Portal de Periódicos da CAPES

Positioning Ganglioside D3 as an Immunotherapeutic Target in Lymphangioleiomyomatosis

2013; Elsevier BV; Volume: 183; Issue: 1 Linguagem: Inglês

10.1016/j.ajpath.2013.04.002

1525-2191

Emily Gilbert, Jonathan M. Eby, Adam M. Hammer, Jared Klarquist, David G. Christensen, Allison Barfuss, Raymond E. Boissy, Maria M. Picken, Robert B. Love, Daniel F. Dilling, I. Caroline Le Poole,

Histiocytic Disorders and Treatments

Tumors that develop in lymphangioleiomyomatosis (LAM) as a consequence of biallelic loss of TSC1 or TSC2 gene function express melanoma differentiation antigens. However, the percentage of LAM cells expressing these melanosomal antigens is limited. Here, we report the overexpression of ganglioside D3 (GD3) in LAM. GD3 is a tumor-associated antigen otherwise found in melanoma and neuroendocrine tumors; normal expression is largely restricted to neuronal cells in the brain. We also observed markedly reduced serum antibody titers to GD3, which may allow for a population of GD3-expressing LAM cells to expand within patients. This is supported by the demonstrated sensitivity of cultured LAM cells to complement mediated cytotoxicity via GD3 antibodies. GD3 can serve as a natural killer T (NKT) cell antigen when presented on CD1d molecules expressed on professional antigen-presenting cells. Although CD1d-expressing monocyte derivatives were present in situ, enhanced NKT-cell recruitment to LAM lung was not observed. Cultured LAM cells retained surface expression of GD3 over several passages and also expressed CD1d, implying that infiltrating NKT cells can be directly cytotoxic toward LAM lung lesions. Immunization with antibodies to GD3 may thus be therapeutic in LAM, and enhancement of existing NKT-cell infiltration may be effective to further improve antitumor responses. Overall, we hereby establish GD3 as a suitable target for immunotherapy of LAM. Tumors that develop in lymphangioleiomyomatosis (LAM) as a consequence of biallelic loss of TSC1 or TSC2 gene function express melanoma differentiation antigens. However, the percentage of LAM cells expressing these melanosomal antigens is limited. Here, we report the overexpression of ganglioside D3 (GD3) in LAM. GD3 is a tumor-associated antigen otherwise found in melanoma and neuroendocrine tumors; normal expression is largely restricted to neuronal cells in the brain. We also observed markedly reduced serum antibody titers to GD3, which may allow for a population of GD3-expressing LAM cells to expand within patients. This is supported by the demonstrated sensitivity of cultured LAM cells to complement mediated cytotoxicity via GD3 antibodies. GD3 can serve as a natural killer T (NKT) cell antigen when presented on CD1d molecules expressed on professional antigen-presenting cells. Although CD1d-expressing monocyte derivatives were present in situ, enhanced NKT-cell recruitment to LAM lung was not observed. Cultured LAM cells retained surface expression of GD3 over several passages and also expressed CD1d, implying that infiltrating NKT cells can be directly cytotoxic toward LAM lung lesions. Immunization with antibodies to GD3 may thus be therapeutic in LAM, and enhancement of existing NKT-cell infiltration may be effective to further improve antitumor responses. Overall, we hereby establish GD3 as a suitable target for immunotherapy of LAM. Lymphangioleiomyomatosis (LAM) is a cystic lung disease that almost exclusively strikes women, most commonly before menopause.1Johnson S.R. Whale C.I. Hubbard R.B. Lewis S.A. Tattersfield A.E. Survival and disease progression in UK patients with lymphangioleiomyomatosis.Thorax. 2004; 59: 800-803Crossref PubMed Scopus (131) Google Scholar Although the lung lesions that develop in patients with LAM are often slow growing, eventually patients need supplemental oxygen and many require lung transplantation.2Johnson S. Rare diseases. 1. Lymphangioleiomyomatosis: clinical features, management and basic mechanisms.Thorax. 1999; 54: 254-264Crossref PubMed Scopus (179) Google Scholar LAM is associated with all common aspects of cancer, including unbridled cell growth and metastasis.3Dilling D.F. Gilbert E.R. Picken M.M. Eby J.M. Love R.B. Le Poole I.C. A current viewpoint of lymphangioleiomyomatosis supporting immunotherapeutic treatment options.Am J Respir Cell Mol Biol. 2012; 46: 1-5Crossref PubMed Scopus (15) Google ScholarA major breakthrough in LAM research has come with the discovery that inactivation of tuberous sclerosis complex 2 (TSC2), or alternatively TSC1, is responsible for disease development.4Strizheva G.D. Carsillo T. Kruger W.D. Sullivan E.J. Ryu J.H. Henske E.P. The spectrum of mutations in TSC1 and TSC2 in women with tuberous sclerosis and lymphangiomyomatosis.Am J Respir Crit Care Med. 2001; 163: 253-258Crossref PubMed Scopus (113) Google Scholar LAM will develop in approximately 30% of patients with tuberous sclerosis predisposed to the disease by an inherited mutation that affects a single allele of a TSC gene. In others, the disease is a manifestation of sporadic mutations that affect both alleles, limiting initial tumor development to the lungs. In either case kidney metastases are a frequent event.5Chorianopoulos D. Stratakos G. Lymphangioleiomyomatosis and tuberous sclerosis complex.Lung. 2008; 186: 197-207Crossref PubMed Scopus (29) Google ScholarThe underlying TSC gene products regulate mammalian target of rapamycin complex (mTORC)-1 activity, thus rapamycin is currently being used as a treatment to inhibit the mTORC-1 in LAM.6Goncharova E.A. Goncharov D.A. Eszterhas A. Hunter D.S. Glassberg M.K. Yeung R.S. Walker C.L. Noonan D. Kwiatkowski D.J. Chou M.M. Panettieri Jr., R.A. Krymskaya V.P. Tuberin regulates p70 S6 kinase activation and ribosomal protein S6 phosphorylation. A role for the TSC2 tumor suppressor gene in pulmonary lymphangioleiomyomatosis (LAM).J Biol Chem. 2002; 277: 30958-30967Crossref PubMed Scopus (365) Google Scholar In recent studies, patients treated with rapamycin showed stabilization of lung function7McCormack F.X. Inoue Y. Moss J. Singer L.G. Strange C. Nakata K. Barker A.F. Chapman J.T. Brantly M.L. Stocks J.M. Brown K.K. Lynch III, J.P. Goldberg H.J. Young L.R. Kinder B.W. Downey G.P. Sullivan E.J. Colby T.V. McKay R.T. Cohen M.M. Korbee L. Taveira-DaSilva A.M. Lee H.S. Krischer J.P. Trapnell B.C. National Institutes of Health Rare Lung Diseases ConsortiumMILES Trial GroupEfficacy and safety of sirolimus in lymphangioleiomyomatosis.N Engl J Med. 2011; 364: 1595-1606Crossref PubMed Scopus (764) Google Scholar and a reduction in size of angiomyolipomas.8Bissler J.J. McCormack F.X. Young L.R. Elwing J.M. Chuck G. Leonard J.M. Schmithorst V.J. Laor T. Brody A.S. Bean J. Salisbury S. Franz D.N. Sirolimus for angiomyolipoma in tuberous sclerosis complex or lymphangioleiomyomatosis.N Engl J Med. 2008; 358: 140-151Crossref PubMed Scopus (978) Google Scholar However, rapamycin analogs are not cytotoxic to LAM cells and cannot reverse existing disease symptoms.9Diaz-Padilla I. Duran I. Clarke B.A. Oza A.M. Biologic rationale and clinical activity of mTOR inhibitors in gynecological cancer.Cancer Treat Rev. 2012; 38: 767-775Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar Additional treatments that address viability of LAM cells without compromising patient health are thus of the essence.The cell type giving rise to LAM tumors remains an enigma. Possible sources include smooth muscle cells, uterine cells, ovarian cells, endothelial cells, or even stem cells.3Dilling D.F. Gilbert E.R. Picken M.M. Eby J.M. Love R.B. Le Poole I.C. A current viewpoint of lymphangioleiomyomatosis supporting immunotherapeutic treatment options.Am J Respir Cell Mol Biol. 2012; 46: 1-5Crossref PubMed Scopus (15) Google Scholar Alternative candidate cell types to transform into LAM are cells of neural crest origin, perhaps melanocytes, because the gp100 reactive HMB45 antibody is commonly used to provide a definitive diagnosis for LAM.10Bonetti F. Chiodera P.L. Pea M. Martignoni G. Bosi F. Zamboni G. Mariuzzi G.M. Transbronchial biopsy in lymphangiomyomatosis of the lung. HMB45 for diagnosis.Am J Surg Pathol. 1993; 17: 1092-1102Crossref PubMed Scopus (178) Google Scholar Gp100 is an antigen expressed by both melanocytes and melanoma cells. Besides this melanosomal antigen, reactivity to melanoma antigen recognized by T cells-1, tyrosinase-related protein-2, and tyrosinase-related protein-1 has been shown in LAM cells, and expression of microphthalmia-associated transcription factor (MITF), the transcription factor responsible for the melanocyte differentiation program, has been observed as well.11Klarquist J. Barfuss A. Kandala S. Reust M.J. Braun R.K. Hu J. Dilling D.F. McKee M.D. Boissy R.E. Love R.B. Nishimura M.I. Le Poole I.C. Melanoma-associated antigen expression in lymphangioleiomyomatosis renders tumor cells susceptible to cytotoxic T cells.Am J Pathol. 2009; 175: 2463-2472Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar In principle, this provides grounds for subjecting patients to immunotherapy that targets melanosomal antigens. However, as we demonstrated recently, expression of these antigens is limited to only minor subsets of LAM cells and is rapidly lost in cell cultures initiated from LAM lung.11Klarquist J. Barfuss A. Kandala S. Reust M.J. Braun R.K. Hu J. Dilling D.F. McKee M.D. Boissy R.E. Love R.B. Nishimura M.I. Le Poole I.C. Melanoma-associated antigen expression in lymphangioleiomyomatosis renders tumor cells susceptible to cytotoxic T cells.Am J Pathol. 2009; 175: 2463-2472Abstract Full Text Full Text PDF PubMed Scopus (27) Google ScholarLimited expression of melanosomal proteins by LAM tumors prompted a search for a more consistent and stably expressed tumor antigen that may be targetable by immunotherapy. Inspired by the melanocytic differentiation pathway and the molecules common to cells of the melanocytic lineage, options included MITF, c-KIT, tetraspanin, and ganglioside D3 (GD3). However, MITF is not considered immunogenic itself,12Kono M. Dunn I.S. Durda P.J. Butera D. Rose L.B. Haggerty T.J. Benson E.M. Kurnick J.T. Role of the mitogen-activated protein kinase signaling pathway in the regulation of human melanocytic antigen expression.Mol Cancer Res. 2006; 4: 779-792Crossref PubMed Scopus (125) Google Scholar and expression of the stem cell factor receptor c-KIT is generally lost in the process of malignant transformation.13Natali P.G. Nicotra M.R. Winkler A.B. Cavaliere R. Bigotti A. Ullrich A. Progression of human cutaneous melanoma is associated with loss of expression of c-kit proto-oncogene receptor.Int J Cancer. 1992; 52: 197-201Crossref PubMed Scopus (194) Google Scholar Both c-KIT and tetraspanin expression are not limited to melanocytes, so targeting these molecules would likely be associated with severe side effects.14Demetrick D.J. Herlyn D. Tretiak M. Creasey D. Clevers H. Donoso L.A. Vennegoor C.J. Dixon W.T. Jerry L.M. ME491 melanoma-associated glycoprotein family: antigenic identity of ME491. NKI/C-3, neuroglandular antigen (NGA), and CD63 proteins.J Natl Cancer Inst. 1992; 84: 422-429Crossref PubMed Scopus (33) Google Scholar GD3, however, is an interesting molecule associated with cellular apoptosis and likely contributes to melanogenesis as well.15Birklé S. Gao L. Zeng G. Yu R.K. Down-regulation of GD3 ganglioside and its O-acetylated derivative by stable transfection with antisense vector against GD3-synthase gene expression in hamster melanoma cells: effects on cellular growth, melanogenesis, and dendricity.J Neurochem. 2000; 74: 547-554Crossref PubMed Scopus (47) Google Scholar In melanoma cells, GD3 expression enhances their proliferation and invasive activity.16Ohkawa Y. Miyazaki S. Hamamura K. Kambe M. Miyata M. Tajima O. Ohmi Y. Yamauchi Y. Furukawa K. Furukawa K. Ganglioside GD3 enhances adhesion signals and augments malignant properties of melanoma cells by recruiting integrins to glycolipid-enriched microdomains.J Biol Chem. 2010; 285: 27213-27223Crossref PubMed Scopus (81) Google ScholarNormal expression of GD3 is found primarily on cells of neuronal origin.17Nakatani Y. Yanagisawa M. Suzuki Y. Yu R.K. Characterization of GD3 ganglioside as a novel biomarker of mouse neural stem cells.Glycobiology. 2010; 20: 78-86Crossref PubMed Scopus (53) Google Scholar Interestingly, mTOR activity, which is up-regulated in LAM, has been linked to GD3 expression.18Yamauchi Y. Furukawa K. Hamamura K. Furukawa K. Positive feedback loop between PI3K-Akt-mTORC1 signaling and the lipogenic pathway boosts Akt signaling: induction of the lipogenic pathway by a melanoma antigen.Cancer Res. 2011; 71: 4989-4997Crossref PubMed Scopus (68) Google Scholar The molecule is a glycosphingolipid, expressed in the membrane compartment.19Kim J.K. Kim S.H. Cho H.Y. Shin H.S. Sung H.R. Jung J.R. Quan M.L. Jiang D.H. Bae H.R. GD3 accumulation in cell surface lipid rafts prior to mitochondrial targeting contributes to amyloid-β-induced apoptosis.J Korean Med Sci. 2010; 25: 1492-1498Crossref PubMed Scopus (12) Google Scholar GD3 can thus be targeted by antibodies to block tumor growth.20Portoukalian J. Carrel S. Doré J.F. Rümke P. Humoral immune response in disease-free advanced melanoma patients after vaccination with melanoma-associated gangliosides. EORTC Cooperative Melanoma Group.Int J Cancer. 1991; 49: 893-899Crossref PubMed Scopus (46) Google Scholar As a processed antigen, GD3 is presented by CD1d, a molecule expressed on the surface of antigen-presenting cells (APCs), which recognizes only lipids and glycolipids.21Park J.E. Wu D.Y. Prendes M. Lu S.X. Ragupathi G. Schrantz N. Chapman P.B. Fine specificity of natural killer T cells against GD3 ganglioside and identification of GM3 as an inhibitory natural killer T-cell ligand.Immunology. 2008; 123: 145-155Crossref PubMed Scopus (41) Google Scholar CD1d-restricted antigens are recognized by natural killer T (NKT) cells.With this in mind, we set out to determine the potential of GD3 as a target for immunotherapeutic treatment of LAM. In the present study we assessed the expression of GD3 in LAM lung and compared its abundance with that found in normal lung tissue and in melanoma tumors. We also confirmed that GD3-overexpressing cells are in fact LAM cells. We performed ELISAs and measured levels of serum GD3 antibodies and also defined potentially ongoing immune responses to GD3 in patients and controls. We established whether surface expression of GD3 can sensitize LAM cells to humoral immune responses. Because so many patients with LAM are now treated with rapamycin analogs, we evaluated GD3 expression in LAM cells exposed to the drug in culture. Finally, we assessed the abundance of tissue-infiltrating NKT cells, as well as CD1d-expressing APCs, in LAM. Taken together, the data serve to identify and establish a tumor antigen that can form the basis for the development of effective immunotherapy for patients with LAM.22Carbone M. Feasibility of immunotherapy for lymphangioleiomyomatosis.Am J Pathol. 2009; 175: 2252-2254Abstract Full Text Full Text PDF PubMed Scopus (4) Google ScholarMaterials and MethodsTissue SourcesFresh LAM tissue was obtained from patients undergoing surgical biopsy at Loyola University Medical Center and also through the National Disease Research Interchange (NDRI) tissue repository (Philadelphia, PA). Control lung samples were obtained from the NDRI; LAM tissues were obtained from necropsies and from patients undergoing wedge resection surgery for other pathology at Loyola University Medical Center. Samples were snap frozen, and 8-μm cryosections were cut and fixed in cold acetone, then stored at −20°C until use. LAM diagnosis provided by the NDRI was confirmed by indirect HMB45 immunostaining (Dako, Glostrup, Denmark). Melanoma samples were obtained as resected surgical tissue. Ten serum samples from patients with LAM were obtained from the NDRI tissue bank. Serum samples from control subjects were obtained from healthy volunteers. Pooled melanoma serum was obtained from Innovative Research (Novi, MI). Both tissue and serum LAM samples were from patients with sporadic LAM. One patient tissue was obtained from a patient with tuberous sclerosis. Tissue from this sample was used only in NKT quantification. All samples included in this study were obtained with prior approval from the Institutional Review Committee in accordance with principles to the Declaration of Helsinki. Strict precautions were taken to protect the patients' identities throughout these studies. Tsc2−/− murine embryo fibroblasts were a generous gift from by Dr. Vera Krymskaya (University of Pennsylvania) and were originally generated by Dr. David Kwiatkowski.23Zhang H. Cicchetti G. Onda H. Koon H.B. Asrican K. Bajraszewski N. Vasquez F. Carpenter C.L. Kwiatkowski D.J. Loss of Tsc1/Tsc2 activates mTOR and disrupts PI3K-Akt signaling through downregulation of PDGFR.J Clin Invest. 2003; 112: 1223-1233Crossref PubMed Scopus (439) Google Scholar Frozen mouse tissues from aged Tsc2+/− and Tsc1+/− mice included in immunohistology analysis were a kind donation by Dr. Jeremy Cheadle.24Wilson C. Idziaszczyk S. Parry L. Guy C. Griffiths D.F. Lazda E. Bayne R.A. Smith A.J. Sampson J.R. Cheadle J.P. A mouse model of tuberous sclerosis 1 showing background specific early post-natal mortality and metastatic renal cell carcinoma.Hum Mol Genet. 2005; 14: 1839-1850Crossref PubMed Scopus (50) Google Scholar For work with fresh tissues to perform fluorescence-activated cell sorting (FACS) analysis, Tsc2+/− mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and aged in house. Animal work performed as part of this investigation was approved by the Institutional Animal Care and Use Committee.Cell Culture and Rapamycin TreatmentFresh tissue samples from transplanted lung of five patients with advanced LAM disease were obtained from the NDRI within 24 hours. Tissue was incubated overnight while shaking in an enzyme cocktail of collagenase type IV (Sigma-Aldrich, St. Louis, MO), thermolysin (Sigma-Aldrich), trypsin (Invitrogen, Carlsbad, CA), and DNaseI (Roche, Madison, WI). Cells were plated in Dulbecco's modified Eagle's medium (Cellgro, Manassas, VA), streptomycin, penicillin, and heat-inactivated fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA). Expression of α-smooth muscle actin with the use of antibody clone 0.N.5 (Santa Cruz Biotechnology Inc., Santa Cruz, CA) and estrogen receptor-α with the use of rabbit polyclonal antibody G-20 (Santa Cruz Biotechnology Inc.; Santa Cruz, CA) was confirmed by immunostaining of adherent cells as described.11Klarquist J. Barfuss A. Kandala S. Reust M.J. Braun R.K. Hu J. Dilling D.F. McKee M.D. Boissy R.E. Love R.B. Nishimura M.I. Le Poole I.C. Melanoma-associated antigen expression in lymphangioleiomyomatosis renders tumor cells susceptible to cytotoxic T cells.Am J Pathol. 2009; 175: 2463-2472Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar Cultured LAM cells and Tsc2−/− cells were incubated for 24 hours with 5-nmol/L, 10-nmol/L, and 20-nmol/L concentrations of rapamycin (Sigma-Aldrich).Immunostaining ProceduresNonspecific staining was averted with SuperBlock (Cytek Labs, Logan, UT), pretreatment with hydrogen peroxide, and dilution of primary antibodies in 10% normal human serum (Valley Biomedical Inc., Winchester, VA). Antibodies used for human tissues include R24 to GD3 (Covance, Princeton, NJ), B-ly6 to CD11c (BD Pharmingen, San Diego, CA), NOR3.2/13.17 to CD1d (Santa Cruz Biotechnology Inc.), HIT-3a to CD3 (BD Pharmingen), HMB45 to gp100 (Dako), B159 to CD56 (BD Pharmingen), 1D5 to estrogen receptor-α (Dako), and 0.N.5 to smooth muscle α-actin (Santa Cruz Biotechnology Inc.). Antibodies used for mouse tissues include T-16 to CD68 (Santa Cruz Biotechnology Inc.) and HL3 to CD11c (BD Pharmingen). Species- and isotype-specific secondary antibodies labeled with alkaline phosphatase, horseradish peroxidase, or fluorescein isothiocyanate, phosphatidylethanolamine, and allophycocyanin cyanine dye 7 fluorochromes were all from Southern Biotech (Birmingham, AL). For single staining, secondary antibodies were labeled with horseradish peroxidase and counterstained with hematoxylin. For fluorescent staining, DAPI counterstaining was used. In double staining procedures, alkaline phosphatase staining was detected first with the use of Fast Blue BB (Sigma-Aldrich) in buffer containing levamisole and naphthol AS-MX phosphate (Sigma-Aldrich). Slides were then exposed to 3-amino-9-ethyl carbazole (Sigma-Aldrich) in the presence of hydrogen peroxide to visualize staining. Quantification of staining was performed by evaluating the density of stained cells, excluding alveolar spaces, per mm2 with the use of an Olympus AX80T Microscope (Melville, NY) and Photoshop software version 12.0 (Adobe, San Jose, CA). Fluorescent microscopy was performed with a Zeiss LSM 510 inverted confocal microscope (Zeiss, Oberkochen, Germany).FACS AnalysisCultured mouse Tsc2−/− and human LAM cells and digested mouse and human lung tissues were subjected to immunostaining with primary antibody R24 to GD3 (Covance), NOR3.2/13.17 to CD1d (Santa Cruz Biotechnology Inc.), T-16 to CD68 (Santa Cruz Biotechnology Inc.), or B-ly6 to CD11c (BD Pharmingen). Antibodies that were not directly labeled were detected with isotype-specific fluorochrome-labeled secondary antibodies (Southern Biotech). Fluorocytometry was used to quantify fluorescence in 50,000 cells with the use of FACScanto II and Fortessa LTR 2 instrumentation (BD Biosciences). Data were analyzed with FlowJo software version 7.6.3 (TreeStar Inc., Ashland, OR).ELISAEnzyme-linked immunosorbent assay (ELISA) plates (R&D Systems, Minneapolis, MN) were coated with bovine brain-derived GD3 antigen (Enzo Life Sciences, Plymouth Meeting, PA) in methanol. Tween and 0.1% bovine serum albumin in PBS was used to block nonspecific binding before incubation with diluted pooled serum from patients with melanoma, healthy subjects, and patients with LAM. To determine relative serum titers, the GD3 antibody R24 (Covance) was used in a range of 4 to 300 ng/mL as a standard alongside the serum samples. Bound GD3 antibodies were detected with horseradish peroxidase-labeled, goat anti-human IgG secondary antibody (Southern Biotech) or HRP-labeled goat anti-mouse IgG3 (Southern Biotech) to detect R24 binding. Peroxidase was detected with tetramethylbenzidine substrate (R&D Systems) and stopped with sulfuric acid. Reaction product in the supernatant fluid was measured in a Polar Star Omega ELISA plate reader (BMG Labtech, Offenburg, Germany) at 450 nm. Data were analyzed with MARS data analysis software version 1.10 (BMG Labtech, Cary, NC).Cytotoxicity AssayFor complement-dependent cytotoxicity assays, target cells were plated in triplicate wells of a 96-well flat bottom plate, then exposed to R24 antibody to GD3 (Covance) in combination with purified human complement (Quidel Corporation, Santa Clara, CA) in Dulbecco's modified Eagle's medium (Cellgro), streptomycin, penicillin, and heat-inactivated fetal bovine serum (Atlanta Biologicals). Wells that contained target cells plus 0.3% saponin permeation buffer served to represent total lysis; wells without R24 antibody served as negative controls. After 16 hours of incubation, supernatant fluids were subjected to MTT assay analysis (Bioassay Systems, Hayward, CA). Briefly, supernatant fluids were incubated with MTT substrate (Bioassay Systems). Solubilization solution (Bioassay Systems) was subsequently added to each well, and plates were incubated overnight. Colorimetric change was measured with a Star Omega plate reader (BMG Labtech) at 570 nm.Statistical AnalysisTo identify treatments that induced a statistically significant difference in results among two groups, Student's t-test was performed with Prism software version 5.01 for Windows (GraphPad Inc., San Diego, CA). Error bars in the bar graph show the SEM.ResultsExpression of GD3 Is Markedly More Abundant in Diseased LungAs shown in Figure 1, GD3 is overexpressed in both LAM lung (Figure 1B) and melanoma (Figure 1C) compared with minimal expression in normal lung (Figure 1A). Although an average of 417 cells/mm2 in LAM lung expressed detectable GD3, only 192 cells/mm2 from a normal lung were GD3 positive (Figure 1D) (P = 0.03). In cells expressing GD3 within LAM tissue, the staining intensity was also appreciably higher than in normal lung, representative of higher levels of expression. GD3 overexpression in LAM lung homogenate is further supported by fluorocytometry (Figure 1E), displaying a mean fluorescent intensity of 5900 in LAM and 158 in normal lung (data not shown). Colocalization of SMA, gp100, and GD3 was observed (Figure 1F). Importantly, GD3 was always associated with SMA-expressing cells. Taken together, these data sets indicate marked GD3 overexpression in LAM tissue.Expression of GD3 Is Intrinsic to Cultured LAM Cells and Persists in Cell Cultures Established from LAM Lung over Multiple PassagesCoexpression of LAM markers, including smooth muscle α-actin, estrogen receptor-α, and GD3, in LAM 10224 chamber slides reveal that LAM cells retain expression of these characteristic molecules in vitro (Figure 2A). To assess antigen retention in cell culture, three samples initiated from LAM lung tissue were maintained over multiple passages. GD3 expression was demonstrated throughout at least eight passages, as shown (Figure 2B). Moreover, fluorocytometry confirms GD3 expression in 27% of LAM cells in the fourth passage (Figure 2C). Importantly, LAM samples treated with various rapamycin doses, including one analogous to the dose administrated to patients, showed an increase in GD3 expression after 24 hours (Figure 2D). These data suggest that targeting GD3 could be even more effective in conjunction with rapamycin treatment. Detectable expression throughout at least eight passages and retention of hallmark surface markers indicate cultured LAM cells can be used for in vitro functional assays to assess GD3-mediated immune targeting.Figure 2GD3 expression is sustained in cell cultures established from LAM lung. A: Chamber slides of LAM10224 (P2) shows comparative expression of GD3, smooth muscle α-actin, and estrogen receptor-α. B: Cell cultures initiated from LAM lung were passaged up to eight times. GD3 expression (red) was retained among adherent cells from three donors. Scale bar = 20 μm. C: FACS analysis of GD3 surface expression in LAM10224 P4 cells. D: LAM10224 (P4) and Tsc2−/− (P12) cells were treated with 5, 10, or 20 nmol/L rapamycin for 24 hours and analyzed for GD3 expression by fluorocytometry and show optimal expression in response to 10 nmol/L rapamycin.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Patients with LAM Exhibit Markedly Reduced Antibody Titers to GD3ELISAs were performed with sera from control human subjects, patients with melanoma, and patients with sporadic LAM. An indirect ELISA measuring GD3 antibody titers demonstrated that GD3 antibodies were significantly more abundant in normal serum (threefold increase) compared with LAM samples with a P value of 0.005. We also found that melanoma serum contained similarly reduced concentrations of GD3 antibodies than normal serum (Figure 3). This indicates that patients with LAM (and patients with melanoma) do not mount a humoral response to antigen after the overexpression of GD3.Figure 3Reduced humoral responses to GD3 in patients with LAM. LAM and melanoma (n = 10) and control (n = 4) serum samples were included in an ELISA to compare GD3 antibody titers. Patients with LAM exhibit reduced antibody titers to GD3 compared with control serum. ∗P = 0.015. Melanoma data similarly show reduced serum antibody levels. ∗∗∗P < 0.0001. Ab, antibody.View Large Image Figure ViewerDownload Hi-res image Download (PPT)LAM Cells Are Sensitive to Complement-Mediated Cytotoxicity Targeting GD3The sensitivity of LAM cells to complement-dependent cytotoxicity responses was measured in the presence and absence of commercial anti-GD3 antibodies. Cytotoxicity was compared with that observed toward GD3-expressing M14 and 624.38 melanoma cells as well as Tsc2−/− cells as positive controls and COS7 cells that served as a GD3-negative control. When analyzed by MTT assays, COS7 cells remained 100% viable, whereas 12% to 42% of LAM cells were killed on the addition of anti-GD3 plus human complement source (P = 0.005) compared with 21% to 45% of melanoma cells (P = 0.02), showing the sensitivity of LAM cells to this effector mechanism (Figure 4).Figure 4LAM cells (white bars) are sensitive to complement-dependent cytotoxicity with antibodies to GD3. Triplicate wells of GD3+ (M14, 624.38, Tsc2−/−, grey bars) and GD3− (COS7, black bar) cells were incubated for 16 hours with antibody R24 and human complement. LAM-cell killing ranged from 12% to 42% in the presence of GD3 antibody and complement compared with only 0.2% of negative control cell. ∗P = 0.005, ∗∗P < 0.01.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Abundance of NKT Cells Is Similar in LAM and Control LungQuantification of tissue-infiltrating NKT (CD3/CD56 double positive) cells in LAM lung was performed (Figure 5). This analysis also allows for a comparison of T-cell and NK-cell infiltration. As exemplified by representative tissue slides in LAM lung (Figure 5A) and control tissue (Figure 5B), LAM lung contained similar numbers of infiltrating NKT cells compared with normal lung tissue (Figure 5C). Trends toward decreased infiltration by NK cells and T cells did not prove significant in this study (P = 0.24 and 0.07, respectively). By combining this data set with the absence of measurable humoral responses shown in Figure 3, G