INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Potomac horse fever (PHF), caused by Ehrlichia risticii, is a well-recognized disease affecting horses (6, 12, 13, 23). The disease was first recognized in 1979 in areas along the Potomac River in Maryland and Virginia. Since then, it has been diagnosed in 41 states in the United States and in Canada and recognized in Europe and other parts of the world (14, 18, 22, 27, 30). PHF occurs seasonally, mostly in the summer months. The disease is characterized by fever, leukopenia, depression, anorexia, and diarrhea (7). Laminitis develops in about 25% of the cases. The mortality may reach as high as 20 to 25%. Recent studies indicate that E. risticii can cross the equine placenta and infect the unborn fetus, causing abortion (5, 17). The natural mode of transmission of the disease remains unknown (11, 14, 16, 26).

Molecular analysis of an E. risticii strain (strain 25-D), originally isolated in 1984 during the early period of recognition of PHF (6), indicated the presence of nine major component antigens (110, 70, 68, 55, 51, 50, 49, 33, and 28 kDa), all of which are apparent surface antigens, as determined by ¹²⁵I surface labeling (9). Humoral immunity is considered important in the host defense against PHF. Infected horses and mice develop a strong immunoglobulin G antibody response and protection against E. risticii infection (8, 15, 21, 24, 28). Passive transfer of horse antisera to E. risticii (25) or mouse antibodies to E. risticii (antiserum or purified immunoglobulin G) (15) protected mice against E. risticii challenge infection, strongly indicating that antibody mediates the immunity. The infected horses develop in vitro neutralizing antibody in their sera by 15 days postinfection, when ehrlichimia starts to decline, and the neutralizing activity continues to rise, reaching a maximum around day 25 postinfection (19, 25). However, there is no correlation between the presence of high antibody titers and the neutralizing capacity of the antisera. Also, the relationship between the presence of in vitro neutralizing antibody and immunoprotection against the infection is not known.

Currently, three inactivated vaccines for PHF are commercially available. All three vaccines are made with inactivated whole organisms of one strain of E. risticii which was isolated from a Maryland horse in 1984 (called the Illinois isolate, it has been deposited with the American Type Culture Collection [ATCC]; this isolate is not the same as strain 25-D). Although the commercial vaccines have been on the market since 1987, and are being widely used in areas of endemicity, the efficacy of one vaccine has been reported to be marginal (20, 32). Systematic studies on the antibody response of horses in the field to vaccination are not available. For the past several years, there have been consistent reports of vaccine failures in the field, particularly in the areas of endemicity (4, 10). E. risticii isolates with different morphologies, antigenic compositions, and 16S rRNA gene sequences have been reported (4, 31). A new strain of E. risticii was isolated in 1990 (90-12 strain) from a vaccinated horse suffering from clinical PHF and with a high titer of antibodies in its acute-phase serum (10). Studies indicated that the 90-12 strain is a variant having pathogenic, immunologic, and molecular differences from the original 25-D strain (28). Mice immunized with the 25-D strain achieved homologous protection but were only partially protected against challenge with the 90-12 strain, whereas mice immunized with the 90-12 strain were completely protected against the homologous and 25-D strain challenge. There was a two- to fourfold difference between the homologous and heterologous antibody titers. In an in vitro neutralization assay, sera from the strain 25-D-infected horse neutralized the homologous strain but did not neutralize the 90-12 strain, whereas sera from the strain 90-12-infected horse neutralized both strains. A major difference in their antigenic composition is that the 25-D strain contains the 50-kDa antigen (but not the 85-kDa antigen), whereas the 90-12 strain contains the 85-kDa antigen (but not the 50-kDa antigen). The recombinant clone-specific antibodies of either of these two antigens react with both of the antigens (2, 29). This indicated that these two antigens are homologs and are considered strain specific. The strain-specific antigen (SSA) is highly immunogenic and has been determined to be a protective antigen (29). Mice immunized with the recombinant 85-kDa SSA are largely protected against challenge with the 90-12 and 25-D strains, whereas mice immunized with the recombinant 50-kDa SSA are protected against challenge with the 25-D strain but not challenge with the 90-12 strain, thus indicating the strain specificity of the protection (2, 29). The genes of the 50-kDa SSA of the 25-D strain and of the 85-kDa SSA of the 90-12 strain differ in size (1.6 and 2.5 kb, respectively), nature of tandem repeats of nucleotide sequences, and profile of deduced amino acid domains. Both 50- and 85-kDa SSAs have eight common amino acid domains but differ in the order of their arrangement, and they have two and six unique amino acid domains, respectively (2). The SSA of the ATCC vaccine strain is mostly similar to that of the 25-D strain, in molecular size (50 kDa) and in nucleotide and deduced amino acid sequences, except that the gene size is 1.5 kb (2).

This paper presents the occurrence of clinical PHF in vaccinated horses and describes the heterogeneity of E. risticii isolates obtained from these infected horses. It also reports the low antibody response in horses to field and experimental PHF vaccination and lack of immunoprotection of the experimentally vaccinated horse.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Collection of specimens from horses suffering from clinical disease. The study was performed for three summer seasons, from 1994 to 1996, in collaboration with the Equine Medical Center, Leesburg, Va., and a few practicing veterinarians in the areas of endemicity in Maryland and Virginia. Heparinized blood and sera were collected from horses suspected of having PHF during the acute stage of the disease. Convalescent-phase sera were collected approximately 2 weeks later. The vaccination history with the commercial PHF vaccine, clinical signs, hematological reports, and treatment records of these horses were provided by the participating veterinarians.

Experimental vaccination and challenge of horse. One horse was experimentally vaccinated three times at 3-week intervals with vaccine 1 (PHF-Vax; Schering-Plough Animal Health, Kenilworth, N.J.). Serum samples were collected at 2-week intervals. Four weeks after the final vaccination, the horse was challenged by intravenous injection of 10 ml of strain 90-12-infected P388D1 cells (10⁶/ml). The horse was observed for clinical signs and hematological profile for 4 weeks postchallenge, and serum samples were collected at 2-week intervals. Heparinized blood samples were collected at the onset of clinical signs for the isolation of E. risticii in cell culture and for PCR detection of the organism.

Laboratory diagnosis of PHF and isolation of E. risticii. PCR with blood mononuclear cells was performed with genomic primers as described previously (1). Isolation of E. risticii from mononuclear cells of infected horses in P388D1 mouse macrophage cells was performed as described before (6). The detection of anti-E. risticii antibodies in the acute- and convalescent-phase sera was done by immunofluorescence assay (IFA) (6).

IFA. IFA was performed by the procedures described previously (6) for the detection of anti-E. risticii antibodies in the sera. The IFA was also used for determining the immunological relationship among the new E. risticii isolates and the known 25-D and 90-12 strains. An infected cell culture of each isolate was reacted with the homologous and heterologous convalescent-phase sera, and end-point antibody titers were determined. The antibody titers of the heterologous convalescent-phase sera in reaction with an isolate were compared to the titers of those sera in reaction with their homologous isolates and determined to be higher, equal, or lower in value. The number of heterologous convalescent-phase serum samples that reacted with each isolate for each of the three categories was determined and expressed in a percentage.

In vitro neutralization assay. In vitro neutralization of cell-free E. risticii with horse antisera by using a mouse macrophage P388D1 cell culture was performed as described previously (28).

Western blotting. The Western blot procedure has been described previously (8). Infected cell culture materials of the known strains and new isolates were reacted with polyclonal mouse or horse antisera and monospecific mouse anti-85-kDa SSA serum.

DAF. DNA amplification fingerprinting (DAF) was performed according to the procedure described previously (3). Briefly, a part of the SSA gene of E. risticii isolates was PCR amplified by using two primers. One primer was selected from a nonvariable upstream region from the start signal of the SSA gene (85-kDa SSA gene of the 90-12 strain). The second primer was selected from unique, highly variable region at the 3' end of the gene. This primer sequence is present in each of these homologs at different locations and does not fall within the repeated sequences. Depending on the number of tandem repeat sequences between the two primer sites, which varies among the E. risticii strains, the target size of the PCR amplification varies for different strains of E. risticii.

Field vaccination of horses with commercial PHF vaccines and collection of sera. A field vaccination study of horses in areas of endemicity was conducted in the 1995 season in collaboration with practicing veterinarians. None of these horses had any previous history of PHF. The vaccination study was done with two commercial PHF vaccines, vaccine 1 (PHF-Vax; Schering-Plough Animal Health) and vaccine 2 (Potomvac; Rhone-Merieux, Athens, Ga.), dictated by their use by the participating veterinarians. Horses were vaccinated by the participating veterinarians independently with the vaccine(s) of their choice used according to the manufacturer's recommendation and their own vaccination schedule. Serum samples were collected by the veterinarians at monthly intervals for 7 and 6 months for vaccines 1 and 2, respectively, and antibody titers were determined by IFA.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Vaccine failure in horses and confirmation of PHF. In limited studies in the past 3 years (1994 to 1996 seasons), 43 cases of PHF were confirmed (Table 1). Laboratory confirmation was made by positive results in at least two of the three following procedures: detection of E. risticii DNA from the mononuclear cells by PCR, isolation of the organism from the mononuclear cell in cell cultures, and rise in the antibody titers of fourfold or greater in serum between the acute and convalescent phases. The horses with confirmed PHF showed typical clinical signs of PHF with a varying degree of severity, but the general opinion of the practicing veterinarians was that in many cases, the clinical signs were relatively less severe than those of nonvaccinated horses with clinical PHF. Altogether, 28 new E. risticii isolations were made, 14 in 1994, 5 in 1995, and 9 in 1996. The titers of anti-E. risticii antibodies in the acute-phase sera at the time of isolation of the organism were relatively high, ranging from 8 to >10,240. There were three cases of mortality from PHF. Of the 43 horses with PHF, 38 (89%) had been vaccinated for the respective season with one of the three commercially available vaccines. Twenty-nine of these horses were vaccinated once, eight were vaccinated twice, and one horse was vaccinated three times in its respective year of study. Since all of these horses were from areas of endemicity, almost all had been routinely vaccinated in the years previous to their respective year of study.

Heterogeneity of the new E. risticii isolates. The 1994 E. risticii isolates have been studied in some detail. Studies of these 14 isolates show that they are heterogeneous in character among themselves and have differences with our two known strains, strains 25-D and 90-12.

View larger version (96K): [in this window] [in a new window]   FIG. 1.   The 50/85-kDa SSA gene-based DAF profile of the new 1994 isolates showing variation in the size of the amplified products. The PCR-amplified products from the 1994 isolates and the 90-12 and 25-D strains of E. risticii were electrophoresed on a 1% agarose gel. Lanes: 1 and 19, DNA molecular size markers; 2, 25-D strain; 3, 94-2; 4, 94-3; 5, 94-24; 6, 94-27; 7, 94-30; 8, 90-12 strain; 9, 94-22; 10, 94-25; 11, 94-29; 12, 94-8; 13, 94-28; 14, 94-31; 15, 94-37; 16, 94-49; 17, 94-50; 18, negative control. The ATCC vaccine strain is not shown due to lack of lanes. However, the molecular size of the DAF product of the strain was verified several times to be 1.75 kb. The tabular data divides the isolates into six groups based on their amplified DNA fragment sizes. Groups representing the three known strains are indicated. The last row of the tabular data indicates the total number of isolates in each group; the percentage representation of the total number of all isolates is shown in parentheses.

View larger version (72K): [in this window] [in a new window]   FIG. 2.   Western blots of new 1994 isolates reacted with the monospecific antisera against the recombinant 85-kDa SSA, showing SSA bands of various sizes ranging from 48 to 85 kDa. Bands are visible for 85-kDa SSA (isolate 94-29 and strain 90-12), 60-kDa SSA (isolates 94-22 and 94-25 [the latter also has an SSA band of 58 kDa]), 58-kDa (isolate SSA 94-31), 50-kDa SSA (isolates 94-2, 94-3, 94-8, 94-24, and 94-27 and strain 25-D and the ATCC vaccine strain), and 48-kDa SSA (isolates 94-28, 94-37, and 94-50). Isolate 94-25 containing 60-kDa SSA was from a mixed culture with an isolate containing 58-kDa SSA. This mixed culture was confirmed by PCR, since amplification of the full-length gene with primers ECP-1 and ECP-2 resulted in two distinct fragments of the corresponding gene sizes. It is important to note that since crude infected cell culture preparations were used, products resulting from proteolytic degradation caused the appearance of multiple bands. In each lane, the major band with the largest molecular size was considered to represent the size of the SSA. To confirm the assigned molecular size of each SSA, the corresponding complete gene was amplified with primers ECP1 and ECP2 (3). The sizes of amplified fragments correlated with the gene sizes calculated on the basis of the assigned molecular size (data not shown).

Competency of the experimental vaccination of a horse. The vaccinated horse developed a maximum antibody titer of 160, as detected by IFA. In vitro neutralization antibody to the homologous ATCC vaccine strain or any of the new field isolates was not detected. Also, antibody to the 50/85-kDa SSA antigen of either the 25-D or 90-12 strain was absent in the sera as determined by Western blotting. Upon challenge with the 90-12 strain, the horse developed clinical signs of PHF. E. risticii DNA was detected in the mononuclear cells by PCR, and the organism was isolated from the mononuclear cells in cell cultures. The IFA antibody in the postchallenge sera increased to a high titer, 10,240. These sera also contained neutralizing antibody, and there was production of antibody to the 50/85-kDa SSA antigen.

Antibody response to field vaccination of horses. The field vaccination studies were performed with two vaccines, 1 and 2 (Tables 4 and 5). Of the total 41 horses, 5 were vaccinated once, 20 were vaccinated twice, and 16 were vaccinated three times. The majority of these horses had also been vaccinated in the previous year(s). The IFA antibody titers were relatively low, ranging from 40 to a maximum of 1,280. In many cases, there was no substantial increase in the antibody titers after the booster vaccination(s). For several horses whose antibody titers were high throughout the collection period, they had entered the collection period with relatively high titers. Horse 24 had the highest titer of all the vaccinated horses, but prevaccination serum from that horse was not available for comparison. Since none of the other horses showed such an antibody response to vaccination, it is highly possible that this horse could have had a clinical or subclinical E. risticii infection in the previous year(s) which resulted in a high antibody response. If there was any increase in the vaccine antibody titer, it occurred within 1 to 2 months postvaccination. Three horses (horses 20, 37, and 38) from both vaccine groups which had never been vaccinated before responded with very little antibody titer upon repeated vaccinations. None of the sera from either vaccine group contained in vitro neutralizing antibody. Overall, there was no significant difference in the antibody responses of horses treated with the two vaccines.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The present study establishes the occurrence of vaccine failure in horses with PHF. Of the PHF-infected horses studied, 89% of them had been vaccinated for the respective season. The vaccinated horses suffered from clinical PHF, although in many cases the clinical signs were reportedly relatively less severe than the signs in unvaccinated horses. Two commercial vaccines (vaccines 1 and 2) were used for the vaccination of these horses. Both of the vaccines are inactivated products of one (ATCC) strain of E. risticii.

In the field vaccination study, in a majority of the cases, the antibody titers determined by IFA were low, ranging from 40 to 1,280, as compared to antibody titers resulting from infection, which are usually 10,240 or higher. Also, in many cases, there was no or a marginal increase in antibody titers from successive vaccinations. None of the sera from the field vaccination study contained in vitro neutralizing antibody. Further, the antibody titer in response to the experimental vaccination, which was given three times, was low. The horse did not produce in vitro neutralizing antibody. Also, the horse did not produce antibody to the 50/85-kDa SSA which is considered to be a protective antigen of E. risticii. Upon challenge infection with the 90-12 strain, the horse was not protected against the disease. But, following challenge infection, the same horse developed antibodies to the 50/85-kDa SSA and also the neutralizing antibody. In addition, in field-vaccinated horses suffering from clinical PHF, E. risticii was isolated at the acute stages of the disease in the presence of vaccine antibodies. All of these facts suggest a deficiency in the inactivated vaccines, which may possibly be due to denaturation of the protective epitopes of the organism by the chemical used to inactivate the E. risticii.

One of the major findings presented here is the heterogeneity of the new E. risticii isolates obtained from the vaccinated horses with clinical PHF. These new isolates show differences from each other and from the 25-D and 90-12 strains as demonstrated by IFA reactivity pattern, DAF profile, neutralization activity, and most importantly the molecular sizes of the SSAs, which varied from 48 to 85 kDa. These characteristics differentiating the isolates or strains were independent of each other, and there was little correlation among them, emphasizing the heterogeneity of the isolates. The culture of the 94-25 isolate showed SSA bands of 60 and 58 kDa due to a mixed culture of two isolates from the same horse. This mixed culture was confirmed by PCR, since amplification of the full-length gene with an appropriate primer pair (3, 28) resulted in two distinct fragments of the corresponding gene sizes (unpublished data). The unique characteristic of the SSA genes of the 25-D and 90-12 strains is the presence of tandem repeat nucleotide sequences which vary in size, number, and type of repeats (2). The generation of heterogeneous strains may be due to the recombination events occurring in the tandem repeat sequences. In a cross-immunoprotectivity study, the 85-kDa SSA of the 90-12 strain provided protection against the 25-D strain, whereas the 50-kDa SSA of the 25-D strain did not provide protection against the 90-12 strain (29). These results indicate that the SSA is important in providing protection and also that the SSAs of smaller molecular size may not provide protection against the strains containing the SSA of larger molecular size. Since the SSA of the ATCC vaccine strain is very similar to that of the 25-D strain, its immunoprotectivity is expected to be very similar to that of the 25-D strain. Thus, by analogy, it appears that the vaccine strain cannot provide protection against the 90-12 strain and other new isolates of larger molecular sizes, which may be a basis for the vaccine failure. In such a scenario, the 85-kDa SSA of the 90-12 strain, which is the largest-molecular-size SSA known at the present time, can provide protection against most of the E. risticii strains.

Thus, it appears from these studies that both the deficiency in the antibody response of the inactivated vaccines in current use and the antigenic variation of the newly emerging variant strains may be responsible for the present vaccine failure in the field. However, further studies with the vaccine antisera, such as Western blotting analysis of the antisera with the whole organism and with the recombinant 50/85-kDa and other SSAs and passive immunoprotection studies with the vaccine antibodies in mice, are necessary to arrive at any definite conclusions. Similarly, immunoprotection and cross-protection studies with the whole organisms and the SSAs of larger or smaller molecular sizes from different strains are necessary to establish the role of heterogeneous strains in the failure of the vaccine.