JOURNAL OF VIROLOGY Feb 2007 p 1162 1173 Vol 81 No 3 0022 538X 07 08 00H110010 doi 10 1128 JVI 01702 06 Copyright 2007 American Society for Microbiology All Rights Reserved Severe Acute Respiratory Syndrome Coronavirus Infection of Mice Transgenic for the Human Angiotensin Converting Enzyme 2 Virus Receptor H17188 Chien Te K Tseng 1 3 Cheng Huang 1 Patrick Newman 2 Nan Wang 1 Krishna Narayanan 1 Douglas M Watts 2 3 Shinji Makino 1 3 Michelle M Packard 4 Sherif R Zaki 4 Teh sheng Chan 1 and Clarence J Peters 1 2 3 Departments of Microbiology and Immunology 1 and Pathology 2 and the Center of Biodefense and Emerging Infectious Disease 3 University of Texas Medical Branch Galveston Texas and Centers for Disease Control and Prevention Atlanta Georgia 4 Received 7 August 2006 Accepted 10 November 2006 Animal models for severe acute respiratory syndrome SARS coronavirus infection of humans are needed to elucidate SARS pathogenesis and develop vaccines and antivirals We developed transgenic mice expressing human angiotensin converting enzyme 2 a functional receptor for the virus under the regulation of a global promoter A transgenic lineage designated AC70 was among the best characterized against SARS coronavirus infection showing weight loss and other clinical manifestations before reaching 100 mortality within 8 days after intranasal infection High virus titers were detected in the lungs and brains of transgene positive Tg H11545 mice on days 1 and 3 after infection Inflammatory mediators were also detected in these tissues coinciding with high levels of virus replication Lower virus titers were also detected in other tissues including blood In contrast infected transgene negative Tg H11546 mice survived without showing any clinical illness Pathologic examination suggests that the extensive involvement of the central nervous system likely contributed to the death of Tg H11545 mice even though viral pneumonia was present Preliminary studies with mice of a second lineage AC63 in which the transgene expression was considerably less abundant than that in the AC70 line revealed that virus replication was largely restricted to the lungs but not the brain Importantly despite significant weight loss infected Tg H11545 AC63 mice eventually recovered from the illness without any mortality The severity of the disease that developed in these transgenic mice AC70 in particular makes these mouse models valuable not only for evaluating the efficacy of antivirals and vaccines but also for studying SARS coronavirus pathogenesis Severe acute respiratory syndrome SARS coronavirus CoV is a highly transmissible human pathogen which emerged in late 2002 in southern China and spread to Asian and western hemisphere countries The disease was success fully contained by July of 2003 following the application of intensive public health measures but not before causing H110118 000 clinical cases with an H1101110 case fatality and tremen dous economic impact worldwide The most likely hypothesis for the emergence of SARS CoV is that the virus from the natural reservoir presumably the Chinese horseshoe bat Rhi nolophus sinicus adapted to infect civets which were permis sive and resulted in an epidemic among civets which were sold in the southern China food markets 19 21 The virus then spread to humans and underwent further genetic adaptation particularly to the spike protein to become more efficiently transmissible among the human population 22 35 It seems unlikely that this first emergence of SARS will be a unique event because many viruses such as Ebola virus Venezuelan equine encephalitis virus and epidemic influenza viruses have all returned after a hiatus in transmission Thus the need for effective antiviral agents and vaccines would be essential should SARS reemerge in the future Animal models are crucial to understanding the pathogen esis of human SARS and evaluating the efficacy of antiviral drugs and vaccines Several animal models for SARS have been proposed namely nonhuman primates i e macaques African green monkeys and marmosets ferrets hamsters and mice including young and aged BALB c and C57BL 6 mice and types lacking components of the immune system i e Stat1 and RAG1 knockout mice 9 18 23 31 32 36 37 These animals were shown to be susceptible to SARS CoV infection and showed viral replication some degree of histo pathology and occasionally limited clinical illness However none exhibited consistent clinical illness or mortality Addi tionally all suffer from some disadvantages including high cost poor availability of reagents and an immunological re sponse profile to the infecting virus quite unlike that observed in the human disease Aged mice in keeping with elderly humans have more pathology than do normal mice However a mild weight loss has been the only clinical manifestation in response to SARS CoV infection Stat1 deficient mice show more pronounced changes than do normal mice but there is no mortality and the pathological changes are not typical of those found with human SARS The tropism of coronaviruses is determined primarily by the interaction of the spike S protein and the cellular receptors for the virus Human angiotensin converting en Corresponding author Mailing address Department of Microbi ology and Immunology University of Texas Medical Branch 301 Uni versity Boulevard G 150 Keiller Building Galveston TX 77555 0609 Phone 409 772 0175 Fax 409 747 0762 E mail sktseng utmb edu H17188 Published ahead of print on 15 November 2006 1162 on April 30 2015 by UNIV OF SYDNEY http jvi asm org Downloaded from zyme 2 hACE2 has been identified as a major receptor for SARS CoV The spike protein of SARS CoV has a much higher binding affinity to hACE2 than do those of mice rats and other animal species which correlates with the much lower level of permissiveness of these animals to this virus 22 Thus one strategy for establishing an economical and suitable animal model for SARS is to establish transgenic mouse lines express ing hACE2 We have generated five lineages of such a trans genic mouse among which three die in response to SARS CoV infection Here we present detailed information about the infectivity and tissue distribution of SARS CoV virus associ ated histopathology inflammatory responses and the clinical manifestations in transgenic mice of the AC70 line one of the three lineages that succumbed to an acute and fatal infection We also present the data from the preliminary studies with transgenic AC63 mice one of the nonlethal lineages showing that despite of the onset of significant weight loss and other signs of illness infected AC63 mice eventually recovered from the infection without any mortality Importantly in contrast to AC70 mice in which both lungs and brains are the major sites of infection virus replication in the AC63 mice is largely re stricted to the lungs Taken together the severity of illness and or the fatal outcome of these transgenic mouse lineages make them attractive models for the evaluation of the prophy lactic and therapeutic efficacy of antiviral drugs and vaccines against SARS CoV infection MATERIALS AND METHODS Construction and expression of the hACE2 transgene The cDNA coding for hACE2 was generated by reverse transcription PCR RT PCR amplification from a human colon carcinoma cell line Caco2 which supported SARS CoV replication 24 The resulting PCR product was cloned into the pSTblue 1 cloning vector Novagen and the entire region corresponding to the ACE2 gene was confirmed by sequencing not shown The cDNA fragment containing the ACE2 sequences was subsequently cloned into a eukaryotic expression vector pCAGGS MCS a gift from Yoshihiro Kawaoka University of Wisconsin at Madison under the control of the CAG promoter a composite promoter consisting of the cytomegalovirus immediate early CMV IE enhancer and the chicken H9252 actin promoter and containing the rabbit globin splicing and polyaden ylation site To verify the expression of hACE2 human embryonic kidney 293 cells were transfected with the resulting plasmid construct designated pCAGGS ACE2 Fig 1A using Lipofectamine 2000 reagent Invitrogen Carlsbad CA per the manufacturer s protocols Cell extracts were prepared 24 h after trans fection and the expression of hACE2 was examined by Western blot analysis using polyclonal antibody against hACE2 R and for RNA 1 forward 5H11032 TCTGCGGATGCATCAACGT 3H11032 and reverse 5H11032 TGTAAGACGGGCTG CACTT 3H11032 with the sequence of CCGCAAACCCGTTTAAA as a detection probe all of these were derived by using the Assays by Design software Ap plied Biosystems The selected primer set and Taq Man probe for 18S rRNA were used as the endogenous control Briefly 80 ng RNA was transferred to separate tubes for amplification of the target genes and endogenous control 18S rRNA respectively by using a TaqMan one step RT PCR master mix reagent kit The cycling parameters for one step RT PCR were as follows reverse tran scription at 48 C for 30 min AmpliTaq activation at 95 C for 10 min denatur ation at 95 C for 15 s and annealing extension at 60 C for 1 min A total of 40 cycles were performed on an ABI PRISM 7000 real time thermocycler Applied Biosystems following the manufacturer s instructions DNA fragments encoding target genes were amplified in triplicate and relative mRNA levels for each sample were calculated as follows H9004C T H11005 C T of target genes H11002 C T of 18S rRNA The relative abundance of the RNA for hACE2 or for SARS CoV was expressed as 2 H11002H20849H9004CT infected H11002H9004CT mockH20850 Histology and IHC Multiple tissues obtained from necropsy were fixed in 10 buffered formalin for 72 h transferred to 70 ethanol and later paraffin em bedded Histopathologic evaluation was performed on deparaffinized sections stained by routine hematoxylin and eosin staining Immunohistochemical IHC testing for SARS CoV was applied using a previously described colorimetric indirect immunoalkaline phosphatase method 36 with a rabbit anti SARS CoV nucleocapsid protein antibody Imgenex catalog no IMG 548 The goat anti human ACE2 antibody R no such expression was detected in Tg H11002 mice and no staining was seen using normal goat serum as a negative con trol data not shown Although two color staining was not performed to colocalize hACE2 and viral antigen expression we noticed that in the lungs and GI tract the viral distribution correlated well with the pattern of expression observed for hACE2 In the lungs of Tg H11001 mice hACE2 was detected pri marily in the pneumocytes vascular smooth muscle and gan glion cells Fig 6A and B Expression was found focally in the muscularis and subserosal ganglia of the GI system in similar areas to those where SARS CoV antigen was present Fig 6H In comparison in the CNS the distribution of viral anti gen and hACE2 was significantly different High levels of hACE2 expression were detected in choroid ventricular lining and vascular endothelial cells while only rare neurons and glial cells showed minimal expression of hACE2 Fig 6C to G However intense staining of SARS CoV was only detected in neuron and glial cells above suggesting that not all of the hACE2 expressing cells are susceptible to the infection SARS CoV induced cytokines and chemokines in the lungs and brains of mice The mechanism of SARS associated lung pathology remains unknown However pathological studies with postmortem specimens of SARS patients reveal diffuse alveolar damage DAD hemophagocytosis and prominent infiltration of activated macrophages in the lungs which sug FIG 5 Histopathology and immunohistochemical analysis of SARS CoV antigen expression in the lungs brain and GI tract of Tg H11001 mice after infection i n Paraffin embedded lung A to G brain H and I and GI tract J sections of infected Tg H11001 mice were analyzed for the pathology and expression of the nucleocapsid protein of SARS CoV by the methodologies described in Materials and Methods SARS CoV antigen red was readily detectable in the cytoplasm of epithelial cells of the bronchial lining A and pulmonary interstitium B at day 2 No staining was seen in the same Tg H11001 mouse when immunohistochemistry was performed with normal mouse ascites fluid C Shown are serial sections hematoxylin and eosin in panel D and IHC in panel E of a bronchus showing intraluminal macrophages and cellular debris in association with viral antigen F Inflammatory cellular infiltrates arrow within smooth muscle of a pulmonary blood vessel associated with SARS CoV antigen G SARS CoV immunostaining of a subepithelial ganglion cell in the lung at day 2 Extensive SARS CoV antigen expression was first detected on day 3 in large numbers of morphologically intact neuronal and glial cells in the CNS H and I In the GI tract the expression of SARS CoV antigen in ganglia within the subserosal layer arrow was detected first at day 4 J Magnifications A to F H and J H11003100 G H11003158 I H1100350 Staining panels A to C and E to J IHC with naphthol red and hematoxylin counterstaining panel D hematoxylin and eosin VOL 81 2007 SARS CoV INFECTION OF MICE TRANSGENIC FOR HUMAN ACE2 1167 on April 30 2015 by UNIV OF SYDNEY http jvi asm org Downloaded from gest that an intense and unregulated inflammatory response within the lungs may be partially responsible for the pathogen esis of human SARS CoV infection 26 The severity of the disease developed in Tg H11001 mice in re sponse to SARS CoV infection prompted us to study the host responses by measuring the contents of various inflammatory mediators in the lungs and brain two of the most affected tissues As shown in Fig 7 among 23 inflammatory mediators measured we were able to detect negligible levels of inflam matory cytokines in the lungs of infected Tg H11002 mice over time compared to the levels in uninfected controls suggesting that SARS CoV infection failed to induce cytokine production in Tg H11002 mice In contrast elevated levels of interleukin 1H9252 IL 1H9252 IL 12 p40 IL 12 p70 CXCL1 KC RANTES and MCP 1 expression were readily detected in the lungs of Tg H11001 mice in at least one time point during the first 3 days of the infection Fig 7 The SARS CoV induced inflammatory response in the brain was similarly evaluated There was no significant expres sion of inflammatory mediators in infected Tg H11001 and Tg H11002 mice on both day 1 and day 2 data not shown However highly elevated levels of IL 6 IL 12 p40 granulocyte colony stimulat ing factor G CSF CXCL1 KC MIP 1H9251 and MCP 1 were detected at day 3 in the brains of Tg H11001 mice but not their Tg H11002 littermates Table 1 Additionally the secretion of IL 1H9251 IL 1H9252 granulocyte macrophage CSF GM CSF IL 12 p70 and RANTES was increased to various extents in the brain of the Tg H11001 mouse Other inflammatory mediators such as IL 2 IL 3 IL 4 IL 5 IL 9 IL 10 IL 13 IL 17 gamma interferon IFN H9253 and tumor necrosis factor alpha TNF H9251 were not signif icantly induced Taken together these results clearly demon strate that Tg H11001 mice which had much higher levels of SARS CoV replication in the lungs and brain than their Tg H11002 littermates as shown in Fig 3 could promptly elicit a strong inflammatory cytokine reaction to acute SARS CoV infection Importantly this acute inflammatory response occurred later and more intensely in the brain than in the lungs consistent with the level of virus replication in respective organs SARS CoV infection in transgenic AC63 mice The suscep tibility of Tg H11001 AC63 mice to SARS CoV infection was initially evaluated by using the same challenge strategy i e 10 3 TCID 50 of SARS CoV via the i n route The Tg H11001 AC63 mice were more susceptible to SARS CoV infection than their Tg H11002 littermates as evidenced by a moderate but progressive weight loss until day 8 Fig 8A However in contrast to the uniform mortality of infected Tg H11001 AC70 mice all infected AC63 mice eventually recovered from the weight loss without any death We next investigated the tissue distribution of infectious virus and whether inoculation with a higher dose of virus could result in a fatal outcome We inoculated i n both AC63 mice and control littermates 10 of each with 10 6 TCID 50 of virus To quantify the viral loads in the lungs and brain five and three mice from each group were sacrificed at days 3 and 8 respec tively The remaining animals were kept to assess morbidity and mortality Infected Tg H11001 but not Tg H11002 mice started to show an progressive weight loss along with other clinical manifes tations between day 3 and day 4 As shown in Fig 8B Tg H11001 mice appeared to be more susceptible to SARS CoV infection than their Tg H11002 littermates as evidenced by a much higher titer of infectious virus in the lungs Additionally four out of five Tg H11001 mice had high virus titers in the lungs whereas only two FIG 6 Expression of hACE2 in the lungs brain and GI tract of Tg H11001 mice The paraffin embedded sections of the lungs brains and GI tract were used to evaluate the expression of the hACE2 by IHC The hACE2 antigen red was readily detectable primarily in the pneumocytes A and vascular smooth muscle in the lung B arrow The hACE2 expression in the brain was also abundantly associated with choroid C ventricular lining D vascular endothelial cells E and patches of neuronal and glial elements F and G Finally hACE2 was also found in the epithelial lining muscularis layer and ganglia of the GI system H arrow Magnification panels A F and G H11003158 panels B D and H H1100350 panel C H1100325 and panel E H11003100 1168 TSENG ET AL J VIROL on April 30 2015 by UNIV OF SYDNEY http jvi asm org Downloaded from had low to moderate titers in the brain at day 3 Infectious virus was no longer detectable in the lungs at day 8 even though one animal still had detectable virus in the brain Re markably despite the severity of the illness as evidenced by the profound weight loss the remaining animals started to show signs of recovery between day 8 and day 9 regained some of the lost weight in 1 week thereafter Fig 8C and recovered completely in the ensuing 1 month when the experiment was terminated DISCUSSION Animal models for SARS in well characterized species that consistently reveal signs of illness pathological findings and mortality are highly desirable not only for studying pathogen esis but also for evaluating the safety and efficacy of antiviral therapeutics and vaccine candidates against SARS CoV infec tion In this study we developed a small animal model for SARS using transgenic mice expressing hACE2 the major cellular receptor for SARS CoV 20 Not only could this transgenic mouse model support more robust viral growth than its nontransgenic littermates but it also manifests respiratory FIG 7 Expression of pulmonary cytokines and chemokines in infected mice Lung homogenates derived from mice at indicated time intervals after infection i n were subjected to Bio Plex analysis for assessment of the concentrations of cytokines and chemokines Among 23 inflammatory mediators tested the expression of IL 1H9252 IL 12 p40 CXCL1 KC RANTES MCP 1 and IL 12 p70 was elevated in infected Tg H11001 but not Tg H11002 mice Duplicate samples of individual specimens were assayed The data shown are the mean H11006 standard error of infected animals n H11005 3 at indicated time points P H11021 0 05 and P H11021 0 01 by Student s t test comparing Tg H11001 mice with aged matched Tg H11002 controls TABLE 1 SARS CoV induced production of cytokines and chemokines in the brains of infected Tg H11001 versus Tg H11002 mice Inflammatory mediator Cytokine or chemokine concn pg ml in mouse brain a Tg H11002 Tg H11001 IL 6 3 2 H11006 1 328 H11006 40 IL 12 p40 17 H11006 2 2 500 H11006 850 G CSF 2 6 H11006 1 523 H11006 58 KC 25 H11006 4 466 H11006 75 MIP 1a ND 311 H11006 35 MCP 1 ND 904 H11006 210 IL 1H9251 7 5 H11006 127H11006 4 IL 1H9252 0 8 H11006 0 5 31 4 H11006 8 GM CSF 2 5 H11006 111H11006 2 IL 12 p70 12 H110