Veterinary Immunology and Immunopathology 156 (2013) 200–204

Contents lists available at ScienceDirect

Veterinary Immunology and Immunopathology

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

Inhibitory effect of interferon gamma on frequency

of Ehrlichia canis-infected cells in vitro
Tomoko Tajima a,∗ , Makoto Wada a,b
a
Department of Veterinary Sciences, Graduate School of Life and Environmental Sciences, Osaka Prefecture University, 1-58
Rinku-orai-kita, Izumisano, Osaka 598-8531, Japan
b
Tani Animal Hospital, 240-8 Hirai, Sakai, Osaka 599-8251, Japan

a r t i c l e i n f o a b s t r a c t

Article history: Ehrlichia canis is an obligate intracellular bacterium that infects the macrophage–monocyte
Received 27 December 2012 cells of dogs, causing canine monocytic ehrlichiosis. Interferon-␥ (IFN-␥), along with other
Received in revised form 2 September 2013
cytokines, mediates the immune response to such intracellular bacterial invasions. To deter-
Accepted 23 September 2013
mine the role of IFN-␥ in the immunity of dogs to E. canis infection, peripheral blood
mononuclear cells (PBMC) and white blood cells (WBC) were collected from E. canis-infected
Keywords:
dogs and added to a culture of E. canis in DH82 cells. The number of E. canis inclusion-positive
Cellular resistance
Dog cells was significantly reduced in cultures containing PBMC and WBC from E. canis-infected
Ehrlichia canis dogs compared to uninfected dogs. However, this resistance was inhibited by the addition
Interferon gamma of an anti-dog IFN-␥ antibody. Resistance was also observed when PBMC were added to
the Cell Culture Inserts, which prohibited contact of PBMC to DH82 cells, while allowed
the diffusion of soluble cell products. The results of this study indicate that resistance was
not dependent on cell to cell contact, but was associated with soluble cell products, such
as IFN-␥. The addition of recombinant canine IFN-␥ to the E. canis culture also reduced the
number of infected cells. A commercial recombinant canine IFN-␥, which is sold in Japan,
was also effective at reducing E. canis-infected cell number. These results indicate that IFN-
␥ has an inhibitory effect on the frequency of E. canis-infected cells in vitro and that contact
between effector and target cells is not necessary for the resistance.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction is inhibited by adding human IFN-␥ during human mono-

Ehrlichia canis is an obligate intracellular bacterium that
infects the macrophage–monocyte cells of dogs, causing
canine monocytic ehrlichiosis (Little, 2010). Phagocytes
play important roles in the immune response against
intracellular bacteria, with interactions being mediated by
various soluble cell products, such as interferon-␥ (IFN-
␥) (Abbas et al., 2012). Several reports have discussed the
role of IFN-␥ in ehrlichial infections. Previously, Barnewall
and Rikihisa (1994) reported that E. chaffeensis infection
∗ Corresponding author. Tel.: +81 72 463 5698; fax: +81 72 463 5093.
E-mail address: tajima@vet.osakafu-u.ac.jp (T. Tajima).
cyte cultivation. In another report, Bitsaktsis et al. (2004)
suggested that CD4 cells and IFN-␥ are responsible for
macrophage activation, and for the elimination of Ixodes
ovatus ehrlichia in mice.
We previously reported that the expression of IFN-␥
mRNA was detected in the peripheral blood mononuclear
cells (PBMC) of E. canis-infected dogs within 3 days of infec-
tion, and that the expression of mRNA continued for more
than 50 days (Tajima and Rikihisa, 2005). The expression of
IFN-␥ mRNA in the leukocytes of E. canis-infected dogs was
also reported by Faria et al. (2011). In another report, the
development of IFN-␥ was examined in E. canis-infected
dog by using a dog IFN-␥ enzyme-linked immunosorbent
spot (ELISpot) assay (Normand et al., 2009). These data led

0165-2427/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.vetimm.2013.09.014
 T. Tajima, M. Wada / Veterinary Immunology and Immunopathology 156 (2013) 200–204
us to hypothesize that canine IFN-␥ mediates resistance to
E. canis infection at the cellular level. To test the hypoth-
esis, we examined the inhibitory effect of IFN-␥ on the
frequency of E. canis-infected cells in vitro.
2. Materials and methods
2.1. Animals and design
E. canis Oklahoma strain was kindly supplied by Dr. Y.
Rikihisa, The Ohio State University, Columbus, OH, USA.
The agent was cultured in DH82 cells, which is a dog
macrophage–monocyte cell line established from canine
malignant histiocytosis (Wellman et al., 1988), using Dul-
becco’s modified Eagle’s MEM (D-MEM), containing 5%
fetal calf serum.
Two female beagles (of 2–3 years in age) were intra-
venously inoculated with 5 × 106 E canis-infected DH82
cells (of which 93% of the cells were infected) in 5 ml
of D-MEM. Before inoculation, both dogs tested nega-
tive for E. canis using indirect fluorescent-antibody (IFA)
testing and Polymerase Chain Reaction (PCR), as previ-
ously described (Wen et al., 1997). Although slight fever
(39–40 ◦ C, 8–10 days post-inoculation [DPI]) was observed
in both dogs, they had good appetites, with no clinical
abnormalities being observed during the experiment. E.
canis infection was confirmed by PCR using DNA extracted
from the buffy coat cells obtained from the peripheral blood
of each dog at 14 DPI. At 16 DPI, blood was drawn from
both dogs. Two additional female beagles (of 4–6 years of
age) were used as uninfected controls. The animal exper-
imentation protocols of this study were approved by the
Osaka Prefecture University’s Animal Research Commit-
tee.
2.2. Preparation of white blood cells (WBC) and
peripheral blood mononuclear cells (PBMC)
To examine the effect of canine peripheral blood cells
on E. canis infection in vitro, WBC and PBMC were pre-
pared from the blood of E. canis-infected and uninfected
dogs. Whole blood was centrifuged at 500 × g for 15 min at
room temperature. Buffy coat was collected, and red blood
cells were lysed with 0.83% NH3 Cl. After washing 3 times
with phosphate-buffered saline (PBS) at 150 × g for 5 min,
the cells were used as WBC. For the preparation of PBMC,
the buffy coat of the blood of each dog was suspended
in 10 ml of PBS, and placed on 10 ml of Histopaque-1077
(Sigma–Aldrich Co., St. Louis, MO, USA). The tubes were
centrifuged at 400 × g for 30 min at room temperature. The
cells at the interface were collected and washed 3 times
with PBS at 150 × g for 5 min, and used as PBMC. Peripheral
blood mononuclear cells contained mostly lymphocytes
and macrophages, while WBC contained PBMC plus gra-
nulocytes.
2.3. Effect of WBC and PBMC on the frequency of E.
canis-positive cells
DH82 cells (5 × 104 cells/well) were infected with E.
canis Oklahoma strain (102.5 TCID50 /well), and cultured on
201
a 24-well culture plate with 5 × 105 WBC or PBMC cells, and
were placed in a total volume of a 1.0 ml/well. On another
plate, 1 ␮g/well of anti-canine IFN-␥ antibody, which was
produced by goats (R & D Systems, Inc., Minneapolis, MN,
USA), was added to the culture. The cells were collected
from 4 separate wells on days 2, 3, and 4 of culture, and
Cytospin smears were prepared as described previously
(Barnewall and Rikihisa, 1994). The smears were stained
with Diff-Quik (Sysmex Co., Kobe, Japan), and examined for
the presence of inclusions. At least 120 cells per well were
examined, with the mean and standard deviation of the
percentage of inclusion-positive cells in the 4 wells being
calculated.
Cell Culture Inserts (BD Falcon, Franklin Lakes, NJ, USA)
were placed into the wells of a 24-well culture plate con-
taining E. canis-inoculated DH82 cells. PBMC from infected
or uninfected dogs, 1 × 105 cells was added to each Cell Cul-
ture Insert, and cultured for 3 days. Smears of each DH82
culture were prepared, stained, and examined for the pres-
ence of inclusions.
2.4. Effect of recombinant canine IFN- (r-IFN-) and
commercial recombinant canine IFN- (com-r-IFN-) on
the frequency of E. canis-positive cells
To examine the effect of r-IFN-␥ (R & D Systems, Inc.),
r-IFN-␥ was added to E. canis-inoculated DH82 cells at a
concentration of 0.625, 1.25, 2.5, and 5.0 ng/ml. A com-r-
IFN-␥, which is named INTERDOGTM (Kyoritsu Seiyaku Co.,
Tokyo, Japan), has been manufactured and sold in Japan
since 2005 to treat atopic dermatitis (eczema) in dogs.
As r-IFN-␥ (R & D Systems, Inc.) is only used for research
purposes, com-r-IFN-␥ was used to examine the utility of
r-IFN- for clinical application. Commercial recombinant
canine IFN-␥ was added to the culture at a concentration
of 0.05, 0.5, 5, and 50 ng/ml. The experiments were carried
out twice for each recombinant canine IFN-␥.
2.5. Statistical analysis
The statistical significance of differences between the 2
groups was determined using SigmaPlot 11.0 (Systat Soft-
ware, Inc., San Jose, CA, USA). When normality tests and
equal variance tests passed, the 2 data were compared with
the t-test. If the normality tests or equal variance tests
failed, the data were compared with the Mann–Whitney
rank sum test.
3. Results
3.1. Effect of WBC and PBMC of E. canis-infected dogs
White blood cells from the 2 infected dogs had a similar
effect on the frequency of E. canis-infected cells in vitro,
as shown in Table 1. There was a significantly reduced
incidence of inclusion-positive cells in the culture con-
taining the WBC of E. canis-infected dogs; however, the
incidence remained high in the culture containing those of
uninfected dogs. The reduction in inclusion-positive cells
was inhibited by adding anti-canine IFN-␥ antibody to
202 T. Tajima, M. Wada / Veterinary Immunology and Immunopathology 156 (2013) 200–204

Table 1
Effect of white blood cells (WBC) collected from E. canis-infected and uninfected dogs on the frequency of E. canis-positive cells. A dog macrophage–monocyte
cell line, DH82 cells, were infected with E. canis and cultured with WBC. The cells were collected from 4 separate wells on days 2, 3 and 4 of culture. Smears
of each culture were prepared, stained, and examined for the presence of E-canis inclusions.

DPIb Frequency of E. canis-positive cellsa

Dog code WBC of E. canis-infected dogs (n = 2) Dog code WBC of uninfected dogs (n = 2)

E. canis E. canis + WBC E. canis + WBC + Abc E. canis E. canis + WBC E. canis + WBC + Ab

2 I-1 16.70 ± 4.26 4.71 ± 0.49 14.88 ± 4.87 N-1 10.24 ± 0.91 13.00 ± 2.81 13.95 ± 1.16
I-2 9.85 ± 0.70 4.70 ± 2.55 8.55 ± 1.44 N-2 14.58 ± 6.13 13.61 ± 1.93 14.97 ± 2.20
Mean 13.27 ± 4.32 4.71 ± 1.59d 11.72 ± 4.43 Mean 12.41 ± 4.88 13.31 ± 2.11d 14.46 ± 1.61

3 I-1 18.94 ± 2.56 3.59 ± 1.43 16.29 ± 4.07 N-1 14.10 ± 1.32 14.83 ± 1.27 12.86 ± 3.03
I-2 12.37 ± 2.25 5.28 ± 0.94 10.51 ± 2.54 N-2 17.99 ± 1.37 16.41 ± 2.93 16.54 ± 1.61
Mean 15.65 ± 3.89 4.44 ± 1.34e 13.40 ± 4.12 Mean 16.05 ± 2.27 15.62 ± 2.11e 14.70 ± 2.80

4 I-1 34.36 ± 2.80 14.37 ± 3.68 30.74 ± 3.50 N-1 25.43 ± 3.90 29.27 ± 6.25 28.35 ± 8.79
I-2 24.17 ± 2.89 12.60 ± 1.76 21.79 ± 3.04 N-2 32.49 ± 1.66 28.59 ± 4.11 32.81 ± 2.48
Mean 29.26 ± 5.66 13.48 ± 2.65f 26.27 ± 5.30 Mean 28.96 ± 4.38 28.93 ± 4.59f 30.58 ± 6.02
a
Percentage of E. canis-positive DH82 cells; mean ± standard deviation of the 4 cultures for each dog, and 8 cultures (total of 2 dogs) for Mean.
b
Days post-inoculation.
c
Anti-dog interferon gamma antibody.
d
There was a statistically significant difference between the infected and uninfected groups (p ≤ 0.001) when using the t-test.
e
There was a statistically significant difference between the infected and uninfected groups (p ≤ 0.001) when using the t-test.
f
There was a statistically significant difference between the infected and uninfected groups (p ≤ 0.001) when using the t-test.

the culture, which indicated that IFN-␥ contributes to the
reduction of E. canis-positive cells. However, this inhibi-
tion effect was limited to 4 days. At 5 and 6 days after
cultivation, the frequency of inclusion-positive cells was
more than 30% in all cultures. On the basis of these results,
all subsequent cultures were conducted for a maximum of
4 days.
The number of E-canis inclusion-positive cells was also
significantly reduced in cultures containing PBMC from
the 2 E. canis-infected dogs compared to the 2 uninfected
dogs (Table 2). Differences were not observed between
WBC and PBMC cultures, indicating that granulocytes
are not essential for inhibiting frequency of E. canis-
positive cells. Hence, PBMC were used in all subsequent
experiments.
3.2. Effect of PBMC in the Cell Culture Inserts
Peripheral blood mononuclear cells from E. canis-
infected dogs inhibited the frequency of E. canis-positive
cells on the culture plate. The Cell Culture Inserts pre-
vented PBMC from forming contact with DH82 cells, while
allowing the diffusion of soluble cell products. The present
results indicate that inhibition is not caused by cell–cell
contact, which is the mechanism of cytotoxic T-cells, but is
caused by soluble cell products, such as IFN- (Table 3).
3.3. Effect of r-IFN-
A significant reduction in the incidence of inclusion-
positive cells was observed in E. canis-infected DH82 cells

Table 2
Effect of peripheral blood mononuclear cells (PBMC) of E. canis-infected and uninfected dogs on the frequency of E. canis-positive cells. A dog macrophage-
monocyte cell line, DH82 cells, were infected with E. canis and cultured with PBMC. The cells were collected from 4 separate wells on days 2, 3 and 4 of
culture. Smears of each culture were prepared, stained, and examined for the presence of E-canis inclusions.

DPIb Frequency of E. canis-positive cellsa

Dog code PBMC of E. canis-infected dogs (n = 2) Dog code PBMC of uninfected dogs (n = 2)
c
E. canis E. canis + PBMC E. canis + PBMC + Ab E. canis E. canis + PBMC E. canis + PBMC + Ab

2 I-1 16.38 ± 3.03 4.66 ± 2.43 14.36 ± 3.22 N-1 16.97 ± 4.86 11.52 ± 2.23 11.05 ± 2.34
I-2 20.89 ± 3.17 5.13 ± 2.73 18.77 ± 2.28 N-2 13.37 ± 3.81 13.79 ± 0.22 14.88 ± 0.80
Mean 18.64 ± 3.51 4.90 ± 2.25d 16.56 ± 3.27 Mean 15.17 ± 4.19 12.65 ± 1.78d 12.79 ± 2.44

3 I-1 19.27 ± 1.36 7.19 ± 1.07 14.79 ± 2.84 N-1 16.76 ± 5.34 11.83 ± 1.14 12.83 ± 1.28
I-2 23.58 ± 2.92 7.13 ± 0.92 19.22 ± 1.29 N-2 18.26 ± 2.40 19.34 ± 1.45 18.79 ± 0.41
Mean 21.42 ± 2.92 7.16 ± 0.86e 17.01 ± 2.93 Mean 17.51 ± 3.67 15.58 ± 3.92e 15.81 ± 3.09

4 I-1 20.94 ± 2.59 12.60 ± 1.34 17.81 ± 1.30 N-1 23.96 ± 5.31 23.78 ± 5.80 19.52 ± 8.99
I-2 27.49 ± 4.00 12.85 ± 3.07 25.87 ± 3.95 N-2 29.32 ± 1.75 31.37 ± 0.92 29.57 ± 2.12
Mean 24.21 ± 4.38 12.73 ± 2.06f 21.84 ± 4.77 Mean 26.64 ± 4.35 27.57 ± 5.23f 24.55 ± 7.56
a
Percentage of E. canis-positive DH82 cells; mean ± standard deviation of the 4 cultures for each dog, and 8 cultures (total of 2 dogs) for Mean.
b
Days post-inoculation.
c
Anti-dog interferon gamma antibody.
d
There was a statistically significant difference between the infected and uninfected groups (p ≤ 0.001) using the t-test.
e
There was a statistically significant difference between the infected and uninfected groups (p ≤ 0.001) using the Mann–Whitney rank sum test.
f
There was a statistically significant difference between the infected and uninfected groups (p ≤ 0.001) using the Mann–Whitney rank sum test.
 T. Tajima, M. Wada / Veterinary Immunology and Immunopathology 156 (2013) 200–204 203

Table 3 Table 5
Effect of peripheral blood mononuclear cells (PBMC) of E. canis-infected Effect of commercial canine recombinant IFN-␥ (com-r-IFN-␥) on the
and uninfected dogs in the Cell Culture Inserts on the frequency of E. canis- frequency of E. canis-positive cells. Each concentration of com-r-
positive cells. Cell Culture Inserts were placed into the wells of a 24-well IFN-␥ was added to E. canis-inoculated DH82 cells, which is a dog
culture plate containing E. canis-inoculated DH82 cells, which is a dog macrophage–monocyte cell line, and cultured for 3 days. Smears of each
macrophage-monocyte cell line. Peripheral blood mononuclear cells were culture were prepared, stained, and examined for the presence of E-canis
added to each Cell Culture Insert and cultured for 3 days. Smears of each inclusions.
culture were prepared, stained, and examined for the presence of E-canis
inclusions. Concentration Experiment Frequency of E. p value
of com-r-IFN-␥ canis-positive
Origin of PBMC Frequency of E. canis-positive cellsa (ng/ml) cellsa

None (DH82 + E. canis) 40.71 ± 1.50 0 (Control) 1 31.52 ± 5.74

2 29.95 ± 3.34
E. canis-infected dog I-1 8.41 ± 2.11
Mean 30.73 ± 4.43
E. canis-infected dog I-2 14.41 ± 3.20
Mean 11.26 ± 3.99b 0.05 1 16.56 ± 3.37
2 29.27 ± 2.59
Uninfected dog N-1 36.13 ± 4.96
Mean 22.92 ± 7.34 p = 0.065b
Uninfected dog N-2 34.35 ± 5.89
Mean 35.31 ± 5.16b 0.5 1 16.46 ± 8.46
2 30.75 ± 2.33
a
Percentage of E. canis-positive DH82 cells; mean ± standard deviation
Mean 23.61 ± 9.56 p = 0.076c
of the 4 cultures for each dog, and 8 cultures (total of 2 dogs) for Mean.
b
There was a statistically significant difference between the infected 5.0 1 18.53 ± 4.80
and uninfected groups (p ≤ 0.001) using the t-test. 2 21.52 ± 0.87
Mean 20.02 ± 3.57 p < 0.001c

50.0 1 5.04 ± 1.59

cultured with r-IFN-␥ at a concentration of 5.0 ng/ml
2 5.13 ± 0.93
(Table 4). This reduction in inclusion-positive cells was Mean 5.08 ± 1.21 p < 0.001b
inhibited by the addition of anti-dog IFN-␥ antibody (data a
Percentage of E. canis-positive DH82 cells; mean ± standard deviation
not shown). of the 4 cultures for each experiment, and 8 cultures (total of 2 experi-
ments) for Mean.
b
3.4. Effect of com-r-IFN- Each mean value was compared to the mean value of the control using
the Mann–Whitney rank sum test.
c
Each mean value was compared to the mean value of the control using
A significant reduction in the incidence of inclusion- the t-test.
positive cells was observed in E. canis-infected DH82 cells
cultured with com-r-IFN-␥ at concentrations higher than
5.0 ng/ml (Table 5). 4. Discussion

The results of the present study confirmed that IFN-␥

Table 4
reduced the frequency of E. canis-infected cells in vitro,
Effect of recombinant canine interferon-␥ (r-IFN-␥) on the frequency of E.
canis-positive cells. Each concentration of r-IFN-␥ was added to E. canis- indicating that IFN-␥ mediates resistance at the cellular
inoculated DH82 cells, which is a dog macrophage-monocyte cell line, and level in E. canis. Previous research demonstrated that the
cultured for 3 days. Smears of each culture were prepared, stained, and growth of E. chaffeensis, which is the pathogen of human
examined for the presence of E-canis inclusions. monocytotropic ehrlichiosis, was inhibited by adding IFN-␥
Concentration Experiment Frequency of E. p Valueb in culture (Barnewall and Rikihisa, 1994). This inhibi-
of r-IFN-␥ canis-positive tion was caused by the down-regulation of transferrin
(ng/ml) cellsa receptors. In another study, E. risticii was killed by an l-
0 (Control) 1 35.35 ± 5.40 arginine-dependent mechanism in IFN-␥-activated mouse
2 29.40 ± 6.93 peritoneal macrophages (Park and Rikihisa, 1992). In com-
Mean 32.38 ± 6.62
parison, Harrus et al. (2003) reported that MHC class II
0.625 1 30.63 ± 9.12 expression was evident in uninfected DH82 cells, with no
2 20.05 ± 3.01 expression being evident in DH 82 cells infected with E.
Mean 25.26 ± 7.25 p = 0.083
canis. MHC class II expression-enhancing activity of recom-
1.25 1 32.15 ± 3.85 binant canine IFN-␥ was exhibited (Okano et al., 2000).
2 21.78 ± 1.70 IFN-␥ might enhance MHC class II expression, which in turn
Mean 26.96 ± 7.25 p = 0.122
might inhibit E. canis growth. Further study is required to
2.5 1 26.33 ± 5.28 clarify the mechanism.
2 25.93 ± 4.24
The E. canis Oklahoma strain used in the present study
Mean 26.13 ± 4.41 p = 0.102
is not a virulent strain. This strain is mildly pathogenic
5.0 1 5.18 ± 1.44 to dogs, with infected dogs only presenting slight fever
2 11.28 ± 1.83
in clinical checks. Unver et al. (2006) reported the weak
Mean 8.23 ± 3.43 p < 0.001
expression of IFN-␥ mRNA in dogs inoculated with the
a
Percentage of E. canis-positive DH82 cells; mean ± standard deviation
New Mexico strain of E. canis. The New Mexico strain is
of the 4 cultures for each experiment, and 8 cultures (total of 2 experi-
ments) for Mean. a virulent strain that develops a severe form of ehrlichio-
b
Each mean value was compared to the mean value of the control using sis (Unvera et al., 2009). Differences in the potency among
the Mann–Whitney rank sum test. strains might be an important factor in the effectiveness of
204 T. Tajima, M. Wada / Veterinary Immunology and Immunopathology 156 (2013) 200–204
IFN-␥. As we only had access to the Oklahoma strain, we
were not able to examine differences between the strains.
The reaction of virulent and avirulent strains against IFN-␥
requires examination in the future.
In the present study, we collected blood from infected
dogs during the early stages of E. canis infection. E. canis
causes subclinical infection, in which dogs remain chroni-
cally infected for months to years, but without any outward
evidence of disease (Little, 2010). It is possible that the
inhibitory effect of PBMC of infected dogs is limited to the
early stages of infection. Hence, it is important to clarify the
role of IFN-␥ at the chronic stage of infection.
Our research group previously examined cytokine
mRNA expression in PBMC from E. canis-infected dogs,
and recorded the persistent expression of IFN-␥ and TNF-␣
mRNA (Tajima and Rikihisa, 2005). The expression of TNF-
␣ mRNA in E. canis-infected dogs was also reported by Faria
et al. (2011). The authors suggested that TNF-␣ plays a role
in the pathogenesis of the acute phase of canine ehrlichio-
sis. In contrast, inhibition of the multiplication of Listeria
monocytogenes by IFN-␥ and TNF-␣ has also been reported
(Haponsaph and Czuprynski, 1996). Therefore, it is impor-
tant to clarify the relationship of IFN-␥ and TNF-␣ when
assessing the pathogenesis and immunity of dogs to E. canis
infection.
In vivo studies have also indicated the important role
of IFN-␥ in ehrlichial infections by using knock-out mice
(Feng and Walker, 2004; MacNamara et al., 2011; Bitsaktsis
et al., 2004). In Japan, recombinant canine IFN-␥ is commer-
cially available for the treatment of atopic dermatitis. The
present results using com-r-IFN-␥ suggest the possibility
of using this product in the treatment of E. canis infection,
with confirmatory in vivo experiments being required.
Conflict of interest statement
The authors declare no conflicts of interest.
Acknowledgements
We would like to thank Dr. Yasuko Rikihisa (The Ohio
State University, USA) for kindly providing the DH82 cells
and the E. canis Oklahoma strain.
This study was supported in part by Grants-in-Aid for
Scientific Research from the Japan Society for the Promo-
tion of Science.
References
Abbas, A.K., Lichtman, A.H., Pillai, S., 2012. Immunity to intracellular
bacteria. In: Abbas, A.K., Lichtman, A.H., Pillai, S. (Eds.), Cellular and
Molecular Immunology. , 7th ed. Elsevier Saunders, Philadelphia, USA.
Barnewall, R.E., Rikihisa, Y., 1994. Abrogation of gamma interferon-
induced inhibition of Ehrlichia chaffeensis infection in human
monocytes with iron transferrin. Infect. Immun. 62, 4804–4810.
Bitsaktsis, C., Huntington, J., Winslow, G., 2004. Production of IFN-␥ by
CD4 cells is essential for resolving ehrlichia infection. J. Immunol. 172,
6894–6901.
Faria, J.L.M., Munhoz, T.D., João, C.F., Vargas-Hernández, G., André, M.R.,
Pereira, W.A.B., Machado, R.Z., Tinucci-Costa, M., 2011. Ehrlichia canis
(Jaboticabal strain) induces the expression of TNF-␣ in leukocytes and
splenocytes of experimentally infected dogs. Rev. Bras. Parasitol. Vet.
20, 71–74.
Feng, H.-M., Walker, D.H., 2004. Mechanisms of immunity to Ehrlichia
muris: a model of monocytotropic ehrlichiosis. Infect. Immun. 72,
966–971.
Haponsaph, R., Czuprynski, C.J., 1996. Inhibition of the multiplication of
Listeria monocytogenes in a murine hepatocyte cell line (ATCC TIB73)
by IFN-␥ and TNF-␣. Microb. Pathog. 20, 287–295.
Harrus, S., Waner, T., Fridmann-Morvinski, D., Fishman, Z., Bark, H.,
Harmelin, A., 2003. Down-regulation of MHC class II receptors of
DH82 cells following infection with Ehrlichia canis. Vet. Immunol.
Immunopathol. 96, 239–243.
Little, S.E., 2010. Ehrlichiosis and anaplasmosis in dogs and cats. Vet. Clin.
Small Anim. 40, 1121–1140.
MacNamara, K.C., Oduro, K., Martin, O., Jones, D.D., McLaughlin,
M., Choi, K., Borjesson, D.L., Winslow, G.M., 2011. Infection-
induced myelopoiesis during intracellular bacterial infection is
critically dependent upon IFN-␥ signaling. J. Immunol. 186,
1032–1043.
Normand, T., Forest, L., Chabanne, L., Richard, S., Davoust, B., Juillard, V.,
2009. Gamma IFN Ehrlichia canis-specific cell responses. Clin. Micro-
biol. Infect. 15 (Suppl. 2), 70–71.
Okano, F., Satoh, M., Ido, T., Okamoto, N., Yamada, K., 2000. Production of
canine IFN-␥ in silkworm by recombinant baculovirus and character-
ization of the product. J. Interferon Cytokine Res. 20, 1015–1022.
Park, J., Rikihisa, Y., 1992. l-Arginine-dependent killing of intracellu-
lar Ehrlichia risticii by macrophages treated with gamma interferon.
Infect. Immun. 60, 3504–3508.
Tajima, T., Rikihisa, Y., 2005. Cytokine responses in dogs infected
with Ehrlichia canis Oklahoma strain. Ann. N.Y. Acad. Sci. 1063,
429–432.
Unver, A., Huang, H., Rikihisa, Y., 2006. Cytokine gene expression by
peripheral blood leukocytes in dogs experimentally infected with
a new virulent strain of Ehrlichia canis. Ann. N. Y. Acad. Sci. 1078,
482–486.
Unvera, A., Rikihisa, Y., Karaman, M., Ozen, H., 2009. An acute severe ehrli-
chiosis in a dog experimentally infected with a new virulent strain of
Ehrlichia canis. Clin. Microbiol. Infect. 15 (Suppl. 2), 59–61.
Wellman, M.L., Krakowka, S., Jacobs, R.M., Kociba, G.J., 1988. A
macrophage–monocyte cell line from a dog with malignant histio-
cytosis. In Vitro Cell Dev Biol. 24, 223–229.
Wen, B., Rikihisa, Y., Mott, J.M., Greene, R., Kim, H.-Y., Zhi, N., Couto,
G.C., Unver, A., Bartsch, R., 1997. Comparison of nested PCR with
immunofluorescent-antibody assay for detection of Ehrlichia canis
infection in dogs treated with doxycycline. J. Clin. Microbiol. 35,
1852–1855.