Contents lists available at ScienceDirect

Livestock Science

Livestock Science 162 (2014) 115–125

http://d1871-14

n CorrE-m

journal homepage: www.elsevier.com/locate/livsci

Efficacy and mode of action of selected non-ionophoreantibiotics and direct-fed microbials in relation toMegasphaera elsdenii NCIMB 41125 during in vitrofermentation of an acidosis-causing substrate

H.H. Meissner a, P.H. Henning b, K.-J. Leeuw a,n, F.M. Hagg b, C.H. Horn b,A. Kettunen c, J.H.A. Apajalahti c

a Agricultural Research Council, Private Bag X2, Irene 0062, South Africab Megastarter Biotech t/a MS Biotech, PO Box 10520, Centurion 0046, South Africac Alimetrics Ltd., Koskelontie 19B, FIN-02920 Espoo, Finland

a r t i c l e i n f o

Article history:Received 4 April 2012Received in revised form22 January 2014Accepted 26 January 2014

Keywords:AntibioticsDirect-fed microbialsMegasphaera elsdeniiIn vitro fermentationAcidosis

x.doi.org/10.1016/j.livsci.2014.01.02613 & 2014 Elsevier B.V. All rights reserved.

esponding author. Tel.: þ27 12 672 9320.ail address: kleeuw@arc.agric.za (K.-J. Leeuw

a b s t r a c t

The efficacy of prominent in-feed antibiotics and direct-fed microbials (DFM) to preventor mitigate ruminal acidosis and lactate accumulation, in addition to whether theirpresence will enhance or inhibit Megasphaera elsdenii strain NCIMB 41125 (Me) werestudied in vitro. The antibiotics studied were aureomycinþsulfamethazine as AS-700 (AS),terramycin as TM-200 (TM), zinc bacitracin (ZB), flavomycin (FM) and tylosin (TS). TheDFM were Bovamine (BM) which contains a propionic bacterium and a lactobacillus,Levucell (LC) which contains a strain of the yeast Saccharomyces cerevisiae and Progut (PG)which contains a hydrolysate of S. cerevisiae. The antibiotics and DFM were introducedalone or in the presence of Me to an in vitro systemwith fermentation vessels containing amedium that promoted rapid gas production and lactate development. Dose sizes of theantibiotics were chosen to inhibit fermentation by 10–20% or 30–40% and for DFM dosesizes were according to the manufacturers. For Me the dose size was 100 ml/40 mlcontaining 2.5�105 colony forming units per ml. Me on average reduced lactate from20.0 mM to 4.89 mM, increased VFA production and shifted VFA proportions to morebutyrate and valerate (respectively from 5.80 to 16.0 mM/100 mM and from 0.51 to4.71 mM/100 mM). The antibiotics moderately reduced lactate (26.7–16.8 mM), and AS,ZB and TS enhanced a VFA proportional shift towards propionate (from 22.6 to 28.7 mM/100 mM). In the presence of Me lactate was reduced to levels of Me alone and the ratiobutyrate to propionate was reduced. None of the antibiotics inhibited the action of Me; onthe contrary the interaction was additive. In contrast to the antibiotics and PG, the DFMBM and LC did not affect fermentation resulting in no response with respect to any of thevariables measured. PG in the presence of Me apparently enhanced the action of Me, asnoticed by an additional increase in butyrate and valerate proportions.

& 2014 Elsevier B.V. All rights reserved.

).

1. Introduction

Ruminal or lactate acidosis is characterised by rapidaccumulation of lactic acid and volatile fatty acids (VFA),

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125116

resulting in sharp declines in ruminal pH to the detrimentof fermentation and physiological function. Gradual adapta-tion of the concentrate-fed ruminant and the inclusion ofbuffers in the diet have been only partially successful,resulting in wide-spread investigations into the usage of anarray of antibiotic and direct-fed microbial (DFM) products.The efficacy and probable mode of action of these productshave recently been reviewed (Brown and Nagaraja, 2009;Krehbiel et al., 2003; Meissner et al., 2010).

Mitigation of digestive disorders such as lactate acido-sis has been done with antibiotics such as ionophores(Mutsvangwa et al., 2002; Nagaraja et al., 1987). Less isknown of the mode of action of non-ionophore antibiotics.Antibiotics have been banned as feed additives for live-stock production in some parts of the world (Cogliani et al.,2011) and their use limited or reconsidered in other parts(Wileman et al., 2009), which prompted many investiga-tions to find alternatives (Allen et al., 2013). Non-ionophore antibiotics nevertheless are still used in certaincountries, thereby justifying investigations into their effi-cacy in mitigating or controlling digestive orders.

The in-feed non-ionophore antibiotic tylosin reduceslactic acid production in vitro (Nagaraja et al., 1987) and iseffective against Fusobacterium necrophorum (Lechtenberget al., 1998) which is the primary aetiologic agent of liverabscesses (Nagaraja and Chengappa, 1998), and which isknown to proliferate in lactate-challenged steers. Thisspecialised action is one reason why tylosin and monensinare often included together in feedlot diets (Harvey et al.,2009). Flavomycin also inhibits F. necrophorum and otherbacterial species, and it reduces VFA concentrations at lowin vitro pH (Edwards et al., 2005). In general, there is a lackof data on the control of lactate acidosis by non-ionophoreantibiotics and their interactions with other rumen modi-fiers (Australian Veterinary Association – RAGFAR, 2007),one reason being that most of these products wereregistered many years ago and for other reasons.

Direct fed microbial products of the bacterial categoryare for example lactate-producing lactobacilli fed alone orin combination with lactate-utilising bacteria (e.g. Propio-nibacterium). They have shown some indication of reducedlactate acidosis through higher ruminal pH, positiveresponses to systemic acid base variables (Ghorbaniet al., 2002; Nocek and Kautz, 2006) and maintenance ofan active lactate-utilizing population (Jouany and Morgavi,2007; Nocek and Kautz, 2006) such as Megasphaeraelsdenii. It is conceivable that this may provide a compe-titive advantage to M. elsdenii. However, the main advan-tage of lactobacillus species appears to be their probioticaction with primary benefit in the lower digestive tract(Brown and Nagaraja, 2009). Propionibacterium as lactate-utiliser converts lactic acid to propionic acid which sup-ports gluconeogenesis in the liver and therefore thecombination with lactobacilli should promote animal pro-duction responses (Jouany and Morgavi, 2007; Krehbielet al., 2004). However, a major impact on control of lactateacidosis is not expected.

Direct fed microbial products of the yeast categorysupport ruminal bacterial growth and may alter fermenta-tion products (Chaucheyras et al., 1996; Mutsvangwa et al.,1992). The yeast Saccharomyces cerevisiae stimulated

growth of M. elsdenii by providing essential nutrients(Chaucheyras et al., 1996; Rossi et al., 2004) and increasedpH in vitro, thereby reducing the risk of lactate acidosis(Jouany and Morgavi, 2007). S. cerevisiae hydrolysates areproviding soluble bio-active oligo- and polysaccharidesand peptides (Provenza and Villalba, 2010; Rossi et al.,2004). Therefore, the primary mode of action of theseproducts in control of lactate acidosis appears to be largelyindirect.

Megasphaera elsdenii may also be developed into auseful DFM. It is an effective lactate-utilising bacterialspecies in the rumen which gives preference to lactic acidas substrate (Marounek et al., 1989; Russell and Baldwin,1978). Several patented strains prevented lactic acid accu-mulation and pH decline in vitro or in the rumen (Hinoet al., 1994; Wiryawan and Brooker, 1995); the robust, fast-growing strain NCIMB 41125 being highly successful incontrolling lactate acidosis (Henning et al., 2010; Meissneret al., 2010). Since M. elsdenii as a DFM will often have tooperate in the presence of ionophores or therapeuticand prophylactic antibiotics, sensitivity or resistanceto these products have been evaluated. In general,M. elsdenii strains are not sensitive to monensin (Callawayet al., 1999; Marounek et al., 1989) and are either resistantto, or not inhibited by other antibiotic products (Marouneket al., 1989) among them tylosin, bacitracin and oxytetra-cycline.

The objective of this investigation was to study theefficacy, compared to M. elsdenii strain NCIMB 41125, andmode of action of selected non-ionophore antibiotics andDFM when introduced to a substrate resulting in highlactate levels during in vitro fermentation. The secondobjective was to establish whether strain NCIMB 41125and these products when introduced together into thein vitro system have enhancing or inhibiting interactions.

2. Materials and methods

2.1. Treatments

2.1.1. Experiment 1The effect of M. elsdenii strain NCIMB 41125 (hereafter

referred to as Me) on in vitro fermentation was tested atfour levels of inoculation to establish an optimum forinclusion in the studies on the non-ionophore antibioticsand DFM. The four levels were respectively 1, 10, 100 and1000 ml Me/40 ml, which are equivalent to 0.025, 0.25, 2.5and 25�105 colony forming units (cfu) Me/ml. The resultsof the four levels were compared with a control (noamendments) (CON) and an autoclaved treatment (aMe;1000 ml Me/40 ml). The experiment revealed that 100 mlMe/40 ml containing approximately 2.5 �105 cfu Me/mlwas optimum (see Section 3), which then became theinoculum level of Me in Experiments 2 and 3.

2.1.1.1. Origin and cultivation of M. elsdenii strain NCIMB41125. M. elsdenii strains were selected from the rumenpopulation of M. elsdenii of dairy cows (Horn et al., 2009).These strains were tested by pH-auxostat enrichmentusing stringent selection criteria of high growth rate and

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125 117

biomass production, preferential use of D- and L-lactate ascarbon source, maintaining multiplication at low pH(o4.5) and insensitivity to ionophores. Strain NCIMBmet all the criteria. Detailed descriptions of selectioncriteria are given by Horn et al. (2009) and Meissneret al. (2010).

2.1.2. Experiment 2The effects of the non-ionophore antibiotics in relation

to Me were tested in two runs: In the first run two levels ofthe therapeutic antibiotics Aureo S-700s (aureomycin(77.2 g active ingredient/kg product)þsulfamethazine(77.2 g active ingredient/kg product) at a 1:1 ratio, fromZoetis, Sandton, South Africa) (AS) were studied. The twolevels being respectively 0.01 mg/40 ml (0.25 mg/ml, AS01)and 0.05 mg/40 ml (1.25 mg/ml, AS05). Similarly, two levelsof TM-200 (terramycin-200, from Zoetis) (TM) were stu-died, at 0.01 mg/40 ml (TM01) and 0.05 mg/40 ml (TM05)respectively. The study also included treatments where Mewas inoculated together with the respective two levels ofAS and TM. In the second run two levels of the prophy-lactic antibiotics zinc bacitracin (ZB)(Virbac, Centurion,RSA), flavomycin (FM)(MSD, Midrand, RSA) and tylosin(TS)(Virbac) were studied and, as in the first run, alsocombined with Me. The levels were: 0.1 mg/40 ml (2.5 mg/ml)(ZB0.1) and 0.5 mg/40 ml (12.5 mg/ml)(ZB0.5) for ZB; 2 mg/40 ml (50 mg/ml)(FM2) and 5 mg/40 ml (125 mg/ml)(FM5)for FM, and 0.2 mg/40 ml (5 mg/ml)(TS0.2) and 1 mg/40 ml(25 mg/ml)(TS1) for TS.

2.1.2.1. Product concentration. Even though there are clearregistered recommended doses of antibiotic compoundsfor ruminants in different production systems andcountries, it proved difficult to estimate the residualconcentration of these products that the rumen microbialpopulations are exposed to. This is probably due to the factthat most antibiotic products have been developed toabsorb rapidly from the rumen to the target tissue,suggesting that their actual concentration in the rumenmay be much lower than anticipated.

To find the most probable concentration range for eachproduct, a ruminal simulation study with a wide range ofconcentrations was conducted by titration, with gas produc-tion as response variable. Gas production is a general measureof bacterial activity in an anaerobic system such as the rumen.The fermentation vessels were run without replications, theassumption being that the accuracy should be sufficient dueto the dose range continuum of between 15 and 17 levels ofconcentration, including the Control. For each antibiotic testproduct two doses were then chosen (the levels given above)for inclusion in the experimental design: 10–20% (low dose)and 30–40% (high dose) inhibition of bacterial cumulative gasproduction. The argument was that these concentrationsshould cover the most likely concentrations that occur inthe rumen, and because similar levels of inhibition werechosen the comparison between the test products, althoughqualitative, is valid.

2.1.3. Experiment 3The DFM investigated were the bacterial combination

product Lactobacillus acidophilus NP 51 and Propionibacterium

freudenreichii NP 24, marketed as Bovamine (BM) (NutritionPhysiology Company, Overland Park, Kansas, USA) the yeaststrain S. cerevisiae CNCM 1-1077, marketed as Levucell (LC)(Lallemand, Toulouse, France) and the yeast hydrolysate S.cerevisiae EP 0946108, marketed as Progut (PG) (SuomenRehu, Espoo, Finland). Treatments were respectively 100 mgBM which contained 2�1010 cfu L. acidophilus NP 51 andP. freudenreichii NP 24 live bacteria per gram of product; 4 mgLC which contained 2�1010 cfu S. cerevisiae CNCM 1-1077live yeast per gram of product, and 40mg PG (yeast hydro-lysate). These levels were in agreement with the respectivemanufacturer's recommendations. As in the case of Experi-ment 2, further treatments were the combinations of Mewiththe above levels of BM, LC and PG.

2.2. Inoculum and substrate

Ruminal fluid was obtained from 2 fistulated cows thatwere fed 3.8 kg DM per day of a commercial concentrate,containing barley, sugar beet pulp, rapeseed, wheat, plusminerals and vitamins (Krossi 100, Suomen Rehu) plusmaize silage for ad libitum intake. The ruminal fluid wasstrained through 4 layers of cheese cloth and immediatelysealed in a preheated thermos container, transported tothe Alimetrics laboratory (Espoo, Finland), and used asinoculum within 1 h.

The composition on a dry matter (DM) basis of thesubstrate used in in vitro fermentation was 50% maizesilage, 25% ground barley grain and 25% soybean meal. Thefeed components were individually weighed into thefermentation vessels to ensure that their relative propor-tion was exactly the same in each vessel. The substrate wasmarkedly higher in energy and non-structural carbohydratesthan the diet of the cows that donated the inoculum. Thiswas to promote fermentation resulting in substantial yieldsof lactic acid.

2.3. Protocol and analyses

Each treatment was replicated in six 120 ml serum bottleswith thick butyl rubber stoppers, hereafter referred to asfermentation vessels. Me, aMe, the respective antibioticproducts and BM were administered into the fermentationvessels at simulation start-up. LC and PG were weigheddirectly into the vessels together with the feed components.Each vessel received the respective test compounds in 1 ml ofanaerobic phosphate buffer (Goering and Van Soest, 1970)whereas the CON vessels received only the 1 ml buffer. TheaMe vessels were autoclaved at 121 1C for 20 minwhere after1 ml of autoclaved Me culture was introduced into the vesselsin the same way as the active cultures were introduced intothe vessels of the other treatments. The filling procedurefollowed was: Substrate (DM basis) consisting of 0.5 g maizesilage, 0.25 g ground barley grain and 0.25 g soybean meal(1 g total) was weighed into the fermentation vessels, thevessels flushed with CO2 and sealed with the butyl rubberstoppers. Anaerobic, reduced buffer solution, fresh ruminalfluid (final dilution 1:40) and 1 ml of test product solutionwere then introduced into the vessels to the final volume of40 ml. After inoculation, the fermentation vessels wereincubated at 37 1C under gentle rotatory movement.

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125118

In Experiments 1 and 3 gas production and the con-centrations of lactic acid and volatile fatty acids (VFA) weremeasured at 3, 6, 9 and 12 h. In Experiment 2 only gasproduction was measured over time, at 3, 5, 7 and 9 h. Gasproduction was measured as follows: at sampling, total gasproduction was measured by manually puncturing therubber stopper with a needle connected to an accurate20 ml glass syringe with a sensitive plunger. The volume ofgas released from the vessels was registered as thedisplacement value.

In all experiments bacterial counts were determinedafter 6 and 12 h incubation. Samples for the analysis(200 ml) were withdrawn from the same vessels as usedfor gas measurement. Since the total volume was 40 ml,the change in volume was considered minute and nottaken into account in calculations. Furthermore, since thesame change in volume resulted with all treatments,between-treatment comparison was not affected. VesselpH was also measured after 6 and 12 h incubation but theresults are not presented as it was recognised that due tobuffering the values recorded may not reflect true treat-ment differences.

Volatile fatty acids and total lactic acid were analysedby gas chromatography using a packed column for theanalysis of free acids (Holben et al., 2002). The acidsanalysed for in Experiments 1 and 3 were lactic, acetic,propionic, butyric, iso-butyric, 2-methyl-butyric, valericand iso-valeric. In case of butyric and valeric acids, thedifferent isomers were added up and in the results pre-sented respectively as total butyric and valeric acids. In thetables and text the terms lactate, acetate, propionate areused, simply to shorten the term lactic acid concentration,acetic acid concentration, propionic acid concentration etc.

To establish how competitive strain NCIMB was in thein vitro simulation system and how bacterial numberswere affected by the studied test products, total bacteriaand strain NCIMB were determined by quantitative real-time PCR (qPCR). The approach adopted was to use thedivergence in the 16S rDNA for the analysis of both theM. elsdenii species identification (using a type strain) andspecifically for strain NCIMB identification. The proceduresfollowed were: The M. elsdenii type strain and strainNCIMB were cultured and DNA extracted. PCR was thenused to multiply the 16S rRNA gene from both strains. ThePCR products were subsequently cloned to a vector and 24clones from both type strain and strain NCIMB sequencedfor the 16S rRNA gene. The resulting sequences werecompared to related sequences from public databasesand to each other to reveal possible differences betweenthe type strain and strain NCIMB, and to see whether the16S rRNA gene operon exists in multiple copy numbersthat diverge from each other. The obtained sequence datarevealed that M. elsdenii strains carry 8 copies of the 16SrRNA gene and both the type strain and strain NCIMBshowed 5 operons where divergent sequences occur.Based on the sequencing, location discrimination betweenthe type strain and strain NCIMB was discovered and astrain-specific qPCR assay could be developed.

In the results total bacterial density values are given inaddition to the density of M. elsdenii. The latter values arethe combination between endemic M. elsdenii present in

the fermentation vessel and the introduced strain NCIMB(Me). The increment in M. elsdenii density in treatmentscontaining Me therefore represents strain NCIMB. Totalbacterial and M. elsdenii density is expressed in chromo-somes/ml, which were calculated from standards in theqPCR method of which the number of 16S gene copies inthe reaction volume was known: Based on the standardsand a log-linear relationship, the number of 16S copies/mlfor the unknown samples were calculated and these wereconverted to chromosomes/ml, using an assumption onthe 16S copy number per chromosome.

2.4. Statistical analysis

In Experiment 1 linear mixed model analysis; forrepeated measurements (REML analysis, Payne et al., 2009)was used to establish the optimum Me concentration to beused in Experiments 2 and 3. An ante-dependence modelof order 1 (AR 1) was used to model the correlation overhours. Two approaches were adopted, (1) where the fixedfactor was: Constantþhourþtreatment (concentration ofMe)þhour� treatment, and (2) where the fixed factorwas: Constantþhourþ log Me concentrationþhour� logMe Concentration. In both cases the random model wasvessel�hour. Estimates of variables in the two approachescorresponded and the results based on Approach 1 aretherefore presented.

In Experiments 2 and 3 a factorial design was used. In thefirst run of Experiment 2 Control (CON) and the twoconcentrations of AS and TM were tested in the absence(Me�) or presence (Meþ) of Me and in the second run CONand the two concentrations of ZB, FM and TS. In Experiment 3CON and BM, LC and PG were tested with Me� and Meþ .As in Experiment 1 linear mixed model analysis was used.The fixed factor in the case of repeated measurements was:ConstantþhourþMeþtreatment (antibiotics or DFM)þhour�Meþhour� treatmentþMe� treatmentþhour�Me� treatment. The random model was: vessel�hour. Inthe case of not repeated measurements the fixed modelwas: ConstantþMeþtreatmentþMe� treatmentþa residualterm.

For gas production the square root had to be calculatedto normalise the data before analysis in Experiments 1and 2, but not in Experiment 3. For total bacteria andM. elsdenii density the log transformation had to be calcu-lated to normalise the data before analysis. The level ofsignificance accepted was Po0.01.

3. Results

3.1. Experiment 1

The trends in gas production and lactate reduction overtime are presented in Tables 1a and 1b respectively, andthe results of treatment differences are presented inTable 2.

The results indicate that Me increased gas productionfrom 3 to 12 h dose dependently (Table 1a). For example,compared to CON the increase for the 1000 μl Me dose wasalready significant at 3 h, whereas for the 10 μl dose theresponse was significant only after 9 h of incubation. The

Table 1aSquare root of gas produced (measured in ml) over time. Control (CON) is compared with M. elsdenii strain NCIMB 41125 (Me) at four concentrationsand autoclaved Me (aMe), s.e.¼0.05.

Treatment Hour

3 6 9 12

CON 2.74a 6.11a 9.39a 11.0a

1Me 2.74a 6.20a 9.59a 11.3a

10Me 2.71a 6.27a 9.96b 12.0b

100Me 2.85a 7.15b 10.6c 12.2b

1000Me 3.89b 7.38b 10.6c 12.2b

aMe 2.72a 6.21a 9.52a 11.2a

s.e, Standard error; 1Me, 1 ml Me/40 ml; 10Me, 10 ml Me/40 ml; 100Me, 100 ml Me/40 ml; 1000Me, 1000 ml Me/40 ml (equivalent to 0.025, 0.25, 2.5and 25�105 cfu Me/ml respectively); a,b,c, numbers in the same column with different superscript letters differ at Po0.001.

Table 1bLactate (mM) development over time Control (CON) is compared withM. elsdenii strain NCIMB 41125 (Me) at four concentrations and autoclaved Me (aMe),s.e.¼0.76.

Treatment Hour

3 6 9 12

CON 7.36b 21.4c 32.6c 24.2c

1Me 7.40b 21.0c 30.4c 18.7b

10Me 6.48b 20.9c 19.9b 0.78a

100Me 7.00b 8.17b 6.04a 0.53a

1000Me 5.35a 4.12a 6.54a 0.59a

aMe 6.83b 22.4c 29.3c 24.3c

s.e, Standard error; 1Me, 1 ml Me/40 ml; 10Me, 10 ml Me/40 ml; 100Me, 100 ml Me/40 ml; 1000Me, 1000 ml Me/40 ml (equivalent to 0.025, 0.25, 2.5and 25�105 cfu Me/ml respectively); a,b,c, numbers in the same column with different superscript letters differ at Po0.001.

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125 119

autoclaved Me (aMe), which was effected at the high doseof 1000 μl Me, did not influence gas production, indicatingthat the observed changes in gas production depended onthe viability of the organism. There were initial ratedifferences between 100 μl Me and 1000 μl Me, but thedifferences were not significant after 6, 9 and 12 h ofincubation, which suggests that the doses of 100 ml and1000 ml, even though they differed tenfold, were almostequally effective.

Lactate was reduced by Me (Table 1b) and similar to gasproduction, the response was dose dependent. The highestdose (1000 μl) significantly reduced the concentration oflactate already after 3–6 h of incubation, with the smallerdose sizes following progressively at later time points.Overall, lactate after 12 h of incubation was reduced from24 mM in CON to 0.5–0.6 mM in the 100 and 1000 μl Metreatments. As with gas production, the autoclaved treat-ment (aMe) did not differ significantly from CON at anyincubation time point and the difference between the 100and 1000 ml doses after 6 h of incubation was small.

Since the differences in M. elsdenii density and VFA inthe fermentation vessels were not significant between100Me and 1000Me (Table 2), which is in line with thesmall difference between these two treatments in gasproduction and lactate reduction (Tables 1a and 1b), itwas decided that 100 ml Me/40 ml was optimal for use inExperiments 2 and 3.

None of the amendments increased total bacterialdensity significantly (Table 2). M. elsdenii density, however,increased with 2–3 log units in response to Me dose sizeand accumulation of the lactate substrate. Compared to

CON, there was no increase in M. elsdenii density in theautoclaved treatment.

Total VFA increased from 43.7 mM in CON to a max-imum of 51.5 mM at 100 Me. The autoclaved treatment(aMe) compared to CON also showed a response, probablyas a result of the fermentation medium (Section 2.2)containing nutrients. The latter explanation also accountsfor results of other parameters of aMe and are not furtherreferred to. The acetate–propionate ratio decreased for1Me and 10Me, but increased for 1000Me. As a proportionof total VFA acetate decreased progressively with Me doselevel, whereas propionate increased for 10Me but decreasedfor 100Me and 1000Me, the difference being associated withincreasing proportions of butyrate and valerate.

3.2. Experiment 2

The effects tested were those of Me (Me� vs. Meþ)and the antibiotics at two levels (Treatment). The test forinteraction between Me and Treatment (MeþTr) indicatedwhether Treatment in the presence of Me (Meþ) had anadditional response. The results of the therapeutic anti-biotics AS-700 (AS) and TM-200 (TM) are shown in Table 3and those of the prophylactic antibiotics zinc bacitracin(ZB), flavomycin (FM) and tylosin (TS) in Table 4.

Me stimulated gas production, whereas AS and TMdepressed fermentation and therefore gas production inaccordance with dose size. The higher concentrations of ASand TM were particularly severe (Table 3) and compared toCON also depressed total bacterial density. Me increased

Table 2Square root of gas production (ml), log 10 of total bacteria and M. elsdenii density (chromosomes/g), lactate (mM), total VFA (mM) and individual VFA asproportion of total VFA (mM/100 M). Control (CON) is compared with M. elsdenii strain NCIMB 41125 (Me) at four concentrations and with autoclavedMe (aMe).

Treatm. Measured parameter

Gas T. bact. Me Lact VFA Ac. Pro Ac:Pro But Val

CON 7.31a 9.01 5.45a 21.4d 43.7a 71.2e 19.1c 2.76c 6.95a 0.94a

1Me 7.45a 9.20 6.55b 19.4c 44.6a 70.7e 19.3c 2.59b 7.12b 1.12a

10Me 7.75b 8.98 7.33c 12.0b 50.3b 66.8d 19.9c 2.39a 9.53c 2.12b

100Me 8.20c 9.14 8.05d 5.43a 51.5b 58.1b 16.4b 2.68bc 18.1d 5.20d

1000Me 8.51d 9.29 8.34d 4.15a 51.8b 50.1a 13.9a 3.04d 26.1e 7.76e

aMe 7.41a 9.00 5.36a 20.7cd 49.1ab 65.2c 22.1d 2.32a 8.34b 2.90c

s.e. 0.01 0.12 0.10 0.35 0.57 0.19 0.17 0.07 0.12 0.07P-value o0.01 0.34 o0.01 o0.01 o0.01 o0.01 o0.01 o0.01 o0.01 o0.01

T. Bact., total bacteria; Lact., lactate; VFA, volatile fatty acids; Ac, acetate; Pro, propionate; Ac:Pro, acetate–propionate ratio; But, butyrate; Val, valerate;1Me, 10Me, 100Me, and 1000Me, respectively 1, 10, 100, and 1000 ml M. elsdenii strain NCIMB/40 ml; s.e., standard error; a,b,c,d,e, numbers in the samecolumn with different superscript letters differ. Level of significance accepted: Po0.01.

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125120

M. elsdenii density from about 4–8 log units, whereas ASand TM had no effect.

Compared to CON, both AS and TM at the highconcentrations reduced lactate, but not to the same extentas Me. The reduction in lactate by AS and TM in thepresence of Me (Meþ) was primarily due to Me. Total VFAwere increased by Me, not affected by AS and reduced byTM (TM05) (Table 3). When Me was added to AS and TM,VFA production was increased. The acetate–propionateratio was lower in all amendments, except TM05 andTM05þMe. The lowest ratio was recorded for AS05 andAS05þMe. As a proportion of total VFA, acetate andpropionate were lower in Me and butyrate and valeratewere increased. Apart from TM05, the proportion ofacetate was unaltered by the two antibiotics, but showedsimilar proportions to that of Me when in the presence ofMe. In contrast, the proportion of propionate wasincreased by AS and TM except for TM05, whereas theproportion of butyrate and valerate was reduced.

Similar to the first trial run, Me increased gas produc-tion whereas the prophylactic antibiotics ZN, FM and TS inthe second run depressed gas production (Table 4) accord-ing to dose size. This in the case of the high antibioticconcentrations caused moderate although significantdepressions in total bacterial density. Addition of strainNCIMB 41125 resulted in an increase of 4 log units in M.elsdenii density in the Me and antibioticþMe treatments.

Lactate concentration was reduced by Me and the highconcentrations of the three antibiotic treatments, and evenfurther reduced when Me was added to the fermentationvessels with the high antibiotic concentration (Table 4), theresponse was additive. Total VFA was increased in the treat-ment Me and reduced in the treatment with the antibiotics,especially at the high concentrations. In the presence of Me,total VFA in the antibiotic treatments resemble the concentra-tions of Me alone (Table 4). Except for FM5, propionate wasproduced at the expense of acetate in the treatment withthese antibiotics as can be seen from the acetate–propionateratio and the proportion of propionate in the total VFA.Propionate as a proportion of total VFA in Me decreased,whereas butyrate and valerate increased in line with theresults in Tables 2 and 3. In the combination treatments with

ZB0.5 and TS the proportion propionate increased. Littlebutyrate and practically no valerate were produced in thetreatments with the three antibiotics, whereas the combina-tion treatments (Me with antibiotics) tended towards thevalues of Me treatment.

3.3. Experiment 3

The results of Me and the DFM products BM, LC and PG,are displayed in Table 5. Me and to a lesser extend PG,stimulated gas production whereas the difference betweenCON, BM and LC was not significant. In the presence of Me,gas production in the fermentation vessels containing BMand LC increased to the same level as in treatment Mealone. For the amendment PGþMe, gas production wasstimulated further.

Total bacterial density was stimulated by Me but notaffected by the DFM treatments, whereas M. elsdeniidensity was increased by 3 log units in the treatmentscontaining Me (Table 5).

Bovamine and Levucell did not affect lactate production,whereas Progut increased lactate production. In treatmentswith Me, lactate was reduced from 21.4 to 5.41 mM (Table 5).In the DFM–Me combination treatments, lactate was affectedsimilarly to Me alone.

Total VFA production was increased by Me and to thesame extend by the DFM–Me combination treatments.Total VFA production was not affected by the DFM alonetreatments. As proportion of total VFA, none of the resultsof acetate and propionate of the DFM products differedsignificantly from CON, but PG increased the proportionbutyrate and valerate. Me containing treatments on theother hand lowered the proportion of acetate and propio-nate and increased the proportion of butyrate and valerate,with PGþMe particularly pronounced.

4. Discussion

4.1. Limitations of in vitro investigation

The results from the in vitro fermentation system canonly provide indicators and act as a screening tool of

Table 3Square root of gas production (ml), log 10 of total bacterial and M. elsdenii density (chromosomes/g), lactate (mM), total VFA (mM) and individual VFA asproportion of total VFA (mM/100 M). The results of the antibiotics AS-700 (AS) and TM-200 (TM) at two concentrations are shown in relation to a Control(CON) and in the absence and presence of M. elsdenii strain NCIMB 41125 (Me).

Treatm. Measured parameter

Gas T. bact. Me Lact VFA Ac. Pro Ac:Pro But Val

Me– 4.50a 8.66a 4.22a 22.3b 39.4a 66.7b 28.2b 2.39b 3.72a 0.09a

Meþ 4.82b 8.80b 7.87b 7.52a 48.5b 56.7a 25.6a 2.22a 12.3b 4.05b

s.e. 0.02 0.03 0.08 0.39 1.72 0.26 0.15 0.01 0.22 0.10P-value o0.001 0.006 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001

CON 5.52e 8.84b 6.21 21.1c 51.1b 60.5a 24.7a 2.45c 11.2d 2.45b

AS01 5.26d 8.82b 6.12 17.9bc 46.3b 59.9a 27.3b 2.19b 8.75c 2.62b

AS05 4.20b 8.62a 5.86 10.5a 42.5b 56.9a 30.0c 2.00a 6.92b 2.07b

TM01 5.04c 8.75ab 6.13 17.5b 49.8b 60.3a 27.2b 2.22b 8.55c 25.0b

TM05 3.30a 8.62a 5.90 8.84a 30.2a 68.1b 25.3a 2.69d 4.66a 0.71a

s.e. 0.03 0.05 0.11 0.63 2.75 0.43 0.15 0.02 0.36 0.16P-value o0.001 0.005 0.15 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001

Me�CON 5.30 8.74 4.42 31.0f 45.1 67.0d 25.4b 2.64e 6.07b 0.20a

AS01 5.09 8.71 4.21 28.0ef 43.4 66.6d 28.7cd 2.32d 3.10a 0.13a

AS05 4.07 8.58 4.11 15.1d 39.3 64.2c 32.2e 1.99a 2.50a 0.00a

TM01 4.85 8.68 4.27 26.5e 43.7 66.0cd 29.0d 2.28d 3.40a 0.13a

TM05 3.21 8.61 4.09 10.8c 25.6 69.6e 25.6b 2.73e 3.53a 0.00a

MeþCON 5.73 8.94 7.99 9.23bc 57.1 53.9ab 23.9a 2.26cd 16.3e 4.69cd

AS01 5.42 8.94 8.03 7.82abc 49.2 53.2a 25.8b 2.06ab 14.4d 5.11d

AS05 4.33 8.67 7.62 5.06a 45.7 55.5b 27.7c 2.00a 11.3c 4.14c

TM01 5.23 8.81 8.00 8.57bc 55.8 54.6ab 25.4b 2.15bc 13.7d 4.88cd

TM05 3.40 8.62 7.71 6.89ab 34.8 66.5d 25.1ab 2.65e 5.80b 1.43b

s.e. 0.04 0.08 0.16 0.93 4.06 0.61 0.35 0.04 0.51 0.23Me�TrP-value 0.023 0.53 0.84 o0.001 0.87 o0.001 o0.001 o0.001 o0.001 o0.001

T. Bact., total bacteria; Lact., lactate; VFA, volatile fatty acids; Ac., acetate; Pro., propionate; Ac:Pro., acetate–propionate ratio; But., butyrate; Val., valerate;Me� , in the absence of M. elsdenii strain NCIMB; Meþ , in the presence of M. elsdenii strain NCIMB (concentration: 100 ml Me/40ml); AS01, 0.01 mg AS-700/40 ml; AS05, 0.05 mg AS-700/40 ml; TM01, 0.01 mg TM-200/40 ml; TM05, 0.05 mg TM-200/40 ml; s.e., standard error; Me�Tr, interaction between Meand the treatments (antibiotics); a,b,c,d,e,f, numbers in the same column with different superscript letters differ. Level of significance accepted: Po0.01.

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125 121

relevance to in vivo. In addition to fermentation, the netresult in the rumen is a function of dilution, passage andabsorption which are absent in in vitro simulation. How-ever, as all treatments were exposed to the same experi-mental conditions, fermentation capacity and endproducts qualitatively should reflect differences betweentreatments. This also applies to interpretation of interac-tion between strain NCIMB and the tested antibioticsand DFM.

Ruminal acidosis is primarily associated with lactateaccumulation and the question is if the in vitro system heresimulated accumulation and reduction satisfactorily. Lac-tate in CON in the present investigations was 20.9–31.0 mM (Tables 2–5) and 3.84–9.23 mM in the strainNCIMB treatments, the reduction value being 20% ofcontrol value. The reduction is less in animal experimentswhich may be 10% of control value (Henning et al., 2010;Meissner et al., 2010), partly because the fermentationmedium contained less rapidly fermentable carbohydrates.The reduction by strain NCIMB nevertheless was of suffi-cient magnitude to allow effective comparison with treat-ments containing antibiotics and DFM.

4.2. M. elsdenii strain NCIMB 41125

M. elsdenii by preference utilises lactic acid as a sourceof carbon and energy (Marounek et al., 1989; Russell andBaldwin, 1978), which is the theoretical basis of theadvantage of the organism in the prevention of lactic acid(ruminal) acidosis. The selected M. elsdenii strain NCIMB(Me) in this study stimulated fermentation of the concen-trate substrate while maintaining or enhancing the totalbacterial population. It also effectively utilised the result-ing lactic acid (Tables 2–5). This indicates the strongpotential of the strain to prevent ruminal fermentationdisorders caused by lactate accumulation and the conse-quent detrimental decline in ruminal pH which has beenreported in previous in vitro and in vivo studies (Henninget al., 2010; Meissner et al., 2010).

A number of other consistent observations were madein the present study compared to previous variable results.Total VFA production was increased by Me in both theantibiotic and DFM trials. Furthermore, the production wasmaintained in the presence of the antibiotics, at leastwith the low dose where fermentation was not inhibited

Table 4Square root of gas production (ml), log 10 of total bacteria and M. elsdenii density (chromosomes/g), lactate (mM), total VFA (mM) and individual VFA asproportion of total VFA (mM/100 M). The results of the antibiotics zinc bacitracin (ZB), flavomycin (FM) and tylosin (TS) at two concentrations are shownin relation to a Control (CON) and in the absence or presence of M. elsdenii strain NCIMB 41125 (Me).

Treatm. Measured parameter

Gas T. bact. Me Lact VFA Ac. Pro Ac:Pro But Val

Me� 4.52a 10.3a 5.52a 14.7b 39.4a 67.6b 27.0b 2.64a 5.46a 0.00a

Meþ 4.82b 10.4b 9.46b 1.21a 47.7b 58.2a 22.8a 2.66b 14.6b 4.27b

s.e. 0.01 0.02 0.09 0.30 0.42 0.08 0.09 0.01 0.07 0.04P-value o0.001 0.008 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001

CON 5.55g 10.7c 7.56 13.1c 50.6e 62.1c 20.5b 3.03e 15.3e 2.13c

ZB0.1 5.07f 10.5c 7.53 11.6c 47.8de 62.6c 23.2d 2.70c 11.5d 2.67d

ZB0.5 4.20b 10.2b 7.38 1.76a 41.8b 60.9b 31.5f 1.94a 6.10b 1.50a

FM2 4.72d 10.3b 7.27 6.23b 44.1bc 64.7d 22.2c 2.92d 10.9c 2.17c

FM5 4.44c 10.3b 7.53 5.06b 38.5a 69.3e 17.9a 3.88f 11.1cd 1.80b

TS0.2 4.99e 10.6c 7.60 16.1d 45.6cd 60.2a 27.9e 2.14b 8.78b 2.88d

TS1 3.72a 10.1a 7.54 1.86a 36.5a 61.4ab 31.6f 1.96a 6.59a 1.82b

s.e. 0.02 0.05 0.19 0.58 0.82 0.16 0.17 0.02 0.14 0.07P-value o0.001 o0.001 0.76 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001

Me�CON 5.39 10.5 5.60 22.3d 44.3ef 68.0h 21.4e 3.18f 10.6f 0.00a

ZB0.1 4.86 10.5 5.66 21.3d 41.2cde 69.4i 25.6g 2.71d 5.06b 0.00a

ZB0.5 4.02 10.2 5.46 3.52b 40.1bcde 62.7e 33.7i 1.86a 3.53a 0.00a

FM2 4.57 10.3 5.15 12.5c 42.0e 69.8i 24.2f 2.89e 5.97c 0.00a

FM5 4.34 10.2 5.69 10.1c 36.0ab 74.7j 19.2b 3.89g 6.05c 0.00a

TS0.2 4.84 10.6 5.55 29.4e 39.9abc 66.6g 30.0i 2.19c 3.43a 0.00a

TS1 3.60 10.0 5.52 3.73b 35.4a 61.9d 34.6j 1.79a 3.54a 0.00a

MeþCON 5.70 10.8 9.51 3.84b 56.9g 56.3b 19.5bc 2.88e 19.9j 4.26d

ZB0.1 5.29 10.6 9.41 1.92ab 54.3g 55.9b 20.8de 2.68d 17.9i 5.34e

ZB0.5 4.39 10.2 9.29 0.00a 43.5ef 59.1c 29.3i 2.02b 8.67d 2.99b

FM2 4.87 10.4 9.39 0.00a 46.2f 59.6c 20.1cd 2.96e 15.9h 4.34d

FM5 4.54 10.3 9.37 0.00a 41.1cde 63.8f 16.5a 3.87g 16.1h 3.60c

TS0.2 5.13 10.6 9.65 2.70ab 54.2g 53.9a 25.7g 2.10bc 14.1g 5.76f

TS1 3.84 10.1 9.56 0.00a 37.5abcd 59.0c 27.7h 2.13bc 9.64e 3.64c

s.e. 0.03 0.07 0.31 0.81 1.16 0.22 0.24 0.03 0.20 0.10

Me�TrP-value 0.028 0.72 0.89 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001

T. Bact., total bacteria; Lact., lactate; VFA, volatile fatty acids; Ac., acetate; Pro., propionate; Ac:Pro., acetate–propionate ratio; But., butyrate; Val., valerate;Me� , in the absence ofM. elsdenii strain NCIMB; Meþ , in the presence ofM. elsdenii strain NCIMB (concentration: 100 ml/40 ml); ZB0.1, 0.1 mg ZB/40 ml;ZB0.5, 0.5 mg ZB/40 ml; FM2, 2 mg FM/40 ml; FM5, 5 mg FM/40 ml; TS0.2, 0.2 mg TS/40 ml; TS1, 1.0 mg TS/40 ml; s.e., standard error; Me� Tr,interaction between Me and the treatments (antibiotics); a,b,c,d,e,f,g,h,i,j, numbers in the same column with different superscript letters differ. Level ofsignificance accepted: Po0.01.

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125122

(Tables 3 and 4). This indicates that it is unlikely that theantibiotics tested will affect the viability of strain NCIMB,which supports results of other M. elsdenii strains(Marounek et al., 1989). Secondly, although proportionallylower, acetate production was maintained which is neces-sary to maintain ideal acetate-propionate ratios in therumen of high concentrate-fed dairy cows. Acetate isnormally an end product of fibre digestion, but since fibredigesters are sensitive to low ruminal pH their functionshould be upheld by the dominant M. elsdenii. Thirdly,compared to CON propionate production was proportion-ally shifted away from propionate to butyrate and valerateby strain NCIMB. A shift to butyrate and valerate was alsoreported by Henning et al. (2010) and Meissner et al.(2010), which supports results on M. elsdenii in general(Horn et al., 2009; Marounek et al., 1989). Butyrate isimportant for rumen epithelium integrity (Mentschel

et al., 2001) and is extensively metabolised in the rumenepithelium (Baldwin and McLeod, 2000), saves propionatefrom being metabolised (Van Soest, 1982) and of all VFA, itshows the highest correlation with milk production(Seymour et al., 2005). Nevertheless, a balance withpropionate is essential since propionate is the only gluco-genic VFA (Reynolds, 2003) and should therefore bemaintained or enhanced. The strong evidence that propio-nate production can be maintained when strain NCIMB isin the presence of specific antibiotics is encouraging forperformance in intensive production systems.

4.3. Antibiotics

The therapeutic antibiotics AS-700 and TM-200 inhib-ited fermentation in the fermentation vessels, moderatelyat the low doses and heavily at the high doses (Table 3).

Table 5Gas production (ml), log 10 of total bacteria and M. elsdenii density (chromosomes/g), lactate (mM), total VFA (mM) and individual VFA as proportion of totalVFA (mM/100 M). The results of the DFM Bovamine (BM), Levucell (LC) and Progut (PG) are presented in relation to a Control (CON) and in the absence orpresence of M. elsdenii strain NCIMB 41125 (Me).

Treatm. Measured parameter

Gas T.bact. Me Lact VFA Ac. Pro Ac:Pro But Val

Me� 65.2a 9.01a 5.28a 21.4b 43.2a 72.4b 19.2b 5.09 7.07a 1.00a

Meþ 82.6b 9.15b 8.04b 5.41a 51.4b 58.6a 16.5a 4.22 18.9b 5.30b

s.e. 0.33 0.05 0.06 0.20 0.37 0.51 0.17 0.05 0.09 0.04P-value o0.001 0.029 o0.001 o0.001 o0.001 o0.001 o0.001 0.13 o0.001 o0.001

CON 72.3a 9.11 6.70 13.2a 47.0 65.7 17.9 4.70b 12.8a 3.11a

BM 71.9a 9.03 6.60 13.5ab 46.9 66.1 17.9 4.72b 12.5a 2.96a

LC 73.2a 9.11 6.61 12.6a 47.7 65.5 17.9 4.76b 12.4a 2.94a

PG 78.2b 9.08 6.73 14.3b 47.4 64.7 17.6 4.43a 14.4b 3.59b

s.e. 0.48 0.07 0.09 0.30 0.55 0.77 0.25 0.07 0.13 0.06P-value o0.001 0.36 0.83 0.004 0.36 0.28 0.56 0.003 o0.001 o0.001

Me�CON 64.6ab 9.09 5.36 20.9b 42.6 72.5 19.0 5.19c 6.99a 0.97a

BM 64.3a 8.88 5.12 21.0b 42.8 72.5 19.1 5.15c 6.89a 0.93a

LC 64.9ab 9.05 5.34 20.2b 44.4 72.0 19.2 5.12c 7.19a 0.98a

PG 67.0b 9.03 5.31 23.5c 42.9 72.6 19.3 4.86c 7.12a 1.12a

MeþCON 79.9c 9.14 8.05 5.42a 51.4 59.0 16.7 4.20ab 18.5c 5.24c

BM 79.5c 9.18 8.08 5.92a 51.1 59.7 16.7 4.28ab 18.0bc 4.99bc

LC 81.5c 9.17 7.88 5.12a 51.0 58.9 16.6 4.39b 17.6b 4.89b

PG 89.5d 9.13 8.15 5.15a 52.0 56.7 15.8 4.00a 21.6d 6.06d

s.e. 0.74 0.11 0.12 0.43 0.82 1.16 0.37 0.10 0.18 0.08

Me� TrP-value 0.002 0.49 0.31 o0.001 0.21 0.56 0.37 0.009 o0.001 o0.001

T. Bact., total bacteria; Lact., lactate; VFA, volatile fatty acids; Ac., acetate; Pro., propionate; Ac:Pro., acetate–propionate ratio; But., butyrate; Val., valerate;Me� , in the absence of M. elsdenii strain NCIMB; Meþ , in the presence of M. elsdenii strain NCIMB (concentration: 100 ml/40 ml); BM, 100 mg BM/40 ml; LC,4 mg LC/40 ml; PG, 40 mg PG/40 ml; s.e., standard error; Me� Tr, interaction between Me and the treatments (DFM); numbers in the same column withdifferent superscript letter differ. Level of significance accepted: Po0.01.

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125 123

The effect on total bacterial density was marginal, possiblydue to the experimental procedures with relatively highinoculums and short incubation time (maximum 12 h).However, the inhibition of microbial metabolism as mea-sured by gas production and VFA was distinct, more so forTM-200 than for AS-700. Notwithstanding this inhibitoryeffect, strain NCIMB was hardly affected, even when fermen-tation was heavily inhibited. The effect of the two antibioticson lactate was noticeable for the high dose. Since thiscoincided with a change in metabolic profile of the VFA suchas a higher propionate proportion, the response probablypartly indicates to inhibition of particular lactate producers,which is a known action of in-feed antibiotics (Nagaraja andTaylor, 1987). However, the response mainly resulted as a co-response to the reduction in total VFA produced. Overall,however, strain NCIMB was the primary instigator in reduc-tion of lactate. The total VFA production of the low doses ofAS-700 and TM-200 was increased to that of the Me alonetreatment if Me was combined with them, which suggeststhat, depending on dose size, it may be possible to optimiseend product output when strain NCIMB is administered inthe presence of selective regulated inhibitors.

The prophylactic antibiotics zinc bacitracin, flavomycinand tylosin also modulated fermentation as noticed by areduction in gas production and total bacterial density(Table 4) and in contrast to the therapeutic antibiotics, they

had major influences on lactate (apart from zinc bacitracinand tylosin at the low dose) total VFA and the proportionalcontribution of fatty acids to total VFA. Lactate was reducedlargely independently of the effect of strain NCIMB. Also, theproportion of propionate was increased, apart from flavomy-cin at the high dose, flavomycin is known for not affectingVFA proportions (Rowe et al., 1982). The modulations causedby zinc bacitracin, flavomycin and tylosin suggest activeinhibition of lactate producers as discussed by Nagarajaet al.et et al. (1987). As with AS-700 and TM-200, none ofthe prophylactic antibiotics influenced M. elsdenii strainNCIMB, since their influence in the combination treatmentsexamined were additive. This implies that strain NCIMBshould operate effectively in the presence of any of the threeantibiotics tested, although the preference would probablybe for zinc bacitracin and tylosin because of their stimulationof (glucogenic) propionate production (Table 4) and theirhigher total VFA production in the presence of Me. Infeedlots tylosin and monensin are used in combination(Harvey et al., 2009), which should further enhance theoutcomes of strain NCIMB administration.

4.4. Direct-fed microbials

As implicated above one way of modifying rumen fer-mentation to obtain more VFA and more propionate, is by

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125124

additive or synergetic products. Some of the tested antibio-tics, notably AS-700, zinc bacitracin and tylosin, have sup-ported total VFA in the presence of strain NCIMB andincreased the propionate proportion. Of the DFM, Propioni-bacterium spp in the absence or presence of lactobacilli isexpected to enhance propionate and reduce butyrate produc-tion (Jouany and Morgavi, 2007; Krehbiel et al., 2003). It wastherefore surprising that the combination of L. acidophilus NP51 and P. freudenreichii NP 24 (Bovamine) compared to CON,or Me or the combination with Me, did not show anyresponse either in this context or in overall fermentation(Table 5). However, either no or non-consistent response infermentation has also been reported in other studies (Jouanyand Morgavi, 2007; Raeth-Knight et al., 2007). Thus, theproduction benefits of the combination product (Brown andNagaraja, 2009; Krehbiel et al., 2004) even though there hasbeen evidence of positive responses to systemic acid basevariables (Ghorbani et al., 2002; Nocek and Kautz, 2006)must be ascribed primarily to its probiotic effects, as thepresent study could find no evidence of beneficial fermenta-tion or control of lactic acid accumulation.

The live yeast product S. cerevisiae CNCM 1-1077(Levucell) showed almost no activity either in the absenceor presence of strain NCIMB (Table 5). The exception was asmall change in the proportion of butyrate and valerate,suggesting modest alteration of ruminal fermentation andend-product formation in the presence of strain NCIMB. Byand large though there was no response by Levucellcompared to Control as can be observed from gas produc-tion, bacterial density, and lactate and VFA yields. Thereasons for the lack of response are unknown. The strictlyanaerobic conditions in the fermentation vessels mighthave been inhibitory as yeasts are aerobic and cannotsurvive for long in an anaerobic environment (Jouany andMorgavi, 2007). In the rumen their removal of oxygenfrom ingested feed supports anaerobic microbial activitycontributing to small increases in pH, cellulolytic bacteria,microbial protein synthesis and fibre digestion, but theyhave no effect on VFA, methane and rumen ammoniaproduction (Jouany and Morgavi, 2007). Yeasts in vitrocompeted with Streptococcus bovis and lactobacilli forglucose use, resulting in less lactate being produced(Chaucheyras et al., 1996). The observations, therefore,support the notion that the product does not directlyinhibit lactate accumulation, but may have positive inter-action with the M. elsdenii population and other lactateutilisers that could be useful. The conclusions of others onmode of action of yeasts (Jouany and Morgavi, 2007;Mutsvangwa et al., 1992; Rossi et al., 2004) are supported.

In contrast to S. cerevisiae CNCM 1-1077, some effects ofthe hydrolysate (EP 0946108) (Progut) were pronounced(Table 5). Compared to Control, Progut increased lactate andin combination with strain NCIMB increased gas production,butyrate and valerate and reduced the acetate-propionateratio. The results in the presence of strain NCIMB are ofparticular interest because synergy is implied. It is knownthat the yeast products provide essential nutrients to M.elsdenii (Chaucheyras et al., 1996; Rossi et al., 2004) and thisalso is the most likely explanation for hydrolysate EP0946108. According to the manufacturer of Progut thehydolysate contains peptides, mannoproteins and β-glucans

and peptides purified from a S. cerevisiae product whichspecifically have been shown to stimulate M. elsdenii growth(Rossi et al., 2004). The results support the hypothesis thatyeast products in general and S. cerevisiae hydrolysate EP0946108 in particular, render support to the lactate utilisersin the ruminal environment and thereby, indirectly, is func-tional in control of lactate acidosis.

5. Conclusions

The study confirmed the strong lactate utilisationcapabilities of M. elsdenii strain NCIMB reported in pre-vious studies. Indicator variables measured here and inprevious in vivo studies suggest that administration of thestrain could be as effective as the antibiotics regularly usedto control lactate acidosis, which is significant in the viewof pressure to ban in-feed antibiotics. In addition, theresults showed that none of the products tested wasantagonistic to strain NCIMB and can therefore be usedin association. Some antibiotics should prove useful whenused together with strain NCIMB because of additiveactions, noteworthy being AS-700, zinc bacitracin andtylosin. Of the direct-fed microbials, the results showedthat it is unlikely that the bacterial DFM L. acidophilusNP51 in combination with P. freudenreichii NP24 and theyeast DFM S. cerevisiae CNCM 1-1077 enhance fermenta-tion or support lactate utilisation on a concentrate typesubstrate, and therefore it is unlikely that they will preventlactate acidosis directly. The same applies for the yeasthydrolysate S. cerevisiae EP 0946108, but the hydrolysateapparently provides essential nutrients that can enhance theefficiency of strain NCIMB and the endogenous M. elsdeniipopulation. As with the promising antibiotics, this positiveresponse should be exploited in future research with strainNCIMB and its commercial application.

Conflict of interest statement

Megasphaera elsdenii NCIMB 41125 is a patentedorganism jointly held by the South African AgriculturalResearch Council (ARC) and Megastarter Biotech. Thisproject was funded equally by the two partners. Alimetricsprovided an analytical service and data interpretation andwere paid for their services by the partners (ARC andMegastarter Biotech). No financial benefit will accrue tothe authors for publishing this article from the partners.

The authors:Dr HH Meissner: Retired Director of the ARC. The ARC is

a government-sponsored Research Institution.Dr PH Henning: Worked as a manager for the ARC, and

later at Megastarter Biotech and is now a technical directorat Meadow Feeds and has no longer any interest in theproduct.

Mr K-J Leeuw: Employee of the ARC.Mr FM Hagg: Former employee of Megastarter Biotech,

is now employed elsewhere and has currently no links tothe organism Megasphaera elsdenii NCIMB 41125.

Dr CH Horn: Former employee of the ARC and Mega-starter Biotech, is now self-employed and has currently nolinks with the product.

H.H. Meissner et al. / Livestock Science 162 (2014) 115–125 125

Mr JHA Apajalahti: Alimetrics in Finland has no linkswith the product. Alimetrics are providers of in vitroservices and DNA analysis.

Mr A Kettunen: Alimetrics, similarly has no links to theproduct. Alimetrics are providers of in vitro services andDNA analysis.

We believe there is no conflict of interest and trust thisinformation is sufficient.

References

Allen, H.K., Levine, U.Y., Looft, T., Bandrick, M., Casey, T.A., 2013. Treat-ment, promotion, commotion: antibiotic alternatives in food produ-cing animals. Trends Microbiol. 21, 114–119.

Australian Veterinary Association – RAGFAR, 2007. Ruminal Acidosis –

Aetiopathogenesis, Prevention and Treatment. A Review for Veterinar-ians and Nutritional Professionals by the Reference Advisory Group onFermentation Acidosis of Ruminants (RAGFAR). AVA Perth, Australia.

Baldwin, R.L., McLeod, K.R., 2000. Effects of diet forage–concentrate ratioand metabolizable energy intake on isolated rumen epithelium cellmetabolism in vitro. J. Anim. Sci. 78, 771–783.

Brown, M., Nagaraja, T.G., 2009. Direct-fed microbials for growing andfinishing cattle. Presentation at the Plains Nutrition Council Meeting.Texas A&M publication AREC 09-18, San Antonio, TX, April 9–10,2009, pp. 42–61.

Callaway, T.R., Adams, K.A., Russell, J.B., 1999. The ability of “low GþCgram-positive” ruminal bacteria to resist monensin and counteractpotassium depletion. Curr. Microbiol. 39, 226–230.

Chaucheyras, F., Fonty, G., Gouet, P., Bertin, G., Salmon, J.-M., 1996. Effectsof a strain of Saccharomyces cerevisiae (Levucells SC), a microbialadditive for ruminants, on lactate metabolism in vitro. Can. J. Micro-biol. 42, 927–933.

Cogliani, C., Goossens, H., Greko, C., 2011. Restricting antimicrobial use infood animals: lessons from Europe. Microbe 6, 274–279.

Edwards, J.E., McEwan, N.R., McKain, N., Walker, N., Wallace, R.J., 2005.Influence of flavomycin on ruminal fermentation and microbialpopulations in sheep. Microbiology 151, 717–725.

Goering, H.K., Van Soest, P.J., 1970. Forage Fiber Analyses: Apparatus,Reagents, Procedures, and Some Applications. Agriculture ResearchService, U.S. Department of Agriculture. [1]. (⟨http://naldc.nal.usda.gov/catalog/CAT87209099⟩).

Ghorbani, G.R., Morgavi, D.P., Beauchemin, K.A., Leedle, J.A.Z., 2002. Effectof bacterial direct-fed microbials on ruminal fermentation, bloodvariables, and the microbial populations of feedlot cattle. J. Anim. Sci.80, 1977–1980.

Harvey, R., Funk, J., Wittum, T.E., Hoet, A.E., 2009. A metagenomicapproach for determining prevalence of tetracycline resistance genesin the fecal flora of conventionally raised feedlot steers and feedlotsteers raised without antimicrobials. Am. J. Vet. Res. 70, 198–202.

Henning, P.H., Horn, C.H., Leeuw, K.-J., Meissner, H.H., 2010. Effect ofruminal administration of the lactate-utilizing strain Megasphaeraelsdenii NCIMB 41125 on abrupt or gradual transition from forage toconcentrate diets. Anim. Feed Sci. Technol. 157, 20–29.

Hino, T., Shimada, K., Maruyama, T., 1994. Substrate preference in a strainof Megasphaera elsdenii, a ruminal bacterium, and its implication inpropionate production and growth competition. Appl. Environ.Microbiol. 60, 1827–1831.

Holben, W.E., Särkilahti, L.K., Williams, P., Saarinen, M., Apajalahti, J.H.A.,2002. Phylogenetic analysis of intestinal microflora indicates a novelMycoplasma phylotype in farmed and wild salmon. Microbe Ecol. 44,175–185.

Horn, C.H., Kistner, A., Fouché, G., 2009. Selective enrichment, isolationand characterization of fast-growing, acid-tolerant and ionophores-resistant lactate utilisers from rumen contents of animals on high-energy diets. In: Chilliard, Y., Glasser, F., Faulconnier, Y., Bocquier, F.,Veissier, I., Doreau., M. (Eds.), Ruminant Physiology – Digestion,Metabolism and Effects of Nutrition on Reproduction and Welfare,Wageningen Academic Publishers, Wageningen, The Netherlands,pp. 216–217.

Jouany, J.-P., Morgavi, D.P., 2007. Use of ‘natural’ products as alternativesto antibiotic feed additives in ruminant production. Animal 1,1443–1466.

Krehbiel, C.R., Rust, S.R., Zhang, G., Gilliland, S.E., 2003. Bacterial direct-fed microbials in ruminant diets: performance response and mode ofaction. J. Anim. Sci. 81 (E. Suppl. 2), E120–E132.

Krehbiel, C.R., Ward, J.S., Step, D.L., McBeth, L.J., Morgan, J.B., Ball, R.A.,2004. Effects of Cultures of Lactobacillus acidophilus (BG2F04) andPropionibacterium freudenreichii P-63 with or without Levucell SB(Saccharomyces cerevisiae) on Feedlot Performance, Carcass Merit,and Escherichia coli 0157:H7 Shedding by Finishing Beef Steers: OklaState University Animal Science Research Report (P-1008). Stillwater,Okla. ⟨http://www.ansi.okstate.edu/research/research-reports-1/2004⟩.

Lechtenberg, K.F., Nagaraja, T., Chengappa, M.M., 1998. Antimicrobialsusceptibility of Fusobacterium necrophorum isolated from bovinehepatic abscesses. Am. J. Vet. Res. 59, 44–47.

Marounek, M., Fliegrova, K., Bartos, S., 1989. Metabolism and somecharacteristics of ruminal strains of Megasphaera elsdenii. Appl.Environ. Microbiol. 55, 1570–1573.

Meissner, H.H., Henning, P.H., Horn, C.H., Leeuw, K.-J., Hagg, F.M., Fouché,G., 2010. Ruminal acidosis: a review with detailed reference to thecontrolling agent Megasphaera elsdenii NCIMB 41125. S. Afr. J. Anim.Sci. 40, 79–100.

Mentschel, J., Leiser, R., Mulling, C., Pfarrer, C., Claus, R., 2001. Butyric acidstimulates rumen mucosa development in the calf mainly by areduction of apoptosis. Arch. Anim. Nutr. 55, 85–102.

Mutsvangwa, T., Edwards, I.E., Topps, J.H., Paterson, G.F.M., 1992. Theeffect of dietary inclusion of yeast culture (Yea-Sacc) on patterns ofrumen fermentation, food intake and growth of intensively fed bulls.Anim. Prod. 55, 35–41.

Mutsvangwa, T., Walton, J.P., Plaisier, J.C., Duffield, T.F., Bagg, R., Dick, P.,Vessie, G., McBride, B.W., 2002. Effects of a Monensin controlled-release capsule or premix on attenuation of subacute ruminalacidosis in dairy cows. J. Dairy Sci. 85, 3454–3461.

Nagaraja, T.G., Chengappa, M.M., 1998. Liver abscesses in feedlot cattle: areview. J. Anim. Sci. 76, 287–298.

Nagaraja, T.G., Taylor, M.B., 1987. Susceptibility and resistance of ruminalbacteria to antimicrobial feed additives. Appl. Environ. Microbiol. 53,1620–1625.

Nagaraja, T.G., Taylor, M.B., Harmon, D.L., Boyer, J.E., 1987. In vitro lacticacid inhibition and alterations in volatile fatty acid production byantimicrobial feed additives. J. Anim. Sci. 65, 1064–1076.

Nocek, J.E., Kautz, W.P., 2006. Direct-fed microbial supplementation onruminal digestion, health, and performance of pre- and postpartumdairy cattle. J. Dairy Sci. 89, 260–266.

Payne, R.W., Welham, S.J., Harding, S.A., 2009. A Guide to REML. In:GenStats for Windows, 12th edition VSN International, HemelHempstead, UK.

Provenza, F.D., Villalba, J.J., 2010. The role of natural plant products inmodulating the immune system: an adaptable approach for combat-ing disease in grazing animals. Small Rumin. Res. 89, 131–139.

Raeth-Knight, M.L., Linn, J.G., Jung, H.G., 2007. Effect of direct-fedmicrobials on performance, diet digestibility, and rumen character-istics of Holstein dairy cows. J. Dairy Sci. 90, 1802–1909.

Reynolds, C.K., 2003. Splanchnic metabolism of dairy cows during thetransition from late gestation through early lactation. J. Dairy Sci. 86,1201–1217.

Rossi, F., Di Luccia, A., Vincenti, D., Cocconcelli, P.S., 2004. Effects ofpeptidic fractions from Saccharomyces cerevisiae culture on growthand metabolism of the ruminal bacteria Megasphaera elsdenii. Ann.Res. 53, 177–186.

Rowe, J.B., Morrell, J.S., Broome, A.W.J., 1982. Flavomycin as a ruminantgrowth promoter – investigation of the mode of action. Proc. Nutr.Soc. 41, A56.

Russell, J.B., Baldwin, R.L., 1978. Substrate preferences in rumen bacteria:evidence of catabolite regulatory mechanisms. Appl. Environ. Micro-biol. 36, 319–329.

Seymour, W.M., Campbell, D.R., Johnson, Z.B., 2005. Relationshipsbetween rumen volatile fatty acid concentrations and milk produc-tion in dairy cows: a literature study. Anim. Feed Sci. Technol. 119,155–169.

Van Soest, P.J., 1982. Nutritional Ecology of the Ruminant. O & B BooksInc., Corvallis Oregon, USA (p. 252).

Wileman, B.W., Thomson, D.U., Reinhardt, C.D., Renter, D.G., 2009.Analysis of modern technologies commonly used in beef cattleproduction: conventional beef production versus nonconventionalproduction using meta-analysis. J. Anim. Sci. 87, 3418–3426.

Wiryawan, K.G., Brooker, J.D., 1995. Probiotic control of lactate accumula-tion in acutely grain-fed sheep. Aust. J. Agric. Res. 46, 1555–1568.