Mitogen-activated protein kinases (MAPK) act as integration points for multiple biochemical signals and are involved in a variety of cellular processes, including cell proliferation, differentiation, transcription regulation and development [1]. Various environmental stimuli, cytokines, and hormones activate MAPK pathways. ERK1 and ERK2 are activated by stimuli such as growth factors and cytokines [1], while the JNK pathway is activated by radiation, environmental stresses, and growth factors. The JNK pathway has been shown to be involved in the development of obesity and type 2 diabetes [2, 3]. The p38 MAPKs have been linked to autoimmunity in humans and are activated by chemical stresses, hormones, cytokines including IL-1 and TNF, and oxidative stress [1, 4].

Activation of MAPK in breast carcinoma has been shown to induce gene expression leading to increased proliferation, invasion and metastasis [5-7]. In this study, we examined the association between variants in MAPK genes and survival after diagnosis with breast cancer in non-Hispanic white (NHW) and U.S. Hispanic/Native American (NA) women, a genetically admixed population. Our analysis is motivated by the differences in breast cancer incidence and survival rates for these groups of women. Hispanic/NA women living in the Southwestern United States have lower incidence rates of breast cancer, although survival is slightly less. Given that MAPK genes are activated by a variety of factors related to growth factors, hormones, and environmental stress, we hypothesized that certain diet and lifestyle factors associated with activation of this signaling pathway would interact with genetic variation in MAPK genes to alter survival in women diagnosed with breast cancer. To our knowledge studies have not reported how MAPK genes relate to breast cancer survival and how diet and lifestyle factor influence those associations.

This analysis from the Breast Cancer Health Disparities Study includes participants with information on survival from two population-based case-control studies, the 4-Corners Breast Cancer Study (4-CBCS) and the San Francisco Bay Area Breast Cancer Study (SFBCS) [8]. In the 4-CBCS, participants were between 25 and 79 years of age, residents of Arizona, Colorado, New Mexico, or Utah with a histologically confirmed diagnosis of first primary invasive breast cancer (n=1391) between October 1999 and May 2004 [9]. The SFBCS included women aged 35 to 79 years from the San Francisco Bay Area diagnosed with a first primary histologically confirmed invasive breast cancer (n= 946) between April 1997 and April 2002 [10, 11]. All participants signed informed written consent prior to participation; this study was approved by the Institutional Review Boards for Human Subjects at each participating institution.

Data were harmonized across study-specific questionnaires[8]. Women were considered post-menopausal if they reported either a natural menopause or if they reported taking hormone therapy (HT) and were still having periods or were at or above the 95th percentile of age for those who reported having a natural menopause (i.e.; ≥ 12 months since their last period); others were classified as pre-menopausal. Lifestyle variables included body mass index (BMI) calculated as self-reported weight (kg) during the referent year divided by measured height squared (m²). Cigarette smoking was classified as ever versus never having smoked cigarettes on a regular basis (> than 100 cigarettes). Having a history of diabetes included those told by a doctor that they had diabetes or high blood sugar; in the SFBCS not all women were asked questions regarding diabetes and are excluded from that portion of the analysis. Dietary information was collected via a computerized validated diet history questionnaire for the 4-CBCS[12, 13] or a modified version of the Block Food Frequency Questionnaire for the SFBCS[14].

DNA was extracted from either whole blood or mouthwash samples (n=345). Genotyping was completed for 933 women from the 4-CBCS who self-identified as NHW, 412 Hispanic, 8 NA, 14 NHW/Hispanic, 10 NHW/NA, 10 Hispanic/NA, and 4 NHW/Hispanic/NA and for 252 women from the SFBCS who self-reported being NHW and 694 who reported being Hispanic. A tagSNP approach was used to characterize variation across candidate genes. TagSNPs were selected using the following parameters: linkage disequilibrium (LD) blocks were defined using a Caucasian LD map and an r²=0.8; minor allele frequency (MAF) >0.1; range of -1500 bps from the initiation codon to +1500 bps from the termination codon; and 1 SNP/LD bin. We used 104 Ancestry Informative Markers (AIMs) to distinguish European and Native American (NA) ancestry in the study population[8]. All markers were genotyped using a multiplexed bead array assay format based on GoldenGate chemistry (Illumina, San Diego, California). The following genes were evaluated: DUSP4 (6 SNPs), DUSP6 (1 SNP), MAP2K1 (6 SNPs), MAP3K1 (7 SNPs), MAP3K2 (3 SNPs), MAP3K3 (2 SNPs), MAP3K7 (6 SNPs), MAP3K9 (19 SNPs), MAPK1 (6 SNPs), MAPK3 (1 SNP), MAPK8 (4 SNPs), MAPK12 (2 SNPs), and MAPK14 (9 SNPs).

Data on survival were available from state cancer registries and included date of death or last follow-up (month and year), cause of death, and stage of disease at time of diagnosis. Survival (in months) was calculated as the difference between diagnosis date and date of death or last follow-up. Survival data in this report includes survival through December 2011 for Utah, Colorado, and California, through 2010 for Arizona, and through 2008 for New Mexico.

The program STRUCTURE was used to compute individual ancestry for each study participant assuming two founding populations[15, 16]. Participants were classified by level of percent NA genetic ancestry. Assessment across categories of ancestry was done using cut-points based on the distribution of genetic ancestry in the control population with the goal of creating distinct ancestry groups that had sufficient power to assess associations. Two strata of ≤28% and >28% were used to evaluate associations by level of NA ancestry since a two population structure best fit the population.

Associations between SNPs and diet and lifestyle factors and all-cause mortality and breast cancer-specific mortality were evaluated using Cox proportional hazards models to obtain multivariate hazard ratios (HR) and 95% confidence intervals (CI) for all women and within strata of genetic ancestry using SAS version 9.3 (SAS Institute, Cary, NC). Individuals were censored if they were lost to follow-up. When examining breast cancer-specific mortality, women who died of other causes also were censored. All SNPs were evaluated as a co-dominant model and if initial analysis suggested too few homozygote variants a dominant model was used. In other instances where a recessive model appeared to fit the data, it was used to evaluate risk estimates. Models were adjusted for age, study center, BMI (continuous), percent NA ancestry (continuous), and SEER stage (local, regional, distant).

We used the adaptive rank truncated product (ARTP) method that utilizes a highly efficient permutation algorithm to determine the significance of each gene and of the overall pathway with survival[17, 18]. We permuted the survival outcome 10,000 times in R version 3.0.1 (R Foundation for Statistical Computing, Vienna, Austria). SNP associations were assessed among the observed and permuted data in R using p values from likelihood-ratio tests comparing Cox proportional hazards models with and without the SNP variable. The P_ARTP is based on assessment of five truncation points for the pathway and each gene. Results included in tables are based on SNPs that had significant (p<0.05) or marginally significant (0.07 or less) gene p values and/or SNPs that were statistically different by ancestry group.

Interactions between genetic variants and lifestyle factors, including cigarette smoking (ever vs never), any alcohol consumption during lifetime (any vs none), a history of diabetes (yes vs no), and dietary factors during referent year (quartiles of nutrient consumption/1000 calories categorized by study-specific control distribution with data presented with middle two quartiles combined) were assessed using p values from one-degree of freedom Wald chi-square tests. We adjusted for multiple comparisons within each gene to determine significance of interactions. We used a step-down Bonferroni correction (i.e., Holm method) to adjust for multiple comparisons and utilized SNP spectral decomposition to take into account the correlated nature of the data[19, 20].

Approximately the same percentage of women who self-reported being NHW and Hispanic/NA died during follow-up (Table 1). Breast cancer was the most common cause of death among NHW (48.9%) and Hispanic/NA (55.3%) women. All cancers contributed to nearly 2/3 of deaths among NHW (64.4%) and 64.8% of deaths among Hispanic/NA women. Lung and pancreatic cancer were the next most common cancer deaths after breast cancer. Cardiovascular diseases accounted for roughly 10.5% of deaths, while respiratory diseases accounted for slightly over 4% of deaths.

Greater age, advanced SEER disease stage, a history of diabetes or high blood sugar, and lower vitamin C intake were associated with decreased survival among both NHW and Hispanic/NA women (Table 2). A history of cigarette smoking and a larger BMI were associated with poorer survival among NHW women, while being pre-menopausal, and both greater amount of long-term alcohol and any long-term alcohol consumption were associated with poorer survival among Hispanic/NA women.

Several MAPK genes were associated with breast cancer-specific mortality among women with low NA ancestry (Table 3). Although the overall pathway was not statistically significant (P_ARTP=0.10), MAP2K1 (P_ARTP=0.04), MAP3K9 (P_ARTP=0.02), and MAPK12 (P_ARTP=0.05) were significantly associated with breast cancer-specific mortality and MAPK1 was marginally significant (P_ARTP=0.07). Several SNPs were significantly different by ancestry (p<0.05), including MAP3K1 rs16886403 (HR_TC/CC for low NA ancestry = 1.48 95%CI 1.01,2.17 and for high NA =0.74 95% CI 0.45,1.19) and MAPK1 rs9610375 (HR_TT= 0.56 95% CI 0.34-0.94 for low NA ancestry and HR 1.13 95% CI 0.59,2.15 for high NA ancestry) that showed stronger associations with breast cancer-specific mortality among women with low NA ancestry; MAP3K3 rs3785574 had a stronger direct association among women with higher NA ancestry.

Stronger associations were observed between MAPK genes and all-cause mortality (Table 4), which also were predominately among women with low NA ancestry. The overall pathway was statistically significant among women with lower NA ancestry (P_ARTP=0.02). At the gene level, MAP2K1 (P_ARTP=0.02), MAP3K1 (P_ARTP=0.02), MAP3K9 (P_ARTP=0.01), and MAPK1 (P_ARTP=0.05) were associated with all-cause mortality. MAP3K2 was significantly associated with all-cause mortality among women with higher NA ancestry (P_ARTP=0.04). Significant differences for all-cause mortality by NA ancestry were also noted for several SNPs in MAP2K1 (rs1432442) (HR_AA 2.93 95% DI 1.46,5.87 for high NA ancestry) and MAP3K9 where rs11625206 was significantly inversely associated with mortality among women with high NA ancestry while rs118447474, rs8010714, and rs11158881 were inversely associated with mortality among women with low NA ancestry.

Several demographic, diet, and lifestyle factors were associated with all-cause mortality and breast cancer-specific mortality (Table 5). Moderate alcohol consumption was significantly associated with lower risk of all-cause mortality. Women diagnosed when they were post-menopausal were less likely to die than women diagnosed when pre-menopausal. Histories of cigarette smoking and diabetes were both significantly associated with increased risk of all-cause mortality although the association with breast cancer-specific mortality did not reach statistical significance. High total caloric intake was associated with increased risk of both all-cause mortality and breast cancer-specific mortality, whereas high intakes of dietary folate, vitamin C, and beta carotene were associated with lower of all-cause mortality.

Assessment of interaction between diet and lifestyle factors and MAPK genes showed significant interaction between DUSP4 and alcohol consumption (increased risk for rs12540995 and rs3824133 and decreased risk for rs47482 with less common genotype and low level of consumption), calories (increased risk with high caloric consumption relative to common genotype with low caloric consumption), dietary fat (both common genotype and high fat and rare genotype and low fat increased risk) and fiber (common genotype and high fiber reduced risk), and cigarette smoking (common genotype and smoking increased risk) for all-cause mortality (Table 6). MAP3K2, MAP3K3, MAP3K7, MAPK1, and MAPK14 interacted with alcohol consumption where a pattern of the more common genotype with high alcohol consumption reduced risk of all-cause mortality. Additionally MAPK14 interacted with calories (high intake with common genotype increased risk), dietary fiber (common genotype and high fiber reduced risk), and folate (common genotype and high folate reduced risk); MAP3K2 interacted with diabetes (less common genotype and having diabetes increased risk); and MAP3K7 interacted with vitamin C (no protective effect for the common genotype with low vitamin C while there was a protective effect for low vitamin C with the less common genotype).

Several MAPK genes were associated with breast cancer-specific and all-cause mortality. Associations were consistently observed among women with low NA ancestry. Diet and lifestyle factors were associated with survival after adjusting for disease stage regardless of genetic ancestry. Furthermore, diet and lifestyle factors interacted with select genes in the MAPK pathway to alter associations with all-cause mortality.

Activation of MAPK, including ERK1/2, JNK, and p38, in breast carcinoma has been shown to induce gene expression leading to increased proliferation, invasion and metastasis [5-7]. Most studies have focused on expression of MAPK in tissue and correlations with survival and prognosis. Activated p38 (MAPK14) has been associated with poorer survival after diagnosis with breast cancer [7]. MAPK1 and MAPK3 (alias p41 and p44 in ERK pathway) expression in breast tumors also has been associated with disease-free survival[21]. Genetic variation in MAPK genes has not been evaluated with survival after breast cancer diagnosis, although biological plausibility for associations exists. Genes in the ERK and JNK pathways were most frequently associated with survival in our study and associations were stronger for women with low NA ancestry. It is possible that women with lower NA ancestry are exposed to a wider range of environmental stress that may activate these pathways, or that they have a genetic profile that alters expression of other genes activated by this pathway. The only gene associated with higher NA ancestry, MAP3K2 (in the ERK pathway) also interacted with a history of diabetes. It is of interest to note that Native American women have high rates of diabetes, which could provide support for this association for both factors.

Since the majority of women included in this study were diagnosed at the local or regional disease stage, assessment of factors that may influence survival is important for clinical recommendations and possible interventions. Other studies have evaluated associations between alcohol and diet and risk of dying from breast cancer and other causes[22, 23]. Diet quality indices were not associated with breast cancer survival in a study of postmenopausal women[24]. Alcohol was not associated with all-cause mortality, but was associated with breast cancer-specific mortality in a cohort of 1,897 women diagnosed at an early-stage breast cancer[23]. Although we observed a slight inverse association for low consumption of alcohol for all-cause mortality, there was no association for breast cancer-specific mortality. This could be the result of low levels of alcohol consumption in our study population. However, the California Breast Cancer Survivorship Consortium, which included over 2000 Latina women observed similar associations between alcohol consumption and all-cause mortality (HR 0.88) [25] as we did. The associations we observed for dietary factors are most similar to those reported by Kwan [22]. While Kwan evaluated dietary patterns and observed an inverse association with a prudent diet and an increased risk of dying with a Western diet, we observed inverse associations between vitamin C, folate, and beta carotene, and direct associations with all-cause mortality for total caloric intake.

An important component of this study is our assessment of the interaction between genes and diet and lifestyle. To maximize power, we assessed all-cause mortality for all women combined. We observed several associations, suggesting that diet and lifestyle factors can further influence survival by modifying risk associated with genetic factors. DUSP4 and MAPK14 (also known as p38) interacted with alcohol, calories, fat, fiber, folate, and cigarette smoking to alter survival; however neither of these genes displayed an independent association with survival. Unfortunately we were not able to look at these interactions by genetic ancestry groups, which could have revealed other unique interactions.

This study represents one of the largest studies to date to report on survival among Hispanic/NA women; however, there is not a similar population available for validation of our findings. Because our population of NA women was restricted to those not living on reservations and included few NA women, we were unable to evaluate these two groups of women separately. However to obtain a better understanding of the associations with NA ancestry, we assessed associations by genetic ancestry. Based on 104 ancestry informative markers, we were able to classify ancestry among all subjects. We have extensive genetic and lifestyle data to assess associations with survival, although some exposures such as alcohol were rare in the study population. Although we were able to assess associations with all-cause and breast cancer mortality, our analysis of all-cause mortality was more robust given the larger sample size. We focused on a unique genetic pathway relevant for breast cancer survival; however, other genes or SNPs within the pathway could also be important predictors of mortality. A limitation is the general unknown functionality of these tagging SNPs. While we have information on disease stage for women, we lack detailed information on complete treatment, although we believe that adjustment for disease stage may help overcome this limitation if women diagnosed at a similar stage and tumor type would receive similar treatment. However, we are unable to evaluate associations by type of treatment, which could be informative. As a whole, we believe that this is a unique and informative population, warranting replication.

Our study supports an association between MAPK genes and survival after diagnosis with breast cancer, with the strongest associations found among women with low NA ancestry. Diet and lifestyle factors importantly modified risk associated with these genes as well as being associated with survival independently. The interaction between genetic variation in the MAPK pathway with diet and lifestyle factors supports the important role of diet and lifestyle on breast cancer survivorship