Salad vegetables dietary pattern protects against HER-2-positive breast cancer: A prospective Italian study
Version of Record online: 23 APR 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 121, Issue 4, pages 911–914, 15 August 2007
How to Cite
Sant, M., Allemani, C., Sieri, S., Krogh, V., Menard, S., Tagliabue, E., Nardini, E., Micheli, A., Crosignani, P., Muti, P. and Berrino, F. (2007), Salad vegetables dietary pattern protects against HER-2-positive breast cancer: A prospective Italian study. Int. J. Cancer, 121: 911–914. doi: 10.1002/ijc.22714
- Issue online: 22 JUN 2007
- Version of Record online: 23 APR 2007
- Manuscript Accepted: 2 FEB 2007
- Manuscript Received: 27 OCT 2006
- Italian Association for Research on Cancer (AIRC)
- breast cancer;
- dietary pattern;
Studies investigating the relation of diet to breast cancer have produced conflicting results. We hypothesized that dietary factors associated with breast cancer risk might differentially influence the HER-2 status of the cancers that develop, and investigated this hypothesis by analyzing the data of the ORDET prospective study. We analyzed 8,861 volunteer women residents of the Varese Province, Italy, for whom we had full data. By December 31, 2001, 238 cases had occurred in which HER-2 status was known. Four dietary patterns had been identified previously by factor analysis: salad vegetables (high consumption of raw vegetables and olive oil), prudent (cooked vegetables, poultry, fish), western (potatoes, meat, eggs, butter), and canteen (pasta, tomato sauce, wine). In our study, relative risks (RRs) of developing HER-2-positive and HER-2-negative breast cancers by tertiles of dietary pattern factor scores were assessed by multinomial logistic regression. The salad vegetables dietary pattern had a protective effect against HER-2-positive cancers (RR = 0.25, 95% CI 0.10–0.64, for the highest tertile; ptrend = 0.001), much stronger than for HER-2-negative cancers (pheterogeneity = 0.039). This important finding that a salad vegetables dietary pattern protects mainly against a specific breast cancer subtype indicates that future studies on environmental/dietary risk factors should explicitly take account of the heterogeneity of breast cancer phenotypes. © 2007 Wiley-Liss, Inc.
The incidence of breast cancer varies markedly between countries, suggesting that dietary habits might be an important cause of this variation. A recent analysis of the ORDET (Hormones and Diet in Etiology of Tumors) cohort suggested that a dietary pattern characterized by high consumption of raw vegetables and olive oil may reduce breast cancer risk, especially in lean women (BMI < 25).1 Other studies of dietary pattern have not shown such an association,2 and epidemiologic studies on vegetable consumption and breast cancer have not provided clear results.3, 4 Etiological relationships between diet and breast cancer may have been masked by lack of consideration of cancer phenotype. Only a few studies have investigated associations between risk factors and cancer phenotype, most of which investigated the expression of receptors for estrogen and progesterone.5, 6, 7 Studies considering HER-2 overexpression, which confers aggressivity8, 9, 10 and is usually associated with an estrogen receptor-negative phenotype,11 have been carried out mainly to investigate the relation of HER-2 overexpression to prognosis and response to treatment.12 The relationship between etiologic risk factors and HER-2 overexpression has not been extensively investigated. It has been suggested that HER-2 overexpression bears little relation to the classic reproductive risk factors for breast cancer.13 It is unclear, however, whether dietary factors can influence the biological characteristics of breast cancers, and there are no data on their influence on HER-2 expression in humans. We investigated the association between dietary patterns and HER-2 expression in the ORDET prospective study.
Material and methods
Women volunteers, aged 34–70, were recruited from 1987 to 1992, from residents in Varese Province, Italy. Participants completed a self-administered dietary questionnaire and anthropometric and other data were collected systematically.
The women were followed-up (mean follow-up of the cohort 11.5 years) through the Varese Cancer Registry, to determine the incidence of breast cancer. Information on the stage, treatment and pathological characteristics of the cancers was also collected.
A total of 8,984 ORDET women had dietary factor scores (the extent to which diet conformed to a particular dietary pattern), previously determined by factor analysis of the food frequency questionnaire completed at recruitment1; of these 8,861 women also had full anthropometric data. A total of 267 breast cancer cases had occurred among these women by December 31, 2001, in 238 of which HER-2 status was available. To estimate the relative risks (RRs) with 95% confidence intervals (CI) of developing HER-2-positive and -negative cancer, we analyzed these 8,861 women, including the 238 cases information on HER-2 status.
Dietary patterns were determined in our previous study1 on the 107 food items in the questionnaire, grouped into 34 predefined food groups based on similarities in nutrient profile and culinary usage. Individual food items were classified into separate food groups if their composition differed substantially from that of other food groups (e.g. eggs or pizza). Furthermore, if a given food item was suspected to represent a particular dietary habit when consumed raw and another habit when consumed cooked, the cooked and raw items were classified separately (e.g. leaf vegetables, tomatoes, carrots).
Exploratory factor analysis was then applied to reduce the food groups to a small number of factors that explained the maximum fraction of the variance. An orthogonal rotation procedure,14 varimax rotation, was performed to simplify the factor structure and render it more easily interpretable. To determine the number of factors to retain (4 factors), the Scree test15 was employed. The next step was to name the factors that emerged. Food groups with an absolute loading greater than 0.25 on a given factor were considered to contribute importantly to that factor; factors were thereby interpreted as dietary patterns and named after the food groups having the highest loadings. Food groups with a positive loading contributed directly to a dietary pattern and food groups with negative loading were inversely associated with a dietary pattern. Factor loadings can be considered as correlation coefficients between food groups and dietary patterns, and take values between −1 and +1.
The factor score for each dietary pattern was then calculated for each woman. Briefly, factor scores were formed by standardizing each variable (food group) to have zero mean and standard deviation of 1, weighting it with a corresponding factor score coefficient, and then summing the terms. The factor score indicated the extent to which the woman's diet conformed to each dietary pattern. Factor scores were categorized into tertiles.
The first dietary pattern obtained was characterized by greatest loadings on raw vegetables and olive oil; it was therefore designated salad vegetables. The second pattern, western, had high loadings on potatoes, ravioli, red and processed meat, eggs, butter, seed oil (as added fat) and cakes. The third pattern, canteen, was characterized by high consumption of pasta, tomato sauce, olive oil and wine. The fourth pattern, prudent, was characterized by high consumption of cooked vegetables, rice, poultry, fish and low consumption of alcohol.
For the present study, we obtained HER-2 status from pathology records of the hospitals where patients were treated. When HER-2 status was not available, immunohistochemical determinations were performed at the Molecular Biology Unit Laboratories, Istituto Nazionale dei Tumori, Milan. Of the 238 cases with known HER-2 status 153 came from clinical records and 85 by analysis of archived paraffin-embedded cancer specimens. In Italy HER-2 assay methodology and results interpretation are standardized.16 HER-2 was considered positive when complete membrane staining was observed in >10% of cancer cells (3+ score), following use of anti-HER-2 CB11 antibody and peroxidase–streptavidin revelation.
Analysis of the 85 archived cases employed a Dako TechMate 1000 automated immunostainer and anti-HER-2 CB11 antibody from Ylem (Avezzano, AQ, Italy).
We carried out multinomial logistic regression analyses17 to estimate RRs with CIs of developing HER-2-positive and HER-2-negative cancers, in relation to tertiles of the factor score of each dietary pattern (lowest tertile as reference), controlling for all the other dietary patterns. RRs were also adjusted for total energy intake, age, years of education, height, weight, parity, age at menarche, smoking and menopausal status. The method allows estimation of risk for HER-2-positive and HER-2-negative cancer in comparison to non-cases and direct significance testing of differences between regression coefficients. Risk estimates obtained using cohort members without breast cancer (non-cases) as comparison group were similar to those provided by proportional hazards regression analysis.
The significance of dietary pattern “exposure” trends was tested by the likelihood ratio test. RR heterogeneity was tested by the Wald test. Statistical analyses were performed using Stata version 7.18
Table I shows RRs of developing HER-2-positive and HER-2-negative cancers, according to tertiles of “exposure” to each dietary pattern. The salad vegetables pattern had a protective effect against HER-2-positive cancers: the protection increased with increasing factor score (RR 0.33, 95% CI 0.15–0.73, and 0.25, 95% CI 0.10–0.64, for medium and high scores, respectively) with a significant linear trend (p = 0.001). The protective effect was much stronger for HER-2-positive cancers than for HER-2 negative cancers (p for heterogeneity = 0.039). The western pattern had no clear association with either HER-2-positive or HER-2-negative cancer. The canteen pattern was associated with nonsignificantly increased RR of developing HER-2-positive disease. Medium exposure to the prudent pattern was associated with reduced RR of HER-2-positive cancer (RR 0.38, 95% CI 0.16–0.90) and with increased RR of HER-2-negative cancer (RR 1.47, 95% CI 1.02–2.11) (p for heterogeneity = 0.017).
|Dietary pattern||RR by tertiles of dietary factor score||P for trend||p for heterogeneity2|
|Low score||Medium score||High score|
|No. of cases/non-cases3||24/2,879||9/2,875||7/2,869||0.001||0.039|
|RR4||1||0.33 (0.15–0.73)||0.25 (0.10–0.64)|
|No. of cases/non-cases||14/2,867||13/2,879||13/2,877||0.584||0.812|
|RR4||1||0.83 (0.37–1.89)||0.75 (0.27–2.08)|
|No. of cases/non-cases||9/2,885||19/2,881||12/2,857||0.530||0.319|
|RR4||1||2.11 (0.92–4.84)||1.39 (0.50–3.84)|
|No. of cases/non-cases||17/2,877||8/2,872||15/2,874||0.372||0.017|
|RR4||1||0.38 (0.16–0.90)||0.72 (0.35–1.48)|
|No. of cases/non-cases||79/2,879||65/2,875||54/2,869||0.072|
|RR4||1||0.84 (0.59–1.18)||0.71 (0.48–1.03)|
|No. of cases/non-cases||69/2,867||74/2,879||55/2,877||0.651|
|RR4||1||1.11 (0.78–1.59)||0.88 (0.55–1.40)|
|No. of cases/non-cases||66/2,885||70/2,881||62/2,857||0.520|
|RR4||1||1.13 (0.79–1.62)||1.14 (0.75–1.75)|
|No. of cases/non-cases||50/2,877||77/2,872||71/2,874||0.126|
|RR4||1||1.47 (1.02–2.11)||1.36 (0.93–1.98)|
To our knowledge this is the first study to investigate dietary risk factors for breast cancer in relation to HER-2 overexpression. In our previous study,1 we found that the salad vegetables dietary pattern was protective against breast cancer, irrespective of HER-2 status (which was not then available) with an RR of 0.66 (95% CI 0.47–0.95) for the highest tertile of exposure. The present study shows that the protective effect of the salad vegetables dietary pattern was much stronger against HER-2-positive cancers, and mainly confined to this cancer type. Only about 20% of breast cancers were HER-2-positive in our study, in accordance with the findings of other studies.9, 13
We performed immunohistochemical determinations on archived paraffin-embedded tissues mostly for cases diagnosed in the earlier years when HER-2 assay was not routine. It is possible that archived tissues degraded over time. However, the percentages of HER-2-positive cancers were similar in the two groups (16% in routinely determined cases and 19% in ad hoc determined cases).
The protective effect of the salad vegetables pattern against HER-2-positive cancers could be related to its high antioxidant content. The raw vegetables and olive oil characteristic of this pattern are rich in antioxidant vitamins (β-carotene, vitamin C, vitamin E) and polyphenols, all of which may prevent DNA damage due to reactive oxygen species.19In vitro studies have shown that hydrogen peroxide stimulates gene amplification, supporting the hypothesis that normal metabolic products of the cell can be a source of genome instability.20
The salad vegetables diet is also rich in folate, which is involved in DNA synthesis, repair and methylation.21 Folic acid's role in DNA repair may limit gene amplification. Cancers that overexpress HER-2 are generally characterized by HER-2 gene amplification. Gene amplification has been shown to be increased in a hamster ovarian cell line lacking folate.22 Folic acid may also alter gene expression via its involvement in cytosine methylation at promoter regions, thereby reducing expression of the related genes.
The protective effect of the medium score of the prudent pattern against HER-2-positive cancers might be due to its relatively high vegetable and low alcohol content. High alcohol consumption and low vegetable intake decrease folate availability,23 and this might explain the (weak) association of HER-2-positive disease with the canteen pattern. Alcohol can also promote cancer by acetaldehyde-induced mutagenesis and induction of oxidative damage.21
In general, however, these diet-based mechanisms would be expected to influence the occurrence of all or most types of breast cancer, and it is unclear why they should specifically protect against HER-2-positive cancers. Interestingly, recent findings show that oleic acid suppresses HER-2 overexpression in human breast cancer cell lines, and also interacts synergistically with the drug herceptin in promoting apoptotic death of these HER-2-overexpressing cells.24 Furthermore a study on induced mammary tumors in the rat indicated that corn oil upregulates and olive oil downregulates the epidermal growth factor receptor and HER-2 signal transduction pathways in the tumor cells.25 Thus the possibility arises that oleic acid, the main monounsaturated fatty acid of olive oil, may be largely responsible for the protective effect against HER-2-positive cancers.
Compared to our previous study,1 the present findings suggest some new insights: first that the effects of diet on breast cancer risk might be underestimated when HER-2-positive and HER-2-negative cancers are considered together; second that factors influencing the occurrence of HER-2-positive cancers may differ from those influencing the occurrence of HER-2-negative cancers; third that the protective effect of a diet rich in raw vegetables and olive oil is much stronger (p for heterogeneity = 0.039) against the small proportion (about 20%) of breast cancers that are HER-2-positive. Future studies on environmental/dietary risk factors should explicitly take account of the heterogeneity of breast cancer phenotypes.
We acknowledge with thanks the invaluable collaboration of pathologists at hospitals in the Province of Varese, where patients were treated. The authors also thank Don Ward for help with the English.
- 3World Cancer Research Fund and American Institute for Cancer Research Food. Nutrition and the prevention of cancer: a global perspective. Washington: American Institute for Cancer Research, 1997.
- 14Applied regression analysis and other multivariate methods. Pacific Grove, CA: Duxbury, 1988., , .
- 16Italian Network for Quality Assurance of Tumor Biomarkers (INQAT) Group. Interobserver reproducibility of immunohistochemical HER-2/neu assessment in human breast cancer: an update from INQAT round III. Int J Biol Markers 2005; 3: 189–94.
- 17Applied logistic regression, 2nd edn. New York: Wiley, 2001., .
- 18Stata Corporation. Stata statistical software: release 7.0. College Station, TX: Stata Corporation, 2001.
- 24Oleic acid, the main monounsaturated fatty acid of olive oil, suppresses Her-2/neu (erbB-2) expression and synergistically enhances the growth inhibitory effects of trastuzumab (Herceptin) in breast cancer cells with Her-2/neu oncogene amplification. Ann Oncol 2005; 16: 359–71., , , .