an enzyme-linked immunosorbent assay for antibodies to hepatozoon canis

Veterinary Parasitology 122 (2004) 131–139

An enzyme-linked immunosorbent assay forantibodies toHepatozoon canis

L. Gonena, D. Strauss-Ayalia, V. Shkapb, N. Vincent-Johnsonc,D.K. Macintirec, G. Banetha,∗

a School of Veterinary Medicine, Hebrew University, P.O. Box 12, Rehovot 76100, Israelb Division of Parasitology, Kimron Veterinary Institute, Beit Degan, Israel

c College of Veterinary Medicine, Auburn University, Auburn, AL 36849-5523, USA

Received 13 October 2003; received in revised form 25 February 2004; accepted 8 March 2004

Abstract

Canine hepatozoonosis is a tick-borne protozoal disease caused in the Old World and SouthAmerica byHepatozoon canis. An enzyme-linked immunosorbent assay (ELISA) using purifiedH.canisgamont antigen was applied for the detection of antibodies reactive withH. canis. Evaluationof the ELISA with sera from naturally infected parasitemic dogs indicated that it was sensitive(86%), specific (97%), and comparable to the indirect fluorescent antibody test (IFAT) for thedetection ofH. canisantibodies. A variable degree of serologic cross-reactivity was found betweensera fromH. americanum-infected dogs and theH. canisantigen. Dogs experimentally infectedwith H. canisseroconverted 1–4 weeks post-infection (PI). Antibody levels peaked at 7–9 weeksPI and gradually declined thereafter remaining above the cut-off value until the conclusion of thestudy 7 months PI. The ELISA will be valuable for serological evaluation of dogs suspected ofexposure toH. canisand for epidemiological studies.© 2004 Elsevier B.V. All rights reserved.

Keywords: Hepatozoon canis; Hepatozoon americanum; Rhipicephalus sanguineus; ELISA

1. Introduction

Hepatozoon canisis a tick-borne protozoan parasite, classified in the phylum Apicom-plexa closely related toPlasmodiumspecies and piroplasms (Baneth et al., 2003). Caninehepatozoonosis caused byH. canis is prevalent in southern Europe (Rioux et al., 1964;Hervas et al., 1995; Kontos and Koutinas, 1990), Asia (Rajamanickam et al., 1985; Murata

∗ Corresponding author: Tel.:+972-3-9688557; fax:+972-3-9604079.E-mail address:[email protected] (G. Baneth).

0304-4017/$ – see front matter © 2004 Elsevier B.V. All rights reserved.doi:10.1016/j.vetpar.2004.03.021

132 L. Gonen et al. / Veterinary Parasitology 122 (2004) 131–139

et al., 1993), Africa (McCully et al., 1975; Ezekoli et al., 1983) and South America (O’Dwyeret al., 2001). It is transmitted mostly by the brown dog tickRhipicephalus sanguineus.Trans-mission of the infection to canine hosts takes place by ingestion of ticks, or parts of tickscontaining matureH. canisoocysts (Baneth et al., 2001). H. canisinfection ranges frombeing asymptomatic in dogs with low parasitemia rates to a severe life-threatening illnesswith fever, lethargy, anemia and emaciation in dogs with high parasitemia (Baneth andWeigler, 1997). Dogs in the southern USA are infected by a different species ofHepato-zoon, H. americanum, which induces mostly musculo-skeletal pathology (Vincent-Johnsonet al., 1997) and is transmitted by the Gulf Coast tick,Amblyomma maculatum(Mathewet al., 1998).

The diagnosis of hepatozoonosis is commonly done by microscopic detection of intra-cellularH. canisgamonts in stained blood smears. The gamonts are ellipsoidal in shape andmeasure about 11�m × 4�m. They are found in the cytoplasm of neutrophils and rarelyin monocytes. Meronts ofH. canisare detectable by cytology or histopathology of infectedhemolymphatic organs. The mature meront visualized in histopathological specimens cre-ates a typical “wheel spoke” form when the micromerozoites contained in it are incised incross-section through their midshaft.

We have previously applied an indirect fluorescent antibody test (IFAT) (Shkap et al.,1994) and western immunoblot (Baneth et al., 2000a) for the detection of anti-H. canisantibodies. The IFAT has been used for epidemiological studies in Israel and Japan (Baneth etal., 1996; Inokuma et al., 1999). An enzyme-linked immunosorbent assay (ELISA) utilizingsporozoite antigens derived fromA. maculatumticks has been reported for the detection ofH. americanumantibodies (Mathew et al., 2001). The goal of this study was to develop andevaluate an ELISA forH. canisantibodies and to compare its performance to the IFAT.

2. Materials and methods

2.1. Purification of H. canis gamonts

Blood (80 ml) was collected in heparin from a 4-year-old male Rottweiler naturallyinfected withH. canis. A complete blood count indicated that the white blood cell (WBC)count was 96,000�l−1 and 79% of the neutrophils were parasitized withH. canisgamonts(60,000 gamonts/�l blood). Separation of leukocytes and purification ofH. canisgamontswas carried out as previously described (Baneth et al., 2000a).

2.2. H. canis antigen preparation and ELISA

Purified gamonts suspended in 3.8 ml of phosphate buffered saline (PBS) were mixedwith a protease inhibitor (Roche Diagnostics, Germany) according to the manufacturer’sprotocol. Degradation was performed by sonication for 45 s and repeated six times. Thiswas followed by centrifugation at 10,000×g for 20 min at 4◦C and supernatant collection.The protein concentration of the soluble antigen was determined by the modified Bradfordprotein assay (Bio-Rad, USA). TheH. canisantigen was divided to aliquots and stored forfurther use at−70◦C. ELISA microassay plates were coated for 48 h with 20 ng/�l H. canis

L. Gonen et al. / Veterinary Parasitology 122 (2004) 131–139 133

antigen dissolved in sodium carbonate buffer 0.1 M, pH= 9.5 at 4◦C (50�l/well). Afterthree successive washes with PBS containing 0.1% Tween 20, blocking was performedwith 2% fetal calf serum (FCS) in PBS for 2 h at room temperature. The plates werewashed again three times, as before, and the tested sera diluted at 1:100 in PBS with 0.1%Tween 20 and 2% FCS, were incubated withH. canisantigen for 1 h at 37◦C. After threeadditional washes, the remaining bound antibodies were incubated for 1 h at 37◦C withHRPO-conjugated protein-A (Zymed Laboratories, Inc., San Francisco, USA) diluted at1:10,000 in PBS with 0.1% Tween 20 and 2% FCS. Excess conjugate was removed bywashing as above and a colorimetric reaction was carried out by addition of the chromogen2,2′-azino-bis(3-ehtylbenzthiazoline-6-sulfonic acid) ammonium salt (ABTS) (BoehringerMannheim, Germany) mixed in ABTS buffer. The plates were read (405 nm) at 10 min andthe optical density (OD) readings were standardized to minimize inter-plate variation byusing the same positive control on every plate and adjusting to a constant positive controlvalue.

2.3. IFAT

Serum samples were analyzed by IFAT for IgG antibodies againstH. canis, as previouslydescribed (Shkap et al., 1994). Antigen slides were prepared from the blood of a naturallyinfected dog with a highH. canisparasitemia. The buffy coat was washed with PBS bycentrifugation three times and the final pellet was resuspended in a mixture of equal volumesof PBS and 3% bovine serum albumin (BSA). Thin blood smears were made on glass slidesand dried at room temperature. The slides were immersed for 10 min in acetone and thenstored at−80◦C until used. Before use, antigen slides were warmed and dried at 37◦C for30 min. A series of successive two-fold dilutions of serum in PBS, from 1:16 to 1:4096 wereapplied to the smears and incubated for 30 min at 37◦C in a humid chamber. The slideswere washed three times in PBS and blotted dry. Rabbit anti-dog IgG fluorescein conjugatewas applied to the wells at 1:70 dilution. The slides were then incubated at 37◦C for 30 min,washed as before and dried. The smears were mounted under cover slips in PBS-bufferedglycerol (pH 8.5) and examined under a fluorescent microscope.

2.4. Dog sera used for the ELISA validation

Dog sera analyzed by ELISA or by IFAT were obtained from: (1) negative dogs (n = 29)raised in a tick-free environment or from regions whereH. canis is not prevalent, andfound to be negative forH. canisby microscopy of blood smear; (2) positive dogs withnaturally occurringH. canisparasitemia (n = 36); (3) dogs with naturally occurringH.americanuminfection admitted to the Auburn University Small Animal Clinic in Alabamaand confirmed forH. americanumby histopathology (n = 26); (4) beagle dogs that wereexperimentally infected withH. canis(n = 3) served for the evaluation of positive andnegative results. Infection was performed as previously described (Baneth et al., 2001) andblood was collected 3 days pre-infection and 6 weeks post-infection, whenH. canisgamontswere evident in the blood smears; (5) sera from dogs naturally or experimentally infectedwith other canine pathogens were evaluated for possible cross-reactivity. All serum sampleswere stored at−20◦C until examined.


2.5. Measurement of antibody response in experimentally infected dogs

Six additional dogs were experimentally infected withH. canisas previously described(Baneth et al., 2001). Briefly, Rhipicephalus sanguineusticks were percutaneously injectedat the nymph stage with buffy coat from a naturally infected dog and then allowed to moltto the adult stage. Inoculation of the dogs was performed by feeding the dogs raw chickenliver mixed with theseH. canis-infected adult ticks. Each dog ingested 30 ticks, of which 15had been cut into pieces and 15 left intact. Sera and blood for detection of parasites in bloodsmears were collected pre-infection and periodically over 156–216 days post-infection (PI).

2.6. Statistical analysis

The ELISA OD values of the groups of sera were compared using a Kruskal–Wallisanalysis of variance. Subsequent pairwise comparisons were made using Mann–Whitneytests with a Boneferroni-type adjustment for multiple comparisons. Logistic regression wasused to calculate the cut-off titer forH. canisseropositivity. The ELISA and IFAT resultsfor parasitemic dogs were compared using the McNemar paired chi-square test.P < 0.05were considered statistically significant.

3. Results

3.1. ELISA for H. canis parasitemic and for H. americanum-infected dogs

The anti-HepatozoonELISA OD values for parasitemic (n = 36) and non-parasitemiccontrol dogs (n = 29) are shown inFig. 1. These were compared to the seroreactivityof dogs with histopathologically confirmedH. americanuminfection (n = 26). All threegroups were significantly different from each other (P < 0.05). A cut-off value of 0.43 ODwith a sensitivity of 86% and specificity of 97% was calculated using logistic regressionanalysis of theH. canisparasitemic and non-parasitemic dogs.

No OD readings equal or higher than the cut-off value were detected in sera from dogsinfected withToxoplasma gondii(n = 1), Neospora caninum(n = 1), Babesia canis(n = 1), Babesia gibsoni(n = 1), Spirocerca lupi(n = 1), Dirofilaria repens(n = 1),Ehrlichia canis(n = 6), Ehrlichia platys(n = 1), Rickettsia rickettsii(n = 1) and caninedistemper virus (n = 1).

3.2. Comparison of ELISA and IFAT for H. canis

A comparison of ELISA and IFAT using sera from parasitemic (n = 36) and controldogs (n = 5) (Table 1) indicated that 31 of 36 dogs (86%) withH. canisparasitemia wereseropositive by ELISA, whereas 28 (78%) were positive by IFAT. Two parasitemic dogs(6%) were negative by both ELISA and IFAT. Of the 28 IFAT-positive sera, 25 (89%)were positive also by ELISA. A statistical analysis comparing the rates of seropositivityin parasitemic dogs by ELISA and IFAT indicated that there was no significant difference


O.D.

0

0.5

1

1.5

2

2.5

H. canis pos. Neg. control H. americanum pos.

------

------------

Fig. 1. Distribution of ELISA optical density (OD) values for sera from dogs withH. canisparasitemia, negativecontrols and dogs infected withH. americanum. The short discontinuous lines represent the median OD value foreach group of dogs. A long discontinuous line represents the cut-off value.

between these two serologic methods (P = 0.508). The three dogs that were seropositiveby IFAT and had an ELISA OD value below the cut-off, had a low to moderate IFAT titerof 1:32 (n = 1) or 1:64 (n = 2). Seventeen of the 19 dogs (89%) with a high value ofOD ≥ 0.8 also had a moderate to high IFAT titer of≥1:64. Of the dogs with an ELISA ODvalue below 0.8, three of six (50%) had an IFAT titer below 1:64.

3.3. Kinetics of antibodies in experimentally infected dogs

Five of the six experimentally infected dogs (nos. 1, 2, 3, 5 and 6) developed an OD valueabove the cut-off (0.43) and seroconverted within the first month PI (Fig. 2). Dog no. 5 hada positive OD value as early as 7 days PI. Dog no. 6 was seropositive on day 18, nos. 1 and

Table 1Comparison of ELISA and IFAT forH. caniswith sera from parasitemic and negative control dogs

H. canisparasitemia

ELISA IFAT

(+) (−) (+) (−)

Seropositive 31 0 28 0Seronegative 5 5 8 5No. of dogs 36 5 36 5

Total seropositive 31/36 (86%) 0/5 (0%) 28/36 (78%) 0/5 (0%)

136L

.Go

ne

ne

tal./V

ete

rina

ryP

ara

sitolog

y1

22

(20

04

)1

31

–1

39

Dog 1

0

0.5

1

1.5

2

2.5

3

0 50 100 150 200 250

Days post infection

Opt

ical

den

sity

Dog 2

0

0.5

1

1.5

2

2.5

3

0 50 100 150 200 250

Day post infection

Opt

ical

den

sity

Dog 3

0

0.5

1

1.5

2

2.5

3

0 50 100 150 200 250

Days post infection

Opt

ical

den

sity

Dog 4

0

0.5

1

1.5

2

2.5

3

0 50 100 150 200 250

Days post infection

Opt

ical

den

sity

Dog 5

0

0.5

1

1.5

2

2.5

3

0 50 100 150 200 250

Days post infection

Opt

ical

den

sity

Dog 6

0

0.5

1

1.5

2

2.5

3

0 50 100 150 200 250

Days post infection

Opt

ical

den

sity

Fig. 2. The kinetics of antibody response toH. canisin six experimentally infected dogs as detected by ELISA.


3 at day 20 and no. 2 at day 28. Dog no. 4 was seronegative during most of the monitoringperiod except for three occasions on days 98, 134 and 216 on which his OD values weremoderately above the cut-off level.

The maximal OD values were reached between days 51–62 PI for dog nos. 1, 2, 5 and6, and on day 98 PI for dog no. 3. The mean maximal OD value for these dogs was 1.98(range 1.38–2.81) which represented a 4.5- to 8.7-fold increase over the pre-infection ODvalues. Over time, the OD values declined gradually in all the five dogs that seroconvertedwithin the first month, however, positive OD values were still evident 156–197 days PI indog nos. 1, 2, 3, 5 and 6 (Fig. 2).

Parasitemia withH. caniswas initially observed in dog nos. 2, 5 and 6 at 78, 46 and 37 daysPI, respectively. The parasitemia appeared in these three dogs following the administration ofan immunosuppressive course of 3 mg/kg of prednisone for 5 days injected sub-cutaneouslystarting at 36 days PI for dog nos. 5 and 6, and at 76 days PI for dog no. 2. Parasitemia wasevident within 48 h of corticosteroid administration in dog nos. 2 and 6, and 10 days afterprednisone in dog no. 5. Dogs 1, 3 and 4 who did not receive prednisone were not observedto have parasitemia. Following the prednisone administration, the OD values of dog nos. 5and 6 remained stable and further elevated within 2–3 weeks, whereas the OD values fordog no. 2 gradually declined over a period of more than 4 months.

4. Discussion

The ELISA developed was shown to be sensitive and specific for the detection of anti-H.canisantibodies in naturally infected dogs with parasitemia. Although there was no sta-tistical difference between the IFAT and ELISA in this study, there are several technicaladvantages to the ELISA over the IFAT in our experience. ELISA is easier to perform, canbe carried out on large numbers of sera rapidly, and is in general more objective than IFAT.Failure to detect anti-H. canisantibodies in parasitemic dogs can be due to several reasonsincluding an early infection prior to the production of a measurable antibody response,chronic infection with a decay below detection level of antibodies, and a possible lack of aconsiderable humoral response to the predominantly intracellularH. canisin some infecteddogs.

The disagreement between the ELISA and the IFAT in the detection of antibodies reactivewith H. canisin some dog’s sera could be due to a difference in the targets of antibodiesdemonstrated by each method. The IFAT principally detects antibodies reactive with anti-gens bound to the parasites’ outer membrane, whereas the ELISA detects antibodies towhole parasite soluble antigens. It is therefore reasonable that the ELISA would allowexposure of serum antibodies to a larger spectrum of putative antigens. The obvious rela-tionship between the IFAT titer and the intensity of the ELISA OD value for the majorityof sera from the parasitemic dogs in this study, further validates the ELISA and verifies itsvalue.

A variable degree of serologic cross-reactivity between sera fromH. americanum-infecteddogs andH. canisantigen was found. This is in agreement with previous studies using IFATand western immunoblotting (Vincent-Johnson et al., 1997; Baneth et al., 2000b). How-ever, theH. canisELISA cannot be recommended for practical use in the diagnosis of


H. americanumdue to the considerable overlap in OD readings between negative non-infected dogs andH. americanum-infected ones, as depicted inFig. 1. A more specificassay employingH. americanumantigen, such as theH. americanumsporozoite ELISA(Mathew et al., 2001), should facilitate serologic diagnosis ofH. americanuminfection.

The experimental infection of dogs withH. canisproduced a marked antibody responsein five of six dogs. Seroconversion occurred 1–4 weeks PI and the antibody levels peaked7–9 weeks PI in four dogs. Antibody levels declined thereafter but remained above thecut-off value for 5–7 months PI when the experiment was concluded. In dogs that developedparasitemia, antibodies were evident long before parasitemia occurred. In dog nos. 6, 5 and2, seroconversion preceded parasitemia by 19, 39 and 50 days, respectively. It is not clearwhether dog no. 4 who only had occasional positive OD values became truly infected. Itdid not have detectable parasitemia and in contrast to the other five dogs, did not developa pattern of peaking antibody response followed by a gradual decline. The low OD valuesfound in the sera of this dog might have resulted from the oral exposure toH. canisantigenduring the inoculation.

Interestingly, parasitemia was only evident in dogs that received immunosuppressivesteroid therapy. This differs from a previous experimental infection that we conducted(Baneth et al., 1998) where spontaneous parasitemia was evident in four of five dogs within28–43 days PI. The fifth dog in the earlier experiment seroconverted by IFAT and hadH. canismeronts in the bone marrow and spleen at necropsy 53 days PI. The fact thatthe three dogs in the present study became parasitemic only after immunosuppressionexemplifies that dogs can probably become carriers and harbor tissue forms ofH. canisuntil immunosuppression or possible concurrent disease facilitate the completion of the lifecycle with the release of mature gamonts to the blood circulation.

In conclusion, the ELISA employing gamont antigen was shown to be a sensitive andspecific method for the detection of antibodies reactive withH. canis. The kinetics ofantibody response in experimentally infected dogs indicated that oral inoculation ofH.canisoocysts is associated with a long-lasting humoral response. The antibody response wasdetectable in dogs that harbored the infection even prior to the appearance of parasitemia.TheH. canisELISA will be a valuable tool for evaluating dogs that have been exposed tovector ticks and are potentially infected with this parasite.

Acknowledgements

The authors thank Dr. Philip H. Kass for his assistance in the statistical analysis of theresults. This study was partially funded by Merial Ltd., Lyon, France.

References

Baneth, G., Shkap, V., Presentey, B.Z., Pipano, E., 1996.Hepatozoon canis: the prevalence of antibodies andgametocytes in dogs in Israel. Vet. Res. Commun. 20, 41–46.

Baneth, G., Weigler, B., 1997. Retrospective case–control study of hepatozoonosis in dogs in Israel. J. Vet. Intern.Med. 11, 365–370.

Baneth, G., Shkap, V., Pipano, E., Savitsky, I., Samish, M., 1998. Antibody response toHepatozoon canisinexperimentally infected dogs. Vet. Parasitol. 74, 297–303.


Baneth, G., Shkap, V., Samish, M., Jaffe, C.L., 2000a. Antigenic analysis of gamonts ofHepatozoon canispurifiedfrom leukocytes. J. Parasitol. 86, 289–294.

Baneth, G., Barta, J.R., Martin, D.S., Macintire, D.K., Vincent-Johnson, N., 2000b. Genetic and antigenic evidencesupports the separation ofHepatozoon canisandHepatozoon americanumat the species level. J. Clin. Microbiol.38, 1298–1301.

Baneth, G., Samish, M., Alekseev, E., Aroch, I., Shkap, V., 2001. Transmission ofHepatozoon canisto dogs bynaturally-fed or percutaneously-injectedRhipicephalus sanguineusticks. J. Parasitol. 87, 606–611.

Baneth, G., Mathew, J.S., Shkap, V., Macintire, D.K., Barta, J.R., Ewing, S.A., 2003. Canine hepatozoonosis—twodisease syndromes caused by separateHepatozoonspecies. Trends Parasitol. 19, 27–31.

Ezekoli, C.D., Ogunkoya, A.B., Abdullahi, R., Tekedec, L.B., Sannusi, A., Ilemobade, A.A., 1983. Clinical andepidemiological studies on canine hepatozoonosis in Zaria, Nigeria. J. Small Anim. Pract. 24, 455–460.

Hervas, J., Carraso, L., Gomez-Villamandos, J.C., Mendez, A., Sierra, A.M., 1995. Acute fatal hepatozoonosis ina puppy: histopathological and ultrastructural study. Vet. Rec. 137, 518–519.

Inokuma, H., Ohno, K., Yamamoto, S., 1999. Serosurvey ofEhrlichia canisandHepatozoon canisinfection indogs in Yamaguchi Perfecture, Japan. J. Vet. Med. Sci. 61, 1153–1155.

Kontos, V., Koutinas, A., 1990. Canine hepatozoonosis: a review of 11 naturally occurring cases. Bull. Hellen.Med. Soc. 41, 73–81.

Mathew, J.S., Ewing, S.A., Panciera, R.J., Woods, J.P., 1998. Experimental transmission ofHepatozoonamericanum(Vincent-Johnson et al., 1997) to dogs by the Gulf Coast tick,Amblyomma maculatumKoch.Vet. Parasitol. 80, 1–14.

Mathew, J.S., Saliki, J.T., Ewing, S.A., Lehenbauer, T.W., Panciera, R.J., Malayer, J.R., Cummings, C.A., Kocan,A.A., 2001. An indirect enzyme-linked immunosorbent assay for diagnosis of American canine hepatozoonosis.J. Vet. Diagn. Invest. 13, 17–21.

McCully, R.M., Basson, P.A., Bigalke, R.D., de Vos, V., Young, E., 1975. Observation on naturally acquiredhepatozoonosis of wild carnivores and dogs in the Republic of South Africa. Onderstepoort J. Vet. Res. 42,117–133.

Murata, T., Shimoda, K., Inoue, M., Shiramizu, K., Kanoe, M., Taura, Y., Nakama, S., 1993. Seasonal periodicalappearance ofHepatozoon canisgamonts in the peripheral blood. J. Vet. Med. Sci. 55, 877–879.

O’Dwyer, L.H., Massard, C.L., Pereira de Souza, J.C., 2001.Hepatozoon canisinfection associated with dog ticksof rural areas of Rio de Janiero State, Brazil. Vet. Parasitol. 94, 143–150.

Rajamanickam, C., Wiesenhutter, E., Zin, F., Hamid, J., 1985. The incidence of canine hematozoa in peninsularMalaysia. Vet. Parasitol. 17, 151–157.

Rioux, J.A., Golvan, Y.J., Honin, R., 1964. MixedHepatozoon canisandLeishmania canisinfection in a dog inthe Sets area, France. Ann. Parasitol. Hum. Comp. 39, 131–135.

Shkap, V., Baneth, G., Pipano, E., 1994. Circulating antibodies toHepatozoon canisdemonstrated byimmunofluorescence. J. Vet. Diagn. Invest. 6, 121–123.

Vincent-Johnson, N.A., Macintire, D.K., Lindsay, D.S., Lenz, S.D., Baneth, G., Shkap, V., Blagburn, B.L., 1997.A new Hepatozoonspecies from dogs: description of the causative agent of canine hepatozoonosis in NorthAmerica. J. Parasitol. 83, 1165–1172.

an enzyme-linked immunosorbent assay for antibodies to hepatozoon canis

Documents