|Who gets it?
What are the symptoms?
|How do we detect it now?
How is it currently treated?
|What are the current challenges?|
Autoantibodies in Breast Cancer
There is strong preclinical evidence that cancer, including breast cancer, undergoes immune surveillance. This continual monitoring, by both the innate and the adaptive immune systems, recognizes changes in protein expression, mutation, folding, glycosylation, and degradation. Local immune responses to tumor antigens are amplified in draining lymph nodes, and then enter the systemic circulation. The antibody response to tumor antigens, such as p53 protein, are robust, stable, easily detected in serum, may exist in greater concentrations than their cognate antigens, and are potential highly specific biomarkers for cancer. However, antibodies have limited sensitivities as single analytes, and differences in protein purification and assay characteristics have limited their clinical application. For example, p53 autoantibodies in the sera are highly specific for cancer patients, but are only detected in the sera of 10-20% of patients with breast cancer. This points to a need to identify new autoantigens that would enable multiplexed testing for cancer. Early studies have shown that using a panel of autoantibodies improves both sensitivity and specificity of detection. Recent advances in proteomic technologies have the potential for rapid identification of immune response signatures for breast cancer diagnosis and monitoring.
The overall goal of this project is to identify autoantibody biomarkers in sera that can be readily used for the early detection of cancers. We are producing protein microarrays using our method and screening this against sera from both cancer patients and healthy controls, in order to find autoantibodies that are specific in patients.
We are particularly interested in breast cancers. In a recent study, we have successfully applied NAPPA to detect novel tumor antigen-specific AAb in the sera of breast cancer patients (Journal of Proteomics Research. 2011 10:85). We used age-, sex- and location-matched sera obtained from both screening and diagnostic mammography clinics. All sera were collected prior to therapy, in the same clinical settings, under standardized collection protocols. Our study followed three phases: Phase 1, eliminate uninformative antigens; Phase 2, identify breast cancer specific AAb; Phase 3, blinded validation of AAb.
Sera from women undergoing routine screening mammography were selected for a test set for Phase 1 (n=53 cases, 53 controls, Cohort 1), which were screened against nearly 5000 antigens. We removed all but the top 761 antigens from an initial set of 4988 based on ROC methodology. In Phase 2 we identified candidate AAb by printing the selected 761 cDNAs in duplicate on single arrays (Fig. 2). These arrays were screened with a separate training set of sera from invasive breast cancer patients (n=51) and from benign breast disease patients (n=39). From these data, 119 antigens were selected as potential biomarkers for further analysis (p<0.05; FDR <13%). In Phase 3 we tested the 119-antigen panel in a blinded independent validation of 51 cases and 38 controls. We tested each antigen using the AUC methods with a focus on highly specific markers. We found evidence (p < 0.05) for 28 of the 119. This represents a statistically significantly higher number of confirmatory findings than would be expected by chance alone (p = 0.0041). For these 28 antigens, we used the threshold that yielded approximately 95% specificity on the training set. In the blinded results, these antigens ranged from 11 - 42% sensitivity with 55 - 100% specificity.
We also used the combined training and validation sets to construct a classifier of patient status using Breiman’s random forests algorithm with 200 trees and 3 random features. We measured the average leave-one-out cross validation performance of the classifier across five random seeds. The average sensitivity of the classifier was 80.0% and the average specificity was 61.3%, with an AUC of 0.756 (range, 0.724-0.789 at the 95% confidence interval). The receiver operating characteristic curve for the classifier using the leave-one-out predictions is shown in Figure 3. When considering the ROC curves one needs to keep in mind how challenging breast cancer is for finding any plasma signature of cancer.
In combined analysis of both the training and validation sets, ATP6AP1 was the most significant individual autoantigen. Using a recombinant protein ELISA with independent sera sera obtained prior to treatment from 102 patients with stage I-III breast cancer and compared with 77 healthy controls divided into two sets. AAb to ATP6AP1 were significantly higher amongst the breast cancer patients (p<0.0002, p<0.0034).
Triple Negative Breast Cancer
In recognition of the heterogeneity of breast cancers, our ongoing efforts focus on identifying subtype-specific AAb in Her2+ breast cancers, triple negative breast cancers (TNBC, ER-PR-HER2-), and cancers associated with increased breast density on mammography. Breast cancers can be classified into different cancer types depending on the expression of ER, PR, and Her2. Hormone-receptor negative cancers (ER-PR-) and Her2+ cancers are more rapidly proliferative, have increased mortality, and occur more frequently in younger women where mammographic screening is less reliable. In addition, these subsets of breast cancers are more prevalent in minority populations. The basal-like TNBC is more frequent in premenopausal African-American women, and Her2+ cancers are more frequent in premenopausal Hispanic women. The 3N and Her2+ breast cancers are frequently missed by mammographic screening due to their rapid proliferative rates and occurrence in younger women. There is also an unmet need for cancers that have limited detection by mammography due to increased breast density. We anticipate that there will be distinct AAb biomarkers for the different breast cancer tumor subtypes.
Tamoxifen Resistance in Breast cancer
Investigators: Laura Gonzalez, Ph.D.
Collaborators: Barbara Pockaj, M.D. (Mayo Clinic), Michael Barret, Ph.D. and Heather Cunliffe, Ph.D. (TGen)
In addition to studying and identifying biomarkers, we are also studying the changes in cells that cause breast cancer’s resistance to treatment. Resistance to tamoxifen in breast cancer patients is a serious therapeutic problem, however it is unknown what underlying mechanisms cause this resistance. To study this problem, we derived a series of cultured breast cancer cells that were either highly sensitive or naturally resistant to tamoxifen to study the factors that lead to drug resistance. Using gene expression microarrays, we identified genes that differentially responded to tamoxifen in resistant vs. sensitive cells. Furthermore, in a screen of over 500 human kinases, we identified 30 kinases whose overexpression resulted in therapy resistance in previously sensitive cells. One kinase, HSPB8 which is also involved in blocking autophagy, was found to confer tamoxifen resistance and predicted poor clinical outcome in one cohort of patients. These results may help overcome the tamoxifen resistance in breast cancer and improve breast cancer prognosis.
More recently, we have started working towards understanding the role of Androgen Receptor (AR) in breast cancer. Though much is known about the role of another hormone receptor, Estrogen Recent (ER), in breast cancer and drugs have been developed that successfully treat ER+ breast tumors, relatively little is known about AR. Furthermore, cancers that don’t respond to drugs against ER have far more AR than those that do respond to drugs, indicating that AR could be a therapeutic target. This study will use breast cancer cells to determine whether AR is a biomarker for drug resistance, and if so, whether blocking AR will effectively treat tumors that don’t respond to drugs targeting ER.
Culturing Human Breast Tumors and Identification of early functional co-drivers of p53-mutant breast cancers
One main challenge when studying cancers is that much of the research is being done on tumor cell lines. These are cells that have been taken from a tumor and grown on a plastic plate, outside of the body, in nutrient supplemented media often for years at a time (in the case of HeLa cells, they have been grown in labs for over 50 years!). Although these cell lines have been “better than nothing” and helpful to develop a general understanding of the general characteristics of cancer cells, they do have considerable drawbacks. Most tumor cells do not grow in culture, which means that scientists have not been able to study a large number of different tumors or from individual patients. What tumor cells that do grow are a selected sub-population of the tumor that may or may not be representative of the actual tumor, and furthermore, this population can mutate and change over time as they are grown on plates over many years and no longer represent the original tumor. To overcome these issues, Dr. Schlegel has developed a method to grow a primary tumor taken directly from a patient using helper cells, non-tumor mouse cells plated on the plastic dish that contribute growth factors and nutrients to the tumor cells, and an inhibitor of a key kinase (publication). Drs. Anderson and Gonzalez have used this method to grow primary breast tumor and normal tissue that they have received from Dr. Barbara Pockaj at the Mayo Clinic (see Figure). These cells will be used for three main purposes, each of which will translate the basic research of the lab into clinical benefit to the patient.
- These cells will be used to identify the key “co-drivers” of cancer that can help develop specific, targeted cancer therapies.
- These cells will be used for providing personalized treatment to the patient. We will expose these cells to a variety of anticancer drugs and identify which drugs kill the cells “in culture.” This information will then help the clinicians to choose a treatment option that will be more likely to kill the tumor cells in the patient.
- We plan on using these cells to develop personalized cancer vaccines.
Gonzalez-Malerva L, Park J, Zou L, Hu Y, Moradpour Z, Pearlberg J, Sawyer J, Stevens H, Harlow E, Labaer J. (2011) High-throughput ectopic expression screen for tamoxifen resistance identifies an atypical kinase that blocks autophagy. Proc Natl Acad Sci Jan 13. [Epub ahead of print]. Abstract
Anderson KS, Sibani S, Wallstrom G, Qiu J, Mendoza EA, Raphael J, Hainsworth E, Montor WR, Wong J, Park JG, Lokko N, Logvinenko T, Ramachandran N, Godwin AK, Marks J, Engstrom P, Labaer J. (2011) Protein Microarray Signature of Autoantibody Biomarkers for the Early Detection of Breast Cancer. J Proteome Res Jan 7;10(1):85-96. Epub 2010 Nov 23. Abstract
Anderson KS, Ramachandran N, Wong J, Raphael JV, Hainsworth E, Demirkan G, Cramer D, Aronzon D, Hodi FS, Harris L, Logvinenko T, LaBaer J. (2008) Application of protein microarrays for multiplexed detection of antibodies to tumor antigens in breast cancer. J Proteome Res. 7(4):1490-9. Epub 2008 Feb 27. PMID: 18311903
Witt AE, Hines LM, Collins NL, Hu Y, Gunawardane RN, Moreira D, Raphael J, Jepson D, Koundinya M, Rolfs A, Taron B, Isakoff SJ, Brugge JS, LaBaer J. Functional proteomics approach to investigate the biological activities of cDNAs implicated in breast cancer. J Proteome Res. 2006 Mar;5(3):599-610. PMID: 16512675