Breast cancer is the most common type of cancer in women, with approximately 2.6 million cases diagnosed annually (Sung et al., 2021). Current therapeutic approaches to treat advanced breast cancer with distant organ metastases are not considered effective, resulting in very low patient survival rates. This strongly contrasts with early-stage non-metastatic disease where the available therapies are able to cure ∼70–80% of patients (Harbeck et al., 2019). Presently, treatment decisions take into account the high level of molecular heterogeneity of these tumours, consequently increasing patient life expectancy. In this context, the molecular heterogeneity of breast cancers should be taken into account when developing new therapeutic approaches (Testa et al., 2020).

The human Lysyl oxidase (LOX), and lysyl oxidase like-1 to 4 (LOXL1–LOXL4) belong to the lysyl oxidase family. The primary function of these enzymes is to catalyse the cross-linking of elastin and collagen in the extracellular matrix (ECM) (Rucker et al., 1998). Altered expression of genes from this family influences the remodelling of ECM components and consequently tissue stiffness. Therefore, these proteins can impact cancer cell proliferation, survival, invasion, migration, epithelial to mesenchymal transition, ultimately influencing tumour progression (Vallet et al., 2021). This is supported by data linking expression dysregulation of LOX enzymes with metastasis and tumour survival rates in some cancers (Ahn et al., 2013; Cox et al., 2015; Salvador et al., 2017; Yu et al., 2020; Ferreira et al., 2021). Consequently, LOX and LOXL1–4 enzymes have been suggested as potential druggable targets to prevent breast cancer metastasis (Cox et al., 2016; Ferreira et al., 2021). Various compounds have been developed with different inhibitory activities against LOXs (Ferreira et al., 2021). Depending on their structure, the inhibitors developed so far can be specific for LOXL2, dual inhibitors for LOX/LOXL2, dual inhibitors for LOXL2/LOXL3, or pan-LOX inhibitors (Ferreira et al., 2021). However, the importance of selectively inhibiting each of these enzymes in the various breast cancer subtypes is not clear.

Here we use The Cancer Genome Atlas (TCGA) Breast Invasive Carcinoma (BRCA) data (https://portal.gdc.cancer.gov/projects/TCGA-BRCA) to explore the relation between the expression of each LOX gene at the mRNA level with breast cancer patient survival, expression of genes involved in cellular mechanisms associated with cancer progression and tumour infiltrates in several breast cancer subtypes. Thus, providing a basis to clarify the potential usefulness of each LOX as progression biomarker or therapeutic target in breast cancer.

To explore the relevance of each LOX gene in BRCA and its subtypes [Basal-like, human epidermal growth factor receptor 2+ (HER2+), Luminal A (LumA) and Luminal B (LumB)], a bioinformatic approach was followed using the TCGA BRCA data set for mRNA expression. The expression profile of LOX family enzymes in BRCA and its subtypes and the association of this expression with patient survival were obtained using GEPIA2–Gene Expression Profiling Interactive Analysis (http://gepia2.cancer-pku.cn/#analysis) (Tang et al., 2019). Differential expression between normal and tumour tissues was determined by a one-way ANOVA test. To explore the association between LOXs individual expression levels and the prognosis of BRCA patients, Kaplan-Meier survival analysis with Log-rank test, and Cox Proportional Hazard (PH) Model were generated using the median cut-offs of tumour LOXs mRNA levels for both overall survival (OS) and disease-free survival (DFS). Cox proportional hazard ratios (HR) from low and high expressions were plotted in a heatmap. Differences between Kaplan-Meier curves were analysed by the Log rank test.

To study the relation between the expression of LOXs and other genes, the one hundred genes with the highest Pearson correlation for each LOX were obtained from the TCGA BRCA data set using both GEPIA2 and UALCAN, a web-portal to perform in-depth analyses of TCGA gene expression data (Chandrashekar et al., 2017) (http://ualcan.path.uab.edu). Non-protein coding mRNAs were excluded from this analysis and only genes commonly found in both platforms were included. To compare the correlation of each selected gene with the remaining LOXs, Pearson´s correlation values were obtained from GEPIA2. Genecard and Uniprot were used to categorize each selected gene based on their functional characterization. Genes were clustered into four groups: 1) cell migration, adhesion, and ECM regulation; 2) cell survival and proliferation; 3) angiogenesis and tumour proliferation; 4) others (other functions).

The association between expression of LOXs and abundance of immune infiltrates in BRCA and its subtypes was evaluated using the TCGA data-based platform TIMER2.0 (http://timer.cistrome.org/) (Li et al., 2020). Spearman’s correlations were calculated based on the algorithm EPIC (Estimating the Proportion of Immune and Cancer cells). This bioinformatic tool predicts the fraction of different cell types from bulk tumour gene expression data, by integrating gene expression profiles from each major non-malignant cell type and renormalizing based on cell-type-specific mRNA content (Racle et al., 2017). The Spearman correlations coefficients obtained were plotted in a heatmap.

Many of the molecular mechanisms involved in breast cancer invasion and metastasis are still unclear. Considering the impact of breast cancer metastasis in patient outcome, increased therapeutic specificity that accounts for the molecular heterogeneity of the tumour is highly desirable. In that sense, LOX and LOXL1-4 are interesting candidates as novel targets to modulate breast cancer progression. Data presented here highlight the importance of Lysyl oxidases gene expression and its association with breast cancer patient survival and relapse. Specifically, the increased expression of LOX, LOXL1 and LOXL2 appears to be correlated with similar trends in terms of patient survival, tumour infiltrates and correlation with expression of genes involved in tumour-related processes. Contrarily to LOXL3, for which strong correlations were not found, or to LOXL4 that was associated with opposite trends. Previous studies have demonstrated that, contrarily to LOXL2, LOXL4 overexpression has an inhibitory effect on cancer proliferation and progression-related events in different cancer models (Wu et al., 2007; Choi et al., 2017; Shao et al., 2019). Thus, the results obtained are in agreement with the limited available data describing the impact of LOXs in cancer.

Several authors have proposed the LOX family proteins, mostly LOXL2 as therapeutic targets in breast cancer treatment (Moreno-Bueno et al., 2011), in accordance with the data presented here. The current targeting strategy focuses on inhibiting the enzymatic activity of LOX proteins. The small molecule inhibitors developed so far present distinct selectivity towards different enzymes of the LOXs family. However, a rational basis to pursue a specific selectivity profile in the drug development process, tailored for each potential therapeutic use, is still missing. The work presented here contributes to fill this gap for the case of breast cancer. Overall, we speculate that while the impact of LOXL3 inhibition may vary with breast cancer subtype, the specific therapeutical inhibition of both LOXL1 and LOXL2 but not of LOXL4 may be beneficial in breast cancer. Therefore, these data provide a rational basis for the drug development process of novel LOXs inhibitors aimed for breast cancer treatment.