d

Virology 274, 65–74 (2000)doi:10.1006/viro.2000.0450, available online at http://www.idealibrary.com on

Identification of a Novel p53 Mutation in JCV-Induced Mouse Medulloblastoma

Barbara Krynska, Luis Del Valle, Jennifer Gordon, Jessica Otte, Sidney Croul, and Kamel Khalili1

Center for Neurovirology and Cancer Biology, Laboratory of Brain Tumor Biology, College of Science and Technology, Temple University,1900 North 12th Street, 015-96, Biology Life Sciences Building, Room 203, Philadelphia, Pennsylvania 19122

Received March 3, 2000; returned to author for revision April 6, 2000; accepted May 31, 2000

Medulloblastoma, a malignant invasive tumor of the cerebellum, is one of the most common neoplasms of the nervoussystem in children. Utilization of the human neurotropic virus JC virus (JCV) early gene T-antigen allowed the developmentof a transgenic animal that models human medulloblastoma. Here we describe the characterization of two distinctpopulations of cells derived from the JCV-induced mouse medulloblastoma. Results from immunohistochemical and bio-chemical studies revealed the expression of T-antigen in some but not all tumor cells. In T-antigen-producing cells, T-antigenwas found in association with wild-type p53 and pRb, two tumor suppressors that control cell growth and differentiation. Incells that lack expression of T-antigen, a novel mutant p53 with a deletion between residues 35 and 123 was detected.Morphological differences were observed between the two populations of cells, though there was no significant differencein their growth rates. However, subcutaneous transplantation of the T-antigen-positive, but not T-antigen-negative, cellsresulted in the development of massive tumors in experimental animals. In light of earlier reports on the association of JCV

with human medulloblastoma, the mouse cell lines described in this study may provide a valuable tool for deciphering thepathways involved in the formation and progression of medulloblastoma. © 2000 Academic Press

nsa

INTRODUCTION

The oncogenic potential of several viruses, includingpolyomaviruses, has been well documented in a varietyof animals, and it has been speculated that some malig-nancies in humans may be induced by these viruses. Inrecent years, the relationship of Simian virus 40 (SV40) tohuman cancer has become a focus of attention as re-sults from a number of studies have led to the detectionof viral sequences in several human tumors, includingependymomas, choroid plexus papillomas, osteosarco-mas, and mesotheliomas (Bergsagel et al., 1992; Car-bone et al., 1996; Lednicky et al., 1995; Testa et al., 1998).JC virus (JCV) is another member of the polyomavirusfamily that has recently received much attention withregard to its possible association with certain brain tu-mors (Gallia et al., 1998; Khalili et al., 1999a). JCV is ahuman neurotropic virus infecting greater than 70% of thehuman population and is considered the etiologic agentof the fatal demyelinating disease of the central nervoussystem progressive multifocal leukoencephalopathy(PML) (for review see Berger and Concha, 1995). Thestructural organization of the JCV genome isolated fromthe brains of PML patients shows substantial similarity tothe SV40 genome, particularly within the region of thevirus that is responsible for expression of the viral earlyprotein T-antigen (Frisque et al., 1984). The striking ho-

1

ot

To whom correspondence and reprint requests should be ad-ressed. Fax: (215) 204-0679. E-mail: kkhalili@astro.temple.edu.

65

mology between SV40 T-antigen, which possesses trans-forming ability (Cole, 1996), and JCV T-antigen suggeststhat the JCV early gene product may also have oncogenicpotential. In support of this notion, earlier results fromseveral laboratories have clearly established the abilityof JCV to induce malignancies in several animal models(London et al., 1978, 1983; Ohsumi et al., 1985; ZuRhein,1983). Perhaps the most direct evidence for the involve-ment of JCV T-antigen in tumor formation is presented bytransgenic animal studies using the JCV early genome(Franks et al., 1996; Small et al., 1986). From these stud-ies it was evident that expression of JCV T-antigen in awhole animal system can result in the appearance ofprimitive neuroectodermal tumors and adrenal neuro-blastomas. More recently, we have developed transgenicmice containing the early genome of the archetype formof the virus, JCVCY. Mice between the ages of 9 and 13months that contain the transgene exhibited signs ofneurological illness and developed cerebellar neuroec-todermal origin tumors that histologically resembled hu-man medulloblastoma (Krynska et al., 1999a). This wasan important observation in light of subsequent studiesthat illustrated the presence of JCV DNA sequences in asignificant number of human medulloblastomas (Khaliliet al., 1999b; Krynska et al., 1999b). Interestingly, immu-

ohistochemical studies of human tumor tissue havehown the expression of JCV T-antigen in some, but notll, of the tumor cells containing the JCV sequence. This

bservation is reminiscent of JCV-induced medulloblas-

omas in transgenic mice in which not all tumor cells

0042-6822/00 $35.00Copyright © 2000 by Academic PressAll rights of reproduction in any form reserved.

mp(l

B14, we

66 KRYNSKA ET AL.

were positive for T-antigen production. Here we havederived T-antigen-positive and T-antigen-negative celllines from a mouse medulloblastoma for biological andmolecular studies. Our results show that T-antigen-pro-ducing cells are morphologically distinct from T-antigen-negative cells and that T-antigen can be found in asso-ciation with wild-type p53 and pRb, two tumor suppres-sors that control cell growth and proliferation. Finally, inT-antigen-negative cells, we identified the presence of anovel mutant p53 with a deletion between residues 35and 123.

RESULTS

A JCV T-antigen transgenic mouse harboring tumorsthat resembled human medulloblastomas was sacrificedand the region of the cerebellum containing the tumorwas dissected for in vitro cell culturing. Tumor tissue was

echanically dissociated and a mixed cell culture wasrepared. By limiting dilution, four clonal cell lines

FIG. 1. Morphological appearance of clonal cell lines derived from a mthat exhibited similar morphological features, including a moderate dderived. Four additional cell lines, BS-1B7, BS-1B8, BS-1B13, and BS-1scant cytoplasm and shorter processes (E through H).

named BS-1a, BS-1b, BS-1c, and BS-1f) were estab-ished. The four clones exhibited similar microscopic

features, which included a moderate degree of cyto-plasm and a process-bearing morphology (Figs. 1A–1D).Four additional clonal cell lines, BS-1B7, BS-1B8, BS-1B13, and BS-1B14, with morphological features distinctfrom those of the initial clones, were isolated and werecharacterized by scant cytoplasm and shorter processes(Figs. 1E–1H).

Immunohistochemical evaluation of BS-1B7, BS-1B8,BS-1B13, and BS-1B14 revealed strong positive nuclearstaining with antibody recognizing JCV T-antigen, whilethe other four cell lines, BS-1a, BS-1b, BS-1c, and BS-1f,showed no evidence of T-antigen expression. Figure 2illustrates representative data showing that T-antigenexpression was absent in BS-1a (Fig. 2A) and was de-tected in the nuclei of BS-1B7 (Fig. 2E). Both T-negativeand T-positive cells expressed neuronal nuclear pro-teins, including neurofilament (Figs. 2B and 2F) and syn-aptophysin (Figs. 2C and 2G). Neither T-negative norT-positive cells expressed the astrocytic marker glial

edulloblastoma. Four clonal cell lines, BS-1a, BS-1b, BS-1c, and BS-1f,f cytoplasm and a process-bearing morphology (A through D), werere similarly derived, although they were morphologically distinct with

ouse megree o

fibrillary acidic protein (GFAP) (Figs. 2D and 2H). Theseresults are consistent with the characteristics of the

n (C an

67NOVEL p53 MUTATION

original medulloblastoma tumor tissue obtained from theJCV T-antigen transgenic mice (Krynska et al., 1999a).

To confirm the presence or absence of T-antigen pro-tein in the individual cell lines, immunoprecipitation/Western blot analysis was performed on whole cell ex-tracts prepared from each line. As shown in Fig. 3A, ahigh level of T-antigen protein was detected in cell linesBS-1B7, BS-1B8, BS-1B13, and BS-1B14 (lanes 1 to 4),while the other four cell lines showed no detectablelevels of the viral protein (lanes 5 to 8). The individual celllines were also examined for the presence of p53. To-ward this end, immunoprecipitation/Western blot analy-sis was performed with antibody specific for mousewild-type p53 protein (pAb246) or with antibody specificfor mouse mutant p53 protein (pAb240). As shown in Fig.3B, wild-type p53, but not mutant p53, was detected inextract prepared from the four T-antigen-positive celllines (BS-1B7, BS-1B8, BS-1B13, and BS-1B14), while nei-ther wild-type nor mutant p53 was detectable in the fourT-antigen-negative cell lines (BS-1a, BS-1b, BS-1c, andBS-1f).

Next, we utilized immunoprecipitation/Western blotanalysis to determine whether the p53 in the T-antigen-

FIG. 2. Immunohistochemical evaluation of clonal cell lines. ImmImmunostaining of the cell lines revealed the presence of JCV T-antigefor neuronal cellular marker neurofilament (B and F) and synaptophysiis shown (D and H).

positive cell lines could be found in complex with JCVT-antigen. Immunoprecipitation was first performed with

T-antigen, followed by Western blot analysis with anti-bodies specific for either p53 or T-antigen. As shown inFig. 4A, following immunoprecipitation with antibody toT-antigen, JCV T-antigen and p53 were detected by West-ern blot analysis in extract prepared from cell lines BS-1B13 and BS-1B14, while neither protein was detectablein extract prepared from the T-antigen-negative cell linesBS-1c and BS-1f (compare lanes 1 and 3 with lanes 5 and7). In addition, the reciprocal experiment of immunopre-cipitation with antibody to p53 followed by Western blot-ting to detect both T-antigen and p53 revealed a similarpattern (compare lanes 2 and 4 with lanes 6 and 8).These results demonstrate that p53 and T-antigen can befound in complex with each other in extract preparedfrom JCV T-antigen-positive cell lines.

Next, sequential immunoprecipitations followed byWestern blot analysis were performed to determine therelative level of the interaction observed betweenT-antigen and p53. First, antibody to p53 was utilized topull down the p53:T-antigen complex. This was followedby Western blot analysis of the immunocomplex usinganti-T-antigen antibody to assess the total amount ofT-antigen bound to p53. To assess the total amount of

ining of BS-1a (A through D) and BS-1B7 (E through H) is shown.e nuclei of BS-1B7 (B) and its absence in BS-1a (A). Positive stainingd G) is depicted. Negative immunoreactivity for the glial marker GFAP

unostan in th

T-antigen in the extract, immunoprecipitation using anti-T-antigen was performed in parallel. As shown in Fig. 4B,

p(gBBspct(lig

68 KRYNSKA ET AL.

immunoprecipitation with antibody to T-antigen revealeda more intense band than that seen with anti-p53, sug-gesting that p53 may not be bound to the entire pool ofT-antigen (compare lanes 1 and 2). Following immuno-precipitation with antibody to T-antigen, the depletedextract was utilized for a second round of immunopre-cipitation with antibody to T-antigen. As shown in lane 4,the primary immunoprecipitation was able to precipitateT-antigen to near completion. Similarly, immunoprecipi-tation was performed with antibody to p53, followed by asecondary precipitation with antibody to T-antigen. Asshown in lane 5, T-antigen in excess of that bound to p53can be detected in extract prepared from one of theT-antigen-positive cell lines, while no T-antigen is de-tected in the reciprocal experiment (lane 6). Of note,sequential immunoprecipitations with antibody to p53revealed no cross-reactivity (lane 3). Thus, these results

FIG. 3. Immunoprecipitation/Western blot analysis demonstrates theresence of T-antigen and wild-type p53 in T-antigen-positive cell lines.

A) Immunoprecipitation/Western blot analysis with antibody to T-anti-en illustrates the presence of JCV T-antigen in cell lines BS-1B7,S-1B8, BS-1B13, and BS-1B14 (lanes 1–4) and its absence in cell linesS-1a, BS-1b, BS-1c, and BS-1f (lanes 5–8). Lane 9 shows the expres-ion of JCV T-antigen in the positive control hamster glial cells, HJC-15,roducing JCV T-antigen. (B) Immunoprecipitation with antibodies spe-ific for either wild-type p53 (pAb246) or mutant p53 (pAb240) revealed

he detection of wild-type p53 in the four T-antigen-positive cell lineslanes 1–4), but not in the T-antigen-negative cell lines (lanes 5–8). Inane 9, preimmune serum was used as a control for the detection of p53n the T-antigen-negative cell line BS-1a. *The positions of immuno-lobulin heavy chain (A).

demonstrate the presence of excess T-antigen, which isnot associated with p53 in T-antigen-positive cell lines.

As JCV T-antigen has previously been shown to beable to interact with the retinoblastoma gene productpRb (Dyson et al., 1990), extracts from T-antigen-positiveand -negative cell lines were utilized to examine the

FIG. 4. Association of T-antigen with tumor suppressor proteins p53and pRb in extract prepared from medulloblastoma cell lines. (A)Immunoprecipitation performed with either antibody to T-antigen or p53as shown above each lane followed by Western blot analysis usinganti-T-antigen or anti-p53 antibodies. Top and bottom panels, respec-tively, indicate coprecipitation of T-antigen and p53 in extract fromT-antigen-positive tumor cell lines (lanes 1 through 4), but not in extractfrom T-antigen-negative cell lines (lanes 5 through 8). (B) Sequentialimmunoprecipitations using antibody to either p53 or T-antigen asshown above each lane demonstrate the presence of excess T-antigennot bound to p53 (lane 5). (C) Immunoprecipitation of T-antigen, p53,and pRb with antibody to these proteins showing the presence ofT-antigen and its association with p53 and pRb in T-antigen-positivecells (lanes 1 to 3), but not in T-antigen-negative cells (lanes 4 to 6).

Lane 7 represents expression of JCV T-antigen in the positive control,JCV T-antigen-producing hamster glial cells (HJC-15).

R and s

69NOVEL p53 MUTATION

possible interaction of T-antigen and pRb. As shown inFig. 4C, the total amount of T-antigen protein was de-tected by immunoprecipitation/Western blot analysis intwo of the T-antigen-positive cell lines, BS-1B13 andBS-1B14, but not in the T-antigen-negative cell lines,BS-1c and BS-1f (compare lanes 1 with lanes 4). Like-wise, the amount of T-antigen, which is in complex withp53, is precipitated with antibody to p53 (lanes 2). Inter-estingly, T-antigen was also detected, although at signif-icantly lower levels, when extract from T-antigen-positivecells was immunoprecipitated with antibody to pRb(lanes 3). Detection of a band corresponding to T-antigenwith anti-pRb more faint than that obtained with anti-p53suggested that JCV T-antigen may have a greater bindingaffinity to p53 than pRb in mouse medulloblastoma cells.As expected, no T-antigen was detected in extract pre-pared from T-antigen-negative cell lines (compare lanes3 with lanes 6).

To further investigate the expression of p53 in themouse medulloblastoma cell lines, Northern blot analy-sis was performed utilizing total RNA prepared from theeight cell lines. As shown in Fig. 5A, a transcript corre-sponding to the wild-type p53 was detected in theT-antigen-positive cell lines, BS-1B7, BS-1B8, BS-1B13,

FIG. 5. Northern blot and sequence analyses of the p53 genome in mfrom T-antigen-positive (lanes 1 to 4) and T-antigen-negative (lanes 5 toshows the position of wild-type p53 and the arrowhead points to thebromide-stained gel (D) demonstrate the integrity of the RNA. (C) RT–PCa deletion of amino acids 35 to 123 in all four cell lines.

and BS-1B14 (lanes 1 to 4, shown by an arrow). Interest-ingly, a smaller transcript was detected in the T-antigen-

negative clones, BS-1a, BS-1b, BS-1c, and BS-1f (lanes 5to 8, shown by an arrowhead). No variation in the size ofthe control GAPDH transcript among these cell lines wasdetected (Fig. 5B). To determine whether the smallertranscript was related to p53, RT–PCR amplification wasperformed on mRNA isolated from the T-antigen-negativecell lines. Sequence analysis of the entire PCR productspanning the smaller transcript revealed that the fourT-antigen-negative cell lines each contained a p53 tran-script with a specific deletion that removes amino acidsequences between 35 and 123 (Fig. 5C). A similar mu-tation was observed in cell cultures derived from tumorsobtained from other animals (data not shown). Theseobservations indicate that the genetic alteration seen inp53 may be frequent in the mouse medulloblastoma andis not restricted to a single animal. Interestingly, inspec-tion of the DNA sequence from the mutant p53 cDNArevealed that this change, which removes the entire exon4, can create an in-frame smaller p53 species. Interest-ingly, this region is located near the T-antigen bindingsite. These observations indicate that while in T-antigen-producing cells wild-type p53 is present and in associ-ation with T-antigen, a specific deletion in p53 inT-antigen-negative cells results in expression of mutant

lastoma cell lines. (A) Northern blot analysis for p53 on RNA preparedls. Radiolabeled mouse p53 cDNA was utilized as a probe. The arrownt p53 transcript. Control blot probed with GAPDH (B) and ethidiumequence analyses of the p53 gene in T-antigen negative cells revealed

edullob8) celmuta

variants of p53.Next, to evaluate the tumorigenecity of the eight cell

oN

a

etp -antigei

70 KRYNSKA ET AL.

lines, 5 3 106 cells from each cell line were subcutane-

FIG. 6. Tumorigenicity of mouse medulloblastoma cell lines in vivo. (ach flank of a Nude mouse and after 2 months tumors were detected a

umor dissected from the Nude mouse. (C) Immunohistochemical staresence of T-antigen in nuclei of the tumor cells. (D and E) H & E and T

n the Nude mouse flank.

usly inoculated into the flanks of immunocompromisedude mice. Interestingly, the T-antigen-positive cell lines

mt

ll produced large tumors at the injection sites within 2

T-antigen-positive cell line BS-1B13 was subcutaneously injected intojection sites. (B) Hematoxylin and eosin (H & E) staining of the BS-1B13f the BS-1B13-induced tumor with anti-T-antigen antibody shows then staining, respectively, of small tumors derived from injection of BS-1a

A) Thet the inining o

onths of inoculation. Figure 6A illustrates results fromhe transplantation of BS-1B13 into Nude mice 8 weeks

agtsptJstbtTprglrcctoTtoTmtaf

ittetTioiafvd(tgtgmrTifap

mweiesthlnsaomTen

71NOVEL p53 MUTATION

after inoculation. Only one of the T-antigen-negative celllines, BS-1a, developed a small, yet palpable tumor 2months after injection (data not shown). Histologicalevaluation of the tumors generated by BS-1B13 andBS-1a revealed characteristics similar to those of thenative medulloblastomas in the transgenic mice, includ-ing small blue cells with scant cytoplasm (Figs. 6B and6D, respectively). Some evidence of neuroblastic differ-entiation was also noted in BS-1B13 (Fig. 6B). Immuno-histochemical analysis of the tumors revealed the pres-ence of JCV T-antigen in the nucleus of the majority oftumor cells generated by T-antigen-positive cell line, BS-1B13. T-antigen was not detectable in the small palpabletumor derived from the T-antigen-negative cell line,BS-1a (compare Fig. 6C with 6E).

DISCUSSION

The paradigm of polyomavirus-induced malignancyprovides a unique model system for dissecting pathwaysinvolved in the formation and progression of tumors. Theviral early protein, T-antigen, has the ability to interactwith and functionally inactivate two classes of tumorsuppressors, i.e., p53 and pRb, both of which are essen-tial for cell growth and differentiation. Further, the role ofp53 in genome stability has recently been well docu-mented (Rotter et al., 1994). This is an important notion,

s inactivation of these tumor suppressors caused byenetic variations in their genomes is a common event in

he genesis of several human tumors. In the context ofeveral recent studies pointing to the association ofolyomavirus and human malignancy, we became par-

icularly interested in the role of the human polyomavirusCV in tumorigenesis. Seroepidemiological studies havehown that infection with JCV occurs in a vast majority of

he human population (Walker and Padgett, 1983). Celliological and biochemical studies have demonstrated

hat the virus possesses an oncogenic early protein,-antigen, that can associate with several tumor sup-ressors (Dyson et al., 1990; Krynska et al., 1997). In this

espect, the development of medulloblastoma in trans-enic mice expressing JCV T-antigen may offer an excel-

ent model system for understanding the changes in theegulatory factors during the path to tumorigenesis. Theoexistence of T-antigen-positive and T-antigen-negativeell populations in the tumor tissue and the derivation of

hese cells from the tumors are particularly interestingbservations. Sequencing of the viral genome in the-antigen-negative cells revealed no sequence varia-

ions in the transgenes, which may contribute to the lackf T-antigen expression in these cells (data not shown).hus, one may speculate that silencing of the JCV pro-oter in the T-antigen-negative cells may be caused by

he presence of transcription suppressors and/or the

bsence of transcription activation in these cells. Our

uture studies are aimed at unraveling the mechanisms

nvolved in the modulation of JCV promoter activity inhese tumor cells. The lack of T-antigen expression in theumor cells, however, shows little, if any, effect on prolif-ration of the cells in culture. Furthermore, results from

rypan blue exclusion assay of the T-antigen-positive and-antigen-negative cells revealed no drastic differences

n the growth rates of the cells (Krynska, unpublishedbservations). However, results from Nude mouse stud-

es revealed that unlike T-antigen-negative cells, the T-ntigen-positive cells can induce large tumors in the

lanks of injected animals. This is an interesting obser-ation in light of earlier studies on a rat model that waseveloped by retrovirus-transduced SV40 T-antigen

Salewski et al., 1999). In that setting, it was reported thathe lack of T-antigen expression resulted in a rapidlyrowing malignant tumor. While the discrepancy be-

ween these two observations remains to be investi-ated, one may speculate that differences in the animalodels, i.e., transgenic mice vs retrovirally transduced

ats, and the variations in the biological activity of-antigen from JCV and SV40 may contribute to this

nconsistency. Nevertheless, in both studies, p53 wasound to be present in the wild-type form and in associ-tion with T-antigen in T-antigen-producing cells, while53 was mutated in T-antigen-negative cells.

Our results have shown the occurrence of a novelutation in the p53 gene which removes a large portionithin the N-terminus of p53. This region overlaps withxon 4 of murine p53 and its deletion leaves the remain-

ng p53 sequence in-frame. Curiously, results from West-rn blot analysis failed to detect this mutant p53 protein,uggesting that similar to the full-length wild-type p53,

his variant of p53 is extremely unstable and has a shortalf-life. Detection of a mutant p53 cDNA sequence that

acks exon 4 suggests that perhaps changes in theucleotide sequence of the genome of p53 altered theplicing signals required for appropriate splicing of p53nd caused exon 4 deletion. A detailed genetic analysisf the p53 genome is required to further determine theechanism for the appearance of a p53 deletion in

-antigen-negative cells. Perhaps it should be noted thatarlier studies from several laboratories reported alter-ative splicing of p53 (Flaman et al., 1996; Kulesz-Martin

et al., 1994; Will et al., 1995). The previously reportedalternative that affects the C-terminal of this protein is aninsertional mutation and is distinct from the deletionmutation that we observed in the N-terminal of this genein mouse medulloblastoma. As mentioned earlier, re-moval of exon 4, which is located in close proximity to theT-antigen binding site, may affect proper folding of theprotein and its binding to T-antigen. Therefore, accordingto one model, a mutation in the p53 genome may resultin an alternative pattern of splicing leading to the gen-eration of a p53 variant with no exon 4 which encom-

passes the T-antigen binding site. This allows dissocia-tion of T-antigen from the p53 mutant in the tumor cells.

gifmsap1p6s

bbas

I

uTmp

72 KRYNSKA ET AL.

The free T-antigen may, in turn, through a known nega-tive autoregulatory pathway, suppress expression of itsown promoter. As such, the result will be proliferation ofT-antigen-negative tumor cells with a deletion mutationin p53. The region of p53 that is deleted in the T-antigen-negative cells, i.e., residues 35–123, possesses severalimportant functional features. For example, the residuesbetween 63 and 97 that are proline-rich have similarity toSH3-binding protein and are required for p53-mediatedapoptosis in some experimental systems (Sakamuro etal., 1997). More interestingly, this region plays a role insuppressing tumor cell growth (Walker and Levine, 1996)and is a target for papillomavirus E6-protein-induced p53degradation (Li and Coffino, 1996). While the importanceof this novel p53 mutant in the tumorigenecity of medul-loblastoma in experimental animals remains to be inves-tigated, one may speculate that such a large deletion inthe N-terminal of p53 may functionally inactivate p53 andresult in rapid cell growth in these tumor cells. Futurestudies are aimed at investigating the biological activityof mutant p53 utilizing p53 null cells and at determiningwhether expression of p53 with a deletion in exon 4affects growth property of these cells. Further, it is inter-esting to determine whether dimerization of mutant p53with wild type affects the activity of wild-type p53. Clini-cally, it is of particular interest to determine the occur-rence of the p53 N-terminal deletion in human medullo-blastomas.

MATERIALS AND METHODS

Generation of medulloblastoma cell lines

The cell lines were established from a tumor dissectedfrom a transgenic mouse brain harboring JCV T-antigen(Krynska et al., 1999a). Tissue specimens were washedseveral times with cold phosphate-buffered saline andcut into 0.5- to 1.0-mm segments. Sections were thencultured in DMEM containing 10% FCS and antibiotics.Adherent cells were subcultured at confluence and indi-vidual medulloblastoma cell lines were established us-ing a limited dilution method.

Histology and immunohistochemistry

Cultured cells were plated onto poly-L-lysine-coatedlass chamber slides (Nunc) for light microscopy and

mmunocytochemistry. Immunohistochemistry was per-ormed on formalin-fixed cells or paraffin-embedded tu-

or tissue using the avidin–biotin–peroxidase complexystem according to the manufacturer’s instructions ands described in detail previously (Vectastain Elite ABCeroxidase kit, Vector Laboratories; Krynska et al.,999b). Antibodies utilized for immunostaining includedAb416 to recognize JCV T-antigen (Oncogene Science),

F2 for glial fibrillary acidic protein (Dako), SY38 forynaptophysin (Roche), NF100 for neurofilament (Stern-

erger Monoclonals, Inc.), and a rabbit polyclonal anti-ody to recognize mouse p53 (Chemicon). Hematoxylinnd eosin staining was performed on paraffin-embeddedamples for routine histological analysis.

mmunoprecipitation/Western blot analysis

Immunoprecipitation/Western blots were performedsing extract made from either T-antigen-positive or-antigen-negative cells in TNN buffer containing 100M Tris (pH 8.0), 100 mM NaCl, and 0.5% NP-40 plus

rotease inhibitors (2 mg/ml aprotinin, 10 mg/ml leupep-tin, 10 mg/ml pepstatin, 100 mg/ml PMSF, 100 mg/mlTPCK). Positive control extracts were made from theHJC-15 cell line (Raj et al., 1995). Fifty to one hundredmicrograms of total protein in 150–300 ml of TNN bufferwas incubated overnight with 5 ml of antibody. The im-munocomplexes were analyzed by SDS–PAGE followedby Western blot using specific antibodies as indicated inthe text. For detection of p53, pAb421, which recognizesboth wild-type and mutant p53, pAb240, which detectsonly mutant p53, or pAb246, which only interacts withwild-type p53, was utilized. For detection of T-antigen,SV40 T-antigen monoclonal antibody pAb416 (OncogeneScience), which recognizes JCV T-antigen, was utilized.Since the immunoglobulin heavy chain of the primaryantibody used for immunoprecipitation may obscure de-tection of the p53 protein, protein A–Sepharose conju-gated to alkaline phosphatase (Pierce) was utilized as asecondary antibody, as it recognizes only native IgGcomplexes, and not the denatured heavy chain. All bandswere visualized by chemiluminescence using CDP-STAR(NEN–DuPont) according to the manufacturer’s protocol.

RNA and DNA extraction from medulloblastoma celllines

Total RNA was isolated from medulloblastoma celllines by acid–guanidinium, thiocyanate–phenol–chloro-form extraction as described previously (Chomczynskiand Sacchi, 1987). Genomic DNA was prepared frommedulloblastoma cells by using the Puregene DNA Iso-lation Kit (GENTRA) according to the manufacturer’s in-structions.

RT and PCR amplification of p53 cDNA forsequencing

One microgram of DNase-treated RNA was incubatedwith 40 U of recombinant Moloney murine leukemia virusreverse transcriptase (RT) (Roche) at 45°C for 30 min ina 20-ml reaction in the presence of 40 U of RNasin, 250mM of each dNTP, and 0.5 mM primer R3 (59-GGCCAG-GAACCACTACTCAG) to generate the cDNA. The sam-ples were then heated to 95°C for 5 min to inactivate thereverse transcriptase. Five microliters of the resulting

cDNA was used for PCR in a total volume of 50 ml in thepresence of 250 mM MgCl2, 200 mM of each dNTP, 0.5

bbFwSX

C

C

C

D

F

F

F

G

K

K

K

K

K

K

L

L

L

L

O

R

73NOVEL p53 MUTATION

mM primers, and 2.5 U of Taq DNA polymerase (Roche).The primer L4 (59-AGTTCATTGGGACCATCCTG-39) wasused for second-strand synthesis. After initial denatur-ation for 5 min at 94°C, PCR was performed for 25 cyclesof 95°C (15 s), 58°C (30 s), and 72°C (45 s) with a finalextension of 7 min at 72°C. The resulting RT–PCR prod-ucts were excised from preparatory agarose gels stainedwith ethidium bromide, and DNA was purified using theQIAquick PCR Purification Kit according to the manufac-turer’s instructions (Qiagen). After extraction, DNA wassequenced by automated fluorescent DNA cycle se-quencing using the Applied Biosystems Prism 377 DNASequencer-XL. The following primers were used for se-quencing of the p53 gene: L4 (59-AGTTCATTGGGAC-CATCCTG-39); L2 (59-CGGGTGGAAGGAAATTTGTA-39);and L5 (59-AGAGACCGCCGTACAGAAGA-39).

Northern blot analysis of p53

Approximately 10 mg of total RNA was size fraction-ated on formaldehyde/agarose gel, stained withethidium bromide, and then transferred to nylon mem-brane (Hybond-N, Amersham). Prehybridization wasperformed in Ultrasensitive Hybridization Solution–ULTRAhyb (Ambion) for 1 h at 65°C followed by hy-

ridization in the same solution containing radiola-eled full-length mouse p53 cDNA overnight at 42°C.ollowing hybridization, the nylon membrane wasashed with 23 SSC and 0.1% SDS followed by 0.13SC and 0.1% SDS at 42°C, air-dried, and exposed to-ray film overnight at 270°C.

ACKNOWLEDGMENTS

The authors thank past and present members of the Center forNeurovirology and Cancer Biology for their continued support, insight-ful discussion, and sharing of ideas and reagents, and CarlosLorenzana for technical support. We also thank Cynthia Schriver foreditorial assistance and preparation of the manuscript. This work wasmade possible by grants awarded by NIH to K.K.

REFERENCES

Berger, J. R., and Concha, M. (1995). Progressive multifocal leukoen-cephalopathy: The evolution of a disease once considered rare.J. NeuroVirol. 1, 5–18.

Bergsagel, D. J., Finegold, M. J., Butel, J. S., Kupsky, W. J., and Garcea,R. L. (1992). DNA sequences similar to those of simian virus 40 inependymomas and choroid plexus tumors of childhood. N. Engl.J. Med. 326, 988–993.

arbone, M., Rizzo, P., Procopio, A., Giuliano, M., Pass, H. I., Gebhardt,M. C., Mangham, C., Hansen, M., Malkin, D. F., Bushart, G., Pompetti,F., Picci, P., Levine, A. S., Bergsagel, J. D., and Garcea, R. L. (1996).SV40-like sequences in human bone tumors. Oncogene 13, 527–535.

homczynski, P., and Sacchi, N. (1987). Single-step method of RNAisolation by acid guanidinium thiocyanate–phenol–chloroform ex-traction. Anal. Biochem. 162, 156–159.

ole, C. (1996). Polyomaviruses: The viruses and their replication. In“Fundamental Virology” (B. N. Fields, D. M. Knipe, and P. N. Howley,

Eds.), 3rd ed., pp. 917–946. Lippincott–Raven, Philadelphia.

yson, N., Bernard, R., Friend, S. H., Gooding, L. R., Hassel, J. A., Major,

E. O., Pipas, J. M., Van Dyke, T., and Harlow, E. (1990). Large T-antigens of many polyomaviruses are able to form complexes withthe retinoblastoma protein. J. Virol. 64, 1353–1356.

laman, J.-M., Waridel, F., Estreicher, A., Vannier, A., Limacher, J.-M.,Gilbert, D., Iggo, R., and Frebourg, T. (1996). The human tumoursuppressor gene p53 is alternatively spliced in normal cells. Onco-gene 12, 813–818.

ranks, R., Rencic, A., Gordon, J., Zoltick, P., Knobler, R., and Khalili, K.(1996). Formation of undifferentiated mesenteric tumors in transgenicmice expressing human neurotropic polyomavirus early protein. On-cogene 12, 2573–2578.

risque, R. J., Bream, G. L., and Cannella, M. T. (1984). Human poly-omavirus JC virus genome. J. Virol. 51, 458–469.

allia, G. L., Gordon, J., and Khalili, K. (1998). Tumor pathogenesis ofhuman neurotropic JC virus in the CNS. J. NeuroVirol. 4, 175–181.

halili, K., Del Valle, L., Krynska, B., Gordon, J., Otte, J., and Croul, S.(1999a). The human neurotropic virus, JCV, and its association withCNS tumors. In “The Decade of Autoimmunity” (Y. Shoenfeld, Ed.), pp.263–273. Elsevier Science, Amsterdam.

halili, K., Krynska, B., Del Valle, L., Katsetos, C., and Croul, S. E.(1999b). Medulloblastomas and the human neurotropic polyomavi-rus, JCV. Lancet 353, 1152–1153.

rynska, B., Gordon, J., Otte, J., Franks, R., Knobler, R., Giordano, A.,DeLuca, A., and Khalili, K. (1997). Role of cell cycle regulators intumor formation in transgenic mice expressing the human neuro-tropic virus, JCV, early protein. J. Cell. Biochem. 67, 223–230.

rynska, B., Otte, J., Franks, R., Khalili, K., and Croul, S. (1999a). Humanubiquitous JCV z CY T-antigen gene induces brain tumors in experi-mental animals. Oncogene 18, 39–46.

rynska, B., Del Valle, L., Croul, S. E., Gordon, J., Katsetos, C., Carbone,M., Giordano, A., and Khalili, K. (1999b). Detection of human neuro-tropic JC virus DNA sequence and expression of the viral oncogenicprotein in pediatric medulloblastomas. Proc. Natl. Acad. Sci. USA 96,11519–11524.

ulesz-Martin, M. F., Lisafeld, B., Huang, J., Kisiel, N., and Lee, L. (1994).Endogenous p53 protein generated from wild-type alternativelyspliced p53 RNA in mouse epidermal cells. Mol. Cell. Biol. 14,1698–1708.

ednicky, J. A., Garcea, R. L., Bergsagel, D. J., Butel, J. S. (1995). Naturalsimian virus 40 strains are present in human choroid plexus andependymoma tumors. Virology 212, 710–717.

i, X., and Coffino, P. (1996). High risk human papillomavirus E6 proteinhas two distinct binding sites within p53 of which only one deter-mines degradation. J. Virol. 70, 4509–4516.

ondon, W. T., Houff, S. A., Madden, D. I., Fucillo, D. A., Gravell, M.,Wallen, W. C., Palmer, A. E., Sever, J. L., Padgett, B. L., Walker, D. L.,ZuRhein, G. M., and Ohashi, T. (1978). Brain tumors in owl monkeysinoculated with a human polyomavirus (JC virus). Science 201, 1246–1248.

ondon, W. T., Houff, S. A., McKeever, P. E., Wallen, W. C., Sever, J. L.,Padgett, B. L., and Walker, D. L. (1983). Viral-induced astrocytomas insquirrel monkeys. Prog. Clin. Biol. Res. 105, 227–237.

hsumi, S., Ikehara, I., Motoi, M., Ogawa, K., Nagashima, K., and Yasui,K. (1985). Induction of undifferentiated brain tumors in rats by ahuman polyomavirus (JCV). Jpn. J. Cancer Res. 76, 429–431.

aj, G., Gordon, J., Logan, T. J., Hall, D., Chang, C.-F., Sala, A., DeLuca,A., Giordano, A., and Khalili, K. (1995). Characterization of clonal celllines derived from hamster glioblastomas induced by the neurotropicpapovavirus, JCV. Int. J. Oncol. 7, 801–808.

Rotter, V., Aloni-Grinstein, R., Schwartz, D., Elkind, N. B., Simons, A.,Wolkowicz, R., Lavigne, M., Beserman, P., Kapon, A., and Goldfinger,N. (1994). Does wild-type p53 play a role in normal cell differentia-tion? Semin. Cancer Biol. 5, 229–239.

Sakamuro, D., Sabbatini, P., White, E., and Prendergast, G. C. (1997).The polyproline region of p53 is required to activate apoptosis but

not growth arrest. Oncogene 15, 887–898.

Salewski, H., Bayer, T. A., Eidhoff, U., Preuss, U., Weggen, S., and

S

T

W

W

W

Z

74 KRYNSKA ET AL.

Scheidtmann, K. H. (1999). Increased oncogenecity of subclones ofSV40 large T-induced neuroectodermal tumor cell lines after loss oflarge T expression and concomitant mutation in p53. Cancer Res. 59,1980–1986.

mall, J., Khoury, G., Jay, G., Howley, P., and Scangos, G. (1986). Earlyregions of JC virus and BK virus induce distinct and tissue-specifictumors in transgenic mice. Proc. Natl. Acad. Sci. USA 83, 8288–8292.

esta, J. R., Carbone, M., Hirvonen, A., Khalili, K., Krynska, B., Linnain-maa, K., Pooley, F. D., Rizzo, P., Rusch, V., and Xiao, G. H. (1998). Amulti-institutional study confirms the presence and expression ofsimian virus 40 in human malignant mesotheliomas. Cancer Res. 58,

4505–4509.

alker, D. L., and Padgett, B. L. (1983). The epidemiology of human

papovaviruses. In “Polyomaviruses and Human Neurological Dis-eases” (J. L. Sever and D. L. Madden, Eds.), pp. 99–106. R. Liss, NewYork.

alker, K. K., and Levine, A. J. (1996). Identification of a novel p53functional domain that is necessary for efficient growth suppression.Proc. Natl. Acad. Sci. USA 93, 15335–15340.ill, K., Warnecke, G., Bergmann, S., and Deppert, W. (1995). Species-and tissue-specific expression of the C-terminal alternatively splicedform of the tumor suppressor p53. Nucleic Acids Res. 23, 4023–4028.

uRhein, G. M. (1983). Studies of JC virus-induced nervous systemtumors in the Syrian hamster: a review. In “Polyomaviruses and

Human Neurological Disease” (J. L. Sever and D. M. Madden, Eds.),pp. 205–221. R. Liss, New York.