Canada: Serotyping of Haemophilus parasuis field isolates from diseased pigs in Quebec by indirect hemagglutination assay and multiplex polymerase chain reaction (PCR)

Haemophilus parasuis, the etiological agent of Glässer’s disease, is one of the most important bacterial pathogens of swine (1). It usually causes polyserositis, but infection may also result in arthritis, meningitis, and septicemia (sudden death) (1). Haemophilus parasuis also contributes to porcine respiratory disease complex, one of the leading causes of mortality in grower-finisher pigs worldwide (1). Although considered as a primary cause of pneumonia (1), strains isolated from lungs have been described as being different from those recovered from systemic infections (2). In addition, low or non-virulent strains have also been isolated from healhy lungs and bronchoalveolar fluid (1,3).

Serotyping of this pathogen is important for decisions on vaccination strategy to prevent future outbreaks as bacterins which are usually used induce mainly serotype-specific immunity. The original serotyping scheme, based on reactions between antisera and surface antigens, classifies H. parasuis into 15 serotypes. Originally, gel immunodiffusion assay (GID) was used to describe these 15 serotypes and to perform epidemiological studies. However, the arrival of the indirect hemagglutination test (IHA) allowed identification of the serotype of many strains that had been non-typable using the GID. An isolate would usually be reported as non-typable if there was no observable antigen-antibody reaction (4). Also, when several antisera reacted with the same isolate, it would be considered as non-typable or it would be assigned the serotype indicated by the strongest agglutination reaction (4), depending on the laboratory in which the serotyping was carried out. Using these methods, distribution of serotypes in various countries showed that, in general, the prelevant serotypes were (the order may vary depending on the country): 2, 5, 4, 7, 13, and 14 (5). More specifically, the sole serotyping report from Canada (Quebec) more than 10 y ago showed that serotypes 4 (27%), 5 (15%), 13 (14%), and 7 (12%) were the most prevalent (6). In the same study, serotypes 4 (25%), 12 (23%), and 5 (15%) were shown to be mainly present among US strains (6). During the last few years, a relatively high percentage of untypable and/or cross-reacting strains have been reported elsewhere (5) as well as in isolates from Quebec (Laboratoire d’expertise en pathologie animale du Québec, LEPAQ, unpublished data). Based on the concept that the capsule locus is responsible for the phenotype of the capsule, a single-step multiplex polymerase chain reaction (mPCR) was developed (4). Fourteen out of 15 serotypes could be individually identified, as no gene could be identified to differentiate between serovars 5 and 12, as shown by whole-genome sequences of these serovars (7). In the present study, strains recovered from diseased pigs in Quebec were serotyped using IHA and a modified version of the mPCR test.

A total of 90 isolates of H. parasuis recovered from diseased pigs in Quebec were included in this study: 25 isolates were recovered from systemic diseases (polyserositis, meningitis, arthritis), 62 isolates were obtained from lungs of animals with respiratory disease, and 3 isolates had no accompanying information. The IHA test was conducted as previously described (6). The mPCR test was a modification of a previously published protocol (4). In our hands, results from the the original single-step mPCR were sometimes difficult to interpret because many bands were closely clustered. Hence, the test was modified in a 3-step mPCR to clearly differentiate the serotypes. Briefly, a loopful of bacteria was suspended in 100 μL of Instagene Matrix (Bio-Rad, Mississauga, Ontario) and manufacturer’s instructions were followed. A 3-μL volume of Instagene extract for each sample was added to each of the 3 mPCR mixtures, which included 12.5 μL of Multiplex PCR kit (2X master mix; Qiagen, Toronto, Ontario), 2.5 μL of the primer mix (PM)1, PM2, or PM3, and 8 μL of UltraPure H₂O. The primer mixes consisted of reverse and forward primers that target the following genes:

• PM1: funB, glyC, wciP, funQ, funAB, and HPS_219690793;

• PM2: wzx, funV, gltP, funI, and HPS_219690793; and

• PM3: wcwK, gltI, scdA, funX, amtA, and HPS_219690793.

The mPCR mixtures were heated at 95°C for 15 min, followed by 30 cycles of 94°C for 30 s, 58°C for 90 s, 72°C for 90 s, and a final extension at 72°C for 10 min. Expected results for each serotype are shown in Figure 1.

Figure 1.

Band patterns for the molecular serotyping polymerase chain reaction for all 15 serovars of H. parasuis with primer mix (PM)1, PM2, and PM3. Lane L, 100-bp DNA ladder; S1 to S15 represent the 15 serotypes of H. parasuis.

Serotyping results for both tests are shown in Figure 2. Of the 53 isolates that could be serotyped by the IHA test, only 3 gave a result that was different from that with the mPCR (94.3% agreement). A high concordance for typable isolates by both techniques has previously been described (4). However, 37.8% of isolates in this study were untypable by IHA, but all of these could be serotyped by mPCR. A drastic reduction in untypable isolates by using PCR has been recently reported (4,5). More than half of serotype 7 isolates (52.9%) could not be originally serotyped by IHA; the remaining isolates belonged to serotypes 2 (17.6%), 5/12 (17.6%), and 3, 6, 8, and 13 (2.9% each serotype). Failure in serotyping of these strains by IHA may be due to factors such as a lower sensitivity of the serological test or lack of phenotypic expression of the capsule (4).

Figure 2.

Serotype distribution of 90 Canadian isolates of H. parasuis as determined by IHA (black) and mPCR (white).

Using the mPCR technique, serotypes 7 and 5/12 (22.2% each) were the most common serotypes found, followed by serotypes 4 (15. 6%) and 13 (11.1%). Although this is not a prevalence study, it is interesting to note that these 4 serotypes are the same that were most commonly detected in the previous study which also included strains from Quebec (6). Serotypes 2 (7.8%), 1, 3, and 9 (5.6% each one) were also detected. Serotypes 10, 11, and 15 were not found.

The site of isolation may also have some importance, as isolates cultured from sytemic diseases survived serum killing and phagocytosis and might be considered as being more virulent than those recovered from the lungs (4). The serotype distribution of strains isolated from systemic and respiratory disease is shown in Figure 3. Serotypes 5/12, 7, and 2 (in decreasing order) were most frequently detected among strains recovered from systemic disease, whereas serotypes 7 and 13, followed by serotypes 4 and 5/12 were frequently identified among respiratory strains. Interestingly, all serotype 13 strains originated from lungs. However, more studies are needed to establish a possible relationship between the serotype and type of infection (systemic versus respiratory). It would also be interesting to conduct the mPCR serotyping on isolates recovered from healthy pigs.

Figure 3.

Serotype distribution of H. parasuis among systemic (white) and respiratory (black) isolates tested by mPCR.

In conclusion, the modified mPCR is a reliable method for serotyping of H. parasuis. As multiple infections of the same individual and within herds can occur (2), a highly sensitive test (such as the mPCR) may identify additional isolates that contribute to disease in animals that are not commonly investigated. Indeed, and using the mPCR, we have already observed that more than one serotype may be present in the same sample (unpublished observations). Therefore, the molecular serotyping assay used herein represents a significant improvement in the tools available to characterize H. parasuis isolates.