Organoids have been used extensively as in vitro models to investigate the development and physiology of normal organs and diseases including cancer. However, improvements to current breast organoid culture methodologies are needed to improve their utility. Here, we describe a modified organoid culture system for more effective expansion of normal breast organoids that maintain bilayered organoid morphology, lineage fidelity, and hormone receptor expression and responsiveness. As a result, our method represents a powerful tool to study estrogen and estrogen receptor regulation in normal tissues and ER positive breast cancer.
The conventional organoid culture system embedding organoids inside Matrigel domes was designed to recapitulate the in vivo extracellular matrix environment. However, passage of conventional dome cultures requires steps to break down Matrigel mechanically or enzymatically to separate organoids from mixtures. This physical stress could negatively affect organoid growth rate because organoids need more time to recover and reorganize following the harsher dissociation steps (ref. 10, 12, 31–33). Moreover, the solid Matrigel dome culture system for normal breast organoids poses many challenges, such as requiring tedious steps for organoid propagation and failure to maintain bilayered organoid structures over time (ref. 12, 14, 34). Our method using organoid culture medium supplemented with a low percentage of Matrigel (5%) simplifies the culture procedures for easier handling while still providing the sufficient extracellular matrix for organoid formation. Although the ratio of three epithelial lineages (HS, LASP and BA) was not significantly distinguished in dome or suspension cultures, organoids formed in suspension were larger and formed more uniform sphere structures than those in Matrigel domes. The differences in structure size may be due to the higher proportion of proliferating cells in suspension cultures (Fig. 2c, e). It is also possible that suspension conditions allow cells to more efficiently migrate and aggregate compared to dome embedding. The maintenance of a bilayered morphology in suspension culture, with BA cells surrounding the luminal cells, could also contribute to the more uniform sphere structures (Supplementary Fig. 1d).
Bulk RNA seq analysis of the three breast epithelial cell types indicated that these three lineages maintain their cell identities in both types of organoid cultures. We detected minimal differences in cell identity signature genes in organoids cultured in domes or suspension. HS cells isolated from organoids in both culture conditions displayed higher lineage fidelity than LASP and BA cells (Fig. 3a). LASP cell populations expressed a few HS and BA genes; however, this is predicted based on the known presence of cell populations that display reduced lineage fidelity in human breasts. These cells, referred to as basal-luminal or BL cells, are enriched for genes specifically associated with BA or HS cells as well as genes that are expressed at very low levels or not at all in breast epithelial cells (BL-unique genes (ref. 15, 20, 21). The heatmap of BL-unique signature genes in Supp. Figure 3a indicates that the organoid cultures maintain BL cells. Thus, some of the BA (KRT17, PTN, SPP1) and HS (FAM102A, HIGD1A, WFDC2) genes that are expressed in LASP cells are likely due to the presence of a BL cell population in this cluster. Other BA genes expressed in LASP cells could be derived from BA cells that have undergone partial differentiation, sufficient for these cells to be sorted with LASP cells (e.g. decreased CD49f and increased EpCAM). Likewise, BA cells expressing LASP markers could have undergone a partial differentiation to LASP cells, without affecting CD49f/EpCAM sorting. BA cells have been reported to express LASP markers after being dissociated from luminal cells which are critical to maintain BA cell identity (6, 35, 36).
Importantly, we were able to maintain purified HS cell organoids in suspension cultures while still maintaining extremely high purity after short-term (Supplementary Fig. 4a) or long-term culture (Fig. 4b-d). For a few HS organoid cultures, we did detect small percentages of LASP and BA cells. This could result from insufficient gating of HS cells during FACS using CD49f and EpCAM. Use of an additional sorting step or inclusion of more surface markers such CD166 for HS cells (ref. 17, 35), CD133 for LASP cells (ref. 15, 37) and CD10 for BA cells (ref. 38) could also increase the purity of HS cells. Overall, we confirmed that HS cells can be isolated and cultured alone without losing cell identity (Supplementary Fig. 4a-b).
We found that the purified HS cells maintain expression of ER and respond to estrogen stimulation after long-term culture (Fig. 5c). Progesterone receptor induction of its target gene WNT4 was only detected when organoids were cultured with additional TGFβ inhibitors (Supplementary Fig. 5a), consistent with previous studies using other medium conditions (28, 29).
One issue that all breast organoid investigations have reported is the heterogeneity of organoids from different donors. As observed previously (ref. 12, 14), we detected widely varying proportions of the three main cell lineages in organoid cultures from different donors. There are several factors that may account for this variation, such as age, menopause status, BMI, or genetic background (e.g., BRCA mutations). This heterogeneity makes it very difficult to identify phenotypic differences in organoids that can be attributed to the factors mentioned above. However, the proportions of each lineage from a single organoid line are maintained for several passages (ref. 12, 14), thus making it feasible to use genetic or pharmacological perturbations within the same organoid lines to address mechanistic questions.
In conclusion, our results demonstrate that organoid suspension cultures represent a valuable in vitro platform to study ER and estrogen regulation in HS cells from normal breast, enabling studies of ER-positive breast cancer initiation that were not previously feasible.