of propionate with beet pulp and corn grain increased (P < 0.001) with each level of corn starch infusion. For pregelatinized corn flour, molar proportion of propionate was greater (P = 0.029) with 3.0 kg corn starch/d compared with 0.0 kg corn starch/d. Molar proportions of butyrate with cellulose, beet pulp and corn grain were greater (P ≤ 0.006) with 3.0 kg corn starch/d compared with 0.0 and 1.5 kg corn starch/d. Molar proportion of isobutyrate with cellulose was greater (P = 0.018) with 0.0 kg corn starch/d compared with 3.0 kg corn starch/d. For beet pulp, corn grain and pregelatinized corn flour, molar proportions of isobutyrate decreased (P < 0.001) with each level of corn starch infusion. Molar proportion of valerate with cellulose, beet pulp and corn grain decreased (P ≤ 0.028) with 3.0 kg corn starch/d compared with 0.0 and 1.5 kg corn starch/d. Molar proportion of isovalerate with cellulose decreased (P = 0.030) with 3.0 kg corn starch/d compared with 0.0 kg corn starch/d. For beet pulp, corn grain and pregelatinized corn flour, molar proportion of isovalerate decreased (P < 0.001) with each level of corn starch. Acetate to propionate molar ratio (A/P) of beet pulp decreased (P < 0.001) with each level of corn starch. Infusion of 1.5 kg and 3.0 kg corn starch/d decreased (P < 0.001) A/P of corn grain compared with 0.0 kg corn starch/d. For pregelatinized corn flour, A/P decreased (P = 0.005) with 3.0 kg corn starch/d compared with 0.0 kg corn starch/d.
Total volatile fatty acids (TVFA), molar proportions of TVFA and acetate to propionate ratio (A/P) after 72 h of in vitro fermentation of cellulose, beet pulp, corn grain and pregelatinized corn flour incubated with faecal inoculum of dairy cows. Cows received abomasal infusions of corn starch (0.0, 1.5 or 3.0 kg/d), β-hydroxybutyrate (BHB; 0.0 or 8.0 mol/d), or combinations of both
SEM, standard error of the mean; VFA, volatile fatty acids.
Regardless of corn starch infusion, BHB infusion did not affect TVFA concentration after 72 h of incubation of any feed substrate. With beet pulp, BHB infusion decreased molar proportion of isobutyrate (P = 0.018) and isovalerate (P = 0.028) and tended to decrease (P = 0.076) molar proportion of valerate. For corn grain, molar proportion of butyrate increased (P = 0.030) and molar proportion of isobutyrate decreased (P = 0.045), and molar proportion of isovalerate tended to decrease (P = 0.079) when infusing 8.0 mol BHB/d compared with 0.0 mol BHB/d.
Discussion
The present study was conducted in order to investigate the effects of abomasal infusion of corn starch and BHB on hindgut microbial fermentation capacity by comparing in vitro GP and MP, fermentation endpoint pH, IVOMD, ammonia concentration, and VFA profiles. In the cows from which faecal inocula were obtained, abomasal corn starch infusion and BHB infusion induced hindgut and compensated metabolic acidosis, respectively. Abomasal infusion of corn starch decreased total tract digestibility of several nutrients and faecal pH and affected the molar proportions of almost all individual VFA (van Gastelen et al., Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a). Abomasal infusion of BHB did not affect faecal pH or total VFA, but did result in increased intestinal permeability associated with decreased total tract digestibility of protein and fat, and increased butyric acid molar proportion in faeces and a tendency to decreased acetic acid molar proportion in faeces. These infusion treatments were thus expected to also affect the microbial composition in the hindgut and its capacity to ferment feed substrates. However, the in vitro results show that GP, MP, fermentation endpoint pH and VFA characteristics of several feed substrates were mainly affected by high starch infusion (3.0 kg corn starch/d) in particular, whereas the effects of BHB infusion were limited or absent. This suggests that, in contrast to starch, BHB did not substantially alter microbial fermentation capacity in the hindgut. These observations support the idea that BHB may act primarily through non-fermentative mechanisms, modulating host physiology or microbial signalling rather than directly contributing to microbial metabolism (Rico and Barrientos-Blanco, Reference Rico and Barrientos-Blanco2024). Nevertheless, greater butyric acid levels in the hindgut, as observed in vivo, may still influence intestinal microbiota composition and activity, potentially supporting a healthy gut environment (Bedford and Gong, Reference Bedford and Gong2018). The increased level of faecal butyrate with BHB infusion reported by van Gastelen et al. (Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a) was not observed in the present study, which might be related to the sampling frequency. Faecal samples were obtained using twice daily sampling by van Gastelen et al. (Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a), whereas the inoculum in the present experiment was obtained based on one sample only. The lack of BHB infusion effects on GP and MP, and the limited effect of BHB infusion on fermentation endpoint pH and VFA profiles in the present in vitro experiment, indicate that the amount of BHB infused into the abomasum hardly affected fermentation capacity of the hindgut of dairy cattle. Therefore, only the effects of starch infusion are discussed below.
Infusion of starch into the abomasum at the highest level likely exceeded small intestinal digestion capacity, and starch fermentation in the hindgut did not fully compensate for this, as evidenced by the higher faecal starch concentration and lower faecal pH. Faecal pH at 1.5 kg corn starch/d was also lower than at 0.0 kg corn starch/d, in line with results reported for these cows (van Gastelen et al., Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a). In agreement with our abomasal infusion model to induce hindgut acidosis, Reynolds et al. (Reference Reynolds, Cammell, Humphries, Beever, Sutton and Newbold2001) reported reduced faecal pH (6.64 to 6.26) in lactating dairy cows abomasally infused with wheat starch (1.2 kg/d). Additionally, van Gastelen et al. (Reference Van Gastelen, Dijkstra, Nichols and Bannink2021 b) reported faecal pH to decrease from 6.86 (control treatment) to 6.00 when 3.0 kg ground corn/d (∼1.5 kg starch/d) was abomasally infused. In the present study, low faecal pH at 3.0 kg corn starch/d was not accompanied by increased TVFA concentration, unlike findings from a grain-based acidosis challenge (Li et al., Reference Li, Khafipour, Krause, Kroeker, Rodriguez-Lecompte, Gozho and Plaizier2012) and from abomasal ground corn infusion (van Gastelen et al., Reference Van Gastelen, Dijkstra, Nichols and Bannink2021 b). However, our results align with the accompanying in vivo study by van Gastelen et al. (Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a), who also observed a lower faecal pH without increased TVFA concentration in faeces upon abomasal infusion of corn starch, attributed to altered buffering capacity of the large intestine. The decreased molar proportions of isobutyrate and isovalerate in faeces at 3.0 kg corn starch/d suggest decreased protein fermentation (despite unchanged ammonia concentration) and increased microbial protein synthesis (Andries et al., Reference Andries, Buysse, De Brabander and Cottyn1987).
In general, at 3 h of incubation, cumulative GP was highest with inoculum of cows receiving 3.0 kg corn starch/d, regardless of feed substrate type. This may reflect increased microbial activity, possibly in combination with the higher faecal starch content providing substrate for early commencement of fermentation. Nagadi et al. (Reference Nagadi, Herrero and Jessop2000) reported increased initial microbial concentration in the ruminal fluid as well as increased volume and rate of GP with higher dietary concentrate-to-hay ratios. Greater faecal starch content with 3.0 kg corn starch/d therefore likely increased substrate content in faeces available for microbes, particularly amylolytic species, potentially enhancing initial GP. In their review, Yáñez-Ruiz et al. (Reference Yáñez-Ruiz, Bannink, Dijkstra, Kebreab, Morgavi, O’Kiely, Reynolds, Schwarm, Shingfield, Yu and Hristov2016) discussed the microbial diversity and activity in rumen fluid to be lowest immediately before feeding, with subsequent impact of faecal sampling time, in particular, on the rate of fermentation and GP. Desrousseaux et al. (Reference Desrousseaux, Santos, Pellikaan, Van der Poel, Cone, Guedes, Ferreira and Rodrigues2012) found that faecal sampling time relative to feeding time influenced fermentative activities of equine faecal bacteria, with the highest asymptotic GP with faecal inocula collected two hours after the morning feeding. In the present study, faecal samples were collected at 0900 h from cows that for a 60-h period had been fed equal portions of feed every 2 h. Whether this faecal collection time in combination with continuous feeding contributed to an increase in microbial diversity and activities is speculation. As we did not measure microbial concentration, we ultimately cannot conclude whether the greater early cumulative GP resulted from higher starch content, greater microbial mass in the inoculum, or both.
A greater cumulative GP with inoculum of cows abomasally infused with 3.0 kg corn starch/d was present only at 3 h of incubation (fast degradable substrates; beet pulp and pregelatinized corn flour), up to 6 h of incubation (slow degradable starch substrate; corn grain) or up to 12 h of incubation (slow degradable fibre substrate; cellulose). Beyond 48 to 72 h, GP with 3.0 kg corn starch/d was similar to or lower than GP of 0.0 and 1.5 kg corn starch/d. In a grain-based subacute rumen acidosis challenge (Plaizier et al., Reference Plaizier, Li, Danscher, Derakshani, Andersen and Khafipour2017), amylolytic bacteria dominated at the expense of fibrolytic bacteria; similar patterns upon abomasal starch infusion were expected in this study, resulting in a greater ability to degrade starch substrates at the expense of fibrolytic capacity. In line with expectation, IVOMD of the slow degradable, fibre-rich cellulose was the most affected by 3.0 kg corn starch/d infusion, reflecting lower hindgut microbial capacity to ferment cellulose. In contrast to expectations, IVOMD of starch-rich corn grain also tended to decrease with infusion of 3.0 kg starch/d, likely due to reduced fermentation of its non-starch polysaccharides, particularly NDF. Fermentation of cellulose with 3.0 kg corn starch/d inoculum incomplete, as the cumulative GP at 72 h had not reached its asymptote yet, whereas for the other substrates, the cumulative GP hardly increased from 48 to 72 h of incubation. This is consistent with the crystalline structure of cellulose resisting enzymatic hydrolysis (Weimer, Reference Weimer1996). Miranda-Romero et al. (Reference Miranda-Romero, Tirado-González, Tirado-Estrada, Améndola-Massiotti, Sandoval-González, Ramírez-Valverde and Salem2020) also observed that GP of cellulose peaked later (between 24 to 81 h of incubation) than glucose (between 0 to 8 h) and starch (between 8 to 24 h). The decreased cumulative GP and IVOMD for several substrates and time points with 3.0 kg corn starch/d did not correspond with the generally observed increased TVFA concentration nor with the lower pH after 72 h of incubation. In dual-flow continuous culture system experiments, Calsamiglia et al. (Reference Calsamiglia, Ferret and Devant2002) and Sari et al. (Reference Sari, Ferret and Calsamiglia2015) changed fermentation media pH by continuously adding buffer or acid, and observed TVFA concentration and OM digestibility to decrease as pH decreases. The greater TVFA concentration after 72 h of incubation with 3.0 kg corn starch/d infusion in this study may have been caused by a higher faecal starch content and, after straining with cheesecloth, fine particulate faecal starch may have been fermented in vitro. We expressed IVOMD per kg of incubated OM of substrate, not including the unknown quantity of OM entering fermentation bottles that may have originated from faecal inoculum. Thus, a part of TVFA may have arisen from faecal starch present in inoculum being fermented. Such an increase in TVFA concentration would have been expected to coincide with increased cumulative GP at 72 h of incubation, but we did not observe this. This lack of increased cumulative GP may stem from shifts in type of VFA formed. Except for substrate cellulose, 3.0 kg corn starch/d infusion increased the molar proportion of propionic acid. In contrast to acetic and butyric acid, the fermentation of glucose units to propionic acid does not yield CO2 directly, but yields CO2 only indirectly from release of gas (CO2) from the buffer (bicarbonate) used (Beuvink and Spoelstra, Reference Beuvink and Spoelstra1992). Besides, fermentation of glucose units to butyric acid rather than to acetic acid or propionic acid yields less CO2 indirectly (Beuvink and Spoelstra, Reference Beuvink and Spoelstra1992), and butyric molar proportion increased for all substrates except pregelatinized corn flour. Thus, increased VFA production, coupled with shifts towards propionic and butyric acid at the expense of acetic acid, may explain the lack of corresponding increase in GP.
At the end of the in vitro fermentation, 3.0 kg corn starch/d infusion reduced fMP for beet pulp and corn grain. Increasing dietary starch content or decreasing dietary NDF content has been found to decrease GP and MP (Getachew et al., Reference Getachew, Robinson, DePeters, Taylor, Gisi, Higginbotham and Riordan2005; Hatew et al., Reference Hatew, Cone, Pellikaan, Podesta, Bannink, Hendriks and Dijkstra2015; Maccarana et al., Reference Maccarana, Cattani, Tagliapietra, Schiavon, Bailoni and Mantovani2016). Van Kessel and Russell (Reference Van Kessel and Russell1996) suggested that low pH (< 6.0) inhibits methanogen activity, and at 72 h, all substrates incubated with inoculum from the 3.0 kg corn starch/d infusion treatment had pH < 6.0, potentially reducing CH4 production. Fermentation of faecal starch in the inoculum in this treatment may have increased production of propionate, a hydrogen sink linked to reduced methanogenesis (Johnson and Johnson, Reference Johnson and Johnson1995). This is supported by increased molar proportions of propionate after 72 h of incubation for all substrates (except cellulose). Such in vitro data on CH4 production under varying inoculum conditions also contribute to modelling approaches aimed at improving quantitative understanding and prediction of this significant greenhouse gas emitted by ruminants (Dijkstra et al., Reference Dijkstra, Bannink, Congio, Ellis, Eugène, Garcia, Niu, Vibart, Yáñez-Ruiz and Kebreab2025).
The decreased A/P ratio for all substrates except cellulose upon 3.0 kg corn starch/d infusion aligns with pH reductions observed in vivo (Li et al., Reference Li, Khafipour, Krause, Kroeker, Rodriguez-Lecompte, Gozho and Plaizier2012) and in vitro (Erfle et al., Reference Erfle, Boila, Teather, Mahadevan and Sauer1982; Mouriño et al., Reference Mouriño, Akkarawongsa and Weimer2001; Calsamiglia et al., Reference Calsamiglia, Ferret and Devant2002). Increased butyrate molar proportion (numerical for pregelatinized corn flour) match findings of increased faecal butyrate molar proportions upon abomasal infusion of corn starch and subsequent partial fermentation of this starch in the hindgut in the in vivo study (van Gastelen et al., Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a). Similar increases were observed in dairy cattle fed high-concentrate diets (Mao et al., Reference Mao, Zhang, Wang and Zhu2012) and in monogastric animals fermenting slowly degradable starch in the hindgut (Lv et al., Reference Lv, Li, Xing, Ma, Huang and Li2006). In an in vitro study using human faecal inoculum, the population of butyrate-producing bacteria Roseburia increased as inoculum pH decreased (Walker et al., Reference Walker, Duncan, McWilliam Leitch, Child and Flint2005). Lower molar proportions of the branched-chain VFA isobutyrate and isovalerate in all feed substrates incubated with 3.0 kg corn starch/d infusion indicates lower protein fermentation and/or increased microbial utilisation of these branched-chain VFA for biosynthesis of the branched amino acids valine, leucine, and isoleucine (Andries et al., Reference Andries, Buysse, De Brabander and Cottyn1987). Branched-chain VFA are the result of deamination of branch-chain amino acids. A reduction in protein fermentation at low pH has been reported (Erfle et al., Reference Erfle, Boila, Teather, Mahadevan and Sauer1982; Calsamiglia et al., Reference Calsamiglia, Ferret and Devant2002). Decreased protein fermentation, or increased microbial protein synthesis including utilisation of the branched chain amino acids, with high starch infusion is also supported by the decrease in ammonia concentration for all substrates. Lana et al. (Reference Lana, Russell and Van Amburgh1998) observed that also ruminal ammonia concentration decreased as in vitro pH decreased from 6.5 to 5.7, and this was due to a decrease in deamination rate by bacteria in inoculum obtained from cattle fed forage. The deamination rate of bacteria from cattle fed a diet containing 90% concentrate (DM), however, was not affected by lower pH, indicating that different bacterial populations had different ability to affect protein fermentation. Thus, reduced deamination is unlikely to explain the lower ammonia concentrations in the present study.
Conclusion
Abomasal infusion of 3.0 kg corn starch/d and to a smaller extent 1.5 kg corn starch/d (both associated with induced hindgut acidosis), but not 8.0 mol BHB/d (associated with induced compensated metabolic acidosis), affected in vitro fermentation characteristics of hindgut inocula from dairy cattle. Incubation of four substrates with inoculum from cows abomasally infused with 3.0 kg corn starch/d increased GP at 3 h for all substrates. After 72 h, this infusion reduced total GP, CH4 production, pH, ammonia concentration, and IVOMD, but increased total VFA concentration and molar proportions of propionate and butyrate, indicating reduced hindgut microbial fermentation capacity and fermentation of the infused starch. Cellulose fermentation showed the lowest IVOMD with inocula from the 3.0 kg corn starch/d infusion, indicating reduced hindgut microbial capacity to ferment cellulose. Regardless of fibre or starch sources serving as substrate, results suggest that hindgut microbial fermentation capacity decreases when more starch enters the hindgut. The findings imply that dietary starch levels should be carefully managed to maintain optimal hindgut fermentation and mitigate CH4 emissions.
Acknowledgements
We are grateful to the animal caretakers of the Carus animal research facilities at Wageningen University & Research (Wageningen, the Netherlands), C. de Wildt (student of Wageningen University & Research), and to the laboratory staff of the Animal Nutrition Group of Wageningen University & Research especially Mrs. S. van Laar for her assistance with GP work and analyses.
Author contributions
S. Abd Rahim, H. van Laar, W. F. Pellikaan, S. van Gastelen and J. Dijkstra conceived and designed the study. S. Abd Rahim conducted the experiment and gathered data. S. van Gastelen provided faecal samples for the research. S. Abd Rahim, H. van Laar, W. F. Pellikaan and J. Dijkstra processed data and performed statistical analysis. S. Abd Rahim wrote the original article and revised the manuscript. H. van Laar, S. van Gastelen, W. F. Pellikaan, A. Bannink and J. Dijkstra reviewed and edited the article. H. van Laar and J. Dijkstra supervised the research.
Funding statement
The study of the first author was funded by Ministry of Higher Education of Malaysia (MoHE). The Ministry of Agriculture, Nature and Food Quality (The Hague, the Netherlands) and the Vereniging Diervoederonderzoek Nederland (Rijswijk, the Netherlands) commissioned and funded the trial of van Gastelen et al. (Reference Van Gastelen, Dijkstra, Alferink, Binnendijk, Nichols, Zandstra and Bannink2021 a) as part of the Policy Support Research theme ‘Feed4Foodure’ (BO-31.03-005-001; TKI-740 AF12039), which provided faecal samples for the in vitro trial. This in vitro research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
The authors declare there are no conflicts of interest.
Ethical standards
The faecal samples were collected in an experiment that was conducted at Wageningen University & Research (Wageningen, the Netherlands), under the Dutch Law on Animal Experiments in accordance with European Union Directive 2010/63, and was approved by the Central Committee of Animal Experiments (The Hague, the Netherlands; 2018.D-0013.002).
Open access
