A comparative study on the effect of subtherapeutic tylosin administration on select feral or domestic porcine gut microflora grown in continuous-flow culture

N. Ramlachan, R.C. Anderson", Kate Andrews and D.J. Nisbet

USDA-ARS, Southern Plains Agricultural Research Center, Food & Feed Safety Research Unit, College Station, TX, 77843, USA, Ph: 979-260-9429 •corresponding author· [email protected]

Abstract Continuous now cultures of feral (culture FC) and domesticated (culture RPCF) p1g gut micronora were established in steady state. Cultures, in duplicate, were continuously infused subtherapeut1c (2S lJg/ml) levels of tylosm and sampled at mtervals to assess effects on total culturable anaerobes, Bacteroides spp. and Enterococcus spp. v1a plating on serial 10-fo ld dilutions to anaerobiC Brucella blood agar, Bacteroides bile esculin agar, and M Enterococcus agar supplemented without or w1th 100 lJg tylosm/ml, the later to assess bacterial sensitivity to tylosin Concentrations of total culturable anaerobes within culture FC decreased (P < O.OS), albeit slightly, following 7 days tylosin administration. Concentrations of Bacteroides and Enterococcus decreased (P < 0 OS) to near or below detectable levels (1 .0 log1 0 CFU/ml) 1n culture FC fo llowing 7 days tylosm adm1n1stration , and tylosin-insensitive colomes were recovered at low numbers (~ 2 log1o CFU/ml) and did not persist. In contrast, concentrations of total cu lturable anaerobes, Bacteroides and Enterococcus 1n culture RPCF, while initially decreased upon initiation of tylosin admmistrat1on, began to increase (P < 0 OS) by as early as 4 days thereafter, w1th tylosin­ Insensitive colonies recovered as one of predominant populations. The results of this study illustrate that under the conditions of th1s test, subtherapeutic administration of tylosin promoted the ennchment of tylosm-1nsensit1ve bacterial populations (capable of growing on media supplemented w1th 100 lJg tylosin/ml) within RPCF cultures (originating from a traditionally reared domesticated p1g) but not from FC cultures (orig1nat1ng from a feral pig).

Introduction Macrolide ant1b1ot1cs are commonly used m human and vetennary med1cme, pnmarily to treat infections caused by Gram-pos1tive bacteria and also as a feed additive to improve product1on effic1ency in swine (Gaynor and Mank1n, 2003). Res1stance can occur via acquisition of erm methyltransferases, wh1ch catalytically Inactivate the macrolide's targeted bmdmg site, via acquisition of multidrug efflux pumps or even, albeit infrequently, via po1nt mutations m the microbe's genome (Chopra and Roberts, 2001 , Gaynor and Mankin, 2003; Karlsson et al. , 2004; Poole, 200S) Recovery of bactena harbouring erm genes from domestic swine and swine production hab1tats is not uncommon (Chee-Sanford et al. , 2001, Wang et al , 200S) but less 1s known regarding the quantitative acquisition and selection of res1stance in bacteria, particularly withm swme not reared traditionally (Stanton and Humphrey, 2004). Contmuous now culture of Intestinal microorganisms has been used to study competitive interactions between commensal and pathogemc micronora (Harvey eta/., 2002, Hume eta/, 2001 , N1sbet eta/ , 2000) as well as to investigate potential factors affecting spontaneous acquisition of antibiotic resistance (Kim et at., 200S) In th1s study, a contmuous now chemostat model established w1th m1xed populations of porcine gut bactena was used to assess the effects of subtherapeutic tylosin adm1mstrat1on on select populations of resident bactena.

Materi als and Methods Two separate m1xed populations of porcine gut bactena were established in continuous now culture as previously described (Harvey et al 2002, Hume et al 2001) The RPCF culture had been previously established w1th cecal contents obtamed from a traditionally reared pig and its Initial characterization has been reported previously (Harvey et al 2002) The other culture defined as FC was established under s1mllar cond1t1ons except with cecal contents from an

s 10n An •microbial res•sl nee Safcpork 2007- Verona (Italy 481 adolescent fera l boar killed near Caldwell, Texas, USA, approximately 2 to 4 h prior to necropsy. Both parent cu ltures of RPCF and FC were established and maintained in BioFio chemostats (New Brunswick Scientific Company, Edison, NJ) with a culture volume of 550 mi. The culture medium was Viande Levure broth which was prepared and maintained anaerobically under a stream of 0 2- free C02 and infused at 0.40 ml/min which corresponds to a 24 h vessel turnover. Cultures were incubated at 39°C and agitated at 100 rpm. Once established in steady state, initial characterization of culture FC was accomplished using traditional bacteriological culture methodologies and antibiotic susceptibility testing was performed as described in the National Committee for Clinical Laboratory Standards (now known as the Clinical and Laboratory Standards Institute [CLSI]) (NCCLS, 2004). Both parent cultures were used to provide inoculum to establish separate RPCF and FC cultures, in duplicate, which after at least 14 vessel turnovers, were continually infused with culture medium containing 25 IJg tylosin/mi. Fluid samples collected immediately before and during tylosin infusion were quantitatively cultured, via plat1ng of 10-fold serial dilutions, to the following media, each prepared with or without 100 IJg tylosin/ml: anaerobic Brucella blood agar and Bacteroides bile esculin agar (Anaerobe Systems, Morgan Hill, CA), for detection of total anaerobes and Bacteroides spp., respectively, and M Enterococcus agar (Becton Dickinson and Company, Sparks, MD) for detection of Enterococcus spp. Inoculated media were incubated 48-72 h at 37 oc and colonies propagated with and without tylosin selection were enumerated. Specific identification of bactena from select colonies was achieved using rapid ID 32 STREP, rapid 20E, 20NE, 20A, and rapid ID 32 A identification strir,s (bioMerieux, Hazelwood, MO). Indole spot tests (Anaerobe Systems, Morgan Hill, CA), E-test M (AB Biodisk, Piscataway, NJ) and gas chromatography were also used in this analysis. Samples containing no detectable colonies of bacteria were g1ven a value of 1.0 log 10 CFU/ml. Log 10 transformations of bacterial concentrations obtained from duplicate cultures were analyzed for main effects of day, culture type (i.e., RPCF or FC culture) and the possible interaction using a repeated measures analysis of variance (Statistix®8 Analytical Software, Tallahassee, FL, USA). Multiple comparison of means was accomplished usmg a Tukeys procedure.

Results and Discussion Characterization of microflora in culture FC. Culture FC achieved steady state after 14 days continuous now culture and the bacteriological composition was found to include Streptococcus bovis, Proteus mirabilis, Staphylococcus epidermidis, Alcaligenes denitrificans, and members of Bacteroides, Lactobacillus, Enterococcus, and Clostridium. Campylobacter and Salmonella were never detected in the feral culture and E. coli that was initially isolated from the cecal contents was never recovered once the culture had achieved steady state. Enterococcus hirae, Streptococcus bovis, Proteus mirabilis, Staphylococcus epidermidis, Alcaligenes denitrificans were susceptible to tylosin and erythromycin; Bacteroides uniformis and Bacteroides stercoris were resistant to gentamicin, c1pronoxacin , ceftriaxone and ampicillin. Clostridium hathewayi showed resistance to tylosin at MIC >512 IJg/ml and was the predominant anaerobe recovered on anaerobiC Brucella blood agar.

Effect of tylosin on select bacterial populations. Recovery of anaerobes from non-selective Brucella blood agar was not affected (P > 0.05) by culture; however, main effects of day and day by culture interaction were observed (P < 0.01) on recovery of anaerobes from nonselective Brucella blood agar due to a decrease in anaerobes in the FC and an increase in RPCF cultures (Figure 1A) . Main effects of culture, day and a day by culture interaction were observed (P < 0.01) on recovery of anaerobes from tylosin-selective Brucella blood agar due to a temporary increase m tylosin-insensitive anaerobes, which were prominent even before tylosm admm1stration, from the FC culture and a gradual enrichment of tylosin-insensitive anaerobes from the RPCF culture (Figure 1A) . The prominent tylosm-1nsensitive anaerobe was Clostndtum hathewayi. Main effects of culture, day and a day by culture interaction were observed (P < 0.05) on recovery of Bacteroides from Bacteroides bile esculin agar supplemented with or without tylosm (Figure 1B) due to higher Bacteroides concentrations in the RPCF cultures prior to administration of tylosin and to an enrichment of tylosin-insensitive Bacteroides spp beginning by day 3 of tylosin administration (Figure 1 B). In the case of total anaerobes and Bacteroides, tylosin-insensitive populations were prominent in RPCF cultures even before initiation of tylosin administration. Conversely, tylosin­ Insensitive Enterococcus spp , were not apparent prior to tylosin admmistratian, however, upon

482 of pork 200 7 V rona (II, ly) 1n1t1at1on of treatment recovery was h1ghly variable between the two RPCF cultures regardless of culturing on M Enterococcus agar supplemented with or without tylosin. This indicates that the two RPCF cultures contained markedly different enterococcal populations. For instance, even though mean concentrations of tylosin-insensitive Enterococcus recovered from the RPCF cultures began to increase markedly beginnmg 4 days after initiation of tylosin administration (Figure 1C) this increase occurred in only one of the cultures. As a consequence, main effects of culture, day or day by culture interactions on quantitative recoveries on M Enterococcus agar supplemented with or without tylosin were not observed (P > 0.05) (Figure 1C) .

In conclusion, results from this study revealed that subtherapeutic administration of tylosin promoted the enrichment of tylosin-insensitive bacterial populations (capable of growing on media supplemented with 100 j.Jg tylos1n/ml) Within RPCF cultures (originating from a traditionally reared domesticated pig) but not from FC cultures (originating from a feral p1g).

Acknowledgements The authors thank M. Reiley Street, Clayton Myers, Jim Snodgrass and Matthew Quatrini for technical support This work was funded in part by the National Pork Board.

References CHEE-SANFORD, J.C., AMINOV, R.I., KRAPAC, I.J., GARRIGUES-JEANJEAN, N., MACKIE, R.I. , 2001 . Occurrence and diversity of tetracycline resistance genes in lagoons and groundwater underlymg two swine production facilities. Applied and Environmental Microbiology, 67,1494-1502. CHOPRA, I , ROBERTS, M., 2001 Tetracycline res1stance: mode of action, applications, molecular biology, and epidemiology of resistance. Microbiology and Molecular Biology Reviews, 65,232-260 GAYNOR, M., MANKIN, A.S., 2003. Macrolide antibiotics: binding s1te , mechan1sm of act1on , resistance. Current Topics in Medicinal Chemistry, 3,949-961 . HARVEY, R.B., DROLESKEY, R.E., HUME, M.E., ANDERSON, R C., GENOVESE, K.J., ANDREWS, K., NISBET, D.J., 2002. In vitro •nh1bit1on of Salmonella enterica serovars Choleraesuis and Typhimurium, Escherichia coli F-18, and Escherichia coli 0157:H7 by a porcine contmuous-flow competitive exclus1on culture. Current Microbiology, 45:226-229. HUME, ME , NISBET, D.J . BUCKLEY, S.A. , ZIPRIN, R L. , ANDERSON, R.C ., STANKER, L.H ., 2001 Inhibition of in vitro Salmonella Typhimurium colonization 1n porcine cecal bacteria continuous-flow compet1t1ve exclusion cultures. Journal of Food Protection, 64,17-22. KARLSSON, M . FELLSTROM, C., JOHANSSON, K.-E., FRANKLIN, A , 2004 Antim1crob1al res1stance m Brachyspiira pilos/coli with spec1al reference to p01nt mutations in the 23S rRNA gene associated with macrolide and lincosam1de resistance. Microbial Drug Res1stance, 10,204-208. KIM, WK., KARABASIL, N., BULAJIC, S , DUNKLEY, K.D , CALLAWAY, T.R , POOLE, T.L , RICKE, S.C., ANDERSON, R.C., NISBET, D.J , 2005 Comparison of spontaneous antibiotic frequency of Salmonella Typh1murium growth in glucose amended contmuous culture at slow and fast dilution rates. Journal of Environmental Science and Health Part 8, 40,4 75-484 NCCLS, 2004 Methods for Antim1crob1al Susceptibility Testing of Anaerobic Bacteria, Approved Standard, 61h Ed., NCCLS document M11-A6, Wayne, PA. National Committee for Cl1n1cal Laboratory Standards NISBET, D J , ANDERSON, R C , CORRIER, DE., HARVEY, R B , STANKER, L H , 2000 Modeling the survivability of Salmonella typh1murium m the ch1cken cecae usmg an anaerobic continuous-culture of ch1cken cecal bacteria Microbial Ecology m Health and Disease, 12,42-47 POOLE, K , 2005. Effiux-med1ated antimicrobial resistance. Journal of Ant1m1croblal Chemotherapy, 56,20-51 STANTON, T.B , HUMPHREY, S B . 2004 Tetracycline resistant bacteria 1n organically ra1sed and feral swme Abstract of the 104th American Soc1ety for Microbiology General Meeting, Z-029 WANG Y WANG, G . SHOEMAKER. N B , WHITEHEAD, T R. , SALYERS, A.A , 2005 Distribution of the ermG gene among bactenal ISolates from porcme mtest1nal contents Applied and Environmental Microbiology, 714930-4934

S s 1on 6 nhmlcrobi I r s•stancc Sa{cpork 2007 -Verona (Italy 483 -<>-Anaerobes recowred on non-selectiw Brucella blood agar A 12 - Anaerobes recowred on tylosin-supplemented (100 ugJml) Brucella blood agar -o-Anaerobes recowred on non-selectl\19 Brucella blood agar - Anaerobes recowred on tylosin-supplemented (100 uglml) B rucella blood agar 10

8 t 6 6 Recowry on Brucella blood agar wolhout !)los in ~ Cullure effect P = 0 4082. SEM =0 19 4 DayeffectP =0 1305 SEM= 0 22 Day X Culture onteractoon P = 0 0006 SEM = 0 31 Recowry on Brucella blood agar with !)los in 2 Culture effect P =0 3293. SEM = 0 30 Day effect P = 0 0889 SEM = 0 33 Day X Culture onteractoon P =0 0001 SEM =0 47 0 -4 0 4 8 12 16 20 24 28

-<>-Bacteroides reco red on Bacteroides bile esculin agar not supplemented woth tylosin B 10 - Bacteroides reco red on tylosin-supplemented (100 ug/ml) Bacteroides bile esculin agar -o-Bacteroldes reco red on Bacteroides bile esc ulin agar not supplemented woth tylosin 9 - Bacteroides reco\lered on tylosin-supplemented (100 ug/ml) Bacteroides bile esculin agar

8

7

6 Recowry on Bacteroodes bole esculin agar wllhout t)4os ln Culture effect P = 0 6362 SEM = 0 12 ~ 5 Day effect P < 0 0001 , SEM = 0 18 Day X Culture Interaction P =0 0006: SEM =0 26 j 4 Recowryon B cteroodes bole esculin ag rwllh !)los on Culture effect P = 0 2796 SEM = 0 10 3 D yeffect P < 0 0001 SEM = 0 19 Day X Culture in tero c~on P = 0 0023. SEM = 0 27 2

0 -4 0 4 8 12 16 20 24 28

-<>-Enterococcus reco red on M Enterococcus agar not supplemented wlth tylosin c 10 - Enterococcu reco r don tylosin-supplemented (100 uglml) M Enterococcus agar -o-Enterococcus r

8

7

t 6 6 5 Recovery on M Enterococcus qor wothoul !)los on j 4 Cullure !feet P - 0 11 36. SEM 1 02 3 D yeffect P 0 0249 SEM 0 69 D y Culture tn t r ClJon P 0 0996. SEM ~ 0 97 2 Recowry on M Enterococcus agar woth t)4o In Culture effect P ~ 0 H60. SEM • 1 06 D yoff ctP-0221 3 SEM 064 D y Cullur Interaction P 0 1777 SEM 0 90 0 -4 0 4 8 12 16 20 24 28 Day or In ub tlon

Figure 1. Effects of sublherapeut1c (25 IJg/ml) tylosm administration on populations of anaerobes, Bacterotdes spp , and Enterococcus spp from continuous now cultures of mixed populations of porcme gut baclena obtained from fera l (culture FC, circles) or domestic (culture RPCF squares) swine Bacteria were quantitatively recovered on anaerobic Brucella blood agar, Bacteroides b1le esculin agar and M Enterococcus agar each supplemented without (open symbols) or with (closed symbols) 100 1-Jg tylosln/ml

484 f pork 2007 V ron, (Italy) s I n 6: Antlrn crobJ I r