ERα signaling is considered a defining and driving event contributing to ER + BC carcinogenesis; ERα overexpression in primary tumors has been linked to disease progression, influencing patient survival (33–35). Nonetheless, approximately 50% of patients with ER + BC fail to respond to endocrine therapy (36). Several reports have shown the intricate relation between response to therapy and TME components, such as fibroblasts (37–39) and ECM components (40,41). Therefore, it is paramount to define the biological determinants of ERα intra-tumoral heterogeneity and the mechanisms underlying therapeutic resistance. However, this knowledge has been hampered by the challenges in developing experimental models recapitulative of intra-tumoral ERα heterogeneity and in which ERα signaling is sustained, essential to address long-term effects of tumor-stromal interactions in ERα signaling and drug response mechanisms against ER.
Here, we propose a culture strategy in which patient-derived tissue microstructures retain ERα + carcinoma cells for at least one month of culture; of note, these cells still respond to ER stimulation and inhibition, therefore constituting a functional ex vivo model of ER-positive BC. Tissue microstructures that were entrapped in alginate capsules and cultured under dynamic conditions maintained high cellularity, low levels of tumor cell proliferation, as reported for human ER + BC (42) and parental tissue architecture (including epithelial, stromal and endothelial cell compartments and deposited fibrillar collagen). Although all interrogation was limited to one month of culture, as we have not detected signs of explant decline in cell viability up to that timepoint, we conjecture that the lifespan of encapsulated tissue microstructures could be extended for even longer periods.
We hypothesized that using tissue microstructures within the millimeter size range would be more favorable to attain an accurate representation of intra-tumoral heterogeneity and TME, than more miniaturized ex vivo models. To overcome the major limitations of ex vivo cultures – the reduced lifespan and zonation due to diffusional gradients (43), we resourced to dynamic culture conditions. Agitation improves mass transfer, promoting nutrient and oxygen diffusion, reducing the formation of gradients typically observed for tissue microstructures within the above mentioned size range (44,45). Moreover, we encapsulated in alginate, a biocompatible, inert hydrogel (46) since it has defined composition and confers support and protection from agitation-induced shear stress (15,47,48). This contributes to the preservation of tissue architecture and cell viability, but also promotes the built-up of relevant cell microenvironment factors. In fact, we have previously shown that cells entrapped in alginate capsules, and cultured under agitation, accumulate secreted soluble factors (e.g., cytokines) and ECM components, promoting homotypic and heterotypic cellular crosstalk, cell migration and reconstruction of cancer-related microenvironments (14,15), such as an immunosuppressive microenvironment in a non-small cell lung cancer model (14). In terms of ECM components, we not only observed the maintenance of TME cellular components in the encapsulated tissue microstructures, such as the stromal cells, which are involved in the secretion of collagen (49), but also ECM components as collagen fibers. These were detected by SHG microscopy, a technique broadly applied to BC tissue (50). In all the encapsulated tissue microstructures analyzed fibrillar collagen presence was observed. Increased collagen density has been shown to directly promote BC tumorigenesis (51). Moreover, collagen is strongly associated with mammographic density used as a measurement of risk of BC (52) and is responsible for drug resistance since it prevents the penetration of therapeutic agents, such as antibodies (53).
The preservation of tumor heterogeneity and TME are critical to closely mimic the in vivo situation (5,54). We observed a high degree of heterogeneity between distinct parental tissues - not only the levels of ER-positivity were different, but also the percentages and physical distribution of carcinoma and stromal cells - that were recapitulated in the derived tissue microstructures. In 5 out of 8 tissue microstructures, infiltrating immune cells were also detected, even after one month of culture, although in very low amounts. This is in accordance with the typically low frequency of immune cell infiltrates in ERα + tumors (55).
After 1 month in culture, p63 was not detected in tissue microstructures, in accordance with what is reported for luminal BC. In fact, the myoepithelial marker p63 is present in basal cells of a variety of healthy epithelial tissues (56), such as in normal breast tissue. However, its expression in BC is rare (57,58). On the other hand, tissue microstructures presented low levels of Ki-67; in fact, ERα-positive subtypes have lower proliferative indexes than other BC subtypes (59). The intrinsic low levels of proliferation and the reduced amount of patient tissue available to set-up tumor microtissue cultures, limit their application in high throughput assays.
The maintenance of ERα + cells in culture is a major accomplishment, as ERα ablation ex vivo has been a major issue in ER + BC research (60). The maintenance of ERα is pivotal for the study of the luminal A BC subtype, as cell proliferation is ER-dependent and targeted therapies typically rely on prolonged treatment with ERα antagonists (61). After one month in culture, we detected ERα + cells in the encapsulated tissue microstructures. ERα functionality was evaluated by challenging encapsulated tissue microstructures, with either activator (17-β-estradiol) or inhibitor (fulvestrant) molecules. Our results show differential expression of PGR, AREG and pS2 in tissue microstructures originated from different ER + BC patients, suggesting that the model reflects inter-patient heterogeneity and differential reactivity and signaling activation in response to 17-β-estradiol challenge (5,62). pS2 is a well-known direct downstream ERα target, which is under the positive control of an ERE consensus sequence located 400 bp before transcription starting site (63). Our results show a higher upregulation of pS2 when comparing with AREG and PGR. In fact, it has been reported for the ERα + MCF-7 BC cell line that, upon estrogen exposure, pS2 expression strongly increases compared to PR, not only at mRNA but also at protein levels (64,65).
Aiming to retain ERα + cells, we employed a culture medium enriched in molecules with reported ER stimulatory effects, such as insulin, hydrocortisone and EGF (17–22). 17-β-estradiol and EGF may also be produced by the breast fibroblasts present in culture (66–68). Our observation of reduced effects upon 17-β-estradiol stimulation in tissue microstructures cultured in complete medium compared with tissue microstructures cultured in depleted medium in the 3 days preceding stimulation, corroborates the presence of soluble ER activators in culture. Further studies are required to understand the signaling events that contribute to the maintenance of ERα signaling under the culture conditions here presented, which will potentially also contribute to further disclose its role in ER + BC.