ORIGINAL ARTICLE Influence of Growth Factors on Tissue-Engineered Pediatric Elastic

Carlos A. Are´valo-Silva, MD; Yilin Cao, MD, PhD; Martin Vacanti, MD; Yulai Weng, DD; Charles A. Vacanti, MD; Roland D. Eavey, MD

Objective: To investigate the influence of growth fac- Results: The average number of cells generated in tors on tissue-engineered pediatric human elastic carti- vitro was as follows: group 1, 12 million; group 2, 40 lage relative to potential clinical application. million; group 3, 7 million; and group 4, 35 million. Group 2 in vivo gross specimens were the largest and Design: Controlled study. heaviest. Histologically, the control group and the basic fibroblast growth factor group (groups 1 and 2) Subjects: Eleven children ranging in age from 5 to 15 exhibited characteristics compatible with normal years provided auricular elastic cartilage specimens mea- auricular cartilage; groups 3 and 4 demonstrated cellu- suring approximately 1ϫ1ϫ0.2 cm and weighing ap- lar disorganization and moderate to severe fibrous tis- proximately 100 mg. sue infiltration.

Interventions: Three million were plated Conclusions: Basic fibroblast growth factor demon- into 4 groups of Ham F-12 culture medium: group 1, Ham strates the greatest positive influence on the in vitro F-12culturemediumonly;nogrowthfactors(controlgroup); and in vivo growth of engineered pediatric human group 2, Ham F-12 culture medium and basic fibroblast auricular cartilage. The results suggest that basic fibro- growth factor; group 3, Ham F-12 culture medium and trans- blast growth factor has the potential for clinical appli- forming growth actor ␤; and group 4, Ham F-12 culture me- cation in which a goal will be to generate a large vol- dium and a combination of both growth factors. At 3 weeks, ume of tissue-engineered cartilage from a small donor the cells were harvested and mixed with a copolymer gel of specimen in a short period of time and of a quality polyethylene glycol and polypropylene oxide (Pluronic similar to native human elastic cartilage. F-127). The cell solution was injected subcutaneously into athymic mice. The constructs were harvested at up to 22 weeks of in vivo culture and histologically analyzed. Arch Otolaryngol Head Neck Surg. 2000;126:1234-1238

ARTILAGE IS vital for re- promising results using freshly harvested constructive surgery and cells to generate a variety of cartilage struc- is used to repair a variety tures.5-8 The future concept of tissue en- of defects within the spe- gineering for humans should require the cialty of otorhinolaryn- harvest of relatively few cells from a small From the Department of gology. The major cartilage source for biopsy specimen. The cells would be al- Anesthesiology, Center C for Tissue Engineering microtia reconstruction is rib fibrocarti- lowed to multiply in vitro and would be (Drs Are´valo-Silva, Cao, lage. Although ribs are a dependable do- added to a preshaped scaffold to generate M. Vacanti, Weng, and nor cartilage source, there are disadvan- new tissue. Although many human pedi- C. A. Vacanti), and the tages to their use. The amount of cartilage atric cells can be generated using current Department of Pathology is limited early in life; intraoperative sculpt- culture techniques,9 months of in vitro cul- (Dr M. Vacanti), University of ing is required; the patient experiences ture may be required to expand the cells Massachusetts Medical Center, discomfort and displays a chest scar; and to a sufficient number to create a full- Worcester; the Department of the reconstructed feels quite firm.1-3 sized , which may result in aged cells Otolaryngology, Massachusetts Tissue engineering, which com- with diminished properties and a tissue of Eye and Ear Infirmary, Boston 10,11 (Drs Are´valo-Silva and Eavey); bines biologic principles with fundamen- suboptimal quality. and the Department of Otology tal engineering and polymer chemistry, In this study, we employed 2 of the and Laryngology, Harvard enables the generation of new tissue most commonly used growth factors for Medical School, Boston replacements in animal models.4 Previ- culture, hypothesizing a rapid (Drs Are´valo-Silva and Eavey). ous animal studies have demonstrated growth effect on the tissue-engineered car-

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, OCT 2000 WWW.ARCHOTO.COM 1234

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 SUBJECTS, MATERIALS, 1-ng/mL TGF-␤. The cells were maintained in vitro at 37°C and 5% carbon dioxide for 3 weeks. The culture AND METHODS medium was changed twice a week. The in vitro growth rate was estimated by photographic records of each CHONDROCYTE ISOLATION group. After 3 weeks of in vitro culture, the cells were harvested using 0.25% trypsin/EDTA (Sigma-Aldrich Excess pediatric elastic cartilage from ear surgery was ob- Corp) and counted using a hemocytometer. Cell viabil- tained from 11 children (age range, 5-15 years) by one of ity was determined using trypan blue vital dye. us (R.D.E.) at the Massachusetts Eye and Ear Infirmary, Bos- ton, with informed consent. was removed CHONDROCYTE POLYMER SUSPENSION under sterile conditions. The isolated cartilage was minced into small fragments; washed in phosphate-buffered sa- After quantification, the cells were suspended in a 30% wt/ line solution containing 100-mg/L penicillin, 100-mg/L vol solution of a copolymer gel of polyethylene glycol and streptomycin, and 0.25-mg/L amphotericin B (Gibco, Grand polypropylene oxide (Pluronic F-127; BASF, Mount Ol- Island, NY); and digested with 0.3% collagenase II (Wor- ive, NJ) at 4°C and constituted in Ham F-12 medium at a thington Biochemical Corp, Freehold, NJ) at 37°C for 8 to cellular concentration of 6ϫ107 cells per milliliter. Ali- 12 hours. The digested cartilage suspension was filtered us- quots containing 100 µL of the above-mentioned mixture ing a sterile 250-mm polypropylene mesh filter (Spectra/ were prepared. Mesh 146-426; Spectrum Medical Industries Inc, Laguna Hills, Calif) and centrifuged at 6000 rpm. The resulting pel- SURGICAL IMPLANTATION let of cells was washed twice with phosphate-buffered sa- line and then resuspended in Ham F-12 medium. The num- Following the animal facility guidelines of the University ber of cells was calculated using a hemocytometer, and the of Massachusetts Medical Center, Worcester, equal ali- viability of the cells was determined using trypan blue vi- quots of the chondrocytes suspended in Pluronic F-127 were tal dye (Sigma-Aldrich, Irvine, Calif). Chondrocyte sus- injected into the dorsal subdermal space of the athymic mice pensions with cell viability in excess of 85% were used. (nu/nu) (Taconic Inc, Boston), which were under general anesthesia. CHONDROCYTE IN VITRO CULTURE HARVEST OF THE SPECIMENS The chondrocytes were plated onto 225-cm2 cell culture flasks (Costar, Cambridge, Mass) at 2400 cells per The specimens were harvested after 4, 6, 8, 10, 12, 14, 16, square centimeter. Four groups of cells from each 18, 20, and 22 weeks of in vivo implantation. After anes- patient were plated at the same concentration. Group 1 thetic overdose, the constructs were carefully dissected from was nourished with Ham F-12 and levoglutamine, the surrounding soft tissue. The specimens were exam- 50-mg/L L-ascorbic acid, 100-mg/L penicillin, 100-mg/L ined grossly, weighed, and measured. streptomycin, 0.25-mg/L amphotericin B, and 10% fetal bovine serum (Sigma-Aldrich Corp, St Louis, Mo). HISTOLOGICAL EXAMINATION Group 2 was nourished with the same Ham F-12 culture medium supplemented with 10-ng/mL b-FGF (R&D Each specimen was fixed in 10% phosphate-buffered for- Systems Inc, Minneapolis, Minn). Group 3 was nour- malin (Fisher Scientific, Fair Lawn, NJ) for histological ished with Ham F-12 culture medium supplemented analysis. The specimens were embedded in paraffin and with 1-ng/mL TGF-␤ (R&D Systems Inc). Group 4 was sectioned. Using standard histochemical techniques, slide nourished with Ham F-12 culture medium and a combi- sections were stained with hematoxylin-eosin, safranin O, nation of the 2 growth factors: 10-ng/mL b-FGF and Masson trichrome blue, and Verhoeff solution.

tilage. One growth factor, basic fibroblast growth factor tion about the effect of growth factors on human auricu- (b-FGF), is a single-chain peptide composed of 146 amino lar cartilage is limited.18 We wished to further evaluate acids.12 Basic fibroblast growth factor does not bind to whether robust chondrocyte expansion might be pos- other growth factor receptors, nor do other growth fac- sible, with an eventual goal of clinical application. tors bind to the b-FGF receptor,12,13 and it can either stimu- 14 late proliferation or delay differentiation. The other RESULTS growth factor, transforming growth actor ␤ (TGF-␤), is a multifunctional peptide dimer from a family of impor- IN VITRO RESULTS tant cell regulators that control growth and differentia- tion, among other functions.15,16 Transforming growth ac- The chondrocytes grown in monolayer cultures in the 4 tor ␤ binds to a specific cell membrane receptor of high different media varied in number, size, and shape. The affinity that is found in essentially all cells. There is no chondrocytes cultured with Ham F-12 culture medium cross-reactivity between TGF-␤ and receptors for oth- (group 1) were polygonal at 1 week, with very few cells ers growth factors.15 Both b-FGF and TGF-␤ act by para- still unattached to the flask bottom. At the beginning of crine and autocrine mechanisms.16 Depending on the cell 2 weeks, cell multiplication became evident. The cells ap- type, TGF-␤ can either inhibit or potentiate b-FGF ac- peared to be of uniform size, polygonal, and distributed tivity, cell proliferation, and differentiation.16,17 Informa- as islets of growth. At 3 weeks, a few cells became elon-

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, OCT 2000 WWW.ARCHOTO.COM 1235

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 4.0 evidence of multiple layers. The average total number of cells was 3.5ϫ107 (Figure 1). Cell viability was more than 3.5 98%; the average cell doublings was about 5 times the 3.0 original number. 7

10 2.5 × 2.0 4 IN VIVO (XENOGRAFT) RESULTS 3.5 1.5

No. of Cells Specimens generated from the 4 experimental groups of 1.0 cells were harvested after 4, 6, 8, 10, 12, 14, 16, 18, 20, 0.5 1.2 and 22 weeks of in vivo culture. The gross appearance 0.7 of the specimens was different among the groups. The 0 largest specimens were from group 2, and the smallest Ham F-12 Culture Medium Plus were from group 3, with those from groups 1 and 4 be- No Growth Factors b-FGF TGF-β b-FGF and TGF-β ing intermediate. The specimens were roughly spheri- cal and ranged in size as follows: group 1, 3.5 to 5.5 mm; Figure 1. Average number of chondrocytes after 3 weeks of in vitro culture. b-FGF indicates basic fibroblast growth factor; TGF-␤, transforming growth group 2, 8.7 to 9.5 mm; group 3, 3.0 to 4.0 mm; and group factor ␤. 4, 3.5 to 4.5 mm. The average weight of the specimens was as follows: group 1, 28 mg (range, 24-36 mg); group 2, 99 mg (range, 97-105 mg); group 3, 16 mg (range, 14-18 gated. Rare areas were devoid of cells. At the time of me- mg); and group 4, 40 mg (range, 35-42 mg). Histologi- dium exchange, a thin liquid with no filaments was evi- cal evaluation demonstrated the differences among the dent. After 3 weeks, the average total cell count was 4 groups. 1.2ϫ107 (Figure 1). The cell viability was greater than Group 1 generated lobules of cartilage with a baso- 98%. The average number of cell doublings was 2.5. philic matrix within evenly spaced lacunae. The lacu- In group 2, in which the chondrocytes were nour- nae varied from round to oval, with slight pleomor- ished with Ham F-12 containing b-FGF, polygonal cells phism. The results of safranin O staining were strongly were seen attached to the plates. At the end of the first positive, correlating with production. Tri- week, cell multiplication was evident. The cells ap- chrome blue staining showed a well-defined perichon- peared to be smaller than those in groups 1 and 3. The drium (Figure 2A). Verhoeff staining revealed elastic shape of the cells persisted, with islands of cell growth. fibers that were homogeneously distributed. At 2.5 weeks, the chondrocytes reached confluence, but The specimens in group 2 demonstrated a variable continued to multiply. The medium was thicker in group matrix with heterochromacia indicative of maturity. The 2 than in any other group, with the formation of visible cartilage was very cellular with round to oval lacunae con- wide filaments, which were evident at the time of me- taining single chondrocytes with single round to oval nu- dium exchange. At the end of 3 weeks, a thick layer of clei, with mild variation. The specimens were strongly cells was present on the bottom of the flasks, with evi- positive for proteoglycans. The perichondrium was thin dence of multiple layers. The average total number of cells and well defined (Figure 2B). The elastic fibers stained was 4ϫ107 (Figure 1). The cell viability was greater than evenly positive (Figure 3). 98%. The average cell doublings for this group was 5.5 The specimens from group 3 had a small ratio of times the original number. round to oval lacunae, containing single chondrocytes. The chondrocytes nourished with Ham F-12 me- There was a predominant presence of fibrous tissue sur- dium containing TGF-␤ (group 3) were polygonal at 1 rounding the cartilage. Moreover, trichrome blue stain- week, with a few unattached cells having a round con- ing revealed fibrous tissue in the engineered cartilage and figuration. At 2 weeks, cell multiplication was evident. a thick perichondrium. The proteoglycans production was The size and shape of the cells persisted within islands strong, as evidenced by safranin O staining. The overall of cell growth. At 3 weeks, the size and shape of the cells quality of cartilage was poor, containing fibrous tissue remained unchanged. Few areas were still devoid of cells. and (Figure 2C). There was moderately posi- The medium at the time of medium exchange was liq- tive staining of the elastic fibers. uid, with no consistency. The average total number of In group 4, the specimens stained basophilic, with cells was 7ϫ106 (Figure 1). The cell viability was more areas of mild heterochromacia. The tissue was very cel- than 98%. An average of 1.5 cell doublings was reached. lular, with some grouping of cells. There was moderate The fourth group of chondrocytes had been nour- variation in the size and shape of the lacunae, which oc- ished with b-FGF and TGF-␤ in Ham F-12 medium. Af- casionally contained bichondrocytes. There was also mild ter the first week, all the cells were attached and polygo- to moderate variation in the size and shape of the nu- nal and showed evidence of cell multiplication. The size clei. Staining results for proteoglycans production were and shape of the cells persisted, with islands of cell growth. strongly positive. The perichondrium was well defined, At 3 weeks, the chondrocytes reached confluence, yet con- with rare ingress of fibrous tissue into the cartilage (Fig- tinued to multiply. These cells were similar in size to the ure 2D). There was moderate to strong staining of the cells in group 3 after the same period of time. At the time elastic fibers. of medium exchange, the medium/serum was very thick, Safranin O staining of the younger specimens (Ͻ6 with evident filament formation. At the end of 3 weeks, weeks) demonstrated a hyperchromic and homoge- a thick layer was present on the bottom of the flasks, with neous matrix with abundant rounded cells and scarce

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, OCT 2000 WWW.ARCHOTO.COM 1236

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 A C

B D

Figure 2. A, Group 1 (control group). The specimen has a well-defined perichondrial layer and no fibrous ingrowth. B, Group 2. The perichondrial layer and the appearance of the cartilage are similar to those of the group 1 specimen. C, Group 3. The specimen is the smallest and shows fibrous ingrowth. D, Group 4. There is visible fibrous tissue ingrowth. All specimens were cultured for 10 weeks in vivo (Masson trichrome, original magnification ϫ8). Arrows indicate the perichondrium [A through D]; arrowheads, fibrous tissue [C and D].

staining for the production of proteoglycans. The older specimens (Ͼ6 weeks) demonstrated a hypochromic and heterogeneous matrix and proportionately fewer cells of shapes from round to oval, with a strong detection of pro- teoglycans as a bright-red stain in all the specimens.

COMMENT Tissue engineering has evolved as a science that may re- duce the problem of scarcity for tissue reconstruction and organ replacement. Although advances in this new field have been encouraging, the clinical applications may not be realized for several years. One major obstacle has been the generation of a sufficient number of healthy young cells in a short period of time. Figure 3. All specimens were positive for elastic fibers, which stain black, This report compares the effect of 2 growth factors similar to native cartilage. This group 2 specimen was cultured in vivo for 10 used in the culture of pediatric auricular chondrocytes. weeks (Verhoeff, original magnification ϫ40). The in vitro findings demonstrate that b-FGF alone re- sults in the generation of more tissue, and tissue of a higher quality, than other combinations of growth factors. For ation of specimens nourished with b-FGF demon- example, the number of in vitro cell divisions in con- strated cartilage that was superior to cartilage from speci- ventional medium alone or in supplemental TGF-␤ was mens that were nourished with TGF-␤ alone; specimens significantly lower than the number in b-FGF alone or from the TGF-␤ group demonstrated a consistent infil- in the 2 growth factors combined. Furthermore, the in tration of fibrous tissue into the elastic cartilage. vivo specimens cultured with b-FGF were notably the The specimens formed from cells nourished only largest and heaviest. Most importantly, histological evalu- with Ham F-12 medium were histologicially of very good

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, OCT 2000 WWW.ARCHOTO.COM 1237

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 quality; however, the in vivo cell number was signifi- atric Otolaryngology, Palm Desert, Calif, April 29, 1999. cantly less than that of specimens from the b-FGF Reprints: Roland D. Eavey, MD, Massachusetts Eye and group, suggesting a persistent effect of this growth fac- Ear Infirmary, 243 Charles St, Boston, MA 02114. tor. Also, even though the absolute number of in vitro cells obtained with the combination of growth factors REFERENCES compared with b-FGF alone was not substantially dif- ferent, the amount of in vivo tissue obtained from the 1. Eavey RD. Surgical repair of the auricle for microtia; In: Bluestone C, Stool S, combination group was significantly smaller and of eds. Atlas of Pediatric Otolaryngology. Philadelphia, Pa: WB Saunders Co; 1995: poorer quality, with moderate fibrous tissue extending chap 2. 2. Eavey RD, Ryan DP. Refinements in pediatric microtia reconstruction. Arch Oto- into the cartilage. Transforming growth factor ␤ can laryngol Head Neck Surg. 1996;122:617-620. inhibit cell proliferation, which could be the reason for 3. Eavey RD. In discussion of: Cao Y, Vacanti JP, Paige KT, Upton J, Vacanti CA. suboptimal neocartilage development. Transplantation of chondrocytes utilizing a polymer cell construct to produce tissue- engineered cartilage in the shape of a human ear. Plast Reconstr Surg. 1997; It was previously demonstrated that a large volume 100:303-304. of bovine chondrocytes could be induced to consis- 4. Langer R, Vacanti JP. Tissue engineering. Science. 1993;260:920-926. 5. Cao Y, Vacanti JP, Paige KT, Upton J, Vacanti CA. Transplantation of chondro- tently create an auricular construct in a nude mouse cytes utilizing a polymer cell construct to produce tissue-engineered cartilage in 5 model. However, for human clinical application, sev- the shape of a human ear. Plast Reconstr Surg. 1997;100:297-302. eral challenges remained. For example, human pediat- 6. Sakata J, Vacanti CA, Schloo B, Healy GB, Langer R, Vacanti JP. Tracheal com- posites tissue-engineered from chondrocytes, tracheal epithelial cells and syn- ric elastic cartilage had to demonstrate properties nec- thetic degradable scaffolding. Transplant Proc. 1994;26:3309-3310. essary for tissue engineering efforts.9 A corollary then arose 7. Puelacher WC, Mooney D, Langer R, Upton J, Vacanti JP, Vacanti CA. Design of as to whether a limited number of chondrocytes from a nasoseptal cartilage replacements synthesized from biodegradable polymers and chondrocytes. Biomaterials. 1994;15:774-778. small piece of human cartilage might be expanded to cre- 8. Cao Y, Lach E, Kim TH, Rodrı´guez A, Are´valo CA, Vacanti CA. Tissue-engineered ate a full-sized auricle. The results of this study demon- nipple reconstruction. Plast Reconstr Surg. 1998;102:2293-2298. 9. Rodriguez A, Cao Y, Ibarra C, et al. Characteristics of cartilage engineered from strate that in vitro chondrocyte enhancement by a growth human pediatric auricular cartilage. Plast Reconstr Surg. 1999;103:1111-1119. factor produces a large volume of healthy-appearing 10. Britt JC, Park SS. Autogenous tissue-engineered cartilage: evaluation as an im- cartilage in a reasonable period of time. Challenges plant material. Arch Otolaryngol Head Neck Surg. 1998;124:671-677. 3 11. Froger-Gaillard B, Charrier AM, Thenet S, Ronot X, Adolphe M. Growth- remain, but the concept of auricular tissue engineering promoting effects of acidic and basic fibroblast growth factor on rabbit articular appears to be closer to reality. chondrocytes aging in culture. Exp Cell Res. 1989;183:388-398. 12. Esch F, Baird A, Ling N, et al. Primary structure of bovine pituitary basic fibro- blast growth factor (basic-FGF) and comparison with the amino terminal se- CONCLUSIONS quence of bovine brain acidic FGF. Proc Natl Acad SciUSA.1985;82:6507- 6511. 13. Neufeld G, Gospodarowicz D. The identification and partial characterization of The results demonstrated that b-FGF enhanced the de- the fibroblast growth factor receptor of baby hamster kidney cells. J Biol Chem. velopment of engineered human pediatric elastic carti- 1985;260:13860-13868. lage in a xenograft model. This growth factor generated 14. Gosporadowicz D, Ferrara N, Schweigerer L, Neufeld G. Structural characteriza- tion and biological functions of fibroblast growth factor. Endocr Rev. 1987;8: a significant amount of engineered cartilage of high qual- 95-114. ity from a small number of pediatric auricular cells in an 15. Villiger PM, Lotz M. Differential expression of TGF beta isoforms by human ar- acceptable period of time. ticular chondrocytes in response to growth factors. J Cell Physiol. 1992;151: 318-325. 16. Sporn MB, Roberts AB, Wakefield LM, Assoian RK. Transforming growth fac- Accepted for publication April 11, 2000. tor-␤: biological function and chemical structure. Science. 1986;233:532-534. 17. Chiang CP, Nilsen-Hamilton M. Opposite effect of epidermal growth factor and This study was awarded the second place William P. human platelet transforming growth factor-␤ on the production of secreted pro- Potsic Prize for Basic Research. teins by murine 3T3 cells and human fibroblasts. J Biol Chem. 1986;261:10478- Presented in part at the 14th Annual Meeting of the 10481. 18. Quatela VC, Sherris DA, Rosier RN The human auricular chondrocyte: re- American Society of Pediatric Otolaryngology and the Third sponses to growth factors. Arch Otolaryngol Head Neck Surg. 1993;119: Biannual Meeting of the International Association of Pedi- 32-37.

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, OCT 2000 WWW.ARCHOTO.COM 1238

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021