Does Cartilage Down-regulate Growth Factor Expression in Tracheal Epithelium? FREE

TRAUMA THAT disrupts intraluminal epithelium, regardless of its cause, often leads to aberrant repair. This process is pathologically manifested by the exuberant proliferation of granulation tissue and replacement of the normal respiratory epithelium with fibroblasts.^{1 - 2} This often leads to scar formation, airway stenosis, and eventual physiologic compromise of the host respiratory tract.

There is currently no effective way to study intraluminal events of reepithelialization after injury. Present approaches to tracheal repair include resection and reanastomosing the injured airway, replacement of the damaged portion by synthetic material, and use of autologous tissue for reconstruction of the tracheal defect.^{3 - 5} Recently, tissue engineering approaches have been taken, including forming an in vivo tracheal cartilaginous scaffolding by injecting dissociated chondrocytes into a preformed synthetic construct.^{6 - 8} Such devices were of limited success owing to lack of reepithelialization. In the case of synthetic replacement, migration of the prosthesis can occur and may result in chronic ulceration, and even fatal hemorrhage.²

We have developed a 3-dimensional in vitro model system that incorporates both cartilaginous and epithelial elements, a feature unique among currently available models of upper airway injury and repair. Using this model, we hypothesized that communication between epithelial cells and underlying stroma would lead to modulation of epithelial cell growth and differentiation through the release of growth factors. To test this hypothesis, we investigated expression of transforming growth factors α and β (TGF-α and TGF-β) in respiratory epithelial cells (REC) cultured on type I collagen or composite cultures of chondrocytes and type I collagen.

Transforming growth factor α is a member of the epidermal growth factor family and plays an important role in wound healing.^{9 - 11} τransforming growth factor β1 is a multifunctional polypeptide with differing cell-specific effects, including stimulation or inhibition of proliferation, and regulation of extracellular matrix production and remodeling.^{12 - 14}

Respiratory epithelial cells in a composite culture system with cartilage represents a 3-dimensional model of the tracheal luminal surface, which facilitates studies of the interactions between epithelium and its underlying stroma. Using this model, we have shown that cartilage may modulate the migration of REC, and changes in expression of TGF-α and TGF-β, when compared with REC grown on type I collagen alone.

Chondrocytes were harvested from bovine articulator cartilage under clean conditions, minced finely, and digested for 12 to 16 hours at 37°C in phosphate-buffered saline with antibiotics and collagenase II (Worthington, Freehold, NJ) and DNAse I (Sigma-Aldrich Corporation, St Louis, Mo), as described elsewhere.¹⁵ Cell viability was determined by trypan blue staining, and cell type was confirmed by staining with hematoxylin-eosin and antibody to extracellular type II collagen. Chondrocytes were plated on collagen inserts (Costar Transwell; VWR, Rochester, NY) at 20 to 40×10⁶ cells per well in Dulbecco modified Eagle medium (DMEM/F-12) (Gibco, Grand Island, NY), with antibiotics, 10% fetal calf serum, and 50-µg/mL ascorbic acid. After the chondrocytes formed a layer of extracellular matrix, 0.5 mL of type I collagen (Vitrogen; Collagen Biomaterials, Palo Alto, Calif) was added.

Human tissue specimens were obtained through Manhattan Eye, Ear, Nose and Throat Hospital, New York, NY, under a human institutional review board–approved protocol. Samples were healthy tissues from surgical procedures performed on adults. Specimens were removed aseptically, rinsed in isotonic sodium chloride solution to remove debris, and shipped at 4°C within 24 hours of procurement. Upper respiratory tract epithelium from bronchus, nasal polyps, or turbinates was dissociated as described elsewhere¹⁶ and REC were added to type I collagen–coated chondrocytes at 2×10⁵/cm², in DMEM/F-12, containing antibiotics, epidermal growth factor (10 ng/mL), hydrocortisone (1 µmol/L), insulin (5 µg/mL), L-isoproterenol (1 µmol/L), and 5% fetal calf serum. Cultures were maintained for 2 weeks and then processed for scanning electron microscopy, histochemical analysis, or RNA isolation. Epithelial cells were also seeded at the same density onto type I collagen–coated culture dishes and kept for the same length of time as the composite cultures.

Fourteen-day-old type I collagen REC cultures, and REC from composite cultures were processed for scanning electron micrcoscopy and examined with a scanning electron microscope (model 35CF; JEOL, Japan) at 10 to 12 kV.

Twenty-eight-day-old chondrocyte cultures were formalin fixed and processed for immunoperoxidase histochemical analysis. Rabbit polyclonal antibody to type II collagen (DAKO, Carpinteria, Calif) was applied, washed, and followed by a biotinylated universal secondary antibody and streptavidin peroxidase (Immunon/Shandon-Lipshaw, Pittsburgh, Pa).

Total cellular RNA was isolated from day 14 REC composite cultures, and REC cultured on type I collagen, using a reagent (TriReagent; Molecular Research Center, Cincinnati, Ohio), as described by the manufacturer. Under a dissecting microscope, the REC layer was removed from composite cultures with a sterile, RNAse-free spatula into TriReagent. One to five micrograms of total RNA was reverse transcribed using murine Moloney leukemia virus reverse transcriptase.¹⁷ Polymerase chain reaction amplification was performed using primer sets for human TGF-α, TGF-β1, and β-actin (Clontech, Palo Alto, Calif), following manufacturer's instructions. One-tenth volumes of polymerase chain reaction products were run on 2.5% or 3% agarose gels and visualized by ethidium bromide staining. The polymerase chain reaction products obtained were of expected sizes (human β-actin, 838 base pairs [bp]; human TGF-β, 161 bp; and human TGF-α, 297 bp).

Bovine chondrocytes established in primary culture were morphologically similar to in vivo bovine cartilage (Figure 1, A and B). Cartilage cultured for less than 2 months did not always form lacunae, but always produced an abundant extracellular matrix of type II collagen (Figure 1, C).

Respiratory epithelial cells grown on type I collagen formed a continuous sheet (Figure 2, A and B). In contrast, REC grown on composite cultures did not spread to confluence, but rather grew as patches of epithelium (Figure 2, C and D). The epithelial cell layer in both the composite cultures and on type I collagen was undifferentiated.

In REC from day 14 composite cultures, expression of both TGF-α and TGF-β was reduced (Figure 3, lane 4, TGF-α; lane 6, TGF-β) compared with REC on type I collagen (Figure 3, lane 3, TGF-α; lane 5, TGF-β). Relative expression of β-actin was equal (Figure 3, lanes 1 and 2).

Paraffin sections of REC grown on type I collagen and on composites were immunostained for TGF-α and TGF-β. Both growth factors were expressed in chondrocytes, and to a lesser extent, in epithelial cells (data not shown).

A frequent problem seen in tracheal repair with synthetic or autologous materials is the failure of luminal surface reepithelialization. Failure of reepithelialization to reestablish luminal integrity is an important reason why no acceptable surgical procedure exists for the repair of extended segments of trachea compromised by inhalation injury, congenital anomalies, or neoplastic disease.

Why the rate of reepithelialization in the large conducting airway is different from that seen within other epithelial-lined or -covered surfaces is unclear. The phenomenon of "slowed" reepithelialization is seen after both ablative surgical reconstruction and denudation injury, where the epithelium and basement membrane are removed with an intact cartilaginous superstructure (eg, inhalation injury).

One of the difficulties in understanding the relationship between respiratory epithelium and its underlying substructure (cartilage and submucosa) is the inaccessibility of the tissue for direct observation. Our model facilitates the examination the tissue interactions and mechanisms resulting in REC migration.

Both TGF-α and TGF-β play crucial roles in new tissue formation and remodeling. Transforming growth factor α stimulates proliferation in cultured epithelial cells,¹⁸ fibroblasts,¹⁹ and endothelial cells.²⁰ It is chemotactic for epithelial cells in vitro²¹ and enhances epithelial wound healing when applied topically.⁹ Transforming growth factor β is mitogenic for cells of mesenchymal origin and plays a role in repair through its ability to modulate extracellular matrix formation and tissue remodeling.

When isolated human REC were cultured on type I collagen, the cells spread to form a confluent layer as has been previously reported by other authors.²² When plated onto composite cultures with a layer of type I collagen on top of cartilage, the cells did not spread efficiently but formed epithelial nests. Complete reepithelialization of the surface did not occur, even after 3 weeks.

We examined TGF-α and TGF-β expression in REC from composite cultures and found that both of these growth factors were reduced in epithelial cells from 14-day composite cultures when compared with the expression of these factors in REC cultured on type I collagen alone. This suggests that the cartilage modulates the behavior of epithelial cells. One hypothesis for the observed diminished expression of TGF-α and TGF-β is the secretion of soluble factors from the cartilage. Transforming growth factor α was expressed in cartilage (not shown), where it may have acted on the epithelium in a paracrine manner to decrease its expression.

Patients with critical large conducting airway injury requiring medical intervention often succumb to their condition because no reliable means of restoring the large conducting airway exists. Lack of REC migration and proliferation remains the key physiologic and biologic issue underlying failure of surgical repair. Understanding the factors that influence epithelial migration and regeneration in the large conducting airway remains a central issue in solving this clinical problem. The model presented demonstrates one potential mechanism to explain the phenomenon of impaired reepithelialization seen in the large conducting airway after injury.

Accepted for publication March 31, 1999.

Reprints: Wesley L. Hicks, Jr, MD, Head and Neck Surgery, Roswell Park Cancer Institute, Elm and Carlton streets, Buffalo, NY 14263.