Recently the number of reports of influenza-associated encephalopathy with multiple organ dysfunction has increased in Japan.1–3 A national survey of the prevalence and clinical features of the disease indicated that the rates of mortality and disability were 31.8 and 27.7%, respectively.4 Brain computed tomography (CT) and/or magnetic resonance imaging findings of fatal cases suggest massive brain edema of the cerebrum, basal ganglia and/or brain stem in the acute phase. Surviving patients with severe sequelae present with brain edema in the acute phase and subsequent cerebral atrophy in the convalescent phase. Findings on pathologic examination suggest that direct viral invasion and inflammation are not likely to cause this disease.1 Vascular damage with a subsequent leakage of plasma proteins and intravascular formation of thrombi is observed in systemic organs, as well as in the brain.4,5 Several studies of inflammatory cytokine concentrations suggest that inflammation in vessels and/or in central nervous system (CNS) has an important role in the pathogenesis of influenza-associated encephalopathy.1,3,6–8 The pathogenesis of vascular damage in influenza-associated encephalopathy is not fully explained, and the mechanisms of subsequent brain atrophy are not known.
Cytochrome c is an intramitchondrial protein normally residing in the intermembrane spaces. It triggers the execution phase of apoptosis by massive translocation into the cytoplasm, leading to Apaf-1-mediated caspase activation.9 We measured cytochrome c values as a marker of apoptosis and tumor necrosis factor (TNF)-α concentrations as a marker of inflammation in the serum and cerebrospinal fluid (CSF) samples collected simultaneously at the acute and convalescent phases and clarified the roles of apoptosis and inflammation in the pathogenesis of influenza-associated encephalopathy.
MATERIALS AND METHODS
Ten children were diagnosed with influenza-associated acute encephalopathy from January 2000 to December 2003 (Table 1). Acute encephalopathy was defined as an acute CNS disorder characterized by seizures and unconsciousness for more than 24 hours and brain edema detected by CT. Diagnosis of type AH3N2 infection (cases 1, 2, 4, 6 and 7) was based on viral isolation from throat swabs. Diagnosis of influenza A infection (cases 3, 5, 8 and 10) and B infection (case 9) was based on viral antigen detection from throat swabs with an enzyme-linked immunosorbent assay kit (Directigen flu A and B; Nippon Becton Dickinson, Tokyo, Japan). Other neurologic, metabolic, endocrine, toxic and drug-induced disorders were excluded. No patients received influenza vaccination before the influenza season in which the patients developed the illness. Three patients died (cases 1–3), 3 patients survived with sequelae (cases 4–6) and 4 patients survived without sequelae (cases 7–10). Amantadine and oseltamivir were administered immediately after the neurologic onset of the illness to cases 1 and 7 and to cases 2, 3, 4, 5, 6, 8, 9 and 10, respectively. Three patients developed brain stem dysfunction, such as extraocular dysmotility and disappearance of the light reflex within 6 hours from neurologic onset, and died within a few days. At the acute phase, whole cerebral edema and localized cerebral edema were observed on CT examination in 6 and 4 patients, respectively. Three patients with a poor prognosis (cases 1–3) had whole cerebral edema with an area of low density in the brain stem and/or bilateral thalami. At the convalescent phase, 3 patients with sequelae (cases 4–6) had cerebral edema followed by cortical atrophy. Serum samples were collected from all patients at neurologic onset (10 samples), from 9 patients at the exacerbation phase (11 samples, 1–2 days after onset) and from the surviving 7 patients at the convalescent phase (10 samples, 3–8 days after onset). CSF samples were collected from 9 patients at neurologic onset (9 samples) and from the surviving 7 patients at the convalescent phase (8 samples, 4–8 days after onset). Lumbar puncture was not performed in 1 patient with a poor prognosis at neurologic onset because of massive brain edema. Serum and CSF samples were also collected for cytochrome c assay and cytokine assay from 8 patients with epilepsy subjected to routine analysis as negative controls. Informed consent for measurement of cytokine and cytochrome c concentrations was obtained from all patients' parents.
A commercially available enzyme-linked immunosorbent assay kit (Endogen, Inc., Woburn, MA) was used to measure the TNF-α concentration. Cytochrome c concentrations in the serum and CSF were measured with an originally developed sandwich electrochemiluminescence immunoassay. Microbeads (Dynabeads M-450 Epoxy; DYNAL AS, Oslo, Norway) coated with anti-cytochrome c monoclonal antibody and anti-cytochrome c polyclonal antibody conjugated with ruthenium chelate (Ru-Ab) were used for the cytochrome c immunoassay. We mixed 50 μL of sample, 150 μL of dilution buffer and 25 μL of coated microbeads. After incubation at 30°C for 9 minutes, the microbeads were washed twice to remove nonreacted specimens, and Ru-Ab was added. The microbeads were then washed twice to remove nonreacted Ru-Ab and placed into magnet-mounted flow cell electrodes to measure the quantity of the emission. The cytochrome c concentration of the sample was calculated with the use of rat cytochrome c solutions in concentrations between 100 ng/mL to 10 pg/mL as a standard. All 3 operations were performed automatically by the Picolumi 8220 (Sanko Junyaku Co., Ltd., Tokyo, Japan), except for the dilution of the sample.
Statistical analysis was performed on a Macintosh computer with a software package for statistical analysis (StatView; Abacus Concepts, Berkeley, CA). Differences among laboratory data of each group were assessed with the Mann-Whitney rank sum test.
TNF-α and Cytochrome c Values in Control Serum and CSF
Mean values and standard deviations of TNF-α concentrations in the serum and CSF were calculated as 2.2 ± 2.1 and 2.0 ± 2.9 pg/mL, respectively. Mean values and standard deviations of cytochrome c concentrations in the serum and CSF were 610 ± 130 and 540 ± 90 pg/mL, respectively. Therefore we established the cutoff values for TNF-α in serum and CSF as 10 pg/mL and those for cytochrome c in the serum and CSF as 1000 pg/mL.
Serum TNF-α Values.
TNF-α values were >50 pg/mL in 3 serum samples collected from patients with a poor prognosis at neurologic onset (67, 56 and 53 pg/mL, respectively). The concentrations decreased 1–2 days after onset, while the patients were still in the exacerbation phase. Three patients with sequelae and 1 patient without sequelae had slightly elevated serum cytokine values at the acute phase (21, 12, 11, and 13 pg/mL, respectively), and the other 3 patients without sequelae had normal serum TNF-α concentrations (<10 pg/mL). TNF-α values in serum samples collected from patients with or without sequelae were not increased during the exacerbation and convalescent phases. Serum TNF-α values at neurologic onset were significantly higher in patients who died than in patients who survived (P = 0.017).
Serum Cytochrome c Values.
(Fig. 1). Cytochrome c values in the serum collected from patients with a poor prognosis at the neurologic onset were >2000 pg/mL. The concentration increased to 8000 pg/mL or more at the exacerbation phase. One patient with severe sequelae had a high serum cytochrome c value (1100 pg/mL) at the acute phase, and the level gradually decreased to the normal range at the convalescent phase. Cytochrome c concentrations of the other 6 patients were not high (<1000 pg/mL) at neurologic onset and did not increase at the exacerbation and convalescent phases. Serum cytochrome c values at the neurologic onset were significantly higher in patients who died than in patients who survived (P = 0.017).
CSF TNF-α Values.
There was no pleocytosis in the 9 CSF samples collected at the neurologic onset, and the 8 samples collected at the convalescent phase. TNF-α values in the CSF samples were not high (<10 pg/mL). TNF-α concentrations in the 2 samples collected from patients with a poor prognosis at the neurologic onset were not detectable (<1 pg/mL).
CSF Cytochrome c Values (Fig. 1)
Cytochrome c values in the 9 CSF samples collected at neurologic onset were <1000 pg/mL. Cytochrome c concentration in the 2 samples collected from patients with a poor prognosis were less than the detection limit (<500 pg/mL). Cytochrome c concentrations at the convalescent phase increased compared with those at the acute phase. In particular, CSF samples collected from patients with subsequent cerebral atrophy had very high cytochrome c concentrations (>10,000 pg/mL) at the convalescent phase. The cytochrome c concentrations in CSF were >20 times higher than those in the serum simultaneously collected at the convalescent phase. CSF cytochrome c values at the convalescent phase were significantly higher in patients who survived with sequelae than in patients who survived without sequelae (P = 0.034).
At the acute phase, serum TNF-α and cytochrome c concentrations of patients with poor prognosis were significantly elevated compared with those with good outcomes. Serum TNF-α and cytochrome c values higher than 50 pg/mL and 8000 pg/mL, respectively, in the acute phase might help predict patients with a poor prognosis. The results indicated that inflammation and apoptosis in vascular vessels were involved in the development of acute encephalopathy. At the convalescent phase, CSF cytochrome c concentrations increased significantly (>10,000 pg/mL) in 3 patients with subsequent brain atrophy, whereas serum cytochrome c values were low. Cytochrome c is produced inside the CNS and is unlikely to be transported out via the blood-brain barrier. These results suggest that apoptosis of the parenchyma in the CNS contributes to the cerebral atrophy observed in patients with sequelae. Brain atrophy is not yet apparent when high concentrations of cytochrome c are detected in the CSF 6–8 days after neurologic onset.
In conclusion, inflammation and apoptosis in vascular vessels and apoptosis in the CNS have significant roles in the pathogenesis of brain edema at the acute phase and subsequent brain atrophy at the convalescent phase, respectively, of influenza-associated encephalopathy.
1. Ito Y, Ichiyama T, Kimura H, et al. Detection of influenza virus RNA by reverse transcription-PCR and proinflammatory cytokines in influenza-virus-associated encephalopathy. J Med Virol
2. Kasai T, Togashi T, Morishima T. Encephalopathy associated with influenza epidemics. Lancet
3. Kawada J, Kimura H, Ito Y, et al. Systemic cytokine responses in patients with influenza-associated encephalopathy. J Infect Dis
4. Morishima T, Togashi T, Yokota S, et al. Encephalitis and encephalopathy associated with an influenza epidemic in Japan. Clin Infect Dis
5. Takahashi M, Yamada T, Nakashita Y, et al. Influenza virus-induced encephalopathy: clinicopathologic study of an autopsied case. Pediatr Int
6. Ichiyama T, Nishikawa M, Yoshitomi T, Hayashi T, Furukawa S. Tumor necrosis factor-α, interleukin-1β, and interleukin-6 in cerebrospinal fluid from children with prolonged febrile seizures: comparison with acute encephalitis/encephalopathy. Neurology
7. Aiba H, Mochizuki M, Kimura M, Hojo H. Predictive value of serum interleukin-6 level in influenza virus-associated encephalopathy. Neurology
8. Ichiyama T, Isumi H, Ozawa H, Matsubara T, Morishima T, Furukawa S. Cerebrospinal fluid and serum levels of cytokines and soluble tumor necrosis factor receptor in influenza virus-associated encephalopathy. Scand J Infect Dis
9. Li P, Najhawan D, Budihardjo I, et al. Cytochrome c
and dATP-dependent formation of Apaf-1/caspase-9 complex initiates an apoptotic protease cascade. Cell