A natural history of melanoma: serial gene expression analysis

Ena Wang and Francesco Marincola propose a novel strategy whereby tumor-host interactions are studied within the melanoma microenvironment by serial gene expression analysis. Methodological constraints and ways to circumvent them are discussed. This
of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  34 Michetti, P. et al. (1999) Oral immunization with urease and Escherichiacoli heat-labile enterotoxin is safe and immunogenic in  Helicobacter  pylori -infected adults. Gastroenterology 116, 804–812 35 Dickinson, B.L. and Clements, J.D. (1995) Dissociation of Escherichiacoli heat-labile enterotoxin adjuvanticity from ADP-ribosyltransferase activity. Infect. Immun. 63, 1617–1623 36 Kim, S.Y. et al. (1999) Oral immunization with  Helicobacter pylori -loaded poly( D  , L -lactide-co-glycolide) nanoparticles.  Helicobacter 4, 33–39 37 Hasan, U.A. et al. (1999) Nucleic acid immunization: concepts andtechniques associated with third generation vaccines.  J. Immunol. Methods 229, 1–22 38 Davis, H.L. (1997) Plasmid DNA expression systems for the purposeof immunization. Curr. Opin. Biotechnol. 8, 635–640 39 Mitchell, H.M. et al. (1992) Epidemiology of  Helicobacter pylori inSouthern China – identification of early childhood as the criticalperiod for acquisition.  J. Infect. Dis. 166, 149–153 40 Blaser, M.J. (1997) Not all  Helicobacter pylori stains are created equal:should all be eliminated? Lancet 349, 1020–1022 41 Raz, E. et al. (1994) Intradermal gene immunization: the possible roleof DNA uptake in the induction of cellular immunity to viruses. Proc.Natl. Acad. Sci. U. S. A. 91, 9519–9523 42 Condon, C. et al. (1996) DNA-based immunization by in vivo transfection of dendritic cells. Nat. Med. 2, 1122–1128 43 Kreig, A.M. (1999) Mechanisms and applications of immunestimulatory CpG oligonucleotides. Biochim. Biophys. Acta 10, 107–116 44 Felgner, P.L. (1999) Prospects for synthetic self-assembling systems ingene delivery.  J. Gene Med. 1, 290–292 45 Scubbert, R. et al. (1997) Foreign (M13) DNA ingested by mice reachesthe peripheral leukocytes, spleen and liver via the intestinal wallmucosa and can be covalently linked to mouse DNA. Proc. Natl. Acad.Sci. U. S. A. 94, 961–966 46 Alm, R.A. et al. (1999) Genomic sequence comparison of two unrelatedisolates of the human gastric pathogen  Helicobacter pylori. Nature 397, 176–180 47  Johnston, S.A. and Barry, M.A. (1997) Genetic to genomic vaccination. Vaccine 15, 808–809 48 Dorrell, N. and Wren, B.W. (1998) From genes to genome biology: a new era in  Helicobacter pylori research. Gut 42, 451–453 eturning hope to a patientwith widespread cancer byinducing tumor rejectionwith immune manipulation isthe goal of many clinicians. However, themolecular events that regulate this relativelyrare occurrence remain elusive. Never-theless, the past three years have witnesseda remarkable breakthrough in the under-standing of the molecular interactions be-tween a tumor and its host: it has beenshown that the immune system naturallymounts humoral and cellular responsesagainst tumor antigens (TAs) expressed byautologous cancer cells 1–3 . Most of these TAsare products of non-mutated genes encodingintracellular proteins that are expressed by cancers from differentpatients 1,2 . Thus, interest in cellular immune responses against these‘shared’ TAs has triggered enthusiasm for the development of immunizations with T-cell epitopes that might benefit a broadpatient population 4 .Despite these advances, progress in the understanding of tumor–host interactions has led to only modest clinical improve-ment. Tumor regression remains capricious and the reason for oc-currence elusive. In particular, vaccination trials have revealed dis-crepancies between efficacy in eliciting T-cell responses and limitedclinical success 4,5 . A review of the literaturedemonstrates an extensive and ever-grow-ing number of hypotheses that explain theerratic behavior of human cancers 6 .Experience has shown that predictions basedon pre-clinical models, although conceptu-ally useful, frequently deviate from the be-havior of human disease. Such models arefacilitated both by the genetic identity of ani-mals and diseases, and the rigorous experi-mental conditions that can be applied. Bycontrast, humans are genetically polymor-phic, their diseases are heterogeneous incause, anatomical srcin and phenotype, andtreatments are often modified according toclinical, practical and/or ethical considera-tions. Indeed, the unpredictability of human cancer depends on theheterogeneity of both host and tumor, and significantly hinders ourunderstanding of tumor rejection.Here, an alternative strategy is proposed for the study of tumor–host interactions in melanoma, based on fine-needle aspi-rates (FNAs) and serial gene expression analysis. This approach per-mits repeated analysis of a tumor over time and limits the experi-mental noise created by tumor–host heterogeneity. It might,therefore, be possible to correlate the molecular basis of tumorregression with the host response to immune manipulation. VIEWPOINT IMMUNOLOGYTODAY 0167-5699/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII:S0167-5699(00)01724-2 DECEMBER2000 Vol.21 No.12 619 DECEMBER2000 A natural history of melanoma: serial geneexpression analysis Ena Wang and Francesco M. Marincola Ena Wang and Francesco Marincola propose a novel strategy wherebytumor–host interactions are studiedwithin the melanomamicroenvironment by serial geneexpression analysis. Methodologicalconstraints and ways to circumventthem are discussed. This approachmight improve our understanding of the molecular basis of tumorregression in response to immunemanipulation. R  The multiple ways to tumor tolerance It has been suggested that inadequate and/or variable immune re-sponses might explain tumor persistence 7–10 . However, TA-specificcytotoxic T lymphocytes (CTLs) can easily be identified in patientswith cancer 11 , and immunization with TA-derived epitopes ap-pearsto benefit rather than hamper immune competence 4,5 .Althoughlocalization of tumor-infiltrating lymphocytes (TILs) at the tumorsite is a prerequisite for a clinical response, it does not seem suf-ficient, as it is not always followed by tumor regression 12 . In ad-dition, the observation of ‘mixed responses’ (characterized by diver-gent behavior of synchronous metastases in response to therapy)points to the importance of factors within the tumor microenviron-ment. For this reason, it is of interest to develop strategies to moni-tor tumor–host interactions within the tumor microenvironmentitself.Such an approach is still complicated by variables associated withhuman polymorphism and tumor heterogeneity. The human leuko-cyte antigen (HLA) molecules that present T-cell epitopes are themost polymorphic structures of the human genome. In addition, thenumber of TAs and T-cell epitopes associated with different HLAclass I alleles is increasing exponentially 6 . In contrast to the pluri-potentiality of tumor–T-cell interactions, HLA–epitope associationsfor the determination of antigenic potential are extremely stringent 13 .Thus, in a given patient, the wealth of the immune response towardsa TA is restricted by (1) its expression level, (2) the subject’s HLAphenotype and (3) the level of HLA expressed by tumor cells 6 . It isreasonable to suggest that each cancer patient represents an inde-pendent immunological entity: different metastases from the samepatient may be separate biological entities, as the phenotype of syn-chronous lesions can be quite different 14 . By tailoring the comparisonof tumor–host interactions to each lesion, direct conclusions can bemade independently of assumptions extrapolated by analogy toother lesions. Dealing with human polymorphism and theheterogeneity of human diseases The use of immunohistochemical (IHC) or molecular analysis of sur-gically removed tumors to characterize tumor–host interactions islimited by the inability to perform functional studies. Similarly, theutilization of cells freshly isolated by in vitro digestion of tumors forfunctional studies is limited by contamination with different celltypes, which confounds quantitative estimates. Although the expan-sion of TIL/tumor-cell pairs from excised tumors has provided an el-egant model for the in vitro characterization of tumor–CTL inter-actions 15 , such characterizations are not necessarily representative of  in vivo conditions.Most importantly, analysis of excised tumor speci-mens yields static information about a disease characterized by ex-treme genetic instability, and the natural course of the removedtumor cannot be studied.Tumor–T-cell interactions could be better studied by followingthe progression of events occurring within the same lesion by re-peated FNAs obtained at time points most relevant to the natural his-tory of the disease (Box 1). In particular, the combination of serialanalysis of a single lesion, which limits the experimental noise cre-ated by tumor heterogeneity, and the monitoring of epitope-specificvaccines restricted to a single TA–HLA allele combination, providesthe unprecedented opportunity to dissect the molecular componentsof tumor–host interactions efficiently. Melanoma and serial gene expression analysis oftumor–host interactions Focusing on treatments restricted to a single allele–epitope combi-nation and studying a single lesion should improve our under-standing of tumor–host interactions. However, several questionsneed to be addressed relating to the genetic profile of cancer cellsand/or the presence and status of activation of immune regulatorycells. Comparison of pre- and post-treatment FNAs has indicatedloss of expression of the TA targeted by the vaccination in a meta-stasis that recurred after initial regression 16 . Expansion of TILs fromthe FNA of this recurring lesion demonstrated a functional dissoci-ation between the immune response at the tumor site and that ob-served systemically, and identified a new cryptic epitope of theMAGE-12 protein whose expression had persisted. Expansion of TIL/tumor-cell pairs from repeated FNAs of single lesions in pa-tients undergoing epitope-specific vaccination demonstrated thatvaccine-elicited T cells can be expanded more readily frommelanoma metastases after treatment, suggesting localization of theeffects of the vaccine at the tumor site 17 . In an effort to obtain infor-mation relevant to the natural history of melanoma, with less laborintensity and more flexibility, serial gene expression analysis was as-sessed from RNA obtained directly from FNAs. This was achieved by quantitative real-time polymerase chain reaction (qRT-PCR) of  VIEWPOINT IMMUNOLOGYTODAY 620Vol.21 No.12 DECEMBER2000 Box 1. Strategies for studying the tumor microenvironment Excisional biopsies a Provides good quantity of material for studyPreserves structural informationDoes not allow serial sampling of the same lesionDoes not allow prospective assessment of the natural history of a given lesion Expansion of tumor/T-cell pairs Allows functional studiesIs not necessarily representative of in vivo conditions Fine-needle aspirates a Provides limited quantity of material for studyDoes not preserve structural informationAllows sampling of the same lesionAllows prospective assessment of the natural history of a givenlesion a Functional information might be obtained with the following new technol-ogies: quantitative real-time polymerase chain reaction (qRT-PCR); intra-cellular fluorescence-activated cell sorter (FACS); human leukocyte antigen(HLA)–epitope tetrameric complexes; microarray analysis.  interferon  (IFN-  ) and/or other cytokinesin pre- and post-vaccination FNAs. Enhance-ment of IFN-  expression was noted aftertreatment in lesions that maintained expres-sion of the targeted TA (Ref.18). Since noneof the lesions responded to treatment, thisstudy excluded lack of tumor localizationand loss of TAs as factors hindering vaccineeffectiveness.Although conceptually demonstratingthe utility of FNAs, these studies were moreeffective at excluding hypotheses that ex-plain tumor escape than at identifying newones. Furthermore, no regressing lesionswere studied because the complexity of themethods restricted the chance of identifyingthese relatively rare events by limiting thenumber of samples that could be studied.Finally, few hypotheses could be studiedowing to the limited amount of RNA obtain-able with FNAs (median yield of approxi-mately 2  g of total RNA). In fact, the smallamount of material available from FNAscould be more valuable when more genes are capable of being studied with the samesmall amount of material.Thus, we con-cluded that a preferred strategy would alloweasy and extensive collection of materialfrom clinical samples to be retrospectivelyfollowed according to clinical relevance (i.e. responding versus non-responding lesions) and to be studied with high-throughputtechnology (Fig. 1).The conservation of FNA material as a future source of RNA is asimple, short, inexpensive procedure that allows the collection of large archives of potentially relevant clinical specimens. Using thisapproach, pre-treatment FNA material from lesions that eventuallydid or did not respond to vaccine-based treatment were gatheredand studied by qRT-PCR for expression of genes potentially relatedto TAs, immunosuppressive cytokines or cell-cycle regulators. In thispivotal study, technological limitations allowed the analysis of ex-pression of only 20–30 genes that did not have predictive signifi-cance (E. Wang and F.M. Marincola, unpublished), leaving an openquestion about the remaining overwhelming number of factors thatcould influence tumor behavior 6 . New techniques for analyzing tumor–host interactions The completion of the Human Genome Project presents the oppor-tunity to expand the frontiers of gene expression analysis to the out-most boundaries of functional genomics 19 . A global view of biologi-cal processes could theoretically be achieved by simultaneouslymonitoring the expression of all human genes using microarray tech-nology 20 . This new approach has been compared with the impact of the discovery of the periodic table on chemistry 21 . Patterns of geneexpression in cancer cell lines 22 , excised surgical specimens 23,24 andother cancerous tissues 25 have been described using this strategy.Although the full genome map is orders of magnitude larger than theperiodic table, powerful statistical methods and supportive softwarehave identified patterns of gene expression and derived functionalassumptions from such extensive information 22,26 . As predicted byLander 21 , global patterns of gene expression may classify tumors ac-cording to differential prognostic or biological behavior 24,25,27 .Furthermore, clustering algorithms can identify discrete expressionpatterns within a complex mixture of cell types, such as cancer fromstromal or circulating cells 24 . Finally, as suggested by Brown andBotstein 28 , the tight correlation existing between the function of agene product and its expression pattern provides a compellingreason for a genome-wide formulation of scientific hypotheses.The ex vivo analysis of tissue samples using cDNA microarraytechnology has been, to various degrees, limited by the amount of RNA necessary for conventional cDNA microarrays [50–200  g of total RNA and 2–5  g poly(A)-RNA] 29 . Duggan et al . calculated thatthe minimal amount of cellular material sufficient for conventionalmicroarray analysis is in the range 10 6 –10 7 cells 29 . This number of cells is easily obtainable from cell cultures or clinical biopsies but isabove, or just at the limit of, the yield from histological samples (e.g.via FNAs or micro-dissection) or material from developmental stud-ies. To broaden the use of cDNA microarrays to experimental con-ditions in which source material is the limiting factor, some methods VIEWPOINT IMMUNOLOGYTODAY Vol.21No.12 621 DECEMBER2000 Fig. 1. Suggested strategy for the collection of large libraries of relevant clinical samples with mini-mum cost and effort. The FNA sample is analyzed by IHC to ensure that adequate material is obtained.Cells from the aspirate are cultured in the presence or absence of IL-2 to develop TIL/tumor-cell pairs,and the majority of the specimen is simply frozen. Because of the relatively low incidence of clinicalresponses, it is expected that a large number of samples need to be collected to identify the few clearlyresponding lesions. These few clinically relevant samples and appropriate controls can then be processed further in gene expression analysis studies. Abbreviations: aRNA, anti-sense RNA; FNA, fine-needle aspirate; IHC, immunohistochemistry; IL-2, interleukin 2; qRT-PCR, quantitative real-time polymerase chain reaction; TIL, tumor-infiltrating lymphocyte. Immunology Today  PCRcDNA micro-arrayqRT-PCRaRNASample selection according to clinical relevanceIHCSnap freezeCulture with IL-2FNAFNATimeUnchangedChanged because of genetic instabilityChanged because of immune selectionChanged because of immune regulation Gene expression:  aim to reduce the detection limits by intensifying the fluorescentsignal, while other methods aim to amplify starting poly(A)-RNA or cDNA and maintain the relative expression of each transcriptpopulation. We devised a procedure that allows 10 5 -fold enrichmentof limiting amounts of source RNA by combining anti-sense RNA(aRNA) amplification with template-switching effect (ClontechLabs, Palo Alto, CA) to generate full-length double-stranded cDNAfor the first round of amplification 30 (for details, see Fig. 2 legend).We have recently described the quantitative comparison of linearityand reproducibility of this method in relation to total RNA orpoly(A)-RNA-based arrays 30 . Grading by ‘hierarchical clusteringdiscrimination logic’ demonstrated the high fidelity of this techniqueand suggested that aRNA-based gene expression patterns maintainlinearity and reproducibility equal to total-RNA- or poly(A)-RNA- based arrays. This technique opens the utilization of cDNA micro-arrays to the analysis of the gene profile in samples obtained fromFNAs or micro-dissection specimens. It is now possible to amplify amedian of 80  g aRNA per FNA, which is sufficient for conven-tional-sized cDNA microarrays utilizing 2  g of pure transcript(Fig.3). Thus, it is now feasible to analyze global gene expressionprofiles in samples from FNAs throughout the natural history of atumor, and/or in response to treatment. Conclusion It is hoped that the combination of serial sampling of the same lesionwith high-throughput methods of analysis will expedite under-standing of the molecular basis of tumor regression in response toimmune manipulation. With this strategy, genetic and/or immuno-logical variability of tumors can be documented. Furthermore, retro-spective correlation of gene expression profiles with the clinical out-come of individual lesions might ultimately yield a global theoryexplaining tumor tolerance in humans and, perhaps, suggest waysto circumvent it. We thank C. Fox, D. Kedes and E. Marincola for helpful discussion of the manuscript. Ena Wang  and  Francesco Marincola (marincola@nih.gov) are at theSurgery Branch, and  Francesco Marincola is also at the Dept of Transfusion Medicine, Clinical Center, National Cancer Institute, NationalInstitutes of Health, Bethesda, MD 20892-1502, USA. References 1 Boon, T. et al. (1997) Tumor antigens recognized by T cells . Immunol.Today 18, 267–268 2 Rosenberg, S.A. (1997) Cancer vaccines based on the identification ofgenes encoding cancer regression antigens. Immunol. Today 18, 175–182 3 Old, L.J. and Chen, Y.T. (1998) New paths in human cancer serology.  J. Exp. Med. 187, 1163–1167 4 Rosenberg, S.A. et al . (1998) Immunologic and therapeutic evaluationof a synthetic tumor associated peptide vaccine for the treatment ofpatients with metastatic melanoma. Nat. Med. 4, 321–327 5 Cormier, J.N. et al. (1997) Enhancement of cellular immunity inmelanoma patients immunized with a peptide from MART-1/MelanA. Cancer J. Sci. Am. 3, 37–44 6 Marincola, F.M. et al . (2000) Escape of human solid tumors from T cellrecognition: molecular mechanisms and functional significance.  Adv.Immunol. 74, 181–273 VIEWPOINT IMMUNOLOGYTODAY 622Vol.21 No.12 DECEMBER2000 Fig. 2. Target preparation method combining template switching and in vitro transcription. The template-switching (TS) effect was applied at the end of reverse transcription (RT) to generate full-length double-stranded (ds) cDNAduring the first round of amplification. The use of the TS effect enriches for full-length transcripts and could also be utilized for the second round of ampli- fication, although this modification has not been tested as yet. When RT reachesthe 5’ end of the mRNA, the terminal transferase activity of RTase results inthe addition of a few dC residues to the cDNA. The TS primer containing threedG residues at the 3   end anneals at 42 o C to the dC string, producing a shortsegment of cDNA duplex. After RNase treatment (with Escherichia coli RNase H), the TS primer initiates ds cDNA synthesis at 75 o C under Advantage cDNA Polymerase (Clontech, Palo Alto, CA, USA), which opti-mizes the amplification of large DNA fragments. This ds cDNA is then used asa template for in vitro transcription. The ds cDNA synthesis in the secondround of amplification is started with random hexamers (dN6) and oligo dT-T7(for details see Ref. 30). Abbreviations: N, unspecified number of RNA copiesobtained with the sequence; 15, number of Ts in the discussed sequence. Immunology Today  AAAAAA(N)AAAAAA(N)TTTTT(15)-T7 -5 ′ AAAAAA(N)TTTTT(15)-T7 -5 ′ RT E.coli   RNase H TTTTT(15)-T7-5 ′ G G G  CCCTS primerGGGCCCGGGAdvantage cDNA PolymeraseCCCTSGGGCCCTTTTT (15)-T7-5 ′ cDNAmRNAds cDNAUUUUU(15)T7 RNA PolymeraseaRNAExtendedcDNAUUUUU(15)RTdAdAdA(15) E.coli   RNaseH dAdAdA(15)Advantage cDNA PolymeraseT7 RNA PolymerasedN6dAdAdA(15)dAdAdA(15) (a)  First-round amplification (b)  Second-round amplification N copy of aRNAdAdAdA(15)-T7-3 ′ dAdAdA(15)-T7-3 ′ TTTTTT(15)- T7-5 ′ UUUUUTTTTT(15)T7-5 ′  7 Toes, R.E. et al. (1997) Activation or frustration of anti-tumor responsesby T-cell-based immune modulation. Semin. Immunol. 9, 323–327 8 Lee, P.P. et al. (1999) Characterization of circulating T cells specific fortumor-associated antigens in melanoma patients. Nat. Med. 5, 677–685 9 Fuchs, E.J. and Matzinger, P. (1996) Is cancer dangerous to the immunesystem? Semin. Immunol. 8, 271–280 10 Ochsenbein, A.F. et al . (1999) Immune surveillance against a solidtumor fails because of immunological ignorance. Proc. Natl. Acad. Sci.U. S. A. 96, 2233–2238 11 Marincola, F.M. et al. (1996) Differential anti-MART-1/MelanA CTLactivity in peripheral blood of HLA-A2 melanoma patients incomparison to healthy donors: evidence for in vivo priming by tumor cells.  J. Immunother. 19, 266–277 12 Pockaj, B.A. et al . (1994) Localization of 111indium-labeled tumorinfiltrating lymphocytes to tumor in patients receiving adoptiveimmunotherapy. Augmentation with cyclophosphamide andcorrelation with response. Cancer 73, 1731–1737 13 Bettinotti, M.P. et al. (1998) Stringent allele/epitope requirements forMART-1/Melan A immunodominance: implications for peptide-basedimmunotherapy.  J. Immunol. 161, 877–889 14 Cormier, J.N. et al. (1998) Heterogeneous expression of melanoma-associated antigens (MAA) and HLA-A2 in metastatic melanoma in vivo . Int. J. Cancer 75, 517–524 15 Pandolfi, F. et al. (1991) Expression of HLA-A2 antigen in humanmelanoma cell lines and its role in T-cell recognition. Cancer Res. 51,3164–3170 16 Panelli, M.C. et al. (2000) Identification of a tumor infiltrating lymphocyte recognizing MAGE-12 in a melanoma metastasis withdecreased expression of melanoma differentiation antigens.  J. Immunol. 164, 4382–4392 17 Panelli, M.C. et al. (2000) Expansion of tumor/T cell pairs from fineneedle aspirates (FNA) of melanoma metastases.  J. Immunol. 164,495–504 18 Kammula, U.S. et al. (1999) Functional analysis of antigen-specific Tlymphocytes by serial measurement of gene expression in peripheralblood mononuclear cells and tumor specimens.  J. Immunol . 163,6867–6879 19 Hieter, P. and Boguski, M. (1997) Functional genomics: it’s all how youread it. Science 278, 601–602 20 Schena, M. et al. (1995) Quantitative monitoring of gene expressionpatterns with a complementary DNA microarray. Science 270, 467–470 21 Lander, E.S. (1996) The new genomics: global view of biology. Science 274, 536–539 22 Derisi, J. et al. (1996) Use of cDNA microarray to analyse geneexpression patterns in human cancer. Nat. Genet. 14, 457–460 23 Perou, C.M. et al. (1999) Distinctive gene expression patterns in humanmammary epithelial cells and breast cancers. Proc. Natl. Acad. Sci. U. S. A. 96, 9212–9217 24 Bittner, M. et al. (2000) Molecular classification of cutaneous malignantmelanoma by gene expression: shifting from a continuous spectrumto distinct biologic entities. Nature 406, 536–540 25 Alizadeh, A.A. et al. (2000) Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403, 467–578 26 Eisen, M.B. et al . (1998) Cluster analysis and display of genome-wideexpression patterns. Proc. Natl. Acad. Sci. U. S. A. 95, 14863–14868 27 Coller, H.A. et al. (2000) Expression analysis with oligonucleotidemicroarrays reveals that MYC regulates genes involved in growth, cellcycle, signaling and adhesion. Proc. Natl. Acad. Sci. U. S. A. 97, 3260–3265 28 Brown, P.O. and Botstein, D. (1999) Exploring the new world of thegenome with DNA microarrays. Nat. Genet. 21, 33–37 29 Duggan, D.J. et al . (1999) Expression profiling using cDNAmicroarrays. Nat. Genet. 21, 10–14 30 Wang, E. et al. (2000) High fidelity mRNA amplification for geneprofiling using cDNA microarrays. Nat. Biotech. 17, 457–459 VIEWPOINT IMMUNOLOGYTODAY Vol.21No.12623 DECEMBER2000 Fig. 3. (a) Scatter plot summarizing fluorescence intensity of fine-needle aspi-rate (FNA)-derived cDNA (3   g) hybridized to an OncoChip microarray 30 .The top plot shows the co-hybridization of Cy5-labeled cDNA (red fluor-escence) from the FNA of a melanoma metastasis (FNA79) with Cy3-labeledcDNA (green fluorescence) from a reference human myeloid leukemia cell line ML-1. The lower plot represents the co-hybridization of the FNA (FNA84) from another patient’s melanoma metastasis. The inserts show amplified sec-tions of the OncoChip with each spot representing a single gene (identical sec-tions are shown for the two FNA samples). In the plot, genes expressed co-ordinately by both melanoma tissue and ML-1 gather close to the linearregression curve. Genes differentially expressed (outliers) are those with fluor-escence biased towards red (expressed preferentially by the melanoma meta-stasis) or towards green (expressed preferentially by ML-1). As can be seen inthe inserts, a very similar pattern of green and red fluorescence is observed foreach individual gene in the two samples. (b) Hierarchical clustering diagramof the data presented in (a). The hierarchical clustering algorithm yields a tableof results in which the expression profiles for various genes are grouped to- gether according to similarities in the overall pattern of expression. The dataare presented graphically as a colored table in which rows represent individual genes and columns represent individual hybridizations (experiments). In thiscase, 467 genes out of 2008 genes present in the array were selected accordingto the following high-stringency parameters: ratio of fluorescence Cy3/Cy5 orCy5/Cy3 had to be equal or above three in at least one of the two experiments.The intensity of fluorescence had to be at least 100 units in both channels (greenand red) unless one of them was equal or above 1000, and the fluorescent spotsize had to be 100 pixels. The columns represent results from hybridizationwith two FNA materials. A very similar pattern between the two samples isnoted with genes differentially expressed by both metastases in comparisonwith ML-1. In spite of the overall similar pattern of expression, differences werenoted in the expression of some genes. It is possible that, among these different patterns of gene expression, the answer(s) to some of the puzzling questionsabout the algorithm governing tumor–host interactions is hiding. Immunology Today  F N A  8  4  F N A 7   9   467 genesThreefold, 1 experimentIntensity 100/1000 Single spot 100 y = 0.688x + 1R2 = 0.235y = 0.4593x + 1R2 = 0.0538Green intensity of ML-1Green intensity of ML-1    R  e   d   i  n   t  e  n  s   i   t  y  o   f   F   N   A   8   4   R  e   d   i  n   t  e  n  s   i   t  y  o   f   F   N   A   7   9 (a)(c)(b) 100000001000000100000100001000100000001000000100000100001000100010000100000100000010000000100010000100000100000010000000
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!