Porcine teschovirus, sapelovirus, and enterovirus in Swiss pigs: multiplex RT-PCR investigation of viral frequencies and disease association

Abstract

Porcine teschovirus (PTV), sapelovirus (PSV-A), and enterovirus (EV-G) are enteric viruses that can infect pigs and wild boars worldwide. The viruses have been associated with several diseases, primarily gastrointestinal, neurologic, reproductive, and respiratory disorders, but also with subclinical infections. However, for most serotypes, proof of a causal relationship between viral infection and clinical signs is still lacking. In Switzerland, there has been limited investigation of the occurrence of the 3 viruses. We used a modified multiplex reverse-transcription PCR protocol to study the distribution of the viruses in Swiss pigs by testing 363 fecal, brain, and placental or abortion samples from 282 healthy and diseased animals. We did not detect the 3 viruses in 94 placental or abortion samples or in 31 brain samples from healthy pigs. In brain tissue of 81 diseased pigs, we detected 5 PSV-A and 4 EV-G positive samples. In contrast, all 3 viruses were detected at high frequencies in fecal samples of both healthy and diseased pigs. In healthy animals, PTV was detected in 47%, PSV-A in 51%, and EV-G in 70% of the 76 samples; in diseased animals, frequencies in the 81 samples were 54%, 64%, and 68%, respectively. The viruses were detected more frequently in fecal samples from weaned and fattening pigs compared to suckling piglets and sows. Co-detections of all 3 viruses were the most common finding. Based on clinical and pathology data, statistical analysis yielded no evidence for an association of virus detection and disease. Further research is required to determine if pathogenicity is linked to specific serotypes of these viruses.

Enteric viruses porcine teschovirus (PTV; Teschovirus A, Teschovirus B), porcine sapelovirus A (PSV-A; Sapelovirus A), and enterovirus G (EV-G)^4,21 belong to the family Picornaviridae (https://talk.ictvonline.org/taxonomy/) and can infect pigs and wild boars. PTV includes 13 serotypes, designated as 1 to 13.⁴⁸ Several novel putative serotypes of PTV have been identified, and a new species, PTV-B, was described in 2019.⁴⁶ For PSV-A, a single serotype (PSV1) has been described; 17 serotypes of EV-G are known to infect pigs (G1–G4, G6, G8–G19).⁴⁸

PTV, PSV-A, and EV-G are distributed worldwide and have been reported in the Americas,^2,3,14,36 Europe,^6–11,34,38 Africa,¹⁷ and Asia.^12,24,29 Infections with PTV, PSV-A, or EV-G have been detected in association with diarrhea^26,28,50 and neurologic,^{3,13,26,36,37,42,44} reproductive,²⁹ and respiratory disorders,^26,28,50 as well as with aphthae-like skin lesions.²⁰ However, subclinical infections have also been described.^{6,7,10,34,40,45} Highly virulent strains of PTV, most notably strains of PTV1, evidentially cause a severe neurologic, and in Switzerland notifiable, disease known as Teschen disease.^1,22 Less-virulent strains of PTV1, as well as other PTV serotypes, can lead to mild neurologic disorders, designated as Talfan disease or benign enzootic paresis.¹ Reported prevalences or frequencies of the 3 enteric viral species have been derived primarily from fecal samples of both domestic pigs and wild boars and show considerable differences: 19.0–61.8% for PTV,^{7,9,14,34,38,47} 2.8–72.8% for PSV-A,^{5,7,9,11,14,34,38,41} and 0–69.4% for EV-G.^{7,14,34,38,41} Despite frequent detection in diseased animals, proof of an association of viral infection with clinical signs is lacking for most serotypes of these 3 viruses.

In Switzerland, only limited investigations of the detection of these viruses have been conducted.^16,35 To gain insight into the occurrence of the 3 viruses, we tested fecal, brain, and placental or abortion samples from healthy and diseased Swiss pigs using an adapted and optimized multiplex reverse-transcription PCR (RT-PCR) protocol that allows simultaneous detection of PTV, PSV-A, and EV-G. Based on the clinical and pathology data on the animals, we investigated a possible statistical association with disease, especially with gastrointestinal, neurologic, and/or reproductive disorders.

Materials and methods

Sample material

Clinical specimens investigated included 3 sample materials from healthy and diseased pigs: fecal, brain, and placental or abortion samples. Because only samples of diseased pigs were available for retrospective testing, healthy animals were newly sampled in 2019. Samples from diseased pigs included paired fecal and brain samples from 81 pigs sent to the Institute of Veterinary Pathology at the University of Zurich between 2014 and 2015. Moreover, 83 placental or abortion samples were analyzed from sows with reproductive disorders, including abortions, mummies, and stillborn and weakly born piglets, that had been acquired for a project in 2009 and were stored at −80°C.¹⁶ Samples from healthy pigs included 31 brain samples from fattening pigs collected by official veterinarians at the municipal slaughterhouse in Zurich in the autumn of 2019, as well as 76 fecal and 11 placental samples from clinically healthy pigs collected by veterinarians of the Swiss Pig Health Service between the summer and winter of 2019. Before RNA was extracted, all sample materials were stored temporarily at −20°C until the samples were processed.

Sample preparation and RNA extraction

Brain or placental or abortion samples (25 mg of each) were mixed with 560 µL of AVL buffer (QIAamp viral RNA mini kit; Qiagen) and disrupted in a TissueLyser II (Qiagen) at 30 Hz for 20 s using a 5-mm stainless steel bead (Qiagen). The tubes were then centrifuged at 6,000 × g for 1 min. For fecal samples, 1 g or 1 mL of feces was mixed with 4 mL of PBS and two 2-mm glass beads (Faust), and vortexed thoroughly. Of this mixture, 1 mL was centrifuged at 6,000 × g for 1 min. RNA was extracted from the supernatant (Qiagen) according to the manufacturer’s instructions. The extracted RNA was stored at −20°C until further use.

PTV, PSV-A, and EV-G multiplex RT-PCR

For amplification of PTV, PSV-A, and EV-G RNA, a nested multiplex RT-PCR protocol⁴⁹ was modified: to reduce the contamination risk, RT-PCR was performed as a single-run assay using only the inner primer pairs (Table 1). For better sensitivity, the cycle number was increased from 35 in the original protocol to 40. Because, according to sequence alignments, the originally published PSV-A primers (pev-8c and pev-8d)⁴⁹ did not fit well on the 3D polymerase gene region of recently detected Swiss PSV-A strains, they were replaced by other PSV-A primers (pev-8g and pev-8h).²³ Like the PTV and EV-G primers,⁴⁹ the pev-8g and pev-8h primers target the more conserved 5′UTR region of the PSV-A genome and aligned perfectly to the genome of Swiss PSV-A strains. Analogous alignments showed that the PTV and EV-G primers⁴⁹ matched the Swiss strains and therefore no adjustment was made for these primers (data not shown).

Table 1.

Primers used for single-run multiplex reverse-transcription PCR for the detection of porcine teschovirus (PTV), porcine sapelovirus A (PSV-A), and enterovirus G (EV-G).

Virus	Gene region	Sequence	Fragment size (bp)	Primer name	Reference
PTV	5′UTR	5′-TGAAAGACCTGCTCTGGCGCGAG-3′	158	pev-1c	49
		5′-GCTGGTGGGCCCCAGAGAAATCTC-3′		pev-1d
PSV-A	5′UTR	5′-ATGGCAGTAGCGTGGCGAGCTAT-3′	212	pev-8g	23
		5′-GTAATGCCAAGAGCATGCGCCA-3′		pev-8h
EV-G	5′UTR	5′-CAAGCACTTCTGTTTCCCCGG-3′	313	pev-9c	49
		5′-GTTAGGATTAGCCGCATTCA-3′		pev-9d

5′UTR = 5′ untranslated region.

The OneStep RT-PCR kit (Qiagen) was used to set up the RT-PCR reaction. The mix had a final volume of 25 µL and included 5 µL of OneStep RT-PCR buffer (5×), 1 µL of dNTP (400 µM dNTP), 0.15 µL of each primer (pev-1c/-1d, pev-8g/-8h, and pev-9c/-9d; 600 nM), 1 µL of OneStep RT-PCR enzyme mix, 0.1 µL of RNase inhibitor (RNasin; Promega), and a top-up of 12 µL of nuclease-free water to which 5 µL of extracted RNA were added. Samples of all groups were tested undiluted and in a 10-fold dilution to exclude PCR inhibition. Thermal cycling was performed (FlexCycler; Analytik Jena) and consisted of a reverse-transcription step at 50°C for 30 min, followed by 95°C for 15 min, 40 cycles of 50 s at 94°C, 50 s at 55°C, and 1 min at 72°C, and a final extension at 72°C for 5 min. After cycling, the samples were cooled to 4°C. The amplified products were analyzed on 2% agarose gel at 80 V for ~70 min. Bands of the expected sizes (Table 1) were excised, DNA was extracted (QIAquick gel extraction kit; Qiagen), and the DNA concentration was measured (NanoDrop 1000 spectrophotometer; Thermo Fisher). Per 100 bp of expected product, 18 ng of DNA were mixed with 3 µL (10 µM) of the respective forward primer, and nuclease-free water was added to achieve a final volume of 15 µL. The samples were sent for sequencing (Microsynth), and the sequences were analyzed with BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi). In case of an unclear sequencing result, the sample was re-sequenced using 3 µL (10 µM) of the corresponding reverse primer, or the RT-PCR was repeated. Given that nonspecific amplifications were observed, only samples that showed a band of the expected size after gel electrophoresis and that yielded sequences with the highest identity to the expected virus were considered positive.

p12S real-time RT-PCR

Amplification of the porcine 12S (p12S) rRNA reference gene served as internal control, to confirm successful RNA extraction. The reaction mix for the real-time RT-PCR (RT-rtPCR) had a final volume of 10 µL and contained 5 µL of RT-PCR (2×) mix (TaqMan RNA-to-CT 1 step kit; Applied Biosystems, Thermo Fisher), 0.5 µL (500 nM) of each primer (Table 2), 0.5 µL (250 nM) of p12S probe (Table 2), 0.25 µL of RT enzyme mix (40×; Applied Biosystems), and a top-up of 2.25 µL nuclease-free water to which 1 µL extracted RNA (10- or 100-fold diluted in nuclease-free water) were added. Thermal cycling included a reverse-transcription step at 48°C for 15 min, followed by 95°C for 10 min, 45 cycles at 95°C for 15 s, and 60°C for 1 min (QuantStudio 7 Flex real-time PCR instrument; Applied Biosystems).

Table 2.

Primers and probe used for the porcine 12S (p12S) rRNA reverse-transcription real-time PCR.

Primer/probe	Sequence
p12S forward	5′-CCACCTAGAGGAGCCTGTTCTATAA-3′
p12S reverse	5′-GGCGGTATATAGGCTGAATTGG-3′
p12S probe	FAM-CGATAAACCCCGATAGACCTTACCAACCC-TAMRA

Clinical and pathology data

Brain samples of healthy pigs were examined histologically to exclude microscopic alterations. For the 81 diseased animals from which paired samples of feces and brains were available, a postmortem examination had been conducted by the Institute of Veterinary Pathology at the University of Zurich. In most cases, autopsy included histologic investigation and further testing for specific pathogens. Clinical data, especially whether the animals had suffered from diarrhea, or neurologic or reproductive disorders, were obtained from the referring veterinarians. Based on the clinical and pathology data, these 81 animals were categorized into 5 main syndrome groups: 1) gastrointestinal tract (GIT) disorders; 2) systemic disorders (SD); 3) musculoskeletal disorders; 4) neurologic disorders; and 5) indeterminate disorders. The GIT and SD groups were further divided into 8 subgroups: 1) GIT, diarrhea; 2) GIT, hemorrhagic bowel syndrome (HBS), 3) GIT, unknown clinical signs/history; 4) SD, with neurologic signs and with diarrhea; 5) SD, without neurologic signs but with diarrhea; 6) SD, with neurologic signs but without diarrhea; 7) SD, without neurologic signs and without diarrhea; 8) SD, without neurologic signs, without diarrhea but with abortion. Postmortem examinations and testing for specific pathogens had also been performed on the fetuses or newborn piglets, from which the 83 placental or abortion samples were derived.¹⁶ Fecal samples from healthy pigs and the 81 diseased pigs with paired fecal and brain samples were divided into 4 categories, according to the age of the animals: 1) suckling piglets < 5 wk old; 2) weaned pigs 5–10 wk old; 3) fattening pigs 11–27 wk old; and 4) adult sows.

Data analysis

Frequencies were calculated on both animal (n = 282) and sample (n = 363) level. On sample level, frequencies were calculated for each of the 6 sample groups (fecal samples of healthy and diseased pigs, brain samples of healthy and diseased pigs, placental samples of sows without reproductive disorders, and placental or abortion samples of sows with reproductive disorders), and where available for the 4 age categories (suckling piglets, weaned pigs, fattening pigs, and sows). Two different kinds of frequencies are described: “overall frequencies” indicate how frequently the individual viruses were detected; “mono-/co-detections” describe the frequencies of all possible detection statuses, including negative, mono-, double- (PTV and PSV-A, PTV and EV-G, or PSV-A and EV-G), and triple- (PTV, PSV-A, and EV-G) detections.

For statistical analyses, we used SPSS Statistics for Windows v.25.0 (IBM); p ≤ 0.05 was considered statistically significant. The Fisher exact test was used to check for differences in the frequencies between the different sample groups, between the different age categories within the same sample group, and between healthy and diseased animals within the same sample material group and age category. To investigate if the presence of viruses may be associated with a specific clinical picture, virus frequencies in fecal samples from diseased pigs, categorized into the above-named main syndrome groups and subgroups, were compared with those from healthy pigs using binary logistic regression analysis. This test was further used to check for an association between etiologic diagnosis within the different syndromes (clear vs. uncertain or unknown etiology) and virus detection to assess whether disease in cases with an uncertain or unknown etiology could be explained by the presence of one or more of the 3 viruses.

Results

Overall frequencies of PTV, PSV-A, and EV-G

We tested 363 samples from 282 animals. PTV was detected in 80 of 282 (28%) animals tested. PTV was only detected in fecal samples, however, at high proportion in both healthy and diseased animals (Table 3). PSV-A and EV-G were detected in 91 of 282 (32%) and 109 of 282 (39%) animals, respectively. Both viruses were detected at high rates in fecal samples from both healthy and diseased pigs. Further, both viruses were detected at low frequencies in brain samples of diseased pigs (Table 3). None of the 3 viruses was detected in 158 of 282 (56%) animals.

Table 3.

Overall virus frequencies within the 6 groups of healthy and diseased pigs.

Virus	Sample group
	Brain samples, healthy pigs (n = 31) (%)	Brain samples, diseased pigs (n = 81) (%)	Fecal samples, healthy pigs (n = 76) (%)	Fecal samples, diseased pigs (n = 81) (%)	Placental samples, sows without reproductive disorders (n = 11) (%)	Placental or abortion samples, sows with reproductive disorders (n = 83) (%)	Total (n = 363) (%)
PTV	ND	ND	47	54	ND	ND	22
PSV-A	ND	6	51	64	ND	ND	26
EV-G	ND	5	70	68	ND	ND	31
Negative	100	89	20	24	100	100	64

EV-G = enterovirus G; ND = not detected; PSV-A = porcine sapelovirus A; PTV = porcine teschovirus.

Comparison of overall frequencies among the age categories

In fecal samples of healthy pigs, the 3 viruses were detected significantly more frequently in weaned and fattening pigs than in suckling piglets and sows (Fig. 1). The same tendency was observed in fecal samples of diseased pigs; however, differences were not always significant (Fig. 2).

Figure 1.

A–F. Comparison of overall frequencies of viral detection in the fecal samples from the 4 age categories of healthy pigs. Significance levels are indicated as follows: *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 2.

A–F. Comparison of overall frequencies of viral detection in the fecal samples from the 4 age categories of diseased pigs. Significance levels are indicated as follows: *p < 0.05, **p < 0.01, ***p < 0.001.

In brain samples of diseased pigs, PSV-A was detected in 1 of 26 (4%) suckling piglets, 2 of 26 (8%) weaned pigs, and 2 of 27 (7%) fattening pigs. EV-G was found in 1 of 26 (4%) weaned pigs, in 2 of 27 (7%) fattening pigs, and in 1 of 2 (50%) sows. No significant difference in the overall frequencies was observed among the different age categories.

Comparison of mono-/co-detection frequencies among the age categories

Regardless of age, in fecal samples of healthy and diseased pigs, coinfections with all 3 viruses were most frequently detected (27 of 76 [36%] and 36 of 81 [44%], respectively). Comparison of mono-/co-detection among the 4 age categories of fecal samples of healthy pigs yielded similar results as described above: again, no significant differences were observed between suckling piglets and sows, nor between weaned and fattening pigs. The only statistically significant differences were observed for triple-detection, with weaned and fattening pigs being positive significantly more often for all 3 viruses than suckling piglets (p < 0.001 and p = 0.001, respectively) and sows (p < 0.001 and p < 0.001, respectively; Fig. 3).

Figure 3.

Comparison of mono-/co-detection frequencies in fecal samples from the 4 age categories (A–D) of healthy pigs.

In fecal samples of diseased pigs, simultaneous detection of the 3 viruses occurred significantly more often in fattening pigs than in suckling piglets (p < 0.001; Fig. 4A, 4C), and the simultaneous presence of PSV-A and EV-G was significantly more likely in weaned pigs than in fattening pigs (p = 0.011; Fig. 4B, 4C). The remaining combinations did not reveal any significant differences.

Figure 4.

Comparison of mono-/co-detection frequencies in fecal samples from the 4 age categories (A–D) of diseased pigs.

Notably, no co-detections were observed in brain samples of diseased pigs.

Comparison of overall and mono-/co-detection frequencies between healthy and diseased animals

When comparing the overall virus frequencies of fecal, brain, and placental or abortion material between healthy and diseased animals, no significant differences were detected, regardless of age. Within the different age categories, only fecal samples of healthy weaned pigs showed a significantly higher EV-G as well as triple-detection frequency than diseased pigs of the same age (p = 0.029 and p = 0.032, respectively; Figs. 3B, 4B).

Assessment of a possible association of virus detection with disease

Binary logistic regression analysis was only performed on fecal samples, given that in the other groups the viruses were either not detected or detected at a very low rate. Virus frequencies within the main syndrome groups and subgroups of diseased pigs were compared with those of healthy pigs (Table 4). The binary logistic regression model did not reveal any association of the main syndromes with the presence of virus. Within the different subgroups of GIT and SD disorders, no statistically significant results were found except for the syndrome “SD, without neurologic disorder and without diarrhea”. The logistic regression model was statistically significant only when including EV-G (χ²(2) = 7.049, p = 0.029), indicating an association with this clinical picture and the presence of EV-G (Wald(1) = 5.173, p = 0.023), with Nagelkerke R² value of 18.5%. For the same syndrome, the regression model also showed statistical significance (χ²(1) = 3.875, p = 0.049) when simultaneous detection of PSV-A and EV-G was included (Wald(1) = 3.910, p = 0.048), with Nagelkerke R² value of 10.4%. However, when including all 3 viruses in the model, significant differences were no longer observed.

Table 4.

Overall virus frequencies in fecal samples from 81 diseased pigs in the 5 main disease syndrome groups and the 8 subgroups.

Virus	Syndrome group
	GIT disease; overall (n = 33) (%)	GIT disease; diarrhea (n = 25) (%)	GIT disease; HBS (n = 7) (%)	GIT disease; unknown (n = 1) (%)	SD; overall (n = 36) (%)	SD; neurologic disease Yes, diarrhea Yes (n = 2) (%)	SD; neurologic disease No, diarrhea Yes (n = 4) (%)	SD; neurologic disease Yes, diarrhea No (n = 22) (%)	SD; neurologic disease No, diarrhea No (n = 7) (%)	SD; neurologic disease No, diarrhea No, abortion Yes (n = 1) (%)	Musculoskeletal disease (n = 5) (%)	Neurologic disease (n = 4) (%)	Indeterminate disorder (n = 3) (%)
PTV	61	56	86	ND	44	ND	50	50	43	ND	80	50	67
PSV-A	67	68	71	ND	56	50	75	59	43	ND	60	100	100
EV-G	82	80	86	100	56	ND	75	68	29	ND	60	100	33
Negative	18	20	14	ND	33	50	25	3	43	100	20	ND	ND

EV-G = enterovirus G; GIT = gastrointestinal tract; HBS = hemorrhagic bowel syndrome; ND = not detected; PSV-A = porcine sapelovirus A; PTV = porcine teschovirus; SD = systemic disease.

For the animals for which a postmortem examination had been conducted, according to pathology data, a clear etiology had been stated in 49 of 81 (60%) cases; in 32 of 81 (40%) cases no etiologic diagnosis (uncertain or unknown) could be stated by the pathologists, including all cases of HBS. No significant differences were found in virus occurrence in cases with an uncertain or unknown etiologic diagnosis compared to cases with a clear etiologic diagnosis within the different syndrome groups and subgroups.

Discussion

We found that PTV, PSV-A, and EV-G are commonly present in fecal samples of healthy as well as diseased Swiss pigs, confirming results from other countries.^{7,11,14,34,38} In brain samples, PSV-A and EV-G were detected at low rates and only in diseased pigs. Notably, despite frequent detection of PTV in feces, no PTV was detected in any of the brain samples tested in our study. Bearing in mind that highly virulent PTV strains most probably would also be detectable in the brain after enteric infection, there was no evidence of the reportable Teschen disease. We therefore conclude that the PTV strains circulating in Swiss pigs are apathogenic or of low virulence and restricted to enteric infection, as has been described in other countries.^1,7,9,30

All placental samples from sows without reproductive disorders, as well as placental or abortion samples, tested negative for the 3 viruses. The placental or abortion samples had been collected previously for a study of infectious reproductive disorders in Swiss pigs.¹⁶ According to the authors, 44 of 286 samples were randomly tested for PTV, PSV-A, and EV-G by RT-PCR.⁴⁹ PTV was found in 1 of 44 (2%) samples and EV-G was found in 4 of 44 (9%) samples; no PSV-A was detected.¹⁶ Unfortunately, we could not trace whether the 83 samples tested in our study were represented in this previous study. In any event, the reported frequencies could not be confirmed here. Suboptimal sample quality and RNA degradation over the years cannot be fully excluded. However, detection of a reference gene in every sample clearly argues against RNA degradation.

Whereas no significant differences in virus frequencies were observed in the brain samples among the different age categories, all 3 viruses were detected more frequently in fecal samples of weaned and fattening pigs compared to suckling piglets and sows. Among weaned and fattening pigs, simultaneous detection of all 3 viruses was the most common finding. Our results are in agreement with previous studies.^{11,14,17,34,41,47} The increasing virus frequency with age may be explained by the presence of protective maternal antibodies in suckling piglets and because sows in general show less infection with virus, most probably as a result of protective immunity, leading to lower infection pressure for the piglets. In contrast, weaned and fattening pigs are more prone to viral infections given stressful or immunocompromising situations, such as the decline of maternal antibody titer after weaning, moving into new premises, or overcrowding. However, in some studies, high EV-G or PSV-A frequencies were described in suckling piglets,^40,45 which might be ascribed to a lack of maternal antibodies.

No significant differences in virus frequencies between healthy and diseased pigs were observed, except for higher frequencies of EV-G and triple-detection in healthy weaned pigs compared to diseased animals. Similar results were observed in a metagenomic study conducted in Swiss piglets, in which EV-G was detected more often in fecal samples of healthy piglets compared to diseased animals,³⁵ and in a study in which higher PTV rates were observed in healthy pigs compared to diarrheic pigs.⁴⁷ However, dilution of the virus in diarrheic feces below the detection limit cannot be ruled out in these cases.

Statistical analyses did not reveal any association of the different syndromes and virus detection in fecal samples. For one syndrome (SD, without neurologic signs and without diarrhea), the logistic regression model was statistically significant when including EV-G only or EV-G and PSV-A. However, when including all 3 viruses in the model, significant differences were no longer found, indicating that these few significant findings with weak associations may be the result of confounders. Also, according to statistical analysis, cases with an unclear etiologic diagnosis could not be explained by detection of 1 or more of the 3 viruses. Our results are in agreement with studies in which no association was found between diarrhea and detection of PTV, PSV-A, or EV-G in fecal samples,^17,34,40 and no association was evident between paraplegia and PSV-A detection in rectal and nasal swabs.⁵ Notably, our results also confirm findings of the metagenomic study of Swiss piglets, in which no significant association was found of PTV or PSV-A with diarrhea.³⁵ In contrast, in other studies, an association was observed of virus detection with diarrhea, or neurologic, respiratory, or reproductive disorders.^13,22,36 Also, after experimental infections with PTV, PSV-A, or EV-G, diarrhea^{19,26,28,31,50} or neurologic,^{15,26,31,43,45} respiratory,^26,50 or reproductive disorders¹⁸ were reported. However, experimental observations must be interpreted with caution given that they do not reflect reality in the field concerning route of infection, infection dose, and other important factors, such as immune status of the animal, other pathogens, and the environment. It has been hypothesized that pathogenicity of PTV, PSV-A, and EV-G may be linked to specific, more virulent serotypes or subtypes.^18,22,27,31 A major limitation of our study lies in our inability to serotype or subtype. This issue must be addressed in future studies by further characterizing samples that initially tested positive by our broad-range multiplex RT-PCR using specific tests targeting less-conserved regions,^{9,10,25,40,47} or by subjecting them to next-generation sequencing (NGS). However, our primary goal was to gain insight into the overall occurrence of the 3 viruses in Swiss pigs.

Next to infection with a more virulent serotype, the clinical outcome may be further influenced by other factors such as coinfections with multiple serotypes, multiple strains, or other enteric pathogenic agents modifying the virulence of the host immune response.^32,33 Indeed, our results suggest the presence of several serotypes or strains in some samples. Among the 81 diseased animals with paired fecal and brain samples, all animals that tested positive for PSV-A in the brain samples (5 animals) had a positive result also in the corresponding fecal sample. The same was true for 3 of the 4 diseased animals that tested positive for EV-G in the brain samples. Viral sequences of these 8 animals were aligned (data not shown). Although some alignments showed no difference in nucleotides between the brain and the fecal sample, other sequences showed up to an 11-nucleotide difference in the highly conserved 5′UTR region, indicating that the respective animal was infected with several different virus strains or serotypes, or mutations or recombinations of the strain occurred in the intestine over time.^27,39

Although PTV, PSV-A, and EV-G have often been described in the context of respiratory disease, no lung samples were screened in our study but would clearly represent another interesting tissue for similar investigations. The assay established in our work allows routine testing of a wide variety of pig tissue for the 3 viruses and will, therefore, contribute to our understanding of their role in health and disease.