http://www.ncku.edu.tw/~biotech/teacher/Jiann-Ruey Hong/18.pdf

A
vailable online at www.sciencedirect.com
Molecular Immunology 45 (2008) 1146–1152
The interferon response is involved in nervous necrosis virus acute and
persistent infection in zebrafish infection model
Ming-Wei Lu a, Yung-Mei Chao a, Tz-Chun Guo b, Nina Santi b, Øystein Evensenb,
Siti Khadijah Kasani a, Jiann-Ruey Hongc, Jen-Leih Wu a,∗
aLaboratory of Marine Molecular Biology and Biotechnology, Institute of Cellular & Organismic Biology, Academia Sinica, Nankang, Taipei 115, Taiwan
bDepartment of Basic Sciences and Aquatic Medicine, Norwegian School of Veterinary Science, Oslo, Norway
cLaboratory of Molecular Virology and Biotechnology, Institute of Biotechnology, National Cheng Kung University, Tainan, Taiwan
Received 29 June 2007; received in revised form 12 July 2007; accepted 18 July 2007
Available online 28 August 2007
Abstract
Betanodavirus, a small positive-sense bipartite RNA virus notoriously affecting marine aquaculture worldwide has been extensively studied in
vitro. However, impending studies in elucidating virus–host interactions have been limiting due to the lack of appropriate animal disease models.
Therefore, in this study, we have attempted to successfully establish NNV infection in zebrafish (Danio rerio) showing typical NNV symptoms and
which could potentially serve as an in vivo model for studying virus pathogenesis. Zebrafish being already a powerful research tool in developmental
biology and having its genome completely sequenced by the end of 2007 would expedite NNV research. We have observed viral titers peaked
at 3 days post-infection and histological study showing lesions in brain tissues similar to natural host infection. Further, we used this infection
model to study the acute and persistence infection during NNV infection. Interestingly, RT-PCR and immunoblotting assays revealed that the acute
infection in larvae and juveniles is largely due to inactive interferon response as opposed to activated innate immune response during persistent
infection in adult stage. This study is the first to demonstrate NNV infection of zebrafish, which could serve as a potential animal model to study
virus pathogenesis and neuron degeneration research.
© 2007 Elsevier Ltd. All rights reserved.
Keywords: Betanodavirus; Zebrafish; Animal model; Interferon; Persistent infection
1. Introduction
Nervous necrosis virus (NNV) is an important fish pathogen
belonging to the virus family Nodaviridae that targets nervous
tissues primarily brain and retina. The manifestation of NNV for
example, Piscine nodaviruses (betanodaviruses), the causative
agents of viral nervous necrosis (VNN) or viral encephalopathy
and retinopathy (VER) (Munday et al., 2002) in a wide range
of host fish species, has resulted in major economic losses for
the marine aquaculture industry. Nodaviruses are small, non-
enveloped and isometric particles containing a bipartite genome
of two positive-sense RNA molecules; RNA1 encoding the RNA
replicase and RNA2 encoding the capsid protein precursor,
which are capped but not polyadenylated (Delsert et al., 1997;
Lin et al., 2001).
∗Corresponding author. Tel.: [phone]; fax: [phone].
E-mail address: [email protected] (J.-L. Wu).
Extensive studies have revealed abnormal swimming behav-
ior and sight defect of NNV infected fish. In addition,
histological examination of tissues from the central nervous
system and the retina of infected fish often reveals areas of con-
spicuous tissue vacuolation and necrosis (Barke et al., 2002;
Dannevig et al., 2000; Johansen et al., 2004). However, the
pathogenesis of NNV infection in which acute infection in
larvae and juvenile stages caused mass mortality while persis-
tently infecting adult remains poorly understood. To this end,
we sought to use zebrafish as a potential NNV disease model
to provide insights into NNV pathogenesis and its host immune
response.
At present, zebrafish is rapidly becoming a valuable molec-
ular genetics model in understanding vertebrate organogenesis
and disease development (Glass and Dahm, 2004; Yee and Pack,
2005). To date, several viruses are known to infect zebrafish,
such as spring viraemia of carp virus (SVCV), a member of
the Rhabdoviridae, that causes significant mortality in common
carp (Cyprinus carpio)(Sanders et al., 2003) and snakehead
0161-5890/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.molimm.2007.07.018

M.-W. Lu et al. / Molecular Immunology 45 (2008) 1146–1152 1147
rhabdovirus (SHRV) which was shown to cause mortalities
exceeding 40% in zebrafish (Phelan et al., 2005).
In view of the differential effect of viral infection, we looked
into the zebrafish interferon response, in which the zebrafish
interferon gene (zIFN) has been recently identified as to having
anti-virus function (Altmann et al., 2003) and may contribute to
both induction and regulation of the innate and adaptive immu-
nity. Downstream interferon activated Mx gene has also been
identified in zebrafish, grouper, salmon, trout, and halibut upon
infection with aquatic viruses (Chen et al., 2006; Kibenge et
al., 2005; Lin et al., 2005; McBeath et al., 2006) suggesting
the importance of the interferon regulatory pathway including
RNA-activated protein kinase (PKR) and the 2-5A proteins dur-
ing viral infection. In our present study, we have observed an
elevated interferon expression in infected adult zebrafish rela-
tive to infected larvae resulting in higher rate of mortality in the
latter. This may indicate that the susceptibility to NNV infection
is dependent on the enhancement of IFN system. However, the
mechanism in which interferon response is activated upon NNV
viral infection needs to be further elucidated.
2. Materials and methods
2.1. Fish
AB (−) inbred strain zebrafish were obtained at 2 months
stage from Institute of Cellular & Organismic Biology,
Academia Sinica, Taiwan. The zebrafish were handled according
to Institutional Animal Care and Use Committee guidelines.
2.2. Cell lines and virus
The SSN-1 derivate cell line E-11 (Iwamoto et al., 2000)
was used to isolate and titrate NNV. Cells were propagated and
maintained at 28 ◦C in L-15 medium (GIBCO) supplemented
with 10% fetal bovine serum (Sigma), 100 I.U./ml penicillin,
and 0.1 mg/ml streptomycin.
The nodavirus used in this study was isolated from mal-
abaricus grouper (Epinephelus malabaricus) juveniles during
an outbreak of VER at a fish farm in southern Taiwan. Sequence
analysis of partial viral genome (data not shown) revealed that
it shares more than 95% homology with other betanodaviruses
isolated from fish in Taiwan (Chi et al., 2003). E-11 cells were
inoculated with a fourth passage virus supernatant and incubated
at 28 ◦C. When a cytopathic effect (CPE) was observed on 3 days
post-infection (p.i.), the culture supernatant was harvested, clar-
ified by centrifugation (3000 ×gfor 5 min) and stored at 4 ◦C
until challenge. The virus titer in the supernatant was determined
by using an infectivity assay and was calculated to be 1 ×108
TCID50 ml−1.
2.3. Fish challenge
Two groups of 60 adult zebrafish, one infected and one mock-
infected, were used in the challenge experiment. The fish were
held in 5 liter tanks at a water temperature of 28 ◦C. Chal-
lenge was performed by intraperitoneal (i.p.) injection of 1 ×105
TCID50 ml−1of NNV in 20 ␮l. The mock-infected group was
injected with PBS. After challenge, the fish were monitored daily
over a 14-day period for signs of disease and mortality. Six fish
in each group were sampled each day and tissue samples from
different organs were isolated for virus detection by RT-PCR and
infectivity assay (TCID50). In a separate experiment, fish were
maintained and challenged similarly, but three fish from each
group were sampled daily for histology and immunohistochem-
istry. Some fish in parallel were also collected for confirmation
of infection by RT-PCR and virus re-isolation.
2.4. Histology and immunohistochemistry (IHC)
The heads of fish used for histology were fixed for at
least 24 h in neutral phosphate-buffered 10% formalin and
embedded in paraffin wax. Five micrometers brain sections
were stained with haematoxylin and eosin (H&E) whereas the
other parallel sections were processed for immunohistochemical
detection of nodavirus protein using anti-NNV antiserum and
a streptavidin-biotin-alkaline-phosphatase complex antibody
detection technique (Chemicon IHC Select Kit). Mock-infected
fish served as negative controls for the experiment.
2.5. RNA isolation and RT-PCR
RNA was extracted using TRIZOL reagent (Life Technolo-
gies) according to the manufacturer’s instructions. The RT-PCR
was performed using a Superscript III one-step RT-PCR sys-
tem (Invitrogen) and each reaction includes 2 ␮l total RNA
(20 ng) and a primer set. The primers were designed to amplify
the variable T4 region of the coat protein gene (Nishizawa
et al., 1994), Mx (sense 5-AGTACCGGGGAAGAGAGCT-
A-3antisense 5-AAGGTGGCATGATTGT CTGT-3), IFN
(sense 5-ATGAGAACTCAAATGTGGAC-3antisense 5-TT-
ACA CTCGAGGATTGAC-3), and ␤-actin (sense 5-ATGGAT-
GAGGAAATCGCTG-3antisense 5-ATGCCAACCATCA-
CTCCCTG-3) gene.
2.6. Microinjection of virus into zebrafish larvae
The NNV titer for microinjection was adjusted to 1 ×103
TCID50 ml−1. Injections were conducted by PLI-100 air-
injection apparatus (Medical System Co.). The amount
(approximately 0.1 ␮l) injected into larvae was estimated by
visualizing the injection volume. After injection, the larvae were
incubated in the Petri-dish at 28 ◦C.
2.7. Immunoblot
Fish samples were collected, homogenized, and solubilized
in disruption buffer containing 50 mM Tris–HCl (pH 7.0), 5%
2-mercaptoethanol, 2% sodium dodecyl sulfate (SDS), and
2.75% sucrose. Samples were then sonicated, boiled, sub-
jected to electrophoresis on denaturing 12% polyacrylamide
gels, transferred to nitrocellulose membranes, blocked with
5% nonfat milk, and reacted with antibodies against phospho-

1148 M.-W. Lu et al. / Molecular Immunology 45 (2008) 1146–1152
rylated eIF2␣protein (Cell Signaling). The membranes were
rinsed in phosphate-buffered saline (PBS) and reacted with anti-
rabbit immunoglobulin conjugated to horseradish peroxidase
and developed with an enhanced chemiluminescence Western
blot detection system kit (Amersham Pharmacia).
2.8. Quantitative RT-PCR for NNV RNA detection
A 100 ng aliquot of total RNA was used to quan-
tify NNV-specific RNA levels using an ABI Prism 7000
sequence detector (Applied Biosystems). Real-time reverse
transcription-polymerase chain reaction (RT-PCR) amplifica-
tions were performed by the High Capacity cDNA Archive kit
(ABI, Applied Biosystems) and primer specific for NNV 5-
CGAGTCAACACGGGTGAAGA-3. RT reactions were incu-
bated for 10 min at 25 ◦C, 2 h at 37 ◦C and cooling to 4 ◦C
for 5 min. The quantitative RCR protocol provided by the ABI
real-time instruments by using Platinum®SYBR®GreenqPCR
SuperMix-UDG, and the primers specific for NNV were 5
NTR: 5-GCCCCTGATGGAGCAGTCT-3(sense 10 ␮M); 5-
AGCACGGTCAACATCTCCAGTT-3(antisense, 10 ␮M); 45
cycles of PCR were performed with cycling conditions of 3 s at
95 ◦C, 30 s at 60 ◦C. Standard curve was generated using vector
(pDA8; Lu et al., 2003) bearing NNVs RNA2 gene of variable
known concentration. The real-time PCR signals were analyzed
in a multiplex format using SDS software (Version 1.7; Applied
Biosystems).
3. Results
3.1. NNV replication in different organs
Six zebrafish were sacrificed each day over 14 days and
pooled tissues from different organs were tested by RT-PCR and
TCID50 titration for NNV. Brain, eye, heart, liver and gut were
tested positive for NNV, while muscle was negative for NNV
by RT-PCR from 3 days p.i. (Fig. 1a). The yields of RT-PCR
product from brain, gut and eye were greater compared to heart
and liver. It was observed to being most abundant in the brain
which is the major target organ for NNV propagation.
Virus titers were detected in brain, eye, and gut at 3.6 ×105,
3.6 ×102, and 3.6 ×102TCID50 g−1, respectively, but not in
the heart, liver or muscle (Fig. 1b). Detection of NNV in gut,
heart and liver from RT-PCR was probably due to excessive virus
introduced during inoculation but had decreased significantly on
Fig. 1. Tissue distribution of virus replication in infected zebrafish. (a) Agarose
gel showing RT-PCR products specific for the NNV coat protein. (b) Virus titers
in different organs from fish exposed to NNV via i.p. injection. B, brain; G, gut;
H, heart; E, eye; L, liver; M, muscle; ND, not-detected.
Fig. 2. NNV replication in infected zebrafish brains at different times post-
infection. (a) Agarose gel showing RT-PCR product specific for the NNV
coat protein. (b) Virus titers in brains from NNV infected and mock-infected
zebrafish.
2 days p.i. (data not shown). In addition, TCID50 assay indicated
no considerable amount of virus was in the heart and liver while
the low copy number of NNV could still be within the detection
of RT-PCR.
NNV titers increased in brain tissue (Fig. 2a) from 1.4 ×103
TCID50 g−1on 1 day p.i. to peak levels of 3.6 ×105TCID50 g−1
after 3 days p.i. (Fig. 2b). The finding of NNV in the target
organs brain and eye implies that the virus can overcome the
blood–brain barrier after i.p. injection.
3.2. Nodavirus detection by histopathology
Histopathological lesions involving vacuolated cells were
observed in brain tissue from 4 days p.i. onwards. Using
immunohistochemistry, positive NNV-specific signal was
observed in brain tissue surrounding these lesions (Fig. 3).
These results demonstrated that the nodavirus not only replicates
but also cause pathological lesions in brain tissue of infected
zebrafish.
3.3. Re-isolation of NNV from zebrafish in E-11 cells
Brain tissue from zebrafish that tested positive by RT-PCR
was used to re-isolate NNV in E-11 cells. Development of CPE
manifest the cell death and detachment was observed in the E-
11 cells 3–4 days after inoculation of the tissue homogenate
(Fig. 4). The presence of NNV in cell cultures displaying CPE
was confirmed by using RT-PCR and sequence analysis of PCR
product.
3.4. Nodavirus infection in zebrafish larvae
The NNV infection can spread both by horizontal and vertical
transmission (Peducasse et al., 1999). Hence, we simulated ver-
tical transmission of NNV in zebrafish by injecting the virus to

M.-W. Lu et al. / Molecular Immunology 45 (2008) 1146–1152 1149
Fig. 3. Detection of NNV antigen in the brain of infected zebrafish by immunohistochemistry. (a) Negative control obtained from the brain of a mock-infected
zebrafish (scale bar = 100 ␮m); (b) brain of an infected zebrafish on day 4 p.i., showing immuno-positive cells associated with brain tissue vacuolization (arrow)
(scale bar = 100 ␮m).
zebrafish larvae. NNV was propagated in E-11 cells and injected
to zebrafish larvae at 1 ×103TCID50 ml−1. The result showed
98% mortality at 24 h p.i. compared to 24% mortality in the
mock injected group (Fig. 5a). Quantitative RT-PCR confirmed
the presence of NNV in dead larvae from the NNV infected
group. Furthermore, we also compared the amounts of viral
RNA between larval and adult stage by quantitative RT-PCR.
The amount of viral RNA2 in larval stage was 0.06 ng, which was
much higher than adult brain sample (1.33 ×10−6ng) (Fig. 5b).
3.5. The interferon response in larval and adult stage after
NNV infection
The mortality of larvae is very distinct from adult fish after
NNV infection. In larval stage, the mortality is higher than 95%
amongst all of marine species which could be infected by NNV
(Barke et al., 2002). In this study, we collected the larval and
adult stage samples which have been infected by NNV and ana-
Fig. 4. Re-isolation of NNV from infected zebrafish. (a) Normal E-11 cell mono-
layer was grown in L-15 with 10% fetal bovine serum for 3 days. Cells were
maintained at 28 ◦C. (b) Monolayer of E-11 cells showing CPE upon infection
with NNV isolated from zebrafish for 3 days.
lyzed for innate immune response. The IFN-␣gene was not
detectable in the larvae stage but expressed normally in the adult
stage (Fig. 6a). The differential expression pattern of activated
Mx gene and phosphorylated eIF2␣gene, which are downstream
genes of interferon pathway, between larval and adult stage at
16 h p.i. (Fig. 6a and b) indicated the importance of interferon
response during NNV pathogenesis. We further confirmed the
observations by introducing interferon in the early stages of
zebrafish larvae followed by NNV infection. The dosage of uni-
Fig. 5. Cumulative mortality for the three groups of zebrafish larvae and the
differential amount of viral RNA between zebrafish larval and adult stage.
(a) Control group was untreated. Mock-injection group was injected with L-
15 medium. NNV injected group was injected with supernatant from a NNV
infected cell culture at a dosage of 1 ×103TCID50 ml−1. (b) Quantitative RT-
PCR of the NNV RNA2 was performed in larval and adult stage after NNV
infection.

1150 M.-W. Lu et al. / Molecular Immunology 45 (2008) 1146–1152
Fig. 6. Interferon related gene expression in larval and adult stage. (a) Semi-
quantitative RT-PCR analysis to detect the temporal expression of interferon
and Mx gene in larval and adult stage after NNV infection. (b) Immuno-blot of
the eIF2␣phosphorylation protein was performed in larval and adult stage after
NNV infection.
versal IFN-␣2a (Roche) that was introduced by microinjection
was 1000U. We designed four experimental groups, the first
group was the positive control for NNV infection, the second
group was co-injected with NNV and interferon, the third group
was pretreated with interferon for 48 h before inoculated with
NNV, and the last group was mock inoculated with PBS. The
group that was co-injected with interferon and NNV displayed
delayed mortality rate, however, after 96 h the mortality still
reached 100%, similar to positive control at 72 h. In addition, the
pretreatment with interferon for 48 h could rescue the survival
rate of NNV infected group from 0 to 76% which is similar to
mock-infected group (Fig. 7a). The viral RNA was also beyond
detectable amounts as shown by RT-PCR (Fig. 7b).
4. Discussion
Betanodavirus, one of the most important causal agent of viral
nervous necrosis resulting in economical losses to the aquacul-
ture industry worldwide. In recent years, betanodaviruses have
been detected in a variety of fish species both in temperate and
tropical areas, as well as in fresh water species (Azad et al.,
2005; Hegde et al., 2003; Johansen et al., 2003, 2004; Sommer
et al., 2004; Starkey et al., 2001; Ucko et al., 2004). Experi-
mental infection models have been established in different host
species, such as halibut, sea bass and wolfish (Grove et al.,
2003; Peducasse et al., 1999; Sommer et al., 2004) but unsuc-
cessful in the mouse system (Banu and Nakai, 2004). In this
study, we demonstrated that zebrafish are also susceptible to
NNV infection. This is also the first NNV infection model in
Fig. 7. The effects of interferon in zebrafish larval stage after NNV infection. (a)
Survival rate of groups with differential interferon treatment. The arrow indicates
the injection time. (b) Semi-quantitative of the viral RNA amount of differential
interferon treatment.
zebrafish in which, its genome project is near completion, thus
will accelerate the mechanistic study of NNV pathogenesis. RT-
PCR showed that viral RNA was most abundant in the brain
followed by eye and gut but was less in liver and heart. Infec-
tivity assay also confirmed the presence of virus in the brain,
eye and gut. The high virus titer in the gut compared to liver
and heart, which do not characteristically support replication of
NNV, may imply a role of the gut in the disease pathogenesis and
as a possible virus entry route via feed, such as contaminated
plankton. In persistently infected juvenile Atlantic halibut, viral
RNA can also be detected in abdominal organs by RT-PCR, but
it is not clear which organ the virus resides in (Johans et al.,
2002).
Herpes simplex virus, and rabies virus are spread by axonal
transport from the skin or muscle to the corresponding dorsal
root ganglion or anterior horn cells and then to populations of
neurons throughout the CNS. In contrast, arboviruses (mainly
togaviruses, flaviviruses, and bunyaviruses) spread to the brain
via the blood. In NNV-infected zebrafish, high titers of virus
occurred in the gut and eye as well as brain. These results indi-
cated a possible spread of virus from the central lacteal via blood
into the brain and eye. However, this hypothesis needs more
detailed investigation.
Beginning from 14 days post-NNV-infection, brain tissue was
collected daily from zebrafish that was in persistent infection.
From 3 days to 14 days p.i., no apparent symptom was observed
in the adult zebrafish. Hand in zebrafish larvae, however, in the
zebrafish larvae, mortality following injection of a brain tis-
sue homogenate from experimentally infected zebrafish reached