PHYSIOLOGY AND MANAGEMENT
Growth Responses of Coliform Bacteria
to Recombinant Bovine Cytoklnes'
J. S. HOGAN, D. A. TODHUNTER, K. L. SMITH, and P. S. SCHOENBERGER
Ohio Agricultural Research and Development Center
ABSTRACT
Growth responses of 10 coliform iso-
lates to recombinant bovine cytokines
were measured in vitro. Six Escherichia
coli and four Klebsiella pneumoniae iso-
lates obtained from bovine MI were
tested for growth responses to recom-
binant bovine interleukin-10, interleukin-
2. and interferon-y. Cytokines were
tested at IO4, lo3, IO2, and 10 U/ml of
media. Media used were a synthetic tis-
sue culture medium, a chemically de-
fined synthetic bacterial growth medium,
and UHT sterilized milk. Bacterial
counts in the synthetic tissue culture
medium and UHT milk increased
slightly as concentration of interferon-y
in the media increased. Recombinant bo-
vine interferon-y increased bacterial
populations during the log growth phase
but did not affect the number of bacteria
in stationary growth phase. Bacterial
growth responses were not related to ei-
ther interleukin-2 or interleukin-10 con-
centrations in any of the three media.
Bacterial growth responses to cytokines
were not related to differences in either
serum susceptibility, growth of isolates
in dry cow secretion, duration of IMI
from which isolates were obtained, or
bacterial species.
Received September 28, 1992.
Accepted November 24, 1992.
'Salaries and research support were provided by state macrophage colony-stimulating factor (4, 8). A
and federal funds appropriated to the Ohio Agricultural
Research and Development Center, The Oh10 State conclusion drawn from both studies (4, 8) was
University. Manuscript Number 278-92.
1993 J Dairy Sci 76:978-982
Department of Dairy Science
The Ohio State University
Wooster 44691
L. M. SORDILLO
Department of Veterinary Science
The Pennsylvania State University
University Park 16802
(Key words: Escherichia coli,cytokines,
growth responses)
Abbreviation key: BSB = Bacto Synthetic
Broth@, IF"-y = interferon-y, IL-16 =
interleukin-I& IL-2 = interleukin-2, rb =
recombinant bovine.
INTRODUCTION
Cytokines are glycoproteins that modulate
activity of cells participating in immune and
inflammatory processes. Therapeutic use of
recombinant bovine (rb) cytokmes appears to
be a promising procedure to enhance non-
specific host defenses against bacterial infec-
tions. Bovine lymphocyte blastogenesis and
neutrophil bactericidal activity were enhanced
by exogenous rb interleukin-1/3 (IL-l@), rb
interleukin-2 (IL-2), and rb interferon-y
(IFN-y) treatments (1, 3). Intramammary infu-
sion of rbIFN-y also enhanced resistance
against experimentally induced coliform masti-
tis. Mammary glands infused with rbIFN-y
had fewer quarters infected, reduced severity
of inflammation, and IMI of shorter duration
than placebo-infused glands following ex-
perimental challenge with Escherichia coli
(10).
Studies determining growth and metabolism
of microbial pathogens in the presence of
cytokines are limited, but data suggest that
recombinant human cytokines may have direct
effects on human pathogens. Human cytokines
that enhanced growth of virulent E. coli in-
cluded IL-2, IL-10, and granulocyte-
that cytokines may act as growth stimulators
978
 COLLFORM GROWTH RESPONSES TO CYTOKINES
for some virulent bacteria. The purpose of this
experiment was to determine in vitro growth
responses of coliform bacteria isolated from
naturally occurring bovine IMI to rb cytokines.
MATERIALS AND METHODS
Cytokines
The rbIL-10 (American Cyanamid, Prince-
ton, NJ), rbIL-2 (American Cyanamid), and
rbIFN-y (Ciba-Geigy Ltd., Basel, Switzerland)
were provided by the manufacturers in sterile
pyrogen-free water. Both rbIL- 10 and rbIL-2
were stored according to manufacturers
specifications at 4"C, and rbIFN-y was frozen
at -2O'C prior to testing. Bioactivity of each
cytokine was reported by the manufacturers as
rbIL-10, 6000 U/pg; rbIL-2, 1250 U/pg; and
rbIFN-y, 5100 Ulpg.
Bacteria
Bacteria tested were six E. coli and four
Klebsiella pneumniae isolated from naturally
occumng cases of clinical mastitis. Bacteria
were identified by Gram-stain reaction, cellular
morphology, ability to ferment lactose, citrate
utilization, motility, and a miniaturized bio-
chemical test system (api-20E@; Analytab
Products, Plainsview, NY). Three E. coli and
two K. pneumoniae strains were susceptible to
the bactericidal activity of bovine serum ( 5 ) .
Isolates were paired within species as either
poorly adapted (<lo2 cfu logldml) or highly
adapted (>lo7 cfu logldml) for growth in dry
cow mammary secretion (12) and by duration
of IMI from which isolates were obtained.
Duration of IMI was the time in days between
the first and final isolation of bacteria. Chronic
IMI were those with durations of greater than
30 d (geometric mean, 78 d; range, 33 to 212
d). Acute IMI were those that resulted in a
single isolation of bacteria from clinically in-
fected quarters.
Bacteria were stored in trypticase soy broth
(BBL Microbiology Systems, Cockeysville,
MD) containing 20% glycerin at -70°C. A
total of .1 ml of thawed stock culture was
inoculated into 12 ml of trypticase soy broth
and incubated overnight at 37°C on a gyratory
shaker at 200 rpm. A .2-ml portion of the
overnight culture was inoculated into 24 ml of
979
fresh trypticase soy broth and incubated 2.5 h
at 37°C and 50 rpm. Bacteria were centrifuged
6500 x g for 30 min at 4°C and resuspended in
PBS (pH 7.1). Bacterial cultures were diluted
in PBS to 75% transmission at 540 nm (Beck-
man DU-50 Spectrophotometer; Beckman In-
struments, Fullerton, CA). Bacterial numbers
were confirmed by removing a portion of as-
say suspension prior to incubation, serially
diluting bacteria in PBS, and plating bacteria
on trypticase soy agar for 24 h at 37°C.
Microassay of Bacterial Growth
A single 96-well U-bottomed tissue culture
plate (Sarstedt Inc., Newton, NC) facilitated
duplicate assays for effects of one cytokine at
four concentrations on growth of four isolates
in two growth media. Each plate well had a
capacity of 300 pl: 200 pl were allotted to the
medium, 25 p1 to bacterial inoculum, and 25 pl
to sterile cytokine. Bacterial growth media
were RPMI-I640 with L-glutamine and with-
out sodium bicarbonate (Sigma Chemical Co.,
St. Louis, MO), Bacto Synthetic Broths (BSB;
Difco Laboratories, Detroit, MI) supplemented
with 10 g/L of dextrose, and UHT sterilized
milk (Farm Best; Dairyman, Inc., Savannah,
GA). Both RPMI-1640 and BSB were adjusted
to pH 7.1 and sterilized by filtration through a
.45-pm pore size membrane. Cytokines were
diluted in sterile PBS and added to the assay at
final concentrations of 104, lo3, lo2, and 10 U/
ml of media. Control cultures were 25 pl of
bacterial inoculum, 25 pl of PBS, and 200 p1
of bacterial growth medium. Viable bacteria
were determined by the procedures outlined by
Nonnecke and Smith (7) after assay plates
were incubated at 37°C for 2 h. Bacteria were
enumerated on trypticase soy agar.
Growth of E. coli 28 in rblFNr
The growth responses of isolate E. coli 28
to rbIFN-y during log and stationary growth
phases were investigated. A 2.5-h culture was
prepared as described for inoculating the bac-
terial growth microassay. Duplicate 5-ml poly-
styrene tubes were inoculated with .1 ml of E.
coli 28 culture into .9 ml of RPMI-1640 con-
taining lo4 U/ml of rbIFN-y. Controls were
lo4 U/ml of rbIFN-y in 1 ml of RPMI- 1640
and 1 ml of RPMI-1640-E. coli culture without
Journal of Dairy Science Vol. 76, No. 4, 1993
980 HOGAN ”*
CI AL’.
rbIFN-y. Growth responses were measured af-
ter 0, 2, 4, 6, and 8 h of incubation at 37’C (7).
After 4 h of incubation, a .4-ml sample from
each tube was removed and centrifuged (2000
i 8.1
t: 1
x g for 1 h at 473, and the supernatant was \
filtered (.45-pm pore size). The rbIFN-y bioac-
tivity of each supernatant was tested by a virus
inhibition assay (1 1) to determine the ability of
E. coli 28 to alter rbIFN-y concentrations in
the medium. All assays were replicated on 2
separate d.

Statistical Analyses
o id io2 io3 io4
rb1FN-y (U/ml UHT)
Distribution of bacterial counts was normal-
ized by log 10 transformation. Linear relation- Figure 2. Mean growth responses of 10 coliform iso-
ships and correlations among bacterial growth lates to recombinant bovine interferon-y (rblFN-y) in
and cytokine concentrations were measured by UHT milk. Bars represent standard error.
Model 1 regression with data blocked by iso-
late (9). Main effects tested by ANOVA were growth were significant, actual differences in
serum susceptibility, growth in dry cow secre- bacterial counts among concentrations of
tion, duration of IMI, and bacterial species (9). rbIFN-y were small. Geometric mean colony-
forming units for isolates exposed to lo4 U/ml
RESULTS AND DISCUSSION of rbIFN-y were 14 x lo7 cfdml compared
with 8 x lo7 cfdml for negative controls in
WMI-1640 media. This increase was similar
rblFNr to the twofold increase in growth of E. coli in
Linear relationships were significant (P < media containing human IL-2 and granulocyte-
.05) among bacterial counts and rbIFN-y con- macrophage colony-stimulating factor (4).
centrations in WMI-1640 (Figure 1) and UHT Bacterial counts in BSB were independent of
milk (Figure 2) when data from all isolates rbIFN-y concentration. Growth responses to
were combined. Although the relationships be- rbIFN-y did not differ between E. coli and K.
tween rbIFN-y concentration and bacterial pneumoniae isolates and were independent of
serum susceptibility, growth in dry cow secre-
tion, and duration of IMI.
Bovine IMI isolates were incubated for 2 h
to determine whether growth responses to rb

I cytokines were comparable with growth

responses of human derived E. coli to human
cytokines after 2 h of incubation (4, 8).
Growth of isolate E. coli 28 in rbIFN-y was
measured during both log and stationary
growth phases to provide information on the
influence of rbIFN-y on growth dynamics of
coliform populations. Escherichia coli 28 was
selected for additional testing because growth
of this isolate was positively correlated with
rbIFN-y concentrations in each medium.
o id io2 io3 io4 Coefficients of determination among E. coli 28
rbIFN-7 (U/ml of RPMI 1640) bacterial counts and rbIFN-y concentrations
were .76, SO, and .94 in WMI-1640, BSB,
Figure 1. Mean growth responses of 10 coliform iso-
lates to recombinant bovine interferon-? (rbIFN-y) in a and UHT milk, respectively. The rbIFN-y in-
synthetic tissue culture medium (RPMI-1640). Bars repre- creased (P c .05) bacterial populations after 2
Sent standard error. and 4 h of incubation but did not affect the
Journal of Dairy Science Vol. 76, No. 4, 1993
 COLIFORM GROWTH RESPONSES TO CYTOKINES 98 1
that result in impaired immune responses and
r increased bacterial growth. However, E. coli
28 did not decrease concentrations of rbIFN-y
when it was cultured in RPMI-1640 (data not
shown). The responses of coliforms to rbIFN-y
varied among growth media, indicating a pos-
sible nutritional role of the cytokine. The
cytokine or excipient carrier may have con-
tained substrates or cofactors that optimized
bacterial growth. In support of this theory, the
greatest increases in growth were measured in
RPMI- 1640, a chemically defined minimal
4 0 2 4 6 8 medium. The growth-promoting effects of
Hours rbIFN-y were the least in BSB, a complex
medium used for the rapid growth of coliform
Figure 3. Growth curves of Escherichia coli 28 in bacteria (12). Enhancement of growth by other
synthetic tissue culture medium (RPMI-1640) with ( 0 ) and cytokines was independent of growth medium
without (V) 104 Ulml of interferon-y. Bars represent
standard error. (4, 8).

rblL-2 and rblL-16

number of bacteria after 6 and 8 h of incuba-
tion (Figure 3). Cultures incubated in the pres-
ence of rbIFN-y reached stationary phase after
4 h of incubation compared with 6 h of incuba-
tion in control cultures. Porat et al. (8) also
reported that human IL-10 enhanced growth of
E. coli isolates during log growth phase but
did not alter the number of bacteria in station-
ary phase.
Previous studies (4) have suggested that
bacteria reduced the concentration of cytokines
in media after only a few hours of incubation.
Consumption of cytokines by pathogens could
have several implications for the host, includ-
ing decreased local concentrations of cytokine
Bacterial growth responses were not related
to either rbIL-2 or rbIL-10 concentrations in
any of the three media. Combined growth
responses to rbIL-2 or rbIL-lP in all three
media are illustrated in Figures 4 and 5 ,
respectively. These results differ from previous
findings, indicating that human cytokines
stimulated the growth of virulent E. coli. Denis
et al. (4) reported that recombinant human IL-2
and recombinant human granulocyte-
macrophage colony-stimulating factor en-
hanced the growth of an E. coli strain in
RPMI-1640. The enhanced growth by human
IL-2 WilI jignificant at concintrationi as low

8.1
1 T
3

E
-E 8.1

l-
4
\
2 8.0
0
3

M
0
M

7.9
o io1 io2 lo3 lo4 o id 102 lo3 io4
rbIL- 2 (U/ml) rbIL-lp (U/ml)
Figure 4. Mean growth responses of 10 coliform iso- Figure 5. Mean growth responses of 10 coliform iso-
lates to recombinant bovine interleukin-2 (rbIL-2) in three lates to recombinant bovine interleukin-lp (rbIL-Ip) in
media. Bars represent standard error. three media. Bars represent standard error.

Journal of Dairy Science Vol. 76, No. 4, 1993

982 HOGAN ET AL
as 1 U/ml. Porat et al. (8) reported that as little
as I O ng/ml of human IL-1/3 enhanced growth
of E. coli strains. The disparity between the
current trial and previous reports may be due
to differences in cytokines tested. For example,
bovine IL-10 is only 62% homologous to the
human IL-lp protein (6). Porat et al. (8) con-
cluded that human IL-lp recognized an IL-
1-like receptor on virulent E. coli to enhance
bacterial growth. Bacterial receptors for human
cytokines may not recognize bovine cytokines
because of the heterogeneity between the pro-
teins.
The lack of response by isolates to rbIL-2
and rbIL-10 in the present study compared
with earlier reports may have been due to
heterogeneity among isolates tested. Growth
responses of E. coli strains to human IL-2 and
IL-I@ were dependent on virulence of strains
(4, 8). Virulent bacteria were defined as serum-
resistant isolates in both studies (4, 8). Growth
responses of bovine mastitis pathogens to
cytokines were not related to serum suscepti-
bility, growth of isolates in dry cow secretion,
duration of IMI, or bacterial species in the
present study. Many traits that are considered
to be virulence factors for coliform bacteria in
human diseases are not related to the patho-
genesis of bovine mastitis (2).
CONCLUSIONS
Little is known concerning the effects of
bovine cytokines on microbial growth and me-
tabolism. Escherichia coli, Mycobacterium
species, and pathogenic protozoan enhanced
growth responses to human cytokines (4, 8). A
significant but small response to rbIFW-y by E.
coli and K. pneumoniae was measured in vitro
in the current study. Denis et al. (4) proposed
that growth stimulation of bacterial pathogens
by cytokines and down-regulation of local im-
mune responses may be responsible for an
unexpectedly high incidence of bacterial infec-
tions in humans treated with exogenous
cytokines. A similar increase in bacterial infec-
tions in cows treated with exogenous cytokines
Journal of Dairy Science Vol. 76, No. 4, 1993
has not been reported. However, monitoring
naturally occurring infections and establishing
growth responses of bacteria to cytokines used
in experimental infections may be warranted in
future studies.
REFERENCES
1 Babiuk. L. A,, L. M. Sordillo, M. Campos, H.P.A.
Hughes, A. Rossi-Campos, and R. Harland. 1991.
Application of interferons in the control of infectious
diseases of cattle. J. Dairy Sci. 74:4385.
2 B m o w . P. A., and A. W. Hill. 1989. The virulence
characteristics of strains of Escherichia coli isolated
from cases of bovine mastitis in England and Wales.
Vet. Microbiol. 20:35.
3 Daley. M. J., P. A. Coyle. T. J. Williams, G. Furda, R.
Dougherty, and P. W. Hayes. 1991. Staphylococcus
aureus mastitis: pathogenesis and treatment with bo-
vine interleukin-la and interleukin-2. J. Dairy Sci. 74:
4413.
4Denis. M., D. Campbell, and E. 0. Gregg. 1991.
Interleukin-2 and granulocyte colony-stimulating fac-
tor stimulate growth of a virulent strain of Escherichia
coli. Infect. Immun. 59:1853.
5 Hogan, J. S., D. A. Todhunter, K. L. Smith, and P. S .
Schoenberger. 1989. Serum susceptibility of coliforms
isolated from bovine intramammary infections. 1.
Dairy Sci. 72:1893.
6Maliszewski. G. R., P. E. Baker, M. A. Schoenborn,
B. S. Davis, D. Cosman, S . Gillis, and D. P. Cemtti.
1988. Cloning, sequence, and expression of bovine
interleukin 1a and interleukin 10 complimentary
DNAs. Mol. Immunol. 25429.
7 Nonnecke, B. J., and K. L. Smith. 1984. Inhibition of
mastitis bacteria by bovine apo-lactofemn evaluated
by in vitro microassay of bacterial growth. J. Dairy
Sci. 67:606.
8 Porat, R., B. D. Clark, S. M. Wolff, and C. A.
Dinarello. 1991. Enhancement of growth of virulent
strains of Escherichia coli by interleukin-I. Science
(Washington, DC) 254:430.
9 S A P User’s Guide: Statistics, Version 5 Edition.
1985. SAS Inst., Inc., Cary, NC.
10Sordillo. L. M.,and L. A. Babiuk. 1991. Controlling
acute Escherichia coli mastitis during the peripar-
turient period with recombinant bovine interferon
gamma. Vet. Microbiol. 28:189.
1 I Sordillo, L. M.,M. T. Redmond, M. Campos, L. War-
ren, and L. A. Babiuk. 1991. Cytokine activity in
bovine mammary gland secretions during the peripar-
turient period. Can. J. Vet. Res. 55:298.
12Todhunter. D. A., K. L. Smith, and J. S. Hogan. 1990.
Growth of Gram-negative bacteria in dry cow secre-
tion. J. Dairy Sci. 73:363.