Expression and localization of CyPD in tg human wtSOD1 and G93A^high-mSOD1 mouse spinal cord.
A. Spinal MNs (open arrows) in tg wtSOD1 mice have high immunoreactivity for CyPD compared to the surrounding neuropil and white matter.
B. In MNs of tg wtSOD1 mice, CyPD immunoreactivity is seen as discrete dots in the cytoplasm (arrows) of the cell body and proximal dendrites and diffusely in cytoplasm. Asterisk marks nucleus.
C. Spinal MNs (open arrows) in tg G93A^high-mSOD1 (early symptomatic shown here) mice have high immunoreactivity for CyPD compared to the surrounding neuropil and white matter matter.
D. In mSOD1 mouse MNs at early symptoms the CyPD-positive particles within the cytoplasm appear to swell.
E. In subsets of MNs showing near end-stage degeneration with severe swelling of the cell body (hatched lines delineate MNs) due to a slow necrotic-like process (Martin and Liu, 2004; Martin et al., 2007), CyPD immunoreactivity is marginated to the periphery of the cell (arrows) and decorates the nuclear surface (asterisk).
F. Western blots of mitochondrial membrane-enriched fractions of spinal cord of G93A^high-mSOD1 mice at pre-symptomatic and early symptomatic disease stages reveal levels of CyPD similar to tg wtSOD1 control, but at end-stage disease CyPD levels appeared slightly lower than control.

Ultrastructural and biochemical changes in G93A^high-mSOD1 tg mice are consistent with permeability transition pore activation.
A and B. EM shows major cristae remodeling and matrix swelling and vesiculation (black asterisks) in spinal cord ventral horn dendrites of G93A^high-mSOD1 mice. Normal mitochondria are marked by white asterisks. Mitochondria delineated by white-dotted boxes are shown at higher magnification in insets. Mitochondria with vacuoles (A, lower left and right; B, inset) have apparent OMM and IMM contacts (darker spots without narrow space separating the OMM and IMM, hatched arrows). The OMM appears breached at some locations (A, lower left and right, open arrows). Some mitochondria with vesiculation of the matrix have buds extending from the surface (A, lower right, double arrow), suggesting a non-replicative fission or fragmentation process. Some dendrites (B, open arrow) have mitochondria with extreme cristae remodeling and matrix vesiculation. A nearby mitochondrion exhibits early swelling (B, hatched arrow). Scale bars = 0.63 μm (A); 0.3 μm (A, insets); 0.3 μm (B); 0.15 μm (B, inset).
C. Graph showing the estimated number (mean ± SD) of contact sites per mitochondria in spinal cord ventral horn dendrites (identified by presynaptic contacts) of wtSOD1 and G93A^high-mSOD1 mice. Asterisk denotes significantly different (p < 0.05).
D. Apparent fragmentation of mitochondria was seen in spinal cord ventral horn dendrites of G93A^high-mSOD1 mice. Shown is a cross-sectional profile of a dendrite with four identifiable mitochondria (1-4). All of the mitochondria contain matrix vacuoles. Mitochondrion 4 has fragmented into several smaller mitochondrial buds with vacuoles (open arrows). Scale bar = 4 μm.
E. Apparent fission of mitochondria was seen in spinal MN cell bodies of G93A^high-mSOD1 mice. Contact sites between mitochondria are distinct (arrows). One mitochondrion is abnormal with pale matrix, but not showing vacuolation, while the other three mitochondria appear normal. Scale bar = 0.2 μm.
F. Western blot analysis for protein carbonyls (Oxyblot) in the mitochondrial membrane-enriched fraction and the soluble fraction of spinal cord from wtSOD1 tg mice and G93A^high-mSOD1 tg mice at different stages of disease. Molecular weight standards are indicated at right.
G. Increased nitration of CyPD and ANT in G93A^high-mSOD1 mouse spinal cord. Nitrated proteins from soluble and mitochondrial membrane-enriched fractions from wtSOD1 and mSOD1 tg mice at different stages of disease were IP with monoclonal mouse 3-nitrotyrosine antibody followed by western blotting with monoclonal CyPD or polyclonal ANT antibodies.

CyPD depletion changes the degenerative process in MNs of G93A-mSOD1 mice.
A and B. Cresyl violet (Nissl) staining shows that degenerating MNs in G93A mice with CyPD have prominent vacuolation of the cytoplasm (A, arrows), while MNs in G93A mice without CyPD have attenuated cytoplasmic vacuolation and the formation of numerous cytoplasmic granule-like structures (B, arrows). Asterisk marks nucleus. Scale bar in A (same for B) = 4 μm.
C and D. Immunoperoxidase staining for VDAC (brown signal) shows that degenerating MNs in G93A mice with CyPD have prominent mitochondrial swelling (C, arrows), while the mitochondrial swelling in MNs in G93A mice without CyPD is blocked (D, arrows). Asterisk marks nucleus. Scale bar in C (same for D) = 3.5 μm.

Expression and localization of mPTP proteins in non-tg mouse spinal cord.
A. Full-length blot demonstrating mono-specific detection of CyPD at ∼18-20 kDa in mitochondrial-enriched membrane fractions of mouse spinal cord. This immunoreactive band is not present in spinal cord of ppif^-/- mice. Molecular weight markers (in kDa) are indicated at left. Blots were re-probed for glyceraldehyde-phosphate dehydrogenase (GAPDH) as a protein loading control.
B. Full-length blot demonstrating mono-specific detection of ANT at ∼33 kDa in mitochondrial-enriched membrane fractions of mouse spinal cord, brainstem, and forebrain. Molecular weight markers (in kDa) are indicated at left. Blots were reprobed for CyPD and actin.
C. Full-length blot demonstrating mono-specific detection of VDAC at ∼35 kDa in mitochondrial-enriched membrane fractions of mouse spinal cord, brainstem, and forebrain. Molecular weight markers (in kDa) are indicated at left. Blots were reprobed for actin as a protein loading control.
D. Immunohistochemical localization of CyPD in mouse spinal cord. Antibody binding was visualized using immunoperoxidase with DAB as chromogen to give brown signal. A hemisection of lumbar spinal cord is shown. MNs in ventral horn are strongly immunopositive (arrows) for CyPD. Faint neuropil immunoreactivity is present in the dorsal horn and ventral horn and low immunolabeling is in white matter. Scale bar (same for F,G) = 100 μm.
E. Immunostaining for CyPD in ppif^-/- mouse spinal cord is negative. Scale bar = 80 μm.
F and G. Immunohistochemical localization of ANT (F) and VDAC (G) in mouse spinal cord. Antibody binding was visualized as in D but with cresyl violet counterstaining. A hemisection of lumbar spinal cord is shown. MNs in ventral horn are strongly immunopositive for ANT and VDAC (arrows) and neuropil immunoreactivity is present in the dorsal horn and ventral horn. Immunolabeling is in white matter is low.
H-J. Localization of CyPD (H), ANT (I) and VDAC (J) in mouse spinal MNs. Immunoreactivities for CyPD and ANT were seen as diffuse labeling in the cytoplasm and as discrete particles (arrows). VDAC immunoreactivity was seen as larger structures in the cytoplasm. The nucleus is marked by the asterisk. Dual immunofluorescent staining for CyPD, ANT, or VDAC and the mitochondrial marker SOD2 and confocal or epifluorescence microscopy (insets) shows co-localization of CyPD, ANT, and VDAC with SOD2 (seen as yellow-orange) in MNs, but there is also non-mitochondrial pools. Scale bar in H (same for I and J) = 7 μm. Scale bars in insets = 1.5 μm.

Formation of megamitochondria in MNs of mSOD1.
A and B. Confocal z-stack composite images showing the immunofluorescent detection of SOD2 (red), used to mark specifically mitochondria, in lumbar spinal MNs of tg wtSOD1 (A) and G93A^high-mSOD1 (B) mice at 10 weeks of age. The nucleus is the dark center. Vast numbers of mitochondria are present in MNs. In control MNs, mitochondria are small (∼0.3-0.5 μm in diameter) and pleiomorphic (round, ellipitical or filamentous). There is no detection of a non-mitochondrial pool. Mitochondria in MNs of mSOD1 mice become essentially all round. Some of these mitochondria are tiny (∼0.1 μm in diameter), suggesting a fragmentation or fission process (see Fig. 4D,E), while other mitochondria are huge (>2 μm in diameter) indicating swelling. There is no detection of a non-mitochondrial pool of SOD2 in mSOD1 mouse MNs indicating no disruption of mitochondria with release of matrix constituents. Scale bar in A (same for B) = 2.3 μm.
C, D, and E. Immunofluorescent localization of SOD2 (red) and CyPD (green) in lumbar spinal MNs of tg wtSOD1 (C) and G93A^high-mSOD1 (D, E) mice at 8 or 10 weeks of age. Colocalization of SOD2 and CyPD is seen as yellow-orange. In control MNs (C), CypD is present in mitochondria and in a non-mitochondrial cytoplasmic pool (green only). Some mitochondria are not positive for CyPD (red only). In mSOD1 mouse MNs before the formation of many mega-mitochondria (D, arrow), most mitochondria, many of which are seen as long vermiform structures, are positive for CyPD. In mSOD1 mouse MNs when many swollen mega-mitochondria are present (E, arrows), essentially all mitochondria are CyPD positive. In some swollen mitochondria CyPD fills the matrix and the SOD2 is marginated to the periphery of the matrix (green cores with red or orange halos). Scale bar in C (same for D and E) = 4 μm.
F and G. DECON analysis of mitochondrial abnormalities in MNs of pre-symptomatic G93A^high-mSOD1 tg mice compared to age-matched control human wtSOD1 tg mice. Mitotracker-labeled mitochondria from the distal axon/terminals in skeletal muscle are red. The nucleus is black profile in center (asterisk). Note the enlargement and clumping of mitochondria in the mSOD1 mouse MN (arrows). Scale bar in F (same for G) = 2 μm. The graphs below each image (generated by measuring pixel fluorescence intensity) are maps of mitochondria within the perikaryon of the actual MNs shown in the upper images. The graphs are projections of the imaged MNs. The empty zone in the center of the graph is the location of the nucleus. The fluorescence intensity was quantified in mid-perikaryal optical sections of MNs (see text). mSOD1 mouse MNs have large aggregates of brightly labeled mitochondria.

CyPD depletion improves survival of G93A^high- and G93A^low-mSOD1 mice and delayed the onset of disease. n=8-10 mice/genotype/gender. PCR genotyping of tail genomic DNA is shown for representative mice (A) above. Values are mean ± SD. Significant difference (*p < 0.05 or **p <0.01) from ppif^+/+ control.