Caspase-mediated cleavage of actin and tubulin is a common feature and sensitive marker of axonal degeneration in neural development and injury

Abstract

Background: Axon degeneration is a characteristic feature of multiple neuropathologic states and is also a mechanism of physiological neurodevelopmental pruning. The vast majority of in vivo studies looking at axon degeneration have relied on the use of classical silver degeneration stains, which have many limitations including lack of molecular specificity and incompatibility with immunolabeling methods. Because Wallerian degeneration is well known to involve cytoskeletal disassembly and because caspases are recently implicated in aspects of this process, we asked whether antibodies directed at caspase-generated neoepitopes of beta-actin and alpha-tubulin would be useful immunohistochemical markers of pathological and developmental axon degeneration.

Results: Here we demonstrate that several forms of axon degeneration involve caspase-mediated cleavage of these cytoskeletal elements and are well-visualized using this approach. We demonstrate the generation of caspase-induced neoepitopes in a) an in vitro neuronal culture model using nerve growth factor-deprivation-induced degeneration and b) an in vivo model using ethanol-induced neuronal apoptosis, and c) during normal developmental pruning and physiological turnover of neurons.

Conclusions: Our findings support recent experimental data that suggests caspase-3 and caspase-6 have specific non-redundant roles in developmental pruning. Finally, these findings may have clinical utility, as these markers highlight degenerating neurites in human hypoxic-ischemic injury. Our work not only confirms a common downstream mechanism involved in axon degeneration, but also illuminates the potential utility of caspase-cleavage-neoepitope antibodies as markers of neurodegeneration.

Figure 1

Demonstration of specificity of caspase cleavage neoepitope antibodies to actin and tubulin.In vitro cleavage of the substrates was performed via incubation of either recombinant active caspase-3 or −6 with permeabilized brain-derived synaptosomes. A western blot was performed using lysate from this preparation. Blot (a) was colabeled with general anti-alpha-tubulin (tubulin, red) and TubulinΔCsp6 (Cleaved-Tub, green). Blot (b) was colabeled with anti-actin (red) and anti-fractin (green). Recombinant cleaved-caspase-3 (rCC3) generates both cleaved tubulin (a) and fractin (b), whereas recombinant cleaved-caspase-6 (rCC6) only generates only cleaved tubulin (a).

Figure 2

NGF-deprivation induces caspase-mediated cleavage of actin and tubulin. (a, b) NGF-replete (+NGF) cultured sympathetic neurons are viable and axons are intact, as visible with β₃ tubulin (Tuj1) staining. NGF-withdrawal-induced apoptosis causes axon degeneration by 12 hours that is marked with antibodies against cleaved actin (fractin) (a) or TubulinΔCsp6 (cTub) (b) and is suppressed by caspase inhibition using ZVAD. Intensity of fractin (c) and TubulinΔCsp6 (d) staining in axons was measured in arbitrary fluorescence units (a.u.). A diagram of the microfluidic chamber used is shown in (e). Data represents three independent experiments. One-way ANOVA, *p < .05, **p < .01. Scale bar = 10 μm.

Figure 3

Fractin and TubulinΔCsp6 labeling highlight degenerating axon tracts after ethanol-induced apoptosis. (a) Low power image of co-labeling in the striatum using anti-tubulinΔCsp6 (cTub; green) along with anti-neurofilament (NFM; red) reveals that tubulinΔCsp6 localizes to axon tracts 24 hours post-ethanol treatment. (b-c) High power image of co-labeling in the striatum using fractin or TubulinΔCsp6 (green) along with a marker for neurofilaments (red) reveals that both fractin and TubulinΔCsp6 highlight blebbing material within axon tracts after ethanol-induced apoptosis (b,d) and are absent in control (saline-treated) brain (c,e). (f-g) While ethanol-induced apoptosis led to CC3-positive cell bodies, CC3 was not usually seen in axon fibers via conventional staining techniques. Tyramide signal amplification (TSA) was required for visualization of CC3 in axon fibers after ethanol-induced apoptosis (shown are striatal pencil fibers). CC3 with TSA (CC3 + TSA) was only detected in tissue after ethanol-induced apoptosis (f), and CC3 staining was not seen in control tissue (g). Scale in (a) is 200 μm. Scale in (b-g) is 50 μm.

Figure 4

Generation of caspase-dependent cleaved actin and tubulin neoepitopes in degenerating axons is BAX-dependent. The amount of CC3, fractin and TubulinΔCsp6 staining was quantified 6 hours after injury via Cell profiler. Data was normalized to the value calculated from the wild type ethanol-treated tissue in order to represent relative levels in each genotype and treatment. BAX deficient mice showed nearly complete loss of ethanol-induced generation of fractin and cleaved tubulin immunoreactivity.

Figure 5

TubulinΔCsp6 highlights degenerating axon tracts during normal developmental pruning. Adjacent sections from embryonic day 14.5 mice were immunolabeled with CC3 (a), fractin (b) and tubulinΔCsp6 (cTub) (c) (green) along with neurofilament (2H3; red). All three antibodies reveal a subset of apoptotic cell bodies in the ganglion, but only cTub prominently labels axons emanating from the ganglia (c,d). Notice that cTub does not label any structures in adjacent non-neural somites (above and below axon tracts). Scale = 200 μm.

Figure 6

Fractin and TubulinΔCsp6 highlight degenerating axon tracts during turnover of olfactory sensory neurons in the adult mouse brain. Axons terminating in the glomerular layer of the adult olfactory bulb were assessed; the region with axon terminals is indicated by the box in (a). Adjacent sections were stained with CC3 (b), fractin (c) and TubulinΔCsp6 (cTub) (d) (green) along with neurofilament-medium (NFM; red) and the glomeruli were imaged. Fractin and TubulinΔCsp6 highlight structures consistent with fragmented neurites in a subset of glomeruli (indicated by filled arrowheads and hatched circles), likely representing neurites from olfactory sensory neurons undergoing physiological turnover. The olfactory nerve layer composed only of sensory axons, contains the highest density of TubulinΔCsp6 immunoreactive neurites (open arrowhead). Other regions of the olfactory bulb were devoid of neoepitope marker labeling. Scale bar = 50 μm.

Figure 7

Areas of ischemic injury in human brain are identified by infiltration of monocytes and microgliosis and there is regional overlap with markers of axon degeneration. Adjacent sections were stained (arrow in (a,b) identifies the same vessel). Areas of injury (to the right of the dashed line) are identified by infiltration of monocytes (a), stained by the monocyte marker CD163, as well as by a marker for degenerating axons, APP (b). In addition, a microglial marker (Iba1) highlights enlarged ameboid microglia in areas of injury, which is suggestive of microglial activation (c). Injured regions identified by monocyte infiltration and microgliosis also showed evidence of axon degeneration; shown here is TubulinΔCsp6 (cTub) staining (d-e). Scale in (a-b) is 200 μm, scale in (c-e) is 25 μm.

Figure 8

Fractin and TubulinΔCsp6 highlight apoptotic bodies and degenerating neurites in areas of ischemic injury in human brain. Adjacent sections were co-labeled with CC3 (green) and SMI32 (red) (a,d,g), or fractin (green) and neurofilament (red) (b,e,h), or TubulinΔCsp6 (cTub) and neurofilament (red) (c,f,i). Areas of injury contained apoptotic neurons and degenerating axons (Row 1 and Row 2). (Row 1) Insets show DAPI-labeled pyknotic nuclei (arrowhead) in apoptotic corpses that are labeled with CC3 (a), fractin (b) and TubulinΔCsp6 (c). (Row 2) Fractin (e) and TubulinΔCsp6 (f) antibodies highlight degenerating axons, whereas CC3 (d) fails to label degenerating axons via conventional staining techniques. (Row 3) Uninjured areas (g-i) have decreased SMI32 staining and no fractin or TubulinΔCsp6-labeled axons. Scale = 25 μm.