Transcriptome dynamics along axolotl regenerative development are consistent with an extensive reduction in gene expression heterogeneity in dedifferentiated cells

Abstract

Although in recent years the study of gene expression variation in the absence of genetic or environmental cues or gene expression heterogeneity has intensified considerably, many basic and applied biological fields still remain unaware of how useful the study of gene expression heterogeneity patterns might be for the characterization of biological systems and/or processes. Largely based on the modulator effect chromatin compaction has for gene expression heterogeneity and the extensive changes in chromatin compaction known to occur for specialized cells that are naturally or artificially induced to revert to less specialized states or dedifferentiate, I recently hypothesized that processes that concur with cell dedifferentiation would show an extensive reduction in gene expression heterogeneity. The confirmation of the existence of such trend could be of wide interest because of the biomedical and biotechnological relevance of cell dedifferentiation-based processes, i.e., regenerative development, cancer, human induced pluripotent stem cells, or plant somatic embryogenesis. Here, I report the first empirical evidence consistent with the existence of an extensive reduction in gene expression heterogeneity for processes that concur with cell dedifferentiation by analyzing transcriptome dynamics along forearm regenerative development in Ambystoma mexicanum or axolotl. Also, I briefly discuss on the utility of the study of gene expression heterogeneity dynamics might have for the characterization of cell dedifferentiation-based processes, and the engineering of tools that afforded better monitoring and modulating such processes. Finally, I reflect on how a transitional reduction in gene expression heterogeneity for dedifferentiated cells can promote a long-term increase in phenotypic heterogeneity following cell dedifferentiation with potential adverse effects for biomedical and biotechnological applications.

Keywords: Cell dedifferentiation; Chromatin compaction; Gene expression heterogeneity; Regenerative development.

Figure 1. Temporal dynamics of the level and the heterogeneity in gene expression along axolotl forearm regenerative development.

Generalized changes in gene expression measures along axolotl forearm regeneration was inspected using data originally produced by Voss and coworkers and unrestricted MCW tests (see Material and Methods for further details). Gene expression bias indexes (GEBIs) measure generalized biases for transcript abundance CV and mean (cvGEBIs and mGEBIs respectively) when comparing data for each post-amputation timepoint (TX) and the day of the amputation (T0), TX vs T0 comparisons. Positive GEBIs represent timepoints for which gene expression measures tend to be higher post-amputationally than in the day of amputation (TX > T0), whereas negative GEBIs represent cases for which gene expression measures tend to be lower post-amputationally than in the day of amputation (TX < T0). Simulated GEBIs were obtained after randomly rearranging gene expression measures for all probesets within TX vs T0 comparison 10,000 times with no restriction. The distribution of simulated GEBIs is summarized using minimums, 5th and 95th percentiles, and maximums. Observed GEBIs were considered significant if they outlied the area of the graph defined by 5th and 95th percentiles (P < 0.05). Grey boxes delimit axolotl forearm regeneration stages defined by morphological changes according to Voss et al. (2015).

Figure 2. Contribution of chance, factors acting on the transcriptome as a whole, and the variation in gene expression level to gene expression heterogeneity dynamics along axolotl forearm regenerative development.

MCW tests using transcript abundance CV for each TX vs T0 comparison were performed by randomly rearranging transcript abundance CV with no restriction (unrestricted), restricted by timepoint (timepoint-restricted), or restricted within probeset bins defined according to their transcript abundance mean values (expression-restricted). Two alternative expression-restricted MCW test designs performed independently are indicated with different superscripts (see Material and Methods for further description). MCW tests were repeated 10,000 times, and simulated cvGEBIs with closer values to observed cvGEBIs are represented in (A). (B) The fraction of observed cvGEBI dynamics that could be explained by chance, or factors acting on the transcriptome as a whole, or the variation in gene expression level were calculated as the proportional cumulative area under the curve (PC_AUC).

Figure 3. Temporal dynamics of the level and the heterogeneity in gene expression along axolotl forearm regenerative development for groups of genes defined by their functionality.

The Gene Ontology Consortium database was used to retrieved genes belonging to “mitotic cell cycle process” (GO:1903047) and “chromatin organization” (GO:0006325) GO terms ( The Gene Ontology Consortium, 2015). The temporal dynamics of the level and heterogeneity in gene expression for probesets associated to “mitotic cell cycle process” and “chromatin organization” genes with regard to the whole transcriptome were inspected using functionally-restricted MCW tests (A and B, respectively). cvGEBI and mGEBI for each group of probesets and TX vs T0 comparison were calculated before and after randomly rearranging functional subset tags for the whole transcriptome 10,000 times. The distribution of simulated GEBIs is summarized using minimums, 5th and 95th percentiles, and maximums. Observed GEBIs were considered significant if they outlied the area of the graph defined by 5th and 95th percentiles (P < 0.05). Grey boxes delimit axolotl forearm regeneration stages defined by morphological changes according to Voss et al. (2015).

Figure 4. Generalized reduction in gene expression heterogeneity upon cell dedifferentiation might reduce genetic capacitance.

Cartoon symbolizing how genetic variation can become phenotypically relevant associated to the chromatin relaxation-dependent reduction in gene expression heterogeneity upon cell dedifferentiation. Blue shades represent the variation in gene expression for a particular gene in a population of cells before and after dedifferentiation. Pre-existing or new genetic mutations might be phenotypically more distinguishable upon cell dedifferentiation because of the generalized reduction in gene expression heterogeneity that cause a narrowing of the spectrum of stochastic phenotypes.