Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Dec 20.
Published in final edited form as: Mol Carcinog. 2009 Dec;48(12):1139–1148. doi: 10.1002/mc.20568

Inactivation of SNF5 Cooperates with p53 Loss to Accelerate Tumor Formation in Snf5+/-;p53+/- Mice

Jessica DelBove 1, Yasumichi Kuwahara 3, E Lorena Mora-Blanco 5, Virginia Godfrey 1,3, William K Funkhouser 1,6, Christopher D M Fletcher 4, Terry Van Dyke 2,3, Charles W M Roberts 5, Bernard E Weissman 1,2,3
PMCID: PMC3527082  NIHMSID: NIHMS146163  PMID: 19676100

Abstract

Malignant rhabdoid tumors (MRTs) are poorly differentiated pediatric cancers that arise in various anatomical locations and have a very poor outcome. The large majority of these malignancies are caused by loss of function of the SNF5/INI1 component of the SWI/SNF chromatin remodeling complex. However, the mechanism of tumor development associated with SNF5 loss remains unclear. Multiple studies have demonstrated a role for SNF5 in the regulation of cyclin D1, p16INK4A and pRbf activities suggesting it functions through the SWI/SNF complex to affect transcription of genes involved in cell cycle control. Previous studies in genetically engineered mouse models (GEMM) have shown that loss of SNF5 on a p53 null background significantly accelerates tumor development. Here, we use established GEMM to further define the relationship between the SNF5 and p53 tumor suppressor pathways. Combined haploinsufficiency of p53 and Snf5 leads to decreased latency for MRTs arising in alternate anatomical locations but not for the standard facial MRTs. We also observed acceleration in the appearance of T-cell lymphomas in the p53+/-;Snf5+/- mice. Our studies suggest that loss of SNF5 activity does not bestow a selective advantage on the p53 spectrum of tumors in the p53+/-;Snf5+/- mice. However, reduced p53 expression specifically accelerated the growth of a subset of MRTs in these mice.

Keywords: SNF5, p53, malignant rhabdoid tumor, SWI/SNF

INTRODUCTION

Although representing only a small percentage of pediatric neoplasms, malignant rhabdoid tumors (MRTs) are particularly devastating, killing the majority of affected children before their second birthday. These aggressive tumors, first described in the kidney, were originally defined as “rhabdoid” due to the morphological resemblance of the tumor cells to rhabdomyoblasts [1]. However, further characterization of these tumor cells including the lack of expression of myogenic transcription factors established them as distinct from rhabdomyoblasts [1-3]. Rhabdoid tumors are now recognized to occur in diverse locations including the brain, spinal cord, liver, chest wall, and face [4-6]

MRTs have always presented a diagnostic challenge given the largely undifferentiated although heterogeneous appearance, polyphenotypic immunoprofiles, and similarities to other tumor types. The designation of MRT historically relied on the histological appearance of large epithelioid cells with eccentric eosinophilic cytoplasm, vesicular nuclei, prominent nucleoli, and whorled bundles of intermediate filaments. These characteristic rhabdoid cells, described above, are frequently not the predominant cell in a malignant rhabdoid tumor, even appearing rarely in some tumors. Therefore, false diagnoses were frequent. The discovery of deletions and mutations at 22q11.2 involving SNF5/INI1 has since facilitated diagnosis and treatment evaluation [7-11]. SNF5 function is now recognized as being lost in almost 100% of MRTs [12]. The finding that genetic alterations in MRTs are usually limited to SNF5 mutations and deletions, without chromosome- or genome-wide losses or instability, strongly implicates the loss of SNF5 function as the primary cause of these tumors [12-14].

Studies using genetically engineered mouse models (GEMM) have verified SNF5’s specific role in tumor suppression. Snf5 nullizygotes are embryonic lethal, dying at peri-implantation stage while loss of SNF5 activity at birth in conditional Snf5 mice cause death within 3 weeks due to hematopoietic failure [15-18]. A reversible conditional mutant that causes only partial penetrance of SNF5 loss avoids the bone marrow failure and death experienced with the fully penetrant conditional mutation [16]. However, it shows a fully penetrant phenotype, developing CD8+ T cell lymphomas or MRTs by 11 weeks [16]. In comparison, loss of either tumor suppressor genes p53 or p16INK4A leads to tumor development at median ages of 20 weeks and 38 weeks, respectively [19,20]. Therefore, the aggressive phenotype observed in the reversible conditional mice makes Snf5 the most rapidly lethal cancer mutation observed in GEMM following inactivation of a single gene.

GEMM offer the opportunity to examine the interactions between different oncogenic and tumor suppressive pathways in an in vivo system. For example, Guidi et al. demonstrated that Snf5 loss may be epistatic to Rb loss [21]. However, a more recent study showed that loss of the entire Rb family accelerated MRT development [22]. These data seem to fit with the etiology of human disease where MRTs do not appear to lose Rb or p16INK4A function. Gene expression array analyses have also shown that inactivation of SNF5 function in normal mouse embryo fibroblasts alters the expression of genes regulated by the E2F transcription factor family [23].

As mentioned above, SNF5 mutations/deletions seem to be the only genetic alterations that appear in MRTs in humans. While most malignancies found in adults harbor inactivating mutations of the p53 tumor suppressor gene, the mutation frequency in MRTs remains uncharacterized [24,25]. Therefore, several groups examined the effects of p53 loss on the development of MRT in different Snf5 GEMM. Isakoff et. al. reported that simultaneous inactivation of p53 and SNF5 activity in GEM accelerated the development of CD8+/CD4- mature T cell lymphomas in 100% of the mice [23]. They further showed an accumulation of p53 in SNF5 deficient cells in culture displaying growth defects followed by apoptosis. In contrast, the SNF5 deficient cells on the p53-null background showed a reduced level of apoptosis. Yaniv and colleagues further demonstrated that p53 nullizygosity accelerated tumor formation in Snf5 heterozygous mice - with complete penetrance within 19 weeks (from 30% penetrant at 60 weeks on wild type p53 background). Neither the tumor spectrum, nor anatomical locations of the resultant MRTs and sarcomas were altered [26]. This implies a true cooperation, not simply additive effect, as all tumors were SNF5-deficient tumors in the same locations found in Snf5+/- mice on a wild type p53 background, but with a much reduced latency and 100% penetrance [26]. Both groups suggested that p53 loss blocked apoptosis associated with SNF5 inactivation in normal cells and accelerated the appearance of the SNF5-deficient tumors [23,26].

While these studies indicate that combined p53 and SNF5 inactivation accelerates the progression of MRT development in GEMM, both studies used p53 nullizygous mice. Therefore, they did not address the potential effects that SNF5 inactivation might have on tumor development in a p53 heterozygous state. To address this question, we characterized tumor development in p53+/-, Snf5+/- and p53+/- ;Snf5+/- mice. Our results reveal that cooperation between these two important tumor suppressor pathways leads to a reduced latency period but not an apparent increase in tumor penetrance. However, with the possible exception of T cell lymphomas, we do not observe a decrease in the latency period for the spectrum of tumors associated with the p53+/- mice. Rather, we find more rapidly arising MRTs in the p53+/- ;Snf5+/- mice developing in anatomical sites infrequently observed in the Snf5+/- mice alone.

MATERIAL AND METHODS

Generation of p53+/-; Snf5+/- mice

The generation, screening and characterization of p53+/- [27] and Snf5+/- [22] mice were described previously. Snf5+/-;p53+/- mice were derived by crossing p53+/- mice with Snf5+/- mice. Mice with resulting genotypes, Snf5+/+;p53+/-, Snf5+/-;p53+/+, Snf5+/-;p53+/- and Snf5+/+;p53+/+ were born with the expected Mendelian frequencies. p53+/- mice were maintained on a BDF1 C57BL/6 background and Snf5+/- mice were maintained on a C57BL/6 × 129/SV mixed background following procedures approved by the Institutional Animal Care and Use Committees. To minimize potential genetic background effects, all experimental mice used in this study were derived from the same 2 female parental Snf5+/- mice. Mice were monitored twice a week for 24 months or until a tumor or evidence of a tumor (paralysis, swelling, extreme lethargy, etc) was observed. Mice were sacrificed by CO2 asphyxiation followed by cervical dislocation as approved by the Institutional Animal Care and Use Committees.

Histology

Tumor samples and selected tissues were fixed in 10% buffered formalin and sent to the UNC Histopathology Core Facility for paraffin embedding and staining with eosin and hematoxylin. Because of the importance of establishing the identity of each tumor, the histopathology was evaluated independently by 2 senior investigators with previous experience in the area of GEMM for soft tissue sarcomas- Dr. C. Fletcher at Harvard University and Dr. V. Godfrey at University of North Carolina [15,16,22]. Each pathologist evaluated H&E and immunohistochemistry slides in a blinded fashion to further reduce bias. There was initial agreement upon the diagnosis in approximately 90% of the tumors. Cases in disagreement centered SNF5-deficient tumors with scant evidence of typical rhabdoid cell morphology being classified as MRT or sarcoma-non otherwise specified or NOS. Therefore, by agreement, we classified all SNF5-deficient tumors as MRT.

Product Forward primer Reverse primer
wild type Snf5 allele CAGGAAAATGGATGCAACTAAGAT CACCATGCCCCCACCTCCCCTACA
mutant Snf5 allele GGCCAGCTCATTCCTCCCACTCAT CACCATGCCCCCACCTCCCCTACA
wild type p53 allele ACAGCGTGGTGGTACCTTAT TATACTCAGAGCCGCCCT
mutant p53 allele TCCTCGTCGTTTACGGTATC TATACTCAGAGCCGGGCCT
Open in a new tab

Genotyping Analyses

DNA was extracted from mouse toes, tumors or organs using the Qiagen DNAeasy Kit for animal tissues. All PCR reactions were carried out as previously described using EasyStart 0.5mL PCR tubes (MolecularBioProducts #6022) [22,27]. The following additions were made to the EasyStart tube: 1-3ul of DNA extracted from mouse toes, Taq DNA polymerase, specific primers for the mutant or wild type alleles, and ddH2O.

Immunohistochemistry

For histology, sections were stained with hematoxylin and eosin as described [28]. Tumors were fixed in 10% formalin for 16-20 hrs and five μ sections were cut after embedding tissues in paraffin. For IHC analysis, a tissue microarray consisting of triplicate 5μm cores of 31 tumors at different locations in the recipient block from p53+/-, Snf5+/- and Snf5+/-;p53+/- mice was constructed by the Anatomic Pathology Translational Core Laboratory using a Pathology Devices (Westminster, MD) semi-automated TMArrayer. We also included normal mouse tissues including lung, kidney and lymph nodes as controls. IHC was carried out using the Vectastain Elite ABC Kit from Vector Laboratories following manufacturer’s directions, as previously described [22]. All sections were quenched in 3% hydrogen peroxide in methanol for 10 min at room temperature, washed in PBS three times and then blocked in 5% goat serum. Sections were incubated with the primary antibody at 4°C overnight. After washing in PBS three times, the sections were incubated with an anti-rabbit antibody (1:333, Vector, Burlingame, CA) at room temperature for 30 min followed by washing in TS-T buffer (50 mM Tris pH 7.6, 150 mM NaCl, and 0.1% Tween-20) three times. Elite ABC reagent (Vector, Burlingame, CA) was then added to the sections at room temperature for 30 min followed by two washes in TS-T buffer. Stains were developed in DAB (diaminobenzidine tetrahydrochloride) (DAKO, Carpinteria, CA) at room temperature for 2-10 minutes, rinsed in PBS, and counter-stained in Light Green Counterstain or hematoxylin (Biomeda, Foster City, CA). Sections were rinsed in H2O and dehydrated before being mounted in Permacent (Fisher Scientific, Pittsburgh, PA). Microscopy was performed with a Zeiss Axioplan 2 Microscope. All antibodies used are included in the table below.

Antibodies.

ANTIBODY Brief Description Source
SNF5 Mouse monoclonal Transduction Labs, 612110
Vimentin Rabbit polyclonal Biomeda V2009
S100ß Rabbit polyclonal Dako Z0311
p53 (CM5) Rabbit polyclonal Novacastra Labs NCL-p53-CM5p
CD3 Rabbit polyclonal Cell Marque CMC363
B220 Mouse monoclonal BD Pharmingen 550286
Open in a new tab

RESULTS

SNF5 heterozygosity leads to increased penetrance and decreased latency of tumor formation on p53+/- background

Previous studies have shown combined nullizygosity of SNF5 and p53 leads to increased tumor formation, but this occurs at a rate no longer compatible with life, with 100% lethality within 7 weeks [23]. We crossed Snf5+/- females with p53+/- males to yield the Snf5+/-;p53+/- double heterozygous mice, as well as control genotypes: Snf5+/-;p53+/+, Snf5+/+;p53+/+, and Snf5+/+,p53+/- (see Materials and Methods for background strain and origin of mice). The overall survival of these genotypes can be seen in Figure 1 and Table 1. In the p53 heterozygote mice, we observed the expected >50% penetrance and 14.5 month median age (range- 11 to 17.5 months) similar to that observed in previous p53+/- mouse studies [19,27]. Sixty percent of Snf5+/- mice developed tumors with a median age of 15 months (range- 9-20 months), somewhat higher than in the original reports [15,17,18]. The mice heterozygous for both p53 and Snf5, however, had a median age of tumor formation of only 9.5 months (range- 2 to 17.5 months), and a total penetrance of approximately 80%. Thus, the Snf5+/-; p53+/- mice show both an increase in tumor penetrance and an overall reduction in latency similar to previous reports using Snf5+/-;p53-/- mice [23,26].

Figure 1. Kaplan-Meier Survival Curve of Snf5+/-, p53+/- and Snf5+/-;p53+/- Mice.

Figure 1

Open in a new tab

Mice were sacrificed when tumor, paralysis or physical distress was observed. Healthy mice at 25 months were sacrificed and recorded as non-events. N values are noted on the figure. Wild type mice experienced 100% survival at 25 months and are not included here.

Table 1. Tumor Incidence in p53+/-, Snf5+/- and Snf5+/-;p53+/- mice.

Genotype No. with Tumors
(Total No.)		 Median Age
(Months)		 Penetrance Tumor Histology (# of
tumors)
Snf5 +/- 6 (n=10) 15.0 60% MRT (5 f, 1s)
p53 +/- 10 (n=15) 14.5 63% Soft Tissue Sarcoma (4)
Lymphoma (5)
Carcinoma (1)
Snf5 +/- ;p53 +/- 32(n=43) 9.5 74% MRT* (5f, 13s, 4other)
Soft Tissue Sarcoma (5)
Osteosarcomas@ (2)
Lymphoma@ (5)
Open in a new tab
*

- one mouse had 2 MRTs, one located in the spinal cord and one located on a limb.

@

- one mouse had a lymphoma and a osteosarcoma. f-facial; s-spinal cord; other sites included limb, back, shoulder and brain.

Snf5+/-;p53+/- mice develop MRTs, Lymphomas, and Osteosarcomas

As shown by Donehower in 1992 and Harvey in 1993, approximately 50% of p53+/- mice develop tumors by 18 months of age [19,29]. The spectrum of tumors includes primarily lymphomas (~30%) and osteosarcomas (~60%) as well as soft-tissue sarcomas and carcinomas. We also observed a similar spectrum of tumors in our mice with the appearance of 5 lymphomas, 3 osteosarcomas, 1 sarcoma NOS and 1 adenocarcinoma (Table 1). The Snf5+/- mice form primarily malignant rhabdoid tumors of the soft tissues of the head and neck, as well as spinal cord tumors, a pattern that we have consistently found in our Snf5+/- mice. There were 5 facial MRTs and one spinal cord MRT resulting in hind limb paralysis (Table 1).

Approximately 80% of the Snf5+/-;p53+/- mice developed tumors by 17.5 months. Twenty-one out of the 32 mice (66%) had tumors that were classified as MRT. The remaining mice had tumors consistent with the p53+/- mouse spectrum including 5 lymphomas, 4 sarcomas NOS, 2 osteosarcomas and 1 fibrosarcoma. Therefore, it appeared the combined Snf5+/- and p53+/- genotypes did not alter the tumor spectrum of either parental genotype as no tumors completely novel to either genotype were observed. However, a significant change in the anatomical sites of MRTs occurred in the Snf5+/-;p53+/- mice. While the MRTs arose only in the face and spinal cords of the Snf5+/- mice, MRTs appeared in sites in the Snf5+/-;p53+/- mice infrequently or rarely found in the Snf5+/- mice including the limbs, back and intestines.

Loss of SNF5 Expression Occurs in Tumors with Decreased Latency in Snf5+/-;p53+/- Mice

Previous studies from several groups using Snf5+/- mouse models have shown virtually 100% LOH in MRTs that arise in these mice [15,17,18]. Due to T-cell infiltration and the high vascularity of these tumors, we could not obtain reliable results by PCR-based assays. Therefore, we generated a tissue array containing 5 tumors from Snf5+/- mice, 6 tumors from p53+/- mice and 18 tumors from Snf5+/-;p53+/- mice (Table 2). We also included normal lung, kidney and lymph node tissue as controls. We then stained and scored each array for expression of SNF5 and p53 protein by IHC. We found that all normal tissues expressed SNF5 protein at high levels (supplemental data).

Table 2. Loss of Heterozygosity (LOH) Analysis of Mouse Tumors.

graphic file with name nihms-146163-t0004.jpg
Open in a new tab

Tissue arrays were stained for p53 and SNF5 expression as described in the Material and Methods. Each core was then scored for average signal strength of positive tumor cells and for percentage of tumor cells demonstrating some positivity by 2 independent pathologists. These two values were multiplied for each score and averaged for each tumor. Total scores were then assessed for significance by plotting average signal strength versus average % positive. Bases on this analysis, we considered less than <70 as negative, 70-160 as +/- and >160 as positive.

We first examined the tumors that arose in the p53+/- mice including 3 lymphomas and 3 osteosarcomas. All tumors were positive for SNF5 expression consistent with their wild-type genetic background (Table 2). In contrast, the MRTs that arose in the Snf5+/- mice showed no SNF5 protein expression indicating that inactivation of the remaining wild-type Snf5 allele had taken place (Table 2, Figure 2).

Figure 2. Representative IHC Analysis of SNF5 and p53 Expression in Primary Malignant Rhabdoid Tumors.

Figure 2

Open in a new tab

The tumor array was stained for p53 or SNF5 expression using the appropriate antibody and counterstained with hematoxylin as described in the Materials and Methods. Representative cores were visualized using a Nikon FXA microscope and captured using Q-Image software on an Apple G4 computer. All images are 100X.

We next examined 18 representative tumors from the Snf5+/-;p53+/- mice including 10 MRTs, 3 sarcoma NOSs, 2 osteosarcomas and 3 lymphomas. As expected, all MRTs arising on the face of the mice (5/5) did not express SNF5 protein similar to the MRTs from the Snf5+/- mice (Table 2; Figure 2). Furthermore, the ages of onset for these facial tumors did not differ significantly between these groups (Table 2- gray shaded areas). The osteosarcomas retained SNF5 protein expression and also arose in a similar timeframe as the osteosarcomas in the p53+/- mice (Table 2). However, the remaining tumors fell into 2 classes. Lymphomas and sarcoma NOSs appearing on the limbs that arose in a similar timeframe as tumors in the p53+/- mice retained SNF5 expression (Table 2). In contrast, those tumors that had lost SNF5 expression showed decreased latency (<9 months, Table 2- stippled areas). Perhaps, tumor #93 proved most informative (Figure 2). Although this limb sarcoma NOS appeared after 13 months, it contained both an area of SNF5 positive staining and an apparently emerging area of SNF5-negative tumor growth. Interestingly, the H&E shows a more “rhabdoid” appearance in the SNF5-negative section of the tumor suggesting that loss of SNF5 was directing the tumor towards a rhabdoid identity.

We also stained the tissue arrays for p53 expression. Immunohistochemistry analyses for p53 protein expression do not provide clear evidence for loss of expression in tumors because normal tissues show variable levels of expression. Therefore, while we could not determine the status of p53 expression in most of our tumor samples, some of our results were suggestive. Both normal lung and lymph node showed low levels of p53-positive staining cells (supplemental data). Interestingly, 3/5 facial MRTs from both Snf5+/- and Snf5+/-;p53+/- mice showed the same low-level p53 staining. In contrast, all 6 tumors from the p53+/- mice and 13/13 of the remaining tumors from the Snf5+/-;p53+/- mice showed no p53 staining.

The Immunoprofile of Mouse Malignant Rhabdoid Tumors Recapitulates Their Human Counterparts

One question about the validity of GEMM for human cancers is how accurately they embody the features of their representative human malignancy. The haploinsufficient Snf5 mouse has been well characterized as producing MRTs with histologies similar to those observed in humans. We also examined a limited number of immunohistochemical markers to determine whether the MRTs in the Snf5+/- mice showed a similar pattern to their human counterparts. We also looked at the immunoprofile in a representative set of MRTs from the Snf5+/-;p53+/- mice to assess whether loss of p53 function affected the phenotype of these tumors. Approximately 60% of the MRTs, from either genotype, were positive for S100ß expression, a somewhat lower frequency than our previous report [22] (Table 3). We also observed a low to high level of expression of vimentin, an intermediate filament found in cells of mesenchymal origin, in all MRTs (Table 3). These results appear consistent with those found in human MRTs [30-33]. Furthermore, the immunophenotype of MRTs that arose in the Snf5+/-;p53+/- mice did not appear to differ from those found in the Snf5+/- mice.

Table 3. Immunophenotype of Tumors from Snf5+/-, p53+/- and Snf5+/- ;p53+/- Mice.

Malignant
Rhabdoid
Tumors		 Vimentin S100ß
<10 10-70 >70 <10 10-70 >70
Snf5+/- mice 0 2 3 2 0 3
Snf5+/-;p53+/-
mice		 0 2 16 1 7 10
Open in a new tab

Each core was scored for average signal strength of positive tumor cells and for percentage of tumor cells demonstrating some positivity by 2 independent pathologists. These two values were multiplied for each score and averaged for the total number of scores. The values are an average of the scores for at least 3 cores.

T-cell Lymphomas Arise in Snf5+/-;p53+/- mice

We had also previously found that induction of SNF5 inactivation in Mx+;Snf5cond/- mice led to the rapid induction of T-cell lymphomas, a process accelerated by the co-inactivation of p53 expression [16,23]. Previous reports indicated that lymphomas arising in p53+/- mice also appeared to arise primarily from T cells [19,27]. Therefore, we determined the origin of the lymphomas that arose in our mice by IHC staining using the well-established B-cell marker, B220, and T-cell marker, CD3. As shown in Figure 3, all lymphomas stained positive for CD3 but showed little or no staining for B220. Therefore, consistent with the majority of previous results, the inactivation of p53 alone or in combination with SNF5 leads to the development of T-cell lymphomas.

Figure 3. Representative IHC Analysis of CD3 and B220 Expression in Primary Lymphomas.

Figure 3

Open in a new tab

The tumor array was stained for CD3 or B220 expression using the appropriate antibody and counterstained with Light Green or hematoxylin as described in the Materials and Methods. Representative cores were visualized using a Nikon FXA microscope and captured using Q-Image software on an Apple G4 computer. All images are 100X.

DISCUSSION

Almost 30 years have passed since the formal recognition and diagnosis of MRT came to exist [1,34]. While the underlying genetic defect, SNF5 loss, associated with MRT development has been identified, this knowledge has not changed the prognosis of individuals diagnosed with this devastating neoplasm. Given the difficulty in translating insights gained from the molecular mechanisms of human cancer initiation and progression to advances in clinical practice, the genetically engineered mouse model has proven invaluable for accelerating this process. However, it has been difficult to find a suitable GEMM for MRT because Snf5 loss results in embryonic lethality and Snf5 heterozygosity leads to MRT development with an extended latency. Complicating matters for conditional deletion studies, the cell of origin for MRTs remains unknown.

In the two previous studies examining the interactions between the p53 and Snf5 tumor suppressor genes in tumor development in GEMM, loss of SNF5 function took place on a p53 null background [23,26]. In the current mouse model, cells could lose either one or both of these tumor suppressor pathways in order to progress to tumorigenicity. In the case of the facial MRTs that presumably correspond to those found in Snf5+/- mice, we did not observe an increase in penetrance or an apparent decrease in latency for the appearance of these tumors compared to the Snf5 heterozygous mice on a p53 wild type background (Table 2). Our data suggest that this lack of difference was associated with retention of wild-type p53 activity in these tumors. In contrast, we found an increase in penetrance and decrease in latency for the spinal cord tumors in the Snf5+/-;p53+/- mice reminiscent of the dramatic increase in the TgT121; Snf5+/- mice (Table 1 & 2) [22]. Therefore, it seems that cooperation between SNF5 loss and reduced p53 expression or pRbf inactivation accelerates the formation of these neuronally-derived tumors. Recent reports have also shown that SNF5 mutations are associated with some types of familial schwannomatosis in humans [35-37].

We also found that loss of SNF5 protein expression correlated with accelerated tumor appearance for MRTs that occurred in locations other than the face (Table 2). This association also held true for the time of onset of lymphomas. In contrast, we did not observe any changes on the development of osteosarcomas in the p53+/-;Snf5+/- mice. Several possibilities may explain the observed increase in tumor formation with the combined haploinsufficiency of p53 and Snf5. First, Vries et al demonstrated in their 2005 study that 10% of MRT cell lines are near tetraploid and over half of the MRT cell lines bear chromosomal aberrations [38]. They went on to show these aberrations could be corrected in the cell population by addition of SNF5 that leads the genetically unstable cells to senesce. Therefore, SNF5 haploinsufficient mice may experience substantial genomic instability leading to secondary genetic events such as loss of p53 activity. This model could account for the acceleration of development of MRTs in Snf5+/-;p53+/- mice. However, molecular and cytogenetic studies of human MRTs suggest that genomic instability occurs infrequently, with chromosomal aberrations localized mainly to region surrounding the SNF5 gene [7,13]. Furthermore, a recent report by McKenna et al. demonstrated that genomic instability does not occur after SNF5 inactivation in genetically engineered mice [14]. Thus, this may not provide the most likely mechanism.

A second explanation for the increase in tumor formation in the double heterozygous mice is that p53 null cells allow genetic instability, increasing the frequency of Snf5 LOH. The role of p53 loss in generating genetic instability in GEMM remains unresolved [39]. Some studies have shown an increase in genetic instability after loss of p53 function while others, including some of our own reports, suggest that its loss does not lead directly to global chromosomal instability [39,40]. Furthermore, loss of p53 has not been found in primary MRTs of children presenting with this disease (J. Biegel, personal communication). In fact these tumors often initially respond well to chemotherapy and radiation and are sometimes ablated. Unfortunately, MRTs frequently return months to years later in a chemotherapy-resistant form in which the p53 status has not been established. Therefore, we do not favor this paradigm to account for our findings.

We prefer a third explanation for observations in human MRT development as well as the increase in MRTs in the double heterozygous mice. Snf5 LOH is likely occurring in the Snf5+/-;p53+/- mice at an equivalent rate and locations as it does in the p53+/+ mice. However, on a p53 wild type background, these Snf5 null cells are arrested or undergo apoptosis via induction of p21 and subsequent cell cycle arrest via the RB pathway. On the p53+/- background, most cells possess enough functional p53 to undergo these protective processes if they lose SNF5 function. However, a few cells are able to escape them due to low p53 levels causing a slight increase in tumor formation. Since tumors developing following Snf5 loss have a shorter latency than those that arise following p53 loss (as determined in previous mouse studies), an increase in Snf5+/- mouse spectrum tumors (spinal cord and possibly limb MRTs) as well as those associated with p53 heterozygosity (lymphomas) occurs.

Whether secondary genetic events drive the progression of malignant rhabdoid tumors remains unresolved. Given the extensive loss of p53 in most human tumors, this study may lead to a deeper understanding of its potential inactivation in MRT patients. The dramatic increase in spinal cord MRTs in the Snf5+/-; p53+/- mice may implicate p53 loss in the progression or reoccurence of atypical teratoid/rhabdoid tumors in the CNS or AT/RTs, in human patients. While primary MRTs rarely show evidence of p53 inactivation, tumors recurring after radiation and chemotherapeutic treatment have not been analyzed. Indeed, several reports have demonstrated a role for p53 in survival and in the control of death checkpoints in central nervous system neurons [41,42]. Further insight into the relationship of classical tumor suppressor gene pathways like p53 in the etiology of MRT and its specific interactions with SNF5 will inevitably allow for more successful treatment avenues for this devastating childhood neoplasm.

Supplementary Material

Supp Data 01

ACKNOWLEDGEMENTS

We thank Dr. Gary Rosson and Dr. Karen Knudsen for helpful discussions. The work was supported by grants from the National Institutes of Health R01CA91048 (BEW), R01CA113794 (CWMR) and R01 CA046283(TVD). Dr. Delbove was supported, in part, by a training grant from the National Institutes of Health T32ES007017. Dr. Roberts gratefully acknowledges support from the Garrett B. Smith Foundation and the Claudia Adams Barr Foundation.

REFERENCES

  • 1.Haas J, Palmer N, Weinberg A, Beckwith J. Ultrastructure of malignant rhabdoid tumor of he kidney: A distinctive renal tumor of children. Hum Pathol. 1981;12:646–657. doi: 10.1016/s0046-8177(81)80050-0. [DOI] [PubMed] [Google Scholar]
  • 2.Kato H, Ohta S, Koshida S, et al. Expression of pericyte, mesangium and muscle markers in malignant rhabdoid tumor cell lines: differentiation-induction using 5-azacytidine. Cancer Sci. 2003;94(12):1059–1065. doi: 10.1111/j.1349-7006.2003.tb01401.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Suzuki A, Ohta S, Shimada M. Gene expression of malignant rhabdoid tumor cell lines by reverse transcriptase-polymerase chain reaction. Diagn Mol Pathol. 1997;6(6):326–332. doi: 10.1097/00019606-199712000-00004. [DOI] [PubMed] [Google Scholar]
  • 4.Gonzalez-Crussi F, Goldschmidt R, Hsueh W, Trujillo Y. Infantile sarcoma with intracytoplasmic filamentous inclusions. Distinctive tumor of possible histiocytic origin. Cancer. 1982;49:2365–2375. doi: 10.1002/1097-0142(19820601)49:11<2365::aid-cncr2820491125>3.0.co;2-i. [DOI] [PubMed] [Google Scholar]
  • 5.Lynch H, Shurin S, Dahms B, Izant RJ, Lynch J, Danes B. Paravertebral malignant rhabdoid tumor in infancey. In vitro studies of a familial tumor. Cancer. 1983;52:290–296. doi: 10.1002/1097-0142(19830715)52:2<290::aid-cncr2820520216>3.0.co;2-l. [DOI] [PubMed] [Google Scholar]
  • 6.Tsuneyoshi M, Daimaru Y, Hashimoto H, Enjoji M. Malignant soft tissue neoplasms with the histologic features of renal rhabdoid tumors: An ultrastructural and immunohistochemical study. Hum Pathol. 1985;16:1235–1242. doi: 10.1016/s0046-8177(85)80036-8. [DOI] [PubMed] [Google Scholar]
  • 7.Biegel J, Zhou J, Rorke L, Stenstrom C, Wainwright L, Fogelgren B. Germ-line and acquired mutations of INI1 in atypical teratoid and rhabdoid tumors. Cancer Res. 1999;59:74–79. [PubMed] [Google Scholar]
  • 8.Rousseau-Merck M, Versteege I, I L. hSNF5/INI1 inactivation is mainly associatd with homozygous deletions and mitotic recombinations in rhabdoid tumors. Cancer Res. 1999;59:3152–3156. [PubMed] [Google Scholar]
  • 9.Sevenet N, Lellouch-Tubiana A, Schofield D, et al. Spectrum of hSNF5/INI1 somatic mutations in human cancer and genotype-phenotype correlations. Hum Mol Genet. 1999a;8:2359–2368. doi: 10.1093/hmg/8.13.2359. [DOI] [PubMed] [Google Scholar]
  • 10.Sevenet N, Sheridan E, Amram D, Schneider P, Handgretinger R, Delattre O. Constitutional mutation of the mutations of the hSNF5/INI1 gene predispose to a variety of cancers. Am J Hum Genet. 1999b;65:1342–1348. doi: 10.1086/302639. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Versteege I, Sevenet J, Lange M, et al. Truncating mutations of hSNF5/INI1 in aggressive paediatric cancer. Nature. 1998;394:201–206. doi: 10.1038/28212. [DOI] [PubMed] [Google Scholar]
  • 12.Biegel J, Kalpana GV, Knudsen E, et al. The role of INI1 and the SWI/SNF complex in the development of rhabdoid tumors. Cancer Research. 2002;61(1):323–328. [PubMed] [Google Scholar]
  • 13.Rousseau-Merck MF, Fiette L, Klochendler-Yeivin A, Delattre O, Aurias A. Chromosome mechanisms and INI1 inactivation in human and mouse rhabdoid tumors. Cancer Genet Cytogenet. 2005;157(2):127–133. doi: 10.1016/j.cancergencyto.2004.06.006. [DOI] [PubMed] [Google Scholar]
  • 14.McKenna ES, Sansam CG, Cho YJ, et al. Loss of the epigenetic tumor suppressor SNF5 leads to cancer without genomic instability. Mol Cell Biol. 2008;28(20):6223–6233. doi: 10.1128/MCB.00658-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Roberts C, Galusha S, McMenamin M, Fletcher C, Orkin SH. Haploinsufficiency of Snf5 (integrase interactor 1) predisposes to malignant rhabdoid tumors in mice. Proc Natl Acad Sci U S A. 2000;97:13796–13800. doi: 10.1073/pnas.250492697. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Roberts C, Leroux M, Fleming M, Orkin SH. Highly penetrant, rapid tumorigenesis through conditional inversion of the tumor suppressor gene Snf5. Cancer Cell. 2002;2:415–425. doi: 10.1016/s1535-6108(02)00185-x. [DOI] [PubMed] [Google Scholar]
  • 17.Guidi C. Disruption of Ini1 leads to peri-implantation lethality and tumorigenesis in mice. Mol Cell Biol. 2001;21:3598–3603. doi: 10.1128/MCB.21.10.3598-3603.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Klochendler-Yeivin A, Fiette L, FBarra J, Muchardt C, Babinet C, Yaniv M. The murine SNF5/INI1 chromatin remodeling factor is essential for embryonic development and tumor suppression. EMBO Reports. 2000;1:500–506. doi: 10.1093/embo-reports/kvd129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Donehower L. Mice deficient for p53 are developmentally normal butsusceptible to spontaneous tumours. Nature. 1992;356:215–221. doi: 10.1038/356215a0. [DOI] [PubMed] [Google Scholar]
  • 20.Sharpless NE, Bardeesy N, Lee KH, et al. Loss of p16Ink4a with retention of p19Arf predisposes mice to tumorigenesis. Nature. 2001;413(6851):86–91. doi: 10.1038/35092592. [DOI] [PubMed] [Google Scholar]
  • 21.Guidi CJ, Mudhasani R, Hoover K, et al. Functional interaction of the retinoblastoma and Ini1/Snf5 tumor suppressors in cell growth and pituitary tumorigenesis. Cancer Res. 2006;66(16):8076–8082. doi: 10.1158/0008-5472.CAN-06-1451. [DOI] [PubMed] [Google Scholar]
  • 22.Chai J, Lu X, Godfrey V, et al. Tumor-Specific Cooperation of Retinoblastoma Protein Family and Snf5 Inactivation. Cancer Res. 2007;67(7):3002–3009. doi: 10.1158/0008-5472.CAN-06-4207. [DOI] [PubMed] [Google Scholar]
  • 23.Isakoff M, Sansam C, Tamayo P, et al. Inactivation of the Snf5 tumor suppressor stimulates cell cycle progression and cooperates with p53 loss in oncogenic transformation. Proc Natl Acad Sci U S A. 2005 doi: 10.1073/pnas.0509014102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Kinoshita Y, Shiratsuchi H, Tamiya S, et al. Mutations of the p53 gene in malignant rhabdoid tumors of soft tissue and the kidney: immunohistochemical and DNA direct sequencing analysis. J Cancer Res Clin Oncol. 2001;127(6):351–358. doi: 10.1007/s004320000217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Rosson GB, Vincent TS, Oswald BW, Wright CF. Drug resistance in malignant rhabdoid tumor cell lines. Cancer Chemother Pharmacol. 2002;49(2):142–148. doi: 10.1007/s00280-001-0398-y. [DOI] [PubMed] [Google Scholar]
  • 26.Klochendler-Yeivin A, Picarsky E, Yaniv M. Increased DNA Damage Sensitivity and Apoptosis in Cells Lacking the Snf5/Ini1 Subunit of the SWI/SNF Chromatin Remodeling Complex. Mol Cell Biol. 2006;26(7):2661–2674. doi: 10.1128/MCB.26.7.2661-2674.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Jacks T, Remington L, Williams BO, et al. Tumor spectrum analysis in p53-mutant mice. Curr Biol. 1994;4(1):1–7. doi: 10.1016/s0960-9822(00)00002-6. [DOI] [PubMed] [Google Scholar]
  • 28.Saenz Robles MT, Symonds H, Chen J, Van Dyke T. Induction versus progression of brain tumor development: differential functions for the pRB- and p53-targeting domains of simian virus 40 T antigen. Mol Cell Biol. 1994;14(4):2686–2698. doi: 10.1128/mcb.14.4.2686. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Harvey M, McArthur MJ, Montgomery CA, Jr., Butel JS, Bradley A, Donehower LA. Spontaneous and carcinogen-induced tumorigenesis in p53-deficient mice. Nat Genet. 1993;5(3):225–229. doi: 10.1038/ng1193-225. [DOI] [PubMed] [Google Scholar]
  • 30.Fanburg-Smith JC, Hengge M, Hengge UR, Smith JS, Jr., Miettinen M. Extrarenal rhabdoid tumors of soft tissue: a clinicopathologic and immunohistochemical study of 18 cases. Ann Diagn Pathol. 1998;2(6):351–362. doi: 10.1016/s1092-9134(98)80038-5. [DOI] [PubMed] [Google Scholar]
  • 31.Kaiserling E, Ruck P, Handgretinger R, Leipoldt M, Hipfel R. Immunohistochemical and cytogenetic findings in malignant rhabdoid tumor. Gen Diagn Pathol. 1996;141(5-6):327–337. [PubMed] [Google Scholar]
  • 32.Tsokos M, Kouraklis G, Chandra RS, Bhagavan BS, Triche TJ. Malignant rhabdoid tumor of the kidney and soft tissues. Evidence for a diverse morphological and immunocytochemical phenotype. Arch Pathol Lab Med. 1989;113(2):115–120. [PubMed] [Google Scholar]
  • 33.Vogel AM, Gown AM, Caughlan J, Haas JE, Beckwith JB. Rhabdoid tumors of the kidney contain mesenchymal specific and epithelial specific intermediate filament proteins. Lab Invest. 1984;50(2):232–238. [PubMed] [Google Scholar]
  • 34.Beckwith J, Palmer N. Histopathology and prognosis of Wilm’s tumor: Results form the First National Wilm’s Tumor Study. Cancer. 1978;41:1937–1948. doi: 10.1002/1097-0142(197805)41:5<1937::aid-cncr2820410538>3.0.co;2-u. [DOI] [PubMed] [Google Scholar]
  • 35.Hulsebos TJ, Plomp AS, Wolterman RA, Robanus-Maandag EC, Baas F, Wesseling P. Germline mutation of INI1/SMARCB1 in familial schwannomatosis. Am J Hum Genet. 2007;80(4):805–810. doi: 10.1086/513207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Hadfield KD, Newman WG, Bowers NL, et al. Molecular characterisation of SMARCB1 and NF2 in familial and sporadic schwannomatosis. J Med Genet. 2008 doi: 10.1136/jmg.2007.056499. [DOI] [PubMed] [Google Scholar]
  • 37.Sestini R, Bacci C, Provenzano A, Genuardi M, Papi L. Evidence of a four-hit mechanism involving SMARCB1 and NF2 in schwannomatosis-associated schwannomas. Hum Mutat. 2008;29(2):227–231. doi: 10.1002/humu.20679. [DOI] [PubMed] [Google Scholar]
  • 38.Vries RG, Bezrookove V, Zuijderduijn LM, et al. Cancer-associated mutations in chromatin remodeler hSNF5 promote chromosomal instability by compromising the mitotic checkpoint. Genes Dev. 2005;19(6):665–670. doi: 10.1101/gad.335805. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Tomasini R, Mak TW, Melino G. The impact of p53 and p73 on aneuploidy and cancer. Trends Cell Biol. 2008;18(5):244–252. doi: 10.1016/j.tcb.2008.03.003. [DOI] [PubMed] [Google Scholar]
  • 40.Lu X, Magrane G, Yin C, Louis D, Gray J, Van Dyke T. Selective Inactivation of p53 Facilitates Mouse Epithelial Tumor Progression without Chromosomal Instability. Mol Cell Biol. 2001;21(17):6017–6030. doi: 10.1128/MCB.21.17.6017-6030.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Aloyz R, Bamji S, Pozniak C, et al. p53 is essential for developmental neuron death as regulated by the TrkA and p75 neurotrophin receoptors. J Cell Biology. 1998;143:1691–1703. doi: 10.1083/jcb.143.6.1691. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Xiang H, Hockman D, Saya H, Fujiwara T, Schwartzkroin P, Morrison R. Evidence for p53-mediated modulation of neuronal viability. J Neurosci. 1996;16:6753–6765. doi: 10.1523/JNEUROSCI.16-21-06753.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp Data 01
NIHMS146163-supplement-Supp_Data_01.doc (2.5MB, doc)

RESOURCES