Cytokine Expression by Head and Neck Squamous Cell Carcinomas* Eric A. Mann, MD, PhD, Jeffrey D. Spiro, MD, Lei L. Chen, MD, PhO, Donald L. Kreutzer, PhD, Farmington,Connecticut
Cytokines are known to play an important role in host defense by regulating the function, growth, and differentiation of the cells of the immune system. We hypothesize that, in the tumor microenvironment, tumor cells and resident tissue eells (e.g., fibroblasts) also produce eytokines that may regulate the local immune response to tumors. Initially, homogenates of eight head and neck squamous cell carcinomas (HNSCC) were assayed for the presence of interleukin- 1 (IL- 1 ), interleukin-4 ( I L - 4 ) , interlenkin-6 (IL-6), and granuloeyte-maerophage colony-stimulating factor (GMCSF) to establish the presence of these eytokines in the tumors in vivo. We detected IL-1 in all tumor homogenates and IL-4, IL-6, and GM-CSF in some homogenates. To assess the ability of HNSCC to produce these eytokines, supernatants of short-term primary cultures of HNSCC were assayed for the same cytokines. No IL-1 was detected, although baseline levels of IL-4, IL-6, and GM-CSF were present. However, the stimulation of primary tumor cultures with exogenous IL-1 induced or significantly enhanced production of IL-4 (p < 0 . 0 1 ) , IL-6 (p < 0 . 0 0 1 ) , and GM-CSF (p < 0 . 0 2 ) . These results support our hypothesis that HNSCC secrete cytokines that may influence the respouse of local immune cells. Our data also suggest that IL-1 may have a central role in regulating the local immune response through the enhancement or induction of cytokine production by tumor and/or resident tissue cells.
From the Divisionof Otolaryngology,Departmentof Surgery(EAM, JDS), Divisionof Hematologyand Ontology,Departmentof Medicine (LLC), Departmentof Pathology(DLK),Universityof Connecticut HealthCenter,Farmington,Connecticut.Fundedby PHS CancerResearch Fund,AmericanCancer SocietyInstitutionalResearchGrant INI52G-100, AmericanCancerSocietyJSRA-288, and the Veterans AdministrationResearchFunds. *Dr. Mannis the recipientofthe 1992ResearchResidentAwardof the Societyof Headand NeckSurgeons. Requestsfor reprintsshouldbe addressedto JeffreyD. Spiro,MD, Divisionof Otolaryngology,Universityof ConnecticutHealthCenter, Farmington,Connecticut06032. Presentedat theThirdInternationalConferenceon Headand Neck Cancer,San Francisco,California,July 26-30, 1992.
he host immune system normally functions to destroy neoplastic cells that continually develop as a result of T somatic mutations. However, patients with head and neck squamous cell carcinoma (HNSCC) have depressed cell-mediated immune function [I-4], which has recently been shown to be most pronounced in the local/regional environment of the primary tumor [5,6]. Two recent studies have actually identified soluble factors produced by HNSCC [7] and regional lymph nodes [6] that profoundly inhibit the tumodcidal ability of lymphokineactivated killer (LAK) cells. Clearly, these studies suggest a local modulation of the host immune response to tumor by secreted immunoregulatory factors such as cytokines. Over the past several years, we have achieved a greater understanding of the cellular regulation of the immune system by small giycoprotein "hormones" known as cytokines. Cytokines can be broadly classified into two groups: (1) pro-inflammatory cytokines (interleukin-1 [IL-1], interleukin-6 [IL-6], interleukin-8 [IL-8], interferons, and tumor necrosis factor) and (2) growth factor cytokines (interleukin-3, interleukin-4 [IL-4], interleukin-5, interleukin-7, interleukin-10, and the colony-stimulating factors), although considerable overlap in their function may exist. Cytokines were originally described as products of macrophages, monocytes, and lymphocytes that served to regulate the function, growth, and differentiation of other leukocytes [8,9]. However, more recent studies have shown that nonimmunologie cells are capable of producing eytokines. Specifically, epithelial and endothelial cells have been shown to produce IL-1 and IL-8 [10-13], and fibroblasts are known to produce interferons, colony-stimulating factors, and IL-6 with appropriate stimuli [14,15]. Other recent studies have shown that nonimmunologic tissue cells can produce cytokines that appear to modulate the function of immune cells in a local immune response [14,16]. For example, fibroblasts from joint synovial cells in patients with rheumatoid arthritis were shown to produce IL-6 and granulocyte-macrophage colony-stimulating factor (GM-CSF), which appear to play a role in the perpetuation of synovitis in this disease [16]. Based on the t'mdings that resident tissue cells can produce cytokines that modulate local inflammation, we hypothesized that, within the tumor microenvironment, tumor cells and resident tissue cells, e.g., fibroblasts, also produce cytokines that regulate immune cell function locally at the site of the tumor. To test this hypothesis, we performed two major sets of experiments. To establish that cytokines are, in fact, present in the tumor microenvironment in oioo, we assayed homogenates of HNSCC for the presence of IL-1, IL-4, IL-6, and GM-CSF. Sec-
THE AMERICAN JOURNALOF SURGERY VOLUME164 DECEMBER1992 567
MANN ET AL
ondly, to assess the ability of HNSCC to produce these cytokines, we assayed superuatants from short-term primary cultures of HNSCC and squamous cell carcinoma cell lines for cytokines under both baseline unstimulated conditions and in response to several exogenous stimuli: IL- 1, epidermal growth factor (EGF), and bacterial lipopolysaccharide (LPS). MATERIALS AND METHODS Tumor specimens: Surgical specimens of primary tumors and lymph node metastases of HNSCC were collected fresh under sterile conditions from the operating room at the University of Connecticut Health Center, New Britain General Hospital, St. Francis Hospital, and the West Haven Veterans Administration Medical Center. Specimens were kept in phosphate-buffered saline (PBS) on ice until processing. Some tumors were immediately placed in n'ficrovials and frozen in liquid nitrogen for tumor homogenate experiments. The remaining tumors were prepared for primary tumor culture as described below. The study was undertaken in a manner approved by the University of Connecticut Institutional Review Board for Human Subjects. Tumor cell lines: A-431 and NCI-H292 squamous cell carcinoma cell lines were obtained from the American Type Tissue Collection (Rockville, MD) and were maintained in a humidified atmosphere containing 5% carbon dioxide at 37~ Tumor homogenates: The tumor tissue samples stored in liquid nitrogen were thawed, and the specimens were quickly weighed, placed in 2 mL of PBS, and homogenized for 10 seconds in a tissue homogenizer. The homogenate was then centrifuged twice at 4~ at 10,000g, and aliquots of the supernatant were prepared for analysis by immunoassay. Short-term primary culture: Squamous cell carcinoma tumor specimens were dissected free from debris or necrotic areas, cut into small pieces in cold PBS with amphotericin B (2.5 ~g/mL) (Gibco, Grand Island, NY), washed twice with cold PBS, and treated with a mixture of enzymes containing coHagenase (1 mg/mL) (Sigma Chemical Co., St. Louis, MO) and deoxyribonuclease (160 mg/mL) (Sigma Chemical Co.) at 37~ with gentle shaking for 40 minutes. The dissociated cells and small clumps were then washed twice with PBS and plated in 24-well plates with Dulbecco's minimum essential medium (Gibeo), supplemented with 10% fetal calf serum (Whittaker Bioproducts, Walkersville, MD), hydrocortisone (5 /zg/mL) (Gibeo), glutamine (2 raM) (Gibco), antibiotics (penicillin 100 IU/mL, streptomycin 100/zg/mL, gentamicin 5 #g/mL, and amphotericin B 2.5 ~g/mL) (Gibeo) in an atmosphere of 5% carbon dioxide. The medium was changed after 48 hours, and dead or unattached cells were removed. The viable cells were incubated for short-term culture analysis. Tumor cells, when growing in aggregates, become slightly more resistant to trypsin treatment. Thus, fibroblast overgrowth was prevented by selective trypsinization (0.05% trypsin-2 mM ethylene diaminetetraacetic acid [EDTA]) (Gibeo). 568
General experimental design: Short-term primary cultures of HNSCC and squamous cell carcinoma cell lines were grown to a uniform degree of confluence (approximately 50% to 60%) in six-well tissue culture dishes. The medium was then changed to one of the following types of fresh media: (1) serum-free medium (Clonetics, Corp., San Diego, CA) (2), serum-free medium supplemented with 20 ng/mL EGF (Collaborative Research, Inc., Bedford, MA), (3) serum-free medium supplemented with 10/~g/mL bacterial LPS (Sigma), or (4) serumfree medium supplemented with 103 U/mL IL-1 (a) (Genzyme Corp., Cambridge, MA). Twenty-four hours after the media change, snpernatants were harvested, centrifuged at 10,000g for 10 minutes, and frozen at -70~ for later analysis. Cells were then trypsinized from the six-well plates and counted, and the percentage of fibroblasts was estimated as described below. Cell s t a i n i n ~ and flow cytometry: To characterize the percentage of fibroblasts and rule out fibroblast overgrowth during short-term tumor culture, a routine monoclonal antibody BR2 [17], which is highly sensitive and specific for human fibroblasts, was used to stain primary tumor cultures. Briefly, after trypsinization of primary tumor cultures, cells were washed three times in PBS at 4~ and resuspended in a 1:1,000 dilution of BR2 ascites fluid at a concentration of 106 cells/mL. The cells were incubated for 30 minutes at 4~ on a rocker platform. After the cells were washed three times in PBS, they were counterstained with fluorescein-conjugated polyclonal goat anti-mouse antibody (Coulter Immunology, Hialeah, FL) at a dilution of 1:40. After a 20-minute incubation, cells were washed three times, resuspended in PBS at 4~ and immediately analyzed on an EPICS flow cytometer (Coulter Immunology) or fixed with 4% paraformaldehyde for analysis within 24 hours. Standard fluorescent antibody and isotype-matched controls (mouse monoclonal anti-KLH [Coulter Immunology]) were performed on each tumor culture sample. Cytokine assays: Immunoassays of tissue culture supernatants and tumor homogenates were performed using the following standardized kits: IL-1 (o0 and GMCSF radioimmunoassay (RIA) kits (Advanced Magnetics, Inc., Cambridge, MA), IL-4 enzyme-linked immunosorbent assay (ELISA) (Genzyme Corp.), and IL-6 (Amgen Diagnostics, Minneapolis, MN). Standard curves were generated from known concentrations of cytokine provided by the manufacturer. Standard curves were then used to determine the quantity of cytokine in the sample based on the level of radioactivity or spectrophotometric absorbance of each sample using regression analysis. All samples were performed twice, and the results were expressed as the number of picograms of each mediator produced per 106 cells in the primary culture and ceil line experiments and in picograms per gram of tumor tissue for the homogenate experiments. Statistical analysis: In these experimental studies, an analysis of variance (ANOVA) was used to determine significant differences between treatment groups. All analyses were performed with the Minitab software package (Addison-Wesley Publishing Co., Inc., Reading,
THE AMERICAN JOURNAL OF SURGERY VOLUME 164 DECEMBER 1992
HEAD AND NECK SQUAMOUSCELL CARCINOMAS
MA). A probability of p
Cytokines are known to play an important role in host defense by regulating the function, growth, and differentiation of the cells of the immune system. We hypothesize that, in the tumor microenvironment, tumor cells and resident tissue eells (e.g., fibroblasts) also produce eytokines that may regulate the local immune response to tumors. Initially, homogenates of eight head and neck squamous cell carcinomas (HNSCC) were assayed for the presence of interleukin- 1 (IL- 1 ), interleukin-4 ( I L - 4 ) , interlenkin-6 (IL-6), and granuloeyte-maerophage colony-stimulating factor (GMCSF) to establish the presence of these eytokines in the tumors in vivo. We detected IL-1 in all tumor homogenates and IL-4, IL-6, and GM-CSF in some homogenates. To assess the ability of HNSCC to produce these eytokines, supernatants of short-term primary cultures of HNSCC were assayed for the same cytokines. No IL-1 was detected, although baseline levels of IL-4, IL-6, and GM-CSF were present. However, the stimulation of primary tumor cultures with exogenous IL-1 induced or significantly enhanced production of IL-4 (p < 0 . 0 1 ) , IL-6 (p < 0 . 0 0 1 ) , and GM-CSF (p < 0 . 0 2 ) . These results support our hypothesis that HNSCC secrete cytokines that may influence the respouse of local immune cells. Our data also suggest that IL-1 may have a central role in regulating the local immune response through the enhancement or induction of cytokine production by tumor and/or resident tissue cells.
From the Divisionof Otolaryngology,Departmentof Surgery(EAM, JDS), Divisionof Hematologyand Ontology,Departmentof Medicine (LLC), Departmentof Pathology(DLK),Universityof Connecticut HealthCenter,Farmington,Connecticut.Fundedby PHS CancerResearch Fund,AmericanCancer SocietyInstitutionalResearchGrant INI52G-100, AmericanCancerSocietyJSRA-288, and the Veterans AdministrationResearchFunds. *Dr. Mannis the recipientofthe 1992ResearchResidentAwardof the Societyof Headand NeckSurgeons. Requestsfor reprintsshouldbe addressedto JeffreyD. Spiro,MD, Divisionof Otolaryngology,Universityof ConnecticutHealthCenter, Farmington,Connecticut06032. Presentedat theThirdInternationalConferenceon Headand Neck Cancer,San Francisco,California,July 26-30, 1992.
he host immune system normally functions to destroy neoplastic cells that continually develop as a result of T somatic mutations. However, patients with head and neck squamous cell carcinoma (HNSCC) have depressed cell-mediated immune function [I-4], which has recently been shown to be most pronounced in the local/regional environment of the primary tumor [5,6]. Two recent studies have actually identified soluble factors produced by HNSCC [7] and regional lymph nodes [6] that profoundly inhibit the tumodcidal ability of lymphokineactivated killer (LAK) cells. Clearly, these studies suggest a local modulation of the host immune response to tumor by secreted immunoregulatory factors such as cytokines. Over the past several years, we have achieved a greater understanding of the cellular regulation of the immune system by small giycoprotein "hormones" known as cytokines. Cytokines can be broadly classified into two groups: (1) pro-inflammatory cytokines (interleukin-1 [IL-1], interleukin-6 [IL-6], interleukin-8 [IL-8], interferons, and tumor necrosis factor) and (2) growth factor cytokines (interleukin-3, interleukin-4 [IL-4], interleukin-5, interleukin-7, interleukin-10, and the colony-stimulating factors), although considerable overlap in their function may exist. Cytokines were originally described as products of macrophages, monocytes, and lymphocytes that served to regulate the function, growth, and differentiation of other leukocytes [8,9]. However, more recent studies have shown that nonimmunologie cells are capable of producing eytokines. Specifically, epithelial and endothelial cells have been shown to produce IL-1 and IL-8 [10-13], and fibroblasts are known to produce interferons, colony-stimulating factors, and IL-6 with appropriate stimuli [14,15]. Other recent studies have shown that nonimmunologic tissue cells can produce cytokines that appear to modulate the function of immune cells in a local immune response [14,16]. For example, fibroblasts from joint synovial cells in patients with rheumatoid arthritis were shown to produce IL-6 and granulocyte-macrophage colony-stimulating factor (GM-CSF), which appear to play a role in the perpetuation of synovitis in this disease [16]. Based on the t'mdings that resident tissue cells can produce cytokines that modulate local inflammation, we hypothesized that, within the tumor microenvironment, tumor cells and resident tissue cells, e.g., fibroblasts, also produce cytokines that regulate immune cell function locally at the site of the tumor. To test this hypothesis, we performed two major sets of experiments. To establish that cytokines are, in fact, present in the tumor microenvironment in oioo, we assayed homogenates of HNSCC for the presence of IL-1, IL-4, IL-6, and GM-CSF. Sec-
THE AMERICAN JOURNALOF SURGERY VOLUME164 DECEMBER1992 567
MANN ET AL
ondly, to assess the ability of HNSCC to produce these cytokines, we assayed superuatants from short-term primary cultures of HNSCC and squamous cell carcinoma cell lines for cytokines under both baseline unstimulated conditions and in response to several exogenous stimuli: IL- 1, epidermal growth factor (EGF), and bacterial lipopolysaccharide (LPS). MATERIALS AND METHODS Tumor specimens: Surgical specimens of primary tumors and lymph node metastases of HNSCC were collected fresh under sterile conditions from the operating room at the University of Connecticut Health Center, New Britain General Hospital, St. Francis Hospital, and the West Haven Veterans Administration Medical Center. Specimens were kept in phosphate-buffered saline (PBS) on ice until processing. Some tumors were immediately placed in n'ficrovials and frozen in liquid nitrogen for tumor homogenate experiments. The remaining tumors were prepared for primary tumor culture as described below. The study was undertaken in a manner approved by the University of Connecticut Institutional Review Board for Human Subjects. Tumor cell lines: A-431 and NCI-H292 squamous cell carcinoma cell lines were obtained from the American Type Tissue Collection (Rockville, MD) and were maintained in a humidified atmosphere containing 5% carbon dioxide at 37~ Tumor homogenates: The tumor tissue samples stored in liquid nitrogen were thawed, and the specimens were quickly weighed, placed in 2 mL of PBS, and homogenized for 10 seconds in a tissue homogenizer. The homogenate was then centrifuged twice at 4~ at 10,000g, and aliquots of the supernatant were prepared for analysis by immunoassay. Short-term primary culture: Squamous cell carcinoma tumor specimens were dissected free from debris or necrotic areas, cut into small pieces in cold PBS with amphotericin B (2.5 ~g/mL) (Gibco, Grand Island, NY), washed twice with cold PBS, and treated with a mixture of enzymes containing coHagenase (1 mg/mL) (Sigma Chemical Co., St. Louis, MO) and deoxyribonuclease (160 mg/mL) (Sigma Chemical Co.) at 37~ with gentle shaking for 40 minutes. The dissociated cells and small clumps were then washed twice with PBS and plated in 24-well plates with Dulbecco's minimum essential medium (Gibeo), supplemented with 10% fetal calf serum (Whittaker Bioproducts, Walkersville, MD), hydrocortisone (5 /zg/mL) (Gibeo), glutamine (2 raM) (Gibco), antibiotics (penicillin 100 IU/mL, streptomycin 100/zg/mL, gentamicin 5 #g/mL, and amphotericin B 2.5 ~g/mL) (Gibeo) in an atmosphere of 5% carbon dioxide. The medium was changed after 48 hours, and dead or unattached cells were removed. The viable cells were incubated for short-term culture analysis. Tumor cells, when growing in aggregates, become slightly more resistant to trypsin treatment. Thus, fibroblast overgrowth was prevented by selective trypsinization (0.05% trypsin-2 mM ethylene diaminetetraacetic acid [EDTA]) (Gibeo). 568
General experimental design: Short-term primary cultures of HNSCC and squamous cell carcinoma cell lines were grown to a uniform degree of confluence (approximately 50% to 60%) in six-well tissue culture dishes. The medium was then changed to one of the following types of fresh media: (1) serum-free medium (Clonetics, Corp., San Diego, CA) (2), serum-free medium supplemented with 20 ng/mL EGF (Collaborative Research, Inc., Bedford, MA), (3) serum-free medium supplemented with 10/~g/mL bacterial LPS (Sigma), or (4) serumfree medium supplemented with 103 U/mL IL-1 (a) (Genzyme Corp., Cambridge, MA). Twenty-four hours after the media change, snpernatants were harvested, centrifuged at 10,000g for 10 minutes, and frozen at -70~ for later analysis. Cells were then trypsinized from the six-well plates and counted, and the percentage of fibroblasts was estimated as described below. Cell s t a i n i n ~ and flow cytometry: To characterize the percentage of fibroblasts and rule out fibroblast overgrowth during short-term tumor culture, a routine monoclonal antibody BR2 [17], which is highly sensitive and specific for human fibroblasts, was used to stain primary tumor cultures. Briefly, after trypsinization of primary tumor cultures, cells were washed three times in PBS at 4~ and resuspended in a 1:1,000 dilution of BR2 ascites fluid at a concentration of 106 cells/mL. The cells were incubated for 30 minutes at 4~ on a rocker platform. After the cells were washed three times in PBS, they were counterstained with fluorescein-conjugated polyclonal goat anti-mouse antibody (Coulter Immunology, Hialeah, FL) at a dilution of 1:40. After a 20-minute incubation, cells were washed three times, resuspended in PBS at 4~ and immediately analyzed on an EPICS flow cytometer (Coulter Immunology) or fixed with 4% paraformaldehyde for analysis within 24 hours. Standard fluorescent antibody and isotype-matched controls (mouse monoclonal anti-KLH [Coulter Immunology]) were performed on each tumor culture sample. Cytokine assays: Immunoassays of tissue culture supernatants and tumor homogenates were performed using the following standardized kits: IL-1 (o0 and GMCSF radioimmunoassay (RIA) kits (Advanced Magnetics, Inc., Cambridge, MA), IL-4 enzyme-linked immunosorbent assay (ELISA) (Genzyme Corp.), and IL-6 (Amgen Diagnostics, Minneapolis, MN). Standard curves were generated from known concentrations of cytokine provided by the manufacturer. Standard curves were then used to determine the quantity of cytokine in the sample based on the level of radioactivity or spectrophotometric absorbance of each sample using regression analysis. All samples were performed twice, and the results were expressed as the number of picograms of each mediator produced per 106 cells in the primary culture and ceil line experiments and in picograms per gram of tumor tissue for the homogenate experiments. Statistical analysis: In these experimental studies, an analysis of variance (ANOVA) was used to determine significant differences between treatment groups. All analyses were performed with the Minitab software package (Addison-Wesley Publishing Co., Inc., Reading,
THE AMERICAN JOURNAL OF SURGERY VOLUME 164 DECEMBER 1992
HEAD AND NECK SQUAMOUSCELL CARCINOMAS
MA). A probability of p
