In vivo prevalence and quantitative significance of miRNA/ceRNA crosstalk at the protein level

microRNAs (miRNAs) repress their target mRNAs, and transcripts that share a miRNA can indirectly regulate one another by competing for miRNA binding — the competing endogenous RNA (ceRNA) effect. A stochastic model of two ceRNAs co-regulated by one miRNA and encoding interacting proteins (Martirosyan et al., 2017) predicted that ceRNA competition (i) stabilizes protein levels across the full range of gene expression rates and (ii) reverses the sign of the correlation between co-regulated interacting proteins from negative to positive, but only within a narrow near-equimolar 'susceptible' parameter window. Whether physiological molecular abundances and kinetic parameters actually populate that window across cell types and conditions — and hence whether protein-level ceRNA effects are quantitatively significant in vivo — remains unresolved, together with the predicted evolutionary tuning of kinetic heterogeneities, direct perturbation-based experimental tests of the correlation signature, and the extension beyond the single-miRNA, two-ceRNA setting to network scale.

Background

Animal microRNAs (miRNAs) are short non-coding RNAs that guide the Argonaute-containing silencing complex to partially complementary sites (miRNA response elements) in target mRNAs, repressing their translation and promoting their decay. Because a single miRNA can bind many transcripts, its targets compete for a shared pool: increasing the level of one target (a 'competitor') sequesters miRNA molecules and thereby derepresses other targets. This indirect cross-regulation defines competing endogenous RNAs (ceRNAs) and was proposed as a broad post-transcriptional regulatory language (Salmena et al., 2011). Quantitative theory of small-RNA titration (Figliuzzi et al., 2013; Martirosyan et al., 2016) shows that effective ceRNA crosstalk requires the total target abundance to be close to the miRNA abundance — the near-equimolar regime, called the 'susceptible' regime — where miRNA-mediated control is most sensitive. Whether endogenous transcripts ever reach this regime has been the crux of a decade-long debate: Denzler et al. (2014, 2016) and Bosson et al. (2014) measured endogenous miRNA and target concentrations and concluded that, for most miRNA families, targets are far too scarce to deplete miRNAs, so ceRNA regulation should be negligible in standard physiological conditions; in vivo support exists mainly for specific cases such as PTEN ceRNAs in a BRAF-driven melanoma model (Karreth et al., 2011). Separately, miRNAs were shown to buffer protein-expression noise for lowly expressed genes (Schmiedel et al., 2015). Martirosyan, De Martino, Pagnani and Marinari (2017) connected these threads with a stochastic model of two ceRNAs (a target mRNA $m_T$ and a competitor mRNA $m_C$) regulated by one shared miRNA, whose proteins $p_T$ and $p_C$ reversibly form a complex $C_p$; the model distinguishes catalytic decay of the miRNA–ceRNA complex (miRNA recycled after target cleavage, rate $\kappa_i$) from stoichiometric decay (miRNA consumed with the target, rate $\sigma_i$). By Gillespie simulation (an exact stochastic simulation algorithm that samples individual reaction events of a chemical reaction network) and mean-field analysis they found that (i) competition can stabilize protein levels across the full range of synthesis rates, with kinetic heterogeneities enhancing the effect, (ii) the correlation between $p_T$ and $p_C$ — negative without shared miRNA regulation, because complex formation depletes both partners — can reverse sign and become positive under ceRNA crosstalk, but only in a narrow window around the quasi-equimolar (susceptible) regime, and (iii) the channel capacity $I^{\max}$ (the maximum mutual information between the input synthesis rate and the output complex level) is modulated, generically increasing, with stronger competition. They proposed that a positive Pearson correlation between the co-regulated subunits of a protein complex may provide, 'for a limited but significant set of cases', the simple and direct proof of active ceRNA crosstalk in standard physiological conditions 'that has been so far lacking', and that for the experimentally observed positive correlations between regulated complex subunits 'kinetic parameters could be globally tuned by evolution so as to fit this narrow window' — a suggestion, not an established result.

Problem Statement

Determine, quantitatively and in vivo, whether miRNA-mediated ceRNA competition significantly shapes protein levels and protein-protein covariation under physiological conditions. Specifically: (1) how broadly do physiological molecular abundances (miRNA, target, competitor) and kinetic parameters — miRNA-ceRNA binding/unbinding rates, catalytic versus stoichiometric decay rates of the miRNA-ceRNA complex, and relative synthesis/decay rates — populate the narrow quasi-equimolar 'susceptible' window in which ceRNA crosstalk and the sign reversal of protein-protein correlations occur, across cell types and conditions; (2) are the kinetic heterogeneities of miRNA-ceRNA networks evolutionarily tuned so as to produce the parameter combinations required for positive protein-protein correlations or enhanced channel capacity of complex synthesis, as speculated by Martirosyan et al. (2017); (3) do the reported perturbation-based in vivo studies of shared-miRNA target pairs, together with measured miRNA, target and competitor abundances and kinetic parameters, exhibit the model's predicted signatures (positive Pearson correlation between interacting partners under shared miRNA regulation, competitor derepression magnitude, and changes in channel capacity) under physiological regulatory complexity; and (4) how do network-level topology and multiple co-expressed miRNAs modify the single-miRNA, two-ceRNA conclusions? These four issues share one scientific objective: establishing the in vivo prevalence and quantitative significance of protein-level ceRNA effects, with the model of Martirosyan et al. (2017) and its successors as the quantitative framework being tested and extended. The standing of these items is declared as follows: (1) and (2) are to be answered quantitatively from existing abundance, kinetic and expression data; (3) is an assessment of the existing in vivo perturbation literature, not a demand for new measurements; and (4) is to be addressed qualitatively as an extension of the single-miRNA, two-ceRNA conclusions; answering all four at these respective levels constitutes a resolution, whereas answering any one in isolation does not.

References

1. Martirosyan, A., De Martino, A., Pagnani, A. and Marinari, E. (2017). ceRNA crosstalk stabilizes protein expression and affects the correlation pattern of interacting proteins. Scientific Reports 7:43673. DOI: 10.1038/srep43673. https://doi.org/10.1038/srep43673
2. Salmena, L., Poliseno, L., Tay, Y., Kats, L. and Pandolfi, P. P. (2011). A ceRNA hypothesis: the Rosetta Stone of a hidden RNA language? Cell 146:353-358. DOI: 10.1016/j.cell.2011.07.014. https://doi.org/10.1016/j.cell.2011.07.014
3. Karreth, F. A., Tay, Y., Perna, D., Ala, U., Tan, S. M., Rust, A. G., DeNicola, G., Webster, K. A., Weiss, D., Perez-Mancera, P. A., Krauthammer, M., Halaban, R., Provero, P., Adams, D. J., Tuveson, D. A. and Pandolfi, P. P. (2011). In vivo identification of tumor-suppressive PTEN ceRNAs in an oncogenic BRAF-induced mouse model of melanoma. Cell 147:382-395. DOI: 10.1016/j.cell.2011.09.032. https://doi.org/10.1016/j.cell.2011.09.032
4. Figliuzzi, M., Marinari, E. and De Martino, A. (2013). MicroRNAs as a selective channel of communication between competing RNAs: a steady-state theory. Biophysical Journal 104:1203-1213. DOI: 10.1016/j.bpj.2013.01.012. https://doi.org/10.1016/j.bpj.2013.01.012
5. Denzler, R., Agarwal, V., Stefano, J., Bartel, D. P. and Stoffel, M. (2014). Assessing the ceRNA hypothesis with quantitative measurements of miRNA and target abundance. Molecular Cell 54:766-776. DOI: 10.1016/j.molcel.2014.03.045. https://doi.org/10.1016/j.molcel.2014.03.045
6. Bosson, A. D., Zamudio, J. R. and Sharp, P. A. (2014). Endogenous miRNA and target concentrations determine susceptibility to potential ceRNA competition. Molecular Cell 56:347-359. DOI: 10.1016/j.molcel.2014.09.018. https://doi.org/10.1016/j.molcel.2014.09.018
7. Schmiedel, J. M., Klemm, S. L., Zheng, Y., Sahay, A., Blüthgen, N., Marks, D. S. and van Oudenaarden, A. (2015). MicroRNA control of protein expression noise. Science 348:128-132. DOI: 10.1126/science.aaa1738. https://doi.org/10.1126/science.aaa1738
8. Martirosyan, A., Figliuzzi, M., Marinari, E. and De Martino, A. (2016). Probing the limits to microRNA-mediated control of gene expression. PLOS Computational Biology 12:e1004715. DOI: 10.1371/journal.pcbi.1004715. https://doi.org/10.1371/journal.pcbi.1004715
9. Denzler, R., McGeary, S. E., Title, A. C., Agarwal, V., Bartel, D. P. and Stoffel, M. (2016). Impact of microRNA levels, target-site complementarity, and cooperativity on competing endogenous RNA-regulated gene expression. Molecular Cell 64:565-579. DOI: 10.1016/j.molcel.2016.09.027. https://doi.org/10.1016/j.molcel.2016.09.027
10. Martirosyan, A., Marsili, M. and De Martino, A. (2017). Translating ceRNA susceptibilities into correlation functions. Biophysical Journal 113:206-213. DOI: 10.1016/j.bpj.2017.05.042. https://doi.org/10.1016/j.bpj.2017.05.042
11. Smillie, C. L., Sirey, T. and Ponting, C. P. (2018). Complexities of post-transcriptional regulation and the modeling of ceRNA crosstalk. Critical Reviews in Biochemistry and Molecular Biology 53:231-245. DOI: 10.1080/10409238.2018.1447542. https://doi.org/10.1080/10409238.2018.1447542
12. Del Giudice, M., Bosia, C., Grigolon, S. and Bo, S. (2018). Stochastic sequestration dynamics: a minimal model with extrinsic noise for bimodal distributions and competitors correlation. Scientific Reports 8:10387. DOI: 10.1038/s41598-018-28647-9. https://doi.org/10.1038/s41598-018-28647-9
13. Park, H. J., Ji, P., Kim, S., Xia, Z., Rodriguez, B., Li, L., Su, J., Chen, K., Masamha, C. P., Baillat, D., Fontes-Garfias, C. R., Shyu, A.-B., Neilson, J. R., Wagner, E. J. and Li, W. (2018). 3' UTR shortening represses tumor-suppressor genes in trans by disrupting ceRNA crosstalk. Nature Genetics 50:783-789. DOI: 10.1038/s41588-018-0118-8. https://doi.org/10.1038/s41588-018-0118-8
14. Miotto, M., Marinari, E. and De Martino, A. (2019). Competing endogenous RNA crosstalk at system level. PLOS Computational Biology 15:e1007474. DOI: 10.1371/journal.pcbi.1007474. https://doi.org/10.1371/journal.pcbi.1007474
15. Liu, Y., Xue, M., Du, S., Feng, W., Zhang, K., Zhang, L., Liu, H., Jia, G., Wu, L., Hu, X., Chen, L. and Wang, P. (2019). Competitive endogenous RNA is an intrinsic component of EMT regulatory circuits and modulates EMT. Nature Communications 10:1637. DOI: 10.1038/s41467-019-09649-1. https://doi.org/10.1038/s41467-019-09649-1
16. Han, Y. and Zhang, F. (2020). Heterogeneity coordinates bacterial multi-gene expression in single cells. PLOS Computational Biology 16:e1007643. DOI: 10.1371/journal.pcbi.1007643. https://doi.org/10.1371/journal.pcbi.1007643
17. Ferro, E., Szischik, C. L., Cunial, M., Ventura, A. C., De Martino, A. and Bosia, C. (2025). Out-of-equilibrium ceRNA crosstalk. In: Computational Biology of Non-Coding RNA. Methods in Molecular Biology, vol. 2883, pp. 167-193. DOI: 10.1007/978-1-0716-4290-0_8. https://doi.org/10.1007/978-1-0716-4290-0_8