Grant Funds the Study of Three Therapeutic Targets in Leiomyosarcoma

An ESUN Announcement

The LeioMyoSarcoma Direct Research Foundation and the Liddy Shriver Sarcoma Initiative have come together to fund a $150,000 grant for promising leiomyosarcoma research at Stanford University.

Leiomyosarcoma (LMS) is a rare and aggressive cancer. Dr. Matt van de Rijn explains the difficulty of treating LMS: "The challenge that leiomyosarcoma poses for the treating physician is that surgery remains the major avenue of therapy for this tumor. This means that when a tumor can not be completely resected, or when a tumor has already metastasized, that there are no targeted therapies that specifically will attack the leiomyosarcoma cells. Instead, post-surgery treatment relies on chemotherapy and radiation therapy, both of which can have significant side effects and also can have variable success."

In this two-year study, Dr. van de Rijn and Santosh Gupta, PhD aim to determine whether three proteins that are expressed at various levels in LMS tumors can be used as targets for novel therapies. The investigators hope that this research will eventually lead to clinical trials of targeted treatments for LMS patients.

Targeting Three Different Proteins

Two of the proteins (ROR2 and CD47) involved in this study are expressed on the surface of LMS cells. Therefore, they can be directly attacked in a variety of ways.

With ROR2, the investigators will focus on antibodies that will render the protein ineffective. By inhibiting this protein they expect to be able to inhibit tumor growth.

The researchers will work with the CD47 protein. They have previously shown in an LMS mouse model that inhibiting the CD47 protein allows immune cells called macrophages to more effectively destroy LMS cells.

The third protein involved in the study, CSF1, is not expressed on the cell surface but instead is secreted by the tumor cells into the surrounding environment. After secretion by the tumor cells, the CSF1 protein attracts macrophages and these macrophages can help the tumor to grow by helping to provide new blood vessels for the tumor. The investigators will look at various ways of inhibiting the CSF1 protein in this setting.

Dr. van de Rijn hopes that it will be possible to combine therapies that target different LMS proteins in order to maximize their effectiveness: "Perhaps most importantly, we will investigate the interplay between therapies directed at each of these proteins. Previous studies have shown that such a collaboration between two therapies can have a dramatic improvement in outcome in tumor model systems."

Meet the Lead Investigator

Matt van de Rijn MD, PhD: As a surgical pathologist I have always been interested in sarcomas. These tumors are rare, and many different subtypes of sarcomas exist. There is always a need for additional markers to improve the ability of surgical pathologists to make the correct diagnosis. It was from this background that I started applying gene expression profiling to sarcomas about a decade ago. After our first paper appeared, I was approached by Iqbal Ahmed who suggested to me that I would specifically focus on leiomyosarcomas. That was an excellent suggestion, as at that time I was orienting myself on which sarcoma I wanted to go into in more detail, and I felt that leiomyosarcomas were relatively understudied within the group of soft tissue tumors, while they constituted a considerable percentage of these tumors. Moreover, no targeted therapies existed at that point of time, and I thought there was an unmet need in leiomyosarcoma. With the help of Iqbal Ahmed and later with the help of Suzanne and Ed Kurtz and Sharon Anderson I started to focus on leiomyosarcomas. My role in this study will be to supervise the work on leiomyosarcoma as we now pursue three potential therapeutic targets simultaneously in the lab.

Funding for this Study

The funding for this $150,000 grant is a result of an ongoing partnership between the LeioMyoSarcoma Direct Research Foundation (LMSdr) and the Liddy Shriver Sarcoma Initiative (LSSI) in supporting leiomyosarcoma research. LMSdr and LSSI have come together to fund three earlier leiomyosarcoma grants, and we are proud to once again be partnering with one another.

The mission of LMSdr, which is based in Tulsa, Oklahoma, is to assure the development of the means to cure and control the disease known as leiomyosarcoma and to improve the quality of life for those with this disease until the cure is found. LSSI supports this mission.

LSSI acknowledges a generous donation from Dr. Laura Somerville to support their $50,000 component of this grant.

V10N4 ESUN Copyright © 2013 Liddy Shriver Sarcoma Initiative.

Developing Novel Therapeutic Options for Leiomyosarcoma

By Matt van de Rijn MD, PhD
and Santosh Gupta, PhD

Introduction

Leiomyosarcoma (LMS) is a rare cancer for which no targeted therapies currently exist. The mainstay of treatment for LMS is surgery, which is sometimes used in combination with local radiation therapy. Adjuvant chemotherapy has had limited success in LMS and the development of novel therapeutic approaches represents a significant unmet medical need.

Over the past decade, we have worked towards developing novel therapeutic targets for LMS and have determined the important role that macrophages play in this disease (1-3, 6, 8, 9). We also described the role of ROR2, a relatively poorly studied tyrosine kinase receptor in the prognosis of LMS (10). In parallel, we defined molecular subsets of LMS tumors (4, 7). The purpose of this research is to support the further study of the therapeutic targets identified in the previous years with a direct goal to lead to one or more clinical trials. We will develop 3 promising targets in LMS and examine the interplay between them:

1. CSF1 is secreted by approximately 30% of LMS tumors and attracts macrophages (see Sidebar 2) that can help tumor growth through stimulation of tumor angiogenesis (new blood vessel formation).
2. CD47 is expressed on essentially all LMS tumors (and many other cancers) and prevents tumor-associated macrophages from destroying the tumor cells. We have shown that inhibition of CD47 limits primary LMS tumor growth and dramatically decreases tumor metastasis in a mouse model.
3. ROR2 is a member of the family of tyrosine kinase receptors and is expressed on a subset of LMS tumors. It can be targeted through inhibitory antibodies, small molecule inhibitors and potentially antibody drug conjugates.

In the past decades single agent therapy has frequently yielded disappointing results as tumors develop resistance to the single agent used. While we will develop each of the three targets described above individually, we will also investigate the interplay between these targets to determine the manner in which they can be combined to decrease tumor growth in complimentary ways. We expect that this will lead to a markedly increased therapeutic effect.

Background and Previous work

In a series of studies, we have shown the association of tumor-associated macrophages (TAMs: macrophages that are present within a tumor mass) with poor clinical outcome in LMS patients. Initially, we showed that the number of TAMs in LMS tumors was associated with poor clinical outcome in extra-uterine LMS (2). We next evaluated the role of CSF1, a protein that attracts and stimulates macrophages. CSF1 is secreted by LMS tumor cells in a significant subset of patients, and we showed that by combining TAM counts with levels of CSF1, we could predict clinical outcome not only in extra-uterine LMS, but also in uterine LMS (3). Importantly, we could confirm these observations in two independent patient groups (6, 8). These studies were consistent with findings in other tumors that showed that macrophages that are attracted to a carcinoma help the tumor grow by aiding in the generation of an active blood circulation within the tumor, thereby providing cancer cells with nutrients and oxygen. We showed that a similar mechanism is active in LMS. Macrophages exert their pro-tumor effect in LMS by increasing the number of blood vessels in LMS tumors, resulting in a nutrient-rich environment for LMS tumors to grow (9). A logical conclusion from these observations is that inhibiting the CSF1 molecule could have a beneficial effect on the patient by preventing macrophages from helping to generate these blood vessels in LMS.

Drs. Matt van de Rijn and Rob West jointly operate a Research Laboratory; within this collaborative group Matt focuses on sarcoma research (with an emphasis on LMS and GIST research) while Rob’s work mostly involves the study of carcinoma. In the past decade we have used novel technologies to identify new diagnostic markers, prognostic markers, and potential therapeutic targets. We initially used gene microarray technology to study the level of expression of tens of thousands mRNAs, which are molecules translated from DNA that tell cells which proteins to make. The ability to study these mRNAs on a large scale was revolutionized by the invention of gene microarray technology by Dr. Pat Brown at Stanford University. For the first time, scientists were able to study in a single experiment the level of not just one, but 20,000 mRNAs. More recently the gene array technology has been replaced with high-throughput “next generation sequencing” (NGS). We developed a technology that allows us to use NGS on formalin-fixed, paraffin-embedded tissue, enabling us to study mRNA levels in vast numbers of archival clinical specimens for which no frozen material is available (5). The work on LMS in our lab has progressed steadily over the past decade, as is shown in the timeline below (Figure 1).

Research Aims and Experimental Plan

We will study three distinct therapeutic approaches to LMS. One may wonder why we would not focus all our attention on the best, most promising, target rather than studying these three at the same time. There are several reasons for this tactic. First, while all three are hopeful leads it is impossible to predict which ones will perform in the end. Second, when working on the three approaches in parallel we can be much more efficient in the design of our experiments, such that each experiment can give answers for not just one but for two or even all three targets. Most importantly, there is a strong possibility that when these targets are attacked simultaneously, there may be an additive effect such that tumor response exceeds the one seen if they would be individually attacked one at a time.

Aim 1: To develop CSF1 as therapeutic target in LMS

In the first aim of the proposal, we will study the efficacy of an anti-human CSF1 antibodies or small molecule inhibitors on inhibiting tumor growth in vivo using xenograft models of human LMS. We have developed three xenograft models in mice using two separate patient-derived LMS cell lines that were established by the laboratory of Dr. Jonathan Fletcher. The first cell line (LMS04, derived from a uterine LMS) grows very well as a subcutaneous tumor mass in NSG mice and also spontaneously gives rise to lung metastases. Human xenograft model systems that spontaneously metastasize are rare and we believe that the study of this model will greatly add to the existing knowledge about LMS. This is especially important since most patients with LMS do not succumb to their primary tumor, but rather to distant organ metastases. The other LMS cell lines (LMS05, derived from a soft-tissue LMS of the thigh and LMS06) also grow well as a subcutaneous tumor mass but do not metastasize to the lung.

In parallel with the in vivo studies discussed above, we will use these inhibitors to determine whether there is evidence for an autocrine stimulatory loop between CSF1 produced by the tumor cells and the CSF1 receptor present on the same tumor cells in LMS. Such a pro-growth, CSF1-CSF1R autocrine loop has been demonstrated in renal cell carcinoma, but no studies on LMS have been reported. This also would have a significant impact on considerations for a potential clinical trial of CSF1 inhibitors for LMS patients.

Aim 2: To further develop CD47 as a therapeutic target in LMS

A complementary approach will look at the ability to “change the behavior” of TAMs such that they transform from “bad” macrophages (that promote tumor growth) into “good” macrophages (that phagocytose, or “eat up” tumor cells, see Sidebar 2). Macrophages can have either “good” or “bad” behavior towards tumor growth – the so-called “bad” macrophages help tumors to grow by increasing their blood supply, as mentioned above. However, in some situations macrophages can be “good” by phagocytosing tumor cells. Recent studies from the Weissman lab showed that macrophages can be coaxed into becoming “good” by exposing tumors to an antibody that allows them to eat tumor cells.

Tumor cells have devised a mechanism to protect themselves from being phagocytosed by macrophages through the expression of CD47 on the surface of the tumor cells. The CD47 protein binds to a protein on the surface of macrophages called SIRPa. This CD47-SIRPa interaction prevents macrophages from eating tumor cells, leading CD47 to be coined as a tumor cell’s “don’t eat me” signal. In our recent article in the Proceedings of the National Academy of Sciences (12), we showed that an antibody against the CD47 protein, which we determined to be expressed on LMS cells, could result in a dramatic increase in the ability of macrophages to phagocytose LMS tumor cells, leading to striking decreases in LMS tumor growth and metastasis (Figure 2).

Most studies on CD47 had been performed in the laboratory of our collaborator Irv Weissman at Stanford University on a wide range of tumors, but LMS was not evaluated. These prior studies have used a murine monoclonal antibody against CD47. This antibody cannot be used in therapy in humans as patients immediately would develop immunity against any immunoglobulin from a foreign species. However, the Weissman laboratory developed a “humanized” version of the anti-CD47 molecule (suitable for human clinical use) and it is this version that we used in our LMS study in PNAS. Our paper appeared back-to-back with a publication from the Weissman laboratory (11) in which a range of other tumors (including breast cancer and glioblastoma multiforme) were studied to determine their response to anti-CD47 treatment. Our LMS study was unique in that it directly addressed the effect of anti-CD47 treatment on tumor metastasis.

The availability of the humanized version of the anti-CD47 antibody allows us to prepare for clinical trials in the near future, and a major effort will be dedicated to have LMS patients represented in these trials. Before clinical trials can start, however, it is crucial we address a number of questions about the interplay between the inhibition of “bad” macrophages by anti-CSF1 antibodies and the ability to change “bad” macrophages to “good” macrophages through CD47 inhibition.

Perhaps the most important question is whether LMS patients are only candidates for treatment with anti-CD47 antibodies if they have high numbers of TAMs, or whether all LMS patients might benefit from anti-CD47 therapy. The second aim of this proposal, therefore, is to use the in vitro and in vivo assays developed in our lab to study combinations of anti-CSF1 and anti-CD47 therapies to model treatment regiments for LMS patients with high or low levels of CSF1. In addition we will generate LMS cell lines in which we over-express CSF1 or decrease CFS1 expression through shRNA. These will be used in studies where we determine the effect of these variable levels of CFS1 in the tumor cells on the ability of macrophages to exert phagocytotic behavior. These studies will be crucial to the success of clinical trials by allowing us to predict response to these novel treatments based on the characteristics of a patient’s LMS tumor. Recently we collaborated on a study that showed that the interaction between CD47 and SIRPa can be disrupted not only by CD47 antibodies but also by synthetically generated protein fragments of SIRPa (13). This novel approach will also be included in our studies of the variables affecting LMS tumor growth.

Aim 3: To further develop ROR2 as a therapeutic target in LMS

We have shown that in LMS (and GIST) expression of the tyrosine kinase receptor ROR2 is associated with increased invasion of tumor in a Matrigel-based invasion assay that it resulted in increased tumor growth of LMS cell in xenografts and conversely that knockdown of ROR2 led to decreased tumor growth. Others had previously shown that ROR2 had similar effects in other cell lines (renal cancer cell lines for example). However, we have access to an antibody that recognizes ROR2 in formalin fixed, paraffin embedded material and this allowed us to use a tissue microarray with >140 samples of LMS and associated clinical outcome to determine that ROR2 expression on LMS and GIST tumors correlates with poor outcome.

Recently we have completed a study on the expression of ROR2 in normal fetal and adult human tissue (manuscript in preparation, Figure 3). While ROR2 is highly expressed in fetal tissues it is only rarely expressed in adult tissue. These results indicate that ROR2 is a promising therapeutic target. We will pursue several approaches to develop reagents that target ROR2 and these include the screening of a library of small molecules for ROR2-specific inhibitors. This approach will be performed with financial support from the “SPARK” fund at Stanford. The funding from the Shriver and LMSdr foundations will be used for two independent approaches.

First, we will have a company generate a custom monoclonal antibody that inhibits activation of ROR2. We already have established a Western blot-based test to determine phosphorylation (indicative of activation) of ROR2. This test will be used to select a monoclonal antibody that can prevent activation of ROR2. This antibody will be tested in vitro for its effect on LMS05 cell line growth and also in xenografts for its effect on tumor growth. Unfortunately, ROR2 is not expressed by LMS04, so far the only cell line that shows metastatic behavior in our lab.

Second, we will determine whether ROR2-positive LMS05 cells are more likely to be phagocytosed by macrophages when they are exposed to both anti-ROR2 and anti-CD47 antibodies. Studies by others have shown that adding an antibody directed against a cell surface antigen in lymphoma markedly increases the effect of CD47 inhibition and demonstrating this in our LMS xenograft model would indicate that this is a hopeful approach to pursue in patients.

Impact Statement

If successful, this work can lead to the development of one or more targeted therapies in LMS, a tumor for which currently no such an opportunity exists. The project involving CD47 will benefit greatly from the enormous amount of work that is currently being performed on this protein in several other tumor types in the Weissman laboratory at Stanford. The funding from the Liddy Shriver Sarcoma Initiative and LMSdr will allow us to ensure that the necessary experiments to incorporate LMS in a possible CD47–based therapy can be performed and will in addition allow us to pursue the ROR2 and CSF1 approaches simultaneously.

V10N4 ESUN Copyright © 2013 Liddy Shriver Sarcoma Initiative.