Gene Expression Profiling Using Nanostring Digital RNA Counting to Identify Potential Target Antigens for Melanoma Immunotherapy

Published OnlineFirst September 10, 2013; DOI: 10.1158/1078-0432.CCR-13-1253

Clinical Cancer Human Cancer Biology Research

Gene Expression Proﬁling using Nanostring Digital RNA Counting to Identify Potential Target Antigens for Melanoma Immunotherapy

Rachel E. Beard, Daniel Abate-Daga, Shannon F. Rosati, Zhili Zheng, John R. Wunderlich, Steven A. Rosenberg, and Richard A. Morgan

Abstract Purpose: The success of immunotherapy for the treatment of metastatic cancer is contingent on the identification of appropriate target antigens. Potential targets must be expressed on tumors but show restricted expression on normal tissues. To maximize patient eligibility, ideal target antigens should be expressed on a high percentage of tumors within a histology and, potentially, in multiple different malignancies. Design: A Nanostring probeset was designed containing 97 genes, 72 of which are considered potential candidate genes for immunotherapy. Five established melanoma cell lines, 59 resected metastatic melanoma tumors, and 31 normal tissue samples were profiled and analyzed using Nanostring technology. Results: Of the 72 potential target genes, 33 were overexpressed in more than 20% of studied melanoma tumor samples. Twenty of those genes were identified as differentially expressed between normal tissues and tumor samples by ANOVA analysis. Analysis of normal tissue gene expression identified seven genes with limited normal tissue expression that warrant further consideration as potential immunotherapy target antigens: CSAG2, MAGEA3, MAGEC2, IL13RA2, PRAME, CSPG4, and SOX10. These genes were highly overexpressed on a large percentage of the studied tumor samples, with expression in a limited number of normal tissue samples at much lower levels. Conclusion: The application of Nanostring RNA counting technology was used to directly quantitate the gene expression levels of multiple potential tumor antigens. Analysis of cell lines, 59 tumors, and normal tissues identified seven potential immunotherapy targets for the treatment of melanoma that could increase the number of patients potentially eligible for adoptive immunotherapy. Clin Cancer Res; 1–10. 2013 AACR.

Introduction and complete response rates of approximately 40% The development of successful immunotherapy for met- reported in trials treating patients with metastatic melano- astatic cancer requires the identification of appropriate ma (1). This strategy necessitates the acquisition of tumor target antigens. As immunotherapeutic strategies become specimens for generation of TIL and has primarily shown increasingly sophisticated and powerful, finding antigens success in treating melanoma. An alternative approach is that are overexpressed in malignancies but have restricted the infusion of lymphocytes that have been harvested from expression in normal tissue becomes challenging. To date, the patient and genetically engineered to recognize tumor- the most successful immunotherapeutic approach is the associated antigens (1, 2). adoptive cell transfer (ACT) of tumor-infiltrating lympho- Tumor antigen-reactive T-cell receptor (TCR) gene ther- cytes (TIL), with objective response rates of more than 70% apies have been used with success in multiple histologies; however, the number of patients that can be treated is somewhat limited as they must express a specific human leukocyte antigen (HLA; e.g. HLA-A0201; refs. 2, 3). The Authors' Affiliation: Surgery Branch, Center for Cancer Research, Nation- al Cancer Institute, Bethesda, Maryland restricted expression of certain target antigens in tumor cohorts can also limit a therapy’s potential use. For example, Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). effective TCR therapies have been reported targeting the cancer-testis antigen (CTA) NY-ESO-1; however, only Corresponding Author: Richard A. Morgan, Surgery Branch, National Cancer Institute, 10 Center Drive, Building 10-Hatfield CRC, Rm 3-5942, around 20% to 30% of melanomas express this antigen Bethesda, MD 20891-1202. Phone: 301-594-9629; Fax: 301-435-5167; (4, 5). Clinical trials using non–MHC-restricted chimeric E-mail: [email protected] antigen receptor (CAR) therapy can potentially expand the doi: 10.1158/1078-0432.CCR-13-1253 number of patients eligible for ACT if the target antigen is 2013 American Association for Cancer Research. expressed on the cell surface (2). Recently, CAR therapies

www.aacrjournals.org OF1

Beard et al.

Immunohistochemistry Translational Relevance All tumors samples were confirmed to be metastatic Accurate quantitation of RNA levels is an essential step melanoma at the time of harvest by pathological evalu- in the identification of potential tumor antigens. Nano- ation including immunohistochemistry. Immunohisto- string is a solution-based gene expression profiling tech- chemical staining was carried out for expression of nology that accurately counts individual RNA molecules the antigen NY-ESO-1 (encoded by gene CTAG1B)with in the small amounts of total RNA (250 ng or less) that the specific anti-NY-ESO-1 monoclonal antibody E978 are often obtained from biopsy samples. We used this (Invitrogen; ref. 5). Immunohistochemical scores were technology to study potential tumor antigen gene assigned for intensity of staining and percentage of expression in 59 metastatic melanoma samples and tumors cells that stained positive. identified seven genes as potential targets for adoptive immunotherapy. Nanostring analysis Using the Nanostring nCounter Analysis System (Nano- string Technologies), gene expression analysis was conducted for each sample as previously described using a custom-designed codeset containing 97 genes (8). Each have shown success in treating non-melanoma and non- reaction contained 250 ng of total RNA in a 5 mL aliquot, solid organ cancers, specifically the ability of the anti-CD19 plus reporter and capture probes, and 6 pairs of positive CAR to effect regression of advanced B-cell malignancies control and 8 pairs of negative control probes. Analysis and (6, 7). normalization of the raw Nanostring data was conducted The most effective way to maximize the number of using nSolver Analysis Software v1.1 (Nanostring Technol- patients potentially eligible for a therapy would be to ogies). Raw counts were normalized to internal levels of target an antigen expressed on a high percentage of 7 reference genes: CNOT2, GAPDH, HPRT1, PHGDH, tumors within a given histology. This study uses Nano- SUMO2, SYS1, and WDR45L. A background count level string technology to achieve gene expression profiling of was estimated using the average count of the 8 negative melanoma cell lines, metastatic melanoma tumors, and control probes in every reaction plus 2 SDs. normal human tissue samples to identify potential target antigens for immunotherapy. This robust technology uses Gene expression analysis unique digital color-coded barcodes that hybridize direct- Principal component analysis (PCA) and ANOVA anal- ly to specific nucleic acid targets and allow for detection ysis were conducted using the Partek Genomic Suite (Partek and quantitation of hundreds of transcripts in a single Incorporated). PCA was used to characterize samples on the reaction. Unlike microarray approaches, it facilitates the basis of their gene expression profiles. ANOVA analysis was direct measurement of mRNA expression levels for sub- used to identify differentially expressed genes (significant sequent gene expression analysis, and has been shown to P value < 0.05) and samples were clustered by hierarchic be highly reproducible and as sensitive as real-time PCR clustering. assays while still allowing for the measurement of multiple genes at one time (8). Flow cytometry and quantitative-PCR Flow cytometry (FACS) was conducted using conjugat- Materials and Methods ed monoclonal antibody (mAb) specific for human chon- Sample collection, RNA isolation, and cell lines droitin sulfate proteoglycan 4 (CSPG4) according to the Patients seen at the Surgery Branch, National Cancer manufacturers’ instructions (R&D Systems). Reverse tran- Institute (NCI; Bethesda, MD) for treatment of metastatic scription (RT) was conducted using High Capacity cDNA melanoma underwent surgical excision of metastatic Reverse Transcription Kit (Applied Biosystems). Quanti- lesions for harvest of TILs in protocols approved by the tative PCR was carried out with the TaqMan Fast Univer- Institutional Review Board and Food and Drug Adminis- sal PCR Master Mix (Applied Biosystems) by the use of a tration. Viable-appearing fragments of these tumors were 7500 Fast Real-Time PCR System (Applied Biosystems). freed from surrounding normal tissue, collected, and either Copy numbers were generated using a standard curve stored in RNAlater (Ambion) or flash-frozen and stored at from CSPG4 plasmid and results were normalized against 80 Celsius, until RNA isolation was conducted using a b-actin (ACTB). RNEasy Mini Kit (Qiagen). Analyzed tumor samples were collected between September 2007 and December 2012. Results RNA isolation was conducted in the same fashion for Samples and Nanostring probeset established human melanoma cell lines initiated at Memo- Gene expression profiling using Nanostring technology rial Sloan-Kettering (SKmel23) or the NCI Surgery Branch was conducted on RNA from five melanoma lines (all others). The lines were grown under standard condi- (mel1300, mel526, mel624.38, mel888, and SKmel23), tions in RPMI-1640 with 10% FBS medium at 37C, 59 resected metastatic melanoma tumor deposits, and 31 5%CO2. For normal tissues, commercially available RNA normal tissue samples. A total of 97 genes were included in samples were used (Agilent, Ambion, Biochain, Clontech). the probeset (Table 1). Immune-related genes were

OF2 Clin Cancer Res; 2013 Clinical Cancer Research

Nanostring Digital RNA Counting for Melanoma Immunotherapy

Table 1. Genes included in Nanostring probeset Table 1. Genes included in Nanostring probeset (Cont'd ) Melanoma-related genes AURKB NM_004217.2 Glioblastoma-related genes B4GALNT1 NM_001478.3 BCAN NM_198427.1 CSPG4 NM_001897.4 CHI3L2 NM_004000.2 DCT NM_001922.3 EGFRvIII NCI_Custom ERBB4 NM_001042599.1 FABP7 NM_001446.3 IGF2BP3 NM_006547.2 GRIN2A NM_000833.3 NRCAM NM_001037132.1 GRM3 NM_000840.2 PTPRZ1 NM_002851.2 MITF NM_000248.3 TNC NM_002160.1 MLANA NM_005511.1 PMEL NM_001200054.1 Other tumor-related genes CEACAM5 NM_004363.2 SOX10 NM_006941.3 ERBB2 NM_004448.2 ST3GAL5 NM_003896.2 FOLH1 NM_001014986.1 ST8SIA1 NM_003034.3 GPC3 NM_004484.2 TYR NM_000372.4 GTF2A1 NM_015859.2 TYRP1 NM_000550.2 KDR NM_002253.2 Cancer testis genes KIF20A NM_005733.2 CSAG2 NM_001080848.2 MSLN NM_013404.3 CT45A1 NM_001017417.1 MUC1 NM_001018017.1 CTAG1B NM_001327.2 PSCA NM_005672.3 CTAG2 NM_172377.3 STAG2 NM_001042749.1 CTCFL NM_080618.2 TG NM_003235.4 CXorf48 NM_001031705.2 TKTL1 NM_001145933.1 GAGE1 NM_001040663.2 WT1 NM_024426.3 GAGE4 NM_001474.1 Immune genes IL13RA2 NM_000640.2 B2M NM_004048.2 MAGEA1 NM_004988.4 CD160 NM_007053.2 MAGEA2 NM_005361.2 CD19 NM_001770.4 MAGEA3 NM_005362.3 CD247 NM_198053.1 MAGEA4 NM_001011548.1 CD274 NM_014143.2 MAGEA5 NM_021049.3 CD3E NM_000733.2 MAGEA8 NM_005364.4 CD4 NM_000616.4 MAGEA9 NM_005365.4 CD8A NM_001768.5 MAGEA10 NM_001011543.1 CD80 NM_005191.3 MAGEA11 NM_005366.4 LILRB2 NM_005874.1 MAGEA12 NM_001166386.1 CD86 NM_175862.3 MAGEB1 NM_002363.4 FOXP3 NM_001114377.1 MAGEB2 NM_002364.4 IL10 NM_000572.2 MAGEB3 NM_002365.4 MS4A1 NM_152866.2 MAGEB6 NM_173523.2 PDCD1 NM_005018.1 MAGEC1 NM_005462.4 PECAM1 NM_000442.3 MAGEC2 NM_016249.3 PTPRC NM_080921.2 POTEF NM_001099771.2 TGFB1 NM_000660.4 PRAME NM_006115.3 Reference genes SAGE1 NM_018666.2 CNOT2 NM_014515.4 SPANXN3 NM_001009609.2 GAPDH NM_002046.3 SPANXA1 NM_013453.2 HPRT1 NM_000194.1 SSX1 NM_005635.2 PHGDH NM_006623.2 SSX2 NM_003147.4 SUMO2 NM_006937.3 SSX3 NM_021014.2 SYS1 NM_033542.3 SYCP1 NM_003176.2 TSPY1 NM_003308.3 NOTE: Gene name and corresponding GenBank Accession Numbers for the 97 genes included in the probeset. (Continued on the following coloumn)

www.aacrjournals.org Clin Cancer Res; 2013 OF3

Beard et al.

included to investigate the immune characteristics of mel- resection had stages III or IV melanoma. They ranged in age anoma samples; however, these would not be considered as from 19 to 66 years (average 47 13 years) and 71% of potential immunotherapy target antigens. After elimination patients were male. of the control genes (n ¼ 7) and the immune genes (n ¼ 18), 72 candidate genes remained. The candidate genes were Validation of Nanostring data grouped as follows for organizational purposes: melanoma- To compare the sensitivity of Nanostring to other meth- related, cancer testis antigen (CTA), glioblastoma-related, ods of gene expression analysis, we used quantitative real- and other tumor-related genes. The sites of resection for the time PCR (RT-PCR) and fluorescence antibody staining. tumor deposits were primarily subcutaneous tissue (46%), Nanostring counts for the high-molecular weight melano- lymph node (24%), lung (15%), and liver (8%). One tumor ma-associated antigen (HMW-MAA) gene CSPG4 for the 5 was resected from each of the following sites: adrenal, pelvic melanoma lines showed high correlation with CSPG4 copy mass, retroperitoneum, and small bowel. Before detailed numbers generated by RT-PCR (R2 ¼ 0.99435; Fig. 1A) and analysis of the data, we used PCA to determine whether mean fluorescence intensity (MFI) values generated by flow anatomic location of the tumor influenced gene expression cytometric analysis (R2 ¼ 0.97418; Fig. 1B). Of the 59 profiling. PCA did not show differentiation of gene expres- tumors samples, 41 had immunohistochemical staining sion profiles based on the site of tumor harvest (Supple- conducted for NY-ESO-1 at the time of resection. Despite mentary Fig. S1). All 59 patients who underwent tumor the fact that immunohistochemical analysis was conducted

A B 10,000 700 9,000 R 2 = 0.99435 600 R 2 = 0.97418 8,000 7,000 500 6,000

ACTB 400 6 5,000 MFI 300 4,000

CSPG4/10 3,000 200 2,000 100 1,000 0 0 0 200 400 600 800 1,000 1,200 1,400 0 200 400 600 800 1,000 1,200 1,400 Nanostring counts for CSPG4 Nanostring counts for CSPG4 C D 10,000.0 10,000.0

1,000.0 1,000.0

100.0 100.0

10.0 10.0

1.0 1.0

Log Nanostring counts for NY-ESO-1 Log Nanostring counts for 0.1 NY-ESO-1 Log Nanostring counts for 0.1 0 0.5 101.5 2 2.5 3 0.51 1.5 2 2.5 3 Tumor intensity score Percentage tumor cells staining score

Figure 1. Validation of Nanostring data. Nanostring counts for CSPG4 gene expression in ﬁve melanoma lines were graphed against results from RT-PCR analysis (A) and MFI from FACS with CSPG4-speciﬁc mAb (B). immunohistochemical results, including scores for intensity of staining (C) and percentage of tumors cells staining (D) for NY-ESO-1 staining in 41 tumors samples, are graphed against the log of the Nanostring counts for NY-ESO-1. Immunohistochemical scores are as follows: for intensity of staining 0 ¼ no reactivity, 1 ¼ weak reactivity, 2 ¼ moderate reactivity, 3 ¼ intense reactivity, and for percentage of tumors cells that stained 0 ¼ 0%, 1 ¼ 0–5%, 2 ¼ 5–50%, and 3 ¼ >50%.

OF4 Clin Cancer Res; 2013 Clinical Cancer Research

Nanostring Digital RNA Counting for Melanoma Immunotherapy

Table 2. Expression of potential target genes Table 2. Expression of potential target genes (n ¼ 72) in tumors (n ¼ 72) in tumors (Cont'd )

Percentage of Average Percentage of Average þ Gene tumors PIRþ count Gene tumors PIR count Melanoma-related genes SSX3 20 796 AURKB 98 461 SYCP1 0 n/a B4GALNT1 15 222 TSPY1 7 4,996 CSPG4 92 2,406 Glioblastoma-related genes DCT 76 10,124 BCAN 56 2,143 ERBB4 7 172 CHI3L2 19 211 GRIN2A 3 120 EGFRvIII 0 n/a GRM3 0 n/a FABP7 47 1,106 MITF 90 3,665 IGF2BP3 61 437 MLANA 83 7,923 NRCAM 37 317 PMEL 85 84,888 PTPRZ1 78 2,437 SOXIO 90 902 TNC 92 2,249 ST3GAL5 100 2,077 Other tumor-related genes ST8SIA1 20 202 CEACAM5 2 618 TYR 76 9,014 ERBB2 75 255 TYRP1 47 23,414 FOLH1 41 539 GPC3 54 1,308 Cancer testis genes GTF2A1 100 1,026 CSAG2 71 1,046 KDR 92 310 CT45A1 24 2,478 KIF20A 69 255 CTAG1B 19 1,787 MSLN 7 124 CTAG2 17 585 MUC1 20 297 CTCFL 10 210 PSCA 12 328 CXorf48 0 n/a STAG2 100 1,004 GAGE1 29 186 TG 3 128 GAGE4 17 6,062 TKTL1 10 1,443 IL13RA2 32 1,468 WT1 3 140 MAGEA1 20 251 þ MAGEA2 49 237 NOTE: The percentage of tumors (n ¼ 59) that were PIR MAGEA3 75 2,567 (potential immune recognitionþ,deﬁned as a Nanostring MAGEA4 15 220 count >100) for each gene in the codeset is shown together MAGEA5 0 n/a with the average Nanostring count of the tumors which were þ MAGEA8 7 250 PIR for that gene. MAGEA9 19 570 MAGEA10 20 443 by multiple individuals over several years, results of both MAGEA11 15 580 the intensity of staining and the percentage of tumor cells MAGEA12 51 308 that stained were shown to compare well with Nanostring MAGEB1 14 289 RNA counts for NY-ESO-1 (Fig. 1C and 1D). MAGEB2 14 528 MAGEB3 2 171 Selection of potential target genes MAGEB6 2 130 On the basis of the counts obtained with the 8 pairs of MAGEC1 27 693 negative control probes (average, 9.0; STD, 5.8), 20 was MAGEC2 39 1,699 chosen as the background level for gene detection. To set a POTEF 8 213 value that would be likely for potential immune recognition PRAME 86 2,899 (PIR), we used our previously published data on the ability SAGE1 2 109 of the engineered T cells to recognize 3 tumor antigens: NY- SPANXN3 20 139 ESO-1 (encoded by gene CTAG1B), MART-1 (encoded by SPANXA1 2 135 gene MLANA), and CSPG4 (formerly known as HMW- SSX1 12 629 MAA, encoded by gene CSPG4; refs. 9–11). These reports SSX2 5 239 used the same melanoma lines analyzed in this study as (Continued on the following coloumn) targets and we observed that tumor cell lines with Nano- string counts greater than 100 for a given target gene

www.aacrjournals.org Clin Cancer Res; 2013 OF5

Beard et al.

consistently showed effector cytokine release when they targeted in previous trials and, although clinical responses were cocultured with genetically modified peripheral blood were achieved, patients suffered associated skin, eye, and ear lymphocytes (PBL), whereas lines with counts lower than toxicities (13, 14). Sixteen genes remained as potential 100 did not reproducibly demonstrate such reactivity (Sup- target antigens for consideration (Table 3). plementary Tables S1 and S2). MITF was expressed in tumors at an average level that is For each gene, the percentage of tumors showing PIR 8.5 times higher than the average expression in normal þ positivity (PIR , defined as a Nanostring count > 100) was tissues (Table 3); however, high levels of expression (counts þ determined as was the average Nanostring count of the PIR >1,000) in diaphragm, muscle, and uterus (Supplementary tumors (Table 2; Supplementary Table S3). Of the 72 Table S4) eliminated it from consideration. ST3GAL5 was potential target genes, 5 were not expressed in any of the eliminated secondary to expression at very high levels in tumor samples (Fig. 2; Supplementary Table S3). Of the multiple tissues including the brain, adrenal, thyroid, þ remaining 67 genes, 33 were PIR in more than 20% of spleen, and artery (Supplementary Table S4). Likewise, tumor samples. Twenty percent was chosen as a cut-off PTPRZ1 was eliminated because of high levels of expression value based on our previous experience targeting NY- in the brain, brainstem, and artery samples (Supplementary ESO-1 (5). RNA from 31 normal tissues was then subject Table S4). AURKB, IGF2BP3, and KIF20A all showed low to Nanostring Analysis with the same codeset (Supplemen- levels of expression on limited normal tissues; however, the tary Table S4). Analysis of these data by PCA showed clear average level of expression detected on tumors was suffi- differences in gene expression profiles between normal ciently low (ratio < 2.0) to render them non-ideal targets tissues and tumor samples (Fig. 3A). The positioning of (Table 3). Of the remaining 10 genes, 8 encode CTAs and the normal tissues testis and spleen were likely attributable exhibit little-to-no normal tissue expression outside of to the expression of the cancer testis genes and immune testis. Of these 8 genes, the highest levels of tumor RNA genes, respectively. We then conducted hierarchic clustering expression (all >1,000) were seen for CSAG2, IL13RA2, of the combined tumor and normal tissue datasets and MAGEA3, MAGEC2, and PRAME, all of which warrant observed that 27 of the initial 72 candidate genes were further consideration as possible targets (Table 3). The differentially expressed between tumors and normal tissues, remaining non-CTA candidate genes were CSPG4 and with at least a two-fold higher expression in tumors (Sup- SOX10. Both exhibited low levels of expression in a number þ plementary Fig. S2). Of these 27 genes, 20 were PIR in of normal tissues (skin, trachea, vein, heart, lung, dia- more than 20% of tumors samples (Table 2, Fig. 3B). Of phragm, muscle, adipose, uterus, prostate, thymus, spleen, those 20, MAGEA12 was eliminated from further consid- bone marrow, and gastrointestinal organs for CSPG4 and eration because of severe reported toxicities in a previous brain, brainstem, trachea, spleen, artery, and breast for clinical trial (12). The melanocyte differentiation antigens SOX10); however, the average levels of expression in tumors MLANA, PMEL, and TYR were eliminated because they were for both were substantially higher than in normal tissues

72 Potential target genes

- PIR PIR+

5 Genes 67 Genes

> 20% of Tumors ≤ 20% of Tumors Figure 2. Selection of potential target genes based on PIR 33 Genes 34 Genes positivity in tumor samples. Sixty- seven of the 72 potential candidate þ genes were PIR (deﬁned as Nanostring count >100) in some of the tumor samples (n ¼ 59). Thirty- Melanoma- Other tumor- Cancer testis Glioblastoma þ related genes related genes genes genes three genes were PIR in more than 20% of the tumor samples. 1. AURKB 1. ERBB2 1. CSAG2 1. BCAN 2. CSPG4 2. FOLH1 2. CT45A1 2. FABP7 3. DCT 3. GPC3 3. GAGE1 3. IGF2BP3 4. MITF 4. GTF2A1 4. IL13RA2 4. NRCAM 5. MLANA 5. KDR 5. MAGEA2 5. PTPRZ1 6. PMEL 6. KIF20A 6. MAGEA3 6. TNC 7. SOX10 7. STAG2 7. MAGEA12 8. ST3GAL5 8. MAGEC1 9. TYR 9. MAGEC2 10. TYRP1 PRAME

OF6 Clin Cancer Res; 2013 Clinical Cancer Research

Nanostring Digital RNA Counting for Melanoma Immunotherapy

A 5

2.9

0.8

–1.3

–3.4

–5.5 Testis

–7.6 PC #2 10.6%

–9.7 Spleen

Figure 3. Differentiation of gene –11.8 expression between normal tissues and tumors. A, PCA for –13.9 normal tissues (n ¼ 31) and tumor samples (n ¼ 59). The color coding –16 is as follows: red ¼ normal tissue, –13 –11.1 –9.2 –7.3 –5.4 –3.5 –1.6 0.3 2.2 4.1 6 blue ¼ tumor. B, hierarchic PC #3 6.0% PC #1 12.5% clustering of 20 genes that differentiate tumors from normal B tissues with at least a twofold higher expression in tumors, and are overexpressed in more than 20% of tumors. Vertical axis shows normal tissues (orange) and tumors (yellow). TYR MITF PMEL SOX10 KIF20A CSPG4 CSAG2 MLANA AURKB GAGE1 PRAME PTPRZ1 IL13RA2 IGF2BP3 MAGEA3 MAGEA2 MAGEC2 MAGEC1 ST3GAL5 MAGEA12

–7.37 0.00 7.37

(6.6 times higher for CSPG4 and 4.6 times higher for studies have examined gene expression profiles of this SOX10). Therefore, seven genes are identified as potential antigen group in melanoma and other cancers using immunotherapy targets: CSAG2, MAGEA3, MAGEC2, microarray technology and RT-PCR (16, 17). Herein, we IL13RA2, PRAME, CSPG4, and SOX10. identified five genes encoding CTAs, which are expressed at high levels on a large percentage of melanoma tumor Discussion samples studied: CSAG2, MAGEA3, MAGEC2, IL13RA2, The identification of target antigens for immunotherapy and PRAME. is a complex process that involves assessing the expression CSAG2, also known as taxol-resistance-associated gene-3 of antigens on tumors and normal tissues. CTAs are of (TRAG3), was overexpressed on 71% of studied melanoma particular interest as immunotherapy targets because they tumors with no significant expression on any normal tis- are expressed in multiple cancers of diverse histologic sues. It is overexpressed in multiple other histologies includ- origin, including breast cancer, prostate cancer, non-small ing carcinoma of the bladder, cervix, breast, esophagus, bile cell lung cancer, gastrointestinal cancers such as colon and duct, stomach, colon, and lung, and its expression has been esophageal cancers, bladder cancer, and melanoma. Aside correlated with poor prognosis in both ovarian cancer and from expression in male germ cells, CTA expression in osteosarcoma (18–21). Preclinical studies have identified a normal human tissues is relatively restricted (15). Previous potential target CSAG2-directed T-cell epitope capable of

www.aacrjournals.org Clin Cancer Res; 2013 OF7

Beard et al.

Table 3. Nanostring counts for potential expressed on 39% of tumors in this study, with no expression on any normal tissues aside from testis. Several pre- candidate genes (n ¼ 16) clinical studies have identified potential target epitopes and have generated CTL that show functionality against target Average Ratio cells (29, 30). count Average (tumors: IL13RA2 was overexpressed on 32% of studied tumors normal count normal and showed low levels of expression on the liver and Gene tissues tumors tissues) adrenal samples. It has been associated with adrenocortical Melanoma-related genes carcinoma, glioblastoma multiforme, and systemic sclero- AURKB 397 461 1.2 sis, and IL13RA2-specific CARs have shown efficacy against CSPG4 367 2,406 6.6 glioma cell targets in preclinical studies (31–33). PRAME MITF 433 3,665 8.5 was overexpressed in 86% of melanoma samples, and its SOXIO 195 902 4.6 average count on overexpressing tumors (2899) was the ST3GAL5 672 2,077 3.1 highest among any of the potential CTA targets. It was Cancer testis genes absent on any normal tissues in this cohort although other CSAG2 - 1,046 - studies have reported some expression on normal tissues GAGE1 - 186 - such as endometrium, ovaries, and adrenals (34, 35). IL13RA2 332 1,468 4.4 Its expression has been widely reported in other cancers MAGEA2 - 237 - including carcinoma of the lung and kidney, squamous cell MAGEA3 453 2,567 5.7 carcinoma of the head and neck, sarcomas, mammary MAGEC1 165 693 4.2 cancer, multiple myeloma, and acute leukemias (34). It is MAGEC2 650 1,699 2.6 one of the few CTAs that is commonly expressed in leukemic PRAME 600 2,899 4.8 malignancies. Preclinical studies have generated PRAME- Glioblastoma-related genes specific T cells that showed activity across multiple histol- IGF2BP3 267 437 1.6 ogies encompassing both solid organ and hematologic PTPRZ1 573 2437 4.3 malignancies (36, 37). Other genes CSPG4 and SOX10 are both melanoma-associated genes KIF20A 192 255 1.3 that were expressed in 92% and 90% of tumors in this study, NOTE: () denotes an average normal tissue count that is respectively. Unlike CTA, they do exhibit expression on reflective solely of expression on testis. () denotes no multiple normal tissues, although at much lower levels overexpression on any normal tissues. The average Nano- than on tumors. This does raise concerns regarding appli- string counts for all normal tissues that were PIRþ for each cation in the clinical setting and argues for safety measures gene are provided as are the average Nanostring counts for such as dose-escalation trials or the implementation of a þ CSPG4 tumors that were PIR . The ratio of average expression in suicide gene (38). is a highly immunogenic cell tumors to average expression in normal tissues is also surface proteoglycan which was identified on melanoma shown. cells in the 1970s, and it has been found on glioblastoma, triple-negative breast cancer, head and neck squamous cell cancer, mesothelioma, renal cell cancer, and sarcoma as well (39–43). It has been targeted with vaccines in clinical inducing cytotoxic lymphocytes (CTL; ref. 22). MAGEA3 is a trials with no reported toxicity, and antitumor effects very appealing target given its overexpression on a high have been reported in preclinical immunotherapy models percentage of metastatic melanoma tumors (75% in this with melanoma and head and neck squamous cell cancers study) and its potential as a target for ACT in other histol- targets (9, 40, 44). SOX10 is a transcription factor that is ogies such as colorectal cancer, lung cancer, breast cancer, expressed on neural crest cells and melanocytes, and has esophageal cancer, and glioblastoma (23, 24). A particular been shown to be crucial for the maintenance of neoplastic epitope, MAGEA3 112–120, was targeted in TCR gene cells (45, 46). In addition to being widely expressed on þ therapy trials in MAGEA3 patients with observed cancer melanomas and other lesions such as giant congenital nevi, regression but severe associated neurologic toxicity and it has also been identified on cancers of the breast and mortality, likely secondary to TCR cross-recognition of this prostate (45, 47–49). Naturally occurring anti-SOX10 CTLs epitope in MAGEA12, which was found to be expressed in were identified in a patient with a dramatic response to the brain (12). The generation of a completely MAGEA3- immunotherapy (50). specific TCR, however, could allow for its targeting in future This study identified genes that have potential as immu- studies. MAGEC2, formerly known as hepatocellular carci- notherapy targets based on their expression in a high noma-associated antigen 587 (HCA587), is also overex- percentage of studied melanoma tumors with limited pressed in a variety of cancers aside from melanoma, expression in normal tissues. One possible study limitation including hepatocellular carcinoma, gallbladder carcino- is the correlation between gene expression and antigen ma, medulloblastoma, multiple myeloma, and squamous expression. We have shown that, for CSPG4 and NY- cell cancer of the head and neck (25–28). It was over- ESO-1, there is a strong association between the level of

OF8 Clin Cancer Res; 2013 Clinical Cancer Research

Nanostring Digital RNA Counting for Melanoma Immunotherapy

gene expression as assessed by Nanostring and the degree Disclosure of Potential Conflicts of Interest of antigen expression, as determined by both FACS and No potential conflicts of interest were disclosed. immunohistochemistry; however, this may not be the case for every gene. Importantly, these data do not directly Authors' Contributions Conception and design: R.E. Beard, D. Abate-Daga, R.A. Morgan address issues involving the potential safety of a given target Development of methodology: R.E. Beard, D. Abate-Daga, Z. Zheng gene. Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R.E. Beard, S.F. Rosati, J.R. Wunderlich To have the potential for clinical application, each new Analysis and interpretation of data (e.g., statistical analysis, biosta- target gene would require the generation of reagents (TCR tistics, computational analysis): R.E. Beard, D. Abate-Daga, S.F. Rosati, or CAR) that can mediate specific antigen recognition S.A. Rosenberg, R.A. Morgan Writing, review, and/or revision of the manuscript: R.E. Beard, D. Abate- and, the target antigen must not be expressed on a vital Daga, S.F. Rosati, S.A. Rosenberg, R.A. Morgan tissue. In the case of CTAs, while it is widely reported that Administrative, technical, or material support (i.e., reporting or orga- these genes are cancer specific, this is not universally true. nizing data, constructing databases): R.E. Beard, S.A. Rosenberg, R.A. Morgan As we recently reported, the MAGEA12 gene is strongly Study supervision: R.A. Morgan expressed in rare neurons in the human brain and expression in these isolated cells was likely sufficient to initiate a Acknowledgments destructiveimmuneresponseleadingtodeathinsome The authors thank Arnold Mixon and Shawn Farid for technical support with FACS analysis and the Laboratory of Pathology Department at the patients (12). On the other hand, we have used the National Cancer Institute for their role in immunohistochemical staining. identical strategy of TCR gene therapy to target NY- ESO-1, and in more than 40 patients treated, have not Grant Support observed any target-related toxicities. Clearly, more This work is supported by the Intramural Research Program of the Center detailed studies (e.g., multiple tissue immunohistochem- for Cancer Research, National Cancer Institute, NIH. The costs of publication of this article were defrayed in part by the istry) would be needed before any of these new antigens payment of page charges. This article must therefore be hereby marked can be targeted in clinical trials; however, Nanostring advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate provides a reliable way to test multiple candidate genes this fact. at once and select attractive potential targets for further Received May 7, 2013; revised June 20, 2013; accepted July 9, 2013; investigation. published OnlineFirst September 10, 2013.

References 1. Turcotte S, Rosenberg SA. Immunotherapy for metastatic solid can- murine-human hybrid MART-1-reactive T cell receptor following cers. Adv Surg 2011;45:341–60. a rapid expansion. Cancer Immunol Immunother 2010;59:1551–60. 2. Park TS, Rosenberg SA, Morgan RA. Treating cancer with genetically 12. Morgan RA, Chinnasamy N, Abate-Daga D, Gros A, Robbins PF, engineered T cells. Trends Biotechnol 2011;29:550–7. Zheng Z, et al. Cancer regression and neurological toxicity following 3. Rosenberg SA. Raising the bar: the curative potential of human cancer anti-MAGE-A3 TCR gene therapy. J Immunother 2013;36:133–51. immunotherapy. Sci Transl Med 2012;4:127ps8. 13. Johnson LA, Morgan RA, Dudley ME, Cassard L, Yang JC, Hughes MS, 4. Lai JP, Robbins PF, Raffeld M, Aung PP, Tsokos M, Rosenberg SA, et al. Gene therapy with human and mouse T-cell receptors mediates et al. NY-ESO-1 expression in synovial sarcoma and other mesen- cancer regression and targets normal tissues expressing cognate chymal tumors: significance for NY-ESO-1-based targeted therapy antigen. Blood 2009;114:535–46. and differential diagnosis. Mod Pathol 2012;25:854–8. 14. Seaman BJ, Guardiani EA, Brewer CC, Zalewski CK, King KA, Rudy S, 5. Robbins PF, Morgan RA, Feldman SA, Yang JC, Sherry RM, Dudley et al. Audiovestibular dysfunction associated with adoptive cell immu- ME, et al. Tumor regression in patients with metastatic synovial cell notherapy for melanoma. Otolaryngol Head Neck Surg 2012;147: sarcoma and melanoma using genetically engineered lymphocytes 744–9. reactive with NY-ESO-1. J Clin Oncol 2011;29:917–24. 15. Scanlan MJ, Simpson AJ, Old LJ. The cancer/testis genes: review, 6. Kochenderfer JN, Dudley ME, Feldman SA, Wilson WH, Spaner DE, standardization, and commentary. Cancer Immun 2004;4:1. Maric I, et al. B-cell depletion and remissions of malignancy along with 16. Segal NH, Blachere NE, Guevara-Patino JA, Gallardo HF, Shiu HY, cytokine-associated toxicity in a clinical trial of anti-CD19 chimeric- Viale A, et al. Identification of cancer-testis genes expressed by antigen-receptor-transduced T cells. Blood 2012;119:2709–20. melanoma and soft tissue sarcoma using bioinformatics. Cancer 7. Porter DL, Levine BL, Kalos M, Bagg A, June CH. Chimeric antigen Immun 2005;5:2. receptor-modified T cells in chronic lymphoid leukemia. N Engl J Med 17. Bredenbeck A, Hollstein VM, Trefzer U, Sterry W, Walden P, Losch 2011;365:725–33. FO. Coordinated expression of clustered cancer/testis genes 8. Geiss GK, Bumgarner RE, Birditt B, Dahl T, Dowidar N, Dunaway DL, encoded in a large inverted repeat DNA structure. Gene 2008; et al. Direct multiplexed measurement of gene expression with color- 415:68–73. coded probe pairs. Nat Biotechnol 2008;26:317–25. 18. Chen Z, Zhu B, Wu Y. Expression of TRAG-3 antigen in non-small-cell 9. Burns WR, Zhao Y, Frankel TL, Hinrichs CS, Zheng Z, Xu H, et al. A high lung carcinomas. Lung Cancer 2002;38:101–2. molecular weight melanoma-associated antigen-specific chimeric 19. Ohta M, Tanaka F, Sadanaga N, Yamaguchi H, Inoue H, Mori M. antigen receptor redirects lymphocytes to target human melanomas. Expression of the TRAG-3 gene in human esophageal cancer: the Cancer Res 2010;70:3027–33. frequent synchronous expression of MAGE-3 gene. Oncol Rep 10. Wargo JA, Robbins PF, Li Y, Zhao Y, El-Gamil M, Caragacianu D, et al. 2006;15:1529–32. Recognition of NY-ESO-1þ tumor cells by engineered lymphocytes is 20. Wu YZ, Zhao TT, Ni B, Zou LY, Liu HL, Zhu B. Expression of TRAG-3 in enhanced by improved vector design and epigenetic modulation of breast cancer. Int J Cancer 2003;107:167–8. tumor antigen expression. Cancer Immunol Immunother 2009;58: 21. Zou C, Shen J, Tang Q, Yang Z, Yin J, Li Z, et al. Cancer-testis 383–94. antigens expressed in osteosarcoma identified by gene micro- 11. Goff SL, Johnson LA, Black MA, Xu H, Zheng Z, Cohen CJ, et al. array correlate with a poor patient prognosis. Cancer 2012;118: Enhanced receptor expression and in vitro effector function of a 1845–55.

www.aacrjournals.org Clin Cancer Res; 2013 OF9

Beard et al.

22. Zhu B, Chen Z, Cheng X, Lin Z, Guo J, Jia Z, et al. Identification of HLA- potent antitumor reactivity useful for therapeutic T-cell receptor gene A0201-restricted cytotoxic T lymphocyte epitope from TRAG-3 anti- transfer. Clin Cancer Res 2011;17:5615–25. gen. Clin Cancer Res 2003;9:1850–7. 37. Quintarelli C, Dotti G, Hasan ST, De Angelis B, Hoyos V, Errichiello S, 23. Shantha Kumara HM, Grieco MJ, Caballero OL, Su T, Ahmed A, Ritter et al. High-avidity cytotoxic T lymphocytes specific for a new PRAME- E, et al. MAGE-A3 is highly expressed in a subset of colorectal cancer derived peptide can target leukemic and leukemic-precursor cells. patients. Cancer Immun 2012;12:16. Blood 2011;117:3353–62. 24. Chinnasamy N, Wargo JA, Yu Z, Rao M, Frankel TL, Riley JP, et al. A 38. Di Stasi A, Tey SK, Dotti G, Fujita Y, Kennedy-Nasser A, Martinez C, TCR targeting the HLA-A0201-restricted epitope of MAGE-A3 recog- et al. Inducible apoptosis as a safety switch for adoptive cell therapy. nizes multiple epitopes of the MAGE-A antigen superfamily in several N Engl J Med 2011;365:1673–83. types of cancer. J Immunol 2011;186:685–96. 39. Rivera Z, Ferrone S, Wang X, Jube S, Yang H, Pass HI, et al. CSPG4 as 25. Oba-Shinjo SM, Caballero OL, Jungbluth AA, Rosemberg S, Old LJ, a target of antibody-based immunotherapy for malignant mesotheli- Simpson AJ, et al. Cancer-testis (CT) antigen expression in medullo- oma. Clin Cancer Res 2012;18:5352–63. blastoma. Cancer Immun 2008;8:7. 40. Wang X, Wang Y, Yu L, Sakakura K, Visus C, Schwab JH, et al. CSPG4 26. Pabst C, Zustin J, Jacobsen F, Luetkens T, Kroger N, Schilling G, et al. in cancer: multiple roles. Curr Mol Med 2010;10:419–29. Expression and prognostic relevance of MAGE-C1/CT7 and MAGE- 41. Wang X, Osada T, Wang Y, Yu L, Sakakura K, Katayama A, et al. C2/CT10 in osteolytic lesions of patients with multiple myeloma. Exp CSPG4 protein as a new target for the antibody-based immunotherapy Mol Pathol 2010;89:175–81. of triple-negative breast cancer. J Natl Cancer Inst 2010;102: 27. Riener MO, Wild PJ, Soll C, Knuth A, Jin B, Jungbluth A, et al. Frequent 1496–512. expression of the novel cancer testis antigen MAGE-C2/CT-10 in 42. Benassi MS, Pazzaglia L, Chiechi A, Alberghini M, Conti A, Cattaruzza hepatocellular carcinoma. Int J Cancer 2009;124:352–7. S, et al. NG2 expression predicts the metastasis formation in soft- 28. Figueiredo DL, Mamede RC, Spagnoli GC, Silva WA Jr, Zago M, Neder tissue sarcoma patients. J Orthop Res 2009;27:135–40. L, et al. High expression of cancer testis antigens MAGE-A, MAGE-C1/ 43. Svendsen A, Verhoeff JJ, Immervoll H, Brogger JC, Kmiecik J, Poli A, CT7, MAGE-C2/CT10, NY-ESO-1, and gage in advanced squamous et al. Expression of the progenitor marker NG2/CSPG4 predicts poor cell carcinoma of the larynx. Head Neck 2011;33:702–7. survival and resistance to ionising radiation in glioblastoma. Acta 29. Wen W, Zhang L, Peng J, Chen J, Hao J, Li X, et al. Identification of Neuropathol 2011;122:495–510. promiscuous HLA-DR-restricted CD4(þ) T-cell epitopes on the can- 44. Wang X, Katayama A, Wang Y, Yu L, Favoino E, Sakakura K, et al. cer-testis antigen HCA587. Cancer Sci 2011;102:1455–61. Functional characterization of an scFv-Fc antibody that immunother- 30. Xing Q, Pang XW, Peng JR, Yin YH, Li Y, Yu X, et al. Identification of new apeutically targets the common cancer cell surface proteoglycan cytotoxic T-lymphocyte epitopes from cancer testis antigen HCA587. CSPG4. Cancer Res 2011;71:7410–22. Biochem Biophys Res Commun 2008;372:331–5. 45. Agnarsdottir M, Sooman L, Bolander A, Stromberg S, Rexhepaj E, 31. Fernandez-Ranvier GG, Weng J, Yeh RF, Khanafshar E, Suh I, BergqvistM,etal.SOX10expressioninsuperficial spreading Barker C, et al. Identification of biomarkers of adrenocortical car- and nodular malignant melanomas. Melanoma Res 2010;20: cinoma using genomewidegeneexpressionprofiling. Arch Surg 468–78. 2008;143:841–6. 46. Shin J, Vincent JG, Cuda JD, Xu H, Kang S, Kim J, et al. Sox10 is 32. Granel B, Allanore Y, Chevillard C, Arnaud V, Marquet S, Weiller PJ, expressed in primary melanocytic neoplasms of various histologies et al. IL13RA2 gene polymorphisms are associated with systemic but not in fibrohistiocytic proliferations and histiocytoses. J Am Acad sclerosis. J Rheumatol 2006;33:2015–9. Dermatol 2012;67:717–26. 33. Kong S, Sengupta S, Tyler B, Bais AJ, Ma Q, Doucette S, et al. 47. Zhong WD, Qin GQ, Dai QS, Han ZD, Chen SM, Ling XH, et al. SOXs in Suppression of human glioma xenografts with second-generation human prostate cancer: implication as progression and prognosis IL13R-specific chimeric antigen receptor-modified T cells. Clin Cancer factors. BMC Cancer 2012;12:248. Res 2012;18:5949–60. 48. Cimino-Mathews A, Subhawong AP, Elwood H, Warzecha HN, 34. Ikeda H, Lethe B, Lehmann F, van Baren N, Baurain JF, de Smet C, et al. Sharma R, Park BH, et al. Neural crest transcription factor Sox10 is Characterization of an antigen that is recognized on a melanoma preferentially expressed in triple-negative and metaplastic breast showing partial HLA loss by CTL expressing an NK inhibitory receptor. carcinomas. Hum Pathol 2012;44:959–65. Immunity 1997;6:199–208. 49. Shakhova O, Zingg D, Schaefer SM, Hari L, Civenni G, Blunschi J, et al. 35. van Baren N, Chambost H, Ferrant A, Michaux L, Ikeda H, Millard I, et al. Sox10 promotes the formation and maintenance of giant congenital PRAME, a gene encoding an antigen recognized on a human mela- naevi and melanoma. Nat Cell Biol 2012;14:882–90. noma by cytolytic T cells, is expressed in acute leukaemia cells. Br J 50. Khong HT, Rosenberg SA. The Waardenburg syndrome type 4 Haematol 1998;102:1376–9. gene, SOX10, is a novel tumor-associated antigen identified in a 36. Amir AL, van der Steen DM, van Loenen MM, Hagedoorn RS, de Boer patient with a dramatic response to immunotherapy. Cancer Res R, Kester MD, et al. PRAME-specific Allo-HLA-restricted T cells with 2002;62:3020–3.

OF10 Clin Cancer Res; 2013 Clinical Cancer Research

Gene Expression Profiling using Nanostring Digital RNA Counting to Identify Potential Target Antigens for Melanoma Immunotherapy

Rachel E. Beard, Daniel Abate-Daga, Shannon F. Rosati, et al.

Clin Cancer Res Published OnlineFirst September 10, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-13-1253

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2013/12/04/1078-0432.CCR-13-1253.DC1

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2013/08/29/1078-0432.CCR-13-1253. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.