34 2006;70:34–42 The Canadian Journal of Veterinary Research 2006;70:000–000
In the mid 1950s, gram-negative organisms identified as
Haemophilus agni were isolated from all affected tissues of feeder
lambs that had died in an epidemic of septicemic disease (1). The
disease was characterized by meningitis, ecchymotic hemorrhage
on serous membranes of subcutaneous tissues and muscles, necrotic
foci in the liver, bacterial thrombosis, and necrotizing vasculitis.
In 1960, Kennedy et al (2) recognized a similar Haemophilus-like
organism as the causative agent of infectious meningoencephalitis in
feedlot cattle. The agent, identified later as H. somnus, has since been
incriminated as the etiologic agent of additional disease conditions in
Haemophilus somnus (Histophilus somni) in bighorn sheep
Alton C.S. Ward, Glen C. Weiser, Bruce C. Anderson, Patrick J. Cummings,
Karen F. Arnold, Lynette B. Corbeil
Respiratory disease and poor lamb recruitment have been identified as limiting factors for bighorn-sheep populations. Haemophilus
somnus (recently reclassified as Histophilus somni) is associated with respiratory disease in American bison, domestic sheep, and
cattle. It is also harbored in their reproductive tracts and has been associated with reproductive failure in domestic sheep and
cattle. Therefore, reproductive tract and lung samples from bighorn sheep were evaluated for the presence of this organism.
Organisms identified as H. somnus were isolated from 6 of 62 vaginal but none of 12 preputial swab samples. Antigen specific to
H. somnus was detected by immunohistochemical study in 4 of 12 formalin-fixed lung tissue samples of bighorn sheep that died
with evidence of pneumonia. Notably, H. somnus was found in alveolar debris in areas of inflammation. The 6 vaginal isolates
and 2 H. somnus isolates previously cultured from pneumonic lungs of bighorn sheep were compared with 3 representative
isolates from domestic sheep and 2 from cattle. The profiles of major outer membrane proteins and antigens for all of the isolates
were predominantly similar, although differences that may be associated with the host–parasite relationship and virulence
were detected. The DNA restriction fragment length profiles of the bighorn-sheep isolates had similarities not shared with the
other isolates, suggesting distinct phylogenetic lines. All of the isolates had similar antimicrobial profiles, but the isolates from
the bighorn sheep produced less pigment than those from the domestic livestock, and growth of the former was not enhanced
by CO2. Wildlife biologists and diagnosticians should be aware of the potential of these organisms to cause disease in bighorn
sheep and of growth characteristics that may hinder laboratory detection.
Les maladies respiratoires et des d’approvisionnement en agneaux ont été identifiées comme étant les facteurs limitants au développement des
populations de mouflons. Haemophilus somnus (reclassifié récemment sous le nom Histophilus somni) est associé aux maladies respira-
toires chez les bisons des USA, les moutons domestiques et les bovins. Cette bactérie est retrouvée dans le tractus reproducteur et a été associée
avec des problèmes de reproduction chez les moutons et les bovins. Ainsi, des échantillons provenant du système reproducteur et des poumons
de mouflons ont été examinés pour détecter la présence de ce microorganisme. Des bactéries identifiées comme étant H. somnus ont été isolées
de 6 des 62 prélèvements vaginaux mais d’aucun des 12 échantillons de prépuce. Un antigène spécifique à H. somnus a été détecté par exa-
men immuno-histochimique dans 4 des 12 échantillons de tissu pulmonaire fixés dans la formaline provenant de mouflons morts avec des
lésions de pneumonie, notamment dans des débris alvéolaires dans des zones d’inflammation. Les 6 isolats vaginaux et 2 isolats de H. somnus
obtenus au préalable de poumons de mouflons avec pneumonie ont été comparés à 3 isolats représentatifs provenant de moutons et 2 de bovins.
Les profils des protéines de la membrane externe et des antigènes de tous les isolats étaient relativement similaires, bien que des différences
pouvant être associées à la relation hôte-parasite et à la virulence ont été détectées. Les profils des patrons obtenus suite à la digestion par des
enzymes de restriction de l’ADN des isolats provenant des mouflons indiquent que ceux-ci ont des similarités qui ne sont pas partagées par
les autres isolats, suggérant ainsi qu’ils proviennent d’une lignée phylogénétique différente. Tous les isolats avaient le même patron de sensi-
bilité antimicrobienne, mais les isolats provenant des mouflons étaient moins pigmentés que ceux provenant des animaux domestiques, et la
croissance des premiers n’était pas influencée par le CO2. Les biologistes et diagnosticien de la faune devraient être au fait du potentiel patho-
gène de ces microorganismes chez le mouflon et des caractéristiques de croissance qui pourraient nuire à sa détection en laboratoire.
(Traduit par Docteur Serge Messier)
University of Idaho, College of Agriculture, Caine Veterinary Teaching Center, 1020 East Homedale Road, Caldwell, Idaho, 83607-8098, USA
(Ward, Weiser, Anderson); Nevada Department of Wildlife, 4747 West Vegas Drive, Las Vegas, Nevada 89108, USA (Cummings); Department of
Pathology, University of California at San Diego, 200 West Arbor Drive, San Diego, California 92103-8416, USA (Arnold, Corbeil); Department of
Population Health and Reproduction, University of California, Davis, California 95616, USA (Corbeil).
Address all correspondence and reprint requests to Dr. Alton C.S. Ward; telephone: (208) 454-8657; fax: (208) 454-8659;
Received March 3, 2005. Accepted August 11, 2005.
The Canadian Journal of Veterinary Research 34
2000;64:0–00 The Canadian Journal of Veterinary Research 35
cattle, including respiratory disease, abortion, arthritis, septicemia,
laminitis, mastitis, and myocarditis (3,4). It is also a common com-
mensal in the upper respiratory tract, prepuce, and vagina of clini-
cally healthy cattle (3). Similar organisms, identified as Histophilus
ovis, were isolated from sheep in Australia and determined to be
the causative agents of mastitis, epididymitis-orchitis, metritis,
septicemia, meningoencephalitis, pneumonia, and polyarthritis (5).
In 1 study, rams in 22 (27.5%) of 80 herds tested were infected with
H. somnus (6); in addition, 12% of ewes in the infected herds failed
to lamb, in comparison with 6% of ewes in noninfected herds, thus
suggesting reduced fertility of ewes in infected herds.
Biberstein (7) found that organisms in these 3 species share several
characteristics; however, growth of only H. somnus was enhanced
by CO2. Stephens et al (8) found that these organisms had similar
morphology and biochemical reactions, as well as common antigens,
and suggested that they should be considered members of a single
taxon within the family Pasteurellaceae. From DNA hybridization
studies, Walker et al (9) concluded that these organisms should
be considered a single species. In 2003, Angen et al (10) proposed
that H. agni, H. ovis, and H. somnus be regarded as a single spe-
cies, Histophilus somni, based on sequencing of the 16S rRNA and
rpoB genes. Because this proposal is not fully accepted, we are using
herein the more usual name, Haemophilus somnus.
Most of the previous reports have presented information related
to the isolation of H. somnus from sheep and cattle. This organism
has also been isolated from tonsillar and reproductive tract samples
from American bison (11). Furthermore, immunohistochemical tests
have demonstrated its presence in lung tissue from bison that died of
respiratory disease (12). The potential for these organisms to infect
other ruminant species indigenous to North America and to cause
disease has not previously been investigated.
Pneumonic epizootics resulting in deaths of free-ranging bighorn
sheep of all ages and poor lamb recruitment have been identified
as factors associated with the precipitous decline in the numbers of
bighorn sheep during the first half of the 20th century and failure
of existing populations to thrive (13,14). Essentially all bighorn
sheep populations harbor multiple strains of Pasteurellaceae (15),
and Pasteurella or Mannheimia spp. have been incriminated as the
cause of some epizootics of respiratory disease (13,16). However,
there are no reports of H. somnus isolation from bighorn sheep.
Because H. somnus is associated with both respiratory disease and
reproductive failure in domestic sheep and cattle and respiratory
disease in bison, we initiated studies to determine if similar organ-
isms are present in bighorn-sheep populations. In this report, we
compare characteristics of organisms isolated from bighorn sheep
with representative H. somnus isolates from cattle and domestic
sheep and present evidence that H. somnus may be associated with
disease in bighorn sheep.
Materials and methods
Reproductive tract swab samples were collected from 74 desert
bighorn sheep (Ovis canadensis nelsoni) captured in southern Nevada
as part of a health-monitoring program. The 12 males (M) and
62 females (F) were approximately 8 mo to more than 4 y of age.
The dates of sampling, numbers of sheep sampled, and locations of
the herds were as follows: October 1996, 1 M, 13 F, Pancake Range;
October 1996, 1 M, 1 F, River Mountains; January 1998, 4 M, 16 F,
Pancake Range; October 1998, 13 F, Muddy Mountains; October
1998, 1 M, 3 F, Arrow Canyon Range; October 1998, 2 M, 6 F, Specter
Range; and October 1999, 3 M, 10 F, Muddy Mountains. All these
sheep were considered to be clinically healthy, but poor lamb recruit-
ment was noted in the Pancake Range population. Sterile Dacron
swab systems with Amies transport medium (Precision Dynamics
Corporation, Van Nuys, California, USA) were used to collect pre-
putial and vaginal samples. The mucosa was exposed by spreading
the preputial or vulvar orifice. The swab was inserted approximately
5 to 6 cm and rotated to collect the sample.
Additional samples included lung tissue from Rocky Mountain
bighorn sheep (Ovis canadensis canadensis) with postmortem evidence
of pneumonia. Fresh lung tissue from an adult ewe and a 5-mo-old
lamb that died in 1997 and 1999, respectively, were received for
bacterial culture. The ewe had been raised and kept in a captive
population, and the lamb was from a free-ranging population in
Hell’s Canyon, which forms portions of the borders between Idaho,
Oregon, and Washington. Formalin-fixed lung tissue from the
lamb and archived paraffin-impregnated lung tissue blocks from
11 other bighorn sheep that had died during a 1995–96 pneumonic
epizootic in Hell’s Canyon (16) were sent to the Prairie Diagnostic
Service Laboratory, Saskatoon, Saskatchewan, for immunohistologic
Bacterial isolation and characterization
Swab and tissue samples were submitted to the University of
Idaho Caine Veterinary Teaching Center for bacterial culture. Samples
were inoculated onto Columbia blood agar (Becton Dickinson
Microbiology Systems, Cockeysville, Maryland, USA) contain-
ing 5% domestic-sheep blood (CBA) and Columbia blood agar
containing 5% bovine blood plus antibiotics (CBAA) selective for
Pasteurellaceae (17). The media were incubated at 37°C with 10%
added CO2 and examined after 24 and 48 h. Colonies characteristic
of Pasteurellaceae were further tested. Organisms that were gram-
negative, nonmotile, oxidase-positive, and catalase-negative and
failed to grow on MacConkey’s agar (Becton Dickinson Microbiology
Systems) or to ferment carbohydrates in standard biochemical tests
for biogrouping Pasteurella spp. (18) were putatively identified as
H. somnus. These isolates were characterized further in test media
used for biogrouping of H. somnus isolates (19) and compared with
representative H. somnus isolates cultured previously from cattle
and domestic sheep (19,20). All isolates and their sources are listed
in Table I.
Each test isolate and M. haemolytica isolate #43270 of the American
Type Culture Collection were streaked for isolation on brain–heart
infusion (BHI) agar containing 5% ovine blood (OBHIA) and BHI
agar containing 5% bovine blood (BBHIA) for evaluation of hemo-
lytic activity, as previously described (19). The agar plates were
incubated in a candle jar at 37°C for 48 h and at room temperature
(approximately 21°C) for an additional 48 h before evaluation.
Isolated colonies were removed with a sterile inoculation loop,
and the underlying medium was evaluated for evidence of lytic
36 The Canadian Journal of Veterinary Research 2000;64:0–00
activity of the test isolates in comparison with that produced by
M. haemolytica. The isolates were evaluated for CO2 requirement by
comparing their growth on separate CBA plates incubated at 37°C
for 48 h with and without 10% added CO2. The diameters of isolated
colonies on CBA were measured after 24 and 48 h of incubation in
CO2. The pigmentation of each isolate was compared with that of
other isolates by streaking an inoculating loop full of 48-h growth on
dry filter paper and then scoring the color on a scale from 1 (white)
to 5 (lemon yellow).
Antimicrobial sensitivity of the isolates was evaluated by a modi-
fied Kirby–Bauer procedure on Mueller–Hinton agar containing 5%
ovine blood: the Mueller–Hinton plates were incubated with 10%
CO2 enrichment to support the growth of CO2-dependent isolates.
All test plates were incubated at 37°C for 24 h under the same
atmospheric conditions regardless of CO2 requirement. Sensitivity
to: ampicillin, amoxicillin/clavulanic acid, ceftiofur, cephalothin,
clindamycin, enrofloxacin, erythromycin, gentamicin, neomycin,
florfenicol, penicillin, sulfachlorpyridazine, tetracycline, tilmicosin,
and trimethoprim/sulfamethoxazole was evaluated. Results were
interpreted according to performance standards of the National
Committee for Clinical Laboratory Standards (21).
Protein and antigen profiles
Isolates were grown on BBHIA, scraped off at 24 h, and suspended
in phosphate-buffered saline. Approximately 108 bacteria per well
were loaded in sodium dodecyl sulfate (SDS) sample buffer onto a
1-mm-thick 8% polyacrylamide gel. After SDS-polyacrylamide gel
electrophoresis (SDS-PAGE), 1 gel was stained with Coomassie
brilliant blue, and a duplicate gel was electrotransferred to polyvi-
nylidene difluoride (immobilon-P; Millipore Corporation, Bedford,
Massachusetts, USA), as previously described (19). Blots were exposed
to a pool of convalescent-phase bovine serum from animals with
experimental H. somnus infection (P3, E5) diluted 1:2000, as previously
described (20). After incubation and washing, blots were developed
with a conjugate of alkaline phosphatase and antibody to bovine IgG
(Kirkegaard & Perry Laboratories, Gaithersburg, Maryland, USA).
Each isolate was propagated on CBA at 37°C with CO2 for 24 h.
Growth was harvested and suspended in 1 mL of physiological
saline to produce approximately 1 108 cells, then centrifuged
(at 13 000 g for 1 min). The DNA was isolated with use of the
protocol and reagents supplied in a kit (Puregene DNA Isolation Kit;
Gentra Systems, Minneapolis, Minnesota, USA). Briefly, this entailed
cell lysis aided by heating to 80°C for 5 min, followed by incubation
with 0.5 U RNAase at 37°C for 1 h. After protein precipitation, DNA
was precipitated from the remaining supernatant with cold 100%
isopropanol and then washed twice with 70% ethanol. The resulting
DNA was hydrated and quantified by means of Hoechst 33258 fluo-
rometry (Sigma Chemical Company, St. Louis, Missouri, USA).
Restriction fragment length polymorphism (RFLP)
Purified DNA (1.5 g) from each isolate was digested with 5 U
of Hinf I restriction endonuclease (Roche Molecular Biochemicals,
Indianapolis, Indiana) for 3 h at 37°C in a final reaction volume
of 20 L. The entire volume was mixed with 3 L of blue/orange
loading dye (Promega, Madison, Wisconsin, USA) and loaded
on a 15-cm-long agarose gel (1% Synergel; Diversified Biotech,
Boston, Massachusetts, USA) for electrophoresis at 30 V for 16 h.
The gel was stained for 1 h with 50 parts per million of ethidium
bromide (J.T. Baker Chemical Company, Phillipsburg, New Jersey,
USA), rinsed in running tap water for 15 min, and imaged with a
computer-coupled gel documentation system (Gel Doc 2000; Bio-Rad
Laboratories, Hercules, California, USA). Banding patterns were used
to calculate indices for the relationship between all isolates (22).
Polymerase chain reaction (PCR)
Isolated DNA from each of the isolates was tested in a PCR pro-
cedure for the lktA gene, which codes for leukotoxin production
and has been demonstrated by Fisher et al (23) to be associated
with production of hemolysis on blood agar by Pasteurella and
Table I. Identification and source of Haemophilus somnus isolates in the United States and Canada in this study
a For bighorn sheep, the number begins with the last 2 digits of the year of isolation
b BHS — bighorn sheep; Mtn — Mountain; bovine — domestic cattle; ovine — domestic sheep
Herd with poor lamb recruitment
Herd with poor lamb recruitment
Herd with poor lamb recruitment
Herd with poor lamb recruitment
Pancake Range, Nevada
Pancake Range, Nevada
Pancake Range, Nevada
Pancake Range, Nevada
Wildlife Center, Caldwell, Idaho
Muddy Mountains, Nevada
Upper Hell’s Canyon, Idaho
Muddy Mountains, Nevada
Desert BHS; vagina
Desert BHS; vagina
Desert BHS; vagina
Desert BHS; vagina
Rocky Mtn BHS; lung
Desert BHS; vagina
Rocky Mtn BHS; lung
Desert BHS; vagina
Ovine; lung/lymph node, spleen, brain
2000;64:0–00 The Canadian Journal of Veterinary Research 37
Thin sections were cut from formalin-fixed, paraffin-embedded
lung tissue samples, stained with hematoxylin and eosin (HE), and
examined for evidence of microscopic lesions. Thin sections were
also cut from the embedded tissue at the Prairie Diagnostic Service
Laboratory and evaluated for H. somnus and M. haemolytica specific
antigens. Antigen detection was conducted with an avidin–biotin
immunoperoxidase technique (24).
Colonies with characteristics of H. somnus were isolated from 4
of the 1996, 1 of the October 1998, and 1 of the October 1999 vaginal
swabs but not from any of the 12 preputial samples (Table I). Similar
organisms were isolated from lung tissue of the 2 Rocky Mountain
bighorn sheep that died with evidence of pneumonia in 1997 and
1999. All reproductive tract isolates and 1 lung isolate (99-1544-10)
from the bighorn sheep produced colonies 2.0 to 2.6 mm in diam-
eter on CBA after 48 h incubation with added CO2; bighorn isolate
97-1205-2a and all isolates from domestic livestock produced smaller
colonies (1.0 to 2.0 mm in diameter).
All the isolates from bighorn sheep and domestic livestock pro-
duced oxidase and indole, reduced nitrate, were catalase-negative,
fermented fructose, glucose, and mannitol, and did not grow
on MacConkey’s agar. From reactions in H. somnus test media,
5 bighorn-sheep isolates and 2 domestic-sheep isolates were identi-
fied as biovariant 4, domestic-sheep isolate 2041 was identified as
biovariant 5, and the bovine isolates were identified as biovariant 1;
bighorn-sheep isolate 97-1205-2a was identified as a new biovariant,
biovariant 7 (Table II). All of the bighorn-sheep isolates grew opti-
mally on CBA in the absence of CO2 but not on CBAA, in contrast
to the isolates from domestic livestock, which all required CO2 for
optimal growth and grew on CBAA (Table III). All bighorn isolates
except 97-1205-2a scored 2 for pigment production, 97-1205-2a
scored 3, all isolates from domestic sheep scored 4, and both bovine
isolates scored 5. Hemolytic activity varied when evaluated on
BBHIA and OBHIA prepared with different lots of blood. None of
the isolates produced hemolysis on 1 lot of each medium, but isolate
99-1544-10 produced weak hemolysis on BBHIA prepared with a
Table II. Biochemical characteristics and biovariant types of the Haemophilus somnus isolates
Test result and source of isolates
I A B C E F G H J K M
a Determined by the results of biochemical utilization tests (20)
4 4 4 4 7 4 4 4 1 1 4 5 4
Table III. Colony pigmentation, hemolytic activity, and CO2 requirement of the Haemophilus somnus isolates
a Scored on a scale of 1 (white) to 5 (lemon yellow)
b Evaluated on brain–heart infusion agar (BHIA) with 5% bovine blood added (BBHIA) or with 5% ovine blood added
(OBHIA); ± indicates that the isolate produced partial clearing of the medium beneath isolated colonies
Test result and source of isolates
38 The Canadian Journal of Veterinary Research 2000;64:0–00
different lot of bovine blood, and all isolates except 97-1205-2a and
the 3 domestic-sheep isolates produced weak hemolysis on OBHIA
prepared with a second lot of ovine blood. The weak hemolysis was
characterized by partial clearing of the medium directly under iso-
lated colonies. None of the isolates produced hemolysis comparable
to that of the M. haemolytica strain.
Growth inhibition zones for all of the isolates equaled or exceeded
those published for all of the antibiotics with the exception of isolate
2336, which had a zone indicative of intermediate sensitivity to
tetracycline, and isolates 98-1451-232 and 99-1598-211, which had
zones indicative of resistance to clindamycin.
Protein profiles for the isolates in SDS-PAGE gels stained with
Coomassie brilliant blue (Figure 1) contained numerous similar
bands, including darker major outer membrane protein (MOMP)
bands for all isolates. As noted previously (25), the relative molecu-
lar weight (rMW) of the MOMPs of bovine preputial isolate 129Pt,
approximately 33 kDa, was much lower than that of most of the
other isolates. Bighorn-sheep isolate 99-1598-211 had a slightly higher
MOMP rMW than most of the other isolates, approximately 41 kDa.
In contrast, the lower rMW bands were approximately 38 kDa in the
isolates from the cattle and domestic sheep but were quite variable
in the bighorn-sheep isolates.
Antigenic profiles after Western blot testing (Figure 2) revealed
a similar 78-kDa antigen band in all isolates. A 76-kDa antigen was
present in the 3 domestic-sheep isolates (67P, 2041, L1203B) and the
virulent bovine isolate (2336) but not in the bovine preputial carrier
isolate (129Pt). Although this band is faint in the depicted profiles
of the four 96-1224 bighorn isolates, it was more clearly defined
in other blots. The 76-kDa antigen band was also detected in the
other 4 bighorn isolates (97-1205-2a, 98-1451-232, 99-1544-10 and
99-1958-211). The 40-kDa antigen was strongly reactive in all iso-
lates, but the bighorn isolates sometimes had 2 bands, and isolates
97-1205-2a and 99-1544-10 had a slightly higher rMW band for this
immunodominant antigen. The 270-kDa antigen was not present in
bovine isolate 129Pt but was variably detectable in the other isolates.
However, the related series of bands at about 120 kDa was detected
in all bighorn-sheep isolates.
The RFLP profiles for the 96-1224 isolates were identical (Figure 3),
which is indicative of isolation of a single clone of H. somnus from
that bighorn population (Pancake Range). Isolates 98-1451-232 and
Figure 1. Protein profiles of Haemophilus somnus isolates after 8% sodium
dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and staining
with Coomassie brilliant blue. The 41-kDa major outer membrane protein is
indicated on the right. Markers for relative molecular weight (rMW) are on
the left and isolate numbers above each profile in all the figures.
Figure 2. Antigenic profile with Western blot testing of the isolates after
exposure to convalescent-phase serum from 2 calves with experimental
Haemophilus somnus infection. The rMWs of key H. somnus antigens are
on the right.
Figure 3. Restriction fragment length polymorphism profiles of DNA from
the isolates after digestion with Hinf I restriction endonuclease, electro-
phoresis on 1% Synergel, and staining with blue/orange loading dye and
2000;64:0–00 The Canadian Journal of Veterinary Research 39
99-1544-10 produced profiles with similarity coefficient values near
1.0, showing close similarity. The remaining isolates each produced
a unique profile. However, all bighorn isolates had DNA fragments
of approximately 6100, 4200, and 2900 kb and a uniform absence of
fragments in the area equivalent to approximately 2700 kb.
The DNA from all test isolates was PCR-negative for the lktA
gene, in contrast to the M. haemolytica control isolate, which was
Necropsy of bighorn sheep 6 to 24 h after death revealed that
essentially all had gross evidence of severe fibrinous pleuritis
and subacute to chronic bronchopneumonia. Lung consolidation,
which generally affected more than 30% of lung tissue, was usually
bilateral; the most severe consolidation involved the anteroventral
lobes, and there was progressive involvement of the posterior lobes.
The advanced consolidation in the anterior lobes suggested that
the lesions could have been 5 to 7 d old. Histologic examination
revealed acute inflammation in the posterior lobes and more chronic
changes of fibrosis and beginning fibroplasia in the alveolar septa
and bronchiolar walls. Inspissated masses of cellular debris were
seen in alveolar lumens. Streaming of alveolar exudates characteristic
of M. haemolytica infection was not observed. Significant vasculitis
was not recognized except in areas of severe inflammation. Although
thrombosis of blood vessels was not seen, sizeable fibrinocellular
masses were present in lymphatic channels in the interlobular
septa. Similar masses of exudate were seen in the afferent lymphatic
channels and the sinus spaces of markedly swollen mediastinal and
bronchial lymph nodes.
Immunohistochemical study with antibody specific for H. somnus
antigens detected positive reactions in a few masses of inspissated
intra-alveolar cellular debris in any given section of 4 of the 12 lung
samples evaluated. In an area of severe fibrinopurulent pleuritis in
case 99-1544, positive staining was seen in thickened pleural fibrous
tissue and in adherent exudate. Blood vessels did not react with the
conjugate. None of the tissue sections were immunohistologically
positive for M. haemolytica antigens.
This is the first known report of isolation of H. somnus from big-
horn sheep. Isolation of the organisms from both the reproductive
and respiratory tracts follows a pattern of colonization similar to
that in domestic livestock. Although colonization of the prepuce is
common in domestic livestock (3,6), colonization was not detected
by culture of any of the 12 samples collected from bighorn rams.
However, isolation of a single clone of H. somnus from 4 vaginal
samples in the Pancake Mountain herd suggests a common source
and transmission by a dominant ram during rut.
The bighorn-sheep isolates of H. somnus had both similarities
to and differences from the isolates from the domestic sheep and
cattle. Although all but 1 of the bighorn isolates and 2 of the
domestic-sheep isolates produced identical biochemical utilization
reactions that were characteristic of biovariant 4, the bighorn-sheep
isolates lacked the yellow pigmentation and CO2 requirement of the
domestic-livestock isolates. Lack of these cardinal characteristics, as
described for domestic-livestock isolates (10), could result in lack of
detection and identification of the organism.
Although the protein and antigenic profiles of the bighorn-sheep,
bovine, and domestic-sheep isolates were predominantly similar,
a few interesting differences were noted. Isolate 99-1598-211 had
an MOMP band with a slightly higher rMR than the usual 41 kDa,
which could be important, as variation in size of the 41-kDa MOMP
has been noted in many H. somnus isolates and may be associated
with virulence (26). The 2 isolates from bighorn sheep with fatal
pneumonia had variation in protein bands below 41 kDa, along
with variations in antigenic profile, especially in the size of the
40-kDa antigen. We previously showed that antibodies to this OMP
protected against experimental H. somnus pneumonia in calves
(27). Therefore, differences in molecular structure of this antigen
may alter the host–parasite relationship. In addition, the bighorn-
sheep isolates had various amounts of the 120/270 and 76-kDa
immunoglobulin-binding proteins (IgBPs). The strains with low
amounts could have low expression, or the IgBPs may have been
shed from their surfaces. Bovine preputial isolates, from asymptom-
atic carriers, not expressing these proteins (129Pt) are sensitive to
killing by bovine complement, whereas isolates that express these
IgBPs, such as 2336, are resistant to complement-mediated killing
and are virulent (28). Therefore, the expression of these antigens
by the bighorn-sheep isolates suggests that they may be virulent.
Although some of the bighorn-sheep isolates were from apparently
normal animals tested in a health-monitoring program, 4 isolates
were from a herd in the Pancake Mountain Range with a history
of poor lamb recruitment. Others have shown that H. somnus (or
H. ovis) is not part of the normal vaginal flora of mature domestic
ewes (29). In studies of natural outbreaks, H. somnus (or H. ovis)
has been reported to be associated with vaginal discharge (30) and
reduced lambing rates (6) in domestic sheep. Therefore, it could be
that the H. somnus isolated from the vaginas of bighorn sheep in the
Pancake Mountain population was one of the causes of poor lamb-
ing or low lamb-survival rates, or both. Whether the vaginal isolates
from bighorn sheep in the Muddy Mountain population were associ-
ated with poor reproductive performance is unknown.
Antimicrobial sensitivity profiles were similar for all the isolates
and correspond with those reported elsewhere for H. somnus and
other members of the Pasteurellaceae family (31). However, the
failure of the bighorn isolates to grow on CBAA, in contrast to the
ability of domestic isolates to grow on that medium, indicated that
1 or more of the incorporated antimicrobials inhibited the bighorn
The multiple similarities in RFLP profiles of the bighorn isolates
and their distinct differences from those of the domestic-livestock
isolates demonstrates that the bighorn-sheep isolates are more closely
related to one another than they are to domestic-livestock isolates.
This was not expected, since all but 4 of the bighorn sheep isolates
were from animals that were geographically widely separated, and it
could suggest distinct phylogenetic lines for H. somnus in the bighorn
sheep and the domestic livestock.
A few studies have shown H. somnus to be host species-specific.
Biberstein (7) reported that H. agni (ovine H. somnus) was pathogenic
for lambs but that H. somnus from cattle was not pathogenic for
lambs. This may be explained by the fact that bovine H. somnus binds
bovine transferrin, but not ovine transferrin, as an iron source (32).
However, the reverse situation may be more complicated: H. agni and
40 The Canadian Journal of Veterinary Research 2000;64:0–00
H. ovis (both considered to be ovine H. somnus) bind ovine, bovine,
and goat transferrins (33). More recently, Ekins et al (34) showed that
strain 649 of bovine H. somnus has 2 transferrin-binding proteins,
1 specific for bovine transferrin and the other binding ovine, caprine,
or bovine transferrin. Although the system is complex, the studies
showing that ovine H. somnus binds bovine transferrin suggest
that this organism may be pathogenic for cattle unless some other
virulence factor is also host-specific. Evidence for host specificity
of isolates from bighorn sheep and for binding of transferrin from
other species is lacking.
In previous studies, we found that bovine isolates varied in abil-
ity to lyse bovine erythrocytes, whereas none of the isolates from
domestic sheep produced hemolysis (19). This is different from the
reactions in this study, in which 7 of the 8 bighorn-sheep isolates
and the 2 bovine isolates produced weak hemolysis on 1 lot of
OBHIA plates. Since hemolysis is reportedly associated with viru-
lence in other gram-negative bacteria (23), the hemolytic activity of
H. somnus may also be associated with a virulence factor. Unlike the
-hemolytic M. haemolytica and P. trehalosi strains that reacted in the
PCR test for the lktA gene (23), none of the H. somnus isolates evalu-
ated in this study was positive for the lktA gene. It is clear, however,
that nonhemolytic H. somnus isolates have enough other virulence
factors to cause disease. We have shown that nonhemolytic bovine
strain 2336 caused pneumonia in experimentally infected calves
(35), and the nonhemolytic domestic-sheep isolates 67P and 2041
reproduced neurologic disease in domestic sheep (20).
The isolation of H. somnus from fatal cases of pneumonia in a
bighorn-sheep epizootic in Hell’s Canyon is likely etiologically
significant, for several reasons. First, H. somnus was detected by
immunohistochemical study in areas of pneumonia or pleuritis.
Detection of antigen specific for H. somnus in alveolar debris is
consistent with our earlier findings of immunostaining lesions in
calves with experimental H. somnus pneumonia (35). The calves
were euthanized at 12 to 24 h after intrabronchial inoculation with
H. somnus, and the lesions were characterized by vasculitis, bron-
chiolitis, lobar necrosis, and dilation and thrombosis of lymphatic
channels. In other studies of bovine pneumonia with a 12-d experi-
mental period, Jackson et al (36) found that mild vasculitis was
detected only occasionally. Rather, pneumonia was characterized
by fibrinopurulent bronchiolitis, peribronchiolar fibrosis, interlobu-
lar fibrosis, and thrombosis of interlobular and pleural lymphatic
channels. From studies of 68 clinical cases of bovine pneumonia
positive only for H. somnus, Andrews et al (37) reported microscopic
lesions of purulent bronchiolitis and bronchopneumonia in 61 cases,
fibrinous pneumonia with bronchiolitis in 2, fibrinous pleuritis in 2,
suppurative interstitial pneumonia with vasculitis in 2, and diffuse
congestion in 1. No vasculitis or vascular thrombosis was detected
in the lungs in 66 of the 68 cases, even though these were predomi-
nant findings in experimentally infected calves killed at 24 h (35).
Perhaps the difference is due to the stage of infection or the extent
of septicemia in experimental versus clinical cases. The histopatho-
logic findings in the pneumonic lungs of the bighorn sheep were
more consistent with the more chronic pneumonia detected in the
12-d bovine experiment and in clinical cases. In domestic sheep, H.
somnus has also been shown to cause pneumonia (38–40). Poonacha
and Donahue (38) reported isolation of H. ovis (H. somnus) from a
lamb (in a flock with multiple lamb deaths) that had died suddenly;
they noted severe diffuse pulmonary congestion and edema, multifo-
cal hemorrhages, interstitial pneumonia, vasculitis, and thrombosis
on histologic examination. Later, Lundberg (39) isolated H. agni
(H. somnus) from 6 ram lambs (4 to 6 mo of age) from different
domestic flocks in Alberta and British Columbia. These lambs had
been submitted for necropsy after sudden illness or death, with or
without prior depression, stiffness, or fever. Septicemia was evident
by gross and histopathologic examination, and the organism was cul-
tured from lung, brain, liver, spleen, and lymph nodes. Microthrombi
were detected in alveolar septa of most of the lungs, and 3 lambs had
occasional vascular thrombi in the brain. The vasculitis was associ-
ated with septicemia rather than subacute or chronic pneumonia.
Rahaley (40) studied experimental H. ovis (H. somnus) infection of
lambs. Intravenous inoculation resulted in a range of findings, from
abscesses to death in 30 or 36 h. In the 2 lambs dying at 30 and 36 h,
there was evidence of septicemia and pulmonary congestion, with
vasculitis of pulmonary arteries in 1 lamb. One lamb inoculated
intranasally died at 72 h, with consolidation of the apical lobes of
the lungs and accumulation of cellular debris in the lumen of bron-
chioles but no reported vasculitis. Again, it appears that vasculitis
was present at the stage of more acute septicemia but was absent in
the more slowly developing chronic stage, as would be suspected in
the bighorn-sheep cases in our study.
In conclusion, persuasive evidence that H. somnus can cause
disease in bighorn sheep was demonstrated by the isolation of
H. somnus from, and the presence of H. somnus antigens in, pneu-
monic lesions. Although the significance of these organisms in the
reproductive tract is only speculative, such infections may cause
poor lamb recruitment or serve as a reservoir from which respiratory
disease may occur, or both. Pasteurella and Mannheimia spp. have
been incriminated as the causes of most pneumonic epizootics in
bighorn sheep. However, since H. somnus is not as readily detect-
able, its presence could have been missed in many of the studies in
which nasal or oropharyngeal samples, or both types, were cultured.
Growth of these organisms can be obscured by the multiple bacte-
rial species that constitute the normal flora of the upper respiratory
tract of ruminants. This is especially true in bighorn sheep, since
H. somnus isolates from that species do not require CO2 for growth,
and the colonies produce less pigment than the more characteristic
isolates from domestic sheep and cattle. Diagnosticians, researchers,
and wildlife managers should be aware that these organisms can be
associated with disease and should conduct procedures adequate
for their detection when evaluating respiratory disease in bighorn
The authors thank all individuals who assisted with the capture of
bighorn sheep and the collection and timely submission of samples
evaluated in this study. The authors also thank Lynna Dibben,
who prepared specialty media for isolation and characterization of
bacteria from the samples, and Lisa Cowan and Tiffany Brush, who
conducted countless laboratory tests.
This research was supported in part by funds provided by the
University of Idaho, College of Agriculture, Idaho Agriculture
2000;64:0–00 The Canadian Journal of Veterinary Research 41
Experiment Station Research Project BGR123, and the Idaho
Department of Fish and Game (Ward), as well as by US Department
of Agriculture National Research Initiative grants #1998-35204-6733
and #2001-35204-10803 (Corbeil). The article was submitted with the
approval of the University of Idaho Experiment Station as manu-
1. Kennedy PC, Frazier LM, Theilen GN, Biberstein EL. A septice-
mic disease of lambs caused by Haemophilus agni (new species).
Am J Vet Res 1958;19:645–654.
2. Kennedy PC, Biberstein EL, Howarth JA, Frazier LM, Dungworth
DL. Infectious meningo-encephalitis in cattle, caused by a
Haemophilus-like organism. Am J Vet Res 1960;21:403–409.
3. Humphrey JD, Stephens LR. ‘Haemophilus somnus’: a review. Vet
4. Guichon PT, Pritchard J, Jim GK. Haemophilus somnus myocarditis
in a feedlot steer. Can Vet J 1988;29:1012–1013.
5. Philbey AW, Glastonbury JRW, Rothwell JT, Links IJ, Searson JE.
Meningoencephalitis and other conditions associated with
Histophilus ovis infection in sheep. Aust Vet J 1991;68:387–390.
6. Lees VW, Meek AH, Rosendal S. Epidemiology of Haemophilus
somnus in young rams. Can J Vet Res 1990;54:331–336.
7. Biberstein EL. “Haemophilus somnus” and “Haemophilus agni”.
In: Kilian M, Frederiksen W, Biberstein EL, eds. Haemophilus,
Pasteurella and Actinobacillus. San Diego, California: Academic
8. Stephens LR, Humphrey JD, Little PB, Barnum DA.
Morphological, biochemical, antigenic, and cytochemical
relationships among Haemophilus somnus, Haemophilus agni,
Haemophilus haemoglobinophilus, Histophilus ovis, and Actinobacillus
seminis. J Clin Microbiol 1983;17:728–737.
9. Walker RL, Biberstein EL, Pritchett RF, Kirkham C.
Deoxyribonucleic acid relatedness among “Haemophilus somnus,”
“Haemophilus agni,” “Histophilus ovis,” “Actinobacillus seminis,”
and “Haemophilus influenzae.” Int J Syst Bacteriol 1985;35:46–49.
10. Angen O, Ahrens P, Kuhnert P, Christensen H, Mutters R.
Proposal of Histophilus somni gen. nov., sp. nov for the
three species incertae sedis ‘Haemophilus somnus’, ‘Haemophilus
agni’ and ‘Histophilus ovis’. Int J Syst Evol Microbiol 2003;53:
11. Ward ACS, Dyer NW, Corbeil LB. Characterization of putative
Haemophilus somnus from tonsils of American bison (Bison bison)
Can J Vet Res 1999;63:166–169.
12. Dyer NW. Haemophilus somnus bronchopneumonia in American
bison (Bison bison). J Vet Diagn Invest 2001;13:419–421.
13. Post G. Pasteurellosis of Rocky Mountain bighorn sheep
(Ovis canadensis canadensis). J Wildl Dis 1962;23:1–14.
14. Foreyt WJ. Pneumonia in bighorn sheep: effects of Pasteurella
haemolytica from domestic sheep and effects on survival and
long-term reproduction. Bienn Symp North Wild Sheep Goat
15. Ward ACS, Jaworski MD, Hunter DL, et al. Pasteurella spp in
sympatric bighorn and domestic sheep. J Wildl Dis 1997;33:
16. Cassirer EF, Oldenburg LE, Coggins VL, et al. Overview and
preliminary analysis of a bighorn sheep dieoff, Hell’s Canyon,
1995–96. Bienn Symp North Wild Sheep Goat Counc 1996;10:
17. Jaworski MD, Ward ACS, Hunter DL, Wesley IV. Use of DNA
analysis of Pasteurella haemolytica biotype T to monitor trans-
mission in bighorn sheep (Ovis canadensis canadensis). J Clin
18. Jaworski MD, Hunter DL, Ward ACS. Biovariants of isolates of
Pasteurella from domestic and wild ruminants. J Vet Diagn Invest
19. Ward ACS, Jaworski MD, Eddow JM, Corbeil LB. A comparative
study of bovine and ovine Haemophilus somnus isolates. Can J Vet
20. Lees VW, Yates WDG, Corbeil LB. Ovine Haemophilus somnus:
experimental intracisternal infection and antigenic comparison
with bovine Haemophilus somnus. Can J Vet Res 1994;58:202–210.
21. Watts JL, Chengappa MM, Doe JR, et al. Performance Standards
for Antimicrobial Disk and Dilution Susceptibility Tests for
Bacteria Isolated from Animals; Approved Standard (NCCLS doc
M31-A). Wayne, Pennsylvania: National Committee for Clinical
Laboratory Standards, 1999.
22. Nei M, Li W-H. Mathematical model for studying genetic varia-
tion in terms of restriction endonucleases. Proc Natl Acad Sci U
S A 1979;76:5269–5273.
23. Fisher MA, Weiser GC, Hunter DL, Ward ACS. Use of a poly-
merase chain reaction method to detect the leukotoxin gene lktA
in biogroup and biovariant isolates of Pasteurella haemolytica and
P. trehalosi. Am J Vet Res 1999;60:1402–1406.
24. Haines DM, Chelack BJ. Technical considerations for developing
enzyme immunohistochemical staining procedures on formalin-
fixed paraffin-embedded tissues for diagnostic pathology. J Vet
Diagn Invest 1991;3:101–112.
25. Widders PR, Dorrance LA, Yarnall M, Corbeil LB. Immunoglobulin-
binding activity among pathogenic and carrier isolates of
Haemophilus somnus. Infect Immun 1989;57:639–642.
26. Tagawa Y, Bastida-Corcuera F, Corbeil LB. Immunological char-
acterization of the major outer membrane protein of Haemophilus
somnus. Vet Microbiol 2000;71:245–254.
27. Gogolewski RP, Kania SA, Liggitt HD, Corbeil LB. Protective
ability of antibodies against 78- and 40-kilodalton outer mem-
brane antigens of Haemophilus somnus. Infect Immun 1988;56:
28. Cole SP, Guiney DG, Corbeil LB. Two linked genes for outer
membrane proteins are absent in four non-disease strains of
Haemophilus somnus. Mol Microbiol 1992;6:1895–1902.
29. Rahaley RS, Edwards LD. Histophilus ovis infection in ewes.
Can Vet J 1983;24:61–62.
30. Higgins R, Godbout-deLaSalle F, Messier S, Couture Y,
Lamothe P. Isolation of Histophilus ovis from vaginal discharge
in ewes in Canada. Can Vet J 1981;22:395–396.
31. Dyer NW, Ward ACS, Weiser GC, White DG. Seasonal incidence
and antibiotic susceptibility patterns of Pasteurellaceae isolated
from American bison (Bison bison). Can J Vet Res 2001;65:7–14.
32. Yu RH, Gray-Owen SD, Ogunnariwo J, Schryvers AB. Interaction
of ruminant transferrins with transferrin receptors in bovine
42 The Canadian Journal of Veterinary Research 2000;64:0–00
isolates of Pasteurella haemolytica and Haemophilus somnus. Infect
33. Yu R, Schryvers AB. Transferrin receptors on ruminant pathogens
vary in their interaction with the C-lobe and N-lobe of bovine
transferrins. Can J Microbiol 1994;40:532–540.
34. Ekins A, Bahrami F, Sijercic A, Maret D, Niven DF. Haemophilus
somnus possesses two systems for acquisition of transferrin-
bound iron. J Bacteriol 2004;186:4407–4411.
35. Gogolewski RP, Leathers CW, Liggitt HD, Corbeil LB.
Experimental Haemophilus somnus pneumonia in calves
and immunoperoxidase localization of bacteria. Vet Pathol
36. Jackson JA, Andrews JJ, Hargis JW. Experimental Haemophilus
somnus pneumonia in calves. Vet Pathol 1987;24:129–134.
37. Andrews JJ, Anderson TD, Slife LN, Stevenson GW. Microscopic
lesions associated with the isolation of Haemophilus somnus from
pneumonic bovine lungs. Vet Pathol 1985;22:131–136.
38. Poonacha KB, Donahue JM. Haemophilus ovis infection in lambs.
Vet Med 1984;79:541–542.
39. Lundberg MS. Isolation of Haemophilus agni from six Alberta ram
lambs with septicemia. Can Vet J 1986;27:501–503.
40. Rahaley RS. Pathology of experimental Histophilus ovis infection
in sheep. I. Lambs. Vet Pathol 1978;15:631–637.