Search

BACKGROUND The functional significance of cytokines expressed in situ by tumor cells and tumor infiltrating lymphocytes (TIL) in human colon carcinomas is largely unknown. METHODS We assessed TIL expression of interleukin-2 (IL-2), IL-4, tumor necrosis factor-alpha (TNF-alpha), interferon-gamma (IFN-gamma), granulocyte-macrophage colony stimulating factor (GM-CSF), IL-10, and transforming growth factor-beta (TGF-beta), and tumor cell expression of IL-10 and TGF-beta in situ in 49 primary colon carcinomas and 20 metastases using immunohistochemistry. RESULTS The percentage of primary colon carcinoma samples in which > 20% of TIL expressed each cytokine was as follows: IL-4: 47%; TNF-alpha: 22%; TGF-beta: 10%; IFN-gamma: 6%; IL-2:2%; IL-10: 0%; and GM-CSF: 0%. Lymphocytes more commonly infiltrated colon carcinoma primaries than metastases, and TIL expression of IL-4 and TNF-alpha was more common in primary than metastastatic carcinomas. Expression of TNF-alpha by even a small proportion (> or = 3%) of the TIL in a colon carcinoma specimen was associated with better overall survival (P = 0.01) when compared with patients with little or no TIL TNF-alpha expression (5-year survival 82% vs. 47%). Expression of IL-4 by > or = 20% of colon carcinoma TIL was also associated with improved survival (P = 0.01; 5-year survival 87% vs. 50%). The expression of IL-10 or TGF-beta by colon carcinoma TIL or colon tumor cells themselves was not associated with impaired survival. Benign epithelial cells stained positively for IL-10 and TGF-beta more frequently than tumor cells (P < 0.001). CONCLUSIONS There are differences between the immune microenvironment of primary tumors and metastases. Although IL-10 is expressed by colon carcinoma cells and TIL, it is unlikely that it plays an important immunosuppressive role. TNF-alpha and IL-4 are commonly expressed by colon carcinoma TIL and both are associated with improved survival.

Mechanisms of coagulation activation in situ were studied by means of immunohistochemical techniques applied to surgically resected primary adenocarcinomas and squamous cell carcinomas of the lung. Findings in these two histologic types were similar. Double-labeling techniques using macrophage-specific antibody together with antibody to either tissue factor, factor VII, factor X, or factor V revealed coincident staining for each of these coagulation factors on tumor-associated macrophages. Staining of tumor cells for these factors was rare and inconsistent. Both macrophages and fibroblasts in the tumor connective tissue stained for the a subunit of factor XIII. Fibrinogen was abundant throughout the tumor connective tissue, but staining for fibrin and D-dimer cross-linked sites of fibrin was restricted to areas adjacent to macrophages, indicating that thrombin was generated in association with tumor macrophages but not with tumor cells. By contrast, tumor cells stained diffusely for urokinase-type plasminogen activator and focally for thrombomodulin. These findings contrast with those reported previously for small cell carcinoma of the lung and suggest that coagulation activation in adenocarcinoma and squamous cell carcinoma of the lung may occur indirectly through activation of certain host cells such as macrophages. By contrast, tumor cell plasminogen activator may mediate certain aspects of the malignant phenotype in these tumor types.

The DNA content of 53 primary and locally metastatic pulmonary neuroendocrine carcinomas was retrospectively determined by flow cytometric analysis of nuclei released from paraffin-embedded tissues and compared with histologic subtypes and clinical survival. Forty-one percent (9 of 22) of tumors classified as well-differentiated neuroendocrine carcinoma (WDNC) were aneuploid. In contrast, 85% (17 of 20) of small cell neuroendocrine carcinoma (SCNC), and 73% (8 of 11) of intermediate cell neuroendocrine carcinoma (ICNC) were aneuploid. DNA content was not a significant independent predictor of patient survival within each histologic subtype (P greater than 0.05 for all subtypes). Yet, when all cases were examined as a single group, diploid cases showed better clinical survival than cases with DNA aneuploidy (P less than 0.02). However, the survival advantage for diploid DNA content did not quite achieve statistical significance when only cases of limited stage disease were analyzed (0.10 greater than P greater than 0.05). Histologic subtype was also a prognostic indicator, with WDNC cases showing a significantly longer survival than either SCNC (P less than 0.05) or ICNC (P less than 0.02), for those cases with limited-stage disease. These results indicate information on clinical staging and the histologic subtype are important parameters to compare with DNA content analysis in determining independent prognostic factors in neuroendocrine or small cell carcinoma of the lung.

To explore mechanisms of coagulation activation in adenocarcinoma of the prostate, the occurrence and distribution of components of coagulation and fibrinolysis pathways in situ were studied by means of immunohistochemical techniques applied to frozen sections of fresh malignant and benign hyperplastic prostatic tissue obtained at transurethral resection. Fibrinogen was distributed throughout the perivascular and tumor connective tissue in both malignant and benign disease but was not present in adjacent areas of normal prostate. Antibodies specific for fibrin and D-dimer crosslink sites stained vascular endothelium focally in both malignant and benign tissues. Both neoplastic cells and benign hyperplastic glandular epithelial cells stained weakly and in a patchy distribution for tissue factor and focally for low-molecular-weight urokinase-type plasminogen activator. Focal staining of vascular endothelium was also observed for tissue plasminogen activator and plasmin-antiplasmin complex neoantigen. By contrast, no tissue staining was observed for factor VII, factor X, factor XIII "a" subunit, high-molecular-weight urokinase-type plasminogen activator, plasminogen activator inhibitors 1 to 3, protein C, and protein S. Thus, the similarity in findings between benign hyperplastic and neoplastic prostate tissue, the lack of either an intact tumor cell-associated coagulation pathway or fibrin formation, and the presence of fibrin on vascular endothelium are consistent with the concept that coagulation activation in prostatic cancer may not be due to a direct effect of the tumor cells on the clotting mechanism. Rather, such activation may be induced by a soluble tumor product that activates procoagulant activity on certain host (for example, vascular endothelial) cells. These findings, together with the lack of effect of warfarin anticoagulation on the clinical course of patients with prostatic cancer, contrast with findings in certain other tumor types and suggest that coagulation activation may not contribute to progression of adenocarcinoma of the prostate.

Components of coagulation and fibrinolysis reactions were identified in situ by immunohistochemical staining in fresh frozen sections of small cell carcinoma of the lung tissue. Tumor cells stained positively for tissue factor, a protein that is capable of activating the extrinsic pathway of coagulation (the components of which have been seen within small cell carcinoma of the lung [SCCL] tissue), and for proteins C and S antigens. Fibrin was seen in a focal distribution at the host-tumor interface, indicating that thrombin had acted upon the fibrinogen found throughout the tumor stroma. Staining with a neoepitope-specific antibody, which does not discriminate between fibrinogen fragment D and fibrin fragment D-dimer, was similar to that of the fibrin antibody. High molecular weight urokinase-type and tissue-type plasminogen activators were seen in vascular endothelium, but neither existed within the tumor. Low molecular weight urokinase was found in rare isolated foci of tumor cells primarily adjacent to areas of necrosis. Plasminogen activator inhibitor-3 occurred in tumor cell cytoplasmic blebs and in necrotic tumor cells, but plasminogen activator inhibitors 1 and 2 were not seen. Our data suggest a mechanism for thrombin generation and fibrin formation within SCCL tissues that could support cell proliferation, stroma formation, and preservation. These features could be conductive to perpetuation of this tumor and conceivably could form the basis of the beneficial effects of antithrombotic therapy seen in SCCL.

Primary neuroendocrine tumors of the larynx appear to be extremely rare. We report a case of well-differentiated neuroendocrine carcinoma arising in the larynx of a 61-year-old white woman. This case was initially diagnosed as a lymph node metastasis in the neck from an unknown primary carcinoma. Extensive workup subsequently revealed that the larynx was the primary site of the tumor. An electron microscopic study of both the primary and metastatic tumors showed numerous neurosecretory-type granules. Immunohistochemical studies revealed positive immunoreactivity against calcitonin, somatostatin, and ACTH. However, the tumor also displayed focal exocrine differentiation and mucosubstance production. The patient underwent a supraglottic laryngectomy and radical neck dissection and was free of tumor 22 months postoperatively.

Through immunohistochemical techniques, blood coagulation factors were identified in situ in fresh frozen sections of small cell carcinoma of the lung. Prothrombin/thrombin, factor VII, factor X, and antithrombin III were present in intercellular spaces and associated with tumor cells. Factor IX, factor XI, prekallikrein, and high molecular weight kininogen were identified as being associated with tumor cells but did not exist in intercellular spaces. Variable connective tissue staining but no tumor-related staining was observed for factor V, factor VIII-related antigen, factor XII, the B subunit of factor XIII, alpha 1-antitrypsin, alpha 2-macroglobulin, or alpha 2-antiplasmin. Neither consecutive tissue nor the tumor manifested platelet Ib and IIbIIIa surface glycoproteins. These divergent staining patterns suggested that the detected clotting factors had not merely diffused from permeabilized blood vessels, but were selectively localized in situ. While conditions may exist within tumor tissue that both retard and promote thrombin generation, we propose that interactions between the observed coagulation factors ultimately lead to local thrombin formation, which is responsible for the conspicuous fibrin deposits already described in small cell carcinoma of the lung. Thrombin formed locally might contribute to progression of this tumor. Inhibition of local thrombin formation by warfarin therapy could explain the beneficial effects of warfarin therapy in treating small cell carcinoma of the lung.

Thrombin-generated cleavage sites of human fibrinogen have been identified adjacent to viable tumor cells in fresh frozen sections of small cell carcinoma of the lung (SCCL) by means of immunohistochemical techniques using mouse monoclonal antibodies to the N-terminal peptides of the fibrinogen alpha and beta chains. These results indicate that thrombin is generated in situ in this tumor type. Based on previous evidence for the existence of an initiator of coagulation together with coagulation factor intermediates associated with viable SCCL tumor cells in situ, and also for the favorable effects of anticoagulant therapy with warfarin in SCCL, we postulate that local tumor cell-induced thrombin formation may contribute to self-regulated progression of SCCL through deposition of fibrin and stimulation of cell proliferation. These results suggest novel treatment strategies for this particular tumor type and justify efforts to identify other tumor types in which similar mechanisms exist.

Immunoperoxidase staining for Regan isoenzyme of alkaline phosphatase was performed on cryostat sections of five human tumor tisssues. With a direct immunoperoxidase staining for the localization of Regan isoenzyme at the light and electron microscope levels, sections previously fixed with 0.05 M phosphate-buffered 4% paraformaldehyde were reacted with rabbit antisera to human placenta alkaline phosphatase conjugated to horseradish peroxidase. Comparison of conventional histochemistry and immunohistochemistry for Regan isoenzyme indicated that strong specific immunoperoxidase staining appeared on the cell membrane surface, and a diffuse one, in the cytoplasm of lung and colon cancer tissue cells showing L-phenylalanine-sensitive alkaline phosphatase. No immunoperoxidase reaction was obtained in tumor cells showing sensitivity to L-homoarginine or lacking aklaline phosphatase activity.

Nodular skin lesions consisting of eosinophilic cell infiltrates developed in a 45-year-old man with metastatic islet cell amphicrine carcinoma of the pancreas. These infiltrates did not fit previously described eosinophilic skin conditions. Although the tumor responded to conventional chemotherapy, the skin lesions did not. They did, however, respond completely to radiation therapy. This clinical constellation has not been reported previously and may represent a new paraneoplastic syndrome.