Neglected Hosts of Small Ruminant Morbillivirus

Eradication of small ruminant morbillivirus (PPRV) is targeted for 2030. PPRV lineage IV is found in much of Asia and Africa. We used PPRV lineage IV strain Kurdistan/2011 in transmission trials to investigate the role of pigs, wild boar, and small ruminants as PPRV reservoirs. Suids were a possible source of infection.

Virus isolation was generally conducted from samples with Cq values <34 (sheep; but including all PCR-positive samples collected 4 to 10 dpi), <32 (wild boar; as detected with Batten PPRV-PCR assay) or ≤32 (pigs; as detected with Bao PPRV-PCR assay). Inocula (used for experimental infection) and animal samples were titrated with endpoint dilution assay in Dulbecco's Modified Eagle's Medium (DMEM) with 7.5% to 10% FBS, antibiotics and antimycotic using a ten-fold dilution series -starting with the 1:10 dilution. After adding VDS or CHS-20 cells, titration assays were incubated for 5 to 7 days at 37°C and 5% CO2. Cells were regularly examined for syncytia caused by cytopathic effect (CPE) to determine 50% tissue culture infective dose (TCID50/ml) for the samples.

Indirect immunofluorescence assay
The results of the titration assays were confirmed by immunofluorescence staining together with a positive control (PPRV Kurdistan/2011/BH15-11_5/1CHS/2VDS_PC 15/09/14) of the same PPRV isolate used for experimental infection of animals using anti-PPRVnucleoprotein (Np) purified monoclonal mouse antibody (Mab anti-PPR, concentration 1 mg/ml, 50% Glycerin, ID.vet) and Alexa 488 (rabbit anti-mouse) fluorophore (Invitrogen, purchased Pathology Selected animals were examined post-mortem for gross pathological lesions. Tissue samples of up to 31 organs were processed for RT-qPCR (Technical Appendix Table 3).
Therefore, a piece of tissue of the size of a grain of rice was collected. Of all hollow organs, tissue samples were taken from the inner side to target the mucosa according to (5). From tonsils and lymph nodes a cross-section of the different areas was chopped. The tissue pieces were homogenized in 500µl of minimal essential medium (MEM) with a 5 mm steal bead at a TissueLyser (Qiagen) at 30 Hz for 2 min. The tissue homogenates were centrifuged at full speed for 1 min with a table centrifuge and subsequently subjected to homogenization, total RNA extraction and PPRV-RNA detection or additionally to virus isolation and indirect immunofluorescence assays (see earlier).
For histopathological (HP) and immunohistochemical (IHC) examination, tissue samples were fixed in 10% buffered formalin (4% solution of formaldehyde) and embedded in paraffin.
Subsequently 3 µm sections were cut, deparaffinised and rehydrated. One section was stained with hematoxylin/eosin, successive sections were used for IHC. Pretreatment included a blocking step for the endogenous peroxidase using 3% H2O2/distilled water followed by an antigen retrieval step using high temperature in citrate buffer (pH 6.0 in microwave). As a primary antibody, the Mab anti-PPR antibody (see earlier) was used in a dilution of 1:100 in Tris-buffered saline (TBS). Negative control sections were treated with TBS alone. The slides were finally developed with biotinylated anti-mouse immunoglobulin and an avidin/biotinylated enzyme complex (VECTASTAIN®ABC Reagent) followed by a visualization with AEC substrate, counterstained with hematoxylin. Infection, BDSL IRVINE LIMITED and The Pirbright Institute, Pirbright, UK) were conducted according to manufacturers' instructions. For the method comparison, only samples positive with Bao and or Batten PPRV-PCR assay were included in this evaluation. To allow a comprehensive comparison of samples containing infectious PPRV, primarily samples positive by virus titration assay were chosen. The analyses of animal samples collected during the trials revealed that RT-qPCR is the more sensitive virological method compared to virus isolation (see also Technical Appendix Table 4). As described earlier, the used RT-qPCR assays showed a high sensitivity of 100 to 1000 copies/ml. Hence, samples positive by virus isolation and/or RT-qPCR were chosen for the comparison of performance characteristics with the other two virological methods (LFD and ag-ELISA).

Statistical analyses
To determine, whether the PPRV-RNA load in excretions collected from animals over time after experimental or contact infection with PPRV lineage IV strain 'Kurdistan 2011' differ statistically significantly by Artiodactyla species, Cq values (obtained with PPRV-N-genespecific RT-qPCR as determined with the PCR assay of Batten et al. (14)) of oronasal, conjunctival and fecal swab samples were statistically analyzed. Therefore, the PCR results of oronasal, conjunctival and fecal swab samples collected from PPRV-infected goats (n = 4), pigs (n = 5), wild boars (n = 4) and sheep (n = 5) were included (Technical Appendix Figure 2, Table   3). For contact infected animals, the day of contact infection with PPRV was estimated by comparison of serologic and virological results of experimentally infected and contact infected animals. Accordingly, the day of contact infection (dpci) after i.n. PPRV infection of the experimentally infected animals was estimated 8 dpi for P1, 4 dpi for P5 and P6, 16 dpi for G4 and 8 dpi for G5 (Technical Appendix Figure 1) and consequently assigned as 0 dpci for statistical analyses and visualization in (Technical Appendix Figure 2 and Table 3). The goodness of fit of Cq values of the swab samples collected over time from the different Artiodactyla species was tested with Shapiro-Wilk normality test using R software package (17), which revealed no normal distribution of these data. Hence, p-values were calculated using i) a linear mixed-effects model (lme) including random effects (individual animal) and fixed effects (animal species and dpi as continuous variables) and ii) independent 2-group Mann-Whitney test with Bonferroni correction to adjust the α-level for multiple comparison using R software (www.r-project.org; packages stats and nlme (17,18)) (Technical Appendix Table 5).
Correlation of virus titration (log10[TCID50/ml]) and RT-qPCR (Cq values) assay results of swab and blood samples collected from pigs, wild boar, goats and sheep experimentally i.n.
infected or contact-infected with PPRV were analyzed using Spearman nonparametric correlation for the calculation of Spearman coefficients (rs) and p values (Technical Appendix Table 6). Therefore, results of oronasal, conjunctival and fecal swab samples and leukocytes were analyzed together, but separately by i) animal species, ii) animal trial, iii) detection method and iv) until versus after seroconversion. Spearman nonparametric correlation analysis was used since D'Agostino & Pearson normality test generally showed no normal distribution of the data (except for PCR results obtained from swab samples collected from goats (trial 3) and pigs (trial 1) after seroconversion). Correlation analysis (Technical Appendix Table 6) and visualization (Technical Appendix Figure 2) and D'Agostino & Pearson normality test were conducted with GraphPad Prism 7.02 software (GraphPad Software, Inc., CA, USA).
For all statistical analyses, p-values of α<0.05 were considered statistically significant.

Suggested terminology and semantic of host status
A reservoir may be a maintenance or spillover host, independent from species, but depending on (multi-)host-pathogen-environment interactions (19)(20)(21). In maintenance hosts, disease persists without an external source of reinfection, while disease in spillover hosts will disappear if the source of infection is eliminated. However, although spillover hosts may be dead-end hosts, they may occasionally play an important epidemiologic role by spilling a pathogen back to a maintenance host or by spilling the pathogen forward to another spillover host (19,20,22).
In a multi-host context, the basic reproductive rate (R0, the number of new or secondary cases of infection within a host population previously naïve to the pathogen of concern) can potentially be topped up by inter-species transmissions from other species that are maintenance or non-maintenance hosts (20). In contrast, if the number of individuals is low, extinction of the pathogen may easily occur even if R0 of the pathogen in the new landscape is greater than R0 (21). Furthermore, R0 may change, e.g., in different environments or due to a change in the pathogen traits (21). Hence, the epidemiologic role of a species may not be definite but is rather dynamic considering persistence and transmission efficiency between host populations in the respective environment (19).
The drivers (risk factors) of pathogen persistence or emergence may include population density, social behavior, animal management (e.g., intra-and inter species interaction in the wild or in a flock of a farmer or a nomad, exploitation of natural resources and sharing of water wholes or feeding places, and defecation habits), encroachment of wildlife habitats at the wildlife/agricultural interface, animal breed and changes in the pathogen resistance or phenotype (e.g., driven by genetic changes or different environmental or host conditions) (19)(20)(21)23).
Another key factor may be that individual animals may act as "superspreader" as, for For PPRV epidemiology in a multi-host system, we therefore propose to use the terminology of a dynamic multi-host-pathogen-environmental system (21) including spillover and maintenance hosts as suggested by Nugent (20) and Palmer et al. (19).

Contact transmission
Of trial 1, two (P2 and P3) of the three experimental inoculated pigs were infected due to the experimental infection with PPRV, while one contact-control goat (G5) and very likely one (P1) of the three pigs were contact-infected about 1 week pi (estimated at 8 dpi) by P2 and/or P3.
The second contact-control goat (G4) was found PPRV-infected about 1 week later (estimated at contact with the pigs and did not obviously feed together with the pigs, but from a separate feeding trough. Contact transmission from the two experimentally PPRV-infected goats (G7 and G8) of trial 3 to the two contact-control pigs (P5 and P6) could be detected in both pigs at 6 dpi ( Figure   1) -shortly after the first detection of infectious PPRV in goat swab samples. According to serologic and virological analyses, the two pigs were contact-infected a few days earlier (estimated at 4 dpi) than the P1 of trial 1 (estimated at 8 dpi).
During the wild boar and sheep trials, PPRV was not transmitted to any of the contact goats or pigs, although infectious PPRV was detected by virus isolation in some swab samples of secretions or excretions.

Clinical signs after PPRV-infection
The progression of clinical signs in the four different animal species are shown schematically in Figure 1 and in detail in Technical Appendix Figure 1.

Pigs and wild boar
The progression of clinical signs in the four different animal species are shown schematically in Figure 1 and in detail in Technical Appendix Figure 1. In trial 1, P1 was very likely refractory to PPRV-infection after experimental inoculation due to clinical, serologic and virological results and was therefore considered a contact control animal in this trial.
Accordingly, the three pigs P1 (trial 1), P5 and P6 (trial 3) were infected with PPRV by contact during the transmission trials at 8 dpi and 4 dpi, respectively. To facilitate the comparison of clinical signs with the experimentally i.n. infected pigs (P2, P3) and wild boar (W1 to W4), the day of contact infection (dpci) is given for the three contact-infected pigs.
All three i.n. inoculated pigs of trial 1 showed a marked transient rise in body temperature >40°C (max. 41.0°C) at single days (P3 at 4 dpi, P2 at 8dpi) or for 4 consecutive days (P1 at 5 to 8 dpci), ruffling bristles (P3 at 8 dpi, P1 at 7 dpci), a reduced activity and food intake / slight emaciation (P3 at 7 dpi / 8dpi), swelling of the eye lids and mild to severe conjunctivitis (P2 at 8 to 12 dpi, P3 at 9 to 16 dpi, P1 at 7 to 15 dpci), as well as mucous to purulent ocular discharge in the first days after infection. Most severe conjunctivitis together with mucopurulent ocular discharge was seen in P2 and P3 at 11 dpi and in P1 at 10 dpci. One of the experimentally PPRV-infected pigs (P3) had diarrhea for a single day (at 8 dpi). A maximum clinical score of 4, 3 or 6 was found at 8 dpi (P2 and 3) and 7dpci (P1). The pigs were reconvalescent after 13 dpi (P2), 17 dpi (P3) or 19dpci (P1). Hence, the clinical signs of the pig PPRV-infected by contact transmission (P1) were presented at a similar stage after PPRVinfection as previously observed in the experimentally PPRV-infected pigs (P2 and P3).
In the noninfected contact control pigs P5 and P6 of trial 2, no to mild clinical signs (clinical score of 0 to 2) possibly due to an increased rectal body temperature (up to 40.1°C) were observed at multiple days during the animal trial. From 6 to 8 dpci, the two pigs were mildly inactive (prolonged recumbency), mildly depressed and mildly inapparent. P5 additionally showed mucopurulent nasal discharge a single day (17 dpci).
In P5 and P6 of trial 3, a marked increase in rectal body temperature was found between 6 and 7 or 6 and 8 dpci, respectively (up to 40.3°C). At the same time both pigs showed a reduced appetite (6 to 8 dpci) and a reduced general condition (prolonged recumbency) (8 dpci).
In wild boars W1, W3 and W4 of trial 2, a marked increase in rectal body temperature was documented during the first weak after infection (40.2 to 40.9°C) and in W2 at 11 dpi (40.4°C) and 15 dpi (41.1°C). Shortly before experimental PPRV infection of the wild boars and after 9 dpi, an increase in body temperature >39.6°C was found at multiple days in three of the wild boars (W1, 2, 3). W4 and W1 showed pasty to watery diarrhea from 4 respectively 8 dpi until the end of the experiment. In addition, W4 had fresh blood in watery diarrhea at 7dpi. In W2 and W3, pasty to watery diarrhea was found at 7 and 14 dpi, respectively. W2 again showed diarrhea at 22 and 24 dpi and W3 at 17 dpi. All wild boar showed a mildly reduced general condition at 5 dpi, W3 additionally between 7 and 9 dpi and W2 at 25 dpi. W1 was mildly inappetent at 20 and 21 dpi, W2 at 24 dpi, and W3 at 15 dpi. W4 had mucopurulent nasal discharge at 8 and 9 dpi, P5 at 21 dpi. Peak clinical scores of 4 to 7 in the four wild boars were found at considerably different days (W1 at 21 dpi, W2 at 26 dpi, W3 at 15 dpi and W4 at 8 dpi).

Sheep
All sheep (S1, 3,5,7,9) of trial no. 4 experimentally infected with PPRV showed a marked increase in rectal body temperature >40.0°C (max. 40.9°C; S9 at 7 and 8 dpi), in particular between 4 and 10 dpi. Mild to moderate clinical signs were found in 4 of the 5 sheep (peak clinical score of 5 between 4 and 10 dpi). In general, clinical signs in these 4 sheep included serous to mucopurulent nasal discharge and pasty feces. One of the 5 sheep (S9) was more severely affected (peak clinical score of 14 at 10 dpi). S9 additionally showed respiratory distress, reddened/congested oral mucosa, open lesions in oral cavity, edematous lips, conjunctivitis, reduced general condition (prolonged recumbency) between 8 to 12 dpi. However, mild but similar clinical signs were also recorded in the contact control sheep (S2, 4, 6, 8, 10) on 2 to 4 single days (between 2 and 17 dpi), while S10 (the contact control sheep of S9) showed up to moderate clinical signs (max. clinical score of 5 at 10 dpi). Hence, except for S9 that showed temporary oral lesions, it remains unclear whether the clinical signs in the other i.n. infected sheep were due to PPRV infection or may have been aggravated by previous infection or vice versa.

Pigs and wild boar
In the pigs, PPRV-RNA was generally detected in whole-blood, serum, leukocytes, urine, oronasal, conjunctival and fecal swabs for up to 4 weeks after experimental or contact infection (Technical Appendix Figure 1.1). In swab samples of pigs P1 to P3, PPRV-RNA copy numbers peaked between 5 and 8 dpi or dpci in oronasal (max. in P3 at 6 dpi, 4.3 × 10 6 copies/ml, Cq 25.99), conjunctival (max. in P3 at 7 dpi, 5.9 × 10 6 copies/ml, Cq 25.51) and fecal (max. in P3 at 7 dpi, 1.5 × 10 7 copies/ml, Cq 24.11) swabs. The median PPRV-RNA copy numbers in swabs from all three pigs between 2 and 30 dpi were 1.4 × 10 4 , 3.9 × 10 3 and 1.6 × 10 4 copies/ml in oronasal, conjunctival and fecal swabs, respectively. Similarly, in whole-blood, serum and leukocytes (highest viral loads with Cq 29.1, 29.5 respectively 35.6), PPRV-RNA was detected during the first 10 days after experimental or contact infection, approximately at the time of seroconversion, as measured with cELISA. In urine, PPRV-RNA was detected at two or three of the 4 days of urine collection (except at post-mortem examination) between 8 and 10 or 14 dpi in P2 and P3, respectively. The highest viral PPRV-RNA loads were found in urine pellets at 8 dpi (Cq 32 to 33). In the urine pellet, viral loads were 0.5 log steps higher than in whole urine or urine supernatant.
PPRV was isolated with VDS and/or CHS-20 cells from conjunctival (peak of 10 3.5 TCID50/ml at 7 dpi) and fecal (peak of 10 3.5 TCID50/ml at 6 dpi) swabs from one of the experimentally infected pigs (P3) and from leukocytes from two of the pigs (P1 and P3) between 4 and 7 dpi, but not from oronasal swabs and urine samples (however, urine was collected around seroconversion). From leukocytes, PPRV was isolated at 4 and 6 dpi (P3) with a peak titer of 2.3 log10(TCID50/10 6 cells) at 6 dpi or at 14 dpi/6 dpci (P1) with a peak titer of 1.1 log10(TCID50/10 6 cells). After seroconversion at 8 dpi, no PPRV was isolated from any of the PPRV-RNA positive samples from the pigs.
In wild boar of trial no. 2 (Technical Appendix Figure 1 In the four in contact-animals P5, P6, G7 and G8 of the wild boar in trial 2, several intermittently PCR-positive results were obtained from oronasal (3 to 17 dpi), conjunctival (7 to 9 dpi) and fecal (7 to 8 dpi) swabs during the peak of viral excretion in wild boar. An explanation is the contamination of the stable with PPRV-(RNA) due to the high viral loads shed during that time by the wild boar.
In the contact-infected pigs P5 and P6 of trial 3, PPRV-RNA was detected in almost all swab samples from 6 dpi until euthanasia at 21 dpi/17 dpci. Viral RNA loads peaked at 10 dpi/8 dpci in oronasal (Cq 28.86 to 29.18), conjunctival (Cq 30.52 to 31.72) and fecal (Cq 30.63 to 31.22) swabs. In whole-blood, PPRV-RNA was detected between 8 and 12 dpi/4 and 8 dpci (up to Cq 34.32 at 10 dpi), and in serum only at the peak of RNAemia in P6 at 10 dpi. No infectious PPRV could be isolated from any of the tested (8 to 14 dpi) PCR-positive swab samples of the contact-infected pigs P5 and P6 of trial 3.

Sheep
In sheep, PPRV-RNA was generally detectable from 2 or 4 dpi until euthanasia, but the period of time varied considerably between individual animals (Technical Appendix Figure 1.4).
PPRV was isolated from sheep swab samples at single or multiple consecutive days between 4 and 10 dpi -until seroconversion at 8 dpi or shortly after seroconversion (10 dpi, S9).

pigs was probably not infected by experimental intranasal PPRV-inoculation but by contact-infection
Technical Appendix Table 2. Clinical Score (CS) sheet for the evaluation of clinical signs in PPRV-infected pigs and wild boar.  Technical Appendix Table 3. Selection of real-time quantitative reverse transcription-PCR (RT-qPCR) and immunohistochemical (IHC) results of tissue samples collected during PPRV-transmission trials from pigs (P), wild boar (W), goats (G) and sheep (S). The animals were experimentally (pi) or contact (pci) infected with PPRV lineage IV strain Kurdistan/2011 (details about the study designs of the transmission trials are given in the table in the main article). RT-qPCR results were determined with the PPRV PCR assay of Batten et al. (14) and IHC results using Mab anti-PPRV-Np (purified monoclonal mouse antibody against PPRV nucleoprotein, ID.vet). Tissue most suitable for PPRV diagnosis in all species are highlighted in bold. The negative control goat and sheep were both PPR-negative with PCR and IHC (data not shown)* Technical Appendix Table 3, part A  animal ID  P1  P2  P3  P5  P6  W1  W2  W3  W4  animal Table 3, part B  animal ID  G4  G5  G7  G8  S1  S3  S5  S7  S9  animal triangle of table). All samples significant with the lme model were also significant with the Mann-Whitney U test and vice versa, except for two values of conjunctival swab samples (marked in bold black). In general, goats secreted and excreted statically significantly (p < 0.05) higher PPRV-RNA loads over time in all swab materials compared to animals belonging to the other 3 Artiodactyla spp. Oronasal secretions were found similarly high between sheep and pigs, but significantly lower for wild boar. Conjunctival secretion patterns were detected significantly different for all species by the Mann-Whitney U test but were similar according to the results of the lme model. For fecal swab samples, PPRV-RNA excretion patterns were found similar for pigs, wild boar and sheep.  Table 3). F) PPRV antibodies were determined with competitive ELISA (cELISA, ID.vet) and by neutralization test. G) Leukocytes (WBC, white blood cell) count was determined relative to 0 dpi. H) The clinical score and I) rectal temperature was documented according to Appendix Table 2 sheep. PPRV could be isolated from samples with Cq values up to Cq 38 and even from one sample negative by PCR. In general, a high number of samples from goats, pigs, wild boar and sheep with Cq ≤31 were positive by virus titration assay, before (red to pink) or after (gray to black) seroconversion (scv).