Breast cancer is the most common cancer in American women, with over 40,000 annual deaths and nearly 200,000 estimated new cases in 2009 1 . Given the high prevalence of breast cancer and associated physical, social, and psychological detriments 2 , reducing cancer incidence through modifiable risk factors such as nutrition is a high public health priority. Estimates suggest that cancer incidence could be reduced by ~1/3 through modifications in diet, including changes in total energy intake and diet content of fat, protein or carbohydrate 3 . With increased attention on higher protein, low carbohydrate diets for weight management, secondary effects of these diets on tumor development are important to evaluate.

Some studies report that dietary protein, and particularly consumption of animal protein, is linked to increased risk of breast cancer 4 5 . However, recent large-scale evaluations have failed to identify intakes of meat, eggs, or dairy products as risk factors 6 . Conversely, epidemiological studies have reported protective effects of dietary protein on breast cancer incidence 7 and mortality 8 . A recent case-controlled study suggests that a dietary pattern including the highest relative levels of animal protein is inversely associated with breast cancer risk 9 .

Evidence is now accumulating that dietary carbohydrates, glycemic load, and hyperinsulinemia are associated with increased cancer risk 10 11 and mortality 12 . Higher levels of circulating and post-prandial insulin have been reported to enhance cellular proliferation and tumor development 13 through upregulation of the PI3K/Akt signal pathway which is implicated in breast cancer promotion and tumorigenesis 14 . Furthermore, insulin stimulates the ovaries to produce androgens 15 , which have been linked to increased breast cancer risk 16 . To our knowledge, no studies have attempted to compare the effects of dietary protein versus carbohydrates in an experimental model of breast cancer.

We have previously demonstrated that feeding rats diets with increased protein and reduced carbohydrates enhances muscle insulin signaling and glucose uptake, attenuates post-prandial hyperinsulinemia, and stimulates skeletal muscle protein synthesis 17 18 . Our experimental model uses a meal-feeding protocol with three discrete meals each day to mimic human eating patterns and allow for isolation of post-prandial meal responses. This study examined the effects of feeding diets differing in ratios of carbohydrate/protein, but within the Dietary Reference Intake (DRI) acceptable macronutrient distribution range, on tumor incidence, size, and development, in a chemically-induced model of breast cancer. We hypothesized that a diet with a reduced carbohydrate:protein ratio would have a favorable impact on breast cancer development by tempering post-prandial insulin response.

The objective of this study was to identify potential effects of manipulating the dietary carbohydrate:protein ratio on the promotion of MNU-induced breast cancer tumors. A major finding of our study is that a HPMC diet with an reduced carbohydrate:protein ratio significantly reduced the rate of incidence of palpable mammary tumors. Although tumor latency was not different between groups, once palpable tumors began to appear, the rate of tumor appearance proceeded more rapidly in the LPHC group. This suggests that either reduced dietary carbohydrate or elevated dietary protein (or both), can attenuate the early development of mammary tumors. In agreement with our previous study 17 , post-prandial insulin at 3 wk was elevated in the LPHC group and not in the HPMC group. Thus, animals in the LPHC group were likely exposed to higher, repeated elevations in serum insulin following each discrete meal during the period leading to the emergence of palpable mammary tumors. Post-prandial insulin at 10 wk was elevated in both groups, although host metabolism was likely confounded by tumor pathology, and this change does not necessarily diminish the early effect of post-prandial insulin on tumor promotion.

Nearly two decades ago, hyperinsulinemia was identified as a significant risk factor for breast cancer, independent of body composition or adiposity 22 . Cohort studies have found that women surgically treated for breast cancer with high fasted insulin levels had substantially increased risk for reoccurrence and death 23 and that the highest tertile of fasted insulin levels was associated with a 2-fold increase in post-menopausal breast cancer risk 24 25 . Most recently, the presence of metabolic syndrome was reported to be associated with breast cancer risk 26 , and insulin resistance and post-prandial hyperinsulinemia are central features of metabolic syndrome 27 , leading to suggestions that metabolic syndrome may serve as a prognostic factor for breast cancer 28 . With respect to dietary factors, a positive association between high glycemic index/load and breast cancer risk has been identified 29 30 . High total carbohydrate intake 31 and sucrose intake have been associated with unfavorable outcomes for breast cancer.

One of the mechanisms by which insulin has been proposed to increase breast cancer risk is via IGF-I. Elevated insulin can increase serum free IGF-I concentrations 32 . Overexpression of IGF-I in mammary epithelium in vivo leads to hyperplasia, spontaneous tumorogenesis, and enhanced susceptibility to chemical carcinogens 33 , and circulating IGF-I appears to be important for the onset and subsequent development of mammary tumorigenesis in in vivo experiments 34 35 . In the present study, the elevated post-prandial insulin observed in the LPHC group was not associated with increased serum IGF-I concentrations.

Clinically, the effect of IGF-I on breast cancer remains equivocal. Elevated serum IGF-I concentrations appear to increase risk of pre-, but not post-menopausal breast cancer, while higher blood glucose levels were associated with increased risk for developing post-menopausal breast cancer 36 . However, a case control study identified a weak association for IGF-I with breast cancer in women under the age of fifty 37 . Most recently, the Women's Health Initiative study found a strong association between fasting insulin levels and breast cancer in post-menopausal women, but no association of free IGF-I 38 .

IGF-I bioactivity is modulated by IGF binding proteins (IGFBP), which can block IGF-I from binding to its receptor. At least 75% of circulating IGF-I is bound to IGFBP-3 39 , though IGFBP-3 levels remain fairly constant within individuals 40 . IGFBP-1 and IGFBP-2 levels are inversely associated with insulin levels 41 , which suggests that chronic hyperinsulinemia could depress binding protein interaction and thus increase free IGF-I. Serum IGF-I levels have also been shown to relate to protein intake 42 . In the present study at 10 wk, there was a trend for post-prandial IGF-I to be higher (p = 0.055) in the HPMC group. Despite the elevation of IGF-I in the HPMC group, the rate of tumor incidence was significantly lower.

Consistent with our findings, early animal experiments investigating the influence of dietary protein on breast cancer found that feeding a higher protein diet prior to administration of a carcinogen resulted in a reduction of tumor prevalence 43 , suggesting a protective effect of protein during the initiation phase. Additionally, increased dietary protein fed after tumor induction had no independent effect on the promotion phase of carcinogenesis 44 . In contrast to the findings of the present study, Hawrylewicz et al. found higher dietary protein (33% vs. 15%) increased the number of tumors and tumor weight in MNU-induced rats, despite no difference in tumor incidence or latency 45 . However, this study differed substantially from the current study in that the rats were exposed in utero from dams fed the test diets and then consumed the same test diets ad libitum after weaning.

A further consideration is that elevated dietary protein was accomplished at the expense of dietary carbohydrate, and since tumors are obligate glucose consumers, HPMC may have been less energetically favorable to neoplastic tissues. Similarly, a ketogenic diet (80% medium chain triglycerides) fed to mice with implanted colon adenocarcinomas reduced tumor weight, while the inclusion of hydroxybutyrate in drinking water counteracted the stimulation of tumor growth by insulin infusion 46 . However, fasted glucose was not different between groups at either 3 or 10 wks, and in reflecting current dietary guidelines, the HPMC diet was not low enough in carbohydrate to induce ketogenesis.

A limitation of the current study is its relatively short length. At 10 wks, tumor incidence had only reached 40% in the HPMC group. Due to the labor-intensive nature of the meal-feeding protocol, this study was designed for 10 wk, with a goal of both groups exceeding 50% tumor incidence. However, within the short period of time, the study still achieved divergent rates of tumor appearance between diet groups.

In total, the present study provides evidence that reducing the dietary carbohydrate:protein ratio attenuates the progression of mammary tumors, and this finding is consistent with reduced post-prandial insulin release potentially diminishing the proliferative environment required for breast cancer tumors to progress. These results warrant additional investigation into the potentially protective effect that elevated dietary protein and reduced carbohydrate may have on breast cancer development.

The authors declare that they have no competing interests.

CM participated in the design of the study, coordinated the study, performed the statistical analysis, and drafted the manuscript. RV assisted in performing the study and helped to draft the manuscript. DL conceived of the study, participated in its design, and helped to draft the manuscript. SD helped in conception of the study and its design. KS helped in the conception of the study and its design. MW performed the histochemical analysis. SMD helped to draft the manuscript. All authors read and approved the final manuscript.