Failure to complete apoptosis following neonatal hypoxia-ischemia manifests as "continuum" phenotype of cell death and occurs with multiple manifestations of mitochondrial dysfunction in rodent forebrain

Abstract

Controversy surrounds proper classification of neurodegeneration occurring acutely following neonatal hypoxia-ischemia (HI). By ultrastructural classification, in the first 24 h after neonatal hypoxia-ischemia in the 7-day-old (p7) rat, the majority of striatal cells die having both apoptotic and necrotic features. There is formation of a functional apoptosome, and activation of caspases-9 and -3 occurring simultaneously with loss of structurally intact mitochondria to 34.7+/-25% and loss of mitochondrial cytochrome c oxidase activity to 34.7+/-12.7% of control levels by 3 h after hypoxia-ischemia. There is also loss of the mitochondrial motor protein, kinesin. This combination of activation of apoptosis pathways simultaneous with significant mitochondrial dysfunction may cause incomplete packaging of nuclear and cytoplasmic contents and a hybrid of necrotic and apoptotic features. Evidence for an intermediate biochemistry of cell death including expression of the 17 kDa isoform of caspase-3 in dying neurons lacking a classic apoptotic morphology and degradation of the neuronal cytoskeletal protein spectrin by caspase-3 and calcium-activated calpains yielding 120 kDa and 145/150 kDa fragments, respectively, is also found. In summary, neonatal hypoxia-ischemia triggers apoptotic cascades, and simultaneously causes mitochondrial structural and functional failure. The presence of a "continuum" phenotype of cell death that varies on a cell-by-cell basis suggests that the phenotype of cell death is dependent on the energy available to drive the apoptotic pathways to completion.

Figure 1. Constituents of the apoptosome Complex accumulate and caspases 9 and 3 are activated following neonatal HI in soluble fraction from injured rat forebrain

A. Immunoblots showing increased levels of both APAF 1 and cytochrome c and the presence of caspase-9 in soluble protein extracts from ipsilateral injured forebrain following neonatal HI in comparison to uninjured controls. B. Increased biochemical activity of caspases 9 and 3 (* p≤ 0.05 vs control, ANOVA with Dunnett’s post hoc test) after HI. C. The active p17 cleaved fragment of caspase-3 is found with increasing abundance at 6 and 24 hours after HI compared to control. Commassie stained gels or ponceau stained nitrocellulose membranes are shown as loading controls for each of the immunoblots in A and C.

Figure 2. Multiple manifestations of mitochondrial dysfunction following neonatal HI

A. Individual mitochondria from striatal neurons were classified as having intact inner cristae and inner and outer membrane or swollen and disrupted inner cristae and separation of the inner and outer membrane as shown. Using these criteria, percent intact mitochondria within individual striatal neurons was calculated in uninjured control neurons and at 3, 6, and 24 hours after HI (graph, mean± SD, n =25 at each time point, * p≤ 0.05 vs control, ANOVA with Dunnett’s post hoc test). B. Mitochondrial complex IV activity rapidly declines within the first 3 hours following neonatal HI (mean ±SD of % control activity, n= 6 samples at each time point, * p≤ 0.05 vs control, ANOVA with Dunnett’s post hoc test). Partial recovery at 6 hours is not sustained and there is more profound secondary decrease at 24 hours. C. Dying striatal neurons with different cell death phenotypes have varying preservation of cytochrome oxidase histochemical activity. Cells with classic apoptotic phenotype maintain intense cytochrome c oxidase activity (Ca, brown immunoreactivity, arrows) compared to cells with incomplete chromatin condensation, consistent with ‘continuum’ phenotype (Cb) and cells with irregularly and minimally condensed chromatin (Cc). D. Immunoblot showing loss of kinesin protein as soon as 3 hours after HI. This coincides with the loss of structural integrity and functional activity of mitochondria following neonatal HI (panels A and B). E. Healthy neurons from control striatum have intact cytoplasmic and nuclear membranes, well dispersed chromatin with nucleoli visible, intact mitochondria dispersed throughout cytoplasm (Ea, scale bar=2μm). In contrast, in injured neurons, perinuclear clustering of mitochondria occurs within darkened and condensed cytoplasm, possibly from loss of anterograde trafficking of mitochondria (Eb) (mitochondria indicated with arrows in Ea, Eb, and Ec) (Al-Abdulla and Martin, 1998, Northington et al., 2001a). Healthy appearing mitochondria are abundant in cross sections of axons of uninjured neurons (Ec) and almost completely absent in swollen injured axons (Ed, asterisks) at 6 hours following HI (scale bars =1μm).

Figure 3. Cell death ultrastructure following neonatal HI is intermediate between necrosis and apoptosis

There is EM evidence for apoptotic, continuum and necrotic cell death within dying striatal neurons at 3 (panels A–E), 6 (panels F–J) and 24 hours (panels K–O) after HI. Necrotic cells with translucent cytoplasm, dispersed and minimal chromatin condensation within an intact nuclear membrane (panel A), and classically apoptotic cells with intact cell membrane, dense cytoplasm, and highly organized chromatin condensation with complete loss of the nuclear membrane (panels E, J, O) are seen but are relatively rare. Neurons with structures associated with excitotoxic neurodegeneration (Ishimaru et al., 1999) are seen (panels B, G, and K). The dying cells in panels C, D, H, I, L, M, and N have an appearance consistent with incomplete packaging of nuclear and cytoplasmic contents including more or less condensed cytoplasm within an intact cell membrane, variably swollen and disrupted mitochondria and vacuolated cytoplasm, irregularly condensed chromatin with at least partial segregation of nuclear and cytoplasmic contents. Representative examples of nuclear membranes are identified with double arrows, irregularly condensed chromatin with single arrows, fully condensed chromatin with asterisks, cell membrane with triple arrows, disrupted mitochondria with m, intact mitochondria with M, vacuoles with v, scale bar=2μm for panels A–O.

Figure 4. Biochemical evidence for incomplete execution of apoptotic cell death following neonatal HI

A. Clusterin. a multimeric 35–40 kDa glycoprotein, thought to be a marker for excitotoxic neurodegeneration (Han et al., 2001) is present in the soluble fraction from ipsilateral forebrain as early as 3 hours, and is markedly elevated at 24 hours following neonatal HI simultaneous with increased p17 caspase 3 fragment (Figure 1C). The commassie stained gel is shown as the loading control. B. Histochemical expression of the 17 kDa active fragment of caspase-3 is found in dying striatal neurons following neonatal HI (brown immunoreactivity, cells counter-stained with CV). The 17 kDa isoform of caspase-3 is present in cells dying with a variety of nuclear morphologies. Cells dying from programmed cell death in control tissue (Ba) as well as cells with classic apoptotic morphology in injured striatum (Bb) are immunoreactive for 17 kDa caspase-3 as expected. Other striatal neurons with immunoreactivity for 17 kDa caspase-3 have nuclear morphologies not consistent with classic apoptosis, but rather show incomplete nuclear fragmentation (Bc) and some have minimal evidence of chromatin condensation (Bd). Spillage of p17 caspase 3 into extracellular space is seen in cells with the least evidence of organized chromatin condensation (arrow) presumably due to disruption of cell membrane. C. Immunoblot showing cleavage of 240 kDa full length spectrin, an important cytoskeletal protein, is shown. Expression of the 145/150 kDa spectrin fragment, produced by excitotoxic- calpain mediated fragmentation occurs as early as 3 hours following HI while expression of the 120 kDa caspase-cleaved fragment appears at 3 and 6 hours but in much more abundance at 24 hours after HI. The commassie stained gel is shown as the loading control.