Gene Expression Differences Associated with Human Papillomavirus Status in Head and Neck Squamous Cell Carcinoma Robbertj.C

Human Cancer Biology

Gene Expression Differences Associated with Human Papillomavirus Status in Head and Neck Squamous Cell Carcinoma RobbertJ.C. Slebos,1, 2 Yajun Yi, 7 Kim Ely,3 Jesse Carter,8 Amy Evjen,1Xueqiong Zhang,4 Yu Shyr,4 Barbara M. Murphy,8 Anthony J. Cmelak,5 Brian B. Burkey,2 James L. Netterville,2 Shawn Levy,6 Wendell G. Yarbrough,1, 2 and Christine H. Chung8

Abstract Human papillomavirus (HPV) is associated with a subset of head and neck squamous cell carcinoma (HNSCC). Between 15% and 35% of HNSCCs harbor HPV DNA. Demographic and exposure differences between HPV-positive (HPV+) and negative (HPV) HNSCCs suggest that HPV + tumors may constitute a subclass with different biology, whereas clinical differences have also been observed. Gene expression profiles of HPV+ and HPV tumors were compared with further exploration of the biological effect of HPV in HNSCC. Thirty-six HNSCC tumors were analyzed using Affymetrix Human 133U Plus 2.0 GeneChip and for HPV by PCR and real-time PCR. Eight of 36 (22%) tumors were positive for HPV subtype 16. Statistical analysis using Significance Analysis of Microarrays based on HPV status as a supervising variable resulted in a list of 91genes that were differentially expressed with statistical significance. Results for a subset of these genes were verified by real-time PCR. Genes highly expressed in HPV+ samples included cell cycle regulators (p16INK4A, p18, and CDC7) and transcription factors (TAF7L, RFC4, RPA2, andTFDP2).The microarray data were also investigated by mapping genes by chromosomal location (DIGMAP).A large number of genes on chromosome 3q24-qter had high levels of expression in HPV+ tumors. Further investigation of differentially expressed genes may reveal the unique pathways in HPV+ tumors that may explain the different natural history and biological properties of these tumors. These properties may be exploited as a target of novel therapeutic agents in HNSCC treatment.

Head and neck cancer remains one of the most devastating subtypes of HPV have been described in humans, with HPV cancers in the United States. Development of the vast majority type 8, 11, 16, and 18 being associated with the majority of of these tumors has been attributed to use of tobacco and human disease. In the cervix, a distinction is made between ethanol products, but a significant portion of these tumors are ‘‘low-risk’’ (types 8 and 11) and ‘‘high-risk’’ (types 16 and 18) associated with human papillomavirus (HPV; refs. 1, 2). HPV, depending on their association with premalignant and Infection with HPV is associated with malignant and prema- malignant lesions, respectively. Reports of the prevalence of lignant lesions of the uterine, cervix, vulva, penis, conjunctiva, HPV infection in head and neck squamous cell carcinoma and upper aerodigestive tract (for review, see ref. 3). Over 100 (HNSCC) indicate that 15% to 35% of HNSCC may harbor HPV sequences, depending on the detection method used (4). DNA amplification by PCR remains the most sensitive technique to detect HPV, with almost 35% of HNSCCs yielding Authors’ Affiliations: Departments of 1Cancer Biology, 2Otolaryngology, HPV-specific amplification products, although this result 3 4 5 6 Pathology, Biostatistics, Radiation Oncology, and Biomedical Informatics; may be biased because of contamination problems associated Divisions of 7Genetic Medicine and 8Hematology/Oncology, Department of Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt University School of with PCR. HPV is most commonly found in tonsillar tumors Medicine, Nashville, Tennessee (45-100%; ref. 5) with HPV type 16 (HPV16) being found in Received 9/14/05; revised 10/24/05; accepted 11/10/05. the vast majority and HPV18 associated with most others (6). Grant support: Barry Baker Research Endowment,Vanderbilt Physician-Scientist There are some indications that HPV-positive (HPV+) Development Award (C.H. Chung), Robert J. Kleberg, Jr. and Helen C. Kleberg Foundation (C.H. Chung and W.G. Yarbrough), and Damon Runyon Cancer HNSCCs may represent a subclass with a different biology Research Foundation (C.H. Chung). and with different clinical properties. Molecular evidence that The costs of publication of this article were defrayed in part by the payment of page HPV status determines a separate class of HNSCC comes from charges. This article must therefore be hereby marked advertisement in accordance studies showing HPV+ tumors are associated with low rates with 18 U.S.C. Section 1734 solely to indicate this fact. INK4A Note: Supplementary data for this article are available at Clinical Cancer Research of p53 or p16 mutations as opposed to HPV-negative INK4A Online (http://clincancerres.aacrjournals.org/). (HPV ) HNSCCs, where p53 and p16 alterations are Requests for reprints: Christine H. Chung, Division of Hematology/Oncology, common (50% and 80%, respectively; refs. 7–9). Comparative Department of Medicine, Vanderbilt University School of Medicine, 2220 Pierce genomic hybridization analysis showed specific patterns of Avenue, 777 Preston Research Building, Nashville,TN 37232-6307. Phone: 615- chromosomal alterations associated with HPV+ tonsillar 322-4967; Fax: 615-343-7602; E-mail: [email protected]. F 2006 American Association for Cancer Research. tumors, which were more likely to have gain on chromosome doi:10.1158/1078-0432.CCR-05-2017 3q, or have an absence of gains on chromosome 7q relative to

www.aacrjournals.org 701 Clin Cancer Res 2006;12(3) February 1, 2006 Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. Human Cancer Biology

HPV tumors (5). HPV status is also associated with specific RNeasy Mini kit according to the manufacturer’s recommendations demographics: patients with HPV+ HNSCCs are usually (Qiagen) using f10 to 20 mg of wet tissue from each sample. Fifty younger and are less likely to have tobacco exposure than nanograms of the total RNA were amplified using NuGEN Ovation those with HPV tumors. Several studies have suggested that Biotin RNA Amplification and Labeling kit (NuGen, San Carlos, CA) according to the manufacturer’s recommendations. The NuGEN HPV+ tumors are associated with favorable survival (4, 10). Ovation amplification methodology uses an isothermal linear Despite these indications that HPV status is associated with amplification using random hexamers. This technology provides molecular and clinical differences, all HNSCCs are clinically sensitive and rapid whole-genome amplification without introducing managed irrespective of their HPV status. Understanding of a significant bias toward the 3V end of the transcripts (14). Fifteen the differences between HPV+ and HPV HNSCC tumors may micrograms of biotin-labeled aRNA were fragmented, and the quality allow us to develop biomarkers for early detection or recurrence of the RNA was reconfirmed using the Agilent RNA 6000 Nano surveillance, to identify therapeutic targets, and to begin LabChip kit and Agilent 2100 bioanalyzer. The fragmented, biotin- individualization of treatment based on the biology of these labeled aRNA was combined with the hybridization mix and loaded tumors. The aim of this study was to identify the differences in on to the Affymetrix Human Genome U133 plus 2.0 GeneChip. After the gene expression profiles of HPV+ and HPV HNSCCs and to hybridization, the GeneChip was washed, stained with streptavidin/ phycoerythrin conjugate and biotinylated antibody, and scanned better understand the biological effect of HPV infection in HNSCC. We found that there is a distinct gene expression profile that is associated with HPV status analyzed by Ta b l e 1. Patient characteristics in HPV+ and HPV Significance Analysis of Microarrays. In addition, the expression cases data was analyzed using differential gene locus mapping (DIGMAP; ref. 11) to investigate the correlation between HPV + HPV Total previously published chromosomal abnormalities and gene (n =8) (n =28) (N =36) expression patterns. These analyses revealed that HPV+ tumors had increased levels of expression of genes on chromosome Age (median, range)* 49 (41-65) 58 (30-89) 55 (30-89) 3q24-qter compared with HPV tumors. Sex Male 8 2 1 29 Female 0 7 7 Materials and Methods Racec White 8 16 24 Patient selection and specimen collection. Thirty-six freshly frozen Black 0 9 9 tumor samples were prospectively collected from patients undergoing Other 0 3 3 surgery or biopsy for HNSCC at the University of North Carolina at Chapel Hill (21 patients) and Vanderbilt University (15 patients; see To b a c c o u s e Table 1 and Supplementary Data). All tissues were snap-frozen in liquid Ever 6 26 32 nitrogen within 30 minutes of surgical resection or biopsy and kept at Never 2 2 4 80jC until further processing. All patients consented to participation Alcohol use in this study under protocols approved by the Institutional Review Ye s 4 1 8 2 2 Boards at the two institutions. A previous gene expression study (12) No 4 10 14 included the 21 tumors from University of North Carolina reported Tu m o r s i t e b here, but to allow comparison with the specimens from Vanderbilt and Oral cavity 0 15 15 because the Agilent platform used in the prior study was discontinued, a Oropharynx 7 2 9 completely new expression analysis was done using the Affymetrix Larynx 1 8 9 platform (see below). HPV detection and DNA sequencing. Tumor DNAs were tested for Hypopharynx 0 3 3 the presence of HPV DNA using a previously established PCR-based Clinical stage method (13). This method employs degenerate PCR primers (MY09 I-II 0 3 3 and MY11, WD72/76 and WD66/67/154) that are designed to III 4 9 13 represent highly conserved HPV L1 and E6 sequences present in all IV 4 16 20 major types of HPV. In addition, all HPV-positive samples were also Clinical cervical lymph nodex tested with a HPV16-specific PCR for E7 (primer A, 5V-GGACCGGTC- Positive 7 18 25 V V GATGTATGTCT-3 and primer B, 3-TAAAACCATCCATTACATCCCG- Negative 1 10 11 5V). As a positive control for amplification, primers for h-globin are Pathologic cervical lymph nodek included with each sample (13). Optimal conditions for this combined Positive 6 17 23 PCR were determined using DNA from the cervical carcinoma cell line SiHa, which harbors on average two copies of HPV16 DNA per cell Negative 2 10 12 (13). Other positive control cell lines were CaSki (HPV16) and HeLa Tu m o r g r a d e (HPV18). For each case, 200 ng of tumor DNA were tested for the Well differentiated 1 3 4 presence of HPV DNA. PCR samples, which showed amplification Moderately differentiated 5 21 26 products indicating the presence of HPV were purified using PCR Poorly differentiated 2 4 6 purification columns (Qiagen, Valencia, CA) and subjected to bidirectional sequence analysis. In all of such cases, a positive *P = 0.08,Wilcoxon rank-sum test. identification of the HPV type could be made. cP = 0.03, Fisher’s exact test (White versus non-White). RNA isolation and DNA microarray analysis. Each tumor was bP < 0.001, Fisher’s exact test. examined by H&E staining to ensure presence of tumor and enriched xP = 0.22, Fisher’s exact test. by macrodissection to achieve a minimum of 70% tumor cells in each kPathologic lymph node status was missing for one patient. preparation. Total RNA was purified from frozen tumors using Qiagen

Clin Cancer Res 2006;12(3) February1,2006 702 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. HPV and Gene Expression in HNSCC according to the manufacturer’s recommendations. The raw micro- statistical analysis package (SAS Institute, Research Triangle Park, NC). array data was normalized using Perfect Match software for further RT-PCR data were analyzed by the 2DDCT method as described statistical analyses. previously (21). Briefly, the average Ct was calculated for the four Gene expression data analysis. The genes that were differentially replicate analyses of the three control genes (18S, PPIA, and GUSB), + expressed in HPV and HPV tumors were selected based on and this value was subtracted from the average Ct calculated from the Significance Analysis of Microarrays (15). The selected genes were four replicate analyses for the genes of interest. Expression differences + verified for statistically significant prediction power using the class were compared using these normalized DCt values between the HPV prediction model based upon the compound covariate method and HPV tumors, and the observed differences were tested using (16, 17). This class prediction model determined whether the patterns Student’s t test. Two-tailed Ps < 0.05 were considered statistically of gene expression can identify two classes of HPV+ versus HPV significant. tumors. The accuracy of the classification rate using the selected genes was estimated using the leave-one-out cross-validation. The pattern among the statistically significant discriminator genes was investigated Results using hierarchical clustering algorithm (18, 19). Chromosome mapping of expression data. DIGMAP was done as HPV detection in HNSCC tumors. A total of 36 DNA described before (11). Briefly, chromosomal locus information for samples obtained from HNSCC specimens, representing all Affymetrix probes was retrieved in a batch mode from our local Gene subclasses except nasopharynx (Table 1), were subjected to Annotation Project database. Genes exhibiting significant differential HPV analysis using PCR amplification of E6, E7, and L1 expression were then identified by T test (significance threshold P < 0.01) using the MEV software package (TMEV2.1; ref. 20). The T scores sequences. An example of an analysis of HPV E6 is shown in 1 Fig. 1. Seven tumors were positive for both E6 and L1 PCR were log-transformed reciprocal Ps [log10(P )]. The output files from these statistical analyses were processed by the DIGMAP Viewer and analyses, whereas one tumor was positive for L1 and not differential flag regions mapping programs. We implemented a sliding for either E6 or E7 (970108), and one tumor was positive for window method using perl scripts to compute total T scores per E6 and E7 but not for L1 (300171). Based on these results, million base pairs (Mbp) from neighboring genes. In this study, we the eight tumors that were positive for E6 and E7 were determined that a window size of 10 genes was optimal for visualizing classified as HPV+, whereas the one specimen that was only differential flag regions without loss of sensitivity and low noise levels. positive for L1 was classified as HPV. DNA sequence The sliding windows overlap each other by one gene locus (i.e., step analysis of the E6 PCR products revealed that all tumors size = 1) to cover the entire chromosome, and the gene expression harbored type 16 HPV. RNA expression of HPV16-E6 was profiles are displayed as a moving average per Mbp. Normalized T + scores for a window size 10 were calculated by summing 10 T scores confirmed by RT-PCR in seven of the eight HPV tumors that (T) from within the window, then divided by the window length in were available for this analysis, including the L1-negative actual genomic distance. Criteria used for identifying differential flag sample (300171), which also showed E6 expression. Patients + regions were a normalized T score >2 SD from the mean of total with HPV tumors were on average younger than those with normalized T score (in this case, the cutoff value is 4.9) for all sliding HPV tumors (median age, 49 versus 58 years), although this windows. difference did not reach statistical significance (P = 0.08, Confirmation of expression data by real-time PCR. Total RNA from Wilcoxon rank sum test; Table 1). HPV was significantly + seven of the eight HPV tumors was available for real-time PCR (RT- overrepresented in tumors originating from the oropharynx PCR) analysis, whereas an equal number of seven RNAs were chosen (seven of eight HPV+ tumors), whereas HPV was also from the HPV tumors for comparisons of expression levels. Fifty observed in one of the nine tumors originating from the nanograms of total RNA were amplified using the NuGEN WT-Ovation RNA Amplification kit (NuGEN; ref. 14). The amplified cDNA was larynx. None of the 15 oral cavity tumors analyzed harbored cleaned using the Qiagen PCR purification kit (Qiagen). Five genes HPV. No significant differences were observed with respect to among the 91 statistically significant genes were analyzed by RT-PCR: race, tobacco use, alcohol use, clinical and pathologic stage, TAF7L, CDKN2A, SYCP2, RFC4, and NAP1L2 using Applied Biosystems or tumor differentiation. Taqman FAM labeled probes (Applied Biosystems, Foster City, CA). An Gene expression differences with HPV status in HNSCC additional RT-PCR assay was done to test for HPV16-E6 expression in tumors. To identify the genes that were differentially expressed + seven of the eight HPV tumors. The endogenous genes 18S, PPIA, and between the eight HPV+ and 28 HPV tumors, statistical GUSB were used as internal calibration standards. The average of these analysis using HPV status as the supervising variable was done three internal genes was used to normalize the RT-PCR results from (15). Among the f25,000 genes on the DNA microarray, 91 the set of five genes and HPV16-E6. Twenty-five nanograms of differentially expressed genes were highly statistically significant amplified cDNA were used per reaction, and the probes were obtained with a false-discovery rate (FDR) of <0.3% in classifying HPV+ from Applied Biosystems. Analysis of each sample was done in quadruplicate on an Applied Biosystems 7900HT instrument (Applied versus HPV tumors (Fig. 2; Supplementary Data). Thus, <1 of Biosystems). the 91 genes (<1.1%) in this list is expected to make the cutoff Statistical analyses. Descriptive statistics were generated and tested to be classified as differentially expressed due to a chance event. with Fisher’s Exact and Wilcoxon rank sum tests using the SAS/STAT The accuracy of the 91-gene set to predict HPV status by

Fig. 1. HPV analysis in HNSCC samples by amplification of E6 and h-globin. Positive control cell lines (lanes1-3); negative controls, no DNA added (lanes 22 and 23). SiHa (lane 1), HeLa (lane 2), CaSki (lane 3), size standard (lanes 4 and 21), DNA samples derived from HNSCC tissue (lanes 5-20). Patient samples (lanes 5, 8,12,and18 ) show the shorter amplification product of HPV E6. The h-globin amplification product was visualized in all samples, indicating the integrity of the starting DNA.

www.aacrjournals.org 703 Clin Cancer Res 2006;12(3) February 1, 2006 Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. Human Cancer Biology

Fig. 2. Cluster diagram of 91genes that are differentially expressed between HPV+ and HPV HNSCC tumors. HPV+ tumors form a separate cluster (right).

Clin Cancer Res 2006;12(3) February1,2006 704 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. HPV and Gene Expression in HNSCC

Ta b l e 2 . Named genes from the 91top classifiers for HPV status by SignificanceAnalysis of Microarrays

Genes with higher expression in HPV+ tumors than HPV tumors Significance Analysis of Microarrays Rank HUGO ID Chromosome Description 1 TCAM1 17q22 Testicular cell adhesion molecule1 2 AL833646 2q21 Unknown protein 3 TAF7L 5q31 TAF7-like RNA polymerase II 4 SYNGR3 16p13 Synaptogyrin 5 CDKN2A 9p21 Cyclin-dependent kinase inhibitor 2A (p16INK4A) 6 FLJ39749 3q29 Unknown protein 7 FLJ37881 16 Unknown protein 8 RPA2 1p35 Replication protein A2 9 MYNN 3q26 Myoneurin 10 FLJ31952 17q21 Unknown protein 11 RIBC2 22q13 RIB43A domain with coiled-coils 2 12 FLJ4628 22q13 Unknown protein 13 BF055370 7Unknownprotein 14 MCM6 2q21 Minichromosome maintenance deficient 6 15 FLJ42662 XUnknownprotein 16 RFC4 3q27 Replication factor C4 17 NR1D2 3p24 Nuclear receptor subfamily1,group D2 18 MGC24665 16p13 Unknown protein 19 EHHADH 3q26 Enoyl-CoA, hydratase 20 FKSG14 5p15 Leucine zipper protein

Genes with lower expression in HPV+ tumors than HPV tumors

1 NAP1L2 Xq Nucleosome assembly protein1-like 2 2 KIRREL 1q21-25 Kin of IRRE- like (nephrin related) leave-one-out cross-validation was 100%. Of the 91 genes, 89 this cluster according to their expression difference between genes had higher average expression levels in the HPV+ tumors, HPV+ and HPV tumors (Table 3). whereas only two genes were on average expressed at lower RNA expression analysis by RT-PCR. Microarray expression levels in the HPV+ tumors (Table 2; Supplementary Data). One results for a subset of differentially expressed genes and HPV of the most significant differentially expressed genes was cyclin- DNA results were confirmed by separate RT-PCR analyses. Seven dependent kinase inhibitor 2A (CDKN2A), which encodes the of the eight HPV+ tumors were available for RT-PCR analysis, p16INK4A tumor suppressor protein. Other genes with higher and all tested positive for E6 expression as measured by RT-PCR, expression in the HPV+ group were cell cycle regulators other whereas none of seven selected HPV tumors expressed detect- than p16INK4A (p18 and CDC7), transcription factors (TAF7L, able levels of HPV16-E6 (see Supplementary Fig. 1). Further- RFC4, RPA2,andTFDP2), the cell adhesion molecule more, expression levels of CDKN2, CDKN2A, TAF7L, SYCP2, (TCAM1), and several sequences defined by expressed sequence and RFC4, which were increased in HPV+ tumors with FDR < tags only (Table 2; Supplementary Data). The genes that had 0.3%, as shown by microarray analyses, were also significantly lower expression in the HPV+ group were NAP1L2 and KIRREL. increased in the RT-PCR analysis (see Supplementary Fig. 1). On Data from this study were deposited in the NIH Gene average, TAF7L expression was increased 220-fold in HPV+ Expression Omnibus database under accession no. GSE3292. tumors [P = 0.0001; 95% confidence interval (95% CI), 28- Gene expression cluster on chromosome 3q. HNSCCs harbor 1,744], SYCP2 increased 56-fold (P < 0.0001; 95% CI, 16-199), a large number of cytogenetic changes, but it is not known if RFC4 increased 3.4-fold (P = 0.019; 95% CI, 1.3-9), and CDKN2 any of these are associated with HPV status of the tumor and (p16INK4A) increased 35-fold (P = 0.006; 95% CI, 3-358). Of differential expression of the genes. To identify potential these genes, RFC4 is located on the chromosome 3 region cytogenetic gains or losses using gene expression differences, identified by DIGMAP (3q27.3). Decreased expression was found we aligned the known chromosomal location of all genes to for NAP1L2, although the results did not reach statistical their expression levels using DIGMAP (11). Clusters of genes significance (7-fold decreased, P = 0.285; 95% CI, 0.2-286). across the full genome were then tested for expression differ- + ences between the HPV and HPV groups. A graphical Discussion representation of this analysis is shown in Fig. 3. The most significant cluster of genes with increased expression in HPV+ New data on the effects of HPV infection as a possible tumors was observed on the long arm of chromosome 3. The etiologic agent in HNSCC has emerged in recent years, region ranged from 3q24-ter, with several subclusters of genes indicating that HPV+ tumors may represent a distinct subset with increased expression levels. We then ranked all genes in that can be identified by specific molecular features (8, 9).

www.aacrjournals.org 705 Clin Cancer Res 2006;12(3) February 1, 2006 Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. Human Cancer Biology

However, a comprehensive analysis of the gene expression is impossible to rule out confounding effects due to tumor differences between HPV+ and HPV HNSCCs has not been site. However, in our previous molecular classification study reported. In this article, we determined global gene expression of 60 HNSCCs, we did not find a specific gene expression in eight HPV+ and 28 HPV tumors and compared the two pattern based on the anatomic subsite (12). Further studies groups to identify differences in gene expression patterns. examining larger numbers of HPV+ HNSCCs from multiple Patient characteristics. In terms of demographics, our study subsites will be needed to clarify this issue. However, it agreed with previously published results showing that HPV is is also reassuring that the current gene list based on mostly associated with tumors from the oropharynx, and that difference in HPV status identifies previously identified patients with HPV+ tumors are generally younger than those HPV-specific features in uterine cervical carcinomas, such with HPV tumors (7, 22). In our patient population, seven of as increased expression of CDKN2A (p16INK4A), RFC4, the eight HPV+ tumors originated in the oropharynx, and MCM2, LIG1, and TFDP2, and decreased expression of patients with HPV+ tumors were significantly younger than NAP1L2 (23). We also found overrepresentation of highly those with HPV tumors. We were unable to determine any expressed genes on chromosome 3q24-qter. Increased copy (negative) association with smoking because of the relatively numbers of this chromosomal region were previously small size of our sample and the fact that all patients had reported in association with positive HPV status (5). For some history of tobacco exposure. From these observations, we these reasons, we believe that the observed gene expression conclude that our patient cohort is representative of the general differences are based on the presence or absence of HPV HNSCC patient population. rather than on any hypothetical differences between oropha- Gene expression analysis. The supervised analyses of the ryngeal carcinomas and other upper airway squamous cell gene expression data showed 91 genes that were differentially carcinomas. expressed in HPV+ and HPV tumors. Because all but one Among the most prominent genes up-regulated in HPV+ of the HPV+ HNSCCs analyzed arose in the oropharynx, it tumors, several are involved in transcription and cell cycle regulation. One such cell cycle inhibitor is CDKN2A, which encodes the p16INK4A tumor suppressor protein. p16INK4A is a cyclin-dependent kinase inhibitor in the Rb tumor suppressor pathway. Increased expression of p16INK4A may potentially reflect loss of a negative feedback loop associated with inactivation of the retinoblastoma tumor suppressor protein (pRB) by HPV E7. Increased levels of p16INK4A are strongly correlated with HPV infection in cervical and in head and neck carcinomas (8, 24). Overexpression of p16INK4A is highly correlated with the presence of HPV and has been used as a surrogate marker for HPV (24). One of the most significantly up-regulated genes was TAF7L, a TATA box binding protein–associated factor, an X-linked gene that is normally only expressed in testis and spermatogonia (25). TAF7L plays a role in regulation of the transcription factor IID during male germ cell differentiation, where it replaces TAF7 in the transcription factor IID protein complex (26). This complex is required for transcription by promoters targeted by RNA polymerase II. TAF7 has been implicated in activator protein transcription regulation as a possible cofactor that binds to c-Jun (27). Expression of the 32-kDa subunit of replication protein A (RPA), encoded by RPA2, was also increased in HPV+ tumors. RPA is a tetrameric protein complex that binds ssDNA and facilitates DNA damage sensing (28). The HPV E1 protein, an ATP-dependent viral DNA helicase, binds and recruits RPA to sites of DNA replication, through binding of the 70-kDa subunit of RPA (RPA1; ref. 29). In the microarray analysis, there were only two genes significantly down-regulated in HPV+ tumors: NAP1L2 and KIRREL (Table 2). However, RT-PCR analysis of NAP1L2 did not reach statistical significance, presumably because only two of the seven HPV tumors showed high NAP1L2 expression levels compared with the HPV+ tumors (see Supplementary Fig. 1). NAP1L2 encodes a member of the nucleosome Fig. 3. Analysis of gene expression in HPV-associated HNSCC across the human assembly protein (NAP) family with high expression levels in genome. A, normalizedT score, indicating areas of significantly higher number of brain and a possible role in neuronal cell cycle regulation genes with high levels of expression in HPV+ versus HPV tumors. B,detailedplotof chromosome 3, indicating the areas on the long arm (3q24-ter) with highly (30). KIRREL (NEPH1) is a nephrin-related member of the significant scores. immunoglobulin superfamily involved in cell-cell interactions

Clin Cancer Res 2006;12(3) February1,2006 706 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. HPV and Gene Expression in HNSCC

Ta b l e 3 . List of genes on chromosome 3q with high expression levels in HPV+ tumors

HUGO ID Description Significance Analysis of Microarrays rank 3q band Statistically significant expression difference with FDR <0.3% FLJ39749 Hypothetical protein 6 29 MYNN Myoneurin zinc finger protein 9 26.2 RFC4 Replication factor C4 16 27.3 EHHADH Enoyl-CoA 19 27.2 MGC15397 Hypothetical protein 21 28 TFDP2 Transcription factor Dp-2 (E2F dimerization partner 2) 30 23 MGC64882 Hypothetical protein 32 29 NDUFB5 NADH dehydrogenase1 hsubcomplex, 5 33 26.3 OPA1 Optic atrophy1 38 28-29 RAP2B Ras-related GTPase 47 25.2 FLJ35794 Hypothetical protein 49 29 HES1 Hairy and enhancer of split 1 50 29 FLJ10560 Hypothetical protein 54 28 SMARCA3 Actin-dependent regulator of chromatin 60 25.1 EIF2B5 Subunit 5q for eukaryotic translation initiation factor 2B 61 27.1 ACTL6 Actin-like protein 6A 65 26.3 PDCD10 Programmed cell death10 67 26.2 POLR2H Polymerase (RNA) II (DNA directed) polypeptide H 77 28

Statistically significant expression difference with FDR <1%

ECT2 Epithelial cell transforming sequence 2 oncogene 96 26.3 PPP1R2 Protein phosphatase1,regulatory (inhibitor) subunit 2 108 29 MCCC1 Methylcrotonoyl-CoA carboxylase1 (a)11027 SHOX2 Short stature homeobox 2 113 25.3 MBNL1 Muscle blind-like 114 25 SFRS10 Splicing factor, arginine/serine-rich10 116 26.2-27 DVL3 Dishevelled, dsh homologue 3 146 27 ABCC5 ATP-binding cassette, subfamily C, member 5 175 27.1 MAP3K13 Mitogen-activated protein kinase kinase kinase13 176 27 BDH 3-Hydroxybutyrate dehydrogenase 179 29 DRE1 DRE1protein 190 27.3 EIF4G1 Eukaryotic translation initiation factor 4 g,1 206 27-ter PIK3CA Phosphoinositide-3-kinase, catalytic, a polypeptide 211 26.3 NCBP2 Nuclear cap binding protein subunit 2 225 29 GMPS Guanine monophosphate synthetase 231 24 RP42 RP42 homologue 239 26.3 LOC442100 Hypothetical protein DKFZp434A128 244 27.2

related to selective ultrafiltration by podocytes in the kidney indicates that either the expression levels are greatly increased (31). However, expression is not limited to the kidney, and in the tumors with chromosome gains on 3q, or that gene KIRREL mRNA is also found in the colon, pancreas, heart, transcription is elevated in the tumors with unaltered DNA and spleen (32). It is currently unclear what role a down- copy numbers. The 3q region identified by DIGMAP includes regulation of expression of these two genes would have in a known amplicon at 3q26 that has been described in HPV-mediated carcinogenesis. HNSCC and lung carcinomas (33–35). Candidate oncogenes DIGMAP. Assignment of expression analyses with known within this amplicon are PIK3CA (35, 36), SCRO (37), the chromosomal location of genes using DIGMAP revealed a RNA component of telomerase (TERC; ref. 38), ZNF639 cluster on chromosome 3q24-qter with a large number of (39), and p63 (40). Of these genes, SCRO and TERC are not genes with higher levels of expression in HPV+ versus HPV represented on the Affymetrix Human 133U Plus 2.0 tumors. Interestingly, a prior study using comparative GeneChip. Amplification and overexpression of PIK3CA has genome hybridization on 25 primary tonsillar carcinomas been reported for HNSCC (35) and cervical carcinomas (41), suggested of a gain on chromosome 3q24-ter in HPV+ although PIK3CA amplification has not been associated with tumors compared with HPV tumors (5). The observed HPV infection (42) and did not reach statistical significance expression difference in our series is highly significant, which in the present study.

www.aacrjournals.org 707 Clin Cancer Res 2006;12(3) February 1, 2006 Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. Human Cancer Biology

The other differentially expressed genes within chromo- Through global gene expression analyses of HNSCC somal band 3q26 (MYNN, NDUFB5, and ECT2) have not tumors, we have identified differences in gene expression previously been associated with amplification of this region. profile of HPV+ and HPV tumors and gained insight into ECT2 belongs to the dbl family of oncogenes that encode the biological effect of HPV infection in HNSCC. The most large, growth-regulatory molecules with guanine nucleotide prominent genes up-regulated in HPV+ tumors were involved exchange factor activity of towards Rho-family GTPases (43). in transcription and cell cycle regulation. These genes are Deregulation and mislocation of ECT2 has been linked to interesting from a clinical targeting standpoint because some malignant transformation in mouse 3T3 cells. Other genes in have already been identified for potential therapeutic benefit the 3q24-ter region identified by DIGMAP do not localize to and novel agents have been developed that are currently in the 3q26 amplicon, which implies that the difference in gene clinical trials. Examples are transcription inhibitors, such as expression profiles between HPV+ and HPV tumors can not histone deacetylase inhibitors (SAHA, MS-275, etc.; ref. 46), be fully explained by just amplification at 3q26. Furthermore, and cell cycle inhibitors, such as CDK and Chk kinase the differentially expressed genes in this chromosomal region inhibitors (flavopiridol, UCN-01, 17-AAG, etc.; ref. 47). were mainly involved in DNA replication and transcription, Combined with previous analyses of chromosomal imbal- such as replication factor C4 (RFC4), TFDP2, POLR2H, ance, analyses such as these may help to further identify SHOX2, and SFRS10. RFC4 encodes the 37-kDa subunit that potential therapeutic targets in altered chromosomal regions. is part of the heteropentamer RFC (also called activator 1), Additionally, gene expression changes driven by HPV status an accessory protein required for the elongation of primed independent of chromosomal abnormalities may also iden- DNA templates by DNA polymerase y and q (44). Prolifer- tify cellular mediators of HPV oncogenesis. ating cell nuclear antigen is also part of the elongation complex. Proliferating cell nuclear antigen is a target of Acknowledgments retinoblastoma signaling, but RFC does not respond to this pathway and is not responsive to E2F signaling (45). We thank Dr. A.L. George for his advice on this project.

References 1. Mork J, Lie AK, Glattre E, et al. Human papillomavi- 13. Resnick RM, Cornelissen MT, Wright DK. Detec- (HPV) DNA detection. Am J Surg Pathol 2004;28: rus infection as a risk factor for squamous-cell carci- tion and typing of human papillomavirus in archival 16 0 ^ 7. noma of the head and neck. N Engl J Med 2001;344: cervical cancer specimens by DNA amplification 25. Wang PJ, McCarrey JR,Yang F, Page DC. An abun- 112 5 ^ 3 1. with consensus primers. J Natl Cancer Inst 1990; dance of X-linked genes expressed in spermatogonia. 2. Dai M, Clifford GM, le Calvez F, et al. Human papillo- 82:1477^ 84. Nat Genet 2001;27:422 ^ 6. mavirus type 16 and TP53 mutation in oral cancer: 14. Kurn N, Chen P, HeathJD, Kopf-Sill A, Stephens KM, 26. Pointud JC, Mengus G, Brancorsini S, et al.The in- matched analysis of the IARC multicenter study. Wang S. Novel isothermal, linear nucleic acid amplifi- tracellular localisation of TAF7L, a paralogue of tran- Cancer Res 2004;64:468 ^ 71. cation systems for highly multiplexed applications. Clin scription factorTFIID subunitTAF7, is developmentally 3. zur Hausen H. Papillomaviruses causing cancer: eva- Chem 2005;51:1973 ^ 81. regulated during male germ-cell differentiation. J Cell sion from host-cell control in early events in carcino- 15. Tusher VG,Tibshirani R, Chu G. Significance analysis Sci 2003;116:1847 ^ 58. genesis. JNatl Cancer Inst 2000;92:690 ^ 8. of microarrays applied to the ionizing radiation re- 27. Munz C, Psichari E, Mandilis D, et al.TAF7 (TAFII55) 4. McKaig RG, Baric RS, Olshan AF. Human papilloma- sponse. Proc Natl Acad Sci U S A 2001;98:5116^21. plays a role in the transcription activation by c-Jun. virus and head and neck cancer: epidemiology and 16. Hedenfalk I, Duggan D, Chen Y, et al. Gene-expres- J Biol Chem 2003;278:21510 ^ 6. molecular biology. Head Neck 1998;20:250 ^ 65. sion profiles in hereditary breast cancer. N Engl J Med 28. Zou L, Elledge SJ. Sensing DNA damage through 5. Dahlgren L, Mellin H, Wangsa D, et al. Comparative 2001;344:539 ^ 48. ATRIP recognition of RPA-ssDNA complexes. Science genomic hybridization analysis of tonsillar cancer 17. Tukey JW. Tightening the clinical trial. Control Clin 2003;300:1542^8. reveals a different pattern of genomic imbalances in Trials 1993;14:266 ^ 85. 29. HanY, LooYM, Militello KT, MelendyT. Interactions human papillomavirus-positive and -negative tumors. 18. Edelman DB, Meech R, Jones FS. The homeodo- of the papovavirus DNA replication initiator proteins, Int J Cancer 2003;107:244 ^9. main protein Barx2 contains activator and repressor bovine papillomavirus type 1 E1and simian virus 40 6. Lundberg AS, Randell SH, Stewart SA, et al. Immor- domains and interacts with members of the CREB large T antigen, with human replication protein A. talization and transformation of primary human airway family. J Biol Chem 2000;275:21737 ^ 45. JVirol1999;73:4899^907. epithelial cells by gene transfer. Oncogene 2002; 19. Eisen MB, Spellman PT, Brown PO, Botstein D. 30. Rogner UC, Spyropoulos DD, Le Novere N, 21:4577 ^ 86. Cluster analysis and display of genome-wide expres- Changeux JP, Avner P. Control of neurulation by the 7. Haraf DJ, Nodzenski E, Brachman D, et al. Human sion patterns. Proc Natl Acad Sci U S A 1998;95: nucleosome assembly protein-1-like 2. Nat Genet papilloma virus and p53 in head and neck cancer: clin- 14 8 6 3 ^ 8. 2000;25:431 ^ 5. ical correlates and survival. Clin Cancer Res 1996;2: 20. Saeed AI, SharovV,WhiteJ, et al.TM4: a free, open- 31. Sellin L, HuberTB, Gerke P, Quack I, Pavenstadt H, 755 ^ 62. source system for microarray data management and Walz G. NEPH1defines a novel family of podocin inter- 8. Olshan AF,Weissler MC, Pei H, et al. Alterations of the analysis. Biotechniques 2003;34:374^ 8. acting proteins. FASEB J 2003;17:115^ 7. p16 gene in head and neck cancer: frequency and as- 21. Livak KJ, Schmittgen TD. Analysis of relative gene 32. Donoviel DB, Freed DD,Vogel H, et al. Proteinuria sociation with p53, PRAD-1 and HPV. Oncogene expression data using real-time quantitative PCR and and perinatal lethality in mice lacking NEPH1, a novel 19 97;14 : 811 ^ 8. the 2(Delta Delta C(T)) method. Methods 2001;25: protein with homology to NEPHRIN. Mol Cell Biol 9. Braakhuis BJ, Snijders PJ, Keune WJ. Genetic pat- 402^8. 2001;21:4829^ 36. terns in head and neck cancers that contain or lack 22. Gillison ML, Koch WM, Capone RB, et al. Evidence 33. Speicher MR, Howe C, Crotty P,du Manoir S, Costa transcriptionally active human papillomavirus. J Natl fora causalassociationbetweenhumanpapillomavirus J, Ward DC. Comparative genomic hybridization Cancer Inst 2004;96:998 ^ 1006. and a subset of head and neck cancers. J Natl Cancer detects novel deletions and amplifications in head and 10. Hoffmann M, GoroghT, Gottschlich S, et al. Human Inst 2000;92:709 ^ 20. neck squamous cell carcinomas. Cancer Res 1995;55: papillomaviruses in head and neck cancer: 8 year- 23. Santin AD, Zhan F, Bignotti E, et al. Gene expression 1010 ^ 3. survival-analysis of 73 patients. Cancer Lett 2005; profiles of primary HPV16- and HPV18-infected early 34. Balsara BR, Sonoda G, du Manoir S, Siegfried JM, 218:199 ^ 206. stage cervical cancers and normal cervical epithelium: Gabrielson E,Testa JR. Comparative genomic hybrid- 11. Yi Y, Mirosevich J, ShyrY, Matusik R, George AL, Jr. identification of novel candidate molecular markers for ization analysis detects frequent, often high-level, Coupled analysis of gene expression and chromosom- cervical cancer diagnosis and therapy.Virology 2005; overrepresentation of DNA sequences at 3q, 5p, 7p, al location. Genomics 2005;85:401 ^ 12. 331:269 ^ 91. and 8q in human non-small cell lung carcinomas. 12. Chung CH, ParkerJS, Karaca G, et al. Molecular clas- 24. Ansari-Lari MA, Staebler A, Zaino RJ, Shah KV, Cancer Res 1997;57:2116^ 20. sification of head and neck squamous cell carcinomas Ronnett BM. Distinction of endocervical and endo- 35. Redon R, Muller D, Caulee K, Wanherdrick K, using patterns of gene expression. Cancer Cell 2004; metrial adenocarcinomas: immunohistochemical p16 Abecassis J, du Manoir S. A simple specific pattern of 5:489^500. expression correlated with human papillomavirus chromosomal aberrations at early stages of head and

Clin Cancer Res 2006;12(3) February1,2006 708 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. HPV and Gene Expression in HNSCC

neck squamous cell carcinomas: PIK3CA but not p63 ZASC1encoding a Kruppel-like zinc finger protein as lignant transformation. J Biol Chem 2004;279: gene as a likely target of 3q26-qter gains. Cancer Res a novel target for 3q26 amplification in esophageal 7169 ^ 79. 2001;61:4122^ 9. squamous cell carcinomas. Cancer Res 2003;63: 44. LeeSH,KwongAD,PanZQ,HurwitzJ.Studies 36. Singh B, Gogineni SK, Sacks PG, et al. Molecular 5691^6. on the activator 1protein complex, an accessory fac- cytogenetic characterization of head and neck squa- 40. Massion PP,Taflan PM, Jamshedur Rahman SM, tor for proliferating cell nuclear antigen-dependent mous cell carcinoma and refinement of 3q amplifica- et al. Significance of p63 amplification and overex- DNA polymerase delta. J Biol Chem 1991;266: tion. Cancer Res 2001;61:4506 ^13. pression in lung cancer development and prognosis. 594^602. 37. Estilo CL, O-Charoenrat P, Ngai I, et al. The role of Cancer Res 2003;63:7113^ 21. 45. Angus SP, Mayhew CN, Solomon DA, et al. RB novel oncogenes squamous cell carcinoma-related 41. MaYY,Wei SJ, LinYC, et al. PIK3CA as an oncogene reversibly inhibits DNA replication via two temporal- oncogene and phosphatidylinositol 3-kinase p110a in cervical cancer. Oncogene 2000;19:2739 ^ 44. ly distinct mechanisms. Mol Cell Biol 2004;24: in squamous cell carcinoma of the oral tongue. Clin 42. Zhang A, Maner S, Betz R, et al. Genetic altera- 5404^20. Cancer Res 2003;9:2300 ^ 6. tions in cervical carcinomas: frequent low-level 46. Johnstone RW, Licht JD. Histone deacetylase inhib- 38. Yokoi S, Yasui K, Iizasa T, Imoto I, Fujisawa T, amplifications of oncogenes are associated with hu- itors in cancer therapy: is transcription the primary tar- Inazawa J. TERC identified as a probable target with- man papillomavirus infection. Int J Cancer 2002;101: get? Cancer Cell 2003;4:13^ 8. in the 3q26 amplicon that is detected frequently in 427 ^ 33. 47. Schwartz GK. Development of cell cycle active non-small cell lung cancers. Clin Cancer Res 43. Saito S, Liu XF, Kamijo K, et al. Deregulation and drugs for the treatment of gastrointestinal cancers: 2003;9:4705 ^ 13. mislocalization of the cytokinesis regulator ECT2 a new approach to cancer therapy. J Clin Oncol 39. Imoto I, Yuki Y, Sonoda I, et al. Identification of activate the Rho signaling pathways leading to ma- 2005;23:4499^508.

www.aacrjournals.org 709 Clin Cancer Res 2006;12(3) February 1, 2006 Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. Gene Expression Differences Associated with Human Papillomavirus Status in Head and Neck Squamous Cell Carcinoma

Robbert J.C. Slebos, Yajun Yi, Kim Ely, et al.

Clin Cancer Res 2006;12:701-709.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/12/3/701

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2006/02/21/12.3.701.DC1

Cited articles This article cites 47 articles, 23 of which you can access for free at: http://clincancerres.aacrjournals.org/content/12/3/701.full#ref-list-1

Citing articles This article has been cited by 29 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/12/3/701.full#related-urls

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/12/3/701. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.