LABORATORY INVESTIGATION J Neurosurg 130:184–196, 2019

Novel experimental surgical strategy to prevent traumatic neuroma formation by combining a 3D-printed Y-tube with an autograft

Anne Bolleboom, BSc,1,2 Godard C. W. de Ruiter, MD, PhD,3 J. Henk Coert, MD, PhD,4 Bastiaan Tuk,2 Jan C. Holstege, MD, PhD,1 and Johan W. van Neck, MD, PhD2

Departments of 1Neuroscience, and 2Plastic and Reconstructive Surgery, Erasmus University Medical Center, Rotterdam; 3Department of Neurosurgery, Medical Center Haaglanden, The Hague; and 4Department of Plastic and Reconstructive Surgery, Utrecht University Medical Center, Utrecht, The Netherlands

OBJECTIVE Traumatic neuromas may develop after nerve injury at the proximal nerve stump, which can lead to neuropathic pain. These neuromas are often resistant to therapy, and excision of the neuroma frequently leads to recur- rence. In this study, the authors present a novel surgical strategy to prevent neuroma formation based on the principle of centro-central anastomosis (CCA), but rather than directly connecting the nerve ends to an autograft, they created a loop using a 3D-printed polyethylene Y-shaped conduit with an autograft in the distal outlets. METHODS The 3D-printed Y-tube with autograft was investigated in a model of rat sciatic nerve transection in which the Y-tube was placed on the proximal sciatic nerve stump and a peroneal graft was placed between the distal outlets of the Y-tube to form a closed loop. This model was compared with a CCA model, in which a loop was created between the proximal tibial and peroneal nerves with a peroneal autograft. Additional control groups consisted of the closed Y-tube and the extended-arm Y-tube. Results were analyzed at 12 weeks of survival using nerve morphometry for the occur- rence of neuroma formation and axonal regeneration in plastic semi-thin sections. RESULTS Among the different surgical groups, the Y-tube with interposed autograft was the only model that did not result in neuroma formation at 12 weeks of survival. In addition, a 13% reduction in the number of myelinated axons regenerating through the interposed autograft was observed in the Y-tube with autograft model. In the CCA model, the authors also observed a decrease of 17% in the number of myelinated axons, but neuroma formation was present in this model. The closed Y-tube resulted in minimal nerve regeneration inside the tube together with extensive neuroma forma- tion before the entrance of the tube. The extended-arm Y-tube model clearly showed that the majority of the regenerat- ing axons merged into the Y-tube arm, which was connected to the autograft, leaving the extended plastic arm almost empty. CONCLUSIONS This pilot study shows that our novel 3D-printed Y-tube model with interposed autograft prevents neu- roma formation, making this a promising surgical tool for the management of traumatic neuromas. https://thejns.org/doi/abs/10.3171/2017.8.JNS17276 KEY WORDS neuroma; sciatic nerve; nerve regeneration; axon; rat; 3D printing; peripheral nerve

ormation of a traumatic neuroma after peripheral ability, which render them sensitive to ectopic stimuli,3,17,25 nerve injury is a clinical problem, often leading to ultimately leading to spontaneous neuropathic pain and long-term functional deficits. Data obtained from abnormal evoked pain.1,13,37 Despite a variety of treat- traumaF populations revealed that an estimated 2%–3% ments, such as simple excision of the neuroma, insertion of patients have experienced a major peripheral nerve in- of the proximal nerve stump into an adjacent anatomical jury28,33 and that approximately 3%–5% of these patients structure, or covering the nerve stump with a cap or epi- develop a painful neuroma at the site of injury.32 The axons neurium,24,35,36 there is presently no uniformly successful trapped in a neuroma show abnormal and increased excit- treatment for symptomatic neuromas.25 The resection of

ABBREVIATIONS CCA = centro-central anastomosis; PO = propylene oxide; PPD = paraphenylenediamine. SUBMITTED March 8, 2017. ACCEPTED August 8, 2017. INCLUDE WHEN CITING Published online February 9, 2018; DOI: 10.3171/2017.8.JNS17276.

184 J Neurosurg Volume 130 • January 2019 ©AANS 2019, except where prohibited by US copyright law

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

FIG. 1. Schematic presentations of the different types of customized 3D-engineered conduits: the open Y-tube (A), the closed Y-tube (B), and the extended-arm Y-tube (C). The dimensions presented are millimeters. Ø = diameter. a neuroma often leads to a transient relief of pain, but the endoneurial tubes are filled,10 and thus the axons would problems often return within a few months because of the penetrate into neighboring tissue where they can form a regenerative capacity of the peripheral nervous system.20,36 neuroma. Experimental studies have demonstrated that As a consequence, surgery often does not lead to sufficient axonal regeneration in the CCA model is not haphazard pain relief in the long term, while pharmaceutical pain but rather arranged, and that the number of regenerating treatment is notoriously ineffective. Therefore, neuromas axons in the CCA model is reduced.9 remain a clinical problem and urgently require a more ef- In the present study, we developed a novel surgical fective approach.18 model to prevent the formation of neuromas by creating a The primary aim of each therapeutic approach after 3D-printed Y-tube that can be customized to exactly fit the nerve injury is to restore the function of the nerve by con- proximal injured nerve end. In this model, the upper arms necting the proximal nerve stump to the distal nerve.12 If a of the Y-tube were connected by a peroneal autograft, nerve defect cannot be directly restored without tension to while the single arm was connected to the proximal sciatic the nerve ends, the current gold standard for repair is inter- nerve stump. With the aid of this Y-tube with autograft, position of an autologous nerve graft to bridge the defect.29 we intended to create a closed loop to induce regrowth However, the distal nerve stump is not always available for of injured axons into the Y-tube and thereafter through reconstruction, for example, after amputation or in cases the graft, which we hypothesized would reduce the pos- of small sensory nerve branches. As an alternative for sibility of an irregular arrangement of regenerating axons different covering and capping methods of the proximal and thereby prevent neuroma formation. An advantage of nerve stump, Slooff and Samii introduced a novel clinical this Y-tube with autograft is that it is applicable to single- technique in the late 1970s, the centro-central anastomo- nerve injuries, with a relatively straightforward surgical sis (CCA).30,31 In this CCA method, each pair of fascicles procedure. The CCA method was used as a control in our of the proximal nerve is reconnected to this nerve—or, experiments. Here, we show that the use of a Y-tube with in cases of multiple severed nerves, with each other—to autograft is an effective approach in preventing neuromas, prevent the occurrence of neuromas.30,31 This method was which also reduces the number of regenerating axons and later investigated experimentally.9,10 Gil-Salú et al. created is easily applicable in nerve surgery. an anastomosis between 2 proximal nerve ends with the interposition of an autograft.9 Recently, there has been a Methods renewed interest in the CCA method by Boroumand et al., who used this model in a clinical setting in patients Preparation of the 3D-Printed Nerve Conduits who underwent lower-limb amputation.4 Both experimen- In this study, 3 types of customized nerve conduits tal9,10 and clinical4,11,31 studies have suggested that the CCA were used: the open Y-tube, the closed Y-tube, and the ex- method is effective in reducing the size of terminal neuro- tended-arm Y-tube (Fig. 1). Conduits were made of poly- mas and in minimizing neuroma-related pain. ethylene-like material (Endur RGD450, Stratasys) using The mechanisms by which the CCA method reduces layer-by-layer 3D printing performed on a commercially neuroma formation are not known. One theory is that the available 3D printer (Objet 30 Pro, Stratasys). During fab- CCA method reduces neuromas by guiding regenerating rication, support material (FullCure 705, Stratasys) was axons into an autograft, protected from surrounding tis- used to prevent any changes in the dimensions of the con- sues, where they grow into the remaining empty endoneu- duits and was removed preoperatively. To fabricate con- rial tubes of the graft.10,31 If an autograft is not interposed duits with anatomically accurate dimensions, the diam- and the terminal stumps of the fascicles are sutured direct- eters of the sciatic and peroneal nerves in a 180-g Lewis ly end to end, it is suspected that the axons of each fascicle rat were measured (0.8 mm and 0.6 mm, respectively) and could not penetrate into the opposite fascicle because the used as tissue templates to design the conduits. To prevent

J Neurosurg Volume 130 • January 2019 185

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

FIG. 2. Illustrations showing the different types of nerve repair. A: The open Y-tube, in which the proximal sciatic nerve (S) is in- serted into the single arm of the conduit and a 10-mm-long peroneal (P) nerve graft is connected to the distal arms of the Y-tube to create a closed loop. B: The CCA, which is formed by the direct coaptation of the proximal tibial (T) and peroneal nerves, placing between them a 10-mm-long peroneal nerve graft. C: The closed Y-tube, in which the proximal sciatic nerve is inserted into the conduit. D: The extended-arm Y-tube, which has an extended plastic arm and a 5-mm extended peroneal nerve graft arm.

nerve compression due to swelling of the nerve, the inner branches of the sciatic nerve, i.e., the tibial, peroneal, and diameters of the conduits (1.0 mm for the common arm sural nerves, were exposed. The animals were randomly and 0.8 mm for the smaller arms) were made slightly larg- assigned to one of 4 experimental groups: the Y-tube with er than the diameters of the donor nerves. With the aid of autograft (n = 4), the CCA (n = 3), the closed Y-tube (n = CAD software (Solidworks, CAD2M), the conduits were 3), and the extended-arm Y-tube (n = 3). designed and 3D printed. All conduits were Y-shaped and consist of 1 common arm (inner diameter 1.0 mm, outer Experimental Groups diameter 1.2 mm) and 2 smaller arms (inner diameter 0.8 Y-Tube With Autograft mm, outer diameter 1.0 mm) to ensure a perfect fit for the After sharply transecting the sciatic nerve 3 mm before donor nerves. The open Y-tube is 4 mm long (Fig. 1A). its trifurcation, the proximal stump of the injured sciatic The closed Y-tube is 8 mm long and includes a plastic cap nerve was inserted into the single arm of the Y-tube. A that fits the 2 smaller arms of the tubes, thereby creating a 10-mm-long nerve graft was taken from the distal pero- continuous loop (Fig. 1B). The extended-arm Y-tube con- neal nerve; each side was inserted into the other tubes and sists of a short arm and a 5-mm extended arm (Fig. 1C). secured with 1 suture in an upper arm of the Y-tube to Two small holes were made in all tubes (0.15 mm wide form a loop (Fig. 2A). and located approximately 1 mm from the tube ending; Fig. 1A) at all insertion sides, which were used to facili- CCA tate pulling the nerve into the conduit and to fix the nerve After exposing the sciatic nerve and its trifurcation, inside the conduit. the tibial and peroneal nerves were sharply transected 3 mm distal from the trifurcation. A 10-mm-long graft was Experimental Animals taken from the distal peroneal nerve and directly sutured Thirteen adult female Lewis rats (160–200 g, Charles to the proximal tibial and peroneal nerve stumps (Fig. 2B) River) were used. The rats were housed in pairs with ad by using 3 sutures at each coaptation side to form an anas- libitum access to food and water. All experiments were tomosis. performed in accordance with the European guidelines for the care and use of laboratory animals and approved Closed Y-Tube by the Dutch Ethical Committee on Animal Welfare. The After a sharp transection of the sciatic nerve, the in- animals were monitored daily for signs of stress, discom- jured proximal nerve stump was inserted in the single arm fort, or autotomy. of the closed Y-tube (Fig. 2C) and single sutured to the conduit. Surgery All surgical procedures were standardized and per- Extended-Arm Y-Tube formed by the same extensively trained surgical team, After sharp transection of the sciatic nerve, the proxi- consisting of a surgeon and a technical assistant, to avoid mal nerve stump was inserted in the lower short arm of variability with regard to the surgical techniques. Under the Y-tube. A 5-mm-long nerve graft was taken from the isoflurane inhalation anesthesia (3%) and by using stan­ distal peroneal nerve and single sutured to the upper short dard aseptic microsurgical techniques with an operating arm of the conduit, leaving the distal part of the graft in microscope (OP-MI 6-SD, Carl Zeiss), the sciatic nerve of open connection with the environment (Fig. 2D). The up- the right hindlimb of the rat was exposed through a lon- per extended arm was left in open connection with the gitudinal incision down the right thigh from where the 3 surrounding tissue.

186 J Neurosurg Volume 130 • January 2019

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

In all groups, a gap of at least 10 mm was maintained mean number of axons. Thus, the density and the total between the transection site of the proximal nerve stump number of axons per nerve were calculated. and the distal denervated nerve stumps to avoid spontane- The presence or absence of a neuroma was macro- ous regeneration to the distal nerve stump. The excised scopically and microscopically determined by an inves- nerve stumps were inserted into the conduits to a depth of tigator blinded to the experimental groups. Macroscopic 1–1.5 mm and fixed to the conduits using a single epineu- observations were based on the presence of a bulb before ral 10-0 monofilament nylon suture (Ethicon) to prevent the entrance of the conduits and the presence of exces- the nerve stumps from shifting and escaping the tube after sive scar and connective tissue. The microscopic presence closing the surgery site. All muscle wound beds and skin or absence of a neuroma was determined on histological incisions were closed using 4-0 Vicryl sutures (Ethicon). evaluation of multiple PPD-stained cross sections of each Immediately postoperatively and at 24 hours, all rats re- transected nerve by looking at the following criteria: dis- ceived intramuscular buprenorphine (0.05–0.1 mg/kg, organized, with regenerating axon sprouts; erosion of the Temgesic). perineurium and epineurium; loss of funicular architec- ture; and the presence of intraneural fibrosis. Nerve Morphometry Twelve weeks after surgery, the animals were killed by Statistical Analysis administering an overdose of sodium pentobarbital (100 Data are expressed as means ± SEM. The 2-tailed in- mg/kg). Peripheral nerve segments, including the surgical dependent Student t-test was used for statistical analysis conduits, were carefully dissected from the rat’s hindlimb, between 2 groups and the paired-samples t-test for statisti- and the conduits were removed from the nerves. Proxi- cal analysis in 1 group for the data, which was normally mal nerve segments were transected 3 mm proximal to distributed, using IBM SPSS software (version 21.0, IBM). the conduit. The excised nerve segments were processed One-way ANOVA was used for comparison among mul- for resin embedding and paraffin embedding. For paraf- tiple groups. The post hoc Tukey test was applied when fin embedding, tissue samples were kept in formalin 10% there was a significant difference; p < 0.05 was considered solution for 24 hours and stored at 4°C. The tissue samples statistically significant. were embedded in paraffin and mounted on plates. Five- micrometer transverse nerve sections were cut and stained Results for 1 hour in Picro Sirius Red (Direct Red 80, Sigma) solu- tion (0.1% solution of Sirius Red F3BA in saturated aque- The different experimental conduits were successfully ous picric acid, pH 2). For resin embedding, tissue seg- fabricated using the 3D printing technique, implanted into ments were postfixed in 2.5% glutaraldehyde (Sigma) in the hind paw of the rats, and adjusted to the severed nerves. phosphate-buffered saline (PBS; pH 7.4) overnight at 4°C, Neuroma formation and nerve regeneration efficiency followed by overnight fixation using 1% osmium tetroxide. through the models was evaluated for the 4 experimental The next day, the tissue was dehydrated in graded ethanol groups (Y-tube with autograft, CCA, closed Y-tube, and followed by propylene oxide (PO) twice for 30 minutes, extended-arm Y-tube) by macroscopic observations and PO/Araldite (dilution 1:1; Huntsman Advanced Materials) histomorphometry. No complications occurred during the for 90 minutes, PO/Araldite (1:3) for 90 minutes, and pure surgical procedures. During the survival period, all rats Araldite overnight. Subsequently, the nerves were embed- exhibited motor deficiencies consistent with a lesion of the ded in Araldite and polymerized at 60°C for 72 hours. sciatic nerve. No overt signs of autotomy were observed. Semi-thin cross sections (0.5 mm) were cut with the aid Table 1 shows the number of myelinated axons in the dif- of a microtome (Ultracut, Leica) using a diamond knife, 1 ferent experimental models. mm proximal from the conduit or the suture lines and at every following 2–3 mm of the nerve and stained with 1% Y-Tube With Autograft paraphenylenediamine (PPD) for 2 minutes at 60°C. Macroscopic Observations Examination of the surgical area 12 weeks after surgery Quantitative Analysis showed neither bulbous tissue nor excessive fibrous or ad- Microscopic images of the PPD-stained cross sections hesive tissue surrounding the surgery site (Fig. 3A). The were digitalized using a digital slide scanner (Nanozoom- epineurium appeared intact without signs of inflammation. er 2.0 series system, Hamamatzu), and examined in the Hamamatsu NPDI file format. Low-magnification images Neuroma Formation were captured for a manual measurement of the cross-sec- The lack of signs of neuroma formation at the macro- tional area of the regenerated tissue using ImageJ and dig- scopic level was confirmed by the microscopic observa- ital microscope NDPI software. At 100× magnification, tions at the different sites in and around the Y-tube cor- randomly selected fields of the regenerated nerve—which, responding to the locations illustrated in Fig. 4A. The taken together, represented at least 10% of the total tissue epineurium was present and not disrupted at any location area—were manually counted for the number of myelin- (Fig. 3B–D). Connective tissue was observed surrounding ated axons by an observer who was blinded to the experi- the intact epineurium (Fig. 3B). mental groups. Beforehand, and on statistical analysis, it was determined that additional images captured beyond Microscopy of Nerve Regeneration 10% of the total tissue area did not significantly alter the Different nerve sections were microscopically ana-

J Neurosurg Volume 130 • January 2019 187

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

lyzed for the number of myelinated axons at positions a–g (Fig. 4A). Abundant nerve regeneration was observed in

PN the Y-tube and the graft (Fig. 4C). Bundles of regenerat- ing axons, from the proximal part of the Y-tube (27,800 ± 3,216 ± 5343,216 1400), split into 2 bundles of approximately uniform axo- nal density (14,000 ± 725 and 13,300 ± 835 for positions d and e, respectively; Fig. 4B). Both bundles entered the au-

TN tograft at their respective side with a significant decrease in the number of axons inside the graft, when comparing 7,828 ± 533 7,828 positions d to f and e to g (p = 0.03 and p = 0.004, respec- tively, Student t-test), with a 10.1% ± 1.2% and 16.3% ± 4.0% decrease in myelinated axons, respectively.

Lt DG CCA 13,100 ± 832 13,100 Macroscopic Observations Examination of the surgical area 12 weeks after surgery showed excessive adhesion of the nerves to the surround- ing tissues and abundant scar tissue covering the surgical

Lt PG area. Bulbous vascularized tissue covered the epineural stitches at the positions where the proximal nerves were 7,680 ± 2,070 7,680 sutured to the interposed autograft (Fig. 3A). vs lt DG, p = 0.027) vs lt DG, p = 0.004) 13,311 ± 406 (lt PG 13,311 14,573 ± 738 (lt14,573 PG Neuroma Formation The macroscopic observations suggesting neuroma formation surrounding the suture lines were confirmed Rt DG Rt by microscopic observations. Random axons that did not 13,365 ± 854 enter the autograft and penetrated the surrounding con- nective tissue were observed especially surrounding the suture lines of the coaptation sites with intraneural fibrosis and disruption of the epineurium (Fig. 3B–D). Connective tissue was observed both intraneural and surrounding the

Rt PG epineurium, as shown in Fig. 3B. 9,300 ± 2,504 rt DG, p = 0.036) vs DGR, p = 0.030) 15,959 (PGR ± 278 Microscopy of Nerve Regeneration 11,118 ± 418 (rt GR vs ± 418 11,118 Different nerve sections were microscopically ana- lyzed at the locations illustrated in Fig. 5A. Sections of the proximal peroneal and tibial nerves, corresponding to positions a and b, respectively, in Fig. 5B and C, showed

Rt DT an axonal morphology and number of axons that reflected 7,935 ± 522 7,935

13,295 ± 835 a noninjured situation (7829 ± 534 and 3216 ± 534, respec- DT, p = 0.005) DT, tively). The number of axons increased after entering the 1,300 ± 822 (PT vs rt autograft corresponding to positions c and d (13,410 ± 625 and 13,213 ± 652, respectively; Fig. 5B). A significant de- crease in the number of axons was found in the middle of the autograft in positions e and f (11,398 ± 795 and 10,839

Lt DT ± 405, respectively) when compared with positions c and d (13,213 [c vs e] ± 652 and 13,410 ± 625 [d vs f], respec- 14,014 ± 725 DT, p = 0.004) DT, DT, p = 0.0001) DT,

97 ± 8497 vs rt (lt DT tively; p = 0.036 left and p = 0.027 right), with a 13.7% ± 1,162 ± 801 (PT vs lt 1,162 2.7% and 19.2% ± 3.2% decrease, respectively.

Closed Y-Tube

PT Macroscopic Observations Examination of the surgical area 12 weeks after surgery 14,918 ± 1,439 14,918 27,802 ± 1,393 27,802 16,288 ± 1,808 showed excessive connective tissue surrounding the surgi- cal area. Bulbous, vascularized tissue surrounded the proxi- mal nerve stump before the entrance of the Y-tube (Fig. 3A). No nerve tissue was observed in the connecting cap of the

Model conduit. The total nerve matrix that grew through the 2 up- Y-tube autograft per arms of the tube was smaller than the nerve matrix that Extended-arm Closed Y-tube Closed Y-tube w/ Y-tube CCA TABLE 1. Mean 1. TABLE ± SEM number of myelinated axons specified for different segments per experimental model DG = distal graft; = distal DT tube; PG = proximal graft; PN = peroneal nerve; PT = proximal tube; TN = tibial nerve. Both the Y-tube with autograft and CCA model demonstrate a significant decrease in the number of regenerated axons in the middle of the autograft(right and left proximalautograft grafts) when (right andcompared left distal to the distal grafts). Insides the closed of the Y-tube a robust decrease is found in the left and right distal tube ends when compared with the proximal tube. The extendedin the number Y-tube demonstrates of regenerated a significant axons between difference the left and right distal tube outlets, showing that the majority of the axons extend through the distal outlet extended with the graft (right distal tube). grew through the proximal part of the conduit (Fig. 3A).

188 J Neurosurg Volume 130 • January 2019

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

FIG. 3. Evaluation of neuroma formation in the different experimental groups, 12 weeks after surgery. A: Ex vivo photographs of the different experimental models, with the regenerated tissue pulled out of the conduits. The dashed circles indicate the areas suspected for neuroma formation. No bulb formation or excessive scar tissue was observed around the Y-tube with autograft (G), while the CCA, closed Y-tube, and extended-arm Y-tube nerves have a well-perfused tissue bulb at the interface of the tubes in the sciatic (S) nerve or at the coaptation sites of the CCA in the tibial (T) and peroneal (P) nerves. B: Representative micro- scopic images of sections stained with Picro Sirius Red from the areas depicted in panel A. The Y-tube with autograft presents an intact epineurium and connective tissue (red) outside the epineurium, while all other models show erosion of the epineurium and the presence of intraneural connective tissue, indicating the presence of neuromas. Bar = 500 µm. C: Representative light micrographs of semi-thin transverse sections stained with PPD at the sites depicted in panel A. In the Y-tube with autograft the epineurium is intact with very few axons outside the epineurium. The CCA, closed Y-tube, and extended-arm Y-tube groups all show morphology that corresponds to the definition of a neuroma, including interruption of the epineurium and randomly regen- erating axons outside the epineurium. The asterisks indicate the unaffected sural nerve. Bar = 500 µm. D: High-magnification images of the areas depicted in the boxes in panel C. The Y-tube with autograft presents an organized funicular architecture with an intact epineurium. The CCA, closed Y-tube, and extended-arm Y-tube show neuromas with interruption of the epineurium (black arrowheads), presence of extraneural fascicles filled with sprouts, and randomly growing axons that are longitudinally sectioned (black arrows). Bar = 50 µm. Figure is available in color online only.

Neuroma Formation nective tissue and the presence of intraneural fibrosis and The macroscopic observations suggesting neuroma connective tissue (Fig. 3B). formation were confirmed by microscopic observations, which showed randomly oriented regenerating axons be- Microscopy of Nerve Regeneration fore the Y-tube entrance (Fig. 3C and D), with loss of the Different nerve sections were microscopically analyzed epineurium and axons penetrating the surrounding con- at the locations illustrated in Fig. 6A. Extensive nerve re-

J Neurosurg Volume 130 • January 2019 189

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

FIG. 4. Nerve regeneration in the Y-tube with autograft. A: Schematic model with indicated levels of analyzed sections. B: Bar graph showing the mean myelinated axon counts in the Y-tube with autograft 12 weeks after surgery. The letters below the bars represent the schematically depicted segments in panel A. The myelinated axon count is significantly reduced in the distal seg- ments of the graft compared with the proximal segments (p = 0.03 and p = 0.004 for the left and right segments, respectively). ​ C: Representative semi-thin transverse sections stained with PPD at 3 different segments in the Y-tube and nerve graft 12 weeks after nerve injury, corresponding to the locations demonstrated in panels A and B. Sections a and b show that the conduit, 12 weeks after injury, is filled with an organized pattern of small myelinated axons. Section f demonstrates that the regenerating ax- ons have invaded the autograft in an organized pattern. Bar = 10 µm. Data are shown as mean ± SEM; data were analyzed using the Student t-test (n = 4 animals). *p < 0.05, **p < 0.01. Figure is available in color online only.

generation was seen in the proximal part of the Y-tube In all rats, nerve matrix was extending beyond the adjusted corresponding to section a in Fig. 6C (14,919 ± 1439). Sig- autograft, making the graft appear longer than the 5-mm nificantly fewer axons were found after the bifurcation of graft connected to the tube during surgery. In 2 rats, no the Y-tube when comparing positions b (1300 ± 823) and c clear endpoint of the nerve matrix extending beyond the (1162 ± 802) to position a (p = 0.005 [b vs a] and p = 0.004 adjusted autograft was identified. A clear difference in size [c vs a], respectively; Fig. 6B), which is a total decrease of was found in the diameter of the regenerated nerve tissue 90.5% of myelinated axons compared with the proximal of position c compared with d (Fig. 3A). arm. No axons were observed through the connecting cap Neuroma Formation of the Y-tube. The macroscopic observations suggesting neuroma Extended-Arm Y-Tube formation were confirmed in 2 rats by microscopic ob- servations, which showed randomly oriented regenerating Macroscopic Observations axons before the extended Y-tube entrance (Fig. 3C and Examination of the surgical area 12 weeks after sur- D). In addition, microscopic observations showed signs of gery showed vascularized bulbous tissue at the distal end loss of epineurium and randomly regenerating nerve fibers of the sciatic nerve before the entrance of the Y-tube (Fig. around the Y-tube with connective tissue in and around the 3A), and connective tissue surrounding the surgical area. epineurium (Fig. 3B and D).

190 J Neurosurg Volume 130 • January 2019

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

FIG. 5. Peripheral nerve regeneration in the CCA. A: Schematic model with indicated levels of analyzed sections. B: Bar graph showing the mean myelinated axon counts in the CCA. The letters below the bars represent the schematically depicted segments in the anastomosis, as demonstrated in panel A. The myelinated axon count is significantly reduced in the distal segments of the graft compared with the proximal segments (p = 0.036 and p = 0.027 for left and right segments, respectively). C: Representa- tive light micrographs of semi-thin sections stained with PPD of the tibial nerve (section a), the peroneal nerve (section b), and the distal part of the graft (section e), corresponding to the levels shown in panels A and B. The tibial and peroneal nerves show an axonal morphology and number of axons that reflect noninjured nerves. Section e shows a more regenerative morphology with the presence of many small myelinated axons that have invaded the graft. Bar = 10 µm. Data are shown as mean ± SEM; morphologi- cal data were analyzed using the Student t-test (n = 3 animals). *p < 0.05. Figure is available in color online only.

Microscopy of Nerve Regeneration in the Y-tube with autograft group, whereas all other Different nerve sections were microscopically ana- groups showed randomly growing axons and loss of epi- lyzed at the locations illustrated in Fig. 7A. Nerve re- neurium, indicating neuroma formation. Histologically generation was observed through the entire conduit and confirmed neuromas were only observed at the position into the autograft (Fig. 7C). The axons extending through where the nerve was about to enter the Y-tube or surround- the proximal tube, corresponding to position a (16,288 ing the coaptation sites in the CCA, as shown in Fig. 3. ± 1808), were split into 2 unequal bundles. Significantly An organized pattern of regenerating axons was observed more axons were present in the graft arm (7936 ± 522) inside the different Y-tubes. when compared with the extended plastic arm (98 ± 84, p = 0.0001; Fig. 7B). No significant differences were found Nerve Regeneration in the number of axons at different positions in the auto- A significant difference, considering the number of my- graft. elinated axons, was observed in the proximal part of the Y-tube between the following groups: the Y-tube with au- Comparing the Different Experimental Groups tograft group, the closed Y-tube group, and the extended- Neuromas arm Y-tube group (1-way ANOVA; F = 22.763, p = 0.001). Macroscopic and microscopic results showed that, 12 A Tukey post hoc test revealed that the number of axons weeks after surgery, no neuroma formation was observed in the proximal part of the Y-tube was significantly lower

J Neurosurg Volume 130 • January 2019 191

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

FIG. 6. Nerve regeneration in the closed Y-tube. A: Schematic model with indicated segments of sections for analysis. B: Bar graph showing the mean numbers of myelinated axons in the model. The letters below the bars represent the segments in the model as demonstrated in panel A. The myelinated axon count is significantly reduced after entering the bifurcation (p = 0.007, Student t-test). C: Representative light micrographs of semi-thin sections stained with PPD of the regenerating nerve at 3 different levels in the tube after 12 weeks of survival. The sections correspond to the locations demonstrated in panels A and B. Sections a and b show that the conduit is filled with small myelinated axons 12 weeks after injury. Section d, however, reveals connective tissue covering the loop without the presence of axons; thus, the axons failed to grow into the loop. The density of axons in section b is comparable to that in section a; however, the total nerve surface of b is much smaller. Scale bars = 10 µm. Data are shown as mean ± SEM; morphological data were analyzed using the Student t-test (n = 3 animals). *p < 0.05. Figure is available in color online only. in the closed Y-tube (14,919 ± 1439, p = 0.001) and the tograft. Following nerve transection, the proximal sciatic extended-arm Y-tube (16,288 ± 1808, p = 0.003) compared nerve stump was inserted into the lower arm of a custom- with the Y-tube with autograft (27,802 ± 1393). There was made Y-tube, while the 2 upper arms of the Y-tube were no statistically significant difference between the number connected by an autologous peroneal autograft to form a of axons in the proximal part of the closed Y-tube and the closed loop. At 12 weeks of survival, we found that re- extended-arm Y-tube (p = 0.826). In the closed Y-tube, no generating axons had extended through the Y-tube and axons were observed in the connecting tube, while many grown into the autograft from both sites. No signs of neu- axons were observed at the same position when using an roma formation were observed in this experimental model, autograft in the Y-tube with autograft model (compare Fig. while the number of regenerating axons in the middle of 6C to Fig. 5C). the autograft was significantly reduced. In contrast, the other experimental models, i.e., the CCA, closed Y-tube, and extended-arm Y-tube models, all showed signs of neu- Discussion roma formation. In the present study, we have developed a novel surgi- The concept of nerve tubes has evolved from an inves- cal strategy that prevents neuroma formation after nerve tigation tool to a surgical device that is used in peripheral injury by using a 3D-printed Y-tube with an interposed au- nerve injury. In recent years, both scientists and the medi-

192 J Neurosurg Volume 130 • January 2019

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

FIG. 7. Nerve regeneration in the extended-arm Y-tube. A: Schematic model with indicated segments of analyzed sections. ​ B: Bar graph showing the mean myelinated axon counts in the model. The letters below the bars represent the segments in the model as demonstrated in panels A and B. A significantly increased number of myelinated axons are present in the graft arm compared with the extended polyethylene arm (p = 0.004). C: Representative light micrographs of semi-thin sections stained with PPD corresponding to the locations demonstrated in panel A. Sections a and b show that the conduit is filled with a large number of regenerating myelinated axons. On the other hand, section c shows hardly any axons regenerating through the extended plastic arm. Thus, the majority of the regenerating axons from the proximal part of the tube extend through the extended graft arm. Sec- tion e indicates that the axons from the short plastic arm extended through the nerve graft. Bar = 10 µm. Data are shown as mean ± SEM; morphological data were analyzed using the Student t-test (n = 3 animals). *p < 0.05. Figure is available in color online only. cal device industry have been working on designing ideal tubes. In addition, this ensured that a closed loop is formed conduits, as they are a promising tool for clinical use.7 Pe- in which regenerating axons are directed in a tension-free ripheral nerve injury may involve one or more nerves, and environment between the nerve ends. The simplicity and the injured nerve may be varied in geometry. The natural flexibility of the 3D printing technique is a clear advan- variance in patient anatomies has motivated the develop- tage, allowing for individual-based treatment after nerve ment of personalized treatment for peripheral nerve inju- injury. ry.16 For this purpose, 3D printing is an advancing technol- Thus far, there have been many attempts to prevent neu- ogy in the field of surgery and enables the fabrication of roma formation and to treat painful terminal neuromas. personalized constructs.19,21 In this study, 3D printing with However, in most cases the results have been ineffective, polyethylene-like material allowed us to successfully fab- as a neuroma often reappears together with neuroma-asso- ricate conduits with dimensions that nicely fit the sciatic ciated pain. In this study, we focused on inducing an en- and peroneal nerves in the rat. The commercially avail- vironment for controlled regrowth of regenerating axons, able material we used to fabricate the conduits was En- which is supposed to reduce irregular outgrowth of the re- dur RGD450 with Fullcure 705 as support material, which generating axons, thus preventing neuroma formation. Our allowed us to accurately print the conduits with the de- Y-tube with autograft model was found to be very well sired small dimensions on our 3D printer. The small holes, suited in preventing neuroma formation in a single tran- which were designed at the nerve insertion sites of the sected nerve by providing a closed and controlled pathway. tubes, made it easy to pull the transected nerves into the The closed loop in this model suppresses the formation tubes using a single suture, which minimized damage to of communication with the surrounding tissues, where in- the epineurium and securely fixed the nerve stumps in the trusion of scar tissue is minimized, but also prevents the

J Neurosurg Volume 130 • January 2019 193

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al. escape of regenerating axons.38 Moreover, a physical bar- cal barrier, which does not allow a perfect-fit connection rier protects the regenerating axons from various external between the nerve stumps and the autograft. Weiss38 pro- stimuli, such as pressure, which reduces scar tissue and posed that open connections at the suture lines might of- also reduces pressure around the nerve.8 In the Y-tube fer opportunities for axons to regenerate outside the nerve with autograft, we found all the regenerating axons from environment and for connective tissue and vessels to pen- the proximal nerve growing into the Y-tube and into the etrate along the sutures. This makes the CCA method less autograft in an organized way, without signs of neuroma suitable for neuroma prevention. formation or scar formation in or around the conduit or In our Y-tube with autograft model, we found that all autograft. regenerating axons sprouted through the Y-tube. These re- An autograft likely provides a gradient of essential generating axons, coming from the proximal sciatic nerve, neurotrophic factors for guiding regenerating axons from extended through the Y-junction where it splits into 2 bun- the proximal nerve stump in the direction of the autograft dles of approximately uniform axonal density. At the point and thereafter into the denervated endoneurial tubes that of entering the autologous nerve graft, we observed a 4% remain inside the autograft following wallerian degenera- increase in the number of myelinated axons, which likely tion.34 Many studies have shown the benefits of the auto- reflects axonal sprouting.38 In contrast, the number of ax- graft in guiding regenerating axons;22,27,38 however, their ons in the middle of the autograft, where the axons from possible role in neuroma prevention has not been clearly opposite directions are expected to encounter each other, documented. Nerve autografts are immunologically inert, was significantly reduced by about 13% compared with possess the extracellular matrix structure meant to host the lateral edges of the graft. This suggests that a form axons, and contain Schwann cells, which are known to se- of axonal growth inhibition occurred in the middle of the crete diffusible factors that guide axonal regeneration.29 In autograft where the axons are supposed to encounter each our study, the attracting influence of the autograft on re- other. It is conceivable that the growth cones of the regen- generating axons was clearly demonstrated in the extend- erating axons are involved in this process, as they react to ed-arm Y-tube, where 99.4% of the regenerating axons in changes in the environment.15 In support of this view, in the Y-tube merged into the Y-tube arm connected to the our extended-arm Y-tube, the number of regenerating ax- autograft, leaving the extended plastic arm almost empty. ons extending through the graft arm did not significantly Furthermore, this is also supported by our observations change at any position in the autologous graft, which fur- of the closed Y-tube, which lacked neurotrophic factors, ther strengthens the idea of axonal growth inhibition in the showing the avoidance of axons to grow into the tube and, middle of the autograft in the Y-tube with autograft. How- consequently, the formation of a neuroma outside the tube. ever, the mechanisms underlying the decrease in axons in Therefore, we can conclude that the autograft is important the middle of the autograft of the Y-tube with autograft for both the guidance of regenerating axons and the mod- model remain unknown. eration of their growth in the absence of a target, making Our observations hold promise for clinical use as they the autograft an important tool in preventing neuromas. indicate that introducing the proximal nerve end into a An early attempt to prevent neuroma formation using perfect-fit structure with an autograft may guide regen- an autograft is the CCA.10,31 The CCA method, in which erating axons and induce axon growth inhibition, thereby 2 nerve branches are sutured to an autograft to create an preventing neuroma formation. In addition, an advantage anastomosis, was found to limit, but not prevent, neuroma is that the autograft can be obtained from the injured nerve formation by offering the regenerating axons an autograft itself by harvesting a piece of the proximal stump, thereby to restore the continuity of the nerve fascicles.4,9,10,31 This preventing the sacrifice of other nerves or the use of donor result may be explained by the mismatch in diameter be- nerves with possible associated morbidity. This surgical tween the proximal nerves and the autograft and in the strategy for preventing neuromas may be suitable for treat- number of axons between the 2 proximal nerve stumps, ment of painful neuromas as well as fresh nerve injuries leading to a bad fit. In our CCA model, we aimed to mini- where end-to-end reconstruction is not an option (e.g., due mize the mismatch between the proximal nerves and to tissue loss or amputation). the autograft by using nerves with a diameter that better Our aim in this work was to produce a surgical model aligned with the diameter of the autograft, thereby in- in which neuroma formation was prevented. In this study, creasing the chance of all axons growing into the empty evaluation of neuroma formation and nerve regeneration endoneurial tubes inside the autograft. The autograft in was assessed using nerve morphometry. Other techniques, this model is approached by regenerating axons from both like behavioral outcomes5,6 in which neuropathic pain be- sides, which, in theory, should create a situation in which havior can be measured, were not assessed in this study. endoneurial channels of the autograft are occupied by re- Behavioral outcomes may often be applied when continu- generating axons from both sides. We found the number ity of the nerve is partially spared. In our model, however, of regenerating axons to be significantly reduced in the we aimed to create an amputation-like situation in which middle of the autograft, making it conceivable that a form the sciatic nerve was transected without the possibility of of axonal growth inhibition occurred at the position where being rejoined with the distal end. the growth cones of the regenerating axons meet. How- From nerve repair surgery in the clinic and from litera- ever, in this model we still found neuromas surrounding ture, it is known that the presence of rigid and semirigid the coaptation sides. This result can likely be explained nerve guidance conduits, such as an interposition graft, by the inequality between the diameters of the proximal should be avoided, as nerve compression is reported over nerve stumps and the autograft and the lack of a physi- time.2,23,26 In our experimental study, we used nondegrad-

194 J Neurosurg Volume 130 • January 2019

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al. able Y-tubes as a proof of principle to assess nerve regen- central anastomosis in the prevention and treatment of pain- eration and neuroma formation in a closed-loop model. ful terminal neuroma. An experimental study in the rat. J However, for clinical translation, biodegradable conduits Neurosurg 63:754–758, 1985 are of interest since they can be used to prevent possible 11. Gorkisch K, Boese-Landgraf J, Vaubel E: Treatment and 14 prevention of amputation neuromas in hand surgery. Plast compression problems. Hu et al. reported biodegradable Reconstr Surg 73:293–299, 1984 cellularized designer conduits, which degraded in 2–4 12. Grinsell D, Keating CP: Peripheral nerve reconstruction months in vivo. Computer-aided design/computer-aided after injury: a review of clinical and experimental therapies. manufacturing (CAD-CAM) and 3D printing warrant the Biomed Res Int 2014:698256, 2014 delivery of nerve repair tubes that can be customized to 13. Hall S: Nerve repair: a neurobiologist’s view. J Hand Surg anatomical dimensions and combined with degradable [Br] 26:129–136, 2001 materials.16 The use of a 3D-printed biodegradable Y-tube 14. Hu Y, Wu Y, Gou Z, Tao J, Zhang J, Liu Q, et al: 3D-engi- neering of cellularized conduits for peripheral nerve regen- with autograft could prevent possible nerve compression eration. Sci Rep 6:32184, 2016 problems, thereby making this an interesting technique for 15. Hur EM, Yang IH, Kim DH, Byun J, Saijilafu, Xu WL, et al: the production of the Y-tube, with the potential for clinical Engineering neuronal growth cones to promote axon regen- translation in preventing and managing neuromas. eration over inhibitory molecules. Proc Natl Acad Sci U S A 108:5057–5062, 2011 16. Johnson BN, Lancaster KZ, Zhen G, He J, Gupta MK, Kong Conclusions YL, et al: 3D printed anatomical nerve regeneration path- Our study demonstrates that the Y-tube with autograft ways. Adv Funct Mater 25:6205–6217, 2015 prevents neuroma formation and significantly reduces the 17. Koch H, Haas F, Hubmer M, Rappl T, Scharnagl E: Treat- number of axons in the middle of the autograft, thereby ment of painful neuroma by resection and nerve stump trans- outperforming the CCA and capping techniques. In our plantation into a vein. Ann Plast Surg 51:45–50, 2003 18. Kryger GS, Kryger Z, Zhang F, Shelton DL, Lineaweaver view, these findings, combined with the feasibility of this WC, Buncke HJ: Nerve growth factor inhibition prevents surgical model, hold promise for a novel clinical approach traumatic neuroma formation in the rat. J Hand Surg Am to both treat painful neuromas and prevent neuroma for- 26:635–644, 2001 mation after nerve injury. 19. Lee SJ, Nowicki M, Harris B, Zhang LG: Fabrication of a highly aligned neural scaffold via a table top stereolithog- raphy 3D printing and electrospinning. Tissue Eng Part A Acknowledgments 23:491–502, 2017 We would like to express our gratitude to E. Fijneman for her 20. Lewin-Kowalik J, Marcol W, Kotulska K, Mandera M, Klim- help with the surgery and anesthesia and thank E. D. Haasdijk for czak A: Prevention and management of painful neuroma. her practical suggestions and help. Furthermore, we would also like Neurol Med Chir (Tokyo) 46:62–68, 2006 to thank M. Manten for fabricating the 3D-printed conduits and his 21. Li C, Cheung TF, Fan VC, Sin KM, Wong CW, Leung GK: involvement in the design of the conduits. Applications of three-dimensional printing in surgery. Surg Innov 24:82–88, 2017 22. Lotfi P, Garde K, Chouhan AK, Bengali E, Romero-Ortega References MI: Modality-specific axonal regeneration: toward selective 1. Atherton DD, Taherzadeh O, Facer P, Elliot D, Anand P: The regenerative neural interfaces. Front Neuroeng 4:11, 2011 potential role of nerve growth factor (NGF) in painful neu- 23. Lundborg G, Rosén B, Dahlin L, Holmberg J, Rosén I: romas and the mechanism of pain relief by their relocation to Tubular repair of the median or ulnar nerve in the human muscle. J Hand Surg [Br] 31:652–656, 2006 forearm: a 5-year follow-up. J Hand Surg Br 29:100–107, 2. Belkas JS, Shoichet MS, Midha R: Peripheral nerve regenera- 2004 tion through guidance tubes. Neurol Res 26:151–160, 2004 24. Mackinnon SE, Dellon AL, Hudson AR, Hunter DA: Altera- 3. Blumberg H, Jänig W: Discharge pattern of afferent fibers tion of neuroma formation by manipulation of its microenvi- from a neuroma. Pain 20:335–353, 1984 ronment. Plast Reconstr Surg 76:345–353, 1985 4. Boroumand MR, Schulz D, Uhl E, Krishnan KG: Tibiope- 25. Mao J, Gold MS, Backonja MM: Combination drug thera- roneal short circuiting for stump neuroma pain in amputees: py for chronic pain: a call for more clinical studies. J Pain revival of an old technique. World Neurosurg 84:681–687, 12:157–166, 2011 2015 26. Merle M, Dellon AL, Campbell JN, Chang PS: Complica- 5. Chaplan SR, Bach FW, Pogrel JW, Chung JM, Yaksh TL: tions from silicon-polymer intubulation of nerves. Microsur- Quantitative assessment of tactile allodynia in the rat paw. J gery 10:130–133, 1989 Neurosci Methods 53:55–63, 1994 27. Nichols CM, Brenner MJ, Fox IK, Tung TH, Hunter DA, 6. Choi Y, Yoon YW, Na HS, Kim SH, Chung JM: Behavioral Rickman SR, et al: Effects of motor versus sensory nerve signs of ongoing pain and cold allodynia in a rat model of grafts on peripheral nerve regeneration. Exp Neurol neuropathic pain. Pain 59:369–376, 1994 190:347–355, 2004 7. de Ruiter GC, Malessy MJ, Yaszemski MJ, Windebank AJ, 28. Noble J, Munro CA, Prasad VS, Midha R: Analysis of upper Spinner RJ: Designing ideal conduits for peripheral nerve and lower extremity peripheral nerve injuries in a population repair. Neurosurg Focus 26(2):E5, 2009 of patients with multiple injuries. J Trauma 45:116–122, 8. Galeano M, Manasseri B, Risitano G, Geuna S, Di Scipio F, 1998 La Rosa P, et al: A free vein graft cap influences neuroma 29. Ray WZ, Mackinnon SE: Management of nerve gaps: auto- formation after nerve transection. Microsurgery 29:568– grafts, allografts, nerve transfers, and end-to-side neurorrha- 572, 2009 phy. Exp Neurol 223:77–85, 2010 9. Gil-Salú JL, González-Darder JM: Study of nerve regenera- 30. Samii M: Centrocentral anastomosis of peripheral nerves. A tion in centrocentral anastomosis. Acta Neurochir (Wien) neurosurgical treatment of amputation neuromas, in Siegfried 105:39–43, 1990 J, Zimmermann M (eds): Phantom and Stump Pain. New 10. González-Darder J, Barberá J, Abellán MJ, Mora A: Centro- York: Springer, 1981, pp 123–125

J Neurosurg Volume 130 • January 2019 195

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC A. Bolleboom et al.

31. Slooff AC: Microsurgical possibilities in the treatment of peripheral pain. Clin Neurol Neurosurg 80:107–111, 1977 Disclosures 32. Stokvis A, van der Avoort DJ, van Neck JW, Hovius SE, Co- ert JH: Surgical management of neuroma pain: a prospective The authors report no conflict of interest concerning the materi- follow-up study. Pain 151:862–869, 2010 als or methods used in this study or the findings specified in this 33. Taylor CA, Braza D, Rice JB, Dillingham T: The incidence of paper. peripheral nerve injury in extremity trauma. Am J Phys Med Rehabil 87:381–385, 2008 Author Contributions 34. Terenghi G: Peripheral nerve regeneration and neurotrophic Conception and design: all authors. Acquisition of data: Bolle- factors. J Anat 194:1–14, 1999 boom, Tuk. Analysis and interpretation of data: all authors. Draft- 35. Tyner TR, Parks N, Faria S, Simons M, Stapp B, Curtis B, ing the article: van Neck, Bolleboom, de Ruiter, Coert, Holstege. et al: Effects of collagen nerve guide on neuroma formation Critically revising the article: van Neck, Bolleboom, de Ruiter, and neuropathic pain in a rat model. Am J Surg 193:e1–e6, Coert, Holstege. Reviewed submitted version of manuscript: van 2007 Neck, Bolleboom, de Ruiter, Coert, Holstege. Approved the final 36. Vernadakis AJ, Koch H, Mackinnon SE: Management of version of the manuscript on behalf of all authors: van Neck. neuromas. Clin Plast Surg 30:247–268, vii, 2003 Statistical analysis: Bolleboom. Administrative/technical/material 37. von Hehn CA, Baron R, Woolf CJ: Deconstructing the neuro- support: Bolleboom, Tuk. Study supervision: van Neck, de Ruiter, pathic pain phenotype to reveal neural mechanisms. Neuron Holstege. 73:638–652, 2012 38. Weiss P: The technology of nerve regeneration: a review. Correspondence Sutureless tubulation and related methods of nerve repair. J Johan W. van Neck: Erasmus MC, University Medical Center, Neurosurg 1:400–450, 1944 Rotterdam, The Netherlands. [email protected].

196 J Neurosurg Volume 130 • January 2019

Unauthenticated | Downloaded 10/11/21 02:53 AM UTC