Nipah virus infection and disease in ferrets
In humans, disease resulting from NiV infection can vary in intensity from an acute febrile illness or one progressing to severe central nervous and respiratory disease. Pathological findings show systemic vasculitis, necrotizing alveolitis and meningoencephalitis 
. The disease in experimentally infected cats and hamsters is similar 
; but in hamsters meninoencephalitis is more prominent, while cats develop an acute respiratory disease 
. Here, we sought to assess a new ferret model of NiV pathogenesis where our preliminary observations had confirmed susceptibility to NiV infection, with development of systemic vasculitis and involvement of the central nervous and respiratory systems. Ferrets have emerged as a model for several viral respiratory diseases including avian influenza 
, severe acute respiratory syndrome 
), and morbilliviruses 
, close relatives of henipaviruses 
. They offer the combined advantages over either of the aforementioned laboratory animal species of being relatively small mammals, while displaying complex behaviors especially in relation to their handlers that may be used to advantage in clinical assessments. They are however also sufficiently large to enable repeated collection of a wide range of clinical samples throughout the course of an experimental infection, as well as administration of potential therapies in a manner similar and consistent with human medicine.
We initiated a NiV minimal infectious dose study (MID50
) for the purpose of determining an appropriate challenge dose for subsequent work that would reliably productively infect naïve ferrets. Doses of 50, 500, 5,000 or 50,000 TCID50
were each administered to groups of two ferrets oral-nasally; the most likely route of natural infection. Based on prior experience with NiV infection in cats 
and using similar parameters, we defined infection endpoints in ferrets for the purpose of humane euthanasia; these were used as surrogates for lethality. Endpoints were fever plus signs of rapidly progressing clinical illness including both generalized (e.g. inappetance, depression) and localizing (e.g.dyspnea, neurological signs) disease signs. Ferrets that received 50 TCID50 (1–50 and 2–50), and one animal that received 500 TCID50 (3–500) remained well throughout the period of observation, did not shed detectable virus or viral RNA, did not seroconvert, and their tissues were normal at post mortem and histological examination. The remaining ferrets developed fever (4–7 days post-infection (dpi)) and rapidly progressive clinical illness (6–8 dpi). Ferrets inoculated with 50,000 TCID50 (7–50000 and 8–50000) were euthanized 6 and 7 dpi, respectively; those inoculated with 5,000 TCID50 (5–5000 and 6–5000) were euthanized 8 and 10 dpi, respectively, and one ferret inoculated with 500 TCID50 (4–500) was euthanized 9 dpi.
Clinical signs in affected ferrets included severe depression, cough, serous nasal discharge, dyspnea and subcutaneous edema of the head (8–50000); severe depression, orthopnea, and expiratory dyspnea (7–50000); severe depression, orthopnea and cutaneous ecchymoses (6–5000); severe depression, vomiting and hypothermia (5–5000); and obtunded with tremor and hind limb paresis (4–500). Gross pathological findings comprised varying degrees of subcutaneous hemorrhagic edema of the head and neck, centered upon local lymph nodes in all diseased ferrets; focal raised pin-head hemorrhagic nodules scattered throughout the pulmonary parenchyma; and scattered intra-abdominal petechial hemorrhages. Histopathology of diseased ferrets revealed acute focal necrotizing alveolitis and pulmonary vasculitis, acute glomerular necrosis, focal necrosis of the spleen, and severe diffuse subacute inflammation of the organs and connective tissues of the head and neck. Less commonly, there was mild focal nonsuppurative meningitis (4–500, 5–5000 and 8–50000), focal cystitis (4–500), severe acute necrotizing salpingitis (4–500), acute focal coagulative necrosis of the adrenal cortex (4–500 and 5–5000), and severe acute thyroiditis (6–5000). Two ferrets demonstrated neurological signs (4–500, 5–5000) and had nonsuppurative meningitis. Syncytia were usually present in lesions and often contained abundant viral antigen (). In mildly affected lymph nodes there was focal mononuclear cell and neutrophilic inflammation of the capsule, accompanied by a zone of subcapsular lymphocyte depletion. In more severely affected lymph nodes, there was severe extensive hemorrhagic and coagulative necrosis, often resulting in the destruction of the entire node with antigen staining mainly in syncytia at the living margin of the necrotic focus (). Clinically healthy ferrets had no gross pathological changes, and no viral antigen or significant histological lesions were found in any tissues.
Nipah viral loads and tissue tropism
To examine virus dissemination and tissue tropism, NiV N gene RNA was measured in pharyngeal and rectal swabs, blood sampled over the course of infection and in a variety of tissues recovered at necropsy. Among ferrets that developed clinical disease, NiV RNA was detected in the blood () at the time of euthanasia and, in two of these (5–5000, 6–5000), also in the sample collected two days earlier. One ferret (3–500) that had remained healthy during the period of observation demonstrated a low level of viral RNA in blood 21 dpi. Interestingly, this ferret shared a cage with ferret 4–500, from which viral RNA was recovered in pharyngeal and rectal swabs 9 dpi suggesting viral shedding. Viral RNA (albeit some at low level) was detected in pharyngeal swabs from all ferrets demonstrating clinical disease () and in rectal swabs from two of these (). Notably, NiV RNA was not detected in any tissues from clinically healthy ferrets. For all other animals with clinical signs, significant levels of viral RNA were detected in adrenal, kidney, lung, bronchial lymph node and spleen tissues (). Although at lower relative levels, NiV RNA was also detected in the bladder, liver, ovary or testes, uterine horn and in the brain (). All infected ferrets had NiV RNA in the olfactory lobe of the brain and ferret 4–500, which showed tremors and hind limb weakness, had the highest level in the occipital lobe.
NiV RNA and virus isolation from ferrets post-challenge.
Although real-time PCR can help discern virus distribution in an infected animal, it does not discriminate between infectious and non-infectious genetic material. For these reasons, virus isolation was attempted from samples and tissues with NiV RNA relative expression levels above 10. NiV was recovered from almost all tissues that had high levels of NiV RNA (, top panel) as indicated by the (+) above individual data bars. Lungs and lymphoid tissues were sites of extensive virus replication. Virus was also isolated from tissues with lower NiV RNA levels, including the occipital lobe, uterine horn and ovary from ferret 4–500 and the olfactory lobe from ferret 8–50000 (, bottom panel). For ferrets where NiV was isolated from brain, histological lesions included nonsuppurative meningitis.
Antibody m102.4 prevents NiV disease in ferrets
Recently, we described a hmAb, m102.4 which engages the receptor-binding site of the viral attachment G glycoprotein, potently neutralizes HeV and NiV in vitro
, and retained its biological activity in vivo 
. For the evaluation of m102.4 efficacy in the ferret model two time points were selected, 24 hrs before (pre-) or 10 hrs after (post-) NiV challenge. Each treatment group contained three ferrets except for the control group (n
2). At the indicated time, 50 mg of m102.4 was administered via intravenous catheter. One ferret in the post-challenge group (28-post) was given m102.4 intraperitoneally due to difficulty in venous catheterization. Control ferrets (22-con and 25-con) were given PBS. Ferrets were challenged oral-nasally with a 5,000 TCID50
dose of NiV (ten-fold the MID50
as determined by the dose-ranging study above). By 6 dpi, ferrets 22-con, 25-con and 23-pre were febrile (>40°C) and control animals started to demonstrate signs of clinical illness similar to those seen in the initial MID50
study. By 8 dpi the controls were severely depressed, had subcutaneous edema of the head and cutaneous hemorrhages and they were euthanized. By 8 dpi, 24-pre and all ferrets in the post-group were also febrile with variable levels of depression and suppression of play activity; a notable sign of abnormal ferret behavior. At 10 dpi, the temperatures of 23-pre and 27-post had started to fall, and each demonstrated moderate (23-pre) or mild (27-post) edema of the throat. 24-pre and 26-post remained febrile with no localizing signs. By 13 dpi, ferrets 23-pre and 24-pre were depressed and inappetant with cutaneous ecchymoses; ferret 23-pre had marked hind limb paresis and generalized tremor and both animals were euthanized. All other ferrets (three in the post- and one in the pre-group) were well and free of any disease signs and remained so until the end of the study (). Upon study completion (20 dpi) all surviving animals were euthanized. For the controls and two ferrets in the pre-challenge group that succumbed to infection, gross and microscopic pathology revealed findings similar to those described above in the MID50
study. However, in ferrets 23-pre and 24-pre the frequency of pinpoint hemorrhagic lesions observed in the pulmonary parenchyma was reduced and lesions were much smaller suggesting the disease progression in the respiratory tract had been dampened (), consistent with their survival to 13 dpi. Histopathological findings in controls and in the two diseased pre-challenge group ferrets were similar to those described in the MID50
study. However, for ferret 23-pre, NiV antigen was present in ependymal tissue () and for ferret 24-pre NiV antigen was present within neurons and neuropil of the cerebellum (), further supporting the suggestion that ferrets could be an appropriate model for NiV-mediated meningoencephalitic disease. No significant pathological abnormalities were found in any of the surviving ferrets.
Clinical scoring and outcome of NiV-infected ferrets from m102.4 efficacy trial.
Gross pathology of the lung post-mortem in ferrets treated with m102.4 prior to NiV challenge.
Immunohistochemistry demonstrating NiV antigen in brain tissue in ferrets treated with m102.4 prior to NiV challenge.
Nipah viral loads and tissue tropism in m102.4 treated ferrets
Blood, swabs and tissues were tested for the presence of NiV RNA and results are shown in . Ferret 21-pre, the only animal to survive in the pre-challenge group, had detectable levels of viral RNA in the blood 3 dpi () whereas all other samples were negative. All other ferrets had significant viremia by 8 dpi with the highest levels of NiV RNA in the controls. By 13 dpi, viral RNA levels in ferrets 23-pre and 24-pre were similar to those found in controls whereas NiV RNA levels had decreased in ferret 26-post and were undetectable in the other post-challenge group ferrets. Viral RNA levels in oral swabs were highest among the controls and pre-challenge group; however, post-challenge group ferrets shed virus over the course of infection (). NiV RNA levels in rectal swabs were low except in 23-pre and 24-pre (). Together these data demonstrated that all treated ferrets (except 21-pre) had significant amounts of systemic and mucosal NiV for at least 10 dpi. Controls and the two diseased pre-group ferrets had similar high levels of viral RNA in all tissues (). Conversely, 21-pre and all post-challenge group ferrets had significantly reduced levels of viral RNA (). Virus isolation was attempted on all samples with detectable NiV RNA (). On day 8 post-infection, NiV was isolated from the blood of one control animal (22-con), from a pharyngeal swab from the other control animal (25-con) and from a rectal swab from ferret 23-pre. All other blood and swab samples were negative for virus isolation despite the high levels of viral RNA detected. Virus was isolated from the majority of control animal tissues and the two diseased pre-group ferret tissues; whereas, for 21-pre and all post-group ferrets, tissues were negative for NiV. Together, these data demonstrate that treatment with m102.4 reduced viral replication and spread, and provided a significant therapeutic benefit leading to the survival of all infected animals in the post-challenge group.
Viral loads in ferrets from m102.4 efficacy trial.
In vivo m102.4 concentration
To evaluate whether a correlation between m102.4 concentration and protection existed, levels of m102.4 in serum were measured (). Coincidently, ferret 21-pre had the highest levels of serum m102.4 pre-challenge along with the longest-lasting serum levels () and was the only animal from this group that survived. By 3 dpi, all pre-challenge group ferrets had significantly lower m102.4 concentrations compared to the post-challenge group (). Importantly, m102.4 concentrations on day 3 correlated with survival and may indicate the existence of a critical therapeutic window following virus exposure. By 6 dpi, m102.4 levels in all animals had dropped and were the lowest among the post-challenge group. Interestingly, on 6 and 8 dpi, 23-pre had higher serum m102.4 levels than those in the post-group and succumbed to disease (13 dpi) whereas 27-post had low levels of m102.4 on these days and survived infection. Antibody specific for NiV G was measured in serum post-challenge. NiV-G specific antibodies were detected in controls by 6 dpi (). Ferret 21-pre had high levels of G-specific antibodies 10 dpi which likely reflected m102.4. For all other ferrets, NiV G-specific antibodies were low on 10 dpi with higher amounts of antibody present in the post-challenge group. By 13 dpi, NiV-G specific antibody responses had increased further in all ferrets with the highest levels found in the 27-post and 28-post ferrets.
In vivo m102.4 concentrations in ferrets following NiV challenge.
Ferret immune responses to NiV G.
Monoclonal antibody resistance and virus strain specificity
In the present study, we have described m102.4, a potent neutralizing human monoclonal antibody. Its mechanism of action involves binding a single epitope within the receptor binding domain of the henipavirus G glycoprotein and affectively blocking receptor engagement and virus infection. HeV and NiV cause acute and often fatal disease in animals and humans and although virus neutralization escape mutants are unlikely in 7–10 days; we evaluated whether resistant virus was present in ferrets 23-pre and 24-pre. Virus isolation was repeated on all tissues homogenates in the presence and absence of m102.4. Virus was isolated from all samples in the absence of m102.4 as before; however, no virus was isolated from tissues when m102.4 was present (data not shown). Together, these data demonstrate that the virus circulating in ferrets 23-pre and 24-pre at the time of euthanasia was still highly sensitive to m102.4 and no evidence of escape mutants was found. In addition, we also examined m102.4 neutralizing activity against different isolates of NiV and HeV, including, NiV-Malaysia, NiV-Bangladesh, HeV-1994 and HeV-Redlands. Importantly, m102.4 neutralized all viruses potently: HeV-Redlands was completely neutralized at 1.25 µg/ml m102.4 and all other isolates were even more sensitive with complete neutralization occurring at 0.63 µg/ml m102.4. These data support the potential broad applicability of m102.4 as a post-exposure therapeutic option for henipavirus-infected individuals.