Regular exercise has been shown to lower one's chances for developing breast cancer, but little is understood about the mechanisms that mediate this effect. To gauge the impact of moderate exercise on the expression of genes in newly-diagnosed breast cancer patients, Jennifer Ligibel, MD, of the Dana-Farber Cancer Institute, and Melinda Irwin of Yale University initiated the Pre-Operative Health and Body Study (PreHAB).
“Several studies have been performed to assess the impact of exercise on tumor growth in animal models; however, thus far little data has been gathered regarding the effect of exercise on human tumor tissue,” Ligibel noted.
The PreHAB study was undertaken to evaluate how exercise affects gene regulation patterns in human tumor samples. To do this, tumor samples were taken from initial biopsies and stored. Then, once the complete tumors were removed, both samples were simultaneously subjected to differential gene expression analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis.
In addition to the comparison of the gene expression levels from the tumors, biomarker analyses were also done for fasting blood samples. Regarding the effect of exercise on the patients' weights, Ligibel clarified, “During the intervention period between enrollment and surgical tumor removal, no significant weight loss was observed.”
In a previous animal study, Thompson, et al, assessed plasma biomarkers in a Sprague-Dawley rat 1-methyl-1-nitrosourea model of breast cancer (Cancer Prev Res (Phila) 2010;3(11):1484-1492). In their study, the researchers evaluated 20 different plasma biomarker levels from three biomarker families: glucose metabolism/homeostasis; sex hormones; and inflammation and cytokines. Both ELISA and BioPlex assays were utilized for biomarker assessment.
Subjects were randomized to one of three groups: sedentary; motorized running wheel; or non-motorized free running wheel. Most biomarkers utilized were affected by exercise; however, the fasting glucose, estradiol, and C-reactive protein levels were not significantly altered.
The results showed exercise clearly provided a protective effect, as only 68 percent of the subjects in the motorized wheel group(n=22) developed cancer, as compared to 79 percent of the free wheel group (n=33) and 98 percent of the sedentary group (n=55).
In Ligibel's window of opportunity study, 49 recently diagnosed stage I-III breast cancer patients who were previously inactive were enrolled in PreHAB from either Yale University or the Dana-Farber Cancer Institute prior to tumor removal.
The patients were randomized to either the exercise group (27 participants) or the control group (22 participants). The exercise group participated in an aerobic and strength training regimen while the control group took part in a mind-body meditation program in between enrollment and tumor removal.
Tumor samples collected at the initial biopsies and at surgical removal were reviewed prior to dissection by a breast pathologist. Areas of 10 percent cellularity or greater were included in samples that were subjected to capture RNA sequencing. This sequencing was done using an RNA access platform for analysis of the transcription coding regions.
The mean age of the participants at baseline was 52.6 years, and the mean body mass index was 30.2 kg/m2. The mean exercise rate at baseline was 49 minutes per week, while the mean time between enrollment and surgical tumor removal was 4.2 weeks. The exercise arm participants dramatically increased their exercise rate during the study period as compared to those in the control arm (203 min/week vs. 23 min/week; p<0.0001).
“Only samples from 32 patients had enough genetic material extracted from their initial biopsy and removed tumors for sequencing; 16 from the exercise arm and 16 from the control arm,” Ligibel noted.
The transcriptomic analyses performed on the patients' initial biopsy and removed tumor samples showed which genes were expressed at those time points. The average read depth was 25 million reads per sample, and of these, approximately 79 percent were mapped to exonic regions. No batch effects or read bias was shown by principal component analysis. Unsupervised clustering revealed that initial biopsy and removed tumor samples were clustered by patient. “Clustering by intervention was not observed,” Ligibel added.
The DEseq2 package (Genome Biology; doi: 10.1186/s13059-014-0550-8) utilized for differential gene expression analysis revealed there was no single gene that showed a significant change in expression between the exercise and control group when comparing baseline and surgical specimens.
KEGG Pathways were analyzed using the Generally Applicable Gene-Set Enrichment for Pathway Analysis (GAGE) package (BMC Bioinformatics 2009;10:161). Of the 214 pathways evaluated, 18 showed upregulation in their excised tumor samples relative to levels present in initial biopsies for participants in the exercise arm of the study. Interestingly, no such upregulation was noted for participants in the mind-body arm (q <0.1).
The most highly ranked upregulated pathway in the exercise arm was for cytokine-cytokine interactions. In this pathway, the interleukin-6 (IL6), the chemokine (C-C motif) ligand 3 (CCL3), and other chemokine-associated genes were among those upregulated. “The pathways that were upregulated in those patients participating in the exercise program were generally focused on immunological and inflammatory responses,” Ligibel commented.
Participants in the exercise arm of the study were also observed to have some pathways downregulated. For these patients, 13 unique pathways involved in DNA replication, RNA transport, and cell cycle processes were downregulated in their post-surgical samples relative to the expression levels present in their initial biopsies(q<0.1).
In addition to transcriptomic analyses, the patients in this study also had blood-based biomarker analyses. “Patients in this study had fasting blood samples drawn and then subjected to ELISA analyses for various metabolic and inflammatory biomarkers,” explained Ligibel.
Among the metabolic biomarkers evaluated were insulin, leptin, and insulin-like growth factor (IGF-1). The inflammatory biomarkers measured included tumor necrosis factor-α (TNFα), C-reactive protein, and the interleukins (e.g., IL6).
“Our results clearly showed that the exercise, even though it was only for a roughly 4-week period for most participants, did have an impact on a number of potentially relevant genes and biomarkers,” Ligibel said. “This study marks the first time data has been obtained for assessing the impact of exercise upon gene expression patterns in human breast cancer tumor tissue.
“The data showed that exercise may directly affect breast cancer tissue, at a minimum would seem to warrant further investigation,” she continued.
Regarding the future direction of the project, Ligibel noted, “We hope to become more familiar with the underlying biological mechanisms by which exercise may lower the risk of developing or dying from breast cancer.
“To better assess this, larger studies will need to be undertaken, as well as a more thorough analysis of which tumor compartments were responsible for the gene expression changes observed,” Ligibel concluded.
Richard Simoneaux is a contributing writer.