A Targeted Therapy for Wound-like Breast Cancers

Institution: Stanford University
Investigator(s): Howard Chang, M.D, Ph.D. -
Award Cycle: 2006 (Cycle 12) Grant #: 12IB-0063 Award: $232,934
Award Type: IDEA
Research Priorities
Detection, Prognosis and Treatment>Innovative Treatment Modalities: search for a cure



Initial Award Abstract (2006)
Each time a person cuts a finger, a remarkable choreography of genes and cells unfolds. Cells that are normally dormant are now invigorated to divide rapidly, remodel the matrix around themselves, migrate across tissue planes, and send forth chemical signals to recruit new blood vessels-all in an effort to close up the wound. These dramatic processes in wound healing may also constitute the ideal genetic tools for cancer metastasis. We discovered that some breast cancers exhibit wound-like features and can be distinguished by a gene expression signature--a specific pattern of 512 genes. In a survey of 295 early stage breast cancers, approximately 40% of tumors exhibited the wound signature, and these patients have a three-fold increased risk of death. The wound signature identified these high risk tumors independently of known risk factors for poor outcome. Therefore, therapies that specifically target cells exhibiting the wound signature are needed for patients with wound-like breast tumors. Recently, we discovered that mutations in a novel combination of two cancer promoting genes, MYC and CSN5, are responsible for wound signature activation in breast tumors. The collaboration between MYC and CSN5 depended on abnormal protein turnover, a process that may be blocked by the FDA-approved drug bortezomib (Velcade®). It is the first in a new class of medicines called proteasome inhibitors. This drug is currently used to treat multiple myeloma in patients who have received at least one treatment that has not helped.

Our specific aims are, first, to test the ability of bortezomib to kill human breast cancer cells with the wound signature. A panel of human breast cancer cells exhibiting different levels of activity of the wound signature will be tested for sensitivity to bortezomib during growth in laboratory cultivation. The results will demonstrate whether the wound signature can predict the sensitivity of human breast cancer cells to bortezomib. Secondly, we will test the ability of bortezomib to inhibit the growth of mouse breast tumors exhibiting the wound signature. Introduction of MYC and CSN5 into mouse mammary glands creates breast tissues mimicking human breast cancers with the wound signature; we will test whether bortezomib treatment can cure or inhibit the growth of such mammary tumors. These studies will use physiological and molecularly precise models of a high-risk breast cancer to test the utility of a therapy

Although genetic studies have detailed increasingly specific subtypes of human breast cancers, there has been no specific treatment for these newly classified patients. Our study of bortezomib is one of the first to test a targeted therapy based on the understanding of the gene expression program of human breast cancers.


Final Report (2007)
Breast cancers are highly heterogeneous. Some patients with breast cancer experience rapid progression to metastatic disease while others have indolent cancers that allow more opportunities for treatment. We and others have previously identified a particular genetic signature, called the “wound signature” that identifies high risk, metastasis-prone breast cancers. These tumors are able to engage in a genetically program normally reserved for wound healing, using the body's normal capacity for regeneration for cancer growth.

The goal of the research project was to identify therapies that might specifically target breast cancer cells exhibiting the wound signature. Our first aim was to elucidate the role of the wound signature in predicting sensitivity to bortezomib, an FDA-approved drug that is the first in a new class of medicines called proteasome inhibitors. This drug is currently used to treat multiple myeloma in patients. Proteosomes are cellular complexes that break down proteins. For our studies we adapted a bioinformatic method, termed a “gene module map”, to identify metabolic and signaling pathways that are coordinately induced or repressed specifically in human breast tumors that activate the wound signature. We found that genes encoding subunit of the proteasome are systematically “activated” in wound-like tumors. Importantly, proteasome blockade by Bortezomib, or its analog MG-132, specifically killed tumor cells exhibiting the wound signature, but not just any rapidly dividing cells, such as those engineered to express the oncogene E2F3 without activating the wound signature. This predictive relationship was observed in cell lines engineered to induce the wound signature and also in a panel of existing human breast cancer cell lines. The relationship between wound signature and proteasome activation was also preserved in six independent datasets of human breast cancers that spanned primary, local disease to metastatic disease. These results identify a possible computational method to identify targeted therapies for human cancers, and pinpoint bortezomib as apotential specific therapy for wound-like breast cancers.

In this project’s second aim, we found that introduction of RAS cooperated with MYC to rapidly induce invasive breast cancer growth by activating the wound signature. We could therefore use such tumors to test the hypothesis that interference with the pathways leading to proteasome activation blocks tumor growth. At this same time, we observed that a point mutation in the CSN5 JAMM protein sequence motif, changing one amino acid (aspartate 151 ? asparagine 151, or D151N), abrogated CSN5 activity and acted as a “dominant negative” with respect to wound signature activation. CSN5 is the catalytic component of the COP9 signalosome, a complex important for maintaining the function of E3 ubiquitin ligases that activate substrates for the proteasome. We found that introduction of CSN5(D151N) strongly blocked tumor growth elicited by RAS and MYC. These results showed that amplification of CSN5, an activator of the ubiquitin proteasome pathway, acts as an authentic human breast oncogene leading to cancers exhibiting the wound signature. They further show that blockade of the ubiquitin proteasome pathway by CSN5 inhibition can abrogate tumor growth, identifying this pathway as a therapeutic target in human breast cancers.

Future studies should determine the consequences of pharmacologic blockade of proteasome and contrast that with genetic blockade of CSN5, and also determine the mechanism of proteasome-induced cell death in this “wound signature” subtype of breast cancer.


Symposium Abstract (2007)
The main goal of this research is to discover treatment strategies for breast cancers that exhibit behaviors like a wound. Each time a person cuts her finger, a remarkable choreography of genes and cells unfolds. Cells that are normally dormant are now invigorated to divide rapidly, remodel the surrounding matrix, migrate across tissue planes, and send forth chemical signals to recruit new blood vessels-all in an effort to close up the wound. These dramatic processes in wound healing may also constitute the ideal genetic tools for cancer metastasis.

We discovered that some breast cancers exhibit wound-like features and can be distinguished by a gene expression signature--a specific pattern of 512 genes. In a survey of 295 early stage breast cancers, approximately 40% of tumors exhibited the wound signature, and these patients have a three-fold increased risk of death. The wound signature identified these high risk tumors independently of known risk factors for poor outcome. Therefore, therapies that specifically target cells exhibiting the wound signature are needed for patients with wound-like breast tumors. Recently, we discovered that mutations in a novel combination of two cancer promoting genes, MYC and CSN5, are responsible for wound signature activation in breast tumors. The collaboration between MYC and CSN5 depended on abnormal protein turnover, a process that may be blocked by the FDA-approved drug bortezomib.


Symposium Abstract (2010)

Each time a person cuts a finger, a remarkable choreography of genes and cells unfolds. Cells that are normally dormant are now invigorated to divide rapidly, remodel the matrix around themselves, migrate across tissue planes, and send forth chemical signals to recruit new blood vessels-all in an effort to close up the wound. These dramatic processes in wound healing may also constitute the ideal genetic tools for cancer metastasis. We discovered that some breast cancers exhibit wound-like features and can be distinguished by a gene expression signature--a specific pattern of several hundred genes. Patients with breast tumors that harbors the "wound signature" have a three-fold increased risk of death. Therefore, therapies that specifically target cells exhibiting the wound signature are needed for patients with wound-like breast tumors. I will present progress in finding the genetic basis of the "wound signature" and leads in specific therapies based on these insights.