$1 Million Study on Leiomyosarcoma Set to Begin

August 8, 2016 - The Liddy Shriver Sarcoma Initiative is pleased to announce the funding of a $1 million international collaborative grant focusing on leiomyosarcoma (LMS). Investigators from France, Germany and the United States will join forces in the two-year research project, aiming to identify new therapies and start promising clinical trials for the most common and clinically challenging types of LMS.

Sarcomas are rare cancers of the connective tissues. Leiomyosarcoma is an aggressive sarcoma of smooth muscle cells. LMS treatment often includes surgery, chemotherapy, and radiation therapy. While research has revealed that many types of sarcoma have a unique responsiveness to newer targeted therapies, scientists have yet to discover successful targeted treatments for LMS.

Dr. Bauer, an investigator in the study, is a practicing oncologist. He writes, "I see patients diagnosed with leiomyosarcomas every week, as LMS represents one of the most common sarcoma subtypes. While some progress has been made with regard to novel chemotherapeutic drugs, the outcome of patients remains poor. Exploiting a 'personalized' approach for LMS as part of a team of some of the best sarcoma researchers is a great honor and a unique chance for me to make a difference for my patients."

This study will focus on two of the most common types of LMS, those that arise in the uterus and soft tissues, and also LMS of the bone. The five investigators will pool their resources and expertise, bringing to bear an unprecedented array of laboratory models and advanced testing on this rare cancer.

According to Dr. Jonathan Fletcher, the study's coordinating investigator, the research team will "pick apart LMS cells functionally, gene by gene, to determine which genes are the most uniquely important in LMS growth and survival." Working with extensive resources, he believes the team should be able to quickly move from identifying therapeutic targets to validating them and starting new clinical trials.

In Dr. Frederic Chibon's lab in France, investigators plan to sequence the whole genome and transcriptome of more than 200 LMS tumors. Armed with this information, the research team will use functional genomics to determine which genes and mutations are important for LMS survival and which are irrelevant. Then specific drugs that target key genes and mutations can be tested in mice and zebrafish with LMS tumors.

At the same time, Dr. van de Rijn will be focusing his work on a specific pathway in LMS tumors. He explains, "The PI3K-AKT-mTOR pathway is a biochemical pathway that plays an important role in cell division, in other words - the process that allows tumors to grow. We hope that by studying this pathway in great detail in LMS we may identify opportunities to affect this pathway in such a manner that tumors stop growing."

As the team discovers new potential therapy approaches, they will test them in LMS cell lines, then mice and zebrafish. Dr. Bauer will be working with mouse models in the study. He says, "My role will be to validate targeted treatment approaches in an in vivo model. That means that my group will test whether or not a treatment that works in cell lines can also help a complex organism, such as a mouse that has a LMS."

Dr. Langenau is excited to utlilize zebrafish in LMS research for the first time. "Remarkably, the same drugs that cause tumor regressions in zebrafish have also been shown to be efficacious in treating human cancers. Our group has had great success in using this approach in other sarcomas to define new classes of drugs that kill human cancer cells. We hope to extend these approaches to the study of LMS."

The Investigative Team

This study will be accomplished by a multidisciplinary team of scientists with a wide range experience, skills and and interests. They are united by a strong commitment to translational sarcoma research. Investigators include:

• Jonathan Fletcher, MD, of Brigham and Women’s Hospital, is a medical oncologist and cancer geneticist whose expertise includes LMS genetics and biology.
• Sebastian Bauer, MD, of the West German Cancer Center, is a medical oncologist focusing on sarcoma with a lab-based research group focused on novel sarcoma models.
• Frederic Chibon, PhD, of France's Bergonie Institute, is a lab-based scientist who focuses on identifying predictive biomarkers and therapeutic targets in sarcoma.
• David Langenau, PhD, of Massachusetts General Hospital, developed the first zebrafish models for embryonal rhabdomyosarcoma, and collaborated in development of the first such models for liposarcoma, malignant peripheral nerve sheath tumor, and angiosarcoma.
• Matt van de Rijn, MD, PhD, of Stanford University Medical Center, is a surgical pathologist and molecular pathologist who has pioneered molecular classification of leiomyosarcoma subgroups.

The Funding

This International Collaborative Grant was co-funded by the Liddy Shriver Sarcoma Initiative ($250,000), the LeioMyoSarcoma Direct Research Foundation ($125,000) and the National LeioMyoSarcoma Foundation ($125,000). These funds, in addition to $500,000 which the investigators are focusing on this research, means that $1 Million is being made available to study these rare and dangerous cancers.

Copyright © 2016 Liddy Shriver Sarcoma Initiative.

Maximizing Therapeutic Response in Leiomyosarcoma

An experimental plan by Jonathan Fletcher, MD; Matt van de Rijn, MD, PhD;
Frederic Chibon, PhD; Sebastian Bauer, MD; and David Langenau, PhD

Introduction

This leiomyosarcoma (LMS) research program spearheads an international collaborative consortium to develop new therapies for LMS. The broad goals of the consortium are to:

1) Identify biologic vulnerabilities in LMS

2) Validate those vulnerabilities as potential therapeutic targets

3) Expeditiously advance the most compelling opportunities into definitive evaluations in clinical trials

The aim of our LMS research program is to identify compelling biological candidates for targeted interventions and then subject those candidates to rigorous preclinical and clinical evaluations using our team approach and resources, which include novel LMS cell lines and animal models. Further, so that our LMS consortium effort has a productive focus and avoids the pitfall of becoming diffuse in biologic themes, we have chosen to focus our initial efforts on the PI3K-AKT-mTOR pathway. This pathway has relevance in LMS tumorigenesis, and several of the consortium investigators have biologic and clinical expertise in this pathway, in addition to relevant in vitro and in vivo LMS models. Therefore, our LMS research program initially focuses on maximizing therapeutic response to PI3K-AKT-mTOR inhibition in LMS, even while our genomic discovery efforts concurrently seek other targets that might prove relevant for new LMS therapies.

Background

Leiomyosarcomas Have Varied Biologic and Clinical Features

Leiomyosarcoma (LMS) is a heterogeneous disease including seven main groups, as follows:

1. Uterine LMS, the most frequent type, is often estrogen and progesterone hormone-receptor positive;
2. Intrabdominal LMS, arising in the retroperitoneum, mesentery, or omentum, has many biologic features in common with uterine LMS, and is sometimes grouped with uterine LMS under the designation “gynecologic-type LMS;”
3. LMS in the subcutaneous or deep soft tissue of the limbs;
4. LMS arising in the bone;
5. Vascular LMS, usually arising from the inferior vena cava or veins in the lower limbs;
6. So-called cutaneous LMS, an occasionally recurrent but essentially benign (subcutaneous non-metastasizing) lesion, also known as “atypical intradermal smooth muscle neoplasm”¹; and
7. LMS in childhood, which is a well-differentiated lesion with favorable prognosis.

Some investigators propose that the clinical behavior of LMS in the deep soft tissues of the limbs that originate from small-caliber vessels is similar to that of vascular LMS.

A dichotomous classification of uterine (gynecologic) versus soft-tissue (extrauterine) LMS has been useful to generate genomic and gene expression data, and allows for meaningful biologic analyses. The research undertaken by this international consortium focuses on uterine and extrauterine soft-tissue LMS, which account for more than 95% of clinically relevant LMS. We will also evaluate bone LMS, which is a heretofore unstudied form of LMS, for which we hope our intensive international LMS research program can provide rapid insights that would be impossible to achieve in any single research center.

Although the abovementioned classifications are useful, they do not reflect the entire spectrum of LMS biologic variations. Notably, gene expression profiling of LMS, pioneered by LMS program investigator Dr. van de Rijn,² revealed a nuanced molecular taxonomy of LMS.

When LMS are classified by gene expression profiles, three major subtypes emerge. These are

1. “Muscle-enriched” subtype, characterized by favorable clinical outcome and a strong myogenic signature
2. Extrauterine subtype expressing genes related to protein metabolism, and
3. A subtype characterized by strong expression of genes involved in inflammatory/CSF-1 response.

Importantly, these biologically-defined LMS subtypes are not restricted to a particular anatomic location. For example, the “muscle enriched” subtype occurs in both the uterine and soft tissue locations, and CSF-1 signature LMS are associated with poorer prognosis in both uterine and soft tissue LMS, possibly due to CSF-1 mediated induction of macrophage response.^3,4 These observations underscore the biologic interrelatedness of uterine vs. soft tissue LMS, even while it is well known that there are also biologic differences in these two major anatomic LMS categories, eg uterine LMS having a higher frequency of cases that express estrogen and progesterone hormone receptors.^5,6 

In summary, the common clinically-lethal types of LMS are generally classified in a dichotomous manner, as arising from uterine (gynecologic) versus soft-tissue sites; however, biologic evidence demonstrates shared mechanisms in these major LMS subtypes, and underscores that they should optimally be studied together. Therefore, our research seeks biologic insights and therapeutic advances in both these major subtypes of LMS, which together account for more than 95% of clinically relevant LMS, including virtually all LMS types that metastasize and require systemic therapies.

New Therapeutic Targets in Leiomyosarcoma

Unlike in some of the other sarcoma histotypes, pathognomonic and diagnostically-specific oncogenic mechanisms have not been identified in LMS. Rather, the highly recurrent oncogenic mutations known in LMS involve genes and pathways that are dysregulated in many types of human cancer, including CDKN2A, RB1, TP53 and PTEN. Few of the oncogenic mutations encountered recurrently in LMS are found in leiomyoma, with the exception of MED12 exon 2 mutations,^7-9 which are found in both gynecologic leiomyoma and LMS. Likewise, the highly recurrent oncogenic events in leiomyoma are rarely found in LMS (Figure 1). This evidence suggests that most LMS does not arise from the usual types of leiomyoma. Therefore, study of the biologic alterations in typical leiomyomas does not necessarily shed light on crucial biologic dysregulations responsible for the genesis of most LMS. Rather, the study of transforming mechanisms and therapeutic concepts in LMS requires a variety of informative and accurate LMS laboratory models. Developing such models is an area of expertise for the labs in this LMS research consortium.

Our LMS consortium will work collaboratively on translational challenges, by identifying LMS biologic vulnerabilities through lab-based discovery studies, and then validating the most promising opportunities in clinical trials. To this end, the consortium sought initially to maximize LMS therapeutic response for a pathway meeting the following conditions:

1. has a convincing oncogenic role in LMS, as evidenced by frequent genomic mutations dysregulating the pathway;
2. relevant to the areas of biologic expertise and LMS modeling in our consortium laboratories;
3. known to be targetable by one or more drugs inhibiting the pathway, such that translational concepts arising from the consortium studies can be evaluated immediately in the clinic.

The aim is to study pathways whose oncogenic importance in LMS is unquestionable and where the consortium can identify strategies to effectively target the pathways, either by identifying optimal targets or through combination therapies. With these goals in mind, our LMS consortium has initially prioritized research and clinical trials on the PI3K-AKT-mTOR pathway, while also evaluating therapies for ER/PR, CDK4/6 and CD47. These options are among the more compelling possibilities for targeted therapies in LMS, and they are relevant to the expertise and LMS models already developed by our consortium.

Targeting PI3K-AKT-mTOR in LMS -
Rationale for initial focus on this pathway

Our LMS consortium has chosen, initially, to focus on maximizing therapeutic response to drugs that inhibit the PI3K-AKT-mTOR pathway.

The considerations behind prioritizing this pathway include the following:

Oncogenic credentialing of the PI3K-AKT-mTOR pathway, in LMS:

The PI3K-AKT-mTOR pathway is dysregulated by PTEN homozygous deletions in a subset of both uterine and soft-tissue LMS, resulting in constitutive AKT and mTOR activation. In a pilot study of 12 LMS, we found PTEN homozygous deletions and/or complete loss of PTEN expression by immunoblotting in 5 cases (Figure 2). These pilot studies jibe well with experimental evidence, in which PTEN homozygous deletion targeted to smooth muscle cells results in smooth muscle hyperplasia and rapid development of LMS with resultant ATK-mTOR axis constitutive activation, and which is responsive to mTOR-inhibition.¹⁰

These collective studies underscore that systematic evaluations of PTEN inactivation (and other PI3K pathway activating events) should be performed in a series of LMS. To this end, we will undertake a rigorous approach to PTEN evaluation, including PTEN immunoblotting and IHC, PTEN FISH and sequencing, and – potentially – PTEN methylation analyses in a core group of 40 snap-frozen LMS (25 uterine and 15 soft-tissue), together with exome sequencing evaluations to identify other PI3K-pathway mutations (Projects 1 and 2 – Drs. Chibon and van de Rijn, PIs). The exome sequencing studies will permit identification of genomic alterations in the PI3K-AKT-mTOR pathway, but will also identify alterations outside the PI3K pathway, which might be useful as targets to maximize LMS response to PI3K pathway inhibition. [Rationale for Targeting the PI3K-AKT-mTOR Pathway]

Consortium expertise with the PI3K-AKT-mTOR pathway:

Our consortium has substantial expertise with PI3K-AKT-mTOR pathway dysregulation in sarcoma, including genomic and biologic evaluations, cell line and xenograft modeling, and preclinical validations and clinical trials of targeted therapies. Dr. van de Rijn studies ROR2-mediated dysregulation of this pathway in LMS; Dr. Fletcher with colleague Dr. Adrian Marino-Enriquez performs functional genomic screens (pooled shRNA and ORF screens) for synthetic lethalities with PI3K-mTOR inhibition in LMS, and has published preclinical validations of PI3K-AKT-mTOR pathway targeting; and Dr. Bauer has also performed preclinical studies and has led clinical trials involving PI3K and/or mTOR inhibition in sarcoma.

Laboratory models for PI3K-AKT-mTOR pathway dysregulation in LMS:

Dr. Fletcher has established three LMS cell cultures (LMS03, LMS04, LMS05) with PI3K-AKT-mTOR constitutive activation due to homozygous PTEN deletions (LMS03 and LMS04) or tuberin/TSC2 inactivating mutation (LMS05). Drs. van de Rijn and Bauer have done work with xenograft models of these LMS lines and xenografts of freshly biopsied human LMS. Dr. Langenau is an expert in developing sarcoma models in zebrafish, which provide a very economical means to evaluating new drugs against sarcomas in a live animal. To develop the first-ever zebrafish LMS resource, Dr. Langenau will introduce an activated form of AKT into the normal cells (smooth muscle cells) that give rise to LMS. In all, therefore, the LMS consortium has novel resources to model PI3K-AKT-mTOR pathway constitutive activation in LMS, and to evaluate novel therapeutic strategies against the pathway.

Therapeutic strategies for PI3K-AKT-mTOR inhibition in LMS:

The PI3K-AKT-mTOR pathway has several therapeutic attack points that have been investigated using single-agent targeted therapies (inhibitors of PI3Ks, AKTs, and mTOR) in LMS and/or other sarcoma histotypes. In these clinical studies, a subset of LMS and other sarcomas responded to mTOR inhibition, but most of the responses were minor.^11,12 Therefore, combination therapies are needed to leverage strategies, such as mTOR inhibition, into longer-duration responses. We believe it is feasible from a toxicity standpoint to pursue such strategies, particularly in the case of targets such as mTOR, where single-agent therapies are relatively well tolerated. In addition, the aim of our functional genomics screens (Project 3) is to identify other PI3K-pathway therapeutic targets in LMS that can be inhibited without undue toxicity to the patient, therefore enabling other combination therapy approaches.

In addition to our initial focus on targeting PI3K/ATK/mTOR, our group will continue to advance our commitments to characterizing and targeting other compelling pathways in LMS. These other pathways provide alternative options, which we will draw upon for combination therapy strategies with PI3K/AKT/mTOR inhibition, or which would replace PI3K/ATK/mTOR as the consortium translational and clinical focus, should the PI3K translational studies not meet our current expectations. These alternate targets include CDK4/6, ER, PDGFRA, and CD47.

The Investigative Team

The team members are: 1) strongly committed to translational sarcoma research, each with funded programs for research in LMS and other sarcomas; 2) complementary in research expertise, with strengths in cell biology, surgical and molecular pathology, genetics, LMS cell lines, LMS animal models, correlative science, and experimental therapeutics; and 3) multidisciplinary, with backgrounds in basic science, surgical and molecular pathology, cancer genetics, and medical oncology.

Jonathan Fletcher, MD is a medical oncologist and cancer geneticist whose expertise includes LMS genetics and biology. His group has established a number of LMS cell lines, and seeks to identify and validate new drug targets in LMS. Dr. Fletcher, working with coinvestigator Dr. Adrian Marino-Enriquez, has initiated functional genomics screens to identify synthetic lethals with PI3K-mTOR inhibition, using pooled shRNA high-throughput libraries. These genome-scale screens will be extended to CRISPR screens, which will be juxtaposed with LMS genomic and gene expression screens to identify targets whose knock-down or activation maximizes response to PI3K pathway inhibition in leiomyosarcoma.

Sebastian Bauer, MD is a medical oncologist focusing on sarcoma with a lab-based research group focused on novel sarcoma models and preclinical validations, and also with responsibility for developing novel clinical trials in LMS. He directs the sarcoma program at the West German Cancer Center, one of the largest sarcoma centers in Europe. He has substantial experience developing sarcoma xenografts from both surgical specimens and patient-derived sarcoma cell lines and has recently successfully translated preclinical target validations in a sarcoma model into a “first in disease”-combination phase I trial that combined a kinase inhibitor with a broader-acting epigenetic inhibitor.¹³

Frederic Chibon, PhD is a lab-based scientist who focuses on identifying predictive biomarkers and therapeutic targets in sarcoma. His basic science perspective adds depth to the consortium’s studies on LMS genomic mechanisms and biology, and his interest in molecular biomarkers dovetails well with those of Dr. Matt van de Rijn in this program. He and Dr. van de Rijn were among the first to identify the different molecular subtypes of LMS. He has also been involved in the development and validation of diagnostic and predictive biomarkers that are used in routine surgical pathology practice. He is currently leading whole genome sequencing and transcriptome sequencing for 200 LMS by the French Sarcoma Group.

David Langenau, PhD is the first researcher to develop a transgenic cancer model in zebrafish, revealing the zebrafish as a new and very cost-effective model for cancer discovery. This work led to the development of a new field of science that now spans hundreds of investigators. Previously, Dr. Langenau developed the first zebrafish models for embryonal rhabdomyosarcoma, and collaborated in development of the first such models for liposarcoma, malignant peripheral nerve sheath tumor, and angiosarcoma. Dr. Langenau has also developed approaches to creating sarcoma metastasis in zebrafish, so that these models can be used to assess efficacy of new therapies against both primary and metastatic sarcomas. Therefore, his research team is uniquely suited to carry out the experiments in this international program, where he will create the first zebrafish model of LMS, and thereby provide an inexpensive model by which new drugs can be screened for efficacy (simply by adding them to small wells in which the fish swim). Advantages of the fish models include: 1) fecundity: each female can produce 100-200 eggs per week; 2) small size: thousands of animals can be reared in a relatively small space; 3) reduced cost: fish per diems are 50-fold less then mice at <$0.01/day; and 4) optical clarity: engraftment of malignant cells can be easily visualized by fluorescent labeling.

Matt van de Rijn, MD, PhD is a surgical pathologist and molecular pathologist who has pioneered molecular classification of leiomyosarcoma subgroups. He has a substantial lab-based research group devoted to sarcoma biology and therapeutics, with a primary focus on LMS. Recent areas of translational focus in LMS include ROR2 receptor tyrosine kinase activation of PI3K-mTOR pathways, and use of anti-CD47 therapies to provoke macrophage-mediated phagocytosis of LMS cells. Dr. van de Rijn also has substantial experience in establishing xenografts from human LMS cells.

Program Structure & Flow of Information

As outlined in Figure 3, this consortium will perform both genomic and functional genomic analyses to identify actionable points of attack in key oncogenic pathways in uterine and soft tissue LMS. The initial studies focus on the PI3K-AKT-mTOR pathway, because this pathway has been implicated in LMS, where constitutive activation results from mutations inactivating PTEN, and because the goal of developing combination therapies is realistic, particularly in conjunction with mTOR inhibitors (we have previous experience administering mTOR inhibitors with RTK inhibitors, as one example, to GIST patients as chronic therapy for more than 3 years). Therefore, our immediate goal is to identify approaches that maximize therapeutic response to PI3K-ATK-mTOR inhibition in LMS. In addition, these initial consortium studies will establish a closely coordinated network of research investigators whose complementary LMS models and expertise will be similarly applied to other crucial LMS oncogenic pathways in the future.

The initial phases of the effort (Figure 3, top) are discovery studies, including (top left) genomic credentialing to identify LMS oncogenes and tumor suppressor genes in the PI3K-AKT-mTOR pathway (Drs. Chibon and van de Rijn), and to determine the percentage of uterine and soft tissue LMS with mutations in this pathway. These genomic findings will then be scrutinized in the Molecular Profiling project (Dr. van de Rijn) to determine whether LMS with PI3K-AKT-mTOR pathway mutations have characteristic histopathological features, and to identify more efficient and universal biomarkers for PI3K-AKT-mTOR pathway dysregulation in LMS. Additional biomarkers, serving as surrogates for pathway activation levels, will be identified by functional genomic screens (Figure 2, top right) in which early-passage LMS cell lines are treated with IC50 concentrations of PI3K-mTOR inhibitors. The functional genomic screens will also identify targets for which drug-induced inhibition is selectively toxic in LMS cells receiving PI3K-mTOR inhibitors. Here the aim is to identify synthetic lethalities with PI3K-mTOR inhibition, which can be targeted in innovative combination therapy strategies for LMS.

The targets identified in the genomic and functional genomic discovery efforts will be vetted in a rigorous series of LMS models that have been created by our LMS research consortium. These include early passage immortal LMS cell lines, established from uterine and soft tissue human LMS (Dr. Fletcher), xenografts from uterine and soft tissue LMS (Drs. van de Rijn and Bauer), and zebrafish LMS (Dr. Langenau). By evaluating novel targets and drugs in these complementary LMS models, we will identify promising candidates for evaluation in clinical trials. Heretofore, there have been few accurate LMS models available to enable drug discovery for this challenging disease, and there have been no cell lines that truly represent uterine leiomyosarcoma. We have developed three new immortal robustly-growing leiomyosarcoma cell lines, one of which is from a PTEN-mutant: PTEN-deficient uterine LMS. Another of these cell lines also contains an inactivating PTEN mutation, but is from an extrauterine soft-tissue LMS. The remaining new cell line, also from a soft-tissue LMS, has an inactivating TSC2 mutation, resulting in constitutive mTOR activation.

Projects

Projects 1-5 are summarized here, with details provided for each in the following sections:

Project 1. Pathology Review and Genomic Mechanisms: this project accomplishes pathology and genomic annotation of a core group of 40 LMS. The studies include pathology review, PI3K/AKT/mTOR pathway genomic evaluations, and genomic assays for other pathways that might be relevant for combination therapies with PI3K pathway inhibition (ER/PR expression, CDKN2A/CDK4 pathway).

Project 2. Biologic Profiles and Biomarkers for LMS with PI3K-pathway Dysregulation: this project accomplishes gene expression annotation for the core group of 40 LMS. These studies extend the genomic evaluations in Project 1, by identifying and validating biomarkers for PI3K/AKT/mTOR pathway dysregulation.

Project 3.  Functional Genomics of LMS and In Vitro Validations: this project identifies targets that might maximize therapeutic response to PI3K-pathway inhibition in LMS, and also identifies alternate therapeutic targets in LMS. The project also performs preclinical validations of therapeutic concepts in LMS in vitro (cell line) models.

Project 4. Zebrafish model of LMS to enhance preclinical drug validations: this project develops the first-ever zebrafish model of LMS, providing a crucial resource for the therapeutic evaluations in Project 5.

Project 5. In vivo validations of new LMS therapies: this project extends the Project 3 preclinical evaluations to in vivo models, in which promising therapies are evaluated rigorously against human LMS xenografted in mice and against zebrafish LMS. The goal of Project 4 (in concert with Project 3) is to provide rigorous preclinical evidence for new therapeutic opportunities, so as to prioritize compounds for evaluation in clinical trials.

Project 1: Pathology Review and Genomic Mechanisms

Principal Investigator: Dr. Chibon

Project 1 provides exome sequencing annotations for 40 snap-frozen LMS (20 soft-tissue and 20 uterine) and companion nonneoplastic tissues, which have already been banked by Drs. Fletcher and Bauer. The immediate aim of these studies is to annotate a representative group of LMS for functional mutations in the PI3K/AKT/mTOR pathway. These studies will be supplemented by protein PI3K pathway protein evaluations, including IHC evaluation of PTEN expression, and immunoblotting evaluations of AKT activation, PTEN expression, and TSC1 and TSC2 expression. In all, these studies will provide a comprehensive assessment of genomic and protein dysregulation in the PI3K/ATK/mTOR pathway. These same LMS will then be further assessed by gene expression profiling in Project 2, to identify molecular profiles and biomarkers for PI3K/ATK/mTOR pathway dysregulation. In sum, these studies will identify LMS with PI3K pathway genomic mechanisms and will generate hypotheses regarding recurrent gene mutations associated with PI3K genomic dysregulation. Biologically- and clinically-important associations with PI3K/PTEN mutations have been found in other sarcomas, such as the linkage of a recurrent MYOD1 mutation in poor-prognosis rhabdomyosarcoma with concurrent mutation of PIK3CA or PTEN.¹⁴

The consortium specimens analyzed in Project 1 fall into three categories: 1) the core group of 40 snap-frozen LMS (20 uterine, 20 soft tissue) and companion nonneoplastic tissues; 2) LMS cell lines; and 3) LMS xenografted into mice. All human specimens are provided with anonymized clinical data that includes clinical stage at presentation, pattern of disease progression/recurrence, therapies administered, and clinical response to therapies up to the time the biopsy was performed.  Collectively, the Project 1 genomic studies serve both to annotate genomic perturbations (in the initial studies, with a focus on the PI3K pathway) and to demonstrate which of our consortium’s LMS models have genomic perturbations in cell pathways relevant to our studies. 

A comprehensive genomic analysis of PI3K/AKT/mTOR pathway components is needed to better understand the potentially diverse mechanism(s) of pathway activation in LMS.  For instance, in GISTs and other sarcomas, we have identified tumors with mutations in multiple regulatory components of the PI3K/AKT/mTOR pathway, and it is likely that these mutations work synergistically to optimize pathway activation.  This has both diagnostic and therapeutic implications.  Clinically, it is important to accurately identify patients with tumors that harbor genetic aberration(s) in the PI3K/AKT/mTOR pathway, as they will more likely respond to targeted inhibition of PIK3CA, mTOR or other pathway proteins.  Furthermore, it is likely that the exact nature (gene target) of an oncogenic mutation in the PI3K/AKT/mTOR pathway will impact the tumor response to a given dose of single-agent or combination PIK3CA/mTOR inhibitor(s). Therefore, a comprehensive genomic analysis of PI3K/AKT/mTOR pathway components in both clinical samples and LMS experimental models will allow us to understand and optimize the therapeutic potential of PIK3CA/mTOR inhibition in LMS.

Because our LMS consortium studies will likely evaluate co-targeting of the PI3K pathway in combination with CDK4/6-inhibition and/or aromatase inhibition, the Project 1-2 pathology evaluations and exome sequencing evaluations will also encompass estrogen receptor (ER) staining and give particular attention (in the exome sequencing datasets) to sequencing and copy number analyses of crucial proteins in the CDK4/6 pathway (CDKN2A, cyclin D isoforms, CDK4, SMAD3, SMAD4, MYC and RB1). Further, as our studies evolve, it is likely that other gene mutations will prove relevant in predicting response vs. resistance to PI3K pathway inhibitors. For all of these reasons, we will use a pan-genome approach (exome sequencing) to evaluate both PI3K-pathway mutations and mutations, generally, in our core group of 40 frozen LMS. Further, these pan-genomic studies will be leveraged by Project 2 IHC evaluations of certain PI3K and CDK4 pathway dysregulation mechanisms, such as PTEN and RB1 inactivation, which can be difficult to capture by genomic sequencing methods. The sarcoma database of the West German Cancer Center (Dr. Bauer) comprises clinical data of 350 patients with LMS. Paraffin-embedded tissues from more than 200 cases are available as validation sets to explore the prognostic relevance of these genomic aberrations.

Project 2: Biologic Profiles and Biomarkers for LMS with PI3K-pathway Dysregulation

Principal Investigator: Dr. van de Rijn

Project 2 aims to identify biomarkers for PI3K pathway oncogenic activation. This will be undertaken by gene expression profiling of the same core group of 40 frozen LMS for which systematic PI3K pathway profiles are obtained in Project 1. In addition, Project 2 will provide gene expression profiles to determine molecular subtypes for all LMS cell line and xenograft models in this research consortium. The hypothesis addressed in this aim is that oncogenic PI3K pathway oncogenic mutations, irrespective of whether in uterine or soft-tissue LMS, might result in characteristic gene expression alterations. Therefore, the translational goal is identification of robust biomarkers that serve as surrogates for PI3K pathway activation in LMS.

Molecular subtypes of LMS with PI3K-pathway activation. Gene expression profiling by RNAseq on a HiSeq Illumina platform will be performed in the same 40 snap-frozen LMS (20 soft-tissue, 20 uterine) for which genomic annotation was obtained in Project 1. Accordingly, the LMS will be assigned to molecular subtypes, based on the molecular profiling definitions that were identified and reported previously by Dr. van de Rijn’s group.^15,16 Here the aims are to: 1) confirm that the major LMS molecular subtypes are represented in this study (and, if not, additional LMS will be added to the study, to accomplish that goal); 2) determine whether particular molecular subtypes are enriched for PI3K-pathway genomic dysregulation (through correlations with the Project 1 analyses); and 3) determine whether PI3K-pathway dysregulation results in expression aberrations for particular genes that might then serve as biomarkers for PI3K-pathway activation. In addition to determining whether certain LMS molecular subtypes are enriched for PI3K-pathway perturbations, Project 2 will broadly assess histopathologic, immunophenotypic, and genetic correlates in LMS harboring PI3K/AKT/mTOR pathway mutations. The translational aims of these studies are to provide mechanistic biomarkers to identify LMS patients who will more likely benefit from therapies targeting the PI3K/AKT/mTOR pathway. To achieve these aims, it is important to develop straightforward algorithms that can be instituted in the clinical setting.  These may involve a combination of clinicopathologic features and diagnostic biomarkers (genomic and/or immunohistochemical).  To this end, we will review the gene expression, histopathologic and immunophenotypic features of the series of genotyped LMS, and examine for significant associations.  Histologic features will include degree of cellular differentiation and mitotic activity, and immunohistochemical markers will include smooth muscle differentiation markers (smooth muscle actin, caldesmon, desmin, calsequestrin 2, human muscle cofilin 2, myosin light chain kinase, and sarcolemmal membrane associated protein) and tumor biology markers (TP53, p16, cyclin D1 and PTEN).

Biomarkers for PI3K-pathway activation in LMS. The gene expression profiles in this project will be triangulated with the genomic sequence alterations (Project 1) and LMS functional genomics (Project 3) to identify priority candidates for biomarkers predictive of PI3K-pathway activation. We will prioritize candidate biomarkers for further validations based on biologic evidence supporting the biomarker as a logical indicator of PI3K-pathway activity, and based on ready availability of one or more reagents capable of assessing the biomarker in clinical specimens, preferably in FFPE materials. For example, we will favor biomarkers where substantial gain or loss of expression indicates PI3K-pathway activity, and for which well-qualified antibodies have already been developed for IHC detection of the protein in FFPE sections.

The aims of these Project 2 efforts are to identify universal biomarkers for PI3K-pathway dysregulation in LMS, with the aim of validating these in future correlative science studies in clinical trials of PI3K-pathway inhibitors. Ultimately, we hope to validate one or more biomarkers that will serve to identify those LMS patients who will benefit most from targeted PI3K-pathway inhibition. In this sense, the studies proposed here establish a model that we hope to extend to other driver genes and pathways in LMS.

Project 3: Functional Genomics of LMS and In Vitro Validations of New Therapies

Principal Investigator: Dr. Fletcher

Project 3 discovers essential genes and pathways that regulate LMS cell proliferation, cell survival, drug resistance, and confer drug synthetic lethalities. This is accomplished using state-of-the-art high-throughput functional genomic platforms developed jointly by Harvard University and Massachusetts Institute of Technology (MIT). The functional screens are innovative in using complementary open reading frame (ORF) and knock-down (lentiviral shRNA and lentiCRISPR-Cas9 libraries). These are unbiased genome-wide approaches, querying the functional relevance of ~16K genes in the ORF screens and ~18K genes in the shRNA screens.^17,18 Importantly, the LMS cell line cultures queried in these studies were created in Dr. Fletcher’s laboratory, and have been vetted in publications.¹⁹ Because Dr. Fletcher’s group created the LMS lines from primary cultures, functional genomic profiling will be performed at low passage numbers (< 20). This is a key advantage in our studies, given that functional genomic screens (and other in vitro studies) are often confounded by the well-known biologic drift that occurs in human cancer cells during repeated cell culture passages due to inevitable culture-related stresses. The studies are innovative because they use a rigorous, complementary approach in which the impact of both knock-down and induced expression for each gene is assessed in optimized early-passage LMS cultures. Essential genes will be identified in untreated LMS cells. In addition, we will particularly focus on identification of synthetic lethalities and resistance mechanisms to drugs targeting PI3K/AKT/mTOR and other crucial oncogenic pathways. This emphasis on existing drugs targeting known oncogenic pathways increases the likelihood of short-term clinical impact in these screens. In the synthetic lethality studies, we seek to maximize response to targeted therapies by identifying new combination therapy strategies. Previous clinical studies have demonstrated only modest activity for mTOR inhibition in LMS, even though a subset of these patients undoubtedly had LMS with PI3K/mTOR constitutive activation due to PTEN and TSC mutations. Hence we seek biologically selective combination therapies that maximize clinical benefit to PI3K/mTOR inhibition. Specific objectives are:

• To characterize essential genes and pathways regulating LMS proliferation and survival
• To identify LMS mechanisms of response and resistance to PI3K pathway inhibitors
• To identify synthetic lethal interactions, discovering additional targets that maximize response to PI3K pathway inhibition in LMS

Functional genomics in PI3K/mTOR-inhibited LMS: To identify genes for synthetic lethalities and drug resistance, the same genome-scale functional genetic technologies described above will be performed, but as modifier screens with IC50 doses of the PI3K/mTOR inhibitor BEZ325. Although BEZ235 is not a clinically useful drug, it is a very useful tool compound for in vitro studies, in which co-inhibition of class I PI3Ks and mTOR are readily achieved at highly-selective drug concentrations. The Fletcher lab has used BEZ235 for modifier functional genomic screens in GIST and Ewing sarcoma, and by using BEZ235 in the LMS studies, we’ll have the added value of identifying synthetic lethals and drug resistance mechanisms that are either unique to LMS, or which are cross-cutting among sarcomas. However, because BEZ235 is not an optimal compound for in vivo studies, in the subsequent preclinical validations we will use other PI3K and MTOR pathway inhibitors.

Prioritization of targets for follow-up. Prioritization of candidate targets will be based on triangulations of the rank lists from these functional studies vs. the genomic screens in Projects 1 and 2. Candidate hits from shRNA screens will be validated by CRISPR-Cas9 experiments, using at least 3 hit-specific RNA guides per gene. Comparisons of untreated and drug-treated experimental replicates will allow us to dissect the genetic correlates of drug response and resistance. The magnitude of the effects observed for each particular comparison will influence the prioritization of candidate hits for follow up. Importantly, the triangulation of datasets corresponding to the same inhibitor (BEZ235) used across different sarcoma types with variable dependencies will provide increased granularity, allowing us to discriminate between tumor type-related signatures and drug-related signatures; tumor type-related vulnerabilities will be favored. Priority rankings will also be influenced substantially by the evidence of active roles of particular candidates on tumor biology. The comprehensive genomic annotations of our LMS cell line models (genome and transcriptome, performed by Projects 1 and 2) will enable prioritization of targets and pathways credentialed by molecular aberrations. From these prioritization analyses, the top 10 candidates will be evaluated individually in single gene experiments. Initial validations will consist of confirmatory functional single-gene experiments, both gain and loss of function, to confirm the effects observed during the pooled screens, but now in an expanded panel of cell lines. Candidate oncogenes and synthetic lethals will be validated in CTG-based cell viability determinations using sets of 3 non-overlapping shRNAs and 3 independent CRISPR-Cas9 constructs targeting each candidate. Hits for drug-resistance will be validated by expressing 1-2 ORFs in drug-sensitive cell lines, followed by dose-response experiments. Possible off-target drug effects will be evaluated by using two or more structurally unrelated inhibitors against each target. We will also assess whether knockdown of putative resistance mechanisms restores sensitivity to the inhibitors in resistant sublines. We anticipate that 4-6 potential targets will be validated during the first 2 years of this grant, with the most promising candidates from this group being evaluated in substantial depth, and moved forward along the preclinical pipeline into Project 5 (in vivo LMS validations).

Project 4: Zebrafish model of LMS to enhance preclinical drug validations

Principal Investigator: Dr. Langenau

Transgenesis and gene inactivation strategies are now routinely employed to model sarcomas in zebrafish. For example, the Langenau group has created a RAS-driven model in rhabdomyosarcoma.²⁰ This model revealed heterogeneity within the sarcomas and identified cells that drive self-renewal and relapse.²¹ We also used this model for large-scale drug screens and to identify novel therapies for sarcoma,²² and we have optimized cell transplantation of fluorescent labeled sarcoma cells into zebrafish, to assess their metastatic potential.²³ 

The development of a zebrafish LMS model has the potential to revolutionize the study of this common subtype of sarcoma. As examples, this model will enable inexpensive, high-throughput in vivo evaluations of the therapeutics strategies implicated by the genetic studies in Projects 1 and 2, and the functional genomics studies in Project 3, including synthetic lethals with PI3K/MTOR-inhibition. Importantly, the zebrafish models will also enable evaluations of drug activity against LMS invasion and metastasis. Therefore, we expect the zebrafish LMS studies will play a key role in the Project 5 preclinical validations of novel therapeutics for metastatic inoperable LMS, vetting therapies that will ultimately be tested in new clinical trials.

Activation of the PI3K/AKT pathway through loss of PTEN drives LMS formation in mice,¹⁰ and PTEN and TP53 genomic inactivations are common in human LMS. Accordingly, we will model LMS in zebrafish using smooth muscle specific promoters including SM22alpha and a-smooth muscle actin (aSMA) to drive expression of myristoylated AKT in the developing zebrafish.²⁴ These constructs will be co-injected with a GFP fluorescent transgene into TP53-deficient zebrafish, allowing robust visualization of LMS development in live animals. Using similar mosaic transgenic approaches, we and others have successfully modeled rhabdomyosarcoma, liposarcoma, T-cell leukemia, melanoma, neuroblastoma, and other cancer types. In a second strategy, we will utilize CRE/LOX approaches to inactivate PTEN in the smooth muscle of developing zebrafish.^25,26 We will create a LOXed allele for PTENa and cross these conditional inactivation mutants into PTENa-/+; PTENb -/- fish. CRE will be specifically expressed in smooth muscle cells using either the SM22alpha or aSMA promoter. PTEN is duplicated in zebrafish and loss of 3 of 4 alleles is well tolerated,²⁷ with tumor development occurring late in life.²⁸ Once these lines become available, we will also breed them into TP53 deficient lines to assess genetic collaboration. Finally, we will assess roles for PI3K-pathway hyperactivation and p53-loss in regulating metastasis in the zebrafish LMS model. Specifically, we will assess tumors for engendering metastatic spread following injection of tumor cells into the vasculature of the eye and/or dorsal musculature as shown in Figure 4.

Project 5: In Vivo Validations of New Therapies

Principal Investigator: Dr. Bauer

Project 5 pursues preclinical validations in LMS xenografts in mice and in the Project 4 zebrafish LMS models. The broad aim of these studies is to finalize preclinical validations of pathway targeting strategies in the context of appropriate molecular subtypes of human LMS. The initial focus is on strategies to maximize LMS response to PI3K pathway targeting. With regard to drug selection for these preclinical in vivo studies, we have already validated two novel dual mTOR inhibitors, PKI587 and MLN0128, in vitro. Both drugs are being tested in clinical trials at the moment and both drugs have shown strong activity in those cell lines that had a proven dysregulation of the PI3K/mTOR pathway. Notably, both PKI587 and MLN0128 show similar activity as BEZ235, a lead compound that was used in our pilot studies, but at doses that can be achieved clinically.

LMS Xenografts. Project 5 uses LMS xenografts that this research consortium has already established from LMS cell lines. In addition, Dr. Bauer will establish primary xenografts from uncultured LMS cell suspensions, which are cryopreserved in his laboratory for 20 LMS, and also from new surgical biopsies of LMS, which are seen on a monthly basis on his sarcoma service at the W. German Tumor Center. Dr. Bauer’s group will also perform xenograft evaluations of small molecule inhibitors and/or gene knockdowns using methods well-validated in his laboratory,²⁹ whereas Dr. Langenau will perform the drug evaluations in zebrafish LMS. The consortium coinvestigators for Project 5 include Drs. Chibon and van de Rijn, who will perform genomic and gene expression profiling annotations for the LMS xenografts, to determine what LMS molecular subtypes they represent and which of these xenograft models harbor genomic PI3K or CDK4 pathway activation events.

In Vivo Preclinical Drug Validations in Mouse LMS. To evaluate the effect of synthetic lethal interactions and drug resistance mechanisms in vivo, we will use xenograft models for the different LMS molecular subtypes (see Projects 1 and 2) to evaluate growth, survival as well as interaction with the tumor microenvironment. The Bauer laboratory has more than 20 cryopreserved LMS surgical specimens for subcutaneous transplantation in adult athymic nude mice (NMRI nu/nu) BALB/c nude mice. In addition, we have validated that the LMS lines established by Dr. Fletcher’s group (Project 3) are tumorigenic in mice.  For initial response evaluation of selected compounds or combination strategies, two models per sarcoma subtype will be selected. Tumor-bearing mice will be treated with selected compounds to determine drug effects on in vivo tumor metabolism, proliferation, apoptosis, angiogenesis, and downstream pathways. Initially, dose response experiments will be performed (8 mice per dose) to determine the optimal dose of a specific inhibitor or combination. Tumor size will be measured twice weekly and growth curves of the different doses will be compared to select the optimal dose for the therapy experiments. For therapy experiments, 10 mice per group will be included and tumor growth curves will be compared with a control group. Two mice per group will be euthanized after 1 week of treatment to ensure availability of non-necrotic material, in case of high treatment efficacy. After 4 weeks of treatment the other mice will be euthanized and tumor will be dissected. Sections of the different tumors will be used for immunohistochemistry to determine the effect of therapy on downstream pathways (e.g. phospho-AKT and phospho-mTOR and novel PI3K-pathway biomarkers, depending on results of Projects 1 and 2); proliferation (Ki67); apoptosis (caspase3); and angiogenesis (VEGFA, VEGFR2, CD34, HIF-1a). Snap-frozen material will be immunoblotted to extend the IHC analyses: in our PI3K-pathway inhibitor evaluations, the immunoblotting analyses will assess drug effects on phospho-AKT, phospho-mTOR, and phospho-S6.

In Vivo Preclinical Drug Validations in Zebrafish LMS. Zebrafish chemical screens have now been widely used to identify drugs that affect development, regeneration and cancer.  The Langenau group has experience performing secondary chemical screens using transplantation and assessing drugs for efficacy in killing sarcomas and leukemias in vivo.  Here, Dr. Langenau treat LMS bearing fish to determine drug effects on in vivo tumor growth, proliferation, apoptosis, angiogenesis, and metastasis. Initial experiments will be performed using 6 animals per drug (or combination) at three log10 dilutions, with drugs being directly added to the water. Drug dosing will be based by available PK and PD measures found in mice and/or human (when possible).  Should drugs not be water-soluble, we have recently optimized IP injection protocols where drugs can be delivered 1x daily IP. Tumor size will be measured by mean tumor fluorescence twice weekly and growth curves of the different doses compared to control treated fish. For drugs with activity in these initial studies, we will repeat our analysis using three independent LMS tumors and expand our sample size to 20 fish per optimized dose treatment.  Animals will be sacrificed, and immunohistochemistry used to determine the effect of therapy on downstream pathways (e.g. phospho-AKT and phospho-mTOR and novel PI3K-pathway biomarkers, depending on results of Projects 1 and 2); proliferation (Ki67); apoptosis (caspase3); 3) and angiogenesis (requiring initial implantation of LMS tumors into Tg(flk-mCherry), capser-strain, rag2E450fs mutant fish). Snap-frozen material will be immunoblotted to extend the IHC analyses: in our PI3K-pathway inhibitor evaluations, the immunoblotting analyses will assess drug effects on phospho-AKT, phospho-mTOR, and phospho-S6.

Translation of LMS Laboratory Studies into New Clinical Trials

The overriding aims, for Projects 1-5, are to validate effective targeted therapies for LMS, and to identify molecular subtypes of LMS against which these therapies will have preferential activity. Accordingly, the success of this LMS consortium will be judged ultimately by effectiveness in moving preclinical concepts into clinical trials. Although the clinical trials themselves are beyond the budgetary scope of this grant, we will make every effort to expedite translation of promising preclinical concepts into clinical evaluations. To this end, Dr. Bauer, a practicing sarcoma medical oncologist, will have primary responsibility (among the LMS program PIs) for ensuring translation of preclinical concepts to formal testing in clinical trials. We therefore prioritize drug validations with compounds that are most advanced in clinical development. Further, the program PIs will invite medical, radiation and surgical oncologists with LMS expertise to an annual research workshop convened by this LMS consortium, so that we will have external advice and input regarding clinical opportunities emerging from the program. As examples of these interactions, we have already had many discussions with our clinical colleagues to develop consensus on the most feasible and promising clinical trial concepts from the present stage of the grant application. Because approximately 75% of LMS have demonstrable oncogenic CDK4-RB1 dysregulation, we have particularly considered combination therapies with mTOR and CDK4/6 inhibitors, in patients with uterine and soft tissue LMS which retain RB1 expression (RB1 inactivation being a well-known resistance mechanism to CDK4 inhibition). Further, those patients with uterine LMS, and either ER or PR expression, could receive an aromatase inhibitor, in addition to mTOR and CDK4/6 inhibitors. We reason that concurrent suppression of the PI3K-mTOR, CDK4/6, and estrogen-dependent pathways will more fully inhibit LMS growth and survival, compared to suppression of any of these pathways alone. To these ends, Projects 1 and 2 will develop a stream-lined NGS panel, using either exon capture or NGS panel, using either Ion Torrent (multiplex PCR), which will be integrated with PTEN IHC to identify patients with PI3K pathway dysregulation. We aim to conduct such trials as investigator-initiated trials that incorporate extensive correlative science (plasma sequencing, central tissue storage, optional biopsy in progressing patients). Completed trials of everolimus with imatinib in GIST,³⁰ and everolimus with IGF1Ri in advanced sarcomas³¹ provide precedence for acceptable tolerability of an mTORi combinatorial strategy in sarcoma. Furthermore, the combination of everolimus and exemestane was well-tolerated and was associated with improved progression-free survival (compared to exemestane, alone) in the Phase 3 Bolero-2 study in hormone-receptor positive breast cancer.^32-34 Therefore, clinical trials of this nature, in LMS, would parallel similar combination studies of mTORi, CDK4/6i, and aromatase inhibition which have shown activity in breast cancer. The aforementioned are examples of trial concepts that we believe already warrant clinical evaluation, and the goal of our studies is to identify strategies that will maximize the selectivity and effectiveness of interventions targeting these pathways. These considerations underscore the urgency of our consortium studies, which aim to identify optimal targets in molecularly-defined LMS subsets (Projects 1-3) coupled with rigorous preclinical validations (Projects 3-5).

The collaborations between our LMS program members have already enabled advances in LMS biology and preclinical studies.^19,35-37 Funding of our LMS translational research consortium will enable expansion of these existing productive collaborations. Further, our close clinical collaborators, Drs. George and Demetri, have a track record in leading single center phase I trials as well as multi-institutional rapidly-accruing LMS studies.^38-40 Therefore, the LMS consortium has the expertise and resources to develop clinical trials with robust correlative science studies, including serial tumor biopsies and other biomarkers to explain the clinical results in a manner that elucidates the molecular mechanisms of aberrant cell signaling and drug resistance.

Consortium Meetings

The project PIs, along with colleagues and post-doctoral fellows from their research programs will interact every-other-month by videoconference to provide updates on each project and to discuss challenges and opportunities. In addition, the PIs and their lab members who work on these projects will meet face-to-face twice yearly. The face-to-face meetings will include a CTOS meeting (Lisbon 2016) for the PIs and their group members, and a larger meeting in 2017, to which we will invite participants (particularly, LMS clinical experts) from outside the immediate consortium, to provide feedback on which preclinical opportunities are considered most attractive and realistic for exploration in clinical trials.

Copyright © 2016 Liddy Shriver Sarcoma Initiative.