Mesenchymal cell activation is the rate-limiting step of granulation tissue induction.

During wound repair a 3-day lag occurs between injury and granulation tissue development. When full-thickness, 8-mm-round, excisional wounds were made in the paravertebral skin of outbred Yorkshire pigs and harvested at various times, no granulation tissue was observed before day 4. Day 4 wounds were 3% filled with granulation tissue, day 5 wounds 48% filled, and day 7 wounds 88% filled. The prerequisites for granulation tissue induction are not known but hypothetically include fibrin matrix maturation or cell activation. To examine whether matrix maturation was necessary, wounds were allowed to heal for 5 or 7 days and then aggressively curetted, resulting in the formation of fresh fibrin clots in the newly formed wound spaces. In contrast to original wounds, no lag phase was observed; wounds curetted on day 5 were 23% filled with granulation tissue 1 day later and 99% filled 3 days later, whereas wounds curetted on day 7 were 47% filled 1 day later and completely filled within 2 days. Thus, granulation tissue formation resumed promptly and independently of fibrin clot matrix maturation. This observation suggested that mesenchymal cell activation might be the rate-limiting step in granulation tissue formation. To address this hypothesis more directly, cultured porcine or human fibroblasts, grown to 80% confluence in Dulbecco's minimal essential medium plus 10% fetal calf serum, were added to new wounds. These wounds were sealed with a freshly made exogenous fibrin clot. In some wounds, platelet releasate was added to the fibrin clot. Granulation tissue did not form in day 3 wounds, which had received either fibrin alone, fibrin and platelet releasate, or fibrin and fibroblasts. In contrast, granulation tissue was observed in wounds receiving fibrin, human fibroblasts, and platelet releasate. By day 4, wounds receiving cultured human fibroblasts, fibrin, and platelet releasate were 14% filled with granulation tissue compared with less than 4% granulation tissue in control wounds. Thus, fibroblast activation is a limiting step of granulation tissue formation, and continued cell stimulation is required for accelerated development.

Vitronectin in colorectal adenocarcinoma--synthesis by stromal cells in culture.

We have investigated the expression and cellular source of vitronectin in colorectal adenocarcinoma. Immunofluorescence staining of tissue sections revealed the presence of vitronectin in the stroma of the 11 tumors studied, but not in adjacent normal colon. A method was devised for the isolation from colorectal adenocarcinomas of fibroblast-like cells that stained positive for vimentin but negative for cytokeratin. These tumor-derived stromal cells synthesized and secreted vitronectin, as revealed by metabolic labeling and immunoprecipitation. This was confirmed by Southern blot analysis of polymerase chain reaction amplification products from reverse-transcribed RNA. Normal skin fibroblasts did not synthesize vitronectin. Immunofluorescence staining showed vitronectin deposited at focal contact sites in the tumor-derived cells, where it colocalized with vinculin and the alpha v integrin subunit. The deposition of vitronectin into focal contact sites was not dependent on the presence of serum. The finding that vitronectin can be synthesized and secreted by tumor-derived fibroblast-like cells in culture indicates that vitronectin expression can be promoted by as yet unknown signals provided in disease states, such as cancer.

An alternative cytoplasmic domain of the integrin beta 3 subunit.

A cDNA encoding a new form of the shared beta subunit (beta 3) of the platelet integrin gpIIb/IIIa and the vitronectin receptor was isolated from a placental cDNA library by screening with a beta 3 (gpIIIa) DNA probe. This beta 3 variant differs from the previously reported beta 3 in that the cytoplasmic domain is 8 amino acids shorter and has an alternative, 13-amino acid COOH-terminal peptide. The 3' untranslated region of the cDNA also differs from the previously reported sequence, while the region coding for the transmembrane domain and extracellular domain is identical to it. Reverse transcription combined with polymerase chain reaction was used to show that human placental tissue and two human cell lines contain the variant mRNA. The sequences of the cDNAs for the previously known beta 3 and the variant beta 3 described here suggest that the difference between the cytoplasmic domains of these subunits arises as a result of an alternative mRNA splicing. These cytoplasmic domains may provide alternative means for the beta 3 integrins to interact with cytoskeletal components.