Clinical signs and gross findings. Pekin ducklings were inoculated with ABBV-1 through the intracranial (IC), intramuscular (IM), and choanal (CH) route, with the control (CO) group receiving carrier only through all routes (Supplementary Figure 1). Within 24 h after inoculation, 1/29 and 2/32 ducks from the CO and IC groups, respectively, were euthanized due to acute cerebral hemorrhage and excluded from the study.
Of the 90 ducks inoculated with ABBV-1, 8 birds (1 IC, 6 IM, and 1 CH) showed neurological signs (head bobbing, ataxia, inability to stand), which occurred most frequently (7/8) between 52 and 78 days postinfection (dpi), while only one IM bird was affected at 124 dpi (1/8). All these birds tested positive for ABBV-1 RNA in the brain and/or spinal cord, with 6/8 (1 IC, 4 IM, and 1 CH) presenting also microscopic lesions in the central nervous system (CNS).
An additional 3 birds died or had severe clinical signs that warranted humane euthanasia; the cause of disease in these birds was identified as heterophilic bacterial meningitis (CH duck, 15 dpi), severe pulmonary aspergillosis (CH ducks, 68 dpi), and septic arthritis (IC duck, 68 dpi). Of these, the CH duck with meningitis tested negative for ABBV-1 RNA in the CNS, while the other two were positive (see below). None of these birds had histologic lesions consistent with ABBV-1 infection, and their demise was considered incidental.
No duck in the entire experimental cohort had gross lesions attributable to ABBV-1 infection, such as proventricular dilatation or poor body condition, as assessed during necropsy.
Infection rates and virus RNA quantification in tissues. Presence of ABBV-1 RNA was quantified in the brain, spinal cord, proventriculus, kidney, gonads, and choanal and cloacal swabs of all ducks by reverse-transcriptase quantitative PCR (RT-qPCR). The frequency of positive tissues per experimental group and time point is reported in Table 1.
In the IC group, virus RNA was first detected at 1 week post-infection (wpi) in the brains of 5/7 (71.4 %) birds, and by 12 and 21 wpi, 100.0 % of brain and spinal cord samples tested positive. Visceral tissues (proventriculus, kidney, gonads) began testing positive starting at 12 wpi, and by 21 wpi all organs had ABBV-1 RNA (Table 1). Quantitative assessment showed an increase in the magnitude of virus RNA copies in all organs over time (Figure 1a). These differences were significant for brain samples at 1 wpi compared to 12 and 21 wpi, brain samples between 12 wpi and 21 wpi, and spinal cord samples between 12 and 21 wpi (p < 0.0001). A significant increase of ABBV-1 RNA concentration was also observed for proventriculus, kidney and gonad between 12 and 21 wpi (p < 0.0001). At every time point, brain and / or spinal cord had the highest virus RNA burden compared to the other tissues (p < 0.0001; Figure 1a). Overall, a total of 28/30 (93.3 %) birds became infected through the IC route, regardless of time point.
In the IM group, no tissues were positive for ABBV-1 RNA at 1 wpi, while by 12 wpi 100.0 % of spinal cord and 88.9 % of brain samples were positive, with similar high frequencies at 21 wpi. Visceral organs became positive by 12 wpi; however, the frequency of positive rates in these tissues was lower compared to the IC group, ranging between 44.4 to 55.5 % and 35.7 to 64.3 % at 12 and 21 wpi, respectively (Table 1). Similar to what observed for the IC group, virus RNA copies in the brain and spinal cord were highest compared to the other organs at 12 and 21 wpi (p < 0.05; Figure 1b). While no significant differences were observed between the average virus RNA copies for the same organ between 12 and 21 wpi, a slight decrease was observed in the brain, spinal cord, and gonads at 21 compared to 12 wpi. This appeared to be associated with a higher data dispersal (Figure 1b). Overall, a total of 23/30 (76.7 %) birds became infected through the IM route, regardless of time point.
In the CH group, 4 birds tested positive for ABBV-1 RNA: 2 at 12 wpi and 2 at 21 wpi (Table 1). At 12 wpi, 1 duck had low-level virus RNA copies in the brain only (2.26 x 101 RNA copies / 150 ng total RNA); this bird was euthanized because of pulmonary aspergillosis at 68 dpi. The other had low-level virus RNA copies in the spinal cord only (8.08 x 101) and was euthanized due to progressively worsening ataxia at 63 dpi. The other 2 ducks were sampled at 21 wpi. One had low virus RNA burden in the brain only (2.08 x 102), while the other was positive in the brain (7.76 x 106), spinal cord (6.75 x 106), and all visceral organs (proventriculus, 1.46 x 103; kidney, 2.02 x 103; gonad, 7.36 x 102).
Blood was collected from 115/118 euthanized birds and consistently tested negative for presence of ABBV-1 N protein by RT-qPCR (data not shown). No virus RNA was detected in any tissue of the CO group.
Virus distribution by immunohistochemistry. Immunohistochemistry (IHC) for ABBV-1 N protein was performed to assess virus distribution in multiple tissues. Tested were 3 ducks that underwent full sampling from the IC, IM, and CH groups at 21 wpi, as well as the one CH duck (partial sampling) that had tested positive in multiple tissues by RT-qPCR, and 1 CO duck from the same time point (negative control). Immunoreactivity was detected in 3/3 IC, 1/3 IM, and 1/4 CH birds. All birds positive by IHC in at least one tissue also showed presence of virus RNA in the CNS, except for one IM duck (#371), which was positive exclusively in the spinal cord by RT-qPCR but showed no immunoreactivity (Table 2). For the IC and IM birds, reactivity was present in 10/20 and 9/20 tissues. For 3 CH birds, no tissues were immunoreactive, while for the only CH bird with evidence of widespread infection by RT-qPCR, 5/8 tissues were immunoreactive. When positive, tissues had similar reactivity patterns across groups. In the CNS, immunoreactivity was seen in the nuclei of neurons randomly throughout the brain with no areas being more affected, Purkinje cells in the cerebellum, and in the grey matter of the spinal cord (Figure 2a-c); while reactivity was not observed in glial cells, occasional ependymal cells in the choroid plexuses showed nuclear and/or cytoplasmic reactivity. In the peripheral nervous system (PNS), immunoreactivity was observed in nerve fibers of the brachial plexus and/or sciatic nerve (Figure 2d), in the nuclei of neuronal bodies in ganglia adjacent to the adrenal gland (Figure 2e) and lungs, as well as in the nucleus and cytoplasm of cells in the adrenal medulla (Figure 2f). In the gastrointestinal tract, immunoreactivity was detected in the intramural plexuses of the proventriculus and various segments of the intestine (Figure 2g), but not in the ventriculus. The CO duck showed no immunolabelling in any of the examined tissues (Table 2). IHC was used to evaluate ABBV-1 N protein expression in the gonads, as part of vertical transmission assessment, and in the CNS of animals with neurological signs. Results are reported in each section.
Virus shedding and environmental dispersal. Virus dispersal was determined by quantification of virus RNA copies by RT-qPCR in the oropharyngeal and cloacal swabs at the time of euthanasia, as well as in the drinking water collected bi-weekly. Virus RNA was detected in the oropharyngeal swabs of 5/9 (55.5 %) IC ducks at 12 wpi, and 11/14 (78.6 %) and 3/14 (21.4 %) IC and IM ducks, respectively, at 21 wpi. The magnitude of virus RNA in the oropharyngeal swabs was highest in the IC group at 21 wpi (Figure 3a). Only one IC duck at 12 wpi had both oropharyngeal and cloacal swabs positive; all other cloacal swabs in the entire experiment tested negative. In the CH group, only 1 oropharyngeal swab tested positive at 21 wpi, although this bird was negative for ABBV-1 RNA in the other tested organs (Table 1). Except for the latter bird, all ducks that tested positive in the swabs were also positive for ABBV-1 RNA in the CNS and variably in the visceral tissues.
Viral RNA was detected in the drinking water starting at 8 wpi in the IC group, and 12 wpi in the IM group. The amount of virus RNA in the drinking water progressively increased over time in the IC group (Figure 3b), consistent with the highest frequency of RNA shedding being at 21 wpi in the IC group. In the IM group, the amount of virus RNA in the drinking water peaked at 14 wpi, although it remained approximately 1 – 1.5 orders of magnitude less concentrated compared to the IC group. No viral RNA was detected in the drinking water of the CO and CH groups (Figure 3b).
Histopathology. In total, 59 ducks underwent histological assessment in multiple tissues (as part of full sampling), and the brain and spinal cord from an additional 31 ducks were processed as part of partial sampling (Supplementary Figure 1). Lesions suggestive of virus infection were mononuclear inflammation of the central and peripheral nervous tissue, and chromatolysis of neurons in the brainstem (Table 3).
Inflammatory lesions. Perivascular cuffs were composed of mononuclear cells with scant cytoplasm and densely arranged chromatin (consistent with lymphocytes) and were randomly distributed throughout the gray and white matter of the cerebrum, cerebellum, brainstem, and spinal cord (Figure 4a-c). In the CNS, mononuclear perivascular cuffs were present mainly in IC and IM ducks at 12 and 21 wpi, with only 1 CH duck (the same bird with widespread infection and marked seroconversion) showing this type of lesion at 21 wpi (Table 3). All birds with inflammation were also positive for ABBV-1 RNA in the CNS. Frequency of inflammation peaked at 12 wpi, when most ducks in the IC (7/9; 77.8 %) and IM (6/8; 75.0 %) groups showed inflammation in the brain and spinal cord, and drastically declined by 21 wpi, when only 1 IC duck and 5 IM ducks had perivascular cuffs in the brain (Table 3). Similarly, lesions scoring showed that severity of inflammation in the CNS peaked at 12 wpi and decreased at 21 wpi (Figure 5). In the IC group, the cerebrum at 12 wpi displayed a significantly higher score compared to the brainstem at the same time point, as well as the cerebrum at 21 wpi, when inflammation in other areas of the CNS was not observed (Figure 5a). In the IM group, a decrease in inflammation severity was seen between 12 wpi and 21 wpi, albeit non-significant, supporting a trend for decreasing inflammation over time (Figure 5b).
Although inflammation was more common and severe in the cerebrum compared to other segments of the CNS in both the IC and IM groups at 12 and 21 wpi (Table 3), pairwise comparisons of proportions revealed no significant differences in the occurrence of inflammation between these CNS areas (p > 0.25).
Peripheral neuritis, characterized by lymphocytes around blood vessels and extending to the surrounding perineurium (Figure 4d), was detected in the brachial plexus and / or sciatic nerve of only two IM ducks at 12 wpi (Table 2), a finding consistent with possible axonal virus trafficking.
Chromatolysis. One IC, 2 IM, and 1 CH duck at 12 wpi showed multifocal neuronal chromatolysis in the brainstem (Table 4). Affected neurons were characterized by markedly enlarged soma, up to 80 mm across, with loss of cytoplasmic basophilia and nucleus often pushed to the side (Figure 4e and f). Immunohistochemistry on these 4 cases showed presence of ABBV-1 N protein in the nucleus (when available on section, due to the larger neuronal soma) and cytoplasm of the chromatolytic neurons in 3/4 birds (1 IC and 2 IM) (Figure 4g, h and i). The CH duck showed no immunoreactivity in the CNS, consistent with the low burden of ABBV-1 RNA detected in the spinal cord of this bird (Table 4).
Incidental lesions. The three ducks that were euthanized due to aspergillosis (CH), septic arthritis (IC), and heterophilic meningitis (CH) did not present perivascular cuffs in either the CNS or PNS, and did not present neuronal chromatolysis.
Serology. Ducks from all groups had not seroconverted at 1 wpi. By 12 wpi, 1/8 (12.5 %) IM and 2/8 (25.0 %) CH ducks seroconverted, albeit with low magnitude just above the threshold. At 21 wpi, 9/14 (64.3 %) IC, 10/14 (71.4 %) IM, and 1/14 (7.1 %) CH ducks showed seroconversion (Figure 6). The only CH duck that seroconverted at 21 wpi had a much higher optical density (OD) value compared to the other positive CH birds, consistent with being the only CH bird with widespread ABBV-1 RNA distribution in the CNS and visceral tissues (bird #461). The IC and IM birds at 21 wpi had the highest magnitude of seroconversion, which was significantly higher compared to both the CO and CH at the same time point, although there were no significant differences between the IC and the IM groups (Figure 6).
Relationship between clinical signs, infectious status, lesions, and seroconversion
Descriptive statistics. Eight birds (1 IC, 6 IM, and 1 CH) presented with neurological signs attributable to ABBV-1 infection, consisting of ataxia, poor balance, and head bobbing (Supplementary material). All these birds were positive for ABBV-1 RNA in the brain and/or spinal cord, and IHC demonstrated presence of N protein in neurons in the brain and spinal cord of 5/8 (62.5 %) birds. Inflammation was observed in 4/8 (50.0 %) (mild, 2/4; moderate, 1/4; severe, 1/4), chromatolysis in 3/8 (37.5 %), and seroconversion in 3/8 (37.5 %) ducks with neurological signs (Table 4).
Multivariable logistic regression analysis. A multivariable logistic regression analysis was used to predict the occurrence of neurological signs, using as explanatory variables time postinfection, inflammation, virus RNA copies in the CNS, serology, and inoculation route. Based on the unconditional association results, cerebrum subscore, spinal cord subscore, and time postinfection were fed into a multivariable regression model, which showed that only time postinfection was significantly associated. Specifically, ducks at 12 wpi had 1.3-time higher odds (1.01-1.67 CI) of developing neurological signs compared to those at 21 wpi (p = 0.042).
Inflammation of the CNS was also predicted using a multivariable logistic regression analysis, using time postinfection, virus RNA copies in brain and spinal cord, serology, and inoculation route as explanatory variables. Based on the unconditional association results, virus RNA copies in the brain and time postinfection were fed into the multivariable regression model. Results indicated that both variables were significantly correlated with the outcome variable (Table 5). Ducks at 12 wpi displayed 1.25 times higher odds of having CNS inflammation than those at 21 wpi, while per each 10-time increase of RNA copies in the brain, the odds of CNS inflammation were 1.81 times higher.
Assessment of vertical transmission
Egg laying and infectious status of embryos. Eighty-eight, 60, 34, and 61 eggs were collected from the CO, IC, IM, and CH groups between 115 and 148 dpi, of which respectively 44.3, 45.0, 35.3, and 55.7 % developed an embryo (Table 6 and Supplementary Table 1). The ratio of fertile to total eggs laid, as well as the total number of eggs normalized to the number of hens were not statistically different between groups, as tested by Chi-square test (p = 0.26) and Kruskal-Wallis test with Dunn’s test for multiple comparisons (p = 0.195), respectively.
A total of 111 embryos were tested, for which 104 brains, 101 pooled organs, and 3 entire early-stage embryos were available (Table 6); none of the tissues from embryos tested positive by RT-qPCR.
Infectious status of gonads. As egg parentage could not be determined for single eggs, we sought to evaluate the infection status of the gonadal tissue present at 21 wpi, since eggs were laid exclusively after 12 wpi. Overall, 100.0 %, 20.0 %, 16.7 %, and 0 % of ovaries, and 100.0 %, 44.4 %, 0 %, and 0 % of testis were positive for ABBV-1 RNA in the IC, IM, CH, and CO groups, respectively (Supplementary Table 2). Given the higher rate of gonadal infection in the IC group, we tested expression of ABBV-1 N protein by IHC in these tissues. When available for histological evaluation, all ovaries presented active folliculogenesis, and all testes active spermatogenesis. By IHC, ABBV-1 N protein was present in 8/8 (100.0 %) ovaries, 1/4 (25.0 %) testes, and in the only epididymis available for assessment (Table 7). In the ovary, immunoreactivity was observed in the interstitium, theca externa, granulosa cells, and rarely within primordial follicles16 (Figure 2h). When available for assessment (3 birds positive in the gonads by RT-qPCR and IHC), no reactivity was observed in the epithelium of the oviduct. Rarely, immunoreactivity was seen in the interstitial cells of the testes (Figure 2i), and in scattered ciliated epithelial cells in the epididymis of one bird. No reactivity was seen in the epithelium of the seminiferous tubules.