Formation of multinucleated giant cells and microglial degeneration in rats expressing a mutant Cu/Zn superoxide dismutase gene

Animal use protocols were approved by the University of Florida Institutional Use and Care of Animals Committee (IUCAC). All transgenic animals used in this study were male Sprague Dawley NTac:SD-TgN(SOD1G93A)L26H rats obtained from Taconic Farms where animals were screened extensively for infections prior to shipping. Upon arrival animals were housed under SPF conditions. Age-matched, wild type Sprague Dawley rats were purchased from Harlan. The time course of disease progression varied among individual animals, but in general once symptoms developed disease progression was quite rapid causing death of most animals by 5 months of age.

To examine microglial morphology, microglial markers were used at three stages of the disease: 1) presymptomatic stage, where animals had no apparent muscle weakness. Animals studied in this group were aged 74–84 days; 2) early symptomatic stage (onset), where animals first showed evidence of hind limb weakness. Animals studied in this group were aged 113–117 days; 3) late symptomatic (end stage), where animals were no longer able to right themselves after 30s. Animals studied in this group were aged 135–156 days. For each of the three disease stages, 4 transgenic and 4 age-matched wild type control animals were used.

Animals were deeply anesthetized with pentobarbital and perfused transcardially with phosphate buffer saline (PBS) followed by a fixative solution containing 4% paraformaldehyde in PBS. The spinal cord and brain were dissected out and fixed overnight in 4% paraformaldehyde at 4°C, transferred to 30% sucrose and then frozen. Lumbar spinal cord, cortical, and brainstem sections were cut in the coronal plane at 20 μm on a cryostat, mounted on slides and air dried. Sections were pretreated in PBS with 0.5% Triton X-100 for 15 min, blocked in 10% normal goat serum for 30 min and incubated overnight at room temperature in the primary antibody diluted in buffer. The primary antibodies included MRC OX-42 (Serotec, Cambridge, UK) and MRC OX-6 (Serotec, Cambridge, UK) at 1:500. The slides were rinsed in PBS and incubated in secondary antibody (1:500) for 1 h. Following incubation, slides were rinsed and Horseradish Peroxidase Avidin D was applied (1:500; Vector, Burlingame, CA) and incubated for 30 min. Slides were washed and immunoreactivity was visualized with 3,3'-diaminobenzidine (DAB)-H₂O₂ substrate. After a brief rinse, slides were dehydrated in increasing concentrations of ethanols, cleared in xylene, and coverslipped using Permount mounting medium (Fisher Scientific).

OX-42 immunoreactivity in the ventral spinal cord was quantified using Image Pro Plus software (version 4.5.1, Media Cybernetics, Carlsbad, CA). The area occupied by stained cells was highlighted and measured for each section of spinal cord (6 sections per animal) then expressed as a percentage of total area of ventral spinal cord. Using GraphPad Prism software (San Diego, CA) a t-test was performed to determine statistical significance between transgenic SOD 1 and control animals at each time point. A one-way ANOVA was performed to compare differences among the transgenic animals followed by a Tukey multiple comparison test.

Animals were deeply anesthetized and transcardially perfused with phosphate buffer saline (PBS) followed by a fixative solution containing 4% paraformaldehyde. The spinal cord and brain were dissected out and fixed 2 h in 4% paraformaldehyde. The tissue was dehydrated through ascending alcohols, cleared in xylenes and embedded in paraffin. Serial 7 μm coronal sections were collected and mounted on slides. Sections were deparaffinized through xylenes, graded alcohols and rinsed in PBS. Next, the slides were trypsin treated (0.1% trypsin, 0.1% CaCl₂) for 12 min at 37°C. Following a 10 min wash the slides were incubated overnight at 4°C in lectin GSA I-B₄-HRP (Sigma Chemical Co.) diluted 1:10 in PBS containing cations (0.1 mM of CaCl₂, MgCl₂ and MnCl₂) and 0.1% Triton X-100. After overnight incubation slides were briefly rinsed in PBS and visualized with 3,3'-diabimobenzidine (DAB)- H₂O₂ substrate. Sections were counterstained with cresyl violet, dehydrated through ascending alcohols, cleared in xylenes and coverslipped with Permount.

The CR3 complement receptor recognized by OX-42 antibody is expressed constitutively by all resting and activated microglial cells [21]. OX-42 immunoreactivity observed in presymptomatic SOD1 transgenic rats was similar to that seen in wild type control, i.e. there was uniform staining of all resting microglia (Figs. 1A,D). Occasionally, in these presymptomatic animals cell fusions involving several microglia were observed (Fig. 1A, inset). The onset of symptoms was associated with a dramatic increase in OX-42 staining in the ventral horn due to much greater microglial cell numbers (Fig. 1B). Many of these seemingly activated microglia were clustered and/or fused into multi-cellular aggregates. In end stage animals, overall immunoreactivity with OX-42 was decreased compared to that seen in animals with disease onset (Fig. 1C). This unexpected diminution in microglial staining was due to widespread degenerative cytoplasmic fragmentation affecting most, if not all microglia within the ventral horn (see below). The qualitatively evident increases and decreases in immunoreactivity were confirmed through quantitative morphometric measurements (Fig. 1E).

With onset of symptoms, there was apparent activation of microglia as judged by the dramatic increase in OX-42 immunoreactivity in the spinal gray matter. Examining sections at low power clearly revealed pronounced spots of enhanced OX-42 staining in the ventral horns (Fig. 2A), and these were judged initially to be due to the formation of microglial phagocytic clusters around dying motor neurons, as this would be a normal response to motor neuron death. However, when spots of intense OX-42 immunoreactivity were examined at higher power (Fig. 2B,D) they appeared unusual in that individual microglial phagocytes were not discernable. Subsequent counterstaining of these sections with cresyl violet allowed us to conclude that the OX-42 reactive structures were, in fact, not phagocytic clusters but represented multinucleated giant cells (Figs. 2C,E). These giant cells were found in all SOD1^G93A transgenic rats studied. They formed apparently as a result of multiple microglial cells fusing together into sizable syncytia (40–50 μm) that often showed a circular arrangement of microglial nuclei about their periphery (Fig. 2E). This kind of nuclear arrangement is classically associated with multinucleated giant cells of the Langhans type. The cytoplasmic interior of Langhans giant cells appeared granular and fragmented, suggesting ongoing deterioration. A few of the giant cells revealed the presence of apoptotic bodies, evident as nuclear fragments (Fig. 2F), but overall apoptotic bodies either inside or outside of giant cells were sparse. Microglia dispersed in between giant cells revealed relatively normal process-bearing morphology and lacked the conspicuous hypertrophy that is characteristic of activated microglia. (Fig. 2E). However, some sections showed ongoing microglial cytorrhexis, i.e. fragmentation of the cytoplasm. Cytorrhexis became conspicuous in animals that were in the terminal stages of the disease process (Fig. 3) and was evident as a loss of discernable microglial cell structure and presence of abundant OX-42 immunoreactive fragments of microglial cytoplasm dispersed throughout the spinal gray matter (Figs. 3A,B,D,E). Occasional giant cells could still be observed during end stage disease, however, most of these showed signs of deterioration evident by increased irregularity of their shape and nuclear arrangement, as well as by increased granularity and fragmentation (Figs. 3A,B). In some sections, neurons remained stained with cresyl violet suggesting residual preservation of neuronal integrity. However, pathological features were evident in motor neurons, including most notably intense hyperchromia with formation of a nuclear cap consisting of condensed chromatin material (Fig. 3C). This neuronal appearance stood in stark contrast to that of normal motor neurons as seen in wild type animals (Fig. 3F).

Sections from the brainstem at the level of cranial nerve VII during disease onset and end stage were marked by changes indicative of severe neuropathology (Fig. 4). They included prominent, widespread vacuolization of the extracellular space and hyperchromia of neuronal processes. Often neurites appeared physically separated (as if torn) from neuronal cell bodies leaving one or more distinct stumps on the perikaryon (Fig. 4D,E). The changes affecting microglia were striking in that multinucleated giant cells were present throughout any given section. These consisted of fused microglial cells that gave rise to a variety of bizarrely shaped cellular fusions which, in some cases, extended for more than one hundred micrometers in length (Figs. 4B,C,F). Microglial fusions varied in size, sometimes involving only a few cells, and other times twenty or more. Although not obviously associated with vascular channels, some microglial giant cells due to their elongated shape seemed to have formed along blood vessels (Fig. 4C). Presence of giant cells was observed in all animals regardless of whether they were at an early or late symptomatic stage of motor neuron disease. They were scattered seemingly at random throughout the brainstem and not limited to any particular nucleus or tract, and often displayed the classic morphological features of Langhans type giant cells (Fig. 4G).

Within vacuolated spaces rounded, shrunken microglia exhibiting nuclear fragmentation or shrinkage (pyknosis) were identified using lectin histochemical staining (Figs. 4H).

Microglial fusions similar to those seen in the spinal cord and brainstem level were found also in the red nucleus of the midbrain (Figs. 5A–D). The specificity with which these microglial fusions were restricted to the red nucleus area was remarkable, as they were visible even at the lowest magnification (Fig. 5A). Microglia outside of the red nucleus displayed normal, ramified morphology. Rubrospinal neurons appeared normal in size and morphology, as well as in number, and there was no evidence to suggest that any of these neurons were undergoing degeneration. Rubrospinal neurons were not encircled by activated microglia. It is noteworthy also that motor neurons in the oculomotor nucleus, which appears with the red nucleus in the same sections, revealed no evidence of degenerative changes, and microglia here were normal and non-activated in appearance. Similarly, microglia in the substantia nigra appeared completely normal (Fig. 5F). Somewhat surprisingly, we also found no evidence at all for microglial activation or abnormalities in the motor cortex of animals, regardless of disease stage, with any of the microglial markers employed (Figs. 5G,H).