Preprint: Endometrial Assembloids Show Context-Dependent Transcriptional Responses to Repurposed Drugs
A preprint compares four drugs in epithelial and stromal monolayers, cell-line assembloids and primary endometrial assembloids. Simvastatin caused the broadest expression changes, but responses depended strongly on sample type and sequencing replication was limited.
Combining Epithelial and Stromal Components
A preprint from the University of California, San Francisco, UC Davis and collaborators tests a transcriptomics-based drug-repurposing approach in human models of endometriosis. The experiments measure changes in epithelial and stromal gene-expression programs after candidate-drug exposure. The comparison includes monolayers, three-dimensional cell-line assembloids and assembloids made from patient endometrial tissue.
The candidates came from the team’s earlier computational and animal work. In the present study, epithelial 12Z cells and stromal iEc cells were combined in agarose micromolds. Histology showed an epithelial outer layer surrounding stromal cells. The authors also isolated epithelial and stromal fractions from eutopic endometrium: tissue within the uterus from patients with surgically confirmed endometriosis. These primary assembloids therefore model patient endometrial cell interactions rather than an intact ectopic lesion.
Cell composition influenced assembly. Increasing the stromal proportion improved the formation of cell-line assembloids, while the primary cultures did not follow the same relationship. Histological characterization used a 1:20 epithelial-to-stromal cell-line assembloid preparation; the drug-treatment method specifies a 1:50 cell-line preparation. The primary-tissue method separately tested epithelial-to-stromal seeding ratios of 3:1, 1:5, 1:10 and 1:20.
Four Drugs, Several Sample Types and Unequal Replication
The researchers exposed cultures for 24 hours to ibuprofen, fenoprofen, simvastatin or primaquine, with DMSO as the vehicle control. Concentrations, selected from Connectivity Map data, were 19.4, 7.2, 9.6 and 8.8 μM, respectively. Bulk RNA sequencing then characterized the response in each sample type.
The sequencing dataset contained three samples per cell-type/drug condition for the individual cell lines, one per condition for cell-line assembloids, and two per condition for primary assembloids. The latter came from two patients. Each cell-line assembloid sequencing sample pooled four wells. These denominators constrain the interpretation of differential-expression patterns, particularly in the three-dimensional cell-line system.
Sample source separated the transcriptomic profiles more strongly than drug treatment in principal-component analysis. Simvastatin nevertheless produced a distinct expression response in epithelial and stromal monolayers and the largest numbers of differentially expressed genes across the reported sample types. Primaquine had a much stronger transcriptional effect in epithelial cells than in stromal cells or primary assembloids. Both nonsteroidal anti-inflammatory drugs produced relatively few expression changes.
Pathway Reversal as a Screening Endpoint
The investigators compared drug-associated pathways with endometriosis-associated signatures from previously published human endometrial microarrays. Primaquine reversed the chemical-carcinogenesis/reactive-oxygen-species pathway signature in assembloids. Pathway reversal was assessed by gene-set enrichment analysis.
Simvastatin reversed the cytokine–cytokine-receptor interaction signature in both epithelial and stromal cell lines. In stromal cells it also reduced expression of genes associated with cytoskeletal remodeling and focal adhesion. These changes support hypotheses about inflammatory and motility-related processes.
Fenoprofen reversed a calcium-signaling signature in stromal cells. Ibuprofen did not reverse any of the dysregulated pathways assessed in this analysis. This limited transcriptional response is an informative comparator: a drug’s known analgesic use does not require a broad, detectable reversal of epithelial or stromal gene expression in a 24-hour culture experiment.
These comparisons show that the molecular response to the same compound changes with the cellular source and culture format. Simvastatin’s broad transcriptional effect and primaquine’s epithelial bias identify distinct hypotheses for follow-up in patient-derived models. The experiment used only two primary donors and one cell-line assembloid sequencing sample per condition, and measured neither invasion nor pain.