“Lipidomic Profiling of Energetics-Associated Cancer Models in Mice”

Project Leader
Dorothy D. Sears, PhD, associate professor of medicine, Department of Medicine, University of California, San Diego

Collaborators
Oswald Quehenberger, PhD, Department of Medicine, University of California, San Diego
Loki Natarajan, PhD, Department of Family & Preventative Medicine, University of California, San Diego

Collaborating TREC Centers
UPenn: Lewis Chodosh, MD, PhD, professor cancer biology, Department of Cancer Biology
WUSTL: Kelle Moley, MD, James P. Crane Professor in Obstetrics and Gynecology, Division of Reproductive Endocrinology; Professor, Cell Biology and Physiology; Vice Chair and Director, Basic Science Research in Obstetrics and Gynecology

Project Summary
Recent advances in broad-range detection and quantitation of lipid subtypes, “lipidomics,” offer exciting discovery pathways regarding the relationship between lipids and disease. Cancer is associated with systemic and tissue-localized inflammation and metabolic alterations and, as a result, the normal distribution of various lipid forms is perturbed in fatty acid-specific ways. Lipids can modulate inflammatory signaling, apoptosis, alter gene transcription, and serve as second messengers. Eicosanoids are metabolites of polyunsaturated fatty acids and are potent mediators of inflammation. Fatty acids and their sphingolipid and eicosanoid derivatives are important modifier variables for cancer risk. Changes in energy balance, e.g. obesity, caloric restriction, and exercise, alter specific lipid levels and expression of lipid metabolism genes and can influence cancer risk. Lysophosphatidic acid is a plasma marker of ovarian cancer and promotes tumor growth. Dietary lipids modulate risk of both breast and prostate cancer, both of which are being studied in TREC@WUSTL mouse model projects.

Our primary aim is to define plasma, tumor and tumor tissue environment lipid signatures from mouse models relating energetics and cancer. Secondarily, we will characterize metabolic pathways and phenotypic characteristics that correlate with the cancer-associated lipid signatures. We will elucidate and characterize lipidomic signatures in plasma, tumor and tumor environments (mammary adipose tissue, prostate) from three mouse models of cancer and specifically query the effects of energetic alterations on these signatures. Overlapping indices of the signatures will be of particular interest as possible metabolic markers of and modifiable therapeutic targets for energetics-influenced cancers.

Findings from these studies will leverage each TREC center’s investigator’s main projects by broadening the scope of disease markers to include lipids and their metabolites. The findings will also inform future lipid profiling analysis studies in samples from breast cancer survivors, obese women at risk for breast cancer, and men at risk for prostate cancer due to in utero exposure to obesity.

“Transgenerational Effect of Maternal Diet on Methylation of Cancer Related Genes”

Project Leader
Jorge Chavarro, MD, ScM, ScD, assistant professor, Department of Nutrition and Department of Epidemiology, Harvard School of Public Health

Collaborators
Myriam Afeiche, PhD, MPH, research fellow, Department of Nutrition, Harvard School of Public Health
Andrea Baccarelli, MD, MPH, PhD, Professor of Environmental Health, Columbia University

Collaborating TREC Centers
UCSD: Dorothy D. Sears, PhD, associate professor of medicine, Department of Medicine
WUSTL: Kelle Moley, MD, James P. Crane Professor in Obstetrics and Gynecology, Division of Reproductive Endocrinology; Professor, Cell Biology and Physiology; Vice Chair and Director, Basic Science Research in Obstetrics and Gynecology

Project Summary
Using a rodent model, project 1 at TREC@WUSTL includes among its aims to identify whether exposing female mice throughout pregnancy to a high fat diet (HFD), is associated to epigenetic alterations of imprinted genes in prostate tissue at 8, 16, 26, 34 and 53 weeks postnatal. Loss of imprinting (LOI) in Igf2, one of the genes under investigation, has been implicated in breast and prostate carcinogenesis. WUSTL’s project 1 will generate data on whether maternal diet can influence LOI in this gene throughout the lifetime. However, whether this mechanism applies to other common cancers and the extent to which the findings from this rodent model are translatable to humans is unknown. In this cross-center developmental project, we are examining the generalizability of this molecular mechanism to breast cancer using a rodent model. In addition, we will evaluate the translatability of the findings of these two rodent models to humans using data from a prenatal cohort. Specifically, we will 1) examine whether the effects of HFD on the methylation of Igf2 and H19 in the mouse prostate are also observed in mouse white blood cells. (WUSTL center); 2) examine whether HFD is related to methylation of Igf2 and H19 in the mouse mammary fat pad and white blood cells (UCSD center); and 3) examine whether 2nd trimester maternal fat intake (total fat intake and intake of major types of fat) is related to methylation of Igf2 and H19 in cord blood white blood cells (Harvard center). We anticipate that the results from this developmental project will serve as preliminary data for future grant applications of translational projects evaluating the role of material diet on breast and prostate cancer in offspring by leveraging on the animal model available at UCSD and TREC@WUSTL together with the large prospective cohorts available at Harvard.

“Inflammation Markers and Body Mass Index in Breast Cancer”

Project Leader
Timothy R. Rebbeck, PhD, professor of epidemiology, Department of Biostatistics and Epidemiology, University of Pennsylvania

Collaborator
Charnita Zeigler-Johnson, PhD, MPH, research assistant professor, Thomas Jefferson University

Collaborating TREC Centers
Harvard: Michelle Holmes, DrPH, MD, MPH, associate professor, Department of Epidemiology
UCSD: Elena Martinez, PhD, professor, Family & Preventive Medicine
WUSTL: Sarah Gehlert, PhD, E. Desmond Lee Professor of Racial and Ethnic Diversity, George Warren Brown School of Social Work

Project Summary
There has been a growing interest in the tumor microenvironment and its association with breast cancer severity. Under chronic inflammation, T lymphocytes, macrophages and neutrophils contribute to increased cell proliferation and inhibition of cell death, potentially advancing cancer. The presence of these immune cells in tumor samples may indicate aggressive tumors that are likely to metastasize. Obesity is a risk factor for postmenopausal breast cancer and reduced response to endocrine therapy that has been associated with chronic inflammation. The majority of breast cancer tumors are estrogen receptorpositive (ER+). Estrogen secreted from adipose tissue promotes ER+ tumor growth, making obesity a prominent factor in ER+ tumorigenesis. African Americans face a heightened risk of obesity and poor breast cancer outcomes, but the relationship between these disparities is not fully understood and it is unclear which mechanisms drive this association. Because it is unlikely that environmental differences account for all of the disparity observed in breast cancer outcomes, biological mechanisms that place groups at high risk for aggressive disease must be considered. Higher rates of obesity which increase underlying inflammation may be a mechanism that contributes to ER+ breast cancer disparities.

We are studying the contribution of obesity to inflammation in the tumor microenvironment that may explain the racial disparity in ER+ breast cancer. We will examine inflammation biomarkers in ER+ breast cancer tumors across a range of body mass indices (BMIs) in postmenopausal African American and Caucasian women. Our study hypotheses are: 1) higher BMI increases the presence of lymphocytes, macrophages and neutrophils within the ER+ breast tumor microenvironment; and 2) race modifies the association of inflammation biomarkers with BMI in ER+ breast cancer tumors. We plan to link pathologic examinations of inflammation in tumor specimens to clinical data.

The long-term objective of this work is to be able to suggest ways to improve treatments for women at high risk for poor outcomes by modifying inflammatory markers in the tumor microenvironment via dietary and/or exercise interventions. If patients can be better classified in terms of risk for metastasis, treatments can be tailored to prevent over- or undertreatment of patients via adjuvant chemotherapy.

“The Use of Obesity Profiles in the Prediction of Breast Cancer”

Project Leader
Bernard Rosner, PhD, professor of medicine (biostatistics), Department of Medicine, Harvard Medical School

Collaborating TREC Centers
UCSD: Loki Natarajan, PhD, Department of Family & Preventative Medicine
WUSTL: Adetunji Toriola, MD, PhD, assistant professor, Division of Public Health Sciences, Department of Surgery

Project Summary
We have previously developed a risk prediction model for breast cancer based on Nurses’ Health Study (NHS) data from 1980-2000+ using information on 22 breast cancer risk factors. This model has been extended within the NHS to include types of benign breast disease. One of the risk factors for breast cancer is obesity. However, the effect of this risk factor is complex with protective effects of obesity before menopause, and deleterious effects of obesity after menopause. Furthermore, effects of obesity after menopause differ according to whether a woman is or is not a hormone replacement therapy (HRT) user. Usually, the effect of obesity after menopause is represented by weight gain since age 18 = current weight minus weight at age 18. However, an unanswered question is what the effect is of recent weight change (e.g., weight change in the past 5 years). A second issue concerns the effect of HRT on the breast cancer-obesity association. Generally, obesity is inversely related to breast cancer risk while a woman is on HRT. An unanswered question is whether this is equally true if a woman is on estrogen alone vs. estrogen + progesterone. A third issue is what is the minimum number of weight measures for accurate risk prediction? We currently estimate average BMI before and after menopause and use these variables as predictors in our model. Can we get by with BMI at age 18 and current BMI for purposes of risk prediction, perhaps adding BMI after important sentinel events (e.g., after specific births for parous women)? If so, then this would make the model more clinically useful in for example a mammography service.

The above model has excellent goodness of fit properties both in terms of discrimination and calibration when developed on a subset of the NHS population and tested in a different NHS subset. However, it has never been tested on an external dataset, particularly among a non-Caucasian population. Hence, this study will address the following aims:

  1. Build an enhanced risk prediction model for breast cancer based on 3,398 NHS cases from 1980-2006 including type of benign breast disease and mammographic density.
  2. Use this model to investigate the effect of: recent weight gain; weight gain among HRT users (while using both estrogen alone and estrogen & progesterone); and using only BMI at age 18 and current weight as predictors rather than lifetime weight profile
  3. Test this model on an external dataset obtained from the Joanne Knight Breast Health Center, Washington University School of Medicine, where 25,000 women have been screened annually and followed from 2009-2012 during which approximately 1,400 invasive incident breast cancers have been confirmed. Thus, a prospective analysis similar to the NHS can be performed, since all the NHS risk factors are collected during routine screenings. In addition, each mammogram is assessed for mammographic density using the BI-RADS approach.