TITLE Signaling and Transcriptional Changes Critical ...

TITLE

Signaling and Transcriptional Changes Critical for Transformation of Human Cells by SV40 Small Tumor Antigen or PP2A B56g Knockdown

Carlos S. Moreno1,2,6, Sumathi Ramachandran1,2,6, Danita G. Ashby2,3, Noelani Laycock1,2, Courtney A. Plattner2,3, Wen Chen4, William C. Hahn4,5, and David C. Pallas2,3

1

Department of Pathology & Laboratory Medicine, Emory University School of Medicine, Atlanta, GA

30322 2

Winship Cancer Institute, Emory University, Atlanta, GA 30322

3

Department of Biochemistry, Emory University School of Medicine, Atlanta, GA 30322

4

Department of Medical Oncology, Dana-Farber Cancer Institute, and Departments of Medicine,

Brigham and Women’s Hospital and Harvard Medical School, Boston, MA 02115 5

Broad Institute, Cambridge, MA 02139

6

These authors contributed equally to this work

Correspondence: [email protected], [email protected]

Running Title: Transformation targets of SV40 ST and B56g Keywords: viral transformation, phosphatases, signal transduction, functional genomics, cell-matrix interactions Footnotes: Supplementary material is available at http://morenolab.whitehead.emory.edu/pubs/ST/

Abstract One set of genes sufficient for transformation of primary human cells utilizes the combination of HaRas-V12, the telomerase catalytic subunit hTERT, SV40 Large T antigen (LT), and SV40 Small T antigen (ST). While SV40 LT inactivates the retinoblastoma (Rb) protein and p53, the contribution of ST is poorly understood. The essential helper function of ST requires a functional interaction with Protein Phosphatase 2A (PP2A). Here we have identified changes in gene expression induced by ST, and show that ST mediates these changes through both PP2A-dependent and PP2A-independent mechanisms. Knockdown of PP2A B56g subunit can substitute for ST expression to fully transform cells expressing LT, hTERT, and Ras-V12. We also identify those genes affected similarly in two cell lines that have been fully transformed from a common parental line by two alternative mechanisms, namely ST expression or PP2A B56g subunit knockdown. ST altered expression of genes involved in proliferation, apoptosis, integrin signaling, development, immune responses, and transcriptional regulation. ST reduced surface expression of MHC Class I molecules, consistent with a need for SV40 to evade immune detection. ST expression enabled cell cycle progression in reduced serum and src phosphorylation in anchorage-independent media, whereas B56g knockdown required normal serum levels for these phenotypes. Inhibitors of integrin and src signaling prevented anchorage-independent growth of transformed cells, suggesting that integrin and src activation are key ST-mediated events in transformation. Our data support a model in which ST promotes survival through constitutive integrin signaling, src phosphorylation, and NF-kB activation, while inhibiting cell-cell adhesion pathways.

2

Introduction The study of DNA tumor viruses and the oncogenes that they encode has provided many critical insights into the essential steps for oncogenesis. Simian Virus 40 (SV40) is a potently oncogenic DNA virus that can cause tumors in rodents (1) and possibly humans (2). The two oncogenic proteins responsible for the transforming activity of SV40 are known as the Large Tumor Antigen (LT) and Small Tumor Antigen (ST). The early region of the SV40 genome encodes the 82 kD LT and the 20 kD ST as two alternatively spliced protein products that share a common amino terminal sequence of 82 amino acids with similarity to the DnaJ family of molecular chaperones (3). LT also contains an LXCXE motif in the central region of the protein that enables it to bind and inactivate members of the Rb-family of tumor suppressor proteins (4), helping to release cells from G1 arrest. The amino terminal DnaJ domain is also necessary to functionally inactivate Rb-family tumor suppressors (5). In addition, the C-terminal domain of LT binds and inactivates the p53 tumor suppressor (6). The unique C-terminal half of ST encodes a domain that enables binding to Protein Phosphatase 2A (PP2A), a heterotrimeric serine-threonine phosphatase that plays important roles in numerous cellular processes including apoptosis, cell cycle regulation, and signal transduction (7, 8).

ST

provides critical helper functions to LT for transformation of both mouse (9) and human cells (10). In addition, ST binding to PP2A is essential for ST to provide its helper function for transformation of human cells (11-13). Efforts to delineate a defined set of genetic changes essential for transformation of primary human cells has demonstrated that one combination of genes sufficient to produce anchorageindependent growth in soft agar and tumors in nude mice includes SV40 LT and ST, the catalytic subunit of human telomerase (hTERT), and the constitutively activated V12 mutant of H-Ras (H-RasV12) (11, 12, 14).

A C-terminal deletion mutant, ST110, that encodes only the first 110 residues of

3

ST, including the DnaJ domain, cannot bind to PP2A, and cannot provide the essential helper function needed to transform human cells (12). To investigate the ST-induced changes in gene expression that are essential for human tumor formation, we used whole genome expression profiling to compare the expression patterns of four human cell lines. Each of these cell lines stably expresses hTERT, H-Ras-V12, and LT (12). In addition to these three stably expressed genes, the tumorigenic HEK-TERST cell line expresses wildtype SV40 ST, while the non-tumorigenic HEK-TERST110 and HEK-TERV cell lines express the ST110 mutant or vector alone, respectively (12). Introduction of an antisense construct to the B56g3 subunit of PP2A into the HEK-TER cell line (HEK-TERASB56g) nearly completely suppresses B56g expression at the protein level, permits anchorage independent growth, and enables tumor formation in nude mice in a manner similar to cells expressing ST (15). Here, we show that introduction of ST or suppression of PP2A B56g subunits impacts expression of a small subset of genes involved in apoptosis, integrin signaling, transcriptional regulation, and cytoskeletal control.

These gene expression and

signaling changes may promote growth and oppose apoptotic signals that prevent growth of normal cells in soft agar, enabling anchorage-independent growth and tumor formation.

Materials and Methods Cell Lines, Culture, & Lysates Stable human embryonic kidney (HEK) cell lines HEK-TERST, HEK-TERST110, and HEK-TERV have been previously described (12). The HEK-TERASB56g cell line is described elsewhere (15). Cells were grown in a-MEM, 10% FBS, 2 mM L-glutamine, 100 U/ml penicillin/streptomycin, and serum-starved in a-MEM, 0.1% FBS, 2 mM L-glutamine, 100 U/ml penicillin/streptomycin for 24 hrs.

4

Western Blots & Antibodies Cell lysates were prepared as previously described (16) or with the additional step of sonication prior to centrifugation to generate a whole cell lysate. Antibodies to SV40 LT (mAb 419), ST, H-Ras-V12, SERPINB2/PAI-2, plakoglobin (all from Santa Cruz Biotech, Santa Cruz, CA), IQGAP-2 (Upstate USA Inc, Lake Placid, NY), thymidine kinase (QED Bioscience, San Diego, CA), cyclin A (BD Transduction Laboratories, Lexington, KY), BNIP3, cyclin B (Oncogene Research, San Diego, CA), Src-phosphoY418 (Biosource Intl, Camarillo, CA), survivin and IkBa (Cell Signaling Technology, Beverly, MA), were used in immunoblots as previously described (16). The src 327 monoclonal antibody was a generous gift from Joan Brugge, Ariad Pharmaceuticals, Inc.

Fluorescent-Activated Cell Sorting (FACS) Analysis For analysis of MHC Class I expression, 106 cells were harvested in PBS, 0.02 % EDTA, incubated in 500 ml of PBS, 10 mg/ml W6/32 mouse mAb (a generous gift of Dr. Charles A. Parkos, Emory University), washed, and then stained with anti-mouse FITC conjugated secondary antibody. For DNA content analysis, cells were incubated in a-MEM, 0.1% FBS for 24 hrs, and harvested immediately or collected after 24 hrs in 10% FBS. For cell synchronization, 0.5 x106 cells were treated with 10 mM aphidicolin in 10% FBS for 24 hours, and then harvested immediately, or released into aMEM, 0.1% FBS without aphidicolin for 24 hrs. Cells were fixed in 70% ethanol at –20 °C overnight, stained with 10 mg/ml propidium iodide, and sorted on a FACScalibur sorter (Becton-Dickinson, Franklin Lakes, NJ).

5

Quantitative Real-time PCR Quantitative real-time PCR (QRT-PCR) was performed in an I-cycler (Bio-Rad, Hercules, CA) using SYBR Green (Molecular Probes, Eugene, OR). The critical cycle threshold (CT) was determined for each gene and the difference relative to the CT for glyceraldehyde-3-phosphate dehydrogenase (GAPDH), or DCT was computed for each RNA sample. Two independent RNA samples from each cell line were analyzed in quadruplicate and the mean and standard deviation were computed. Primers were designed using Primer Express software to amplify across splice junctions. The sequences of the primer sets used are given in Supplementary Table S7.

Microarrays and Data Analysis Total RNA was prepared using the RNeasy kit (Qiagen Inc., Valencia, CA) according to the manufacturer’s instructions. RNA was reverse transcribed and labeled probes were fragmented and hybridized to the Human Genome U133 Chip Set (Affymetrix Inc., Santa Clara, CA) according to the manufacturer’s protocols. Analysis of four cell lines in duplicate on both the U133A and U133B GeneChips resulted in sixteen microarray hybridizations that generated eight combined U133AB whole genome datasets (ST-1, ST-2, ST110-1, ST110-2, B56-1, B56-2, TERV-1, and TERV-2). Scanned images were analyzed and all twelve possible comparison files against the TERV-1 and TERV-2 datasets were generated using Microarray Suite 5.0 software. Each chip was normalized with a target value of 150.

Genes called absent in all hybridizations and genes that were called no change (NC) in

more than one ST-TERV Affymetrix comparison file were filtered out leaving 2545 probes for SAM analysis. Data from affymetrix CEL files was then normalized using the robust multiarray average (RMA) method (17). After data normalization, SAM analysis was performed on the remaining 2545 probe sets using the following relevant parameters: D = 0.26, fold-change = 1.5, number permutations =

6

1000, RNG seed = 1234567, median FDR = 3%, significant genes = 555, predicted false positives = 17. Hierarchical clustering with average linkage based on Euclidian distance was performed with Spotfire Decision Site 7.0 software.

Annotations were obtained from the NetAffx website

(http://www.affymetrix.com/analysis/index.affx) and the April 2003 assembly of the Human Genome at UCSC (http://genome.ucsc.edu).

P-values for GO categories were computed using GOstat (18) and

corrected to false discovery rate using the method of Benjamini and Hochberg (19).

Soft Agar Assays Soft Agar assays were performed essentially as described (14). Briefly, 10,000 cells were seeded in 0.3% Noble Agar and either 10 mg/ml RGD or RAD peptide (Biomol, Plymouth Meeting, PA) (20), 10 mM PP1 (21) or PP3 inhibitor (22), or 1 mM wortmannin. The c-src-specific kinase inhibitor PP1 (4amino-5-(4-methylphenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine) was obtained from Biomol (Plymouth Meeting, PA) and the inactive structural analog, PP3 (4-amino-7- phenylpyrazol[3,4-d]pyrimidine) was purchased from CalBiochem (San Diego, CA). Cell viability was determined using MTT dye staining (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide) as described (23).

Methylcellulose Assays Methylcellulose (MC) assays were performed as described (24). Briefly, MC culture medium (final concentration of 1.3% w/v) was made by diluting autoclaved 2.6% MC (Sigma-Aldrich, St. Louis, MO) in water with an equal volume of 2X concentrated MEM followed by stirring overnight at 4oC. Essentially, 3 x 106 cells were incubated in 10 ml of MC culture medium in a flask at 37oC. After incubation for 24 hours, the cells were recovered by solubilizing the MC medium with 5 volumes of chilled PBS followed by gently mixing and centrifugation at 1500 rpm for 5 mins.

7

Results SV40 ST alters the expression of 452 genes. We used a whole genomic profiling approach of ST-expressing (HEK-TERST), ST110 mutantexpressing (HEK-TERST110), PP2A B56g knockdown (HEK-TERASB56g), and vector control (HEKTERV) cell lines to investigate the transcriptional changes in transformation of human cells. Each of these cell lines is actually a pool of hundreds of individual clones derived from retroviral infection of a human embryonic kidney (HEK) cell line that stably expresses hTERT, H-Ras-V12, and LT (HEKTER) (12). We reasoned that high levels of serum might cause some of the same changes in gene expression as ST expression or B56g down-regulation. Therefore, all cell lines were grown in low serum for 24 hours to prevent serum effects from masking the impact of ST expression or B56g downregulation on global gene expression patterns. Total RNA was prepared from two independent biological replicates and analyzed using the Affymetrix U133 Human Genome GeneChip Array Set. To compensate for multiple testing issues, we have used the Significance Analysis of Microarrays (SAM) software (25), which computes false discovery rates (FDR’s). SAM analysis resulted in a total of 555 probe sets corresponding to 452 unique genes that exhibited at least 1.5 fold-change between HEKTERST and HEK-TERV cell lines with a predicted FDR of 3% (q < .03) (see Materials and Methods and Supplementary Table S1). Although a 1.5 fold change in mRNA levels may not be biologically significant for some genes, microarrays often underestimate the actual fold change when compared to QRTPCR and northern blots (26), and SAM analysis produces fewer false positives than fold change criteria (25). Moreover, a cutoff of 1.5 fold has been used in several studies (27, 28), and the data seen at the 1.5 fold level included genes such as Cyclin B which we confirmed at the protein level were induced much more than 1.5 fold.

8

We analyzed the gene ontology (GO) consortium annotations of the genes that were significantly affected by ST using the GOstat software (18) which finds GO terms that are statistically overrepresented in a gene list compared to the rest of the genome. GOstat analysis of the genes affected by ST expression found five major categories of genes that were significantly overrepresented relative to all annotated human genes using the Benjamini & Hochberg correction for FDR (Supplementary Table S2; Figures 1A and 1B). Of the 452 unique genes, only 274 had GO annotations (Figure 1A). Of the 274 annotated genes, 109 genes (or 40%) had GO annotations relating to cellular proliferation (Figure 1B) compared to 4622 of 19085 total genes (or 24%) in these categories for the entire genome (Supplementary Table S2; p = 2.55 x 10-7). The other major functional categories significantly overrepresented in the genes affected by ST expression included development and morphogenesis (p = 4.8 x 10-5), inflammation (p = 3.5 x 10-4), regulation of transcription (p = 6.2 x 10-4), and antigen presentation (p = 2.6 x 10-3). Although some of the 452 genes affected by ST such as osteopontin (29), and thymidine kinase (30, 31) agree with earlier studies, the vast majority of these ST targets have not been reported previously (Supplementary Table S1). ST appears to affect proliferation, possibly through changes in developmental programs that are controlled by sequence-specific transcription factors. Our data show for the first time that ST induces expression of the developmental HOX genes HOXA9, HOXB6, and HOXB3; the forkhead family factors FOXD1, FOXM1, and FOXG1b; and the cancer-associated developmental transcription factors c-MYC, ETS1, ID2, and SMAD3. Our observation that ST induces c-myc mRNA complements recent findings that PP2A inhibition stabilizes c-myc at the protein level, contributing to transformation (32). Moreover, several early developmental markers from multiple tissue types such as endothelial cell-specific molecule 1 (ESM1), epithelial membrane protein 1 (EMP1), stathmin-like 3 (STM3), osteopontin (SPP), and CD24 were induced by ST, suggesting that the

9

changed expression of these developmental transcription factors results in downstream differences between HEK-TERV and HEK-TERST cells. While changes in mRNA levels alone do not necessarily translate into phenotypic alterations, the coordinated alterations of so many genes involved in similar processes, the alterations in cell cycle progression reported here, and previous literature all support the hypothesis that ST-induced changes in mRNA expression likely affect these cellular functions. ST may also repress immune responses and impair antigen presentation by repressing expression of MHC class I molecules HLA-A, HLA-B, HLA-C, and b2-microglobulin (B2M); the CD74 invariant chain necessary for folding of MHC class II heterodimers; and pro-inflammatory cytokines such as IL-8, IL-1b, consistent with the need for SV40 to avoid immune detection. FACS analysis of MHC Class I expression in three independent experiments demonstrated that surface expression of HLA molecules was repressed in the HEK-TERST, HEK-TERST110, and HEK-TERASB56g cell lines compared to the HEK-TERV line (Figures 1C-1E). While MHC Class I expression was reduced in all three lines, the repression was most effective in the cells that express wild-type ST.

SV40 ST alters the expression of 171 genes in a PP2A-Independent Manner and 281 genes in a PP2A-Dependent Manner Expression data from the HEK-TERST110 and HEK-TERASB56g cell lines was also compared to the HEK-TERV data, and signal-log ratios for all cell lines were then analyzed. Comparison of the HEKTERST110 line with the HEK-TERV line identified 411 genes that were altered in expression levels (Supplementary Table S3). PP2A-independent genes were defined as the intersection of the 452 genes altered by ST expression and the set of 411 genes affected by ST110 expression that changed in the same direction relative to vector control cells. The intersection of these two gene sets resulted in a set of 171 genes that are represented by the magenta and cyan sections of the Venn diagram in Figure 2A. The

10

expression pattern of all 452 ST-regulated genes was analyzed by hierarchical clustering (Figure 2B), and the expression pattern of the 171 PP2A-independent genes is shown in Figure 2C. Strikingly, 37% of ST-regulated genes (Supplementary Table S4) can be regulated by the N-terminal domain of SV40 ST independently of PP2A modulation. Because these changes are induced by ST in cells that already express LT, these data suggest that the amino-terminal half of ST may have effects that LT, which also contains a DnaJ domain, does not. Given that the transforming helper function of ST is known to be PP2A-dependent, genes that are affected by ST in a PP2A-dependent manner are of particular interest. The 281 genes affected by ST that were not similarly affected by the ST110 mutant represent this PP2A-dependent gene set (red and orange sections of Figure 2A). Figure 2B shows that many of the genes affected by ST are unaffected by down-regulation of the PP2A B56g subunit in the HEK-TERASB56g cell line, or are even affected in the opposite direction. Thus, the subset of genes similarly affected by B56g antisense and ST expression (Figure 2D) are PP2A-dependent via two different molecular mechanisms that transform of human cells. The fact that knockdown of PP2A B56g subunits produced changes in 843 genes (green section of Figure 2A) that are not affected by ST or ST110 is not surprising, given the large number of cellular functions of PP2A. It is difficult to know what mechanism would affect the 161 genes represented in blue in Figure 2A that were affected by ST110 alone. One could speculate that the truncated N-terminal portion of ST could gain new functions that are not necessarily biologically relevant. More likely, the effects of the C-terminal portion of ST on PP2A activity could counteract or attenuate effects that are mediated by the N-terminal portion of ST. The 79 genes affected by both ST110 and B56g knockdown (but not by wild-type ST) are interesting (Supplementary Table S5), since several of them are altered in human cancers (AMACR, TWIST, TGFbR2, MMP-10) or are involved in the wnt signaling pathway (SFRP1, GSK3b, Frizzled). Two genes (SFRP1 and tropomysin 1) were repressed by ST110 and B56g

11

knockdown but induced by ST, while one gene (stem cell growth factor) was induced by ST110 and B56g knockdown but repressed by ST.

Western blotting and Real-Time PCR data confirm the microarray data To confirm our microarray observations, we prepared total RNA from independently treated samples and performed QRT-PCR on 29 PP2A-dependent genes and 12 PP2A-independent genes and compared expression changes relative to GAPDH. The magnitude of the changes observed in the QRT-PCR assays varied somewhat from that observed in the microarray experiments (Figures 3A-D and 3G-I). Nevertheless, the directionality of the QRT-PCR changes confirmed the microarray results for 40/41 (98%) of the genes assayed in the ST vs. TERV ratios, in 34/41 (83%) of the genes in the ST110 vs. TERV ratios, and in 22/25 (88%) of the genes in the ASB56g vs. TERV ratios. Thus, the QRT-PCR data confirmed our microarray observations in 96 out of 107 comparisons, or 90% of the time. Among the numerous novel ST targets that we confirmed by QRT-PCR are two thrombin protease-activated receptors (PARs), F2R and F2R2 which are expressed in prostate cancers and are implicated in motility, metastasis, angiogenesis, and src kinase activation (33-35). To confirm changes at the protein level, we first immunoblotted for three genes representative of different expression patterns. In agreement with mRNA data, junction plakoglobin (JUP) decreased in a PP2A-dependent manner, IQGAP2 decreased in a PP2A-independent manner, and thymidine kinase 1 (TK-1) increased in HEK-TERST110 cells and even more so in the HEK-TERST cells (Figure 3E). We also confirmed changes at the protein level for BNIP3 and SERPINB2 (Figure 3F). Nevertheless, one of the limitations of microarray analyses is that changes at the mRNA level are not always reflected at the protein level, and thus it is not possible to be certain that the mRNA changes reported here are all reflected at the protein level. Protein levels of LT and H-Ras-V12 were also measured in the HEK-

12

TERST, HEK-TERST110, HEK-TERASB56g and HEK-TERV by immunoblotting, quantitated using a chemiluminescence imager as described (36), and determined to be within 2-fold for each of the matched lines (data not shown).

Expression Profiling of PP2A B56g knockdown cells identifies 99 potentially critical genes correlated with tumorigenicity Antisense knockdown of PP2A B56g subunits has previously been shown to nearly completely suppress B56g expression at the protein level (15). However, our comparison of microarray expression data in the HEK-TERASB56g and HEK-TERV cells showed only a 1.5 fold decrease of B56g at the mRNA level, suggesting that much of the loss of B56g protein may be due to translational inhibition. Because downregulation of the PP2A B56g subunit can substitute for ST (15), we reasoned that the set of genes similarly affected by ST expression and B56g knockdown would include some of the ST targets that are the most relevant to tumorigenesis. The intersection of the genes affected by B56g knockdown with those genes affected by ST (shown in orange and cyan in Figure 2A), were 128 probe sets corresponding to 99 unique genes (Supplementary Table S6 and Figure 2D). Among these 99 genes were 37 genes (cyan section of Figure 2A) that were similarly affected in the HEK-TERASB56g, HEK-TERST, and HEK-TERST110 cell lines. These 37 genes included matrix metalloproteinase MMP-1, the apoptosisrelated genes TRAIL, MFGE8, and BNIP3, and the microfilament-associated protein palladin. The fact that this set of genes were similarly affected by the ST110 mutant and by knockdown of B56g suggests that these genes can be regulated by both PP2A-dependent and PP2A-independent mechanisms. Of these 99 newly-described targets that are similarly affected by ST expression and B56g knockdown, 62 genes (orange section of Figure 2A) were differentially affected in the HEK-TERST110 cell line (orange bar, Figure 2D), suggesting that they are strongly dependent on ST-PP2A interactions (Table 1). Among

13

these genes were the transcription factors FOXD1 and HOXB3; the apoptosis-related genes, gelsolin, ALDH1A3, and SERPINB2; and cell-cell adhesion molecule protocadherin gamma A11. It has been previously shown that ST activates NF-kB via PKCz and PI3K signaling (37). Our observation that the anti-apoptotic NF-kB targets ALDH1 and SERPINB2 were upregulated in HEKTERST and HEK-TERASB56g cells (Table 1) prompted us to test the steady-state protein levels of the NF-kB inhibitor IkBa. Quantitation of three independent immunoblots (Figure 3F) by chemilumimager demonstrated that IkBa levels are reduced to 55% and 28% of HEK-TERV levels in the HEK-TERST and HEK-TERASB56g cells, respectively, while IkBa in the HEK-TERST110 cells was unchanged. In addition, the NF-kB targets SERPINB2 and survivin are increased at the protein level in HEK-TERST and HEK-TERASB56g cells compared to the HEK-TERV controls (Figure 3F). Activation of NF-kB might be expected to be due to downstream effects of enhanced integrin and src signaling (see below), since NF-kB mediates endothelial cell survival signals from the integrin avb3 - src pathway (38), or via inhibition of PP2A dephosphorylation of the IkB kinase (IKK) (39), or a combination of mechanisms. Thus, expression of ST and knockdown of B56g PP2A subunits both appear to result in PP2A-dependent decreased IkBa protein and subsequent NF-kB activation. Nevertheless, the increase in SERPINB2 and survivin expression in the HEK-TERST110 line suggests that some anti-apoptotic NF-kB targets can also be activated by ST by PP2A-independent mechanisms. ST repressed expression of the pro-inflammatory NF-kB targets IL-8 and IL-1b while enhancing expression of the anti-apoptotic targets of NF-kB, ALDH1, survivin, and SERPINB2. QRT-PCR and western blotting confirmed microarray observations for twenty-two common targets of ST and B56g including SERPINB2, TRABID, DNER, ALDH1, SMAD3, gelsolin (GSN), MMP1, FOXD1, CTGF, PRSS11, and ICAM1 (Figures 3F-I).

14

Normally, NF-kB activation results in activation of both pro-inflammatory and anti-apoptotic target genes. Consistent with this typical response, the knockdown of PP2A B56g subunits in HEKTERASB56g cells increased expression of both pro-inflammatory (IL-8, IL-1b) and anti-apoptotic targets of NF-kB (ALDH1, survivin, and SERPINB2).

However, ST effects on expression of NF-kB

targets were unusual in that pro-inflammatory and anti-apoptotic targets are affected in opposite directions. These differences were due to ST’s general repression of immune system genes such as MHC Class I genes, invariant chain (CD74), and TRAIL among others (Supplementary Table S1).

ST affects expression of cell cycle genes and cell cycle progression in low serum As expected, ST induced a general pattern of increased expression of genes associated with cell cycle progression and decreased expression of genes associated with cell cycle arrest. In agreement with published studies (30, 31) in which co-expression of SV40 LT and ST antigen drives cells into S phase, thymidine kinase (TK-1) and dihydrofolate reductase (DHFR) were upregulated by ST. Notably, some S phase and cell cycle regulated genes, such as TK-1 (Figures 3A and 3C), were also upregulated by ST110. Immunoblotting for cyclin A showed high expression in the HEK-TERV cells, demonstrating that LT and H-Ras-V12 can activate the cyclin A promoter in the absence of ST (Figure 4A). Cyclin A levels were decreased in HEK-TERST and HEK-TERST110 cells compared to HEK-TERV cells at both the protein and mRNA levels (Figure 4A and Supplementary Table S1). This may be due to the fact that protein and mRNA were prepared after 24 hours in low serum, causing an arrest of HEK-TERV cells at the G1/S phase transition, while HEK-TERST and HEK-TERST110 cells progressed into G2/M. Consistent with this hypothesis, the HEK-TERST and HEK-TERST110 cells have elevated cyclin B levels, demonstrating the ability of these lines to progress through the cell cycle under conditions of low serum. The HEK-TERASB56g cells showed intermediate changes, with a partial decrease in cyclin A

15

levels and partial increase in cyclin B levels, suggestive of fewer cells progressing through the cell cycle.

Moreover, T K - 1 and DHFR mRNA were not increased in HEK-TERASB56g cells

(Supplementary Table S1), indicating that unlike ST or ST110 expression, the reduction of B56g is not sufficient to enable cell cycle progression in low serum even in the presence of activated Ras. To ascertain the cell cycle distribution of these cell lines after 24 hours in low serum, DNA content was measured by FACS analysis (Figure 4B). HEK-TERST and HEK-TERST110 cells exhibited highly similar cell cycle distributions, with over 35% in S phase and little sign of cell cycle arrest. However, the 4N peak composed of cells in G2/M dropped dramatically in HEK-TERASB56g cells, indicative of a strong dependence on serum. To further investigate the serum dependence of the four cell lines, all lines were synchronized in the G1 phase by aphidicolin treatment in 10% FBS for 24 hours, and then either harvested, or released into low serum conditions without aphidicolin. The HEKTERST and HEK-TERST110 cells progressed through the cell cycle in a serum-independent manner, whereas most of the HEK-TERV and HEK-TERASB56g cells remained arrested in G1 in a serumdependent manner (Figure 4C). Thus, ST and ST110 can drive HEK cells through the cell cycle in the presence of LT and activated Ras, whereas knockdown of B56g subunits cannot.

ST-induced gene expression patterns suggest increased integrin signaling and reduced cell-cell adhesion A large number of genes involved in cellular adhesion, cytoskeletal structure, and motility were affected by the presence of ST. In particular, several genes known to affect the integrin signaling pathway such as osteopontin, paxillin, f-spondin, gelsolin, and matrix metalloproteinase-1 (MMP-1) were changed in directions consistent with integrin activation by microarray analysis (Supplementary Table S1) and by QRT-PCR (Figures 3A and 3B).

Our data confirmed previous studies (29) that SV40 activates

16

expression of osteopontin (SPP1), and show that ST does this in a PP2A-dependent manner. In addition, ST upregulated expression of integrin signaling targets such as MMP-1, alpha collagens, c-myc, and paxillin by both PP2A-dependent and PP2A-independent mechanisms. These data suggest, but do not prove, that ST may activate integrin signaling directly or indirectly. In contrast to the apparent up-regulation of integrin signaling, expression of many genes important for cell-cell adhesion such as ICAM-1 and VCAM-1 were downregulated. Components of junctional adhesion complexes were also repressed, such as b-catenin, plakoglobin, junctional adhesion molecule 1 (JAM), claudin 11, and protocadherin gamma family members. In addition, secreted-frizzled related protein 1 (SFRP-1) which binds directly to wingless and can inhibit wnt signaling through destabilization of b-catenin (40), was upregulated by ST.

HEK-TERST activation of src shows less serum dependence than HEK-ASB56g The tyrosine kinase c-src and the phosphatidiylinositol-3 kinase (PI3K) are downstream components of integrin signaling pathways (41, 42), and integrin activation of c-src can block proper assembly of cellcell contacts (43). To determine whether activation of integrin signaling through c-src and PI3K is essential for the anchorage independent growth phenotype of the HEK-TERST cell line, soft agar assays were performed in the presence of inhibitors of integrin, c-src, or PI3K signaling. To test the effect of inhibition of integrin signaling on growth in soft agar, HEK-TERST and HEK-ASB56g cells were plated in the presence of 10 mg/ml of a circularized arginine-glycine-aspartic acid (RGD) peptide that acts as an integrin a vb3 antagonist, or an equal concentration of a control arginine-alanine-aspartic acid (RAD) peptide. To determine whether c-src signaling is essential for HEK-TERST and HEK-ASB56g growth in soft agar, a c-src-specific kinase inhibitor PP1 and an inactive structural analog, PP3, were used. HEKTERST and HEK-ASB56g cells were also plated in the presence of wortmannin to test for the

17

requirement of PI3K signaling in anchorage-independent growth. The integrin RGD inhibitor, c-src PP1 inhibitor, and wortmannin all dramatically interfered with the HEK-TERST and the HEK-ASB56g cell line’s ability to form colonies in soft agar (Figure 5A), whereas the control inhibitors had minimal effects. These data show that integrin signaling is essential for transformation by ST expression and by B56g knockdown, and suggest that both of these transformation events activate integrin signaling directly or indirectly. The viability of HEK-TERST cells was determined by MTT dye assay after two weeks of treatment with each of these inhibitors in soft agar, demonstrating that none of the inhibitors had a direct killing effect, but rather prevented colony growth (Figure 5B). To determine whether ST expression or B56g knockdown affects src activation and phosphorylation on tyrosine 418, HEK-TERST and HEK-ASB56g cells were also grown in anchorageindependent 1.3% methylcellulose in both 5% and 10% FBS. After 24 hours in anchorage independent media, cells were spun down and harvested as described (24), lysed and whole cell lysates were probed for the presence of phospho-Y418-src (Figure 5C). ST induced phosphorylation and activation of src under all conditions, whereas src-Y418 phosphorylation in HEK-ASB56g cells was detected only in 10% serum. The dependence of HEK-ASB56g cells on serum for src phosphorylation is consistent with the observation that B56g knockdown does not transform human cells as efficiently as ST transformation in soft agar or tumor formation assays (15). Src-Y418 was not phosphorylated in the HEK-TERV or HEK-TERST110 cells under any conditions, demonstrating that ST-induced activation of src is PP2A-dependent.

While it is not possible to determine whether the changes in src

phosphorylation are direct or indirect effects of ST based on these data, we hypothesize that ST may indirectly activate src through PP2A-dependent activation of integrins (Figure 5D).

18

Discussion Here we have identified the changes in gene expression that are generated by SV40 ST and determined that many of these targets are highly relevant to cancer growth, survival, motility, and metastasis. Our data provide new insights that support a model (Figure 5D) in which ST promotes growth and prevents apoptosis through constitutive integrin signaling and NF-kB activation, while inhibiting components of cell-cell adhesion pathways that might provide cell cycle arrest and pro-differentiation signals. While many of the effects of ST were shown to be independent of PP2A binding and inhibition, the key changes shared by the tumorigenic HEK-TERST and HEK-TERASB56g cell lines were largely PP2Adependent and included upregulation of anti-apoptotic effectors like SERPINB2/PAI-2, as well as developmental homeobox and forkhead box transcription factors. Moreover, ST appeared to be able to differentially regulate pro-inflammatory and anti-apoptotic targets of NF-kB. The 137 genes that were affected by ST in a PP2A-independent manner but were unaffected by B56g knockdown may enhance, but not be essential, for transformation. Many of these genes were cell cycle regulated genes, and may reflect the inability of the HEK-TERASB56g line to progress through S phase under conditions of low serum. In agreement with published literature (10, 30, 31, 44-48), coexpression of SV40 LT and ST antigen drove cells into S phase in low serum, and we observed corresponding increases in expression of S phase genes such as TK-1, DHFR, and G0S2. In contrast to earlier studies in which S phase entry mediated by polyoma or SV40 LT and ST was PP2A-dependent (13, 49), we observed that the ST110 mutant could also support S-phase entry, although to a lesser degree than wild-type ST. One potential reason for the ability of the ST110 mutant to support cell cycle progression in our model system is that our cell lines also express the constitutively activated H-RasV12 mutant. Thus, in the presence of LT and activated Ras signaling, the N-terminal domain of ST appears to be sufficient to drive cells into S phase. Consistent with this hypothesis, mutations in the

19

DnaJ domain of polyoma ST has been shown to strongly inhibit activation of the cyclin A promoter (50). The set of 99 genes that were regulated similarly by ST and B56g knockdown may represent some of the most critical ST-induced changes for transformation of human cells. Many of these expression changes likely result from previously identified effects of SV40 ST in signal transduction pathways such as p27/Kip1 downregulation (30), AKT and telomerase activation (51), MAPK pathway activation (52), PKCz activation of NFkB (37, 53), and induction of cyclins (46, 47). The remainder of the expression alterations may result from still unidentified effects of ST on other signal transduction pathways or from direct effects of ST on transcription factors. It is important to note that HEKTERASB56g cells are serum-dependent, grow more slowly, and are less potently transformed than the HEK-TERST cells.

Thus, while the set of 99 genes are probably the most critical ones for

tumorigenesis, other ST-regulated genes outside of this group may account for the rapidly proliferating, serum-independent phenotype of the HEK-TERST cells. While 99 genes were affected similarly by ST and B56g knockdown, several hundred genes were not. Several different reasons could account for these observations. First, B56g antisense downregulates B56g to a greater extent than ST (15). Second, ST is known to target PP2A isoforms other than B56 (52), although knockdown of B55 subunits cannot fully transform human cells as B56g knockdown does (15). Third, the N-terminal domain of ST may influence the pattern of gene expression caused by PP2A inhibition. Finally, since our microarray experiments were performed in low serum and HEK-ASB56g cells are serum-dependent, more similarities with ST expression may be identified if these cells were compared in normal serum conditions. Several of the genes regulated by ST have roles in prevention or induction of apoptosis.

ST

upregulated ALDH1 which protects cells by metabolizing oxidized lipids; moreover, inhibitors of

20

ALDH1 can drive Bcl-2 overexpressing cells into apoptosis (54). ST also increased expression of SERPINB2, which inhibits TNF-induced apoptosis, and strongly repressed expression of TRAIL. ST also inhibits apoptosis through repression of gelsolin, a regulator and effector of apoptosis. Gelsolin plays a key role in actin remodeling and motility, and associates with integrin av, c-src, focal adhesion kinase (FAK), PI3K, and paxillin in response to integrin activation by osteopontin (55). Using an RGD peptide inhibitor, we showed that integrin signaling was essential for anchorage independent growth of both HEK-TERST and HEK-TERASB56g cell lines. Besides increases in integrin signaling targets, we also observed increased expression of three protease-activated receptors (PAR-1, PAR-2, and PAR-3), which could also contribute towards activation of PI3K and AKT. Consistent with recent work showing that constitutive PI3K signaling can substitute for ST to fully transform human cells (56), we have shown that integrin signaling is critical for ST-helper function in tumorigenesis.

Our data demonstrated that ST expression induces activation of src in low serum, while

knockdown of B56g subunits results in serum-dependent src phosporylation. Thus, the effects of ST on the integrin-src-PI3K pathway are critical for transformation of human cells. Recently, it has been shown that a 6b4 integrin signaling can confer resistance to apoptosis in mammary epithelium via NF-kB activation (57). It is known that ST activates NF-kB via PKCz and PI3K signaling (37), and that PP2A regulates NF-kB activation by dephosphorylation of the IkB kinase (IKK) (39).

Thus, ST may be impacting NF-kB activation in multiple ways, by mimicking and/or

stimulating growth factor and integrin signaling, and by modulation of PP2A activity. ST affected expression of several developmental transcription factors, including HOXA9, HOXB3, HOXB6, Ets-1, FOXD1, FOXG1, and FOXM1. Additional developmental markers induced by ST included markers of pre-B and pro-B lymphocytes, cardiac, epithelial, endothelial, and neuronal tissues. The expression of early differentiation makers from such a wide variety of tissue types (see

21

Supplementary Table S1) suggests that part of the helper function of ST results in a “de-differentiated” phenotype of transformed cells. The observations that we report here suggest that ST may achieve this function in part by repression of components of junctional adhesion complexes such as b-catenin, plakoglobin, JAM1, and protocadherin family members. Our conclusions are consistent with reports that overexpression of plakoglobin can suppress tumorigenicity of SV40 transformed cells (58) and that ST can alter distribution of and reduce levels of tight junction proteins such as occludin and claudin in polarized epithelial cells (59). In conclusion, we have identified the changes in gene expression induced by ST in transformation of human cells, and determined that many of the critical changes are in genes that influence cellular adhesion, apoptosis, proliferation, development, and transcriptional regulation. Many of these factors may regulate pathways essential for tumor formation in human cells and could represent potential therapeutic targets.

22

Acknowledgements The authors would like to thank Suresh Karanam, Wenjian Li, Annise Chung for technical assistance, Dr. Joan Brugge, Ariad Pharmaceuticals, for src 327 monoclonal antibody, Dr. Charles Parkos for the W6/32 MHC Class-I monoclonal antibody, Dr. Andrew Young for assistance with QRT-PCR primer design, and Drs. Paul Wade, Andrew Neish, Jeremy Boss, and Guy Benian for critical reading of this manuscript. This research was supported in part by NIH grant K22 CA96560 to CSM, by NIH Grant CA57327 to DCP, and NIH Grants K01 CA94223 and P01 CA50661, a Doris Duke Clinical Scientist Development Award, a Dunkin Donuts Rising Stars Award, and a Kimmel Scholar Award to WCH.

References

1. 2. 3. 4. 5. 6. 7. 8. 9.

Altstein, AD, Vassiljeva, NN, and Sarycheva, OF Neoplastic transformation of rat embryo cells by simian papovavirus SV40 Nature 1967; 213;931-932. Carbone, M, Rizzo, P, and Pass, HI Simian virus 40, poliovaccines and human tumors: a review of recent developments Oncogene 1997; 15;1877-1888. Kelley, WL and Landry, SJ Chaperone power in a virus? Trends Biochem Sci 1994; 19;277-278. DeCaprio, JA, Ludlow, JW, Figge, J, et al. SV40 large tumor antigen forms a specific complex with the product of the retinoblastoma susceptibility gene Cell 1988; 54;275-283. Zalvide, J, Stubdal, H, and DeCaprio, JA The J domain of simian virus 40 large T antigen is required to functionally inactivate RB family proteins Mol Cell Biol 1998; 18;1408-1415. Montenarh, M, Vesco, C, Kemmerling, G, Muller, D, and Henning, R Regions of SV40 large T antigen necessary for oligomerization and complex formation with the cellular oncoprotein p53 FEBS Lett 1986; 204;51-55. Janssens, V and Goris, J Protein phosphatase 2A: a highly regulated family of serine/threonine phosphatases implicated in cell growth and signalling Biochem J 2001; 353;417-439. Pallas, DC, Shahrik, LK, Martin, BL, et al. Polyoma small and middle T antigens and SV40 small t antigen form stable complexes with protein phosphatase 2A Cell 1990; 60;167-176. Bikel, I, Montano, X, Agha, ME, et al. SV40 small t antigen enhances the transformation activity of limiting concentrations of SV40 large T antigen Cell 1987; 48;321-330.

23

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

Rundell, K, Gaillard, S, and Porras, A Small-t and large-T antigens cooperate to drive cell proliferation Dev Biol Stand 1998; 94;289-295. Yu, J, Boyapati, A, and Rundell, K Critical role for SV40 small-t antigen in human cell transformation Virology 2001; 290;192-198. Hahn, WC, Dessain, SK, Brooks, MW, et al. Enumeration of the simian virus 40 early region elements necessary for human cell transformation Mol Cell Biol 2002; 22;2111-2123. Porras, A, Bennett, J, Howe, A, et al. A novel simian virus 40 early-region domain mediates transactivation of the cyclin A promoter by small-t antigen and is required for transformation in small-t antigen-dependent assays J Virol 1996; 70;6902-6908. Hahn, WC, Counter, CM, Lundberg, AS, et al. Creation of human tumour cells with defined genetic elements Nature 1999; 400;464-468. Chen, W, Possemato, R, Campbell, KT, et al. Identification of specific PP2A complexes involved in human cell transformation Cancer Cell 2004; 5;127-136. Moreno, CS, Park, S, Nelson, K, et al. WD40 repeat proteins striatin and S/G(2) nuclear autoantigen are members of a novel family of calmodulin-binding proteins that associate with protein phosphatase 2A J Biol Chem 2000; 275;5257-5263. Bolstad, BM, Irizarry, RA, Astrand, M, and Speed, TP A comparison of normalization methods for high density oligonucleotide array data based on variance and bias Bioinformatics 2003; 19;185-193. Beissbarth, T and Speed, T GOstat: Find statistically overrepresented Gene Ontologies within a group of genes Bioinformatics 2004. Hochberg, Y and Benjamini, Y More powerful procedures for multiple significance testing Stat Med 1990; 9;811-818. Castel, S, Pagan, R, Garcia, R, et al. Alpha v integrin antagonists induce the disassembly of focal contacts in melanoma cells Eur J Cell Biol 2000; 79;502-512. Hanke, JH, Gardner, JP, Dow, RL, et al. Discovery of a novel, potent, and Src family-selective tyrosine kinase inhibitor. Study of Lck- and FynT-dependent T cell activation J Biol Chem 1996; 271;695-701. Altmann, C, Steenpass, V, Czyborra, P, Hein, P, and Michel, MC Comparison of signalling mechanisms involved in rat mesenteric microvessel contraction by noradrenaline and sphingosylphosphorylcholine Br J Pharmacol 2003; 138;261-271. Kuppumbatti, YS, Rexer, B, Nakajo, S, Nakaya, K, and Mira-y-Lopez, R CRBP suppresses breast cancer cell survival and anchorage-independent growth Oncogene 2001; 20;7413-7419. Kume, K, Jinno, S, Miwatani, H, et al. Oncogenic signal-induced ability to enter S phase in the absence of anchorage is the mechanism for the growth of transformed NRK cells in soft agar New Biol 1992; 4;504-511. Tusher, VG, Tibshirani, R, and Chu, G Significance analysis of microarrays applied to the ionizing radiation response Proc Natl Acad Sci U S A 2001; 98;5116-5121. Yuen, T, Wurmbach, E, Pfeffer, RL, Ebersole, BJ, and Sealfon, SC Accuracy and calibration of commercial oligonucleotide and custom cDNA microarrays Nucleic Acids Res 2002; 30;e48. Schmalbach, CE, Chepeha, DB, Giordano, TJ, et al. Molecular profiling and the identification of genes associated with metastatic oral cavity/pharynx squamous cell carcinoma Arch Otolaryngol Head Neck Surg 2004; 130;295-302. Xin, W, Rhodes, DR, Ingold, C, Chinnaiyan, AM, and Rubin, MA Dysregulation of the annexin family protein family is associated with prostate cancer progression Am J Pathol 2003; 162;255261.

24

29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47.

Salih, V, Knowles, JC, O'Hare, MJ, and Olsen, I Retroviral transduction of alveolar bone cells with a temperature- sensitive SV40 large T antigen Cell Tissue Res 2001; 304;371-376. Porras, A, Gaillard, S, and Rundell, K The simian virus 40 small-t and large-T antigens jointly regulate cell cycle reentry in human fibroblasts J Virol 1999; 73;3102-3107. Kawasaki, S, Diamond, L, and Baserga, R Induction of cellular deoxyribonucleic acid synthesis in butyrate- treated cells by simian virus 40 deoxyribonucleic acid Mol Cell Biol 1981; 1;10381047. Yeh, E, Cunningham, M, Arnold, H, et al. A signalling pathway controlling c-Myc degradation that impacts oncogenic transformation of human cells Nature Cell Biology 2004; Advanced Online Publication. Sabri, A, Muske, G, Zhang, H, et al. Signaling properties and functions of two distinct cardiomyocyte protease-activated receptors Circ Res 2000; 86;1054-1061. Yin, YJ, Salah, Z, Maoz, M, et al. Oncogenic transformation induces tumor angiogenesis: a role for PAR1 activation Faseb J 2003; 17;163-174. Cooper, CR, Chay, CH, Gendernalik, JD, et al. Stromal factors involved in prostate carcinoma metastasis to bone Cancer 2003; 97;739-747. Yu, XX, Du, X, Moreno, CS, et al. Methylation of the Protein Phosphatase 2A Catalytic Subunit Is Essential for Association of Balpha Regulatory Subunit But Not SG2NA, Striatin, or Polyomavirus Middle Tumor Antigen Mol Biol Cell 2001; 12;185-199. Sontag, E, Sontag, JM, and Garcia, A Protein phosphatase 2A is a critical regulator of protein kinase C zeta signaling targeted by SV40 small t to promote cell growth and NF-kappaB activation Embo J 1997; 16;5662-5671. Scatena, M, Almeida, M, Chaisson, ML, et al. NF-kappaB mediates alphavbeta3 integrininduced endothelial cell survival J Cell Biol 1998; 141;1083-1093. DiDonato, JA, Hayakawa, M, Rothwarf, DM, Zandi, E, and Karin, M A cytokine-responsive IkappaB kinase that activates the transcription factor NF-kappaB Nature 1997; 388;548-554. Uren, A, Reichsman, F, Anest, V, et al. Secreted frizzled-related protein-1 binds directly to Wingless and is a biphasic modulator of Wnt signaling J Biol Chem 2000; 275;4374-4382. Tian, B, Lessan, K, Kahm, J, Kleidon, J, and Henke, C beta 1 integrin regulates fibroblast viability during collagen matrix contraction through a phosphatidylinositol 3-kinase/Akt/protein kinase B signaling pathway J Biol Chem 2002; 277;24667-24675. Ojaniemi, M, Martin, SS, Dolfi, F, Olefsky, JM, and Vuori, K The proto-oncogene product p120(cbl) links c-Src and phosphatidylinositol 3'-kinase to the integrin signaling pathway J Biol Chem 1997; 272;3780-3787. Avizienyte, E, Wyke, AW, Jones, RJ, et al. Src-induced de-regulation of E-cadherin in colon cancer cells requires integrin signalling Nat Cell Biol 2002; 4;632-638. Howe, AK, Gaillard, S, Bennett, JS, and Rundell, K Cell cycle progression in monkey cells expressing simian virus 40 small t antigen from adenovirus vectors J Virol 1998; 72;9637-9644. Rundell, K and Parakati, R The role of the SV40 ST antigen in cell growth promotion and transformation Semin Cancer Biol 2001; 11;5-13. Watanabe, G, Howe, A, Lee, RJ, et al. Induction of cyclin D1 by simian virus 40 small tumor antigen Proc Natl Acad Sci U S A 1996; 93;12861-12866. Gaillard, S, Fahrbach, KM, Parkati, R, and Rundell, K Overexpression of simian virus 40 smallT antigen blocks centrosome function and mitotic progression in human fibroblasts J Virol 2001; 75;9799-9807.

25

48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59.

Goetz, GS, Dean, FB, Hurwitz, J, and Matson, SW The unwinding of duplex regions in DNA by the simian virus 40 large tumor antigen-associated DNA helicase activity J Biol Chem 1988; 263;383-392. Schuchner, S and Wintersberger, E Binding of polyomavirus small T antigen to protein phosphatase 2A is required for elimination of p27 and support of S-phase induction in concert with large T antigen J Virol 1999; 73;9266-9273. Schuchner, S, Nemethova, M, Belisova, A, et al. Transactivation of murine cyclin A by polyomavirus large and small T antigens J Virol 2001; 75;6498-6507. Yuan, H, Veldman, T, Rundell, K, and Schlegel, R Simian Virus 40 Small Tumor Antigen Activates AKT and Telomerase and Induces Anchorage-Independent Growth of Human Epithelial Cells J. Virol. 2002; 76;10685-10691. Sontag, E, Fedorov, S, Kamibayashi, C, et al. The interaction of SV40 small tumor antigen with protein phosphatase 2A stimulates the map kinase pathway and induces cell proliferation Cell 1993; 75;887-897. Johannessen, M, Olsen, PA, Sorensen, R, et al. A role of the TATA box and the general coactivator hTAF(II)130/135 in promoter-specific trans-activation by simian virus 40 small t antigen J Gen Virol 2003; 84;1887-1897. Canuto, RA, Muzio, G, Salvo, RA, et al. The effect of a novel irreversible inhibitor of aldehyde dehydrogenases 1 and 3 on tumour cell growth and death Chem Biol Interact 2001; 130-132;209218. Chellaiah, MA, Biswas, RS, Yuen, D, Alvarez, UM, and Hruska, KA Phosphatidylinositol 3,4,5trisphosphate directs association of Src homology 2-containing signaling proteins with gelsolin J Biol Chem 2001; 276;47434-47444. Zhao, JJ, Gjoerup, OV, Subramanian, RR, et al. Human mammary epithelial cell transformation through the activation of phosphatidylinositol 3-kinase Cancer Cell 2003; 3;483-495. Weaver, VM, Lelievre, S, Lakins, JN, et al. beta4 integrin-dependent formation of polarized three-dimensional architecture confers resistance to apoptosis in normal and malignant mammary epithelium Cancer Cell 2002; 2;205-216. Simcha, I, Geiger, B, Yehuda-Levenberg, S, Salomon, D, and Ben-Ze'ev, A Suppression of tumorigenicity by plakoglobin: an augmenting effect of N-cadherin J Cell Biol 1996; 133;199209. Nunbhakdi-Craig, V, Craig, L, Machleidt, T, and Sontag, E Simian Virus 40 Small Tumor Antigen Induces Deregulation of the Actin Cytoskeleton and Tight Junctions in Kidney Epithelial Cells J Virol 2003; 77;2807-2818.

26

Accession NM_002575 NM_002276 NM_032873 NM_017580 H28999 NM_007173 AI653107 NM_139072 AI143879 BG165011 NM_000693 NM_007203 NM_152322 NM_007107 AW664964 NM_032333 NM_006717 NM_002128 NM_006572 NM_015440 AU144882 NM_004472 NM_001034 NM_006559 NM_005996 NM_017829 AW575374 NM_007368 NM_052865 NM_152330

Symbol SERPINB2 KRT19 KIAA1959 TRABID FLJ36748 SPUVE ESTs DNER FLJ25677 ESTs ALDH1A3 AKAP2 FLJ33957 SSR3 ESTs MGC4248 SPIN HMGB1 GNA13 DKFZP586G1517 FLJ13545 FOXD1 RRM2 KHDRBS1 TBX3 CECR5 FLJ22425 GAP1IP4BP C20orf72 C14orf31

AK023585

FLJ13523

Fold Change 3.6 2.5 2.3 2.2 2.2 2.2 2.1 2.1 2.0 2.0 1.9 1.9 1.9 1.9 1.8 1.8 1.8 1.8 1.8 1.8 1.8 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7

Accession BE856336 NM_032883 AV724183 NM_001673 NM_152329 NM_022731 NM_020390 NM_018482 NM_153689 NM_018312 NM_015934 NM_005238 NM_002588 U20489 NM_001560 NM_001946 NM_000393 NM_018912 NM_018914 NM_004995 NM_018916 NM_001814 NM_006033 NM_005610 NM_001219 NM_004342 NM_000935 U40053 NM_006227 NM_001901

Symbol C8orf13 C20orf100 FLJ31362 ASNS PPIL5 NUCKS EIF5A2 DDEF1 FLJ38973 C11orf23 NOP5/NOP58 ETS1 PCDHGC3 ESTs IL13RA1 DUSP6 COL5A2 PCDHGA1 PCDHGA11 MMP14 PCDHGA3 CTSC LIPG RBBP4 CALU CALD1 PLOD2 ESTs PLTP CTGF

1.6

NM_004105

EFEMP1

Fold Change 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.5 -1.5 -1.5 -1.6 -1.6 -1.6 -1.6 -1.6 -1.6 -1.7 -1.7 -1.7 -1.7 -1.7 -1.7 -1.9 -2.0 -2.0 -2.4 -2.9

Table 1: PP2A Dependent Genes with similar profiles in tumorigenic HEK-TERST and HEK-TERASB56g lines that are clustered in Figure 2D. Fold change values are for HEK-TERST relative to HEK-TERV.

27

FIGURE LEGENDS Figure 1: SV40 ST alters expression in genes involved in proliferation, development, transcriptional regulation, and inflammation. A: Summary of number of GO annotations and annotated genes used for GOstat analysis.

B: GO

categories that were significantly overrepresented in the 452 genes affected by ST expression as determined by GOstat analysis included genes involved in proliferation, inflammation, antigen presentation, development, and transcription. C: FACS analysis of HEK-TERV and HEK-TERST cells using the pan-HLA mAb W6/32 shows decreased surface expression of MHC Class I molecules in the HEK-TERST cells as a shift in the peaks to the left relative to the HEK-TERV cells. A representative result from three independent experiments is shown. D: FACS analysis of HEK-TERV and HEKTERST110 cells as in (C). E: FACS analysis of HEK-TERV and HEK-TERASB56g cells as in (C).

Figure 2: Hierarchical clustering of gene subsets affected by SV40 ST expression. A: Venn diagram of gene sets analysed. The entire set of 452 ST-regulated genes is represented by the large red circle. Numbers indicate the number of unique genes in each subset. PP2A-independent genes, shown in magenta and cyan, were similar in ST110 and ST. Strongly PP2A-dependent genes, shown in orange, were similar in ASB56g and ST, but not in ST110. Genes shown in cyan were affected similarly in ST, ST110, and ASB56g. B: Hierarchical clustering of 555 probe sets affected at least 1.5-fold by wt ST expression. Signal log ratios for all twelve possible comparisons of ST/TERV (HEK-TERST vs. HEK-TERV cells), B56/TERV (HEK-TERASB56g vs. HEK-TERV cells), and ST110/TERV (HEKTERST110 vs. HEK-TERV cells) were clustered by gene and by sample. As an example, the dataset label of ST-1/TERV-1 corresponds to a comparison of HEK-ST (repeat 1) vs. HEK-TERV (repeat 1).

28

Increased and decreased expression levels are indicated by red and green intensities, respectively. C: Hierarchical clustering of 205 probe sets corresponding to 171 genes significantly affected by wt ST expression in a PP2A-independent manner. This gene set shows similar expression effects for wildtype ST and the ST110-mutant. D: Hierarchical clustering of 128 probe sets corresponding to 99 genes with similar expression patterns in fully transformed HEK-TERST and HEK-TERASB56g cells.

Figure 3: QRT-PCR and immunoblotting data confirm microarray observations. Expression ratios from QRT-PCR and microarray data are shown for 41 sample genes compared to mRNA levels in HEK-TERV cells. The average fold change and standard error for four independent samples are shown on a log2 scale. A: Expression ratios are shown for four genes upregulated by ST in a PP2A Independent manner. B: Expression ratios are shown for eight genes downregulated by ST in a PP2A Independent manner. C: Expression ratios are shown for ten genes upregulated by ST in a PP2A Dependent manner. D: Expression ratios are shown for six genes downregulated by ST in a PP2A Dependent manner. E: Immunoblot verification of microarray data at the protein level for plakoglobin, IQGAP2, and thymidine kinase. PP2A C subunit is shown as a loading control. IQGAP2 was difficult to detect, indicating that it is expressed at a very low level even in the HEK-TERV cells. Blotting of IQGAP2 immunoprecipitates from these cell lines (data not shown) confirmed the reduction in IQGAP2 protein seen in this figure. F: IkBa levels are decreased in the tumorigenic HEK-TERST and HEKTERASB56g cell lines, but not in HEK-TERST110 cells. Quantitation of three independent experiments (not shown) by chemilumimager demonstrated that IkBa levels were reduced to 55% and 28% of HEKTERV levels in the HEK-TERST and HEK-TERASB56g cells, respectively, while IkBa in the HEKTERST110 cells was unchanged. Immunoblotting also confirms of expression changes at the protein level for BNIP3 and the NF-kB targets SERPINB2/PAI-2 and survivin. G: Expression ratios are shown

29

in HEK-TERST, HEK-TERST110, and HEK-TERASB56g cell lines for ten genes regulated by ST in a PP2A Independent manner. H: Expression ratios are shown in HEK-TERST, HEK-TERST110, and HEK-TERASB56g cell lines for eight genes upregulated by ST in a PP2A Dependent manner. I: Expression ratios are shown in HEK-TERST, HEK-TERST110, and HEK-TERASB56g cell lines for seven genes downregulated by ST in a PP2A Dependent manner.

Figure 4: ST and ST110 expressing cells progress through S phase, whereas B56g knockdown and vector control cell lines are serum-dependent. A: Immunoblots show decreased cyclin A levels and increased cyclin B levels in HEK-TERST and HEK-TERST110 cells.

B: Cell cycle distribution of each cell line is shown as determined by

propidium iodide staining and FACS. Peaks corresponding to cells with 2N DNA content in G1 phase or 4N DNA content in G2/M are indicated on the x-axis in arbitrary units. Sub-2N peaks presumably correspond to subpopulations of aneuploid cells in lines that express LT, but not ST. Shown are cell populations corresponding to asynchronous cells in 10% FBS (black) or to cell populations after 24 hours in low serum conditions (gray). C: Similar FACS analysis to that shown in (B) except that all lines were synchronized in the G1 phase by 24 hours of aphidicolin treatment in 10% FBS, and then immediately harvested (black), or released into low serum conditions without aphidicolin for an additional 24 hours (gray).

Figure 5: Integrin-src-PI3K signaling is essential for anchorage-independent growth. A: Soft agar assays of HEK-TERV, HEK-TERST110, HEK-TERST and HEK-TERASB56g cells either untreated, or with an inhibitor of integrin signaling (RGD peptide), a control peptide (RAD), a src 30

inhibitor (PP1), a control inhibitor (PP3), or a PI3K inhibitor (wortmannin). Results for the HEKTERASB56g line were similar to the HEK-TERST line except that HEK-TERASB56g colonies were generally smaller than HEK-TERST colonies (data not shown). The mean and standard deviation of data from six replicates are shown. B: Viability of HEK-TERST cells after two weeks in soft agar in the presence of treatments used in (A) as measured by MTT assay. The mean and standard deviation of four replicates are shown. C: HEK-TERV, HEK-TERST110, HEK-TERST and HEK-TERASB56 were grown in 1.3% methycellulose for 24 hours in either 5% or 10% FBS and whole cell lysates prepared. Immunoblots were probed for total src and with antibodies specific to src kinase phosphorylated at tyrosine 418. PP2A C subunit is shown as a loading control. ST induced phosphorylation and activation of Src under all conditions, whereas in B56g knockdown cells src phosphorylation was detected only in 10% serum. Src was not phosphorylated in the HEK-TERV or HEK-TERST110 cells under any conditions, demonstrating that ST-induced activation of src is PP2A-dependent. D: Model of the role of ST and B56g in anchorage-independent growth. HEK-TERST cells: Integrin signaling and ST together activate src and NFkB to increase expression of anti-apoptotic genes. ST represses pro-inflammatory genes by an unknown mechanism. HEK-TERASB56 cells: Lack of B56g together with integrin signaling and growth factors from serum enable src activation, NFkB activation, and increases in both pro-inflammatory and anti-apoptotic genes.

31