VEGF is upregulated in a neuroblastoma and hepatocyte coculture model.

BACKGROUND: We hypothesize that angiogenic factors are altered by the interaction between neuroblastoma cells and host tissues. MATERIALS AND METHODS: Human Chang hepatocytes and human neuroblastoma cells are cultured separately and in a noncontact, coculture system. Immunostaining for VEGF is performed on the cells. ELISA is used to detect vascular endothelial growth factor (VEGF), basic fibroblast growth factor, and interleukin-8 in the conditioned media. Human umbilical vein endothelial cells (HUVEC) are cultured with standard medium (control) and hepatocyte, neuroblastoma, and coculture conditioned media. After 48 and 72 h, cells are counted to determine proliferation. Finally, VEGF-blocking antibody is added to the HUVEC cultures with the conditioned media. RESULTS: VEGF is markedly elevated in the coculture medium compared to the media from hepatocytes or neuroblastoma grown alone [412.2 +/- 52 vs 235 +/- 35 or 74.5 +/- 28.5 (pg/10(6) cells), P < 0.05]. Other growth factors are almost undetectable in any of the media. Immunostaining for VEGF in the cocultured hepatocytes is decreased by almost 50%, but VEGF immunostaining is increased fourfold in the cocultured neuroblastoma cells. A significant increase in cell proliferation is seen at both 48 and 72 h when HUVEC are cultured with the coculture media. Cell proliferation is blocked with the addition of anti-VEGF antibody. CONCLUSION: The interaction of neuroblastoma with hepatocytes results in an increased production of VEGF. It stimulates endothelial cell proliferation and may enhance the tumor's metastatic potential in an autocrine fashion.

Neuroblastoma and hepatocyte coculture conditioned media alter apoptosis.

BACKGROUND: Neuroblastoma is a childhood tumor that often displays unusual biological behavior. The tumor may present with widespread metastases that are unresponsive to aggressive treatment. At other times, both the metastases and the primary tumor may spontaneously regress without treatment. Apoptosis, or programmed cell death, is thought to play a role in the dichotomous behavior of neuroblastoma. We hypothesize that neuroblastoma cells will interact with host tissues to release mediators that affect apoptosis. MATERIALS AND METHODS: Human neuroblastoma cells and human Chang hepatocytes are grown in a noncontact, coculture system. After incubation for 4 days, the medium from the coculture system is collected. Neuroblastoma cells and Chang hepatocytes are then plated separately with the conditioned medium and their own standard growth medium as controls. After 4 days, these cells are harvested and cytospins made for immunostaining. Tumor necrosis factor alpha (TNF-alpha), Fas ligand, and Bcl-2, are measured with immunohistochemistry. Apoptosis is detected with the TUNEL method. Immunostaining data are interpreted with computer image analysis and reported as stain index. TUNEL data are reported as percentage apoptotic cells. All data are reported as means +/- SEM. Statistical analysis is performed and P < 0.05 considered significant. RESULTS: Chang hepatocytes grown in the coculture conditioned media have an increase in TNF-alpha and Fas ligand. The neuroblastoma cells have a significant decrease in Fas ligand. There is a significant increase in the number of apoptotic hepatocytes when they are cultured in the conditioned media. In contrast, the neuroblastoma cells grown in the coculture conditioned media show no increase in apoptosis. Finally, Bcl-2 is significantly increased in the neuroblastoma cells cultured in the conditioned media. CONCLUSIONS: Neuroblastoma cells grown in coculture conditioned media show increased expression of Bcl-2 and decreased Fas ligand levels. These changes should diminish apoptosis activity in the tumor cells. In contrast, the conditioned media induce elevated levels of proapoptotic mediators in the Chang hepatocytes. A tumor's ability to successfully metastasize may be dependent on mediators generated in the tumor-host interaction, and may not be just an independent characteristic of the tumor itself.

The effect of leukocyte infiltration on apoptosis in an in vitro thermal injury bioartificial living skin equivalent model.

An in vitro bioartificial skin construct (BSC) model was studied to see how inflammatory infiltration affects apoptosis in skin that has been thermally injured. The BSC was used as a target organ. Control BSCs without leukocytes (CON) were burned (BCON) by scalding with phosphate-buffered saline heated to 70 degrees C for 6 seconds, and they were then cooled with room temperature phosphate-buffered saline for 15 seconds. Human alloimmunocytes were added to CON to create rejection cultures (REJ) and to BCON to create burned rejection cultures (BREJ). Slides were stained with hematoxylin and eosin and anti-Lewis antibody. In situ labeling of apoptosis was measured by terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL). Those sections that were immunostained for Lewis Y were analyzed for intensity stain index (ISI = [sigma P0 x I0]/ total tissue area in pixels). TUNEL was quantified with the following equation: no. of total apoptotic cells / total tissue area. Necrosis and blister formation in the epidermal layer were evident in BCON and BREJ. Pyknosis and nuclear fragmentation-indicators of apoptosis-were also present. Phagocytosis of keratinocytes by leukocytes was seen in REJ and BREJ. Immunostaining showed greater expression of Lewis Y antigen, as determined by ISI, in REJ as opposed to CON (58.2+/-2.3 vs. 36.4+/-2.3, respectively, P<.001), but no significant difference was found between BCON and BREJ (55.0+/-5.7 vs. 60.5+/-3.4, respectively) and REJ and BREJ (58.3+/-2.3 vs. 60.5+/-3.4, respectively). TUNEL staining indicated the presence of apoptosis as follows: REJ versus CON (0.0015+/-0.0002 vs. 0.0003+/-0.0001, respectively, P<.001); BREJ versus BCON (0.0031+/-0.0006 vs. 0.0018+/-0.0004, respectively, P<.05); REJ versus BREJ (0.0015+/-0.0002 vs. 0.0031+/-0.0006, respectively, P = .007). The presence of leukocytes and thermal injury induces apoptosis in BSC. The combination of these two variables results in increased apoptosis as determined by TUNEL. These findings suggest that a common pathway for skin injury may include inappropriate regulation of apoptosis exacerbated by a mechanism that includes inflammatory cellular infiltration.

Fas-mediated induction of hepatocyte apoptosis in a neuroblastoma and hepatocyte coculture model.

BACKGROUND: We have previously demonstrated an increase in hepatocyte apoptosis when they are cocultured with neuroblastoma cells. Death receptors in the tumor necrosis factor (TNF) family such as TNFR1 and Fas have been identified as regulators of apoptosis and may be responsible for the altered regulation of apoptosis seen in our coculture model. To evaluate the effects of released factors and remove the potential alterations induced by direct contact, a noncontact coculture system was used to study the interaction between hepatocytes and neuroblastoma cells. METHODS: Human Chang hepatocytes (HC) were plated onto Falcon cell culture inserts with 0.45-micrometer pores in the permeable membrane. Human neuroblastoma cells (NB-IMR-32) were seeded into wells of the Falcon companion plate. After 24 h, inserts containing HC were placed into wells containing NB cells and incubated for 4 days. This provided a coculture environment without actual cellular contact. Immunohistochemical staining for TNFalpha, Fas, and Fas ligand (Fas-L) was performed. Apoptosis was detected via the TUNEL method. Images were analyzed with ImagePro-Plus. Statistical analyses were done with significance determined at P < 0.05. RESULTS: Chang hepatocytes demonstrated a significant increase in the levels of TNF, Fas, and Fas-L when cocultured with neuroblastoma cells (P < 0.005). In addition, the cocultured hepatocytes had a 20-fold increase in the apoptotic rate (P < 0.001). Neuroblastoma cells had no demonstrable level of Fas or TNF when grown alone and in cocultures. Neuroblastoma cells that were grown alone had an elevated level of Fas-L, but this level diminished by 44% when cocultured with hepatocytes (P < 0.001). CONCLUSION: An upregulated TNF/Fas receptor-ligand system may be responsible for increased apoptosis in hepatocytes when cocultured with neuroblastoma. This upregulation may be due to release of neuroblastoma-derived Fas ligand into the media. Tumors may alter the regulation of apoptosis in surrounding tissues via the death receptors.

Pharmacologic inhibition of collagen in an experimental model of subglottic stenosis.

Subglottic stenosis occurs as a complication of prolonged endotracheal intubation secondary to inflammation with development of scar tissue and subsequent fibrosis. Collagen I and III levels increase during the healing process. Steroids alter the inflammatory response, decreasing recruitment of macrophages and fibroblasts. Beta-aminopropionitrile (betaAPN) inhibits the development of collagen cross-linking. A mechanism that would minimize hypertrophic scarring was sought. Eighteen dogs were anesthetized, had tracheostomies performed, and later had cautery of the mucosa and inner layer of the cricoid cartilage. Of 18 survivors, 6 animals were used as controls, 6 animals received oral Decadron, 2 mg/d, and 6 animals received oral betaAPN, 40 mg/d. There were 9 early deaths--5 in the steroid group. Animals were painlessly sacrificed, and the specimens were sectioned at the cricoid cartilage level and were stained immunohistochemically for antibodies to collagen types I to VI. Analysis of the area of scar and the intensity of stain was performed with Mocha image analysis software. Collagen III increased in control animals to 14.38 +/- 1.85 (intensity stain index), but this reaction was reduced by betaAPN (5.77 +/- 1.78, p < .01). Steroids had no significant effect on formation of any type of collagen. Lathyrogens (betaAPN) may offer a pharmacologic tool to reduce scar tissue.

In vitro bioartificial skin culture model of tissue rejection and inflammatory/immune mechanisms.

We hypothesized that an in vitro bioartificial skin rejection model using living LSEs grown in tissue culture could be developed for the study of autologous, allogenic, and/or xenogeneic inflammatory/immune mechanisms and topical immunosuppressive drugs. Human fibroblasts were mixed with type 1 rat-tail collagen to form a matrix (4 to 5 days), on which human keratinocytes were seeded. After a keratinocyte monolayer formed, CT cultures were raised to the air-liquid interface for continued growth. In the REJ LSE model, immunocytes isolated from human blood were seeded on top of the NHEK monolayer at the time of air-lifting. Thickness measurements of the acellular keratin and keratinocyte layers, and nuclear/cytoplasmic ratios, in both CT and REJ were made using digital image analysis. Immunostaining with anticytokeratin demonstrated a viable, keratin-producing epidermal layer; staining with anti-TGF-beta suggested a role for this cytokine in the rejection or wound-healing process. The LSE appeared histologically similar to normal human epidermis. Immunocytes added to the REJ cultures caused an obvious rejection response and were clearly identifiable in the gels as CD45+ staining cells. The LSE model appears promising for the study of immune/inflammatory mechanisms, thermal injury, screening antirejection agents that might be applied topically and as an in vitro replacement for skin graft studies in animals.

Graft-versus-host disease in extremity transplantation: digital image analysis of bone marrow in situ.

The development of either unstable immune chimerism and lethal graft-versus-host disease or stable immune chimerism and alloimmune tolerance can result from extremity transplantation. LBN rats served as recipients of Lewis vascularized extremity (limb) transplants. Recipients received no immune suppression and were immunologically unmodified. The bone marrow of transplanted and contralateral limbs was analyzed in situ for distribution of nuclei, nuclear area, and staining intensity by digital image analysis and computerized morphometry. Cellularity was significantly increased, and fat content was significantly decreased in the graft-versus-host disease animals' marrow versus the tolerant animals' marrow for both the transplanted and contralateral limbs. Tolerant animals demonstrated significantly increased nuclear staining compared with graft-versus-host disease animals for both transplanted and contralateral limbs. Additionally, there were significant changes between the host and the transplanted limbs for marrow intensity and cellularity within tolerant and graft-versus-host disease groups. The significant differences in the graft-versus-host disease-positive recipients suggested that both autoimmune dysregulation and alloimmune reactions were in effect for both donor and host bone marrow compartments. Cellular alterations in the tolerant recipients' marrow were suggestive of subtle subclinical graft-versus-host responses.