Human Small Cell Lung Carcinoma and Carcinoid Tumor Regulate Dendritic Cell Maturation and Function Nora S. Katsenelson, Galina V. Shurin, Svetlana N. Bykovskaia, Jeffrey Shogan, Michael R. Shurin Department of Pathology (NSK, GVS, MRS), University of Pittsburgh Medical Center, University of Pittsburgh Cancer Institute, Pittsburgh, Pennsylvania; and Bone Marrow Transplantation Program (SNB, JS), Allegheny General Hospital, Pittsburgh, Pennsylvania KEY WORDS: Apoptosis, Bronchial carcinoid tumor, The induction of apoptosis in dendritic cells (DC) is Dendritic cell, Immunohistochemistry, Small-cell a key mechanism by which tumors escape immune lung carcinoma. recognition and elimination. In fact, a number of Mod Pathol 2001;14(1):40–45 studies have showed the correlation between the number of DC within the tumor and the clinical Immune responses to tumor antigens are the result prognosis, suggesting that increased infiltration of of reciprocal communication between antigen- tumor tissue by DC was associated with better pa- presenting cells, T cells, and B cells. The most po- tient survival and low incidence of metastatic dis- tent professional antigen-presenting cells are den- ease. We compared the number of DC and their dritic cells (DC), which are considered to be solely distribution pattern in human small-cell lung car- responsible for initiating primary immune re- cinoma and bronchial carcinoid tumor (CT) tissues. sponses, including antitumor immunity (1, 2). DC Immunohistochemical analysis revealed the pres- recognize and take up antigen, process it, and ence of cells expressing DC markers CD1a and CD83 present antigenic peptides to T cells in a major in small-cell lung carcinoma tissues and the com- histocompatibility complex (MHC)–restricted man- plete absence of these cells in CT samples. Next, we ner. Thus, interaction between DC and tumor cells examined whether human lung tumor cells produce with the tumor tissue, where DC recognize and take soluble factors that inhibit differentiation of hema- up tumor-associated antigens, should play a key topoietic precursors into mature DC. The addition role in the induction of specific antitumor immune of small-cell lung carcinoma–conditioned medium responses. In fact, it has been shown that the pres- -to CD34؉ precursor cell cultures significantly inhib- ence of greater numbers of DC within various tu ited colony-forming units of DC formation when mors is often associated with a better prognosis and compared with nontreated control DC cultures. longer patient survival (3, 4). Furthermore, DC generation and differentiation Varieties of tumor-produced factors exhibit im- was completely abrogated in CD34؉ cell cultures munomodulatory properties and play an important treated with CT-conditioned medium, suggesting role in the regulation of tumor growth and escape that CT-derived factors blocked CD34؉ cell differ- from immune recognition and elimination. Some of entiation into DC or induced their apoptosis. Fi- these molecules inhibit natural killer– and cytotoxic T lymphocyte–mediated cytotoxicity and macro- nally, flow cytometry analysis of cultured DC con- phage functions, induce apoptosis of T cells and firmed these results. Thus, analysis of our data natural killer cells, and cause apoptotic death of DC suggests that human lung tumors produce factors (5, 6). After the identification of DC as Langerhans that inhibit DC generation or maturation and may cells in the epidermis, interstitial cells within non- also induce apoptotic death of DC precursors in lymphoid tissues and lymph nodes, and veiled cells vitro. in lymph and blood, it was subsequently recog- nized by clinical pathologists that there was a rela- tionship between the presence and number of DC Copyright © 2001 by The United States and Canadian Academy of within the tumor tissue and prognosis. Pathology, Inc. VOL. 14, NO. 1, P. 40, 2001 Printed in the U.S.A. The identification of DC in the lung and airway Date of acceptance: July 25, 2000. Funding: This study was supported by RO1 CA80126 and RO1 CA84270 (to tissues of human is based on pleomorphic pathol- MRS). ogy in conjunction with a high level of MHC Class II Address reprint requests to: Michael R. Shurin, M.D., Ph.D., Immuno- pathology-CLSI, 5725 CHP-MT, 200 Lothrop Street, Pittsburgh, PA 15213; expression and lack of expression of characteristic fax: 412-692-4997. B-cell, T-cell, and macrophage markers. DC express 40 variable levels of S-100 protein, CD1a, CD11c, based fixation of the tissue. Rehydration and washing CD83, CD68, RFD1, and a number of adhesion mol- were done with standard phosphate buffered saline ecules (4, 7). The epithelial surfaces of the conduct- (PBS; Gibco BRL, Grand Island, NY) and a 3% PBS ing airways and the deep lung are continuously solution of goat serum (Gibco) prepared ex tempore. exposed to various antigens and pathogens from Antibody binding was localized by biotinylated sec- the inhaled air, suggesting that DC in the lung ondary antibodies followed by an avidin-conjugated should have specific features caused by the unique horseradish peroxidase and diaminobenzidine (3,3Ј- lung anatomy and physiology (8). DC are present in diaminobenzidine) substrate with an auidin biotinyl- the normal airway epithelium, lung parenchyma, ated enzyme complex detection kit (Vector Laborato- and epithelial surface of the lower respiratory tract ries, Burlingame, CA). Isotype-matched antibodies (9–11). The distribution of DC within the epithe- were used as a negative control. DC infiltration per lium of conducting airways is similar to the Lang- high power field was graded 0 through 4 as follows: 0, erhans cell network in the epidermis (12). The pres- absence of DC; 1, a small amount of DC (fewer than ence of DC in human respiratory tract tissues was 10 cells); 2, a medium infiltration (10 to 50 cells); 3, a also noted in the context of several diseases such as medium to high amount of DC (more than 50 cells); histiocytosis X (13), carcinoma (14), and various and 4, a high DC number (widely distributed cells granulomatous disorders (15, 16). Pulmonary neu- throughout the whole slide). roendocrine tumors, in particular carcinoid tumor (CT) and small-cell lung carcinoma (SCLC), are of significant interest because of their ability to se- Human DC Cultures crete a number of hormones and hormone-like Human DC were generated from CD34ϩ precursors substances, peptides, and mediators of inflamma- isolated from umbilical cord blood and cultured with tion, leading to a specific clinical picture, as well as recombinant human granulocyte-macrophage colony- their ability to cause distant metastases of unpre- stimulating factor (1000 U/mL) and recombinant dictable location several months after successful interleukin-4 (1000 U/mL) in AIM-V medium (BioWhi- surgery. tacker, Walkersville, MD) supplemented with 2.5% heat- A growing body of evidence indicates that tumor- inactivated human type AB serum (Sigma, St. Louis, induced immunologic unresponsiveness may be MO). Briefly, fresh cord blood samples were diluted 1:4 primarily the result of an insufficient number or with PBS, and 25 mL of cell suspension was loaded onto function of immune cells, including DC (5, 17, 18). 15 mL HistoPrep solution (Sigma). Mononuclear cells However, to our knowledge, the interaction be- were isolated from the interface after centrifugation (400 tween DC and tumor cells in the lung has not yet g, 30 min, 20°C) and washed twice in PBS. CD34ϩ cells been investigated. The aim of this study was to were separated with magnetic cell sorting (MACS, Milte- compare the distribution of DC in human SCLC and nyi Biotech, Sunnyvale, CA). According to the manufac- CT tissues and determine whether SCLC- and CT- turer’s protocol, mononuclear cells were labeled with derived soluble factors regulate DC generation and QBEND10, and anti-CD34 antibodies attached to the maturation. We have demonstrated that CT cells superparamagnetic microparticles. The cell suspension were significantly stronger in elimination of DC was then passed through a MiniMacs separation col- from the tumor microenvironment and inhibition umn in a highly magnetized field. The percentage of of DC generation (dendropoiesis) when compared CD34ϩ cells was determined in each experiment by a with SCLC cells. These results might serve as a basis FACScan analysis, and only cell suspensions enriched for the further development of immunohistochem- with up to 90% and more of CD34ϩ cells were selected ical criteria for the differential diagnosis of lung for further experiments. The cell viability was more than tumors and provide new insights in our under- 95% in all experiments. standing of the immunobiology of lung cancer. Tumor Cell Lines MATERIALS AND METHODS Bronchial CT and SCLC lines CRL-5815 and CRL- 5837, respectively, were obtained from American Immunohistochemistry Type Culture Collection (Rockville, MD) and cultured Samples of bronchial CT and SCLC tissues frozen in a complete RPMI 1640 medium supplemented optimal cutting temperature compound were ob- with 2 mML-glutamine, 100 U/mL penicillin, 100 tained from the University of Pittsburgh Cancer Insti- ␮g/mL streptomycin, 1 mM sodium pyruvate, 0.1 mM tute Tumor Tissue Bank. For the staining of human nonessential amino acids, and 5% heat-inactivated DC, anti-CD1a, anti-CD83, anti–S-100, or anti–hu- fetal bovine serum (all from Gibco). Cell-free super- man leukocyte antigen–DR antibodies (Immunotech, natants were harvested from nonconfluent cultures, Westbrook, ME) were used. Staining