Complexities of the pathogenesis of Mannheimia haemolytica and Haemophilus somnus infections: challenges and potential opportunities for prevention?

Progress in producing improved vaccines against bacterial diseases of cattle is limited by an incomplete understanding of the pathogenesis of these agents. Our group has been involved in investigations of two members of the family Pasteurellaceae, Mannheimia haemolytica and Haemophilus somnus, which illustrate some of the complexities that must be confronted. Susceptibility to M. haemolytica is greatly increased during active viral respiratory infection, resulting in rapid onset of a severe and even lethal pleuropneumonia. Despite years of investigation, understanding of the mechanisms underlying this viral-bacterial synergism is incomplete. We have investigated the hypothesis that active viral infection increases the susceptibility of bovine leukocytes to the M. haemolytica leukotoxin by increasing the expression of or activating the beta2 integrin CD11a/CD18 (LFA-1) on the leukocyte surface. In vitro exposure to proinflammatory cytokines (i.e. interleukin-1beta, tumor necrosis factor-alpha and interferon-gamma) increases LFA-1 expression on bovine leukocytes, which in turn correlates with increased binding and responsiveness to the leukotoxin. Alveolar macrophages and peripheral blood leukocytes from cattle with active bovine herpesvirus-1 (BVH-1) infection are more susceptible to the lethal effects of the leukotoxin ex vivo than leukocytes from uninfected cattle. Likewise, in vitro incubation of bovine leukocytes with bovine herpesvirus 1 (BHV-1) potentiates LFA-1 expression and makes the cells more responsive to leukotoxin. A striking characteristic of H. somnus infection is its propensity to cause vasculitis. We have shown that H. somnus and its lipo-oligosaccharide (LOS) trigger caspase activation and apoptosis in bovine endothelial cells in vitro. This effect is associated with the production of reactive oxygen and nitrogen intermediates, and is amplified in the presence of platelets. The adverse effects of H. somnus LOS are mediated in part by activation of endothelial cell purinergic receptors such as P2X7. Further dissection of the pathways that lead to endothelial cell damage in response to H. somnus might help in the development of new preventive or therapeutic regimens. A more thorough understanding of M. haemolytica and H. somnus virulence factors and their interactions with the host might identify new targets for prevention of bovine respiratory disease.     

Ovine Haemophilus somnus: experimental intracisternal infection and antigenic comparison with bovine Haemophilus somnus.

Experimental infection was produced by two of four isolates of ovine Haemophilus somnus given by intracisternal inoculation into two to three-month-old lambs. Isolate 2041 (originally obtained from a septicemic lamb in Alberta) caused lethal infection in eight of nine lambs, isolate 67p from the prepuce of a normal lamb produced less acute disease in four of nine lambs, and the other two isolates (93p and 1190) caused no detectable disease. Significant lesions were limited to the brain and spinal cord. Purulent meningitis was characteristic but vasculitis or septicemia were not detected, perhaps due to the route of inoculation. Since a difference in virulence was noted among strains, we analyzed surface proteins thought to be virulence factors of bovine H. somnus. Protein profiles of bovine and ovine H. somnus done by sodium dodecyl sulphate-polyacrylamide gel electrophoresis showed similar patterns for virulent bovine isolates and ovine septicemic isolates. Preputial isolates showed a lower molecular mass major outer membrane protein than septicemic isolates. Antigenic analysis revealed that outer membrane proteins p270, p78, p76, p40, and p39 were detected in both ovine and bovine isolates except for 1190, which was probably not a true H. somnus isolate. Thus the preputial and septicemic isolates of ovine H. somnus were similar to bovine H. somnus in pathogenicity and in surface antigens.     

A comparative study of bovine and ovine Haemophilus somnus isolates.

Bacterial isolates (including 17 Haemophilus somnus isolates and an H. somnus-like isolate) from asymptomatic or diseased cattle and sheep, were evaluated for markers associated with virulence and host predilection. The isolates were separated into 6 distinct biovariants, 3 for sheep and 3 for cattle, based on reactions in a battery of 21 test media. Three bovine isolates associated with disease caused hemolysis of bovine blood. The rest of the isolates did not hemolyze either bovine or ovine erythrocytes. Protein profiles of all H. somnus isolates were similar with the exception of the major outer membrane proteins (MOMPs). The MOMPs of isolates associated with disease in cattle had a relative molecular weight of approximately 41 kDa compared with 33 kDa for the MOMPs of isolates from asymptomatic cattle. The MOMPs from sheep isolates were either slightly higher or lower than the 41 kDa MOMPs of bovine isolates. Major antigens detected by Western blotting were similar in all isolates except the H. somnus-like isolate. An immunodominant 40 kDa antigen was conserved in all H. somnus isolates. Antibodies to this antigen have previously been found to be protective in cattle and may also be protective for sheep. Marked differences between cattle and sheep isolates were revealed by use of restriction enzyme analysis, which separated the isolates into 12 ribotypes and 15 unique DNA profiles. Thus, cattle and sheep isolates in this collection had distinctive differences in biochemical reactions, MOMP profiles, and DNA analyses. Such differences have potential value for epidemiological studies and may also be used to evaluate host specificity of H. somnus isolates.     

Update on the diagnosis of Haemophilus parasuis infection in pigs and novel genotyping methods

Haemophilus parasuis causes Gl?sser's disease as well as a number of other diseases in pigs. The diagnosis of H. parasuis-associated disease is usually established by clinical signs, pathological findings and bacterial isolation but diagnosis is complicated by the existence of non-virulent strains and the early colonisation of the upper respiratory tract of healthy piglets. Moreover, several strains can be found on a farm and even within a single animal so it is important to determine the specific strain that is causing the clinical outbreak. Recently, genotyping methods have been developed with the goal of correlating genotype with the degree of virulence of H. parasuis strains. The association between genotype and virulence in H. parasuis is challenging due to the lack of knowledge of the complete genomic sequence and virulence factors of this bacterium.     

Haemophilus parasuis exhibits IgA protease activity but lacks homologs of the IgA protease genes of Haemophilus influenzae

Haemophilus parasuis, the bacterium responsible for Gl?sser's disease, is a pathogen of significant concern in modern high-health swine production systems but there is little information regarding the identity or function of its virulence factors. Several important human mucosal pathogens, including the closely related bacterium Haemophilus influenzae, utilize IgA proteases to aid in defeating the host immune response and facilitate disease but it is unknown whether H. parasuis synthesizes any product with IgA protease activity. To investigate potential virulence mechanisms of H. parasuis, we evaluated five strains for their ability to digest purified IgA. Western blotting demonstrated cleavage of swine IgA, but not human IgA1, following incubation with culture supernatants from three strains, two of which are known to cause invasive disease. No genes with homology to the H. influenzae IgA protease genes iga and igaB could be identified in any H. parasuis strain using either PCR or Southern blotting. These results demonstrate that a novel IgA protease produced by some strains of H. parasuis cleaves the swine IgA heavy chain at a site not found in human IgA1.     

Haemophilus somnus (Histophilus somni) in bighorn sheep.

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.     

Binding of bovine IgG2a and IgG2b allotypes to protein A, protein G, and Haemophilus somnus IgBPs.

Immunoglobulin binding proteins (IgBPs) are thought to be virulence factors which enable pathogens to evade the host's immune response. Since bovine IgG2 is important in protection against pyogenic infections, the binding characteristics of Staphylococcus aureus protein A (PrA), streptococcal protein G (PrG), or Haemophilus somnus high molecular weight IgBPs to the two bovine IgG2 allotypes were examined. For PrA or PrG binding of IgG2, guinea pig red blood cells coated with specific IgG2a or IgG2b antibodies were used in a competitive binding inhibition assay with unlabeled and horseradish peroxidase-labeled PrA or PrG. To determine which sizes of H. somnus. IgBPs bind to the two IgG2 allotypes, immunoblots with H. somnus culture supernatant were probed with anti-DNP IgG2a and IgG2b. This detects only Fc binding because anti-DNP does not cross-react with H. somnus antigens. Both IgG2 allotypes bound equally well to PrA and PrG. However, IgG2b but not IgG2a bound to H. somnus high molecular weight IgBPs. The lack of differential binding of bovine IgG2 allotypes to PrA and PrG means that these IgBPs can be considered to be unbiased reagents in assays for detection of bovine IgG2 or for immunoaffinity purification. The differential binding of H. somnus IgBPs to the IgG2 allotypes indicates that animals having one allotype may be more resistant to H. somnus infection than animals having the other allotype.     

First isolation of Histophilus somni from goats

Histophilus somni (former name: Haemophilus somnus) is a Gram-negative, facultative pathogen bacterium that colonises the mucous membranes of cattle and sheep, however it was also described in American bison and bighorn sheep. It can cause local or generalised diseases and asymptomatic carriers can also occur. The presence and the etiological role of this microorganism have not been confirmed in any other domesticated species yet. The purpose of this study was to prove the presence of H. somni in goats by bacterial isolation. Nasal, vaginal or praeputial swab samples were collected from 205 goats in 10 flocks. H. somni strains were isolated from 2 out of 10 flocks; in one flock 10 H. somni strains were isolated from the genital mucosa of 17 goats, while a single H. somni strain was cultured from a vagina of 26 animals in the other flock. Partial amplification and sequencing of the 16S rRNA gene of three H. somni strains verified the identification. The comparative examination of carbon source metabolism using the Biolog Microstation ID System (Biolog, Ca) showed a close relationship of the caprine strains, while they were less related to H. somni type strain CCUG-36157 of bovine origin. H. somni strains were isolated only in the oestrus season from goat flocks with sheep contact. This is the first paper on isolation of H. somni from goats.     

Selective medium for isolation of Haemophilus somnus from cattle and sheep

Incorporation of vancomycin (5 micrograms/ml), neomycin (5 micrograms/ml), sodium azide (50 micrograms/ml), nystatin (100 iu/ml) and cyclohexamide (100 micrograms/ml) into 5 per cent horse blood agar results in a selective medium for the primary isolation of Haemophilus somnus from cattle and sheep. Addition of thiamine monophosphate (1 microgram/ml) to the medium enhanced growth of this bacterium. Gram-positive bacteria did not grow on the medium and colonies of many Gram-negative bacteria were eliminated or reduced in numbers and size. Colonies of H somnus were larger on the selective medium than on sheep blood agar but retained typical morphology. Recovery of 18 laboratory strains was 73 to 166 per cent (mean 112) on selective medium compared to sheep blood agar. H somnus was isolated from the vagina of a total of 136 (28.6 per cent) of 476 cows surveyed, 79 (16.6 per cent) on sheep blood agar and 129 (27.1 per cent) on selective medium. The selective agents and thiamine were stable indefinitely as a freeze dried mixture while prepared plates were stable for two weeks.     

Development of improved methods for the transport and isolation of Haemophilus somnus

The survival of several strains of Haemophilus somnus under various simulated transport conditions was investigated. Recovery of H somnus from alginate swabs kept at room temperature was possible for up to 27 hours after sampling, but this period could be extended to 72 hours if the swabs were refrigerated. Storage of swabs in transport media did not prolong survival time significantly but did increase the number of bacterial contaminants, thus making recovery of H somnus less likely. The sensitivity of several strains of H somnus to a number of dyes, antibiotics and other antimicrobial agents was determined and a selective isolation medium then formulated. This medium which consisted of brain heart infusion agar supplemented with 5 per cent ovine blood, 5 per cent equine serum, 0.5 per cent yeast extract, cycloheximide (100 micrograms ml-1) and lincomycin (3 micrograms ml-1), facilitated the isolation of H somnus from contaminated material. However, while this medium was effective against many contaminants, it did not prevent swarming by Proteus species.     

Molecular characterization of the major outer membrane protein of Haemophilus somnus

The major outer membrane protein (MOMP) of Haemophilus somnus shows antigenic and molecular mass diversity that forms the basis of a preliminary grouping system for H. somnus strains. In this study, the gene encoding MOMP of H. somnus strain 8025 was cloned in three overlapping fragments by PCR techniques, and then sequenced. The gene consists of a 1164-bp open reading frame encoding a deduced 380-amino acid protein with a 19-amino acid signal sequence, giving a mature protein with a calculated molecular mass of 39,913 Da. Significant homology was found between MOMP and porin protein sequences of bacteria in Pasteurellaceae species. When expressed in Escherichia coli, the protein from the MOMP gene directed by the T7 promoter was identical in size (approximately 40 kDa) to native MOMP and reacted with MOMP-specific antibodies. Comparisons of the MOMP gene sequences from six unrelated strains of H. somnus to that of strain 8025 revealed that the genes of three MOMP type 1 strains were highly conserved with that of strain 8025 in length and sequence. However, two MOMP type 3c strains and one MOMP type 3a strain differed markedly from the MOMP of strain 8025 in their 3'-terminal halves. Their deduced MOMP amino acid sequences differed in sequence (3c, 80.5 and 82.7% identity; 3a, 62.4% identity) and in length (3c, 384 and 376; 3a, 316), indicating that the molecular differences are the basis of antigenicity and molecular mass differences of H. somnus MOMP. In the predicted MOMP secondary structure, the variable sequences primarily mapped to putative surface-exposed loops, and a variable and surface-exposed epitope of MOMP-specific antibody was identified in the seventh-largest loop. These findings are useful for understanding the structural and immunological characteristics of H. somnus.     

Virulence differences among three strains of Haemophilus somnus following intratracheal inoculation of calves.

Pneumonia was induced in four month old Holstein calves by intratracheal inoculation of 1 x 10(9) colony forming units of Haemophilus somnus. Twenty calves were divided into four groups of five and challenged with a pneumonic strain (Group 1), an encephalitic strain (Group 2), a preputial strain (Group 3), or a placebo (Group 4). The clinical score, neutrophil count, respiratory rate, and temperature were significantly increased in group 1 by day 1 postinoculation (P less than 0.05) and maintained until day 6 postinoculation (P less than 0.05). The macroscopic pathological changes were significantly greater in group 1 (P less than 0.05). Haemophilus somnus was consistently isolated from pneumonic tissue of group 1 only. Groups 2 and 3 had mild transient increases in all parameters measured and macroscopically only small focal lesions were present. It is concluded that virulence differences exist between H. somnus strains following intratracheal challenge of bovine lungs.     

Structural features of the avian influenza virus hemagglutinin that influence virulence

Analysis of the structure of the avian influenza (AI) virus hemagglutinin (HA) gene and protein has yielded a wealth of information on the virulence mechanisms of influenza viruses. The AI hemagglutinin appears to be unique in its capacity to accept basic amino acids at its proteolytic cleavage site (PCS). The association of multiple basic (MB) amino acids, HA cleavage, tissue spread and virulence by AI strains first proposed in the late 1970s and early 1980s [Klenk, H.D., Rott, R., Orlich, M., 1977. J. Gen. Virol. 36, 151-161; Bosch, F.X., Garten, W., Klenk, H.D., Rott, R., 1981. Virology 113, 725-735] has held fast for two decades now. While other structural characteristics and other genes can certainly influence virulence, the presence of MB amino acids at the PCS has provided a hallmark structural feature which justifies continuing sequence analysis of emerging field isolates of AI strains. In addition to this structural feature, the distal tip of the HA is prone to appearance and disappearance of glycosylation sites, some of which have been associated with virulence.The recent outbreaks of highly pathogenic AI in Mexico, Australia, Pakistan, Hong Kong and in the ongoing outbreak of moderately pathogenic H7 avian influenza in the northeast US have all provided new and useful information regarding the role of HA RNA and protein structure in both virulence and host adaptation. We have previously noted that stable RNA secondary structure near the PCS is related to the acquisition of virulence and have proposed that the secondary structure may promote the insertion of basic amino acids. In this report we evaluate the phylogenetic relationships for three recent isolates of highly pathogenic avian influenza viruses and the possible virulence factors associated with their primary and secondary structure.     

Specificity of IgG and IgE antibody responses to Haemophilus somnus infection of calves

Haemophilus somnus is an important cause of bovine respiratory disease and septicemia with all it's sequelae. The role of immune responses in protection and immunopathogenesis is not well understood. We showed that infection with bovine respiratory syncytial virus (BRSV) 6 days before H. somnus increased clinical scores and levels of IgE antibody to H. somnus over that of infection with H. somnus alone. To determine whether antigenic specificity of IgE responses differed from IgG responses, Western blots were done with sera from the infected calves, at 0 time and at 21 days post infection. Thus each calf was its own control. IgG antibodies recognized primarily a 40 kDa outer membrane protein (OMP) in whole cell H. somnus preparations and a 270 kDa immunoglobulin binding protein (IgBPs) in culture supernatants but generally not the 41 kDa major OMP (MOMP). IgE antibodies recognized primarily the 41 kDa MOMP in whole cell pellet preparations. Results were consistent among calves. With culture supernatants, IgE antibodies recognized both the 270 kDa IgBPs and the MOMP. Since some H. somnus strains from asymptomatic carriers (including strain 129Pt), do not have IgBPs and express a truncated MOMP (33 kDa rather than 41 kDa), reaction of strain 129Pt cells with serum from calves infected with H. somnus or BRSV and H. somnus was studied. IgE did not react with the truncated MOMP even at much lower (1:100) dilutions than in Western blots with virulent strain 2336 (serum dilution of 1:500). Reactions of IgE with the 40 and 78 kDa antigens in strain 129Pt were noted but since the major reactivities with the IgBPs and the MOMP were not detected, this strain may be useful for inducing protective rather than immunopathogenic responses.     

Effects of Hemophilus somnus on the pregnant bovine reproductive tract and conceptus following cervical infusion

Three strains of Hemophilus somnus were infused into the posterior cervix of six pregnant cows. The organism persisted in the cervicovaginal region for eight to 87 days, and at parturition H. somnus was isolated from chorioallantois in four of six cows; placentitis developed, and fetal membranes were retained. All calves were born alive and no H. somnus was recovered from them. One cow died 14 days after parturition. The death was attributed to severe necrotizing metritis; H. somnus was not isolated from the uterus at death, but was isolated from the placenta at parturition and cervicovaginal mucus two days days later.     

Histamine production by Haemophilus somnus

Ten Haemophilus somnus isolates were grown on blood agar plates under a 5% CO2 atmosphere for 48 h. Harvested whole cells were washed and evaluated for the presence of histamine by ELISA. All H. somnus isolates had cell-associated histamine concentrations of between 18.5 and 200 ng/ml. In a separate study, the ability of H. somnus to secrete histamine into BHI growth medium was evaluated using H. somnus strains 8025 and 156A as well as a recent 156A respiratory isolate. Each strain or isolate was grown under various concentrations of CO2 to approximate the CO2 concentration in the bronchi. The histamine content of washed whole cells and medium supernatant were determined at various stages of incubation. Highest histamine concentrations were detected in the recent respiratory isolate; whole cells (225 ng/ml) after 120 h incubation in 15% CO2 and supernatant (1721 ng/ml) after incubation for 41 h in 25% CO2. This study indicates that different H. somnus isolates can produce and secrete histamine which may be enhanced by CO2 concentrations which approximate those in the bronchial tree. Results of this study may partially explain some of the post-vaccination reactions occasionally observed with H. somnus bacterins. Additional studies are needed to determine the actual role of H. somnus-derived histamine in the pathogenesis of bovine respiratory disease and airway hyperresponsiveness.     

An evaluation of the API ZYM system as a means of identifying Haemophilus somnus and related taxa.

The commercially available API ZYM microbiological identification system was evaluated for the rapid identification of Haemophilus somnus. Eighty-seven isolates of the organism had API ZYM profiles which were characteristic. The API ZYM profiles demonstrate clear differences between H. somnus and other genera but suggest a close association to three related organisms. Enzyme activity of H. somnus isolates were similar to organisms identified as Histophilus ovis, Haemophilus agni and strains UQV of Actinobacillus actinoides and Actinobacillus seminis but was clearly different from isolates of Pasteurella haemolytica, Pasteurella multocida, Bordetella bronchiseptica and group EF4. The API ZYM system allowed more rapid identification of H. somnus than conventional biochemical tests and may be a useful adjunct to conventional methods used for identification of H. somnus isolates. The test did not reveal obvious differences between isolates from various anatomic locations.     

Reproductive difficulties in cattle with antibody titers to Haemophilus somnus

A herd of Holstein cows was examined because of suspected embryonic death. Four cows had embryonic loss, and 2 cows had aborted. Paired serum samples were tested for antibodies to infectious bovine rhinotracheitis virus, bovine viral diarrhea virus, and Haemophilus somnus. Of 16 cows, 8 had antibody titers to H somnus greater than 1:1,024, and 3 had greater than or equal to four-fold changes in antibody titers to H somnus. Haemophilus somnus infection was active in this herd and may have been responsible for the herd's reproductive problems.     

Immunological characterization of the major outer membrane protein of Haemophilus somnus

Antigens and molecular mass diversity of the Haemophilus somnus major outer membrane protein (MOMP) were investigated. The molecular mass of the MOMP of 53 strains of H. somnus varied from 43 to 33 kDa and four MOMP MAb reactivity patterns were detected in immunoblot analysis and immunodot assay. The molecular mass and MAb reactivity data were used for preliminary grouping of H. somnus strains. Disease strains fell into groups 1 and 3, including two of three Group 3 subgroups, whereas strains from asymptomatic carriers were found in all the four groups and three subgroups. Immunoblot analysis with convalescent phase serum showed strain specific reactivity with MOMPs from three isolates used to reproduce disease in cattle. The reaction with the MOMP was only detectable at dilutions of 1:100 or less, whereas the same convalescent sera showed strong reactivity at dilutions of 1:1000 (or more) with other H. somnus antigens. The data suggest that the bovine immune response to the MOMP during infection is weak and is directed to antigenically variable determinants in a strain-specific manner. This may be important in evaluating the role of the antibody response to MOMPs in protective immunity.     

Adherence of Haemophilus somnus to tumor necrosis factor-alpha-stimulated bovine endothelial cells in culture.

Vascular thrombosis and tissue infarction is a principal lesion in Haemophilus somnus septicemia known also as thrombotic meningoencephalitis. This study was undertaken to examine whether tumor necrosis factor-alpha (TNF-alpha) can influence the adherence of H. somnus to cultured bovine aortic endothelial cells (BAEC). Confluent BAEC were exposed to 0-100 nM of human recombinant TNF-alpha for 12-48 h. Suspensions of different strains of H. somnus (approximately 1.5-3 x 10(8) labelled with [methyl-3H]-thymidine, were added to BAEC and incubated for 1.5 h. Initial studies with one pathogenic (P) strain and one non-pathogenic (NP) strain revealed that both strains adhered to normal endothelial cells but minimally to subendothelial matrix remaining after removal of BAEC. Adherence to BAEC was reduced by an excess of unlabelled H. somnus of the same strain. Adherence was enhanced for both strains by exposure of BAEC to TNF-alpha in a manner that increased with TNF-alpha concentration and with duration of exposure to TNF-alpha prior to addition of bacteria. A survey of adherence of six live P strains and six NP strains demonstrated considerable variation but no difference in adherence between P and NP strains to normal or to TNF-alpha-stimulated BAEC. However, TNF-alpha consistently increased adhesion of each strain to BAEC. Both P and NP strains caused more severe cytotoxic changes in TNF-alpha-treated BAEC. Tumor necrosis factor-alpha also increased adhesion of formalin-killed bacteria of P and NP strains.(ABSTRACT TRUNCATED AT 250 WORDS)