Tumor biology: use of tiled images in conjunction with measurements of cellular proliferation and death in response to drug treatments.

Tumor growth is dependent on the balance between cell proliferation and cell death, and these events occur heterogenously within an individual tumor. We present a methodology that provides integrative information about cell kinetics, cell death, and cell growth within individual tumors in animals treated with cytotoxic chemotherapeutic agents. Using HCT-116 and NCI-H460 cells, human colonic adenocarcinoma and non-small cell lung cells, respectively, traditional xenograft studies were performed. The tumor-bearing animals were treated with cyclophosphamide (Cytoxan), gemcitabine (Gemzar), or mitomycin C, and extensive analysis of the tumors was studied. Cell kinetics were evaluated by measuring the apoptotic and proliferation indices. The ability to image an entire tumor section using "tiling" by creating a large montage from many high-resolution images makes it possible to identify regional differences within areas of tumor and to demonstrate differences in these tumor regions after treatment with selected chemotherapeutic agents. Two specific areas within tumors have been identified: (a) areas of viable cells within the cell cycle, determined by bromodeoxyuridine and/or morphological characteristics determined by hematoxylin staining; and (b) areas of necrosis determined by the absence of bromodeoxyuridine and proliferating cell nuclear antigen-labeled cells coupled with morphological changes. By standardizing the tumor size to 100 mm2, different patterns of tumor responses to chemotherapeutic agents were determined. By creating such tiled images and by quantitating cell cycle kinetics, it is possible to gain a more complete understanding of tumor growth and response to treatment, leading to the development of more reliable methods for assessing the clinical behavior of anticancer drugs.

Novel small molecule alpha v integrin antagonists: comparative anti-cancer efficacy with known angiogenesis inhibitors.

Recent evidence supports the involvement of integrins in angiogenesis: blockade of alpha v beta 3 and alpha v beta 5 integrins disrupts angiogenesis leading to decreased blood vessel formation and hence decreased tumor growth. We hypothesized that av antagonists could inhibit tumor growth in tumor cells devoid of alpha v beta 3 integrins. We evaluated SM256 and SD983, novel small molecules that are specific av antagonists in mouse models of angiogenesis and tumorigenesis, and compared them with standards: TNP470, a fumagillin analog in the clinic, and flavopiridol, a cell cycle kinase inhibitor. In vitro SM256 was a selective alpha v beta 3 inhibitor with an IC50 = 4nM, and the affinity of SD983 against the mouse endothelial alpha v beta 3 integrin yielded an IC50 = 2nM and an IC50 = 54nM against alpha v beta 5. In the mouse Matrigel model of angiogenesis SM256 decreased blood vessel formation (hemoglobin content) with an ED50 = 0.055 ug/kg/day, tenfold more potent than TNP470. SG545, an ester of SD983, decreased blood vessel formation with an ED50 = 6 ug/kg/day, while flavopiridol ED50 = 18 ug/kg/day. In the mouse xenograft model, using human colon carcinoma RKO cells that do not express alpha v beta 3 but express alpha v beta 5, tumor growth was inhibited by SG545 (10 mg/kg/day) and flavopiridol (5 mg/kg/every other day) 40% and 70%, respectively (p < 0.05). Although the proliferative index (measured by BrdU incorporation) was not significantly changed with SM256, SG545 or flavopiridol (29-32%), the apoptotic index increased significantly (p < 0.05) in the SM256 and SG545-treated groups (2.3-2.7%) compared with controls (1.1%), suggesting increased cell death contributed to decreased tumor volumes. Neovascularization decreased with SM256 and SG545 treatment. The data demonstrate that potent selective av antagonist can target endothelial cells, tumor cells, inhibit angiogenesis and inhibit tumor growth.

Calcium release from rod outer segments: evidence for a cGMP-sensitive calcium binding protein.

Numerous studies investigating the cGMP-gated cation conductance in rod disk membranes have purported to measure efflux of Ca2+ entrapped in rod disk membrane vesicles. We have utilized sonication and osmotic shock as additional tests for sensitivity of cGMP- and A23187-induced Ca2+ release to elimination of the transvesicular Ca2+ gradient. We find that 1) Treatment with sonication or osmotic shock in low Ca2+ medium does not release Ca2+ from either native cGMP/Ca2(+)-loaded vesicles or solubilized, reconstituted "Ca2(+)-loaded" vesicles, 2) 70-100% of the cGMP-induced "flux" and 90-100% of the A23187-induced Ca2+ "flux" is insensitive to elimination of the Ca2+ gradient by sonication or osmotic shock in low Ca2+ medium, and 3) total amount of releasable Ca2+ is related to membrane surface area rather than vesicle entrapment volume. We conclude that 1) A23187 disrupts binding of Ca2+ to proteins and phospholipids as well as releasing entrapped Ca2+ and 2) a large fraction of the cGMP-induced release observed in rod disk vesicles is due to release of bound Ca2+.

Transport of immunoglobulin G by endothelial vesicles in isolated capillaries.

The ability of capillary endothelial vesicles to transport immunoglobulin G (IgG) across the capillary wall was investigated. Isolated capillary endothelial cells were found to endocytose fluorescent-labeled IgG via a bulk fluid-phase mechanism. When placed in marker-free medium the cells released all previously ingested label. Cationization of IgG causes greater amounts of the protein to be ingested as compared to normal IgG. Cationized IgG is probably endocytosed via an adsorptive mechanism, allowing greater efficiency in its uptake. It is concluded that IgG can be transcytosed by isolated endothelial cells of capillary origin and that the charge of the IgG molecule is important in determining its mode of uptake.

Micropinocytic ingestion of glycosylated albumin by isolated microvessels: possible role in pathogenesis of diabetic microangiopathy.

Microvessels isolated from rat epididymal fat exhibit differential vesicular ingestion rates for unmodified and non-enzymatically glycosylated rat albumin. While unmodified rat albumin is excluded from ingestion by endothelial micropinocytic vesicles, glycosylated albumin is avidly taken up by endocytosis. Interaction of albumin and glycosylated albumin with endothelium was studied with a double-label fluorescence assay of micropinocytosis. When glycosylated albumin was present at a concentration of 6% with respect to total albumin (the level found in "non diabetic" serum), only glycosylated albumin was ingested. At higher concentrations of glycosylated albumin (those found in diabetic serum), both albumin and glycosylated albumin are ingested. Glycosylation of endothelial membrane components results in stimulated ingestion of glycosylated albumin, persistent exclusion of unmodified albumin, and unaltered micropinocytic ingestion of native ferritin. These results indicate that nonenzymatic glycosylation of serum albumin may result in rapid vesicle-mediated extravasation of albumin. Chronic microvascular leakage of glycosylated albumin could contribute to the pathogenesis of diabetic microangiopathy.

Structural and functional consequences of increased tubulin glycosylation in diabetes mellitus.

The extent of in vitro nonenzymatic glycosylation of purified rat brain tubulin was dependent on time and glucose concentration. Tubulin glycosylation profoundly inhibited GTP-dependent tubulin polymerization. Electron microscopy and NaDodSO4/polyacrylamide gel electrophoresis showed that glycosylated tubulin forms high molecular weight amorphous aggregates that are not disrupted by detergents or reducing agents. The amount of covalently bound NaB3H4-reducible sugars in tubulin recovered from brain of streptozotocin-induced diabetic rats was dramatically increased as compared with tubulin recovered from normal rat brain. Moreover, tubulin recovered from diabetic rat brain exhibited less GTP-induced polymerization than tubulin from nondiabetic controls. The possible implications of these data for diabetic neuropathy are discussed.

Endocytosis and exocytosis of transferrin by isolated capillary endothelium.

Vesicular transport of solutes across capillary walls may be regulated by specific solute-endothelial interactions. Little data is available on the vesicular transport of serum proteins which may transit the capillary wall in situ. Capillaries were isolated from epididymal fat and incubated in fluorescent-labeled transferrin and radiolabeled sucrose. Endocytosis and exocytosis of these tracers were quantitated on a picomolar basis over timed intervals and standardized against the amount of endothelial DNA present in the isolate. The rate of vesicular endocytosis of transferrin was 6-7 times greater than that of sucrose indicating a mechanism of selection for transferrin. Endocytosis as a function of external concentration exhibited complex kinetics for transferrin that was consistent with an adsorptive component and a fluid component. Sucrose uptake appeared to be simple fluid endocytosis but with a rate-limiting concentration at 500-600 microM. Vesicular exocytosis of both solutes from preloaded capillaries appeared to occur more rapidly than their endocytosis. This was probably not due to different rates of filling and emptying of attached vesicles nor to an intrinsic difference in rates of vesicle interiorization and refusing with the plasma membrane. Different rates of endocytosis and exocytosis may only be apparent since exocytosis of marker before the capillaries reach ingestion equilibrium would reduce the measured uptake rate.