Cell adhesion and prostate tumor-suppressor activity of TSLL2/IGSF4C, an immunoglobulin superfamily molecule homologous to TSLC1/IGSF4


The TSLL2/IGSF4C encodes an immunoglobulin (Ig) superfamily molecule showing significant homology with a lung tumor suppressor, TSLC1. The TSLL2 protein of 55 kDa is mainly expressed in the kidney, bladder, and prostate in addition to the brain. Here, we report the biological significance of TSLL2 in the urinary tissues. An immunohistochemical study reveals that TSLL2 is expressed at the cell–cell attachment sites in the renal tubules, the transitional epithelia of the bladder, and the glandular epithelia of the prostate. Confocal microscopy analysis demonstrates that TSLL2 is localized in the lateral membranes in polarized Mardin-Darby canine kidney (MDCK) cells. TSLL2 forms homo-dimers and its overexpression induces aggregation of suspended MDCK cells in a Ca2+/Mg2+-independent manner, suggesting that it is involved in cell adhesion through homophilic trans-interaction. The TSLL2 gene is mapped on the chromosomal region 19q13.2, whose loss of heterozygosity has been frequently reported in prostate cancer. TSLL2 protein is lost in nine of nine primary prostate cancers and in a prostate cancer cell, PPC-1. Introduction of TSLL2 into PPC-1 strongly suppresses subcutaneous tumor formation in nude mice. These results suggest that TSLL2 is a new member of the Ig superfamily cell adhesion molecules and is a tumor-suppressor candidate in prostate cancer.


Immunoglobulin superfamily cell adhesion molecules (IgCAMs) are involved in not only cell–cell or cell–matrix interaction but also in many aspects of normal cell behavior, such as cell motility, proliferation, and differentiation (Hynes, 1999; Benson et al., 2000; Takai et al., 2003). On the other hand, alterations of these functions cause disruption of cell adhesion and could lead cancer cells to invasion or metastasis in some instances. TSLC1 is an IgCAM originally identified as a lung tumor suppressor and is involved in cell adhesion physiologically and tumor progression when inactivated (Kuramochi et al., 2001; Fukami et al., 2003a). TSLC1 forms homo-dimers, localizes along the lateral membrane of polarized epithelial cells, and harbors cell aggregation activity in a Ca2+/Mg2+-independent manner (Masuda et al., 2002). Restoration of TSLC1 expression in TSLC1-lacking cancer cells strongly suppresses subcutaneous tumor formation in nude mice (Kuramochi et al., 2001; Fukuhara et al., 2002; Ito et al., 2003). These findings suggest that the disruption of TSLC1 function would participate in the invasion or metastasis of human tumors. Recently, TSLC1 was found to play specific roles in cell adhesion of various tissues in addition to epithelium and has been termed TSLC1/IGSF4/RA175/Necl2/SgIGSF/SynCAM1 (Gomyo et al., 1999; Kuramochi et al., 2001; Wakayama et al., 2001; Biederer et al., 2002; Fujita et al., 2003; Shingai et al., 2003; Murakami, 2005).

We have previously isolated two additional immunoglobulin (Ig) superfamily molecules, TSLL1/IGSF4B/Necl1 and TSLL2/IGSF4C/Necl4, homologous to TSLC1 (Fukuhara et al., 2002). TSLC1, TSLL1, and TSLL2 form a unique subfamily (the TSLC1-gene family) that contains three Ig loops in the extracellular domain, a single transmembrane domain, and a short cytoplasmic domain (Figure 1a) (Murakami, 2002). In humans, both TSLL1 and TSLL2 display 37% identity in the amino-acid sequences with TSLC1 and are highly conserved during the evolution of vertebrates (Fukuhara et al., 2002; Fukami et al., 2003b). Interestingly, these molecules display distinct patterns of expression. Whereas TSLC1 expression was observed in the brain and almost all epithelial tissues and TSLL1 expression was restricted exclusively in the brain and spinal cord by Northern blotting, TSLL2 is expressed in the kidney, bladder, and prostate in addition to the brain. These findings suggest that TSLL2 is implicated in cell adhesion and, possibly, in tumor progression in urinary tissues, including prostate, as TSLC1 is.

Figure 1

Expression of TSLL2 protein in various organs. (a) Schematic representation of the TSLL2 protein. Ig1, Ig2, and Ig3 indicate immunoglobulin-like domains, while branches indicate possible N-glycosylation sites. S–S indicates a disulfide bond, while EX, TM, and CT indicate the extracellular, transmembrane, and cytoplasmic domains, respectively. (b) Western blotting of TSLL2 in mouse tissues or cell lines. Lysates from tissues or cells were treated with or without N-glycosidase F, subjected to SDS–PAGE (1 μg per lane) and detected by anti-TSLL2 pAb (BC2). Lane 1, heart; 2, brain; 3, spleen; 4, lung; 5, liver; 6, skeletal muscle; 7, kidney; 8, spinal cord; 9, prostate; 10, testis; 11, bladder; 12, MDCK; 13, MDCK transfected with TSLL2; 14, Caco-2; 15, brain; 16, spleen; 17, kidney; 18, spinal cord; 19, prostate; 20, testis; and 21, bladder. (c) Immunohistochemical analysis of TSLL2 and TSLC1 in human prostate, kidney, and bladder detected by BC2 and CC2. All sections were counterstained with hematoxylin. The original magnification is 400 × . Arrows and arrowheads in the kidney indicate proximal and distal convoluted tubules, respectively.

Prostate cancer is one of the most common malignancies among men in western countries and the incidence is increasing in Japan (Parker et al., 1997). Previous studies have demonstrated that alterations of multiple tumor-suppressor genes and/or metastasis-suppressor genes are involved in human prostate cancers, including the RB, TP53, E-cadherin, CDKN4, Kai-1, PTEN1, and TSLC1/IGSF4 genes (Ozen and Pathak, 2000; Fukuhara et al., 2002). Moreover, a recent study has reported that loss of heterozygosity on chromosomal region 19q is a frequent event in prostate cancer (Neville et al., 2003). The TSLL2/IGSF4C gene is mapped on 19q13.2 and its mRNA is downregulated in a subset of prostate cancer cells (Fukuhara et al., 2001).

In the present study, we examine the biological significance of TSLL2 in urinary tissues and cell lines using a specific antibody against TSLL2 and propose that TSLL2 protein is involved in cell–cell adhesion and implicated in tumor suppression in prostate cancer.


TSLL2 expression in the urinary tissues

We have previously reported by Northern blot analysis that TSLL2 is expressed in several specific tissues including the brain, kidney and prostate (Fukuhara et al., 2001). In the present study, we generated a polyclonal antibody (pAb) against TSLL2 (BC2) and examined the expression of TSLL2 protein in mouse tissue by Western blot analysis. As shown in Figure 1b, TSLL2 pAb detected a single protein with a molecular weight of 55 kDa in the spleen and urinary tissues including the kidney, prostate, and bladder and of 52 kDa in the brain and spinal cord. In the testis, three additional signals of about 50, 48, and 30 kDa were detected (Figure 1b, lane 10). Although no splicing variants were detected in the TSLL2 transcripts (data not shown), TSLL2 harbors four possible N-glycosylation sites in its extracellular domain, suggesting that the difference in the molecular weight is due to its post-translational modification. We, therefore, carried out the enzymatic deglycosylation of the TSLL2 protein from seven tissues expressing TSLL2 by treating with N-glycosidase F to release putative N-linked oligosaccharides. As shown in the right panel in Figure 1b, a single signal of approximately 45 kDa was observed in all samples, including the testis, by Western blotting after N-glycosidase F treatment, indicating that the molecular weight of TSLL2 protein in a deglycosylated form is about 45 kDa. These results strongly suggest that the signals of 55, 52, 50, and 48 kDa detected in the left panel are specific to the TSLL2 protein and generated by distinct post-translational modification. On the other hand, the signal of about 30 kDa observed in the testis in the left panel might be derived from a cleaved TSLL2 protein or due to a nonspecific reaction.

We focused on the urinary tissues and examined the expression of TSLL2 as well as TSLC1 by immunohistochemistry using specific antibodies against TSLL2 and TSLC1. As shown in Figure 1c, both TSLL2 and TSLC1 were expressed in the renal tubules, the transitional epithelia of the bladder and the glandular epithelia of the prostate at the cell–cell attachment sites (Figure 1c). In contrast, none of these proteins was expressed in the connective tissues. Interestingly, TSLL2 and TSLC1 showed distinct distribution in the kidney. TSLL2 was expressed in the proximal convoluted tubules and Henle's loops, while TSLC1 was mainly expressed in the distal convoluted tubules. On the other hand, neither TSLL2 nor TSLC1 was detected in the collecting ducts or Bowman's capsules (data not shown).

Lateral localization of TSLL2 in polarized epithelial cells

We next examined the subcellular distribution of TSLL2 in polarized epithelial cells in culture. For this purpose, Mardin–Darby canine kidney (MDCK) cells lacking endogenous TSLL2 were transfected with a plasmid expressing a full length of TSLL2 tagged with GFP (TSLL2-GFP) and stable transfectants were subsequently obtained. These cells were grown on collagen-coated filters to confluence, incubated with an anti-ZO-1 pAb, and analyzed by confocal laser scanning microscopy. TSLL2-GFP was directly detected by green fluorescence. In the X–Y section of the cells, TSLL2-GFP, as well as ZO-1, a marker protein of the tight junction (Stevenson et al., 1986), was distributed to a honeycomb-like structure at the cell–cell boundaries (Figure 2a). In the X–Z vertical cross-section, TSLL2-GFP was observed diffusely along the lateral membrane of the cells, showing distinct localization from that of ZO-1, which was detected at the tight junction as intense dots. In addition, subcellular localization of the endogenous TSLL2 was examined by anti-TSLL2 antibody, BC2, using human colon cancer cells, Caco-2, expressing a high amount of TSLL2 (Figure 2b). An endogenous TSLL2 protein was also detected predominantly at the lateral membrane of the cells. Therefore, we speculated that TSLL2 could be involved in the adhesion of epithelial cells, as was TSLC1.

Figure 2

Lateral localization of TSLL2 in the polarized epithelial cells. The representative sets of X–Y and X–Z sections are demonstrated. (a) MDCK-TSLL2-GFP cells on a collagen filter were immunostained with anti-ZO-1 antibody (center, red) and examined by laser scanning confocal microscopy. Green signals in the left panel are derived from the green fluorescence from TSLL2-GFP. (b) Endogenous expression of TSLL2 in Caco-2 cells. Cells on a collagen filter were doubly immunostained with anti-TSLL2-antibody, BC2 (left, green) and anti-ZO-1 antibody (center, red). Merged images of the left two columns were shown in the right columns (a and b). Scale bars indicate 10 μm.

Homophilic interaction of TSLL2 molecule in vitro

Previous studies have demonstrated that the majority of IgCAMs generate homophilic and/or heterophilic interaction. To examine possible homophilic interaction of TSLL2 molecules, human embryonic kidney (HEK) 293 cells were transfected with the expression vectors of TSLL2 tagged with Flag (TSLL2-Flag) and/or TSLL2 tagged with HA (TSLL2-HA). When the lysates from cells expressing both TSLL2-Flag and TSLL2-HA were immunoprecipitated with an anti-Flag antibody and blotted with an anti-HA antibody, a signal of approximately 55 kDa corresponding to TSLL2-HA was detected (Figure 3a). An additional signal of approximately 50 kDa was that of IgG derived from a cross-reaction of the two antibodies. Inversely, when the same lysates were immunoprecipitated with an anti-HA antibody and blotted with an anti-Flag antibody, a signal corresponding to TSLL2-Flag was also detected (Figure 3a). These results suggest that TSLL2 molecules form homophilic interaction.

Figure 3

Homophilic interaction of TSLL2 in vitro. (a) Coimmunoprecipitation of TSLL2-Flag and TSLL2-HA. HEK293 cells were transfected with either pTSLL2-Flag and pTSLL2-HA or both. Cell lysates were then immunoprecipitated with anti-Flag antibody or anti-HA antibody and subjected to Western blotting using anti-Flag antibody or anti-HA antibody. (b) Chemical cross-linking of TSLL2. A single-cell suspension of HEK293 cells was treated with or without a chemical cross-linker, BS3. Total cell lysates (1 μg) were subjected to Western blot analysis using BC2. A thin and a bold arrow indicate monomeric and possible dimeric forms, respectively.

We further investigated whether TSLL2 forms cis-homophilic dimers and/or multimers on the cell surface as TSLC1 does. A single cell suspension of HEK293 cells expressing endogenous TSLL2 was incubated in the absence or presence of the cell surface cross-linker, BS3. Total cell lysates were then subjected to Western blotting using anti-TSLL2 pAb, BC2. In the absence of BS3, a single band of 55 kDa was detected, corresponding to a monomeric form of TSLL2 (a thin arrow). On the other hand, a signal of approximately 110 kDa (a bold arrow), which appears to be a dimeric form of TSLL2, was detected in cell lysates treated with BS3 in addition to the signal of monomeric TSLL2 (Figure 3b). These findings suggest that at least a portion of TSLL2 proteins forms cis-homophilic dimers.

Cell aggregation activity of TSLL2 in a Ca2+/Mg2+ -independent manner

We next tested cell–cell adhesion activity of TSLL2 by a cell aggregation assay. We obtained two independent clones of MDCK expressing relatively high and low amount of TSLL2 protein (MDCK-TSLL2 #2 and MDCK-TSLL2 #5, respectively) and these two cells as well as parental MDCK were dissociated and incubated in HBSS without Ca2+ and Mg2+. Both MDCK-TSLL2 #2 and #5 cells formed large aggregates after 60 min of incubation, whereas parental MDCK formed few aggregates at the same time point (Figure 4b). When an aggregate of cells was counted as a single particle, the number of particles decreased significantly in the population of MDCK-TSLL2 #2 and #5 cells (Figure 4c). Cell aggregation was more prominently observed in MDCK-TSLL2 #2 than MDCK-TSLL2 #5, suggesting that the expression level of TSLL2 could correlate with the cell adhesion activity, although the difference was not statistically significant. Cell aggregation activity was not changed in the presence or absence of Ca2+ and Mg2+ in HBSS (data not shown). In contrast, the number of particles was not essentially changed within 60 min in the population of parental MDCK cells (Figure 4b). These results suggest that TSLL2 mediates intercellular adhesion through homophilic trans-interaction in a Ca2+/Mg2+ -independent manner.

Figure 4

Cell aggregation activity of TSLL2. (a) Western blot analysis of TSLL2 in parental MDCK cells and two independent MDCK derivatives expressing TSLL2 (MDCK-TSLL2 #2 and #5). (b) Parental MDCK and MDCK-TSLL2 #2 cells in a phase-contrast microscopy at 60 min of incubation. (c) Aggregation activity of parental MDCK (circles), MDCK-TSLL2 #2 (squares), and MDCK-TSLL2 #5 (triangles). Cell particles were counted at 0, 20, 40, and 60 min after dissociation and resuspension in Ca2+/Mg2+- free HBSS. Nt and No are the total number of particles at incubation times t and 0, respectively. The rate of aggregation was represented by the index Nt/No (mean±s.d. of three independent experiments).

As shown in Figure 1c, TSLL2 and TSLC1 are coexpressed in the bladder and prostate epithelia. This finding led us to examine the possible heterophilic interaction of TSLL2 with TSLC1. An equal number of MDCK expressing TSLL2 (MDCK-TSLL2 #2) and MDCK expressing TSLC1 (MDCK-TSLC1) were mixed together to test the formation of mixed cell aggregates. Prior to the mixture, MDCK-TSLL2 #2 cells were labeled with a lysophilic fluorescent dye, DiI. As shown in Figure 5, most of the MDCK-TSLL2 #2 and MDCK-TSLC1 cells formed aggregates by themselves. When over 100 cell aggregates composed of four cells were examined for the expression of DiI, 40 and 38% of the aggregates were found to be composed of MDCK-TSLL2 #2 cells alone and MDCK-TSLC1 cells alone, respectively, whereas only 22% of them were found to be a mixture of MDCK-TSLL2 #2 and MDCK-TSLC1 cells (Figure 5c). These results suggest that the majority of TSLL2 and TSLC1 molecules participate in homophilic trans-interaction by themselves to mediate cell–cell adhesion, while only a small portion of TSLL2, if any, generates heterophilic trans-interaction with TSLC1.

Figure 5

Mixed cell aggregation of MDCK-TSLL2 and MDCK-TSLC1. (a) Mixture of Dil-labeled MDCK-TSLL2 cells and MDCK-TSLC1 cells in phase-contrast microscopy at 30 min of incubation with Ca2+/Mg2+-free HBSS. (b) Representative pictures of four-cell aggregates with various compositions of MDCK-TSLL2 (with fluorescence) and MDCK-TSLC1 (without fluorescence) cells. (c) Frequency of each type of four-cell aggregates represented with the mean±s.d. of three independent experiments.

Loss of TSLL2 protein in primary prostate cancer

Given that the TSLL2 protein is involved in cell adhesion of normal epithelial cells, its function might be abrogated in cancer cells. Thus, we focused on the prostate cancer and examined TSLL2 expression in nine primary prostate cancers as well as four normal prostate tissues by immunohistochemistry using an anti-TSLL2 antibody, BC2. As expected, TSLL2 protein was expressed along the cell membrane in all four normal prostate epithelia (Figure 6a). In contrast, six of nine primary prostate cancers showed absence of TSLL2 protein expression (Figure 6b), while the other three tumors showed marked reduction of TSLL2 protein with an aberrant pattern of expression, in which weak signals of TSLL2 protein were detected diffusely in the cytoplasm but not at the cell membrane (Figure 6c). Totally, nine of nine primary prostate cancers showed inactivation of TSLL2 as summarized in Supplementary Table 1, suggesting that TSLL2 could be a possible tumor suppressor and its dysfunction participates in the malignant feature of prostate cancer.

Figure 6

Inactivation of TSLL2 in primary tumors and cell lines from prostate cancer. (ac) Immunohistochemical analysis of TSLL2 protein in human normal prostate (a) and primary prostate cancers (b and c). (a) Expression of TSLL2 is detected along the cell membrane in normal glandular epithelia of the prostate. (b) TSLL2 expression is absent in a prostate adenocarcinoma with grade III. (c) Reduced expression of TSLL2 in the cytoplasm in a prostate adenocarcinoma with grade II. (d) Western blot analysis of TSLL2 protein in normal prostate and human prostate cancer cell lines. (e) RT-PCR analysis of TSLL2 mRNA in normal prostate and PPC-1. (f) Allelic status of the polymorphic loci, D19S570, D19S422, D19S421, D19S417, and D19S223, around the TSLL2 locus on the chromosomal region 19q13.2 in the genomic DNA from PPC-1 and SW1783 cells.

Suppression of tumorigenicity of a prostate cancer cell line, PPC-1, by TSLL2

Next, we screened the expression of TSLL2 in prostate cancer cells by Western blotting and found that the TSLL2 protein was absent in PPC-1 cells and greatly reduced in Du145 cells (Figure 6d). The increased size of TSLL2 in LNCaP cells might be due to aberrant post-translational modification. RT-PCR revealed that TSLL2 mRNA was completely lost in PPC-1 cells (Figure 6e). PPC-1 cells were shown to have strong tumorigenicity when injected into the flank of nude mice (Brothman et al., 1991). We, therefore, examined whether TSLL2 could have a tumor-suppressor activity in PPC-1 cells. As the TSLL2 gene was mapped on 19q13.2, we studied the allelic status of five polymorphic microsatellite markers, D19S570, D19S422, D19S421, D19S417, and D19S223, around the TSLL2 in PPC-1 cells and found that all five markers showed hemizygosity or homozygosity (Figure 6f). Since these markers are highly polymorphic, as observed in DNA from another cell line, SW1783, it is strongly suggested that PPC-1 has lost one of the chromosomal regions, 19q13.2, including the TSLL2 locus, although no information is available as to the heterozygosity in normal DNA corresponding to PPC-1.

To examine the possible tumor-suppressor activity of TSLL2 in nude mice, we transfected a plasmid containing a full-length TSLL2 into PPC-1 (PPC-1-TSLL2) and eight independent cell clones were subsequently obtained. All these cell clones expressed a significant amount of TSLL2 protein of 55 kDa as shown in Figure 7b. These clones, however, showed no significant difference in morphology and growth rate in vitro when compared with parental PPC-1 cells or PPC-1 transfected with control plasmid, pcDNA3.1 (P3.1) (Figure 7a and c). PPC-1-TSLL2 cells, as well as P3.1 and the parental PPC-1 cells, were then injected subcutaneously into BALB/cA-nu (nu/nu) mice. As shown in Figure 7d, both the parental PPC-1 and P3.1 cells formed tumors at eight of eight sites of injection. The tumors continued to grow aggressively until the mice were killed on Day 35. In contrast, only 23 of 63 injection sites (37%) of PPC-1-TSLL2 cells developed notable tumors by Day 35. In addition, the average volume of tumors developed by PPC-1-TSLL2 cells was less than 40% of those by PPC-1 or P3.1 cells. These results indicate that TSLL2 significantly suppresses tumor formation of PPC-1 in nude mice.

Figure 7

Suppression of tumorigenicity of PPC-1 in nude mice by TSLL2. (a) Morphology of parental PPC-1, PPC-1 transfected with control plasmid, pcDNA3.1 (P3.1), and a PPC-1 cell clone transfected with TSLL2 (PPC-1-TSLL2) under phase-contrast microscopy. (b) Western blot analysis of TSLL2 protein in PPC-1, P3.1, and a PPC-1-TSLL2 clone. (c) In vitro cell growth assay of PPC-1, P3.1, and PPC-1-TSLL2 clones. Cell number was measured at Days 1, 3, 5, and 7. •, PPC-1; ▪, P3.1; □, PPC-1-TSLL2. (d) Tumor formation in nude mice. The average volume of tumors that formed at eight sites was determined at the indicated times after injection of 105 cells from the following PPC-1 derivatives: •,PPC-1; ▪, P3.1; □, PPC-1-TSLL2.


We have demonstrated in the present study that TSLL2 is a new member of IgCAM in the urinary tissues based on the following observations: (1) the TSLL2 protein is mainly expressed in the urinary tissues in addition to the brain and localized at the cell–cell attachment sites of the renal proximal tubules, the transitional epithelia of the bladder and the glandular epithelia of the prostate; (2) TSLL2 is predominantly expressed along the lateral membrane in polarized epithelial cells, MDCK; (3) TSLL2 forms homophilic dimers when expressed in HEK293 cells; (4) TSLL2 harbors cell aggregation activity in a Ca2+/Mg2+-independent manner. Although other members of IgCAMs are frequently cleaved at their extracellular domains and generate multiple bands, most tissues except for the testis expressed only a single molecule of TSLL2, suggesting that the proteolytic cleavage is not common in TSLL2. Further studies using antibody against the N-terminal portion of TSLL2 would be required. On the other hand, the molecular weights of TSLL2 varied in several tissues. This variation appears to be due to distinct post-translational modification, because N-glycosidase F treatment generated a single TSLL2 signal of about 45 kDa in all tissues by Western blotting.

These features of TSLL2 are similar to those of TSLC1 as a cell adhesion molecule. However, TSLL2 appears to function predominantly in the urinary tissues in contrast to TSLC1. It is extremely interesting that TSLL2 and TSLC1 are differentially expressed in the proximal and distal convoluted tubules, respectively (Figure 1c). TSLL2 and TSLC1 could be implicated in the specific cell–cell interactions or in some differentiated functions in each cell type. Immunoprecipitaion coupled with Western blotting revealed that the TSLL2 protein formed homophilic interaction. Moreover, the short-term suspension aggregation assay showed that the homophilic trans-interaction of TSLL2 causes cell aggregation. MDCK cells we used in the present study are renal epithelial cells and not the fibroblastic cells widely used in the aggregation assay (Yoshida and Takeichi, 1982). The advantage of using epithelial cells in this assay is that they are expected to express groups of downstream proteins related to the cell adhesion molecule, TSLL2. In fact, aggregation activity of TSLL2 was well demonstrated using MDCK. On the other hand, the disadvantage would be that epithelial cells, like MDCK, might express other cell adhesion molecules, which could modify the function of TSLL2. We confirmed, however, that TSLC1 was not expressed in MDCK cells. These results strongly suggest that TSLL2 mediates cell adhesion through homophilic trans-interaction.

In contrast to the renal convoluted tubules, the transitional epithelium of the bladder and the glandular epithelium of the prostate express both TSLL2 and TSLC1 proteins in the same cells. It is possible, therefore, that TSLL2 could interact with TSLC1 on the membranes of these cells. However, mixed cell aggregation assay revealed that the homophilic trans-interaction of TSLL2 and that of TSLC1 were much stronger than the heterophilic trans-interaction between TSLL2 and TSLC1. Therefore, even in the cells expressing both TSLL2 and TSLC1, the homophilic trans-interaction is mainly involved in the cell adhesion through TSLL2.

In the last part of this study, we showed that TSLL2 could act as a tumor-suppressor candidate in human prostate cancer. This proposal is based on the following observations. (1) The TSLL2 gene is mapped on the chromosomal region 19q13.2 and is included in the candidate region harboring a possible tumor suppressor or susceptibility gene of prostate cancer identified by several previous studies. Chromosome transfer experiments suggested that a tumor-suppressor gene was present between 19q13.1–q13.2 (Gao et al., 1999; Astbury et al., 2001). Whole-genome linkage analyses in combination with allelic imbalance studies also demonstrated that a chromosomal fragment, 19q or 19q12–q13, could be linked to prostate cancer aggressiveness (Neville et al., 2003; Slager et al., 2003). (2) TSLL2 protein was lost or markedly reduced in nine of nine primary prostate cancers and two of four prostate cancer cell lines, PPC-1 and Du145, in comparison with that in normal human prostate. (3) The tumorigenicity of PPC-1 was strongly suppressed by the restoration of TSLL2 expression. In this study, all eight independent PPC-1 derivatives expressing TSLL2 showed significant suppression of tumorigenicity, while parental PPC-1, PPC-1 carrying a vector alone (Figure 7d), or 10 additional independent subclones of parental PPC-1 (data not shown) showed 100% tumorigenicity. Thus, the clonal variation in terms of tumorigenicity could be excluded.

It is noteworthy that the restoration of TSLL2 expression into PPC-1 cells did not induce significant cell death or growth inhibition in vitro, suggesting that the inactivation of TSLL2 would be involved not in the direct cell growth but in the aberrant cell–cell contact in the prostate carcinogenesis, as seen in TSLC1 (Fukuhara et al., 2002). In this connection, our preliminary experiments showed that reduction of TSLL2 expression by RNAi caused flat morphology with immature actin-bundle formation in a mouse breast cancer cell, BALB-MC, which is overexpressing TSLL2, suggesting that TSLL2 could be implicated in mature cell–cell adhesion. Involvement of TSLL2 in cell adhesion of normal epithelia was also supported by the observation that TSLL2 protein was lost or greatly reduced in all nine primary prostate cancers with grade II to grade IV. Inactivation of TSLL2 might be a relatively late event in the prostate carcinogenesis. Further studies would be required to elucidate the clinico-pathological significance of TSLL2 in prostate cancer as a potential tumor-suppressor gene.

Materials and methods

Cell lines and animals

A human prostate cancer cell line, PPC-1, was kindly provided by Dr AR Brothman at the University of Utah and maintained in RPMI1640 with 10% fetal bovine serum. Human prostate cancer cells, PC-3, LNCaP, and Du145, an astrocytoma cell, SW1783, and a colorectal cancer cell, Caco-2, were purchased from the American Type Culture Collection (Rockville, MD, USA). A canine kidney cell, MDCK, was obtained from the Human Science Research Resources Bank (Osaka, Japan). BALB/cA Jcl and BALB/cA-nu (nu/nu) mice were from Japan Crea (Tokyo, Japan). All animal experiments were performed according to the National Cancer Center Guidelines for Animal Experiments.


A rabbit anti-TSLL2 pAb (BC2) was raised against the synthetic peptides of 13 amino acids at the carboxyl terminus of TSLL2 coupled with keyhole limpet hemocyanin and purified with an affinity column. These 13 amino acids are completely identical between human and mouse (Fukami et al., 2003b). A rabbit anti-TSLC1 pAb (CC2) was previously described (Masuda et al., 2002). A mouse anti-ZO-1 mAb and secondary antibodies were purchased from Zymed (San Francisco, CA, USA) and Amersham Biosciences (Buckinghamshire, UK), respectively.

Plasmids and transfection

A whole coding sequence of TSLL2 was cloned into the expression vector pcDNA3.1 (Invitrogen, Carlsbad, CA, USA) to obtain pcTSLL2 and cloned into pEGFP-N3 (Clontech, Palo Alto, CA, USA) to obtain GFP-tagged TSLL2 (TSLL2-GFP). Expression vectors for TSLL2-Flag and TSLL2-HA were obtained by introducing the tagged sequence of Flag (5′-IndexTermGACTACAAGGATGACGATGACAAG-3′) and that of HA (5′-IndexTermTACCCATACGACGTCCCAGACTACGCT-3′) into the 3′ end of TSLL2 cDNA in the pcTSLL2, respectively. A plasmid, pcTSLC1, containing a whole coding sequence of TSLC1, was previously described (Kuramochi et al., 2001). Transfection was carried out using LipofectAMINE plus (Invitrogen) and stable cell clones with plasmids carrying the neomycin resistance gene were selected using 300 μg/ml of G418.


Sequential paraffin sections of the normal human kidney, bladder, and prostate were purchased from Dako (Carpinteria, CA, USA), while a human tissue array of prostate cancer was obtained from Cybrdi (Gaithersburg, MD, USA). The sections were incubated with the indicated primary antibodies and visualized by Envision kit/HRP (DAB) (Dako). All sections were counterstained with hematoxylin.

Confocal microscopy

For the polarized cell culture, cells were seeded at a density of 1 × 105 cells perφ24 mm trans-well-COL collagen-coated filter insert (Costar, Cambridge, MA, USA) and allowed to grow to confluence for 6 days. The cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton, and stained with indicated antibodies. Confocal microscopy analysis was performed as described previously (Masuda et al., 2002) using a Bio-Rad Radiance 2000 laser confocal scanning system equipped with a 488/524-nm argon and a 543-nm helium-neon laser.

Immunoprecipitation and Western blotting

Cell lysates were prepared from tissues of BALB/cA Jcl mice or cultured cells as described previously (Masuda et al., 2002). For direct Western blotting, an aliquot (1 μg) was applied in each lane in a 4–12% gradient SDS–PAGE (Invitrogen). For immunoprecipitation, cell lysates were incubated with an appropriate primary antibody overnight at 4°C; then, protein A-Sepharose (Amersham Biosciences) was added. Immunoprecipitates were rinsed with the lysis buffer, suspended in a sample buffer (Invitrogen) containing 50 mM DTT, and incubated for 10 min at 70°C. The samples were fractionated in 4–12% gradient SDS–PAGE, transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA, USA), and reacted with an appropriate primary antibody. The binding of the primary antibody was detected with Lumi-lightPLUS Western blotting Substrate (Roche Diagnostics, Basel, Switzerland) using a peroxidase-conjugated secondary antibody.


Digestion of N-linked glycosylation was carried out using PNGase F (New England Biolabs, Breverly, MA, USA) according to the manufacturer's instructions. Briefly, 10 μg of protein from cell lysates in a 90 μl lysis buffer was denatured with 10 μl of a glycoprotein denaturing buffer (5% SDS, 10% beta-mercaptoethanol) at 100°C for 10 min and then incubated with 10 μl of a G7 Buffer (0.5 M sodium phosphate, pH 7.5 and 10 μl of 10% NP-40 containing 1500 U of PNGase F) at 37°C for 1 h.

Cross-linking with BS3

Cells were trypsinized and resuspended in PBS at 5 × 105 cells/ml, transferred to 15-ml polypropylene tubes (Corning Inc., Corning, NY, USA), and incubated in the presence or absence of 3 mM bis(sulfosuccinimidyl) suberate (BS3, Pierce, Rockford, IL, USA) at room temperature with constant agitation for 20 min. The reaction was quenched with the addition of 20 mM Tris-HCl, pH7.5. To assure a single-cell suspension, the cells were subjected to phase-contrast microscopy before and after the chemical cross-linking reaction. The cells were then spun down and added to a lysis buffer containing 10 mM iodoacetamide (Sigma Chemical, St Louis, MO USA) to obtain protein samples.

Cell aggregation assay

A cell aggregation assay was performed as described previously (Masuda et al., 2002). Briefly, dissociated cells were suspended in normal HBSS containing Ca2+/Mg2+ or Ca2+/Mg2+-free HBSS, reseeded in 12-well plates (5 × 105 cells/well) precoated with bovine serum albumin fraction V (Sigma Chemical), and rotated on a gyrator shaker at 37°C for 20, 40, or 60 min. Aggregation was stopped with the addition of 2% glutaraldehyde. The extent of cell aggregation was represented by the ratio of the total particle number at time t of incubation (Nt) to the initial particle number (No). The mixed cell aggregation assay was performed as described previously (Satoh-Horikawa et al., 2000). MDCK-TSLL2 were labeled with a lysophilic fluorescent dye, DiI (Molecular Probes, Eugene, OR, USA), mixed with MDCK-TSLC1 at a ratio of 1:1, and incubated in Ca2+- and Mg2+-free HBSS for 30 min in 12-well plates (5 × 105 cells/well). Cell aggregation was examined by a phase-contrast microscopy, and MDCK-TSLL2 cells were identified for their fluorescence. Over 100 four-cell aggregates were examined for their composition of MDCK-TSLL2 and MDCK-TSLC1 cells.

Reverse transcription-polymerase chain reaction (RT-PCR)

Total cellular RNA was extracted with RNeasy Mini Kit (Qiagen). Of total cellular RNA, 1 μg was reverse-transcribed using the SuperScript II reverse transcriptase (Invitrogen) in the presence of RNase inhibitors, and an aliquot was subjected to PCR using Ex Taq DNA polymerase (Takara Bio, Kyoto, Japan) to obtain a 451-bp fragment of TSLL2cDNA. The primers used for PCR were 5′-IndexTermCAAGGAGCTGAAAGGAGTGAG-3′ (forward) and 5′-IndexTermACCAGGGTCGTAGACCACAA-3′ (reverse).

Analysis of the allelic status

Genomic DNA was extracted by the Proteinase K-phenol-chloroform extraction methods. DNA fragments containing five microsatellite STS markers on 19q13.2, namely D19S570, D19S422, D19S421, D19S417, and D19S223, were amplified by PCR using pairs of primers, one of which was end labeled with Texas Red. Amplified fragments were subjected to electrophoresis in polyacrylamide gels containing 7 M urea for 120 min at 45°C using SF5200 (Hitachi Electronics Engineering, Japan).

Cell growth assay in vitro

In vitro cell growth was examined using CellTiter 96 Nonradioactive cell proliferation assay (Promega, Madison, WI, USA) according to the manufacture's instruction. Briefly, 5 × 103 cells were cultured in 96-well plates in triplicate for 1, 3, 5, and 7 days. Samples were incubated with Dye solution containing the tetrazolium salt for 4 h, and the absorbance of the formazan product was measured at 595 nm using a 96-well plate reader.

Tumorigenicity analysis

A suspension of 1 × 105 cells in PBS (0.2 ml) was subcutaneously injected into 1–4 sites on the flanks of 5–6-week-old female BALB/cA-nu (nu/nu) mice. Tumor growth was assessed by measuring the xenografts in three dimensions once a week.


  1. Astbury C, Jackson-Cook CK, Culp SH, Paisley TE, Ware JL . (2001). Genes Chromosomes Cancer 31: 143–155.

  2. Benson DL, Schnapp LM, Shapiro L, Huntley GW . (2000). Trends Cell Biol 10: 473–482.

  3. Biederer T, Sara Y, Mozhayeva M, Atasoy D, Liu X, Kavalali ET et al. (2002). Science 297: 1525–1531.

  4. Brothman AR, Wilkins PC, Sales EW, Somers KD . (1991). J Urol 145: 1088–1091.

  5. Fujita E, Soyama A, Momoi T . (2003). Exp Cell Res 287: 57–66.

  6. Fukami T, Fukuhara H, Kuramochi M, Maruyama T, Isogai K, Sakamoto M et al. (2003a). Int J Cancer 107: 53–59.

  7. Fukami T, Satoh H, Williams YN, Masuda M, Fukuhara H, Maruyama T et al. (2003b). Gene 323: 11–18.

  8. Fukuhara H, Kuramochi M, Fukami T, Kasahara K, Furuhata M, Nobukuni T et al. (2002). Jpn J Cancer Res 93: 605–609.

  9. Fukuhara H, Kuramochi M, Nobukuni T, Fukami T, Saino M, Maruyama T et al. (2001). Oncogene 20: 5401–5407.

  10. Gao AC, Lou W, Ichikawa T, Denmeade SR, Barrett JC, Isaacs JT . (1999). Prostate 38: 46–54.

  11. Gomyo H, Arai Y, Tanigami A, Murakami Y, Hattori M, Hosoda F et al. (1999). Genomics 62: 139–146.

  12. Hynes RO . (1999). Trends Cell Biol 9: M33–7.

  13. Ito T, Shimada Y, Hashimoto Y, Kaganoi J, Kan T, Watanabe G et al. (2003). Cancer Res 63: 6320–6326.

  14. Kuramochi M, Fukuhara H, Nobukuni T, Kanbe T, Maruyama T, Ghosh HP et al. (2001). Nat Genet 27: 427–430.

  15. Masuda M, Yageta M, Fukuhara H, Kuramochi M, Maruyama T, Nomoto A et al. (2002). J Biol Chem 277: 31014–31019.

  16. Murakami Y . (2002). Oncogene 21: 6936–6948.

  17. Murakami Y . (2005). Cancer Sci 96: 543–552.

  18. Neville PJ, Conti DV, Krumroy LM, Catalona WJ, Suarez BK, Witte JS et al. (2003). Genes Chromosomes Cancer 36: 332–339.

  19. Ozen M, Pathak S . (2000). Anticancer Res 20 (3B): 1905–1912.

  20. Parker SL, Tong T, Bolden S, Wingo PA . (1997). Cancer Statist 47: 5–27.

  21. Satoh-Horikawa K, Nakanishi H, Takahashi K, Miyahara M, Nishimura M, Tachibana K et al. (2000). J Biol Chem 275: 10291–10299.

  22. Shingai T, Ikeda W, Kakunaga S, Morimoto K, Takekuni K, Itoh S et al. (2003). J Biol Chem 278: 35421–35427.

  23. Slager SL, Schaid DJ, Cunningham JM, McDonnell SK, Marks AF, Peterson BJ et al. (2003). Am J Hum Genet 72: 759–762.

  24. Stevenson BR, Siliciano JD, Mooseker MS, Goodenough DA . (1986). J Cell Biol 103: 755–766.

  25. Takai Y, Irie K, Shimizu K, Sakisaka T, Ikeda W . (2003). Cancer Sci 94: 655–667.

  26. Wakayama T, Ohashi K, Mizuno K, Iseki S . (2001). Mol Reprod Dev 60: 158–164.

  27. Yoshida C, Takeichi M . (1982). Cell 28: 217–224.

Download references


We thank Dr Tesshi Yamada at NCC for his generous help with the Bio-Rad Radiance 2000 confocal scanning system and Drs Yoshiyuki Ishii and Satoshi Sekiguchi at the University of Tokyo for their generous support of our pathological studies. This work was supported in part by a Grant-in-Aid for the Third-Term Comprehensive Control Research for Cancer from the Ministry of Health, Labor, and Welfare of Japan; a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Culture, Sports, Science, and Technology, Japan; and a Grant from the Program for the Promotion of Fundamental Studies in Health Sciences of Pharmaceutical and Medical Devices Agency (PMDA) of Japan; MM is a recipient of a Research Fellowship from PMDA. MS-Y is a recipient of a Research Resident Fellowship from the Foundation for Promotion of Cancer Research (Japan).

Author information

Correspondence to Y Murakami.

Additional information

Supplementary Information accompanies the paper on Oncogene website (http://www.nature.com/onc).

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Williams, Y., Masuda, M., Sakurai-Yageta, M. et al. Cell adhesion and prostate tumor-suppressor activity of TSLL2/IGSF4C, an immunoglobulin superfamily molecule homologous to TSLC1/IGSF4. Oncogene 25, 1446–1453 (2006). https://doi.org/10.1038/sj.onc.1209192

Download citation


  • cell adhesion
  • tumor-suppressor gene
  • prostate cancer

Further reading