Transcriptional Proﬁling of Melanoma Sentinel Nodes Identify … · Microenvironment and Immunology Transcriptional Proﬁling of Melanoma Sentinel Nodes Identify Patients with

Microenvironment and Immunology

Transcriptional Profiling of Melanoma Sentinel NodesIdentify Patients with Poor Outcome and Reveal anAssociation of CD30þ T Lymphocytes with Progression

Viviana Vallacchi1, Elisabetta Vergani1, Chiara Camisaschi1, Paola Deho1, Antonello D. Cabras2,Marialuisa Sensi3,4, Loris De Cecco3, Niccol�o Bassani5, Federico Ambrogi5, Antonino Carbone2,Federica Crippa6, Barbara Vergani7, Paola Frati1, Flavio Arienti8, Roberto Patuzzo6, Antonello Villa7,Elia Biganzoli5, Silvana Canevari3, Mario Santinami6, Chiara Castelli1, Licia Rivoltini1, and Monica Rodolfo1

AbstractSentinel lymph nodes set the stance of the immune system to a localized tumor and are often the first site to be

colonized by neoplastic cells that metastasize. To investigate how the presence of neoplastic cells in sentinellymph nodes may trigger pathways associated with metastatic progression, we analyzed the transcriptionalprofiles of archival sentinel node biopsy specimens obtained from melanoma patients. Biopsies from positivenodes were selected for comparable tumor infiltration, presence or absence of further regional node metastases,and relapse at 5-year follow-up. Unsupervised analysis of gene expression profiles revealed immune response to bea major gene ontogeny represented. Among genes upregulated in patients with progressing disease, the TNFreceptor family member CD30/TNFRSF8 was confirmed in biopsy specimens from an independent group ofpatients. Immunohistochemical analysis revealed higher numbers of CD30þ lymphocytes in nodes fromprogressing patients compared with nonprogressing patients. Phenotypic profiling demonstrated that CD30þ

lymphocytes comprised a broad population of suppressive or exhausted immune cells, such as CD4þFoxp3þ orPD1þ subpopulations and CD4�CD8� T cells. CD30þ T lymphocytes were increased in peripheral bloodlymphocytes of melanoma patients at advanced disease stages. Our findings reinforce the concept that sentinelnodes act as pivotal sites for determining progression patterns, revealing that the presence of CD30þ lymphocytesat those sites associate positively with melanoma progression. Cancer Res; 74(1); 130–40. �2014 AACR.

IntroductionSentinel lymph node (SN) biopsy (SNB) represents the

standard of care for the staging and management of clinicallylocalized cutaneous melanomas. SNB detects metastasis inapproximately 20% of patients with primary melanoma T1B-

T4B. Patients with tumor-positive SNBs are treated withregional lymphadenectomy by completion lymph node dissec-tion (CLND), and approximately 20% showmetastatic involve-ment of additional nodes. Patients with lymph node metasta-ses are classified as stage III, and their estimated 5-year survivalrates are 78%, 59%, and 40% for stages IIIA, IIIB, and IIIC,respectively (1). However, this subclassification, which is basedon themost predictive independent prognostic factors (i.e., thenumber of tumor-bearing nodes, the relative tumor burden,and the thickness, ulceration, and mitosis of the primarytumor), is insufficient to provide an individualized patientassessment to determine prognosis and establish treatmentplans. Amore precise assessment of recurrence risk on a single-patient basis would help to guide adjuvant treatment options,and to reduce the rate of morbidity caused by overtreatmentand the associated costs.

The SN is the first lymph node to receive lymphatic drainagefrom a tumor and represents a relevant immunologic barrieragainst metastasis. Although providing the optimal environ-ment for the generation of a tumor-specific immune response,in melanoma patients, the SN is usually the first site ofmetastasis and is the node in the regional basin that is mostpowerfully influenced by tumor-derived factors and cells (2).In fact, pathologic and functional characterizations of theSN have revealed profound structural alterations, even in the

Authors' Affiliations: 1Immunotherapy Unit; 2Department of Pathology;3Functional Genomics Unit; 4Immunobiology of Human Tumors Unit;5Department of Medical Statistics and University of Milan, Medical Statis-tics Biometry and Bioinformatics Unit; 6Department of Surgery, MelanomaandSarcomaUnit, Fondazione IRCCS IstitutoNazionale dei Tumori, Milan;7Microscopy and Image Analysis, University of Milan Bicocca, Milan;8Department of Immunohematology and Transfusion Medicine, Fonda-zione IRCCS Istituto Nazionale dei Tumori, Milan, Italy

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

V. Vallacchi, E. Vergani, andC.Camisaschi contributed equally to thiswork.

Current address for A. Carbone: Division of Pathology, Centro di Rifer-imento Oncologico, Aviano, Italy.

Corresponding Author: Monica Rodolfo, Department of ExperimentalOncology and Molecular Medicine, Fondazione IRCCS Istituto Nazionaledei Tumori, via Venezian 1, 20133 Milan, Italy. Phone: 39-02-23903235;Fax: 39-02-23902154; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-13-1672

�2014 American Association for Cancer Research.

CancerResearch

Cancer Res; 74(1) January 1, 2014130

on May 1, 2021. © 2014 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

absence of tumor cells, including an increase in lymphaticvessels and amajor reduction in immune functions, facilitatingthe growth of metastatic tumors and promoting further dis-semination of the disease (3). Hence, the SN represents a keycomponent of the tumor microenvironment that potentiallypromotes immune tolerance. Increasing evidence supports theview that the triggering of immunosuppressive pathways atthe microenvironment level, which inactivate tumor-specificT-cell responses, is associated with systemic immune dysfunc-tion (4). These signs of local and systemic immunologicdysfunction, including the accumulation of myeloid-derivedsuppressor cells (5), tumor-specific CTLs with an anergicphenotype (6), regulatory T cells (Treg) and Th2 cells (7, 8),and dendritic cells (DC) with diminished functionality (9), aredefinitively detected in those patients with a poor clinicaloutcome (10, 11).To obtain insight into the molecular modifications occur-

ring in the SNs that are associated with disease progression, wedesigned an exploratory study to analyze the genome-widetranscriptional profiles of SNB samples derived from melano-ma patients with different outcomes. Here, we report that abroad downmodulation of immune-related genes, togetherwith the upregulated expression of lymphocyte markers defin-ing immunosuppressive and exhausted lymphocytes, identifiespatients with progressing melanoma.

Materials and MethodsClinical samplesSNB samples isolated from 42 patients were studied. A

diagram depicting the study design and the analyzed caserecords is shown in Fig. 1A and is detailed in SupplementaryMaterials and Methods. The histologic data obtained from thereviewof the diagnostic sections by a single pathologist and theclinical data for the studied cases are reported in Supplemen-tary Table S1. Lymph node lymphocytes were obtained frompatients recurring or not at the 5-year follow-up, whereas bloodlymphocytes were obtained from stage IIIC–IV patients andage- and gender-matched healthy donors (n ¼ 25). The studywas reviewed and approved by the Institutional Review Boardand the Independent Ethics Committee, and written informedconsent was obtained from the patients.

RNA microarraysGene expression profiles were generated using Illumina

HumanRef-8 WG-DASL v3.0 (Illumina). The Illumina BeadAr-ray Reader was used for scanning the arrays, and IlluminaBeadScan software was used for image acquisition and therecovery of primary data. The data were normalized usingBeadStudio software and the quantile method. The data weredeposited in the Gene Expression Omnibus repository (acces-sion number GSE39945; 12). Data analysis is detailed in Sup-plementary Materials and Methods.

ImmunohistochemistryImmunohistochemistry (IHC) staining was performed on

consecutive sections on a semiautomated platform by stan-dard diagnosticmethods and examined in a blinded fashion for

semiquantitative estimation (see Supplementary Materialsand Methods).

Cytofluorimetric analysesThe fluorescence intensity was measured using a Live/Dead

Fixable Dead Cell Stain Kit (Invitrogen) by a Gallios flowcytometer and analyzed with Kaluza software (Beckman Coul-ter). The monoclonal antibodies (mAb) used are indicated inthe Supplementary Materials and Methods.

Statistical analysisThe association between the CD30 IHC staining score and

progressionwas evaluated using the conditional odds ratio andits exact 95% confidence interval (Fisher test R function).Statistical analyses were performed by a Student t test orANOVA using GraphPad Prism v5.0.

ResultsGene expression profiles distinguish tumor-positiveSNBs isolated from patients with or without diseaseprogression

We performed microarrays to obtain gene expression pro-files, with the goal of exploring the molecular modificationsoccurring in SNBs isolated from patients with progressingdisease. We selected two groups of cases with tumor-positiveSNBs representing the extremes of the survey (see MaterialsandMethods): (i) patients with further metastatic nodes at thetime of CLND (stages IIIB–C) and disease recurrence within 5years of follow-up (group PP, positive SNB-positive CLND/progressing) and (ii) patients with tumor-positive SNBs whowere melanoma negative at the time of CLND (stages IIIA–B)and did not relapse within 5 years (group PN, positive SNB-negative CLND/nonprogressing). In addition, a group of SNB-negative patients who did not have disease recurrence at 5years (group N, negative SNB/nonprogressing) was also select-ed to compare the tumor-positive and tumor-negative SNBs(Fig. 1A). After reviewing the pathology records and assessingthe quality of the extracted RNA, 24 SNBs were selected for thegene expression profiling study.

Illumina WG-DASL-V3 microarrays were used for sampleprofiling. The data were normalized using quantile normali-zation with BeadStudio software and filtered by excludingprobes whose detection P values were larger than 0.05 in morethan 95% of the samples, thereby resulting in 23,468 probesmapping to 17,811 single genes. Cluster analysis using theaffinity propagation method indicated that the SNB sampleswere clustered into two groups. The first cluster consisted of 16samples and included eight N and eight tumor-positive SNBsamples, seven of which were from the PN nonprogressing set,and one of which was from the PP progressing group. Thesecond cluster comprised tumor-positive SNB samples, ofwhich seven were PP and one was PN. Therefore, the PPsamples from patients with progressing metastatic diseasedisplayed gene expression profiles that were significantlydifferent from those of all other SNB samples from patientswith nonprogressing disease, independent of the presence orabsence of tumormetastases. A three-dimensional scatter plot

Progression-Related Expression Profiles of Sentinel Nodes

www.aacrjournals.org Cancer Res; 74(1) January 1, 2014 131



PrimaryA

B

C E

DTumorSurgery

Sentinel CompletionNode Lymph node 5-year follow-up

SNB-NNED 91%

NED 86%

NED 43%

AWD/DOD

9%

SNB-negative

80%

SNB-positive

20%

CLND-

negative 70%

CLND-

positive 30%

AWD/DOD

14%

AWD/DOD

57%

SNB-PN

N+PN PP

SNB-PP

Biopsy Dissection

NPNPP

–1

50

–1

00

–5

00

50

10

0

–100–100

–50

–50

0

0

50

50

100

100

150

150

200

PC1

PC

3 PC

2

N vs. PN PN vs. PP

PP vs. N+PN N vs. PN+PP

Figure 1. Gene expression profiling of the SNBsamples. A, study sample design, frequencies, and outcomes at the 5-year follow-up of 752melanoma patientsundergoing SNBs in 2001–2004. The groups selected for the study are indicated in yellow. NED, no evidence of disease; AWD, alive with disease; DOD,dead of disease. B, unsupervised clustering, shown as principal component analysis, displaying the spatial separation of PP from the other samples.C, class comparison analysis displayed as volcanoplots. The comparedSNBgroups are indicated. D, heatmapof the genes thatwere differentially expressedbetween the PP and PNþN samples. E, heat map of the genes that were differentially expressed between the PP and PN samples. Red, upregulation;blue, downregulation. N, tumor-negative SNBs from cases with nonrecurring disease at the 5-year follow-up (n ¼ 8); PN, tumor-positive SNBsfrom cases with tumor-negative nodes at the time of CLND and NED at the 5-year follow-up (n¼ 8); PP, tumor-positive SNBs from cases with tumor-positivenodes at the time of CLND and relapse by the 5-year follow-up (n ¼ 8).

Vallacchi et al.

Cancer Res; 74(1) January 1, 2014 Cancer Research132



illustrating the characteristics of the samples based on the firstthree principal components, as generated by principal com-ponent analysis, further illustrates the spatial separation of thePP samples from the other samples, resulting in the clusteringof the transcript expression (Fig. 1B).Because the tumor-positive SNB samples were separated

into two clusters, the tumor burden of the SNBs was carefullyconsidered. The number, size, and localization of the micro-metastases in the nodes, as identified by IHC staining for themelanoma markers S-100, HMB-45, and Mart-1, were hetero-geneous in the two groups. When the tumor burden wasquantified on the basis of histopathology by ranking met-astatic deposits according to the Rotterdam classificationcriteria (13), each class was represented in both the PP andPN groups (Supplementary Table S1). To evaluate whetherdiffering tumor loads could explain the variation in expres-sion profiles between the PP and PN samples at the molec-ular level, the expression levels of the genes that encode themelanoma diagnostic markers S-100, Mart-1, HMB-45/PMEL, MITF, and tyrosinase (S100B, MLANA, SILV, MITF,and TYR) were evaluated in the array data. These melanoma-associated genes were expressed at similar levels in the PNand PP samples, thereby indicating that the different expres-sion profiles were unlikely to be a reflection of varying tumorburdens. Eleven other genes involved in melanin biosynthe-sis or highly expressed in melanoma (ASIP, DCT, GPR143,MC1R, MIA, PRAME, RAB27A, SLC24A5, SLC45A2, TRPM, andTYRP1) did not significantly differ between the PN and PPsamples (Supplementary Fig. S1).Class comparison analysis of the gene expression data is

represented as volcano plots in Fig. 1C, which demonstratesthat a large number (n ¼ 3,078) of genes were differentiallyexpressed between the PP samples and all other samples (Fig.1D), that a small number (n ¼ 58) of genes were differentiallyexpressed between the N samples and all of the tumor-positiveSNB samples, and that a smaller number of genes (n¼ 18) weredifferentially expressed between the N and PN samples.In contrast, a class comparison analysis of the PN and PPtumor-positive SNB samples yielded a manageable gene listconsisting of 337 probe sets identifying 333 genes, which werenearly equally distributed as up- and downregulated genes, asdepicted in the heat-map representation (Fig. 1E and Supple-mentary Table S2). A summary scheme of the number of genesdifferentially expressed among the different groups is reportedin Supplementary Fig. S2.Taken together, these results indicate that significant varia-

tions in gene expression profiles were detected in the tumor-positive SNB samples with a diverse disease course, therebyimplying a remarkable influence of tumor aggressiveness ormicroenvironment-modulating effects once melanoma cellshave reached the SN. Therefore, we next focused our study onthe transcripts that distinguish the PP samples from the PNsamples.

Differential transcriptional patterns in tumor-positiveSNB samples involve immune response genesTo gain insight into the biologic processes influencing the

different gene expression patterns, we used two approaches.

First, coordinated expression changes of genes in the PP andPN samples were assessed at the pathway level by Gene SetEnrichment Analysis (GSEA), a statistical functional enrich-ment analysis. We examined the enrichment of the C2

-curated catalog of functional gene sets (minimum gene setsize of 18) in the rank-ordered gene list obtained by GSEAusing the quantile-normalized data set. The heat map of thetop 100 GSEA-ranked, significant individual genes that dis-tinguish the PP and PN samples is shown in Fig. 2A.Interestingly, compared with PN, 356 gene sets were down-regulated and none were upregulated in PP [false discoveryrate (FDR) < 0.25, P < 0.05]. Gene signatures biased towardthe regulation of immune responses and gene expressionwere associated with the most representative gene setsshowing downregulation in the PP samples, accounting for27% and 12%, respectively, of all enriched gene sets (Sup-plementary Table S3). Among the most significant immunegene sets downregulated in PP, the enrichment plots andheat maps of core enrichment genes for genes associatedwith antigen processing and presentation (PELLICC-IOTTA_HDAC_IN_ANTIGEN_PRESENTATION_UP), genesassociated with chemokines, T- and B-cell receptor signaling(BIOCARTA_ CXCR4_PATHWAY), and genes dampened byFoxp3 in Tregs (MARSON_FOXP3_TARGETS_DN) are dis-played in Fig. 2B. Genes overlapping in the 16 gene setsassociated with immune response regulation were identifiedusing the leading-edge analysis tool and belonged tothe KEGG categories "KEGG_T_CELL_RECEPTOR_SIGNA-LING_PATHWAY" and "KEGG_ B_CELL_RECEPTOR_ SIG-NALING_PATHWAY," and mapped to the REACTOMEterm "REACTOME_SIGNALING_BY_THE_B_CELL_RECEPT-OR_BCR" (P < 0.00E). GSEA, therefore, revealed a coordinateddownregulation of components of adaptive immunity poten-tially associated with immunosuppression regulated by Foxp3signaling, in the PP samples compared with the PN SNBs.

In the second approach, the molecular networks involvingthe 333 genes that were differentially expressed between thetwo conditions were analyzed using Ingenuity PathwayAnalysis (IPA) software. The top IPA network was associatedwith "inflammatory response, cell death and cell-mediatedimmune response" functions (Fig. 2C). In accordance withthe GSEA results, most of the genes in the immune-responsecategory were downregulated in the PP samples comparedwith the PN samples, thereby indicating that genes associ-ated with immune modulation principally display reducedexpression in SNB samples from patients with progressingdisease. Other network functions displaying enrichmentwere the "cell growth and proliferation," "cell death, post-translational modification, protein folding," and "carbohy-drate metabolism, molecular transport, small molecule bio-chemistry" networks, which all displayed a pattern ofreduced gene expression in PP SNBs. Taken together, theresults of the combined GSEA and IPA analyses demonstratethat in the presence of comparable tumor infiltration, in PPSNBs, an impaired microenvironment was detected, withdysregulation of the genes that are involved in severalaspects of the cellular immune response, including cellsurvival, proliferation, and metabolism.





Validation studies confirm differential gene expressionpatterns in additional SNB samples

Validation of the gene-profiling results was performed bytesting the expression levels of selected genes in the PP and PNSNB samples by quantitative real-time (qRT) PCR. Genesselected from the 'Inflammation and Immune Response' listdisplayed expression patterns that confirmed the results of themicroarray profiling both in the training and in independentsamples (Supplementary Table S4 and Fig. S3). Among theimmune-response genes, TNFRSF8 and BSG attracted ourinterest because these genes were upregulated in the PPsamples (Fig. 3A) and encode the activation markers CD30and CD147, which have been associated with regulatory T cells(14, 15), Th2-like response, and chronic T-cell stimulation (16,17), although these genes are also detected in other immunecell subsets (18). In addition, TNFRSF8 upregulation wasassociated with several downregulated genes in the "inflam-matory response, cell death and cell-mediated immuneresponse" IPA pathway network shown in Fig. 2C, suggestinga potential role in the major signaling pathway identifiedby IPA.

To investigate the presence and role of lymphocytes expres-sing CD30 in SNBs, we first examined CD30 expression by IHCin PP and PN sections. The staining of sparse lymphocyteslocalized to the paracortical CD3-positive region of the nodeswas observed in both samples (Fig. 3B), whereas themetastaticmelanoma cells were not stained, in agreement with thepublished data (19). A total of 26 SNB samples were thenanalyzed by assigning the CD30þ cells a score ranging from 0 to3. The results demonstrated that most of the PP samplesscored 2 to 3, whereas most of the PN samples had a scoreof 0 to 1, indicating that higher numbers of CD30þ cells werepresent in the PP samples than in the PN SNBs (SupplementaryTable S5). To evaluate whether the CD30þ lymphocytes coex-pressed CD147 and additional immunoregulatory markers,such as Foxp3, a marker of activated Tregs, or PD1, a markerof anergic lymphocytes, the sections were double-stained withspecific mAbs. Although these markers were represented inseveral areas of the nodes, subsets of CD30þ cells were shownto be positive, indicating that the CD30þ population encom-passed different lymphocyte subsets sharing immunoregula-tory and dysfunctional traits (Fig. 3C).

PELLICCIOTTA_HDAC-IN-ANTIGEN-PRESENTATION-UP

BIOCARTA_CXCR4_PATHWAY

MARSON_FOXP3_TARGETS_DN

En

rich

me

nt sco

re (

ES

)E

nri

ch

me

nt sco

re (

ES

)E

nri

ch

me

nt

sco

re (

ES

)

A B C

ES

NES

–0.51

–1.80

FDR q-val 0.1

ES

NES

–0.66

–1.78

FDR q-val 0.1

ES

NES

–0.57

–1.76

FDR q-val 0.1

Figure 2. Bioinformatics analysis of the genes differentially expressed between the PP and the PN tumor-positive SNB samples. A, heat map of the top100 ranked genes distinguishing the PN and PP samples based on GSEA. B, profiles of the GSEA running enrichment score (ES), positions of thegene set members in the rank-ordered list, and heat maps and lists of leading-edge genes of three representative pathways among the top gene setsdownregulated in the PP group (Supplementary Table S3). In the heat maps, the gray columns represent the PP samples, whereas the yellow columnsrepresent the PN samples. C, the main gene network "inflammatory response, cell death, cell-mediated immune response" was identified by the IPA pooleddata analysis; this network includes the TNFRSF8 gene.

Vallacchi et al.




CD30þ lymphocytes display features of tolerogenic andexhausted T cells in lymph nodes and increase inperipheral blood of patients with progressing diseaseThe CD30 receptor is expressed by activated lymphocytes,

whether healthy or neoplastic, and is upregulated duringchronic inflammation and autoimmunity (14, 20), but a poten-tial function as a tolerogenic marker in cancer immunity hasnot been reported. A detailed characterization of CD30þ

lymphocytes was performed by fluorescence-activated cellsorting (FACS) analysis of cell suspensions obtained fromregional nodes at lymphadenectomy. Interestingly, a higherfrequency of CD30þ lymphocytes was observed in tumor-invaded nodes compared with tumor-free nodes (Fig. 4A).CD30þ or CD30� lymphocytes were then sorted fromtumor-invaded lymph nodes and functionally analyzed ex vivofor proliferative capacity and cytokine secretion upon T-cellreceptor triggering by stimulation with anti-CD3 and anti-

CD28 antibodies. As reported in Fig. 4B and C, the CD30þ cellsdisplayed a lower proliferative response compared with theirCD30� counterparts. Moreover, CD30þ lymphocytes displayeda limited production of Th1 and Th2 cytokines. Thus, thesedata strongly suggest that the CD30þ T cells in the tumor-invaded nodes include an exhausted or anergic cell populationwith poor effector function.

To fully explore the nature of cells that express CD30 inthe nodes of progressing melanoma patients, lymphocytesobtained from tumor-positive nodes were analyzed by multi-parametric flow cytometry. Staining for CD30 was performedin combination with staining for lineagemarkers specific for T,B, and NK cells and the differentiation/activation markersCD25, CD147, PD1, and Foxp3. The gating strategy and thedistribution of the CD30þ cells are illustrated in Fig. 5 A and B,respectively. Approximately 13% of the whole lymph nodederived lymphocytes expressed CD30 (panel I). The CD30þ

TrainingA

B C

TNFRSF8

PP PN PP PN PP PN PP PN

BSG

TrainingValidation Validation

3.1¥10-02

log

2 2

-DC

t

log

2 2

-DC

t

3.1¥10-02

1.0¥1000 0.125

0.063

0.031

0.016

0.008

0.004

9.8¥10-04

3.1¥10-05

9.5¥10-07

9.8¥10-04

9.8¥10-04

3.1¥10-05

9.5¥10-07

3.0¥10-08

3.1¥10-05

9.5¥10-07

3.0¥10-08

SNB-PP44 CD30

SNB-PN11 CD30

CD30-CD147

CD30-FOXP3

CD30-PD1

CD20 CD3

CD20 CD3

Figure 3. Expression of TNFRSF8/CD30 in PP and PN SNBs asassessed by qRT-PCR and IHCstaining. A, comparative geneexpressionanalysis byqRT-PCRofthe TNFRSF8/CD30 and BSG/CD147 genes, showingupregulation in the PP samplescompared with the PN samples(P < 0.05). The graphs display theresults obtained with the trainingand validation samples. Therelative quantification for eachsample was determined using2�DCt, andRPL13Awas used as aninternal reference. B, staining withanti-CD30, anti-CD20, and anti-CD3 antibodies is shown for tworepresentative samples. Thearrows indicate the CD30þ cells.CD30-immunostainedlymphocytes were detected in theparacortical CD3-positive area ofthe nodes (magnification, �20).C. double staining for CD30 andCD147, Foxp3, and PD1 in arepresentative PP sample showingcells expressing one or bothmarkers (magnification, �100).Confocal images of sectionsstained for CD30 in green and forCD147 or PD1 in red are shown(merge). In theCD30-Foxp3 doublestaining, CD30þ cells are indicatedby red, and nuclear staining forFoxp3 is indicated by brown.





lymphocytes included CD19þ B cells (panels II–III, 35% of theCD30þ-gated population) and a population of CD4/CD8double-negative (DN) CD3þ T cells. The latter population wasfurther characterized as CD56�CD45RAþCD25�PD1�Foxp3�

T cells (panel III, 24% of the CD30þ cells), thereby displayinga phenotype compatible with that of tolerogenic DN Tcells (21, 22). Approximately one third of CD30þ cells wereCD4þ T lymphocytes (panel II), including Foxp3þ

PD1�CD25þCD45RA� Tregs, in line with the IHC dataobtained for the SNBs (panel IV, 9.7%). Within the Foxp3�

CD4þ T conventional population, most displayed the pheno-type of exhausted/anergic T cells (23), such as the loss of theactivation marker CD25 and high expression of PD1 (panel IV,13.2%), whereas cells negative for PD1 (panel IV, 5.9%) wereequally composed of naive and antigen-experienced T cells.BothTregs and conventional T cells expressed high levels of theactivation marker CD147 (data not shown). Only a minority ofthe CD30þ cells were CD8þ cells (panel II, 5%). Half of thesecells were Foxp3�CD25�CD45RA� and positive for PD1, thusdisplaying a phenotype compatible with an exhausted/anergicfunctional state (panel V, 2.6%). The remaining cells, negativefor PD1 expression (panel V, 1.9%), were mainly in the na€�vecompartment.

The enhanced expression of CD30 on lymphocytes withsuppressive/regulatory functions in tumor invaded lymphnodes of patients with worse prognosis suggests a possibledirect role of tumor cells in CD30 upregulation. In vitroexperiments have been conducted by coculturing healthydonor peripheral blood mononuclear cells (PBMC) with mel-anoma cells. A significant CD30 upregulation was detected inthe CD3þCD4�CD8� DN cells and in Tregs gated in PBMCs asearly as after 48 and 96 hours of coculture withmelanoma cells(data not shown). These in vitro data are in agreement with thefindings reported in the Figs. 4A and 5. To determine whetherCD30þ cells could also be detected in the peripheral blood,PBMCs obtained from patients with regional or systemicdisease (stages IIIC–IV) were analyzed. CD30þ lymphocytenumbers significantly increased in melanoma patients com-pared with age- and gender-matched healthy donors, thusindicating a systemic accumulation of CD30þ lymphocytes aswell (Fig. 6). The increase in CD30þ cells was primarilyobserved in the CD3þ subpopulation and the CD4þ subsetand included both activated Tregs and conventional T cells,whereas for CD8þ-, CD3þCD4�CD8�-, CD3�CD4�CD8�-, andPD1-expressing subpopulations, the increase did not reachstatistical significance.

In sum, the phenotypic characterization depicted hereclearly shows that CD30 might represent a common markerfor both regulatory and anergic lymphocytes, identifying dys-functional immune responses both in tumor-neighboringlymph nodes and in the peripheral circulation of melanomapatients.

DiscussionOur study demonstrates that SNBs isolated from patients

with progressing melanoma can be distinguished fromSNBs derived from patients with good prognoses on the

P = 0.059

P = 0.029

P = 0.0121P = 0.054

15

10

5

0

A

B

C

TUM+LN TUM–LN

% o

f C

D30-p

ositiv

e c

ells

CD3/CD28 stimulation

CD30+CD30–

CD30– CD30+

lL-10 lL-6

lL-13 lL-5

CD30– CD30+

CD30– CD30+

CD30– CD30+ CD30– CD30+

CD30– CD30+

73% 40.6%

CFSE

TNF-α lL-2

100

80

60

40

20

0

100

80

60

40

20

0

Unstimulation

100 101 102 103 100 101 102 103

300

200

100

0

50

40

30

20

10

0

8

6

4

2

0

pg/m

Lpg/m

L

pg/m

L

pg/m

L

pg/m

Lpg/m

L

3,000

2,000

1,000

0

2,000

1,500

1,000

500

0

150

100

50

0

Figure 4. Melanoma-invaded nodes are enriched in CD30þ cells,displaying a low proliferation rate and cytokine production. A, CD30þ

cells as detected by FACS analysis of lymphocytes obtained from lymphnodes. TUMþ LN, tumor-invaded nodes (n ¼ 6); TUM- LN, tumor-freenodes (n ¼ 5). B, CD30þ and CD30� fractions were FACS-sorted from amelanoma-invaded node. CD30þ and CD30� cells were labeled withCFSE, and their proliferation rates upon activation with anti-CD3/CD28microbeads were evaluated. The percentages of proliferating cells arereported in the histograms. The black line represents the stimulated cells,whereas thedottedgray line represents the cells thatwere not stimulated.The results from one of two tested patients are shown. C, thesupernatants of the stimulated cells were collected and analyzed for theircytokine content by FlowCytomix assay.

Vallacchi et al.




basis of defined transcriptional profiles. By contrast, fewdifferences were detected between the profiles of tumor-negative and tumor-positive SNBs derived from patientswith good outcomes, thereby suggesting that the merepresence of melanoma cell deposits is not sufficient to drive

local immune dysfunction, and that defined tumor aggres-siveness or immunologic features are required. Genes relat-ed to immune response comprised the major ontology of thegenes that were differentially expressed between the tumor-positive SNB samples with a good or poor prognosis, again

Figure 5. CD30 expression indifferent T cell subsets inmetastatic lymph nodes frommelanoma patients. A, CD30þ cellswere identified in the population ofvital lymphocytes (I), as describedin Materials and Methods. On thebasis of the expression patterns ofCD4 and CD8, three CD30þ

subsets were identified (II), whichare indicated in the dot plot asCD4�CD8� (III), CD4þCD8� (IV),and CD8þCD4� (V). Each subsetwas characterized for itsexpression of Foxp3, PD1, CD25,and CD45RA. The CD4�CD8�

subset (III) was furthercharacterized for CD3 and CD56/CD19 expression. The percentageof positive cells in the gatedpopulation is reported in each dotplot. Thedata froma representativepatient of three tested patientsare shown. B, schematicrepresentation of lymphocytesubpopulations expressing CD30.Sections in red indicate cellsubsets with acknowledgedtolerogenic or anergic phenotypes.





indicating that immune changes are not solely driven bylocal tumor invasion.

The gene expression profiles of melanoma-positive SNBsassociated with disease progression were analyzed to identifypotential novel markers of poor prognosis in stage III mela-noma patients. We identified TNFRSF8/CD30, which displayedupregulated expression in SNBs from patients with dissemi-nated disease and was correlated with the downregulation of acomplex network of immune-associated genes. The geneexpression data were confirmed at the protein level by IHC,which revealed a higher number of CD30þ T cells in SNBs fromprogressing patients compared with samples from nonpro-gressing patients. Interestingly, CD30 expression seemed toencompass a panel of T cells mostly exhibiting tolerogenic orexhausted features, thus representing a potential marker of animmune system that is 'permissive' to tumor dissemination inthe draining lymph nodes. In fact, CD30 was expressed in SNBsby cells that also stained positive for CD147, an upregulatedmarker according to mRNA profiles, and either nuclear Foxp3or PD1. This latter evidence suggests that the CD30þ T cellsincluded both Tregs and exhausted lymphocytes.

Regional nodes from a separate group of stage III melanomapatients were used for in-depth phenotypic and functionalcharacterization, which revealed a higher frequency of CD30þ

cells in the presence of tumor metastases compared withtumor-free nodes. Multiparametric flow cytometry confirmedthat approximately one third of the CD30þ T cells were CD4 Tlymphocytes, most of which coexpressed Foxp3 and CD25 asactivated Tregs, or PD1, hence representing exhausted T cells.Another one third of the CD30þ T cells included CD4/CD8 DNlymphocytes. DN T cells were recently shown to exert aregulatory/tolerogenic function (21, 24, 25) and to be enrichedin the peripheral blood of melanoma patients (26). Functionalstudies performed with immunosorted CD30þ lymphocytesfrom the nodes of stage III melanoma patients have demon-strated that these cells are endowed with low proliferativeability and a limited capacity to produce Th1 and Th2 cyto-kines. Last, approximately one third of the CD30þ lymphocyteswere identified as a subpopulation of CD19þ B cells, which alsodisplayed a low proliferative rate compared with their CD30�

counterparts (data not shown). Indeed, CD30/CD30L interac-tions have been reported to play an important role in humoralimmune responses (27). Interestingly, we observed that CD30þ

T lymphocyte numbers also increased in the peripheral bloodof patients with metastatic melanoma compared with PBMCsfrom healthy donors, further indicating a potential role of thisimmune cell subset in melanoma progression.

No information is available about the role of CD30 in cancer-related immune responses. First described in the Reed–Sten-berg cells of Hodgkin lymphoma, CD30 is expressed in normaland neoplastic T lymphocytes, B cells, NK cells, and in severalnon-lymphoid cells (28). CD30 engagement by its ligand CD30Lhas been reported to provide costimulatory signals to T cells,leading to the activation of NF-kB, cytokine production, andproliferative responses (16). In addition, CD30 signaling hasbeen found to limit the proliferative potential of autoreactiveCD8 effector T cells and protects against autoimmunity inmurinemodels (20). In a human setting, the expression of CD30

Figure 6. Increased CD30 in the blood lymphocytes of melanoma patients.Freshly thawed PBMCs obtained from melanoma patients (Me Pts) andfrom age- and gender-matched healthy donors (HD) were analyzed forCD30þ subpopulations by FACS. A gate for the CD3þ and CD3� cells wascreated, and among the CD3þ-gated cells, the CD4þ, CD8þ, and CD4�

CD8� regionswere identified.TheCD4þ regionwas further analyzed for theexpression of CD25, CD45RA, and Foxp3 to identify Tregs(CD4þFoxp3hiCD25hiCD45RA�) and conventional T cells (T conv, CD4þ

Foxp3loCD25loCD45RA�). Both CD4 and CD8 subsets were analyzed forthe expression of PD1 and CD25 to identify the exhausted subpopulation(CD25�PD1þ). Among the CD3�-gated cells, the CD4�CD8� populationwas identified.Samplesobtained from25subjectswere tested.Statisticallysignificant differences between patients and controls are indicated.

Vallacchi et al.




marks suppressive Foxp3þ Tregs (14), and the presence ofCD30þ T cells in synovial fluid has been correlated with amorefavorable course in patients with juvenile arthritis (29). Thesedata support the involvement of CD30 in the negative controlof immune responses directed against self-antigens, a settinglargely including tumor-specific T cells.In addition to CD30, other genes were upregulated in SNBs

from progressing patients: HMGA1, a transcriptional factorregulating the IFN, IL2, and IL4 genes and recently reported toplay a role in transcriptional silencing of the CD4/CD8 loci in theT-cell lineage (30); BSG, which encodes the CD4þFoxp3þ Treg-activationmarker CD147 (15), here resulting expressed by nodalCD30þ lymphocytes; and DEFA1, which encodes defensin a,which has been reported to attract immature DCs and sustainchronic inflammation (31). In addition to the upregulation ofdefined immune-related genes, GSEA demonstrated that SNBsfrom patients with poor outcomes displayed a prominentpattern of the reduced expression of genes associated withimmunologic responses, thus indicating transcriptional down-regulation associated with immune-function impairment.According to our analysis, the different expression profiles

detected in tumor-positive SNBs are unlikely to be a reflectionof a difference in tumor burden, as clearly indicated byhistopathology and, at the molecular level, by the comparableexpression levels of severalmelanoma-associated genes. None-theless, the melanoma genes in the PP and PN samples dis-played a different pattern in the array results, for example,PRAME, RAB27A, and TRPM1 showed higher expression in thePP samples than in the N samples. By contrast, DCT, SLC24A5,and MC1R expression were higher in the PN samples (Supple-mentary Fig. S1), suggesting that the metastatic melanomacells in the two SNB groups may have different molecularfeatures.This is the first study to analyze whole-genome gene expres-

sion profiling in SNB samples. Broad defects affecting differentimmunologic components have been extensively documentedin SNs with respect to nontumor-draining nodes, including adecreased frequency of CD8þ effector T cells, a reducedpresence of stimulatory CD86þ and CD11cþ DCs, andincreased Foxp3 density and Th2 cytokine levels (7, 11, 32–37). Our findings partially contradict such observationsbecause we found that immune dysfunction is detected intumor-positive SNBs in association with disease progression.Interestingly, limited differential expression of genes wasdetected between tumor-negative and tumor-positive SNBs(n ¼ 58), and even fewer (n ¼ 18) genes were detected when

negative SNBs were compared with positive SNBs frompatients with a favorable clinical course (Supplementary Fig.S2). By contrast, patients with tumor-positive SNBs and pro-gressing disease displayed a relatively different gene expressionprofile compared with nonprogressing patients (n ¼ 333) andpatients with tumor-negative SNBs (n ¼ 3708).

Although in this study a small number of samples was testedin the discovery and validation experiments (n ¼ 24 and 18,respectively), our results encourage further studies of tran-scriptional profiles of formalin-fixed paraffin-embedded SNBsas a potential informative tool in the clinical setting, forpersonalized patient treatment. In addition, the strong corre-lation that we observed between the presence of exhausted/regulatory CD30þ T cells in SNBs and disease progressionsuggests a potential role for this marker in the prognosticevaluation and therapeutic targeting of melanoma.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: V. Vallacchi, E. Vergani, C. Camisaschi, M. Sensi,R. Patuzzo, S. Canevari, C. Castelli, L. Rivoltini, M. RodolfoDevelopment of methodology: C. Camisaschi, P. Deho, M. Sensi, F. Ambrogi,A. Carbone, B. Vergani, A. Villa, S. Canevari, C. CastelliAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): V. Vallacchi, E. Vergani, C. Camisaschi, P. Deho,M. Sensi, L.D. Cecco, F. Ambrogi, A. Carbone, F. Crippa, B. Vergani, P. Frati,R. Patuzzo, A. Villa, S. Canevari, M. SantinamiAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): V. Vallacchi, E. Vergani, P. Deho, A.D. Cabras,M. Sensi, L.D. Cecco, N. Bassani, F. Ambrogi, E. Biganzoli, S. Canevari, M. RodolfoWriting, review, and/or revision of themanuscript: V. Vallacchi, E. Vergani,C. Camisaschi, P. Deho, M. Sensi, N. Bassani, F. Ambrogi, B. Vergani, A. Villa,E. Biganzoli, S. Canevari, C. Castelli, L. Rivoltini, M. RodolfoAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): V. Vallacchi, E. Vergani, C. Camisaschi,B. Vergani, P. Frati, F. Arienti, R. Patuzzo, M. RodolfoStudy supervision: M. Rodolfo

AcknowledgmentsThe authors thank Valeria Beretta and Maria Daniotti for essential

contributions.

Grant SupportThis work was supported by the Italian Association for Cancer Research

(AIRC) grant IG-9030 and the Ministry of Health (RF30504).The costs of publication of this article were defrayed in part by the payment of

page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received June 13, 2013; revised September 27, 2013; accepted October 21, 2013;published online January 6, 2014.

References1. BalchCM,Gershenwald JE, SoongSJ, Thompson JF, AtkinsMB,Byrd

DR, et al. Final version of 2009 AJCC melanoma staging and classi-fication. J Clin Oncol 2009;27:6199–206.

2. Cochran AJ, Huang RR, Lee J, Itakura E, LeongSP, Essner R. Tumour-induced immune modulation of sentinel lymph nodes. Nat Rev Immu-nol 2006;6:659–70.

3. Takeuchi H, Kitajima M, Kitagawa Y. Sentinel lymph node as atarget of molecular diagnosis of lymphatic micrometastasis andlocal immunoresponse to malignant cells. Cancer Sci 2008;99:441–50.

4. ZuckermanNS, Yu H, Simons DL, Bhattacharya N, Carcamo-CavazosV, Yan N, et al. Altered local and systemic immune profiles underlielymph node metastasis in breast cancer patients. Int J Cancer2013;132:2537–47.

5. Filipazzi P, Huber V, Rivoltini L. Phenotype, function and clinicalimplications of myeloid-derived suppressor cells in cancer patients.Cancer Immunol Immunother 2012;61:255–63.

6. Baitsch L, Baumgaertner P, Devevre E, Raghav SK, Legat A, Barba L,et al. Exhaustion of tumor-specific CD8þ T cells in metastases frommelanoma patients. J Clin Invest 2011;121:2350–60.





7. Viguier M, Lemaitre F, Verola O, Cho MS, Gorochov G, Dubertret L,et al. Foxp3 expressing CD4þCD25high regulatory T cells are over-represented in human metastatic melanoma lymph nodes and inhibitthe function of infiltrating T cells. J Immunol 2004;173:1444–53.

8. Nevala WK, Vachon CM, Leontovich AA, Scott CG, Thompson MA,Markovic SN, et al. Evidence of systemic Th2-driven chronic inflam-mation in patients with metastatic melanoma. Clin Cancer Res2009;15:1931–9.

9. McCarter MD, Baumgartner J, Escobar GA, Richter D, Lewis K,Robinson W, et al. Immunosuppressive dendritic and regulatory Tcells are upregulated in melanoma patients. Ann Surg Oncol 2007;14:2854–60.

10. Jensen TO, Schmidt H, Møller HJ, Høyer M, Maniecki MB, Sjoegren P,et al.Macrophagemarkers in serumand tumor have prognostic impactin American Joint Committee on Cancer stage I/II melanoma. J ClinOncol 2009;27:3330–7.

11. Mohos A, Sebesty�en T, Liszkay G, Pl�ot�ar V, Horv�ath S, Gaudi I, et al.Immune cell profile of sentinel lymph nodes in patients with malignantmelanoma – FOXP3þ cell density in cases with positive sentinel nodestatus is associated with unfavorable clinical outcome. J Transl Med2013;11:43.

12. Gene expression profiles of tumor-positive sentinel lymph node biop-sies from cutaneous melanoma patients. Available from: http://www.ncbi.nlm.nih.gov/geo/. Accessed Aug 01, 2013.

13. van Akkooi AC, Nowecki ZI, Voit C, Schafer-Hesterberg G, Michej W,de Wilt JH, et al. Sentinel node tumor burden according to theRotterdam criteria is the most important prognostic factor for survivalin melanoma patients: a multicenter study in 388 patients with positivesentinel nodes. Ann Surg 2008;248:949–55.

14. de Kleer I, Vercoulen Y, Klein M, Meerding J, Albani S, van der Zee R,et al. CD30 discriminates heat shock protein 60-induced FOXP3þCD4þ T cells with a regulatory phenotype. J Immunol 2010;185:2071–9.

15. Solstad T, Bains SJ, Landskron J, Aandahl EM, Thiede B, Tasken K,et al. CD147 (Basigin/Emmprin) identifies FoxP3þCD45ROþCTLA4þ-activated human regulatory T cells. Blood 2011;118:5141–51.

16. Tarkowski M. Expression and a role of CD30 in regulation of T-cellactivity. Curr Opin Hematol 2003;10:267–71.

17. Pellegrini P, Berghella AM, Contasta I, Adorno D. CD30 antigen: not aphysiological marker for TH2 cells but an important costimulatormolecule in the regulation of the balance between TH1/TH2 response.Transpl Immunol 2003;12:49–61.

18. Pistol G, Matache C, Calugaru A, Stavaru C, Tanaseanu S, Ionescu R,et al. Roles of CD147 on T lymphocytes activation and MMP-9secretion in systemic lupus erythematosus. J Cell Mol Med 2007;11:339–48.

19. Polski JM, Janney CG. Ber-H2 (CD30) immunohistochemical stainingin malignant melanoma. Mod Pathol 1999;12:903–6.

20. HeathWR,Kurts C, Caminschi I, Carbone FR,Miller JF. CD30 preventsT-cell responses to non-lymphoid tissues. Immunol Rev 1999;169:23–9.

21. Ellis TM, Simms PE, Slivnick DJ, Jack HM, Fisher RI. CD30 is a signal-transducing molecule that defines a subset of human activatedCD45ROþ T cells. J Immunol 1993;151:2380–9.

22. Fischer K, Voelkl S, Heymann J, Przybylski GK, Mondal K, Laumer M,et al. Isolation and characterization of human antigen-specific TCRalpha betaþ CD4-CD8- double-negative regulatory T cells. Blood2005;105:2828–35.

23. Day CL, Kaufmann DE, Kiepiela P, Brown JA, Moodley ES, Reddy S,et al. PD-1 expression on HIV-specific T cells is associated with T-cell exhaustion and disease progression. Nature 2006;443:350–4.

24. Juvet SC, Zhang L. Double negative regulatory T cells in transplanta-tion and autoimmunity: recent progress and future directions. J MolCell Biol 2012;4:48–58.

25. Voelkl S, Gary R, Mackensen A. Characterization of the immunoreg-ulatory function of human TCR-alphabeta þCD4- CD8- double-neg-ative T cells. Eur J Immunol 2011;41:739–48.

26. Voelkl S, Moore TV, Rehli M, Nishimura MI, Mackensen A, Fischer K.Characterization of MHC class-I restricted TCRalphabetaþ CD4-CD8- double negative T cells recognizing the gp100 antigen from amelanoma patient after gp100 vaccination. Cancer Immunol Immun-other 2009;58:709–18.

27. Kennedy MK, Willis CR, Armitage RJ. Deciphering CD30 ligandbiology and its role in humoral immunity. Immunology 2006;118:143–52.

28. de Bruin PC, Gruss HJ, van der Valk P, Willemze R, Meijer CJ. CD30expression in normal and neoplastic lymphoid tissue: biologicalaspects and clinical implications. Leukemia 1995;9:1620–7.

29. de Kleer IM, Kamphuis SM, Rijkers GT, Scholtens L, Gordon G, DeJager W, et al. The spontaneous remission of juvenile idiopathicarthritis is characterized by CD30þ T cells directed to human heat-shock protein 60 capable of producing the regulatory cytokine inter-leukin-10. Arthritis Rheum 2003;48:2001–10.

30. Xi Y, Watanabe S, Hino Y, Sakamoto C, Nakatsu Y, Okada S, et al.Hmga1 is differentially expressed and mediates silencing of the CD4/CD8 loci in T cell lineages and leukemic cells. Cancer Sci 2012;103:439–47.

31. Lehrer RI, Lu W. Alpha-defensins in human innate immunity. ImmunolRev 2012;245:84–112.

32. Ma MW, Medicherla RC, Qian M, Vega-Saenz de Miera E, FriedmanEB, Berman RS, et al. Immune response in melanoma: an in-depthanalysis of the primary tumor and corresponding sentinel lymph node.Mod Pathol 2012;25:1000–10.

33. Mansfield AS, Holtan SG, Grotz TE, Allred JB, Jakub JW, Erickson LA,et al. Regional immunity in melanoma: immunosuppressive changesprecede nodal metastasis. Mod Pathol 2011;24:487–94.

34. Lee JH, Chen Y, Chan JL, Qian YW, Goydos JS. Molecular analysis ofmelanoma-induced sentinel lymph node immune dysfunction. CancerImmunol Immunother 2011;60:685–92.

35. Torisu-Itakura H, Lee JH, Scheri RP, Huynh Y, Ye X, Essner R, et al.Molecular characterization of inflammatory genes in sentinel andnonsentinel nodes in melanoma. Clin Cancer Res 2007;13:3125–32.

36. Botella-EstradaR, Dasi F, RamosD,Nagore E, HerreroMJ,Gimenez J,et al. Cytokine expression and dendritic cell density in melanomasentinel nodes. Melanoma Res 2005;15:99–106.

37. Grotz TE, Mansfield AS, Jakub JW, Markovic SN. Regional lymphaticimmunity in melanoma. Melanoma Res 2012;22:9–18.

Vallacchi et al.




2014;74:130-140. Cancer Res Viviana Vallacchi, Elisabetta Vergani, Chiara Camisaschi, et al. Lymphocytes with Progression

T+Patients with Poor Outcome and Reveal an Association of CD30Transcriptional Profiling of Melanoma Sentinel Nodes Identify

Updated version

http://cancerres.aacrjournals.org/content/74/1/130

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2014/01/09/74.1.130.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/74/1/130.full#ref-list-1

This article cites 36 articles, 9 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/74/1/130.full#related-urls

This article has been cited by 3 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/74/1/130To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/content/suppl/2014/01/09/74.1.130.DC1

http://cancerres.aacrjournals.org/content/74/1/130.full#ref-list-1

http://cancerres.aacrjournals.org/content/74/1/130.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

Transcriptional Proﬁling of Melanoma Sentinel Nodes Identify … · Microenvironment and Immunology Transcriptional Proﬁling of Melanoma Sentinel Nodes Identify Patients with