- Email: [email protected]
Detection of viral antigens in the encephalopathy brain by influenza A virus
International Congress Series 1219 (2001) 615 – 622
Detection of viral antigens in the encephalopathy brain by inf luenza A virus Mitsuo Takahashia,*, Tatsuo Yamadab, Tetsuya Toyodac a
Department of Clinical Pharmacology, Faculty of Pharmaceutical Sciences, Fukuoka University, 1-19-8 Nanakuma, Jonan, Fukuoka 814-0180, Japan b Fifth Department of Internal Medicine, School of Medicine, Fukuoka University, 1-45-7 Nanakuma, Jonan, Fukuoka 814-0180, Japan c Department of Virology, School of Medicine, Kurume University, Kurume, Fukuoka, Japan
Abstract Rapid progressive encephalopathy showing a high fever, consciousness loss and recurrent convulsions has been occasionally reported in childhood during influenza epidemics in Japan since 1995. A clinicopathological study of a 2-year-old female diagnosed with hemorrhagic shock and encephalopathy syndrome associated with influenza A virus (A/Nagasaki/76/98; H3N2) infection showed that the virus antigens are present in CD8-positive T lymphocytes from the lung and spleen, suggesting the possible route of the virus spread. The virus antigens are confined to a very limited part of the brain, especially Purkinje cells in the cerebellum and many neurons in the pons, without inducing an overt immunological reaction of the host. RT-PCR for detecting the hemagglutinin gene demonstrated definite positive bands in all frozen tissues and cerebrospinal fluids taken at autopsy but not in samples on admission. The pathological change induced by the direct viral invasion cannot be sufficient for the rapid and severe clinical course of the disease within 24 – 48 h after the initial respiratory symptoms. The rapid production of several inflammatory cytokines together with the breakdown of the blood – brain barrier, which will induce severe brain edema, may be the major pathological processes. Any therapeutic strategy to control this multi-step progression could be effective. D 2001 Elsevier Science B.V. All rights reserved. Keywords: Influenza A virus; Encephalopathy; Autopsy; RT-PCR; Immunohistochemistry
*
Corresponding author. Tel.: +81-92-871-6631x6645; fax: +81-92-863-0389. E-mail address: [email protected] (M. Takahashi).
0531-5131/01/$ – see front matter D 2001 Elsevier Science B.V. All rights reserved. PII: S 0 5 3 1 - 5 1 3 1 ( 0 1 ) 0 0 3 7 4 - 0
616
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
1. Introduction Influenza virus is one of the most common causes of respiratory tract infection during the winter season. A wide spectrum of central nervous system (CNS) involvement has been observed during influenza A virus infection [1]. Rapid progressive encephalopathy cases with a high fever, consciousness disturbances and recurrent convulsions have been occasionally reported in children during influenza epidemics in Japan since 1995. Isolation of viruses or viral genomes from throat swabs or other clinical specimens strongly suggests that the encephalitic condition could be associated with influenza virus infection. After reviewing the severe or fatal cases reported previously, we conclude that the clinical signs and symptoms following influenza infection are superimposed by three suggested entities: Rye’s syndrome [2], acute necrotizing encephalopathy [3], and hemorrhagic shock and encephalopathy (HSE) syndrome [4]. In this paper, we examined a 2-year-old girl with HSE syndrome, trying to answer several questions for which no histological or virological data exist.
2. Patient and methods 2.1. Case report A previously healthy 2.5-year-old girl, who had no history of immunization for influenza virus, became pyrexial at 38 – 39 C and had a cough 1 day before admission despite having been treated with a cold remedy. She managed to speak a few words and felt uneasy. She looked uncomfortable and sometimes trembled. She could hardly sleep during the following night owing to a continuous feeling of discomfort. She could stand and walk at night. On the following morning, when she asked her mother to take her to the lavatory, her mother noticed that her locomotion was quite unstable and that she could not walk straight. She fell down after voiding and stopped breathing. Her heart beat recovered 30 min after cardiopulmonary resuscitation and she was then transferred to Sasebo Municipal General Hospital. On admission, she was comatose without spontaneous respiration. As shown in Table 1, her clinical course was characterized by coma, DIC, subsequent hemorrhagic diathesis, shock and severe rhabdomyolysis. She died 7 days after admission. 2.2. Nested RT-PCR and Southern blot analysis One microgram of total RNA was reverse-transcribed using H3 specific primer H3F-7 (ACTATCATTGCTTTGAGC) and avian myeloblastosis virus reverse transcriptase. PCR of the H3 specific HA gene was carried out for 30 cycles with the primer pair of H3F-7 (ACTATCATTGCTTTGAGC) and H3R-1184 (ATGGCTGCTTGAGTGCTT); the expected product size was 1178 base pairs (bp). PCR cycling conditions included denaturation at 95 C for 5 min, annealing for 2 min at 42 C, extension for 3 min at 72 C, and additional extension for 10 min at 72 C. One microliter of the RT-PCR
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
617
Table 1 Variable
On admission
Third day
Sixth day
Body temperature (C) Blood pressure (mm Hg) Urine volume (ml) Platelet count/mm3 Prothrombin time (s) Aspartate aminotransferase (U/l) Alanine aminotransferase (U/l) Lactate dehydrogenase (U/l) Creatine kinase (U/l) Ammonia (ml/dl)
30.5 65/40 640 297,000 20.7 1019 243 2053 2127 38
31.5 100/80 1110 33,000 65.4 1988 552 17,630 197,850 22
30.5 60/50 190 16,000 97.0 645 387 7466 9066
products was further PCR-amplified for 30 cycles with the primer pairs of H3-HA50n (GCACACTGATAGATGCTCTATTGGG) and H3-HA30n (GGTGCATCTGACCTCATTATTGAG) using a Ready-to-Go PCR kit (Pharmacia Biotech); the expected size of the nested PCR product was 629 bp. The conditions of the nested PCR included denaturation at 94 C for 1 min, annealing for 2 min at 55 C, and extension for 1 min at 72 C. As a control, PCR was performed exactly as above but without reverse-transcription. RT-PCR of b-actin was also performed with the primer pairs of b-actin 50 (ATCATGTTTGAGACCTTCAA) and b-actin 30 (CATCTCTTGCTCGAAGTCCA). RT-PCR cycling included reverse transcription at 42 C for 30 min, followed by denaturation at 95 C for 1 min, annealing for 1 min at 44 C, and extension for 1 min at 72 C using a Ready-toGo RT-PCR kit (Pharmacia); the expected product size was 308 bp. Digoxygenin-11dUTP (Boehringer Mannheim Biochemica)-labeled H3 HA of influenza virus A/Nagasaki/ 93/98 (H3N2) was used as a Southern hybridization probe. 2.3. Viral inoculation of cynomolgus monkey brain Each monkey received an inoculation into the bilateral frontal lobes of 1.0 ml PBS containing 25 plaque forming units of A/Nagasaki/76/98 (H3N2) with a 22-gauge needle. On days 3 and 7 after inoculation, monkeys were anesthetized and were killed by bilateral common carotid perfusion with 4% PFA and brains were removed. All experimental procedures were approved by the ethical committee for animal experiments at National Institute of Infectious Disease, Tokyo, Japan. 2.4. Immunohistochemistry Sections were pretreated with 0.2% Triton X-100 (Sigma) in TBS (50 mmol/l Tris – HCl; 150 mmol/l NaCl; pH 7.5) for 20 min to increase permeability to primary antibodies. Peroxidase-like activity in tissue samples was blocked by incubating with 3% hydrogen peroxide in TBS for 10 min. They were then incubated with a primary antibody for 48 h at 4 C, with 1:1000 biotinylated horse anti-mouse IgG (Vector) preabsorbed by 10% normal human or monkey serum for 2 h, and with 1:1000 avidin-
618
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
Table 2 Primary antibodies Specificity
Species and dilution
Source
Human glial fibrillary acidic protein Microtubule associated protein-2 A/Kumamoto/22/76 (H3N2) A/Memphis/96 (H3N2) CD3, CD4, CD8, CD20 Human insulin Human glucagon Human somatostatin
Mouse monoclonal, 1:1000 Mouse monoclonal, 1:1000 Chicken polyclonal, 1:2000 Chicken polyclonal, 1:2000 Mouse monoclonal, 1:1000 Guinea pig polyclonal, 1:100 Rabbit polyclonal, 1:100 Rabbit polyclonal, 1:100
DAKO Boehringer Kida Kida DAKO Lipshaw Lipshaw Lipshaw
biotinylated horse-radish peroxidase complex (Vector) for 2 h at room temperature. Color development of peroxidase-labeled areas was performed for 15 min in a freshly prepared solution consisting of 20 ml TBS, 1 ml of 10 mg/ml 3,30-diaminobenzidine (DAB; Dojindo), 3 ml 1% nickel ammonium sulfate and 3.5 ml hydrogen peroxide (30% stock solution). For double immunostaining, the second cycle was carried out in a manner similar to the first after treating sections for 30 min with 0.5% H2O2 solution, except that the nickel ammonium sulfate was omitted from the DAB solution, yielding a brown precipitate. The primary antibodies tried in this study and the dilutions used are shown in Table 2. Several negative controls were simultaneously performed: normal sera corresponding to the applied primary antibody as a primary antibody, intrinsic peroxidase detection by DAB solution without any further process, and the procedure without a primary antibody.
3. Results 3.1. Viral isolation and detection of the HA gene by nested RT-PCR Influenza A virus was isolated in MDCK cells from a throat swab sample taken on admission, and its serotype was identified as influenza A virus (H3N2). Throat swab and cerebrospinal fluid samples were tested for antibody titer to anti-A/Wuhan/ 359/95 (H3N2). The hemagglutinin inhibition (HI) titer of the throat swab sample was 1280. In nested RT-PCR of HA, positive bands of corresponding sizes, 629 bp for the HA gene segment and 308 bp for b-actin, were obtained in all samples taken at autopsy, and not in CSF and PBMC taken on admission. Specificity of the nested PCR products was confirmed in all samples by Southern blot analysis (Fig. 1). 3.2. Immunohistochemical study of the autopsied samples Immunohistochemistry of the lung with anti-A/Kumamoto/22/76 (H3N2) showed linear dense staining, probably corresponding to type I alveolar epithelial cells, as well
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
619
Fig. 1. RNAs from tissues and CSF taken at autopsy were subjected to RT-PCR. The outer and inner pairs of PCR primers were designed to detect the HA gene segment. Positive bands can be seen in all tissues and CSF (A), and the specificity of the amplified PCR products was confirmed by Southern blot (B). The band in lane V is from a positive control using a template of HA cDNA transcribed from RNA of influenza virus A/Nagasaki/76/98 isolated from this case. M: marker; C: CSF; F: frontal lobe; T: temporal lobe; H: hippocampus; Lu: lung; K: kidney; Cr: cerebellum; Li: liver.
as positively stained round cells in alveoli and vessels. Double immunostaining with antiA/Kumamoto and anti-CD3 showed that there were some double positive cells. In the pancreas, although there was relatively high background staining, islets of Langerhans were positively stained compared to negative controls. In immunohistochemistry of the spleen with anti-A/Memphis/96 (H3N2) and some lymphocyte markers, anti-A/Memphis/96-positive cells are located in the marginal area of the white pulp, where T cells are ordinarily abundant. In the center of the white pulp, there is no germinal center formation. Residual round, mononuclear cells are stained with antiA/Memphis/96 and anti-CD4. Doubly stained cells are hardly observed. In contrast, the majority of the virus antigen-positive cells are also positive with anti-CD8. No fatty degeneration or virus antigen was observed in liver tissues. The brain hemisphere was softened and edematous after 1 week of mechanical respiratory assist. There was no obvious hemorrhage. We could not obtain any parts of the thalamus and basal ganglia due to massive morphological destruction at autopsy. Immunohistochemical studies were shown in Fig. 2. No pathological alterations were seen in the vessels and perivascular spaces. 3.3. Immunohistochemical study of the infected cynomolgus monkey brain Immunoreactivity with anti-A/Memphis/96 was seen only in the vicinity of injection site of the virus. Adjacent neurons in the pyramidal layer were focally virus antigen
620
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
Fig. 2. Immunohistochemistry of the brain tissues. Antigenicity in the fixed brain tissues was checked by staining with anti-MAP-2 (A) and with anti-GFAP (B). Fairly good preservation of the antigenicity was confirmed in each area of the brain examined. Astrocytes seen in the vicinity of blood vessels had the morphology of activated ones (B). However, clear astrocytosis or diffusely activated astrocytes were not observed in all examined areas. (C) Many Purkinje cells in the cerebellum were strongly stained with anti-A/Memphis/96 (H3N2) (inset). Unstained Purkinje cells were sometimes seen even in areas very close to the positively stained cells, reflecting focal invasion of the virus. (D) Various neurons in the pons were positive for virus antigen.
positive. Astrocytic activation and proliferation were seen in the injection site and were positive with anti-A/Memphis/96.
4. Discussion In the autopsied brain, Purkinje cells in the cerebellum and various neurons in the pons were positive for influenza virus antigen. There was no positive staining in the hippocampus, midbrain, and neocortex. When we consider this parenchymal tissue involvement, the characteristic clinical course of influenza virus-induced encephalopathy is very important. The syndrome is diagnosed clinically by a sudden onset of fever followed by disturbances of consciousness ranging from somnolence to deep coma and convulsions to some degree within a very short period, mostly 24 –48 h after the onset of respiratory symptoms during an influenza epidemic [5]. RT-PCR demonstrated that the influenza A virus genome RNA was present in brain tissues as well as other organs and CSF. In contrast, CSF had little or no characteristics corresponding to meningoencephalitic conditions (pleocytosis and protein level elevation) in most reported cases. Immunohis-
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
621
tochemical study of the infected cynomolgus monkey brain showed that over 1 week, the infection was restricted to the injection site. The time between the onset of respiratory tract symptoms and CNS involvement is not long enough for the virus to replicate and spread to parts of the brain, which we previously observed in mice showing neurological symptoms due to encephalitis. Therefore, we can assume that the syndrome may occur mainly not by the direct invasion of the brain parenchymal tissues by the virus. Reports measuring proinflammatory cytokine levels in CSF from children with acute encephalitis showed that soluble TNF receptor 1 (sTNF-R1) and IL-6 levels were significantly higher in those who died or were left with sequelae than in those who survived without neurological deficits. The level of sTNF-R1 during the acute stage of encephalitis is an important index for predicting the neurological outcome. The serum level of IL-6 is also higher in patients with acute influenza virus-induced encephalopathy [6]. Furthermore, hyperactivated coagulation factors associated with DIC would participate in the pathogenesis [7,8]. Some residual mononuclear blood cells in the splenic vessels were double-positive for anti-virus and CD8 corresponding to suppressor T lymphocytes [9,10]. The barrier between the blood and susceptible extra-respiratory tissues is very strong under physiological condition because of the integrity and the insusceptibility of the endothelial cells lining blood vessels. In this study, we observed vascular endothelial cells under light microscopy and found no clear pathological alterations. Future studies are needed to look for ultrastructural changes of the endothelial cells as well as for the measurement of the humoral factors capable of acting on these cells [11].
Acknowledgements Drs. A. Matsuo and K. Iwasaki at Sasebo Municipal General Hospital kindly provided us a clinical record and autopsied samples. We thank Drs. K. Ito, K. Yamada and K. Yamada at Kurume University for generous gifts of the antibodies shown in Table 2. We also thank Drs. H. Noguchi and T. Ueda at Nagasaki Prefectural Institute on Public Health and Environmental Science for virus isolation. We are grateful to Dr. K. Ami at National Institute of Infectious Disease for animal experiments.
References [1] Y. Hayase, K. Tobita, Influenza virus and neurological disease, Psychiatry Clin. Neurosci. 51 (1997) 181 – 184. [2] R.D.K. Rye, G. Morgan, J. Baral, Encephalopathy and fatty degeneration of the viscera-a disease entity in childhood, Lancet II (1963) 749 – 752. [3] M. Mizuguchi, J. Abe, K. Mikkaichi, S. Noma, K. Yoshida, T. Yamanaka, S. Kamoshita, Acute necrotising encephalopathy of childhood: a new syndrome presenting with multifocal, symmetric brain lesions, J. Neurol., Neurosurg. Psychiatry 58 (1995) 555 – 561. [4] M. Levin, M. Hjelm, J.D.S. Kay, J.R. Pincott, J.D. Gould, R. Dinwiddie, D.J. Matthew, Haemorrhagic shock and encephalopathy: a new syndrome with a high mortality in young children, Lancet II (1983) 64 – 67. [5] S. Kimura, N. Ohtuki, A. Nezu, M. Tanaka, S. Takeshita, Clinical and radiological variability of influenzarelated encephalopathy or encephalitis, Acta Paediatr. Jpn. 40 (1998) 264 – 270.
622
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
[6] H. Tsuruoka, H. Xu, K. Kuroda, K. Hayashi, O. Yasui, A. Yamada, T. Ishizaki, Y. Yamada, T. Watanabe, Y. Hosaka, Detection of influenza virus RNA in peripheral blood mononuclear cells of influenza patients, Jpn. J. Med. Sci. Biol. 50 (1997) 27 – 34. [7] Y. Nagai, Protease-dependent virus tropism and pathogenicity, Trends Microbiol. 1 (1993) 81 – 87. [8] A. Nishino, M. Suzuki, T. Yoshimoto, H. Otani, H. Nagura, A novel aspect of thrombin in the tissue reaction following central nervous system injury, Acta Neurochir. (Suppl.) 60 (1994) 86 – 88. [9] I. Mori, T. Komatsu, K. Takeuchi, K. Nakakuki, M. Sudo, Y. Kimura, Viremia induced by influenza virus, Microb. Pathog. 19 (1995) 237 – 244. [10] P.G. Stevenson, S. Hawke, C.R.M. Bangham, Protection against lethal influenza virus encephalitis by intranasally primed CD8 + memory T cells, J. Immunol. 157 (1996) 3065 – 3073. [11] I. Joris, G. Majno, E.J. Corey, R.A. Lewis, The mechanism of vascular leakage induced by leukotriene E4, Am. J. Pathol. 126 (1987) 19 – 24.
Detection of viral antigens in the encephalopathy brain by inf luenza A virus Mitsuo Takahashia,*, Tatsuo Yamadab, Tetsuya Toyodac a
Department of Clinical Pharmacology, Faculty of Pharmaceutical Sciences, Fukuoka University, 1-19-8 Nanakuma, Jonan, Fukuoka 814-0180, Japan b Fifth Department of Internal Medicine, School of Medicine, Fukuoka University, 1-45-7 Nanakuma, Jonan, Fukuoka 814-0180, Japan c Department of Virology, School of Medicine, Kurume University, Kurume, Fukuoka, Japan
Abstract Rapid progressive encephalopathy showing a high fever, consciousness loss and recurrent convulsions has been occasionally reported in childhood during influenza epidemics in Japan since 1995. A clinicopathological study of a 2-year-old female diagnosed with hemorrhagic shock and encephalopathy syndrome associated with influenza A virus (A/Nagasaki/76/98; H3N2) infection showed that the virus antigens are present in CD8-positive T lymphocytes from the lung and spleen, suggesting the possible route of the virus spread. The virus antigens are confined to a very limited part of the brain, especially Purkinje cells in the cerebellum and many neurons in the pons, without inducing an overt immunological reaction of the host. RT-PCR for detecting the hemagglutinin gene demonstrated definite positive bands in all frozen tissues and cerebrospinal fluids taken at autopsy but not in samples on admission. The pathological change induced by the direct viral invasion cannot be sufficient for the rapid and severe clinical course of the disease within 24 – 48 h after the initial respiratory symptoms. The rapid production of several inflammatory cytokines together with the breakdown of the blood – brain barrier, which will induce severe brain edema, may be the major pathological processes. Any therapeutic strategy to control this multi-step progression could be effective. D 2001 Elsevier Science B.V. All rights reserved. Keywords: Influenza A virus; Encephalopathy; Autopsy; RT-PCR; Immunohistochemistry
*
Corresponding author. Tel.: +81-92-871-6631x6645; fax: +81-92-863-0389. E-mail address: [email protected] (M. Takahashi).
0531-5131/01/$ – see front matter D 2001 Elsevier Science B.V. All rights reserved. PII: S 0 5 3 1 - 5 1 3 1 ( 0 1 ) 0 0 3 7 4 - 0
616
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
1. Introduction Influenza virus is one of the most common causes of respiratory tract infection during the winter season. A wide spectrum of central nervous system (CNS) involvement has been observed during influenza A virus infection [1]. Rapid progressive encephalopathy cases with a high fever, consciousness disturbances and recurrent convulsions have been occasionally reported in children during influenza epidemics in Japan since 1995. Isolation of viruses or viral genomes from throat swabs or other clinical specimens strongly suggests that the encephalitic condition could be associated with influenza virus infection. After reviewing the severe or fatal cases reported previously, we conclude that the clinical signs and symptoms following influenza infection are superimposed by three suggested entities: Rye’s syndrome [2], acute necrotizing encephalopathy [3], and hemorrhagic shock and encephalopathy (HSE) syndrome [4]. In this paper, we examined a 2-year-old girl with HSE syndrome, trying to answer several questions for which no histological or virological data exist.
2. Patient and methods 2.1. Case report A previously healthy 2.5-year-old girl, who had no history of immunization for influenza virus, became pyrexial at 38 – 39 C and had a cough 1 day before admission despite having been treated with a cold remedy. She managed to speak a few words and felt uneasy. She looked uncomfortable and sometimes trembled. She could hardly sleep during the following night owing to a continuous feeling of discomfort. She could stand and walk at night. On the following morning, when she asked her mother to take her to the lavatory, her mother noticed that her locomotion was quite unstable and that she could not walk straight. She fell down after voiding and stopped breathing. Her heart beat recovered 30 min after cardiopulmonary resuscitation and she was then transferred to Sasebo Municipal General Hospital. On admission, she was comatose without spontaneous respiration. As shown in Table 1, her clinical course was characterized by coma, DIC, subsequent hemorrhagic diathesis, shock and severe rhabdomyolysis. She died 7 days after admission. 2.2. Nested RT-PCR and Southern blot analysis One microgram of total RNA was reverse-transcribed using H3 specific primer H3F-7 (ACTATCATTGCTTTGAGC) and avian myeloblastosis virus reverse transcriptase. PCR of the H3 specific HA gene was carried out for 30 cycles with the primer pair of H3F-7 (ACTATCATTGCTTTGAGC) and H3R-1184 (ATGGCTGCTTGAGTGCTT); the expected product size was 1178 base pairs (bp). PCR cycling conditions included denaturation at 95 C for 5 min, annealing for 2 min at 42 C, extension for 3 min at 72 C, and additional extension for 10 min at 72 C. One microliter of the RT-PCR
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
617
Table 1 Variable
On admission
Third day
Sixth day
Body temperature (C) Blood pressure (mm Hg) Urine volume (ml) Platelet count/mm3 Prothrombin time (s) Aspartate aminotransferase (U/l) Alanine aminotransferase (U/l) Lactate dehydrogenase (U/l) Creatine kinase (U/l) Ammonia (ml/dl)
30.5 65/40 640 297,000 20.7 1019 243 2053 2127 38
31.5 100/80 1110 33,000 65.4 1988 552 17,630 197,850 22
30.5 60/50 190 16,000 97.0 645 387 7466 9066
products was further PCR-amplified for 30 cycles with the primer pairs of H3-HA50n (GCACACTGATAGATGCTCTATTGGG) and H3-HA30n (GGTGCATCTGACCTCATTATTGAG) using a Ready-to-Go PCR kit (Pharmacia Biotech); the expected size of the nested PCR product was 629 bp. The conditions of the nested PCR included denaturation at 94 C for 1 min, annealing for 2 min at 55 C, and extension for 1 min at 72 C. As a control, PCR was performed exactly as above but without reverse-transcription. RT-PCR of b-actin was also performed with the primer pairs of b-actin 50 (ATCATGTTTGAGACCTTCAA) and b-actin 30 (CATCTCTTGCTCGAAGTCCA). RT-PCR cycling included reverse transcription at 42 C for 30 min, followed by denaturation at 95 C for 1 min, annealing for 1 min at 44 C, and extension for 1 min at 72 C using a Ready-toGo RT-PCR kit (Pharmacia); the expected product size was 308 bp. Digoxygenin-11dUTP (Boehringer Mannheim Biochemica)-labeled H3 HA of influenza virus A/Nagasaki/ 93/98 (H3N2) was used as a Southern hybridization probe. 2.3. Viral inoculation of cynomolgus monkey brain Each monkey received an inoculation into the bilateral frontal lobes of 1.0 ml PBS containing 25 plaque forming units of A/Nagasaki/76/98 (H3N2) with a 22-gauge needle. On days 3 and 7 after inoculation, monkeys were anesthetized and were killed by bilateral common carotid perfusion with 4% PFA and brains were removed. All experimental procedures were approved by the ethical committee for animal experiments at National Institute of Infectious Disease, Tokyo, Japan. 2.4. Immunohistochemistry Sections were pretreated with 0.2% Triton X-100 (Sigma) in TBS (50 mmol/l Tris – HCl; 150 mmol/l NaCl; pH 7.5) for 20 min to increase permeability to primary antibodies. Peroxidase-like activity in tissue samples was blocked by incubating with 3% hydrogen peroxide in TBS for 10 min. They were then incubated with a primary antibody for 48 h at 4 C, with 1:1000 biotinylated horse anti-mouse IgG (Vector) preabsorbed by 10% normal human or monkey serum for 2 h, and with 1:1000 avidin-
618
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
Table 2 Primary antibodies Specificity
Species and dilution
Source
Human glial fibrillary acidic protein Microtubule associated protein-2 A/Kumamoto/22/76 (H3N2) A/Memphis/96 (H3N2) CD3, CD4, CD8, CD20 Human insulin Human glucagon Human somatostatin
Mouse monoclonal, 1:1000 Mouse monoclonal, 1:1000 Chicken polyclonal, 1:2000 Chicken polyclonal, 1:2000 Mouse monoclonal, 1:1000 Guinea pig polyclonal, 1:100 Rabbit polyclonal, 1:100 Rabbit polyclonal, 1:100
DAKO Boehringer Kida Kida DAKO Lipshaw Lipshaw Lipshaw
biotinylated horse-radish peroxidase complex (Vector) for 2 h at room temperature. Color development of peroxidase-labeled areas was performed for 15 min in a freshly prepared solution consisting of 20 ml TBS, 1 ml of 10 mg/ml 3,30-diaminobenzidine (DAB; Dojindo), 3 ml 1% nickel ammonium sulfate and 3.5 ml hydrogen peroxide (30% stock solution). For double immunostaining, the second cycle was carried out in a manner similar to the first after treating sections for 30 min with 0.5% H2O2 solution, except that the nickel ammonium sulfate was omitted from the DAB solution, yielding a brown precipitate. The primary antibodies tried in this study and the dilutions used are shown in Table 2. Several negative controls were simultaneously performed: normal sera corresponding to the applied primary antibody as a primary antibody, intrinsic peroxidase detection by DAB solution without any further process, and the procedure without a primary antibody.
3. Results 3.1. Viral isolation and detection of the HA gene by nested RT-PCR Influenza A virus was isolated in MDCK cells from a throat swab sample taken on admission, and its serotype was identified as influenza A virus (H3N2). Throat swab and cerebrospinal fluid samples were tested for antibody titer to anti-A/Wuhan/ 359/95 (H3N2). The hemagglutinin inhibition (HI) titer of the throat swab sample was 1280. In nested RT-PCR of HA, positive bands of corresponding sizes, 629 bp for the HA gene segment and 308 bp for b-actin, were obtained in all samples taken at autopsy, and not in CSF and PBMC taken on admission. Specificity of the nested PCR products was confirmed in all samples by Southern blot analysis (Fig. 1). 3.2. Immunohistochemical study of the autopsied samples Immunohistochemistry of the lung with anti-A/Kumamoto/22/76 (H3N2) showed linear dense staining, probably corresponding to type I alveolar epithelial cells, as well
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
619
Fig. 1. RNAs from tissues and CSF taken at autopsy were subjected to RT-PCR. The outer and inner pairs of PCR primers were designed to detect the HA gene segment. Positive bands can be seen in all tissues and CSF (A), and the specificity of the amplified PCR products was confirmed by Southern blot (B). The band in lane V is from a positive control using a template of HA cDNA transcribed from RNA of influenza virus A/Nagasaki/76/98 isolated from this case. M: marker; C: CSF; F: frontal lobe; T: temporal lobe; H: hippocampus; Lu: lung; K: kidney; Cr: cerebellum; Li: liver.
as positively stained round cells in alveoli and vessels. Double immunostaining with antiA/Kumamoto and anti-CD3 showed that there were some double positive cells. In the pancreas, although there was relatively high background staining, islets of Langerhans were positively stained compared to negative controls. In immunohistochemistry of the spleen with anti-A/Memphis/96 (H3N2) and some lymphocyte markers, anti-A/Memphis/96-positive cells are located in the marginal area of the white pulp, where T cells are ordinarily abundant. In the center of the white pulp, there is no germinal center formation. Residual round, mononuclear cells are stained with antiA/Memphis/96 and anti-CD4. Doubly stained cells are hardly observed. In contrast, the majority of the virus antigen-positive cells are also positive with anti-CD8. No fatty degeneration or virus antigen was observed in liver tissues. The brain hemisphere was softened and edematous after 1 week of mechanical respiratory assist. There was no obvious hemorrhage. We could not obtain any parts of the thalamus and basal ganglia due to massive morphological destruction at autopsy. Immunohistochemical studies were shown in Fig. 2. No pathological alterations were seen in the vessels and perivascular spaces. 3.3. Immunohistochemical study of the infected cynomolgus monkey brain Immunoreactivity with anti-A/Memphis/96 was seen only in the vicinity of injection site of the virus. Adjacent neurons in the pyramidal layer were focally virus antigen
620
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
Fig. 2. Immunohistochemistry of the brain tissues. Antigenicity in the fixed brain tissues was checked by staining with anti-MAP-2 (A) and with anti-GFAP (B). Fairly good preservation of the antigenicity was confirmed in each area of the brain examined. Astrocytes seen in the vicinity of blood vessels had the morphology of activated ones (B). However, clear astrocytosis or diffusely activated astrocytes were not observed in all examined areas. (C) Many Purkinje cells in the cerebellum were strongly stained with anti-A/Memphis/96 (H3N2) (inset). Unstained Purkinje cells were sometimes seen even in areas very close to the positively stained cells, reflecting focal invasion of the virus. (D) Various neurons in the pons were positive for virus antigen.
positive. Astrocytic activation and proliferation were seen in the injection site and were positive with anti-A/Memphis/96.
4. Discussion In the autopsied brain, Purkinje cells in the cerebellum and various neurons in the pons were positive for influenza virus antigen. There was no positive staining in the hippocampus, midbrain, and neocortex. When we consider this parenchymal tissue involvement, the characteristic clinical course of influenza virus-induced encephalopathy is very important. The syndrome is diagnosed clinically by a sudden onset of fever followed by disturbances of consciousness ranging from somnolence to deep coma and convulsions to some degree within a very short period, mostly 24 –48 h after the onset of respiratory symptoms during an influenza epidemic [5]. RT-PCR demonstrated that the influenza A virus genome RNA was present in brain tissues as well as other organs and CSF. In contrast, CSF had little or no characteristics corresponding to meningoencephalitic conditions (pleocytosis and protein level elevation) in most reported cases. Immunohis-
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
621
tochemical study of the infected cynomolgus monkey brain showed that over 1 week, the infection was restricted to the injection site. The time between the onset of respiratory tract symptoms and CNS involvement is not long enough for the virus to replicate and spread to parts of the brain, which we previously observed in mice showing neurological symptoms due to encephalitis. Therefore, we can assume that the syndrome may occur mainly not by the direct invasion of the brain parenchymal tissues by the virus. Reports measuring proinflammatory cytokine levels in CSF from children with acute encephalitis showed that soluble TNF receptor 1 (sTNF-R1) and IL-6 levels were significantly higher in those who died or were left with sequelae than in those who survived without neurological deficits. The level of sTNF-R1 during the acute stage of encephalitis is an important index for predicting the neurological outcome. The serum level of IL-6 is also higher in patients with acute influenza virus-induced encephalopathy [6]. Furthermore, hyperactivated coagulation factors associated with DIC would participate in the pathogenesis [7,8]. Some residual mononuclear blood cells in the splenic vessels were double-positive for anti-virus and CD8 corresponding to suppressor T lymphocytes [9,10]. The barrier between the blood and susceptible extra-respiratory tissues is very strong under physiological condition because of the integrity and the insusceptibility of the endothelial cells lining blood vessels. In this study, we observed vascular endothelial cells under light microscopy and found no clear pathological alterations. Future studies are needed to look for ultrastructural changes of the endothelial cells as well as for the measurement of the humoral factors capable of acting on these cells [11].
Acknowledgements Drs. A. Matsuo and K. Iwasaki at Sasebo Municipal General Hospital kindly provided us a clinical record and autopsied samples. We thank Drs. K. Ito, K. Yamada and K. Yamada at Kurume University for generous gifts of the antibodies shown in Table 2. We also thank Drs. H. Noguchi and T. Ueda at Nagasaki Prefectural Institute on Public Health and Environmental Science for virus isolation. We are grateful to Dr. K. Ami at National Institute of Infectious Disease for animal experiments.
References [1] Y. Hayase, K. Tobita, Influenza virus and neurological disease, Psychiatry Clin. Neurosci. 51 (1997) 181 – 184. [2] R.D.K. Rye, G. Morgan, J. Baral, Encephalopathy and fatty degeneration of the viscera-a disease entity in childhood, Lancet II (1963) 749 – 752. [3] M. Mizuguchi, J. Abe, K. Mikkaichi, S. Noma, K. Yoshida, T. Yamanaka, S. Kamoshita, Acute necrotising encephalopathy of childhood: a new syndrome presenting with multifocal, symmetric brain lesions, J. Neurol., Neurosurg. Psychiatry 58 (1995) 555 – 561. [4] M. Levin, M. Hjelm, J.D.S. Kay, J.R. Pincott, J.D. Gould, R. Dinwiddie, D.J. Matthew, Haemorrhagic shock and encephalopathy: a new syndrome with a high mortality in young children, Lancet II (1983) 64 – 67. [5] S. Kimura, N. Ohtuki, A. Nezu, M. Tanaka, S. Takeshita, Clinical and radiological variability of influenzarelated encephalopathy or encephalitis, Acta Paediatr. Jpn. 40 (1998) 264 – 270.
622
M. Takahashi et al. / International Congress Series 1219 (2001) 615–622
[6] H. Tsuruoka, H. Xu, K. Kuroda, K. Hayashi, O. Yasui, A. Yamada, T. Ishizaki, Y. Yamada, T. Watanabe, Y. Hosaka, Detection of influenza virus RNA in peripheral blood mononuclear cells of influenza patients, Jpn. J. Med. Sci. Biol. 50 (1997) 27 – 34. [7] Y. Nagai, Protease-dependent virus tropism and pathogenicity, Trends Microbiol. 1 (1993) 81 – 87. [8] A. Nishino, M. Suzuki, T. Yoshimoto, H. Otani, H. Nagura, A novel aspect of thrombin in the tissue reaction following central nervous system injury, Acta Neurochir. (Suppl.) 60 (1994) 86 – 88. [9] I. Mori, T. Komatsu, K. Takeuchi, K. Nakakuki, M. Sudo, Y. Kimura, Viremia induced by influenza virus, Microb. Pathog. 19 (1995) 237 – 244. [10] P.G. Stevenson, S. Hawke, C.R.M. Bangham, Protection against lethal influenza virus encephalitis by intranasally primed CD8 + memory T cells, J. Immunol. 157 (1996) 3065 – 3073. [11] I. Joris, G. Majno, E.J. Corey, R.A. Lewis, The mechanism of vascular leakage induced by leukotriene E4, Am. J. Pathol. 126 (1987) 19 – 24.