Physioxia rewires mitochondrial complex composition to protect stem cell viability

Abstract

Human induced pluripotent stem cells (hiPSCs) are an invaluable tool to study molecular mechanisms on a human background. Culturing stem cells at an oxygen level different from their microenvironmental niche impacts their viability. To understand this mechanistically, dermal skin fibroblasts of 52 probands were reprogrammed into hiPSCs, followed by either hyperoxic (20 % O₂) or physioxic (5 % O₂) culture and proteomic profiling. Analysis of chromosomal stability by Giemsa-banding revealed that physioxic -cultured hiPSC clones exhibited less pathological karyotypes than hyperoxic (e.g. 6 % vs. 32 % mosaicism), higher pluripotency as evidenced by higher Stage-Specific Embryonic Antigen 3 positivity, higher glucose consumption and lactate production. Global proteomic analysis demonstrated lower abundance of several subunits of NADH:ubiquinone oxidoreductase (complex I) and an underrepresentation of pathways linked to oxidative phosphorylation and cellular senescence. Accordingly, release of the pro-senescent factor IGFBP3 and β-galactosidase staining were lower in physioxic hiPSCs. RNA- and ATAC-seq profiling revealed a distinct hypoxic transcription factor-binding footprint, amongst others higher expression of the HIF1α-regulated target NDUFA4L2 along with increased chromatin accessibility of the NDUFA4L2 gene locus. While mitochondrial DNA content did not differ between groups, physioxic hiPSCs revealed lower polarized mitochondrial membrane potential, altered mitochondrial network appearance and reduced basal respiration and electron transfer capacity. Blue-native polyacrylamide gel electrophoresis coupled to mass spectrometry of the mitochondrial complexes detected higher abundance of NDUFA4L2 and ATP5IF1 and loss of incorporation into complex IV or V, respectively. Taken together, physioxic culture of hiPSCs improved chromosomal stability, which was associated with downregulation of oxidative phosphorylation and senescence and extensive re-wiring of mitochondrial complex composition.

Keywords: Complexes; HIF1α; Human induced pluripotent stem cells; Mitochondrial function; NDUFA4L2; Senescence.

Graphical abstract

Fig. 1

Generation and characterization of human induced pluripotent stem cells (hiPSCs) cultured at physioxia (5 % O₂) or hyperoxia (20 % O₂). A) Experimental setup. Fibroblasts from 52 unrelated patients were reprogrammed by a Sendai virus containing pluripotency factors and subsequently incubated at 5 % O₂ (physioxia) or 20 % O₂ (hyperoxia). Between passage 18 and 30, a master cell bank (MCB) was generated and used for different approaches. B) Karyotype analysis by Giemsa banding, C) quantification as parts of whole representing normal karyotypes (white) and abnormal karyotypes (grey) at physioxic or hyperoxic culture and D) representative images of physioxic and hyperoxic cultured hiPSCs. Scale bars are equal to 400 μm. E) Comparison of passage numbers of hiPSC-lines. F) Stage-Specific Embryonic Antigen 3 (SSEA3) quantification as a marker for pluripotency. Paired t-test, #### = p < 0.0001. G) Assessment of hiPSC proliferation rate under hyperoxic (black) or physioxic (green) conditions monitored over 10 days and three passages. H) Glucose consumption (mM) and lactate production (mM) were measured in culture media at 12 h, 36 h, 48 h and 72 h incubation of hiPSCs under hyperoxia (black dots) or physioxia (green dots). I) Western immunoblot analysis for HIF1α and β-actin, and J) quantification of HIF1α by Western blot normalised to β-actin. n = 8 independent hyperoxic or physioxic hiPSC batches. For statistical analysis unpaired t-test was performed with # = p < 0.05; ## = p < 0.01; ### = p < 0.001 and #### = p < 0.0001.

Fig. 2

Proteomic analysis of physioxic (5 % O₂) and hyperoxic (20 % O₂) cultured hiPSCs shows enhanced cellular senescence in hyperoxic-cultured hiPSCs. A) Volcano plot representing the most significantly up- and downregulated genes in hiPSCs that had been cultured under physioxic conditions (n = 4 independent hiPSC batches) in comparison to hyperoxic (n = 4 independent hiPSC batches)-cultured hiPSCs. Subunits are highlighted in color: complex I (yellow); complex II (orange); complex III (pink); complex IV (green); ATP synthase (turquoise). B) Heatmap for hiPSC samples cultured under physioxic or hyperoxic conditions with top 10 up- and downregulated genes at physioxia and hyperoxia. C) Pathway analysis of processes that were associated with genes that were significantly downregulated in physioxic-compared to hyperoxic-cultured hiPSCs. Enrichment analysis was performed with the online WEB-based GEne SeT AnaLysis Toolkit (WebGestalt) (n = 4). D) Schematic experimental overview. 12 h after standard medium exchange, physioxic (green arrow) and hyperoxic (black arrow) medium was exchanged to the respective other hiPSCs and cultured at the same oxygen tension for another 24 h before the procedure was repeated once more. E) Representative images for quantification of senescence-associated β-galactosidase activity (n = 3) and the same images after background subtraction with ImageJ. The positive control was treated twice for 2 h with 75 μM H₂O₂. Scale bar is equal to 100 μm, and F) quantification of the SA-associated β-galactosidase activity with ImageJ after removal of the background (n = 3). G) Concentrations of the senescence marker IGFBP3 in endpoint media from F). Green symbols represent culture at physioxic conditions while black symbols represent hyperoxic culture with physioxic (green bars) or hyperoxic (white bars) medium. For statistics, unpaired t-test compared to physioxic-cultured hiPSCs (# = p < 0.05, #### = p < 0.0001) or one-way ANOVA, Tukey's post-hoc test (∗∗∗ = p < 0.001, ∗∗∗∗ = p < 0.0001) were performed.

Fig. 3

RNA- and ATAC sequencing of physioxic (5 % O₂) and hyperoxic (20 % O₂)-cultured hiPSCs. A) Volcano plot representing the most significantly up- and downregulated genes in hiPSCs that had been cultured under physioxic conditions (n = 3 independent batches) in comparison to hyperoxic (n = 3 independent batches)-cultured hiPSCs. B) mRNA fold change of NDUFA4L2, NDUFA4 and HIF1A normalised to GUSB. C-E) Plots of RNAseq and ATACseq for C)NDUFA4L2, D)NDUFA4, and E)HIF1A (n = 3). For statistics, unpaired t-test compared to physioxic-cultured hiPSCs was performed (# = p < 0.05, ### = p < 0.001).

Fig. 4

Analysis of the composition of the mitochondrial respiratory chain complexes and function in physioxia (5 % O₂) and hyperoxia (20 % O₂)-cultured hiPSCs by blue-native (BN)-polyacrylamide gel electrophoresis (BNE). Mitochondrial complex analysis of A) complex I (yellow); B) complex II (orange); C) complex III (red); D) complex IV (green) with NDUFA4 highlighted in blue and NDUFA4L2 in red; E) complex V (blue) with ATP5IF1 highlighted in light blue in physioxic- (n = 3 independent batches) versus hyperoxic- (n = 3 independent batches) cultured hiPSCs. The Coomassie stained gels (top of each panel) represent the fractionated strip for each complexome. The sum of IBAQ values of the complexomes were used to normalize the physioxia to the hyperoxia data set. The heatmap illustrates the maximum normalization between both conditions.

Fig. 5

Quantification of mitochondrial complex abundance. A) Quantification of mitochondrial complex I–V in hyperoxia (left) versus physioxia (right). The sum of all subunits was used to summarize the abundance of each complex. The heatmap represents a maximum normalization between both conditions. B) Quantification and mapping of NDUFA4L2 (red arrows) in hyperoxia (left) versus physioxia (right) in relation to NDUFA4 (blue). C) Quantification and mapping of ATP5IF1 (IF1; blue arrows) in hyperoxia (left) versus physioxia (right).

Fig. 6

HIF1α localization and mitochondrial function in physioxic (5 % O₂) and hyperoxic (20 % O₂) cultured hiPSCs. A) Quantification of genomic and mitochondrial DNA (mt-DNA) content by RT-qPCR analysis (n = 3 independent batches of each physioxic- or hyperoxic-cultured hiPSCs). Gene expression of mitochondrial-encoded NADH dehydrogenase-1 (mt-ND1) and −2 (mt-ND2) were normalised to the nuclear encoded globular actin (g-βactin; n = 6 independent batches of each physioxic- or hyperoxic-cultured hiPSCs). B) Assessment of mitochondrial membrane potential (mtMP) by TMRM fluorescence emitted at 580–700 nm and normalised to MitoTrackerGreen fluorescence at 500–530 nm. C) Mitochondria network visualization in physioxic and hyperoxic hiPSCs by MitoSpy (orange) immunofluorescence. MitoSpy images were transformed to threshold images (TI; black and white) by ImageJ software. Merged images show localization of nuclei by DAPI staining. Scale bars are equal to 20 μm. D) Oxygen flux measurements in intact, non-permeabilized hiPSCs cultured under physioxic (green) or hyperoxic (black) conditions were performed in an O2k oxygraph. For each measurement, 750000 hiPSCs/mL buffer were used. HiPSCs were sequentially treated with pyruvate (5 mM), oligomycin (5 nM), CCCP (0.5 μM titration steps) and rotenone (0.5 μM). Representative trace. E) Scatter dot plot shows data corrected for residual oxygen consumption (ROX) after treatment with rotenone. R': routine respiration (R′) was evaluated after addition of pyruvate, L': leak respiration after addition of oligomycin, E': electron transfer capacity after titration with CCCP. For statistics, unpaired t-test was performed (# = p < 0.05, ## = p < 0.01). Immunofluorescence analysis of hyperoxic (upper panel) or physioxic (bottom panel) cultured hiPSCs for F) HIF1α (green) or G) hydroxy-(Pro564)- HIF1α (green), the actin-cytoskeleton by phalloidin (orange), the nuclei by DAPI (blue). Scale bar 20 μm; white arrow represents zoom-in area, scale bar 10 μm.