Dose-dependent effect of aerobic exercise on inflammatory biomarkers in a randomized controlled trial of women at high risk of breast cancer.

The WISER (Women In Steady Exercise Research) Sister study was a randomized controlled trial of premenopausal women at high risk for breast cancer that enrolled participants from 2008–2012. Detailed information on study recruitment, randomization, the intervention, and main results (improvements in body composition, hormone levels, breast density, adiponectin and leptin) have been previously reported. 24 , 25 , 26 , 27 Briefly, eligible participants were considered high risk if they possessed any of the following: 1) confirmed BRCA1/2 mutation, 2) ≥18% lifetime risk according to the Gail or Claus prediction models, or 3) documented deleterious mutation in a relative that conferred a >25% risk on the participant. Eligibility criteria also included: > 18 years of age; normal menstrual cycles; not currently smoking; body mass index (BMI) 18–50 kg/m 2 and weight stable; no history of: fibroids, endometriosis, or polycystic ovary syndrome; no recent use of hormonal contraception; no contraindications for exercise training; controlled hypertension; and sedentary (<75 minutes of aerobic exercise per week). The University of Pennsylvania Human Subjects Review Committee approved this study, and written informed consent was obtained from all subjects before beginning study activities. Women were randomized to control or intervention groups following stratification of 1) baseline BMI above and below 30 kg/m 2 and 2) time since first menstrual cycle (> vs ≤ 10 yrs). In the current study, 132 participants had available samples at baseline and 116 of those women also had follow-up samples available. Sixteen women were lost to follow-up. Baseline analyses of TNFα and IL-12 included all 132 women (Control n=43, Low-dose n=44, High-dose n=45). Baseline analysis of CCL2 and IL-10 included 131 participants. Two data points at baseline were excluded due to technical error (CCL2 (Control n=43, Low-dose n=43, High-dose n=45) and IL-10 (Control n=42, Low-dose n=44, High-dose n=45)). A 5-menstrual-cycle long exercise intervention was selected to be long enough to induce changes in urinary hormones (primary outcome of the WISER Sister Trial), while balancing the feasibility of recruitment, complexity of delivering a distance-based intervention, and completing the study within the funding period. 24 , 27 , 28 Participants in the low-dose and high-dose aerobic exercise groups were asked to complete 150 minutes per week and 300 minutes per week, respectively. Participants within an intervention group were shipped treadmills (Smooth Fitness, model 5.65, King of Prussia, PA) to their homes and asked to complete home-based aerobic workouts for five menstrual cycles. Exercise intensity progressed in the first four weeks of the intervention for both groups until 70–80% of age predicted heart rate maximum (220-age) was attained. 24 , 26 , 27 Following the first four weeks, the high-dose group increased exercise duration by 20–25 minutes every two weeks until reaching 300 min/wk. Exercise intensity did not differ between the intervention groups. Participants in the control group were asked to maintain their regular physical activity levels (<75min/wk). Adherence to the exercise protocol was monitored by review of the participants’ exercise logs and heart rate monitor data as previously reported. 24 , 26 , 27 Data collection was performed at baseline and follow-up clinic visits. Participants completed 3-day dietary logs. Body composition variables were measured using dual-energy x-ray absorptiometry (DXA; Hologic) and all scans were analyzed by APEX 3.3. Fitness capacity was assessed by time completed in a Bruce protocol treadmill exercise test. Energy expenditure was measured with a modified version of the Modifiable Activity Questionnaire. 29 Biomarkers were assessed in fasted plasma samples stored at −80°C. Levels of TNFα, IL-12, CCL2 and IL-10 were assessed using ELLA Simple Plex technology (Protein Simple) following the manufacturer’s protocol. 30 Plasma samples were thawed at room temperature and then diluted 1:2 in Sample Diluent. 50ul of diluted plasma was added to each sample well and then directed down four separate channels that contain a capture antibody for one of the biomarkers. Each channel assessed antibody capture in triplicate and then is quantified using a pre-loaded calibration curve. Intra-assay CV% for each analyte was: TNFα (3.2%), IL-12 (7.2%), CCL2 (3.3%), and IL-10 (2.6%). The limit of detection for each analyte was: TNFα (0.278 pg/ml), IL-12 (0.39 pg/ml), CCL2 (0.35 pg/ml), and IL-10 (0.14 pg/ml). Concentrations were measured and analyzed using Simple Plex Runner software version v.3.3.0.98. Normality was first tested and either Wilcoxon signed rank tests or paired t tests were used accordingly to assess changes within group from baseline to follow-up. Independent two-sample t tests assessed baseline differences between women that completed the study and women that dropped out of the study. Correlations between variables were performed using Pearson’s correlation test. A one-way ANOVA was used to test for differences in the percent change in biomarkers between intervention groups after logarithmic transformation on the ratio. Baseline-adjusted linear regression was used to evaluate the absolute change in biomarker concentration following the intervention period. An extension of the Wilcoxon rank-sum test was used to test for linear trends across groups. Statistical analyses were conducted using STATA version 12 (Stata Corp.). False discovery rate due to multiple testing was controlled by Benjamini and Hochberg’s method 31 . Statistical significance was set at an alpha level of 0.05.

Demographic information is shown in Table 1 . The study population (n=132) had a mean age of 34.4 years and was predominantly white (85.6%), non-Hispanic (92.4%), and overweight (BMI: 26.8 kg/m 2 ). Compared to the control and high-dose groups, women in the low-dose group were significantly more likely to be married/partnered and to have children. Additionally, it was previously reported that BRCA1/2 mutation status, caloric intake, body fat percentage and intervention adherence (% of total prescribed minutes completed: low-dose = 85%, high-dose = 81%; % with ≥ 80% adherence: low-dose = 76%, high-dose = 75%) 24 – 27 were not significantly different between groups. A sensitivity analysis was conducted to compare baseline measurements between the 16 women who dropped out of the study prior to follow up and the 116 women who completed the study (data not shown). Reasons for drop out were previously reported 25 with time pressures and family issues given as the primary reasons. The majority of women who dropped out of the study were from the intervention groups (87.5%). Women who dropped out of the study were significantly different than study participants in that they were younger (30.6 years vs 34.8 years, p=0.02), had higher BMI (31.5 kg/m 2 vs 26.2 kg/m 2 , p<0.01), and were less fit (treadmill time of 6.96 minutes vs 8.27 minutes, p<0.01). They were also significantly less likely to be married/partnered (p<0.01), to be Caucasian (p<0.01) and to have completed higher levels of education (p<0.01). Concentration of biomarkers were not different between study participants and women that dropped out of the study. A correlation analysis was performed at baseline to assess associations between inflammatory biomarkers and measures of body composition, fitness capacity and energy balance ( Table 2 ). We observed a significant positive association between levels of TNFα and all other biomarkers (CCL2: r=0.275, IL-10: r=0.268, and IL-12 r=0.261). Higher levels of TNFα were also significantly associated with higher BMI (r=0.231). IL-12 was significantly associated with higher caloric intake (r=0.237) and higher fitness capacity (r=0.206). Energy expenditure and body fat percentage were not significantly correlated with any of the biomarkers. Absolute change in biomarker concentration from baseline to follow-up was assessed within and between groups ( Table 3 ). Control group participants experienced significant decreases of the pro-inflammatory biomarkers CCL2 (−11.9 pg/ml), IL-12 (−0.4 pg/ml) and TNFα (−0.4 pg/ml). The high-dose group experienced a significant increase in TNFα (0.67 pg/ml) levels following the intervention period whereas the other pro-inflammatory biomarkers didn’t change. Levels of pro-inflammatory biomarkers did not significantly change among women in the low-dose group. Finally, levels of the anti-inflammatory biomarker IL-10 did not significantly change within any group. Between groups, the change in IL-12 and TNFα concentrations in the low-and high-dose groups were significantly different from the control group, whereas the change in CCL2 and IL-10 was not significantly different between groups. Percent change of biomarker levels was calculated and compared between groups. We further tested if the observed changes in biomarker levels changed linearly as the dose of exercise increased ( Figure 1 ). The percent change of both IL-12 (low-dose: 36.6% ± 65.6, high-dose: 25.8% ± 50.2) and TNFα (low-dose: 12.4% ± 31.7, high-dose: 15.7% ± 15.9) was significantly different from the change in control group (IL-12: −22.0% ± 30.3, TNFα: −4.8% ± 20.0). These changes were also significantly dependent on exercise dose (P trend , p<0.01) ( Figure 1C and 1D ). The percent change in CCL2 was not significantly different between groups (control: −5.3% ± 20.7, low-dose: −0.1% ± 20.7, high-dose: 1.54 ± 19.7), however there was a significant dose response effect of the intervention (P trend , p<0.05) ( Figure 1A ). The percent change in IL-10 (control: 5.05% ± 31.7, low-dose: 6.14% ± 20.9, high-dose: 10.5% ± 37.8) was not significantly different between groups, nor was there a significant dose dependent effect of exercise ( Figure 1B ). Percent change of biomarkers between the low-and high-dose intervention groups were not significantly different.

This study is the first randomized controlled trial to examine the effect of a dosed 5-menstrual-cycle-long moderate-to-vigorous aerobic exercise intervention on circulating levels of CCL2, IL-10, IL-12, and TNFα in a national cohort of pre-menopausal women at risk for breast cancer. We observed significant exercise dose-dependent increases in the concentrations of the pro-inflammatory biomarkers CCL2, IL-12 and TNFα. Control group participants experienced significant decreases in the levels of pro-inflammatory cytokines whereas the intervention groups experienced increases in the levels of pro-inflammatory biomarkers. The increase in inflammatory biomarker levels was most striking in TNFα levels among participants in the high-dose exercise group (15.7% ± 15.9). Additionally, the increase in TNFα and IL-12 levels within both the low-and high-dose groups were significantly different from the decreases observed in the control group. Despite the conventional wisdom that structured exercise decreases levels of pro-inflammatory biomarkers, there have been studies that also report increases. Hayase et al observed that premenopausal women (n=9), but not postmenopausal women (n=9), had significantly increased levels of TNFα following 10 weeks of aqua aerobics (120 min per week) and resistance training (60 min per week). 32 They also showed that this increase in TNFα was significantly correlated (r= −0.816, p<0.05) with a decrease in visceral fat tissue. Brown et al reported that participants of the WISER Sister study from the low-and high-dose groups had significantly decreased levels of visceral adipose tissue compared to control group participants, 29 which may have influenced our increase in TNFα levels. There is also evidence in the literature that inflammatory biomarkers are not changed following an exercise intervention. In one example, 100 women (54.4 ± 7.1 yrs, 34% premenopausal,) that participated in a 12-month aerobic exercise intervention (6 days/wk, 60 min/day, 60–85% of maximal heart rate) experienced unchanged levels of inflammation, measured via CRP, despite improvements in body composition and fitness capacity. 33 Following a stratified analysis of women above and below a BMI of 30 kg/m 2 , however, a trend was observed that women with higher baseline BMI reduced their CRP levels more than women with a lower BMI. This observation may help explain why women in our study with an average BMI of 26.8 kg/m 2 did not experience a decrease in the levels of pro-inflammatory biomarkers. While this study did not assess CRP, the inflammatory biomarkers were chosen to represent various facets of the inflammatory response. CCL2 aids in the recruitment and activation of monocytes and macrophages in mammary tissue, 14 , 34 and it has also been associated with increased stromal density of mammary glands which is a risk factor for breast cancer. 35 TNFα signals through the TNF receptor 1 which is expressed by all human tissue, and intracellular responses to TNFα include regulation of apoptosis and production of cytokines. 36 IL-12 is highly involved in the functional activity of natural killer cells and T lymphocytes, and for this reason this pro-inflammatory cytokine can have beneficial or detrimental effects. 37 , 38 IL-10 has a major role in resolving the inflammatory response. Our study population included adult (34.4 ± 7.0 years of age), slightly overweight (BMI: 26.8 ± 6.3 kg/m 2 ), but otherwise healthy, pre-menopausal women. It was not anticipated that this cohort would have elevated levels of inflammation at baseline, but it was hypothesized that our exercise intervention could further decrease biomarkers of inflammation. All participants began the study with similar levels of each biomarker suggesting that changes in the measured biomarkers of inflammation were an effect of the intervention, and that consistent (>80% of total prescribed minutes of exercise were completed), 24 – 27 chronic (approximately 6 months on average), exercise training increased levels of inflammatory biomarkers. These results need to be taken in context with prior reports from the WISER Sister study. Our previous work has demonstrated an exercise dose-dependent decrease in leptin levels, estrogen sensitive breast tissue, and body fat. 25 , 26 , 27 Thus, given the full scope of physiological changes observed from this interventional study, we can speculate that in healthy women at risk for breast cancer, exercise induced changes in adipokines, body composition, and fibroglandular breast tissue may be more meaningful mechanistically for breast cancer protection than exercise induced changes in systemic inflammation. A major strength of this study was the high level of adherence to the exercise intervention in both the low-and high-dose group. This study allowed for a more direct interpretation of the effect of exercise on inflammation by focusing on volume of exercise rather than varying exercise intensity or including a caloric restriction component. The inclusion of a control sedentary group for comparison of inflammation levels with the intervention groups was an additional strength. Finally, this study focused solely on a national sample of women at high genetic risk of breast cancer who were otherwise healthy without other known diseases. 24 This population of women is underrepresented in the literature on inflammation and structured exercise. This study has several limitations, one of which is the selection of four cytokines. Inclusion of additional pro-and anti-inflammatory biomarkers and their receptors would provide a more complete inflammatory profile that may explain the changes we observed in our panel. Future studies will also assess a more stable marker of inflammation, such as CRP. Lastly, our study population was predominantly educated white women, which limits the application of our findings to other women more broadly. Overall, our study showed clear increases in levels of pro-inflammatory biomarkers following a 5-menstrual-cycle long moderate-to-vigorous aerobic exercise intervention. Despite the increase in levels of inflammation in these high-risk women, the WISER Sister study has been shown to improve hormone levels, breast density, body composition, fitness capacity, and energy expenditure. 28 – 31 Overall, the information gained from the WISER Sister study indicates that for healthy premenopausal women, the mechanism of reduced breast cancer risk observed in physically active individuals may not be through reduced levels of inflammation.

On average, women have a 12% lifetime risk of developing breast cancer 1 , and there are several factors involved in carcinogenesis such as sex hormones 2 , growth factors 3 , and adipokines 4 . Women may be at an elevated risk of developing breast cancer if they possess additional risk factors including breast cancer-associated mutations, 5 , 6 or increased levels of inflammatory biomarkers. 7 Physical activity has been shown to reduce breast cancer risk. 8 Pre-menopausal women who exercise (39+ MET-h/wk) can reduce their risk of breast cancer by 23% compared to physically inactive women. 9 One mechanism through which physical activity is hypothesized to decrease cancer risk is by decreasing levels of inflammation. Inflammation is associated with an increased risk of disease. 10 , 11 Specifically, increased levels of pro-inflammatory biomarkers, such as tumor necrosis factor alpha (TNFα), interleukin 12 (IL-12, IL-12p70 active heterodimer) and chemokine (C-C motif) ligand 2 (CCL2) have been implicated in the development and progression of cancer. 12 , 13 , 14 Anti-inflammatory biomarkers, like IL-10, are responsible for controlling the inflammatory response by inhibiting production of pro-inflammatory biomarkers by inflammatory macrophages. 15 , 16 Failure to resolve an inflammatory response can result in a chronic state of inflammation that is characterized by persistent, low levels of pro-inflammatory biomarkers and inflammatory cells, even in the absence of injury or infection, that can cause harm to healthy tissue. Mitigating chronic inflammation may be particularly important for women at high risk for breast cancer. Cells in an inflammatory environment are known to produce free radicals like reactive oxygen and nitrogen species. 17 Through oxidative stress, these free radicals are capable of damaging DNA and preventing DNA repair. In addition, mutations in genes that are associated with an increased risk of breast cancer (BRCA1/2 and PALB2) are also key components of the DNA repair process. 18 Furthermore, mutations in genes associated with the DNA damage response can also increase a woman’s risk of breast cancer. 2 , 3 Thus, reducing chronic inflammation, via implementation of healthy lifestyle changes, in women with deficiencies in DNA repair processes may be particularly important. Several studies in women have observed physical activity-induced decreases in pro-inflammatory biomarkers and increases in anti-inflammatory biomarker levels. 19 , 20 , 21 , 22 Despite data from observational studies on physical activity and interventional studies using structured exercise interventions 19 , 21 , 22 showing that both can improve levels of inflammatory biomarkers, there hasn’t been a randomized controlled intervention trial that has assessed whether structured aerobic exercise is able to change levels of inflammation in a dose-dependent manner in a population of women at high risk for breast cancer. Studies that have been performed in high risk women, specifically BRCA1/2 carriers, have primarily focused on observational data exploring the relationship between levels of physical activity and breast cancer risk. This data has shown that physical activity early in life can reduce breast cancer risk. 23 It is important to note these observational studies have assessed general physical activity (defined as any bodily movement produced by skeletal muscles that requires energy expenditure). Exercise is a subtype of physical activity. Exercise is planned, structured and intentional movement with a goal to improve or maintain physical fitness. Exercise physiology indicates that the body reacts and adapts to unique (frequency, intensity, and duration) physical stressors in unique ways. Thus, the objective of this prospective, randomized controlled trial, was to assess the effect of a dosed aerobic exercise intervention on levels of inflammation in pre-menopausal women at high risk of breast cancer.