Plasminogen supports tumor growth through a fibrinogen-dependent mechanism linked to vascular patency by Joseph S. Palumbo, Kathryn E. Talmage, Hong Liu, Christine M. La Jeunesse, David P. Witte, and Jay L. Degen Blood Volume 102(8):2819-2827 October 15, 2003 ©2003 by American Society of Hematology

Plasminogen is a determinant of LLC tumor growth through a mechanism linked to anatomic location. Plasminogen is a determinant of LLC tumor growth through a mechanism linked to anatomic location. (A) Tumor mass after LLCGFP transplantation into the dorsal subcutis of control (•, n = 4) and plasminogen-deficient (○, n = 5) mice was estimated by serial calipation. Tumor sizes were not significantly different between the cohorts at any time point. Direct tumor weight measurements made at the time of animal killing were also not significantly different (data not shown). Tumor size after LLCGFP transplantation into the footpad was estimated by serial calipation of tumor width (B) and direct tumor weight at day 14 (C) in the footpad of control (•, n = 6) and plasminogen-deficient (○, n = 7) mice. Footpad tumors in plasminogen-deficient mice were significantly smaller than in control animals at every time point, including the final weight measurement (P < .02 at each comparison; Mann-Whitney U test). Horizontal lines in panel C indicate median values. Error bars indicate standard deviation from the mean. Joseph S. Palumbo et al. Blood 2003;102:2819-2827 ©2003 by American Society of Hematology

Gross appearance of LLC footpad tumors in mice with single and combined deficits in plasminogen and fibrinogen. Gross appearance of LLC footpad tumors in mice with single and combined deficits in plasminogen and fibrinogen. A normal, non-tumor–bearing foot (A) is shown along with examples of LLCGFP footpad tumors 14 days after inoculation into control (B), Plg– (C,F), Fib– (D), and Plg–/Fib– (E) mice. Despite the minimal difference in size between the Plg– tumor-bearing foot (C) and a non-tumor–bearing foot (A), the presence of tumor tissue could be easily appreciated using a fluorescence-equipped stereomicroscope (F). Note that the foot in panel F is the same foot shown in panel C and is photographed at the same magnification. Joseph S. Palumbo et al. Blood 2003;102:2819-2827 ©2003 by American Society of Hematology

LLC tumor growth is restored in the footpad of plasminogen-deficient mice by genetic elimination of fibrinogen. LLC tumor growth is restored in the footpad of plasminogen-deficient mice by genetic elimination of fibrinogen. (A) Weights of footpad tumors in cohorts of control •, Plg– ○, Fib– ▪, and Plg–/Fib– □ mice analyzed in parallel are shown 14 days after transplantation of LLCGFP. The weight distribution of footpad tumors was significantly smaller in Plg– mice than in control animals (P < .02), consistent with 2 other independent experiments. The elimination of fibrinogen relieved the impediment in tumor growth in Plg– mice. Indeed, footpad tumors from Plg–/Fib– mice were larger than those derived from control mice, though this difference was not statistically significant (P = .1). Plasminogen was not a significant determinant of tumor mass in the absence of fibrinogen (P = .7). Horizontal lines indicate median values. (B) Serial tumor width measurements demonstrated similar tumor expansion in fibrinogen-deficient mice with (▪, n = 7) and without (□, n = 7) plasminogen (P > .1 at each time point). All P values were established using the Mann-Whitney U test. Error bars indicate standard deviation from the mean. Joseph S. Palumbo et al. Blood 2003;102:2819-2827 ©2003 by American Society of Hematology

Histologic features of footpad tumors are similar in plasminogen-deficient and control mice. Histologic features of footpad tumors are similar in plasminogen-deficient and control mice. Shown are representative hematoxylin-eosin–stained sections of footpad LLCGFP tumors collected from control (A), Plg– (B), Fib– (C), and Plg–/Fib– (D) mice 14 days after initial transplantation. Note that the overall microscopic features of the tumor tissue were unremarkable within animals of each genotype and suggest no obvious basis for the impediment in tumor growth in Plg– mice. LLCGFP tumors grew as sheets of anaplastic cells that often invaded normal structures, including adjacent muscle and bone. Small areas of necrosis (*) were also present, but these were not appreciably different between animals of each genotype. Dermal surfaces of the footpads are indicated by arrows in panels A to D. Immunohistochemical staining for the endothelial marker PECAM (brown stain) revealed abundant vessels in footpad tumors derived from control (E) and plasminogen-deficient mice (F). Fibrin(ogen) immunostains revealed relatively scant and focal fibrin(ogen) deposition in control (G) and Plg– (H) mice (brown staining highlighted by arrows). Sparse fibrin(ogen) deposition was generally peritumoral or was associated with small areas of necrosis. Sections used for immunohistology were counterstained with hematoxylin. Joseph S. Palumbo et al. Blood 2003;102:2819-2827 ©2003 by American Society of Hematology

Occlusive microvascular thrombi are prominent in LLC footpad tumors derived from plasminogen-deficient mice. Occlusive microvascular thrombi are prominent in LLC footpad tumors derived from plasminogen-deficient mice. Electron microscopic evaluation of footpad tumor vasculature revealed widespread microvascular thrombi within tumors derived from Plg– mice (A-C) but not tumors collected from control (D), Fib– (E), or Plg–/Fib– (F) animals 14 days after initial tumor cell transplantation. These representative sections from Plg– mice include examples of vessels with small platelet aggregates (A) and vessels with advanced, highly occlusive thrombi (B-C). Occlusive platelet thrombi (arrowheads) were found in areas populated with viable tumor cells (*) and where the endothelium appeared essentially intact (arrows). Note that occlusive intravascular thrombi were not a feature of the rapidly growing tumors of control mice and Fib– mice. More significantly, no such platelet aggregates were observed in plasminogen-deficient mice lacking fibrinogen (F), animals that exhibited no impediment in tumor growth. Horizontal bars in each micrograph indicate 8 μm. Joseph S. Palumbo et al. Blood 2003;102:2819-2827 ©2003 by American Society of Hematology

Microvascular patency is maintained within LLC tumors of plasminogen-deficient mice when the tumors develop within the dorsal subcutis. Microvascular patency is maintained within LLC tumors of plasminogen-deficient mice when the tumors develop within the dorsal subcutis. Representative transmission electron micrographs illustrating dorsal skin tumor vasculatures of control (A), Plg– (B), Fib– (C), and Plg–/Fib– (D) mice. Note that in contrast to tumors grown in the footpad, tumors within the dorsal subcutis had no evidence of microvascular thrombus formation, regardless of animal genotype. Viable tumor cells are highlighted with asterisks, and endothelium is indicated with arrows. Horizontal bars in each micrograph indicate 8 μm. Joseph S. Palumbo et al. Blood 2003;102:2819-2827 ©2003 by American Society of Hematology

Tumor cell mitotic index is diminished in footpad tumors derived from plasminogen-deficient mice relative to control animals. Tumor cell mitotic index is diminished in footpad tumors derived from plasminogen-deficient mice relative to control animals. Mitotic figures (arrowheads) were easily discernible in hematoxylin-eosin–stained sections of footpad tumors from control (A) and plasminogen-deficient (B) mice. (C) Quantitative analysis of mitotic indices within footpad tumors of mice with single and combined deficits in plasminogen and fibrinogen. The mitotic index observed in plasminogen-deficient (*) mice was significantly less than that observed in control animals (P < .02). In contrast, no significant difference in mitotic index was observed in comparative analyses of control animals and plasminogen-deficient mice that also lacked fibrinogen (P > .8). Mitotic indices were also similar in control and fibrinogen-deficient mice (P > .9). All P values were established using the Mann-Whitney U test. Error bars indicate standard deviation from the mean value for 6 mice. Joseph S. Palumbo et al. Blood 2003;102:2819-2827 ©2003 by American Society of Hematology