Expression levels of serine/glycine metabolism-related proteins in triple negative breast cancer tissues

To evaluate the expression levels of serine/glycine metabolism-related proteins (PHGDH, PSAT, PSPH, SHMT, and GLDC) in six different subtypes of triple negative breast cancer (TNBC) patients and gain insight into their implications. Formalin-fixed, paraffin-embedded tissues from 129 TNBC patients were assembled into tissue microarrays. Immunohistochemical staining was performed for serine/glycine metabolism-related proteins (PHGDH, PSAT, PSPH, SHMT, and GLDC) and surrogate immunohistochemical markers (CK5/6, EGFR, claudin 3, claudin 4, claudin 7, E-cadherin, STAT1, interleukin-8 [IL-8], AR, and GGT-1) for identifying the molecular subtype of TNBC. TNBC subtype classifications included the following: basal-like (CK5/6-positive and/or EGFR-positive), molecular apocrine (AR-positive and/or GGT-1-positive), claudin-low (claudin 3-, claudin 4-, claudin 7-negative and/or E-cadherin-negative), immune-related (IL-8-negative and stromal STAT1-positive), mixed (features from two or more of the four subtypes), and null (no features from any of the four subtypes). Tissues from basal marker-positive patients showed increased expression levels of tumoral PHGDH compared with those from basal marker-negative patients (p=0.029); lack of stromal SHMT1 expression was significantly correlated with T stage (p=0.016). Multivariate Cox analysis revealed that a lack of stromal SHMT1 expression was an independent prognostic factor for predicting a shorter disease-free survival period (hazard ratio 4.002, 95% confidence interval [CI] 1.077-14.83, p=0.038); furthermore, a lack of tumoral PHGDH expression was predictive of a shorter overall survival rate (hazard ratio 3.053, 95% CI 1.002-9.305, p=0.050). In conclusion, the most abundantly expressed serine/glycine metabolism-related protein in basal-like TNBC tissues was tumoral PHGDH, and expression levels of stromal SHMT1 and tumoral PHGDH were inversely correlated with clinical prognostic factors. Also, this study is the first to assess serine/glycine relationships at the protein level in regards to clinical outcomes.