Mechano-dependent regulation of alternative splicing machinery in cancer

Introduction: Tumor progression is associated with a gradual increase in extracellular matrix (ECM) stiffness. For instance, while physiological breast tissue is soft (with a stiffness of 0.5 kPa), tumor tissue gradually stiffens as it advances to more advanced stages (>20 kPa) [1]. This stiffening plays a central role in tumor progression by facilitating invasion, migration, and dissemination of tumor cells. Our previous work has shown that certain alternative splicing (AS) events are influenced by stiffness. Alternative splicing (AS) increases protein diversity by allowing the production of different version of the same proteins with distinct functions [2, 3]. Besides, an abnormal AS profile is considered a hallmark of cancer cells: AS misregulation was shown to contribute to tumor initiation, progression, and invasion by modifying the relative expression of isoforms of various oncogenes and tumor suppressors [4]. Several splicing factors (SF) work in concert to modulate AS patterns [3]. The lab has also shown that matrix stiffening decreases the expression of a specific SF: ESRP1 [5]. It is one key factor included in a bigger network. Thus, a stiffness-dependent modulation of SFs could contribute to the change in AS profile in tumors. Our objective is to identify and characterize mechanosensitive AS events in the ECM as well as the molecular mechanisms involved in their regulation.

Materials and

Methods: Here, we utilized metastatic MDA-MB-231 and non-tumorigenic MCF10a cells. The cells were seeded onto collagen-coated polyacrylamide gels of 1 kPa and 10 kPa. Once cultures reached 80% confluency, mRNA was extracted and sequenced on a NovaSeq 6000 platform. Sequencing data were aligned using STAR, and AS events were detected and quantified based on their Percent Spliced-In (PSI) using PSI-Sigma [6]. PSI changes were calculated with signal-to-noise ratio (SNR) approach to compare the level of a signal (maximum difference between the two measurements) to the level of background noise (maximum standard deviation between replicates). The association between SF and AS events was evaluated using expression analysis (FeatureCount). De novo motif enrichment was performed with MEMEsuite to detect motif enrichments showing similarities with SF recognition motifs. A database of motifs recognized by each SF was compiled to examine the enrichments within 50 base pairs flanking the mechanosensitive AS events. To take into account heterogeneity within the cell population, we studied specific splicing events at cellular resolution using an in situ hybridization technique, Splice RCA [7]. This high-resolution strategy allows multiplexed profiling of the RNA splicing variants in single cells by targeting the splice-junction sequence.

Results, Conclusions, and Discussions: AS events were first classified by AS type (single exon skipping, intron retention, alternative 5' splicing, alternative 3' splicing or mutual exclusion of exons). We have shown that there are no differences in terms of proportion of the different AS events between the two cell lines and the two stiffness. AS events were sorted and classified based on stiffness-dependent splicing events (measured as PSI changes). 230087 AS events were identified to undergo stiffness-mediated changes in term of PSI. Among those, 3876 mechanosensitive AS events were observed in both cell lines (Figure 1A). The top stiffness-mediated splicing events were then assessed with RT-PCR to validate the sequencing data in both cell lines (Figure 1B-D). Analysis using GO Term on the commons mechanosensitive AS events reveals that impacted genes are enriched in the cadherin binding pathway in both cell lines. Of note, several of these genes are involved in the epithelial-to-mesenchymal transition. To identify the splicing machinery component that mediated these stiffness-mediated splicing events, we focused on the group of commons mechanosensitive AS events to obtain the correlation between SF protein expression and AS events (Figure 2A). Interestingly, there were clear groups of SF protein that clustered together, and which were associated with common AS events. To further test which ones were the main drivers, we performed gene expression correlation and enrichment motif analysis. The motifs found within the sequences of these events identified several SF proteins, including QK1, HNRPA1, SRSF2 and HNRPAB, whose expression correlated with increased stiffness (Figure 2B-F). Overall, this bioinformatic approach helped us dissect the splicing machinery linked to the mechanosensitive AS events. To validate our findings at the cellular level, we will use SpliceRCA in situ hybridization to visualize AS events. Our current data show that we can use this approach to quantify AS events, as exemplified by the detection of the exclusion or inclusion of PTK2 exon 4 in cells expressing GFP-PTK2 splice variants. Future experiments will assess the functional impact of SF on mechano-dependent AS and its importance in cell response to tumor stiffness.

Acknowledgements (Optional): [1] Emon B, Bauer J, Jain Y, Jung B, Saif T. Biophysics of Tumor Microenvironment and Cancer Metastasis - A Mini Review. Computational and Structural Biotechnology Journal. 1 janv 2018;16:27987.
[2] Liu Q, Fang L, Wu C. Alternative Splicing and Isoforms: From Mechanisms to Diseases. Genes (Basel). 24 févr 2022;13(3):401.
[3] Bradley RK, Anczuków O. RNA splicing dysregulation and the hallmarks of cancer. Nat Rev Cancer. 10 janv 2023;
[4] Hanahan D. Hallmarks of Cancer: New Dimensions. Cancer Discov. janv 2022;12(1):31‑46.
[5] Wang W, Taufalele PV, Millet M, Homsy K, Smart K, Berestesky ED, et al. Matrix stiffness regulates tumor cell intravasation through expression and ESRP1-mediated alternative splicing of MENA. Cell Reports [Internet]. 25 avr 2023 [cité 8 mai 2023];42(4).
[6] Lin KT, Krainer AR. PSI-Sigma: a comprehensive splicing-detection method for short-read and long-read RNA-seq analysis. Bioinformatics. 1 déc 2019;35(23):5048‑54.
[7] Ren X, Deng R, Zhang K, Sun Y, Teng X, Li J. SpliceRCA: in Situ Single-Cell Analysis of mRNA Splicing Variants. ACS Cent Sci. 27 juin 2018;4(6):680‑7.