Introduction

Greater than 80% of the population are infected with the human polyomavirus, JC virus (JCV) during childhood, although in the majority of infected individuals the virus establishes latency in the kidney and does not induce any overt signs of disease (Berger and Concha, 1995). In immunocompromised individuals such as AIDS patients, transplant recipients, and individuals with lymphoproliferative disorders, however, reactivation of JCV results in the fatal demyelinating disease, progressive multifocal leukoencephalopathy (PML) (Berger and Concha, 1995). Over the last several years, studies have suggested a role for JCV in human cancer, as a broad range of CNS tumors have been found to harbor JCV DNA sequences and to express the viral protein, T-antigen, including medulloblastoma, glioblastoma, astrocytoma, oligodendroglioma, and other tumors of neural crest origin (for review, see Del Valle et al., 2001a). More recently, JCV has been detected in colorectal carcinoma and in vitro studies suggest that T-antigen may target the Wnt signaling pathway via -catenin in the settings of colon cancer and medulloblastoma (Gan et al., 2001; Enam et al., 2002).

Early attempts to develop animal models of the demyelinating disease, PML, were unsuccessful. However, inoculation of nonhuman primates, neonatal hamsters, and rats resulted in the development of a broad range of tumors, including astrocytoma, glioblastoma, medulloblastoma, and adrenal neuroblastomas (London et al., 1978; ZuRhein, 1983; Ohsumi et al., 1986). Several transgenic mouse lines have been generated using JCV T-antigen under the control of natural isolates of the JCV promoter that develop tumors of neural crest origin, including medulloblastoma and adrenal neuroblastoma, as well as pituitary tumors (Small et al., 1986a, 1986b; Franks et al., 1996; Krynska et al., 1999b; Gordon et al., 2000). In each of these animal models, p53 protein was detectable in tumor tissue and was shown to interact with JCV T-antigen, demonstrating further evidence that targeting of the p53 pathway is a major mechanism of JCV T-antigen-mediated tumor induction (Krynska et al., 2000). Inactivation of the p53 pathway is one of the most commonly affected alterations in human cancer. In addition to mutations or deletions, wild-type p53 may be inactivated through association with polyomavirus T-antigens. Similar to its well-characterized counterpart, SV40, JCV T-antigen has multifunctional domains that control the viral lytic cycle (Kim et al., 2001). In addition, several of these models have also shown upregulation of p21/WAF-1 likely due to p53-independent mechanisms (Gordon et al., 2000).

Neurofibromatosis types 1 (NF1) and 2 (NF2) are autosomal dominant disorders affecting one in 4000 and one in 40 000 live births, respectively (Louis et al., 2000; von Deimling et al., 2000). These complex disorders result in the development of a broad range of neoplastic and dysplastic lesions of neural crest origin (Louis et al., 2000; von Deimling et al., 2000). Approximately 50% of all patients with NF1 and NF2 and inherit the disorders; the remaining 50% acquire new germline mutations often at sites of prior radiation. The NF1 gene product, neurofibromin, encodes a large protein of approximately 250 kDa, which is localized to the cytoplasm and has homology with GAP-related domains, thus suggesting that it functions as a RasGTPase-activating protein. Loss of heterozygosity (LOH) and frequent mutations in the remaining NF1 allele have suggested that NF1 functions as a tumor suppressor gene. Likewise, the NF2 gene exhibits many germline and somatic mutations in NF2 patients, suggesting that it, too, may function as a tumor suppressor gene. NF2 is localized to the cytoplasm and thought to interact with actin to effect cytoskeletal organization (Sun et al., 2002).

Malignant tumors arising from peripheral nerves or having evidence of nerve sheath differentiation, the so-called malignant peripheral nerve sheath tumors (mpnst), are uncommon neoplasms that are most frequently associated with NF1, as approximately 50% of all cases arise in NF1 patients (Woodruff et al., 2000). The remaining individuals develop mpnst in the third to sixth decades of life, with approximately 10% arising at sites of prior irradiation. Approximately two-thirds of all mpnsts are thought to develop from neurofibromas, while the rest develop de novo from peripheral nerves or, rarely, from other benign lesions of the neural crest (Woodruff et al., 2000). The majority of mpnst demonstrate the expression of p53 and have high levels of cellular proliferation (as determined by Ki-67 immunolabeling), compared with benign schwannomas and neurofibromas.

Further information on the functions of NF1 and NF2 can be gleaned from mouse models with targeted disruption of their genes. For example, while mice lacking NF1 die in utero due to defects in cardiac development, some chimeric NF1-/- mice develop neurofibromas that mimick the phenotype observed in NF1 patients, suggesting that the complete inactivation of NF1 is sufficient for the induction of neurofibromas (Side et al., 1997). Furthermore, mice heterozygous at the NF1 locus (i.e. NF1+/-) develop some features associated with NF1, including myeloid leukemia, pheochromocytomas and cognitive deficits (McClatchey and Cichowski, 2001). Interestingly, mice heterozygous for both NF1 and p53 developed mpnst, although it should be noted that, unlike in humans, these two genes are closely linked on chromosome 11 in the mouse and were deleted from the same chromosome (Cichowski et al., 1999; Vogel et al., 1999). Furthermore, expression of both remaining alleles of NF2 and p53 were lost in the tumors; however, animals with deletion of NF1 and p53 on different chromosomes did not develop neoplasias, suggesting that positional effects may play a role in this animal model. As in the case with the NF1+/- model, mice heterozygous for NF2 exhibit several neoplastic tendencies including osteosarcoma, fibrosarcoma, and hepatocellular carcinoma that show evidence of metastasis, suggesting a role for NF2 in cell adhesion, migration, and metastasis (McClatchey et al., 1998). By generating mice heterozygous for NF1, NF2, and p53, acceleration and cooperativity of the neoplasms was observed (McClatchey and Cichowski, 2001).

Here, we report a novel transgenic mouse model of mpnst induced by the expression of JCV T-antigen in which T-antigen has been shown to interact physically with NF2, but not NF1. Tumor tissue also shows evidence of dysregulation of the p53 pathway and suggests the possible interaction between NF2 and p53 in the presence of T-antigen. This animal model may shed light on the function of NF2 as a tumor suppressor protein and additional mechanisms whereby JCV T-antigen induces cellular transformation.

Results

In order to expand our knowledge on the tissue tropism and transforming ability of JCV, we developed a transgenic mouse model that harbors sequences coding for the JCV early protein, T-antigen, under the control of the natural JCV Mad-4 promoter. The DNA fragment was microinjected into fertilized embryos and 10 founder animals were positive for the transgene by PCR of genomic DNA extracted from the tail. The founder mice exhibited a broad range of phenotypes, which consisted mainly of neoplasias (see Table 1). One founder exhibited behavioral characteristic of demyelination such as that previously described for JCV T-antigen transgenic mice (Small et al., 1986a). This animal also harbored an abdominal neuroblastoma, a phenotype frequently observed in mice expressing JCV T-antigen (Small et al., 1986b; Franks et al., 1996). A second founder succumbed to a neuroblastoma at the age of 4 months. Interestingly, one founder exhibited a primitive neuroectodermal tumor in the ventricular region of the brain, similar to the PNET/medulloblastoma phenotype described with JCV T-antigen under the control of the archetype promoter (Krynska et al., 1999b). Three of the founder mice developed pituitary adenomas and one of these (E30) was successfully mated resulting in an F1 generation carrying the transgene. The progeny of the F1 generation exhibited Mendelian inheritance of the transgene and a male to female ratio of 1 : 1, indicating germline transmission. As described previously, approximately 50% of the E30 line of mice develop pituitary adenomas by 1 year of age (Gordon et al., 2000).

Table 1 - Phenotypes observed in founder JCV T-antigen transgenic mice.

Full table

Interestingly, a subset of E30 progeny developed large masses along the extremities or the neck at ages of greater than 9 months. Occasionally, the masses were seen to arise in the vestibular region (Figure 1a) or adjacent to the salivary gland (panel c). Morphologically, the tumors were very pale, solid, and well circumscribed while, histologically, these highly cellular interdigitating spindle cell tumors showed frequent mitotic figures and a sheet-like appearance (panel e). Immunohistochemical evaluation of the tumors detected strong labeling for the neuronal marker, S-100 (panel f), but were negative for desmin, cytokeratin, epithelial membrane antigen, glial fibrillary acidic protein, and synaptophysin (data not shown). These observations are consistent with a diagnosis of mpnst (Figure 1, panel f), rare neoplasms that occur in individuals with neurofibromatosis.

Figure 1.

JCV T-antigen induces malignant peripheral nerve sheath tumors in T-antigen transgenic mice. (a) and (b) A solid, well-circumscribed pale mass in the vestibular region is observed. (c) At low magnification, a highly cellular neoplasm is seen adjacent to the salivary gland. (d) and (e) Higher magnification reveals a sheet-like appearance characteristic of malignant peripheral nerve sheath tumors (e) and regions which are mixoid (d). Frequent mitotic figures are observed. (f) Immunohistochemical analysis with antibody that recognizes the neuronal-specific protein, S100, confirms the origin of the tumor. Panels b–e, hematoxylin and eosin staining; panel f, hematoxylin counterstain. Original magnification: c: 100; d and f: 200; e: 400

Full figure and legend (577K)

The tumors were further characterized by immunohistochemistry to detect the transgene, JCV T-antigen. As mpnst occur in individuals with neurofibromatosis type 1 (NF1), immunohistochemistry was also performed with antibodies that recognize the gene products of neurofibromatosis 1 and 2. Owing to previous observations indicating dysregulation of the p53 pathway in JCV T-antigen-induced neoplasms, tumors were also evaluated for the presence of p53 and its downstream effector, p21/WAF-1 (Gordon et al., 2000). As shown in Figure 2, panels a, T-antigen was detected in the nucleus of some, but not all mpnst tumor cells. More specifically, approximately 25% of the tumor cells expressed nuclear T-antigen to varying degrees. Immunolabeling with anti-p53 antibody revealed a similar pattern in that some cells showed strong nuclear expression, others showed weak staining, while overall, the majority of the cells did not express either protein (Figure 2, panels a and b, respectively). Similarly, p21/WAF-1 protein was detected in the nucleus of some cells (panel b, inset). As anticipated, NF1 and NF2 showed a cytoplasmic distribution in most cells (panels c and d, respectively).

Figure 2.

Immunohistochemical detection of T-antigen; tumor suppressor proteins p53, NF1, NF2; and cellular regulatory factors in malignant peripheral nerve sheath tumors. Immunolabeling was performed on sections of paraffin-embedded tumor tissue with antibodies that recognize JCV T-antigen (panel a), p53 (panel b) the p53 downstream factor p21/WAF1 (panel b, inset), NF1 (panel c), NF2 (panel d). All panels, original magnification 400; hematoxylin counterstain

Full figure and legend (383K)

Owing to the fact that NF1 and NF2 act as tumor suppressor proteins, and the demonstrated ability of JCV T-antigen to inactivate tumor suppressor proteins such as p53 and pRb, immunoprecipitation followed by Western blot analysis was performed to determine whether T-antigen was able to associate with either NF1 or NF2. Toward this end, cellular extracts from mpnst tumor tissue or normal mouse brain were incubated with antibodies to T-antigen, NF1, or NF2, complexes were immunoprecipitated, and were subjected to Western blot analysis. As shown in Figure 3, T-antigen could be immunoprecipitated from extracts of the tumor tissue but could not be precipitated from the normal brain (compare lanes 1 and 4). Interestingly, antibody to NF2 was able to co-precipitate T-antigen from tumor tissue extracts, but not the normal brain, indicating that these proteins may form a complex in the tumor tissue (compare lanes 3 and 6). In contrast, while NF1 could be detected by immunoprecipitation of an extract from the normal brain as well as tumor tissue, T-antigen did not co-precipitate with NF1 (panel a, lanes 2 and 5; panel b, lanes 1, 2, 4, and 5).

Figure 3.

Association of JCV T-antigen with NF2, but not NF1, in T-antigen-induced malignant peripheral nerve sheath tumors. Immunoprecipitation was performed with extract prepared from the tumor and normal brain tissue using antibodies recognizing NF1, neurofibromatosis (NF2), and T-antigen followed by Western blot analysis with antibodies to T-antigen (panel a) and NF1 (panel b). The region of T-antigen that may interact with NF2 was mapped using GST-T-antigen deletion mutants shown in panel c. Extract prepared from the normal mouse brain tissue was incubated with eluted GST (lane 1), GST-T-antigen full length (lane 2), and GST-T-antigen deletion mutants (lanes 3–7). Samples were then immunoprecipitated with antibodies to NF2 and immunoblotted with antibodies against GST (panel d)

Full figure and legend (97K)

To further characterize the interaction of NF2 with JCV T-antigen, immunoprecipitation followed by Western blot analysis was performed in an extract from the normal mouse brain supplemented with bacterially expressed T-antigen mutants. More specifically, wild-type and mutant GST-T-antigen protein, as well as GST alone, were expressed and purified from bacteria, as shown in panel c. GST or GST-T-antigen mutants were added to the normal mouse brain extract along with anti-NF2 antibody and incubated overnight. The following day, NF2:T-antigen complexes were immunoprecipitated with Pansorbin and were analysed by Western blotting using anti-GST antibody. As shown in Figure 3 (panel d), anti-NF2 antibodies co-precipitated with wild-type GST-T-antigen and mutants 1–411, 1–265, and 412–688 (lanes 2, 3, 4, and 6), but not mutants representing the N- and C-termini of T-antigen (1–81 and 629–688, respectively) or GST alone. Thus, sequences of T-antigen between amino acids 81 and 265 or 412 and 629 are sufficient for binding to NF2. These two regions of T-antigen contain several important functional domains, including pRb binding, the T-antigen nuclear localization signal, and DNA binding (mt 81–265); ATPase (mt 412–688); and p53 binding and helicase activity (both mutants).

Analysis of the interaction between T-antigen and p53 was next performed by immunoprecipitation with anti-T-antigen or anti-p53 antibodies in extracts prepared from tumor, as well as the brain, heart, kidney, liver, lung, and spleen tissues from the same animal. Western blotting with anti-T-antigen antibodies revealed high levels of T-antigen in extract from the tumor tissue, but not in the normal brain or other organs of the transgenic mice (Figure 4, panel A, upper). Furthermore, anti-p53 antibody was able to co-precipitate T-antigen from the tumor extract, indicating the T-antigen and p53 form a complex in mpnst (panel a, lower). As both NF2 and p53 were able to co-precipitate with T-antigen, immunoprecipitation was performed with anti-T-antigen, anti-NF2, and anti-p53 antibodies followed by Western blotting for T-antigen and p53. As shown in panel b, reciprocal binding of T-antigen and p53 was observed using specific antibodies (lanes 1 and 3). Interestingly, anti-NF2 antibody was able to precipitate both T-antigen and p53, while the normal mouse and normal rabbit serum (NMS and NRS, respectively) did not precipitate either protein (compare lane 2 with lanes 4 and 5). Of note, the precipitated proteins were detected with equal intensity, suggesting that the three proteins may form a ternary complex.

Figure 4.

T-antigen expressed in tumor tissue is associated with p53 and NF2 of transgenic mice. Extract prepared from tumor and normal tissues of a T-antigen transgenic mouse were immunoprecipitated with antibody to T-antigen (panel a, upper) or p53 (panel a, lower) followed by Western blot analysis to detect T-antigen. Arrows indicate the position of T-antigen. Extracts prepared from a cell line derived from mpnst tumor tissue of a JCV T-antigen transgenic mouse were subjected to immunoprecipitation/Western blot analysis (panel b). Whole-cell extract was immunoprecipitated with antibodies to p53 (lane 1), NF-2 (lane 2) and T-antigen (lane 3) and separated by SDS–PAGE. After transfer to nylon membrane, the membrane was cut between the 82 and 51 kDa marker and the top half of the membrane was subjected to Western blot analysis with antibodies used to detect JCV T-antigen (panel b, top), while the bottom half was immunoblotted with antibodies to p53 (panel b, bottom). Whole-cell extracts were prepared from the p53-null cell line, Saos-2, transiently transfected with expression plasmids encoding JCV T-antigen, HA-epitope-tagged NF2, and mouse p53, alone or in combination. Immunoprecipitation was performed with anti-NF2 antibody (lanes 1–5) followed by Western blotting to detect T-antigen (panel c, top) or the HA epitope tag of HA-NF2 (panel c, bottom). Anticytochrome c antibody serves as a negative control (lane 6), while whole-cell extract from Saos-2 transfected with T-antigen, HA-NF2, and p53 serves as a positive control (lane 7)

Full figure and legend (174K)

To determine the contribution of p53 to the interaction of T-antigen and NF2, the p53-null cell line, Saos-2, was transfected with expression vectors for JCV T-antigen, mouse p53, and HA-tagged NF2. Whole-cell extracts harvested from cells 36 h after transfection were immunoprecipitated with anti-NF2 antibody followed by Western blotting to detect T-antigen and HA-NF2. As shown in panel c, T-antigen and NF2 could be precipitated with anti-NF2 antibodies from the extracts in the presence and absence of p53, although the T-antigen:NF2 complex was much more efficiently precipitated in the presence of p53 (compare lanes 2 and 5). Of note, faint binding was detected when T-antigen was immunoprecipitated with anti-NF2 antibody in the absence of HA-NF2, which may be due to the precipitation of low levels of endogenous NF2 (lane1). Immunoprecipitation with anticytochrome c antibody serves as a nonspecific control (lane 6).

JCV T-antigen usually is detected with robust intensity in the nucleus of cells, while NF2 is considered to be a cytoplasmic protein. To determine the subcellular localization of these two proteins in mpnst, a clonal cell line was established from mpnst tumor tissue by limiting dilution and utilized for fluorescence microscopy. Toward this end, cells were incubated with mouse monoclonal anti-T-antigen and rabbit polyclonal anti-NF1 or anti-NF2 antibodies followed by detection with FITC-conjugated anti-mouse and rhodamine-conjugated anti-rabbit secondary antibodies. As shown in Figure 5, T-antigen localized to the nucleus of the tumor cells (panels a and d), while NF1 and NF2 localized to the cytoplasm (panels b and e, respectively). Interestingly, in some cells, NF2 localized to both the nucleus and the cytoplasm (panel e, arrow). Overlay of the images demonstrates colocalization of T-antigen and NF2 in the nucleus of some cells (panel f), while NF1 and T-antigen do not colocalize (panel c). In fact, close examination of these cells immunolabeled with anti-NF2 antibody detected with DAB chromogen revealed cells with varied staining patterns, including cytoplasmic/perinuclear (panel g), cytoplasmic (H), or cytoplasmic/nuclear staining within the same cell (panel i). Examination of paraffin sections labeled with NF2 in Figure 2 showed occasional cells within the tumor tissue in which NF2 could be localized to the nucleus as well (data not shown). In order to confirm the nuclear localization of NF2, mpnst cells were transfected with an expression vector encoding HA epitope-tagged NF2. Cellular fractionation was performed as described previously (Chestukhin et al., 2002). As shown in Figure 5 (panel j), T-antigen was predominantly detected in the nucleus, while NA-NF2 could be found in both cellular fractions. p53 and class III -tubulin were used to confirm the purity of the nuclear and cytoplasmic fractions, respectively.

Figure 5.

Double immunolabeling of cells cultured from T-antigen transgenic mouse tumor for T-antigen, NF1, and NF2 and detection of NF2 in the nucleus of mpnst cells. mpnst Tumor tissue harvested from T-antigen transgenic mice was cultured in vitro to generate a clonal cell line. Cells were stained with antibody recognizing T-antigen and detected with rhodamine-conjugated secondary antibody (panels a and d), or antibodies recognizing NF1 and NF2 detected with fluorescein-conjugated secondary antibodies (panels b and e, respectively). Superimposed images shown in panels c and f detect NF2 colocalizes with T-antigen in the nuclei of some tumor cells. Additional images of the mpnst cell line demonstrate cytoplasmic/perinuclear (panel g), nuclear (panel h, center), as well as both nuclear and cytoplasmic staining in the same cell (panel i). Nuclear localization of NF2 was confirmed in fractionated extracts from mpnst cells transfected with an HA-tagged NF2 expression plasmid. As shown in panel j, T-antigen was predominantly localized to the nucleus, while NF2 could be detected in both the nuclear and cytoplasmic fractions. p53 and class III -tubulin serve as positive controls for nuclear and cytoplasmic proteins, respectively, demonstrating the purity of the fractions

Full figure and legend (249K)

Discussion

We have previously described the development of pituitary tumors in this line of JCV T-antigen transgenic mice in which approximately 50% of the mice developed the neoplasias by 9 months of age. As there appears to be a longer latency for the mice to develop mpnst (i.e. greater than 9 months of age), we were not able to determine the absolute penetrance of this phenotype; however, mpnst were detected in approximately 30% of animals surviving beyond 1 year of age. Interestingly, we detected several other tumor types in founder animals of the same strain, including PNET, adrenal neuroblastomas, and lesions in the pituitary, all of which have been previously detected in our various JCV T-antigen transgenic mouse lines. In addition, these tumor types have all been previously described in JCV-inoculated newborn Golden Syrian hamsters (ZuRhein, 1983). Of note, mpnsts were also detected by ZuRhein (1983) in the JCV hamster model. With the exception of the pituitary tumors, all such tumors occurred in cells derived from the neural crest, suggesting that JCV has a strong tropism for neural-origin tissues. We also detected angiosarcomas in some animals, although these tumors were negative for T-antigen (Del Valle, et al, unpublished observations).

In fact, JCV T-antigen has been detected in malignant human tumors of neural origin, including glioblastoma multiforme and medulloblastoma (Krynska et al., 1999a; Del Valle et al., 2001b). In human tumors, as well as the animal models, T-antigen is not detected in the nucleus of every tumor cell, but rather in a random subset. mRNA for T-antigen may be detected in small amounts by RT–PCR in normal transgenic tissues, although the levels may be too low to allow the protein to be detected (Franks et al., 1996). In addition, studies in newborn hamsters inoculated with JCV have detected T-antigen by immunohistochemistry in cells within preneoplastic lesions in the brain (Ressetar et al., 1990).

Interestingly, nuclear NF2 was detected in a subset of tumor cells in culture colocalized with T-antigen as well as in vivo, suggesting that NF2 may be retained in the nucleus. Previous studies have detected NF2 in the nucleus and have predicted that the protein contains a nuclear export domain in addition to a cytoplasmic retention domain (Schmucker et al., 1999; Kressel and Schmucker, 2002). The consequence of NF2 translocation to the nucleus is not yet understood. It is intriguing that wild-type NF1 was detected in the tumors. Inactivation of NF1 has been found in a number of mpnst in patients with NF1; however, not all mpnst display LOH at the NF1 locus, suggesting other mechanisms of cellular transformation may be involved. Of note, although mpnst do not occur in individuals with neurofibromatosis type 2, inactivation of NF2 via deletion of its gene locus has been reported previously in mpnst from patients with NF1 (Rey et al., 1993; Frank et al., 2003).

Studies have also suggested a potential role for NF2 in mesothelioma (Bianchi et al., 1995; Baser et al., 2002) and SV40 DNA and T-antigen protein have been detected in complex with p53 in human mesothelioma (Carbone et al., 1997; Testa et al., 1998). As mentioned above, p53 is often overexpressed in mpnst, although it should be noted that overexpression of p53 in this model is not indicative of its mutation, as T-antigen binds to wild-type p53. In addition, recent studies have suggested a role for NF2 in the inhibition of the p53 regulatory protein, Mdm2, whereby NF2 promotes degradation of Mdm2, thus stabilizing p53. Such findings parallel those described in our mouse model and suggest potential functional consequences for NF2 interaction with T-antigen and/or p53 (Kim et al., 2004). One may envision a scenario in which T-antigen may bind to p53 and/or NF2, thus inactivating the G1/S checkpoint and possibly other cell cycle regulatory pathways downstream of NF2. These steps may occur in NF1+/- tumor cells and bypass the need for LOH at the NF1 locus and could ultimately result in the transformation of cells that express a malignant phenotype, that is, mpnst.

Owing to the demonstrated ability of T-antigen to simultaneously affect multiple pathways of cell cycle control and its ability to induce chromosomal instability (Theile and Grabowski, 1990), we have suggested a hit and run mechanism, whereby T-antigen expression is not required to maintain the transformed phenotype once certain checkpoints are permanently inactivated through acquired mutations. For example, cell lines cultured from a JCV T-antigen transgenic mouse model of medulloblastoma show loss of T-antigen expression coincident with p53 mutation and aneuploidy (Krynska et al., 2000). Based on our current studies, we propose that T-antigen interaction with NF2 and/or p53 may provide such 'hits' toward cellular transformation. Through the study of this animal model, we hope to gain further insight into the mechanisms of JCV T-antigen-mediated cellular transformation and the roles of p53 and NF2 in the pathogenesis of neurofibromatosis.

Materials and methods

Generation and analysis of transgenic animals

Transgenic mice were generated using a 3.2 kb Bal1/NciI fragment from the plasmid, pBJC, containing the JCV Mad-4 early promoter sequence and the coding region for the viral early genes. The DNA was microinjected into the pronucleus of fertilized mouse oocytes generated by mating of C57BL/6J SJL mice by DNX Chrysalis. Founder animals were mated with C57BL/6J mice and all mice were screened for the presence of the transgene by DNA extraction from tail tissue and PCR analysis with primers specific to the transgene as described previously (Franks et al., 1996; Gordon et al., 2000)

Generation and maintenance of cell lines

Tumor tissue harvested from transgenic mice was mechanically disrupted by mincing and pipetting in HBSS containing antibiotics, followed by enzymatic digestion in 0.005%. Trypsin–EDTA and 50 g/ml DNaseI for 1 h at 37°C with constant agitation. After digestion, 10% FBS was added to inactivate the trypsin, the cells were washed and twice passed through a 75 m nylon mesh filter. The cells were collected, washed, resuspended in DMEM containing 10% FBS and antibiotics, and plated in tissue culture flasks. The cells were subcultured when confluent for several passages. Clonal cell lines were isolated from low passage primary cultures by limiting dilution. Cell lines were evaluated by immunocytochemistry, as described below, for the presence of S-100 to confirm neuronal origin.

Antibodies utilized

The following antibodies were used for immunoprecipitation (5 l) and Western blot analysis at a dilution of 1 : 1000: mouse monoclonal anti-T-antigen, Oncogene Science pAb416; mouse monoclonal anti-p53, Oncogene Science pAb421; rabbit polyclonal anti-NF1, Santa Cruz, D; rabbit polyclonal anti-NF2, Santa Cruz C-18 (IP only); mouse monoclonal anti-HA11, BAbCo (Western only), rabbit polyclonal anti-GST, Santa Cruz Z-5 (Western only); and mouse monoclonal anti-Class III -tubulin was a kind gift from Dr Christos Katsetos (Western only). The following antibodies and dilutions were used for immunohistochemistry and immunocytochemistry: mouse monoclonal anti-T-antigen, Oncogene Science pAb416 (1 : 100); mouse monoclonal anti-p53, Dako D07 (1 : 100); rabbit polyclonal anti-p21/WAF-1, Oncogene Science AB-5 (1 : 200); rabbit polyclonal anti-NF1, Santa Cruz, N-20 (1 : 500); rabbit polyclonal anti-NF2, Santa Cruz A-19 (1 : 500); and rabbit polyclonal anti-cow S100, Dako (1 : 250).

Histology, immunohistochemistry, and immunocytochemistry

The normal and tumor tissue collected from transgenic mice were fixed in formalin, embedded in paraffin, and sectioned at 6m for histological and immunohistochemical analysis. Routine light microscopy was performed on sections stained with hematoxylin and eosin (H&E). Immunohistochemistry was performed using the avidin–biotin–perioxidase complex system according to the manufacturer's instructions (Vectastain Elite ABC-Peroxidase Kit, Vector Laboratories). Briefly, sections were deparaffinized in xylene and rehydrated in graded ethanol to water. Nonenzymatic antigen retrieval in citrate buffer was performed for all antibodies by heating the slides to 95°C in 0.01 M citrate (pH 6.0) for 30 min. The sections were rinsed in water followed by PBS, then incubated in methanol containing 3% H₂O₂ for 20 min for the inhibition of endogenous peroxidase. For mouse primary antibodies, an additional step using the mouse-on-mouse kit (MOM Kit, Vector Laboratories) was performed immediately prior to blocking with 5% normal horse serum for 2 h at room temperature. After blocking, sections were incubated with primary antibodies described above overnight at room temperature in a humidified chamber. After washing in PBS, secondary antibodies and avidin–biotin peroxidase complex steps were performed according to the manufacturer's instructions (Vector Laboratories). Finally, sections were developed in 0.02% diaminobenzidine and 0.005% hydrogen peroxide, counterstained with hematoxylin, and mounted. For immunocytochemistry, cells plated on chamber slides were fixed in formalin and processed as described above.

Double labeling immunocytochemistry

Cells were plated on poly-L-lysine-coated glass chamber slides and allowed to attach overnight. Cells were then fixed for 3 min in ice-cold acetone, followed by washing with PBS. After blocking in 2% normal horse serum for 2 h, slides were incubated in the first primary antibody overnight at room temperature. Cells were then washed in PBS, incubated in anti-mouse FITC-conjugated antibody for 2 h, washed again, and blocked in 2% normal goat serum for 2 h. The second primary antibody was then added followed by incubation O/N at room temperature. After incubation, cells were washed in PBS, incubated in anti-rabbit rhodamine-conjugated antibody for 2 h, washed again, and mounted in aqueous mounting medium.

Immunoprecipitation/Western blot analysis

Fractionated nuclear and cytoplasmic protein extracts were prepared using NE-PER (Nuclear and Cytoplasmic Extraction Reagent, Pierce) as described previously in accordance with the manufacturer's instructions (Chestukhin et al., 2002). Nuclear and cytoplasmic proteins (20 and 40 g), respectively, were separated by SDS–PAGE and blotted as described below. Whole-cell extracts were prepared from the normal and tumor tissue by homogenization in TNN buffer containing 100 mM Tris (pH 8.0), 100 mM NaCl, and 0.5% NP-40 plus protease inhibitors (1 mM AEBSF, 0.8 M aprotinin, 20 M leupeptin, 40 M bestatin, 15 M pepstatin A, and 14 M E-64). In all, 150 or 300 g of total protein in 300 or 600 l of TNN buffer, respectively, were incubated overnight at 4°C with 5 l of primary antibody. For GST-T-antigen and NF2 mapping, 500 g of whole-cell extract from the normal mouse brain in 1 ml of TNN was incubated with approximately 500 ng of purified GST or GST-T-antigen mutants plus 5 l of anti-NF2 antibody overnight at 4°C. GST-T-antigen constructs and protein production has been described previously (Safak et al., 1999). All samples were then incubated with 25 l prewashed Pansorbin (Calbiochem) at 4°C for 1 h. Next, the pansorbin was pelleted by centrifugation, the pellets were washed in TNN, and proteins were eluted by resuspension in 25 l of SDS–PAGE sample buffer. The precipitated proteins were fractionated by SDS–PAGE and transferred to PVDF membrane (Hybond-P, Amersham) in transfer buffer containing 192 mM glycine, 25 mM Tris base, and 20% methanol. The blot was incubated in 1 TTBS (0.1% Tween-20, 100 mM Tris (pH 7.5), 0.9% NaCl) containing 5% nonfat dry milk for 1 h followed by incubation in TTBS containing primary antibody and 0.1% milk for 2 h. After washing with TTBS, the blot was incubated in TTBS containing alkaline phosphatase-conjugated secondary antibodies against the appropriate species and 0.1% milk for 2 h. For Western blotting of GST and p53 proteins, protein A sepharose conjugated to alkaline phosphatase was used as the secondary antibody to prevent detection of the IgG heavy and light chains, which would obscure visualization of comigrating GST T-antigen mutants. All blots were then washed in TTBS and proteins were visualized using the CDP-STAR chemiluminscence kit according to the manufacturer's instructions (NEN-DuPont).

References

1. Baser ME, De Rienzo A, Altomare D, Balsara BR, Hendrick NM, Gutmann DH, Pitts LH, Jackler RK and Testa JR. (2002). Neurology, 59, 290–291. | PubMed |
2. Berger JR and Concha M. (1995). J. Neurovirol., 1, 5–18. | PubMed | ISI | ChemPort |
3. Bianchi AB, Mitsunaga SI, Cheng JQ, Klein WM, Jhanwar SC, Seizinger B, Kley N, Klein-Szanto AJ and Testa JR. (1995). Proc. Natl. Acad. Sci. USA, 92, 10819–10820. | PubMed |
4. Carbone M, Rizzo P, Grimley PM, Procopio A, Mew DJ, Shridhar V, de Bartolomeis A, Esposito V, Guiliano MT, Steinberg SM, Levine AS, Giordano A and Pass HI. (1997). Nat. Med., 3, 839–840. | Article | PubMed |
5. Chestukhin A, Litovichick L, Rudich K and DeCaprio JA. (2002). Mol. Cell. Biol., 22, 453–468. | Article | PubMed | ISI | ChemPort |
6. Cichowski K, Shih TS, Schmitt E, Santiago S, Reilly K, McLaughlin ME, Bronson RT and Jacks T. (1999). Science, 286, 2172–2176. | Article | PubMed | ISI | ChemPort |
7. Del Valle L, Gordon J, Assimakopoulou M, Enam S, Geddes JF, Varakis J, Katsetos CD, Croul S and Khalili K. (2001b). Cancer Res., 61, 4287–4293. | PubMed | ISI | ChemPort |
8. Del Valle L, Gordon J, Ferrante P and Khalili K. (2001a). Human Polyomaviruses: Molecular and Clinical Perspectives, Khalili K and Stoner GL (eds). Wiley Liss: New York, NY, pp. 409–430.
9. Enam S, Del Valle L, Lara C, Gan DD, Ortiz-Hidalgo C, Palazzo JP and Khalili K. (2002). Cancer Res., 62, 7093–7101. | PubMed |
10. Frank D, Bunawan B, Holtrup M and Fuzesi L. (2003). Cancer Genet. Cytogenet., 144, 18–22. | Article | PubMed |
11. Franks RR, Rencic A, Gordon J, Zoltick PW, Curtis M, Knobler RL and Khalili K. (1996). Oncogene, 12, 2573–2578. | PubMed |
12. Gan DD, Reiss K, Gorrill T, Del Valle L, Croul S, Giordano A, Fishman P and Khalili K. (2001). Oncogene, 20, 4864–4870. | Article | PubMed |
13. Gordon J, Del Valle L, Otte J and Khalili K. (2000). Oncogene, 19, 4840–4846. | Article | PubMed | ISI | ChemPort |
14. Kim H-S, Henson JW and Frisque RJ. (2001). Human Polyomaviruses: Molecular and Clinical Perspectives, In Khalili K, Stoner GL (eds). Wiley Liss: New York, NY, pp. 73–126.
15. Kim H, Kwak NJ, Lee JY, Choi BH, Lim Y, Ko YJ, Kim YH, Huh PW, Lee KH, Rha HK and Wang YP. (2004). J. Biol. Chem, 279, 7812–7818. | Article | PubMed | ISI | ChemPort |
16. Kressel M and Schmucker B. (2002) 11, 2269–2278.
17. Krynska B, Del Valle L, Croul S, Gordon J, Katsetos K, Carbone M, Giordano A and Khalili K. (1999a). Proc. Natl. Acad. Sci. USA, 96, 11519–11524. | Article | PubMed | ChemPort |
18. Krynska B, Del Valle L, Gordon J, Otte J, Croul S and Khalili K. (2000). Virology, 274, 65–74. | Article | PubMed | ChemPort |
19. Krynska B, Otte J, Franks R, Khalili K and Croul S. (1999b). Oncogene, 18, 39–46. | Article | PubMed | ChemPort |
20. London WT, Houff SA, Madden DL, Fuccillo DA, Gravell M, Wallen WC, Palmer AE, Sever JL, Padgett BL, Walker DL, Zu Rhein GM and Ohashi T. (1978). Science, 201, 1246–1249. | PubMed | ChemPort |
21. Louis DN, Stemmer-Rachamimov AO and Wiestler OD. (2000). Pathology and Genetics of Tumors of the Nervous System, Kleihues P and Cavenee WK (eds). IARC Press: Lyon, France, pp. 219–222.
22. McClatchey AI and Cichowski K. (2001). Biochim. Biophsy. Acta, 1471, M73–M80. | Article |
23. McClatchey AI, Saotome I, Ramesh V, Gusella JF and Jacks T. (1998). Genes Dev., 11, 1121–1233.
24. Ohsumi S, Motoi M and Ogawa K. (1986). Acta Pathol. Jpn., 36, 815–825. | PubMed |
25. Ressetar HG, Walter DL, DeF Webster H, Braun DG and Stoner GL. (1990). Lab. Invest., 62, 287–296. | PubMed |
26. Rey JA, Bello MJ, Kusak ME, de Campos JM and Pestana A. (1993). Cancer Genet. Cytogenet., 66, 28–32. | Article | PubMed |
27. Safak M, Gallia GL, Ansari SA and Khalili K. (1999). J. Virol., 73, 10146–10157. | PubMed |
28. Schmucker B, Tang Y and Kressel M. (1999). Hum. Mol. Genet., 8, 1561–1570. | Article | PubMed | ChemPort |
29. Side L, Taylor B, Cayouette M, Conner E, Thompson P, Luce M and Shannon K. (1997). N. Engl. J. Med., 336, 1713–1720. | Article | PubMed | ISI | ChemPort |
30. Small JA, Khoury G, Jay G, Howley PM and Scangos GA. (1986b). Proc. Natl. Acad. Sci. USA, 83, 8288–8292. | Article | PubMed | ChemPort |
31. Small JA, Scangos GA, Cork L, Jay G and Khoury G. (1986a). Cell, 46, 13–18. | Article | PubMed | ChemPort |
32. Sun CX, Robb VA and Gutmann DH. (2002). J. Cell Sci., 115, 3991–4000. | Article | PubMed | ISI | ChemPort |
33. Testa JR, Carbone M, Hirvonen A, Khalili K, Krynska B, Linnainmaa K, Pooley FD, Rizzo P, Rusch V and Xiao GH. (1998). Cancer Res., 58, 4505–4509. | PubMed | ISI | ChemPort |
34. Theile M and Grabowski G. (1990). Arch. Virol., 113, 221–233. | PubMed |
35. Vogel KS, Klesse LJ, Velasco-Miguel S, Meyers K, Rushing EJ and Parada LF. (1999). Science, 286, 2176–2179. | Article | PubMed | ISI | ChemPort |
36. von Deimling A, Foster R and Krone W. (2000). Pathology and Genetics of Tumors of the Nervous System, Kleihues P and Cavenee WK (eds). IARC Press: Lyon, France, pp. 216–218.
37. Woodruff JM, Kourea HP, Louis DN and Scheithauer BW. (2000). Pathology and Genetics of Tumors of the Nervous System, In Kleihues P and Cavenee WK (eds). IARC Press: Lyon, France, pp. 172–174.
38. ZuRhein GM. (1983). Polyomaviruses and Human Neurological Diseases, In Sever JL and Madden DL (eds). Alan R Liss, Inc.: New York, NY, pp. 205–221.

Acknowledgements

We thank the members of the Center for Neurovirology and Cancer Biology for their helpful suggestions and sharing of reagents. In particular, we thank Dr Kamel Khalili for his generous support for our initial studies and for his continued insight and guidance. We thank Cynthia Schriver for editorial assistance. We thank Dr James Gusella for his generous gift of the NF2 cDNA and Dr Christos D Katsetos for his kind gift of mouse monoclonal anti-Class III -tubulin antibody. This work was supported by grants awarded by the National Institutes of Health to KK and JG.