In vitro comparison of bovine mastitis and fecal Escherichia coli isolates

ELSEVIER

Veterinary Microbiology40 (1994) 231-238

In vitro comparison of bovine mastitis and fecal Escherichia coli isolates J. Nemeth, C.A. Muckle, C.L. Gyles* Department of Veterinary Microbiology and Immunology, Ontario Veterinary College, Universityof Guelph, Guelph, Ont., Canada, NI G 2W1 (Received 10 March 1993; accepted 26 October 1993)

Abstract In vitro methods were used to test the hypothesis that Escherichia coli from bovine mastitis are essentially no different from isolates from bovine feces. Fifty E. coli isolates from bovine mastitic milk, 50 from feces of mastitic cows and 50 from feces of healthy cows were compared with respect to biochemical properties and certain potential virulence factors. There were no significant differences among the groups in tests for biotype; production of colicins, colicin V, or Veto cell cytotoxicity; and growth in 90% gnotobiotic calf serum or 90% normal milk whey. Resistance to killing in 90% gnotobiotic calf serum varied from 66 to 84%. Most isolates grew in normal whey: the percentage in a group varied from 86 to 96. Mastitic milk isolates were significantly different from the fecal isolates in adonitol fermentation (P < 0.006), production of aerobacffn (P < 0.026), and ability to grow in 90% mastitic whey (P<0.00004). However, only 40% of mastitis E. coli fermented adonitol and only 20% produced aerobacffn. Ninety-six percent of mastitic milk E. coli grew in mastitic whey, whereas 64% and 60%, respectively, of mastitic fecal and normal fecal isolates grew in this medium. It is concluded that none of the properties that were investigated constitute potential virulence factors or markers for ability to induce mastiffs; the data are consistent with the hypothesis that mastitic E. coli are simply opportunistic pathogens. Keywords: Escerichiacoli; Cow; Mastitis,bovine;Fecal isolate

1. Introduction The proportion of mastitic infections due to Escherichia coli has increased in recent years, mainly as a result of improved control programs against mastitis caused by Gram-positive bacterial pathogens (Jones, 1986). This development has sparked considerable interest in *Correspondingauthor. 0378-1135/94/$07.00 1994ElsevierScienceBN. SSD10378-1135 (93)E0154-A

232

J. Nemeth et al. / VeterinaryMicrobiology 40 (1994) 231-238

seeking potential virulence factors of mastitic E. coli, but none have been discovered (Wray et al., 1983; Sanchez-Carlo et al., 1984a, 1984b; Barrow and Hill, 1989; Nemeth et al., 1991 ). The only potential virulence factor associated with E. coli from bovine mastitis was the ability to resist serum-killing. E. coli isolated from bovine mastitis have been examined for properties associated with enteric or extraintestinal disease. Besides serum resistance, factors that have been investigated include production of enterotoxin and verotoxin (Sanchez-Carlo et al., 1984a; Barrow and Hill, 1989); colicin and aerobactin (Linggood et al., 1987); capsule (Wray et al., 1983; Barrow and Hill, 1989; Nemeth et al., 1991) and adhesive structures (Frost, 1975; Harper et al., 1978; Frost et al., 1982) ; growth in normal and mastitic whey (Breau and Oliver, 1986; Kaartinen et al., 1989; Todhunter et al., 1990); and association with serotype (Linton et al., 1979; Sanchez-Carlo et al., 1984b). In no case was there an association of one property with a high percentage of E. coli from mastitis. Furthermore, there was no clustering of isolates within a limited number of serotypes as has been observed for E. coli that cause enteric, urinary tract, or septicemic disease; instead, the mastitic isolates belonged to a large number of serotypes. Studies cited above examined E. coli isolates from bovine mastitis for common properties that may be expected to be virulence factors or virulence markers. The objective of the present study was to compare E. coli isolated from mastitic milk with those from the feces of mastitic and healthy cattle to determine whether there were any differences in biochemical reactions and/or potential virulence properties among the groups. Isolates were examined for adonitol fermentation, biotype, colicinogeny, colicin V production, presence of aerobactin receptor, production of Vero cell cytotoxins, and the ability to resist killing by serum and by normal and mastitic whey.

2. Materials and methods 2.1. Bacterial strains

Fifty E. coli from mastitic milk, 50 from feces of mastitic dairy cows, and 50 from feces of healthy cows were isolated from samples collected from dairy farms in Southwestern Ontario. One mastitic milk isolate and one fecal isolate were obtained from a single animal and a fecal isolate was collected from a healthy animal of approximately the same age and stage of lactation in the same herd. All isolates were maintained at - 70°C in a medium that consisted of 10% skim milk, 7.5% glucose, 10% sucrose, and 1% (w/v) bovine plasma albumin in distilled water. 2.2. Biotypes o f isolates

The Replianalyzer system (Cathra Gram Negative Identification System, MCT Medical, Inc., St. Paul, Minn.) was used to identify isolates and to determine their biotypes. This system determined the following: esculin and arginine hydrolysis; fermentation of glucose, sucrose, lactose, arabinose, cellobiose, inositol, mannitol, rhamnose, and sorbitol; production of indole, lysine decarboxylase, ornithine decarboxylase, and hydrogen sulfide; utili-

J. Nemethet al. / VeterinaryMicrobiology40 (1994) 231-238

233

zation of citrate or malonate as the sole source of carbon; and ability to grow in the presence of bile, nalidixic acid, nalidixic acid and colistin, and cetrimide and kanamycin. 2.3. Adonitol fermentation Adonitol fermentation was determined in phenol red broth base containing 1% adonitol (Difco, Detroit, Mich.). 2.4. Colicin and colicin V production Colicin production was determined by the soft-agar overlay technique of Fredericq (1957), with E. coli K-12 as the indicator strain. Colicin-producing isolates were tested for ability to produce colicin V by repeating the assay with the colicin V-positive E. coli strain B 188 as the indicator strain. 2.5. Aerobactin production E. coli isolates were tested for the ability to provide aerobactin to a mutant strain of E. coli, which requires aerobactin for growth on iron-deficient medium. The method was as described by Carbonetti and Williams (1985). 2.6. Vero cell cytotoxicity of culture supernatants arm polymyxin B extracts Isolates were tested for Vero cell cytotoxicity following the protocol of Gannon et al. (1988). Cell culture supernatants of all isolates were tested, then polymyxin B extracts of cells were tested for isolates whose supernatants were positive. 2. 7. Serum resistance The isolates were tested for resistance to gnotobiotic calf serum according to the protocol of Helmuth et al. (1985). Approximately 105 cfu of each culture in L-broth was incubated in 90% serum for 3 h. The numbers of viable bacteria in the original culture (time 0) and in the serum culture (time 3 h) were determined by plating dilutions onto L-agar plates and incubating the plates overnight at 37°C. Heat-inactivated serum was used as a control to assess whether non-complement serum factors were affecting the growth of the isolates. The numbers of cfu/ml in the 3-hour unheated and the original inoculum were compared. Isolates for which the 3-hour cfu/ml in unheated serum or whey was greater than that of the 0-hour control culture were determined to be resistant to serum. Isolates for which the 3-hour cfu/ml in unheated serum was less than that of the 0-hour control sample were determined to be sensitive to serum. E. coli B 117, a serum-resistant strain and E. coli K- 12 (PACRM), a serum-sensitive organism, were used as controls. 2.8. Growth in whey Approximately 1 L of normal milk was collected from the mammary gland of a healthy, lactating cow with a somatic cell count of less than 250000 cells/ml. Mastitic milk was

J. Nemeth et al. / Veterinary Microbiology 40 (1994) 231-238

234

prepared by infusing 10 ml of saline containing 10/xg of endotoxin (Sigma Chemical Co., St. Louis. Mo.) into one quarter of the mammary gland of the same cow and collecting the mammary secretion 24 h later. Both samples were transported on ice, then tested for California Mastiffs Test (CMT) scores and streaked onto blood agar plates, which were incubated overnight at 37°C. The normal milk had a negative CMT value, while the mastitic milk had a CMT value of 2. No microbial growth was found on the blood agar plates after the incubation period. Whey was immediately prepared from both milk samples following the protocol of Kaartinen and Pyorala (1989). Resistance to normal and mastitic whey was determined as described for resistance to serum bactericidal effects.

2.9. Statistical analysis Differences among the three groups with respect to numbers of isolates possessing various traits were assessed using a binomial test for proportions ( a = 0.05) (Fleiss, 1981 ).

3. Results and discussion

E. coli from bovine mastitis did not belong to unique biotypes, when compared with fecal isolates. Most isolates belonged to three of nine biotypes that were identified (Fig. 1). Significantly more of the mastitic milk (60%) and the control fecal isolates (48%) belonged

60 ILl

..~50 0 cn 40 It.

0 14.1

0 30 z

U.I

0

20

nILl 13. 10

2

I

I

3

4

I

I

I

I

5 6 BIOTYPES

I

7

8

9

Fig. 1. Distribution of biotypes among E. coli from mastitic milk ([]), from the feces of mastitic cows ( [] ) and from the feces of normal cows ( Q ) . There were 50 samples in each group.

J. Nemeth et al. / Veterinary Microbiology 40 (1994) 231-238

235

Table 1 Comparison of three groups of bovine E. coli isolates in selected in vitro tests Group of E. coli isolates

Percentageof isolates in a group (n=50) that were positive for Adonitol Colicin ColicinV

Mastitic milk 40~ Fecal, mastitic 10b

Aerobactin V e r o S e r u m Growthin toxin resistance normal whey

Growthin mastitic whey

14 14

6 8

20e 4d

8 2

76~ 84f

96 90

968 64"

26

14

0d

0

66e

86

60n

COWS

Fecal, healthy

14b

COWS

1Foreach property,absenceof superscripts indicatesno significantdifferencesamongthe threegroups;superscripts indicate that one group is significantlydifferentfrom the two which have the same letter.

to biotype 1 than did mastitic fecal isolates (18%) ( P = 0 . 0 0 0 , 0.002 respectively). The distinguishing feature of biotype 1, compared with biotype 2, was that the isolates failed to decarboxylate ornithine. Interestingly, 74.4% of isolates from mastitis in herds in Iowa decarboxylated ornithine, indicating that strains may vary with location. Also, the biotype data suggest that the fecal and mastitic isolates from the same cow were often different. Table 1 shows the ability of the isolates to ferment adonitol, produce colicin, colicin V, aerobactin, and verotoxin, and to grow in serum, normal whey, and mastitic whey. Adonitol fermentation was detected for 40% of the E. coli from mastitic milk; this was a significantly higher percentage than for isolates from the other sources. The 40% is similar to that reported by Sanchez-Carlo et al. (1984a), but the findings do not support the suggestion of these workers that ability to ferment adonitol may be a marker for bovine fecal E. coli. Possibly, this characteristic is a marker for a subset of masfftis isolates. Colicin V was of particular interest because the genes for its synthesis are carried on plasmids which also carry genes for properties associated with pathogenic E. coli, including the aerobacffn iron uptake system, increased survival in serum, resistance to phagocytosis, hydrophobicity, and intestinal epithelial cell adherence. However, colicin production was not common among the E. coli examined, a finding similar to that of other researchers who examined E. coli from bovine mastiffs (Wray et al., 1983; Barrow and Hill, 1989). The results of the present and previous studies indicate that colicin V-associated factors are not important in the virulence of mastitic E. coli. Only 20% of E. coli from bovine mastiffs produced aerobacffn but this was significantly higher than for E. coli from the other groups. Similarly, Linggood et al. (1987) found that only 12% of 148 E. coli from mastiffs in cattle produced aerobactin while 85% of 54 isolates from septicemic cattle did. In 73 fecal isolates from healthy cows they found only 4% that produced aerobactin. Although aerobactin has been associated with extraintestinal E. coli pathogens, enterochelin or some unidentified siderophore may be sufficient for mastitic E. coli to obtain iron from lactoferrin. As verotoxigenic E. coli are isolated from the feces of cattle, the possibility that Vero cell cytotoxins may be produced by mastitic E. coli was explored; the cytotoxicity could account for the epithelial lesions seen in the mammary glands. However, only 8% of mastitic

236

J. Nemeth et al. / Veterinary Microbiology40 (1994) 231-238

isolates examined in the present study were cytotoxic. Barrow and Hill (1989) had previously found that only one strain (0.5%) of 237 mastitic E. coli produced verocytotoxins. Cytotoxic necrotizing factor (CNF) (DeRycke et al., 1990) is another cytotoxin which could be responsible for the epithelial cell necrosis, but no one has examined this possibility. The findings on serum resistance (Table 1) are consistent with previous reports that serum resistance varies between 64% and 100% for E. coli isolated from bovine mastitis, and between 53% and 100% for fecal isolates (Carroll and Jasper, 1977; Sanchez-Carlo et al., 1984; Barrow and Hill, 1989; Nemeth et al., 1991). All isolates grew well in heatinactivated serum, indicating that the bactericidal components were heat-labile. Despite variations in percentages, a consistent pattern has emerged: the majority of E. coli isolates from bovine mastitis and from bovine feces are serum-resistant. Compared with serum, whey and mastitic whey more closely approximate the environment in which bacteria must grow during mastitis. Normal and mastitic whey differ in their anti-bacterial properties, including availability of iron (Rainard, 1986), levels of lysozyme and immunoglobulin (Nickerson, 1985), and amount of complement (Mueller et al., 1982). Increases in antimicrobial components in mastitis may suffice to inhibit the growth of nonpathogenic bacteria while pathogens would presumably possess factors for survival and growth in this hostile environment. There were no significant differences in resistance to normal whey among the three groups of isolates (Table 1 ). All organisms susceptible to normal whey were also susceptible to mastitic whey except for one mastitis fecal isolate. However, the mastitic milk isolates were significantly more resistant to mastitic whey than were the mastitis fecal (P = 0.00004) and control fecal isolates (P = 0.00001). The mastitis fecal and control fecal isolates demonstrated significantly greater growth in normal whey than in mastitic whey (P = 0.006, 0.004, respectively). Isolates that were susceptible to serum or whey had reductions in numbers (compared to time 0) which ranged from one-fourth to > 10 -5 at 3 hours. Compared with the growth in unheated serum, the reductions ranged from approximately 10 -1 to 10 -6. There are conflicting reports on the ability of E. coli from bovine mastitis to grow in mastitic secretions. Malkamaki et al. (1986) found that whey samples from inflamed quarters promoted the growth of two mastitis E. coli isolates when compared with whey from control quarters. Kaartinen and Pyorala (1989) examined growth of a mastitis E. coli isolate in 80% normal whey samples taken throughout the lactation period and found a slight inhibition of growth. Carroll et al. (1973) reported that only serum resistant mastitis pathogens grew in 90% normal whey. Lohuis et al. ( 1988, 1990) found inhibition of two strains of E. coli from bovine mastitis in 90% milk and in 90% whey collected 18 h and 36 h after infusion of the mammary gland with 0.1 mg lipopolysaccharide. Mattila et al. (1985) found an initial bacteriostasis followed by growth promotion after infusion of 0.1 mg of endotoxin into the quarters of dairy cows over several hours. These studies used a single E. coli isolate. Variations among these reports and in comparison with the present study are undoubtedly due to differences in type of mastitis (acute or chronic), strains of E. coli, stage at which mastitic milk was collected, and duration of exposure of bacteria to whey. In the present study, normal whey was generally not inhibitory to bacterial growth, whereas mastitic whey collected 24 hours after infusion of endotoxin was often inhibitory. Furthermore, E. coli from mastitic milk usually grew in mastitic whey, whereas significantly

J. Nemeth et al. / Veterinary Microbiology 40 (1994) 231-238

237

lower percentages ofE. coli from the other groups were able to grow. Since the fecal isolates from mastitic cows and those from control fecal samples were similarly sensitive to mastitic whey, it cannot be argued that cows develop E. coli mastitis because of their fecal carriage of E. coli resistant to mastitic whey. Perhaps growth in mastitic secretions requires the selection of resistant organisms. In the absence of concurrent infections, bacteria entering the gland would be exposed to normal milk. Initially, therefore, most isolates would be able to multiply and cause mastitis. However, once a response was initiated in the gland, changes in composition resulting in mastitic milk would take place and only resistant bacteria would survive in the gland. In conclusion, this study found that E. coli isolated from mastitic milk did not differ from fecal E. coli except in their ability to ferment adonitol, to produce aerobactin production, and to grow in mastitic whey. However, only low percentages of mastitis isolates fermented adonitol or produced aerobactin, and these properties appear to be neither virulence markers nor potential virulence factors. Ability of a high percentage of E. coli from mastitic milk to grow in mastitic whey may represent selection of fecal isolates with this ability. The data from this study are consistent with the notion that E. coli from bovine mastits are not a group of isolates which possess special virulence attributes.

References Barrow, P.A. and Hill, A.W., 1989.The virulencecharacteristicsof Escherichia coli isolatedfromcases of bovine mastiffs in Englandand Wales. Vet. Microbiol.,72: 3548. Breau, W.C. and Oliver. S.P., 1986. Growth inhibitionof environmentalmastitispathogens during physiologic transitionsof the bovine mammarygland. Am. J. Vet. Res., 47: 218-222. Carbonetti,N.H. and Williams,P.H., 1985. Detectionof synthesisof the hydroxamatesiderophoreaerobactinby pathogenicisolatesof Escherichia coli. In: M. Sussman(Editor). The Virulenceof Escherichia coli. pp. 419424. Carroll, E.J. and Jasper, D.E., 1977. Bactericidal activity of standard bovine serum against coliformbacteria isolated from udders and the environmentof dairy cows. Am. J. Vet. Res., 38: 2019-2022. Carroll, E.J., Jain, W.C., Schalm, O.W. and Lasmanis, M.J., 1973. Experimentallyinduced coliformmastitis: Inoculationof adders with serum-sensitiveand serum-resistantorganisms.Am. J. Vet. Res., 34:1143-1146. De Rycke, J., Gonzalez,E.A., Blanco,J., Oswald, E., Blanco, M. and Boivin,R., 1990. Evidencefor two types of cytotoxic necrotizingfactor in human and animalclinicalisolates of Escherichia coli. J. Clin. Microbiol., 28: 694499. Fleiss, J.L., 1981. Assessingsignificancein a four-foldtable. In: StatisticalMethods for Rates and Proportions. 2nd edn. J. Wileyand Sons, Toronto,p. 19-32. Fredericq, P., 1957. Colicins.Ann. Rev. Microbiol., 11: 7-22. Frost, A.J., 1975. Selectiveadherenceof microorganismsto the ductularepitheliumof the bovinemammarygland. Infect. Immun., 12:1154-1156. Frost, A.J., Hill, A.W. and Brooker, B.E., 1982. Pathogenesisof experimentalbovine mastitisfollowinga small inoculumofEscherichia coli. Res. Vet. Sci., 33:105-112. Gannon,V.P.J., Gyles,C.L. and Friendship,R.W., 1988. Characteristicsof VerotoxigenicEscherichia coli from pigs. Can, J. Vet. Res., 52: 331-337. Harper, M., Turvey, A. and Bramley, A.J., 1978. Adhesionof fimbriateEscherichia coli to bovine mammarygland epithelialcell in vitro. J. Med. Mierobiol.,11:117-123. Helmuth, R., Stephen,R., Bunge,C., Hoog, B., Steinbeck,A. and Bulling,E., 1985. Epidemiologyof virulenceassociated plasmidsand outer membraneproteinpatterns withinsevencommonSalmonella serotypes. Infect. Immun.,48: 175-182. Jones, T.O., 1986. A review of teat factors in bovineE. coli mastitis.Vet. Rec., 118: 507-509.

238

J. Nemeth et al. / Veterinary Microbiology 40 (1994) 231-238

Kaartinen, L. and Pyorala, S., 1989. Changes in milk plasminogen, plasmin and in vitro bacterial growth in whey during early lactation. J. Dairy Res., 56: 719-725. Kaartinen, L., Ali-Vehmas, T., Mattila, T. and Sandholm M., 1989. Bacterial growth in mastitic whey in relation to bacterial association with mastiffs. Vet. Microbiol., 2t: 155-163. Linggood, M.A., Roberts, M., Ford, S., Parry, S.H. and Williams, P.H., 1987. Incidence of the aerobactin iron uptake system among Escherichia coil isolates from infections of farm animals. J. Gen. Microbiol., 133: 835842. Linton, A.H., Howe, K., Sojka, W.J. and Wray, C., 1979. A note on the range of Escherichia coli O-serotypes causing clinical bovine mastitis and their antibiotic resistance spectra. J. Appl. Bacteriol., 46: 585-590. Lohuis, J.A.C.M., VanLeeuwen, V., Verheijden, J.H.M., Smit, J.A.H., Van Brand, A. and Van Miert, A.S.J.P.A.M., 1988. Growth ofEscherichia coli in whole and skim milk from endotoxin-induced mastitic quarters: In vitro effects of deferoxamine, zinc and iron supplementation. J. Dairy Sci., 71: 2772-2781. Lohuis, J.A.C.M., Kremer, W., Schukken, Y.H., Smit, J.A.H., Verheijden, J.H.M., Brand, A. and Van Miert, A.S.J.P.A.M., 1990. Growth of Escherichia coli in milk from endotoxin-induced mastitic quarters and the course of subsequent experimental Escherichia coli mastitis in the cow. J. Dairy Sci., 73: 1508-1514. Maikamaki, M., Mattila, T. and Sandholm, M., 1986. Bacterial growth in mastitic milk and whey. J. Vet. Med. B, 33: 174-179. Mattila, T., Honkanen-Buzalski, T. and Sandholm, M. 1985. Depressed bacterial growth in whey during endotoxin induced mastitis. Zentralbl. Veterinarmed. B, 32: 85-92. MueUer, R., Carroll, E.J. and Panico, L. 1982. Complement C3 levels and haemolytic activity in normal and mastitic milk whey. Zentralbl. Veterinarmed. B, 29: 99-106. Nemeth, J., Muckle, C.A. and Lo, R.Y.C., 1991. Serum resistance and the traT gene in bovine mastitis-causing Escherichia coli. Vet. Microbiol., 28:343-351. Nickerson, S.C. 1985. Immune mechanisms of the bovine udder: An overview. J. Am. Vet. Med. 187: 41-45. Rainard, P. 1986. Bacteriostatic activity of bovine milk lactoferrin against mastitic bacteria. Vet. Microbiol. 11: 387-392. Sanchez-Carlo, V., McDonald, J.S. and Packer, R.A., 1984a. Virulence factors of Escherichia coli isolates from cows with acute mastitis. Am. J. Vet. Res., 45: 1775-1777. Sanchez-Carlo, V., Wilson, R.A., McDonald, J.S. and Packer, R.A., 1984b. Biochemical and serological properties ofEscherichia coli isolated from cows with acute mastitis. Am. J. Vet Res., 45: 1771-1774. Todhunter, D., Smith, K.L. and Hogan, J.S., 1990. Growth of Gram-negative bacteria in dry cow secretion. J. Dairy Sci., 73: 363-372. Wray, C., Callow, R.J. and Sojka, W.J., 1983. An examination of Escherichia coli strains isolated from cases of bovine mastiffs for possible virulence determinants. Vet Micro., 8: 141-145.