Site-specific metabolic phenotypes in metastatic breast cancer

Breast cancer has a high morbidity and mortality, mainly because it can easily metastasize to distant organs. The main metastatic sites from breast cancer are the lung, bone, brain, and liver [1],[2]. However, most studies have focused on bone and brain metastases [3]-[8]. The main mechanism of tumor metastasis is the reciprocal interaction between tumor cells and the host tissue, involving cell adhesion, proteolysis, invasion, and angiogenesis [2],[9]. Because different cancers display distinct metastatic patterns, the seed and soil hypothesis has been proposed, which dictates that the specific tumor (seed) can survive only in specific visceral organs (soil) [10]. Accordingly, metastatic breast cancer cells show different characteristics according to the metastatic site. For example, brain metastasis is associated with young age, estrogen receptor (ER) negativity, prior lung metastasis, HER-2 overexpression, EGFR overexpression, and the basal subtype [5]-[7], while bone metastasis is associated with lower histologic grade, ER positivity, ER positivity/progesterone receptor (PR) negativity, strand growth pattern, and the presence of fibrotic foci in invasive ductal carcinoma [4],[11],[12]. Therefore, metastatic breast cancer is also likely to display distinct characteristics according to metastatic site.

According to the Warburg effect theory, while normal cells gain energy from oxidative phosphorylation, cancer cells obtain energy from glycolysis, making glycolysis an important component in cancer metabolism [13]. However, this theory cannot fully explain the energy usage of all cancer cells [14]. Glutamine and mitochondrial metabolism, along with glucose metabolism, are also important components in cancer cell metabolism. Tumor cells under active glycolysis have higher levels of glycolytic intermediates, and the metabolism of glycolytic intermediates has been recently shown to be involved in tumorigenesis. A representative metabolic pathway of glycolytic intermediates is the glycine and serine metabolic pathway [15]-[18], which has been recently studied as a new possible target for tumor therapy [19]. Targeted therapy can be used in metastatic cancer, as well as in primary cancer, making the identification of metabolic phenotypes in metastatic cancer clinically important. However, metastatic cancer displays distinct characteristics according to metastatic site, but the site-specific metabolic features have not yet been fully identified. The purpose of this study was to examine the expression of metabolism-related proteins according to their metastatic site in metastatic breast cancer and their implication.

In this study, differences in metabolic features of tumors were seen according to metastatic site among cases of metastatic breast cancer. Briefly, brain and lung metastasis showed higher expression levels of glycolysis-related proteins (Glut-1, hexokinase II, CAIX, and MCT4) than did bone and liver metastasis. Thus, there are several possible reasons for the different metabolic features according to metastatic site. First, the molecular subtypes showed site specificity in metastatic breast cancer, with a high proportion of TNBC in brain and lung metastases and luminal A in bone and liver metastases. In previous studies, the expression levels of glycolysis-related proteins were higher in TNBC or basal-like type [25],[26]. Such data are in agreement with the results from this study that the expression of glycolysis-related proteins is higher in brain and lung metastases, which consist of more TNBC cases. Another possible reason is the variety of influences from circulation tumor cells (CTC). Given that CTC, defined as cancer cells in the blood of cancer patients, plays a significant role in the metastatic process, CTC and its metabolites have an influence on metastatic properties. This may result in site specificity according to metastatic site; however, further study is required [27]. One other reason is the different metabolic characteristics of metastatic sites. For example, given that bone tissue creates the hematopoietic cells, the metabolites of bone tissue are expected to differ from those of liver, brain, or lung. This is supported by the fact that immune-responsive tissue and immune-privileged tissues are reported to show different cellular compositions, as well as different metabolic and immunological responses [28].

The stromal expression levels of hexokinase II, MCT4, GLS1, GDH, SDHA, PHGDH, and SHMT1 were higher in bone metastasis compared to that at other sites. In terms of histologic features, bone metastasis forms highly rich tumor stroma with prominent fibroblasts. The reverse-Warburg effect theory, which describes the metabolic interaction between tumor cells and the stroma, supports the expression of metabolism-related proteins in the stroma in bone metastasis. The theory insists that lactate created by glycolysis in the stroma is transferred to tumor cells and metabolized as the substrate by oxidative phosphorylation in tumor cells [29],[30]. Therefore, the reverse Warburg effect phenotype may be applicable to bone metastasis, in which glycolysis-related molecules or glycolytic metabolism intermediates are highly expressed. In previous studies, luminal type tumors were more likely to have the reverse Warburg effect phenotype [31]. This may explain why metastatic tumors with the reverse Warburg effect phenotype are more likely to occur in the bone, since luminal type tumors are most commonly found in the bone. However, further validation studies are needed to confirm these findings.

Primary tumors were negative for Glut-1 and MCT4, but some positivity was seen in lung metastasis. Previous studies reported differential expression of most important biomarkers of breast cancer (ER, PR, and HER-2) between primary cancer and metastatic cancer, with 21–50% showing HER-2 loss, about 30% showing HER-2 gain [8],[32], 3.2–44% showing ER loss [33]-[35], 24% showing PR loss [35], and ER or PR gain not reported. In other words, when primary breast cancer progresses to metastatic cancer, expression of ER/PR can be lost. The loss of ER/PR may be because metastatic cancer exhibits more aggressive features compared to primary tumors; thus, ER/PR, a good prognostic marker, presents as a loss rather than as a gain. In this study, Glut-1 expression was associated with ER negativity and MCT4 expression was associated with PR negativity. In the progression of primary to metastatic cancer, ER/PR is lost and the expression of metabolism-related proteins like Glut-1 and MCT4 appear. Further validation is required to generalize the findings of this study.

The clinical significance of this study is that inhibition of the metabolic pathway may be a potential treatment target. The expression of metabolism-related proteins, especially glycolysis-related proteins differed according to metastatic site. Previous preclinical studies reported that Glut-1 inhibitor [36],[37], CAIX inhibitor [38], and MCT4 inhibitor [39] suppress tumor growth in several tumor types. Thus, these proteins are possible targets for chemotherapy in brain and lung metastasis, which showed higher expression levels of glycolysis-related proteins. However, it should be noted that a compensating response may appear if one or two molecules are inhibited in metabolic pathway targeted therapy [40].

In conclusion, differences in metabolic features according to metastatic site were seen in metastatic breast cancer, with the glycolysis phenotype found predominantly in the brain and lung since the expression of glycolysis-related protein was higher and the non-glycolysis phenotype in the bone and liver.