Inherent mitochondrial activity influences specification of the germ line in pluripotent stem cells

Abstract

Herein we investigated whether inherent differences in mitochondrial activity in mouse pluripotent cells could be used to identify populations with an intrinsic ability to differentiate into primordial germ cells (PGCs). Notably, we determined that stem cells sorted based on differences in mitochondrial membrane activity exhibited altered germline differentiation capacity, with low-mitochondrial membrane potential associated with an increase in PGC-like cells. This specification was not further enhanced by hypoxia. We additionally noted differences between these populations in metabolism, transcriptome, and cell-cycle. These data contribute to a growing body of work demonstrating that pluripotent cells exhibit a large range of mitochondrial activity, which impacts cellular function and differentiation potential. Furthermore, pluripotent cells possess a subpopulation of cells with an improved ability to differentiate into the germ lineage that can be identified based on differences in mitochondrial membrane potential.

Keywords: Biological sciences; Cell biology; Mitochondria; PGC; Primordial germ cell; Stem cell; Stem cell research.

Figure 1

Inherent differences in mitochondrial activity in pluripotent stem cells. (A) Imaging of undifferentiated mESC and miPSC colonies. Mitochondria with high membrane potential were stained by TMRM (red) and total mitochondria stained in MitoTracker Green FM (MTG, green), with nuclei stained in Hoechst (blue). Scale bars = 50 μm. (B) mESCs and miPSCs stained with TMRM and MTG were treated with FCCP (400 nM). TMRM staining alone was used to isolate viable cells from the highest 5% of TMRM signal (high-Δψ_m) and lowest 5% of TMRM signal (low-Δψ_m). Density plots and histograms represent n = 3 replicates, and a representative plot is shown for the TMRM gating strategy. Data represent mean ± SEM for n = 6 experiments with representative FACS plots shown for each sample set. Significant differences between populations was calculated by student's T-test, with P-values shown.

Figure 2

Low- and high-Δψ_m undifferentiated cells have altered ROS and ATP production, but similar mtDNA copy number and differentiation capacity. (A) ATP generation from high-Δψ_m and low-Δψ_m cells. (B) ROS detection by H2DCFDA dye incorporation and FACS analysis in high-Δψ_m and low-Δψ_m cells. Plots represent n = 3 and values shown are average mean fluorescence intensity ± SEM for each population with P-values. Insets show H2DCFDA signal for unstained cell populations. (C) mtDNA copy number was detected by quantitative PCR of mitochondrial genes Nd1 and Nd4 and normalized to nuclear gene Tert. Fold change was calculated in comparison to low-Δψ_m cells. (D) 0.5 × 10⁶ cells from each population were injected subcutaneously in Matrigel into female 6-week old NOD/SCID mice to form teratomas (n = 4). Teratomas were removed after 22 or 25 days and processed for H&E staining to reveal cells from all three germ layers. Ectoderm: neural rosettes, endoderm: gland like/epithelium, mesoderm: cartilage. Scale bars = 50 μm. Data for ATP, ROS, and mtDNA copy number analyses represent mean ± SEM for triplicate replicates. Significant differences between populations were calculated by student's T-test, with P-values shown.

Figure 3

Transcriptome changes in low- and high-Δψ_m cells reveal that mitochondrial activity correlates to cell cycle regulation. (A) Principal component analysis (PCA) was computed across all genes from microarray data of undifferentiated mESCs and miPSCs sorted by membrane potential. (B) Genes that were significantly over-expressed in low-Δψ_m and high-Δψ_m cells were compared to identify genes that were over-expressed in both cell lines. (C) Proliferation was assessed by BrdU incorporation and was used in conjunction with DAPI DNA staining to determine cell cycle by FACS. Mean ± SEM are shown for triplicate replicates. Representative FACS plots, controls, and gating strategy are shown in Figure S3. (D) Apoptosis was assessed by caspase 3/7 induction in vehicle (veh) or doxorubicin (dox) treated cells after sorting and 6 h of treatment. Fold change calculated as doxorubicin signal over vehicle for each condition. Mean ± SEM are shown for triplicate biological replicates.

Figure 4

Inherent differences in mitochondrial activity influences the ability of pluripotent stem cells to differentiate into PGCLCs. (A) After sorting by TMRM staining intensity and undergoing directed differentiation to PGCLCs, viability was assessed by DAPI incorporation and FACS analysis. (B) Aggregate diameter was compared for differentiated cells from high- or low-Δψ_m cells in comparison to the total population. (C) PGCLC differentiation was analyzed by assessing the proportion of SSEA1+/CD61 + cells by FACS. Data represent mean ± SEM for n = 6 experiments with representative FACS plots shown for each sample set. Significant differences between populations was calculated by student's T-test, with P-values shown.

Figure 5

Hypoxic culture conditions to decrease mitochondrial activity does not increase germ cell differentiation. (A) L-Lactate production was assessed in undifferentiated cells that had been incubated under hypoxic conditions (5% O₂) in comparison to control cells cultured in standard cell culture conditions (~21% O₂). Cells were cultured with glucose analog 2-DG to detect cells with decreased L-Lactate production. (B) Glycolysis-related genes and common downstream targets of HIF1Awere evaluated for increase in response to hypoxic culture at 24, 48, and 72 h. (C) ESCs and iPSCs were cultured in normoxic or hypoxic conditions and underwent PGCLC differentiation maintained in each condition. The resulting cells were assessed for viability and imaged for aggregate morphology. (D) ESCs and iPSCs were assessed for PGCLC formation after culture under normoxic or hypoxic conditions. Data represent mean ± SEM for n = 3 replicates. Significant differences between populations was calculated by student's T-test, with P-values shown (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001).