Role of IFI 16, a Member of the Interferon-Inducible P200-Protein Family, in Prostate Epithelial Cellular Senescence

Role of IFI 16, a member of the interferon-inducible p200-protein family, in prostate epithelial cellular senescence

Hong Xin1, Jonathan Curry1, Ricky W Johnstone2, Brian J Nickoloff1 and Divaker Choubey*,1

1Departments of Pathology and Radiation Oncology, Stritch School of Medicine, Loyola University Medical Center, 2160 South First Avenue, Mail code: 114B, Maywood, IL 60153, USA; 2Cancer Immunology Program, The Peter MacCallum Cancer Institute, Trescowthick Research Laboratories, St Andrews Place, East Melbourne, 3002 Victoria, Australia

Recent studies have implicated interferon signaling in the senescence represents a major barrier, analogous to regulation of cellular senescence. However, the role of apoptosis, which cells must circumvent to become specific interferon-inducible proteins in cellular senescence malignant (Campisi, 2001). Therefore, identification of remains to be defined. Here we report that IFI 16, an pathways, which regulate cellular senescence, and interferon-inducible transcriptional modulator from the elucidation of the molecular mechanisms by which cells p200-protein family, contributes to cellular senescence of can bypass cellular senescence, are critical to the prostate epithelial cells. Normal human prostate epithelial understanding of development of prostate cancer. cells (PrEC) in culture expressed detectable levels of IFI Interferons (IFNs), a family of cytokines, are potent 16, and the levels increased more than fourfold when cells inhibitors of cell growth both in vitro and in vivo approached cellular senescence. Consistent with a role of (Gutterman, 1994; Borden et al., 2000). IFNs also IFI 16 in cellular senescence, human prostate cancer cell exhibit antitumor activity (Borden et al., 2000; Jonasch lines either did not express IFI 16 or expressed a variant and Haluska, 2001). As IFN-inducible proteins mediate form, which was primarily detected in the cytoplasm of the growth-inhibitory activities of IFNs (Sen and prostate cancer cells and not in the nucleus. Moreover, Lengyel, 1992; Stark et al., 1998), it is important to overexpression of functional IFI 16 in human prostate define their role in pathways contributing to irreversible cancer cell lines inhibited colony formation. Additionally, cell growth arrest. ectopic expression of IFI 16 in clonal prostate cancer cell Recent studies involving unbiased approaches, such lines was associated with a senescence-like phenotype, as cDNA-microarray analyses (Shou et al., 2002) or production of senescence-associated b-galactosidase (a serial analysis of gene expression (SAGE) (Untergasser biochemical marker for cellular senescence), and reduc- et al., 2002), have revealed that IFN signaling may play tion of S-phase cells in culture. Importantly, upregulation an important role in cellular senescence of human of p21WAF1 and inhibition of E2F-stimulated transcription prostate epithelial cells. For example, expression of a set accompanied inhibition of cell growth by IFI 16 in of IFN-activatable genes is upregulated during the onset prostate cancer cell lines. Collectively, our observations of cellular senescence in human prostate epithelial cells support the idea that increased levels of IFI 16 in PrECs (HPrECs) (Untergasser et al., 2002). Moreover, the loss contribute to senescence-associated irreversible cell of expression of IFN-activatable genes is correlated with growth arrest. development of prostate cancer (Shou et al., 2002). Oncogene (2003) 22, 4831–4840. doi:10.1038/sj.onc.1206754 Although these studies suggest the role of IFN-inducible proteins in cellular senescence of PrECs, role of specific Keywords: interferon-inducible IFI 16; prostate; senes- IFN-inducible proteins in cellular senescence remains to cence; p21WAF1; E2F; Rb be defined. The 200-gene family encodes one family (the p200 family) of IFN-inducible proteins (Lengyel et al., 1995; Landolfo et al., 1998; Johnstone and Trapani, 1999). The gene family includes in mice Ifi202a, Ifi202b, Ifi203, Introduction Ifi204, and D3 genes; and in humans IFI16, MNDA,and AIM2 genes (Lengyel et al., 1995; Wang et al., 1999). Normal somatic cells have a limited capacity for cell The murine genes are located on the long arm of mouse division in culture and eventually enter a state of chromosome 1 (Lengyel et al., 1995). Interestingly, this irreversible proliferative arrest, termed replicative senes- part of mouse chromosome is syntenic to human cence or cellular senescence (Hayflick, 1965; Smith and chromosome 1 (Kingsmore et al., 1989). Notably, the Pereira-Smith, 1996). Accumulating data suggest that human genes, including the IFI 16 gene, are located in the irreversible cell growth arrest associated with cellular the 1q22 region, predicted to harbor a gene(s) with a role in cellular senescence (Hensler et al., 1994; Karlsson *Correspondence: D Choubey; E-mail: [email protected] et al., 1996). Consistent with this prediction, expression Received 29 November 2002; revised 1 May 2003; accepted 1 May 2003 of the AIM2 gene is lost in human fibroblasts derived Role of IFI 16 in prostate epithelial cellular senescence HXinet al 4832 from Li-Fraumeni patients during immortalization gene is located in the region 1q22, predicted to harbor a (Kulaeva et al., 2003). senescence gene(s) (Hensler et al., 1994; Karlsson et al., The p200-family proteins share at least one repeat of 1996), we investigated the role of IFI 16 in cellular partially conserved 200-amino-acid residues (HIN-200 senescence of prostate epithelial cells. Our observations domain) (Lengyel et al., 1995; Johnstone and Trapani, provide support for the idea that the loss of IFN- 1999). The protein p202a (hereafter referred as p202) is inducible IFI 16 function in normal human prostate the best-characterized family member and contains two epithelial cells contributes to bypassing of cellular 200-amino-acid repeats (Choubey, 2000; Choubey and senescence, thus, contributing to development of pros- Kotzin, 2002). Of note, these repeats in p202 bind tate cancer. several transcription factors (Choubey, 2000), raising the possibility that 200-amino-acid repeats serve as scaffolds to assemble large protein complexes modulat- Results ing transcription of the target genes. Additionally, some of these proteins, including p204 and IFI 16, also share a Expression of IFI 16 is detectable in normal human newly identified protein domain named DAPIN/PYR- prostate IN/CARD in their N-terminus (Choubey and Kotzin, 2002). The domain is thought to be involved in protein– Recent immunohistochemical analyses of IFI 16 protein protein interactions and is found in several death- in normal human tissues revealed that its expression is associated proteins (Fairbrother et al., 2001; Staub et al., not restricted to hematopoietic cells (Gariglio et al., 2001). The presence of a CARD domain and two HIN- 2002; Wei et al., 2003). Additionally, hybridization of 200 domains in IFI 16 protein is consistent with its IFI 16 cDNA probe to a multiple tissue Northern blot potential role as a scaffold protein. revealed that IFI 16 mRNA expression is relatively high Studies have indicated that ectopic expression of in several human tissues, including the prostate (Wei p200-family proteins, such as p202 and p204, in a et al., 2003). Therefore, we tested whether expression of variety of cultured cells results in retardation of cell IFI 16 protein is detectable in normal human tissues. proliferation (Choubey, 2000). Interestingly, inhibition Our experiments using immunoblotting indicated that of cell growth by p202 (Choubey and Kotzin, 2002) and expression of IFI 16 is readily detectable in extracts p204 (Hertel et al., 2000) appears to depend on Rb derived from normal human prostate, ovary, and breast tumor-suppressor protein. Consistent with a role for tissues (see Figure 1). As a control, we also determined p200-family proteins in cell-growth regulation, there are levels of double-stranded RNA-dependent protein indications that viral oncoproteins functionally inacti- kinase (PKR), a well-known IFN-inducible protein vate p202 (Xin et al., 2001). Expression of AIM2 is lost (Sen and Lengyel, 1992), which were also detectable in by frame-shift mutations in colorectal tumors (Mori human liver-, prostate-, ovary-, and breast-derived et al., 2001), and loss of MNDA expression in prostate protein extracts. carcinoma is linked to progression to more aggressive metastatic prostate cancer (Varambally et al., 2002). Expression of IFI 16 is detectable in normal PrECs and These observations support the idea that the loss of its levels increase after IFN treatment or during cellular function of p200-family proteins, by providing the senescence growth advantage to the affected cells, may contribute to the development of cancer. As expression of IFI 16 was detectable in extracts The protein IFI 16 is primarily a nuclear phospho- derived from normal prostate, we sought to determine protein (80–85 kDa), which contains two repeats of the whether expression of IFI 16 was detectable in normal 200-amino acids (one a-type and one b-type) and a human PrECs in culture. For this purpose, we pur- serine–threonine–proline (S/T/P)-rich spacer region se- chased PrECs in culture (from two different donors). parates the two repeats (Johnstone and Trapani, 1999). These PrECs have been shown to exhibit several The size of the spacer region in IFI 16 is regulated by immunophenotypic features consistent with prostate mRNA splicing and can contain one, two, or three copies of highly conserved 56-amino-acid S/T/P domain encoded by distinct exons, resulting in three different- size proteins: A, B, and C (Johnstone et al., 1998). The protein IFI 16 is primarily nuclear in lymphoid cells (Dawson and Trapani, 1995). Importantly, recent analyses of IFI 16 expression in several human tissues by immunohistochemistry revealed that expression of IFI 16 is not restricted to cells of hematopoietic origin (Gariglio et al., 2002; Wei et al., 2003). Of note, IFN induction of IFI 16 in hematopoietic cells appears to be cell-type- and donor-specific (Schlaak et al., 2002). A study has revealed that IFN induction of IFI 16 in Figure 1 Detection of IFI 16 in various human tissues. Total human melanoma cell lines correlates well with inhibi- lysates derived from the indicated human tissues were analysed by tion of cell growth (Certa et al., 2001). As the IFI 16 immunoblotting using specific antibodies to IFI 16, PKR, and actin

Oncogene Role of IFI 16 in prostate epithelial cellular senescence HXinet al 4833 cells (65–75%) appearing large, ﬂattened, and round (passage 8). As shown in Figure 3, basal levels of IFI 16 increased with increases in the number of cell passages. Furthermore, with increases in the passage number of prostate epithelial cells, the levels of cyclin D1 decreased (not shown) and the levels of p21WAF1 and p16INK4a increased (see Figure 3). The above observations suggest that IFI 16 may play a role in cellular senescence of PrECs. Figure 2 PrECs constitutively express IFI 16, which is increased by IFN treatment. Levels of IFI 16 were analysed by immunoblotting total cell extracts prepared from normal prostate epithelial Prostate cancer cell lines either do not express detectable cells in culture (passage 4) without any treatment (lane 1), after 24 h levels of IFI 16 or express a variant form of IFN-a treatment (lane 2) or IFN-g treatment (lane 3). The three isoforms (A, B, and C) of IFI 16 are indicated. Extracts were also To investigate a role for IFI 16 in prostate epithelial analysed for PKR and actin levels. IFN induction in IFI 16 levels cellular senescence, we asked whether expression of IFI (in fold) is indicated 16 or its function is lost during the development of prostate cancer. For this purpose, we compared the basal as well as IFN-induced levels of IFI 16 in cells of basal epithelial cells (Garraway et al., 2003). As shown three well-characterized human prostate tumor cell lines: in Figure 2, IFI 16B form (the middle protein band) was the androgen-responsive LNCaP cell line and androgen predominantly detected in proliferating PrECs at an nonresponsive DU-145 and PC-3 cell lines (Webber early passage (passage 4). Notably, treatment of cells et al., 1997). As shown in Figure 4, basal levels of IFI 16 with IFN-a or g resulted in further induction of IFI 16 mRNA and protein were not detectable in two prostate (compare lane 3 with 1). As a positive control for IFN cancer cell lines: LNCaP and DU-145. Similarly, treatment, we also determined levels of PKR. As shown expression of IFI 16 was not detectable in an in Figure 2, IFN treatment of cells also produced immortalized (immortalized by HPV18) prostate epithe- increases in the levels of PKR. lial cell line RWPE-1 (data not shown). However, To further investigate the role of IFI 16 in prostate expression of IFI 16 mRNA and protein was detectable epithelial cell growth, we compared basal levels of IFI 16 in prostate cancer cell line PC-3. in extracts derived from normal PrECs at different Interestingly, IFN treatment of cells with either IFN-a passages: cells proliferating actively (passage 4; cells (1000 U/ml) or IFN-g (10 ng/ml) for 24 h triggered the were small and polygonal), cells (30–40%) appearing induction of IFI 16 in DU-145 and PC-3 cells. However, ﬂattened and round (passages 6 and 7), and most of the IFN treatment of LNCaP cells did not result in the

Figure 3 Levels of IFI 16 increase when normal PrECs in culture approach replicative senescence. Left panel: morphology of normal PrECs proliferating actively at passage 4 (P4; also see left top panel for cell-cycle analyses). After reaching passage 8 (P8), most of the cells in culture morphologically appeared senescent (see arrows). Right panel: Levels of IFI 16, p16INK4a, p21WAF1, and actin were analysed by immunoblotting in extracts derived from normal PrECs at the indicated passage (from passage 4 to 8). Increase in IFI 16 levels (in fold) is indicated

Oncogene Role of IFI 16 in prostate epithelial cellular senescence HXinet al 4834

Figure 4 Expression of IFI 16 in human prostate cancer cell lines. Left panel: levels of IFI 16 mRNA were analysed by Northern blotting total RNA isolated from indicated human prostate cancer Figure 6 Human PrECs (passage 6; left panel) or PC-3 cells (right cell lines. The same blot was rehybridized to a probe specific to panel), cultured on glass coverslips, were fixed with methanol, and actin. The bottom panel shows the amount of RNA in an agarose stained for IFI 16 expression and localization by indirect gel used for Northern blotting. Right panel: levels of IFI 16, PKR, immunofluorescence microscopy as described in Methods and actin were analysed by immunoblotting in extracts derived from indicated prostate cancer lines without any treatment (lanes 1, 4, and 7), after 24 h of IFN-a treatment (lanes 2, 5, and 8) or after in the IFI 16 protein sequence may affect its structure IFN-g treatment (lanes 3, 6, and 9) and/or function. Consistent with this prediction, we found that IFI 16 protein was primarily detected in the induction of IFI 16. Furthermore, expression of PKR cytoplasm of PC-3 cells and not in the nucleus (Figure 6, was detectable in all the three human prostate cancer cell right panel). As expected, in PrECs IFI 16 protein was lines and IFN treatment resulted in very modest detected exclusively in the nucleus (Figure 6, left panel). increases in DU-145 and PC-3 cells, but not in LNCaP This observation supports the idea that mislocalization cells. Together, these observations indicated that ex- of IFI 16 in the cytoplasm of human prostate cancer pression of IFI 16 is silenced in cells of LNCaP and DU- cells could inactivate IFI 16 function. 145 cell lines. Since the PC-3 cells express detectable levels of IFI 16 Ectopic expression of IFI 16 in prostate cancer cell lines mRNA and protein, we isolated cDNA coding for the inhibits colony formation IFI 16B protein (because the B form of IFI 16 is predominant in these cells) and determined its sequence Inhibition of cell growth by murine members of p200- (GenBank accession number for the sequence is family proteins depends on the retinoblastoma tumor- AY138863). Next, we compared the sequence with IFI suppressor protein pRb (Choubey and Kotzin, 2002). 16 gene sequences in the GenBank to identify single Moreover, IFI 16 stimulates the transcriptional activity nucleotide polymorphisms (SNPs) in the IFI 16 gene of p53 (Johnstone et al., 2000). Therefore, we chose to expressed in cells of PC-3 cell line. This approach compare the inhibition of colony formation by IFI 16 in revealed that IFI 16B coding sequence from the PC-3 three well-characterized human prostate cancer cell lines cell line contains six previously identified SNPs in the (Webber et al., 1997), differing in their status of Rb and GenBank, resulting in the substitution of six amino p53 functions: DU-145 (harboring nonfunctional Rb acids in the IFI 16B protein (see Figure 5). and p53), PC-3 (harboring functional Rb but mutant Although the functional significance of SNPs in the p53), and LNCaP (harboring functional Rb and p53). IFI 16 gene, which is expressed in the PC-3 prostate For this purpose, we first compared the transfection cancer cell line, remains unknown, the nonconservative efficiencies between the vector (pCMV) and pCMV-IFI amino-acid substitutions, such as His to Asp (in the 16B. Using FuGene-6 as a transfection reagent, we CARD domain), Arg to Ser, Ser to Tyr, and Asn to Thr, transfected pCMV or pCMV IFI16 plasmid along with pCMV-b-gal plasmid (encoding b-galactosidase; in 10 : 1 ratio) into DU-145 cells and stained cells for b- galactosidase activity. Counting of the blue cells indicated that the transfection efficiencies between the two plasmids were comparable (B34 and B38%, respectively) in two independent experiments. Next, the above three prostate cancer cell lines were transfected either with vector pCMV or pCMV-IFI16B plasmid in the same experiment. Transfectants were selected in G418 for about 2 weeks and colonies stained, Figure 5 Schematic presentation of genetic polymorphisms in the and counted. As shown in Figure 7, transfection of IFI 16 gene, which is expressed in prostate cancer cell line PC-3, pCMV-IFI 16B plasmid in the three cell lines resulted in resulting in the substitution of indicated amino-acid residues in IFI inhibition of colony formation, as compared to vector- 16B protein (bottom panel) and its comparison with the wildtype transfected cells (the number of colonies was normalized IFI 16B (top panel). Various domains (CARD, HIN-200) in IFI 16B are also indicated. The altered amino-acid residues in IFI 16, for differences in the transfection efficiencies). However, with potential to affect the function of the variant form of IFI 16, the extent of colony inhibition varied: the maximum are indicated in bold inhibition (490%) was seen in LNCaP cells; minimum

Oncogene Role of IFI 16 in prostate epithelial cellular senescence HXinet al 4835

Figure 7 Ectopic expression of IFI 16 in human prostate cancer cell lines inhibits colony formation. Cells of human prostate cancer cell lines (LNCaP, PC-3, and DU-145) were either transfected with pCDNA3 plasmid or equal amounts of pCDNA3-IFI16B plasmid. Transfected cells were selected in G418 for about 2 weeks and the resistant colonies (size42 mm) in plates were stained (see left panel). Right panel shows percent colony formation by IFI 16B (as compared to vector-transfected cells) in the indicated cell lines. An average of two independent experiments is shown inhibition (B10%) was seen in DU-145 cells. The percent inhibition in cells of PC-3 cell line was B50%. We found similar results in two sets of experiments.

Restoration of IFI 16 function in the PC-3 cell line results in a senescence-like phenotype Overexpression of IFI 16 in cells of PC-3 cell line resulted in a significant (B50%) inhibition of colony formation. Moreover, the levels of IFI 16 increased in cultured normal PrECs when they approached replicative senescence. Therefore, to determine whether the restoration of IFI 16 function in PC-3 cells results in a senescence-like phenotype, we isolated several (430) independent cell clones of PC-3 cells (overexpressing IFI 16B) and maintained them in culture. Interestingly, culture of these cell clones after about 18–34 population doublings exhibited morphological changes in cells: cells appeared enlarged and flattened with increased granu- larity (see Figure 8a). (It is known that, after expression of a senescence-inducing gene in human immortal cell lines, not all cell clones cease cell proliferation in a synchronous manner; see Campisi, 2001.) We chose three independent cell clones (indicated as S-1, S-2, and S-3, respectively) for further characterization with respect to the expression levels of IFI 16 (and other Figure 8 Ectopic expression of functional IFI 16 in cells of PC-3 cell-cycle-regulatory proteins), expression of senescence- prostate cancer cell line results in a senescence-like phenotype. (a) Expression of functional IFI 16 in cells of PC-3 cell line resulted in associated-b-gal (pH 6.0), and irreversible cell growth morphological changes. As compared to cells transfected with arrest in the G1 phase of cell cycle. vector, most cells transfected with IFI 16 expression plasmid Overexpression of IFI 16 in three cell clones (S-1, S-2, morphologically appeared senescent (see arrows). (b) Expression of and S-3) resulted in a senescence-like phenotype over the IFI 16, p21WAF1, cyclin E, and actin by immunoblotting in extracts prepared from cells transfected with vector (V) and three timescale of the experiment. Importantly, cells of all the independent clones (indicated as S1, S2, and S3), isolated after three clones expressed significantly higher levels of IFI transfection of IFI 16-encoding plasmid (a short exposure is shown 16 as compared to vector-transfected cells (pooled cell for IFI 16; therefore, basal levels are not seen in lane 1). (c) clones) and IFI 16 was primarily detected in the nuclear Detection of senescence-associated b-galactosidase activity in cells fraction of cells (data not shown). Of note, we detected of PC-3 cell line after expression of functional IFI 16B, but not in vector-transfected control cells (see arrows). (d) Percent inhibition two IFI 16 protein bands in extracts derived from cell of BrdU incorporation (as compared to vector-transfected cells) in clones (see Figure 8b), raising the possibility that the three PC-3 cell clones after expression of functional IFI 16B upper protein band arises from the phosphorylated form of IFI 16B. Moreover, the increased levels of IFI 16 correlated with increases in levels of CDKI p21WAF1 and of cell growth arrest in the G1 phase of cell cycle decreases in the levels of cyclin E. The latter is indicative (Harbour and Dean, 2000; DeGregori, 2002).

Oncogene Role of IFI 16 in prostate epithelial cellular senescence HXinet al 4836 Staining of three cell clones for expression of SA-b-gal cell line also resulted in upregulation of p21WAF1 (data (pH 6.0) revealed that the majority of cells (493%) in not shown). Since both PC-3 and DU-145 prostate culture, which appeared morphologically senescent, also cancer cell lines do not harbor a functional p53 (Webber stained positive for SA-b-gal (see Figure 8c). However, et al., 1997), these observations are consistent with p53- vector-transfected cells only occasionally (o1%) stained independent regulation of p21WAF1 by IFI 16. Together, positive for SA-b-gal. these observations support the idea that the interferon- Staining of cell clones for BrdU incorporation (to inducible IFI 16 inhibits cell growth of PrECs, in part, determine cell growth arrest in the G1 phase of cell by upregulating the expression of p21WAF1. cycle), in two independent experiments, revealed that, as compared to vector-transfected cells, all the three IFI 16 inhibits E2F1-stimulated transcription and binds to cultures of cell clones had significantly fewer cells in E2F1 and Rb the S phase (see Figure 8d). This observation was also consistent with our flow-cytometry data in which we The tumor suppressor Rb is known to inhibit cell compared the percentage of cells in the G1 phase of cell growth, in part, by binding to E2Fs (E2F1, 2, and 3) and cycle between cultures of vector-transfected cells and inhibiting E2F-stimulated transcription (Harbour and cultures of three cell clones (data not shown). Thus, Dean, 2000; DeGregori, 2002). Since IFI 16 contains an these approaches indicated that ectopic expression of LxCxE-like motif to bind Rb, and predicted to inhibit functional IFI 16 in PC-3 cells resulted in irreversible the E2F-mediated transcription (Gribaudo et al., 1999), cell growth arrest primarily in the G1 phase of cell cycle in order to elucidate the molecular mechanism(s) by and a senescence-like phenotype. which IFI 16 inhibits growth of prostate epithelial cells, Together, the above observations indicated that the we tested whether IFI 16 inhibits E2F1-stimulated expression of IFI 16 in normal prostate epithelial cells transcription. As shown in Figure 10a, expression of contributes to a senescence-associated irreversible cell IFI 16 in promoter–reporter assays significantly (70– growth arrest. 80%) inhibited E2F1-stimulated increases in the activity of the E2F-luc-reporter in PC-3 cell line (these cells IFI 16 upregulates expression of p21WAF1 express functional Rb). Since binding of p202 to both E2F1 and Rb correlates Since ectopic expression of IFI 16 in PC-3 cells was with inhibition of E2F-stimulated transcriptioin accompanied by increases in the steady-state levels of (Choubey et al., 1996), we tested whether IFI 16 could p21WAF1, we tested whether IFI 16 upregulates expres- bind to E2F1 and/or Rb. For this purpose, we used the sion of p21WAF1. As shown in Figure 9a, expression of full-length (amino-acid residues 1–729) or truncated (1– IFI 16 in promoter–reporter assays (in several experi- 159 and 127–729) IFI 16 proteins fused with the ments) resulted in a dose-dependent increase in the glutathione-S-transferase (GST) protein. The fusion activity of p21-luc-reporter in DU-145 cells. Consistent proteins bound to GST beads were used to study with this observation, increased constitutive expression interactions of IFI 16 with Rb and E2F1 proteins in of IFI 16 in stable cell clones of DU-145 cell line resulted vitro (in GST-pull-down assays). As shown in Figure in upregulation of p21WAF1 (see Figure 9b). Again, we 10b and c, the full-length IFI 16 (1–729) fused with detected two IFI 16 B protein bands, which probably GST, but not GST alone, selectively bound to Rb and represent hypo and hyperphosphorylated forms of IFI E2F1 present in the extracts prepared from PC-3 cell 16B. Similarly, increased expression of IFI 16B in HeLa line. Interestingly, the N-terminal segment (amino acids 1–159) of IFI 16, comprising the CARD/DAPIN domain, was sufficient to associate with E2F1 selectively (compare lane 3 with 2), but not Rb (compare lane 3 with 2) (see Figure 10c). Consistent with earlier observation that deletion of the N-terminal amino acids (1–127) in IFI 16 results in increased binding of p53 to IFI 16 (Johnstone et al., 2000), we also noted that deletion of the N-terminal amino acids (1–127) resulted in increased binding of E2F1 and Rb to IFI 16 (compare signal in lane 4 from lane 5 in Figure 10b). Thus, these observations raise the possibility that the N- terminal region (amino acids 1–127) of IFI 16 comprising a CARD/DAPIN protein–protein interaction Figure 9 IFI 16 upregulates expression of p21WAF1 in DU-145 domain may regulate the binding of IFI 16 to p53, prostate cancer cell line. (a) In promoter–reporter assays, expres- Rb, and E2F. sion of IFI 16 in DU-145 cells stimulates the activity of p21-luc- Together, these observations provide support for the reporter in a dose-dependent manner. Low amounts of p53 plasmid hypothesis that the binding of IFI 16 to both E2F1 and (100 ng) were used in column 2 to stimulate the p21-luc-reporter Rb potentiates the Rb-E2F-mediated transcriptional activity. Bars indicated the standard deviation. (b) Overexpression of IFI 16 in a stable cell line of DU-145 cell line results in repression of the growth-promoting genes. This could upregulation of p21WAF1. Two IFI 16 bands probably represent account for the IFI 16-mediated irreversible cell growth hypo- and hyperphosphorylated forms arrest in the G1 phase of cell cycle.

Oncogene Role of IFI 16 in prostate epithelial cellular senescence HXinet al 4837 levels of IFI 16 increase significantly when cells approach cellular senescence. Consistent with its potential role in cellular senescence, prostate cancer cell lines either do not express detectable levels of IFI 16 (LNCaP and DU-145 cell lines), or express a variant form (PC-3 cell line), which was primarily detected in the cytoplasm and not in the nucleus (Figure 6). These observations are consistent with the possibility that the loss of IFI 16 function contributes to bypassing of cellular senescence. Importantly, ectopic expression of IFI 16 in three prostate cancer cell lines resulted in inhibition of colony formation (see Figure 7). Albeit the inhibition by IFI 16 varied significantly between the two cell lines that do not express IFI 16, the maximum inhibition (490%) was seen in LNCaP cells (these cells express functional Rb and p53), and minimum inhibition (B10%) was seen in DU-145 cells (these cells do not express functional Rb and p53). Interestingly, an intermediate inhibition (B50%) was seen in PC-3 cells (these cells express functional Rb but mutated p53), which express a variant form of IFI 16 (see Figure 6). Since these three human prostate cancer cell lines differ with respect to expression of functionals of Rb and p53 (Webber et al., 1997), together, these observations supported an exciting possibility that IFI 16-mediated inhibition of cell growth, in part, depends on the presence of functional Rb and p53. Consistent with this possibility, we found that ectopic expression of IFI 16 in two prostate cancer cell lines (DU-145 and PC-3) resulted in upregulation of cyclin-dependent kinase p21WAF1 (see Figures 8 and 9). Figure 10 IFI 16 inhibits E2F1-stimulated transcription and binds Increases in p21WAF1 levels inhibit phosphorylation of to E2F1 and Rb. (a) Expression of functional IFI 16 in PC-3 cells Rb by Cdks, resulting in inhibition of E2F-mediated inhibits E2F1-stimulated transcription of a reporter gene. Bars transcription of its target genes (Weinberg, 1995; indicate the standard deviation. (b) IFI 16 binds to E2F1 and Rb in GST-pull-down assays. Extracts prepared from PC-3 cells were Harbour and Dean, 2000). Moreover, increased levels WAF1 incubated with indicated GST-IFI 16 fusion proteins. As a negative of p21 can also inhibit E2F-stimulated transcription control, extracts were also incubated with GST alone (lane 2). The independent of Rb (Dimri et al., 1996). Therefore, it is bound proteins were analysed by immunoblotting using antibodies conceivable that the IFI 16-mediated inhibition of cell to E2F1 or Rb. As a positive control for protein size, proteins were growth, in part, depends on p21WAF1. immunoprecipitated from cell extracts with antibodies to E2F1 or Rb and subjected to immunoblotting (lane 1). (c) Schematic Ectopic expression of IFI 16 in promoter–reporter presentation of various GST-IFI 16 fusion proteins, their protein assays potentiates the p53-mediated stimulation of the domains, and their relative ability to bind E2F1 and Rb in the activity of a reporter (pww-luc) whose transcription is experiment shown in panel b. þþ, strong binding; þ , binding; À, driven by the 50-regulatory region of the p21-gene no binding (Johnstone et al., 2000). Moreover, the binding of IFI 16 to p53 in the C-terminus results in increased binding Discussion of p53 to DNA in gel mobility shift assays (Johnstone et al., 2000). Although these observations suggest that Ectopic expression of IFI 16 in promoter–reporter IFI 16 transcriptionally upregulates expression of assays stimulates the transcriptional activity of p53, by p21WAF1 by activating the transcriptional activity of binding to p53 in the C-terminus potentiating the p53, it remains unclear whether IFI 16 also upregulates specific DNA-binding activity (Johnstone et al., 2000). expression of p21WAF1 by activating the transcriptional The amino-acid sequence of IFI 16 contains an LxCxE- activity of other p53 family members, such as p73. like Rb-binding motif in the b-repeat (Lengyel et al., Notably, we found that ectopic expression of IFI 16 in 1995; Johnstone and Trapani, 1999). When ectopically promoter–reporter assays as well as in stable prostate expressed in human cell lines that do not express the cancer cell lines, which do not harbor functional p53 protein, IFI 16 retards cell growth by delaying progres- (Webber et al., 1997), also resulted in upregulation of WAF1 sion from G1 to S phase of cell cycle (Johnstone and p21 . Thus, IFI 16 joins p202, the other member of Trapani, 1999). However, the molecular mechanisms by the p200-family, which also upregulates p21WAF1 by which IFI 16 inhibits cell proliferation remain to be mechanisms independent of p53 (Gutterman and elucidated. Choubey, 1999). Here we report that expression of IFI 16 is detectable Although it remains to be determined how IFI 16 in normal human prostate epithelial cells in culture and upregulates expression of p21WAF1, our observations

Oncogene Role of IFI 16 in prostate epithelial cellular senescence HXinet al 4838 have revealed that ectopic expression of IFI 16 in HeLa gene. Sequencing of this genomic DNA fragment and its cells results in downregulation of c-Myc (data not comparison with sequences in the GenBank revealed the shown). Since c-Myc represses the transcription of p21 presence of SNPs, raising the possibility that these SNPs gene (Gartel et al., 2001), it is conceivable that IFI 16 might contribute to differential regulation of IFI 16 gene transcriptionally activates the p21 gene (independent of in individuals. Consistent with this possibility, the IFN- p53), in part, by relieving c-Myc-mediated transcrip- induction of IFI 16 gene has been shown to be cell-type- tional repression of p21 gene. Additionally, Rb is shown and donor-specific (Schlaak et al., 2002). to activate the transcription of p21 gene (Decesse et al., In summary, our novel observations support the idea 2001). Therefore, binding of IFI 16 to Rb in prostate that IFN-inducible IFI 16 from the p200-protein family epithelial cells with functional Rb (for example, in PC-3 contributes to cellular senescence of prostate epithelial and LNCaP) is expected to potentiate the Rb-mediated cells. Since irreversible cell growth arrest associated with transcriptional activation of p21 gene. cellular senescence represents a major barrier, analogous It remains to be seen whether a variant form of IFI to apoptosis, which cells must circumvent to become 16, in which six amino acids of IFI 16 are substituted malignant, our observations support the idea that loss of and is expressed in PC-3 prostate cancer cell line, is IFI 16 function in prostate epithelial cells may provide a functionally defective in its ability to inhibit prolifera- proliferation advantage to cells by bypassing a senes- tion of prostate epithelial cells. Notably, phosphoryla- cence checkpoint. We suggest that IFI 16 inhibits tion status of IFI 16 is linked to its appropriate nuclear proliferation of prostate epithelial cells, in part, by localization (Briggs et al., 2001). Therefore, it is upregulating expression of p21WAF1, and by inhibiting conceivable that the variant form of IFI 16 with E2F-stimulated transcription. Importantly, our obser- additional amino acids, such as Ser409 and Thr667 (see vations will serve as a basis to determine whether loss of Figure 5), which could be phosphorylated by protein IFI 16 function could serve as a molecular marker for kinases, behaves differently with respect to its nuclear development of prostate carcinoma. localization and/or protein–protein interactions. Con- sistent with this prediction, we report that IFI 16 is mislocalized in the cytoplasm of PC-3 prostate cancer Materials and methods cells (Figure 6). Further work will be needed to address these interesting possibilities. PrECs and prostate cancer cell lines Studies have provided evidence that bypassing of Normal human PrECs were purchased (in culture) from replicative senescence in PrECs requires alterations in Clonetics (San Diego) at passage 3. Cells were split as the p16/Rb pathway and all prostate cancer cell lines, suggested by the supplier and maintained in PrEBM culture which bypass senescence, show defects in either p16 or medium (cat # CC-3165; Clonetics), supplemented with Rb function (Jarrard et al., 1999). Our observations suggested reagents/growth factors and antibiotics. After indicate that IFI 16 levels increase in normal PrECs passage 8, PrECs in culture exhibited a phenotype consistent when they approach cellular senescence (see Figure 3). with replicative senescence: cells became large, flattened with Additionally, restoration of functional IFI 16 in cells of granular cytoplasm (Campisi, 2001). Human prostate cancer PC-3 cell line resulted in a senescence-like phenotype cell lines (LNCaP, DU-145, and PC-3) were purchased from (Figure 8). Together, these observations support the ATCC and maintained in 1640 culture medium supplemented idea that loss of IFI 16 function in normal prostate with 10% fetal bovine serum (FBS) and antibiotics. epithelial cells contributes to bypassing of replicative senescence. Remarkably, the accumulation of senescent Plasmids and antibodies prostate epithelial cells, which stain for senescence- Mammalian expression vector (pCMV-IFI 16B) to express IFI associated marker b-galactosidase, has been reported to 16B, bacterial expression vectors to express GST-IFI 16 fusion play a role in the development of the prostatic proteins (GST-IFI 16(1–729), GST-IFI 16(127–729), and enlargement associated with benign prostatic hyperpla- GST-IFI 16(1–159)) and monoclonal antibodies to IFI 16 sia (BPH) (Choi et al., 2000; Castro et al., 2003). have been described (Johnstone et al., 1998). Consistent with these reports, our preliminary work indicated that expression of IFI 16 is detectable by RT–PCR and sequencing immunohistochemistry in prostate tissues derived from Total RNA isolated from PC-3 cells was subjected to one-step BPH but not prostate adenocarcinoma. Further work is RT–PCR, using IFI 16B-specific primers and a kit from in progress to determine whether expression of IFI 16 Gibco-BRL, as suggested by the supplier. The PCR product function in prostate epithelial cells could serve as a was cloned in a TOPO cloning vector (Invitrogen, Carlsbad, molecular marker to follow the severity of BPH in CA, USA), to perform DNA sequencing. patients and more importantly, to determine whether the loss of IFI 16 function could serve as a reliable Colony-formation assays marker for development of prostate cancer in an early Colony-formation assays were performed essentially as de- stage. scribed previously (Xin et al., 2001). In brief, human prostate To elucidate the molecular mechanisms by which cell lines (B105 cells in 60-mm plate) were transfected with expression of the IFI 16-gene is regulated in normal either empty vector (pCDNA3; 2 mg) or equal amounts of PrECs, we have isolated a genomic fragment (1.68 Kb) pCDNA3-IFI 16B plasmid (encoding IFI 16B protein) using corresponding to the 50-regulatory region of the IFI 16 FuGENE 6 transfection agent. Following 24 h of transfec-

Oncogene Role of IFI 16 in prostate epithelial cellular senescence HXinet al 4839 tions, cells were split into two 100 mm plates and selected in Immunoblotting G418 for about 2 weeks. Colonies exhibiting G418 resistance Tissue lysates (Protein Medleys) derived from various human were stained with crystal violet and counted as described tissues were purchased from Clontech (Palo Alto, CA, USA). previously (Xin et al., 2001). To prepare cell lysates, cells were collected from plates in PBS and resuspended in a modified RIPA lysis buffer (50 mm Tris- Reporter assays HCl, pH 8.0; 250 mm NaCl, 1% NP-40, 0.5% sodium Luciferase reporter assays were performed as described deoxycholate, 0.1% SDS) supplemented with protease inhibi- previously (D’Souza et al., 2001). In brief, subconfluent tors (leupeptin, 50 mg/ml; Pepstatin A, 50 mg/ml; PMSF, 1 mm), cultures of PrECs were transfected using FuGENE 6 transfec- and incubated at 41C for 30 min. The cell lysates were tion reagent (Rosch, Indianapolis, IN, USA) with desired p- sonicated briefly before centrifugation at 14 000 r.p.m. in a luc-reporter plasmid (1 mg) and a second reporter pRL-TK microfuge for 10 min. The supernatants were collected and (0.1 mg; encoding the Renella luciferase) with appropriate equal amounts of proteins were processed for immunoblotting amounts of plasmid encoding the gene of interest. The cells as described previously (Xin et al., 2001). The IFI 16 were harvested 45 h after transfections and the firefly luciferase monoclonal antibodies have been described (Johnstone et al., and Renilla luciferase activities were determined (in triplicates) 1998). Antibodies to PKR, p16, p21WAF1, pRb, E2F1, and using Dual-Luciferase Reporter Assay kit (from Promega). cyclin E were purchased from Santa Cruz Biotech. Inc. (Santa The firefly-luciferase activity was normalized to Renilla- Cruz, CA, USA). luciferase in order to control for variations in transfections. The luciferase activity in control vector-transfected cells will be Northern blotting considered as 1. Standard error was calculated using three experimental readings. Total cytoplasmic RNA was isolated from cells and subjected to Northern blotting followed by hybridization with a HindIII Flow cytometry 50-fragment of IFI 16B cDNA (nucleotides 1–1192) (Johnstone et al., 1998). Hybridizations were performed as described Flow cytometry was performed on single-cell suspensions on previously (D’Souza et al., 2001). adherent (after trypsin and EDTA treatment) as well as the floating cells after pooling them. Briefly, for cell-cycle analysis, cells were stained with propidium iodide (50 mg/ml, Sigma) and Senescence-associated b-galactosidase assays subjected to flow cytometry using Coulter Epics XL-MCL flow These assays were performed essentially as described pre- cytometer as described previously (D’Souza et al., 2001). viously (Dimri et al., 1995). In brief, cells in culture were fixed Apoptosis was measured by the accumulation of cells with a using in situ b-galactosidase staining kit (cat # 200384; from sub-G DNA content. 1 Stratagene, La Jolla, CA, USA). Fixed cells were incubated with the staining solution prepared in a buffer (pH 6.0) Indirect immunofluorescence overnight at 371C. Cells, which stained blue, were considered Human PrECs or PC-3 cells were cultured on glass coverslips positive for the activity of senescence-associated b-galactosi- in a six-well plate. Cells were fixed with 100% methanol at dase and were counted. room temperature and processed for indirect immunofluorescence microscopy to localization of IFI 16, as described Cell-proliferation assays previously (Choubey and Lengyel, 1992). Antibodies to IFI 16 were detected by the secondary anti-mouse IgG–conjugated to Cell proliferation was studied by monitoring the incorporation rhodamine fluorochrome (Sigma, Cat # T6528). of 5-bromo-20-deoxy-uridine (BrdU) into cellular DNA using BrdU labeling and detection kit II (from Roche, Indianapolis, Binding to GST-fusion proteins IN, USA). We followed the protocol suggested by the supplier. Cells staining positive for the incorporation of BrdU as well as Fusion proteins, in which glutathione transferase protein was total number of cells were counted in each sample. fused to various segments of IFI 16B proteins, were produced in bacteria and purified as described previously (Choubey et al., 1996). The purified fusion proteins bound to glutathione– Acknowledgements sepharose beads were used in GST-pull-down assays. The We thank Dr Qin for technical advice and Dr M Tainsky for assays were performed essentially as described previously sharing observations before publication. We also thank Drs O (Choubey et al., 1996). In brief, cell lysates containing equal Pereira-Smith, S Campbell, N Mohideen, and E Wojcik for amounts of proteins were incubated with beads for 30 min at their suggestions and advise during the course of this work. 41C. Beads were washed with wash buffer and the bound This work was supported by NIH Grants CA69031 and proteins were analysed by immunoblotting. AG022124 (to DC), and AR47307 (to BJN).

References

Borden EC, Lindner D, Dreicer R, Hussein M and Peereboom Choi J, Shendrik I, Peacocke M, Peehl D, Buttyan R, D. (2000). Semin. Cancer Biol., 10, 125–144. Ikeguchi EF, Katz AE and Benson MC. (2000). Urology, Briggs LJ, Johnstone RW, Elliot RM, Xiao CY, Dawson M, 56, 160–166. Trapani JA and Jans DA. (2001). Biochem. J., 353, 69–77. Choubey D. (2000). J. Biol. Regul. Homeost. Agents, 14, 187– Campisi J. (2001). Trends Cell. Biol., 11, S27–S31. 192. Castro P, Giri D, Lamb D and Ittmann M. (2003). Prostate, Choubey D and Kotzin BL. (2002). Front. Biosci., 7, e252– 55, 30–38. e262. Certa U, Seiler M, Padovan E and Spagnoli GC. (2001). Br. J. Choubey D and Lengyel P. (1992). J. Cell. Biol., 116, 1333– Cancer, 85, 107–114. 1341.

Oncogene Role of IFI 16 in prostate epithelial cellular senescence HXinet al 4840 Choubey D, Li SJ, Datta B, Gutterman JU and Lengyel P. Johnstone RW, Wei W, Greenway A and Trapani JA. (2000). (1996). EMBO J., 15, 5668–5678. Oncogene, 19, 6033–6042. Dawson MJ and Trapani JA. (1995). Biochem. Biophys. Res. Jonasch E and Haluska FG. (2001). Oncologist, 6, 34–55. Commun., 214, 152–162. Karlsson C, Stenman G, Vojta PJ, Bongcam-Rudoloff E, Decesse JT, Medjkane S, Datto MB and Cremisi CE. (2001). Barret JC, Westermark B and Paulsson Y. (1996). Cancer Oncogene, 20, 962–971. Res., 56, 241–245. DeGregori J. (2002). Biochim. Biophys. Acta, 1602, 131–150. Kingsmore SF, Watson ML, Woward TA and Seldin MF. Dimri GP, Lee X, Basile G, Acosta M, Scott G, Roskelley C, (1989). EMBO J., 8, 4073–4080. Medrano EE, Linskens M, Rubelj I and Pereira-Smith O. Kulaeva OI, Draghici S, Tang L, Kraniak JM, Land SJ and (1995). Proc. Natl. Acad. Sci. USA, 92, 9363–9367. Tainsky MA. (2003). Oncogene, 25, 4118–4127. Dimri GP, Nakanishi M, Desprez PY, Smith JR and Campisi Landolfo S, Gariglio M, Gribaudo G and Lembo D. (1998). J. (1996). Mol. Cell. Biol., 16, 2987–2997. Biochimie, 80, 721–728. D’Souza S, Xin H, Walter S and Choubey D. (2001). J. Biol. Lengyel P, Choubey D, Li SJ and Datta B. (1995). Semin. Chem., 276, 298–305. Virol., 6, 203–213. Fairbrother WJ, Gordon NC, Humke EW, O’Rourke KM, Mori JY, Rashid A, Leggett BA, Young J, Simms L, Kuehl Starovasnik MA, Yin JP and Dixit VM. (2001). Protein Sci., PM, Langenberg P, Meltzer SJ and Stine OC. (2001). Cancer 10, 1911–1918. Res., 61, 6046–6049. Gariglio M, Azzimonti B, Pagano M, Palestro G, De Andrea Schlaak JF, Hilkens CM, Costa-Pereira AP, Strobl B, M, Valente G, Voglino G, Navino L and Landolfo S. (2002). Frischauf AM and Kerr IM. (2002). J. Biol. Chem., 277, J. Interferon Cytokine Res., 22, 815–821. 49428–49437. Garraway LA, Lin D, Signoretti S, Walteregny D, Dilks J, Sen GC and Lengyel P. (1992). J. Biol. Chem., 267, 5017–5020. Bhattacharya N and Loda M. (2003). Prostate, 55, 206–218. Shou J, Soriano R, Hayward SW, Cunha GR, Williams PM Gartel AL, Ye X, Goufman E, Shianov P, Hay N, Najmabadi and Gao WQ. (2002). Proc. Natl. Acad. Sci. USA, 99, 2830– F and Tyner AL. (2001). Proc. Natl. Acad. Sci. USA, 98, 2835. 4510–4515. Smith JR and Pereira-Smith OM. (1996). Science, 273, 63–67. Gribaudo G, Riera L, De Andrea M and Landolfo S. (1999). Stark GR, Kerr IM, Williams BR, Silverman RH and FEBS Lett., 456, 31–36. Schreiber RD. (1998). Annu. Rev. Biochem., 67, 227–264. Gutterman JU. (1994). Proc. Natl. Acad. Sci. USA, 91, 1198– Staub E, Dahl E and Rosenthal A. (2001). Trends Biochem. 1205. Sci., 26, 83–85. Gutterman JU and Choubey D. (1999). Cell Growth Differ., Untergasser G, Koch HB, Menssen A and Hermeking H. 10, 93–100. (2002). Cancer Res., 62, 6255–6262. Harbour JW and Dean DC. (2000). Genes Dev., 14, 2393–2409. Varambally S, Dhanasekaran SM, Zhou M, Barrette TR, Hayﬂick L. (1965). Exp. Cell. Res., 37, 614–636. Kumar-Sinha C, Sanda MG, Ghosh D, Pienta KJ, Sewalt Hensler PJ, Annab LA, Barret JC and Pereira-Smith OM. RG, Otte AP, Rubin MA and Chinnaiyan AM. (2002). (1994). Mol. Cell. Biol., 14, 2291–2297. Nature, 419, 624–629. Hertel L, Rolle S, De Andrea M, Azzimonti B, Osello R, Wang H, Chatterjee G, Meyer JJ, Liu CJ, Manjunath NA, Gribaudo G, Gariglio M and Landolfo S. (2000). Oncogene, Bray-Ward P and Lengyel P. (1999). Genomics, 60, 281–294. 19, 3598–3608. Webber MM, Bello D and Quader S. (1997). Prostate, 30, Jarrard DF, Sarkar S, Shi Y, Yeager TR, Magrane G, 58–64. Kinoshita H, Nassif N, Meisner L, Newton MA, Waldman Wei W, Clarke CJ, Somers GR, Cresswell KS, Loveland KA, FM and Reznikoff CA. (1999). Cancer Res., 59, 2957–2964. Trapani JA and Johnstone RW. (2003). Histochem. Cell. Johnstone RW, Kershaw MH and Trapani JA. (1998). Biol., 119, 45–54. Biochemistry, 37, 11924–11931. Weinberg RA. (1995). Cell, 81, 323–330. Johnstone RW and Trapani JA. (1999). Mol. Cell. Biol., 19, Xin H, D’Souza S, Fang L, Lengyel P and Choubey D. (2001). 5833–5838. Oncogene, 20, 6828–6839.

Oncogene