A, Facial nerve schematic. B, Facial nerve schematic under natural light. C, Facial nerve schematic under UV light. D, Histological analysis of neuromuscular junctions of the levator labii superioris muscle. E, Staining of motor end plates with α-bungarotoxin. F, Staining of native facial nerve sections. Facial nerve branches: T indicates temporal; Z, zygomatic; B, buccal; M, marginal mandibular.
A, Schematic of the crush sites on the buccal and marginal mandibular branches (arrowheads). B, Initial crush (30 seconds) to predegenerate the distal nerve. C, Incomplete predegeneration after 2 weeks, second crush injury (30 seconds). D, Macroscopic regeneration 4 weeks after second crush injury.
A, Surgical model, suture site marked schematically. B, After nerve transection, the proximal and distal stumps were tied back. C, Direct repair after 2-week predegeneration of the distal stump. D, After 4 weeks of axon regeneration into the distal stump.
A, Surgical model: CFNG from right buccal to left buccal and marginal mandibular branches. Because of the strong fluorescence of the brain through the skull (B), the graft was imaged on a blue background (C and D).
A, Predegenerated CFNG. A1: White arrowhead: proximal end. A2: recipient branches: left buccal and marginal mandibular branches (yellow arrowhead); A3: Donor branch: right buccal branch (red arrowhead). B, Regeneration after 2 weeks. B1: CFNG. B2: Recipient branches. B3: Donor branch. White arrowhead: frontal view of advancing regeneration front. Red arrowhead: lateral view of regeneration front. C, Regeneration after 4 weeks. C1: CFNG. C2: Recipient branches. C3: Donor branch. White arrowhead: frontal view of advancing regeneration front. Red arrowhead: lateral view of regeneration front. D, Regeneration after 8 weeks. D1: CFNG. D2: Recipient branches. D3: Donor branch. White arrowhead: frontal view of advancing regeneration front (5B to 5D). Red arrowhead: lateral view of regeneration front. Asterisk: regeneration front reaching the recipient side after 8 weeks of regeneration.
Placheta E, Wood MD, Lafontaine C, Frey M, Gordon T, Borschel GH. Macroscopic In Vivo Imaging of Facial Nerve Regeneration in Thy1-GFP Rats. JAMA Facial Plast Surg. 2015;17(1):8–15. doi:10.1001/jamafacial.2014.617
Facial nerve injury leads to severe functional and aesthetic deficits. The transgenic Thy1-GFP rat is a new model for facial nerve injury and reconstruction research that will help improve clinical outcomes through translational facial nerve injury research.
To determine whether serial in vivo imaging of nerve regeneration in the transgenic rat model is possible, facial nerve regeneration was imaged under the main paradigms of facial nerve injury and reconstruction.
Design, Setting, and Participants
Fifteen male Thy1-GFP rats, which express green fluorescent protein (GFP) in their neural structures, were divided into 3 groups in the laboratory: crush-injury, direct repair, and cross-face nerve grafting (30-mm graft length). The distal nerve stump or nerve graft was predegenerated for 2 weeks. The facial nerve of the transgenic rats was serially imaged at the time of operation and after 2, 4, and 8 weeks of regeneration. The imaging was performed under a GFP-MDS-96/BN excitation stand (BLS Ltd).
Intervention or Exposure
Facial nerve injury.
Main Outcome and Measure
Optical fluorescence of regenerating facial nerve axons.
Serial in vivo imaging of the regeneration of GFP-positive axons in the Thy1-GFP rat model is possible. All animals survived the short imaging procedures well, and nerve regeneration was followed over clinically relevant distances. The predegeneration of the distal nerve stump or the cross-face nerve graft was, however, necessary to image the regeneration front at early time points. Crush injury was not suitable to sufficiently predegenerate the nerve (and to allow for degradation of the GFP through Wallerian degeneration). After direct repair, axons regenerated over the coaptation site in between 2 and 4 weeks. The GFP-positive nerve fibers reached the distal end of the 30-mm–long cross-face nervegrafts after 4 to 8 weeks of regeneration.
Conclusions and Relevance
The time course of facial nerve regeneration was studied by serial in vivo imaging in the transgenic rat model. Nerve regeneration was followed over clinically relevant distances in a small number of experimental animals, as they were subsequently imaged at multiple time points. The Thy1-GFP rat model will help improve clinical outcomes of facial reanimation surgery through improving the knowledge of facial nerve regeneration after surgical procedures.
Level of Evidence
Facial nerve injury leads to functional and aesthetic deficits, which include impairment of speech, eating, and emotional expression.1,2 Deficits in eyelid function result in lagophthalmus, reduced corneal protection, hyperlacrimation, keratitis, and, in severe cases, impairment of vision.3Facial palsy of all etiologies affects 15 to 35 per 100 000 people per year.1,4,5 In (partially) recovered facial palsy, aberrant facial nerve regeneration may lead to severe sequelae, such as involuntary synkinetic movements, which deteriorate facial function.6- 8 In severe cases of facial nerve deficits, facial reanimation surgery is performed when there is no potential for (sufficient) spontaneous facial nerve regeneration.9 Dynamic facial reanimation includes nerve grafts that innervate the denervated mimetic muscles or free muscle transplants.10- 12 Given the complexity of facial reanimation, an interdisciplinary team is required for optimal patient management and for the success of multistage surgical concepts.11,13,14
Different facial nerve injury15- 17 and reconstruction models18- 22 are established in small (murine model23,24) and large animals, such as rats,15- 22 rabbits,25- 27 dogs,28 and primates.29 Outcome measures of facial nerve function and recovery include histological30- 35 and functional metrics.36- 40 In recent years, transgenic animals that express fluorescent proteins under the neuron-specific Thy1 promotor41- 43 have broadened the spectrum of metrics in the investigation of nerve regeneration. Transgenic mouse models that express green fluorescent protein (GFP) in their neurons provide detailed imaging options of nerve regeneration, which include in vivo imaging and high-resolution imaging of the neuromuscular junction.23,43,44 Transgenic mice that express spectral variants of fluorescent proteins in their neural structures are used to study different subsets of neurons.23,43,45 In double-transgenic mice, the interaction of Schwann cells and regenerating axons can be imaged in vivo.44,46,47 Recently, a transgenic rat that expresses GFP in neural structures through the Thy1 promoter was introduced.48- 50 The Thy1-GFP rat is raised on a Sprague-Dawley background and provides advantages compared with the murine model owing to its larger size and suitability for behavioral analysis.48- 50 Kemp et al50 showed that the neural architecture and functional recovery of Thy1-GFP rats do not differ from wild-type Sprague-Dawley rats. In this rat, surgical models of facial nerve crush injury48 and sciatic nerve reconstruction (with clinically relevant nerve graft lengths)49 demonstrated that nerve regeneration can be visualized in great quality in vivo and also in the histological analysis.
In vivo imaging of nerve regeneration in transgenic animals that express fluorescent proteins in their neural structures was performed in mouse44,47 and rat models.48,49 Most studies performed the in vivo imaging during nonsurvival interventions.48,49 Previously, the regeneration of GFP-positive axons was imaged transcutaneously in the same animals at multiple time points.47,51 The aim of the current study is to (1) characterize the serial in vivo imaging of nerve regeneration in the Thy1-GFP rat model and (2) to study the regeneration of GFP-positive axons in a facial palsy model for the main paradigms of facial nerve injury and repair. Serial in vivo imaging of nerve regeneration will provide insight into the time course of facial nerve regeneration after crush injury, direct repair, and cross-face nerve grafting, while reducing the number of experimental animal required for the investigation.
Fifteen male transgenic Thy1-GFP rats,48- 50 which express GFP in their neural structures, were used in this study (Figure 1). All experimental animals weighed 300 to 400 g. The rats were divided into 3 groups with 5 animals per group: (1) crush injury, (2) direct repair, and (3) cross-face nerve grafting. The animal care and all interventions were performed according to animal care and safety guidelines. The study protocol was approved by the animal committee of The Hospital for Sick Children, Toronto, Ontario, Canada.
All surgical procedures were performed with aseptic technique with the animals under 2.5% isoflurane gas anesthesia. Standard microsurgical technique was used, and all procedures were performed under the operative microscope. Meloxicam (1.0 mg/kg subcutaneously) was administered as intraoperative analgesic. The surgical procedures are described in the following Methods subsections. Before returning to the animal housing facility, the rats recovered from anesthesia on a warming pad.
The GFP-positive axons of the transgenic Thy1-GFP rats were macroscopically imaged under a GFP-MDS-96/BN excitation stand (BLS Ltd), which is a camera stand with 8 light sources. The images were captured under natural and UV light (Figure 1).
The imaging was performed under a short 2.5% isoflurane gas anesthesia using aseptic technique with Meloxicam as analgesic, as described in the previous subsection, during the initial surgery and after 2, 4, and 8 weeks of regeneration.
In the crush-injury group, the left facial nerve was exposed through a preauricular incision that was extended toward the neck to expose the whole cheek. The parotid gland was elevated to allow for imaging of the facial nerve branches. The buccal and marginal mandibular branches were crushed 10 mm distal to the division of the main facial trunk for 30 seconds with a jeweler’s forceps52 to predegenerate the distal part of the nerve branches (Figure 2A). The crush site was marked with a single epineural 10-0 Ethilon suture. After irrigation, the skin was closed with interrupted 5-0 Vicryl sutures. After 2 weeks, the facial nerve branches were reexposed as described and imaged in vivo. After the initial 2-week predegeneration of the distal part of the buccal and marginal mandibular nerve branches, a second crush was administered to the marked crush site. The skin was closed as described herein.
In the cut-and-repair group, the left facial nerve was exposed as described in the previous subsection. The buccal and marginal mandibular branches were cut 10 mm distal to the division of the main facial nerve trunk (Figure 3A). The distal and proximal nerve stumps of each nerve branch were sutured to the muscle with 5-0 single silk sutures and subsequently imaged. After irrigation, the skin was closed with interrupted 5-0 Vicryl sutures. After 2 weeks, the buccal and marginal mandibular branches were reexposed. After imaging, the distal and proximal stumps of the respective nerves were coapted end-to-end with interrupted epineural 10-0 Ethilon sutures. The skin was closed as described herein.
The right common peroneal (fibular) nerve was used as the cross-face nerve graft, which was predegenerated for 2 weeks prior to the cross-face nerve grafting procedure. The common peroneal nerve was exposed through a gluteal muscle-splitting approach, cut proximally (just distally to the trifurcation of the sciatic nerve) and sutured to the muscle with one 5-0 silk suture. The wound was irrigated with 0.9% saline and the muscle and skin closed with interrupted 5-0 Vicryl sutures. After 2 weeks, the common peroneal nerve was exposed as described herein as a 30-mm–long graft (the orientation of the nerve graft was marked). Through a bicoronary incision, both facial nerves were exposed. The right buccal branch (donor nerve branch) was dissected and cut as distally as possible. The left buccal and marginal mandibular branches (recipient branches) were dissected and cut proximal to their branching division. The nerve graft was placed distoproximally over the forehead (Figure 4A). The nerve coaptations were performed with interrupted epineural 10-0 Ethilon sutures. At the end of the procedure, the facial nerve branches and cross-face nerve graft were imaged. After irrigation, the skin was closed with interrupted 5-0 Vicryl sutures.
All animals survived the procedures. In 1 rat, a seroma in the face was drained on postoperative day 2 after the cross-face nerve grafting. No major complications were observed postoperatively.
The facial nerve of Thy1-GFP rats was exposed (Figure 1A-C) and imaged under natural light (Figure 1B) and blue light (Figure 1C). After removal of the parotid gland, the main trunk of the facial nerve and the division into the main branches could be imaged without further dissection (Figure 1A-C). The fluorescence captured in the macroscopic images corresponded with the fluorescence seen under the fluorescent microscope, for example, in muscle sections (Musculus levator labii superioris, Figure 1D), after α-Bungarotoxin staining of muscle end plates (Figure 1E) or in nerve sections (Figure 1F).
The reexposure of the nerves for serial in vivo imaging was increasingly challenging because of scar tissue formation, which interfered with the fluorescent signal and had to be removed.
The left buccal and marginal mandibular branches of Thy1-GFP rats were crushed for 30 seconds to predegenerate the distal stump of the facial nerve branches to allow for imaging of the regenerative front of GFP-positive axons after crush injury (Figure 2B). After 2 weeks, the intensity of the fluorescence was reduced but had not entirely vanished (Figure 2C) because the GFP-positive axons regenerated across the crush site. This demonstrates that the regenerating axons in the distal nerve stump preclude complete predegeneration of the distal nerve. At 4 weeks after the second crush injury (Figure 2D), the GFP signal had macroscopically returned to the intensity seen preoperatively, as the axons regenerated across the crush site.
The 2-week predegeneration of facial nerve branches after tying the distal and proximal stumps to the muscle (Figure 3B) cleared the GFP out of the distal nerve stumps (Figure 3C). After 4 weeks, GFP-positive axons regenerated across the repair site. At this time point, the difference of the intensity of fluorescence of the nerve proximal and distal to the repair site was macroscopically visible (Figure 3D).
The regeneration of GFP-positive axons across the (predegenerated) cross-face common peroneal nerve graft (Figure 4A) was imaged over sterile blue fabric (Figure 4C and D) because the fluorescence of the brain through the skull bone was too strong (Figure 4B) to allow for imaging of the regenerative front.
The common peroneal nerve was harvested as a 30-mm–long nerve graft (Figure 4A) after a 2-week predegeneration period. At the time of cross-face nerve grafting (Figure 5A), little or no GFP was visible in the graft, having been cleared by Wallerian degeneration of the distal common peroneal nerve over the 14-day period of denervation. The difference in fluorescence of the predegenerated nerve graft and the freshly cut facial nerve branches, on the donor and recipient side, is clearly visible (Figure 5A1-A3). After 2 weeks of regeneration, the axons merely crossed the proximal coaptation site of the cross-face nerve graft (Figure 5B2 and B3). In the lateral view of the left (recipient) side, the fluorescent signal of the facial nerve branches had disappeared as the denervated branches underwent Wallerian degeneration for 2 weeks (Figure 5B2). The GFP-positive axons regenerated across the cross-face graft and reached the distal coaptation site within 4 to 8 weeks (Figure 5C1 and C3 and Figure 5D1-D3). At 8 weeks of regeneration after placing the CFNG between the right buccal to the left buccal and marginal mandibular branches, the intensity of the fluorescent signal of the cross-face nerve graft had macroscopically reached the level of the donor nerve (Figure 5D1 and D3).
In this study, we demonstrated serial macroscopic in vivo imaging of facial nerve regeneration after 3 paradigms of facial nerve injury and repair: crush injury, direct nerve repair, and nerve grafting. Serial in vivo imaging provides an additional outcome measurement of facial nerve regeneration and provides evidence for the advantage of using the transgenic Thy1-GFP rat model,48- 50 which expresses GFP in its neural structures, for analysis of the outcomes of nerve injury.
Serial in vivo imaging provides insight into the time course of nerve regeneration and can reduce the number of experimental animals used because each animal can be studied at multiple time points.49,53 The Thy1-GFP rat has previously been used to study crush injury of the facial nerve.48 The time course of nerve regeneration in the Thy1-GFP rat has been further established in sciatic nerve crush, direct repair, and nerve grafting.49 These studies did not provide data on serial in vivo imaging because the animals were killed and perfused before imaging. Serial life imaging of transgenic mice that express fluorescent proteins in their sensory and motor axons has been applied to image the tibial nerve after crush injury.51 In similar murine models, the rate of axonal regeneration of sensory nerves, like the saphenous53 and sural nerves,47 was imaged transcutaneously.
Pan et al53 described advantages and limitations of serial in vivo imaging of nerve regeneration in the transgenic mouse model: first, the ovoids,54,55 which persist in the distal stump during Wallerian degeneration,27,28 produce a slight fluorescent signal. Second, the different rate of regeneration within the axon population can be differentiated, which is often not possible with traditional techniques. And third, in whole mounted samples, the thin axons extending from the coaptation site, which are generally lost during dissection, can be imaged.53 We found that, consistent with the findings of Pan et al,53 the remaining fluorescence of the distal nerve stump due to ongoing Wallerian degeneration during the first 2 weeks after nerve injury made the imaging of early regenerated nerve difficult. To overcome the remaining fluorescence of the distal nerve stump or nerve graft, 2 different strategies are established: either predegeneration of the nerve, as described by Hayashi et al,47 or the use of nerve grafts harvested from GFP-negative littermates.49 We chose to predegenerate the autografts to allow for better comparison with the predegenerated distal nerve stumps in the crush and direct repair group.
Technical challenges of serial live imaging included increased scar formation after multiple reexposure of the nerve branches. The scar tissue had to be removed because it obscured the fluorescence and biased the imaging. Moore et al49 took images after perfusion because scar tissue and neovascularization obstruct optimal visualization of the regeneration front. We found that with careful dissection, the imaging results could be improved profoundly, although this increased the time required for the imaging procedure.
The time course of nerve regeneration was followed by macroscopic imaging with a portable GFP-MDS-96/BN excitation stand (BLS Ltd), which was used directly in the animal operating room during the interventions. In a study of axonal sprouting after end-to-side nerve repair, Hayashi et al47 used both macroscopic (transcutaneous) and microscopic imaging of GFP-positive axon regeneration in the transgenic mouse model. Transcutaneous macroscopic imaging was used to study the time course of sural nerve regeneration over several weeks. In a subpopulation of transgenic mice, which express GFP in under 10% of axons, detailed images of axon sprouting were taken at the coaptation site.47 Recently, Kale et al56 described detailed imaging of regeneration of GFP-positive axons in Thy1-GFP rats across a reverse end-to-side coaptation under the confocal microscope. The Thy1-GFP–positive rats were killed on postoperative days 3, 7, 10, and 35. Axons crossed the coaptation site as early as 7 days postoperatively.56
Macroscopic serial imaging was previously used in the transgenic mouse model to transcutaneously follow regeneration of sensory nerves and to determine the time course of regeneration.47 The technique proposed in the current study involves macroscopic serial imaging of nerve regeneration during short survival procedures (during which the nerves and nerve grafts are reexposed), assessing the outgrowth of GFP positive axons in the rat model. The advantage of this method is the assessment of nerve regeneration at the site of intervention and distally (not only analysis of the cutaneous branches). Compared with the detailed imaging of GFP-positive axons under the confocal microscope,56 the serial macroscopic imaging is performed in the same animal at different time points. While lacking the detailed resolution, it follows regeneration over greater, clinically relevant distances and provides an additional outcome measure to detailed microscopic imaging of axonal outgrowth.56
The advantages of the rat model over the murine model include its larger size and the ability to study longer nerve defects and more clinically relevant methods of nerve reconstruction.49 The rat model is also superior for behavioral analysis, as Kemp et al50 recently reviewed: (1) rats participate more readily in behavioral studies57- 59 and (2) results of behavioral studies in transgenic murine models were not reproducible in different laboratories.60- 62
The current surgical model for facial nerve crush injury (Figure 2), direct repair (Figure 3), and cross-face nerve grafting (Figure 4) was designed for in vivo imaging of nerve regeneration of the facial nerve branches63 (buccal and marginal mandibular branches), which supply the whisker pad.64 The selection of the buccal and marginal mandibular branches for this facial nerve injury model allows for subsequent functional analysis of whisker function21,22,36,37,65- 70 while preserving the eyelid function, and which reduces the strain on the experimental animal.
Matsuda et al19 described a cross-face nerve graft model in the rat, in which the nerve grafts are placed over the forehead through a subcutaneous tunnel. We adopted the placement of the cross-face nerve graft over the forehead but chose an open approach to allow for imaging of the regeneration front. The placement of the cross-face nerve graft ruled out techniques of functional analysis, which require the implantation of plates into the skull for head fixation.37,38
Apart from the functional analysis of whisker function, the Thy1-GFP model offers the option for all histological and electrophysiological outcome measurements of nerve regeneration,18,30 such as histomorphometry of myelinated axons,31,32 retrograde labeling of the neurons that regenerate their axons,33,34,71,72 immunohistochemical analysis, muscle histological analysis,35 and motor end plate staining.36 Serial in vivo imaging of GFP-positive axons enriches the established spectrum of outcome measurements, especially during early nerve regeneration and with regard to analysis of the time course of regeneration.
The findings of serial macroscopic in vivo imaging of nerve regeneration in the transgenic Thy1-GFP rat enhance the field of facial nerve research but are not limited to this application. The imaging will add to the analysis of nerve regeneration in other areas and with other surgical models.
Serial in vivo imaging in the Thy1-GFP rat provides real-time insight into early peripheral nerve regeneration. The time course of nerve regeneration over clinically relevant distances is can be studied in this transgenic rat model while reducing the number of required experimental animals because the rats are studied over multiple time points.
Accepted for Publication: June 3, 2014.
Corresponding Author: Gregory H. Borschel, MD, Division of Plastic and Reconstructive Surgery, The Hospital for Sick Children, 555 University Ave, 5408, Toronto, ON M5G 1X8, Canada (firstname.lastname@example.org).
Published Online: October 9, 2014. doi:10.1001/jamafacial.2014.617.
Author Contributions: Dr Placheta had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Placheta, Wood, Frey, Borschel.
Acquisition, analysis, or interpretation of data: Placheta, Wood, Lafontaine, Gordon, Borschel.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: Wood. Lafontaine, Frey, Borschel.
Obtained funding: Placheta, Frey, Borschel.
Administrative, technical, or material support: Placheta, Wood, Lafontaine.
Study supervision: Frey, Gordon, Borschel.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was funded by Dr Borschel’s grant from the Canadian Institutes for Health Research (CIHR), a scholarship from the International Society for Science in Plastic Surgery, Vienna, Austria (Dr Placheta), and the Hospital for Sick Children Foundation (Dr Gordon).
Role of the Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Previous Presentations: This study was presented at the European Association of Plastic Surgeons (EURAPS) Research Council Meeting; May 24-26, 2012; Munich; Germany; at the EURAPS Meeting; May 23-25, 2013; Antalya, Turkey; and at the 12th International Facial Nerve Symposium; June 28–July 1, 2013; Boston, Massachusetts.