“Single cell-derived clonal analysis of desmoid tumors to investigate tumor-stroma interactions.” Year 1 of 3

The tumor stroma includes many non-neoplastic cells and the extracellular matrix, making up the microenvironment in which neoplastic cells grow. The non- neoplastic cells are predominantly stromal fibroblasts, sometimes called “tumor associated fibroblasts”. These cells are important for the maintenance and remodeling of the microenvironment, providing the appropriate conditions for neoplastic cell growth and invasion. The role of stromal fibroblasts in promoting tumor invasion has recently been highlighted in a number of cancers such as breast cancer, gastric carcinoma, non-small cell lung carcinoma, and colorectal cancer. This interaction between the stromal fibroblasts and the neoplastic cells can occur indirectly through secreted paracrine factors, or directly by physical cell-cell contact. Desmoid tumors (DTs) are characterized by proliferating and invading fibroblastic cells embedded in depositions of extracellular matrix driven by mutations that activate β-catenin signaling. We hypothesize that desmoid tumor cells interact with the stromal fibroblasts and that this interaction is responsible for maintaining the neoplastic phenotype of tumor cells. To test our hypothesis, we propose to answer the following questions:

Aim 1: How do the neoplastic and stromal populations of desmoid tumors differ?

Aim 2: Do desmoid tumor cells interact with the surrounding stromal cells by paracrine signals?

Aim 3: Do desmoid tumor cells interact with the surrounding stromal cells by direct cell-cell contact?

By elucidating the composition of desmoid tumors, and how the tumor cells interact with the normal stromal cells, we can greatly enhance our knowledge of how they grow. DTs are invasive tumors characterized by abundant extracellular matrix deposition. Therefore, it is very likely that tumor- stromal interactions are important in the way they grow and infiltrate the surrounding tissue. By isolating and characterizing individual cell populations from the tumor mass, we can gain a better understanding of what distinguishes the mutant cells from the normal cells. Our method of separating individual cells will also allow us to study variability between patient samples at the cellular level, without the confounding factor of having a different proportion of normal cells in each tumor sample. Since DTs are known to be heterogeneous, this method is key at understanding the source of this heterogeneity, which can influence treatment decisions. Elucidating what signaling pathways are involved in the process of tumor-stroma interaction will also allow us to potentially identify novel therapeutic targets that we can use to inhibit tumor growth and infiltration. We anticipate that the established single cell-derived clones, and the knowledge gained from this project of desmoid tumor biology, will be the basis for future experiments that build on our knowledge of desmoid tumor heterogeneity and tumor-stromal interactions.

______________

In neoplasia, tumor cells interact with the normal stromal cells, such as the surrounding fibroblasts. This makes studying desmoid tumors difficult as both tumor and stromal cells display a mesenchymal phenotype and there are no well-established markers to distinguish the two populations. To elucidate the composition of desmoid tumors, we isolated and expanded single cells derived from patient desmoid tumor samples. Sequencing of individual clones derived from the same patient sample revealed that desmoid tumors do consist of mutant sub-populations carrying the beta-catenin activating mutation, and normal sub-populations lacking the mutation. To improve our isolation methodology, we performed a high throughput flow cytometry surface antigen screen to study the expression pattern of over 300 surface markers on mutant and non-mutant cells. From this analysis, we identified CD142 as a unique positive marker for the mutant population. Having isolated mutant and non-mutant cell populations, we next looked into the expression of secreted factors that could potentially play a role in cell-cell communication. From this study, we found that CTHRC1 is a factor secreted by the mutant cells. Gain- and loss-of-function experiments showed that CTHRC1 can influence the proliferation rate of desmoid tumor primary cultures. Studying desmoid tumors at the clonal level will enhance our understanding of the intratumoral heterogeneity of these tumors. Identifying unique surface markers allows for the rapid isolation of mutant and non-mutant subpopulations while minimizing cell divisions. In addition to studying tumor-stroma communication, measuring tumor composition may also be useful for ongoing drug screening efforts and as a potential post-treatment readout.

LAY VERSION OF ABSTRACT- Single cell-derived clonal analysis of desmoid tumors to investigate tumor-stroma interactions

Desmoid tumors are proliferations of relatively benign appearing fibroblasts. Despite their histologic bland appearance, a significant subset of these tumors recurs aggressively and requires often debilitating surgery. Currently there are no molecular markers that predict the behavior of desmoid tumors.

The aims:

AIM 1: To generate an extensive molecular dataset covering gene expression profiling and gene mutation analysis of desmoid tumors and scar samples.

AIM 2: To identify gene changes in expression level (for mRNA and non-coding RNAs) and gene mutations that correlate with risk for recurrence.

AIM 3: To use the same approach to identify genetic markers that can distinguish scar tissue from recurrent desmoid.

AIM 4: To explore the molecular heterogeneity of desmoid tumors based on the analysis of multiple regions sampled from a single tumor.

AIM 5: To search for low frequency CTNNB1/APC mutations in “wild-type” tumors.

The purpose of this study will be to perform a very broad molecular search for markers that can be used to address two clinically highly relevant questions. Using a carefully selected group of cases with known clinical outcome, we will use stranded RNA-Seq data (for mutation and expression level analysis) to generate a dataset that will characterize desmoid tumors in the broadest sense possible. First, this dataset will be searched for markers that can distinguish desmoid tumors that can be followed by “watchful waiting” at the time of their initial biopsy from tumors that require aggressive surgical therapy. When desmoid tumors recur, the distinction between recurrent tumor and scar tissue from prior surgery can be very difficult by histology, and immunohistochemistry for beta-catenin is insufficient. In the second approach that will use the dataset mentioned above, we will compare RNA-Seq findings on scar tissue with those seen in desmoid tumors to identify markers that distinguish scar from desmoid in an effort to find a novel immunohistochemistry marker to discern these two entities and that can be used in surgical pathology practice.

We also plan to perform deep targeted sequencing of CTNNB1 and APC genes in samples with no detectable mutations in Sanger sequencing/RNA-Seq data. According to the recent findings of Crago et al. (Genes Chromosomes Cancer. 2015 Oct;54(10):606-15), many of desmoid tumors initially considered as wild-type appear to carry low frequency mutations in CTNNB1/APC genes that can be detected using next generation sequencing. Our set of samples comprises relatively high portion of wild-type cases (27%) therefore we aim to further investigate this issue in order to better characterize our samples.

This work will build on an already existing large dataset in the van de Rijn laboratory and will benefit from the experience that we have generated in our laboratory for the use of archival formalin-fixed, paraffin-embedded tissue in high throughput studies using next generation sequencing. The project greatly benefits from the collaboration with Dr. Kristen Ganjoo at Stanford University Hospital and from the collaboration with Dr. Raffi Avedian, orthopedic surgeon at Stanford. We have also started a new collaboration with Dr. Justin Cates and Dr. Thomas Stricker from Vanderbilt University, and with Dr. Chiara Colombo from Milan, Italy. We have previously performed quantitative measurement of the expression levels for all know human protein-encoding genes in 9 desmoid tumor samples, 4 scars and 42 non-desmoid fibroblastic lesions. The number of scar samples that we have analyzed was too low but the preliminary results were encouraging. Our current dataset includes 29 primary, 20 recurrent desmoid tumors and 15 scars from 26 patients.

LAY VERSION OF ABSTRACT- Next generation sequencing approach to desmoid tumors

Desmoid tumors are proliferations of relatively benign appearing fibroblasts. Despite their histologic bland appearance, a significant subset of these tumors recurs aggressively and requires often debilitating surgery. Currently there are no molecular markers that predict the behavior of desmoid tumors.

The aims:

AIM 1: To generate an extensive molecular dataset covering gene expression profiling and gene mutation analysis of desmoid tumors and scar samples.

AIM 2: To identify gene changes in expression level (for mRNA and non-coding RNAs) and gene mutations that correlate with risk for recurrence.

AIM 3: To use the same approach to identify genetic markers that can distinguish scar tissue from recurrent desmoid.

AIM 4: To explore the molecular heterogeneity of desmoid tumors based on the analysis of multiple regions sampled from a single tumor.

AIM 5: To search for low frequency CTNNB1/APC mutations in “wild-type” tumors.

The purpose of this study will be to perform a very broad molecular search for markers that can be used to address two clinically highly relevant questions. Using a carefully selected group of cases with known clinical outcome, we will use stranded RNA-Seq data (for mutation and expression level analysis) to generate a dataset that will characterize desmoid tumors in the broadest sense possible. First, this dataset will be searched for markers that can distinguish desmoid tumors that can be followed by “watchful waiting” at the time of their initial biopsy from tumors that require aggressive surgical therapy. When desmoid tumors recur, the distinction between recurrent tumor and scar tissue from prior surgery can be very difficult by histology, and immunohistochemistry for beta-catenin is insufficient. In the second approach that will use the dataset mentioned above, we will compare RNA-Seq findings on scar tissue with those seen in desmoid tumors to identify markers that distinguish scar from desmoid in an effort to find a novel immunohistochemistry marker to discern these two entities and that can be used in surgical pathology practice.

We also plan to perform deep targeted sequencing of CTNNB1 and APC genes in samples with no detectable mutations in Sanger sequencing/RNA-Seq data. According to the recent findings of Crago et al. (Genes Chromosomes Cancer. 2015 Oct;54(10):606-15), many of desmoid tumors initially considered as wild-type appear to carry low frequency mutations in CTNNB1/APC genes that can be detected using next generation sequencing. Our set of samples comprises a relatively high portion of wild-type cases (27%) therefore we aim to further investigate this issue in order to better characterize our samples.

This work will build on an already existing large dataset in the van de Rijn laboratory and will benefit from the experience that we have generated in our laboratory for the use of archival formalin-fixed, paraffin-embedded tissue in high throughput studies using next generation sequencing. The project greatly benefits from the collaboration with Dr. Kristen Ganjoo at Stanford University Hospital and from the collaboration with Dr. Raffi Avedian, orthopedic surgeon at Stanford. We have also started a new collaboration with Dr. Justin Cates and Dr. Thomas Stricker from Vanderbilt University, and with Dr. Chiara Colombo from Milan, Italy. We have previously performed quantitative measurement of the expression levels for all known human protein-encoding genes in 9 desmoid tumor samples, 4 scars and 42 non-desmoid fibroblastic lesions. The number of scar samples that we have analyzed was too low but the preliminary results were encouraging. Our current dataset includes 29 primary, 20 recurrent desmoid tumors and 15 scars from 26 patients.

LAY VERSION OF ABSTRACT- Next generation sequencing approach to desmoid tumors

This project is to develop the first validated PRO tool in desmoid tumors. The PRO tool will be developed and validated in partnership between MSKCC and Quintiles, a leading PRO company. Desmoid tumor (DT/DF) is a sarcoma of fibroblastic origin that afflicts children, adolescents and adults and has a median age of onset of 30 years. Desmoid tumors can occur in any anatomic location and threaten vital organs and limbs which often results in significant morbidity from organ failure, mutilating surgeries (and amputations), loss of limb function, pain and death. Surgery is associated with high rates of recurrences (~40%). In advanced disease, systemic therapies range from anti-estrogens to cytotoxic chemotherapies with variable response rates however no systemic therapies have been approved since pivotal trials evaluating survival or quality of life has never been conducted. In the last few years, a deeper biological understanding of this disease has led to promising drugs such as sorafenib, mTOR and Notch inhibitors that are currently in clinical trials. These and other drugs cause both tumor reduction and improvement in quality of life. Tumor reduction is measured by standard images and by established criteria (RECIST or WHO). However, the improvement in quality of life described by patients is lost as there are no validated tools to measure this drug effect. The classical regulatory (FDA) pathway to approving new drugs in cancer is demonstration of improvement in overall survival. Surrogates of overall survival include progression free survival and durable tumor responses. An emerging endpoint for regulatory approval is patient reported outcomes (PRO) tools. More recently, the FDA approved JakafiTM, a drug to treat myelodysplastic syndrome based on a PRO tool developed in MDS. We hypothesize that development of a PRO tool in desmoid tumors, will facilitate a new regulatory endpoint that can be used in all future clinical trials and will be a great investment of resources to advance the field for researchers worldwide.”

LAY VERSION OF ABSTRACT- Development of a patient reported outcome (PRO) tool in desmoid tumors

Desmoid tumors (DT) are locally invasive lesions that are difficult to treat using conventional therapies. While there is some success with chemotherapy in people who have recurrence of their tumors after surgery and/or radiation, much of the time the results from chemotherapy are temporary, and regrowth of tumor is seen. One approach to treatment is to develop a combination of medications that target different aspects of tumor growth, but have few side effects. Since desmoids are locally invasive but do not metastasize to other locations in the body, it is not necessary to eliminate every tumor cell, but instead to stabilize or shrink the tumor. We assembled a consortium of researchers from different backgrounds, and with complementary skills, whose ultimate goal is to develop a treatment to therapeutically target tumor cells in desmoid patients without causing serious side effects.

To achieve this goal, over the past two years, we screened drugs in the laboratory on cell lines from desmoid tumor patients to see which agents specifically target tumor cell viability and tested a first cohort of drugs in mice that develop desmoid tumors, identifying several agents already in use or close to being in use for patient care that could be repurposed for use in desmoid tumor treatment.

Over the remaining course of the proposal we plan to:

1. Complete the testing of drugs identified in the screen in a mouse model that is engineered to develop desmoids.
2. The agents identified will then be rescreening for drug combinations with synergistic effects.
3. A pilot multi-drug regimen will be tested in mice.

In the past two years we screened eight cell lines using the Maybridge HitFinderTM Collection; DIVERSetTM Collection (Chembridge); LOPAC (Sigma); BIOMOL collection; Prestwick collection (Prestwick Chemical); Seminatural compounds (AnalytiCon Discovery); and the NIH Clinical Collection (BioFocusDPI). This screening identified 45 compounds that inhibited cell proliferation in desmoid tumors by greater than 50% while not affecting normal fibroblast proliferation, and that also are either in use in patients or that have the potential to be rapidly developed for patient care. We then tested these agents using a collection of desmoid tumor cell cultures of different genetic etiologies, and based on this, as well as how available the selected agents are for clinical use today, prioritized 18 for testing in mice that develop desmoid tumors. We completed testing of nine agents, and found that five: Dasatinib (an oral Bcr-Abl tyrosine kinase inhibitor); A multi-target PI3K inhibitor; an FAK inhibitor; Letrozole; and a gamma secretase inhibitor. An additional nine drugs are currently being tested in the mice.

In the next year, we will undertake a rescreening of the five drugs that were effective in mice, by undertaking a sensitization test, in which we first treat the cells with the positively identified agents, and then rescreening them using the libraries. Agents that have a synergistic or additive effect will show lower cell viability than either agent alone. This approach will rapidly identify a multi-drug combination. We will then test the drug combinations in mice. Depending on the number of agents identified, we will either test all combinations, or if the number if too large, we will prioritize which to test based on a pharmacokinetic and then undertake analysis prioritizing agents that can be used in patients at the present time.

This data will identify new treatment regiments, which can rapidly be brought to patient care. In addition, it will identify new pathways activated by beta-catenin and new agents not yet in use for patient care that target desmoid tumors, both of which can be developed in future work which might ultimately identify even more effective therapeutic approaches.

LAY VERSION OF ABSTRACT- Collaboration for a Cure: Identifying new therapeutic targets for desmoid tumors

Desmoid-type fibromatosis (DF) is a rare connective tissue disease, characterized by the presence of one or more benign non metastasizing myofibroblastic neoplasms with a high local recurrence rate. DF can be diagnosed at any age in both genders but it is more common in young women.

A minority of DF is inheritable and is associated with mutations of the Adenomatous Polyposis Coli (APC) gene, while most of the DF arises sporadically. On the molecular level, 50-80% of the sporadic DF cases have somatic mutations in the codons encoding serine 41 and threonine 45 in exon 3 of the β-catenin gene, CTNNB1. These sites are required for β-catenin phosphorylation and subsequent degradation. Consequently alterations in these sites lead to the Wnt pathway alteration with β-catenin accumulation and its translocation to the nucleus.

Nuclear β-catenin is detected in almost 90% of the desmoid cells and it is considered a positive marker of DF cells once malignant cell transformation is excluded. In addition to this marker, we recently demonstrated that GSK-3β, also involved in Wnt pathway regulation, translocates in the nucleus of DF cells as well, colocalizing and interacting with β-catenin. This event occurs regardless of the presence of CTNNB1 gene mutations suggesting that the β-catenin/GSK-3β interaction and the nuclear translocation can be unrelated to the genetic alterations of CTNNB1. This led us to formulate the hypothesis that the alteration Wnt/β-catenin signaling pathway in desmoid tumor is modulated by the interaction with microenvironmental factors.

The aim of this project is the identification and the characterization of the microenvironment factors responsible for the proliferation and aggressiveness of DF cells. Since the deregulation of the inflammatory process can lead to aberrant fibroblast activation and accumulation of ECM proteins with subsequent tissue fibrosis that can evolve in fibrotic disease and tumor development, we will focus our attention on the inflammatory and growth factors and their interaction with the Wnt/β-catenin pathway. For this purpose cells derived from desmoid tumor and control cell samples will be incubated with different cytokines, growth factors and hormones. Their effect on the growth rate and aggressiveness will be evaluated by immunofluorescence using specific antibodies for cell cycle markers, for the receptors of the microenvironmental factors, for the proteinases degrading the extracellular matrix, for cysteine proteases and for the markers of the basal membrane, lysosomes and vesicles.

We will use a custom Ion AmpliSeq RNA panel to analyze the gene expression of DF cells under treatment, in particular focusing on genes of the Wnt pathway and of the other correlated pathways. Modulating the microenvironment we expect to identify potential factors that promote the DF tumor growth and aggressiveness. This is a critical milestone in defining the potential targets for a therapy aimed at reducing the proliferation, local aggressiveness and recurrence rate of DF tumors.

This study is part of a long term multidisciplinary research program that is ongoing since 2011 and it aims to improve the knowledge of the Desmoid-type fibromatosis and to identify a viable treatment. To date the study has led to the identification of a second potential positive marker of desmoid tumor cells, the nuclear translocation of GSK-3β.