The prion protein is not required for peripheral nerve repair after crush injury

The cellular prion protein (PrP) is essential to the long-term maintenance of myelin sheaths in peripheral nerves. PrP activates the adhesion G-protein coupled receptor Adgrg6 on Schwann cells and initiates a pro-myelination cascade of molecular signals. Because Adgrg6 is crucial for peripheral myelin development and regeneration after nerve injury, we investigated the role of PrP in peripheral nerve repair. We performed experimental sciatic nerve crush injuries in co-isogenic wild-type and PrP-deficient mice, and examined peripheral nerve repair processes. Generation of repair Schwann cells, macrophage recruitment and remyelination were similar in PrP-deficient and wild-type mice. We conclude that PrP is dispensable for sciatic nerve regeneration after crush injury. Adgrg6 may sustain its function in peripheral nerve repair independently of its activation by PrP.


Introduction
The reciprocal interaction between axons, Schwann cells and the extracellular matrix is crucial for development, maintenance and repair of the peripheral nervous system (PNS) (1,2). The adhesion G-protein coupled receptor Adgrg6 (formerly called Gpr126) was shown to be required for Schwann cell development (3), peripheral nerve repair (4,5) and possibly myelin maintenance (4). Three natural ligands for Adgrg6 have been described: collagen IV (6), laminin-211 (7) and the prion protein (PrP) (8). Collagen IV and laminins are essential components of the Schwann cell basal lamina, which plays an important role in developmental myelination and remyelination after injury (9)(10)(11). PrP is a glycophosphoinositol-anchored glycoprotein highly expressed in the nervous system (12) and mainly known for its role in prion diseases (13,14). The development of chronic peripheral demyelinating neuropathy in PrP knockout mice has led to the identification of cellular PrP as necessary for myelin maintenance (15,16). This finding raises the question whether PrP may also play a role in PNS repair. Here, we hypothesized that PrP may represent one of the ligands required to activate Adgrg6 after nerve injury. If so, PrP-deficient mice might develop similar defects in nerve regeneration as were reported in Adgrg6 knockout mice. To address this hypothesis, we performed nerve crush injuries in female Prnp knockout mice (Prnp ZH3/ZH3 ) and compared the morphological and molecular stages of repair to wild type (WT) mice. Contrary to our expectations, we detected no differences in peripheral nerve repair between Prnp ZH3/ZH3 and WT mice. These findings suggest that PrP is dispensable for peripheral nerve repair.

Mice
Breeding and maintenance of mice was performed in laboratory animal facilities (optimal hygienic conditions)at the University Hospital Zurich. Mice were housed in groups of 3 with unlimited access to food and water. Animal care and experiments were performed in accordance with the Swiss Animal Protection Law. All experiments were approved by the Veterinary Office of the Canton of Zurich (permit ZH168/2019).

Nerve crush surgery
Surgery was performed on the right hind limb of two-month-old female WT and Prnp ZH3/ZH3 mice as described previously (17,18). Briefly, mice were injected with buprenorphinum (Temgesic, 0.1 mg/kg of bodyweight) prior to surgery. The surgery was performed under isoflurane anaesthesia. Fur was removed with an electric trimmer and the sciatic nerve was exposed at the height of the hip by making a small incision. The sciatic nerve was exposed by blunt dissection and crushed by squeezing firmly for 30 seconds with a Dumont S&T JF-5 forceps (FST tools) at the height of the sciatic notch. The tissue was repositioned, and the wound was sealed with a suture. Mice were administered buprenorphinum for analgesia during the first 2 postoperative days. Sciatic nerves from crushed and contralateral side were harvested at 5, 10, 12, 16 and 30 days post crush.

Nerve harvesting
Mice were sacrificed by cervical dislocation in deep anaesthesia. Sciatic nerves were embedded in the required fixation solution for morphological analysis or frozen in liquid nitrogen for protein analysis. After crush injury, the sciatic nerve distal to the crush site was divided in 3 parts (see Supplementary Fig. 1). A segment distal to the crush site (3 mm) was used for electron microscopy (EM). The distal side of this segment was embedded facing the front of the block face for sectioning. Sections from 2 mm distal to the crush site were analysed by EM. The remaining sciatic nerve was cut in half and the more proximal segment was processed for immunofluorescence (IF) as described below. the nerve segment was positioned with the proximal side facing the front of the block face for sectioning. Thereby, the EM and IF images derived from a similar distance from the crush side. The most distal segment of the sciatic nerve was frozen in liquid nitrogen and used for protein analysis. On the contralateral side, the corresponding segments were collected in the same manner.

Morphological analysis by EM
Sciatic nerves were immersed in 4% glutaraldehyde in 0.1 M sodium phosphate buffer pH 7.4 immediately after dissection and incubated at 4 °C overnight. Tissue was embedded in Epon using standard procedures. Further steps were performed as described (18). Briefly, 99 nm sections were collected on ITO coverslips (Optics Balzers). Imaging for EM reconstruction of the entire sciatic nerve section was performed with either a Carl Zeiss Gemini Leo 1530 FEG or Carl Zeiss Merlin FEG scanning electron microscope attached to Atlas modules (Carl Zeiss).
Adobe Photoshop CS5 was used for image analysis. The g-ratio corresponds to the ratio between axon diameter and fibre diameter. The axon diameter was derived from the axon area. The myelin thickness was measured at two different locations of the myelin ring. The average myelin thickness was added twice to the axon diameter to obtain the fibre diameter.
For g-ratio quantification, three different locations on the cross section were chosen and at least 100 fibres per sample were analysed. The number of intact appearing myelin profiles, remyelinated fibres and the area covered by myelin debris was assessed manually on the entire cross section. The investigator was blinded as to the genotype of the mice for all analyses.

Immunofluorescence (IF)
Sciatic nerves were isolated as described above. The tissue was fixed in 4% paraformaldehyde overnight at 4°C. Then, the tissue was incubated in 30% sucrose solution and frozen in OCT compound. Cross-sections of 8 µM thickness were cut at the cryostat. After drying at room temperature (RT), sections were incubated in blocking buffer (10% normal goat serum, 0.5% bovine serum albumin (BSA), 0.3% Triton X-100 in PBS) for 1 h at RT. Blocking buffer was

Experimental design and statistical analysis
Statistical analysis was performed with GraphPad Prism software (version 8.4.2). We assumed normal distribution and equal variances of data but did not formally test this assumption due to small n values. Unpaired two-tailed t-test was used for comparison of two groups. For comparison of three or more groups, two-way ANOVA followed by Sidak's multiple comparison test was used and multiplicity adjusted p-values were reported. P-values below 0.05 were considered statistically significant. P-values are indicated in graphs as *: p < 0.05. ns: not significant, p > 0.05. Error bars in graphs show SEM. No samples or data were omitted during the analyses. R (version 3.5.2) was used to generate the scatterplot visualizing the g-ratio analysis.

PrP is not required for demyelination, repair Schwann cell generation and proliferation following nerve injury
To study the role of PrP in peripheral nerve regeneration, we performed sciatic nerve crush injuries in female Prnp ZH3/ZH3 and WT mice at 2 months of age and investigated the morphological and molecular stages of peripheral nerve repair. We crushed the sciatic nerve at the sciatic notch, and harvested the nerve segments distal to the injury site at 5, 10, 12, 16 and 30 days post crush (d.p.c). Additionally, we collected the uninjured contralateral sciatic nerves as controls. To ensure comparability between the groups, the sections harvested for the various analyses were taken from the same distance from the crush site as depicted in supplementary figure 1.
First, we assessed if Schwann cells in Prnp ZH3/ZH3 mice were able to properly transdifferentiate to repair Schwann cells. We investigated c-Jun as a molecular marker of repair Schwann cells (20) by immunofluorescence (IF) and western blotting (Fig. 1a-d). At 5 d.p.c. a strong upregulation of c-Jun was detected both in WT and Prnp ZH3/ZH3 mice. c-Jun levels decreased at later time points post injury. In uninjured contralateral nerves, c-Jun protein levels were low and nuclear c-Jun staining in IF was very faint. We did not detect any significant differences in c-Jun levels between WT and Prnp ZH3/ZH3 mice at 5, 10, 16 and 30 d.p.c.
GFAP, another marker of repair Schwann cells, peaked one week later than c-Jun at 10-16 d.p.c (Fig. 1e,f). Quantification of GFAP upregulation after crush (Fig. 1f) revealed lower GFAP upregulation in Prnp ZH3/ZH3 mice compared to WT mice, which reached the threshold for statistical significance at 16 d.p.c. (p = 0.0425). This differential upregulation might be related to the fact that Prnp ZH3/ZH3 mice had on average a two-fold higher baseline GFAP level when compared to WT mice (Fig. 1g, relative increase to WT GFAP level as quantified from western blot: WT 1.00 ± 0.17 (6); Prnp ZH3/ZH3 2.05 ± 0.09 (6); mean ± SEM (n); p = 0.0003, unpaired ttest). Compatible with this explanation, the absolute GFAP protein levels after nerve crush as assessed by western blotting were similar in WT and Prnp ZH3/ZH3 mice at all investigated time points (Fig. 1h). In conclusion, Prnp ZH3/ZH3 mice required less GFAP upregulation to reach WT levels after nerve crush due to their higher baseline levels.
Ki67 staining in crushed sciatic nerves showed that proliferation was highest at 5 d.p.c. both in WT and Prnp ZH3/ZH3 mice, and no significant difference in the proliferation index was detected between WT and Prnp ZH3/ZH3 mice at any of the time points investigated (Fig. 2a,b).
In the early stages after nerve crush, Schwann cells are the main effectors of myelin breakdown and digestion (21). Using electron microscopy (EM) we investigated the extent of demyelination by counting the number of myelin profiles that still appeared intact at 5 d.p.c. (Fig. 2c). The number of intact appearing myelin profiles per sciatic nerve cross section was not significantly altered in Prnp ZH3/ZH3 mice when compared to WT mice (WT 191 ± 38 (3); Prnp ZH3/ZH3 244 ± 68 (3); mean ± SEM (n); p = 0.5277; unpaired t-test), indicating that the initial myelin breakdown is neither slowed nor accelerated in the absence of PrP.
Collectively, these results suggested that PrP is not required for upregulation of repair Schwann cell markers, proliferation and early myelin breakdown after nerve crush injury.
PrP is not required for macrophage recruitment after peripheral nerve injury Blood derived macrophages are recruited to the injured nerve via chemokines secreted by Schwann cells and mesenchymal cells (22). Together with resident endoneurial macrophages, they contribute to debris clearance and repair after injury (23) . We performed IF for the macrophage marker CD68 to investigate if PrP is playing a role in macrophage recruitment ( Fig. 3a-b). While uninjured nerves were basically devoid of CD68-positive cells (Fig. 3b), crushed sciatic nerves were infiltrated by numerous macrophages (Fig. 3a). In both genotypes, macrophages were rounded and displayed a phagocytic, foamy appearance. We quantified the area covered by CD68-positive cells in relation to the S100-positive area (Fig. 3c).
Macrophage accumulation in the endoneurium peaked at 10 d.p.c. in both genotypes, which is consistent with the time course of macrophage infiltration described in the literature (23,24).
No differences in macrophage levels were detected between WT and Prnp ZH3/ZH3 sciatic nerves in this analysis, suggesting that the lack of PrP does not affect injury-induced macrophage recruitment to the peripheral nerves.

Discussion
Adgrg6 is expressed in Schwann cells and was shown to have autonomous and nonautonomous functions in peripheral nerve repair (4,5). PrP is an agonist of Adgrg6 on Schwann cells (8). We therefore hypothesized that the lack of PrP causes similar defects in peripheral nerve repair as the knockout of Adgrg6. To evaluate our hypothesis, we performed sciatic nerve crush injuries in female WT and Prnp ZH3/ZH3 mice and investigated the stages of repair by EM, IF and western blotting. However, we found that peripheral nerve repair was not altered in Prnp ZH3/ZH3 mice when compared to WT mice. The extent and time point of c-Jun upregulation, demyelination, Schwann cell proliferation, macrophage recruitment and remyelination were similar in Prnp ZH3/ZH3 mice when compared to WT mice at all investigated time points. Our results thus suggest that PrP is dispensable for peripheral nerve repair. This may be either because PrP plays no major role in the peripheral nerve repair processes, or because other molecules can compensate for the lack of PrP. Alternatively, Adgrg6 may exhibit a basal constitutive activity sufficient to guarantee nerve repair in the absence of any agonist.
Regarding the first point, PrP is dispensable for normal development of the PNS and only required for long-term maintenance of the myelin sheath (15). Many of the repair processes after nerve injury recapitulate developmental processes (25) and it is possible that PrP is simply not required for peripheral nerve repair. As to the second point, PrP is thought to function in myelin maintenance via activation of Adgrg6 on Schwann cells. Adgrg6 is involved in peripheral nerve repair, as was suggested by deficits in macrophage recruitment and remyelination after peripheral nerve injury in conditional Adgrg6 knockout mice. Thus, Adgrg6 must sustain its function in nerve repair independent of PrP, for example via activation by other ligands such as collagen IV (6) and laminin-211 (7).
In a further possible scenario, PrP might play contrasting roles in different cells types during the repair processes. Theoretically, the effects of PrP in different cell types may cancel each other out in a transgenic mouse with global PrP knockout as was used in the present study.
However, considering that the peripheral nerve repair processes are regulated by a highly orchestrated, balanced interaction between many cell types, we feel that this is an unlikely scenario.
We have previously described that Prnp ZH3/ZH3 mice exhibit an early upregulation of repair Schwann cell markers in sciatic nerves, which probably represents an early sign of the peripheral nerve disease (26). In the present study, we confirmed that the 2-3 months old Prnp ZH3/ZH3 mice undergoing crush injury showed a two-fold upregulation of GFAP in their uninjured sciatic nerves when compared to WT mice. The absolute GFAP levels after crush injury were similar in both genotypes, and the elevated baseline GFAP level in Prnp ZH3/ZH3 mice did not interfere with regeneration. Whereas defective Schwann cell proliferation and delayed regeneration after nerve crush injury have been described in GFAP knockout mice (27,28), the effect of GFAP overexpression on PNS development, maintenance or repair has not been investigated (29,30). GFAP gain-of-function mutations cause Alexander disease, a rare astrogliopathy characterized by accumulation of GFAP in the central nervous system (31).
Development of a peripheral demyelinating neuropathy was reported in an 8-year-old patient with Alexander disease (32), but apart from this case, there is currently no compelling evidence that GFAP accumulation or overexpression contributes to or causes PNS degeneration. The finding that sciatic nerve repair is not impaired in Prnp ZH3/ZH3 mice despite increased baseline GFAP levels might be interpreted as an indication that GFAP overexpression is compatible with peripheral nerve regeneration.
In conclusion, the molecular and morphological analyses presented here consistently show that PrP is dispensable for peripheral nerve repair. While this finding disproves the hypothesis that PrP is required to activate Adgrg6-mediated macrophage recruitment and remyelination after nerve crush injury, it does not rule out an ancillary, non-essential role for such interactions.
Furthermore, it is possible that additional, hitherto undiscovered Adgrg6 ligands exist. The importance of such ligands to peripheral-nerve repair may vary in different species. Therefore, the negative results reported here should not discourage researchers from exploring Adgrg6 activation as a possible adjuvant therapy for peripheral nerve repair.     Straight lines in (c) show linear regression. No significant difference in mean g-ratio (d) was detected between genotypes (unpaired t-test, p > 0.05). e) Quantification of area covered by myelin debris at 30 d.p.c. showed no significant difference between WT (n = 3) and Prnp ZH3/ZH3 (n = 3) mice (unpaired t-test, p > 0.05). n.s. = not significant.

Supplementary Figures
Supplementary Figure S1: Nerve harvest after crush injury. The sciatic nerve was crushed using a forceps at the sciatic notch. For harvesting, the nerve was cut at 3 mm distal to the crush side. The proximal segment was embedded for electron microscopy (EM), the distal segment was used for immunofluorescence (IF) and western blotting.