Developing a Porcine Model for Study of Vocal

Fold Scar

Gayle Woodson, Springfield, Illinois

Summary: The porcine larynx is very similar in size and structure to that of humans, and wound healing in pigs is very

AccepFrom t

sity SchoAddre

yngologyfield, ILJourna0892-1� 201doi:10

similar to that of humans. However, the pig is not often used in vocal fold scar research because it is difficult to view thevocal folds endoscopically. To further assess the pig as a model for studying vocal scar, we compared the plane of sur-gical dissection in the mucosa of four porcine vocal folds with that in eight human cadaver larynges. The plane of dis-section was quite similar in porcine and human larynges, occurring within the loose layer of the superficial laminapropria. We also compared healing of porcine vocal folds after elevation and replacement of an epithelial flap versusexcision of epithelium, leaving an open wound. After 6 weeks, larynges were harvested for histologic examination.There was no significant difference between the mucosa of the normal vocal fold and that of the healed microflap. How-ever, after healing of epithelial excision, there was a depressed scar, with average lamina propria thickness of 302 mmversus 864 mm for the normal fold (P < 0.05). Finally, to document that the mucosal wave can be evaluated in the porcinelarynx, we developed a preparation that removes the false vocal folds, to allow ex vivo phonation. Experimentally cre-ated scar in the porcine larynx is a favorable model for the study of vocal fold healing and for assessment of treatmentsfor vocal fold scar.Key Words: Larynx–Vocal fold–Scar–Animal model–Pig–Wound healing.

INTRODUCTION

Vocal fold scar is one of the most challenging clinical problemscurrently encountered in laryngology. Many treatment ap-proaches have been used, but no treatment reliably rehabilitatesand restores the voice. Clinical study of the pathogenesis andtreatment of vocal fold scar is very difficult. Vocal scar is notcommon, and consequently, in any study, treatment groupsare small. Additionally, the defects are varied. Sometimesa scarred vocal fold is grossly normal, but stroboscopy revealsdiffuse stiffness, adynamic segments, or asymmetric mucosalwave. In other patients, the scar is a divot that results in a glottalgap on phonation. Scar may also result in a loss of vocal foldvolume or a bulky scar that impairs glottic closure. Becauseof low incidence and heterogeneous pathology, statistical eval-uation of treatment outcomes is difficult.1

Animal models offer the opportunity for controlled experi-ments to create standardized injuries, study wound healing,and quantify the effects of different therapeutic interventions.Vocal fold scar has been studied in rabbits, dogs, and rats.The dog has been a preferred model because of the ease of view-ing the larynx in vivo to assess vocal fold vibration.2 However,the structure of the canine vocal fold differs significantly fromthat of humans. Although no animal studied to date has a vocalligament such as that seen in humans, the microscopic architec-ture of the pig is more similar to that of humans than that ofdogs, rabbits, or rats.

The human vocal fold is uniquely structured for phonation:the epithelium is separated from underlying muscle and liga-

ted for publication March 16, 2012.he Department of Surgery, Division of Otolaryngology, Southern Illinois Univer-ol of Medicine, Springfield, Illinois.ss correspondence and reprint requests to Gayle Woodson, Division of Otolar-, Southern Illinois University School of Medicine, P.O. Box 19662, Spring-62794-9662. E-mail: gwoodson@siumed.edul of Voice, Vol. 26, No. 6, pp. 706-710997/$36.002 The Voice Foundation.1016/j.jvoice.2012.03.003

ment by specialized lamina propria in which tissue density isstratified to allow the epithelium to vibrate with greater ampli-tude than deeper structures.3 Collagen density is sparse in thesuperficial lamina propria and increases with depth into thevocal fold.4 The lamina propria is a continuous layer about1 mm thick that wraps around the free edge of the vocal fold.Epithelium can be easily stripped from the human vocal fold,and the plane of separation in humans occurs at a consistentlevel within the lamina propria.5 But in the canine larynx, thepattern of collagen distribution is inverted. Density is greatestin the superficial lamina propria, near the epithelium, anddecreases toward the underlying muscle.4 The shape of thecanine vocal fold is quite different from that of humans. Thelayer of lamina propria is much deeper and projects into thelumen as a long shelf, rather than wrapping around the edgeof the vocalis muscle.4 And when epithelium is removed fromthe vocal fold, the level at which tissue shears off is unpredict-able.2 Thus, the structure of the canine vocal fold differs signif-icantly from that of humans.The structure of the pig vocal fold is quite similar to that of

humans.4 As in humans, the lamina propria is approximately1 mm thick along the medial edge of the vocal fold. Collagenis loosely organized in the superficial lamina propria andmore densely concentrated at deeper levels.2 In addition, thereis extensive literature indicating that wound healing in pigs isvery similar to that in humans, making the pig an attractivemodel for the study of therapeutic interventions for wound heal-ing and scar.4 A recent review of 180 published studies testing25 skin wound therapies concluded that results in the pig havethe closest concordance with humans.6 Anatomically and phys-iologically, porcine skin is quite similar to that of humans. Por-cine dermal collagen is biochemically quite similar to humandermal collagen, so much so that porcine collagen is used ther-apeutically in human wound products.7 Porcine skin and humanskin also have very similar concentration and distribution ofseveral antigens including keratins 16 and 10, filaggrin,

Gayle Woodson Porcine Vocal Scar 707

collagen IV, fibronectin, and vimentin.8 Physiologically, heal-ing of skin wounds in humans and pigs is similar. Althoughwounds in most small animals occur primarily through contrac-ture, partial thickness wounds in humans and swine heal largelythrough re-epithelialization.9 Given the parallels of porcine andhuman skin, it is not unreasonable to assume that the vocal foldmucosa would also have biological similarity, making the piga good model for the study of vocal fold healing.

As mentioned above, the major objection that has previouslybeen cited against the pig as an animal model for vocal fold scaris that it is very difficult to view the vocal folds with laryngos-copy. The false vocal folds in pigs are very prominent andobscure the true vocal folds. Thus, it is not possible to viewthe mucosal wave in vivo to document the functional effectsof vocal fold scar and the efficacy of treatment. Therefore, wedeveloped a protocol for assessing function by ex vivo phona-tion in the excised larynx, by debulking the false vocal folds.

The present study was intended to demonstrate the feasibilityand value of the pig as a model for studying vocal fold scar andthe potential for testing the efficacy of surgical and medicaltreatment of vocal fold scar. The objectives of the experimentswere to (1) compare the histology and surgical dissection planein human and porcine vocal folds, (2) establish a protocol forcreating experimental scar in the porcine larynx, and (3) docu-ment the feasibility of assessing the mucosal wave in the ex-cised porcine larynx.

METHODS

Dissection plane in cadaver larynges

Four fresh porcine cadaver larynges were bisected in the mid-line sagittal plane. In one half of each larynx, an epithelialflap was developed with microscissors and excised from thevocal fold edge. The opposite fold was left intact as a control.

Epithelium was also excised from one vocal fold in each ofseven human cadaver larynges, using microscissor dissection.In four larynges, dissection was as superficial as possible, andthe tips of the scissors were oriented toward the surface, withthe goal of separating the epithelium from the underlying lam-ina propria. In three larynges, the tips of the scissors weredirected deeply, with the goal of separating the lamina propriafrom the vocal ligament. As with the porcine larynges, onehemilarynx was left intact as a control.

Each hemilarynx was decalcified, fixed in 10% formalin, andembedded in paraffin. Coronal sections of 0.7 mm were pre-pared and stained alternately withH&EorVerhoeff-vanGieson.In addition, the excised mucosa was similarly fixed and stained.Slides were digitally imaged and then processed with Image J(NIH) to measure the thickness of the lamina propria on the ex-cised epithelium, the larynx after excision of epithelium, and inthe intact specimen.

Vocal fold healing

The study was approved by Animal Care Committee. Subjectswere eight minipigs. Under general anesthesia, a vertical mid-line incision was made to in the anterior neck of each pig toexpose the larynx and trachea. A small tracheal cannula was

placed for ventilation during the procedure. The larynx wasopened by amidline thyrotomy to expose the vocal folds. A lon-gitudinal incision was made laterally on the superior aspect ofthe left vocal mucosa. Mucosa was elevated from the vibratoryedge by blunt dissection in the subepithelial plane. In fouranimals, this flap was excised. In the other four animals, themedially based flap was replaced and secured with 5-0 chromicsutures. In all animals, the thyroid cartilage was reapproxi-mated, carefully aligning the anterior commissure, and thetracheal cannulawas removed.All pigs had a good laryngeal air-way at the end of the procedure and tolerated a normal diet afterrecovery from anesthesia. However, two pigs died a few daysafter surgery because of stress-related pulmonary edema. Thestudy was completed in six pigs.

Thirty days after surgery, the animals were sacrificed. Laryn-ges were excised and bisected in a sagittal plane. Each hemilar-ynx was grossly inspected and photographed, then decalcified,fixed in 10% formalin, and embedded in paraffin. Coronal sec-tions of 0.7 mm were prepared from the anterior, middle, andposterior third of each vocal fold. Alternate sections werestained with H&E or Verhoeff-van Gieson technique.

Slides were examined for collagen and elastin fiber distribu-tion and density, comparing the experimental fold with the con-trol side. Computerized image (NIH Scion) analysis was used tomeasure mucosal thickness, from the free edge of the vocal foldto the outer surface of the vocalis muscle. Data were analyzedusing a paired student t test, comparing values on the experi-mental and control sides.

Feasibility of ex vivo phonation in excised porcine

larynges

Two fresh porcine larynges were prepared for ex vivo phona-tion. For each larynx, approximately 4 cm of trachea were pre-served in continuity with the larynx. The thyroid cartilage wasresected above the glottis, along a line beginning posteriorlyabove the arytenoid cartilages, and angling anteriorly and cau-dally to a point just above the anterior commissure. The epiglot-tis and submucosal tissue of the false vocal folds were excised.The mucosa of the false folds was then sutured laterally to thethyroid cartilage, exposing the superior surface of the vocalfolds. Bilateral arytenoid adduction sutures were placed toapproximate the vocal processes. The larynx was mounted ona platform, and a polyethylene tube was inserted into the distaltrachea and secured with circumferential plastic ligatures.A stroboscopic video camera was mounted above the larynxto view the vocal folds. Phonation was elicited when com-pressed air was passed through the larynx with variable lateralcompression of the thyroid cartilages, and mucosal vibrationwas recorded.

RESULTS

Dissection plane

In all four porcine larynges, and all human larynges, the planeof dissectionwaswithin the lamina propria.Histologic examina-tion of the excised epithelium always included a portion of thesuperficial lamina propria along its deep surface (Figure 1A).

FIGURE 1. Epithelial strip excised from vocal fold mucosa by blunt

microscissor dissection. Note attached portion of superficial lamina

propria (103). A. Porcine epithelium. B. Human cadaver epithelium.

Journal of Voice, Vol. 26, No. 6, 2012708

Thiswas also true in both groups of human larynges (Figure 1B).Figure 2 displays the measurements of the thickness of the ex-cised specimens. Because there was no difference between thetwo groups of human larynges, that data are combined. The av-erage thickness of the excised strip was 145 m in the human lar-ynx and 194 m in the porcine larynx.

FIGURE 2. Bar graph comparing thickness of excised epithelium

from excised porcine and cadaver human larynges.

Histologic examination of the residual laryngeal specimensafter excision of epithelium always revealed a layer of thesuperficial lamina propria. However, the amount of tissue couldnot be quantified by measuring the thickness of this layer, asthere was distortion at the interface, with distraction of the col-lagen fibers and patchy separation of lamina propria fromunderlying soft tissue (Figure 3).

Vocal fold healing

Gross inspection. After healing of an epithelial flap, thegross appearance of the operated vocal fold was not noticeablydifferent from that of the unoperated side. The surface of theepithelium was smooth and even, and the vibratory edge wassharp. However, after resection of vocal fold epithelium, heal-ing resulted in a depressed scar, with an uneven surface anda rounded edge.

Histologic analysis. In the vocal folds that had healed afterreplacement of a microflap, the edge of the vocal fold, as seen incoronal section, was similar to the control vocal folds(Figure 4A and B). However, vocal folds that had healed by sec-ondary intention had a rounded edge, with very thin laminapropria (Figure 4C). Table 1 displays the thickness of the mu-cosa in the specimens. The average thickness of the mucosain the control vocal folds was 919.15 m. After replacement ofa microflap, mucosal thickness was 879.86 m, compared withan average thickness of 919 m. This difference was not signifi-cantly significant. In vocal folds that healed by secondary inten-tion, the average thickness of mucosa was 301.95 m,significantly thinner than the thickness in the control vocal folds(864.33), P < 0.05. The concentration of collagen was some-what increased in the lamina propria of the scarred vocal foldsbut not appreciably different in the vocal fold that healed afterelevation and replacement of an epithelial flap.

Ex vivo phonation in excised larynges

The resection of supraglottic cartilage and retraction of the falsevocal folds provided a good view of the vocal fold when viewedfrom above (Figure 5). Phonation could be produced by apply-ing medial compression to both thyroid ala as air was passedthrough the trachea across the glottis, and the resultant glotticwave was visible during stroboscopic illumination.

DISCUSSION

In both porcine and human vocal folds, the excised epitheliumalways contained a portion of the superficial lamina propria.This is consistent with a study by Gray et al,5 who used bluntinstruments to dissect the lamina propria in eight human laryn-ges. The plane of dissection was consistently within the laminapropria. This strong attachment of the epithelium to underlyinglamina propria in humans has been attributed to anchoringfibers, analogous to the fibers in skin that secure the epidermisto dermis. Anchoring fibers are abnormal in blistering skin dis-orders, such as epidermolysis bullosa. The vocal fold has nodermis; however, electron microscopy of the basement mem-brane zone of the vocal fold showed a similar structure to thatof the epidermal dermal interface.10 Collagen fibers of the

FIGURE 3. Coronal sections through larynges after excision of epithelium by blunt dissection (23). Note fragmentation of residual lamina propria

and separation from underlying tissues. A. Porcine larynx. B. Human cadaver larynx.

Gayle Woodson Porcine Vocal Scar 709

superficial lamina propria pass through loops created by an-choring fibers that originate in the lamina densa of the basementmembrane.

It is likely that porcine vocal fold mucosa has a similar ultra-structure, given the remarkable similarity of porcine and humanskin. Pigskin is anatomically very similar to human skin, withthick and well-developed rete pegs and dermal papillary bodies.There are also similarities in keratinous proteins and lipid com-position of the stratum corneum.11–13

Although the superficial lamina propria vocal fold epithe-lium seems firmly anchored to epithelium, this attachment totissue is easily disrupted. The results of this study confirmedwhat has long been recognized clinically: that epithelium canbe easily stripped from the vocal fold. It has also been observedthat after stripping, the vocal fold tends to heal with scarring,and the vibratory edge becomes stiff. In this study, the histology

FIGURE 4. Coronal sections through porcine hemilarynx. A. Control. B.

month after excision of epithelium, with healing by secondary intention.

of both human and porcine specimens after removal of epithe-lium showed that the residual lamina propria was disrupted anddistorted during this process, with dispersion of fibers and sep-aration from the underlying muscle (Figure 3). In the porcinevocal folds that healed after removing epithelium, the areahealed by re-epithelialization, there was very little subepithelialtissue, indicating that most of the residual lamina propria hadsloughed or undergone necrosis (Figure 4C). However, whenthe flap was replaced, healing resulted in a nearly normal layerof lamina propria, indicating that the residual lamina propriawas either preserved or had regenerated.

This porcine model of scar provides support for the benefitsof the epithelial preservation in laryngeal microsurgery. Thelaryngeal microflap technique was developed based on recogni-tion of the layered structure of the vocal fold mucosa. Toremove submucosal pathology, the epithelium is elevated and

One month after elevation and replacement of epithelial flap. C. One

TABLE 1.

Mean Thickness of Vocal Fold Mucosa

Microflap Replaced (P > 0.05) Excised (P < 0.05)

Surgical side (m) 879.86 301.95Control side (m) 919.15 864.34Difference (m) 39.29 562.39

Journal of Voice, Vol. 26, No. 6, 2012710

then replaced at the end of the procedure. This theoreticallycovers and preserves the lamina propria.14 Although microflapsurgery has better results than leaving a wound to close second-arily, Sataloff et al15 have pointed out that the vocal fold is nottotally normal following healing of a microflap and presenteda case series showing that vocal function is better after the‘‘micromini-flap’’ technique, which limits epithelial elevationto the smallest possible area. Although Sataloff et al proposethat the impaired function after microflap surgery is becauseof disruption of the basement membrane, the results from thehuman specimens in this study suggest that the basement mem-brane is robust and not easily disrupted by surgery. Although itis true that the pig lacks a vocal ligament, similarities in the dis-section plane of the lamina propria suggest that the pig is a goodmodel for the study of vocal fold healing and for the study oftherapeutic interventions. In particular, the model would be use-ful for the study of procedures to restore the lamina propria inscarred vocal folds and to rehabilitate the larynx after endo-scopic resection of laryngeal cancers.

FIGURE 5. Excised porcine larynx, prepared for ex vivo phonation,

viewed from above. Thyroid cartilage has been removed above level of

vocal folds. Epiglottis removed. Submucosal fat removed from false

vocal folds and mucosa sutured laterally to expose vocal folds.

SUMMARY

The pig is a very logical choice for the study of vocal fold scar, asthe size of the larynx and histologic structure of thevocal foldmu-cosa of the pig is more similar to humans than that of other ani-mals studied. Additionally, the process of wound healing inpigs is very similar to that of humans. The current experiments in-dicate that the dissection plane in the lamina propria is very sim-ilar in the porcine and human larynx. This may be because ofsimilarities in the ultrastructure of the basement membranezone. Experimental scar can be created via a laryngofissureapproach by excising epithelium. Elevation and replacement ofa microflap has little impact on the structure of the vocal fold.An openvocal foldwound heals by secondary intention, resultingin a depressed scar,with loss of lamina propria. This experimentalscar can be used in the future to study therapeutic interventions.Phonatory function can be studied in the excised porcine larynx.We conclude that the pig is an excellent model for the study

of vocal fold scar and has excellent potential for assessing theefficacy of treatment.

REFERENCES1. Hansen JK, Thibeault SL. Current understanding and review of the litera-

ture: vocal fold scarring. J Voice. 2006;20:110–120.

2. Garrett CG, Coleman JR, Reinisch L. Comparative histology and vibration

of the vocal folds: implications for experimental studies in microlaryngeal

surgery. Laryngoscope. 2000;110:814–824.

3. Hirano M. Morphological structure of the vocal cord as a vibrator and its

variations. Folia Phoniatr (Basel). 1974;26:89–94.

4. Kurita S, Nagata K, HiranoM. A comparative study of the layer structure of

the vocal fold. In: Bless DM, Abbs JH, eds. Vocal Fold Physiology:

Contemporary Research and Clinical Issues. San Diego, CA: College-

Hill Press; 1983:3–21.

5. Gray SD, Chan KJ, Turner B. Dissection plane of the human vocal fold lam-

ina propria and elastin fibre concentration. Acta Otolaryngol. 2000;120:

87–91.

6. Sullivan TP, Eaglstein WH, Davis SC, Mertz P. The pig as a model for

human wound healing. Wound Repair Regen. 2001;9:66–76.

7. HeinrichW, Lange PM, Stirz T, Iancu C, Heidermann E. Isolation and char-

acterization of the large cyanogen bromide peptides from burns skin graft

donor sites. J Burn Care Rehabil. 1996;17:246–251.

8. Wollina U, Berger U, Mahrle G. Immunohistochemistry of porcine skin.

Acta histochem. 1991;90:87–91.

9. Hartwell SW. The Mechanism of Healing in Human Wounds. Springfield,

IL: Charles Thomas; 1955.

10. Gray SD, Pignatari SN, Harding P. Morphologic ultrastructure of anchoring

fibers in normal vocal fold basement membrane zone. J Voice. 1994;8:

48–52.

11. Weinstein GD. Comparison of turnover time and of keratinous protein frac-

tions in swine and human epidermis. In: Bustad LK, McClellan RD, eds.

Swine in Biomedical Research. Seattle, WA: Frayn; 1966:287–289.

12. Nicolaides N, Fu HC, Rice GR. The skin surface lipids of man compared

with those of eighteen species of animals. J Invest Dermatol. 1968;51:

83–89. 15:886–9.

13. Gray GM, White RJ, Williams RH, Yardley HJ. Lipid composition of the

superficial stratum corneum of pig epidermis. Br J Dermatol. 1982;106:

59–63.

14. Zeitels SM, Hillman RE, Bunting GW, Vaughn T. Reinke’s edema: phona-

tory mechanisms and management strategies. Ann Otol Rhinol Laryngol.

1996;105:533–543.

15. Sataloff RT, Spiegel JR, Heuer RJ, et al. Laryngeal mini-microflap: a new

technique and reassessment of the microflap saga. J Voice. 1995;9:

198–204.