International Collaborative Grant Funds Research on the p53 Pathway

An ESUN Announcement

The Liddy Shriver Sarcoma Initiative is funding a $40,000 International Collaborative Grant for the study of the p53 pathway in sarcoma susceptibility. In the study, investigators from the United Kingdom, Germany and Australia will work together to identify inherited genetic markers that can detect individuals at increased risk of developing sarcomas.

Single nucleotide polymorphisms (SNPs)

The DNA sequence of the human genome is peppered with tiny variations that help account for many of the differences between people, from the color of their eyes to the curliness of their hair to their risk of obesity. These mutations, known as single nucleotide polymorphisms (SNPs), are also associated with risk for a wide variety of diseases, including cancers.

One use of cancer causing SNPs is to help estimate an individual’s risk of developing certain cancers. For example, many high-grade tumors are resistant to currently available adjuvant therapies, resulting in few effective treatment options. For this reason, inherited genetic markers, like SNPs, that could detect individuals with an increased risk for these cancers, could increase the likelihood of detecting low-grade cancers, when curative surgical intervention is possible.

The effects of any one SNP on an individual are too small to affect clinical decisions in oncology. Hence, to incorporate inherited genetics into risk assessment and surveillance protocols, it will be important to identify groups of SNPs in genes within the same signalling pathway whose individual effects combine to impact upon the whole pathway in a significant and predictable manner.

P53 mutations and the p53 pathway

The tumor suppressor protein p53 is the cells' most important defense mechanism against cancer. It is mutated and silenced in about half of all cancers, and it has been exhaustively studied by researchers. What scientists have learned about mutations of the p53 tumor suppressor gene has led to the identification of one small subset of the population (those with Li-Fraumeni Syndrome) who benefit from proactive surveillance to detect and treat sarcomas early in their development. The researchers involved in this study hope to discover SNPs in the p53 pathway that can help estimate the risk of sarcoma development in larger populations, leading to earlier diagnosis and more effective treatment.

About the Lead Investigator

Gareth L. Bond, PhD: Dr. Bond is a molecular geneticist with a primary interest in understanding the effect of high-frequency genetic variants on human cancer. His lab is located at the Ludwig Institute for Cancer Research in Great Britain.

V10N5-6 ESUN Copyright © 2013 Liddy Shriver Sarcoma Initiative.

Evaluation of Genetic Biomarkers in the p53 Pathway for Sarcoma Susceptibility

An ESUN Experimental Plan
By Gareth L. Bond, PhD; Helge Taubert, PhD; Peter Würl, MD;
Mandy Ballinger, PhD; and David Thomas, FRACP, PhD

Many sarcomas are resistant to adjuvant therapies, resulting in few effective treatment options. For this reason, it is important to identify inherited genetic markers to detect individuals with an increased risk for sarcoma to increase the likelihood of detecting low-grade sarcomas, when curative surgical intervention is possible. Our understanding of the impact of inherited p53 gene mutations on the development of many different cancers, including sarcomas, has led to their use as critical genetic variants in risk-stratification strategies to determine asymptomatic surveillance for a very small subset of the population. Although these low-frequency, highly penetrant p53 gene mutations clearly increase the risk for sarcoma development little is known about the influence of common genetic variances in the p53 pathway, which could aid in estimating risk of sarcoma development in a larger portion of the population. Only the single nucleotide polymorphism, MDM2 SNP309 has been clearly shown to affect sarcoma in mouse and man. Here, we propose to begin to attack this deficit through a thorough evaluation of four SNPs that have been shown to associate with altered expression of genes, whose protein products have been shown to directly regulate or be regulated by MDM2, namely, estrogen receptor alpha, FOXO3, p21 and E-cadherin. We expect these studies will both deepen our understanding of the molecular defects that disrupt the p53 pathway in the onset of sarcoma, potentially aid in estimating risk in a larger portion of the population and in selecting individuals who could benefit from asymptomatic surveillance protocols.

Introduction to Risk Biomarkers

There are many lines of evidence that sarcoma can have a strong genetic component. For example, sarcomas are more common in people with recognized hereditary cancer syndromes, including retinoblastoma, Li-Fraumeni syndrome (LFS), Gardner's syndrome, Werner's syndrome, neurofibromatosis type 1, and some immunodeficiency syndromes.¹ Additionally, up to 33% of paediatric sarcomas are estimated to be associated with a significant family history of cancers.^2,3 Surgery is still the most effective single modality available to patients with sarcoma. Unfortunately, many sarcomas are resistant to adjuvant therapies, resulting in few effective treatment options for those patients either with unresectable or residual tumours following surgery, or metastatic disease. Therefore, inherited genetic markers that can detect individuals with an increased risk for sarcoma could increase the likelihood of both providing an opportunity for tumour resection with clear histological margins and detecting low-grade sarcomas before they progress to a more malignant state. A recent prospective observational study of asymptomatic LFS individuals with p53 mutations provides strong support for this hypothesis.⁴

LFS is an autosomal-dominant inherited cancer predisposition disorder defined by early onset and family history of multiple cancers, whereby an early onset sarcoma is the key diagnostic feature. LFS has an estimated prevalence of 1:5,000 in the United Kingdom, and 30–60% of affected individuals carry inherited mutations in the p53 tumour suppressor gene. Around 600 families with 395 identified p53 mutations have been identified worldwide (IARC Database). The p53 gene encodes a central node of a cellular stress response pathway that is crucial in tumour suppression, and mediates the responses to numerous cancer therapies. In the above-mentioned prospective observational study, asymptomatic p53 mutation carriers were surveyed using a protocol of non-invasive biochemical and imaging techniques.⁴ Of the 33 p53 mutation carriers identified, 18 underwent surveillance. Significantly, 3-year survival was 100% in the surveillance group compared to only 21% in the non-surveillance group. These results strongly suggest that genetic markers that can identify individuals with an attenuated p53 stress response and, consequently, an increased risk for sarcoma, could dramatically improve outcome through asymptomatic surveillance protocols.

Background: Single Nucleotide Polymorphisms (SNPs)

A higher frequency, albeit lower penetrant inherited genetic variant in the p53 pathway, the single nucleotide polymorphism (SNP) MDM2 (murine double-minute 2) SNP309, can modify sarcoma risk in both p53 mutation carriers and non-carriers in both mice and man.^5-8 MDM2 SNP309 (T/G) is found in a transcriptional enhancer of the MDM2 oncogene that encodes for a key negative regulator of p53. The MDM2 protein directly binds to p53 and disruption of this interaction through phosphorylation is key to the activation of the pathway. We and others demonstrated in mice and man that the different alleles of MDM2 SNP309 are subject to differential transcriptional activation by the transcription factor Sp1, thereby resulting in differential cellular levels of MDM2 and p53 activity. Specifically, the G-allele that is found at a very high frequency in many populations (e.g. 33% in Caucasians and 50% in Chinese) increases the affinity of Sp1, which leads to the increased transcription and expression of MDM2 and inhibition of the p53 stress response. These differences in transcriptional regulation of MDM2 significantly affect p53's role in cancer susceptibility, progression, and survival in many cancer types.^7-9 In our initial study, the MDM2 SNP309 locus served as an age of onset modifier of LFS. LFS p53 mutation carriers with the G-allele of MDM2 SNP309 and more cellular MDM2 were diagnosed with tumours on average 7 years earlier than those that were T/T in genotype.⁶ This observation was reproduced in three independent studies, in which p53 mutation carriers with the G-allele were diagnosed with cancer on average 10, 16, and 12.5 years earlier.⁷ The age of onset modifying effect of MDM2 SNP309 is not restricted to LFS individuals; similar differences arose in soft tissue sarcoma patients with no known germline p53 mutations, whereby patients G/G in genotype were diagnosed 12 years earlier than those T/T in genotype. The age of onset modifying effect of MDM2 SNP309 is not restricted to sarcomas, as similar differences were seen in cohorts of lymphoma, leukemia, head, neck, and oral squamous cell carcinoma, breast, colon, bladder, ovary, brain, melanoma, and liver cancer patients.⁷ More recently, we generated mice that carry either the G- or the T-allele of MDM2 SNP309, and found that cells from animals with a G/G genotype have elevated MDM2 levels, reduced p53 levels, and decreased apoptosis.⁸ G/G mice have shorter tumour latency and decreased survival due to tumour burden for multiple cancers including sarcomas, both in animals with two copies of wild-type p53 and in animals with one copy mutated to model LFS. MDM2 SNP309 can therefore modify sarcoma risk in both p53 mutation carriers and non-carriers. These data also suggest that common genetic variations in the p53 pathway could prove useful in estimating sarcoma risk in a larger portion of the population compared to the currently utilized low-frequency, highly penetrant mutations in the p53 gene.

The effects of any one SNP on an individual are too small to affect clinical decisions in oncology. Hence, to incorporate inherited genetics into risk assessment and surveillance protocols, it will be important to identify groups of SNPs in genes within the same signaling pathway whose individual effects combine to impact upon the whole pathway in a significant and predictable manner. A crucial step will be the identification of SNPs that contain modifying alleles for well-documented functional SNPs, like MDM2 SNP309. Unfortunately, little is known about the effects of other functional p53 pathway SNPs on sarcoma risk, much less about their potential interactions with the MDM2 SNP309 locus. However, the p53 pathway and its key proteins are some the most intensely researched genes in the genome. SNPs worthy of further study in sarcoma are easily identified. Over 175 proteins directly interact with either the MDM2 gene or its RNA or protein products (NCBI). Interestingly, 4 of these genes have been shown to contain SNPs that associate with altered gene expression in cells or tissues, similar to MDM2 SNP309. Furthermore, these SNPs have also been shown to associate significantly with altered phenotypes in humans, such as cancer risk and longevity (Table 1).

Estrogen receptor alpha, ESR1 SNP rs2234693

The ESR1 SNP rs2234693 resides in intron 1 of the ESR1 gene that encodes for estrogen receptor alpha (ER-alpha). Previous reports have correlated this SNP to allelic differences in breast and endometrial cancer risk, whereby the major alleles associate with a higher risk for those malignancies.^10,11 Moreover, allelic difference for this SNP have also been seen for altered risk for other human conditions linked to estrogen signaling such as osteoarthritis, myocardial infarction, schizophrenia, differences in body height, and bone mass density. Importantly, one study provides data that suggest that this SNP or a linked SNP could affect ESR1 transcript levels.¹² Specifically, the major alleles of the SNPs were found to also associate with higher ESR1 transcript levels in human brain tissues. Importantly for this study, ER-alpha has also been shown to inhibit p53 signaling, through the transcriptional activation of MDM2, and the direct inhibition the transcriptional activity of p53. Furthermore, the MDM2 SNP309 locus has been shown to alter the estrogen-mediated transcriptional regulation of MDM2.¹³ Consequently, an increase in ER-alpha expression through the major alleles of one or more linked SNPs could lead to a heightened proliferation of cells and an increased cancer risk at least in part through MDM2 up-regulation, which could, in turn, be influenced by MDM2 SNP309.

Forkhead box O3, FOXO3 SNP rs2802292

The FOXO3 SNP rs2802292 resides in intron 1 of FOXO3 that encodes for a member of the Forkhead Box O transcription factor family. This SNP and its linked SNPs have been shown to associate with allelic differences in longevity and insulin sensitivity.^14-19 For example, centenarians from seven independent cohorts were significantly enriched for the G-allele of FOXO3 SNP rs2802292 with odds ratios ranging from 1.16 to 2.75 compared to controls. Importantly, one study provides data that suggest that one or more of the SNPs could also affect FOXO3 transcript levels. Specifically, the G-allele was shown to associate with increased FOXO3 mRNA levels in skeletal muscles compared to the T-allele mdm2.¹⁵ The role of FOXO3 as a tumor suppressor is well documented and its activity is tightly regulated by many key cancer-signaling pathways. MDM2 has also been shown to regulate FOXO3 through mediating it ubiquitination and poteosomal degredation, thereby inhibiting its tumor suppressive activity. Consequently, a potential increase in FOXO3 through the G-allele or one of its linked alleles could lead to increased tumor suppression, which could be influenced by an increase or decrease in cellular MDM2 levels by the MDM2 SNP309 locus.

Cyclin-dependent kinase inhibitor (p21, Cip1), CDKN1A SNP rs1321311

The CDKN1A SNP rs1321311 resides lies approximately 23kb upstream of the p21 gene. It is found in a region of linkage disequilibrium (LD) comprising of 14 SNPs in over 20kb of upstream sequence, the well-defined promoter and the transcribed region. A recent report demonstrated a clear association of SNP rs1321311 with differential risk of colorectal cancer (CRC).²⁰ The minor T allele at rs1321311 is associated with CRC susceptibility (p=2.32x10-10). Furthermore, three other GWASs clearly demonstrated that rs1321311 associates with altered p21 gene expression, whereby cells with minor T allele associated with lower p21 RNA.^21-23 p21 is centrally involved in determining whether cells live or die. For example, depending on its molecular and microenvironmental context, p21 transcriptional activation by proteins such as the p53 tumor suppressor may lead to apoptosis, senescence or cell cycle arrest. MDM2 has also been shown to directly regulate cellular p21 through regulating its proteosomal degradation.²⁴ Consequently, a potential decrease in p21 through the T-allele or one of its linked alleles could lead to decreased tumor suppression, which could be influenced by an increase or decrease in cellular MDM2 levels by the MDM2 SNP309 locus.

E-cadherin, CDH1 SNP rs16260

CDH1 SNP rs16260, also referred to as -160C/A, is located 160 base pairs upstream from the transcriptional start site of CDH1 that encodes for E-cadherin. The original report demonstrated reduced transcriptional enhancer activity and nuclear protein binding with sequences containing the A-allele compared to the C-allele.²⁵ Consistent with E-cadherin's role as a tumor suppressor, individuals with the A-allele have been noted to have an increased risk of urothelial and prostate cancers.²⁶ Interestingly, MDM2 has also been shown to interact with E-cadherin and stimulate its degradation.²⁷ Consequently, a potential decrease in E-cadherin through the A-allele could lead to decreased tumor suppression, which could be influenced by an increase or decrease in cellular MDM2 levels by the MDM2 SNP309 locus.

Purpose of the Study

We expect that these SNPs could aid in estimating risk of sarcoma development in a larger portion of the population, and in selecting individuals that could benefit from asymptomatic surveillance protocols. Therefore, we propose to utilize two well-characterized sarcoma patient cohorts that we have successfully utilized to characterize MDM2 SNP309, to begin to explore the abilities of these high frequency genetic variants in the p53 pathway to function either in combination or alone to affect sarcoma risk. During this six-month project, we aim to generate the necessary data to apply for funding to more fully examine the influence of these polymorphisms on gene expression, p53 signaling, cancer risk and sarcomagenesis. We expect these studies will both deepen our understanding of the molecular defects that disrupt the p53 pathway in the onset of sarcoma, and potentially aid in estimating risk in a larger portion of the population.

Research Plan

Single nucleotide polymorphisms that would alter cellular expression of ESR1, FOXO3, p21 or e-cadherin could alter cancer risk at least in part through MDM2, which could, in turn, be influenced by MDM2 SNP309, which could allow for a potential interaction between the genotypes of these SNPs in their abilities to affect sarcoma. If true, together, the genotype of these loci could be utilized to aid in estimating risk of sarcoma development in a larger portion of the population, and in selecting individuals who could benefit from asymptomatic surveillance protocols. To begin to ascertain the role of these SNPs in sarcoma risk either alone and/or in combination with MDM2 SNP309, we will utilize two independent cohorts of sarcoma patients and control populations.

i) The German Cohort

One hundred thirty patients (73 females and 57 males; ages 14–87 y; mean 56.3 y) diagnosed with sporadic soft-tissue sarcomas (STS) in the years 1991 to 2001 at the Surgical Clinic 1, University of Leipzig, Germany, and at the Institute of Pathology of the Martin-Luther-University Halle, Germany. All patients had a RO-resection of their primary tumour done by the same team. Each person gave written and informed consent. Approval from the local ethics committee was obtained. One thousand and seven hundred of healthy blood donors of Caucasian ethnicity from Germany will serve as controls.

ii) The ISKS Cohort

ISKS is a global initiative with the Global Study Centre located at the Peter MacCallum Cancer Centre. ISKS Australia began recruitment in 2009 at several sites across Australia. Currently there are federated Local Study Centres at the Centre Leon Berard in Lyon, France, the Huntsman Cancer Institute in Utah, USA, the Mt Sinai Hospital in New York, USA, the Christchurch Hospital in Christchurch, New Zealand, the Vancouver General Hospital in British Columbia, Canada and the Tata Memorial Hospital in Mumbai, India. Further sites in France and the UK are currently being established. The study aims to recruit 3000 families to the study worldwide. Family history, clinical, epidemiological, pathological and mutation information are stored on the central ISKS database. Other ISKS data are available, in a de-identified manner, to researchers for approved research projects. Investigators using ISKS material agree to submit new information found by that project to the central database so that molecular and biological information can be built up on these families and specimens. The database contains family history, clinical, epidemiological, pathologic and mutation information. Nine hundred and seventy sarcoma probands, 1890 family members and 302 controls have been enrolled. The average age of sarcoma diagnosis is 46.5 years and the cohort comprises of 70% soft tissue subtypes and 30% bone sarcomas.

We have genotyped and analyzed the significant associations of MDM2 SNP309 with differential age-dependent sarcoma incidence in both the German and ISKS cohorts, and have noted their striking similarities to each other and to our mouse models of this locus. These results have either been published or are currently being prepared for submission for publication.^5,6,8 Therefore, we propose to utilize these two well-characterized sarcoma patient cohorts, to begin to explore the abilities of these high frequency genetic variants in the p53 pathway to function either in combination or alone to affect sarcoma risk. To do this, we propose to:

1. Genotype the four p53 pathway SNPs in both cohorts.
2. Explore potential allelic differences sarcoma onset and risk.

The ESR1 SNP rs2234693, the FOXO3 SNP rs2802292, the CDKN1A SNP rs1321311, and the CDH1 SNP rs16260 will be genotyped in these 775 cases and 1944 controls using validated allelic-discrimination TaqMan assays from Applied Biosystems that are commonly utilized in our laboratory with great success. Validation of genotypes will be sought in a subset of DNAs through PCR amplification and direct sequencing. Once successful genotyping is achieved, we will explore potential allelic differences in age of sarcoma onset and sarcoma risk as measured by allele enrichment in sarcoma cases versus controls.

Although, the two well-annotated cohorts are limited in number, they should be sufficient to give a preliminary indication of associations in order to justify application for further funding to study additional and larger cohorts, powered to show significance. However, in order to begin to asses the ability to detect modifiers of the age-dependent incidence for these SNPs, we undertook a pilot experiment to asses the feasibility of proposed analysis. Specifically, we first genotyped the ESR1 SNP in the 130 sarcoma patients from the German cohort, which we had previously genotyped for the MDM2 SNP309 locus. As mentioned above, we had previously published that the G-allele of MDM2 SNP309 associated with a 14 year earlier onset of in female sarcoma but not in male sarcoma, most likely due to the fact that the MDM2 SNP309 locus can alter the estrogen-mediated transcriptional regulation of MDM2. Interestingly and similarly to MDM2 SNP309, females showed significant allelic differences in the age of STS diagnosis for the ESR1 SNP, whereby women homozygous for the major alleles were diagnosed on average 12 years earlier than those homozygous for the minor alleles (p=0.0374; T-Test), and no significant differences between genotypes in the male patients were observed.

As mentioned above, we have reasoned that a polymorphic variant in the estrogen receptor alpha gene that regulates estrogen signaling could interact with MDM2 SNP309 to affect cancer and therefore be utilized to more accurately predict an individual's cancer risk, as estrogen has been shown to preferentially increase MDM2 levels from the MDM2 SNP309 G-allele.¹³ In order to begin to explore the ability to note potential functional interactions with MDM2 SNP309, we first explored the possibility that the above-noted allelic differences in the age of sarcoma onset for the ESR1 SNPs could be restricted to female individuals of a specific genotype of MDM2 SNP309. We stratified the female patients according to their MDM2 SNP309 genotypes into two groups, one comprising G-allele carriers and one including those T/T in genotype, and analyzed the allelic differences in the age of STS diagnosis for ESR1 SNP in each group. Interestingly, all significant allelic differences for the ESR1 SNP were found in females harboring the G-allele of MDM2 SNP309. Specifically, female G-allele carriers homozygous for the major allele (T/T) were diagnosed on average 22.5 years earlier (32.5 years median difference) than female G-allele carriers homozygous for the minor allele (C/C) (p=0.031, Junckheere Terpstra Test, Table 2). Interestingly, no significant allelic differences were observed for ESR1 SNP in females T/T in genotype for MDM2 SNP309. In order to asses if similar differences could be noted eventually in the ISKS cohort, we determined the ESR1 SNP and MDM2 SNP309 genotypes in 167 Australian sarcoma patients. Similar to the observations made in the German STS cohort, only female G-allele carriers for MDM2 SNP309, who harbor the T-allele of ESR1 SNP were diagnosed on average 16.1-years earlier in life than G-allele carriers, C/C in genotype for the ESR1 SNP (p=0.0207, Mann-Whitney Test, two-sided, Table 3). Together, these results of our pilot experiment suggest that the ESR1 SNP could interact with MDM2 SNP309 to affect sarcoma incidence and suggest that such interaction will be measurable in our cohorts.

Impact Statement

As summarized elsewhere in this application, the most profound impact of our study will be in the clinical setting of sarcoma surveillance and treatment. However, we also anticipate the results generated from this study will be of wide general interest for scientists and health care professionals that work on the p53 pathway, estrogen signaling, biomarkers in oncology, and for those interested in how cancers arise and what influences susceptibility, age of onset. The work will also be of substantial interest to clinicians, pathologists and the pharmaceutical industry interested both in developing novel therapeutics that take into account genetic variability and tumour susceptibility and in understanding the impact of current therapies. However, our work may also lead to benefits to patients. The proposed identification and characterization of genetic biomarkers that aid in estimating risk of sarcoma development in a larger portion of the population would help in selecting individuals that could benefit from asymptomatic surveillance protocols by increasing the likelihood of both providing an opportunity for tumour resection with clear histological margins and detecting low-grade sarcomas before they progress to a more malignant state.

Conflict of Interest

ESUN adheres to the US National Institutes of Health definition of Conflict of Interest. The authors have no conflicts of interest to declare.

V10N5-6 ESUN Copyright © 2013 Liddy Shriver Sarcoma Initiative.