Influenza D in Italy: towards a better understanding of an emerging viral infection in swine

Influenza D virus (IDV), a new member of the Orthomyxoviridae family, was first reported in 2011 in swine in Oklahoma, and consequently found in cattle across North America and Eurasia. To investigate the circulation of IDV among pigs in Italy, in the period between June 2015 and May 2016, biomolecular and virological tests were performed on 845 clinical samples collected from 448 pig farms affected by respiratory distress located in the Po Valley. Serological tests were conducted on 3698 swine sera, including archive sera collected in 2009, as well as samples collected in 2015 from the same region. Viral genome was detected in 21 (2.3%) samples from 9 herds (2%), while virus was successfully isolated from 3 samples. Genetic analysis highlighted that Italian swine IDVs are closely related to the D/swine/Oklahoma/1334/2011 cluster. Sera collected in 2015 showed a high prevalence of IDV antibody titers (11.7%), while archive sera from 2009 showed statistically significant lower positivity rates (0.6%). Our results indicate an increasing epidemiological relevance of the pathogen and the need for in-depth investigations towards understanding its pathogenesis, epidemiology and possible zoonotic potential of this emerging virus.

isolation 15 . Moreover, in the same period and region, IDV virus was detected in 27 (8.1%) out of 332 cattle herds investigated for respiratory pathogens 20 .
This study was conducted in the area of Northern Italy, the Po Valley, one of the most important swine-producing regions in Europe with more than 5 million pigs reared annually in approximately 7,200 herds, where IDV was isolated in 2015 15 (Table 1). The farms were located in Emilia Romagna (n.3), Lombardia (n.4) and Veneto (n.2). Viral isolation was obtained from 1 oral fluid and 2 nasal swabs from 3 herds. Accordingly, virus strains were named D/swine/Italy/199724-3/2015, D/swine/Italy/354017/2015 and D/swine/ Italy/173287/2016. Virus isolation was obtained in Human Rectal Tumor (HRT-18G) cells both with and without trypsin treatment.
To better correlate respiratory disease with IDV infection, samples from IDV-positive herds, were also tested against other common swine respiratory diseases pathogens, while other epidemiological information were also incorporated in the investigation. The type and the location of these farms, as well as the disease reported and the investigations performed for the detection of other pathogens are summarized in Table 1.
An additional 54 clinical specimens from farms with respiratory disease collected in 2013 and 2014 were also screened by Real Time RT-PCR. No IDV viral genome was detected in any sample from this period.
Serological results. A total of 3106 swine sera collected at slaughter in 2015 from 143 herds were screened by hemagglutination inhibition (HI) test for antibodies against IDV. The HI test performed on the 3106 sera revealed the presence of antibodies against IDV in 364 samples (between-animal prevalence = 11.7% [CI95%: 10.6-12.9%]) from 74 herds (between-herd prevalence = 51.8% [CI95%: 43.6-59.8%]). As shown in Fig. 1, positive samples had HI antibody titers between 20 and 640. To exclude possible presence of non-specific inhibitors in low dilution of the tested sera, HI antibody specificity of titers = 20 was confirmed by microneutralization assays.
In addition, HI tests were performed on 90 sera collected from two herds (namely, farms B and C in Table 1) that were previously tested positive for IDV. In these herds 26 out of 90 sera were found sero-positive (prevalence = 28.9% [CI95%: 20.5-39.0%]). Among the positive samples, 15 sera had HI antibody titers of 20 (57%), 6 samples had titers of 40 (23%) and 5 showed titers of 80 (19%). The distribution of antibody titers in farms tested positive for IDV did not display significant difference with respect to the distribution of antibody titers in farms with positive sera from the serological screening (Wilcoxon rank sum test; W = 76324; p-value = 0.20).
Finally, HI test performed on 502 sera collected in 2009 from herds that had experienced respiratory symptoms identified only 3 samples (from 3 different herds) with low HI antibody titers (20 to 40) against IDV.

Discussion
Although IDV was first isolated in swine, cattle were subsequently suggested as its main reservoir host and all ensuing studies have focused on this species. Some recent publications suggest that IDV has a global distribution in cattle, with high prevalence of sero-positive herds 11,12 , while the virus is commonly detected in diagnostic submissions for bovine respiratory disease 12,17,18 . On the other hand, studies on the diffusion of this newly emerged pathogen in swine are generally missing both in Europe and North America.
Our results clearly demonstrate that IDV is currently circulating in swine herds in Northern Italy, in an area with high concentrations of both swine and cattle herds. However, it is important to notice that the two species are not mixed by farming practices, but one cannot exclude virus mechanical or airborne transportation. Passive surveillance revealed virus presence in 9 of the 448 pig farms investigated (2%). Furthermore, parallel investigation of swine and bovine archival samples should be conducted to fully clarify when the virus was introduced in cattle and whether its presence in swine might be correlated with periodical outbreaks in cattle.
Although similar quantities of low (lungs) and upper (nasal swabs) respiratory tract samples (361 to 350) were examined, the majority of IDV-positive samples were nasal swabs (14 against 3). This finding is in agreement with the results of experimental infections of pigs with IDV conducted by Hause and coworkers 10 , in which virus was only detected in the upper respiratory tract. This confirms that IDV replication occurs mainly in the upper respiratory tract and indicates that nasal swabs are the preferred sample for diagnosis. On the other hand, it is important to underline IDV was also detected in three lung samples, suggesting that infection can also reach the lower respiratory tract either by active virus replication, or possibly passively by muco-ciliary transportation.
The presence and circulation of IDV in Italian pig farms is further highlighted by the serological results presented in this study. Specifically, 36.5% (27/74) of the positive farms examined in 2015, showed a positive rate >25% of the tested sera (Fig. 1). Among these, 17 (63%) showed sera with titer ≥80. Overall, 30 of the 74 positive farms showed sera with positive titers not higher than 20, which is in agreement with the findings of Hause and coworkers 10,11 . This low seroconversion was confirmed by the HI test results in farms known to be infected by IDV, with 57% of the positive sera showing a titer of 20.
Serological data on samples collected from 2009 suggest that IDV emerged in the region recently, while its prevalence has been steadily increasing over the past three to four years. In fact, only 3 of the 502 sera from 2009 showed detectable titers to IDV (prevalence = 0.6% CI95%: 0.2-1.7%). This significant increase in virus circulation is also confirmed by the results of the retrospective virological study that examined samples collected in 2013 and 2014, and in which all 54 samples examined were negative.
The etiological role of the IDV in respiratory distress in the species is an important question posed by the frequent detection of virus and IDV-specific antibodies in swine herds where respiratory is reported. In this study, co-infections with other swine respiratory pathogens were identified in eight out of nine herds IDV-positive herds, while in one farm, otherwise healthy gilts were found positive for IDV alone (Table 1). To offer a definite answer to this question, more pathogenesis and transmission experiments in pigs must be conducted, including infections with IDV alone or combined with other pathogens to investigate possible synergy. While IDV is a novel virus, its importance and potential for interspecies transmission should not be underestimated.
The circulation of at least two distinct IDV lineages has been identified in the U.S. cattle population 12 . Phylogenetic analysis revealed that Italian IDV strains isolated in this study and previous investigations 15 belong to D/swine/Oklahoma/1334/2011 genetic cluster. However, the possible introduction of a second IDV lineage in the country, the endemic status of IDV in cattle and, to a lower degree, in pigs, together with the proven capacity of the virus to reassort 12,14 , underline the need to continuously monitor susceptible species and to track virus evolution. Additionally, the virus seems able to replicate in other animals, including small ruminants 21 , ferrets 10 , guinea pigs 19 and possibly humans. A study by Hause et al. 10 showed a low percentage (1.3%) of positivity in a set of 316 human sera, while a more recent survey in Scotland was unsuccessful in identifying IDV from archived human respiratory samples 22 . However, the sero-prevalence of IDV among cattle workers in Florida was found to be as high as 91% 23 . Therefore, the zoonotic aspect of this emerging pathogen cannot be excluded or neglected and further investigations are required, especially in people with occupational frequent exposure to cattle and swine.  samples (n.361). From the laboratory's archive, an additional 54 nasal swabs collected from the same region in 2013 (n.32) and 2014 (n.22) respectively, were also selected for virological testing.
Serum samples. In the serological screening, 3106 samples coming from 143 herds in Northern Italy (representing ~5% of the 2900 finishing farms in the same region), were randomly collected at slaughter as part of a routine screening program for Aujeszky virus infection, between June and December 2015. The median number of samples examined in each herd was 24 (range 10-30). By assuming 9.5% as the expected seroprevalence in pigs as indicated by Hause et al. 10 , our study design provides a desired precision equal to 1% in the estimate of between-animal seroprevalence and (considering 24 sera/herd) a probability higher than 90% to find positive sera in a herd with seropositive animals.
In addition, to estimate the distribution of antibody titers in farms positive to IDV, we collected 90 samples in 2015 from two farms (namely, B and C Table 1), just after IDV had been there detected.
Confidence intervals in the observed prevalence were estimated by using binomial approximation. Differences in the empirical distributions of antibody titers between farms tested positive for IDV and farms with positive sera from the routine screening through were identified with the Wilcoxon rank sum test. Statistical analyses were performed in the R 3.2.0 environment.
Finally, to evaluate whether IDV was previously circulating in Northern Italy, we analyzed 502 samples collected in 2009 from 25 herds (range: 15-25 samples/herd) where respiratory diseases occurred.
Real-Time RT-PCR. Viral RNA was extracted from clinical samples using One-For-All Vet Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The subsequent Real-Time PCR was performed as previously described by our group 24 .
Virus isolation. Samples positive by Real-Time RT-PCR were tested for virus isolation in Human Rectal Tumor cells (HRT-18G) (ATCC, Manassas, VA) as described by Ferguson et al. 25 . To maximize chances of virus isolation, samples were tested both with and without trypsin added to the culture medium. Following infection, incubation was prolonged up to 5 days, in the absence of cytopathic effect. Two serial passages were performed for each sample. Confirmation of viral replication was performed using hemagglutination test and Real-Time RT-PCR test at each passage.
Other diagnostic tests. In order to correlate IDV with respiratory disease and to exclude other causative agents, clinical specimens from IDV positive herds were also subjected to standard bacteriological and/or virological PCR tests against the most common swine respiratory pathogens. These included: Actinobacillus pleuropneumoniae (APP) 26  Hemagglutination and hemagglutination inhibition tests. Hemagglutination (HA) and hemagglutination inhibition (HI) tests were performed as described in standard protocols 29 . HI assay was performed using 0.5% turkey red blood cells in U-bottom 96 well plates. Briefly, sera were treated 1:5 with receptor-destroying enzyme (RDE) (Sigma-Aldrich, Milan, Italy) at 37 °C overnight, followed by heat inactivation at 56 °C for 30 min. After treatment with 50% turkey red blood cells, the sera were diluted to a final concentration of 1:10 with sodium citrate. The assay was conducted at room temperature starting from dilution 1:20 to 1:640 for detection of D/swine/ Italy199724-3/2015 specific antibodies. Samples showing antibody value ≥ 20 were considered positive. HI titers were expressed as the reciprocal of the highest dilution of serum that completely inhibited hemagglutination (4 HA units were used). A negative serum, as well as a swine polyclonal antiserum generated using D/swine/ Italy199724-3/2015 were used as controls in the HI assay. The antiserum was produced in IAV and IDV negative pigs by intra-tracheal inoculation of live virus (2 ml 10 −5,5 TCID 50 ) and boosting by intramuscular injection (2 ml with Freund's complete adjuvant) two weeks later. The polyclonal positive control antiserum was tested against swine IAVs circulating in Italy (A/swine/Italy/257605/2010 H1N1, A/swine/Italy/284922/2009 H1N2, A/swine/ Italy 312583/2009 H3N2) showing no cross-reactivity. Serological cross reactivity against ICV has not been considered because circulation of ICV in swine in Italy has not been demonstrated. Moreover Hause et al. 11 showed no cross reactivity between IDV and ICV in human sera.
Microneutralization test. A microneutralization assay was performed as described for Influenza A virus 30 . Virus stock was titrated with HRT-18G cells using a lysis by boiling of the cell lysate 31 followed by the Real-Time RT-PCR assay 24 to identify infected and non-infected culture wells. The TCID 50 was calculated by the method of Reed-Muench 32 . Infection was performed as described for Influenza A virus 33 . An inoculum containing 1000 TCID 50 /50 µl of IDV was mixed in duplicate, in a 96-well culture plate, with a two-fold dilution series (from 1/10 to1/160) of 15 sera showing HI titer 1/20. Incubation was performed for 1 h at 37 °C and then 100 µl of HRT-18G cells (1.5 × 10 4 per well) were added. After incubation of 72 h at 37 °C viral RNA was extracted from cell culture supernatant diluted 1:1 v/v in water and boiled 10 min at 95 °C as previously described 31 . Viral infectivity was assessed by RT-PCR 24 using 5 µl of boiled sample. Sera dilution with at least 90% inhibition of the RT-PCR signal were considered as positive.