The second loop of occludin is required for suppression of ... - Nature

Oncogene (2005) 24, 4412–4420

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

The second loop of occludin is required for suppression of Raf1-induced tumor growth Zili Wang1, Kenneth J Mandell1, Charles A Parkos1, Randall J Mrsny2 and Asma Nusrat*,1 1 2

Epithelial Pathobiology Research Unit, Department of Pathology, Emory University, 615 Michael Street, Atlanta, GA 30322, USA; Unity Pharmaceuticals, 11620 Buena Vista Dr., Los Altos Hills, CA 94022, USA

Tight junctions (TJs) regulate epithelial cell polarity and paracellular permeability. Loss of functional TJs is commonly associated with epithelial cell-derived cancers. Raf1-mediated transformation of rat salivary gland epithelial cells (Pa4-Raf1) induces transcriptional downregulation of the TJ protein occludin and forced reexpression of occludin rescues polarized phenotype of epithelial cells. In the present study, we used this model to examine how specific structural modifications in the occludin protein affect its function in vitro and influence tumor growth in vivo. Our results revealed that neither the C-terminal nor the N-terminal half of occludin alone were sufficient to rescue cells from transformation by Raf1. However, forced expression of an occludin mutant lacking the first extracellular loop was sufficient to rescue cells from Raf1-mediated transformation. Interestingly, forced expression of an occludin mutant lacking the second extracellular loop did not rescue the epithelial phenotype in vitro nor did it prevent tumor growth in vivo. These results demonstrate that the TJ protein occludin has a potent inhibitory effect on the Raf1-mediated tumorigenesis, and the second extracellular loop of occludin appears to be critical for this function. Oncogene (2005) 24, 4412–4420. doi:10.1038/sj.onc.1208634 Published online 4 April 2005 Keywords: tight junction; occludin; raf

Introduction A crucial function of epithelial cells is the maintenance of a regulated barrier separating the external from internal environments. The apical junctional complex (AJC) consisting of the tight junction (TJ) and adherens junction (AJ) is important in determining epithelial polarity and barrier properties. In addition, TJs are involved in signal transduction pathways linked to cell *Correspondence: A Nusrat, Epithelial Pathobiology Research Unit, Department of Pathology and Laboratory Medicine, Emory University School of Medicine, 615 Michael Street, Room 105E, Atlanta, GA 30322, USA; E-mail: [email protected] Received 22 November 2004; revised 15 February 2005; accepted 17 February 2005; published online 4 April 2005

growth and proliferation. Since disruption of barrier function, cell polarity and cell growth are common characteristics of epithelial cancers, TJ proteins appear to be attractive candidates for study of events associated with epithelial oncogenic transformation. The TJ complex is comprised of transmembrane proteins, scaffolding proteins, regulatory enzymes, and transcription factors (Nakamura et al., 2000; GonzalezMariscal et al., 2003; Matter and Balda, 2003). Occludin is a TJ integral membrane protein that is predicted to be comprised of four transmembrane domains with two extracellular loops, cytosolic amino and carboxy termini (Furuse et al., 1993; Ando-Akatsuka et al., 1996). Evidences suggest that occludin is directly involved in TJ barrier and fence function (Balda et al., 1996; McCarthy et al., 1996; Bamforth et al., 1999; Tsukita et al., 2001), and in cell adhesion events (Van Itallie and Anderson, 1997). Various studies have investigated how specific occludin domain deletions or substitutions affect its localization at TJs and its role in barrier function (Furuse et al., 1994; Balda et al., 1996; Chen et al., 1997). Overexpression of occludin mutants has been shown to affect TJ function in cultured epithelial cells (Ando-Akatsuka et al., 1996; Bamforth et al., 1999) and synthetic peptides corresponding to the extracellular loops of occludin were observed to disrupt TJs and inhibit cell adhesion (Wong and Gumbiner, 1997; Lacaz-Vieira et al., 1999). These studies largely agree that occludin plays a central role in TJ function, and that mutation of specific occludin domains affects its localization and function in cells. No study to date, however, has investigated how specific structural modifications of the occludin protein affect cell growth and proliferation, particularly with respect to suppressing oncogenic transformation. Disruption of TJ structures is a common feature of many human epithelial cancers. Downregulation of specific TJ proteins has been shown to correlate with staging and metastatic potential in various cancers. In endometrial cancers, downregulation of occludin was shown to correlate with tumor grade and invasiveness (Tobioka et al., 2004), and occludin expression was also observed to decrease in poorly differentiated gastrointestinal adenocarcinomas (Kimura et al., 1997). Besides occludin, other TJ-associated proteins such as ZO-1 and claudins have been shown to be important indicators of malignant potential. In some breast

Second occludin loop required for Raf1 tumor suppression Z Wang et al

4413

cancers, ZO-1 expression was shown to decrease in more malignant forms (Hoover et al., 1998), and expression of claudin-7 was correlated with histological grade (Kominsky et al., 2003). In addition to TJ proteins serving as prognostic indicators, there is mounting evidence to suggest that introduction of TJ proteins into cancer cells can prevent tumor invasion and metastasis. Overexpression of claudin-4 has been observed to decrease invasiveness of pancreatic cancer cells in vitro and to decrease metastases in animal models (Michl et al., 2003). These studies provide promising evidence that TJ proteins may serve as useful molecular targets for both diagnosis and treatment of epithelial cancers. Raf-1 is a downstream effector of the ras oncogene. Approximately 10–20% of all human cancers involve mutations in Ras proteins (de Vries et al., 1996), and mutations resulting in a constitutively active Ras have been linked to various epithelial cancers (Vogelstein et al., 1988). Ras proteins are known to control cellular proliferation/differentiation by activation of Raf1, which in turn activates the mitogen-activated protein kinase (MEK)/extracellular-regulated kinase (ERK) pathway (Cobb et al., 1994). Given the prevalence of Ras mutations in human cancers and its known role in cellular growth and proliferation, Ras and its downstream effectors, such as Raf1, are prime candidates for diagnosis and treatment of epithelial cancers (Nottage and Siu, 2002). Our previous studies have demonstrated that constitutively active Raf1-induced transformation of Pa4 cells is associated with transcriptional downregulation of occludin. Introduction of exogenous occludin into Raf1transformed cells was observed to rescue the epithelial

phenotype and induce reassembly of functional TJs (Li and Mrsny, 2000). We applied this rescue-offunction approach to investigate whether occludin expression can inhibit Raf1-transformed tumor growth in vivo and examined the contribution of specific occludin domains on their ability to suppress Raf1induced transformation of Pa4 cells. Our results suggest that the second loop of occludin is essential for its role in rescuing the epithelial phenotype, and its absence permits tumor growth in vivo. These results provide new insight into the structural basis for occludinmediated tumor suppression. Such knowledge may lead to novel molecular approaches for restoring occludin function in the treatment of epithelial cancers.

Results Occludin rescues the epithelial phenotype in Raf-1-transformed cells Our previous study (Li and Mrsny, 2000) demonstrated that introduction of constitutively active Raf1 induces transcriptional downregulation of endogenous occludin in Pa4 cells that coincides with acquisition of a transformed phenotype. Introduction of the human occludin gene induces forced expression of human occludin protein and reversal of the Raf1-induced oncogenic phenotype. To investigate the roles of specific structural elements within the occludin protein on this Raf1-occludin dynamic, experiments were first performed to validate this rescue-of-function model in Pa4-Raf1 cells. Pa4 cells formed tight colonies of polarized epithelial cells (Figure 1a) and revealed a

Figure 1 Occludin rescues the epithelial phenotype in Raf1-transformed Pa4 cells. Pa4 cells were stably transfected to express either constitutively active Raf1 (Pa4-Raf1) or Raf1 and full-length human occludin (Occ). (a) Phase-contrast images of subconfluent cultures of Pa4 and Pa4-Raf1 and Occ cells. (b) Immmunofluorescence labeling of occludin, claudin-1, ZO-1 and E-cadherin in Pa4, Pa4-Raf1 and Occ cells. (c) Immunoblots illustrating the changes in junctional protein expression in Raf-1-transformed Pa4 cells. (d) Transepithelial resistance (TER) in Pa4, Pa4-Raf1 and Occ cells grown on permeable transwell filters Oncogene


4414

highly organized ‘chicken wire’ staining pattern of TJ and AJ proteins typical of polarized epithelial cells (Figure 1b). Stable expression of constitutively active Raf1, however, induced transformation of Pa4 epithelial cells into spindle-shaped mesenchyme-type cells (Figure 1a) that no longer expressed junctional proteins at the plasma membrane (Figure 1b). Western blots showed that Pa4-Raf1 cells significantly downregulated the expression of occludin (Figure 1c) and decreased the expression of claudin-1 while E-cadherin and ZO-1 were not significantly changed (Figure 1c). Since Pa4-Raf1 cells were generated by fusing the catalytic domain of the Raf-1 with the hormone-binding domain of the estrogen receptor (DRaf-1: ER) (Li et al., 1997), we could markedly increase Raf1 expression by incubating cells with 1 mM estradiol. Under such conditions, Pa4Raf1 cells completely downregulate the expression of occludin (data not shown). Pa4-Raf1 cells no longer developed measurable transepithelial electrical resistance (TER) values when cultured on semipermeable supports (Figure 1d). Consistent with previous studies (Li and Mrsny, 2000), forced expression of human occludin protein into Pa4-Raf1 cells (Occ) recovered epithelial morphology (Figure 1a), junctional staining patterns of TJ and AJ proteins (Figure 1b) and barrier function (Figure 1d) similar to that observed for Pa4 cells. Additionally, expression of an intermediate filament, vimentin that is abundantly expressed in cells of mesenchymal origin, was markedly upregulated in Pa4-Raf1 cells and downregulated by forced expression of occludin in Pa4-Raf1 cells (data not shown). These results reaffirm that forced expression of full-length occludin in Raf1-transformed cells can promote reacquisition of an epithelial phenotype and the formation of functionally intact TJs in Pa4 cells. The second extracellular loop of occludin is required for rescue of epithelial cell morphology Occludin has been predicted to possess four transmembrane domains with two extracellular domains and cytosolic amino and carboxy termini (Furuse et al., 1993). Given the observation that full-length occludin can revert Raf1-transformed Pa4 cells back to an epithelial phenotype, experiments were conducted to investigate whether specific occludin domain deletions affect this function. As illustrated in Figure 2a, a panel of occludin deletion constructs were made, all with a 10amino-acid myc tag engineered at the C-terminus to facilitate immunodetection. These occludin deletion constructs were then stably transfected into Pa4-Raf1 cells by double selection for G418 and Zeocin resistance. Transgene expression in Pa4-Raf1 cells was confirmed by immunoblotting with the anti-myc antibody (Figure 2b) and anti-occludin antibody (data not shown). Western blotting of cell lysates revealed that the transgenes were expressed at a comparable level in all transfected cells examined. As illustrated in Figure 2c, introduction of wild-type occludin into Raf1-transformed Pa4 cells (Occ) completely rescued epithelial morphology. In contrast, expression of conOncogene

structs lacking the N-terminal half of occludin (OccDN), C-terminal half of occludin (OccDC), or the second loop of occludin (OccDL2) failed to rescue epithelial cell morphology (Figure 2c, top panel) and organized distribution of ZO-1 in intercellular contacts was not observed (Figure 2c, bottom panel). Interestingly, introduction of occludin lacking the first extracellular loop into Raf1-transformed Pa4 cells (OccDL1) resulted in normal epithelial cell morphology (Figure 2c, top panel) and lateral staining pattern of TJ and AJ proteins consistent with intact intercellular junctions (ZO-1, Figure 2c, bottom panel). These results suggest that both the N-terminal and C-terminal domains of occludin are critical for rescue of epithelial cell morphology in Raf1-transformed Pa4 cells. In addition, it appears that, within the extracellular domain of occludin, the second loop but not the first loop is essential for this process. The second loop of occludin is required for assembly of junctional complexes Having observed differences in the morphology and staining patterns of Pa4-Raf1 cells expressing the OccDL1 and OccDL2 mutants, experiments were performed to further investigate the staining patterns of specific intercellular junction proteins in these cell lines. As shown in Figure 3a, expression of the OccDL1 mutant resulted in well-organized junctional staining with typical ‘chicken wire’ patterns for occludin, ZO-1, claudin-1 and E-cadherin. In addition, the confocal z-sections shown in Figure 3b illustrate that OccDL1 cells were able to form polarized monolayers with distinct junctional complexes visible at regular intervals near the apical surface of the monolayer. In contrast, cells expressing the OccDL2 mutant exhibited highly disorganized staining for junctional proteins occludin, ZO-1, claudin-1 and E-cadherin (Figure 3a) and failed to form polarized monolayers with well-defined AJCs (Figure 3b). These results suggest that the second loop of occludin is essential for the assembly of intact junctional complexes and formation of polarized epithelial monolayers. Nonionic detergent insolubility is considered an indicator of protein incorporation into cytoskeletonassociated junctional complexes (Sakakibara et al., 1997). To complement the morphologic and confocal data (Figures 2 and 3), we performed biochemical characterizations of junctional protein expression and association with the nonionic detergent Triton X-100 (TX-100) soluble (S) or insoluble (I) fractions of cell homogenates prepared from Pa4, Pa4-Raf1, OccDL1 or OccDL2 cells (Figure 4). Hyperphosphorylated occludin expressed in Pa4 cells is mostly associated with the I fraction of these cells while claudin-1, ZO-1 and E-cadherin are equally present in the I and S fractions. Introduction of constitutively active Raf1 (Pa4-Raf1) results in loss of normal junction organization exemplified by the loss of expressed occludin and a shift of claudin-1, ZO-1 and E-cadherin out of the I fraction of these cells (Figure 4). Occludin protein lacking the first


4415

Figure 2 The second loop of occludin is essential for rescue of epithelial phenotype. (a) Full-length occludin and constructs encoding specific occludin deletions were expressed in Pa4-Raf1 cells. All constructs contain a C-terminal myc-tag. Occ is full-length human occludin; OccDN lacks the N-terminal tail and both extracellular loops; OccDC lacks C-terminal cytoplasmic domain; OccDL1 lacks the first extracellular loop; OccDL2 lacks the second extracellular loop. (b) Expression of myc-tagged occludin mutants in stably transfected Pa4-Raf1 cells was detected with anti-myc mAb. Myc-tagged occludin mutants OccDN, OccDC, OccDL1 and OccDL2 with the expected molecular masses were obtained. (c) Top panel, phase-contrast images of Pa4-Raf1 cells expressing full-length occludin or occludin mutants. Bottom panel, immmunofluorescence staining in Pa4-Raf1 cells expressing full-length occludin or occludin mutants with antibodies against ZO-1 (red) and myc-tagged occludin constructs (green). Scale bar: 20 mm

loop (OccDL1) expressed in Pa4-Raf1 cells was found to distribute both I and S fractions similarly to that observed for Pa4 cells. Importantly, the solubility profile for claudin-1, ZO-1 and E-cadherin in the OccDL1 cells greatly resembled that of the parental Pa4 cells (Figure 4). In contrast, Pa4-Raf1 cells expressing OccDL2 demonstrated proportionately less occludin, claudin-1, ZO-1 and E-cadherin in the detergentinsoluble fractions, and this pattern closely resembled that of oncogenic Pa4-Raf1 cells (Figure 4). The differences in solubility profiles of AJC proteins observed for the OccDL1 and OccDL2 cell lines are consistent with the differences in cell morphology and junctional staining patterns.

The second loop of occludin is required for TJ barrier function The extracellular loops of occludin have been reported to be important for regulation of epithelial TJ barrier function. Experiments were performed to investigate the effect of the OccDL1 and OccDL2 mutants on transepithelial resistance (TER) and FITC-dextran flux. As shown in Figure 5a, Pa4-Raf1 cells expressing wild-type occludin (Occ) formed monolayers with TER values of >900 O cm2, and cells expressing the OccDL2 mutant did not produce TER values above background (o80 O cm2). Interestingly, Pa4-Raf1 cells expressing the OccDL1 mutant produced resistance as high as Oncogene


4416

Figure 3 The second loop of occludin is essential for localization of occludin, claudin-1, E-cadherin and ZO-1 at intercellular junctions. (a) En face immunofluorescence confocal images of Pa4-Raf1 cells expressing the occludin lacking the first loop (OccDL1) and the second loop (OccDL2). (b) Confocal images in the xz plane illustrate localization of Pa4 cells and Pa4-Raf1 cells expressing full-length occludin (Occ), OccDL1 and OccDL2. Scale bar: 20 mm

Figure 5 The second extracellular loop of occludin is required for TJ barrier function. Pa4-Raf1 cells expressing full-length occludin or occludin mutants lacking either the first extracellular loop or the second extracellular loop were grown on permeable transwell filters and subjected to assays of barrier function. (a) TER to passive ion flux, (b) paracellular flux of FD-3 (fluorescent dextran; 3000 Da)

The second loop of occludin is required for anchorage-dependent growth in soft agarose

Figure 4 Detergent solubility profiles for junctional proteins. Triton X-100-soluble (S) and -insoluble (I) protein fractions were prepared from Pa4-Raf1 cells expressing occludin lacking the loop1 (OccDL1) and loop2 (OccDL2) and compared with extracts from Pa4 cells and Pa4-Raf1 cells. Proteins fractions were immunoblotted for the junctional proteins occludin, claudin-1, E-cadherin, and ZO-1 and actin

500 O cm2, suggesting development of functional TJs. This observation is supported by FITC-dextran flux assays that demonstrated that OccDL1 cells have threefold higher paracellular flux compared to cells expressing wild-type occludin (Figure 5b). As expected, forced expression of OccDL2 in Pa4-Raf1 cells did not exhibit any restriction to FD-3 flux (data not shown). These results suggest that the second loop of occludin is necessary for TJ barrier function, but the first loop of occludin may play a role in establishing and/or maintaining TJ barrier function. Oncogene

Loss of anchorage-dependent growth on soft agarose is considered an indicator of oncogenic transformation. In the case of Pa4 cells, transformation with constitutively active Raf1 has been shown to confer the capacity for anchorage-independent growth, while overexpression of exogenous occludin was observed to restore anchorage dependence (Li and Mrsny, 2000). To investigate whether specific occludin domains are required to suppress anchorage-independent growth, Pa4-Raf1 cells expressing various occludin mutants or full-length occludin were subjected to growth assays in soft agar (Figure 6a). Pa4-Raf1 cells expressing the OccDC, OccDN and OccDL2 mutants formed numerous colonies on soft agarose, averaging 105, 100, and 80 colonies, respectively. In contrast, expression of full-length occludin (Occ) or the OccDL1 mutant significantly inhibited growth on soft agarose, with only two and seven colonies on average, respectively (Figure 6b). These results suggest that the second loop of occludin is essential for suppression of anchorage-independent growth. The second loop of occludin is required for suppression of Raf1-mediated tumor formation in nude mice Having observed that expression of exogenous occludin can rescue the epithelial phenotype of Raf1-transformed Pa4 cells in vitro, experiments were performed to further investigate whether the occludin can suppress


4417

Figure 6 The second extracellular loop of occludin is required to suppress Raf1-induced anchorage-independent growth. Pa4-Raf1 cells expressing occludin constructs with specific domain deletions were plated into 35-mm dishes in soft agar. After 4 weeks of culture, colonies >0.3 mm in diameter were counted. (a) Representative photographs from cultured dishes from each cell line. Size bar: 2 mm, (b) the average number of colonies from cultured dishes for each cell line

Raf1-mediated transformation in vivo and whether specific domain deletions affect its ability to do so. Pa4-Raf1 cells expressing full-length occludin or various mutants (Figure 2a) were injected subcutaneously into the flanks of nude mice. Tumor formation was assayed by visual inspection. Injection of Pa4 cells or Pa4-Raf1 cells expressing either full-length occludin (Occ) or the OccDL1 failed to produce tumors 6 weeks postinjection (Figure 7a). Conversely, injection of Pa4-Raf1 cells expressing OccDL2 or OccDC or OccDN, or Pa4-Raf1 cells produced large visible tumors. Histologic analysis of tumor sections revealed that all tumors consisted predominantly of spindle-shaped tumor cells (Figure 7b). The morphological features of tumor cells were identical in all the groups examined (Pa4-Raf1, OccDL2, OccDC, OccDN). Furthermore, immunohistochemical analysis demonstrated that the tumors derived from Pa4-Raf1 cells downregulated the expression of TJ protein occludin (Figure 7c). In contrast to occludin, tumor cells expressed ZO-1 and E-cadherin. Additionally, tumor cells expressing occludin deletion constructs expressed ZO-1 and E-cadherin analogous to that observed in Pa-Raf1 cells (data not shown). Together,

Figure 7 The second extracellular loop of occludin is required for preventing tumor formation in nude mice. Pa4 cells, Pa4-Raf1 cells and Pa4-Raf1 cells expressing full-length occludin and various occludin mutants were injected subcutaneously into the flanks of nude mice, respectively, at 1 106 cells per injection site. Tumor formation was monitored daily, and mice were killed 6 weeks after injection. Tumors were isolated and analysed by histology and immunohistochemistry. (a) Table documenting the expression of active Raf-1 and average tumor weights from four mice injected with each cell line. (b) Representative hematoxylin & eosin (H&E) staining of Pa4-Raf1 tumor sections. (c) Occludin, ZO-1 and Ecadherin expression in tumors derived from Pa4-Raf1 cells was determined by immunohistochemistry and light microscopy

these results support the in vitro studies by demonstrating that Raf1 mediates transformation of Pa4 cells by downregulation of TJ protein occludin, expression of exogenous occludin (or occludin lacking only the first extracellular loop) can suppress Raf1-induced tumor growth in nude mice, and the second loop of occludin is required for growth suppression of Raf1-transformed epithelial cells.

Discussion Oncogene Raf-1 is a key regulator of the MAP kinase cascade. Raf-1 activation involves sequential phosphorylation and activation of mitogen-activated protein kinase (MEK) and extracellular-regulated kinase (ERK) kinases and regulates a variety of cellular processes including proliferation, differentiation, apoptosis, and transformation (Cobb et al., 1994; Morrison and Cutler, 1997). Recently, it has been reported that active Raf-1 induces transition of epithelial cells to a Oncogene


4418

mesenchymal phenotype (EMT), which is associated with loss of intercellular junctions, loss of epithelial phonotype and acquisition of an invasive phenotype (Lehmann et al., 2000; Lan et al., 2004). These studies support our present results showing Raf-1 induced transformation of epithelial cells into spindle shaped mesenchymal type cells. Epithelial TJs are important in regulating cell polarity. While disruption of TJ proteins has been reported in epithelial cancers, their role in regulation of oncogenic transformation remains largely unexplored. We demonstrated that overexpression of TJ protein occludin completely rescued the epithelial phenotype of Raf1-transformed Pa4 cells (Pa4-Raf1) in vitro as evidenced by recovery of normal epithelial morphology that coincided with re-expression of critical TJ proteins and their localization to TJ structures. Occludincorrected Pa4-Raf1 cells lost their ability of anchorage-independent growth in soft agarose and their capacity to form tumors in nude mice following subcutaneous injection. In addition, we present direct evidence to show that the second loop of occludin is required for repression of Raf1-mediated transformation in Pa4 cells. Previous studies have shown that the extracellular domains of occludin are directly involved in cell–cell adhesion (Van Itallie and Anderson, 1997), and homophilic occludin interactions between cells (Furuse et al., 1998). Administration of a peptide corresponding to the second loop of occludin disrupted TER and resulted in the disappearance of occludin from cell junctions (Wong and Gumbiner, 1997). Thus, the second extracellular loop of occludin appears to be important for concentrating and/or stabilizing the occludin protein within TJ complexes. Consistent with these reports, we observed that expression of occludin lacking the second extracellular loop in Pa4-Raf1 cells did not rescue epithelial morphology or restore TJ structure/function. It is interesting, however, that the occludin mutant lacking the first loop localized at TJs (Figure 2) and promoted assembly of TJ complexes (Figures 3 and 4) and yet only partially rescued barrier function (Figure 5). These results suggest that the second loop is required for assembly of occludin into the TJ complex, but the first loop may play a role in establishing and/or maintaining paracellular permeability properties of the TJ. Various studies have investigated the effect of specific mutations on occludin localization and TJ function. Although many of these studies agree that the occludin protein is a critical component of the TJ, some inconsistencies have arisen regarding whether specific occludin domains are required for its targeting and formation of functional TJs. For instance, occludin constructs lacking the entire C-terminal tail were observed to localize at the TJs (Balda et al., 1996; Chen et al., 1997) and the C-terminal tail alone is insufficient for localization (Matter and Balda, 1998). These results, however, are at variance with another study that reported that a C-terminally truncated occludin construct failed to localize at TJs (Furuse et al., 1994; Mitic et al., 1999; Medina et al., 2000). Some of these apparent Oncogene

inconsistencies may be accounted for by subtle differences in the occludin constructs and the origin of the cell lines used in these studies. In particular, presence of endogenous occludin and its binding partners may influence the localization of transfected occludin constructs and their affect on TJ function. A unique aspect of our study is the use of the Raf1-transformed Pa4 cells. The Pa4-Raf1 cells were generated by stably transfected Pa4 cells with constitutively active Raf1, which was constructed by fusing the catalytic domain of the Raf-1 with the hormone binding domain of the estrogen receptor (DRaf-1:ER) (Li et al., 1997; Li and Mrsny, 2000). Although the kinase activity of DRaf-1:ER can be further increased in the presence of added estradiol, the DRaf-1:ER fusion protein has a high level of basal activity. This basal activity of DRaf-1:ER is sufficient to downregulate expression of occludin and claudin-1 in Pa4 cells, and to induce disruption of TJs and oncogenic transformation. Thus, most of our studies were carried out in the absence of added estradiol. TJ disruption has been observed in many epithelial cancers and this downregulation has been shown to correlate with invasiveness of cancer. Thus, for example, downregulation of occludin was shown to correlate with tumor grade and tendency to metastasize in endometrial cancers (Kimura et al., 1997; Tobioka et al., 2004). These clinical observations are well supported by celland animal-based studies that suggest expression of TJ proteins can be protective against tumor invasion and metastasis in epithelial cancers. Overexpression of claudin-4 has been shown to decrease invasiveness of pancreatic cancer cells in vitro and to decrease metastasis in animal models (Kominsky et al., 2003; Michl et al., 2003). Evidences from our studies suggest that expression of the TJ protein occludin can suppress Raf1transformation of epithelial cells and the second loop of occludin is required for reversing EMT phenotype changes associated with activation of Raf1. EMT reversal by occludin is also associated with formation of functional TJs and re-acquisition of a polarized phenotype characteristic of functional epithelia. Importantly, these results demonstrate that TJs play a pivotal role in regulation of epithelial carcinomas and suggest that the TJ protein occludin participates in a functional dynamic with the Raf1 oncogene to control events associated with EMT. Further studies will expound the molecular mechanism by which occludin inhibits Raf1induced transformation.

Materials and methods Occludin deletion constructs Occludin deletion constructs were generated by PCR-based mutagenesis using cDNA coding for human occludin as template. PCR amplifications were performed with PFU DNA polymerase. Products were subcloned into the pCRBlunt II-TOPO vector, confirmed by sequence analysis, and then cloned into the pcDNA4-myc vector (Invitrogen, Carlsbad, CA, USA). A 10-amino-acid myc tag was engineered at the C-terminus of each occludin deletion construct to


4419 facilitate immunodetection. The first extracellular loop of human occludin is predicted to include 46 residues encompassing residues D90 and R135. The truncated first extracellular loop, OccDL1, contained amino acids D90RGYGGYTDPR135 resulting from deletion of residues 94–128. The second loop of human occludin is predicted to include 48 residues from G196 to E243. The truncated second extracellular loop, OccDL2, contained the amino acids G196VNPVDPQE243 resulting from deletion of residues 200–238. The construct, OccDC, consists of the N-terminal domain and extracellular loops (amino acids 1–266), and lacks the C-terminal domain. The construct, OccDN, contains only the C-terminal domain (amino acids 260–522). Cell culture and generation of stable cell lines Pa4 cells, Pa4-Raf1cells and Occ (Pa4-Raf1-Occ) cells were established and cultured as described previously (Li et al., 1997; Li and Mrsny, 2000). Pa4-Raf1 cells were transfected with various occludin deletion constructs using a Lipofectaine Plus reagent (Invitrogen, Carlsbad, CA, USA). Stable cell lines were selected in G418 (500 mg/ml) and Zeocin (200 mg/ml). Resistant colonies were isolated and maintained in Dulbecco’s modified Eagle’s/F12 (1:1, phenol red-free) medium supplemented with 2.5% charcoal-stripped fetal bovine serum, G418 (500 mg/ml), and Zeocin (200 mg/ml). Immunohistochemistry and microscopy Cells grown on polycarbonate filters (Costar, NK) were fixed/ permeabilized in absolute methanol for 20 min at 201C followed by incubation in HBSS þ containing 5% normal goat serum (blocking buffer) for 1 h at room temperature and incubation for 60 min with antibodies to human occludin, claudin-1 and ZO-1, E-cadherin (Zymed Labs, San Francisco, CA, USA). Cells were washed, incubated for 60 min with Alexa-conjugated secondary antibodies followed by rinsing and mounting on slides with ProLong Antifade medium (Molecular Probes, Eugene, OR, USA). Cells were analysed using a Zeiss LSM510 laser scanning confocal microscope (Zeiss Microimaging, Thornwood, NY, USA) coupled to a Zeiss Axioplan2e with 63 or 100 Pan-Apochromat oil lenses. Formalin-fixed and paraffin-embedded tissue sections were deparaffinized in xylene, immersed in 3% hydrogen peroxide for 30 min, and rehydrated through graded ethanol. Antigen retrieval was performed by immersing sections in 0.01 M sodium citrate, pH 6.0, and exposure in a microwave for 20 min. Sections were then cooled to room temperature, and incubated in blocking buffer (2% milk, 0.05% Tween-20 in PBS). Sections were then incubated with the primary antibodies (occludin, ZO-1, E-cadherin) for 1 h, followed by sequential incubation with biotinylated second antibody (DakoCytomation, Inc, Carpinteria, CA, USA) for 45 min and peroxidase labeled streptavidin for 30 min. The reaction was visualized using 3,30 -diaminobenzamidine (DAB) as per the manufacturer’s instructions (DakoCytomation, Inc., Carpinteria, CA, USA). Sections were then counterstained in hematoxylin (Richard-Allan Scientific, Kalamazoo, MI, USA) for 4 min. Lastly, sections were dehydrated through graded ethanol, cleared in xylene, mounted, and coverslipped. Western blots and detergent solubility assays To determine expression of the occludin deletion constructs in Pa4-Raf1 cells, cell lines were plated onto six-well tissue culture plates. Lysed cells were analysed by Western blotting with a mouse anti-myc antibody (Invitrogen, Carlsbad, CA,

USA) and mouse anti-occludin antibody (Zymed Laboratories, San Francisco, CA, USA). HRP-conjugated secondary antibodies and the enhanced chemiluminescence detection system (NENt Life Science Products, Inc.) were used to detect bound antibodies. To isolate Triton X-100-soluble and -insoluble pools, cells were lysed in 1% Triton X-100 buffer containing 100 mM NaCl, 10 mM HEPES, 2 mM EDTA, and a cocktail of protease inhibitors. Lysates were incubated at 41C for 10 min, and then centrifuged at 15 000 g for 30 min at 41C. This supernatant was considered the Triton X-100-soluble pool. The pellet was solubilized in 1% SDS and referred to as the Triton X-100insoluble pool. Protein concentrations were determined by Pierce BCA assay. Equal amounts of cellular lysate protein were run on SDS–PAGE, electroblotted onto PVDF membranes, blocked with 5% milk solution for 1 h, incubated with primary antibodies, and then HRP-conjugated secondary antibodies. The enhanced chemiluminescence detection system (NENt Life Science Products, Inc.) was used to detect bound antibodies. Measrement of TER and paracellular permeability TER was measured across cells grown on polycarbonate filters using an eithelial voltometer (EVOM) (Precision World Instruments, Sarasota, FL, USA). TER values were normalized for the area of the filter and were obtained by subtracting the contribution of the filter and bathing solution. Paracellular permeability to fluoresceinated dextran (FD-3; MW 3000) was assessed as previously described (Sanders et al., 1995; Hopkins et al., 2003). Numerical values from individual experiments were pooled and expressed as mean7standard error of the mean. Data shown are representative of at least three experiments. Measurement of cloning efficiency in soft agarose Pa4 cells, Pa4-Raf1 cells and Pa4-Raf1 cells expressing fulllength occludin or various occludin mutants were seeded at a density of 1 104 cells per 35-mm culture dish in 1 ml of 0.35% (wt/vol) low melting point agarose in complete medium. Dishes were coated with 1 ml of 0.7% (wt/vol) low melting point agarose before cell plating, and 1 ml of overlay medium was added after cell plating. The dishes were incubated at 351C in 5% CO2 and 95% air for 4 weeks. The overlay medium was changed every 3 days. After 4 weeks, the cells were stained with 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT; 0.05 mg/ml). The stained dishes were photographed, and colonies >0.3 mm in diameter were counted and analysed. Tumorigenicity in nude mice Nude mice were used to assay tumorigenicity in vivo. Pa4 cells, Pa4-Raf1 cells and Pa4-Raf1 cells expressing full-length occludin or various occludin mutants were grown to logarithmic growth phase, harvested, washed and resuspended in PBS for injection. Each cell line (1 106 cells in 200 ml PBS) was injected subcutaneously into the flanks of nude mice. Tumor formation was monitored daily. All mice were killed at 6 weeks. Tumors were isolated and analysed. Acknowledgements We thank GT Brown and M Utech for help in manuscript preparation, A Akyildiz for technical assistance and D Hunt for both. NIH Grants DK 59888, DK64399, DK61379, DK72564 and the Crohn’s and Colitis Foundation of America supported this work. Oncogene


4420 References Ando-Akatsuka Y, Saitou M, Hirase T, Kishi M, Sakakibara A, Itoh M, Yonemura S, Furuse M and Tsukita S. (1996). J. Cell Biol., 133, 43–47. Balda MS, Whitney JA, Flores C, Gonzalez S, Cereijido M and Matter K. (1996). J. Cell Biol., 134, 1031–1049. Bamforth SD, Kniesel U, Wolburg H, Engelhardt B and Risau W. (1999). J. Cell Sci., 112, 1879–1888. Chen Y, Merzdorf C, Paul DL and Goodenough DA. (1997). J. Cell Biol., 138, 891–899. Cobb MH, Hepler JE, Cheng M and Robbins D. (1994). Semin. Cancer Biol., 5, 261–268. de Vries JE, ten Kate J and Bosman FT. (1996). Pathol. Res. Pract., 192, 658–668. Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura S and Tsukita S. (1993). J. Cell Biol., 123, 1777–1788. Furuse M, Itoh M, Hirase T, Nagafuchi A, Yonemura S and Tsukita S. (1994). J. Cell Biol., 127, 1617–1626. Furuse M, Sasaki H, Fujimoto K and Tsukita S. (1998). J. Cell Biol., 143, 391–401. Gonzalez-Mariscal L, Betanzos A, Nava P and Jaramillo BE. (2003). Prog. Biophys. Mol. Biol., 81, 1–44. Hoover KB, Liao SY and Bryant PJ. (1998). Am. J. Pathol., 153, 1767–1773. Hopkins AM, Walsh SV, Verkade P, Boquet P and Nusrat A. (2003). J. Cell Sci., 116, 725–742. Kimura Y, Shiozaki H, Hirao M, Maeno Y, Doki Y, Inoue M, Monden T, Ando-Akatsuka Y, Furuse M, Tsukita S and Monden M. (1997). Am. J. Pathol., 151, 45–54. Kominsky SL, Argani P, Korz D, Evron E, Raman V, Garrett E, Rein A, Sauter G, Kallioniemi OP and Sukumar S. (2003). Oncogene, 22, 2021–2033. Lacaz-Vieira F, Jaeger MM, Farshori P and Kachar B. (1999). J. Membr. Biol., 168, 289–297. Lan M, Kojima T, Osanai M, Chiba H and Sawada N. (2004). Carcinogenesis, 25, 2385–2395. Lehmann K, Janda E, Pierreux CE, Rytomaa M, Schulze A, McMahon M, Hill CS, Beug H and Downward J. (2000). Genes Dev., 14, 2610–2622.

Oncogene

Li D and Mrsny RJ. (2000). J. Cell. Biol., 148, 791–800. Li D, Lin HH, McMahon M, Ma H and Ann DK. (1997). J. Biol. Chem., 272, 25062–25070. Matter K and Balda MS. (1998). J. Cell Sci., 111, 511–519. Matter K and Balda MS. (2003). Nat. Rev. Mol. Cell Biol., 4, 225–236. McCarthy KM, Skare IB, Stankewich MC, Furuse M, Tsukita S, Rogers RA, Lynch RD and Schneeberger EE. (1996). J. Cell Sci., 109, 2287–2298. Medina R, Rahner C, Mitic LL, Anderson JM and Van Itallie CM. (2000). J. Membr. Biol., 178, 235–247. Michl P, Barth C, Buchholz M, Lerch MM, Rolke M, Holzmann KH, Menke A, Fensterer H, Giehl K, Lohr M, Leder G, Iwamura T, Adler G and Gress TM. (2003). Cancer Res., 63, 6265–6271. Mitic LL, Schneeberger EE, Fanning AS and Anderson JM. (1999). J. Cell Biol., 146, 683–693. Morrison DK and Cutler RE. (1997). Curr. Opin. Cell Biol., 9, 174–179. Nakamura T, Blechman J, Tada S, Rozovskaia T, Itoyama T, Bullrich F, Mazo A, Croce CM, Geiger B and Canaani E. (2000). Proc. Natl. Acad. Sci. USA, 97, 7284–7289. Nottage M and Siu LL. (2002). Curr. Pharm. Des., 8, 2231–2242. Sakakibara A, Furuse M, Saitou M, Ando-Akatsuka Y and Tsukita S. (1997). J. Cell Biol., 137, 1393–1401. Sanders SE, Madara JL, McGuirk DK, Gelman DS and Colgan SP. (1995). Epithelial. Cell Biol., 4, 25–34. Tobioka H, Isomura H, Kokai Y, Tokunaga Y, Yamaguchi J and Sawada N. (2004). Hum. Pathol., 35, 159–164. Tsukita S, Furuse M and Itoh M. (2001). Nat. Rev. Mol. Cell Biol., 2, 285–293. Van Itallie CM and Anderson JM. (1997). J. Cell Sci., 110, 1113–1121. Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM and Bos JL. (1988). N. Engl. J. Med., 319, 525–532. Wong V and Gumbiner BM. (1997). J. Cell Biol., 136, 399–409.