Novel Markers for Disease Outcome in Breast Cancer

Ambs, Stefan

Abstract

In the past, most molecular epidemiology studies examined the association between genotypes and cancer risk. Relatively few studies have explored the association between common genetic variations and disease outcome, although the concept of gene-environment interactions strongly suggests that genotypes may influence disease outcome because of their modifying effects on tumor biology and therapeutic outcome. It is well known that large differences exist in patient response to treatment. Several recent studies showed that polymorphisms influence therapy response and breast cancer survival, but more studies are needed to define the relative contribution of gene polymorphisms to poor outcome after therapy. Some of these polymorphisms could become biomarkers to guide cancer therapy. A functional single nucleotide polymorphism (SNP) in the manganese superoxide dismutase (SOD2) gene, characterized by a T to C transition in codon 16 (rs4880), leads to a substitution of valine by alanine (Val16Ala). This substitition is common and 20% to 25% of the general population have the Ala/Ala genotype. Carriers of the Ala allele were found to have an increased risk of breast and prostate cancer if they consume low amounts of antioxidants. We studied the association of the Val16Ala genotype with breast cancer survival in two independent populations, one from the greater Baltimore, Maryland area (n=244) and one from Norway (n=329). We observed an association between the Ala allele and poor disease outcome in both patient populations with an additive effect of the Ala allele. In the combined analysis, homozygous carriers of the Ala allele had twice the risk of a poor outcome when compared with homozygous carriers of the Val allele (adjusted HR = 2.19; 95% CI, 1.40-3.34). Stratification by therapy type showed that homozygous carriers of the Ala allele had a greater risk of poor survival, when compared with homozygous carriers of the Val allele, when they received therapies containing either adriamycin or cyclophosphamide. The Ala allele was not associated with disease outcome among those patients who did not receive chemotherapy. Our preliminary results indicate that the SOD2 Val16Ala genotype is a potential predictor of response to chemotherapy. Inducible nitric oxide synthase (NOS2) is a signature gene of the inflammation response and has key functions in host defense. A major physiological role of this enzyme is the release of nitric oxide to support the wound healing process. It has been hypothesized that the wound healing properties of nitric oxide could turn NOS2 into an oncogene that promotes the metastatic spread of human cancer. Nitric oxide (NO) may also select for a mutant p53 tumor status and activate oncogenic pathways such the Akt and HIF1 pathways, leading to increased cell survival and resistance to therapy. We hypothesized that NOS2 expression leads to poor survival in breast cancer by activating these pathways, and by inducing a poor outcome gene signature. Previous research has shown that NOS2 is expressed in breast tumors and is associated with poor outcome markers.
We aim ed our research to understand the function of NOS2 in estrogen receptor (ER)-negative breast cancer because previous data from our laboratory indicated that these tumors have an inflammation signature. We examined immunohistochemically the expression of NOS2 in 248 surgically resected tumors with follow-up for survival. NOS2 protein was moderately to strongly expressed in 173 of the 248 breast tumors (70%). Because of the previously observed relationships between NOS2 expression and tumor p53 status, tumor grade, and tumor angiogenesis, we examined those relationships in our samples set. Consistent with the previous reports, NOS2 expression was found to significantly correlate with an increased p53 mutation frequency, high tumor grade, and increased microvessel density. We next examined the effect of NOS2 expression on predicting patient survival, and how this would be affected by the tumor estrogen receptor (ER) status. While NOS2 was not associated with breast cancer survival in the unstratified analysis, we found that the estrogen receptor status modified the association between NOS2 and breast cancer survival, with high NOS2 expression being significantly associated with poor survival in estrogen receptor-negative breast cancer patients, but not estrogen receptor-positive breast cancer patients. We further investigated why NOS2 is associated with poor survival in estrogen receptor-negative but not estrogen receptor-positive breast cancer and analyzed the gene expression profiles of 32 microdissected breast tumors (8 NOS2 low/ER-negative; 9 NOS2 low/ER-positive; 9 NOS2 high/ER-negative; 6 NOS2 high/ER-positive). We could not generate a list of differentially expressed genes at an acceptable false-discovery rate for the estrogen receptor-positive tumors, suggesting that NOS2 may not produce a gene signature in these tumors. In contrast, we found that 49 transcripts corresponding to 44 genes were differentially expressed in estrogen receptor-negative tumors. Among the genes most highly up-regulated were the cytokeratins 6 and 17, and P-cadherin, which are marker genes of the basal-like breast cancer phenotype. To determine whether this gene signature had further links with basal-like breast cancer, we examined two previously published basal-like breast cancer gene signatures for similarities. Cross referencing the NOS2/estrogen receptor-negative gene signature with these data revealed that many transcripts (19/44) in the NOS2 signature have previously been identified as being associated with basal-like breast cancer. We conducted additional experiments to further determine whether the gene expression profile of NOS2 in estrogen receptor-negative breast tumor is at least partly caused by nitric oxide and examined the effects of nitric oxide on the protein expression of four genes: interleukin-8 (IL-8), S100 calcium binding protein A8 (S100A8) and P-cadherin (CDH3), which are both markers of basal-like breast cancer, and the hyaluronate receptor (CD44), which is a marker of breast cancer stem cells. Using the slow release nitric oxide donor, DETA/NO, we found that nitric oxide induces IL-8, S100A8, CDH-3 and CD44 protein expression in the estrogen receptor-negative breast cancer cell lines, but not in the estrogen receptor-positive cell lines, suggesting that these effects of nitric oxide in breast cancer are restricted to estrogen receptor-negative tumors. Both IL-8 and S100A8 are known poor outcome markers for breast cancer. This project is currently being completed with additional experiments that examine the effect of nitric oxide on the phenotype of estrogen receptor-negative and estrogen receptor-positive breast cancer cells, and whether expression of the estrogen receptor in ER-negative cell lines can reverse the nitric oxide-induced phenotype.