- Email: [email protected]
Ovine Fetal Necrobacillosis
ARTICLE IN PRESS J. Comp. Path. 2007,Vol. 136, 213^221
www.elsevier.com/locate/jcpa
Ovine Fetal Necrobacillosis J. S. Agerholm, M. Boye* and B. Aalbæk Department of Veterinary Pathobiology, Faculty of Life Sciences, University of Copenhagen, Copenhagen, and *NationalVeterinary Institute,Technical University of Denmark, Copenhagen, Denmark
Summary The association of Fusobacterium necrophorum with late term abortion in sheep is reported. The bacterium was not culturable, but was identi¢ed in ¢ve cases by £uorescence in-situ hybridization (FISH) with an oligonucleotide probe speci¢cally targeting16S rRNA in F. necrophorum. Gross lesions were found in several tissues. Histologically, placental lesions were characterized by locally di¡use in¢ltration of neutrophils, closely associated with abundant small Gram-negative and FISH-positive rods, thrombosis and necrosis. Lesions in the fetal^maternal interface were multifocal and consisted of villous necrosis and suppurative in£ammation. Spread to the fetus from the placenta appeared to occur in two ways. Some fetuses had multifocal necrotizing hepatitis consistent with haematogenous spread through the umbilical vein; further dissemination to other organs occurred. Transplacental spread and infection of the fetus through the amniotic £uid was characterized by development of multifocal suppurative dermatitis and suppurative bronchopneumonia. Localization of FISH-positive bacteria in necrotic lesions was restricted to the periphery. F. necrophorum would seem to have been unrecognized previously as a cause of abortion. The value of culture-independent diagnostic methods is emphasized. r 2007 Elsevier Ltd. All rights reserved. Keywords: Fusobacterium necrophorum; abortion; necrobacillosis; sheep
Introduction Fusobacterium necrophorum, a Gram-negative, strictly anaerobic pleomorphicbacterium, is an opportunistic pathogen in several animal species (Langworth,1977). In ruminants, the bacterium may act as a principal pathogen or as an important constituent in mixed bacterial infections.The most signi¢cant diseases associated with F. necrophorum in ruminants are the ‘‘rumenitis-liver abscess complex’’ (Scanlan and Edwards, 1990; Nagaraja and Chengappa, 1998), necrobacillosis of the foot (Parsonson et al., 1967; Yager and Scott, 1985), and necrotizing stomatitis (Barker and Van Dreumel, 1985; McCourtie et al., 1990). In neonatal ruminants, necrotizing hepatitis may be the result of ascending infection from omphalophlebitis (Kelly,1985). Abortion due to F. necrophorum was reported in a bovine fetus showing widespread serosal in£ammation (Otter, 1996); other reported cases (Pattnaik et al., 1994), have
lacked histopathological con¢rmation. Isolation of the bacterium from aborted fetuses is not su⁄cient to establish a diagnosis, as F. necrophorum is a common invader of the bovine uterus post partum (Bekana et al.,1994,1997). Opening of the cervical canal before abortion may lead to colonization of the fetus and its membranes, or these structures may become contaminated during passage through the caudal parts of the genital tract. Consequently, diagnosis must be based on histopathological recognition of in£ammation in association with F. necrophorum. Kirkbride et al. (1989) reported ovine abortion due to Fusobacterium nucleatum, but fetal infection with F. necrophorum in sheep has not been reported. The present report describes ovine fetuses spontaneously infected with F. necrophorum.
Materials and Methods Fetuses Infected with F: necrophorum
Correspondence to: J.S. Agerholm, Department of Veterinary Pathobiology, Faculty of Life Sciences, University of Copenhagen, Ridebanevej 3, DK-1870 Frederiksberg C, Denmark (e-mail: [email protected]). 0021-9975/$ - see front matter
doi:10.1016/j.jcpa.2007.01.012
The animals were obtained from a larger study on the causes of prenatal mortality in Danish sheep (Agerholm r 2007 Elsevier Ltd. All rights reserved.
ARTICLE IN PRESS 214
J.S. Agerholm et al.
et al., 2006). Initially, six fetuses consisting of three pairs of twins (cases 1A,B ^ 3A,B) with an unidenti¢ed bacterial infection were selected, forming a group of cases with similar in£ammatory lesions closely associated with microcolonies of Gram-negative bacteria. These bacteria were identi¢ed as F. necrophorum by laser capture microdissection of bacterial colonies from formalin-¢xed tissue, followed by polymerase chain reaction (PCR) ampli¢cation of bacterial 16S rDNA and sequencing. An rRNA-targeting oligonucleotide probe speci¢c for F. necrophorum was designed and used in a £uorescent in-situ hybridization (FISH) assay (see below). This examination identi¢ed a further pair of F. necrophorum-infected twin fetuses (cases 4A,B) and an infected placental specimen from a sheep (case 5) that aborted in late gestation. Placental material was not available from cases 2A, 3A and 4A. Pathology
The fetal age was calculated from breeding data and compared with the stage of development. The fetus and corresponding fetal membranes were examined post mortem and tissues including placenta, lung, myocardium, liver, kidney, skeletal muscle, brain and cervical spinal cord were sampled for histopathology. Multiple areas were sampled from placenta, lung, liver, and brain and other tissues were taken if showing gross lesions. The specimens were ¢xed in 10% neutral buffered formalin, dehydrated through graded alcohols, embedded in para⁄n wax and sectioned at 7 mm (brain) or 3 mm (other tissues). Initial histological examination was performed on haematoxylin and eosin (HE)-stained sections. Additional staining methods were applied when considered appropriate. These included Mallory’s phosphotungstic acid haematoxylin method (PTAH), Masson’s trichrome method, the periodic acid-Schi¡ (PAS) technique, and the BrownHopps method for Gram-positive and Gram-negative bacteria. Immunohistochemistry
Immunohistochemical examination for Listeria monocytogenes antigen was performed on all tissues of any pair of twins from which this organism was cultured. Selected sections from the other cases were also examined. A commercially available polyclonal antibody against L. monocytogenes types 1 and 4 (Difco Laboratories, Detroit, MI, USA) was used in a streptavidinbiotin immunoperoxidase assay. Bacteriology
This was performed on samples of lung, liver, fetal placenta and abomasal contents, as previously described
(Agerholm et al., 2006). Brie£y, each sample was spread on two blood agar plates (Blood Agar Base CM0055; Oxoid, Basingstoke, UK, containing sterile bovine blood 5%) of which one was incubated aerobically and the other anaerobically for 48 h at 37 1C. Colonies were subcultured and identi¢ed by standard methods. Selected isolates were characterized by the API-20-E kit (BioMeŁrieux, Marcy l’Etoile, France) according to the manufacturer’s recommendations. Examination for Campylobacter spp. was performed as previously described (Agerholm et al., 2006). In-situ Hybridization
This was performed with the general bacterial probe eub338: S-D-Bact-0338-a-A-18 50 -GCTGCCTCCCGTAGGAGT-30 (Amann et al., 1990) and a probe speci¢c for F. necrophorum (F.necrophorum183: S-SF.necrophorum-0183-a-A-18, 50 -GATTCCTCCATGCGAAAA-30 [Boye et al., 2006]). Epifluorescence Microscopy and Scanning
An Axio Imager M1 epi£uorescence microscope (Carl Zeiss, Oberkochen, Germany) equipped for epi£uorescence with a 100-W HBO lamp and ¢lter set XF53, (Omega Optical, Brattleboro, VT, USA) was applied for simultaneous detection of red and green £uorescence. Filter sets 43 HE and 38 HE (Carl Zeiss) were used to ‘‘visualize’’ CY3 and £uorescein, respectively. Images were obtained with an AxioCam MRm version 3 FireWire monochrome camera (Roper Scienti¢c, Ottobrunn, Germany) and the software AxioVision version 4. Entire tissue sections were examined by scanning £uorescence in-situ hybridization using an ArrayWoRxe scanner (pixel resolution of 3.25 mm) (Applied Precision, Issaquah,WA, USA). PCR and Sequencing
DNAwas extracted from dermal lesions of fetus1A kept at 20 1C until used. Subcutaneous tissue was removed and material scraped with a scalpel from the dermis was suspended in 1ml of 0.9% NaCl. DNA was extracted with the Easy-DNA kit (InVitrogen, Paisley, UK) and ‘‘protocol #3,’’according to the manufacturer’s instructions. Primers (MWG-BIOTECH AG, Ebersberg, Germany) used for 16S rRNA gene ampli¢cation consisted of S-D-Bact-0008-a-S-20 (50 -AGAGTTTGATCMTGGCTCAG-30) and S-Univ-*-1492-a-A-21 (50 -GTTACCTTGTTACGACTTCAG-30), modi¢ed fromWeisburg et al. (1991). Oligonucleotide probes were designated according to Alm et al. (1996). Reaction conditions were as follows:1 PCR Bu¡er II (Roche Applied Science, Indianapolis, IN, USA), 3 mM MgCl2,
ARTICLE IN PRESS Ovine Fetal Necrobacillosis
2.5 mM of each deoxynucleotide triphosphate (Amersham Biosciences, Piscataway, NJ, USA), 0.4 mM of each primer, 0.5 U of AmpliTaq DNA Polymerase (Roche Applied Science) and 2.0 ml of template DNA in a total volume of 50 ml. Thermocycling with a T3 Thermocycler (BioMetra, Goettingen, Germany) was performed as follows: denaturation at 94 1C for 3 min followed by 35 cycles of denaturation at 94 1C for 1min, annealing at 55 1C for 1min, extension at 72 1C for 1ˆ˜ min, and a ¢nal extension at 72 1C for 10 min. Ampli¢ed PCR products were visualized by electrophoresis in 1.5% agarose gels. The amplicons were puri¢ed with the QIAquick spin PCR puri¢cation kit (Qiagen, Hilden, Germany) and sequenced by cycle sequencing on an ABI PRISM 377 DNA Sequencer (PE Biosystems, Foster City, CA, USA) with the ABI PRISM Big Dye Terminator Cycle Sequencing Kit (PE Biosystems), according to the manufacturer’s instructions and with the following primers: S-D-Bact-0008-a-S-20, S -*-Univ-0519-a-A-18 (50 GTATTACCGCGGCTGCTG-30), S-D-Bact-0786-a-A-21 (50 -GACTACCNGGGTATCTAATCC-30), S-D-Bact-1054-a-A-20 (50 -ACGAGCTGACGACRRCCATG’-30) and S-*Univ-1492-a-A-21. The sequence was compared with published 16S rDNA sequences in GenBank (www.ncbi.nlm.nih. gov/) by means of a Blast search for homology. The near-full length 16S rDNA sequence has been deposited in the GenBank database under accession number DQ486127. Other Laboratory Examinations
The animals were examined for other infectious organisms including Toxoplasma gondii, Border disease virus (BDV), Coxiella burnetii and Chlamydophila abortus and tested for antibodies againstT. gondii and BDV in pleural e¡usions (Agerholm et al., 2006).
215
part of a mixed culture. A few L. monocytogenes bacteria were identi¢ed by immunohistochemistry in the liver of case 4A and on the placental surface of case 4B. F. necrophorum was identi¢ed in close association with lesions in most cases (1A,B; 2A,B; 3A,B; 4B; 5) by means of the FISH as previously described (Boye et al., 2006). Additionally, 16S rDNA was isolated, PCR-ampli¢ed, and sequenced from frozen skin of fetus 1A; a BLAST search revealed 99% nucleotide homology with both the type strain for F. necrophorum (ATCC 25286) and a ruminant F. necrophorum strain isolated in New Zealand (GenBank Acc. No. AF044948). No other pathogens and no fetal antibodies against T. gondii or BDV were found. Gross Lesions
Placentitis, seen in four cases (1A,B; 2B; 4B), was characterized by slight necrosis of cotyledons and intercotyledonary areas, oedema, haemorrhage and variable amounts of purulent exudate (Fig. 1). Fetus 1A had extensive skin lesions, spread over the entire body and consisting of multifocal hyperaemic elevated nodules, with a diameter of p 5 mm and occasional central ulceration (Fig. 2). Pulmonary lesions were observed in four fetuses, either as pale disseminated pinpoint lesions (4B) (Fig. 3), or as widespread consolidation of the lung, sometimes with elevated hyperaemic lobules (2A,B; 3A). Hepatomegaly was observed in six fetuses. In addition, two of these (1A; 2A) had focal areas of necrosis in the liver, either as separate lesions or as multiple lesions grouped in severely autolysed and emphysematous areas (Fig. 4). Super¢cial lesions were covered by perihepatitis. Multiple necrotic areas (diameter p 5 mm) were found in the spleen of case 1A. Only in case 4B were in£ammatory lesions absent; this
Results Clinical Signs
The abortions occurred in four herds during the ¢nal month of gestation. One sheep had a depressed appetite before abortion, but the others were clinically normal. One lamb (4B) was born weak and died within 24 h. Agent Identification
Except for case 4B, lung, liver and abomasal content were found to be bacteriologically sterile by cultural examination, while mixed bacterial cultures were obtained from the placentas. In fetus 4B, L. monocytogenes was isolated from the liver and abomasal content as
Fig. 1. Local purulent placentitis with a single in£amed cotyledon (arrowhead) and a purulent exudate in the intercotyledonal placenta (arrow) due to F. necrophorum. Ovine placenta, case 2B.
ARTICLE IN PRESS 216
J.S. Agerholm et al.
fetus had leptomeningeal haemorrhage in the caudal cerebrum, softening of the cerebral white matter, and multifocal pinpoint haemorrhages in the lung.
Fig. 2. Multifocal in£amed nodules in the skin of an ovine fetus due to F. necrophorum. Case 1A.
Histopathology
In mono-infected fetuses, lesions in the cotyledonary villi were mild and consisted of focal necrosis and mild neutrophil in¢ltration. However, in the deep stroma of the cotyledons and of the intercotyledonary placenta, the tissue had severe surface-orientated in£ammation, characterized by extensive, locally di¡use, neutrophil in¢ltration, closely associated with numerous small Gram-negative rods, necrosis of the connective tissue and amnionic/allantoic surface, oedema, and acute haemorrhage (Fig. 5).The walls of arteries and veins located in in£amed tissue showed peripheral necrosis, neutrophil in¢ltration, and sometimes transmural in£ammation and developing thrombosis (Fig. 6). Bacterial microcolonies occasionally ¢lled the lumen of non-in£amed thin-walled vessels. Bronchopneumonia, found in all fetuses, was characterized by widespread, intense hyperaemia, haemorrhages, and accumulations of neutrophils and debris in bronchi, bronchioles and alveoli. The distribution of bacteria varied from single organisms to multiple mi-
Fig. 3. Disseminated lesions of pinpoint pulmonary necrosis (arrowheads) due to F. necrophorum septicaemia. Lung, ovine fetus, case 4B.
ARTICLE IN PRESS Ovine Fetal Necrobacillosis
217
Fig. 4. Two major hepatic lesions consisting of multiple necrotic areas are present in autolysed and emphysematous areas. Additionally, pinpoint necrotic lesions are disseminated in the liver. F. necrophorum septicaemia, ovine fetus, case 1A.
Fig. 5. Intercotyledonary placenta with di¡use purulent in£ammation (I) and oedema (O) in the connective tissue and super¢cially located colonies of F. necrophorum (arrowhead). Ovine placenta, case 1A. HE. Bar, 50 mm.
Fig. 6. Large numbers of F. necrophorum (arrow) surrounding an in£amed artery undergoing thrombosis. Ovine placenta, case 4B. HE. Bar,100 mm.
ARTICLE IN PRESS 218
J.S. Agerholm et al.
Fig. 7. Large numbers of F. necrophorum (arrowheads) located in the centre (C) and in the periphery (P) of a large chronic hepatic lesion. Note the intense bacterial staining in the periphery, in contrast to weak staining of bacteria in the central area. Ovine fetus, case 2A. HE. Bar,100 mm.
Fig. 8. Acute focal hepatic lesion with a central area occupied by large numbers of F. necrophorum surrounded by purulent in£ammation. Ovine fetus, case 2A. HE. Bar,100 mm.
crocolonies, plugging and distending the alveoli. As compared with the bacteria in the placenta, those in the lungs were longer, and microcolonies occasionally contained intermingled, ¢lamentous structures. In several cases, large bacterial colonies were not associated with marked in£ammation, suggesting continued bacterial growth after fetal death. In all cases lesions tended to be widespread, a¡ecting both cranial and diaphragmatic lobes. Multiple necrotic areas with large numbers of intralesional Gram-negative bacteria and peripheral neutrophil in¢ltration were present in the liver of cases 1A and 2A. Peripherally located bacteria showed intense
staining, but staining decreased in intensity towards the centre of major lesions, probably due to bacterial death (Fig. 7). Major areas of necrosis were partly surrounded by ¢brosis, but bacterial invasion of the surrounding parenchyma and corresponding neutrophil in¢ltration were seen locally. Adjacent to major lesions, multiple acute foci of ¢brinous and suppurative in£ammation surrounding bacterial colonies were present (Fig. 8). The spleen of case 1A showed multifocal areas of acute necrosis, with large bacterial colonies, central debris, and peripheral in£ammation. In the skin of case 1A, multiple areas of severe suppurative in£ammation extending into the deep layer of the dermis and occasionally into the subcutis were present. The epithelium at the periphery of each lesion showed hydropic degeneration and necrosis. Large numbers of Gram-negative rods were present at the margin of the lesions. Twin fetuses often di¡ered from each other in respect of the severity of lesions and degree of autolysis. Severe lesions were usually associated with large numbers of bacteria and pronounced autolysis. The lesions in the fetuses infected with both F. necrophorum and L. monocytogenes (cases 4A,B) di¡ered from the general ¢ndings. Extensive necrosis of the stroma with oedema and suppurative in£ammation was present in the placenta of fetus 4B. In necrotic areas, veins and small- and medium-sized arteries showed transmural necrosis and thrombosis, while large arteries showed only peripheral necrosis and in£ammation. In the periphery of necrotic areas, multiple colonies of Gram-negative rods were present. In some areas lesions pierced the placenta. Lesions in the interface were multifocal, consisting of villous necrosis and suppurative in£ammation, occasionally with extension into the underlying connective tissue. Abundant Gram-negative bacteria were seen in close association with the lesions and also in the cytoplasm of trophoblasts. Immunolabelling did not reveal L. monocytogenes in such areas; these organisms were seen exclusively on the cotyledonal surface in association with bedding debris and a mixed bacterial and fungal population. In the lung, disseminated foci were present. Minor lesions consisted of accumulations of Gram-negative bacteria interspersed with neutrophils, while larger lesions had a central area of necrosis, surrounded by large numbers of bacteria spreading through the alveoli, and peripheral suppurative in£ammation (Fig. 9). In some areas suppurative bronchio-alveolitis with bacteria in the bronchi was seen. Hepatic lesions were mild and dominated by interstitial lymphocytic in¢ltration. Colonies of Gram-negative bacteria were seen in the sinusoids, without associated in£ammation. Additionally, multifocal necrotizing pyelonephritis with numerous bacteria in the renal pelvis was observed.
ARTICLE IN PRESS 219
Ovine Fetal Necrobacillosis
Fig. 9. Acute focal pneumonia with numerous F. necrophorum and debris surrounded by a purulent exudate. Ovine fetus, case 4B. HE. Bar, 50 mm.
Fig. 11. Margin of a chronic hepatic lesion showing intense peripheral orange £uorescence (P) and weaker orange £uorescence towards the centre (C). Ovine fetus, case 4B. FISH. Bar, 20 mm.
periphery of large necrotizing lesions, while £uorescence like Gram-staining (see above), was absent in the central areas (Figs 10 and 11). This was probably due to bacterial death, as the signal from the rRNA-targeting probe is dependent on the metabolic activity of the cells. Except for case 1A, F. necrophorum was present only in close association with lesions, and was absent in tissues without lesions. In case 1A, F. necrophorum was seen in association with lesions in placenta, skin, liver, lung and spleen. However, occasional bacteria not associated with lesions were seen in the kidney, skeletal muscle and central nervous system, suggesting terminal haematogenous spread or growth post mortem.
Fig.10. Demonstration of metabolically active F. necrophorum bacteria (orange) in the periphery of a large chronic hepatic lesion. A few associated acute lesions are seen as foci of intense orange £uorescence. Ovine fetus, case 2A. FISH. Bar, 2.5 mm.
In fetus 4A, the most prominent lesion was an extensive subacute cerebral leucoencephalomalacia with massive in¢ltration of macrophages and meningeal haemorrhage. Other tissues showed acute congestion, but the placenta was not available for examination. A few immunolabelled L. monocytogenes-like structures were seen in the liver, without adjacent in£ammation. Epifluorescence Microcopy
The in-situ hybridization demonstrated a close association of F. necrophorum with lesions. An intralesional FISH signal was seen in areas corresponding to the location of the Gram-negative bacteria as seen in light microscopy. An intense FISH signal was present in the
Discussion The study demonstrated that F. necrophorum infection in ovine fetuses was associated with severe in£ammation of multiple tissues. The close association of the bacterium with pyonecrotizing lesions strongly indicated that F. necrophorum was the cause of fetal or neonatal death. The sequencing of 16S rDNA obtained from intralesional bacterial colonies by laser capture microdissection (Boye et al., 2006) and by sequencing of 16S rDNA extracted from the in£amed skin of case 1A, showed a 99% nucleotide homology with F. necrophorum. The fetuses were obtained from a study on the causes of prenatal mortality in Danish sheep, which demonstrated that F. necrophorum frequently caused congenital infection (Agerholm et al., 2006). That such infections have not been reported previously may re£ect di⁄culties in bacteriological culture or exclusion of anaerobic culturing in routine diagnostic bacteriology. Isolation of the bacterium in the present study was unsuccessful, even though a method that regularly identi¢es
ARTICLE IN PRESS 220
J.S. Agerholm et al.
F. necrophorum in pathological specimens (Aalbk, personal observation) was used. As multiple specimens containing large numbers of bacteria were examined, the unsuccessful attempts to isolate the bacterium probably re£ected decreased viability upon arrival at the laboratory.This was not due to antibiotic treatment, as the sheep were not medicated before abortion. In aborted fetuses showing pyonecrotizing in£ammation associated with Gram-negative rods, the use of diagnostic methods other than bacteriological culture, e.g., FISH, would seem advisable. The localization of the most severe lesions in the deep stroma of the fetal placenta may indicate placental infection other than by maternal bacteraemia, which is the generally assumed route of bacterial infection in the pregnant uterus of ruminants (Jubb et al., 1985). F. necrophorum bacteraemia in sheep is known to occur following bacterial penetration of the ruminal epithelium, which causes a local portal bacteraemia with subsequent hepatitis (‘‘rumenitis-liver abscess complex’’) (Nagaraja and Chengappa, 1998), but systemic bacteraemia may also occur (Ramos-Vara et al., 1997). Initial placentomal localization is expected to occur in the interdigitations created by villi and crypts. In aborted placentas this is seen as in£ammation of the villous portion of the cotyledons; later the in£ammation may penetrate into the underlying stroma. In this study, cotyledonary changes suggested initial lesion development in the deep connective tissue, with subsequent expansion and eruption to the placentomal interface, causing localized in£ammation and necrosis of villi. It could not be determined whether the necrosis of the amniotic or allantoic surface occurred as a result of an expanding lesion or from bacteria present in the £uids. Equally it might be argued that the fetal maternal interface may have an oxygen-rich environment unfavourable for strict anaerobes as F. necrophorum, or that representative specimens were not examined. Alternatively, it is possible that an ascending transcervical infection caused a dissecting in£ammation in the fetal placenta and a transplacental spread to the amniotic and allantoic £uids, causing widespread surface in£ammation. A prequisite for this hypothesis is the presence of F. necrophorum on the mucosal membranes of the ovine genital tract. Although not thoroughly studied, this does not seem to be the case in ewes, in contrast to cows (Bekana et al., 1994, 1997; Tzora el al., 2002). Examination of further cases, in which the localization of lesions should be compared with the position of the cervical canal, is needed. The distribution of lesions in the fetus itself and the occurrence of bacteria in di¡erent tissues, as shown by FISH, strongly suggested that the fetus could be infected through the amniotic £uid and haematogenously. Infection of the fetus through the amniotic £uid
caused dermatitis and bronchopneumonia, while haematogenous spread through the umbilical vein caused hepatitis. In a single fetus (1A), systemic spread was followed by localization in the spleen, and fetus 4B had a pyelonephritis of probable haematogenous origin. In the dually infected fetuses (4A,B), lesions were de¢nitely associated with F. necrophorum, while immunolabelled L. monocytogenes bacteria were mainly found on the placental surface in association with bedding materials. The low number of L. monocytogenes and lack of association with in£ammation suggests contamination and post-mortem invasion of the fetus. That post-mortem invasion actually occurred was suggested by the presence of mixed bacterial culture in the liver and abomasal contents. Alternatively, the infection may have been present as a local infection in fetus 4A, from which the placenta was unavailable. The placental lesions in fetus 4B were more severe than in the other cases, and were associated with fetal septicaemia. The leucoencephalomalacia present in fetus 4A was probably of a secondary nature and not speci¢cally due to F. necrophorum. The lesion probably re£ected a placental insu⁄ciency, causing cerebral ischaemia and subsequent necrosis, as reported in association with ovine fetal toxoplasmosis (Hartley and Kater,1963).
References Agerholm, J. S., Aalbk, B., Fog-Larsen, A. M., Boye, M., Holm, E., Jensen, T. K., Lindhardt, T., Larsen, L. E. and Buxton, D. (2006).Veterinary and medical aspects of abortion in Danish sheep. Acta Pathologica, Microbiologica et Immunologica Scandinavica, 114,146^152. Alm, E. W., Oerther, D. W., Larsen, N., Stahl, D. A. and Raskin, L. (1996). The oligonucleotide probe database. Applied Environmental Microbiology, 62, 3557^3559. Amann, R. I., Binder, J., Olson, R. J., Chisholm, S. W., Devereux, R. and Stahl, D. A. (1990). Combination of 16S rRNA-targeted oligonucleotide probes with £ow cytometry for analyzing mixed microbial populations. Applied Environmental Microbiology, 56,1919^1925. Barker, I. K. and Van Dreumel, A. A. (1985). The alimentary system. In: Pathology of Domestic Animals, vol. 2, K.V.F. Jubb, P.C. Kennedy and N. Palmer, Eds, Academic Press, San Diego, p. 15. Bekana, M., Jonsson, P., Ekman, T. and Kindahl, H. (1994). Intrauterine bacterial ¢ndings in postpartum cows with retained fetal membranes. Journal of Veterinary Medicine Series A, 41, 663^670. Bekana, M., Jonsson, P. and Kindahl, H. (1997). Bacterial isolates associated with retained fetal membranes and subsequent ovarian activity in cattle. Veterinary Record, 140, 232^234. Boye, M., Aalbk, B. and Agerholm, J. S. (2006). Fusobacterium necrophorum determined as abortifacient in sheep by laser capture microdissection and £uorescence
ARTICLE IN PRESS 221
Ovine Fetal Necrobacillosis
in situ hybridization. Molecular and Cellular Probes, 20, 330^336. Hartley, W. J. and Kater, J. C. (1963). The pathology of Toxoplasma infection in the pregnant ewe. Research inVeterinary Science, 4, 326^332. Jubb, K. V. F., Kennedy, P. C. and Palmer, H. (1985). The female genital system. In: Pathology of Domestic Animals, vol. 3, K.V.F. Jubb, P.C. Kennedy and N. Palmer, Eds, Academic Press, San Diego, pp. 339^359. Kelly, W. R. (1985). The liver and biliary system. In: Pathology of Domestic Animals, vol. 2, K.V.F. Jubb, P.C. Kennedy and N. Palmer, Eds, Academic Press, San Diego, p. 278. Kirkbride, C. A., Gates, C. E. and Libal, M. C. (1989). Ovine and bovine abortion associated with Fusobacterium nucleatum. Journal of Veterinary Diagnostic Investigation, 1, 272^273. Langworth, B. F. (1977). Fusobacterium necrophorum: its characteristics and role as an animal pathogen. Bacteriological Reviews, 41, 373^390. McCourtie, J., Poxton, I. R., Brown, R., Whittaker, C. R., Spence, J. A. and Aitchison, G. U. (1990). A longitudinal study of the cultivable subgingival anaerobic bacteria isolated from sheep during the development of broken mouth periodontitis. Journal of Medical Microbiology, 31, 275^283. Nagaraja, T. G. and Chengappa, M. M. (1998). Liver abscess in feedlot cattle: a review. JapaneseJournal of Animal Science, 76, 287^298. Otter, A. (1996). Fusobacterium necrophorum abortion in a cow. Veterinary Record, 139, 318^319.
Parsonson, I. M., Egerton, J. R. and Roberts, D. S. (1967). Ovine interdigital dermatitis. Journal of Comparative Pathology, 77, 309^313. Pattnaik, N., Mohanty, B. N., Ray II., S. K., Mohanty, D. N. and Mishra, P. R. (1994). Isolation of Fusobacterium necrophorum from bovine abortion. Indian Journal of Animal Sciences, 64, 355^357. Ramos-Vara, J. A., Rook, J., Scanlan, C. M., Mugli, F. and Yamini, B. (1997). Fusobacterium necrophorum septicemia in a lamb: pathological and microbiologic characterization. Journal of Veterinary Diagnostic Investigation, 9,79^82. Scanlan, C. M. and Edwards, J. F. (1990). Bacteriologic and pathologic studies of hepatic lesions in sheep. American Journal of Veterinary Research, 51, 363^366. Tzora, A., Leontides, L. S., Amiridis, G. S., Manos, G. and Fthenakis, G. C. (2002). Bacteriological and epidemiological ¢ndings during examination of the uterine content of ewes with retention of fetal membranes. Theriogenology, 57,1809^1817. Weisburg,W. G., Barns, S. M., Pelletier, D. A. and Lane, D. J. (1991). 16S ribosomal DNA ampli¢cation for phylogenetic study. Journal of Bacteriology, 173, 697^703. Yager, J. A. and Scott, D.W. (1985).The skin and appendages. In: Pathology of Domestic Animals, vol. 1, K.V.F. Jubb, P.C. Kennedy and N. Palmer, Eds, Academic Press, San Diego, pp. 480^483.
Received, May 2nd, 2006 Accepted, January 31st, 2007
www.elsevier.com/locate/jcpa
Ovine Fetal Necrobacillosis J. S. Agerholm, M. Boye* and B. Aalbæk Department of Veterinary Pathobiology, Faculty of Life Sciences, University of Copenhagen, Copenhagen, and *NationalVeterinary Institute,Technical University of Denmark, Copenhagen, Denmark
Summary The association of Fusobacterium necrophorum with late term abortion in sheep is reported. The bacterium was not culturable, but was identi¢ed in ¢ve cases by £uorescence in-situ hybridization (FISH) with an oligonucleotide probe speci¢cally targeting16S rRNA in F. necrophorum. Gross lesions were found in several tissues. Histologically, placental lesions were characterized by locally di¡use in¢ltration of neutrophils, closely associated with abundant small Gram-negative and FISH-positive rods, thrombosis and necrosis. Lesions in the fetal^maternal interface were multifocal and consisted of villous necrosis and suppurative in£ammation. Spread to the fetus from the placenta appeared to occur in two ways. Some fetuses had multifocal necrotizing hepatitis consistent with haematogenous spread through the umbilical vein; further dissemination to other organs occurred. Transplacental spread and infection of the fetus through the amniotic £uid was characterized by development of multifocal suppurative dermatitis and suppurative bronchopneumonia. Localization of FISH-positive bacteria in necrotic lesions was restricted to the periphery. F. necrophorum would seem to have been unrecognized previously as a cause of abortion. The value of culture-independent diagnostic methods is emphasized. r 2007 Elsevier Ltd. All rights reserved. Keywords: Fusobacterium necrophorum; abortion; necrobacillosis; sheep
Introduction Fusobacterium necrophorum, a Gram-negative, strictly anaerobic pleomorphicbacterium, is an opportunistic pathogen in several animal species (Langworth,1977). In ruminants, the bacterium may act as a principal pathogen or as an important constituent in mixed bacterial infections.The most signi¢cant diseases associated with F. necrophorum in ruminants are the ‘‘rumenitis-liver abscess complex’’ (Scanlan and Edwards, 1990; Nagaraja and Chengappa, 1998), necrobacillosis of the foot (Parsonson et al., 1967; Yager and Scott, 1985), and necrotizing stomatitis (Barker and Van Dreumel, 1985; McCourtie et al., 1990). In neonatal ruminants, necrotizing hepatitis may be the result of ascending infection from omphalophlebitis (Kelly,1985). Abortion due to F. necrophorum was reported in a bovine fetus showing widespread serosal in£ammation (Otter, 1996); other reported cases (Pattnaik et al., 1994), have
lacked histopathological con¢rmation. Isolation of the bacterium from aborted fetuses is not su⁄cient to establish a diagnosis, as F. necrophorum is a common invader of the bovine uterus post partum (Bekana et al.,1994,1997). Opening of the cervical canal before abortion may lead to colonization of the fetus and its membranes, or these structures may become contaminated during passage through the caudal parts of the genital tract. Consequently, diagnosis must be based on histopathological recognition of in£ammation in association with F. necrophorum. Kirkbride et al. (1989) reported ovine abortion due to Fusobacterium nucleatum, but fetal infection with F. necrophorum in sheep has not been reported. The present report describes ovine fetuses spontaneously infected with F. necrophorum.
Materials and Methods Fetuses Infected with F: necrophorum
Correspondence to: J.S. Agerholm, Department of Veterinary Pathobiology, Faculty of Life Sciences, University of Copenhagen, Ridebanevej 3, DK-1870 Frederiksberg C, Denmark (e-mail: [email protected]). 0021-9975/$ - see front matter
doi:10.1016/j.jcpa.2007.01.012
The animals were obtained from a larger study on the causes of prenatal mortality in Danish sheep (Agerholm r 2007 Elsevier Ltd. All rights reserved.
ARTICLE IN PRESS 214
J.S. Agerholm et al.
et al., 2006). Initially, six fetuses consisting of three pairs of twins (cases 1A,B ^ 3A,B) with an unidenti¢ed bacterial infection were selected, forming a group of cases with similar in£ammatory lesions closely associated with microcolonies of Gram-negative bacteria. These bacteria were identi¢ed as F. necrophorum by laser capture microdissection of bacterial colonies from formalin-¢xed tissue, followed by polymerase chain reaction (PCR) ampli¢cation of bacterial 16S rDNA and sequencing. An rRNA-targeting oligonucleotide probe speci¢c for F. necrophorum was designed and used in a £uorescent in-situ hybridization (FISH) assay (see below). This examination identi¢ed a further pair of F. necrophorum-infected twin fetuses (cases 4A,B) and an infected placental specimen from a sheep (case 5) that aborted in late gestation. Placental material was not available from cases 2A, 3A and 4A. Pathology
The fetal age was calculated from breeding data and compared with the stage of development. The fetus and corresponding fetal membranes were examined post mortem and tissues including placenta, lung, myocardium, liver, kidney, skeletal muscle, brain and cervical spinal cord were sampled for histopathology. Multiple areas were sampled from placenta, lung, liver, and brain and other tissues were taken if showing gross lesions. The specimens were ¢xed in 10% neutral buffered formalin, dehydrated through graded alcohols, embedded in para⁄n wax and sectioned at 7 mm (brain) or 3 mm (other tissues). Initial histological examination was performed on haematoxylin and eosin (HE)-stained sections. Additional staining methods were applied when considered appropriate. These included Mallory’s phosphotungstic acid haematoxylin method (PTAH), Masson’s trichrome method, the periodic acid-Schi¡ (PAS) technique, and the BrownHopps method for Gram-positive and Gram-negative bacteria. Immunohistochemistry
Immunohistochemical examination for Listeria monocytogenes antigen was performed on all tissues of any pair of twins from which this organism was cultured. Selected sections from the other cases were also examined. A commercially available polyclonal antibody against L. monocytogenes types 1 and 4 (Difco Laboratories, Detroit, MI, USA) was used in a streptavidinbiotin immunoperoxidase assay. Bacteriology
This was performed on samples of lung, liver, fetal placenta and abomasal contents, as previously described
(Agerholm et al., 2006). Brie£y, each sample was spread on two blood agar plates (Blood Agar Base CM0055; Oxoid, Basingstoke, UK, containing sterile bovine blood 5%) of which one was incubated aerobically and the other anaerobically for 48 h at 37 1C. Colonies were subcultured and identi¢ed by standard methods. Selected isolates were characterized by the API-20-E kit (BioMeŁrieux, Marcy l’Etoile, France) according to the manufacturer’s recommendations. Examination for Campylobacter spp. was performed as previously described (Agerholm et al., 2006). In-situ Hybridization
This was performed with the general bacterial probe eub338: S-D-Bact-0338-a-A-18 50 -GCTGCCTCCCGTAGGAGT-30 (Amann et al., 1990) and a probe speci¢c for F. necrophorum (F.necrophorum183: S-SF.necrophorum-0183-a-A-18, 50 -GATTCCTCCATGCGAAAA-30 [Boye et al., 2006]). Epifluorescence Microscopy and Scanning
An Axio Imager M1 epi£uorescence microscope (Carl Zeiss, Oberkochen, Germany) equipped for epi£uorescence with a 100-W HBO lamp and ¢lter set XF53, (Omega Optical, Brattleboro, VT, USA) was applied for simultaneous detection of red and green £uorescence. Filter sets 43 HE and 38 HE (Carl Zeiss) were used to ‘‘visualize’’ CY3 and £uorescein, respectively. Images were obtained with an AxioCam MRm version 3 FireWire monochrome camera (Roper Scienti¢c, Ottobrunn, Germany) and the software AxioVision version 4. Entire tissue sections were examined by scanning £uorescence in-situ hybridization using an ArrayWoRxe scanner (pixel resolution of 3.25 mm) (Applied Precision, Issaquah,WA, USA). PCR and Sequencing
DNAwas extracted from dermal lesions of fetus1A kept at 20 1C until used. Subcutaneous tissue was removed and material scraped with a scalpel from the dermis was suspended in 1ml of 0.9% NaCl. DNA was extracted with the Easy-DNA kit (InVitrogen, Paisley, UK) and ‘‘protocol #3,’’according to the manufacturer’s instructions. Primers (MWG-BIOTECH AG, Ebersberg, Germany) used for 16S rRNA gene ampli¢cation consisted of S-D-Bact-0008-a-S-20 (50 -AGAGTTTGATCMTGGCTCAG-30) and S-Univ-*-1492-a-A-21 (50 -GTTACCTTGTTACGACTTCAG-30), modi¢ed fromWeisburg et al. (1991). Oligonucleotide probes were designated according to Alm et al. (1996). Reaction conditions were as follows:1 PCR Bu¡er II (Roche Applied Science, Indianapolis, IN, USA), 3 mM MgCl2,
ARTICLE IN PRESS Ovine Fetal Necrobacillosis
2.5 mM of each deoxynucleotide triphosphate (Amersham Biosciences, Piscataway, NJ, USA), 0.4 mM of each primer, 0.5 U of AmpliTaq DNA Polymerase (Roche Applied Science) and 2.0 ml of template DNA in a total volume of 50 ml. Thermocycling with a T3 Thermocycler (BioMetra, Goettingen, Germany) was performed as follows: denaturation at 94 1C for 3 min followed by 35 cycles of denaturation at 94 1C for 1min, annealing at 55 1C for 1min, extension at 72 1C for 1ˆ˜ min, and a ¢nal extension at 72 1C for 10 min. Ampli¢ed PCR products were visualized by electrophoresis in 1.5% agarose gels. The amplicons were puri¢ed with the QIAquick spin PCR puri¢cation kit (Qiagen, Hilden, Germany) and sequenced by cycle sequencing on an ABI PRISM 377 DNA Sequencer (PE Biosystems, Foster City, CA, USA) with the ABI PRISM Big Dye Terminator Cycle Sequencing Kit (PE Biosystems), according to the manufacturer’s instructions and with the following primers: S-D-Bact-0008-a-S-20, S -*-Univ-0519-a-A-18 (50 GTATTACCGCGGCTGCTG-30), S-D-Bact-0786-a-A-21 (50 -GACTACCNGGGTATCTAATCC-30), S-D-Bact-1054-a-A-20 (50 -ACGAGCTGACGACRRCCATG’-30) and S-*Univ-1492-a-A-21. The sequence was compared with published 16S rDNA sequences in GenBank (www.ncbi.nlm.nih. gov/) by means of a Blast search for homology. The near-full length 16S rDNA sequence has been deposited in the GenBank database under accession number DQ486127. Other Laboratory Examinations
The animals were examined for other infectious organisms including Toxoplasma gondii, Border disease virus (BDV), Coxiella burnetii and Chlamydophila abortus and tested for antibodies againstT. gondii and BDV in pleural e¡usions (Agerholm et al., 2006).
215
part of a mixed culture. A few L. monocytogenes bacteria were identi¢ed by immunohistochemistry in the liver of case 4A and on the placental surface of case 4B. F. necrophorum was identi¢ed in close association with lesions in most cases (1A,B; 2A,B; 3A,B; 4B; 5) by means of the FISH as previously described (Boye et al., 2006). Additionally, 16S rDNA was isolated, PCR-ampli¢ed, and sequenced from frozen skin of fetus 1A; a BLAST search revealed 99% nucleotide homology with both the type strain for F. necrophorum (ATCC 25286) and a ruminant F. necrophorum strain isolated in New Zealand (GenBank Acc. No. AF044948). No other pathogens and no fetal antibodies against T. gondii or BDV were found. Gross Lesions
Placentitis, seen in four cases (1A,B; 2B; 4B), was characterized by slight necrosis of cotyledons and intercotyledonary areas, oedema, haemorrhage and variable amounts of purulent exudate (Fig. 1). Fetus 1A had extensive skin lesions, spread over the entire body and consisting of multifocal hyperaemic elevated nodules, with a diameter of p 5 mm and occasional central ulceration (Fig. 2). Pulmonary lesions were observed in four fetuses, either as pale disseminated pinpoint lesions (4B) (Fig. 3), or as widespread consolidation of the lung, sometimes with elevated hyperaemic lobules (2A,B; 3A). Hepatomegaly was observed in six fetuses. In addition, two of these (1A; 2A) had focal areas of necrosis in the liver, either as separate lesions or as multiple lesions grouped in severely autolysed and emphysematous areas (Fig. 4). Super¢cial lesions were covered by perihepatitis. Multiple necrotic areas (diameter p 5 mm) were found in the spleen of case 1A. Only in case 4B were in£ammatory lesions absent; this
Results Clinical Signs
The abortions occurred in four herds during the ¢nal month of gestation. One sheep had a depressed appetite before abortion, but the others were clinically normal. One lamb (4B) was born weak and died within 24 h. Agent Identification
Except for case 4B, lung, liver and abomasal content were found to be bacteriologically sterile by cultural examination, while mixed bacterial cultures were obtained from the placentas. In fetus 4B, L. monocytogenes was isolated from the liver and abomasal content as
Fig. 1. Local purulent placentitis with a single in£amed cotyledon (arrowhead) and a purulent exudate in the intercotyledonal placenta (arrow) due to F. necrophorum. Ovine placenta, case 2B.
ARTICLE IN PRESS 216
J.S. Agerholm et al.
fetus had leptomeningeal haemorrhage in the caudal cerebrum, softening of the cerebral white matter, and multifocal pinpoint haemorrhages in the lung.
Fig. 2. Multifocal in£amed nodules in the skin of an ovine fetus due to F. necrophorum. Case 1A.
Histopathology
In mono-infected fetuses, lesions in the cotyledonary villi were mild and consisted of focal necrosis and mild neutrophil in¢ltration. However, in the deep stroma of the cotyledons and of the intercotyledonary placenta, the tissue had severe surface-orientated in£ammation, characterized by extensive, locally di¡use, neutrophil in¢ltration, closely associated with numerous small Gram-negative rods, necrosis of the connective tissue and amnionic/allantoic surface, oedema, and acute haemorrhage (Fig. 5).The walls of arteries and veins located in in£amed tissue showed peripheral necrosis, neutrophil in¢ltration, and sometimes transmural in£ammation and developing thrombosis (Fig. 6). Bacterial microcolonies occasionally ¢lled the lumen of non-in£amed thin-walled vessels. Bronchopneumonia, found in all fetuses, was characterized by widespread, intense hyperaemia, haemorrhages, and accumulations of neutrophils and debris in bronchi, bronchioles and alveoli. The distribution of bacteria varied from single organisms to multiple mi-
Fig. 3. Disseminated lesions of pinpoint pulmonary necrosis (arrowheads) due to F. necrophorum septicaemia. Lung, ovine fetus, case 4B.
ARTICLE IN PRESS Ovine Fetal Necrobacillosis
217
Fig. 4. Two major hepatic lesions consisting of multiple necrotic areas are present in autolysed and emphysematous areas. Additionally, pinpoint necrotic lesions are disseminated in the liver. F. necrophorum septicaemia, ovine fetus, case 1A.
Fig. 5. Intercotyledonary placenta with di¡use purulent in£ammation (I) and oedema (O) in the connective tissue and super¢cially located colonies of F. necrophorum (arrowhead). Ovine placenta, case 1A. HE. Bar, 50 mm.
Fig. 6. Large numbers of F. necrophorum (arrow) surrounding an in£amed artery undergoing thrombosis. Ovine placenta, case 4B. HE. Bar,100 mm.
ARTICLE IN PRESS 218
J.S. Agerholm et al.
Fig. 7. Large numbers of F. necrophorum (arrowheads) located in the centre (C) and in the periphery (P) of a large chronic hepatic lesion. Note the intense bacterial staining in the periphery, in contrast to weak staining of bacteria in the central area. Ovine fetus, case 2A. HE. Bar,100 mm.
Fig. 8. Acute focal hepatic lesion with a central area occupied by large numbers of F. necrophorum surrounded by purulent in£ammation. Ovine fetus, case 2A. HE. Bar,100 mm.
crocolonies, plugging and distending the alveoli. As compared with the bacteria in the placenta, those in the lungs were longer, and microcolonies occasionally contained intermingled, ¢lamentous structures. In several cases, large bacterial colonies were not associated with marked in£ammation, suggesting continued bacterial growth after fetal death. In all cases lesions tended to be widespread, a¡ecting both cranial and diaphragmatic lobes. Multiple necrotic areas with large numbers of intralesional Gram-negative bacteria and peripheral neutrophil in¢ltration were present in the liver of cases 1A and 2A. Peripherally located bacteria showed intense
staining, but staining decreased in intensity towards the centre of major lesions, probably due to bacterial death (Fig. 7). Major areas of necrosis were partly surrounded by ¢brosis, but bacterial invasion of the surrounding parenchyma and corresponding neutrophil in¢ltration were seen locally. Adjacent to major lesions, multiple acute foci of ¢brinous and suppurative in£ammation surrounding bacterial colonies were present (Fig. 8). The spleen of case 1A showed multifocal areas of acute necrosis, with large bacterial colonies, central debris, and peripheral in£ammation. In the skin of case 1A, multiple areas of severe suppurative in£ammation extending into the deep layer of the dermis and occasionally into the subcutis were present. The epithelium at the periphery of each lesion showed hydropic degeneration and necrosis. Large numbers of Gram-negative rods were present at the margin of the lesions. Twin fetuses often di¡ered from each other in respect of the severity of lesions and degree of autolysis. Severe lesions were usually associated with large numbers of bacteria and pronounced autolysis. The lesions in the fetuses infected with both F. necrophorum and L. monocytogenes (cases 4A,B) di¡ered from the general ¢ndings. Extensive necrosis of the stroma with oedema and suppurative in£ammation was present in the placenta of fetus 4B. In necrotic areas, veins and small- and medium-sized arteries showed transmural necrosis and thrombosis, while large arteries showed only peripheral necrosis and in£ammation. In the periphery of necrotic areas, multiple colonies of Gram-negative rods were present. In some areas lesions pierced the placenta. Lesions in the interface were multifocal, consisting of villous necrosis and suppurative in£ammation, occasionally with extension into the underlying connective tissue. Abundant Gram-negative bacteria were seen in close association with the lesions and also in the cytoplasm of trophoblasts. Immunolabelling did not reveal L. monocytogenes in such areas; these organisms were seen exclusively on the cotyledonal surface in association with bedding debris and a mixed bacterial and fungal population. In the lung, disseminated foci were present. Minor lesions consisted of accumulations of Gram-negative bacteria interspersed with neutrophils, while larger lesions had a central area of necrosis, surrounded by large numbers of bacteria spreading through the alveoli, and peripheral suppurative in£ammation (Fig. 9). In some areas suppurative bronchio-alveolitis with bacteria in the bronchi was seen. Hepatic lesions were mild and dominated by interstitial lymphocytic in¢ltration. Colonies of Gram-negative bacteria were seen in the sinusoids, without associated in£ammation. Additionally, multifocal necrotizing pyelonephritis with numerous bacteria in the renal pelvis was observed.
ARTICLE IN PRESS 219
Ovine Fetal Necrobacillosis
Fig. 9. Acute focal pneumonia with numerous F. necrophorum and debris surrounded by a purulent exudate. Ovine fetus, case 4B. HE. Bar, 50 mm.
Fig. 11. Margin of a chronic hepatic lesion showing intense peripheral orange £uorescence (P) and weaker orange £uorescence towards the centre (C). Ovine fetus, case 4B. FISH. Bar, 20 mm.
periphery of large necrotizing lesions, while £uorescence like Gram-staining (see above), was absent in the central areas (Figs 10 and 11). This was probably due to bacterial death, as the signal from the rRNA-targeting probe is dependent on the metabolic activity of the cells. Except for case 1A, F. necrophorum was present only in close association with lesions, and was absent in tissues without lesions. In case 1A, F. necrophorum was seen in association with lesions in placenta, skin, liver, lung and spleen. However, occasional bacteria not associated with lesions were seen in the kidney, skeletal muscle and central nervous system, suggesting terminal haematogenous spread or growth post mortem.
Fig.10. Demonstration of metabolically active F. necrophorum bacteria (orange) in the periphery of a large chronic hepatic lesion. A few associated acute lesions are seen as foci of intense orange £uorescence. Ovine fetus, case 2A. FISH. Bar, 2.5 mm.
In fetus 4A, the most prominent lesion was an extensive subacute cerebral leucoencephalomalacia with massive in¢ltration of macrophages and meningeal haemorrhage. Other tissues showed acute congestion, but the placenta was not available for examination. A few immunolabelled L. monocytogenes-like structures were seen in the liver, without adjacent in£ammation. Epifluorescence Microcopy
The in-situ hybridization demonstrated a close association of F. necrophorum with lesions. An intralesional FISH signal was seen in areas corresponding to the location of the Gram-negative bacteria as seen in light microscopy. An intense FISH signal was present in the
Discussion The study demonstrated that F. necrophorum infection in ovine fetuses was associated with severe in£ammation of multiple tissues. The close association of the bacterium with pyonecrotizing lesions strongly indicated that F. necrophorum was the cause of fetal or neonatal death. The sequencing of 16S rDNA obtained from intralesional bacterial colonies by laser capture microdissection (Boye et al., 2006) and by sequencing of 16S rDNA extracted from the in£amed skin of case 1A, showed a 99% nucleotide homology with F. necrophorum. The fetuses were obtained from a study on the causes of prenatal mortality in Danish sheep, which demonstrated that F. necrophorum frequently caused congenital infection (Agerholm et al., 2006). That such infections have not been reported previously may re£ect di⁄culties in bacteriological culture or exclusion of anaerobic culturing in routine diagnostic bacteriology. Isolation of the bacterium in the present study was unsuccessful, even though a method that regularly identi¢es
ARTICLE IN PRESS 220
J.S. Agerholm et al.
F. necrophorum in pathological specimens (Aalbk, personal observation) was used. As multiple specimens containing large numbers of bacteria were examined, the unsuccessful attempts to isolate the bacterium probably re£ected decreased viability upon arrival at the laboratory.This was not due to antibiotic treatment, as the sheep were not medicated before abortion. In aborted fetuses showing pyonecrotizing in£ammation associated with Gram-negative rods, the use of diagnostic methods other than bacteriological culture, e.g., FISH, would seem advisable. The localization of the most severe lesions in the deep stroma of the fetal placenta may indicate placental infection other than by maternal bacteraemia, which is the generally assumed route of bacterial infection in the pregnant uterus of ruminants (Jubb et al., 1985). F. necrophorum bacteraemia in sheep is known to occur following bacterial penetration of the ruminal epithelium, which causes a local portal bacteraemia with subsequent hepatitis (‘‘rumenitis-liver abscess complex’’) (Nagaraja and Chengappa, 1998), but systemic bacteraemia may also occur (Ramos-Vara et al., 1997). Initial placentomal localization is expected to occur in the interdigitations created by villi and crypts. In aborted placentas this is seen as in£ammation of the villous portion of the cotyledons; later the in£ammation may penetrate into the underlying stroma. In this study, cotyledonary changes suggested initial lesion development in the deep connective tissue, with subsequent expansion and eruption to the placentomal interface, causing localized in£ammation and necrosis of villi. It could not be determined whether the necrosis of the amniotic or allantoic surface occurred as a result of an expanding lesion or from bacteria present in the £uids. Equally it might be argued that the fetal maternal interface may have an oxygen-rich environment unfavourable for strict anaerobes as F. necrophorum, or that representative specimens were not examined. Alternatively, it is possible that an ascending transcervical infection caused a dissecting in£ammation in the fetal placenta and a transplacental spread to the amniotic and allantoic £uids, causing widespread surface in£ammation. A prequisite for this hypothesis is the presence of F. necrophorum on the mucosal membranes of the ovine genital tract. Although not thoroughly studied, this does not seem to be the case in ewes, in contrast to cows (Bekana et al., 1994, 1997; Tzora el al., 2002). Examination of further cases, in which the localization of lesions should be compared with the position of the cervical canal, is needed. The distribution of lesions in the fetus itself and the occurrence of bacteria in di¡erent tissues, as shown by FISH, strongly suggested that the fetus could be infected through the amniotic £uid and haematogenously. Infection of the fetus through the amniotic £uid
caused dermatitis and bronchopneumonia, while haematogenous spread through the umbilical vein caused hepatitis. In a single fetus (1A), systemic spread was followed by localization in the spleen, and fetus 4B had a pyelonephritis of probable haematogenous origin. In the dually infected fetuses (4A,B), lesions were de¢nitely associated with F. necrophorum, while immunolabelled L. monocytogenes bacteria were mainly found on the placental surface in association with bedding materials. The low number of L. monocytogenes and lack of association with in£ammation suggests contamination and post-mortem invasion of the fetus. That post-mortem invasion actually occurred was suggested by the presence of mixed bacterial culture in the liver and abomasal contents. Alternatively, the infection may have been present as a local infection in fetus 4A, from which the placenta was unavailable. The placental lesions in fetus 4B were more severe than in the other cases, and were associated with fetal septicaemia. The leucoencephalomalacia present in fetus 4A was probably of a secondary nature and not speci¢cally due to F. necrophorum. The lesion probably re£ected a placental insu⁄ciency, causing cerebral ischaemia and subsequent necrosis, as reported in association with ovine fetal toxoplasmosis (Hartley and Kater,1963).
References Agerholm, J. S., Aalbk, B., Fog-Larsen, A. M., Boye, M., Holm, E., Jensen, T. K., Lindhardt, T., Larsen, L. E. and Buxton, D. (2006).Veterinary and medical aspects of abortion in Danish sheep. Acta Pathologica, Microbiologica et Immunologica Scandinavica, 114,146^152. Alm, E. W., Oerther, D. W., Larsen, N., Stahl, D. A. and Raskin, L. (1996). The oligonucleotide probe database. Applied Environmental Microbiology, 62, 3557^3559. Amann, R. I., Binder, J., Olson, R. J., Chisholm, S. W., Devereux, R. and Stahl, D. A. (1990). Combination of 16S rRNA-targeted oligonucleotide probes with £ow cytometry for analyzing mixed microbial populations. Applied Environmental Microbiology, 56,1919^1925. Barker, I. K. and Van Dreumel, A. A. (1985). The alimentary system. In: Pathology of Domestic Animals, vol. 2, K.V.F. Jubb, P.C. Kennedy and N. Palmer, Eds, Academic Press, San Diego, p. 15. Bekana, M., Jonsson, P., Ekman, T. and Kindahl, H. (1994). Intrauterine bacterial ¢ndings in postpartum cows with retained fetal membranes. Journal of Veterinary Medicine Series A, 41, 663^670. Bekana, M., Jonsson, P. and Kindahl, H. (1997). Bacterial isolates associated with retained fetal membranes and subsequent ovarian activity in cattle. Veterinary Record, 140, 232^234. Boye, M., Aalbk, B. and Agerholm, J. S. (2006). Fusobacterium necrophorum determined as abortifacient in sheep by laser capture microdissection and £uorescence
ARTICLE IN PRESS 221
Ovine Fetal Necrobacillosis
in situ hybridization. Molecular and Cellular Probes, 20, 330^336. Hartley, W. J. and Kater, J. C. (1963). The pathology of Toxoplasma infection in the pregnant ewe. Research inVeterinary Science, 4, 326^332. Jubb, K. V. F., Kennedy, P. C. and Palmer, H. (1985). The female genital system. In: Pathology of Domestic Animals, vol. 3, K.V.F. Jubb, P.C. Kennedy and N. Palmer, Eds, Academic Press, San Diego, pp. 339^359. Kelly, W. R. (1985). The liver and biliary system. In: Pathology of Domestic Animals, vol. 2, K.V.F. Jubb, P.C. Kennedy and N. Palmer, Eds, Academic Press, San Diego, p. 278. Kirkbride, C. A., Gates, C. E. and Libal, M. C. (1989). Ovine and bovine abortion associated with Fusobacterium nucleatum. Journal of Veterinary Diagnostic Investigation, 1, 272^273. Langworth, B. F. (1977). Fusobacterium necrophorum: its characteristics and role as an animal pathogen. Bacteriological Reviews, 41, 373^390. McCourtie, J., Poxton, I. R., Brown, R., Whittaker, C. R., Spence, J. A. and Aitchison, G. U. (1990). A longitudinal study of the cultivable subgingival anaerobic bacteria isolated from sheep during the development of broken mouth periodontitis. Journal of Medical Microbiology, 31, 275^283. Nagaraja, T. G. and Chengappa, M. M. (1998). Liver abscess in feedlot cattle: a review. JapaneseJournal of Animal Science, 76, 287^298. Otter, A. (1996). Fusobacterium necrophorum abortion in a cow. Veterinary Record, 139, 318^319.
Parsonson, I. M., Egerton, J. R. and Roberts, D. S. (1967). Ovine interdigital dermatitis. Journal of Comparative Pathology, 77, 309^313. Pattnaik, N., Mohanty, B. N., Ray II., S. K., Mohanty, D. N. and Mishra, P. R. (1994). Isolation of Fusobacterium necrophorum from bovine abortion. Indian Journal of Animal Sciences, 64, 355^357. Ramos-Vara, J. A., Rook, J., Scanlan, C. M., Mugli, F. and Yamini, B. (1997). Fusobacterium necrophorum septicemia in a lamb: pathological and microbiologic characterization. Journal of Veterinary Diagnostic Investigation, 9,79^82. Scanlan, C. M. and Edwards, J. F. (1990). Bacteriologic and pathologic studies of hepatic lesions in sheep. American Journal of Veterinary Research, 51, 363^366. Tzora, A., Leontides, L. S., Amiridis, G. S., Manos, G. and Fthenakis, G. C. (2002). Bacteriological and epidemiological ¢ndings during examination of the uterine content of ewes with retention of fetal membranes. Theriogenology, 57,1809^1817. Weisburg,W. G., Barns, S. M., Pelletier, D. A. and Lane, D. J. (1991). 16S ribosomal DNA ampli¢cation for phylogenetic study. Journal of Bacteriology, 173, 697^703. Yager, J. A. and Scott, D.W. (1985).The skin and appendages. In: Pathology of Domestic Animals, vol. 1, K.V.F. Jubb, P.C. Kennedy and N. Palmer, Eds, Academic Press, San Diego, pp. 480^483.
Received, May 2nd, 2006 Accepted, January 31st, 2007