We established a combination of an in vivo double-labeling and an in vivo double-tracing approach by applying two fluorescent dyes in living animals. We also were able to detect the temporal/spatial pattern of labeling repetitively in vivo. The goal was to study the relationship between the fate of the cell (death/survival) and the recovery of axonal transport in living tissue.
In comparison to the kinematics of (double)-labeling in unoperated animals, optic nerve damage significantly influenced the axonal transport dynamics:
-
There were two waves of increased retrograde axonal transport: within the first days after ONC and approximately 3 weeks post lesion;
-
There was a delayed wave of increased cell death around 3 weeks post ONC; and
-
Recovery of axonal transport was not associated with improved survival but rather enhanced cell death.
The fact that there is preserved trafficking between days 2 to 4 post ONC is surprising, as so far mainly the deteriorating effects of impaired/interrupted axonal transport after traumatic brain injury have been observed.
21 However, our data suggest that during this early time-window axonal transport may be even accelerated or, another possibility, slow transport mechanisms are specifically destroyed. These hypotheses are based on the observations that maximum labeling was accomplished already 2 days after injection (in unlesioned animals it took 6 days until maximum labeling was achieved). This effect is unlikely due to completely interrupted axonal transport after day 2 as we did see retrograde transport in the very same cells when a second dye (green) was later injected.
Results from other laboratories investigating peripheral nerve injury
22 are in line with our hypothesis of a preserved/enhanced axonal transport after injury. For example, retrograde axonal transport of growth factors is accelerated after injury and this increased trafficking is not only restricted to the growth factor protein family but applies more generally to axonal transport of different substances.
23,24 Similar dynamics of axonal transport are actually also seen after nerve crush damage in the central nervous system.
3,25 However, labeled biologic or injected artificial tracers did not cross the lesion site. Traumatic axonal injury is even used as a “stop-crush” model where the interrupted axonal transport is used as a tool to quantify the kinetics of accumulating substances.
26 However, on the other hand, it has been demonstrated that different substances are retrogradely transported by different mechanisms.
The fluorescent dyes used in the present study are basically nanoparticles. It might be that such nanoscale beads are transported by different mechanisms—maybe independent of the microtubules/axonal cytoskeleton which is already severely deranged during the first days after ONC and which causes blockage of other tracers. This assumption of different ways/kinetics of axonal transport is supported by data showing that, for example, modifications of the horseradish peroxidase lead to different trafficking kinetics.
27 We therefore propose that early after ONC a window is open for retrograde (nonclassical) axonal transport, but this window is not only temporally limited but also restricted to selected chemical/particular entities. Another possible explanation of retrograde RGCs labeling in a 2-day period is that the posttraumatic “transport window” is exclusively open for the fast transport mechanism. This would mean that we actually do not see the effects of acceleration/enhancement but a “cutoff” of the slow transport. Clearly, this issue requires further study.
However, it is now also widely accepted that traumatic axon injury is a progressive event and does not only lead to an immediate but also delayed axonal disconnection.
28 As this process is probably dependent on the characteristics and severity of the damage,
29 the functional disturbance may vary widely and the fact that we observed preserved axonal transport on day 2 to 4 post ONC may be a situation that is unique to our neurotrauma crush model.
In addition to the effect of a measurable transport early after ONC our data reveal a decreased/interrupted trafficking (with a minimum at approximately 2 weeks after lesion) and a partial recovery of axonal transport at 3 weeks post ONC. This temporal pattern of decrease and subsequent recovery is in good agreement with earlier data from our laboratory. Here we showed initial damage to be followed by repair after ONC with classical histologic/immunohistochemical approaches.
25 Also in that study indicators of axonal transport and integrity transiently decreased and then recovered at 3 weeks. Similar trends within this 3-week time window have been found with other parameters: visual function is lost transiently after a lesion but spontaneously improves within a few weeks,
30,31 cellular activity in the deafferented tectum is decreased after traumatic nerve injury but recovers within this period,
32 and structural repair of the axon has been shown with immunohistochemical techniques 3 weeks after ONC.
33 Especially these immunohistochemical studies provide interesting results regarding the recovery mechanisms of axonal transport. It was shown that the structure of neurofilament-H (an element of the axonal cytoskeleton) was partially repaired on day 21 after ONC and that this was accompanied by a recovery of axonal transport. Therefore, the temporal profile of RGC labeling in the present study confirms the hypothesis that there is internal axon repair or regeneration of the damaged cytoskeleton within 3 weeks after ONC which is why axonal transport can recover as well.
Another interesting aspect of our study is that this recovery of axonal transport was not associated with increased neuronal survival; rather the contrary was seen: the extent of delayed cell death was similar (moderate crush group) or even higher (mild crush group) in the population of cells with recovered trafficking (i.e., which were double labeled). Although ongoing neuronal death has been detected already in previous studies,
4,19,20 the precise temporal pattern of cell loss was not systematically analyzed. The current results suggest that restoration of axonal transport does not prevent delayed neuronal death after diffuse traumatic brain injury but rather accelerates it. This might point toward the existence of a cell death signal in the deafferented target, but such a potential association between restoration of nerve function and cell degeneration clearly needs to be explored further.
A secondary cell death scenario has already been described previously with experiments using classic labeling/staining techniques. They found signs of “secondary degeneration” (i.e., delayed death) as a function of the severity of the primary insult.
34
On the one hand, the increased cell death rate during the period of restoration of axonal transport is rather surprising as it is generally assumed that retrograde transport of growth factors (e.g., brain-derived neurotrophic factor [BDNF]) may be beneficial to prevent cell death caused by apoptosis.
35,36 On the other hand, there are multiple sources of growth factors (e.g., Müller cells
37 ) and it is not exactly clear whether the RGCs die from a lack of trophic supply after ONC. The existing data actually do not seem to support the hypothesis that a preserved axonal transport or even the existence of the axon is a condition sine qua non for neuron survival. Experiments using paclitaxel (Taxol; Sigma Aldrich Chemie GmbH, Munich, Germany), an axonal transport inhibitor, showed that this changed the state of the deranged cytoskeleton for the worse and deteriorated axonal trafficking.
25,38 Nevertheless, neurons appeared to have survived even better when this compound was applied.
38 Furthermore, studies of the pathophysiology of optic nerve cut report that a certain percentage (10% to 15%
39 ) of RGCs survive, even though no retrograde supply via axonal transport is possible. In a recent elegant study, Leung et al.
40 followed the slow degenerating process of RGCs after mild ONC. They showed that although neurons lost their axons within 3 weeks post lesion, some of these cells were still alive more than 2 months later. In summary, the data of all these studies are compatible with the concept that while axonal damage may trigger retrograde cell death, repair of axonal transport does not support neuronal survival. But, to the contrary, restored trafficking may actually accelerate cell death.
We would like to point out that the current protocol was designed to analyze the dynamics of labeling/axonal transport post ONC in vivo. It was not designed to compare the absolute number of labeled (surviving) cells, as this depends on the individual retinal region chosen for imaging. However, the fact that after mild crush we find 190 red cells labeled and in the moderate crush group we find only 97 red RGCs 2 days after the first dye injection is in line with the assumption that after a more severe lesion axonal transport is more impaired or, alternatively, impaired in more neurons.
In summary, our study demonstrates the usefulness of in vivo tracing studies with two different dyes and subsequent in vivo ICON imaging as a sensitive tool to relate cell death with the axon transport ability of individual cells. In our model of diffuse and incomplete nerve trauma axonal transport recovers within 3 weeks after the impact. We take this as a sign of intrinsic axon repair.
Our second main finding is a delayed, secondary wave of neuronal death which not only occurred despite the axonal transport repair but actually perhaps because of it. Thus, recovered axon transport does not enhance neuronal survival but, to the contrary, actually increases cell death. This possibility of an axon transport-dependent cell death suggests the existence of a delayed cell death signal (“kiss of death”) which may be transported retrogradely from the deafferented target.
41 Uncovering such factors will likely lead not only to new insights into mechanisms of neuronal death but also provide new clues in our search for means to enhance restoration (recovery) of vision.
Supported by the Ministry of Education and Research (BMBF, 01ZZ0407) and Sibylle Assmus Foundation (PH-N).
The authors thank Steffi Matzke for administrative assistance and Uta Werner for excellent technical assistance.