The LaryngoscopeVC 2010 The American Laryngological,Rhinological and Otological Society, Inc.

Characterization of Mesenchymal Stem CellsFrom Human Vocal Fold Fibroblasts

Summer E. Hanson, MD; Jaehyup Kim, MD; Beatriz H. Quinchia Johnson, DDS, PhD;

Bridget Bradley; Melissa J. Breunig; Peiman Hematti, MD; Susan L. Thibeault, PhD

Objectives/Hypothesis: Mesenchymal stemcells (MSCs) originally isolated from bone marrow(BM), are fibroblast-looking cells that are nowassumed to be present in the stromal component ofmany tissues. MSCs are characterized by a certainset of criteria, including their growth culture charac-teristics, a combination of cell surface markers, andthe ability to differentiate along multiple mesenchy-mal tissue lineages. We hypothesized that humanvocal fold fibroblasts (hVFF) isolated from the laminapropria meet the criteria established to define MSCsand are functionally similar to MSCs derived fromBM and adipose tissue.

Study Design: In vitro study.Methods: hVFF were previously derived from

human vocal fold tissues. MSCs were derived fromadipose tissue (AT), and BM of healthy donors basedon their attachment to culture dishes and their mor-phology and expanded in culture. Cells were ana-lyzed for standard cell surface markers identified onBM-derived MSCs and the ability to differentiateinto cells of mesenchymal lineage (i.e., fat, bone, andcartilage). We investigated the immunophenotype ofthese cells before and after interferon-c (INF-c)stimulation.

Results: hVFF displayed cell surface markersand multipotent differentiation capacity characteristicof MSCs. Furthermore, these cells exhibited similarpatterns of expression of human leukocyte antigenand costimulatory molecules, after stimulation withINF-c compared to MSCs derived from BM and AT.

Conclusions: Based on our findings, hVFFderived from lamina propria have the same cell sur-face markers, immunophenotypic characteristics, anddifferentiation potential as BM- and AT-derivedMSCs. We propose that vocal fold fibroblasts areMSCs resident in the vocal fold lamina propria.

Key Words: Mesenchymal stem cells, humanvocal fold stem cells, immune modulation, vocal foldfibroblasts.

Laryngoscope, 120:546–551, 2010

INTRODUCTIONOver the last several years, much attention has

been directed at the potential of cell therapy in regener-ative medicine, specifically mesenchymal stem cells(MSCs). Mesenchymal stem cell therapy involves thetransplantation of autologous or allogeneic cells topatients through local delivery or systemic infusion andoffers a novel approach for tissue repair, augmentation,or reconstruction. This new era of cell-based therapiesstems from the observation that MSCs, originally iso-lated from bone marrow (BM), possess the ability todifferentiate along multiple tissue lineages, participatein the tissue repair and regeneration process through avariety of paracrine mechanisms, and suppress activa-tion and proliferation of immune and inflammatorycells.1 It is now thought that these stem cells can be iso-lated from most tissues of the body, including fat, andthat MSCs isolated from these diverse tissues possesssimilar biological characteristics, differentiation poten-tial, and immunological properties.2

The vocal fold lamina propria is composed of a com-bination of extracellular matrix (ECM) components thataccount for not only its unique biomechanical character-istics but also challenges that are not adequatelymatched by current replacement modalities.3,4 Some ofthe potential treatment modalities proposed/reported forvocal fold scar include injection of biomaterials/

From the Division of Plastic and Reconstructive Surgery, School ofMedicine and Public Health, University of Wisconsin–Madison (S.E.H.); theDivision of Otolaryngology–Head and Neck Surgery, University ofWisconsin–Madison (B.H.Q.J., S.L.T.); the Department of Medicine, Schoolof Medicine and Public Health, University of Wisconsin–Madison (J.K. P.H.);the College of Letters and Sciences, University of Wisconsin–Madison(B.B.); the College of Agriculture and Life Sciences, University ofWisconsin-Madison (M.J.B.) ; and the University of Wisconsin CarboneCancer Center (P.H., S.L.T.), Madison, Wisconsin, U.S.A.

Editor’s Note: This Manuscript was accepted for publicationNovember 3, 2009.

This work was supported in part by NIH/NIDCD Grant R01DC4336 (S. L. Thibeault), NIH/NHLBI Grant HL081076 K08 (P. Hema-tti), and NIH T32 Physician-Scientist Training Grant CA009614 (S. E.Hanson).

Summer Hanson, MD, Jaehyup Kim, MD, and Beatriz H. Quin-chia Johnson, DDS, PhD contributed equally to this work.

Send correspondence to Susan L. Thibeault, PhD or PeimanHematti, MD, Otolaryngology Office Hematology Office, WIMR 5107WIMR 4033, 1111 Highland Avenue, Madison, WI 53705-2725. E-mail:thibeault@surgery.wisc.edu or pxh@medicine.wisc.edu

DOI: 10.1002/lary.20797

Laryngoscope 120: March 2010 Hanson et al.: Vocal Fold MSCs

546

hydrogels,5–7 injection of cells (autologous or nonautolo-gous mesenchymal stem cells), autologous fibroblasts,and potentially derivatives of human embryonic stemcells.8–10 Mesenchymal stem cells exhibit fibroblastic mor-phology and are generally derived using culturemethodologies similar to what has been used to derivefibroblasts from different tissues. We hypothesized thathuman vocal fold fibroblasts (hVFF) are indeed MSCsderived from these tissues. In this study we characterizedpreviously isolated hVFF derived from adult human vocalfold lamina propria based on the consensus criteria devel-oped to define MSCs11 and compared them to MSCsderived from adipose tissue (AT) and BM.

MATERIALS AND METHODS

Human Mesenchymal Stem Cell Isolationand Culture

Human MSCs were derived from BM and AT of healthydonors based on protocols approved by the University of Wiscon-sin Health Sciences Institutional Review Board (IRB) afterobtaining informed consent from the donors. We used discardedBM filters after BM harvest from normal human leukocyte anti-gen (HLA)-matched sibling donors. Briefly, BM-MSCs wereisolated from cell pellet fractions collected from filters andwashed with phosphate buffered saline (PBS) by centrifugation.Mononuclear cells were isolated by Ficoll Hypaque 1.073 (GEHealthcare Bio-Sciences, Piscataway, NJ) and Leucosep tube(Greiner Bio-One, Monroe, NC) according to manufacturer’sprotocol. Red blood cells (RBC) were lysed with a 3-minute incu-bation in RBC lysis buffer, and mononuclear cells weresuspended in alpha minimum essential media (a-MEM) supple-mented with 10% fetal bovine serum (FBS) (Hyclone, Logan,UT), 1� nonessential amino acids (NEAA) (Sigma-Aldrich, St.Louis, MO), 4 mM L-glutamine (Sigma-Aldrich), and 1� penicil-lin/streptomycin (Sigma-Aldrich). Human AT-MSCs wereisolated from tissues excised or from aspirates from abdomino-plasty flaps. The excised tissue was rinsed with PBS, minced,and digested with type I collagenase/PBS (Sigma-Aldrich) at37�C for 45 minutes. The sample was neutralized with an equalvolume of a-MEM and centrifuged at 1,200 rpm for 5 minutes.Red blood cells were lysed with 3-minute incubations in RBClysis buffer, and mononuclear cells were suspended in supple-mented a-MEM as above and plated on culture flasks. Afteradherence of stromal cells to plastic plates, the culture mediawas changed to remove nonadherent cells. The cells were cul-tured in a 37�C incubator with a 5% CO2, 95% humidifiedatmosphere. Media was changed every 3 days and the cellswere expanded until passage 4, at which time they were usedfor flow cytometry, differentiation experiments, and immuno-phenotype assays.

Human Vocal Fold Fibroblast Isolation and CellLine Derivation

Two primary, normal hVFF lines—p59 and p21—wereobtained from specimens excised from noncancerous donors inaccordance with our IRB-approved protocol; their derivation hasbeen previously described.4 Additionally, two immortalizedhVFF lines—T59 and T21—whose transduction is reported else-where12 were used in this investigation. For currentexperiments, primary and immortalized hVFF were grown inDulbecco’s Modified Eagle’s Medium (DMEM) with 10% FBSand 1� NEAA. All experiments performed on primary andimmortalized cells ranged between passages 4 through 6.

Fluorescent Activated Cell Sorting AnalysisFor fluorescent activated cell sorting (FACS) analysis, sin-

gle cell suspensions of cells were stained and analyzed within24 hour of staining. Data was acquired using FACScalibur (Bec-ton Dickinson, Franklin Lakes, NJ) or Accuri C6 flow cytometer(Accuri, Ann Arbor, MI), and acquired data were analyzed usingFlowJo software (Tree Star, Ashland, OR) or CFlow software(Accuri). Antibodies used were human-specific fluorescently la-beled mouse monoclonal antibodies as follows: anti-CD3allophyocyanin (APC), anti-CD14 APC, anti-CD31 APC, anti-CD45 phycoerithrin (PE) labeled, anti-CD90 PerCP-Cy5.5, anti-CD105 APC, anti-CD163 PE, anti-HLA-ABC fluorescein isothio-cyanate (FITC) (all eBioscience, San Diego, CA), anti-CD14FITC, anti-CD29 PE, anti-CD34 FITC, anti-CD44 PE, anti-CD73 PE, anti-CD90 APC, and anti-HLA-DR FITC (all BDPharmingen, San Diego, CA). For surface staining, Fc receptorswere blocked with Fc Receptor Blocking agent (Miltenyi Bio-tech, Auburn, CA) for 15 minutes at 4�C. Surface antibodieswere added and incubated for 30 minutes at 4�C in the dark,and then cells were washed with autoMACS separation buffer(PBS, 2 mM ethylenediaminetetraacetic acid, 0.5% bovine se-rum albumin, pH 7.2, from Miltenyi Biotech) and fixed with 1%paraformaldehyde in PBS until ready for analysis within 24hours. Control staining with directly labeled-isotype-matchedmonoclonal antibodies was included in all FACS experiments.

Differentiation of Cells Toward Adipogenic,Osteogenic, and Chondrogenic Lineages

Cell differentiation into adipogenic, osteogenic, and chon-drogenic lineages and subsequent detection was performedusing established reported methodologies.2 Briefly, humancells—hVFF (primary and immortalized lines) and AT- and BM-derived MSCs—were treated with osteogenic medium or adipo-genic medium (Miltenyi Biotech) for 21 days, with mediachanges every 3 to 4 days. Cultures were assayed for mineralcontent by alizarin red S staining (Acros Organics, MorrisPlains, NJ) and for lipid accumulation by oil red O staining(Sigma-Aldrich). For chondrogenic differentiation, we used cellpellet three-dimensional culture methodology with chondrogenicmedia (Miltenyi Biotech) changes every 3 to 4 days for 28 days.Paraffin-embedded sections were stained with safranin O(Sigma-Aldrich) to detect glycosaminoglycans.

Cell Surface Marker Change AfterInterferon-c Challenge

Bone marrow derived MSCs, primary hVFF p59 and p21,and immortalized hVFF T59 and T21 were treated with 200 IUof interferon-c (IFN-c) for 2 and 4 days, respectively. Cells werethen harvested by trypsinization and stained with HLA-ABCFITC, HLA-DR PE, CD80 FITC, and CD86 PE (all from eBio-science) by using the protocol described above. Cells untreatedwith IFN-c were used as the control. Staining for each groupwas done in triplicate.

RESULTS

Cell Surface Phenotype Characterizationof hVFF

hVFF primary lines p59 and p21previously isolatedfrom normal healthy tissue donors, and their telomerasetransduced counterparts T59 and T21, respectively, wereexpanded to passage 4 and compared to BM- and AT-MSCs expanded under similar culture conditions

Laryngoscope 120: March 2010 Hanson et al.: Vocal Fold MSCs

547

(Fig. 1). All cell types were used for characterization byflow cytometry, differentiation studies, and immunophe-notype assays. All hVFF were positive for MSC markerssuch as CD29, CD44, CD73, CD90, and CD105 and neg-ative for hematopoietic markers such as CD14, CD31,CD34, CD45, similar to MSCs derived from BM or AT(Fig. 2A). Figure 2B shows the mean values 6 standarderrors of the percentages of MSCs and hVFF that weretested for each cell surface antigen expression by flowcytometry.

Differentiation of Cells Toward Osteogenic,Adipogenic, and Chondrogenic Lineages

Undifferentiated hVFF, both primary and immortal-ized lines, and BM- and AT-derived MSCs weresubjected to adipogenic, osteogenic, and chondrogenicculture conditions. hVFF, both parent cells and immor-talized lines, differentiated into adipocytes, osteocytes,and chondrocytes. Oil red O staining shows lipidvacuoles stained red (Fig. 3, left column); alizarin red Sstaining shows deposits of calcium crystals stained or-ange to brown (Fig. 3, middle column), and safranin Ostaining shows cartilage-specific glycosaminoglycansstained pink (Fig. 3, right column).

Effect of IFN-c on HLA Antigen andCostimulatory Cell Surface Markers

Unstimulated BM-MSCs and hVFF were positivefor HLA-ABC (BM-MSCs, 99.4%; hVFF-p59, 98.8%;hVFF-p21, 99.6%; hVFF-T59, 99.5%; and hVFF-T21,86.1%) although negative for HLA-DR and costimulatorymolecules CD80 and CD86. However, after 2 days ofstimulation with 200 IU of IFN-c, cultured cells began toexpress HLA-DR (BM-MSCs, 69.1%; hVFF-p59, 84.9%;hVFF-p21, 91.5%; hVFF-T59, 79.4%; and hVFF-T21,77.8%), but remained negative for CD80 or CD86. Stim-

ulation for 4 days resulted in almost all the cellsstaining positive for HLA-DR (BM-MSCs, 99.4%; hVFF-p59, 99.5%; hVFF-p21, 100%; hVFF-T59, 99.9%; andhVFF-T21, 99.5%); however, CD80 and CD86 remainednegative overall. Figure 4 summarizes the data fromthree different sets of experiments.

DISCUSSIONWe investigated whether hVFF would be similar to

AT- and BM-derived MSCs in their cell surface markers,differentiation potential, and importantly, their immuno-logical phenotypes. We found that hVFF meet thedefinition of MSCs based on a strict set of criteria widelyaccepted for defining MSCs.11 From a clinical stand-point, the similarity in BM- and AT-derived MSCs andhVFF, now characterized as VF-MSCs in this study, indi-cates that either BM- or AT-derived MSCs could besuitable for novel therapies in vocal fold repair andregeneration. Some of the functional characteristics ofMSCs include their ability to migrate to the site ofinjury or inflammation, stimulate proliferation and dif-ferentiation of resident progenitor cells, and promoterecovery of injured cells through growth factor secretionand matrix remodeling. Furthermore, there is an emerg-ing body of evidence that MSCs possess the ability tosuppress activation and proliferation of T-lymphocytes,B-lymphocytes, natural killer cells, and dendritic cells.Such properties have been the basis of using allogeneicMSCs without HLA matching in a variety of clinicalapplications, such as for treatment of graft versus hostdisease and other inflammatory or immune system dis-orders.1 Importantly, we propose the apparent lack ofimmunogenicity and their potential applicability as auniversal stem cell donor without the need for tissuematching make MSCs a desirable cell-based therapeuticmodality in a variety of pathophysiologic processes,

Fig. 1. Morphology of human vocal fold fibroblasts (hVFF). Representative microscopic views (5�) of human bone marrow-derived mesen-chymal stem cells (MSCs) (upper left), human adipose tissue-derived MSCs (lower left), hVFF p59 (upper middle), hVFF p21 (lower middle),hVFF T59 (upper right), and hVFF T21 (lower right).

Laryngoscope 120: March 2010 Hanson et al.: Vocal Fold MSCs

548

including voice disorders such as vocal fold scarring orparalysis.

Voice disorders affect 3% to 9% of Americans yearlyand have a significant impact on quality of life. Vocalfold scarring may cause a deformity of the vocal foldedge, a disruption of the viscoelastic layered structure ofthe lamina propria, and increase in stiffness of the vi-bratory structure and glottic incompetence. Vocal foldscarring has been suggested by Hirano13 as one of themajor problems awaiting improvement in the future.Treatment for dysphonia caused by connective tissue orECM injury or loss has been limited. The foremost rea-son for the inability to adequately treat thesedysphonias is that present surgical options do notadequately address the ECM biomechanical tissue prop-erties and do not mimic the complex composition of theECM. The composition of the ECM is a central issuebecause of the crucial contributions of this component tothe biomechanical properties of the tissue and the re-sultant voice quality. Collagen injections, fat injections,steroid injections, microflaps with dissection, and fat

implants have all been tried for ECM disorders withvarying success.13 Of unique interest is autologous fatinjection, as it potentially manipulates the undifferenti-ated MSC population. First reported in 1991, intracordalfat injection14 has been clinically useful, although theamount of adipose resorption is unpredictable. Althoughthere are inherent advantages to using autologous fatfor head and neck reconstruction, the technique is alsorestricted by mechanical damage during harvest, cystformation, and localized necrosis. Mesenchymal stemcells derived from AT or BM offer the same regenerativepotential as autologous fat with the additional immuno-modulatory effects described above and also of a highernumber of cells generated through ex vivo cultureexpansion. Nevertheless, it is quite possible that MSCsfrom these different sources will have different geneexpression signatures. These investigations are cur-rently underway in our lab.

Not only are MSCs being investigated in novel cel-lular therapies, this cell type offers many favorablecharacteristics for tissue engineering strategies as well.

Fig. 2. Phenotype of human vocal fold fibroblasts (hVFF). (A) Representative fluorescent activated cell sorting analysis of hVFF cell linesp59, p21, T59, and T21 compared to bone marrow (BM)- and adipose tissue (AT)-derived mesenchymal stem cells (MSCs) for different cellsurface markers. (B) A table of the mean value (6 standard error) of the expression of corresponding markers. These experiments were per-formed in triplicate.

Laryngoscope 120: March 2010 Hanson et al.: Vocal Fold MSCs

549

Recently, Macchiarini and colleagues published theirresults with the transplantation of a tissue-engineeredairway.15 The construct reported was a decellularized al-lograft seeded with autologous chondrocytes derivedfrom BM-MSCs. Such reports provide novel solutions forserious clinical disorders; however, there is much poten-tial for cell-based therapies in other disorders of thehead and neck. Given the unique composition of thevocal folds and laryngotracheal tree, there is much to belearned about cellular differentiation and immunobiol-ogy of MSCs residing in the tissues of interest whenconsidering such cellular therapies. Fibroblasts arederived from the mesodermal germ layer and, therefore,may have an immunophenotype similar to mesenchymalcells. We have shown that fibroblasts cultured fromhuman vocal folds are indeed, by definition, MSCs thatpossess properties similar to BM- and AT-derived MSCs.The results of this study provide insight to the biology of

stromal components of vocal folds, the true identity offibroblasts isolated from these tissues, and the immuno-modulatory phenotype of these cells. Research of thistype begins to set the foundation to address challengingclinical problems in the head and neck population, bothreconstructive and cosmetic, through identification of anoptimal MSC tissue source for head and neck bioengin-eering strategies from direct cell injection to complexcell-scaffold constructions.

CONCLUSIONMesenchymal stem cells have unique functional

and immunomodulatory characteristics that makethem attractive for regenerative medicine. This work,for the first time, shows that hVFF indeed are MSCsas defined by cell surface markers and differentiationpotential; furthermore, these VF-MSCs express an

Fig. 3. Differentiation potential. Rep-resentative microscopic images viewsof the differentiation of human vocalfold fibroblasts (hVFF) cell lines p59,p21, T59, and T21 compared to bonemarrow (BM)- and adipose tissue(AT)-derived mesenchymal stem cellsinto adipogenic (left column), osteo-genic (middle column), and chondro-genic lineages (right column). Oil redO staining shows lipid vacuolesstained red, alizarin red S stainingshows deposits of calcium crystalsstained orange to brown, and safraninO staining shows cartilage-specificglycosaminoglycans.

Laryngoscope 120: March 2010 Hanson et al.: Vocal Fold MSCs

550

immunophenotype similar to BM-MSCs. In-depthgenetic characterization of MSCs from hVFF and theircomparison with BM and AT-MSCs is needed before wecan optimize a cell population for development of tissuebioengineered constructs or other cellular therapymodalities for vocal fold lamina propria reconstruction.

BIBLIOGRAPHY

1. Hematti P. Role of mesenchymal stromal cells in solid organtransplantation. Transplant Rev (Orlando) 2008;22:262–273.

2. Trivedi P, Hematti P. Derivation and immunological charac-terization of mesenchymal stromal cells from human em-bryonic stem cells. Exp Hematol 2008;36:350–359.

3. Thibeault SL, Klemuk SA, Smith ME, Leugers C, PrestwichG. In vivo comparison of biomimetic approaches for tissueregeneration of the scarred vocal fold. Tissue Eng Part A2009;15:1481–1487.

4. Thibeault SL, Li W, Bartley S. A method for identificationof vocal fold lamina propria fibroblasts in culture. Otolar-yngol Head Neck Surg 2008;139:816–822.

5. Hansen JK, Thibeault SL, Walsh JF, Shu XZ, PrestwichGD. In vivo engineering of the vocal fold extracellularmatrix with injectable hyaluronic acid hydrogels: earlyeffects on tissue repair and biomechanics in a rabbitmodel. Ann Otol Rhinol Laryngol 2005;114:662–670.

6. Duflo S, Thibeault SL, Li W, Shu XZ, Prestwich GD. Vocalfold tissue repair in vivo using a synthetic extracellularmatrix. Tissue Eng 2006;12:2171–2180.

7. Duflo S, Thibeault SL, Li W, Shu XZ, Prestwich G. Effect ofa synthetic extracellular matrix on vocal fold laminapropria gene expression in early wound healing. TissueEng 2006;12:3201–3207.

8. Kanemaru S, Nakamura T, Omori K, et al. Regeneration ofthe vocal fold using autologous mesenchymal stem cells.Ann Otol Rhinol Laryngol 2003;112:915–920.

9. Chhetri DK, Head C, Revazova E, Hart S, Bhuta S, BerkeGS. Lamina propria replacement therapy with culturedautologous fibroblasts for vocal fold scars. OtolaryngolHead Neck Surg 2004;131:864–870.

10. Cedervall J, Ahrlund-Richter L, Svensson B, et al. Injectionof embryonic stem cells into scarred rabbit vocal foldsenhances healing and improves viscoelasticity: short-term results. Laryngoscope 2007;117:2075–2081.

11. Dominici M, Le Blanc K, Mueller I, et al. Minimal criteriafor defining multipotent mesenchymal stromal cells. TheInternational Society for Cellular Therapy position state-ment. Cytotherapy 2006;8:315–317.

12. Chen X, Thibeault SL. Novel isolation and biochemicalcharacterization of immortalized fibroblasts for tissue en-gineering vocal fold lamina propria. Tissue Eng Part CMethods 2009;15:201–212.

13. Hirano S, Bless DM, Heisey D, Ford CN. Effect of growthfactors on hyaluronan production by canine vocal foldfibroblasts. Ann Otol Rhinol Laryngol 2003;112:617–624.

14. Mikaelian DO, Lowry LD, Sataloff RT. Lipoinjection for uni-lateral vocal cord paralysis. Laryngoscope 1991;101:465–468.

15. Macchiarini P, Jungebluth P, Go T, et al. Clinical transplan-tation of a tissue-engineered airway. Lancet 2008;372:2023–2030.

Fig. 4. Immunophenotype of human vocal fold fibroblasts (hVFF). Unstimulated bone marrow (BM)-mesenchymal stem cells (MSCs) andhVFF cells are positive for human leukocyte antigen (HLA)-ABC but do not express HLA-DR, CD80, or CD86. Treatment of cells with inter-feron-c induces the cell surface expression of HLA-DR, whereas CD80 and CD86 remain unchanged. Shown is the trend in cell surfacemarker expression over 4 days of treatment with IFN-c for BM-MSCs, hVFF-p59, hVFF-p21, hVFF-T59, and hVFF-T21. These experimentswere performed in triplicate.

Laryngoscope 120: March 2010 Hanson et al.: Vocal Fold MSCs

551