The Yin and Yang of Breast Cancer: Ion Channels as Determinants of Left-Right Functional Differences

Breast cancer is a complex and heterogeneous disease that displays diverse molecular subtypes and clinical outcomes. Although it is known that the location of tumors can affect their biological behavior, the underlying mechanisms are not fully understood. In our previous study, we found a differential methylation profile and membrane potential between left (L)- and right (R)-sided breast tumors. In this current study, we aimed to identify the ion channels responsible for this phenomenon and determine any associated phenotypic features. To achieve this, experiments were conducted in mammary tumors in mice, human patient samples, and with data from public datasets. The results revealed that L-sided tumors have a more depolarized state than R-sided. We identified a 6-ion channel-gene signature (CACNA1C, CACNA2D2, CACNB2, KCNJ11, SCN3A, and SCN3B) associated with the side: L-tumors exhibit lower expression levels than R-tumors. Additionally, in silico analyses show that the signature correlates inversely with DNA methylation writers and with key biological processes involved in cancer progression, such as proliferation and stemness. The signature also correlates inversely with patient survival rates. In an in vivo mouse model, we confirmed that KI67 and CD44 markers were increased in L-sided tumors and a similar tendency for KI67 was found in patient L-tumors. Overall, this study provides new insights into the potential impact of anatomical location on breast cancer biology and highlights the need for further investigation into possible differential treatment options.

Ion channels (ICH) are transmembrane proteins that regulate the flow of ions across cell membranes, thereby modulating membrane potential, intracellular signaling, and gene expression. Rather than being confined to the nervous systems alone, these abilities are shared by all cells and are considered adaptations of an ancient bioelectric communication system that employed similar strategies to tackle challenges in the past [1]. The ability of cell groups to approach a specific morphology is largely reliant on bioelectric communication across tissues. This information processing occurs through the utilization of identical, well-conserved molecular components like those used by neurons [2]. ICH play crucial roles in many physiological processes, including cell proliferation, migration, invasion, and apoptosis [3]. Dysregulation of ICH can lead to altered membrane potential, intracellular signaling, and gene expression, promoting tumorigenesis and therapy resistance. Any bioelectric disruption can trigger cancerous characteristics [4,5], even in the absence of DNA damage or mutation. Essentially, when cells lose their tight informational connections with neighboring cells, they revert back to a primitive, unicellular state [6], breaking away from the multicellular group and regarding it as a mere external environment. This results in the re-emergence of behaviors appropriate to the unicellular state, such as proliferation. So cancer can be seen as primarily a disorder of establishing informational boundaries between the cellular self and the outside world [7], which has already led to the development of effective treatments that can reverse and prevent cancer in amphibians [8] and is currently undergoing testing in human tissues [9].

Emerging evidence suggests that ICH also participate in the crosstalk between the tumor microenvironment (TME) and cancer cells [10,11], contributing to tumor initiation, progression, and therapy resistance (reviewed in [12,13,14]). For example, ICH expressed by cancer cells can sense and respond to changes in the TME, such as alterations in pH, oxygen tension, and ion concentration, leading to changes in intracellular signaling and gene expression that promote tumorigenesis [3]. Moreover, ICH expressed by TME components, such as immune cells and endothelial cells, can modulate their function and secretion of cytokines and growth factors, affecting tumor cell behavior [14]. Targeting ICH that are involved in TME-tumor interaction may provide new opportunities for developing effective and specific anticancer therapies.

Breast cancer (BC) is a complex disease that exhibits heterogeneity with diverse subtypes that vary in their molecular and clinical characteristics. One aspect of this heterogeneity is the laterality of breast tumors, i.e., whether the tumor arises in the left (L) or right (R) breast [15]. Although breast tumors are generally thought to be similar regardless of laterality, several studies have suggested that the surrounding tissue of L-R breast tumors may differ [16]. The mammary glands are not identical [17,18,19]; each breast has its unique set of ducts, lobules, and supporting stroma. These structures can differ in size, shape, and branching patterns. In fact, all paired organs present physiological and molecular differences, as shown by microRNA profiling in paired eyes, lung, and testes normal tissues [19]. Therefore, the normal differences between both breasts can lead to potential differences in the microenvironment surrounding tumors impacting tumor biology differently in each breast, thus making it worthwhile to consider when studying BC.

Interestingly, L-sided BC incidence is slightly increased [20], and even though the underlying causes are unclear [21], the tendency is consistent across populations. Similar differences are observed in other paired organs which laterally differ in cancer incidence and progression [22]. Although these differences seemed not to have a relevant clinical impact [23], growing evidence from research findings indicates they are more important than was considered [24]. For example, in animal studies using mouse models of BC, it was shown that L-R-sided tumors had distinct gene expression profiles [25]. In humans, recent studies discovered that L-R-sided tumors may respond differently to specific therapies depending on the local TME, with tumors in different regions of a single organ showing different expression profiles and properties [26,27]. Furthermore, L-sided BC was found to be associated with a more aggressive biology and a worse outcome as compared to R-BC [28]. Remarkably, recent literature has raised new questions about the relationship between BC laterality and prognosis [29]. Understanding these differences may have significant implications for diagnosis and treatment.

In our previous publications [30,31], we discovered a distinct methylation profile between L-R-sided breast tumors, which we associated with differences in membrane potential and ion concentration. Integrating the facts that L-R normal breast tissues have differences in gene expression and that the bioelectricity is known to play a crucial role in tumorigenesis, we hypothesized that bioelectric differences of both sides of the breast could contribute to the development of BC in a site-specific manner. In our current work, we sought to determine the primary ICH involved in these differences and explore the potential impact on the functional characteristics of the tumor cells. By investigating the role of specific ICH in L-R tumor development, we hope to uncover novel therapeutic targets for BC treatment.

We chose to use a mouse model of breast cancer (MMTV-PyMT) which spontaneously develops mammary tumors on both the L and R side, to investigate the role of laterality in regulating the bioelectric properties of tumors. Pairwise comparisons of L-R tumors within the same mouse were possible using this model. After extracting paired tumors, we used two fluorescent probes with opposite charges to measure the bioelectric state establishing a ratio between the fluorescence of both probes (MT/DB). The paired comparison of four tumors, all located on the L-R 4th gland, revealed by confocal microscopy (Figure 1A and Supplementary Figure S1A,B) a difference in the MT/DB ratio: L-sided tumors showed a decreased MT/DB ratio (Figure 1B and Supplementary Files S1A–D).

In parallel, samples from 3/5 mice were analyzed by flow cytometry, confirming the same trend as seen by microscopy: L-sided tumors had a decreased MT/DB ratio compared to R-sided tumors (Figure 2 and Supplementary File S2).

This suggests that the L and R glands generate a difference in the bioelectric state of growing tumor cells. The consistent shift into the same direction suggests that L-sided tumors have a more depolarized state than R.

Ion channels play a major role in regulating the bioelectrical properties of tumors. To further investigate which ICH may play a relevant role, we switched to a bioinformatic approach. We obtained a list of human ICH genes from the HGNC database. The list contained 330 genes, of which the expression levels of 319 were downloaded from 781 invasive breast tumors in The Cancer Genome Atlas (TCGA) dataset, comprising 393 L-sided and 388 R-sided tumors.

To determine whether a set of ICH genes contributed in a relevant way to the L-R differences, we used the computational method Gene Set Enrichment Analysis (GSEA). We selected the four gene sets that showed significant difference, particularly a consistent decrease of their expression in L-sided tumors (p < 0.05; FDR < 0.1): WP-ENERGY-METABOLISM (collection: C2), REACTOME-PHASE-0-RAPID-DEPOLARISATION (collection: C2), REACTOME_CARDIAC-CONDUCTION (collection: C2), and CANCER-MODULES (collection: C4). By combining these gene sets, we identified the leading genes that were associated with the tumor laterality: CACNA1C, CACNA2D2, CACNB2, KCNJ11, SCN3A, and SCN3B, signature that we called: 6-ICH signature.

We first confirmed in the Human Protein Atlas (www.proteinatlas.org, accessed on 16 June 2023) that the six channels had been detected at protein level in normal breast tissue cells (glandular epithelial and mesenchymal cells). Next, we analyzed the expression of the 6-ICH signature in 101 normal vs. 1097 primary breast tumor tissues and in 571 left-sided vs. 526 right-sided breast tumors, using the Xena Functional Genomics Explorer [32] from UCSC (accessed on 16 June 2023). We obtained the expression levels of the signature from an equation where we assumed an equal contribution of each gene, i.e., = CACNA1C + CACNA2D2 + CACNB2 + KCNJ11 + SCN3A + SCN3B.

We found a decreased expression of the signature in tumoral cells as compared to normal tissue (Welch’s t-test, p < 0.001). And among the tumors, we found a decreased expression in L-sided compared to R-sided ones (Mann–Whitney test for non-normal data, p = 0.005, Supplementary File S3). Adding more genes to the signature did not significantly increase the association with laterality. However, removing genes one by one resulted in a loss of association. Interestingly, when we compared the expression levels of the individual genes of the 6-ICH signature in a one-by-one manner, we found a reduced or complete loss of association with the side for some genes. Therefore, we considered the 6-ICH signature as robust for further studies.

ICH play a role in a wide range of biological processes and can regulate opposing cellular phenotypes. Therefore, the 6-ICH signature did not suggest a clearly defined phenotypic impact. To gain insight, we used a bioinformatic approach. We dichotomized the 6-ICH signature expression data into High and Low, using the mean expression value as a cut-off, and performed a differential gene expression (DGE) analysis visualized in a volcano plot. In the Low-expressing group (as compared to the High-expressing one), a significant number of genes were up or down regulated (Figure 3A).

Next, we conducted an enrichment analysis to determine the principal biological processes the DGE were participating in. Interestingly, we found that in the Low-expressing subgroup, mitotic genes were overexpressed, and membrane potential regulating processes were less expressed (Figure 3B). These findings suggest that L-sided tumors with a low-expressing 6-ICH signature may exhibit increased mitotic activity compared to R-sided tumors.

From an in silico cohort of 1082 primary breast tumors from the Xena Functional Genomics Explorer-UCSC, we randomly generated a Discovery Cohort (533 tumors) and a Validation Cohort (549 tumors). We choose the markers “KI67” and “stemness scores” to analyze, as indicative of pro-mitotic behavior, with KI67 being a protein expressed throughout the cell cycle, and stemness scores measuring the extent to which a tissue is capable of self-renewal. So, we studied the correlation between the expression of the 6- ICH signature and the expression of KI67 and an RNA-seq based stemness score [33], both available in the TCGA Pan-cancer Atlas. In the Discovery Cohort, we observed a strong inverse correlation between the 6-ICH signature and both markers, (r = −0.47 and r = −0.62 respectively, Spearman correlation test, p < 0.0001) (Figure 4A). Receiver operating characteristic curves (ROC analyses) demonstrated an acceptable discrimination value of the 6-ICH signature to distinguish between high and low expression of KI67 (AUC = 0.74, SE = 0.02, 95% CI = 0.698 to 0.782, p < 0.0001) and high or low stemness scores (AUC = 0.79, SE = 0.01, 95% CI = 0.761 to 0.837, p < 0.0001). We confirmed these results in the Validation Cohort of 549 primary breast tumors (r = −0.51 and r = −0.62 respectively, Spearman correlation test, p < 0.0001; ROC curves for KI67: AUC = 0.74, SE = 0.02, 95% CI = 0.707 to 0.788, and for CD44: AUC = 0.81, SE = 0.01, 95% CI = 0.776 to 0.848, p < 0.0001) (Figure 4B). We also confirmed these correlations in a pan-cancer analysis on 9546 tumors (r = −0.51 and r = −0.62 respectively, Spearman correlation test, p < 0.0001) (Figure 4C).

We then analyzed the 5-year (1800 days) rates of overall survival, disease-specific survival, disease-free survival, and progression-free survival by comparing high vs. low 6- ICH signature expression in 1094 BC and in 9546 pan-cancer patients. Remarkably, all survival rates were increased in high-expressing tumors (Figure 4D).

We also examined the correlation between the 6-ICH signature and epigenetic regulators, particularly histone modifiers, methylation erasers, and methylation writers, by performing an expression signature of each group of epigenetic enzymes. We found a strong inverse correlation only between the 6-ICH signature and the signature of DNA methylation writers composed of DNMT1, DNMT3A, DNMT3B, DNMT3L, TRDMT1, and DMAP1 (r = −0.42, Spearman correlation test, p < 0.0001) (Figure 5 and Supplementary File S4).

Taken together, these results suggest that the bioelectric signature is associated with the side of the tumor and with cancer hallmarks such as proliferation and stemness markers and is also associated with regulators of DNA methylation (which are known to respond to microenvironmental factors). Moreover, it reveals an impact on the survival rate of cancer patients. The results are robust across different cohorts.

After establishing in silico that the 6-ICH signature was associated with the side and with proliferation and stemness divergences, we progressed to in vivo models in mice to confirm these findings. We utilized paired L-R tumors of four female mice and analyzed the expression of the proliferation marker KI67 and the stemness marker CD44.

Our analyses revealed a significant increase in the expression of both markers in L-sided tumors compared to R-sided tumors (paired t-test, p < 0.05) (Figure 6 and Supplementary File S5).

We also investigated whether these differences were present in human breast tumor RNA samples. We analyzed the expression of the markers KI67 (in 6 L and 5 R) and CD44 (in 5 L and 4 R) by qPCR. Although not statistically significant, it is notable that we found a similar trend: an increased mean expression of both markers in L-sided tumors (unpaired t-test, p = 0.2) (Supplementary File S6).

Taken together, our observations in mice and humans strongly suggest that L-R breast tumors are distinct in two critical cancer hallmarks, namely proliferation and stemness, which are shaped by the differential expression of ICH.

Breast cancer is a heterogeneous disease that encompasses a broad spectrum of molecular subtypes and clinical outcomes [34]. The association between the anatomical location of tumors and their biological behavior is still an open question. Diverse studies across different species have addressed the L-R issue of bilateral organs, like RNA differential profiles in mice paired eyes, lungs, and testes [19], L-R differences acquired during embryonic development [35], L-R metabolic differences in Xenopus [36], in avian [37] and mice [38] models, and in humans [24]. It has been known for many decades that breast tumors occur more frequently in L sides [39], however the causes are still unclear [17,21,24]. In our previous research, we addressed the question of whether epigenetic differences existed between L-R human and mice-xenografted breast tumors. We found differential DNA methylation profiles in genomic regions [30], among which ICH genes were included. Furthermore, we found how L and R extracts from human mammary glands were able to generate membrane potential differences in cultured cells and how L-extract treated cells developed a more depolarized state [31]. So, we proposed that the different L-R environments generated methylation differences which ended up provoking different bioelectric states in the L-R tumors. In this study, we aimed to address the L-R asymmetry question at a whole organism level, in mice and in humans. Our results show that L-sided tumors consistently exhibit a more depolarized state, an increased expression of proliferation and stemness markers, and that the lower level of ICH signature expression in L-tumors is associated with reduced survival rates, as compared to the R-sided ones (Figure 7).

This suggests that the microenvironments of the L and R sides generate a consistent difference in the bioelectric state of growing tumor cells, with a direct impact on tumor behavior. To our knowledge, prior to our research, there were no published studies informing proliferation differences between L and R breast tumors. However, a recent publication supports our findings, demonstrating (in line with our observations) that L-sided BC is associated with more aggressive biology and worse outcome as compared to R BC. As far as we know, these are the first conclusive findings differentiating breast tumors in bioelectric, proliferative, and survival rate terms, based on their side.

We identified a 6-ICH gene expression signature that was consistently associated with the tumor side. The signature is composed of 3 calcium, 2 sodium, and 1 potassium channels. CACNA1C, CACNA2D2, and CACNB2, are all subunits of voltage-gated calcium channels, which are involved in a wide range of physiological processes, including muscle contraction, neurotransmitter release, and gene expression [40]. KCNJ11, on the other hand, codes for a potassium channel that is involved in regulating insulin secretion in pancreatic beta cells [41]. Bioelectric signaling driven by potassium channels has been proposed as metastasis regulators in triple-negative breast cancer [42]. Finally, SCN3A and SCN3B are subunits of voltage-gated sodium channels, which are important for the initiation and propagation of action potentials in neurons and other excitable cells [43]. Recent findings have revealed the clinical importance of sodium channels and the ionic microenvironment of breast tumors [44]. Cell cycle checkpoints, which rely on being heavily dependent on bioelectricity, have been linked to errors that occur when the number or current of sodium channels increases [45].

Predicting the outcome of a gene expression shift in these six ICH entities, however, is challenging. For example, reduced expression of CACNA1C, CACNA2D2, and CACNB2 in cardiac myocytes could result in decreased calcium influx and reduced contractility, which could lead to a hyperpolarized state. Additionally, reduced expression of SCN3A and SCN3B in neurons could result in decreased sodium influx and reduced excitability, which could also lead to a hyperpolarized state. However, reduced expression of KCNJ11 in pancreatic beta cells could lead to decreased potassium efflux and increased insulin secretion, which could result in a depolarized state. Overall, the impact of changed expression of ICH on membrane potential and cellular excitability is complex and context-dependent and requires consideration of the specific ion channel and cell type involved. Therefore, the 6-ICH signature did not suggest a clearly defined phenotypic impact. However, what was clear was the observation that the distinctive L/R expression had an influence on membrane potentials. This phenomenon can be attributed to the fact that the membrane potential is reliant on the collective actions of ion channels and transporters operating at the cellular membrane. Our findings indicate that the less the signature expression, the more depolarized is the membrane potential. Speculating on the function of the genetic signature channels, one could propose in general terms that less function of voltage independent potassium channels (like KCNJ11) would result in membrane depolarization, which would trigger the activation of voltage gated calcium and sodium channels. This sequential process would maintain a more depolarized membrane potential involved in cell cycle progression.

Cancer cells are known to tend to be more depolarized than their normal counterparts [46]. Evidence from both rodent and human tissues has shown that fast proliferating tumor cells exhibit a depolarized membrane potential, whereas non- or less-proliferating are distinguished by their hyperpolarized membrane potential [47]. This has allowed it to be stated that depolarization serves as a signal to initiate mitosis and DNA synthesis, independently of the fluctuation of membrane potential across the cell cycle [47]. Membrane potential has also a role in normal stem cell differentiation; in particular, a hyperpolarization is required during cell maturation, while on the contrary, depolarization reduces the differentiated phenotype [48].

By bioinformatic analyses in TCGA breast tumors, the lower 6-ICH expression was revealed to be associated with increased proliferation and stemness, and with the worst survival in 5 years. Proliferation and stemness are key biological processes in tumorigenesis. Importantly, higher values for stemness indices are associated with an increased number of cancer stem cells and with a greater de-differentiated phenotype [33]. Stemness is defined as the potential of a cell for self-renewal and differentiation. Cancer progression involves gradual loss of differentiated phenotype and acquisition of stem cell-like features. Less differentiated tumors are more likely to spread to distant organs and make prognosis worse, particularly because metastases are usually resistant to available therapies [49]. Proliferation is also a critical aspect to consider when evaluating BC, and KI67 expression stands out as one of the most important and cost-effective surrogate markers for assessing tumor cell proliferation [50]. Our results suggest that the 6-ICH signature may play a role in modulating these processes in breast cancer, probably by changing the bioelectric state in a side-dependent manner.

The TME is a complex and dynamic ecosystem that plays a critical role in tumor initiation, progression, and response to therapy. It comprises various cell types, extracellular matrix components, and signaling molecules that interact with tumor cells in a bidirectional manner [51]. Growing evidence suggests that the TME can alter the epigenetic landscape of cancer cells, leading to changes in gene expression patterns that promote tumorigenesis and therapy resistance. For example, TME-derived factors, such as growth factors, cytokines, and extracellular matrix proteins, can activate oncogenic signaling pathways that promote DNA methylation, histone modifications, and chromatin remodeling in cancer cells [52]. Moreover, TME components, such as cancer-associated fibroblasts, immune cells, and endothelial cells, can secrete exosomes and micro-vesicles that transfer epigenetic modifiers, such as micro-RNAs, DNA methyltransferases, and histone deacetylases, to cancer cells, further altering their epigenetic landscape [10,11]. The impact of TME on the tumoral epigenome and its effects on tumor behavior are of great interest because epigenetic changes are reversible and potentially targetable by therapeutic agents.