Cell-Cycle-Dependent ERK Signaling Dynamics Direct Fate Specification in the Mammalian Preimplantation Embryo

Abstract

Despite the noisy nature of single cells, multicellular organisms robustly generate different cell types from one zygote. This process involves dynamic cross regulation between signaling and gene expression that is difficult to capture with fixed-cell approaches. To study signaling dynamics and fate specification during preimplantation development, we generated a transgenic mouse expressing the ERK kinase translocation reporter and measured ERK activity in single cells of live embryos. Our results show primarily active ERK in both the inner cell mass and trophectoderm cells due to fibroblast growth factor (FGF) signaling. Strikingly, a subset of mitotic events results in a short pulse of ERK inactivity in both daughter cells that correlates with elevated endpoint NANOG levels. Moreover, endogenous tagging of Nanog in embryonic stem cells reveals that ERK inhibition promotes enhanced stabilization of NANOG protein after mitosis. Our data show that cell cycle, signaling, and differentiation are coordinated during preimplantation development.

Keywords: ERK; NANOG; blastocyst; cell cycle; embryonic stem cells; preimplantation development; signaling dynamics.

Figure 1 |. ERK-KTR^LoxP Mice reveal ERK signaling dynamics in mouse preimplantation development.

(A) Schematic of KTR reporter construct for ERK-KTR^LSL animals (upper) and mating scheme to generate mice with germline expression of transgene (lower). Representative image of p0 pups with visible GFP fluorescence is shown. (B) MEFs were derived from ERK-KTR^LoxP animals and imaged before and after (30 min.) stimulation with FGFb (5 ng/ml). Representative images are shown. Scale bar = 50 μm. (C) Schematic of ERK activity quantification method. For each cell, a nuclear and cytoplasmic region of interest was drawn and KTR intensity was measured. ERK activity was reported as the log₂ of the cytoplasmic / nuclear (log₂(C/N)) intensity ratio. (D) E3.5 blastocysts from ERK-KTR^LoxP animals were collected, cultured, and imaged as described in methods. Representative images containing ERK low cells (arrows) are shown in ICM and TE compartments. Scale bar = 20 μm. (E) Embryos were isolated from ERK-KTR^LoxP animals at the 8-cell, 16-cell, E3.5 blastocyst, and E4.5 peri-implantation blastocyst stages and mounted for imaging in KSOM as described in methods. Representative images of single confocal planes are shown for clarity. Scale bar = 20 μm. (F) Single cell ERK activities from embryos collected in (E) were quantified as described in methods. Cells with predominantly cytoplasmic (log₂(C/N)>0) or nuclear (log₂ (C/N)<0) KTR intensity were represented as blue or red points respectively. From left to right, N = 8 embryos, 55 cells; 4 embryos, 51 cells; 10 embryos, 339 cells; and 3 embryos, 190 cells. (G) E3.5 blastocysts were isolated from ERK-KTR^LoxP animals and treated with indicated inhibitors and growth factors (ERKi, 5 μM Ulixertinib; MEKi, 1 μM PD0325901; FGFRi, 1 μM AZD4547, 1000 ng/ml FGF4) for 45 min. Embryos were mounted for imaging in KSOM, maintaining the same concentration of inhibitors and growth factors, and embryos were imaged via confocal microscopy. Representative images of a single confocal plane are shown. Scale bar = 20 μm. (H) Single cell ERK activities from embryos collected in (G) were quantified and plotted as in (F). From left to right, N = 10 embryos, 339 cells; 5 embryos, 216 cells; 6 embryos, 256 cells; 3 embryos, 103 cells; and 2 embryos, 105 cells. Untreated KSOM control corresponds to E3.5 blastocysts from (E). Significant differences between treated and untreated ERK activity distributions were determined by Kolmogorov-Smirnov test. * p < 0.05 and *** p < 0.001.

Figure 2 |. ERK activity bifurcates after mitosis in single cells of developing blastocysts.

(A) E3.5 blastocysts were isolated from ERK-KTR^LoxP animals, mounted in KSOM, and imaged every 15 min. for 12 hr. Single cell ERK activity was quantified and represented as a heat map (see methods for details). Mitotic events and lost tracks are represented in blue and green, respectively. Predominantly cytoplasmic (log₂(C/N)>0) or nuclear (log₂ (C/N)<0) values are depicted with gray or red scale colormaps respectively. Black represents C=N intensity. Data depicts traces from a single embryo representative of 2 independent experiments. (B) Representative images of ERK high and ERK low mitotic events observed from the experiment in (A). For each example, the mother cell is shown immediately before entering mitosis and the daughter cells are shown 90 min. after anaphase. Yellow arrows indicate featured cells. Scale bar = 10 μm. (C) Representative traces of ERK high (Pair 1) and ERK low (Pair 2) mitotic events are plotted. Grey box indicates mitosis. Traces left of mitosis indicate mother cell ERK activity and right of mitosis indicate corresponding daughter cell activities. Each sister cell pair is shown as a solid and dotted line pair. (D) ERK-KTR^LoxP blastocysts were isolated and imaged as in (A). Daughter cell pairs were assigned as ERK high (C>N, blue) or ERK low (C<N, red). Based on 4 individual embryos. (E) ERK-KTR^LoxP blastocysts were isolated and imaged as in (A). The duration of ERK inactivity in ERK low divisions was measured as described in methods. Histogram reflects measurements from 28 ERK low cells from 4 embryos obtained from 2 independent experiments. (F) E3.5 blastocysts were isolated from ERK-KTR^LoxP animals, incubated in KSOM ± MEKi (1 μM PD0325901), and imaged every 15 min. for 9 hr. ERK activity was quantified pre- and post-mitosis as described in methods. ERK activity is represented as in (A) and in silico synchronized to mitosis. Left and right sides of the heat map indicate mother and daughter cell activities respectively. Data represent 22 (KSOM) and 13 (MEKi) mitotic events from 4 and 3 embryos, respectively, from 3 independent experiments. (G) Scatter plot of mean ERK activity 90 min. before (mother) and after (daughter) mitosis from data in (D). Significant linear correlations between mother and daughter cell activity were observed for both the KSOM and MEKi groups (p = 2.58 × 10⁻⁶ and 0.00063 by F-test, respectively) with R² values of 0.413 and 0.405, respectively. (H) E3.5 blastocysts were isolated from ERK-KTR^LoxP animals, incubated in KSOM and KSOM with MEKi (1 μM PD0325901) or APCi (2 μM ProTAME), and imaged every 15 min. for 9 hr. Δ ERK activity was defined as the mean ERK activity post-mitosis subtracted from the mean ERK activity pre-mitosis. ERK low (blue) and ERK high (red) cells were defined as daughter cells whose mean ERK activity was less than 0.1 and greater than or equal to 0.1, respectively. Dark red lines indicate mean ± standard deviation of measurements on ERK low cells only. Data represent 22 mitotic events from 4 blastocysts (KSOM), 13 mitotic events from 3 blastocysts (MEKi), and 7 mitotic events from 2 blastocysts (APCi). *** p<0.001 and N.S. p>0.05 by Student’s t-test comparing ERK low cells in each group to the KSOM condition. (I) Data obtained in (H) is shown as in silico synchronized ERK activity traces. ERK low divisions only are shown for clarity. Grey vertical bar indicates mitosis.

Figure 3 |. ERK activity at mitotic exit governs EPI and PrE lineage commitment.

(A) E3.5 blastocysts were isolated and cultured in KSOM plus indicated inhibitors or growth factors (MEKi, 1 μM PD0325901; APCi, 2 μM ProTAME; 1000 ng/ml FGF4) for 9 hr. Embryos were fixed, immunostained, imaged, and quantified as described in methods. TE cells were excluded from analysis for clarity (see Figure S4A and S4B for details). Data reflects 47 to 153 individual ICM cells from 3 to 8 embryos from 4 independent experiments. (B) Schematic of experimental workflow to investigate ERK signaling dynamics and fate selection in the same cells. E3.5 ERK-KTR^LoxP blastocysts were imaged in KSOM for 9 hr., retrieved, fixed and immunostained as described in methods. H2B-mCherry signal from the live imaging and fixed embryos were aligned in silico to correlate ERK signaling dynamics and marker expression in the same cells. For each cell that was unequivocally identified in both data sets, ERK activity was measured at mitotic exit (60 min.) and at the end of the time lapse period (End Point) (60 min). GATA6 and NANOG signals were then measured and matched with ERK activity measurements for the corresponding cells (see methods for details). (C-E) ERK-KTR^LoxP blastocysts were imaged and immunostained as described in (B). ERK activity at mitotic exit (left) and end point (right) were plotted against normalized NANOG intensity (C), normalized GATA6 intensity (D), normalized log₁₀(GATA6/NANOG) intensity ratio (E). p values of linear correlation by F-test are shown. Data represent measurements from 40 mitotic cells from 6 embryos from 3 independent experiments. (F) Data from (C-E) was plotted as normalized NANOG vs. normalized GATA6 with each point colored according to mean ERK activity at mitotic exit (left) or at end point (right).

Figure 4 |. ERK-KTR ES cells recapitulate ICM signaling dynamics.

(A) ES cells were derived from ERK-KTR^LoxP animals as described in methods and seeded into imaging plates coated with fibronectin with or without ERKi (5 μM Ulixertinib). Images were obtained every 5 min. Representative images before or after (15 min.) inhibitor addition are shown. Scale bar = 20 μm. (B) ERK-KTR ES cells were seeded to imaging plates coated with fibronectin and cultured in complete 2i growth media. Next day, MEK and GSK3β inhibitors were removed and cells were imaged every 5 min. for 18 hr. ERK activity in single cells was measured and quantified as described in methods. ERK high and low cells were defined as cells whose mean ERK activity after mitosis was greater than −0.1 or less than −0.1, respectively. Following this sorting, randomly selected traces were plotted from each group. Data represents 4 experimental replicates. (C) Mean ERK activity before (mother) and after (daughter) mitosis (90 min.) was calculated from ERK activity traces obtained in (B). Significant linear correlation was observed (p = 3.95 × 10⁻¹² and R² = 0.265). Data represent 159 mitotic events. (D) ERK-KTR ES cells were seeded as described in (B) and imaged every 5 min. for 20 hr. in growth media without inhibitors or growth media with MEKi (2 μM PD0325901). Fraction of ERK low divisions was calculated as described in methods. Data represent >700 mitotic events from >8 replicates for each condition. *** p < 0.001 by a Student’s t-test.

Figure 5 |. ERK inhibition promotes rapid NANOG stabilization at mitotic exit.

(A) Schematic representation of the NANOG NV2C reporter inserted at the endogenous Nanog locus in mouse ES cells (see methods for details). (B) NV2C ES cells were seeded to imaging plates and cultured in 2i/LIF media or XEN differentiation media for 24 hr. before imaging. Representative colonies from 16 replicates are shown. Scale bar = 30 μm. (C) Quantification of Venus and mCherry intensities from single cells obtained in (B). n>1000 cells for each condition from 16 replicates. ***p<0.001 by Kolmogorov-Smirnov tests. (D) NV2C ES cells were seeded onto imaging plates and cultured in ES growth media without inhibitors. Cells were imaged every 5 min. for 3 hr. and MG132 (10 μM) or Cycloheximide (CHX, 20 ug/ml) were added after 35 min. of imaging. Single cell Venus and mCherry intensities were quantified and analyzed as described in methods. Mean fold change (solid lines) with 75^th and 25^th percentiles (shaded area) are plotted for 108, 73, and 85 cells from 4 replicates each for the vehicle, MG132, and CHX groups, respectively. (E) Data from (D) plotted as mean fold change (solid lines) and 75^th and 25^th percentiles (shaded area) of Venus (left) and mCherry (right) intensities from each condition. (F) NV2C ES cells treated as in (D) were imaged every 5 min. for 20 hr. Venus and mCherry intensities were quantified as described in methods and in silico synchronized to mitotic exit. Data are plotted as fold change in intensity for the first 10 min. after anaphase. Overlaid grey dashed line indicates cells that upregulate NANOG at mitotic exit (lower) and cells that maintain or downregulate NANOG expression (upper). Data represents 122 mitotic events from 6 experimental replicates. (G) Average traces are plotted for data in (F). Plots show mean fold change (solid lines) and 75^th and 25^th percentiles (shaded area) of Venus and mCherry intensities for cells that maintain or downregulate NANOG (upper) and cells that upregulate NANOG (lower). (H) NV2C ES cells were seeded as in (E) and treated with or without ERKi (5 μM Ulixertinib). Single cell traces of Venus and mCherry signals were quantified as in (F). Mitotic events were synchronized in silico and mean fold change of Venus and mCherry intensities for all cells analyzed are plotted. Data represents >122 cells from 6 replicates each. Student’s T-test was performed for means at each time point and p values are plotted according to color bar in Figure 5I. (I) ERKi (5 μM Ulixertinib) treated cells obtained in (H) are plotted to compare average Venus and mCherry fold changes in asynchronous (black) and mitotic exit in silico synchronized (red) individual cells. Data represents >140 individual cells from 6 replicates for each condition. Student’s T-test was performed for means at each time point and p values are plotted according to color bar.