key: cord-0275853-z1xnwsk3
authors: Holwerda, Melle; Portmann, Jasmine; Stalder, Hanspeter; Dijkman, Ronald
title: Assessment of the zoonotic potential of the ruminant-associated influenza D virus
date: 2018-08-23
journal: bioRxiv
DOI: 10.1101/395517
sha: b719a435dd3d72d82bc366d316f8e3e0a388ff88
doc_id: 275853
cord_uid: z1xnwsk3

Influenza viruses are notorious pathogens that frequently cross the species barrier with often severe consequences for both animal and human health. In 2011, a novel member of the Orthomyxoviridae family, Influenza D virus (IDV), was identified in the respiratory tract of pigs with influenza-like symptoms and subsequently also in cattle, a species that previously never was associated with influenza virus infection. Epidemiological surveys among livestock demonstrated that IDV is worldwide distributed among ruminants, but the most striking observation is the detection of IDV-directed antibodies among humans with occupational exposure to livestock. As a first step toward identifying the zoonotic potential of the newly emerging IDV we determined the replication kinetics and cell tropism at the primary site of replication using an in vitro respiratory epithelium model of humans. The inoculation of IDV on human airway epithelial cell (hAEC) cultures revealed efficient replication kinetics and apical progeny virus release of IDV at different body temperatures. Intriguingly, the replication characteristics of IDV revealed many similarities to the human-associated Influenza C virus, including the predominant cell tropism for ciliated cells. Moreover, analysis of the host response during IDV infection revealed only a pronounced upregulation of Type III interferon (IFN) transcripts. Nevertheless, viral progeny virus is replication competent and can be efficiently sub-passaged in hAEC cultures from different donors. Highlighting, that there is no intrinsic impairment of IDV replication within the human respiratory epithelium and might explain why IDV-directed antibodies can be detected among humans with occupational exposure to livestock. Importance Influenza viruses are notorious pathogens that frequently cross the species barrier with often severe consequences for both animal and human health. In 2011, a novel member of the Orthomyxoviridae family, Influenza D virus (IDV), was identified among pigs with influenza-like symptoms and subsequently also in cattle. IDV infections in humans have not yet been described, although IDV-directed antibodies have been found among people with occupational exposure to livestock. This observation suggests a possible spillover from livestock to humans. Using an in vitro human respiratory epithelium model we demonstrate there is no inherent restriction for IDV to replicate within the human respiratory epithelium and this might explain why IDV-directed antibodies are detected among humans with occupational exposure to livestock.

Since the initial discovery of Influenza D virus (IDV) in 2011, among swine with symptoms, knowledge about this new genus in the family of Orthomyxoviridae is increasing (1, 2). 67

Epidemiological studies have shown that the virus has a worldwide distribution that can be divided into 68 at least two distinct cocirculating lineages which reassort (3-10). Because of the high seroprevalence 69 the proposed natural reservoir of IDV is cattle, which in an experimental setting is shown to cause a 70 mild respiratory infection (11) . In addition to cattle, virus specific-antibodies towards IDV have also 71 been detected in swine, feral swine, equine, ovine, caprine and camelid species, suggesting a broad-72 host tropism for IDV (3, 9, 12, 13) . However, the most striking observation is the detection of IDV-73 directed antibodies among humans with occupational exposure to livestock (14) . 74

The hemagglutinin-esterase fusion (HEF) glycoprotein of IDV utilizes the receptor determinant 75 9-O-acetyl-N-acetylneuraminic acid for cell entry, which is similar to that of the closely related human 76 Influenza C virus (ICV) (15, 16) . The fusion of the IDV HEF glycoprotein with the host cell membrane 77 is efficient at both 33°C and 37°C, which is in contrast to the HEF glycoprotein of ICV that is restricted 78 to 33°C (10, 17) . This discrepancy between both viruses is mediated by the open receptor-binding 79 cavity in the HEF of IDV (15). Consequently, due to this temperature insensitivity, IDV can replicate 80 efficiently at both 33°C and 37°C in various immortalized cell lines, including those derived from 81 humans (1). Interestingly, like ICV, the HEF glycoprotein of IDV can bind to the luminal surface of 82 epithelium from the human upper respiratory tract (15). However, this does not necessarily imply that 83 cells within the human respiratory epithelium can be infected by IDV. Therefore, it remains unclear 84 whether the ruminant-associated IDV can infect cells within the human respiratory epithelium, and thus 85 whether it has a zoonotic potential. 86

The respiratory epithelium is the main entry port for respiratory pathogens and is therefore an 87 important first barrier towards intruding viruses. During the past 15 years, the human airway epithelial 88 cell (hAECs) culture model has been applied as an in vitro surrogate model of the in vivo respiratory 89 epithelium to study a wide range of respiratory viruses (18) (19) (20) (21) (22) (23) (24) (25) (26) (Figure 1A and B) . However, some temperature dependent differences were 130 observed when we analyzed the apical washes for infectious virus. For the hAEC cultures incubated at 131 33°C, viral titers were detected for every donor at 48 and 72 hpi, but only for one donor we could 132 detect viral titers at 24 hpi ( Figure 1C ). In contrast, for the IDV infection at 37°C we observed viral 133 titers as early as 24 hpi for every donor that increased over time ( Figure 1D ). These results indicate 134 that IDV kinetics seems to be more efficient at ambient temperatures corresponding to the human 135 lower respiratory tract. This, most likely, reflect the necessity for IDV to replicate at the body 136 temperature of cattle, which is between 37 -39°C. 137

After having demonstrated that IDV is able to replicate in hAEC cultures from different donors 138 at both 33°C and 37°C, we wanted to confirm these results through visualization of IDV-infected cells 139 via immunofluorescence analysis. However, because commercial antibodies against IDV are currently 140 unavailable, we ordered a custom generated antibody directed against the nucleoprotein ( signal from the NP-positive cells has a cytoplasmic distribution pattern, however some of those also 146 appeared to have a nuclear staining pattern. Suggesting that, like other orthomyxoviruses, the NP of 147 IDV is actively translocated to the nucleus during viral replication (29, 30) . 148

These results combined demonstrate that IDV is able to efficiently replicate in hAEC cultures 149 from different donors at temperatures corresponding to both the upper and lower respiratory tract of 150 humans. 151 152

Influenza C virus (ICV) is a well-known common cold virus that is able to cause mild upper respiratory 154 tract infections in humans (31). Because of the structural similarity of the HEF of IDV and ICV, and the 155 fact that we showed that IDV is able to replicate in hAECs, we wondered how this relates to replication 156 efficiency in our hAEC cultures. To address this question, we inoculated hAECs with equal amounts of hemagglutination units for ICV (C/Johannesburg/1/66) and IDV and incubated the cultures at 33°C. 158

This due to the previous reported temperature restriction of the ICV HEF glycoprotein fusion efficiency. 159

The replication kinetics were monitored as before, by collecting apical washes every 24 hours for a 160 duration of 72 hours. Here, we observed similar replication kinetics for both viruses, although the viral 161 RNA yield for ICV was higher compared to IDV (Figure 2A and B) . However, the replication kinetics of 162 the IDV-infected hAEC cultures were similar compared to the previous experiment at 33°C ( Figure  163 1A). Revealing that the replication kinetics for IDV in hAEC cultures is robust and independent from 164 the donor that is used. More importantly, we show that the replication kinetics of IDV are almost 165 identical to that of well-known common cold ICV. 166

In addition to the replication kinetics analysis for ICV and IDV in hAEC cultures, we wanted to 167 determine their respective cell tropism as both viruses utilize the 9-O-acetyl-N-acetylneuraminic acids 168 as receptor determinant. For this we formalin-fixed the previous infected hAEC cultures to analyze the 169 cell tropism for both viruses via immunostaining. To discriminate between the ciliated and non-ciliated 170 cell types we stained the cultures with well-defined antibodies to visualize the cilia (β-tubulin IV) and 171 tight junction borders (ZO-1) at the apical surface between the different cells, while the nucleus was 172 visualized using DAPI. For detection of IDV infected cells we used our previous generated NP-173 antibody, whereas for ICV there are unfortunately no commercial antibodies available. Therefore, we 174 used intravenous immunoglobulins (IVIg) that contains polyclonal immunoglobulin G from over a 175 thousand of healthy donors, as most people encounter one or multiple ICV infections during their life 176 and generate antibodies directed against ICV (31, 32). By overlaying the different cellular marker 177 stains with that of the virus antigen we observed that for both ICV and IDV the virus-positive signal 178 overlaps with that of the ciliated cell marker ( Figure 2C and 2D). 179

To accurately define the cell tropism, we counted all cell types among ten random fields per Figure 2E ). This is in line with our initial observation, and shows that IDV and 185 ICV both have a predominant preference for ciliated cells. This is most likely due to the usage of the 186 same 9-O-acetyl-N-acetylneuraminic acids as receptor determinant to enter the host cell. In addition to 187 the cellular tropism, we also calculated the overall infection rate for IDV and ICV, which is 4.1 and 3.3 percent, respectively (Supplementary table 1). Showing that the overall infection rate of IDV and ICV 189 are almost identical, which is in accordance with the previous observed replication kinetics . 190 These results establish that the ruminant-associated IDV has almost identical replication 191 kinetics and overall infection rate characteristics in hAEC cultures to that of the human-associated 192 ICV. Furthermore, both viruses show to exhibit a predominant preference for ciliated cells, which is 193 most likely due to the usage of the same receptor determinant for cell entry. To address whether IDV provokes a similar host response as ICV we inoculated hAEC 207 cultures of different donors as described previous. However, for the analysis of the host response we 208 lysed the hAEC cultures at 18, 36, 48 and 72 hpi for relative quantification of interferon (IFN)-β, IFN-λ1 209 and IFN-λ2/3 mRNA-transcripts. Interestingly, for both IDV and ICV there was no upregulation of IFN-210 β detected. This is in contrast to the IFN-λ1 and IFN-λ2/3 transcripts that all increase overtime ( Figure  211 3A and 3B). The observed increase was most pronounced for the IDV-infected hAEC cultures, as for 212 ICV the amplitude in IFN-λ1 and IFN-λ2/3 transcripts increase was approximately 10-fold lower 213 ( Figure 3A and 3B) . This seemed to correlate with the difference in the amount of cellular associated 214 viral transcripts for both viruses (data not shown). Because we observed an upregulation in the type III 215

IFNs mRNA transcripts we wondered whether an induction of downstream ISGs transcripts could be 216 detected. To this end we also monitored the expression levels for several well-known ISGs, namely 217

MxA, 2`-5`OAS and IFIT1. Interestingly, for both viruses we observed only a minor upregulation of 2`-218 5`OAS and IFIT1 transcripts beyond 48 hpi, while for MxA the transcriptional levels remained constant ( Figure 3C and 3D) . The minor upregulation of 2`-5`OAS and IFIT1 at 48 hpi seems to correlate with 220 the observed increase level of Type III IFN transcripts. However, in which extend this would impair 221 virus propagation and spread remains to be elucidated. 222

Nonetheless, these results reveal that both the human-associated ICV and ruminant-223 associated IDV provoke a similar type of host response during infection, which is characterized by an 224 induction of only type III IFNs. Figure 4B ). However, we observed that the viral yield at 37°C was approximately one order lower 245 in the first round compared to the viral yield at 33°C, however at 96 hpi this difference was slightly 246 reduced. Interestingly, in the second passaging experiment we did not observe any difference 247 between the different incubation temperatures ( Figure 4B ). In addition, we observed no pronounced 248 differences in the viral titers between the different temperatures or passage numbers at 96 hpi ( Figure  249 4C). However, at 48 hpi we could only detect infectious virus in the apical wash from the last passaging experiment that was performed at 37°C (P3; Figure 4C ). These results show that the viral 251 progeny from the initial experiments on hAEC cultures is replication competent and that IDV can be 252 sub-passaged on hAEC cultures from different donors at both 33°C and 37°C. 253 This observation raises the question if IDV is potentially circulating among the general 254 population. To address this question, we performed a cross-sectional serological survey using IVIg in 255 a hemagglutination inhibition (HI) assay to determine whether IDV-directed antibodies can be detected 256 among pooled immunoglobulin G from thousands of healthy donors. For the HI assay we used ICV as 257 a positive control, as the majority of adults have antibodies towards this virus and we could previously 258 detect ICV-positive cells by immunofluorescence. Here we could readily observe that IVIg inhibited red 259 blood cell hemagglutination by ICV, however for IDV no inhibition was observed (Figure 4D and 4E) . 260

Indicating that, unlike ICV, no IDV-directed antibodies can be detected among the general population 261 and that pre-existing antibodies against ICV are not cross-reactive against IDV. 262

Combined these results highlight that there is no intrinsic barrier for IDV to replicated within 263 the human respiratory epithelium. However, thus far, there is no evidence that IDV is circulating 264 among the general population. individuals whom have occupationally exposure to livestock (14). However, we cannot assess whether 285 IDV can be transmitted among humans with our model. Nonetheless, it is interesting to note that, IDV 286 can be transmitted among infected and naïve ferrets, an animal model often used as surrogate model 287 to assess transmission potential of influenza A viruses among humans (1, 38, 39). However, besides 288 individuals whom have occupationally exposure to livestock there is currently no epidemiological 289 evidence that IDV is circulating among the general population. Suggesting that beyond the spillover 290 from livestock to humans virus transmission might be restricted due to unknown host factors. 291

Within the respiratory epithelium, the host innate immune system plays a pivotal role in the 292 disease outcome during viral infection. Interestingly, in our study we observed a pronounced 293 upregulation for Type III IFNs, but not for Type I IFN. This data suggest that Type III IFNs play a more 294 pronounced role in comparison to Type I IFNs in the context of IDV infection in the human respiratory 295 epithelium. In contrast to the IFN expression, we only detected a mild upregulation for some ISGs 296 transcripts during IDV infection. This suggests that certain proteins of IDV, such as the non-structural 297 protein 1 (NS1), might efficiently antagonize the induction of ISGs in the human airway epithelium.

Conversely, thus far, we only determined the host response up to 72 hpi and therefore it might be that 300 a more pronounced ISGs induction can be observed during later time points. Therefore expanding our 301 knowledge on the dynamics of the innate immune response at multiple stages during IDV infection and 302 the role of host and viral proteins herein remains warranted. 303

During our study we used the prototypic D/Bovine/Oklahoma/660/2013 strain representing one 304 of the two distinct cocirculating lineages of IDV (10). The prototypic D/Swine/Oklahoma/1334/2011 is 305 the representative strain of the other lineage (10). Both lineages have greater than 96% identity from 306 which the HEF glycoprotein (96.7 to 99.0% identity) is the most divergent of all 7 segments (2). 307

However, due to the high similarity it is likely that the replication characteristics observations from our 308 experiments are similar between both lineages. Especially, as the previous described IDV replication 309 in human cell lines and the binding of the HEF glycoprotein to the human respiratory epithelium have 310 all been performed in the context of the prototypic D/Swine/Oklahoma/1334/2011 strain (1, 15). 311

Nonetheless, whether both cocirculating lineages of IDV indeed exhibit similar characteristics in 312 human respiratory epithelium remains formally to be elucidated. 313

Both IDV and ICV utilize the 9-O-acetyl-N-acetylneuraminic acid as their receptor determinant 314 for host cell entry (15, 16) . We have shown that both viruses have a predominant affinity towards 315 ciliated cells, suggesting that the distribution of this type of sialic acid is limited to ciliated cells within 316 our in vitro model of the human airway epithelium. This tropism is similar to what we previously 317 observed for the human coronavirus OC43, from which it has been reported to also utilize the 9-O-318 acetyl-N-acetylneuraminic acid as receptor determinant (22, 40) . Nonetheless, whether this cell 319 tropism for both IDV and ICV corresponds to that of in vivo airway epithelium remains to be 320 determined. Although, previous studies have shown that the hAEC cultures recapitulates many 321 characteristics of the in vivo airway epithelium, including receptor distribution (18, 22) . 322

In summary, we demonstrate that IDV replicates efficiently in an in vitro surrogate model of the 323 in vivo respiratory epithelium. Highlighting, that there is no intrinsic impairment of IDV propagation 324 within the human respiratory epithelium. These results might explain why IDV-directed antibodies are 325 detected among individuals whom have occupationally exposure to livestock. BE78/17. The hemagglutination agglutination and hemagglutination inhibition assays were performed 382 using 1% chicken red blood cells diluted in ice-cold PBS as described previous (43) Formalin-fixed ICV and IDV infected hAEC cultures and their respective controls were immunostained 608 with antibodies to visualize the cilia (β-tubulin IV, green), tight junction borders (ZO-1, purple). 609 Human airway epithelial cell cultures were inoculated with tenfold-diluted apical wash and sequentially 627 propagated upon new hAEC cultures to assess whether IDV viral progeny is replication competent and 628 can be sub-passaged (A). The monitored viral RNA yield is given as genomic equivalents (GE) per 2 629 μ L of isolated RNA (y-axis) at indicated hours post-inoculation (x-axis) for each of the conditions (B). 630

Whereas the viral titer is given as TCID 50 /mL (y-axis) for each condition, at indicated hours post-631 inoculation (x-axis) (C). The results are displayed as means and SD from duplicates from three 632 independent donors. A cross-sectional survey was performed using 8 hemagglutination units of virus 633 antigen in combination with intravenous immunoglobulins (IVIg) in a hemagglutination inhibition (HI) 634 assay to detect ICV-and IDV-directed antibodies among the general population (D). The HI-titer was 635 calculated and displayed as mean HI-titer (E). Results are displayed as means and SD from duplicates 636 from four independent experiments. 637

Isolation of a Novel Swine Influenza Virus 452 from Oklahoma in 2011 Which Is Distantly Related to Human Influenza C Viruses. PLoS 453 Pathog 9

Characterization of a novel influenza 455 virus strain in cattle and swine: proposal for a new genus in the Orthomyxoviridae family

Influenza D in Italy : towards a better understanding 459 of an emerging viral infection in swine

Influenza D Virus in Animal Species in Guangdong province

Nationwide distribution of bovine influenza D virus infection in 465 Japan

Serological 467 evidence for high prevalence of Influenza D Viruses in Cattle

Influenza d virus in cattle

Influenza D Virus in Cattle

Cocirculation of Two Distinct 478 Genetic and Antigenic Lineages of Proposed Influenza D Virus in Cattle

Pathogenesis of Influenza D virus in Cattle

Influenza D Virus 484 Infection in Feral Swine Populations , United States. Emerg Infect Dis 24

Serological evidence for the co-circulation of two 487 lineages of influenza D viruses in equine populations of the Midwest United States. Zoonoses 488 Public Health 1-7

Serologic evidence of exposure 490 to influenza D virus among persons with occupational contact with cattle

Influenza D Virus: Basis for Its Broad Cell Tropism

Influenza C virus uses 9-O-acetyl-N-495 acetylneuraminic acid as a high affinity receptor determinant for attachment to cells

Intrinsic Temperature Sensitivity of Influenza C 499 Virus Hemagglutinin-Esterase-Fusion Protein

Dipeptidyl peptidase 4 is a functional receptor for the emerging human coronavirus-503 EMC

Human and avian 507 influenza viruses target different cell types in cultures of human airway epithelium

Culturing the unculturable: human coronavirus HKU1 infects, 511 replicates, and produces progeny virions in human ciliated airway epithelial cell cultures

Isolation and characterization of current human 515 coronavirus strains in primary human epithelial cell cultures reveal differences in target cell 516 tropism

Severe acute 518 respiratory syndrome coronavirus infection of human ciliated airway epithelia: role of ciliated 519 cells in viral spread in the conducting airways of the lungs

Respiratory syncytial virus 521 infection of human airway epithelial cells is polarized, specific to ciliated cells, and without 522 obvious cytopathology

524 Infection of Ciliated Cells by Human Parainfluenza Virus Type 3 in an In Vitro Model of Human 525 Airway Epithelium

In vitro susceptibility to rhinovirus infection 528 is greater for bronchial than for nasal airway epithelial cells in human subjects

Efficient replication of the 532 novel human betacoronavirus EMC on primary human epithelium highlights its zoonotic 533 potential

Link of a ubiquitous 537 human coronavirus to dromedary camels

The NPI-1/NPI-3 Nucleoprotein, (karyopherin α ) binding 539 site on the influenza a virus NP is a nonconventional nuclear localization signal

Contributions of two nuclear localization signals of influenza A virus 543 nucleoprotein to viral replication

Clinical Features of Influenza C Virus Infection in Children

Influenza C virus high seroprevalence rates observed in 3 different population groups

Respiratory epithelial cells in innate immunity to 555 influenza virus infection

INTERFEROME: The database of 557 interferon regulated genes

Identification of genes differentially 559 regulated by interferon alpha, beta, or gamma using oligonucleotide arrays

Host 561 Species Barriers to Influenza Virus Infections

Airborne Transmission of Influenza A/H5N1 Virus Between Ferrets

Experimental adaptation of an influenza H5 HA confers respiratory droplet 569 transmission to a reassortant H5 HA/H1N1 virus in ferrets

Human and bovine coronaviruses recognize 571 sialic acid-containing receptors similar to those of influenza C viruses

Differentiated Human Airway Epithelial Cell Cultures

Coronaviruses: Methods and Protocols

WHO Global Influenza Surveillance Network. 2011. Manual for the laboratory diagnosis and 579 virological surveillance of influenza

Analysis of relative gene expression data using real-time 581 quantitative PCR and the 2-ΔΔCT method

A DNA transfection 583 system for generation of influenza A virus from eight plasmids

Whereas virus-infected cells (red) were visualized with either a custom generated IDV NP-antibody or 610 intravenous immunoglobulins (IVIg) for ICV (C&D). Magnification 60x, the scale bar represent 10 μ M

The cell tropism of ICV (Black bars) and IDV (white bars) was quantified by calculating the percentage 612 of viral antigen-positive signal co-localization with either ciliated or non-ciliated cells (E). The mean 613 percentage and SEM from ten random fields from three independent donors are displayed

Figure 3. Transcriptional host response during ICV and IDV infection in hAECs

Human airway epithelial cell cultures were inoculated with 32 HAU of ICV and IDV after which the 618 transcriptional host response was quantified for Type I and III interferon (IFN

Stimulated Genes (ISG) mRNA-transcripts (C, D) at 18, 36, 48 and 72 hours post-inoculation using the 620 Δ Δ Ct-method (44). The results are displayed as means and SD from three technical replicates from

Total cellular RNA from infected hAECs was extracted using the NucleoMag RNA (Macherey-Nagel) 392 according to manufacturer guidelines on a Kingfisher Flex Purification system (Thermofisher). Reverse 393 transcription was performed with GoScript™ reverse transcriptase mix random hexamers according to 394 the manufacturer's protocol (A2800; Promega) using 200 ng of total RNA. Two microliters of tenfold 395 diluted cDNA was amplified using Fast SYBR™ Green Master Mix (Thermofisher) according to the 396 manufacturer's protocol using primers targeting 18S and MxA, 2'-5'-OAS, IFIT1 as described 397previously (27). Measurements and analysis were performed using an ABI7500 instrument and 398 software package (ABI). Relative gene expression was calculated using the 2-∆∆Ct method and is 399shown as fold induction of compared to that of non-infected controls (44). 400For quantification of the viral kinetics of IDV and ICV, a total of 50 μL of apical wash was used to 401 extract viral RNA using the NucleoMag VET (Macherey-Nagel) according to manufacturer guidelines 402 on a Kingfisher Flex Purification system (Thermofisher). Two microliters of extracted RNA was 403 amplified using TaqMan™