the double-stranded rna-activated protein kinase mediates ...the double-stranded rna-activated...
Embed Size (px)
The double-stranded RNA-activated protein kinase mediatesviral-induced encephalitis
Donalyn Scheuner,a Matthias Gromeier,b,1 Monique V. Davies,c Andrew J. Dorner,c
Benbo Song,a Rupali V. Patel,d Eckard J. Wimmer,b Roger E. McLendon,e andRandal J. Kaufmana,d,*
a Howard Hughes Medical Institute, University of Michigan, Ann Arbor, MI 48109, USAb Department of Molecular Genetics and Microbiology, State University of New York at Stony Brook, NY 11794-8621, USA
c The Genetics Institute, Inc., Cambridge, MA 02140, USAd The Department of Biological Chemistry, University of Michigan, Ann Arbor, MI 48109, USA
e Department of Pathology, Duke University Medical Center, Durham, NC 27710, USA
Received 24 March 2003; returned to author for revision 30 April 2003; accepted 1 August 2003
The double-stranded (ds) RNA-activated protein kinase (PKR) plays an important role in control of viral infections and cell growth. Wehave studied the role of PKR in viral infection in mice that are defective in the PKR signaling pathway. Transgenic mice were derived thatconstitutively express a trans-dominant-negative kinase-defective mutant PKR under control of the �-actin promoter. The trans-dominant-negative PKR mutant expressing transgenic mice do not have a detectable phenotype, similar to observations with PKR knock-out mice.The requirement for PKR in viral pathogenesis was studied by intracerebral infection of mice with a mouse-adapted poliovirus.Histopathological analysis revealed diffuse encephalomyelitis with severe inflammatory lesions throughout the central nervous system(CNS) in infected wild-type mice. In contrast, histopathological evaluation of virus-injected trans-dominant-negative PKR transgenic miceas well as PKR knock-out mice yielded no signs of tissue damage associated with inflammatory host responses. However, the virus didreplicate in both models of PKR-deficient mice at a level equal to that observed in wild-type infected mice. Although the results indicatea clear difference in susceptibility to poliovirus-induced encephalitis, this difference manifests clinically as a slight delay in fatal neuropathyin trans-dominant-negative PKR transgenic and PKR knock-out animals. Our observations support the finding that viral-induced PKRactivation may play a significant role in pathogenesis by mediating the host response to viral CNS infection. They support PKR to be aneffective target to control tissue damage due to deleterious host responses to viral infection.© 2003 Elsevier Inc. All rights reserved.
Keywords: Interferon; Poliovirus; Eukaryotic Translation initiation factor 2; CNS infection
The interferon-induced double-stranded RNA-activatedprotein kinase (PKR) was first identified as a component ofthe host-defense mechanism induced by type I interferons
(interferon-� and -�) (Samuel, 1991). PKR is a serine/threonine protein kinase ubiquitously expressed in mamma-lian cells (Levin and London, 1978; Kaufman, 2000). Al-though it has long been established that treatment withinterferon and dsRNA is cytotoxic to cell cultures in vitro(Lengyel, 1987), it is now recognized that this antiprolif-erative and proapoptotic response is mediated through thePKR signaling pathway (Chong et al., 1992; Koromilas etal., 1992; Meurs et al., 1993). The mechanism by whichPKR inhibits cell growth and promotes apoptosis in inflam-mation is unknown. However, several signaling pathwaysimplicated include phosphorylation of eukaryotic transla-
* Corresponding author. Howard Hughes Medical Institute, 4570MSRBII, 1150 W. Medical Center Dr., University of Michigan MedicalCenter Ann Arbor, MI 48109. Fax: �1-734-763-9323
E-mail address: [email protected] (R.J. Kaufman).1 Present address: Dept. of Molecular Genetics and Microbiology,
Duke University Medical Center, Durham, NC 27710.
Available online at www.sciencedirect.com
Virology 317 (2003) 263–274 www.elsevier.com/locate/yviro
0042-6822/$ – see front matter © 2003 Elsevier Inc. All rights reserved.doi:10.1016/j.virol.2003.08.010
tion initiation factor 2 on the alpha subunit (eIF2�) (Donzeet al., 1995; Srivastava et al., 1998), and activation oftranscription factors NF�B (Kumar et al., 1994; Chu et al.,1999), STAT1 (Ramana et al., 2000), interferon regulatoryfactor 1 (IRF1) (Kumar et al., 1997), and the tumor sup-pressor p53 (Cuddihy et al., 1999a,b). In addition, recentobservations suggest that signaling through PKR for acti-vation of NK�B does not require its protein kinase activity(Chu et al., 1999; Bonnet et al., 2000; Ishii et al., 2001).
Given the negative role of PKR on cell growth, it wassurprising that mice deleted of the PKR gene did not have areadily detectable phenotype (Yang et al., 1995; Abraham et
al., 1999). Two PKR-deficient mouse models were created;one deletion was targeted to the amino-terminal dsRNAbinding domain (Yang et al., 1995) and another was tar-geted to the carboxy-terminal kinase domain (Abraham etal., 1999). Curiously, these mice do not form spontaneoustumors as observed with mice harboring knock-outs in othertumor suppressor genes such as p53 (Donehower et al.,1992). In addition, the defect in induction of type 1 inter-feron and activation of NF�B in response to poly(I:C) inembryonic fibroblasts from PKR-deficient mice was cor-rected by pretreatment with interferon (Yang et al., 1995).Therefore, it was suggested that the deletion creates a partial
Fig. 1. Derivation and characterization of PKR(K296R) transgenic mice. (A) The PKR(K296R) expression vector. The mutant K296R catalytically inactivehuman mutant PKR cDNA was isolated on a HindIII-PstI restriction fragment and cloned into the SalI site of the human �-actin promoter vector (pHBAPr-1)(Gunning et al., 1987) to yield pBapPKR(K296R). The human �-actin promoter is followed by 50 bp of �-actin 5� UTR plus intervening sequence 1. TheSV40 polyadenylation signal and plasmid sequences are derived from pcDV1 (Okayama and Berg, 1983). ClaI digestion was used to liberate the 6.4-kbtransgene construct DNA from plasmid sequence. (B) Analysis of PKR(K296R) transgene copy number. Genomic tail DNA was isolated from F1 mice (1–15,2–28, 3–45, 4–57, 5–3, 6–20) and digested with AccI in comparison against NIH3T3 fibroblast control DNA and NIH3T3 fibroblast DNA spiked withpBapPKR(K296R). Genomic Southern hybridization was performed using the random primed transgene construct DNA (6.4 kb ClaI) as a probe. The F1analyzed presented three major hybridizing bands (0.8, 2.6, and 3.0 kb) representative of multicopy head-to-tail integration of the transgene construct.Quantitation of the hybridizing bands against the pBapPKR standards yielded an estimate of gene copy number (Table 1). (C) Expression of PKR(K296R)mRNA in transgenic mouse tissues. Total RNA was isolated from tissues of F1 mice obtained from the six founder mice (In, intestine; Kd, kidney; Gd, gonad;Lg, lung; Sp; spleen; Br; brain; Lr, liver; Ms, muscle; Ht, heart; Sk, sketetal muscle). Reverse transcription and PCR products were quantitated after Southernblotting and hybridization with an internal oligonucleotide probe (see Materials and methods). Representative blots F1 3–45 and 6–20 are shown and relativeexpression levels were quantitated and are summarized in Table 1. Competitive RT-PCR was performed using a constant amount of RNA (1 �g) isolatedfrom heart tissue, but altering the ratio of transgenic:nontransgenic RNA in each reaction (see Materials and methods). (D) Analysis of PKR K296P hPKRexpression in transgenic mice and MEFs. Western blot analysis of murine heart tissues from control and transgenic mice was performed with monoclonalanti-PKR antibody 71–10. Anti-human PKR polyclonal antibody was used for Western blot detection of PKR in extracts from MEFs. (E) Phosphorylationof eIF2� in PKR(K296R) transgenic and wild-type MEFs. MEFs were stimulated for 16 h with 100 �g/ml pIC in the presence or absence of actinomycinD (50 ng/ml) added to culture medium. Phosphorylation of eIF2� was measured by Western blotting with antiphosphorylated Ser51 eIF2� antibody followedby stripping and reprobing with antitotal eIF2� antibody.
264 D. Scheuner et al. / Virology 317 (2003) 263–274
loss-of-function in PKR consistent with the potential for thedeleted allele to produce an amino-terminal truncated PKRprotein in these mice (Yang et al., 1995). Alternatively, aredundant PKR pathway might exist in the PKR-deletedmice. Recently, a mammalian homologue of yeast GCN2gene (Sood et al., 2000) and an endoplasmic reticulum (ER)localized eIF2� kinase that is activated upon ER stress wereidentified (Shi et al., 1998; Harding et al., 1999). Therefore,it is possible that another PKR-related eIF2� kinase existsthat may respond to dsRNA and provide tumor suppressoractivity in the absence of PKR. To directly test the require-ment for PKR kinase activity and the potential existence ofadditional PKR homologues, we derived mice that are de-fective in the PKR pathway by expression of a transgenethat encodes a trans-dominant negative mutant of PKR that
is defective in kinase activity (K296R). Transgenic micethat express this trans-dominant-negative PKR mutant willelucidate the requirement for PKR kinase activity in PKR-dependent responses. Whereas wild-type mice display se-vere encephalitis upon intracerebral viral infection, thesemutant PKR transgenic mice, as well as PKR-null mice,were protected from the encephalitis lesions.
To evaluate the role of PKR in viral pathogenesis, weused a transgenic mouse strain that constitutively expressesthe K296R mutant of human PKR under control of theconstitutive �-actin promoter (Fig. 1A) (Gunning et al.,
Fig. 1 (continued)
265D. Scheuner et al. / Virology 317 (2003) 263–274
1987). The K296R mutation was previously demonstratedto act in a trans-dominant negative manner to inhibit theendogenous PKR (Chong et al., 1992). Southern blot anal-ysis was used to quantitate that six different founder micehad between 2 and 25 copies of transgene/genome (Fig. 1B;Table 1). The transgene expression unit was intact in allfounders and multiple copies were found in a “head-to-tail”tandem repeat arrangement, typical of transgene insertion.Northern blot analysis detected a 2.1-kb human PKRmRNA transcript in tissues from transgenic mice (data notshown). RT-PCR detected transgene mRNA in multipletissues (Fig. 1C shows data for founders 3–45 and 6–20).
PKR (K296R) mRNA expression for all founders is sum-marized in Table 1. To quantitate transgene expressionrelative to the endogenous murine PKR, competitive RT-PCR was performed. This analysis demonstrated that trans-gene mRNA was present at levels approximately five-foldgreater than that of the endogenous PKR transcript in twodifferent founders (3–45, 5–3) (Fig. 1C, bottom, shows datafor founder 3–45). These mice have been under continualobservation for over 4 years and appear phenotypicallynormal. Subsequent studies were performed with mice de-rived from founder 3-45.
PKR transgene expression was monitored by Western
Table 1Expression of PKR(K296R) in transgenic mouse tissues
Founder F1 Genecopyno.
PKR (K296R) mRNA
In Kd Gd Lg Sp Br Lr Ms Ht Sk
1 15 25 � � �� ��� � � � ��� ��� �2 28 3 �� �� ��� ��� � �� � � ��� �3a 45 5 � �� �� �� �� �� � �� �� �4 57 20 � � � � � � � � � �5 3 2 � �� ��� �� � � � ��� ��� �,�6 20 3 �� � ��� ��� � � �� � �� �
Note. Expression of PKR(K296R) in transgenic mouse tissues. Gene copy number was determined by genomic Southern analysis (Fig. 1B). Total RNAwas isolated from F1 intestine, kidney, gonads, lung, spleen, brain, liver, muscle, heart, and skin. Relative mRNA expression levels were derived fromquantitative RT-PCR. Expression levels were highest in heart, lung, and gonadal tissue.
a All subsequent studies of the PKR transgenic mice were performed with mice derived from 3–45.
Fig. 1 (continued)
266 D. Scheuner et al. / Virology 317 (2003) 263–274
blot analysis of heart tissue and murine embryonic fibro-blasts (MEFs) (Fig. 1D). Human PKR was expressed at asignificant level in transgenic murine heart tissue (Fig. 1D;lanes 2–4). Although endogenous murine PKR in MEFswas expressed at very low levels under basal conditions,after induction by overnight treatment with IFN� (1000u/ml), PKR was detected in both wild-type and PKR(K296R) fibroblasts. No induced product was detected inPKR “knock-out” fibroblasts. A considerable level of trans-genic human PKR (K296R) protein was detected in thetransgenic MEFs. The apparent molecular weight of thehuman PKR (68 kDa) is slightly higher than the murineform (65 kDa), as expected.
To ascertain whether expression of the PKR transgenealtered the level of eIF2� phosphorylation in vivo, theeIF2� phosphorylation status was measured in MEFs.Western immunoblot analysis was performed with an anti-body specific for the phosphorylated form of eIF2� (Sriv-astava et al., 1998). The same blot was then reprobed withan eIF2� antibody that reacts with total eIF2�. The com-parison of the intensity from these two probings indicatesthe relative level of eIF2� phosphorylation. The basal levelof eIF2� phosphorylation was not altered in MEFs isolatedfrom PKR (K296R) transgenic mice and mice that harborthe amino-terminal PKR deletion compared to wild-typemice (Yang et al., 1995) (Fig. 1E). This is consistent withfindings that constitutive eIF2� phosphorylation is notmaintained by the PKR eIF2� kinase, but rather reflects theactivities of GCN2 and/or PERK (Harding et al. 2000;Zhang et al., 2002). The level of eIF2� phosphorylation wasalso measured after stimulation with dsRNA, poly(I):poly(C). Where treatment with poly(I):poly(C) increasedeIF2� phosphorylation in the control MEFs, phosphoryla-tion of eIF2� in the MEFs from the homozygous trans-dominant-negative PKR mutant mice was not increased(Fig. 1E, right). The level of eIF2� phosphorylation wasalso measured after stimulation with the apoptotic stimuli ofpoly(I):poly(C) in the presence of actinomycin D. Here,phosphorylation of eIF2� was again enhanced in the controlMEFs, whereas phosphorylation was not increased in MEFsfrom the heterozygous or homozygous mutant PKR mice.Interestingly, there was a 2.5-fold increase in the level ofeIF2� phosphorylation in the MEFs harboring thePKR(K296R) transgene when treated with interferon andpoly(I):poly(C) that was comparable to the increase in phos-phorylation observed in wild-type MEFs under these con-ditions (data not shown). It is likely that the interferon-induced expression of the endogenous PKR gene can titrateout the ability of the mutant transgene to inhibit eIF2�phosphorylation under the activating conditions of pICtransfection.
To elucidate whether PKR might play a role in viralencephalitis, we used a poliovirus model system for viralencephalitis. Although poliovirus cannot infect rodents nat-urally, a serotype 2 poliovirus strain isolated from a fatalcase of poliomyelitis in 1942 in Cairo, Egypt [PV2(MEF-
1)] has proven to express a mouse-neuropathogenic pheno-type after intracerebral inoculation (Schlesinger et al.,1943). The clinical and histopathological features of diffuseencephalomyelitis in rodents following infection with po-liovirus type 2 isolates are fundamentally different frompoliomyelitis occurring in primates (Gromeier et al., 1995).These differences can be ascribed to the absence of thehuman poliovirus receptor CD155 in rodents. CD155 me-diates the characteristic neuropathological features of polio-virus, restricting tropism to brain stem and spinal cordanterior horn motor neurons in primates (Gromeier et al.,1996). We have chosen PV2 infection as a model for ex-perimental encephalitis because PV2 encephalomyelitis inwild-type mice is a highly reproducible, rapidly progressivesyndrome dominated by severe and extensive inflammatorylesions in circumscribed areas of the brain resulting inirreversible tissue damage (Gromeier et al., 1995). Sincehistopathological damage appears to be mediated mainly byhost responses to viral infection, this model presents aunique opportunity to study the effect of PKR on inflam-matory responses to viral CNS infection.
Homozygous PKR (K296R) transgenic mice and non-transgenic control mice of the same genetic backgroundwere injected with PV2. The transgenic mice developedneurological symptoms after infection of PV2(MEF-1), butin contrast to age-matched nontransgenic control animals,the onset of disease was slightly delayed and progression ofsymptoms was protracted (Fig. 2). Histopathological eval-uation of brain and spinal cord tissue from both groups ofinfected animals found the differences in clinical suscepti-bility to correlate with a significant divergence in the sever-ity and histopathological pattern of CNS lesions.
In comparison with uninfected nontransgenic mice (Fig.3; left panels), infected nontransgenic C57Bl/6J mice (Fig.3; middle panels) developed a diffuse encephalomyelitiswith extensive lesions throughout the cerebral hemispheresand spinal cord similar to ICR–or Swiss–Webster micedescribed earlier (Gromeier et al., 1995). Damage, althoughdistributed throughout the brain (Fig. 3B; frontal cortex; E,hippocampal formation; and H, lumbosacral spinal cord),focused on the hippocampal formations bilaterally, with
Fig. 2. Progression to terminal neurological disease was protracted inPKR(K296R) mice. The % survival after intracerebral injection of PV2was evaluated in C57B1/6J (open symbols) and trans-dominant-negativePKR(K296R) (solid symbols) mice. Virus-injected mice were evaluateddaily after virus administration (day 0). All mice in the study succumbedto PV2 infection, but progression was protracted in the transgenic mice.
267D. Scheuner et al. / Virology 317 (2003) 263–274
severe involvement of adjacent occipital and piriform cor-tical areas, the basal ganglia, as well as the spinal cord.
In accordance with the histopathological characteristicsof viral encephalitis in humans, poliovirus-induced enceph-alomyelitis in C57Bl/6J mice was dominated by extensivemicrogliosis (Figs. 4C and E show characteristic lesions ofthe hippocampal formation and cortical areas at higher mag-nification). Microglial proliferation and invasion were mostevident in subependymal and periventricular areas, the hip-pocampal formation and the basal ganglia bilaterally, aswell as the spinal cord. The extent and intensity of enceph-alitis and its predominant feature microgliosis were equallygrave in all infected control C57Bl/6J animals studied. Inparticular, bilateral severe involvement of the hippocampal
formation was universally observed in infected mice (Figs.4A and C).
In sharp contrast, the transgenic mutant PKR (K296R)mice did not develop significant histopathological lesions inthe areas typically affected by PV2 (Fig. 3; right panels).The hippocampal formations, neighboring cortical areas,and all other brain regions of infected animals appearedrelatively unaffected (Figs. 3C, F, and J). Minor signs ofmicrogliosis in PKR(K296R) mice indicated that, despitethe quantitative differences to PV2-induced encephalitis inwild-type C57Bl/6J mice, the histopathological characteris-tics of lesions were similar (Figs. 4B, D, and F). Thehallmark signs of severe microglial profileration affectingthe hippocampal formations was absent in PKR(K296R)
Fig. 3. PKR(K296R) transgenic mice are protected from PV2-induced destruction of neural tissue. After intracerebral inoculation of 1 � 106 PFU and clinicalsymptoms had established in control or transgenic mice (see Materials and methods). All animals were sacrificed and neural tissues were processed forhistopathology. Histological sections showing the frontal cortex (A–C), the hippocampal formation (D–F), and the lumbosacral spinal cord (G–I) wereobtained and stained as described previously (Gromeier et al., 1996). Sections in the left column (A,D,G) were obtained from uninfected C57B1/6J mice;specimens in the middle column (B,E,H) were derived from C57B1/6J mice inoculated intracerebrally with 1 � 106 PFU of PV2 and those in the right column(C,F,I) originated from dominant-negative transgene PKR mice virus-injected with PV2 as in Fig. 2. A shows a horizontal section through the frontal cortexof an uninfected C57B1/6J mouse. The same region of an infected littermate shows an area of cortical damage with destruction of neuronal components(arrows) and small perivascular infiltrates (B). Cortical lesions within infected transgenic mice were absent (C). D shows horizontal section through thehippocampal formation from an uninfected mouse. The hippocampal formation invariably harbored severe lesions in PV2-injected C57B1/6J mice (E).Intense infiltration (black arrow) and cortical neurons in a zone of destruction (red arrow) are shown. F shows the hippocampal formation of an infectedtransgenic mouse. G shows a transversal section through a normal cervical spinal cord. Lesions within the upper regions of the spinal cord were always presentin C57B1/6J mice injected with PV2 strains (H). The black arrow indicates a region of severe infiltration, cell death, and complete tissue destruction. Sectionsfrom transgenic mice occasionally contained cytopathic neurons (I). Larger infiltrates were absent.
268 D. Scheuner et al. / Virology 317 (2003) 263–274
mice (Fig. 4B) but, on close examination, minor foci ofmicrogliosis and neuronal cell death could be detected inthese animals as well (Figs. 4D and F).
These findings indicated a difference in the degree ofsusceptibility toward poliovirus-induced encephalitis in thetrans-dominant-negative PKR expressing transgenic ani-mals. This difference was evident as a relative resistancetoward virus-induced tissue damage. Histopathological ex-
amination revealed the overall extent of encephalitis intrans-dominant-negative PKR expressing animals to be re-duced by approximately 90%. The difference in suscepti-bility to virus infection did not manifest itself as a distinctentity histopathologically or clinically (apart from slightlyprotracted progression), since all transgenic animals diddevelop fatal CNS disease (Fig. 2).
These observations prompted us to study PV2(MEF-1)
Fig. 4. Microscopic detail of hippocampal lesions in PV2-injected C57B1/6J (A, C, E) and PKR(K296R) mice (B, D, F). (A, B) Overview of the hippocampalformation; �12. Red and blue rectangles indicate the location of hippocampal and cortical regions shown at higher magnification, respectively. (C, D)Extensive microgliosis affecting the hippocampal formation (arrowheads) in wild-type animals was absent in PKR(K296R) mice. (E, F) Detail of corticalregions adjacent to the hippocampal formation is shown. Microglial proliferation extended cortically in wild-type mice (E, arrowhead). Occasionally,minuscule lesions could be detected in PKR(K296R) mice (F, arrowhead).
269D. Scheuner et al. / Virology 317 (2003) 263–274
infection in the amino-terminal PKR knock-out mouse(Yang et al., 1995). To this end, the spinal cords of PV2-infected C57Bl/6J, and PKR�/� mice were analyzed his-topathologically (Fig. 5). Since the spinal cord harbored themost severe virus-induced lesions in PV2-injected mice (seeFig. 3), we focused our analysis on this compartment of thecentral nervous system (CNS). In accordance with previousobservations, compared to uninfected animals (Figs. 5A andD), there was extensive damage to the spinal cord upon PV2infection (Figs. 5B and E). As previously described, thelesions were characterized by extensive microgliosis,perivasular microglial and macrophage invasion, and neu-ronophagia (Figs. 5B and E).
In contrast, PKR�/� mice were not free of virally in-duced damage but demonstrated a consistently much morebenign pattern of lesions (Figs. 5C and F). In this study, wealso compared poliovirus replication kinetics with tissuedamage to the brain and spinal cord. After inoculation withviral titers exceeding the LD100 100-fold, poliovirus repli-cation was observed in the PKR�/� brain and spinal cord atlevels comparable to those seen in wild-type mice (Fig. 6).
Animal viruses and their hosts have coevolved complexinterrelationships to permit virus reproduction without de-struction of the host organism (Shen and Shenk, 1995;
Kaufman, 1999). The interferon-induced cellular antiviralresponse is the first line of defense against viral infectionwithin the infected cell. Upon viral infection, the expressionof interferon is induced at the transcriptional level. Theinterferon is secreted to protect adjacent cells from second-ary infection, thereby limiting viral spread. The double-stranded (ds)RNA-activated protein kinase is the most well-characterized interferon-induced gene product that mediates
Fig. 5. PKR�/� mice are protected from poliovirus PV2-induced spinal cord destruction. PKR�/� mice and control mice were injected intracerebrally with1 � 108 PFU of PV2. After 4.5 days, all animals were sacrificed and their spinal cords were processed for histopathology. In infected controls, spinal corddestruction was severe and infiltration was extensive in the meninges and around spinal blood vessels (B, E). Uninfected controls are shown for comparison(A, D). In contrast, the PKR�/� animals appeared either normal (C, F) or contained minor infiltrative lesions. Arrowheads point to neurons in the anteriorhorn to show there are signs of neuronal damage also in the PKR knock-out mice. Cervical (A–C) and lumbar (D–F) sections of the spinal cord are shown.
Fig. 6. PV2(MEF-1) replication in PKR�/� and PKR�/� mice. Brain andspinal cord tissues were obtained from PKR�/� or wild-type mice atdifferent stages after intracerebral administration of PV2. Tissues werehomogenized and the viral titer of the homogenate was determined in aplaque assay. Initial viral recovery from brain tissue is high due to theintracerebral route of administration. The variability at time 0 is due todifferences in viral spread and recovery and is typical of these experiments.Replication of PV2(MEF-1) in wild-type mouse brain (�), spinal cord (E);and PKR�/� mouse brain (■ ), spinal cord (F).
270 D. Scheuner et al. / Virology 317 (2003) 263–274
the antiviral actions of type I interferons. However, themechanism(s) of protection provided by the PKR responseto the host upon viral infection is unclear and has onlyrecently been studied. It was surprising that initial studiesdemonstrated deletion of PKR had little effect upon infec-tion with EMCV and vaccinia virus (Yang et al., 1995;Abraham et al., 1999; Stojdl et al., 2000). Amino-terminal-deleted PKR-null mice injected with EMCV (1000 plaque-forming units (PFU) iv) did not show decreased survival inthe absence of interferon-� and only a slight effect uponpretreatment with interferon-� or poly(IC) (Yang et al.,1995). Similarly, infection of carboxy-terminal-deletedPKR-null mice with EMCV (100 PFU iv) showed only asmall difference in survival (Stojdl et al., 2000). Recently, itwas demonstrated that PKR is required for mice to surviveinfection with either VSV or influenza virus and to mount aresponse to interferon-�/� (Stojdl et al., 2000; Balachand-ran et al. 2000). In addition, MEFs from PKR-null miceshow increased replication of HSV-1 (Khabar et al., 2000).Therefore, the cellular resistance to VSV, influenza virus,and HSV-1 relies on a functional PKR pathway. Based onthese studies, we expected that deficiency in the PKR path-way may sensitize the host to viral infection because of theinability to mount an interferon protective response. How-ever, it was surprising that PKR deficiency in both murinetrans-dominant-negative PKR transgene and PKR knock-out genetic models provided protection of host tissue dam-age from viral-induced encephalitis.
Our results demonstrate that the PKR pathway may pro-vide an essential role in the inflammatory response to viralinfection in vivo. Although there are limitations from draw-ing conclusions from transgenic mice as well as from thePKR knock-out mice, we believe it is significant that bothmodels display similar histopathological evidence for re-duced inflammation in an encephalitis model based on po-liovirus infection. These results strongly implicate a role forPKR functional activity in encephalitis as a result of the hostresponse to poliovirus infection, although the clinical out-come was for the most part unchanged. The clinical out-come was probably expected for the acute, very rapidlyprogressive viral CNS infection studied here. In the enceph-alitis model used, the clinical symptoms produced by thehost response to infection are overshadowed by the detri-mental effect of rapid virus replication on the host CNS.Changes in the clinical outcome of infection due to repres-sion of the PKR response might be more obvious in slowlyprogressive viral encephalitides. Inflammatory host re-sponses (e.g., those mediated by PKR) play a more pro-nounced role in the progression of clinically overt diseasewhere the impact of viral-induced neuronal cell death andconsequent loss of critical neurological function is lessacute.
Presently, there are four known downstream effectorsthat may mediate PKR-induced proinflammatory responsesand apoptosis; eIF2�, NF�B, STAT1, and p53 (for review,see Kaufman 1999). Given the role of NF�B in the proin-
flammatory response (Rossi et al., 2000), it is tempting tosuggest that PKR mediates encephalitis in this modelthrough NF�B. PKR interaction with I�B kinase (IKK)subunit � signals activation of NF�B, although this re-sponse does not require the dsRNA binding activity or PKRkinase activity of PKR (Chu et al., 1999; Bonnet et al.,2000; Ishii et al., 2001). These results demonstrate theprotein kinase activity of PKR is not required for NF�Bsignaling. Since mice that express the trans-dominant-neg-ative kinase-mutant PKR displayed a defective inflamma-tory response to poliovirus infection, it would appear un-likely that the PKR-dependent inflammatory response toviral infection in this model is mediated through NF�B.Previous studies demonstrated that poliovirus infection ac-tivates PKR to induce eIF2� phosphorylation (Black et al.,1989). Future experiments should identify whether eIF2�phosphorylation or another PKR target is important to me-diate inflammation and disease symptoms.
To date, studies indicate that the status of PKR activationin the cell may have variable consequences that depend onthe virus species and the host. In some instances, PKRactivation will elicit early apoptosis, thereby limiting viralspread (Yeung et al., 1999; Bilgin et al., 2003). Viruses thathave mechanisms to counteract PKR activation may estab-lish persistent or latent infections that will eventually bedetrimental to the host (Gale and Katze, 1998). Alterna-tively, infection by some viruses, such as poliovirus studiedhere, may be aggravated by viral-induced PKR-mediatedencephalitis. Although our experimental model of viral en-cephalitis-induced inflammation supports a critical role forPKR, it is also possible that PKR may contribute to otherphysiological inflammatory responses. In this regard, it isinteresting to note that PKR-mediated eIF2� phosphoryla-tion is associated with neuronal degeneration in Alzhei-mer’s disease (Chang et al., 2002) and some aspects of thisdisease may reflect an abnormal inflammatory response.The ability to target and inactivate the PKR pathway mayprovide an attractive therapeutic strategy to interfere withharmful inflammatory responses to viral infections.
Materials and methods
Generation PKR(K296R) transgenic mice
The plasmid pBS-8.4 that contains wild-type humanPKR cDNA was kindly provided by Dr. B.R.G. Williams(The Cleveland Clinic, OH) (Meurs et al., 1990). An XbaI-SalI fragment subcloned for site-directed oligonucleotide-mediated mutagenesis was performed by the heteroduplexprocedure to introduce the K296R mutation. The transgeneconstruct plasmid was generated by cloning the HindIII-PstIPKR(K296R) cDNA into the SalI site of the human consti-tutive �-actin promoter (Bap) vector pHBAPr-1 (Gunninget al., 1987) (Fig. 1A).
The transgenic construct DNA (6.4-kb ClaI) was isolated
271D. Scheuner et al. / Virology 317 (2003) 263–274
from plasmid sequences and oocyte (C57Bl/6J � SJL/J)injections and implantations were performed in conjunctionwith the University of Massachusetts at Worcester Trans-genic Core Mouse Facility. Genomic tail DNA was pre-pared from over 100 progeny and analyzed by PCR forpresence of the transgene. Six independent founders wereidentified by PCR and confirmed by Southern blot analysis.Founders were crossed with C57BL/6J to yield F1, whichwas analyzed similarly (Fig. 1B, Table 1). Copy numberswere estimated assuming one gene copy of a 5- to 10-kbsequence per diploid murine genome is 1/106. Homozygoteswere selected by quantitative dot blot analysis of genomictail DNA using SV40 poly(A) and �-actin probes. Trans-genic mice used in these studies were derived from F1 3-45.
Expression of the transgene was evaluated in varioustissues of F1 mice from each of six founders using reversetranscription (Perkin–Elmer, random hexamer priming) andPCR amplification using a 5� primer from the 5�UTR of thehuman albumin promoter (bp 31–50) and a 3� primer fromthe hPKR cDNA (bp 454–477) (3� Primer) to generate a383-bp fragment. A portion of the human 5�UTR was notincluded in the transgene construct. PCR products wereanalyzed by agarose gel electrophoresis, Southern blot anal-ysis, and hybridization against an internal oligonucleo-tide spanning the initiation codon (gaaatggctg gtgatc). Den-sitometry scanning of film exposures was used to quantitateand normalize expression as summarized in Table 1. Aninternal control sample of mRNA isolated fromPKR(K296R)-transfected NIH3T3 cells was used to com-pare relative expression between founders (Fig. 1C). Inquantifying expression by densitometry, the band intensitieswere normalized to the internal control. Total RNA isolatedfrom heart tissue was used for competitive reverse transcrip-tion and PCR using primers that react with both human andmurine PKR mRNA but yield different sized products (Fig.1C, Table 1). When the amount of total RNA in the reactionis held constant (1 �g), the level of transgene transcriptrelative to the endogenous PKR mRNA determines whichtemplate is preferred for amplification. When the amplifi-cation from each template is equal, the ratio of nontrans-genic mRNA/transgenic mRNA indicates the fold overex-pression of the transgenic mRNA.
PKR knock-out mice
PKR knockout mice were kindly provided by Dr. B.R.G.Williams (The Cleveland Clinic, Cleveland, OH) (Yang etal., 1995). The mouse PKR gene is disrupted by a BalI-PstIrestriction fragment deletion removing 2 kb including partsof exons 2 and 3 and intron 2 (deleting 17 nucleotides ofnoncoding region and the coding sequence for the first 79amino acids). The deleted region is replaced by PGKNEOinsertion in the antisense orientation followed by untrans-lated mouse sequence (UMS).
Preparation and analysis of mouse embryo fibroblasts
Mouse embryonic fibroblasts were isolated at E14.5 us-ing standard procedures with culture in high-glucoseDMEM supplemented with 10% fetal calf serum (Hogan etal., 1994). PKR monoclonal antibody 71-10 was kindlyprovided by Dr. Ara Hovanessian. Western analysis of hearttissues was performed using a 1:250 dilution of monoclonalantibody with alkaline phosphatase labeled secondary fordetection. Anti-human PKR polyclonal antibody was kindlyprovided by Dr. Brian Williams (Cleveland Clinic, Cleve-land, OH). Western analysis of MEF extracts was per-formed using a 1:500 dilution of primary antibody followedby HRP-labeled secondary incubation and ECL detection.eIF2� phosphorylated at Ser51 was detected using affinity-purified antibody from Cell Signaling Tech. Inc. TotaleIF2� was detected as previously described (Srivastava etal., 1998).
The PKR-null mice were derived from a background ofC57Bl/6J � SV129ev. The dominant-negative PKR micewere derived from a background of C57Bl/6J � SJL. Bothlines were back-crossed one time into C57Bl/6J. Therefore,both the PKR-null and dominant-negative PKR mice have asubstantial contribution of C57Bl/6J in their background.Therefore, wild-type C57Bl/6J mice were used as controls.Since PV2 infection elicits a highly reproducible responsein inbred (C57Bl/6J; BALBc; Gromeier et al., unpublishedobservations) as well as outbred mice (ICR; Swiss–Web-ster; Gromeier et al., 1995), the differences in genetic back-ground in our experiments are unlikely to account for thedifferences observed. Homozygous PKR(K296R) trans-genic mice (from generation F3 of founder line 3-45),PKR�/� knock-out mice, and C57Bl/6J controls weretreated with intracerebral inoculation of 1 � 106 or 108 PFUPV2(MEF-1) (Figs. 2–4, 5, and 6, respectively). For thehistopathological analysis of CNS tissues, animals weresacrificed at a preterminal stage. A preterminal stage ofsevere neurological dysfunction was defined as paraplegiawith involvement of the upper extremities, and/or extrapy-ramidal signs and respiratory strain. Because of the slightlydelayed progression of clinical disease in trans-dominant-negative PKR mutant and PKR-null mice, these animalswere sacrificed for histopathological analysis 24–32 h laterthan their wild-type peers. At the time of sacrifice, allanimals displayed a preterminal stage of severe neurologicaldysfunction. Thus, the histological comparison of infectedPKR mutant or null mice and infected control mice reflectsthe brain pathology at a time when clinical symptoms aresevere and comparable. Neural tissues were processed forhistopathological analysis according to standard protocols.All tissues were embedded in paraffin, cut at a thickness of10 �m, and stained with luxol fast blue/PAS/hematoxylin.
272 D. Scheuner et al. / Virology 317 (2003) 263–274
Analysis of viral replication
To measure viral titers in neural tissues of mice virus-injected with PV2, brain and spinal cord samples weredissected and suspended in phosphate-buffered saline(PBS). Samples were then weighed and dounce homoge-nized. The tissue homogenate was subjected to a plaqueassay following standard procedures and the amount ofplaque forming units per milligram of tissue was deter-mined.
M.G. is a recipient of a Burroughs Wellcome FundCareer Award in the Biomedical Sciences. Portions of thiswork were supported by NIH Grant DK 42394 (R.J.K.).
Abraham, N., Stojdt, D.F., Duncan, P.I., Methot, N., Ishii, T., Dube, M.,Vanderhyden, B.C., Atkins, H.L., Gray, D.A., McBurney, M.W.,Koromilas, A.E., Brown, E.G., Sonenberg, N., Bell, J.C., 1999. Char-acterization of transgenic mice with targeted disruption of the catalyticdomain of the double-stranded RNA-dependent protein kinase, PKR.J. Biol. Chem. 274, 5953–5962.
Balachandran, S., Roberts, P.C., Brown, L.E., Truong, H., Pattnaik, A.K.,Archer, D.R., Barber, G.N., 2000. Essential role for the dsRNA-depen-dent protein kinase PKR in innate immunity to viral infection. Immu-nity 13, 129–141.
Bilgin, D.D., Liu, Y., Schiff, M., Dinesh-Kumar, S.P., 2003. P58IPK, aplant ortholog of double-stranded RNA-denendent protein kinase PKRinhibitor, functions in viral pathogenesis. Dev. Cell 4, 651–661.
Black, T.L., Safer, B., Hovanessian, A., Katze, M.G., 1989. The cellular68,000-Mr protein kinase is highly autophosphorylated and activatedyet significantly degraded during poliovirus infection: implications fortranslational regulation. J. Virol. 63, 2244–2251.
Bonnet, M.C., Weil, R., Dam, E., Hovanessian, A.G., Meurs, E.F., 2000.PKR stimulates NF-kappaB irrespective of its kinase function by in-teracting with the IkappaB kinase complex. Mol. Cell Biol. 20, 4532–4542.
Chang, C.C., Wong, A.K.Y., Ng, H-K., Hugon, J., 2002. Phosphoryation ofeukaryotic initiation factor-2� (eIF2�) is associated with neuronaldegeneration in Alzheimer’s disease. Clin. Neurosci. Neuropathol. 13,2429–2432.
Chong, K.L., Feng, L., Schappert, K., Meurs, E., Donahue, T.F., Friesen,J.D., Hovanessian, A.G., 1992. Human p68 kinase exhibits growthsuppression in yeast and homology to the translational regulator GCN2.EMBO J. 11, 1553–1562.
Chu, W.M., Ostertag, D., Li, Z.W., Chang, L., Chen, Y., Hu, Y., Williams,B., Perrault, J., Karin, M., 1999. JNK2 and IKKbeta are required foractivating the innate response to viral infection. Immunity 11, 721–731.
Cuddihy, A.R., Li, S., Tam, N.W., Wong, A.H., Taya, Y., Abraham, N.,Bell, J.C., Koromilas, A.E., 1999a. Double-stranded-RNA-activatedprotein kinase PKR enhances transcriptional activation by tumor sup-pressor p53. Mol. Cell Biol. 19, 2475–2484.
Cuddihy, A.R., Wong, A.H., Tam, N.W., Li, S., Koromilas, A.E., 1999b.The double-stranded RNA activated protein kinase PKR physicallyassociates with the tumor suppressor p53 protein and phosphorylateshuman p53 on serine 392 in vitro. Oncogene 18, 2690–2702.
Donehower, L.A., Harvey, M., Slagle, B.L., McArthur, M.J., MontgomeryJr., C.A., Butel, J.S., Bradley, A., 1992. Mice deficient for p53 are
developmentally normal but susceptible to spontaneous tumors. Nature356, 215–221.
Donze, O., Jagus, R., Koromilas, A.E., Hershey, J.W., Sonenberg, N.,1995. Abrogation of translation initiation factor eIF-2 phosphorylationcauses malignant transformation of NIH 3T3 cells. EMBO J. 14,3828–3834.
Gale Jr., M., Katze, M.G., 1998. Molecular mechanisms of interferonresistance mediated by viral-directed inhibition of PKR, the interferon-induced protein kinase. Pharmacol. Ther. 78, 29–46.
Gromeier, M., Alexander, L., Wimmer, E., 1996. Internal ribosomal entrysite substitution eliminates neurovirulence in intergeneric poliovirusrecombinants. Proc. Natl. Acad. Sci. USA 93, 2370–2375.
Gromeier, M., Lu, H.H., Wimmer, E., 1995. Mouse neuropathogenic po-liovirus strains cause damage in the central nervous system distinctfrom poliomyelitis. Microb. Pathog. 18, 253–267.
Gunning, P., Leavitt, J., Muscat, G., Ng, S.Y., Kedes, L., 1987. A humanbeta-actin expression vector system directs high-level accumulation ofantisense transcripts. Proc. Natl. Acad. Sci. USA 84, 4831–4835.
Harding, H.P., Zhang, Y., Bertolotti, A., Zeng, H., Ron, D., 2000. PERKis essential for translation regulation and cell survival during the un-folded protein response. Mol. Cell 5, 897–904.
Harding, H.P., Zhang, Y., Ron, D., 1999. Protein translation and foldingare coupled by an endoplasmic-reticulum-resident kinase. Nature 397,271–274.
Hogan, B., Beddington, R., Costantini, F., Lacy, E., 1994. Manipulating theMouse Embryo. Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY.
Ishii, T., Kwon, H., Hiscott, J., Mosialos, G., Koromilas, A.E., 2001.Activation of the I�B� kinase (IKK) complex by double-strandedRNA-binding defective and catalytic inactive mutants of the interferon-inducible protein kinase PKR. Oncogene 20, 1900–1912.
Kaufman, R.J., 1999. Double-stranded RNA-activated protein kinase me-diates virus-induced apoptosis: a new role for an old actor. Proc. Natl.Acad. Sci. USA 96, 11693–11695.
Kaufman, R.J., 2000. Double-stranded RNA-activated protein kinase PKR,in: Sonenberg, N., Hershey, J.W.B., Mathews, M.B. (Eds.), Transla-tional Control of Gene Expression, Cold Spring Harbor LaboratoryPress, Cold Spring Harbor, NY, pp. 503–528.
Khabar, K.S., Dhalla, M., Siddiqui, Y., Zhou, A., Al-Ahdal, M.N., Der,S.D., Silverman, R.H., Williams, B.R., 2000. Effect of deficiency of thedouble-stranded RNA-dependent protein kinase, PKR, on antiviral re-sistance in the presence or absence of ribonuclease L: HSV-1 replica-tion is particularly sensitive to deficiency of the major IFN-mediatedenzymes. J. Interferon Cytokine Res 20, 653–659.
Koromilas, A.E., Roy, S., Barber, G.N., Katze, M.G., Sonenberg, N., 1992.Malignant transformation by a mutant of the interferon-inducible dou-ble-stranded RNA dependent protein-kinase. Science 257, 1685–1689.
Kumar, A., Haque, J., Lacoste, J., Hiscott, J., Williams, B.R., 1994.Double-stranded RNA-dependent protein kinase activates transcriptionfactor NF-kappa B by phosphorylating I kappa B. Proc. Natl. Acad. Sci.USA 91, 6288–6292.
Kumar, A., Yang, Y.L., Flati, V., Der, S., Kadereit, S., Deb, A., Haque, J.,Reis, L., Weissmann, C., Williams, B.R., 1997. Deficient cytokinesignaling in mouse embryo fibroblasts with a targeted deletion in thePKR gene: role of IRF-1 and NF-kappaB. EMBO J. 16, 406–416.
Lengyel, P., 1987. Double-stranded RNA and interferon action. J. Inter-feron Res. 7, 511–519.
Levin, D., London, I.M., 1978. Regulation of protein synthesis: activationby double-stranded RNA of a protein kinase that phosphorylates eu-karyotic initiation factor 2. Proc. Natl. Acad. Sci. USA 75, 1121–1125.
Meurs, E., Chong, K., Galabru, J., Thomas, N.S.B., Kerr, I., Williams,B.R.G., Hovanessian, A.G., 1990. Molecular cloning and characteriza-tion of the human double stranded RNA-activated protein kinase in-duced by interferon. Cell 62, 379–390.
273D. Scheuner et al. / Virology 317 (2003) 263–274
Meurs, E.F., Galabru, J., Barber, G.N., Katze, M.G., Hovanessian, A.G.,1993. Tumor suppressor function of the interferon-induced double-stranded RNA-activated protein kinase. Proc. Natl. Acad. Sci. USA 90,232–236.
Okayama, H., Berg, P., 1983. A cDNA cloning vector that permits expres-sion of cDNA inserts in mammalian cells. Mol. Cell Biol. 3, 280–289.
Ramana, C.V., Grammatikakis, N., Chernov, M., Nguyen, H., Goh, K.C.,Williams, B.R., Stark, G.R., 2000. Regulation of c-myc expression byIFN-gamma through Stat1-dependent and -independent pathways.EMBO J. 19, 263–272.
Rossi, A., Kapahi, P., Natoli, G., Takahashi, T., Chen, Y., Karin, M.,Santoro, M.G., 2000. Anti-inflammatory cyclopentenone prostaglan-dins are direct inhibitors of IkappaB kinase. Nature 403, 103–108.
Samuel, C.E., 1991. Antiviral actions of interferon: interferon regulatedcellular proteins and their surprisingly selective antiviral activities.Virology 183, 1–11.
Schlesinger, R.W., Morgan, J.M., Olitsky, P.K., 1943. Transmission torodents of Lansing type poliomyelitis virus originating in the MiddleEast. Science 98, 452–454.
Shen, Y., Shenk, T.E., 1995. Viruses and apoptosis. Curr. Opin. Genet.Dev. 5, 105–111.
Shi, Y., Vattem, K.M., Sood, R., An, J., Liang, J., Stramm, L., Wek, R.C.,1998. Identification and characterization of pancreatic eukaryotic ini-tiation factor 2 alpha-subunit kinase, PEK, involved in translationalcontrol. Mol. Cell Biol. 18, 7499–7509.
Sood, R., Porter, A.C., Olsen, D.A., Cavener, D.R., Wek, R.C., 2000. Amammalian homologue of GCN2 protein kinase important for transla-tional control by phosphorylation of eukaryotic initiation factor-2alpha.Genetics 154, 787–801.
Srivastava, S.P., Kumar, K.U., Kaufman, R.J., 1998. Phosphorylation ofeukaryotic translation initiation factor 2 mediates apoptosis in responseto activation of the double-stranded RNA-dependent protein kinase.J. Biol. Chem. 273, 2416–2423.
Stojdl, D.F., Abraham, N., Knowles, S., Marius, R., Brasey, A., Lichty,B.D., Brown, E.G., Sonenberg, N., Bell, J.C., 2000. The murine dou-ble-stranded RNA-dependent protein kinase PKR is required for resis-tance to vesicular stomatitis virus. J. Virol. 74, 9580–9585.
Yang, Y.-L., Reis, L.F., Pavlovic, J., Aguzzi, A., Schäfer, R., Kumar, A.,Williams, B.R.G., Aguet, M., Weissmann, C., 1995. Deficient signalingin mice devoid of double-stranded RNA-dependent protein kinase.EMBO J. 14, 6095–6106.
YeungM.ChangD., L., R. Camanitigue, E., Lau, A.S., 1999. Inhibitory roleof the host apoptogenic gene PKR in the establishment of persistentinfection by encephalomyocarditis virus in U937 cells. Proc. Natl.Acad. Sci. USA 96;11860–11865.
Zhang, P., McGrath, B.C., Reinert, J., Olsen, D.S., Lei, L., Gill, S., Wek,S.A., Vattem, K.M., Wek, R.C., Kimball, S.R., Jefferson, L.S.,Cavener, D.R., 2002. The GCN2 eIF2alpha kinase is required foradaptation to amino acid deprivation in mice. Mol. Cell Biol. 22,6681–6688.
274 D. Scheuner et al. / Virology 317 (2003) 263–274
The double-stranded RNA-activated protein kinase mediates viral-induced encephalitisIntroductionResultsDiscussionMaterials and methodsGeneration PKR(K296R) transgenic micePKR knock-out micePreparation and analysis of mouse embryo fibroblastsPoliovirus infectionAnalysis of viral replication