Journal of Plastic, Reconstructive & Aesthetic Surgery (2013) 66, 251e259

Natural conduits for bridging a 15-mm nerve defect:Comparison of the vein supported by muscle andbone marrow stromal cells with a nerve autograft*

Tim H.J. Nijhuis a,*, Corneel W.J. Bodar a, Johan W. van Neck a,Erik T. Walbeehm a, Maria Siemionow b, Maria Madajka b, Joanna Cwykiel b,Joleen H. Blok c, Steven E.R. Hovius a

aDepartment of Plastic, Reconstructive and Hand Surgery, Erasmus MC, University Medical Center, Rotterdam,The Netherlandsb Institute of Dermatology and Plastic Surgery, Cleveland Clinic, Cleveland, OH, USAcDepartment of Clinical Neurophysiology, Erasmus MC, University Medical Center, Rotterdam, The Netherlands

Received 16 October 2011; accepted 13 September 2012

KEYWORDSVeinemuscle graft;BMSCs;Autograft;Nerve reconstruction

* Portions of this work were presentMexico, 15e16 January 2011.* Corresponding author. Erasmus MC,

Postal Box 2040, 3000 CA, Rotterdam,E-mail address: t.nijhuis@erasmus

1748-6815/$-seefrontmatterª2012Brihttp://dx.doi.org/10.1016/j.bjps.2012.0

Summary Object: The gold standard for reconstructing large nerve defects, the nerveautograft, results in donor-site morbidity. This detrimental consequence drives the searchfor alternatives. We used a vein filled with a small piece of fresh muscle to prevent the veinfrom collapsing and with bone marrow stromal cells (BMSCs) to enhance regeneration.Methods: In 60 rats, a 15-mm sciatic nerve defect was bridged with a nerve autograft, a veinfilled with muscle or a vein filled with muscle and BMSCs. Toe spread and pinprick were used toevaluate motor and sensory function. Compound muscle action potentials (CMAPs) and thegastrocnemius muscle index (GMI) were recorded to assess conduction properties and denerva-tion atrophy. Extensive histology was performed to confirm presence of BMSCs and to evaluateregeneration by staining neural tissue for Schwann cells and neural growth factor.Results: After 12 weeks, all animals responded with toe spread and pinprick reaction;significant differences were found between groups.

Six weeks post grafting no difference was found comparing the GMI between the groups.Group I had a significant increase in GMI at 12 weeks compared to group II and group III.The CMAP measurements showed comparable results at 6 weeks post grafting. Twelve weeksafter reconstruction, group I had significantly better results compared to group II and group

ed in Abstract form at the annual meeting for the American Society for Peripheral Nerve in Cancun,

University Medical Center, Department of Plastic, Reconstructive and Hand Surgery, Room Ee 15.91,The Netherlands. Tel.: þ31 10 7043291.

mc.nl (T.H.J. Nijhuis).

tishAssociationofPlastic,ReconstructiveandAestheticSurgeons.PublishedbyElsevierLtd.All rightsreserved.9.011

252 T.H.J. Nijhuis et al.

Figure 1 Surgical techniques usedand Group III a vein-muscle graft w

III. Group III showed a tendency to outperform group II at 12 weeks postoperatively. Immuno-fluorescence analysis showed an increased number of Schwann cells in group III compared togroup II. The BMSCs were visible 6 and 12 weeks postoperatively.Conclusions: This study is a step forward in the search for an alternative to the nerve autograftbecause it demonstrates the beneficial effect of BMSCs to a conduit. However, our data do notdemonstrate sufficient benefit to warrant clinical implementation at this stage.ª 2012 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published byElsevier Ltd. All rights reserved.

The autologous non-vascularised perineural nerve graft canbe considered the gold standard of treatment in the clinicalsetting.1 However, harvesting a donor nerve requires anextra operating site and the removal of a sensory nerve(usually the sural nerve) thus resulting in a sensory deficitat the donor site. Further, the limited length available isnot always sufficient to warrant an adequate length forreconstruction. Furthermore, scarring, neuroma formation,neuropathic pain and other symptoms have all beencorrelated to donor-site morbidity.2 Veins form a frequentlydescribed alternative for nerve grafting.3e8 They can beharvested with limited donor-site morbidity and providea metabolically supportive environment for the regenerat-ing axons. However, the tendency of an empty vein tocollapse when used for bridging gaps hinders the recoveryprocess.4e7 This problem may be solved by placing a pieceof muscle in the graft, which prevents its collapse, and hasbeen shown to yield promising results.3,9,10 Furthermore,the muscle appears to provide an adequate matrix for thenerve fibres growing inside the vein.11

The use of bone marrow stromal cells (BMSCs) as anextra stimulus for nerve regeneration has gained muchinterest over the last decade. For example, Cui et al. havedemonstrated enhanced nerve regeneration using embry-onic stem cells and revealed successful regeneration whentransplanted in an epineural tube.12

The present study was designed to compare a modifica-tion of a well-known grafting technique (the muscle-supported vein graft with additional injected BMSCs) tothe autologous nerve graft.

Methods

The experimental protocol was approved by the AnimalExperiments Committee (Protocol number: EUR 1896, 133-09-07). Eighty isogenic adult female Lewis rats, weighing

in this study. Group I comprisesith BMSCs.

180e200 g, were used. Surgical procedures and electro-physiological evaluations were performed under generalanaesthesia (Isoflurane, 1e2% in a mixture of O2/N2O). Sixtyanimals were randomly allocated to one of two groups. Thefirst group, consisting of 30 animals, was evaluated 6 weekspost-surgery, the other group of 30 animals after 12 weeks.These two groups were subdivided into three groups of 10rats each. Group I was treated with a donor autograft.Group II was treated with a veinemuscle graft and group IIIwith the veinemuscle graft and additional injection ofBMSCs. The remaining 20 isogenic animals served as auto-graft donors for the animals in group I and for harvestingbone marrow to isolate the BMSCs.

BMSC preparation

BMSCs were isolated and cultured following exactly thesame protocol as previously described by the same group.13

Prior to injection, the BMSCs were labelled with a redmembrane dye PKH-26 (SigmaeAldrich, UK) in order toevaluate their capacity to differentiate into neuronal andother tissue types as well as to track their migration intolymphoid organs and the contralateral sciatic nerve. PKH-26 staining was performed in accordance with the manu-facturer’s instruction and a final concentration of 3 � 106

cells in 0.05 ml was prepared and injected in the two endsof the vein using a small syringe (27G), injecting 0.025 ml ineach side.

Fluorescence-activated cell sorting analysis

The fluorescence-activated cell sorting (FACS) analysis wasperformed with an analogue BMSC culture from the micro-surgical laboratory in the Cleveland Clinic. The first inves-tigator has been trained there and exactly the sameisolation and culture method was implemented in this

an autologous nerve graft, Group II an empty vein-muscle graft

Veinemuscle graft supported with BMSCs for nerve reconstruction 253

study. The cells used for this analysis were cultured ata different time by a co-author of this paper and shouldillustrate the quality of the cells used in this experiment.Cells from a 14 days’ old culture were trypsinised (0.05%trypsin for 5 min at room temperature), washed three timesin washing buffer (1% bovine serum albumin (BSA), 0.1%sodium azide) and fluorescently labelled. Different markerswere used to evaluate the phenotype (CD90, CD34 andCD31) and CD45RA, CD8a and CD3 were used to evaluatethe expression of B lymphocytes, natural killer (NK) cells,macrophages and peripheral T-lymphocytes.

Surgical technique

The surgical procedure was performed on the sciatic nerveof the left hind limb. The right limb sciatic nerve served e ifpossible e as a pairwise control. In the donor animals,a skin incision was made in the left gluteal region, exposingthe sciatic nerve through a gluteal muscle-splitting incisionand externally dissected to excise a 15-mm segment of thenerve for use as a nerve graft in group I. The same proce-dure was performed in the recipient animals. For group IIand group III, the jugular vein was harvested througha longitudinal mid-line incision of 35 mm in the neck. Theleft external jugular vein was dissected and both theproximal and the distal ends of the vein were ligated. Amuscle fragment of 1½ � 1½ mm2 was cut out of the glutealmuscle and placed inside the lumen of the vein usinga straight irrigator. Next, the vein graft was connected tothe nerve stumps using six 10/0 Ethilon sutures (Ethicon,Johnson & Johnson, Amersfoort, the Netherlands) at eachcoaptation side (Figure 1). In all groups, the muscle andskin were closed using 6/0 Vicryl Rapide sutures (Ethicon,Johnson & Johnson, Amersfoort, the Netherlands).

Functional assessment

Sensory recovery was evaluated at 1, 3, 6 and 12 weeks’regeneration time, by means of stimulating three differentpoints along the lateral side of the foot (pinprick test).14,15

Reflexes were considered positive when the paw waswithdrawn. The toe-spread test is a commonly used test toassess motor function recovery.13,16 Toe-spread was eval-uated in the same sessions as the pinprick test. Voluntaryreaction of the toes was observed and graded on a 3-pointscale. An independent investigator, who was blinded to thetreatment of the rodents, performed both behaviouralstudies.

Table 1 Detailed overview of the FACS analysis for the analogu

Marker Fluorochrome Antibody (com

CD90 (stromal cells) PacBlue BioLegend, 20CD34 (hematopietic marker) PE SantaCruze, SCD31 (endothelial marker) PE Pharmingen, 5CD45RA (B lymphocytes) FITC Pharmingen, 5CD8a (NK cells, macrophages) PerCP BioLegend, 20CD3 (peripheral T-lymphocytes) APC Pharmingen, 5

Electrophysiology

The CMAP is a validated way of assessing nerve regenera-tion.17 We used a minimally invasive technique to evoke theCMAPs, as described in more detail elsewhere.18 Briefly, weemployed a 10e18 MHz probe (LA435) and an Esaote MyLab-Five ultrasound system (Esaote Europe, the Netherlands) tovisualise the nerve. The CMAPs that were subsequently eli-cited were recorded in batches of 25 responses and thenaveraged. These averages were quantified bymeans of signalonset latency, amplitude and duration.

Gastrocnemius muscle index

The Gastrocnemius Muscle Index (GMI) was measured toassess denervation atrophy at 6 and 12 weeks’ regenerationtime. Following the electrophysiological evaluation, bothmuscles were excised from the lower leg and the wetmuscle weight was measured immediately using a digitalscale to calculate the GMI.

Immunofluorescence staining

The conduits were divided into proximal, medial and distalparts. A nerve section from the contralateral nerve wastaken as control. All parts were snap-frozen in liquidnitrogen. Tissue sections were subsequently cut for 4-mmslides and fixed for 10 min in acetone. Next, the slides wereincubated with rabbit anti-rat NGF (Sigma) and S-100(Sigma) monoclonal antibody for 30 min. The binding ofprimary antibodies was detected using a fluorescent goatanti-rabbit immunoglobulin (FITC, Sigma). Slides weremounted in Vectashield mounting medium with DAPI andwere analysed using a camera (DFC 350 FX R2 v1.9.0; Leica,Wetzlar, Germany) connected to a microscope (DM5500B;Leica, Wetzlar, Germany), using imaging software LeicaApplication Suite Advanced Fluorescence (LASAF v2.0.0developed 1929, Leica, Wetzlar, Germany). The nerve areaswere captured in different magnifications (10�, 20� and40�). Before analysis, all the images were filtered exactlythe same way with the imaging program (Cell̂D v3.1 build1276, Olympus, Tokyo, Japan) to reduce noise and sharpenthem. Expression of the NGF and Schwann cell (S-100)marker immunoreactivity were assessed for staining inten-sity by three independent observers who were blinded tothe treatment. Intensity of staining was scored as absent(0), weak (1); moderate (2); strong (3) and powerful (4).The inter-rater reliability of the observers was evaluatedprior to the scoring, which gave a Cronbach’s Alpha of

e BMSC population.

pany, catalogue number) % Of positive cells in population

2521 49.33C7324 0.1055027 0.9054883 0.101712 0.1057030 0.03

Figure 2 Representative examples of multicolour FACS analysis of the different markers used for an analogue BMSC culture. Each marker was analysed by dot plot and histogram inorder to show the purity and characteristics of BMSC. Each histogram shows the shift of positively labelled BMSC population (red colour line) in comparison to unstained BMSC (bluecolour line). These data are derived from an analogue BMSC culture and not from the original injected cells.

254T.H

.J.Nijh

uis

etal.

Figure 3 Mean results of the Pinprick test at week 1, 3, 6,and 12 after grafting the three groups: Group I (Autograft),Group II (Muscle) and Group III (Muscle þ BMSC). Presence ofwithdrawal during stimulus above the ankle, the metatarsalarea, and at the level of the toes was graded 1, 2 and 3,respectively. No withdrawal during skin stimulation wasmarked with ‘0’. The error bars indicate the Standard Error ofthe Mean (SEM).

Figure 4 Mean results of the Toe-spread test at week 1, 3, 6,and 12 after grafting the three groups: Group I (Autograft),Group II (Muscle) and Group III (Muscle þ BMSC). No reaction:0 points. Any sign of toe movement: 1 point. Abduction of thetoes: 2 points. Extension and abduction: 3 points. The errorbars indicate the Standard Error of the Mean (SEM).

Veinemuscle graft supported with BMSCs for nerve reconstruction 255

0.973. Therefore, the scores of the three observers wereaveraged and used for statistical analysis. Double stainingof PKH (red) with neurotrophic factors was used to assessco-localisation of BMSC expression for NGF and Schwanncells (S-100) (both green), respectively.

Histology

To evaluate the presence of muscle in the vein graft, weexamined our frozen cross-sectional samples by haema-toxylin and eosin (H&E) staining.

Statistical analysis

Unpaired Student’s t-tests were used to determine statis-tical significance. p values of less than 0.05 were consid-ered significant. Spearman’s r was calculated to assess thecorrelation between the CMAP and the GMI.

Results

We lost two animals during surgery due to anaesthesiacomplications. One animal was lost in group I in the 6-weeksurvival group; the second animal was lost in group II in the12-week survival group.

FACS analysis

A summary of the quantitative FACS analysis of theanalogue cell culture from the Cleveland Clinic is presentedin Table 1 and in Figure 2.

The BMSCs expressed CD90 highly with a positive cellpopulation of 49.33%; the haematopoietic marker (CD34)and endothelial marker (CD31) were expressed in 0.10% and0.90% of the cells, respectively.

No presence of NK cells, macrophages, B-lymphocytesand peripheral T-lymphocytes was detected in the BMSCpopulation.

Pinprick test

Six weeks after surgery, there was little reaction tothe pinprick test and no sign of toe movement in mostanimals (Figure 3). There were no significant differencesbetween the three groups. Twelve weeks post-surgery,pinprick reaction had improved. Group I (nerve graft)had a significantly better reaction than groups II and III(vein grafts) (p � 0.004). Furthermore, group III (withBMSCs) tended to have better reactions than group II(p Z 0.081).

Toe-spread test

For all three groups, significant improvement of toe spreadover time was found (p < 0.001; Figure 4). The autograftresulted in a significantly better toe spread than the veingrafts at 12 weeks (p � 0.004). Group III performed worsethan group I (p Z 0.054), but better than group II(p Z 0.049).

Electrophysiology

After 6 weeks of recovery, the onset latency of the CMAP ingroup I was significantly longer than that in groups II and III(p Z 0.014 and p Z 0.029, respectively) (Table 2).Furthermore, in all three groups this latency was signifi-cantly different compared to the control value (p Z 0.001,p Z 0.014 and p Z 0.007, respectively). There were nodifferences in duration between groups or between thegroups and their controls. However, amplitude was signifi-cantly decreased in all three groups compared to thecontrol values (p < 0.001) and was significantly lower ingroup II than in group I (p Z 0.008).

Twelve weeks after surgery onset latency had normal-ised, but amplitude remained significantly decreased in all

Table 2 Mean CMAP latencies [ms], amplitudes [mV] and duration [ms] of the operated side at 6 and 12 weeks. Controlvalues e unoperated side (mean � SD).

6 weeks 12 weeks

Onset latency Peak latency Duration Amplitude Onset latency Peak latency Duration Amplitude

Group I 1.89 � 0.92 5.2 � 0.6 3.3 � 0.8 9.8 � 6.4 1.04 � 0.11 4.8 � 1.0 3.7 � 0.9 13.3 � 2.8Group II 1.06 � 0.11 4.8 � 0.7 3.7 � 0.6 9.7 � 2.6 1.13 � 0.24 3.9 � 0.6 2.9 � 0.6 10.0 � 2.8Group III 1.04 � 0.08 4.6 � 0.3 3.6 � 0.3 12.4 � 1.3 1.00 � 0.29 4.0 � 0.5 3.1 � 0.5 9.5 � 2.3Control 1.23 � 0.19 4.7 � 0.8 3.5 � 0.8 36.8 � 10.9 1.08 � 0.13 4.2 � 0.3 3.2 � 0.4 40.0 � 9.6

256 T.H.J. Nijhuis et al.

three groups compared to control values (p < 0.001).Amplitude was significantly higher in group I than in groupsII and III (p < 0.05). Both peak latency and duration in groupI were significantly longer than in group II and group III andwhen compared to the control values (p < 0.008). Betweengroups II and III, no differences were found.

GMI

In group I, GMI increased between 6 and 12 weeks(p < 0.001) whereas it decreased in groups II and III(p < 0.001). At 12 weeks, the GMI was significantly higherfor group I than for groups II and III (p < 0.001) (Figure 5).Moreover, there was a high correlation between CMAPamplitude and wet muscle weight (R Z 0.774, p < 0.001).

Immunofluorescence staining

NGFTypical NGF staining examples are depicted in Figure 6.Compared to the healthy control nerve (group IV) allsegments, except the middle section of group I, showedsignificantly increased staining intensity of NGF (p < 0.010).Twelve weeks after grafting, the NGF staining intensity wasdecreased in all groups compared to the 6 weeks sections(again with the exception of the middle section of group I).However, in all operated groups the staining was stillsignificantly more than in group IV (p < 0.050).

Figure 5 Results of the Gastrocnemius Muscle Index at 6 and12 weeks after grafting the three groups: Group I (Autograft),Group II (Muscle) and Group III (Muscle þ BMSC).

S-100Figure 7 provides typical examples of the sections taken. At6 weeks, the S-100 expression was significantly decreased inall segments except the distal segment in group II,compared to the healthy control nerve (group IV)(p < 0.050). Significant differences were found betweengroups in the proximal and middle sections, but not in thedistal segments. Twelve weeks post grafting, expression ofS-100 in the proximal and distal sections was similar in allgroups and all sections, with relatively low values only inthe middle section in group III.

BMSCs were clearly visible at both 6 and 12 weeks(Figure 8). More importantly, evaluation of the S-100staining revealed double staining in all sections in group III.

Histology

H&E stained sections did not show any remaining muscletissue at 6 and 12 weeks postoperatively.

Discussion

This study suggests that recovery following our modifiedtechnique, using a vein filled with a small fragment ofmuscle and injected with BMSCs, appears to have a benefi-cial effect when compared to the vein model with musclesupport only (group II). However, both conduits are not onpar with the recovery after grafting the defect using anautologous nerve graft (group I).

Differences in the recovery process between the groupsappear most pronounced in the GMI. These results suggestthat the nerve reconstructed with the nerve autograft hadalready re-innervated the muscle through newly formedaxons, whereas in the other two groups muscle re-innervation had not yet been established. Despite theincrease of atrophy between 6 and 12 weeks, the differ-ence 3 months postoperative between group II and group IIIindicates that re-innervation has been established and themuscle weight is increasing again, favouring group III.

The objective behavioural studies (i.e., the toe-spreadand pinprick assessments) showed a significant beneficialeffect of the BMSCs as a supportive cellular therapy.Although the power analysis predicted that e with thechosen group size e a significant effect could be measured,unfortunately, the GMI and CMAPs did not demonstratea significant effect. Another limitation of our study design isthe lack of a morphological analysis. Quantitative estima-tion of the number of myelinated nerve fibres remains a keyinvestigation tool in peripheral nerve research. However,

Figure 6 An overview of the NGF expression in the different sections (P Z proximal, M Z middle, D Z distal) at 6 and 12 weeksafter grafting groups: Group I (Autograft), Group II (Muscle), Group III (Muscle þ BMSC), Control (healthy control nerve). Afterstaining the tissue with NGF and Fit C (green staining), all the slides were mounted with vectamount medium with Dapi (bluestaining) for visualizing the epineurium and vessel wall. Pictures are taken with a 10� magnification.

Figure 7 An overview of the S100 expression in the different sections (P Z proximal, M Z middle, D Z distal) at 6 and 12 weeksafter grafting groups: Group I (Autograft), Group II (Muscle), Group III (Muscle þ BMSC), Control (healthy control nerve). Afterstaining the tissue with S100 and Fit C (green staining), all the slides were mounted with vectamount medium with Dapi (bluestaining) for visualizing the epineurium and vessel wall. Pictures are taken with a 10� magnification.

Veinemuscle graft supported with BMSCs for nerve reconstruction 257

Figure 8 Typical example of the red PKH-26 dye staining at 6 and 12 weeks post grafting in Group III (Muscle þ BMSC). Picturesare taken with a 20� magnification.

258 T.H.J. Nijhuis et al.

the enormous variability in assessments of the number ofaxons between different laboratories illustrates the limi-tations of this technique. Raimondo et al. summed at leastfive different sources of bias, relating to variability in thenumber of axons present and to counting errors introducedby the measurement approach.19 Based on these conclu-sions, we decided not to include this analysis in our study.

The staining for NGF and S-100 (a Schwann cell marker)in three sections of the reconstructed nerve at two timepoints allowed us to evaluate the regeneration progresshistologically. One of the findings was that in the regener-ation process an increased expression of NGF was encoun-tered. NGF is critical in the regeneration, survival andmaintenance of neurons.20 Moreover, as such, aftera complete axotomy and reconstruction has taken place anincreased expression of NGF is expected first proximal. Thishypothesis was confirmed by our results, showing analready normal expression of the autograft group in theproximal and medial sections and an increased expressionin the distal section. Three months after reconstruction,a normalised expression of NGF was encountered in allsections. Consequently, the results demonstrated a benefi-cial effect of the BMSCs especially in early regeneration.Not only could we still find the labelled BMSCs inside thevein, but e more importantly e we also found a strongerstaining intensity of NGF and S-100 in group III compared togroup II, 6 weeks postoperatively. Twelve weeks post-operatively, group I revealed the least NGF expression. Asdiscussed above, possibly, the newly formed axonal sproutsin the autologous donor nerve had already progressedbeyond the evaluated sections. S-100 expression, predom-inantly present in differentiated Schwann cells, was similarto the healthy control nerve in the proximal and distalsection 12 weeks postoperatively. The middle section,however, continued to show significantly less expression ingroup II compared to the other two groups, which might bean indication of an incomplete regeneration of the Schwanncell population at this site in the graft.

Regarding the FACS analysis, a high expression of CD90(i.e., 49.33%) was encountered. CD90 is reported in theliterature as predominant in the BMSC population andcharacteristic for cellular adhesion and migration.21e24 Inaddition, as reported by Polisetti (2010), BMSCs do notexpress haematopoietic stem cell markers such as CD34.23

The lack of CD34 expression in the BMSC population was

confirmed (only 0.1% cells carried CD34). Consequently, theanalysis showed that only 0.9% cells were carrying theendothelial marker CD31. This observation was also re-ported during investigation of the therapeutic effects ofBMSC, for example by Corcos et al.21

In the literature, two explanations for the contributionof BMSCs to nerve regeneration are provided. The first isthat BMSCs may produce cytokines and growth factors thatpositively impact neural cell survival.8,25e33 The secondexplanation is that BMSCs can differentiate into neurallineages including neurons, astrocytes, oligodendrocytes,microglia and, most important in this context, Schwanncells.26,34,35 In line with the work of Tohill et al.,32e34 ourresults provide support for the latter possibility, becausethe reaction of the labelled BMSCs to the S-100 stainingshows that they can have the capability to trans-differentiate into cells with a Schwann cell-like phenotype.A similar finding has been published by Tohill et al., whofound mesenchymal cells differentiating into a Schwanncell-like phenotype,36 and later demonstrated that differ-entiated mesenchymal stem cells can have the samecharacteristics as Schwann cells in rats.37,38

Regarding the muscle fragment, our H&E stainingconfirmed that the muscle became atrophic. Hence, thesupportive role could be explained by the remainingmuscle’s extracellular matrix, which was also confirmed byanother group.39

The fact that we did not find significant beneficialeffects of the BMSCs implies that, for the future, this modelunder investigation is not the ideal reconstruction method.However, the limited supply of available nerve graftsremains a problem and therefore we must continue thesearch for alternative interposition conduits.

Conflict of interest

None.

Acknowledgements

The authors are grateful to Ineke Hekking for practicalsuggestions and assisting with the surgery and anaesthesia.The authors would like to thank Dr. James Phillips for theconstructive advice regarding the histological analysis.

Veinemuscle graft supported with BMSCs for nerve reconstruction 259

References

1. J IJ-P, Jansen K, Gramsbergen A, Meek MF. Transection ofperipheral nerves, bridging strategies and effect evaluation.Biomaterials 2004;25:1583e92.

2. Matsuyama T, Mackay M, Midha R. Peripheral nerve repair andgrafting techniques: a review. Neurol Med Chir (Tokyo) 2000;40:187e99.

3. Battiston B, Tos P, Cushway TR, Geuna S. Nerve repair bymeans of vein filled with muscle grafts I. Clinical results.Microsurgery 2000;20:32e6.

4. Benito-Ruiz J, Navarro-Monzonis A, Piqueras A,Baena-Montilla P. Invaginated vein graft as nerve conduit: anexperimental study. Microsurgery 1994;15:105e15.

5. Chiu DT, Janecka I, Krizek TJ, Wolff M, Lovelace RE.Autogenous vein graft as a conduit for nerve regeneration.Surgery 1982;91:226e33.

6. Kelleher MO, Al-Abri RK, Eleuterio ML, Myles LM, Lenihan DV,Glasby MA. The use of conventional and invaginated autologousvein grafts for nerve repair by means of entubulation.Br J Plast Surg 2001;54:53e7.

7. Rice DH, Berstein FD. The use of autogenous vein for nervegrafting. Otolaryngol Head Neck Surg 1984;92:410e2.

8. Zhang F, Blain B, Beck J, et al. Autogenous venous graft withone-stage prepared Schwann cells as a conduit for repair oflong segmental nerve defects. J Reconstr Microsurg 2002;18:295e300.

9. Battiston B, Tos P, Geuna S, Giacobini-Robecchi MG,Guglielmone R. Nerve repair by means of vein filled withmuscle grafts. II. Morphological analysis of regeneration.Microsurgery 2000;20:37e41.

10. Brunelli GA, Battiston B, Vigasio A, Brunelli G, Marocolo D.Bridging nerve defects with combined skeletal muscle and veinconduits. Microsurgery 1993;14:247e51.

11. Battiston B, Tos P, Conforti LG, Geuna S. Alternative tech-niques for peripheral nerve repair: conduits and end-to-sideneurorrhaphy. Acta Neurochir Suppl 2007;100:43e50.

12. Cui L, Jiang J, Wei L, et al. Transplantation of embryonic stemcells improves nerve repair and functional recovery after severesciatic nerve axotomy in rats. Stem Cells 2008;26:1356e65.

13. Nijhuis TH, Brzezicki G, Klimczak A, Siemionow M. Isogenicvenous graft supported with bone marrow stromal cells asa natural conduit for bridging a 20 mm nerve gap. Microsurgery2010;30(8):639e45.

14. Siemionow M, Demir Y, Mukherjee AL. Repair of peripheralnerve defects with epineural sheath grafts. Ann Plast Surg2010;65:546e54.

15. Dai Y, Kondo E, Fukuoka T, Tokunaga A, Miki K, Noguchi K. Theeffect of electroacupuncture on pain behaviors and noxiousstimulus-evoked Fos expression in a rat model of neuropathicpain. J Pain 2001;2:151e9.

16. Ozmen S, Ayhan S, Latifoglu O, Siemionow M. Stamp and papermethod: a superior technique for the walking track analysis.Plast Reconstr Surg 2002;109:1760e1.

17. Werdin F, Grussinger H, Jaminet P, et al. An improved elec-trophysiological method to study peripheral nerve regenera-tion in rats. J Neurosci Methods 2009;182:71e7.

18. Nijhuis TH, Smits ES, van Neck JW, et al. Ultrasound-guidedneedle positioning near the sciatic nerve to elicit compoundmuscle action potentials from the gastrocnemius muscle of therat. J Neurosci Methods 2010;194(2):283e6.

19. Raimondo S, Fornaro M, Di Scipio F, Ronchi G, Giacobini-Robecchi MG, Geuna S. Chapter 5: methods and protocols inperipheral nerve regeneration experimental research: part II-morphological techniques. Int Rev Neurobiol 2009;87:81e103.

20. Freeman RS, Burch RL, Crowder RJ, et al. NGF deprivation-induced gene expression: after ten years, where do westand? Prog Brain Res 2004;146:111e26.

21. Corcos J, Loutochin O, Campeau L, et al. Bone marrowmesenchymal stromal cell therapy for external urethralsphincter restoration in a rat model of stress urinary inconti-nence. Neurourol Urodyn 2011;30:447e55.

22. TanQ, Lui PP, Rui YF,Wong YM. Comparison of potentials of stemcells isolated from tendon and bonemarrow formusculoskeletaltissue engineering. Tissue Eng Part A 2012;18:840e51.

23. Polisetti N, Chaitanya VG, Babu PP, Vemuganti GK. Isolation,characterization and differentiation potential of rat bonemarrow stromal cells. Neurol India 2010;58:201e8.

24. Zhao Y, Guan Y, Xu Y, Li Y, Wu W. Sodium Ferulate combinedwith bonemarrow stromal cell treatment ameliorating rat brainischemic injury after stroke. Brain Res 2012;1450:157e65.

25. Caplan AI, Dennis JE. Mesenchymal stem cells as trophicmediators. J Cell Biochem 2006;98:1076e84.

26. Chen X, Wang XD, Chen G, Lin WW, Yao J, Gu XS. Study ofin vivo differentiation of rat bone marrow stromal cells intoSchwann cell-like cells. Microsurgery 2006;26:111e5.

27. Chopp M, Li Y. Treatment of neural injury with marrow stromalcells. Lancet Neurol 2002;1:92e100.

28. Lu L, Chen X, Zhang CW, et al. Morphological and functionalcharacterization of predifferentiation of myelinating glia-likecells from human bone marrow stromal cells through activa-tion of F3/notch signaling in mouse retina. Stem Cells 2008;26:580e90.

29. Munoz-Elias G, Woodbury D, Black IB. Marrow stromal cells,mitosis, and neuronal differentiation: stem cell and precursorfunctions. Stem Cells 2003;21:437e48.

30. Neuhuber B, Timothy Himes B, Shumsky JS, Gallo G, Fischer I.Axon growth and recovery of function supported by humanbone marrow stromal cells in the injured spinal cord exhibitdonor variations. Brain Res 2005;1035:73e85.

31. Suzuki H, Taguchi T, Tanaka H, et al. Neurospheres inducedfrom bone marrow stromal cells are multipotent for differen-tiation into neuron, astrocyte, and oligodendrocyte pheno-types. Biochem Biophys Res Commun 2004;322:918e22.

32. Wislet-Gendebien S, Hans G, Leprince P, Rigo JM, Moonen G,Rogister B. Plasticity of cultured mesenchymal stem cells:switch from nestin-positive to excitable neuron-like pheno-type. Stem Cells 2005;23:392e402.

33. Woodbury D, Schwarz EJ, Prockop DJ, Black IB. Adult rat andhuman bone marrow stromal cells differentiate into neurons. JNeurosci Res 2000;61:364e70.

34. Cuevas P, Carceller F, Dujovny M, et al. Peripheral nerveregeneration by bone marrow stromal cells. Neurol Res 2002;24:634e8.

35. Cuevas P, Carceller F, Garcia-Gomez I, Yan M, Dujovny M. Bonemarrow stromal cell implantation for peripheral nerve repair.Neurol Res 2004;26:230e2.

36. Tohill M, Mantovani C, Wiberg M, Terenghi G. Rat bone marrowmesenchymal stem cells express glial markers and stimulatenerve regeneration. Neurosci Lett 2004;362:200e3.

37. Caddick J, Kingham PJ, Gardiner NJ, Wiberg M, Terenghi G.Phenotypic and functional characteristics of mesenchymalstem cells differentiated along a Schwann cell lineage. Glia2006;54:840e9.

38. Mahay D, Terenghi G, Shawcross SG. Growth factors inmesenchymal stem cells following glial-cell differentiation.Biotechnol Appl Biochem 2008;51:167e76.

39. Pagnotta A, Tos P, Fornaro M, Gigante A, Geuna S, Battiston B.Neurotrophins and their receptors in early axonal regenerationalongmuscle-vein-combinedgrafts.Microsurgery2002;22:300e3.